JPS6234136B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an electron beam exposure method, and more particularly to a method for forming a highly accurate electron beam exposure pattern by correcting the so-called proximity effect.

In pattern forming technology using electron beam exposure, correction of the so-called proximity effect is essential to improve pattern accuracy. As is well known, the proximity effect is caused by electron beam scattering (forward scattering) in the resist layer coated on the exposed object and electron beam scattering (backward scattering) from the substrate, which is the exposed object. , is based on the phenomenon that the actual electron beam intensity distribution in the resist layer is much wider than the external electron beam irradiation pattern.
In particular, when the distance between patterns is 2 μm or less, the negative effect of resulting in significant distortion of pattern shape and deterioration of accuracy becomes significant. The electron beam intensity distribution in the resist due to this scattering is expressed as a function of the distance r from the center of the beam irradiated from the outside using the following formula: It is known that the first term is given by forward scattering and the second term is given by backward scattering. Note that in formula (1), A, B, and C are constants determined by conditions such as the thickness of the resist and the material of the substrate.

The most common method for correcting the proximity effect is to set the optimal dose for each pattern in advance, taking into consideration its shape and distance from adjacent patterns, or to transform the drawn pattern. This is a method of keeping it in place. In either case, the amount of correction must be determined in advance at the time of creating pattern data, but conventionally this has been determined empirically based on experimental results. This makes reliable correction of the proximity effect extremely difficult as required exposure patterns become finer and more complex.

Therefore, in order to quantitatively determine the amount of correction, in order to determine the optimal dose for each pattern, the influence due to scattering from all other patterns is determined for each pattern based on the above formula, and this is integrated. It is considered that the amount of external electron beam irradiation can be reduced by a value equal to

Before explaining this, in order to facilitate understanding, the basic configuration of a typical electron beam exposure apparatus will be briefly explained using the conceptual diagram of FIG. The electron beam exposure apparatus main body 1 includes an electron gun 2, a convergent electron lens system 3,
It has an XY deflector 4 and irradiates a finely focused electron beam onto a sample 5 such as a substrate coated with resist.The position of the electron beam spot on the sample 5 is determined by pattern data from an electronic computer 6. is controlled by driving the XY deflector 4 via the DA converter 7 and amplifier 8. Although not shown, the electron beam is irradiated or deflected onto the sample 5 by a blanking device in response to a signal from the computer 6. Then, the electron beam spot on the sample 5 advances at the cycle of a clock signal controlled by the computer 6, and irradiates the sample so as to fill in a predetermined pattern to perform drawing. Therefore, by controlling the period of the clock signal, the amount of electron beam irradiation per unit area can be controlled.

Next, a method for quantitatively determining the amount of electron beam scattering and strictly correcting the proximity effect will be explained. Second
The figure shows an example of an exposure pattern to explain this, and patterns 11 to 13 are each rectangular,
Points a to d shown on the pattern 11 are sample points for determining the optimum exposure amount. The scattered electron beam intensity distribution during exposure of a certain pattern at a specific sample point is determined by integrating the intensity distribution of one electron beam spot expressed by the above equation (1) within the pattern. Expressing this generally in a formula, the scattered electron beam intensity distribution based on exposure of pattern j at sample point i is expressed by the following formula Fij=∫∫(r)dxdy (2).

Here, it is assumed that the irradiation amount per unit area is constant within one pattern. This is to facilitate control during actual exposure operations, and is based on the consideration that it is not necessary to significantly increase the amount of pattern data if only a constant clock cycle is set for each pattern. If this irradiation amount is Qj for rectangular pattern j, and the optimum exposure amount (development energy amount) determined by the resist is E, then at all sample points (m) on the rectangular pattern, each of the rectangular patterns 1~n
If it is assumed that the sum of the scattered electron beam intensities given by the exposure is made equal to the optimum exposure amount (development energy amount) E, the following simultaneous equations hold true.

F ₁₁ Q ₁ +F ₁₂ Q ₂ +...+F ₁ nQn=E...... FiQ ₁ +Fi ₂ Q ₂ +...+FinQn=E......(3)...... Fm ₁ Q ₁ +Fm ₂ Q ₂ +... +FmnQn=E That is, each equation in simultaneous equations (3) is a rectangular pattern 11
Sample points a, b, c, d, ... (m pieces) inside
This indicates that the total scattered electron beam intensity in each case is equal to E.

By establishing this simultaneous equation for all other patterns, a simultaneous equation consisting of m×n equations can be obtained. When multiple sample points are set on one pattern as shown in Figure 2, the total number of equations m x n is greater than the number of variables Qn, so the simultaneous equations generally do not have a specific solution. It turns out.
It is possible to set only one sample point for each pattern, but in that case, only a part of the electron beam intensity distribution within the actual pattern is taken into account, which tends to cause large errors for some patterns. There is. Therefore, when a plurality of sample points are set for each pattern, each dose Qn is determined by taking the least squares solution of the minimum norm as an approximate solution to the simultaneous equations. Even though this method uses electronic computers to calculate the above approximate solution, as the number of sample points increases to improve accuracy, and the total number of required patterns increases dramatically, it takes a long time. This method requires a considerable amount of processing time, making it difficult to apply in reality.

Therefore, the present invention eliminates these conventional drawbacks, and provides an electron beam that can determine the electron beam irradiation amount in a relatively short period of time, and that can obtain an exposure pattern that is sufficiently accurate for practical use, although it is an approximate correction. The present invention aims to provide an exposure method.

The basic principle of the present invention is that when determining the electron beam irradiation amount for each pattern from the simultaneous equations as described above, at least two sample points are set on one pattern, and the scattered electrons from the surrounding patterns are The beam intensity is determined by associating only one sample point with the representative point of one pattern, thus specifying the solution of the simultaneous equations by assuming that the total number of equations in the simultaneous equations is equal to the total number of patterns,
Each pattern is drawn using the electron beam irradiation amount obtained thereby.

More preferably, the sample points corresponding to one pattern are selected far from the pattern, and the pattern required for drawing with an electron beam with a determined dose during actual exposure processing is selected. By drawing a pattern in which the outer edge of the wafer is removed by a certain fine width, preferably 0.1 to 0.3 μm, a highly accurate pattern can be obtained. In practice, the above-mentioned electron beam irradiation amount is determined at the time of pattern data creation, and if the device is like the one shown in FIG. It is used to control the period of the clock signal for stepping the beam spot.

The procedure for setting the electron beam irradiation amount according to the present invention will be described in detail below. FIG. 3 shows a pattern for explaining this, in which two sample points k ₁ and k ₂ are set on the pattern 21. Sample point k ₁ ,
_k2 is set on the longitudinal side of the pattern 21, but this is because, in integrated circuit patterns, the line width is generally smaller than the longitudinal tolerance, so the line width, In other words, this is because higher accuracy is often required for the distance between k ₁ and k ₂ , and it is generally extremely appropriate to set sample points at the locations where the line width is determined in this way. .

When calculating the actual scattered electron beam intensity in the pattern 21 using the sample points k ₁ and k ₂ , only one sample point k ₁ or k ₂ is associated with each of the patterns 21 to 23. At this time, as shown in the figure, the more distant sample point k ₂ is associated with the pattern 22, and the sample point k ₁ is similarly associated with the pattern 23. Pattern 21 with sample points
Regarding itself, it is the same no matter which one is associated with it, but here, sample point _k1 is selected for convenience. When the electron beam irradiation amounts of the patterns 21 to 23 to be determined are respectively Q ₁ , Q ₂ , Q ₃ and the distances to the corresponding sample points are respectively r ₁ , r ₂ , r ₃ , the method of the present invention Then, the following formula F(r ₁ )Q ₁ +F(r ₂ )Q ₂ +F(r ₃ )Q ₃ =E
...(4) is established. Here, for example, F(r ₃ ) is the scattered electron beam intensity at the sample point _K1 , which is scattered from the pattern 23, and is determined by the following numerical calculation.

F(r ₃ )=∫∫(r)dxdy...(5) Here, r indicates the distance from the sample point _K1 to the center of each electron beam spot on the pattern 23,
Integration is performed within pattern 23.

In other words, each term on the left side represents the intensity at the corresponding sample point when the electron beam irradiated to each pattern spreads due to scattering, and the sum total does not have a strict physical meaning, but This closely approximates the average electron beam intensity at any point within the pattern 21.

Similarly, for pattern 23, if we set two sample points each and create an equation similar to equation (4) using the same procedure as above, we can obtain 3 for the three variables Q ₁ to _{Q 3} .
Since a system of equations consisting of two equations is obtained, which has a specific solution, Q ₁ to _{Q 3} can be determined.

As mentioned above, since the scattered electron beam intensity from each pattern is obtained at a sample point further away, the result is Q ₁ ~ obtained from the simultaneous equations.
The value of Q ₃ is uniformly slightly larger than the optimum exposure amount E. However, this is generally preferable to establishing simultaneous equations by correlating closer sample points. In the case of patterns for integrated circuits, as shown in Fig. 4, the spacing between adjacent patterns 24 and 25 is extremely narrow, or there are many adjacent patterns that are in contact with each other. When drawing with
This is because, as shown in the figure, the peripheral area 26 on the side where no adjacent pattern exists often causes undesired results.

Therefore, in order to compensate for the tendency of overexposure, the fifth
As shown in the figure, when writing with the determined irradiation doses Q ₁ to _{Q 1} , a pattern is created by removing the peripheral portions of the target patterns 21 to 23 over a certain minute width. An electron beam is applied to 21', 22', and 23' to perform drawing. Actual results have shown that the fine width is preferably in the range of 0.1 to 0.3 μm.

The electron beam irradiation amount for each pattern is determined by the above procedure, and this is stored in a storage device together with the pattern data. For example, in an electron beam exposure system as shown in FIG. A deflection signal based on the pattern data is given to the deflector 4 at each time, and the DA converter 7
The period of the input deflection data signal is controlled to perform writing using an electron beam. In this way, when PMMA is used as the electron beam resist, the large-scale integrated circuit pattern exposed to electron beam has a pattern accuracy of 0.1μ after development.
m, and it was confirmed that pattern distortion due to the proximity effect did not occur locally.

As described above, according to the method of the present invention, when determining the amount of electron beam irradiation for each pattern, although a plurality of sample points are set for each pattern, the amount of electron beam irradiation for each pattern is determined. In order to find the scattered electron beam intensity, one sample point is associated with one pattern, so the electron beam irradiation amount can be determined from a set of simultaneous equations with a specific solution, and the process is relatively simple. It can be done in a short time. Although the electron beam irradiation amount determined in this way is obtained approximately, when applied to pattern formation by actual electron beam exposure, the least squares method described as a conventional example is used. It is possible to achieve pattern accuracy that is comparable to that in the case where the In particular, as explained above, if a sample point that corresponds to each pattern is selected from a farther distance and the electron beam irradiation amount is determined so as to cause a slight overexposure, the resist film thickness at the exposed pattern area may be reduced. Another advantage is that small particles (in the case of negative resist) and residues (in the case of positive resist) can be completely eliminated.

In the explanation of FIG. 3 or FIG. 5, only three patterns are shown for the sake of simplicity, but the method of the present invention is particularly suitable for forming integrated circuit patterns having hundreds of thousands to millions of patterns. This has an extremely high practical advantage, and it goes without saying that the number of samples for each pattern may be three or more.

[Brief explanation of the drawing]

Fig. 1 shows an example of the basic configuration of an electron beam exposure system, Fig. 2 shows a pattern for explaining the conventional procedure for determining the electron beam and irradiation amount, and Figs. 3 to 5 is an example of a pattern for explaining the procedure for determining the electron beam and irradiation amount in the method of the present invention.

Claims (1)

[Scope of Claims] 1. When exposing a sample by irradiating an electron beam to draw a large number of independent patterns and performing exposure processing, the amount of electron beam irradiation per unit area of a specific pattern is applied to a large number of surrounding patterns. This is a method in which the intensity of the scattered electron beam in a particular pattern during electron beam irradiation is set to a corrected value, and drawing is performed with the irradiation amount set for each pattern. , a unique sample point is set on a specific pattern in association with the specific pattern and a number of patterns around it, and the scattered electron beam during electron beam irradiation is set based on the associated pattern for each sample point. For each specific pattern, perform the steps of determining the intensity and then determining the total sum of the determined scattered electron beam intensities.
An electron beam exposure method characterized by setting the electron beam irradiation amount for each pattern and performing exposure so that the sum of the electron beam intensities determined for each specific pattern becomes a predetermined electron beam exposure intensity. 2. When determining the intensity of the scattered electron beam at a sample point on a specific pattern when irradiating the plurality of patterns with the electron beam, a sample point farther from the pattern is selected as the sample point corresponding to each pattern. An electron beam exposure method according to claim 1. 3. When each pattern is drawn based on the determined electron beam irradiation amount, the pattern is drawn by irradiating the pattern with a predetermined minute width removed from the periphery of each pattern. The electron beam exposure method described in . 4. The electronic device according to any one of claims 1 to 3, wherein each pattern is rectangular, and each sample point is set on a longitudinal side of each rectangular pattern. Beam exposure method.