TW201736284A - Method and apparatus for making glass substrate capable of reducing the partial plate-thickness deviation generated by a glass substrate - Google Patents

Abstract

The purpose of the present invention is to provide a method for making a glass substrate with which it is possible to reduce the partial plate-thickness deviation generated by a glass substrate. A method of manufacturing the glass substrate, wherein the molten glass is supplied from the glass supply pipe to the molded body having a supply tank, is to use the molded body to form a flat-plate glass by the overflow pull-down method, and an uniform shape for the bottom surface is provided along the extending direction of the above-described supply tank and the width direction orthogonal to the extending direction of the supply tank where the molten glass is supplied thereto and overflown therefrom. The method for producing the glass substrate includes: a forming step of forming a maximum temperature difference of 30 DEG C or less from the glass supply pipe to the molten glass of the supply tank and the molten glass having a viscosity of 22,000 dPa.s or more and 38,000 dPa.s or less is supplied to the supply tank, and the molten glass is converged at the lower end of the molded body to form the flat glass; and an end-part cooling step in which the partial plate-thickness deviation generated by a glass substrate in the forming step is reduced, so that two ends of the flat glass along the width direction are cooled.

Description

Method for producing glass substrate and device for manufacturing glass substrate

The present invention relates to a method for producing a glass substrate and a device for producing a glass substrate.

In order to manufacture a glass substrate (hereinafter referred to as a "glass substrate for display") for a flat panel display such as a liquid crystal display or a plasma display, an overflow down-draw method is used. The overflow down-draw method includes a step of forming a plate-shaped flat glass under the molded body by overflowing the molten glass from the upper portion of the molded body in the forming furnace, and cooling the flat glass in the slow cooling furnace step. In the slow cooling furnace, the flat glass is introduced into the pair of rolls, and the flat glass is pulled to the lower side by a roll and pulled to a desired thickness, and then the flat glass is slowly cooled. Thereafter, the glass plate is formed by cutting the flat glass into a specific size. The molten glass flowing downward along the side surface of the molded body is caused to contract from the molded body while being contracted in the width direction of the flat glass by the surface tension. Patent Document 1 discloses a method of adjusting a flat plate between a molded body and a tension roller below the molded body in the vicinity of the edge portion in the width direction of the flat glass by using a cooling unit provided separately from the flat glass. The temperature at the edge of the glass suppresses the shrinkage of the flat glass. Thereafter, the flat glass whose contraction is suppressed is formed by passing through the slow cooling space. In the slow cooling space, the ambient temperature is controlled so as to become a desired temperature distribution (temperature distribution in which no strain occurs in the glass sheet), thereby suppressing variations in sheet thickness, warpage, and strain of the glass sheet. [Prior Art Document] [Patent Document] [Patent Document 1] Japanese Patent Laid-Open No. Hei 5-124827

[Problems to be Solved by the Invention] In recent years, the requirements for specifications (quality) have become increasingly strict in glass substrates for liquid crystal display devices. The flatness of the surface of the glass substrate is required to be high, and in order to meet the required specifications, it is particularly necessary to suppress the occurrence of streaks or partial thickness variations of the steep concave or convex streaks. The stripe is formed by the unevenness due to the variation in the thickness (height) of the flat glass in a specific width, and is caused by the surface tension due to the surface tension being contracted in the width direction of the flat glass, and is conveyed on the flat glass. The grain shape is continuously generated in the direction. Accordingly, an object of the present invention is to provide a method for producing a glass substrate and a glass substrate manufacturing apparatus which can suppress variation in thickness of a portion including streaks which are generated on a flat glass. [Technical means for solving the problem] One aspect of the present invention is a method of manufacturing a glass substrate in which molten glass is supplied from a glass supply pipe to a molded body having a supply groove, and the flat body glass is formed by an overflow down-draw method using the above-described molded body. method. In the manufacturing method, the supply tank has a uniform bottom surface shape in which the molten glass supplied to the supply tank overflows from the supply tank in a direction in which the supply groove extends and a width direction orthogonal to the extending direction. The method for producing a glass substrate includes a molding step of a maximum temperature difference of 30° C. or less between the molten glass supplied from the glass supply tube to the supply tank, and a viscosity of the molten glass of 22,000 dPa·s or more and 38,000 dPa·s or less. The molten glass is supplied to the supply tank, and the molten glass is condensed at a lower end of the molded body to form a flat glass; and an end portion cooling step is performed to suppress localized generation in the flat glass obtained by molding in the forming step. In the manner of the thickness deviation of the plate, both ends of the flat glass in the width direction are cooled. In the end portion cooling step, it is preferable that the tension applied when the molded body is not deformed in the width direction of the flat glass and the cross-sectional shape of the flat glass is a target shape is used as a reference tension, and the molded body is not At the time of deformation, the both ends of the flat glass in the width direction are cooled to be the reference tension, and when the molded body is deformed, the flat glass is subjected to deformation according to the molded body. The tension is added by the above reference tension. In this case, it is preferable that the deformation of the molded body is a creep deformation which changes with time with the use of the molded body, and a specific position of the molded body due to the creep deformation is added to the reference tension. The amount of displacement corresponds to the tension. Further, it is preferable that the larger the deformation is, the more the cooling of the both end portions is enhanced. Preferably, the thickness deviation is 10 μm or less. Preferably, in the forming step, the molten glass is heated so that the temperature of the molten glass flowing downward in the formed body is increased by 10 to 150 ° C from the liquidus temperature of the molten glass. Another aspect of the present invention is a manufacturing apparatus for supplying molten glass from a glass supply pipe to a molded body having a supply groove, and forming a glass substrate of the flat glass by the overflow down-draw method using the molded body. The molded body has a supply groove for receiving a supply of molten glass having a maximum temperature difference of 30 ° C or less and a viscosity of 22,000 dPa·s or more and 38,000 dPa·s or less, and a molten groove for forming a molten glass at a lower end of the molded body. The wall of the flat glass. The supply tank has a uniform bottom surface shape in which the molten glass supplied to the supply tank overflows from the supply tank in a direction in which the supply groove extends and a width direction orthogonal to the extending direction. Further, the manufacturing apparatus further includes an end portion cooling device that cools both end portions in the width direction of the flat glass so as to suppress variation in thickness locally generated in the flat glass obtained by molding the molded body. [Effect of the Invention] According to the method for producing a glass substrate and the apparatus for producing a glass substrate of the above aspect, variation in local thickness caused in the sheet glass can be suppressed.

Hereinafter, a method of producing the glass substrate of the present embodiment will be described. (Overall Outline of Manufacturing Method of Glass Substrate) Fig. 1 is a view showing an example of a procedure of a method for producing a glass substrate of the present embodiment. The method for producing a glass substrate mainly includes a dissolution step (ST1), a clarification step (ST2), a homogenization step (ST3), a supply step (ST4), a molding step (ST5), a slow cooling step (ST6), and a cutting step. (ST7). Others may have a grinding step, a grinding step, a washing step, an inspection step, a packing step, and the like. The glass substrate obtained by the production is stacked in the packing step as needed, and then transferred to the orderer. The dissolving step (ST1) is performed by heating a glass raw material to obtain a molten glass. The clarification step (ST2) produces oxygen and CO contained in the molten glass by heating the molten glass. ₂ Or SO ₂ Bubble. The bubble absorbs oxygen generated by a reduction reaction of a clarifying agent (tin oxide or the like) contained in the molten glass, and is floated to the liquid surface of the molten glass to be released. Thereafter, in the clarification step, since the temperature of the molten glass is lowered, the reducing substance obtained by the reduction reaction of the clarifying agent is subjected to an oxidation reaction. Thereby, the gas component such as oxygen in the bubble remaining in the molten glass is again absorbed into the molten glass, and the bubble is broken. The oxidation reaction and the reduction reaction carried out by the clarifying agent are carried out by controlling the temperature of the molten glass. Further, the clarification step may be a vacuum defoaming method in which the bubbles present in the molten glass are grown in a reduced pressure environment and then defoamed. The vacuum defoaming method is effective in not using a clarifying agent. However, the vacuum defoaming method causes the device to be complicated and enlarged. Therefore, it is preferred to use a clarifying method using a clarifying agent to raise the temperature of the molten glass. The homogenization step (ST3) performs homogenization of the glass component by stirring the molten glass using a stirrer. Thereby, the composition unevenness of the glass which is a cause such as a stripe can be reduced. The homogenization step is carried out in a stirred tank as described below. The supplying step (ST4) supplies the agitated molten glass to the forming apparatus. The forming step (ST5) and the slow cooling step (ST6) are carried out in a forming apparatus. The forming step (ST5) forms the molten glass into a flat glass to form a flat glass flow. In the forming process, an overflow down-draw method is employed. The slow cooling step (ST6) is a method in which the flat glass which is formed after the formation is formed to have a desired thickness and does not generate internal strain, and is cooled without causing warpage. In the cutting step (ST7), a plate-shaped glass substrate is obtained by cutting the slow-cooled sheet glass to a specific length. The obtained glass substrate was cut and cut into a specific size to prepare a glass substrate of a target size. Fig. 2 is a schematic view showing a manufacturing apparatus of a glass substrate in which the dissolving step (ST1) to the cutting step (ST8) in the present embodiment are performed. As shown in FIG. 2, the manufacturing apparatus of a glass substrate mainly has the dissolution apparatus 100, the shaping apparatus 200, and the cutting apparatus 300. The dissolution apparatus 100 has a dissolution tank 101, a clarification pipe 102, a stirring tank 103, transfer pipes 104 and 105, and a glass supply pipe 106. A heating mechanism such as a burner (not shown) is provided in the dissolution tank 101 shown in Fig. 2 . A glass raw material to which a clarifying agent has been added is introduced into the dissolution tank, and a dissolution step (ST1) is performed. The molten glass that has been melted in the dissolution tank 101 is supplied to the clarification pipe 102 via the transfer pipe 104. In the clarification pipe 102, the temperature of the molten glass MG is adjusted, and the clarification step (ST2) of the molten glass is performed by the oxidation-reduction reaction of the clarifying agent. Specifically, by heating the molten glass in the clarification pipe 102, the oxygen and CO contained in the molten glass are contained. ₂ Or SO ₂ The bubble absorbs oxygen generated by the reduction reaction of the clarifying agent to grow, floats to the liquid surface of the molten glass, and is released into the gas phase space. Thereafter, by reducing the temperature of the molten glass, the reducing substance obtained by the reduction reaction of the clarifying agent is subjected to an oxidation reaction. Thereby, the gas component such as oxygen in the bubble remaining in the molten glass is again absorbed into the molten glass, and the bubble is broken. The clarified molten glass is supplied to the stirring tank 103 via the transfer pipe 105. In the stirring tank 103, the molten glass is stirred by the stirring bar 103a, and the homogenization process (ST3) is performed. The molten glass which is homogenized in the stirring tank 103 is supplied to the molding apparatus 200 via the glass supply pipe 106 (supply step ST4). In the molding apparatus 200, the sheet glass SG is formed from the molten glass by the overflow down-draw method (forming step ST5), and the slow cooling is performed (slow cooling step ST6). In the cutting device 300, a plate-shaped glass substrate obtained by cutting out the sheet glass SG is formed (cutting step ST7). The supply step S4 controls the temperature of the molten glass flowing in the glass supply tube 106. Specifically, the glass supply pipe 106 is electrically heated to heat the molten glass flowing through the glass supply pipe 106, and the glass supply pipe 106 is surrounded by the refractory material, thereby suppressing the flow in the glass supply pipe 106. The heat dissipation of the molten glass. In the supply step S4, the temperature of the molten glass is controlled such that the temperature of the molten glass flowing through the glass supply pipe 106 gradually decreases from the upstream side toward the downstream side. The glass supply pipe 106 is divided into a plurality of zones, and the temperature of the molten glass is controlled by zones. The electric heating device for heating the glass supply pipe 106 controls the current and voltage flowing into the respective regions of the glass supply pipe 106 so that the temperature of the molten glass changes in accordance with the measurement data of the measuring device. The temperature of the molten glass supplied to the molding apparatus 200 can be appropriately changed by controlling the current and voltage in the glass supply pipe 106. Here, at the end portion on the downstream side of the glass supply pipe 106, the tube temperature and the center temperature of the molten glass are preferably from 1235 ° C to 1265 ° C, more preferably from 1240 ° C to 1260 ° C. (Configuration of the molded body) Next, the configuration of the molded body 1 included in the molding apparatus 200 will be described with reference to Figs. 3 and 4 . Fig. 3 shows an example of a molded article 1 which can be used in the manufacturing method of the present embodiment, and Fig. 4 shows an example of a forming step in the manufacturing method of the present embodiment using the molded article 1 shown in Fig. 3. The molded body 1 includes an upper surface 3 in which a supply tank 2 for supplying molten glass is formed, and a pair of wall surfaces 5 (only one wall surface is illustrated in FIGS. 3 and 4) which are guided from both sides of the supply tank 2 After the overflow, the molten glass flowing downward from between the end portions 3a, 3b of the upper surface 3 in the direction in which the supply groove 2 extends is merged into the flat glass SG at the lower end 4 of the molded body 1, and a pair of guides 6a And 6b are formed at positions of both end portions 5a and 5b in the width direction of the wall surface 5. The guides 6a and 6b are formed to face each other so as to protrude from the wall surface 5 at the positions of the end portions 5a and 5b. The molten glass overflowing from the supply tank 2 flows downward in each of the pair of wall faces 5. The wall surface 5 has a vertical wall surface in which the molten glass overflowing from the supply tank 2 flows downward in the vertical direction, and an inclined wall surface which is connected to the lower end 4 of the molten glass guide forming body 1 which flows downward from the vertical wall and is connected to the vertical wall surface. One of the convection systems of the molten glass flowing downward on the wall surface 5 merges at the lower end 4 of the formed body 1 so as to meet each other. At this time, the width of the molten glass flowing downward along the wall surface 5 is restricted by the guides 6a and 6b, thereby continuously forming, for example, the sheet glass SG having a high thickness uniformity in the width direction. The lower end 4 of the molded body 1 is formed as a linear ridge line in which a pair of wall surfaces 5 are connected to each other (inclined wall surfaces). The reference numeral 2a shown in Figs. 3 and 4 is the bottom surface 2a of the supply tank 2, and the reference numeral 7 shown in Fig. 3 is supplied to the liquid surface 7 of the molten glass of the supply tank 2. Here, the supply groove 2 of the molded body 1 has a shape of a bottom surface 2a which is a shape in which the molten glass supplied to the supply tank 2 overflows from the supply tank 2 in an extending direction of the supply tank 2 (melting glass) The flow direction) and the supply groove 2 orthogonal to the extending direction are uniform in the width direction. The flow rate of the molten glass flowing in the supply tank 2 is based on the viscosity based on the molten glass, the density of the molten glass, the depth from the liquid surface of the molten glass flowing in the supply tank 2 to the bottom surface 2a, and the width of the bottom surface 2a. Calculated. By the equation, the linear density of the flow rate of the molten glass is fixed in the flow direction from the groove start point side to the groove end point side of the glass supply pipe 106, that is, the overflow amount is uniform. The shape of the bottom surface 2a of the supply tank 2 is taken out. Further, the supply groove 2 at the position of the both end portions 3a, 3b of the molded body 1 has overflow from the molten glass to both sides of the supply groove 2 and is the same as the other portions from the positions of the both end portions 3a, 3b of the upper surface 3 The height of the ground surface 2a to the upper surface 3 is uniformly overflowed. When the molten glass overflows from both end portions 3a, 3b of the upper surface 3, the molten glass has a height from the upper surface 3 to the liquid surface of the molten glass. The intersection of the groove curve and the upper surface 3 including the shape of the bottom surface 2a obtained by subtracting the height from the bottom surface 2a to the liquid surface of the molten glass at the time of overflow from the upper surface 3 to the liquid surface of the molten glass becomes the supply tank 2 The end of the trough. Thereby, the distance from the groove start point to the groove end point of the supply groove 2 to which the glass supply pipe 106 is connected is determined, and the shape of the molded body 1 is determined. The cooling roll 8 is a unit which heat-treats both ends of the flat glass SG in the width direction. The cooling roll 8 is disposed on the downstream side of the lower end 4 of the molded body 1. Further, the cooling rolls 8 are disposed on both sides in the thickness direction of the sheet glass SG, and both sides in the width direction of the sheet glass SG. That is, the cooling roll 8 heat-treats the sheet glass SG which leaves the molded object 1 just below the molded object 1. The cooling rolls 8 disposed on both sides in the thickness direction of the sheet glass SG are operated in a paired shape. Therefore, both end portions in the width direction of the sheet glass SG are sandwiched by the pair of cooling rolls 8. The cooling roller 8 is air-cooled by an air cooling pipe leading to the inside. The cooling roller 8 is in contact with the end portion SGa of the sheet glass SG, and the end portion SGa of the sheet glass SG is quenched by heat conduction (end cooling step). The cooling roller 8 has a viscosity of 10 at the end portion SGa of the flat glass SG. ^9.0 In the manner of dPa·s or more, the end portion SGa of the sheet glass SG is quenched. Further, the cooling roller 8 preferably has a viscosity of 10 at the end portion SGa of the flat glass SG. ^9.0 ~10 ^14.5 In the range of dPa·s, the end portion SGa of the sheet glass SG is quenched. In the vicinity of each of the guides 6a and 6b, a heater is disposed so as to extend downward from the upper surface 3 side to the lower end 4 side of the molded body 1, and the molten glass which will flow downward toward the pair of wall faces 5 by the heater The portion in the vicinity of the guides 6a and 6b in the middle and the molten glass flowing downward in the wall surface 5 are heated. This heating is based on the viscosity of the portion near the guide members 6a, 6b in the molten glass flowing downward from the wall surface 5 from the upper surface 3 to the lower end 4 of the molded body 1 (the portion of the molten glass is from the upper surface of the molded body 1) 3, the downward flow reaches the lower end 4), and the liquid phase viscosity (hereinafter, also referred to simply as "liquidus viscosity") of the glass composition constituting the molten glass is not carried out along the guides 6a and 6b. In the formation of the sheet glass SG (and the production of the glass substrate obtained by cooling the sheet glass SG) by the overflow down-draw method using the molded article 1 having the guide member, it is easy to form the flat glass in the vicinity of the guide member. The end of the SG is devitrified. In this case, it is considered that the molten steel is formed into a viscosity suitable for molding at the lower end of the molded body 1 by the forming furnace for accommodating the molded body 1, and is usually set not only by the formation of the flat glass SG. For the purpose and also for the purpose of cooling the molten glass, that is, lower than the temperature of the molten glass, the heat of the molten glass is taken from the guides 6a, 6b, resulting in the temperature of the molten glass near the guides 6a, 6b. It is easy to be lower than the temperature of other portions in the molten glass; and, due to such a temperature drop and physical resistance caused by contact with the guide members 6a, 6b, the downward flow velocity of the molten glass in the vicinity of the guide members 6a, 6b is low. The other portion of the molten glass requires a long time or the like from the contact with the guides 6a, 6b until it leaves the molded body 1. According to Japanese Laid-Open Patent Publication No. 2010-215428, there is a possibility that the devitrification generated at the lower end of the guide can be suppressed. However, in the technique of this document, it is difficult to suppress the devitrification which occurs in the region upstream of the lower end of the guide member, especially in the initial stage of the contact between the molten glass and the guide member, and thus cannot be heated by the lower end of the guide member. One generation of devitrification is eliminated. Further, a glass composition having a liquidus temperature and a low liquid phase viscosity, such as an alkali-free glass and a trace amount of alkali glass, which are suitable for use in a glass substrate for a flat panel display, for example, is used in the production method of the present embodiment. Such a devitrification is particularly likely to occur in the case of a flat glass in which a glass composition having a liquidus viscosity of 80,000 dPa·s or more and 100,000 dPa·s or less and a liquidus temperature of 1200 ° C to 1,220 ° C is formed. In the manufacturing method of the present embodiment, the viscosity of the portion near the guide members 6a, 6b in the molten glass flowing downward on the wall surface 5 of the molded body 1 is maintained from the upper surface 3 to the lower end 4 of the molded body 1 The manner in which the viscosity of the liquid phase is reached (the temperature of the portion becomes higher than the liquidus temperature from the upper surface 3 to the lower end 4 of the molded body 1), and the portion in the molten glass is heated along the guides 6a, 6b. Thereby, a high effect of suppressing devitrification in a portion (the end portion of the molten glass) in the vicinity of the guide for the molten glass is obtained, and the glass composition constituting the molten glass has a density of 80,000 dPa·s or more and 100,000 dPa·s or less. In the case of a small liquid phase viscosity and a liquidus temperature in the range of 1200 ° C to 1220 ° C, the occurrence of devitrification in the end portion can also be suppressed. In the present specification, the liquid phase temperature refers to the equilibrium temperature between the melt and the initial phase of the crystallization, and if the temperature exceeds the temperature, there is no crystallization temperature, and the liquid phase viscosity means that the glass becomes the viscosity of the liquid phase temperature. . Fig. 5 is a view showing a cross section of the glass supply pipe 106 connected to the supply tank 2 of the molded body 1. In the glass supply pipe 106, if the temperature of the molten glass flowing in the central region 106a of the glass supply pipe 106 is compared with the temperature of the molten glass flowing in the peripheral region 106b, the temperature of the molten glass flowing in the central region 106a rises. high. When the molten glass is supplied to the supply tank 2 of the molded body 1 in a state where the temperature difference (viscosity difference) between the central region 106a and the peripheral region 106b is present, the molten glass is heated even in the space in which the molded body 1 is provided. The temperature difference of the molten glass is not improved from the glass supply tube 106 to the upper surface 3 of the molded body 1, but overflows from the upper surface 3 of the molded body 1 toward the lower end 4 in a state where the temperature difference of the residual molten glass is maintained. When there is a temperature difference in the molten glass at the time when the upper surface 3 of the molded body 1 overflows, the flow of the molten glass locally changes (stagnation), so that the molten glass does not uniformly overflow, and the wall surface 5 of the molded body 1 The thickness (amount) of the molten glass flowing downward changes, and the thickness of the sheet glass SG formed at the lower end 4 locally differs. Thereby, in the sheet glass SG, a local thickness variation including streaks occurs. Since both end portions SGa of the flat glass are cooled by the cooling roll 8, and the flat glass SG is applied with tension toward the both end portions SGa, the variation in the thickness of the sheet glass SG is reduced. In order to suppress the variation in the thickness of the sheet glass SG produced by the above-described principle to 10 μm or less, the maximum temperature difference between the molten glass in the glass supply tube 106 when the supply tank 2 of the molded body 1 is supplied, and the molten glass are used. Viscosity becomes more important. In the manufacturing method of the present embodiment, the maximum temperature difference of the molten glass when the glass supply tube 106 is supplied to the molding apparatus 200 (the supply tank 2 of the molded body 1) (the central region 106a of the glass supply tube 106 and the peripheral region 106b) The temperature difference is preferably 30° C. or lower, more preferably 20° C. or lower, and still more preferably 10° C. or lower. Further, the maximum viscosity difference of the molten glass (the difference in viscosity between the central region 106a of the glass supply tube 106 and the peripheral region 106b) is preferably 19,000 dPa·s or less, more preferably 12,500 dPa·s or less, and more preferably The best setting is 6200 dPa·s or less. Since the molten glass supplied to the supply tank 2 is heated in the space in which the molded body 1 is provided, the temperature difference of the molten glass is further smaller than the supply tank 2 from the supply tank 2 to the upper surface 3 of the molded body 1. The temperature difference at the time of supply is, for example, 10 ° C or lower. When the molten glass overflows from the upper surface 3 in such a temperature difference state, the molten glass uniformly overflows, and the thickness (amount) of the molten glass flowing downward on the wall surface 5 is uniform. The molten glass formed by confluence at the lower end 4 is formed into a sheet glass SG. The plate thickness of the sheet glass SG in the lower end 4 is greater than 10 μm, but the end portions SGa of the flat glass are cooled by the cooling roller 8 by applying tension to the both end portions SGa of the flat glass, the flat glass SG The local plate thickness deviation generated in the middle is 10 μm or less. The amount of variation in the thickness of the sheet which can be reduced by cooling the both end portions SGa of the flat glass affects the viscosity of the molten glass. In the production method of the present embodiment, when the molten glass is supplied to the molding apparatus 200 (the supply tank 2 of the molded body 1) via the glass supply pipe 106, the viscosity of the molten glass is preferably 22,000 dPa·s or more and 38,000 dPa·s or less. More preferably, it is 25,000 dPa·s or more and 38,000 dPa·s or less, and more preferably 25,000 dPa·s or more and 35,000 dPa·s or less. When the viscosity of the molten glass supplied to the supply tank 2 of the molded body 1 is lowered, that is, if the temperature of the molten glass is raised, the creep phenomenon of the molded body 1 becomes remarkable, and the time accompanying the start of the self-forming is also generated. After passing through, the central part of the flat glass is drooping and other problems. On the other hand, when the viscosity of the molten glass supplied to the supply tank 2 of the molded body 1 is raised, that is, when the temperature of the molten glass is lowered, the thickness variation of the flat glass tends to occur, and devitrification is likely to occur. Therefore, it is necessary to supply the molten glass which can suppress the creep phenomenon of the molded object 1 to the molded object 1 while preventing the variation in thickness and the occurrence of devitrification. The viscosity of the molten glass when the molding apparatus 200 is supplied is preferably 22,000 dPa·s or more and 38,000 dPa·s or less. The viscosity is determined by the glass composition due to the average viscosity of the molten glass. Hereinafter, the viscosity is referred to as an average viscosity-based viscosity. When the liquid phase viscosity of the glass composition constituting the molten glass is 80,000 dPa·s or more and 100,000 dPa·s or less, the lower end of the molded body 1 in which the viscosity of the molten glass formed by the molded body 1 is the highest is prevented. The viscosity of the molten glass is controlled so that the viscosity of the molten glass becomes less than 80,000 dPa·s. In order to suppress the creep phenomenon of the molded body 1, the viscosity of the molten glass supplied to the supply tank 2 of the molded body 1 is increased, and at the same time as the lower end of the molded body 1, the viscosity of the molten glass is less than 80,000 dPa·s. The molten glass is supplied to the supply tank 2 of the formed body 1. In the manufacturing method of the present embodiment, the viscosity of the molten glass supplied to the supply tank 2 of the molded body 1 (viscosity based on the average temperature) is from 22,000 dPa·s to 25,000 dPa·s, and the upper limit is from 35000 dPa.・s to 38000 dPa·s. Since the viscosity of the molten glass supplied to the supply tank 2 becomes small, the time from the supply of the molten glass to the supply tank 2 to overflow from the upper surface 3 becomes short. Therefore, during this time, the amount of heat received by the molten glass is reduced. The molten glass is heated in the supply tank 2 so that the temperature difference becomes small. However, if the time from the supply to the overflow is short, the temperature difference cannot be eliminated, and the temperature difference may occur. As a result, a portion where the speed of the downward flow locally changes occurs, and this causes a variation in the thickness of the plate. The sheet glass SG formed by converging at the lower end 4 of the molded body 1 by the viscosity (the viscosity based on the average temperature) of the molten glass supplied to the supply tank 2 of the molded body 1 is 22,000 dPa·s to 38000 dPa·s. The low viscosity is easily affected by the cooling effect of the cooling roll 8, and is pulled to the end side in the width direction, so that the effect of suppressing the thickness deviation of the sheet glass SG becomes large. On the other hand, when the molten glass has a low viscosity, the time until the overflow becomes short. Therefore, if there is a temperature difference (viscosity difference) of the molten glass, the thickness difference (viscosity difference) causes a variation in the thickness. Therefore, the maximum temperature difference of the molten glass supplied to the supply tank 2 is 30 degrees C or less. By supplying the molten glass satisfying the two conditions to the supply tank 2, the sheet thickness variation of the sheet glass SG can be 10 μm or less. In order to set the maximum temperature difference of the molten glass supplied to the supply tank 2 to 30 ° C or less, the temperature management of the molten glass flowing through the glass supply pipe 106 is important, and in the longitudinal direction of the tube of the glass supply pipe 106, such as As shown in Fig. 6, the plurality of cells SC1 to SC9 and the plurality of pipe segments PP1 to PP3 are divided to adjust the temperature. Fig. 6 is a graph showing changes in temperature of the molten glass flowing in the glass supply pipe 106 in the longitudinal direction of the tube of the glass supply pipe 106. In Fig. 6, the solid line L1 indicates the change in the "tube temperature" which is the temperature of the molten glass which is in contact with the inner peripheral surface of the glass supply tube 106, that is, the temperature of the glass supply tube 106, and the broken line L2 indicates the glass supply tube. The change in the "central temperature" of the temperature of the molten glass at the center of the profile of 106. In Fig. 6, the chain line L3 represents the average glass temperature obtained by weighted averaging of the mass flow rate of the molten glass per unit sectional area. Using the average temperature, the viscosity based on the average temperature of the molten glass was determined. The temperature change of the molten glass shown in FIG. 6 is described, and the molten glass obtained by homogenization in the molten glass homogenization step ST3 flowing into the glass supply pipe 106 flows into the first pipe section PP1 (zones SC1 to SC5). The difference between the tube temperature of the molten glass and the center temperature is zero. The first pipe section PP1 is used to cool the molten glass to a level not lower than the devitrification temperature of the glass. In the first pipe section PP1, the pipe temperature and the center temperature gradually decrease, and the difference between the pipe temperature and the center temperature gradually increases due to heat dissipation from the glass supply pipe 106. At the boundary between the first pipe section PP1 and the second pipe section PP2, the difference between the pipe temperature and the center temperature is preferably 100 ° C or less. In Fig. 6, in the first tube section PP1, the average glass temperature is lowered from 1470 °C to 1260 °C. In the second pipe section PP2, the drop in the tube temperature is suppressed. The current flowing in the second pipe section PP2 is higher than the current flowing in the first pipe section PP1. Therefore, the amount of heat supplied to the second pipe section PP2 by the energization heating is larger than the amount of heat given to the first pipe section PP1 by the energization heating. Therefore, in the second pipe section PP2, heat dissipation from the glass supply pipe 106 is suppressed, and the temperature of the glass supply pipe 106 is maintained substantially constant. At this time, in the second pipe section PP2, the molten glass from the center of the cross section of the glass supply pipe 106 is transferred toward the molten glass which is in contact with the inner peripheral surface of the glass supply pipe 106, and the center temperature gradually decreases. As a result, in the second pipe section PP2, there is a tendency that the difference between the tube temperature and the center temperature gradually decreases. At the boundary between the second pipe section PP2 and the third pipe section PP3, the difference between the pipe temperature and the center temperature is preferably 50 ° C or less. In Fig. 6, in the second tube section PP2, the average glass temperature is lowered from 1260 ° C to 1250 ° C. Further, in the seventh region SC7 of the second pipe segment PP2, the inner diameter of the glass supply pipe 106 is reduced. Therefore, in the second pipe section PP2, the area of the outer peripheral surface of the glass supply pipe 106 is gradually reduced, so that the heat dissipation of the molten glass via the glass supply pipe 106 is suppressed. That is, in the second pipe section PP2, the difference between the tube temperature and the center temperature is gradually reduced due to the two main factors of the application of the high current and the reduction of the inner diameter. In the third pipe section PP3, the maximum temperature difference between the tube temperature of the molten glass and the center temperature is 30 ° C or less. The refractory material surrounding the third pipe section PP3 is superior to the refractory material 106 surrounding the first pipe section PP1 and the second pipe section PP2. Therefore, the third pipe section PP3 is further suppressed from being radiated by the molten glass of the glass supply pipe 106 as compared with the first pipe section PP1 and the second pipe section PP2. Further, the current flowing in the third pipe section PP3 is lower than the current flowing in the second pipe section PP2, and the amount of heat supplied to the third pipe section PP3 by the energization heating is smaller than the amount of heat given to the second pipe section PP2 by the energization heating. Therefore, the temperature rise of the molten glass flowing in the third pipe section PP3 is suppressed. As a result, in the third pipe section PP3, the difference between the tube temperature and the center temperature is further reduced by the heat transfer in the molten glass in a state where the average glass temperature is substantially constant. In Fig. 6, in the third pipe section PP3, the average glass temperature is maintained at 1,250 °C. Further, a preferred range of the temperature of the molten glass passing through the glass supply tube 106 is as follows. The tube temperature and the center temperature are preferably 1420 ° C to 1470 ° C at the end portion on the upstream side of the glass supply pipe 106. Preferably, the tube temperature is 1210 ° C to 1260 ° C and the center temperature is 1300 ° C to 1350 ° C at the junction of the first tube section PP1 and the second tube section PP2. Preferably, the tube temperature is 1210 ° C to 1260 ° C and the center temperature is 1250 ° C to 1300 ° C at the junction of the second tube section PP2 and the third tube section PP3. The tube temperature and the center temperature are preferably 1235 ° C to 1265 ° C at the end portion of the downstream side of the glass supply pipe 106. The temperature difference of the molten glass by the glass supply pipe 106 is adjusted so that the maximum temperature difference of the molten glass supplied to the supply tank 2 is 30 degrees C or less. Further, since the measurement of the center temperature of the molten glass is difficult to use the thermometer, in this case, information on the amount of heat per unit time and unit area from the glass supply tube 106, and the unit of the glass supply tube 106 can be used. The information on the heating amount of time and unit area, and the temperature and flow rate of the molten glass when flowing into the glass supply pipe 106 are obtained by computer simulation based on the measurement results of the tube temperature of the glass supply pipe 106. Further, the tube temperature is measured by a thermometer (not shown) attached to each position of the glass supply tube 106. The viscosity is measured by a viscometer (not shown) attached to a portion where the glass supply tube 106 is connected to the supply tank 2 of the molded body 1. The viscometer uses, for example, a thin tube viscometer or a rotary viscometer. In the thin tube viscometer, the molten glass to be measured is passed through a thin tube, and the viscosity of the molten glass is measured according to the time (flow rate) of the molten glass passing through the thin tube and the pressure difference between both ends of the thin tube. The rotary viscometer measures the viscosity of the molten glass by reading the resistance which the molten glass receives from the rotating body, that is, the viscous resistance, based on the torque of the rotating body or the like. In the manufacturing method of the present embodiment, it is preferable that the temperature of the portion near the guide members 6a, 6b in the molten glass flowing downward from the wall surface 5 of the molded body 1 becomes the upper surface 3 to the lower end 4 of the molded body 1. The portion is heated so as to be higher than the liquidus temperature by 10 ° C or higher, and it is more preferable to heat the portion so as to be higher than the liquidus temperature by 15 ° C or higher. In such cases, the occurrence of devitrification in the end portion of the formed flat glass is more reliably suppressed. The specific liquidus temperature will vary depending on the composition of the glass composition. In the manufacturing method of the present embodiment, in the forming step, the temperature of the portion near the guide members 6a, 6b in the molten glass flowing downward from the wall surface 5 of the molded body 1 becomes the upper surface 3 of the molded body 1 to The lower end 4 is higher than the liquidus temperature by 10 ° C to 150 ° C (to be higher than the liquidus temperature by 10 ° C or more, and the liquidus temperature is increased by 150 ° C or less), along the guide Heat this part. Thereby, deformation of the molded body 1 and shrinkage in the width direction of the sheet glass SG after molding can be suppressed. Further, it is preferable that the temperature of the portion near the guide members 6a, 6b in the molten glass flowing downward from the wall surface 5 of the molded body 1 becomes higher than the liquidus temperature from the upper surface 3 to the lower end 4 of the molded body 1 The portion is heated along the guide in a manner of °C to 100 °C. The combination of the quenching of the end portion of the molten glass leaving the molded body 1 (quenching of the end portion SGa of the sheet glass SG) causes the sheet thickness variation of the sheet glass SG to be 10 μm or less. Moreover, the occurrence of devitrification in the end portion SGa is further reliably suppressed. Even if the temperature of the portion near the guides 6a, 6b in the molten glass flowing downward from the wall surface 5 of the molded body 1 is sufficiently higher than the liquidus temperature from the upper surface 3 to the lower end 4 of the molded body 1, The temperature of the entire molten glass flowing downward in the molded body 1 is sufficiently higher than the liquidus temperature, and theoretically, devitrification can be suppressed. However, in the case of producing a glass having a higher liquidus temperature, in reality, such an approach cannot be applied in the overflow down-draw method. The reason for this is that there is a viscosity of the molten glass which is suitable for the flat glass forming by the overflow down-draw method (the molten glass in the lower end 4 of the formed body 1 is used to avoid the problem of the relaxation of the flat glass described below or the width shrinkage of the flat glass). The viscosity is preferably 40,000 dPa·s or more, more preferably 70,000 dPa·s or more). When the temperature of the portion in the vicinity of the guide member in the molten glass becomes sufficiently higher than the liquidus temperature, the temperature of the entire molten glass flowing downward in the molded body 1 is sufficiently higher than the liquidus temperature, or When the lower end 4 of the molded body 1 is excessively heated, there is a possibility that the viscosity of the molten glass in the lower end 4 of the molded body 1 becomes smaller than the above-mentioned appropriate range. In this case, the viscosity of the sheet glass after leaving the molded body 1 is not sufficiently increased, and the flat glass falls at a speed higher than the stretching speed of the conveying roller disposed on the downstream side of the molded body 1, resulting in a lower speed. The flat glass is relaxed on the roll or the width of the flat glass is shrunk. Further, as the temperature of the molded body is higher, the creep phenomenon which changes with time with the use of the molded body becomes more remarkable, and the central portion of the flat glass also has a problem of sagging accompanying the passage of time from the start of the forming. Considering the thickness required for the glass substrate and the temperature control of the flat glass performed in the slow cooling step after forming, there is a limit to the increase in the stretching speed of the conveying roller (if the temperature control of the flat glass implemented in the slow cooling step is considered) The conveying speed of the flat glass is preferably from 50 to 500 m/hr, preferably from 100 to 400 m/hr, and preferably from 120 to 300 m/hr. Further, when the temperature of the portion in the vicinity of the guide in the molten glass is sufficiently higher than the liquidus temperature, the width of the formed flat glass is shrunk, and the product width as the glass substrate cannot be secured. In addition, when the temperature of the entire molten glass flowing downward in the molded body 1 is sufficiently higher than the liquidus temperature, the creep phenomenon of the molded body 1 occurs, and if the creep phenomenon becomes remarkable, the manufacturing is performed. The uniformity of the thickness of the glass substrate is lowered. The height of the guide protruding from the wall surface from which the molten glass flows in the molded body 1 is preferably as low as the position below the forming device. Preferably, the lower end 4 of the molded body 1 is a linear ridge line in which the inclined wall faces on both sides are connected to each other, and the height of the pair of guide members in the inclined wall surface is 0 (zero) at the position of the ridge line. Thereby, the end portion (ear portion) of the flat glass can be further inhibited from being opened in a bifurcated manner, so that the glass substrate can be continuously produced more stably. The amount of cooling and the amount of rotation of the cooling roll 8 are controlled by a control device (not shown). The control device mainly includes a computer such as a CPU, a RAM, a ROM, and a hard disk. The control device controls the drive motor that drives the cooling roller 8, and acquires and adjusts the contact load between one of the end portions SGa of the width direction of the holding plate glass SG and the cooling roller 8 and the sheet glass SG. The control device individually controls the amount of cooling of each of the cooling rolls 8. Further, the control device aims to set the cross-sectional shape in the thickness direction of the sheet glass SG to be described below as the target shape, and the tension applied to the sheet glass SG by the cooling control of the cooling roll 8 is at least the transfer unit and the acquisition unit. The four programs that the determination unit and the control unit function are stored and executed. In the conveyance unit, the sheet glass SG formed by the molded body 1 is conveyed to the lower side at a specific conveyance speed in a slow cooling space by using a conveyance roller provided below the molded body 1. The conveyance unit controls the drive motor that drives the conveyance roller to adjust the rotation speed of the conveyance roller, thereby adjusting the conveyance speed of the sheet glass SG. The acquisition unit acquires shape data relating to the current shape of the molded body 1 by obtaining a temporal change in the shape of the molded body 1 by computer simulation. Specifically, the acquisition unit acquires the shape data based on the creep characteristic parameter. The creep characteristic parameter is a parameter for reproducing the relationship between the stress applied to the formed body 1, the temperature of the formed body 1, and the strain rate of the formed body 1 due to creep deformation. Here, the stress applied to the molded body 1 compresses the force of the molded body 1 in the longitudinal direction of the molded body 1 (the extending direction of the supply groove 2). Further, the strain rate of the molded body 1 is assumed to be fixed without changing with time. First, the acquisition unit measures the strain rate of the molded body 1 under the condition that the stress applied to the molded body 1 is fixed, and changes the temperature of the molded body 1 in dependence. Then, the acquisition unit measures the strain rate of the molded body 1 under the condition that the temperature of the molded body 1 is fixed, and changes the stress depending on the stress applied to the molded body 1. Then, the acquisition unit determines a creep characteristic parameter that can reproduce the measured value of the temperature dependence change and the stress dependence change of the strain velocity of the molded body 1. Then, the acquisition unit calculates the strain velocity of the molded body 1 at a specific temperature and stress by computer simulation using the determined creep characteristic parameter, and obtains a time change of the shape of the molded body 1 to obtain a molded body. 1 shape information. Fig. 7 is an example of shape data of the molded body 1 obtained by the acquisition unit. Fig. 7 shows the obtained molded body 1 as viewed in a direction perpendicular to the surface of the sheet glass SG obtained by molding the formed body 1. In Fig. 7, the creep deformation of the formed body 1 is intensified as compared with the actual case. In Fig. 7, the shape of the molded body 1 which is not used, that is, the shape of the molded body 1 before the creep deformation is indicated by a broken line, and the current shape of the molded body 1 after the creep deformation is indicated by a solid line. The acquisition unit acquires at least the displacement amount in the vertical direction of the upper surface 3 of the molded body 1 from the shape data based on the creep deformation of the molded body 1, that is, the amount of displacement of the upper surface. In Fig. 7, the upper surface displacement amount is the dimension in the vertical direction between the upper surface 3 before the creep deformation and the upper surface 3 after the creep deformation. In addition, in FIG. 7, the maximum upper surface displacement amount L which is the maximum value of the surface displacement amount in the longitudinal direction of the molded object 1 is shown. Moreover, the acquisition unit acquires the thickness data of the glass substrate measured by the glass substrate shape measuring device (not shown). The thickness data is, for example, a distribution in the width direction of the thickness of the glass substrate manufactured by the molding apparatus 200. The determination unit determines whether or not the displacement amount L acquired by the acquisition unit has reached the reference amount. Here, the reference amount means a sheet thickness deviation when a flat tension (initial tension) is applied to the sheet glass SG, and the sheet glass SG (glass substrate) is formed into a predetermined thickness (for example, 0.1 mm to 0.8 mm). Can meet the amount of ± 10 μm. When the tension applied to the sheet glass SG is not changed from the initial value, if the displacement amount L exceeds the reference amount, the sheet thickness deviation of the sheet glass SG exceeds ±10 μm. Therefore, the thickness of the sheet glass SG is controlled so that the sheet thickness of the sheet glass SG becomes ±10 μm or less by increasing the tension applied to the sheet glass SG. The reference amount can be arbitrarily changed depending on the initial tension, the predetermined plate thickness of the sheet glass SG, the thickness deviation, and the like, and is, for example, 3 mm to 30 mm. The control unit is configured to apply a tension when the molded body 1 is not displaced in the width direction of the sheet glass SG (formed in the longitudinal direction of the molded body 1) and the cross-sectional shape in the thickness direction of the sheet glass SG is the target shape. The reference tension (initial tension) is controlled by controlling the cooling amount of the cooling roller 8 to cool both end portions SGa in the width direction of the sheet glass SG, whereby the tension applied to the sheet glass SG becomes the reference tension. . In a state in which the molded body 1 is not displaced, the sheet glass SG is formed into a predetermined sheet thickness by applying a reference tension in the width direction of the sheet glass SG, and the sheet thickness variation satisfies ±10 μm. When the tension applied to the sheet glass SG is always the reference tension in the state where the molded body 1 is creep-deformed, the target shape cannot be formed, for example, the sheet thickness cannot be formed into a predetermined shape, and the sheet thickness variation cannot satisfy ±10 μm. . Therefore, the control unit applies not only the reference tension but also the tension corresponding to the displacement of the formed body 1 to the sheet glass SG. Here, the displacement of the molded body 1 is, for example, the upper surface displacement of the molded body 1 in the longitudinal direction. The control unit controls the cooling so that the thickness of the flat glass SG becomes a predetermined thickness based on the shape data of the molded body 1 obtained by the acquisition unit, and the variation in the thickness of the flat glass SG in the width direction is small. The amount of cooling of the roller 8 thereby controls the tension applied to the sheet glass SG. The shape data of the molded body 1 is, for example, a distribution of the surface displacement amount in the longitudinal direction of the molded body 1, that is, a shape distribution. The control unit controls the amount of cooling of the cooling roller 8 so that the amount of displacement of the upper surface 3 is larger as the shape is distributed, and the tension in the width direction of the sheet glass SG is increased. As the displacement amount of the upper surface 3 obtained from the shape distribution, for example, the maximum upper surface displacement amount L is used. After the sheet glass SG formed at the lower end 4 of the molded body 1 is separated from the lower end 4, the central region SGb starts to contract toward the center in the width direction due to its surface tension. Therefore, the cooling roller 8 cools the both end portions SGa of the sheet glass SG to increase the viscosity of the both end portions SGa, and suppresses the contraction of the sheet glass SG in the width direction so that the tension is applied from the central portion SGb toward the both end portions SGa. The thickness of the central portion SGb of the sheet glass SG is uniform. However, when the molded body 1 is subjected to creep deformation, the amount of molten glass in the vicinity of the central portion SGb of the sheet glass SG increases, and the thickness of the central portion SGb changes. That is, the cross-sectional shape of the sheet glass SG in the thickness direction is no longer the target shape. Fig. 8 is a view showing the sheet glass SG in which the thickness in the vicinity of the central portion SGb is increased by the creep deformation of the molded body 1. When the molded body 1 is subjected to creep deformation, the amount of molten glass overflowing from between the end portion 3a of the upper surface 3 and the end portion 3b is increased, so that the thickness of the vicinity of the central portion SGb of the sheet glass SG is increased. In Fig. 8, the thickness in the vicinity of the central region SGb is at most thicker than the predetermined thickness D1, so that the thickness of the central region SGb becomes non-uniform. Therefore, the control unit changes the amount of cooling of the cooling roller 8 in accordance with the shape data of the molded body 1, and suppresses the contraction of the sheet glass SG in the width direction so as to apply tension from the central portion SGb of the sheet glass SG toward the both end portions SGa. The thickness of the central region SGb of the sheet glass SG is made uniform. Fig. 9 is a view showing the relationship between the maximum upper surface displacement amount L of the molded body 1 and the tension T applied to the sheet glass SG. In Fig. 9, the maximum upper surface displacement amount L is referred to as the displacement amount L. When the determination unit determines that the maximum upper surface displacement amount L of the molded body 1 does not exceed L1, the thickness of the central region SGb of the sheet glass SG caused by the creep deformation of the molded body 1 can be ignored. The tension T applied to the sheet glass SG is changed from the initial value T1 (the range of the displacement amount L: 0 or more and less than L1). When the displacement amount L of the molded body 1 is less than L1, the control unit does not change the cooling amount of the cooling roll 8, and maintains the tension T at the initial value T1, whereby the plate thickness deviation of the formed flat glass SG satisfies ±10 μm. . When the determination unit determines that the displacement amount L of the molded body 1 exceeds L1, the control unit controls the tension T corresponding to the maximum upper surface displacement amount L to be applied to the sheet glass SG as shown in FIG. 9 . If the maximum upper surface displacement amount L reaches L1 or more, as shown in Fig. 8, the thickness of the central portion SGb of the sheet glass SG is increased, and the thickness is no longer uniform. Therefore, the control unit controls the tension T=T1+A×maximum upper surface displacement amount L (the range of the displacement amount L: L1 or more and less than Lm, A) corresponding to the displacement amount L in such a manner as to correspond to the displacement amount L. The coefficient is applied to the sheet glass SG from the central portion SGb of the sheet glass SG toward both end portions SGa. The larger the deformation of the control portion molded body 1, the more the cooling of the both end portions SGa is enhanced. Specifically, the amount of cooling of the cooling rolls 8 is increased to increase the viscosity of the both end portions SGa. When the viscosity of the both end portions SGa is increased, the tension T from the central portion SGb toward the both end portions SGa is increased, and the molten glass located in the central portion SGb of the sheet glass SG is stretched toward both end portions SGa, so that the central region SGb The thickness is close to the predetermined thickness of the forming, and the thickness is uniform. The control unit is configured such that the viscosity of both end portions SGa is from, for example, 10 ^9.0 dPa·s increased to 10 ^14.5 From dPa·s, the tension T is increased to control. Further, when the maximum upper surface displacement amount L is in the range of L1 or more and less than Lm, by controlling the tension T to T1 to Tm, the thickness of the central portion SGb is close to a predetermined thickness, so that the thickness is uniform. However, when the displacement amount L exceeds Lm and is displaced, if only the tension T is controlled, it is difficult to make the thickness of the central region SGb close to the predetermined thickness, and the thickness is uniform. Therefore, it is determined by the determination unit. The periodic replacement period of the formed body 1 is reached. Further, due to the creep deformation of the molded body 1, the variation in the thickness of the plate glass SG (the surface unevenness) also changes. The volume shrinkage amount of the sheet glass SG increases as the end portion SGa from the sheet glass SG faces the center region SGb. Therefore, the tensile stress acts in the central portion SGb of the sheet glass SG. Since the thickness in the vicinity of the central region SGb is increased, the tension from the both end portions SGa toward the central region SGb is increased, so that the surface unevenness of the sheet glass SG is increased. Fig. 10 (a) is an enlarged view of a cross section taken along line AA of Fig. 4, and Fig. 10 (b) is an enlarged view of a cross section taken along line BB of Fig. 4. Before the tension T is applied to the sheet glass SG by the cooling roller 8, the sheet glass SG is shrunk toward the central portion SGb. Therefore, the surface unevenness of the sheet glass SG becomes D2, and after the tension T is applied to the sheet glass SG by the cooling roller 8, The surface unevenness of the sheet glass SG becomes D3 which is smaller than D2. When the molded body 1 is subjected to creep deformation, the surface unevenness D2 and D3 of the sheet glass SG also become large. Therefore, by applying the tension T from the central portion SGb toward the both end portions SGa so as to correspond to the maximum upper surface displacement amount L, the sheet glass SG is stretched toward the both end portions SGa, so that the surface of the flat glass SG is uneven. The difference D3 becomes smaller. In order to make the thickness of the central portion SGb close to the predetermined thickness, the surface unevenness D3 of the sheet glass SG is made smaller by applying the tension T in a manner corresponding to the maximum upper surface displacement amount L, so that the central portion of the flat glass SG The thickness of the SGb is uniform. Moreover, the control unit can suppress the presence of the streaks which are likely to occur in the conveyance direction of the sheet glass SG by applying the tension T to the sheet glass SG. The stripe is one of strains due to variations in thickness (height) of the sheet glass SG in a specific width range, and is continuously generated in a grain-like manner in the conveyance direction of the sheet glass SG. Moreover, the main factors of the stripe also include the difference in viscosity of the glass. When the tension is applied in the width direction of the sheet glass SG by the control unit controlling the amount of cooling of the cooling rolls 8, the locally generated streaks of the surface unevenness of the sheet glass SG are stretched to the both end sides of the sheet glass SG. SGa, so that the surface unevenness of the formed surface becomes small and the partial thickness deviation of the plate satisfies ±10 μm. As described above, the tension T in the width direction of the sheet glass SG applied to the sheet glass SG can be changed by the displacement amount due to the creep deformation of the formed body 1 by the lower end 4 of the molded body 1, while The thickness of the central portion SGb is close to the predetermined thickness, and the thickness is uniform. When the central portion of the longitudinal direction of the molded body 1 is sagged downward and is bent by the creep deformation of the molded body 1, the amount of cooling of the cooling roller 8 can be increased to apply the flat glass SG. The tension T in the width direction of the sheet glass SG becomes large, and the sheet thickness deviation in the width direction of the sheet glass SG is lowered. As a result, the variation in the thickness of the glass substrate as the final product can be reduced. Further, in the production step of using a glass substrate having a high liquidus temperature and a glass having a high strain point, the creep deformation of the molded body 1 is particularly likely to be a problem because the temperature of the molded body 1 is likely to increase. In addition, in recent years, as the size of the glass substrate is increased, the dimension of the longitudinal direction of the molded article is continuously increased. Therefore, the deflection of the molded body 1 due to creep deformation tends to be more remarkable. In the present embodiment, by adjusting the amount of cooling of the cooling roll 8, the tension T applied to the sheet glass SG can be changed, and the sheet in the width direction of the sheet glass SG due to the creep deformation of the formed body 1 can be effectively reduced. Thick deviation. According to the production method of the present embodiment, even when the liquid phase temperature of the glass composition constituting the molten glass is high and the liquidus viscosity is small, for example, when the glass composition is an alkali-free glass or a trace amount of alkali glass, Also, the effect of suppressing the devitrification in the end portion of the formed flat glass can be obtained. That is, when the temperature of the liquid phase of the glass composition constituting the molten glass is high and the viscosity of the liquid phase is small, the advantage obtained by the production method of the present embodiment is large. In the production method of the present embodiment, the glass composition constituting the molten glass has a liquidus viscosity of 10,000 dPa·s or less. Such glass compositions have previously been susceptible to devitrification problems in the ends of the flat glass of the overflow down draw process. However, in the production method of the present embodiment, the effect of suppressing devitrification can be obtained. The liquid glass viscosity of the molten glass used in the production method of the present embodiment is 100000 dPa·s or less. In the glass composition having a liquidus viscosity of 100,000 dPa·s or less, the above problem of devitrification becomes more remarkable, but the production method of the present embodiment has an effect of suppressing devitrification. The liquidus viscosity is preferably 80,000 dPa·s or more from the viewpoint of forming the flat glass which can stably perform the overflow down-draw method. The liquidus temperature of the glass composition constituting the molten glass used in the production method of the present embodiment is preferably 1200 ° C or more and 1220 ° C or less. Such glass compositions have previously been susceptible to devitrification problems in the ends of the flat glass of the overflow down draw process. However, the manufacturing method of the present embodiment has an effect of suppressing devitrification. In the production method of the present embodiment, the molten glass may contain zirconia and/or tin oxide. In the molten glass containing zirconia, the liquidus temperature of the glass composition rises as compared with the case where zirconium oxide is not contained. Such a molten glass is prone to devitrification problems in the end portion in the formation of the flat glass of the overflow down-draw method. However, the manufacturing method of the present embodiment has an effect of suppressing devitrification. Zirconium oxide can be eluted in the molten glass by using a dissolution tank and a molding apparatus using a high zirconia refractory material, even when the component of the glass composition is originally contained in the molten glass. In particular, when such a dissolution tank is used to electrolytically dissolve the glass raw material, the concentration of zirconia in the molten glass tends to increase. That is, the production method of the present embodiment is more suitable for the case where the glass raw material is electrolytically dissolved by using a dissolution tank composed of a high zirconia refractory. Further, the dissolution tank composed of the high zirconia refractory material is less likely to be corroded by the glass than the dissolution tank composed of the alumina electroformed refractory material which has been widely used in the past, and has a long service life as a dissolution tank. Moreover, foaming of the molten glass can also be suppressed. Therefore, it is suitable to form the melting temperature (the viscosity of the glass composition reaches 10) ^2.5 The glass composition having a higher temperature is, for example, an alkali-free glass and a molten glass containing a trace amount of alkali glass. Further, when the molten glass formed by the dissolution tank contains an alkali-free glass or a trace amount of alkali glass, the electrical resistivity of the glass composition tends to become high, so that current flows into the high-zirconia refractory material without flowing into the glass raw material. tendency. When a current flows into the refractory, the zirconia is eluted in the molten glass formed by the dissolution tank. In other words, the production method of the present embodiment is further suitable for a case where an alkali-free glass or a molten glass containing a small amount of alkali glass is formed by electrolytic dissolution using a dissolution tank composed of a high zirconia-based refractory. In the glass substrate for an FPD (Flat Panel Display) such as a liquid crystal display or an organic EL display, a glass substrate containing an alkali-free glass or a trace amount of alkali glass is preferable. This is because if the alkali component is eluted from the glass substrate in the panel manufacturing step, the characteristics of the electronic component such as a thin film transistor (TFT) deteriorate. In other words, the production method of the present embodiment is particularly suitable for the case where a glass substrate for a flat panel display is produced by electrolytically dissolving a glass raw material using a dissolution tank composed of a high zirconia refractory material and using the obtained molten glass by an overflow down-draw method. . In addition, the alkali-free glass means a glass composition which does not substantially contain an alkali metal oxide (it is less than 0.05 mass % in terms of content rate). The alkali-containing glass refers to a glass composition containing 0.05 to 2.0% by mass of an alkali metal oxide. In molten glass containing tin oxide, devitrification is likely to occur due to crystallization of tin oxide. Further, in the case of coexistence with zirconia, tin oxide has a function of crystallizing zirconia. Such a molten glass is particularly prone to devitrification problems in the end portion during the process of forming the flat glass by the overflow down-draw method. However, the manufacturing method of the present embodiment has an effect of suppressing devitrification. In the production method of the present embodiment, the glass composition constituting the molten glass may be an alkali-free glass or a trace amount of alkali glass. Compared with an alkali glass containing more than 2.0% by mass of an alkali metal oxide, such an alkali-free glass or a trace amount of alkali glass tends to have a high liquidus temperature and a low liquid phase viscosity, but the production method of the present embodiment is obtained. The effect of suppressing devitrification. This effect is particularly remarkable as described above in the case where a molten glass containing an alkali-free glass or a small amount of alkali glass is formed by electrolytic dissolution using a dissolution tank composed of a high zirconia-based refractory. Further, the alkali-free glass is suitable for a glass substrate for a flat panel display from the viewpoint of preventing deterioration of characteristics of an electronic component such as a TFT (Thin Film Transistor). Among them, the alkali glass containing a trace amount is suitable for a glass substrate for flat panel displays from the viewpoint of solubility and clarification. The solubility and clarity of the glass composition are improved by making a trace amount of alkali glass by intentionally containing a trace amount of an alkali metal oxide. The alkalinity of the glass increases due to the presence of the alkali metal oxide, so that the metal whose valence is changed becomes easy to oxidize and contributes to clarification. Further, even when molten glass is formed by electrolytic dissolution of a glass raw material in a dissolution tank composed of a high zirconia refractory, the electrical resistivity of the glass can be made smaller than that of the alkali-free glass, so that the zirconia direction can be suppressed. The molten glass is eluted to suppress an increase in devitrification of the molten glass. In the production method of the present embodiment, in the case of the glass composition constituting the molten glass, 10 is presented. ^2.5 The temperature of the viscosity of the poise (melting temperature) may also be 1500 ° C to 1750 ° C. Since such a glass composition requires a high temperature at the time of melting, when a molten glass is formed by the dissolution tank which consists of a high zirconia-type refractories, zirconia is melted easily. Even in such a glass composition, the production method of the present embodiment has an effect of suppressing devitrification. Examples of the glass component contained in the glass substrate produced by the production method of the present embodiment include SiO. ₂ Al ₂ O ₃ , B ₂ O ₃ , MgO, CaO, SrO, BaO, Li ₂ O, Na ₂ O, K ₂ O, ZrO ₂ TiO ₂ , ZnO, and P ₂ O ₅ . SiO ₂ It is a skeleton component of glass and is therefore an essential component. When the content is small, the strain point is lowered and the coefficient of thermal expansion tends to increase. Also, if SiO ₂ When the content is too small, it is difficult to reduce the density of the glass substrate. On the other hand, if SiO ₂ When the content is too large, the electrical resistivity of the molten glass MG increases, and the melting temperature remarkably increases, which tends to be difficult to dissolve. If SiO ₂ If the content is too large, the devitrification temperature rises and the devitrification resistance tends to decrease. Further, if SiO ₂ If the content is too much, the etching rate becomes slow. According to this point of view, SiO ₂ The content is preferably in the range of, for example, 60 to 80 mol%. SiO ₂ The content is more preferably 64 to 73 mol% or 65 to 75 mol%, still more preferably 66 to 72 mol%, still more preferably 67 to 71 mol%. Al ₂ O ₃ It is necessary to raise the strain point. If Al ₂ O ₃ If the content is too small, the strain point is lowered. Further, if Al ₂ O ₃ When the content is too small, there is a tendency that the Young's modulus and the etching rate by oxygen are also lowered. On the other hand, if Al ₂ O ₃ When the content is too large, the devitrification temperature of the glass increases, and the devitrification resistance decreases, so that the formability tends to be deteriorated. According to this view, Al ₂ O ₃ The content is in the range of 8 to 20 mol%. Al ₂ O ₃ The content is preferably from 10 to 17 mol%, more preferably from 10.5 to 17 mol%, still more preferably from 11 to 15 mol%, still more preferably from 12 to 15 mol%. B ₂ O ₃ It reduces the high temperature viscosity of the glass and improves the composition of the meltability. That is, since the viscosity in the vicinity of the melting temperature is lowered, the solubility is improved. Also, B ₂ O ₃ It is also a component that lowers the devitrification temperature. If B ₂ O ₃ When the content is small, there is a tendency for solubility and devitrification resistance to decrease. If B ₂ O ₃ If the content is too much, the strain point and Young's modulus decrease. Also, since the glass is formed, B ₂ O ₃ Volatilization causes easy devitrification. In particular, since the glass having a high strain point tends to have a high molding temperature, the above-described volatilization is promoted, and the occurrence of devitrification becomes a significant problem. Also, when the glass dissolves, B ₂ O ₃ When it is volatilized, the heterogeneity of the glass becomes remarkable, and it becomes easy to generate streaks. According to this point of view, B ₂ O ₃ The content is 0 to 15 mol%, preferably 0 to 8 mol%, more preferably 0 to 7 mol%, still more preferably 0.1 to 6 mol%, still more preferably 1 to 5 mol%, still more preferably 1.5 to 4.5 mol%. The scope. MgO is a component that enhances solubility. Further, among the MgO-based alkaline earth metals, a component which is difficult to increase in density is required. Therefore, when the content is relatively increased, it is easy to achieve a low density. The resistivity and the melting temperature of the molten glass MG can be lowered by containing MgO. Among them, when the content of MgO is too large, the devitrification temperature of the glass sharply rises, so that devitrification is likely to occur particularly in the molding step. From such a viewpoint, the MgO content is 0 to 15 mol%, preferably 1 to 15 mol%, more preferably 0 to 6 mol%, still more preferably 1 to 6 mol%. Alternatively, the MgO content is preferably from 0 to 15 mol%, more preferably from 0 to 6 mol%, still more preferably from 1 to 6 mol%. The CaO system is a component which is effective in not increasing the devitrification temperature of the glass and increasing the solubility of the glass. Further, among the CaO-based alkaline earth metal oxides, a component which is difficult to increase in density is required. Therefore, when the content is relatively increased, it is easy to achieve a low density. When the content is too small, the electrical resistivity of the molten glass MG tends to increase and the devitrification resistance tends to decrease. If the CaO content is too large, the coefficient of thermal expansion increases and the density tends to increase. From such a viewpoint, the CaO content is 0 to 20 mol%, preferably 1 to 15 mol%, more preferably 2 to 11 mol%, still more preferably 4 to 9 mol%. The SrO system is capable of lowering the composition of the devitrification temperature of the glass. Although SrO is not essential, if it contains SrO, the devitrification resistance and solubility are improved. However, if the SrO content is too large, the density is increased. From such a viewpoint, the SrO content is 0 to 15 mol%, preferably 0 to 8 mol%, more preferably 0 to 3 mol%, still more preferably 0 to 1 mol%, still more preferably 0 to 0.5 mol%, Further preferably, it is substantially not contained. BaO is an essential component capable of effectively lowering the devitrification temperature of the glass and the electrical resistivity of the molten glass MG. When BaO is contained, the devitrification resistance and solubility are improved. However, if the content of BaO is too large, the density is increased. Further, the BaO content is 0 to 15 mol% or 0.1 to 15 mol%, preferably 1 to 15 mol%, more preferably 1 to 10 mol%, more preferably from the viewpoint of environmental load and a tendency to increase the coefficient of thermal expansion. It is in the range of 1.5 to 6 mol%. Li ₂ O and Na ₂ The O system has a component which increases the thermal expansion coefficient of the glass and causes damage to the substrate during heat treatment. Also, Li ₂ O and Na ₂ O is also a component that lowers the strain point. On the other hand, since the resistivity of the molten glass MG can be lowered, Li can be contained. ₂ O and Na ₂ O inhibits the dissolution tank from being eroded. According to the above point of view, Li ₂ The content of O is preferably from 0 to 0.5 mol%, more preferably substantially not contained. Na ₂ The content of O is preferably from 0 to 0.5 mol%, more preferably from 0 to 0.2 mol%. Furthermore, Na ₂ O system and Li ₂ O is preferably Na compared to a component which makes the strain point more difficult to reduce. ₂ O>Li ₂ O. Furthermore, Li is protected from the viewpoint of preventing deterioration of TFT characteristics due to elution from the glass substrate. ₂ O and Na ₂ O is preferably substantially not contained. K ₂ The O system enhances the alkalinity of the glass to promote the clarifying component. Also, K ₂ O is a component which lowers the specific resistance of the molten glass MG. If containing K ₂ O, since the electrical resistivity of the molten glass MG is lowered, it is possible to prevent current from flowing into the refractory material constituting the dissolution tank, thereby suppressing erosion of the dissolution tank. Further, when the refractory material constituting the dissolution tank contains zirconia, the dissolution tank can be prevented from being eroded, and zirconia is eluted from the dissolution tank to the molten glass MG, so that devitrification caused by zirconia can also be suppressed. Further, since the viscosity of the glass in the vicinity of the dissolution temperature can be lowered, the solubility and clarity are improved. On the other hand, if K ₂ When the content of O is too large, there is a tendency that the coefficient of thermal expansion increases and the strain point decreases. According to this point of view, K ₂ The O content is preferably from 0 to 0.8 mol%, more preferably from 0.01 to 0.5 mol%, still more preferably from 0.1 to 0.3 mol%. ZrO ₂ And TiO ₂ A component that raises the strain point of the glass. However, if ZrO ₂ Amount and TiO ₂ If the amount is too large, the devitrification temperature rises remarkably, so that the devitrification resistance tends to decrease. Especially ZrO ₂ The problem of refractory melting due to the high melting point causes a part of the raw material to be deposited on the bottom of the dissolution tank. When these undissolved components are mixed into the glass body, the quality of the glass is deteriorated as an inclusion. Also, TiO ₂ Since the glass is colored, the substrate for the display is less preferable. According to such a viewpoint, in the glass substrate of the present embodiment, ZrO ₂ And TiO ₂ The content is preferably from 0 to 5 mol%, more preferably from 0 to 2 mol%, even more preferably substantially not contained. ZnO is a component that enhances solubility. But not a necessary ingredient. When the ZnO content is too large, the devitrification temperature rises, the strain point decreases, and the density tends to increase. From such a viewpoint, the ZnO content is preferably from 0 to 5 mol%, more preferably from 0 to 2 mol%, and further preferably substantially not contained. P ₂ O ₅ It is a component that lowers the viscosity of the high temperature and improves the solubility. But not a necessary ingredient. If P ₂ O ₅ If the content is too much, the strain point is lowered. Also, when the glass dissolves, P ₂ O ₅ The volatilization causes the non-homogeneity of the glass to become remarkable, and streaks are easily generated. According to this point of view, P ₂ O ₅ The content is preferably 0 to 3 mol%, more preferably 0 to 1 mol%, still more preferably 0 to 0.5 mol%, still more preferably substantially no content. The glass substrate to which the present embodiment is applied contains, for example, an alkali-free glass having the following composition. SiO ₂ :55-80% by mass Al ₂ O ₃ : 8-20% by mass B ₂ O ₃ : 0-18 mass% RO 0 to 17 mol % (RO is the total amount of MgO, CaO, SrO and BaO) R' ₂ O 0~2 mol% (R' ₂ O is Li ₂ O, Na ₂ O and K ₂ The total amount of O). From the viewpoint of reducing the heat shrinkage rate, SiO is preferred. ₂ It is 60 to 75% by mass, and further 63 to 72% by mass. In the RO, MgO is preferably 0 to 10% by mass, CaO is 0 to 10% by mass, SrO is 0 to 10% by mass, and BaO is 0 to 10% by mass. Also, it may also contain at least SiO ₂ Al ₂ O ₃ , B ₂ O ₃ And RO and Moer than (2 × SiO ₂ )+Al ₂ O ₃ )/((2×B ₂ O ₃ ) +RO) is a glass of 4.5 or more. Further, it is preferable to contain at least one of MgO, CaO, SrO, and BaO, and the molar ratio (BaO+SrO)/RO is 0.1 or more. Also, preferably expressed in mass% B ₂ O ₃ The total content of the RO and the content of RO represented by the mass % are 30% by mass or less, preferably 10 to 30% by mass. Further, it is preferable that the oxide of the metal (tin oxide, iron oxide) whose valence varies in the molten glass is 0.05 to 1.5% by mass in total. Preferably, substantially no AS is included ₂ O ₃ , Sb ₂ O ₃ , PbO, but may also contain such arbitrarily. Further, the metal oxide (tin oxide, iron oxide) having a valence in the glass is contained in an amount of 0.05 to 1.5% by mass in total, and substantially does not contain As. ₂ O ₃ , Sb ₂ O ₃ And the case of PbO is arbitrary and not necessary. The glass substrate produced in the present embodiment is suitable for a glass substrate for a display including a glass substrate for a flat panel display. In the glass substrate produced in the present embodiment, a glass substrate for an oxide semiconductor display having an oxide semiconductor such as IGZO (indium, gallium, zinc, or oxygen) and a glass substrate for an LTPS display using an LTPS (low temperature polysilicon) semiconductor are suitably used. Further, the glass substrate produced in the present embodiment is preferably a glass substrate for a liquid crystal display which requires a very small content of an alkali metal oxide. Moreover, it is also suitable for a glass substrate for an organic EL display. In other words, the method for producing a glass substrate of the present embodiment is suitable for producing a glass substrate for a display, and is particularly suitable for producing a glass substrate for a liquid crystal display. Others can also be used as a cover glass for a mobile terminal device or the like, a cover glass for a housing, a touch panel, a glass substrate for a solar cell, or a cover glass. The glass substrate produced in the present embodiment is particularly preferably a glass substrate for a liquid crystal display using a polycrystalline germanium TFT. Further, the glass substrate produced in the present embodiment can also be applied to a cover glass, a glass for a disk, a glass substrate for a solar cell, or the like. The glass substrate manufacturing method and the glass substrate manufacturing apparatus of the present embodiment have been described in detail above. However, the present invention is not limited to the above-described embodiments, and it is needless to say that various modifications can be made without departing from the scope of the invention. Or change. (Example) A glass frit prepared by a composition having the following composition was electrolytically dissolved by a dissolution tank using a high zirconia refractory to form a molten glass. Then, the formed molten glass was clarified by a clarification tube made of a platinum alloy, and then stirred by a stirring tank. Then, the molten glass is supplied to the molding apparatus 200 (molded body 1), and the flat glass is formed by the overflow down-draw method. The end of the flat glass is made up of 10 by the chill roll 8 and the viscosity of the end is 10 ^12.5 The flat glass formed by cooling in the form of dPa·s was subjected to slow cooling and then cut to obtain a glass substrate for a flat panel display having a thickness of 0.4 mm and a size of 2200 mm × 2500 mm. Further, the glass composition had a liquidus viscosity of 50,000 dPa·s and a strain point of 715 ° C. SiO ₂ : 61.5 mass%, Al ₂ O ₃ : 20% by mass, B ₂ O ₃ : 8.4% by mass, CaO: 10% by mass, SnO ₂ : 0.1% by mass. The maximum temperature difference of the molten glass supplied from the glass supply pipe 106 to the supply tank 2 of the molded body 1 and the viscosity of the molten glass (viscosity based on the average temperature) were changed, and the variation in thickness of the flat glass (glass substrate) was measured. The results are shown in Table 1. [Table 1] As shown in Table 1, in the examples 1 to 6 in which the maximum temperature difference of the molten glass was 30 ° C or less and the viscosity of the molten glass (viscosity based on the average temperature) was 22,000 dPa·s or more and 38,000 dPa·s or less, the thickness was The deviation is less than 10 μm, so that the thickness deviation can be suppressed. On the other hand, when the maximum temperature difference of the molten glass exceeds 30 ° C, the viscosity of the molten glass (the viscosity based on the average temperature) is less than 22,000 dPa·s, and the viscosity of the molten glass exceeds 38,000 dPa·s. In the case of Comparative Examples 1 to 7, the sheet thickness deviation became larger than 10 μm. In this way, it is confirmed that the variation in the thickness of the flat glass is 10 μm or less, and the maximum temperature difference of the molten glass supplied to the supply tank 2 of the molded body 1 is 30° C. or less, and the viscosity of the molten glass is based on The viscosity of the average temperature is set to 22,000 dPa·s or more and 38,000 dPa·s or less.

1‧‧‧Formed body
2‧‧‧ supply slot
2a‧‧‧ bottom
3‧‧‧ upper surface
3a, 3b‧‧‧ (top surface) end
4‧‧‧Bottom
5‧‧‧ wall
5a, 5b‧‧‧ end
6a, 6b‧‧‧ Guides
7‧‧‧ liquid level
8‧‧‧Cooling roller
100‧‧‧ Dissolving device
101‧‧‧Dissolution tank
102‧‧‧clarification tube
103‧‧‧Stirring tank
103a‧‧‧ stir bar
104, 105‧‧‧Transport tube
106‧‧‧Glass supply tube
106a‧‧‧Central area
106b‧‧‧ surrounding area
200‧‧‧Forming device
300‧‧‧cutting device
D1‧‧‧ thickening
D2, D3‧‧‧ surface glass SG surface unevenness
L (L1 ~ Lm) ‧ ‧ displacement
MG‧‧‧ molten glass
PP1～PP3‧‧‧Multiple pipe sections
SC1～SC9‧‧‧Multiple districts
SG‧‧ ‧ flat glass
SGa‧‧‧ (flat glass) end
Central area of SGb‧‧‧ (flat glass)
T (T1 ~ Tm) ‧ ‧ tension

Fig. 1 is a view showing the flow of the manufacturing method of the embodiment. 2 is a schematic view showing a manufacturing apparatus of a glass substrate. Fig. 3 is a perspective view showing an example of a molded body which can be used in the production method of the embodiment. Fig. 4 is a view showing an example of a manufacturing method of the present invention using the apparatus shown in Fig. 3. Fig. 5 is a view showing a cross section of a glass supply pipe connected to a supply groove of a molded body. Fig. 6 is a graph showing the temperature change of the molten glass flowing in the glass supply tube in the longitudinal direction of the glass supply tube used in the present embodiment. Fig. 7 is a view showing an example of a shape change of a molded body obtained by an acquisition unit. Fig. 8 is a view showing an example of a cross section of a glass ribbon obtained by molding a molded body which has been subjected to creep deformation. Fig. 9 is a view showing the relationship between the displacement amount of the molded body and the tension T applied to the glass ribbon. Fig. 10(a) is an enlarged view of an example of a section of a plate glass along the line AA shown in Fig. 4, and (b) is an enlarged view of a section of a plate glass along the line BB shown in Fig. 4. The resulting map.

1‧‧‧Formed body

2‧‧‧ supply slot

2a‧‧‧ bottom

3‧‧‧ upper surface

4‧‧‧Bottom

5‧‧‧ wall

6a, 6b‧‧‧ Guides

7‧‧‧ liquid level

8‧‧‧Cooling roller

106‧‧‧Glass supply tube

SG‧‧ ‧ flat glass

SGa‧‧‧ (flat glass) end

Claims (7)

A method for producing a glass substrate, which is characterized in that molten glass is supplied from a glass supply tube to a molded body having a supply groove, and the flat glass is formed by an overflow down-draw method using the molded body, and the supply tank is supplied to the molten glass. The molten glass overflowing from the supply tank is formed in a direction in which the supply groove extends and a uniform bottom surface shape in a width direction orthogonal to the extending direction, and the method for producing the glass substrate includes a molding step. The molten glass having a maximum temperature difference of 30° C. or less from the molten glass supplied from the glass supply pipe to the supply tank and having a viscosity of 22000 dPa·s or more and 38,000 dPa·s or less of the molten glass is supplied to the supply tank, and a flat glass is formed by converging the molten glass at a lower end of the molded body, and an end cooling step of suppressing a thickness variation locally generated in the flat glass obtained by molding in the forming step. Both ends in the width direction are cooled. The method of producing a glass substrate according to claim 1, wherein the end portion cooling step is a reference tension when a cross-sectional shape of the flat glass applied to the flat glass in the width direction of the flat glass is a target shape. When the molded body is not deformed, it is controlled such that the both ends of the flat glass in the width direction are cooled to be the reference tension, and when the molded body is deformed, the flat glass is applied according to the above. The deformation of the molded body and the tension added to the above reference tension. The method for producing a glass substrate according to claim 2, wherein the deformation of the molded body is a creep deformation which changes with time with the use of the molded body, and the forming of the reference tension is caused by the creep deformation. The amount of displacement of a particular position of the body corresponds to the tension. The method for producing a glass substrate according to claim 2 or 3, wherein the larger the deformation, the more the cooling of the both end portions is enhanced. The method for producing a glass substrate according to any one of claims 1 to 4, wherein the thickness deviation is 10 μm or less. The method for producing a glass substrate according to any one of claims 1 to 5, wherein, in the forming step, the temperature of the molten glass flowing downward from the molded body is increased by 10 ° C from the liquidus temperature of the molten glass. The above molten glass was heated at 150 °C. A glass substrate manufacturing apparatus characterized in that molten glass is supplied from a glass supply pipe to a molded body having a supply groove, and the flat glass is formed by an overflow down-draw method using the molded body, and the molded body has a maximum receiving amount. a supply groove for supplying molten glass having a temperature difference of 30 ° C or less and a viscosity of 22,000 dPa·s or more and 38,000 dPa·s or less, and a wall surface for forming the flat glass by converging the molten glass at the lower end of the molded body, the supply The groove has a uniform bottom surface shape in which the molten glass supplied to the supply tank overflows from the supply tank in a direction in which the supply groove extends and a width direction orthogonal to the extending direction, and the glass substrate manufacturing apparatus further includes An end portion cooling device that cools both end portions in the width direction of the flat glass so as to suppress variation in thickness locally generated in the flat glass obtained by molding the molded body is suppressed.