Semiconductor laser source and method for emitting with this laser source

Abstract

A semiconductor laser source including a Mach-Zehnder interferometer including first and second arms. Each of these arms being divided into a plurality of consecutive sections. The first and second arms each include a gain-generating section forming first and second gain-generating waveguides, respectively. The laser source includes power sources able to deliver currents through the gain-generating waveguides such that the following condition is met:

Description

The invention relates to a semiconductor laser source and to a method for emitting an optical signal at a wavelength λ_Li close to a desired wavelength λ_Si.

Known semiconductor laser sources comprise:

• a set of waveguides optically coupled to one another and forming an optical cavity able to make an optical signal resonate at a plurality of possible resonant frequencies, the possible wavelengths λ_Rj of these possible resonant frequencies being regularly spaced apart from one another by an interval Δλ_R, this set comprising:
  • at least one waveguide in which a bandpass filter is produced, the transmission spectrum of this bandpass filter comprising a passband able to select, among the various possible wavelengths λ_Rj, the wavelength λ_Li closest to the wavelength λ_Si, and
  • at least one gain-generating waveguide able to generate optical gain at each wavelength λ_Li selected by the bandpass filter, this gain-generating waveguide comprising a III-V gain medium.

Such a known laser source is for example disclosed in patent application US2018261976A.

Prior art is also known from US20140254617A1, US20030123784A1, CN103368678B, US2016301191A1, US2012057079A1 and WO2007107187A1.

Patent application US20140254617A1 discloses a bandpass filter containing a Mach-Zehnder interferometer. However, the arms of the Mach-Zehnder interferometer are devoid of gain-generating sections.

Patent application US20030123784A1 discloses a multi-mode demultiplexer using an interferometer.

Patent application CN103368678B describes an optical switch using a Mach-Zehnder interferometer. This optical switch is devoid of resonant cavity.

It is known that the wavelength λ_Li of a semiconductor laser source varies as a function of temperature. It is desirable to limit this variation as much as possible. Many solutions have already been proposed to limit this variation. For example, in patent application US2018261976A, the bandpass filter of the laser source is made of silicon nitride in order to limit this variation as much as possible. However, the fabrication of the bandpass filter from silicon nitride complexifies the fabrication of the laser source.

It is also desirable to limit the bulk of such a semiconductor laser source as much as possible and to simplify the production thereof.

The invention aims to provide an architecture for the laser source that minimizes the length of the resonant cavity without however decreasing the power of the emitted optical signal. By way of secondary objective, the invention also aims to provide a laser source the variations in the emission wavelength λ_Li of which as a function of temperature are limited without it being necessary to make the bandpass filter from a material other than silicon.

One of the subjects thereof is therefore a semiconductor laser source according to Claim 1.

Embodiments of this laser source may comprise one or more of the features of the dependent claims.

Another subject of the invention is a method for emitting an optical signal at a wavelength λ_Li close to a desired wavelength λ_Si.

The invention will be better understood on reading the following description, which is given merely by way of nonlimiting example with reference to the drawings, in which:

FIG. 1 is a schematic illustration, seen from above, of the architecture of a semiconductor laser source;

FIG. 2 is a schematic illustration in vertical cross section of the laser source of FIG. 1;

FIG. 3 is a graph illustrating the transmission spectrum of a bandpass filter of the laser source of FIG. 1;

FIGS. 4, 5, 8, 10, 11, 13 and 15 are schematic illustrations, seen from above, of various other possible architectures of the laser source of FIG. 1;

FIGS. 6 and 7 are graphs illustrating transmission spectra of filters of the laser source of FIG. 5;

FIG. 9 is a graph illustrating the transmission spectrum of filters of the laser source of FIG. 8;

FIG. 12 is a flowchart of a method for emitting an optical signal at a wavelength λ_Li using the laser source of FIG. 11;

FIG. 14 is a graph illustrating the transmission spectrum of filters of the laser source of FIG. 13.

In these figures, the same references have been used to reference elements that are the same. In the rest of this description, features and functions well known to those skilled in the art will not be described in detail.

Section I: Definitions and Notations

The wavelength λ_Li is the wavelength at which the laser source emits. When the laser source is a monomode or monochromatic laser source, there is only a single wavelength λ_Li. When the laser source is a multimode or polychromatic laser source, a plurality of different wavelengths λ_Li exist simultaneously. In the latter case, the index i is an identifier of the wavelength λ_Li among the various wavelengths at which the laser source emits. Typically, the wavelength λ_Li is comprised between 1250 nm and 1590 nm.

The wavelengths λ_Rj are the various wavelengths at which a resonant cavity is capable of resonating in the absence of bandpass filter. The bandpass filter is the filter that selects a small number of wavelengths λ_Rj. These wavelengths λ_Rj are regularly spaced apart from one another by an interval Δλ_R. The index j is an identifier of one particular wavelength λ_Rj.

The interval ΔR is the smallest wavelength range containing all the possible wavelengths λ_Rj. Typically, when the resonant cavity is bounded by reflectors of optical signals, the interval ΔR is equal to the reflective band of these reflectors. The interval ΔR may also be limited by the width of the amplification band or “gain band” of the optical amplifier used to amplify the optical signals that resonate in the interior of the cavity. The gain band is a frequency band that contains all the frequencies of the optical signals capable of being amplified by the optical amplifier. Optical signals the frequencies of which are outside of this gain band are not amplified by the optical amplifier.

The reflective band of a reflector is the −3 dB reflective band. It is a question of the wavelength range containing all the wavelengths λ_Rj able to be reflected by the reflector with a power higher than or equal to 50% of the maximum power I_max reflected by this reflector. The power I_max is equal to the power of the reflected optical signal for the wavelength λ_Rj, at which this power is maximum.

The wavelength λ_CR is a wavelength corresponding to a maximum of the reflection spectrum of the reflector. Typically, the wavelength λ_CR is the wavelength located at the middle of the reflective band of the reflector. This wavelength λ_CR moves at the same time as the reflective band moves.

Here, the widths of the passbands are the widths of the −3 dB passbands. Below, the width of the passband of a filter is denoted Δλ_FX, where the index FX is an identifier of the filter.

The wavelength that corresponds to a maximum in the transmission spectrum of a filter is denoted λ_CFX, where the index FX is an identifier of the filter. Typically, the wavelength λ_CFX is the wavelength located at the middle of a passband of the filter.

The distance between the maximums of two successive passbands of the transmission spectrum of a bandpass filter is denoted d_FSRX, where the index FX is an identifier of the filter. This distance is known as free spectral range (FSR).

By “athermal” filter, what is meant is the fact that the coefficient dλ_CFX/dT associated with this filter is comprised between L_min and L_max, where:

• • λ_CFX is the wavelength that corresponds to a maximum in the transmission spectrum of the filter,
  • the coefficient dλ_CFX/dT is the variation in the wavelength λ_CFX as a function of temperature,
  • L_min is equal to −50 pm/° C. and, preferably, equal to −15 pm/° C. or −7 pm/° C., and
  • L_max is equal to +50 pm/° C. and, preferably, equal to +15 pm/° C. or +7 pm/° C.

For a given filter, the coefficient dλ_CFX/dT may be determined experimentally or by numerical simulation. For example, the variation in the wavelength λ_CFX as a function of temperature is measured or simulated in a temperature range extending from 20° C. to 100° C. Typically, the wavelength λ_CFX varies approximately linearly as a function of temperature in this temperature range. It is therefore possible to approximate the relationship that relates the wavelength λ_CFX to the temperature T in this range using the following linear relationship: λ_CFX=aT+b, where the coefficients a and b are constants. The values of the coefficients a and b are established by determining the straight line that minimizes the deviations, for example in the least-squares sense, between this straight line and each measured or simulated point. Each measured or simulated point has as abscissa a temperature and as ordinate the wavelength λ_CFX measured or obtained by simulation for this temperature. The value of the coefficient dλ_CFX/dT is then set equal to the value of the coefficient a.

The thermo-optical coefficient of a material is the coefficient dn_r/dT, where:

• • n_r is the refractive index of this material at the wavelength λ_Si, and
  • dn_r/dT is the variation in the refractive index n_r as a function of temperature in a temperature range extending from 20° C. to 100° C.Similarly to what was indicated for the coefficient dλ_CFX/dT, the coefficient dn_r/dT is approximated by a constant in the temperature range extending from 20° C. to 100° C.

Section II: Examples of Embodiments

FIG. 1 schematically shows the general architecture of a semiconductor laser source 10 that emits at at least one wavelength λ_Li that is as close as possible to a desired wavelength λ_Si. The wavelength λ_Si is constant and either set once and for all during the design of the laser source or adjustable by a user. In any case, once set, the wavelength λ_Si remains constant provided that the settings of the laser source are not modified.

Below, only the particularities of the laser source 10 are described in detail. For general information on the production and operation of a semiconductor laser source using waveguides made of silicon and of a III-V gain medium, the reader may consult the following article: B. Ben Bakir et al., “Hybrid Si/III-V lasers with adiabatic coupling”, 2011.

The laser source 10 comprises a back reflector 12 and a front reflector 14 that define the ends of a Fabry-Pérot resonant cavity. In the interior of this cavity, an optical signal may resonate, in the absence of bandpass filter, at a plurality of possible wavelengths λ_Rj. The wavelengths λ_Rj are regularly spaced apart from one another by an interval Δλ_R. In this embodiment, the interval ΔR that contains all the wavelengths λ_Rj is equal to the reflective band of the reflectors 12 and 14.

For example, the reflector 12 has a reflectance strictly higher than that of the reflector 14. The reflectance is equal to the ratio of the power of the optical signal reflected by the reflector to the power of the optical signal incident on this reflector. Typically, the reflectance of the reflector 12 is higher than or equal to 90% or 95% for the wavelength λ_Li. The reflectance of the reflector 14 is generally comprised between 30% and 80% and is typically equal to 50%.

The reflectors 12 and 14 are wideband reflectors. In this embodiment, this means that the width of the reflective band of the reflectors 12 and 14 is larger than a preset lower limit. This lower limit is here:

• • equal to Δλ_f+DT.(dλ_CR/dT) if the bandpass filter is arranged to select solely a single wavelength λ_Li, and
  • equal to Δλ_f+max {DT.(dλ_CR/dT); N.d_SFRf+DT.(dλ_Cf/dT)} if the bandpass filter is arranged to select N wavelengths λ_Li,where:
  • Δλ_f is the width of the passband of the bandpass filter of the laser source,
  • DT is the width of a preset operating temperature range of the laser source,
  • dλ_CR/dT is the variation in the central wavelength λ_CR of the reflectors as a function of temperature expressed in nm/° C.,
  • dλ_Cf/dT is the variation in the central wavelength λ_Cf of the bandpass filter as a function of temperature expressed in nm/° C.,
  • N is an integer number higher than or equal to two,
  • d_SFRf is the interval between two successive passbands of the bandpass filter expressed in nanometres,
  • the symbol “·” designates in this description the operation of scalar multiplication, and
  • max{ . . . } is the function that returns the highest of the elements that are located between the curly brackets and that are separated from each other by a semi-colon. Various embodiments of the bandpass filter of the laser source are described below.

The operating temperature range of a laser source is often chosen higher than 10° C. or 30° C. Here the operating temperature range is chosen as being comprised between +20° C. and +100° C. The width DT is therefore here equal to 80° C. Below, the lowest and highest temperatures of the operating temperature range are denoted T_min and T_max, respectively. In this embodiment, the reflectors 12 and 14 are produced in a waveguide the core of which is made of silicon. Thus, the variation dλ_CR/dT is here equal to 0.07 nm/° C. Typically, the width of the reflective band is larger than 6 nm or 15 nm or 30 nm.

The reflectors 12 and 14 are also designed so that, at the temperature T_min, the wavelength λ_Li is closer to the upper limit λ_Rmax of the reflective band of the reflectors 12 and 14 than to its lower limit λ_Rmin. For example, at the temperature T_min, the wavelength λ_Li is comprised between 0.9 λ_Rmax and λ_Rmax. This constraint may be relaxed if the width of the reflective band is very much larger than its lower limit such as defined above.

Here, the reflectors 12 and 14 are for example Bragg gratings.

Between the reflectors 12 and 14, the laser source includes in succession the following photonic components in order from the reflector 12 to the reflector 14:

• • an optical waveguide 15 made of silicon in which the reflector 12 is produced,
  • a bandpass filter 22 able to select the wavelength λ_Li of operation of the laser source 10 from the various wavelengths λ_Rj that are able to be supported in the interior of the Fabry-Pérot cavity,
  • a coupler 26 that optically connects the waveguide 15 to an entrance of a waveguide 28 made of III-V gain medium,
  • a semiconductor optical amplifier 30 (SOA) produced in the waveguide 28 and able to generate and to amplify the optical signal resonating in the interior of the Fabry-Pérot cavity at each of the wavelengths λ_Rj,
  • a coupler 32 that optically connects an exit of the waveguide 28 to a waveguide 25, and
  • the optical waveguide 25 made of silicon in which the reflector 14 is produced.

Here, by waveguide made of silicon, what is meant is a waveguide the core of which is made of silicon. For example, the cladding of the waveguides made of silicon is made from another material such as, typically, silicon oxide.

The couplers 26 and 32 are for example adiabatic couplers. For a detailed description of an adiabatic coupler, the reader is referred to the following article: Amnon Yariv et al., “Supermode Si/III-V hybrid Lasers, optical amplifiers and modulators: proposal and analysis” Optics Express 9147, vol. 14, No. 15, Jul. 23, 2007.

Such an adiabatic coupler is, for example, obtained by modifying the width of the waveguide made of silicon with respect to the width of the waveguide made of III-V gain medium. Typically, for an adiabatic coupling of a waveguide made of silicon to a waveguide made of III-V material, the width of the waveguide made of silicon is gradually decreased as the waveguide made of III-V material is approached. In the opposite direction, to transfer by adiabatic coupling an optical signal from the waveguide made of III-V material to a waveguide made of silicon, the width of the waveguide made of silicon is for example gradually increased. In addition, typically toward the middle of the adiabatic coupler, the dimensions of the cross sections of the waveguides made of silicon and made of III-V material are generally such that their respective effective indices are equal. This is also true for an adiabatic coupling between hybrid Si/III-V material waveguides.

Preferably, the amplifier 30 is a wideband amplifier, i.e. one capable of generating and amplifying a large range of wavelengths. This range comprises the wavelength λ_Si. Typically, it is centred on this wavelength λ_Si at the temperature (T_max+T_min)/2. The width of this wavelength range at −3 dB is for example at least 10 nm or 25 nm or 35 nm and remains wide with the increase in temperature. For example, the III-V materials from which the amplifier 30 is made are those described in the following article: Dimitris Fitsios et al. “High-gain 1.3 μm GaInNAs semiconductor optical amplifier with enhanced temperature stability for all-optical processing at 10 Gb/s”, Applied Optics, vol. 54, n° 1, 1 Jan. 2015. The fact of producing the amplifier 30 as described in this article in addition allows a wideband amplifier that is stable in temperature to be obtained. This improves the operation of the laser source and notably this allows the power emitted by the laser source to be kept almost constant in all the operating temperature range [T_min; T_max]. In this case, the waveguide 28 and the amplifier 30 take the form of a stack 34 (FIG. 2) of alternating sublayers made of GaInNAs and of GaNAs interposed between a lower sublayer 35 (FIG. 2) and an upper sublayer 36 (FIG. 2) made of p-doped GaAs. The sublayer 35 is a sublayer made of III-V material of opposite dopant type to the upper layer 36. Here, it is a question of a sublayer made of n-doped GaAs.

The amplifier 30 comprises, in addition to the waveguide 28, a connection 37 (FIG. 2) making direct mechanical and electrical contact with the segment of the sublayer 35. The sublayer 36 makes mechanical and electrical contact with a connection 38 (FIG. 2) in order to electrically connect the lower portion of the amplifier 30 to a potential. When a DC current, called the “injection current”, higher than the threshold current of the laser is applied between the connections 37 and 38, the amplifier 30 generates and amplifies the optical signal that resonates in the interior of the Fabry-Pérot cavity.

In this embodiment, the filter 22 comprises solely an athermal primary filter 40. The filter 40 is here composed of a Mach-Zehnder interferometer 42.

The interferometer 42 comprises:

• • two arms 44 and 46 that are each able to guide the resonant optical signal, and
  • two forks 48 and 50.

In the direction of propagation F (FIG. 1) of the optical signal travelling from the reflector 12 to the reflector 14, the fork 48 distributes equally the optical signal reflected by the reflector 12 to the arms 44 and 46.

In FIG. 1 and the following figures, the vertical direction has been represented by a direction Z of an orthogonal coordinate system XYZ, where X and Y are horizontal directions. Here, the direction X is essentially parallel to the direction F of propagation of the optical signal.

In the direction F, the fork 50 combines the optical signals output from the arms 44 and 46 together to obtain an optical signal that propagates in the waveguide 25 to the reflector 14. In the direction of propagation opposite to the direction F, the roles of the forks 48 and 50 are inverted. For example, the forks 48 and 50 are multi-mode interferometers (MMIs).

The arm 44 is divided into three consecutive sections S_1,1, S_1,2 and S_1,3 in order in the direction F. Below, section n of arm m is designated S_m,n, where:

• • the index m is an identifier of the arm of the interferometer, and
  • the index n is an identifier of the section of this arm.

In this text, when the index m is equal to 1, this designates the lower arm 44. When the index m is equal to 2, this designates the upper arm 46. The index n is the numerical order of the section, starting at the fork 48 and travelling in the direction F. Thus, an index n equal to 1 corresponds to the first section encountered after the fork 48, the index n equal to 2 corresponds to the second section encountered starting from the fork 48 and so on.

Each section S_m,n is a waveguide. Each of these sections therefore has an effective index Neff_m,n and a nonzero length L_m,n. Typically, each section differs from the section that precedes it and from the section that follows it in:

• • the dimensions of its cross section,
  • the one or more materials used to produce its core, and/or
  • the one or more materials used to produce its cladding.

The effective index of a waveguide depends on the materials used to form the core and the cladding of this waveguide and also on the dimensions of the core and notably on the width and thickness of the core. The effective indices of two successive sections in the direction F are therefore different. Here, the effective index Neff_m,n of each section S_m,n is considered to be constant over all its length L_m,n.

The effective index of a waveguide may be determined, for a given temperature and for a given wavelength, by numerical simulation or experimentally. Here, when the effective indices of various sections are compared, it is a question of the effective indices determined at the same temperature and at the same wavelength. Typically, this wavelength is equal to λ_Si.

The effective index of a waveguide varies as a function of the temperature of this waveguide. Below, the coefficient that expresses the degree of variation in the index neff_m,n of the section S_m,n as a function of temperature is denoted dneff_m,n/dT. To a first approximation, in the temperature range DT, this coefficient dneff_m,n/dT may be considered to be constant.

Here, the section S_1,1 extends from the fork 48 to the coupler 26. The section S_1,1 is a waveguide made of silicon.

The section S_1,2 extends from the coupler 26 to the coupler 32. The section S_1,2 therefore corresponds to the waveguide 28 made of III-V gain media, in which waveguide the amplifier 30 is produced. The section S_1,2 is here called the “gain-generating section” because it has the capacity to amplify the optical signal.

The section S_1,3 extends from the coupler 32 to the fork 50. The section S_1,3 is a waveguide made of silicon that is identical to the waveguide made of silicon of section S_1,1 except that the length L_1,3 made it be different from the length L_1,1.

The arm 46 is divided into four consecutive sections S_2,1, S_2,2, S_2,3 and S_2,4 in order in the direction F. The sections S_2,1 and S_2,2 are identical to the sections S_1,1 and S_1,2, respectively. The arm 46 therefore comprises, just like the arm 44:

• • an optical coupler 52 that optically connects the waveguide 15 to an entrance of the waveguide 54 made of III-V gain media,
  • an optical amplifier 56 produced in the waveguide 54, and
  • an optical coupler 58 that optically connects an exit of the waveguide 54 to the waveguide 25.

The couplers 52, 58, the waveguide 54 and the amplifier 56 are identical to the couplers 26, 32, to the waveguide 28 and to the amplifier 30, respectively. In addition, the injection current that flows through the amplifier 56 here systematically has the same magnitude as the injection current that flows through the amplifier 30 at the same time. For example, to this end, the lower sublayer 35 and upper sublayer 36 of the amplifier 30 and of the amplifier 56 are common to these two amplifiers or are directly electrically connected to each other. Under these conditions, the indices neff_1,2 and neff_2,2 are identical and the coefficients dneff_1,2/dT and dneff_2,2/dT are also identical.

The section S_2,4 is identical to the section S_1,3 except that its length L_2,4 is larger than the length L_1,3. In FIG. 1, the fact that the length L_2,4 is larger than the length L_1,3 is schematically represented by a zigzag.

The section S_2,3 has a coefficient dneff_2,3/dT lower than the coefficients dneff_2,1/dT, dneff_2,2/dT and dneff_2,4/dT. To this end, according to a first solution, section S_2,3 is a waveguide made of silicon the width of the core of which is smaller than the width of the core of sections S_2,1 and S_2,4. According to a second solution, section S_2,3 is a waveguide the core of which is made from a material different from silicon such as, for example silicon nitride. In this case, the dimensions of the core of section S_2,3 may be identical to the dimensions of the cores made of silicon of sections S_2,1 and S_2,4. A third solution consists in using an another material for the cladding of section S_2,3, this material being different from that used for the cladding of sections S_2,1 and S_2,4. In addition, it is possible to combine these various solutions to obtain the desired coefficient dneff_2,3/dT.

In order for the interferometer 42 to have a passband that allows the wavelength λ_Li to be selected, the sections S_m,n are configured to meet the following condition (1):

$\sum _{n=1}^{N_2}{L}_{2,n}{\mathrm{neff}}_{2,n}-\sum _{n=1}^{N_1}{L}_{1,n}{\mathrm{neff}}_{1,n}=k_f{\lambda }_{\mathrm{Si}}$ where:

• • k_f is a preset integer number higher than or equal to 1,
  • N₁ and N₂ are the numbers of sections in the arms 44 and 46, respectively,
  • L_1,n and L_2,n n are the lengths of the nth sections of the arms 44 and 46, respectively,
  • neff_1,n and neff_2,n are the effective indices of the nth sections of the arms 44 and 46, respectively.

A high number of lengths L_m,n and of indices Neff_m,n allow condition (1) to be satisfied.

Here, the filter 40 is also to designed to be athermal. Thus its coefficient dλ_CF1/dT is close to zero, i.e. comprised between L_min and L_max. Below, a wavelength corresponding to a maximum of the transmission spectrum of the primary filter 40 is denoted λ_CF1. To this end, the sections S_m,n are configured to also meet the following condition (2):

$L_{\min }\le \sum _{n=1}^{N_2}{L}_{2,n}\frac{d{\mathrm{neff}}_{2,n}}{\mathrm{dT}}-\sum _{n=1}^{N_1}{L}_{1,n}\frac{d{\mathrm{neff}}_{1,n}}{\mathrm{dT}}\le L_{\max }$ where dneff_1,n/dT and dneff_2,n/dT are the variations in the effective indices of the nth sections of the arms 44 and 46, respectively.

For example, here the lengths L_m,n and the coefficients dneff_m,n/dT meet condition (2) with L_min=−7 pm/° C. and L_max=+7 pm/° C. and, preferably, with L_min=−2 pm/° C. and L_max=+2 pm/° C.

There are a high number of sets of values for the lengths L_m,n and the indices neff_m,n that meet both of the above conditions (1) and (2). Here, among all of these sets, it is the set of values or one of the sets of values that in addition allows a width Δλ₄₀ to be obtained for the passband of the primary filter 40 larger than Δλ_R that is chosen.

Under these conditions, because the interferometer 42 meets condition (2), the passband of the filter 22 does not move or practically does not move as a function of temperature. In contrast, in this embodiment, the wavelengths λ_Rj move as a function of temperature. In particular, the variation in temperature may be such that a wavelength λ_Rj precedingly located in the interior of the passband of the filter 22 moves and exits from this passband. In this case, since the width Δλ₄₀ is larger than or equal to the interval Δλ_R, when a wavelength λ_Rj exits from the passband of the filter 22, another wavelength λ_Rj−1 or λR_Rj+1 enters into the interior of this passband. Thus, in this embodiment, it is not necessary to employ a tuning device that moves the wavelengths λ_Rj so that there permanently exists a wavelength λ_Rj located at the centre of the passband of the filter 22.

To configure the sections S_m,n so that they meet the various above conditions, one solution consists, in a first step, in choosing the transverse dimensions of each section and the materials used for the core and the cladding of each section. Thus, the effective indices neff_m,n and the coefficients dneff_m,n/dT of each section are defined. Next, in a second step, the lengths L_m,n allowing the various preceding conditions to be met are sought. If it proves to be impossible to find satisfactory lengths, then the method returns to the first step in order to modify the effective indices of one or more sections. Such sections are for example dimensioned in accordance with the teaching given, in a different context, in the following article: Biswajeet Guha et al: “Minimizing temperature sensitivity of silicon Mach-Zehnder interferometers”, Optics Express, Jan. 15, 2010, pages 1879-1887.

FIG. 2 shows the laser source 10 in a vertical cross section passing through the arm 44. In this figure, the path followed by the optical signal between the reflectors 12 and 14 is represented by a double-headed arrow.

The laser source 10 comprises a substrate 70 that mainly lies in a horizontal plane called the “plane of the substrate”. The thickness of the substrate 70 is for example larger than 500 μm.

A layer 72 of single-crystal silicon encapsulated in silicon oxide is stacked on the upper face of the substrate 70. The reflectors 12 and 14, the waveguides 15 and 25 and the sections S_1,1, S_1,3, S_2,1, S_2,3 and S_2,4 and the lower portions of the couplers 26, 32, 52 and 58 are produced in the encapsulated silicon of this layer 72.

A thin layer 74, for example made of silicon oxide, is deposited on the layer 72. Lastly, the laser source 10 comprises a layer 76 made of gain medium encapsulated in, for example, silicon oxide. The waveguides 28 and 54, the amplifiers 30 and 56 and the upper portions of the couplers 26, 32, 52 and 58 are produced in this layer 76.

FIG. 3 shows the transmission spectrum 80 of the primary filter 40. The ordinate axis represents the transmittance of the filter 40 and the abscissa axis the wavelength associated with this transmittance. In this figure and the other figures showing a transmission spectrum, the transmittance of the filter is plotted on a logarithmic scale. Thus, a maximum of the transmission spectrum corresponds to the value 0 or to a value close to 0 in these graphs. In this example, the width Δλ₄₀ of a passband of the filter 40 is for example 5 nm. Since this width λ₄₀ is larger than the interval Δλ_R, the laser source 10 is multimode.

FIG. 4 shows a laser source 100 that is identical to the laser source 10, except that the filter 22 has been replaced by a filter 102. The filter 102 is identical to the filter 22, except that the primary filter 40 has been replaced by a primary filter 104 that comprises an interferometer 106 instead of the interferometer 42. The interferometer 106 is identical to the interferometer 42 except that the arm 46 has been replaced by an arm 108.

The arm 108 is identical to the arm 46 except that it comprises only three sections S_2,1, S_2,2 and S_2,3. Section S_2,1 of the arm 108 is identical to section S_1,1 of the arm 44. Section S_2,2 is identical to section S_2,2 of the interferometer 42 except that the amplifier 56 has been replaced by an amplifier 110 shorter than the amplifier 30. Thus, the length L_2,2 of section S_2,2 of the arm 108 is shorter than the length L_1,2. Section S_2,3 is identical to section S_1,3 except that the length L_2,3 is larger than or equal to the length L_1,3.

As in the preceding embodiment, the various sections S_m,n are arranged so that the lengths L_m,n and the indices neff_m,n meet both of conditions (1) and (2) described above. This laser source 100 operates just like the laser source 10 but without it being necessary to implement an additional section in one of the arms the effective index of which is lower than the indices neff_2,1 and neff_2,2.

FIG. 5 shows a laser source 130 that is identical to the laser source 10, except that the filter 22 has been replaced by a filter 132 that comprises a primary filter 134 instead of the primary filter 40. The filter 134 comprises a Mach-Zehnder interferometer 136 and a secondary filter 138.

The filter 138 is a bandpass filter typically comprising a plurality of passbands. The width Δλ₁₃₈ of each of these passbands is smaller than or equal to the interval Δλ_R. Here, the width Δλ₁₃₈ is equal to Δλ_R. This filter 138 has a coefficient dneff₁₃₈/dT higher than zero and generally higher than 15 pm/° C. or 50 pm/° C., where neff₁₃₈ is the effective index of the waveguide in which the filter 138 is produced. Here, the filter 138 is formed in a waveguide made of silicon. Thus, its coefficient dneff₁₃₈/dT is higher than 50 pm/° C.

In this embodiment, the filter 138 comprises a ring resonator 140. The ring 140 is optically coupled to section S_2,3 of the arm 46 by way of an optical coupling 142. Thus, a portion of the optical signal that passes through the arm 46 also passes through the filter 138. The coupling 142 is for example an evanescent coupling.

FIG. 6 shows the transmission spectrum 150 of the filter 138. This spectrum 150 comprises a plurality of passbands that are regularly spaced apart from one another by the distance d_FSR138. FIG. 6 also shows the transmission spectrum 152 of the interferometer 136. The spectrum 152 presents a plurality of transmission maximums that are regularly spaced apart from one another by the distance d_FSR136.

Here, the filter 138 is arranged so that the distance d_FSR138 is such that at a given time, in the interior of the interval ΔR, it is possible for there to be only a single passband of the filter 138 centred on a transmission maximum of the spectrum 152. To achieve this, either the filter 138 is configured so that the distance d_FSR138 is larger than ΔR or the filter 138 is configured so that the distance d_FSR138 is a non-integer multiple of the distance d_FSR136, as shown in FIG. 6. In FIG. 6, this band of the filter 138 centred on a transmission maximum of the spectrum 152 has been framed by a rectangle 154. Under these conditions, the filter 132 selects a single wavelength λ_Li contained in the interior of the passband framed by the rectangle 154. The laser source 130 is therefore a monomode laser source.

Here, the interferometer 136 is identical to the interferometer 42 except that the various sections S_m,n are arranged so as to meet condition (1) and the following condition (3):

$L_{\min }-\frac{d{\mathrm{neff}}_{138}}{\mathrm{dT}}\le \sum _{n=1}^{N_2}{L}_{2,n}\frac{d{\mathrm{neff}}_{2,n}}{\mathrm{dT}}-\sum _{n=1}^{N_1}{L}_{1,n}\frac{d{\mathrm{neff}}_{1,n}}{\mathrm{dT}}\le L_{\max }-\frac{d{\mathrm{neff}}_{138}}{\mathrm{dT}}$

In this embodiment, condition (2) is replaced by condition (3) above. When conditions (1) and (3) are met, the interferometer 136 compensates to a large extent the movement of the transmission spectrum of the filter 138 due to the temperature variations. The filter 134, which combines this interferometer 136 and the filter 138, is therefore athermal. Such sections are for example dimensioned in accordance with the teaching given, in a different context, in the following article: Biswajeet Guha et Al: “CMOS-compatible athermal silicon microring resonators”, Optics Express, Feb. 3, 2010, pages 3487-3493.

Here, the filter 134 also comprises a device 144 for adjusting the position of the passbands of the filter 138 with respect to the transmission maximum of the spectrum 152. For example, the device 144 is a heater able to locally heat the ring 140 without heating the arms 44 and 46 of the interferometer 136. As illustrated in FIG. 7, this device 144 allows the spectrum 150 to be moved with respect to the spectrum 152 until another passband of the filter 138 is centred on a transmission maximum of the spectrum 152. This allows, with a small movement of the spectrum 150, the value of the desired wavelength λ_Si of the laser source 130 to be substantially changed. In the example of FIGS. 6 and 7, a movement of less than 4 nm of the spectrum 150 allows the value of the wavelength λ_Si to be made to vary by about 10 nm.

FIG. 8 shows a laser source 150 that is identical to the laser source 10, except that the filter 22 has been replaced by a filter 152 that comprises, in addition to the primary filter 40, a tertiary filter 154 that is optically connected in series with the filter 40. To simplify FIG. 8, the details of the filter 40 have not been illustrated again.

The filter 154 is an athermal filter the transmission spectrum 156 of which (FIG. 9) comprises a plurality of passbands that are regularly spaced apart from one another by a distance d_SFR154. Preferably, the filter 154 is arranged so that the distance d_SFR154 is equal to p.d_SFR40, where p is an integer number higher than or equal to 2.

In addition, the filter 154 is arranged so that at least one transmission maximum of its spectrum 156 is coincident with a transmission maximum of the spectrum 80 of the filter 40. Under these conditions, the distance between two consecutive maximums of the transmission spectrum 158 (FIG. 9) of the filter 152 is equal to the distance d_SFR154. The width Δλ₁₅₂ of the passband of the filter 152 remains equal to the width Δλ₄₀ of a passband of the filter 40. Thus, the use of the filter 154 allows the passbands of the filter 152 to be spaced apart without increasing the width of these passbands. This therefore allows a number of wavelengths λ_Li lower than the number of wavelengths λ_Li selected by the filter 22 of the laser source 10 to be selected. By virtue of this, the laser source 150 may be made monomode.

Here, the filter 154 is an athermal Mach-Zehnder interferometer. For example, the filter 154 is identical to the filter 40 except that:

• • it is devoid of gain-generating section comprising a core made of gain medium,
  • the various sections are adjusted to meet conditions (1) and (2) and in addition to obtain a distance d_SFR154 equal to p times the distance d_SFR40.

Thus, the arm 44 of the filter 154 comprises a single section S_1,1 between the forks 48 and 50. This section S_1,1 is a waveguide the core of which is made of silicon. This section S_1,1 is identical, for example, to the concatenation of sections S_1,1 and S_1,3 of filter 40. The arm 46 of the filter 154 comprises three consecutive sections S_2,1, S_2,2 and S_2,3. Sections S_2,1, S_2,2 and S_2,3 of the filter 154 are for example identical to sections S_2,1, S_2,3 and S_2,4 of the filter 40, respectively.

FIG. 10 shows a laser source 170 that is identical to the laser source 10, except that the reflectors 12 and 14 have been replaced by the reflectors 172 and 174, respectively. The reflectors 172 and 174 are reflectors the reflective bands of which are narrow and narrower than the reflective bands of the reflectors 12 and 14. Here, the width of the reflective bands of the reflectors 172 and 174 is comprised between Δλ_R and d_SFR40. Thus, the reflectors 172 and 174 permit only wavelengths λ_Rj contained in the interior of a single passband of the filter 22. The reflectors 172 and 174 are, for example, Bragg gratings. The production of such narrow-reflective-band reflectors is for example described in the following article: G. H. Duan et al.: “Hybrid III-V on Silicon Lasers for Photonic Integrated Circuits on Silicon”, Journal of Selected Topics in Quantum Electronics, vol. 20, n° 4, July 2014, pages 1-13.

Preferably, the reflectors 172 and 174 are also insensitive to temperature variations. For example, to this end, the core of the reflectors 172 and 174 is made from a material of low thermo-optical coefficient such as silicon nitride.

FIG. 11 shows a laser source 180 that is identical to the laser source 10, except that the filter 22 has been replaced by a filter 182 in which the primary filter 40 has been replaced by a primary filter 184. The filter 184 is identical to the filter 40 except that the interferometer 42 has been replaced by an interferometer 186. The interferometer 186 is identical to the interferometer 42 except that the arm 46 is structurally identical to the arm 44. Thus, the arm 46 comprises only three consecutive sections S_2,1, S_2,2 and S_2,3. These sections S_2,1, S_2,2 and S_2,3 of the interferometer 186 are identical to sections S_1,1, S_1,2 and S_1,3 of the arm 44 of the same interferometer, respectively. Under these conditions, the total lengths of the arms 44 and 46 are equal. In addition, the transverse dimensions of the cores of sections S_1,2 and S_2,2 are identical. Here, by “transverse dimensions”, what is meant is the dimensions in a plane orthogonal to the direction of propagation of the optical signal. It is therefore in particular a question of the width and height of the core.

In contrast, in this embodiment, the injection currents I₁ and I₂ that flow through the amplifiers 30 and 56, respectively, are not identical. To this end, here, the connections 37 of the amplifiers 30 and 56 are electrically connected to respective power sources 190 and 192. Under these conditions, even though sections S_1,2 and S_2,2 are structurally identical, their effective indices neff_1,2 and neff_2,2 are different. Specifically, since the magnitudes of the currents I₁ and I₂ are different, the temperatures of the amplifiers 30, 56 are different and therefore their effective indices are different. Here, the sources 190 and 192 are adjusted so the condition (1) is met. In the particular case of the laser source 180, condition (1) is then written in the following form:L_2,2neff_2,2−L_1,2neff_1,2=k_fλ_Si where the lengths L_2,2 and L_1,2 are equal.

Since sections S_1,2 and S_2,2 are structurally identical, the coefficients dneff_1,2/dT and dneff_2,2/dT are equal. Thus, condition (2) described above is also met provided that the lengths L_2,2 and L_1,2 are equal. Therefore, the primary filter 184 is also athermal. The way in which the laser source 180 works may be deduced from the explanations given with respect to the laser source 10.

More precisely, as shown in FIG. 12, in a step 200 of emitting an optical signal at the wavelength λ_Li close to the desired wavelength λ_Si, the sources 190 and 192 inject the currents I₁ and I₂ of different magnitudes through the amplifiers 30 and 56, respectively.

FIG. 13 shows a laser source 220 that is identical to the laser source 130 except that the bandpass filter 132 has been replaced by a bandpass filter 222. The filter 222 is identical to the filter 132 except that the primary filter 134 has been replaced by a primary filter 224. The primary filter 224 is identical to the primary filter 134 except that the secondary filter 138 has been replaced by a secondary filter 228. The filter 228 is identical to the filter 138 except that the ring 140 has been replaced by a ring 230. The ring 230 is configured so that:

• • the width Δλ₂₂₈ of each passband of the filter 228 is equal to Δλ_R,
  • a maximum of the transmission spectrum 234 (FIG. 14) of the filter 228 coincides with a maximum of the transmission spectrum 152 of the interferometer 136, and
  • the distance d_FSR228 between two consecutive maximums of the spectrum 234 (FIG. 14) is equal to p.d_SFR136, where p is an integer number.

FIG. 14 shows the transmission spectra 152 and 234 in the particular case where p is equal to 1. As illustrated in FIG. 14, the filter 222 simultaneously selects a plurality of wavelengths λ_Li that are separated from one another by the distance d_FSR228. The laser source 220 is therefore multimode.

FIG. 15 shows a laser source 240 that is a combination of the embodiments of FIGS. 5 and 11. The laser source 240 is identical to the laser source 130 except that the bandpass filter 132 has been replaced by a bandpass filter 242. The filter 242 is identical to the filter 132 except that the primary filter 134 has been replaced by a primary filter 244. The primary filter 244 is identical to the primary filter 134 except that the interferometer 136 has been replaced by an interferometer 246.

The interferometer 246 is identical to the interferometer 186 except that the lengths L_1,2 and L_2,2 are different. As described above with reference to FIG. 11, the magnitudes of the DC currents I₁ and I₂ are different. Thus, the effective indices neff_1,2 and neff_2,2 are different. In contrast, the coefficients dneff_1,2/dT and dneff_2,2/dT are equal since the materials and the cross section of the cause of sections S_1,2 and S_2,2 are identical. In this particular case, condition (1) is then written in the following form:L_2,2neff_2,2−L_1,2neff_1,2=k_fλ_Si

Condition (3) for the primary filter 244 to be athermal is written in the following simplified form:

$L_{\min }-\frac{d{\mathrm{neff}}_{138}}{\mathrm{dT}}\le L_{2,2}\frac{d{\mathrm{neff}}_{2,2}}{\mathrm{dT}}-L_{1,2}\frac{d{\mathrm{neff}}_{1,2}}{\mathrm{dT}}\le L_{\max }-\frac{d{\mathrm{neff}}_{138}}{\mathrm{dT}}$ where the coefficients dneff_1,2/dT and dneff_2,2/dT are equal.

To find the magnitudes of the currents I₁ and I₂ and the lengths L_1,2 and L_2,2 that meet the above conditions (1) and (3), it is possible, for example, to first set the magnitudes of the currents I₁ and I₂ and then to seek the lengths L_1,2 and L_2,2 that meet the above conditions. It is also possible to proceed in the opposite way, i.e. for the lengths L_1,2 and L_2,2 to be set first and then the magnitudes of the currents I₁ and I₂ sought.

In the embodiment of FIG. 15, the tuning device 144 has been omitted.

Section III: Variants

Variants of the Bandpass Filter

As a variant, the laser source comprises a plurality of primary filters that are connected in series one after the other.

In another embodiment, the primary filter is not athermal. In this case, the various sections S_m,n have no need to be arranged to meet condition (2) or (3). If the primary filter is not athermal and this primary filter is connected in series with a tertiary filter, this tertiary filter is itself also not athermal. More precisely, in this case, the tertiary filter is arranged so that its coefficient dλ_CF3/dT is equal or practically equal to the coefficient dλ_CF1/dT of the primary filter.

The optical coupling 142 between the ring 140 and the arm 46 may occur elsewhere. For example, as a variant, this optical coupling occurs between section S_2,1 or S_2,4 and the ring 140, or even between section S_2,2 and the ring 140.

A tertiary filter such as the filter 154 may be implemented in the other embodiments described here and, in particular, in the embodiments of FIGS. 1, 4, 5, 10, 11 and 13.

Other embodiments of the tertiary filter are possible. For example, as a variant, the tertiary filter is identical to the primary filter 134 but the sections S_1,2 and S_2,2 comprising gain media are omitted. In another example embodiment, the tertiary filter is made athermal not using a Mach-Zehnder interferometer but using other means. For example, the core of the tertiary filter is made from a material the thermo-optical coefficient of which is low, i.e. at least two times lower than the thermo-optical coefficient of silicon. Thus, in another example, the tertiary filter is a ring resonator formed in a waveguide made of silicon nitride and coupled to the waveguide 15 or 25.

Other embodiments of the secondary filter 138 are possible. More precisely, any other filter having a transmission spectrum similar to the spectrum 150 and capable of being optically coupled to one of the sections of the arm 46 may be employed.

The secondary filter may also be produced in waveguides the core of which is made from a material the thermo-optical coefficient of which is low. For example, as a variant, the filter 138 is produced in a waveguide the core of which is made of silicon nitride and the cladding of which is made of silicon oxide.

Other Variants

The fact of injecting currents I₁ and I₂ of different magnitudes into the amplifiers 30 and 56 may be implemented in all the embodiments described here. In particular, currents I₁ and I₂ of different magnitudes may be injected through the amplifiers 30 and 56 of the embodiments of FIGS. 1, 4, 5, 8, 10 and 13. In this case, the other sections of the arms 44 and 46 are arranged so that conditions (1) and (2) or conditions (1) and (3) are met. The way in which these variants work is then the same as that of the laser source 180.

In the laser source 180, it is not necessary for sections S_1,1 and S_1,3 to be identical to sections S_2,1 and S_2,3, respectively. For example, as a variant, section S_1,1 is longer than section S_2,1 and, to compensate, section S_1,3 is shorter than section S_2,3. In fact, it is enough for the following condition (4) to be met for the laser source 180 to operate correctly: L_1,1neff_1,1+L_1,3neff_1,3=L_2,1neff_2,1+L_2,3neff_2,3. This condition (4) may be generalized without difficulty to the case where each arm comprises more than three sections S_m,n.

In the laser source 180, if it is not necessary for the bandpass filter to be athermal, then it is also not necessary for the lengths L_1,2 and L_2,2 to be equal. In this case, the laser source 108 operates correctly provided that sections S_1,2 and S_2,2 are configured so that the following condition is met:L_2,2neff_2,2−L_1,2neff_1,2=k_fλ_Si The above condition may be met even with different lengths L_1,2 and L_2,2.

The core of the gain-generating sections such as sections S_1,2 and S_2,2 may be produced using other materials. For example, the stack 34 may also be a stack of sublayers in alternation made of InP and InGaAsP. As a variant, the gain medium is deposited directly on an extension made of silicon of the waveguide 15 or 25. This superposition of the gain medium on the extension made of silicon then forms the core of an optical-gain-generating waveguide. In this case, the optical-gain-generating waveguide comprises III-V material but also silicon.

There are also other embodiments of the resonant cavity and, in particular, embodiments devoid of reflectors. For example, the cavity may take the form of a ring in which the optical signal rotates in such a way as, in each rotation, to pass through the gain medium of the interferometer. For example, this amounts to replacing the reflectors 12 and 14 with an additional waveguide that connects the left end of the resonant cavity to the right end while circumventing the gain media. In this case, the additional waveguide is, for example, produced in the layer 72.

Whatever the embodiment described here, to limit the consequences of the shift in the reflective band of the reflectors as a function of temperature, these reflectors may be made from a material of low thermo-optical coefficient such as silicon nitride.

The reflectors are not necessarily Bragg gratings. For example, a reflector may also be produced using a Sagnac loop.

Lastly, the system allowing the variations in the wavelength λ_Li as a function of temperature to be limited is not necessarily a passive system as described in the preceding embodiments. Thus, as a variant, the laser source may comprise:

• • a tuning device able to move the wavelengths λ_Rj as a function of an electrical control signal,
  • a sensor able to measure a physical quantity representative of the difference between the wavelength λ_Cf of the bandpass filter and the closest of the wavelengths λ_Rj, and
  • an electronic circuit able to generate the electrical control signal so as to permanently keep one of the wavelengths λ_Rj at the centre of the passband of the bandpass filter.The wavelength λ_Cf of the bandpass filter corresponds to a maximum of the transmission spectrum of this filter. Embodiments and implementations of these various components in a laser source are, for example, identical to those described in patent application US20180261976A.

Section IV: Advantages of the Embodiments

The fact of producing the gain-generating waveguide in each of the arms of an interferometer allows the length of the resonant cavity to be substantially decreased with respect to the case of a resonant cavity that allows an identical optical signal of the same power to be generated, but in which the gain-generating waveguide is located outside of the bandpass filter. Specifically, here, the functions of filtering and amplifying the optical signal are interleaved and carried out by one and the same component, namely a Mach-Zehnder interferometer. Thus, at equal characteristic and, in particular, at equal power, the resonant cavity of the laser sources described here is much shorter. This therefore allows, at equal characteristic, the bulk of the laser source to be decreased.

The fact of using a Mach-Zehnder interferometer arranged so as to meet condition (2) or (3) allows the movement of the passband of this filter as a function of temperature to be limited. This limits the variations in the wavelength λ_Li as a function of temperature. In particular, this limitation is obtained without it being necessary to make some of this bandpass filter from materials of low thermo-optical coefficient such as silicon nitride. Production of this bandpass filter is therefore simpler. In addition, the limitation of the movement of the passband of the filter as a function of temperature that it is possible to obtain is better than the limitation that it is possible to obtain by making, at least partially, the bandpass filter from materials of low thermo-optical coefficient.

When the arms of the interferometer comprise no optical coupling with a secondary filter such as the filter 138 and when this interferometer is arranged to meet condition (2), its transmission spectrum varies very little as a function of temperature. Therefore, the position of the passband of its transmission spectrum is practically constant. By virtue of this, it is possible to limit the variations in the wavelength λ_Li of the laser source as a function of temperature.

When the arm of the interferometer comprises an optical coupling with a secondary filter such as the filter 138 and when this interferometer is arranged to meet condition (3), the obtained primary filter is also athermal. Thus, this embodiment of the bandpass filter for its part also allows the variations in the wavelength λ_Li as a function of temperature to be limited. In addition, the secondary filter allows the width of the one or more passbands of the bandpass filter to be decreased or certain passbands to be limited with respect to the case where this bandpass filter is devoid of such a secondary filter. This therefore allows a monomode laser source to be obtained.

The fact that the secondary filter is a ring resonator allows the bulk of the laser source to be decreased.

The use of an athermal tertiary filter connected in series with the primary filter allows the passbands of the bandpass filter to be spaced apart while preserving small variations in the wavelength λ_Li as a function of temperature.

The use of an athermal Mach-Zehnder interferometer to produce the tertiary filter allows a tertiary filter the spectrum of which varies very little as a function of temperature to be obtained and therefore the variations in the wavelength λ_Li as a function of temperature to be very effectively limited.

The fact that the width of the passband of the bandpass filter is smaller than or equal to the interval Δλ_R allows a monomode laser source to be obtained. In contrast, the fact that the width of the passband of the bandpass filter is larger than the interval Δλ_R allows a multimode laser source to be obtained.

The use of reflectors the reflective band of which is narrow allows a monomode laser source to be obtained.

The presence in the transmission spectrum of the bandpass filter of a plurality of passbands spaced apart from one another by an interval equal to an integer multiple of the interval Δλ_R allows a multimode laser source to be obtained.

The fact that the passband of the bandpass filter has a width substantially identical to the interval Δλ_R allows a passive thermal stabilization of the wavelength λ_Li to be obtained. Specifically, if following heating of the resonant cavity the wavelength λ_Rj selected by the bandpass filter moves and exits from this passband, at the same time the preceding wavelength λ_Rj−1 or the following wavelength λ_Rj+1 enters into the interior of this passband. Thus, even in the absence of an active component for keeping one of the wavelengths λ_Rj at the centre of the passband of the bandpass filter, the variations in the wavelength λ_Li as a function of temperature are limited.

Advantages of the Currents I₁ and I₂ of Different Magnitudes

The fact of injecting currents I₁ and I₂ of different magnitudes into the amplifiers 30 and 56 makes it possible to obtain sections S_1,2 and S_2,2 the effective indices of which are different without having to modify the transverse dimensions of these sections or the materials from which they are made. This therefore simplifies the production of the bandpass filter.

If, in addition, the arms 44 and 46 are structurally identical and these arms are not coupled to a secondary filter, then the simple fact of meeting condition (1) necessarily leads to condition (2) being met. This therefore facilitates the production of a primary filter that is in addition athermal.

Claims

1. A semiconductor laser source able to emit at a wavelength λLi close to a desired wavelength λSi, this laser source comprising: a set of waveguides optically coupled to one another and forming an optical cavity able to make an optical signal resonate at a plurality of possible resonant frequencies, the possible wavelengths λRj of these possible resonant frequencies being regularly spaced apart from one another by an interval ΔλR, this set comprising: at least one waveguide in which a bandpass filter is produced, the transmission spectrum of this bandpass filter comprising a passband able to select, among the various possible wavelengths λRj, the wavelength λLi closest to the wavelength λSi, andat least one gain-generating waveguide able to generate optical gain at each wavelength λLi selected by the bandpass filter, this gain-generating waveguide comprising a III-V gain medium,

2. The source according to claim 1, wherein: the primary filter comprises a secondary filter the transmission spectrum of which comprises a narrow passband located in the interior of a wider passband of the interferometer, this secondary filter having a coefficient dλCF2/dT higher than Lmax, where:λCF2 is the wavelength corresponding to a maximum of the transmission spectrum of the secondary filter,the coefficient dλCF2/dT is the variation in the wavelength λCF2 as a function of temperature, andLmax is equal to 50 pm/° C. or 15 pm/° C.,one of the first and second arms of the interferometer comprises an optical coupling to the secondary filter, andthe lengths of the various sections of the first and second aims are such that a coefficient dλCMZI/dT is comprised between −dλCF2/dT+Lmin and −dλCF2/dT+Lmax, where:λCMZI is a wavelength corresponding to a maximum of the transmission spectrum of the interferometer,the coefficient dλCMZI/dT is the variation in the wavelength λCMZI as a function of temperature,Lmin is equal to −50 pm/° C. or −15 pm/° C., andLmax is equal to +50 pm/° C. or +15 pm/° C.

3. The source according to claim 2, wherein the secondary filter comprises a ring resonator optically coupled to one of the first and second arms of the interferometer, the narrow passband of the transmission spectrum of this ring resonator being centred on the wavelength λCMZI and having a width ΔλF2 smaller than or equal to the interval ΔλR.

4. The source according to claim 2, wherein: the passbands of the secondary filter are spaced apart from one another by a distance such that there exists, in the interior of an interval ΔR, only a single passband of the secondary filter centred on a transmission maximum of the transmission spectrum of the interferometer of the primary filter, where the interval ΔR is the smallest wavelength range containing all the possible wavelengths λRj permitted by the resonant cavity, andthe width of the passband of the secondary filter centred on the transmission maximum of the transmission spectrum of the interferometer of the primary filter is smaller than or equal to the interval ΔλR.

5. The source according to claim 2, wherein: the passbands of the secondary filter are separated from one another by a distance equal to an integer multiple of the interval ΔλR and smaller than an interval ΔR, where the interval ΔR is the smallest wavelength range containing all the possible wavelengths λRj permitted by the resonant cavity, anda passband of the secondary filter is centred on a transmission maximum of the transmission spectrum of the interferometer of the primary filter.

6. The source according to claim 1, wherein: none of the arms of the interferometer comprises an optical coupling to a secondary filter,the lengths of the various sections of the first and second aims are such that a coefficient dλCMZI/dT is comprised between Lmin and Lmax in a temperature range extending from 20° to 100°, where:λCMZI is a wavelength corresponding to a maximum of the transmission spectrum of the interferometer,the coefficient dλCMZI/dT is the variation in the wavelength λCMZI as a function of temperature,Lmin is equal to −50 pm/° C. or −15 pm/° C., andLmax is equal to +50 pm/° C. or +15 pm/° C.

7. The source according to claim 6, wherein the lengths of the sections of the first and second arms meet the following athermicity condition for the wavelength λSi and in a temperature range extending from 20° to 100° C.:

8. The source according to claim 6, wherein the first and second aims are structurally identical.

9. The source according to claim 1, wherein: the bandpass filter comprises a tertiary filter that is optically connected in series with the primary filter, the transmission spectrum of this tertiary filter comprising a plurality of passbands that are spaced apart from one another by a distance dSFR154 at least two times larger than a distance dSFR40, where the distance dSFR40 is the distance that separates from one another the passbands of the transmission spectrum of the interferometer of the primary filter,this tertiary filter has a coefficient dλCF3/dT comprised between Lmin and Lmax, where: λCF3 is a wavelength corresponding to a maximum of the transmission spectrum of the tertiary filter,the coefficient dλCF3/dT is the variation in the wavelength λCF3 as a function of temperature,Lmin is equal to −50 pm/° C. or −15 pm/° C., andLmax is equal to +50 pm/° C. or +15 pm/° C.

10. The source according to claim 9, wherein: the passbands of the tertiary filter are spaced apart from one another by a distance such that there exists, in the interior of an interval ΔR, only a single passband of the tertiary filter centred on a transmission maximum of the transmission spectrum of the interferometer of the primary filter, where the interval ΔR is the smallest wavelength range containing all the possible wavelengths λRj permitted by the resonant cavity, andthe width of the passband of the tertiary filter centred on the transmission maximum of the transmission spectrum of the interferometer of the primary filter is smaller than or equal to the interval ΔλR.

11. The source according to claim 9, wherein: the passbands of the tertiary filter are separated from one another by a distance equal to an integer multiple of the interval ΔλR and smaller than an interval ΔR, where the interval ΔR is the smallest wavelength range containing all the possible wavelengths λRj permitted by the resonant cavity, anda passband of the tertiary filter is centred on a transmission maximum of the transmission spectrum of the interferometer of the primary filter.

12. A method for emitting an optical signal at a wavelength λLi close to a desired wavelength λSi, this method comprising: a) providing a laser source comprising:a set of waveguides optically coupled to one another and forming an optical cavity able to make an optical signal resonate at a plurality of possible resonant frequencies, the possible wavelengths λRj of these possible resonant frequencies being regularly spaced apart from one another by an interval ΔλR, this set comprising: at least one waveguide in which a bandpass filter is produced, the transmission spectrum of this bandpass filter comprising a passband able to select, among the various possible wavelengths λRj, the wavelength λLi closest to the wavelength λSi, andat least one gain-generating waveguide able to generate optical gain at each wavelength λLi selected by the bandpass filter, this gain-generating waveguide comprising a III-V gain medium,wherein:the bandpass filter comprises a primary filter containing a Mach-Zehnder interferometer, this interferometer comprising: first and second arms,a first fork via which the optical signal is distributed between the first and second arms, anda second fork via which the optical signals of the first and second arms are combined together,each of the first and second arms being divided into a plurality of consecutive sections that are immediately contiguous to one another, each section comprising a core able to guide the optical signal and each section having an effective index at the wavelength λSi, the effective index of each section located immediately after a preceding section being different from the effective index of this preceding section,the first and second arms each comprise a gain-generating section, the cores of each of these gain-generating sections being made from the same III-V gain medium and the transverse dimensions of the cores of these gain-generating sections being identical, these gain-generating sections thus forming first and second gain-generating waveguides, respectively,b) during the emission of the optical signal, first and second DC currents are injected through the first and second gain-generating waveguides, respectively, the magnitudes of these first and second currents being constant and different and such that the following condition is met: