JP7596191B2 - Method for modifying the surface of silicon wafer - Google Patents

Description

The present invention relates to the repair of surface defects, which are layers that have been altered due to processing on the surface of a silicon wafer, and in particular to a method for modifying the surface of a silicon wafer using laser heat treatment after alkaline etching (AE).

Semiconductor wafers such as silicon wafers used in the production of semiconductor devices are surface-treated using mechanical processes such as cutting, grinding, lapping, and polishing. However, a process-induced deterioration layer is formed on the surface and inside of the wafer, and some of the process-induced deterioration layers contain microcracks. These internal cracks are mainly removed by chemical and mechanical methods such as etching and chemical mechanical polishing (CMP).

For example, Patent Document 1 describes a method of removing distortions in a silicon wafer that has been subjected to chamfering processing by alkaline etching using an aqueous potassium hydroxide solution, an aqueous sodium hydroxide solution, or the like, to remove distortions that have occurred in the silicon wafer due to processing up to the previous process.

Patent Document 2 also describes how, after grinding such as chamfering, the damaged layer on the surface of the silicon wafer is repaired and the roughness is smoothed efficiently and effectively by irradiating it with a pulsed laser.

JP 2018-101698 A JP 2020-131218 A

In the above-mentioned conventional technology, the alkaline etching described in Patent Document 1 strongly manifests the anisotropy of the silicon wafer, and it is difficult to prevent the cracks caused by this. In addition, there is a risk that the existing peripheral edge grinding, etching, and chemical mechanical polishing (CMP) cannot completely remove internal cracks, etc. Furthermore, the internal cracks, etc. progress as cracks and cause damage, resulting in a decrease in the yield of surface processing of the silicon wafer.

The technique described in Patent Document 2 makes it possible to repair damage such as grinding marks caused by grinding and perform flattening treatment. However, the technique described in Patent Document 2 has limitations in terms of laser irradiation conditions according to the surface condition after grinding.

In particular, when flattening a surface after alkaline etching (AE) processing by laser irradiation, if an appropriate irradiation method is not used, internal damage such as dislocations will occur, resulting in reduced quality and a lower yield in subsequent processes. Furthermore, there is a limit to the improvement of the quality of surface modification for internal cracks, surface roughness (especially roughness caused by anisotropy), etc.

The object of the present invention is to solve the problems of the conventional technology described above, and to perform suitable heat treatment using a laser on the surface of a silicon wafer after etching, thereby eliminating the effects of processing stress and modifying the surface to a uniform and flat surface. This improves strength and increases the yield in subsequent processes. In particular, the object is to perform suitable laser treatment on the wafer edge, which has a shape in which the crystal orientation changes.

To achieve the above objective, the present invention is a method for modifying the surface of a silicon wafer using laser heat treatment, in which after alkaline etching treatment, a cumulative irradiation energy corresponding to the crystal orientation of the irradiated area is determined and a nanosecond pulse laser is irradiated.

In addition, in the above-mentioned method for modifying the surface of a silicon wafer using laser heat treatment, it is desirable to determine irradiation conditions so that the cumulative irradiation energy at the irradiated location is equal to or less than a predetermined threshold value.

Furthermore, in the above-mentioned method for modifying the surface of a silicon wafer using laser heat treatment, it is preferable that after the alkaline etching process, the irradiation is performed by changing at least one of the energy density, the scan pitch, and the number of irradiations in accordance with the curvature of the irradiation area.

Furthermore, in the above-mentioned method for modifying the surface of a silicon wafer using laser heat treatment, it is desirable to use CW (continuous wave) laser irradiation in addition to the nanosecond pulse laser irradiation after the alkaline etching treatment.

Furthermore, in the above-mentioned method for modifying the surface of a silicon wafer using laser heat treatment, it is desirable to divide the irradiation area having a curvature into sections, change the angle of incidence, irradiate perpendicularly to the irradiation surface of each section, and increase the scan pitch.

Furthermore, in the above-mentioned method for modifying a silicon wafer surface using laser heat treatment, when the irradiated portion is a curved surface, it is preferable that a ratio of the irradiation energy on the curved surface to the irradiation energy on a flat surface is defined as C _i , and the scan pitch of the curved surface is defined as the scan pitch of the flat surface × 1/C _i .

Furthermore, in the above-mentioned method for modifying the surface of a silicon wafer using laser heat treatment, it is desirable that the scan pitch is set to the nanosecond pulse laser spot diameter D x 0.5 to 0.7, and that the irradiation position i where the first irradiation energy E1 is at its peak and the irradiation position i' where the second irradiation energy E2 is at its peak are set to i + the scan pitch/2, so that the cumulative irradiation energy at the irradiated location is equal to or less than a predetermined threshold value.

Furthermore, in the above-mentioned method for modifying the surface of a silicon wafer using laser heat treatment, it is desirable to set the scan pitch to be smaller than 1/2 the spot diameter D of the nanosecond pulse laser, set the number of irradiations to multiple, and set the cumulative irradiation energy of the irradiated area to be equal to the sum of the non-overlapping area and the overlapping area.

According to the present invention, after alkaline etching, the cumulative irradiation energy corresponding to the crystal orientation of the irradiated area is determined and then the nanosecond pulse laser is irradiated, which eliminates the effects of processing stress and modifies the surface to a uniform one, thereby improving strength and improving the yield in subsequent processes.

A perspective view showing an example of the distribution of crystal orientations on a silicon wafer surface. Upper half cross-sectional view showing an example of crystal orientation distribution at the wafer edge Flowchart showing the laser irradiation procedure FIG. 13 shows the difference between the flat portion and the R portion of the wafer edge in terms of laser irradiation. Graph showing the relationship between spot diameter D and laser radiation intensity I(r) Graph explaining how to determine spot diameter D and scan pitch SP according to an embodiment of the present invention. Graph for explaining how to determine spot diameter D and scan pitch SP according to another embodiment

Figure 1 is a perspective view showing an example of the distribution of crystal orientation on the surface of a silicon wafer 1, and Figure 2 is a cross-sectional view of the upper half showing an example of the distribution of crystal orientation at the edge of the wafer. Note that Figure 2 shows a cross-section of the upper half of the circled portion of Figure 1 (Si (100) plane). In general, even for the surface of the exact same material, the properties differ depending on the direction of the plane along which the crystal is cut. Crystal planes are specified by Miller indices; for example, a cut surface of a silicon single crystal cut along a lattice plane with Miller indices (111) is called a Si (111) plane with a crystal orientation.

For example, the crystal structure of the Si(100) surface is one in which atoms are packed almost uniformly in the depth direction, whereas the Si(111) surface is layered in the depth direction. Therefore, it is believed that the propagation of kinetic energy in the depth direction due to lattice vibration, i.e., heat conduction in the depth direction, is easier on the Si(111) surface. In other words, it can be said that the Si(111) surface is more likely to melt to a greater depth than the Si(100) surface due to its poorer heat conduction. The order of magnitude is Si(110) surface > Si(100) surface > Si(111) surface.

In addition, the propagation speed of the laser shock wave differs between the Si(100) surface and the Si(111) surface. This is because the atomic arrangement on the Si(111) surface is layered in the depth direction (Si(111) direction), so lattice vibrations are more easily transmitted. It is believed that heat conduction on the scale of several microns is dominated by the propagation of lattice vibrations (waves), with ballistic heat conduction occurring due to phonons (quanta of lattice vibrations in crystals).

Compared to the Si(100) and Si(110) surfaces, the Si(111) surface has a higher thermal conductivity due to the heat input from the laser, and the heat (lattice vibration) is transferred quickly, so that the deformation (contraction) cannot keep up during the subsequent cooling process, resulting in cracks. In other words, distortion occurs due to thermal stress, and internal damage such as dislocations occurs.

As shown in Fig. 1, the distribution of crystal orientations on the surface of a silicon wafer 1 is such that the same orientations are distributed at every 90 degree rotation angle in three dimensions. As shown in Fig. 2, even if the end surface 2 is a Si(100) surface, the distribution of crystal orientations on the wafer edge changes at the R portion 3, and for example, the Si(111) surface appears at 36 degrees from the horizontal, and becomes a Si(100) surface on the top surface 4. The surface state of each crystal orientation after alkaline etching is as follows:
Si (110) surface: Etch pits are widely distributed and the surface is rough overall.
Si(100), Si(111) surface: some etch pits scattered
Etch pits are corrosion holes that appear on the surface during etching, and correspond to the points where dislocations, which are linear crystal lattice defects, intersect with the crystal surface. Etch pits occur because the area near the lattice defect is more chemically reactive or corroded than other areas.

In one embodiment, in a method for modifying the surface of a silicon wafer 1 using laser heat treatment, irradiation conditions are determined according to the crystal orientation and shape so that the cumulative value of the laser irradiation energy (cumulative irradiation energy) at the irradiated location is equal to or less than a predetermined threshold value in order to reduce the effects of thermal stress. In another embodiment, a CW (continuous wave) laser is also used to prevent sudden temperature changes, reduce the load caused by repeated thermal stress, and suppress the occurrence of dislocations.

Figure 3 is a flow chart showing the procedure for laser irradiation. In step S1, the surface condition (roughness, waviness) of the silicon wafer 1 after grinding is measured and mapped to obtain distribution data corresponding to the shape of the silicon wafer 1. It is preferable to measure the surface condition and/or shape of the silicon wafer 1 using a non-destructive method. The non-destructive method may be any one selected from image measurement means, capacitance change measurement means, droplet contact angle measurement means, Raman spectroscopy measurement means, surface roughness meter, acoustic measurement means, eddy current characteristic measurement means, reflectance measurement means, X-ray Lang method measurement means, electron beam diffraction measurement means, SEM measurement means, etc., or a combination of two or more of these may be used.

Next, in step S2, the internal state of the irradiated area is inspected by Raman spectroscopy. In other words, in step S2, the change in the crystalline state inside the silicon wafer 1 is detected by measuring the distortion from the Raman shift. Here, those with significant internal damage are excluded.

In step S3, irradiation conditions are determined and irradiation is performed. In step S3-1, irradiation conditions are determined according to the crystal orientation and shape. In other words, conditions must be determined that flatten the surface of the irradiated area, which has a different surface condition depending on the orientation and shape, while not causing internal damage. As already mentioned, the surface condition after alkaline etching (roughness due to the distribution of etch pits) and thermal conductivity differ for Si(110), Si(100), and Si(111) surfaces, and the irradiation energy required for the desired flattening differs.

The required irradiation energy has a magnitude relationship of Si(110)>Si(100)>Si(111), and irradiation conditions are determined according to the crystal orientation. For example, the energy required for a Si(110) surface to flatten a surface of several hundred nanometers to a roughness of a few nanometers without generating internal dislocations by laser irradiation is approximately 1.3 times that required for a Si(100) surface. Note that the Si(111) surface requires irradiation energy that is approximately the same as or 0.7 times that of a Si(100) surface.

In terms of laser conditions according to the shape, for irradiation points that can be considered flat, such as the inclined surface 5, the end surface 2, and the top surface 4, it is desirable for the laser to be nearly perpendicular to the irradiation surface, with an incidence angle of 10 to 15 degrees or less. For the curved portion 3, it is difficult to irradiate the entire surface with a laser perpendicularly, and the effective laser radiation intensity of the laser is lower than for flat surfaces.

The laser irradiation is a nanosecond pulsed laser, which melts the amorphous layer of the processing-induced deterioration layer at nanosecond speeds. The melting caused by the nanosecond pulsed laser irradiation promotes recrystallization (epitaxial growth) with aligned crystal orientation, eliminating crystal defects caused by mechanical processing. In addition, the amorphous silicon layer has a strong absorption of light with a wavelength of 532 nm.

Therefore, the nanosecond pulse laser has a wavelength λ of 355, 532, or 785 nm, for example, λ=532 nm, and the pulse irradiation time is preferably within a range of 3 to 4 nanoseconds.The pulse width, energy per pulse, and energy density are preferably 0.5 to 30 μJ ^and 0.125 to 7.5 J/ ^cm2 .

In addition, after the alkaline etching process, in areas with large etch pits or waviness, the nanosecond pulsed laser is irradiated together with a continuous wave (CW) laser. For example, before irradiating with the nanosecond pulsed laser, a continuous wave (CW) laser with a wavelength of 1080 nm, which has a high absorption rate for _SiO2, is irradiated.

Furthermore, the nanosecond pulse laser is irradiated by changing at least one of the energy density, scan pitch SP, and number of irradiations in response to the curvature of the irradiated area. This allows the irradiation conditions to be adjusted according to the crystal orientation and shape, and the surface of the silicon wafer 1 is modified more uniformly.

In step S3-2, the laser is irradiated by overlapping scanning with a predetermined spot diameter D (Gaussian laser spot diameter) so that the cumulative irradiation energy is averaged regardless of the irradiation position and is below the threshold. In other words, in step S3-2, the area to be irradiated is scanned overlappingly multiple times to uniformly flatten the entire edge without any gaps, resulting in cumulative irradiation. Cumulative irradiation with a predetermined spot diameter D means that there are areas that are irradiated more than twice, i.e., there are overlapping areas of irradiation.

In the overlapping areas, there is a risk that the cumulative irradiation energy will exceed the threshold at which dislocations do not occur. Therefore, important conditions for reducing waviness in the laser-irradiated areas are the spot diameter D, the scan pitch SP, and multiple irradiations, for example the phase between the first and second irradiations in two irradiations, and the cumulative irradiation energy is set below the threshold at which dislocations do not occur and averaged.

In step S4, a process of checking whether the surface of the silicon wafer 1 has been repaired is performed in the same manner as in step S2. If the surface has been repaired, the laser irradiation is terminated, and if the surface has not been repaired, the laser irradiation is resumed or continued until repair is confirmed.

Figure 4 shows the difference between the flat portion and the R portion 3 of the wafer edge portion in laser irradiation. Figure 4 shows an example of step S3-1. The inclined surface 5, which is the flat portion, is approximately perpendicular to the irradiation surface, but the R portion 3 is an irradiation point with curvature, so it is difficult to irradiate all irradiation points with the laser perpendicularly. Therefore, as shown in Figure 4, the irradiation points with curvature are divided (discretized) into appropriate sections, the angle of incidence is changed, and the irradiation surface of each section is irradiated perpendicularly. Also, for irradiation points with curvature, it is desirable to increase the scan pitch SP (or reduce the cumulative number of irradiations) to shorten the throughput and improve efficiency.

5 is a graph showing the relationship between spot diameter D and laser radiation intensity I(r), and shows the relationship between spot diameter D (radial distance r from the beam center = D/2) and laser radiation intensity I(r) when the irradiated portion is a curved surface and is R portion 3. The vertical axis is laser radiation intensity I(r), the horizontal axis is radial distance r (= D/2) from the beam center, and R _i is the radius of curvature of the irradiated portion of R portion 3. Since angle α _O at spot diameter D is D/R _i , angle α(r) at any r is α _O /2 × r/(D/2) = r/R _i .

The relationship between the irradiation energy when the irradiation surface is a flat surface and when it is a curved surface is as follows: Since the laser radiation intensity I(r) is a Gaussian laser, it can be expressed as (Equation 1).
When the irradiation surface is flat, the irradiation energy E ₀ is expressed by (Equation 2) with the integral interval being from −D/2 to D/2.
however,
I ₀ : Peak radiant intensity at the center of the beam
r: The radial distance from the beam center.
The ratio of the irradiation energy on a curved surface to the irradiation energy on a flat surface, that is, (irradiation energy on a curved surface)÷(irradiation energy on a flat surface) C _i is given by α(r)=r/R _i and is expressed by (Equation 3).
Similarly, when the curvature changes (R _i ≠ R _i+1 ), E _i,1 is given by (Equation 4) with C _i,i+1 = (C _i + C _i+1 )/2 between adjacent irradiation points.
however,
_Ei ,1: the jth irradiation energy Ei,j actually received at the irradiation position i (i') on the curved surface
E ₁ , E ₂ , E _n : nth irradiation energy
It is.

From the above, when the curvature of the R portion 3 is constant,
Scan pitch of R portion 3 = scan pitch of flat portion × 1/C _i
If the curvature changes,
Scan pitch of R portion 3 = scan pitch of plane portion × 1 / C _{i,i + 1}
It is preferable to set the above.

Figures 6 and 7 are graphs explaining how to determine the spot diameter D and scan pitch SP in step S3-2. Figure 6 shows the case where the scan pitch SP is the spot diameter D x 0.5 to 0.7, and Figure 7 shows the case where the scan pitch SP is the spot diameter D x to 0.5 (smaller than 1/2 the spot diameter D), with the vertical axis representing the irradiation energy and the horizontal axis representing the irradiation position. In order to uniformly flatten the entire surface, the irradiation energy needs to be averaged regardless of the irradiation position, and the cumulative irradiation energy needs to be below a threshold value.

In Figure 6, E1 is the peak of the first irradiation energy, and the cumulative irradiation energy (point b indicated by arrow) of the overlapping portion between irradiation position i where this peak occurs and the next irradiation position (i+1) is slightly smaller than E1 even when they are overlapped. The second irradiation is then set to the irradiation energy of E2. (However, if the scan pitch SP is such that the cumulative irradiation energy of the overlapping portion is approximately the same as E1, then it is sufficient to irradiate only E1.) In this case, the irradiation position i' where E2 peaks is shifted in phase so that it becomes i+SP/2.

As a result, the cumulative irradiation energy at points a and b indicated by the arrows is equalized and averaged. The cumulative irradiation energy at the irradiated area is set to a predetermined threshold or less that does not cause dislocations even if the number of irradiations is multiple, e.g., two times (E1+E2), or many times (E1+E2+...). Furthermore, in order to reduce throughput (processing time), it is desirable to perform the cumulative irradiation no more than about two times.

In FIG. 7, E1 is the peak of the first irradiation energy, and the second irradiation is performed with an irradiation energy of E2. The scan pitch SP is smaller than 1/2 of the spot diameter D. The cumulative irradiation energy of E1 at point b is greater than E1 because of overlap, and similarly, the cumulative irradiation energy of E2 at point b is greater than E2.

The cumulative irradiation energy due to irradiation with E1 and E2 is then averaged so that it is equal to the sum of E1+E2 at point a, which is the non-overlapping portion, and point b, which is the overlapping portion. As in FIG. 6, the cumulative irradiation energy at the irradiated portion is set to a predetermined threshold value or less at which no dislocation occurs even if it is E1+E2 or E1+E2+....

REFERENCE SIGNS LIST 1: silicon wafer 2: end face 3: R portion 4: upper surface 5: inclined surface SP: scan pitch

Claims (8)

A method for modifying a surface of a silicon wafer using laser heat treatment, comprising the steps of:
A method for modifying a surface of a silicon wafer, comprising: determining a cumulative irradiation energy corresponding to a crystal orientation of an irradiated portion after an alkaline etching process; and irradiating the portion with a nanosecond pulse laser. The method for modifying the surface of a silicon wafer according to claim 1, characterized in that the irradiation conditions are determined so that the cumulative irradiation energy at the irradiated location is equal to or less than a predetermined threshold value. The method for modifying the surface of a silicon wafer according to claim 1 or 2, characterized in that after the alkaline etching process, irradiation is performed by changing at least one of the energy density, the scan pitch, and the number of irradiations in response to the curvature of the irradiation area. The method for modifying the surface of a silicon wafer according to any one of claims 1 to 3, characterized in that after the alkaline etching process, irradiation with the nanosecond pulsed laser is combined with irradiation with a CW (continuous wave) laser. The method for modifying the surface of a silicon wafer according to claim 3, characterized in that the irradiation area having a curve is divided into sections, the angle of incidence is changed, the irradiation surface of each section is irradiated perpendicularly, and the scan pitch is increased. 6. The method for modifying a surface of a silicon wafer according to claim 3 or 5 _, characterized in that, when the irradiated portion is a curved surface, a ratio of the irradiation energy on the curved surface to the irradiation energy on a flat surface is defined as C _i , and the scan pitch of the curved surface is defined as the scan pitch of the flat surface × 1/C i. The surface modification method of silicon wafers according to claim 6, characterized in that the scan pitch is set to 0.5 to 0.7 times the spot diameter D of the nanosecond pulse laser, and the irradiation position i where the first irradiation energy E1 is at its peak is set to i + the scan pitch/2 for the irradiation position i' where the second irradiation energy E2 is at its peak, so that the cumulative irradiation energy at the irradiated location is equal to or less than a predetermined threshold value. The method for modifying the surface of a silicon wafer according to claim 6, characterized in that the scan pitch is smaller than 1/2 the spot diameter D of the nanosecond pulse laser, the number of irradiations is multiple, and the cumulative irradiation energy of the irradiated area is equal to the sum of the non-overlapping area and the overlapping area.