Semiconductor wafer generation method

Abstract

The invention relates to a semiconductor wafer generation method, which comprises the following steps of: firstly, setting laser scanning times, and then respectively setting a scanning path and a dot spacing of each laser scanning; determining a laser scanning speed and a laser pulse repetition frequency based on a calculation rule preset for each laser scanning, and determining a corresponding modified point diameter, laser pulse energy and an offset distance of a laser focus relative to a preset stripping surface; the pulse laser is focused on a preset stripping surface in the crystal ingot or below the preset stripping surface, and n times of laser scanning are carried out, so that modified points and cracks are sequentially formed on the preset stripping surface; and stripping the crystal ingot along the preset stripping surface to obtain a wafer and a residual crystal ingot. The method not only can eliminate burrs at the edge of the wafer, but also can further reduce the material loss of laser stripping, and has the advantages of simple process conditions, good modified layer morphology, low wafer stripping difficulty and suitability for large-scale industrial application.

Description

Technical Field

[0001] The present invention relates to the technical field of wafer processing, and in particular to a method for producing a semiconductor wafer. Background Art

[0002] In recent years, with the rapid development of optoelectronics and microelectronics industries, aerospace, aviation, machinery, light industry, chemical industry and other industries have also been moving towards integration and miniaturization, which requires semiconductor devices to be more and more integrated, more and more complex in function and smaller in size. Silicon carbide substrate is the core material of the newly developed wide bandgap semiconductor. The devices made of it have the characteristics of high temperature resistance, high voltage resistance, high frequency, high power, and radiation resistance. It has the advantages of fast switching speed and high efficiency, which can greatly reduce product power consumption, improve energy conversion efficiency and reduce product size. Wafer processing is a key process in the manufacture of silicon carbide substrates. The quality of slicing directly affects the performance of silicon carbide substrates.

[0003] Laser slicing is a laser technology that separates silicon carbide ingots into individual wafers. The process uses a precision laser beam to form a modified layer inside the ingot, so that the wafers can be accurately separated along the laser scanning path with a slight external force. Laser scanning is the process of forming the modified layer. During this process, the laser focuses at a predetermined depth inside the silicon carbide ingot, inducing the formation of microcracks that extend along the peeling surface. When uniformly distributed microcracks exist in the material, the stress field will produce a concentration effect around the microcracks. When the peeling force is applied mechanically, the stress will induce the expansion of cracks to the specified position due to the presence of the modified layer, thereby completing the peeling of the wafer. The above method can greatly reduce material loss during the slicing process and improve wafer production efficiency.

[0004] A Chinese patent with authorization announcement number CN107790898B discloses a method for producing a SiC wafer, comprising the following steps: a peeling surface producing step, in which the following separation layer forming process is performed: a focal point of a pulsed laser beam having a wavelength that is transparent to SiC is positioned at a depth from the first surface that is equivalent to the thickness of the wafer to be produced, and while the single crystal SiC ingot and the focal point are relatively processed and fed in a first direction perpendicular to a second direction in which the deviation angle is formed, the single crystal SiC ingot is irradiated with a pulsed laser beam, thereby forming a separation layer composed of a modified layer and cracks; and a wafer producing step, in which a portion of the single crystal SiC ingot is peeled off with the peeling surface as an interface to produce a SiC wafer, in which the processing feed includes forward movement and return movement, and the forward movement causes the focal point to move relatively from one end of the single crystal SiC ingot to the other end. The return movement causes the focal point to relatively move from the other end to the one end while the depth of the focal point is maintained at the same depth as that of the focal point in the forward movement, thereby tracing back the separation layer that has been formed. In the forward movement, the modified layer initially formed is formed by the focal point. The modified layer formed subsequent to the initial modified layer is formed at a position slightly shallower than the focal point. A rise of the modified layer occurs at the one end of the single crystal SiC ingot where the irradiation of the pulsed laser light starts. When the modified layer reaches a depth inside the single crystal SiC ingot at which the power density of the pulsed laser light becomes a prescribed value, a modified layer is formed near the focal point and at a depth at which the power density is the prescribed value. In the return movement, a modified layer is formed near the focal point and at a depth at which the power density is the prescribed value from the other end of the single crystal SiC ingot to the one end.

[0005] The above-mentioned prior art solution has the following defects: in the peeling surface generation process, the above method avoids the burr problem at the laser scanning end face of the single crystal SiC ingot by means of laser reciprocating scanning to reduce material loss; wherein, the burr at the laser scanning end face is mainly caused by the self-organization change of the modified layer during the formation process, that is, after the laser enters the single crystal SiC ingot, the modified point generated by the first laser pulse is located near the geometric focus of the focusing lens, and as the laser moves along the scanning direction, the modified point gradually rises until the overlap rate of adjacent modified points reaches a balance with the laser power density, and a stable modified layer is formed at a predetermined depth; refer to Figure 1This also causes the modified layer to rise in an area about tens of μm away from the end of one side of the single crystal SiC ingot during the forward movement. After actual testing, the lifting amplitude is about 40~100μm. The above method only reduces the burrs left on the wafer to a certain extent, but leads to the problem that the modified points at the laser scanning end face gradually rise toward the peeling surface. Therefore, after the wafer is peeled off, it is still necessary to grind or thin a considerable thickness of material to completely remove the traces left by the laser scanning at the end face, and the material loss is as high as 30%.

[0006] A Chinese patent with publication number CN115635183A discloses a method for laser stripping a workpiece, comprising: using a pulse laser with a short pulse width to focus on a preset stripping surface inside the workpiece to form a modified point on the preset stripping surface inside the workpiece, using a pulse laser with a long pulse width to focus on the modified point to form a modified zone at the modified point and a crack extending in the radial direction of the preset stripping surface; dividing the workpiece into a first workpiece unit and a second workpiece unit along the preset stripping surface.

[0007] The above-mentioned prior art solutions have the following defects: the above method uses a pulse laser with a short pulse width to form a modified point on a preset peeling surface inside the silicon carbide, and then uses a pulse laser with a long pulse width to focus on the modified point generated by the short pulse laser and form a modified area, thereby forming a crack extending in the radial direction of the preset peeling surface, so that the workpiece can be easily peeled off. However, this requires that the pulse lasers with short and long pulse widths must not only be strictly aligned with the points, but also have strictly synchronized timing, which has poor actual operability and is not conducive to large-scale industrial applications. Summary of the invention

[0008] The problem to be solved by the present invention is to provide a semiconductor chip production method in view of the above-mentioned deficiencies existing in the prior art, which solves the modification point climbing phenomenon existing in the existing laser reciprocating scanning method and the problem of harsh process conditions existing in the pulse laser connection scanning with short pulse width and long pulse width. It can not only eliminate the burrs on the edge of the chip, but also further reduce the material loss of laser peeling. It has the advantages of simple process conditions, good morphology of the modified layer, low difficulty of chip peeling, and suitability for large-scale industrial applications.

[0009] The above-mentioned object of the present invention is achieved through the following technical solutions:

[0010] A method for producing a semiconductor wafer comprises the following steps:

[0011] S1 first sets the number of laser scans to n, where n is an integer ≥ 2, and then sets the scanning path of each laser scan and the point spacing P between two adjacent modified points on the scanning path;

[0012] S2 determines the laser scanning speed V and laser pulse repetition frequency F required to achieve the point spacing P based on the preset calculation rules for each laser scan, and determines the corresponding modified point diameter D, and then determines the laser pulse energy E required to achieve the modified point diameter D, and the offset distance S of the laser focus relative to the preset peeling surface;

[0013] S3: focusing a pulsed laser on a preset peeling surface inside the ingot or below the preset peeling surface, and performing n laser scans to sequentially form modified points on the preset peeling surface, and forming an overlapping area between the modified points formed by at least two laser scans, so as to form a crack extending laterally along the preset peeling surface in the overlapping area;

[0014] S4: peeling the crystal ingot along the preset peeling surface to obtain a wafer and a remaining crystal ingot.

[0015] Specifically, in the present invention, the meaning of the modified point refers to the independent or continuous laser action point formed on the preset peeling surface. Correspondingly, the point spacing P, the line distance L and the modified point diameter D also refer to the point spacing between two adjacent modified points at the preset peeling surface position, the line distance between two scanning segments spaced from each other, and the modified point diameter.

[0016] Furthermore, in S1, the scanning path of each laser scanning is one or a combination of the following: a line-by-line scanning path, a grid-interlaced scanning path, a concentric circle scanning path, and a spiral line scanning path.

[0017] Furthermore, in S1, the line distance L between two scanning sections spaced apart from each other on the scanning path is controlled to be 0.05-1.00 mm. The scanning sections on the scanning path are arranged with equal or unequal spacing.

[0018] Furthermore, in the step S1, the point spacing P is controlled to be 0.5-50 μm.

[0019] Furthermore, in S2, the preset calculation rule for the 1st to nth laser scans is: n′ =V n′ / F n′ , P n′ ≥0.7D n′ , E n′ ≥1μJ, S n′ 0~5μm; among them, P n′ V is the distance between two adjacent modified points on the 1st to nth scanning path, n′ is the laser scanning speed from the 1st to the nth time, F n′ is the repetition frequency of the 1st to nth laser pulses, E n′ is the energy of the 1st~n′th laser pulse, Sn′ is the offset distance of the 1st to nth laser focus relative to the preset peeling surface, and n' is an integer ≥1.

[0020] Furthermore, in S2, the preset calculation rule for the n′+1~n″th laser scan is: n′+1 =V n′+1 / F n′+1 , P n′+1 <D n′+1 , E n′+1 ≥5μJ, S n′+1 0~20μm; among them, P n′+1 is the point distance between two adjacent modified points on the n′+1~n″th scanning path, V n′+1 is the laser scanning speed of the n′+1~n″th time, F n′+1 is the laser pulse repetition frequency of the n′+1~n″th time, E n′+1 is the laser pulse energy of the n′+1~n″th time, S n′+1 It is the offset distance of the laser focus from the n′+1 to n″th time relative to the preset peeling surface, where n″ is an integer ≥2.

[0021] Furthermore, in S2, 0.2D is controlled n′+1 <P n′+1 <0.6D n′+1 .

[0022] Furthermore, in S3, the laser wavelength range λ of each laser scan is controlled to be 780-2500 nm, and the laser pulse width Δt is controlled to be 10 fs-100 ns.

[0023] Furthermore, in S3, the pulse laser is a pulse train composed of a plurality of sub-pulses, wherein the time interval between two adjacent sub-pulses does not exceed 100 ns.

[0024] Furthermore, in S3, the pulse laser is pre-corrected for aberration before the 1st to n′th laser scans.

[0025] Furthermore, in S3, the longitudinal thickness of the modified point of the 1st to n'th laser scans is controlled to be 5 to 40 μm.

[0026] Furthermore, in the S3, a first laser, a first beam combining and expanding unit, an aberration correction unit, a first workpiece reflector, a first objective lens, and a crystal ingot are sequentially arranged on the pulse laser emission path of the 1st to n′th laser scans; wherein, at least two first lasers are arranged in parallel, the first beam combining and expanding unit includes a first beam combining mirror arranged on the pulse laser emission path of one of the first lasers, first laser reflectors respectively arranged between the pulse laser emission paths of the remaining first lasers and the first beam combining mirror, and a beam expanding mirror arranged on the pulse laser emission path of the first beam combining mirror, and the aberration correction unit is configured as an adaptive optical element, a diffractive optical element, or a deformable lens.

[0027] Furthermore, in S3, the aberration correction process includes shaping the wavefront of the incident light beam of the pulsed laser through an adaptive optical element, compensating the aberration of the focusing lens of the pulsed laser at a specific depth through a diffractive optical element, or presetting the aberration and compensating the aberration of the focusing lens of the pulsed laser at a specific depth by adjusting the deformable lens. The adaptive optical element may be, for example, a spatial light modulator and a digital micromirror array, the diffractive optical element may be, for example, a beam shaper, a beam splitter, a diffractive cone lens, a spiral phase plate, a homogenizer, and a multi-focus telephoto depth, and the deformable lens may be, for example, a lens having a movable lens or lens group inside and / or outside.

[0028] Furthermore, in S3, the pulse laser is beam shaped before the n′+1~n″th laser scan.

[0029] Furthermore, in S3, the diameter D of the modified spot of the n′+1~n″th laser scan is controlled. n′+1 15~100μm.

[0030] Furthermore, in the S3, a second laser, a second beam combining unit, a beam shaping unit, a second workpiece reflector, a second objective lens, and a crystal ingot are sequentially arranged on the pulse laser emission path of the n′+1~n″th laser scan; wherein, at least two second lasers are arranged in parallel, the second beam combining unit includes a second beam combining mirror arranged on the pulse laser emission path of one of the second lasers, and second laser reflectors respectively arranged between the pulse laser emission paths of the remaining second lasers and the second beam combining mirrors, and the beam shaping unit is set as an adaptive optical element or a diffractive optical element.

[0031] Furthermore, in S3, the beam shaping process includes shaping the incident beam of the pulsed laser through an adaptive optical element, shaping the incident beam of the pulsed laser through a diffractive optical element, or increasing the beam quality factor of the incident beam of the pulsed laser. The adaptive optical element may be, for example, a spatial light modulator and a digital micromirror array, and the diffractive optical element may be, for example, a beam shaper, a beam splitter, a diffractive aconic lens, a spiral phase plate, a homogenizer, and a multi-focus telephoto depth.

[0032] Specifically, the number of laser scans is first set to 2, and then the scanning paths of the two laser scans are respectively set to be line-by-line scanning paths, and the scanning paths of the two laser scans are completely parallel and overlapped, parallel and overlapped at intervals, or cross-vertically overlapped. The specific implementation method is as follows.

[0033] First, during the first laser scanning process, the pulsed laser needs to be focused on the preset peeling surface inside the ingot, that is, the specific depth inside the ingot where the wafer is to be generated, and the drive module drives the ingot and the laser focus to move relative to each other laterally according to the scanning path, so that the laser scanning can form modified points covering the entire wafer. Among them, the modified points formed by the first laser scanning are independent of each other, that is, there is no interaction between two adjacent modified points. At this time, the depth of the modified points will not gradually rise to self-organize to form a modified layer, and the modified points are uniformly located near the laser focus. To achieve this goal, the processing parameters of the pulsed laser need to meet the following conditions: the point spacing P1 between two adjacent modified points on the same scanning section is not less than 0.7 times the diameter D1 of the modified point, that is, P1 ≥ 0.7D1. It should be noted that this condition is derived based on a deep understanding of the principle of self-organization formation of the modified layer, that is, when P1 < 0.7D1, the next laser pulse will be affected by the modified point generated by the previous laser pulse, so that it will no longer propagate to the laser focus, but the position of its modified point will gradually move upward until the overlap rate of two adjacent modified points reaches a balance with the laser power density, and a stable and thin modified layer is formed at a specific depth.

[0034] In the calculation rules preset for the first laser scan, the laser pulse repetition frequency of the laser is recorded as F1, and the laser scanning speed (the relative movement speed between the ingot and the laser focus) is recorded as V1. Based on P1=V1 / F1, the point spacing P1 can be set by adjusting the laser pulse repetition frequency F1 and the laser scanning speed V1. At the same time, the diameter of the modified point generated by a single laser pulse is recorded as D1, and D1 can be obtained by measuring the diameter of the modified point through a microscope under the condition that P1 is set to be much larger than D1. D1 is related to the laser pulse energy E1 that forms the modified point. The larger the E1, the larger the D1 within a certain range, but the extension length of the modified point in the longitudinal direction also increases accordingly. The smaller the E1, the smaller the D1 within a certain range, but too small an E1 will cause the laser pulse to fail to form a stable modified point. After selecting the appropriate E1 and P1, the entire surface of the ingot is laser scanned, and the line distance between the two scanning segments spaced apart from each other on the scanning path is recorded as L1.

[0035] Preferably, in order to obtain the smallest possible material peeling loss, the focusing lens in this first step needs to be aberration corrected to eliminate the spherical aberration effect caused by the refraction of the laser at the air-ingot interface, so that the light field inside the ingot is as concentrated as possible in the transmission direction, thereby reducing the longitudinal thickness of the modified point scanned by the first laser.

[0036] Preferably, the laser wavelength λ1 of the first laser scan is selected as a pulse laser in the range of 780-2500 nm, the laser pulse width Δt1 is 10 fs-5 ns, and the laser pulse energy E1 ≥ 1 μJ.

[0037] Preferably, the first laser scan should make adjacent modified points as close as possible, that is, P1 is as small as possible, while ensuring that the depth of the modified points does not gradually rise to self-organize and form a modified layer; or, for the same processing mark, the number of relative movements of the laser and the crystal ingot is set to m1≥1, such as processing the same processing mark line back and forth (the number of relative movements m1 is 2); or, each laser pulse for generating a modified point is set to a pulse train with several sub-pulses, wherein the time interval between adjacent sub-pulses does not exceed 100ns, so as to obtain the densest processing effect possible.

[0038] Secondly, during the second laser scanning process, the pulse laser needs to be focused below the preset peeling surface inside the ingot, and the pulse laser is made to act on the local or entire area of ​​the modified point generated by the first laser scanning with a larger spot area, and the drive module is used to drive the ingot and the laser focus to move relative to each other laterally according to the scanning path, so that the laser scanning can form a modified layer composed of modified points covering the entire wafer, and form cracks extending along the dissociation surface of the ingot at the modified location. In this process, when the laser spot overlaps with the modified point generated by the first laser scanning, the modified point generated by the first laser scanning (including amorphous carbon and silicon) will strongly absorb the laser energy, and then form a higher local internal pressure at the overlapping part, causing cracks to be generated along the dissociation surface of the ingot and the cracks to expand laterally. By performing a second laser scan on the entire width of the ingot, a crack layer covering the entire ingot is generated to form a preset peeling surface. Therefore, the key to the second laser scan is the generation and expansion of cracks, which mainly depends on the modified points generated by the first laser scan, that is, the cracks are mainly generated in the overlapping area of ​​the two laser scans. If the processing parameters generated by the second laser scan are used independently for processing, a good modified layer and cracks cannot be formed. This avoids the problem of gradual upward movement of the modified points at the laser scanning end face during the process of generating cracks by separate processing.

[0039] In the calculation rules preset for the second laser scan, let the pulse repetition frequency of the laser be F2, and the scanning speed (relative moving speed between the ingot and the laser focus) be V2. Then the interval between adjacent modified points on the same scanning line in this second step is P2=V2 / F2; let the spot diameter at the depth of the modified point generated in the first step in this second step be D2. In order to form a good crack layer, P2 and D2 should satisfy: P2<D2. More specifically, 0.2D2 is preferred. <P2<0.6D2。

[0040] Preferably, in order to allow the laser to have a larger spot area at the depth of the wafer thickness to be generated after being focused by the focusing lens, the laser focus of the second laser scan is preferably moved downward by a certain amount, and the offset distance S2 is preferably 5-20 μm. In order to allow the laser to be focused by the focusing lens and to perform processing independently with the second laser scan without forming a good modified layer and cracks, the focused beam of the second laser scan needs to have a larger spot area at its laser focus.

[0041] Preferably, the laser wavelength λ2 of the second laser scan is selected as a pulse laser in the range of 780-2500 nm, the laser pulse width Δt2 is 10 ps-100 ns, and the laser pulse energy E2 ≥ 5 μJ.

[0042] Preferably, for the same processing mark, the relative movement times m2 between the laser and the crystal ingot are set to ≥ 1, such as processing the same processing trace line back and forth (the relative movement times m2 is 2); or, each laser pulse for generating a modified point is set to a pulse train with several sub-pulses, wherein the time interval between adjacent sub-pulses does not exceed 100ns, so as to obtain the densest processing effect possible.

[0043] Finally, during the stripping process, the upper and lower end faces of the ingot that has been scanned twice by lasers can be fixed to two hard fixtures with adhesives, and external forces in opposite directions are applied to the two fixtures to separate the wafer along the stripping surface. The wafer and the remaining ingot are then removed from the fixtures, the stripping surface is ground or thinned, and the remaining ingot is used for the next cycle of wafer generation.

[0044] In summary, the beneficial technical effects of the present invention are: it solves the problem of modified point climbing in the existing laser reciprocating scanning method and the harsh process conditions of pulse laser connection scanning with short pulse width and long pulse width, which can not only eliminate the burrs on the edge of the chip, but also further reduce the material loss of laser peeling. It has the advantages of simple process conditions, good morphology of the modified layer, low difficulty of chip peeling, and suitability for large-scale industrial applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the main structure of the crystal ingot during the laser scanning process of the prior art of the present invention.

[0046] Figure 2 It is a schematic diagram of the connection relationship between the first laser, the first beam combining and expanding unit, the aberration correction unit, the first workpiece reflector, the first objective lens and the crystal ingot in Example 2 of the present invention.

[0047] Figure 3 It is a schematic diagram of the connection relationship between the second laser, the second beam combining unit, the beam shaping unit, the second workpiece reflector, the second objective lens and the crystal ingot in Example 2 of the present invention.

[0048] Figure 4 It is a schematic diagram of the main structure of the crystal ingot in Example 3 of the present invention during the first laser scanning process.

[0049] Figure 5 It is a schematic diagram of the top view structure of the crystal ingot in Example 3 of the present invention during the first laser scanning process.

[0050] Figure 6 It is a schematic diagram of the locally enlarged structure of the crystal ingot in Example 3 of the present invention during the first laser scanning process.

[0051] Figure 7It is a schematic diagram of the main structure of the crystal ingot in Example 3 of the present invention during the second laser scanning process.

[0052] Figure 8 It is a schematic diagram of the top view structure of the crystal ingot in Example 3 of the present invention during the second laser scanning process.

[0053] Fig. 9 It is a schematic diagram of the locally enlarged structure of the crystal ingot in Example 3 of the present invention during the second laser scanning process.

[0054] Fig.10 It is a side view representation diagram of the chip produced in Example 3 of the present invention.

[0055] Fig.11 It is a schematic diagram of the top view structure of the crystal ingot in Example 4 of the present invention during the second laser scanning process.

[0056] Fig.12 It is a schematic diagram of the locally enlarged structure of the crystal ingot in Example 4 of the present invention during the second laser scanning process.

[0057] In the figure, 1, wafer; 2, ingot; 3, preset peeling surface; 4, modified point; 5, crack; 61, first laser; 62, first beam combining and expanding unit; 621, first beam combining mirror; 622, first laser reflection mirror; 623, beam expanding mirror; 63, aberration correction unit; 64, first workpiece reflection mirror; 65, first objective lens; 71, second laser; 72, second beam combining unit; 721, second beam combining mirror; 722, second laser reflection mirror; 73, beam shaping unit; 74, second workpiece reflection mirror; 75, second objective lens. DETAILED DESCRIPTION

[0058] In order to make the technical means, creative features, objectives and functions achieved by the present invention clearer and easier to understand, the present invention is further explained below in conjunction with the accompanying drawings and specific implementation methods.

[0059] Embodiment 1: A method for producing a semiconductor wafer disclosed in the present invention comprises the following steps:

[0060] S1 first sets the number of laser scans to n, where n is an integer ≥ 2, and then sets the scanning path of each laser scan and the point spacing P between two adjacent modified points 4 on the scanning path;

[0061] S2 determines the laser scanning speed V and laser pulse repetition frequency F required to achieve the point spacing P based on the preset calculation rules for each laser scan, and determines the corresponding diameter D of the modified point 4, and then determines the laser pulse energy E required to achieve the diameter D of the modified point 4, and the offset distance S of the laser focus relative to the preset peeling surface 3;

[0062] S3: focusing the pulsed laser on the preset peeling surface 3 inside the ingot 2 or below the preset peeling surface 3, and performing n laser scans to sequentially form modified points 4 on the preset peeling surface 3, and forming an overlapping area between the modified points 4 formed by at least two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area;

[0063] S4: peeling the crystal ingot 2 along the preset peeling surface 3 to obtain the wafer 1 and the remaining crystal ingot 2.

[0064] Embodiment 2: A semiconductor chip generation method disclosed in the present invention is different from Embodiment 1 in that, in S3, the pulse laser is pre-corrected for aberration before the first laser scan, and the pulse laser is pre-beam shaped before the second laser scan.

[0065] Reference Figure 2 First, a first laser 61, a first beam combining and expanding unit 62, an aberration correction unit 63, a first workpiece reflector 64, a first objective lens 65, and a crystal ingot 2 are sequentially arranged on the pulse laser emission path of the first laser scan; wherein, two first lasers 61 are arranged in parallel, and the first beam combining and expanding unit 62 includes a first beam combining mirror 621 arranged on the pulse laser emission path of one of the first lasers 61, a first laser reflector 622 respectively arranged between the pulse laser emission paths of the remaining first lasers 61 and the first beam combining mirror 621, and a beam expanding mirror 623 arranged on the pulse laser emission path of the first beam combining mirror 621, and the aberration correction unit 63 is set as an adaptive optical element, a diffractive optical element, or a deformable lens. This embodiment is preferably an SLM beam shaper.

[0066] Specifically, the aberration correction process includes shaping the wavefront of the incident light beam of the pulsed laser through an adaptive optical element, compensating the aberration of the focusing lens of the pulsed laser at a specific depth through a diffractive optical element, or presetting the aberration and compensating the aberration of the focusing lens of the pulsed laser at a specific depth by adjusting the deformable lens. Among them, the adaptive optical element can be, for example, a spatial light modulator and a digital micromirror array, the diffractive optical element can be, for example, a beam shaper, a beam splitter, a diffractive conic lens, a spiral phase plate, a homogenizer, and a multi-focus telephoto depth, and the deformable lens can be, for example, a lens with a movable lens or lens group inside and / or outside.

[0067] Reference Figure 3, secondly, a second laser 71, a second beam combining unit 72, a beam shaping unit 73, a second workpiece reflector 74, a second objective lens 75, and a crystal ingot 2 are sequentially arranged on the pulse laser emission path of the second laser scan; wherein, two second lasers 71 are arranged in parallel, and the second beam combining unit 72 includes a second beam combining mirror 721 arranged on the pulse laser emission path of one of the second lasers 71, and second laser reflectors 722 arranged respectively between the pulse laser emission paths of the remaining second lasers 71 and the second beam combining mirrors 721, and the beam shaping unit 73 is set as an adaptive optical element or a diffractive optical element. This embodiment is preferably an SLM beam shaper.

[0068] Specifically, the process of beam shaping includes shaping the incident beam of the pulsed laser through an adaptive optical element, shaping the incident beam of the pulsed laser through a diffractive optical element, or increasing the beam quality factor of the incident beam of the pulsed laser. The adaptive optical element may be, for example, a spatial light modulator and a digital micromirror array, and the diffractive optical element may be, for example, a beam shaper, a beam splitter, a diffractive aconic lens, a spiral phase plate, a homogenizer, and a multi-focus telephoto depth.

[0069] Example 3: Reference Figures 4 to 9 , is a semiconductor wafer production method disclosed in the present invention, which is different from Example 2 in that it includes the following steps:

[0070] S1 first sets the number of laser scans to 2, and then sets the scanning path of each laser scan to be a line-by-line scanning path, and the scanning paths of the two laser scans are completely parallel and overlapped, wherein the point spacing between two adjacent modified points 4 on the first scanning path is P1=6μm, and the line distance between two scanning segments spaced apart from each other is L1=0.15mm, and the point spacing between two adjacent modified points 4 on the second scanning path is P2=6μm, and the line distance between two scanning segments spaced apart from each other is L2=L1;

[0071] S2 is based on the calculation rules preset in the first laser scan: P1=V1 / F1, P1≥0.7D1, E1≥1μJ, and selects a pulse laser with a laser wavelength range λ1 of 1030nm and a laser pulse width Δt1 of 300fs; determines the laser scanning speed V1=300mm / s and laser pulse repetition frequency F1=50kHz required to achieve the point spacing P1, and determines the corresponding modified point 4 diameter D1=7μm, and then determines the laser pulse energy E1=8μJ required to achieve the modified point 4 diameter D1, and the offset distance S1=0μm of the laser focus relative to the preset peeling surface 3;

[0072] S3 focuses the pulsed laser on the preset peeling surface 3 inside the ingot 2, and performs the first laser scanning to sequentially form modified points 4 on the preset peeling surface 3. At this time, the modified points 4 do not gradually increase in depth to form a modified layer by self-organization, and the modified points 4 are uniformly located near the laser focus of the focusing lens, and the longitudinal thickness of the modified points 4 is within the range of 25-30 μm;

[0073] S4 is based on the preset calculation rules of the second laser scan: P2=V2 / F2, P2<D2, E2≥5μJ, and selects a pulse laser with a laser wavelength range λ2 of 1030nm and a laser pulse width Δt2 of 50ps; determines the laser scanning speed V2=300mm / s and laser pulse repetition frequency F2=50kHz required to achieve the point spacing P2, and determines the corresponding modified point 4 diameter D2=25μm, and then determines the laser pulse energy E2=20μJ required to achieve the modified point 4 diameter D2, and the offset distance S2=10μm of the laser focus relative to the preset peeling surface 3;

[0074] S5 focuses the pulse laser below the preset peeling surface 3 inside the ingot 2, and performs a second laser scan to sequentially form modified points 4 on the preset peeling surface 3, and an overlapping area is formed between the modified points 4 formed by the two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area; Fig.10 , the wafer 1 processed according to the above parameters is characterized, and the total thickness of the laser action area is about 50μm. There is no problem of gradual rise of the modified point 4 at the laser scanning end face, which can effectively reduce the material loss of laser peeling.

[0075] Example 4: Reference Figures 11-12 , is a semiconductor wafer generation method disclosed in the present invention, which is different from Example 3 in that, in S1, the scanning paths of the two laser scans cross and overlap vertically, wherein the line distance L1 and the line distance L2 have no proportional relationship and are independently selected from 0.10~1.00mm. In this embodiment, preferably, the line distance L1 between the two scanning segments spaced apart from each other on the first scanning path is 0.10mm, and the line distance L2 between the two scanning segments spaced apart from each other on the second scanning path is 0.10mm.

[0076] Embodiment 5 is a semiconductor wafer generation method disclosed in the present invention. The difference from Embodiment 3 is that in S3, the pulse laser is a pulse train consisting of three sub-pulses, wherein the time interval between two adjacent sub-pulses is 50ns.

[0077] Embodiment 6 is a semiconductor wafer generation method disclosed in the present invention. The difference from Embodiment 3 is that in S5, the pulse laser is a pulse train consisting of three sub-pulses, wherein the time interval between two adjacent sub-pulses is 50ns.

[0078] Example 7: A semiconductor chip generation method disclosed in the present invention, which differs from Example 3 in that, in S3, for the same processing trace, the relative movement times m1 of the laser and the crystal ingot are set to 2 times, that is, the same processing trace line is processed back and forth according to the scanning path of the first laser scan.

[0079] Embodiment 8: A semiconductor wafer production method disclosed in the present invention, which is different from Embodiment 2 in that it includes the following steps:

[0080] S1 first sets the number of laser scans to 2, and then sets the scanning path of each laser scan to be a line-by-line scanning path, and the scanning paths of the two laser scans are completely parallel and overlapped, wherein the point spacing between two adjacent modified points 4 on the first scanning path is P1=5μm, and the line distance between two scanning segments spaced apart from each other is L1=0.10mm, and the point spacing between two adjacent modified points 4 on the second scanning path is P2=15μm, and the line distance between two scanning segments spaced apart from each other is L2=2L1;

[0081] S2 is based on the calculation rules preset in the first laser scan: P1=V1 / F1, P1≥0.7D1, E1≥1μJ, and selects a pulse laser with a laser wavelength range λ1 of 800nm ​​and a laser pulse width Δt1 of 10fs; determines the laser scanning speed V1=250mm / s and laser pulse repetition frequency F1=50kHz required to achieve the point spacing P1, and determines the corresponding modified point 4 diameter D1=7μm, and then determines the laser pulse energy E1=8μJ required to achieve the modified point 4 diameter D1, and the offset distance S1=0μm of the laser focus relative to the preset peeling surface 3;

[0082] S3 focuses the pulsed laser on the preset peeling surface 3 inside the ingot 2, and performs the first laser scanning to sequentially form modified points 4 on the preset peeling surface 3. At this time, the modified points 4 do not gradually increase in depth to form a modified layer by self-organization, and the modified points 4 are uniformly located near the laser focus of the focusing lens, and the longitudinal thickness of the modified points 4 is within the range of 25-30 μm;

[0083] S4 is based on the preset calculation rules of the second laser scan: P2=V2 / F2, P2<D2, E2≥5μJ, and selects a pulse laser with a laser wavelength range λ2 of 800nm ​​and a laser pulse width Δt2 of 100ps; determines the laser scanning speed V2=750mm / s and laser pulse repetition frequency F2=50kHz required to achieve the point spacing P2, and determines the corresponding modified point 4 diameter D2=50μm, and then determines the laser pulse energy E2=40μJ required to achieve the modified point 4 diameter D2, and the offset distance S2=20μm of the laser focus relative to the preset peeling surface 3;

[0084] S5 focuses the pulsed laser below the preset peeling surface 3 inside the ingot 2, and performs a second laser scan to sequentially form modified points 4 on the preset peeling surface 3, and an overlapping area is formed between the modified points 4 formed by the two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area.

[0085] Embodiment 9 is a semiconductor wafer production method disclosed in the present invention, which is different from Embodiment 2 in that it includes the following steps:

[0086] S1 first sets the number of laser scans to 2, and then sets the scanning path of each laser scan to be a line-by-line scanning path, and the scanning paths of the two laser scans are completely parallel and overlapped, wherein the point spacing between two adjacent modified points 4 on the first scanning path is P1=8μm, and the line distance between two scanning segments spaced apart from each other is L1=0.80mm, and the point spacing between two adjacent modified points 4 on the second scanning path is P2=30μm, and the line distance between two scanning segments spaced apart from each other is L2=L1;

[0087] S2 is based on the calculation rules preset in the first laser scan: P1=V1 / F1, P1≥0.7D1, E1≥1μJ, and selects a pulse laser with a laser wavelength range λ1 of 1030nm and a laser pulse width Δt1 of 200ps; determines the laser scanning speed V1=400mm / s and laser pulse repetition frequency F1=50kHz required to achieve the point spacing P1, and determines the corresponding modified point 4 diameter D1=10μm, and then determines the laser pulse energy E1=12μJ required to achieve the modified point 4 diameter D1, and the offset distance S1=0μm of the laser focus relative to the preset peeling surface 3;

[0088] S3 focuses the pulsed laser on the preset peeling surface 3 inside the ingot 2, and performs the first laser scanning to sequentially form modified points 4 on the preset peeling surface 3. At this time, the modified points 4 do not gradually increase in depth to form a modified layer by self-organization, and the modified points 4 are uniformly located near the laser focus of the focusing lens, and the longitudinal thickness of the modified points 4 is within the range of 30-35 μm;

[0089] S4 is based on the preset calculation rules of the second laser scan: P2=V2 / F2, P2<D2, E2≥5μJ, and selects a pulse laser with a laser wavelength range λ2 of 1064nm and a laser pulse width Δt2 of 20ns; determines the laser scanning speed V2=1500mm / s and laser pulse repetition frequency F2=50kHz required to achieve the point spacing P2, and determines the corresponding modified point 4 diameter D2=100μm, and then determines the laser pulse energy E2=80μJ required to achieve the modified point 4 diameter D2, and the offset distance S2=20μm of the laser focus relative to the preset peeling surface 3;

[0090] S5 focuses the pulsed laser below the preset peeling surface 3 inside the ingot 2, and performs a second laser scan to sequentially form modified points 4 on the preset peeling surface 3, and an overlapping area is formed between the modified points 4 formed by the two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area. Embodiment 10: A semiconductor wafer production method disclosed in the present invention is different from Embodiment 2 in that it includes the following steps:

[0091] S1 first sets the number of laser scans to 2, and then sets the scanning path of each laser scan to be a line-by-line scanning path, and the scanning paths of the two laser scans are completely parallel and overlapped, wherein the point spacing between two adjacent modified points 4 on the first scanning path is P1=10μm, and the line distance between two scanning segments spaced apart from each other is L1=0.40mm, and the point spacing between two adjacent modified points 4 on the second scanning path is P2=8μm, and the line distance between two scanning segments spaced apart from each other is L2=2L1;

[0092] S2 is based on the calculation rules preset in the first laser scan: P1=V1 / F1, P1≥0.7D1, E1≥1μJ, and selects a pulse laser with a laser wavelength range λ1 of 1064nm and a laser pulse width Δt1 of 100ps; determines the laser scanning speed V1=500mm / s and laser pulse repetition frequency F1=50kHz required to achieve the point spacing P1, and determines the corresponding modified point 4 diameter D1=5μm, and then determines the laser pulse energy E1=6μJ required to achieve the modified point 4 diameter D1, and the offset distance S1=0μm of the laser focus relative to the preset peeling surface 3;

[0093] S3 focuses the pulsed laser on the preset peeling surface 3 inside the ingot 2, and performs the first laser scanning to sequentially form modified points 4 on the preset peeling surface 3. At this time, the modified points 4 do not gradually increase in depth to form a modified layer by self-organization, and the modified points 4 are uniformly located near the laser focus of the focusing lens, and the longitudinal thickness of the modified points 4 is within the range of 15-20 μm;

[0094] S4 is based on the preset calculation rules of the second laser scan: P2=V2 / F2, P2<D2, E2≥5μJ, and selects a pulse laser with a laser wavelength range λ2 of 1340nm and a laser pulse width Δt2 of 5ns; determines the laser scanning speed V2=400mm / s and laser pulse repetition frequency F2=50kHz required to achieve the point spacing P2, and determines the corresponding modified point 4 diameter D2=15μm, and then determines the laser pulse energy E2=12μJ required to achieve the modified point 4 diameter D2, and the offset distance S2=10μm of the laser focus relative to the preset peeling surface 3;

[0095] S5 focuses the pulsed laser below the preset peeling surface 3 inside the ingot 2, and performs a second laser scan to sequentially form modified points 4 on the preset peeling surface 3, and an overlapping area is formed between the modified points 4 formed by the two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area. Embodiment 11: A semiconductor wafer production method disclosed in the present invention is different from Embodiment 2 in that it includes the following steps:

[0096] S1 first sets the number of laser scans to 2, and then sets the scanning path of each laser scan to be a line-by-line scanning path, and the scanning paths of the two laser scans are completely parallel and overlapped, wherein the point spacing between two adjacent modified points 4 on the first scanning path is P1=13μm, and the line distance between two scanning segments spaced apart from each other is L1=1.00mm, and the point spacing between two adjacent modified points 4 on the second scanning path is P2=40μm, and the line distance between two scanning segments spaced apart from each other is L2=L1;

[0097] S2 is based on the calculation rules preset in the first laser scan: P1=V1 / F1, P1≥0.7D1, E1≥1μJ, and selects a pulse laser with a laser wavelength range λ1 of 2100nm and a laser pulse width Δt1 of 5ns; determines the laser scanning speed V1=650mm / s and laser pulse repetition frequency F1=50kHz required to achieve the point spacing P1, and determines the corresponding modified point 4 diameter D1=15μm, and then determines the laser pulse energy E1=18μJ required to achieve the modified point 4 diameter D1, and the offset distance S1=0μm of the laser focus relative to the preset peeling surface 3;

[0098] S3 focuses the pulsed laser on the preset peeling surface 3 inside the ingot 2, and performs the first laser scanning to sequentially form modified points 4 on the preset peeling surface 3. At this time, the modified points 4 do not undergo the phenomenon of gradually increasing depth and self-organizing to form a modified layer, and the modified points 4 are uniformly located near the laser focus of the focusing lens, and the longitudinal thickness of the modified points 4 is within the range of 35-40 μm;

[0099] S4 is based on the preset calculation rules of the second laser scan: P2=V2 / F2, P2<D2, E2≥5μJ, and selects a pulse laser with a laser wavelength range λ2 of 1550nm and a laser pulse width Δt2 of 100ns; determines the laser scanning speed V2=2000mm / s and laser pulse repetition frequency F2=50kHz required to achieve the point spacing P2, and determines the corresponding modified point 4 diameter D2=80μm, and then determines the laser pulse energy E2=60μJ required to achieve the modified point 4 diameter D2, and the offset distance S2=15μm of the laser focus relative to the preset peeling surface 3;

[0100] S5 focuses the pulsed laser below the preset peeling surface 3 inside the ingot 2, and performs a second laser scan to sequentially form modified points 4 on the preset peeling surface 3, and an overlapping area is formed between the modified points 4 formed by the two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area. Embodiment 12: A semiconductor wafer production method disclosed in the present invention is different from Embodiment 2 in that it includes the following steps:

[0101] S1 first sets the number of laser scans to 2, and then sets the scanning path of each laser scan to be a line-by-line scanning path, and the scanning paths of the two laser scans are completely parallel and overlapped, wherein the point spacing between two adjacent modified points 4 on the first scanning path is P1=6μm, and the line distance between two scanning segments spaced apart from each other is L1=0.30mm, and the point spacing between two adjacent modified points 4 on the second scanning path is P2=7μm, and the line distance between two scanning segments spaced apart from each other is L2=2L1;

[0102] S2 is based on the calculation rules preset in the first laser scan: P1=V1 / F1, P1≥0.7D1, E1≥1μJ, and selects a pulse laser with a laser wavelength range λ1 of 1340nm and a laser pulse width Δt1 of 50ps; determines the laser scanning speed V1=600mm / s and laser pulse repetition frequency F1=100kHz required to achieve the point spacing P1, and determines the corresponding modified point 4 diameter D1=7μm, and then determines the laser pulse energy E1=5μJ required to achieve the modified point 4 diameter D1, and the offset distance S1=0μm of the laser focus relative to the preset peeling surface 3;

[0103] S3 focuses the pulsed laser on the preset peeling surface 3 inside the ingot 2, and performs the first laser scanning to sequentially form modified points 4 on the preset peeling surface 3. At this time, the modified points 4 do not undergo the phenomenon of gradually increasing depth and self-organizing to form a modified layer, and the modified points 4 are uniformly located near the laser focus of the focusing lens, and the longitudinal thickness of the modified points 4 is within the range of 5-10 μm;

[0104] S4 is based on the preset calculation rules of the second laser scan: P2=V2 / F2, P2<D2, E2≥5μJ, and selects a pulse laser with a laser wavelength range λ2 of 2100nm and a laser pulse width Δt2 of 10ns; determines the laser scanning speed V2=700mm / s and laser pulse repetition frequency F2=100kHz required to achieve the point spacing P2, and determines the corresponding modified point 4 diameter D2=25μm, and then determines the laser pulse energy E2=20μJ required to achieve the modified point 4 diameter D2, and the offset distance S2=10μm of the laser focus relative to the preset peeling surface 3;

[0105] S5 focuses the pulsed laser below the preset peeling surface 3 inside the ingot 2, and performs a second laser scan to sequentially form modified points 4 on the preset peeling surface 3, and an overlapping area is formed between the modified points 4 formed by the two laser scans, so as to form a crack 5 extending laterally along the preset peeling surface 3 in the overlapping area.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for producing a semiconductor wafer, characterized in that: The following steps are included: S1 first sets the number of laser scans to n, where n is an integer ≥ 2, and then sets the scanning path of each laser scan and the point spacing P between two adjacent modified points on the scanning path; S2 determines the laser scanning speed V and laser pulse repetition frequency F required to achieve the point spacing P based on the preset calculation rules for each laser scan, and determines the corresponding modified point diameter D, and then determines the laser pulse energy E required to achieve the modified point diameter D, and the offset distance S of the laser focus relative to the preset peeling surface; S3: focusing a pulsed laser on a preset peeling surface inside the ingot or below the preset peeling surface, and performing n laser scans to sequentially form modified points on the preset peeling surface, and forming an overlapping area between the modified points formed by at least two laser scans, so as to form a crack extending laterally along the preset peeling surface in the overlapping area; S4: peeling the crystal ingot along the preset peeling surface to obtain a wafer and a remaining crystal ingot.

2. A semiconductor wafer production method according to claim 1, characterized in that: In S1, the scanning path of each laser scanning is one or a combination of the following: a line-by-line scanning path, a grid-interlaced scanning path, a concentric circle scanning path, and a spiral line scanning path.

3. A semiconductor wafer production method according to claim 2, characterized in that: In S1, the line distance L between two scanning sections spaced apart from each other on the scanning path is controlled to be 0.05-1.00 mm.

4. A semiconductor wafer production method according to claim 1, characterized in that: In S2, the preset calculation rule for the 1st to nth laser scans is: n′ =V n′ / F n′ , P n′ ≥0.7D n′ , E n′ ≥1μJ, S n′ 0~5μm; among them, P n′ V is the distance between two adjacent modified points on the 1st to nth scanning path, n′ is the laser scanning speed from the 1st to the nth time, F n′ is the repetition frequency of the 1st to nth laser pulses, E n′ is the energy of the 1st~n′th laser pulse, S n′ is the offset distance of the 1st to nth laser focus relative to the preset peeling surface, and n' is an integer ≥1.

5. A semiconductor wafer production method according to claim 1, characterized in that: In S2, the preset calculation rule for the n′+1~n″th laser scan is: n′+1 =V n′+1 / F n′+1 , P n′+1 <D n′+1 , E n′+1 ≥5μJ, S n′+1 0~20μm; among them, P n′+1 is the point distance between two adjacent modified points on the n′+1~n″th scanning path, V n′+1 is the laser scanning speed of the n′+1~n″th time, F n′+1 is the laser pulse repetition frequency of the n′+1~n″th time, E n′+1 is the laser pulse energy of the n′+1~n″th time, S n′+1 It is the offset distance of the laser focus from the n′+1 to n″th time relative to the preset peeling surface, where n″ is an integer ≥2.

6. A semiconductor wafer production method according to claim 5, characterized in that: In S2, control 0.2D n′+1 <P n′+1 <0.6D n′+1 .

7. A semiconductor wafer production method according to claim 1, characterized in that: In S3, the laser wavelength range λ of each laser scan is controlled to be 780-2500 nm, and the laser pulse width Δt is controlled to be 10 fs-100 ns.

8. A semiconductor wafer production method according to claim 1, characterized in that: In S3, the pulse laser is a pulse train composed of a plurality of sub-pulses, wherein the time interval between two adjacent sub-pulses does not exceed 100 ns.

9. A semiconductor wafer production method according to claim 1, characterized in that: In the above S3, the pulse laser is pre-corrected for aberration before the 1st to n′th laser scans.

10. A semiconductor wafer production method according to claim 1, characterized in that: In S3, the pulse laser is beam shaped before the n′+1-n″th laser scan.

Images