US20100291844A1

US20100291844A1 - Dresser for abrasive cloth

Info

Publication number: US20100291844A1
Application number: US12/812,643
Authority: US
Inventors: Hiroaki Sakamoto; Toshiya Kinoshita
Original assignee: Nippon Steel Materials Co Ltd
Current assignee: Nippon Steel Chemical and Materials Co Ltd
Priority date: 2008-02-20
Filing date: 2008-10-28
Publication date: 2010-11-18
Also published as: JP2009196025A; WO2009104224A1; TW200936317A; TWI455794B; JP5255860B2; MY153268A

Abstract

The invention provides a dresser for an abrasive cloth, which has a smaller abrasive grain diameter than the conventional dresser and has an abrasive grain spacing regulated in a predetermined range depending upon the abrasive grain diameter to simultaneously meet a high level of pad grinding power and a pad flatness and, at the same time, is less likely to cause dropout of abrasive grains. The dresser comprises a plurality of abrasive grains fixed as a single layer on a surface of a metallic support material. The dresser is characterized in that the metallic support material on its surface on which the abrasive grains are fixed has a convex shape, the difference in height between the end part and the central port on the surface is not less than 3 μm and not more than 40 μm, and the center-to-center spacing between at least one set of adjacent grains is d≦L<2d wherein d represents the diameter of the abrasive grains; L represents the center-to-center spacing between adjacent abrasive grains. The diameter (d) of the abrasive grains is preferably 3 μm≦d<100 μm.

Description

TECHNICAL FIELD

The present invention relates to a dresser used for planarization of an abrasive cloth and for removing clogging or debris from the abrasive cloth by a chemical mechanical planarization (hereinafter “CMP”) process.

BACKGROUND ART

Semiconductor wafer polishing machines, machines for planarizing the surface of the interconnection or insulating layer during manufacture of integrated circuits, and machines for planarizing the surface of Al or glass magnetic hard disk substrates use a CMP process. CMP is the process whereby the surface to be polished is planarized by pressing a rotating head equipped with a resin (e.g., urethane) polishing pad against the surface in the presence of a slurry containing fine abrasives. Naturally, the performance of the polishing pad decreases with the duration of use. To avoid this, dressing is regularly carried out in which the pad surface is ground to restore a flat surface. A tool used for dressing is called a “dresser.” The dresser is prepared by bonding abrasive grits to a metal substrate by electroplating, brazing, etc.
In CMP processes it has been required to prevent the generation of scratches on the surface being polished. Moreover, with a recent tendency where the depth of field of photolithography equipment is lowered for much smaller line/space widths in the integrated circuitry or where the magnetic hard disk storage capacity is increased, the CMP process is encountering a stricter planarity requirement characterized by, for example, reduced waviness across the polished surface. To meet these requirements, high pad planarity needs to be maintained by uniformly grinding the pad surface. Further, the dressing process requires pad grinding power that enables removal of pad clogging or debris from the polishing operation.
There are disclosed dressers directed to uniform pad grinding. Patent Document 1 discloses a dresser in which ultrafine abrasive grits of 5-1,000 μm diameter are bonded to the core disc at the intersections of the lines of an imaginary mesh consisting of equilateral triangles with each side 0.2-10 mm long. Patent Document 2 discloses a dresser in which abrasive grits are substantially concentrically arranged at substantially regularly spaced intervals. More specifically, Patent Document 2 discloses a dresser which includes diamond grits of about 1 mm diameter arranged at center-to-center spacing of 3 mm. Patent Document 3 discloses a dresser which aims to reduce the generation of scratches on the substrate by preventing pull-out of diamond grit particles. More specifically, Patent Document 3 discloses a dresser in which an abrasive grit layer is electroplated onto the surface of a disc-shaped substrate such that the center-to-center spacings between nearby abrasive grits are each 2-10 times as large as the average grit diameter.
Patent Document 4 discloses a dresser in which deformation of the metallic support by the heat of brazing is suppressed. This is achieved by reducing variations in melting temperature of a brazing metal to lower brazing process temperature. Also, Patent Document 5 discloses a dresser in which deformation of the metallic support by the heat of brazing is suppressed. This is achieved by employing a brazing metal with a specific composition that can reduce variations in melting temperature thus allowing for lower brazing process temperature. Patent Documents 4 and 5 remain silent with respect to the surface shape of the metallic support.
[Patent Document 1] Japanese Patent Application Laid-Open No. 2006-55944
[Patent Document 2] Japanese Patent Application Laid-Open No. 2000-141204
[Patent Document 3] Japanese Patent Application Laid-Open No. 2001-121418
[Patent Document 4] Japanese Patent Application Laid-Open No. 2006-305659
[Patent Document 5] Japanese Patent Application Laid-Open No. 2007-83352

DISCLOSURE OF INVENTION

Problems to be Solved by the Invention

As described above, dressers which uniformly grind the polishing pad surface have been disclosed. However, as there is a trade-off between pad grinding power and pad planarity, it has been difficult in the art to achieve good pad grinding power and good pad planarity at the same time. Specifically, for example, in order to increase yields of IC substrate production or Al or glass magnetic hard disk substrate production, the polishing pad needs to be regularly dressed to restore a flat surface. This however requires a certain level of pad grinding rate, and the high pad grinding rate inevitably leads to poor pad planarity. In particular, magnetic hard disk substrates require a stricter planarity requirement than IC substrates. Thus, the conventional dressers are encountering the problem of low productivity caused by the necessity to lower pad grinding power for attaining high planarity.
The present invention has been accomplished in order to solve the foregoing problems, and an object of the present invention to provide a dresser which achieves high pad grinding power and high pad planarity at the same time, and to provide a dresser in which pullout of grit particles is prevented.

Means for Solving the Problem

The gist of the present invention is as follows:
[1] A dresser for an abrasive cloth including:
a metallic support; and
abrasive grits secured to a surface of the metallic support in a single layer,
wherein the surface to which the abrasive grits are secured has a convex shape,
the height difference between the center and the periphery of the surface is 3 μm to 40 μm, and
at least one pair of the nearby abrasive grits satisfies the relationship d≦L<2d, where d is the abrasive grit diameter and L is the center-to-center spacing between the nearby abrasive grits.
[2] The dresser according to [1], wherein the height difference between the center and the periphery of the surface of the metallic support to which the abrasive grits are secured is 5 μm to 20 μm.
[3] The dresser according to [1] or [2], wherein the abrasive grit diameter satisfies the relationship 3 μm≦d<100 μm.
[4] The dresser according to any one of [1] to [3], wherein the abrasive grits are at least one of diamond grits, cubic boron nitride grits, boron carbide grits, silicon carbide grits, and aluminum oxide grits.
[5] The dresser according to any one of [1] to [4], wherein the abrasive grits are secured to the surface of the metallic support by brazing.
[6] The dresser according to any one of [1] to [5], wherein the metallic support is made of stainless steel.

ADVANTAGEOUS EFFECT OF THE INVENTION

The dresser according to the present invention can ensure high pad planarity while attaining sufficient pad grinding power. Thus, when the dresser is employed as a CMP pad conditioner for Al or glass magnetic hard disk substrates, it provides advantageous effects of achieving high product quality due to improved product surface planarity, as well as high productivity.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 is a graph of pad grinding rate vs. height difference across grit securing surface;

FIG. 2 is a graph of planarity vs. height difference across grit securing surface;

FIG. 3 is a graph of pad grinding rate vs. center-to-center grit spacing; and

FIG. 4 is a graph of planarity vs. grit diameter.

BEST MODE FOR CARRYING OUT THE INVENTION

When an Al or glass magnetic hard disk substrate is to be polished by CMP, the pad's surface planarity is a particularly important key factor. It is thus desirable to reduce the grit diameter as much as possible. However, the dresser's pad grinding power decreases with decreasing grit diameter. To overcome this problem the inventors conducted extensive studies to develop a dresser which can attain both high pad planarity and high pad grinding power. The inventors prepared various types of test dressers—which vary in grit diameter, grit spacing or grit pattern as well as in the surface shape of their metallic support on which the grits are bonded. The inventors then conducted detailed evaluations for their pad grinding rate—a measure of pad grinding power—and pad planarity. More specifically, the inventors investigated in detail the thickness reduction across the surface of a resin (e.g., urethane) polishing pad in a grinding process.
In these studies, the inventors established that, in order to ensure a sufficient pad grinding rate and pad planarity, it is effective to form a convex shape for the surface of the metallic support on which abrasive grits are to be secured (hereinafter also simply referred to as “grit securing surface”) as well as to set the grit-to-grit spacing to fall within a specific range.
High pad grinding power and high pad planarity are ensured when the convex grit securing surface of the metallic support is so configured that the height difference between the center and periphery of the convex surface (hereinafter simply referred to as “height difference” in some cases) is 3-40 μm. The lower limit of the height difference is particularly important; when it is less than 3 μm, it results in low pad grinding rates. The reduction is significant particularly when the abrasive grit size is smaller. By shaping the grit securing surface of the metallic support to form a convex surface and setting the height difference to 3 μm or greater, the slurry and grinding debris can be smoothly removed from the dresser surface during a CMP pad grinding process. When the height difference exceeds 40 μm, on the other hand, it results in poor pad planarity. Further, when the height difference is from 5-30 μm, it is possible to obtain higher pad grinding rate and pad planarity. Moreover, when the height difference is from 5-20 μm, it is possible to obtain much higher pad grinding rate and pad planarity.
The top of the convex shape of the grit securing surface preferably comes at the center thereof. The convex shape be of substantially conical shape, substantially circular truncated conical shape, substantially pyramid shape or substantially truncated pyramid shape. The term “substantially conical shape” refers to a conical shape or somewhat distorted conical shape. The term “somewhat distorted conical shape” includes a conical shape which has been somewhat distorted by the grinding of the polishing pad. Thus, the sides of the vertical cross section of the conical shape may be linear or curved. The same is true for substantially circular truncated conical shape, substantially pyramid shape and substantially truncated pyramid shape. When the grit securing surface is made flat or concave, the pad grinding rate and pad planarity decrease. When the grit securing surface has a substantially circular truncated conical shape or substantially truncated pyramid shape, the ratio of the top surface diameter to the bottom surface diameter is preferably 1:2 to 1:20. In the case of substantially truncated pyramid shape, the bottom surface preferably has a polygonal shape such as octagon or icosagon. In the present invention, the convex grit securing surface is preferably a substantially conical shape in view of easiness of formation.
The inventors obtained results which totally runs counter to the conventional concept that pad grinding rate can be enhanced simply by increasing the downforce applied to each of the abrasive grits bonded to the metallic support, which can be suitably achieved by setting large grit spacing and reducing the total number of the abrasive grits.
In cases where pad planarity is not required to the same extent as that for polishing of magnetic hard disk substrates, it is possible to employ relatively large abrasive grits of around 120-200 μm diameter for pad grinding by CMP. In these cases it is known that the pad grinding rate increases with increasing grit spacing. However, the inventors found that when the abrasive grit diameter is as small as less than 100 μm, the pad grinding rate decreases as the grit spacing decreases, contrary to what has been known in the art. Specifically, the inventors was able to obtain high pad grinding rates when center-to-center spacing L between at least one pair of nearby abrasive grits (also referred to as “center-to-center grit spacing L) and grit diameter d satisfied the relationship d≦L<2d. When L is equal to or greater than 2d, it results in poor pad grinding power. When L is less than d, abrasive grits cannot be arranged in a single layer. When abrasive grits which are arranged so as to satisfy the above relationship account for 50% or more of the total abrasive grits on the dresser, the effects of the present invention can be obtained. It is more preferable that such abrasive grits account for 70% or more of the total abrasive grits.
Even when L≧2d, pad planarity tends to increase when grit diameter d is small. However, the inventors found that when center-to-center grit spacing L and grit diameter d satisfy the relationship d≦L<2d, pad planarity significantly improved compared to the case where grit diameter d is set smaller when L≧2d. When center-to-center grit spacing L satisfies the relationship d≦L<1.5d, pad planarity improves further. As described above, the effect brought about by this becomes significant when the grit diameter is less than 100 μm. Nevertheless, grit diameter d is preferably 3 nm≦d<100 nm because the pad grinding rate may decrease if d is less than 3 μm. With the above dresser according to the present invention, it is possible to achieve high pad grinding power and high pad planarity at the same time.
It is more preferable that grit diameter d be 50 μm or less because pad planarity further improves. Moreover, pad planarity further improves when d is 20 μm or less. It should be noted, however, that larger diameter abrasive grits are easy to handle.
Grit diameter may be measured with any suitable method; however, it is preferable in the present invention to employ an average grit diameter as measured in the manner described below. Measurement may be made to either abrasive grits before attached to a metallic support or abrasive grits scraped off from the metallic support. A number-average grit diameter may be employed which is measured by the sieve classification method, laser diffraction method, centrifugation, direct observation with a scanning electron microscope (SEM), etc. When secured abrasive grits are measured, as grit diameter, an average grit diameter can be used which is obtained by averaging circle equivalent diameters of the abrasive grits measured by direct observation under SEM.
The abrasive grits may be arranged either regularly or irregularly. In the case of regular arrangement, they can be arranged in various patterns, e.g., triangle, square, pentagonal or hexagonal lattice pattern.
The abrasive grits for the dresser according to the present invention preferably have high hardness as well as are less reactive to acidic or alkaline slurries; for example, diamond grit particles, cubic boron nitride grit particles, boron carbide grit particles, silicon carbide grit particles, aluminum oxide grit particles, silicon oxide grit particles, and cerium oxide grit particles may be used. Of these, diamond grit particles, cubic boron nitride grit particles, boron carbide grit particles, silicon carbide grit particles, and aluminum oxide grit particles are preferable. These types of abrasive grit may be used either alone or in combination. Further, the abrasive grits may be covered with at least one metal selected from titanium, zirconium and chrome. It is also possible to cover the abrasive grits with at least compound selected from titanium carbide, zirconium carbide and chrome carbide. Generally, each type of abrasive grit is used alone. However, it is possible to further improve pad grinding power while ensuring pad planarity by combining two or more different types of equally-sized abrasive grits.
The dresser according to the present invention is manufactured in the manner below. Firstly, a brazing metal is temporarily attached to a metallic support. For the metallic support, a stainless steel is preferable because it is less reactive to acidic or alkaline slurries as are materials of abrasive grits. Suitable examples of stainless steels for the metallic support include SUS304, SUS316 and SUS430, which are representative stainless steels. It is also possible to employ steels for general structural purposes (e.g., carbon steels) plated with Ni or other metal.
There are no particular limitations to the shape of the metallic support; it may have a polygonal shape such as octagon or icosagon as described above. However, because pad grinding takes place by rotating the metallic support itself, the metallic support preferably has a disc shape in order to ensure uniform grinding across the pad surface. A convex grit securing surface may be formed by machining process or deformation process. In the deformation process the metallic support is pushed at the center while being supported at the periphery. Moreover, any desired convex shape can be obtained by thermal deformation of the metallic support by heat treatment in combination with the machining process and/or deformation process. The above processing may be effected after abrasive grits have been secured to the surface of the metallic support. To form a convex shape by thermal deformation, in a cooling process after heat treatment, the rate of cooling the grit securing surface may be set high compared to the rate of cooling the opposite surface of the metallic support from the grit securing surface. To achieve this, for example, the opposite surface of the metallic support may be in contact with a heat insulating member such as alumina.
For the brazing metal, Ni—Cr—Fe—Si—B metals, Ni—Si—B metals and Ni—Cr—Si—B metals, as represented by JIS BNi-2 and JIS BNi-5, can be employed. When using a brazing foil, it can be temporarily attached to the metallic support by spot welding. When using a powder brazing metal, it can be kneaded with a cellulose binder or the like and applied to the metallic support. The abrasive grits may be evenly spaced in a single layer over the brazing metal layer. The abrasive grits are temporarily immobilized with a paste to avoid displacement. Under vacuum condition of around 10⁻³Pa, the temperature is raised to the melting point of the brazing metal. Most of the binder, paste and other materials are vaporized during the course of heat rising. Preferably, the temperature at which a brazing metal melts (also referred to as “brazing process temperature”) is not below the melting point of the brazing metal and is preferably as close to the melting point as possible. Brazing process temperature is preferably within the liquidus temperature plus 20° C. or so, at the highest. This is because too high brazing process temperature causes large thermal deformation in the metallic support. Only around 5-30 minutes will suffice for the time during which the metallic support is exposed to brazing process temperature. Instead of brazing, the abrasive grits may be secured to the metallic support by electroplating of Ni or the like.

EXAMPLES

The present invention will be described in detail with reference to Examples.

Example 1

Dressers were manufactured using metallic supports with different surface shapes while using diamond abrasive grits of two different average particle diameters (grit diameters): 15 μm and 70 μm. The dressers were evaluated for their pad grinding rate and pad planarity. For the metallic supports, SUS304 stainless discs of 100 mm diameter and 4 mm thickness were used. The metallic supports were machined to form a convex (conical), flat or concave grit securing surface thereon. The height difference between the top and periphery of the dresser surface thus machined was measured as follows. A coordinate system in millimeters was established on the surface of the metallic support so that the x-axis and y-axis intersect at the center of the metallic support surface. Using a dial gauge, the height at the origin (0, 0) was measured. Similarly, the heights at the coordinates (0, 49), (0, −49), (−49, 0) and (49, 0)—four points near the periphery of the disc, each located 1 mm away from the edge toward the center—were measured. The height differences between the origin point and each of the four periphery points were calculated. The obtained four differences were then averaged and defined as the height difference between the top and periphery of the dresser surface.
Thirteen metallic supports with different grit securing surface shapes were prepared. Diamond abrasive grits with average grit diameter d of 15 μm were secured to the respective metallic supports in a square lattice pattern at center-to-center grit spacing L of 1.5d. More specifically, the diamond abrasive grits were applied in a ring area (48 mm outer diameter and 35 μm-inner inner diameter) along the top surface of the disc. Here, the ring area is divided into 6 segments by 2 mm-width spaces on which no abrasive grits were applied, so that the segments are oriented at equal angles relative to the center of the ring. On the respective metallic supports with different height differences, the diamond abrasive grits were arranged at center-to-center grit spacing L of 1.5d. In this way 13 dressers were prepared, where the length of each side of the square lattice is equal to center-to-center grit spacing L.
In practice, diamond grits were secured to the surface of the stainless support as follows. Firstly, by spot welding, a brazing foil was temporarily attached to the surface of the stainless support where diamond abrasive grits are to be arranged. A sieve was then fabricated which includes through holes which are arranged in a square lattice pattern and are small enough for diamond grit particles to pass through. The sieve was placed over the support and diamond grits were arranged thereon through the sieve. The used brazing metal was a 30 μm-thick brazing foil with a composition of: 84.98 wt % Ni, 0.12 wt % Fe, 7.4 wt % Cr, 4.0 wt % Si, 3.0 wt % B and 0.5 wt % P. An organic adhesive was applied on the brazing metal to prevent displacement of the diamond grits. Brazing was carried out at 980° C. for 15 minutes in a vacuum.
Similarly, 13 metallic support with different grit securing surface shapes were prepared. On the respective metallic supports, diamond grits with average grit diameter d of 70 μm were arranged in a square lattice pattern at center-to-center grit spacing L of 1.5d. In this way 13 dressers were manufactured.
Tables 1 and 2 show measurement results for the grit securing surfaces of the metallic supports of the dressers prepared above. Table 1 shows results for 15 μm-diameter diamond abrasive grits, and Table 2 shows results for 70 μm-diameter diamond abrasive grits. In the tables, the dressers with height difference of 3-40 μm were grouped into Examples, whereas those outside this range were grouped into Comparative Examples. Note that dresser Nos. 1 and 21 grouped into Comparative Example had negative values for height difference, which indicates that their grit securing surfaces were concaved.
Polishing pads were ground using the dressers thus manufactured. Pad grinding rates were measured based on the pad thickness reductions after grinding. Further, the pads were measured for their planarity. Polyurethane foam pads of 250 mm diameter were employed. Specifically, the pad was attached to the surface of a polishing platen. The dresser was attached to a machine equipped with a rotation mechanism and a mechanism which moves back and forth across the radius of the pad surface. The dresser was pressed against the pad at a downforce of 1 kg with a pressing mechanism. The dresser was then caused to move back and forth across the radius of the rotating pad over a stroke of 30-90 mm (one back-and-forth motion constitutes one stroke). The pad rotational rate was set at 90 rpm, the dresser rotational rate was set at 80 rpm, and the stroke count was set at 10 strokes per minute. The pad and dresser were caused to rotate in the same direction. During grinding water was supplied to an extent that the pad surface being ground is covered with water. The grinding operation was stopped 5 minutes after the start of grinding. The pad thickness was measured along the two mutually orthogonal diameters with a length-measuring microscope. Specifically, along each diameter the thickness of the pad was measured at 10 points that are substantially equidistant from each other, with the measurement points lying at nearly the center of each of 10 equally divided diameter line segments. In this way pad thickness was measured at 20 points in total, followed by averaging of the measured values. Grinding was started again, and 15 hours after the restart of grinding, pad thickness measurements were made in the same manner. Based on the averaged value of pad thickness, the average pad grinding rate during 15 hour-grinding was found. Pad planarity was found by subtracting the minimum value of pad thickness from the maximum value of pad thickness among the 20 pad thicknesses measured after 15-hour grinding.

TABLE 1

(15 μm-diameter diamond abrasive grit)

	No. 5	No. 6	No. 7	No. 8	No. 9	No. 10	No. 11
	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.

Height	3.0	4.5	5.9	16.8	24.7	29.5	38.5
difference
(μm)
Pad	1.9	2.0	2.2	2.2	2.4	2.6	2.7
grinding
rate
(μm/min)
Pad	0.13	0.15	0.18	0.20	0.39	0.50	0.72
planarity
(μm)

	No. 1	No. 2	No. 3	No. 4	No. 12	No. 13
	Comp.	Comp.	Comp.	Comp.	Comp.	Comp.
	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.

Height	−2.1	0	1.8	2.1	45.8	52.8
difference
(μm)
Pad	0.7	0.9	1.2	1.4	2.8	3.1
grinding
rate
(μm/min)
Pad	0.28	0.12	0.12	0.13	0.96	1.28
planarity
(μm)

TABLE 2

(70 μm-diameter diamond abrasive grit)

	No. 25	No. 26	No. 27	No. 28	No. 29	No. 30	No. 31
	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.

Height	3.1	4.7	6.1	17.2	24.1	29.8	36.5
difference
(μm)
Pad	2.1	2.1	2.3	2.4	2.6	2.9	3.2
grinding
rate
(μm/min)
Pad	0.46	0.51	0.55	0.65	0.68	0.73	0.78
planarity
(μm)

	No. 21	No. 22	No. 23	No. 24	No. 32	No. 33
	Comp.	Comp.	Comp.	Comp.	Comp.	Comp.
	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.

Height	−2.3	0	1.7	2.3	46.3	51.8
difference
(μm)
Pad	0.8	0.9	1.2	1.5	3.4	3.6
grinding
rate
(μm/min)
Pad	0.62	0.42	0.43	0.45	1.11	1.35
planarity
(μm)

From the results shown in Tables 1 and 2 (FIGS. 1 and 2), it was confirmed that dresser Nos. 5-11 and 25-31, where height difference falls within a range of 3 μm to 40 μm, showed good pad grinding rates (1.9 μm/min or more) and good planarities (0.8 μm or less). It was also confirmed that dresser Nos. 7-10 and 27-30, where height difference falls within a range of 5 μm to 30 μm, showed much better pad grinding rates and planarities. On the other hand, dresser Nos. 12, 13, 32 and 33, where height difference exceeds 40 μm, showed poor planarity (0.95 μm or more). In the case of dresser Nos. 1, 2, 21 and 22 with flat or concave surface, the pad grinding rate decreased to less than 1.0 μm/min and there was a tendency concave surface show poor pad planarity.

Example 2

SUS304 stainless discs of 100 mm diameter and 4 mm thickness were employed as metallic supports. The metallic supports were machined to form a convex (conical) grit securing surface thereon. The measured value of the height difference between the center and the periphery of the surface was 17 μm. Height measurements were made as in Example 1. The minimum height difference was 16.1 μm, and the maximum height difference was 17.8 μm. Diamond abrasive grits with an average diameter of 8 μm were applied in a ring area (48 mm outer diameter and 35 μm-inner inner diameter) along the top surface of the disc. Here, the ring area is divided into 6 segments by 2 mm-width spaces on which no abrasive grits were applied, so that the segments are oriented at equal angles relative to the center of the ring. In this way dressers were manufactured in which diamond abrasive grits were arranged in a square lattice pattern at different center-to-center grit spacings. The length of each side of the square lattice pattern is equal to center-to-center grit spacing L. As shown in FIG. 3, for the dressers with diamond abrasive grits with average diameter d of 8 μm, L varied from 9 μm to 47 μm.
Similarly, dressers with abrasive grits with average diameter d of 12 μm, 48 μm, 70 μm, 95 μm and 150 μm were manufactured. For the abrasive grits where d is 12 μm, dressers where L is adjusted from 12 μm to 120 μm were manufactured. For the abrasive grits where d is 48 μm, dressers where L is adjusted from 52 μm to 230 μmm were manufactured; for the abrasive grits where d is 70 μm, dressers where L is adjusted from 73 μm to 320 μm were manufactured; for the abrasive grits where d is 95 μm, dressers where L is adjusted from 100 μm to 360 μm were manufactured;
and for the abrasive grits where d is 150 μm, dressers where L is adjusted from 157 μm to 410 μm were manufactured. Here, when d is 12 μm for example, diamond abrasive grits are substantially in contact with each other when L is also 12 μm. The diamond abrasive grits were arranged in the same manner as in Example 1.
Using the dressers thus manufactured above, polishing pads were ground as in Example 1, followed by measurement of their pad grinding rates based on the pad thickness reduction after grinding, and their pad planarities.
FIG. 3 shows the results of the measurement of pad grinding rate (μm/min). From the results of FIG. 3, it was confirmed that the use of diamond abrasive grits with average diameter d of 8 μm succeeded in obtaining a pad grinding rate of as high as 2.0 μm/min when center-to-center grit spacing L was set to 9-16 μm, a range that satisfies the claimed range d≦L<2d. However, when L≧2d, there was a tendency that the pad grinding rate decreases with increasing L.
The use of diamond abrasive grits with average diameter d of 12 μm resulted in obtaining a pad grinding rate of as high as 2.1 μm/min when center-to-center grit spacing L was set to 12-23 μm so as satisfy the claimed range d≦L<2d. However, when L≧2d, there was a tendency that the pad grinding rate decreases with increasing L.
The use of diamond abrasive grits with average diameter d of 48 μm resulted in obtaining a pad grinding rate of as high as 2.2 μm/min when center-to-center grit spacing L was set to 52-95 μm so as satisfy the claimed range d≦L<2d. However, when L≧2d, there was a tendency that the pad grinding rate decreases with increasing L.
The use of diamond abrasive grits with average diameter d of 70 μm resulted in obtaining a pad grinding rate of as high as 2.4 μm/min when center-to-center grit spacing L was set to 73-120 μm so as satisfy the claimed range d≦L<2d. However, when L≧2d, there was a tendency that the pad grinding rate decreases with increasing L.
The use of diamond abrasive grits with average diameter d of 95 μm succeeded in obtaining a pad grinding rate of as high as 2.6 μm/min when center-to-center grit spacing L was set to 100-185 μm so as satisfy the claimed range d≦L<2d. However, when L≧2d, there was a tendency that the pad grinding rate decreases with increasing L.
FIG. 4 shows the results of measurement of pad planarity for the dressers to which diamond abrasive grits with average diameter d ranging from 5 μm to 148 μm were secured by brazing at center-to-center grit spacing L of 1.3d to 4 d. From the results of FIG. 4, it was confirmed that when L=1.3d and L=1.9d, each satisfying the claimed range d≦L<2d, it succeeded in obtaining good planarities of as low as less than 0.8 μm over the entire d range. Moreover, when L=1.3d which satisfies L<1.5d, pad planarity further improved.
Further, when L=1.3d or L=1.9d and d is less than 100 μm, pad planarity significantly increased compared to the case where d is made smaller with L≧2d. In particular, when d<50 μm, it resulted in good pad planarities of as low as 0.6 μm or less. When d<20 μm, it resulted in much better pad planarities of as low as 0.3 μm or less. When d<10 μm, it resulted in even much better pad planarities of as low as 0.2 μm or less. Although excellent pad planarity of as low as 0.04 μm was obtained when the diamond abrasive grit diameter was 2 μm (L=1.8d), it showed poor pad grinding rate of as low as 1.6 μm/min (not shown).
When diamond abrasive grits with average diameter d of 150 μm were used, there was a tendency that pad grinding rate increases with increasing L (see FIG. 3). When L=2.4d and L=4d, which fall outside the claimed range, pad planarity increased to 0.9 μm or more over the entire grit diameter range (see FIG. 4).
From the above results it is clear that the dressers according to the present invention can ensure sufficient pad grinding power as well as excellent pad planarity.

Example 3

SUS304 stainless discs of 100 mm diameter and 4 mm thickness were emoloyed as metallic supports. The metallic supports were machined to form a flat grit securing surface thereon. Diamond abrasive grits with average diameter d of 30 μm were applied in a ring area (48 mm outer diameter and 35 μm-inner inner diameter) along the top surface of the disc. Here, the ring area is divided into 6 segments by 2 mm-width spaces on which no abrasive grits were applied, so that the segments are oriented at equal angles relative to the center of the ring. The diamond abrasive grits were arranged in the following way. A grid was drawn on the diamond abrasive grit arrangement surface, and then diamond abrasive grits were applied at the intersections of the lines of the grid. Grid spacing corresponds to center-to-center grit spacing L. Two different grid cell side lengths were employed: 50 μm (L=1.7d) and 100 μm (L=3.3d). In the grid, 50 μm-grid cell sides and 100 μm-grid cell sides were randomly provided in predetermined proportions.
In practice, diamond grits were secured to the surface of the stainless support as follows. Firstly, by spot welding, a brazing foil was temporarily attached to the surface of the stainless support where diamond abrasive grits are to be arranged. A sieve was then fabricated which includes through holes which are arranged in a square lattice pattern and are small enough for diamond grit particles to pass through. The sieve was placed over the support and diamond grits were arranged thereon through the sieve. The through holes were made by drawing a grid on the sieve plate and by drilling holes at the intersections of the lines of the grid. In the grid, two different grid cell side lengths were employed: 50 μm (L=1.7d) and 100 μm (L=3.3d), and 50 μm-grid cell sides and 100 μm-grid cell sides were randomly provided in predetermined proportions. Various through hole patterns were designed that had values of “[N50/(N50+N100)]×100(%)” shown in Table 3, where N50 is the number of 50 μm-grid cell sides, and N100 is the number of 100 μm-grid cell sides.
The brazing metals, diamond grit securing method and brazing treatment are the same as those in Example 1. To alter the surface shape of the dressers after brazing, the steel supports were pushed at the center while being supported at the periphery, forming a conical surface shape. The measured value of height difference obtained in the same manner as in Example 1 was 8 μm. In the height difference measurement, the height of the grit secured surface was measured while taking care not to include the height of the diamond grit. Pad grinding rate and pad planarity were evaluated in the same manner as in Example 1. The results are shown in Table 3 below.

TABLE 3

								No. 128
	No. 121	No. 122	No. 123	No. 124	No. 125	No. 126	No. 127	Comp.
No.	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.	Ex.

[N50/(N50 +	100	85	75	65	55	35	20	0
N100)] × 100
(%)
Pad grinding rate	2.1	2.1	2.1	2.1	2.1	2.0	2.0	2.0
(μm/min)
Pad planarity	0.30	0.37	0.45	0.62	0.70	0.78	0.80	1.1
(μm)

From the results of Example Nos. 121-127, it is clear that it is possible to obtain excellent pad planarity of as low as 0.8 μm or less when center-to-center spacing L between at least one pair of nearby abrasive grits satisfies the claimed range d≦L<2d. It is also clear that it is possible to obtain excellent pad planarity of as small as 0.7 μm or less when the abrasive grits which satisfy d≦L<2d account for 50% or more of the total abrasive grits on the dresser surface, and pad planarity of as small as 0.5 μm or less when such abrasive grits account for 70% or more of the total abrasive grits. It should be noted that because a large number of abrasive grits are used in the present invention, the value of ““[N50/(N50+N100)]×100(%)” can be considered as the proportion of abrasive grits arranged at center-to-center grit spacing of 50 μm to the total number of abrasive grits. Specifically, assume that the abrasive grits are arranged at the intersections of the lines of a grid consisting of randomly arranged grid cells having different grid cell side lengths of L B₁, B₂. . . and B_n(where n is a positive integer), the value obtained from the equation [NL/(NL+NB₁+NB₂+ . . . +NB_n)]×100″ (where NL is the number of grid cell sides with length L, and NB₁, NB₂and NB_nare the numbers respectively of grid cell sides with length of B₁, B₂and B_n) can be considered as the proportion of abrasive grits arranged at center-to-center grit spacing of L μm to the total number of abrasive grits.

Example 4

Dressers were manufactured which are identical to dresser No. 121 of Example 3 except that the diamond grits were replaced by cubic boron nitride grits, boron carbide grits, silicon carbide grids, aluminum oxide grits, 50:50 (by weight) mixture of the boron carbide grits and silicon carbide grits, or silicon oxide grits. These abrasive grits were 30 μm in average grit diameter.
The method of arranging abrasive grits, brazing method, and evaluation methods of pad grinding rate and pad planarity are the same as those described in Example 1. The results are shown in Table 4 below.

TABLE 4

		Pad planarity	Pad grinding rate
No.	Abrasive grit	(μm)	(μm/min)

Example No. 131	Cubic boron nitride	0.50	1.7
Example No. 132	Boron carbide	0.41	1.6
Example No. 133	Silicon carbide	0.32	1.8
Example No. 134	Aluminum oxide	0.45	1.5
Example No. 135	Boron carbide +	0.39	1.9
	Silicon carbide
Example No. 136	Silicon oxide	0.60	1.5

From the above results it was confirmed that the use of at least one of cubic boron nitride grits, boron carbide grits, silicon carbide grids, aluminum oxide grits and silicon oxide grits resulted in good pad planarity. In particular, the use of at least one of cubic boron nitride grits, boron carbide grits, silicon carbide grits and aluminum oxide grits resulted high performance. Moreover, it is clear that the combined use of two different types of abrasive grits enhanced pad grinding power while ensuring pad planarity.

Example 5

A dresser was manufactured which is identical to dresser No. 121 of Example 3 except that the following method of forming a convex grit securing surface was employed. Specifically, after brazing heat treatment as conducted in Example 1, alumina was brought in contact with the opposite surface of the metallic support from the grit securing surface, causing deformation of the metallic support to form a convex (conical) grit securing surface. The convex was adjusted to give a desired convex shape by application of stress to the metallic support. The measured value of the height difference between the center and the periphery of the grit securing surface was 10 μm. Height measurements were made as in Example 1. The height of the grit secured surface was measured while taking care not to include the height of diamond grit secured by brazing. The method of arranging abrasive grits, and evaluation methods of pad grinding rate and pad planarity are the same as those described in Example 1. It was confirmed that the dresser showed a sufficient pad grinding rate and pad planarity, with pad grinding rate being 2.2 μm/min and pad planarity being 0.31 μm.
The present application claims the priority of Japanese Patent Application No. 2008-039218 filed on Feb. 20, 2008, the entire contents of which are herein incorporated by reference.

Claims

1. A dresser for an abrasive cloth comprising:

a metallic support; and

abrasive grits secured to a surface of the metallic support in a single layer,

wherein the surface to which the abrasive grits are secured has a convex shape,

the height difference between the center and the periphery of the surface is 3 μm to 40 μm, and

at least one pair of the nearby abrasive grits satisfies the relationship d≦L<2d, where d is the abrasive grit diameter and L is the center-to-center spacing between the nearby abrasive grits.

2. The dresser according to claim 1, wherein the height difference between the center and the periphery of the surface of the metallic support to which the abrasive grits are secured is 5 μm to 20 μm.

3. The dresser according to claim 1 or 2, wherein the abrasive grit diameter satisfies the relationship 3 nm≦d<100 nm.

4. The dresser according to any one of claims 1 to 3, wherein the abrasive grits are at least one of diamond grits, cubic boron nitride grits, boron carbide grits, silicon carbide grits, and aluminum oxide grits.

5. The dresser according to any one of claims 1 to 4, wherein the abrasive grits are secured to the surface of the metallic support by brazing.

6. The dresser according to any one of claims 1 to 5, wherein the metallic support is made of stainless steel.