JPH0427688B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to an electron beam annealing method. When forming a three-dimensional integrated circuit using an SOI (Semiconductor on Insulator) structure, a method using an electron beam is considered to be a promising method for forming the SOI. As methods for forming SOI, methods using laser beams, strip heaters, and lamps have been considered to date, but from the perspective of forming three-dimensional integrated circuits, the effects of heat on the underlying devices are It is thought that it is necessary to satisfy at least the following conditions: (heating time is several tens of milliseconds or less) and that crystal growth does not overlap during SOI formation (the width of growth during SOI formation is about the size of a chip). A forming means that satisfies the above conditions is a method using an electron beam or a laser beam. However, when laser beams are used, the current maximum output is about 20 W and the maximum beam diameter is about 100 μm, and the beams must be overlapped when forming a large-area SOI. When the beams are overlapped at this time, grain boundaries occur at the overlapped portion, making it difficult to form single-crystal SOI over a large area, which does not satisfy the conditions. On the other hand, when an electron beam is used, it is possible to extract a beam with a large area and high power, and an area about the size of a chip can be processed at once. Considering this point, the method using an electron beam is considered to be the method that satisfies the current conditions. Possible parameters when heat-treating a semiconductor using an electron beam include power density, scanning speed, and substrate temperature. Among these, when scanning the beam, using an electromagnetic coil provides stability of several percent, and regarding the substrate temperature,
If the temperature is higher than 500℃, heating by radiation can be used, and if the temperature is lower than 500℃, it is possible to make the temperature uniform within the surface by attaching a low melting point metal such as gallium or indium to the back surface of the semiconductor. , these parameters have not become a particularly big problem. On the other hand, controlling power density with good reproducibility is currently a major problem. The power density is determined by (acceleration voltage) x (beam current)/(beam area). Regarding the accelerating voltage and beam current, their stability is within a few percent, so no major problems have arisen at present. However, there are very big problems in that the area of the beam changes significantly and the power density is not reproducible. An object of the present invention is to improve the drawbacks of the conventional methods as described above and to provide a method for controlling the power density of an electron beam with good reproducibility. In the present invention, when performing electron beam annealing, at least one type of gas such as an inert gas such as argon gas or nitrogen gas is introduced into the vacuum chamber of the electron beam annealing apparatus to maintain a constant degree of vacuum in the passage path of the electron beam. It is characterized by maintaining the According to the present invention, when heat-treating a sample with an electron beam, the power density of the electron beam can be controlled with good reproducibility, making the heat treatment easier. Furthermore, by introducing gas and worsening the degree of vacuum,
It has been found that the beam can be narrowed down to a small diameter, and that if a partial heat treatment is to be performed using an electron beam with a small beam diameter, it is better to use a lower degree of vacuum. A more detailed explanation will be given below based on examples. First, we will discuss the case of semiconductor processing using a spot-shaped electron beam. The performance of the electron beam accelerator used was an accelerating voltage of 20 kV and a maximum beam current of 500 μA, and the substrate was fixed on a water-cooled copper holder using thermal compound. The beam diameter was determined using a Faraday gauge in two orthogonal axes directions. The electron beam has a Gaussian distribution, and the beam diameter was determined as the width at which the intensity of the electron beam becomes 1/e. Using such an apparatus, a predetermined power density was determined by changing the lens, changing the beam diameter, and changing the beam current. After that, the semiconductor was heated by irradiating an electron beam, and after replacing the sample, the heat treatment was performed again under the same conditions, but the way the sample was heated was very different. Although it did not, it did melt in the latter case. When we investigated the cause of this, we found that the beam diameter varied greatly depending on the degree of vacuum. An example of the results is shown in the table below.

[Table] As is clear from the table, as the degree of vacuum worsens,
The beam diameter tends to become smaller, and there is a nine-fold difference in power density between 4×10 ^-6 torr and 2×10 ^-4 torr even when the accelerating voltage and beam current are constant.
It can thus be seen that the beam diameter of the electron beam is easily influenced by the degree of vacuum. Possible causes of this are as follows. When an electron beam passes through a vacuum with a poor vacuum, it ionizes the residual gas around it, and because the ions are heavy, they move slowly and the electron beam is surrounded by positive ions. Therefore, the positive ions cancel out the effect of the space charge on the electron beam, thereby suppressing the spread of the electron beam and reducing the beam diameter. Therefore, in order to keep the beam diameter of the electron beam constant, it is considered necessary to keep the degree of vacuum in the passage of the electron beam constant. In order to maintain a constant degree of vacuum, gas can be controlled and introduced into the vacuum chamber and column, and
A vacuum gauge was fixed to the side of the sample chamber and electron gun. First, nitrogen gas was used as the introduced gas. After the sample chamber and column were evacuated to 2×10 ^-6 torr, nitrogen gas was gradually introduced to 2×10 ^-4 torr, and the beam diameter corresponding to each degree of vacuum was measured. Figure 1 shows the relationship between the beam diameter and the degree of vacuum in the sample chamber. At this time, acceleration voltage 2KV, beam current
The current was 400 μA, and all other parts such as the lens were measured under the same conditions. This shows that the beam diameter does indeed change depending on the degree of vacuum.
It was also confirmed that the beam diameter was indeed kept constant when the degree of vacuum was kept constant. As described above, it has been found that when heat treatment is performed using an electron beam, the beam diameter changes depending on the degree of vacuum, so it is necessary to maintain the degree of vacuum constant. Furthermore, it has been found that there is no problem in using an inert gas such as argon gas or a mixed gas thereof other than nitrogen gas as the introduced gas for controlling the degree of vacuum. In the above embodiment, a spot beam with a Gaussian distribution was used. Next, an embodiment will be described in which a linear electron beam is used. The conditions of the accelerator used were an acceleration voltage of 15KV.
The beam current was 15mA. When the degree of vacuum in the sample chamber is 4 × 10 ^-6 torr, the length of the beam in the long side direction is 3.5 mm, and the beam width is 800 μm (the width in this case is the distance at which the beam intensity decreases by 10%). I asked for it). In this state, introduce nitrogen gas,
When the degree of vacuum was worsened to 2×10 ^-5 torr, the beam length remained unchanged at 3.5 mm, but the beam width changed to 500 μm, which was affected by the degree of vacuum only in the short side direction. In the case of a linear beam, it is thought that the effect of canceling the space charge was stronger along the long side direction, and thus the effect of the degree of vacuum appeared strongly only in the short side direction of the beam. As mentioned above, in order to keep the beam diameter of a large current electron beam constant and to control it with good reproducibility, it is necessary to keep the degree of vacuum constant. It turns out that it can be controlled using a method. Furthermore,
It was found that the beam diameter of the electron beam could be made extremely small by introducing gas and reducing the degree of vacuum.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the relationship between the beam diameter of the electron beam and the degree of vacuum in the sample chamber when nitrogen gas is introduced into the sample chamber and the column.

Claims (1)

[Claims] 1. Introducing a certain amount of at least one type of inert gas or nitrogen gas into the vacuum chamber of the electron beam annealing device and irradiating the sample with the electron beam while maintaining a constant degree of vacuum in the electron beam passage path. An electron beam annealing method characterized by: