JPH0964054A - Manufacture of bipolar transistor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the hydrogenating rate of carbon impurity by etching except an emitter region, exposing a base layer, heat treating it at a high temperature in an inert gas atmosphere after an emitter mesa is formed, and releasing the hydrogen contained in the base layers in and out of the emitter region by lateral diffusion. SOLUTION: An HBT layer structure is formed on an InP substrate. Then, an emitter electrode 7 is formed of metal, with the electrode as a mask an InGaAs emitter cap layer 1 and an InP emitter layer 2 are etched, and a base layer 3 is exposed. Then, it is heat treated at a range of 500 to 600 deg.C for one minute or longer in a nitrogen atmosphere. This heat treatment is conducted to laterally diffuse hydrogen existed directly under an emitter mesa, the hydrogen arrives at the mesa end, and then releases out of the emitter mesa. Thus, the activation of carbon in the layer 3 can be attempted to be improved, and the element characteristics can be attempted to be improved by the realization of higher base layer carrier concentration.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for manufacturing an ultrafast bipolar transistor using a III-V group compound semiconductor.

[0002]

2. Description of the Related Art A heterojunction bipolar transistor (HBT) using a wide-gap semiconductor for an emitter layer is
It has excellent characteristics as a high-speed and high-frequency element and is expected to be used in various applications. When configuring an HBT,
Combination of GaAs and AlGaAs, InP and In
The constitution of the semiconductor material based on the combination with GaAs is the most important in terms of practicality, and is therefore widely used.

Since carbon (C), which is a group IV atom, has a small diffusion constant in the crystal and is capable of high-concentration doping, it has recently attracted attention as a p-type dopant. . Several methods are used as a method for growing a semiconductor layer to which carbon is added as an impurity. Among them, a vapor phase growth method such as metal organic chemical vapor deposition (MOCVD) or vapor phase epitaxial growth (VPE) is an apparatus. It is superior to other growth methods such as molecular beam epitaxial growth (MBE) and liquid phase epitaxial growth (LPE) in that it is large in size and easy to replenish raw materials, and is widely used industrially. . Usually, hydrogen is often used as a carrier gas for vapor phase growth because it has a characteristic that a gas of higher purity than that of other gases can be obtained. Also, III-V
AsH is used as a group V raw material in vapor phase growth of group semiconductors.
Many hydrides such as ₃ and PH ₃ are used. Therefore, the dissociated active hydrogen is present in a high concentration in the growth chamber. In this situation, TMAs, TBAs, T
Even if an organic metal material such as MP or TBP is used, dissociated hydrogen is generated, and the same occurs. As one of the adverse effects on carbon impurities, there is known a problem of inactivation due to hydrogenation of a dopant. This is a phenomenon in which hydrogen, which is present in a high concentration in the growth atmosphere, bonds with carbon impurities in the crystal to terminate the bond of carbon and cancel the carrier supply from carbon. This phenomenon is particularly noticeable in crystals grown at low temperatures. For example, InGaA
Regarding s, it becomes more remarkable in the high In composition region where lower temperature growth is required. Therefore, when trying to add C at a high concentration to the InGaAs layer lattice-matched to the InP substrate, most of the impurity carbon is inactivated by hydrogenation, and the resistivity of the carbon-added layer becomes high. There was a point.

On the other hand, it is known that a heat treatment in nitrogen is effective as a method for reactivating carbon deactivated by hydrogenation. However, even if another semiconductor layer is grown on the carbon-added layer and then heat-treated in nitrogen, the desorption of hydrogen is hindered by the space charge in the other semiconductor layer. The problem is that it cannot be simpled. This phenomenon becomes remarkable especially when the other semiconductor layer is n-type doped. A method of interrupting the growth and performing heat treatment in nitrogen is also conceivable, but in general, the purity of nitrogen is considerably lower than that of hydrogen, and when introduced into a growth apparatus, the inside of the growth apparatus and the crystal surface are contaminated. It causes problems such as. In addition, since nitrogen has a lower thermal conductivity than hydrogen, the actual temperature of the substrate surface deviates from a desired value, and the in-plane uniformity of the substrate temperature deteriorates. Furthermore, when the growth is restarted after dehydrogenation in a nitrogen atmosphere, active hydrogen is regenerated from various raw materials, so that hydrogenation of the carbon-added layer occurs again, and as a result, the resistivity of the carbon-added layer is reduced. The problem was that it could not be lowered.

[0005]

In order to solve the above problems, in a method of manufacturing a heterojunction bipolar transistor having a base layer containing carbon as an impurity,
The purpose is to obtain a technique for reducing the hydrogenation rate of carbon impurities.

[0006]

In order to achieve the above-mentioned object, the present invention performs etching at a portion other than the emitter region to expose the base layer to form an emitter mesa, and then performs high temperature heat treatment in an inert gas atmosphere to form the emitter region. The hydrogen contained in the inner and outer base layers is desorbed by lateral diffusion.

The base layer semiconductor material is an InGaAs layer, and the shape and dimensions of the emitter region are set so that the shortest distance to the mesa edge is 3 μm or less, or the heat treatment is performed within a range of 500 ° C. to 600 ° C. in a nitrogen atmosphere. This is done for 1 minute or more.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION When the present invention is applied to the manufacturing process of InP / InGaAs-HBT, it will be explained.
As emitter cap layer 1, n-InP emitter layer 2,
C-doped p-InGaAs base layer 3, i-InGaA
The s collector layer 4 and the n-InGaAs collector electrode layer 5 are stacked on the InP substrate 6 to form the emitter electrode 7, the base electrode 8 and the collector electrode 9, respectively. After each layer is grown on the semi-insulating InP substrate 6, an emitter pattern is formed of metal to serve as a mask, and the InGaAs emitter cap layer 1 and the InP emitter layer 2 are etched to expose the base layer 3. At this time, if an appropriate solution is used, only InP can be selectively etched.
At the P / InGaAs interface, etching can be stopped with good controllability on the upper surface of the base layer.

Next, heat treatment is carried out in a nitrogen atmosphere to remove hydrogen remaining in the carbon-added crystal.
Improves the activation rate of carbon as an acceptor. WS
By masking a refractory metal such as i, alloy reaction between the emitter electrode material and the semiconductor during heat treatment can be suppressed.
In addition, the use of the metal mask enables the emitter electrode and the emitter mesa to be formed in a self-aligned manner, which can improve the performance of the device. After that, a pattern is formed with a photoresist, and electrodes are formed by etching, lift-off or the like to complete the device.

In the heat treatment step, the heat treatment temperature is 500.
It was confirmed that the dehydrogenation can be effectively carried out by performing the treatment at a temperature of from 600 ° C to 600 ° C for 1 minute or more. In addition, it was confirmed that the hydrogen atom bonded to the carbon atom in the crystal was released from the crystal. Hydrogen existing immediately below the emitter mesa is prevented from being desorbed due to the n-type doped emitter layer and diffuses laterally. After reaching the mesa edge, hydrogen is desorbed outside the crystal. Also, under the above heat treatment conditions, the diffusion distance of hydrogen is 3
The above effect can be obtained in the same manner as long as the distance is at least μm and the shortest distance to the mesa end is at most 3 μm.

[0011]

EXAMPLE Next, a method for manufacturing a bipolar transistor according to the present invention will be described based on an example applied to a manufacturing process of InP / InGaAsHBT. 1A to 1C are cross-sectional views for explaining the above-described embodiment, and FIGS. 1A to 1C are views showing respective steps.

In FIG. 1, 1 is an n-InGaAs emitter cap layer, 2 is an n-InP emitter layer, 3 is a C-doped p-InGaAs base layer, and 4 is i-InGaAs.
Collector layer, 5 is an n-InGaAs collector electrode layer, 6
Is an InP substrate, 7 is an emitter electrode, 8 is a base electrode, 9
Indicates a collector electrode, and the layer structure shown in FIG. 1A has a semi-insulating property I by a method such as MOCVD or MBE.
Grows on nP substrate. Next, an emitter pattern is formed of metal, and the InGaAs emitter cap layer 1 and the InP emitter layer 2 are etched using this as a mask to expose the base layer 3 as shown in FIG. 1B. At this time, by using an appropriate solution such as a mixed solution of hydrochloric acid and water, only the InP layer can be selectively etched, and etching is stopped on the upper surface of the base layer with good controllability at the InP / InGaAs interface. be able to.

Next, heat treatment is performed at 500 ° C. for 5 minutes in a nitrogen atmosphere. By performing this heat treatment, the alloy reaction between the emitter electrode material and the semiconductor during the heat treatment can be suppressed. The structure using metal for this mask is
Since the emitter electrode and the emitter mesa can be formed in a self-aligned manner, it is suitable for improving the performance of the device. After the above steps, the element is completed as shown in FIG. 1C through a pattern forming step using photoresist, a wet or dry etching step, an electrode forming step such as lift-off and the like.

The inventors of the present invention conducted a careful study on the assumption that the conditions under which the dehydrogenation can be effectively carried out in the above heat treatment step have a deep correlation with the temperature and time thereof. It was confirmed to be effective as shown in FIG. 2 when the temperature was in the range of 600 ° C. It was also confirmed that the same effect was obtained when the heat treatment time was 1 minute or more. The carrier concentration of the sample used here was about 3 × 10 ¹⁸ / cm ³ immediately after growth, but the carrier concentration was improved to about 3 × 10 ¹⁹ / cm ³ by performing heat treatment at 500 ° C. for 5 minutes. . Also,
From the result of the atomic concentration analysis in the crystal, it was confirmed that these changes were caused by the desorption of the hydrogen atom bonded to the carbon atom in the crystal from the crystal. On the other hand, hydrogen existing directly under the emitter mesa is desorbed by the presence of the n-type doped emitter layer, and therefore diffuses in the lateral direction to reach the mesa edge and then desorbs to the outside of the crystal. As a result of experiments conducted by the inventors, when the above heat treatment conditions are used, the hydrogen diffusion distance becomes 3 μm or more, that is, the emitter mesa size is such that the shortest distance to the mesa end is 3 μm.
It has been found that the above effect can be obtained similarly if the thickness is less than or equal to μm. For example, in the case of forming a rectangular emitter mesa as shown in FIG. 3, if the length of the shorter side is 6 μm or less, the distance to the mesa end is 3 in all the emitter mesa regions.
Since the thickness is less than μm, the dehydrogenation from the base layer can be completely performed.

In this embodiment, WSi is used as the metal for the emitter mesa mask, but other heat resistant metals such as W and Mo can be used. Although the case of forming the emitter mesa by using the metal mask is described, a photoresist can be used as the mask. In that case, the photoresist may be removed after the mesa formation and before the heat treatment. The emitter electrode can be formed separately after the heat treatment. On the other hand, regarding the material system, InP / InGaAsHB
Although the case of T has been described, the method of the present invention can be widely used for HBTs having a double hetero structure and HBTs using other materials such as AlGaAs / GaAs. Although the case where nitrogen is used as the atmosphere gas used for the heat treatment is described, the same effect can be obtained by using another inert gas such as argon or helium.

[0016]

As described above, according to the method of manufacturing a bipolar transistor of the present invention, a collector layer, a base layer having a p-type conductivity type in which carbon is added as an impurity, and an n-type conductivity type are formed on a substrate. In a method for manufacturing a bipolar transistor in which an emitter layer having the above is sequentially deposited by a vapor phase growth method and then formed using photolithography and an etching technique, the base layer is exposed by etching except for the emitter region, and then an inert layer is formed. A heat treatment for activating the carbon impurities is performed in a gas atmosphere, and hydrogen contained in the base layer inside and outside the emitter region is desorbed by lateral diffusion to form a compound semiconductor in which carbon is added as an impurity to the base layer. When manufacturing the HBT used, it is possible to improve the activation rate of carbon in the base layer and It is possible to improve the device characteristics due realization of the base layer carrier concentration.

[Brief description of drawings]

FIG. 1 is a cross-sectional view of an embodiment in which a method of manufacturing a bipolar transistor according to the present invention is applied to a manufacturing process of InP / InGaAsHBT, and (a), (b), (c) are views showing respective manufacturing processes.

FIG. 2 is a diagram showing the heat treatment temperature dependence of carrier concentration in a C-doped InGaAs layer.

FIG. 3 is a plan view of an emitter mesa shape for explaining an emitter mesa dimension to which the present invention can be applied.

[Explanation of symbols]

 2 n-type emitter layer 3 p-type base layer 4 collector layer 6 substrate

Claims (3)

[Claims] 1. A collector layer, a base layer having a p-type conductivity type doped with carbon and impurities, and n on a substrate.
In a method for manufacturing a bipolar transistor in which an emitter layer having a conductivity type is sequentially deposited by a vapor phase epitaxy method and then formed by using photolithography and an etching technique, a portion other than the emitter region is etched to expose a base layer. After that, activation heat treatment of the carbon impurities is performed in an inert gas atmosphere, and hydrogen contained in the base layer inside and outside the emitter region is desorbed by lateral diffusion. . 2. The base layer doped with carbon as an impurity is an InGaAs layer, and the heat treatment step is performed for 1 minute or more within a range of 500 ° C. to 600 ° C. in a nitrogen atmosphere. Manufacturing method of bipolar transistor. 3. The bipolar transistor according to claim 1, wherein the shape of the emitter region left by the etching is such that the shortest distance from an arbitrary point in the emitter region to the edge of the emitter region is 3 μm or less. Manufacturing method.