A die singulation method and a silicon-on-insulator wafer

Abstract

The present disclosure relates to the technical field of semiconductor manufacturing. The present disclosure provides a splitting method and a silicon-on-insulator wafer. The splitting method comprises: heating and holding a bonded wafer in an atmosphere with a pressure of 0.2 MPa-2 MPa to grow hydrogen bubbles in an ion implantation layer, wherein the bonded wafer comprises a donor silicon wafer and a base silicon wafer, the donor silicon wafer is bonded to the base silicon wafer through a silicon dioxide layer on the surface of the donor silicon wafer, and the ion implantation layer is formed by implanting hydrogen ions into the donor silicon wafer; after the holding is completed, the bonded wafer is cooled, and the pressure of the atmosphere is reduced to normal pressure after the cooling is completed; and the bonded wafer after the cooling and the pressure reduction is peeled along the ion implantation layer, so that a silicon layer between the ion implantation layer and the silicon dioxide layer of the donor silicon wafer is transferred to the base silicon wafer to obtain a silicon-on-insulator wafer. Thus, the surface roughness of the silicon-on-insulator wafer after peeling can be reduced.

Description

Technical Field

[0001] This disclosure relates to the field of semiconductor manufacturing technology, and more particularly to a dicing method and silicon-on-insulator wafers. Background Technology

[0002] A silicon-on-insulator (SiI) wafer can include a silicon layer, an insulating layer, and a substrate silicon wafer, with the insulating layer located between the silicon layer and the substrate silicon wafer. Smart Cut, also known as ion implantation lift-off, is a process for fabricating SiI wafers. In this process, an oxide layer is first formed on the surface of a donor silicon wafer. Then, hydrogen or helium ions are implanted into the donor silicon wafer to a predetermined depth to form an ion implantation layer. After the donor silicon wafer is bonded to the substrate silicon wafer, the bonded wafer is heat-treated, causing the implanted ions to accumulate and form bubbles in the ion implantation layer, which in turn causes the donor silicon wafer to cleave along the ion implantation layer.

[0003] During the exfoliation heat treatment at normal pressure, hydrogen bubbles in the ion-implanted layer expand at high temperatures. The growth state of these bubbles may differ at different locations within the ion-implanted layer, leading to uneven exfoliation interfaces as cracks propagate along the ion-implanted layer. This results in potentially high surface roughness on the silicon-on-insulator wafer, requiring subsequent surface treatment to reduce this roughness, thus adding an extra processing step. Therefore, existing dicing processes suffer from the problem of high surface roughness in the resulting silicon-on-insulator wafers. Summary of the Invention

[0004] In a first aspect, this disclosure provides a dicing method. The method includes: heating and holding a bonding wafer in an atmosphere with a pressure of 0.2 MPa to 2 MPa to allow hydrogen bubbles to grow in an ion-implanted layer, wherein the bonding wafer includes a donor silicon wafer and a substrate silicon wafer, the donor silicon wafer being bonded to the substrate silicon wafer via a silicon dioxide layer on its surface, and the ion-implanted layer being formed by hydrogen ion implantation into the donor silicon wafer; after holding the temperature, cooling the bonding wafer and reducing the pressure of the atmosphere to ambient pressure; and peeling the cooled and depressurized bonding wafer along the ion-implanted layer, transferring the silicon layer of the donor silicon wafer located between the ion-implanted layer and the silicon dioxide layer onto the substrate silicon wafer, thereby obtaining a silicon-on-insulator wafer.

[0005] Secondly, this disclosure provides a silicon-on-insulator (SBI) wafer. The SBI wafer includes a substrate silicon wafer, a silicon dioxide layer, and a silicon layer. The silicon dioxide layer is disposed on one side of the substrate silicon wafer, and the thickness of the silicon dioxide layer is 50 nm to 3 μm. A silicon layer is disposed on the side of the silicon dioxide layer away from the substrate silicon wafer. The silicon layer is the portion of the donor silicon wafer located between the ion implantation layer and the silicon dioxide layer in the dicing method. The dicing method includes: heating and holding a bonding wafer in an atmosphere with a pressure of 0.2 MPa to 2 MPa. The bonding wafer includes a donor silicon wafer and a substrate silicon wafer. The donor silicon wafer is bonded to the substrate silicon wafer through a silicon dioxide layer located on the surface of the donor silicon wafer. The donor silicon wafer has an ion implantation layer formed by hydrogen ion implantation. Heating and holding cause hydrogen bubbles to grow in the ion implantation layer. After holding, the bonding wafer is cooled down, and then the pressure of the atmosphere is reduced to atmospheric pressure. The cooled and depressurized bonding wafer is peeled off along the ion implantation layer, so that the silicon layer between the ion implantation layer and the silicon dioxide layer of the donor silicon wafer is transferred to the substrate silicon wafer, resulting in a silicon-on-insulator wafer. The surface roughness Ra of the silicon wafer on insulator is less than 3 nm in a 10 μm × 10 μm region, and the total thickness variation of the silicon wafer on insulator is less than 1 nm.

[0006] The cleaving method disclosed herein applies pressure from outside the bonded wafer during the cleaving process using an atmosphere of 0.2 MPa to 2 MPa to constrain the expansion state of the hydrogen bubbles and cause the cracks to propagate along the ion implantation layer, thereby reducing the surface roughness of the silicon wafer on the insulator after peeling. Attached Figure Description

[0007] Figure 1 A flowchart of the dicing method provided in this disclosure.

[0008] Figure 2 A schematic diagram of a bonded wafer implanted with hydrogen ions.

[0009] Figure 3 for Figure 2 A schematic diagram of the ion implantation layer.

[0010] Figure 4(a) is a schematic diagram of a flat, disc-shaped bubble.

[0011] Figure 4(b) is a schematic diagram of an ellipsoidal bubble.

[0012] Figure 5 This is a schematic diagram of a silicon-on-insulator wafer provided in this disclosure.

[0013] Figure 6 This is a dot plot showing the surface roughness of a silicon wafer on an insulator as a function of atmospheric pressure.

[0014] Figure 7 This is a dot plot showing the change in the total thickness of a silicon wafer on an insulator as a function of atmospheric pressure. Detailed Implementation

[0015] The technical solutions in this disclosure will now be clearly and completely described with reference to the accompanying drawings.

[0016] Reference Figure 1 The dicing method may include steps S120 to S180. Step S120 involves heating the bonding wafer in an atmosphere with a pressure of 0.2 MPa to 2 MPa and holding it at that temperature to allow hydrogen bubbles to grow in the ion implantation layer; step S140 involves cooling the bonding wafer after the holding temperature is reached; step S160 involves reducing the pressure of the atmosphere to atmospheric pressure; and step S180 involves peeling the cooled and depressurized bonding wafer along the ion implantation layer.

[0017] In this disclosure, both atmosphere pressure and furnace tube pressure refer to absolute pressure. The atmosphere pressure during the heating and holding stages is 0.2 MPa to 2 MPa, meaning that the atmosphere pressure is within this range during this stage; it is not required that the atmosphere pressure remain at the same fixed value at all times. This range is not used to limit the vacuuming process before atmosphere establishment or the subsequent process of reducing the atmosphere pressure to atmospheric pressure.

[0018] In this disclosure, atmospheric pressure refers to the atmospheric pressure of the process environment, which is typically about 0.1 MPa; reducing the atmospheric pressure to atmospheric pressure means reducing it to a level comparable to the atmospheric pressure of the process environment.

[0019] As Figure 1 The processing object of the method shown includes a donor silicon wafer and a substrate silicon wafer. The surface of the donor silicon wafer has a silicon dioxide layer, which is bonded to the substrate silicon wafer. The donor silicon wafer has an ion implantation layer formed by hydrogen ion implantation, which is spaced apart from the silicon dioxide layer. The silicon material between them constitutes the silicon layer to be transferred.

[0020] Reference Figure 2 In step S120, heating causes the injected hydrogen to accumulate within the ion implantation layer 222, and the already formed hydrogen bubbles 224 continue to grow. Heating maintains the bonded wafer at a specific temperature for a certain period, providing heat treatment conditions for bubble growth and crack formation along the ion implantation layer 222. An atmosphere of 0.2 MPa to 2 MPa is applied from outside the bonded wafer to constrain the expansion state of the hydrogen bubbles 224 and the crack propagation state within the ion implantation layer 222.

[0021] In steps S140 and S160, after the heat preservation is completed, the bonding wafer is first cooled down, and after the cooling is completed, the atmosphere pressure is reduced to atmospheric pressure. When the cooling and depressurization are completed, the donor silicon wafer 220 has been cleaved along the ion implantation layer 222, but the remaining part 228 has not been completely separated from the silicon layer 226 to be transferred.

[0022] In step S180, a dicing device can be used to peel off along the location of the ion implantation layer 222, completely separating the remaining portion 228 from the silicon layer 226 to be transferred that is bonded to the substrate silicon wafer 260 through the silicon dioxide layer 240.

[0023] After stripping, the silicon layer located between the ion implantation layer 222 and the silicon dioxide layer in the donor silicon wafer is retained on the side of the silicon dioxide layer away from the substrate silicon wafer. The substrate silicon wafer, the silicon dioxide layer, and the silicon layer retained on the side of the silicon dioxide layer away from the substrate silicon wafer together form a silicon on insulator (SOI) wafer.

[0024] Reference Figure 2 The bonding wafer may include a donor silicon wafer 220, a silicon dioxide layer 240, and a substrate silicon wafer 260. The silicon dioxide layer 240 is located between the donor silicon wafer 220 and the substrate silicon wafer 260, and the donor silicon wafer 220 is bonded to the substrate silicon wafer 260 through the silicon dioxide layer 240.

[0025] The donor silicon wafer 220 has a first surface 221 facing the substrate silicon wafer 260 and a second surface 223 facing away from the substrate silicon wafer 260. A silicon dioxide layer 240 is formed on one side of the first surface 221 and is located between the donor silicon wafer 220 and the substrate silicon wafer 260 after bonding.

[0026] The donor silicon wafer 220 has an ion implantation layer 222 inside. The ion implantation layer 222 is separated from the first surface 221 by a predetermined depth and can extend in a direction generally parallel to the first surface 221. The portion of the donor silicon wafer 220 located between the ion implantation layer 222 and the silicon dioxide layer 240 is the silicon layer to be transferred 226; the portion of the ion implantation layer 222 located on the side opposite to the silicon dioxide layer 240 is the remaining portion 228 of the donor silicon wafer 220.

[0027] The ion implantation layer 222 can be formed by hydrogen ion implantation. During the ion implantation process, hydrogen ions enter the donor silicon wafer 220 and form hydrogen-rich plate-like defects at a predetermined depth corresponding to the implantation energy. The hydrogen-rich plate-like defects can serve as nucleation sites for hydrogen bubbles 224 during subsequent heat treatment, so that bubbles and cracks are mainly distributed within the depth range of the ion implantation layer 222.

[0028] In embodiments of this disclosure, the hydrogen ion implantation energy can be from 20 keV to 200 keV, and the implantation dose can be 5 × 10¹. 6 cm - ² (cm) -2 ) ~ 1×10¹ 7 cm - ².

[0029] The donor silicon wafer 220 can be a 12-inch diameter silicon wafer with a crystal orientation of (100) and a resistivity of 1 Ω·cm to 100 Ω·cm, and its front-side silicon dioxide layer 240 has a thickness of 50 nm to 3 μm. H₂ is implanted using an ion implanter at an energy of 40 keV. + (Hydrogen ions), injected at a dose of 6 × 10¹ 6 cm - ², thereby forming an ion implantation layer 222 inside the donor silicon wafer 220.

[0030] The substrate silicon wafer 260 can be a 12-inch diameter single-crystal silicon wafer. A donor silicon wafer 220 with a silicon dioxide layer 240 is placed opposite the substrate silicon wafer 260 and bonded together to form a bonded wafer. The silicon dioxide layer 240 serves as an insulating layer between the silicon layer and the substrate silicon wafer in the final insulator-on-silicon wafer. The thickness of the silicon layer to be transferred 226 is related to the depth of the ion implantation layer 222 relative to the first surface 221.

[0031] After the bonding wafer is formed, the donor silicon wafer 220, the silicon dioxide layer 240, and the substrate silicon wafer 260 remain in a stacked state. In the subsequent heat treatment stage, the hydrogen-rich defects in the ion implantation layer 222 serve as bubble nucleation and growth regions, while the bonding relationship between the silicon dioxide layer 240 and the substrate silicon wafer 260 allows the silicon layer 226 to be transferred to remain on the substrate silicon wafer 260 side after peeling.

[0032] The bonded wafers can be housed inside a furnace tube. Before heating and holding, the furnace tube can be evacuated to reduce the pressure inside to less than 10 Pa. Evacuation is used to reduce the pressure of the existing gas inside the furnace tube before introducing process gas. Once the pressure condition of less than 10 Pa is achieved, gas is continuously introduced into the furnace tube to create an atmosphere suitable for heat treatment. The introduced gas can be N2 (nitrogen) or Ar (argon).

[0033] An atmosphere of 0.2 MPa to 2 MPa can be formed by introducing gas into the furnace tube and controlling the pressure inside the furnace tube. This atmosphere acts on the outer surface of the bonded wafer during the heating and holding process, and together with the internal pressure of the hydrogen bubbles 224 in the ion implantation layer 222, it forms the pressure boundary conditions for the bubbles and cracks.

[0034] As specific heat treatment conditions, the furnace tube can be heated to 500 °C at a heating rate of 3 °C / min to 5 °C / min, and held at 500 °C for 1 hour. During the heating process, the injected hydrogen diffuses, accumulates, and forms bubbles in the ion implantation layer 222; during the 500 °C holding stage, the bubbles continue to accumulate and grow, forming cracks within the ion implantation layer 222.

[0035] A pressurized atmosphere can be maintained at a corresponding pressure during the heating and holding processes. During heating, the injected hydrogen may diffuse along the silicon lattice to the outside of the donor silicon wafer and may escape from the side or back of the wafer. A pressurized N2 or Ar atmosphere can reduce the driving force for hydrogen diffusion outward, allowing more hydrogen to remain in the region where the ion implantation layer 222 is located, providing conditions for the accumulation of internal pressure in bubbles at different locations of the ion implantation layer 222.

[0036] After the heat treatment is completed, the temperature of the furnace tube can be reduced to below 200°C at a cooling rate of 3°C / min to 5°C / min. Once the corresponding temperature is reached, the pressure of the atmosphere inside the furnace tube can be reduced to atmospheric pressure at a depressurization rate of 0.1 MPa / min. After completing the cooling and depressurization, the bonding wafer can be removed from the furnace tube.

[0037] The atmospheric pressure inside the furnace tube can be controlled within the range of 0.2 MPa to 2 MPa during heating and heat preservation. Specifically, when establishing a pressurized atmosphere, the atmospheric pressure inside the furnace tube can be increased to 0.2 MPa first, and heat treatment can be carried out under the corresponding pressure.

[0038] In one embodiment, the atmosphere pressure is 2 MPa during heating and holding. The furnace tube temperature can be raised to 500 °C at a heating rate of 3 °C / min to 5 °C / min, and held at 500 °C for 10 min to 60 min. After holding, the furnace tube temperature can be lowered to below 200 °C at a cooling rate of 3 °C / min to 5 °C / min, and then the atmosphere pressure inside the furnace tube can be reduced to atmospheric pressure at a depressurization rate of 0.1 MPa / min.

[0039] Reference Figure 3 Multiple hydrogen bubbles 224 are distributed within the ion implantation layer 222, and a pressurized atmosphere applies pressure to the donor silicon wafer 220 from the outside of the bonding wafer. Figure 3 The arrows in the diagram indicate the direction in which external atmospheric pressure acts on the bonded wafer.

[0040] After hydrogen ion implantation, the ion-implanted layer 222 contains hydrogen-rich plate-like defects, vacancies, and dislocations. During heating, hydrogen accumulates at these locations and forms microbubbles. As the temperature increases and the holding time lengthens, the volume of the hydrogen bubbles 224 increases, the internal pressure of the bubbles increases, and the silicon material between adjacent bubbles is subjected to the internal pressure of the bubbles.

[0041] The growth direction of hydrogen bubble 224 is influenced by the defect distribution of ion implantation layer 222, silicon lattice state, internal pressure of the bubble, and external atmospheric pressure. When hydrogen bubble 224 grows along the plane of ion implantation layer 222, it can utilize the relatively low resistance path formed by hydrogen-rich defect regions; when hydrogen bubble 224 grows in a direction perpendicular to ion implantation layer 222, it requires displacement of adjacent relatively intact silicon lattices.

[0042] After an external pressurized atmosphere is applied to the bonding wafer, the external pressure constrains the expansion of bubbles along the thickness direction. This external pressure constrains the longitudinal expansion of bubbles more strongly than it constrains lateral growth, making bubbles more inclined to grow along the plane of the ion implantation layer 222. The resulting bubbles can have a larger diameter-to-height ratio.

[0043] Under normal pressure, bubbles at different locations in the ion implantation layer 222 may grow at different rates. When a local bubble first reaches the internal pressure capable of triggering a crack, a crack will form at that location first, while bubbles at other locations have not yet reached the corresponding state. The depth of the first-formed local crack may differ from that of the later-formed crack, resulting in steps or undulations on the peeling surface.

[0044] After an external pressurized atmosphere is applied to the bonded wafer, it creates a common external pressure boundary for each bubble within the ion implantation layer 222. Under similar conditions, as the external high-pressure atmosphere increases, the internal pressure required for bubble-induced cracking also increases, allowing the bubble to grow further before cracking. When adjacent bubbles connect at similar depths, cracks can propagate along the ion implantation layer 222, thereby reducing the undulations of the peeling interface along the thickness direction of the donor silicon wafer 220. The specific conditions for bubble-induced cracking also depend on the size and location of the bubble and the local state of the surrounding material.

[0045] By allowing bubbles at different locations in the ion implantation layer 222 to continue accumulating internal pressure under external pressure conditions, a more sufficient growth process can be provided for bubbles at different locations, reducing the possibility of premature cracking caused by local bubbles. When adjacent bubbles connect at similar depths, cracks can propagate along the ion implantation layer 222, thereby reducing the undulation of the peeling interface along the thickness direction of the donor silicon wafer 220.

[0046] Referring to Figures 4(a) and 4(b), the bubble morphology in the ion implantation layer 222 can include flat disc-shaped (shown in Figure 4(a)) and spherical or ellipsoidal (shown in Figure 4(b)). Figures 4(a) and 4(b) are used to compare the relationship between the size of the bubble along the plane of the ion implantation layer 222 and the size along the thickness direction of the donor silicon wafer 220, as well as the influence of different bubble morphologies on the connectivity of adjacent bubbles and crack paths.

[0047] Referring to Figure 4(a), the flat, dish-shaped bubbles have a relatively large aspect ratio, which is the ratio of the bubble diameter D to the bubble height H. That is, the size of the flat, dish-shaped bubble along the plane of the ion implantation layer 222 is larger than its size along the thickness direction of the donor silicon wafer 220. This type of bubble preferentially grows laterally, with the main parts of adjacent bubbles located at similar depths. After the bubble edges approach each other, a connection region extending along the planar direction can be formed within the ion implantation layer 222. After multiple flat, dish-shaped bubbles connect at similar depths, the resulting crack can continue to propagate along the plane of the ion implantation layer 222. Due to the small height difference between adjacent bubbles, the degree to which the crack deflects along the thickness direction of the donor silicon wafer 220 when passing through different bubble regions is relatively small, and the height change at the peeling interface is correspondingly reduced.

[0048] Referring to Figure 4(b), spherical or ellipsoidal bubbles have a relatively small aspect ratio, and their dimensions along the thickness direction of the donor silicon wafer 220 are proportionally large to the bubble diameter. The centers or edges of adjacent spherical or ellipsoidal bubbles may be located at different depths. When bubbles connect, cracks need to propagate at different heights, which may create local height differences at the peeling interface.

[0049] The pressurized atmosphere constrains the expansion of bubbles along the thickness direction, while the hydrogen-rich defects within the ion-implanted layer 222 provide a path for the lateral growth of bubbles. When these two factors work together, they can guide the bubbles to form flat, dish-shaped bubbles with a large aspect ratio, making it easier for adjacent bubbles to connect within the plane of the ion-implanted layer 222. This reduces the undulation of the peeling interface and decreases the surface roughness of the silicon wafer on the insulator.

[0050] The ion implantation layer 222 itself has a certain thickness, and the depth distribution of hydrogen ion implantation and the distribution of the implantation dose within the wafer surface also affect the nucleation location of bubbles. Therefore, even in a pressurized atmosphere, the morphology and location of different bubbles may still differ. This disclosure constrains the bubble growth direction by external pressure, causing the overall bubble distribution and crack path to change towards propagation along the ion implantation layer 222.

[0051] The ion-implanted layer 222 may include hydrogen-rich plate-like defects, vacancies, and dislocations distributed along the planar direction of the donor silicon wafer 220, where multiple hydrogen bubbles 224 nucleate and grow. The ion-implanted layer 222 is not an ideal plane with zero thickness, but rather a damaged region occupying a certain depth range within the donor silicon wafer 220. After hydrogen ion implantation, hydrogen accumulates in this region to form hydrogen-rich defects, and bubbles formed during heat treatment are also distributed within the corresponding depth range. The depth difference of the bubble nucleation sites constitutes one of the sources of undulations at the peeling interface.

[0052] In traditional atmospheric pressure Smart Cut processes, the surface roughness of a freshly cleaved silicon-on-insulator wafer typically ranges from 5 nm to 10 nm. Under an atomic force microscope, this surface can exhibit irregular nanoscale wavy undulations; micro-pits can form at the locations of bubble ruptures, and the density of pits may vary at different locations. Ion-implanted regions may also retain crystal damage, including dislocations and point defects.

[0053] The surface roughness of crack fragments is also related to the initial distribution of bubbles, the growth rate and direction of bubbles during heat treatment, and the local state during bubble rupture. Crack propagation paths may deviate along adjacent bubble interfaces and defect locations, the spalling surface may consist of multiple micro-cleavage planes, and the impact of bubble rupture may also cause localized plastic deformation of the surface. These factors can cause crack paths to deviate from the ideal plane and create local height differences.

[0054] The ion-implanted layer 222 is enriched with hydrogen-rich plate-like defects, vacancies, and dislocations in the planar direction, while still needing to penetrate a relatively intact silicon lattice in the direction perpendicular to this plane. Therefore, the fracture toughness corresponding to crack propagation along the plane of the ion-implanted layer 222 is lower than the fracture toughness corresponding to crack propagation into the intact silicon lattice in the perpendicular direction. For example, the transverse fracture toughness of the ion-implanted layer 222 is approximately 0.3 MPa·m¹. ² (megapascals) 1 / 2 The longitudinal fracture toughness of a perfect silicon lattice is approximately 0.9 MPa·m¹ ². External atmospheric pressure constrains the expansion of bubbles along the 220 thickness direction of the donor silicon wafer, which helps reduce the deflection of cracks in this thickness direction.

[0055] Under the condition of the same internal pressure of the bubble, increasing the external atmospheric pressure will reduce the pressure difference between the inside and outside of the bubble, thereby weakening the effect of bubble expansion on the surrounding silicon material. The external pressure changes the boundary conditions at the crack tip, reduces the crack propagation rate, and decreases the possibility of the crack deviating from the ion implantation layer 222 and entering the longitudinally intact silicon lattice. The crack can then propagate more along the ion implantation layer 222, which has lower fracture toughness.

[0056] During hydrogen ion implantation, the standard deviation σ of the hydrogen ion implantation depth can range from 15 nm to 25 nm, giving the ion implantation layer 222 a certain distribution range in the thickness direction. Furthermore, beam scanning deviations may cause lateral differences in the implantation dose within the wafer plane. These lateral differences in implantation depth distribution and implantation dose affect the nucleation location, size, and distribution of bubbles within the ion implantation layer 222, thereby influencing the surface roughness of the silicon layer after lift-off.

[0057] The size of the bubble along the thickness direction is affected by the thickness of the ion implantation damage zone. When the size of the bubble along the thickness direction is close to the thickness of the ion implantation damage zone, the effect of increasing the external atmosphere pressure on the further adjustment of the bubble morphology gradually decreases, and the decrease in the surface roughness of the silicon wafer on the insulator after peeling with increasing external atmosphere pressure tends to be smaller.

[0058] In the method disclosed herein, the absolute pressure of the atmosphere during the heating and holding stages is 0.2 MPa to 2 MPa. To examine the performance changes when the atmosphere pressure is further increased, a reference test pressure higher than the above range can be set, and the results obtained can be compared with the test results within the above range.

[0059] like Figure 5 As shown, this disclosure also provides a silicon-on-insulator (SBI) wafer 200, which can be fabricated using the dicing method provided in this disclosure. The SBI wafer 200 includes a substrate silicon wafer 260, a silicon dioxide layer 240, and a silicon layer 226, with the silicon dioxide layer 240 located between the substrate silicon wafer 260 and the silicon layer 226. The silicon layer 226 is composed of… Figure 2 A silicon layer 226 to be transferred is formed in the donor silicon wafer 220. Specifically, the silicon layer 226 to be transferred is located between the ion implantation layer 222 and the silicon dioxide layer 240, and is separated from the remaining portion 228 along the ion implantation layer 222 during the stripping process. After separation, the silicon layer 226 to be transferred is still bonded to the substrate silicon wafer 260 through the silicon dioxide layer 240, and constitutes the silicon layer 226 of the silicon-on-insulator wafer 200.

[0060] The surface of silicon layer 226 facing away from silicon dioxide layer 240 is formed by peeling along ion implantation layer 222. The morphology, distribution, and crack propagation path of bubbles in ion implantation layer 222 are reflected in the microscopic undulations of this surface. By heating and holding in an atmosphere of 0.2 MPa to 2 MPa, the expansion of bubbles along the thickness direction and the deflection of cracks are constrained, thereby reducing the surface roughness.

[0061] The surface roughness Ra (Roughness Average, profile arithmetic mean deviation) of the silicon-on-insulator (SiO2) wafer 200 can be measured using an atomic force microscope (AFM) with a scanning area of ​​10 μm × 10 μm. An ellipsometry can also be used to perform 49-point measurements on the surface of the SiO2 wafer 200 to determine the total thickness variation (TTV). In this disclosure, the surface roughness Ra of the SiO2 wafer 200 refers to the surface roughness Ra of the side of the silicon layer 226 facing away from the silicon dioxide layer 240. The surface roughness Ra characterizes the microscopic undulations of the stripped surface, and the total thickness variation characterizes the thickness variation of the SiO2 wafer 200 between measurement locations.

[0062] Within the test pressure range of 0.2 MPa to 2 MPa, the surface roughness Ra and total thickness change of the silicon wafer on the insulator generally decrease with increasing atmosphere pressure. Under the test pressure condition of 2 MPa, the surface roughness Ra of the silicon wafer on the insulator is less than 3 nm, for example, 2.8 nm; and the total thickness change of the silicon wafer on the insulator is less than 1 nm, for example, 0.92 nm.

[0063] Compared to bulk silicon wafers, silicon-on-insulator (SiI) wafers have the advantages of lower parasitic capacitance, stronger radiation resistance, and better device isolation. SiI wafers 200 can be used in high-frequency radio frequency devices, low-power CMOS (Complementary Metal-Oxide-Semiconductor) circuits, power semiconductor devices, or MEMS (Micro-Electro-Mechanical System) sensors.

[0064] The thickness of the silicon-on-insulator wafer 200 is related to the depth of the ion implantation layer 222 relative to the surface of the donor silicon wafer 220. The thickness uniformity of the silicon-on-insulator wafer 200 can be characterized by the total thickness variation. The 10 μm × 10 μm scanning results of the atomic force microscope correspond to the local surface region, and the 49-point measurement results of the ellipsometer correspond to multiple measurement positions within the wafer surface. These two methods are used to characterize the surface quality and thickness uniformity of the silicon-on-insulator wafer 200, respectively.

[0065] This disclosure introduces external atmospheric pressure during the wafer cleaving heat treatment stage without altering the basic layered relationship of the silicon-on-insulator wafer 200, which consists of a substrate silicon wafer 260, a silicon dioxide layer 240, and a silicon layer 226. The pressure conditions primarily affect the bubble growth and crack propagation processes before the formation of the silicon layer 226 to be transferred, and their impact is reflected in the changes in surface roughness and total thickness of the silicon-on-insulator wafer 200 after peeling.

[0066] The following describes the changes in surface roughness and total thickness obtained under different atmospheric pressure conditions, using specific embodiments and comparative examples.

[0067] Example 1 uses a 12-inch diameter silicon wafer with a crystal orientation of (100), a resistivity of 1 Ω·cm to 100 Ω·cm, and a 50 nm to 3 μm thick silicon dioxide layer on the front side as the donor silicon wafer. The donor silicon wafer is placed in an ion implanter and H+ is implanted at an energy of 40 keV. + The injection dose was 6×10¹ 6 cm - ², an ion implantation layer 222 is formed inside the donor silicon wafer.

[0068] Prepare a 12-inch diameter single-crystal silicon wafer as the substrate silicon wafer, and perform low-temperature hydrophilic bonding between the donor silicon wafer and the substrate silicon wafer, so that the silicon dioxide layer is located between the donor silicon wafer and the substrate silicon wafer to form a bonded wafer.

[0069] The bonding wafer is placed in the furnace tube, and the furnace tube is evacuated to a pressure of less than 10 Pa. Then, N2 or Ar gas is introduced into the furnace tube and the gas is continuously introduced. When establishing the pressurized atmosphere, the atmosphere pressure inside the furnace tube is increased to 0.2 MPa.

[0070] In the aforementioned atmosphere, the furnace tube temperature is increased to 500 °C at a heating rate of 3 °C / min to 5 °C / min, and held at 500 °C for 60 min. During this process, hydrogen bubbles 224 in the ion implantation layer 222 grow under the external atmospheric pressure and form cracks that extend along the ion implantation layer 222.

[0071] After the heat treatment is completed, the furnace tube temperature is reduced to below 200 °C at a rate of 3 °C / min to 5 °C / min, and then the atmosphere pressure inside the furnace tube is reduced to atmospheric pressure at a rate of 0.1 MPa / min. The bonded wafer that has undergone heat treatment is removed, and a dicing device is used to peel it off along the ion implantation layer 222 to obtain a silicon-on-insulator wafer.

[0072] The surface roughness Ra of the silicon wafer on insulator was measured using an atomic force microscope within a 10 μm × 10 μm scanning area, and the result was 6.9 nm. An ellipsometry was used to perform 49-point measurements on the silicon wafer on insulator, and the total thickness variation was found to be 2.7 nm.

[0073] Except for the adjustment of atmospheric pressure, Examples 2 to 6 used the same donor silicon wafer, substrate silicon wafer, ion implantation parameters, bonding method, heating rate, holding temperature, holding time, cooling rate, cooling endpoint, pressure reduction rate, peeling method, and measurement method as Example 1 to compare the effect of atmospheric pressure on the surface roughness Ra and total thickness change of silicon wafers on insulators.

[0074] In Example 2, the atmospheric pressure was 0.5 MPa, the surface roughness Ra of the silicon wafer on the insulator was 5.7 nm, and the total thickness variation was 2.4 nm.

[0075] In Example 3, the atmospheric pressure was 1 MPa, the surface roughness Ra of the silicon wafer on the insulator was 4.3 nm, and the total thickness variation was 1.7 nm.

[0076] In Example 4, the atmospheric pressure was 1.5 MPa, the surface roughness Ra of the silicon wafer on the insulator was 3.6 nm, and the total thickness variation was 1.31 nm.

[0077] In Example 5, the atmospheric pressure was 2 MPa, the surface roughness Ra of the silicon wafer on the insulator was 2.8 nm, and the total thickness variation was 0.92 nm.

[0078] Example 6 serves as a reference test. The atmosphere pressure was 2.5 MPa, and the resulting silicon wafer on insulator had a surface roughness Ra of 2.6 nm and a total thickness variation of 0.83 nm. The 2.5 MPa pressure was used as a reference test pressure condition, exceeding the pressure range defined in this disclosure, to examine performance changes when the pressure was further increased.

[0079] Compared with Example 1, Examples 7 and 8 only changed the heat preservation time, while other implementation conditions remained the same.

[0080] In Example 7, the heat preservation time was 10 min, the surface roughness Ra of the silicon wafer on the insulator was 7.3 nm, and the total thickness change was 3.2 nm.

[0081] In Example 8, the heat preservation time was 30 min, the surface roughness Ra of the silicon wafer on the insulator was 7.2 nm, and the total thickness change was 3.1 nm.

[0082] The comparative example used the same donor silicon wafer, hydrogen ion implantation parameters, substrate silicon wafer, and bonding method as the example. The bonded wafer was placed in a furnace tube, and the furnace temperature was raised to 500 °C at a heating rate of 3 °C / min to 5 °C / min and held for 1 hour under a N2 or Ar atmosphere of 0.1 MPa. Subsequently, the temperature was lowered to below 200 °C at a rate of 3 °C / min to 5 °C / min. The silicon layer thickness of the silicon-on-insulator wafer prepared in the comparative example was 200 nm.

[0083] The bonded wafer that had undergone heat treatment in the comparative example was removed and peeled off in a dicing apparatus to obtain a silicon-on-insulator (SiI) wafer. The surface roughness Ra of the SiI wafer was measured to be 7.6 nm in a 10 μm × 10 μm scanning area using an atomic force microscope; the total thickness variation was measured to be 3.6 nm using an ellipsometry with 49 points.

[0084] Reference Figure 6 Comparing the test results at 0.1 MPa, 0.2 MPa, 0.5 MPa, 1 MPa, 1.5 MPa, 2 MPa, and 2.5 MPa, the surface roughness Ra of the silicon wafer on the insulator decreases with increasing atmosphere pressure; (Refer to...) Figure 7 The total thickness change also decreased as the atmospheric pressure increased. When the atmospheric pressure increased from 0.1 MPa to 2 MPa, the surface roughness Ra of the silicon wafer on the insulator decreased from 7.6 nm to 2.8 nm, and the total thickness change decreased from 3.6 nm to 0.92 nm.

[0085] When the atmosphere pressure is further increased from 2 MPa to 2.5 MPa, the surface roughness Ra of the silicon wafer on insulator decreases from 2.8 nm to 2.6 nm, and the total thickness change decreases from 0.92 nm to 0.83 nm, with smaller changes than some changes in the lower pressure range. Further pressure increases will place higher pressure resistance requirements on the furnace tube and may reduce the driving force for crack propagation. Therefore, using 2 MPa as the upper limit can balance surface roughness, total thickness change, and equipment pressure resistance requirements.

[0086] It should be noted that the technical solutions described in this disclosure can be combined arbitrarily as long as they do not conflict.

[0087] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A method for dicing, characterized in that, include: The bonding wafer is heated and held at a pressure of 0.2 MPa to 2 MPa to allow hydrogen bubbles to grow in the ion implantation layer. The bonding wafer includes a donor silicon wafer and a substrate silicon wafer. The donor silicon wafer is bonded to the substrate silicon wafer through a silicon dioxide layer on the surface of the donor silicon wafer. The ion implantation layer is formed by hydrogen ion implantation into the donor silicon wafer. After the heat preservation is completed, the bonded wafer is cooled down, and after the cooling is completed, the pressure of the atmosphere is reduced to atmospheric pressure. as well as After cooling and depressurization, the bonded wafer is peeled off along the ion implantation layer, so that the silicon layer of the donor silicon wafer located between the ion implantation layer and the silicon dioxide layer is transferred to the substrate silicon wafer, resulting in silicon-on-insulator wafer.

2. The dicing method according to claim 1, characterized in that, The heating and holding are performed inside a furnace tube containing the bonding wafer, and the heating and holding include raising the temperature of the furnace tube to 500°C and holding it at 500°C for 10 to 60 minutes.

3. The dicing method according to claim 2, characterized in that, During the heating and heat preservation process, the pressure of the atmosphere is 2 MPa.

4. The dicing method according to claim 2, characterized in that, The temperature of the furnace tube is increased to 500°C at a heating rate of 3°C / min to 5°C / min.

5. The dicing method according to claim 2, characterized in that, The cooling is carried out inside the furnace tube; after the heat preservation is completed, the temperature of the furnace tube is reduced to below 200°C at a cooling rate of 3°C / min to 5°C / min, and then the pressure of the atmosphere is reduced to atmospheric pressure.

6. The dicing method according to claim 5, characterized in that, The pressure of the atmosphere is reduced to atmospheric pressure at a rate of 0.1 MPa / min.

7. The dicing method according to claim 2, characterized in that, Before the heating and heat preservation, the furnace tube is evacuated until the pressure inside the furnace tube is less than 10 Pa, and then gas is continuously introduced into the furnace tube to form the atmosphere.

8. The dicing method according to claim 7, characterized in that, The gas continuously introduced into the furnace tube is nitrogen or argon.

9. The dicing method according to claim 1, characterized in that, The ion implantation layer is implanted with an energy of 20 keV to 200 keV and a dose of 5 × 10¹. 6 cm - ²~1×10¹ 7 cm - ² is formed by hydrogen ion implantation.

10. A silicon-on-insulator wafer, characterized in that, include: Substrate silicon wafer; A silicon dioxide layer is disposed on one side of the substrate silicon wafer, and the thickness of the silicon dioxide layer is 50 nm to 3 μm; as well as A silicon layer is disposed on the silicon dioxide layer on the side away from the substrate silicon wafer; The silicon layer is the portion of the donor silicon wafer located between the ion implantation layer and the silicon dioxide layer in the dicing method, wherein the dicing method includes: The bonding wafer is heated and held at a pressure of 0.2 MPa to 2 MPa. The bonding wafer includes a donor silicon wafer and a substrate silicon wafer. The donor silicon wafer is bonded to the substrate silicon wafer through a silicon dioxide layer located on the surface of the donor silicon wafer. The donor silicon wafer has an ion implantation layer formed by hydrogen ion implantation. The heating and holding cause hydrogen bubbles to grow in the ion implantation layer. After the heat preservation is completed, the bonded wafer is cooled, and then the pressure of the atmosphere is reduced to atmospheric pressure; and After cooling and voltage reduction, the bonded wafer is peeled off along the ion implantation layer, transferring the silicon layer of the donor silicon wafer located between the ion implantation layer and the silicon dioxide layer onto the substrate silicon wafer, thus obtaining the silicon-on-insulator wafer. The surface roughness Ra of the silicon wafer on the insulator is less than 3 nm in a 10 μm × 10 μm region, and the total thickness variation of the silicon wafer on the insulator is less than 1 nm.

Images