JPS6358829A - Pattern formation using electron beam - Google Patents

Abstract

PURPOSE:To reduce fog effect of an electron beam and to form a pattern of high dimensional accuracy by a method wherein a correction exposure is performed in such a manner that the amount of exposure in the electron beam patterning area becomes uniform by changing a beam current and a beam diameter by the automatic adjustment of an electrooptical system utilizing the frame-to-frame shifting time. CONSTITUTION:The first process in which the relation between the white-black ratio, which is the area ratio of the exposure part of the pattern to be formed and the non-exposure part and the amount of exposure, is computed, the second process in which the relation between the distance from the edge of pattern and the dimensional deviation is computed, the third proces in which the amount of exposure correction for the dimensional deviation of pattern is computed based on the rate of dimensional variation per unit of the amount of exposure, and the fourth proces, in which an electron beam is made to irradiate while the amount of exposure is being corrected based on the amount correction exposure for every patterning frame, are provided. When the correction of fog effect in the plane is going to be performed, the change of pattern size from the left end of the pattern area is measured in advance experimentally, the result is formed into numerical formula, and the dimensional variation can be suppressed to the minimum, if a correction dosage is given. When a patterning is actually performed, it is conducted by changing the beam current for every frame. The amount of the correction exposure is indicated by the slunt-lined part in the diagram.

Description

Detailed Description of the Invention [Object of the Invention] (Industrial Field of Application) The present invention relates to the exposure dimension of an electron beam, and particularly to wells of so-called electron beam fog that occurs during electron beam exposure.
The present invention relates to a pattern forming method using an electron beam that reduces the

(Prior Art) Technology for forming fine patterns on semiconductor substrates is essential to the manufacturing process of semiconductor integrated circuits.

A commonly used pattern forming method for forming such fine patterns is to apply a resist onto a substrate, irradiate this resist layer with an electron beam, and expose only the required pattern. This method involves dissolving the unexposed areas with a developer. That is, when a positive resist is used, the exposed portion is dissolved, and when a negative resist is used, the unexposed portion is dissolved to form a pattern.

Currently, a method in which a pentafold reticle drawn with an electron beam is reduced and exposed on a wafer is generally used mainly for memory products. Furthermore, as patterns become finer, a method of directly drawing patterns on a wafer with an electron beam is also becoming popular.

As described above, in the semiconductor integrated circuit manufacturing process, the method of forming a pattern using an electron beam occupies an extremely important position.

However, when electron beams are used for pattern formation, the fogging effect of the electron beams poses a problem. This fogging effect will be explained with reference to FIG. FIG. 6 is an explanatory diagram showing a state in which a semiconductor substrate is irradiated with an electron beam to form a pattern. The electron beam 2 passes through the objective lens 1 and the objective lens aperture 3 and is irradiated onto the semiconductor substrate 5. FIG. The semiconductor substrate 5 is fixed to a cassette 6, and the cassette 6 is fixed to an XY stage 7. Since the XY stage 7 moves in the horizontal direction, the entire surface of the semiconductor substrate 5 can be scanned by 1 m 2 of electrons. However,
A part of the irradiated electron beam 2 is reflected by the surface of the semiconductor substrate 5, and the reflected electrons 4 repeat the reflection between the shielding plate 8 and the semiconductor substrate 5 until their energy is dissipated. Therefore, the diameter portion in the figure is affected by the exposure by the reflected electrons 4, and generally L=20 seconds. This is an electron fogging effect, which causes dimensional variations in the pattern.

FIG. 7 shows the results of measuring pattern dimensional fluctuations caused by the fogging effect of electron beams. Figure (a) shows the original pattern to be formed, the white part is the exposed part,
Black areas indicate unexposed areas. Diameter 6 in the center of the pattern
A circular exposure pattern of 0al+ is located, and a stripe-like exposure pattern of a constant width (typically 10μ) is formed from the center to the right. When such a pattern is actually formed on a semiconductor substrate using a resist pattern forming method using electron beam irradiation, dimensional errors occur due to the fogging effect of the electron beam. Now, the dimensional error will be expressed by the width of the white part of the formed stripe pattern. That is, let the width of the white part of the original stripe pattern shown in FIG. 7(a) be W, and let the width of the white part of the stripe pattern actually formed on the semiconductor substrate be W', and this difference Δw-w' Consider w as the dimensional error of the pattern. FIG. 7(b) is a graph showing the results of measuring the dimensional error ΔW of this pattern with respect to the position of the pattern. Here, the position of the pattern is shown in Fig. 7 (
The horizontal position of the stripe pattern in a) is shown, and the horizontal positions in FIGS. 7(a) and (b) correspond. Although the dimensional error varies depending on the developing conditions, here the developing conditions were set so that the dimensional error ΔW of the white portion of the stripe pattern located on the circumference of the circular pattern in FIG. 7(a) becomes 0. As shown in Figure 7(b), the closer the result is to the center of the circular pattern, the wider the width becomes (ΔWho).
, conversely, the further away the width becomes, the narrower the width becomes (ΔW<0). This is because the center of a circular pattern is strongly affected by the electron beam fogging effect because most of the surrounding area is exposed (white area), whereas as it moves away from the center, the surrounding area is mostly unexposed (black area). This is because the influence of the electron fogging effect becomes weaker.

FIG. 8 is a diagram showing a pattern mask used to actually form a resist pattern on a semiconductor chip. The pattern mask 9 is a square mask with a part size of about 120M1, and the original pattern 10 is formed in the center part. Generally, one semiconductor wafer is composed of a plurality of semiconductor chips. The example shown in FIG. 8 is an example composed of semiconductor chips, and the original pattern 10 is also composed of the same four patterns 10-1 to 10-4. Note that here, X is the electron beam scanning direction, and Y is XY.
Indicates the direction of stage movement.

When electron beam irradiation is performed using such a pattern mask 9 to form a resist pattern, a dimensional error occurs between the central portion and the peripheral portion of the pattern 10 due to the fogging effect of the electron beam described above. That is, in the pattern mask 9, the exposed portion exists only inside the pattern 10, so that the peripheral portion of the pattern 10 is less affected by the fogging effect of the electron beam. Therefore, after forming the resist pattern, the semiconductor wafer is diced.
If it is divided into individual semiconductor chips, a dimensional error will occur between the pattern at point A and the pattern at point B in FIG. That is, if the pattern 10 is made up mostly of exposed parts, even if electron beam irradiation is performed at a proper focus, a dimensional difference of about 0.2 μm will occur between point A and point B. When the focus is shifted, this difference reaches about 0.7 μm.

Dimensional errors caused by the fogging effect of electron beams reduce dimensional accuracy, and have become an extremely serious problem in the manufacturing technology of semiconductor devices, which increasingly requires higher density.

(Problems to be Solved by the Invention) As described above, the electron beam fogging effect reduces dimensional accuracy and becomes an obstacle to high integration.

As a method of suppressing such a fogging effect, it has been proposed to change the structure of the shielding plate 8 shown in FIG. 6, but the processing is complicated and it has not been possible to completely eliminate the fogging effect.

In addition, the material of the shielding plate 8 is beryllium (Beryllium), which has a low reflection coefficient.
), charcoal 1 (C), etc. have been proposed, but these tend to have problems such as Be being toxic and dangerous, C causing cold, and the magnetic shielding being incomplete. It is not possible to completely eliminate the effect.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a pattern forming method using an electron beam, which can reduce the influence of electron beam fogging and form a pattern with high dimensional accuracy.

[Structure of the invention]

(Means for Solving the Problems) According to the present invention, the first step of determining the relationship between the black and white ratio, which is the area ratio of the exposed area and the unexposed area of the pattern to be formed, and the exposure amount; A second step of determining the relationship between the distance from the edge and pattern dimensional deviation, a third step of determining the corrected exposure value for pattern dimensional deviation based on the dimensional change rate per unit exposure ω, and exposure for each drawing frame. The method is characterized by comprising a fourth step of irradiating the electron beam while correcting the exposure amount using a corrected exposure amount.

(Function) As described above, in the present invention, correction exposure is performed so that the exposure within the electron beam writing area is uniform. Therefore, the corrected exposure amount is determined based on the black-and-white ratio of the entire pattern collected in advance and dimensional variation data according to the pattern area. The exposure amount during drawing is corrected by automatically adjusting the electron optical system to change the beam current and beam diameter using the time it takes to move from frame to frame.

Therefore, by performing pattern forming according to the present invention, a pattern can be formed with good reproducibility regardless of the pattern area or the black/white ratio of the pattern.

(Example) Examples of the present invention will be described below using poly(trifluoroethyl-alpha-O-roacrylate) as a positive electron beam sensitive resist and MIBK (methyl-isobutyl-ketone) at 25°C as a developer. A case in which the present invention is applied to an electron beam mask will be described in detail as an example.

First, an example will be described in which the exposure amount of the electron beam is corrected by changing the beam current.

10 having the pattern A-F shown in FIG. 9(b)
:1 or 5:1 compound reticle and Figure 9 (C)
When drawing using a raster scan drawing type electron beam exposure apparatus, the drawing area abcd of the master mask shown in FIG. Exposure is performed while scanning the beam. However, as described above, due to the fogging effect of the electron beam, variations occur in the energy density absorbed within the plane of the main pattern.

Figure 10 shows main pattern 15 and alignment mark 2.
FIG. 3 is an explanatory diagram showing in-plane variations in the amount of exposure absorbed in the resist in a pattern having a pattern of 0 and 0, in terms of energy bonds. In the main pattern 15, the energy irradiated to the central part is the largest and the energy irradiated to the peripheral part becomes smaller. Therefore, an energy gap of Δε occurs between the peripheral bracket and the central portion. The fogging effect is on the exposed area (W)
It also changes depending on the area ratio (hereinafter referred to as black-and-white ratio) of the unexposed area (B) and the unexposed area (B). .

FIG. 2 is a characteristic diagram showing the relationship between the black-and-white ratio and pattern dimensions, which are determined in advance from pattern data.

If such a change in pattern dimensions is converted into a change in exposure amount (dose amount), characteristic diagrams as shown in FIG. 3 are obtained.

The relationship between pattern dimensions and dose amount was calculated using the relationship between beam current and dimensional deviation ΔW shown in FIG. The relationship between pattern dimensions and dose changes linearly, with 0.2
It can be seen that a dimensional change of 0.1 μm occurs for a dose of μC/cj.

The relationship between black and white ratio W/(W+8) and exposure ff1D is D.

(μC/ci)=4.65-0.5(W/ (W+
B)')-(1) However, O<W/<W+8)≦
It becomes 1. In the case of a 0.5 μm address unit, the beam current Ao is expressed as A,=100D0.

Next, in-plane fogging effect correction is performed. The method will be explained using FIG. First, the pattern dimension change from the left end of the pattern area is experimentally measured in advance, and the results are expressed mathematically. In this case, the initial dose m is given by equation (1). In general, the distance r from the pattern edge (jllllI)
The relationship between and the pattern dimension deviation ΔW(r) (μm) is approximated by a quadratic equation, (i) 0≦r<j /2-15: β-j-30 ( 2)(ii)
l r-j /2 l≦15=ΔW(r)-α
・・・・・・・・・(3)(iii) j /
2+15<r≦J:ΔW(r)−ΔW(J)−r)
−・−・・・<4) is given. Here, 1 is the pattern size in the X direction (30 shoes or more), and α and β are constants.

For example, the initial exposure amount is 4.4μC/aj (beam electric 1440
nA, black and white ratio W/(W+8)=50%) and pattern size in the X direction j=60m+, (i) 0≦r<
15: ΔW(r)=-6,67X10' x r (r + 30) "..." (5) (
ii) 15≦r≦45: ΔW (r) =0. 15”-=(6)
(iii) 45<r≦60: It is expressed as ΔW (r) = W (60-r) = (7).

Figure 5 shows the relationship between the distance r from the pattern edge and the pattern dimensional deviation ΔW, both when it is given by the approximation formulas shown in equations (5) to (7) and when it is given by actual measured values. be. As is clear from FIG. 5, the approximation formula and the actual measured values almost match. Therefore, for this dimensional deviation, the dimensional change rate per unit dose is expressed as γ(μC/aj)/μm, and γ・(ΔWl/2)−ΔW(r)) (μC/d)...・(8)
If a correction dose ΔT (r) of 2 is given, the dimensional variation can be suppressed to a minimum.

During actual writing, dose correction cannot be performed continuously, so it is performed while changing the beam current for each frame. Representative dose of n frames ff1Dc (n
) is set to r-128 (2n-1), D (n)=Dg +D (r)-D, +Δ
D (128(2n-1))・・・・・・・・・(9
) is given by. The shaded area in FIG. 1 indicates this corrected exposure amount, and this dose amount correction is executed when moving from frame to frame.

For example, Toshiba electron beam exposure device EBM130/40
If you use , the frame-to-frame travel time (
The time required for frame positioning is approximately 1.5 seconds,
It is sufficient to automatically adjust the electron optical system within this time and change the beam current value by the amount of correction. Since this correction is carried out in a very short time, there is no effect of lengthening the actual drawing time at all.

Next, a case where correction is performed by changing the beam diameter will be described.

A correction dose amount is determined in advance using a method similar to that used when changing the beam current described above. Dose correction during actual writing is performed by automatically adjusting the beam diameter using the frame-to-frame movement time. This takes advantage of the fact that when the beam current is constant, the effective dose decreases as the beam diameter increases. When using this method, it is necessary to obtain a relational expression between the beam diameter and the effective dose amount in advance.

(Effects of the Invention) As described above in detail based on the embodiments, by using the pattern forming method according to the present invention, it is possible to reduce the dimensional variation within the substrate surface due to the fogging effect to half that of the conventional method, and to improve the pattern area and pattern. It is possible to form a resist pattern with good reproducibility regardless of the black-and-white ratio.Furthermore, since the dimensions of the developed area are constant, there is also the effect that the development time can be kept constant.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram for explaining a frame-by-frame corrective exposure method according to the present invention, and the shaded area indicates the corrected exposure amount. Figure 2 is a characteristic diagram showing the relationship between black and white ratio of a pattern and pattern dimensions, Figure 3 is a characteristic diagram showing the relationship between black and white ratio of a pattern and exposure amount, and Figure 4 is a diagram showing the relationship between beam current and pattern dimension deviation. FIG. 5 is a characteristic diagram showing the relationship between the distance from the pattern etch and pattern dimensional deviation. FIG. 6 is an explanatory diagram for explaining the fogging effect associated with the conventional method. Explanatory diagram showing pattern dimensional errors at each pattern position for testing the influence of fogging effect, No. 8
The figure is a plan view of a mask pattern used to form a resist pattern on a semiconductor chip, Figure 9 is an explanatory diagram for explaining the drawing method using electron beam exposure, and Figure 10 is the amount of exposure absorbed in the resist. FIG. DESCRIPTION OF SYMBOLS 1... Objective lens, 2... Electron beam, 3... Objective lens aperture, 4... Backscattered electrons, 5... Semiconductor substrate, 8
... Shielding plate, 9... Pattern mask, 10... Original pattern. Applicant's agent Mr. Sato Figure 1 Figure 3 Figure 4 Figure 5 Figure 6 Pattern 1 Figure 7 Figure 8 (b)' (C) Figure 9 Figure 10

Claims (1)

[Scope of Claims] 1. A first step of determining the relationship between the black and white ratio, which is the area ratio between exposed and unexposed areas of the pattern to be formed, and the exposure amount;
a second step of determining the relationship between the distance from the pattern edge and the pattern dimensional deviation; a third step of determining the corrected exposure amount for the pattern dimensional deviation based on the dimensional change rate per unit exposure amount; A fourth step of irradiating the electron beam while correcting the exposure amount using the corrected exposure amount.
A method for forming a pattern using an electron beam, comprising the steps of: 2. A pattern forming method using an electron beam according to claim 1, characterized in that the correction includes changing a beam current. 3. A pattern forming method using an electron beam according to claim 1, characterized in that the beam diameter is changed as the correction.