Field
[0001] The present invention relates to a polishing pad. More particularly, the present
invention relates to a polishing pad preferably used in order to form a flat surface
in a semiconductor, a dielectric/metallic composite, an integrated circuit, and the
like.
Background
[0002] As the density of a semiconductor device becomes higher, the importance of technologies
such as multilayer wiring, and formation of interlayer insulating films and electrodes
(such as a plug and a damascene structure) associated with the multilayer wiring is
increasing. At the same time, the importance of planarization processes of the interlayer
insulating films and the electrode metal films is increasing. As an efficient technology
for the planarization processes, a polishing technology called CMP (Chemical Mechanical
Polishing) is widespread.
[0003] The CMP apparatus generally includes a polishing head that holds a semiconductor
wafer as a subject to be processed, a polishing pad for performing a polishing process
of a subject to be processed, and a polishing platen that holds the polishing pad.
In a polishing process of a semiconductor wafer using a slurry, a semiconductor wafer
and a polishing pad move relative to each other, so that projections of a semiconductor
wafer surface layer are removed to planarize the wafer surface layer. A pad surface
is updated by dressing with a diamond dresser and the like for clogging prevention
and setting.
[0004] Conventionally, there is known a technology to improve wafer flatness and a polishing
rate by providing a groove that is arranged on a polishing layer surface and has a
concentric circular pattern and a substantially rectangular cross-sectional shape
(for example, see Patent Literature 1).
[0005] However, in this technology, corners in a cross-sectional shape of a groove and burr-like
materials formed in the corners caused by dressings performed prior to, following
to, or during polishing may sometimes cause generation of scratches. To solve this
problem, there is disclosed a technology of providing an inclined surface at a boundary
between a polishing surface and a groove (for example, see Patent Literatures 2 and
3).
Citation List
Patent Literature
[0006]
Patent Literature 1: Japanese Patent Application Laid-Open No. 2002-144219
Patent Literature 2: Japanese Patent Application Laid-Open No. 2004-186392
Patent Literature 3: Japanese Patent Application Laid-Open No. 2010-45306
Summary
Technical Problem
[0007] Here, the inventors have found that an inclined surface is provided at a boundary
between a polishing surface and a groove, so that not only scratches are reduced,
but also improvement in suction and slurry flow between a wafer and a polishing pad
is developed to increase the polishing rate. However, the inventors have also found
that variation of a polishing rate cannot be suppressed at some angle of the inclined
surface. Furthermore, the inventors have also found that provision of such an inclined
surface reduces a polishing surface area to increase a pad cut rate, resulting in
shortened pad life.
[0008] In view of the above problems associated with conventional technologies, an object
of the present invention is to provide a polishing pad that, among other polishing
properties, has a long life and can suppress variation of a polishing rate while maintaining
a high polishing rate.
Solution to Problem
[0009] The inventors considered that an inclination from a polishing surface to a groove
bottom influences a pad cut rate, and that an angle at a boundary between a polishing
surface and a groove influences a polishing rate. To balance them, the inventors considered
that the problems could be solved by combining an angle at which a pad cut rate decreases
and an angle at which variation of a polishing rate decreases.
[0010] Therefore, the present invention employs the following means to solve the above problems.
That is, a polishing pad includes at least a polishing layer, wherein the polishing
layer includes a groove on a polishing surface, the groove having side surfaces, at
least one of the side surfaces includes a first side surface that extends continuously
to the polishing surface and forms an angle α with the polishing surface, and a second
side surface that extends continuously to the first side surface and forms an angle
β with a plane parallel to the polishing surface, the angle α formed with the polishing
surface is larger than 95 degrees, the angle β formed with the plane parallel to the
polishing surface is larger than 95 degrees, and the angle β formed with the plane
parallel to the polishing surface is smaller than the angle α formed with the polishing
surface, and a bending point depth from the polishing surface to a bending point between
the first side surface and the second side surface is more than 0.2 mm and not more
than 3.0 mm.
Advantageous Effects of Invention
[0011] According to the present invention, a polishing pad that has a long life and can
suppress variation of a polishing rate while maintaining a high polishing rate can
be provided.
Brief Description of Drawings
[0012]
FIG. 1 is a partial cross-sectional view illustrating a configuration of a main part
of a polishing pad according to an embodiment of the present invention.
FIG. 2 is a partial cross-sectional view illustrating the configuration (second example)
of a main part of a polishing pad according to an embodiment of the present invention.
FIG. 3 is a partial cross-sectional view illustrating the configuration (third example)
of a main part of a polishing pad according to an embodiment of the present invention.
FIG. 4 is a partial cross-sectional view illustrating the configuration (fourth example)
of a main part of a polishing pad according to an embodiment of the present invention.
Description of Embodiments
[0013] Embodiments for carrying out the present invention will be described below.
[0014] The inventors extensively studied a polishing pad that has a long life and can suppress
variation of a polishing rate while maintaining a high polishing rate. As a result,
the inventors found that the problem described above can be solved once for all by
configuring a polishing pad having at least a polishing layer, wherein the polishing
layer includes a groove on a polishing surface, and the groove has side surfaces;
at least one of the side surfaces includes a first side surface that extends continuously
from the polishing surface and forms an angle α with the polishing surface and a second
side surface that extends continuously from the first side surface and forms an angle
β with a plane parallel to the polishing surface; the angle α formed with the polishing
surface is larger than 95 degrees, the angle β formed with the plane parallel to the
polishing surface is larger than 95 degrees, and the angle β formed with the plane
parallel to the polishing surface is smaller than the angle α formed with the polishing
surface; and a bending point depth from the polishing surface to a bending point between
the first side surface and the second side surface is more than 0.2 mm and not more
than 3.0 mm.
[0015] In the present invention, the polishing pad preferably has at least a cushion layer
in addition to the polishing layer. When a cushion layer is not provided, distortion
caused by, for example, water absorption of the polishing layer cannot be buffered.
Therefore, a polishing rate and in-plane uniformity of a material to be polished unstably
vary. A distortion constant of the cushion layer is preferably not lower than 7.3
× 10
-6 µm/Pa and not higher than 4.4 × 10
-4 µm/Pa. From a viewpoint of polishing rate variation and local flatness of a material
to be polished, the upper limit of the distortion constant is preferably not higher
than 3.0 × 10
-4 µm/Pa, and more preferably not higher than 1.5 × 10
-4 µm/Pa. Also, the lower limit of the distortion constant is preferably not lower than
1.0 × 10
-5 µm/Pa, and more preferably not lower than 1.2 × 10
-5 µm/Pa. When polishing rate variation is large, a polishing amount of a material to
be polished varies. As a result, a film thickness of a material to be polished varies,
thereby adversely affecting performance of a semiconductor device. Therefore, the
polishing rate variation is preferably not more than 20%, more preferably not more
than 15%.
[0016] A distortion constant in the present invention was calculated according to the following
equation:

wherein T1 (µm) is a thickness when a pressure of 27 kPa is applied for 60 seconds
with a dial gauge using an indenter having a leading end diameter of 5 mm, and T2
(µm) is a thickness when a pressure of 177 kPa is applied for 60 seconds thereafter.
[0017] Examples of such a cushion layer may include, but are not limited to, natural rubber,
nitrile rubber, "Neoprene (registered trademark)" rubber, polybutadiene rubber, thermosetting
polyurethane rubber, thermoplastic polyurethane rubber, silicone rubber, non-foamed
elastomer such as "Hytrel (registered trademark)", a polyolefin foamed body such as
"Toraypef (registered trademark, PEF manufactured by Toray Industries, Inc.)", and
non-woven fabric such as "Suba 400" manufactured by Nitta Haas Incorporated.
[0018] The distortion constant of the cushion layer can be adjusted depending on a material
thereof. For example, when the cushion layer is a foamed body, increasing a foaming
degree tends to cause the foamed body to become soft. Therefore, the distortion constant
tends to increase. Also, when the cushion layer is non-foamed, hardness can be controlled
by adjusting a crosslinking degree in the cushion layer.
[0019] The thickness of the cushion layer is preferably 0.1 to 2 mm. From a viewpoint of
in-plane uniformity on a whole surface of a semiconductor substrate, the thickness
is preferably not less than 0.25 mm, and more preferably not less than 0.3 mm. Moreover,
from a viewpoint of local flatness, the thickness is preferably not more than 2 mm,
and more preferably not more than 1 mm.
[0020] The polishing layer surface (polishing surface) of the polishing pad according to
the present invention has a groove. Examples of a shape of the groove as seen from
the polishing layer surface may include, but are not limited to, lattice, radial,
concentric circular, and spiral shapes. When the groove is an open-type and extends
in a circumferential direction, slurry can be efficiently updated. Therefore, a lattice
shape is the most preferable.
[0021] According to the present invention, at least one of the side surfaces of the groove
includes a first side surface that extends continuously from a polishing surface and
forms an angle α with the polishing surface, and a second side surface that extends
continuously from the first side surface and forms an angle β with a plane parallel
to the polishing surface. Each of the first side surface and the second side surface
may be plane (linear in a cross-sectional shape of the groove) or curved (curved in
a cross-sectional shape of the groove).
[0022] In the present invention, the angle α is larger than 95 degrees, the angle β is larger
than 95 degrees, and the angle β is smaller than the angle α. Thus, variation of a
polishing rate can be suppressed while maintaining a high polishing rate. This can
be explained as below. Variation of a polishing rate is generally large in initial
and middle stages of polishing. However, by providing an inclined surface having an
angle larger than 95 degrees at a boundary between the polishing surface and the groove,
not only a polishing rate increases, but also such variation of a polishing rate in
initial and middle stages can be effectively suppressed.
[0023] On the other hand, in such a configuration, a contact area between a material to
be polished and a polishing pad surface is small. Accordingly, there is a concern
that a pad cut rate is high. Therefore, a configuration in which the groove provides
a larger contact area when the depth of the groove is equal to or deeper than a certain
depth is preferable. By adjusting the angle α and the angle β as described above,
such an object can be achieved. The difference between the angle α and the angle β
is more preferably not larger than 55 degrees, and further preferably not larger than
50 degrees.
[0024] From a viewpoint of retention and fluidity of slurry, the lower limit of the angle
α is preferably not smaller than 105 degrees, and more preferably not smaller than
115 degrees. Also, the upper limit of the angle α is preferably not larger than 150
degrees, and more preferably not larger than 140 degrees. Both the side surfaces forming
a groove and facing each other may have a similar shape. However, since slurry flows
due to a centrifugal force, it is more effective that, of the side surfaces forming
a groove and facing each other, at least the side surface on a circumferential side
has an inclination. The angle β is not limited as long as the angle β is smaller than
the angle α. However, the upper limit of the angle β is preferably smaller than 150
degrees, and further preferably smaller than 140 degrees.
[0025] Here, a side surface (side surface 3) that extends continuously from the side surface
2 in a direction opposite to the side surface 1. In such a case, an angle (angle 3)
formed between the side surface 3 and the polishing surface is preferably larger than
95 degrees and smaller than the angle β.
[0026] Similarly, when n is a natural number equal to or more than 3, a side surface (side
surface (n + 1)) that extends continuously in a direction opposite to a side surface
(n - 1) with respect to a side surface n can be provided. In such a case, an angle
(angle (n + 1)) formed between the side surface (n + 1) and the polishing surface
is preferably larger than 95 degrees and smaller than an angle n.
[0027] When the polishing layer is scraped off as a material to be polished is polished
and the polishing surface passes the bending point that is a boundary between the
first side surface and the second side surface, variation of a polishing rate can
occur. Furthermore, a pad cut rate differs depending on whether a groove side surface
in the shallowest part is the first side surface or the second side surface. Therefore,
the depth from the polishing surface to the bending point is preferably equal to or
deeper than a level of inhibiting reduction in effects of an inclined groove part
on the polishing surface side. In view of this, and considering that the life of a
polishing pad is preferably long, a specific depth from the polishing surface to the
bending point is preferably not less than 10% and not more than 95% of the depth of
the entire groove, and more preferably not less than 20% and not more than 90% thereof.
[0028] Since it is important that a pad life is long and that suppression of variation of
a polishing rate is balanced with, the bending point depth from the polishing surface
to the bending point between the first side surface and the second side surface is
more than 0.2 mm and not more than 3.0 mm. The polishing surface described here means
a polishing surface before the polishing layer is scraped off. When the bending point
depth is deep, a pad life becomes short. When the bending point depth is shallow,
a polishing rate varies. The upper limit of the bending point depth from the polishing
surface to the bending point between the first side surface and the second side surface
is preferably not more than 2.5 mm, more preferably not more than 2.0 mm, and further
preferably not more than 1.8 mm. Also, the lower limit of the bending point depth
from the polishing surface to the bending point between the first side surface and
the second side surface is preferably not less than 0.3 mm, more preferably not less
than 0.4 mm, and further preferably not less than 0.5 mm.
[0029] A specific shape of the groove according to the present invention as described above
will be described with reference to the drawings. FIG. 1 is a partial cross-sectional
view illustrating the configuration of a main part of a polishing pad according to
an embodiment of the present invention. A polishing pad 1 illustrated in FIG. 1 has
a polishing layer 10. A groove 12 is formed on a polishing surface 11 of the polishing
layer 10. The groove 12 has a first side surface 13 that extends continuously to the
polishing surface 11 and inclines at an angle α formed with respect to the polishing
surface 11, a second side surface 15 that extends continuously to the first side surface
13 and bends with respect to the first side surface 13 at a bending point 14, and
a deepest groove part 16. An angle β of the second side surface with respect to a
plane parallel to the polishing surface 11 is smaller than the angle α of the first
side surface 13 with respect to the polishing surface 11.
[0030] Here, a groove shape configured by the second side surfaces 15 and the deepest part
16 is not limited to the shape illustrated in FIG. 1. For example, like a groove 17
of a polishing pad 2 illustrated in FIG. 2, a deepest part 18 may have a bottom surface
substantially parallel to the polishing surface 11. Also, like a groove 19 of a polishing
pad 3 illustrated in FIG. 3, a boundary part between the second side surface 15 and
a deepest part 20 may constitute a curved surface. Also, like a groove 21 of a polishing
pad 4 shown in FIG. 4, the cross-sectional shape of second side surfaces 15 and a
deepest part 22 may constitute a U-shape.
[0031] As the polishing layer constituting the polishing pad, a closed cell structure is
preferable, because a flat surface is formed in a semiconductor, a dielectric/metallic
composite, an integrated circuit, and the like. The hardness of the polishing layer
measured by an Asker D hardness meter is preferably 45 to 65 degrees. When the Asker
D hardness is less than 45 degrees, as the wafer in-plane uniformity of a polishing
rate for a material to be polished decreases, the uniformity of wafer in-plane planarization
properties (planarity) tends to decrease.
[0032] Examples of a material for forming such a structure may include, but are not particularly
limited to, polyethylene, polypropylene, polyester, polyurethane, polyurea, polyamide,
polyvinyl chloride, polyacetal, polycarbonate, polymethyl methacrylate, polytetrafluoroethylene,
epoxy resin, ABS resin, AS resin, phenol resin, melamine resin, "Neoprene (registered
trademark)" rubber, butadiene rubber, styrene butadiene rubber, ethylene propylene
rubber, silicone rubber, fluorine rubber, and resins including these as a main component.
Two or more of these may be used. Also in these resins, since a closed cell diameter
can be relatively easily controlled, a material including polyurethane as a main component
is more preferable.
[0033] Polyurethane is a macromolecule synthesized by a polyaddition reaction or a polymerization
reaction of polyisocyanate. A compound used as a reference of polyisocyanate is an
active hydrogen-containing compound that is a compound containing two or more polyhydroxy
groups or an amino group-containing compound. Examples of polyisocyanate may include,
but are not limited to, tolylene diisocyanate, diphenylmethane diisocyanate, naphthalene
diisocyanate, hexamethylene diisocyanate, and isophorone diisocyanate. Two or more
of these may be used.
[0034] A compound containing a polyhydroxy group is representatively polyol. Examples thereof
may include polyether polyol, polytetramethylene ether glycol, epoxy resin-modified
polyol, polyester polyol, acrylic polyol, polybutadiene polyol, and silicone polyol.
Two or more of these may be used. Combination and optimum amounts of polyisocyanate
and polyol, and a catalyst, a foaming agent and a foam stabilizer are preferably determined
depending on hardness, a cell diameter and a foaming ratio.
[0035] As a method of forming closed cells in the polyurethane, a chemical foaming method
in which various foaming agents are blended into a resin during production of polyurethane
is generally used. However, a method including foaming a resin by mechanical stirring
and thereafter curing the foamed resin may also preferably be used.
[0036] The average cell diameter of closed cells is preferably not less than 30 µm in order
to reduce scratches. Also, in view of flatness of local unevenness of a material to
be polished, the average cell diameter is preferably not more than 150 µm, more preferably
not more than 140 µm, and further preferably not more than 130 µm. The average cell
diameter is obtained as follows. Of cells observed in one field of view when observing
a sample section at a magnification of 400 times using an ultra-deep microscope VK-8500
manufactured by Keyence Corporation, circular cells excluding cells that are observed
in a circle in a state of being deficient in the field end are measured using an image
processing apparatus to obtain a circle-equivalent diameter from the cross-sectional
area. Then, a number average value is calculated.
[0037] A preferred embodiment of the polishing pad according to the present invention is
a pad that contains a polymer of a vinyl compound as well as polyurethane and has
closed cells. With only a polymer from a vinyl compound, toughness and hardness can
be improved, but a uniform polishing pad having closed cells is unlikely to be obtained.
Furthermore, polyurethane becomes brittle when hardness is brought to be higher. By
impregnating a vinyl compound into polyurethane, a polishing pad containing closed
cells and having high toughness and hardness can be obtained.
[0038] A vinyl compound is a polymerizable compound having a carbon-carbon double bond.
Specific examples of the vinyl compound may include methyl acrylate, methyl methacrylate,
ethyl acrylate, ethyl methacrylate, n-butyl acrylate, n-butyl methacrylate, 2-ethylhexyl
methacrylate, isodecyl methacrylate, n-lauryl methacrylate, 2-hydroxy ethyl methacrylate,
2-hydroxy propyl methacrylate, 2-hydroxy butyl methacrylate, dimethylaminoethyl methacrylate,
diethylaminoethyl methacrylate, glycidyl methacrylate, ethylene glycol dimethacrylate,
acrylic acid, methacrylic acid, fumaric acid, dimethyl fumarate, diethyl fumarate,
dipropyl fumarate, maleic acid, dimethyl maleate, diethyl maleate, dipropyl maleate,
phenylmaleimide, cyclohexyl maleimide, isopropyl maleimide, acrylonitrile, acrylamide,
vinyl chloride, vinylidene chloride, styrene, α-methylstyrene, divinylbenzene, ethylene
glycol dimethacrylate, and diethylene glycol dimethacrylate. Two or more of these
may be used.
[0039] Among the above-described vinyl compounds, CH
2=CR
1COOR
2 (R
1: a methyl group or an ethyl group, R
2: a methyl group, an ethyl group, a propyl group, or a butyl group) is preferable.
Especially, methyl methacrylate, ethyl methacrylate, n-butyl methacrylate, and isobutyl
methacrylate are preferable. This is because closed cells can be easily formed into
polyurethane; monomers can be favorably impregnated; polymerization curing can be
easily performed; and a foaming structure containing a polymer of a polymerization-cured
vinyl compound and polyurethane has high hardness and favorable planarization properties.
[0040] Examples of a polymerization initiator preferably used for obtaining these polymers
of vinyl compounds may include a radical initiator such as azobisisobutyronitrile,
azobis(2,4-dimethylvaleronitrile), azobis cyclohexane carbonitrile, benzoyl peroxide,
lauroyl peroxide, and isopropyl peroxy dicarbonate. Two or more of these may be used.
Also, a redox-based polymerization initiator, for example, a combination of peroxide
and amines can be used.
[0041] A method of impregnating a vinyl compound into polyurethane may include a method
including immersing polyurethane in a vessel containing a vinyl compound. At that
time, treatments such as heating, pressurizing, pressure-reducing, stirring, shaking,
and ultrasonic vibration are preferably performed in order to increase an impregnation
speed.
[0042] The impregnation amount of the vinyl compound into polyurethane should be determined
depending on types of the vinyl compound and polyurethane to be used and properties
of a polishing pad to be manufactured. Therefore, the impregnation amount cannot be
completely defined. However, for example, the content ratio of the polymer obtained
from the vinyl compound and polyurethane in a polymerization-cured foamed structure
is preferably 30/70 to 80/20 in terms of weight. When the content ratio of the polymer
obtained from the vinyl compound is not less than 30/70 in terms of weight, hardness
of the polishing pad can be made sufficiently high. Also, when the content ratio is
not more than 80/20, elasticity of the polishing layer can be made sufficiently high.
[0043] Here, the content ratio of the polymer obtained from the polymerization-cured vinyl
compound in polyurethane can be measured by a pyrolysis gas chromatography/mass spectrometry
technique. An apparatus that can be used in this technique may include a double-shot
pyrolyzer "PY-2010D" (manufactured by Frontier Laboratories Ltd.) as a thermal decomposition
apparatus and "TRIO-1" (manufactured by VG) as a gas chromatography and mass spectrometry
apparatus.
[0044] In the present invention, from a viewpoint of flatness of local unevenness of a semiconductor
substrate, a phase of the polymer obtained from the vinyl compound and a phase of
polyurethane are preferably contained without being separated from each other. When
expressed quantitatively, it is preferable that an infrared spectrum obtained when
the polishing pad be observed using an infrared microspectrometer with a spot size
of 50 µm have an infrared absorption peak of the polymer polymerized from the vinyl
compound and an infrared absorption peak of polyurethane, and that infrared spectra
in various locations be approximately the same. An infrared microspectrometer to be
used here may include IRµs manufactured by SPECTRA-TEC.
[0045] In order to improve properties, the polishing pad may contain various additives such
as an abrasive, an antistatic agent, a lubricant, a stabilizer, and a dye.
[0046] In the present invention, in order to reduce poor local flatness and global steps,
the density of the polishing layer is preferably not less than 0.3 g/cm
3, more preferably not less than 0.6 g/cm
3, and further preferably not less than 0.65 g/cm
3. On the other hand, in order to reduce scratches, the density is preferably not more
than 1.1 g/cm
3, more preferably not more than 0.9 g/cm
3, and further preferably not more than 0.85 g/cm
3. Here, the density of the polishing layer in the present invention is a value measured
using a Harvard-type pycnometer (in accordance with JIS R-3503 standard) with water
as a medium.
[0047] Examples of a material to be polished in the present invention may include a surface
of an insulating layer or a metal wiring formed on a semiconductor wafer. The insulating
layer may include an interlayer insulating film of a metal wiring, a lower-layer insulating
film of a metal wiring, and a shallow trench isolation layer used for element isolation.
The metal wiring may be made from aluminum, tungsten, copper, or an alloy thereof.
Examples of a structure of the metal wiring may include damascene, dual damascene,
and a plug. When copper is used as the metal wiring, barrier metal such as silicon
nitride also becomes a subject to be polished. Currently, silicon oxide is mainly
used as the insulating film. However, a low dielectric constant insulating film is
also used. In addition to a semiconductor wafer, a magnetic head, a hard disk, sapphire,
SiC, MEMS (Micro Electro Mechanical Systems), and the like may be used as a subject
to be polished.
[0048] The polishing method according to the present invention is suitably used in order
to form a flat surface of glass, a semiconductor, a dielectric/metallic composite,
an integrated circuit, and the like.
Examples
[0049] The present invention will be further described in detail by examples. However, the
present invention should not be interpreted to be limited by the examples. Measurement
was performed as below.
<Measurement of cell diameter>
[0050] Of cells observed in one field of view when observing a sample section at a magnification
of 400 times using an ultra-deep microscope VK-8500 manufactured by Keyence Corporation,
circular cells excluding cells that are observed in a circle in a state of being deficient
in the field end are measured using an image processing apparatus to obtain a circle-equivalent
diameter from the cross-sectional area. A number average value is calculated to serve
as an average cell diameter.
<Measurement of hardness>
[0051] Measurement was performed in accordance with JIS K6253-1997. The produced polyurethane
resin was cut out into a piece having a size of 2 cm × 2 cm (thickness: optional).
The piece was used as a hardness measurement sample, and left to stand for 16 hours
in an environment of a temperature of 23°C ± 2°C and a humidity of 50% ± 5%. During
measurement, samples were superimposed on each other to have a thickness of not less
than 6 mm. Hardness was measured using a hardness meter (manufactured by Kobunshi
Keiki Co., Ltd., Asker D-type hardness meter).
<Measurement of micro rubber A hardness>
[0052] A cushion layer was cut out into a piece having a size of 3 cm × 3 cm. The piece
was used as a hardness measurement sample, and left to stand for 16 hours in an environment
of a temperature of 23°C ± 2°C and a humidity of 50% ± 5%. Different three points
in one piece of sample were measured using a micro rubber hardness meter MD-1 manufactured
by Kobunshi Keiki Co., Ltd. An average value was calculated to serve as a micro rubber
A hardness.
<Measurement of inclination angle>
[0053] A pad having a groove formed on a polishing layer surface was disposed so that a
razor blade was vertical to a groove direction. Then, the pad was sliced in a groove
depth direction. The obtained groove section was observed by an ultra-deep microscope
VK-8500 manufactured by Keyence Corporation. An angle (angle α) formed between a polishing
surface and a side surface extending continuously to the groove polishing surface
was measured. At locations of 1/3 and 2/3 of a radius from a pad center, the closest
grooves were measured. An average of one each location, two locations in total, was
calculated to serve as an inclination angle. An angle β was measured in a similar
manner thereto.
<Measurement of bending point depth>
[0054] A pad having a groove formed on a polishing layer surface was disposed so that a
razor blade was vertical to a groove direction. Then, the pad was sliced in a groove
depth direction. The obtained groove section was observed by an ultra-deep microscope
VK-8500 manufactured by Keyence Corporation. A vertical distance from the polishing
surface, to a midpoint between two bending points each including a first side surface
and a second side surface and both facing each other, was measured. At locations of
1/3 and 2/3 of a radius from a pad center, the closest grooves were measured. An average
of one each location, two locations in total, was calculated to serve as a bending
point depth.
<Measurement of initial inter-bending point distance>
[0055] A pad having a groove formed on a polishing layer surface was disposed so that a
razor blade was vertical to a groove direction. Then, the pad was sliced in a groove
depth direction. The obtained groove section was observed by an ultra-deep microscope
VK-8500 manufactured by Keyence Corporation. A distance between two bending points
each having an angle α and including a polishing surface and a first side surface
and both facing each other was measured to obtain a bending point distance. Also,
the inter-bending point distance in the initial stage of polishing was determined
as an initial inter-bending point distance.
<Calculation of distortion constant>
[0056] A distortion constant was calculated according to the following equation:

wherein T1 (µm) is a thickness when a pressure of 27 kPa was applied for 60 seconds
with a dial gauge using an indenter having a leading end diameter of 5 mm, and T2
(µm) is a thickness when a pressure of 177 kPa was applied for 60 seconds thereafter.
<Calculation of average polishing rate>
[0057] Using Mirra 3400 manufactured by Applied Materials, Inc., polishing was performed
while performing end point detection under a given polishing condition. Polishing
properties were measured in a diameter direction, excluding a region of 10 mm from
the outermost circumference of an 8-inch wafer. Measurement was performed at 37 points
per 5 mm on a surface within a radius of 90 mm from the center. Then, an average polishing
rate (nm/minute) was calculated.
<Calculation of polishing rate variation>
[0058] After 1000 wafers were polished and an average polishing rate was measured wafer
by wafer, a polishing rate variation of the first to 700th wafers was calculated according
to the following equation:

[0059] When the variation of a polishing rate is large, insufficient polishing or excess
polishing can cause device failure. Therefore, the polishing rate variation is suitably
low, preferably not more than 30%, and more preferably not more than 20%.
<Measurement of average pad cut rate>
[0060] Using Mirra 3400 manufactured by Applied Materials, Inc., polishing was performed
while performing end point detection under a given polishing condition. With a depth
gauge, a groove depth (D1) mm after polishing 30 workpieces and a groove depth (D2)
mm after polishing 1000 workpieces were measured. Calculation was made from dress
time (t
d) minutes by a dresser.

[0061] The average pad cut rate depends on the inter-bending point distance as well as the
angle α and the angle β. The inter-bending point distance changes as polishing proceeds.
When the average inter-bending point distance from the initial stage to the final
stage of polishing is smaller, the average pad cut rate is lower.

<Calculation of polishing pad life>
[0062] A groove depth in the polishing initial stage was measured. Then, an effective groove
depth (D3) mm that is shallower by 0.3 mm from the deepest part was calculated. Calculation
was made from a time (t
p) minute during which a wafer was polished and the average pad cut rate.

[0063] The polishing pad life is preferably not less than 15 hours.
[0064] Examples 1 to 12 and Comparative Examples 1 to 4 will be described below.
(Example 1)
[0065] In a RIM molding machine, 30 parts by weight of polypropylene glycol, 40 parts by
weight of diphenylmethane diisocyanate, 0.5 parts by weight of water, 0.3 parts by
weight of triethylamine, 1.7 parts by weight of a silicone foam stabilizer, and 0.09
parts by weight of tin octylate were mixed. The mixture was discharged into a mold
and subjected to pressure molding. Thus, a foamed polyurethane sheet containing closed
cells was produced.
[0066] The foamed polyurethane sheet was immersed in methyl methacrylate added with 0.2
parts by weight of azobisisobutyronitrile for 60 minutes. Next, the foamed polyurethane
sheet was immersed in a solution including 15 parts by weight of polyvinyl alcohol
"CP" (polymerization degree: about 500, manufactured by Nacalai Tesque Inc.), 35 parts
by weight of ethyl alcohol (special grade chemical, manufactured by Katayama Chemical
Co., Ltd.), and 50 parts by weight of water, and then dried. Thus, a surface layer
of the foamed polyurethane sheet was coated with polyvinyl alcohol.
[0067] Next, the foamed polyurethane sheet was placed between two glass plates via vinyl
chloride gaskets, and then heated for 6 hours at 65°C and for 3 hours at 120°C to
be polymerization-cured. The sheet was released from between the glass plates, washed
with water, and then vacuum-dried at 50°C. The hard foamed sheet obtained as above
was subjected to a slicing process into a piece having a thickness of 2.00 mm. Thus,
a polishing layer was produced. The content ratio of methyl methacrylate in the polishing
layer was 66% by weight. The polishing layer had a D hardness of 54 degrees and a
density of 0.81 g/cm
3. An average cell diameter of closed cells was 45 µm.
[0068] Both surfaces of the obtained hard foamed sheet were ground. Thus, a polishing layer
having a thickness of 2.4 mm was produced.
[0069] Thermoplastic polyurethane (cushion layer thickness: 0.3 µm) manufactured by Nihon
Matai Co., Ltd. having a distortion constant of 0.15×10
-4 µm/Pa (micro rubber A hardness 89) as a cushion layer was laminated on the polishing
layer obtained by the above method via an MA-6203 adhesive layer manufactured by Mitsui
Chemicals Polyurethanes, Inc. using a roll coater. Furthermore, a double-sided tape
5604TDM manufactured by Sekisui Chemical Co., Ltd. as a rear surface tape was bonded
to the rear surface thereof. This laminate was punched into a circle having a diameter
of 508 mm. A groove having a groove pitch of 15 mm, an angle α of 135 degrees, an
angle β of 120 degrees, and a groove depth of 1.9 mm was formed in an XY grid pattern
on the polishing layer surface. Thus, a polishing pad was obtained. At this time,
the bending point depth was 0.69 mm, and the initial stage inter-bending point distance
was 3 mm.
[0070] The polishing pad obtained by the above method was pasted on a platen of a polishing
machine ("Mirra 3400" manufactured by Applied Materials, Inc.). Under a retainer ring
pressure = 41 kPa (6 psi), an inner tube pressure = 28 kPa (4 psi), a membrane pressure
= 28 kPa (4 psi), a platen revolution = 76 rpm, a polishing head revolution = 75 rpm,
and a slurry (manufactured by Cabot Corporation, SS-25) flow of 150 mL/minute, 1000
8-inch wafers as oxide films were polished using a dresser manufactured by Saesol
at a load of 17.6 N (4 lbf), a polishing time of 1 minute, and an in-situ dressing
time of 30 seconds after polishing started.
[0071] The average polishing rate of the 1000th oxide film was 192.2 nm/minute. The polishing
rate variation for 1000 oxide films was 8.5%. The average pad cut rate was 1.22 µm/minute,
and the polishing pad life was 22 hours. Thus, the results were favorable.
(Example 2)
[0072] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 145 degrees, the
polishing layer thickness was changed to 2.25 mm, and the groove depth was changed
to 1.75 mm. At this time, the bending point depth was 0.46 mm, and the initial stage
inter-bending point distance was 3 mm. The average polishing rate was 195.2 nm/minute,
and the polishing rate variation was 13.2%. The average pad cut rate was 1.15 µm/minute,
and the polishing pad life was 21 hours. Thus, the results were favorable.
(Example 3)
[0073] Polishing was performed in the same manner as that in Example 1, except that the
angle β of the groove on the polishing layer surface was changed to 100 degrees, the
polishing layer thickness was changed to 3.15 mm, and the groove depth was changed
to 2.65 mm. At this time, the bending point depth was 1.37 mm, and the initial stage
inter-bending point distance was 3.4 mm. The average polishing rate was 184.1 nm/minute,
and the polishing rate variation was 17.2%. The average pad cut rate was 1.22 µm/minute,
and the polishing pad life was 32 hours. Thus, the results were favorable.
(Example 4)
[0074] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 100 degrees, the
angle β was changed to 98 degrees, the polishing layer thickness was changed to 2.0
mm, and the groove depth was changed to 1.5 mm. At this time, the bending point depth
was 0.3 mm, and the initial stage inter-bending point distance was 3 mm. The average
polishing rate was 187.8 nm/minute, and the polishing rate variation was 17.8%. The
average pad cut rate was 1.20 µm/minute, and the polishing pad life was 16 hours.
Thus, the results were favorable.
(Example 5)
[0075] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 150 degrees, the
angle β was changed to 145 degrees, the polishing layer thickness was changed to 2.0
mm, and the groove depth was changed to 1.5 mm. At this time, the bending point depth
was 0.27 mm, and the initial stage inter-bending point distance was 5 mm. The average
polishing rate was 201.9 nm/minute, and the polishing rate variation was 18.9%. The
average pad cut rate was 1.24 µm/minute, and the polishing pad life was 16 hours.
Thus, the results were favorable.
(Example 6)
[0076] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 160 degrees, the
angle β was changed to 110 degrees, the polishing layer thickness was changed to 2.5
mm, and the groove depth was changed to 2.05 mm. At this time, the bending point depth
was 0.79 mm, and the initial stage inter-bending point distance was 5 mm. The average
polishing rate was 183.8 nm/minute, and the polishing rate variation was 16.4%. The
average pad cut rate was 1.35 µm/minute, and the polishing pad life was 21 hours.
Thus, the results were favorable.
(Example 7)
[0077] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 115 degrees, the
angle β was changed to 100 degrees, the polishing layer thickness was changed to 2.0
mm, and the groove depth was changed to 1.5 mm. At this time, the bending point depth
was 0.27 mm, and the initial stage inter-bending point distance was 3 mm. The average
polishing rate was 182.5 nm/minute, and the polishing rate variation was 17.5%. The
average pad cut rate was 1.22 µm/minute, and the polishing pad life was 16 hours.
Thus, the results were favorable.
(Example 8)
[0078] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 165 degrees, the
angle β was changed to 155 degrees, the polishing layer thickness was changed to 2.2
mm, and the groove depth was changed to 1.7 mm. At this time, the bending point depth
was 0.5 mm, and the initial stage inter-bending point distance was 5 mm. The average
polishing rate was 190.2 nm/minute, and the polishing rate variation was 15.6%. The
average pad cut rate was 1.36 µm/minute, and the polishing pad life was 17 hours.
Thus, the results were favorable.
(Example 9)
[0079] Polishing was performed in the same manner as that in Example 1, except that the
polishing layer thickness was changed to 2.9 mm, and the groove depth was changed
to 2.4 mm. At this time, the bending point depth was 2.1 mm, and the initial stage
inter-bending point distance was 3 mm. The average polishing rate was 185.7 nm/minute,
and the polishing rate variation was 14.4%. The average pad cut rate was 1.23 µm/minute,
and the polishing pad life was 28 hours. Thus, the results were favorable.
(Example 10)
[0080] Polishing was performed in the same manner as that in Example 1, except that the
polishing layer thickness was changed to 3.5 mm, and the groove depth was changed
to 3.0 mm. At this time, the bending point depth was 2.6 mm, and the initial stage
inter-bending point distance was 3 mm. The average polishing rate was 183.3 nm/minute,
and the polishing rate variation was 15.1%. The average pad cut rate was 1.24 µm/minute,
and the polishing pad life was 36 hours. Thus, the results were favorable.
(Example 11)
[0081] Polishing was performed in the same manner as that in Example 1, except that two
angles α that face each other via the groove on the polishing layer surface were changed
to 135 degrees and 130 degrees so that the two angles facing each other differ from
each other. At this time, the bending point depth was 0.69 mm, and the initial stage
inter-bending point distance was 3 mm. The average polishing rate was 191.8 nm/minute,
and the polishing rate variation was 9.0%. The average pad cut rate was 1.20 µm/minute,
and the polishing pad life was 22 hours. Thus, the results were favorable.
(Example 12)
[0082] Polishing was performed in the same manner as that in Example 1, except that a polyester
film having a thickness of 188 µm was bonded to the rear surface of the polishing
layer via an adhesive, and a cushion layer was bonded to the polyester film surface.
At this time, the bending point depth was 0.69 mm, and the initial stage inter-bending
point distance was 3 mm. The average polishing rate was 192.8 nm/minute, and the polishing
rate variation was 9.3%. The average pad cut rate was 1.22 µm/minute, and the polishing
pad life was 22 hours. Thus, the results were favorable.
(Comparative Example 1)
[0083] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 93 degrees, the
angle β was changed to 90 degrees, the polishing layer thickness was changed to 2.0
mm, and the groove depth was changed to 1.5 mm. At this time, the bending point depth
was 0.27 mm, and the initial stage inter-bending point distance was 1.5 mm. The average
polishing rate was 180.1 nm/minute, and the polishing rate variation was 45.1%. Thus,
the polishing rate variation was large. The average pad cut rate was 1.12 µm/minute,
and the polishing pad life was 18 hours. Thus, the results were favorable.
(Comparative Example 2)
[0084] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 93 degrees, the
angle β was changed to 90 degrees, the polishing layer thickness was changed to 2.0
mm, and the groove depth was changed to 1.5 mm. At this time, the bending point depth
was 0.27 mm, and the initial stage inter-bending point distance was 3 mm. The average
polishing rate was 189.5 nm/minute, and the polishing rate variation was 30.8%. Thus,
the polishing rate variation was large. The average pad cut rate was 1.5 µm/minute,
and the polishing pad life was 13 hours. Thus, the life was short.
(Comparative Example 3)
[0085] Polishing was performed in the same manner as that in Example 1, except that the
angle β was changed to 98 degrees, the polishing layer thickness was changed to 2.0
mm, and the groove depth was changed to 1.5 mm. At this time, the bending point depth
was 0.15 mm, and the initial stage inter-bending point distance was 3 mm. The average
polishing rate was 190.1 nm/minute, and the polishing rate variation was 36.2%. Thus,
the polishing rate variation was large. The average pad cut rate was 1.42 µm/minute,
and the polishing pad life was 14 hours. Thus, the life was short.
(Comparative Example 4)
[0086] Polishing was performed in the same manner as that in Example 1, except that the
angle α of the groove on the polishing layer surface was changed to 160 degrees, the
angle β was changed to 100 degrees, the polishing layer thickness was changed to 2.5
mm, and the groove depth was changed to 2.0 mm. At this time, the bending point depth
was 0.60 mm, and the initial stage inter-bending point distance was 4 mm. The average
polishing rate was 184.6 nm/minute, and the polishing rate variation was 31.0%. Thus,
the polishing rate variation was large. The average pad cut rate was 1.32 µm/minute,
and the polishing pad life was 21 hours. Thus, the results were favorable.
[0087] The results obtained in Examples 1 to 12 and Comparative Examples 1 to 4 described
above are shown in Table 1.
Table 1
| |
Example 1 |
Example 2 |
Example 3 |
Example 4 |
Example 5 |
Example 6 |
Example 7 |
Example 8 |
Example 9 |
Example 10 |
Example 11 |
Example 12 |
Comp. Ex. 1 |
Comp. Ex. 2 |
Comp. Ex. 3 |
Comp. Ex. 4 |
| α (degrees) |
135 |
145 |
135 |
100 |
150 |
160 |
115 |
165 |
135 |
135 |
135, 130 |
135 |
93 |
93 |
135 |
160 |
| β (degrees) |
120 |
120 |
100 |
98 |
145 |
110 |
100 |
155 |
120 |
120 |
120 |
120 |
90 |
90 |
98 |
100 |
| Polishing Layer Thickness (mm) |
2.4 |
2.25 |
3.15 |
2.0 |
2.0 |
2.5 |
2.0 |
2.2 |
2.9 |
3.5 |
2.4 |
2.4 |
2.0 |
2.0 |
2.0 |
2.5 |
| Groove Depth (mm) |
1.9 |
1.75 |
2.65 |
1.5 |
1.5 |
2.05 |
1.5 |
1.7 |
2.4 |
3.0 |
1.9 |
1.9 |
1.5 |
1.5 |
1.5 |
2.0 |
| Bending Point Depth (mm) |
0.69 |
0.46 |
1.37 |
0.3 |
0.27 |
0.79 |
0.27 |
0.5 |
2.1 |
2.6 |
0.69 |
0.69 |
0.27 |
0.27 |
0.15 |
0.60 |
| Initial Stage Groove Inter-Bending Point Distance (mm) |
3 |
3 |
3.4 |
3 |
5 |
5 |
3 |
5 |
3 |
3 |
3 |
3 |
1.5 |
3 |
3 |
4 |
| Average Polishing Rate (nm/min.) |
192.2 |
195.2 |
184.1 |
187.8 |
201.9 |
183.8 |
182.5 |
190.2 |
185.7 |
183.3 |
191.8 |
192.8 |
180.1 |
189.5 |
190.1 |
184.6 |
| Polishing Rate Variation (%) |
8.5 |
13.2 |
17.2 |
17.8 |
18.9 |
16.4 |
17.5 |
15.6 |
14.4 |
15.1 |
9.0 |
9.3 |
45.1 |
30.8 |
36.2 |
31.0 |
| Average Pad Cut Rate (µm/min.) |
1.22 |
1.15 |
1.22 |
1.20 |
1.24 |
1.35 |
1.22 |
1.36 |
1.23 |
1.24 |
1.20 |
1.22 |
1.12 |
1.5 |
1.42 |
1.32 |
| Polishing Pad Life (hours) |
22 |
21 |
32 |
16 |
16 |
21 |
16 |
17 |
28 |
36 |
22 |
22 |
18 |
13 |
14 |
21 |
| Example 12 ... Polyester film having thickness of 188 µm bonded to polishing layer
rear surface + cushion layer bonded to polyester film surface |
Reference Signs List
[0088]
1, 2, 3, 4 Polishing pad
10 Polishing layer
11 Polishing surface
12, 17, 19, 21 Groove
13 First side surface
14 Bending point
15 Second side surface
16, 18, 20, 22 Deepest part