JP5068357B2 - Semiconductor device manufacturing method, photomask pattern design method, and photomask manufacturing method - Google Patents

Description

The present invention relates to a method of manufacturing a semiconductor device, which relates to the production how the pattern design method and photomask of the photomask, a method of manufacturing a semiconductor device in photolithography technique particularly forming a fine pattern in a semiconductor device, a photomask those related to the pattern design method and manufacturing how photomask.

  In recent years, there has been remarkable progress in high integration and miniaturization in semiconductor integrated circuits. Along with this, miniaturization of circuit patterns formed on a semiconductor substrate (hereinafter simply referred to as a wafer) is also progressing rapidly.

  Among them, photolithography technology is widely recognized as a basic technology in pattern formation. Therefore, various developments and improvements have been made to date. However, the miniaturization of patterns is unknown and demands for improving the resolution of patterns are becoming stronger.

  This photolithography technique is a technique in which a pattern on a photomask (original image) is transferred to a photoresist applied on a wafer, and an underlying etching target film is patterned using the transferred photoresist.

  At the time of transfer of the photoresist, the photoresist is subjected to a development process. The type in which the photoresist in the portion exposed to light is removed by this development processing is a positive type. The type to be removed is called negative photoresist.

In general, the resolution limit R (nm) in the photolithography technique using the reduced exposure method is
R = k ₁ · λ / (NA)
It is expressed as Here, the wavelength (nm) of light using λ, NA is the numerical aperture of the projection optical system of the lens, and k ₁ is a constant depending on the resist process.

As can be seen from the above equation, in order to improve the resolution limit R, that is, to obtain a fine pattern, a method of decreasing the values of k ₁ and λ and increasing the value of NA is conceivable. That is, it is only necessary to reduce the constant depending on the resist process and to shorten the wavelength and increase the NA.

However, it is technically difficult to improve the light source and the lens, and the depth of focus δ (δ = k ₂ · λ / (NA) ² ) of light becomes shallower by shortening the wavelength and increasing the NA. There is also a problem that leads to a decline.

  Under such circumstances, in manufacturing a semiconductor integrated circuit, it is necessary to form a fine pattern with a large process margin. The modified illumination method is effective for the formation of dense patterns and has been widely put into practical use. On the other hand, as a method for forming an isolated line pattern with a large process margin, there is a method using a Levenson type phase shift mask.

  However, in the case of the Levenson type phase shift mask, since it is necessary to produce a phase shifter for converting the phase of exposure light by 180 °, there is a problem that it is difficult to produce the mask. In addition, the Levenson-type phase shift mask improves resolution by positively interfering transmitted light with different phases, so it is easily affected by lens aberration of the projection exposure apparatus, and should be able to be obtained without aberrations. There is also a problem that characteristics cannot be obtained. For this reason, the method using the Levenson type phase shift mask has not been put into practical use.

  In addition, a process margin improvement method (so-called auxiliary pattern method) is considered by arranging lines having a line width that is not resolved on the mask along the original line pattern. However, this method has a problem that it is difficult to inspect a defect of the mask because the mask pattern dimension is extremely small.

Therefore, an object of the present invention is to provide a semiconductor device manufacturing method, a photomask pattern design method, and a photomask which can form a fine pattern without using an auxiliary pattern method or a phase shift mask and can easily inspect a defect of the mask. it is to provide a manufacturing how the mask.

The photomask pattern design method of the present invention includes the following steps.

  First, a fine line pattern figure portion is extracted from the design pattern layout. The mask dark line width W2 of the fine line pattern figure portion satisfies the relationship of 0.35 <W2 / (λ / NA) where λ is the wavelength of the exposure light and NA is the numerical aperture of the projection optical system. Adjusted to Then, two sets of light transmission aperture patterns having a line width W1 satisfying a relationship of 0.35 <W1 / (λ / NA) <0.65 are arranged so as to sandwich a mask dark line having a line width W2. .

  This makes it possible to design a photomask pattern that can form a fine pattern with high process latitude and high accuracy.

  According to the photomask manufacturing method of the present invention, at least two light-transmitting opening patterns are included as at least a part of the entire pattern based on the line widths W1 and W2 calculated by the above-described photomask pattern design method. A photomask is manufactured.

Thereby, the photomask which has the said mask pattern can be manufactured.
A method of manufacturing a semiconductor device according to the present invention includes a substrate having a main surface, and formed on the main surface of the substrate and running parallel to each other with substantially the same line width and spaced apart from each other. A light-shielding film having two light transmission aperture patterns isolated from the opening pattern , and the line width of the two light transmission aperture patterns is W1, the interval between the light transmission aperture patterns Is W2, and the minimum distance between the two light transmission aperture patterns and the other light transmission aperture patterns is W3, W1, W2, and W3 each have 0.54 <W2 / W1 and 1.08. < A step of preparing a photomask satisfying the relationship of W3 / W1, a first exposure step of exposing a first photoresist on the wafer surface by a projection exposure method through the photomask, and an exposed first photo By developing resist, Patani A step of processing the first film to be processed under the first photoresist using the patterned first photoresist as a mask, and removing the second photoresist after removing the first photoresist. A step of applying, a second exposure step of exposing a region of the second photoresist corresponding to a region other than a region sandwiched between the two light transmission aperture patterns, and a second exposed photo A step of patterning by developing the resist, a step of processing the first film to be processed under the second photoresist using the patterned second photoresist as a mask, and the first and second photoresists And a step of patterning a second processed film under the first processed film using the first processed film processed as a mask as a mask.

Thereby , a complicated fine pattern can be accurately transferred to a film to be processed using a photomask capable of forming a fine pattern with high process latitude and high accuracy .

  According to the photomask pattern design method of the present invention, it is possible to design a mask pattern capable of forming a fine pattern with high process tolerance and high accuracy.

  According to the mask manufacturing method of the present invention, a photomask having two sets of light transmission opening patterns can be manufactured based on the line widths W1 and W2 calculated by the above-described photomask pattern design method.

According to the method for manufacturing a semiconductor device of the present invention , a complex fine pattern can be accurately transferred to a film to be processed using a photomask capable of forming a fine pattern with high process latitude and high accuracy .

It is a top view which shows roughly the structure of the photomask in Embodiment 1 of this invention. It is a schematic sectional drawing in alignment with the II-II line of FIG. It is a figure for demonstrating the formation method of the pattern of the semiconductor device using the photomask in Embodiment 1 of this invention. It is a figure which shows the relationship between the pattern (a) of a photomask, and relative optical image intensity distribution (b). It is a figure which shows the relationship between exposure energy distribution (a) in case exposure energy is small, and a resist pattern (b). It is a figure which shows the relationship between exposure energy distribution (a) in case exposure energy is large, and a resist pattern (b). It is a figure for demonstrating normal illumination. It is a figure for demonstrating modified illumination. It is a top view which shows the structure of an annular illumination stop. It is a top view which shows the structure of a quadrupole illumination stop. It is a top view which shows the structure of a dipole illumination stop. It is a figure which shows the structure (a) at the time of using the photomask in Embodiment 1 of this invention as a halftone type phase shift mask, and the intensity distribution (b) of transmitted light. It is a figure for demonstrating the definition of each part in deformation | transformation illumination. It is a figure which shows relative optical image intensity distribution at the time of overexposing the photomask in Embodiment 1 of this invention. It is an enlarged view of the part enclosed by the thick line square of FIG. It is a figure which shows the relationship between CD value at the time of overexposing the photomask in Embodiment 1 of this invention, and a focus offset. It is the SEM photograph which image | photographed the upper surface of the resist pattern when a focus offset and an exposure amount are changed when the photomask in Embodiment 1 of this invention is overexposed. It is a figure which shows the relationship between the line width of a resist, and a focus offset. It is a figure which shows the relative optical image intensity distribution in 0.18 micromL / S. It is a figure which shows the relative optical image intensity distribution in the 0.18 micrometer width isolated dark line. It is a figure which shows the relative optical image intensity distribution in the 0.18 micrometer-wide isolated bright line. It is a figure which shows the relationship between CD value in 0.18 micromL / S, and a focus offset. It is a figure which shows the relationship between CD value and focus offset in the 0.18 micrometer-wide isolated dark line. It is a figure which shows the relationship between CD value and focus offset in the 0.18 micrometer width isolated bright line. It is a figure which shows the relationship between CD value and focus offset when the dimension W2 of the photomask in Embodiment 1 of this invention is changed. It is a figure which shows the relationship between CD value and focus offset when the dimension W3 of the photomask in Embodiment 1 of this invention is changed. It is a figure which shows the relationship between CD value and focus offset when the dimension W1 of the photomask in Embodiment 1 of this invention is changed. It is a figure which shows the relative optical image intensity distribution at the time of exposing the photomask in Embodiment 2 of this invention using normal illumination. It is a figure which shows the relationship between CD value when a photomask in Embodiment 2 of this invention is exposed using normal illumination, and a focus offset. It is a figure which shows relative optical image intensity distribution when it exposes using the halftone type phase shift mask in Embodiment 3 of this invention. It is a figure which shows the relationship between an image width and a relative exposure amount in a binary mask and a halftone type phase shift mask. It is a figure which shows the relationship between an image width and the line | wire width of a mask in a binary mask and a halftone type phase shift mask. It is a figure which shows the relationship between the line | wire width of a dark line image when it exposes using the photomask in Embodiment 3 of this invention, and the line | wire width of a mask. It is a figure which shows the relationship between the line width of an image and focus offset when it exposes using the photomask in Embodiment 3 of this invention when there is no lens aberration. It is a figure which shows the relationship between the line width of an image when it exposes using a Levenson type | mold phase shift mask when there is no lens aberration, and a focus offset. It is a figure which shows the relationship between the line width of an image and focus offset when it exposes using the photomask in Embodiment 3 of this invention in case there exists a lens aberration. It is a figure which shows the relationship between the line width of an image when it exposes using a Levenson type | mold phase shift mask in case there exists a lens aberration, and a focus offset. FIG. 3A is a plan view showing a gate pattern of an SRAM, a plan view of a first photomask (b), and a plan view of a second photomask (c). It is a schematic sectional drawing which follows the XXXIX-XXXIX line | wire of Fig.38 (a). It is a schematic sectional drawing which follows the XL-XL line | wire of FIG.38 (b). It is a schematic sectional drawing which follows the XLI-XLI line | wire of FIG.38 (c). It is a schematic sectional drawing which shows the 1st process of the formation method of the pattern using the photomask in Embodiment 4 of this invention. It is a schematic sectional drawing which shows the 2nd process of the formation method of the pattern using the photomask in Embodiment 4 of this invention. It is a schematic sectional drawing which shows the 3rd process of the formation method of the pattern using the photomask in Embodiment 4 of this invention. It is a schematic sectional drawing which shows the 4th process of the formation method of the pattern using the photomask in Embodiment 4 of this invention. It is a schematic sectional drawing which shows the 5th process of the formation method of the pattern using the photomask in Embodiment 4 of this invention. It is a schematic sectional drawing which shows the 1st process of the formation method of the pattern using a hard mask. It is a schematic sectional drawing which shows the 2nd process of the formation method of the pattern using a hard mask.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(Embodiment 1)
FIG. 1 is a plan view schematically showing a configuration of a photomask according to Embodiment 1 of the present invention, and FIG. 2 is a schematic sectional view taken along line II-II in FIG.

  Referring to FIGS. 1 and 2, photomask 5 of the present embodiment has a transparent substrate 1 made of, for example, quartz and a light shielding film 2 made of, for example, chromium. The light-shielding film 2 has a pair of light transmission aperture patterns 2a formed so as to run in parallel with each other with a substantially equal line width W1 and an interval W2 at the center in the drawing.

  The light-shielding film 2 is not only this pattern, but also a pair of light transmission aperture patterns formed so as to run in parallel with each other with the substantially same line width W1a shown in the left side of the drawing and the interval W2a. 2a, or a pair of light transmission aperture patterns 2a formed so as to run in parallel with each other with the interval W2b being substantially the same line width W1b shown on the right side in the drawing. .

  Each of the two pairs of light transmission opening patterns 2a has a wide interval W3 between the other light transmission opening patterns 2a and is isolated.

  The line width W1 (or W1a, W1b) of the light transmission opening pattern 2a, the interval W2 (or W2a, W2b) between the two light transmission opening patterns 2a, and the two light transmission opening patterns 2a Each of the intervals W3 with the other light transmitting opening pattern 2a satisfies the relationship of 0.54 <W2 / W1 and 1.08 <W3 / W1.

  The line width W1 (or W1a, W1b) of each of the two sets of light transmission aperture patterns 2a is 0. When the exposure light wavelength during exposure is λ and the numerical aperture of the projection optical system is NA. The relationship 35 <W1 / (λ / NA) <0.65 is satisfied. Further, the interval W2 (or W2a, W2b) between the two light transmission aperture patterns 2a satisfies the relationship of 0.35 <W2 / (λ / NA), and the other light transmission aperture patterns 2a and others. The distance W3 from the light transmission opening pattern 2a satisfies the relationship 0.70 <W3 / (λ / NA).

  Further, the length L of each of the two sets of light transmission opening patterns 2a satisfies the relationship of 1.3 <L / (λ / NA).

  Note that the line widths W1a and W1b of the two light transmission aperture patterns 2a on the left and right in the figure may be the same as the line width W1 of the light transmission aperture pattern 2a in the center, Different dimensions may also be used. However, both of these line widths W1a and W1b must satisfy the relationship of 0.35 <W1 / (λ / NA) <0.65. Also, the line widths W1a and W1b may have the same dimensions as long as the above relationship is satisfied, or may have different dimensions.

  In addition, the intervals W2a and W2b between the two light transmission opening patterns 2a on the left and right in the drawing may be the same as or different from the interval W2 between the two light transmission opening patterns 2a in the center. It may be a dimension. However, both of these intervals W2a and W2b need to satisfy the relationship of 0.35 <W2 / (λ / NA). Further, the distances W1a and W1b may have the same dimensions or different dimensions as long as the above relationship is satisfied.

  Further, the interval W3 between the two light transmission aperture patterns 2a at the center and the left side in the drawing and the interval W3 between the two light transmission opening patterns 2a at the center and the right side in the drawing are also 0.70 <W3. As long as the relationship of / (λ / NA) is satisfied, the dimensions may be the same or different.

  In addition, the length L of each of the two sets of light transmission opening patterns 2a at the center, the left side, and the right side in the drawing may be the same or different. However, all of these lengths L need to satisfy the relationship of 1.3 <L / (λ / NA).

  Next, a pattern forming method of the semiconductor device using the photomask shown in FIGS. 1 and 2 will be described.

  FIG. 3 is a diagram schematically showing a configuration of a projection exposure apparatus using a mask in one embodiment of the present invention. Referring to FIG. 3, this projection exposure apparatus reduces a pattern on a photomask and projects it onto a photoresist 21b on the surface of a wafer 21. The projection exposure apparatus includes an illumination optical system from the light source 11 to the pattern of the photomask 5 and a projection optical system from the pattern of the photomask 5 to the wafer 21.

  The illumination optical system includes a mercury lamp 11 as a light source, a reflecting mirror 12, a condensing lens 18, a fly-eye lens 13, a diaphragm 14, condensing lenses 16a, 16b, and 16c, a blind diaphragm 15, and a reflection. And a mirror 17. The projection optical system includes telephoto lenses 19a and 19b and a pupil plane stop 25.

  In the exposure operation, first, the light 11a emitted from the mercury lamp 11 is reflected by the reflecting mirror 12, for example, only g-line (wavelength: 436 nm), and becomes a single wavelength light. Next, the light 11 a passes through the condenser lens 18, enters each fly-eye constituent lens 13 a of the fly-eye lens 13, and then passes through the diaphragm 14.

  Here, the light 11b indicates an optical path created by one fly-eye component lens 13a, and the light 11c indicates an optical path created by the fly-eye lens 13.

  The light 11a that has passed through the diaphragm 14 passes through the condenser lens 16a, the blind diaphragm 15, and the condenser lens 16b, and is reflected by the reflecting mirror 17 at a predetermined angle.

  The light 11a reflected by the reflecting mirror 17 passes through the condenser lens 16c and then uniformly irradiates the entire surface of the photomask 5 on which a predetermined pattern is formed. Thereafter, the light 11a is reduced to a predetermined magnification by the projection lenses 19a and 19b, and the photoresist 21b on the semiconductor substrate 21a is exposed.

  In the present embodiment, the above exposure is performed by overexposure. That is, the exposure amount when exposing the photoresist 21b is the exposure amount at the boundary where the photoresist 21b becomes soluble or insoluble in the developer by the exposure, or the exposure amount at the boundary where the photoresist 21b becomes soluble from the insoluble property. 4 times to 20 times. In normal exposure, the exposure amount is about 2.5 to 3.5 times the boundary exposure amount.

  The photoresist 21b thus exposed is patterned by development. In this development, when the photoresist 21b is a positive type, only a portion of the photoresist to which an exposure energy of a predetermined value or more is input is removed, and in the case of a negative type, an exposure energy of a predetermined value or less is input. Only the marked part is removed. In this way, a pattern of the photoresist 21b is formed.

  Thereafter, the film to be processed is patterned by processing the underlying film to be processed, such as etching, using the pattern of the photoresist 21b as a mask.

Next, the intensity of the exposure light defined above will be described in detail.
For example, the relative light intensity distribution of the exposure light transmitted through the photomask 5 shown in FIG. 4A is as shown in FIG. That is, the light intensity of the exposure light transmitted through the sufficiently large opening pattern 2b is the highest, and the light intensity of the exposure light transmitted through the two pairs of openings 2a is smaller than that.

  Here, as shown in FIG. 5A, when the exposure energy input to the photoresist with respect to the sufficiently large opening pattern 2b becomes the exposure energy of the above boundary (here, 1.0), As shown in FIG. 5B, only the portion of the photoresist 21b corresponding to the opening pattern 2b is removed, and a pattern corresponding to the two pairs of opening patterns 2a cannot be obtained.

  Therefore, as shown in FIG. 6A, the exposure energy input to the photoresist for a sufficiently large opening pattern is, for example, five times the exposure energy input to the photoresist for a sufficiently large opening pattern in FIG. By doing so, the exposure energy input to the photoresist by the light transmitted through the two pairs of opening patterns 2a can be made larger than the exposure energy at the boundary (here, 1.0). Thereby, as shown in FIG. 6B, a pattern corresponding to the two sets of opening patterns 2a can be formed in the photoresist 21b.

  In other words, the exposure energy is the exposure energy given to the corresponding pattern on the wafer by the exposure light transmitted through the sufficiently large opening 2b of the photomask 5, because the photoresist 21b is soluble in the developer. This means that the exposure energy is 4 times or more and 20 times or less the exposure energy of the boundary that becomes soluble or the boundary that becomes insoluble to soluble.

  In this pattern forming method, the exposure may be performed by ordinary illumination, but is preferably performed by modified illumination. In the case of normal illumination, as shown in FIG. 7, exposure light is irradiated perpendicularly to the photomask 5, and the wafer 21 is exposed with three light beams of zero-order light and ± first-order light. However, if the pattern of the photomask 5 becomes finer, the diffraction angle increases, so that ± first-order light does not enter the lens in vertical illumination and may not be resolved.

  Therefore, as shown in FIG. 8, the illumination light beam is incident on the photomask 5 obliquely by the modified illumination. As a result, exposure can be performed with only two light beams of 0th-order light and + 1st-order or -1st-order light diffracted by the photomask 5, and resolution can be obtained.

  As the diaphragm 14 used for this modified illumination, an annular illumination diaphragm having an annular transmission part 14a as shown in FIG. 9 or a quadrupole illumination diaphragm having four transmission parts 14a as shown in FIG. May be. Further, as shown in FIG. 11, a dipole illumination stop having two transmission portions 14a may be used.

  Further, the photomask 5 shown in FIGS. 1 and 2 may be a halftone phase shift mask as shown in FIG. In this case, instead of the light shielding film 2, a semi-transmissive light shielding film 2 that transmits exposure light to some extent is used. The transflective light-shielding film 2 has a phase so that the phase of the exposure light after passing through the transflective light-shielding film 2 is substantially 180 ° different from the phase of the exposure light after passing through the light transmission aperture pattern 2a. A function of having a shifter function and attenuating the exposure light so that the intensity of the exposure light after passing through the transflective light-shielding film 2 is smaller than the intensity of the exposure light after passing through the light transmission aperture pattern 2a Have The transmittance of exposure light of the semi-transmissive light shielding film 2 is preferably 2% or more and 10% or less.

  As a result, as shown in FIG. 12B, the light of the opposite phase overlaps at the boundary between the light transmission opening pattern 2a and the light shielding portion, so that the lights cancel each other and the light intensity at the edge of the exposure pattern is reduced. And the resolution of the pattern image can be increased.

Further, as shown in FIG. 13, the radius R of the pupil stop is proportional to the sine (= NA) of the maximum incident angle of the light beam in imaging. The position a at the pupil of the maximum incident angle illumination beam to the photomask 5 is proportional to the sine of the maximum illumination angle of illumination, and the position b at the pupil of the minimum incident angle illumination beam to the photomask 5 is the minimum illumination angle of illumination. Proportional to sine. The illumination coherence index σ (coherency) is given by σ = a / R in the case of conventional illumination. The shape of the modified illumination is also the ratio of the maximum / minimum incident angle sine and NA, and is expressed as σ _out = a / R and σ _in = b / R. In this description, the projection is an equal magnification, and the reduction ratio r in the projection optical system is 1.

  In exposure using an annular illumination stop, the ratio (a / R) of the maximum incident angle sine a to the maximum incident ray angle sine R of the projection optical system is 0.6 or more and 0.9 or less. Is preferred. In the exposure using the annular illumination stop, it is preferable that the sine b of the minimum incident angle is ½ or more of the sine a of the maximum incident angle.

  In exposure using a quadrupole illumination stop, the ratio (a / R) between the maximum incident angle sine a and the maximum incident ray angle sine R of the projection optical system is 0.6 to 0.9. Preferably there is. In exposure using a quadrupole illumination stop, the ratio (b / R) between the sine b of the minimum incident angle and the sine R of the maximum incident light angle of the projection optical system is preferably 0.3 or more. .

  In the present embodiment, since the photoresist 21b is exposed by overexposure through the photomask 5 having the two pairs of light transmission opening patterns 2a, the resist size hardly changes even when the focus is changed. Can be formed with high process tolerance and high accuracy. The inventor of the present application has confirmed that the above effect can be obtained by conducting the following experiments.

The photomask 5 shown in FIGS. 1 and 2 is a binary mask using chromium (Cr) as the light-shielding film 2, the line width W1 is 170 nm, the interval W2 is 170 nm, and the interval W3 is 360 nm (that is, the pitch is 870 nm). . When this photomask 5 is exposed with KrF excimer laser light (wavelength: 248 nm), NA of 0.65, and 2/3 annular illumination (σ _out / σ _in = 0.80 / 0.53), FIG. An optical image having a relative optical image intensity distribution as shown in FIG. 14 was obtained.

  In FIG. 14, the high light intensity portions on the left and right sides of the optical image intensity distribution correspond to the two light transmission aperture patterns 2a, and the low light intensity portion between them corresponds to the two light transmission aperture patterns 2a. Corresponds to the light shielding part sandwiched. This optical image intensity distribution is shown by changing the focus position in the range of 0 to 0.5 μm. FIG. 15 shows an enlarged view of a portion surrounded by a thick line square in the figure.

  Referring to FIG. 15, it can be seen that light intensity (Iso-Focal Slice Level) that does not change the pattern size of the photoresist even when the focus is changed can be obtained by adjusting the exposure amount. It can also be seen that the size of the image (pattern) can be as fine as about 90 nm with this light intensity without dimensional variation. That is, by performing overexposure using the photomask 5 shown in FIGS. 1 and 2, a fine pattern with little dimensional variation due to defocusing can be formed.

  In addition, in order to form a pattern in a positive photoresist, the light intensity at the light shielding pattern portion sandwiched between the two light transmission aperture patterns is higher than a certain level (Resolution Criteria). It is necessary to be small. In the case of this imaging, since the light intensity becomes larger than that level at 0.5 μm defocus, it can be seen that a depth of focus of ˜1.0 μm can be obtained.

  That is, it can be seen that overexposure using the photomask 5 shown in FIGS. 1 and 2 can ensure a large depth of focus in forming a fine pattern with little dimensional variation.

  FIG. 16 is a diagram illustrating a relationship (CD-Focus characteristic) between a CD (critical dimension) value and a focus offset in the case of FIGS. 14 and 15. As is apparent from FIG. 16, it can be seen that when the CD value is in the range of 80 to 90 nm, the CD value hardly changes even when the focus changes, and the CD-Focus characteristics are good.

  FIG. 17 is an SEM photograph showing the upper surface of the photoresist pattern when the pattern of the photomask shown in FIGS. 1 and 2 is transferred to the photoresist under the above exposure conditions while changing the focus offset and the exposure dose. It is. Referring to FIG. 17, the numerical values given in the photograph are measured values of the CD value. This result also shows that the pattern dimension actually transferred to the photoresist hardly changes even when the focus changes.

  Further, FIG. 18 shows a resist line width (resist line width) when a photomask (binary mask) 5 having a Cr light-shielding film 2 shown in FIGS. 1 and 2 is used for exposure with modified illumination. ) And the focus offset. Referring to FIG. 18, it can be seen that according to the pattern forming method of the present embodiment, a photoresist pattern having a line width of ˜100 nm can be formed in the range of the focal depth of ˜1.0 μm.

  Next, when the pattern formed on the photomask is a line-and-space (L / S) pattern, an isolated dark line pattern, or an isolated bright line pattern, the present embodiment can be implemented even if the line is thinned by changing the exposure amount. It will be described that good characteristics such as form cannot be obtained.

  19, 20, and 21 show focus in each of 0.18 μmL / S (both line width and space width is 0.18 μm), 0.18 μm wide isolated dark line, and 0.18 μm wide isolated bright line. It is a figure which shows relative optical image intensity distribution which used as parameter.

  The broken lines in FIG. 19 to FIG. 21 indicate the light intensity (Exp. Level to Mask Width) corresponding to the exposure amount when the best focus is set to the mask size (0.18 μm) and the dimension of 0.10 μm. The light intensity corresponding to the exposure amount (Exp. Level to 0.10 μm Width) is shown. From FIG. 19 to FIG. 21, in any of the three patterns, it is expected that when the pattern width is narrowed, the dimensional variation increases with the focus change.

  22, FIG. 23 and FIG. 24 are diagrams showing CD-Focus characteristics when the exposure level is finely changed in each of 0.18 μmL / S, 0.18 μm wide isolated dark line and 0.18 width isolated bright line. It is. The relationship between the CD value and the focus offset shown in FIGS. 22 to 24 is the behavior expected from the image intensities shown in FIGS. That is, it can be seen that the smaller the dimension (CD value) is, the larger the CD value fluctuates with respect to the focus offset fluctuation, and the better the CD-Focus characteristic. In addition, 0.18 μmL / S has less CD value variation with respect to focus offset variation than 0.18 μm width isolated dark line and 0.18 width isolated bright line, but still 0.3 μm defocused. On the other hand, since a dimensional change of .about.0.02 .mu.m occurs, the CD-Focus characteristic is worse than that of the present embodiment.

  From the above, it can be seen that, in the present embodiment, a good CD-Focus characteristic that cannot be obtained with other patterns can be obtained by using a pattern having two isolated light transmission aperture patterns 2a.

  Next, the dimensions of each part of the light transmission opening pattern 2a shown in FIGS. 1 and 2 will be considered.

  25, 26 and 27 are diagrams showing CD-Focus characteristics when the dimensions W1, W2 and W3 of the respective parts of the photomask shown in FIGS. 1 and 2 are changed.

  In the measurements of FIGS. 25 to 27, the exposure level was adjusted for each mask pattern so that the image contrast was such that the pattern could be resolved within a focus range of 0 to 0.5 μm. Further, W2 / W1 was adjusted to be constant in the smallest dimension. Further, assuming that the exposure amount may be changed when the mask pattern is changed, the CD-Focus characteristic of the minimum dimension (resist) pattern that can be resolved up to a defocus of 0.5 μm is obtained.

First, referring to FIG. 25, when the dimension W2 is 0.16 to 0.20 [mu] m, the CD value is small, and the variation of the CD value due to the variation of the focus offset is small (that is, the focus characteristic is good). On the other hand, when the dimension W2 is 0.14 to 0.12 μm, the CD value becomes small, but the CD value fluctuates with respect to the focus offset fluctuation, resulting in poor focus characteristics. On the other hand, when the dimension W2 is 0.22 to 0.24 μm, it can be seen that the CD value can be increased with excellent focus characteristics in which the CD value varies little with respect to the focus offset variation. In view of such a result, taking into consideration the wavelength λ of the exposure light at the time of exposure and the numerical aperture NA of the projection optical system, a preferable range of the dimension W2 is
0.35 <W2 / (λ / NA)
It becomes.

Next, referring to FIG. 26, when the dimension W3 is larger than 0.32 μm, the dimension (CD value) of the dark line sandwiched between the two pairs of light transmission opening patterns 2a becomes small, and the focus offset varies. As a result, the CD characteristic does not fluctuate and the focus characteristic is improved. On the other hand, when the dimension W3 is 0.28 and 0.24, the focus characteristic is good, but the CD value becomes large. In view of this result, taking into account the wavelength λ of the exposure light during exposure and the numerical aperture NA of the projection optical system, the preferred range of the dimension W3 is
W3> 0.70 × (λ / NA)
It becomes.

Next, referring to FIG. 27, when the dimension W1 is 0.24 or more, the CD value becomes too large. When the dimension W1 is 0.10 or less, the CD value becomes large and the variation of the CD value with respect to the variation of the focus offset increases. Deteriorate. In view of this result, taking into account the wavelength λ of the exposure light during exposure and the numerical aperture NA of the projection optical system, the preferred range of the dimension W1 is
0.35 <W1 / (λ / NA) <0.65
It becomes.

From the above relationship,
W2 / W1> 0.35 / 0.65≈0.54
W3 / W1> 0.70 / 0.65 ≒ 1.08
It becomes.

As for the pattern length L of the two pairs of light transmission aperture patterns 2a shown in FIGS. 1 and 2, the dimensions change in the region of 0.3 μm or less from both ends in the longitudinal direction of the pattern. If correction (OPC: changing the mask dimension at the end) is not performed, at least the pattern length L needs to be 0.6 μm or more. In consideration of this, when a preferable range of the pattern length L is obtained,
1.3 <d / (λ / NA)
It becomes.

(Embodiment 2)
Relative by calculation when exposure is performed with normal illumination (σ = 0.85) with a KrF excimer laser beam and NA of 0.60 using the same photomask 5 as in the first embodiment (FIGS. 1 and 2) The optical image intensity distribution is shown in FIG. 28, and the CD-Focus characteristic is shown in FIG.

  28 and 29, by overexposure using the photomask 5 of FIGS. 1 and 2, a line having a line width of 140 nm has a CD value with respect to the focus without using any super-resolution method. As can be seen from FIG.

(Embodiment 3)
Although the dimensions are the same as those of the first and second embodiments, as shown in FIG. 12, a half-tone type phase shift mask (Atten-PSM) in which the transflective light-shielding film 2 transmits 3% of the exposure light is used. The numerical image NA of the projection optical system is 0.65, and the calculated optical image intensity distribution when exposed by 2/3 annular illumination (σ _out / σ _in = 0.80 / 0.53) is used. As shown in FIG. From the results of FIG. 30, it can be seen that the dimension at which the photoresist pattern dimension does not vary with respect to the focus variation is 80 μm, which is smaller than that of the first embodiment. It can also be seen that the image quality that can be resolved is maintained even with a defocus of 0.5 μm.

  FIG. 31 shows the result of examining the exposure amount margin for the binary mask 5 used in the first embodiment and the halftone phase shift mask of the present embodiment. The horizontal axis in FIG. 31 is the relative exposure level, and the vertical axis is the CD value (image width) of the image. The exposure margin defined by the equation (ΔCD (%) / ΔExp. (%)) Shown in FIG. 31 is 1.5 for the binary mask and 1.2 for the halftone phase shift mask. Both exposure margins can withstand practical use, but it is clear that the exposure margin is improved in the halftone phase shift mask.

  Note that ΔCD (%) in the above equation represents the variation of the CD value, and ΔExp. (%) Indicates the variation of the exposure amount.

  FIG. 32 shows how the CD value changes due to changes in the mask dimension (mask line width) in the binary mask used in the first embodiment and the halftone phase shift mask of the present embodiment. Referring to FIG. 32, in fine photolithography, MEF (mask error enchancement factor) becomes large, and a large technical barrier (strict dimensional uniformity of the mask becomes severe). However, according to the technique of the present application, the MEF is smaller than that of other techniques, which is ~ 1.5 in the case of a binary mask and ~ 1.3 in a halftone phase shift mask. Note that the normal MEF is 10 or more.

  FIG. 33 shows a mask between two light transmission aperture patterns 2a when only the interval W2 between the two light transmission aperture patterns 2a shown in FIG. 1 and FIG. 2 is changed and other dimensions are made constant. It is a figure which shows the change of the line width of the dark line image corresponding to it by the change of a dimension. Referring to FIG. 33, according to the technique of the present embodiment, the line width of the dark line image is proportional to the line width of the mask up to 80 nm. In the conventional method, since the line width of the image is 200 nm and is not proportional to the line width of the mask, it can be seen that the technique of the present embodiment is much more suitable for miniaturization than the conventional method.

  Further, the deterioration of the CD-Focus characteristics due to lens aberration was compared between the technique of the present embodiment and the case of using a Levenson type phase shift mask (Alt-PSM).

  FIG. 34 and FIG. 35 are diagrams showing CD-Focus characteristics in the case of using the technique of the present embodiment when there is no lens aberration and when using the Levenson type phase shift mask. FIG. 36 and FIG. 37 are diagrams showing CD-Focus characteristics between the technique of the present embodiment when there is lens aberration and the case where a Levenson type phase shift mask is used. The aberration is a low-order spherical aberration and is assumed to be 0.05λ.

  35 and 37, when the Levenson type phase shift mask is used, it is understood that the line width of the image remarkably fluctuates with the fluctuation of the focus due to the aberration, and the CD-Focus characteristic is greatly deteriorated. . On the other hand, from FIGS. 34 and 36, according to the technique of the present embodiment, even when there is lens aberration, the amount of change in the line width of the image due to focus fluctuation is small, and the CD-Focus characteristic is hardly degraded. I understand that there is no.

(Embodiment 4)
In this embodiment, a method for forming an actual pattern by overexposing the photomask 5 of FIGS. 1 and 2 will be described.

  FIG. 38A is a plan view showing a gate pattern of each transistor constituting a SRAM (static random access memory) memory cell. FIGS. 38B and 38C are schematic plan views showing patterns of the first and second photomasks used for forming the pattern of FIG.

  39 is a schematic sectional view taken along line XXXIX-XXXIX in FIG. 38 (a), FIG. 40 is a schematic sectional view taken along line XL-XL in FIG. 38 (b), and FIG. It is a schematic sectional drawing which follows the XLI-XLI line of ().

  First, a method for designing the patterns of the first and second photomasks 5 and 55 shown in FIGS. 38B and 38C will be described.

  Only the fine line portion is extracted from the design pattern shown in FIG. The line width of the extracted fine line portion is enlarged. At this time, a dimension of 0.35 <W1 / (λ / NA) <0.65 is added to both sides of the fine line portion to enlarge. By reducing the dimension W2 satisfying the relationship of 0.35 <W2 / (λ / NA) from the central portion of the pattern having the enlarged line width, two pairs of light transmission openings shown in FIG. Pattern 2a is designed.

  For this reason, as shown in FIG. 40, the first photomask 5 has a light shielding film (or a semi-transparent light shielding film) 2 having a plurality of two light transmission opening patterns 2a formed on the transparent substrate 1. It has a configuration.

  Further, the design pattern shown in FIG. 38A is provided with a two-pair pattern having a line width W1 satisfying the relationship of 0.35 <W1 / (λ / NA) <0.65 as shown in FIG. ) Is designed.

  Therefore, the second photomask 55 has a configuration in which the light shielding film 52 constituting the above-described light shielding pattern is formed on the transparent substrate 51 as shown in FIG.

  Next, a method of forming a gate pattern using the first and second photomasks 5 and 55 will be described.

  42 to 46 are schematic cross-sectional views corresponding to the cross-section of FIG. 39 showing the gate pattern manufacturing method in the order of steps. Referring to FIG. 42, an insulating layer 102 to be a gate insulating film is formed on a semiconductor substrate 101 made of silicon or the like. A conductive layer 103 serving as a gate electrode is formed on the insulating layer 102. For example, a positive photoresist 111 is applied on the conductive layer 103. Insulating layer 102 is made of, for example, a silicon oxide film, and conductive layer 103 is made of, for example, a polycrystalline silicon film doped with impurities.

  The photoresist 111 is developed after being subjected to the first exposure using the first photomask 5 shown in FIG. In the first exposure, the exposure is performed by overexposure with an exposure amount larger than that of normal exposure. As described in the first embodiment, this overexposure means that the exposure amount when exposing the photoresist 111, that is, the exposure energy to the pattern having a sufficiently large transmission opening is caused by the exposure of the photoresist 111 to the developer. Thus, the exposure energy is 4 to 20 times the boundary exposure energy from insoluble to soluble.

  As a result, a fine opening pattern 111 a corresponding to the two light transmission opening patterns is formed in the photoresist 111.

  Referring to FIG. 43, conductive layer 103 and insulating layer 102 thereunder are sequentially etched using patterned photoresist 111 as a mask to form opening pattern 103a. Thereafter, the photoresist 111 is removed by, for example, ashing.

Referring to FIG. 44, the upper surface of conductive layer 103 is exposed by this ashing or the like.
Referring to FIG. 45, for example, positive type photoresist 112 is applied to the entire surface, and then, photoresist 112 is subjected to second exposure using second photomask 55 shown in FIG. It is developed after. As a result, the photoresist 112 remains so as to cover the pair of opening patterns 103a and the portion sandwiched therebetween. The conductive layer 103 and the insulating layer 102 are removed using the pattern of the photoresist 112 as a mask.

  Referring to FIG. 46, this exposes the surface of semiconductor substrate 101 in the region where the pattern of photoresist 112 is not formed. Thereafter, the pattern of photoresist 112 is removed by, for example, ashing to form a gate pattern made of conductive layer 103 shown in FIGS.

  The first exposure may be performed a plurality of times until the photoresist 111 is developed. The second exposure may be performed a plurality of times until the photoresist 112 is developed.

  In the above description, the conductive layer to be a gate pattern is directly patterned using a photoresist pattern. However, the conductive layer to be a gate pattern may be patterned using a hard mask. This will be described below.

  47 and 48 are schematic cross-sectional views showing a method of forming a gate pattern using a hard mask in the order of steps. First, a hard mask pattern 121 is formed as shown in FIG. 47 instead of the gate pattern by the method shown in FIGS. Using this hard mask pattern 121 as a mask, conductive layer 122 serving as a gate electrode underneath is etched.

  Referring to FIG. 48, the conductive layer 122 is patterned by this etching to form a gate pattern.

  Note that an insulating layer serving as a gate insulating layer under the conductive layer 122 serving as a gate pattern is omitted for convenience of description.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  DESCRIPTION OF SYMBOLS 1 Board | substrate, 2 Light-shielding film, 2a Light transmission opening part, 2b Sufficient opening pattern, 5,55 Photomask, 14 Diaphragm, 14a Transmission part, 21 Wafer, 21a, 101 Semiconductor substrate, 21b, 111, 112 Photoresist, 51 transparent substrate, 52 light shielding film, 102 insulating layer, 103, 122 conductive layer, 103a opening pattern, 111a fine opening pattern, 121 hard mask pattern.

Claims (3)

Extracting a fine line pattern figure from the design pattern layout; and
The mask dark line width W2 of the fine line pattern figure portion satisfies the relationship of 0.35 <W2 / (λ / NA) where λ is the wavelength of exposure light and NA is the numerical aperture of the projection optical system. Adjusting the process to
Arranging two sets of light transmission aperture patterns having a line width W1 satisfying a relationship of 0.35 <W1 / (λ / NA) <0.65 so as to sandwich the mask dark line having the line width W2. A photomask pattern design method comprising:   A photomask having the two sets of light transmission aperture patterns as at least a part of all patterns is manufactured based on the line widths W1 and W2 calculated by the photomask pattern design method according to claim 1. , Photomask manufacturing method. A substrate having a main surface and two sets of light formed on the main surface of the substrate and running parallel to each other with substantially the same line width and spaced apart from each other, and isolated from other light transmitting aperture patterns A light shielding film having a transmission aperture pattern, the line width of the two light transmission aperture patterns is W1, the interval between the two light transmission aperture patterns is W2, and the two light beams When the minimum distance between the transmission aperture pattern and the other light transmission aperture pattern is W3, each of W1, W2, and W3 has a relationship of 0.54 <W2 / W1 and 1.08 <W3 / W1. Preparing a photomask to fill,
A first exposure step of exposing the first photoresist on the wafer surface by a projection exposure method through the photomask;
Patterning by developing the exposed first photoresist;
Processing the first film to be processed under the first photoresist using the patterned first photoresist as a mask;
Applying a second photoresist after removing the first photoresist;
A second exposure step of exposing a region of the second photoresist corresponding to a region other than a region sandwiched between the two sets of light transmission aperture patterns;
Patterning by developing the exposed second photoresist;
Processing the first film to be processed under the second photoresist using the patterned second photoresist as a mask,
As the first and second masking the first film to be processed that has been processed using the photoresist as a mask, further comprising the step of patterning the first second film to be processed under the film to be processed A method for manufacturing a semiconductor device.

Images