JPH03185736A - Manufacture of semiconductor device - Google Patents

Abstract

PURPOSE:To reduce the interface level at a substratum interface of a semiconductor thin film and to enhance the electrical characteristic of a TFT by a method wherein hydrogen is introduced into a crystalline semiconductor thin film, light is irradiated and only a local part near a grain boundary or near an interface of the crystalline semiconductor thin film is heated. CONSTITUTION:A first insulating film and a second insulating film 22, 27 which act as barriers against the diffusion of hydrogen are formed on both the upper side and the lower side of a crystalline semiconductor thin film 23; hydrogen is introduced into the crystalline semiconductor thin film 23. In addition, light is irradiated; only a local part near a grain boundary or near an interface of the crystalline semiconductor thin film 23 is heated. In this case, the hydrogen is introduced into the crystalline semiconductor thin film 23 and the light is irradiated; as a result, the crystal grain boundary or a defect inside the crystalline semiconductor thin film 23 or the interface between the crystalline semiconductor thin film 23 and a substrate 21 is heated selectively, a trap level existing in these places is terminated effectively at the hydrogen, the number of trap levels can be reduced. Thereby, it is possible to enhance the electrical characteristic of a semiconductor device such as a TFT or the like.

Description

[Detailed description of the invention] [Field of use in business a]' 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 3 made of glass or the like, as shown in FIG.
1, a semiconductor thin film 32 is formed on top of the semiconductor thin film 32, and elements are 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.

[X that the invention attempts to solve! Title]

However, in the conventional example described above, as shown in FIG. For example, when a MOSFET is formed in the crystalline semiconductor thin film 32 as shown in the figure, carriers in the channel part are moved to the interface state 34 due to the influence of this interface state.
35, forming a so-called back channel, resulting in deterioration of device characteristics such as fluctuations in threshold voltage and reduction in signal ON10ff ratio.

Further, polycrystalline silicon thin films are often used as crystalline semiconductor thin films, but there are many interface states 37 in the crystal grain boundaries 36 that exist within the polycrystalline silicon thin film, and these are carriers. By trapping
This reduces carrier mobility in the channel portion.

Furthermore, if an inexpensive material such as glass with a high alkali ion content is used for the substrate, alkali ions such as Na'' contained in C in the substrate material will migrate toward the semiconductor thin film due to heat treatment during the manufacturing process. They exist in the form of mobile ions at interfaces with semiconductors and in thin films of semiconductors such as silicon, causing deterioration of device characteristics and reliability problems.

As a method to solve these problems, for example, after the device is formed, a silicon nitride film is formed as a protective film of the device by plasma CVD method, and hydrogen passivation by this protective film is used to improve localized levels within the silicon thin film. reduce,
Efforts have been made to increase mobility. Additionally, in order to prevent alkali ion contamination, a method of using high-purity quartz, alkali-free glass, or the like as a substrate has been used.

However, even with the above method, the passivation of the grain boundaries and the passivation of the interface with the substrate are not perfect. That is, in the above method, a process in which hydrogen is introduced into the semiconductor thin film and a process in which hydrogen is released from the semiconductor thin film coexist. Hydrogen diffuses even in areas with good properties, and there remains much room for further improvement in order to effectively passivate grain boundaries and interfaces.

Further, substrates made of high-purity quartz or non-alkali glass are expensive, and one of the inherent advantages of TPT, which is employed to form inexpensive elements on large-area substrates, is lost.

[Means for Solving nB] The method for manufacturing a semiconductor device of the present invention includes forming a crystalline semiconductor thin film on an insulating substrate, in which: on both upper and lower sides of the crystalline semiconductor thin film, A step of forming first and second ams that act as a barrier against hydrogen diffusion, a step of introducing hydrogen into the crystalline semiconductor thin film, and a step of irradiating light to form near the grain boundaries of the crystalline semiconductor thin film. or a step of heating only a local part near the interface;
It is characterized by including.

[Function] In the present invention, by introducing hydrogen into a crystalline semiconductor thin film and irradiating it with light, crystal grain boundaries and defects in the crystalline semiconductor thin film, or the interface between the crystalline semiconductor thin film and the substrate, etc.
By selectively heating, the trap levels existing in these locations can be effectively terminated with hydrogen and the number of trap levels can be reduced, so it is expected that the electrical characteristics of semiconductor devices such as TPTs will be improved.

In addition, it is possible to eliminate crystal grain boundaries, defects, etc. without changing the spatial position of impurities introduced into the source and drain of TPT.
Alternatively, since the interface between the semiconductor thin film and the substrate can be selectively heated, manufacturing of fine transistors becomes easier.

Furthermore, since only the crystalline semiconductor thin film can be heated, it is possible to prevent alkali ions such as Na''' from entering from a substrate such as glass, and an improvement in reliability can be expected.

Furthermore, since an insulating film is formed on the semiconductor thin film to act as a barrier against hydrogen diffusion, it is expected that out-diffusion of hydrogen diffused into the thin film will be prevented and the above-mentioned effects will be obtained more stably. .

Note that when a silicon nitride film is formed between the substrate and the crystalline semiconductor thin film as an insulating film that acts as a barrier against hydrogen diffusion, it also has the effect of blocking alkali ions such as Na'' from the substrate such as glass. As a result, further improvements in reliability can be expected.

[Embodiment Examples] Preferred embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 is a sectional view of an example of a semiconductor device manufactured by the method of the present invention.

In the first embodiment of the present invention, first, a first insulating film that serves as a barrier against hydrogen diffusion is formed on an insulating substrate 21 made of glass or the like by, for example, plasma CVD, at a substrate temperature of 200 t. Silicon nitride 1i2 at 300℃
In this example, several percent to several tens of percent of hydrogen is contained in the silicon nitride film 22.

Thereafter, a crystalline silicon thin film 23, for example, is formed as a crystalline semiconductor thin film at a temperature comparable to or lower than the temperature at which silicon nitride 1II22 was formed. For crystalline silicon thin@23, sputtering method, low pressure CVD method,
Polycrystalline silicon formed by plasma CVD method,
In the sputtering method, CVD method (in which amorphous silicon formed by annealing and recrystallization, and the plasma CVD method separately proposed by the applicant), hydrogen halides such as HCf are added to the film-forming atmosphere. Large-grain polycrystalline silicon obtained by the effect of gas addition, and large-grain polycrystalline silicon proposed by the present applicant in Japanese Patent Application No. 62-73629 and Japanese Patent Application No. 62-73630. Alternatively, single crystal silicon formed on an amorphous substrate, as proposed by the present applicant in Japanese Unexamined Patent Publication No. 107016/1983, can be used.

Next, a second insulating film 24 that serves as a barrier against hydrogen diffusion is formed on the crystalline silicon 23. The second insulation 1]i24 that acts as a barrier against hydrogen diffusion is as follows:
A silicon nitride film formed by a low pressure CVD method, or a silicon nitride film or a silicon nitride oxide film formed by a plasma CVD method in the same manner as the first insulating film can be used.

Next, as irradiation light, for example, near-infrared light is irradiated near the interface and crystal grain boundaries of the crystalline silicon thin film 1i23. Near-infrared light is e.g. a xenon lamp (wavelength 0.8 μm)
etc. are suitable. When irradiated with light having a wavelength of about 0.8 μm, electron-hole pairs are generated in the silicon thin film 23.

Electrons and holes generated by light irradiation travel within the conductive band and valence band, respectively, within the silicon thin film crystal with a relatively long lifetime. Among these, those that reach grain boundaries, defects, or interfaces cause recombination through trap levels caused by dagging bonds existing there, and the energy of electrons and holes lost due to recombination. is converted into lattice vibration, i.e. thermal energy. In other words, grain boundaries, defects,
Alternatively, the interface will be selectively heated by electron/hole recombination. This thermal energy will cause trap levels in grain boundaries, defects, or interfaces that already exist in the film to be heated. It can be terminated with hydrogen that is present, and crystal defects can be recovered.

Here, light with a wavelength of 0.8 μm corresponds to 1.5 keV when converted into energy, the absorption coefficient is about I x 10” am-' in the case of crystalline silicon, and the penetration depth is 1.5 keV from the surface. /e intensity is approximately 10 μm.

Light with a wavelength sufficiently shorter than 0.8 μm (for example, visible light, infrared light, etc.) has high energy, and electrons generated by light irradiation exist in one high-energy direction of the conductive band. As a result, the entire thin film is heated due to collisions of electrons in the conductive band, causing heating of the substrate, release of hydrogen, and diffusion of doped impurities.

Conversely, light with a wavelength sufficiently longer than 0.8 μm (for example,
For light with a wavelength of 1 to several μm), if the film is thick, the light penetrates quite deep from the surface, and this generates electron-hole pairs not only at the surface but also at a considerable depth, causing recombination. Therefore, it may not be possible to efficiently heat only the area near the surface of the thin film, which is necessary for device operation.

From the above, light irradiation with near-infrared light having a wavelength larger than 0.7 μm and smaller than 1.0 μm, preferably about 0.8 μm, is optimal in the present invention.

In addition, it is preferable that the heat generated at this time be higher than 300°C, where hydrogen in the thin film begins to diffuse, and lower than 600°C, at which hydrogen in the thin film does not diffuse out again. This can be achieved by adjusting the amount of irradiation light.

The atmospheric gas used for this light irradiation is N, Ar.
, H, or a mixture thereof.

During this light irradiation, the hydrogen present in the silicon nitride film is
By diffusing into the crystalline silicon thin film, the interface states exist at the base interface that is selectively heated, the defect levels in the crystalline silicon thin film, and the interface states that exist at the grain boundaries of the crystalline silicon. terminating, suppressing the occurrence of back channels at the base interface, and reducing the potential of grain boundaries.
Increase mobility.

In addition, by forming a silicon nitride film between the substrate and the crystalline silicon thin film, it is possible to have a blocking effect against alkali ions such as Na'' from the glass substrate, improving reliability. Improve further.

In addition, by forming an insulating film that acts as a barrier against hydrogen diffusion on both sides of the crystalline silicon thin film, it is possible to prevent out-diffusion from the surface of the crystalline silicon thin film when hydrogen diffuses through heat treatment. The passivation effect can be further enhanced.

(Second Embodiment) In the first embodiment, silicon nitride wA22 contains several percent
In this example, hydrogen gas was introduced into the chamber using a plasma CVD device, and then a discharge was generated to cause the hydrogen plasma to contain hydrogen in the crystalline silicon thin film. Introducing hydrogen into

In this example, light irradiation is also performed, and during this light irradiation, hydrogen introduced from the plasma diffuses into the crystalline silicon thin film, creating an interface state that exists at the selectively heated base interface. It terminates defect levels in crystalline silicon thin films and interface levels existing at grain boundaries of crystalline silicon, suppresses the generation of back channels at the underlying interface, and reduces the potential of grain boundaries. Increase mobility. In addition, the heat generated during this light irradiation begins to cause hydrogen diffusion.
It is necessary to carry out the process at a temperature higher than 00°C and lower than 600°C, at which hydrogen in the film will not diffuse out again.
What can be achieved by adjusting the amount of irradiation light is the same as in the first embodiment.

(Third Embodiment Example) In this example, instead of including several percent to several tens of percent hydrogen in the silicon nitride film 22, hydrogen is added into the crystalline silicon thin film by a normal ion implantation method. to be introduced.

In this example, light irradiation is also performed, and during this light irradiation, hydrogen implanted by ion implantation diffuses into the crystalline silicon thin film, resulting in interface units existing at the selectively heated base interface. terminating defect levels in crystalline silicon thin films and interface levels existing at grain boundaries of crystalline silicon,
It suppresses the generation of back channels at the base interface, reduces the potential of grain boundaries, and increases mobility.

*fS7e> The trajectory of the reaction must be performed at a temperature higher than 300°C, where hydrogen diffusion begins, and lower than 600°C, at which hydrogen in the film does not diffuse out again. Similar to the first embodiment, this can be achieved by adjusting the amount of irradiated light.

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

FIG. 2 is a schematic cross-sectional view showing the manufacturing process of a MOSFET according to the present invention.

(Example 1) SiH was deposited on the glass substrate 2'1 by plasma CVD as a first insulating film that acts as a barrier against hydrogen diffusion.
,/NH3 mixed gas system, the silicon nitride film 22 is
000 people deposited. The deposition conditions were 5in4 (10% H2 dilution) using a parallel plate plasma CVD device.
Flow rate 15scan, NH2 flow rate 10105c, pressure 0.
16Torr, discharge power 3.5W, substrate temperature 300℃
Deposition was carried out for 35 minutes under these conditions. In the silicon nitride film deposited under these conditions, as a result of IR (infrared spectroscopy) analysis,
It was found that about 10% hydrogen was contained (Figure 2 (a)).

Next, SiH was obtained using the FtF Blaze vCVD method.

A polycrystalline silicon thin film 23 is deposited 10 times on the silicon nitride film 22 in a Cftz/HCJl/H2 mixed gas system.
00 people deposited. The deposition conditions are 5i)(, (u2
:Q, 9sccm, HCl2:130sec, H,:
200sccm, pressure crab 2, 0Torr%RFpowe
r: 60W, substrate temperature = 230°C. Under this condition, the grain size on the silicon nitride film 22 is approximately 1°0.
A μm thick polycrystalline silicon thin film was deposited (Figure 2(b))
.

Next, by sputtering, 5i02
After 500 layers of film 24 were deposited, An was deposited by sputtering.
was deposited and patterned to form a gate electrode 25 (FIG. 2(C)).

Next, P9 was implanted by ion implantation to form source/drain regions 26 (FIG. 2(d)).

Next, 5,000 silicon nitride films 27 were deposited by plasma CVD in the same manner as the first insulating film 22 as a second insulating film to serve as a barrier against hydrogen diffusion.

Next, a xenon lamp with a light wavelength of 0.8 μm and a power of 1000 W was irradiated from a distance of about 10 cm in an H gas atmosphere. At this time, the effective power of the xenon lamp is considered to be about 40 W/cm' (Figure 2 (e))
.

Next, contacts were opened in desired regions, and AJ2 was deposited and patterned to form source and drain electrodes 28 and gate electrodes 29.

A comparison of the electrical characteristics of a MOS FET formed on a glass substrate with a polycrystalline silicon thin film formed directly on the substrate and a MOS FET created according to this example shows that the electron mobility is twice as high in this example. As a result, the fluctuation width of the threshold voltage was reduced to 1/2 or less.

This means that due to the heat treatment, hydrogen diffuses from the silicon nitride film 22 into the polycrystalline silicon thin film 23, and the underlying interface and polycrystalline silicon thin film II! This is thought to be because the interface levels existing at the grain boundaries in No. 23 were terminated, the number of levels decreased, the generation of back channels at the underlying interface was suppressed, and the potential barrier at the grain boundaries was lowered. As a result of ESR (electron spin resonance) measurement, this shows that the density of dangling bonds in polycrystalline silicon fil@
This is clear from the fact that the heat treatment caused a decrease of more than one digit.

Furthermore, it was confirmed that impurity diffusion in the source/drain region 26 did not occur because the vicinity of the interface state was selectively heated at this time.

Further, regarding the effect of silicon nitride 1i22, 27 which acts as a barrier against hydrogen diffusion, for example, depending on the presence or absence of silicon nitride 1i27, the polycrystalline silicon thin film 23
Since the density of hydrogen present in the film has decreased from the order of I x 10"am-' to below the order of 1 x 10" cm-3, this film
It was found that it acts as a barrier against USion.

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.

Furthermore, in this example, it was revealed from the measurement of electrical characteristics that it was possible to diffuse hydrogen and activate the source/drain regions by light irradiation at the same time.

(Example 2) Gate insulation @ 24 5i025 used in the first example
As a second example, instead of 00 Å, a Si 02 film was formed by sputtering, and then a plasma CVD film was formed as an insulating film to act as a barrier against hydrogen diffusion.
After depositing 300 silicon nitride films by the method, the gate electrode 25 was formed. As is well known, what was deposited first in s i 02 @ was only a silicon nitride film, which resulted in a gate failure of +1! This is to prevent deterioration of the electrical characteristics of the MOSFET due to polarization in the film when the film is constructed.

Other steps are the same as in the first example.

In this example, even if the thickness of the silicon nitride film is 300 mm thick as the second insulating film that acts as a barrier against hydrogen diffusion, the density of hydrogen in the polycrystalline silicon thin film is 5000 mm thick. There was no change at all from when I did it.

In addition, there was no change in the density of hydrogen even when using a film of 5,000 layers of Sin as a protective film.
It was found that even silicon nitride 11 [300λ] sufficiently acts as a barrier.

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 0211i was used, almost no deterioration in electrical characteristics was observed.

In addition, in this example, plasma CVD film was used as the barrier film.
Although a silicon nitride film deposited by a low-pressure CVD method was used, similar effects could be obtained using a silicon nitride film deposited by a low-pressure CVD method.

(Example 3) 5i025 of the gate insulating film 24 described in the first example
In the third example, instead of 0.00 people, a silicon nitride oxide film with a thickness of 500 λ was deposited by plasma CVD method C as a gate insulating film and an insulating film that acts as a barrier against hydrogen diffusion. As is well known, silicon nitride oxide films can be produced by carefully selecting the composition ratio of nitrogen and oxygen in the film.
It is possible to have the properties of both a silicon film and a silicon nitride oxide film. Here, SiH4/NH3/
By reoptimizing the deposition conditions using an N20 mixed gas system, the composition ratio of the film was reduced to 1 for Si and 3.0 to 2 for N.
I made it so that

Other steps are the same as in the first embodiment.

In this example, even if a silicon nitride oxide film is used as the second insulating film that acts as a barrier against hydrogen diffusion,
The density of hydrogen in the polycrystalline silicon thin film did not change at all compared to when a silicon nitride film was used.

Further, although a silicon nitride oxide film was used as the gate insulating film, almost no deterioration in electrical characteristics was observed in this film as compared to the case where a 5i02 film was used.

(Example 4) Gate insulation II described in the first example! 24 Sins
.

The hydrogen plasma conditions were a pressure of 0.16 Torr, a discharge output of 600 W, a substrate temperature of 300° C., and an irradiation time of 30 minutes.

Other steps are the same as in the first example.

In this example as well, the same effect as in the first example was obtained in terms of improving the electrical characteristics of the MOSFET.

(Example 5) As described in the first example, Gate Zettai! 1WA24's 5i0
After depositing 2,500 layers and forming the gate electrode 25, hydrogen was introduced by ion implantation at a dose of IE16 cm as a fifth example.
-'' and an acceleration voltage of 2°keV, the implantation was carried out over the entire surface of the polycrystalline silicon thin film.

Other steps are the same as in the first example.

In this example, the same effect as in the first example was obtained in terms of improving the electrical characteristics of the MOSFET.

As described above, in this example, the crystalline semiconductor thin film is
This effect was demonstrated for large-grain polycrystalline silicon formed by the plasma CVD method proposed by the present applicant, and for polycrystalline silicon obtained by crystallizing amorphous silicon formed by the plasma CVD method by heat treatment. Other crystalline semiconductor thin films, such as polycrystalline silicon formed by low-pressure CVD, amorphous silicon made amorphous by implanting Si1 into polycrystalline silicon, and then annealed and recrystallized; Large-grain polycrystalline silicon, which the present applicant proposed in Japanese Patent Application No. 62-73629 and Japanese Patent Application No. 62-73630;
Needless to say, the same effect could be obtained by using single crystal silicon formed on an amorphous substrate as proposed in Publication No. 6.

[Effects of the Invention] According to the present invention, it is possible to reduce the interface state at the semiconductor thin film base interface and suppress the back channel effect, and
It is also possible to reduce the levels existing within the semiconductor thin film, and as a result,
Change in threshold voltage! It was possible to improve the electrical characteristics of TPT, such as reducing /r[ and improving carrier mobility.

Note that when using a silicon nitride film as an insulating film that acts as a barrier against hydrogen diffusion, Na"
It was possible to block the intrusion of alkali ions such as, and improve the reliability of TPT.

As a result, it has become possible to easily form a TPT with excellent electrical properties and reliability on an inexpensive glass substrate.

[Brief explanation of drawings]

FIG. 1 is a sectional view of a semiconductor device that characterizes the present invention. FIG. 2 is a schematic cross-sectional view showing the manufacturing process of a MOSFET according to the present invention. FIG. 3 is a schematic diagram for explaining the problems of the prior art. (Explanation of symbols) 21...Substrate, 22...First insulating film serving as a barrier against hydrogen diffusion, 23.32...Semiconductor thin film,
27...Second insulating film that acts as a barrier against hydrogen diffusion. Fig. l Now μ Fig. (d) (e) (f) Fig.

Claims (8)

[Claims] (1) In a method for manufacturing a semiconductor device in which a crystalline semiconductor thin film is formed on an insulating substrate, a first layer, which serves as a barrier to hydrogen diffusion, is provided on both upper and lower sides of the crystalline semiconductor thin film.
, a step of forming a second insulating film, a step of introducing hydrogen into the crystalline semiconductor thin film, and a step of irradiating light to heat only a local portion of the crystalline semiconductor thin film near the grain boundaries or near the interface. A method for manufacturing a semiconductor device, comprising the steps of: (2) The step of forming the first and second insulating films is performed by low pressure CV
2. The method of manufacturing a semiconductor device according to claim 1, wherein the step is to form a silicon nitride film by a D method or a plasma CVD method. (3) The method of manufacturing a semiconductor device according to claim 1, wherein the step of forming the first and second insulating films is a step of forming a silicon nitride oxide film by a plasma CVD method. (4) The method of manufacturing a semiconductor device according to any one of claims 1 to 3, wherein the crystalline semiconductor thin film material is silicon. (5) The method of manufacturing a semiconductor device according to any one of claims 1 to 4, wherein the hydrogen is introduced by diffusion of hydrogen from the insulating film containing hydrogen. (6) The method of manufacturing a semiconductor device according to any one of claims 1 to 4, wherein the hydrogen is introduced using plasma containing hydrogen. (7) The method of manufacturing a semiconductor device according to any one of claims 1 to 4, wherein the hydrogen is introduced by ion implantation. (8) The method of manufacturing a semiconductor device according to any one of claims 1 to 7, wherein the irradiation light is near-infrared light.