JP4093995B2 - Method for manufacturing magnetic recording medium substrate - Google Patents

Description

  The present invention relates to a magnetic recording medium substrate suitable for a magnetic recording medium substrate, preferably a small-diameter substrate having a diameter of 55 mm or less, more preferably 50 mm or less.

The increase in recording density (surface density) of magnetic recording is very rapid, and during the last 10 years, the rapid increase of 50% to 200% per annum has progressed continuously. A product of 70 Gbit / inch ² is shipped at the mass production level, and a surface recording density of 160 Gbit / inch ² which is more than double that at the laboratory level is reported. The surface recording density at the mass production level is equivalent to 80 Gbytes per platter in a 3.5 "HDD (""" represents an inch), and equivalent to 40 Gbytes per platter in a 2.5 "HDD. This recording capacity is sufficient when a recording medium of one platter is installed in a use application of a normal desktop personal computer (3.5 "HDD installed) or a notebook personal computer (2.5" HDD installed).

Recording density is expected to continue to improve. However, the conventional horizontal magnetic recording method is approaching the limit of thermal fluctuation recording, and when it reaches a recording density of 100 Gbit / inch ² to 200 Gbit / inch ² , it is considered that it will gradually shift to perpendicular magnetic recording. It has been. Although it is not certain at this time which recording limit of perpendicular magnetic recording is present, 1000 Gbit / inch ² (1 Tbit / inch ² ) is considered achievable. If such a high recording density can be achieved, a recording capacity of 600 to 700 Gbytes per 2.5 "HDD platter can be obtained.

  Since there is a high possibility that the large capacity up to this point will not be used up in ordinary PC applications, recording media with a smaller diameter than 2.5 "are gradually being used. "Board, 1.3" HDDs have been released in the past. Currently, HDDs of 2 "or less are very small in quantity. However, if the magnetic recording density is improved in the future, a 1.8" HDD can secure a sufficient recording capacity in a personal computer (especially a notebook personal computer). The recording capacity of a 1 "HDD is currently about 1 to 4 Gbytes, but if the capacity increases several times, not only digital cameras, but also PCs, digital video cameras, information terminals, portable music devices, mobile phones, etc. There is a possibility that it can be used in mobile applications. Small-diameter HDDs of 2 "or less and small-diameter recording media / substrates are promising applications in the future.

  As a substrate for HDD recording media, an Al alloy substrate is mainly used for the 3.5 "substrate, and a glass substrate is mainly used for the 2.5" substrate. In mobile applications such as notebook computers, HDDs are highly susceptible to shocks, and the 2.5 "HDDs mounted on them are highly susceptible to data destruction due to scratching of the recording media and heads due to head hitting. Therefore, a glass substrate having high hardness has been used. Therefore, a glass substrate is likely to be used even in a small-diameter substrate of 2 "or less.

  However, since small-diameter substrates of 2 "or less are mainly used for mobile applications, impact resistance is more important than 2.5" substrates mounted on notebook computers. In addition, in order to reduce the size, it is required to reduce the size and thickness of the entire component including the board. A board thickness thinner than 0.635 mm, which is the standard thickness of a 2.5 "substrate, is required for a substrate of 2" or less. From the specifications required for such a small-diameter substrate, a substrate having a high Young's modulus and sufficient strength even with a thin plate and easy to manufacture is desired. There are several problems with glass substrates in this regard.

  First, in the currently used crystallized glass substrate, the Young's modulus is not sufficient when the plate thickness is 0.635 mm or less, and the resonance frequency during rotation exists in the practical rotation range. Therefore, it is difficult to reduce the thickness further. In addition, a glass plate having a thickness of about 0.8 mm is used. However, it is difficult to make the plate further thinner in production with the glass composition required for the HDD plate. Therefore, it is necessary to adjust the thickness by lapping from a thickness of 0.8 mm to a thickness of 0.5 mm or less. Since the thickness is adjusted, the polishing time becomes considerably long, which causes an increase in processing time and processing cost.

  In addition, since the glass substrate is naturally a non-conductor, there is a problem of charge-up on the substrate in sputter deposition, so that a buffer metal is provided between the substrate and the magnetic film in order to ensure good contact with the magnetic film. It is necessary to put a membrane. Although this technical problem has been basically overcome, it is one of the factors that make it difficult to use a glass substrate in the sputter deposition process. Therefore, if conductivity can be imparted to the substrate, it will not be over, but a glass substrate is difficult.

  As glass substrates are mainly used in 2.5 "HDDs, Al alloy substrates are completely unsuitable for mobile applications. As mentioned above, the substrate hardness is insufficient, but the substrate rigidity is insufficient. Therefore, the only way to make the resonance frequency higher than the practical rotation range is to increase the plate thickness, and it cannot be a candidate substrate for mobile applications.

  Several other alternative substrates such as sapphire glass, SiC substrate, engineering plastic substrate and carbon substrate have been proposed, but small-diameter substrates have been evaluated based on evaluation criteria such as strength, workability, cost, surface smoothness, and film formation affinity. Any of these alternative substrates is insufficient.

  It has been proposed to use a Si single crystal substrate as an HDD recording film substrate (Patent Document 1). The Si single crystal substrate is excellent as an HDD substrate because it is excellent in substrate smoothness, environmental stability, and reliability, and has higher rigidity than a glass substrate. Unlike glass substrates, conductivity is at least a semiconductor property. Further, since ordinary wafers often contain some P-type or N-type dopant, the conductivity is even higher. Therefore, there is no charge-up problem at the time of sputtering film formation as in the case of a glass substrate, and it is possible to directly deposit a metal film on a Si substrate. In addition, since the thermal conductivity is good, the substrate can be easily heated, the film can be formed even at a high temperature of 300 ° C. or higher, and the affinity with the sputter film forming process is very good. Si single crystal substrates are mass-produced for semiconductor ICs with a diameter of 100 mm to 300 mm.

JP-A-6-176339 Japanese Patent Laid-Open No. 10-334461

However, it is difficult to obtain a small diameter wafer having a diameter of 100 mm or less at present. Therefore, it is practical to cut out a desired small-diameter substrate from a 6 "to 8" wafer with a large circulation volume by core extraction. Since the price of the Si single crystal substrate is not inexpensive, it is important to cut out as many HDD substrates as possible from one wafer.
The present invention provides a method for improving the productivity of core cutting.

The present invention provides a magnetic core comprising a core-cutting process in which a single crystal silicon wafer having a diameter of 150 mm to 300 mm is cored to obtain a donut-shaped substrate having an outer diameter of 55 mm or less, a preferable inner diameter of 20 mm or less, and a more preferable inner diameter of 12 mm or less. There is provided a method for manufacturing a recording medium substrate, wherein in the core extraction step, a core for magnetic recording medium is cored so that remaining wafers other than the plurality of extracted substrates are integrated.
According to the present invention, in the coring process, to overlooked before as possible disconnect the core grinding wheel feed rate using a cup grinding machining, a plurality of the remaining wafers other than the substrate withdrawn come together, the remaining wafer after multiple cored The core is cored so that the shortest surface width is 3 to 5 times the wafer thickness, and the remaining wafers other than the multiple substrates to be extracted are integrated using laser processing or water jet processing. Thus, the core is cored so that the shortest width of the remaining wafer surface is at least 1.2 times and not more than 5 times the wafer thickness. In addition, the core removal step is preferably a method in which the core pressure is less likely to be applied to the single crystal silicon substrate than the cup grindstone processing. The core removal step preferably uses laser processing or water jet processing so that the shortest width of the remaining wafer surface after core removal is at an interval of 1.5 to 2.5 times the wafer thickness. To be cored.

  According to the present invention, by leaving the cullet which is the remaining wafer as one body without damaging it, the productivity of the core removal process is improved.

  The present invention relates to a manufacturing method for efficiently coring as many small-diameter substrates as possible from an Si single crystal wafer by coring from an HDD recording film substrate.

FIG. 1 is a schematic process showing an example of manufacturing a magnetic recording medium substrate for HDD using a Si single crystal wafer as an original plate.
After slicing the single crystal silicon rod 1 to obtain a single crystal Si wafer 2 having a diameter of 200 mm, a plurality of donut-shaped substrates 3 having an outer diameter of 55 mm or less are obtained in the core removal step.
The doughnut substrate 3 is preferably chamfered on the inner peripheral end face and the outer peripheral end face, and the end face is polished. Thereafter, a small-diameter substrate is usually manufactured through an alkali etching process, a double-side polishing process, and a cleaning process.
Preferably, before or after the coring step, for example, before the coring step, between the coring step and the chamfering step, between the chamfering step and the end surface polishing step, or more preferably after the end surface polishing step, Before the core removal step, between the chamfering step and the end surface polishing step, or after the end surface polishing step, a lapping step of grinding and removing the surface of the single crystal silicon wafer or the doughnut-shaped substrate preferably by 10 μm to 100 μm may be included.

The single crystal silicon wafer used for the core removal step preferably has a plane orientation (100), a diameter of 150 mm or more and 300 mm or less, and a thickness of 0.4 to 1 mm.
A semiconductor grade Si single crystal wafer is expensive, and even if a 65 mm diameter substrate is manufactured using the single crystal original plate, the cost is several times to nearly 10 times that of a glass substrate. No matter how excellent the characteristics of the Si single crystal substrate are, it is difficult to put it to practical use with such a cost difference.

  In the coring step, for example, as shown in FIG. 2, seven 2.5 "HDD substrates can be cored from an 8" wafer. This method can be achieved by using the technique proposed in Patent Document 2. In this case, by making the machining allowance portion at the time of 2.5 "substrate core removal overlap between adjacent core removal substrates, a maximum of 7 2.5" substrates can be cored from 8 "wafers. However, the remaining remaining portions of the seven cores (hereinafter referred to as “cullet”) are not connected to each other, and are broken apart during processing. I want to increase the number of cores removed from the wafer as much as possible. However, if the cullet is broken during the processing, the work process becomes complicated and troublesome, and the broken cullet becomes a disc. Causing inconveniences such as chipping and damage to the substrate surface.

  First, air adsorption to the lower surface of the substrate is effective for holding the wafer, but the cullet part is small below 2.5 "substrate, and it is difficult to suppress it from being scattered only by air adsorption. If it is removed, there is no problem, but it takes a lot of human power, but it can be automated by a robot or the like, but it is not easy to remove because the unconnected cullet moves during processing.

  Further, if the number of small-diameter substrates to be cored is reduced, the cullets are connected to each other and can be handled as a single unit even after the cores are removed. Accordingly, if at least three locations of the wafer are fixed, the cullet can be held without falling apart, so that the manufacturing process without core can be greatly simplified. However, a decrease in the number of cores is not desirable because it increases the cost of a small-diameter substrate. The cullet process in the core removal process is a big problem on how to make conflicting elements compatible.

In the case of core removal using a conventional cup grindstone, pressure is applied to the original wafer during core removal. Therefore, in order for the cullet to remain intact without being damaged, the minimum cullet width between adjacent cores after core removal is about 5 mm (wafer thickness It was absolutely necessary. Below that, the cullet breaks at a constant rate. Even after the core was devised, it was devised to make it an integral caret and make the shortest width as small as possible. The shortest width refers to the shortest width of the cutlet surface, and is the shortest of the distance between the cores and the distance between the core and the original wafer. When the distance between the cores is shorter than the distance between the core and the original wafer, the shortest width is d1 in FIG. 3 where the distance d1 between the three cores is the same width. is there.
The present inventors have found that by applying a laser processing method or a water jet processing method to the core removal method, the core removal can be completed while keeping the cullet integral even if the shortest width of the cullet is 5 times or less of the wafer.

Laser processing is a method of processing by condensing laser light from an oscillator such as a CO ₂ gas laser, YAG laser, or laser diode.
In the case of thermal core processing such as laser processing, since no pressure is generated on the wafer substrate, it has been found that the shortest width may be 1.5 to 2.5 times the wafer thickness.
However, since the temperature of the laser irradiation part is raised instead of the pressure, if the shortest width is less than one time of the wafer thickness, it may not be able to withstand the heat shock. Further, if the shortest width exceeds 5 times the wafer thickness, the cost increases due to the decrease in the number of cores, which is not preferable. Also in the laser processing method, when the CO ₂ laser is used as the light source, the power density is low although the total power is large, so that heat is easily applied to the cored substrate and cullet, and cracking due to heat shock is likely to occur. A solid-state laser (such as a YAG laser) having a high power density is more desirable because the laser power is used for core removal itself, and heat flows out to surrounding members.

  The water jet processing is a processing method in which high pressure water of 100 MPa or more is mixed with an abrasive such as garnet having an average particle diameter of 20 to 200 μm and irradiated. Water jet processing is advantageous because the processing width is small, no large pressure is applied to the substrate, and there is almost no thermal effect. The width of the shortest part of the cullet may be substantially the same as that of the laser processing method, and is preferably 1 to 5 times the wafer thickness, more preferably 1.5 to 2.5 times the wafer thickness.

  Of course, even with a conventional cup grindstone, it is possible to leave it as an integral cullet even with a minimum width of 5 times or less by devising the processing process. For example, in cup cutting, the shortest part of the cullet is distorted immediately before the core is cut out, and it is easy to break. Therefore, from the point immediately before the cutting (for example, a thickness of about 0.1 mm to 0.2 mm is left), By reducing the grindstone feed rate to less than one half and greatly reducing the cutting pressure, the productivity is significantly reduced, but an integrated cullet processing is possible. However, since it is not pressure-free like the laser processing method, the shortest width cannot be reduced to 2.5 times or less.

  Since the cullet can be left integrally, it is not necessary to add an extra process until the end of processing if the original wafer is fixed at least at three points. In addition, in the case of a small-diameter substrate having a core diameter of 2 "or less, the size of each part of the remaining cullet becomes smaller. Since a method such as air adsorption becomes difficult to use, the cullet integration makes it possible to Simplification is very advantageous.

  The core removal step includes outer diameter core removal (outer peripheral core removal) and inner diameter core removal (inner peripheral core removal). Either the inner diameter core removal or the outer diameter core removal may be performed first.

It may be either before or after the core removal process, but it is preferable to provide a lapping process for grinding and removing the wafer surface preferably by 10 to 100 μm. After the core removal step, for example, a lapping step is also provided between the core removal step and the chamfering step, between the chamfering step and the end surface polishing step, or preferably after the end surface polishing step, and preferably the chamfering step and the end surface polishing step. A lapping step is provided during or after the end surface polishing step.
By the lapping process, warpage and undulation of the wafer original plate or the donut-shaped circular substrate can be suppressed, and the thickness can be adjusted to set an appropriate polishing amount in the subsequent process.

In the manufacture of the HDD substrate shown in FIG. 1, a chamfering process and an end surface polishing process for the inner and outer peripheral end faces may be provided after the core removal process for the original plate such as a wafer.
Chamfer angles and dimensions are generally defined as standard dimensions. Usually, it can be made into a product by a chamfering process. However, abrasive grains and processing debris attached to the end face can cause the substrate strength to decrease and may cause the substrate to break down, so the end face is polished after the chamfering process, and then the strained layer is removed by etching. Is preferred. The end face refers to the inner diameter side face and the outer diameter side face portion of the donut-shaped substrate.

After the end surface polishing step or after the lapping step after the end surface polishing step, preferably, further, a step of alkali etching the substrate, a step of polishing the front and back surfaces of the alkali etched substrate, and a subsequent cleaning step, May be included.
The alkali etching step is performed by immersing in an aqueous solution of 2 to 60% by weight NaOH adjusted to 40 to 60 ° C., for example, in order to remove processing distortion in the lapping step and the end surface polishing step.
The step of polishing the front and back surfaces of the alkali-etched substrate may be performed by a known method. For example, a substrate mounted on a carrier may be sandwiched and rotated between an upper surface plate and a lower surface plate and polished using colloidal silica as abrasive grains.
The cleaning step may be performed by a known method. For example, brush cleaning, alkali and / or acid solution cleaning.

The magnetic recording medium substrate of the present invention can be handled in the same manner as a conventional substrate. For example, a soft magnetic layer and a recording layer can be provided to form a perpendicular magnetic recording medium. In order to improve the adhesion of the soft magnetic layer, an underlayer may be provided prior to the formation of the soft magnetic layer.
A protective layer and a lubricating layer may be provided on the recording layer.

EXAMPLES Hereinafter, although this invention is demonstrated based on an Example, this invention is not limited to this.
The following is an overview of the examples.
Slicing is performed from a large-diameter single crystal silicon rod to form a wafer. Next, lapping is performed using abrasive grains to adjust the thickness and surface of the wafer. Next, a donut-shaped circular substrate is cut out from the wafer by laser light from the YAG laser oscillation device or by cup grindstone processing. Thus, a plurality of substrates are formed. Next, chamfering of the inner peripheral end surface and the outer peripheral end surface of the substrate with a grindstone is performed. Subsequently, the front and back surfaces of the substrate are polished, and a desired substrate is completed. Next, the polishing agent and the like attached to the substrate in the cleaning process are removed to complete the manufacture of the substrate.

Example 1
A large-diameter single crystal silicon ingot was sliced to obtain a wafer having a diameter of 200 mm. Eleven doughnut-shaped circular substrates having a diameter of 48 mm and an inner diameter of 12 mm were obtained with a cup grindstone processing apparatus. At this time, the shortest width d1 between each doughnut-shaped circular substrate was set to 5 times the wafer thickness, and the grinding wheel feed speed was reduced by half before 0.2 mm before the core was pulled out. Remained in one piece without damage. Subsequently, the core was removed, and it took 400 minutes to process five wafers, and 55 substrates were obtained. The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Example 2
Except that the shortest width d1 was set to 3 times the wafer thickness, the same process as in Example 1 was performed, and the cullet as the remaining wafer remained integrally without being damaged. Subsequently, core removal was performed, and it took 440 minutes to process five wafers, and 60 substrates were obtained.
The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Reference example 3
Although the shortest width d1 was set to 8 times the wafer thickness and the grindstone feed rate was kept normal, the same processing as in Example 1 was performed, and the cullet that was the remaining wafer remained integrally without breakage, The number of doughnut-shaped circular substrates obtained was as few as eight. Subsequently, core removal was performed, and it took 370 minutes to process five wafers, and 40 substrates were obtained. The obtained substrate did not show chipping or surface damage due to cullet.

Comparative Example 1
The same processing as in Example 1 was performed except that the shortest width d1 was set to 0.5 times the wafer thickness, and the cullet that was the remaining wafer was damaged and broken apart. The damaged cullet was removed, and then the core was removed. However, it took 560 minutes to process 5 wafers, and 60 substrates were obtained. There were 40 sheets.

  As described above, it can be seen that when the shortest width is increased from 2.5 times to 5 times the wafer thickness in the cup grindstone processing, the cullet remains integrally and the substrate can be obtained efficiently.

Example 4
Using a large-diameter single crystal silicon rod 1, a wafer having a diameter of 200 mm was obtained. Twelve doughnut-shaped circular substrates 3 having a diameter of 48 mm and an inner diameter of 12 mm were obtained with a YAG laser processing apparatus (YAG laser). At this time, the shortest width d1 between the doughnut-shaped circular substrates was made twice the wafer thickness, and the cullet as the remaining wafer remained integrally without being damaged. Subsequently, core removal was performed, and it took 50 minutes to process five wafers, and 60 substrates were obtained.
The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Example 5
A wafer having a diameter of 200 mm was obtained using a large-diameter single crystal silicon rod. Thirty doughnut-shaped circular substrates having a diameter of 26 mm and an inner diameter of 7 mm were obtained with a YAG laser processing apparatus (YAG laser). At this time, the shortest width d1 between the doughnut-shaped circular substrates was set to three times the wafer thickness, and the cullet as the remaining wafer remained integrally without being damaged. Subsequently, core removal was performed, and it took 60 minutes to process five wafers 2, and 150 substrates were obtained.
The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Example 6
The same processing as in Example 5 was performed except that the shortest width d1 was set to 1.2 times the wafer thickness, and the cullet that was the remaining wafer remained integrally without being damaged. Subsequently, core removal was performed, and it took 70 minutes to process five wafers, and 180 substrates were obtained.
The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Comparative Example 2
A wafer having a diameter of 200 mm was obtained using a large-diameter single crystal silicon rod, except that the core was removed with a cup grindstone, and the same treatment as in Example 5 was performed. An attempt was made to obtain 30 doughnut-shaped circular substrates having a diameter of 26 mm and an inner diameter of 7 mm using a cup grindstone processing apparatus. However, the wafers were broken and separated, and no substrate was obtained.

Comparative Example 3
The same process as in Example 5 was performed except that the shortest width d1 was set to 0.5 times the wafer thickness, and the cullet that was the remaining wafer was partially damaged. The damaged cullet was removed, and then the core was removed, but it took 100 minutes to process 5 wafers, and 200 substrates were obtained. However, when the cullet was damaged, the substrate was damaged and could actually be used. It was 140 sheets.

  As described above, it can be seen that when the shortest width is made 1 to 2.5 times the wafer thickness by laser processing, the cullet remains integrally and a substrate can be obtained more efficiently.

Example 7
A wafer having a diameter of 200 mm was obtained using a large-diameter single crystal silicon rod. Eleven doughnut-shaped circular substrates having a diameter of 48 mm and an inner diameter of 12 mm were obtained by a water jet processing apparatus (garnet particle # 220). At this time, the shortest width d1 between the doughnut-shaped circular substrates was set to three times the wafer thickness, and the cullet as the remaining wafer remained integrally without being damaged. Continue,
It took 40 minutes to remove the core and process five wafers, and 55 substrates were obtained.
The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Example 8
The same processing as in Example 7 was performed except that the shortest width d1 was set to 1.2 times the wafer thickness, and the cullet that was the remaining wafer remained integrally without breakage. Subsequently, the core was removed, and it took 45 minutes to process 5 wafers, and 60 substrates were obtained.
The obtained substrate was free from chipping due to cullet and surface damage, and many substrates were obtained in a short time.

Comparative Example 4
The same process as in Example 6 was performed except that the shortest width d1 was set to 0.5 times the wafer thickness, and the cullet that was the remaining wafer was partially damaged. The damaged cullet was removed and then the core was removed. However, it took 60 minutes to process 5 wafers, and 56 substrates were obtained. However, when the cullet was damaged, the substrate was damaged and could actually be used. It was 50 sheets.

  As described above, it can be seen that when the shortest width is made 1 to 2.5 times the wafer thickness by water jet processing, the cullet remains integrally and a substrate can be obtained more efficiently.

6 is a schematic process showing an example of manufacturing a magnetic recording medium substrate for HDD using a Si single crystal wafer as an original plate. A method of core-deforming 65 mm HDD substrates from a 200 mm diameter single crystal Si wafer 2 is shown. It is a figure which shows the shortest width d1 in a core removal process.

Explanation of symbols

1 Single crystal silicon rod 2 Single crystal Si wafer 3 Donut substrate

Claims (3)

A method for manufacturing a substrate for a magnetic recording medium, comprising a core-drawing step of obtaining a plurality of donut-like substrates having an outer diameter of 55 mm or less by core-cutting a single crystal silicon wafer having a diameter of 150 mm to 300 mm. in punching process, to overlooked before as possible disconnect the core grinding wheel feed rate using a cup grinding machining, a plurality of the remaining wafers other than the substrate withdrawn come together, shortest width of said residue wafer surface, the wafer thickness Manufacturing method of the substrate for magnetic recording media cored so that it may become a space | interval of 3 times or more and 5 times or less.   A method for manufacturing a substrate for a magnetic recording medium, comprising a core-drawing step of obtaining a plurality of donut-like substrates having an outer diameter of 55 mm or less by core-cutting a single crystal silicon wafer having a diameter of 150 mm to 300 mm. In the punching process, using laser processing or water jet processing, the remaining wafers other than the plurality of substrates to be extracted are integrated, and the shortest width of the remaining wafer surface is 1.2 times to 5 times the wafer thickness. Manufacturing method of the substrate for magnetic recording media cored so that it may become the space | interval of.   The method for manufacturing a substrate for a recording medium according to claim 2, wherein the core is cored so that the shortest width of the remaining wafer surface is 1.5 to 2.5 times the wafer thickness.

Images