WO2018079105A1

WO2018079105A1 - Wafer manufacturing method and wafer

Info

Publication number: WO2018079105A1
Application number: PCT/JP2017/032814
Authority: WO
Inventors: 田中　利幸; 敏又川; 中島　亮
Original assignee: 株式会社Sumco
Priority date: 2016-10-31
Filing date: 2017-09-12
Publication date: 2018-05-03
Also published as: TW201826367A; JP2018074018A

Abstract

A resin layer forming process for forming a resin layer (RH) on one surface (W1) of a wafer (W) includes: a holding step for holding, through suction, the other surface (W2) of the wafer (W) by a holding means; a flat surface forming step for sandwiching a curable resin having a viscosity not greater than 1000 mPa·s between the flat part of a flat plate and one surface (W1) of the wafer (W) so as to form a pre-cured flat surface conforming to the flat part on the curable resin; a hold releasing step for releasing the holding, through suction, of the other surface (W2); a curing step for forming the resin layer (RH) by curing the curable resin; and a separation step for separating the flat part from the resin layer (RH).

Description

Wafer manufacturing method and wafer

The present invention relates to a wafer manufacturing method and a wafer.

In the semiconductor device manufacturing process, multiple layers of metal and insulating films are formed on the wafer. Since the film thickness uniformity of each layer formed on the wafer affects the device performance, planarization is performed by CMP (Chemical Mechanical Polishing) immediately after the formation of each layer. However, if the wafer has waviness, the accuracy of CMP decreases and a layer with a non-uniform film thickness is formed. Conventionally, the following is known as a technique for planarizing a wavy wafer.

First, a curable resin is applied to one surface of the wafer, and the curable resin is processed flat and cured to form a resin layer. After that, the flat surface of the resin layer is held and the other surface of the wafer is ground and flattened, and after the resin layer is removed or not removed, the other flat surface is held and one of the wafers is held. Grind and flatten the surface. Hereinafter, the technique may be referred to as “resin pasting”.
And further flattening which applied such resin pasting grinding is examined (for example, refer to patent documents 1).

In the method of Patent Document 1, the other surface of the wavy wafer is sucked and held by holding means. Further, a curable resin is dropped onto the film on the stage, and one surface of the wafer is pressed against the dropped curable resin to form a pre-curing flat surface on the curable resin following the film. Thereafter, the suction holding of the other surface is released to cure the curable resin, and a resin layer having a flat surface after curing is formed.

JP 2011-249652 A

However, in the method as disclosed in Patent Document 1, the flat surface after curing of the resin layer after being separated from the film is not sufficiently flattened, and the nanotopography on the wafer surface after mirror polishing cannot be sufficiently reduced. The device may not be properly manufactured.

An object of the present invention is to provide a wafer manufacturing method and a wafer that can be planarized without affecting the manufacture of semiconductor devices.

As a result of intensive studies to solve the above problems, the present inventor can sufficiently flatten the flat surface after curing of the resin layer by appropriately setting the viscosity of the curable resin before curing. I got the knowledge.
The present invention has been completed based on the above findings.

That is, the method for producing a wafer of the present invention includes a resin layer forming step of forming a resin layer on one surface of a wafer cut from a single crystal ingot or a lapped wafer, and the one surface via the resin layer. A first surface grinding step of surface grinding the other surface of the wafer, a resin layer removal step of removing the resin layer, and holding the other surface and subjecting the one surface to surface grinding. A second surface grinding step, wherein the resin layer forming step includes a holding step of sucking and holding the other surface of the wafer by a holding means, a flat portion of the flat surface forming means, and one surface of the wafer A flat surface forming step in which a curable resin having a viscosity of 1000 mPa · s or less is sandwiched between and a pre-curing flat surface following the flat portion is formed on the curable resin, and the holding release for releasing the suction holding of the other surface Craft When, characterized in that it includes a hardening step to form the resin layer by curing the curable resin, and a spacing step of separating the flat portion from the resin layer.

According to the present invention, by forming a resin layer using a curable resin having a viscosity before curing of 1000 mPa · s or less, a sufficiently flattened resin layer having a flat surface after curing can be formed. The waviness of the wafer after the first surface grinding step, the resin layer removal step, and the second surface grinding step can be sufficiently reduced on the wafer provided with the resin layer. As a result, the nanotopography of the wafer surface after mirror polishing can be made sufficiently small, and a wafer capable of appropriately manufacturing a semiconductor device can be provided.

In the wafer manufacturing method of the present invention, it is preferable that the resin layer forming step forms the resin layer so as to satisfy the following formula (1).
V / T ≦ 10 (1)
V: Viscosity before curing of the curable resin (mPa · s)
T: Thickness (μm) of the thickest portion of the curable resin after the holding release step and before the curing step

According to the present invention, by appropriately setting the thickness of the curable resin with respect to the viscosity of the curable resin and setting V / T to 10 or less, deformation of the flat surface after curing of the resin layer can be suppressed, The maximum value of nanotopography on the wafer surface after mirror polishing can be made 5 nm or less.

The wafer of the present invention is characterized in that the maximum value of nanotopography measured in a plurality of regions of 10 mm × 10 mm on the surface is 5 nm or less.

The flowchart of the manufacturing method of the wafer which concerns on one Embodiment of this invention. Explanatory drawing of the manufacturing method of the said wafer. Explanatory drawing of the manufacturing method of the said wafer. Explanatory drawing of the manufacturing method of the said wafer. FIG. 2 is an explanatory diagram of the wafer manufacturing method, showing a state following FIGS. 2A to 2C. FIG. 2 is an explanatory diagram of the wafer manufacturing method, showing a state following FIGS. 2A to 2C. FIG. 2 is an explanatory diagram of the wafer manufacturing method, showing a state following FIGS. 2A to 2C. The flowchart of the resin layer formation process in the manufacturing method of the said wafer. Explanatory drawing of the effect | action of using curable resin whose viscosity V before hardening is 1000 mPa * s or less. Explanatory drawing of the effect | action of using curable resin whose viscosity V before hardening is 1000 mPa * s or less. Explanatory drawing of the effect | action of using curable resin whose viscosity V before hardening is 1000 mPa * s or less. FIG. 6 is an explanatory diagram of the action of using a curable resin having a pre-curing viscosity V of 1000 mPa · s or less, and shows a state following FIGS. 5A to 5C. FIG. 6 is an explanatory diagram of the action of using a curable resin having a pre-curing viscosity V of 1000 mPa · s or less, and shows a state following FIGS. 5A to 5C. The figure which shows the result of the experiment 1 in the Example of this invention. The graph which shows the result of the experiment 2 in the said Example. The graph which shows the result of the experiment 3 in the said Example.

An embodiment of the present invention will be described with reference to the drawings.
[Wafer manufacturing method]
As shown in FIG. 1, in the wafer manufacturing method, first, a single crystal ingot (hereinafter simply referred to as “ingot”) such as silicon, SiC, GaAs, or sapphire is cut with a wire saw to obtain a plurality of wafers ( Step S1: Slicing step).
Next, both surfaces of the wafer are simultaneously planarized by a lapping apparatus (step S2: lapping process) and chamfered (step S3: chamfering process).
At this time, since it is difficult to achieve sufficient planarization of the wafer only by the lapping process, the wafer W in which the undulations W11 and W21 are generated on one surface W1 and the other surface W2, as shown in FIG. can get.
Thereafter, as shown in FIG. 1, a resin layer forming step (step S4) in which a curable resin R (see FIG. 2B) is applied to one surface W1 of the wafer W to form a resin layer RH (see FIG. 2B). A first surface grinding step (step S5) for holding one surface W1 via the resin layer RH and surface grinding the other surface W2 of the wafer W, and a resin layer removal step (step S5) for removing the resin layer RH ( A resin pasting and grinding step including step S6) and a second surface grinding step (step S7) for holding the other surface W2 and surface grinding one surface W1 is performed.

In the resin layer forming step, as shown in FIG. 4, using the holding and pressing device 10 as shown in FIG. 2B, the coating step (step S11), the holding step (step S12), and the flat surface forming step (step S13). ), A holding release process (step S14), a curing process (step S15), and a separation process (step S16).
In the applying step, the curable resin R is applied onto the flat plate 11 as a flat surface forming unit having the flat portion 11A that has been highly flattened. At this time, as the curable resin R, one having a viscosity V before curing (hereinafter simply referred to as “viscosity before curing”) of 1000 mPa · s or less is used. Here, the viscosity V before curing is preferably 100 mPa · s or more.

In the holding step, the holding means 12 sucks and holds the other surface W2 of the wafer W with the holding surface 121, as indicated by a solid line in FIG. 2B. At this time, the swell W21 on the other surface W2 of the wafer W is corrected following the holding surface 121, and the swell W11 on the one surface W1 is also reduced.
In the flat surface forming step, the holding means 12 is lowered, and the curable resin R is sandwiched between the flat portion 11A and one surface W1 of the wafer W, as shown by a two-dot chain line in FIG. By pressing the curable resin R, a pre-curing flat surface R1 that follows the flat portion 11A is formed on the curable resin R.
In the holding release process, the suction holding of the other surface W2 of the wafer W by the holding unit 12 is released.
In the curing step, the curable resin R is cured to form a resin layer RH in which the surface opposite to the surface in contact with the one surface W1 becomes the flat surface RH1 after curing.
In the separation step, the flat portion 11A is separated from the resin layer RH.

Here, the operation of using the curable resin R having a pre-curing viscosity V of 1000 mPa · s or less will be described with reference to FIGS. 5A to 5C in which the shapes of the wafer W and the curable resin R are exaggerated and simplified.
When the holding process is performed on the wafer W as shown in FIG. 5A, the waviness W21 of the other surface W2 is corrected following the holding surface 121 as shown in FIG. The swell W11 of the surface W1 is reduced to a swell W111 indicated by a solid line. Next, when a flat surface forming step is performed, a pre-curing flat surface R1 that follows the flat portion 11A is formed in the curable resin R.

Thereafter, when the holding release process is performed, the waviness of the wafer W returns to the original state. However, as shown in FIG. 5C, when the waviness W111 of one surface W1 indicated by a two-dot chain line returns to the waviness W11 indicated by a solid line. Following this return, the wafer contact surface R2 of the curable resin R is deformed. However, the pre-curing flat surface R1 remains in close contact with the flat portion 11A. For this reason, in the portion of the curable resin R where the wafer contact surface R2 is deformed so as to be separated from the pre-curing flat surface R1, an elastic force (hereinafter simply referred to as “compressive elastic force”) F1 in the compression direction is generated. In a portion that is deformed so as to approach the flat surface R1 before curing, an elastic force (hereinafter simply referred to as “tensile elastic force”) F2 in the tensile direction is generated.

When the pre-curing viscosity V is greater than 1000 mPa · s, the deformation of the wafer contact surface R2 is difficult to be absorbed by the curable resin R, so that the compression elastic force F1 and the tensile elastic force F2 remain. In this state, after the curing process is performed and the resin layer RH is formed, when the separation process is performed, the cured flat surface RH1 of the resin layer RH is caused by the remaining elastic forces F1 and F2 as shown in FIG. 6A. It will be deformed and its flatness will be reduced.
In this embodiment, since the pre-curing viscosity V is set to 1000 mPa · s or less, the deformation of the wafer contact surface R2 is easily absorbed by the curable resin R, and the compression elastic force F1 and the tensile elastic force F2 hardly remain. As a result, when the separation step is performed after the formation of the resin layer RH, as shown in FIG. 6B, the deformation of the flat surface RH1 after the curing of the resin layer RH is suppressed, and the sufficiently flattened flat surface RH1 is flattened. The resin layer RH can be formed.

In the resin layer forming step, the resin layer RH is preferably formed so as to satisfy the following formula (1). However, the viscosity V before curing is 1000 mPa · s or less.
V / T ≦ 10 (1)
V: Viscosity before curing of curable resin R (mPa · s)
T: Thickness (μm) of the thickest part of the curable resin R after the holding release process and before the curing process (see FIG. 5C)

When V / T exceeds 10 (the expression (1) is not satisfied), the curable resin R is too thin with respect to the pre-curing viscosity V, so that the wafer contact surface R2 is deformed by the holding release process. When deformed following the movement, the elastic forces F1 and F2 may remain in the curable resin R, although the viscosity V before curing is sufficiently smaller than the case where the viscosity V is greater than 1000 mPa · s. As a result, the flatness of the flat surface RH1 after curing of the resin layer RH after the separation step is very small, but the nanotopography on the surface of the wafer W after mirror polishing described later may not be 5 nm or less.
On the other hand, when V / T is 10 or less (the expression (1) is satisfied), the thickness of the curable resin R is appropriate with respect to the pre-curing viscosity V. Even if it is deformed, this deformation can be sufficiently absorbed by the curable resin R, and residual compression elastic force F1 and tensile elastic force F2 can be suppressed. As a result, the flatness of the flat surface RH1 after curing of the resin layer RH after the separation step can be sufficiently maintained, and the nanotopography of the surface of the wafer W after mirror polishing can be reduced to 5 nm or less.

As a method for applying the curable resin to the wafer W, the other surface W2 is sucked and held by the holding surface 121 so that the one surface W1 of the wafer W faces upward, and the curable resin is set on the one surface W1. The resin is dripped and the wafer W is rotated to spin the curable resin over the entire surface W1. The spin coating method is used to place the screen plate on one surface W1, and the curable resin is placed on the screen plate and applied with a squeegee. A method in which the highly flattened flat plate 11 is pressed against the curable resin after applying the curable resin by a screen printing method, an electric spray deposition method, or the like by spraying the entire surface of the one surface W1 can be applied. The curable resin is preferably a curable resin such as a photosensitive resin in terms of ease of peeling after processing. In particular, the photosensitive resin is preferable in that it is not subjected to heat stress. In the present embodiment, a UV curable resin is used as the curable resin. Other specific curable resin materials include adhesives (such as wax).

In the first surface grinding step, the other surface W2 is surface ground using a surface grinding device 20 as shown in FIG. 2C.
First, when the wafer W is placed on the highly flattened holding surface 211 of the vacuum chuck table 21 with the flat surface RH1 after curing facing downward, the vacuum chuck table 21 sucks and holds the wafer W.
Next, as shown by a solid line in FIG. 2C, the surface plate 23 provided with the grindstone 22 on the lower surface is moved above the wafer W. Then, while rotating the surface plate 23, the vacuum chuck table 21 is rotated, and as shown by a two-dot chain line in FIG. 2C, the grindstone 22 and the other surface W2 are brought into contact with each other. Surface grinding. When the machining allowance is equal to or greater than the machining allowance minimum value P, the surface grinding is finished. By the above process, the other surface W2 becomes a flat surface from which the undulation is sufficiently removed.

In the resin layer removing step, the resin layer RH formed on one surface W1 of the wafer W is peeled off from the wafer W as shown in FIG. 3A. At this time, the resin layer RH may be chemically removed using a solvent.

In the second surface grinding step, as shown in FIG. 3B, one surface W1 is surface ground using the same surface grinding device 20 as in the first surface grinding step.
First, when the wafer W is placed on the holding surface 211 with the other flat surface W2 facing down, the vacuum chuck table 21 sucks and holds the wafer W, as shown by a solid line in FIG. 3B. In addition, the surface plate 23 moved above the wafer W is lowered while being rotated, and the vacuum chuck table 21 is rotated, so that one surface W1 is surface ground as indicated by a two-dot chain line in FIG. 3B. When the machining allowance is equal to or greater than the machining allowance minimum value P, the surface grinding is finished, so that one surface W1 becomes a flat surface from which the undulation is sufficiently removed.

Through the above resin pasting grinding process, the undulations W11 and W21 are sufficiently removed, and as shown in FIG. 3C, a wafer W in which one surface W1 and the other surface W2 are highly planarized is obtained.

Next, as shown in FIG. 1, etching is performed in order to remove a work-affected layer that occurs during chamfering or resin pasting grinding and remains on the wafer W (step S8: etching process).
Thereafter, mirror polishing including a primary polishing step (step S9) for polishing both surfaces of the wafer W using a double-side polishing device and a final polishing step (step S10) for polishing both surfaces of the wafer W using a single-side polishing device. A process is performed and the manufacturing method of a wafer is complete | finished.
Resin-grinding is performed under conditions satisfying the above formula (1), and the wafer W obtained after this mirror polishing step has a characteristic that the maximum value of nanotopography measured in a plurality of regions of 10 mm × 10 mm on the surface is 5 nm or less. Have.

[Effects of Embodiment]
As described above, by forming the resin layer RH using the curable resin R having a pre-curing viscosity V of 1000 mPa · s or less, the resin layer RH having the sufficiently flattened post-curing flat surface RH1 can be formed. . Therefore, the nanotopography on the surface of the wafer W after mirror polishing can be made sufficiently small, and a wafer capable of appropriately manufacturing a semiconductor device can be provided.

[Modification]
It should be noted that the present invention is not limited to the above-described embodiment, and various improvements and design changes can be made without departing from the scope of the present invention. The general procedure and structure may be other structures as long as the object of the present invention can be achieved.

For example, the resin pasting and grinding step may be performed under the above conditions without performing the lapping step. Even in such a case, the wafer W having the above-described characteristics can be obtained.
Further, the removal of the resin layer RH may be performed by grinding in the second surface grinding process as the resin layer removing process instead of peeling off.

Next, the present invention will be described in more detail with reference to examples, but the present invention is not limited to these examples.

[Experiment 1: Relationship between viscosity of curable resin and nanotopography]
[Create sample]
First, UV curable resins A to C were prepared. The pre-curing viscosities V of the resins A to C were 350 mPa · s, 700 mPa · s, and 1050 mPa · s as shown in FIG.
Moreover, the slice process shown in FIG. 1 was performed, and a wafer having a diameter of 300 mm and a thickness of about 900 μm was prepared.
Next, a chamfering process and a resin pasting grinding process were performed on these wafers.

In the resin layer forming step, the thickness A of the thickest portion of the curable resin (hereinafter simply referred to as “resin thickness”) T is 70 μm after the holding release step and before the curing step using the resin A. The coating process, the holding process, and the flat surface forming process were performed. After the holding release step, the resin A was cured by UV irradiation in the curing step to form a resin layer, and a separation step was performed. The value of V / T was 5, as shown in FIG. 7, satisfying the above formula (1).
Also, the resins A to C were applied to other wafers in the combinations as shown in FIG. 7 to form a resin layer having the resin thickness shown in FIG.

And the 1st surface grinding process, the resin layer removal process, and the 2nd surface grinding process were performed with respect to each wafer provided with the resin layer. In the first and second surface grinding steps, surface grinding was performed using a grinding machine (DFG8000 series) manufactured by DISCO Corporation with a machining allowance of 20 μm.
Thereafter, an etching process, a mirror polishing process, and a cleaning process were performed. In the mirror polishing step, a double-side polishing device was used as the primary polishing step, and polishing was performed in a total of 5 μm to 20 μm on both sides. A single-side polishing device was used as the final polishing step to polish less than 1 μm on only one side.
One sample was prepared for each condition.

[Evaluation]
Using an optical interference type flatness measuring device Wafersight 2 (manufactured by KLA-Tencor), the height distribution (height difference) of a plurality of 10 mm × 10 mm regions on each wafer surface was measured to obtain a nanotopography map.
The nanotopography map is obtained by filtering the measurement result of the wafer to remove a wavelength component of 20 mm or more, and then illustrating the measurement result of the nanotopography in shades of color. The nanotopography map represents the height difference of the surface in the non-adsorption state of the wafer.
In addition, binarization processing with a threshold of 5 nm was performed on the nanotopography map to generate a binarized image in which the area where the nanotopography is equal to or greater than the threshold and the height difference is large is expressed in white.
FIG. 7 shows a maximum value of nanotopography, a nanotopography map, and a binarized image in each wafer. Further, the case where a white area exists in the binarized image is described as “with pattern”, and the case where it does not exist is described as “without pattern”.

As shown in FIG. 7, the maximum value of nanotopography was almost the same between Resin A and Resin B having a pre-curing viscosity V of 1000 mPa · s or less. On the other hand, the maximum value of nanotopography in the resin C having a pre-curing viscosity V larger than 1000 mPa · s was larger than those in the resin A and the resin B. Also for the binarized image, it was confirmed that the white area of the resin C was larger than the resin A and the resin B.

From the above, it was confirmed that the nanotopography on the wafer surface after mirror polishing can be made sufficiently small by forming the resin layer using a curable resin having a viscosity V before curing of 1000 mPa · s or less.

[Experiment 2: Relationship between nanotopography of wafer and unevenness of flat surface after curing of resin layer]
Regarding the sample in which the resin layer having the resin thickness T of 25 μm (V / T = 14) was formed using the resin A in Experiment 1, the relationship between the nanotopography in the radial direction of the wafer and the unevenness of the flat surface after curing was examined. .
For measurement of unevenness on the flat surface after curing, a linear gauge (manufactured by Mitutoyo Corporation, model LGF) was used, and the unevenness on the flat surface after curing was measured in a linear range. For the measurement of the nanotopography profile, the above-described flatness measuring device Wafersight 2 was used, and the surface after mirror polishing was measured in the above range. Each measurement was performed in a range passing through the center of the wafer.
The measurement results are shown in FIG.
As shown in FIG. 8, it was found that the shape of the nanotopography and the flat surface after curing were very similar. From this, it was found that the uneven shape of the flat surface after curing was transferred to the wafer as waviness.

[Experiment 3: Relationship between V / T and nanotopography]
As shown in FIG. 7, even when the pre-curing viscosity V is 1000 mPa · s or less, a pattern may be generated in the binarized image depending on the value of V / T.
Therefore, the relationship between V / T and nanotopography was examined.

First, the same UV curable resins A and B as used in Experiment 1 and a wafer were prepared. As shown in Table 1 below, three samples were prepared under the same conditions as in Experiment 1 except that the resin layers were formed by combining the resins A and B and the resin thickness T. Table 1 shows V / T for each wafer.
Thereafter, as in Experiment 1, the measurement results of nanotopography of each wafer were obtained. The relationship between V / T and the maximum value of nanotopography is shown in FIG.

As shown in FIG. 9, it was confirmed that in both cases of resins A and B, the maximum value of nanotopography was 5 nm or less when V / T was 10 or less.
From the above, the maximum value of nanotopography after mirror polishing is as high as 5 nm or less by forming a resin layer under conditions where the viscosity V before curing is 1000 mPa · s or less and V / T is 10 or less. It was confirmed that a quality wafer could be obtained.

DESCRIPTION OF SYMBOLS 11 ... Flat plate (flat surface formation means), 11A ... Flat part, 12 ... Holding means, R ... Curable resin, R1 ... Flat surface before hardening, RH ... Resin layer, W ... Wafer, W1 ... One side, W2 ... The other side.

Claims

A resin layer forming step of forming a resin layer on one surface of a wafer cut from a single crystal ingot or a lapped wafer;
A first surface grinding step of holding the one surface via the resin layer and surface grinding the other surface of the wafer;
A resin layer removing step of removing the resin layer;
A second surface grinding step of holding the other surface and surface grinding the one surface;
The resin layer forming step includes
A holding step of sucking and holding the other surface of the wafer by a holding means;
A flat surface in which a curable resin having a viscosity of 1000 mPa · s or less is sandwiched between a flat portion of the flat surface forming means and one surface of the wafer, and a pre-cured flat surface that follows the flat portion is formed on the curable resin. Forming process;
A holding release step for releasing suction holding of the other surface;
A curing step of curing the curable resin to form the resin layer;
And a separation step of separating the flat portion from the resin layer.
In the manufacturing method of the wafer according to claim 1,
The said resin layer formation process forms the said resin layer so that the following formula | equation (1) may be satisfy | filled, The manufacturing method of the wafer characterized by the above-mentioned.
V / T ≦ 10 (1)
V: Viscosity before curing of the curable resin (mPa · s)
T: Thickness (μm) of the thickest portion of the curable resin after the holding release step and before the curing step
A wafer characterized in that the maximum value of nanotopography measured in a plurality of regions of 10 mm × 10 mm on the surface is 5 nm or less.