Glass plates, disc-shaped glass, and glass substrates for magnetic disks

Abstract

The present invention provides a glass substrate for magnetic disks that can suppress deterioration of flatness due to heat treatment for forming a magnetic recording layer, a disc-shaped glass used in such a glass substrate for magnetic disks, a glass plate, and a method for manufacturing a glass plate. [Solution] A rectangular glass plate with a thickness of less than 0.68 mm, wherein the flatness of a 100 mm square area to be measured, cut from the central region of the glass plate excluding the end regions on both sides in the short and long directions, is 30 μm or less; the thermal shrinkage rate of the area to be measured after a first heat treatment in which the area to be measured is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour is 130 ppm or less; and when the glass transition temperature of the glass plate is expressed as Tg (°C), the change in the flatness of the area to be measured after a second heat treatment in which the area to be measured is maintained at Tg-160°C for 60 seconds and then cooled to room temperature in the air is 10 μm or less.

Description

Technical Field

[0001] The present invention relates to a glass substrate for a magnetic disk used in a hard disk drive device, a disk-shaped glass, a glass plate, and a method for manufacturing a glass plate.

Background Art

[0002] With the recent prosperity of cloud computing, many hard disk drive (HDD) devices are used in cloud-oriented data centers to increase the storage capacity. As a storage medium in an HDD device, a magnetic disk in which a magnetic layer is provided on an annular non-magnetic glass substrate for a magnetic disk is used. In order to increase the storage capacity of an HDD device, in addition to increasing the recording density of the magnetic disk, it is preferable to increase the number of mounted magnetic disks by mounting a large number of thin magnetic disks.

[0003] For increasing the recording density, as a recording method on a magnetic disk, in addition to the conventional perpendicular magnetic recording method, a heat-assisted magnetic recording method (HAMR) and a microwave-assisted magnetic recording method (MAMR) have been studied. In recent years, in order to form a magnetic recording layer suitable for these recording methods, heat treatment of a magnetic film has been performed. The heat treatment is performed, for example, by heat-treating the glass substrate after forming the magnetic film at a high temperature, or by forming the magnetic film while heating the glass substrate at a high temperature. At this time, the temperature of the magnetic film may exceed 600 ° C by far and reach 700 ° C or higher. In this heat treatment, since the glass substrate is also heated together with the magnetic film, the glass substrate for a magnetic disk is required to have high heat resistance, that is, a high glass transition temperature (Tg), so as not to be thermally deformed.

[0004] Methods for manufacturing glass plates that serve as the raw material for glass substrates for magnetic disks include the press method, float method, Fulcol method, Pittsburgh method, downdraw method, Colburn method, and redraw method. Of these, the float method, Fulcol method, Pittsburgh method, downdraw method, Colburn method, and redraw method are suitable for manufacturing glass substrates for flat panel displays (FPDs) such as liquid crystal displays because they make it easier to produce larger glass plates compared to the press method.

[0005] Flat panel displays (FPDs) use glass substrates on which electronic elements such as thin-film transistors (TFTs) are mounted. In the TFT manufacturing process, the glass substrate is heated to high temperatures, which causes thermal shrinkage and makes it prone to dimensional changes. Therefore, when manufacturing glass plates using the float method or other methods described above, the glass plate is slowly cooled while being formed, and the residual stress in the glass plate is reduced and the thermal shrinkage rate is reduced by adjusting the slow cooling conditions. Furthermore, as a method to further reduce the thermal shrinkage rate of glass plates, offline annealing is known, in which glass plates cut to a predetermined size from a molded long glass sheet are subjected to heat treatment (Patent Document 1). [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2017-178711 [Overview of the project] [Problems that the invention aims to solve]

[0007] It has been found that when a glass substrate for magnetic disks is fabricated from a glass plate with a high glass transition temperature (Tg) manufactured by the float method or other methods described above, and a magnetic disk is fabricated using this substrate, the glass substrate deforms by shrinking and bending due to the heat treatment of the magnetic film, resulting in a deterioration of the flatness of the magnetic disk. Furthermore, it has become clear that this deterioration of flatness is particularly pronounced when using a thin glass substrate for magnetic disks. Poor flatness of the magnetic disk makes fluttering more likely to occur in HDD devices, making stable reading impossible.

[0008] Therefore, the present invention aims to provide a glass substrate for magnetic disks that can suppress deterioration of flatness due to heat treatment for forming the magnetic recording layer of a magnetic disk, and a disc-shaped glass, a glass plate, and a method for manufacturing a glass plate used in such a glass substrate for magnetic disks. [Means for solving the problem]

[0009] One aspect of the present invention is a glass plate. The glass plate is a rectangular glass plate with a thickness of less than 0.68 mm, The flatness of a 100mm square area to be measured, cut from the central region of the glass plate excluding the end regions on the inside of the glass plate that are 5-20% of the length of the short side of the glass plate from each end on the short side of the glass plate, and the end regions on the inside of the glass plate that are 5-20% of the length of the long side of the glass plate from each end on the long side of the glass plate, is 30μm or less. The thermal shrinkage rate of the area to be measured is 130 ppm or less when the area to be measured is subjected to a first heat treatment in which the area to be measured is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour. When the glass transition temperature of the glass plate is expressed in Tg (°C), the change in the flatness of the region to be measured due to a second heat treatment, in which the region to be measured is maintained at Tg -160°C for 60 seconds and then cooled to room temperature in air, is 10 μm or less.

[0010] It is preferable that the shorter side length exceeds 900 mm.

[0011] Preferably, the glass plate is a portion cut from a long glass sheet formed using one of the following methods: float method, full col method, Pittsburgh method, downdraw method, Colburn method, and redraw method.

[0012] Preferably, the difference between the amount of thermal shrinkage S1 of the area to be measured in the direction where the thermal shrinkage rate is minimum and the amount of thermal shrinkage S2 of the area to be measured in the direction where the thermal shrinkage rate is maximum is 1.0 μm or less.

[0013] The aforementioned glass plate has been annealed to reduce its thermal shrinkage rate. The glass plate before the annealing treatment exhibits anisotropy in its thermal shrinkage rate, where the magnitude of the thermal shrinkage rate differs depending on the in-plane direction of the region of the glass plate corresponding to the region being measured. The difference between the amount of thermal shrinkage S1 of the region in the direction where the thermal shrinkage rate is minimum and the amount of thermal shrinkage S2 of the region in the direction where the thermal shrinkage rate is maximum may be greater than 1.0 μm.

[0014] Another aspect of the present invention is a glass plate. The glass plate is a rectangular glass plate with a thickness of less than 0.68 mm, a flatness of 30 μm or less, and the lengths of two orthogonal sides are each 95 to 120 mm. The thermal shrinkage rate after the first heat treatment, in which the material is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour, is 130 ppm or less. The glass transition temperature of the glass plate is expressed as Tg(°C), and the change in flatness associated with the second heat treatment, which involves maintaining the glass plate at Tg-160°C for 60 seconds and then cooling it to room temperature in air, is 10 μm or less.

[0015] The aforementioned rectangular glass plate is a base plate that forms the basis of a disc-shaped glass having a circular outer circumference. It is preferable that the area of the main surface of the rectangular glass plate is 1.6 times or less the area inside the outer periphery of the disc-shaped glass.

[0016] Another aspect of the present invention is a disc-shaped glass. The disc-shaped glass is a disc-shaped glass having a plate thickness of less than 0.68 mm, a flatness of 30 μm or less, and a circular outer periphery with a diameter of 95 to 100 mm, when a first heat treatment is performed in which it is maintained at 700°C for 4 hours and then cooled at a rate of 50°C / hour from 700°C to 400°C, the thermal shrinkage rate is 130 ppm or less, when the glass transition temperature of the disc-shaped glass is represented by Tg (°C), after maintaining at Tg - 160°C for 60 seconds and then cooling to room temperature in the atmosphere, the change amount of flatness accompanying the second heat treatment is 10 μm or less, which is characterized.

[0017] Another aspect of the present invention is a glass substrate for a magnetic disk. The glass substrate for a magnetic disk is a glass substrate for a magnetic disk having a plate thickness of less than 0.68 mm, a flatness of 30 μm or less, and a diameter of 95 to 100 mm, when a first heat treatment is performed in which it is maintained at 700°C for 4 hours and then cooled at a rate of 50°C / hour from 700°C to 400°C, the thermal shrinkage rate is 130 ppm or less, when the glass transition temperature of the glass substrate for a magnetic disk is represented by Tg (°C), after maintaining at Tg - 160°C for 60 seconds and then cooling to room temperature in the atmosphere, the change amount of flatness accompanying the second heat treatment is 10 μm or less, which is characterized.

[0018] It is preferable that the change amount of circularity accompanying the first heat treatment is 0.5 μm or less.

[0019] Another aspect of the present invention is a method for manufacturing a glass plate. The method for manufacturing the glass plate includes a step of annealing heat-treating the glass plate material that becomes the glass plate, The glass plate A rectangular plate with a thickness less than 0.68 mm, From each of both ends in the short side direction of the glass plate, an end region having a length of 5 to 20% of the short side of the glass plate is formed inside the glass plate, and from each of both ends in the long side direction of the glass plate, an end region having a length of 5 to 20% of the long side of the glass plate is formed inside the glass plate. The flatness of a square measurement region with a side length of 100 mm cut out from the central region of the glass plate excluding these is 30 μm or less. After maintaining the measurement region at 700 °C for 4 hours, when a first heat treatment of cooling from 700 °C to 400 °C at a rate of 50 °C / hour is performed, the thermal shrinkage rate of the measurement region is 130 ppm or less. When the glass transition temperature of the glass plate is represented by Tg (°C), after maintaining the measurement region at Tg - 160 °C for 60 seconds and then cooling it to room temperature in the atmosphere, the change amount of the flatness accompanying the second heat treatment is 10 μm or less. This is the feature.

[0020] Another aspect of the present invention is a method for manufacturing a glass plate. The method for manufacturing the glass plate is as follows. A step of annealing a glass plate material that is the source of the glass plate, A step of taking out the glass plate from the glass plate material after the annealing treatment, and includes. The glass plate is A rectangular plate with a thickness less than 0.68 mm, a flatness of 30 μm or less, and the lengths of two orthogonal sides are 95 to 120 mm each. After maintaining at 700 °C for 4 hours, when a first heat treatment of cooling from 700 °C to 400 °C at a rate of 50 °C / hour is performed, the thermal shrinkage rate is 130 ppm or less. When the glass transition temperature of the glass plate is represented by Tg (°C), after maintaining at Tg - 160 °C for 60 seconds and then cooling it to room temperature in the atmosphere, the change amount of the flatness accompanying the second heat treatment is 10 μm or less. This is the feature.

[0021] The rectangular glass plate is square and is a base plate that is the source of a disk-shaped glass having a circular outer periphery. Preferably, the area of ​​the main surface of the rectangular glass plate is 1.6 times or less the area of ​​the inner surface of the outer circumference of the disc-shaped glass.

[0022] Another aspect of the present invention is a method for manufacturing a glass plate. The method for manufacturing the glass plate is as follows: A method for manufacturing disc-shaped glass, The process of annealing the glass plate material that will form the disc-shaped glass, The process includes the step of removing the disc-shaped glass from the glass plate material after the annealing treatment, The glass plate material is a rectangular glass plate, The aforementioned disc-shaped glass is The plate thickness is less than 0.68 mm, the flatness is 30 μm or less, and the diameter is 95 to 100 mm. The thermal shrinkage rate after the first heat treatment, in which the material is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour, is 130 ppm or less. The glass transition temperature of the disc-shaped glass is expressed in terms of Tg (°C), and the change in flatness associated with the second heat treatment, which involves maintaining the glass at Tg -160°C for 60 seconds and then cooling it to room temperature in air, is 10 μm or less.

[0023] The aforementioned rectangular glass plate is square, Preferably, the area of ​​the main surface of the rectangular glass plate is 1.6 times or less the area of ​​the inner surface of the outer circumference of the disc-shaped glass. [Effects of the Invention]

[0024] The above-described glass substrate for magnetic disks can suppress the deterioration of flatness that occurs during heat treatment to form the magnetic recording layer of the magnetic disk. Furthermore, the above-described glass plate and disc-shaped glass can be used to obtain such a glass substrate for magnetic disks. Moreover, the above-described method for manufacturing the glass plate can be used to obtain the above-described glass plate. [Brief explanation of the drawing]

[0025] [Figure 1] (a) is an external view of a glass plate (large glass plate) according to one embodiment, and (b) is a plan view illustrating the measurement area of ​​the glass plate. [Figure 2] (a) is an external view of a glass plate (fragmented glass) according to one embodiment, and (b) is a plan view of the glass plate showing the portion that will become a disc-shaped glass. [Figure 3] This is an external view of a disc-shaped glass, which is one embodiment of the present invention. [Figure 4] This is an external view of a glass substrate for a magnetic disk, which is one embodiment of the present invention. [Modes for carrying out the invention]

[0026] The following describes in detail an embodiment of a glass plate, a method for manufacturing a glass plate, a disc-shaped glass, and a glass substrate for a magnetic disk.

[0027] (Large sheet glass) Figure 1(a) shows an external view of a glass plate 10 according to one embodiment. Figure 1(b) shows a plan view illustrating the measurement area 13 of the glass plate 10, which will be described later.

[0028] The glass plate 10 is a rectangular plate with a thickness of less than 0.68 mm.

[0029] By having a glass plate thickness of less than 0.68 mm, the thickness of the glass substrate for magnetic disks (hereinafter also referred to as glass substrate) made from the glass plate 10 can be reduced, thereby increasing the number of disks that can be mounted in the HDD device. The thickness of the glass plate 10 is preferably less than 0.61 mm, and more preferably less than 0.58 mm. There is no particular lower limit to the thickness of the glass plate 10, but for example, it is 0.2 mm.

[0030] The shorter side length of the glass plate 10 is preferably greater than 900 mm. This allows for the production of many glass substrates for magnetic disks from the glass plate 10, thereby reducing the manufacturing cost of the glass substrates for magnetic disks. The longer side length of the glass plate 10 may be longer than the shorter side length, as shown in the example in Figure 1, or it may be equal to the shorter side length. In other words, the glass plate 10 is rectangular or square. When the glass plate 10 has a shorter side and a longer side, the ratio of the length of the longer side to the length of the shorter side (length of the longer side ÷ length of the shorter side) is preferably 1.2 or less. By performing the precision annealing treatment described later on the glass plate material that forms the basis of the glass plate 10 with the above ratio of 1.2 or less, it becomes easier to reduce the thermal shrinkage rate and also easier to reduce the anisotropy of the thermal shrinkage rate (described later). The reason for this is thought to be that a workpiece that is closer to a square is less likely to experience differences in thermal history due to differences in position within the workpiece surface during the annealing treatment. In this specification, a glass plate with a shorter side length exceeding 900 mm is sometimes referred to as a "large glass plate." Furthermore, it is preferable that the length of the long side of the glass plate 10 is 2000 mm or less. If the length of the long side exceeds 2000 mm, it may become difficult to maintain temperature uniformity inside the furnace when performing the precision annealing process described later.

[0031] The flatness of the measurement area 13 cut from the central region of the glass plate 10 is 30 μm or less. The flatness of the measurement area 13 is 30 μm or less, which reduces the amount of material that needs to be removed by grinding or polishing when manufacturing a glass substrate for a magnetic disk from the glass plate 10, resulting in a higher yield of glass substrates for magnetic disks. Furthermore, the flatness of the measurement area 13 is 30 μm or less, which reduces fluttering when a magnetic disk made from the glass plate 10 is rotated at high speed, allowing for stable reading by the head of the HDD device's reading unit. In particular, when the thickness of the magnetic disk is thin, the low rigidity of the glass substrate can cause deflection, which can lead to fluttering. However, the low flatness of the glass plate 10 suppresses fluttering even with a thin plate thickness. In this specification, flatness refers to flatness in accordance with JIS B0621-1984. Flatness can be measured, for example, using an interferometric flatness measuring instrument and phase-shift interferometry at a predetermined measurement wavelength (e.g., 680 nm). The flatness of the area to be measured 13 is preferably 20 μm or less, more preferably 10 μm or less. The above flatness refers to the flatness of the area to be measured 13 before the first or second heat treatment described later is performed. The same applies hereafter unless otherwise specified.

[0032] The central region 12 of the glass plate 10 refers to the region of the glass plate 10 excluding the end region 11a, which is 5-20% of the length L of the short side 10a of the glass plate 10 and is located inside the glass plate 10, or the end region 11b, which is 5-20% of the length W of the long side 10b of the glass plate 10 and is located inside the glass plate 10, from each end of the short side direction of the glass plate 10. The length of the central region 12 in the short side direction is Lc, and the length in the long side direction is Wc.

[0033] The measurement area 13 is a square area with sides of 100 mm, cut out from the central area 12 of the glass plate 10. The measurement area 13 does not have to be cut out as shown in Figure 1(b), and can be cut out arbitrarily from the central area 12. The size and shape of the measurement area 13 are close to the size and shape of the glass plate (described later as "fragmented glass") that will be the basis for the glass substrate for the magnetic disk.

[0034] According to the inventors' investigations, when the glass transition temperature of the glass plate 10 is expressed in Tg (°C), the thermal shrinkage rate of the measured region 13 when a first heat treatment is performed, in which the measured region 13 is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour, is 130 ppm or less, and the change in the flatness of the measured region 13 accompanying a second heat treatment, in which the measured region 13 is maintained at Tg -160°C for 60 seconds and then cooled to room temperature in air, is 10 μm or less, resulting in the following effect: When a magnetic film is heat-treated on a glass substrate for a magnetic disk made from a glass plate 10 with a plate thickness of less than 0.68 mm and a flatness of the measured region 13 of 30 μm or less, thermal shrinkage of the glass substrate is suppressed, thereby suppressing deformation such as bending while thermally shrinking the glass substrate, and as a result, deterioration of the flatness of the glass substrate can be suppressed. As a result of these effects, deterioration of flatness in glass plates 10 with a thickness of 30 μm or less is suppressed in the glass substrate after heat treatment of the magnetic film, and fluttering is suppressed when the magnetic disk is rotated at high speed.

[0035] For the reasons stated above, in this embodiment, the glass plate 10 is configured such that the thermal shrinkage rate of the area to be measured 13 after a first heat treatment in which the area to be measured 13 is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour is 130 ppm or less, and the change in flatness of the area to be measured 13 during a second heat treatment in which the area to be measured 13 is maintained at Tg-160°C for 60 seconds and then cooled to room temperature in the atmosphere is 10 μm or less. If the thermal shrinkage rate of the area to be measured 13 after the first heat treatment exceeds 130 ppm, and the change in flatness of the area to be measured 13 during the second heat treatment exceeds 10 μm, deformation of the glass substrate that bends while shrinking due to heat cannot be suppressed, and the flatness of the glass plate deteriorates. The high flatness of 30 μm or less is greatly impaired even by slight thermal shrinkage. The conditions for the first and second heat treatments are determined by referring to the treatment conditions when heat-treating a magnetic film. The conditions for the first heat treatment are determined from the perspective of providing temperature conditions that allow evaluation of the thermal shrinkage rate during high-temperature and long-duration heating, which is presumed to be related to the deterioration of the flatness of the glass substrate during heat treatment of a magnetic film at 600°C or higher. The conditions for the second heat treatment are determined from the perspective of providing temperature conditions that allow direct evaluation of the amount of deterioration of the flatness of the glass substrate during heat treatment of a magnetic film at 600°C or higher. This is because the temperature required to form a magnetic film with an L10 structure, which is considered optimal for energy-assisted magnetic recording methods (EAMR) such as heat-assisted magnetic recording (HAMR) and microwave-assisted magnetic recording (MAMR), can reach far above 600°C, sometimes exceeding 700°C. In this specification, unless otherwise specified, the thermal shrinkage rate refers to the thermal shrinkage rate before and after the first heat treatment, and means the maximum value of the thermal shrinkage rate measured in 25 directions parallel to the main surface of the object being measured, passing through the center of the object and changing by 7.2 degrees in the circumferential direction. Furthermore, by measuring the thermal shrinkage rate using the above method, it is possible to measure the thermal shrinkage rate in all directions (360 degrees), thus enabling a more accurate evaluation of the thermal shrinkage rate than conventional methods.

[0036] The heating (increasing temperature), temperature maintenance, and cooling (decreasing temperature) in the first heat treatment are preferably carried out continuously in an atmosphere with the same conditions except for the temperature (for example, in the atmosphere of a single annealing furnace). The heating in the first heat treatment is preferably carried out from room temperature over 2.5 hours. In other words, it is preferable that the substrate to be treated is heated from room temperature to 700°C at a rate of 270°C / hour. Furthermore, it is preferable that the substrate is cooled from 700°C to room temperature at a rate of 50°C / hour. Furthermore, when performing the first heat treatment, in order to avoid a significant deterioration in the flatness of the substrate to be treated (which may be the measurement area 13 of the glass plate 10, or a glass plate 20, a disc-shaped glass 30, or a glass substrate 40 for magnetic disks, as described later), it is preferable to perform the treatment by placing the substrate flat and sandwiching it from above and below using two setters, as described later. By avoiding a significant deterioration in flatness, it becomes possible to accurately measure the thermal shrinkage rate. At this time, the size of the setters should be equal to or greater than that of the substrate to be treated. The thickness of the setter placed on top of the substrate to be treated should be set to a weight that does not hinder the thermal shrinkage of the substrate to be treated and maintains the flatness to a certain extent (for example, maintaining a flatness of 30 μm or less). Needless to say, a setter that is so heavy that it thins the thickness of the substrate to be treated is inappropriate. By maintaining the flatness of the substrate to be treated to a certain extent, it is possible to evaluate the change in flatness due to the second heat treatment, as described later, after evaluating the thermal shrinkage rate due to the first heat treatment, using the same substrate to be treated. Furthermore, the evaluation of the thermal shrinkage rate due to the first heat treatment and the evaluation of the change in flatness due to the second heat treatment, described later, may be performed using separate substrates to be treated.

[0037] Regarding the second heat treatment, cooling in the atmosphere means allowing the substrate to cool in a room temperature atmosphere without temperature control for controlling the cooling rate. Room temperature is, for example, 25°C. Heating and maintaining the temperature of the substrate in the second heat treatment is performed, for example, in the atmosphere between heaters in a heating device equipped with two panel-shaped heaters, with the substrate held in a substrate holder. This heating device is modeled after a substrate heating chamber provided in a known single-wafer vacuum film deposition apparatus used for depositing magnetic films on magnetic disks, etc. The substrate is set in a known film deposition substrate holder (also called a carrier) in an orientation perpendicular to the ground. The substrate holder (for example, the substrate holder described in paragraph 0045, Figure 4, etc. of Japanese Patent Application Publication No. 2011-117019) has three or four L-shaped leaf spring support members fixed to it, and the substrate is fixed to the substrate holder by the elasticity of the leaf springs by pressing their tips against the outer peripheral end face of the substrate. This system can hold circuit boards not only when they are circular, but also when they are rectangular or other shapes. By adjusting the specifications of the circuit board holder, circuit boards of various shapes can be heated in the same way. However, it is important to note that the circuit board is constantly subjected to a force from the support member due to the elasticity of the leaf spring, which causes the board to bend. Therefore, it is more prone to bending compared to heating without a support member (for example, when the board is placed flat).

[0038] The thermal shrinkage rate can be determined, for example, by measuring the change in length in the measurement area 13 before and after heat treatment and calculating it according to the following formula. C (thermal contraction coefficient) = (L0 - L) / L0 Here, L0 is the length before heat treatment, and L is the length after heat treatment. The sign of C is positive if it shrinks due to heat treatment, and negative if it expands. L0 and L can be determined, for example, by making two markings on the surface of the cut-out measurement area 13 and measuring the distance between the two markings before and after heat treatment. Alternatively, L0 and L may be the lengths of the measurement area 13 before and after heat treatment. Preferably, these lengths pass through the center of the measurement area 13. If the object to be measured is a disc-shaped glass or a glass substrate for a magnetic disk, the diameter may be used. Furthermore, when evaluating the anisotropy of thermal shrinkage, for example, lengths (e.g., diameter) in 25 directions, varying by 7.2 degrees in the circumferential direction with respect to the center of the object being measured, can be used. The absolute value of the difference in thermal shrinkage obtained by measuring the thermal shrinkage in 25 directions and subtracting the minimum value from the maximum value can be used as an indicator of the anisotropy of thermal shrinkage. Alternatively, instead of thermal shrinkage (C), thermal shrinkage amount (S) may be used, and the absolute value of the difference between the maximum and minimum thermal shrinkage amounts in the above 25 directions may be calculated and used as the indicator of anisotropy. S (thermal shrinkage amount) is given by S = (L0 - L).

[0039] The thermal shrinkage rate of the measured region 13 after the first heat treatment is preferably 90 ppm or less, more preferably 50 ppm or less. The change in flatness of the measured region 13 after the second heat treatment is preferably 7.5 μm or less, more preferably 5 μm or less.

[0040] Preferably, the glass plate 10 is a portion cut from a long glass sheet formed using one of the following methods: float method, Fulcol method, Pittsburgh method, downdraw method, Colburn method, and redraw method. Since a large glass plate 10 can be obtained from a glass sheet formed by these methods, a large amount of individual glass fragments, which are the basis for glass substrates for magnetic disks, can be obtained from the glass plate 10, thereby reducing the manufacturing cost of glass substrates for magnetic disks. Furthermore, since these methods are advantageous for forming glass sheets with a high glass transition temperature (Tg), the manufacturing cost of glass plates 10 with a high glass transition temperature (Tg) can be reduced. Specific examples of the downdraw method include the slot downdraw method and the overflow downdraw method. It is also preferable that at least one of the main surfaces of the glass plate 10 is a forged surface. This makes it possible to omit some of the grinding and polishing processes of the main surface of the substrate, which are generally required when manufacturing glass substrates for magnetic disks, or to reduce the material removal amount. In other words, it is preferable that at least one of the main surfaces of the glass plate 10 is an unground surface and / or an unpolished surface.

[0041] In glass sheets obtained by the float method or other methods described above, the thickness of the sheet is usually greater at both ends in the width direction perpendicular to the longitudinal direction of the glass sheet (the direction in which the glass flows out of the melting furnace) compared to the center in the width direction. Therefore, the glass sheet material that will become the glass plate 10 is cut from the remaining portion of the glass sheet after both ends in the width direction have been cut off. Typically, the glass sheet material that will become the glass plate 10 is cut so that the width direction of the glass sheet coincides with the short side direction or the long side direction of the glass plate 10.

[0042] The glass plate 10 may exhibit anisotropy in its thermal shrinkage rate. Anisotropy in thermal shrinkage rate refers to the characteristic where the magnitude of the thermal shrinkage rate differs depending on the direction within the plane of the main surface of the measurement area 13. The inventors' studies have revealed that if the glass plate 10 exhibits anisotropy in its thermal shrinkage rate, the roundness of the glass substrate (JIS B0621-1984) may deteriorate when the magnetic film on the glass substrate for magnetic disks obtained from the glass plate 10 is heat-treated. In particular, if the glass plate material from which the glass plate 10 is made is a portion cut from a glass sheet formed using the float method or the methods described above, differences in thermal shrinkage rate are likely to occur depending on the direction within the plane of the glass sheet, and anisotropy in thermal shrinkage rate is likely to occur. Furthermore, the inventors have also found that the direction in which the thermal shrinkage rate is maximum and the direction in which it is minimum are not necessarily 90 degrees perpendicular to each other in the in-plane direction of the main surface. In other words, conventionally, when evaluating the anisotropy of thermal shrinkage, the thermal shrinkage rates in two directions—the longitudinal direction of the glass sheet and the width direction perpendicular to it—were measured, and the difference between them was used as an indicator of anisotropy. However, it has been found that this method may not accurately evaluate the maximum and minimum values ​​of thermal shrinkage, or the difference between them. The reason for this is not entirely clear, but it is presumed that in methods that continuously produce long glass sheets, such as the float method and the down-draw method, the glass flowing out of the melting or softening furnace is pulled in the flow direction as it moves, and is also stretched in the width direction to form a glass sheet. Therefore, it is also pulled in an oblique direction which is a combination of the two perpendicular directions mentioned above. Furthermore, the direction of this pulling differs depending on the molding conditions and the position within the glass sheet, and also changes over time. In addition, the thermal history differs depending on the position, so it is thought that the direction in which the thermal shrinkage rate is maximum, the direction in which it is minimum, and the magnitude of the thermal shrinkage rate all change in various ways. Therefore, when precisely evaluating anisotropy, it is necessary to cut the desired glass from a long glass sheet and examine it in all directions.

[0043] As described above, if the roundness of the outer edge of the glass substrate for magnetic disks deteriorates, wobble occurs when the magnetic disk is rotated at high speed, making fluttering more likely. Therefore, it is extremely important to precisely understand the values ​​of the thermal shrinkage in the direction in which the thermal shrinkage rate is maximum and minimum, as well as the difference between them, for the glass substrate for magnetic disks, the disc-shaped glass that forms its base, and the glass plate from which the disc-shaped glass is made. From the viewpoint of suppressing such deterioration of the roundness of the glass substrate, it is preferable that the difference (absolute value) between the thermal shrinkage rate C1 in the direction in which the thermal shrinkage rate is minimum and the thermal shrinkage rate C2 in the direction in which the thermal shrinkage rate is maximum, among the in-plane directions of the measured region 13, be 10 ppm or less. Furthermore, it is preferable that the difference (absolute value) between the amount of thermal shrinkage S1 of the measured region in the direction where the thermal shrinkage rate is minimum and the amount of thermal shrinkage S2 of the measured region in the direction where the thermal shrinkage rate is maximum is 1.0 μm or less.

[0044] Preferably, the glass plate 10 has been subjected to an annealing treatment (for example, "precision annealing" described later) to reduce the thermal shrinkage rate. The glass plate before annealing (the glass plate material that forms the basis of the glass plate 10) has anisotropy in thermal shrinkage rate, where the magnitude of the thermal shrinkage rate differs depending on the in-plane direction of the region of the glass plate corresponding to the measurement region 13. In some cases, the difference (absolute value) between the thermal shrinkage rate C1 in the direction with the minimum thermal shrinkage rate and the thermal shrinkage rate C2 in the direction with the maximum thermal shrinkage rate may be greater than 10 ppm. Also, the difference (absolute value) between the amount of thermal shrinkage S1 in the region in the direction with the minimum thermal shrinkage rate and the amount of thermal shrinkage S2 in the region in the direction with the maximum thermal shrinkage rate may be greater than 1.0 μm. Even when such anisotropy in thermal shrinkage rate is present, the annealed glass plate 10 satisfies the range of changes in thermal shrinkage rate and flatness described above, so when a magnetic film is heat-treated on a glass substrate for magnetic disks obtained from the glass plate 10, deterioration of flatness and roundness is suppressed. The change (deterioration) in roundness is preferably 0.5 μm or less, and more preferably 0.2 μm or less.

[0045] According to one embodiment, it is preferable that a 100 mm square area to be measured, cut out from the entire glass plate 10 including the edge regions 11a and 11b, has flatness, thermal shrinkage rate, and change in flatness that satisfy the above-mentioned ranges, similar to the area to be measured 13. From such a glass plate 10, more glass substrates for magnetic disks can be manufactured compared to when a glass substrate for magnetic disks is manufactured from the central region 12.

[0046] For the material of the glass plate 10, it is preferable to use, for example, aluminosilicate glass, soda-lime glass, soda-aluminosilicate glass, aluminoborosilicate glass, or borosilicate glass.

[0047] The glass transition temperature (Tg) of the glass plate 10 is preferably 750°C or higher, and more preferably 770°C or higher. A glass substrate for a magnetic disk made from such a glass plate 10 with a high glass transition temperature (Tg) is less prone to deformation at high temperatures, and therefore has a significant effect in suppressing deterioration of flatness when the magnetic film is heat-treated at, for example, 700°C. There is no particular upper limit to the glass transition temperature (Tg) of the glass plate 10, but it is preferably 850°C or lower. If the glass transition temperature (Tg) exceeds 850°C, it may become difficult to form a thin sheet of glass. Furthermore, the Young's modulus of the glass plate 10 is preferably 80 GPa or higher. If the Young's modulus is less than 80 GPa, for example, when the magnetic film is heat-treated at 700°C, warping due to elastic stress from the support member for holding the substrate may occur, and this, combined with the warping caused by the heat treatment, may significantly worsen the flatness. If the flatness deteriorates too much, problems such as the substrate falling from the substrate holder during film formation may occur. Furthermore, the average coefficient of linear thermal expansion of the glass plate 10 at 100-300°C is 45 × 10⁻⁶. -7 It is preferable that the temperature is below / ℃. The average coefficient of linear thermal expansion is 45 × 10 -7 When temperatures exceed [temperature]°C, there is a higher risk of the substrate cracking when rapidly heating or cooling it to improve productivity. Furthermore, the density of the glass plate 10 is 2.65 g / cm³. 3 Preferably, it is 2.60 g / cm³. 3 The following is more preferable. If the density is too high, the weight will increase when used as a glass substrate for magnetic disks, which tends to increase the power consumption of the HDD.

[0048] (Method of manufacturing glass plates) The glass plate 10 described above can be manufactured by a glass plate manufacturing method that includes an annealing process in which the glass plate material that will become the glass plate 10 is heated under predetermined conditions. In the following description, this annealing process of the glass plate material performed under predetermined conditions will be referred to as "precision annealing". Precision annealing is performed on the glass plate material so that the measurement area 13 of the glass plate 10 satisfies the range of changes in thermal shrinkage rate and flatness described above.

[0049] According to the inventors' research, as described above, when a glass substrate for a magnetic disk is made by extracting a small piece of glass from a conventional glass plate manufactured by the float method or other methods described above, and the magnetic film that will become the magnetic recording layer is heat-treated, the glass substrate deforms by shrinking and bending due to the heat, resulting in a deterioration of the flatness of the magnetic disk. Glass sheets formed by the float method or other methods described above are stretched in various directions and rapidly cooled over a large area, making it difficult to maintain constant stress, temperature, and thermal history across the entire glass sheet, and making it difficult to uniformly reduce the thermal shrinkage rate across the entire glass sheet. Therefore, it is thought that variations in the thermal shrinkage rate occur due to differences in the in-plane position of the glass sheet. As a result, when a portion is extracted from the glass sheet and heated later, there are cases where the amount of thermal shrinkage is large. In other words, when using a glass sheet as a glass substrate for an FPD, it is used in its large area state, so it is sufficient if a single value of the thermal shrinkage rate measured for the entire glass sheet is within the acceptable range, and it was not necessary to consider in-plane variation in thermal shrinkage rate. However, since glass substrates for magnetic disks are much smaller in size than glass substrates for FPDs, it has been found that in some cases, glass substrates for magnetic disks with large thermal shrinkage rates are produced due to in-plane variation in thermal shrinkage rate. In addition, the heat treatment temperature of magnetic films formed on glass substrates for magnetic disks has been increasing in recent years, sometimes reaching over 700°C. This temperature is close to the glass transition temperature (Tg) of high-heat-resistant glass substrates. Since the heat treatment conditions for such magnetic films are much stricter than the heating conditions for glass substrates when depositing TFTs on FPD glass substrates (e.g., 350-600°C), it has been found that even a small thermal shrinkage rate that is not a problem for FPD glass substrates due to slow cooling during molding can have a significant adverse effect when the magnetic film is heat treated. In other words, it has become clear that the heat treatment of the magnetic film can cause the glass substrate to shrink significantly due to the heat, or to deform in a way that causes it to bend while shrinking due to the heat.The inventors have found that by performing the precision annealing described above on the glass plate material that forms the basis of the glass plate 10, it is possible to precisely remove in-plane variations in thermal shrinkage while maintaining the flatness of the glass plate 10 below a predetermined value, that is, a glass plate 10 can be obtained in which the amount of change in thermal shrinkage and flatness of the measured region 13 is within a predetermined range.

[0050] Furthermore, as the inventors continued their investigations, they discovered that even when conventional annealing treatments, such as offline annealing, were performed to address the aforementioned problem of the glass substrate deforming by shrinking and bending during heat treatment of the magnetic film, the following problem arose. Specifically, when a general annealing treatment was performed on the glass plate material that forms the basis of a large glass plate, the effect of annealing (such as the effect of reducing the thermal shrinkage rate) did not spread uniformly throughout the entire surface of the glass plate material. As a result, even if the thermal shrinkage rate of the large glass plate as a whole was below a predetermined value after the first heat treatment described above, when multiple rectangular glass plates, each with a side size of, for example, 95 to 120 mm, were cut (fragmented) from the large glass plate, some of the fragmented glass plates had a thermal shrinkage rate that did not fall below the predetermined value, or warped when the magnetic film was heat-treated as a glass substrate for a magnetic disk. This resulted in variations in properties among the fragmented glass plates. In particular, when the glass plate material that forms the basis of the glass plate 10 is a portion cut from a glass sheet formed using the above-mentioned methods such as the float method or the down-draw method, it was found that such variations may occur in the glass substrate for magnetic disks that is made by cutting from the central region, excluding the edge regions near each side of the glass sheet. In the case of the central region of a large glass plate, it is not possible to directly measure the annealing effect in a narrower region within the central region, so even if there are areas with a low annealing effect, it was not possible to notice them. When the inventors investigated the cause of such variations even after performing a general annealing process, they inferred that the slight difference in thermal history between the outer and central parts of the glass plate material during the annealing process was the main influencing factor. Furthermore, they found that the thermal shrinkage rate of the central part (central region) of a large glass plate cannot be accurately determined unless the target portion is cut out, as long as it is connected to the surrounding outer part (end region), the outer part can restrict the movement of the central part and hinder its thermal shrinkage, or the central part can shrink excessively as it is dragged along by the thermal shrinkage of the outer part. The inventors found that with a glass plate 10 obtained by performing the precision annealing described above, the variations in the effect of annealing are eliminated, and even in the fragmented glass (fragmented glass) taken from the central region 12, variations in properties among such fragmented glass are suppressed. Therefore, the measurement area 13 is cut out from the central region 12 of the glass plate 10 as described above. The sizes of the end regions 11a and 11b, which determine the size of the central region 12, are determined from the viewpoint of suppressing the variation in the amount of change in thermal shrinkage and flatness between multiple individual glass fragments.

[0051] For the reasons stated above, precision annealing is performed on the glass plate material such that the area 13 to be measured on the glass plate 10 satisfies the range of changes in thermal shrinkage and flatness described above. Preferably, precision annealing is performed by a heat treatment at Tg-110°C or higher for 4 hours or more, more preferably by a heat treatment at Tg-80°C or higher for 4 hours or more. The conditions for such a heat treatment are determined by referring to the treatment conditions when heat treating a magnetic film. Preferably, heating (increasing temperature), maintaining temperature, and cooling (decreasing temperature) in precision annealing are performed continuously in an atmosphere with the same conditions except for the temperature (for example, in the atmosphere of a single annealing furnace). Preferably, the heating in precision annealing is performed from room temperature over 2.5 hours. In other words, it is preferable that the glass plate material is heated from room temperature to a temperature of Tg-110°C or higher, more preferably to a temperature of Tg-80°C or higher, at a rate of 270°C / hour. Furthermore, regarding temperature reduction, it is preferable to cool (defrost) from a temperature of Tg-110°C or higher, more preferably Tg-80°C or higher, to room temperature at a rate of 50°C / hour.

[0052] Precision annealing is preferably performed using the annealing plate material (hereinafter referred to as "setter") described below. This makes it possible to efficiently obtain a glass plate 10 that satisfies the above-mentioned ranges for flatness, thermal shrinkage rate, and change in flatness.

[0053] The setter has a plate-like form with a pair of main surfaces, and at least one of these surfaces is configured to be in contact with the main surface of the glass plate material that will become the glass plate 10. The main surface of the setter is wider than the main surface of the glass plate material that will become the glass plate 10, and extends beyond the entire circumference of the glass plate material. The length of the extension is, for example, 5 centimeters or more in the direction away from the center of the glass plate material.

[0054] The flatness of the setter is preferably less than 30 μm, more preferably 20 μm or less, and even more preferably 10 μm or less, in order to efficiently obtain glass plates 10 with a flatness of 30 μm or less.

[0055] The thermal conductivity of the setter is, for example, 1 to 200 W / (m·K) at 20°C. Having the thermal conductivity of the setter within this range allows for uniform heating and cooling of the glass plate material when precision annealing is performed on the glass plate material that forms the basis of the glass plate 10. This effectively suppresses variations in the magnitude of thermal shrinkage after precision annealing depending on the in-plane position of the glass plate 10.

[0056] Examples of setter materials include alumina (Al2O3), silicon carbide (SiC), silicon nitride (Si3N4), zirconia (ZrO2), Sialon (Si3N4·Al2O3), steatite, spinel, and cordierite. Among these, alumina (Al2O3) and silicon carbide (SiC) are preferred.

[0057] One example of precision annealing using setters is a method using two setters and an insulating material. In one embodiment, precision annealing is preferably performed with a glass plate sandwiched between two setters having a larger main surface area than the glass plate, and surrounded by an insulating material placed in the gap between the setters. The insulating material is preferably made of a fibrous material with high heat resistance. In addition to rock wool, which will be described below, inorganic fibers such as ceramic fibers and glass fibers are preferably used as the fibrous material. In this example of precision annealing, the setters and glass plate are stacked so that one glass plate is sandwiched between two setters with a larger surface area than the glass plate, and the gap between the two setters adjacent to the side (end face) of the glass plate is filled with high heat-resistant rock wool. Here, the two setters are made to be the same shape, and are arranged so that they almost perfectly overlap each other in the thickness direction while sandwiching the glass plate (i.e., there is no misalignment in the in-plane direction), and the outer periphery of the setters protrudes almost uniformly from the glass plate all around, so that a gap is created between the two setters near the entire edge of the glass plate. By lightly filling this gap with high heat-resistant rock wool, the entire glass plate is covered with setters and rock wool, and the load of the setters can be applied appropriately and evenly to the main surface of the glass plate. In this example of precision annealing, the weight on the glass plates is lighter compared to the case where multiple setters and multiple glass plates are stacked alternately and then precision annealed. Therefore, the influence of the weight of the setters and glass plates can suppress the inhibition of in-plane expansion and contraction of the lower glass plates, and contribute to reducing the thermal shrinkage rate regardless of the in-plane position of the glass plate 10. Because rock wool has both thermal insulation and breathability, it can effectively seal the gaps between setters. As a result, it becomes easier to heat or cool the entire glass plate evenly (uniform heating), and the thermal shrinkage rate can be reduced regardless of the in-plane position of the glass plate. Furthermore, the rock wool is used in an amount that does not hinder the load on the glass plate by the setter placed on the glass plate. Therefore, since the load of the setter is applied appropriately and evenly to the main surface of the glass plate, it is possible to suppress deterioration of the flatness of the glass plate during precision annealing, and in some cases, reduce the flatness. In other words, the above method makes it possible to reduce the effect of the weight of the setter on the glass plate material, thereby reducing the thermal shrinkage rate of the glass plate 10 regardless of its position in the plane, while also enhancing the effect of reducing the flatness of the glass plate material. The precision annealing described above is not limited to being performed on the sheet material that forms the basis of the glass plate 10, but can also be performed on the individual pieces of glass sheet material that form the basis of the glass plate 20, as will be described later.

[0058] The glass plate 10 can be manufactured by the glass plate manufacturing method, which includes the precision annealing described above.

[0059] (Fragmented glass) Figure 2(a) shows an external view of a glass plate 20 according to one embodiment. Figure 2(b) shows a plan view of the glass plate 20, with the disc-shaped glass portion indicated by a dashed line.

[0060] The glass plate 20 has a thickness of less than 0.68 mm, a flatness of 30 μm or less, and is a rectangular plate with two orthogonal sides each measuring 95 to 120 mm in length. This glass plate is smaller in size than the glass plate (large glass plate) 10 described above, and in this specification, it may be referred to as "fragmented glass."

[0061] The glass plate 20 is a rectangular plate measuring 95 to 120 mm in diameter, and has a suitable size for manufacturing the disc-shaped glass (described later) that serves as the base plate for the glass substrate for magnetic disks, and consequently, the glass substrate for magnetic disks, as it requires less material to be removed during the manufacturing process. Preferably, the glass plate 20 is a square plate. By performing a precision annealing treatment on the glass plate material that forms the basis of the square glass plate 20, the thermal shrinkage rate can be reduced for the same reasons as described above, and the anisotropy of the thermal shrinkage rate can also be reduced. Note that even if the ratio of the length of the vertical and horizontal sides is slightly different (for example, if the ratio is between 0.95 and 1.05), it is still included within the range of a square plate as described above. The thickness of the glass plate 20 is preferably less than 0.61 mm, and more preferably less than 0.58 mm. The lower limit of the thickness is not particularly limited, but for example, it is 0.2 mm.

[0062] The thermal shrinkage rate of the glass plate 20 after a first heat treatment in which it is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour is 130 ppm or less. Preferably, this thermal shrinkage rate is 90 ppm or less, and more preferably 50 ppm or less. Furthermore, when the glass transition temperature of the glass plate 20 is expressed as Tg(°C), the change in the flatness of the glass plate 20 accompanying a second heat treatment in which it is maintained at Tg-160°C for 60 seconds and then cooled to room temperature in air is 10 μm or less. Preferably, this change in flatness is 7.5 μm or less, and more preferably 5 μm or less.

[0063] The glass plate 20 is the base plate from which the disc-shaped glass is made. The area of ​​the main surface of the glass plate 20 is preferably 1.6 times or less the area of ​​the inner circumference of the disc-shaped glass, and more preferably 1.5 times or less. In this way, by having a main surface area of ​​the glass plate 20 that is close to the area of ​​the inner circumference of the disc-shaped glass (the effect of internal pores is negligible), the disc-shaped glass is more likely to possess the two characteristics of thermal shrinkage rate and change in flatness that the glass plate 20 possesses. In other words, the above characteristics of the disc-shaped glass do not deviate significantly from the above characteristics of the glass plate 20 before cutting. This effect is particularly effective in glass plates 20 made by performing precision annealing (precision annealing after fragmentation) on the fragmented glass plate material that forms the basis of the glass plate 20. In other words, because the individualized glass plate material has a small surface area, the effect of precision annealing easily reaches even the corners of the outer periphery of the individualized glass plate material, and the variation in the annealing effect depending on the in-plane position is small. Therefore, the effect of suppressing the deterioration of flatness when the magnetic film is heat-treated in the disc-shaped glass cut from the glass plate 20 is greatly increased. Thus, the closer the shape of the glass plate material subjected to precision annealing is to the shape of a glass substrate for a magnetic disk, the greater the effect of precision annealing. Furthermore, fragmentation refers to obtaining fragmented glass 20 from a large sheet of glass 10, or obtaining glass sheets of the same dimensions as fragmented glass 20 from the glass sheet material that forms the basis of the large sheet of glass 10.

[0064] The glass plate 20 can be obtained, for example, by cutting it from the glass plate 10 (which has undergone precision annealing), or by performing precision annealing on individual pieces of the glass plate material that forms the basis of the glass plate 10.

[0065] Preferably, the glass plate 20 has been subjected to an annealing treatment (for example, the precision annealing described above) to reduce the thermal shrinkage rate. The glass plate before annealing (the glass plate material that forms the basis of the glass plate 20) has anisotropy in thermal shrinkage rate, where the magnitude of the thermal shrinkage rate differs depending on the in-plane direction of the glass plate. The difference (absolute value) between the thermal shrinkage rate C1 in the direction with the minimum thermal shrinkage rate and the thermal shrinkage rate C2 in the direction with the maximum thermal shrinkage rate may be greater than 10 ppm. Also, the difference (absolute value) between the amount of thermal shrinkage S1 of the glass plate in the direction with the minimum thermal shrinkage rate and the amount of thermal shrinkage S2 of the glass plate in the direction with the maximum thermal shrinkage rate may be greater than 1.0 μm.

[0066] Preferably, the difference (absolute value) between the thermal shrinkage rate C1 in the direction with the minimum thermal shrinkage rate and the thermal shrinkage rate C2 in the direction with the maximum thermal shrinkage rate in the in-plane direction of the glass plate 20 is 10 ppm or less. Furthermore, it is preferable that the difference (absolute value) between the amount of thermal shrinkage S1 in the direction with the minimum thermal shrinkage rate and the amount of thermal shrinkage S2 in the direction with the maximum thermal shrinkage rate is 1.0 μm or less.

[0067] The glass plate 20 described above is manufactured, for example, by cutting it from a large sheet of glass 10 and separating it into individual pieces. The cutting method may be by forming incisions and cleaving using a well-known scriber (cutter), or by irradiating the large sheet of glass 10 with laser light to form defects at regular intervals, and then separating and cutting the glass plate from the large sheet of glass 10 by connecting these defects. It is preferable that the glass plate 20 is cut from the central region 12 of the large sheet of glass 10. Therefore, it is preferable that at least one of the main surfaces of the glass plate 20 is a forged surface. This makes it possible to omit some of the grinding and polishing processes of the main surface of the substrate, which are generally required when manufacturing glass substrates for magnetic disks, or to reduce the amount of material removed. In other words, it is preferable that at least one of the main surfaces of the glass plate 20 is an unground surface and / or an unpolished surface.

[0068] Furthermore, the glass plate 20 can be manufactured, for example, by performing precision annealing on a glass plate material that will become the large glass plate 10 after it has been fragmented. In other words, the glass plate 20 can be manufactured by a glass plate manufacturing method that includes precision annealing of the fragmented glass plate material that will become the glass plate 20. In this method, precision annealing is performed on the large glass plate 10 in the same manner as the precision annealing described above, using the individualized glass plate material that will become the glass plate 20 as the target of heating. The individualized glass plate material used in this method that will become the glass plate 20 is, for example, a plate cut and individualized from the glass plate material that will become the large glass plate 10 without precision annealing, and has approximately the same dimensions and shape as the glass plate 20. According to the inventors' research, it was found that by performing precision annealing on the individualized glass plate material that forms the basis of the glass plate 20, a glass plate 20 can be obtained with even smaller changes in thermal shrinkage rate and flatness compared to a glass plate 20 cut from a large glass plate 10 that has already undergone precision annealing. Therefore, by performing precision annealing on the individualized glass plate material that forms the basis of the glass plate 20, a glass plate 20 can be obtained with even smaller changes in thermal shrinkage rate and flatness.

[0069] (Disc-shaped glass) Figure 3 shows an external view of a disc-shaped glass 30, which is one embodiment of the product.

[0070] The disc-shaped glass 30 is, for example, a base plate that forms the basis of a glass substrate for a magnetic disk. The disc-shaped glass 30 has a circular outer circumference. A hole (internal hole) is provided in the center of the disc-shaped glass, penetrating in the thickness direction. The disc-shaped glass may have an annular shape, but it may not have an internal hole, as shown in the example disc-shaped glass 30 in Figure 3.

[0071] The disc-shaped glass 30 has a thickness of less than 0.68 mm and a flatness of 30 μm or less. The disc-shaped glass 30 has a thermal shrinkage rate of 130 ppm or less when subjected to a first heat treatment in which it is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour. The thermal shrinkage rate is preferably 90 ppm or less, more preferably 60 ppm or less, and even more preferably 50 ppm or less. When the glass transition temperature of the disc-shaped glass 30 is expressed as Tg (°C), the change in flatness accompanying the second heat treatment in which it is maintained at Tg-160°C for 60 seconds and then cooled to room temperature in air is 10 μm or less. The change in flatness is preferably 7.5 μm or less, more preferably 6 μm or less, and even more preferably 5 μm or less. The thickness of the disc-shaped glass 30 is preferably less than 0.61 mm, and more preferably less than 0.58 mm. The lower limit of the thickness is not particularly limited, but for example, it is 0.2 mm.

[0072] In the disc-shaped glass 30, it is preferable that the difference (absolute value) between the thermal shrinkage rate C1 in the direction where the thermal shrinkage rate is minimum and the thermal shrinkage rate C2 in the direction where the thermal shrinkage rate is maximum is 10 ppm or less. The above difference (C2-C1) is more preferably 7 ppm or less, and even more preferably 5 ppm or less. Furthermore, it is preferable that the difference (absolute value) between the amount of thermal shrinkage S1 in the direction where the thermal shrinkage rate is minimum and the amount of thermal shrinkage S2 in the direction where the thermal shrinkage rate is maximum is 1.0 μm or less. The above difference (S2-S1) is more preferably 0.7 μm or less, and even more preferably 0.5 μm or less.

[0073] The disc-shaped glass 30 can be obtained, for example, by cutting it from a glass plate 20. The disc-shaped glass 30 can be extracted from the glass plate 20 by, for example, a known scribing method or coring method. The scribing method can be carried out using, for example, a diamond scriber, a scribing wheel, or a laser. Therefore, it is preferable that at least one of the main surfaces of the disc-shaped glass 30 is a forged surface. This makes it possible to omit some of the grinding and polishing processes of the main surface of the substrate, which are generally required when manufacturing glass substrates for magnetic disks, or to reduce the amount of material removed. In other words, it is preferable that at least one of the main surfaces of the disc-shaped glass 30 is an unground surface and / or an unpolished surface. When the disc-shaped glass 30 is used as a base plate (intermediate) for a magnetic disk glass substrate, it is preferable to adjust the diameter according to the final size of the magnetic disk glass substrate to be manufactured. The values ​​and ranges shown below are all examples. When used as a base plate for a magnetic disk glass substrate with a nominal diameter of 3.5 inches, the outer diameter can be 95 to 100 mm. Also, when an inner hole is provided, the inner diameter can be 23 to 25 mm. On the other hand, when used as a base plate for a magnetic disk glass substrate with a nominal diameter of 2.5 inches, the outer diameter can be 65 to 70 mm. Also, when an inner hole is provided, the inner diameter can be 18 to 20 mm.

[0074] (Glass substrate for magnetic disks) Figure 4 shows an external view of a glass substrate 40 for a magnetic disk, which is one embodiment of the device. The glass substrate 40 for a magnetic disk shown in Figure 4 has an internal hole in its center.

[0075] The size of the glass substrate 40 is not limited, but examples include glass substrates for magnetic disks with nominal diameters of 3.5 inches or 2.5 inches. For a glass substrate for a magnetic disk with a nominal diameter of 3.5 inches, the outer diameter can be, for example, 95 to 100 mm, and the inner bore diameter can be, for example, 24 to 26 mm. Specifically, for example, the outer diameter can be 95 mm or 97 mm, and the inner bore diameter can be, for example, 25 mm. On the other hand, for a glass substrate for a magnetic disk with a nominal diameter of 2.5 inches, the outer diameter can be, for example, 65 to 70 mm, and the inner bore diameter can be, for example, 19 to 21 mm. Specifically, for example, the outer diameter can be 65 mm or 67 mm, and the inner bore diameter can be, for example, 20 mm.

[0076] The glass substrate 40 for magnetic disks has a thickness of less than 0.68 mm and a flatness of 30 μm or less. The glass substrate 40 for magnetic disks has a thermal shrinkage rate of 130 ppm or less when subjected to a first heat treatment in which it is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour. The thermal shrinkage rate is preferably 90 ppm or less, more preferably 60 ppm or less, and even more preferably 50 ppm or less. When the glass transition temperature of the glass substrate 40 for magnetic disks is expressed as Tg (°C), the change in flatness accompanying a second heat treatment in which it is maintained at Tg-160°C for 60 seconds and then cooled to room temperature in air is 10 μm or less. The change in flatness is preferably 7.5 μm or less, more preferably 6 μm or less, and even more preferably 5 μm or less. The thickness of the glass substrate 40 for the magnetic disk is preferably less than 0.61 mm, and more preferably less than 0.58 mm. The lower limit of the thickness is not particularly limited, but for example, it is 0.2 mm.

[0077] In the glass substrate 40 for magnetic disks, it is preferable that the difference (absolute value) between the thermal shrinkage rate C1 in the direction with the minimum thermal shrinkage rate and the thermal shrinkage rate C2 in the direction with the maximum thermal shrinkage rate is 10 ppm or less. The above difference (C2-C1) is more preferably 7 ppm or less, and even more preferably 5 ppm or less. Furthermore, it is preferable that the difference (absolute value) between the amount of thermal shrinkage S1 in the direction with the minimum thermal shrinkage rate and the amount of thermal shrinkage S2 in the direction with the maximum thermal shrinkage rate is 1.0 μm or less. The above difference (S2-S1) is more preferably 0.7 μm or less, and even more preferably 0.5 μm or less.

[0078] The glass substrate 40 for magnetic disks is obtained by a method for manufacturing a glass substrate for magnetic disks, which includes, for example, grinding and / or polishing the main surface of the disc-shaped glass 30. In addition to grinding and / or polishing the main surface of the disc-shaped glass 30, the method for manufacturing a glass substrate for magnetic disks may also include processes such as forming chamfered surfaces, grinding and / or polishing the end faces, chemical strengthening, and cleaning. Furthermore, when manufacturing a glass substrate 40 for magnetic disks from a disc-shaped glass 30 that does not have internal holes, the method for manufacturing a glass substrate for magnetic disks may include a process to form internal holes in the disc-shaped glass 30 (for example, the scribing method or the core drilling method described above).

[0079] A method for manufacturing a glass substrate for a magnetic disk is, in one example, carried out as follows: Chamfered surfaces are formed on the inner and outer end faces of an annular disc-shaped glass 30. Next, the main surface of the disc-shaped glass with the chamfered surfaces is ground. In the grinding process, the main surface of the disc-shaped glass is ground using a grinding member in which fixed abrasive grains are formed in a sheet shape, or a slurry containing free abrasive grains. Next, the main surface of the grinding disc-shaped glass is polished. In the polishing process, polishing is carried out using a slurry containing free abrasive grains with a smaller particle size than the free abrasive grains used in the grinding process, and a polishing pad. The polishing process can be divided into multiple processes and carried out using abrasive grains of different particle sizes or polishing pads of different hardness.

[0080] When performing chemical strengthening, it is preferable to perform it, for example, before or after the final polishing step. In the chemical strengthening step, for example, the disc-shaped glass is immersed in a molten solution of a mixed salt of several types of nitrates. In the cleaning step, the disc-shaped glass is washed with a cleaning solution after chemical strengthening or after the final polishing step. Note that additional cleaning steps may be added between each of the above steps as appropriate.

[0081] (Experimental Example 1) To investigate the effects of the present invention, various glass substrates for magnetic disks (Conventional Example 1 and Examples 1-5) shown in the table below were fabricated, and the change in thermal shrinkage rate and roundness when subjected to the first heat treatment described above was evaluated, and the degree of deterioration in flatness when subjected to the second heat treatment described above was evaluated. Regarding the thermal shrinkage rate, the difference in thermal shrinkage rate (C2-C1), the difference in thermal shrinkage amount (S2-S1), and the angle between the direction of maximum and minimum thermal shrinkage rate were also evaluated as an anisotropy evaluation of the thermal shrinkage rate. However, if the difference in thermal shrinkage amount (S2-S1) was 0.5 μm or less, the anisotropy was judged to be extremely small, and therefore the angle was excluded from the measurement.

[0082] [Table 1]

[0083] [Table 2]

[0084] The glass substrates for magnetic disks in Conventional Example 1 and Examples 1-5 all had the following specifications. • Aluminosilicate glass with a glass transition temperature (Tg) of 810°C. • Outer diameter 97 mm, inner diameter 25 mm, plate thickness 0.5 mm, main surface flatness 5 μm. Furthermore, the average retardation value on the main surface of all the glass substrates for magnetic disks fabricated in the examples was 0.5 nm or less. In other words, the residual stress of the glass substrates for magnetic disks fabricated in the examples is sufficiently small, and therefore, it is considered that the effect of residual stress is almost negligible in various evaluations.

[0085] The glass substrates for Conventional Example 1 and Examples 1-5 were prepared in the following manner. (Conventional example 1) Using the overflow downdraw method, a glass sheet was formed while slowly cooling. The ends, which were thicker than the center in the width direction, were cut off, and a predetermined area of ​​the glass sheet was cut out from the remaining portion to obtain a rectangular glass plate material with a short side of 1000 mm and a long side of 1200 mm and a thickness of 0.6 mm. The long side direction of the glass plate material corresponded to the width direction of the glass sheet material. From the obtained glass plate material, individual pieces of glass plate material (fragmented glass) with sides of 109 mm were cut from the central region, excluding the 200 mm end region on the inside of the glass plate material from both the short side and the long side. Subsequently, a disc-shaped glass with a diameter of 99 mm was cut from the aforementioned fragmented glass using the scribing method. At this time, the area ratio of the fragmented glass to the disc-shaped glass was approximately 1.54. Then, using known methods, circular holes were formed, chamfered surfaces were formed, the outer and inner diameters were adjusted, the end faces were polished, the main surface was ground and polished, and cleaning was performed to obtain a glass substrate for magnetic disks with the above specifications. (Example 1) A glass substrate for magnetic disks was obtained in the same manner as in Conventional Example 1, except that the above-described precision annealing was performed on a rectangular glass plate (large plate) with dimensions of 1000 mm on the short side and 1200 mm on the long side, and a plate thickness of 0.6 mm. (Example 2) A glass substrate for magnetic disks was obtained in the same manner as in Conventional Example 1, except that the individualized glass plate material was subjected to precision annealing. (Example 3) A glass substrate for a magnetic disk was obtained in the same manner as in Example 2, except that the size of the individualized glass was 106 mm × 106 mm and the area ratio of the individualized glass to the disc-shaped glass was approximately 1.46. (Example 4) A glass substrate for a magnetic disk was obtained in the same manner as in Example 2, except that the size of the individualized glass was 112 mm × 112 mm and the area ratio of the individualized glass to the disc-shaped glass was approximately 1.63. (Example 5) A glass substrate for a magnetic disk was obtained in the same manner as in Example 4, except that the individual glass fragments were made into rectangles measuring 118.3 mm × 106 mm.

[0086] Precision annealing was performed by placing glass plates in an annealing furnace with the ambient temperature adjusted to 700°C (Tg-110°C) and holding them there for 4 hours. More specifically, the temperature was raised to 700°C over 2.5 hours, held at 700°C for 4 hours, and then cooled at a rate of 50°C / hour. During this process, the glass plates were stacked between two setters, each with a larger main surface area than the glass plate and extending beyond its entire circumference, with the outer edges of the setters extending 5 cm beyond the entire circumference of the glass plate. Rock wool was then lightly packed into the gap between the two setters that were in contact with the sides of the glass plate, so that the entire glass plate was covered with setters and rock wool.

[0087] <Measurement of thermal shrinkage rate and change in roundness> The following first heat treatment was performed on the glass substrate to be measured. (First heat treatment) The glass substrate is placed in an annealing furnace at room temperature and heated to 700°C. The glass substrate is maintained at 700°C for 4 hours, and then cooled from 700°C to 400°C at a rate of 50°C / h.

[0088] For determining the thermal shrinkage rate, the rate was calculated from the change in diameter before and after the first heat treatment in 25 directions, each passing through the center of the glass substrate and spaced at a central angle of 7.2 degrees in the circumferential direction around the center. The maximum value of this calculation was taken as the thermal shrinkage rate of the glass substrate. Furthermore, the difference in thermal shrinkage rates (C2-C1) is the absolute difference between the thermal shrinkage rate C1 in the direction with the minimum thermal shrinkage rate and the thermal shrinkage rate C2 in the direction with the maximum thermal shrinkage rate among the 25 directions mentioned above. And the difference in thermal shrinkage amount (S2-S1) is the absolute difference between the thermal shrinkage amount S1 in the direction with the minimum thermal shrinkage rate and the thermal shrinkage amount S2 in the direction with the maximum thermal shrinkage rate among the 25 directions mentioned above.

[0089] The change in roundness was calculated by subtracting the roundness before the first heat treatment from the roundness after the first heat treatment. All values ​​of the change in roundness were expressed as absolute values. Roundness was measured using a roundness measuring instrument. In principle, the roundness after the first heat treatment was greater than the roundness before the first heat treatment. The degree of deterioration in roundness was evaluated by comparing the difference in roundness of the outer circumference of the glass substrate measured before and after the first heat treatment. A was rated as A if the difference was 0.2 μm or less, B if it was greater than 0.2 μm but 0.5 μm or less, and C if it was greater than 0.5 μm. A and B were evaluated as indicating that the deterioration of roundness had been suppressed.

[0090] <Measurement of change in flatness> The following second heat treatment was performed on the glass substrate to be measured. (Second heat treatment) A glass substrate is placed in a heating device at room temperature and heated to 650°C (corresponding to a Tg of -160°C) in 50 seconds. The glass substrate is then maintained at 650°C for 60 seconds, after which it is removed from the device and allowed to cool naturally to room temperature in the atmosphere. The heating device used is equipped with two panel heaters that are spaced apart and arranged parallel to each other, as described above. The glass substrate is mounted in a holder (substrate holder) and can be positioned vertically within the gap between the panel heaters. Furthermore, the holder (substrate holder) with the glass substrate mounted on it can move between the outside of the heating device (where it is exposed to the atmosphere) and the inside of the heating device.

[0091] The change in flatness was calculated by subtracting the flatness before the second heat treatment from the flatness after the second heat treatment. Both the flatness values ​​and the change in flatness were expressed as absolute values. The degree of deterioration in flatness was determined by the difference in flatness measured before and after the second heat treatment. If the difference was 7 μm or less, it was classified as A; if it was between 7 μm and 10 μm or less, it was classified as B; and if it was greater than 10 μm, it was classified as C. Of these, A and B were evaluated as having successfully suppressed deterioration in flatness.

[0092] A comparison of Conventional Example 1 and Example 1 shows that by performing precision annealing on the large plate before individual piece formation, the thermal shrinkage rate during the first heat treatment becomes 130 ppm or less, and the change in flatness during the second heat treatment can be suppressed to 10 μm or less. Furthermore, the anisotropy of the thermal shrinkage rate is also reduced to 10 ppm or less, which suppresses the deterioration of the roundness of the glass substrate. A comparison of Example 1 and Example 2 shows that precision annealing of the glass plate material after it has been cut into individual pieces is more effective in suppressing the deterioration of the flatness of the glass substrate compared to precision annealing of the large plate material before it has been cut into individual pieces. It also shows that the effect of suppressing the deterioration of the roundness of the glass substrate is also greater. A comparison of Examples 2 to 4 shows that when the area ratio of the individualized glass to the disc-shaped glass is 1.6 or less, preferably 1.5 or less, the thermal shrinkage rate during the first heat treatment is reduced, and the effect of suppressing the change in flatness during the second heat treatment is improved. Furthermore, the anisotropy of the thermal shrinkage rate is also improved, as the difference in thermal shrinkage rates (C2-C1) is reduced, thereby reducing the effect of reducing the change in the roundness of the glass substrate. A comparison of Example 4 and Example 5 shows that, compared to the case where the individualized glass is not square, the square shape of the individualized glass reduces the thermal shrinkage rate during the first heat treatment and improves the effect of suppressing the change in flatness during the second heat treatment. Furthermore, it can be seen that the anisotropy of the thermal shrinkage rate is also improved, as the difference in thermal shrinkage rates (C2-C1) is reduced, which reduces the effect of reducing the change in the roundness of the glass substrate.

[0093] Furthermore, when measuring the thermal shrinkage rate of other glass substrates for magnetic disks manufactured using the conventional method of Example 1, the conditions for the first heat treatment were changed to raise the temperature from room temperature to 600°C at a rate of 100°C / hour, hold at 600°C for 80 minutes, and then cool from 600°C to room temperature at a rate of 100°C / hour. The thermal shrinkage rate was measured at 20-40 ppm, which was considerably smaller than that of Conventional Example 1 when the first heat treatment was performed. This is likely because both the heating temperature and holding time in the thermal shrinkage rate measurement conditions were relaxed. From this, it can be seen that the value of the thermal shrinkage rate is significantly affected by the heat treatment conditions in the measurement conditions.

[0094] (Experimental Example 2) Twenty glass substrates for magnetic disks were manufactured under the conditions of Conventional Example 1, and the thermal shrinkage rate (the maximum value of the thermal shrinkage rate in the 25 directions mentioned above) was measured. The difference (variation) between the maximum and minimum values ​​among the 20 substrates was calculated to be 188 ppm. In the same manner as described above, glass substrates for magnetic disks were prepared under the conditions of Example 1, and the difference (variation) between the maximum and minimum values ​​of the thermal shrinkage rate among 20 substrates was calculated to be 37 ppm. In the same manner as described above, glass substrates for magnetic disks were prepared under the conditions of Example 2, and the difference (variation) between the maximum and minimum values ​​of thermal shrinkage among 20 substrates was calculated to be 9 ppm.

[0095] Furthermore, except that individualized glass pieces immediately before cutting out the disc-shaped glass were used instead of a glass substrate for magnetic disks, the variation in thermal shrinkage rate (difference between maximum and minimum values) of Conventional Example 1, Example 1, and Example 2 was compared in the same manner as described above (Experimental Example 2), and the results were generally the same as those described above (Experimental Example 2). From the above results, it can be seen that (1) precision annealing reduces the variation in thermal shrinkage rate between individual glass fragments, and (2) precision annealing is more effective in reducing the variation in thermal shrinkage rate when performed on glass fragments than when performed on large glass sheets.

[0096] Although the glass substrate for magnetic disks, the disc-shaped glass plate, the glass plate, and the method for manufacturing the glass plate of the present invention have been described in detail above, the present invention is not limited to the above embodiments and examples, and various improvements and modifications may be made without departing from the spirit of the present invention. [Explanation of symbols]

[0097] 10 Glass plates (large glass sheets) 10a Short side 10b Long side 11a,11b End area 12 Central area 13 Measured area 20 Glass plates (fragmented glass) 30 disc-shaped glass 40 Glass substrates for magnetic disks

Claims

1. A rectangular glass plate with a thickness of less than 0.68 mm, The flatness of a 100 mm square area to be measured, cut from the central region of the glass plate excluding the end regions on the inside of the glass plate that are 5 to 20% of the length of the short side of the glass plate from each end in the short side direction of the glass plate, and the end regions on the inside of the glass plate that are 5 to 20% of the length of the long side of the glass plate from each end in the long side direction of the glass plate, is 30 μm or less. The thermal shrinkage rate of the area to be measured is 130 ppm or less when the area to be measured is subjected to a first heat treatment in which the area to be measured is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour. The glass plate is characterized in that the thermal shrinkage rate of the area to be measured is the maximum value of the thermal shrinkage rates measured in 25 directions, each passing through the center of the area to be measured and varying by 7.2 degrees in the circumferential direction, parallel to the main surface of the area to be measured.

2. The glass plate according to claim 1, wherein the glass transition temperature of the glass plate is 750°C or higher.

3. A glass plate according to claim 1, wherein the shorter side length exceeds 900 mm.

4. The glass plate according to claim 1, wherein the glass plate is a portion cut from a long glass sheet formed using one of the float method, the Furcol method, the Pittsburgh method, the downdraw method, the Colburn method, and the redraw method.

5. The glass plate according to any one of claims 1 to 4, wherein the difference between the amount of thermal shrinkage S1 of the region to be measured in the direction in which the thermal shrinkage rate is smallest and the amount of thermal shrinkage S2 of the region to be measured in the direction in which the thermal shrinkage rate is maximum is 1.0 μm or less.

6. A rectangular glass plate having a thickness of less than 0.68 mm, a flatness of 30 μm or less, and the lengths of two orthogonal sides being 95 to 120 mm each, The thermal shrinkage rate after the first heat treatment, in which the material is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour, is 130 ppm or less. The glass plate is characterized in that the thermal shrinkage rate when the first heat treatment is performed is the maximum value of the thermal shrinkage rates measured in 25 directions parallel to the main surface of the glass plate, passing through the center of the glass plate and changing by 7.2 degrees in the circumferential direction.

7. The glass plate according to claim 6, wherein the glass transition temperature of the glass plate is 750°C or higher.

8. The aforementioned rectangular glass plate is a base plate that forms the basis of a disc-shaped glass having a circular outer circumference. The glass plate according to claim 6 or 7, wherein the area of ​​the main surface of the rectangular glass plate is 1.6 times or less the area of ​​the inner surface of the outer circumference of the disc-shaped glass.

9. A disc-shaped glass having a thickness of less than 0.68 mm, a flatness of 30 μm or less, and a circular outer circumference with a diameter of 95 to 100 mm, The thermal shrinkage rate after the first heat treatment, in which the material is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour, is 130 ppm or less. The disc-shaped glass is characterized in that the thermal shrinkage rate when the first heat treatment is performed is the maximum value of the thermal shrinkage rates measured in 25 directions, each passing through the center of the disc-shaped glass and varying by 7.2 degrees in the circumferential direction, parallel to the main surface of the disc-shaped glass.

10. The disc-shaped glass according to claim 9, wherein the glass transition temperature of the disc-shaped glass is 750°C or higher.

11. The disc-shaped glass according to claim 9 or 10, wherein the difference between the amount of thermal shrinkage S1 in the direction with the minimum thermal shrinkage rate and the amount of thermal shrinkage S2 in the direction with the maximum thermal shrinkage rate is 1.0 μm or less among the 25 directions.

12. A glass substrate for magnetic disks having a thickness of less than 0.68 mm, a flatness of 30 μm or less, and a diameter of 95 to 100 mm, The thermal shrinkage rate after the first heat treatment, in which the material is maintained at 700°C for 4 hours and then cooled from 700°C to 400°C at a rate of 50°C / hour, is 130 ppm or less. A glass substrate for magnetic disks, characterized in that the thermal shrinkage rate when the first heat treatment is performed is the maximum value of the thermal shrinkage rates measured in 25 directions, each passing through the center of the glass substrate for magnetic disks in a direction parallel to the main surface of the glass substrate for magnetic disks and varying by 7.2 degrees in the circumferential direction.

13. The glass substrate for magnetic disks according to claim 12, wherein the glass transition temperature of the glass substrate for magnetic disks is 750°C or higher.

14. The glass substrate for a magnetic disk according to claim 12, wherein the difference between the amount of thermal shrinkage S1 in the direction with the minimum thermal shrinkage rate and the amount of thermal shrinkage S2 in the direction with the maximum thermal shrinkage rate is 1.0 μm or less among the 25 directions.

15. A glass substrate for a magnetic disk according to any one of claims 12 to 14, wherein the amount of change in roundness due to the first heat treatment is 0.5 μm or less.

Images