JP2012507870A - Method and apparatus for producing semiconductor-on-insulator structures utilizing directional debonding - Google Patents

Abstract

  The implantation surface of the donor semiconductor wafer is subjected to an ion implantation process to produce a fragile flake in cross section to define a release layer of the donor semiconductor wafer; at least one of the parameters of the fragile flake is at least one of the X and Y axes A semiconductor-on-insulator comprising each step of subjecting the donor semiconductor wafer to a spatial variation step before, during or after the ion implantation step so as to vary spatially across the fragile flakes on one side Methods and apparatus for forming SOI) structures are provided.

Description

Cross-reference of related applications

  This application claims the benefit of priority of US Patent Application No. 12 / 290,362 filed on October 30, 2008 and US Patent Application No. 12 / 290,384 filed on October 30, 2008; The contents of both are hereby incorporated by reference.

  The present invention relates to the manufacture of semiconductor-on-insulator (SOI) structures, such as those with non-circular cross-sections and / or with relatively large cross-sectional areas.

  Semiconductor-on-insulator devices are becoming more attractive as market demand continues to increase. SOI technology is becoming increasingly important for high performance thin film transistors (TFTs), solar cells, and displays such as active matrix driven displays, organic light emitting diode (OLED) displays, liquid crystal displays (LCDs), integrated circuits. And photovoltaic devices. An SOI structure can include a thin layer of a semiconductor material, such as silicon, over an insulating material.

  Various methods for obtaining SOI structures include epitaxial growth of silicon (Si) on a lattice-matched substrate and bonding of a single crystal silicon wafer to another silicon wafer. Further methods include ion implantation, where either hydrogen or oxygen ions are implanted, in the case of oxygen ion implantation, forming an oxide layer buried in a silicon wafer covered with Si, or In the case of hydrogen ion implantation, the Si thin layer is separated (peeled) and attached to another Si wafer having an oxide layer.

  Patent Document 1 discloses a method for producing an SOG (semiconductor-on-glass) structure using a peeling technique. The process includes (i) exposing the silicon wafer surface to hydrogen ion implantation to produce an attachment surface; (ii) bringing the attachment surface of the wafer into contact with a glass substrate; (iii) applying pressure, temperature and voltage to the wafer and the glass substrate. To promote adhesion between them; (iv) separating the glass substrate and the thin layer of silicon from the silicon wafer.

  The above methods are susceptible to undesirable effects in some environments and / or when used in certain applications. With reference to FIGS. 1A-1D, ions such as hydrogen ions are implanted into the semiconductor wafer 20 through the surface 21 such that the implantation dose is uniform with respect to the density and depth of the semiconductor wafer 20.

Referring to FIG. 1A, when an ion such as H ion is implanted into a semiconductor material such as silicon, a damaged site is generated. The layer at the damaged site defines a release layer 22. Some of these lesions nucleate in specific platelets with very high aspect ratios (they have a very large effective diameter and almost no height). Gas derived from implanted ions such as H ₂ diffuses into the platelet to form bubbles with an equally high aspect ratio. The gas pressure of these bubbles may be extremely high and is estimated to be as high as about 10 kbar.

 As indicated by the double arrows in FIG. 1B, the platelets or bubbles grow at an effective diameter until they are close enough to each other that the remaining silicon is too weak to withstand the high gas pressure. Since there is no priority point to start on the separation front, multiple separation fronts occur randomly and many cracks propagate through the semiconductor wafer 20.

 Near the edge of the semiconductor wafer 20, a greater percentage of the implanted hydrogen can escape from the hydrogen-rich surface. This is because of the sink that is close (ie, the sidewall of the wafer 20). More specifically, during implantation, ions (eg, hydrogen protons) decelerate through the lattice structure (eg, silicon) of the semiconductor wafer 20 to move some silicon atoms from the lattice site, A defective surface is produced. When hydrogen ions lose kinetic energy, they become atomic hydrogen, and further define an atomic hydrogen surface. Both the defect surface and the atomic hydrogen surface are not stable in the silicon lattice at room temperature. Thus, the defects (vacancies) and atomic hydrogen move in the direction of each other to form a thermally stable vacancy-hydrogen species. Complex species collectively produce hydrogen-rich surfaces. (On heating, the silicon lattice generally splits along a hydrogen-rich surface).

 Not all vacancies and hydrogen decay into hydrogen-vacancy species. Some atomic hydrogen species diffuse from the hole surface and finally leave the silicon wafer 20. Therefore, some atomic hydrogen does not cause cracks in the release layer 22. Near the edge of the silicon wafer 20, the hydrogen atoms have yet another way to escape the lattice. Therefore, the edge region of the silicon wafer 20 may have a low hydrogen concentration. Lowering the hydrogen concentration results in a higher temperature or longer time required to generate sufficient force to support the separation.

 Thus, during the separation process, a tent-like structure 24 is produced having unseparated edges. At critical pressure, crushing of the remaining semiconductor material occurs along a relatively fragile surface such as the {111} surface (FIG. 1C), completing the separation of the release layer 22 from the semiconductor wafer 20 (FIG. 1D). However, the edges 22A, 22B are outside the major crack plane defined by the damaged site. This non-planar crack is undesirable. Other features of separation include those described as the release layer 22 having “mesas”, where the platelets or bubbles are surrounded by “valleys” where fracture occurs. Note that these mesas and valleys shown in FIG. 1D are not accurate as these details exceed the replication capability at the illustrated scale.

  Without limiting the present invention to any theory of operation, the inventors of the present application consider that the time from the start of separation to the completion of separation is approximately tens of microseconds using the techniques described above. Yes. In other words, the random initiation and propagation of separation is generally about 3000 meters / second. Again, although the present invention is not limited to any theory of operation, the inventors of the present application believe that this separation rate contributes to the undesirable characteristics of the cleaved surface of the release layer 22 described above. Is thinking (FIG. 1D).

  In US Pat. No. 6,057,059, to achieve a “controlled split front”, the wafer is brought to a temperature below that at which separation can be initiated, and multiple impulses of energy are captured at the edge of the substrate 10 near the implantation depth Z0. A uniform ion implantation method is described which includes each step and reaches a uniform depth Z0 into the semiconductor substrate 10. Patent Document 2 describes that the above method improves so-called “random” cracks at least with respect to surface roughness. The present invention employs a directed separation method that is significantly different from the “controlled split front” method of US Pat.

  A problem associated with the separation of the release layer 22 from the semiconductor wafer 20 described above is that it gets worse as the size of the SOI structure increases, especially when the shape of the semiconductor wafer is rectangular. These rectangular semiconductor wafers can be used in applications where multiple semiconductor tiles are connected to an insulating substrate. Further details regarding the manufacture of tiled SOI structures may be obtained from U.S. Patent No. 6,053,077, the complete disclosure of which is hereby incorporated by reference.

US Pat. No. 7,176,528 US Pat. No. 6,010,579 US Patent Application Publication No. 2007/0117354

  To simplify the description, the following description will often relate to the SOI structure. Reference to this particular type of SOI structure is made to facilitate the description of the invention and is not intended to limit the scope of the invention in any way and should not be construed as such. Absent. The abbreviation SOI is used herein to refer to a semiconductor-on-insulator structure and generally includes, but is not limited to, a silicon-on-insulator structure. Similarly, the abbreviation SOG is used to refer to a semiconductor-on-glass structure and generally includes, but is not limited to, a silicon-on-glass structure. The abbreviation SOI includes the SOG structure.

  In accordance with one or more embodiments of the present invention, a method and apparatus used to form a semiconductor-on-insulator (SOI) structure provides an implantation surface of a donor semiconductor wafer for an ion implantation process and is fragile in cross section. Generating a flake to define a release layer of the donor semiconductor wafer; such that one or more parameters of the fragile flake vary spatially across at least one of the X and Y axes of the wafer Each step is provided to subject the donor semiconductor wafer to a spatial variation step before, during, or after.

  The spatial variation process facilitates the separation feature of the release layer from the donor semiconductor wafer so that these separations can be directed and / or temporarily controlled.

  As parameters, one or more of the following may be included alone or in combination: (i) density of nucleation sites obtained from the ion implantation step; (ii) weakness from the implantation surface (or reference surface). The depth of the flakes; (iii) an artificially created lesion location (e.g., blind hole) through the injection surface; and (iv) nucleation of defect sites and / or using temperature gradients; Increased pressure across a fragile flake.

  The method and apparatus further provide for raising the donor semiconductor wafer to a temperature sufficient to initiate separation of the fragile flakes from points, edges, and / or regions in the fragile flakes. The donor semiconductor wafer may be subjected to additional temperatures sufficient to continue the separation, substantially along the fragile flakes, as a function of the variation parameter.

  Other aspects, features, advantages, etc. will become apparent to those skilled in the art from the description of the invention herein taken in conjunction with the accompanying drawings.

  For the purpose of illustrating various aspects of the invention, the presently preferred form is shown in the drawings, but it will be understood that the invention is not limited to the precise arrangements and instrumentalities shown.

The block diagram which shows the peeling method according to a prior art. 1 is a block diagram illustrating a stripping method according to one or more aspects of the present invention. FIG. 4 is a top view of a donor semiconductor wafer having spatially varying parameters associated with a weakened layer or a piece therein according to one or more aspects of the present invention. A plot showing the spatial variation parameters of FIG. 3A. FIG. 3B is a plot showing that the spatial variation parameter of FIG. FIG. 6 is a top view of each donor semiconductor wafer having additional spatial variation parameters in accordance with one or more further aspects of the present invention. FIG. 3 is a simplified diagram of some ion implanters that can be adapted to achieve spatial variation parameters of a donor semiconductor wafer. An ion implantation method that can be adapted to achieve a spatial variation density of nucleation sites in a donor semiconductor wafer. An ion implantation method that can be adapted to achieve a spatially varying implantation depth in a donor semiconductor wafer. An ion implantation method that can be adapted to achieve a spatially varying implantation depth in a donor semiconductor wafer. The graph which illustrates the relationship between the inclination angle of ion implantation and the implantation depth. The graph which illustrates the relationship between the inclination angle of ion implantation and the implantation depth. An ion implantation method that can be adapted to achieve a spatially varying ion implantation distribution width in a donor semiconductor wafer. An ion implantation method that can be adapted to achieve a spatially varying ion implantation distribution width in a donor semiconductor wafer. The graph which illustrates the relationship between an ion implantation inclination angle and scattering. Further ion implantation methods that can be adapted to achieve a spatially varying ion implantation depth in the donor semiconductor wafer. Further ion implantation methods that can be adapted to achieve a spatial variation distribution of defect sites in the donor semiconductor wafer. Further ion implantation methods that can be adapted to achieve a spatial variation distribution of defect sites in the donor semiconductor wafer. A time-temperature profile technique that can be adapted to achieve a profile of spatially varying parameters in a donor semiconductor wafer.

  Referring to these figures, wherein like numerals indicate like elements, FIGS. 2A-2B illustrate intermediate SOI structures (especially SOG structures) according to one or more embodiments of the present invention. . The intermediate SOG structure comprises an insulating substrate such as a glass or glass-ceramic substrate 102 and a donor semiconductor wafer 120. The glass or glass-ceramic substrate 102 and the donor semiconductor wafer 120 are bonded using any method recognized in the art, such as bond formation, melting, and adhesion.

  Prior to bonding the glass or glass-ceramic substrate 102 and the donor semiconductor wafer 120, the donor semiconductor wafer 120 includes an exposed implant surface 121. The implantation surface 121 of the donor semiconductor wafer 120 is subjected to an ion implantation process to produce a fragile flake 125 in the cross section that defines the release layer 122. The fragile flakes 125 are located substantially parallel to a reference plane (which can be anywhere, not shown) defined by XY orthogonal axes. The X axis is shown from left to right in FIG. 2A, and the Y axis is orthogonal to the X axis into the page (and thus not shown).

  The donor semiconductor wafer 120 is subjected to a spatial variation process before, during or after the ion implantation process so that the separation characteristics of the release layer 122 from the donor semiconductor wafer 120 can be directed and / or temporarily controlled. To be served. While not intending to limit the invention to any theory of operation, these directed and / or temporary controllability results in release layer 122 and donor semiconductor wafer 120 (after separation). It is believed that separation characteristics such as a smoother exposed surface can be improved. These directed and / or temporary controllability improves, for example, the generation of the release layer 122 and the exposed surface edge of the donor semiconductor wafer 120 at the major crack surface defined by the fragile flakes 125. As a result, it may be possible to improve the edge characteristics.

  A directed and / or temporarily controllable feature for separation of the release layer 122 from the donor semiconductor wafer 120 is a spatial parameter that spans at least one of the X and Y axes of the fragile flake 125. Can be achieved in many ways, such as changing The parameters may include one or more of the following, alone or in combination: (i) density of nucleation sites obtained from said ion implantation step; (ii) from implantation surface 121 (or reference surface). The depth of the fragile flake 125; (iii) an artificially created damage location (eg, blind hole) through the injection surface 121 to at least the fragile flake 125; and (iv) nucleation and / or temperature of the defect site. Pressure rise across the fragile flake 125 using a gradient.

  As shown by arrows A in FIGS. 2A-2B, the directed and / or temporarily controllable feature of separation of the release layer 122 from the donor semiconductor wafer 120 results in a fragile flake as a function of time. Propagation of separation from one point, edge, and / or region of 125 to another point, edge, and / or region. This is generally achieved as follows: First, as described above, one or more parameters are spatially varied across the fragile flake 125, and then the donor semiconductor wafer 120 is moved to these points. , From the edges and / or regions to a temperature sufficient to initiate separation in the fragile flakes 125. Thereafter, the donor semiconductor wafer 120 is raised along the fragile flake 125 along the substantially directed fragile flake 125 to a temperature sufficient to continue further separation as a function of spatial variation parameters. Let The variation parameter preferably has an elevated temperature time-temperature profile of about a few seconds, and the propagation of separation along the fragile flake 125 occurs for at least longer than one second.

  Referring to FIGS. 3A-3C, more details regarding the spatial variation of one or more parameters across the fragile flake 125 are illustrated. FIG. 3A is a top view of the donor semiconductor wafer 120 shown through the implantation surface 121. Changes in shading on the X axis are spatial variations in parameters (eg, density of nucleation sites, pressure within the site, degree of nucleation, distribution of artificially created damaged sites (holes), injection depth, etc.) Represents. In the illustrated example, one or more parameters change from one edge 130A of the X axis to the opposite edge 130B of donor semiconductor wafer 120 (and its fragile flake 125), and vice versa.

  Referring to FIG. 3B, the separation parameter graph shows a cross-sectional profile as a function of the X-axis, such as the density of nucleation sites in the fragile flake 125, for example. Alternatively or additionally, the separation parameter can be one of the following: pressure in the nucleation site, degree of nucleation, distribution of artificially created damage sites (holes), etc. It can represent more than one. Referring to FIG. 3C, the separation parameter graph represents, for example, a cross-sectional profile of the fragile flake 125 depth (eg, corresponding to ion implantation depth) as a function of the X-axis.

  Although it is not intended to limit the invention to any of the theory of operation, the propagation of separation in the direction of edge 130A to edge 130B (shown by the dashed arrows) indicates that the density of nucleation sites is at edge 130A. It occurs when it is relatively high, and is believed to reduce to a low density of nucleation sites at spatial locations in the direction of edge 130B. This theory can also be applied in relation to other parameters such as the gas pressure in the nucleation site, the degree of binding of the nucleation site before separation, and the distribution of artificially created damage sites (holes). it is conceivable that. However, with respect to parameters related to the depth of the fragile flake 125, the propagation of separation from the edge 130B to the edge 130A (shown by the solid arrow) is the first of the flakes 125 where the substantially lower depth is weak. It is believed to occur when present along edge 130B and when a relatively high depth exists at a continuous further distance in the direction of edge 130A.

  With reference to FIGS. 4A-4C, further details are shown in connection with the spatial variation of one or more parameters across the fragile flake 125. The figure shows a top view of the donor semiconductor wafer 120 seen through the implantation surface 121. Again, the change in shading in the X and Y axes is a space in parameters such as density of nucleation sites, pressure in the site, degree of nucleation, distribution of artificially created damaged sites (holes), depth of injection, etc. This represents a dynamic fluctuation. In each case shown, the parameters vary spatially on both the X and Y axes.

  With particular reference to FIG. 4A, the shading starts from two edges 130A, 130D and proceeds in the direction of the other edges 130B, 130C and varies over successive further distances in both the X and Y axes. Can represent spatial changes. Consistent with the discussion above, when considering the density parameter of the nucleation site, the propagation of separation (shown by the dashed arrows) is higher for the edges 130A, 130D when the higher density begins at the edges 130A, 130D. It is considered that it spreads radially from the corner toward the center of the wafer 120 and toward the other edges 130B and 130C. This theory is also applied in relation to other parameters such as gas pressure within the nucleation site, degree of fusion of the nucleation site prior to separation, and the distribution of artificially created damage sites (holes). It is considered possible. However, with respect to parameters related to the depth of the fragile flake 125, the propagation of separation (indicated by solid arrows) will cause the edges 130B, 130C when a shallower depth is initiated along the edges 130B, 130C. From the corners of the wafer 120 toward the center of the wafer 120 and toward the other edges 130A and 130D, it is considered that they spread radially.

  With particular reference to FIGS. 4B and 4C, the shading can represent a spatial change in parameters that starts at all edges 130 and changes toward the center of the donor semiconductor wafer 120, and vice versa. .

Further details are provided on specific parameters of the spatial variation in density of nucleation sites resulting from ion implantation across the fragile flake 125 in one or both of the X and Y axes. Whatever technique is used to achieve these spatial variations, the highest density of nucleation sites is about 5 × 10 ⁵ sites / cm ² of one or more edges, points, or regions of the fragile flake 125 Preferably, the nucleation sites have a minimum density of about 5 × 10 ⁴ sites / cm ² of fragile flakes 125 spaced from them. Judging from variations in other methods, the difference between the highest density of nucleation sites and the lowest density of nucleation sites will be up to about 10 times.

  According to one or more aspects of the present invention, the density of nucleation sites within the fragile flake 125 can be spatially varied by changing the dose of the ion implantation process. For background purposes, the fragile flake 125 (and thus the release layer 122) is created by subjecting the implantation surface 121 to one or more ion implantation steps. There are a number of ion implantation techniques, machines, etc. that can be used in this context, but one suitable method is that the implantation surface 121 of the donor semiconductor wafer 120 is subjected to a hydrogen ion implantation process and the donor semiconductor wafer 120. It shows that the generation of the release layer 122 at least starts.

  Referring to FIG. 5A, an Axcelis NV-10 type batch implanter that can be modified for use in spatially varying the density of nucleation sites in the fragile flake 125 by changing the dose of implanted ions. A simplified diagram is shown.

  The multi-donor semiconductor wafer 120, which in this case is a rectangular tile, can be azimuthally distributed with a fixed radius on the platen 200 relative to the accompanying ion beam 202 (facing the page). The rotation of the platen 200 provides a pseudo-X-scan (dX / dt), while the mechanical movement of the entire platen 200 provides a Y-scan (dY / dt). The term pseudo-X-scan means that in a small radius platen 200, the X scan is somewhat bent compared to a larger radius platen 200 and thus rotates like this. The platen 200 is used because a complete straight scan cannot be obtained. Adjustment of the X scan rate and / or Y scan rate results in spatial variation in dose. Increasing the Y scan speed as the ion beam 202 moves radially toward the center of the platen 200 has traditionally been used to ensure a uniform dose. In fact, the conventional idea in the art is to achieve a spatially uniform dose, and since the angular velocity for the donor semiconductor wafer 120 decreases as it approaches the center of the platen 200, the Y scan rate is corresponding. Must be increased. However, according to the present invention, spatially varying doses can be achieved without a conventional scanning protocol, resulting in, for example, the patterns of FIGS. 3A and 4A. For example, when the ion beam 202 moves in the radial direction toward the center of the platen 200, the Y scan speed is made uniform. Alternatively, the Y scan speed could be reduced as the ion beam 202 moves radially toward the center of the platen 200. Those skilled in the art will recognize other possibilities from the disclosure herein. Another method is to change the beam energy as a function of scan speed and position. These changes can be achieved by modifying the injector control algorithm in software, the electronic interface between the control software, and the end station drive, or other mechanical modifications.

  Referring to FIG. 5B, a single substrate XY implanter that can be modified for use in spatial variation in the density of nucleation sites in the fragile flake 125 by varying the dose of implanted ions. A simplified diagram is shown. In this case, the electron beam 202 scans much faster than a mechanical substrate scan (of FIG. 5A). Again, the conventional idea in the art is to achieve a spatially uniform dose and set the X and Y scan rates and beam energy so that a uniform dose is achieved. Again, spatially varying doses can be achieved without performing conventional scanning protocols. Significant spatial variation in the injected dose can be achieved through multiple combinations of scan rates and / or beam energy of variables X and Y. Univariate or bivariate slopes can occur, for example, vertically or horizontally through these variations resulting in the patterns of FIGS. 3A, 4A, 4B and 4C.

  Referring to FIG. 5C, a simplified diagram of an implanter according to the ion shower technique is shown. Ribbon beam 204 originates from an expanded ion source. According to the prior art, a single uniform velocity scan (proportional to uniform beam energy in orthogonal directions) can achieve the conventional ideal, i.e. spatially uniform dose. However, according to various aspects of the present invention, a univariate gradient (eg, 90 degrees rotated of FIG. 3A) is generated through variations in the mechanical scan speed of the donor semiconductor wafer 120 through the ribbon beam 204. Can be done. The twist of the donor semiconductor wafer 120 by some angle relative to the ribbon beam 204 can be combined with variations in mechanical scan speed to produce a spatial variation in dose in a manner similar to that of FIG. 4A. Alternatively or additionally, the spatially varying beam current along the beam source provides a gradient orthogonal to the scan direction and provides an additional free angle to provide a spatially varying dose of interest. Will generate.

Substantially independent of the particular infusion technique used to achieve variation in dose and regardless of the location of the highest dose (eg, along one or more initial edges, initial points, or initial regions) The highest dose is within some desirable range in units of atoms / cm ² , and the minimum dose therefrom in at least one of the X and Y axes is some other in units of atoms / cm ² . Within the desired range. The difference between the highest dose and the lowest dose can be about 10-30% with a maximum tolerance of about 3 times. In some applications, a difference of at least about 20% has been found to be important.

  According to one or more further aspects of the present invention, the density of nucleation sites within the fragile flake 125 is determined in a substantially uniform manner to achieve the fragile flake 125 in a substantially uniform distribution. The first ionic species may be spatially changed by implantation. Thereafter, the donor semiconductor wafer 120 can be implanted with a second ion species in a substantially non-uniform manner. The non-uniform implantation is accomplished such that the second ionic species results in the transfer of atoms to the fragile flake 125, resulting in a spatial density change of the nucleation sites across the fragile flake 125.

  As an example, the first ionic species can be hydrogen ions and the second ionic species can be helium ions.

  Non-uniform implantation can be performed using the techniques described above, as described later herein, or obtained from other sources. For example, the dose of the second ionic species can vary spatially. Changing the dose of a second ionic species (such as He ions) results in a subsequent non-uniform migration of the second ionic species to the location of the first ionic species, thereby causing a non-uniform density of nucleation sites. To achieve. This change probably also changes the platelet pressure, which can be beneficial.

  Alternatively, the non-uniform implantation of the second ionic species can include implanting the second ionic species across the donor semiconductor wafer 120 to a spatially varying depth. Any known technique for ion implantation to a uniform depth can be modified by those skilled in the art in accordance with the teachings herein to achieve a non-uniform depth profile. As a background, it is known that He ions can be implanted deeper than H, for example, twice or more deeply. As the wafer temperature increases, most of the He ions move to the shallower H ion implantation site, providing a gas pressure for later separation. According to the first aspect of the present invention, the damage caused by the deeper buried He is located at a depth away from the shallow H ion implantation of the donor semiconductor wafer 120 and reaches it within a certain time. The number of will decrease. The opposite is true for He ions that are not implanted too deep, resulting in a spatial variation density of nucleation sites across the fragile flake 125.

 Theoretically, a spatially varied density of nucleation sites can be achieved regardless of the order of the first and second ion species (eg, whether He is implanted first or H is implanted). The order of the multiple ion implantation steps can contribute to the desired result. In fact, at the same time that the density varies spatially, the implantation order can also have a general effect on the density, depending on the ion species. Intuitively and surprisingly for many skilled technicians, it has been found that the initial implantation of H produces more nucleation sites. It is recognized by skilled technicians that at a given dose, He causes about 10 times more damage than H ions. However, it should be noted that damage caused by He ions (vacancies and interstitial semiconductor atoms, or Frenkel pairs) rapidly self-anneal even at room temperature. Thus, most if not all of the He damage is repaired. H ions, on the other hand, bind to semiconductor atoms such as Si atoms (form Si—H bonds) and stabilize the resulting damage. If H is present prior to He injection, additional nucleation sites are generated.

  With reference to FIGS. 6A-6B, there is shown a further example that may be suitable to achieve a spatial variation in the density of nucleation sites. In this example, as shown in FIG. 6A, spatial variation in the density of nucleation sites is achieved by adjusting the beam angle of the ion beam during the ion implantation process. Although the beam angle can be adjusted in many ways, one of these methods is to tilt the donor semiconductor wafer 120 with respect to the ion beam (eg, dot beam 202) as shown in FIG. 6A. The donor semiconductor wafer 120 has a width (left and right of the page), depth (into the page) and height (up and down of the page). The width and depth can define the X and Y axes, and the height can define the longitudinal axis Lo perpendicular to the injection surface 121. During the ion implantation process, the donor semiconductor wafer 120 is inclined such that its longitudinal axis Lo is at an angle Φ with respect to the direction axis of the ion implantation beam (indicated by a solid arrow). The angle Φ can be about 1 to 45 degrees.

  As the beam source scans from position A to position B under tilt conditions, the width W of the beam 202 changes from width Wa to width Wb at the implantation surface 121 of the donor semiconductor wafer 120, and vice versa. The change in the width W contributes to the density change of the nucleation site obtained from ion implantation in the scan direction (can be set to change along at least one of the X and Y axes).

  The implantation beam 202 can include hydrogen ions having the same (positive) charge. Since particles having the same charge repel each other, the beam 202 becomes wider as the distance from the ion source increases (position A), and becomes narrower when the distance from the ion source is short (position B). The more focused (narrow width Wb) ion beam at position B heats the local area of donor semiconductor wafer 120 more than the less focused (wide width Wa) ion beam at position A. Under high temperatures, hydrogen ions diffuse further from these localities and the proportion of hydrogen ions remains low compared to other regions. As shown in FIG. 6B, this results in a non-uniform distribution of hydrogen (and the density of nucleation sites) in the lateral direction in the fragile flake 125 of the donor semiconductor wafer 120.

  Similar spatial variations in nucleation site density may be achieved by adjusting the angle of the beam source or incorporating some of the known mechanisms for adjusting the collimation of the ion beam 202.

  A further technique suitable for achieving spatial variation in the density of nucleation sites is the use of a two-step ion implantation process. The first ion implantation is performed to implant ions having an effect of attracting the second ion species. Thereafter, a second ion species is implanted. The first ionic species is implanted in a spatially non-uniform manner using any of the appropriate techniques described herein above or below. Thus, when the second ion species is implanted and moves toward the first species, the resulting fragile flake 125 exhibits a non-uniform density of nucleation sites.

  For example, the first ionic species can be based on the material of the donor semiconductor wafer 120, such as the use of silicon ions to implant the silicon donor semiconductor wafer 120. Such Si ions may have a property of capturing a second ionic species such as hydrogen ions. As described above, H ions are bonded to some semiconductor atoms such as Si atoms to form Si—H bonds. By way of example, silicon-in-silicon implantation is described in US Pat. No. 7,148,124, the entire disclosure of which is hereby incorporated by reference. Can be performed at doses and energies known in the art. However, unlike the prior art, the spatial density distribution of trapped ion species (Si in this case) is non-uniform (eg, maximum at one edge of donor semiconductor wafer 120 and minimum at the opposite edge). Other variations discussed or discussed herein). Next, a second ion species, such as hydrogen, that can be uniformly distributed is implanted. The amount of hydrogen remaining in the fragile flake 125 of the donor semiconductor wafer 120 will depend on two factors: (1) a concentrated distribution of sites capable of capturing the second ion species, hydrogen, and (2) Available hydrogen (injected hydrogen remaining from the injected dose).

  Note that the non-uniform spatial distribution of ionic species may be reversed to achieve similar results. For example, after the first ionic species are uniformly implanted, the second ionic species are non-uniformly implanted. Alternatively, both injections may be spatially non-uniform. The non-uniform distribution of the second ionic species (eg, hydrogen) within the fragile flake 125 results in the highest concentration point, edge or region of hydrogen, which in other words is the lowest temperature at which cracking begins. Position.

  Again, referring to FIGS. 2A-2B, arrow A illustrates a directed and / or temporarily controllable feature of separation of the release layer 122 from the donor semiconductor wafer 120, where fragile Separation propagating from one point, edge, and / or region of a thin flake 125 to another point, edge, and / or region is achieved as a function of time. In the context of spatial variation in the density of nucleation sites, the donor semiconductor wafer 120 is raised to a temperature sufficient to initiate separation in the fragile flakes 125 from the highest density points, edges, and / or regions. The It was found that high-concentration hydrogen in silicon can be separated at a temperature of 350 ° C. or lower, while silicon containing low-concentration hydrogen is separated at a high temperature such as 450 ° C. or higher. The donor semiconductor wafer 120 is raised over the fragile flake 125 along the substantially directed fragile flake 125 to a temperature sufficient to continue further separation as a function of spatial density variation. The

  Further details are provided in relation to specific parameters of the spatial variation in the depth of the fragile flake 125 obtained from one or both of the ion implanted X-axis and Y-axis. Whatever technique is used to achieve these spatial variations, the substantially shallow depth is about 200-380 nm and the maximum depth is about 400-425 nm. Judging from the change in another method, the difference between the maximum depth and the minimum depth can be about 5-200%.

  According to one or more aspects of the present invention, the depth of the fragile flake 125 can be spatially varied by adjusting the beam angle of the ion beam during the ion implantation process. Indeed, the method discussed with respect to FIGS. 6A-6B may have applicability to the adjustment of the depth of the fragile flake 125 (the mechanism of temperature change as a function of beam width is a function of the change in depth of the fragile flake 125. Note that this is not considered to be the reason for achieving this).

  Referring to FIGS. 6A and 7A-7B, spatial variation in the depth of the fragile flake 125 can be achieved by changing at least one of the following: (1) Tilt angle Φ (shown with respect to FIG. 6A) And (2) a twist about the longitudinal axis Lo of the donor semiconductor wafer 120 with respect to the direction axis of the ion implantation beam 202. Tilt and / or twist adjustments are made by adjusting the degree of channeling through the lattice structure of the donor semiconductor wafer 120, where these channels tend to align and the ion beam 202 is on the implantation surface 121. , The ion beam 202 tends not to be aligned. Since the degree of channeling varies spatially, the depth of the fragile flake 125 also varies.

  The angle Φ can be about 1 to 10 degrees, and the twist angle can be about 1 to 45 degrees.

  As previously estimated and with further reference to FIGS. 7C and 7D, the injection depth decreases as the slope increases. At relatively small angles (eg, 0-10 degrees), the relationship between implantation depth and tilt is dominated by channeling. At relatively larger angles, the cosine effect dominates. In other words, the thickness of the obtained release film is essentially proportional to the cosine of the implantation angle.

  Alternatively, or in addition, the spatial variation process may be such that when the ion beam 202 scans over the implantation surface 121 of the donor semiconductor wafer 120, the depth of the fragile flake 125 from the implantation surface 121 extends across the donor semiconductor wafer 120. Changing the energy level of the ion beam 202 to vary spatially may be included.

  As shown in FIG. 7B, the technique results in a non-uniform depth in the lateral direction of the fragile slice (or implantation depth) of the donor semiconductor wafer 120.

  In connection with adjusting the tilt of the donor semiconductor wafer 202, a further parameter that can be utilized to achieve the spatial variation is the width of the ion deposition distribution (or dispersion). As shown in FIG. 8A, the width (up and down) of the ion distribution through the fragile flake 125 varies as a function of the tilt angle (or more generally the beam angle) of the donor semiconductor wafer 120. Thus, by changing the tilt angle, a spatially varying distribution width can be achieved in the fragile flake 125 (shown in FIG. 8B). While not intending to be limited to any theory of operation, the portion of the fragile flake 125 having a narrower distribution width is separated at a lower temperature than the portion of the fragile flake 125 having a wider distribution width. it is conceivable that. Thus, the directed and / or temporarily controllable feature of separation of the release layer 122 from the donor semiconductor wafer 120 is from one point, edge, and / or region of the fragile flake 125 to another point, edge. It is believed that separation that propagates into and / or into the region can be achieved as a function of time and temperature.

  Referring to FIG. 8C, the additional data regarding the effect of tilting in the dispersion again has some influence on the width of the injection profile. The dose used for both infusions shown in FIG. 8C is the same. Although the peak H concentration is different, both implants delaminate. Thus, the difference in slope change between ± 0.1 degrees and ± 3 degrees is significant for dispersion.

  Referring to FIGS. 9A-9D, another technique for spatial variation in the depth of the fragile flake 125 is such that the depth of the fragile flake 125 from the implantation surface 121 varies spatially across the donor semiconductor wafer 120. The donor semiconductor wafer 120 is subjected to a material removal step after implantation. As shown in FIG. 9A, the donor semiconductor wafer 120 can be subjected to several deterministic polishing steps or plasma-assisted chemical etching (PACE). These techniques allow in-situ control of the amount of material removed during the polishing stage. Other methods including reactive ion etching (RIE), chemical mechanical polishing (CMP), and wet chemical etching may also have regular and reproducible removal of non-uniform material from exposed surfaces. . Using one or more of these or other techniques, capture slight changes in the depth of the fragile flake 125 from the injection surface 121, such as those shown in FIGS. 3A, 4A, 4B, 4C, and others. But you can. The ion implantation step prior to material removal may be spatially uniform or non-uniform.

  Referring to FIGS. 9B and 9C, the spatial variation process is performed in a spatially non-uniform manner so that ion penetration is prevented to varying degrees as the ion beam 202 scans over the implantation surface 121. It may include using a mask 220A or 220B on the implantation surface 121 of the donor semiconductor wafer 120. Masking film 220 includes silicon dioxide, organic polymers such as photoresist, and others. Possible deposition techniques include plasma enhanced chemical vapor deposition (PECVD), spin coating, polydimethylsiloxane (PDMS) stamping, and the like. The thickness of the masking film 220 can be less than or equal to the depth of the fragile flake 125 of interest. Since the depth at which ions are implanted is determined by the accompanying ion energy, the disturbing effect of the mask 220 is a spatial modulation of the depth in the donor semiconductor wafer 120, primarily in the implanted ion species. Depending on the characteristics of the deposited mask 220, the desired characteristics may be achieved by adding length to the ion path, scattering of ions to change the degree of channeling, or other phenomena.

  As shown in FIG. 9D (illustrated to be shallow at all edges of the fragile flake 125 and deeper towards its center), either after bonding to the substrate 102 or during bonding. The donor semiconductor wafer 120 is raised to a temperature sufficient to initiate separation in the fragile flake 125 from the minimum depth points, edges, and / or regions. The donor semiconductor wafer 120 follows a substantially directed fragile flake 125 to a temperature sufficient to continue further separation as a function of depth spatial variation from minimum to maximum depth. Be raised.

  Referring to FIGS. 10A-10D and 11, the spatial variation process includes drilling one or more blind holes 230 in at least the fragile flake 125 through the injection surface 121, preferably through the fragile flake 125. (FIG. 10B). While not intending to limit the present invention to any theory of operation, during or after bonding to the substrate 102 (FIG. 10C), the elevated temperature of the donor semiconductor wafer 120 is free of such holes. It is believed that separation begins in blind hole 230 (FIG. 10D) prior to separation in position. As shown in FIG. 11, opening many blind holes 230 through the injection surface 121 can result in an uneven spatial distribution of these holes. Thus, the rise of donor semiconductor wafer 120 to a temperature sufficient to initiate and continue separation along substantially fragile flakes 125 is a function of the distribution of many blind holes 230 from the highest concentration to the lowest concentration. Can be achieved directionally.

  Referring to FIGS. 12A-12B, the spatial variation process is performed such that the density or temperature of nucleation sites at each spatial location throughout the fragile slice 125 varies spatially across the donor semiconductor wafer 120. Subjecting to a non-uniform time-temperature profile. For example, the temperature gradient shown in FIG. 12A applies a higher temperature on the left side compared to the right side of the wafer 120. This temperature gradient can be applied in-situ before or during bonding to the substrate 102. Over time, if the processing time is kept below the separation threshold at a given processing temperature, at least one of the defect site nucleation and the gas pressure there is spatially across the wafer 120 as a function of the temperature gradient. , Increase to various degrees across the fragile flake 125 (see FIG. 12B). The separation threshold time for a given processing temperature is predicted according to the Arrhenius relationship, where the separation threshold time is exponentially proportional to the inverse of the processing temperature. The target parameter is the ratio of the processing time to the separation threshold time at the processing temperature. The data described herein, the aforementioned spatially varying parameter profile, or another desired parameter profile may be achieved by preparing a processing time-separation time ratio profile. Thus, the donor semiconductor wafer 120 is raised from the maximum processing time-separation time ratio point, edge, and / or region to a temperature 125 sufficient to initiate separation in the fragile flakes. In the illustrated example, the maximum processing time-separation time ratio is shown on the left side of the wafer 120. Next, the donor semiconductor wafer 120 is directed along the substantially fragile flake 125 as a function of the variation time-temperature profile from the maximum processing time-separation time ratio to the shortest processing time-separation time ratio. Further, the temperature is raised to a temperature sufficient to continue the separation. Depending on the material characteristics and other factors including ionic species, dosage, and depth of implantation, the substantially high processing time-separation time ratio is about 0.9-0.5, with minimal processing. The time-separation time ratio is about 0-0.5.

  Various mechanisms can be used prior to bonding or in-situ during bonding to achieve a spatially varying time-temperature profile. For example, using one or more spatially non-uniform conductive, convective, or radiant heating techniques (hotplate, laser irradiation, visible / infrared lamp, or others), the donor semiconductor wafer 120 may be It can be heated. Control time / temperature gradients can be achieved by direct or indirect thermal contact (conduction) to achieve any desired profile. Various profiles based on computer control or programming can be achieved using an addressable two-dimensional array of hotplate elements. For example, localized infrared radiation using a lamp such as that used for rapid thermal annealing (radiation) can be used and / or visible or near infrared laser radiation is localized and spatially non-uniform May be used to provide free heating (radiation). Alternatively, application of a uniform or non-uniform temperature profile through any means and application of spatially non-uniform cooling mechanisms, such as direct contact (conductive), or gas or fluid jets (conductive / convective) May be used to achieve the desired time-temperature gradient.

  Again, these heating / cooling techniques can be used prior to bonding or in-situ. In the context of in-situ bonding technology, for example, U.S. patent application entitled HIGH TEMPERATURE ANODIC BONDING APPARATUS, the complete disclosure of which is incorporated herein by reference. The joining device described in 11 / 417,445 can be adapted for use in accordance with the present invention. Management of thermal radiation loss is controlled in the joining device, thus achieving a time-temperature gradient through the incorporation of infrared reflective elements around the joining device to minimize radiation loss and maximize edge temperature Can be used. Conversely, management of thermal radiation loss in the bonding device can be controlled through the incorporation of a cooled infrared absorber to maximize radiation loss and minimize edge temperature. Many variations on the above themes can be used to achieve the desired time-temperature gradient.

  Although the invention has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. Accordingly, many modifications may be made to the illustrated embodiments, and other arrangements may be devised as defined by the appended claims without departing from the spirit and scope of the present invention. It should be understood.

Claims (5)

A method for forming a semiconductor-on-insulator (SOI) structure, comprising:
The implantation surface of the donor semiconductor wafer is subjected to an ion implantation process, a fragile flake is generated in the cross section to define a release layer of the donor semiconductor wafer,
Donor before, during or after the ion implantation step so that the density of nucleation sites obtained from the ion implantation step varies spatially across the fragile flake in at least one of the X and Y axes Subjecting the semiconductor wafer to a spatial variation process;
A method comprising each step. The highest local density of nucleation sites is present in the first region of the fragile flakes of about 5 × 10 ⁵ sites / cm ² ;
A minimum density of nucleation sites is present in the second region of the fragile flakes of about 5 × 10 ⁴ sites / cm ² ;
The method of claim 1, wherein the second region is spaced from the first region in at least one of the X and Y axes.   3. The highest density of nucleation sites in the first region of the fragile flakes is at least 10 times the lowest density of nucleation sites in the second region of the fragile flakes. The method described.   Further comprising raising the donor semiconductor wafer from the highest density point, edge and / or region of the nucleation site to a temperature sufficient to initiate separation of the fragile flakes. The method according to any one of claims 1 to 3.   Raise the donor semiconductor wafer from the highest density to the lowest density, substantially as a function of density change at the nucleation site, to a further temperature sufficient to continue the separation along the directional fragile flake The method according to claim 4, further comprising the step of:

Images