JP5375983B2 - Manufacturing method of semiconductor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a thickness of a metamorphic layer caused by diffusion/segregation of In in the hetero-interface between a compound semiconductor including In and a compound semiconductor not including In. <P>SOLUTION: An LD 40 of the present invention comprises: an n-semiconductor substrate 12; a first n-cladding layer 42a that is disposed above the n-semiconductor substrate 12 and is formed of n-AlGaInP or n-GaInP; a second n-cladding layer 42b of n-AlGaAs disposed above the first n-cladding layer 42a; and an insertion layer 42c that is interposed between the second n-cladding layer 42b and the first n-cladding layer 42a, is formed of n-AlGaAsP corresponding to the first n-cladding layer 42a, has the same composition of Al and Ga of the group III element as that of the second n-cladding layer 42b, and in which the amount of P is larger than the amount of As in the composition of As and P of the group V element. <P>COPYRIGHT: (C)2012,JPO&amp;INPIT

Description

  The present invention relates to a semiconductor device and a manufacturing method thereof, and more particularly to a semiconductor device having a hetero interface between a compound semiconductor containing In and a compound semiconductor not containing In and a manufacturing method thereof.

In recent years, as communication broadbandization has progressed and public communication networks using optical fibers have become widespread, it has been increasingly required to transmit a large amount of information at a low cost. In order to increase the amount of information transmitted in response to such a request, it is necessary to increase the transmission rate, and the transmission rate is gradually increased from 600 Mbps to 2.5 Gbps, and further to 10 Gbps. Yes.
In addition, with the background of improving the communication speed of optical communication devices, the optical communication network is expanding not only to trunk line systems but also to access systems such as offices and homes. As one of the semiconductor devices, light emission and light reception used in optical transceivers Devices are required to be fast, cheap and efficient.
As a semiconductor optical device, for example, a semiconductor laser is formed by crystal growth of a compound semiconductor on an InP substrate, a GaAs substrate, or a GaN substrate.

A typical compound semiconductor is a group III-V compound semiconductor in which a group III element and a group V element are bonded, and a mixed crystal compound having various composition ratios by bonding a plurality of group III atoms or group V atoms. A semiconductor is obtained.
For example, there are InGaAsP, GaAsP, GaPN, GaNAs, InGaN, AlGaN, AlGaInP, InGaP, and the like.
These compound semiconductors are formed by the vapor phase growth process on the above-described compound semiconductor substrate. However, different LDs (Light Emitting Diodes) and LED (Light Emitting Diodes) are always adjacent when different compound semiconductors are formed. It is necessary to form a heterointerface that
Examples of the heterointerface include, for example, an interface between InGaN and AlGaN formed on the InGaN (hereinafter, the heterointerface having such a positional relationship is expressed as AlGaN / InGaN), AlGaAs / InGaP, and AlGaAs. / GaAs.
In the heterointerface composed of these two compound semiconductors, it is ideal that the first compound semiconductor changes rapidly from one compound semiconductor to the second compound semiconductor. However, in practice, the first compound semiconductor is transformed from the first compound semiconductor. It changes into a 2nd compound semiconductor through a layer. Here, this metamorphic layer means a layer whose composition cannot be controlled, which is unintentionally generated when the second compound semiconductor is formed on the first compound semiconductor to form a heterojunction.

In general, in a semiconductor device that requires a heterojunction, the thinner the metamorphic layer, the better the characteristics of the semiconductor device. Ideally, it is desirable that the composition ratio is changed at one interface from the first compound semiconductor to the second compound semiconductor at the heterointerface and there is no metamorphic layer.
For example, AlGaAs / GaAs is an example of a heterointerface used for an optical element. In this AlGaAs / GaAs, since the first compound semiconductor is GaAs and the second compound semiconductor is AlGaAs, this heterointerface is formed only of Al and Ga as group III elements and As as a group V element. Yes. At the time of crystal growth, at such a hetero interface, group III elements can be switched sharply, so that there is no or very little metamorphic layer. Therefore, the optical element is deteriorated due to the metamorphic layer. Does not occur.
However, in the case of AlGaN / InGaN, AlGaAs / InGaP, etc.
(I) The presence or absence of In changes at these heterointerfaces. That is, the group III element changes from In and Ga to Al and Ga, and In is not contained in the upper layer. Also,
(Ii) The type of group V element further changes at the heterointerface in the latter case. That is, the group V element changes from P to As.
When crystal growth is performed in the above cases (i) and (ii), it is difficult for the group III element and the group V element to switch sharply at the heterointerface and to perform crystal growth. This is because of diffusion of In from a compound semiconductor containing In to a compound semiconductor not containing In, segregation of In at the heterointerface, mutual diffusion of In and V group atoms, mutual interaction between V group atoms promoted by the presence of In. This is because diffusion occurs.

  As a known example of a semiconductor device that has been treated to prevent diffusion of In from a group III-V compound semiconductor layer containing In and a manufacturing method thereof, Te is contained on the group III-V compound semiconductor layer containing In. There is an example in which one In diffusion prevention film made of a II-IV group compound semiconductor layer is formed and a II-IV group compound semiconductor layer is formed thereon (for example, Patent Document 1, paragraph number [0006] ] And FIG.

JP 2000-91707 A

A semiconductor device, for example, an LD, has an n-type GaAs semiconductor substrate (hereinafter, “n-type” is “n−” and “p-type” is “p-”, in particular, undoped with no impurities. Is represented by “i-”), n-AlGaAs n-buffer layer, n-AlGaInP n-first cladding layer, n-AlGaAs n-second cladding layer, and i-AlGaAs n-side guide. Active layer structure including quantum well and barrier layer of AlGaAs, p-side guide layer of i-AlGaAs, p-first cladding layer of p-AlGaAs, p-second cladding layer of p-GaInP, p-AlGaInP The p-third cladding layer, the p-InGaP p-BDR layer (BDR: Band Distinuity Reduction), and the p-GaAs contact layer are formed by vapor phase growth. It is stacked.
In this LD laminated structure, the junction surface between the n-first cladding layer of n-AlGaInP and the n-second cladding layer of n-AlGaAs is a heterojunction surface, and the n-first cladding layer has a group III. In is included as an element, and the n-second cladding layer does not include In as a group III element. Further, in this case, the n-first cladding layer contains P as a group V element, and the n-second cladding layer does not contain P as a group V element and is completely different from the n-first cladding layer. Is replaced by As.
In such a hetero interface, a metamorphic layer is formed, and light emitted from the LD is absorbed to reduce the luminous efficiency. Depending on the metamorphic layer having a high electrical resistance value, the threshold current of the LD is increased. When the metamorphic layer has high strain, it may cause deterioration of the LD.

  The present invention was made to solve the above-mentioned problems, and a first object is to provide a heterojunction surface between a group III-V compound semiconductor layer containing In and a group III-V compound semiconductor layer not containing In. A second object is to form a semiconductor device in which diffusion and segregation of In is reduced as much as possible. When forming a semiconductor device, III-V containing In by reducing the diffusion and segregation of In at the heterojunction interface as much as possible. It is an object of the present invention to provide a production method for easily laminating a group III-V compound semiconductor layer not containing In on a group compound semiconductor layer.

In the method for manufacturing a semiconductor device according to the present invention, a first compound semiconductor layer containing a first group III element containing In and a first group V element and lattice-matching with a semiconductor substrate is formed on a semiconductor substrate. Following the step and the step of forming the first compound semiconductor layer, a heterojunction with the first compound semiconductor layer is formed on the first compound semiconductor layer, and In of the first group III elements is added. forming a third compound semiconductor layer and a second III-group element and a first V-group element different from the second group V element of the first group V element Toko not containing a third compound following the step of forming the semiconductor layer, on the third compound semiconductor layer, a well as the third compound semiconductor layer and the hetero-junction, and the second group III element and a second V-group element first and forming a second compound semiconductor layer seen including, third In the step of forming the compound semiconductor layer, the flow rate of the second group III element source gas is set to be the same as the flow rate of the second group III element source gas in the next step of forming the second compound semiconductor layer. And setting the flow rate of the second group V element source gas higher than the flow rate of the first group V element source gas to form a third compound semiconductor layer having a predetermined film thickness, In the step of forming the semiconductor layer, the flow rate of the source gas of the second group III element in the step of forming the third compound semiconductor layer is maintained, the supply of the source gas of the first group V element is stopped, The second compound semiconductor layer is formed by setting the flow rate of the second group V element to a predetermined flow rate for forming the second compound semiconductor layer .

  In the semiconductor device according to the present invention, the diffusion and segregation of In is suppressed by disposing the third compound semiconductor layer, and the metamorphic layer is hardly formed on the heterojunction surface. As a result, deterioration of the optical characteristics, electrical characteristics, reliability, and the like of the semiconductor device can be suppressed.

1 is a perspective view of a semiconductor laser according to an embodiment of the present invention. It is a fragmentary sectional view of the hetero interface vicinity in the II-II cross section of the semiconductor laser of FIG. It is a graph which shows the SIMS profile in the hetero interface vicinity of LD which concerns on one embodiment of this invention. It is a graph which shows the SIMS profile which raised the sensitivity of the density | concentration of the vertical axis | shaft of FIG. It is a graph which shows the SIMS profile in the hetero interface vicinity of the conventional LD. It is a graph which shows CL spectrum in the hetero interface vicinity of LD which concerns on one embodiment of this invention. It is a graph which shows CL spectrum in the hetero interface vicinity in conventional LD. It is a perspective view of the semiconductor laser which shows the modification of the semiconductor laser which concerns on one embodiment of this invention. 1 is a perspective view of a semiconductor laser according to an embodiment of the present invention. FIG. 10 is a partial cross-sectional view in the vicinity of a heterointerface in the XX cross section of the semiconductor laser of FIG. 9. It is a graph which shows the SIMS profile in the hetero interface vicinity of LD which concerns on one embodiment of this invention. It is a graph which shows the SIMS profile which raised the sensitivity of the density | concentration of the vertical axis | shaft of FIG.

In the following embodiments, an optical semiconductor device, for example, an LD will be described as an example. However, the present invention is not limited to an optical semiconductor device, but a semiconductor device having a hetero interface between a compound semiconductor containing In and a compound semiconductor containing no In. The same effect is achieved when applied to the whole.
Embodiment 1 FIG.
FIG. 1 is a perspective view of a semiconductor laser according to an embodiment of the present invention. FIG. 2 is a partial cross-sectional view in the vicinity of the heterointerface in the II-II cross section of the semiconductor laser of FIG. In each figure, the same reference numerals indicate the same or equivalent ones.
In FIG. 1, this LD 10 is a waveguide ridge type LD, a buffer layer 14 formed of an n-semiconductor on an n-semiconductor substrate 12, an n-cladding layer 16 lattice-matched with the n-semiconductor substrate 12, The active region 18 is sequentially stacked.
A p-cladding layer 20 is disposed on the active region 18. In the LD 10 of this embodiment, a part of the p-cladding layer 20 is formed in a ridge shape, and the widthwise center of both end faces of the LD 10 is formed. The part extends between the end faces.

A p-BDR (BDR: Band Distinuity Reduction) layer 22 (hereinafter referred to as a BDR layer) is disposed on the p-cladding layer 20 formed in the ridge shape. A contact layer 24 is provided.
A waveguide ridge is formed by a part of the p-cladding layer 20 formed in a ridge shape, the p-BDR layer 22 and the p-contact layer 24.
Further, a p-side electrode 26 is disposed on the p-contact layer 24, and an n-side electrode 28 is disposed on the back surface of the n− semiconductor substrate 12.
In this LD 10, the n-semiconductor substrate 12 is formed of n-GaAs. For example, the impurity is Si, the impurity concentration is about 0.5-1.5 × 10 ¹⁸ / cm ³ , and the layer thickness is about 500-700 nm. is there.
The buffer layer 14 is formed of n-GaAs, the impurity is Si, the impurity concentration is about 0.5-1.5 × 10 ¹⁸ / cm ³ , and the layer thickness is about 100-300 nm.

The hetero interface on the n side of the LD 10 is included in the n-cladding layer 16 and includes the characteristic configuration of the present invention.
Referring to FIG. 2, the n-cladding layer 16 is laminated in close contact with the first n-cladding layer 16a as the first compound semiconductor layer laminated in close contact with the buffer layer 14 and the first n-cladding layer 16a. The insertion layer 16c as the third compound semiconductor layer and the second n-cladding layer 16b as the second compound semiconductor layer stacked in close contact with the insertion layer 16c are formed. The original hetero interface is formed by the first n-cladding layer 16a and the second n-cladding layer 16b.
Here, in the LD 10, the first n-cladding layer 16a is formed of, for example, n-AlGaInP, the impurity is Si, the impurity concentration is about 0.1-0.5 × 10 ¹⁸ / cm ³ , and the layer thickness is 2000- It is about 4000 nm. Here, the first n-cladding layer 16a is formed of n-AlGaInP, but may be formed of GaInP not containing Al.
The insertion layer 16c is formed of n-AlGaInP corresponding to the n-AlGaInP of the first n-cladding layer 16a. For example, the impurity is Si and the impurity concentration is about 0.1-0.5 × 10 ¹⁸ / cm ^3. Thus, the layer thickness is about 5-40 nm, more preferably about 5-10 nm.
In the n-AlGaInP constituting the insertion layer 16c, the composition of Al, Ga, P is the same as the composition of Al, Ga, P of the first n-cladding layer 16a, and only the composition of In is the first n-cladding layer 16a. It is less than the composition of In.
That is, when the In composition ratio x1 of the first n-cladding layer 16a is set to 0.5, the insertion layer 16c has the same composition values of Al, Ga, and P as those of the first n-cladding layer 16a. The composition ratio x2 is less than x1 (= 0.5). In this case, X2 may be 0.

The insertion layer 16c is formed of n-AlGaInP corresponding to the first n-cladding layer 16a. When the first n-cladding layer 16a is formed of n-GaInP, the insertion layer 16c is formed of n- The n-GaInP formed of GaInP and constituting the insertion layer 16c has the same composition of Ga and P as the composition of Ga and P of the first n-cladding layer 16a, and only the composition of In is the first n-cladding layer 16a. It is the same as the case where the insertion layer 16c described above is formed of n-AlGaInP that the composition of In is less than that of In.
The second n-cladding layer 16b is made of, for example, n-AlGaAs, the impurity is Si, the impurity concentration is about 0.05-0.15 × 10 ¹⁸ / cm ³ , and the layer thickness is about 100-200 nm.
That is, in the configuration near the heterointerface of the LD 10, the first n-cladding layer 16a is formed of n-AlGaInP or n-GaInP, the second n-cladding layer 16b is formed of n-AlGaAs, and the insertion layer 16c is The n-AlGaInP or n-GaInP is formed corresponding to the 1n-cladding layer 16a, and Al and Ga of the insertion layer 16c have the same composition as the first n-cladding layer 16a, and only the composition of In is the first n-cladding layer 16a. When the In composition ratio x1 of the first n-cladding layer 16a is set to x1 = 0.5, the insertion layer 16c has a composition value of Al, Ga, and P of the first n-cladding layer 16a. The In composition ratio x2 is less than x1 (= 0.5) while keeping the same as the layer 16a.

In other words, in the configuration near the heterointerface, the first n-cladding layer 16a is formed of n-AlGaInP or n-GaInP, and the second n-cladding layer 16b is formed of n-AlGaAs. The layer 16c is formed of n-AlGaInP or n-GaInP corresponding to the first n-cladding layer 16a, and Al and Ga of the insertion layer 16c have the same composition as the first n-cladding layer 16a, and only the composition of In is the first. The strain is less than the In composition of the 1n-cladding layer 16a and exceeds −1.0% with respect to the GaAs semiconductor substrate 12 for crystal growth (for example, −1.5%, −2%, etc.). Distortion amount).
The active region 18 includes a first guide layer stacked in close contact with the second n-cladding layer 16b, a first active layer stacked in close contact with the first guide layer, and stacked in close contact with the first active layer. The second guide layer, the second active layer stacked in close contact with the second guide layer, and the third guide layer stacked in close contact with the second active layer.
The first guide layer, the second guide layer, and the third guide layer are each formed of i-AlGaAs, and the layer thickness is 5-15 nm. The first active layer and the second active layer are each formed of i-AlGaAs, and the layer thickness is 7-9 nm.

The p-cladding layer 20 is in close contact with the first p-cladding layer stacked in close contact with the third guide layer, the second p-cladding layer stacked in close contact with the 1p-cladding layer, and the 2p-cladding layer. The third p-cladding layer is laminated.
The first p-cladding layer is made of p-AlGaAs, and the impurities are Zn, C, Mg, etc., for example, the impurity concentration is about 1.0-2.0 × 10 ¹⁸ / cm ³ and the layer thickness is about 100-200 nm. It is.
The second p-cladding layer is made of p-InGaP, and the impurities are Zn, C, Mg, etc., for example, the impurity concentration is about 1.0-2.0 × 10 ¹⁸ / cm ³ and the layer thickness is about 20-100 nm. It is.
The third p-cladding layer is made of, for example, p-AlInGaP, the impurities are Zn, C, Mg, etc., the impurity concentration is about 1.0-2.0 × 10 ¹⁸ / cm ³ , and the layer thickness is about 1000-2000 nm. It is. The third p-cladding layer may be formed of p-InGaP.
The p-BDR layer 22 is formed of, for example, p-InGaP, the impurity is Zn, C, Mg, etc., the impurity concentration is about 0.1-7.0 × 10 ¹⁸ / cm ³ , and the layer thickness is about 10-50 nm. It is.
The p-contact layer 24 is formed of, for example, p-GaAs, the impurities are Zn, C, Mg, etc., the impurity concentration is about 20-40 × 10 ¹⁸ / cm ³ , and the layer thickness is about 200-300 nm.

Next, a manufacturing method will be described.
The laminated structure of the LD 10 is such that a buffer layer 14 is formed on an n-semiconductor substrate 12, and an n-cladding layer 16, an active region 18, and a p-cladding layer lattice-matched with the n-semiconductor substrate 12 on the buffer layer 14. 20, the p-BDR layer, and the p-contact layer 24 are formed by metal organic chemical vapor deposition (hereinafter referred to as MOCVD method), molecular beam epitaxy (hereinafter referred to as MBE method), or the like. It is formed by.
Here, the MOCVD method will be described as an example. An n-GaAs substrate as an n-semiconductor substrate 12 on which crystal growth is desired is introduced into a reactor of an MOCVD apparatus.
Thereafter, thermal energy is applied to the n-GaAs substrate, and crystal growth is performed at a growth temperature (ie, temperature of the GaAs substrate) of, for example, 700 ° C.
As raw materials supplied for crystal growth of the laminated structure of LD10, trimethylindium (TMI), trimethylgallium (TMG), trimethylaluminum (TMA), phosphine (PH3), arsine (AsH3), silane (SiH4) diethyl zinc These materials are thermally decomposed in a reaction furnace using (DEZn) or the like, and a group III-V compound semiconductor containing Al, Ga, In, As, and P is grown on a GaAs substrate.
These source gases are adjusted in flow rate using a mass flow controller or the like to adjust the composition of each layer, and the n-clad layer 16, the active region 18, the p-clad layer 20, the p-BDR layer, and the p-contact layer 24. Are stacked.
Among these layers, the buffer layer 14 excluding the n-cladding layer 16, the active region 18, the p-cladding layer 20, the p-BDR layer, and the p-contact layer 24 are formed by a known manufacturing process.

The n-cladding layer 16 is formed as follows.
An n-GaAs substrate as an n-semiconductor substrate 12 is doped with Si as a buffer layer 14 and has an impurity concentration of about 0.5-1.5 × 10 ¹⁸ / cm ³ and a layer thickness of about 500-700 nm. A GaAs layer is formed; As the buffer layer 14, the first n-cladding layer 16a has an impurity concentration of about 0.1-0.5 × 10 ¹⁸ / cm ³ and a layer thickness of about 2000-4000 nm on the n-GaAs layer. n-AlGaInP is formed. So far, it is performed by a known manufacturing method.
Next, the insertion layer 16c is formed when n-AlGaInP has a predetermined layer thickness as the first n-cladding layer 16a. The insertion layer 16c maintains the flow rates of trimethylaluminum (TMA), trimethylgallium (TMG), and phosphine (PH3) among the raw materials for forming n-AlGaInP as the first n-cladding layer 16a. Only the flow rate of (TMI) is made smaller than the flow rate when forming the first n-cladding layer 16a, and each source gas is supplied to the reactor, so that In is more than n-AlGaInP as the first n-cladding layer 16a. N-AlGaInP having a small composition only is formed as the insertion layer 16c. The n-AlGaInP layer as the insertion layer 16c has an impurity of Si, an impurity concentration of about 0.1-0.5 × 10 ¹⁸ / cm ³ , and a layer thickness of about 5-40 nm, more preferably about 5-10 nm. Is done.

In the case where the insertion layer 16c is formed, the amount of In when the In composition ratio x1 of the first n-cladding layer 16a is 0.5, the insertion layer 16c has a composition value of Al, Ga, and P of the first n-th cladding layer 16c. The In composition ratio x2 is set to less than x1 (= 0.5) while keeping the same as the cladding layer 16a.
In other words, when the insertion layer 16c is formed, a strain amount exceeding −1.0% (for example, −1.5%, −2%, etc.) with respect to the GaAs semiconductor substrate 12 on which crystal growth is performed. The amount of In is reduced so as to have a distortion amount.
When the insertion layer 16c having a predetermined thickness is formed, an impurity is Si and the impurity concentration is about 0.05 to 0.15 × 10 ¹⁸ / cm ³ as the second n-cladding layer 16b on the insertion layer 16c. N-AlGaAs having a thickness of about 100 to 200 nm is formed by a known manufacturing process.

Next, the effect of including the insertion layer 16c in the n-cladding layer of the LD according to the embodiment of the present invention will be described with reference to a graph showing SIMS profiles of the constituent elements of the n-cladding layer as an example.
FIG. 3 is a graph showing a SIMS profile in the vicinity of the heterointerface of the LD according to one embodiment of the present invention. FIG. 4 is a graph showing a SIMS profile in which the sensitivity of the concentration on the vertical axis in FIG. 3 is increased. FIG. 5 is a graph showing a SIMS profile in the vicinity of the hetero interface of a conventional LD.
FIGS. 3 and 4 are graphs showing SIMS (Secondary Ion Mass Spectrometry) profiles of the n-cladding layer 16 in the LD 10.
For comparison, FIG. 5 shows a conventional hetero interface, that is, the first n-cladding layer and the second n-cladding layer are formed without forming the insertion layer 16c, and the hetero interface is formed.
3, 4, and 5, the vertical axis indicates the concentration of each element, the unit is an arbitrary unit (au), and has a logarithmic scale. The horizontal axis is the depth from the surface of the second n-cladding layer 16b, and the second n-cladding layer 16b, the insertion layer 16c, and the first n-cladding layer 16a are arranged in the direction in which the scale of the horizontal axis increases. Has been. The n-GaAs substrate as the n-semiconductor substrate 12 is arranged in the direction in which the horizontal axis increases.
3, 4, and 5, the second n-cladding layer 16 b is formed of n-AlGaAs, but a graph in the case where the first n-cladding layer 16 a is formed of n-GaInP will be described. ing.
Although the material configuration is different from that of the first n-cladding layer 16a of the LD 10 described above, the technical features are essentially the same as when the first n-cladding layer 16a is made of n-AlGaInP.
Further, as the first n-cladding layer 16a is formed of n-GaInP, the insertion layer 16c is also formed of n-GaInP having a smaller In composition than the first n-cladding layer 16a.

In the SIMS profile of FIG. 3, the In and P concentrations suddenly rise from the noise region in the region where the depth from the surface is near 95 nm, and the In and P concentrations are almost stable at a little shallower than 110 nm. The depth from the surface to the vicinity of 95 nm is the region of the AlGaAs layer of the second n-cladding layer 16b, and the portion deep from the vicinity of 110 nm is the region of the InGaP layer of the first n-cladding layer 16a.
The transition region is between the AlGaAs layer of the second n-cladding layer 16b and the InGaP layer of the first n-cladding layer 16a, and the transition region includes an n-GaInP layer as the insertion layer 16c. Become.
FIG. 4 is a SIMS profile in which the sensitivity of the concentration of the vertical axis of In and P in FIG. 3 is increased, and portions from 5 to 7 on the scale of the vertical axis in FIG. 3 are shown.
In FIG. 4, the concentration of In gradually decreases from the InGaP layer side toward the AlGaAs layer, and in the portion A surrounded by a circle, a shallow valley is once formed and then rapidly rises to form a peak. The In concentration rapidly decreases from the depth of the peak position toward the surface of the AlGaAs layer. The value of In concentration at the peak position is slightly higher than the In concentration level of the InGaP layer.
Further, when the level of P at the depth at which the In concentration shows a peak is observed, the level of P in the InGaP layer is almost reached, and the concentration of In showing the peak at the A portion is the depth at this peak position. Considering the fact that the concentration of In rapidly decreases as it moves toward the surface of the AlGaAs layer, and that the concentration of P also decreases rapidly, the In concentration reaches a peak in the A portion. It is considered that the vicinity of the depth from the surface shown is close to the interface between the AlGaAs layer of the second n-cladding layer 16b and the n-GaInP layer as the insertion layer 16c.

FIG. 5 is a SIMS profile of an n-cladding layer of a conventional LD. The scale of the vertical axis in FIG. 5 matches the scale of the vertical axis in FIG. 4, and is compared with the SIMS profiles of In and P in FIG. Is shown in
That is, in the case of FIG. 5, the n-AlGaAs layer as the second n-cladding layer is formed directly on the n-GaInP layer as the first n-cladding layer, and the insertion layer 16c in the LD 10 is not formed. It is a SIMS profile in the case of.
In FIG. 5, the concentration of In gradually increases without decreasing from the InGaP layer toward the AlGaAs layer side, and as shown in the B portion surrounded by a circle, it rapidly increases to form a peak, As the depth from the peak position goes to the surface of the AlGaAs layer, the concentration of In decreases rapidly.
That is, the shallow In concentration valley shown in the A portion of the In SIMS profile in FIG. 4 does not appear in the In SIMS profile in FIG. Further, the height of the In concentration peak shown in the A portion of the In SIMS profile in FIG. 4 is almost the same as the In concentration level in the InGaP layer, whereas the In concentration in FIG. The height of the In concentration peak shown in part B of the SIMS profile is considerably higher than the In concentration level of the InGaP layer.

When the In SIMS profile in FIG. 4 is compared with the In SIMS profile in FIG. 5, the following is recognized.
That is, an n-GaInP insertion layer 16c having a smaller In composition than the first n-cladding layer 16a is formed on the n-GaInP layer of the first n-cladding layer 16a, and a second n-AlGaAs second n-cladding layer is formed thereon. When the layer 16b is formed, the In concentration in the insertion layer 16c near the interface between the second n-cladding layer 16b and the insertion layer 16c is approximately the same as the In concentration level in the InGaP layer.
Further, the n-GaInP insertion layer having a smaller In composition than the first n-cladding layer 16a so that a shallow valley of In concentration is observed in the SIMS profile near the interface between the second n-cladding layer 16b and the insertion layer 16c. It is considered that 16c remains.
Therefore, even if there is a layer in which In concentration is increased due to In diffusion or segregation in the insertion layer 16c in the vicinity of the interface between the second n-cladding layer 16b and the insertion layer 16c, the extent of the influence as the metamorphic layer is It is considered that the degree of influence of the metamorphic layer formed when the n-AlGaAs layer as the second n-cladding layer is formed directly on the n-GaInP layer as the first n-cladding layer is considerably smaller. It is done. Further, it is possible to form a heterointerface that rapidly changes from a material having the composition of the n-GaInP layer of the first n-cladding layer 16a to a material having the composition of n-AlGaAs of the second n-cladding layer 16b.

Therefore, an n-GaInP insertion layer 16c having a smaller In composition than the first n-cladding layer 16a is formed on the n-GaInP layer of the first n-cladding layer 16a, and a second n-AlGaAs second n-cladding layer is formed thereon. When the layer 16b is formed, stimulated emission light and spontaneous emission light from the active region 18 are absorbed to reduce the light emission efficiency of the LD, or depending on the metamorphic layer having a high electrical resistance value, the threshold current of the LD. In the case where the thickness becomes high or the metamorphic layer has a high strain, the deterioration of the LD is suppressed.
Next, the presence or absence of the metamorphic layer will be described using an example of a CL (Cathode Luminescence) spectrum in the n-cladding layer of the LD according to one embodiment of the present invention.

FIG. 6 is a graph showing a CL spectrum in the vicinity of the heterointerface of the LD according to one embodiment of the present invention. FIG. 7 is a graph showing a CL spectrum in the vicinity of a hetero interface in a conventional LD.
6 and 7, the horizontal axis represents the wavelength, and the vertical axis represents the light intensity of CL expressed in arbitrary units (au).
FIG. 6 shows an example in which a GaAs layer is formed as a buffer layer 14 on an n-GaAs substrate as an n-semiconductor substrate 12, and an AlGaInP layer as a first n-cladding layer 16a is formed on the GaAs layer. An insertion layer 16c made of n-GaInP having a smaller In composition than the first n-cladding layer 16a is formed on the first n-cladding layer 16a, and a second n-cladding layer of AlGaAs is formed on the insertion layer 16c. The layer 16b is formed.
In the CL measurement, the second n-clad layer 16b of AlGaAs is removed by wet etching, and the interface of the insertion layer 16c is exposed to obtain a CL spectrum.
FIG. 7 shows an example in which a GaAs layer is formed as a buffer layer 14 on an n-GaAs substrate as an n-semiconductor substrate 12, and an AlGaInP layer as a first n-cladding layer 16a is formed on the GaAs layer. An AlGaAs second n-cladding layer 16b is formed directly on the first n-cladding layer 16a.
In the CL measurement, as in the case of FIG. 7, the second n-cladding layer 16b of AlGaAs is removed by wet etching, and the interface of the first n-cladding layer 16a is exposed to obtain the CL spectrum.

In FIG. 6, the CL spectrum is observed only in the energy band of the optical band gap, but in FIG. 7, the spectrum is observed at a position other than the energy band of the optical band gap (position indicated by A in the figure).
That is, when an AlGaInP layer is formed as the first n-cladding layer 16a and the second n-cladding layer 16b of AlGaAs is formed directly on the first n-cladding layer 16a, a metamorphic layer is observed at the heterointerface. On the other hand, an AlGaInP layer is formed as the first n-cladding layer 16a, and an insertion layer 16c formed of n-GaInP having a lower In composition than the first n-cladding layer 16a is formed on the first n-cladding layer 16a. It can be seen that when the second n-cladding layer 16b made of AlGaAs is formed on the insertion layer 16c, no metamorphic layer is observed.

  The LD 10 according to this embodiment is disposed on an n-GaAs n-semiconductor substrate 12 and the n-semiconductor substrate 12, lattice-matched with the n-semiconductor substrate 12, and n-AlGaInP or n-GaInP. A first n-cladding layer 16a formed on the first n-cladding layer 16a, an n-AlGaAs second n-cladding layer 16b disposed on the first n-cladding layer 16a, and the second n-cladding layer 16b and the first n-cladding layer 16b. The n-AlGaInP or n-GaInP is formed corresponding to the first n-cladding layer 16a, and Al and Ga have the same composition as the first n-cladding layer 16a, and only the In composition. Includes an insertion layer 16c having a smaller In composition than the In composition of the first n-cladding layer 16a, and this configuration suppresses the formation of a metamorphic layer having a large In composition. Is, it is possible to construct a LD having a hetero-interface that changes sharply to a material having a composition of the 2n- cladding layer 16b of a material having a composition of the 1n- cladding layer 16a. As a result, it is possible to provide an LD in which deterioration of optical characteristics, electrical characteristics, reliability, and the like of the LD is suppressed.

  In addition, the manufacturing method of the LD 10 according to this embodiment includes a first n formed on an n-GaAs n-semiconductor substrate 12 and lattice-matched with the n-semiconductor substrate 12 and made of n-AlGaInP or n-GaInP. Among the raw materials for forming n-AlGaInP as the first n-cladding layer 16a on the first n-cladding layer 16a and the step of forming the cladding layer 16a, trimethylaluminum (TMA), trimethylgallium (TMG) And the flow rate of phosphine (PH3) is maintained as they are, or among the raw materials for forming n-GaInP as the first n-cladding layer 16a, the flow rates of trimethylgallium (TMG) and phosphine (PH3) are maintained as they are. Only the flow rate of indium (TMI) is applied to the first n-cladding layer 16a. A step of forming n-AlGaInP as the first n-cladding layer 16a with a lower In composition than the n-AlGaInP as the first n-cladding layer 16a as the insertion layer 16c, and a step of forming the insertion layer 16c. And forming a second n-cladding layer 16b of n-AlGaAs on the substrate, and by this manufacturing method, diffusion of In in the insertion layer 16c near the interface between the second n-cladding layer 16b and the insertion layer 16c is achieved. And an increase in In concentration due to segregation is suppressed, and the material having the composition of the n-GaInP layer of the first n-cladding layer 16a is sharply changed from the material having the composition of n-AlGaAs of the second n-cladding layer 16b. A heterointerface can be formed. As a result, an LD in which deterioration of optical characteristics, electrical characteristics, reliability, and the like of the LD is suppressed can be manufactured by a simple process of forming the insertion layer 16c.

Modification 1
FIG. 8 is a perspective view of a semiconductor laser showing a modification of the semiconductor laser according to one embodiment of the present invention.
FIG. 8 shows a buried ridge type LD 30. The laminated structure of the LD 30 is basically the same as that of the LD 10, but n-current blocking layers 32 are disposed on both sides of the waveguide ridge. The p-electrode extends not only on the waveguide ridge but also on the surface of the n-current blocking layer 32.
Since the n-cladding layer 16 of the LD 30 has the same configuration as that of the LD 10 and includes the insertion layer 16c, the same effect as that of the LD 10 can be obtained.

Modification 2
The blue LD in the second modification is an example in which the substrate material in the LD 10 is changed to a GaN substrate.
Also in the blue LD of this modification, the configuration of the LD 10 shown in FIG. 1 and the partial cross-sectional view of the semiconductor laser in the vicinity of the heterointerface shown in FIG. 2 are similar.
In this case, however, the n-semiconductor substrate 12 is made of n-GaN, and the buffer layer 14 formed thereon is made of n-GaN.
The heterointerface on the n-side of the blue LD is included in the n-cladding layer 16. As shown in FIG. 2, the n-cladding layer 16 is a first compound semiconductor layer stacked in close contact with the buffer layer 14. The first n-cladding layer 16a, the insertion layer 16c as a third compound semiconductor layer stacked in close contact with the first n-cladding layer 16a, and the second compound semiconductor stacked in close contact with the insertion layer 16c. The second n-cladding layer 16b as a layer is formed. The original hetero interface is formed by the first n-cladding layer 16a and the second n-cladding layer 16b as in the case of the LD10.
The first n-cladding layer 16a of the blue LD is made of, for example, n-AlGaInN. In this modification, the first n-cladding layer 16a is formed of n-AlGaInN, but may be formed of GaInN not containing Al.
The insertion layer 16c is made of n-AlGaInN corresponding to the first n-cladding layer 16a.

The n-AlGaInN constituting the insertion layer 16c has the same composition of Al, Ga, N as the composition of Al, Ga, N of the first n-cladding layer 16a, and only the composition of In is the first n-cladding layer 16a. It is less than the composition of In. That is, when the In composition ratio x1 of the first n-cladding layer 16a is set to 0.5, the insertion layer 16c has the same composition values of Al, Ga, and P as those of the first n-cladding layer 16a. The composition ratio x2 is less than x1 (= 0.5). In this case, X2 may be 0.
The insertion layer 16c is formed of n-AlGaInN corresponding to the first n-cladding layer 16a. When the first n-cladding layer 16a is formed of n-GaInN, the insertion layer 16c is formed of n− In the n-GaInN formed of GaInN and constituting the insertion layer 16c, the composition of Ga and P is the same as the composition of Ga and N of the first n-cladding layer 16a, and only the composition of In is the first n-cladding layer 16a. It is the same as the case where the insertion layer 16c described above is formed of n-AlGaInN. The second n-cladding layer 16b is made of, for example, n-AlGaN.
Accordingly, in the configuration in the vicinity of the hetero interface, the first n-cladding layer 16a is formed of n-AlGaInN or n-GaInN, the second n-cladding layer 16b is formed of n-AlGaN, and the insertion layer 16c is formed of the first n-cladding layer 16c. -Formed of n-AlGaInN or n-GaInN corresponding to the cladding layer 16a, Al and Ga of the insertion layer 16c have the same composition as the first n-cladding layer 16a, and only the composition of In is the first n-cladding layer 16a. When the In composition ratio x1 = 0.5 of the first n-cladding layer 16a is less than the In composition, the insertion layer 16c has the composition values of Al, Ga, and N as the first n-cladding layer. The In composition ratio x2 is made less than x1 (= 0.5) while keeping the same as 16a.
In this case, the presence or absence of In changes at the heterointerface, and the type of group V element does not change.

The active region 18 is formed of i-AlGaN, and the p-cladding layer 20 disposed on the active region 18 includes an AlGaN layer disposed on the active region side and an InGaN disposed on the AlGaN layer. And a heterojunction interface. Further, a p-BDR layer 22 and a p-contact layer 24 are disposed on the p-BDR layer 22.
Since the insertion layer 16c is also disposed at the hetero interface included in the n-cladding layer 16 of the blue LD, similarly to the LD 10, the formation of the metamorphic layer having a large In composition is suppressed, and the first n-cladding layer 16a is suppressed. An LD having a heterointerface that changes sharply from a material having the following composition to a material having the composition of the second n-cladding layer 16b can be configured. As in the case of the LD 10, it is possible to provide an LD in which deterioration of the optical characteristics, electrical characteristics, reliability, and the like of the LD is suppressed.

  The blue LD manufacturing method is basically the same as the LD 10 manufacturing method, in which ammonia is used in place of phosphine (PH3) and arsine (AsH3) as raw materials, and an n-GaN n-semiconductor substrate 12 is used. A step of forming a first n-cladding layer 16a formed of n-AlGaInN or n-GaInN, and a step of forming n-AlGaInN as a first n-cladding layer 16a on the first n-cladding layer 16a. Among these raw materials, the flow rates of trimethylaluminum (TMA), trimethylgallium (TMG), and ammonia are maintained as they are, or among the raw materials when forming n-GaInP as the first n-cladding layer 16a, trimethylgallium (TMG) And the flow rates of ammonia and trimethylindium (TM N-AlGaInN having a smaller In composition than n-AlGaInN or n-GaInN as the first n-cladding layer 16a. The method includes a step of forming as the insertion layer 16c and a step of forming the second n-clad layer 16b of n-AlGaN on the insertion layer 16c.

  By this manufacturing method, an increase in In concentration due to In diffusion and segregation in the insertion layer 16c in the vicinity of the interface between the second n-cladding layer 16b and the insertion layer 16c is suppressed, and the n-GaInN of the first n-cladding layer 16a is suppressed. It is possible to form a heterointerface that changes abruptly from the material having the layer composition to the material having the n-AlGaN composition of the second n-cladding layer 16b. As a result, as in the case of the LD 10, it is possible to manufacture an LD in which deterioration of optical characteristics, electrical characteristics, reliability, and the like of the LD is suppressed by a simple process of forming the insertion layer 16c.

As described above, a semiconductor device according to an embodiment of the present invention includes a semiconductor substrate, a group III element and a group V element that are disposed on the semiconductor substrate, lattice-matched with the semiconductor substrate, and include In. A first compound semiconductor layer including the first compound semiconductor layer, a second compound semiconductor layer including a group III element and a group V element other than In disposed on the first compound semiconductor layer, and the second compound The first compound semiconductor layer is interposed between the semiconductor layer and the first compound semiconductor layer, has the same constituent elements as the first compound semiconductor layer, and has only a smaller composition of In than the first compound semiconductor layer. And a third compound semiconductor layer that is heterojunction with the second compound semiconductor layer. With this configuration, the third compound semiconductor layer is disposed, so that In diffusion and segregation occur. Suppressed, metamorphic layer on heterojunction surface Less likely to be formed. As a result, deterioration of the optical characteristics, electrical characteristics, reliability, and the like of the semiconductor device can be suppressed.
According to another aspect of the present invention, there is provided a method for manufacturing a semiconductor device, comprising: forming a first compound semiconductor layer including a group III element including In and a group V element on a semiconductor substrate and lattice-matching with the semiconductor substrate; The first compound semiconductor layer is heterojunctioned with the first compound semiconductor layer, and has the same constituent elements as the first compound semiconductor layer, and only the composition of In is less than that of the first compound semiconductor layer. And a second compound comprising a group III element and a group V element excluding In that are heterojunction with the third compound semiconductor layer and formed on the third compound semiconductor layer. Forming a semiconductor layer, and forming the third compound semiconductor layer makes it difficult for In to diffuse and segregate, and makes it difficult to form a metamorphic layer on the heterojunction surface. As a result, a semiconductor device in which deterioration of optical characteristics, electrical characteristics, reliability, and the like is suppressed can be formed by a simple process.

Embodiment 2. FIG.
FIG. 9 is a perspective view of a semiconductor laser according to an embodiment of the present invention. FIG. 10 is a partial cross-sectional view in the vicinity of the heterointerface in the XX cross section of the semiconductor laser of FIG.
In FIG. 9, the LD 40 has basically the same configuration as the LD 10 in the first embodiment, but only the n-cladding layer 42 of the LD 40 having a heterointerface has a structure different from the n-cladding layer 16 of the LD 10. ing.
That is, the hetero interface on the n side of the LD 40 is included in the n-cladding layer 42, and the characteristic configuration of the present invention is included.
Referring to FIG. 10, the n-cladding layer 42 is laminated in close contact with the first n-cladding layer 42a as the first compound semiconductor layer laminated in close contact with the buffer layer 14 and the first n-cladding layer 42a. In addition, an insertion layer 42c as a third compound semiconductor layer and a second n-cladding layer 42b as a second compound semiconductor layer stacked in close contact with the insertion layer 42c are formed. The original heterointerface is formed by the first n-cladding layer 42a and the second n-cladding layer 42b.
Here, in the LD 40, the first n-cladding layer 42a is formed of, for example, n-AlGaInP, the impurity is Si, the impurity concentration is about 0.1-0.5 × 10 ¹⁸ / cm ³ , and the layer thickness is 2000− It is about 4000 nm. Here, the first n-cladding layer 42a is formed of n-AlGaInP, but may be formed of GaInP not containing Al.

The insertion layer 42c is formed of n-AlGaAsP corresponding to n-AlGaInP of the first n-cladding layer 42a. For example, the impurity is Si and the impurity concentration is about 0.1-0.5 × 10 ¹⁸ / cm ^3. Thus, the layer thickness is about 5-40 nm, more preferably about 5-10 nm.
The composition of n-AlGaAsP constituting the insertion layer 42c is determined by the composition of n-AlGaAs in the second n-cladding layer 42b formed subsequently on the insertion layer 42c.
That is, the composition ratio of Al and Ga of the group III element constituting the insertion layer 42c is made the same as the composition ratio of Al and Ga of the group III element constituting the second n-cladding layer 42b, and As and P of the group V element The composition of As is set so that As is greater than P, and P is 30% or less of the composition ratio of n-AlGaInP constituting the first n-cladding layer 42a at the maximum.
When the insertion layer 42c is formed with such a composition, the n-semiconductor substrate 12 on which crystal growth is performed has a strain amount exceeding -1.0% (for example, a strain amount of -1.5%, -2%, etc.). It will be.
The second n-cladding layer 42b is formed of n-AlGaAs, the impurity is Si, the impurity concentration is about 0.05-0.15 × 10 ¹⁸ / cm ³ , and the layer thickness is about 100-200 nm.

Therefore, in the configuration of the LD 40 near the n-side heterointerface, the first n-cladding layer 42a is formed of n-AlGaInP or n-GaInP, the second n-cladding layer 42b is formed of n-AlGaAs, and the insertion layer 42c is formed of n-AlGaAsP, and the composition ratio of the group III element Al and Ga is the same as the composition ratio of the group III element Al and Ga constituting the second n-cladding layer 42b, and the group V element As and The composition of P is such that As is greater than P, and P is 30% or less of the composition ratio of n-AlGaInP constituting the first n-cladding layer 42a at the maximum.
The other layers, which are the buffer layer 14, the active region 18, the p-cladding layer 20, the p-BDR layer 22, and the p-contact layer 24, have the same configuration as the LD 10.
Next, a manufacturing method will be described.
Since the manufacturing method of the LD 40 other than the n-cladding layer 42 is the same as that of the LD 10, the manufacturing method of the n-cladding layer 42 will be described.
The buffer layer 14 is formed on the n-GaAs substrate, and Si is used as the first n-cladding layer 42a on the buffer layer 14 so that the impurity concentration is about 0.1-0.5 × 10 ¹⁸ / cm ³ . N-AlGaInP having a layer thickness of about 2000 to 4000 nm is formed.

Next, the insertion layer 42c is formed when the n-AlGaInP has a predetermined thickness as the first n-cladding layer 42a.
In forming the insertion layer 42c, the flow rates of trimethylaluminum (TMA) and trimethylgallium (TMG), which are source gases of Group III elements Al and Ga constituting the insertion layer 42c, are set to the second n-cladding layer 42b. The flow rate is the same as that of trimethylaluminum (TMA) and trimethylgallium (TMG) when forming.
The raw material is such that the composition ratio of the group V elements As and P is greater than P, for example, P is 30% or less of the composition ratio of n-AlGaInP constituting the first n-cladding layer 42a at the maximum. Set the gas flow rate. As a result, the insertion layer 42 c has a strain amount exceeding −1.0% (for example, a strain amount of −1.5%, −2%, etc.) with respect to the n− semiconductor substrate 12.
The n-AlGaAsP layer as the insertion layer 42c has an impurity of Si, an impurity concentration of about 0.1-0.5 × 10 ¹⁸ / cm ³ , and a layer thickness of 5-40 nm, more preferably about 5-10 nm. The
When the insertion layer 42c having a predetermined thickness is formed, the second n-cladding layer 42b is formed on the insertion layer 42c as an impurity of Si and an impurity concentration of about 0.05 to 0.15 × 10 ¹⁸ / cm ^3. N-AlGaAs having a thickness of about 100 to 200 nm is formed by a known manufacturing process. At this time, the flow rates of trimethylaluminum (TMA) and trimethylgallium (TMG), which are source gases of Group III elements Al and Ga, are maintained at the insertion layer 42c, and the supply of phosphine (PH3) is stopped. The amount of arsine (AsH3) is increased to a predetermined value.

Next, the effect obtained by providing the insertion layer 42c in the n-cladding layer of the LD according to one embodiment of the present invention will be described with reference to a graph showing SIMS profiles of the constituent elements of the n-cladding layer as an example.
FIG. 11 is a graph showing a SIMS profile in the vicinity of the heterointerface of the LD according to one embodiment of the present invention. FIG. 12 is a graph showing a SIMS profile in which the sensitivity of the concentration on the vertical axis in FIG. 11 is increased.
11 and 12, the vertical axis indicates the concentration of each element, the unit is an arbitrary unit (au), and is a logarithmic scale. The horizontal axis is the depth from the surface of the second n-cladding layer 42b, and the second n-cladding layer 42b, the insertion layer 42c, and the first n-cladding layer 42a are arranged in the direction in which the scale of the horizontal axis increases. Has been. The n-GaAs substrate as the n-semiconductor substrate 12 is arranged in the direction in which the horizontal axis increases.
In the SIMS profile of FIG. 11, when the depth from the surface is in the vicinity of 95 nm, the concentration of In and P increases rapidly from the noise region, and when the depth from the surface exceeds 100 nm and becomes a little deeper, Therefore, the depth from the surface to the vicinity of 95 nm is the AlGaAs layer region of the second n-cladding layer 42b, and the deeper part from the vicinity where it is slightly deeper than 100 nm is the first n−. This is a region of the AlGaInP layer of the cladding layer 42a.
The transition region is between the AlGaAs layer of the second n-cladding layer 42b and the InGaP layer of the first n-cladding layer 42a, and the transition region includes an n-AlGaAsP layer as the insertion layer 42c. Become.

FIG. 12 is a SIMS profile in which the sensitivity of the concentration on the vertical axis of In and P in FIG. 11 is increased, and portions from 3 to 7 on the scale on the vertical axis in FIG. 11 are shown.
In FIG. 12, the In concentration is maintained at the same level from the AlGaInP layer side to the AlGaAs layer, and the In concentration is slightly higher in the AlGaAsP layer in the boundary region between the AlGaInP layer and the AlGaAsP layer surrounded by a circle. After the low peak is formed, the concentration of In decreases rapidly from the depth of the peak position toward the surface of the AlGaAs layer.
When the In concentration at this peak is compared with the In concentration at the In peak at B in the SIMS profile of the n-cladding layer of the conventional LD shown in FIG. Considering the slight difference and the difference in sensitivity on the vertical axis, the concentration of In in the n-AlGaAsP layer as the insertion layer 42c is considered to be relatively low.
That is, Al is a group III element that is formed of n-AlGaAsP on the n-AlGaInP layer of the first n-cladding layer 42a and has a composition ratio of Al and Ga of the group III element constituting the second n-cladding layer 42b. An insertion layer 42c in which the composition of the group V elements As and P is larger than that of P is formed, and the n-AlGaAs second n-cladding layer 42b is formed on the insertion layer 42c. Is formed, the In concentration in the insertion layer 42c in the vicinity of the interface between the insertion layer 42c and the first n-cladding layer 42a is slightly higher, but is much higher than the In concentration in the first n-cladding layer 42a. Must not.

Even if there is a layer in which the formation of the metamorphic layer is suppressed in the insertion layer 42c in the vicinity of the interface between the first n-cladding layer 42a and the insertion layer 42c, there is a diffusion or segregation of In, and the In concentration is slightly increased. The degree of influence as the metamorphic layer is determined by the effect of the metamorphic layer generated when the n-AlGaAs layer as the second n-cladding layer is formed directly on the n-AlGaInP layer as the first n-cladding layer. It is considered to be considerably smaller than the degree.
Further, it is possible to form a heterointerface that changes sharply from a material having the composition of the n-AlGaInP layer of the first n-cladding layer 42a to a material having the composition of n-AlGaAs of the second n-cladding layer 42b.
Therefore, like the LD 40, it is formed on the n-AlGaInP layer of the first n-cladding layer 42a by n-AlGaAsP, and the composition ratio of the group III elements Al and Ga is made of the group III group constituting the second n-cladding layer 42b. An insertion layer 42c having the same composition ratio of Al and Ga as the elements and the composition of As and P of the group V element in which As is larger than P is formed, and the second n-clad layer 42b of n-AlGaAs is formed thereon. Is formed, the stimulated emission light and spontaneous emission light from the active region 18 are absorbed to reduce the luminous efficiency of the LD, or the metamorphic layer having a high electrical resistance value, as in the LD 10 in the first embodiment. In some cases, the threshold current of the LD is increased, and when the metamorphic layer has a high strain, the deterioration of the LD is suppressed.

  In the LD 40 according to this embodiment, an n-GaAs substrate as the n-semiconductor substrate 12 is disposed on the n-semiconductor substrate 12, lattice-matched with the n-semiconductor substrate 12, and n-AlGaInP or a first n-cladding layer 42a formed of n-GaInP, a second n-cladding layer 42b of n-AlGaAs disposed on the first n-cladding layer 42a, a second n-cladding layer 42b, It is interposed between the first n-cladding layer 42a, the composition ratio of Al and Ga is the same as the composition ratio of Al and Ga, which is a group III element constituting the second n-cladding layer 42b, and the group V element As and The composition of P is such that As is more than P and the first n-cladding layer 42a and the insertion layer 42c heterojunction with the second n-cladding layer 42b are provided. Since the formation of the metamorphic layer is suppressed in the insertion layer 42c in the vicinity of the interface between the first n-cladding layer 42a and the insertion layer 42c, the second n− is formed from a material having the composition of the n-AlGaInP layer of the first n-cladding layer 42a. It is possible to configure an LD having a heterointerface that changes sharply to a material having the n-AlGaAs composition of the cladding layer 42b. As a result, it is possible to provide an LD in which deterioration of optical characteristics, electrical characteristics, reliability, and the like of the LD is suppressed.

  In addition, the LD 40 manufacturing method according to this embodiment includes a first method in which n-AlGaInP or n-GaInP is formed on the n-GaAs substrate as the n-semiconductor substrate 12 and lattice-matched with the n-semiconductor substrate 12. A step of forming a 1n-cladding layer 42a and a step of forming an insertion layer 42c on the first n-cladding layer 42a, and trimethylaluminum (TMA) which is a source gas of a group III element Al and Ga; The flow rate of trimethylgallium (TMG) is set to the same flow rate as that of trimethylaluminum (TMA) and trimethylgallium (TMG) when the second n-cladding layer 42b to be formed on the insertion layer 42c later is formed. And the step of forming the insertion layer 42c such that the composition ratio of P and As is larger than P, and the top of the insertion layer 42c. As the flow rate of trimethylaluminum (TMA) and trimethylgallium (TMG) which are source gases of Group III elements Al and Ga, the flow rate at the time of the insertion layer 42c is maintained, and the supply of phosphine (PH3) is stopped. Correspondingly, the step of forming the second n-cladding layer 42b of n-AlGaAs by increasing the amount of arsine (AsH3) to a predetermined value. According to the method for manufacturing a semiconductor device, n-AlGaInP and In the vicinity of the heterojunction surface with n-AlGaAs, an increase in In concentration due to In diffusion or segregation is suppressed, and a heterointerface that changes sharply from n-AlGaInP to n-AlGaAs can be formed. As a result, an LD in which deterioration of optical characteristics, electrical characteristics, reliability, and the like of the LD is suppressed can be manufactured by a simple process of forming the insertion layer 42c.

As described above, the semiconductor device according to this embodiment includes a semiconductor substrate, a group III element and a first group V element that are disposed on the semiconductor substrate, lattice-matched with the semiconductor substrate, and include In. A first compound semiconductor layer including the first compound semiconductor layer, and a second group V element disposed on the first compound semiconductor layer and including a group III element excluding In and a second group V element excluding the first group V element. A second compound semiconductor layer, a second compound semiconductor layer interposed between the second compound semiconductor layer and the first compound semiconductor layer, wherein the composition ratio of the group III element is the same as the composition ratio of the second compound semiconductor layer. And a first group V element having a composition ratio equal to or smaller than the composition ratio of the compound semiconductor layer and the second group V element, and the first compound semiconductor layer and the second compound A third compound semiconductor layer heterojunction with the semiconductor layer As hereinbefore, the third compound semiconductor layer In the diffusion and segregation by is arranged by the configuration is suppressed, hard metamorphic layer is formed at the heterojunction surface. As a result, deterioration of the optical characteristics, electrical characteristics, reliability, and the like of the semiconductor device can be suppressed.
In the semiconductor device manufacturing method according to this embodiment, a first compound semiconductor layer containing a group III element containing In and a first group V element and lattice-matching with the semiconductor substrate is formed on the semiconductor substrate. And a step of forming a heterojunction with the first compound semiconductor layer on the first compound semiconductor layer, and a group III element and a first group V element excluding In constituting the first compound semiconductor layer, And forming a third compound semiconductor layer containing a second group V element different from the first group V element that contains more than the composition of the first group V element, and a third compound semiconductor layer In addition to the heterojunction with the third compound semiconductor layer, the third compound semiconductor layer includes a group III element and a second group V element except for In and the first group V element. A step of forming a compound semiconductor layer of 2. Diffusion and segregation of In hardly occurs by forming a third compound semiconductor layer, hard metamorphic layer is formed at the heterojunction surface. As a result, a semiconductor device in which deterioration of optical characteristics, electrical characteristics, reliability, and the like is suppressed can be formed by a simple process.
In the above example, the case where GaAs and GaN are used as the substrate has been described, but InP may be used as the substrate.

  As described above, the semiconductor device and the manufacturing method thereof according to the present invention are suitable for a semiconductor device having a hetero interface between a compound semiconductor containing In and a compound semiconductor not containing In and a manufacturing method thereof.

  12 n-semiconductor substrate, 16a, 42a First n-cladding layer, 16b, 42b Second n-cladding layer, 16c, 42c Insertion layer.

Claims (3)

Forming a first compound semiconductor layer including a first group III element including In and a first group V element on the semiconductor substrate and lattice-matching with the semiconductor substrate;
Following the step of forming the first compound semiconductor layer, the first compound semiconductor layer is heterojunctioned with the first compound semiconductor layer and the first group III element does not contain In. forming a third compound semiconductor layer containing a second III-group element and a first V-group element different from the second group V element of the first group V element Toko,
Subsequent to the step of forming the third compound semiconductor layer , the second compound group element and the second group V element are heterojunctioned with the third compound semiconductor layer on the third compound semiconductor layer. look including a step of forming a second compound semiconductor layer consisting of,
In the step of forming the third compound semiconductor layer, the flow rate of the source gas of the second group III element is equal to the flow rate of the source gas of the second group III element in the step of forming the second compound semiconductor layer. Set the same, the flow rate of the source gas of the second group V element is set higher than the flow rate of the source gas of the first group V element, and a third compound semiconductor layer having a predetermined thickness is formed,
In the step of forming the second compound semiconductor layer, the flow rate of the source gas of the second group III element in the step of forming the third compound semiconductor layer is maintained, and the source gas of the first group V element is supplied. A method of manufacturing a semiconductor device, wherein the second compound semiconductor layer is formed by stopping and setting the flow rate of the second group V element to a predetermined flow rate for forming the second compound semiconductor layer . The method for manufacturing a semiconductor device according to claim 1 , wherein the third compound semiconductor layer is stacked to a thickness of 40 nm or less. In addition to In, Ga is included as the group III element constituting the first compound semiconductor layer, P is included as the first group V element, and Al and Ga are included as the group III element constituting the second compound semiconductor layer. the method of manufacturing a semiconductor device according to claim 1 or 2, characterized in that the containing as as the second group V element.

Images