This invention relates to lasers in general, and more particularly to widely tunable, single mode emission semiconductor lasers.
Single mode emission semiconductor lasers provide light sources for many modern applications, ranging from spectroscopy to telecommunications and far beyond. These single mode emission semiconductor lasers allow for electronic control of the light output and the emitted wavelength λ.
For many applications, it is desirable to have a wide tuning range of the laser's emitted wavelength across an interval λmin . . . λmax, whose width is measured by the tuning ratio r, where r is
A typical semiconductor laser emits many wavelengths, so additional wavelength selection mechanisms are necessary in order to achieve single mode operation (i.e., the emission of a single wavelength λ). There exists a wide range of wavelength selection mechanisms, but fully electronic wavelength selection mechanisms are typically limited to a tuning ratio of r≈0.01 or less. Wavelength selection mechanisms that support a wider range of wavelengths typically possess a very large parameter space (i.e., they have a large number of tuning parameters), which complicates the characterization process and the controlled operation of these wavelength selection mechanisms, and thus renders these approaches unfeasible for mass production.
More particularly, semiconductor lasers without a wavelength selection mechanism emit light in many modes (i.e., at many wavelengths). For many applications it is highly desirable to have a laser that emits only a single light mode (i.e., light at only a single wavelength). There exist many wavelength selection mechanisms, but it is often highly desirable to have a monolithic (i.e., on-chip) wavelength selection mechanism. Monolithic (i.e., on-chip) solutions are preferred for several reasons, such as their size and their mechanical stability. Several monolithic wavelength selection mechanisms are known, e.g., distributed feedback (DFB) gratings, distributed Bragg reflector (DBR) structures and Vernier tuning of coupled cavities or resonators, etc.
It is often highly desirable to construct a laser that possesses a wide tuning range for its wavelength. To achieve this, one needs to be able to control the wavelength selection mechanism. For many applications it is highly desirable to have a fully electronic wavelength selection mechanism. These fully electronic wavelength selection mechanisms are typically based on a tuning of the refractive index of the laser's waveguide material using variations of (i) the temperature of the waveguide material, and (ii) the density of the injected current. The best side mode suppression ratios are achieved with lasers having distributed feedback (DFB) gratings and lasers with distributed Bragg reflectors (DBRs). With these wavelength selection mechanisms, it is possible to achieve a tuning ratio of up to 1 percent for typical operating conditions (i.e., r≈0.01).
Wider tuning ranges have been demonstrated where the wavelength selection mechanisms (i.e., the DFB grating or the DBR period) are adjusted not only to select a single wavelength, but also to select two or more preferred wavelengths. It is then possible to build multi-segment lasers such that the absolute tuning of the effective refractive index of the waveguide material (achieved by variations in the temperature of the waveguide material and variations in the density of the injected current) essentially allows for a continuous tuning of the output wavelength, and the relative tuning (e.g., through the selection of phase relations) allows selection of the grating or reflector that is effectively turned on. The combination of these two effects (continuous tuning and relative tuning) achieve, in principle, much higher tuning ranges but, in practice, fabrication and characterization complexities increase significantly as the tuning range increases.
A different approach to wavelength selection uses the Vernier effect. In this approach, one typically defines two cavities, each possessing its own Fabry-Perot wavelength comb. Tuning is then achieved by manipulating the refractive index of the two cavities individually (i.e., using variations of the temperature of the waveguide material and the density of the injected current). This selects an emission wavelength that is contained in both Fabry-Perot wavelength combs and within the range of the gain of the gain material used in the laser. This typically allows for a much wider tuning range than where wavelength selection is based on DFB or DBR wavelength selection mechanisms.
Thus, known monolithic single mode semiconductor lasers with high side mode suppression ratios (i.e., DFB or DBR lasers) possess small tuning ratios of about 1 percent (i.e., r≈0.01). Those lasers that possess DFB gratings or DBRs and which are able to select more than one wavelength are usually very complex in their operation. This makes the operating characteristics of the laser very susceptible to variations in the fabrication process. Finding the desired protocol of operating parameters leads to a very complex characterization process in a high-dimensional parameter space (i.e., it is a highly complex procedure to characterize the particular waveguide material temperature and the particular injected current density for each segment of the laser). This leads to a characterization process that requires the collection of spectra on a fine grid in a high dimensional parameter space. The number of elements of this grid increases exponentially with the number of independent parameters, which makes it unfeasible to fabricate and control DFB or DBR lasers with the number of segments that is necessary to stably select many wavelengths.
Furthermore, known mechanisms that implement wavelength selection through the Vernier tuning principle suffer from a similar problem: since one wants to simultaneously control the output wavelength, output intensity and side mode suppression ratio, one needs many control parameters (in particular, waveguide material temperature and injection current densities). This again leads to a very complex characterization process in a very high dimensional parameter space, which makes mass production of devices with full simultaneous control of output wavelength, output intensity and side mode suppression ratio unfeasible.
The present invention provides a widely tunable, single mode emission semiconductor laser with fully electronic control of the light intensity and an electronically tunable wavelength with a typical tuning ratio r≈0.1 or more, i.e., more than ten times the tuning ratio of conventional single mode semiconductor lasers. In contrast to known devices with wide tuning ranges, the laser of the present invention possesses a very simple characterization process and an established process route and is thus suitable for mass production.
Generally speaking, the present invention comprises a monolithic semiconductor laser that provides single mode light emission, a wide tuning range for the wavelength, and simultaneous independent control of the light intensity. The semiconductor laser is characterized by a semiconductor material that contains a layer structure suitable for laser emission, with two or more linearly-aligned ridge waveguides structured thereon so as to provide two linearly-aligned coupled cavities. Furthermore, the laser is characterized by (i) the individual controllability of three heating currents, wherein two power resistors are structured in close proximity to the two linear ridge waveguides, and one power resistor is mounted on the base of the chip (i.e., the monolithic semiconductor laser), and (ii) the individual controllability of the laser currents in the two coupled cavities. Furthermore, the laser is characterized by the use of a gain material that possesses wide gain tuning with variations in the temperature of the waveguide material.
More particularly, the present invention is based on the Vernier tuning principle, but the laser is structured in such a way that the characterization process needs very few measured tuning parameters for each laser, while enabling the simultaneous control of wavelength and light output at a high side mode suppression ratio.
The preferred embodiment of the present invention comprises a laser which comprises two linearly-arranged coupled cavities which contain contacts to individually control the two laser currents (i.e., the injection current densities for each coupled cavity). In addition, the laser comprises three heaters: the first two heaters are resistors that are structured on surfaces which are in close proximity to the two coupled cavities, such that each resistor possesses an effective heat contact with only one of the two coupled cavities. The third heater is mounted on the bottom of the laser chip and allows for the control of the temperature of the bulk of the laser chip on a millisecond timescale.
The key observation regarding this design is that the wavelength of a mode is given by a simple relation of the form:
λ=λi+aT+b1I1+b2I2+c1H12+c2H22, (Eq. 2)
where a, b1, b2, c1, c2 are the tuning parameters of the coupled cavities that describe the laser tuning with the base temperature T of the chip, the laser currents I1, I2 which are the injected current densities through the two coupled cavities, and the heating currents H1, H2 which are applied to the two coupled cavities through their adjacent heating resistors. The base wavelength λi depends on the relative refractive index of the two coupled cavities and depends, essentially, on only the effective temperatures of the two coupled cavities. Note that the tuning parameters a, b1, b2, c1, c2 are the same for all base modes λi. The tuning parameters (a, b1, b2, c1, c2) can be measured by measuring spectra around a reference point; both the main mode and the side modes will tune with the tuning parameters (a, b1, b2, c1, c2). The base wavelengths λi are measured after knowing the tuning parameters (a, b1, b2, c1, c2), by measuring series with a larger step size. The gist is that one needs to measure only a few tuning parameter lines with high precision to find the tuning parameters (a, b1, b2, c1, c2), and identify a fairly coarse grid to find the base wavelengths λi. This drastically reduces the number of operation points that need to be characterized in order to characterize the laser. In this way, one is able to characterize the laser by measuring only five tuning parameters (i.e., a, b1, b2, c1, c2) and a discrete set of base wavelengths λi. The number of base wavelengths depends on the geometry of laser and the particular application of the laser and can be as small as four, but can also reach a number of about 20 base modes for particular purposes. This allows for an effective characterization of the lasers and thus facilitates mass production of the devices.
In connection with the aforementioned Eq. 2, it should be appreciated that:
Thus, where a given base mode λi is known for a particular combination of settings T, I1, I2, H1 and H2, and where the coefficients a, b1, b2, c1 and c2 have been determined for the laser, tuning the laser from a known base mode λi to a target wavelength λ may be achieved by appropriately adjusting the values of one or more of T, I1, I2, H1 and/or H2 so as to shift the output of the laser from a base wavelength λ to the target wavelength λ. Of course, it will be appreciated that as one or more of T, I1, I2, H1 and/or H2 is adjusted, the refractive index of one or both of the coupled cavities may change, which may in turn require further changes in the values of T, I1, I2, H1 and/or H2 in order to achieve the target wavelength λ.
It will also be appreciated that many combinations of T, I1, I2, H1 and H2 may be used to produce a given base wavelength λi, and that many different base wavelengths λi may be identified for a laser of a particular construction. In practice, it is possible to meaningfully characterize the laser by determining five or so base wavelengths λi, or in some cases more base wavelengths λi (perhaps up to as many as twenty base wavelengths λi for particular purposes). It will be appreciated that each specific base wavelength λi is produced by using a particular combination of T, I1, I2, H1 and H2, and the particular combination of T, I1, I2, H1 and H2 which is chosen to produce a base wavelength λi reflects factors which are relevant to the intended use of the laser.
Thus, in use, where it is desired to tune the laser to some wavelength λ, and where the laser has already been characterized to identify a plurality of base wavelengths λi, the user selects a base wavelength λi which is close to the desired wavelength λ, and sets the laser to that base wavelength λi by setting the values of T, I1, I2, H1 and H2 to those values which produce the selected base wavelength λi, and then the user adjusts T, I1, I2, H1 and/or H2 as needed in order to move the output of the laser from the selected base wavelength λi to the desired wavelength λ.
Two advantages of the present invention over the state of the art are that the present invention provides a widely tunable, single mode emission semiconductor laser that is produced as a monolithic device, and features fully electronic simultaneous control of wavelength and light output at a high side mode suppression ratio. Monolithic devices are preferred as they are small, stable and very robust when mounted in a suitable package. Fully electronic control is highly responsive (i.e., very fast), reliable and controllable with relatively inexpensive, readily-available electronic controls. In particular, wavelength tuning based on the cavity heaters and laser currents can be achieved on a sub-millisecond time scale, while wavelength tuning based on the substrate heater can be achieved on a millisecond time scale.
Another advantage of the present invention over the state of the art is that the laser achieves a wide tuning range while allowing for a rather compact characterization process based on the measurement of only five tuning parameters (i.e., the aforementioned a, b1, b2, c1, c2 in Eq. 2) and a few base wavelengths (i.e., λi in Eq. 2).
And another advantage of the present invention stems from the utilization of a monolithic fabrication process: the laser can be easily integrated in a photonic integrated circuit which may contain an on-chip wavelength- and side mode-analyzer. The simple characterization process then allows for an online calibration mode.
In one preferred form of the invention, there is provided a widely tunable, single mode emission semiconductor laser comprising:
In another preferred form of the invention, there is provided a method for generating light of a selected wavelength, the method comprising:
These and other objects and features of the present invention will be more fully disclosed or rendered obvious by the following detailed description of the preferred embodiments of the invention, which is to be considered together with the accompanying drawings wherein like numbers refer to like parts, and further wherein:
In one preferred form of the invention, and looking now at
The two linear ridge waveguides 15, 20 preferably have a width and height that are comparable to the target wavelengths, and the two linear ridge waveguides 15, 20 are preferably spaced from one another by a distance D which is about one-half the target wavelength. The two linear ridge waveguides 15, 20 are structured such that they guide the laser mode to the four facets 25, 30, 35, 40. The two linear ridge waveguides 15, 20 define two coupled cavities 45, 50, respectively, with coupled cavity 45 comprising facets 25, 30, and with coupled cavity 50 comprising facets 35, 40. A gap 55 separates the two coupled cavities 45, 50.
The two linear ridge waveguides 15, 20 are preferably generated through a material removal process effected from the top (such as chemical or physical etching). The remaining material (i.e., the material remaining after etching) then defines the two linear ridge waveguides 15, 20. The lengths of the coupled cavities 45, 50 are typically defined through a second material removal process, typically etching, such that facets 25, 30, 35, 40, as well as the gap 55 between the two linear ridge waveguides 15, 20, are defined with a precision on the order of 10 nanometers. The typical length of the two coupled cavities 45, 50 is between from about 80 nm to about 800 nm. It is also possible to structure Distributed Bragg Reflectors (DBRs) on facets 25, 30, 35, 40 to control the reflectivity of the facets and gap 55 beyond the values that are obtainable with a single etching step.
Two heating resistors 60, 65 are structured in close proximity to the sides of the two linear ridge waveguides 15, 20, i.e., heating resistor 60 extends along linear ridge waveguide 15 and heating resistor 65 extends along linear ridge waveguide 20. The distance is typically one or a few micrometers (note: this refers to the lateral distance between the linear ridge waveguides 15, 20 and the heating resistors 60, 65, respectively—ideally one would like these to be as close as possible, but the minimum distance is limited by the fact that one needs to isolate the heating resistors 60, 65 from the laser contacts (i.e., p-contacts 70, 75, respectively, see below), therefore a distance of a few, e.g., two, micrometers, is necessary). The two heating resistors 60, 65 are typically processed from a highly conducting material (e.g., titanium (Ti), platinum (Pt), or gold (Au)) and their dimensions are arranged such that the total resistance is on the order of a few Ohms. It is particularly preferred that the heating resistors 60, 65 possess a “meandering” structure, which enhances the heat contact between the two linear ridge waveguides 15, 20 and the two heating resistors 60, 65, respectively, at constant resistivity (see, for example,
The laser 5 comprises two p-contacts 70, 75 for receiving the laser currents on the tops of the two linear ridge waveguides 15, 20, i.e., p-contact 70 extends along linear ridge waveguide 15 and p-contact 75 extends along linear ridge waveguide 20. The two p-contacts 70, 75 are typically processed with a highly conductive material, e.g., gold (Au), and linked to two laser current contact pads 80, 85, respectively (i.e., contact pad 80 is connected to the p-contact 70 for linear ridge waveguide 15 and contact pad 85 is connected to the p-contact 75 for linear ridge waveguide 20). The two laser current contact pads 80, 85 are positioned on the opposite side of laser 5 to the two heating resistors 60, 65. The two n-contacts 86, 87 for the two coupled cavities 45, 50 are processed on the bottom of the chip (i.e., at the bottom of semiconductor substrate 10, see
The two heating resistors 60, 65 are provided with their own contact pads, preferably on the side opposite to the laser current contact pads 80, 85, and the two heating resistors 60, 65 may share a common ground contact pad. By way of example but not limitation, three heating resistor contact pads 90, 95, 100 may be provided, with contact pad 95 being a common ground contact pad, so that contact pads 90, 95 are used to supply current to heating resistor 60, and contact pads 100, 95 are used to supply current to heating resistor 65.
The semiconductor substrate 10 of the laser 5 is preferably mounted on a heat spreader plate 105 that contains a heating resistor 110 for heating the bulk of semiconductor substrate 10. Contact pads 115, 120 are used to supply electrical current to heating resistor 110.
If desired, contact pads 125, 130, 135, 140, 145 may be provided on heat spreader 105, with the contact pads providing easy electrical connection to various components of semiconductor laser 5, e.g., various ones of contact pads 125, 130, 135, 140, 145 may be connected to various ones of contact pad 80 for p-contact 70 of coupled cavity 45, contact pad 85 for p-contact 75 of coupled cavity 50, contact pad 90 for heating resistor 60, contact pad 95 (the common ground) for heating resistors 60, 65, and contact pad 100 for heating resistor 65, whereby to provide easy electrical connection to these components. Note that the various contact pads do not require a particular bonding scheme, and may be adjusted according to a particular application. However, it is generally preferred that the common ground for the heating resistors 60, 65 is bonded to the ground for the chip heating resistor 110. In other words, it is generally preferred that contact pad 120 (the ground) of chip heating resistor 110 is connected to contact pad 140 for the common ground 95 for heating resistors 60, 65.
Thus it will be seen that, in the preferred embodiment of the present invention, there is provided a semiconductor laser 5 which comprises a semiconductor substrate 10, a first linear ridge waveguide 15 which forms a first coupled cavity 45, and a second linear ridge waveguide 20 which forms a second coupled cavity 50, with first coupled cavity 45 being separated from second coupled cavity 50 by a gap 55. Coupled cavities 45, 50 comprise p-contacts 70, 75 and n-contacts (not shown, and preferably in the form of a common n-contact) for allowing laser currents I1, I2 to be injected into coupled cavities 45, 50, respectively. Coupled cavities 45, 50 also comprise heating resistors 60, 65, respectively, for heating the coupled cavities when heating currents H1, H2 are applied to heating resistors 60, 65, respectively. A heating resistor 110 is provided for heating the semiconductor substrate 10 of laser 5 so as to regulate the base temperature T of the chip (i.e., semiconductor substrate 10).
The foregoing construction provides semiconductor laser 5 with five “controls” which may be used to regulate the output wavelength of laser 5, i.e., the base temperature T of the chip (which is controlled by current passed through heating resistor 110), the laser currents I1, I2 which are injected through the two coupled cavities 45, 50 (i.e., by means of the p-contacts and n-contacts for the two coupled cavities), and the heating currents H1, H2 which are applied to the two coupled cavities through the adjacent heating resistors 60, 65.
In essence, in the preferred embodiment of the present invention, semiconductor laser 5 comprises two coupled cavities 45, 50 which are heated by heating resistors 60, 65, respectively. One of the coupled cavities 45, 50 possesses a length that ensures that the distance between two of its Fabry-Perot modes is smaller than the tuning range. This is typically the shorter cavity (i.e., cavity 50 in the construction shown in
In the preferred form of the invention, the semiconductor laser 5 is disposed (e.g., soldered) on a heat spreader 105.
The preferred mode of operation of the laser is to characterize the laser according to Eq. 2. Each shape of the base mode possesses a distinct base wavelength λi. Then a combination of the heater currents H1, H2 can be tuned to sweep a range of the wavelengths. Thereafter, the base temperature T of the laser is adjusted by changing the current of the heater which heats the semiconductor substrate (i.e., the chip). This shifts the gain maximum. The laser currents I1, I2 are then adjusted to achieve the desired light output level; this is necessary since the laser efficiency depends critically on the base temperature T of the chip (i.e., semiconductor substrate 10). The ratio between the two laser currents I1, I2 is then adjusted to achieve an optimal side mode suppression ratio.
Then another sweep may be performed, e.g., by varying the combination of heater currents H1, H2. This process may be repeated until the entire gain of the laser material is covered. The advantage of fixing the base temperature T and then performing another sweep is that this tuning process is very fast, since the slowest part in tuning is achieving a change in the base temperature T of the chip (i.e., semiconductor substrate 10).
Additional Information on the Characterization of the Semiconductor Laser
1. Background
There exists a plethora of possibilities for the construction of widely tunable semiconductor lasers that are able to achieve stable single mode operation over a wide range of wavelengths. The requirement that these lasers are suitable for series production sets imposes a number of constraints, in particular:
2. Effective Model
Inspection of the equilibrium states reveals that in a given laser they depend only on a small set of specific combinations of certain macroscopically accessible quantities. These quantities are the center ko and width Δk of the gain of the laser material, as well as the effective refractive indices ni of the cavity segments. These quantities are in turn functions that depend essentially only on the base temperature T of the laser chip as well as the laser Ii and heater Hi current levels of the laser segments, i.e.,
ko(T,I1, . . . ,In) and ni(T,I1, . . . ,In),
while Δk is practically constant. A sufficiently good model for the gain maximum is
while it is practically sufficient to assume that the gain width Δk remains constant. The effective refractive indices of the laser segments can be described by
ni=nio+aiT+biIi+ciHi2, (2)
while the effective refractive index of the gap is described by
The effective model of the laser then consists of investigating which light modes, labelled by their vacuum wave number
of the laser minimize the mirror losses, while lying within
from the gain maximum ko. The modes that minimize the mirror losses are the so-called Fabry-Perot modes. These are standing waves whose modulus of the electric field possess minima at both laser facets. These modes are exponentially amplified through stimulated emission in the gain material with an effective exponent that is given by the average number of gain length lg that a single photon in the mode is reflected between the cavities. See
3. Characterization
The identification of the modes with the highest reflectivities as the modes that appear in the spectra of the laser allows one to characterize the laser effectively by measuring a few spectra. It is fairly simple to identify the tuning of the entire cavity by considering the side mode spectrum of a laser. The Fabry-Perot modes of the entire laser appear as side modes of the coupled cavity laser. The vacuum wave number of the Fabry Perot modes is
These modes are seen as the main and side modes (modes with signal above −50 dB) in
The measurement of the tuning of the heater currents H1 and H2 for the linear ridge waveguides 15 and 20, respectively, is analogous to the measurement of the laser currents I1 and I2.
The final piece of information that one needs to characterize the laser is to determine its gain tuning. This can be achieved by taking a temperature series.
4. Characterization of a New Laser Design Vs. Characterization of Mass Produced Lasers of the Same Design
It should be appreciated that there is typically a difference between the characterization of a new laser design and the characterization of mass produced lasers of the same design. A new laser design typically requires a detailed characterization, because one has to understand how process variations and material properties transform into the effective model parameters. However, once the process variations of a particular design and process route are understood and lasers of the same design are mass produced, one has a much simpler characterization problem. In this case one only needs to measure the “fine tuning”. The simple laser characterization can then be reduced to an even shorter program (but of the same type as the program described above) which only measures a few properties of the laser to find the tuning parameters and locations of the base wavelengths with sufficient precision.
Alternative Constructions and Modes of Operation
In alternative constructions, one can consider lasers with more than two coupled cavities, lasers that possess only one heater at a coupled cavity, and lasers with several coupled cavities that possess heaters at each cavity or at only a subset of the cavities. The important part of the design is that the mode equation for the light mode possesses a scaling symmetry, such that a simple formula, analogous to that of Eq. 2, effectively describes the tuning behavior of the laser.
In alternative modes of operation, the operating mode may be adjusted due to the needs of the application. In particular, it may be desired to only scan a discrete set of wavelengths rather than to perform a number of wavelength sweeps.
It should be understood that many additional changes in the details, materials, steps and arrangements of parts, which have been herein described and illustrated in order to explain the nature of the present invention, may be made by those skilled in the art while still remaining within the principles and scope of the invention.
This patent application claims benefit of prior U.S. Provisional Patent Application Ser. No. 63/004,816, filed Apr. 3, 1920 by Automotive Coalition For Traffic Safety, Inc. and Tim Koslowski et al. for WIDELY TUNABLE, SINGLE MODE EMISSION SEMICONDUCTOR LASER, which patent application is hereby incorporated herein by reference.
Number | Name | Date | Kind |
---|---|---|---|
20080025355 | Hu et al. | Jan 2008 | A1 |
20080159341 | Patel | Jul 2008 | A1 |
20140010248 | Larson | Jan 2014 | A1 |
20150139256 | Dallesasse et al. | May 2015 | A1 |
20160211646 | Avramescu et al. | Jul 2016 | A1 |
20170104307 | Lin | Apr 2017 | A1 |
20170194763 | Treese | Jul 2017 | A1 |
20210036482 | Möhrle | Feb 2021 | A1 |
Number | Date | Country |
---|---|---|
WO 2017119944 | Jul 2017 | WO |
WO 2019211203 | Nov 2019 | WO |
Entry |
---|
Debrégeas, H., Widely Tunable Laser Diodes, Springer Series in Optical Sciences, Feb. 2017, vol. 161, pp. 168-177. |
Gao, H. et al., Cleaved-coupled nanowire lasers, Proceedings of the National Academy of Sciences of the United States of America, vol. 110, No. 3, Jan. 15, 2013, pp. 865-869. |
Number | Date | Country | |
---|---|---|---|
20210313771 A1 | Oct 2021 | US |
Number | Date | Country | |
---|---|---|---|
63004816 | Apr 2020 | US |