The invention relates to a semiconductor laser source able to emit at least one wavelength λLi, and to a process for manufacturing this laser source.
Laser sources known to the inventors include:
In these laser sources known to the inventors, the filter and the reflectors are produced in waveguides made of silicon and the filter selects only a single wavelength.
When the operating temperature of the known laser sources varies, the emission wavelength of these laser sources also varies at a rate of about 0.07 nm/° C. For this reason, it is said that these known laser sources are sensitive to temperature and that the emission wavelength of the laser source drifts as a function of temperature.
To remedy this drawback, it is known to associate the laser source with a device for regulating its operating temperature in order to keep it at a constant temperature. However, such regulating devices are bulky and consume a lot of power.
Prior art is also known from:
US2015/180201A1,
NOBUHIDE FUJIOKA et al: “Compact and Low Power Consumption Hybrid Integrated Wavelength Tunable Laser Module Using Silicon Waveguide Resonators”, Journal of Lightwave Technology, IEEE Service Center, US, vol. 28, no. 21, Jan. 11, 2010,
US2014/153600A1,
US2014/153601A1,
GB2522252A,
US2009/135891A1.
The invention aims to remedy this drawback by providing a semiconductor laser source that is less sensitive to temperature variations while remaining simple to manufacture. One of its subjects is therefore a semiconductor laser source according to Claim 1.
To limit the temperature drift of the claimed laser source, the proposed solution consists in producing the filter in a waveguide that instead of being made of silicon is made from a material that is less sensitive to temperature, and in integrating this filtering to the structure of the laser source. Specifically, the fact that the waveguide in which the filter is produced is made from a material the variation in the refractive index as a function of temperature of which is at least two times less rapid than the variation in the refractive index of silicon allows the temperature drift of the filter to be limited. In addition, the tuning device, the sensor and the electronic circuit allow one wavelength λRj to be kept permanently at the centre of each passband of the filter that selects a wavelength λLi. The consequence of this is to decrease the amplitude of the drift in the emission wavelength λLi of the laser source with temperature without it being necessary to have recourse to a device for regulating the temperature of the laser source.
The integration into the laser source of this filter is made possible by the use of a laser-source structure comprising three waveguides, namely a waveguide made of silicon, a waveguide made of III-V gain material and a waveguide made of material that is less sensitive to temperature. This three-waveguide structure especially makes it possible to get around the difficulty according to which it is difficult to achieve adiabatic coupling directly between the waveguide made of material that is less sensitive to temperature and the waveguide made of III-V gain material. Specifically, the materials of these two waveguides generally have large differences in optical refractive index. In the claimed laser source, it is therefore proposed to use a waveguide made of silicon to couple, on the one hand, the waveguide made of III-V gain material to the waveguide made of silicon and, on the other hand, the waveguide made of silicon to the waveguide made of material that is less sensitive to temperature. Thus, the waveguides made of III-V gain material and of material that is less sensitive to temperature are coupled to each other by way of the waveguide made of silicon. In addition, the presence of the waveguide made of silicon may be taken advantage of to manufacture a particularly effective tuning device. Specifically, the variation in the refractive index of silicon as a function of temperature is substantial and it is also possible to generate free carriers in a waveguide made of silicon simply. These two effects modify the phase of any optical signal propagating through the silicon waveguide
Embodiments of this laser source may have one or more of the features of the dependent claims.
These embodiments of the laser source furthermore have the following advantages:
Another subject of the invention is a process for manufacturing the claimed laser source.
The invention will be better understood on reading the following description, which is given merely by way of nonlimiting example and with reference to the appended drawings, in which:
In these figures, the same references are used to designate the same elements. In the rest of this description, features and functions that are well known to those skilled in the art are not described in detail.
The reason why the production of the filter in a waveguide made of material that is less sensitive to temperature decreases the sensitivity of the laser source to temperature will firstly be explained in detail with reference to
The reflection band 4 is characterized by its width ΔR and its central wavelength λCR. The width ΔR is the width of the reflection band 4 at −3 dB. It is a question of the band of wavelengths containing all the wavelengths λRj able to be reflected by the front and back reflectors with a power higher than or equal to 50% of the maximum power Imax reflected by these reflectors. The power Imax is equal to the power of the optical signal reflected at the wavelength λRj at which this power is maximum.
The central wavelength λCR is the wavelength located in the middle of the reflection band 4.
In the context of a Fabry-Pérot cavity, the various wavelengths λRj at which this cavity is liable to resonate are regularly spaced apart from one another by an interval ΔλR. In
In the context of a monochromatic laser source that emits at a single wavelength λLi, the filter selects from all of the possible wavelengths λRj one of these wavelengths λRj. The selected wavelength λRj is equal to the emission wavelength λLi of the laser source. To this end, the filter has a single narrow passband 6 centred on the wavelength λLi. This passband 6 is characterized by its width Δλf and by its central wavelength λCf. The width Δλf is the width of the passband 6 at −3 dB. The central wavelength λCf is the wavelength located in the middle of the passband 6. To select the single wavelength λLi, the width Δλf is smaller than or equal to the interval ΔλR. To select the wavelength λLi, one of the wavelengths λRj, typically the closest to λCf, is tuned to be equal to this wavelength λCf. This wavelength λRj that is equal to the wavelength λCf corresponds to the wavelength λLi.
The graph in
In known laser sources, the optical signal is guided between the front and back reflectors by waveguides made of silicon and of III-V gain materials. The reflectors and the filter are produced in waveguides made of silicon, and the gain function is provided in the waveguide made of III-V materials. Under these conditions, it is known that the wavelengths λRj and the wavelength λCR move as a function of temperature at about 0.07 nm/° C. The wavelength λCf also moves at about 0.07 nm/° C.
The graph in
To limit this drift in the wavelength λLi, the proposed solution consists in producing the filter in a waveguide that instead of being made of silicon is made from a material that is less sensitive to temperature, and in integrating this filter into the structure of the laser source. Here, a material that is less sensitive to temperature is a material the variation dnf/dT of the refractive index of which as a function of temperature is at least two times lower, under the same conditions, as the variation dnSi/dT in the refractive index of silicon as a function of temperature. Typically, the variation dnSi/dT is equal to 2.3×10−4/° C. to within plus or minus 20%. Thus, the variation dnf/dT is at least lower than 1×10−4/° C. and, preferably, lower than 0.5×10− 4/° C. Below, the particular case where the material that is less sensitive to temperature is silicon nitride (Si3N4), the variation dnf/dT of which is equal to 0.4×10−4/° C., is described. However, as described in detail at the end of this description, other materials are possible. The refractive index of silicon nitride varies as a function of temperature 6 to 7 times less rapidly than that of silicon. Therefore, the central wavelength λCf of the filter produced in a waveguide made of silicon nitride moves, as a function of temperature, only at 0.01 nm/° C.
As will be seen below, in the laser source described, a tuning device, a sensor and an electronic controlling circuit are used to automatically and correctly centre the closest of the wavelengths λRj on the centre of the passband 6 of the filter in order to obtain a laser-source emission at the wavelength λLi. Specifically, since the passband 6 of the filter moves much less rapidly than the wavelengths λRj, if nothing is done to correctly centre one of the wavelengths λRj on the centre of the passband 6, this passband 6 may contain no wavelength λRj or a wavelength λRj that is poorly centred, thereby suppressing the emission of the laser source or greatly limiting the power of the emitted optical signal.
The reflectors 12 and 14 are wideband reflectors. In this embodiment, this means that the width ΔR of the reflection band 4 of the reflectors 12 and 14 is strictly larger than Δλf+DT×(dλCR/dT), where:
The range of operating temperatures of a laser source is often at the present time chosen so that, whatever the operating temperature of the laser source in this range, the wavelength λLi does not vary by more than 0.35 nm. Thus, for example, the width DT of this range may be larger than 10° C. or 30° C. To meet this criterion, here the range of operating temperatures is chosen to be comprised between +20° C. and +55° C. The width DT is therefore here equal to 35° C. Below, the lowest and highest temperatures of the range of operating temperatures are denoted Tmin and Tmax, respectively. In this embodiment, the reflectors 12 and 14 are produced in a waveguide made of silicon. Thus, the variation dλCR/dT is here equal to 0.07 nm/° C. As described with reference to
The reflectors 12 and 14 are also designed so that, at the temperature Tmin, the wavelength λLi is closer to the upper limit λRmax of the reflection band 4 than its lower limit λRmin. For example, at the temperature Tmin, the wavelength λLi is comprised between 0.9 ΔRmax and λRmax. This constraint may be relaxed if the width ΔR is very much larger than Δλf+DT×(dλCR/dT). Specifically, in the latter case, in order for the wavelength λLi to be contained inside the reflection band 4 whatever the operating temperature of the laser source 10 between Tmin and Tmax, it is not necessary for the wavelength λLi to be close to λRmax.
Here, the reflectors 12 and 14 are for example reflectors such as Bragg gratings.
Between the reflectors 12 and 14, the laser source includes the following photonic components in succession from the reflector 12 to the reflector 14:
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 III-V gain material, the reader may consult the following article: B. Ben Bakir et al., “Hybrid Si/III-V lasers with adiabatic coupling”, 2011.
For a detailed description of an adiabatic coupler, the reader may refer 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.
In particular, it will be recalled that an adiabatic coupler is able to convert an optical mode guided in a first waveguide made of silicon into an optical mode guided in a second waveguide made of III-V gain material. In particular, an adiabatic coupler is able to transfer almost all the optical signal present in the first waveguide to a second waveguide located above or below, without reflection. For example, the power, of the optical signal, transferred to the second waveguide is higher than or equal to 95% of the power of the optical signal circulating in the first waveguide. Such an adiabatic coupler is, for example, obtained by modifying the width of the first waveguide with respect to the width of the second waveguide. Typically, for adiabatic coupling from 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 it approaches the waveguide made of III-V material. Conversely, to transfer by adiabatic coupling an optical signal from the waveguide made of III-V material to the first waveguide made of silicon, the width of the waveguide made of silicon is for example gradually increased. In addition, the waveguides made of silicon and of III-V material generally have a width such that their respective propagation indices are equal. This is also true for adiabatic coupling between hybrid Si/III-V waveguides.
In this embodiment, the filter 22 is a ring resonator filter, the ring of which is produced in a waveguide 50 (
Evanescent coupling between a first and a second waveguide transfers a fraction of the power of the optical signal present in the first waveguide to the second waveguide. This fraction of the power is adjusted, for example, by adjusting the distance between the first and second waveguides. Here, the geometry and distance between the waveguides 15, 25 and the waveguide 50 in which the ring of the filter 22 is produced are adjusted so that the power, of the optical signal, transferred between the waveguide made of silicon and the waveguide 50 is comprised between 5% and 25% and, preferably, equal to 10% to within plus or minus 3% or 1%. The geometries of the waveguides are modified to make the propagation indices equal in both waveguides. Here, the width of the waveguides 15 and 25 made of silicon is for example decreased over a length L1 and L2, respectively, and that of the waveguide 50 made of Si3N4 is not modified. To adjust the coupled power, the distance between the two waveguides 15 and 50 and 25 and 50 is adjusted in a horizontal direction to a calculated distance d1, d2, respectively, as is described in more detail with reference to
The use of evanescent coupling between the waveguide 50 of the ring and the waveguides 15 and 25 made of silicon allows the bulk of the laser source to be limited while avoiding recourse to adiabatic couplers and to additional waveguides made of silicon nitride.
Adiabatic or evanescent coupling also has the advantage of limiting the reflection of the optical signals at the optical interface between two waveguides.
To generate the electrical signal for controlling the tuning device 16, the laser source 10 also includes:
For this purpose, here, the sensor 40 measures the power of the optical signal emitted by the laser source 10. In this embodiment, the resonant optical signal may exit the laser source 10 in two possible ways, namely either by passing through the reflector 14, or by passing through the reflector 12. Given that the reflection coefficient of the reflector 12 is higher than that of the reflector 14, the optical signal that exits by passing through the reflector 12 has a much lower power than the signal exiting through the reflector 14. Conventionally, the optical signal that exits by passing through the reflector 14 is called the “useful optical signal” and the signal that exits by passing through the reflector 12 is called the “optical control signal” or the “optical monitoring signal”. Here, the sensor 40 measures the power of the optical signal that passes through the reflector 12. For example, the sensor 40 includes a photo detector that measures the power of the optical signal. The power thus measured is transmitted to the electronics circuit 42. This measured power is representative of the distance between the wavelengths λCf and the closest of the wavelengths λRj. Specifically, the measured power is maximum when the wavelength λCf is aligned with one of the wavelengths λRj. This power decreases monotonically as the wavelength λRj gets further from the wavelength λCf provided that it does not leave an interval of width ΔλR centred on the wavelength λCf.
The electronics circuit 42 constructs the controlling signal that allows a wavelength λRj to be kept at the centre of the passband 6 depending on the physical quantity measured by the sensor 40. For this purpose, it is electrically connected to the sensor 40. In addition, it is electrically connected to the tuning device 16 in order to transmit thereto the generated electrical controlling signal. Typically, the sensor 40 is produced on the same substrate as the other photonic components of the laser source 10 and for example in the III-V gain material. The electronics circuit 42 is often added to this substrate.
As shown in
Here, the dimensions of the filter 22 are determined by numerical simulation or experimentally so that the interval Δvf is strictly wider than the width ΔR of the reflection band 4. Thus, only one of the passbands of the filter 22, i.e. here the passband 6, is located inside the reflection band 4, whatever the operating temperature. Thus, the filter 22 allows just one single wavelength ΔλRj to be selected and hence the laser source 10 is a monochromatic laser source. In the graph in
For example, the values of the wavelength λCf and of the interval Δvf are set on the basis of the following parameters: the perimeter of the ring, the propagation index neff of the waveguide 50, the group index ng of the waveguide 50 (i.e. the refractive group index) and the order of resonance K. By way of illustration, the wavelength λCf and the interval Δvf are estimated using the following relationships:
where “perimeter” is the perimeter of the ring produced in the waveguide 50 and λ is the wavelength, of the optical signal, for which the wavelength λCf and the interval Δvf are calculated.
The width Δλf may be determined from the propagation losses in the waveguide 50 of the ring and the coupling coefficients of the evanescent coupling between the waveguide 50 and the waveguides 15 and 25.
The graph in
The dimensions of the filter 22 are also defined to obtain the properties desired for the evanescent coupling between the waveguide 50 of the ring and the waveguides 15 and 25 and in particular the desired degree of transfer of power between the waveguides 15, 25 and the waveguide 50. For example, the defined dimensions of the filter 22 are especially chosen from the following dimensions:
By way of illustration, typically:
For example, to obtain a degree of transfer of power of 10% between the waveguide 15 and the waveguide 50 of the ring, the following are possible dimensions: R=30 μm, L1=60 μm, eG=300 nm, eAG=50 nm, LG=180 nm, eA=500 nm, eAF=50 nm, LA=700 nm, d1=1.1 μm.
In
The adiabatic couplers 26 and 32 are partially produced in the waveguide 25 and in the waveguide 28.
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 λLi. Typically, it is centred on this wavelength λLi at the temperature (Tmax+Tmin)/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 large with the increase in temperature. For example, the III-V materials constituting the amplifier 30 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 and 10 Gb/s” Applied Optics, May 2015 vol. 54, no. 1, 1 Jan. 2015. Producing the amplifier 30 in the way described in this article allows, in addition, a wideband amplifier that is stable with temperature to be obtained. This improves the operation of the laser source and in particular this allows the power emitted by the laser source to be kept almost constant over the entire range of operating temperatures [Tmin; Tmax]. In this case, the waveguide 28 and the amplifier 30 take the form of a stack of alternating sublayers made of GaInNAs and of GaNAs, which sublayers are interposed between a lower sublayer 70, and an upper sublayer made of p-doped GaAs. The sublayer 70 is a sublayer made of III-V material doped oppositely to the upper sublayer. For example, here, it is a question of a sublayer made of n-doped GaAs.
The amplifier 30 includes, in addition to the waveguide 28, a contact 74 making direct mechanical and electrical contact with the section of the sublayer 70 located under the stack of sublayers made of GaInNAs and GaNAs. The sublayer made of p-doped GaAs makes mechanical and electrical contact with a contact 76 for electrically connecting the upper portion of the amplifier 30 to a potential. When a current higher than the threshold current of the laser is applied between the contacts 74 and 76, the amplifier 30 generates and amplifies the optical signal that resonates inside the Fabry-Pérot cavity.
The tuning device 16 is here a heater able to heat the waveguide 15 in order to move the wavelengths λRj.
In this embodiment, the tuning device 16 comprises a resistor 80 electrically connected to two electrical contacts 82 and 84 allowing a current to be made to flow through this resistor 80 so as to convert electrical energy into heat. These contacts 82 and 84 are electrically connected to a voltage or current source that is controlled by the electronics circuit 42 depending on the measurements of the sensor 40. The control of the tuning device 16 therefore here consists in adjusting the electrical power that passes through the resistor 80. Making a current flow through the resistor 80 allows the waveguide 15 to be heated and therefore the wavelengths λRj to be moved.
The resistor 80 is here a strip produced in the sublayer 70. This strip is therefore a strip made of n-doped GaAs. In this embodiment, it is logically located above the waveguide 15 the index variation dnSi/dT of which is clearly higher than the index variation dnf/dT.
The tuning device 16, the waveguide 28 and the amplifier 30 are covered with a protective covering 90 that mechanically insulates them from the exterior. Only the contacts 74, 76, 82 and 84 protrude beyond the covering 90. For example, the covering 90 is made of silicon nitride.
The path of the resonant optical signal in the laser source 10 is represented in
A process for manufacturing the laser source 10 will now be described with reference to the process in
In a step 100, the process starts with the supply of an SOI (silicon-on-insulator) substrate 102. This substrate 102 (
In a step 110, the waveguides 15 and 25 and the reflectors 12 and 14 are manufactured in the layer 66 made of silicon. For example, they are manufactured by photolithography and etching of this layer 66. In this step, the portions of the couplers 26 and 32 that are located in the layer 66 are also produced.
In a step 112, the layer 66 is encapsulated in a layer 114 (
In a step 116, a layer made of silicon nitride is deposited on the top face of the layer 114. Next, this layer made of silicon nitride is etched to form the waveguide 50 then encapsulated in silicon oxide. In
In a step 120, a substrate 122 (
In a step 124, the layer 104 made of silicon is removed, and the layer 106 is thinned in order to leave only a thin intermediate layer 126 (
In a step 128, the layer 68 made of III-V gain material is bonded or deposited on the layer 126. For example, the layer 68 (
Once step 128 has been carried out, in a step 130, the layer 68 is etched to manufacture the waveguide 28, the amplifier 30 and the resistor 80. Typically, in a first etch, the upper sublayers of the layer 68 are etched to structure the amplifier 30. Next, in a second etch, the sublayer 70 is etched to finalize the structuring of the amplifier 30 and to produce the resistor 80.
Lastly, in a step 132, the covering 90 and the contacts 74, 76, 82 and 84 are produced. The laser source, a cross-sectional view of which is shown in
Next, it continues with a step 152 in which a layer that is identical to the layer 68 described above is bonded to or deposited on the layer 114 (
Next, in a step 158, this layer is etched to manufacture the waveguide 28, the amplifier 30, the resistor 80 and an additional resistor 154 (
The resistor 154 is used to produce a tuning device 156 (
The resistor 154 is located above the reflector 14.
In a step 160, the protective covering 90 is produced. In this step, the waveguide 50 is manufactured in the covering 90, which is made of silicon nitride. For example, the waveguide 50 is manufactured by photolithography and etching of the covering 90 in proximity to the ends of the waveguides 15 and 25 (
In a step 162, the contacts 74, 76, 82 and 84 are manufactured. The contacts 164 and 166 that mechanically and electrically connect the resistor 154 are also manufactured in order to allow a current to be made to flow through this resistor 154. The combination of the resistor 154 and the contacts 164 and 166 forms the tuning device 156. Thus, the laser source 150 shown in
In
The laser source 150 is identical to the laser source 10, except that the waveguide 50 and the waveguide 28 are both placed on the same side of the layer 66. In addition, in the laser source 150, the reflectors 12 and 14 are turned upward and not downward, as in the laser source 10. This allows the output of the reflector 14 to be connected to an optical fibre positioned above the laser source 150.
The laser source 150 operates in the same way as the laser source 10, except that the electronic circuit 42 is here in addition suitable for controlling the tuning device 156 so as to tune, in addition, the reflectors 12 and 14.
The reflectors 182 and 184 are identical to the reflectors 12 and 14, respectively, except that they are produced in the same layer made of silicon nitride as the waveguide 50. For example, here the reflector 182 is produced at one end of a waveguide 186 made of silicon nitride, the other end of which waveguide is optically connected to the waveguide 50 of the ring of the filter 22 by evanescent coupling. Therefore, the waveguide 15 is omitted. The waveguide 50 of the ring is optically connected to the waveguide 25 by evanescent coupling as described above. Similarly, the reflector 184 is produced at the end of a waveguide 188 made of silicon nitride, the other end of which waveguide is optically connected, by adiabatic coupling, to the waveguide 25.
In this case, preferably, the width ΔR of the reflection band of the reflectors 182 and 184 is larger than DT×(dλCf/dT).
The couplers that optically connect the waveguides 25 and 28 together are adiabatic couplers.
It will be noted that producing the reflectors 182 and 184 in a waveguide made of silicon nitride allows their bandwidth to be decreased. This also allows the interval Δvf of the filter 22 needed to obtain a monochromatic laser source to be decreased.
In
The laser source 180 has the advantage that the reflection band 4 of the reflectors 182 and 184 moves much less rapidly than in the laser sources 10 or 150. Specifically, in this embodiment, the variation dλCR/dT in the central wavelength λCR is equal to the variation dλCf/dT. Thus, whatever the variation in temperature, the passband 6 always remains inside the reflection band 4 and does not move with respect to the upper and lower limits of this reflection band 4. The width ΔR of the reflection band 4 may thus be small. Typically, the width ΔR is strictly larger than DT× dλCf/dT+Δλf.
The filter 222 is identical to the filter 22 except that the dimensions of the filter 22 have been modified so that it has a plurality of passbands simultaneously located inside the reflection band of the reflectors 224 and 226. In this embodiment, the filter 222 is a ring resonator produced in the waveguide 50, which is optically connected to the waveguides 15 and 25 by evanescent coupling.
In the graph in
Preferably, the width ΔR of the reflection band 234 of the reflectors 224 and 226 is:
The operation of the laser source 220 is identical to that of the laser source 10 except that it simultaneously emits N wavelengths that are here denoted λL1, λL2 and λL3.
The heater 237 in addition includes a lateral arm 244 produced from silicon in the layer 60. The arm 244 is in thermal continuity with the waveguide 15, that is the arm 244 is directly thermally coupled with the waveguide 15. This arm 244 extends in a transverse horizontal direction from the waveguide 15 leftward. Its thickness in the vertical direction is smaller than that of the waveguide 15. More precisely, its dimensions are such that the optical signal cannot propagate inside the arm 244 and remains mainly confined inside a zone 245 that encircles the waveguide 15. At its end opposite the waveguide 15, in this embodiment, the arm 244 has a distal section 246 facing the resistor 240. The distal section 246 is separated from the resistor 240 by a thin layer of silicon oxide of thickness eSP. typically, the thickness eSP is equal to the thickness of silicon oxide that separates the waveguide 28 from the waveguide 25. Thus, the distal portion 246 forms a bulge. Here, the thickness of the distal section 246 is equal to the thickness of the waveguide 15. The arm 244 forms a single block of material with the waveguide 15.
During the operation of the device 236, the resistor 240 preferentially heats the distal section 246, then, by thermal conduction via the arm 244, the heat is transmitted to the waveguide 15. The waveguide 15 is therefore heated in turn, thereby making its refractive index vary. In this embodiment, the resistor 240 is placed outside of the zone 245 in which the resonant signal circulates. In particular, the resistor 240 is not located, in the vertical direction, above and facing the waveguide 15. Thus, optical losses are limited even though it is still possible to effectively heat this waveguide. Specifically, by way of comparison, in the device 16 the resistor 80 is facing the waveguide 15 and located inside the zone 245, thereby causing optical losses. Here, the heater 237 does not heat the waveguide 15 directly but mainly heats it via the arm 244 that conveys no optical signal.
Typically, the coupling factors of the evanescent coupling between the waveguide 262 and the waveguides 264, 266 are the same as those described in the case of the filter 22. The two waveguides 264, 266 are optically connected via adiabatic couplers 268 and 270, respectively, to the waveguides 15 and 25. In this embodiment, the waveguide 262 is not directly optically connected to the waveguides 15 and 25, but connected to these waveguides via waveguides 264 and 266 made of silicon nitride.
Many other embodiments are possible. For example, the tuning device 16 or 156 may be omitted. Conversely, an additional tuning device, such as the tuning device 156, may be added to any of the embodiments described above to move the reflection band of the reflectors.
The tuning device is not necessarily a heater. For example, it is also possible to use by way of tuning device a p-n junction produced in one of the waveguides made of silicon. The refractive index of the silicon level with this p-n junction varies depending on the bias of this junction. The electronic signal for controlling the tuning device then makes the bias of this p-n junction vary. This way of making the refractive index of the silicon vary is for example described in more detail in the following article: G. T. Reed et al., “Silicon optical modulators”, Nature Photonics, vol. 4, August 2010.
The reflection coefficients of the front and back reflectors may also be equal.
If the filter 22 of the laser source 180 is replaced by the filter 222, then the width ΔR of the reflection band of the reflectors 182 and 184 may be decreased. However, preferably, it is larger than N× Δvf+DT×(dλCf/dT), where N is an integer higher than or equal to two and equal to the number of wavelengths λRj selected by the filter 222.
Other embodiments of the waveguide 28 and the amplifier 30 are possible. Typically, the waveguide 28 includes in succession, starting from the substrate 60: a doped lower sublayer, a stack of quantum wells made of quaternary material and a doped upper sublayer the doping of which is of opposite type to the doping of the lower sublayer. The doped sublayer 70 may be produced from other material such as from p- or n-doped InP. In this case, the stack of quantum wells includes sublayers made of InGaAsP or of GaInNAs, etc. For an exemplary embodiment of a wideband amplifier that is stable with temperature and in which the sublayer 70 is made of InP, the reader may refer to the following article: K. Morito et al. “GaInNAs/InP Tensile-Strained Bulk Polarization-Insensitive SOA”, ECOC2006, IEEE. When the lower and upper sublayers are made of GaAs, the stack of quantum wells may be produced with sublayers made of AlGaAs. The doping type of the upper and lower sublayers may be inverted.
As a variant, the waveguide 28 and the amplifier 30 are optically connected to the waveguides 15 and 25 by evanescent coupling. In this case, the adiabatic couplers 26 and 32 are omitted. Such an optical connection between the amplifier and waveguides made of silicon is for example described in the following article: A. W. Fang et al., “Electrically pumped hybrid AlGaInAs-silicon evanescent laser”, Optics Express, vol. 14, pp. 9203-9210 (2006).
Other embodiments of the sensor 40 are also possible. For example, as a variant, the sensor 40 may be replaced by a sensor that measures the operating temperature of the laser source. Specifically, with each operating temperature of the laser source, it is possible to associate a difference between the wavelengths λCf1 and λRj. In this case, for example, the electronics circuit 42 includes a pre-recorded table that associates, with a plurality of operating temperatures of the laser source, the characteristics of the electrical controlling signal to be generated to keep the wavelength λLi at the centre of the passband 6 of the filter. For example, this pre-recorded table is constructed experimentally. The temperature of the laser source may be measured using a transducer, such as a p-n junction. Specifically, the electrical properties of a p-n junction vary as a function of temperature.
As a variant, the heater 238 is omitted. In this case, the device 236 is no longer symmetric with respect to the plane 239. In another variant, the bulge 246 is omitted.
Other embodiments of the filter are possible. For example, the ring resonator may be replaced by an array waveguide grating (AWG). In the latter case, the filter is coupled to the waveguide made of silicon by adiabatic couplers such as the adiabatic couplers 268 and 270 shown in
It is also possible to use materials other than silicon nitride. For example, it is also possible to use, by way of material that is less sensitive to temperature, aluminium nitride.
A laser source that emits it a plurality of wavelengths λLi may also be produced as described, for example, with reference to
Other manufacturing processes are possible. For example, the process described with reference to
In another embodiment, the intermediate layer 126 is obtained by removing the layer 106 in its entirety, then by depositing on the exposed layer 66 a layer of silicon oxide.
Lastly, the order in which the various photonics components are arranged inside the Fabry-Pérot cavity may be modified. For example, the filter 22 may be placed between the coupler 32 and the reflector 14.
The tuning device 236 may be used in any semiconductor photonics system in which it is necessary to heat an optical waveguide while limiting losses. In particular, it may be used to heat a waveguide of a system other than a laser source. This other system may be a modulator for modulating the phase or intensity of an optical signal.
Number | Date | Country | Kind |
---|---|---|---|
16 50171 | Jan 2016 | FR | national |