Phase shift mask and method for producing a body formed by a pattern using the same

Abstract

Halftone phase shift mask (200) onto which ArF excimer laser exposure light is to be directed, wherein the mask (200) comprises a transparent substrate (201) and a semi-transparent layer pattern (202) formed on the transparent substrate (201) and consisting only of Si and N, or a semi-transparent layer pattern (202) formed on the transparent substrate (201) and consisting only of Si, N and O, characterized in that the semi-transparent layer pattern (202) has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light and a transmittance of 15% to 38% at the wavelength of the ArF excimer laser exposure light and furthermore has a layer thickness of 57 nm to 67 nm, the semi-transparent layer pattern (202) has a main pattern (202a) that is to be resolved onto a wafer, and an auxiliary pattern (202b) that is used to assist in the resolution of the main pattern (202a) and is not to be resolved onto the wafer, the auxiliary pattern (202b) comprises one or more convex pattern elements, each with a width or length of 60 nm or less, and the main pattern (202a) has the same thickness as the auxiliary pattern (202b).

Description

Technical field

[0001] The present invention relates to a phase-shifting mask, for example, for the fabrication of a semiconductor device; and to a method for fabricating a patterned body using the mask. More specifically, the present invention relates to a phase-shifting mask that enables the transfer of a pattern of the mask onto a wafer using a high NA exposure device employing ArF excimer laser exposure light with a wavelength of 193 nm in photolithography, resulting in a pattern with high transfer properties and excellent resistance to irradiation with ArF excimer laser exposure light and cleaning resistance; and to a method for fabricating a patterned body using the same. Technical background

[0002] A halftone phase-shift mask, consisting of transparent and semi-transparent regions, is known as a means of improving the resolution of a phase-shift mask used in photolithography. A typical example of a halftone phase-shift mask is one with a 6% transmittance, in which a MoSi layer is used as the semi-transparent layer.

[0003] When the spacing of intermediates to be formed in a wafer is finely defined, a semi-transparent layer forming regions that should transmit light halfway in a halftone phase-shift mask must be a layer with properties that include a smaller EMF bias and OPC bias and a larger exposure latitude and depth of focus (EL-DOF) than desired properties when the intermediates are transferred to the wafer.

[0004] In the case of a halftone phase-shift mask, the interference of light at its pattern element boundary regions, based on the phase effect, reduces the intensity of the light at the interference-caused regions to zero, thus improving the contrast of the transferred images. It is expected that if the mask's transmittance is 15% or higher, this phase effect will become more pronounced, further enhancing the contrast of the transferred images.

[0005] To adjust the transmittance of a semi-transparent layer in a halftone phase-shift mask to a target range, a metal is embedded in the semi-transparent layer to adjust the light transmittance of the layer (Patent Literature 1, 2, 3 and 4).

[0006] However, with conventional halftone phase-shift masks, each featuring a semi-transparent layer with an embedded metal, it is known that the inclusion of the metal in the semi-transparent layer causes problems with regard to the resistance of their pattern to irradiation with ArF excimer laser exposure light and their cleaning resistance. As described in examples in patent literature 1 and 2, Mo is frequently used as the metal to be embedded in the semi-transparent layer. When Mo is used, ArF excimer laser exposure light is applied to the Mo for a long period of time, causing water to be generated by the humid atmosphere. The MoSi layer is oxidized by the generated water, resulting in the growth of a silicon (Si) oxide layer.It is known that the growth causes a problem regarding resistance to laser irradiation, whereby this problem is a problem that causes the masks to be altered in the critical dimension of the pattern. In this case, it is also known that a problem regarding cleaning resistance is caused, whereby this problem is a problem that occurs in the same way during the cleaning step of the halftone phase-shift masks.

[0007] However, in the case of an attempt to use a metal-free semi-transparent layer so that a halftone phase-shift mask avoids such problems regarding resistance to irradiation with ArF excimer laser exposure light and cleaning resistance, regions of the mask where the semi-transparent layer is formed have an excessively high transmittance (patent literature 5 and 6).

[0008] As a result, for example in patent literature 5, it is necessary to laminate a metal-free phase-setting layer (half-transparent layer) and a transmittance-setting layer that is different from the phase-setting layer and contains a metal, thereby forming a halftone phase-shift mask, since only the metal-free phase-setting layer (half-transparent layer) gives the mask too high a transmittance.

[0009] In the case of using a halftone phase-shift mask, which can be used to fabricate a semiconductor device, to transfer a fine pattern for contact holes, conductors, etc., onto a wafer, the following method is also known to give the mask a greater depth of focus at the time of exposure for the transfer: a method for forming a primary pattern, which is a region that is actually resolved to correspond to the fine pattern, and an auxiliary pattern, which is not actually resolved onto a wafer, using a semi-transparent layer of the halftone phase-shift mask. According to this method, variation in the critical dimension (CD) of the pattern can be reduced when the pattern is defocused, since support of the diffraction light by the auxiliary pattern can improve the exposure latitude for the primary pattern region.

[0010] In modern wafer transfer techniques for a halftone phase-shift mask used to fabricate a semiconductor device, when transferring a fine pattern, as described above, for contact holes, traces, or other features onto the wafer, it is necessary to form a semi-transparent layer pattern in the halftone phase-shift mask. In this layer pattern, the width or length of each pattern element of a primary pattern, as described above, is specifically set to the range of 100 to 300 nm. Since the width or length of each pattern element of an auxiliary pattern, as described above, is larger in this case, variations in the critical dimension (CD) of the pattern can be further reduced when the pattern is defocused. However, if the width or length is too large, an undesired pattern will be resolved onto the wafer.If, for example, the width or length of the pattern element of the main pattern is set to the range of 100 to 300 nm, as described above, then the width or length of the pattern element of the auxiliary pattern is preferably set to 60 nm or less.

[0011] Furthermore, in a wafer process using a positive development to transfer a mask pattern of a halftone phase-shift mask, which is used to manufacture a semiconductor element, onto a wafer, in the case of forming a pattern for openings, such as contact holes or gaps, a main pattern and an auxiliary pattern, as described above, are formed onto a resist on the wafer, each as a concave element pattern in which a semi-transparent layer in the halftone phase-shift mask is partially hollowed out.In contrast, in a wafer development process using negative development to transfer a mask pattern of a halftone phase-shift mask, used to fabricate a semiconductor device, to a wafer, when forming a pattern for openings, such as contact holes or gaps, onto a resist on the wafer, it is necessary that a primary pattern and an auxiliary pattern, as described above, are each formed as a convex element pattern consisting of a semi-transparent layer in the halftone phase-shift mask. Accordingly, the auxiliary pattern becomes a convex element pattern consisting of one or more pattern elements, each consisting of the semi-transparent layer and having a width or length of 60 nm or less.

[0012] When using a halftone phase-shift mask to fabricate a complex semiconductor device, removing foreign materials through cleaning is a critical issue. In one field where this technique is employed, the following method is specifically used for the physical removal of foreign material: a process in which, during cleaning the halftone phase-shift mask, ultrasonic waves are applied to a chemical cleaning solution to utilize impacts based on bubble breakup.

[0013] However, if the power of the ultrasonic waves is increased to achieve a high range, a problem remains: the waves damage a fine convex element pattern, as described, which consists of a semi-transparent layer. It is assumed that the damage to the convex element pattern is primarily caused by a downward pulling force exerted on it by impacts generated when the bubbles produced by the ultrasonic waves are broken. Accordingly, if the thickness of the semi-transparent layer is large, the convex element pattern covers a large area. When bubbles of the same density are generated in the cleaning solution, a region of the convex element pattern becomes broad where the bubbles are generated.This situation causes a problem: the damage to the convex element pattern increases, leading to its flaking. Furthermore, if the thickness of the semi-transparent layer is large, the convex element pattern also experiences impacts from bubbles generated at a higher level. This increases the impact time, causing significant damage to the convex element pattern and ultimately leading to flaking. Therefore, reducing the thickness of the semi-transparent layer is a very effective way to prevent damage to the convex element pattern during ultrasonic cleaning.

[0014] In a procedure for calculating optical near-field correction (OPC processing) for a mask pattern, an approximation was primarily used. When the accuracy of this approximation needed to be increased, a time-domain finite difference (FDTD) method was occasionally employed to compute an exact solution. The FDTD method for calculating an exact solution involves developing a Maxwell equation directly into a differential equation for space-time regions to perform a sequential computation, thereby determining electric and magnetic fields. It is a method for performing a calculation that takes into account the thickness of a semi-transparent layer pattern. In the FDTD method for calculating an exact solution, a space region is divided into finite elements to perform a computation at each of the individual grid points within it.The calculation period depends on the region being calculated. Therefore, if this method is used to perform a calculation for the entirety of a mask, the calculation period becomes extremely long. Consequently, for calculations involving the entirety of a mask, without using any FOTO method to calculate an exact solution, an approximation is used.

[0015] The approximation method is a simplified approach that does not account for the thickness of a semi-transparent layer pattern, unlike the FDTD method for calculating an exact solution, and is not a method for performing an exact calculation. For example, in the case of calculating an optical near-field correction (OPC processing) for a phase-shift mask used to create a fine pattern on a wafer, an approximation method—unlike the FOTO method for obtaining an exact solution—cannot include a shielding effect caused by the thickness of the semi-transparent layer pattern in the calculation results. Previously, the thickness of a semi-transparent layer pattern was large, thus increasing the shielding effect caused by its thickness.As a result, in the conventional case of calculating optical near-field correction (OPC processing) for a phase-shift mask used to create a fine pattern on a wafer, the shielding effect caused by the layer thickness of its semi-transparent layer pattern is significantly influenced by an approximation, leading to a difference between results from any exact solution calculation and those from the approximation. This can cause the fine pattern produced on the wafer to be introduced into or separated from the intended area by the effect of the pattern where the approximate solution calculation is used to apply optical near-field correction (OPC processing).

[0016] Because of this, when designing a semi-transparent layer pattern in a phase-shift mask used to form a fine pattern on a wafer, and when an optical near-field correction (OPC) calculation is performed using an approximation, the formation of the semi-transparent layer pattern must be subjected to more stringent constraints to prevent contact or separation of the fine pattern on the wafer, contrary to the intended purpose. Consequently, the flexibility of the semi-transparent layer pattern formation is low.

[0017] Furthermore, according to patent literature 7, a photomask with an intermediate examination film layer is known from the prior art.

[0018] Patent literature 8 also describes a method for producing a lithography mask blank, a lithography mask and a halftone phase shift mask blank. Citation list for patent literature Patent literature 1: JP 2003 - 322 948 A Patent literature 2: JP 2009 - 217 282 A Patent literature 3: JP 2005 - 284 213 A Patent literature 4: JP 2010 - 009 038 A Patent literature 5: JP 2002 - 351 049 A Patent literature 6: JP 2008 - 310 091 A Patent literature 7: US 2004 / 0 043 303 A1 Patent literature 8: JP 2003 - 322 955 A Summary of the present invention Technical problem

[0019] The present invention was made in light of the above-mentioned actual situation. A main objective of this invention is to make a phase-shift mask pattern excellent with regard to its transfer properties and to give it high resistance to irradiation with ArF excimer laser exposure light and high cleaning resistance during photolithography, to prevent any flaking of a semi-transparent layer pattern in the phase-shift mask based on ultrasonic cleaning, and to increase the formation flexibility of the semi-transparent layer pattern in the phase-shift mask. Solution to the problem

[0020] To solve the above problem, the present invention provides a halftone phase-shift mask onto which ArF excimer laser exposure light is directed, wherein the mask comprises a transparent substrate and a semi-transparent layer pattern formed on the transparent substrate and consisting only of Si and N, or a semi-transparent layer pattern formed on the transparent substrate and consisting only of Si, N and O, characterized in that the semi-transparent layer pattern has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light, and a transmittance of 15% to 38% at the wavelength of the ArF excimer laser exposure light, and furthermore has a layer thickness of 57 nm to 67 nm, the semi-transparent Layer patterns have a main patternthat is to be resolved onto a wafer, and an auxiliary pattern used to assist in the resolution of the main pattern, which is not to be resolved onto the wafer, the auxiliary pattern comprising one or more convex pattern elements, each with a width or length of 60 nm or less, and the main pattern having the same thickness as the auxiliary pattern.

[0021] According to the present invention, the phase-shift mask is used in a method for producing a patterned body to reduce the light intensity to zero at its pattern element boundaries through light interference based on the phase effect, thereby improving the contrast of the resulting transferred image. When a patterned body is produced with this enhancement, the phase effect can be made more pronounced due to the high transmittance of the semi-transparent layer. Furthermore, the semi-transparent layer contains no metal; thus, no silicon (Si) oxide layer grows, even when ArF excimer laser exposure is applied to this layer for an extended period. Consequently, changes to the critical dimensions of the pattern are prevented.Similarly, during the cleaning step of the phase-shift mask, changes to the critical dimensions of the pattern can be prevented. Accordingly, the present invention enables the phase-shift mask pattern to be excellently designed with regard to its transfer properties in photolithography, and to be endowed with high resistance to irradiation with ArF excimer laser exposure light and high cleaning resistance.

[0022] Insofar as, according to the present invention, the semi-transparent layer pattern comprises a main pattern to be resolved onto a wafer and an auxiliary pattern used to assist the resolution of the main pattern, which is not to be resolved onto the wafer, and furthermore, the auxiliary pattern is a convex element pattern with one or more convex pattern elements (each) having a width or length of 60 nm or less, the phase-shift mask in the present invention can clearly exhibit an advantageous effect in preventing pattern flaking caused by cleaning the pattern using ultrasonic waves with an intense removal force in a cleaning solution.

[0023] Furthermore, in the present invention, the semi-transparent layer pattern is preferably formed directly on the transparent substrate. The phase-shift mask has no etch barrier layer between the transparent substrate and the semi-transparent layer; thus, it becomes unnecessary to perform an etching step multiple times. Accordingly, the etching process is not complicated, and it is also possible to prevent deterioration of the shape of the semi-transparent layer or the transparent substrate, as well as deterioration of the shape uniformity of the semi-transparent layer, these problems being based on the difficulty of etching the etch barrier layer.

[0024] Furthermore, in the present invention, the phase-shift mask is preferably a negative-type phase-shift mask. With respect to the semi-transparent layer, its transmittance is in the range of 15% to 38% at the wavelength of the ArF excimer laser exposure light, thus being higher than that of conventional phase-shift masks. This results in a greater phase effect during negative development at the edges of light-blocking layer regions that correspond to a fine pattern, such as that for contact holes. This makes it easier to transfer the fine pattern, such as that for the contact holes, onto the wafer via negative development than in conventional cases.

[0025] Furthermore, in the present invention, the phase-shift mask preferably comprises a light-blocking layer pattern formed on the semi-transparent layer pattern, wherein the optical density (OD value) of this light-blocking layer pattern is adjusted at the wavelength of the ArF excimer laser exposure light to yield a desired overall value (OD value) when combined with the optical density (OD value) of the semi-transparent layer pattern. This case, where the mask further comprises the light-blocking layer pattern, allows the mask to achieve a more advantageous effect when it has a large semi-transparent layer pattern element, thus avoiding the problem of the transferred image becoming indistinct due to exposure light transmitted through this pattern element.

[0026] Furthermore, in the present invention, the light-blocking layer pattern preferably comprises a single-layer structure comprising a light-absorbing layer pattern formed on top of the semi-transparent layer pattern. This layer serves as an etch barrier for the semi-transparent layer pattern and absorbs the ArF excimer laser exposure light. This configuration allows the light-blocking layer pattern to be formed by using a pattern created by etching the hard mask layer in the light-blocking layer with the aforementioned double-layer structure, instead of a resist pattern that would otherwise be required when etching the light-absorbing layer. As a result, a fine pattern of the light-blocking layer pattern is easily formed.

[0027] Furthermore, in the present invention, the light-blocking layer pattern preferably has a bilayer structure comprising: an etch barrier layer pattern formed on top of the semi-transparent layer pattern, which serves as an etch barrier for the semi-transparent layer pattern; and a light-absorbing layer pattern formed on top of the etch barrier layer pattern, which serves to absorb the ArF excimer laser exposure light. This arrangement allows the use of a pattern formed by etching the hard mask layer in the light-blocking layer with the aforementioned three-layer structure, instead of a resist pattern that would otherwise be required when etching the light-absorbing layer. As a result, a fine pattern of the light-blocking layer can be easily formed.Regarding the raw material of the layer with a light-absorbing function for absorbing the ArF excimer laser exposure light, the selection flexibility is increased, allowing for a smaller layer thickness of the light-blocking layer pattern. Furthermore, the light-absorbing layer and the etch barrier layer can each be etched with different reactive etching gases. Thus, the etch barrier layer can be designed to act as an etch barrier for the semi-transparent layer, advantageously preventing damage to the semi-transparent layer when the light-absorbing layer is etched.

[0028] Furthermore, in the present invention, the light-absorbing layer pattern preferably comprises simple silicon (Si). This makes it possible to prevent the light-absorbing layer from being damaged when the hard mask layer is etched.

[0029] Furthermore, in the present invention, the optical density (OD value) of the light-blocking layer pattern is preferably set at the wavelength of the ArF excimer laser exposure light, resulting in an overall value of 3.0 or more when combined with the optical density (OD value) of the semi-transparent layer pattern. This makes it possible to impart the necessary light-blocking property to a desired region of the mask when the mask is exposed to light.

[0030] As mentioned above, the present invention also provides a method for producing a pattern-formed body. The method for producing a pattern-formed body of the present invention comprises a step of forming a resist pattern by exposing a resist layer of a positive-type resist composition with ArF excimer laser exposure light using the phase-shift mask of the present invention, and resolving and removing an unexposed region by negative development.

[0031] According to the present invention, the transmittance of the semi-transparent layer is in the range of 15% to 38% at the wavelength of the ArF excimer laser exposure light, thus being higher than that of conventional semi-transparent layers. This results in a greater phase effect during negative development at the edges of light-blocking layer regions that correspond to a fine pattern, such as that for contact holes. This makes it easier to transfer the fine pattern, such as that for the contact holes, onto the wafer via negative development than in conventional cases. Advantageous effects of the present invention

[0032] The present invention produces advantageous effects, enabling the pattern of a phase-shift mask to be excellently formed in terms of transfer properties and to give it high resistance to irradiation with ArF excimer laser exposure light and high cleaning resistance in photolithography, avoiding any flaking of a semi-transparent layer pattern in the phase-shift mask based on ultrasonic cleaning and increasing the formation flexibility of the semi-transparent layer pattern in the phase-shift mask. Brief description of the drawings Fig. Figure 1 is a schematic sectional view illustrating an example of a mask blank. Fig.Figure 2 is a graphical representation that represents the layer thickness (nm) of a semi-transparent layer required to generate a counter-phase with a retardation of 180° relative to the respective values ​​of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at a wavelength of the ArF excimer laser exposure light. Fig. Figure 3 is a graphical representation that represents the transmittance (%) of a semi-transparent layer formed with a layer thickness necessary to generate a counter-phase with a retardation of 180°, relative to the respective values ​​of the extinction coefficient (k) and the refractive index (n) of the semi-transparent layer at a wavelength of the ArF excimer laser exposure light. Fig.Figure 4 is a graphical representation that shows the value of the extinction coefficient (k) and the value of the refractive index (n) of each of the known materials at the wavelength of the ArF excimer laser exposure light. Fig. Figure 5 is a schematic sectional view illustrating another example of a mask blank. Fig. Figure 6 is a schematic sectional view illustrating another example of a mask blank. Fig. Figure 7 is a schematic sectional view illustrating another example of a mask blank. Fig. Figure 8 is a schematic sectional view illustrating another example of a mask blank. Fig. Figure 9 is a schematic sectional view illustrating an example of a mask blank attached to a negative type resist layer. Fig.Figure 10 is a schematic sectional view illustrating an example of the phase shift mask of the present invention. Fig. 11 is an AA sectional view of Fig. 10. Fig. Figure 12 is a schematic top view illustrating another example of the phase shift mask of the present invention. Fig. 13 is an AA sectional view of Fig. 12. Fig. Figures 14A to 14C are schematic process diagrams illustrating an example of a method for producing a patterned body using the phase shift mask of the present invention. Fig. Figure 15 is a graphical representation that represents OPC bias value simulation results versus transmittance. Fig.Figure 16 is a representation showing XY images of exposure light intensity distributions, each on a wafer, the distributions being obtained by a simulator, and further showing graphs representing the respective exposure light intensities of the distributions. Fig. 17-1 is a graphical representation which, as simulation results, represents a relationship between the focus depth and the exposure latitude of a phase shift mask of Example 1 and Comparison Example 1 respectively, when a 180 nm hole spacing pattern of the mask is transferred. Fig. 17-2 is a graphical representation which, as simulation results, represents a relationship between the focus depth and the exposure latitude of a phase shift mask of Example 1 and Comparison Example 1 respectively, when a 240 nm hole spacing pattern of the mask is transferred. Fig.17-3 is a graphical representation which, as simulation results, represents a relationship between the focus depth and the exposure latitude of a phase shift mask of Example 1 and Comparison Example 1 respectively, when a 300 nm hole spacing pattern of the mask is transferred. Fig. Figure 18 is a graphical representation that represents the contrast of wafer-transferred optical spatial images resulting from the calculation under conditions that assume a phase shift mask where the transmittance of the semi-transparent layer is 38% and a phase shift mask where the transmittance of the semi-transparent layer is 6%. Fig.Figure 19 is a graphical representation that represents the OPC bias of a phase shift mask resulting from the calculation under conditions that assume the phase shift mask where the transmittance of the semi-transparent layer is 38% and the phase shift mask where the transmittance of the semi-transparent layer is 6%. Description of embodiments

[0033] A detailed description of the phase-shift mask of the present invention is given, as well as of the method for producing a body formed by a pattern using the same invention. For this purpose, a mask blank, which serves to produce the phase-shift mask of the present invention, is also described below. A. Mask blank

[0034] The mask blank is a mask blank used to produce the halftone phase-shift mask of the present invention, onto which ArF excimer laser exposure light is directed, and this mask comprises a transparent substrate and a semi-transparent layer formed on the transparent substrate and consisting only of Si (silicon) and N (nitrogen), or a semi-transparent layer formed on the transparent substrate and consisting only of Si (silicon), N (nitrogen), and O (oxygen), and is characterized in that the semi-transparent layer has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light, and a transmittance of 15% to 38% at the wavelength of the ArF excimer laser exposure light, and furthermore has a layer thickness of 57 nm to 67 nm.

[0035] Fig. Figure 1 is a schematic sectional view illustrating an example of the mask blank. A Fig. One illustrated mask blank 100 is a mask blank used to produce the halftone phase-shift mask onto which the ArF excimer laser exposure light is directed. The one in Fig.The illustrated mask blank 100 comprises a transparent substrate 101 and a semi-transparent layer 102, which is formed on the transparent substrate 101 and has a single-layer structure consisting only of Si and N, or Si, N, and O. The semi-transparent layer 102 has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at a wavelength of the ArF excimer laser exposure light, and a transmittance of 15 to 38% at a wavelength of the ArF excimer laser exposure light. The semi-transparent layer 102 also has a thickness of 57 to 67 nm.

[0036] In the mask blank, the semi-transparent layer has a high transmittance in the range of 15 to 38% of the transmittance at the wavelength of the ArF excimer laser exposure light. Accordingly, a phase-shifting mask fabricated from the mask blank is used to reduce the light intensity to zero at its pattern element boundaries through light interference based on the phase effect, thereby improving the contrast of the resulting transferred image. When a patterned object is created with this enhancement, the phase effect can be made more pronounced due to the high transmittance of the semi-transparent layer. Furthermore, the semi-transparent layer contains no metal; thus, no silicon (Si) oxide layer forms, even when ArF excimer laser exposure light is applied to this layer for an extended period.Accordingly, changes to the critical dimension of the pattern can be prevented. Similarly, changes to the critical dimension of the pattern can be prevented during the cleaning step of the phase-shift mask. Therefore, the present invention enables photolithography to achieve excellent design of the phase-shift mask pattern with respect to its transfer properties and to impart to it high resistance to irradiation with ArF excimer laser exposure light and high cleaning resistance.

[0037] Regarding the mask blank, the thickness of the semi-transparent layer is in the range of 57 to 67 nm, in order to be smaller than that of conventional semi-transparent layers. Thus, in a halftone phase-shift mask produced from the mask blank, the height of a convex element pattern, consisting of one or more pattern elements that (each) consist of the semi-transparent layer and have a width or length of, for example, 60 nm or less, is lower than that in conventional halftone phase-shift masks.This leads, for example, to the following in the cleaning step of the phase-shift mask produced from the mask blank: a reduction in the area of ​​a region of the convex element pattern that absorbs impacts from bubble breakup when the convex element pattern is cleaned by ultrasonic waves; and a lowering of the level of the position of the convex element pattern that absorbs the impacts from bubble breakup. As a result, for example, flaking of the convex element pattern can be avoided, where flaking is caused by cleaning the convex element pattern using ultrasonic waves with a strong displacement force in a cleaning solution.

[0038] Such an expression, that flaking of the pattern is caused by ultrasonic wave cleaning, means that the pattern, which is in a convex element shape, collapses or is lost due to the downward pulling force generated by impacts caused when bubbles generated by the ultrasonic waves break up.

[0039] Regarding the mask blank, the thickness of the semi-transparent layer is in the range of 57 to 67 nm, which is smaller than that of conventional semi-transparent layers. When using a phase-shift mask fabricated from the mask blank to form a fine pattern on a wafer, the following can be avoided when calculating near-field optical correction (OPC) for the phase-shift mask: the random error of a correction applied to the data when the mask is formed is significantly increased, as in the prior art, by a shielding effect caused by the thickness of the semi-transparent pattern.Avoiding this makes it possible to limit the occurrence of a problem where the fine pattern on the wafer is brought into contact with or separated from any other part contrary to the training intention.

[0040] Because of this, when using a phase-shift mask fabricated from the mask blank to form a fine pattern on a wafer, it prevents the fine pattern from coming into contact with or separating from any other part of the wafer contrary to the intended design, even if optical near-field correction (OPC) processing is performed using an approximation during the formation of the mask's semi-transparent layer pattern. This reduces the need to impose stricter constraints on the formation of the semi-transparent layer pattern, even when the OPC calculation is performed using an approximation. This increases the flexibility of the semi-transparent layer pattern formation.

[0041] Regarding the mask blank, the thickness of the semi-transparent layer is in the range of 57 to 67 nm, which is smaller than that of conventional semi-transparent layers. Therefore, it is also easier to form the semi-transparent layer pattern by etching on this mask blank than on conventional mask blanks. The time required for etching is also sufficiently short. Even if the mask blank does not have an etch barrier layer between the transparent substrate and the semi-transparent layer to prevent damage to the transparent substrate, as described below, damage to the transparent substrate during the etching process can be adequately avoided.

[0042] The following section describes parts of the mask blank and the structure of the mask blank separately. 1. Parts of the mask blank

[0043] First, the components of the mask blank are described. The mask blank comprises at least one transparent substrate and one semi-transparent layer. (1) Semi-transparent layer

[0044] The semi-transparent layer in the present invention is a semi-transparent layer formed on a transparent substrate, which is described in detail below, and consists only of Si and N, or a semi-transparent layer formed on a transparent substrate, which is described in detail below, and consists only of Si, N, and O. The semi-transparent layer has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light, and a transmittance of 15 to 38% at the wavelength of the ArF excimer laser exposure light, and furthermore has a layer thickness of 57 to 67 nm. a. Transmittance and layer thickness

[0045] The transmittance of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light is selected from the range of 15 to 38%, which is higher than that of conventional semi-transparent layers at this wavelength. The thickness of the semi-transparent layer is selected from the range of 57 to 67 nm, which is less than that of conventional semi-transparent layers, making the layer durable for practical applications.

[0046] The reason the transmittance of the semi-transparent layer is selected within the specified range above for the wavelength of the ArF excimer laser exposure light is as follows: If the transmittance does not reach this range in a modern wafer process technique for transferring a mask pattern of a halftone phase-shift mask with a semi-transparent layer onto a wafer, the phase-shift mask cannot achieve the desired phase-shift effect. Specifically, the phase-shift mask is insufficient in the antiphase light transmittance, so the mask cannot achieve the desired light-blocking property. If the transmittance exceeds this range, the light-blocking property of the semi-transparent layer will be lower than required.Specifically, the transmittance of out-of-phase light increases, causing pattern elements to be formed unfavorably by the transmitted light, even in a region of the wafer where the formation of pattern elements is not desired. If the transmittance is not selected within this range, a problem with the transfer properties of a desired pattern will ultimately arise, as described here.

[0047] The semi-transparent layer is not particularly restricted. The layer is preferably a semi-transparent layer with a transmittance of 18 to 38%, and more preferably one with a transmittance of 20 to 38%. If the transmittance does not reach this range, the resulting phase-shift mask may not clearly achieve the desired phase-shift effect. If the transmittance exceeds this range, the light-blocking property of the semi-transparent layer will be less pronounced than required, thus causing a problem with the transfer properties of a desired pattern.

[0048] The reason the layer thickness is selected from the range specified above is as follows: If the height of the convex element pattern, consisting of one or more pattern elements (each) made of the semi-transparent layer and with a width or length of 60 nm or less, exceeds this range, then, during the cleaning step of a phase-shift mask produced from the mask blank, this convex element pattern is cleaned using ultrasonic waves with a strong removal force in a cleaning solution, causing the convex element pattern to flake off. If the thickness of the semi-transparent layer exceeds this range, it becomes difficult to form the semi-transparent layer pattern by etching in the same way as in the prior art. The time required for etching will not be short.If the mask blank lacks an etch barrier layer between the transparent substrate and the semi-transparent layer, the transparent substrate will be damaged when the semi-transparent layer pattern is formed by etching. Furthermore, if the layer thickness does not reach the required range, it will be difficult to create a counter-phase in the semi-transparent layer with an extinction coefficient (k) in the range specified above, as described in detail below.

[0049] Furthermore, selecting a layer thickness within the range specified above allows for the following advantageous effects. The semi-transparent layer consisting only of Si and N, or the semi-transparent layer consisting only of Si, N, and O, is a layer in which a residual defect, an unnecessary superfluous part, is not easily corrected by an etching removal process based on focused ion beam (FIB) radiation using, for example, Ga ions, which are used to correct an existing ordinary photomask, or by an etching removal process based on electron beam (EB) radiation. In the mask blank, however, the layer thickness of the semi-transparent layer is in the range of 57 to 67 nm, to be smaller than that of conventional semi-transparent layers.Although the semi-transparent layer consists only of Si and N, or only of Si, N, and O, it can more easily correct its residual defect area than conventional semi-transparent layers composed of only the same material. Furthermore, if the thickness of the semi-transparent layer is small, the lateral etching that occurs during etch removal is reduced. This improves the precision of the corrected position, thus further enhancing the cross-sectional shapes of the semi-transparent layer. Consequently, the halftone phase-shift mask with the corrected semi-transparent layer exhibits improved transfer properties.Even if a residual defect, which is an unnecessary superfluous part, forms in the semi-transparent layer pattern of a phase-shift mask produced from the mask blank, in a region where narrow gaps and holes are to be formed, each with a smaller width than in the prior art, the residual defect can be corrected more easily in these narrow gaps and holes than if the same residual defect were formed in conventional semi-transparent layer patterns made of the same material. If the correction is to be made by a mechanical method, such as a method of shearing off a foreign material with a fine needle, the fine needle is specifically placed lightly between pattern elements; thus, a residual defect in such narrow gaps or holes can be easily corrected.

[0050] The thickness of the semi-transparent layer is not particularly restricted. The layer thickness is preferably in the range of 57 to 64 nm, and more preferably from 57 to 62 nm. Since the thickness of the semi-transparent layer is smaller, flaking of the pattern can be significantly reduced. This flaking is a phenomenon where a convex element pattern, as described above, flakes off when cleaned with ultrasonic waves using a strong force in a cleaning solution. Alternatively, the semi-transparent layer is easier to process, or the semi-transparent layer pattern is easier to correct. b. Extinction coefficient and refractive index, and material

[0051] The respective ranges of the extinction coefficient and the refractive index of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light are calculated as ranges that allow the layer thickness of the semi-transparent layer required to achieve an opposite phase to be set in the range of 57 to 67 nm, and which further allow the transmittance of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light to be set in the range of 15 to 38%. The semi-transparent layer is a layer selected as a new, non-metal semi-transparent layer from a semi-transparent layer consisting only of Si and N, and from a semi-transparent layer consisting only of Si, N, and O.The ranges of the extinction coefficient and the refractive index of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light are obtained as respective ranges of the coefficient and the index that the semi-transparent layer consisting only of Si and N, or only of Si, N and O, can exhibit.

[0052] The following describes how the above-mentioned respective ranges of the extinction coefficient and the refractive index of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light and the raw material of the semi-transparent layer were obtained and selected.

[0053] Initially, Fresnel's formula was used to calculate the thickness of a semi-transparent layer required to generate an antiphase with respect to the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light. In the present invention, the term "antiphase" means that in a halftone phase-shift mask with a transparent substrate and a semi-transparent layer pattern formed on the transparent substrate, the following retardation occurs in the range of 160 to 200°: the retardation between the ArF excimer laser exposure light, which has a wavelength of 193 nm and is transmitted only through the transparent substrate, and the laser exposure light, which is transmitted through the transparent substrate and the semi-transparent layer pattern.

[0054] Fig.Figure 2 shows a graphical representation of the layer thickness (nm) of the semi-transparent layer required to generate an antiphase with a retardation of 180° relative to the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at a given wavelength of the ArF excimer laser exposure light. For example, the layer thickness of the semi-transparent layer required to produce an antiphase with a retardation of 180°, as shown in Fig.Figure 2 shows the values ​​of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light. The thickness of the semi-transparent layer required to generate an antiphase with a retardation other than 180° is also obtained in the same way, by comparing the values ​​of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light.

[0055] Next, calculations were made from the extinction coefficient (k) regarding the transmittance of a semi-transparent layer at the wavelength of the ArF excimer laser exposure light, where this semi-transparent layer is a layer which was formed for a layer thickness which is necessary to give an opposite phase to the respective values ​​of the extinction coefficient (k) and the refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light. Fig.Figure 3 shows a graphical representation of the transmittance (%) of the semi-transparent layer formed at a thickness required to achieve an antiphase with a retardation of 180°, relative to the respective values ​​of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light. For example, the transmittance of the semi-transparent layer formed at a thickness required to achieve an antiphase with a retardation of 180°, as shown in Figure 3, is... Fig.Figure 3 shows the values ​​of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light. The transmittance of the semi-transparent layer, formed in a thickness necessary to produce an antiphase with a retardation other than 180°, is obtained in the same way, compared to the values ​​of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light.

[0056] Next, the following was obtained, calculated from the results of the calculations for the thickness of the semi-transparent layer required to generate an antiphase with respect to the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light, and the results of the calculations for the transmittance, at the wavelength of the ArF excimer laser exposure light, of the semi-transparent layer formed to a thickness required to produce an antiphase: respective ranges of the extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light, whereby these ranges allow the layer thickness required to produce an antiphase to lie in the range of 57 to 67 nm, and allowthat the transmittance of the layer at the wavelength of the ArF excimer laser exposure light is in the range of 15 to 38% when the layer is formed to the thickness necessary to produce the opposite phase.

[0057] As a result, a range of extinction coefficient (k) from 0.2 to 0.45 and a range of refractive index (n) from 2.3 to 2.9 were obtained as the respective ranges of extinction coefficient (k) and refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light, whereby these ranges allow the layer thickness necessary to generate an antiphase to be in the range of 57 to 67 nm, and allow the transmittance of the layer at the wavelength of the ArF excimer laser exposure light to be in the range of 15 to 38% when the layer is formed at the thickness necessary to produce the antiphase.

[0058] Next, known materials were examined and selected to obtain a material that contains no metal and has an extinction coefficient of 0.2 to 0.45 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.3 to 2.9 at the wavelength of the ArF excimer laser exposure light.

[0059] Fig.Figure 4 shows a graph representing the extinction coefficient (k) and refractive index (n) values ​​of each of the known materials at the wavelength of the ArF excimer laser exposure light. The respective extinction coefficient (k) and refractive index (n) values ​​of the known materials represented in this graph are available in non-patent literature (Refractive Index List / Refractive Index List for Thin Film Measurement [online]. [Accessed 2014-07-03]. Retrieved from the Internet:<URL: http: / / www.filmetricsinc.jp / refractive-index-database> ) and patent literature (JP 2007 - 017 998 A). Table 1 shows the respective values ​​of the extinction coefficient (k) and the refractive index (n) of the known materials described in the Fig.4 shown in the graphic representation; and the respective partial composition ratios (in atomic %) of Si (silicon), N (nitrogen) and O (oxygen) in the composition of each of the known materials. [Table 1-1] n k AG 1,03 1,18 Al 0,11 2,22 Al2O3 0,14 1,65 Mon 0,79 2,34 Ni 1,01 1,46 Si 0,85 2,73 Si3N4 2,70 0,20 SiO2 1,56 0,00 Ta 1,32 2,26 Cr 1,43 1,70 MoSi 2,34 0,59 SiO 1,80 0,72 [Table 1-2] n k Composition ratio (at%) Si N O Material A based on SiO 2,22 1,37 60,7 0,0 39,3 Material A based on SiON 1,68 0,01 33,4 4,7 61,9 Material B based on SiON 1,92 0,04 36,3 15,0 48,7 Material C based on SiON 2,16 0,21 38,9 29,6 31,5

[0060] As in Fig. Figure 4 and Table 1 show that Si3N4 is a known material used for a semi-transparent layer, and the extinction coefficient (k) and refractive index (n) values ​​of Si3N4 are 2.70 and 0.20, respectively. Si3N4 is a metal-free material with an extinction coefficient (k) of 0.2 to 0.45 and a refractive index (n) of 2.3 to 2.9.

[0061] As in Fig.Figure 4 and Table 1 show that known materials containing Si, O, or N are: Material A based on Si (silicon), SiO₂, SiO₂, SiO₂, SiO₂, Material A based on SiON₄, Material B based on SiON₄, and Material C based on SiON₄. The respective values ​​of the extinction coefficient (k) and refractive index (n) of the known materials, and the respective partial composition ratios (in atomic %) of Si, N, and O in the composition of each of the known materials, are as shown in Figure 4 and Table 1. Fig. 4 and Table 1 shown.

[0062] Furthermore, in a material containing only Si and N, such as the compound Si3N4, or in a material containing only Si, N, and O, such as SiON-based material A, SiON-based material B, or SiON-based material C, the values ​​of the extinction coefficient (k) and the refractive index (n) at the wavelength of the ArF excimer laser exposure light change when the partial composition ratio of Si, N, or O is altered. For each of these materials, it is clear how the values ​​of the extinction coefficient (k) and the refractive index (n) tend to change according to a change in the partial composition ratio of Si, N, or O.

[0063] According to these circumstances, in the material containing only Si and N, or in the material containing only Si, N and O, the value of the extinction coefficient (k) and the value of the refractive index (n) thereof at the wavelength of the ArF excimer laser exposure light can be changed within the range of 0.2 to 0.45 and that of 2.3 to 2.7, respectively, by changing the partial composition ratio of Si, that of N or that of O.

[0064] Under these circumstances, a material containing only Si and N, or a material containing only Si, N, and O, was selected from known materials as the material for the semi-transparent layer. As a result, the respective ranges of the extinction coefficient (k) and the refractive index (n) of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light were obtained as the ranges that a semi-transparent layer consisting only of Si and N, or only of Si, N, and O, can exhibit. It was thus determined that the extinction coefficient (k) and the refractive index (n) lie in the range of 0.2 to 0.45 and 2.3 to 2.7, respectively. C. Semi-transparent layer

[0065] The semi-transparent layer can have a single-layer or multi-layer structure, and preferably a single-layer structure, since the effects and advantages of the present invention can be easily achieved. Specifically, this simpler structure allows the resulting phase-shift mask pattern to be excellently formed in terms of transfer properties and to give it high resistance to irradiation with ArF excimer laser exposure light and high cleaning resistance, as described above, in order to prevent flaking of the pattern and, furthermore, to simplify the processing of the semi-transparent layer.

[0066] The semi-transparent layer consisting only of Si and N is a layer that contains essentially no other element besides Si and N. However, the layer may contain any impurity, provided that the impurity does not impair the functions or properties of the semi-transparent layer. The functions and properties of the semi-transparent layer include the extinction coefficient and the refractive index of the layer at the wavelength of the ArF excimer laser exposure light. Examples of impurities that do not impair the functions or properties of the semi-transparent layer include carbon, oxygen, boron, helium, hydrogen, argon, and xenon.The proportion of impurities that do not impair the functions or properties of the semi-transparent layer is preferably 5% or less, more preferably 2% or less, and in particular preferably 1% or less.

[0067] The semi-transparent layer consisting only of Si, N, and O is a semi-transparent layer that contains essentially no other element besides Si, N, and O. However, the layer may contain any impurity, provided that the impurity does not impair the functions or properties of the semi-transparent layer. The functions and properties of the semi-transparent layer are equivalent to those of the semi-transparent layer consisting only of Si and N. The type and proportion of the impurity, which does not impair the functions or properties of the semi-transparent layer consisting only of Si, N, and O, in this semi-transparent layer are equivalent to those in the semi-transparent layer consisting only of Si and N.

[0068] The semi-transparent layer is a semi-transparent layer consisting only of Si and N, for which the refractive index specified above is in the range of 2.3 to 2.7 (high-refractive-index SiN layer), or a semi-transparent layer consisting only of Si, N, and O, for which the refractive index specified above is in the range of 2.3 to 2.7 (high-refractive-index SiON layer). For example, the high-refractive-index SiN layer and the high-refractive-index SiON layer have a higher refractive index than any semi-transparent layer consisting only of Si, N, and O, for which the refractive index specified above is less than 2.3 (low-refractive-index SiON layer).This case makes it possible to reduce the thickness of the semi-transparent layer, which is a thickness to achieve a retardation of 180° at the wavelength of the ArF excimer laser exposure light.

[0069] The following section provides a more detailed description of why the high-refractive-index SiN-based layer and the high-refractive-index SiON-based layer are preferred over the low-refractive-index SiON-based layer. Table 2 shows the properties of each of the following: a high-refractive-index SiN-based layer, a high-refractive-index SiON-based layer, a low-refractive-index SiON-based layer, and a conventional semi-transparent MoSiON-based layer. [Table 2] Layer type Characteristics Blank: Sheets- Reini- Resistant- Quality of external appearance resistance of the fine pattern resistance to contamination resistance to irradiation with exposure light SiN-based layer with high refractive index ◯ ⊚ ◯ ⊚ SiON-based layer with high refractive index ◯ ⊚ ◯ ⊚ SiON-based layer with low refractive index △ ◯ ◯ ⊚ MoSiON-based layer ◯ △ △ △

[0070] As shown in Table 2, the layer based on high refractive index SiN or the layer based on high refractive index SiON has better flaking resistance of the fine pattern than the layer based on low refractive index SiON for the following reason: the layer based on low refractive index SiON has a low refractive index, so the layer has a large thickness to achieve a 180° retardation; the layer based on high refractive index SiN or the layer based on high refractive index SiON, on the other hand, can have a high refractive index, so this layer with high refractive index can be made small in thickness to achieve a 180° retardation.The MoSiON-based layer has a greater thickness to achieve a 180° retardation than the low-refractive-index SiON-based layer, resulting in poor fine pattern peel resistance for the MoSiON-based layer.

[0071] The following section provides a further description of the effects and advantages gained from the fact that the layer based on high refractive index SiN or high refractive index SiON exhibits good flaking resistance of the fine pattern. To improve the resolving power in a wafer transfer process for a halftone phase-shift mask pattern, obtained from a mask blank as described above, it is generally necessary to increase the transmittance of its halftone layer. To increase the transmittance of the halftone layer, a method is employed to minimize its extinction coefficient. However, minimizing the extinction coefficient requires a reduction in the refractive index of the halftone layer to increase its thickness.

[0072] Increasing the thickness of the semi-transparent layer makes it essential, when a pattern of the semi-transparent layer is finely formed in the phase-shift mask, to clean the pattern consisting of the semi-transparent layer, which is a convex element pattern and has become highly reflective, in a cleaning solution using intense ultrasonic waves during the phase-shift mask cleaning step. This can cause pattern flaking or other problems. Pattern flaking or other problems become particularly noticeable when aiming for high resolution with a maximum limit, such as a resolution for forming a 10 nm knot pattern in the wafer. As described below, especially under point "D.In a process for producing a patterned body using a phase-shift mask, it is necessary in a wafer process to transfer a mask pattern of a phase-shift mask, as described above, onto a wafer by negative development. This requires forming not only a primary pattern, corresponding to a fine pattern for contact holes, conductors, or other features, in the phase-shift mask to be actually resolved, but also an auxiliary pattern, located close to the primary pattern and not actually resolved, as convex element patterns consisting of the semi-transparent layer. As described below in section "D.In the method described for producing a patterned body using a phase-shifting mask, such an auxiliary pattern is a convex element pattern consisting of one or more pattern elements, each formed from a semi-transparent layer and having a width or length of 60 nm or less in modern wafer processing techniques. For example, in this auxiliary pattern, or any other convex element pattern consisting of one or more pattern elements, each formed from a semi-transparent layer and having a width or length of 60 nm or less, pattern peeling, as described above, or similar problems become particularly apparent.

[0073] In other words, to improve the resolving power in the wafer process, it is necessary to maximize the transmittance of the semi-transparent layer. However, increasing the transmittance of the semi-transparent layer also increases its thickness; thus, promoting an increase in the transmittance of the semi-transparent layer conflicts with promoting a decrease in its thickness, which is necessary to prevent pattern peeling or other problems in the phase-shift mask where the semi-transparent layer pattern is finely formed.

[0074] In contrast, as described above, the high-refractive-index SiN or high-refractive-index SiON layer can have a high refractive index, allowing it to be thinner than the low-refractive-index SiON layer to achieve good flaking resistance of the fine pattern. Furthermore, the high-refractive-index SiN or high-refractive-index SiON layer can promote increased transmittance in the same way as the low-refractive-index SiON layer by adjusting the proportions of Si, N, or O.

[0075] Accordingly, a layer based on high refractive index SiN or high refractive index SiON exhibits good resistance to pattern peeling, thus preventing pattern peeling and other problems more effectively than a layer based on low refractive index SiON. Particularly when targeting high resolution with a maximum limit, such as the resolution required to create a 10 nm knot pattern on a wafer, a high refractive index layer offers superior resistance to pattern peeling and other issues compared to a layer based on low refractive index SiON.This makes it possible for the layer based on high refractive index SiN or the layer based on high refractive index SiON to achieve the high resolution with the maximum limit that cannot be achieved by the layer based on low refractive index SiON due to the occurrence of pattern peeling or other problems.

[0076] If the auxiliary pattern is, in particular, a convex element pattern formed from one or more pattern elements, each consisting of the semi-transparent layer and having a width or length of 60 nm or less in modern wafer processing techniques, the high refractive index SiN-based layer or the high refractive index SiON-based layer can better prevent pattern peeling or other problems in a fine auxiliary pattern than the low refractive index SiON-based layer, since the high refractive index layer exhibits good peel resistance of the fine pattern.This fact allows the layer based on high refractive index SiN or the layer based on high refractive index SiON to achieve a high resolution that cannot be achieved by the layer based on low refractive index SiON due to the occurrence of pattern peeling or other problems.

[0077] Accordingly, a layer based on high refractive index SiN or a layer based on high refractive index SiON can better accommodate both an increase in the transmittance of the semi-transparent layer and a reduction in the thickness of that layer than a layer based on low refractive index SiON, thereby achieving good resistance to flaking of the fine pattern. Thus, among the semi-transparent layers usable in the present invention, the high refractive index layer can achieve the highest resolving power.

[0078] As shown in Table 2, the layer based on high refractive index SiN or the layer based on high refractive index SiON exhibits a better external appearance of the mask blank than the layer based on low refractive index SiON. This is because the layer based on high refractive index SiN or the layer based on high refractive index SiON contains a lower proportion of oxygen (O) than the layer based on low refractive index SiON, resulting in greater stability of the external appearance during formation. In contrast, the layer based on low refractive index SiON contains a higher proportion of oxygen (O), making it more susceptible to the formation of impurities during formation and thus exhibiting lower stability of the external appearance.Although the layer based on MoSiON contains oxygen in its composition, the quality of the outer appearance of the mask blank is well developed through repeated improvements under different production conditions.

[0079] As shown in Table 2, each of the high-refractive-index SiN-based layer, the high-refractive-index SiON-based layer, and the low-refractive-index SiON-based layer contains no transition metal, thus exhibiting higher chemical resistance than the MoSiON-based layer, which exhibits better cleaning resistance. Furthermore, it is not easily denatured by exposure to light, thus exhibiting good resistance to light exposure. In contrast, the MoSiON-based layer contains a transition metal, which causes it to vary in dimensions, retardation, transmittance, and other properties after repeated cleaning. Therefore, this layer has poor cleaning resistance.Furthermore, the transition metal is replaced by the exposure to light, so that the layer is progressively oxidized, to vary in dimensions and other properties, to exhibit poor resistance to exposure to light.

[0080] The partial composition ratio of Si and that of N in the semi-transparent layer consisting only of Si and N, and the partial composition ratio of Si, that of O and that of N in the semi-transparent layer consisting only of Si, N and O, can be adjusted by appropriately selecting the following when the layers are each formed by a known layer formation process, such as sputtering, as described below: a device and materials to be used, layer formation conditions or other.In the case of using, for example, a parallel-plate GS magnetron sputtering device to form the semi-transparent layer on a transparent substrate by sputtering, the settings can be adjusted by appropriately selecting a target, the distance between the target and the transparent substrate (TS distance), the ratio between the gas flow rates, the gas pressure, the electrical energy (current) to be applied and the rotational speed of the substrate, a sputtering gas such as Ar, N2 or O2, and other sputtering process conditions.

[0081] The semi-transparent layer is not particularly restricted, provided the extinction coefficient defined above is in the range of 0.2 to 0.45, and is preferably a layer in which the extinction coefficient is in the range of 0.2 to 0.4, particularly preferably from 0.2 to 0.35. If the extinction coefficient does not reach this range, the semi-transparent layer is high in transmittance to achieve a low extinction coefficient and refractive index, thus resulting in an unfavorably large layer thickness necessary to produce a predetermined retardation at the wavelength of the ArF excimer laser exposure light. If the extinction coefficient exceeds this range, the semi-transparent layer is low in transmittance to avoid a desired phase shift effect at the wavelength of the ArF excimer laser exposure light.

[0082] The value of the extinction coefficient can be adjusted by adjusting the partial composition ratio of Si and that of N in the semi-transparent layer consisting only of Si and N, or the partial composition ratio of Si, that of O and that of N in the semi-transparent layer consisting only of Si, N and O.

[0083] The extinction coefficient of the semi-transparent layer can be measured and calculated using an ellipsometer VUV-VASE™, manufactured by JA Woolam Co., Inc.

[0084] The semi-transparent layer is not particularly restricted, provided the refractive index defined above is in the range of 2.3 to 2.7, and is preferably a layer in which the refractive index is in the range of 2.5 to 2.7, particularly preferably from 2.6 to 2.7. If the refractive index does not reach this range, the semi-transparent layer's thickness increases disadvantageously in order to achieve a retardation of 180° at the wavelength of the ArF excimer laser exposure light.

[0085] The value of the refractive index can be adjusted by adjusting the partial composition ratio of Si and that of N in the semi-transparent layer consisting only of Si and N, or the partial composition ratio of Si, that of O and that of N in the semi-transparent layer consisting only of Si, N and O.

[0086] The refractive index of the semi-transparent layer can be measured and calculated using the VUV-VASE™ ellipsometer, manufactured by JA Woolam Co., Inc. Alternatively, the refractive index can be measured using a method that simulates a reflection curve obtained by measuring the layer with a spectrometer.

[0087] The semi-transparent layer is not particularly restricted as long as the transmittance defined above is in the range of 15% to 18%, and is preferably a layer in which the transmittance is in the range of 18% to 38%, particularly preferably 20% to 38%. If the transmittance does not reach this range, the resulting phase-shift mask will not have the desired phase-shift effect. If the transmittance exceeds this range, the light-blocking property of the semi-transparent layer will be less pronounced than necessary to cause a problem with regard to the transfer properties of a desired pattern.

[0088] The transmittance value can be adjusted by adjusting the partial composition ratio of Si and N in the semi-transparent layer consisting only of Si and N, or the partial composition ratio of Si, O and N in the semi-transparent layer consisting only of Si, N and O.

[0089] Using, for example, an instrument for measuring the magnitude of the phase shift (MPM 193™, manufactured by Lasertec Corp.), measurements can be taken regarding the transmittance of the semi-transparent layer, the retardation between ArF excimer laser exposure light having a wavelength of 193 nm and transmitted only through the transparent substrate, and the same laser exposure light transmitted through the transparent substrate and the semi-transparent layer, and other properties.

[0090] The thickness of the semi-transparent layer is not particularly restricted, provided it is in the range of 57 nm to 67 nm, and is preferably in the range of 57 to 64 nm, particularly preferably from 57 to 62 nm, for the following reasons: Since the thickness of the semi-transparent layer is smaller, flaking of the pattern can be more significantly reduced. This flaking is a phenomenon in which a convex element pattern, as described above, flakes off when the pattern is cleaned with ultrasonic waves using a strong removal force in a cleaning solution. Alternatively, the semi-transparent layer is easier to process, or the semi-transparent layer pattern is easier to correct.

[0091] The thickness of the semi-transparent layer can be measured and calculated using the VUV-VASE™ ellipsometer, manufactured by JA Woollam Co., Inc.

[0092] The semi-transparent layer can be induced in reverse phase by adjusting the retardation between the ArF excimer laser exposure light, which has a wavelength of 193 nm and is transmitted only through the transparent substrate, and the same laser, which is transmitted through the transparent substrate and the semi-transparent layer, to a range of 160 to 200°. The semi-transparent layer is not particularly restricted. Its retardation is preferably between 170 and 190°, and more preferably 177°.When a phase-shift mask is formed from the mask blank by etching its semi-transparent film, this preferred retardation makes it possible to set the following retardation to 180° in the phase-shift mask, even when the transparent substrate is etched beneath a region of the semi-transparent layer where this layer is to be etched, in order to form an etched area: the retardation between the ArF excimer laser exposure light, which has a wavelength of 193 nm and is sent through a translucent region where the etched area is formed, and the same laser sent through a semi-transparent region where the semi-transparent layer remains. Accordingly, a halftone phase-shift mask can be produced. (2) Transparent substrate

[0093] The transparent substrate in the present invention is not particularly limited. Examples include synthetic quartz glass, fluorspar, and calcium fluoride, each of which is optically polished to transmit the exposure light with high transmittance. Among these examples, synthetic quartz glass is usually preferred, as it is used in many cases and exhibits stable quality and high transmittance for exposure light with a short wavelength. (3) Light-shielding layer

[0094] The mask blank is not particularly restricted with regard to its layer structure, material, and optical density (OD value) at the wavelength of the ArF excimer laser exposure light, provided that the mask blank comprises the semi-transparent layer and the transparent substrate defined above. The mask blank is preferably a mask blank that further comprises a light-blocking layer formed on the semi-transparent layer, and the optical density (OD value) of the light-blocking layer is adjusted to a desired optical density (OD value) at the wavelength of the ArF excimer laser exposure light when combined with the optical density (OD value) of the semi-transparent layer.

[0095] Fig. 5 is a schematic sectional view illustrating another example of the mask blank. One in Fig.Figure 5 illustrates a mask blank 100 comprising a transparent substrate 101; a semi-transparent layer 103 formed on the transparent substrate 101 and having a single-layer structure consisting only of Si and N, or Si, N, and O; and a light-blocking layer 103 with a single-layer structure formed on the semi-transparent layer 102. The optical density (OD value) of the light-blocking layer 103 is adjusted at the wavelength of the ArF excimer laser exposure light to be 3.0 or higher when combined with that of the semi-transparent layer 102. The light-blocking layer 103 performs all the functions of a light-absorbing function for absorbing the ArF excimer laser exposure light and an etching barrier function for the semi-transparent layer.

[0096] When a phase-shift mask fabricated from a mask blank contains a semi-transparent layer pattern with a large area, a problem can arise where the resulting transferred image becomes blurred due to exposure light passing through the pattern. Forming a light-shielding layer pattern on top of the semi-transparent layer in the phase-shift mask blocks an unnecessary component of the exposure light passing through this semi-transparent layer pattern, thus overcoming this problem. This problem becomes more pronounced when the transmittance of the semi-transparent layer is between 15% and 38% at the wavelength of the ArF excimer laser exposure light, as is evident in the mask blank, and is particularly high compared to a typical transmittance of 6%.Accordingly, the fact that the present invention further comprises the light-shielding region allows the present invention to more clearly avoid this problem.

[0097] The light-blocking layer is not particularly restricted, provided that the layer is formed on top of the semi-transparent layer and its optical density (OD value) is adjusted at the wavelength of the ArF excimer laser exposure light to achieve a desired overall optical density (OD value) when combined with that of the semi-transparent layer. The light-blocking layer is preferably a layer with an antireflection function to prevent multiple reflections between the phase-shift mask and a lens when the mask pattern is transferred to a wafer, and with an etch barrier function for the semi-transparent layer to advantageously prevent damage to the semi-transparent layer when it is etched.The light-shielding layer is in particular preferably a layer with an electrically conductive function to prevent this layer from becoming electrically charged when an image is drawn on it by an electron beam.

[0098] As it is in Fig. In the mask blank 100 illustrated in Figure 5, the light-blocking layer preferably has a single-layer structure comprising a light-absorbing layer formed on the semi-transparent layer, which serves as an etching barrier for the semi-transparent layer and as a light-absorbing layer for absorbing the ArF excimer laser exposure light. This structure allows a mask with the necessary functions to be obtained in a smaller number of steps.

[0099] If the light-blocking layer has a single-layer structure, as described in Fig.As illustrated in Figure 5, the raw material for the light-blocking layer is not particularly restricted and is preferably, for example, Cr, Ta, W, or Mo. Of these materials, Cr is preferred. In this case, a reactive etching gas used when etching the light-blocking layer is different in nature from a reactive etching gas used when etching the semi-transparent layer; thus, the light-blocking layer can be expected to have the etch barrier function specified above.

[0100] If the light-blocking layer has a single-layer structure, its thickness is varied according to the raw material to avoid excessive limitations. The thickness is preferably in the range of 30 to 80 nm. This allows the optical density (OD value) of this light-blocking layer to be combined with that of the semi-transparent layer at the wavelength of the ArF excimer laser exposure light, making it easy to achieve an overall value of 3.0 or higher. Additionally, this layer thickness facilitates etching of the light-blocking layer.

[0101] The optical density (OD value) of this light-blocking layer pattern, combined with that of the semi-transparent layer, is measured and calculated using a product, MCPD 3000™, manufactured by Otsuka Electronics Co., Ltd.; and the reflectance of the light-blocking layer is measured using a product, MCPD 7000™, manufactured by Otsuka Electronics Co., Ltd.

[0102] Fig. Figure 6 is a schematic sectional view illustrating yet another example of the mask blank. In a mask blank 100, which is in Fig.As illustrated in Figure 6, a light-blocking layer 103 has a bilayer structure comprising: a light-absorbing layer 103a formed on a semi-transparent layer 102, and a hard mask layer 103b formed on the light-absorbing layer 103a. The light-absorbing layer 103a has two functions: it absorbs the ArF excimer laser exposure light and acts as an etch barrier for the semi-transparent layer 102. The hard mask layer 103b also acts as an etch barrier for the light-absorbing layer 103a.

[0103] As it is in Fig.As illustrated in Figure 6, the light-blocking layer in the mask blank has a double-layer structure, comprising: a light-absorbing layer formed on top of the semi-transparent layer, which serves as an etch barrier for the semi-transparent layer and as a light-absorbing layer to absorb the ArF excimer laser exposure light; and a hard mask layer formed on top of the light-absorbing layer, which also serves as an etch barrier for the light-absorbing layer. This structure allows for the use of a pattern formed by etching the hard mask layer instead of a resist pattern, which would otherwise be required when etching the light-absorbing layer. This enables a small resist layer thickness to be used, thus facilitating the creation of a fine pattern in the light-blocking layer.

[0104] If the light-blocking layer has a single-layer structure, as is the case in Fig. As illustrated in Figure 5, the light-shielding layer of the single-layer structure is thick, so a thick resist pattern is used to etch the light-shielding layer to form a pattern. Therefore, it is difficult to create a fine pattern of the light-shielding layer based on the relationship to the aspect ratio of the resist pattern. However, if the light-shielding layer has a double-layer structure, as in Figure 5, the light-shielding layer can be etched in a more precise way. Fig.As illustrated in Figure 6, the hard mask layer is thinner than the light-blocking layer of the single-layer structure. By using a thinner resist pattern than the one used to etch the light-blocking layer of the single-layer structure, it is possible to etch the hard mask layer to form a pattern. This pattern can be used instead of the resist pattern required when etching the light-absorbing layer, allowing the resist layer thickness to be reduced. In this way, a fine pattern can be more easily formed in the present light-blocking layer than in the light-blocking layer of the single-layer structure.

[0105] If the light-shielding layer has a bilayer structure, the hard mask layer is not particularly restricted. The hard mask layer can have the electrically conductive function described above. If the light-shielding layer has a bilayer structure, the raw material for the hard mask layer is not particularly restricted, provided the raw material has an etch barrier function for the light-absorbing layer. Examples include Si, SiN, SiON, SiO2, MoSi, Cr, CrO, and CrON. Of these materials, Si, SiN, SiON, SiO2, MoSi, and others are preferred.When a Cr-containing raw material is used as the raw material for the light-absorbing layer, these materials make a reactive etching gas used for etching the light-absorbing layer different in nature from a reactive etching gas used for etching the hard mask layer; when the hard mask layer is etched, this can prevent damage to the light-absorbing layer.

[0106] If the light-blocking layer has a bilayer structure, the raw material for the light-absorbing layer is not particularly restricted, provided the raw material has an etch barrier function for the semi-transparent layer and a light-absorbing function for absorbing the ArF excimer laser exposure light. Examples include Cr, Si, SiO, SiON, and MoSi. Of these materials, Cr is preferred. In this case, a reactive etching gas used for etching the light-absorbing layer will differ in nature from a reactive etching gas used for etching the hard mask layer; this prevents damage to the light-absorbing layer when the hard mask layer is etched.Furthermore, Cr has a high extinction coefficient; since the light-blocking layer has a smaller thickness, the optical density (OD value) of this layer at the wavelength of the ArF excimer laser exposure light can be combined with that of the semi-transparent layer to be set to 3.0 or more overall.

[0107] If the light-blocking layer has a double-layer structure, the thickness of the hard mask layer varies depending on the type of raw material and is not particularly limited. The thickness is preferably in the range of 4 to 15 nm, more preferably 4 to 10 nm, and particularly preferably 4 to 7 nm. Since the layer thickness is smaller, the layer can be machined with greater precision.

[0108] If the light-blocking layer has a double-layer structure, the thickness of the light-absorbing layer varies depending on the type of raw material and is not particularly limited. The thickness is preferably in the range of 30 to 80 nm, more preferably 30 to 70 nm, and particularly preferably 30 to 60 nm. Since the layer thickness is smaller, the layer is easier to machine or a defective pattern in the layer is easier to correct.

[0109] Fig. 7 is a schematic sectional view illustrating another example of the mask blank. In one in Fig.The illustrated mask blank 100 has a light-blocking layer 103 and a three-layer structure comprising: an etch barrier layer 103c formed on a semi-transparent layer 102, a light-absorbing layer 103a formed on the etch barrier layer 103c, and a hard mask layer 103b formed on the light-absorbing layer 103a. The etch barrier layer 103c has an etch barrier function for the semi-transparent layer 102; the light-absorbing layer 103a has a light-absorbing function for absorbing the ArF excimer laser exposure light; and the hard mask layer 103b has an etch barrier function for the light-absorbing layer 103a.

[0110] Fig. Figure 8 is a schematic sectional view illustrating another example of the mask blank. In one in Fig.The illustrated mask blank 100 has a light-blocking layer 103 and a three-layer structure comprising: an etch barrier layer 103c formed on a semi-transparent layer 102, a light-absorbing layer 103a formed on the etch barrier layer 103c, and a hard mask layer 103b formed on the light-absorbing layer 103a. The etch barrier layer 103c has an etch barrier function for the semi-transparent layer 102; the light-absorbing layer 103a has a light-absorbing function for absorbing the ArF excimer laser exposure light; and the hard mask layer 103b has an etch barrier function for the light-absorbing layer 103a.

[0111] How it was achieved in the mask blank 100, which is in each of Fig. 7 and Fig.As illustrated in Figure 8, the light-blocking layer in the mask blank preferably has a three-layer structure comprising: an etch barrier layer formed on its semi-transparent layer, which acts as an etch barrier for the semi-transparent layer; a light-absorbing layer formed on the etch barrier layer, which acts as a light-absorbing layer for the absorption of the ArF excimer laser exposure light; and a hard mask layer formed on the light-absorbing layer, which also acts as an etch barrier for the light-absorbing layer. Similar to the case where the light-blocking layer has a double-layer structure, this case allows the use of a pattern formed by etching the hard mask layer instead of a resist pattern, which is required when etching the light-absorbing layer.As a result, a fine pattern of the light-blocking layer is easily formed.

[0112] In the mask blank, a three-layer structure is preferred for the light-blocking layer over a single-layer or double-layer structure. This allows for the use of a light-absorbing layer for absorbing the ArF excimer laser exposure light. This layer can be made from a raw material that can be etched with a highly reactive etching gas to a semi-transparent layer consisting of either Si and N or Si, N, and O. Consequently, the flexibility in selecting the raw material for this light-absorbing layer is increased, allowing for a thinner light-blocking layer.Furthermore, this layer, which has a light-absorbing function for absorbing the ArF excimer laser exposure light, and a layer with an etch barrier function for the semi-transparent layer can be made into the light-absorbing layer and the etch barrier layer, respectively, consisting of different raw materials. Accordingly, the light-absorbing layer and the etch barrier layer can be etched with different etching gases, so that the etch barrier layer can be made into a layer with an etch barrier function for the semi-transparent layer, advantageously preventing damage to the semi-transparent layer when the light-absorbing layer is etched.

[0113] For example, in Fig.Figure 7 illustrates that the mask blank 100, as a layer with a light-absorbing function for absorbing the ArF excimer laser exposure light, can be used as the light-absorbing layer 103a, which is a layer consisting of MoSi, a material that can be etched with a highly reactive etching gas to the semi-transparent layer, which consists only of Si and N, or only of Si, N, and O. Furthermore, this layer, which has a light-absorbing function for absorbing the ArF excimer laser exposure light, and a layer with an etch barrier function for the semi-transparent layer can be made into the light-absorbing layer 103a, which is a layer consisting of MoSi, or into the etch barrier layer 103c, which is a layer consisting of Cr.This case makes it possible to etch the light-absorbing layer 103a and the etch barrier layer 103c with a fluorine-containing gas and a chlorine-containing gas, respectively, which are different from each other. Accordingly, the etch barrier layer 103c can be used as a layer with an etch barrier function for the semi-transparent layer 102, advantageously preventing damage to the semi-transparent layer when the light-absorbing layer 103a is etched.

[0114] In the Fig.The mask blank 100 illustrated in Figure 8 can be used as a layer with a light-absorbing function for absorbing the ArF excimer laser exposure light. This layer consists of Si, a material that can be etched with a highly reactive etching gas to the semi-transparent layer, which consists only of Si and N, or only of Si, N, and O. Furthermore, this layer, which has a light-absorbing function for absorbing the ArF excimer laser exposure light, and a layer with an etch barrier function for the semi-transparent layer can be combined to form the light-absorbing layer 103a, which consists of Si, or the etch barrier layer 103c, which consists of Cr.This case makes it possible to etch the light-absorbing layer 103a and the etch barrier layer 103c with a fluorine-containing gas and a chlorine-containing gas, respectively, which are different from each other. Accordingly, the etch barrier layer 103c can be used as a layer with an etch barrier function for the semi-transparent layer 102, advantageously preventing damage to the semi-transparent layer when the light-absorbing layer 103a is etched.

[0115] If the light-shielding layer has a three-layer structure, the hard mask layer is not particularly restricted and can have the electrically conductive function described above. If the light-shielding layer has a three-layer structure, the raw material for the hard mask layer is not particularly restricted, provided the raw material has an etch barrier function for the light-absorbing layer. Examples include Cr, CrO, CrON, SiN, SiON, and SiO2. Of these materials, Cr is preferred.

[0116] If the light-shielding layer has a three-layer structure, the raw material for the light-absorbing layer is not particularly restricted, provided the raw material has the light-absorbing function specified above. Examples include metals such as MoSi, materials containing silicon (Si), and simple substances consisting of a single metal, such as tungsten (W) or tin (Ta), or of silicon (Si). Of these examples, silicon (Si) is particularly preferred. This makes it possible to design a reactive etching gas used for etching the light-absorbing layer differently from a reactive etching gas used for etching the hard mask layer, thus preventing damage to the light-absorbing layer when the hard mask layer is etched.Additionally, simple silicon (Si) has a high extinction coefficient; thus, the thickness of the light-shielding layer can be reduced if the light-absorbing layer consists of simple silicon (Si).

[0117] If the light-blocking layer has a three-layer structure, the raw material for the etch barrier layer is not particularly restricted, provided the raw material possesses the etch barrier function specified above. Examples include Cr and CrON. This allows for the use of a different reactive etching gas for etching the etch barrier layer compared to a reactive etching gas used for etching the light-absorbing layer, thus preventing damage to the etch barrier layer when the light-absorbing layer is etched. For example, if the light-absorbing layer is etched with a fluorine-containing gas, damage to the etch barrier layer can be prevented, and the light-absorbing layer can be etched with a chlorine-containing gas.Furthermore, this case makes it possible to design a reactive etching gas used for etching the etch barrier layer differently from a reactive etching gas used for etching the semi-transparent layer, thus preventing damage to the semi-transparent layer when the etch barrier layer is etched.

[0118] If the light-blocking layer has a three-layer structure, the thickness of the hard mask layer is varied according to the type of raw material and is not particularly limited. The thickness is preferably in the range of 4 to 10 nm, more preferably 4 to 7 nm, and particularly preferably 4 to 6 nm. This configuration makes it possible to ensure that the etch barrier function described above is sufficiently demonstrated and to adequately prevent damage to the light-absorbing layer when the hard mask layer is etched.

[0119] If the light-shielding layer has a three-layer structure, the thickness of the light-absorbing layer varies depending on the type of raw material and is not particularly limited. The thickness is preferably in the range of 20 to 70 nm, more preferably 20 to 60 nm, and particularly preferably 30 to 50 nm. This configuration ensures that the light-absorbing function described above is sufficiently effective and adequately prevents damage to the etch barrier layer when the light-absorbing layer is etched. Furthermore, the thickness of the light-shielding layer can be smaller if the light-absorbing layer consists of simple silicon (Si) than if it consists of MoSi, due to the higher extinction coefficient of silicon (Si).

[0120] If the light-blocking layer has a three-layer structure, the thickness of the etch barrier layer varies depending on the type of raw material and is not particularly limited. The thickness is preferably in the range of 2 to 6 nm, and more preferably 2 to 4 nm. In this case, the etch barrier function can be adequately demonstrated.

[0121] The light-blocking layer is not particularly restricted. The layer is preferably a light-blocking layer for which the optical density (OD value) is set at the wavelength of the ArF excimer laser exposure light to yield a total value of 3.0 or more in combination with that of the semi-transparent layer. This allows the resulting mask to acquire the light-blocking property necessary for a desired region when the mask is exposed to light. (4) Other parts

[0122] The mask blank is not particularly restricted, provided it includes the semi-transparent layer and the transparent substrate defined above. Apart from these components, any optional parts can be suitably added.

[0123] The mask blank is preferably a mask blank that includes a resist layer on the semi-transparent layer or on the light-blocking layer. Without the need to create a new resist layer, this allows a predetermined resist pattern to be developed after the resist layer is exposed to light. In this way, a phase-shift mask can be more easily created from the mask blank. 2. Structure of the mask blank

[0124] Next, the structure of the mask blank is described. The mask blank is a mask blank in which the semi-transparent layer defined above is formed on the transparent substrate defined above. The structure of the mask blank and a manufacturing process for it are described below. (1) Structure of the mask blank

[0125] The mask blank is not particularly restricted and is preferably a mask blank in which the semi-transparent layer is formed directly on the transparent substrate.

[0126] In a semi-transparent layer with high transmittance, such as the semi-transparent layer in the present invention, its refractive index tends to be low. In the case of attempting to generate a 180° retardation, it is therefore necessary in a halftone phase-shift mask, between a region where this semi-transparent layer is formed and a region through which exposure light is sent, to achieve a suitable phase shift, to make the thickness of the semi-transparent layer large (Patent References 1 and 2). If the thickness of the semi-transparent layer is made large, an etch barrier layer can be formed on the transparent substrate to prevent damage to the transparent substrate during etching of the semi-transparent layer (Patent References 1 and 2).An etch barrier layer is formed that is not etched with the same type of gas (such as CF4 or SF6) as the semi-transparent layer, in order to increase the etch selectivity between the barrier layer and the semi-transparent layer. However, in the case of forming such an etch barrier layer, for example, a Ta-Hf layer is formed in patent literature 1 and then etched with a Cl-containing gas, as shown in the examples therein. This involves performing the etching step multiple times, thus complicating the etching process. Furthermore, the layer is not easily etched with the Cl-containing gas, resulting in a poorly shaped and uniform finish. Additionally, patent literature 2 also describes a structure in which the etch barrier layer consists of Zr and Hf.However, the etching step must be performed several times, and the same problem based on etching difficulty is caused in the same way.

[0127] According to the present invention, however, the semi-transparent layer is thinner than conventional semi-transparent layers, making it easier to etch into a transparent layer pattern. This reduces the time required for etching. Even if the mask blank does not have an etch barrier layer between the transparent substrate and the semi-transparent layer, damage to the transparent substrate during etching is sufficiently prevented. Accordingly, the mask blank becomes one in which the semi-transparent layer is formed directly on the transparent substrate. Therefore, it is unnecessary to form an etch barrier layer as described above; thus, it is unnecessary to repeat the etching step multiple times.Accordingly, the etching process is not complicated, and furthermore, the following can be prevented: a deterioration of the shape of the semi-transparent layer or the transparent substrate, and a deterioration of the uniformity of the shape of the semi-transparent layer, these inconveniences being based on the difficulty of etching the etch barrier layer. (2) Method for producing a mask blank

[0128] The process for manufacturing the mask blank is not particularly restricted, provided that the process is one that, as a mask blank, can yield a desired mask blank. In one example of the process for manufacturing the mask blank, the transparent substrate defined above is initially produced. Next, the semi-transparent layer defined above is formed on the transparent substrate by a known layering process, such as sputtering. Next, the light-blocking layer defined above is formed on the semi-transparent layer by a known layering process, such as sputtering.If the light-blocking layer has a bilayer structure, the light-absorbing layer defined above is formed on the semi-transparent layer by a known layering process, such as sputtering, and then the hard mask layer defined above is formed on the light-absorbing layer by a known layering process, such as sputtering. If the light-blocking layer has a three-layer structure, the etch barrier layer defined above is formed on the semi-transparent layer by a known layering process, such as sputtering, and then the light-absorbing layer defined above is formed on the etch barrier layer by a known layering process, such as sputtering. Finally, the hard mask layer defined above is formed on the light-absorbing layer by a known layering process, such as sputtering. In this way, the mask blank is obtained.

[0129] In this method for producing a mask blank, the process for forming the semi-transparent layer is not particularly restricted. For example, the method involves using a silicon (Si) target as a sputtering target, selecting a suitable sputtering gas, and then using sputtering to form a semi-transparent layer consisting only of Si and N, or a semi-transparent layer consisting only of Si, N, and O, under conditions to adjust the respective component composition ratios to desired proportions.Among such methods, a preferred method involves using a silicon (Si) target as the sputtering target and a sputtering gas containing nitrogen but no oxygen, or containing oxygen only in a small proportion, to form a high-refractive-index SiN layer or a high-refractive-index SiON layer. This method makes it possible to form a semi-transparent layer with excellent properties among semi-transparent layers, as defined above. Furthermore, the term "high-refractive-index SiN layer," that is, the semi-transparent layer consisting only of Si and N, refers to a semi-transparent layer formed using a silicon (Si) target as the sputtering target and a sputtering gas containing nitrogen but no oxygen, in an oxygen-free atmosphere. B. Mask blank attached to a negative-type resist layer

[0130] The following describes a mask blank attached to a negative-type resist layer. The mask blank attached to a negative-type resist layer comprises the mask blank defined above and a negative-type resist layer formed on the mask blank.

[0131] Fig. Figure 9 is a schematic sectional view illustrating an example of the mask blank attached to a negative-type resist layer. A mask blank 110 attached to a negative-type resist layer, which is in Fig.As illustrated in Figure 9, a mask blank attached to a negative-type resist layer is used to produce a halftone phase-shift mask onto which ArF excimer laser exposure light is directed. The mask blank 110 attached to the negative-type resist layer comprises a mask blank 100 with a transparent substrate 101, a semi-transparent layer 102 formed on the transparent substrate 101 and having a single-layer structure consisting only of Si and N or only of Si, N, and O, and a light-blocking layer 103 formed on the semi-transparent layer 102 and having a single-layer structure for which the optical density (OD value) of the semi-transparent layer 103 is set to be 3.0 or more at a wavelength of the ArF excimer laser exposure light, in combination with that of the semi-transparent layer 102.The mask blank 110 attached to the negative-type resist layer also has a negative-type resist layer 111 formed on the mask blank 100.

[0132] As described below under point “D. Method for producing a pattern-formed body using a phase-shift mask”, if a halftone phase-shift mask with a semi-transparent layer is used to transfer a fine pattern for contact holes, conductors or other onto a wafer, the fine pattern for the contact holes, conductors or other onto the wafer can be transferred by negative development while easily avoiding a sidelobe phenomenon.

[0133] When a fine pattern for contact holes, conductors or other features is transferred to the wafer by negative development, it is necessary in a halftone phase-shift mask to form a fine convex element pattern consisting of a semi-transparent layer and corresponding to the fine pattern for the contact holes, conductors or other features.

[0134] According to the present invention, a fine convex element pattern is formed from the semi-transparent layer using a pattern of the above-mentioned negative-type resist layer for etching the semi-transparent layer, and thus a negative-type phase-shift mask can be produced, which is described in detail below. When the negative-type phase-shift mask is produced, the area exposed to light from the negative-type resist layer in the mask blank attached to the negative-type resist layer is a very narrow area corresponding to the fine convex element pattern from the semi-transparent layer. Thus, the mask attached to the negative-type resist layer can be produced in a shorter time.

[0135] A negative-type resist composition used to form the negative-type resist layer is not particularly restricted. Examples include product NEB-22A™, manufactured by Sumitomo Chemical Co., Ltd.; and products SEBN-1637™, SEBN-1702™, and SEBN-2014™, manufactured by Shin-Etsu Chemical Co., Ltd. Of these products, SEBN-2014™, manufactured by Shin-Etsu Chemical Co., Ltd., is preferred because it is suitable for forming a finer pattern.

[0136] The thickness of the negative-type resist layer is not particularly restricted and is preferably in the range of 50 to 150 nm, particularly 80 to 100 nm. This makes it possible that, when a pattern is formed on the mask, this pattern is fine, while the mask blank still provides sufficient etch barrier function.

[0137] The method for forming the negative-type resist layer is not particularly restricted and includes, for example, application by centrifugal coating. C. Phase shift mask

[0138] The following describes the phase shift mask of the present invention. The phase shift mask of the present invention is a halftone phase shift mask onto which ArF excimer laser exposure light is directed.The mask comprises a transparent substrate and a semi-transparent layer pattern formed on the transparent substrate and consisting only of Si (silicon) and N (nitrogen), or a semi-transparent layer pattern formed on the transparent substrate and consisting only of Si (silicon), N (nitrogen) and O (oxygen), and is characterized in that the semi-transparent layer pattern has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light and a transmittance of 15 to 38% at the wavelength of the ArF excimer laser exposure light and furthermore has a layer thickness of 57 to 67 nm.The semi-transparent layer pattern features a primary pattern that is to be resolved onto a wafer and an auxiliary pattern that is used to assist in the resolution of the primary pattern and is not to be resolved onto the wafer. The auxiliary pattern comprises one or more convex pattern elements, each with a width or length of 60 nm or less. The primary pattern has the same thickness as the auxiliary pattern.

[0139] Fig. Figure 10 is a schematic top view illustrating an example of the phase shift mask of the present invention. Fig. 11 is an AA sectional view of Fig. 10. One in Fig. The illustrated phase shift mask 200 is a halftone phase shift mask onto which ArF excimer laser exposure light is to be directed. The one in Fig.The illustrated phase-shift mask 200 comprises a transparent substrate 201 and a semi-transparent layer pattern 202 formed on the transparent substrate 201. The semi-transparent layer pattern 202 has a single-layer structure consisting of either Si and N or Si, N, and O. It has an extinction coefficient of 0.2 to 0.45 at the wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light, and a transmittance of 15 to 38% at the wavelength of the ArF excimer laser exposure light. The semi-transparent layer pattern 202 also has a thickness of 57 to 67 nm.Furthermore, the semi-transparent layer pattern 202 has a main pattern 202a, which is resolvable onto a wafer and is a pattern element with a width or length of 100 to 300 nm, and an auxiliary pattern 202b, which is not resolvable onto the wafer but supports the resolution of the main pattern 202a and consists of pattern elements, each with a width or length of 60 nm or less. The main pattern 202a pattern element and the auxiliary pattern 202b pattern elements are each convex pattern elements.

[0140] Fig. Figure 12 is a schematic top view illustrating another example of the phase shift mask of the present invention. Fig. 13 is an AA sectional view of Fig. 12. The following section describes the differences between the Fig. 12 illustrated phase shift mask 200 compared to the one in Fig.Figure 10 illustrates phase-shift mask 200. In a semi-transparent layer pattern 202, as the main pattern 202a, which is to be resolved onto a wafer and is a pattern element with a width or length of 100 to 300 nm, a concave pattern region is formed, which is obtained by partially hollowing out the semi-transparent layer; thus, the transparent substrate 201 is freed from this pattern. In the semi-transparent layer pattern 202, as the auxiliary pattern 202b, which is not to be resolved onto the wafer but supports the resolution of the main pattern 202a, and which consists of pattern elements each with a width or length of 60 nm or less, concave pattern regions are formed, which are obtained by partially hollowing out the semi-transparent layer; thus, the transparent substrate 201 is freed from the auxiliary pattern.

[0141] In the phase-shift mask of the present invention, the semi-transparent layer pattern has a high transmittance in the range of 15 to 38% at the wavelength of the ArF excimer laser exposure light. Accordingly, the phase-shift mask of the present invention is used to adjust the intensity of the light to zero at its pattern element boundaries by light interference based on the phase effect, thereby improving the contrast of the resulting transferred image. When a patterned body is produced with this improvement, the phase effect can be made more pronounced due to the high transmittance of the semi-transparent layer pattern. Furthermore, the semi-transparent layer pattern contains no metal; thus, no silicon (Si) oxide layer grows, even when the ArF excimer laser exposure light is irradiated onto this layer pattern for a long period of time.Accordingly, it is possible to prevent changes to the critical dimension of the pattern. Similarly, during the cleaning step of the phase-shift mask, changes to the critical dimension of the pattern can also be prevented. Therefore, the present invention enables photolithography to achieve excellent transfer properties for the phase-shift mask pattern and to impart high resistance to irradiation with ArF excimer laser exposure light and to cleaning resistance.

[0142] With regard to the phase-shift mask of the present invention, the layer thickness of the semi-transparent layer pattern is in the range of 57 to 67 nm, in order to be smaller than that of conventional semi-transparent layer patterns. Thus, for the same reasons as described with regard to the mask blank, flaking of the convex element pattern can be avoided in the phase-shift mask of the present invention, as described above, wherein flaking is caused by cleaning the convex element pattern using ultrasonic waves with a strong removal force in a cleaning solution.

[0143] With regard to the phase-shift mask of the present invention, the layer thickness of the semi-transparent layer pattern is in the range of 57 to 67 nm, in order to be smaller than that of conventional semi-transparent layer patterns. Thus, for the same reasons as described with regard to the mask blank, the design flexibility of the semi-transparent layer pattern can be increased.

[0144] With regard to the phase-shift mask of the present invention, the transmittance of the semi-transparent layer pattern at the wavelength of the ArF excimer laser exposure light is in the range of 15 to 38%, which is higher than in the prior art. For this reason, particularly when a negative-type resist process or negative development is used in a wafer process, the phase-shift mask of the present invention can reduce the OPC bias value compared to conventional phase-shift masks. Thus, when using a negative-type resist process or negative development to form a fine pattern on a wafer using the phase-shift mask of the present invention, the advantages stated above can be clearly achieved.In other words, by keeping the OPC bias value small, a significant advantage of the restriction that causes the following result can be clearly preserved: When an optical near-field effect correction (OPC processing) calculation is performed by approximation in a semi-transparent layer pattern design, the pattern, which is a fine pattern, is brought into contact with or separated from any other part of the wafer contrary to the design intent, as described above. Accordingly, the aforementioned advantage of increasing the design flexibility of the semi-transparent layer pattern can be clearly preserved.

[0145] In the phase-shifting mask of the present invention, the thickness of the semi-transparent layer pattern is in the range of 57 to 67 nm, in order to be smaller than that of conventional semi-transparent layer patterns. Thus, for the same reasons as described with regard to the mask blank, damage to the transparent substrate can be sufficiently prevented.

[0146] In the following, parts of the phase shift mask and the structure of the phase shift mask of the present invention are described separately. 1. Parts of the phase shift mask

[0147] First, the components of the phase-shift mask of the present invention are described. The phase-shift mask of the present invention comprises at least one transparent substrate and a semi-transparent layer pattern. (1) Semi-transparent layer pattern

[0148] The semi-transparent layer pattern in the present invention is a semi-transparent layer pattern formed on a transparent substrate, which is described in detail below, and consists only of Si and N, or a semi-transparent layer pattern formed on a transparent substrate, which is described in detail below, and consists only of Si, N, and O. The semi-transparent layer pattern has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light, and a transmittance of 15 to 38% at the wavelength of the ArF excimer laser exposure light. The semi-transparent layer pattern furthermore has a layer thickness of 57 to 67 nm.

[0149] With regard to modern halftone phase-shift masks onto which ArF excimer laser exposure light is directed, positional accuracy within the phase-shift masks is very important. In the phase-shift mask of the present invention, the layer thickness of the semi-transparent layer pattern is in the range of 57 to 67 nm, which is smaller than that of conventional semi-transparent layer patterns. This makes it possible to reduce the stresses exerted on the phase-shift mask of the present invention by the semi-transparent layer pattern. In this way, the stress on the phase-shift mask can be limited, thus improving positional accuracy within the phase-shift mask.

[0150] In modern wafer processing techniques for transferring a mask pattern of a halftone phase-shift mask with a semi-transparent layer onto a wafer, it is necessary, when transferring a finer pattern than conventional patterns onto a wafer, to take measures such as the following to limit the variation in the shape of a wafer resist from a regular shape, in order to reduce variation in the wafer dimensions from a regular dimension: Measures to achieve this used a wafer resist with low sensitivity to make the irradiation period of the ArF excimer laser exposure light longer than in the prior art.Currently, a fine pattern, corresponding to a finer pattern than conventional patterns, is transferred to the wafer and consists of the semi-transparent layer in the phase-shift mask, making it finer than in the prior art; therefore, it is necessary to increase the resistance of this finer pattern to irradiation with ArF excimer laser exposure light.

[0151] In contrast to the techniques described above, the layer thickness of the semi-transparent layer pattern in the phase-shift mask of the present invention is in the range of 57 to 67 nm, making it smaller than that of conventional semi-transparent layer patterns. Thus, in the phase-shift mask of the present invention, the fine pattern of the semi-transparent layer occupies a small area. This reduces the adsorption of residual ions from a cleaning solution, as well as ions and organic substances from the surrounding environment, onto the fine pattern. This, in turn, reduces the risk of the formation of impurities, referred to as a haze, by illuminating these adsorbed impurities or substances with the ArF excimer laser light.

[0152] Accordingly, the phase-shift mask of the present invention reduces the risk of veiling in the fine pattern of the semi-transparent layer in the phase-shift mask compared to conventional phase-shift masks. This makes the following possible, even if the fine pattern of the semi-transparent layer in the phase-shift mask of the present invention is finer than in conventional phase-shift masks: the phase-shift mask of the present invention becomes a mask with such resistance to irradiation with ArF excimer laser exposure light that the mask can withstand irradiation with ArF excimer laser exposure light for a longer period than in the prior art.

[0153] In the phase-shift mask of the present invention, the layer thickness of the semi-transparent layer pattern is in the range of 57 to 67 nm, in order to be smaller than that of conventional semi-transparent layer patterns. Thus, in the phase-shift mask of the present invention, the semi-transparent layer pattern covers a small area, in particular the sidewall area, so that the mask's resistance to irradiation with the ArF excimer laser exposure light and its cleaning resistance (resistance to denaturation based on a chemical cleaning solution) can be improved.

[0154] In the cleaning step of a phase-shift mask used to form a 10 nm node pattern in a wafer, conventional cleaning conditions with high physical removal force cannot be selected to avoid flaking off the pattern of its convex element pattern consisting of one or more pattern elements, each of which is composed of a semi-transparent layer with a width or length of 60 nm or less, such as an auxiliary pattern as described above. This leads to the problem that it becomes difficult to remove a foreign substance trapped in a narrow gap or a large foreign substance covering a line grid pattern of the mask. In the phase-shift mask of the present invention, however, the thickness of the semi-transparent layer pattern is in the range of 57 to 67 nm, which is smaller than that of conventional semi-transparent layer patterns.This minimizes the difference in level between the concave and convex shapes of the grid pattern, allowing foreign material trapped in its narrow gaps to be easily removed. Furthermore, a large foreign material covering the grid pattern is given a small area where it fits into the pattern. This allows any foreign material covering the grid pattern to be easily removed.

[0155] In the phase-shift mask of the present invention, the layer thickness of the semi-transparent pattern is in the range of 57 to 67 nm, which is smaller than that of conventional semi-transparent layer patterns. This results in a small surface area for the semi-transparent layer pattern in the phase-shift mask of the present invention. This makes it possible to shorten the time required for cleaning and drying the phase-shift mask. Furthermore, the fact that the layer thickness of the semi-transparent layer pattern is smaller than that of conventional semi-transparent layer patterns makes it possible to keep raw material costs for manufacturing the phase-shift mask low. a. Convex element pattern

[0156] The semi-transparent layer pattern is not particularly restricted and is preferably a convex element pattern consisting of one or more pattern elements, each having a width or length of 60 nm or less. The height of the convex pattern element(s) with a width or length of 60 nm or less is in the range of 57 to 67 nm, to be lower than those in conventional halftone phase-shift masks. This reduces the area of ​​this pattern that experiences bubble-breaking impacts when the pattern is cleaned using ultrasonic waves in the phase-shift mask cleaning step, thus lowering the position of this pattern that experiences bubble-breaking impacts.Accordingly, flaking of this pattern can be avoided, whereby this flaking is caused by cleaning the pattern using ultrasonic waves with a strong removal force in a cleaning solution. b. Main pattern and auxiliary pattern(a) Pattern to be transferred to a wafer by negative development

[0157] The semi-transparent layer pattern is not particularly restricted and can be a semi-transparent layer pattern, such as the one in Fig. 10 illustrated examples of transferring a fine pattern for contact holes, conductors, and other features onto a wafer by negative development. As shown in Fig.As illustrated in Figure 10, this semi-transparent layer pattern has a primary pattern that is to be resolved onto the wafer and an auxiliary pattern that is not to be resolved onto the wafer but assists in the resolution of the primary pattern. The auxiliary pattern is a convex element pattern consisting of one or more pattern elements, each with a width or length of 60 nm or less. The primary pattern corresponds to the fine pattern that is to be transferred onto the wafer. If the semi-transparent layer pattern is such a semi-transparent layer pattern, flaking of the pattern can be avoided, whereas this flaking is caused by cleaning the auxiliary pattern using ultrasonic waves with a strong removal force in a cleaning solution.

[0158] Furthermore, as described below, even with a pattern to be transferred to a wafer by positive development, pattern flaking can be caused by cleaning the pattern using ultrasonic waves. However, when comparing a pattern to be transferred to a wafer by negative development with one to be transferred to a wafer by positive development, the length of its convex pattern elements, each with a width of 100 nm or less, is generally shorter in the pattern to be transferred to the wafer by negative development. This allows the advantage of avoiding pattern flaking caused by cleaning the pattern using ultrasonic waves to be clearly demonstrated.

[0159] The main pattern shown above is a convex element pattern from the semi-transparent layer and can actually be resolved onto a wafer.

[0160] The auxiliary pattern is a convex element pattern from the semi-transparent layer and is not actually resolved onto the wafer. The auxiliary pattern is positioned to assist diffraction light generated by the main pattern, thus improving the mask's exposure latitude. When the mask is defocused, variations in the critical dimension (CD) of the pattern can be reduced. Therefore, the auxiliary pattern is a pattern that can assist in resolving the main pattern onto the wafer. The width or length of one or more elements of the auxiliary pattern is in the range of 10 to 60 nm. The auxiliary pattern is not particularly restricted and is preferably a pattern in which one or more elements have a width or length of 20 to 60 nm, and more preferably 30 to 60 nm.In this case, the auxiliary pattern can be produced with a good yield ratio, with the element(s) functioning as the auxiliary pattern. However, a suitable size of the element(s) of the auxiliary pattern is comprehensively determined according to the lighting conditions at the time of exposure, the size of the main pattern, and other factors. (b) Pattern to be transferred to a wafer by positive development

[0161] The semi-transparent layer pattern described above is not particularly limited and can be used to transfer a fine pattern for contact holes, conductors, and other features onto a wafer by positive development, as in the [reference to be added]. Fig. As can be seen in the illustrated example 12. As in the one in Fig.As can be seen in the illustrated example 12, this semi-transparent layer pattern is a pattern in which a primary pattern, which is to be resolved onto a wafer, and an auxiliary pattern, which assists in the resolution of the primary pattern and is not to be resolved onto the wafer, are formed; and concave region patterns, as primary and auxiliary patterns, are formed by partially hollowing out the semi-transparent layer. The primary pattern corresponds to the fine pattern that is to be transferred onto the wafer.

[0162] The main pattern is a concave region pattern in which the semi-transparent layer is partially hollowed out, and is actually resolvable onto a wafer.

[0163] The auxiliary pattern is a concave region pattern in which the semi-transparent layer is partially hollowed out and is not actually resolved onto a wafer. Through diffraction light generated by the auxiliary pattern, this pattern assists in resolving the primary pattern onto the wafer, thus improving the mask's exposure latitude. When the mask is defocused, this can reduce variation in the critical dimension (CD) of the pattern. Accordingly, the auxiliary pattern is a pattern that can support the resolution of the primary pattern onto the wafer. It should also be noted here that even with positive development, pattern flaking can be caused by ultrasonic cleaning if the spacing between the concave regions of the concave region pattern is in the range of 10 to 100 nm, particularly 10 to 60 nm. c. Other structural factors

[0164] The structure of the semi-transparent layer pattern in the present invention is equivalent to that of the semi-transparent layer in the present invention described under point “A. Mask blank, 1. Parts of the mask blank, (1) Semi-transparent layer”, except for the points mentioned above. Therefore, any description thereof is omitted here. (2) Transparent substrate

[0165] The structure of the transparent substrate in the present invention is equivalent to that of the transparent substrate in the present invention described under point “A. Mask blank, 1. Parts of the mask blank, (2) Transparent substrate”. Therefore, any further description of it is omitted here. (3) Light-shielding layer pattern

[0166] The phase-shift mask of the present invention is not particularly restricted with respect to the layer structure and the raw material thereof, nor with respect to its optical density (OD value) at the wavelength of the ArF excimer laser exposure light, provided that the phase-shift mask comprises the transparent substrate and semi-transparent layer pattern defined above. The phase-shift mask is preferably a mask that further comprises a light-blocking layer pattern formed on the semi-transparent layer pattern and that pattern has an optical density (OD value) at the wavelength of the ArF excimer laser exposure light that is adjusted to yield a desired overall optical density (OD value) when combined with that of the semi-transparent layer pattern.

[0167] The structure of the light-blocking layer pattern in the present invention is equivalent to that of the light-blocking layer in the present invention described under point “A. Mask blank, 1. Parts of the mask blank, (3) Light-blocking layer”, except that the structure is formed in a pattern shape. Therefore, any description thereof is omitted here. (4) Other parts

[0168] The phase-shift mask of the present invention is not particularly limited, provided that the mask comprises the semi-transparent layer and the transparent substrate defined above. Other optional components can be suitably added. 2. Structure of the phase shift mask

[0169] The following describes the structure of the phase-shift mask of the present invention. The phase-shift mask of the present invention is a mask in which the semi-transparent layer pattern defined above is formed on the transparent substrate defined above. The structure of the phase-shift mask of the present invention and a method for manufacturing the mask are described below. (1) Structure of the phase shift mask

[0170] The phase shift mask is not particularly restricted and is preferably a negative-type phase shift mask.

[0171] The term "negative-type phase-shift mask" refers to a phase-shift mask used in a negative-type resist process or negative development process in a wafer process, and this is a mask where monochrome inversion pattern data are used to form a semi-transparent layer pattern, this situation being different from that of a phase-shift mask used in a wafer process using a positive-type resist.

[0172] As described below under point “D. Method for producing a pattern-formed body using the phase-shift mask”, when a halftone phase-shift mask with a semi-transparent layer is used to transfer, for example, a fine pattern for contact holes, conductors, or other features onto a wafer, negative development makes it possible to transfer the fine pattern for the contact holes, conductors, or other features onto the wafer while easily avoiding a sidelobe phenomenon.

[0173] According to the present invention, the transmittance of the semi-transparent layer pattern is in the range of 15 to 38% at the wavelength of the ArF excimer laser exposure light, which is higher than in the prior art. Thus, during negative development, the phase effect is greater at the edges of the light-blocking regions, which correspond to a fine pattern for contact holes, conductors, or other features. This makes it easier to transfer the fine pattern for the contact holes, conductors, or other features onto the wafer via negative development than in conventional methods.

[0174] As described below under point "D. Method for producing a pattern-formed body using the phase-shift mask", a method is known in which, when a halftone phase-shift mask with a semi-transparent layer is used to transfer a fine pattern for contact holes, conductors, or other features onto a wafer, the semi-transparent layer is used as raw material in the halftone phase-shift mask to form not only a primary pattern corresponding to the fine pattern that actually needs to be resolved, but also an auxiliary pattern located close to the primary pattern that does not actually need to be resolved. This method reduces the variation in the critical dimension (CD) of the pattern when the mask is defocused.

[0175] In modern wafer transfer techniques for a halftone phase-shift mask pattern with a semi-transparent layer, it is necessary to create a semi-transparent layer pattern to adjust the width or length of each element of the primary pattern, as described above, within the range of 100 to 300 nm. Furthermore, in this case, the width or length of each element of an auxiliary pattern, as described above, must be 60 nm or less, as the auxiliary pattern will be poorly resolved if the width or length is too large.

[0176] If this phase-shift mask is a negative-type phase-shift mask as in the present invention, the main pattern and the auxiliary pattern must be formed as convex element patterns, each consisting of the semi-transparent layer. Accordingly, the auxiliary pattern becomes a convex element pattern consisting of one or more pattern elements, each having a width or length of 60 nm or less, and consisting of the semi-transparent layer. Conventionally, the thickness of any semi-transparent layer is large.For example, in the case of the formation of this auxiliary pattern, or any other convex element pattern consisting of one or more pattern elements, each having a width or length of 60 nm or less and consisting of the semi-transparent layer, the phase-shift mask cleaning step causes the pattern to flake off. This flake-off is a phenomenon in which the convex element pattern flakes off during cleaning using ultrasonic waves with a strong removal force in a cleaning solution. However, the present invention makes it possible to adjust the thickness of the semi-transparent layer to the range of 57 to 67 nm, which is smaller than the thickness of conventional semi-transparent layers.Accordingly, in the phase-shift mask, the height of the convex element pattern, which consists of pattern elements each with a width or length of 60 nm or less and each composed of the semi-transparent layer, is lower than in conventional halftone phase-shift masks. This makes it possible to avoid pattern flaking during the cleaning step of the phase-shift mask. Flaking is a phenomenon in which the convex element pattern flakes off when cleaned using ultrasonic waves with a strong force in a cleaning solution.

[0177] The structure of the phase-shift mask of the present invention is equivalent to the structure of the mask blank described under point “A. Mask blank, 2. Structure of the mask blank, (1) Structure of the mask blank”, with the exception of the points specifically described. Therefore, any description of these points is omitted here. (2) Method for producing the phase shift mask

[0178] The method for producing the phase-shift mask of the present invention is not particularly restricted, provided that the method, as this mask, can yield a desired phase-shift mask. In one example of the method for producing the phase-shift mask, a mask blank with the light-blocking layer defined above is initially produced. Next, an electron beam resist is applied to the light-blocking layer, and an electron beam imaging device is used to pattern-expose the workpiece with light. The workpiece is developed with a developing solution exclusive to the resist to form a resist pattern with a desired shape.Next, the resist pattern, shaped as desired, is used as a mask to dry-etch the light-blocking layer in a dry-etching apparatus using a desired gas. This shapes the light-blocking layer into a semi-transparent layer pattern, as described below. The light-blocking layer, shaped into the semi-transparent layer pattern, is then used as a mask to dry-etch the semi-transparent layer to form a semi-transparent layer pattern. Finally, an electron beam resist is applied to the light-blocking layer, which has been shaped into the semi-transparent layer pattern. An electron beam imaging apparatus is then used to expose the workpiece to light pattern by pattern.The workpiece is developed using a development solution exclusively for the resist to form a resist pattern with a desired shape. Next, the resist pattern with the desired shape is used as a mask to dry-etch the light-blocking layer, which has been processed into the shape of the semi-transparent layer pattern, in a dry-etching apparatus using a desired gas, thereby forming a light-blocking layer pattern. In this way, the phase-shifting mask of the present invention is obtained. D. Method for producing a pattern-formed body using the phase-shifting mask.

[0179] The following describes a method for producing a patterned body using the phase-shift mask of the present invention. The method for producing a patterned body using the phase-shift mask of the present invention comprises a step of forming a resist pattern by exposing a resist layer of a positive-type resist composition with ArF excimer laser exposure light using the phase-shift mask of the present invention, and resolving and removing an unexposed region by negative development.

[0180] Fig.Figures 14A to 14C are schematic process diagrams illustrating an example of a method for producing a pattern-formed body using the phase-shift mask of the present invention. In the method for producing a pattern-formed body using the phase-shift mask of the present invention, a positive-type resist composition is initially applied directly or via an intermediate layer 302 onto or over a substrate 301 to be processed, thereby forming a resist layer 303 on / over the substrate ( Fig. 14A). Next, a phase-shifting mask 200, made from the mask blank, is used to expose the resist layer 303 to light ( Fig. 14B).

[0181] The in Fig.The phase shift mask 200 illustrated in Figure 14B is a halftone phase shift mask onto which ArF excimer laser exposure light is to be directed. The one in Fig.The phase-shift mask 200, illustrated in Figure 14B, comprises a transparent substrate 201 and a semi-transparent layer pattern 202 with a single-layer structure formed on the transparent substrate 201 and consisting either of Si and N or Si, N, and O. The semi-transparent layer pattern 202 has an extinction coefficient of 0.2 to 0.45 at the wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light, and a transmittance of 15 to 38% at the wavelength of the ArF excimer laser exposure light. The semi-transparent layer pattern 202 also has a layer thickness of 57 to 67 nm.Furthermore, the semi-transparent layer pattern 202 has a main pattern 202a, which can be resolved onto a wafer and has a width or length of 100 to 300 nm, and an auxiliary pattern 202b, which cannot be resolved onto the wafer but supports the resolution of the main pattern 202a onto the wafer, and which consists of pattern elements, each with a width or length of 60 nm or less.

[0182] Next, the light-exposed resist layer 303 is developed with an organic solvent to dissolve and remove an unexposed region 303a of the resist layer 303, thereby forming a resist pattern 403 ( Fig. 14C). During this process, the phase shift mask 200 is used to form the resist pattern 403 by negative development.

[0183] A halftone phase-shift mask with a semi-transparent layer is used, for example, to transfer contact holes onto a wafer via positive development. The edges of the contact holes can be sharply defined due to the phase effect. However, at positions further away than the edges of the contact holes, the definition effect diminishes. Thus, even a region of the semi-transparent layer where the exposure light should be blocked transmits the exposure light, causing a sidelobe phenomenon that sensitizes any region of the layer other than the resist layer where the contact holes are to be formed.One method for preventing this sidelobe phenomenon is, for example, a method of placing a light-blocking layer on the semi-transparent layer in such a way that no exposure light is transmitted through a region of the semi-transparent layer where the exposure light should be blocked. However, this method requires the light-blocking layer to be positioned at locations from the edges of the contact holes up to a predetermined distance in order to prevent the sidelobe phenomenon while maintaining the phase effect at the edges of the contact holes. Specifically, it is necessary to position the light-blocking layer at precise locations based on a predetermined rule after optical near-field correction (OPC) processing has been applied.This requires complex data processing, making it difficult to create a halftone phase shift mask.

[0184] When a halftone phase-shift mask with a semi-transparent layer is used, for example, to transfer contact holes onto a wafer via negative development, parts of the resist composition onto which the exposure light is directed are not easily resolved. Accordingly, in the halftone phase-shift mask, regions corresponding to contact holes become light-shielding regions, composed of the semi-transparent layer to block the exposure light. The size of the light-shielding region is only a few tens of nanometers on the wafer. Even though parts of the light-shielding regions corresponding to the contact holes transmit the exposure light, rays of the exposure light interfere with each other at the edges due to the phase effect and are canceled out. As a result, the intensity of the exposure light becomes zero.This prevents the resist composition region portions, where the exposure light should be blocked by the light-shielding region, from being exposed to the light. Therefore, it is possible to avoid the sidelobe phenomenon and easily transfer the contact holes onto the wafer.

[0185] According to the present invention, the transmittance of the semi-transparent layer is in the range of 15 to 38% at a wavelength of the ArF excimer laser exposure light, which is higher than that of conventional semi-transparent layers. Thus, during negative development, the phase effect is greater at the edges of the light-blocking regions, which correspond to a fine pattern, such as that for contact holes. This makes it easier to transfer the fine pattern, such as that for contact holes, onto a wafer by negative development than in conventional cases.

[0186] A method is known in which, when a halftone phase-shift mask with a semi-transparent layer is used to transfer a fine pattern for contact holes, conductors, or other features onto a wafer, the semi-transparent layer is used as raw material in the halftone phase-shift mask to form not only a primary pattern corresponding to the fine pattern that actually needs to be resolved, but also an auxiliary pattern located close to the primary pattern that does not actually need to be resolved. This method makes it possible to assist pattern resolution of the primary pattern by diffraction light generated by the auxiliary pattern, thereby improving the exposure latitude of the mask. When the mask is defocused, this reduces variation in the critical dimension (CD) of the pattern.

[0187] In modern wafer transfer techniques for a halftone phase-shift mask with a semi-transparent layer, when a finer pattern than described above is transferred to the wafer for contact holes, conductors, and other features to form a semi-transparent layer pattern within the halftone phase-shift mask, the width or length of each element of the primary pattern, as described above, must be set within the range of 100 to 300 nm. Furthermore, in this case, the variation in the critical dimension (CD) of the pattern at the time of mask defocusing can be reduced more significantly because the width or length of each element of the auxiliary pattern, as described above, is larger. However, if the width or length is too large, the auxiliary pattern will be poorly resolved; therefore, the width or length must be 60 nm or less.

[0188] In a wafer-based process for transferring a mask pattern of a halftone phase-shift mask with a semi-transparent layer onto a wafer by positive development, a primary pattern and an auxiliary pattern, as described above, are formed as concave region patterns where the semi-transparent layer is partially hollowed out. In contrast, in a wafer-based process for transferring a mask pattern of a halftone phase-shift mask with a semi-transparent layer onto a wafer by negative development, a primary pattern and an auxiliary pattern, as described above, must be formed as convex element patterns from the semi-transparent layer. Therefore, the auxiliary pattern is a convex element pattern consisting of one or more elements with a width or length of 60 nm or less, formed from the semi-transparent layer.

[0189] The thickness of conventional semi-transparent layers is large; this causes the pattern to flake off, a phenomenon that occurs when such a convex element pattern is cleaned using ultrasonic waves with a strong removal force in a cleaning solution.

[0190] In contrast, the thickness of the semi-transparent layer in the mask blank is in the range of 57 to 67 nm, which is smaller than that of conventional semi-transparent layers. Thus, in a phase-shift mask used for a method of producing a pattern-formed body according to the present invention, the height of the auxiliary pattern, which is a convex element pattern, is lower than in conventional halftone phase-shift masks, as described above. For this reason, pattern flaking can be avoided in the method of producing a pattern-formed body according to the present invention. Flaking is a phenomenon in which the auxiliary pattern, which is the convex element pattern, flakes off when this auxiliary pattern is cleaned using ultrasonic waves with a strong removal force in a cleaning solution, for the same reasons as described above for the mask blank.

[0191] In the present invention, the resist composition used to form the resist pattern is not particularly restricted, provided that the composition is one from which the resist pattern can be formed by negative development. The resist composition used to form the resist pattern can be a positive-type resist composition or a negative-type resist composition. The positive-type resist composition is preferred because it has a higher resolving power than the negative-type resist composition.

[0192] The positive-type resist composition is not particularly restricted and is, for example, a product, TOK 6063™, manufactured by Tokyo Ohka Kogyo Co., Ltd.

[0193] When the positive-type resist composition is used, a resist pattern is formed by developing the positive-type resist composition with an organic solvent developer to cause its exposed region to react with the organic solvent, thereby reducing the solution rate of it, and its unexposed region to be dissolved and removed.

[0194] Furthermore, the present invention is not limited to the embodiments specified above. These embodiments serve only as examples. Examples

[0195] The present invention will be described in more detail below with reference to examples and comparative examples. [Example 1]

[0196] Example 1 is a halftone phase-shift mask onto which ArF excimer laser exposure light is directed, and is a mask comprising a transparent substrate and a semi-transparent layer pattern formed on the transparent substrate. The phase-shift mask of Example 1 further comprises a light-blocking layer pattern formed on the semi-transparent layer pattern, and is a pattern in which the optical density (OD value) is set at a wavelength of the ArF excimer laser exposure light, resulting in a total of 3 when combined with that of the semi-transparent layer pattern.

[0197] In Example 1, the semi-transparent layer pattern consists solely of Si3N4. As shown in Table 3 below, the pattern has an extinction coefficient of 0.20 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.70 at the wavelength of the ArF excimer laser exposure light. As a result, for the semi-transparent layer pattern, the layer thickness required to achieve opposite phase is 57 nm, and the transmittance at the wavelength of the ArF excimer laser exposure light is 38%.

[0198] Furthermore, in non-patent literature (Refractive Index List / Refractive Index List for Thin Film Measurement [online]. [Accessed 2014-07-03]. Retrieved from the Internet: <URL: (http: / / www.filmetricsinc.jp / refractive-index-database>) shows that the extinction coefficient of Si3N4 at the wavelength of the ArF excimer laser exposure light is 0.20 and the refractive index of Si3N4 at the wavelength of the ArF excimer laser exposure light is 2.70. Therefore, it is clear that for the present semi-transparent layer pattern, the extinction coefficient at the wavelength of the ArF excimer laser exposure light is 0.20 and the refractive index at the wavelength of the ArF excimer laser exposure light is 2.70. [Example 2]

[0199] Example 2 is a halftone phase-shift mask onto which ArF excimer laser exposure light is directed, and is a mask comprising a transparent substrate and a semi-transparent layer pattern formed on the transparent substrate. The phase-shift mask of Example 2 further comprises a light-blocking layer pattern formed on the semi-transparent layer pattern, and is a pattern in which the optical density (OD value) is set at the wavelength of the ArF excimer laser exposure light, resulting in a total of 3 when combined with that of the semi-transparent layer pattern.

[0200] In Example 2, the extinction coefficient and refractive index of the semi-transparent layer pattern at the wavelength of the ArF excimer laser exposure light are calculated to yield values ​​that allow the layer thickness of the semi-transparent layer pattern required to achieve antiphase in the range of 57 to 67 nm, and furthermore allow the transmittance of the semi-transparent layer pattern to be in the range of 15 to 38% at the wavelength of the ArF excimer laser exposure light. In Example 2, the following semi-transparent layer patterns are selected: a semi-transparent layer pattern consisting only of Si and N (SiN-based layer pattern); or a semi-transparent layer pattern consisting only of Si, N, and O (SiON-based layer pattern).Furthermore, in Example 2, the extinction coefficient and refractive index values ​​of the semi-transparent layer pattern at the wavelength of the ArF excimer laser exposure light are obtained as values ​​that the semi-transparent layer pattern consisting only of Si and N (SiN-based layer pattern) or the semi-transparent layer pattern consisting only of Si, N and O (SiON-based layer pattern) can exhibit.

[0201] In Example 2, the semi-transparent layer pattern consists only of Si and N, or only of Si, N, and O (SiN-based layer pattern or SiON-based layer pattern). As shown in Table 3 below, the pattern has an extinction coefficient of 0.45 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.70 at the wavelength of the ArF excimer laser exposure light. As a result, for the semi-transparent layer pattern, the layer thickness required to achieve opposite phase is 58 nm, and the transmittance at the wavelength of the ArF excimer laser exposure light is 15%. [Example 3]

[0202] Example 3 differs from Example 2, as shown below in Table 3, in that the semi-transparent layer pattern has an extinction coefficient of 0.35 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.60 at the wavelength of the ArF excimer laser exposure light. As a result, Example 3 differs from Example 2 in that, for the semi-transparent layer pattern, the layer thickness required to achieve a counter-phase is 60 nm and the transmittance at the wavelength of the ArF excimer laser exposure light is 20%. Example 3 is identical to the phase-shift mask of Example 2 except for these points. [Example 4]

[0203] Example 4 differs from Example 2, as shown below in Table 3, in that the semi-transparent layer pattern has an extinction coefficient of 0.30 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.50 at the wavelength of the ArF excimer laser exposure light. As a result, Example 4 differs from Example 2 in that, for the semi-transparent layer pattern, the layer thickness required to achieve opposite phase is 63 nm, and the transmittance at the wavelength of the ArF excimer laser exposure light is 25%. Example 4 is identical to the phase-shift mask of Example 2 except for these points. [Example 5]

[0204] Example 5 differs from Example 2, as shown below in Table 3, in that the semi-transparent layer pattern has an extinction coefficient of 0.25 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.40 at the wavelength of the ArF excimer laser exposure light. As a result, Example 5 differs from Example 2 in that, for the semi-transparent layer pattern, the layer thickness required to achieve an opposite phase is 67 nm, and the transmittance at the wavelength of the ArF excimer laser exposure light is 30%. Example 5 is identical to the phase-shift mask of Example 2 except for these points. [Comparison example 1]

[0205] Comparative Example 1 is a halftone phase-shift mask onto which ArF excimer laser exposure light is directed, and is a mask comprising a transparent substrate and a semi-transparent layer pattern formed on the transparent substrate. The phase-shift mask of Comparative Example 1 further comprises a light-blocking layer pattern formed on the semi-transparent layer pattern, and is a pattern in which the optical density (OD value) is set at a wavelength of the ArF excimer laser exposure light, resulting in a total of 3 when combined with that of the semi-transparent layer pattern.

[0206] In comparative example 1, the semi-transparent layer pattern consists of a MoSiON-based material. As shown in Table 3 below, the pattern has an extinction coefficient of 0.59 at the wavelength of the ArF excimer laser exposure light and a refractive index of 2.34 at the wavelength of the ArF excimer laser exposure light. As a result, for the semi-transparent layer pattern, the layer thickness required to achieve opposite phase is 68 nm, and the transmittance at the wavelength of the ArF excimer laser exposure light is 6%.

[0207] Furthermore, in examples 1 to 5, it is assumed that the light-blocking layer is a layer consisting of a simple chromium substance. As in any of Fig. 6 and Fig.Figure 7 illustrates that if the light-shielding layer is composed of several layers, such as two or three layers, the same phase-shifting mask as described above can be produced by selecting an etching gas, etching conditions and other suitable options to etch each of the layers. [Evaluation 1]

[0208] For the phase-shift masks of Examples 1 to 5 and Comparison Example 1, the maximum exposure latitude (maxEL) and the maximum depth of focus (maxDoF) were evaluated by simulations. Specifically, under the simulation evaluation conditions described below, the evaluations were performed by calculations using Kirchhoff's method as the algorithm to obtain the respective transfer properties of the mask patterns, and EM-Suite™, manufactured by Panoramic Technology Inc., as the simulation software. <simulationsevaluierungsbedingungen> • NA: 1.35 • Sigma: c-quad 0.95 / 0.80-30 degrees • Polarization: X / Y • Target: 60 nm hole (NTD) • Spacing: 180, 240 and 300 nm respectively

[0209] NA, sigma, and polarization were used as effective illumination conditions that are realistically applicable for transferring a target shape (60 nm hole (NTD)) and spacing (180, 240, and 300 nm, respectively). The evaluation results are shown in Table 3 below. For the phase-shift mask of each of Examples 1 to 5 and the comparison example 1, Table 3 also shows the raw material of the semi-transparent layer pattern, the refractive index (n) and extinction coefficient (k) of the semi-transparent layer pattern at the wavelength of the ArF excimer laser exposure light, the layer thickness (d) of the semi-transparent layer pattern to provide a counter-phase, and the transmittance (trans) of the semi-transparent layer pattern at the wavelength of the ArF excimer laser exposure light. [Table 3] Raw material n k d Trans[%] maxEL[%] maxDoF[um] Example 1 Si3N4 2,70 0,20 57nm 38 8,95 0,052 Example 2 based on SiN or SiON 2,70 0,45 58nm 15 8,23 0,049 Example 3 based on SiN or SiON 2,60 0,35 60nm 20 8,51 0,050 Example 4 based on SiN or SiON 2,50 0,30 63nm 25 9,23 0,051 Example 5 based on SiN or SiON 2,40 0,25 67nm 30 9,15 0,053 Comparison Example 1 based on MoSiON 2,34 0,59 68nm 6 7,69 0,047 [Evaluation 2]

[0210] Fig. Figure 15 is a graphical representation that represents a simulation result of the OPC bias value of each of the phase shift masks versus its transmittance.

[0211] Out of Fig. 15 shows that in a suitable illumination system for 1 × nodes, the OPC bias value of the phase shift masks, in each of which the transmittance of the semi-transparent layer is 15% or more, is smaller than in the phase shift mask of comparison example 1, in which the transmittance of the semi-transparent layer is 6%. (Evaluation 3)

[0212] Fig. Figure 16 is a representation showing respective XY images of exposure light intensity distributions, each present on a wafer, where the distributions are obtained by a simulator, and further showing graphs that each depict the respective exposure light intensities of the distributions. The lower positions of Fig. The 16 XY images shown are respective XY images of the exposure light intensity distributions present on the wafer, the distributions being obtained by a calculation for each of the phase-shift masks and further for each of their different spacings. The above in Fig. The 16 graphs shown are each a graph representing the intensity of the exposure light sent through each of the phase-shift masks at each position along the transverse axis of the XY image of the exposure light intensity distribution shown in the lower position, assuming that the intensity of the exposure light not sent through the semi-transparent layer is considered to be 1.0.

[0213] Out of Fig. 16 shows that for the individual distances the image contrast calculated under conditions of assuming one of the phase shift masks described above, in which the transmittance of the semi-transparent layer is 38%, is higher than that calculated under conditions of assuming the other of the phase shift masks, in which the transmittance of the semi-transparent layer is 6%. [Evaluation 4]

[0214] Fig. Figures 17-1 to 17-3 are each a graph that, as simulation results, represents a relationship between the depth of focus and the exposure latitude of the phase-shift mask of each of Example 1 and Comparison Example 1 when the pattern is transferred. The in Fig. The graphs shown in Figures 17-1 to 17-3 are graphs for the phase-shift masks where the hole spacing is 180 nm, 240 nm, and 300 nm, respectively, and where their transverse axis represents the depth of focus (DOF) and their vertical axis represents the exposure latitude (EL). The EL (%) is shown in Table 4 when the DOF is 0 nm, and the DOF (nm) is shown in Table 5 when the EL is 10%. [Table 4] Distance 180nm 240nm 300nm 6% PSM 16,08 14,43 14,00 38% PSM 25,13 23,80 23,28 EL improvement ratio 56 % 65 % 66 % [Table 5] Distance 180nm 240nm 300nm 6% PSM 37,2 26,5 24,0 38% PSM 58,7 45,3 44,3 DOF improvement ratio 58 % 71 % 85 %

[0215] Out of Fig. Reference 17-1 and Tables 4 and 5 show that, in a case where the spacing is 180 nm, for the EL, when the DOF is 0 nm, the calculated result of the mask with a transmittance of 38% is approximately 56% greater than that of the mask with a transmittance of 6%; and for the DOF, when the EL is 10%, the calculated result of the mask with a transmittance of 38% is approximately 58% greater than that of the mask with a transmittance of 6%. Fig. Reference 17-2 and Tables 4 and 5 show that even in a case where the spacing is 240 nm, for the EL, when the DOF is 0 nm, the calculated result for the mask with a transmittance of 38% is 65% greater than that for the mask with a transmittance of 6%; and for the DOF, when the EL is 10%, the calculated result for the mask with a transmittance of 38% is 71% greater than that for the mask with a transmittance of 6%. Furthermore, it is shown from Fig. 17-3 and Tables 4 and 5 show that even in a case where the distance is 300 nm, for the EL, when the DOF is 0 nm, the calculated result of the mask in which the transmittance is 38% is 66% greater than that of the mask in which the transmittance is 6%; and for the DOF, when the EL is 10%, the calculated result of the mask in which the transmittance is 38% is 85% greater than that of the mask in which the transmittance is 6%.

[0216] Therefore, it is clear that the calculation results of the mask in which the transmittance is 38% are greater, both in terms of DOF and EL, than those of the mask in which the transmittance is 6%. [Evaluation 5]

[0217] Fig. Figure 18 is a graphical representation that depicts the contrast of optical spatial images transferred to a wafer, calculated under conditions of assuming each phase-shift mask in which the semi-transmittance of the layer is 38% and the phase-shift mask in which the transmittance of the semi-transmittance is 6%. In the graphical representation of Fig. 18 its transverse axis represents the spacing of the pattern formed on the wafer, and its vertical axis represents the image contrast.

[0218] Out of Fig. Figure 18 shows that for each of the spacings of the pattern formed on the wafer, the phase-shift mask in which the transmittance of the semi-transparent layer is 38% has a higher image contrast than the phase-shift mask in which the transmittance of the semi-transparent layer is 6%. In other words, it is shown that the greater image contrast leads to the following: even if the exposure quantity is changed (the spatial images in the cutting height are changed), a variation in the critical dimension of the pattern formed on the wafer is advantageously small (the EL is large). [Evaluation 6]

[0219] Fig. Figure 19 is a graphical representation of the OPC bias resulting from a calculation under conditions assuming each phase-shift mask in which the transmittance of the semi-transparent layer is 38%, and the phase-shift mask in which the transmittance of the semi-transparent layer is 6%. In the graphical representation of Fig. 19 represents its transverse axis the spacing of the pattern formed on the wafer, and its vertical axis represents the critical dimension (CD) of the phase shift mask pattern with the OPC bias.

[0220] Out of Fig. Figure 19 shows that for each of the spacings of the pattern formed on the wafer, the phase-shift mask in which the transmittance of the semi-transparent layer is 38% has a smaller critical dimension (CD) of the pattern of the mask with the OPC bias than the mask in which the transmittance of the semi-transparent layer is 6%. In other words, it is shown that the smaller OPC bias makes it possible to form a fine pattern on the wafer, thereby increasing the formation flexibility of a pattern to be formed on the wafer. [Evaluation 7]

[0221] Individual semi-transparent layer patterns of phase-shift masks of the present invention were evaluated with respect to their resistance to flaking. Specifically, before and after each of the phase-shift masks, each having a semi-transparent layer pattern with approximately one billion convex pattern elements, was cleaned using ultrasonic waves in a cleaning solution, a mask testing apparatus was used to perform a comparative examination. According to this examination, before and after cleaning, it was investigated whether the convex pattern elements had flaked off. The number of chips of the convex pattern elements was counted, and the size of the chips of the convex pattern elements was examined.

[0222] The semi-transparent coating patterns tested for peel resistance had two thicknesses (60 nm and 75 nm). The semi-transparent coating patterns tested for peel resistance at any of the thicknesses had convex pattern elements of six sizes (width × length: 60 nm × 150 nm, 60 nm × 300 nm, 60 nm × 600 nm, 85 nm × 150 nm, 85 nm × 300 nm, and 85 nm × 600 nm). The peel resistance evaluation was performed under the cleaning conditions described below. Furthermore, a high level of ultrasonic cleaning resulted in twice the physical removal force of a low level. <reinigungsbedingungen> • Cleaning condition: Ultrasonic wave cleaning • Ultrasound conditions: two levels (low and high level) • Number of cleaning cycles: Estimate the number in intervals of five times.

[0223] For the semi-transparent layer patterns, each with a layer thickness of 60 nm, Table 6 shows results obtained by examining whether their cleaned convex element pattern had flaked off. For each of the convex pattern element sizes, Table 6 shows whether the convex pattern elements had flaked off under each of the ultrasonic conditions. In Table 6, each case where flaking of the convex element pattern was induced is represented by ×; and each case where flaking of the convex element pattern was not induced is represented by ◯. [Table 6] Size of the convex pattern element Ultrasound conditions Layer- Width (nm) length Low High thickness (nm) (nm) level level 60 60 150 ◯ ◯ 60 60 300 ◯ ◯ 60 60 600 ◯ ◯ 60 85 150 ◯ ◯ 60 85 300 ◯ ◯ 60 85 600 ◯ ◯

[0224] Similarly, for the semi-transparent layer patterns, each with a layer thickness of 75 nm, Table 7 shows results obtained by examining whether their cleaned convex element pattern had flaked off. For each of the convex pattern element sizes, Table 7 shows whether the convex pattern elements had flaked off under each of the ultrasonic conditions. In Table 7, each case where flaking of the convex element pattern was induced is represented by ×; and each case where flaking of the convex element pattern was not induced is represented by ◯. [Table 7] Size of the convex pattern element Ultrasound conditions Layer thickness (nm) Width (nm) Length (nm) Low level High level 75 60 150 ◯ × 75 60 300 ◯ × 75 60 600 ◯ × 75 85 150 ◯ ◯ 75 85 300 ◯ ◯ 75 85 600 ◯ ◯

[0225] Table 6 shows that, in the semi-transparent layer patterns with a layer thickness of 60 nm, the convex pattern elements of all sizes did not delaminate under either ultrasound condition. However, Table 7 shows that, in the semi-transparent layer patterns with a layer thickness of 75 nm, the convex pattern elements of the 60 nm size delaminated to a high degree under the ultrasound condition. The incidence of delamination of the convex pattern elements ranged from a few parts per billion to several tens of parts per billion.

[0226] Such results indicate that if the layer thickness of a semi-transparent layer pattern, such as the semi-transparent layer pattern of the phase-shift mask of Example 3, is 60 nm (within the range of 57 to 67 nm), flaking of the pattern can be avoided, where flaking is a phenomenon in which its convex pattern elements, each having a width or length of 60 nm (60 nm or less) and consisting of the semi-transparent layer, are cleaned in a cleaning solution using ultrasonic waves with a strong removal force, causing the convex pattern elements to flake off.However, it is clear that if the layer thickness of a semi-transparent layer pattern is 75 nm (outside the range of 57 to 67 nm), flaking of the pattern cannot be avoided, where flaking is a phenomenon in which its convex pattern elements, each having a width or length of 60 nm (60 nm or less) and consisting of the semi-transparent layer, are cleaned using ultrasonic waves with a strong removal force in a cleaning solution, causing the convex pattern elements to flake off. List of reference figures

[0227] 100: Mask blank, 101: Transparent substrate, 102: Semi-transparent layer, 103: Light-blocking layer, 110: Mask blank attached to a negative-type resist layer, 200: Phase-shift mask, 201: Transparent substrate, and 202: Semi-transparent layer pattern.< / reinigungsbedingungen> < / simulationsevaluierungsbedingungen>

Claims

[1] Halftone phase shift mask (200) onto which ArF excimer laser exposure light is directed, wherein the mask (200) comprises a transparent substrate (201) and a semi-transparent layer pattern (202) formed on the transparent substrate (201) and consisting only of Si and N, or a semi-transparent layer pattern (202) formed on the transparent substrate (201) and consisting only of Si, N and O, characterized by , that the semi-transparent layer pattern (202) has an extinction coefficient of 0.2 to 0.45 at a wavelength of the ArF excimer laser exposure light, a refractive index of 2.3 to 2.7 at the wavelength of the ArF excimer laser exposure light and a transmittance of 15% to 38% at the wavelength of the ArF excimer laser exposure light and furthermore has a layer thickness of 57 nm to 67 nm, the semi-transparent layer pattern (202) has a main pattern (202a) that is to be resolved onto a wafer, and an auxiliary pattern (202b) that is used to assist in the resolution of the main pattern (202a) and is not to be resolved onto the wafer, the auxiliary pattern (202b) comprises one or more convex pattern elements, each with a width or length of 60 nm or less, and the main pattern (202a) has the same thickness as the auxiliary pattern (202b). [2] Phase shift mask (200) according to claim 1, characterized by , that the semi-transparent layer pattern (202) is formed directly on the transparent substrate (201). [3] Phase shift mask (200) according to claim 1 or 2, which is a negative-type phase shift mask. [4] Phase shift mask (200) according to one of claims 1 to 3, further comprising a light-shielding layer pattern formed on the semi-transparent layer pattern (202), wherein an optical density (OD value) of this light-shielding layer pattern is set at the wavelength of the ArF excimer laser exposure light, to give a desired overall value (OD value) when combined with an optical density of the semi-transparent layer pattern (202). [5] Phase shift mask (200) according to claim 4, characterized by , that the light-shielding layer pattern has a single-layer structure comprising a light-absorbing layer pattern formed on the semi-transparent layer pattern (202) and having an etching barrier function for the semi-transparent layer pattern (202) and a light-absorbing function for absorption of the ArF excimer laser exposure light. [6] Phase shift mask (200) according to claim 4, characterized by , that the light-shielding layer pattern has a double-layer structure comprising: an etch barrier layer pattern formed on the semi-transparent layer pattern (202) and having an etch barrier function for the semi-transparent layer pattern (202); and a light-absorbing layer pattern formed on the etch barrier layer pattern and having a light-absorbing function for absorbing the ArF excimer laser exposure light. [7] Phase shift mask (200) according to claim 6, characterized by , that the light-absorbing layer pattern comprises simple silicon (Si). [8] Phase shift mask (200) according to one of claims 4 to 7, characterized by, that the optical density (OD value) of the light-shielding layer pattern is set at the wavelength of the ArF excimer laser exposure light, resulting in a total value of 3.0 or more when combined with the optical density of the semi-transparent layer pattern (202). [9] Method for producing a pattern-formed body, comprising a step of forming a resist pattern (403) by exposing a resist layer (303) of a positive-type resist composition with ArF excimer laser exposure light using the halftone phase-shift mask (200) according to any one of claims 1 to 8, and dissolving and removing an unexposed region (303a) by negative development.

Images