JP5879777B2 - Polishing pad, polishing apparatus, polishing method - Google Patents

Description

  The present invention relates to a polishing pad, a polishing apparatus, and a polishing method.

  As this type of technology, Patent Document 1 discloses a two-layer polishing pad used in a semiconductor wafer polishing apparatus by a chemical mechanical polishing (CMP) method. The two-layer polishing pad is affixed to the surface of the platen, and is composed of a soft pad made of a soft material and a hard pad made of a hard material hard to the soft pad. In this document 1, by introducing this two-layer polishing pad, a pressure distribution is realized such that the pressure gradually decreases from the central portion to the outer peripheral portion of the semiconductor wafer.

  Patent Document 2 discloses a CMP polishing pad having a specified surface hardness.

JP 2000-77370 A JP 11-90809 A

  However, in the two-layer polishing pad of Patent Document 1, the polishing amount of the outer peripheral portion of the semiconductor wafer may be larger than the polishing amount of the central portion.

  Patent Document 2 does not mention the above problem at all.

  An object of the present invention is to provide a technique for suppressing the amount of polishing of the outer peripheral portion of an object to be polished when performing polishing by a CMP method using a polishing pad.

According to a first aspect of the present invention, there is provided a polishing pad for polishing a polishing object disposed on the inner peripheral side of a retainer ring in a chemical mechanical polishing (CMP) method, comprising: a hard layer having a polishing surface; And a soft layer that is disposed on the opposite side of the object to be polished across the hard layer, and is softer than the hard layer, and the polishing surface when a different load is applied to the polishing surface. A polishing pad is provided that has a dent amount difference of 86 micrometers or less. According to the above configuration, when performing polishing by the CMP method using the polishing pad, it is possible to suppress the polishing amount of the outer peripheral portion of the object to be polished.
According to a second aspect of the present invention, there is provided a polishing pad for polishing a polishing object disposed on the inner peripheral side of the retainer ring and the retainer ring in a chemical mechanical polishing (CMP) method, comprising: a polishing surface; And a soft layer that is disposed on the opposite side of the object to be polished across the hard layer and is softer than the hard layer, and when different loads are applied to the polishing surface A polishing pad including at least a polishing pad having a difference in the amount of depressions on the polishing surface of 86 micrometers or less.
According to a third aspect of the present invention, there is provided a polishing method for polishing an object to be polished by a chemical mechanical polishing (CMP) method, comprising: a hard layer having a polished surface; and the object to be polished across the hard layer. A polishing layer that is disposed on the opposite side and has a soft layer that is softer than the hard layer, and the difference in the amount of depression of the polishing surface when a different load is applied to the polishing surface is 86 micrometers or less. There is provided a polishing method using a pad to polish the object to be polished disposed on the inner peripheral side of the retainer ring.
Preferably, the number of rotations of the object to be polished is adjusted according to the temperature rise of the polishing pad.

  According to the present invention, when performing polishing by the CMP method using the polishing pad, it is possible to suppress the polishing amount of the outer peripheral portion of the object to be polished.

FIG. 1 is a schematic view of a polishing apparatus. FIG. 2 is an enlarged view of the polishing portion. FIG. 3 is an enlarged view of a portion A in FIG. FIG. 4 is a comparative example and is an enlarged view of a portion A in FIG. FIG. 5 is a diagram illustrating a method of measuring the specific depression amount difference value. FIG. 6 is a distribution graph of the polishing amount. FIG. 7 is a time transition graph of the platen temperature. FIG. 8 is a distribution graph of the polishing amount. FIG. 9 is a distribution graph of the polishing amount.

(Polishing device 1)
As shown in FIG. 1, the polishing apparatus 1 includes a head 2, a platen 3, a control unit 4, a motor driver 5, a head drive motor 6, and a platen drive motor 7.

  The head 2 sucks and holds the semiconductor wafer 8 as an object to be polished. An annular retainer ring 9 is disposed on the outer periphery of the head 2. The semiconductor wafer 8 is disposed on the inner peripheral side of the retainer ring 9.

  A polishing pad 10 is attached to the surface of the platen 3 on the head 2 side. The polishing pad 10 is for polishing the semiconductor wafer 8 held by suction by the head 2.

  The control unit 4 drives the head drive motor 6 and the platen drive motor 7 via the motor driver 5. The head drive motor 6 rotates the head 2. The platen drive motor 7 rotates the platen 3.

  The platen 3 is equipped with a thermocouple 11 for measuring the temperature of the platen 3. The thermocouple 11 indirectly measures the temperature of the polishing pad 10 by measuring the temperature of the platen 3. The output voltage of the thermocouple 11 is input to the control unit 4.

  Next, the polishing pad 10 will be described in detail with reference to FIG. In the present embodiment, the polishing pad 10 is a so-called two-layer polishing pad. That is, the polishing pad 10 includes a hard layer 12 having a polishing surface 10 a and a soft layer 13 that is disposed on the opposite side of the semiconductor wafer 8 with the hard layer 12 interposed therebetween and is softer than the hard layer 12. Yes. The hard layer 12 is made of, for example, a polyurethane resin material. The soft layer 13 is made of, for example, polyurethane foam. And the hard layer 12 and the soft layer 13 are mutually adhere | attached by the contact bonding layer which is not shown in figure. A release paper 14 is provided on the opposite side of the hard layer 12 across the soft layer 13. The soft layer 13 and the release paper 14 are bonded to each other by an adhesive layer (not shown).

  With the above configuration, the platen 3 and the head 2 shown in FIG. 1 are rotationally driven at a desired rotational speed. At that time, a slurry as an abrasive is appropriately supplied onto the platen 3. Then, an appropriate amount of slurry flows between the head 2 and the platen 3, so that the semiconductor wafer 8 attached to the head 2 is polished by the platen 3. At this time, the retainer ring 9 is pressed against the polishing pad 10 to temporarily secure a high flatness of a portion of the polishing pad 10 located in the retainer ring 9. The retainer ring 9 suppresses only the outer peripheral portion of the semiconductor wafer 8 from being overpolished.

  However, even if the retainer ring 9 is used, the outer peripheral portion of the semiconductor wafer 8 may be overpolished. This problem will be described below with reference to FIGS.

  As shown in FIG. 3, when the retainer ring 9 is pressed against the polishing pad 10, the polishing pad 10 is deformed so as to be recessed in the vicinity of the retainer ring 9, thereby forming a recessed portion P on the polishing surface 10 a. Due to the presence of the depression P, a slight gap g is formed between the outer peripheral portion of the semiconductor wafer 8 and the polishing surface 10a. During polishing, a large amount of slurry flows into the gap g. On the other hand, the inflow amount of the slurry and the polishing rate (degree of progress of polishing) have a positive correlation. Therefore, due to the presence of the gap g, the outer peripheral portion of the semiconductor wafer 8 is slightly overpolished. However, since the range of about 5 mm from the outer periphery of the semiconductor wafer 8 is not taken out as a chip, there is no problem if it is overpolished in the range shown in FIG.

  Here, according to the earnest study of the present inventor, the hard layer 12 of the polishing pad 10 is not so dominant when it comes to the depth of the recess P in FIG. It turns out to be strong and dominant. That is, the depth of the depression P cannot be managed even if the hardness of the hard layer 12 of the polishing pad 10 is measured.

  On the other hand, FIG. 4 shows a state where the polishing pad 10 is entirely softer than that of FIG. 3 and is polished. That is, there is a great difference in the degree to which the gap g is formed depending on whether the polishing pad 10 is hard or soft as a whole. If the gap g is formed large as shown in FIG. 4, the amount of slurry flowing into the gap g inevitably increases. As described above, there is a positive correlation between the inflow amount of the slurry and the polishing amount. As a result, if the polishing pad 10 is soft as a whole, the outer peripheral portion of the semiconductor wafer 8 is greatly excessive. It can be mentioned that it has been polished.

Therefore, the inventor of the present application conducted a specific indentation amount difference value measurement test shown in FIG. 5 in order to quantitatively evaluate the degree of hardness of the polishing pad 10 as a whole. The specific dimple amount difference value measurement test conforms to JIS L-1096. Specifically, in this test, the truncated cone-shaped weight 15 having the contact bottom surface 15a is gently placed on the hard layer 12 of the polishing pad 10 with the contact bottom surface 15a facing downward. Then, one minute later, with the weight 15 placed on the polishing pad 10, how much the weight 15 is sinking in the polishing pad 10 is measured using, for example, a laser displacement meter. That is, the specific depression amount T as the amount by which the contact bottom surface 15a sinks is measured with reference to the polishing surface 10a of the polishing pad 10 before deformation. When measuring the specific depression amount T, two weights 15 are used. One is a weight 15 having a mass w1 of 641 [g / cm ² ], and one is a weight 15 having a mass w2 of 5161 [g / cm ² ]. When the specific depression amount T is measured twice using these two weights 15, a specific depression amount difference value ΔT “μm”, which is a difference value between the two specific depression amounts T, is calculated. 1).
ΔT [μm] = | T1-T2 | ... (1)
However, the specific hollow amount T1 [μm] is the specific hollow amount T when the weight 15 of the weight w1 is used, and the specific hollow amount T2 [μm] is the specific hollow amount when the weight 15 of the weight w2 is used. T. In addition, as the specific depression amount difference value ΔT “μm”, an average value of measurement results at two different measurement positions is adopted.

FIG. 6 is a graph showing a polishing amount distribution after polishing by performing a polishing test using a plurality of polishing pads 10 having different specific depression amount difference values ΔT. The vertical axis represents the polishing amount [nm], and the horizontal axis represents the distance [mm] from the center of the semiconductor wafer 8. The polishing conditions are as follows.
・ Polishing time: 240 [sec]
Wafer type: DXA (D: dummy, X: no polarity, A: lowest grade) wafer • A 1800 nm plasma CVD film is laminated on the polished surface of the semiconductor wafer 8 prior to the polishing test.
-Number of rotations of head 2: 90 [rpm]
・ Number of rotations of platen 3: 90 [rpm]
-Diameter of semiconductor wafer 8: 200 [mm]

  In FIG. 6, the black circle plot is the specific depression amount difference value ΔT = 81 [μm]. The black rhombus plot shows the specific depression amount difference value ΔT = 86 [μm]. The white square plot shows the specific depression amount difference value ΔT = 93 [μm]. The white triangle plot shows the specific depression amount difference value ΔT = 84 [μm].

  According to the graph of FIG. 6, if the polishing pad 10 having the specific depression amount difference value ΔT of 86 [μm] or less is used, the polishing amount at the outer peripheral portion of the semiconductor wafer 8 can be suppressed (compared to the central portion). I understand that I can do it. More specifically, according to the graph of FIG. 6, if the polishing pad 10 having a specific dent amount difference value ΔT of 86 [μm] or less is used, the polishing amount at a position 5 mm inside from the outer peripheral edge of the semiconductor wafer 8 is the semiconductor. It can be seen that the polishing amount is smaller than the polishing amount at the center of the wafer 8.

  FIG. 6 shows that the smaller the specific depression amount difference value ΔT is, the more the polishing amount at the outer peripheral portion of the semiconductor wafer 8 can be suppressed. However, from another viewpoint, that is, the polishing pad 10 and the semiconductor wafer. In order to secure the familiarity with 8, the specific depression amount difference value ΔT is desirably 55 [μm] or more.

  Although the polishing pad 10 of the present embodiment has been described above, the polishing pad 10 basically has the following features.

  That is, the polishing pad 10 for polishing the semiconductor wafer 8 (polishing object) disposed on the inner peripheral side of the retainer ring 9 in the chemical mechanical polishing (CMP) method is a hard layer having the polishing surface 10a of the polishing pad 10. 12 and a soft layer 13 that is disposed on the opposite side of the semiconductor wafer 8 with the hard layer 12 interposed therebetween and is softer than the hard layer 12. The specific dent amount difference value ΔT, which is the difference in the dent amount of the polishing surface 10a when different loads are applied to the polishing surface 10a, is 86 [μm] or less. When the above polishing pad 10 is used, as shown in FIGS. 3, 4, and 6, the polishing amount of the outer peripheral portion of the semiconductor wafer 8 is suppressed when the polishing pad 10 is used to perform CMP polishing. Can do.

(About the rotation speed of the head 2)
Next, the rotation speed of the head 2 will be described. FIG. 7 is a graph showing the relationship between the polishing time and the platen temperature (that is, the temperature of the polishing pad 10) when the rotational speed of the head 2 is 45 [rpm] and 90 [rpm]. ing. The vertical axis represents the temperature of the platen 3 at a position of 20 [mm] inner peripheral side from the outer peripheral edge of the platen 3 as shown in FIG. Specifically, the temperature [° C.] of the platen 3 means a temperature in the vicinity of the surface of the platen 3 in contact with the polishing pad 10. “The temperature in the vicinity of the surface of the platen 3 in contact with the polishing pad 10” can be equated with the temperature of the polishing pad 10. The temperature of the platen 3 is measured by a thermocouple 11. The horizontal axis represents the polishing time [sec]. According to FIG. 7, it can be seen that the temperature of the platen 3 increases as the rotational speed of the head 2 is increased. Further, as the temperature of the platen 3 increases, the temperature of the polishing pad 10 also increases. As the temperature of the polishing pad 10 increases, the polishing pad 10 tends to soften. This is because the polishing pad 10 is generally formed of a resin material. When the polishing pad 10 is softened, the amount of polishing of the outer peripheral portion of the semiconductor wafer 8 is increased according to FIG.

The above consideration is supported by the results of the polishing test shown in FIG. FIG. 8 is a graph showing a polishing amount distribution after polishing by performing a polishing test while changing the rotation speed of the platen 3. The vertical axis represents the polishing amount [nm], and the horizontal axis represents the distance [mm] from the center of the semiconductor wafer 8. The polishing conditions are as follows.
・ Polishing time: 240 [sec]
Wafer type: DXA wafer / Plasma CVD film having a thickness of 1800 nm is laminated in advance on the polished surface of the semiconductor wafer 8 prior to the polishing test.
-Number of rotations of head 2: 90 [rpm]
-Number of rotations of platen 3: 45 [rpm] or 90 [rpm]
-Diameter of semiconductor wafer 8: 200 [mm]

  According to FIG. 8, it can be seen that when the rotational speed of the platen 3 is set to an extremely high rotational speed such as 90 [rpm], the polishing amount at the outer peripheral portion of the semiconductor wafer 8 is dramatically increased. This is presumably because the rotation speed of the platen 3 was too high and the polishing pad 10 was remarkably heated, so that the polishing pad 10 was excessively softened. Therefore, it can be seen from FIG. 8 that it is important not to increase the rotation speed of the platen 3 too much in order to suppress the polishing amount at the outer peripheral portion of the semiconductor wafer 8. Further, preferably, as shown in FIG. 1, the temperature of the platen 3 is measured during polishing, and when the temperature of the platen 3 exceeds a predetermined value, the number of rotations of the platen 3 is reduced by 20%, etc. It is preferable to adjust the rotation speed of the platen 3 according to the temperature rise of the platen 3 while monitoring the temperature of the platen 3.

  Further, according to FIG. 8, it can be seen that there is a positive correlation between the rotational speed of the platen 3 and the variation width of the polishing amount. That is, when the rotational speed of the platen 3 is 90 [rpm], the difference (= variation width) between the maximum value and the minimum value of the polishing amount of the semiconductor wafer 8 is 150 [nm], whereas the platen 3 Was 45 [rpm], the difference between the maximum value and the minimum value of the polishing amount of the semiconductor wafer 8 was about 100 [nm]. From this, if the number of rotations of the platen 3 is suppressed, the variation width of the polishing amount of the semiconductor wafer 8 can be suppressed, so that stable polishing can be realized.

  Lastly, the usefulness of the technique for reducing the polishing amount of the outer peripheral portion of the semiconductor wafer 8 disclosed above will be added.

FIG. 9 is a graph showing a polishing amount distribution after polishing by performing a polishing test using an SOI (Silicon On Insulator) wafer and a normal DXA wafer. The vertical axis represents the polishing amount [nm], and the horizontal axis represents the distance [mm] from the center of the semiconductor wafer 8. The polishing conditions are as follows.
・ Polishing time: 240 [sec]
A 1800 nm plasma CVD film is laminated in advance on the surface to be polished of the semiconductor wafer 8 prior to the polishing test.
-Number of rotations of head 2: 90 [rpm]
・ Number of rotations of platen 3: 90 [rpm]
-Diameter of semiconductor wafer 8: 200 [mm]

  Unlike normal DXA wafers, SOI wafers are generally subjected to terrace polishing, which is mechanical polishing that scrapes off the outer periphery at an angle, prior to CMP polishing. As already explained, the cause of the polishing amount of the outer peripheral portion of the semiconductor wafer 8 is the gap g shown in FIG. 3 and FIG. 4. As can be seen from FIG. 9, the SOI wafer is more than the normal DXA wafer. The amount of polishing at the outer periphery is remarkable. Therefore, it can be mentioned that the above-described technique that defines the specific depression amount difference value ΔT is particularly useful for a wafer subjected to terrace polishing (particularly, an SOI wafer).

DESCRIPTION OF SYMBOLS 1 Polishing apparatus 2 Head 3 Platen 10 Polishing pad (DELTA) T Specific hollow amount difference value

Claims (3)

A polishing pad for polishing a polishing object disposed on an inner peripheral side of a retainer ring in a chemical mechanical polishing (CMP) method,
A hard layer having a polished surface;
A soft layer disposed on the opposite side of the object to be polished with the hard layer interposed therebetween, and being softer than the hard layer; and
Have
The specific depression amount difference value ΔT is 55 to 86 micrometers ,
Polishing pad.
However, the specific depression amount difference value ΔT is obtained as follows. That is, the truncated cone-shaped weight having the contact bottom surface is gently placed on the hard layer of the polishing pad with the contact bottom surface facing downward. Next, after 1 minute, with the weight resting on the polishing pad, it is measured how much the weight is sinking in the polishing pad. That is, the specific depression amount T as the amount of the contact bottom sinked is measured on the basis of the polishing surface of the polishing pad before deformation. When measuring the specific depression amount T, two weights are used. One has a load W1 of 641 [g / cm2], and one has a load W2 of 5161 [g / cm2]. When the specific depression amount T is measured twice using the two weights, the specific depression amount difference value ΔT “μm”, which is a difference value between the two specific depression amounts T, is calculated. (1).
ΔT [μm] = | T1-T2 | ... (1)
However, the specific depression amount T1 [μm] is the specific depression amount T when the weight of the load w1 is used, and the specific depression amount T2 [μm] is the identification when the weight of the load w2 is used. The amount of depression T. Further, as the specific depression amount difference value ΔT “μm”, an average value of measurement results at two different measurement positions is adopted. Retainer ring,
A polishing pad for polishing a polishing object disposed on an inner peripheral side of the retainer ring in a chemical mechanical polishing (CMP) method, comprising a hard layer having a polishing surface, and the polishing target sandwiching the hard layer A polishing pad that is disposed on the opposite side of the object and has a soft layer that is softer than the hard layer, and a specific indentation difference value ΔT is 55 to 86 micrometers ,
Including at least
Polishing equipment.
However, the specific depression amount difference value ΔT is obtained as follows. That is, the truncated cone-shaped weight having the contact bottom surface is gently placed on the hard layer of the polishing pad with the contact bottom surface facing downward. Next, after 1 minute, with the weight resting on the polishing pad, it is measured how much the weight is sinking in the polishing pad. That is, the specific depression amount T as the amount of the contact bottom sinked is measured on the basis of the polishing surface of the polishing pad before deformation. When measuring the specific depression amount T, two weights are used. One has a load W1 of 641 [g / cm2], and one has a load W2 of 5161 [g / cm2]. When the specific depression amount T is measured twice using the two weights, the specific depression amount difference value ΔT “μm”, which is a difference value between the two specific depression amounts T, is calculated. (1).
ΔT [μm] = | T1-T2 | ... (1)
However, the specific depression amount T1 [μm] is the specific depression amount T when the weight of the load w1 is used, and the specific depression amount T2 [μm] is the identification when the weight of the load w2 is used. The amount of depression T. Further, as the specific depression amount difference value ΔT “μm”, an average value of measurement results at two different measurement positions is adopted. A polishing method for polishing an object to be polished by a chemical mechanical polishing (CMP) method,
A hard layer having a polished surface, and a soft layer disposed on the opposite side of the object to be polished across the hard layer and softer than the hard layer, and applying different loads to the polished surface. Using a polishing pad having a specific dent amount difference value ΔT of 55 to 86 μm, wherein the difference in the dent amount of the polishing surface is 86 μm or less ,
Polishing the polishing object disposed on the inner peripheral side of the retainer ring;
Polishing method.
However, the specific depression amount difference value ΔT is obtained as follows. That is, the truncated cone-shaped weight having the contact bottom surface is gently placed on the hard layer of the polishing pad with the contact bottom surface facing downward. Next, after 1 minute, with the weight resting on the polishing pad, it is measured how much the weight is sinking in the polishing pad. That is, the specific depression amount T as the amount of the contact bottom sinked is measured on the basis of the polishing surface of the polishing pad before deformation. When measuring the specific depression amount T, two weights are used. One has a load W1 of 641 [g / cm2], and one has a load W2 of 5161 [g / cm2]. When the specific depression amount T is measured twice using the two weights, the specific depression amount difference value ΔT “μm”, which is a difference value between the two specific depression amounts T, is calculated. (1).
ΔT [μm] = | T1-T2 | ... (1)
However, the specific depression amount T1 [μm] is the specific depression amount T when the weight of the load w1 is used, and the specific depression amount T2 [μm] is the identification when the weight of the load w2 is used. The amount of depression T. Further, as the specific depression amount difference value ΔT “μm”, an average value of measurement results at two different measurement positions is adopted.