JPS6381305A - Optical integrated circuit - Google Patents

Abstract

PURPOSE:To obtain an optical integrated circuit which permits easy manufacture and has high coupling efficiency between passive waveguide and active elements by forming waveguide layers so as to be larger in band gap and higher in refractive index than multiple quantum well structure. CONSTITUTION:An n<+>-GaAs buffer layer 2, n<+>-AlGaAs clad layer 3, an n- AlGaAs guide layer 4, an i-GaAs/AlAs multiple quantum well (MQW) clad layer 5, a p-AlGaAs layer 6, and p<+>-GaAs cap layer 7 are continuously grown on an n<+>-GaAs substrate 1 and after an ohmic electrode 8 is formed on the p side, an optical circuit pattern 9 is formed as the rib shape arriving at the middle of the n-AlGaAs guide layer 4. The layers are then removed down to the n<+>-AlGaAs clad layer 3 so as to form a reflection surface 10 in the bent parts of the respective optical circuit parts. The ohmic electrode 8, the p<+>-GaAs cap layer 7, the p-AlGaAs layer 6 and the MQW clad layer 5 are thereafter removed in the parts except switches 11a, 11d at the center, and an n side ohmic electrode 12 is formed on the rear surface of the substrate; finally, inlet and exit end faces are formed by cleavage.

Description

[Detailed description of the invention] (Industrial application field) The present invention relates to optical integrated circuits made of semiconductor materials.

(Prior art) Optical integrated circuits integrate various optical elements on one substrate to realize functions, and are expected to be realized as optical communication systems become more popular in recent years. . Especially G
■-V group compound semiconductor materials such as aAQAs and InGaAsP are excellent as materials for monolithic optical integrated circuits because they can realize all functions such as light emission, light reception, switching, and waveguide.

In conventional optical integrated circuits using semiconductor materials, methods for coupling active elements such as light emitting, light receiving, and switching devices with passive optical waveguides include (1) Directional coupling (Japanese Patent Laid-Open No. 50-15928
7) (2) Due to leakage coupling (11 journals ``IEEE Journal of Quantum Electronics (IEEE Journal of Quantum Electronics)
tum Electronics) Vol. QE-15, 1979, pp. 72-82)
Lied Physics Letter-rs) No. 27
Vol., 1975, pp. 241-243).

(Problems to be solved by the invention) Among these methods, (1) has strict conditions for setting the layer thickness and refractive index, and (3) uses buried growth so that the positions of the active and passive waveguides match. On the other hand, in (2), an active waveguide is partially formed on a passive waveguide and is relatively easy to manufacture, but the active waveguide has a higher refractive index. is high, and the coupling efficiency is low because it utilizes light leakage from the active waveguide to the passive waveguide.

An object of the present invention is to eliminate such problems and provide an optical integrated circuit that is easy to manufacture and has high coupling efficiency between a passive waveguide and an active element.

(Means for Solving the Problems) The optical integrated circuit according to the present invention includes: a multiple quantum well structure and a means for applying an electric field or injecting a current to the multiple quantum well structure, which is a waveguide formed on a semiconductor substrate. The multi-quantum well structure has a first semiconductor layer having a thickness of approximately the de Broglie wavelength sandwiched between second semiconductor layers having a larger band gap than the first semiconductor layer. It is composed of multiple quantum wells stacked in the layer thickness direction; the waveguide layer is characterized by having a larger band gap and higher refractive index than the multiple quantum well structure.

(Function) The present invention utilizes the characteristics of the multiple quantum well (MQW) structure: ■ The band gap and refractive index can be controlled independently. ■ Both the real and imaginary parts of the complex refracting weight n-jk change by applying an electric field. This is what I did. First, this point will be explained.

The bandgap of the MQW structure will be explained with reference to FIG. FIG. 2(a) shows an energy band diagram of one quantum well of the MQW. In the case of MQW as well, if the barrier layer thickness is made sufficiently thick, each well can be considered as an isolated state, and can be considered using the same energy band diagram as in FIG. 2(a). This quantum well has a depth of ΔE3 and Δ for electrons in the conduction band and holes in the valence band, respectively.
It acts as a quantum well for E7, forming discontinuous energy levels in each. For simplicity, we will use electrons and heavy holes ()1. H,) and light holes (L, H,), respectively.
It is assumed that levels are formed one by one. FIG. 2(b) shows the density of states, and in the bulk state, what is represented by a parabolic function as shown by the dashed-dotted line becomes a step-like distribution as shown by the solid line when it becomes a quantum well. According to this, the band gap also changes from E, which is in the pulck state, to E,. Changes to W.

The magnitude of this bandgap change is determined by the quantum well depth Δ
Eg.

It can be controlled by ΔE, the effective mass of holes and electrons in the well, and the well thickness. In other words, once the barrier and well materials are determined, they can be controlled primarily by the well thickness.

On the other hand, the refractive index of the MQW structure can be approximated by the refractive index of a mixed crystal composed of all atoms constituting the MQW.

That is, the refractive index can be controlled by the ratio of barrier and well materials and their thicknesses. Therefore, in the MQW structure, the band gap and refractive index can be controlled independently. Normal mixed crystals have a large bandgap and a low culm refractive index, but MQW
With this method, a layer with a small band gap and a low refractive index can be obtained.

Next, changes in the complex refraction weight due to the electric field of the MQW will be explained.

FIG. 3 is a characteristic diagram obtained by measuring changes in the optical absorption spectrum when an electric field E is applied perpendicularly to the QW structure in a GaAs/AQGaAsM QW structure. The electric field E tilts the potential structure of the QW, causing quantum level movement and electron/hole wave functions to shift within the well, resulting in an increase in the absorption coefficient at wavelengths longer than the absorption edge, and an increase in the absorption coefficient at wavelengths shorter than the absorption edge. Conversely, a decrease in the absorption coefficient occurs. On the other hand, complex refractometer -n
-jk has a relationship with the absorption coefficient α of -λα/4π (where λ is the wavelength), and n and k are related by the Kramers-Kronig relationship. Therefore, a change in the absorption coefficient α as described above also brings about a change in the real part n of the refractive index.

FIG. 4 shows an outline of the spectrum of absorption coefficient change (Δα) and refractive index change (Δn) at a certain electric field intensity based on such a relationship. Note that λ is the bandgap wavelength of the MQW, and it can be seen that both the refractive index and the absorption coefficient positively change or increase on the wavelength side longer than λ.

Considering the characteristics of MQW as described above, MQW can be used as a cladding when no electric field is applied to a waveguide layer with a higher refractive index and a larger band gap, and when an electric field is applied, it can be used as an absorption type optical gate. - Can be used as a switch or photodetector.

Furthermore, if a layer with a wider band gap than the MQW and a conductivity type different from that of the optical waveguide layer is laminated on top of the MQW on the optical waveguide layer, a double hetero p-n structure can be formed, and the MQW can be formed by injecting forward current. can give a gain. Most of the power of the light is in the optical waveguide layer, so the required current is larger than that of a normal LD, but this structure allows laser oscillation,
Amplifier operation can be obtained.

The present invention utilizes such characteristics of MQW, and will be explained in detail below with reference to Examples.

(Embodiment) FIG. 1(a) is a perspective view showing an embodiment of the optical integrated circuit according to the present invention, and shows an example applied to a two-manpower, two-output branch gate type matrix switch. Here, a case where GaAQAs/GaAs-based material is used will be explained as an example. First, the manufacturing method of this embodiment will be explained.

An n ”-GaAs buffer layer 2 and an n ”-AQGaAs cladding layer 3 (A1
1 molar ratio x = 0.31 refractive index n = 3.43), n-AQ
GaAs guide layer 4 (x-0, 21, n=3.46),
i-GaAs/AQAsM QW cladding layer 5 (
Gaas well: 115 people thick, AQAs barrier: 115 people thick x 30 cycles), p-AQGaAs layer 6 (xwO
, 31, n=3.43), a p4''-GaAs cavity layer 7 is continuously grown by the MBE method, and an ohmink electrode 8 is formed on the zero-order p side. By etching (RIBE), an optical circuit pattern 9 as shown in FIG.
1(b) to form a reflective surface (for example, 10) at the bent portion of each optical circuit section by RIBE.
) is removed until it reaches the n''-AQGaAs cladding layer 3. In the RIBE method, an etching surface with high perpendicularity is obtained at almost constant speed regardless of the A1 molar ratio. After this, the central switch 11a is etched.

11b, lie, lid, ohmic electrode 8, p''-GaAs skive layer 7 p-A
The QGaAs layer 6 and the MQW cladding layer 5 were removed, an n-side ohmic electrode 12 was formed on the back surface of the substrate, and finally, input and output end surfaces were formed by cleaving.

Next, the operation of this embodiment will be explained. The optical circuit pattern remains on the n-AQGaAs guide layer 4 as a lip guide. The light incident on the two input-side lip guides 20a and 20b from the left side of FIG. 1(a) is branched by two light branches 13a and 13b using mirror surfaces, respectively, and the central switches lla, llb, llc
, enters the lid. When no electric field is applied to each switch, the light leaking into the MQW cladding layer 5 is small and the absorption coefficient is also small, so almost no light is absorbed, and the incident light is output as is.

On the other hand, when an electric field is applied, the absorption coefficient and refractive index of the MQW cladding layer 5 increase and no light is emitted, as described above. Each switch lla, llb.

The lights emitted from the light beams 11c and 11c and the lid are combined by merging units 14a and 14b using mirrors, and output to output side lip guides 20c and 20d. Here each switch 1
1a.

By changing the combination of operating states of 11b, lie, and lid, the input side and output side guides 20a, 2
Any connection state between 0b, 20c, and 20d can be realized.

The production of this example does not require buried growth or strict control of layer thickness and refractive index, which is required for the production of directional couplers.
Furthermore, although the light distribution within the waveguide differs somewhat between the switch portion and other portions, the light is distributed around the n-GaAQAs guide layer 4, and loss at the connection portion can be suppressed to a small level.

In this embodiment, a case using GaAs/AQGaAs-based materials has been described, but it is clear that the present invention is applicable to other material systems as well.

The optical switch of this embodiment has a p-1-n structure and can also be used as a photodetector when an electric field is applied, and in that case, an electrical signal corresponding to the optical signal can be extracted.

In addition, the i-MQW cladding layer 5 has a wider bandgap than that and has a double hetero pn structure sandwiched between Al1GaAs layers 4.6 of different conductivity types, so the electrode 8,
By applying a forward bias between 12 and 12, gain can be given to the optical signal. In this case, the MQW layer 5 is not located at the center of the light distribution, and gain cannot be obtained unless the injection current is increased. However, it is known that in the MQW structure, the value at which laser oscillation does not occur decreases; can be kept smaller than when it is not used as an active layer. It is also possible to compensate for loss by installing an MQW clad optical amplifier having such a structure in the output section of the optical switch. In this case as well, the manufacturing of the optical amplifier section is the same as that of the optical switches lla, llb, lie, and lid described above, and no major difficulties arise due to the optical amplifier convergence.

In this embodiment, the present invention was used to realize a 2×2 optical branch/gate type switch, but it goes without saying that it can also be applied to realize other optical circuits and functions.

(Effects of the Invention) As described above in detail, according to the present invention, it is possible to obtain an optical integrated circuit that is easy to manufacture and has a high coupling efficiency between a passive optical waveguide and an active element portion.

[Brief explanation of the drawing]

FIG. 1(a) is a perspective view showing an embodiment of the optical integrated circuit according to the present invention, FIG. 1(b) is a diagram showing a circuit pattern formed in the manufacturing process of this embodiment, and FIG. MQW
The energy band diagram of one of the quantum wells, FIG. 2(b) is a diagram showing the relationship between the density of states and energy in the quantum well, and FIGS. 3 and 4 are the MQs used in the present invention.
FIG. 3 is a characteristic diagram showing the characteristics of a W layer. In the figure, 1, 2, 3, 4, 5, 6.7 are semiconductors, 8.
12 are electrodes, lla, llb, lie, l
id is a switch, 10 is a mirror, 13a, 13b are optical branches, 14a, 14b are combiners, 20a, 20b
, 20c, and 20d are guides, and 9 is an optical circuit pattern. (b) Figure 1 Figure 2 (b) Figure 4

Claims (1)

[Claims] A multiple quantum well structure and means for applying an electric field or injecting a current to the multiple quantum well structure are partially formed on a waveguide layer formed on a semiconductor substrate; , a first semiconductor layer having a thickness of approximately the de Broglie wavelength is replaced with a second semiconductor layer having a larger band gap than this first semiconductor layer.
1. An optical integrated circuit comprising quantum wells sandwiched between semiconductor layers stacked in multiple layers in the layer thickness direction; said waveguide layer having a larger band gap and higher refractive index than said multi-quantum well structure.