JPH03181137A - Manufacturing method of semiconductor device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to a method for manufacturing a semiconductor device.

[Prior Art] Conventionally, a thin film transistor on an insulating substrate (hereinafter referred to as TPT) has an insulating substrate 41 made of glass or the like, as shown in FIG.
A semiconductor thin film 42 was formed thereon, and elements were built therein.

Furthermore, in recent years, crystalline semiconductor thin films are often used as semiconductor thin films in order to improve the characteristics of TPT. The term "crystalline semiconductor" as used herein refers to a single-crystal semiconductor, which has more defects than a normally used single-crystal wafer, or a polycrystalline semiconductor, which has one or more internal crystal grain boundaries.

[Problems to be Solved by the Invention] However, in the above conventional example, the crystalline semiconductor thin film 42
A large number of interface states 46,44 exist at the interface between the crystalline semiconductor thin film 42 and the gate insulating film 43, and at the interface between the crystalline semiconductor thin film 42 and the gate insulating film 43. For example, when a MOSFET is formed, the channel Carriers are trapped at a level in the semiconductor layer, forming a so-called back channel, resulting in deterioration of device characteristics such as fluctuations in threshold voltage and reduction in on/off ratio.

Furthermore, although polycrystalline silicon is often used as a crystalline semiconductor thin film, there are many interface states 45 at the grain boundaries that exist in polycrystalline silicon, and these trap carriers. This reduces carrier mobility in the channel portion.

Additionally, if an inexpensive material such as glass is used for the substrate, alkali ions such as Na''?z contained in the substrate material will migrate due to heat treatment during the process and exist as mobile ions at the interface with the substrate or in the silicon thin film. However, this has caused problems with deterioration of device characteristics and reliability.

To solve these problems, for example, hydrogen passivation using a silicon nitride film by plasma CVD is used as a protective film for the device after the device is formed, thereby reducing the levels in the silicon thin film and increasing the mobility. Ta. Further, in order to prevent alkali ion contamination, high-purity quartz, alkali-free glass, or the like may be used as the substrate.

However, even with the above method, the problem of the interface with the substrate has not been solved. This means that the hydrogen concentration distribution on the surface of a relatively crystalline semiconductor thin film after hydrogen passivation is
This is because, as shown in FIG. 5, the distribution has a peak on the surface of a relatively crystalline semiconductor thin film, which is not sufficient to passivate the levels inside the thin film and at the interface between the thin film and the underlying layer.

Further, substrates made of high-purity quartz, non-alkali glass, etc. are expensive, and there is a problem in forming TPT on a large area substrate at low cost.

[Means for Solving the Problems] A method for manufacturing a semiconductor device of the present invention is a method for manufacturing a semiconductor device comprising forming a crystalline semiconductor thin film on an insulating substrate. After forming the absolute IA ILM, the crystalline semiconductor thin film is formed at a temperature that is about the same as or lower than the temperature at which the hydrogen-containing insulating film was formed, and then the crystalline semiconductor thin film containing hydrogen is formed on the upper surface of the crystalline semiconductor thin film. Second
The method is characterized in that an insulating film is formed, and then heat treatment is performed at a temperature higher than the temperature at which the first insulating film and the second insulating film containing hydrogen are formed.

[Function] In the present invention, by heat-treating the insulating film containing hydrogen, hydrogen diffuses into the crystalline semiconductor thin film, and the interface states existing at the interface with the semiconductor thin film substrate are trapped by the hydrogen. As a result, the number of interface states can be reduced, the generation of back channels can be prevented, and the electrical characteristics of semiconductor devices such as TFTs can be improved.

In addition, hydrogen diffused into the antagonistic semiconductor Fit film is trapped in defect levels in the thin film and interface levels of crystal grain boundaries, lowering the barrier hydride and causing the electrical current in semiconductor devices such as TPT. Improve characteristics.

Note that by using a silicon nitride film as the insulating film containing hydrogen, an effect of blocking alkali ions such as Na3 from the substrate is produced, and an improvement in reliability is expected.

In addition, since a silicon nitride film is used as an insulating film containing hydrogen, an insulating film that acts as a barrier against hydrogen diffusion is formed on both the top and bottom surfaces of the crystalline semiconductor thin film, and the hydrogen diffused into the thin film is removed from the outside. Diffusion (out-diff)
The above-mentioned effects can be obtained more stably.

[Embodiment] FIG. 1 is a cross-sectional view of a semiconductor device that characterizes the present invention.

In the first embodiment of the present invention, first, a first insulating film containing hydrogen is formed on an insulating substrate 11 made of inexpensive glass or the like at a substrate temperature of 200° C. or more by, for example, plasma CVD.
Silicon nitride [12] is formed at 300°C. This silicon nitride film 12 contains several percent to several tens of percent of hydrogen.

Thereafter, a crystalline semiconductor thin film 13 is formed at a temperature comparable to or lower than the temperature at which the silicon nitride film 12 was formed. As a crystalline semiconductor thin film, low pressure CVD method, plasma CV
Polycrystalline silicon formed by method D or large-grain polycrystalline silicon obtained by the effect of adding hydrogen halide gas such as HCl2 to the film-forming atmosphere in the plasma CVD method proposed by the applicant is used. be able to.

(For example, Japanese Patent Application No. 1983-73630). The large-grain polycrystalline silicon thin film proposed by the present applicant (Japanese Patent Application No. 73629/1982) proposed by the present applicant is suitable for the present invention from the viewpoints of low process salinity and electrical properties.

Next, a second insulating film containing hydrogen is formed on the crystalline semiconductor thin film by the same method as the first insulating film containing hydrogen.

Next, in an atmosphere of N2, Ar2, N2, or a mixed gas thereof, the temperature (30
Heat treatment is performed at a temperature of 0°C to 600°C.

During this heat treatment, the hydrogen present in the silicon nitride film is
It is trapped by defect levels in the crystalline semiconductor thin film and interface states existing at the grain boundaries of the crystalline semiconductor thin film, suppresses the generation of back channels at the base interface, and suppresses the threshold voltage, and Reduce grain boundary potential and increase mobility.

In addition, the temperature of the heat treatment is 300°C, at which hydrogen diffusion begins to occur.
The temperature is higher than 600° C. and lower than 600° C. at which hydrogen diffused into the crystalline semiconductor thin film does not diffuse out again.

Further, by forming a silicon nitride film between the substrate and the crystalline semiconductor thin film, a blocking effect is provided against alkali ions such as Na + from the substrate, and reliability is improved.

In addition, by forming a silicon nitride film on the surface of the crystalline semiconductor thin film, out-diffusion of hydrogen diffused into the crystalline semiconductor thin film by heat treatment can be prevented.
can be prevented, and the above-mentioned effects can be obtained more stably.

As a second embodiment of the present invention, first, a first insulating film containing hydrogen is formed on a substrate 11 made of inexpensive glass or the like.
For example, in the plasma CVD method, the substrate temperature is 200°C to 30°C.
Silicon nitride wA12 is formed at 0°C. This silicon nitride film 12 contains several percent to several tens of percent of hydrogen.

Thereafter, an amorphous silicon thin film 13 is formed at a temperature comparable to or lower than the temperature at which the silicon nitride film 12 was formed.

As the amorphous silicon thin film, amorphous silicon formed by low pressure CVD or plasma CVD, or polycrystalline silicon made amorphous by ion implantation of S and + is used.

Next, a second insulating film containing hydrogen is formed on the crystalline semiconductor thin film by the same method as the first insulating film containing hydrogen.

Next, N2. In an atmosphere of Ar, N2, or a mixture thereof, the temperature is higher than the temperature at which an insulating film containing hydrogen, such as a silicon nitride film, is formed (300°C to SOO).
Heat treatment is performed at ℃).

Regarding the temperature of this heat treatment, it is desirable to set the temperature at which the formed amorphous silicon undergoes solid phase crystal growth and becomes polycrystalline, in order to produce a higher performance TPT. Therefore, the temperature of the heat treatment is more preferably 500°C.
It is desirable to set the temperature to ~600°C in order to produce a higher performance TPT.

During this heat treatment, the hydrogen present in the silicon nitride film is
Due to heat treatment, amorphous silicon diffuses into polycrystalline silicon, resulting in interface states existing at the underlying interface, defect levels in polycrystalline silicon thin films, and interfaces existing at grain boundaries of polycrystalline silicon. It is trapped in the level, suppresses the generation of back channels at the base interface, suppresses fluctuations in threshold voltage, reduces the potential of grain boundaries, and increases mobility.

In addition, the temperature of the heat treatment is 30°C, at which hydrogen diffusion begins to occur.
It is preferable to carry out the process at a temperature higher than 0° C. and lower than 600° C. at which it will not diffuse into polycrystalline silicon.

Furthermore, by forming a silicon nitride film between the substrate and the polycrystalline silicon thin film, a blocking effect is imparted to alkali ions such as Na'' from the substrate, thereby improving reliability.

Furthermore, by forming a silicon nitride film on the surface of the crystalline semiconductor thin film, out-diffusion of hydrogen diffused onto the surface of the crystalline semiconductor thin film by heat treatment can be prevented, and the above-mentioned effects can be obtained more stably.

[Example] Embodiments of the present invention will be described in detail using the drawings.

FIG. 2 is a cross-sectional view of a MOSFET made using the present invention.

(Example 1) As a first insulating film containing hydrogen on a glass substrate 21,
With plasma CVD method, SiH4/NH! 1000 silicon nitride films 22 were deposited using a mixed gas system. As for the deposition conditions, a parallel plate plasma CVD apparatus was used,
SiH4 (10% H2 dilution) flow rate 15 S CCm,
N H, flow rate 10105e, pressure 0.16 Torr,
Under the conditions of discharge power 3.5W and substrate temperature 300℃, 35
Deposition was carried out for minutes. As a result of IR (infrared spectroscopy) analysis, it was found that the silicon nitride film deposited under these conditions contained about 10% hydrogen.

Next, 1000 polycrystalline silicon thin films 23 were deposited on the silicon nitride film 22 using a SiH2Cf12/HC/H2 mixed gas system by RF plasma CVD. The deposition conditions are SiH2C1
120.9sec, HC11130sec, pressure 2
．． 0Torr, RF power 60W, substrate temperature 2
It was carried out at 30°C. Under this condition, the silicon nitride film 22
A polycrystalline silicon thin film with a grain size of about 1.0 μm was deposited on top.

Next, by sputtering, 5i02
After 500 layers of film 24 were deposited, gate electrode 25 was formed.

Next, P4 was implanted by ion implantation to form source/drain regions 26.

Next, a second insulating film containing hydrogen is formed by plasma CVD.
A silicon nitride film 27 was deposited by 5,000 people under the same conditions as when the first silicon nitride film was formed.

Next, heat treatment was performed at 550° C. in an N2 atmosphere.

As a result of this heat treatment, the concentration distribution of hydrogen in the polycrystalline silicon thin film 23 was confirmed to be almost uniformly introduced into the polycrystalline silicon thin film as shown in FIG. 3, as confirmed by SIMS depth direction analysis. Ta.

Next, a contact hole is opened in a desired area, and AJ2
An electrode 28 was formed.

In this example, a comparison of electrical characteristics measurements of a MOSFET formed on a glass substrate with a polycrystalline silicon thin film formed directly on the glass substrate and a MOSFET created according to this example revealed that the electron mobility was more than 1.5 times higher than the threshold. The fluctuation width of the value voltage was reduced to less than 1/2.

In addition, in this example, the heat treatment temperature was 550°C.
was used, but the same effect was obtained even at 300°C.

In addition, in the case where no heat treatment was applied at all, almost no effect was observed on both the electron mobility and the threshold voltage variation.

This can be confirmed by heat treatment.
Hydrogen diffuses into the polycrystalline silicon thin film wA 23 from 7 and is trapped in interface states existing at the interface with the gate insulating film, the interface with the base insulating film, and the grain boundaries in the polycrystalline silicon thin film 23, This is considered to be because the number of levels decreased, the generation of packed channels at the base interface was suppressed, and the potential barrier at the grain boundary was lowered. This means that E
It is clear from the results of SR (electron spin resonance) measurements that the density of dangling bonds in the polycrystalline silicon thin film was reduced by more than one digit due to the heat treatment.

In addition, silicon nitride I as a second insulating film containing hydrogen
I! There was a difference of about 1.5 times in electron mobility depending on the presence or absence of 27, and this effect was confirmed.

This means that in the absence of the silicon nitride film 27 as the second insulating film containing hydrogen, the hydrogen concentration distribution in the polycrystalline silicon thin film is mainly distributed near the base interface.
This was confirmed by MS measurement, and it is thought that this is due to the lack of an absolute amount of hydrogen for passivating.

In addition, in the reliability test, there was almost no change in the electrical characteristics even in the high-temperature test, and the reliability was sufficient.

This is considered to be because the silicon nitride 1lj22 blocks the diffusion of alkali ions from the glass substrate.

In addition, by heat treatment at 550° C., hydrogen can be diffused into the polycrystalline silicon and the source/drain regions can be activated at the same time, which further simplifies the process.

(Example 2) On the glass substrate 21, as a first insulating film containing hydrogen,
A silicon nitride film 22 was deposited by 1000 people using a plasma CVD method using an S I H4/N)is mixed gas system. As for the deposition conditions, a parallel plate plasma CVD device was used, SiH4 (10% dilution) flow rate was 15 sec, and N
H, flow rate 10sec m, pressure 0.16Torr,
Under the conditions of discharge power 3.5W and substrate temperature 300℃, 35
Deposition was carried out for minutes. As a result of IR (infrared spectroscopy) analysis, it was found that the silicon nitride film deposited under these conditions contained about 10% hydrogen.

Next, 1000 amorphous silicon thin films 23 were deposited on the silicon nitride IN122 using a s i H4/H2 mixed gas system by plasma CVD.

The deposition conditions were a SiH4 flow rate of 2 seconds.

H2 flow rate 18sec, pressure 0.12Torr.

Deposition was carried out at a discharge power of 5 W for 30 minutes.

Next, a gate insulating film of 5i0
After 500 deposits of the two films 24, a gate electrode 25 was formed.

Next, Po was implanted by ion implantation to form source/drain regions 26.

Next, a second insulating film containing hydrogen is formed by plasma CVD.
5,000 silicon nitride films 27 were deposited using the same method used to form the second insulating film.

Next, heat treatment was performed at 600° C. in a N2 atmosphere.

SIMS depth direction analysis confirmed that by this heat treatment, the concentration distribution of hydrogen in the polycrystalline silicon thin film 23 was almost uniformly introduced into the polycrystalline silicon thin film, as shown in FIG. Ta.

Next, a contact hole is opened in a desired area, and the Au electrode 28
was formed.

In this example, it was confirmed by cross-sectional TEM (transmission electron microscopy 112ilt) that the amorphous silicon thin film 23 underwent solid phase crystal growth and became polycrystalline due to the heat treatment at 600°C.

In this example, an amorphous silicon thin film was directly formed on a glass substrate, and a MOSFET was formed on a polycrystalline substrate by heat treatment, and a MOS fabricated according to this example.
A comparison of measurements of the electrical characteristics of FETs revealed that the electron mobility was reduced by more than 1.5 times, and the fluctuation range of the threshold voltage was reduced to less than 1/2.

This can be confirmed by heat treatment.
7, hydrogen diffuses into the polycrystalline silicon thin film 23 in which amorphous silicon has become polycrystalline, and is trapped in the interface states existing at the base interface and at the grain boundaries in the polycrystalline silicon thin film 23. It is thought that this is because the number of particles decreased, the generation of back channels at the base interface was suppressed, and the potential barrier at the grain boundaries was lowered. This means that ESR
(electron spin resonance) measurement results show that the density of dangling bonds in a polycrystalline silicon thin film increases as a result of heat treatment.
This is clear from the fact that it has decreased by more than one digit.

Further, in this embodiment, the amorphous silicon film 23 was made polycrystalline by heat treatment, but even if the heat treatment temperature is lowered, for example, 300° C., and the amorphous silicon thin film is used as it is, the amorphous silicon film 23 becomes polycrystalline. Compared to a silicon thin film formed directly on a glass substrate, an improvement in electrical properties was observed.

Furthermore, the electron mobility differed by about 1.5 times depending on the presence or absence of the silicon nitride film 27 as the second insulating film containing hydrogen, and this effect was confirmed.

This is because it has been confirmed from SIMS measurements that the hydrogen concentration distribution in the silicon nitride thin film as a second barrier containing hydrogen is mainly distributed near the base interface. This is thought to be due to the absolute amount of hydrogen being insufficient.

Further, in the reliability test, there was almost no change in the electrical characteristics even in the high-temperature test, and the reliability was sufficient.

This is considered to be because the silicon nitride film 22 blocks the diffusion of alkali ions from the glass substrate.

Further, by heat treatment at 600° C., it was possible to activate the source/drain regions at the same time that hydrogen diffused into the polycrystalline silicon, thereby further simplifying the process.

(Example 3) On the glass substrate 21, as a first insulating film containing hydrogen,
SiH4/NH by plasma CVD method.

A silicon nitride film 22 having a thickness of 1000λ was deposited using a mixed gas system in the same manner as in the first and second embodiments. As a result of IR (infrared spectroscopy) analysis, this silicon nitride film contains
It was found that it contained about 10% hydrogen.

Next, by RF plasma method, 5iH2Cn2/HC
A polycrystalline silicon thin film 23 was deposited by 1000 people on the silicon nitride film 22 using an It/H2 mixed gas system.

Next, as the gate insulating film 24, 200 5 oztli films were deposited in succession by sputtering, and as a second insulating film containing hydrogen, 300 silicon nitride films were deposited by plasma CVD. Formed. As is well known, the reason why Sin,ll is deposited first is to prevent the electrical characteristics of the MOSFET from deteriorating due to polarization in the film if the gate insulating film is composed of only a silicon nitride film. .

Next, P+ was implanted by ion implantation to form source/drain regions 26.

Next, heat treatment was performed in a N atmosphere at 550°C.

Next, as a protective film, a silicon oxide film was deposited to a thickness of 5000 by plasma CVD.

Next, a contact hole is opened in a desired area, and the Aj2 electrode 2
8 was formed.

In this example, even if the thickness of the silicon nitride film as the second insulating film containing hydrogen is 300, the density of hydrogen in the polycrystalline silicon thin film is completely different from when the silicon nitride film is 5000. There wasn't.

Furthermore, although SiO2@5,000 layers were used as a protective film, there was no change in the density of hydrogen, indicating that even a silicon nitride film of 300 layers was sufficient to act as a barrier against hydrogen diffusion.

In addition, a two-layer structure of a silicon nitride film and a silicon oxide film was used as the gate insulating film;
Compared to the case where a membrane was used, almost no deterioration in electrical properties was observed.

In addition, by heat treatment at 550° C., hydrogen can be diffused into the polycrystalline silicon and the source/drain regions can be activated at the same time, which further simplifies the process.

(Example 4) On a glass substrate 21, a first insulating film containing hydrogen and a sheet,
A silicon nitride film 22 was deposited by 1000 people using a plasma CVD method using a S t O2/N H3 mixed gas system.

Next, 500 silicon nitride oxide films were deposited as a gate insulating film 24 and a second insulating film containing hydrogen by plasma CVD, and then a gate electrode 8i25 was formed.

As is well known, a silicon nitride oxide film can have the properties of both a silicon nitride film and a silicon oxide film by appropriately selecting the ratio of nitrogen and oxygen in the film. Here, S i H4/NH3/N20
By optimizing the deposition conditions using a mixed gas system, the composition ratio of the film was set to N~3 and O~2 to Si.

Next, Po was implanted by ion implantation to form source/drain regions 26.

Next, heat treatment was performed at 550° C. in a N2 atmosphere.

Next, 5,000 silicon oxide films were deposited as a protective film by plasma CVD.

Next, a contact hole is opened in the desired area, and the AJ2 electric 8i
28 was formed.

In this example, a silicon nitride oxide film was used as the second insulating film containing hydrogen, but the density of hydrogen in the polycrystalline silicon thin film was almost the same as when a silicon nitride thin film was used.

Furthermore, although a silicon nitride oxide film was used as the gate insulating film, almost no deterioration in electrical characteristics was observed compared to the case where a 5in2 film was used.

In addition, by heat treatment at 550° C., hydrogen can be diffused into the polycrystalline silicon and the source/drain regions can be activated at the same time, which further simplifies the process.

Claims (5)

[Claims] (1) In a method for manufacturing a semiconductor device comprising forming a crystalline semiconductor thin film on an insulating substrate, a first insulating film containing hydrogen is formed on the substrate, and then the insulating film containing hydrogen is formed. The crystalline semiconductor thin film is formed at a temperature comparable to or lower than the above temperature, and then a second insulating film containing hydrogen is formed on the upper surface of the crystalline semiconductor thin film, and then a second insulating film containing hydrogen is formed on the top surface of the crystalline semiconductor thin film. 1. A method of manufacturing a semiconductor device, comprising performing heat treatment at a temperature higher than the temperature at which the insulating film and the second insulating film were formed. (2) The method of manufacturing a semiconductor device according to claim 1, wherein a silicon nitride film formed by plasma CVD is used as the insulating film containing hydrogen. (3) The method of manufacturing a semiconductor device according to claim 1, wherein the temperature of the heat treatment is 300°C to 600°C. (4) The method for manufacturing a semiconductor device according to claim 1, wherein the crystalline semiconductor thin film material is silicon. (5) The method of manufacturing a semiconductor device according to claim 1, 3 or 4, wherein the temperature of the heat treatment is a temperature at which amorphous silicon becomes polycrystalline.