JP3967359B2 - Pattern formation method - Google Patents

Description

  The present invention relates to a pattern forming method using a photomask for forming a fine pattern used for manufacturing a semiconductor integrated circuit device.

  In recent years, miniaturization of circuit patterns has become more and more necessary for high integration of large-scale integrated circuit devices (hereinafter referred to as LSIs) realized using semiconductors. As a result, it has become very important to make the wiring pattern constituting the circuit thinner.

  Hereinafter, the thinning of the wiring pattern by the conventional light exposure system will be described by taking as an example the case of using a positive resist process. Here, the line pattern is a portion of the resist film that is not exposed to exposure light, that is, a resist portion (resist pattern) remaining after development. The space pattern is a portion exposed by exposure light in the resist film, that is, an opening portion (resist removal pattern) formed by removing the resist by development. In addition, when using a negative resist process instead of a positive resist process, the definitions of the above-described line pattern and space pattern may be interchanged.

  Conventionally, when pattern formation is performed using an optical exposure system, a photomask in which a completely light-shielding pattern made of Cr or the like is made to correspond to a desired pattern on a transparent substrate (transparent substrate) made of quartz or the like is used. I used it. In such a photomask, the region where the Cr pattern exists is a light-shielding portion that does not transmit exposure light of a certain wavelength at all (substantially 0% transmittance), while the region where no Cr pattern exists (opening) Part) is a light-transmitting part having a transmittance (substantially 100%) equivalent to that of the transmissive substrate with respect to the exposure light described above. When exposure is performed using this photomask, the light shielding portion corresponds to the non-photosensitive portion of the resist, and the opening (translucent portion) corresponds to the photosensitive portion of the resist. Therefore, such a photomask, that is, a photomask composed of a light shielding portion and a light transmitting portion with respect to exposure light having a certain wavelength is called a binary mask.

  In general, in an optical exposure system, the contrast of an image (energy intensity distribution generated on an exposed material by exposure) formed when exposure is performed using the binary mask described above is inversely proportional to λ / NA. Here, λ is the wavelength of the exposure light emitted from the light source, and NA is the numerical aperture of the reduction projection optical system (specifically, the projection lens) of the exposure machine. For this reason, the dimension that can be formed as a resist pattern is proportional to λ / NA. Therefore, in order to realize a fine pattern, it is effective to shorten the wavelength λ of the exposure light and to increase the NA so as to increase the numerical aperture NA.

On the other hand, the image formed by the above-described optical exposure system may deviate from the ideal focus because of a step caused by the formation of elements constituting the LSI, or because the substrate surface is not flat. For this reason, the dimension of the pattern formed in the defocused state must also be kept within a predetermined range. The limit of the defocus value at which the dimension of the pattern is maintained within a predetermined range, that is, the limit of the defocus value at which the dimensional accuracy of the pattern is guaranteed is called the depth of focus (DOF). That is, in order to make the pattern finer, it is necessary to enhance the contrast of the image and improve the DOF value. However, since the DOF is proportional to λ / NA ² , the DOF value decreases when the wavelength is shortened and the NA is increased in order to improve the contrast.

  As described above, a technique capable of simultaneously improving the contrast by a method other than shortening the wavelength and increasing the NA and improving the DOF without changing the wavelength λ and the numerical aperture NA has become important. Yes.

  Among the methods for greatly improving contrast and DOF, there is a method of performing oblique incidence exposure on a periodic pattern on a photomask. However, depending on the oblique incidence exposure, a great effect can be obtained only in the case of a pattern arranged with a short period of λ / NA or less, so that it is not an effective method for miniaturization of an arbitrary pattern. There is a method using an auxiliary pattern (hereinafter referred to as an auxiliary pattern method) as a method for compensating for the disadvantage of the oblique incidence exposure.

  Hereinafter, the auxiliary pattern method disclosed in Patent Document 1 (hereinafter referred to as a first conventional example) will be described. FIG. 35 shows a plan view of a photomask used in the first conventional example. The photomask shown in FIG. 35 is used in a stepper that performs 1/5 reduction projection exposure. As shown in FIG. 35, a light shielding film 11 made of chromium is provided on the surface of a transparent glass substrate 10 to be a mask substrate. The light shielding film 11 is provided with a first opening 12 which is a main pattern (circuit pattern) transferred by exposure. A pair of second openings 13 which are auxiliary patterns that are not transferred by exposure and improve the transfer accuracy of the main pattern are provided on both sides of the first opening 12 in the light shielding film 11. Here, the width of the first opening 12 is set to 1.5 μm, for example, and the width of each second opening 13 is set to 0.75 μm, for example. Moreover, the distance between the center of the 1st opening part 12 and the center of each 2nd opening part 13 is set, for example to 4.5 micrometers. That is, in the photomask used in the first conventional example, an auxiliary pattern having a size smaller than the size of the circuit pattern is provided so as to be adjacent to the circuit pattern as the main pattern. However, although the DOF is slightly improved by the auxiliary pattern method according to the first conventional example, the same effect as in the case of the original periodic pattern cannot be obtained.

Hereinafter, an auxiliary pattern method (hereinafter referred to as a second conventional example) disclosed in Patent Document 2, which is an improved method of the first conventional example, will be described. FIG. 36 shows a plan view of a photomask used in the second conventional example. As shown in FIG. 36, a main pattern 21 is provided on a transparent glass substrate 20 serving as a mask substrate, and auxiliary patterns 22 are periodically arranged on both sides of the light shielding portion 21 on the glass substrate 20. ing. The main pattern 21 is composed of a laminated film of a lower low-transmittance film and an upper light-shielding film (chrome film). The auxiliary pattern 22 is composed of the remaining low-transmittance film from which the upper light-shielding film of the laminated film is removed. Here, the auxiliary pattern 22 made of a low-transmittance film does not form a non-photosensitive portion of the resist (that is, a resist pattern) during exposure. Accordingly, the DOF can be improved by periodically arranging the auxiliary pattern 22 having a low transmittance with respect to the main pattern 21 and performing oblique incidence exposure.
JP-A-5-165194 JP-A-9-73166

  By the way, by using the phase shifter, the contrast and the DOF can be greatly improved, but this is limited to the following cases. In other words, on the photomask, only when a light transmitting part (opening) and a phase shifter that transmits exposure light with a phase difference of 180 ° with respect to the light transmitting part can be arranged on both sides of the fine line pattern, respectively. It is. Therefore, even if a phase shifter is used, it is not possible to obtain an improvement effect of contrast and DOF over all fine portions of a wiring pattern in a general LSI.

  Further, by using the oblique incidence exposure, an effect of greatly improving the contrast and DOF can be obtained for a complete periodic pattern. However, the above-described effects cannot be obtained over the entire fine portion of the wiring pattern including an isolated pattern in a general LSI. At this time, the use of the auxiliary pattern shows a slight improvement in the DOF or the like (first conventional example), but the effect is slight compared to the complete periodic pattern. Further, by using a low transmittance pattern as an auxiliary pattern, it is possible to improve the degree of freedom of auxiliary pattern arrangement, thereby improving the periodicity in the pattern arrangement (second conventional example). Even in this case, the following problems occur. That is, the substantial effect of the second conventional example is only that the assist pattern can be easily processed by increasing the assist pattern. That is, with respect to the improvement of the contrast and DOF, only the same effect as that of the first conventional example (when a thin auxiliary pattern is used) can be obtained. The reason is that the improvement effect of contrast and DOF is not determined depending on whether the mask pattern composed of the main pattern and the auxiliary pattern is a periodic pattern, but an image (energy intensity distribution) formed by the mask pattern at the time of exposure. This is because it is determined depending on whether the periodicity of) is high.

  In view of the above, an object of the present invention is to provide a photomask, a pattern formation method, and a mask data creation method that can improve contrast and DOF in the formation of a pattern having an arbitrary shape.

  In order to achieve the above object, a first photomask according to the present invention includes a mask pattern formed on a transmissive substrate, and a light transmitting portion on the transmissive substrate where the mask pattern is not formed. It is assumed that the photomask has. Specifically, the mask pattern has a main pattern transferred by exposure and an auxiliary pattern that diffracts exposure light and is not transferred by exposure. The main pattern has a first transmittance that partially transmits the exposure light, and a first semi-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion, and the light transmitting portion as a reference. And a phase shifter that transmits exposure light in the opposite phase. The auxiliary pattern has a second transmittance that partially transmits the exposure light, and includes a second semi-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion.

  In the present application, the same phase means a case where the phase difference is in the range of (−30 + 360 × n) degrees or more and (30 + 360 × n) degrees or less (where n is an integer), and the opposite phase is This means that the phase difference is in the range of 150 + 360 × n) degrees or more and (210 + 360 × n) degrees or less (where n is an integer).

  According to the first photomask, since the main pattern is composed of the semi-light-shielding portion and the phase shifter, a part of the light transmitted through the light-transmitting portion and the semi-light-shielding portion can be canceled by the light transmitted through the phase shifter. . For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern can be enhanced. In addition, since an auxiliary pattern with low transmittance is provided separately from the main pattern, diffracted light that interferes with light transmitted through the phase shifter of the main pattern is generated by arranging the auxiliary pattern at an appropriate position. be able to. Therefore, the defocus characteristic in the transfer image of the main pattern is improved, and as a result, the DOF characteristic is improved.

  In addition, according to the first photomask, since the auxiliary pattern is a semi-light-shielding portion, the degree of freedom of auxiliary pattern arrangement is improved, thereby improving the periodicity in the pattern arrangement including the main pattern. The characteristics are further improved. Further, since the auxiliary pattern is a semi-light-shielding portion, the auxiliary pattern can be thickened under the condition that the auxiliary pattern is not transferred by exposure, so that the processing becomes easy.

  In the first photomask, the first transmittance is preferably 15% or less.

  In this way, it is possible to prevent resist film loss or optimize resist sensitivity during pattern formation. That is, this effect can be made compatible with the DOF improvement effect and the contrast improvement effect.

  In the first photomask, the second transmittance is preferably 6% or more and 50% or less.

  In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of the auxiliary pattern being too high.

  The second photomask according to the present invention is a photomask having a mask pattern formed on a transmissive substrate and a translucent portion where the mask pattern is not formed on the transmissive substrate. Specifically, the mask pattern has a main pattern transferred by exposure and an auxiliary pattern that diffracts exposure light and is not transferred by exposure. Further, a translucent portion is interposed between the main pattern and the auxiliary pattern, and defined as sin φA = NA × SA when a predetermined oblique incident position is SA (0.4 ≦ SA ≦ 0.8). With respect to the oblique incident angle φA, the center of the auxiliary pattern is arranged at or near the position of M × (λ / (2 × sin φA)) from the center of the main pattern (where λ is the wavelength of the exposure light) M and NA are the reduction magnification and numerical aperture of the reduction projection optical system of the exposure machine).

  According to the second photomask, in addition to the main pattern, the auxiliary pattern is provided at a position separated from the main pattern by M × (λ / (2 × sinφA)) or in the vicinity thereof. For this reason, the diffracted light generated by the auxiliary pattern improves the defocus characteristic in the transferred image of the main pattern, and as a result, improves the DOF characteristic.

  In the second photomask, the main pattern may be composed of a light shielding portion, or may be composed of a phase shifter that transmits the exposure light in the opposite phase with the light transmitting portion as a reference.

  In the second photomask, the main pattern has a transmittance that partially transmits the exposure light, and a semi-light-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion, and a light transmitting portion as a reference. A phase shifter that transmits the exposure light in the opposite phase is preferable.

  In this case, since the main pattern is composed of the semi-light-shielding portion and the phase shifter, a part of the light transmitted through the light-transmitting portion and the semi-light-shielding portion can be canceled by the light transmitted through the phase shifter. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern can be enhanced.

  When the main pattern includes a semi-light-shielding portion and a phase shifter, the phase shifter is preferably disposed so as to be surrounded by the semi-light-shielding portion at the center of the main pattern.

  In this way, the contrast of the light intensity distribution at the center of the light-shielded image corresponding to the main pattern can be enhanced, so that, for example, a fine line pattern can be formed while maintaining good defocus characteristics. The dimension of the portion sandwiched between the phase shifter and the translucent portion in the semi-light-shielding portion is 20 nm or more and (0.3 × λ / NA) × M or less, or four times the wavelength of the exposure light. 1 or more and (0.3 × λ / NA) × M or less (where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and aperture of the reduction projection optical system of the exposure machine) Number). Note that the main pattern may include a light-shielding portion that replaces the semi-light-shielding portion and a phase shifter.

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the phase shifter is preferably arranged so as to be surrounded by the semi-light-shielding part at the peripheral part of the main pattern.

  In this way, the contrast of the light intensity distribution in the vicinity of the main pattern of the image of the light transmitted through the translucent portion can be enhanced, so that, for example, a fine contact pattern can be formed while maintaining good defocus characteristics. .

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the transmissivity of the semi-light-shielding part is preferably 15% or less.

  In this way, it is possible to prevent resist film loss or optimize resist sensitivity during pattern formation. That is, this effect can be made compatible with the DOF improvement effect and the contrast improvement effect.

  In the second photomask, the auxiliary pattern may include a light shielding portion, or may have a transmittance that partially transmits the exposure light and transmits the exposure light in the same phase with respect to the light transmitting portion. You may comprise from the semi-light-shielding part to make it. If the auxiliary pattern is composed of a semi-light-shielding portion, the degree of freedom of auxiliary pattern arrangement is improved, and thereby the periodicity in the pattern arrangement including the main pattern can be improved, so that the DOF characteristics are further improved. Further, if the auxiliary pattern is a semi-light-shielding portion, the auxiliary pattern can be thickened under the condition that the auxiliary pattern is not transferred by exposure, so that the processing becomes easy. In the case where the auxiliary pattern is composed of a semi-light-shielding part, the transmissivity of the semi-light-shielding part is preferably 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of the auxiliary pattern being too high.

  A third photomask according to the present invention is a photomask having a mask pattern formed on a transmissive substrate and a translucent portion on which the mask pattern is not formed on the transmissive substrate. Specifically, the mask pattern has a main pattern transferred by exposure and an auxiliary pattern that diffracts exposure light and is not transferred by exposure. Further, a translucent portion is interposed between the main pattern and the auxiliary pattern, and defined as sin φB = NA × SB when a predetermined oblique incident position is SB (0.4 ≦ SB ≦ 0.8). With respect to the oblique incident angle φB, the center of the auxiliary pattern is arranged at a position away from the center of the main pattern by M × ((λ / (2 × sin φB)) + (λ / (NA + sin φB))) or in the vicinity thereof. (Where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and numerical aperture of the reduction projection optical system of the exposure machine).

  According to the third photomask, in addition to the main pattern, the auxiliary pattern is provided at or near the position away from the main pattern by M × ((λ / (2 × sin φB)) + (λ / (NA + sin φB))). ing. For this reason, the diffracted light generated by the auxiliary pattern improves the defocus characteristic in the transferred image of the main pattern, and as a result, improves the DOF characteristic. Further, the first auxiliary pattern is provided at or near the position M × (λ / (2 × sin φB)) from the main pattern, and the second auxiliary pattern is M × ((λ / (2 × sinφB)) + (λ / (NA + sinφB))) When provided at or near the position, the following effects are obtained. That is, in addition to the first auxiliary pattern functioning as the first-order diffracted light generation pattern, the second auxiliary pattern functions as the second-order diffracted light generation pattern, so that the DOF improvement effect can be further increased.

  In the third photomask, the main pattern may be composed of a light-shielding portion, or may be composed of a phase shifter that transmits exposure light in an opposite phase with the light-transmitting portion as a reference.

  In the third photomask, the main pattern has a transmittance that partially transmits the exposure light, and a semi-light-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion, and a light transmitting portion as a reference. A phase shifter that transmits the exposure light in the opposite phase is preferable.

  In this case, since the main pattern is composed of the semi-light-shielding portion and the phase shifter, a part of the light transmitted through the light-transmitting portion and the semi-light-shielding portion can be canceled by the light transmitted through the phase shifter. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern can be enhanced.

  When the main pattern includes a semi-light-shielding portion and a phase shifter, the phase shifter is preferably disposed so as to be surrounded by the semi-light-shielding portion at the center of the main pattern.

  In this way, the contrast of the light intensity distribution at the center of the light-shielded image corresponding to the main pattern can be enhanced, so that, for example, a fine line pattern can be formed while maintaining good defocus characteristics. The dimension of the portion sandwiched between the phase shifter and the translucent portion in the semi-light-shielding portion is 20 nm or more and (0.3 × λ / NA) × M or less, or four times the wavelength of the exposure light. 1 or more and (0.3 × λ / NA) × M or less (where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and aperture of the reduction projection optical system of the exposure machine) Number). Note that the main pattern may include a light-shielding portion that replaces the semi-light-shielding portion and a phase shifter.

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the phase shifter is preferably arranged so as to be surrounded by the semi-light-shielding part at the peripheral part of the main pattern.

  In this way, the contrast of the light intensity distribution in the vicinity of the main pattern of the image of the light transmitted through the translucent portion can be enhanced, so that, for example, a fine contact pattern can be formed while maintaining good defocus characteristics. .

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the transmissivity of the semi-light-shielding part is preferably 15% or less.

  In this way, it is possible to prevent resist film loss or optimize resist sensitivity during pattern formation. That is, this effect can be made compatible with the DOF improvement effect and the contrast improvement effect.

  In the third photomask, the auxiliary pattern may be formed of a light shielding portion, or has a transmittance that partially transmits the exposure light and transmits the exposure light in the same phase with respect to the light transmission portion. You may comprise from the semi-light-shielding part to make it. If the auxiliary pattern is composed of a semi-light-shielding portion, the degree of freedom of auxiliary pattern arrangement is improved, and thereby the periodicity in the pattern arrangement including the main pattern can be improved, so that the DOF characteristics are further improved. Further, if the auxiliary pattern is a semi-light-shielding portion, the auxiliary pattern can be thickened under the condition that the auxiliary pattern is not transferred by exposure, so that the processing becomes easy. In the case where the auxiliary pattern is composed of a semi-light-shielding part, the transmissivity of the semi-light-shielding part is preferably 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of the auxiliary pattern being too high.

  In the second or third photomask, the auxiliary pattern is M × (λ / (2 × sinφ)) or M × ((λ / (2 × sinφ)) + (λ / (NA + sinφ)) from the phase shifter. ) When provided at a distant position, the oblique incident angle φ is not less than φ1 and not more than φ2, the oblique incident angle φ is (φ1 + φ2) / 2, or the oblique incident angle φ is (ξ + φ2) / 2 (Where φ1 and φ2 are the minimum oblique incidence angle and the maximum oblique incidence angle of the oblique incidence illumination system of the exposure machine, and ξ is sinξ = 0.4 × NA (where NA is a reduced projection of the exposure machine) It is an angle that satisfies the numerical aperture of the optical system).

  A fourth photomask according to the present invention is a photomask having a mask pattern formed on a transmissive substrate and a translucent portion on which the mask pattern is not formed on the transmissive substrate. Specifically, the mask pattern has a main pattern transferred by exposure and an auxiliary pattern that diffracts exposure light and is not transferred by exposure. In addition, the auxiliary pattern includes a first auxiliary pattern that is disposed at or near the position of the distance X from the center of the main pattern and sandwiches the translucent portion between the auxiliary pattern and the first auxiliary pattern as viewed from the main pattern. The second auxiliary pattern is arranged at a position of the distance Y from the center of the first auxiliary pattern in the outer side of the first auxiliary pattern or in the vicinity thereof, and sandwiches the translucent portion with the first auxiliary pattern. Here, X is longer than Y.

  In the present application, the distance between the main pattern and the auxiliary pattern means the distance between the centers. For example, when an auxiliary pattern having a similar shape is provided in parallel to the line-shaped main pattern, it means a distance between center lines of the main pattern and the auxiliary pattern.

  According to the fourth photomask, in addition to the main pattern, the first auxiliary pattern is provided at the position of the distance X from the main pattern or in the vicinity thereof, and the second auxiliary pattern is separated from the first auxiliary pattern. , At a position of a distance Y shorter than the distance X or in the vicinity thereof. For this reason, the diffracted light generated by each auxiliary pattern improves the defocus characteristic in the transferred image of the main pattern, and as a result, the DOF characteristic improves.

  In the fourth photomask, it is preferable that X / Y = (1 + S) / (2 × S), where S (0.4 ≦ S ≦ 0.8) is a predetermined oblique incident position. In this way, the effect of improving DOF characteristics can be maximized.

  In the fourth photomask, X = M × with respect to the oblique incident angle φA defined by sin φA = NA × SA when the predetermined oblique incident position is SA (0.4 ≦ SA ≦ 0.8). (Λ / (2 × sin φA)) may also be used.

  In the fourth photomask, the main pattern may be composed of a light shielding portion, or may be composed of a phase shifter that transmits exposure light in an opposite phase with respect to the light transmitting portion.

  In the fourth photomask, the main pattern has a transmittance that partially transmits the exposure light, and a semi-light-shielding portion that transmits the exposure light in the same phase with respect to the light-transmitting portion, and a light-transmitting portion as a reference. A phase shifter that transmits the exposure light in the opposite phase is preferable.

  In this case, since the main pattern is composed of the semi-light-shielding portion and the phase shifter, a part of the light transmitted through the light-transmitting portion and the semi-light-shielding portion can be canceled by the light transmitted through the phase shifter. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern can be enhanced.

  When the main pattern includes a semi-light-shielding portion and a phase shifter, the phase shifter is preferably disposed so as to be surrounded by the semi-light-shielding portion at the center of the main pattern.

  In this way, the contrast of the light intensity distribution at the center of the light-shielded image corresponding to the main pattern can be enhanced, so that, for example, a fine line pattern can be formed while maintaining good defocus characteristics. The dimension of the portion sandwiched between the phase shifter and the translucent portion in the semi-light-shielding portion is 20 nm or more and (0.3 × λ / NA) × M or less, or four times the wavelength of the exposure light. 1 or more and (0.3 × λ / NA) × M or less (where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and aperture of the reduction projection optical system of the exposure machine) Number). Note that the main pattern may include a light-shielding portion that replaces the semi-light-shielding portion and a phase shifter.

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the phase shifter is preferably arranged so as to be surrounded by the semi-light-shielding part at the peripheral part of the main pattern.

  In this way, the contrast of the light intensity distribution in the vicinity of the main pattern of the image of the light transmitted through the translucent portion can be enhanced, so that, for example, a fine contact pattern can be formed while maintaining good defocus characteristics. .

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the transmissivity of the semi-light-shielding part is preferably 15% or less.

  In this way, it is possible to prevent resist film loss or optimize resist sensitivity during pattern formation. That is, this effect can be made compatible with the DOF improvement effect and the contrast improvement effect.

  In the fourth photomask, the first and second auxiliary patterns may be composed of a light shielding portion, or have a transmittance that partially transmits exposure light and is exposed with the light transmitting portion as a reference. You may be comprised from the semi-light-shielding part which permeate | transmits light with the same phase. If each auxiliary pattern is composed of a semi-light-shielding portion, the degree of freedom of auxiliary pattern arrangement is improved, and thereby the periodicity in the pattern arrangement including the main pattern can be improved, so that the DOF characteristics are further improved. . Further, if each auxiliary pattern is a semi-light-shielding portion, each auxiliary pattern can be thickened under the condition that it is not transferred by exposure, and thus the processing becomes easy. When each auxiliary pattern is composed of a semi-light-shielding part, the transmissivity of the semi-light-shielding part is preferably 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of each auxiliary pattern being too high.

  A fifth photomask according to the present invention is a photomask having a mask pattern formed on a transmissive substrate and a translucent portion on which the mask pattern is not formed on the transmissive substrate. Specifically, the mask pattern has a main pattern transferred by exposure and an auxiliary pattern that diffracts exposure light and is not transferred by exposure. The auxiliary pattern includes a first auxiliary pattern having a width D1 provided so as to sandwich the light transmitting portion between the auxiliary pattern and the first auxiliary pattern in the outer direction of the first auxiliary pattern as viewed from the main pattern. And a second auxiliary pattern having a width D2 provided so as to sandwich the light transmitting portion between the pattern and the pattern. Here, D2 is larger than D1.

  According to the fifth photomask, since the first and second auxiliary patterns are provided separately from the main pattern, the diffracted light generated by each auxiliary pattern improves the defocus characteristic in the transferred image of the main pattern. As a result, the DOF characteristics are improved. Further, since the width D2 of the second auxiliary pattern far from the main pattern is larger than the width D1 of the first auxiliary pattern close to the main pattern, the above-mentioned effect of improving the DOF characteristic can be achieved while keeping the exposure margin high. can get.

  In the fifth photomask, D2 / D1 is preferably 1.2 or more and 2 or less. In this way, the above-described effect of improving the DOF characteristics can be obtained while preventing the non-photosensitive portion of the resist from being formed by the auxiliary pattern.

  In the fifth photomask, the main pattern may be composed of a light shielding portion, or may be composed of a phase shifter that transmits exposure light in an opposite phase with the light transmitting portion as a reference.

  In the fifth photomask, the main pattern has a transmittance that partially transmits the exposure light, and a semi-light-shielding portion that transmits the exposure light in the same phase with respect to the light-transmitting portion, and a light-transmitting portion as a reference. A phase shifter that transmits the exposure light in the opposite phase is preferable.

  In this case, since the main pattern is composed of the semi-light-shielding portion and the phase shifter, a part of the light transmitted through the light-transmitting portion and the semi-light-shielding portion can be canceled by the light transmitted through the phase shifter. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern can be enhanced.

  When the main pattern includes a semi-light-shielding portion and a phase shifter, the phase shifter is preferably disposed so as to be surrounded by the semi-light-shielding portion at the center of the main pattern.

  In this way, the contrast of the light intensity distribution at the center of the light-shielded image corresponding to the main pattern can be enhanced, so that, for example, a fine line pattern can be formed while maintaining good defocus characteristics. The dimension of the portion sandwiched between the phase shifter and the translucent portion in the semi-light-shielding portion is 20 nm or more and (0.3 × λ / NA) × M or less, or four times the wavelength of the exposure light. 1 or more and (0.3 × λ / NA) × M or less (where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and aperture of the reduction projection optical system of the exposure machine) Number). Note that the main pattern may include a light-shielding portion that replaces the semi-light-shielding portion and a phase shifter.

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the phase shifter is preferably arranged so as to be surrounded by the semi-light-shielding part at the peripheral part of the main pattern.

  In this way, the contrast of the light intensity distribution in the vicinity of the main pattern of the image of the light transmitted through the translucent portion can be enhanced, so that, for example, a fine contact pattern can be formed while maintaining good defocus characteristics. .

  When the main pattern is composed of a semi-light-shielding part and a phase shifter, the transmissivity of the semi-light-shielding part is preferably 15% or less.

  In this way, it is possible to prevent resist film loss or optimize resist sensitivity during pattern formation. That is, this effect can be made compatible with the DOF improvement effect and the contrast improvement effect.

  In the fifth photomask, the first and second auxiliary patterns may be composed of a light shielding portion, or have a transmittance that partially transmits the exposure light and is exposed on the basis of the light transmitting portion. You may be comprised from the semi-light-shielding part which permeate | transmits light with the same phase. If each auxiliary pattern is composed of a semi-light-shielding portion, the degree of freedom of auxiliary pattern arrangement is improved, and thereby the periodicity in the pattern arrangement including the main pattern can be improved, so that the DOF characteristics are further improved. . Further, if each auxiliary pattern is a semi-light-shielding portion, each auxiliary pattern can be thickened under the condition that it is not transferred by exposure, and thus the processing becomes easy. When each auxiliary pattern is composed of a semi-light-shielding part, the transmissivity of the semi-light-shielding part is preferably 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of each auxiliary pattern being too high.

  In the first to fifth photomasks, the phase shifter is preferably formed by digging a transmissive substrate. In this way, a very excellent defocus characteristic is exhibited in pattern formation.

  In the first to fifth photomasks, the semi-light-shielding portion is preferably made of a metal thin film formed on a transmissive substrate. In this way, since the semi-light-shielding portion can be easily formed, the photomask can be easily processed.

  A first pattern formation method according to the present invention is a pattern formation using a photomask having a mask pattern formed on a transparent substrate and a light transmitting portion on the transparent substrate where the mask pattern is not formed. A method comprising: forming a resist film on a substrate (a); irradiating the resist film with exposure light through the photomask (b); and the resist film irradiated with the exposure light And developing a resist pattern to form a resist pattern, wherein the mask pattern includes a main pattern transferred by exposure and an auxiliary pattern that diffracts the exposure light and is not transferred by exposure. The main pattern has a first transmittance that partially transmits the exposure light, and the exposure light is in phase with respect to the light transmitting portion. A first semi-light-shielding portion to be passed, and a phase shifter that transmits the exposure light in an opposite phase with the light-transmitting portion as a reference, and the auxiliary pattern is a first portion that partially transmits the exposure light. 2 having a transmittance of 2 and transmitting the exposure light in the same phase with respect to the light transmitting portion as a reference, the pattern width of the auxiliary pattern being the pattern width of the main pattern Smaller than.

  Further, the second pattern forming method according to the present invention uses a photomask having a mask pattern formed on a transmissive substrate and a translucent portion on the transmissive substrate where the mask pattern is not formed. A pattern forming method comprising: a step (a) of forming a resist film on a substrate; a step (b) of irradiating the resist film with exposure light through the photomask; and the step of irradiating the exposure light And (c) forming a resist pattern by developing a resist film, wherein in the photomask, the mask pattern diffracts the exposure light and is not transferred by the exposure. The main pattern has a first transmittance that partially transmits the exposure light, and the exposure light is the same with respect to the light transmitting portion. A first semi-light-shielding portion that is transmitted in phase and a phase shifter that transmits the exposure light in an opposite phase with respect to the light-transmitting portion, and the auxiliary pattern is disposed between the light-transmitting portion and the main pattern. The translucent portion is sandwiched between the first auxiliary pattern having a width D1 provided so as to sandwich the portion and the first auxiliary pattern in the outer direction of the first auxiliary pattern when viewed from the main pattern. And a second auxiliary pattern having a width D2 provided in such a manner that D2 is larger than D1.

According to each pattern forming method of the present invention, the same effect as each photomask of the present invention can be obtained. Moreover, in each pattern formation method of this invention, it is preferable to use an oblique incidence illumination method in the process of irradiating exposure light. In this way, in the light intensity distribution of the light transmitted through the photomask, the contrast between the main pattern and the portion corresponding to the light transmitting portion is improved. Also, the focus characteristic of the light intensity distribution is improved. Therefore, the exposure margin and focus margin in pattern formation are improved. In other words, it is possible to form a fine pattern with excellent defocus characteristics. Moreover, in each pattern formation method of this invention, it is preferable to irradiate exposure light by annular illumination at the process of irradiating exposure light. In this case, in particular, when the average value of the outer diameter and the inner diameter in the illumination shape used for annular illumination is 0.58 or more and 0.8 or less (however, the values of the outer diameter and the inner diameter are exposure The value normalized by the numerical aperture of the machine is used), and the effect of improving the DOF characteristics by the second or third photomask of the present invention is surely obtained. Furthermore, in each pattern formation method of this invention, it is preferable to irradiate exposure light by quadrupole illumination at the process of irradiating exposure light. In this way, in particular, the distance from the light source center of the center position of each of the four polarized illumination shapes used for quadrupole illumination is 0.4 / (0.5) ^0.5 or more and 0.6 / ( 0.5) When it is ^0.5 or less (however, the values normalized by the numerical aperture of the exposure machine are used as the values of the outer diameter and inner diameter), the improvement of the DOF characteristics by the second or third photomask of the present invention The effect is definitely obtained.

  According to the present invention, since the auxiliary pattern for diffracting the exposure light is provided on the photomask separately from the main pattern, the defocus characteristic in the transferred image of the main pattern is improved by the diffracted light generated by the auxiliary pattern. As a result, the DOF characteristics are improved.

(Prerequisite)
The premise for describing each embodiment of the present invention will be described.

  Usually, since a photomask is used in a reduction projection type exposure apparatus, a reduction ratio must be taken into consideration when discussing a pattern dimension on the mask. However, when describing the following embodiments, in order to avoid confusion, when describing the pattern dimensions on the mask in correspondence with a desired pattern to be formed (for example, a resist pattern), unless otherwise specified, the pattern is reduced. A value obtained by converting the dimension by a magnification is used. Specifically, in a 1 / M reduction projection system, even when a resist pattern having a width of 100 nm is formed by a mask pattern having a width of M × 100 nm, both the mask pattern width and the resist pattern width are assumed to be 100 nm.

  In each embodiment of the present invention, unless otherwise specified, M and NA represent the reduction magnification and numerical aperture of the reduction projection optical system of the exposure machine, λ represents the wavelength of exposure light, and φ represents oblique incidence. The oblique incident angle in exposure is expressed.

(First embodiment)
Hereinafter, a photomask according to a first embodiment of the present invention will be described with reference to the drawings.

  FIG. 1A is a plan view of the photomask according to the first embodiment, and FIG. 1B is an example of a cross-sectional view taken along the line I-I in FIG. c) is another example of a cross-sectional view taken along the line II of FIG.

  As shown in FIGS. 1A and 1B, a line-shaped main pattern 101 transferred by exposure is provided on a transmissive substrate 100. The main pattern 101 includes a first semi-light-shielding portion 101A having a first transmittance that partially transmits exposure light, and a phase shifter 101B. The first semi-light-shielding portion 101A is formed so as to surround the line-shaped phase shifter 101B. In other words, the phase shifter 101B is disposed at the center of the main pattern 101. The phase shifter 101B is formed by digging down the transmissive substrate 100, for example. On both sides of the main pattern 101 on the transmissive substrate 100, a pair of auxiliary patterns 102 that diffract the exposure light and are not transferred by the exposure are provided so as to sandwich the light-transmitting portion between the main pattern 101. The auxiliary pattern 102 is composed of a second semi-shielding portion having a second transmittance that partially transmits the exposure light. Here, the first semi-light-shielding part 101 </ b> A and the second semi-light-shielding part are, for example, a semi-light-shielding film 106 formed on the transmissive substrate 100. The first semi-light-shielding portion 101A and the auxiliary pattern 102 may be light-shielding portions.

  In the photomask shown in FIGS. 1A and 1B, a mask pattern is composed of a main pattern 101 and an auxiliary pattern 102. Further, the portion of the transmissive substrate 100 where the mask pattern is not formed is a light transmitting portion (opening).

  Further, the relationship between the light transmitted through the phase shifter 101B and the light transmitted through the light transmitting portion is opposite in phase (specifically, the phase difference between them is (150 + 360 × n) degrees or more and (210 + 360 × n) degrees or less). (Where n is an integer)).

  Further, the light transmitted through each of the first semi-light-shielding part 101A and the second semi-light-shielding part (auxiliary pattern 102) and the light transmitted through the light-transmissive part have the same phase relationship (specifically, the level of both). The phase difference is not less than (−30 + 360 × n) degrees and not more than (30 + 360 × n) degrees (where n is an integer).

  According to the first embodiment, since the main pattern 101 includes the first semi-light-shielding part (or the first light-shielding part) 101A and the phase shifter 101B, the light transmitted through the phase shifter 101B (with the opposite phase) A portion of the light (light having the same phase) transmitted through the light transmitting portion and the first semi-light-shielding portion 101A can be canceled by the light, thereby realizing a strong light-shielding property. This effect is particularly remarkable when the main pattern 101 becomes a fine pattern and the light shielding property is reduced. On the other hand, since the auxiliary pattern 102 is composed of a semi-light-shielding portion, the light-shielding property is low. Therefore, the light shielding degree of the auxiliary pattern 102 can be weakened while the light shielding degree in the light shielding image corresponding to the main pattern 101 is higher than that of the normal light shielding pattern, so that the contrast of the light intensity distribution is achieved by the mask configuration of this embodiment. Can be emphasized. In addition, since the auxiliary pattern 102 having a low transmittance is provided separately from the main pattern 101, the auxiliary pattern 102 is arranged at an appropriate position to interfere with light transmitted through the phase shifter 101B of the main pattern 101. Diffracted light can be generated. Therefore, the defocus characteristic in the transfer image of the main pattern 101 is improved, and as a result, the DOF characteristic is improved.

  In the following description, it is assumed that the main pattern is composed of a phase shifter and a semi-light-shielding part, but the same effect can be obtained even if the main pattern is composed of a phase shifter and a light-shielding part unless otherwise specified. Shall.

  That is, according to the first embodiment, since the auxiliary pattern 102 is a semi-light-shielding portion having a low light shielding property, it is difficult to transfer the auxiliary pattern 102 by exposure, and the auxiliary pattern can be freely arranged under a constraint condition that the auxiliary pattern 102 is not transferred. The degree is improved. For this reason, since the periodicity in the pattern arrangement including the main pattern 101 can be enhanced, the DOF characteristics are further improved. Further, since the auxiliary pattern 102 is a semi-light-shielding portion, the auxiliary pattern 102 can be thickened under the condition that the auxiliary pattern 102 is not transferred by exposure, so that the photomask can be easily processed and a dimensional error may occur in the auxiliary pattern processing. The influence of the main pattern 101 on the transferred image is also reduced. Furthermore, according to the present embodiment, when the peripheral portion of the main pattern 101 is composed of a semi-light-shielding portion, an effect that the influence of the dimensional error of the main pattern 101 in mask processing on pattern formation is reduced.

  Further, according to the first embodiment, since the phase shifter 101B is arranged at the center of the main pattern 101, the contrast of the light intensity distribution at the center of the light-shielded image corresponding to the main pattern 101 can be enhanced. For example, a fine line pattern can be formed while maintaining good focus characteristics.

  Further, according to the first embodiment, the phase shifter 101B is formed by creating an opening in the semi-light-shielding part (semi-light-shielding film 106) and digging up the transmissive substrate 100 in the opening. For this reason, a highly transmissive phase shifter can be realized. In addition, the intensity of the light having the opposite phase transmitted through the main pattern (that is, the phase shifter 101B) can be adjusted by the opening size in the semi-light-shielding portion. For this reason, the optimization of the light of the opposite phase which permeate | transmits the main pattern 101 is easily realizable, Therefore The very outstanding defocus characteristic is exhibited in pattern formation. That is, the mask size can be adjusted by the width of the semi-light-shielding portion surrounding the phase shifter, and the light intensity of the opposite phase can be adjusted by the opening size of the semi-light-shielding portion. It is possible to obtain a unique effect that it can be adjusted to As a result, a desired pattern dimension can be easily realized while reliably realizing the effect of adjusting the light of the opposite phase, for example, the effect of improving the focus characteristics and the effect of improving the contrast of the fine pattern.

  In the first embodiment, it is preferable that the first transmittance of the first semi-light-shielding portion 101A constituting the main pattern 101 is 15% or less. In this way, at the time of pattern formation, it is possible to prevent a reduction in the thickness of the resist film due to an excessive increase in light transmitted through the semi-light-shielding portion, and it is possible to achieve optimization of resist sensitivity. That is, it is possible to achieve both these effects, the DOF improvement effect, and the contrast improvement effect.

  In the first embodiment, the second transmittance of the auxiliary pattern 102 (that is, the second semi-light-shielding portion) is preferably 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of the auxiliary pattern 102 being too high.

  In the first embodiment, the first semi-light-shielding portion 101A and the second semi-light-shielding portion serving as the auxiliary pattern 102 are the same semi-light-shielding film 106, for example, a metal thin film formed on the transmissive substrate 100. It may be formed from. In this case, each semi-light-shielding portion can be easily formed, so that the photomask can be easily processed. As the metal thin film, a thin film (thickness of about 50 nm or less) made of a metal such as Cr (chromium), Ta (tantalum), Zr (zirconium), Mo (molybdenum), Ti (titanium), or an alloy thereof is used. be able to. Specific alloy materials include Ta-Cr alloy, Zr-Si alloy, Mo-Si alloy, Ti-Si alloy and the like. Further, a thick film containing silicon oxide such as ZrSiO, CrAlO, TaSiO, MoSiO or TiSiO may be used instead of the metal thin film.

  Further, in the first embodiment, even when only the first semi-light-shielding portion 101A of the main pattern 101 is replaced with the light-shielding portion, the contrast improvement effect by the main pattern 101 and the auxiliary pattern 102 can be obtained. Specifically, for example, as shown in FIG. 1C, a mask structure in which a light shielding film 107 is further deposited on the semi-light shielding film 106 constituting the first semi-light shielding part 101A may be used.

  Here, FIGS. 2A to 2D show variations in the structure of the photomask according to the first embodiment. That is, instead of the mask structure shown in FIG. 1B, the phase shift film 108 and the semi-light-shielding film 106 made of a material with high transmittance are sequentially formed on the transparent substrate 100 shown in FIG. In the stacked structure, the phase shift film 108 in the formation region of the phase shifter 101B in the main pattern 101 is removed, and the formation region of the phase shifter 101B and the semi-light-shielding film 106 in the light transmitting portion formation region are removed. It may be used. Also in this case, as shown in FIG. 2B, the light shielding film 107 may be deposited on the semi-light shielding film 106 constituting the first semi-light shielding portion 101 </ b> A of the main pattern 101. In the mask structure shown in FIGS. 2A and 2B, a region where only the phase shift film 108 is formed on the transparent substrate 100 is a light transmitting portion. According to such a structure, since the phase of the phase shifter 101B can be controlled by the film thickness of the phase shift film 108, the phase of the phase shifter 101B can be accurately controlled.

  Further, instead of the mask structure shown in FIG. 1B, a semi-light-shielding film 106 and a phase shift film 108 made of a material having high transmittance are sequentially formed on the transparent substrate 100 shown in FIG. 2C. In the laminated structure, a structure in which the semi-light-shielding film 106 in the light-transmitting portion forming region is removed and the phase shift film 108 in other regions other than the region for forming the phase shifter 101B in the main pattern 101 is removed may be used. Good. In this case, as shown in FIG. 2D, a laminated structure of the phase shift film 108 and the light shielding film 107 is formed on the semi-light shielding film 106 constituting the first semi-light shielding part 101A of the main pattern 101. It may be. According to such a structure, the phase of the phase shifter 101B can be controlled by the film thickness of the phase shift film.

  Next, a mask (hereinafter referred to as a mask enhancer structure) in which a phase shifter (phase shifter 101B) is provided at the center of a linear light shielding pattern (main pattern 101) found by the inventors of the present application is used. A method for enhancing the resolution of the line pattern by enhancing the light shielding property of the pattern (hereinafter referred to as the center line enhancement method) will be described. Hereinafter, a case where a fine line pattern is formed by a positive resist process will be described as an example. However, when a negative resist process is used, the center line emphasis method is similarly established if the minute line pattern (resist pattern) in the positive resist process is replaced with a minute space pattern (resist removal pattern). For simplicity of explanation, it is assumed that the light shielding pattern other than the phase shifter portion is composed of a light shielding portion.

  In the mask enhancer, if the pattern width and the phase shifter width are adjusted so that the intensity of the light that travels from the periphery of the light shielding pattern to the back side of the light shielding pattern and the intensity of the light that passes through the phase shifter are just balanced, the mask enhancer is transmitted. The amplitude intensity of light has a distribution that becomes 0 at a position corresponding to the center of the mask enhancer. At this time, the intensity of the light transmitted through the mask enhancer (the square of the amplitude intensity) also has a distribution that becomes 0 at a position corresponding to the center of the mask enhancer. That is, an image with high contrast is formed by the mask enhancer. Here, even if the light shielding portion is a semi-light shielding portion that transmits light in the same phase as the light transmitting portion (transparent substrate) and has a finite transmittance, the same effect can be obtained. In other words, in consideration of the weakness of the light shielding property, it is a semi-light-shielding portion that is not preferable as a line-shaped mask pattern, but by providing a phase shifter inside the mask pattern, that is, by using a mask enhancer structure, contrast is improved. High image can be formed. In other words, the semi-light-shielding portion can be used for forming a fine pattern.

  As described above, the centerline enhancement method is extremely effective in the situation where it is difficult to form a pattern (completely light-shielding pattern) consisting of only a complete light-shielding portion on the mask due to its principle. That is, the center line enhancement method is effective in a mask pattern width where it is difficult to block light by a complete light blocking pattern due to the diffraction phenomenon, that is, a mask pattern width of 0.8 × λ / NA or less. The center line emphasis method is more effective when the mask pattern width is 0.5 × λ / NA or less where the influence of the above becomes large. Further, the center line emphasis method exhibits an extremely remarkable effect in a mask pattern width of 0.4 × λ / NA or less that makes it very difficult to form a pattern using a complete light-shielding pattern. Therefore, the present embodiment in which the main pattern has a mask enhancer structure exerts its effect particularly when the mask width of the main pattern is as described above, and thus has a very high effect in forming a fine pattern. However, in the present application, the mask pattern width in the mask enhancer structure means the width of the entire outer shape of the light shielding part or semi-light shielding part containing the phase shifter.

  Hereinafter, it will be described that by using a mask enhancer structure for a main pattern with an auxiliary pattern, the influence of dimensional errors in the mask on pattern formation can be reduced.

  When an auxiliary pattern is added in the vicinity of the main pattern, the mask pattern is densely arranged. In general, in a densely arranged mask pattern, the influence of dimensional errors in the mask on pattern formation becomes large. However, if a mask enhancer structure is used for the main pattern, this effect can be reduced.

  FIG. 3A shows a mask pattern including a main pattern 101 having a mask enhancer structure including a phase shifter 101B and a semi-light-shielding portion 101A, and an auxiliary pattern 102 including a semi-light-shielding portion. Specifically, the mask pattern includes a main pattern 101 with a width of 140 nm and an auxiliary pattern 102 with a width of 90 nm having a center at a position 300 nm away from the center of the main pattern 101. In the mask enhancer structure of the main pattern 101, the width of the phase shifter 101B is 70 nm. At this time, as shown in FIG. 3B, the configuration of the auxiliary pattern 102 is the same as the mask pattern of FIG. 3A, and the main pattern 101 is formed of a simple light shielding portion (light shielding pattern) 109. In this case, a light shielding pattern with a width of 180 nm is necessary to realize a light shielding property similar to that of the mask enhancer in FIG.

  3 (c) and 3 (d) show the results of simulating the effect on pattern formation when the main pattern width and auxiliary pattern width of the mask pattern shown in FIGS. 3 (a) and 3 (b) change simultaneously by 10 nm. FIG. Specifically, FIG. 3C shows a simulation result of the light intensity distribution formed at a position corresponding to the IIIA-IIIA line (width direction of the main pattern 101) of FIG. FIG. 3D shows a simulation result of the light intensity distribution formed at a position corresponding to the IIIB-IIIB line (width direction of the main pattern 101) of FIG. 3B during pattern exposure. However, in the simulation, the wavelength λ of the light source was 193 nm (the wavelength of the ArF light source), and the lens numerical aperture NA was 0.6. Further, the transmittance of all the semi-light-shielding portions used for the mask pattern was set to 6%. 3C and 3D, the results when each pattern width (mask dimension) is increased by 10 nm (+10 nm) are shown by a solid line, and when each pattern width is decreased by 10 nm (−10 nm). The result is shown by a dotted line. 3C and 3D, the position corresponding to the center position of the main pattern 101 is set to zero.

  From the simulation results shown in FIGS. 3C and 3D, it can be seen that when the mass enhancer structure is used for the main pattern, the light intensity at the center of the main pattern hardly changes with the increase or decrease of the mask dimension. Further, when the evaluation is performed using the dimension of the pattern to be formed, the main pattern width and the auxiliary pattern width, that is, the mask dimension is increased by 10 nm under the condition that the pattern having a width of 100 nm is formed using the mask configuration of FIG. The width of the pattern to be formed is 106 nm. Similarly, when the mask dimension is reduced by 10 nm, the width of the formed pattern becomes 95 nm. That is, even if the mask dimension is increased or decreased by about 10 nm, the dimension of the pattern to be formed is affected only by about 5 nm.

  On the other hand, if the mask dimension of the main pattern and the auxiliary pattern is increased by 10 nm under the condition that a pattern having a width of 100 nm is formed using the mask configuration of FIG. 3B, the dimension of the formed pattern becomes 116 nm. Similarly, when the mask dimension is reduced by 10 nm, the pattern dimension to be formed is 86 nm. That is, when the mask dimension increases or decreases by 10 nm, the dimension of the pattern to be formed increases or decreases by about 15 nm. In other words, according to the mask configuration of FIG. 3B, it can be seen that the pattern dimension fluctuates more than the mask dimension fluctuation amount, and the influence of the mask dimension error is likely to occur in pattern formation.

  In this way, by using the mask enhancer structure for the main pattern adjacent to the auxiliary pattern, an effect that the influence of the dimensional variation of the mask is less likely to occur in pattern formation, which is not found in the prior art, can be obtained. This effect is particularly noticeable in the case of a mask enhancer structure in which the main pattern is composed of a phase shifter and a semi-light-shielding part, but the same effect is also obtained by a mask enhancer structure in which the main pattern is composed of a phase shifter and a light-shielding part. An effect is obtained.

  In the present embodiment, the auxiliary pattern is not necessarily provided on both sides of the main pattern. Specifically, when another main pattern is close to one side of the main pattern, the auxiliary pattern may be provided only on the side of the main pattern opposite to the side where the other main pattern is close.

(First modification of the first embodiment)
Hereinafter, a photomask according to a first modification of the first embodiment of the present invention will be described with reference to the drawings.

  FIG. 4 is a plan view of a mask pattern in a photomask according to a first modification of the first embodiment. In FIG. 4, the same components as those of the photomask according to the first embodiment shown in FIGS. 1A and 1B are denoted by the same reference numerals, and the description thereof is omitted.

  The first feature of the present modification is that the auxiliary pattern 102 (width: D1) from the center of the phase shifter 101B (width: W) of the main pattern 101 (width: L) with respect to a predetermined oblique incident angle φ. (Λ / (2 × sinφ)).

  The second feature of this modification is that the auxiliary pattern (in other words, a position away from the center of the phase shifter 101B of the main pattern 101 by (λ / (2 × sinφ)) + (λ / (NA + sinφ)). A second auxiliary pattern 103 (width: D2) that diffracts the exposure light and is not transferred by exposure is provided at a position (λ / (NA + sinφ)) away from the center of 102 (hereinafter referred to as the first auxiliary pattern). It is that. Note that a light transmitting portion is interposed between the first auxiliary pattern 102 and the second auxiliary pattern 103. The second auxiliary pattern 103 includes a semi-shielding portion similar to the first auxiliary pattern 102.

  According to this modification, the DOF improvement effect by having arrange | positioned the auxiliary pattern can be implement | achieved reliably.

  In the present modification, either one of the first auxiliary pattern 102 and the second auxiliary pattern 103 may not be arranged.

  In the present modification, the above-described effect is produced to some extent even if the distance between the center of the phase shifter 101B and the center of the first auxiliary pattern 102 is a value in the vicinity of (λ / (2 × sinφ)). .

  In this modification, the distance between the center of the phase shifter 101B and the center of the second auxiliary pattern 103 is a value in the vicinity of (λ / (2 × sin φ)) + (λ / (NA + sin φ)). However, the above-mentioned effects occur to some extent.

  In the present modification, the oblique incident angle φ is preferably 0.4 × NA or more and 0.80 × NA or less, particularly 0.58 × NA or more and 0.7 × NA. It is preferable that it is NA or less. When exposure is performed using annular illumination, the oblique incident angle φ is preferably 0.6 × NA or more and 0.80 × NA or less. When exposure is performed using quadrupole illumination, the oblique incident angle φ is preferably 0.4 × NA or more and 0.60 × NA or less (details are given in the first embodiment). (Refer to a second modification of the embodiment).

  Further, in this modification, the width L of the main pattern 101 is preferably at least 2 × 20 nm (actual dimension on the mask) or more than the width W of the phase shifter 101B, and in particular, 4 of the exposure wavelength (exposure light wavelength). It is preferably at least twice as large as one part. That is, in the mask enhancer structure of the main pattern, the width of the semi-light-shielding part (or light-shielding part) sandwiched between the phase shifter and the light-transmitting part is preferably at least 20 nm (actual dimension on the mask). It is preferable that it is 1/4 or more of the exposure wavelength. However, since it is a photomask using a mask enhancer structure, the width of the main pattern is preferably 0.8 × λ / NA or less. Therefore, a semi-light-shielding portion (sandwiched between a phase shifter and a light-transmitting portion ( Alternatively, it is preferable that the width of the light shielding portion does not exceed 0.4 × λ / NA (for details, refer to the third modification of the first embodiment).

  In this modification, the width D2 of the second auxiliary pattern 103 is preferably larger than the width D1 of the first auxiliary pattern 102, and more preferably, D2 is 1.2 times or more of D1. (For details, see the fourth modification of the first embodiment).

  Hereinafter, the light transmitted through the opening (phase shifter 101B) in the mask enhancer by disposing the diffracted light generation pattern (auxiliary pattern 102) at the specific position with respect to the main pattern 101 having the mask enhancer structure. The reason why the diffracted light that interferes with the light can be generated, thereby improving the defocus characteristic during pattern formation will be described.

  FIG. 5A is a diagram for explaining a diffraction phenomenon that occurs when exposure is performed on a mask in which patterns are periodically arranged. As shown in FIG. 5A, light 141 is irradiated from a light source 140 onto a mask 150 in which a plurality of light-shielding patterns (hereinafter referred to as pitch patterns) 151 are arranged substantially infinitely at a predetermined pitch P. Has been. Here, the substantially infinite periodic arrangement means that the same effect as that obtained by each pitch pattern when an infinite number of pitch patterns are actually arranged is the pitch pattern at the center of the mask. 151 is obtained. In other words, it means that the pitch pattern 151 is arranged so that the distance from the pitch pattern 151 at the center of the mask to the pitch pattern 151 at the edge of the mask is about 4 × λ / NA or more.

  FIG. 5A shows a diffraction phenomenon assuming oblique incidence exposure. That is, the light source 140 is located at a distance S from a normal line (a dashed line in the figure) passing through the center of the lens 152. At this time, the incident angle (oblique incident angle) φ of the light 141 from the light source 140 with respect to the mask 150 is expressed by sin φ = S × NA. Here, S defining the oblique incident angle φ is referred to as an oblique incident position. However, the coordinates of the light source 140 are represented using values normalized by the numerical aperture NA. Further, the diffraction angle θn of the n-th order diffracted light (n is an integer) of the light 141 that has passed through the pitch pattern 151 arranged at the pitch P is expressed as sin θn = n × λ / P. The 0th-order diffracted light 142 of the light 141 incident on the mask 150 at an oblique incident angle φ is a coordinate on the lens 152 (coordinate on a one-dimensional coordinate system with the lens center as the origin. The same applies hereinafter) r0 = −sin φ = A position represented by S × NA is reached. The first-order diffracted light (+ 1st-order diffracted light) 143 of the light 141 reaches a position represented by coordinates r1 = r0 + sin θ1 = r0 + λ / P on the lens 152. In general, the position where the nth-order diffracted light reaches on the lens 152 is represented by coordinates rn = r0 + sin θn = r0 + n × λ / P. However, when the absolute value of rn exceeds NA, the nth-order diffracted light does not become diffracted light that passes through the lens 152, and therefore is not imaged on the wafer.

  By the way, in the case of an ideal lens, the phase change at the time of defocus in the diffracted light that passes through the lens and forms an image on the wafer depends only on the distance (radius) from the lens center of the passing position of the diffracted light in the lens. Determined. When light is incident perpendicular to the lens, the 0th-order diffracted light always passes through the center of the lens, and the 1st-order or higher-order diffracted light passes through a position away from the lens center. Therefore, at the time of defocusing, a phase difference is generated between the 0th-order diffracted light that passes through the lens center and the higher-order diffracted light that passes through a position away from the lens center, so that image blur occurs.

  FIG. 5B shows a case where the mask shown in FIG. 5A is subjected to oblique incidence exposure under the condition that only the 0th-order diffracted light and the 1st-order diffracted light pass through the lens and r0 = −r1. It is a figure explaining the diffraction phenomenon which arises. As shown in FIG. 5B, both the 0th-order diffracted light 142 and the 1st-order diffracted light 143 pass through a position away from the center of the lens 152 by the same distance. For this reason, the phase changes received during defocus coincide with each other between the zero-order diffracted light 142 and the first-order diffracted light 143. That is, there is no phase difference between the two diffracted lights, and therefore image blur due to defocus does not occur. Here, when r1 = r0 + sin θ1 is modified in consideration of the condition of r0 = −r1, −2 × r0 = sin θ1 is obtained. Further, when considering the relationship of sin θ1 = λ / P and r0 = −sin φ, 2 × sin φ = λ / P is obtained. Therefore, by performing oblique incidence exposure at an oblique incident angle φ represented by sin φ = λ / (2 × P), it is possible to form a pattern with excellent defocus characteristics. In other words, for oblique incidence exposure at an oblique incidence angle φ, good defocus characteristics can be obtained when the periodic pattern is provided on the mask at a pitch of P = λ / (2 × sinφ). The reason why the defocus characteristic is improved by the pitch pattern arranged substantially infinitely in the oblique incidence exposure is as described above.

  However, the above-described good defocus characteristics can be obtained only when the pitch P is λ / (2 × sinφ) or a value close thereto. The defocus characteristic improvement effect can be obtained only when the diffracted light passing through the lens is in the condition that the diffracted light is one of the 0th order diffracted light, the + 1st order diffracted light, and the −1st order diffracted light. It is.

  FIG. 5 (c) shows a diffraction phenomenon that occurs when the mask shown in FIG. 5 (a) is subjected to oblique incidence exposure under conditions where 0th-order diffracted light, + 1st-order diffracted light, and −1st-order diffracted light pass through the lens. FIG. As shown in FIG. 5C, the zero-order diffracted light 142 passes through a position represented by coordinates r0 = −sin φ on the lens 152, and the + 1st-order diffracted light 143 is represented by r1 = sin φ on the lens 152. Even when the first-order diffracted light 144 passes through the lens 152, r0-sin θ1, which is the position where the first-order diffracted light 144 passes through the lens 152, cannot be obtained. Here, the condition that the −1st order diffracted light 144 passes outside the lens 152 is expressed by r0−sin θ1 <−NA. Considering r0 = −sinφ and sinθ1 = r1−r0 = 2 × sinφ, the condition that the −1st order diffracted light 144 passes outside the lens 152 is −sinφ−2 × sinφ <−NA, that is, 3 × sinφ. > NA.

  That is, with respect to the oblique incident angle φ, the condition that both the + 1st order diffracted light and the −1st order diffracted light pass through the lens is represented by 3 × sinφ <NA. At an oblique incident angle φ corresponding to this condition, both the + 1st order diffracted light and the −1st order diffracted light pass through the lens, and the distance from the lens center at a position where each of the + 1st order diffracted light and −1st order diffracted light passes through the lens. Therefore, image defocusing occurs due to the phase difference between both diffracted lights during defocusing. Therefore, the lower limit of the oblique incident angle φ that can improve the defocus characteristic is defined by sin φ> NA / 3. In consideration of the fact that 0th-order diffracted light does not pass through the lens when sinφ is larger than NA, the condition of the oblique incident angle φ that can improve the defocus characteristic is expressed by NA> sinφ> NA / 3. Will be.

  Next, a simulation result of DOF characteristics when oblique incidence exposure is performed on the mask shown in FIG. 5A while changing the pitch of the pitch pattern will be described. 6A is a diagram showing a point light source used in the simulation, FIG. 6B is a diagram showing a pitch pattern used in the simulation, and FIG. 6C is a diagram showing a simulation result of DOF characteristics. It is. In the simulation of DOF characteristics, a pair of light sources (point light sources) 140 shown in FIG. 6A, specifically, coordinates in a plane (two-dimensional coordinate system) perpendicular to the normal passing through the center of the lens 152 are used. (Light source coordinates) A pair of point light sources 140 positioned at (−0.8, 0) and (0.8, 0) were used. Here, the point light source 140 is an ArF (wavelength λ = 193 nm) light source. Further, in the simulation of the DOF characteristics, a slanted oblique line having a lens numerical aperture NA of 0.6 and an oblique incident angle φ of sin φ = 0.48 (that is, the distance S in FIG. 5A is 0.8). Exposure was used. Further, the pitch pattern 151 shown in FIG. 6B, that is, each pitch pattern 151 arranged infinitely at the pitch P is a line-shaped light shielding pattern having a width of 100 nm, and the width of the pitch pattern 151 is determined by using the pitch pattern 151. A simulation is performed in the case of forming a pattern of 100 nm. Note that the DOF value shown in FIG. 6C is a defocus range in which a pattern with a width of 100 nm is formed with a dimensional error of 10 nm or less. As shown in FIG. 6C, when the pitch P of the pitch pattern 151 is λ / (2 × sinφ) (that is, about 0.20 μm) or a value close to λ / (2 × sinφ), A high DOF is obtained. However, even if the pitch P is smaller or larger than λ / (2 × sinφ), the value of DOF rapidly decreases.

  Here, the inventor of the present application indicates that the pitch P values at which the DOF is locally improved exist at equal intervals with respect to the change in the pitch P in the pitch dependence of the DOF shown in FIG. Focused on. Further, the inventor of the present application indicates that the value of the pitch P at which the DOF is locally improved is an integer multiple of λ / (NA + sinφ) (that is, about 0.18 μm) in which high-order diffracted light affects the image formation. It was found as explained below.

  In the light diffraction phenomenon, adjacent orders of diffracted light are in the relationship of having opposite phases. FIG. 6D shows the diffraction phenomenon that occurs when the mask shown in FIG. 5A is subjected to oblique incidence exposure under conditions where the 0th-order diffracted light, the first-order diffracted light, and the second-order diffracted light pass through the lens. It is a figure explaining. At this time, the second-order diffracted light 145 has an opposite phase to the first-order diffracted light 143 shown in FIG. When an image formed by the second-order diffracted light 145 having the opposite phase interferes with an image formed by the zero-order diffracted light 142 and the first-order diffracted light 143 in the oblique incidence exposure, a phenomenon occurs in which the intensity of the counterpart image cancels each other. . However, at the time of defocusing, the intensities of both images are reduced, but the phases of the first-order diffracted light 143 and the second-order diffracted light 145 are opposite to each other. That is, the defocus characteristic in the image generated by the interference of the image formed by the second-order diffracted light 145 with the opposite phase to the image formed by the first-order diffracted light 143 is improved, and as a result, the DOF characteristic is improved.

  As shown in FIG. 6D, the condition for producing such an effect of improving the DOF characteristic is a condition for allowing the second-order diffracted light 145 to pass through the lens 152, which is expressed by sin θ2 <NA−r0. Here, considering sin θ2 = 2 × λ / P and r0 = −sinφ, the above-described condition is expressed by 2 × λ / P <NA + sinφ. Therefore, the pitch P at which the lens 152 changes from the state where the first order diffracted light 143 passes to the state where the second order diffracted light 145 passes is expressed by 2 × λ / (sin φ + NA). In general, the condition for the nth-order diffracted light to pass through the lens is expressed as sin θn = n × λ / P <NA−r0 = NA + sinφ. In other words, the n-th order diffracted light can pass through the lens end when the pitch P is n × λ / (sin φ + NA) or more. Therefore, the DOF characteristics are locally improved when the pitch P = n × λ / (sin φ + NA) (n is an integer of 2 or more).

  As described above, in oblique incidence exposure, the light in the opposite phase interferes with each other, so that the defocus characteristic can be improved. Therefore, the inventor of the present application considered that the defocus characteristic can be improved by an operation that causes an action equivalent to increasing the intensity of the second-order diffracted light having an opposite phase to the first-order diffracted light. Specifically, the defocus characteristics can be improved by introducing a zero-order diffracted light component having an opposite phase to the first-order diffracted light into the diffracted light formed by periodically arranging the light-shielding pattern. It becomes possible. This can be realized by using the above-described mask enhancer structure. That is, in the mask enhancer, the order of the diffracted light passing through the lens is changed while adjusting the degree of light shielding by the mask enhancer by controlling the size of the phase shifter (opening) provided in the mask enhancer. This is because the light of the opposite phase can be controlled. The inventor of the present application, when using a mask enhancer, improved DOF which was slightly improved locally when the pitch P = n × λ / (sin φ + NA) (n is an integer of 2 or more) (see FIG. 6C). However, it was expected to improve significantly in proportion to the size of the phase shifter of the mask enhancer.

  Then, next, the inventor of the present application will explain the result of performing a simulation similar to the simulation shown in FIGS. 6A to 6C using a mask enhancer as a pitch pattern. FIG. 7A is a diagram showing a mask enhancer used as a pitch pattern in the simulation, and FIG. 7B is a diagram showing a simulation result of DOF characteristics by the pitch pattern shown in FIG. 7A. The point light source used for the simulation is the same as the point light source shown in FIG. 6A, and the other simulation conditions are the same as those shown in FIGS. 6A to 6C. FIG. 7B also shows the result when the size (opening width) of the phase shifter of the mask enhancer is changed.

  As shown in FIG. 7A, a plurality of mask enhancers 110 are arranged at a predetermined pitch P substantially infinitely. Each mask enhancer 110 includes a semi-light-shielding portion 111 having an outer shape, and a phase shifter 112 provided at the center of the mask enhancer 110 so as to be surrounded by the semi-light-shielding portion 111. Here, the width (pattern width) of the mask enhancer 110 is L, and the width (opening width) of the phase shifter 112 is W.

  Similarly to the DOF characteristic shown in FIG. 6C, in the DOF characteristic shown in FIG. 7B, when the pitch P is n × λ / (sin φ + NA) (n is an integer of 2 or more), the DOF is locally Is increasing. Further, as the ratio of the opening width W to the pattern width L (W / L) increases, the DOF improvement effect becomes very large. That is, as an effect peculiar to the mask enhancer, when oblique incidence exposure is performed on a pattern in which the mask enhancer is periodically arranged, the pitch P is λ / (2 × sinφ) (about 0.20 μm under simulation conditions) or Not only when the value is in the vicinity, but also when the pitch P is a large dimension of n × λ / (sin φ + NA) (approximately (n × 0.18) μm under simulation conditions, where n is an integer of 2 or more). Pattern transfer with excellent DOF characteristics is possible.

  Furthermore, the present inventor replaced three mask enhancers arranged in parallel instead of the mask enhancer arranged substantially infinitely as shown in FIG. 7A as a mask enhancer as a pitch pattern. The results of simulation similar to the simulation shown in FIGS. 6A to 6C will be described. FIG. 8A is a diagram showing a mask enhancer used in the simulation, and FIG. 8B is a diagram showing a simulation result of DOF characteristics by the mask enhancer shown in FIG. 8A. The point light source used for the simulation is the same as the point light source shown in FIG. 6A, and the other simulation conditions are the same as those shown in FIGS. 6A to 6C. That is, oblique oblique exposure with a lens numerical aperture NA of 0.6, a distance S (see FIG. 5A) of 0.8, and an oblique incident angle φ of sin φ = S × NA = 0.48. Was used.

  As shown in FIG. 8A, three mask enhancers 110 are arranged in parallel at equal intervals (inter-pattern distance R). Each mask enhancer 110 includes a semi-light-shielding portion 111 having an outer shape, and a phase shifter 112 provided at the center of the mask enhancer 110 so as to be surrounded by the semi-light-shielding portion 111. Here, assuming that the width (pattern width) of the mask enhancer 110 is L and the width (opening width) of the phase shifter 112 is W, W / L = 0.4.

  FIG. 8B shows DOF characteristics by the mask enhancer 110 located at the center of the three mask enhancers 110 when the inter-pattern distance R is changed. Similar to the DOF characteristics shown in FIG. 7B, the pitch P, that is, the distance between patterns (more precisely, the distance between the phase shifters 112 of each mask enhancer 110) also in the DOF characteristics shown in FIG. 8B. When R is λ / (2 × sinφ) (approximately 0.20 μm under simulation conditions) or a value in the vicinity thereof, the DOF characteristics are locally improved. Further, as shown in FIG. 8B, the inter-pattern distance R is a value obtained by adding a multiple of λ / (sin φ + NA) (about 0.18 μm) to λ / (2 × sin φ) (about 0.20 μm). In some cases, the maximum DOF value appears.

  That is, when oblique incidence exposure is performed on a pattern in which three mask enhancers 110 are arranged, if the inter-pattern distance R is λ / (2 × sin φ) (about 0.20 μm), the maximum DOF value is appear. Furthermore, even if the inter-pattern distance R is λ / (2 × sin φ) + n × λ / (sin φ + NA) (n is a natural number), the DOF maximum value can be obtained. This is because a phase shift (phase change) generated in each of the 0th-order diffracted light and the 1st-order diffracted light due to defocusing when oblique incidence exposure is performed on a pattern composed of mask enhancers arranged infinitely. This phenomenon is considered to be caused by a synchronized diffraction phenomenon and a phenomenon in which high-order diffracted light in the opposite phase interferes with the imaging of the 0th-order diffracted light and the first-order diffracted light. Therefore, it is estimated that the relational expression representing the inter-pattern distance R from which the DOF maximum value is obtained generally holds.

  As described above, when the inventor of the present application uses a pattern having a phase shifter such as the mask enhancer 110 as the main pattern 101 in the photomask of the present embodiment shown in FIG. As described above, by placing a pattern (diffracted light generation pattern) that is not transferred by exposure and generates diffracted light at a predetermined position, the DOF characteristics when the main pattern 101 is transferred by exposure can be greatly improved. I found what I can do. Here, the predetermined position is a position away from the center of the phase shifter 101B of the main pattern 101 (that is, the phase shifter 112 of the mask enhancer 110) by λ / (2 × sinφ) and λ / (2 × sinφ) + n. Xλ / (sin φ + NA) (n is a positive integer).

  FIGS. 9A to 9C are plan views showing mask patterns of the present invention in which diffracted light generation patterns (auxiliary patterns) are arranged so that the DOF characteristics of the main pattern made of the mask enhancer are greatly improved. It is.

  As shown in FIG. 9A, for example, λ / (2 from the center of the phase shifter 112 in the mask enhancer 110 (width: L) in which the phase shifter 112 (width: W) is provided inside the semi-light-shielding portion 111. By placing a first-order diffracted light generation pattern (first auxiliary pattern) 113 (width: D) composed of a semi-light-shielding portion having a dimension that is not transferred by exposure at a position separated by × sinφ), the DOF of the mask enhancer 110 Improved characteristics.

  9B, for example, λ / (2 × sinφ) + λ / (sinφ + NA) from the center of the phase shifter 112 in the mask enhancer 110 in which the phase shifter 112 is provided inside the semi-light-shielding portion 111. The DOF characteristic of the mask enhancer 110 is improved by disposing a second-order diffracted light generation pattern (second auxiliary pattern) 114 (width: D) made of a semi-light-shielding portion having a dimension that is not transferred by exposure at a distant position. To do.

  Furthermore, the inventor of the present application, because the DOF improvement effect is generated by diffracted light, as shown in FIG. 9C, the auxiliary pattern shown in FIG. 9A and the auxiliary pattern shown in FIG. It has been found that the DOF characteristics of the mask enhancer 110 are greatly improved by synthesizing. That is, the first-order diffracted light generation pattern 113 is arranged at a position away from the center of the phase shifter 112 of the mask enhancer 110 by λ / (2 × sinφ), and λ / (2 × sinφ) + λ / (( By arranging the second-order diffracted light generation pattern 114 at a position distant from (sin φ + NA), the effect of improving the DOF characteristic is further increased.

  That is, as compared with the periodic arrangement in which the distance (X) from the main pattern to the first auxiliary pattern and the distance (Y) from the first auxiliary pattern to the second auxiliary pattern are equal, as described above. The non-periodic arrangement in which the distance X from the main pattern to the first auxiliary pattern is longer than the distance Y from the first auxiliary pattern to the second auxiliary pattern is superior in improving the DOF characteristics. Yes. Here, the optimum ratio of X and Y is represented by X / Y = (sin φ + NA) / (2 × sin φ). When this is expressed by an oblique incident position S having a relationship of sin φ = NA × S with respect to the oblique incident angle φ, X / Y = (sin φ + NA) / (2 × sin φ) = (1 + S) / (2 × S) Thus, the value does not depend on NA.

  Although not shown in the figure, a semi-light-shielding portion having a dimension that is not transferred by exposure at a position away from the center of the second auxiliary pattern 114 (in the direction away from the mask enhancer 110) by λ / (sin φ + NA). It is also preferable to arrange a third diffracted light generation pattern (third auxiliary pattern). Similarly, fourth, fifth, sixth,... Diffracted light composed of semi-light-shielding portions having dimensions that are not transferred by exposure at positions separated by multiples of .lamda ./ (sin.phi. + NA) from the center of the third auxiliary pattern. It is also preferable to arrange the generated pattern as an auxiliary pattern.

  Hereinafter, the effect of disposing the diffracted light generation pattern at the predetermined position will be confirmed by simulation.

  First, the inventor of the present application has proved by simulation that a good DOF characteristic can be obtained for the main pattern composed of the mask enhancer by arranging the diffracted light generation pattern at the positions shown in FIGS. The results will be described. FIG. 10A shows a pattern (mask pattern) used in the simulation. Specifically, as shown in FIG. 10A, transfer is performed by exposure to a position X distance away from the center of the phase shifter 112 in the mask enhancer 110 in which the phase shifter 112 is provided inside the semi-light-shielding portion 111. A first-order diffracted light generation pattern (first auxiliary pattern) 113 composed of a semi-light-shielding portion having a dimension that is not used is arranged. In addition, a second-order diffracted light generation pattern (second-order light generation pattern) including a semi-light-shielding portion having a dimension that is not transferred by exposure at a position Y away from the center of the first-order diffracted light generation pattern 113 (in a direction away from the mask enhancer 110) Auxiliary pattern) 114 is arranged. Here, the width of the mask enhancer 110 is L, the width of the phase shifter 112 is W, and the widths of the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114 are D. FIG. 10B is a diagram showing a simulation result of the DOF characteristic by the pattern shown in FIG. Specifically, FIG. 10B is formed corresponding to the mask enhancer 110 when exposure is performed while changing the distance X and the distance Y with respect to the pattern shown in FIG. It is the figure which calculated | required DOF of the pattern by simulation and mapped the result about the distance X and the distance Y. The simulation conditions are L = 100 nm, W = 60 nm, D = 70 nm, wavelength λ = 193 nm of the light source (ArF light source), lens numerical aperture NA = 0.6, sin φ (φ: oblique incident angle) = 0.8 × NA. Further, the transmittance of the semi-light-shielding portion constituting each of the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114 is set to 6%, and the dimension (width) of the pattern formed corresponding to the mask enhancer 110 ) Is 100 nm. In FIG. 10B, the contour line represents DOF, and the point A has a distance X = λ / (2 × sinφ) (about 0.20 μm) and a distance Y = λ / (sinφ + NA) (about 0). .18 μm). As shown in FIG. 10B, it can be seen that at point A, the maximum value of DOF is obtained. That is, it was demonstrated that good DOF characteristics can be obtained with the photomask of this modification shown in FIG.

  Further, in order to prove that the DOF can be maximized by arranging the diffracted light generation pattern at the position shown in FIG. 4, the inventor of the present application uses the pattern shown in FIG. The result of performing the same simulation under different optical conditions will be described.

  FIG. 11A shows a simulation result of DOF characteristics in the case of using optical conditions of lens numerical aperture NA = 0.6 and sin φ = 0.7 × NA (other conditions are the same as those in FIG. 10B). FIG. In FIG. 11A, a point A represents a point of distance X = λ / (2 × sin φ) (about 0.23 μm) and distance Y = λ / (sin φ + NA) (about 0.19 μm). As shown in FIG. 11A, it can be seen that at point A, the maximum value of DOF is obtained.

  FIG. 11B shows the DOF characteristics simulation results when using the optical conditions of lens numerical aperture NA = 0.6 and sin φ = 0.6 × NA (the other conditions are the same as in FIG. 10B). FIG. In FIG. 11B, point A represents a point of distance X = λ / (2 × sin φ) (about 0.268 μm) and distance Y = λ / (sin φ + NA) (about 0.20 μm). As shown in FIG. 11B, it can be seen that at point A, the maximum DOF value is obtained.

  FIG. 11C shows a simulation result of DOF characteristics in the case of using optical conditions of lens numerical aperture NA = 0.7 and sin φ = 0.7 × NA (other conditions are the same as those in FIG. 10B). FIG. In FIG. 11C, point A represents a point of distance X = λ / (2 × sin φ) (about 0.196 μm) and distance Y = λ / (sin φ + NA) (about 0.162 μm). As shown in FIG. 11C, it can be seen that at point A, the maximum value of DOF is obtained.

  In the above description, the case where a point light source is used has been presupposed. However, hereinafter, the results of evaluating the improvement effect of the DOF characteristics by the diffracted light generation pattern by simulation when a light source having an area is used will be described. 12A to 12C are diagrams showing patterns (mask patterns) used in the simulation.

  Specifically, the mask pattern shown in FIG. 12A is composed of only the mask enhancer 110 having a width L. Here, the mask enhancer 110 includes a semi-light-shielding portion 111 having an outer shape, and a phase shifter 112 (width W) provided at the center of the mask enhancer 110 so as to be surrounded by the semi-light-shielding portion 111. . The mask pattern shown in FIG. 12B causes the mask enhancer 110 shown in FIG. 12A to generate diffracted light that is diffracted at an angle θ represented by sin θ = 2 × sin φ (φ: oblique incident angle). A first-order diffracted light generation pattern 113 is added. Here, the first-order diffracted light generation pattern 113 is a semi-light-shielding pattern, and its width is D. The mask pattern shown in FIG. 12C is obtained by adding sin η = 2 × (NA + sin φ) (φ: oblique incidence angle, NA: lens numerical aperture) to the mask enhancer 110 and the first-order diffracted light generation pattern 113 shown in FIG. The second-order diffracted light generation pattern 114 for generating diffracted light diffracted at an angle η represented by () is added. Here, the second-order diffracted light generation pattern 114 is a semi-light-shielding pattern, and its width is D.

  FIG. 12D is a diagram showing a light source used in the simulation, specifically, an annular light source. As shown in FIG. 12D, the outer diameter and inner diameter of the annular light source are 0.8 and 0.6 (normalized by the lens numerical aperture NA), respectively. In this case, with respect to the oblique incident angle φ, there is oblique incident light of 0.6 × NA <sin φ <0.8 × NA. In this case, it is considered that the optimal arrangement positions of the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114 can be determined by the oblique incident angle that is the main component of the light source. That is, in the case of the above light source, it can be considered that the average value of the oblique incident angle φ is the main component of the light source, and therefore, the oblique incident angle φ as the main component is sin φ = NA × (0.6 + 0. 8) /2=0.7×NA. In the following, the explanation of the confirmation result by simulation including this content will be continued.

  FIG. 12E shows a simulation result of the light intensity distribution generated corresponding to each mask enhancer 110 when the mask patterns shown in FIGS. 12A to 12C are exposed under predetermined conditions. FIG. In FIG. 12F, the mask pattern shown in each of FIGS. 12A to 12C is exposed under predetermined conditions, and a pattern having a width of 0.1 μm corresponding to each mask enhancer 110 is formed. It is a figure which shows the result of having simulated the defocus characteristic of CD of the pattern in the case of forming. Here, CD is a critical dimension and represents a finished dimension of a pattern. In the simulation, L = 180 nm, W = 60 nm, D = 90 nm, and the position of the diffracted light generation pattern was determined based on the oblique incident angle φ where sin φ = 0.7 × NA = 0.42. . The wavelength λ of the light source (ArF light source) was 193 nm, and the lens numerical aperture NA was 0.6. Further, the transmissivity of the semi-light-shielding portion constituting each of the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114 is set to 6%. In FIGS. 12 (e) and 12 (f), (a), (b) and (c) showing the types of lines are shown in FIGS. 12 (a), 12 (b) and 12 (c), respectively. This corresponds to the mask pattern shown in FIG.

  As shown in FIG. 12E, a very high contrast image (light intensity distribution) is formed by all the mask patterns shown in FIGS. 12A to 12C, that is, by the mask pattern having the mask enhancer 110. Has been. Further, the portion related to the formation of the 0.1 μm pattern in each light intensity distribution is hardly affected by the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114. At this time, since the critical light intensity in forming a 0.1 μm (100 nm) pattern is about 0.2, the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114 are resolved and not transferred to the resist. I understand that.

  In addition, as shown in FIG. 12F, the defocus characteristic of the mask enhancer 110 is dramatically improved by adding the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114 to the mask pattern. That is, by using the mask enhancer 110 provided with the phase shifter 112 together with the first-order diffracted light generation pattern 113 and the second-order diffracted light generation pattern 114, a pattern with excellent defocus characteristics can be formed.

  Next, a description will be given of the result of the simulation of the present inventor through simulation that the calculation method of the arrangement position of the diffracted light generation pattern that optimizes the defocus characteristic theoretically derived as described above. Specifically, when the position of the diffracted light generation pattern 113 or 114 is changed in the mask pattern shown in FIG. 12B or FIG. The DOF value of a 1 μm line pattern was determined by simulation. Here, DOF is a defocus range in which the width of the line pattern changes from 0.1 μm to 0.09 μm. The simulation conditions are the same as those in FIGS. 12 (e) and 12 (f).

  FIG. 13A is a diagram illustrating a state in which the first-order diffracted light generation pattern 113 is disposed at a distance P1 from the center of the phase shifter 112 in the mask pattern illustrated in FIG. FIG. 13B is a diagram showing the above-described change in DOF when exposure is performed while changing the distance P1 with respect to the mask pattern shown in FIG. As shown in FIG. 13B, when the distance P1 is about 230 nm obtained by the theoretical formula λ / (2 × sinφ), the DOF value is almost peaked.

  FIG. 13C is a diagram showing a state in which the second-order diffracted light generation pattern 114 is arranged at a position away from the center of the first-order diffracted light generation pattern 113 in the mask pattern shown in FIG. is there. Here, the distance between the phase shifter 112 and the first-order diffracted light generation pattern 113 is λ / (2 × sin φ) (about 230 nm). FIG. 13D is a diagram showing the change in the DOF described above when exposure is performed while changing the distance P2 with respect to the mask pattern shown in FIG. 13C. As shown in FIG. 13D, when the distance P2 is about 190 nm obtained by the theoretical formula λ / (sin φ + NA), the value of DOF is almost a peak.

  As described above, the phase shifter or mask enhancer is obliquely exposed by an exposure machine having a wavelength of the light source (a light source that is a distance S away from the normal passing through the center of the lens by a distance S) and a lens numerical aperture of NA. When performing the above, it is possible to optimize the DOF characteristics of the pattern formed corresponding to the phase shifter or the mask enhancer by arranging the diffracted light generation pattern as follows. That is, the first-order diffracted light generation pattern is arranged at a position away from the center of the phase shifter or mask enhancer (in any case, the center of the phase shifter) by λ / (2 × sinφ), and the center of the first-order diffracted light generation pattern The second-order diffracted light generation pattern is arranged at a position away from λ / (sin φ + NA) from the center of the phase shifter, that is, at a position away from the center of the phase shifter (λ / (2 × sin φ)) + (λ / (sin φ + NA)).

  In the above description, the optimum arrangement position of the diffracted light generation pattern has been shown with respect to the main component of the oblique incident angle φ determined from the illumination shape. Next, the allowable range of the arrangement position of the diffracted light generation pattern will be described. . As shown in FIG. 8B, a position where the DOF is lowest (hereinafter referred to as the worst arrangement position) exists between the optimum arrangement positions of the diffracted light generation patterns. Here, when an intermediate position between the optimum arrangement position and the worst arrangement position is defined as a position where an average DOF improvement effect is obtained (hereinafter, referred to as an average arrangement position), a pair of average arrangement positions sandwiching the optimum arrangement position is determined. It is preferable that the diffracted light generation pattern is in between. Alternatively, it is more preferable that the center of the diffracted light generation pattern is located between intermediate positions between the optimum arrangement position and each of the pair of average arrangement positions sandwiching the optimum arrangement position.

  More specifically, the optimum arrangement position of the second-order diffracted light generation pattern is a position away from the center of the phase shifter by λ / (2 × sinφ) + λ / (sinφ + NA). When this position is referred to as a point OP, the worst arrangement position on both sides of the point OP is a position away from the point OP by (λ / (sin φ + NA)) / 2 on both sides. Further, the average arrangement position on both sides of the point OP is a position away from the point OP on both sides by (λ / (sin φ + NA)) / 4. Assuming that the second-order diffracted light generation pattern is preferably contained in a region sandwiched between the pair of average arrangement positions, the distance from the center of the phase shifter is λ / (2 × sinφ) + (λ / (sinφ + NA)) × It is preferable that the second-order diffracted light generation pattern is within the range of (3/4) to λ / (2 × sin φ) + (λ / (sin φ + NA)) × (5/4). When numerical conversion is performed under the same conditions as in FIG. 13C, the second-order diffracted light generation pattern is within the range of 143 to 238 nm from the center of the first-order diffracted light generation pattern. preferable.

  In addition, the intermediate position between the optimum arrangement position of the second-order diffracted light generation pattern and each of the pair of average arrangement positions sandwiching the second-order diffracted light generation pattern is a position separated from the point OP by (λ / (sin φ + NA)) / 8 on both sides. Here, the second-order diffracted light generation pattern is an auxiliary pattern that does not form a resist pattern, and the center of the second-order diffracted light generation pattern is most likely to exist between the pair of intermediate positions that sandwich the optimum arrangement position. preferable. That is, the distance from the center of the phase shifter is λ / (2 × sinφ) + (λ / (sinφ + NA)) × (7/8) to λ / (2 × sinφ) + (λ / (sinφ + NA)) × (9 The center of the second-order diffracted light generation pattern is preferably present in the range of / 8). When numerical conversion is performed under the same conditions as in FIG. 13C, the center of the second-order diffracted light generation pattern exists within the range of 166 to 214 nm from the center of the first-order diffracted light generation pattern. Is preferred.

  Similarly to the second-order diffracted light generation pattern, the first-order diffracted light generation pattern is also sandwiched between a pair of average positions separated by (λ / (sinφ + NA)) / 4 on both sides of the optimum position of the first-order diffracted light generation pattern. It is preferable that the first-order diffracted light generation pattern fits in the region. Alternatively, (λ / (sin φ + NA)) / between the optimum arrangement position of the first-order diffracted light generation pattern and each of the pair of average arrangement positions sandwiching it, that is, both sides of the optimum arrangement position of the first-order diffracted light generation pattern It is preferable that the center of the first-order diffracted light generation pattern is located in a region sandwiched between a pair of intermediate positions that are separated by eight.

  Specifically, the first-order diffracted light generation pattern has a distance from the center of the phase shifter of λ / (2 × sinφ) − (λ / (sinφ + NA)) / 4 to λ / (2 × sinφ) + (λ / It is preferable to be within the range of (sin φ + NA)) / 4. When numerical conversion is performed under the same conditions as in FIG. 13A, the first-order diffracted light generation pattern is preferably within a range of 183 to 278 nm from the center of the phase shifter. Alternatively, the center of the first-order diffracted light generation pattern has a distance from the center of the phase shifter of λ / (2 × sinφ) − (λ / (sinφ + NA)) / 8 to λ / (2 × sinφ) + (λ / ( sin φ + NA)) / 8. When numerical conversion is performed under the same conditions as in FIG. 13A, the center of the first-order diffracted light generation pattern is preferably in the range of 206 to 254 nm from the center of the phase shifter.

  Note that the concept of the allowable range of the arrangement position of the second-order diffracted light generation pattern described above is not limited to the case where the second-order diffracted light generation pattern as shown in FIG. 4 exists, for example. . That is, when there is a third-order diffracted light generation pattern, a fourth-order diffracted light generation pattern, or the like, the third-order diffracted light generation pattern or A fourth-order diffracted light generation pattern or the like is preferably arranged.

(Second modification of the first embodiment)
Hereinafter, as a second modification of the first embodiment of the present invention, a preferable range of the oblique incidence angle with respect to the photomask according to the first embodiment (or the first modification) will be described. So far, the preferred auxiliary pattern arrangement position with respect to the oblique incident angle used for exposure has been described, but in this modification, it is shown that the preferred oblique incident angle actually exists. That is, a photomask in which the auxiliary pattern is arranged at an optimum position with respect to the preferable oblique incident angle is the photomask that is most excellent in pattern formation.

  By the way, the illumination used for actual exposure is not a point light source but has a certain area. Therefore, in the exposure, a plurality of oblique incident angles φ exist simultaneously. Therefore, in considering a preferable oblique incidence angle, oblique incidence illumination is classified into groups such as annular illumination, dipole illumination, and quadrupole illumination. This is because, for example, in annular illumination, all illumination components become oblique incidence components with respect to the mask surface, while there are illumination components that do not substantially become oblique incidence components with respect to each line pattern on the mask. On the other hand, in dipole illumination and quadrupole illumination, there is no illumination component other than the oblique incidence component.

  Specifically, in annular illumination, light incident from a direction perpendicular to the direction in which the line pattern extends (hereinafter referred to as the line direction) becomes oblique incident light with respect to the line pattern, Light incident from a direction parallel to the line direction is substantially perpendicular incident light to the line pattern. Therefore, the illumination in which such a normal incidence component exists can be considered as a group of annular illumination regardless of its shape.

  On the other hand, in dipole illumination polarized in a direction perpendicular to the line direction, substantially perpendicular incident component light does not exist with respect to the line pattern. Therefore, the illumination in which only such a grazing incidence component exists can be considered as a group of dipole illuminations regardless of its shape.

Further, in the quadrupole illumination, as in the dipole illumination, there is no incident component (that is, substantially perpendicular incident light) from a direction parallel to the line direction. Furthermore, in the quadrupole illumination, only oblique incident light that is incident from a diagonal direction with respect to the line direction (a direction that is 45 degrees with respect to the line direction) exists. For example, when light having a maximum oblique incident angle φ _MAX defined by sin φ _MAX = NA with respect to the numerical aperture NA is incident from a diagonal direction of 45 degrees with respect to the line direction, this light is perpendicular to the line direction. When projecting in any direction, it is substantially equivalent to an oblique incident component of sin φ = NA × 0.5 ^0.5 . Because of this particularity, quadrupole illumination has different characteristics from dipole illumination.

  Hereinafter, the dependency of the DOF on the oblique incident angle when the diffracted light generation pattern is arranged as the auxiliary pattern with respect to the main pattern will be described using simulation results. Here, the mask patterns shown in FIGS. 9A to 9C were used for the simulation. In FIG. 9A, only the first auxiliary pattern 113 is arranged. The first auxiliary pattern 113 is composed of a semi-light-shielding portion, and its width is D. In FIG. 9B, only the second auxiliary pattern 114 is arranged. The second auxiliary pattern 114 is also composed of a semi-light-shielding portion, and its width is D. In FIG. 9C, the first auxiliary pattern 113 and the second auxiliary pattern 114 described above are arranged together. Further, the mask enhancer 110 in FIGS. 9A to 9C is composed of a semi-light-shielding portion 111 and a phase shifter 112, the mask width is L, and the phase shifter width is W.

  FIGS. 14A to 14D show simulation results for performing oblique incidence exposure using annular illumination on the mask patterns shown in FIGS. 9A to 9C. FIG. 14A is a diagram showing annular illumination used for the simulation. In FIG. 14A, the inner diameter of the illumination shape is represented by S1, and the outer diameter is represented by S2. In FIG. 14A, an XY coordinate system is also shown. Each of the line patterns (mask enhancer 110, first and second auxiliary patterns 113 and 114) shown in FIGS. 9A to 9C is arranged in parallel to the Y axis of the XY coordinate system. And When simulating the dependency of DOF on the oblique incidence angle, when exposure is performed using the illumination shown in FIG. 14A, S is defined as S = (S1 + S2) / 2, and sin φ = The auxiliary patterns 113 and 114 shown in FIGS. 9A to 9C are arranged based on S × NA, and pattern formation characteristics at that time are simulated. Here, in the simulation, L = 100 nm, W = 60 nm, D = 60 nm, λ = 193 nm, and NA = 0.7. Moreover, the shape of the annular illumination was determined so that S2-S1 = 0.02. Under these conditions, dependence of the DOF on the oblique incident angle φ when a pattern having a width of 80 nm is formed using the mask patterns shown in FIGS. The simulation results of the incident position S = dependency on sinφ / NA) are shown in FIGS. In FIGS. 14B to 14D, for comparison, the results when a light shielding pattern having a width of 100 nm is used as the main pattern are also shown. That is, in FIGS. 14B to 14D, the result when the main pattern has a mask enhancer structure is indicated by a solid line, and the result when the main pattern is a light shielding pattern is indicated by a dotted line.

  In all the results shown in FIGS. 14B to 14D, the DOF is maximized when the illumination condition of S≈0.7 is used. In addition, under the illumination condition where S is 0.58 or more and 0.8 or less, the DOF is improved by using the mask enhancer structure for the main pattern as compared with the case where the light shielding pattern is used for the main pattern. That is, the auxiliary pattern is arranged at a position defined by sin φ and NA with respect to the main pattern composed of the mask enhancer, and the mask pattern of FIGS. 9A to 9C has sin φ = 0.7 × NA. By using the satisfying oblique incident angle φ as the exposure condition, it becomes a mask pattern that realizes the most excellent pattern formation characteristics. Although the optimum illumination conditions are determined as described above, it is also possible to introduce a mask enhancer structure into the main pattern and use illumination conditions where sin φ is 0.6 × NA or more and 0.8 × NA or less. The DOF can be improved compared to the case where the main pattern is composed of a light shielding pattern.

  The case of the above-mentioned simulation conditions will be specifically described as an example as follows. That is, since the optimum illumination condition is sin φ = 0.7 × NA and NA = 0.7, λ / (2 × sin φ) = 0.193 / (2 × from the center of the phase shifter constituting the main pattern. An optimal photomask can be obtained by disposing the first auxiliary pattern (more precisely, the center thereof) at a position separated by 0.7 × 0.7) = 197 nm. Similarly, from the center of the phase shifter constituting the main pattern, λ / (2 × sin φ) + λ / (NA + sin φ) = 0.193 / (2 × 0.7 × 0.7) + 0.193 / (0.7 + 0. An optimal photomask can be obtained by arranging the second auxiliary pattern (more precisely, the center) at a position separated by 7 × 0.7) = 359 nm.

  In addition, a mask enhancer structure is introduced into the main pattern, and the distance from the center of the phase shifter of the main pattern is 0.193 / (2 × 0.8 × 0.7) = 172 nm or more and 0.193 / (2 By arranging the first auxiliary pattern (more precisely, the center thereof) in the range of × 0.58 × 0.7) = 238 nm or less, the DOF improvement effect is produced. Similarly, the distance from the center of the phase shifter of the main pattern is 0.193 / (2 × 0.8 × 0.7) + 0.193 / (0.7 + 0.8 × 0.7) = 325 nm or more and 0 .193 / (2 × 0.58 × 0.7) + 0.193 / (0.7 + 0.58 × 0.7) = 412 nm or less in the second auxiliary pattern (more precisely, the center) Thereby, the improvement effect of DOF arises.

  Next, FIGS. 15A to 15D show the results of a simulation in which oblique incidence exposure is performed on the mask patterns shown in FIGS. 9A to 9C using dipole illumination. FIG. 15A is a diagram illustrating the dipole illumination used for the simulation. In FIG. 15A, an XY coordinate system is also shown. In FIG. 15A, the inner coordinates of the polarized illumination shape are represented by x1, and the outer coordinates are represented by x2. Here, the polarization direction of the dipole illumination is assumed to be a direction perpendicular to the line direction of the mask enhancer 110 and the like. That is, the polarization direction of the dipole illumination is parallel to the X axis of the XY coordinate system, and each line pattern in FIGS. 9A to 9C is parallel to the Y axis of the XY coordinate system. It shall be arranged in. When simulating the dependency of DOF on the oblique incidence angle, when exposure is performed using the illumination shown in FIG. 15A, S is defined as S = (x1 + x2) / 2, and sin φ = The auxiliary patterns 113 and 114 shown in FIGS. 9A to 9C are arranged based on S × NA, and the pattern formation characteristics at that time are simulated. Here, in the simulation, L = 100 nm, W = 60 nm, D = 60 nm, λ = 193 nm, and NA = 0.7. Further, the shape of the dipole illumination was determined so that x2−x1 = 0.02. As in the case of annular illumination shown in FIGS. 14B to 14D, a pattern having a width of 80 nm is formed using the mask patterns shown in FIGS. 9A to 9C under these conditions. FIGS. 15B to 15D show the results of simulating the dependence of the DOF on the oblique incident angle φ (more precisely, the dependence on the oblique incident position S = sin φ / NA) in the case of performing the above. 15 (b) to 15 (d) also show the results when a light-shielding pattern having a width of 100 nm is used as the main pattern for comparison. That is, in FIGS. 15B to 15D, the result when the main pattern has a mask enhancer structure is indicated by a solid line, and the result when the main pattern is a light shielding pattern is indicated by a dotted line.

  In all the results shown in FIGS. 15B to 15D, the DOF is the maximum when the illumination condition of S≈0.58 is used. Also, under the illumination conditions where S is 0.5 or more and 0.7 or less, DOF is improved by using a mask enhancer structure for the main pattern as compared with the case where a light shielding pattern is used for the main pattern. That is, the mask pattern of FIGS. 9A to 9C in which the auxiliary pattern is arranged at the position defined by sin φ and NA with respect to the main pattern composed of the mask enhancer has sin φ = 0.58 × NA. By using the satisfying oblique incident angle φ as the exposure condition, it becomes a mask pattern that realizes the most excellent pattern formation characteristics. Although the optimum illumination conditions are determined as described above, it is also possible to introduce a mask enhancer structure into the main pattern and use an illumination condition in which sinφ is 0.5 × NA or more and 0.7 × NA or less. The DOF can be improved as compared with the case where the main pattern is composed of a light shielding pattern.

  Next, simulation results of performing oblique incidence exposure using quadrupole illumination on the mask patterns shown in FIGS. 9A to 9C are shown in FIGS. FIG. 16A is a diagram illustrating the quadrupole illumination used for the simulation. In FIG. 16A, the XY coordinate system is also shown. Each of the line patterns (mask enhancer 110, first and second auxiliary patterns 113 and 114) shown in FIGS. 9A to 9C is arranged in parallel to the Y axis of the XY coordinate system. And In FIG. 16A, the quadrupole illumination is polarized in a diagonal position with respect to the direction of the line pattern (line direction), and with respect to the line direction (Y-axis direction). The inner coordinate of the polarized illumination shape in the vertical direction (X-axis direction) is represented by x1, and the outer coordinate is represented by x2. Further, when simulating the dependency of the DOF on the oblique incidence angle, when exposure is performed using the illumination shown in FIG. 16A, as in the case of the dipole illumination shown in FIG. S is defined as S = (x1 + x2) / 2, and the auxiliary patterns 113 and 114 shown in FIGS. 9A to 9C are arranged based on sin φ = S × NA. Perform a simulation. Here, in the simulation, L = 100 nm, W = 60 nm, D = 60 nm, λ = 193 nm, and NA = 0.7. Moreover, the shape of the quadrupole illumination was determined so that x2-x1 = 0.02. Similarly to the case of the dipole illumination shown in FIGS. 15B to 15D, under these conditions, a pattern having a width of 80 nm is used using the mask patterns shown in FIGS. 9A to 9C. FIGS. 16B to 16D show the simulation results of the dependence of DOF on the oblique incidence angle φ (more precisely, dependence on the oblique incidence position S = sin φ / NA) when the formation is performed. 16 (b) to 16 (d) also show the results when a light-shielding pattern having a width of 100 nm is used as the main pattern for comparison. That is, in FIGS. 16B to 16D, the result when the main pattern has a mask enhancer structure is indicated by a solid line, and the result when the main pattern is a light shielding pattern is indicated by a dotted line.

In all the results shown in FIGS. 16B to 16D, that is, in the quadrupole illumination, the DOF is maximum when the illumination condition of S≈0.50 is used. That is, in quadrupole illumination, the optimum oblique incident angle φ is determined by sin φ = 0.50 × NA. This is because the projected position on the X-axis in quadrupole illumination is 0.5 ^0.5 times the original oblique incident position, so sin φ = 0.7, which is the optimum condition for annular illumination, is reduced to 0. .5 Corresponds to a value multiplied by ^0.5 . Further, the preferable oblique incident angle φ is defined by 0.4 × NA ≦ sin φ ≦ 0.6 × NA. At this time, the distance from the light source center (the origin of the XY coordinate system) of each of the four polarized illumination areas is 0.4 / (0.5 ^0.5 ) × NA or more and 0.6 / (0.5 ^0.5 ). × NA or less.

  As described above, when annular illumination is used, the preferable oblique incident position S is 0.6 or more and 0.8 or less, and the condition that S = 0.7, that is, sin φ = 0.7 × NA is satisfied. It becomes the optimum value. In addition, when using dipole illumination, the preferable oblique incident position S is 0.5 or more and 0.7 or less, and the condition that S = 0.58, that is, sin φ = 0.58 × NA is the optimum value. . When quadrupole illumination is used, the preferable oblique incident position S is 0.4 or more and 0.6 or less, and the condition that S = 0.5, that is, sin φ = 0.5 × NA is the optimum value. . That is, in the annular illumination and the dipole illumination, there is a common part in the range of the preferred oblique incident position S, and the oblique incident position S is defined by a value of 0.58 or more and 0.7 or less, The mask patterns shown in FIGS. 9A to 9C are mask patterns having excellent DOF characteristics for any illumination belonging to the annular illumination group and the dipole illumination group. Further, not the ideal quadrupole illumination as shown in FIG. 16A, but a modified quadrupole illumination in which the illumination area is distributed in a wide angle direction not limited to 45 degrees with respect to the line direction. Is substantially belonging to the annular illumination group and the dipole illumination group, and is configured to correspond to the preferred oblique incidence angle common to both the annular illumination and the dipole illumination. This photomask is the most preferable photomask for practical use.

  A photomask configured to correspond to an oblique incident position S of 0.4 or more and 0.8 or less is a photomask having excellent DOF characteristics by adapting illumination conditions to the photomask. .

  The most practical illumination is annular illumination. This is because the dipole illumination has almost no effect on the line pattern parallel to the polarization direction of the illumination shape, and the pattern to which the dipole illumination is applied is limited. In addition, quadrupole illumination causes an undesirable phenomenon such that the shape of the pattern is greatly deformed with respect to the mask shape in forming a pattern such as a T-type or an L-type configured by bending one line. It is.

(Third Modification of First Embodiment)
Hereinafter, as a third modification of the first embodiment of the present invention, the mask enhancer structure of the photomask according to the first embodiment (or the first or second modification thereof) has an auxiliary pattern that produces a more preferable effect. The arrangement will be described.

  Up to this point, it has been shown that various effects can be obtained by introducing a mask enhancer structure into the main pattern. That is, in the mask enhancer structure, the pattern formation characteristics such as DOF and contrast can be improved by adjusting the mask width of the main pattern and the width of the phase shifter therein.

  However, even if any one of the structures shown in FIGS. 1B and 1C and FIGS. 2A to 2D is used, the width of the semi-light-shielding part or the light-shielding part surrounding the phase shifter has a certain size. It is preferable. This is because if the semi-light-shielding part or light-shielding part surrounding the phase shifter becomes too fine, it becomes difficult to process this fine part in mask processing, and the pattern may be peeled off in post-processing such as cleaning performed after processing. A problem occurs. Further, when the transmittance of the phase shifter is approximately the same as the transmittance of the light transmitting part, it is impossible to distinguish the phase shifter from the light transmitting part by mask inspection using light transmittance. On the other hand, if there is a semi-light-shielding part or a light-shielding part with a low transmittance that can be recognized by the mask inspection apparatus at the boundary between the phase shifter and the light-transmitting part, the photomask can be easily inspected.

  From the viewpoint of mask processing, in the mask enhancer structure, the width of the semi-light-shielding part or the light-shielding part surrounding the phase shifter is preferably at least 20 nm (actual size on the photomask). This is because the resolution limit that can be realized by using a technique such as double exposure in an electron beam exposure apparatus used for photomask processing is considered to be about 20 nm.

  In addition, since it is ideal to perform mask inspection using light having the same wavelength as the exposure wavelength, the dimension that can be recognized by the mask inspection apparatus is a dimension that is at least 1/4 times the exposure wavelength (actual dimension on the photomask). It is preferable that This is because a dimension smaller than that cannot be recognized by light. Here, the actual size on the photomask means an actual size on the photomask that is not converted by the mask magnification with respect to the size on the wafer.

  However, in order to obtain the effect as a mask enhancer, the light transmitted through the phase shifter and the light transmitted through the translucent part must interfere with each other. The dimension of the light-shielding part (or light-shielding part) is preferably 0.3 × λ / NA or less, which is the distance at which the two lights interfere strongly. However, since this is the distance in the transfer image on the wafer, it is preferable that it is a distance obtained by multiplying this dimension by the mask magnification M on the mask (0.3 × λ / NA) × M or less.

  Hereinafter, it will be described based on simulation results that an excellent pattern forming characteristic can be realized by using a mask enhancer structure that produces a preferable effect by the arrangement of the auxiliary pattern according to this modification.

  FIG. 17A shows a mask pattern used for simulation. In FIG. 17A, the same components as those of the photomask according to the first embodiment shown in FIG.

  As shown in FIG. 17A, in the main pattern 101 having a mask enhancer structure composed of a light shielding portion 101A (first semi-light shielding portion 101A in the first embodiment) and a phase shifter 101B, The width is L, and the width of the phase shifter 101B is W. A pair of auxiliary patterns 102 that diffract the exposure light and are not transferred by exposure are provided on both sides of the main pattern 101. Here, the distance from the center of the phase shifter 101B to the center of the auxiliary pattern 102 is G. The auxiliary pattern 102 is formed of a semi-light-shielding portion having a width D.

  FIGS. 17B to 17D show the results of evaluating the DOF and the exposure margin by simulation when a pattern having a width of 90 nm is formed while changing W and G with L = 90 m and D = 70 nm. Yes. Here, the exposure margin means the ratio (unit:%) of the exposure amount change necessary for changing the pattern dimension by 10%. In other words, the larger the exposure margin, the more stable the pattern dimension with respect to the change in exposure amount, and therefore the preferable situation is that the pattern dimension is less likely to vary with respect to the exposure amount variation in the actual pattern forming process. In the simulation, λ = 193 nm, NA = 0.7, and quadrupole illumination shown in FIG. 16A was used. However, in the configuration of FIG. 16A, x1 = 0.45 × NA and x2 = 0.6 × NA.

  Specifically, FIG. 17B shows the dependence of DOF on W (phase shifter width) in each case of G = 240 nm and G = 500 nm. FIG. 17C shows the dependence of the exposure margin on W (phase shifter width) in each case of G = 240 nm and G = 500 nm.

  As can be seen from FIGS. 17B and 17C, in the simulation result of G = 500 nm, when W≈90 nm, that is, the phase shifter width W is substantially equal to the mask pattern width (width of the main pattern 101) L. Then, the DOF and the exposure margin are maximized. Further, since DOF increases only slowly in proportion to the phase shifter width W, if the phase shifter width W is made narrower than the mask pattern width L, a sufficient margin cannot be obtained.

  In order to obtain a sufficient margin in pattern formation, it is necessary to set the phase shifter width W to substantially the same dimension as the mask pattern width L. In this case, however, the width of the light shielding portion surrounding the phase shifter becomes very fine. The mask enhancer structure is not preferable.

  On the other hand, as can be seen from FIGS. 17B and 17C, in the simulation result of G = 240 nm, when W≈60 nm, the exposure margin is maximized, and the DOF is also a simple light shielding with W = 0. It is sufficiently improved compared to the pattern. Therefore, by making the phase shifter width W about 30 nm thinner than the mask pattern width L of 90 nm, a sufficient margin is realized in pattern formation. In this case, the width of the light shielding portion surrounding the phase shifter is 15 nm. In this case, assuming that the reduction ratio M is 4 times, when converted to the actual size on the mask, it becomes 60 nm, so a dimension of 1/4 or more of the exposure wavelength (193 nm) is secured. In FIGS. 17B and 17C, W = 90 nm means that the main pattern 101 is composed only of the phase shifter.

  From the results described above, when the mask width of the main pattern is the same, the auxiliary pattern is arranged using the mask enhancer structure for the main pattern, compared to the case where a simple phase shifter is used for the main pattern. It can be seen that the exposure margin of the photomask is improved. If NA × sinφ = (x1 + x2) / 2, G = 240 nm corresponds to λ / (2 × sinφ), which is the optimum arrangement position of the first auxiliary pattern (auxiliary pattern 102). Although the description is omitted, the inventor of the present application confirmed the same result for the second auxiliary pattern (when G = 440 nm corresponding to λ / (2 × sinφ) + λ / (NA + sinφ) which is the optimum arrangement position). did.

  Further, as can be seen from FIGS. 17B and 17C, the exposure margin increases rapidly in proportion to the phase shifter width W until reaching the maximum value. On the other hand, if a sufficient exposure margin can be ensured, the phase shifter width W does not necessarily have to be a value that maximizes the exposure margin. Therefore, whether or not a sufficient margin can be secured in pattern formation while sufficiently securing the width of the semi-light-shielding portion or the light-shielding portion surrounding the phase shifter depends on the increase in the phase shifter width W in the mask enhancer structure. It depends on whether or not it improves greatly.

  FIG. 17D shows the result of plotting the dependence of DOF on the matrix of the phase shifter width W and the distance G from the center of the phase shifter to the center of the auxiliary pattern. From this result, it can be seen that in the case of G = 240 nm (0.24 μm) and G = 440 nm (0.44 μm), the DOF increases rapidly in proportion to the value of W. As described above, these are positions corresponding to the optimal position of the first auxiliary pattern and the optimal position of the second auxiliary pattern in this modification, respectively. Therefore, by arranging the auxiliary pattern at these positions, it is possible to secure a sufficient margin in pattern formation while sufficiently securing the width of the semi-light-shielding portion or the light-shielding portion surrounding the phase shifter in the mask enhancer structure.

(Fourth modification of the first embodiment)
Hereinafter, as a fourth modification of the first embodiment of the present invention, the width of the auxiliary pattern that causes a more preferable effect of the photomask according to the first embodiment (or the first to third modifications) will be described. .

  First, the results of simulating the dependency of the auxiliary pattern width, such as DOF and exposure margin, when the mask pattern shown in FIG. 4 is exposed will be described. In the simulation, λ = 193 nm, NA = 0.7, and annular illumination (inner diameter: 0.65, outer diameter: 0.75) was used. Further, it is assumed that the transmissivity of the light transmitting part and the phase shifter is 1, and the mask pattern part other than the phase shifter is composed of a semi-light-shielding part (transmittance: 6%). Furthermore, each auxiliary pattern is arranged at the optimum position shown in FIG. 4 according to sin φ = 0.7 × NA.

  In general, when the auxiliary pattern is thickened, the DOF of the main pattern increases while the exposure margin decreases. Therefore, how this phenomenon actually depends on the first and second auxiliary pattern widths was evaluated by simulating DOF and exposure margin in main pattern transfer. Specifically, the DOF and exposure margin when the widths D1 and D2 of the first and second auxiliary patterns 102 and 103 are changed from 40 nm to 100 nm with respect to the main pattern 101 of L = 140 nm and W = 80 nm, respectively. Asked.

  FIGS. 18A and 18B show simulation results of DOF and exposure margin when D2 is fixed to 70 nm and D1 is changed from 40 nm to 100 nm. FIGS. 18C and 18D show the simulation results of DOF and exposure margin when D1 is fixed to 70 nm and D2 is changed from 40 nm to 100 nm. Note that the DOF and the exposure margin shown in FIGS. 18A to 18D are for forming a pattern with a width of 90 nm.

  From the results shown in FIGS. 18A and 18B, it can be seen that the DOF in the main pattern transfer greatly increases in proportion to the increase in the width D1 of the first auxiliary pattern 102. On the other hand, it can also be seen that the exposure margin in the main pattern transfer greatly decreases as the width D1 increases. That is, when the first auxiliary pattern 102 is thickened, the DOF is improved but the exposure margin is reduced. Therefore, the only way to improve the margin is to trade off the two.

  On the other hand, as can be seen from the results shown in FIGS. 18C and 18D, the DOF in the main pattern transfer greatly increases as the width D2 of the second auxiliary pattern 103 increases. This is the same as the case of one auxiliary pattern 102. However, even if the width D2 is increased, the exposure margin in the main pattern transfer is only slightly reduced. Therefore, by increasing the width D2, the DOF can be improved without reducing the exposure margin.

  As described above, by making the second auxiliary pattern width D2 larger than the first auxiliary pattern width D1, it is possible to improve the DOF while keeping the exposure margin high. In addition, the inventor of the present application stated that the second auxiliary pattern has a thickness about 1.2 times that of the maximum width of the first auxiliary pattern in a range where the non-photosensitive portion of the resist is not formed. It has been empirically obtained that no non-photosensitive portion is formed. Therefore, even if the thickness of the second auxiliary pattern is 1.2 times that of the first auxiliary pattern, the phenomenon that the non-photosensitive portion of the resist is formed does not occur. Accordingly, by making the second auxiliary pattern width D2 1.2 times larger than the first auxiliary pattern width D1, the second auxiliary pattern forms the non-photosensitive portion of the resist while avoiding the phenomenon described above. An effect can be obtained reliably. However, in order to obtain the DOF improvement effect by the auxiliary pattern, the auxiliary pattern width is preferably half or more of the minimum dimension that the auxiliary pattern resolves. That is, when the first auxiliary pattern width D1 is large enough to produce the DOF improvement effect, the second auxiliary pattern width D2 is set to the first auxiliary pattern width in order to prevent the second auxiliary pattern from being resolved. It is preferably 2 times or less of D1.

  In this modification, the description has been made on the assumption that a pair of auxiliary patterns serving as diffracted light generation patterns are provided on both sides of the main pattern. However, when another main pattern is close to one side of the main pattern, the same effect as this modification can be obtained even if the auxiliary pattern is provided only on the side of the main pattern opposite to the side where the other main pattern is close Occurs.

(Second Embodiment)
Hereinafter, a photomask according to a second embodiment of the present invention will be described with reference to the drawings.

  FIG. 19 is a plan view of a photomask according to the second embodiment.

  As shown in FIG. 19, a main pattern 201 to be transferred by exposure is provided on a transparent substrate 200. The main pattern 201 includes a first semi-shielding portion 201A having a first transmittance that partially transmits exposure light, and a phase shifter 201B. The first semi-light-shielding portion 201A has the outer shape of the main pattern 201. The phase shifter 201B is provided on the periphery of the main pattern 201 so as to be surrounded by the first semi-light-shielding portion 201A. The phase shifter 201B is formed by digging up the transmissive substrate 200, for example. The first auxiliary pattern 202 (width: D1) that diffracts the exposure light and is not transferred by exposure is located at a position away from the center of the phase shifter 201B of the main pattern 201 on the transparent substrate 200 by λ / (2 × sinφ). However, the light-transmitting portion is provided between the main pattern 201 and the main pattern 201. Further, a second light that diffracts the exposure light and is not transferred by exposure at a position away from the center of the first auxiliary pattern 202 on the transparent substrate 200 (in a direction away from the main pattern 201) by λ / (NA + sinφ). The auxiliary pattern 203 (width: D2) is provided so as to sandwich the light-transmitting portion between the auxiliary pattern 203 and the first auxiliary pattern 202. Here, the first auxiliary pattern 202 and the second auxiliary pattern 203 are composed of a second semi-shielding portion having a second transmittance that partially transmits the exposure light.

  In the present embodiment, the distance between the phase shifter 201B and the first auxiliary pattern 202 may be a value in the vicinity of (λ / (2 × sinφ)) (first in the first embodiment). See Variation).

  In the present embodiment, the distance between the phase shifter 201B and the second auxiliary pattern 203 may be a value in the vicinity of (λ / (2 × sinφ)) + (λ / (NA + sinφ)) ( (Refer to a first modification of the first embodiment).

  Further, in the present embodiment, the above sin φ (φ: oblique incidence angle) is preferably 0.4 × NA or more and 0.80 × NA or less, particularly 0.58 × NA or more and 0. It is preferable that it is 7 × NA or less. When exposure is performed using annular illumination, the above sin φ is preferably 0.6 × NA or more and 0.80 × NA or less. Further, when exposure is performed using quadrupole illumination, the above sin φ is preferably 0.4 × NA or more and 0.60 × NA or less (see the second modification of the first embodiment). ).

  In the present embodiment, the width D2 of the second auxiliary pattern 203 is preferably larger than the width D1 of the first auxiliary pattern 202. In particular, it is preferable that D2 is equal to or greater than 1.2 times D1 (see the fourth modification of the first embodiment).

  Here, the characteristics of the phase shifter arrangement in the present embodiment will be described. First, when the dimension of a pattern to be formed is 0.3 × λ / NA or less, it is preferable to arrange a phase shifter at the center of a semi-light-shielding portion (semi-light-shielding pattern) corresponding to the pattern. When the dimension of the pattern to be formed is λ / NA or more, it is preferable to arrange a phase shifter at the peripheral edge of the semi-light-shielding pattern corresponding to the pattern. When the dimension of the pattern to be formed is larger than 0.3 × λ / NA and smaller than λ / NA, the phase shifter may be arranged at the center of the semi-light-shielding pattern corresponding to the pattern, or You may arrange | position to a peripheral part.

  The reason why the phase shifter is arranged at the peripheral portion in the mask pattern portion for forming a pattern having a dimension of λ / NA or more is to obtain an improvement effect of pattern formation characteristics by the “contour emphasis method” described later. This is because the diffracted light generation pattern can be arranged at an optimum position. Specifically, the phase shifter is preferably present in a range where the distance from the outer periphery of the semi-light-shielding pattern is λ / (2 × sin φ) or less. That is, in order to arrange the first-order diffracted light generation pattern, the phase shifter needs to exist at a position where the distance from the outer periphery of the semi-light-shielding pattern (main pattern) is λ / (2 × sinφ) or less. is there. Further, considering that the maximum value of sinφ is NA, in the case of a semi-light-shielding pattern having a dimension of λ / NA or more, it is preferable to arrange a phase shifter at the peripheral portion.

  In the photomask shown in FIG. 19, a mask pattern is composed of a main pattern 201, a first auxiliary pattern 202, and a second auxiliary pattern 203. Further, a portion of the transmissive substrate 200 where the mask pattern is not formed is a light transmitting portion (opening portion).

  Further, the relationship between the light transmitted through the phase shifter 201B and the light transmitted through the light transmitting portion is opposite in phase (specifically, the phase difference between the two is (150 + 360 × n) degrees or more and (210 + 360 × n) degrees or less). (Where n is an integer)).

  Further, the light transmitted through each of the first semi-light-shielding part 201A and the second semi-light-shielding part (first and second auxiliary patterns 202 and 203) and the light transmitted through the light-transmissive part have the same phase relationship. Specifically, the phase difference between the two is (−30 + 360 × n) degrees or more and (30 + 360 × n) degrees or less (where n is an integer).

  According to the second embodiment, since the main pattern 201 includes the first semi-light-shielding part 201A and the phase shifter 201B, the light-transmitting part and the first semi-light-shielding part 201A are transmitted by the light transmitted through the phase shifter 201B. A part of the light transmitted through can be canceled out. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern 201 can be enhanced.

  According to the second embodiment, the first and second auxiliary patterns 202 and 203 having low transmittance are provided separately from the main pattern 201. Specifically, a first auxiliary pattern (first-order diffracted light generation pattern) 202 is provided at a position away from the center of the phase shifter 201B of the main pattern 201 by λ / (2 × sin φ). A second auxiliary pattern (second order diffracted light generation pattern) 203 is provided at a position away from the center of the first auxiliary pattern 202 by λ / (NA + sinφ). Therefore, it is possible to reliably generate diffracted light that interferes with light transmitted through the phase shifter 201B of the main pattern 201. Therefore, the defocus characteristic in the transfer image of the main pattern 201 is improved, and as a result, the DOF characteristic is improved.

  However, in the present embodiment, even if the distance between the phase shifter 201B and the first auxiliary pattern 202 is a value in the vicinity of (λ / (2 × sinφ)), the above-described effect is produced to some extent. Similarly, even if the distance between the phase shifter 201B and the second auxiliary pattern 203 is a value in the vicinity of (λ / (2 × sin φ)) + (λ / (NA + sin φ)), the above-described effect is produced to some extent. (Refer to the first modification of the first embodiment). Further, the above sin φ (φ: oblique incidence angle) is preferably 0.4 × NA or more and 0.80 × NA or less, particularly 0.6 × NA or more and 0.7 × NA or less. There is preferably (see a second variant of the first embodiment).

  In the present embodiment, the width of the semi-light-shielding portion (first semi-light-shielding portion 201A) sandwiched between the phase shifter 201B and the light transmitting portion of the main pattern 201 is at least 20 nm (actual dimension on the mask) or more. It is preferable that the exposure wavelength is not less than one-fourth of the exposure wavelength (see the third modification of the first embodiment).

  In the present embodiment, the width D2 of the second auxiliary pattern 203 is preferably larger than the width D1 of the first auxiliary pattern 202, and more preferably, D2 is 1.2 times or more of D1. (For details, see the fourth modification of the first embodiment).

  In addition, according to the second embodiment, since the first and second auxiliary patterns 202 and 203 are semi-light-shielding portions, the degree of freedom of auxiliary pattern arrangement is improved, and thereby in the pattern arrangement including the main pattern 201. Since the periodicity can be increased, the DOF characteristics are further improved. Further, since the first and second auxiliary patterns 202 and 203 are semi-light-shielding portions, each auxiliary pattern can be thickened under the condition that it is not transferred by exposure, so that the processing becomes easy.

  Further, according to the second embodiment, since the phase shifter 201B is arranged at the peripheral edge of the main pattern 201, the contrast of the light intensity distribution in the vicinity of the main pattern 201 of the light image transmitted through the light transmitting portion is enhanced. Therefore, pattern formation can be performed while maintaining good defocus characteristics.

  Further, according to the second embodiment, since the phase shifter 201B is formed by digging the transmissive substrate 200, very excellent defocus characteristics are exhibited in pattern formation.

  In the second embodiment, only one of the first and second auxiliary patterns 202 and 203 may be arranged.

  In the second embodiment, it is preferable that the first transmittance of the first semi-light-shielding portion 201A constituting the main pattern 201 is 15% or less. In this way, it is possible to prevent resist film loss or optimize resist sensitivity during pattern formation. However, the first transmittance is preferably 3% or more in order to achieve both of these effects, the DOF improvement effect, and the contrast improvement effect.

  In the second embodiment, the second transmittance of the first and second auxiliary patterns 202 and 203 (that is, the second semi-light-shielding portion) is preferably 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of each auxiliary pattern being too high.

  In the second embodiment, the first semi-light-shielding part 201A and the second semi-light-shielding parts to be the first and second auxiliary patterns 202 and 203 are the same semi-light-shielding film, for example, the transmissive substrate 200. You may form from the metal thin film formed on it. In this case, each semi-light-shielding portion can be easily formed, so that the photomask can be easily processed. As the metal thin film, a thin film (thickness of about 50 nm or less) made of a metal such as Cr (chromium), Ta (tantalum), Zr (zirconium), Mo (molybdenum), Ti (titanium), or an alloy thereof is used. be able to. Specific alloy materials include Ta-Cr alloy, Zr-Si alloy, Mo-Si alloy, Ti-Si alloy and the like. Further, a thick film containing silicon oxide such as ZrSiO, CrAlO, TaSiO, MoSiO or TiSiO may be used instead of the metal thin film.

  In the second embodiment, the phase shifter 201B may be formed by forming a phase shift film made of a material having a high transmittance on the transparent substrate 200.

  Next, a method for improving the resolution of an isolated space pattern (hereinafter, contour) by a structure in which the phase shifter (phase shifter 201B) is provided at the peripheral edge of the light shielding pattern (main pattern 201), as found by the inventors of the present application. Will be described. Here, the “contour emphasis method” is a principle that holds true regardless of the shape of a minute space pattern in a positive resist process. Hereinafter, a case where a contact pattern is formed by a positive resist process will be described as an example. However, even when a negative resist process is used, the edge enhancement method is similarly established if the minute space pattern (resist removal pattern) in the positive resist process is replaced with a minute pattern (resist pattern). Further, in the following description, it is assumed that the light shielding pattern other than the phase shifter portion is composed of a semi-light shielding portion unless otherwise specified.

  For example, in a photomask in which a light shielding pattern is provided so as to surround an opening and a phase shifter is provided in the periphery of the light shielding pattern (hereinafter referred to as an outline emphasis mask), the periphery of the light shielding pattern, that is, the opening (translucent portion) The light transmitted through the phase shifter arranged in the vicinity of () can cancel out part of the light transmitted through the opening and the semi-light-shielding part. Therefore, if the intensity of the light transmitted through the phase shifter is adjusted so that the light in the region (contour) surrounding the opening is canceled, the light intensity distribution in which the light intensity at the contour decreases to a value close to 0 is obtained. Can be formed. Further, the light transmitted through the phase shifter strongly cancels the light at the contour portion, while weakly cancels the light near the center of the opening. As a result, the effect of increasing the slope of the profile from the opening to its periphery in the intensity distribution of the light transmitted through the contour enhancement mask is also obtained. Accordingly, since the intensity distribution of the light transmitted through the contour enhancement mask has a sharp profile, an image (image) with high contrast and light intensity is formed. This is the principle by which the image of light intensity can be enhanced in the contour enhancement method. That is, by arranging the phase shifter in the vicinity of the opening in the mask pattern composed of the semi-light-shielding portion having a low transmittance, an emergency image corresponding to the contour of the opening is formed in the image of the light intensity formed by the photomask. It is possible to form a strong dark portion. As a result, a light intensity distribution with enhanced contrast can be formed between the light intensity of the opening and the peripheral light.

  The higher the transmittance of the semi-light-shielding part used for the contour enhancement method, the better. However, due to the presence of the semi-light-shielding part, the transmitted light is originally present in the region that becomes the light-shielding part. In view of prevention of film loss of the resist film (resist pattern corresponding to the semi-light-shielding portion) or optimization of resist sensitivity, the maximum value of the transmittance of the semi-light-shielding portion is preferably about 15%. On the other hand, in order to obtain the effect by the contour enhancement method, it is preferable to set the minimum value of the transmittance of the semi-light-shielding portion to about 3%. Therefore, it can be said that the optimum value of the transmittance of the semi-light-shielding portion in the contour enhancement mask is 3% or more and 15% or less. In the contour emphasis mask, the phase shifter may be disposed so as to be in contact with the opening, or may be disposed so as to sandwich the semi-light-shielding portion between the opening. Further, the phase shifter may be arranged along the entire outline of the opening, or may be arranged along only a part of the outline.

  Further, by using a semi-light-shielding pattern that transmits light with the same phase as the light-transmitting portion as the light-shielding pattern, formation of a mask pattern using the centerline enhancement method (see the first embodiment) and the contour enhancement method simultaneously. Is possible. That is, a phase shifter is arranged at the center of a semi-light-shielding pattern for forming a fine line pattern. On the other hand, a phase shifter is disposed on the peripheral edge of the semi-light-shielding pattern for forming a large pattern. Thereby, the contrast of the image of the light intensity corresponding to the end portion of the large pattern is improved according to the contour emphasis method, so that the contrast of the image of the light intensity can be enhanced for every part in all the patterns to be formed. As described above, a semi-light-shielding pattern (semi-light-shielding portion that transmits light at the same phase as the light-transmitting portion) that has not been conventionally used as a mask pattern in pattern formation can form a pattern having an arbitrary shape. Further, by using a semi-light-shielding pattern, that is, a semi-light-shielding film, the following merit is also produced. That is, in a conventional mask, a thick metal film has to be used as a mask pattern in order to ensure light shielding properties. On the other hand, the semi-light-shielding pattern can be formed by a thin metal film having semi-light-shielding properties. In other words, the metal film for forming the mask pattern becomes thin, so that the mask can be easily processed. Specifically, when a Cr film is used, a thickness of about 100 nm is necessary as a mask pattern of a conventional mask, but a thickness of about 10 nm is sufficient as a semi-light-shielding pattern. For this reason, defects such as peeling do not occur when a fine mask pattern is formed by etching or when cleaning is performed after the mask pattern is formed.

  In the second embodiment, light transmitted through the opening (phase shifter 201B) in the mask enhancer (main pattern 201) using the diffracted light generation pattern (first and second auxiliary patterns 202 and 203). The reason why it is possible to generate diffracted light that interferes with the light and thereby improve the defocus characteristic (DOF characteristic) at the time of pattern formation is the same as in the first embodiment.

  That is, in the mask enhancer structure of the edge enhancement method (see FIG. 19) in which the phase shifter (phase shifter 201B) is arranged at the peripheral edge of the semi-light-shielding portion (first semi-light-shielding portion 201A), it is surrounded by the phase shifter. The semi-light-shielding portion transmits only light that is not exposed to the resist, but is optically the same as the light-transmitting portion. Accordingly, since the phase shifter of the contour enhancement method produces the same effect as the phase shifter of the centerline enhancement method, it is located at a position away from the center of the phase shifter by λ / (2 × sinφ) as in the case of the centerline enhancement method. Defocusing characteristics are improved by arranging the first-order diffracted light generation pattern and (or) arranging the second-order diffracted light generation pattern at a position away from the center of the first-order diffracted light generation pattern by λ / (NA + sinφ). Is possible. However, the allowable range of the arrangement position of the diffracted light generation pattern based on the position of the phase shifter in the contour enhancement method is the allowable range of the arrangement position of the diffracted light generation pattern in the center line enhancement method described in the first embodiment. Similar to range.

(Third embodiment)
Hereinafter, a photomask according to a third embodiment of the present invention will be described with reference to the drawings.

  FIG. 20A is a plan view of a photomask according to the third embodiment, and FIG. 20B is a cross-sectional view taken along the line XX-XX in FIG. FIGS. 21A to 21C are variations of the cross-sectional view taken along the line XX-XX in FIG.

  As shown in FIGS. 20A and 20B, a line-shaped main pattern 301 transferred by exposure is provided on the transparent substrate 300. The main pattern 301 includes a first light shielding portion 301A and a phase shifter 301B. The first light shielding portion 301A is formed so as to surround the line-shaped phase shifter 301B. In other words, the phase shifter 301B is disposed at the center of the main pattern 301. The phase shifter 301B is formed by digging up the transmissive substrate 300, for example. On both sides of the main pattern 301 on the transmissive substrate 300, a pair of auxiliary patterns 302 that diffract the exposure light and are not transferred by the exposure are provided so as to sandwich the light transmitting portion between the main pattern 301. The auxiliary pattern 302 includes a second light shielding portion.

  In the present embodiment, the first light shielding portion 301A and the second light shielding portion serving as the auxiliary pattern 302 are the same light shielding film 307, for example, a Cr (chromium) film formed on the transparent substrate 300, or the like. It is formed from the metal film.

  That is, the third embodiment is different from the first embodiment in that a mask enhancer including a light shielding portion and a phase shifter is used, and an auxiliary pattern (diffracted light generation pattern) including only the light shielding portion. It is using. In the photomask shown in FIGS. 20A and 20B, the main pattern 301 and the auxiliary pattern 302 constitute a mask pattern. Further, a portion of the transmissive substrate 300 where the mask pattern is not formed is a light transmitting portion (opening portion). In addition, the relationship between the light transmitted through the phase shifter 301B and the light transmitted through the light transmitting portion is opposite in phase (specifically, the phase difference between the two is (150 + 360 × n) degrees or more and (210 + 360 × n) degrees or less). (Where n is an integer)).

  According to the third embodiment, since the main pattern 301 includes the phase shifter 301B, a part of the light transmitted through the light transmitting portion can be canceled by the light transmitted through the phase shifter 301B. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern 301 can be enhanced. Further, since the auxiliary pattern 302 is provided separately from the main pattern 301, diffracted light that interferes with the light transmitted through the phase shifter 301B of the main pattern 301 is generated by arranging the auxiliary pattern 302 at an appropriate position. be able to. Therefore, the defocus characteristic in the transfer image of the main pattern 301 is improved, and as a result, the DOF characteristic is improved.

  Further, according to the third embodiment, since the phase shifter 301B is arranged at the central portion of the first light shielding portion 301A having the outer shape of the main pattern 301, the central portion of the light shielding image corresponding to the main pattern 301. Thus, for example, a fine line pattern can be formed while maintaining a good defocus characteristic.

  Further, according to the third embodiment, since the phase shifter 301B is formed by digging down the transmissive substrate 300, a very excellent defocus characteristic is exhibited in pattern formation. Specifically, since the phase shifter 301B is formed by creating an opening in the light-shielding film 307 (light-shielding portion) and digging up the transparent substrate 300 in the opening, the phase shifter 301B is made of a transparent material. High phase shifter. Further, since the intensity of the light with the opposite phase transmitted through the inside of the main pattern 301 can be adjusted by the opening size of the light shielding portion, the optimization of the light with the opposite phase transmitted through the main pattern 301 can be easily realized. In this case, a very excellent defocus characteristic is exhibited. That is, the mask dimension (main pattern dimension) can be adjusted by the width of the light-shielding part surrounding the phase shifter, and the intensity of light in the opposite phase that transmits the main pattern can be adjusted by the opening dimension of the light-shielding part. Therefore, a specific effect is obtained that the mask dimension and the intensity of light in the opposite phase can be adjusted independently. Accordingly, as in the first embodiment, the desired pattern dimensions can be easily achieved while reliably realizing the effect of adjusting the light of the opposite phase, for example, the effect of improving the focus characteristic and the effect of improving the contrast of the fine pattern. Can be realized.

  In the third embodiment, for example, as shown in FIG. 21A, the light shielding film 307 in the phase shifter forming region and the light transmitting portion forming region is formed on the laminated structure of the transparent substrate 300 and the light shielding film 307. A photomask that produces the same effect as the photomask shown in FIG. 20B can also be realized by removing the transmissive substrate 300 in the light-transmitting portion forming region.

  In the third embodiment, for example, as shown in FIGS. 21B and 21C, a light shielding film 307 is formed on a transmissive substrate 300 with a phase shift film 308 made of a material having high transmittance interposed therebetween. The main pattern 301 having the phase shifter 301B and the auxiliary pattern 302 may be formed by using the structured. Specifically, as shown in FIG. 21B, in a mask structure in which a phase shift film 308 having a high transmittance is deposited on a transparent substrate 300 and a light shielding film 307 is deposited thereon, a phase shifter is used. A photomask that produces an effect equivalent to that of the photomask shown in FIG. 20B also by removing the light shielding film 307 in the formation region and the light transmitting portion formation region and removing the phase shift film 308 in the phase shifter formation region. realizable. Further, according to the photomask shown in FIG. 21B, the phase of the phase shifter 301B can be controlled with high accuracy. On the other hand, as shown in FIG. 21C, in the same layered mask structure as in FIG. 21B, the light shielding film 307 in the phase shifter formation region and the light transmission portion formation region is removed and the phase of the light transmission portion formation region By removing the shift film 308, a photomask that produces the same effect as the photomask shown in FIG. 20B can be realized.

  Here, as in this embodiment, when a photomask is formed by forming a light shielding portion made of a metal film on a transmissive substrate and digging the transmissive substrate to form a phase shifter, the mask inspection It will be briefly described that an easy photomask is realized. The transmittance of a material having transparency to light changes depending on the wavelength of light. For this reason, mask inspection cannot be performed unless light having the same wavelength as the exposure light is used in mask inspection. That is, when a material having a low transmittance with respect to exposure light is inspected using light having a wavelength larger than that of the exposure light, for example, the material has a very high transmittance with respect to the wavelength of the inspection light. As a result, the light shielding property of the mask pattern may not be inspected. However, when a sufficiently thick metal film that can almost completely block light is used as the light-shielding portion as in this embodiment, the metal film can be used for most light except for wavelengths in the X-ray region. Almost complete shading film. Therefore, even if the wavelength of the exposure light and the wavelength of the light of the mask inspection apparatus are different, the mask inspection can be easily performed on the photomask as in this embodiment.

(Modification of the third embodiment)
Hereinafter, a photomask according to a modification of the third embodiment of the present invention will be described with reference to the drawings.

  FIG. 22 is a plan view of a mask pattern in a photomask according to a modification of the third embodiment. In FIG. 22, the same components as those of the photomask according to the third embodiment shown in FIGS.

  The first feature of this modification is that the auxiliary pattern 302 (width: D1) is provided at a position (λ / (2 × sinφ)) away from the center of the phase shifter 301B of the main pattern 301 (width: L). It is that.

  In addition, the second feature of this modification is, in other words, a position away from the center of the phase shifter 301B (width: W) of the main pattern 301 by (λ / (2 × sin φ)) + (λ / (NA + sin φ)). Then, the second auxiliary pattern 303 (width: diffracted by exposure light) is diffracted at a position away from the center of the auxiliary pattern (hereinafter referred to as the first auxiliary pattern) 302 by (λ / (NA + sinφ)). D <b> 2) is provided so as to sandwich the translucent portion with the first auxiliary pattern 302. The second auxiliary pattern 303 includes a light shielding part similar to the first auxiliary pattern 302.

  According to this modification, the DOF improvement effect by the diffracted light in the third embodiment can be realized with certainty.

  In this modification, any one of the first auxiliary pattern 302 and the second auxiliary pattern 303 may not be arranged.

  In the present modification, the above-described effect is produced to some extent even if the distance between the phase shifter 301B and the first auxiliary pattern 302 is a value in the vicinity of (λ / (2 × sinφ)).

  In the present modification, even if the distance between the phase shifter 301B and the second auxiliary pattern 303 is a value in the vicinity of (λ / (2 × sinφ)) + (λ / (NA + sinφ)), Some effect is produced.

  Here, the above-mentioned “neighboring value” regarding the arrangement position of the first auxiliary pattern 302 or the second auxiliary pattern 303 that the above-described effect occurs to some extent will be described in the first modification of the first embodiment. The allowable range of the arrangement position of the diffracted light generation pattern.

  In the present modification, the oblique incident angle φ is preferably 0.4 × NA or more and 0.80 × NA or less, particularly 0.58 × NA or more and 0.7 × NA. It is preferable that it is NA or less. When exposure is performed using annular illumination, the oblique incident angle φ is preferably 0.6 × NA or more and 0.80 × NA or less. In addition, when performing exposure using quadrupole illumination, the oblique incident angle φ is preferably 0.4 × NA or more and 0.60 × NA or less (in the first embodiment). (Refer to the second modification).

  In this modification, the width L of the main pattern 301 is preferably at least 2 × 20 nm (actual dimension on the mask) larger than the width W of the phase shifter 301B, and in particular, an exposure wavelength (exposure light wavelength) of 4 is used. It is preferably at least twice as large as one part. That is, in the mask enhancer structure of the main pattern, the width of the semi-light-shielding part (or light-shielding part) sandwiched between the phase shifter and the light-transmitting part is preferably at least 20 nm (actual dimension on the mask). It is preferable that it is 1/4 or more of the exposure wavelength. However, since it is a photomask using a mask enhancer structure, the width of the main pattern is preferably 0.8 × λ / NA or less. Therefore, a semi-light-shielding portion (sandwiched between a phase shifter and a light-transmitting portion ( Alternatively, it is preferable that the width of the light shielding portion does not exceed 0.4 × λ / NA (for details, refer to the third modification of the first embodiment).

  In this modification, the width D2 of the second auxiliary pattern 303 is preferably larger than the width D1 of the first auxiliary pattern 302, and more preferably, D2 is 1.2 times or more of D1. (For details, see the fourth modification of the first embodiment).

(Fourth embodiment)
A photomask according to the fourth embodiment of the present invention will be described below with reference to the drawings.

  FIG. 23A is a plan view of a photomask according to the fourth embodiment, and FIG. 23B is a cross-sectional view taken along line XXIII-XXIII in FIG.

  As shown in FIGS. 23A and 23B, a line-shaped main pattern 401 to be transferred by exposure is provided on the transparent substrate 400. The main pattern 401 is composed of a phase shifter. The phase shifter is formed, for example, by digging up the transmissive substrate 400. On both sides of the main pattern 401 on the transmissive substrate 400, a pair of auxiliary patterns 402 that diffract the exposure light and are not transferred by the exposure are provided so as to sandwich the light transmitting portion between the main pattern 401. The auxiliary pattern 402 includes a semi-light-shielding portion having a transmittance that partially transmits the exposure light.

  That is, the fourth embodiment is different from the first embodiment in that the main pattern 401 uses a structure including only a phase shifter instead of the mask enhancer structure. In the photomask shown in FIGS. 23A and 23B, a main pattern 401 and an auxiliary pattern 402 constitute a mask pattern. Further, the portion of the transmissive substrate 400 where the mask pattern is not formed is a light transmitting portion (opening).

  Further, the light transmitted through the phase shifter serving as the main pattern 401 and the light transmitted through the translucent portion have an opposite phase relationship (specifically, the phase difference between them is (150 + 360 × n) degrees or more and (210 + 360 ×)). n) degrees or less (where n is an integer)).

  Further, the light transmitted through the semi-light-shielding portion serving as the auxiliary pattern 402 and the light transmitted through the light-transmitting portion have the same phase relationship (specifically, the phase difference between them is (−30 + 360 × n) degrees or more and ( 30 + 360 × n) degrees or less (where n is an integer)).

  According to the fourth embodiment, since the main pattern 401 is composed of a phase shifter, a part of the light transmitted through the light transmitting portion can be canceled out by the light transmitted through the phase shifter. For this reason, the contrast of the light intensity distribution in the light-shielded image corresponding to the main pattern 401 can be enhanced. In addition, since the auxiliary pattern 402 having a low transmittance is provided separately from the main pattern 401, the auxiliary pattern 402 is arranged at an appropriate position to interfere with light transmitted through the phase shifter that is the main pattern 401. Diffracted light can be generated. Therefore, the defocus characteristic in the transfer image of the main pattern 401 is improved, and as a result, the DOF characteristic is improved.

  Further, according to the fourth embodiment, since the auxiliary pattern 402 is a semi-light-shielding portion, the degree of freedom of auxiliary pattern arrangement is improved, thereby improving the periodicity in the pattern arrangement including the main pattern 401. , DOF characteristics are further improved. Further, since the auxiliary pattern 402 is a semi-light-shielding portion, the auxiliary pattern 402 can be thickened under the condition that the auxiliary pattern 402 is not transferred by exposure, so that the processing becomes easy.

  Further, according to the fourth embodiment, since the phase shifter to be the main pattern 401 is formed by digging the transmissive substrate 400, very excellent defocus characteristics are exhibited in pattern formation.

  In the fourth embodiment, since the main pattern 401 is composed only of the phase shifter, the effect of the mask enhancer structure used in the first to third embodiments, that is, the dimension of the mask enhancer and the phase shifter provided thereto. By adjusting the size of the (opening), it is not possible to obtain a specific effect of easily realizing a desired size in pattern formation while controlling both contrast and defocus characteristics. However, if only the defocus characteristic is improved, the mask enhancer may be replaced by a simple phase shifter.

  In the fourth embodiment, it is preferable that the transmittance of the semi-light-shielding portion serving as the auxiliary pattern 402 is 6% or more and 50% or less. In this way, it is possible to reliably realize the DOF improvement effect by diffracted light while preventing the non-photosensitive portion of the resist from being formed due to the light shielding property of the auxiliary pattern 402 being too high.

  In the fourth embodiment, the mask cross-sectional structure shown in FIG. 23C may be used instead of the mask cross-sectional structure shown in FIG. That is, as shown in FIG. 23C, in the structure in which the semi-light-shielding film 406 and the phase shift film 408 made of a material with high transmittance are sequentially laminated on the transparent substrate 400, the phase shifter formation region (main pattern formation) A structure in which the phase shift film 408 other than (region) is removed and the semi-light-shielding film 406 in the light transmitting portion formation region is removed may be used. According to the structure shown in FIG. 23C, it is possible to easily realize that the transmittance of the phase shifter serving as the main pattern 401 is lower than the transmittance of the semi-light-shielding portion serving as the auxiliary pattern 402. In this case, if the transmittance of the phase shifter is set to 15% or less, not only the fine pattern but the entire main pattern having an arbitrary dimension can be configured by the phase shifter. A photomask in which patterns are mixed can be easily formed.

(Modification of the fourth embodiment)
Hereinafter, a photomask according to a modification of the fourth embodiment of the present invention will be described with reference to the drawings.

  FIG. 24 is a plan view of a mask pattern in a photomask according to a modification of the fourth embodiment. In FIG. 24, the same components as those in the photomask according to the fourth embodiment shown in FIGS. 23A and 23B are denoted by the same reference numerals, and the description thereof is omitted.

  The first feature of this modification is that the auxiliary pattern 402 (width: D1) is provided at a position away from the main pattern 401 (width: W), that is, the center of the phase shifter (λ / (2 × sinφ)). It is that you are.

  The second feature of the present modification is that the auxiliary pattern (in other words, a position away from the center of the phase shifter that is the main pattern 401 by (λ / (2 × sinφ)) + (λ / (NA + sinφ)). The second auxiliary pattern 403 (width: D2) that diffracts the exposure light and is not transferred by exposure at a position away from the center of (402) (hereinafter referred to as the first auxiliary pattern) (λ / (NA + sinφ)) In other words, the light transmitting portion is provided between the auxiliary pattern 402 and the auxiliary pattern 402. The second auxiliary pattern 403 includes a semi-light-shielding part similar to the first auxiliary pattern 402.

  According to this modification, the DOF improvement effect by the diffracted light in the fourth embodiment can be realized with certainty.

  In this modification, any one of the first auxiliary pattern 402 and the second auxiliary pattern 403 may not be arranged.

  Further, in this modification, the above-described effect is produced to some extent even when the distance between the phase shifter serving as the main pattern 401 and the first auxiliary pattern 402 is a value in the vicinity of (λ / (2 × sinφ)). .

  In this modification, the distance between the phase shifter serving as the main pattern 401 and the second auxiliary pattern 403 is a value in the vicinity of (λ / (2 × sinφ)) + (λ / (NA + sinφ)). However, the above-mentioned effects occur to some extent.

  Here, the above-mentioned “neighbor values” relating to the arrangement position of the first auxiliary pattern 402 or the second auxiliary pattern 403 that the above-described effect occurs to some extent will be described in the first modification of the first embodiment. The allowable range of the arrangement position of the diffracted light generation pattern.

  In the present modification, the oblique incident angle φ is preferably 0.4 × NA or more and 0.80 × NA or less, particularly 0.58 × NA or more and 0.7 × NA. It is preferable that it is NA or less. When exposure is performed using annular illumination, the oblique incident angle φ is preferably 0.6 × NA or more and 0.80 × NA or less. In addition, when performing exposure using quadrupole illumination, the oblique incident angle φ is preferably 0.4 × NA or more and 0.60 × NA or less (in the first embodiment). (Refer to the second modification).

  In this modification, the width D2 of the second auxiliary pattern 403 is preferably larger than the width D1 of the first auxiliary pattern 402, and more preferably, D2 is 1.2 times or more of D1. (For details, see the fourth modification of the first embodiment).

  In this modification, the semi-light-shielding portion is used for each of the auxiliary patterns 402 and 403, but a light-shielding portion may be used instead. In this case, the contrast between the main pattern and the auxiliary pattern is reduced as compared with the case where the semi-light-shielding portion is used for the auxiliary pattern. On the other hand, by arranging the auxiliary pattern at the position described in this modification, a large DOF is obtained. Needless to say, a photomask capable of realizing the above can be obtained.

  In addition, when a simplified mask structure as described below is used instead of the mask structure of the present modification, the contrast and DOF enhancement effects in pattern formation are reduced compared to the present modification, The enhancement effect can be improved as compared with the conventional example.

  Specifically, in place of the auxiliary patterns 402 and 403 (see FIG. 24) made of the semi-light-shielding portion of this modification, auxiliary patterns 412 and 413 made of the light-shielding portion are used as shown in FIG. In this case, since the light shielding property of the auxiliary pattern is increased, the degree of freedom in arranging the auxiliary pattern is lowered. However, also in the mask structure shown in FIG. 25A, since the main pattern 401 is composed of phase shifters, the auxiliary patterns 412 and 413 are arranged at the same arrangement positions as the auxiliary patterns 402 and 403 of this modification. By this, the improvement effect of DOF is acquired. In addition, by replacing the semi-light-shielding portion with the light-shielding portion, the performance in pattern formation is reduced, while the effect that the mask inspection or the like can be performed reliably is obtained (see the third embodiment).

  25 (b) and 25 (c) are diagrams showing variations in the cross-sectional configuration along the line XXV-XXV in FIG. 25 (a).

  In the configuration shown in FIG. 25B, the phase shifter to be the main pattern 401 is formed by digging up the transmissive substrate 400. Each auxiliary pattern (the second auxiliary pattern 413 is not shown) is composed of a light shielding film 407 formed on the transmissive substrate 400. According to the configuration shown in FIG. 25B, a phase shifter having a sufficiently high transmittance can be realized, so that the DOF improvement effect can be sufficiently obtained.

  In the structure shown in FIG. 25C, the light shielding film 407 other than the auxiliary pattern formation region is removed and the light transmission film 407 is removed from the structure in which the phase shift film 408 and the light shielding film 407 are sequentially stacked on the transparent substrate 400. This can be realized by removing the phase shift film 408 in the optical part formation region. That is, the main pattern 401 has a single-layer structure of the phase shift film 408, and the auxiliary patterns 412 and 413 (the second auxiliary pattern 413 is not shown) are formed between the phase shift film 408 and the light shielding film 407. It consists of a laminated structure. If the transmittance of the phase shifter is set to 15% or less in the configuration shown in FIG. 25C, not only a fine pattern but also a large pattern can be configured by the phase shifter. A photomask in which patterns having arbitrary dimensions are mixedly arranged as a main pattern can be easily formed.

  In addition, instead of the auxiliary patterns 402 and 403 (see FIG. 24) made of the semi-light-shielding portion of this modification, as shown in FIG. 26, when auxiliary patterns 422 and 423 made of phase shifters are used, the light shielding of the auxiliary pattern is performed. While the effect of lowering the lightness of the main pattern is lower than that of the main pattern, the DOF improvement effect can be sufficiently obtained. The configuration shown in FIG. 26 can be realized only by patterning the phase shift film on the transmissive substrate, so that photomask processing becomes easy.

  Further, instead of the main pattern 401 made of phase shifter and the auxiliary patterns 402 and 403 made of the semi-light-shielding portion (see FIG. 24) in this modification, as shown in FIG. 27, the main pattern 411 made of the light-shielding portion and the light-shielding portion. Auxiliary patterns 412 and 413 composed of parts may be used. In this case, as compared with the present modification, the light shielding enhancement effect and the DOF improvement effect by the main pattern 411 are reduced, but a certain degree of DOF improvement effect is obtained. According to the mask configuration shown in FIG. 27, since the mask pattern is composed of only the light-shielding portion, photomask processing and inspection are very easy.

(Fifth embodiment)
Hereinafter, a pattern forming method according to the fifth embodiment of the present invention, specifically, a pattern forming method using the photomask according to any one of the first to fourth embodiments (hereinafter referred to as the photomask of the present invention). Will be described with reference to the drawings.

  28A to 28D are cross-sectional views showing respective steps of the pattern forming method according to the fifth embodiment.

  First, as shown in FIG. 28A, a film to be processed 501 such as a metal film or an insulating film is formed on a substrate 500, and then, on the film to be processed 501 as shown in FIG. Then, a positive resist film 502 is formed.

  Next, as shown in FIG. 28C, the photomask of the present invention, for example, the photomask according to the first embodiment shown in FIG. The resist film 502 is exposed by the transmitted light 504 that has passed through.

  A linear main pattern 101 transferred by exposure is provided on the transparent substrate 100 of the photomask used in the step shown in FIG. The main pattern 101 includes a first semi-light-shielding portion 101A having a first transmittance that partially transmits exposure light, and a phase shifter 101B. The first semi-light-shielding portion 101A is formed so as to surround the line-shaped phase shifter 101B. The phase shifter 101B is formed by digging up the transmissive substrate 100, for example. On both sides of the main pattern 101 on the transmissive substrate 100, a pair of auxiliary patterns 102 that diffract the exposure light and are not transferred by the exposure are provided so as to sandwich the light-transmitting portion between the main pattern 101. The auxiliary pattern 102 is composed of a second semi-shielding portion having a second transmittance that partially transmits the exposure light.

  In the exposure step shown in FIG. 28C, the resist film 502 is exposed using an oblique incidence exposure light source. At this time, since the semi-light-shielding portion having a low transmittance is used for the mask pattern, the entire resist film 502 is exposed with weak energy. However, as shown in FIG. 28C, only the latent image portion 502a of the resist film 502 corresponding to the area other than the main pattern 101 is irradiated with exposure energy sufficient to dissolve the resist in the development process. .

  Next, as shown in FIG. 28D, the resist film 502 is developed to remove the latent image portion 502a, thereby forming a resist pattern 505 corresponding to the main pattern 101.

  According to the fifth embodiment, since it is a pattern formation method using the photomask of the present invention (specifically, the photomask according to the first embodiment), the same effects as those of the first embodiment can be obtained. . Specifically, oblique incidence exposure is performed on a substrate (wafer) coated with a resist through the photomask of the present invention. At this time, since the mask enhancer (main pattern 101) having the phase shifter (opening) has a very strong light-shielding property, only the resist corresponding to other regions other than the mask enhancer is dissolved in the development process. Sufficient exposure energy is irradiated. Further, since the contrast of the light-shielded image formed by the mask enhancer is very high and the defocus characteristic of the light-shielded image is excellent, it is possible to form a fine pattern with a high DOF.

  In the fifth embodiment, the photomask according to the first embodiment is used. Alternatively, when the photomask according to any of the second to fourth embodiments is used, The same effect as each embodiment is acquired.

  In the fifth embodiment, the positive resist process is used. However, the same effect can be obtained by using a negative resist process instead.

  In the fifth embodiment, it is preferable to use an oblique incidence illumination method (oblique incidence exposure method) in the step of irradiating the exposure light 503 shown in FIG. In this way, in the light intensity distribution of the light transmitted through the photomask of the present invention, the contrast between the main pattern and the portion corresponding to the light transmitting portion is improved. Also, the focus characteristic of the light intensity distribution is improved. Therefore, the exposure margin and focus margin in pattern formation are improved. In other words, it is possible to form a fine pattern with excellent defocus characteristics.

  Next, a method of calculating an oblique incidence angle that plays an important role in oblique incidence exposure using the photomask of the present invention having an auxiliary pattern (diffracted light generation pattern) will be described.

  When a point light source is used as the oblique incident exposure light source, the oblique incident angle is clearly defined (see FIG. 5). However, in the case of illumination by a normal light source having an area, there are a plurality of oblique incident angles.

  FIGS. 29A to 29E are main calculation methods of the oblique incidence angle defined by the inventors of the present application so that an appropriate arrangement position of the diffracted light generation pattern can be calculated even in the case of illumination with a light source having an area. FIG.

  FIG. 29A shows a method of calculating the oblique incident angle when performing annular illumination. As shown in FIG. 29 (a), in annular illumination, the inner diameter S1 of the annular light source corresponds to the light source having the minimum oblique incident angle φ1, and the outer diameter S2 of the annular light source is the light source having the maximum oblique incident angle φ2. Correspond. Therefore, the oblique incident angle φ used for calculating the arrangement position of the diffracted light generation pattern is defined based on the light source that emits light from the position of S = (S1 + S2) / 2 calculated by the inner diameter S1 and the outer diameter S2. . That is, the oblique incident angle φ = (φ1 + φ2) / 2. If S1 and S2 are values normalized by NA, the oblique incident angle φ in the photomasks of the first to fourth embodiments is based on sin φ = S × NA = (S1 + S2) × NA / 2. May be set. However, when annular illumination is used, it is preferable to use illumination and a photomask in which the oblique incident position S is 0.6 or more and 0.8 or less in the pattern forming method. It is most preferable to use an illumination and a photomask having a value in the vicinity of 0.7 (see the second modification of the first embodiment).

  Needless to say, the oblique incident angle φ may be set to an arbitrary value not less than φ1 and not more than φ2. In other words, it goes without saying that the oblique incident angle φ may be set to any value satisfying the relationship of S1 × NA ≦ sin φ ≦ S2 × NA.

  By the way, as described in the first embodiment, in the oblique incidence exposure, when the oblique incidence angle φ is sinφ <NA / 3, the effect of improving the defocus characteristic cannot be obtained. Therefore, even in the case of illumination by a light source having an area, it is desirable that the oblique incident angle φ is composed only of a value that can obtain a sufficient defocus improvement effect. If the light source used for exposure contains an oblique incident angle φ such that sin φ <NA / 3, a mask provided with an optimum diffracted light generation pattern ignoring the oblique incident light from that portion. When the exposure is performed using the above, a defocus characteristic superior in the pattern formation is exhibited, compared with the case where the diffracted light generation pattern is provided in consideration of the oblique incident light from the aforementioned portion. Therefore, the minimum angle ξ of the oblique incident angle φ is preferably a value defined by sin ξ = 0.4 × NA. That is, the oblique incident angle φ used for calculating the arrangement position of the diffracted light generation pattern is (ξ + φ2) / 2. In other words, the oblique incident angle φ is defined by sin φ = (0.4 + S2) × NA / 2.

  FIGS. 29B to 29E each show a method of calculating the oblique incidence angle when performing quadrupole illumination. In the case of quadrupole illumination, as shown in each figure, the oblique incidence angle is calculated using an XY coordinate system with the center of the quadrupole light source (four-eye light source) (hereinafter referred to as the light source center) as the origin. Specifically, in the case of quadrupole illumination, the oblique incident angle is optimized for each pattern parallel to the X axis and the Y axis in the XY coordinate system. That is, the oblique incident angle is not defined by the distance from the light source center to each light source, but by the coordinate value of each light source on the X axis or Y axis. Hereinafter, the case where the oblique incident angle is optimized for the pattern parallel to the Y axis will be described. However, the same applies to the case where the oblique incident angle is optimized for the pattern parallel to the X axis. First, the minimum oblique incident angle is defined by the value closest to the origin among the absolute values of the X coordinate of each light source of the quadrupole light source. That is, the minimum oblique incident angle is defined by x1 shown in each of FIGS. Similarly, the maximum oblique incidence angle is defined by the value farthest from the origin among the absolute values of the X coordinate in each light source of the quadrupole light source, that is, x2 shown in each of FIGS. 29 (b) to (e). Therefore, in the case of quadrupole illumination, the oblique incident angle φ used for calculating the arrangement position of the diffracted light generation pattern may be set according to sin φ = S × NA = (x1 + x2) × NA / 2. However, when using quadrupole illumination, it is preferable to use illumination and a photomask in which the oblique incident position S is 0.4 or more and 0.6 or less in the pattern forming method. It is most preferable to use illumination and a photomask having a value in the vicinity of 0.5 (see the second modification of the first embodiment).

  Needless to say, the oblique incident angle φ may be set to any value satisfying the relationship of x1 × NA ≦ sin φ ≦ x2 × NA.

  Also, in the case of the quadrupole illumination shown in FIGS. 29B to 29E, the minimum angle ξ of the oblique incident angle φ is sinξ = 0. A value defined by 4 × NA is preferable. That is, the oblique incident angle φ used for calculating the arrangement position of the diffracted light generation pattern is defined by sin φ = (0.4 + x2) × NA / 2.

(Sixth embodiment)
Hereinafter, the mask data creation method according to the sixth embodiment of the present invention, specifically, any one of the first to fourth embodiments using the centerline enhancement method, the contour enhancement method, and the diffracted light generation pattern. A mask data creation method for a photomask according to the present invention (hereinafter referred to as a photomask of the present invention) will be described with reference to the drawings. In the present embodiment, the functions, properties, and the like of the constituent elements of the photomask are the same as the corresponding constituent elements in the above-described photomask of the present invention unless otherwise specified.

  Before describing specific processing contents, important points in the photomask mask data creation method of the present invention will be described. In the photomask of the present invention, even when forming an isolated pattern, the phase shifter, the light shielding portion or the semi-light shielding portion surrounding the phase shifter, and the diffracted light generation pattern (auxiliary pattern) arranged around the phase shifter are related. Therefore, in order to set the pattern dimension at the time of pattern formation, that is, CD (Critical Dimension) to a desired value, the width of the phase shifter, the width of the light-shielding portion or the semi-light-shielding portion, and the position and width of the diffracted light generation pattern. It is necessary to determine values of a plurality of elements such as. In many cases, the combination of the values of the above elements for realizing a desired CD is not limited to one, and there are a plurality of combinations. Therefore, in the mask data creation method of the present embodiment, the value of an important element is preferentially determined to maximize the margin in pattern formation, and then the pattern dimension is determined by an element that has a small effect on the margin in pattern formation. Make adjustments.

  Specifically, in the present embodiment, as an element having a high influence on the margin in pattern formation, firstly, the position and width of the phase shifter are determined, and secondly, the position and width of the auxiliary pattern are determined, Finally, by adjusting the width of the light-shielding portion or semi-light-shielding portion surrounding the phase shifter, that is, the width of the region sandwiched between the phase shifter and the light transmitting portion, mask data for realizing a desired CD can be created. preferable. Specific processing contents will be described below.

  FIG. 30 is a flowchart of a mask data creation method according to the sixth embodiment, specifically, a method of creating an LSI mask pattern that becomes a light-shielding pattern on a mask based on a fine desired pattern. FIGS. 31A to 31G are diagrams showing specific mask pattern creation examples in each step of the mask data creation method according to the sixth embodiment.

  FIG. 31A shows a desired pattern to be formed by a mask pattern. That is, a pattern 600 shown in FIG. 31A is a pattern corresponding to a region where the resist is not desired to be exposed in the exposure using the photomask of the present invention. When pattern formation is described in the present embodiment, the description will be made on the assumption that a positive resist process is used unless otherwise specified. That is, the description will be made on the assumption that the photosensitive portion of the resist is removed by development and the non-photosensitive portion of the resist remains as a resist pattern. Therefore, in the case of using the negative resist process, it is exactly the same except that the photosensitive portion of the resist remains as a resist pattern and the non-photosensitive portion of the resist is removed.

  First, in step S1, a desired pattern 600 shown in FIG. 31A is input to a computer used for creating mask data.

  Next, in step S2, the desired pattern shown in FIG. 31A is used depending on whether overexposure or underexposure is used when performing exposure on the photomask created according to the present embodiment. Resize to enlarge or reduce Alternatively, a desired pattern may be resized in order to intentionally adjust the dimensions in response to dimensional changes that occur in various processes during pattern formation. The resized pattern is a main pattern 601 composed of a semi-light-shielding portion as shown in FIG.

  Next, in step S3, as shown in FIG. 31 (c), the shape of the phase shifter 602 (width, etc .; the same applies hereinafter) arranged at the center of the region where the dimension of the main pattern 601 is a predetermined value or less is determined. . At this time, the phase shifter 602 is completely contained in the main pattern 601, that is, the semi-light-shielding pattern. That is, the outermost edge of the main pattern 601 is made to be the edge of the semi-light-shielding pattern.

  Here, the width of the phase shifter to be generated is preferably adjusted as follows. That is, the width of the region sandwiched between the phase shifter and the light transmitting portion in the semi-light-shielding portion surrounding the phase shifter is changed for CD adjustment performed later, and as a result, the width of the region is less than or equal to a predetermined width. The phase shifter width is adjusted in advance so that it does not occur. The predetermined width is preferably 20 nm or more in actual dimensions on the mask or a quarter or more of the exposure wavelength. Therefore, at this time, the phase shifter is secured so that the width of the region sandwiched between the phase shifter and the light transmitting portion is not less than the predetermined value and the predicted CD is not thicker than a desired value. Set the width. Specifically, the phase shifter width that realizes the desired CD in the above state is defined as the maximum phase shifter width, and the contrast and DOF of each pattern are optimized within the range of the phase shifter width that is equal to or less than the maximum phase shifter width. A phase shifter is arranged so that This eliminates the need to adjust the pattern size by changing the phase shifter width in a later step. The description so far has been made on the premise that the main pattern has a mask enhancer structure. However, in the case of creating mask data so that the main pattern is composed of only the phase shifter, the above processing contents are described. It may be omitted.

  Next, in step S4, as shown in FIG. 31 (d), the shape of the phase shifter 602 disposed at the periphery of the region where the dimension of the main pattern 601 is larger than a predetermined value is determined. At this time, the phase shifter 602 is completely contained in the main pattern 601, that is, the semi-light-shielding pattern, so that the outermost edge of the main pattern 601 becomes the edge of the semi-light-shielding pattern.

  Next, in step S5, as shown in FIG. 31 (e), each is determined based on a predetermined distance from the phase shifter 602 arranged in steps S3 and S4 (an oblique incident angle of illumination by a light source used during exposure). The first-order diffracted light generation pattern 603 and the second-order diffracted light generation pattern 604 composed of a semi-shielding portion are arranged as auxiliary patterns for diffracting the exposure light at positions separated by a predetermined distance. For example, when the phase shifter 602 is a line pattern, the line-shaped diffracted light generation pattern is arranged at a position away from the phase shifter 602 by a predetermined distance so as to be parallel to the phase shifter 602. However, when another pattern exists at the arrangement position of the diffracted light generation pattern, the diffracted light generation pattern is not arranged in a region where the other pattern exists.

  With the above processing, the position and width of the phase shifter having a large influence on the margin in pattern formation and the position and width of the diffracted light generation pattern are determined to be optimum values.

  Next, in step S6, preparation for processing for adjusting the dimension of the mask pattern is performed so that a pattern having a desired dimension corresponding to the mask pattern is formed when exposure is performed using the photomask of the present invention. Do. That is, preparation for a process called normal OPC (Optical Proximity Correction) process is performed. In this embodiment, a mask pattern region whose dimension is adjusted, that is, a CD at the time of pattern formation and whose size is adjusted based on the result is limited to only the edge of the main pattern 601, that is, the edge of the semi-light-shielding pattern. That is, as shown in FIG. 31 (f), the outermost edge of the main pattern 601 is set to the CD adjustment edge 605. In other words, the adjustment of the CD, which is the dimension of the pattern to be formed, is adjusted only by the outermost edge of the semi-light-shielding portion constituting the main pattern. As a result, for the main pattern in which the phase shifter is arranged, the CD can be adjusted by the width of the semi-light-shielding portion sandwiched between the phase shifter and the light transmitting portion. Therefore, it is possible to create a mask pattern capable of realizing a desired CD without modifying the phase shifter 602 and the diffracted light generation patterns 603 and 604 arranged at the optimum positions with respect to the phase shifter 602.

  Next, in step S7, the transmissivities of the semi-light-shielding portion and the phase shifter used for the mask pattern are set.

  Next, in step S8, step S9, and step S10, OPC processing (for example, model-based OPC processing) is performed. Specifically, in step S8, the dimensions of the resist pattern formed by the main pattern 601 on which the phase shifter 602 is arranged and the diffracted light generation patterns 603 and 604 are predicted by simulation in consideration of the optical principle and resist development characteristics. To do. At this time, in the simulation, not only the lithography process but also a process related to pattern formation such as dry etching may be considered. Subsequently, in step S9, it is checked whether or not the predicted dimension of the resist pattern matches the desired dimension. If it does not coincide with the desired dimension, the CD adjustment edge 605 is moved based on the difference between the predicted dimension of the resist pattern and the desired dimension in step S10, thereby deforming the main pattern 601.

  A feature of the present embodiment is that a mask pattern capable of forming a resist pattern having a desired dimension is realized by changing only the CD adjustment edge 605 set in step S6. That is, by repeating Steps S8 to S10 until the predicted size of the resist pattern matches the desired size, finally, in Step S11, a mask pattern that can form a resist pattern having the desired size is output. FIG. 31G shows an example of the mask pattern output in step S11.

  By the way, the parameters that naturally affect the pattern (resist pattern) dimensions in the photomask of the present invention are very large, such as the width of the phase shifter, the width of the mask pattern (main pattern), and the width and position of the auxiliary pattern.

  On the other hand, according to the sixth embodiment, in order to realize a mask having excellent contrast and defocus characteristics, which are important pattern formation characteristics, first, the width of the phase shifter 602 and the diffracted light which are important parameters. The arrangement positions of the generation patterns 603 and 604 are determined. Thereafter, only the outermost edge of the main pattern 601 set as the CD adjustment edge 605 is moved to perform pattern dimension control, thereby realizing a mask pattern having excellent pattern formation characteristics.

  Therefore, a photomask is created based on the mask data created by the method of the present embodiment, and further, oblique incidence exposure is performed using the photomask, thereby forming a fine pattern or a fine space. High contrast and very good DOF characteristics can be obtained.

  Further, according to the sixth embodiment, since the phase shifter 602 is disposed at the center of the region where the dimension of the main pattern 601 is a predetermined value or less, a finer desired pattern can be formed and excellent pattern formation characteristics A mask pattern having

  Further, according to the sixth embodiment, since the phase shifter 602 is arranged at the peripheral edge of the main pattern 601, a desired pattern having an arbitrary shape can be formed and a mask pattern having excellent pattern formation characteristics can be realized. .

  So far, the mask data creation method has been described assuming that the diffracted light generation pattern can be arranged at a position where optimal diffracted light can be generated. Next, the mask data when another main pattern is close to the main pattern. The creation method (particularly the processing in step S5 above) will be described in detail. In the following description, a mask pattern creation example shown in FIG. 32 is used instead of the mask pattern creation example shown in FIG. In FIG. 32, reference numeral 701 denotes a main pattern of interest, and reference numerals 702, 703, 704, and 705 denote other main patterns that are close to the main pattern 701, respectively. In the following description, the optimum position of the center of the first-order diffracted light generation pattern is set to the position G0 from the center of the phase shifter, and the allowable position range of the center of the first-order diffracted light generation pattern is G1 to G2. (G1 <G0 <G2). At this time, it is preferable to match G1 and G2 with the allowable range of the arrangement position of the first-order diffracted light generation pattern described in detail in the first embodiment. Further, the optimum position of the center of the second-order diffracted light generation pattern is set to the position of H0 from the center of the phase shifter, and the allowable position range of the center of the second-order diffracted light generation pattern is from H1 to H2 (where H1 < H0 <H2). At this time, it is preferable that H1 and H2 also coincide with the allowable range of the arrangement position of the second-order diffracted light generation pattern described in detail in the first embodiment.

  Hereinafter, a method for generating a diffracted light generation pattern in consideration of the relationship between the main pattern 701 of interest and the other main patterns 702 to 705 adjacent thereto will be described in detail. Each main pattern 701 to 705 has a mask enhancer structure. That is, the main pattern 701 includes a phase shifter 701B and a semi-shielding portion 701A surrounding it, the main pattern 702 includes a phase shifter 702B and a semi-shielding portion 702A surrounding it, and the main pattern 703 includes a phase shifter 703B and the same. The main pattern 704 is composed of a phase shifter 704B and a semi-light-shielding portion 704A surrounding it, and the main pattern 705 is composed of a phase shifter 705B and a semi-light-shielding portion 705A surrounding it.

  As shown in FIG. 32, it is assumed that the main pattern 701 and the main pattern 702 are close to each other so that the distance p1 between the centers is p1 <2 × G1. In this case, a diffracted light generation pattern is not disposed between the main pattern 701 and the main pattern 702.

  The main pattern 701 and the main pattern 703 are close to each other so that the distance p2 between the centers is 2 × G1 ≦ p2 <2 × G2. In this case, the first-order diffracted light generation pattern 801 is arranged at the center between the main pattern 701 and the main pattern 703.

  The main pattern 701 and the main pattern 704 are close to each other so that the distance p3 between the centers is 2 × G2 ≦ p3 <2 × H1. In this case, a first-order diffracted light generation pattern 802 having a center at a position distant from G0 from the center of the main pattern 701 is disposed between the main pattern 701 and the main pattern 704, and G0 away from the center of the main pattern 704. A first-order diffracted light generation pattern 803 having a center at a certain position is arranged.

  Further, it is assumed that the main pattern 701 and the main pattern 705 are close to each other so that the distance p4 between the centers is 2 × H1 ≦ p4 <2 × H2. In this case, a first-order diffracted light generation pattern 804 having a center at a position away from G0 from the center of the main pattern 701 is disposed between the main pattern 701 and the main pattern 705. A second-order diffracted light generation pattern 805 is disposed at the center between them, and a first-order diffracted light generation pattern 806 having a center at a position away from G0 from the center of the main pattern 705 is disposed.

  In addition, when the center of the main pattern of interest and the center of another main pattern adjacent to it are separated by 2 × H2 or more, a position separated by G0 from the center of each main pattern is between each main pattern. A pair of first-order diffracted light generation patterns each having a center may be disposed, and a pair of second-order diffracted light generation patterns each having a center may be disposed at a position away from the center of each main pattern by H0.

  According to the diffracted light generation pattern generation method described above, it is possible to reliably generate a preferable diffracted light generation pattern even when the main pattern is close to another main pattern with an arbitrary size.

  In the sixth embodiment, the mask pattern having the mask enhancer structure using the semi-light-shielding portion has been described. However, instead of this, the mask pattern having the mask enhancer structure using the light-shielding portion is targeted. It is good. Specifically, all portions described as semi-light-shielding portions in this embodiment may be replaced with light-shielding portions. In this case, the step of placing the phase shifter 602 on the peripheral edge of the main pattern 601 in step S4 may be omitted. When a light shielding part is used instead of the semi-light shielding part, a great improvement in contrast or DOF can be obtained when a pattern having a predetermined dimension or less is formed by a mask pattern created according to the method of this embodiment. Specifically, in the formation of a minute space, the effect of improving contrast or DOF is small. However, for example, in the pattern formation of the gate layer of an LSI circuit for high-speed operation, that is, in the formation of a pattern in which only the transistor pattern dimension is extremely small and does not include a minute space pattern, the above-described effect is obtained. Extremely expensive.

  By the way, it is important to increase a margin for forming a transistor pattern in creating mask pattern data for a normal LSI. For this reason, when the main patterns are close to each other, and the diffracted light generation pattern cannot be placed at the optimum position at the same time for both main patterns, the diffracted light generation pattern is at the optimum position with respect to the main pattern as the transistor portion On the other hand, the diffracted light generation pattern may be generated without regard to the optimum position for the main pattern serving as the wiring portion. Hereinafter, a specific description will be given with reference to the flowchart shown in FIG. The improvement process flow shown in FIG. 33 is different from the process flow shown in FIG. 30 in that the process of arranging the diffracted light generation pattern based on the arrangement position of the phase shifter of the main pattern is performed in two steps. It is. Specifically, the process of step S5 in the process flow of FIG. 30 is divided into two steps of steps S51 and S52 in the process flow of FIG. Specifically, first, in step S51, an optimal diffracted light generation pattern is generated and arranged for the phase shifter of the main pattern in the transistor region. Next, in step S52, a diffracted light generation pattern is generated for the phase shifter of the main pattern in the region other than the transistor region. According to this method, since the main pattern in the transistor region and the main pattern in other regions (such as the wiring region) are close to each other, there arises a situation in which an optimal diffracted light generation pattern cannot be placed on both main patterns at the same time. Even so, an optimal diffracted light generation pattern can be arranged for the main pattern in the transistor region. The extraction of the main pattern of the transistor region can be easily performed by a process such as extracting an overlap between the gate layer and the active layer based on the LSI design data.

  FIG. 34 shows an example of mask pattern creation, which is a specific processing result according to the processing flow shown in FIG. As shown in FIG. 34, the main patterns 710 to 712 include phase shifters 710B to 712B and light shielding portions 710A to 712A, respectively. In addition, first-order diffracted light generation patterns 811 to 815 are arranged for these main patterns 710 to 712. Here, the phase shifter 710B is a phase shifter arranged in the transistor region, and the other phase shifters 711B and 712B are phase shifters arranged in a region other than the transistor region. The optimal arrangement position of the first-order diffracted light generation pattern with respect to the phase shifter is a position away from G0 from the center of the phase shifter, and the allowable position range of the first-order diffracted light generation pattern is from G1 to G2. . Further, it is assumed that the main pattern 710 and the main pattern 711, and the main pattern 711 and the main pattern 712 are close to each other, and the distance between the centers of the adjacent main patterns is p. Here, it is assumed that p is 2 × G1 or more and less than 2 × G2. Further, the diffracted light generation patterns 812 and 815 are diffracted light generation patterns arranged in a region between main patterns that are close to each other. The other diffracted light generation patterns 811, 813, and 814 are diffracted light generation patterns arranged with respect to the main pattern that is not close to the other main patterns. As shown in FIG. 34, the diffracted light generation pattern 815 arranged between the main patterns 711 and 712 arranged in the region other than the transistor region is arranged in the center between the two main patterns. On the other hand, the diffracted light generation pattern 812 arranged between the main pattern 710 arranged in the transistor region and the main pattern 711 is arranged so as to have a center at the position of G0 from the center of the phase shifter 710B of the main pattern 710. Is done. That is, the diffracted light generation pattern is preferentially arranged at an optimal position with respect to the phase shifter arranged in the transistor region.

  In the improved process flow shown in FIG. 33, the process of step S5 in the process flow of FIG. 30 is divided into two steps, so that the diffracted light generation pattern is preferentially arranged at the optimal position with respect to the main pattern in the transistor region. This is easily realized. However, it goes without saying that the arrangement of the diffracted light generation pattern in each of the transistor region and other regions may be performed simultaneously in one step. In addition, the transistor region has been described as an important region in pattern formation. However, when the region important in pattern formation is a region other than the transistor region, the transistor region in the above-described improvement process flow is changed to another region. Think of it as a replacement for the area.

  In the sixth embodiment, the mask data creation method for the photomask whose main pattern has the mask enhancer structure has been mainly described. However, the mask data for the photomask whose main pattern does not have the mask enhancer structure can also be created by the above processes. That is, for example, when the main pattern is composed of only the phase shifter, the edge of the phase shifter becomes the edge of the main pattern, so CD adjustment may be performed according to the width of the phase shifter. Further, for example, when the main pattern is composed only of the light shielding pattern, the edge of the light shielding pattern becomes the edge of the main pattern. In this case, one or both of steps S3 and S4 in the processing flow of FIG. 30 may be omitted.

  In the first to sixth embodiments, description has been made assuming a transmission type photomask. However, the present invention is not limited to this, and the present invention can also be applied to a reflective mask if the transmission phenomenon of exposure light is replaced with a reflection phenomenon, for example, by replacing transmittance with reflectance. Is.

  As described above, the present invention relates to a pattern forming method used for manufacturing a semiconductor integrated circuit device, and is particularly useful when applied to fine pattern formation.

(A) is a top view of the photomask which concerns on the 1st Embodiment of this invention, (b) And (c) is sectional drawing in the II line | wire of (a). (A)-(d) is a figure which shows the variation in sectional structure of the photomask which concerns on the 1st Embodiment of this invention. (A)-(d) is a figure for demonstrating that the influence which the dimensional error in a mask has on pattern formation can be made small by using a mask enhancer structure for the main pattern with an auxiliary pattern. It is a top view of the mask pattern in the photomask which concerns on the 1st modification of the 1st Embodiment of this invention. (A) is a figure explaining the diffraction phenomenon produced when it exposes with respect to the mask in which the pattern is arrange | positioned periodically, (b) is a 0th-order diffracted light and 1 with respect to the mask shown to (a). It is a figure explaining the diffraction phenomenon which arises when only the next diffracted light passes through a lens and oblique incidence exposure is performed under the condition of r0 = −r1, and (c) shows the mask shown in (a). It is a figure explaining the diffraction phenomenon which arises when an oblique incidence exposure is performed on the conditions which 0th order diffracted light, + 1st order diffracted light, and -1st order diffracted light pass a lens. (A) is a figure which shows the point light source used for the simulation of DOF characteristic at the time of performing oblique incidence exposure, changing the pitch of a pitch pattern with respect to the mask shown in FIG. 5 (a), (b) It is a figure which shows the pitch pattern used for simulation, (c) is a figure which shows the result of this simulation, (d) is 0th order diffracted light, 1st order diffracted light with respect to the mask shown in FIG. It is a figure explaining the diffraction phenomenon which arises when a grazing incidence exposure is performed on the conditions through which a 2nd-order diffracted light passes a lens. (A) is a figure which shows the pitch pattern (the mask enhancer arrange | positioned substantially infinitely periodic) used for the simulation of DOF characteristic at the time of performing oblique incidence exposure, changing the pitch of the pitch pattern on a mask. (B) is a figure which shows the result of this simulation. (A) is a figure which shows the pitch pattern (three mask enhancers arrange | positioned in parallel) used for the simulation of DOF characteristic at the time of performing oblique incidence exposure, changing the pitch of the pitch pattern on a mask, (B) is a figure which shows the result of this simulation. (A)-(c) is a top view which shows the mask pattern of this invention by which a diffracted light generation pattern (auxiliary pattern) is arrange | positioned so that the DOF characteristic of the main pattern which consists of a mask enhancer may improve significantly, respectively. . (A) shows the pattern (mask pattern) used in the simulation for demonstrating that a favorable DOF characteristic can be obtained by arranging the diffracted light generation pattern at the positions shown in FIGS. It is a figure and (b) is a figure which shows the result of this simulation. 9A shows a lens numerical aperture NA = 0.6 in a simulation for demonstrating that a good DOF characteristic can be obtained by arranging a diffracted light generation pattern at the positions shown in FIGS. It is a figure which shows the result at the time of using the optical condition of sinphi = 0.7 * NA, (b) uses the optical condition of lens numerical aperture NA = 0.6 and sinphi = 0.6 * NA in this simulation. (C) is a diagram showing the results when the optical conditions of lens numerical aperture NA = 0.7 and sin φ = 0.7 × NA are used in the simulation. (A)-(c) is a figure which shows the pattern (mask pattern) used for the simulation for evaluating the improvement effect of the DOF characteristic by the diffracted light generation pattern of this invention when using the light source which has an area, respectively. (D) is a diagram showing a light source used in the simulation, and (e) is a mask enhancer when the mask patterns shown in (a) to (c) are exposed under predetermined conditions. (F) is a diagram showing the simulation result of the light intensity distribution generated corresponding to the above, and (f) corresponds to the mask enhancer by exposing the mask patterns shown in (a) to (c) under predetermined conditions. It is a figure which shows the result of having simulated the defocus characteristic of CD of the pattern at the time of forming a pattern with a width of 0.1 micrometer. (A) is a figure which shows a mode that the 1st-order diffracted light generation pattern is arrange | positioned in the position which left | separated the distance P1 from the center of the phase shifter in the mask pattern shown in FIG.12 (b), (b) is (a). FIG. 13C is a diagram showing a change in DOF when exposure is performed while changing the distance P1 with respect to the mask pattern shown in FIG. 12C. FIG. 12C shows the second-order diffracted light generation pattern in the mask pattern shown in FIG. It is a figure which shows a mode that it arrange | positions in the position away from the center of the 1st-order diffracted light generation pattern by distance P2, (d) is the case where it exposes, changing distance P2 with respect to the mask pattern shown in (c). It is a figure which shows the change of DOF in. (A)-(d) is a figure for demonstrating the result of the simulation which performs oblique incidence exposure using annular illumination with respect to the mask pattern shown to Fig.9 (a)-(c). (A)-(d) is a figure for demonstrating the result of the simulation which performs oblique incidence exposure using dipole illumination with respect to the mask pattern shown to Fig.9 (a)-(c). (A)-(d) is a figure for demonstrating the result of the simulation which performs oblique incidence exposure using quadrupole illumination with respect to the mask pattern shown to Fig.9 (a)-(c). (A)-(d) demonstrates that the outstanding pattern formation characteristic is realizable using the mask enhancer structure which produces a favorable effect by arrangement | positioning of the auxiliary pattern which concerns on the 3rd modification of the 1st Embodiment of this invention. It is a figure for doing. (A)-(d) is a figure for demonstrating the result of having simulated the dependence of the DOF, the exposure margin, etc. with respect to the width | variety of an auxiliary pattern at the time of exposing with respect to the mask pattern shown in FIG. . It is a top view of the photomask which concerns on the 2nd Embodiment of this invention. (A) is a top view of the photomask which concerns on the 3rd Embodiment of this invention, (b) is sectional drawing in the XX-XX line of (a). (A)-(c) is the variation of sectional drawing in the XX-XX line of Fig.20 (a). It is a top view of the photomask which concerns on the modification of the 3rd Embodiment of this invention. (A) is a top view of the photomask which concerns on the 4th Embodiment of this invention, (b) And (c) is sectional drawing in the XXIII-XXIII line of (a). It is a top view of the photomask which concerns on the modification of the 4th Embodiment of this invention. (A) is a top view which shows an example of the simplified photomask which concerns on the modification of the 4th Embodiment of this invention, (b) And (c) is sectional drawing in the XXV-XXV line | wire of (a) It is. It is a top view which shows an example of the simplified photomask which concerns on the modification of the 4th Embodiment of this invention. It is a top view which shows an example of the simplified photomask which concerns on the modification of the 4th Embodiment of this invention. (A)-(d) is sectional drawing which shows each process of the pattern formation method which concerns on the 5th Embodiment of this invention. (A) to (e) are the main methods for calculating the oblique incident angle defined by the inventors of the present application so that the appropriate arrangement position of the diffracted light generation pattern can be calculated even in the case of illumination with a light source having an area. FIG. It is a flowchart of the mask data creation method which concerns on the 6th Embodiment of this invention. (A)-(g) is a figure which shows the specific example of mask pattern creation in each process of the mask data creation method which concerns on the 6th Embodiment of this invention. It is a figure which shows the detailed preparation example of the diffracted light generation pattern by the mask data preparation method concerning the 6th Embodiment of this invention. It is a figure which shows the process flow improved in the mask data creation method which concerns on the 6th Embodiment of this invention. It is a figure which shows the detailed preparation example of the diffracted light generation pattern by the improved process flow in the mask data preparation method concerning the 6th Embodiment of this invention. It is a top view of the conventional photomask. It is a top view of the conventional photomask.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Transparent substrate 101 Main pattern 101A Semi-light-shielding part or light-shielding part 101B Phase shifter 102 1st auxiliary pattern (semi-light-shielding part)
103 2nd auxiliary pattern (semi-light-shielding part)
106 Semi-light-shielding film 107 Light-shielding film 108 Phase shift film 109 Light-shielding part 110 Mask enhancer 111 Semi-light-shielding part 112 Phase shifter 113 First-order diffracted light generation pattern 114 Second-order diffracted light generation pattern 140 Light source 141 Light 142 0th-order diffracted light 143 First-order diffracted light 144 First-order diffracted light 145 Second-order diffracted light 150 Mask 151 Pitch pattern 152 Lens 160 Transparent substrate 161 Phase shifter 170 Transparent substrate 171 Semi-shielding pattern 200 Transparent substrate 201 Main pattern 201A Semi-shielding portion 201B Phase shifter 202 First auxiliary Pattern (semi-shading part)
203 2nd auxiliary pattern (semi-shading part)
300 transmissive substrate 301 main pattern 301A light shielding portion 301B phase shifter 302 first auxiliary pattern (light shielding portion)
303 2nd auxiliary pattern (light-shielding part)
307 Light-shielding film 308 Phase shift film 400 Translucent substrate 401 Main pattern (phase shifter)
402 1st auxiliary pattern (semi-shading part)
403 Second auxiliary pattern (semi-shielding portion)
406 Semi-shield film 407 Shield film 408 Phase shift film 411 Main pattern (shield part)
412 1st auxiliary pattern (light-shielding part)
413 2nd auxiliary pattern (light-shielding part)
422 First auxiliary pattern (phase shifter)
423 Second auxiliary pattern (phase shifter)
500 Substrate 501 Processed film 502 Resist film 502a Latent image portion 503 Exposure light 504 Transmitted light 505 Resist pattern 600 Desired pattern 601 Main pattern 602 Phase shifter 603 First-order diffracted light generation pattern 604 Second-order diffracted light generation pattern 605 For CD adjustment Edge 701 Main pattern 701A Semi-shield part 701B Phase shifter 702 Main pattern 702A Semi-shield part 702B Phase shifter 703 Main pattern 703A Semi-shield part 703B Phase shifter 704 Main pattern 704A Semi-shield part 704B Phase shifter 705 Main pattern 705 Shifter 710 Main pattern 710A Shading part 710B Phase shifter 711 Main pattern 711A Shading part 711B Phase shifter 712 Main pattern 712 A Shielding part 712B Phase shifter 801 First-order diffracted light generation pattern 802 First-order diffracted light generation pattern 803 First-order diffracted light generation pattern 804 First-order diffracted light generation pattern 805 Second-order diffracted light generation pattern 806 First-order diffracted light generation pattern 811 First First-order diffracted light generation pattern 812 First-order diffracted light generation pattern 813 First-order diffracted light generation pattern 814 First-order diffracted light generation pattern 815 First-order diffracted light generation pattern S Distance from normal passing through lens center to light source (oblique incident position)
φ Oblique incident angle NA Lens numerical aperture P Pitch θ1 Diffraction angle of first-order diffracted light θ2 Diffraction angle of second-order diffracted light λ Light wavelength r0 Arrival position of zero-order diffracted light on lens r1 Arrival position of first-order diffracted light on lens L pattern Width (line width)
W aperture width R distance between patterns D diffracted light generation pattern width D1 first order diffracted light generation pattern width D2 second order diffracted light generation pattern width Ia light intensity corresponding to the center of the phase shifter Ib light intensity corresponding to the center of the semi-light-shielding pattern X Distance from the phase shifter to the first order diffracted light generation pattern Y Distance from the first order diffracted light generation pattern to the second order diffracted light generation pattern P1 Distance from the phase shifter to the first order diffracted light generation pattern P2 Generation of the second order diffracted light from the first order diffracted light generation pattern Distance to pattern S1 Inner diameter of annular light source S2 Outer diameter of annular light source x1 Value closest to origin among X coordinates of each quadrupole light source x2 Most origin of X coordinates of each light source of quadrupole light source Far from

Claims (21)

A pattern forming method using a photomask having a mask pattern formed on a transmissive substrate and a light-transmitting portion in which the mask pattern is not formed on the transmissive substrate,
Forming a resist film on the substrate (a);
Irradiating the resist film with exposure light through the photomask; and
Developing the resist film irradiated with the exposure light to form a resist pattern (c),
In the photomask, the mask pattern has a main pattern transferred by exposure, and an auxiliary pattern that diffracts the exposure light and is not transferred by exposure,
The main pattern has a first transmittance that partially transmits the exposure light, and a first semi-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion, and the light transmitting A phase shifter that transmits the exposure light in an opposite phase with respect to the part,
The auxiliary pattern has a second transmittance that partially transmits the exposure light and includes a second semi-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion. ,
The pattern forming method, wherein a pattern width of the auxiliary pattern is smaller than a pattern width of the main pattern. In the pattern formation method of Claim 1,
In the photomask, the first transmittance is 15% or less. In the pattern formation method of Claim 2,
In the photomask, the first semi-light-shielding part and the second semi-light-shielding part are formed of the same semi-light-shielding film, and the second transmittance is 6% or more and 15% or less. There is provided a pattern forming method. In the pattern formation method of Claim 1 or 2,
In the photomask, the first semi-light-shielding part and the second semi-light-shielding part are formed from the same semi-light-shielding film. In the pattern formation method of any one of Claims 1-4,
In the photomask, the light transmitting portion is interposed between the main pattern and the auxiliary pattern, and sin φA = NA × when a predetermined oblique incident position is SA (0.4 ≦ SA ≦ 0.8). With respect to the oblique incident angle φA defined by SA, the center of the auxiliary pattern is arranged at a position away from the center of the main pattern by M × (λ / (2 × sin φA)) or in the vicinity thereof. Characteristic pattern formation method (where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and numerical aperture of the reduction projection optical system of the exposure machine). In the pattern formation method of any one of Claims 1-4,
In the photomask, the light transmitting portion is interposed between the main pattern and the auxiliary pattern, and sin φB = NA × when a predetermined oblique incident position is SB (0.4 ≦ SB ≦ 0.8). The center of the auxiliary pattern is M × ((λ / (2 × sin φB)) + (λ / (NA + sin φB))) away from the center of the main pattern with respect to the oblique incident angle φB defined by SB. Alternatively, the pattern forming method is characterized in that it is arranged in the vicinity thereof (where λ is the wavelength of the exposure light, and M and NA are the reduction magnification and numerical aperture of the reduction projection optical system of the exposure machine). In the pattern formation method of any one of Claims 1-4,
In the photomask, the auxiliary pattern is disposed in the outer direction of the first auxiliary pattern as viewed from the first auxiliary pattern, the first auxiliary pattern sandwiching the translucent portion between the auxiliary pattern and the main pattern, and A pattern forming method comprising: a second auxiliary pattern sandwiching the translucent portion with a first auxiliary pattern. In the pattern formation method of Claim 7,
In the photomask, the first auxiliary pattern with respect to an oblique incident angle φA defined by sin φA = NA × SA when a predetermined oblique incident position is SA (0.4 ≦ SA ≦ 0.8). The pattern forming method (where λ is the wavelength of the exposure light), wherein the center of the pattern is arranged at or near the center of the main pattern by M × (λ / (2 × sin φA)) M and NA are the reduction magnification and numerical aperture of the reduction projection optical system of the exposure machine). In the pattern formation method of Claim 8,
In the photomask, the second auxiliary pattern with respect to an oblique incident angle φA defined by sin φA = NA × SA when a predetermined oblique incident position is SA (0.4 ≦ SA ≦ 0.8). The pattern forming method is characterized in that the center of the pattern is arranged at a position away from the center of the main pattern by M × ((λ / (2 × sin φA)) + (λ / (NA + sin φA))) or in the vicinity thereof. . A pattern forming method using a photomask having a mask pattern formed on a transmissive substrate and a light-transmitting portion in which the mask pattern is not formed on the transmissive substrate,
Forming a resist film on the substrate (a);
Irradiating the resist film with exposure light through the photomask; and
Developing the resist film irradiated with the exposure light to form a resist pattern (c),
In the photomask, the mask pattern has a main pattern transferred by exposure, and an auxiliary pattern that diffracts the exposure light and is not transferred by exposure,
The main pattern has a first transmittance that partially transmits the exposure light, and a first semi-shielding portion that transmits the exposure light in the same phase with respect to the light transmitting portion, and the light transmitting A phase shifter that transmits the exposure light in an opposite phase with respect to the part,
The auxiliary pattern includes a first auxiliary pattern having a width D1 provided so as to sandwich the light transmitting portion between the auxiliary pattern and the outer side of the first auxiliary pattern as viewed from the main pattern. A second auxiliary pattern having a width D2 provided so as to sandwich the light-transmitting portion between the first auxiliary pattern,
A pattern forming method, wherein D2 is larger than D1. In the pattern formation method of Claim 10,
In the photomask, D2 / D1 is 1.2 or more and 2 or less. In the pattern formation method of Claim 10 or 11,
In the photomask, the phase shifter is disposed at the center of the main pattern so as to be surrounded by the first semi-light-shielding portion. In the pattern formation method of Claim 12,
In the photomask, the main pattern includes a light shielding portion instead of the first semi-light shielding portion, and the phase shifter. In the pattern formation method of any one of Claims 10-12,
In the photomask, the transmittance of the first semi-light-shielding portion is 15% or less. In the pattern formation method of any one of Claims 10-14,
In the photomask, the first auxiliary pattern and the second auxiliary pattern are configured by a light shielding portion, or have a second transmittance that partially transmits the exposure light, and the light transmission A pattern forming method comprising: a second semi-light-shielding portion that transmits the exposure light in the same phase with respect to the portion. In the pattern formation method of Claim 15,
In the photomask, the second semi-light-shielding portion has a transmittance of 6% or more and 50% or less. In the pattern formation method of any one of Claims 1-16,
In the step (b), an oblique incidence illumination method is used. In the pattern formation method of any one of Claims 1-17,
In the step (b), the exposure light is irradiated by annular illumination. The pattern forming method according to claim 18,
An average value of an outer diameter and an inner diameter in the illumination shape used for the annular illumination is 0.58 or more and 0.8 or less (however, the values of the outer diameter and the inner diameter) As a value normalized by the numerical aperture of the exposure machine). In the pattern formation method of any one of Claims 1-17,
In the step (b), the exposure light is irradiated by quadrupole illumination. The pattern forming method according to claim 20,
The distance from the light source center of the center position of each of the four polarized illumination shapes used for the quadrupole illumination is 0.4 / (0.5) ^0.5 or more and 0.6 / (0.5) ^0.5 or less. A pattern forming method (however, a value standardized by a numerical aperture of an exposure machine is used as the distance value).

Images