JP6140330B2 - Mask blank manufacturing method, transfer mask manufacturing method, and semiconductor device manufacturing method - Google Patents

Description

  The present invention relates to a method for manufacturing a mask blank having a low-stress thin film and a method for manufacturing a transfer mask. In particular, the present invention relates to a method for manufacturing a mask blank and a method for manufacturing a transfer mask in which a change in stress of a thin film with time is reduced. The present invention also relates to a method for manufacturing a semiconductor device using the transfer mask.

  In general, in a manufacturing process of a semiconductor device, a fine pattern is formed using a photolithography method. Further, a number of substrates called transfer masks are usually used for forming this fine pattern. This transfer mask is generally provided with a fine pattern made of a metal thin film on a translucent glass substrate, and the photolithographic method is also used in the production of this transfer mask.

  In manufacturing a transfer mask by photolithography, a mask blank having a thin film (for example, a light shielding film) for forming a transfer pattern (mask pattern) on a light-transmitting substrate such as a glass substrate is used. In the manufacture of a transfer mask using this mask blank, an exposure process for drawing a desired pattern on the resist film formed on the mask blank, and developing the resist film in accordance with the desired pattern drawing, the resist pattern is developed. The development step is formed, the etching step is to etch the thin film in accordance with the resist pattern, and the step is to remove and remove the remaining resist pattern. In the developing step, a resist pattern formed on the mask blank is subjected to a desired pattern drawing (exposure), and then a developing solution is supplied to dissolve a portion of the resist film that is soluble in the developing solution. Form. Further, in the etching step, using this resist pattern as a mask, the portion where the resist pattern is not formed and the thin film is exposed is dissolved by dry etching or wet etching, whereby the desired mask pattern is formed on the translucent substrate. To form. Thus, a transfer mask is completed.

  When miniaturizing a pattern of a semiconductor device, it is necessary to shorten the wavelength of an exposure light source used in photolithography in addition to miniaturization of a mask pattern formed on a transfer mask. In recent years, as an exposure light source for manufacturing semiconductor devices, the wavelength has been shortened from an KrF excimer laser (wavelength 248 nm) to an ArF excimer laser (wavelength 193 nm).

As a transfer mask, a binary mask having a light-shielding film pattern made of a chromium-based material on a translucent substrate has been known.
In recent years, a binary mask for an ArF excimer laser using a material containing a molybdenum silicide compound (MoSi-based material) as a light shielding film has also appeared (Patent Document 1). In addition, a binary mask for ArF excimer laser using a material containing a tantalum compound (tantalum-based material) as a light shielding film has appeared (Patent Document 2). In Patent Document 3, when a photomask made of a light-shielding film using a metal containing at least two of tantalum, niobium, vanadium or tantalum, niobium, and vanadium is subjected to acid cleaning or hydrogen plasma cleaning, It is disclosed that the film becomes hydrogen embrittled and the light shielding film may be deformed. Further, as a means for solving this problem, it is disclosed that after forming a pattern on the light shielding film, a hydrogen blocking film is formed that airtightly covers the upper surface and side surfaces of the light shielding film.

On the other hand, Patent Document 4 describes a manufacturing method of a mask substrate for vacuum ultraviolet light made of synthetic quartz glass. Here, in order to improve the light transmittance of the synthetic quartz glass in the vacuum ultraviolet region, the necessity of reducing the OH groups in the synthetic quartz glass is shown. As one of the solutions, it is disclosed to reduce the Si—H content or the H ₂ content in the synthetic quartz glass to a predetermined value or less.

JP 2006-78807 A JP 2009-230112 A JP 2010-192503 A JP 2004-26586 A

  In recent years, the required level of pattern position accuracy for a transfer mask has become particularly strict. One factor for achieving high pattern position accuracy is to improve the flatness of a mask blank that is an original for producing a transfer mask. In order to improve the flatness of the mask blank, first, it is necessary to improve the flatness of the main surface on the side where the thin film of the glass substrate is formed. The glass substrate for manufacturing the mask blank starts from manufacturing a glass ingot as described in Patent Document 4 and cutting it into the shape of the glass substrate. The glass substrate immediately after being cut out has poor main surface flatness and a rough surface. For this reason, a multi-step grinding process and polishing process are performed on the glass substrate, and the surface is finished with high flatness and good surface roughness (mirror surface). In addition, after the polishing process using the abrasive grains, cleaning with a cleaning solution containing a hydrofluoric acid solution or a silicic hydrofluoric acid solution is performed. Further, cleaning with a cleaning solution containing an alkaline solution may be performed before the step of forming a thin film.

  However, it is not enough to produce a mask flat with high flatness. When the film stress of the thin film for forming the pattern formed on the main surface of the glass substrate is high, the substrate is deformed and the flatness is deteriorated. For this reason, various measures have been taken during film formation or after film formation in order to reduce the film stress of the thin film for forming the pattern. Until now, a mask blank that has been adjusted to have a high flatness by taking such measures will be flat if it is stored in a case even after being stored for a relatively long period of time (for example, about half a year). Was not expected to change significantly. However, in the case of a mask blank using a material containing tantalum for the thin film for pattern formation, it is confirmed that the flatness of the main surface deteriorates as time elapses even if it is hermetically stored in the case. It was done. Specifically, with the passage of time, the flatness of the main surface on the side where the thin film is formed has deteriorated in a direction in which the tendency of the convex shape becomes stronger.

  This means that when the glass substrate is not the cause, the film stress of the thin film gradually increases the tendency of compressive stress. In the case of a mask blank having a thin film using a chromium-based material or a material containing a molybdenum silicide compound, such a remarkable phenomenon does not occur. For this reason, this phenomenon that occurs in mask blanks using tantalum-containing materials for pattern forming thin films is not due to the deformation of the glass substrate itself, but the compressive stress of the thin film increases with time. It is guessed that it will become. On the other hand, when a transfer mask is produced using such a thin film mask blank having a high compressive stress, there is a problem that a large positional shift of the pattern occurs in the thin film area freed from the film stress by the pattern formation. is there. Furthermore, even when a transfer mask is manufactured in a short period of time after the mask blank is manufactured, there is a problem in that pattern displacement occurs as time passes after the manufacture.

  The present invention has been made under such circumstances, and the object of the present invention is to provide a mask blank using a material containing tantalum as a thin film for pattern formation. An object of the present invention is to provide a method for manufacturing a mask blank and a method for manufacturing a transfer mask, which solve the problem that the tendency of compressive stress becomes strong and suppress the deterioration of flatness. It is another object of the present invention to provide a method for manufacturing a semiconductor device using the transfer mask.

  In order to achieve the above object, the mask blank manufacturing method of the present invention includes a step of preparing a glass substrate that has been mirror-polished on the main surface, and a step of performing a heat treatment on the prepared glass substrate. And a step of forming a thin film made of a material containing tantalum and substantially not containing hydrogen on the main surface of the glass substrate after the heat treatment. Disclosed is a mask blank manufacturing method according to the present invention, which eliminates the problem that the film stress of a thin film has a tendency of compressive stress to increase with time and suppresses deterioration of flatness. Can do.

  Moreover, the manufacturing method of the mask blank of this invention is the process of preparing the glass substrate which mirror-polished with respect to the main surface, the process of heat-processing with respect to the prepared said glass substrate, and after heat processing A step of placing a glass substrate in a deposition chamber, using a tantalum-containing target, introducing a sputtering gas not containing hydrogen into the deposition chamber, and forming a thin film on the main surface of the glass substrate by a sputtering method; It is characterized by providing. In the mask blank manufacturing method of the present invention, by using a predetermined sputtering gas, the problem that the film stress of the thin film has a tendency to increase in compressive stress with the passage of time is eliminated, and deterioration of flatness is suppressed. A method for manufacturing a mask blank can be provided.

  In each mask blank manufacturing method, the heat treatment performed on the glass substrate is preferably a treatment of heating the glass substrate to 300 ° C. or higher. By heating the glass substrate to 300 ° C. or higher, hydrogen can be sufficiently removed out of the substrate.

  In each mask blank manufacturing method, the thin film is preferably formed in contact with a main surface of a glass substrate. In the case where the thin film made of a material containing tantalum is in contact with the main surface of the glass substrate after the heat treatment, hydrogen or the like present in the glass substrate directly enters the thin film. In the case of such a configuration, the effect obtained by eliminating OH groups, hydrogen, water, and the like from the glass substrate by heat treatment can be obtained more significantly.

  In each of the mask blank manufacturing methods, it is preferable that the thin film is made of a material containing tantalum and nitrogen and substantially not containing hydrogen. By including nitrogen in tantalum, oxidation of tantalum can be suppressed.

  In particular, in this mask blank manufacturing method, it is desirable that a highly oxidized layer containing 60 atomic% or more of oxygen is formed on the surface layer of the thin film. Since the high oxide film of the thin film material has a high binding energy, hydrogen in the gas surrounding the mask blank can be prevented from entering the thin film from the surface layer of the thin film.

  In each mask blank manufacturing method, the thin film has a structure in which a lower layer and an upper layer are laminated from the glass substrate side, and the lower layer contains tantalum and nitrogen, and does not substantially contain hydrogen. The upper layer is preferably made of a material containing tantalum and oxygen. With such a configuration, the upper layer can function as a film (antireflection film) having a function of controlling the surface reflectance with respect to the exposure light of the thin film.

  In particular, in this mask blank manufacturing method, it is desirable that a high oxide layer containing 60 atomic% or more of oxygen be formed on the upper surface layer. Since the high oxide film of the thin film material has high binding energy, hydrogen can be prevented from entering the thin film from the surface layer of the thin film.

  The transfer mask manufacturing method of the present invention is characterized in that a transfer pattern is formed on a thin film of the mask blank using the mask blank manufactured by the above-described mask blank manufacturing method. Since the flatness of the mask blank of the present invention is maintained at the required high level, the transfer mask manufactured using the mask blank having such characteristics can also have the required high flatness. .

  In particular, in this transfer mask manufacturing method, it is desirable that a highly oxidized layer containing 60 atomic% or more of oxygen be formed on the surface layer of the transfer pattern formed of the thin film. Since the high oxide film of the thin film material has high binding energy, it is possible to prevent hydrogen from entering the thin film of the transfer mask from the surface layer. Therefore, the transfer mask can have higher flatness.

  The semiconductor device manufacturing method of the present invention is characterized in that a transfer pattern is exposed and transferred onto a resist film on a semiconductor substrate using the transfer mask manufactured by the transfer mask manufacturing method. By using the transfer mask of the present invention and exposing and transferring a transfer pattern onto a resist film on a semiconductor substrate, a semiconductor device having a highly accurate pattern can be manufactured.

  According to the mask blank manufacturing method of the present invention, even in a mask blank in which a material containing tantalum is used for a thin film for pattern formation, the film stress of the thin film tends to increase in compressive stress with time. Absent. Thereby, after manufacturing a mask blank, it can suppress that flatness deteriorates with progress of time. Moreover, since the mask blank manufactured by the mask blank manufacturing method of the present invention can suppress the increase in the film stress of the thin film for pattern formation with the passage of time, the film stress of the thin film is the level at the time of manufacture. Can be maintained. Thereby, the large position shift of the pattern which arises when a transfer mask is produced from the mask blank which has a thin film with high film | membrane stress can be suppressed. Furthermore, when a transfer mask is produced from a mask blank produced by the method for producing a mask blank of the present invention, it is possible to suppress the occurrence of pattern displacement with the passage of time after production. Furthermore, the transfer pattern can be transferred to the resist film on the semiconductor substrate using a transfer mask in which deterioration of the flatness of the main surface due to the film stress of the thin film is suppressed and the positional deviation of the pattern formed on the thin film is also suppressed. . As a result, a semiconductor device having a fine and highly accurate circuit pattern on the semiconductor substrate can be manufactured.

It is sectional drawing which shows the structure of the mask blank concerning embodiment of this invention. It is sectional drawing which shows the structure of the transfer mask concerning embodiment of this invention. It is sectional drawing which shows the process until it manufactures the transfer mask from the mask blank concerning embodiment of this invention. It is a figure which shows the result by the HFS / RBS analysis of the mask blank in an Example. It is a figure which shows the result by the HFS / RBS analysis of the mask blank in a comparative example.

Embodiments of the present invention will be described below.
The present inventor has intensively studied the cause of the compressive stress increasing with time in a thin film containing tantalum immediately after film formation on a glass substrate. First, in order to confirm that there is no cause in the storage method of the mask blank after film formation, it was verified with various storage cases and storage methods. In either case, the flatness of the main surface of the mask blank deteriorated. Thus, no clear correlation was obtained. Next, the mask blank whose main surface flatness deteriorated in the convex direction was subjected to heat treatment using a hot plate. The heat treatment was performed at 200 ° C. for about 5 minutes. When this heat treatment was performed, the convex shape of the main surface temporarily changed in a slightly better direction. However, it was found that the flatness of the main surface of the mask blank deteriorates again as time passes after the heat treatment, and it does not lead to a fundamental solution.

  Next, the present inventor examined the possibility that the material containing tantalum is related to the property of easily taking in hydrogen. That is, it was hypothesized that hydrogen was gradually taken into the tantalum-containing thin film over time, and the compressive stress increased. However, this mask blank, in which the phenomenon that the compressive stress increases with the passage of time, has not found a factor for hydrogen incorporation in the conventional knowledge. The substrate used for the mask blank is formed of synthetic quartz glass, and a process that suppresses the inclusion of hydrogen was used when manufacturing the synthetic quartz ingot. In addition, a tantalum-containing thin film is formed by laminating a lower layer made of a material containing tantalum and nitrogen and an upper layer made of a material containing oxygen in tantalum formed on the lower layer on the main surface side of the substrate. It was a structure. Since a film made of a material containing oxygen in tantalum has an effect of suppressing the entry of hydrogen from the outside air, it has been thought that hydrogen in the outside air hardly enters the thin film containing tantalum.

  In order to confirm whether hydrogen has been taken into the tantalum-containing thin film over time after the film formation, the following verification was performed. Specifically, for a mask blank including a thin film made of a material containing tantalum, the film is stored in a case after the film is formed, and about two weeks have not passed so much, and the flatness of the thin film has deteriorated. 4 months have passed since the film was formed and housed in the case after film formation, and the compressive stress of the thin film increased and the flatness deteriorated (the side on the basis of the center of the substrate main surface was 142 mm) The film composition was analyzed for each of the mask blanks in which the amount of change in flatness was about 300 nm in the flatness by Coordinate TIR in the square inner region. For the membrane analysis, HFS / RBS analysis (hydrogen forward scattering analysis / Rutherford backscattering analysis) was used. As a result, in the thin film about 2 weeks after the film formation, the hydrogen content was below the lower limit of detection, whereas in the thin film after 4 months from the film formation, about 6 at% hydrogen was contained. Turned out to be.

  From these results, it was confirmed that the film stress was changed by hydrogen being taken into the tantalum-containing thin film after film formation. Next, the inventor suspected the substrate as a hydrogen generation source. After polishing is performed until the flatness and surface roughness of the main surface exceed the level required for the mask blank substrate, a heat-treated substrate is prepared, and a thin film made of a material containing tantalum is prepared. The film was deposited and verified in the same manner as described above. As a result, in the case of a mask blank using a substrate subjected to heat treatment, even when 4 months have elapsed after film formation, the degree of deterioration of flatness was small, and the hydrogen content in the film was also suppressed.

  Factors that exist in the mask blank glass substrate produced using a glass ingot produced by a manufacturing method in which OH groups and hydrogen are less likely to be mixed, include OH groups, hydrogen, water, and the like that are sources of hydrogen. There is a possibility.

  Usually, the conditions of the flatness and surface roughness of the main surface required for the substrate used for the mask blank are strict, and the conditions for a glass substrate for the mask blank are as it is cut out from the glass ingot into the shape of the glass substrate. It is difficult to satisfy. It is necessary to perform a grinding process and a polishing process in a plurality of stages on the cut glass substrate to finish the main surface with high flatness and surface roughness. Further, the polishing liquid used in the polishing step contains colloidal silica abrasive grains as an abrasive. Since colloidal silica abrasive grains are likely to adhere to the substrate surface, it is also usual to use a cleaning solution containing hydrofluoric acid or silicic hydrofluoric acid having an action of etching the substrate surface during or after a plurality of polishing steps. Has been done.

  In the grinding process or the polishing process, a work-affected layer is easily formed on the surface layer of the substrate, and there is a possibility that OH groups and hydrogen are taken into the work-affected layer. At this time, OH groups and hydrogen may be further taken into the substrate from the work-affected layer. There is a possibility that OH groups and hydrogen are taken in even when the substrate surface is finely etched during cleaning such as during the polishing process. Further, a hydration layer may be formed on the surface of the substrate.

  The present invention has been made in consideration of the above. That is, the manufacturing method of the mask blank of the present invention includes a step of preparing a glass substrate that is mirror-polished on the main surface, a step of performing a heat treatment on the prepared glass substrate, and a step after the heat treatment. Forming a thin film made of a material containing tantalum on the main surface of the glass substrate and containing substantially no hydrogen.

  Moreover, the manufacturing method of the mask blank of this invention is the process of preparing the glass substrate which mirror-polished with respect to the main surface, the process of heat-processing with respect to the prepared said glass substrate, and after heat processing A step of installing a glass substrate in a film formation chamber, using a tantalum-containing target, introducing a hydrogen-free sputtering gas into the film formation chamber, and forming a thin film on the main surface of the glass substrate by a sputtering method. It is characterized by that.

  After the mirror polishing is performed in the polishing step, the glass substrate in a stage satisfying the conditions as the mask blank substrate is subjected to heat treatment, so that OH groups and hydrogen incorporated in the surface layer or inside of the substrate And water can be forced out. And by forming a thin film containing tantalum on the glass substrate after the heat treatment, it is possible to suppress the incorporation of hydrogen into the thin film containing tantalum, and to increase the compressive stress of the thin film. Can be suppressed.

  The step of preparing a glass substrate that has been mirror-polished on the main surface is a square inner region with a side of 142 mm on the basis of the center of the main surface of the glass substrate after this step (hereinafter referred to as a 142 mm square region). The surface roughness in the square inner region having a side of 1 μm on one side is preferably Rnm (hereinafter simply referred to as “surface roughness Rq”) of 0.2 nm or less. The flatness in the 132 mm square area of the main surface of the glass substrate is more preferably 0.3 μm or less. A glass substrate in a state of being cut out from a glass ingot cannot satisfy such high flatness and surface roughness conditions. In order to satisfy the conditions of high flatness and surface roughness, it is essential to perform mirror polishing at least on the main surface of the glass substrate. In this mirror polishing, it is preferable that both main surfaces of the glass substrate are simultaneously polished by a double-side polishing with a polishing liquid containing colloidal silica abrasive grains. Further, it is desirable to finish the main surface satisfying the required flatness and surface roughness conditions by performing a plurality of stages of grinding and polishing on the cut state glass substrate. In this case, at least in the final stage of the polishing process, a polishing liquid containing colloidal silica abrasive grains is used.

  It is preferable that the heat processing performed with respect to the prepared glass substrate is a process which heats a glass substrate to 300 degreeC or more. In the heat treatment at less than 300 ° C., since the temperature is insufficient, the effect of discharging hydrogen in the substrate out of the substrate cannot be sufficiently obtained. The heat treatment is more effective when the temperature is 400 ° C. or higher. Further, when the temperature is 500 ° C. or higher, a sufficient effect of removing hydrogen from the substrate can be obtained even if the heating time is shortened. On the other hand, the heat treatment needs to be less than 1600 ° C. This is because synthetic quartz generally has a softening point of 1600 ° C., and if the heating temperature is 1600 ° C. or higher, the substrate softens and deforms. The heat treatment is preferably 1000 ° C. or lower, and more preferably 800 ° C. or lower. Moreover, although the processing time of the heat processing performed with respect to a glass substrate is based also on heating temperature, it is desirable that it is at least 10 minutes or more. Moreover, it is preferable that the processing time of the heat processing performed with respect to a glass substrate is 30 minutes or more, and it is more preferable that it is 40 minutes or more.

  The heat treatment is preferably performed in a state where a gas from which hydrogen is excluded as much as possible exists around the glass substrate. Although the amount of hydrogen itself is small in the air, a large amount of water vapor is present. Although the humidity in the air in the clean room is controlled, a relatively large amount of water vapor is present. By performing the heat treatment on the glass substrate in dry air, hydrogen can be prevented from entering the glass substrate due to water vapor. Furthermore, it is more preferable to heat-treat the glass substrate in a gas that does not contain hydrogen or water vapor (an inert gas such as nitrogen or a rare gas).

  It is desirable to perform a predetermined cleaning process on the glass substrate to be heat-treated. If the heat treatment is performed in a state where the abrasive grains used in the polishing step are attached, the heat treatment may stick to the surface of the substrate and may not be removed even if a normal cleaning step is performed after the heat treatment. In particular, when the material is similar to a glass substrate such as colloidal silica, it may adhere firmly to the surface of the glass substrate, so a cleaning solution containing hydrofluoric acid or silicic acid having an action of etching the substrate surface. It is desirable to wash with.

Examples of the material of the glass substrate in the mask blank include synthetic quartz glass, quartz glass, aluminosilicate glass, soda lime glass, low thermal expansion glass (SiO ₂ —TiO ₂ glass, etc.) and the like. In particular, since synthetic quartz glass has high transmittance with respect to ArF excimer laser light (wavelength 193 nm), it is preferable to use synthetic quartz glass as a material for the glass substrate. The exposure light applied to the mask blank and transfer mask of the present invention is not particularly limited, such as ArF excimer laser light, KrF excimer laser light, and i-line light. Mask blanks and transfer masks that apply ArF excimer laser to exposure light have very high required levels such as the flatness of the main surface and the positional accuracy of a transfer pattern formed of a thin film. Therefore, it is particularly effective to use the manufacturing method of the present invention for manufacturing a mask blank and a transfer mask for applying ArF excimer laser to exposure light.

  The heat treatment for the glass substrate after mirror polishing is particularly effective when forming a thin film made of a material containing tantalum. This is because the material containing tantalum has a property of easily taking in hydrogen. Since tantalum has a characteristic of becoming brittle when hydrogen is taken in, it is desired to suppress the hydrogen content even immediately after the thin film is formed. For this reason, in this invention, the material which contains a tantalum and does not contain hydrogen substantially is selected for the thin film formed on the main surface of the glass substrate of heat processing. “Substantially no hydrogen” means that the hydrogen content in the thin film is at least 5 at% or less. The preferable range of the hydrogen content in the thin film is preferably 3 at% or less, and more preferably the detection lower limit value or less. For the same reason, in the present invention, in the step of forming a thin film, a heat-treated glass substrate is placed in a film formation chamber, a tantalum-containing target is used, and a hydrogen-free sputtering gas is added to the film formation chamber. And a thin film can be formed by sputtering on the main surface of the glass substrate.

  The material containing tantalum that forms a thin film provided on the glass substrate and does not substantially contain hydrogen is, for example, tantalum metal and one or more selected from tantalum, nitrogen, oxygen, boron, and carbon. Examples include materials that contain elements and do not substantially contain hydrogen. For example, examples of the material containing tantalum and substantially not containing hydrogen include Ta, TaN, TaON, TaBN, TaBON, TaCN, TaCON, TaBCN, and TaBOCN. About the said material, you may contain metals other than a tantalum in the range with which the effect of this invention is acquired.

  In addition to tantalum, there is a metal having a property of easily taking in hydrogen. However, when tantalum in the material of the thin film is replaced with another metal having a property of easily taking in hydrogen, the same effect as the present invention can be obtained. It is done. Examples of other metals having the property of easily incorporating hydrogen include niobium, vanadium, titanium, magnesium, lanthanum, zirconium, scandium, yttrium, lithium, and praseodymium. The same effect can be obtained for an alloy composed of tantalum and two or more metals selected from the metal group having the property of easily taking in hydrogen. The heat treatment for the glass substrate is also effective when a thin film made of a material containing these metals or alloys is formed on the glass substrate.

  When the thin film made of a material containing tantalum has an oxygen content of 50 at% or more, an effect of preventing hydrogen from being taken into the material can be obtained to some extent. For this reason, when the material of the thin film on the main surface side of the substrate is a material containing tantalum having an oxygen content of less than 50 at%, an effect obtained by heat-treating the substrate before forming the thin film can be further obtained. Therefore, it is desirable.

  In addition, when the thin film made of a material containing tantalum is formed in contact with the main surface of the glass substrate after the heat treatment, hydrogen or the like present in the glass substrate directly enters the thin film. . In the case of such a configuration, the effect obtained by eliminating OH groups, hydrogen, water, and the like from the glass substrate by heat treatment can be obtained more significantly.

  The thin film of the mask blank is preferably formed of a material containing tantalum and nitrogen and substantially not containing hydrogen. Tantalum is a material that easily oxidizes naturally. As tantalum progresses, the light shielding performance (optical density) against exposure light decreases. In addition, from the viewpoint of forming a thin film pattern, a material in which oxidation of tantalum has not progressed is an etching gas containing fluorine (fluorine-based etching gas) and an etching gas containing chlorine and not containing oxygen (oxygen-free) It can be said that dry etching is possible with any of the above chlorine-based etching gases. However, tantalum that has undergone oxidation is a material that is difficult to dry etch with a chlorine-based etching gas that does not contain oxygen from the viewpoint of forming a thin film pattern, and it can be said that only a fluorine-based etching gas can be dry-etched. By including nitrogen in tantalum, oxidation of tantalum can be suppressed. Moreover, when the thin film which consists of a material containing a tantalum is formed in contact with the main surface of a glass substrate, compared with the case where oxygen is included, reducing the back surface reflectance with respect to exposure light by containing nitrogen. This is preferable because a decrease in optical density can be suppressed. The nitrogen content in the thin film is preferably 30 at% or less, more preferably 25 at% or less, and further preferably 20 at% or less from the viewpoint of optical density. In addition, the nitrogen content in the thin film is desirably 7 at% or more when the back surface reflectance needs to be less than 40%.

  The thin film of the mask blank preferably has a highly oxidized layer containing 60 atomic% or more of oxygen on the surface layer (the surface layer of the thin film opposite to the main surface of the substrate). As described above, hydrogen enters not only from the substrate, but also hydrogen in the gas surrounding the mask blank from the surface of the thin film. A high oxide film of a thin film material has a high binding energy and a property of preventing hydrogen from entering the thin film. Moreover, the high oxidation layer (tantalum high oxidation layer) of the material containing tantalum has excellent chemical resistance, warm water resistance, and light resistance against ArF exposure light.

The thin film in the mask blank or transfer mask is desired to have a crystal structure of microcrystals, preferably amorphous. For this reason, the crystal structure in the thin film is unlikely to be a single structure, and a plurality of crystal structures are likely to be mixed. That is, in the case of a high tantalum oxide layer, TaO bonds, Ta ₂ O ₃ bonds, TaO ₂ bonds, and Ta ₂ O ₅ bonds tend to be mixed. As the abundance ratio of Ta ₂ O ₅ bonds increases in a predetermined surface layer in the light-shielding film, the properties of preventing hydrogen intrusion, chemical resistance, hot water resistance and ArF light resistance are all increased, and the abundance ratio of TaO bonds. These properties tend to decrease with increasing.

In the high tantalum oxide layer, when the oxygen content in the layer is 60 at% or more and less than 66.7 at%, the bonding state between tantalum and oxygen in the layer tends to be mainly composed of Ta ₂ O ₃ bonds. It is done. In the case of the oxygen content in this layer, TaO bonds, which are the most unstable bonds, are considered to be much less than when the oxygen content in the layer is less than 60 at%. In the high tantalum oxide layer, when the oxygen content in the layer is 66.7 at% or more, it is considered that the bonding state between tantalum and oxygen in the layer tends to be mainly composed of TaO ₂ bonds. In the case of the oxygen content in this layer, it is considered that both the TaO bond which is the most unstable bond and the Ta ₂ O ₃ bond which is the next unstable bond are very few.

When the oxygen content in the tantalum highly oxidized layer is 68 at% or more, it is considered that not only TaO ₂ bonds are the main component but also the ratio of Ta ₂ O ₅ bonding state is increased. At such an oxygen content, the “Ta ₂ O ₃ ” and “TaO ₂ ” bonding states rarely exist, and the “TaO” bonding state cannot exist. When the oxygen content in the high tantalum oxide layer is 71.4 at%, it is considered that the tantalum high oxide layer is formed substantially only in a bonded state of Ta ₂ O ₅ . When the oxygen content in the high tantalum oxide layer is 60 at% or more, not only the most stable bonding state “Ta ₂ O ₅ ” but also bonding states of “Ta ₂ O ₃ ” and “TaO ₂ ” Will also be included. In addition, when the oxygen content in the layer is 60 at% or more, TaO bond, which is at least the most unstable bond, has the effect of preventing hydrogen entry, chemical resistance, and ArF light resistance. The amount is very small so as not to give. Therefore, it is considered that the lower limit value of the oxygen content in the layer is 60 at%.

The abundance ratio of Ta ₂ O ₅ bonds in the high tantalum oxide layer is preferably higher than the abundance ratio of Ta ₂ O ₅ bonds in the thin film excluding the high oxidation layer. The Ta ₂ O ₅ bond is a bonded state having very high stability, and by increasing the abundance ratio of the Ta ₂ O ₅ bond in the high oxide layer, the characteristics of preventing hydrogen intrusion, chemical resistance, Mask cleaning resistance such as warm water and ArF light resistance are greatly increased. In particular, it is most preferable that the high tantalum oxide layer is formed only by a bonded state of Ta ₂ O ₅ . The content of nitrogen and other elements in the high tantalum oxide layer is preferably in a range that does not affect the operational effects such as the property of preventing hydrogen intrusion, and is preferably not substantially contained.

  The thickness of the high tantalum oxide layer is preferably 1.5 nm or more and 4 nm or less. If it is less than 1.5 nm, it is too thin to prevent the effect of blocking hydrogen penetration, and if it exceeds 4 nm, the influence on the surface reflectance increases, and a predetermined surface reflectance (reflectance for exposure light or light of each wavelength) Control for obtaining a reflectance spectrum becomes difficult. Moreover, since the optical density with respect to ArF exposure light is extremely low, the optical density that can be secured by the surface antireflection layer is lowered, and the high tantalum oxide layer works negatively from the viewpoint of reducing the thickness of the thin film. . In consideration of the balance between securing the optical density of the entire thin film and the characteristics of preventing hydrogen intrusion, chemical resistance and ArF light resistance, the thickness of the highly oxidized layer is 1.5 nm or more and 3 nm. The following is more desirable.

The method for forming the high tantalum oxide layer includes a hot water treatment, an ozone-containing water treatment, a heat treatment in a gas containing oxygen, and a gas containing oxygen on the mask blank after the thin film is formed. Performing ultraviolet irradiation treatment and / or O ₂ plasma treatment, and the like. The high oxide layer is not limited to the metal high oxide film forming the thin film. Any metal high oxide film may be used as long as it has the property of preventing hydrogen intrusion, and a structure in which the high oxide film is laminated on the surface of the thin film may be used. Further, as long as the material has a property of preventing hydrogen from entering the thin film, the material may not be a high oxide, and the material film may be stacked on the surface of the thin film.

  The thin film of the mask blank has a structure in which a lower layer and an upper layer are laminated from the glass substrate side, the lower layer is made of a material containing tantalum and nitrogen and substantially not containing hydrogen, and the upper layer is Preferably, it is made of a material containing tantalum and oxygen. With such a configuration, the upper layer can function as a film (antireflection film) having a function of controlling the surface reflectance with respect to the exposure light of the thin film.

  Furthermore, it is desirable to form a highly oxidized layer containing 60 at% or more of oxygen on the upper surface layer (surface layer opposite to the lower layer side). The mode and effect of the high oxide layer and the high oxide layer of the material containing tantalum are the same as described above. In consideration of the ease of control of the surface reflectance characteristics (reflectance with respect to ArF exposure light and reflectance spectrum with respect to light of each wavelength), the upper layer has an oxygen content of less than 60 at% excluding the surface layer portion. Preferably there is.

  The upper layer is formed of a material containing oxygen in tantalum. From the viewpoint of forming a thin film pattern, dry etching of the upper layer (a material containing oxygen in tantalum) is difficult with a chlorine-based etching gas not containing oxygen, and can be dry-etched only with a fluorine-based etching gas. On the other hand, the lower layer is formed of a material containing tantalum and nitrogen. From the viewpoint of forming a thin film pattern, dry etching of a lower layer (a material containing tantalum and nitrogen) can be performed using either a fluorine-based etching gas or an oxygen-free chlorine-based etching gas. Therefore, when a pattern is formed by dry etching a thin film using a resist pattern (resist film on which a transfer pattern is formed) as a mask, the pattern is formed by performing dry etching with a fluorine-based etching gas on the upper layer. Using this pattern as a mask, a process of forming a pattern by performing dry etching with a chlorine-based etching gas containing no oxygen in the lower layer can be used. By applying such an etching process, the resist film can be thinned.

  In order to make it difficult to dry-etch the upper layer with an oxygen-free chlorine-based etching gas, it is preferable to increase the amount of bonds between tantalum and oxygen having a relatively high binding energy in the upper layer. In order to make dry etching of the upper layer more difficult, the ratio of the oxygen content to the tantalum content of the upper layer is preferably 1 or more. When the upper layer is formed only of tantalum and oxygen, the oxygen content in the upper layer is preferably 50 at% or more.

  About the material which forms the lower layer of the said thin film, it is the same as that of what was enumerated in the material which contains the said tantalum and does not contain hydrogen substantially. The material for forming the upper layer is preferably a material containing tantalum and oxygen, and further containing nitrogen, boron, carbon and the like. Examples of the material for forming the upper layer include TaO, TaON, TaBO, TaBON, TaCO, TaCON, TaBCO, and TaBOCN.

  The thin film of the mask blank is not limited to the above laminated structure. A three-layer structure may be used, a single-layer composition gradient film may be used, or a film structure having a composition gradient between the upper layer and the lower layer may be employed. The thin film of the mask blank is desirably used as a light-shielding film that functions as a light-shielding pattern when a transfer mask is produced, but is not limited thereto. The mask blank thin film can also be used as an etching stopper film or an etching mask film (hard mask film). If it is within the constraints required for the thin film, a halftone phase shift film or an optical half film can be used. It can also be applied to a permeable membrane. In addition, since the mask blank can suppress the time-dependent change of the compressive stress of the thin film, a double patterning technique (a double patterning technique (DP technique) in a narrow sense, a double patterning technique) that requires high positional accuracy for a pattern formed by the thin film. This is particularly suitable when a transfer mask set to which an exposure technique (DE technique, etc.) is applied is produced.

  In the mask blank thin film, a material having etching selectivity between the glass substrate and the thin film (a material containing Cr, such as Cr, CrN, CrC, CrO, CrON and CrC) between the glass substrate and the thin film. An etching stopper film or an etching mask film may be formed. Further, a halftone phase shift film having a predetermined phase shift effect and transmittance with respect to exposure light, or a light semi-transmissive film having only a predetermined transmittance may be formed between the glass substrate and the thin film. (In this case, the thin film is used as a light shielding film for forming a light shielding band, a light shielding patch, or the like). However, these films in contact with the thin film are basically made of a material that does not contain hydrogen (does not contain positively, and the process performed until the thin film is formed is a means that prevents hydrogen from being taken in as much as possible. )).

  The transfer mask produced by the production method of the present invention is preferably produced by forming a transfer pattern on the thin film of the mask blank using the mask blank produced by the production method. Since the mask blank that has been produced for a certain period of time or more after being manufactured is suppressed from increasing in compressive stress of the thin film over time, the flatness of the mask blank is maintained at the required high level. If a mask blank having such characteristics is used, the completed transfer mask can have the required high flatness. Further, since the compressive stress of the thin film is suppressed, it is also possible to suppress the amount of displacement on the main surface caused by each pattern of the thin film released from the surrounding compressive stress after the etching process for producing the transfer mask.

  On the other hand, when a transfer mask is produced using a mask blank that has not been manufactured for a long time, a transfer mask (conventional transfer mask) immediately after being produced by a conventional production method is required to be high. It has flatness. However, when a conventional transfer mask is stored and stored in a mask case without being used, or when it is continuously used after being set in an exposure apparatus, the compressive stress of the thin film increases. As a result, the pattern of the thin film may be largely displaced. If the transfer mask of the present invention produced using the mask blank produced by the production method of the present invention is used, an increase in the compressive stress of the thin film over time can be suppressed, so that the mask case can be used without being used after production. Even when stored and stored or when it is set and continuously used in an exposure apparatus, the required high flatness can be maintained, and the positional deviation of each pattern of the thin film can be suppressed.

  Further, it is desirable that a highly oxidized layer containing 60 atomic% or more of oxygen is formed on the surface layer of the transfer pattern formed by the thin film of the transfer mask. The mode and effect of the high oxide layer and the high oxide layer of the material containing tantalum are the same as described above.

  The transfer mask can be used as a binary mask, and is particularly suitable when ArF excimer laser light is applied to exposure light. The present invention is also applicable to a digging Levenson type phase shift mask, a halftone phase shift mask, an enhancer type phase shift mask, a chromeless phase shift mask (CPL mask), and the like. The transfer mask is excellent in pattern position accuracy, and is particularly suitable for a transfer mask set to which a double patterning technique (DP technique, DE technique, etc.) is applied.

As etching performed on the mask blank when the transfer mask is manufactured, dry etching effective for forming a fine pattern is preferably used. For dry etching of the thin film with the etching gas containing fluorine, fluorine-based gas such as SF ₆ , CF ₄ , C ₂ F _6, and CHF ₃ can be used. For dry etching of the thin film with the etching gas containing chlorine and not containing oxygen, a chlorine-based gas such as Cl ₂ and CH ₂ Cl ₂ , or at least one of these chlorine-based gases, A mixed gas with He, H ₂ , N ₂ , Ar and / or C ₂ H ₄ or the like can be used.

  A semiconductor device having a highly accurate pattern can be manufactured by exposing and transferring a transfer pattern to a resist film on a semiconductor substrate using the transfer mask manufactured by the above manufacturing method. This is because the transfer mask has high flatness and pattern position accuracy required during production. In addition, the transfer mask is not used after production, but is stored in a mask case and stored for a certain period, and then set in an exposure apparatus and used for exposure transfer. This is because, even when it is set in an exposure apparatus and used for exposure transfer, the required high flatness can be maintained and positional deviation of each pattern of the thin film can be suppressed.

  As shown in FIG. 1, in the mask blank according to the present embodiment, a lower layer (light-shielding layer) 2 mainly composed of tantalum and nitrogen having a thickness of 42.5 nm is formed on a glass substrate 1 made of synthetic quartz. An upper layer (antireflection layer) 3 mainly composed of tantalum and oxygen having a thickness of 5.5 nm is formed on the lower layer 2, and a tantalum highly oxidized layer 4 is formed on the surface layer of the upper layer 3. It will be. The light shielding film 30 is composed of the lower layer 2 and the upper layer 3 including the tantalum high oxide layer 4.

  Further, as shown in FIG. 2, the transfer mask according to this embodiment causes the light shielding film 30 to remain on the glass substrate 1 by patterning the light shielding film 30 of the mask blank shown in FIG. The light-shielding part 30a and the removed light-transmitting part 30b form a fine pattern. A high tantalum oxide layer 4a is formed on the surface layer of the light shielding film pattern (thin film pattern) 30a. Further, on the side wall of the light shielding film pattern 30a, a tantalum high oxide layer 4b is formed on the surface layer of the pattern 3a of the upper layer 3, and a tantalum high oxide layer 4c is formed on the surface layer of the pattern 2a of the lower layer 2. . The method for forming the high tantalum oxide layers 4b and 4c on the side walls of the patterns 2a and 3a of the lower layer 2 and the upper layer 3 is the same as the method for forming the high tantalum oxide layer in the mask blank.

Next, an example of manufacturing a mask blank and a transfer mask according to the present embodiment will be described as an example with reference to FIG.
Example 1
[Manufacture of mask blanks]
A synthetic quartz ingot manufactured by a conventionally used manufacturing method (soot method) was subjected to heat treatment at 1000 ° C. or more for 2 hours or more. Next, a glass substrate made of synthetic quartz having a length and width of about 152 mm × 152 mm and a thickness of 6.85 mm was cut out from the synthetic quartz ingot after the heat treatment, and a chamfered surface was appropriately formed.

  The main surface and the end face were ground on the cut glass substrate. A rough polishing process and a precision polishing process were sequentially performed on the main surface of the ground glass substrate. Specifically, a glass substrate supported by a carrier is sandwiched between upper and lower surface plates of a double-side polishing apparatus having polishing pads (hard polishers) on the upper and lower surface plates, and cerium oxide abrasive grains (grain size coarse polishing step) : The glass substrate was moved in a planetary gear while a polishing liquid containing 2 to 3 μm and a precision polishing step: 1 μm was introduced into the surface of the platen. The glass substrate after each process was immersed in a cleaning tank (applied with ultrasonic waves) and cleaned to remove the attached abrasive grains.

  An ultraprecision polishing step and a final polishing step were performed on the main surface of the glass substrate after the cleaning. Specifically, a glass substrate supported by a carrier is sandwiched between upper and lower surface plates of a double-side polishing apparatus having a polishing pad (ultra-soft polisher) on each upper and lower surface plate, and colloidal silica abrasive grains (particle size ultra-precision) The glass substrate was moved in a planetary gear while a polishing liquid containing a polishing step: 30 to 200 nm and a final polishing step: 80 nm on average was allowed to flow into the surface plate surface. The glass substrate after the ultra-precision polishing step was immersed in a cleaning tank containing hydrofluoric acid and silicic hydrofluoric acid (ultrasonic application) to perform cleaning to remove the abrasive grains.

  The glass substrate after the final polishing step has a thickness of 6.35 mm and a flatness of a 142 mm square inner region (hereinafter referred to as a 142 mm square inner region) with reference to the center of the main surface is 0.3 μm or less. The roughness Rq was 0.2 nm or less, which was a sufficient level as a glass substrate used in a 22 nm node mask blank.

Next, this glass substrate was placed in a heating furnace, and the gas in the furnace was changed to the same gas as outside the furnace (air in the clean room), and a heat treatment was performed at a heating temperature of 500 ° C. for 40 minutes. Furthermore, the glass substrate after the heat treatment is rinsed with a detergent and rinsed with pure water, and further irradiated with a Xe excimer lamp in the atmosphere. The main surface is irradiated with ultraviolet rays and O ₃ generated by the ultraviolet rays. Was washed.

Next, the cleaned glass substrate was introduced into a DC magnetron sputtering apparatus. A mixed gas of Xe and N ₂ was introduced into the sputtering apparatus, and a TaN layer (lower layer) 2 having a thickness of 42.5 nm was formed in contact with the main surface of the glass substrate 1 by a sputtering method using a tantalum target. (See FIG. 3 (a)). Further, the gas in the sputtering apparatus was replaced with a mixed gas of Ar and O _2, and a TaO layer (upper layer) 3 having a thickness of 5.5 nm was formed by sputtering using a tantalum target (FIG. 3A )reference). Immediately after the film formation, the flatness on the surface of the light shielding film 30 on the main surface of the substrate was measured with a flatness measuring device UltraFLAT 200M (manufactured by Corning Toropel) for the mask blank including the light shielding film 30. Moreover, as a result of performing HFS / RBS analysis with respect to the mask blank manufactured on the same conditions, the hydrogen content in a TaN film | membrane was below the detection lower limit.

  Next, the glass substrate on which the light-shielding film 30 having a laminated structure of the TaN layer (lower layer) 2 and the TaO layer (upper layer) 3 is formed is placed in a heating furnace, the gas in the furnace is replaced with nitrogen, and a heating temperature of 300 An accelerated test was performed in which heat treatment was performed at 1 ° C. for 1 hour to forcibly incorporate hydrogen in the substrate into the light-shielding film. With respect to the mask blank including the light shielding film 30 after the acceleration test, the flatness of the surface of the light shielding film 30 on the main surface of the substrate was measured with a flatness measuring apparatus UltraFLAT 200M (Corning TOROPEL). The difference between the flatness measured after the acceleration test and the flatness measured before the acceleration test was 30 nm (the surface shape changed in a convex direction after the acceleration test) in the 142 mm square area. The numerical value of the flatness difference is within the measurement error range, and almost no change was observed in the flatness before and after the heat treatment. Moreover, the result of having performed the HFS / RBS analysis with respect to the mask blank provided with the light shielding film 30 after an acceleration test is shown in FIG. From the result of FIG. 4, it can be said that the TaN layer contains about 2 at% of hydrogen. From this result, it can be said that since the inside of the heating furnace is a nitrogen atmosphere, the supply source of hydrogen taken into the TaN layer is a glass substrate. In addition, it can be seen from the flatness difference before and after the acceleration test that the flatness of the main surface is hardly affected as long as about 2 at% of hydrogen is taken into the TaN layer.

Next, under the same conditions as described above, a mask blank was manufactured in which a light shielding film 30 having a laminated structure of TaN layer 2 and TaO layer 3 was formed in contact with the main surface of glass substrate 1. This mask blank was placed on a hot plate and heat-treated at 300 ° C. in the atmosphere to form a highly oxidized tantalum layer 4 on the surface layer in the TaO layer 3 (see FIG. 3B). HFS / RBS analysis was performed on the mask blank after forming the high oxide layer 4, and it was confirmed that the high oxide layer 4 having a high oxygen content was formed from the surface of the TaO layer 3 to a depth of 2 nm. It was done. Further, when XPS analysis (X-ray photoelectron spectroscopy) was performed, a high peak was observed at the position of the binding energy (25.4 eV) of Ta ₂ O ₅ in the narrow spectrum of the outermost layer of the light shielding film 30. Further, in the narrow spectrum in the layer having a depth of 1 nm from the surface of the light-shielding film 30, the peak at the position of the binding energy of Ta ₂ O ₅ (25.4 eV) and the binding energy of Ta (21.0 eV) A peak was observed from Ta ₂ O ₅ . From these results, it can be said that the highly oxidized layer 4 having Ta ₂ O ₅ bonds is formed on the surface layer of the TaO layer 3. As described above, the mask blank of Example 1 including the light-shielding film 30 having the stacked structure of the TaN layer 2 and the TaO layer 3 including the tantalum high oxide layer 4 on the surface layer is formed on the main surface of the glass substrate 1. Obtained.

  The reflectance (surface reflectance) at the film surface of the light shielding film 30 manufactured as described above was 30.5% in ArF exposure light (wavelength 193 nm). The reflectance (back surface reflectance) of the surface of the glass substrate 1 on which the light-shielding film was not formed was 38.8% in ArF exposure light. The optical density in ArF exposure light was 3.02.

[Preparation of transfer mask]
Next, using the obtained mask blank, a transfer mask of Example 1 was produced by the following procedure.

  First, a 100 nm-thickness chemically amplified resist 5 for electron beam drawing was applied by spin coating (see FIG. 3C), and electron beam drawing and development were performed to form a resist pattern 5a (FIG. 3 ( d)). In addition, the pattern which performed the electron beam drawing used one of what divided | segmented the fine pattern of 22 nm node into two comparatively sparse transfer patterns using the double patterning technique.

Next, dry etching using a fluorine-based (CF ₄ ) gas was performed to produce a pattern 3a of the TaO layer (upper layer) 3 including the highly oxidized layer 4 (see FIG. 3E). Subsequently, dry etching using a chlorine-based (Cl ₂ ) gas was performed to produce a TaN layer (lower layer) 2 pattern 2a, and a light-shielding film pattern 30a was produced on the substrate 1 (see FIG. 3F). Subsequently, the resist on the light shielding film pattern 30a was removed to obtain a light shielding film pattern 30a having a function as a transfer mask (see FIG. 3G). Thus, a transfer mask (binary mask) was obtained.

  Next, with respect to the produced transfer mask, before natural oxidation proceeds (for example, within 1 hour after film formation) or after storage in an environment in which natural oxidation does not proceed after film formation, It was immersed in DI water for 120 minutes, and warm water treatment (surface treatment) was performed. Thereby, the transfer mask of Example 1 was obtained.

  In the transfer mask of this Example 1, formation of highly oxidized tantalum layers 4a, 4b and 4c as schematically shown in FIG. 2 was confirmed on the surface layer of the light shielding film pattern 30a. Specifically, high-oxide layers 4a, 4b, and 4c having a thickness of 3 nm were confirmed by cross-sectional observation using a scanning transmission electron microscope (STEM). Further, HFS / RBS analysis was performed on a portion of the light shielding film pattern 30a where the light shielding film is present. According to the analysis result of the profile in the depth direction of the light shielding film, it was confirmed that the tantalum highly oxidized layer 4a as the surface layer of the TaO layer 3 has an oxygen content of 71.4 to 67 at%. On the other hand, for the pattern side wall portions (4b and 4c), it is difficult to confirm the oxygen content by HFS / RBS analysis. For this reason, EDX (energy dispersive X-ray spectroscopy) analysis is used in the observation with the STEM, and compared with the result of HFS / RBS analysis of the highly oxidized layer 4a on the surface layer of the light shielding film pattern 30a analyzed earlier, It was confirmed that the oxygen content was the same in the layer 4a and the highly oxidized layers 4b and 4c.

  An acceleration test was performed on the produced transfer mask of Example 1 under the same conditions as those for the mask blank (replacement of the gas in the furnace with nitrogen and heat treatment at 300 ° C. for 1 hour). It was. In addition, before and after the acceleration test, the pattern width and the space width at a predetermined portion in the surface of the transfer mask were measured. The variation widths of the pattern width and the space width before and after the acceleration test were all within an allowable range. In the same procedure, using the mask blank of the first embodiment, transfer having the other transfer pattern of the 22 nm node fine pattern divided into two relatively sparse transfer patterns using the double patterning technique A mask was prepared. Through the above procedure, a transfer mask set capable of transferring a fine pattern of 22 nm node to a transfer object was obtained by exposure transfer using a double patterning technique using two transfer masks.

[Manufacture of semiconductor devices]
Using the produced transfer mask set, an exposure apparatus using ArF excimer laser as exposure light was applied, and a double patterning technique was applied to expose and transfer a 22 nm node fine pattern onto the resist film on the semiconductor device. A predetermined development process was performed on the resist film on the semiconductor device after exposure to form a resist pattern, and the lower layer film was dry-etched using the resist pattern as a mask to form a circuit pattern. When the circuit pattern formed on the semiconductor device was confirmed, there was no wiring short circuit or disconnection of the circuit pattern due to insufficient overlay accuracy.

(Comparative Example 1)
[Manufacture of mask blanks]
A glass substrate after the final polishing step was prepared by the same procedure as in Example 1. The glass substrate is not subjected to the heat treatment as in Example 1, but is washed with a detergent and rinsed with pure water. Further, the glass substrate is irradiated with a Xe excimer lamp, and ultraviolet rays and ultraviolet rays thereof are irradiated. The main surface was cleaned with O ₃ generated by. A light shielding film 30 was formed on the cleaned glass substrate under the same film forming conditions as in Example 1. Immediately after the film formation, the flatness on the surface of the light shielding film 30 on the main surface of the substrate was measured with a flatness measuring device UltraFLAT 200M (Corning TOROPEL) for the mask blank including the light shielding film 30. Moreover, as a result of performing HFS / RBS analysis with respect to the mask blank manufactured on the same conditions, the hydrogen content in a TaN film | membrane was below the detection lower limit.

  Next, a light-shielding film 30 having a laminated structure of TaN layer (lower layer) 2 and TaO layer (upper layer) 3 is formed, the glass substrate is placed in a heating furnace, the gas in the furnace is replaced with nitrogen, and the heating temperature is 300 ° C. Then, an acceleration test was performed in which heat treatment for 1 hour was performed to forcibly incorporate hydrogen in the substrate into the light-shielding film. With respect to the mask blank including the light shielding film 30 after the acceleration test, the flatness of the surface of the light shielding film 30 on the main surface of the substrate was measured with a flatness measuring apparatus UltraFLAT 200M (Corning TOROPEL). The difference between the flatness measured after the acceleration test and the flatness measured before the acceleration test is 219 nm (the surface shape changes in a convex direction after the acceleration test) in the 142 mm square region, and the flatness is It changed a lot. This flatness difference was unacceptable with a mask blank for at least 22 nm node. Moreover, the result of having performed the HFS / RBS analysis with respect to the mask blank provided with the light shielding film 30 after an acceleration test is shown in FIG. From the results shown in FIG. 5, the TaN layer contains about 6.5 at% hydrogen. From this result, it can be said that since the inside of the heating furnace is a nitrogen atmosphere, the supply source of hydrogen taken into the TaN layer is a glass substrate. Further, from the difference in flatness before and after the acceleration test, it was found that when about 6.5 at% of hydrogen was taken into the TaN layer, the flatness of the main surface was greatly deteriorated.

  Next, under the same conditions as described above, a mask blank was manufactured in which a light shielding film 30 having a laminated structure of TaN layer 2 and TaO layer 3 was formed in contact with the main surface of glass substrate 1. The mask blank was placed on a hot plate and heat-treated at 300 ° C. in the atmosphere in the same manner as in Example 1 to form a highly oxidized tantalum layer 4 on the surface layer in the TaO layer 3 (FIG. 3B). )reference). As described above, the mask blank of Comparative Example 1 including the light-shielding film 30 having the laminated structure of the TaN layer 2 and the TaO layer 3 including the highly oxidized tantalum layer 4 on the surface layer is formed on the main surface of the glass substrate 1. Obtained.

[Preparation of transfer mask]
Next, using the obtained mask blank, a transfer mask of Comparative Example 1 was produced in the same procedure as in Example 1. An acceleration test was performed on the produced transfer mask of Comparative Example 1 under the same conditions as in the mask blank (replace the gas in the furnace with nitrogen and heat treatment at 300 ° C. for 1 hour). It was. In addition, before and after the acceleration test, the pattern width and the space width at a predetermined portion in the surface of the transfer mask were measured. The variation widths of the pattern width and the space width before and after the acceleration test were both large, and it was clearly outside the allowable range in the transfer mask to which the double patterning technology for at least 22 nm node was applied. For this reason, in the same procedure, a transfer mask having the other transfer pattern, which is obtained by dividing a fine pattern of 22 nm node into two relatively sparse transfer patterns using double patterning technology, is transferred. Even if a mask set for use is produced, the overlay accuracy is low and cannot be used as a transfer mask set for double patterning.

  Moreover, the mask for transfer of the comparative example 1 was produced in the procedure similar to Example 1 using the mask blank after performing the said acceleration test. As a result, since the flatness has already deteriorated greatly in the mask blank state, the movement on the main surface of the pattern is remarkable when chucked on the mask stage of the exposure apparatus, and the double patterning technology for at least 22 nm node is applied. It was clearly out of the acceptable range for the transfer mask. Further, since the compressive stress of the light shielding film was remarkably large, the pattern of the light shielding film after the dry etching was largely deviated from the electronic drawing pattern.

1 Glass substrate 2 Lower layer (TaN layer)
2a Lower layer pattern 3 Upper layer (TaO layer)
3a Upper layer pattern 4, 4a, 4b, 4c High tantalum oxide layer 5 Resist film 30 Light shielding film 30a Light shielding part 30b Light transmitting part

Claims (11)

A step of performing a heat treatment for heating the glass substrate to 300 ° C. or higher ;
A step of rinsing and cleaning the glass substrate after the heat treatment;
The main surface of the glass substrate after the rinse cleaning contains one or more metals selected from tantalum, niobium, vanadium, titanium, magnesium, lanthanum, zirconium, scandium, yttrium, lithium and praseodymium , and substantially contains hydrogen. Forming a thin film made of a material that does not contain;
A method for manufacturing a mask blank, comprising:   2. The method of manufacturing a mask blank according to claim 1, wherein the heat treatment is a treatment of heating the glass substrate in dry air or in a gas not containing hydrogen and water vapor. The thin film manufacturing method of the mask blank according to claim 1 or 2, characterized in that it is formed in contact with the main surface of the glass substrate. The thin film is tantalum containing nitrogen, and method for producing a mask blank according to any one of claims 1 to 3, characterized in that it consists substantially free material hydrogen. The surface layer of the thin film, the mask blank manufacturing method according to any one of claims 1 4, characterized in that highly oxidized layer is formed containing oxygen 60 atom% or more. The thin film has a structure in which a lower layer and an upper layer are laminated from the glass substrate side, the lower layer is made of a material containing tantalum and nitrogen and substantially free of hydrogen, and the upper layer is made of tantalum. The mask blank manufacturing method according to any one of claims 1 to 3 , wherein the mask blank is made of a material containing oxygen. The method for manufacturing a mask blank according to claim 6 , wherein a highly oxidized layer containing 60 atomic% or more of oxygen is formed on the upper surface layer.   The method for manufacturing a mask blank according to claim 1, further comprising a step of forming an etching stopper film made of a material containing chromium between the glass substrate and the thin film.   9. A method for producing a transfer mask, comprising: using a mask blank produced by the method for producing a mask blank according to claim 1; and forming a transfer pattern on a thin film of the mask blank.   10. The method for manufacturing a transfer mask according to claim 9, wherein a highly oxidized layer containing 60 atomic% or more of oxygen is formed on a surface layer of the transfer pattern formed of the thin film.   A method for manufacturing a semiconductor device, comprising: transferring and transferring a transfer pattern onto a resist film on a semiconductor substrate using the transfer mask manufactured by the method for manufacturing a transfer mask according to claim 9.