JP2001296647A - Photomask and exposure method using the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a photomask superior in the depth of a focus and MEF (mask error factor) and effective in suppressing the occurrence of a side lobe and shortening exposure time. SOLUTION: The photomask consists of a transparent substrate 1, made of quartz, a halftone film 3 having about 180 deg. for phase shift amount and 3-10% transmittance and a light-shielding film 2 comprising Cr. The halftone film 3 and the light-shielding film 2 have a common opening 4 on the substrate 1. The light-shielding film 2 is formed on the halftone film in a width (Cr width) equidistant from the four sides of the opening in directions perpendicular to the sides. The Cr width is set smaller than the width of the halftone film.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a photomask for a projection exposure apparatus used in a lithography process in a semiconductor device manufacturing process and an exposure method using the same.
In particular, the present invention relates to a phase shift mask using a translucent film and an exposure method using the same.

[0002]

2. Description of the Related Art In a semiconductor manufacturing process, a lithography technique for forming a resist pattern based on a circuit design using a material (= resist) sensitive to light or radiation, for example, a photosensitive resin is used. What is required for this lithography technique is resist size control suitable for design, a wide depth of focus (DOF (Depth of Focus)), high alignment accuracy, as well as throughput.
In lithography, an exposure method by direct writing on a resist and a photomask (a mask for a reduction projection device is generally called a reticle, but in the present specification, a mask or a photomask including a reticle) are used. There is an exposure method. Advantageous in terms of product cost throughput is lithography with reduced projection exposure,
Currently, it is the mainstream in the semiconductor device manufacturing process. In recent years, with the miniaturization of semiconductor devices, the minimum line width called a design rule has been steadily miniaturized, and as a result,
The required minimum line width approaches or exceeds the resolution limit of the exposure apparatus. To cope with this, a phase shift mask technique for improving the resolution of the photolithography technique by forming a phase shift film for controlling the phase of transmitted light on a mask has been adopted as a useful means. Of the shift mask,
The phase of transmitted light is controlled using a halftone film (semi-transparent film) that has some transmittance for exposure light,
A halftone mask for forming a desired pattern has been put into practical use as a technology applicable to mass production due to ease of mask design and mask fabrication.

FIG. 12 shows an ordinary photomask in which a portion other than a transmission portion of a conventionally used photomask is completely shielded from light. FIG. 12A is a cross-sectional view of the photomask.
FIG. 12B is a plan view as viewed from the surface having the light shielding film. As shown in FIG. 12, the photomask includes a transparent substrate 1 made of quartz (Qz) and a light shielding film 2 made of chromium (Cr). The light-shielding film has an opening 4
[Hereinafter, a photomask having this structure is referred to as a binary mask.] Also,
FIG. 13 shows a phase shift halftone mask (HTPSM: Half Tone Ph) having a halftone film having a phase shift amount of about 180 degrees and a light transmittance of about 3 to 10% on a transparent substrate.
ase Shift Mask). FIG. 13A is a cross-sectional view of the photomask, and FIG. 13B is a plan view of the photomask viewed from a surface having a halftone film. As shown in FIG. 13, a halftone film 3 is formed on a transparent substrate 1 made of quartz. The halftone film 3 has an opening 4
Are formed. The resolution of the leaked light from the halftone film can be improved because the phase of the leaked light from the halftone film is inverted by 180 degrees with respect to the light transmitted through the opening.

It is known that the phase shift type halftone mask shown in FIG. 13 particularly has a larger DOF of the contact hole pattern than the normal mask shown in FIG. However, DOF is used for hole patterning.
Is not only desired, but the ME defined by (resist pattern dimensional error) / (mask dimensional error)
It is desired that F (Mask Error Factor) is small. Therefore, it is difficult to obtain a mask excellent in both DOF and MEF. Further, as a specific problem of the phase shift type halftone mask, a side lobe generated at a position where no transmission pattern exists on the mask due to the influence of light transmitted through the halftone film may expose the resist.

FIG. 14 is a diagram for explaining side lobes generated during exposure of a phase shift type halftone mask. FIG. 14A is a cross-sectional view of a phase shift halftone mask, and FIG. 14B is a distribution diagram of the light intensity of the light transmitted through the mask. As shown in FIG. 14B, a light intensity peak appears at the center position of the diameter of the opening 4. And
A side lobe, which is a local maximum of the light intensity, is generated around the transmitted light pattern. When the side lobe exceeds a threshold value for exposing the resist indicated by a dotted line in FIG. 14B, the side lobe is transferred to the photoresist. When this side lobe is transferred onto the photoresist, unnecessary "resist digging" occurs after development in the case of a positive resist, and unnecessary resist residue occurs in the case of a negative resist. Occur. There is a mask bias as a method for preventing the generation of unnecessary patterns due to side lobes. When the bias is increased, the optimum exposure amount at the time of transfer is reduced, so that the side lobe intensity is reduced. However, according to this method, transfer of the unnecessary pattern can be surely prevented, but at the same time, the optical contrast is lowered, so that it is not a fundamental solution.

Further, when a conventional photomask is used, the following problem occurs when a random layout pattern in which an isolated hole pattern and a dense hole pattern are mixed is formed. When forming a fine hole pattern group having a random layout, if the coherence factor σ is small, a pattern distortion occurs due to a difference in pitch in each of the horizontal (X) and vertical (Y) directions with respect to an adjacent hole pattern. It has been found that a coherence factor σ effective for suppressing pattern distortion is large σ ordinary illumination of 0.6 or more. However, in the case of a large σ illumination having a coherence factor σ of 0.6 or more, a change in the dimension itself is inevitable due to the pitch between adjacent hole patterns (density of the pattern). This is due to the effect of the optical proximity effect. Generally, the size of dense holes tends to be larger than that of isolated holes.

FIG. 15 is a graph showing the pitch dependency of the dimension generated when exposure is performed using a binary mask when the pitches in the X and Y directions are equal. In FIG.
The horizontal axis indicates the pattern pitch, and the vertical axis indicates the size of the formed resist pattern. At this time, the mask size is 18
0 nm (value converted to the size on the wafer. Hereinafter, all such dimensions are shown as values converted to the size on the wafer. The size on the reticle is 4 to 5 times the size on the wafer). FIG. 16 is a graph showing the Y-direction pitch dependency of the dimension (the geometric mean of the X-direction dimension and the Y-dimension, but the difference between them is less than 20 nm) when the pitch in the X-direction is fixed at 340 nm. . In any case (when the exposure amount is set to the isolated hole), the size of the dense hole greatly deviates from the standard.

It is known that mask dimension correction (Bias OPC) is performed as a means for correcting a dimensional change caused by pattern density. Even when this method is used, it is necessary to suppress the dimensional variation. It is difficult. The reason is shown below. The mask dimension correction is performed at a fixed step size (grid size). In a fine hole smaller than the exposure wavelength, the ratio of the resist pattern size result to the mask step size (ie, MEF) is as large as 3 or more. For this reason, the dimensional correction cannot be performed accurately with a grid size of about 10 nm which is generally used at present. To increase the accuracy, it is necessary to reduce the grid size of the electron beam lithography system used to make the mask. However, the finer grid results in a dramatic increase in the drawing data and drawing time, and the burden required for mask making increases significantly. I do. Although the above description of the dimensional change caused by the pattern density is related to the binary mask, the same applies to the ordinary phase shift type halftone mask.

[0009]

As described above, the phase shift type halftone mask is improved in DOF as compared with a normal photomask, but there is a problem that an unnecessary pattern is generated due to a side lobe. . In recent years, it has been required to form a hole having a size smaller than the exposure wavelength.
It is difficult to secure both DOF and MEF to a satisfactory level using any of the TPSMs. Also,
With regard to lithography technology, reduction in exposure time is required from the viewpoint of cost reduction and throughput. further,
When patterning a random pattern in which an isolated pattern and a dense pattern are mixed, it is difficult to prevent a dimensional change depending on the pattern using a binary mask or a normal halftone mask. Was. The present invention has been made to solve the above-described problems, and aims to prevent side lobe transfer, improve DOF, reduce MEF, shorten exposure time, and suppress dimensional change caused by pattern density. It is an object of the present invention to provide a photomask and an exposure method that can perform the method.

[0010]

According to the present invention, a semi-transparent film for shifting the phase of exposure light transmitted therethrough is formed on a transparent substrate transmitting the exposure light. In a photomask configured such that a phase difference of 180 degrees is generated between light transmitted through the translucent film and light transmitted only through the transparent substrate, light is completely shielded around the opening of the translucent film. A photomask characterized by having a film formed thereon is provided. In order to achieve the above object, according to the present invention, a translucent film that shifts the phase of the transmitted exposure light is formed on a transparent substrate that transmits the exposure light, and the light transmitted through the translucent film is formed. And a photomask configured such that a phase difference of 180 degrees occurs between light transmitted only through the transparent substrate and
A photomask is provided, wherein the translucent film around the opening of the translucent film is replaced with a light-shielding film over the entire periphery of the opening. Preferably, the width of the light-shielding film is formed to a size of 50 nm or more and 150 nm or less. More preferably, the transparent substrate is formed of any of quartz, CaF ₂ , and MgF ₂ , and the translucent film is formed of CrON, CrO, MoSiON, MoSiO, a-
C: formed of one of H and SiN, and the light-shielding film is formed of Cr or CrO. According to another aspect of the present invention, there is provided an exposure method using the photomask, wherein the exposure method uses normal illumination light. When using a photomask including a pattern with a random pitch, preferably,
The exposure is performed using normal illumination light having a coherence factor σ of 0.6 or more.

[0011]

Next, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 (a), (b)
1 shows a cross-sectional view and a plan view of a photomask according to an embodiment of the present invention. As shown in the figure, the photomask according to the present embodiment has a transparent substrate 1 made of quartz, shifts the phase of exposure light by about 180 degrees, and has a light transmittance of 3 to 10.
% Halftone film 3 and light-shielding film 2 made of Cr. The halftone film 3 and the light shielding film 2 have a common opening 4 on the transparent substrate 1. Here, the light shielding film 2 is formed on the halftone film with an equal distance (Cr width) in the vertical direction from the opening end to each side of the opening. The dimensions of this Cr width are
The width is less than the width of the halftone film in FIG. As a material for forming the halftone film, for example, a chromium-based (Cr
ON, CrO), molybdenum (MoSiON, MoS)
One of iO), amorphous carbon (aC: H), and silicon nitride (SiN) is used. As a material of the transparent substrate, besides quartz, calcium fluoride (CaF
₂ ) or magnesium fluoride (MgF ₂ ) can be used. As a material of the light shielding film, CrO may be used in addition to Cr.

The use of the photomask having the structure of the present embodiment can achieve the effects of shortening the exposure time and reducing the MEF as compared with the phase shift type halftone mask. The reason that the exposure time can be shortened is that in the conventional phase shift type halftone mask shown in FIG. 13, the phase of the light passing through the opening 4 and the phase of the light passing through the halftone film 3 are 1 unit.
Since it is shifted by 80 degrees, the peak value of the light intensity (Ima
x) is reduced. However, in the photomask according to the embodiment of the present invention shown in FIG. 1, the light-shielding film provided on the halftone film around the opening weakens the effect of the cancellation by the halftone film, so that the reduction of Imax is suppressed. be able to. That is, it is possible to shorten the exposure time.

In recent years, M is used as an index of lithography accuracy.
EF is considered important. As described above, the MEF is defined by (resist pattern dimension error) / (mask dimension error). When a dimensional error occurs in a mask, a pattern having the error is transferred onto a photoresist through a projection optical system. The smaller the error to be transferred, the better, so the ideal value of the MEF is 1. The larger the MEF, the worse the dimensional accuracy of the transferred pattern. However, with the current photography technology, it is difficult to obtain an ideal value of the MEF especially in a phase shift type halftone mask.

FIG. 2 shows that the photomask according to the present invention is ME.
It is sectional drawing for demonstrating the effect which suppresses F. FIG. 2A shows a case where the mask is formed with a dimensional error of 0, and FIGS. 2B and 2C show a case where the mask dimensional error is positive and a case where the mask is negative, respectively. Even if there is an error in the opening width, the outer shape of the light shielding film 2 does not change so much. Therefore, when the opening width is formed large, as shown in FIG. 2B, the Cr width is formed narrow. And conversely, if the opening width is formed narrow,
As shown in FIG. 2C, the Cr width is formed widely. In the case where exposure is performed using a photomask having a positive mask error shown in FIG. 2B, the width of the mask opening increases, so that Imax increases and the resist pattern corresponding to the mask opening is formed wider. (That is, the dimensional error of the resist pattern is positive). However, since the light-shielding film width is formed to be narrow and the halftone film transmitted light that cancels Imax exists near the aperture transmitted light,
ax is suppressed, and an increase in the resist pattern to be resolved is suppressed. In the case where exposure is performed using a photomask having a negative mask error shown in FIG. 2C, the mask opening width becomes narrower, so that Imax decreases and the resist pattern corresponding to the mask opening becomes narrower. It is formed. However, since the width of the light-shielding film is widened and the halftone film transmitted light that cancels Imax is transmitted away from the aperture transmitted light, the effect of canceling the transmitted light by the halftone film transmitted light is reduced. Reduction of the resist pattern to be suppressed and resolved is suppressed. That is, in any case, the increase in the dimensional error of the resist pattern is suppressed, and the MEF is reduced.

FIG. 3 shows the results of simulating the MEF, Imax and DOF for a 130-nm isolated hole pattern with a binary mask, HTPSM and the photomask of the present invention. Exposure conditions are
Exposure light source: ArF excimer laser (wavelength 193 nm),
Aperture ratio: NA = 0.6, Coherence factor: σ =
0.30. Mask bias is +20 nm, +60
2 (that is, the opening diameter is 150 nm and 190 nm). In FIG. 3, the mask bias was +2 in FIG.
The case of 0 nm is indicated by a black line, and the case of +60 nm is indicated by a white line. Then, for the photomask of the present invention, C
r width is set to 155 nm in increments of 20 nm from 15 nm.
Were determined for each case. FIG.
In (a), (b) and (c), the left end is HTPSM,
The right end is the data related to the binary mask, and the rest is the data according to the present invention. FIG. 3A shows the MEF of each mask. As shown in FIG. 3A, the values of all the MEFs corresponding to the Cr width of the photomask of the embodiment of the present invention are the same as those of the HTPSM. It is as follows. In addition, when compared to binary masks, Crwidth
Is 155 nm and the mask bias is +60 nm.
The values of EF are almost equal, but all other values are low. Then, under the conditions of this simulation, the photomask of the embodiment of the present invention is Cr wi
By optimizing dth, the MEF is biased +
It can be seen that it can be reduced to a good value of about 2.5 at 20 nm and about 2.0 at bias + 60 nm.

FIG. 3B shows Imax of each mask. As shown in FIG. 3B, Imax corresponding to all values of Cr width of the photomask of the present invention is shown.
Exceeds that of HTPSM. Also, when the Cr width is 15 nm, Imax is slightly lower than that of the binary mask.
Indicates a high value. That is, if the photomask of the embodiment of the present invention is used, Imax higher than that of the conventional mask can be obtained by adjusting the Cr width. Therefore, according to the present invention, the exposure time can be reduced.

FIG. 3C shows the DOF of each mask, where the HTPSM and the Crwidth are 15n.
m, sidelobe transfer occurred with the mask of the present invention having a mask bias of +20 nm. However, the mask bias +2
At 0 nm, when the Cr width is in the range of 35 nm to 155 nm, the photomask of the present invention exhibits a good value higher than the value of about 0.33 μm at the bias of the binary mask + 20 nm. At a mask bias of +60 nm, a value intermediate between the HTPSM and the binary mask was shown. That is, if the photomask of the embodiment of the present invention is used,
By adjusting the Cr width, it is possible to obtain a photomask having a DOF superior to that of the binary mask and having no side lobe transferred.

Next, a mask bias of +20 nm, +40 nm, and +6 is applied to each of three types of masks, that is, a binary mask, an HTPSM, and a photomask according to the embodiment of the present invention.
MEF and DOF when opening width is changed over 0 nm
FIG. 4 shows the results obtained by simulation. The exposure wavelength, the aperture ratio, and the coherence factor were the same as those described above, and the Cr width was 75 nm. The symbol x in the figure is a binary mask, the symbol Ｈ is HTPSM, ●
Marks plot the output results of the photomask of the present invention. FIG. 4 also shows the results (I in the figure) for a mask bias of 0 nm, that is, a 130 nm arc-shaped hole pattern for the photomask of the present invention.
B, C, and D in the figure denote mask biases of +20 nm, +40 nm, and +60 nm for a binary mask, respectively.
This is the output result when multiplied. The DOF is always constant, about 0.33 μm, and the MEF decreases as the bias amount increases. Also, F, G, and H in the figure are HTTPS
M has a mask bias of +20 nm and +40 n, respectively.
m, output result when +60 nm is applied. When the bias was +20 nm, side lobe transfer occurred. Also,
In the figure, I, J, K, and L indicate that the mask bias of the photomask of the present invention is 0 nm, +20 nm, +40 nm, respectively.
This is an output result at +60 nm. As the amount of bias increases, the numerical values of both MEF and DOF decrease.

FIG. 4 shows that the mask bias is 0 nm, +2
The MEFs of the photomasks of the present invention of 0 nm, +40 nm, and +60 nm are about 2.8, 2.6, 2.3, 2..
It is one. This indicates that the mask bias is smaller than the MEFs of the binary mask and the HTPSM having the same size. That is, if the photomask of the present invention is used under the conditions of this simulation, lithography having higher dimensional accuracy than the conventional two types of masks is possible. For DOF, the mask bias is 0.
The values of the photomask of the present invention of nm, +20 nm, +40 nm, +60 nm are about 1.1, 0.7, 0.5, 0.4.
It is. It has properties that are much greater than any of the corresponding values, as compared to binary masks. In the HTPSM, side lobe transfer occurs when the mask bias is +20 nm, but in the photomask of the present invention, a good DOF is obtained under the same bias condition. This is because, in the photomask of the present invention, Cr formed on the halftone film has an effect of reducing the light intensity of the side lobe. And the bias is +40
In the case of nm and +60 nm, the DOF of the photomask of the present invention shows a lower value than that of HTPSM. In FIG. 4, a point particularly noticed is that any of the points B to H shown in the conventional example is lower right (from the broken line connecting the points I, J, K, and L) according to the present invention (indicated by “better” in the figure). (The direction indicated by the white arrow). This indicates that the photomask of the present invention has an advantage over the conventional example in a comprehensive manner.

In the photomask of the present invention, since the DOF and the MEF are in a trade-off relationship with respect to the mask bias amount, these two conditions cannot be satisfied at the same time. However, when designing, the Cr width and the mask bias amount are adjusted. By doing, the optimal MEF, DO
What is necessary is just to find F. In other words, when designing a mask in which DOF is to be emphasized, the mask bias should be reduced and FIG.
In a mask in which importance is placed on the MEF in the direction of the right arrow on the paper, the mask bias may be increased so as to proceed in the downward direction on the paper. In addition, as described above, the adjustment of the MEF and the DOF is not limited to the bias amount but also to the Cr width.
Can also be adjusted.

The simulation results and effects when the photomask of one embodiment of the present invention is applied to the isolated hole pattern have been described above. For dense holes, the Cr width is the same as in the isolated hole pattern.
MEF reduction by optimizing th and mask bias amount,
It was confirmed by simulation that the effect of DOF expansion can be obtained. In this simulation, the standard pitch was determined to be three times or less the size of the hole to be formed.

The photomask of the above-described embodiment shown in FIG. 1 has a halftone film formed on a flat transparent substrate and a light-shielding film formed thereon. It is not limited to such a form.
5 to 7 are a cross-sectional view and a plan view showing the structure of a photomask according to another embodiment of the present invention. In the photomask shown in FIG. 5, a dug portion 5 having a pattern opposite to that of the halftone film 3 is formed on the transparent substrate 1. In this mask, the light transmitted through the dug portion 5 and the halftone film 3
The depth of the dug portion 5 and the thickness of the halftone film 3 are selected so that a phase difference of 180 degrees is generated between the transmitted light and the halftone film 3. In the photomask shown in FIG. 6, the opening 4 side of the halftone film 3 is removed by a fixed width.
The light shielding film 2 is formed in the removed portion. In the photomask shown in FIG. 7, the transparent substrate 1 in the region where the light shielding film 2 is formed is dug by the thickness of the light shielding film 2, and the light shielding film 2 is embedded in the dug portion. .

FIG. 8 is a plan view showing an embodiment of the present invention in which dense holes and isolated holes coexist. FIGS. 9A and 9B are AA 'in FIG. Line and B
It is sectional drawing in the -B 'line. In the present embodiment, the width Cr width of the light shielding film 2 is set to 50 to 150 nm. Therefore, in the isolated pattern, the light shielding film 2 is also an isolated pattern, but in the dense hole pattern portion, the light shielding film 2 is a continuous pattern. Further, the light transmittance of the halftone film 3 that generates a phase difference of 180 degrees is set to 3 to 10%.

In the case of the isolated hole, when the outer periphery of each opening of the photomask is shielded from light, the optimum exposure amount at which the resist pattern dimension is as intended is smaller than that of the conventional complete light-shielding mask [ FIG. 3 (b)]. However, when the pitch between adjacent holes is narrow, shielding the outer periphery is the same as in the case of conventional complete light shielding, and the optimal exposure amount is also substantially equal to that of the conventional complete light shielding mask. Therefore, by shielding the outer periphery of each opening of the photomask from light, the size of the narrow pitch portion (when the exposure amount is adjusted to a wide pitch) is reduced. However, as described with reference to FIGS. 15 and 16, when exposure is performed by a large σ normal illumination having a coherence factor σ of 0.6 or more, the size of the dense hole tends to be larger than that of the isolated hole. is there. Therefore, by using the large σ normal illumination of 0.6 or more, it is possible to reduce the dimensional deviation between the narrow pitch portion and the wide pitch portion without correcting the mask size of each hole pattern.

FIG. 10 is a graph showing the pitch dependence of the dimension that occurs when exposure is performed using the photomask of the present invention with Cr width = 100 nm when the pitches in the X and Y directions are equal. In FIG. 10, the horizontal axis represents the pattern pitch, and the vertical axis represents the dimension of the resist pattern actually formed with respect to the mask dimension of 180 nm. Also,
FIG. 11 is a graph showing the Y-direction pitch dependence of the dimension (the geometric mean of the X-direction dimension and the Y-direction dimension, but the difference between them is less than 20 nm) when the pitch in the X-direction is fixed at 340 nm. 10 and FIG. 11 also show the pitch dependence of the dimensions in the conventional example for reference.) The exposure conditions were as follows: light source: KrF excimer laser, NA:
0.68, σ: 0.70 Normal illumination, no mask bias. In each case, the pitch dependence of the dimensions is significantly reduced. In particular, even when the X pitch is fixed, the difference between the dimension in the X direction and the dimension in the Y direction is less than 20 nm, and no distortion of the hole pattern is observed.

The preferred Cr width is 50 to 150 nm, more preferably 100 to 150 nm.
nm. According to the present invention, pitch dependency can be suppressed without employing a mask bias, but a design considering a mask bias may be performed as necessary. 8 and 9 may be applied to the mask structure shown in FIGS.

[0027]

As described above, the photomask of the present invention is formed by forming a light-shielding film smaller than the width of the halftone film around the opening end of the phase shift type halftone mask. Reduction of effects on resist,
It is possible to shorten the exposure time and avoid the transfer of the side lobe. In addition, the photomask having the above structure can obtain a wider depth of focus as compared with the conventional photomask by appropriately setting the mask bias, so that the defocus can be reduced as compared with the case where the conventional mask is used. It can be kept small. Further, when the dense pattern and the isolated pattern are mixed, the light shielding film width is set to 50 to 15
By performing the large σ normal illumination exposure at 0 nm, it is possible to suppress the pitch dependence of the pattern dimension in addition to the above effects.

[Brief description of the drawings]

FIG. 1 is a cross-sectional view and a plan view illustrating a structure of a photomask according to an embodiment of the present invention.

FIG. 2 is a cross-sectional view for explaining that the photomask of the present invention has an effect of reducing MEF.

FIG. 3 HTPSM for an isolated hole pattern,
FIG. 7 is a diagram showing simulation results of MEF, Imax, and DOF of three types of masks of the present invention, a photomask and a binary mask.

FIG. 4 shows the HTPS when the mask bias is changed.
M, ME of binary mask and photomask of the present invention
7 is a graph showing changes in F and DOF.

FIG. 5 is a cross-sectional view and a plan view illustrating a structure of a photomask according to another embodiment of the present invention.

FIG. 6 is a cross-sectional view and a plan view showing a structure of a photomask according to still another embodiment of the present invention.

FIG. 7 is a cross-sectional view and a plan view showing a structure of a photomask according to still another embodiment of the present invention.

FIG. 8 is a plan view showing a structure of a photomask according to still another embodiment of the present invention.

FIG. 9 is a sectional view taken along lines AA ′ and BB ′ in FIG. 8;

FIG. 10 is a graph showing pitch dependency of a pattern dimension for explaining an effect of the present invention.

FIG. 11 is a graph showing pitch dependence of pattern dimensions for explaining the effect of the present invention.

FIGS. 12A and 12B are a cross-sectional view and a plan view illustrating a structure of a normal photomask (binary mask) in which light is completely shielded except for a conventional transmission portion.

FIG. 13 is a cross-sectional view and a plan view showing the structure of a conventional phase shift halftone mask (HTPSM).

FIG. 14 is a view for explaining side lobes generated in a conventional phase shift type halftone mask.

FIG. 15 is a graph showing pitch dependency of a pattern dimension for explaining a problem of the conventional example.

FIG. 16 is a graph showing pitch dependency of pattern dimensions for explaining a problem of the conventional example.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Transparent substrate 2 Light shielding film 3 Halftone film 4 Opening 5 Engraved part

Claims (12)

[Claims] A semi-transparent film that shifts the phase of the transmitted exposure light is formed on a transparent substrate that transmits the exposure light;
In a photomask configured such that a phase difference of 180 degrees occurs between light transmitted through the translucent film and light transmitted only through the transparent substrate, the entire periphery of the opening of the translucent film is A photomask having a light-shielding film formed over the same. 2. A translucent film that shifts the phase of the transmitted exposure light is formed on a transparent substrate that transmits the exposure light,
In a photomask configured to generate a phase difference of 180 degrees between light transmitted through the translucent film and light transmitted only through the transparent substrate, a semitransparent film around an opening of the translucent film Is a photomask which is replaced by a light shielding film over the entire periphery of the opening. 3. The light-shielding film has a width of 50 nm or more and 150 nm or more.
3. The photomask according to claim 1, wherein the photomask is not more than nm. 4. The photomask according to claim 1, wherein the translucent film has a light transmittance of 3 to 10%. 5. The photomask according to claim 1, wherein the light-shielding film has a constant width from an end of the opening over the entire periphery of the opening. 6. A mask bias is applied to at least a part of the pattern of the photomask, and a width or a diameter of the opening is a reciprocal multiple of a reduction ratio of an exposure apparatus than a pattern formed by photolithography. The photomask according to any one of claims 1 to 5, wherein the photomask is formed to be large as described above. 7. The photomask according to claim 1, wherein a concave portion is formed in an opening forming region of the translucent film of the transparent substrate. 8. The transparent substrate is made of quartz, CaF ₂ , MgF
_2. The method according to claim 1, wherein the first member is formed of any one of the first and _second members.
8. The photomask according to any one of items 1 to 7. 9. The translucent film is made of CrON, CrO, Mo.
The photomask according to any one of claims 1 to 8, wherein the photomask is formed of one of SiON, MoSiO, aC: H, and SiN. 10. The photomask according to claim 1, wherein said light-shielding film is formed of Cr or CrO. 11. An exposure method using the photomask according to claim 1, wherein an ordinary illumination light is used. 12. The exposure method according to claim 11, wherein the photomask includes a pattern having a random pitch, and uses ordinary illumination light having a coherence factor σ of 0.6 or more.