JPH02210820A - Manufacture of bipolar transistor - Google Patents

Abstract

PURPOSE:To install a hole barrier of silicon oxide film whose thickness is precisely controlled, on the interface between a base layer and an emitter layer, by continuously performing a first process wherein a silicon oxide film is formed by simultaneously projecting silicon molecular beam and oxygen molecular beam, and a second process wherein polysilicon is formed by projecting silicon molecular beam, in the same vacuum vessel. CONSTITUTION:By simultaneously projecting Si molecular beam and O2 molecular beam on the clean surface of single crystal Si 12, or by performing low pressure thermal oxidation wherein oxygen is sent on a heated substrate, thin SiO2 13 is formed. After O2 is discharged without taking out the object from the vacuum vessel to the air, the Si molecular beam is continuously projected. Thus polysilicon 14 is formed on the SiO2 film in the manner in which the polysilicon is buried in the thin SiO2 film. Said oxide film functions as a hole barrier. Thereby, the SiO2 film whose thickness is precisely controlled and thinner than a natural oxide film can be buried between the base and the polysilicon emitter of an NPN transistor, so that the current amplification factor can be increased.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a method for manufacturing a bipolar transistor using molecular beam growth, and more particularly to a method for manufacturing a polysilicon emitter.

(Prior Art) In recent years, polysilicon has been used as an emitter material for NPN bipolar transistors. Baren is eye
E.E. Transactions on Electron Devices ED-32, page 1307 (P.

Halen and D. L. Pu1frey.
IEBE Transaction on Electr
on Devices, ED-32 (1985) 1
307. ), it is known that when a polysilicon emitter is used, the current amplification factor is improved by up to 20 times compared to when a single crystal emitter is used. This is Balaton's IE Transactions on Electron Devices ED-33 Volume 17
Page 54 (G, L, Patton, J, C,
Bravman and J, D, Plummer
.

IEEE Transaction on Elect
ron Devices, ED-33 (1986) 1
754. ) Vapor phase growth method (CVD)
This is thought to be due to the fact that when polysilicon is deposited, the natural oxide film on the surface of the single crystal base is embedded in the interface between the single crystal base and the polysilicon emitter, and this extremely thin oxide film acts as a hole barrier. ing. However, the tunneling probability of the oxide film barrier for electrons and holes is highly dependent on the oxide film thickness, and for example, as described by Green, Solid State Electronics, Vol.
, F., D., King and J.

Shewchun, 5solid 5tate Ele
ctron, 17 (1974) 551. ), if the oxide film thickness changes by two degrees, the tunneling current changes by an order of magnitude. Therefore, in order to use an oxide film as a hole barrier in a bipolar transistor,
It is necessary to control the film thickness to 2 Å or less in a region as thick as a natural oxide film so that it does not act as a barrier to electrons but acts as a barrier to holes. In the present circumstances,
Such precise control is difficult, and there is a problem in that the current amplification factor varies greatly depending on the method of cleaning the substrate before depositing polysilicon.

In addition, in recent years, growth has been performed at lower temperatures than conventional silicon thin film growth techniques, and growth at the atomic layer level has been developed for the purpose of application to high-speed bipolar devices, microwave devices, superlattice structure devices, etc. Silicon molecular beam growth (S) in high vacuum is characterized by its controllable properties.
iMBE) technology is being actively researched and developed. This MB
In the E technology, by simultaneously irradiating an oxygen molecular beam and a silicon molecular beam, it is possible to form a thin SiO3 film whose thickness is extremely precisely controlled. Also,
A similar film can also be formed by low-pressure thermal oxidation in which a heated substrate is irradiated with oxygen. However, if a 5102 film with a precisely controlled thickness equal to or thinner than a native oxide film is formed on the base layer by these methods, and then a polysilicon emitter is formed by the conventional method of CVD, , when the wafer is taken out into the atmosphere from the MBE chamber, oxidation on the base layer progresses, and C
There is also a problem in that there is no difference compared to the case where a natural oxide film is used because it is exposed to high temperature and high pressure inside the VD device.

(Problems to be Solved by the Invention) The purpose of the present invention is to eliminate such conventional drawbacks,
An object of the present invention is to provide a method for producing a polysilicon emitter in which a hole barrier of an extremely thin silicon oxide film whose thickness is precisely controlled is provided at the interface with a base layer.

(Means for solving the problem) The present invention forms a silicon oxide film by simultaneously irradiating silicon molecular beams and oxygen molecular beams onto the base layer surface after cleaning the substrate surface where the base layer is exposed. The first step is to form polysilicon by irradiating silicon molecular beams onto the substrate containing the silicon oxide film.
A method for manufacturing a bipolar transistor is characterized in that the above steps are performed consecutively in the same vacuum chamber or in vacuum chambers connected by a vacuum, and after cleaning the substrate surface where the base layer is exposed, the substrate is heated. The first step is to form a silicon oxide film on the surface of this base layer by low-pressure thermal oxidation with oxygen irradiation, and the second step is to form polysilicon by irradiating a silicon molecular beam onto the substrate containing the silicon oxide film. The present invention provides a method for manufacturing a bipolar transistor, characterized in that the above steps are successively carried out in the same vacuum chamber or in vacuum chambers connected by a vacuum.

(Operation) First, the principle of the present invention will be explained. In Figure 2, the single crystal 5i which is the base layer introduced into the ultra-high vacuum chamber
A natural oxide film 11 exists on the layer 12 (a).

When this is heated in an ultra-high vacuum, the natural oxide film 11 evaporates, and a clean Si surface appears (b). By simultaneously depositing Si molecular beams and 0□ molecular beams on this clean surface, or by performing low-pressure thermal oxidation that sends oxygen to the heated substrate, an extremely thin 5i
o213 is formed (C). When taken out from the vacuum chamber into the atmosphere in this state, oxidation on the base surface progresses and the oxide film thickness becomes the same as the natural oxide film.

Therefore, the inventor discovered that if the Si molecular beam was continuously irradiated after evacuating 0□ without taking it out of the vacuum chamber, a thin 5102 film could be embedded on the 5102 film as shown in (d). It was discovered for the first time that polysilicon 14 was formed and this oxide film functioned as a hole barrier. Once buried, there is no change in 5102 even if it is exposed to the atmosphere, and if As is diffused into the upper polysilicon layer, an emitter is formed. Since the effective mass for electrons of the thin SiO2 barrier between the emitter and base formed by the above method is smaller than the effective mass for holes, if the thickness of the barrier is appropriately selected, only electrons can pass through and holes can be suppressed. That's why. Note that the process of forming SiO2 and the process of forming polysilicon can be performed in separate vacuum chambers, but in this case, connect the re-vacuum chambers with a vacuum tunnel to prevent the substrate from being exposed to the atmosphere and oxidizing. There must be.

(Example) Next, embodiments of the invention will be specifically described.

First, the formation of an extremely thin 5102 film embedded in Si will be described. The experiment was conducted using a 40cc electron gun type Si
This was carried out using an MBE apparatus equipped with a vapor deposition device. Sample wafer has 4 inch n and p type 5i (100) 0.01~
A 0.02 Ωcm substrate was used. The sample wafer is a normal R
After CA cleaning, it was transported into the formation chamber and 10 layers of A-8i were deposited, followed by a cleaning process for 800901 minutes, the clean surface was exposed, and an epitaxial layer as a buffer layer was grown at a growth temperature of 500°C for 3000 layers. . After lowering the substrate temperature to room temperature, oxygen with a purity of 99.9999% was leaked into the formation chamber from a nozzle, and a Si molecular beam of 0.55A/8 in terms of SiO2 formation rate was irradiated from an electron gun type Si evaporator. An oxide film was formed on the clean surface. Oxygen partial pressure is 5 x 10 To
rr, and the formed film thickness was varied from 2.5 to 60. Furthermore, the oxygen leaked into the growth chamber was evacuated, the substrate temperature was raised again to 500°C, and the vacuum degree was reduced to I x 10.
Growth rate of 1.0A on this oxide film in Torr atmosphere
/S was deposited, and changes in the crystallinity of the upper Si layer depending on the thickness of the lower oxide film were investigated by RHEED.

After five SiO2 layers were grown on the (100) clean surface, changes in the RHEED pattern were observed when Si was grown again. The superlattice pattern indicating a clean surface is S
After io2 growth, it completely disappeared and became a single cutting pattern.

It was found that when the thickness of the upper Si layer became 70 mm, the RHEED pattern became a pattern peculiar to three-dimensional epitaxial growth, and the upper Si layer was grown epitaxially.

The growth pattern is island growth rather than layered growth in homoepitaxial growth. Furthermore, the upper layer S
It was found that when i is 500 Å or more, a 2×1 two-dimensional pattern appears again, and the shrimp islands coalesce to form a flat surface again. On the other hand, when growing 7.5 layers of SiO on a clean surface and then growing Si again, the upper layer Si
It was found that the RHEED pattern was a ring pattern, and the upper layer Si was polycrystalline. The SiO film thickness is 7
．． Above 5 Å, the upper Si layer became polycrystalline.

Further, the formation temperature of the upper layer Si must be 1006C or higher. At temperatures below 100°C, the upper layer Si does not become polycrystalline but becomes amorphous. Furthermore, the formation temperature of the upper layer Si is 86
When the temperature exceeds 0°C, the irradiated Si and the SiO2 of the substrate
reacts to become SiO and evaporates, so the formation temperature is 86
Must be below 0°C.

A similar phenomenon was also found in low-pressure thermal oxidation in which only O2 molecules were irradiated onto a heated Si substrate. 5i (100) clean surface, substrate temperature 500°C, oxygen partial pressure 5 x 10
It was found that when Si was grown again after 20 minutes of oxidation at Torr, the RHEED pattern became a pattern specific to three-dimensional epitaxial growth, and the upper layer Si grew as an epitaxial island. However, when the oxidation time exceeded 30 minutes, the RHEED pattern of the upper layer Si became a ring pattern, and the upper layer Si was polycrystalline. FIG. 3 shows the oxidation time dependence of the AES oxygen peak measured after low-pressure thermal oxidation but before forming the upper Si layer. As can be seen from this figure, the oxygen peak was saturated, and there was no significant change in the oxygen peak at the point of change from epitaxial growth to polysilicon formation. This indicates that the oxygen concentration on the surface is saturated, and it is considered that the upper Si layer becomes polycrystalline when the thickness of the oxide film formed by low-pressure thermal oxidation exceeds a certain value.

As mentioned above, both molecular beam growth and low pressure thermal oxidation
When 2 is thin, epitaxial information is transmitted onto 5iOZ. There are three possible explanations for this. The first is that there is a pinhole in the thin 8102 film, and epitaxial information is transmitted through this pinhole.The second is that Si and 5tO2 react during the initial growth of the upper layer S, resulting in a thin SiO□
The third case is when the film nucleates and the underlying Si is exposed and epitaxial information is transmitted.The third case is when the thin 5102 has some epitaxial information. However, Sio2
If the film thickness is 7.5 Å or more, epitaxial information cannot be transmitted, so at least above this film thickness, the lower Si layer and the upper Si layer are considered to be separated by SiO□, and above this film thickness, it functions effectively as a hole barrier. It is thought that then. SiO actually embedded in Si by Futatsu et al.
When the lattice image of the cross-sectional TEM was observed, 5102 was a precipitate with a diameter of about 12 mm when the SiO film thickness was 5 mm, and areas with no SiO□ were observed between these precipitates. Epitaxial information was transmitted to the upper mother layer from the opening of this SiO□. Further, in this case, the 5i02/Si interface was not flat and had an interfacial layer of 20 layers. On the other hand, in the sample in which 7.5 SiO layers were added and the upper Si became polysilicon, the amorphous S
An image peculiar to iO2 was observed between the lower single crystal Si and the upper polysilicon, and the upper Si and lower Si were completely separated by this SiO2. Furthermore, the SiO2/Si interface was extremely flat. Furthermore, the width of this 8102 image was calculated from the Si lattice spacing to be approximately 8, which approximately corresponded to the design value. The width of this 5102 image is 5 due to molecular beam growth.
It corresponded to the thickness of 102 images and could control from 8 people to 6OA.

Next, we investigated whether the thin 5102 film embedded in Si acts as an electrical hole barrier. The experiment consists of P (
After growing aoooA P-type and N-type epitaxial buffer layers on the 100) and N (100) substrates and cooling them to room temperature, 8102 was formed by co-evaporating Si and 0, and the substrate temperature was raised to 500°C again. Then, P-type and N-type doped Si layers were grown aoooA. P
B was used for type doping and sb was used for N type doping. B doping is done by heating HBO2 in a PBN crucible, and sb doping is done by heating sb in a PBN crucible.
This was done by heating the sample substrate in a crucible and applying a voltage of -400 V to the sample substrate to increase the adhesion coefficient of sb. The film thickness of 8102 was 20 to 60A. At this time, it was confirmed that the upper Si layer was entirely polycrystalline Si.

A contact was made on the upper Si layer with AI for the P type and AuSb for the N type, and using this electrode as a mask, the surrounding Si was etched and removed until the substrate was exposed. The diameter of the electrode was 0.6 mm. Back contact was made with In. Figures 4a) and 4b) compare the dependence of tunnel current on the SiO2 film thickness when holes and electrons are injected from the underlying single crystal Si. For both N-type and P-type, the amount of tunnel current decreases as the SiO□ film thickness increases.

In the N-type, the amount of sb doping was small and the resistance of the upper layer poly-Si was high, so when the 8102 film thickness was 20 people, the value was lower than the P-type, but when the thickness was 30 or more, the 8102 barrier was stronger than the series resistance. Therefore, the value of N-type tunneling current is one order of magnitude higher than that of P-type. The above experiments revealed that this oxide film functions as a hole barrier. However, as the oxide film thickness increases, electron injection also decreases, so the appropriate oxide film thickness is IOA~20. Experiments similar to those described above were conducted on an 8102 film obtained by low-pressure thermal oxidation in which only 02 molecules were irradiated onto a heated Si substrate, and it was confirmed that it functions as a hole barrier.

Finally, the results of actually fabricating a prototype bipolar transistor using this polysilicon emitter fabrication method will be explained. A conceptual diagram of the transistor prototype manufacturing process is shown in Figure 1. Collector 43 has a thickness of 4000 mm and a carrier concentration of 5X
A 10 cm phosphorus-doped, high-concentration buried layer 4, which will become an electrode, is grown by CVD epitaxial growth.
Formed on 2. After element separation using the Locos method, the base portion was opened (FIG. 1(a)). Next, SiMB
Boron-doped base layer 4 grown by E method at a temperature of 700°C with a thickness of 500 cm and a carrier concentration of 5 x 10 cm.
After removing the polysilicon on the field oxide film 4, a passivation oxide film 45 was formed by the CVD method, and the part that would become the emitter was opened (
Figure 1(b)). Furthermore, after cleaning the surface in SiMBE, a Si molecular beam of 0.2 A/S and 5X
810247 was formed by simultaneous irradiation with an oxygen molecular beam of 10 Torr, and after exhausting O2, the substrate temperature was raised to 500 °C, and a Si molecular beam was irradiated to form 3000 polysilicon 48, and the oxide film was The polysilicon on 45 was removed (FIG. 1(C)). Arsenic was implanted into the polysilicon emitter portion 48 in an amount of 5×10 cm by ion implantation at an acceleration of 100 KeV, and forced diffusion was performed by lamp annealing. The base electrode portion was opened and 5×10 cm of boron was implanted using a 15 KeV ion implantation method to form an ohmic contact with the base layer 46. When we evaluated the static characteristics of the prototype transistor as described above, the current amplification factor was 15 when no SiO□ was sandwiched, but when the thickness of the 8102 layer was 10 people, it was 150, which is about 10 times as large. An increase in current amplification factor was observed. Furthermore, when the thickness of the SiO2 layer was 5, the current amplification factor did not increase and it did not function as a hole barrier. When the thickness of the 8102 layer increased to 20 layers, the emitter resistance increased and the collector current decreased. onto a heated Si substrate. An experiment similar to the above was conducted on the 5102 film obtained by low-pressure thermal oxidation in which only 0□ molecules were irradiated, and it was confirmed that the current amplification factor increased. The above experiments revealed that the polysilicon emitter made by this method effectively functions as a hole barrier and increases the current amplification factor of the bipolar transistor.

The method of cleaning the base surface is not limited to the one described above, but also methods such as evaporation of the natural oxide film on the surface by heating alone, ion sputtering with Ar, Si
I tried etching with molecular beams, but the results remained the same.

Although silicon wafers were used as targets in this example, the method of the present invention is applicable to SO8 (Sil) where silicon exists only on the surface.
icon on 5apphire) substrate and more generally 5OI (Silicon On In5ulator)
Naturally, this also applies to substrates, etc.

(Effects of the Invention) As described in detail above, according to the present invention, a SiO2 film thinner than a natural oxide film whose film thickness is extremely accurately controlled is
Such a thin 5-layer film can be embedded between the base and polysilicon emitter of an N-type bipolar transistor.
Since the 102 film acts as a hole barrier, reverse injection of holes from the base can be suppressed and the current amplification factor can be increased.

[Brief explanation of the drawing]

Figures 1 (a) to (C) are conceptual diagrams of the transistor prototype process, Figures 2 (a) to (d) are diagrams for explaining the present invention in detail, and Figure 3 is a diagram showing the process of low-pressure thermal oxidation. Oxidation time and AES
FIGS. 4(a) and 4(b) are diagrams showing the relationship between oxygen peaks, and are diagrams showing the dependence of the tunnel current on the 8102 film pressure when holes and electrons are injected from the lower single crystal Si, respectively. In the figure, 11... Natural oxide film 1.12... Silicon single crystal base layer, 13... Oxide film for hole barrier, 14... Polysilicon emitter, 41... P-type silicon substrate, 42 ...High concentration buried layer, 43...
Collector layer, 44...Field oxide film, 45...
Passivation oxide film, 46... base layer, 47.
...Hole barrier oxide film, 48...Polysilicon emitter. Figure 1

Claims (2)

[Claims] (1) A first step of forming a silicon oxide film by simultaneously irradiating the surface of the base layer with silicon molecular beams and oxygen molecular beams after cleaning the substrate surface having the base layer; and a second step of forming polysilicon by irradiating the substrate with silicon molecular beams, and the first and second steps are performed consecutively in the same vacuum chamber or in vacuum chambers connected by a vacuum. A method for manufacturing a bipolar transistor, characterized in that the method is performed by: (2) After cleaning the surface of the substrate having the base layer, a first step of forming a silicon oxide film on the surface of the base layer by low-pressure thermal oxidation in which oxygen is irradiated while heating the substrate; and a second step of forming polysilicon by irradiating silicon molecular beams onto the substrate, and the first and second steps are performed in the same vacuum chamber or in vacuum chambers connected by a vacuum. A method for manufacturing a bipolar transistor characterized by consecutive steps.