JP2009043790A - Semiconductor laser array - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To uniquely and precisely control the gain peak wavelengths in a plurality of light emission regions for generating the plurality of laser light beams respectively in a semiconductor laser array for outputting the plurality of laser light beams having mutually different wavelengths so that the gain peak wavelengths correspond to the wavelengths of the laser light beams respectively. <P>SOLUTION: This semiconductor laser array 10 outputs the plurality of laser light beams having mutually different wavelengths. The semiconductor array 10 includes a semiconductor substrate, and the plurality of light emission regions 17A-17E formed on the semiconductor substrate and for generating the plurality of laser light beams respectively. The plurality of light emission regions 17A-17E have the plurality of quantum fine lines 19A-19E provided aligned with periods Λa-Λe corresponding to the wavelengths of the laser light beams respectively. The fine line widths Wa-We of the quantum fine lines 19A-19E in the light emission regions 17A-17E are the widths corresponding to the wavelengths of the laser light beams generated by the light emission regions, and the fine line widths in at least two counterpart light emission regions are mutually different. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a semiconductor laser array.

An example of a semiconductor laser array that outputs a plurality of laser beams having different wavelengths is described in Patent Document 1, for example. The semiconductor optical device described in Patent Document 1 employs the following method in order to make the wavelengths of a plurality of laser beams different from each other. That is, when a plurality of waveguide structures including a light emitting region are formed using a selective growth technique, a dummy growth region is provided between the waveguide structures, and the width and position thereof are different between the waveguide structures. . Thus, the density of the semiconductor material around each waveguide structure during selective growth is changed, the composition and layer thickness of the light emitting region (quantum well structure) are changed for each light emitting region, and the wavelength at which the laser gain is maximized ( The gain peak wavelength) is made different for each light emitting region. In addition, a diffraction grating having a period corresponding to the gain peak wavelength of each light emitting region is formed in each waveguide structure to form a distributed feedback (DFB) laser array.
JP 2002-289971 A

  However, in the method described in Patent Document 1, since both the composition of the quantum well structure and the layer thickness change according to the width and position of the dummy growth region, it is difficult to control the gain peak wavelength to an arbitrary value. . That is, the gain peak wavelength varies depending on the composition of the quantum well and also varies depending on the layer thickness of the quantum well. However, since the method described in Patent Document 1 cannot control these independently, the dummy growth region It is difficult to predict the relationship between the width and position of the signal and the gain peak wavelength. In the method described in Patent Document 1, since the composition and film thickness of the quantum well that determines the gain peak wavelength are indirectly controlled by the width and position of the dummy growth region, the accuracy of the gain peak wavelength ( Reproducibility) is kept low.

  The present invention has been made in view of the above-described problems. In a semiconductor laser array that outputs a plurality of laser beams having different wavelengths, gain peak wavelengths in a plurality of light emitting regions that respectively generate a plurality of laser beams are calculated. It is an object to control the laser light uniquely and accurately so as to correspond to the wavelength of the laser light.

  In order to solve the above-described problems, a semiconductor laser array according to the present invention is a semiconductor laser array that outputs a plurality of laser beams having different wavelengths, and a semiconductor substrate having a main surface and a main surface of the semiconductor substrate. And a plurality of light emitting regions that generate each of the plurality of laser beams, and each of the plurality of light emitting regions has a wavelength of the laser light generated by the light emitting region. It has a plurality of quantum wires provided side by side in the longitudinal direction of each light emitting region with a corresponding period, and the thin wire width in the longitudinal direction of each quantum wire in each of the plurality of light emitting regions generates the light emitting region. It is formed in a dimension corresponding to the wavelength of the laser light, and the thin line width is different between at least two light emitting regions among the plurality of light emitting regions.

  The semiconductor laser array described above has a so-called complex coupled DFB laser configuration. That is, each light emitting region has a plurality of quantum wires, and the plurality of quantum wires are arranged in the longitudinal direction of each light emitting region with a period corresponding to the wavelength of the laser light to be output by the light emitting region. Is provided. Further, in the semiconductor laser array described above, the thin line width of each quantum wire in the longitudinal direction of the light emitting region is formed to have a dimension corresponding to the wavelength of the laser light (Bragg wavelength) generated by the light emitting region. As described above, at least two of the light emitting regions have different thin line widths.

  There is a direct correlation between the fine line width of the quantum fine line and the gain peak wavelength of the light emitting region. For example, when the narrow line width of the quantum thin line is narrowed, the gain peak wavelength of the light emitting region changes to the short wavelength side. On the contrary, when the quantum wire width becomes wider, the gain peak wavelength of the light emitting region is hardly shortened, and becomes a value close to the gain peak wavelength when the quantum wire structure is not provided. From this, it is easy to predict the relationship between the fine wire width of the quantum wire and the gain peak wavelength. In addition, even when a plurality of light emitting regions are provided on a single substrate, the quantum wire can be formed with extremely high accuracy in different dimensions for each light emitting region by an electron beam exposure method or the like. is there. Therefore, if the gain peak wavelength is controlled by the fine line width of the quantum thin line, the gain peak wavelength in the plurality of light emitting regions can be uniquely and accurately controlled so as to correspond to the wavelength of the laser light generated in the light emitting region. Furthermore, by forming the fine line width of the quantum wire so as to correspond to the wavelength of the laser light (Bragg wavelength) determined by the period of the quantum wire, single mode oscillation in each light emitting region can be suitably realized.

  Further, the semiconductor laser array is formed such that the fine line width and period of each quantum wire in at least a part of the light emitting regions are substantially equal to the gain peak wavelength and the Bragg wavelength in the light emitting region. It may be characterized by being. Thereby, the gain and the oscillation wavelength are matched in the at least part of the light emitting region, and the single mode oscillation can be realized more preferably.

  In addition, the semiconductor laser array is formed such that the fine line width and period of each quantum wire in a part of the light emitting regions are substantially equal to the gain peak wavelength and the Bragg wavelength in the light emitting region. The thin line widths in two or more light emitting regions other than some light emitting regions may be equal to each other in the light emitting regions. When the laser gain obtained by a certain thin line width is sufficiently large over a wide wavelength range, single mode oscillation can be sufficiently realized if the Bragg wavelength exists in the vicinity of the gain peak wavelength. Therefore, the thin line widths of the two or more light emitting regions may be set to be equal to each other with such a thin line width, and the quantum wire period may be set so that the Bragg wavelengths of these light emitting regions are different from each other. Single mode oscillation in the light emitting region can be suitably realized.

  According to the present invention, in a semiconductor laser array that outputs a plurality of laser beams having different wavelengths, gain peak wavelengths in a plurality of light emitting regions that respectively generate a plurality of laser beams correspond to the wavelengths of the laser beams. Can be controlled uniquely and accurately.

  Embodiments of a semiconductor laser array according to the present invention will be described below in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted.

  FIG. 1 is a partially cutaway perspective view showing the structure of a semiconductor laser array according to an embodiment of the present invention. 2 is a plan sectional view showing a section taken along line II-II shown in FIG. Referring to FIG. 1, the semiconductor laser array 10 includes a semiconductor substrate 37 having a main surface 37a and a plurality of semiconductor laser portions 11A to 11E formed on the main surface 37a. As the semiconductor substrate 37, for example, a p-type InP substrate is preferably used.

  Each of the plurality of semiconductor laser units 11A to 11E has a configuration of a complex coupled DFB laser, and outputs a plurality of laser beams having different wavelengths. The plurality of semiconductor laser portions 11 </ b> A to 11 </ b> E are formed with a predetermined direction along the main surface 37 a of the semiconductor substrate 37 as a longitudinal direction, and are formed side by side in a direction intersecting the longitudinal direction. The semiconductor laser portions 11A to 11E have first light confinement layers 13A to 13E, second light confinement layers 15A to 15E, and light emitting regions 17A to 17E, respectively. As a constituent material of the first light confinement layers 13A to 13E and the second light confinement layers 15A to 15E, for example, GaInAsP is suitable. The first light confinement layers 13A to 13E, the second light confinement layers 15A to 15E, and the light emitting regions 17A to 17E constitute semiconductor mesas 25A to 25E extending along a predetermined direction.

  The light emitting regions 17A to 17E are respectively formed with the direction along the main surface 37a of the semiconductor substrate 37 as the longitudinal direction, and generate laser beams output from the respective semiconductor laser portions 11A to 11E. The light emitting regions 17A to 17E are provided between the first light confinement layers 13A to 13E and the second light confinement layers 15A to 15E, respectively. As shown in FIG. 2, the light emitting region 17A includes a plurality of quantum wires 19A and an intermediate semiconductor region 21A. Similarly, the light emitting region 17B includes a plurality of quantum wires 19B and an intermediate semiconductor region 21B, and the light emitting region 17C includes a plurality of quantum wires 19C and an intermediate semiconductor region 21C. The region 17D includes a plurality of quantum wires 19D and an intermediate semiconductor region 21D, and the light emitting region 17E includes a plurality of quantum wires 19E and an intermediate semiconductor region 21E.

  As shown in FIG. 2, the quantum wires 19A to 19E have periods Λ (lambda) a to Λe (Λa to Λe having different lengths corresponding to the wavelengths of the laser beams to be generated by the light emitting regions 17A to 17E. In the embodiment, Λa> Λb> Λc> Λd> Λe) are provided side by side in the longitudinal direction of each light emitting region. The widths (thin line widths) Wa to We of the quantum wires 19A to 19E in the longitudinal direction of the light emitting regions 17A to 17E are the wavelengths of the laser beams generated by the light emitting regions 17A to 17E (that is, the quantum wires 19A to 19E). It is set to a width corresponding to the Bragg wavelength determined by the periods Λa to Λe. For example, in the present embodiment, the thin line widths Wa and Wb of the light emitting regions 17A and 17B are set to a width W1 corresponding to both the periods Λa and Λb, and the thin line width Wc of the light emitting region 17C is a width corresponding to the period Λc. W2 (<W1) is set, the thin line width Wd of the light emitting region 17D is set to a width W3 (<W2) corresponding to the period Λd, and the thin line width We of the light emitting region 17E is a width corresponding to the period Λe. W4 (<W3) is set. As described above, when the thin line widths Wa to We are set to widths corresponding to the Bragg wavelengths of the light emitting regions 17A to 17E, as a result, at least two light emitting regions (light emitting regions in the present embodiment) among the light emitting regions 17A to 17E. 17A (or 17B), 17C, 17D and 17E) have different thin line widths.

  Further, as shown in FIG. 2, the gaps between the plurality of quantum wires 19A are filled with the intermediate semiconductor region 21A. Similarly, the gaps between the plurality of quantum wires 19B are filled with the intermediate semiconductor region 21B, the gaps between the plurality of quantum wires 19C are filled with the intermediate semiconductor region 21C, and the gaps between the plurality of quantum wires 19D are filled with the intermediate semiconductor region 21D. The gaps between the plurality of quantum wires 19E are filled with the intermediate semiconductor region 21E. As a constituent material of the quantum wires 19A to 19E, for example, undoped GaInAsP is suitable. Moreover, as a constituent material of the intermediate semiconductor regions 21A to 21E, for example, undoped InP is suitable.

  FIG. 3 is a side sectional view showing a section along the longitudinal direction of the light emitting region 17A. Note that the sectional structure of the light emitting regions 17B to 17E is the same as the sectional structure of the light emitting region 17A shown in FIG. As already described, the plurality of quantum wires 19A are formed side by side with the period Λa, and the gaps between the plurality of quantum wires 19A are filled with the intermediate semiconductor region 21A. As shown in FIG. 3, the quantum wires 19A (and 19B to 19E) have a multiple quantum well structure, and are formed by alternately laminating barrier layers (barrier layers) 19f and well layers 19g.

  The multiple quantum well structure of the quantum wires 19A (and 19B to 19E) is preferably a strain compensated quantum well structure including a compressive strain well layer 19g and a tensile strain barrier layer 19f. In this case, it is more preferable that the compressive strain amount of the well layer is about 1.0% and the tensile strain amount of the barrier layer is about -0.15%. In order to compensate for a decrease in effective gain due to a decrease in active layer volume accompanying the formation of the quantum wire structure, the number of well layers is preferably 4 or more.

  Referring again to FIGS. 1 and 2, the semiconductor laser array 10 further includes a p-type (first conductivity type) cladding layer 27 and an n-type (second conductivity type) cladding layer 29. The p-type cladding layer 27 is located between the first optical confinement layers 13A to 13E and the semiconductor substrate 37, and the n-type cladding layer 29 is located on the second optical confinement layers 15A to 15E. . The p-type cladding layer 27 and the n-type cladding layer 29 are provided over the entire main surface 37a of the semiconductor substrate 37, and sandwich the semiconductor mesas 25A to 25E from above and below.

  Further, the side surfaces of the semiconductor mesas 25 </ b> A to 25 </ b> E are embedded by the embedded region 31. The buried region 31 is constituted by, for example, a first p-type current confinement layer 31a, an n-type current confinement layer 31b, and a second p-type current confinement layer 31c. The first p-type current confinement layer 31a is provided in a region on the p-type cladding layer 27 excluding a region in which the semiconductor mesas 25A to 25E are provided, and the region of the p-type cladding layer 27 and the semiconductor mesas 25A to 25E. Covers the sides. The n-type current confinement layer 31b is provided on the first p-type current confinement layer 31a. The second p-type current confinement layer 31c is provided between the n-type current confinement layer 31b and the n-type clad layer 29 described above. These current confinement layers 31a to 31c are made of, for example, n-type (or p-type) InP.

  An insulating film 33 and a first (cathode) electrode film 35 are formed on the n-type cladding layer 29. The insulating film 33 has openings corresponding to the semiconductor mesas 25A to 25E, and the n-type cladding layer 29 and the electrode film 35 are in ohmic contact through the openings. A second (anode) electrode film 39 is provided on the back surface opposite to the main surface 37 a of the semiconductor substrate 37, and the second electrode film 39 and the semiconductor substrate 37 are in ohmic contact.

  Next, the relationship between the periods Λa to Λe and the thin line widths Wa to We of the quantum wires 19A to 19E formed in the light emitting regions 17A to 17E will be described in more detail. FIG. 4 is a perspective view schematically showing semiconductor mesas 25A to 25E provided on the main surface 37a of the semiconductor substrate 37, illustrating the periods Λa to Λe and the thin wire widths Wa to We of the quantum wires 19A to 19E. ing. 5A and 5B are diagrams showing the relationship between the gain wavelength characteristics Ga to Ge of the light emitting regions 17A to 17E and the Bragg wavelengths λBa to λBe of the light emitting regions 17A to 17E. FIG. 5A shows gain wavelength characteristics Ga to Ge, where the horizontal axis represents wavelength and the vertical axis represents gain. FIG. 5B shows the positions of the Bragg wavelengths λBa to λBe on the wavelength axis in FIG.

  There is a direct correlation between the fine line width of the quantum fine line and the wavelength (gain peak wavelength) at which the gain of the light emitting region is maximized. For example, when the narrow line width of the quantum thin line is narrowed, the gain peak wavelength of the light emitting region changes to the short wavelength side. On the contrary, when the quantum wire width becomes wider, the gain peak wavelength of the light emitting region is hardly shortened, and becomes a value close to the gain peak wavelength when the quantum wire structure is not provided. Specifically, in FIG. 5A, the gain peak wavelength λGa of the gain wavelength characteristic Ga has a value corresponding to the thin wire width Wa (see FIG. 4) of the quantum wire 19A. Similarly, the gain peak wavelengths λGb to λGe of the gain wavelength characteristics Gb to Ge are values corresponding to the thin wire widths Wb to We of the quantum thin wires 19B to 19E. Further, the bandwidth of the gain wavelength characteristics Ga to Ge is determined according to the thin wire widths Wa to We of the quantum wires 19A to 19E, and the bandwidth tends to be narrowed as the thin wire width is narrowed. In this embodiment, since Wa = Wb> Wc> Wd> We, λGa = λGb> λGc> λGd> λGe.

As shown in FIGS. 5A and 5B, in the present embodiment, the gain peak wavelength λGc, the Bragg wavelength λBc, the gain in some of the light emitting regions 17C to 17E among the light emitting regions 17A to 17E. The periods Λc to Λe and the thin line widths Wc to We of the quantum wires 19C to 19E are set so that the peak wavelength λGd and the Bragg wavelength λBd, and the gain peak wavelength λGe and the Bragg wavelength λBe substantially coincide with each other. In the other light emitting regions 17A and 17B excluding these light emitting regions 17C to 17E, the thin line widths Wa and Wb are set to the same length (W1), and the periods Λa and Λb are determined by the thin line width W1. In the gain wavelength characteristic Ga (Gb) to be determined, the laser gain at the Bragg wavelengths λBa and λBb is set to be sufficiently large. Specific examples of these numerical values are shown below. The values shown below indicate the cases where the wavelengths λBa to λBe correspond to the C band and the L band in the CWDM (Coarse Wavelength-Division Multiplexing) communication system.
Light emitting region 17A: λBa = 1610 [nm], Wa = 90 [nm], Λa = 251.6 [nm]
Light emitting region 17B: λBb = 1590 [nm], Wb = 90 [nm], Λb = 248.4 [nm]
Light emitting region 17C: λBc = 1570 [nm], Wc = 47 [nm], Λc = 245.3 [nm]
Light emitting region 17D: λBd = 1550 [nm], Wd = 28 [nm], Λd = 242.2 [nm]
Light emitting region 17E: λBe = 1530 [nm], We = 20 [nm], Λe = 239.1 [nm]

  The effects obtained by the semiconductor laser array 10 according to this embodiment will be described. The semiconductor laser array 10 has a so-called complex coupled DFB laser configuration. That is, each of the light emitting regions 17A to 17E has quantum wires 19A to 19E, and the quantum wires 19A to 19E have periods Λa to Λe corresponding to the wavelengths λBa to λBe of the laser light to be output by the light emitting regions. Is provided. Further, the thin line widths Wa to We in the respective light emitting regions 17A to 17E are formed to have dimensions corresponding to the Bragg wavelengths λBa to λBe of the light emitting region, and as a result, at least two of the plurality of light emitting regions 17A to 17E. The thin line widths of the two light emitting regions are different from each other.

  As described above, there is a direct correlation between the fine line width of the quantum fine line and the gain peak wavelength of the light emitting region. Therefore, it is easy to predict the relationship between the fine line width of the quantum fine line and the gain peak wavelength, and the thin line widths Wa to We corresponding to the desired gain peak wavelengths λGa to λGe can be easily determined. Further, the thin wire widths Wa to We of the quantum thin wires 19A to 19E can be formed with extremely high accuracy in different dimensions in each of the plurality of light emitting regions provided on one semiconductor substrate 37. Therefore, if the gain peak wavelengths λGa to λGe are controlled by the thin wire widths Wa to We of the quantum wires 19A to 19E, the gain peak wavelengths λGa to λGe are uniquely and accurately controlled to a desired value on one semiconductor substrate 37. it can. Further, by setting the thin line widths Wa to We of the quantum thin lines 19A to 19E so that the gain peak wavelengths λGa to λGe and the Bragg wavelengths λBa to λBe correspond to each other, single mode oscillation in the respective light emitting regions 17A to 17E Can be suitably realized. That is, according to the semiconductor laser array 10 according to the present embodiment, it is possible to realize a broadband laser array that can cover both the C band and the L band in the CWDM communication system, for example.

  Further, as in the present embodiment, the thin line widths Wc to We and the periods Λc to Λe in at least some of the light emitting regions 17A to 17E (the light emitting regions 17C to 17E in the present embodiment) correspond to the light emission. The gain peak wavelength and the Bragg wavelength in the region are preferably formed so as to substantially match. Thereby, the gain and the oscillation wavelength are matched in at least a part of the light emitting region, and single mode oscillation can be realized more suitably.

  Further, as in the present embodiment, the thin line widths Wc to We and the periods Λc to Λe in some of the light emitting regions 17C to 17E among the plurality of light emitting regions 17A to 17E are the gain peak wavelength and the Bragg wavelength in the light emitting region. May be formed so as to substantially coincide with each other, and the thin line widths Wa and Wb in the other light emitting regions 17A and 17B may be formed to be equal to each other. When the laser gain obtained by a certain thin line width (W1) is sufficiently large over a wide wavelength region as in the gain wavelength characteristics Ga and Gb shown in FIG. 5A, the gain peak wavelength and the Bragg wavelength Even if they do not match, single mode oscillation can be realized sufficiently. Therefore, even when the thin line widths Wa and Wb are set to be equal to each other with such a thin line width (W1), the periods Λa and Λb are set so that the Bragg wavelengths λBa and λBb of the light emitting regions 17A and 17B are different from each other. Even with such a configuration, single mode oscillation in the plurality of light emitting regions 17A to 17E can be suitably realized.

  Subsequently, main steps in the method of manufacturing the semiconductor laser array according to the embodiment will be described with reference to FIGS.

  First, as shown in FIG. 6A, a p-type InP cladding layer 53 is grown on a p-type InP semiconductor substrate 51. This growth is performed using, for example, a metal organic vapor phase growth furnace. Next, a semiconductor stack 55 for a quantum well structure is formed on the p-type InP cladding layer 53. The semiconductor stack 55 includes a p-type GaInAsP light confinement layer 57, a multiple quantum well structure 59 made of GaInAsP, and an undoped GaInAsP light confinement layer 61. The semiconductor film in the semiconductor stack 55 is grown using, for example, a metal organic chemical vapor deposition furnace. The GaInAsP multiple quantum well structure 59 has a plurality of well layers and barrier layers that are alternately stacked. In one example of the semiconductor stack 55, the multiple quantum well structure 59 is preferably a strain compensated quantum well structure including a compressive strain well layer and a tensile strain barrier layer. Thereby, the non-light-emitting recombination current component at the regrowth interface can be reduced. In this case, it is more preferable that the compressive strain amount of the well layer is 1.0% and the tensile strain amount of the barrier layer is −0.15%. In order to compensate for a decrease in effective gain due to a decrease in active layer volume accompanying the formation of the quantum wire structure in a later step, the number of well layers is preferably 4 or more. When the gain peak wavelengths λGa to λGe are set in the range of 1610 [nm] to 1530 [nm] as in the above embodiment, the photoluminescence (PL) peak wavelength of the GaInAsP multiple quantum well structure 59 in this step. It is preferable to adjust the composition and the film thickness so that becomes 1610 [nm].

  Subsequently, a mask for forming quantum wires is formed on the semiconductor stack 55. As shown in FIG. 6B, an insulating film 63, for example, a silicon-based inorganic compound film such as a silicon oxide film is deposited for the mask. This deposition is performed by, for example, chemical vapor deposition. A resist 65 is applied on the insulating film 63.

  Subsequently, as shown in FIG. 7A, a resist mask 67 is formed by transferring a plurality of periodically arranged patterns to the resist 65. In this step, the resist mask 67 is formed to have a pattern corresponding to the quantum wires 19A to 19E shown in FIG. The resist mask 67 is formed using, for example, an electron beam exposure method or a lithography technique such as nanoimprint.

Subsequently, the insulating film 63 (silicon oxide film) is etched using the resist mask 67. For this etching, for example, reactive ion etching using CF ₄ gas can be used. The resist mask 67 and the insulating film 63 (silicon oxide film, for example, SiO ₂₎ in this process in order to ensure the selectivity between, the thickness of the insulating film 63 and 15-25 [nm] in the previous step It is preferable. When the resist mask 67 is removed after the etching, a plurality of mask patterns 69 for forming quantum wires are formed as shown in FIG. 7B. The width of the mask pattern 69 is set equal to the thin line widths Wa to We of the quantum thin lines 19A to 19E shown in FIG. The arrangement period of the mask pattern 69 is set equal to the periods Λa to Λe of the quantum wires 19A to 19E.

Subsequently, as shown in FIGS. 8A and 8B, the semiconductor stack 55 is etched using the mask pattern 69 to form a plurality of quantum wires 71 </ b> A to 71 </ b> E all together. 8A is a plan view of the quantum wires 71A to 71E viewed from the main surface side of the substrate, and FIG. 8B is a side view along the line VIII-VIII shown in FIG. 8A. A cross-sectional view is shown. In this step, the optical confinement layer 61 and the multiple quantum well structure 59 of the semiconductor stack 55 are etched. In this example of etching, RIE using CH ₄ / H ₂ is used. For example, by repeating the O ₂ ashing for removing carbon polymer deposited on the semiconductor surface during the etching and RIE etching using CH 4 _/ H _2, forming an excellent multi-layer quantum wire structure perpendicularity it can. As a result of this etching, a plurality of quantum wires 71 </ b> A to 71 </ b> E are arranged on the optical confinement layer 57. The quantum wires 71 </ b> A to 71 </ b> E include an optical confinement layer 61 and a multiple quantum well structure 59. The quantum wires 71A to 71E are arranged with periods Λa to Λe, respectively, and the thin wire widths of the quantum wires 71A to 71E are set to Wa to We, respectively.

  Thereafter, wet etching is performed to remove a damaged layer caused by dry etching. For this etching, for example, a sulfuric acid-based solution is used. After the wet etching, the mask pattern 69 is removed. For example, the mask pattern 69 made of silicon oxide is removed with buffered hydrofluoric acid.

  Subsequently, as shown in FIG. 8A, the semiconductor 73 for burying the quantum wires 71A to 71E is regrown. This embedding is performed by undoped InP. In this embedding process, the InP growth rate for embedding is as low as about 250 [nm / h] in order to obtain a flat re-growth interface by uniformly embedding the gaps between the quantum wires 71A to 71E having different thin wire widths and periods. It is desirable that By the embedding step, a periodic structure is formed in which the gain regions provided by the quantum wires 71A to 71E and the intermediate semiconductor regions without gain are alternately arranged in a predetermined direction.

  Subsequently, as shown in FIG. 8B, the n-type GaInAsP light confinement layer 75 and the n-type InP cladding layer 77 are regrown. These growth rates are normal growth rates, for example, about 1 [μm / h]. Thereafter, if necessary, a refractive index waveguide structure such as a buried heterostructure may be formed. That is, after forming a semiconductor mesa (for example, a width of 1.0 [μm]), the side surface of the semiconductor mesa is formed by the first p-type InP current confinement layer, the n-type InP current confinement layer, and the second p-type InP current confinement layer. Embed. Thus, the semiconductor laser array having the configuration shown in FIGS. 1 and 2 is completed.

  The semiconductor laser array according to the present invention is not limited to the above-described embodiment, and various other modifications are possible. For example, in the above-described embodiment, an InP-based semiconductor is exemplified as an example of a constituent material. However, the present invention can also be applied to a semiconductor laser array made of another constituent material such as a GaAs-based semiconductor or a GaN-based semiconductor.

FIG. 1 is a partially cutaway perspective view showing the structure of a semiconductor laser array according to an embodiment of the present invention. FIG. 2 is a plan sectional view showing a section taken along line II-II shown in FIG. FIG. 3 is a side sectional view showing a section including the longitudinal direction and the thickness direction of the light emitting region. FIG. 4 is a perspective view schematically showing a semiconductor mesa provided on the main surface of the semiconductor substrate, and illustrates quantum wire periods Λa to Λe and thin wire widths Wa to We. FIGS. 5A and 5B are diagrams showing the relationship between the gain wavelength characteristic and the Bragg wavelength in each light emitting region. FIG. 5A shows gain wavelength characteristics, and FIG. 5B shows the position of the Bragg wavelength on the wavelength axis of FIG. FIG. 6 is a drawing showing main steps in the method of manufacturing the semiconductor laser according to the present embodiment. FIG. 7 is a drawing showing main steps in the method of manufacturing the semiconductor laser according to the present embodiment. FIG. 8A and FIG. 8B are drawings showing main steps in the method of manufacturing the semiconductor laser according to the present embodiment. FIG. 9 is a drawing showing main steps in the method of manufacturing the semiconductor laser according to the present embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Semiconductor laser array, 11A-11E ... Semiconductor laser part, 13A-13E ... 1st light confinement layer, 15A-15E ... 2nd light confinement layer, 17A-17E ... Light-emitting region, 19A-19E ... Quantum wire, 21A to 21E: Intermediate semiconductor region, 25A to 25E ... Semiconductor mesa, 27 ... p-type cladding layer, 29 ... n-type cladding layer, 31 ... buried region, 31a ... first p-type current confinement layer, 31b ... n-type current Narrowing layer, 31c ... second p-type current confinement layer, 33 ... insulating film, 35 ... first electrode film, 37 ... semiconductor substrate, 39 ... second electrode film.

Claims (3)

A semiconductor laser array that outputs a plurality of laser beams having different wavelengths from each other,
A semiconductor substrate having a main surface;
A plurality of light emitting regions that are formed on the main surface of the semiconductor substrate with the direction along the main surface as a longitudinal direction, and generate each of the plurality of laser beams, and
Each of the plurality of light emitting regions has a plurality of quantum wires provided side by side in the longitudinal direction of each light emitting region with a period corresponding to the wavelength of the laser light generated by the light emitting region,
The fine line width in the longitudinal direction of each quantum thin line in each of the plurality of light emitting regions is formed in a dimension corresponding to the wavelength of the laser light generated by the light emitting region, and at least two of the plurality of light emitting regions A semiconductor laser array, wherein the thin line widths are different between light emitting regions.   The fine line width and the period of each quantum wire in at least a part of the plurality of light emitting regions are formed such that the gain peak wavelength and the Bragg wavelength in the light emitting region substantially coincide with each other. The semiconductor laser array according to claim 1. The fine line width and the period of each quantum wire in a part of the light emitting regions of the plurality of light emitting regions are formed such that the gain peak wavelength and the Bragg wavelength in the light emitting region substantially coincide with each other.
2. The semiconductor laser array according to claim 1, wherein the thin line widths in two or more light emitting regions other than the part of the light emitting regions are equal to each other in the light emitting regions.

Images