Disclosure of Invention
The invention aims to provide a monolithic integration of a Q-modulation semiconductor laser with an electro-absorption grating structure and a high-speed low-chirp modulator aiming at the defects of the prior art, and solves the problems of high cost, difficulty in integration, chirp in wavelength, complex manufacturing and the like of the traditional semiconductor laser and modulator.
The purpose of the invention is realized by the following technical scheme:
the technical scheme 1: a Q-modulation semiconductor laser with an electro-absorption grating structure comprises a phase-shift distribution feedback grating embedded in an active waveguide structure, a first upper electrode and a second upper electrode which are separated from each other, a lower electrode which is used as a common ground plane, and a first upper electrode and a second upper electrode which are respectively covered on the tops of a gain area and a modulator area; the phase shift distributed feedback grating is sequentially divided into a first part, a second part and a third part, wherein the first part and the second part are separated by a phase shift region; the first upper electrode deposited in the gain area covers the first part, the second part and the phase shift area between the first part and the second part of the grating, and injects a constant current into the active optical waveguide below the electrode to provide the required optical gain for the laser.
The phase shift in the phase-shifted distributed feedback grating is equal to a quarter wavelength.
The phase shift in the phase shift distributed feedback grating is formed by flipping the grating pattern on one side of the phase shift region relative to the other side.
The phase shift in the phase-shifted distributed feedback grating is formed by a section of phase-shifting region waveguide having a different effective index of refraction.
The period of the third part of the grating enables the working wavelength of the laser to be located in the central area of the forbidden band of the grating.
The optical waveguide losses of the modulator region are modulated by forward biased current injection.
The optical waveguide loss of the modulator region is modulated by the electro-absorption effect of the reverse bias.
The technical scheme 2 is as follows: a Q-modulation semiconductor laser with an electro-absorption grating structure comprises a first distributed Bragg reflection grating, a second distributed Bragg reflection grating and a gain region positioned between the two gratings; the gain region is sandwiched between a first pair of electrodes for injecting a constant current to provide optical gain to the laser; the second distributed bragg reflector grating comprises a modulator region having electrically controlled absorption properties, the modulator region being sandwiched between a second pair of electrodes for applying an electrical signal to alter optical losses of the modulator region, thereby altering the threshold and output power of the laser.
The modulator region and the gain region are separated by a portion of a second distributed Bragg reflection grating.
The modulator region includes a grating having a period such that the operating wavelength of the laser is located in a central region of the grating forbidden band.
The invention has the beneficial effects that:
1. the invention utilizes a new principle mechanism to integrate the semiconductor laser and the high-speed low-chirp Q-modulator in a single chip, realizes a high-performance small-size laser transmitter, and has the advantages of low cost and simple manufacture similar to those of a direct modulation laser.
2. The invention has a plurality of different specific structural forms, including different structures based on quarter-wavelength phase shift distributed feedback lasers, distributed Bragg grating lasers and the like.
3. The present invention separates the modulation function from the region of the gain function, which is pumped by a constant current, which not only reduces the wavelength chirp, but also increases the modulation speed, so that compared with a direct modulation or external electroabsorption modulator, the modulator of the present invention has a much shorter length, thus having smaller capacitance and higher speed.
4. The invention has the advantages of integration, high speed, high extinction ratio, low wavelength chirp, low cost and the like.
Detailed Description
The present invention will be described in detail below with reference to the drawings and examples.
The Q-modulated semiconductor lasers of the present invention have a number of different specific configurations in which the laser resonator may be based on a Distributed Feedback (DFB) grating or a Distributed Bragg Reflector (DBR) grating, respectively, with a phase shift.
The Q-factor or quality factor of a laser resonator is a measure of how much light from the laser gain medium is fed back through the optical resonator, and a high Q-factor means that the light propagates through the resonator with each coming back at a small loss. The principle of Q-modulation is to vary the laser output power using a device that can vary the Q-factor of the resonator, which has been applied in Q-switched dyes or solid state lasers that produce periodic short pulses. Prior art methods of implementing Q-switching generally include: a rotating mirror is used in the optical resonator, or an electro-optic or acousto-optic modulator is used. However, these methods are not feasible for small semiconductor lasers.
For modulation of semiconductor lasers, reducing the wavelength chirp is a very important aspect to consider. R.c. alferness et al in U.S. patent 4,667,331 issued at nineteen-seventy-seven fife describes a method of placing an electrical modulator in a laser resonator, but this method is not advantageously feasible because, in addition to the increased fabrication complexity, it also introduces significant wavelength chirp similar to that of a directly modulated laser.
In U.S. patent 6,519,270 issued to two-zero three years, february-a-ten-one, H.B. Kim and J.J. Hong, a composite cavity laser is described that is formed by integrating a single-mode distributed feedback laser with a passive optical waveguide region. The phase of the effective reflectivity of the cleavage surface behind the passive waveguide is modulated by modulating the refractive index of the passive waveguide, so as to modulate the laser frequency, and then a narrow-band optical filter formed by a Mach-Zehnder interferometer is arranged in front of the laser, so that the frequency modulation is converted into intensity modulation. Although this modulator is also placed at the back end of the laser, it does not change the Q value of the laser, but only the phase, resulting in a modulation of the frequency rather than the intensity. The narrow band filters required to convert frequency modulation to intensity modulation make it difficult to practically apply to conventional communication systems, and the active-passive waveguide integration required also makes fabrication of the device difficult and expensive.
A paper entitled "Q-modulation of a surface emitting laser and an integrated determined cavity", S.R. A.Dods, and M.Ogura, IEEE Journal of Quantum Electronics, vol.30, pp.1204-1211, 1994 describes and analyzes surface emitting vertical cavity lasers vertically integrated with a detuned resonant cavity, the modulation of the laser intensity being achieved by varying the refractive index in the detuned resonant cavity. The same principle is applied in U.S. patent No. 6,215,805 issued to b.sartorius and m.moehrel, two and zero years, april and april. In both of the above prior art, a reflector of the laser cavity is a slightly detuned resonator which is highly dispersive in reflectivity at the operating wavelength of the laser, i.e. the reflectivity spectrum exhibits a sharp negative spike near the lasing wavelength. High reflectivity dispersion is necessary, and small refractive index changes in the detuned cavity can cause large changes in reflector reflectivity, thereby modulating the laser output. However, this prior art approach has significant drawbacks: 1) In near resonance conditions, the reflectivity is highly wavelength dependent, and therefore it is necessary to correct their resonant wavelength accurately according to a predetermined detuning requirement between the two resonators, which is very difficult and sensitive to fabrication. 2) The change in reflectivity due to the change in refractive index in the detuned cavity is accompanied by a large phase change, which results in a large chirp of the laser wavelength.
To overcome the drawbacks of the prior art methods, the present inventors have proposed a new structure in a related chinese patent application (application No. 200610050484.6) by using an anti-resonant cavity as a back reflector of a semiconductor laser, the reflectivity of which can be changed by changing the optical absorption of the waveguide material in the anti-resonant cavity. The reflectivity of the anti-cavity and the change in reflectivity due to optical loss modulation are much less wavelength dependent than the cavity, and the phase change caused by the change in reflectivity is also quite small, so the wavelength chirp is very low. The structure usually needs a vertically deep etched air slot to realize the monolithic integration of the laser resonant cavity and the modulator anti-resonant cavity.
The invention discloses another Q-modulation semiconductor laser structure, which realizes the modulation of reflectivity and a laser quality factor by changing the absorption coefficient of a part of a grating of a back reflector without an air groove which is vertically deeply etched. The laser and its back reflector are designed such that the phase of the reflection is nearly constant as the reflectivity is changed, so this implementation introduces only a small wavelength chirp. The modulation mechanism does not require the use of wavelength sensitive resonator structures. The variation in optical loss can be achieved by current injection, and the same material as the laser gain medium can be used, thus greatly simplifying fabrication. The details of a monolithic Q-modulated semiconductor laser structure implementing the above mechanism will be described in detail below.
Fig. 1 is a first embodiment of the present invention, a single mode distributed feedback semiconductor laser with quarter-wave phase shift monolithically integrated with an electro-absorption Q-modulator, comprising a λ/4 phase shifted DFB grating 130 divided into a gain region and a modulator region. The gain region includes a phase shift region 100 and regions of waveguides with gratings 101 and 102 located on either side of the phase shift region. The gain region is covered by the first upper electrode 108 and sandwiched between the first upper electrode 108 and the ground electrode 120. The gain region provides optical gain to the laser when a direct current is injected through the first upper electrode 108. The modulator region 105 is a remaining grating portion away from the phase shift region and acts as a Q-modulator for the laser. An electrical signal is applied to the modulator region via the second upper electrode 110 to vary the Q of the laser by varying the absorption coefficient of the waveguide in that region, thereby varying the threshold current and output power. The light beam 140 exits the front end face of the gain region, i.e., the end face on the side opposite the modulator.
The waveguide structure generally includes a buffer layer 116, a waveguide core layer 114 providing optical gain when electrically pumped, and a surface cladding layer 112, all deposited on a substrate 118. The waveguide core layer 114 preferably comprises a multiple quantum well structure with appropriate doping in each layer as in conventional laser structure layers. In cross-section, the waveguide is machined as a standard ridge waveguide to also achieve confinement of the optical mode in the horizontal direction. Isolated upper electrodes 108 and 110 are deposited on the upper surfaces of the gain and modulator regions, respectively, and a metal electrode 120 is also deposited on the back side of the substrate as a common ground electrode. The electrode pairs 108/120 are used to provide current injection to the active gain region to provide optical gain. The electrode pair 110/120 is used to change the absorption coefficient of the waveguide (either by current injection or reverse bias) in the modulator region to change the Q of the laser.
Different waveguide material structures can be used in the gain and modulator regions so that the two regions are optimized separately. In actual fabrication, this can be achieved by etch-regrowth techniques or post-growth bandgap engineering such as quantum well intermixing techniques. A simpler approach is to use the same laser layer structure but apply different voltages or currents to obtain different performance in the two regions. The gain region is strongly current pumped to produce optical gain, and the modulator region is varied between transparent (low current injection) and absorptive (zero current injection) states.
To clarify the working principle of the Q-modulated laser of the invention, we consider a specific example. In this example, the refractive index of the grating is rectangularly distributed, and n 1 =3.215,n 2 =3.21 (Δ n = 0.005), grating period Λ =0.2412 μm, operating wavelength λ =1550nm. Modulation zone length L m =150 μm. The total length of the gain region is 400 μm and the λ/4 phase shift region is 100 μm from the modulator (i.e. the regions 101 and 102 have a length L, respectively 1 =300 μm, and L 2 =100 pm). The laser cavity Q value can be derived from Q = λ/Δ λ, where Δ λ is the line width of the resonant peak of the transmittance or reflectance spectrum when the gain region is in the transparent state.
Fig. 2 is a reflection spectrum of light incident on the laser structure from the gain region side when the modulator region is in the transparent (on) and absorptive (off) states. Wherein, let the absorption coefficients of the modulation regions be α =0 (on), and α =500cm, respectively -1 (off). The full width at half maximum (FWHM) of the reflection peak for both states was 0.1nm and 0.37nm, respectively, and the corresponding Q values were 15500 and 4189.
A phase shifted DFB grating can also be seen as a fabry-perot cavity with two mirrors formed by a distributed bragg reflector grating (DBR). The first DBR is the section 101 to the right of the phase shift section and the second DBR is comprised of the section 102 to the left of the phase shift section and the modulator region 105. The laser wavelength is determined by the following resonance condition:
where n is the average effective refractive index of the phase shifting region, Λ = λ/2n is the grating period, L p Is the amount of phase shift (i.e., the length of region 100), phi 1 Phi of 2 Is the reflected phase change of the first and second DBRs with respect to the phase shifting region, and m is an integer. DBR gratings have a wavelength window, called the forbidden band, within which light of a wavelength is mostly reflected. For a wavelength at the center of the DBR forbidden band, phi 1 =Ф 2 And =0. When m =1, L p = λ/4n, which corresponds to a quarter-wave phase shift. The above-described quarter-wave phase shifted DFB structure can be realized by inverting the grating pattern on one side of the phase shift location relative to the other, which can be realized using photoresist of opposite polarity during the grating fabrication process.
Fig. 3 shows the reflectivity spectrum (a) and its corresponding phase change (b) of a second DBR grating consisting of the above sections 102 and 105. Light enters from the phase shift region 100The absorption coefficients of the modulator region 105 are α =0 and α =500cm, respectively -1 . From the figure we can see that the absorption of the modulator results in a large change in the reflection peak with an accompanying phase change that is minimal at the peak wavelength (center of the forbidden band). The change in reflectivity causes a change in the Q value of the laser cavity, thereby changing the lasing threshold. As can be seen from equation (1), the minimum phase change corresponds to the minimum wavelength chirp, which is very important.
In semiconductor materials, the change in absorption is always accompanied by a change in refractive index according to the Kramer-Kronig relationship, which can be very large in some working cases. It can be used to enhance the modulation of the laser threshold. However, the change in refractive index causes a shift in the peak value and a change in the reflection phase, increasing the wavelength chirp. The DBR grating region 102 (L in the above example) is added in the modulator region and phase shift region 2 =100 μm) the peak shift and phase variation can be minimized. Fig. 3 also shows when the modulator region is in the absorbing state (α =500 cm) -1 ) While the reflectance spectrum and phase of the DBR grating changes with an accompanying increase in refractive index of 0.005. The shift in the reflection peak due to the change in refractive index is only 0.35nm. Without the DBR section 102, the shift of the modulator section reflection peak is calculated by equation (2):
calculated using the parameters in the above example, the resulting wavelength shift is Δ λ =1550 × 0.005/3.215=2.4nm. Thus, the addition of the DBR section 102 to the modulator and phase shift section can greatly reduce the wavelength shift. In addition, as can be seen from fig. 5 (b), the phase change in the central region of the forbidden band is also not significant. With L 2 The peak shift and phase change increase accordingly. On the other hand, the efficiency of modulation varies with L 2 Is increased and decreased. Thus, in selecting L 2 When the value of (A) is needed to be considered comprehensively, the method also comprises the step ofDependent on lightThe refractive index difference of the grating.
Compared with the directly modulated quarter-phase shifted DFB laser, the Q-modulated laser of the present invention has significant advantages in reducing the wavelength chirp as can be seen from equations (1) and (2). For those directly modulated lasers, the wavelength fluctuation is 2.4nm in the above example, since the refractive index of the entire laser structure changes with the modulation current, as can be derived from equation (2). In the structure of the present invention, however, the refractive index n of the phase shift region in the formula (1) and the phase Φ of the first DBR region are lost due to the modulation of only the portion of the grating away from the phase shift region 1 Will remain unchanged with only the phase of the second DBR section 2 Will vary slightly with the modulation current. But according to fig. 5 (b), this phase variation can be minimized by designing the length of region 102 between the modulator region and the phase shift region. Therefore, the wavelength chirp can be greatly reduced.
FIG. 4 shows the gain factor g =9.25cm in the above example -1 Absorption coefficients of α =0 and α =500cm, respectively -1 The laser structure's transmission small signal gain spectrum in both modulator states. The lasing wavelength is in the center of the forbidden band due to the quarter wavelength phase shift in the DFB grating. When the modulator is in the transparent state (α = 0), the threshold gain factor of the lasing mode is 9.25cm -1 . When the modulator is in the absorption state, the absorption coefficient is alpha =500cm -1 When the threshold gain factor is increased to 38cm -1 While the wavelength remains unchanged for λ =1549.711 nm. If the refractive index variation is taken into account in the calculation, the threshold gain factor becomes 41.5cm -1 And the laser wavelength is 1549.745nm, and the drift is only 0.034nm. This figure is reduced by 2 orders of magnitude compared to the several nm wavelength chirp of a conventional directly modulated DFB laser.
The large difference in lasing mode thresholds for the two states of the modulator region indicates that it is an efficient way to implement Q-modulation using the Q-modulator loss variation of the present invention. When the constant current of the pump gain region produces an optical gain that is lower than the threshold of the laser in the absorption state of the modulator but much higher than the threshold of the laser in the transparent state, the output of the laser is modulated by an electrical signal applied across the modulator. The phase change accompanying the Q modulation causes only a very low wavelength chirp and is small enough to be almost negligible, which is an important advantage of the present invention.
Fig. 5 shows the threshold gain factor of a laser as a function of the absorption factor of the modulator. It can be seen that the absorption coefficient of the modulator is only 200cm -1 This threshold varies by up to 300%. In the above embodiments, the effective refractive index of the modulator region in the on state can preferably be the same as the gain region. The gain region is pumped by a relatively strong current during operation of the device to provide gain to the laser. The modulation region can inject current at the same current density in the on state if the waveguide material, cross-sectional shape, and grating period are the same for both the modulator region and the gain region. However, even in the on state, it is generally not necessary for the modulator region to inject such a large current, since a large current injection results in a large total drive power. Generally, in the on state, it is sufficient to apply a current that makes the waveguide sufficiently transparent. Effective refraction in the gain and modulator regions due to different current densitiesThere will also be a slight difference in the rate. This effect can be compensated by changing the shape of the cross-section of the waveguides in the modulator region (e.g. the ridge width), i.e. by using different ridge widths in the gain region and modulator region, and by using a width-graded structure to reduce the transition loss.
Fig. 6 shows the intensity distribution in the laser structure when the modulator is in the on (a) and off (b) states, respectively, for a gain factor g =8.8cm -1 Absorption coefficients =0 and α =500cm, respectively -1 And (4) calculating. It can be seen that in the on state, the light intensity increases exponentially from both ends towards the centre until it reaches a maximum at the phase shift position. When the modulator is turned to the off state, the light intensity is significantly reduced,and the distribution will also change. This extremely inhomogeneous field distribution, especially the sharp peaks of the phase shifting regions in the on-state, leads to strong spatial hole burning effects and gain saturation.
To mitigate the spatial hole burning effect, the phase shift can be achieved by a waveguide region of a specific length with slightly different effective refractive index. Let us consider another example, in which the modulation region length L is m =150 μm, the gain region is composed of two gain regions of length L 1 =250 μm and L 2 =100 μm DBR sections, which are each formed by a length L p =50 μm phase shift zones apart. The phase shifting region has the same grating period Λ =0.2412 μm, but the effective index is reduced to 3.204, while the effective index in the other regions is 3.2125.
Fig. 7 (a) and (b) show the modulator in the on state (α = 0) and the off state (α =500 cm), respectively -1 ) Intensity distribution of light, gain factor g =8.2cm used in calculation -1 . The variation in light intensity of the phase shift region becomes less significant compared to fig. 8. At a wavelength λ =1549.75nm, the laser threshold gain factor is 8.6cm in the on state of the modulator -1 Off state (α =500 cm) -1 ) Lower is 29cm -1 . In this example, the phase shifting regions of the DFB grating may be implemented by waveguides of different ridge widths, or by injecting different current densities through a separate electrode.
In the present invention, the mechanism of Q-switching can also be applied to conventional DFB lasers with uniform gratings (i.e., no phase shifting regions are present). In this case, however, the DBR grating of the modulation region requires a forbidden band that is detuned from the DFB region. To enable single mode, partially gain-coupled DFB gratings can be used, similar to those described in G.P.Li, T.Makino, and H.Lu in their paper "Simulation and interpretation of longitudinal-mode waveguide in partial gate-coupled InGaAsP/InP multi-quantum-well DFB lasers", IEEE Photonics Technology Letters, vol.4, no.4, pp.386-388, 1993. In this case, the laser wavelength is on the long wavelength side of the DFB forbidden band. In accordance with the inventive concept, the grating of the modulator region needs to operate in a high reflectivity state, near the center of its forbidden band. This point is very important because the phase difference between the on state and the off state is minimal when the wavelength is at the center of the forbidden band, as shown in fig. 3 (b). Therefore, a detuning of the wavelengths of the DBR modulator region and the DFB gain region is necessary to reduce the wavelength chirp. The detuning of the wavelength can be achieved by adjusting the structure of the cross-sectional waveguide, e.g. the waveguide ridge width or the grating period. A fixed or adjustable phase region may also be added between the DFB region and the modulation region to tune the laser wavelength to the forbidden center of the DBR grating in the modulator region. To reduce phase variations and wavelength drift due to refractive index variations accompanying loss modulation, another fixed current injection DBR section may be added between the modulator and the phase/DFB section. The DBR section and the phase and DFB sections of the fixed current injection may together form a gain section using a common electrode, similar to the embodiment of fig. 1.
The Q-modulated semiconductor laser of the present invention may also take the form of a distributed bragg laser. Figure 8 shows another embodiment of the present invention. The laser consists of two DBR gratings 231 and 232 and a gain waveguide section 200 between the two gratings. The waveguide region 201 containing the DBR grating 231 and the waveguide region 202 containing a portion of the DBR grating 232 are passive and substantially transparent. The gain region does not include a grating and is sandwiched between a pair of electrodes 208/120 to provide optical gain. The modulator region 205, which is comprised of another portion of the DBR grating 232, is also located between a pair of electrodes 110/120 that are used to change the optical loss of the optical waveguide therebetween, thereby changing the Q value and thus the lasing threshold and output power.
It is apparent that the DBR grating 201 of the arrangement shown in fig. 8 may be replaced by a partially reflective cleaved surface, which may or may not be coated with a dielectric film.
The Q-modulated laser of the present invention has many advantages. Since the modulation function is separated from the gain region, which is constant-current pumped, this not only reduces the wavelength chirp, but also increases the modulation speed, since the modulator of the present invention is much shorter in length, and thus has smaller capacitance and higher speed, compared to a directly modulated or externally mounted electro-absorption modulator. The extinction ratio of the modulator of the present invention is also much higher due to the use of a Q-switching mechanism and does not require a long modulator length, relative to an electro-absorption modulator placed in the path of the output laser beam. Furthermore, it does not inevitably produce energy losses as does an external electro-absorption modulator.
The embodiments of the present invention are merely to illustrate the present invention and not to limit the present invention, and any modifications and changes made within the spirit of the present invention and the scope of the claims fall within the scope of the present invention. For example, the structural principle of the Q-modulated semiconductor laser in the present invention can also be applied to a vertical cavity surface radiation laser.