Optical waveguide element, chromatic dispersion compensator, methods for designing chromatic dispersion compensator, optical filter, methods for designing optical filter, optical resonator and methods for designing optical resonator

[0001] This is a Continuation Application of International Application No. PCT/JP2009/053731, filed on Feb. 27, 2009, which claims priority to Japanese Patent Application No. 2008-049841, filed Feb. 29, 2008. The contents of the aforementioned applications are incorporated herein by reference.

[0002] Description herein related to an optical waveguide element which has a Bragg grating pattern, and which can be used in various applications such as a chromatic dispersion compensator, an optical filter and an optical resonator and the like, and also related to a design method thereof.

[0003] The following are examples of chromatic dispersion compensation in an optical waveguide structure for which polarization dependence is not considered.

[0004] A dispersion compensator which has a Bragg grating pattern on the waveguide is disclosed in Patent document 1. The dispersion compensator disclosed in Patent document 1 has a plurality of Bragg grating elements in which the period changes spatially such that chromatic dispersion is compensated in a plurality of wavelength channels, and a refractive index distribution n (z) of the Bragg grating which is formed by a plurality of elements extending in the direction of the optical axis of the waveguide shows sinusoidal changes as shown in the following formula (wherein z is the position of a point on the light propagation axis).

[0005] $\begin{array}{cc}n\left(z\right)=n_{\mathrm{eff}}\left(z\right)+\sum _{i=1}^m\Delta n_i\left(z\right)\sin \left({\int }_0^z\frac{2\pi }{p_i\left(z^{\prime }\right)}d{z}^{\prime }+{\varphi }_i\right)& \left[\mathrm{Expression}1\right]\end{array}$

[0006] In a sine wave component which corresponds to the Bragg grating pattern of each wavelength channel, the pitch (local period) p_i gradually changes (i.e., chirps) together with z. In FIG. 3 of Patent document 1, the pitch chirps in a direction in which it decreases in response to increases in z. In addition, an origin phase φ_i changes discretely in each grating element i. As in the above described formula, the Bragg grating pattern which corresponds to each channel is defined independently, and a Bragg grating pattern is formed by superimposing these patterns. In Patent document 1, a case is illustrated in which a Bragg grating pattern is formed in an optical fiber.

[0007] In Patent document 2, a chromatic dispersion compensator is described in which a Bragg grating having one period is formed on the waveguide path, and a sampling structure is formed on the waveguide path which is superimposed on this Bragg grating, so that chromatic dispersion compensation is performed in a plurality of wavelength channels. The sampling structure is formed by a pattern that has undergone phase sampling in one period which is longer than the period of the Bragg grating. Each period of the phase sampling is divided into a plurality of spatial areas in a direction along the optical axis of the waveguide, and the phase of the Bragg grating changes discontinuously at a boundary where mutually adjacent spatial areas are in contact with each other. As is shown in FIG. 1A through FIG. 1D of Patent document 2, there are no discontinuous phase changes within a single spatial area.

[0008] In Patent document 3, a two-input and two-output light dispersion equalizer is described that performs chromatic dispersion compensation. The optical dispersion equalizer has a structure as a basic component element in which two optical waveguides are coupled by a plurality of directional couplers, the optical path lengths of two waveguides in a region sandwiched by two adjacent directional couplers are mutually different, and a phase controller is provided in at least one of the two waveguides. In this document, a device is illustrated that compensates a dispersion slope using these waveguides, and an element that compensates chromatic dispersion is provided in an optical input section. Furthermore, this document also described that the number of stages formed by connecting the aforementioned basic component elements in series is increased in order to enhance the compensation effect.

[0009] In Patent document 4, a design method of an optical signal processor is described in which a structure provided with a directional coupler having an amplitude coupling ratio ranging from a positive value to a negative value on one side of two waveguides having an optical path difference is used as a basic component element, and these basic component elements are combined in a series so as to form a two-input and two-output optical circuit with no feedback (namely, no reflection). In this design technique, the structure of the optical circuit is decided by expressing the characteristics of the optical circuit using a two-row two-column unitary matrix, imparting the desired output characteristics of the cross-port, and calculating amplitude parameters of the directional coupler in which the amplitude parameters are unknown parameters of the optical circuit. An example of the design of a chromatic dispersion compensator that is based on this design method is given in the Examples.

[0010] In Patent document 5, a broadband chromatic dispersion compensator that employs a high refractive index waveguide that uses photonic crystals is described, and in which chromatic dispersion compensation is performed by a transmission type of optical waveguide structure. The sign of the chromatic dispersion can be changed.

[0011] In Patent document 6, an optical reflector is described in which the optical waveguide has a structure in which the equivalent refractive index of the optical waveguide shows macroscopic monotone changes in the light propagation direction over a constant long-term period, while within this long-term period, the equivalent refractive index exhibits microscopic variations so as to follow the macroscopic monotone changes. This optical reflector uses grating with super periodic structure. A grating with super periodic structure indicates a type of sampled grating. In this optical reflector, when it is necessary to change the pitch of the grating in a chirped shape in order to achieve targeted optical characteristics, it is possible to obtain an equivalent effect by changing the width of the waveguide without changing the pitch of the grating itself.

[0012] In Patent document 7, a wavelength-variable element is described which comprises a first super periodic grating in which the reflection wavelength is variable, a second super periodic grating in which the reflection wavelength is variable and that has the same center wavelength as the first super periodic grating, and a ¼ period shifting structure that connects the first and second super periodic gratings together in a state in which these gratings are offset from each other by ¼ the grating period which corresponds to the center wavelength.

[0013] In Non-patent document 1, an actual fiber Bragg grating chromatic dispersion compensator is prepared using a design technique similar to that of Patent document 2, and the result of this is described. Firstly, a Bragg grating pattern of a single channel in a center wavelength is designed using the information in Non-patent document 2. The grating pattern is derived using an inverse scattering solution from the spectrum characteristics of the desired reflection and chromatic dispersion. However, in the fiber Bragg grating, because there are limits to the range over which the refractive index can be changed in order to manufacture a grating pattern, an operation in which the aforementioned spectrum characteristics are apodized by applying an inverse Fourier transform is also carried out so that these limits are not exceeded. As a result of this, a pattern is obtained in which the pitch of the Bragg grating changes continuously together with the position. Thereafter, Bragg grating patterns are designed using phase sampling for a plurality of channels. In a fiber Bragg grating, because there are limits on the range of refractive index change, phase sampling is effective.

[0014] In Non-patent document 2, an algorithm of a solution for the problem of inverse scattering which is based on layer peeling solution is described, and an example of the analysis of a chromatic dispersion compensator using a fiber Bragg grating is illustrated.

[0015] In Non-patent document 3, a chromatic dispersion compensator that is based on a chirped Bragg grating waveguide on a substrate is described. In this chromatic dispersion compensator, a rectangular optical waveguide core is formed by silver ion exchange on a silica glass substrate, and a Bragg grating pattern is formed in silica cladding on a top portion of the core. Because the grating pitch is gradually changed, the propagation axis of the core of the optical waveguide is bent. Laser light pulses having a wavelength of 800 nm are incident to the waveguide so that 58 ps/nm is obtained for an optical waveguide having a 7 mm grating length. Using a grating having a length of 50 mm, it is possible to perform chromatic dispersion compensation for an optical fiber equivalent to 50 km at a wavelength of 1550 nm.

[0016] [Patent document 1] U.S. Pat. No. 6,865,319

[0017] [Patent document 2] U.S. Pat. No. 6,707,967

[0018] [Patent document 3] Japanese Patent No. 3262312

[0019] [Patent document 4] Japanese Patent No. 3415267

[0020] [Patent document 5] Japanese Patent No. 3917170

[0021] [Patent document 6] Japanese Unexamined Patent Application, First Publication No. H6-313818

[0022] [Patent document 7] Japanese Unexamined Patent Application, First Publication No. H9-129958

[0023] [Non-patent document 1] “Phase-Only Sampled Fiber Bragg Gratings for High-Channel-Count Chromatic Dispersion Compensation” H. Li, Y. Sheng, Y. Li and J. E. Rothenberg, Journal of Lightwave Technology, Vol. 21, No. 9, September 2003, pp. 2074-2083

[0024] [Non-patent document 2] “An Efficient Inverse Scattering Algorithm for the Design of Nonuniform. Fiber Bragg Gratings” R. Faced, M. N. Zervas and M. A. Muriel, IEEE Journal of Quantum Electronics, Vol. 35, No. 8, 1999, pp. 1105-1115

[0025] [Non-patent document 3] “Integrated-Optic Dispersion Compensator that uses Chirped Gratings” C. J. Brooks, G. L. Vossler and K. A. Winick, Optics Letters, Vol. 20, No. 4, 1995, pp. 368-370

[0026] The amount of information that can be transmitted by optical communication has been constantly increasing. In response to this, measures such as (I) increasing the transmission rate of a signal, and (II) increasing the number of channels in wavelength division multiplexing communication are being promoted.

[0027] In optical communication, optical signals are transmitted using light pulses. Consequently, the following problems occur in the aforementioned (I). If the transmission rate increases, the time width of the light pulses is shortened, and the intervals between adjacent light pulses on a time axis becomes narrower. Because of this, it is crucial for the temporal waveform of the light pulses to be controlled. On a transmission line formed by an optical fiber, because of chromatic dispersion in which the propagation speed differs depending on the wavelength of the light, the time width of the light pulses becomes broader as they travel along the optical fiber. Because of this, technology for chromatic dispersion compensation is required in which au optical element having the opposite sign of chromatic dispersion against that of the optical fiber is provided on the optical fiber transmission line, and chromatic dispersion is eliminated for the light pulses after they are transmitted through the transmission line. The above described Patent documents 1 to 5 provide technology for a chromatic dispersion compensator for (I). In particular, in Patent documents 1 and 2, technology is described which relates to a multichannel chromatic dispersion compensator that deals with a plurality of channels of wavelength division multiplexing optical fiber communication.

[0028] On the other hand, measures to deal with (II) have the problems that, because the transmission line becomes more complex as the number of optical components increases, they lead to increased size, increased complexity, and increased costs for the optical communication equipment.

[0029] In order to avoid increases in the size and complexity of optical communication equipment, it is necessary to miniaturize the component elements such as the circuits and components of the devices that constitute the optical communication equipment, and to then integrate the miniaturized component elements so as to avoid any increase in the number of components. In order to achieve a miniaturization of the optical components, it is essential to miniaturize the optical elements that are the fundamental elements constituting the optical components. Optical elements that are used for optical communication are commonly fabricated using optical waveguides. Because of this, miniaturization of the optical waveguide is important to miniaturize the optical components. In order to miniaturize the optical waveguide, it is essential to use a material having a high refractive index such as silicon (Si) or the like. This is because, as the wavelength of light in a medium is inversely proportional to the refractive index of that medium, dimensions such as the core width of the optical waveguide become smaller as the refractive index becomes higher. The refractive index of Si is approximately 3.5, which is more than 2.3 times the refractive index of silica (SiO₂), which is approximately 1.5. Because the materials having a high refractive index such as Si and the like can be formed on a planar substrate, coupling of a plurality of optical waveguides becomes easier, and these materials are suitable for the task of integrating a plurality of optical components.

[0030] In order to avoid increased costs in optical components, it is important to reduce the manufacturing costs of the optical elements. If the optical waveguides are miniaturized, the cost of the raw materials needed for each optical element also decreases, and unit costs can be reduced. Because materials having a high refractive index such as Si and the like can be formed on a planar substrate, a large number of optical elements can be manufactured on one substrate by using a large area substrate, and the manufacturing costs can be reduced even further.

[0031] In order to obtain chromatic dispersion elements that are small in size and suitable for integration by patterning optical waveguides on a planar substrate using Si and Si-based high refractive index materials, in the design of the optical waveguide, it is necessary for the effective refractive index of the optical waveguide in a polarization state which is parallel to the substrate surface to be equivalent to the effective refractive index in a polarization state which is orthogonal to the substrate surface. This is because the cross-sectional configuration of the core of a high refractive index optical waveguide on the planar substrate is different from a circular core cross-section of a silica optical fiber. If the effective refractive index of the optical waveguide is different due to the polarization, then the chromatic dispersion generated in the optical waveguide is also changed by the polarization. In this case, the performance of the chromatic dispersion compensator ends up being affected by the polarization of the light pulses traveling along the optical fiber.

[0032] Solving the above described problems is difficult using conventional technology. A description will now be given of each element of the above mentioned conventional technology.

(1) Technology of Patent Document 1

[0033] In the technology described in Patent document 1, as an example of device manufacturing, there is only a description of a case in which a Bragg grating that uses an optical fiber is formed. Namely, the main subject of this technology is an optical fiber Bragg grating. The cross-section of an optical fiber is circular, and the optical characteristics thereof do not depend on the polarization direction of the propagated light. Accordingly, absolutely no mention is made about providing technology that relates to the design of an optical waveguide intended to reduce polarization dependence. In a design that considers polarization dependence, the effective refractive index is independently defined for each of polarization that is parallel to the substrate surface and polarization that is perpendicular to the substrate surface, and it is necessary to optimize the waveguide structure such that these two effective refractive indices coincide with each other. However, in this document, as is described above formula, only a single effective refractive index n (z) is defined irrespective of the polarization. Accordingly, it is not possible for the technology of this document to be applied to the design of chromatic dispersion compensators that are formed from high refractive index optical waveguides in which the polarization dependence on the substrate has been reduced.

[0034] Moreover, the design method of the chromatic dispersion compensator according to Patent document 1 follows a procedure in which the parameters in the formula such as the element length and amplitude of the effective refractive index are determined such that the chromatic dispersion characteristics obtained by simulation from that structure approach predetermined characteristics, while determining the profile of the effective refractive index pattern of the Bragg grating by the formula in advance. In this design method, a Bragg grating optical waveguide is constructed only by superimposing the Bragg grating patterns that correspond to each wavelength channel. Accordingly, the removal of interference between wavelength channels has not been considered in this design method, therefore the problem arises that the chromatic dispersion characteristics are deteriorated by the interference between wavelength channels. Furthermore, the procedure of this design method is a reverse flow of a procedure of a design method that specifies the effective refractive index pattern of a Bragg grating from predetermined element dimensions or optical characteristics. In order to achieve a miniaturization of the elements, it is essential for the element length to be decided in advance. However, this is not possible in the design method of Patent document 1.

(2) Technology of Patent Document 2

[0035] Patent document 2 did not mention a design in which polarization dependence is considered as not in Patent document 1. Accordingly, it is not possible for the technology of this document to be applied to the design of chromatic dispersion compensators that are formed from high refractive index optical waveguides in which the polarization dependence on a substrate is reduced.

[0036] In this document, a principle of designing a grating waveguide based mainly on phase sampling of the grating is employed. The reason for this is that this document is intended for low refractive index optical waveguides such as optical fibers whose refractive index is within a range of 1.4 to 1.5, and the effective refractive index of the optical waveguide which can be changed is restricted in a narrow range. Patent document 2 mentioned that the technology can also be applied to waveguides on a substrate. However, the technology of Patent document 2 is suitable only for the same type of waveguides as the low refractive index optical waveguides. Accordingly, the technology of Patent document 2 is not suitable for the purpose of attaining miniaturization by shortening the grating length as much as possible without reducing the reflectance by widely changing the effective refractive index in a reflective optical waveguide.

[0037] Furthermore, Patent document 2 disclosed that a phase sampling pattern is effective in avoiding performance deterioration due to voids when a grating structure is manufactured. The reason for this is that this document is concerned with manufacturing optical fiber gratings, and intended for a manufacturing method in which a grating pattern is printed on an optical fiber using ultraviolet irradiation. If it were intended for a high refractive index optical waveguide on a substrate, then it should be expected that there would be no restrictions such as performance deterioration due to the voids.

(3) Technology of Patent Document 3

[0038] Patent document 3 did not mention any technology to reduce polarization dependence. In the simple optical waveguide structure of this document, it is only possible to compensate a dispersion slope, and it is not possible to compensate chromatic dispersion. In order to compensate chromatic dispersion, it is necessary to form a structure in which another optical element is connected to the optical waveguide. Because of this, it is not possible to achieve miniaturization by using the technology of this document

(4) Technology of Patent Document 4

[0039] Patent document 4 did not mention any technology to reduce polarization dependence. In the chromatic dispersion compensator of this document, because the phase characteristics are anti-symmetrical to the point of origin, the chromatic dispersions in adjacent spectrum regions end up being inverted. Accordingly, this chromatic dispersion compensator can only be used for chromatic dispersion compensation that targets a particular limited spectrum region, namely, that targets specific spectrum region channels. This chromatic dispersion compensator cannot be used to compensate chromatic dispersion in a plurality of channels for the purpose of application to wavelength multiplexing optical fiber communication.

(5) Technology of Patent Document 5

[0040] With the technology of Patent document 5, chromatic dispersion can be compensated in a broad wavelength band. However, the technology of Patent document 5 cannot deal with multichannel chromatic dispersion. Because of this, the chromatic dispersion value thereof is not large. Accordingly, this technology cannot be used to compensate chromatic dispersion in a long distance (e.g., 40 km) optical fiber transmission line for the purpose of application to wavelength division multiplexing optical fiber communication.

(6) Technology of Patent Document 6

[0041] The technology of Patent document 6 proposes a structure in which, in order to obtain an optical reflector for a plurality of wavelength channels, instead of changing the pitch of the grating smoothly into a chirped shape, a number of uniform-pitch grating waveguides corresponding to the number of wavelength channels are connected together in series. Namely, a structure is proposed in which optical reflectors are connected in series in which the optical reflectors correspond to different wavelength channels. The pitches are different between gratings that correspond to different channels. In this type of series-connected grating, the following two problems arise.

[0042] (a) The waveguide length becomes longer in proportion to the number of wavelength channels. Because there is an increase in optical loss as the distance over which light is propagated on the waveguide increases, optical loss in wavelength channels located on the inward side when viewed from the light incident end is greater than that in channels located closer to the viewer. The more wavelength channels there are, the greater the difference in optical loss between channels becomes.
(b) Because the waveguide length becomes longer in proportion to the number of wavelength channels, the dimensions of the waveguide increase in proportion to the number of channels. Accordingly, it is not possible to reduce the size of an optical reflector that is used for wavelength division multiplexed transmissions.

[0043] Moreover, in the technology described in Patent document 6, by smoothly changing the width of the waveguide, equivalent optical characteristics are obtained to those obtained when the pitch is changed smoothly. The following problems are created by this.

[0044] (c) If the width of the waveguide increases, a plurality of horizontal modes are generated, and a multi-mode waveguide is formed. Accordingly, there is an upper limit to the waveguide width. Consequently, the number of channels to be multiplexed is limited by the waveguide width.
(d) Because any change in the waveguide width affects not only the reflection intensity, but also the wavelength of the reflection band, the deterioration in the reflection spectrum characteristics which is caused by fluctuations in the waveguide width occurring during manufacturing becomes more serious. Furthermore, in connecting portions between gratings that correspond to different wavelength channels, because changes in width are discontinuous and have a large amount of change, the following problems occur.
(e) The effects of fluctuations in the side walls of the waveguide increase in the connecting portions, independence between different channels is lost, the reflection spectrum bands between wavelength channels overlap and this causes cross-talk between channels.
(7) Technology of Patent Document 7

[0045] In the technology of Patent document 7, a super periodic grating is one in which a plurality of Bragg grating waveguides whose periods change gradually (i.e., chirp) are connected in series. In order for them to be used in wavelength division multiplexing communication, it is necessary for the wavelength to be changed between a plurality of wavelength channels. In this case, it is necessary to connect the same number of grating waveguides as there are channels, so that the problem arises that the length of the super periodic grating increases in proportion to the number of channels, so that a reduction in size cannot be achieved. At the same time as waveguide length increases, the problem of increasing optical loss also arises. Moreover, in the technology of Patent document 7, instead of a super periodic grating, it is also possible to use a sampled grating. However, in a sampled grating, because there are areas between each Bragg grating region where the grating amplitude is continuously zero, in order to increase the reflectance efficiency of the Bragg grating it is necessary to lengthen the waveguide length by an amount which corresponds to the areas where the grating amplitude is zero. As a result, a reduction in size cannot be achieved.

[0046] Moreover, in the technology of Patent document 7, a super periodic grating is created using a waveguide having a rectangular cross-section, however, there is no description relating to any design or structure that is intended to reduce polarization dependence. Accordingly, it is not possible to cancel out polarization dependence.

[0047] Moreover, in the technology of Patent document 7, a high refractive index optical waveguide which uses an InP-based compound which is typically used for optical communication semiconductor lasers is used. In this case, the waveguide is affected by roughness in the waveguide side walls which is generated in the manufacturing process.

(8) Technology of Non-Patent Document 1

[0048] The technology of Non-patent document 1 has the same problems as those of Patent document 2.

(9) Technology of Non-Patent Document 3

[0049] Although this is a Bragg grating optical waveguide formed on a planar substrate, a grating pattern is only formed in the cladding region on top of the optical waveguide core. Accordingly, the effective refractive indices of the respective optical waveguides are different for linearly polarized light in a direction parallel to the substrate surface and in a direction perpendicular to the substrate surface. Because of this, the chromatic dispersion performance differs greatly due to the polarization state. Experiments described in this document were performed using a Ti:sapphire laser as a light source. A Ti:sapphire laser normally emits linearly polarized light. In this document, there is no description of polarization states, and no consideration is given as to how to solve the problem of effective refractive index differences that are caused by disparities in polarization. Accordingly, it is not possible for the technology of this document to be applied to the design of chromatic dispersion compensators that are formed by high refractive index optical waveguides in which the polarization dependence in a substrate is reduced.

[0050] Moreover, conventionally, in waveguides having a Bragg grating pattern, a single pitch type of grating structure and a chirped type of grating structure are known. Namely, in the single pitch type of grating structure, as is shown in FIG. 62, the pitches p_G of protruding and recessed portions 201 and 202 which are provided in side walls of an optical waveguide 200 are uniform. In the chirped type of grating structure, as is shown in FIG. 63, the pitches of protruding and recessed portions 301 and 302 which are provided in side walls of an optical waveguide 300 gradually change (in FIG. 63, p_Gⁱ>p_G^j>p_G^k>p_G^l>p_G^m>p_G^m).

[0051] However, it is not possible to impart high-level functionality such as dispersion compensation (in particular, compensations for chromatic dispersion and dispersion slopes are performed simultaneously in a plurality of channels) in a simple grating structure such as that shown in FIG. 62.

[0052] Moreover, as is shown in FIG. 63, in a chirped type of grating structure, because each pitch is different, if too large a number of patterns are formed extending in the propagation direction of guided light, then it becomes difficult to control the fabrication accuracy of each pitch.

[0053] The exemplary embodiments described herein were conceived in view of the above described circumstances and it is an exemplary object thereof to provide an optical waveguide element that can achieve high-level functionality such as dispersion compensation, while also can simplify control of fabrication accuracy in the manufacturing process, and to a chromatic dispersion compensator that uses this optical waveguide element, and to a design method thereof.

[0054] In order to solve the above described problems, exemplary embodiments include the following.

[0055] Namely,

[0056] (1) An aspect of an exemplary embodiment is related to an optical waveguide element that comprises: a core of an optical waveguide; and a Bragg grating pattern that is provided on the core, wherein a pitch of the Bragg grating pattern takes a value from among three or more predetermined discrete values; the pitches that take the respective discrete values are present in a plurality of locations over an entire length of the optical waveguide respectively; and if a value from among all of the discrete values which has the highest distribution frequency is taken as M, and if the closest value to the M which is larger than the M is taken as A, and if the closest value to the M which is smaller than the M is taken as B, then a difference expressed as A−M is equal to a difference expressed as M−B.
(2) The Bragg grating pattern may comprise a plurality of isolated single coordinate points where a sign of a gradient of an envelope curve of an amplitude of the Bragg grating is inverted.
(3) The optical waveguide element may be a planar optical waveguide element in which the core of the optical waveguide is formed on a substrate; the core of the optical waveguide may comprise an inner side core which consist of a rib structure, and an outer side core which covers three directions of protruding portions of the rib structure on a top side of the inner side core; and the outer side core may be formed from a material having a lower refractive index than an average refractive index of the inner side core, and the Bragg grating pattern may be provided on the outer side core.
(4) The optical waveguide element may be a planar optical waveguide element in which the core of the optical waveguide is formed on a substrate; a gap portion that is formed from a material having a lower refractive index than that of the core may be provided in a center of a width direction of the core of optical waveguide so as to extend in a propagation direction of guided light; and the core may comprise two areas that are separated by the gap portion, and a single mode optical waveguide, in which a single mode is propagated so as to span across these two areas, may be formed.
(5) The optical waveguide element may be a planar optical waveguide element in which the core of the optical waveguide is formed on a substrate; a first Bragg grating pattern and second Bragg grating pattern, which are mutually parallel in the propagation direction of guided light, may be provided in the core; the first Bragg grating pattern may comprise recessed and protruding portions that are formed on both side walls of the core of the optical waveguide along the propagation direction of guided light; the second Bragg grating pattern may comprise a groove that is formed in the center of width direction of the core and is formed in a top portion in a perpendicular direction which is perpendicular to the propagation direction of guided light and also perpendicular to the width direction, and recessed and protruding portions may be formed in both side walls of the groove along the propagation direction of guided light; and when viewed in the propagation direction of guided light, portions of the first Bragg grating pattern where a core width is wide may correspond with portions of the second Bragg grating pattern where a groove width is narrow, and portions of the first Bragg grating pattern where the core width is narrow may correspond with portions of the second Bragg grating pattern where the groove width is wide.
(6) Another aspect of an exemplary embodiment is related to a chromatic dispersion compensator in which a chromatic dispersion and a dispersion slope in an optical transmission line are compensated by differing a distance over which signal light is propagated in the optical waveguide between entering the optical waveguide and being reflected in accordance with a wavelength, in a plurality of wavelength channels, wherein the chromatic dispersion compensator may comprise the above described optical waveguide element.
(7) Another aspect of an exemplary embodiment is related to a design method of the above described chromatic dispersion compensator which comprises a Bragg grating pattern design process, in which after a predetermined complex reflectance spectrum is calculated by specifying a chromatic dispersion, a dispersion slope, and a reflectance as parameters, a profile of an effective refractive index along the propagation direction of guided light of the optical waveguide is obtained from the complex reflectance spectrum and a desired optical waveguide length.
(8) Another aspect of an exemplary embodiment is related to a design method of the above described chromatic dispersion compensator, which comprises: an optical waveguide cross-sectional structure design process in which, by changing dimensions, in a cross-section which is perpendicular to the propagation direction of guided light, of two areas which form the first Bragg grating pattern and the second Bragg grating pattern, and thus equalizing effective refractive indices of the optical waveguide for two mutually independent polarizations that are guided on the optical waveguide, and by then determining this effective refractive index as a common effective refractive index for the two polarizations, a relationship between dimensions of the two areas and the common effective refractive index is obtained; a Bragg grating pattern design process in which, after a predetermined complex reflectance spectrum is calculated by specifying a chromatic dispersion, a dispersion slope, and a reflectance as parameters, a profile of an effective refractive index along the propagation direction of guided light of the optical waveguide is obtained from the complex reflectance spectrum and the desired optical waveguide length; and a chromatic dispersion compensator design process in which, by converting the profile of the effective refractive index obtained in the Bragg grating pattern design process into a profile of dimensions of the two areas based on the relationship between the dimensions of the two areas and the common effective refractive index obtained in the optical waveguide cross-sectional structure design process, the first Bragg grating pattern and the second Bragg grating pattern which are formed by the changes in the dimensions of the two areas are obtained.
(9) The Bragg grating pattern design process may further comprise a coarse graining process in which a resolution of discretization of a coordinate axis is taken as more than an amount of change in a pitch which corresponds to a half value of a width of a reflection band, in the other words, taken as more than the maximum value of the amount of change from the center value of the pitch in a chirped Bragg grating; and an optical waveguide may be created by the coarse graining process, which includes a plurality of isolated single coordinate points where a sign of a gradient of an envelope curve of an amplitude of a Bragg grating is inverted.
(10) Another aspect of an exemplary embodiment is related to an optical filter that may comprise the above described optical waveguide element.
(11) Another aspect of an exemplary embodiment is related to a design method of the above described optical filter, which may comprise a Bragg grating pattern design process in which, after a predetermined complex reflectance spectrum is calculated by specifying a reflectance and phase as parameters, a profile of an elective refractive index along the propagation direction of guided light of the optical waveguide is obtained from the complex reflectance spectrum and a desired optical waveguide length.
(12) Another aspect of an exemplary embodiment is related to a design method of the above described optical filter, which may comprise: an optical waveguide cross-sectional structure design process in which, by changing dimensions, in a cross-section which is perpendicular to the propagation direction of guided light, of two areas which form the first Bragg grating pattern and the second Bragg grating pattern, and thus equalizing effective refractive indices of the optical waveguide for two mutually independent polarizations that are guided on the optical waveguide, and by then determining this effective refractive index as a common effective refractive index for the two polarizations, a relationship between the dimensions of the two areas and the common effective refractive index is obtained; a Bragg grating pattern design process in which, after a predetermined complex reflectance spectrum is calculated by specifying a reflectance and phase as parameters, a profile of an effective refractive index along the propagation direction of guided light of the optical waveguide is obtained from the complex reflectance spectrum and a desired optical waveguide length; and an optical filter design process in which, by converting the profile of the effective refractive index obtained in the Bragg grating pattern design process into a profile of the dimensions of the two areas based on the relationship between the dimensions of the two areas and the common effective refractive index obtained in the optical waveguide cross-sectional structure design process, the first Bragg grating pattern and the second Bragg grating pattern which are formed by the changes in the dimensions of the two areas are obtained.
(13) The Bragg grating pattern design process may further comprise a coarse graining process in which a resolution of discretization of a coordinate axis is taken as more than an amount of change in a pitch which corresponds to a half value of a width of a reflection band; in the other words, taken as more than the maximum value of the amount of change from the center value of the pitch in a chirped Bragg grating; and an optical waveguide may be created by the coarse graining process, which includes a plurality of isolated single coordinate points where a sign of a gradient of an envelope curve of an amplitude of a Bragg grating is inverted.
(14) Another aspect of an exemplary embodiment is related to an optical resonator which may comprises: a first optical waveguide which forms a first reflection mirror; a second optical waveguide which forms a second reflection mirror: and a third optical waveguide which is sandwiched between the first optical waveguide and the second optical waveguide, wherein the first optical waveguide, the third optical waveguide, and the second optical waveguide may be connected in series, so that a single planar optical waveguide is formed; and wherein the first optical waveguide and the second optical waveguide may comprise the above described optical waveguide element.
(15) Another aspect of an exemplary embodiment is related to a design method of the above described optical resonator, which may comprise: a Bragg grating pattern design process in which, after a predetermined complex reflectance spectrum is calculated by specifying a reflectance and phase as parameters, a profile of an effective refractive index along the propagation direction of guided light of the optical waveguide is obtained from the complex reflectance spectrum and a desired optical waveguide length.
(16) Another aspect of an exemplary embodiment is related to a design method of the above described optical resonator, which may comprise: an optical waveguide cross-sectional structure design process in which, by changing dimensions, in a cross-section which is perpendicular to the propagation direction of guided light, of two areas which form the first Bragg grating pattern and the second Bragg grating pattern, and thus equalizing effective refractive indices of the optical waveguide for two mutually independent polarizations that are guided on the optical waveguide, and by then determining this effective refractive index as a common effective refractive index for the two polarizations, a relationship between the dimensions of the two areas and the common effective refractive index is obtained; a Bragg grating pattern design process in which, after a predetermined complex reflectance spectrum is calculated by specifying both a reflectance and phase as parameters, a profile of an effective refractive index along the propagation direction of guided light of the optical waveguide is obtained from the complex reflectance spectrum and a desired optical waveguide length; and a reflection mirror design process in which, by converting the profile of the effective refractive index obtained in the Bragg grating pattern design process into a profile of dimensions of the two areas based on the relationship between the dimensions of the two areas and the common effective refractive index obtained in the optical waveguide cross-sectional structure design process, the first Bragg grating pattern and the second Bragg grating pattern which are formed by the changes in the dimensions of the two areas are obtained.
(17) The Bragg gating pattern design process may further comprise a coarse graining process in which a resolution of discretization of a coordinate axis is taken as more than an amount of change in a pitch which corresponds to a half value of a width of a reflection band, in the other words, taken as more than the maximum value of the amount of change from the center value of the pitch in a chirped Bragg grating; and an optical waveguide is created by the coarse graining process, which comprises a plurality of isolated single coordinate points where a sign of a gradient of an envelope curve of an amplitude of a Bragg grating is inverted.

[0057] According to the exemplary aspect described in (1) above, compared with a conventional chirped-type grating in which the pitch changes gradually, tolerance control in the manufacturing process is easier and this contributes to an improvement in the manufacturing yield.

[0058] According to the exemplary aspect described in (2) above, it is possible to make the waveguide length shorter than in a sampled grating.

[0059] According to the exemplary aspect described in (3) above, because the confinement of light in an inner side core which is formed from a high refractive index material is weaker, compared to a conventional embedded type of optical waveguide with high relative refractive index difference which is only formed from two portions, namely, a core and a cladding which are formed from high refractive index materials, it is possible to inhibit the effects (i.e., scattering loss) on the optical characteristics from roughness on the inner-side core side walls, in which the roughness is unavoidably generated in the manufacturing process.

[0060] According to the exemplary aspect described in (4) above, because the mode field diameter of the fundamental mode expands, it is possible to inhibit the effects (i.e., scattering loss) on the optical characteristics from roughness on the inner-side core side walls which is unavoidably generated in the manufacturing process.

[0061] According to the exemplary aspect described in (5) above, in a planar optical waveguide element having a Bragg grating pattern, it is possible to reduce the polarization dependence of the optical characteristics.

[0062] According to the exemplary aspect described in (6) above, because a small-sized chromatic dispersion compensator that collectively compensates chromatic dispersion and dispersion slopes in a plurality of channels can be achieved by a Bragg grating structure having pitches (local period) to which only a limited small number of, namely, three or more different discrete values have been assigned, in the manufacturing process of an element or phase mask, control of the fabrication accuracy is easier and this contributes to an improvement in the manufacturing yield.

[0063] According to the exemplary aspects described in (7) and (8) above, it is possible to easily design a chromatic dispersion compensator in which a Bragg grating pattern has pitches (local period) to which only a limited small number of, namely, three or more different discrete values have been assigned.

[0064] According to the exemplary aspect described in (9) above, it is possible to design a chromatic dispersion compensator which is miniaturized even further.

[0065] According to the exemplary aspect described in (10) above, because a small-sized optical filter that collectively compensates chromatic dispersion and dispersion slopes in a plurality of channels can be achieved by a Bragg grating structure having pitches to which, only a limited small number of, namely, three or more different discrete values have been assigned, in the manufacturing process of an optical filter or phase mask, control of the fabrication accuracy is easier and this contributes to an improvement in the manufacturing yield.

[0066] According to the exemplary aspects described in (11) and (12) above, it is possible to easily design an optical filter in which a Bragg grating pattern has pitches to which only a limited small number of namely, three or more different discrete values have been assigned.

[0067] According to the exemplary aspect described in (13) above, it is possible to design an optical filter which is miniaturized even further.

[0068] According to the exemplary aspect described in (14) above, because a small-sized optical resonator that collectively compensates chromatic dispersion and dispersion slopes in a plurality of channels can be achieved by a Bragg grating structure having pitches to which only a limited small number of, namely, three or more different discrete values have been assigned, in the manufacturing process of an optical resonator or phase mask, control of the fabrication accuracy is easier and this contributes to an improvement in the manufacturing yield.

[0069] According to the exemplary aspects described in (15) and (16) above, it is possible to easily design an optical resonator in which a Bragg grating pattern has pitches to which only a limited small number of, namely, three or more different discrete values have been assigned.

[0070] According to the exemplary aspect described in (17) above, it is possible to design an optical resonator which is miniaturized even further.

• • [0145] M . . . Value having the highest distribution frequency (i.e., center value)
  • [0146] A . . . Value which is closest to the center value from among values larger than the center value
  • [0147] B . . . Value which is closest to the center value from among values smaller than the center value
  • [0148] 1, 10 . . . Core
  • [0149] 2, 12 . . . Side wall
  • [0150] 2a, 12a . . . Recessed portion (Narrow portion of core width)
  • [0151] 2b, 12b . . . Protruding portion (Wide portion of core width)
  • [0152] 13 . . . Groove
  • [0153] 13a . . . Recessed portion (Wide portion of groove width)
  • [0154] 13b . . . Protruding portion (Narrow portion of groove width)
  • [0155] 15 . . . Projection
  • [0156] 15a . . . Recessed portion (Narrow portion of projection width)
  • [0157] 15b . . . Protruding portion (Wide portion of projection width)
  • [0158] 20, 30 . . . Planar optical waveguide element
  • [0159] 21, 31 . . . First rib of inner side core
  • [0160] 22, 32 . . . Second rib of inner side core
  • [0161] 23 . . . Center gap
  • [0162] 24, 34 . . . Outer side core
  • [0163] 25, 35 . . . Substrate
  • [0164] 26, 36 . . . Bottom cladding
  • [0165] 27, 37 . . . Top cladding
  • [0166] 101 . . . Chromatic dispersion compensator
  • [0167] 102 . . . Optical circulator
  • [0168] 150 . . . Optical resonator
  • [0169] 151 . . . First optical waveguide (Reflection mirror)
  • [0170] 152 . . . Second optical waveguide (Reflection mirror)
  • [0171] 153 . . . Third optical waveguide

[0172] Hereinafter, the present invention will be described based on embodiments thereof.

[0173] A first exemplary embodiment of the optical waveguide element is shown in typical view in FIG. 1A through FIG. 1C. FIG. 1A is a perspective view of a portion of a core 1 of an optical waveguide, FIG. 1B is a plan view of the same portion of the core 1, and FIG. 1C is a cross-sectional view of an optical waveguide element. This optical waveguide element is a planar optical waveguide element in which an optical waveguide is formed on a substrate 5. The optical waveguide comprises a bottom cladding 6 which is formed on the substrate 5, a core 1 that is formed on the bottom cladding 6, and a top cladding 7 that is formed on the core 1 and on the bottom cladding 6. The grating structure comprises recessed portions 2a and protruding portions 2b that are formed in both side walls of the core 1 as periodical changes in the width w of the core 1. A top surface 3 and bottom surface 4 of the core 1 are flat.

[0174] In a high-function element such as a chromatic dispersion compensator or the like, in a Bragg grating pattern which is formed by repeating a single pitch such as that shown in FIG. 62, it is not possible to express desired characteristics. A conventional known technology developed in order to solve this problem is a chirped type of grating pattern such as that shown in FIG. 63. In a chirped type of grating pattern, as is described above, the pitch of the grating changes consecutively together with the position thereof in the propagation direction of guided light (i.e., in a left-right direction in FIG. 63). Accordingly, the dimensions of the pitch vary greatly as the optical waveguide length becomes longer, and the problem arises that it is difficult to control fabrication accuracy during the manufacturing process.

[0175] Control of the fabrication accuracy is performed, for example, by making measurements using the photoresist width and pitch image analysis after development as the DICD (Development Inspection Critical Dimensions) in a photolithographic process, and by making measurements using the structural object width and pitch image analysis after etching as the FWD (Final Inspection Critical Dimensions) in a etching process.

[0176] In this type of fabrication accuracy control, it is difficult to confirm whether an observed grating structure such as a chirped-type grating structure in which the pitch of the grating structural objects changes gradually is being machined to a predetermined pitch.

[0177] In contrast to this, in the optical waveguide element of the present embodiment, the amplitude of the Bragg grating is changed so that the Bragg grating pattern is formed as an amplitude modulation type. The amplitude modulation type of Bragg grating pattern of the present embodiment has the characteristic that the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted. Namely, in the present invention, the change when the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted is called amplitude modulation.

[0178] The sign inversions exhibit stepped precipitousness or discontinuousness that is generated at a single isolated coordinate point, and the feature of a sampled Bragg grating that a waveguide region is interposed where the amplitude changes continuously to zero between two points where the sign is inverted is not appeared. In the amplitude modulation-type grating of the present embodiment, because the amplitude becomes zero only at an isolated coordinate point where the sign of the gradient of the envelope curve is inverted, essentially, there are no regions where the amplitude is zero. Accordingly, it is possible for the waveguide length to be made shorter than a sampled Bragg grating.

[0179] A plurality of isolated coordinate points where the sign of the gradient of the envelope curve is inverted axe present on the waveguide. In each of these coordinate points, a discontinuous change in the phase is brought about incidentally. If the phase changes discontinuously the pitch (local period) changes. Consequently, the pitch takes on a value which is different from half the value obtained by dividing the center wavelength of the targeted spectrum at the relevant coordinate point by the average value n_av of the effective refractive index of the optical waveguide. The accuracy with which the coordinate point where the sign of the gradient of the envelope curve is inverted is specified depends on the interval of discretization of the coordinate z of the waveguide which is shown on the horizontal axis. If this interval is taken as ΔP, then the accuracy with which the coordinate point is specified is within a range of ±ΔP. In this manner, in the amplitude modulation-type Bragg grating of the present embodiment, the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted so that, as a result, coordinate points are present where the pitch changes discretely.

[0180] In the present exemplary embodiment, in this example and in all the other examples, the resolution of discretization of the coordinate z refers to the interval of discretization ΔP of the coordinate z.

[0181] Furthermore, in the optical waveguide element of the present embodiment, in the Bragg grating pattern, the pitches are each assigned only one of a limited small number of, namely, three or more different discrete values. Moreover, the respective discrete values are present in a plurality of locations over the entire length of the optical waveguide.

[0182] For example, in the example shown in FIG. 1B, there are only the three kinds of pitch: M, A, and B. Accordingly, because only a limited small number of, namely, three or more different discrete values are assigned to the pitches of an observed grating structure, it is easy to match together the measurement values and design values for an observation subject, and inspections of fabrication accuracy can be easily achieved.

[0183] If the number of discrete values is only one (i.e., single pitch) or two, then it is not possible to obtain the desired optical characteristics. Consequently; three or more different kinds are necessary. There is no particular restriction on the upper limit, however, an upper limit of several kinds or several tens of kinds may be considered as this provides the advantage that control of the fabrication accuracy in the manufacturing process is made easier. Moreover, it is preferable that the discrete value is an odd number such as 3, 5, 7, 9, . . . , 19, . . . , 49, . . . , 99, etc.

[0184] In the optical waveguide element of the present embodiment, if the value from among all the discrete values which has the highest distribution frequency is taken as M, the value which is closest to M from among values larger than M is taken as A, and the value which is closest to M from among values smaller than M is taken as B, then a difference shown by A−M is preferably equal to a difference shown by M−B. For example, if there are three kinds of discrete value, these correspond to M, M+α, and M−α (wherein α=A−M=M−B). Even if there are five kinds of discrete value, then it is preferable for at least three values, namely, the center value M and the two values A and B on either side thereof to have the above described relationship.

[0185] Furthermore, if there are five kinds of discrete value, then these are preferably M, M+α, M−α, M+β, and M−β (wherein α≠β). If there are seven kinds of discrete value, then these are preferably M, M+α, M−α, M+β, M−β, M+γ, and M−γ (wherein α, β, and γ are mutually different and are also smaller than M).

[0186] To generalize this, preferably, there are (2n+1) kinds of discrete value, and each discrete value (always being a positive value) is M, M+α₁, M−α₁, M+α₂, M−α₂, . . . , M+α_n, M−α_n (wherein α₁, α₂, . . . , α_n are mutually different and are also smaller than M). However; n is a limited small integer of 1 or more. Namely, it is preferable two particular discrete values on both side of the center value M to form a group, and the differences (the absolute values) between the center value and the two discrete values making up each group are equal.

[0187] By employing this type of structure, superior structure design results that enable the desired optical characteristics to be achieved can be obtained in addition, controlling the fabrication accuracy in the manufacturing process becomes even easier. Moreover, by varying the pitch of the width w of the core 1, high-level optical functions can be achieved.

[0188] An example of a method used to design a Bragg grating pattern comprises a method that specifying three parameters, namely, chromatic dispersion, dispersion slope, and reflectance, and then calculating a predetermined complex reflectance spectrum, and then the profile (shape distribution) of the effective refractive index extending in the waveguide direction of an optical waveguide having a Bragg grating is obtained from this complex reflectance spectrum and from the desired optical waveguide length.

[0189] A relationship between the effective refractive index n_eff and the waveguide width w for the optical waveguide element of the present embodiment is shown in FIG. 2. The thickness t of the core is 1.4 μm. Mode 1 and mode 2 correspond respectively to a TE mode and a TM mode. In the present embodiment, the TE mode is used as the subject, however, the same procedure can be applied for the TM mode as well. An example of the method used to obtain the correspondence relationship between the effective refractive index n_eff and the waveguide width w is a method in which the field distribution of the horizontal mode within the waveguide core is calculated using a film mode matching method (i.e., an FFM method), a finite element method, or a beam propagation method, and an effective refractive index for the eigen mode is then determined.

[0190] The example shown in FIGS. 1A through 1C is an example of a planar optical waveguide element, however, the technology ideas of the present invention can also be applied to an optical fiber Bragg grating. In an optical fiber Bragg grating, the grating is formed by a plurality of rising refractive index regions. Unlike a planar optical waveguide element, there is no direct control of the width and pitch of the rising refractive index regions during the manufacturing process, however, in the process to manufacture a phase shift mask which is used when the optical fiber Bragg grating is being manufactured, it is necessary to control the width and pitch of those portions of the mask that correspond to the rising refractive index regions. Accordingly, the above described technology ideas of the present invention have significance in an optical fiber Bragg grating as well.

[0191] An example of an optical waveguide element of the second embodiment of the present invention is shown in FIG. 3A through FIG. 3C. The optical waveguide element is a planar optical waveguide element in which an optical waveguide is located on the substrate. If the width or thickness of the waveguide in the light propagation direction is changed periodically in an optical waveguide, then the effective refractive index of the optical waveguide also changes periodically, and a Bragg grating can be constructed. In FIG. 3A through FIG. 3C, only the core 10 is shown and the claddings are not shown, however, it is to be assumed that cladding surrounds the periphery of the core 10. In addition, a substrate (not shown) is located below the cladding, and a bottom surface 14 of the core; 10 is parallel with the substrate surface. The term “horizontal direction” refers to a direction which is parallel to this substrate surface, and the term “vertical direction” refers to a direction which is perpendicular to the substrate surface.

[0192] FIG. 3A is a plan view of a portion of the core 10. The symbol C represents a single center axis within a horizontal plane of the optical waveguide core 10, and light is propagated along this center axis C within the optical waveguide. This optical waveguide has a Bragg grating pattern (described below in detail), and at least one reflection band is evident in the spectrum of this optical waveguide. A center wavelength λ₀ of this reflection band is determined by λ₀=2 p_G/n_eff when the period of the Bragg grating is taken as p_G, and the effective refractive index of the optical waveguide is taken as n_eff. Here, the effective refractive index n_eff is the value when the width of the core 10 of the optical waveguide is taken as an average width w₀.

[0193] The average width w₀ of the core 10 is equivalent to the average value in one period of the horizontal width w_out of the core 10, and is a constant value (same value) along the center axis C over the entire optical waveguide. Recessed portions 12a and protruding portions 12b are formed alternatingly in a side wall 12 of the core 10, and the horizontal width w_out oscillates alternatingly for each period p_G so as to form a first Bragg grating pattern. This Bragg grating pattern is regarded as one in which the width in the horizontal direction (namely, the horizontal width w_out) of an optical wave guide having a rectangular cross-section (see FIG. 3B) changes alternatively.

[0194] Eigen waveguide modes exist respectively in both cases when an electric field of linearly polarized light extends principally in a horizontal direction (referred to below as TE-type polarization), and when it extends principally in a vertical direction (referred to below as TM-type polarization) in an optical waveguide having a rectangular cross-section. In addition, polarization dependence that eigen (intrinsic) effective refractive indices are present in each waveguide mode also exists.

[0195] An effective refractive index n_eff^TE of the eigen mode in TE-type polarization changes more sensitively to changes in the width of the optical waveguide compared to an effective refractive index n_eff^TM of the eigen mode in TM-type polarization. In contrast, the effective refractive index n_eff^TM of the eigen mode in TM-type polarization changes more sensitively to changes in the height (namely, the thickness) of the optical waveguide compared to the effective refractive index n_eff^TE of the eigen mode in TE-type polarization.

[0196] Accordingly, as is shown in FIG. 1A through FIG. 1C, if no Bragg grating pattern is provided on the top surface 3 of the optical waveguide core 1, and the recessed portions and protruding portions 2a and 2b are provided in the side wall 2 so that only the width of the core 1 is changed periodically, then there is a considerable increase in polarization dependence. Accordingly, in order to reduce the polarization dependence of the Bragg grating, not only is the width of the optical waveguide periodically changed, but to also the height of the optical waveguide is periodically changed.

[0197] Because of this, in this planar optical waveguide element, two Bragg grating patterns are positioned in mutually different areas on a cross-section that is perpendicular to the propagation direction of guided light. The features, such as the pitches (local period) are each assigned only one of a limited small number of, namely, three or more discrete values, of the pitch of this Bragg grating pattern is the same as that of the first embodiment.

[0198] In addition, two Bragg grating patterns are formed in areas that extend in parallel with the propagation direction of guided light. Namely, the ranges in which the respective Bragg grating patterns are located extending in the direction of the center axis C are the same.

[0199] As a result, by combining a first Bragg grating pattern with a second Bragg grating pattern, the effect thereof on TE-type polarization and the effect thereof on TM-type polarization are equalized, and polarization dependence can be reduced.

[0200] If the application of this to a rectangular optical waveguide (i.e., an optical waveguide having a substantially rectangular cross-section) is considered, then the first Bragg grating pattern may be provided on either one or both side walls of the core, and the second Bragg grating pattern may be provided on at least one of the top surface and bottom surface (on the top surface and/or the bottom surface) of the core. In the present embodiment, in order to simplify the fabrication of the core on a substrate, the first Bragg grating pattern is provided on both side walls of the core, and the second Bragg grating pattern is provided on the top surface of the core. The shape of the core 10 is symmetrical (i.e., symmetrical above and below the center axis C in FIG. 3A) in a horizontal direction relative to a plane that extends in the vertical direction and includes the center axis C.

[0201] In order to form an optical waveguide having a Bragg grating on a substrate, the following manufacturing process is followed.

[0202] Firstly, a material used to form the bottom cladding is formed (coated) on one surface of the substrate as a film. Next, a material used to form the core is formed on the bottom cladding as a film, and this film is then processed into the shape of the Bragg grating. Thereafter, a material used to form the top cladding is formed on the bottom cladding and the core as a film, so that when viewed in cross-section, the core is enclosed by a bottom cladding and a top cladding.

[0203] As is described below, the amplitude and period of periodic variations in the Bragg grating become non-uniform in order to provide chromatic dispersion compensation in a plurality of wavelength channels. Accordingly, a fabrication process may be performed on the core to form it into a shape that is able to deal with this type of non-uniform periodic variations. Fabrication process of the core in the width is achieved by etching and lithography (pattern drawing) using an optical mask that contains a grating pattern (i.e., a periodic variation of the horizontal width w_out) that provides chromatic dispersion compensation for a plurality of channels.

[0204] In contrast, it is difficult for the depth of etching which is intended to form the height of the core to be changed in accordance with the Bragg grating pattern. Namely, it is necessary to achieve a periodic variation in etching depth in order to form a grating pattern (i.e., a periodic variation in the core height) on the top portion of the core by etching. However, if depth variations in etching within a horizontal plane extending along the substrate surface which are non-uniform and non-controllable are ignored, then etching depths are substantially uniform (same) under the same conditions. Accordingly, it is difficult to forming process the height of a core in accordance with a Bragg grating pattern.

[0205] FIG. 3B shows a cross-section of a core in a plane that intersects the center axis C. In the core 10 of the present embodiment, instead of changing the height of the core, as is shown in FIGS. 3A through 3C, the width w_in of a groove 13 provided in a top portion of the core is periodically changed. The core height is t_out, and the depth of the grooves 13 is t_in. As is shown in FIG. 3A, the groove 13 extends in a direction along the center axis C, and coordinates in the horizontal direction of the center point of the width w_in of the groove 13 are positioned on the center axis C.

[0206] As a result of this, it is possible to make changes to the effective refractive index that is the equivalent of periodically changing the height of the core 10. Recessed portions 13a and protruding portions 13b are formed alternatively in side walls of the groove 13, and the groove width w_in oscillates alternatively each one period p_G so as to form a second Bragg grating pattern. Because the depth t_in is constant (uniform) within the groove 13, it is possible to achieve a groove 13 having a periodic variation in the width by using etching and lithography using an optical mask.

[0207] According to this type of method, by performing fabrication process on the width of a groove provided in a top portion of the core in the same way as the core width w_out, it is possible to construct an optical waveguide having a Bragg grating in both a width direction and height direction. Accordingly, it is possible to reduce polarization dependence by matching changes in the effective refractive index which are generated by the first Bragg grating pattern in the width direction with changes in the effective refractive index which are generated by the second Bragg grating pattern in the height direction.

[0208] In the structure shown in FIGS. 3A through 3C, in the propagation direction of guided light, the wide portions (i.e., the protruding portions 12b) of the core width w_out of the side walls 12 correspond to the narrow portions (i.e., the protruding portions 13b) of the groove width w_in of the inner side walls of the grooves 13, and the narrow portions (i.e., the recessed portions 12a) of the core width w_out of the side walls 12 correspond to the wide portions (i.e., the recessed portions 13a) of the groove width of the inner side walls of the groove 13. In this manner, the protrusions and recesses of the first Bragg grating pattern are synchronized with the protrusions and recesses of the second Bragg grating pattern, so that the pitches (local period) p_c of each one coincide with each other (same). Employing this simplifies the designing of the optical waveguide dimensions.

[0209] The effective refractive indices calculated for the TE polarization and the TM polarization will now be described based on a cross-sectional structure corresponding to the Bragg grating optical waveguide shown in FIGS. 3A through 3C. The calculations of the effective refractive indices can be performed using a film mode matching method (FMM method), a finite element method, or a beam propagation method or the like.

[0210] The thickness of the claddings positioned below and above the core is 2 μm for both. t_in is 0.1 μm and t_out is 1.4 μm. Mode 1 shows TE polarized light while mode 2 shows TM polarized light. As is shown in FIG. 4B, if the relationship between w_in and w_out is provided, then, as is shown in FIG. 4A, it is possible to reduce the polarization dependence of the effective refractive index of the waveguide. If the effective refractive index of the TE polarization is regarded as the effective refractive index of the waveguide, then a relationship between the effective refractive index and w_in and w_out is calculated and is plotted on the graph shown in FIG. 5.

[0211] In order to generate changes that are equivalent to changes in the height of the core, it is also possible to provide projections (ridges) 15, as is shown in FIGS. 6A through 6C, instead of the grooves 13 as the structure which is provided on the core top portion. The grooves 13 are preferable from the standpoint of ease of controlling the effective refractive index, however, if restrictions on materials or processing conditions or the like exist, then it is also possible to select the projections 15. The projections 15 can be constructed by forming another layer of film from the same material which is used to form the core, and then forming periodic variations in the width direction using optical lithography and etching.

[0212] In the structure shown in FIGS. 6A through 6C, in the propagation direction of guided light, the wide portions (i.e., the protruding portions 12b) of the core width w_out of the side walls 12 correspond with the wide portions (i.e., the protruding portions 15b) of the width w_in of the projections 15, and the narrow portions (i.e., the recessed portions 12a) of the core width w_out of the side walls 12 correspond with the narrow portions (i.e., the recessed portions 15a) of the width w_in of the projections 15. In this manner, the period of the protrusions and recessed portions of the first Bragg grating pattern is synchronized with the period of the protrusions and recessed portions of the second Bragg grating pattern, so that the pitches p_u of each coincide. Employing this method is preferable as it simplifies the designing of the optical waveguide dimensions.

[0213] The grooves 13 and/or the projections 15 are preferably formed in the center in the width direction of the core 10 and on the top portion in the vertical direction thereof. In this case, the coordinates in the horizontal direction of the center of the width w_in of the grooves 13 and/or projections 15 are located on the center axis C of the core 10. Moreover, it is also possible to form the second Bragg grating pattern by combining grooves with projections.

[0214] The grooves 13 and projections 15 shown in FIGS. 3A through 3C and FIGS. 6A through 6C are connected in the propagation direction of guided light, however, it is also possible to provide a periodic change in the height direction of the core by forming recessed portions and/or protruding portions in each pitch. The grooves 13 and projections 15 that are formed on the top surface 11 of the core 10 are formed in a portion of the center in the width direction of the core 10, however, it is also possible for the thickness of the core 10 itself to be changed.

[0215] Among these structures, it is preferable from the standpoint of the manufacturing process to provide a periodic change in the height direction of the core using changes in the width direction of the structural elements of the core top portion, in particular, as is shown in FIGS. 3A through 3C and FIGS. 6A through 6C, it is preferable for the second Bragg grating pattern to be formed by the projections 15 and/or the grooves 13 which are formed in the center in the width direction of the core 10 and on the top portion in the vertical direction thereof. It is preferable because the grooves 13 can be obtained simply by forming a single layer of film of the material used to form the core.

[0216] An optical wave guide having a cross-sectional structure such as that shown in FIG. 7 can be given as an example of a Bragg grating optical waveguide structure in which polarization dependence has been reduced. On account of the simplified description of the principle of reducing polarization dependence, the cross-sectional structures of the cores 10 are all the same in, the planar optical waveguide elements of FIGS. 3A through 3C and FIGS. 6A through 6C. However, if the effective refractive index is changed by changing the dimensions of the optical waveguide, then an optical waveguide having a composite core structure such as that shown in FIG. 7 is preferable in order to improve the accuracy of the effective refractive index.

[0217] The core of the planar optical waveguide element 20 having the cross-sectional structure shown in FIG. 7 is a composite core formed by two areas, namely, inner side cores 21 and 22 and an outer side core 24.

[0218] In this example, the inner side core comprises two areas, namely, a first rib 21 and a second rib 22, and a center gap 23 is provided between these two ribs. The first rib 21 and second rib 22 are formed from a material having a higher refractive index than that of the outer side core 24. It is not necessary for the center gap 23 to be formed from a material having a higher refractive index than that of the outer side core 24. The heights of the first rib 21 and second rib 22 and of the center gap 23 are equal, and are indicated by t₁+t₂ in FIG. 7. If the center gap is inserted between the first rib 21 and second rib 22, then it is possible, while maintaining the condition in which only a single mode is present in a single polarization state, to increase the cross-sectional area of the region in which light is confined in the inner side core. Moreover, because it is possible to reduce the accuracy deterioration of the effective refractive index which is due to fabrication errors of the Bragg grating which is formed on the outer side core 24 (described below), this example is also effective in reducing the polarization dependence of the effective refractive index.

[0219] The first rib 21 and second rib 22 each have the same shape, and are shaped so as to be the reverse of each other in a horizontal direction. Specifically, the first rib 21 and second rib 22 are formed by planar portions 21a and 22a which have a thickness of t₂, and by rectangular parallelepiped portions 21b and 22b which have a height t₁ and a width w₁ and which are positioned on edges of the planar portions 21a and 22a. The material used to form the rectangular parallelepiped portions 21b and 22b is the same as the material used to form the planar portions 21a and 22a. The width of the center gap 23 is w₂, and the center gap 23 is formed from a material having a lower refractive index than that of the first rib 21 and second rib 22.

[0220] Examples of t₁, t₂, w₁, and w₂ are: t₁=250 nm, t₂=50 nm, w₁=280 nm, and w₂=160 nm, however, the present invention is not limited to these dimensions.

[0221] One example is a combination in which the first rib 21 and second rib 22 are formed from silicon (Si), and the center gap 23 is formed from silica (SiO₂). Instead of forming the center gap 23 from silica (i.e., silicon oxide), it may also be formed from silicon oxynitride (SiO_xN_y) or silicon nitride (Si_xN_y). Materials in which the composition ratio x:y is controlled such that, for example, in SiO_xN_y the refractive index is 1.5, while in Si_xN_y the refractive index is 2.0 may be used, however, other composition ratios may also be used provided that they have a lower refractive index than that of the Si of the high refractive index ribs 21 and 22.

[0222] P-conductivity or N-conductivity may be imparted to the first rib 21 or the second rib 22 by doping the medium with suitable impurities. Namely, the first rib 21 may be set as a P-type area, while the second rib 22 may be set as an N-type area. Conversely, the first rib 21 may be set as an N-type area, while the second rib 22 may be set as a P-type area.

[0223] The impurities (i.e., dopants) that impart conductivity to the high refractive index core which is formed by a semiconductor can be appropriately selected in accordance with the host medium. For example, if the host medium is a group-IV semiconductor such as silicon, then a group-III element such as boron (B) can be used as a dopant to impart P-type polarity, while a group-V element such as phosphorus (P) or arsenic (As) can be used as a dopant to impart N-type polarity.

[0224] In this manner, by making the first rib 21 and second rib 22 of the core high refractive index ribs which are formed from a semiconductor such as Si or the like, and by making one of these ribs a P-type and the other one a N-type and mutually separating these by the center gap 23 which is formed from an insulating material, a P-I-N bond is formed within a plane represented by the thickness t₂ of the inner side cores 21 and 22. In addition, by providing electrode pads through which electric voltage is applied to both the first rib 21 and the second rib 22, and thereby generating an electric potential difference between the two ribs 21 and 22, it is possible to induce a change in the refractive index which is caused by changes in the carrier density, and to thereby make it possible to vary the optical characteristics of an electrode element. Moreover, by providing the center gap 23 which is formed from an insulating material between the two ribs 21 and 22 that make up P-type and N-type areas, the effect is obtained that it is possible to prevent leakage current between the P-type and the N-type areas, and it is also possible to considerably reduce the current consumption. Specifically, when voltage of several V was applied between the two ribs in a structure in which there was no center gap, a sub-milliamp (sub-mA) current flowed between the P-type and N-type areas. In contrast to this, when a center gap was provided, even when 30 to 40 V of voltage was applied, the leakage current between the P-type and N-type areas are only a sub-nanoamp (sub-nA) range.

[0225] Note that the imparting of conductivities of opposite polarities (i.e., P-type or N-type) to the first rib 21 and the second rib 22, and also the provision of the electrode pad to impart voltage are not essential structure in the present embodiment, and it is also possible to use the inner side cores 21 and 22 without imparting any external voltage thereto.

[0226] The outer side core 24 is placed on top of the inner side cores 21 and 22. The refractive index of the outer side core 24 is lower than the average refractive index of the inner side cores 21 and 22. Examples of the material used for the outer side core 24 include Si_xN_y, however, other materials may also be used. Although omitted from FIG. 7, the same first and second Bragg grating patterns as those formed on the core 10 shown in FIGS. 3A through 3C are formed on a top surface 24a and side walls 24b of the outer side core 24. The features, such as the pitches (local period) are each assigned only one of a limited small number of; namely, three or more discrete values, of the pitch of this Bragg grating pattern is the same as that of the first embodiment.

[0227] Specifically, there are provided a first Bragg grating pattern in which the width w_out of the outer side core 24 is changed periodically, and a second Bragg grating pattern in which the width w_in of a groove 24c which is formed on the top surface 24a of the outer side core 24 is changed periodically. The thickness of the outer side core 24 is t_out, and the depth of the groove 24c is t_in.

[0228] Examples of t_out and t_in are: t_out=600 nm, and t_in=100 nm, however, the present invention is not particularly limited to these dimensions. w_in and w_out change periodically.

[0229] Note that, in the example shown in FIG. 7, the Bragg grating pattern on the top surface 24a is formed by the groove 24c, however, as is described above it is also possible to employ the projections 15 (see FIGS. 6A through 6C).

[0230] The above described composite core is located on the bottom cladding 26 which is formed as a film on the substrate 25. The top portion and side walls of the composite core are covered by the top cladding 27. The top cladding 27 and bottom cladding 26 are formed from a material having a lower refractive index than an average refractive index of the composite core. The same material or different materials may be used for the top cladding 27 and the bottom cladding 26. Specific examples thereof include using Si for the material of the substrate 25, and using SiO₂ as the material for the top cladding 27 and bottom cladding 26, however, the present invention is not particularly limited to this. The top cladding 27 and bottom cladding 26 may have a suitable thickness to correspond to the thickness of the composite core. For example, if the dimensions of the composite core are as described above, then the thickness of the bottom cladding 26 is approximately 2000 nm, while the maximum thickness of the top cladding 27 (i.e., the thickness on the top of the planar portions 21a and 22a) is approximately 2000 nm.

[0231] If the average refractive index (i.e., the overall average refractive index obtained from the two ribs and the center gap combined) of the inner side core is higher than the average refractive index of the outer side core 24, then when light is guided through the composite core, because a greater electric field is present within the inner side core, the proportion of the effective refractive index that is changed by w_out and w_in is reduced compared with when the core is uniform. Accordingly, even if there are errors in the fabrication dimensions of the Bragg grating pattern formed on the outer side core, any effects of this on the effective refractive index are minimal. Accordingly, it is possible to increase the accuracy of the effective refractive index. In microfabrication on a planar substrate, it is generally necessary to consider errors of approximately 10 nm. According to a composite core such as that shown in FIG. 7, it is possible to hold the effects on errors in the effective refractive index which are due to the fabrication accuracy to less than 80 ppm of the average value of the effective refractive index. Note that this term “average value of the effective refractive index” indicates the effective refractive index of the optical waveguide at the average width w₀, as is shown in FIG. 3A.

[0232] The effective refractive indices calculated for the TE polarization and the TM polarization will now be described based on a cross-sectional structure corresponding to the Bragg grating optical waveguide shown in FIG. 7.

[0233] Note that the specific example given below corresponds to Step 1 of the following [Example 1 of a chromatic dispersion compensator].

[0234] In an optical waveguide structure having the composite core shown in FIG. 7, the first rib 21 and second rib 22 are formed from Si, the center gap 23 is formed from SiO₂, the outer side core 24 is formed from Si_xN_y, the substrate 25 is formed from Si, the bottom cladding 26 is formed from SiO₂, and the top cladding 27 is formed from SiO₂. In addition, t₁=250 nm, t₂=50 nm, w₁=280 nm, w₂=160 nm, t_out=600 nm, and t_in=100 nm, and the thickness of the bottom cladding 26 is 2000 μm, and the maximum thickness of the top cladding 27 is 2000 nm.

[0235] Moreover, w_out and w_in are specified, and the field distribution of the eigen mode is calculated using a film mode matching method (i.e., an PPM method), a finite element method, or a beam propagation method, and an effective refractive index that corresponds to the eigen mode is determined. As a result of this, the correspondence relationship between w_out and w_in and the effective refractive index is obtained.

[0236] A portion of this result is shown in FIG. 8A and FIG. 8B. FIG. 8A is a graph showing changes in the effective refractive index relative to w_in, while FIG. 8B is a graph showing changes in w_out which accompanies the changes in w_in. Here, w_in and w_out are changed simultaneously. Mode 1 shown in FIG. 8A is TE-type polarization (with a polarization level of 98% or more), and mode 2 is TM-type polarization (with a polarization level of 97% or more). According to FIG. 8A, the difference between n_eff^TE and n_eff^TM is 20 ppm or less, which is smaller than the amount of change in the effective refractive index caused by fabrication errors. Accordingly, polarization dependence can be ignored.

[0237] As is shown in FIG. 8B, if the relationship between w_in and w_out is provided, then, as is shown in FIG. 8A, it is possible to reduce the polarization dependence of the effective refractive index of the waveguide. If the effective refractive index of the TE polarization is regarded as the effective refractive index of the waveguide, then a relationship between the effective refractive index and w_in and w_out is calculated and is plotted on the graph in FIG. 9.

[0238] In the present embodiment, the average value of n_eff is taken as 2.3480. The horizontal axis is set to n_eff, the left vertical axis is set to w_in, and the right vertical axis is set to w_out. A graph showing this relationship of w_in and w_out relative to n_eff is shown in FIG. 9. Using these settings, if the n_eff of a particular position is supplied, w_in and w_out can be determined, and the cross-sectional structure of the optical waveguide in that position is determined.

[0239] Next, a description will be given of a procedure introduced for the first time by the present exemplary embodiment for the design of a chromatic dispersion compensator using a Bragg grating optical waveguide having reduced polarization dependence, based on the waveguide shown on the third embodiment of the optical waveguide element. The respective steps given below provide an itemized outline of the design flow of this procedure.

[0240] [1] The dimensions of the cross-sectional structure of an optical waveguide having reduced polarization dependence are specified, and the field distribution of the eigen mode of the TE-type polarization and TM-type polarization in the cross-section are calculated. The effective refractive index is calculated from the field distribution of each eigen mode, and correspondence relationships between these effective refractive indices and the optical waveguide dimensions which are used to determine the cross-sectional structure are obtained from the effective refractive indices.
[2] The desired chromatic dispersion characteristics and reflection characteristics are specified, and the necessary data required for determine the structure of the optical waveguide is prepared.
[3] The optical waveguide length is provided, and the shape distribution (profile) of the effective refractive index in a direction extending along the center axis C of the optical waveguide is calculated from the chromatic dispersion characteristics and reflection characteristics of the above described [2] using an inverse scattering problems solution method.
[4] Based on the correspondence relationships between the effective refractive indices and the optical waveguide dimensions obtained in the above described [1], the shape (i.e., the profile of the optical waveguide dimensions in a direction extending along the center axis C of the optical waveguide) of the Bragg grating optical waveguide is decided from the shape distribution of the effective refractive index obtained in [3].

[0241] The step [1] through [4] will now be described.

[0242] Note that because step [1] only needs to be completed before step [4], it is of course possible for the steps to be performed in the sequence [1]→[2]→[3]→[4], or [2]→[3]→>[1]→>[4], or [2]→[1]→[3]→[4], or for [1], [2], and [3] to each be performed in parallel.

[0243] Namely, this design method includes an optical waveguide cross-sectional structure design process (a) which comprises step [1], a Bragg grating pattern design process (b) which comprises steps [2] and [3], and a chromatic dispersion compensator design process which comprises step [4], and the sequence in which process (a) and process (b) are performed is not restricted.

[0244] Note that the design method of exemplary embodiments can also be applied to optical waveguide structures having a uniform core such as those in FIGS. 3A through 3C and FIGS. 6A through 6C.

(Step [1])

[0245] In the case of the present embodiment, the term “optical waveguide dimensions for determining the cross-sectional structure” refers to w_out for a first Bragg grating pattern which consists of recessed portions and protruding portion formed in the side wall 24h, and to w_in for a second Bragg grating pattern which is formed in the groove 24c in the top surface 24a. Therefore, w_out and w_in are specified, and the field distribution of the eigen mode is calculated using a film mode matching method (i.e., an FMM method), a finite element method, or a beam propagation method, and an effective refractive index that corresponds to the eigen mode is determined. As a result of this, the correspondence relationship between w_out and w_in and the effective refractive index is obtained.

[0246] In the case of the present embodiment, step 1 is the same as the example shown in the above described [second embodiment of the optical waveguide element]. The same is performed as shown in FIGS. 8A and 8B, and FIG. 9, and then the effective refractive index is calculated. Hereinafter, the effective refractive index of mode 1 is taken as effective refractive index n_eff of this optical waveguide.

[0247] Note that, here, it is assumed that the effective refractive index of mode 1 is used in the following design process. However, because the difference between n_eff^TE and n_eff^TM is less than an error, instead, the effective refractive index of mode 2 may be taken as the effective refractive index n_eff of the optical waveguide, or the average of the effective refractive indices of both mode 1 and mode 2 may be taken as the effective refractive index n_eff of the optical waveguide.

(Step [2])

[0248] The chromatic dispersion characteristics required in the chromatic dispersion compensator cancel out (counteract) the chromatic dispersion of the intended optical fiber transmission line, and have the opposite sign against the chromatic dispersion of the optical fiber transmission line, and have an equivalent absolute value. In the present example, the wavelength band of optical signals being transmitted is in the L band region (1566.31 to 1612.65 mm), and the optical transmission line is formed by a dispersion shifted fiber (G653) having a length of 40 km, so that the chromatic dispersion to be assigned to the chromatic dispersion compensator is stipulated. Moreover, in the intended optical transmission line, it is to be assumed that optical signals of 50 channels of an L band ITU grid in which the frequency interval is 100 GHz (which is approximately 0.84 nm if converted into a wavelength interval) are being transmitted. The bit rate of the transmitted optical signals is 40 Gbit/s, and the bandwidth usage of each channel is 80 GHz. Outside this bandwidth usage, the delay time is stipulated as a fixed value.

[0249] A dispersion shifted fiber exhibits anomalous dispersion in the L band, and the group delay time increases as the wavelength lengthens. If the length of the optical transmission line is 40 km, and the center wavelength of the bandwidth usage is approximately 1590 nm, then the chromatic dispersion value is 116 ps/nm, and the dispersion slope (i.e., the higher order chromatic dispersion) value is 2.8 ps/nm².

[0250] The chromatic dispersion compensator generates normal dispersion in which the group delay time decreases as the wavelength increases. In order to perform chromatic dispersion compensation on a dispersion shifted fiber, it is necessary to make the absolute values of the chromatic dispersion and the dispersion slope of the chromatic dispersion compensator equal to these values in the dispersion shifted fiber.

[0251] If, using the above, the wavelength dependence of the group delay time requested in the chromatic dispersion compensator are depicted in a graph, it would be like that shown in FIG. 10. Within one wavelength channel, because it is necessary to generate a fixed chromatic dispersion and dispersion slope, it is necessary for the group delay time to be continuous. However, between the wavelength channels, the spectrums of the optical signals are mutually isolated from each other and are independent. Because of this, there is no obstacle even if the group delay time changes discontinuously at the boundaries between the respective wavelength channels. Using the characteristic of the group delay time that discontinuous changes are repeated between wavelength channels, it is possible to superimpose the Bragg grating patterns of a plurality of wavelength channels in the same area of a single optical waveguide.

[0252] The characteristics required for a design are the spectrum of the phase of the reflectance and reflectance magnitude of the Bragg grating optical waveguide, namely, the complex reflectance spectrum thereof. The reflectance magnitude is flat in the 1570 to 1612.2 nm wavelength region, and is set to 85%. The characteristics of the chromatic dispersion of the Bragg grating optical waveguide are reflected in the phase of the reflectance. The relationship in the following Formula 1 is established between the group delay time τ_d and the phase φ.

[0253] $\begin{array}{cc}\left[\mathrm{Expression}2\right]& \\ {\tau }_d\left(\lambda \right)=-\frac{{\lambda }^2}{2\pi c}\frac{d\varphi }{d\lambda }& \left(\mathrm{Formula}1\right)\end{array}$

[0254] Here, the variable λ is the wavelength, the constant π is pi (circumference ratio), and c is the speed of light (within a medium). By integrating the two sides of Formula 1, the phase φ is determined from the group delay time τ_d. As a result of the above, the complex reflectance spectrum is obtained and this is used as a predetermined characteristic in the following step [3].

[0255] In the present embodiments, design which utilizes an amplitude modulation-type of Bragg grating in which the amplitude of the Bragg grating changes and the phase changes subordinately to the amplitude is implemented using processing known as coarse graining (described below). In order to simplify this coarse graining, all of the frequency regions fox which predetermined group delay time characteristics are determined from the frequency point of origin, namely, from 0 Hz are included in the complex reflectance spectrum used as input data in the design.

(Step [3])

[0256] In this step, an effective refractive index profile in a direction along the center axis C of a Bragg grating optical waveguide is derived from the predetermined complex reflectance spectrum obtained in step [2]. Hereinafter, this derivation process will be described.

[0257] Firstly, the following formula is obtained using Maxwell's equations for an electrical field E (z) and a magnetic field H (z) in an optical waveguide. Here, z is a coordinate along the center axis C of the Bragg grating optical waveguide, and the coordinate point of origin (z=0) is placed at the starting end of the optical waveguide, while z is at a maximum value at the termination end. Accordingly, the maximum value of z is the overall length of the Bragg grating optical waveguide.

[0258] $\begin{array}{cc}\left[\mathrm{Expression}3\right]& \\ \frac{dE\left(z\right)}{dz}=\mathrm{i\omega }{\mu }_{o}H\left(z\right)& \left(\mathrm{Formula}2\right)\\ \left[\mathrm{Expression}4\right]& \\ \frac{dH\left(z\right)}{dz}=i{\mathrm{\omega \varepsilon }}_o{n}_{\mathrm{eff}}^2\left(z\right)E\left(z\right)& \left(\mathrm{Formula}3\right)\end{array}$

[0259] i is an imaginary unit, ω is the frequency, μ₀ is the magnetic permeability (in a vacuum), and ∈₀ is the dielectric constant (in a vacuum). In order to construct coupled mode equations from Formula 2 and Formula 3, as in Formula 4 and Formula 5, E (z) and H (z) are converted into a forward wave A₊ (z) and a backward wave A₋ (z) in the coupled mode equations. Reflected waves correspond to A₋ (z).

[0260] $\begin{array}{cc}\left[\mathrm{Expression}5\right]& \\ A_{+}\left(z\right)={\frac{1}{2}\left[\frac{n_{\mathrm{eff}}\left(z\right)}{n_{\mathrm{av}}}\right]}^{1/2}\left[E\left(z\right)+\sqrt{\frac{{\mu }_o}{{\varepsilon }_o}}\frac{H\left(z\right)}{n_{\mathrm{eff}}\left(z\right)}\right]& \left(\mathrm{Formula}4\right)\\ \left[\mathrm{Expression}6\right]& \\ A_{-}\left(z\right)={\frac{1}{2}\left[\frac{n_{\mathrm{eff}}\left(z\right)}{n_{\mathrm{av}}}\right]}^{1/2}\left[E\left(z\right)-\sqrt{\frac{{\mu }_o}{{\varepsilon }_o}}\frac{H\left(z\right)}{n_{\mathrm{eff}}\left(z\right)}\right]& \left(\mathrm{Formula}5\right)\end{array}$

[0261] n_av is the average refractive index of the Bragg grating optical waveguide, and in the present example n_av=2.3480. If the forward wave A₊ (z) and backward wave A₋ (z) are used, then the coupled mode equations are expressed as are shown in the following Formula 6 and Formula 7.

[0262] $\begin{array}{cc}\left[\mathrm{Expression}7\right]& \\ \frac{d{A}_{+}\left(z\right)}{dz}-ik\left(z\right)A_{+}\left(z\right)=-q\left(z\right)A_{-}\left(z\right)& \left(\mathrm{Formula}6\right)\\ \left[\mathrm{Expression}8\right]& \\ \frac{d{A}_{-}\left(z\right)}{dz}+ik\left(z\right)A_{-}\left(z\right)=-q\left(z\right)A_{+}\left(z\right)& \left(\mathrm{Formula}7\right)\end{array}$

[0263] Here, the wave number k (z) is expressed in the following Formula (8), and the potential q (z) in the coupled mode equations is expressed in Formula (9). c_light is the speed of light (within a vacuum).

[0264] $\begin{array}{cc}\left[\mathrm{Expression}9\right]& \\ k\left(z\right)=\frac{\omega }{c_{\mathrm{light}}}{n}_{\mathrm{eff}}\left(z\right)& \left(\mathrm{Formula}8\right)\\ \left[\mathrm{Expression}10\right]& \\ q\left(z\right)=-\frac{1}{2}\frac{d}{dz}\ln \left[n_{\mathrm{eff}}\left(z\right)\right]& \left(\mathrm{Formula}9\right)\end{array}$

[0265] If the potential q (z) is determined, then the effective refractive index profile of the Bragg grating optical waveguide is provided by the following Formula 10.

[0266] $\begin{array}{cc}\left[\mathrm{Expression}11\right]& \\ n_{\mathrm{eff}}\left(z\right)=n_{\mathrm{av}}\exp \left[-2{\int }_0^{z}q\left(s\right)ds\right]& \left(\mathrm{Formula}10\right)\end{array}$

[0267] The overall length z of the Bragg grating optical waveguide is specified as 10.2 mm. An estimation of the overall length is performed in the following manner. The speed of light in a vacuum is multiplied by the maximum value of the group delay time that should be generated in the Bragg grating optical waveguide, and the result is then divided by the average value of the effective refractive indices. The inverse scattering method of Non-patent document 2 is applied to the design of a high refractive index optical waveguide grating pattern, and the potential q (z) is determined from the complex reflectance spectrum R (λ) using the following procedure.

[0268] Firstly, the solutions of Formula 4 and Formula 5 are expressed as in the following Formula 11 and Formula 12.

[0269] $\begin{array}{cc}\left[\mathrm{Expression}12\right]& \\ A_{+}\left(z\right)=e^{i\mathrm{kz}}+{\int }_{\infty }^z{e}^{i{\mathrm{kz}}^{\prime }}{B}_{-}\left(z,z^{\prime }\right)d{z}^{\prime }& \left(\mathrm{Formula}11\right)\\ \left[\mathrm{Expression}13\right]& \\ A_{-}\left(z\right)=e^{-i\mathrm{kz}}+{\int }_{\infty }^z{e}^{-i{\mathrm{kz}}^{\prime }}{B}_{+}\left(z,z^{\prime }\right)d{z}^{\prime }& \left(\mathrm{Formula}12\right)\end{array}$

[0270] A₊ (z) and A₋ (z) are propagated in the +z direction and the −z direction respectively. The integral terms in Formula 11 and Formula 12 show the effects of reflection. The coupled equations are converted into the following Gel'fand-Levitan-Marchenko equations (Formula 13 and Formula 14) from Formula 11 and Formula 12.

[0271] $\begin{array}{cc}\left[\mathrm{Expression}14\right]& \\ B_{+}\left(z,y\right)+{\int }_{-\infty }^z{B}_{-}\left(z,z^{\prime }\right)r\left(z^{\prime }+y\right)d{z}^{\prime }=0& \left(\mathrm{Formula}13\right)\\ \left[\mathrm{Expression}15\right]& \\ r\left(z+y\right)+B_{-}\left(z,y\right)+{\int }_{-\infty }^z{B}_{+}\left(z,z^{\prime }\right)r\left(z^{\prime }+y\right)d{z}^{\prime }=0& \left(\mathrm{Formula}14\right)\end{array}$

[0272] Here, y=c_lightt (wherein t is time), and y<z. r (z) is an inverse Fourier transform of the complex reflectance spectrum R (k) which uses the wave number as a variable, and corresponds to the impulse response. By providing r (z) and solving Formula (13) and Formula (14), q (z) can be determined. q (z) is provided by the following Formula (15)
[Expression 16]
q(z)=−2B(z,z)  (Formula 15)

[0273] If the determined q (z) is substituted in Formula (10), then the effective refractive index profile n_eff(z) is obtained. A graph in which the effective refractive index profile of the present example is plotted over the entire length of a Bragg grating optical waveguide is shown in FIG. 11. z=0 mm corresponds to the start point (i.e., the incident end and the emission end) of the Bragg grating optical waveguide, while z=10.2 mm corresponds to the termination end of the Bragg grating optical waveguide. In addition, the amplitude of the effective refractive index changes over the entire length of the optical waveguide.

[0274] The potential q (z) in Formula 10 and Formula 15 is a real number. As a result of this the calculation to transform from the complex reflectance spectrum R (k) to r (z) which provides an impulse response (in other words, a ‘temporal response’) is a real number type, so that the amplitude changes and the phase changes subordinately to the amplitude.

[0275] Note that the inverse scattering method which is based on the Gel'fand-Levitan-Marchenko equations of the coupled mode equations is described in the following document.

[0276] “An Efficient Algorithm for Solving Zakharov-Shabat Inverse Scattering Problem”, G. Xiao and K. Yashiro, IEEE Transactions on. Antennas and Propagation, 2002, Vol, 50 Issue 6 pp. 807-811.

[0277] FIG. 12 is an enlargement of the horizontal axis of FIG. 11, and displays a portion of a refractive index profile. As is shown in FIG. 12, the effective refractive index oscillates as a function of the coordinate z, and it is shown that a Bragg grating pattern is formed.

[0278] In the amplitude modulation-type Bragg grating of the embodiments of the present invention, it has the characteristic that the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted.

[0279] In order to show an example of amplitude modulation, a portion of the effective refractive index distribution shown in FIG. 11 is enlarged and shown in FIG. 13 together with the envelope curve (the dotted line) of the Bragg grating amplitude. The envelope curve is only displayed fox the maximum values of the amplitude. Because the sign is inverted for the envelope curve for the minimum values of the amplitude at identical points as fox the envelope curve for the maximum values, it is only necessary to consider the envelope curve for the maximum values. An arrow indicates the coordinate point on the waveguide where the sign of the gradient of the envelope curve becomes inverted. The sign inversions exhibit precipitous stepped changes or discontinuous changes that are generated at a single isolated coordinate point.

[0280] In contrast to this, in a sampled Bragg grating, when a sign inversion is generated, it is generated via two points and the precipitous stepped changes or discontinuous changes do not appear. Furthermore, a waveguide region where the amplitude changes continuously to zero is present between these two points. In the amplitude modulation-type grating of the present example, the amplitude of the envelope curve does not become zero at an isolated coordinate point where the sign of the gradient of the envelope curve is inverted, and there are no regions where the amplitude is continuously zero. Accordingly, it is possible for the waveguide length to be made shorter than for a sampled Bragg grating.

[0281] A plurality of isolated coordinate points where the sign of the gradient of the envelope curve is inverted are present on the waveguide. In each of these coordinate points, a discontinuous change in the phase is brought about incidentally. If the phase changes discontinuously, then the pitch (local period) changes. Consequently, the pitch takes on a value which is different from half the value obtained by dividing the center wavelength at that coordinate point (1590.83 nm) by n_av. The accuracy with which the coordinate point where the sign of the gradient of the envelope curve is inverted is specified depends on the interval of discretization of the coordinate z of the waveguide which is shown by the horizontal axis. If this interval is taken as ΔP, then the accuracy for specifying the coordinate point is within a range of ±ΔP. In this manner, in the amplitude modulation-type Bragg grating of the present invention, the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted so that, as a result, coordinate points are present where the pitch changes discretely.

[0282] If the pitch of variations in the effective refractive index over the entire length of the optical waveguide is measured in the effective refractive index distribution of the present example, then it can be understood that the pitch changes discretely as shown in FIG. 14. Here, the pitch is determined by extracting all of the maximum values of the changes in the effective refractive index that regulate the pattern of the Bragg grating, and using the distance between respective adjacent maximum values as the pitch. The pitch of the vertical axis is set within a range of 200 nm through 450 nm. The pitch value where the frequency of occurrence is the highest is the main pitch or pitch center value, and corresponds to half a value determined by dividing the center wavelength (1590.83 nm) by n_av. In the present example, the discrete changes in the pitch are set using ΔP as the minimum unit of change, and with the amount of increase or decrease from the main pitch being an integer multiple of ΔP. Accordingly, the amount of discrete change in the pitch changes in accordance with any changes in the interval of discretization of the coordinates on the waveguide shown on the horizontal axis.

[0283] The discrete changes in the pitch have the feature that they do not appear in a chirped Bragg grating. In a chirped Bragg grating, the pitch changes continuously in the propagation direction of guided light. In a chirped Bragg grating, the amplitude of the Bragg grating also changes at the same time, however, the changes in the amplitude are limited to usage for achieving secondary characteristics such as apodization. Important characteristics such as the phase characteristics and number of channels of the filter reflection spectrum, are achieved by changing the frequency of the Bragg grating in the propagation direction of guided light. In the present step, it is not possible to construct a chirped Bragg grating. In order to construct a chirped Bragg grating, it is necessary to switch the conversion from the complex reflectance spectrum R (ν) to a temporal response (an impulse response) to a complex number type of conversion. As a result of this, the q (z) obtained from Formula 15 becomes a complex number. If q (z) is a complex number, then when n_eff(z) is determined from q (z), because n_eff(z) is a real number, it is necessary for only the real portion of q (z) to be used. Accordingly, an amplitude modulation type of Bragg grating has a different design method from that used for a chirped Bragg grating and the two are classified into mutually different categories. Because it is contrasted with an amplitude modulation type, a chirped Bragg grating is classified as a frequency modulation type.

[0284] In the present description, in this example and in all the other examples, the calculation used for the conversion from the complex reflectance spectrum to an impulse response is a real number-type, and is targeted at amplitude modulation-type Bragg gratings. The conditions (details thereof are given in the following supplement) that are used to select an amplitude modulation-type Bragg grating are set such that, the resolution of discretization of the coordinate axis, namely, the sampling period is not less than the amount of change in the pitch which corresponds to a half value of the width of the reflection band, using coarse graining. In other words, the sampling period is not less than the maximum value of the amount of change from the center value of the pitch in a chirped Bragg grating.

[0285] At this time, the following two conditions may be met: (I) The frequency range of the specified spectrum characteristics includes all the regions from the point of origin (i.e., a frequency of 0) to the region where the relevant spectrum channel is located; (II) A real number type is selected in the above described conversion from the complex reflectance spectrum to an impulse response.

[0286] Here, the reason for this is that, because (I) simplifies the coarse graining, and because (II) is not targeted at chirped Bragg gratings, it is not necessary to select a complex number type in which the procedure is complicated.

[0287] Five discrete values are taken for the pitch value, and the frequency with which these values are taken is concentrated in the three values formed by the center value and by the values above and below the center value. In FIG. 14, regions which include the three values are shown on the vertical axis. Among these, the frequency with which the center pitch (340 nm) is taken is the highest, and this becomes the main pitch. An average value of the minimum value (272 nm) and the maximum value (408 nm) of the pitch within the range of the vertical axis shown in FIG. 14 coincides with the main pitch. If the center wavelength is calculated with the product of the average value (2.3480) of the effective refractive index and the main pitch providing the half value of the center wavelength of the reflection band of the Bragg grating, then this center wavelength is 1597 nm which substantially coincides with the center of the wavelength band shown in FIG. 10. Accordingly, the repetition of a change either +68 nm or −68 nm from the main pitch becomes the principal cause for chromatic dispersion being generated in a plurality of wavelength channels around the center wavelength. As a result of this, the Bragg grating pattern of the present example is formed as a result of the pitch changing discretely at the same time as the amplitude changes continuously.

[0288] Taking a limited number (i.e., a low number) of discrete values as the pitches is effective in maintaining the frequency of work in the manufacturing process on a planar substrate. A grating pattern is manufactured based on pattern lithography using an optical mask. If the pitch changes continuously, it is difficult to maintain the accuracy of the optical lithography over all the pitches, and there is a concern that the pattern of a chirped Bragg grating will differ from the design. If the changes in pitch are limited to a small discrete value, then optimizing the lithography conditions is easy and the accuracy of the lithography is not adversely affected. Accordingly, the design method of the present example is suitable for applications in which an optical waveguide is manufactured on a planar substrate.

[0289] The chromatic dispersion characteristics of a Bragg grating optical waveguide having the effective refractive index profile shown in FIG. 11 is reproduced by simulation, and it is confirmed that they matched the characteristics (see FIG. 10) used as input data. The confirmed simulation is executed as a direct problem in which the effective refractive index profile shown in FIG. 11 is assigned to the coupled mode equations of Formula 6 and Formula 7. When Formula 1 is applied to the phase component of the complex reflectance spectrum which is the result thereof, then the wavelength dependence of the group delay time was obtained as is shown in FIG. 15A through FIG. 15D. When FIG. 15A through FIG. 15D are compared with FIG. 10, it is found that predetermined chromatic dispersion characteristics were reproduced.

(Step [4])

[0290] Based on a correspondence relationship between the dimensions and the effective refractive index of the optical waveguide prepared in Step [1], the effective refractive index profile determined in Step [3] is converted into distribution data (i.e., profile) of the dimensions of the optical waveguide. When the effective refractive index is provided, w_out and w_in, which are the dimensional parameters to be decided, are determined. Accordingly, by associating w_out and w_in with the effective refractive index at each point of the coordinates z, distribution data is obtained for the optical waveguide dimensions w_out and w_in.

[0291] The Bragg grating pattern shown in FIG. 12 has a sinusoidal shape. In a pattern transfer process which is based on dry-etching and lithography using an optical mask, if a square-like shape is employed in which a line which has a fixed width and a space whose width changes in accordance with the pitch are arranged alternatively, there are few fluctuations in the shape after the dry etching. Therefore, after profile data for the optical waveguide dimensions w_out and w_in are obtained from the profile of the effective refractive index, it is converted into a square-like shaped profile. However, in this conversion to a square-like shape, the following two restrictions must be observed.

(1) The line width is fixed to 180 nm (the space changes in accordance with the pitch).

(2) The line amplitude of the square-like shape line is adjusted so as to match the core surface area covered by the sinusoidal Bragg grating pattern.

[0292] In accordance with the above flow, the profiles of w_out and w_in shown in FIG. 16 are obtained. The range of the horizontal axis in FIG. 16 is in the same region as the horizontal axis in FIG. 12. Because a groove is provided in the top portion of the core, and the width of the groove is changed in accordance with the Bragg grating pattern, an antiphase type of change is exhibited in which as w_out increases, w_in decreases. If projections are provided on the top portion of the core, and the width of the projections is changed in accordance with the Bragg grating pattern, then a normal phase type of change is exhibited in which as w_out increases, w_in also increases.

[0293] A procedure for manufacturing a Bragg grating optical waveguide having reduced polarization dependence is described above using Steps [1] through [4]. It is estimated that the element length is less than half the length when an optical fiber Bragg grating is used. By manufacturing an optical waveguide based on the above description, it is possible to provide a small-sized chromatic dispersion compensator having reduced polarization dependence. Because the optical waveguide is a reflection type, light which is input from z=0 is propagated along the Bragg grating optical waveguide, and is propagated reversely from the input direction and is output from z=0. Note that as long as Step [1] is executed prior to Step [4], Step [1] may be executed after either one of Step [2] or Step [3].

[0294] If both w_out and w_in are not changed simultaneously and only one is changed, then it is not possible to reduce the polarization dependence, and the difference between n_eff^TE and n_eff^TN increases to approximately 1000 ppm at maximum, which is approximately 50 times the difference in the present example. As is shown in FIG. 15B through FIG. 15D, a linear relationship is maintained between the group delay time and the wavelength, and variations in the group delay time which accompany variations in the change increase by approximately a factor of 50 compared with the present example. Namely, using the present example, in a chromatic dispersion compensator which employs a high refractive index optical waveguide, it is possible to reduce the polarization dependence of the chromatic dispersion to approximately 1/50^th.

[0295] The Bragg grating optical waveguide described in the present example can also be used for chromatic dispersion compensation in other wavelength regions. As an example of a chromatic dispersion compensator in another wavelength band, a case in which a C-band region is targeted is provided in another example.

[0296] The cross-sectional structure of a planar optical waveguide element of the present exemplary embodiment is shown in FIG. 17. The core of a planar optical waveguide element 30 which has the cross section shown in FIG. 17 is a composite core formed by two areas, namely, inner side cores 31 and 32 and an outer side core 34. The present embodiment is the same as the third embodiment shown in FIG. 7 except for the fact that the inner side cores 31 and 32 do not have a center gap. The structures of the outer side core 34, of a first Bragg grating pattern which is formed on a side wall 34b of the outer side core 34, of a second Bragg grating pattern which is formed in a groove 34c in a top surface 34a, and of a substrate 35, a bottom cladding 36, and a top cladding 37 are the same as in the third embodiment shown in FIG. 7. The characteristic such as the pitch of the Bragg grating pattern only takes a value from among a limited small number of, namely, three or more discrete values, is the same as that in the above described first embodiment.

[0297] In this embodiment, the inner side cores 31 and 32 are formed as two areas, namely, a first rib 31 and second rib 32, and a central gap is not provided between the two ribs. The first and second ribs 31 and 32 are formed from a material having a higher refractive index than that of the outer side core 34. The first rib 31 and second rib 32 are the same height, and this height is indicated by t₁+t₂ in FIG. 17. The first and second ribs 31 and 32 each have an identical shape, and are shaped so as to be the inverse of each other in a horizontal direction. Specifically, the first and second ribs 31 and 32 are formed by planar portions 31a and 32a which each have a thickness of t₂, and by rectangular parallelepiped portions 31b and 32b which have a height t₁ and a width w₁, and which are positioned on edges of the planar portions 31a and 32a. The material used to form the rectangular parallelepiped portions 31b and 32b is the same as the material used to form the planar portions 31a and 32a. The first rib 31 and the second rib 32 are joined via a center junction portion 33.

[0298] Because the cross-sectional area of the inner side core is reduced due to there being no center gap, variations in the effective refractive index which are caused by size variations in the first and second ribs 31 and 32 are increased. However, because it is possible to omit the manufacturing process which is performed in order to provide the center gap, the manufacturing process can be simplified and both a shortening of the manufacturing time and a reduction in costs can be achieved. If a shortening of the manufacturing time and a reduction in costs are given precedence over the performance of the element, then the structure of the present embodiment is preferred.

[0299] In the planar optical waveguide element of the present embodiment, in the same way as in the above described third embodiment, P-conductivity or N-conductivity may be generated in the first rib 31 and the second rib 32 by doping the medium with suitable impurities. It is also possible to provide electrode pads through which electric voltage is applied to both the first rib 31 and the second rib 32, and by generating an electric potential difference between the two ribs 31 and 32, to make it possible to induce changes in the refractive index which are caused by changes in the carver density, and to thereby make it possible to vary the optical characteristics of an electrode element.

[0300] Note that the generation of conductivities having opposite polarities (i.e., P-type or N-type) in the first rib 31 and the second rib 32, and also the provision of the electrode pads to apply voltage are not essential structure in the present embodiment, and it is also possible to use the inner side cores 31 and 32 without imparting any external voltage thereto.

[0301] Moreover, it is also possible for the first rib 31 and the second rib 32 to be formed from an identical material including whether or not dopants axe added thereto. In this case, the center junction portion 33 is not provided, and the two ribs 31 and 32 are able to form an timer side core as a single integrated layer.

[0302] In the present example, a case was calculated as Example 2 relating to a chromatic dispersion compensator in which, in, an optical waveguide structure having the composite core shown in FIG. 17 (having no center gap), the first and second ribs 31 and 32 are formed from Si, the outer side core 34 is formed from Si_xN_y, the substrate 35 is formed from Si, the bottom cladding 36 is formed from SiO₂, the top cladding 37 is formed from SiO₂, and in which t₁=250 nm, t₂=50 nm, w₁=100 nm, w₂=160 nm, t_out=600 nm, and t_in=100 nm, and in which the thickness of the bottom cladding 36 is 2000 inn, and the maximum thickness of the top cladding 37 is 2000 nm.

[0303] In this Example 2 as well, in the same way as in Example 1, changes in w_out and w_in relative to the effective refractive index were calculated in accordance with Step [1]. The results thereof are shown in FIG. 18. In this Example 1, the average value of the effective refractive index is 2.2225.

[0304] Furthermore, in the same way as in Example 1, a chromatic dispersion compensator for 50 channels of an L band ITU grid in which the frequency interval was 100 GHz is designed in accordance with Steps [2] through [4] for a dispersion shifted fiber having a length of 40 km. The wavelength dependence of the group delay time to be assigned to the chromatic dispersion compensator is the same as that shown in FIG. 10 for Example 1. The reflectance was also 85% in a wavelength region of 1570-1612.2 nm. Accordingly, as for the predetermined characteristics, the complex reflectance spectrum is the same as in Example 1. The bit rate of the transmitted optical signals is also the same as in Example 1 at 40 Gbit/s, and the usage band of each wavelength channel is defined as 80 GHz.

[0305] The overall length of the Bragg grating optical waveguide is 10.737 mm, and the effective refractive index profile shown in FIG. 19 is obtained. This profile is the same as the profile shown in FIG. 11 other than in the following points. In Example 2, the profile is enlarged in the center axis direction and the overall length of the optical waveguide is extended by the amount that the average value of the effective refractive index is smaller than it is in Example 1.

[0306] From the relationship shown in FIG. 18, in the same way as in Step [4] of Example 1, profiles of w_out and w_in are obtained. A portion of this is shown in enlargement in FIG. 20. The Bragg grating optical waveguide of the present example can also be designed so as to deal with wavelength bands outside the L band. In this case, the complex reflectance spectrum to be determined in the relevant wavelength band is determined in accordance with Step [2] described in Example 1, and the shape is designed in accordance with Steps [3] and [4].

[0307] Next, in the same way as in Example 1, an example of a chromatic dispersion compensator for 40 channels ITU grid in which the frequency interval is 100 GHz in a C band (1528.77 to 1577.03 nm) which was designed in accordance with Steps [2] through [4] of Example 1 using a Bragg grating optical waveguide having the cross-sectional structure described in FIG. 7, is described as Example 3.

[0308] The material forming the optical waveguide is the same as in Example 1. The intended optical fiber is a standard dispersion single mode fiber (G652) having a length of 30 km. In a wavelength of 1550 nm, the chromatic dispersion value is 510 ps/nm, and the dispersion slope value is 1.74 ps/nm². The bit rate of the transmitted optical signals is 10 Gbit/s, and the usage band of each wavelength channel is 20 GHz. Outside the usage band, the group delay time is defined to be uniform. A graph of the wavelength dependence of the group delay time required in the chromatic dispersion compensator is shown in FIG. 21. Here, the range of the horizontal axis is 1533.85 through 1565.58 nm. The reflectance is flat within the wavelength region shown by the horizontal axis in FIG. 21 and is 85%.

[0309] The length of the Bragg grating optical waveguide was 9.9 mm, and the effective refractive index profile (shape distribution) shown in FIG. 22 is obtained. The peak distributed in the vicinity of z=2 mm and the peak distributed in the vicinity of z=6.5 through 7 mm are present in order to flatten the reflectance and group delay time in regions separate from the usage band. Accordingly, the Bragg grating length which contributes to changing the group delay time within the 20 GHz usage band at maximum corresponds to the distance differential between these two peaks, and is thought to be 5 mm or less. However, if the length required by an optical fiber Bragg grating having the equivalent functions is estimated based on the results shown in Non-patent document 1, then it is thought to be approximately 10 mm. Accordingly, according to the present example, the optical waveguide length required for chromatic dispersion compensation is shortened to less than half that required by an optical fiber Bragg grating.

[0310] FIG. 23 shows changes in pitch over the entire length in a pitch range of 200 to 450 nm. In the same way as in FIG. 14, the frequency with which the pitch in the center is taken is the highest, and this forms the main pitch. Moreover, the average value of the minimum value (i.e., the next smaller discrete value from the center value) and the maximum value (i.e., the next larger discrete value from the center value) of the pitch within the vertical axis range shown in FIG. 23 coincides with the main pitch.

[0311] In the same way as in Example 1, by assigning the shape distribution of the effective refractive index supplied in FIG. 22 to the coupled mode equations of Formula 6 and Formula 7 and then solving these equations, the chromatic dispersion characteristics shown in FIG. 24 are obtained. If FIG. 24 is compared with FIG. 21, it can be seen that predetermined chromatic dispersion characteristics are reproduced.

[0312] In the same way as in Example 1, by deciding the dimensions of a Bragg grating optical waveguide from the relationships between w_in and w_out relative to the effective refractive index n_eff shown in FIG. 9, it is possible to manufacture a small-size chromatic dispersion compensator for the C band having reduced polarization dependence.

[0313] Next, in the same way as in Example 1, an example of a chromatic dispersion compensator for an L band single channel designed in accordance with Steps [2] through [4] of Example 1 using a Bragg grating optical waveguide having the cross-sectional structure described in FIG. 7 is described as Example 4.

[0314] The material forming the optical waveguide is the same as in Example 1. The intended optical fiber is a dispersion shifted fiber having a length of 30 km. The reflectance is 85% and the characteristics shown in FIG. 25 are specified for the wavelength dependence of the group delay time so that a predetermined complex reflectance spectrum is obtained.

[0315] The length of the Bragg grating optical waveguide is 8.13 mm, and the effective refractive index profile (shape distribution) shown in FIG. 26 was obtained. The peak of the envelope curve of the changes in the effective refractive index in the effective refractive index profile shown in FIG. 26 is in the vicinity of z=4.2 mm.

[0316] FIG. 27 shows changes in pitch over the entire length in a pitch range of 200 to 450 nm. In the case of the present example, the pitch takes only three values (discrete values outside the range of the vertical axis do not appear). In the same way as in FIG. 14, the frequency with which the pitch in the center (340 nm) is taken is the highest, and this forms the main pitch. Moreover, the maximum value (i.e., the next larger discrete value from the center value) is 68 nm bigger than the center value and the minimum value (i.e., the next smaller discrete value from the center value) is 68 nm smaller than the center value. The average value of the maximum value and the minimum value coincides with the center value which forms the main pitch.

[0317] As is shown in FIG. 27, on the front end side and the rear end side of the peak position of the effective refractive index, the trend of the discrete changes in pitch is inverted. On the front end side of the peak position, the pitch only takes two values, namely, the center value and the maximum value. In other words, a binary change is exhibited on the long wavelength side of the center value. In contrast, on the rear end side of the peak position, the pitch only takes two values, namely, the center value and the minimum value, and a binary change is exhibited on the short wavelength side of the center value. Chromatic dispersion compensation is possible by a pitch change Bragg grating which is simpler than chirped type of Bragg grating in which the pitch changes continuously. The Bragg grating of Example 1 may be thought to be constructed by combining a plurality of patterns to the pattern of the present example which, is used as a basis.

[0318] In the same way as in Example 1, by assigning the profile of the effective refractive index supplied in FIG. 26 to the coupled mode equations of Formula 6 and Formula 7 and then solving these equations, the chromatic dispersion characteristics shown in FIG. 28 are obtained. If FIG. 28 is compared with FIG. 25, it can be seen that predetermined chromatic dispersion characteristics are reproduced.

[0319] As a result of the above, it is possible to manufacture a chromatic dispersion compensator which has reduced polarization dependence in a single wavelength channel within the L band. The manufacture of elements used in different wavelength bands can also be achieved by considering the characteristics of the chromatic dispersions that correspond to each wavelength band, and designing Bragg grating optical waveguides using the ideas of the present example.

[Method of Connecting a Chromatic Dispersion Compensator and an Optical Transmission Line]

[0320] In the chromatic dispersion compensators of Examples 1 through 4, optical signals emitted from a Bragg grating optical waveguide are propagated in the opposite direction along the path of incident optical signals. Namely, because emitted signal light is propagated along the same path as incident signal light, a method of separating the emitted signal light from the incident signal light is required. In the present example, as is shown in FIG. 29, a description is given of the structure of a chromatic dispersion compensator by connecting an optical circulator 102 to a chromatic dispersion compensator 101, in, which the chromatic dispersion compensator has a port where incident signal light is input into the chromatic dispersion compensator, and a port where emitted signal light is output from the chromatic dispersion compensator.

[0321] If the chromatic dispersion compensator 101 of the present example corresponds to the chromatic dispersion compensator of exemplary embodiments described herein, then the chromatic dispersion compensator 101 of any of the Examples 1 through 4 may be used, or another element may be used. The optical circulator 102 is connected to the front end portion side of the chromatic dispersion compensator 101. An incident optical fiber 103 that propagates incident signal light, a coupling optical fiber 104 that connects together the chromatic dispersion compensator 101 and the optical circulator 102, and an emission optical fiber 105 that propagates emitted signal light are connected to the optical circulator 102.

[0322] Incident signal light is transferred by the optical circulator 102 from the incident optical fiber 103 to the coupling optical fiber 104, and is input into the chromatic dispersion compensator 101. Emitted signal light that is reflected inside the chromatic dispersion compensator 101 is transferred from the coupling optical fiber 104 to the emission optical fiber 105 via the optical circulator 102. In order to reduce loss which is caused by the connection between the coupling optical fiber 104 and the chromatic dispersion compensator 101, it is preferable for lens fabrication to be performed on the distal end of the coupling optical fiber 104 (i.e., the distal end thereof on the chromatic dispersion compensator 101 side), or for a micro lens to be placed between the coupling optical fiber 104 and the chromatic dispersion compensator 101, or for the coupling optical fiber 104 to be connected by adhesion to the front end portion of the Bragg grating optical waveguide of the chromatic dispersion compensator 101. The loss created by the connection is, for example, approximately 1 dB. Because the loss within the optical circulator 102 is approximately 1 dB, the total optical loss created by the connections of the optical circulator 102 is approximately 2 dB.

[0323] In order to install the structure 100 shown in FIG. 29 on an optical fiber transmission line that is intended for chromatic dispersion compensation, it is sufficient for the incident optical fiber 103 to be connected to the transmitter side of the optical fiber transmission line, and for the emission optical fiber 105 to be connected to the receiver side of the optical fiber transmission line. By doing this, it is possible to construct a small-sized chromatic dispersion compensator that can be installed on an optical fiber transmission line, and that has a low level of optical insertion loss.

[0324] An optical filter having reflection bands in ten mutually different wavelength channels is constructed using the optical waveguide in the above described firth embodiment of an optical waveguide element. The method used to design the optical filter comprises the following steps [1] to [4].

[0325] [1] The width (w) of the cross-section of the optical waveguide core is specified, and the field distribution of the eigen mode in the TE-type polarization and TM-type polarization in the cross-section are calculated, and a correspondence relationship between the effective refractive index and the optical waveguide width w is then obtained. This step is the optical waveguide cross-sectional structure design process of the present example.
[2] The reflection characteristics desired for the optical filter are specified, and the necessary data required to determine the structure of the optical waveguide is obtained. What are specified as the reflection characteristics are the reflectance and phase in each wavelength. All of the frequency regions that include the desired reflection characteristics from the point of origin (i.e., from a frequency of 0 Hz) are included in the frequency range.
[3] The optical waveguide length is provided, and the profile (shape distribution) of the effective refractive index extending in the waveguide direction of the optical waveguide is derived from the complex field reflectance spectrum obtained in Step [2] using an inverse scattering problem solution. This step includes a calculation process to convert the complex field reflectance spectrum into a temporal response, however, this is a real number type of conversion.

[0326] Steps [2] and [3] are the Bragg grating pattern design process. The Bragg grating obtained in this design process is an amplitude modulation-type in which only the amplitude of the Bragg grating changes.

[0327] [4] Based on the correspondence relationships between the effective refractive index and the cross-sectional dimensions of the optical waveguide core obtained in Step [1], the profile (shape) extending in the propagation direction of guided light of the Bragg grating optical waveguide is decided from the profile (shape distribution) of the effective refractive index obtained in Step [3]. This step is the optical filter design process.

[0328] Note that in the same way as in the above described chromatic dispersion compensator design step, the order of these steps can also be switched.

[0329] Hereinafter, each step of designing the optical filter will be described in detail when Si₃N₄ (having a refractive index of 2.05) is used for the material of the core, and SiO₂ (having a refractive index of 1.45) is used for the material of the cladding.

[0330] Step [1]

[0331] A relationship between the effective refractive index n_eff and the waveguide width w is shown in FIG. 2. The cross-sectional structure of the waveguide is the same as that shown in the insertion diagram in FIG. 2. The portion indicated by diagonal lines is the core, and cladding surrounds both the top and bottom as well as side surfaces thereof. The thickness t of the core is 1.4 μm. Mode 1 and mode 2 correspond respectively to a TE mode and a TM mode. In the present embodiment, the TE mode is used for the subject, however, the same procedure can be applied for the TM mode as well.

[0332] Step [2]

[0333] The optical characteristics of an optical filter having reflection bands in ten mutually different wavelength channels are specified. In optical communication, it is common for distinctions to be made between spectrum regions using frequency instead of wavelength. In the present example, hereinafter, the spectrum characteristics of an optical filter as a function of frequency will be discussed. A complex field reflectance spectrum R (ν) is calculated from the reflectance and phase in each frequency. In an orthogonal coordinate system, R (ν) is formed by a real number component and an imaginary number component, however, converting the coordinates into a polar coordinate system and splitting the complex field reflectance into phase and absolute value of the field reflectance simplifies dealing with the optical filter characteristics. Therefore, as in the following Formula A, the complex field reflectance is expressed using a polar coordinate display.
[Expression 17]
R(ν)=|R(ν)|exp[−φ(ν)]  (Formula A)

[0334] Here, R is the electric field, ν is the frequency, |R (ν)| is the absolute value of the field reflectance, and φ(ν) is the phase. The absolute value of the reflectance is normalized to 1 (namely 100%). The absolute value of the field reflectance is set to 0.95 (i.e. 95%) such that the power reflectance |R (ν)|² becomes 0.9 (90%), within the reflection band of each channel.

[0335] In the optical filter of the present example, the chromatic dispersion in the reflection band of each channel is set to zero. If the chromatic dispersion is zero, the phase forms a linear function relative to the frequency. As a result of the above, the optical characteristics specified for the optical filter of the present example are shown in FIG. 30. In FIG. 30, the left vertical axis shows the absolute value of the field reflectance |R (ν)|, the right vertical axis shows the phase φ(ν), and these are plotted respectively by a solid line and a broken line. The horizontal axis shows the frequency ν in units of THz, and specifies optical characteristics by dividing the frequency from 192.6 THz to 193.6 THz into ten equal channels at intervals of 0.1 THz. The center frequency is 193.1 THz. If this is converted into a center wavelength, it becomes 1552.52 μm. The spectrum width of the reflection band in each channel is 0.01 THz, and it can be seen that the phase changes linearly within this range.

[0336] If the spectrum shape of the rectangular reflection band of each channel, such as those shown in FIG. 30, is converted into a temporal waveform using an inverse Fourier transform, a sine function type of impulse waveform is obtained. If the width of the spectrum of the reflection band is taken as Δν, then the main peak of the sine function type impulse waveform is contained within a time region of approximately Δt=3/(Δν). Accordingly, in an optical waveguide that generates a reflection band in each channel shown in FIG. 30, the propagation time that is required between when light is incident and when it is reflected must be approximately Δt or even greater than this. The phase, which changes linearly within the frequency region of each reflection band shown in FIG. 30, reflects phase delay which is caused by this propagation time.

[0337] FIG. 30 only displays the frequency bands in the vicinity of channels where a reflection band is present. All of the frequency bands in which a reflection channel is present from the point of origin (i.e., from 0 THz) are included as desired optical characteristics in the optical characteristics being target of an inverse scattering method. However, because no reflection channels are present in frequency regions outside those shown in FIG. 30, the value of the field reflectance is zero.

[0338] Step [3]

[0339] The effective refractive index distribution in the waveguide direction of the optical waveguide forming an optical filter is derived based on the inverse scattering problem solution. This procedure is described in Step [3] in the designing direction of the above described chromatic dispersion compensator.

[0340] When specifying the overall length of the optical waveguide, the optical path length in accordance with Δt in Step [2] is taken as the minimum value, and the specification is made based on the loss and allowable dimensions of the optical waveguide. After the optical waveguide length has been specified, the potential q (z) is determined using an inverse scattering problem solution. q (z) is substituted to the above described Formula 10, and the effective refractive index distribution n_eff(z) is obtained. Here, the conversion which is used when an impulse response is derived from the complex reflectance spectrum R (ν) is a real number type.

[0341] As a result of this, the q (z) which is obtained from the above described Formula 15 is also a real number, and an effective refractive index distribution for an amplitude modulation-type Bragg grating is obtained in which the amplitude of the Bragg grating changes and the phase changes incidentally to the amplitude. A definition of amplitude modulation in the present invention is given below.

[0342] n_eff(z) is plotted in FIG. 31 and FIG. 32. The horizontal axis z shows coordinates in the propagation direction of guided light. z=0 mm is the starting end of the Bragg grating optical waveguide, and z=39.8194 mm is the termination end thereof. n_av which corresponds to the average value of the refractive index distribution of the grating optical waveguide is 1.94945 in the present example.

[0343] FIG. 32 shows an enlargement of the effective refractive index distribution shown in FIG. 31 for a portion of the optical waveguide. It can be seen that n_eff oscillates at a period of half the value obtained by dividing the center wavelength (1552.52 nm), which corresponds to the center frequency (193.1 THz), by n_av, so as to exhibit a pattern that provides the Bragg grating.

[0344] One of the features of the amplitude modulation-type Bragg grating of the present invention is that the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted. Namely, in the present invention, the change when the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted is known as amplitude modulation.

[0345] In order to show an example of amplitude modulation, a portion of the effective refractive index distribution shown in FIG. 31 is shown in enlargement in FIG. 33 together with the envelope curve (i.e., the dotted line) of the Bragg grating amplitude. The envelope curve is only displayed for the maximum values of the amplitude. Because the sign is inverted for the envelope curve for the minimum values of the amplitude at identical points as for the envelope curve for the maximum values, it is only necessary to consider the envelope curve for the maximum values. An arrow indicates the coordinate point on the waveguide where the sign of the gradient of the envelope curve becomes inverted.

[0346] The sign inversions exhibit stepped precipitousness (changes) or discontinuousness (discontinuous changes) that are generated at a single isolated coordinate point, and the feature of a sampled Bragg grating that a waveguide region is interposed where the amplitude changes continuously to zero between two points where the sign is inverted do not appear in the amplitude modulation-type grating of the present invention, because the amplitude becomes zero only at an isolated coordinate point where the sign of the gradient of the envelope curve is inverted, essentially, there are no regions where the amplitude is zero. Accordingly, it is possible for the waveguide length to be made shorter than for a sampled Bragg grating.

[0347] A plurality of isolated coordinate points where the sign of the gradient of the envelope curve is inverted are present on the waveguide. In each of these coordinate points, a discontinuous change in the phase is brought about incidentally. If the phase changes discontinuously, the pitch (local period) changes. Consequently, the pitch takes on a value which is different from half the value obtained by dividing the center wavelength (1552.52 nm) at the relevant coordinate point by n_av. The accuracy with which the coordinate point where the sign of the gradient of the envelope curve is inverted is specified depends on the interval of discretization of the coordinate z of the waveguide which is shown on the horizontal axis. If this interval is taken as ΔP, then the accuracy with which the coordinate point is specified is within a range of ±ΔP.

[0348] In this manner, in the amplitude modulation-type Bragg grating of the present example, the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted so that, as a result, coordinate points are present where the pitch changes discretely.

[0349] In the effective refractive index distribution of the present example, if the pitch is measured over the entire length of the optical waveguide, it can be seen that it changes discretely as is shown in FIG. 34. Here, the pitch is determined by extracting all of the maximum values of the changes in the effective refractive index that regulate the pattern of the Bragg grating, and using the distance between respective adjacent maximum values as the pitch. The pitch value where the frequency of occurrence is the highest is the main pitch or pitch center value, and corresponds to half a value determined by dividing the center wavelength (1552.52 nm) by n_av. In the present example, the main pitch is approximately 398.2 nm. The discrete changes in the pitch take ΔP as the minimum unit of change, and take the amount of increase or decrease from the main pitch as an integer multiple of ΔP. Accordingly, the amount of discrete change in the pitch changes in accordance with any changes in the interval of discretization of the coordinates on the waveguide shown on the horizontal axis.

[0350] The discrete changes in the pitch have the feature that they do not appear in a chirped Bragg grating. In a chirped Bragg grating, the pitch changes continuously in the propagation direction of guided light. In a chirped Bragg grating, the amplitude of the Bragg grating also changes at the same time, however, the changes in the amplitude are confined to being used for achieving secondary characteristics such as apodization. Important characteristics such as the phase characteristics and number of channels of the filter reflection spectrum are achieved by changing the frequency of the Bragg grating in the propagation direction of guided light. In the present step, it is not possible to construct a chirped Bragg grating. In order to construct a chirped Bragg grating, it is necessary to switch the conversion from the complex reflectance spectrum R (ν) to a temporal response (an impulse response) to a complex number type of conversion. As a result of this, the q (z) obtained from Formula 15 becomes a complex number. If q (z) is a complex number, then when n_eff(z) is determined from q (z), because n_eff(Z) is a real number, it is necessary for only the real portion of q (z) to be used. Accordingly, an amplitude modulation type of Bragg grating has a different design method from that used for a chirped Bragg grating and the two are classified into mutually different categories. Because it is contrasted with an amplitude modulation type, a chirped Bragg grating is classified as what is known as a frequency modulation type.

[0351] Step [4]

[0352] Based on a mutual correspondence relationship between the optical waveguide width w and the effective refractive index n_eff prepared in Step [1], the effective refractive index distribution (z) obtained in Step [3] is converted into distribution data (i.e., a profile) of the optical waveguide width w. The optical waveguide width w is determined using the correspondence relationship of the effective refractive index of the TE mode shown in FIG. 2. As is shown in FIG. 32, the Bragg grating pattern in the effective refractive index distribution has a sinusoidal shape.

[0353] In a pattern transfer process which is based on dry-etching and lithography using an optical mask, if a square-like shape is employed in which a line which has a fixed width and a space whose width changes in accordance with the pitch are arranged alternatively, there are few fluctuations in the shape after the dry etching. Therefore, after profile data for the optical waveguide width w has been obtained from the profile of the effective refractive index, it is converted into a square-like shaped profile. However, in this conversion to a square-like shape, the following two restrictions must be observed.

(1) In the present example, the line width is fixed at 140 nm. In contrast, the spaces change in accordance with the pitch of the grating. A larger value than the threshold value of the fabrication accuracy is set for the line width.

(2) The line amplitude of the square-like shape is adjusted so as to match the core area covered by the sinusoidal Bragg grating pattern.

[0354] In accordance with the above flow, the profiles of the square-like shaped w shown in FIG. 35 are obtained. The range of the horizontal axis in FIG. 35 is in the same region as the horizontal axis in FIG. 35. Moreover, FIG. 36 shows the profile of the waveguide width in the vicinity of the coordinate points where the sign of the grading of the envelope curve is inverted.

[0355] Applications for the optical filter of the present example include, for example, using it to extract as reflection light only the signal light of wavelength-multiplexed channels after this light is transmitted through an optical amplifier, and to remove spontaneous emission optical noise which is present in wavelength regions adjacent to the signal light. However, no consideration is given to reducing polarization dependence. Because the pitch changes discretely, it becomes easy to manufacture a photolithographic mask which is used in a process to manufacture an optical waveguide, and it is possible to manufacture this type of optical filter with a high level of fabrication accuracy. Furthermore, because Bragg grating patterns that correspond to different wavelength channels are mutually superimposed as a single grating pattern in a common location on the waveguide without being connected together in series, the problems occurring in Patent document 6 can be solved.

[0356] In order to confirm that Bragg gratings for different wavelength channels are superimposed, a distribution in the waveguide direction of the grating frequency is shown in FIG. 37. In FIG. 37, the left-side horizontal axis displays the grating frequency. The right-side horizontal axis displays the effective refractive index. The left-side vertical axis shows the delay time. The right-side vertical axis corresponds to the position from the incident end of the optical waveguide. The left-side graph shows changes in the grating frequency in the waveguide direction. The right side is an effective refractive index profile. The graph showing the changes in the grating frequency in the waveguide direction is a spectrogram obtained by a procedure known as a short-term Fourier transform.

[0357] The procedure to obtain a spectrogram showing changes in the Bragg grating frequency in the propagation direction of guided light by a short-term Fourier transform will now be described using Formulae. Firstly, a position z from the incident end of the optical waveguide is divided by the speed of light c in a vacuum. This is then multiplied by the average refractive index n_av so that, as is shown in the following Formula 16, a propagation time t on the optical waveguide is obtained.

[0358] $\begin{array}{cc}\left[\mathrm{Expression}18\right]& \\ t=\frac{n_{\mathrm{av}}z}{c}& \left(\mathrm{Formula}16\right)\end{array}$

[0359] If the effective refractive index is expressed as a function of t based on Formula 16, then the effective refractive index distribution n_eff(z) of the Bragg grating can be regarded as the impulse response n_eff (t) of the effective refractive index. If a short-term. Fourier transform is executed for n_eff(t), then the spectrogram S (ν,τ) shown in the following Formula 17 is obtained.

[0360] $\begin{array}{cc}\left[\mathrm{Expression}19\right]& \\ S\left(v,\tau \right)={\int }_{-\infty }^{\infty }dt\left[g\left(\tau -t\right)n_{\mathrm{eff}}\left(t\right)\exp \left(i2\pi \mathrm{vt}\right)\right]& \left(\mathrm{Formula}17\right)\end{array}$

[0361] In Formula 17, ν is the frequency, τ is the delay time, and g is the gate function. In the present example, the gate function is a Gauss function and is expressed by the following Formula 18.

[0362] $\begin{array}{cc}\left[\mathrm{Expression}20\right]& \\ g\left(\tau -t\right)=\exp \left[-{\left(\frac{\tau -t}{\Delta \tau }\right)}^2\right]& \left(\mathrm{Formula}18\right)\end{array}$

[0363] In Formula 18, Δτ corresponds to the gate width. The gate width is equal to or greater than a reciprocal of the frequency interval of the specified reflection band. In the present embodiment, because the frequency interval of the reflection band is 0.1 THz, the gate width is 10 ps or greater. Because the delay time is equivalent to the propagation time, it corresponds to the position from the incident end of the optical waveguide from Formula 16. Accordingly, the point where a straight line drawn parallel with the horizontal axis from the point of a particular delay time value on the left-side vertical axis in FIG. 37 intersects with the right-side vertical axis is the waveguide position that corresponds to that delay time. It can be seen from the spectrogram in FIG. 37 that, in this Bragg grating, grating, components for different wavelength channels coexist in a common area on a single waveguide. The spectrogram intensity at delay times that correspond to the vicinities of the incident end and the rear end of waveguide increases, however, this is not caused by the Bragg grating and is due to the intermittent discontinuity of the impulse response at both ends of the waveguide becoming a broadband response and appearing in numerical calculations.

[0364] Note that the number of channels, the channel intervals, and the spectrum width of the reflection band are not limited to the numerical values of the present example, and an optical filter can be designed with the optimum numerical values specified in accordance with the application.

[0365] It should also be noted that, even if different optical characteristics are specified, an optical filter can still be designed and manufactured in the same way. For example, if the optical filter is to be used as a beam splitter, then it is also possible for the power reflectance within the reflection band of each channel, to be specified as 0.4 (40%).

[0366] In the present example, an optical filter structure targeting rectangular waveguides has been described. However, similar optical filters are not limited to rectangular waveguides and can also be constructed using optical fibers.

[0367] In the present example, an optical filter having the optical characteristics (see FIG. 30) described in Example 1 of an optical filter is designed using the planar optical waveguide (see FIGS. 3A through 3C) in the above described second embodiment of an optical waveguide element. The method used to design the optical filter comprises the following steps [1] to [4].

[0368] [1] The dimensions (w_in/w_out) of the cross-sectional structure of the optical waveguide core are specified, and the field distribution of the eigen mode of the TE-type polarization and TM-type polarization in the cross-section are calculated. These dimensions are then adjusted such that the effective refractive indices in the two polarizations are equal. w_in/w_out are then decided such that polarization dependence is cancelled out in different effective refractive indices. Correspondence relationships between the effective refractive indices and w_in/w_out are then obtained so that it is possible to decide the dimensions of the cross-sectional structure of the optical waveguide core from the effective refractive indices. This step is the design process for the cross-sectional structure of the optical waveguide core.
[2] The reflection characteristics desired for the optical filter are specified, and the necessary data required to determine the structure of the optical waveguide is obtained. What are specified as the reflection characteristics are the reflectance and phase in each wavelength. All of the frequency regions that include the desired reflection characteristics from the point of origin (i.e., from a frequency of 0 Hz) are included in the frequency range.
[3] The optical waveguide length is provided, and the profile (shape distribution) of the effective refractive index along the waveguide direction of the optical waveguide is derived from the complex field reflectance spectrum obtained in Step [2] using an inverse scattering problem solution. This step includes a calculation process to convert the complex field reflectance spectrum into a temporal response, and this is a real number type of conversion.

[0369] Steps [2] and [3] form the Bragg grating pattern design process.

[0370] [4] Based on the correspondence relationships between the effective refractive indices and the cross-sectional dimensions of the optical waveguide core obtained in Step [1], the profile extending in the propagation direction of guided light of the Bragg grating optical waveguide is decided from the profile of the effective refractive index obtained in Step [3]. This step forms the optical filter design process.

[0371] Note that in the same way as in the above described chromatic dispersion compensator design step, the order of these steps can also be switched.

[0372] Step [1]

[0373] The cross-sectional structure of the waveguide is the same as that shown in the insertion diagram in FIG. 4B. The portion indicated by diagonal lines is the core, and cladding surrounds both the top and bottom as well as sides thereof. The thickness of the cladding positioned at the top portion and at the bottom portion of the core is 2 μm in both cases. t_in is 0.1 μm, while t_out is 1.4 μm. As is shown in FIG. 4B, if the relationship between w_in and w_out is provided it is possible to reduce the polarization dependence of the effective refractive index of the waveguide as is shown in FIG. 4A.

[0374] If the effective refractive index of the TE polarization is regarded as the effective refractive index of the waveguide, then the graph shown in FIG. 5 is obtained when relationships between the effective refractive index and w_in, and w_out are calculated and plotted.

[0375] Step [2] and Step [3]

[0376] Step [2] and step [3] form a Bragg grating pattern design process, and can be implemented in the same way as in the above described Example 1 of an optical filter. Because of this, any duplicate description thereof is omitted. An effective refractive index profile of an optical waveguide derived in these steps is shown in FIG. 38 and FIG. 39.

[0377] The Bragg grating obtained in this design process is an amplitude modulation type in which only the amplitude of the Bragg grating changes.

[0378] Step [4]

[0379] Based on a correspondence relationship between the dimensions w_in and w_out and the effective refractive index n_eff of the optical waveguide prepared in Step [1], the effective refractive index distribution n (z) obtained in Step [3] is converted into distribution data (i.e., a profile) of w_in and w_out. When the effective refractive index is determined, w_out and w_in, which are the dimensional parameters to be decided, are determined by using the correspondence relationship shown in FIG. 4A and FIG. 4B. As is shown in FIG. 39, the Bragg grating pattern in the effective refractive index distribution has a sinusoidal shape.

[0380] In a pattern transfer process which is based on dry-etching and lithography using an optical mask, if a square-like shape is employed in which a line which has a fixed width and a space whose width changes in accordance with the pitch are arranged alternatively, there are few fluctuations in the shape after the dry etching. Therefore, after profile data for the optical waveguide dimensions w_out and w_in has been obtained from the profile of the effective refractive index, it is converted into a square-like shaped profile. However; the two restrictions that must be observed in this conversion to a square-like shape are the same as in the above described Example 1 of an optical filter.

[0381] In accordance with the above flow, the profiles of w_out and w_in shown in FIG. 40 are obtained. The range of the horizontal axis in FIG. 40 is in the same region as the horizontal axis in FIG. 39.

[0382] Applications for the optical filter of the present example include, for example, using it to extract, as non-polarization dependent reflection light, only the signal light of wavelength-multiplexed channels after this light has passed through an optical amplifier, and to remove spontaneous emission optical noise which is present in wavelength regions adjacent to the signal light. Note that the number of channels, the channel intervals, and the spectrum width of the reflection band are not limited to the numerical values of the present example, and the optical filter can be designed with the optimum numerical values specified in accordance with the application.

[0383] The present example is an example of the design of an optical filter having a single reflection band. The design steps are the same as in Example 2. The power reflectance of the reflection band is set to approximately 90%. The optical waveguide structure which is used is the planar optical waveguide in the above described second embodiment of an optical waveguide element (see FIGS. 3A through 3C), and the correspondence relationship between the effective refractive index and w_in and w_out is identical to Example 2. The specified optical characteristics are shown in FIG. 41. The spectrum width of the reflection band is 0.01 THz.

[0384] The effective refractive index distribution derived based on an inverse scattering problem solution using these optical characteristics is shown in FIG. 42 and FIG. 43. Discrete changes in the pitch are shown in FIG. 44. In the present example, there are three discrete values for the pitch. An absolute value of the difference between the pitch of the center value and the pitch of the next smallest value is equal to an absolute value of the difference between the pitch of the center value and the pitch of the next largest value.

[0385] In the present example, the results obtained when the effective refractive index was converted to a square-like shaped profile are shown in FIG. 45. The optical filter of the present example can be used to extract signal, light of a single specific channel as reflection light.

[0386] Note that the spectrum width of the reflection band is not limited to 0.01 THz and it is also possible to design an optical filter with an optional width specified.

[0387] The present example is an example of the design of an interleaver for a wavelength channel having intervals of 0.1 THz. In the present example, an optical filter is designed with the channel interval set to 0.2 THz, and with the width of the reflection band of each channel set to 0.1 THz. The optical waveguide structure which is used is the planar optical waveguide in the above described second embodiment of an optical waveguide element (see FIGS. 3A through 3C), and the correspondence relationship between the effective refractive index and w_in and w_out is identical to Example 2. The specified optical characteristics are shown in FIG. 46. The effective refractive index distribution derived based on an inverse scattering problem solution using these optical characteristics is shown in FIG. 47 and FIG. 48. The results obtained when the effective refractive index is converted to a square-like-shaped profile are shown in FIG. 49.

[0388] The optical filter (i.e., the interleaver) of the present example is able to split the signal light of each channel having an interval of 0.1 THz into two paths made up of odd-numbered and even-numbered channels.

[0389] In the present example, an optical filter having the optical characteristics (see FIG. 30) described in Example 1 of an optical filter is designed using the planar optical waveguide (see FIG. 7) in the above described third embodiment of an optical waveguide element. The correspondence relationship between the effective refractive index and w_in and w_out of the optical waveguide are shown in FIGS. 8A and 8B, and FIG. 9, in this example.

[0390] In step [3], apart from n_av being taken as 2.348, the effective refractive index distribution is derived in the same way as in Example 2 of an optical filter. The effective refractive index distribution derived using these optical characteristics based on an inverse scattering problem solution is shown in FIGS. 50 and 51. The results obtained when the effective refractive index is converted to a square-like shaped profile are shown in FIG. 52.

[0391] In the optical waveguide of this example, because it is possible to reduce the deterioration in the accuracy of the effective refractive index which is caused by fabrication errors, it is possible to reduce the polarization dependence generated by fabrication errors, and it is possible to extract only signal light, as reflection light, from wavelength multiplexed channels irrespective of the polarization thereof. Furthermore, by applying voltage to the inner side core, it is possible for the wavelength of the reflection band to be made variable.

[0392] Note that the third embodiment (see FIG. 7) of an optical waveguide element can also be applied to an optical filter that achieves the optical characteristics of Examples 1 through 4.

[0393] In the present example, an optical filter (i.e., an interleaver) having the optical characteristics described in Example 4 of an optical filter (see FIG. 46) is designed using the planar optical waveguide (see FIG. 7) of the above described third embodiment of the planar optical waveguide element. The effective refractive index distribution derived using these optical characteristics based on an inverse scattering problem solution is shown in FIG. 53. Discrete changes in the pitch are shown in FIG. 54. The results obtained when the effective refractive index is converted to a square-like shaped profile are shown in FIG. 55.

[0394] The optical filter (i.e., the interleaver) of the present example is able to switch between extraction directions of odd-numbered and even-numbered (i.e., transmission and reflection) by applying voltage (i.e., is able to perform direction reversal).

[0395] In the present example, an optical filter having the optical characteristics described in Example 1 of an optical filter (see FIG. 30) is designed using the planar optical waveguide (see FIG. 17) of the above described forth embodiment of the planar optical waveguide element. In the present example, the relationships between the effective refractive index and w_in and w_out are shown in FIG. 18.

[0396] In step [3], apart from n_av being taken as 2.22252, the effective refractive index distribution is derived in the same way as in Example 2 of an optical filter. The effective refractive index distribution derived using these optical characteristics based on an inverse scattering problem solution is shown in FIG. 56. The results obtained when the effective refractive index is converted to a square-like shaped profile are shown in FIG. 57.

[Optical Resonator]

[0397] As is shown in FIG. 58, an optical resonator 150 has a structure in which optical waveguides which are formed by reflection mirrors 151 and 152 (i.e., a first optical waveguide 151 and a second optical waveguide 152) are placed at the two ends thereof, and a third optical waveguide 153 which includes an optical resonator medium is sandwiched between the reflection mirrors 151 and 152. In the present example, the first optical waveguide 151, the third optical waveguide 153, and the second optical waveguide 152 are connected together in series so that a single planar optical waveguide is formed, and optical waveguides that have a Bragg grating pattern and that have a reflection function are used for the reflection mirrors 151 and 152 at the two ends thereof. The design of an optical waveguide having a reflection function can then be achieved in accordance with the above described optical filter design method by setting the desired reflection characteristics. The third optical waveguide 153 which is formed by an optical resonator medium may have a predetermined optical path length in order for light to resonate between the reflection mirrors 151 and 152.

[0398] Because it is desirable to extract light to the outside of the resonator, the reflectance of at least one mirror is lower than 1 (i.e., 100%). For example, as is shown in FIG. 58, in order to emit a portion of the light transmitted from the reflection minor of the second optical waveguide 152, a fourth optical waveguide 154 for emissions is provided. The fourth optical waveguide 154 is connected in series to the first through third optical waveguides so that a single planar optical waveguide is formed.

[0399] An optical resonator is designed so as to have a function of selecting any one of a plurality of wavelength channels. An example of a plurality of wavelength channels is an ITU grid having frequency intervals of 100 GHz. A description of the optical characteristics of the component elements of an optical resonator having the required function will now be described based on FIG. 59 and FIG. 60. In the graph in the bottom portion of FIG. 59, the power reflection spectrum (the solid line) of the first reflection mirror and the power reflection spectrum (the broken line) of the second reflection mirror are shown.

[0400] A spectrum obtained as the product of the power reflection spectra of the first and second reflection mirrors is shown in the graph in the top portion of FIG. 59. The power reflectance of the reflection bands of the first and second reflection mirrors is set at 0.9 (i.e., 90%). The wavelength of the light resonant in the optical resonator is limited to the region where the reflection regions of the two spectra overlap. This is generally called a vernier function, and is used in applications in which a specific wavelength component is extracted by combining two optical filters having mutually different comb-like power reflection spectra, and, furthermore, is used to make it possible to vary the wavelength components being extracted by varying the characteristics of one optical fiber.

[0401] The resonance characteristics of an optical resonator that includes the first and second optical waveguides having the characteristics shown in FIG. 59 are shown in the graph (total) in the top portion in FIG. 60. The vertical axis is shown in a normal logarithmic scale. If it is assumed that the power reflectances of the two end mirrors are not wavelength-dependent and are fixed at 0.9, then the resonance characteristics are shown in the graph (FP) in the bottom portion of FIG. 60. In these resonance characteristics, the resonance peak is normalized to 1. The optical length of the optical resonator is set to 1000 μm. If the effective refractive index of the optical waveguide is 1.94945, then the waveguide length obtained by conversion from the optical length is approximately 513 μm. The resonance characteristics shown in the graph in the top portion of FIG. 60 are based on a spectrum obtained by superimposing the characteristics of the graph shown in the bottom portion of FIG. 60 onto the spectrum in the graph shown in the top portion of the FIG. 59. The resonance characteristics in the graph shown in the top portion of FIG. 60 have a peak at 193.1 THz (1552.52 nm).

[0402] By fixing the effective refractive index of the first optical waveguide, and changing the effective refractive index of the second optical waveguide, and changing the pitch of the Bragg grating pattern in the second reflection mirror, it is possible to select wavelength components of different single channels for the reflection spectrum of the first reflection mirror using a vernier function. Namely, by changing the effective refractive index of the second optical waveguide, the selected wavelength can be varied. Of course, it is also possible to change the effective refractive index of the first reflection mirror, or to change the effective refractive indices of both reflection mirrors. In the graph shown in the top portion of FIG. 60, the side channel suppression ratio is approximately 24 dB.

[0403] In order to maximize the resonance power in a selected wavelength channel, it is sufficient to adjust the phase shift which is generated when light is propagated along the third optical waveguide which is an optical resonator medium. Namely, it is sufficient to adjust the effective refractive index of the third optical waveguide. In the graph shown in the top portion of FIG. 60, the phase shift is 0.477π.

[0404] Hereinafter, the procedure to design the first optical waveguide which is formed by the first reflection mirror will be described.

[0405] The method used to design, the first reflection mirror in the present example comprises the following steps [1] to [4].

[0406] [1] The width (w) of the cross-section of the optical waveguide core is specified, and the field distribution of the eigen mode in the TE-type polarization and TM-type polarization in the cross-section are calculated, and a correspondence relationship between the effective refractive index and the optical waveguide width w is then obtained. This step is the optical waveguide cross-sectional structure design process of the present example.
[2] The reflection characteristics desired for the reflection mirror are specified, and the necessary data required to decide the structure of the optical waveguide is obtained. What are specified as the reflection characteristics are the reflectance and phase in each wavelength. All of the frequency regions that include the desired reflection characteristics from the point of origin (i.e., from a frequency of 0 Hz) are included in the frequency range.
[3] The optical waveguide length is provided, and the profile (shape distribution) of the effective refractive index extending in the waveguide direction of the optical waveguide is derived from the complex field reflectance spectrum obtained in Step [2] using an inverse scattering problem solution. This step includes a calculation process to convert the complex field reflectance spectrum into a temporal response, and this is a real number type of conversion.

[0407] Steps [2] and [3] are the Bragg grating pattern design process. The Bragg grating obtained in this design process is an amplitude modulation-type in which only the amplitude of the Bragg grating changes.

[0408] [4] Based on the relationships between the effective refractive index and the cross-sectional dimensions of the optical waveguide core obtained in Step [1], the shape (profile) extending in the propagation direction of guided light of the Bragg grating optical waveguide is decided from the shape distribution (profile) of the effective refractive index obtained in Step [3]. This step is the reflection mirror design process.

[0409] Note that in the same way as in the above described chromatic dispersion compensator design step, the order of these steps can also be switched.

[0410] Hereinafter, each step of designing the first reflection mirror will be described in detail when Si₃N₄ (having a refractive index of 2.05) is used for the material of the core, and SiO₂ (having a refractive index of 1.45) is used for the material of the cladding.

[0411] Step [1]

[0412] A relationship between the effective refractive index n_eff and the waveguide width w is shown in FIG. 2. The cross-sectional structure of the waveguide is the same as that shown in the insertion diagram in FIG. 2. The portion indicated by diagonal lines is the core, and cladding surrounds both the top and bottom as well as side surfaces thereof. The thickness t of the core is 1.4 μm. Mode 1 and mode 2 correspond respectively to a TE mode and a TM mode. In the present embodiment, the TE mode is used for the subject, however, the same procedure can be applied for the TM mode as well.

[0413] Step [2]

[0414] Using the desired phase characteristics and the power reflection spectrum in the graph shown in the bottom portion of FIG. 59, the complex field reflectance spectrum R(ν) of the grating optical waveguide is calculated. In an orthogonal coordinate system, R (ν) is formed by a real number component and an imaginary number component, however, converting the coordinates into a polar coordinate system and splitting the complex field reflectance into phase and absolute value of the field reflectance simplifies dealing with the reflection mirror characteristics. Therefore, as in the above mentioned Formula A, the complex field reflectance is expressed using a polar coordinate system.

[0415] The absolute value of the reflectance is normalized to 1 (namely 100%). Within the reflection band of each channel, the absolute value of the field reflectance is set to 0.95 (i.e. 95%) such that the power reflectance |R (ν)|² becomes 0.9 (90%).

[0416] In the reflection mirror of the present example, the chromatic dispersion in the reflection band of each channel is set to zero. If the chromatic dispersion is zero, the phase provides a linear function relative to the frequency. As a result of the above, the optical characteristics specified for the reflection mirror of the present example are shown in FIG. 30. In FIG. 30, the left vertical axis shows the absolute value of the field reflectance |R (ν)|, the right vertical axis shows the phase φ(ν), and these are plotted respectively by a solid line and a broken line. The horizontal axis shows the frequency ν in units of THz, and specifies optical characteristics by dividing the frequency from 192.6 THz to 193.6 THz into ten equal channels at intervals of 0.1 THz. The center frequency is 193.1 THz. If this is converted into a center wavelength, it becomes 1552.52 nm. The spectrum width of the reflection band in each channel is 0.01 and it can be seen that the phase changes linearly within this range.

[0417] If the spectrum shape of the rectangular reflection band of each channel, such as those shown in FIG. 30, is converted into a temporal waveform using an inverse Fourier transform, a sine function type of impulse waveform is obtained. If the width of the spectrum of the reflection band is taken as Δν, then the main peak of the sine function type impulse waveform is contained within a time region of approximately Δt=3/(Δν). Accordingly, in an optical waveguide that generates a reflection band in each channel shown in FIG. 30, the propagation time that is required between when light is incident and when it is reflected must be approximately Δt or even greater than this. The phase, which changes linearly within the frequency region of each reflection band shown in FIG. 30, reflects phase delay which is caused by this propagation time.

[0418] In FIG. 30, only the frequency bands in the vicinity of channels where a reflection band is present are displayed. All of the frequency bands in which a reflection channel is present from the point of origin (i.e., from 0 THz) are included as desired optical characteristics in the optical characteristics being target of an inverse scattering method. However, because no reflection channels are present in frequency regions outside those shown in FIG. 30, the value of the field reflectance is zero.

[0419] Step [3]

[0420] The effective refractive index distribution in the waveguide direction of the optical waveguide forming a reflection mirror is derived based on the inverse scattering problem solution. This procedure is described in Step [3] in the design direction of the above described chromatic dispersion compensator.

[0421] When specifying the entire length of the optical waveguide, the optical path length in accordance with Δt in Step [2] is taken as the minimum value, and the specification is made based on the loss and allowable dimensions of the optical waveguide. After the optical waveguide length is specified, the potential q (z) is determined using the inverse scattering problem solution, q (z) is substituted to the above described Formula 10, and the effective refractive index distribution n_eff(z) is obtained. Here, the conversion which is used when an impulse response is derived from the complex reflectance spectrum R (ν) is a real number type.

[0422] As a result of this, the potential q (z) which is obtained from the above described Formula 15 is also a real number, and an effective refractive index distribution for an amplitude modulation type Bragg grating in which the amplitude of the Bragg grating changes and the phase changes incidentally to the amplitude is obtained. A definition of amplitude modulation in the present invention is given below.

[0423] n_av(z) is plotted in FIG. 31 and FIG. 32. The horizontal axis z shows coordinates in the propagation direction of guided light. z=0 nun is the starting end of the Bragg grating optical waveguide, and z=39.8194 mm is the termination end thereof. n_av which corresponds to the average value of the refractive index distribution of the grating optical waveguide is 1.94945 in the present example.

[0424] FIG. 32 shows an enlargement of the effective refractive index distribution shown in FIG. 31 for a portion of the optical waveguide. It can be seen that n_eff oscillates at a period of half the value obtained by dividing the center wavelength (1552.52 nm), which corresponds to the center frequency (193.1 THz), by n_av, so as to exhibit a pattern that regulates the Bragg grating.

[0425] A feature of the amplitude modulation-type Bragg grating of the present example is that, as was described in [Example 1 of an optical filter], the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted. In order to show an example of amplitude modulation, a portion of the effective refractive index distribution shown in FIG. 31 is enlarged and shown in FIG. 33 together with the envelope curve (the dotted line) of the Bragg grating amplitude. If the pitch is measured over the entire length of the optical waveguide in the effective refractive index distribution of the present example, it can be seen that it changes discretely as is shown in FIG. 34.

[0426] Step [4]

[0427] Based on a correspondence relationship between the optical waveguide width w and the effective refractive index n_eff prepared in Step [1], the effective refractive index distribution n_eff(z) obtained in Step [3] is converted into distribution data (i.e., a profile) of the optical waveguide width w. The optical waveguide width w is determined using the correspondence relationship of the effective refractive index for the TE mode shown in FIG. 2. As is shown in FIG. 32, the Bragg grating pattern in the effective refractive index distribution has a sinusoidal shape.

[0428] In a pattern transfer process which is based on dry-etching and lithography using an optical mask, if a square-like shape is employed in which a line which has a fixed width and a space whose width changes in accordance with the pitch are arranged alternatively, there are few fluctuations in the shape after the dry etching. Therefore, after profile data for the optical waveguide width w has been obtained from the profile of the effective refractive index, it is converted into a square-like shaped profile. However, in this conversion to a square-like shape, the following two restrictions must be observed.

(1) In the present example, the line width is fixed at 140 nm. In contrast, the spaces change in accordance with the pitch of the grating. A larger value than the threshold value of the fabrication accuracy is set for the line width.

(2) The line amplitude of the square-like shape is adjusted so as to match the core area covered by the sinusoidal Bragg grating pattern.

[0429] In accordance with the above flow, the profiles of the square-like shaped w shown in FIG. 35 are obtained. The range of the horizontal axis in FIG. 35 is in the same region as the horizontal axis in FIG. 35. Moreover, FIG. 36 shows the profile of the waveguide width in the vicinity of the coordinate points where the sign of the grading of the envelope curve is inverted.

[0430] In order to confirm that Bragg gratings for different wavelength channels are superimposed, a distribution in the waveguide direction of the grating frequency is shown in FIG. 37. In FIG. 37, the left-side horizontal axis displays the grating frequency. The right-side horizontal axis displays the effective refractive index. The left-side vertical axis shows the delay time. The right-side vertical axis corresponds to the position from the incident end of the optical waveguide. The left-side graph shows changes in the grating frequency in the waveguide direction. The right side is an effective refractive index profile. The graph showing the changes in the grating frequency in the waveguide direction is a spectrogram obtained by a procedure known as a short-term Fourier transform.

[0431] FIG. 31 through FIG. 37 shown in the present example are the same as those in the above described Example 1 of an optical filter.

[0432] A design procedure for the first optical waveguide which forms the first reflection mirror has been described above. However, the second optical waveguide which forms the second reflection mirror can also be designed in the same way based on the predetermined phase characteristics and the power reflection spectrum shown in the graph in the bottom portion of FIG. 59.

[0433] The third optical waveguide is connected in series between the first optical waveguide and the second optical waveguide. The length of the third optical waveguide is as described above. If an optical waveguide on a substrate is used, then an optical waveguide in which the first optical waveguide, the third optical waveguide, and the second optical waveguide are connected in series can be defined on an optical mask.

[0434] The optical resonator of the present example can be used in applications such as optical filters that extract a specific frequency component, and lasers resonators. If it is used for laser resonators, it is necessary for the third optical waveguide to have an optical amplification function which utilizes optical gain.

[0435] In the present example, a reflection mirror structure targeting rectangular waveguides has been described. Similar reflection mirrors are not limited to rectangular waveguides and can also be constructed using optical fibers. In this case, the third optical waveguide can also be constructed using an optical fiber.

[0436] In the present example, a first reflection mirror having the optical characteristics described in Example 1 of an optical resonator (see FIG. 59 and FIG. 60) was designed using the planar optical waveguide in the above described second embodiment of an optical waveguide element (see FIGS. 3A through 3C). In the present example, in the same way as in the above described Example 2 of an optical filter, w_in and w_out are specified as the dimensions of the cross-sectional structure of the optical waveguide core. A correspondence relationship between the effective refractive index and w_in and w_out of the optical waveguide is shown in FIG. 5.

[0437] In the present example, in the same way as in the above described Example 2 of an optical filter, the effective refractive index distribution is derived. The effective refractive index distribution derived based on an inverse scattering problem solution using these optical characteristics is shown in FIG. 38 and FIG. 39. The results obtained when the effective refractive index was converted to a square-like shaped profile are shown in FIG. 40. Note that FIG. 38 through FIG. 40 are the same as those shown in the above described Example 2 of an optical filter.

[0438] According to the present example, by reducing polarization dependence, it is possible to manufacture an optical resonator for a desired polarization.

[0439] In the present example, a first reflection mirror having the optical characteristics described in Example 1 of an optical resonator (see FIG. 59 and FIG. 60) is designed using the planar optical waveguide (see FIG. 7) of the above described third embodiment of the optical waveguide element. In the present example, the correspondence relationships between the effective refractive index and w_in and w_out are shown in FIGS. 8A and 8B and in FIG. 9.

[0440] In step [3], apart from being taken as 2.348, the effective refractive index distribution is derived in the same way as in Example 2 of an optical resonator. The effective refractive index distribution derived using these optical characteristics based on an inverse scattering problem solution is shown in FIGS. 50 and 51. The results obtained when the effective refractive index is converted to a square-like shaped profile are shown in FIG. 52. Note that FIG. 51 through FIG. 52 are the same as those illustrated in the above described Example 5 of an optical filter.

[0441] In the optical waveguide of this example, because it is possible to reduce the deterioration in the accuracy of the effective refractive index which is caused by fabrication errors, it is possible to reduce the polarization dependence generated by fabrication errors, and it is possible to extract only signal light, as reflection light, from wavelength multiplexed channels irrespective of the polarization thereof. Furthermore, by applying voltage to the inner side core, it is possible for the wavelength of the reflection band to be made variable.

[0442] In the present example, a first reflection mirror having the optical characteristics described in Example 1 of an optical resonator (see FIG. 59 and FIG. 60) is designed using the planar optical waveguide (see FIG. 17) of the above described fourth embodiment of the optical waveguide element. In the present example, the correspondence relationships between the effective refractive index and w_out of the optical waveguide are shown in FIG. 18.

[0443] In step [3], apart from n_av being taken as 2.22252, the effective refractive index distribution is derived in the same way as in Example 2 of an optical resonator. The effective refractive index distribution derived using these optical characteristics based on an inverse scattering problem solution is shown in FIG. 56. The results obtained when the effective refractive index is converted to a square-like shaped profile are shown in FIG. 57. Note that FIG. 56 through FIG. 57 are the same as those illustrated in the above described Example 7 of an optical filter.

[Amplitude Modulation-Type Bragg Gratings]

[0444] In the above description, it was described that an amplitude modulation-type Bragg grating based on exemplary embodiments differs from a chirped Bragg grating. In contrast, according to the sampling theorem described below, a Bragg grating pattern is defined exclusively (i.e., as being unique), and differences such as amplitude modulation types or chirped types of Bragg grating pattern did not appear. However, this applies to continuous effective refractive index distributions, and does not apply to discrete effective refractive index distributions that have undergone coarse graining. This point is supplemented below.

[0445] The effective refractive index distribution of a Bragg grating is obtained as an effective refractive index distribution for discrete points sampled at fixed intervals on a coordinate axis which extends in the light propagation direction. If the sampling theorem derived by Nyquist, Shannon, and Someya is applied to the effective refractive index distribution of a Bragg grating, then if the coordinate interval, namely, the sampling period of discrete points in the effective refractive index distribution obtained from a design is set to not more than half the local period (i.e., the pitch) of the sinusoidal changes in the effective refractive index of the targeted Bragg grating, then the continuous effective refractive index distribution which corresponds to the discrete effective refractive index distribution is determined as being unique. In order to determine a continuous effective refractive index distribution, as is shown in the following Formula B, a Whittaker-Shannon interpolation formula which utilizes a sine function is used.

[0446] $\begin{array}{cc}\left[\mathrm{Expression}21\right]& \\ q\left(z\right)=\sum _{n=-\infty }^{+\infty }\left\{q\left(n{Z}_{\mathrm{IS}}\right)·\sin c\left[\pi \left(\frac{z}{Z_{\mathrm{IS}}}-n\right)\right]\right\}& \left(\mathrm{Formula}B\right)\end{array}$

[0447] Mere, z is the continuous coordinates, q (z) expresses a potential that provides an effective refractive index distribution defined by continuous coordinates, q (nZ_IS) expresses a potential that provides an effective refractive index defined by discrete coordinates, and Z_IS is the sampling period. In the present example, because the sampling period is ⅕^th the pitch, the effective refractive index is determined uniquely. n is an integer specifying a discrete coordinate point. In reality, because the Bragg grating length is n is finite. Reproducing the (original) continuous waveform from the discrete waveform is called “reconstruction”. In order to achieve a Bragg grating having predetermined optical characteristics, it is necessary for the effective refractive index distribution to be able to be reconstructed uniquely. The above Formula 10 is used to obtain the effective refractive index n_eff from the potential.

[0448] It should be noted that, in order to form a Bragg grating pattern by optical exposure, it is necessary to prepare data of a Bragg grating pattern for an optical mask. The pattern data for an optical mask is prepared as a digital file in GDS format or the like. Because the number of data points is unlimited in a continuous effective refractive index distribution, the file capacity is infinite. Accordingly, it is necessary to use a discrete effective refractive index distribution which has a limited number of data points as the optical mask pattern data. As a result of this, even if a continuous effective refractive index distribution is reconstructed, it is necessary to convert it to a discrete distribution. Because of this, the discrete effective refractive index distribution before the reconstruction is used for the mask pattern data. In a discrete effective refractive index distribution, the shape of the effective refractive index distribution differs depending on the discretized sampling period and the mode of discretization. This fact generates the differences between amplitude modulation-type and chirped Bragg gratings. If there is a need to further improve the accuracy with which predetermined optical characteristics are reproduced, then it is possible to use an effective refractive index distribution which has been discretized after undergoing the reconstruction as the mask data.

[0449] For example, a case in which optical characteristics such as those shown in FIG. 61 are specified as the predetermined characteristics is considered. In FIG. 61, an example of the characteristics of an optical element having a single reflection channel is shown. In the top portion of the graph in FIG. 61, the frequency dependence of the delay time is plotted, while in the bottom portion of the graph in FIG. 61, the absolute value of the complex field reflectance and phase are plotted. The frequency width of the reflection channel is approximately 1.244 THz. The center frequency is 193.1 THz. The spectrum occupancy of the half value of the width of the reflection channel is approximately 0.32% of the center frequency, which is a narrow band. Note that in each example of the present invention, the width of each channel is even narrower.

[0450] If a Bragg grating that provides the above characteristics is formed by a chirped Bragg grating, then a resolution that corresponds to changing the pitch by a maximum of only 0.32% is required for the discretization of the coordinate axis of the Bragg grating. Namely, the number of segments used to discretize each pitch is at least 313 which is an inverse number of 0.32%. In order to reproduce the fact that the pitch changes continuously in the propagation direction of guided light, it is necessary to raise the resolution even further, so that the number of data points increases even further. Accordingly, if an attempt to accurately construct a chirped Bragg grating is made, the number of data items is enormous, and the processing itself of the mask data becomes difficult. Moreover, a maximum amount of change in the pitch of only 0.32% is only approximately 1 nm if the pitch center value is taken as 340 nm. In order to chirp this, it is necessary to segment this even further, however, it is difficult to accurately manufacture an optical mask pattern at a sub-nanometer accuracy level.

[0451] Accordingly, it can be said that an amplitude modulation-type Bragg grating is advantageous from the standpoints of improving the accuracy of the manufacturing process, and of both reducing processing time and costs. As has been described above, in order to select an amplitude modulation-type Bragg grating pattern, it is sufficient to perform coarse graining in which the resolution of discretization of the coordinate axis is taken as more than an amount of change in the pitch which corresponds to the half value of the width of the reflection band, in other words, the resolution of discretization of the coordinate axis is taken as more than (not less than) the maximum value of the amount of change from the pitch center value in a chirped Bragg grating. As a result of this, the continuous changes in pitch in a chirped Bragg grating are accumulated, and the feature is evident that the sign of the gradient of the envelope curve of the amplitude of the Bragg grating is inverted at a single isolated coordinate point, and that the phase changes continuously incidentally to this.

[0452] According to exemplary embodiments described herein, it is possible to provide an optical waveguide element that can achieve high-level functionality such as dispersion compensation, and also can simplify control of fabrication accuracy in the manufacturing process, and also possible to provide a chromatic dispersion compensator that uses this optical waveguide element, and also possible to provide a design method thereof.

[0453] Although a few exemplary embodiments have been shown and described, it would be appreciated by those skilled in the art that changes may be made in these embodiments without departing from the principles and spirit of the inventive concept, the scope of which is defined in the claims and their equivalents.