CN116544174A - Wafer clamping and overturning device and method - Google Patents

Abstract

The invention discloses a wafer clamping and overturning device, which comprises: the clamping assembly comprises an upper clamping claw disc and a lower clamping claw disc, and at least one pair of clamping blocks meshed with each other are arranged at the edges of the upper clamping claw disc and the lower clamping claw disc; the clamping driving piece is connected with the upper clamping claw disc, can drive the upper clamping claw disc to move towards the direction close to or far away from the lower clamping claw disc, and is connected with the clamping assembly, and can drive the clamping assembly to overturn. The wafer clamping and overturning device and method provided by the invention can realize centering and overturning of the wafer at the same time, are compatible with the cavity, and are flexible in control mode.

Description

Method and apparatus for preparing heterostructures with strain reduction by radial compression

The present application is a divisional application of patent application with application date 2013, 12, 30, application number 201810287918.7, entitled "method and apparatus for preparing heterostructures with strain reduction by radial compression".

Cross Reference to Related Applications

The present application claims U.S. provisional application No. 61/747,613 filed on 12/31 2012; U.S. provisional application No. 61/793,999 filed on 3.3.15; U.S. provisional application No. 61/790,445 filed on 3/15 in 2013; priority to U.S. provisional application No. 61/788,744 filed on 3, 3 and 15 in 2013, each of which is incorporated herein by reference.

Technical Field

The present disclosure relates generally to the fabrication of semiconductor heterostructures with reduced strain, and in particular, to heterostructures with semiconductor substrates that conform to surface layers having different lattice constants than the substrate, thereby forming relatively low defect heterostructures.

Background

Multilayer structures comprising a device layer having a device quality surface and a substrate having a lattice structure different from the material of the device layer are useful for many different purposes. These multilayer structures typically comprise multiple layers of materials having different lattice constants. The lattice mismatch between the layers results in the layers being strained. Misfit dislocations form spontaneously in the device layers to relax (relax) the strain between the layers. Such dislocations reduce the quality and utility of the multi-layer semiconductor structure.

There is a continuing need for methods for relaxing the strain between lattice-mismatched semiconductor layers and for methods that result in substrates and device layers that are substantially free of dislocations.

Disclosure of Invention

An aspect of the present disclosure is directed to a method for relaxing strain in a heterostructure comprising a substrate, a surface layer disposed on the substrate, and an interface between the substrate and the surface layer. The substrate includes a central axis, a back surface generally perpendicular to the central axis, and a diameter extending across the substrate through the central axis. A dislocation source layer is formed in the substrate. The substrate is radially compressed to generate dislocations and the dislocations are slid from the dislocation source layer toward the surface layer.

Another aspect of the present disclosure is directed to a method for preparing a relaxed heterostructure. A surface layer is deposited on a front surface of the semiconductor substrate, creating a strain between the surface layer and the substrate. A dislocation source layer is formed in the substrate. The strain in the surface layer and the substrate is relaxed by radially compressing the substrate to generate dislocations and to slip the dislocations from the dislocation source layer toward the surface layer.

Another aspect of the present disclosure is directed to a method for radially compressing a semiconductor structure in a device. The structure has a front surface, a back surface, and a circumferential edge. The device includes a structure holder including a top plate and a back plate for contacting the structure adjacent a circumferential edge of the structure. The top plate is adapted to contact the front surface of the structure and the back plate is adapted to contact the back surface of the structure. A peripheral cavity is formed between the top plate, the back plate, and the peripheral edge of the structure. The pressure in the peripheral cavity is varied to radially compress the structure.

Drawings

FIG. 1 is a schematic cross-sectional view of a silicon heterostructure;

FIG. 2 is a flow chart illustrating a method for fabricating a heterostructure;

FIGS. 3-4 are cross-sectional views of a semiconductor structure and a structure clamp for compressing the semiconductor structure;

FIG. 5 is a cross-sectional view of another embodiment of a semiconductor structure and a structure clamp for compressing the semiconductor structure;

FIGS. 6-7 are cross-sectional views of a semiconductor structure having a coating thereon and the structure clamp of FIG. 3;

FIG. 8 is a schematic cross-sectional view of a semiconductor structure and an apparatus for compressing the structure having the structure clamp of FIG. 3 disposed therein;

FIG. 9 is a cross-sectional view of another embodiment of a semiconductor structure and a structure clamp for compressing the semiconductor structure;

FIG. 10 is a cross-sectional view of a plurality of semiconductor structures and a structure clamp for compressing the semiconductor structures;

FIG. 11 is a top view of another embodiment of a structure clamp for compressing a semiconductor structure;

FIG. 12 is a cross-sectional view of another embodiment of a semiconductor structure and a structure clamp for compressing the semiconductor structure;

FIG. 13 is a cross-sectional view of another embodiment of a semiconductor structure having a trench and a structure clamp for compressing the semiconductor structure;

FIG. 14 is a cross-sectional view of a semiconductor structure and the structure clamp of FIG. 13 with a top plate;

FIG. 15 is a cross-sectional view of another embodiment of a semiconductor structure having two slots and a structure clamp for compressing the semiconductor structure;

FIG. 16 is a cross-sectional view of another embodiment of a semiconductor structure and a structure clamp including a presser;

fig. 17 is a top view of the semiconductor structure and structure holder of fig. 16;

FIG. 18 is a top view of another embodiment of a structural clip and a structural clip having segments; and

fig. 19 is a cross-sectional view of another embodiment of a semiconductor structure and a structure holder having a flange.

Corresponding reference characters indicate corresponding parts throughout the several views of the drawings.

Detailed Description

In accordance with one or more aspects of the present disclosure, a heterostructure having reduced strain between a substrate and a surface layer having a different lattice constant than the substrate is prepared, such as by the method of fig. 2. The surface layer is also referred to herein as an "epitaxial layer", "heteroepitaxial layer", "deposited film", "heterolayer" or "deposited layer". Heterostructures can be formed with a substantially relaxed surface layer and a reduced concentration of misfit dislocations, also known as threading dislocations.

Generally, the methods of the present disclosure include forming a dislocation source layer in a semiconductor substrate, depositing a hetero-layer on the substrate either before or after formation of the dislocation source layer, and radially compressing the hetero-structure to create (i.e., "activate") dislocations, and sliding the dislocations from the dislocation source layer toward the surface layer. By applying compression to the substrate, activation of the source layer and sliding of dislocations from the source layer toward the interface with the deposited layer simultaneously occur. In one or more steps, and in various combinations, stresses are applied to activate and slip dislocations, plastically compressing the heterostructure.

The hetero layer has a lattice constant a _Sl Which is different from the natural lattice constant a of the substrate _S To form a film on the surface of the substrate. In general, the lattice constant a of the heterogeneous layer _Sl Less than the natural lattice constant a of the substrate _S By controlling the creation and sliding of dislocation loops by compression in the substrate, the substrate is plastically deformed and more properly aligned with the crystal lattice of the film, allowing the film to be fully relaxed and to have a reduced density of threading dislocations on the substrate.

The method of the present disclosure has several advantages over conventional methods for relaxing heterogeneous layers. Conventional methods create a large asymmetry in the stress between the film and the substrate, which results in dislocation generation where the stress is greatest (i.e., the film). By confining the dislocation loops to the film, the dislocations leave segments as reduced threading dislocations. Many efforts have been made to minimize the density of such threading dislocations.

In contrast, the methods of the present disclosure induce stress asymmetry with dislocation generation occurring in the substrate (e.g., by weakening the substrate and using a relatively thin film therein when weakening the substrate to avoid dislocation generation). This allows confinement of dislocations to the substrate while forming a dislocation layer at the interface between the substrate and the film. When the dislocation weakens the substrate by various controlled means, an external stress is applied to the system to activate the dislocation. This differs from conventional methods, which result in self-relaxation (i.e., relaxation without the application of external stress) due to relatively large intrinsic, internal stresses. The methods of the present disclosure involve relaxation, rather than by self-relaxation, by weakening and applying external stresses with a relatively thin film at an appropriate temperature, so that no self-relaxation occurs.

I. Semiconductor substrate

Referring to fig. 1, a semiconductor substrate 1 may be any single crystal semiconductor material suitable for use as a substrate for supporting a surface layer, such as deposition of an epitaxial layer by chemical vapor deposition. Typically, the semiconductor substrate is composed of the following materials: silicon, silicon carbide, sapphire, germanium, silicon germanium, gallium nitride, aluminum nitride, gallium arsenide, indium gallium arsenide, or any combination thereof. Typically, the semiconductor substrate is composed of silicon.

The semiconductor substrate 1 is of any shape suitable for use as a substrate for depositing a surface layer and for applying stress to a substrate material as described in more detail below. Typically, the semiconductor substrate has a central axis 2; an interface 3 having a deposited layer 7 and a back surface 4, the substrate-surface layer interface 3 and back surface 4 being generally perpendicular to the central axis 2; a thickness t corresponding to the distance from the interface to the back surface of the substrate; a circumferential edge 5; and a diameter D extending across the substrate through the central axis. It should be noted that for purposes of illustration, the back surface 4 is described as the opposing surface 4 at or near which the dislocation source layer will be formed, and as such is referred to herein as the "opposing surface" and/or the "damaged surface". In this respect, the heterostructure itself and the deposited layer 7 described below are generally concentric with the substrate 1 and also comprise a central axis 2; a circumferential edge 5; and a diameter D, extending across the heterostructure (and surface layer) and through the central axis.

The substrate 1 has any suitable diameter for acting as a substrate on which a semiconductor layer is to be deposited. Typically, the substrate 1 has a diameter of about 150mm or more. Typically, the substrate 1 has a diameter of about 200mm or more, about 300mm or more, or even about 450mm or more. It should be noted that the substrate diameter is the diameter before plastically deforming the heterostructure, in which case the diameter increases or decreases from the illustrated value after plastic deformation as described in more detail below. Alternatively, the substrate has a diameter less than or greater than the stated value before plastic deformation, so that the diameter after plastic deformation is approximately equal to the stated value.

Similarly, the substrate 1 has any thickness t suitable for use as a substrate upon which a semiconductor layer is deposited. For example, the substrate has a thickness t of from about 500 microns to about 1000 microns, typically from about 600 microns to about 1000 microns, from about 700 microns to about 900 microns, or even from about 700 microns to about 800 microns.

In some embodiments, for example, substrate 1 is a single crystal silicon wafer that has been sliced from a single crystal silicon ingot by a Czochralski crystal growing method, has a diameter of about 150mm or more, about 200mm or more, about 300mm or more, or even about 450mm or more, and has a thickness of from about 675 microns to about 1000 microns or even from about 725 microns to about 925 microns.

The surface of the substrate on which the epitaxial layer is deposited is polished to be suitable for depositing the epitaxial layer or is further conditioned prior to chemical vapor deposition. The opposing surface is also polished or alternatively unpolished, i.e., as ground, or as ground and etched, without departing from the scope of the disclosure. In various embodiments, the opposing surface is left in an unpolished state in which, for example, a grinding, such as grinding, or a surface such as grinding and etching will be employed as the dislocation source layer. Alternatively or in addition, the opposing surfaces are damaged to form a dislocation source layer, as described in more detail below.

It should be noted that Czochralski grown silicon typically has a silicon growth on the order of 5×10 ¹⁷ Up to about 9X 10 ¹⁷ Atoms/cm ³ (ASTM standard F-121-83) oxygen concentration in the range. Generally, single crystal silicon wafers for substrates in the present disclosure have oxygen concentrations that fall within or even outside the range typically achieved by the czochralski method, with no excess oxygen concentration provided to prevent dislocation activation and slip.

II deposited surface layer

The surface layer 7 is positioned on the front surface of the substrate 1. The deposited layer 7 is any single crystal semiconductor material suitable for deposition as an epitaxial layer by chemical vapor deposition. In general, the heterogeneous layer includes a lattice constant a _Sl Which is smaller than the natural lattice constant a of the substrate _S . The deposited layer is composed of any suitable material and, as in some embodiments, is composed of the following materials: silicon, silicon carbide, sapphire, germanium, silicon germanium, gallium nitride, aluminum nitride, gallium arsenide, indium gallium arsenide, or any combination thereof. In embodiments where the substrate is comprised of silicon, the heterogeneous layer having a smaller lattice constant comprises, for example, gallium nitride.

Basically any technique generally known in the art is used to form the deposited layer, such as one of the known epitaxial deposition techniques. In general, the thickness of the deposited layers varies greatly without departing from the scope of the present disclosure. For example, the thickness has a substantially uniform thickness and an average thickness of at least about 0.1 microns, at least about 0.5 microns, at least about 1.0 microns, and even at least about 2.0 microns. Alternatively, it is desirable to express thickness in terms of range. For example, the average thickness is typically in the range from about 0.1 microns to about 2.0 microns, such as from about 0.5 microns to about 1.0 microns.

It should be noted that growing the deposited layer on a substrate having a different lattice constant creates equal but opposite stresses in the deposited layer and the substrate. The relative amounts of stress in the deposited layer and the substrate above and below the interface are proportional to the relative thicknesses of the deposited layer and the substrate. As a result, the stress in the deposited layer above the interface is several orders of magnitude greater than the stress in the substrate below the interface. During growth, the stress in the deposited layer may increase until self-relaxation in the deposited layer by forming a layer of misfit or threading dislocations. In order to avoid self-relaxation of the deposited layer, it is therefore preferred to grow a thin deposited layer at least initially on the substrate. The thin layer is then relaxed or partially relaxed to or adjacent to its natural lattice constant by activating or extending dislocations in the substrate, as described in more detail below. If a thicker deposited layer is desired, additional material is deposited after the layer has been sufficiently relaxed.

Essentially any technique generally known in the art is used to form a deposited layer on a substrate. For example, epitaxial deposition techniques (e.g., atmospheric Pressure Chemical Vapor Deposition (APCVD)) may be used; low or reduced pressure CVD (LPCVD); ultra High Vacuum CVD (UHVCVD); molecular Beam Epitaxy (MBE); or Atomic Layer Deposition (ALD)). Epitaxial growth systems include single-wafer or multi-wafer batch reactors.

The surface layer 7 comprises a surface forming the front surface 8 of the heterostructure. As shown in fig. 1, the surface layer 7 extends continuously across the entire diameter of the substrate 1. In some embodiments, the surface layer 7 does not extend continuously across the substrate 1, but rather comprises several discrete segments or "islands" of semiconductor material deposited on the substrate as described further below. For example, the surface layer is disposed over less than about 95% of the substrate, or as in other embodiments, less than about 80%, less than about 60%, less than about 40%, or less than about 20% of the substrate.

III preparation of dislocation Source layer

A dislocation source layer 6 is positioned within the substrate 1 and spaced from the substrate surface over which the epitaxial layer is to be deposited. Typically, the dislocation source layer 6 is located at or near the opposite surface, over which the epitaxial layer has been or will be deposited. For example, if an epitaxial layer were to be deposited on the front surface of the substrate, the dislocation source layer 6 would be located at or near the back surface 4 of the substrate. In such an example, the front surface of the substrate would become the interface between the substrate and the deposited layer 7.

The source layer 6 is present or arranged across the main radial width of the substrate 1. As in the embodiment shown in fig. 1, the source layer 6 extends across the entire diameter of the substrate 1. Although this embodiment is preferred, in other embodiments the source layer does not extend across the entire diameter. Generally, therefore, the source layer 6 will have a radial width that is typically at least about 75%, more typically at least about 85% and still more typically at least about 95% of the radius of the wafer or even at least about 99% of the radius of the wafer. In some embodiments, the source layer 6 extends to within some millimeters of the circumferential edge, for example, to within 1mm of the circumferential edge.

Typically, the source layer 6 comprises any portion of the substrate on which the source layer is provided, provided that it does not comprise a surface on which an epitaxial layer is to be deposited. Typically, the source layer 6 has a thickness of about 100 microns or less, about 50 microns or less, about 25 microns or less, or about 10 microns or less (e.g., from about 1 micron to about 100 microns, from about 1 micron to about 50 microns, from about 1 micron to about 25 microns, or from about 5 microns to about 25 microns). The source layer 6 comprises the back surface of the substrate and extends therefrom. It should be noted that the source layer 6 need not include the back surface of the wafer and extends from the depth of the back surface towards the front surface of the substrate.

The dislocation source layer 6 is any layer capable of generating a measurable concentration of dislocations when subjected to a sufficiently high stress at a sufficiently high temperature. Generally, as described in more detail below with respect to activation of dislocations within the substrate, the dislocation source layer 6 is capable of generating a measurable concentration of dislocations when subjected to compressive stress between about 5MPa and about 100MPa (typically about 15MPa at a temperature between about 500 ℃ and about 1000 ℃).

The dislocation source layer 6 is formed in the substrate 1 either before or after deposition of the surface layer 7. In embodiments where the substrate is a wafer sliced from a single crystal ingot, the dislocation source layer 6 is mechanically damaged by a slicing method, a grinding method, or a grinding method, including as part of the overall wafer process.

Alternatively or additionally, the dislocation source layer 6 is formed partially or entirely by mechanically damaging the back surface of the substrate by one or more of the following methods: grinding the back surface, setting soft lesions by sandblasting the back surface, forming indentations on the back surface, implanting ions in the back surface, and/or combinations thereof.

In some embodiments, dislocation source layer 6 is formed by pressing an array of pins onto the back surface of the wafer to form indentations in the back surface. The indentations are formed unevenly across the surface or in a predetermined pattern. Such patterns are arranged in a specific relationship to the crystal direction of the wafer. For example, the square matrix pattern is disposed at a shallow angle to the 110 direction. This allows dislocations generated at these locations to slip along parallel slip planes and not interact with each other. Moreover, by such a process, there is precise control of dislocation loop density.

In some embodiments, the source layer 6 is formed by implanting ions through the back surface of the substrate. The implanted ions may be electrically plasma, neutral or inert to minimize any impact on the electrical characteristics of the substrate. For example, the implanted ions are selected from the group consisting of silicon, germanium, hydrogen, helium, neon, argon, xenon, and combinations thereof.

Implanting ions to a target depth D relative to the back surface _i . However, as a practical matter, some implanted ions will not travel that distance and others will travel even further distances (i.e., to greater depths relative to the back surface). Actual ion implantation depth from D _i The change is about 5%, 10%, 15%, 20%, 25% or more. This results in a product contained in D _i A region or layer of relatively high concentration ion-implanted amorphous material at or near, wherein the concentration of the implanted ions is from D in the direction of the front surface of the substrate 1 and in the opposite direction _i Descending. Target depth D _i Also known as the projection range of the implanted ions.

For a given implantation energy,the implantation depth is at least partially affected by ion species implantation, as lighter ions tend to penetrate further into the substrate. Thus, for example, at an implantation energy of 50keV, the silicon ions will have an ion energy of about Is to have an average implantation depth of germanium ions +.> Is used for the implantation of the semiconductor device. Generally, it is preferred to implant ions at an energy of at least about 30keV, such as at least about 40keV or even at least about 50 keV. In one application, ions are implanted at an energy of at least about 45keV and less than about 55 keV. The ions and implantation energies selected should be sufficient to form an amorphous layer in the substrate that serves as a dislocation source layer.

Generally, if sufficient energy is used to implant ions of sufficient concentration to form an amorphous layer of silicon, dislocation loops are formed at the ends of the range of implanted ions when subsequently annealed. Typically, although the exact depth is more or less, about under the implanted ionsDislocation loops are formed to a depth of about 300 a. Generally, it is more difficult to form amorphous materials using lower quality elements. Therefore, a higher concentration of low-quality elements must be used to promote adequate damage, while a lower concentration of high-quality elements is sufficient to form amorphous silicon. For example, when the implanted ions are silicon ions, the implanted dose is preferably at least about 2 x 10 ¹⁴ Atoms/cm ² Such as at least about 5 x 10 ¹⁴ Atoms/cm ² Or even at least about 1 x 10 ¹⁵ Atoms/cm ² . In a preferred embodiment, the implanted ions are implanted at a dose of at least about 2X 10 ¹⁵ Atoms/cm ² . By contrast, when the implanted ions are higher mass germanium ions, the implant dose is preferably at least about6×10 ¹³ Atoms/cm ² Such as at least about 1 x 10 ¹⁴ Atoms/cm ² Or even at least about 5 x 10 ¹⁴ Atoms/cm ² . In a preferred embodiment, the implanted ion dose is at least about 1X 10 ¹⁵ Atoms/cm ² 。

In some preferred embodiments, the source layer 6 is formed by grinding the back surface of the substrate. After slicing from a czochralski grown single crystal silicon ingot, the surface is polished to form a silicon wafer using any polishing method typically used in the semiconductor silicon industry. In a particularly preferred embodiment, the back surface is abraded using an abrasion method using a particle size of about 600.

Activation and sliding of misalignment

The dislocation source layer is activated to form dislocations at or near the source layer that can be slid toward the substrate-surface layer interface. In accordance with embodiments of the present disclosure, activation and slip of dislocations is performed after a surface layer has been deposited on a substrate, such that the substrate and/or the surface layer is subjected to strain.

The dislocation source layer is activated by subjecting the dislocation source layer (and typically the substrate) to stress by compressing the substrate at an elevated temperature to cause the formation of dislocations. Compression is applied to the entire substrate in a direction perpendicular to the axis, i.e., in a radial direction, using one or more suitable devices. That is, the wafer is compressed radially inward from the peripheral edge. In this way, dislocations will form at or near the source layer and will slip towards the opposite surface.

Generally, more severely damaged dislocation source layers will be activated at lower stress levels and lower temperatures, while less severely damaged dislocation source layers will be activated at higher stress levels and temperatures. Generally, stress applied by compression of at least about 5MPa, typically from about 5MPa to about 100MPa or from about 10MPa to about 100MPa, is applied to the dislocation source layer at a temperature of between about 550 ℃ and about 1000 ℃. More typically, the stress is from about 10MPa to about 50MPa or from about 10MPa to about 25MPa. Typically, the activation and/or sliding of dislocations is performed at a temperature of from about 650 ℃ to about 1000 ℃, or even from about 700 ℃ to about 1000 ℃. For example, typical stresses will be applied to activate dislocation source layers formed by milling and/or milling, at temperatures greater than about 600 ℃ and even more typically at temperatures greater than about 700 ℃, the stresses being about 15MPa. Otherwise, at even lower stress levels, the higher damaged layer may activate.

The substrate is maintained under stress at an elevated temperature for a sufficient time to activate and slip dislocations. Typically, as described above, the substrate is maintained under stress and at elevated temperature for a period of at least about 10 seconds, and the substrate is maintained under those conditions for a period of at least about 5 hours, at least about 10 hours, or even longer. Typically, the substrate is maintained under stress for a period of time of at least about 1 minute, from about 5 minutes to about 60 minutes, more typically from about 10 minutes to about 45 minutes, and in some embodiments, from about 10 minutes to about 20 minutes, at an elevated temperature. It should be noted that higher stress levels and higher temperatures each tend to reduce the duration required to activate and slip dislocations.

Compression is applied to the substrate alone or, as in other embodiments, to the entire heterostructure (i.e., substrate and heterostructure layer). Furthermore, preferably, the stress applied relatively uniformly (in direction and/or magnitude) by the stress penetrates the heterostructure (e.g., radially and circumferentially). It should be noted that limiting the degree of uniformity of the stress and some variation (radial or circumferential) by the means for compressing the substrate may result from non-uniform distribution of the stress. In some embodiments, the stress is applied at least about 5MPa along the entire circumference of the substrate, or as in other embodiments, at least about 10MPa along the entire circumference of the substrate.

When sufficient stress is applied, dislocations form continuously at the dislocation source layer and slip toward the substrate-surface layer interface. At a given point in time during the application of stress, dislocations are typically uniformly distributed throughout the thickness of the substrate. When the substrate-surface layer interface is reached, dislocations form misfit interface dislocations at the interface. During compression of the substrate, the misfit dislocation density increases at the interface and continues to relax the strain between the surface layer and the substrate. When sufficient mismatch density is established, the strain is eventually balanced.

Dislocations are generated from the dislocation source layer and slip toward the substrate-surface layer interface, the dislocations being substantially parallel to the back and front surfaces of the heterostructure (i.e., being disposed laterally). It is believed that relatively little or even no threading dislocations are generated from the dislocation source layer.

Preferably, as the dislocations are further generated and slip, compression of the substrate stopping at or near the point at which the strain is balanced causes the dislocations to penetrate the surface layer. Once the compression of the substrate is stopped, dislocations in transit in the substrate stop sliding to the interface and no further dislocations are generated (i.e., the dislocations become frozen).

At any given point in the application of stress and heat, the number of dislocations present in the substrate is at least about 1 x 10 ⁵ Dislocation/cm ² Or even at least about 5 x 10 ⁵ Dislocation/cm ² (e.g., from about 1X 10) ⁵ Dislocation/cm ² Up to about 5X 10 ⁷ Dislocation/cm ² Or from about 5X 10 ⁵ Dislocation/cm ² Up to about 1X 10 ⁷ Dislocation/cm ² ). The number density of dislocations is determined using any dislocation loop detection method including, for example, sampling the substrate and subjecting the sample to a delineating (delineating) etchant prior to viewing and counting dislocation loops by microscopy.

In some embodiments, the stress applied to the substrate by compression of the substrate is reduced to a value less than the threshold for dislocation generation from the dislocation source layer, but at this value at a sufficient magnitude to allow existing dislocations to slide further upward toward the interface. In this way, a heterostructure having a substrate substantially free of dislocations is produced. In such an embodiment, the initial stress S is induced by compression of the substrate ₁ Is applied to the substrate to generate and slip dislocations from the source layer to the substrate-surface layer interface. Then, the applied stress is reduced to S ₂ (i.e., S ₂ Less than S ₁ ). Stress S ₂ To be smaller than that generated from dislocation source layerThe stress of the threshold of dislocations and allows existing dislocations to slip further upward toward the interface to produce a substrate that is substantially free of dislocations. S is S ₁ At least about 5MPa, at least about 10MPa, or at least about 25MPa (e.g., from about 5MPa to about 100MPa or from about 10MPa to about 100 MPa). S is S ₂ Is less than about 10MPa, less than about 5MPa, or even less than about 1MPa. Typically, even at a stress of about 1MPa, dislocations will slip at a speed of about 1 micron per second at a temperature of about 850 ℃ or at a speed of about 2.5 microns per second at a temperature of about 900 ℃.

According to lattice constant a at the substrate _S Lattice constant a of semiconductor material with surface layer _SL The difference between the two, the magnitude of the stress, the time and/or the temperature of application of the stress applied to the substrate. A, depending on the substrate material selected and the semiconductor material deposited thereon _SL And a _S Possibly changed. In general, when a _SL Less than a _S When, i.e. when the ratio a _SL /a _S At less than 1, compression effectively relaxes the heterogeneous layer. Ratio a _SL /a _S From about 0.84 to about 0.99, or as in other embodiments, from about 0.90 to about 0.99, from about 0.95 to about 0.99, from about 0.84 to about 0.95, from about 0.84 to about 0.90, or from about 0.90 to about 0.95.

By slip dislocation to the interface, the surface layer is at least about 85% relaxed, at least about 90% relaxed, at least about 95% relaxed, or even fully relaxed, i.e., 100% relaxed. The surface layer is substantially free of threading dislocations or has a dislocation density of less than about 10 ⁴ Threading dislocation/cm ² Threading dislocations at a concentration of (a).

In an embodiment, the surface layer is not continuous but includes discontinuous segments (i.e., islands) disposed on the surface of the substrate that become relaxed by creating and sliding dislocations from the dislocation source layer to interfaces with the islands to create misfit interface dislocations between each island and the substrate. Dislocations reaching the surface of the substrate between the islands dissipate at such surfaces, which allow the regions between the islands to be substantially free of dislocations when compression is complete. After relaxation of the islands, the semiconductor material is further deposited to create a surface layer that extends continuously across the entire diameter of the substrate. In such embodiments, dislocations below the islands propagate laterally at the interface between the newly deposited material and the substrate, thereby relaxing the newly deposited material and the continuous surface layer as a whole.

The relaxed heterostructure fabricated by any of the methods described above is used to fabricate integrated circuits using wafer bonding and layer transfer methods for silicon-on-insulator structures, or for subsequent fabrication of strained silicon-on-insulator structures.

An additional layer is deposited on the relaxed surface layer, forming a heteroepitaxial structure with a strained layer on top of the relaxed layer on top of the substrate. Such structures are also used to transfer the relaxed layer and the strained layer to another substrate, forming a heteroepitaxial structure with a buried strained layer or alternatively a buried strained layer on an insulator. That is, the heteroepitaxial structure has a relaxed layer of semiconductor material on top of the strained layer of semiconductor material on top of the insulating layer on the substrate or on top of the substrate.

Furthermore, the structure manufactured by the method of the present disclosure is used for manufacturing a semiconductor device, such as a Field Effect Transistor (FET) or a modulation doped field effect transistor (MODFET) layer structure.

V. device for compression

In this regard, any of the apparatus described below is used to perform the methods described herein that involve compression of a substrate.

Referring now to fig. 3-10, compression of a substrate is achieved through the use of a substrate holder that includes a chamber and/or fluid channel for applying a pressure differential across the substrate.

Referring now to fig. 3-4, compression of the structure 9 is achieved through the use of a structure clamp 11. The structural clip 11 includes a top plate 13. As shown in fig. 3-4, the top plate 13 is a ring. The top plate 13 may have other shapes and may extend completely across the substrate 9 without limitation. The top plate 13 is adapted to contact the front surface of the substrate 9 at the peripheral edge 5 of the structure.

The structure holder 11 comprises a back plate 15 for contacting the back surface of the structure 9 adjacent the circumferential edge 5. The back plate 15 includes a peripheral ring 20, which peripheral ring 20 extends upwardly toward the top plate 13. However, in other embodiments, the peripheral ring 20 may be part of the top plate 13 or may be separate from the top plate 13 and the back plate 15. The back plate 15, top plate 13 and peripheral ring 20 are each adapted to form a peripheral cavity 18 between the top plate 13, back plate 15 (including the peripheral ring) and the peripheral edge 5 of the structure 9. In general, the back plate 15 and top plate 13 form a seal with the structure 9 that allows the pressure in the peripheral cavity 18 to be increased relative to the pressure outside the clamp 11 as described below. The peripheral cavity 18 is formed by positioning the semiconductor structure 9 on the back plate 15 and lowering the top plate 13 onto the back plate 15 until a seal is formed between the top plate 13, the back plate 15 and the peripheral edge 5 of the structure 9.

The holder 11 comprises an outlet 22 in the back plate 15 for adjusting the pressure in the peripheral cavity 18. Alternatively, the outlet extends through the front plate 13 and/or the peripheral ring 20. The outlet 22 is in fluid communication with a pump (not shown) for increasing the pressure in the peripheral chamber 18.

Referring now to fig. 8, the grip 11 is part of a device 36 for compressing the structure 9. The device 36 also comprises a housing 35, the housing 35 defining a main cavity 27, in which main cavity 27 the holder 11 is arranged. The device 36 includes an outlet 32, the outlet 32 being in fluid communication with a pump (not shown) for regulating the pressure P in the main chamber ₁ . The outlet 22 in the structural holder 11 extends through the shell 35. In this way, the pressure P is maintained in the main chamber 27 ₁ While maintaining a different pressure P in the peripheral cavity 18 of the structural clamp 11 ₂ . By pressure P in main chamber 27 ₁ Maintained at less than the pressure P in the peripheral chamber 18 ₂ The structure 9 is compressed (i.e., the relaxation radius of the substrate is reduced).

In this respect, the pressure P in FIGS. 3-10 ₁ And/or P ₂ The associated arrows are provided for exemplary purposes and should not be considered limiting the device to a particular pressure profile (i.e., use of vacuum or pressure in the peripheral or main chamber).

During compression of structure 9, P ₁ Ratio P ₂ At least about 10MPa less, or as in other embodiments, P ₂ At least about 20MPa, at least about 50MPa, or at least about 75MPa (e.g., from about 10MPa to about 100MPa, from about 10MPa to about 50MPa, or from about 10MPa to about 25 MPa). In some embodiments, P ₁ Is ambient pressure. In such an embodiment, the main chamber 27 and the shell 35 are eliminated and the shell is exposed to the ambient environment (i.e., atmospheric pressure).

The heating element 30 is used to heat the structure 9 during compression to activate the dislocation source layer. As described above, the structure is heated to a temperature of from about 650 ℃ to about 1000 ℃ or from about 700 ℃ to about 1000 ℃.

Another embodiment of a structural grip 111 is shown in fig. 5. It should be noted that the holder components shown in fig. 5 are similar to those in fig. 3, designated by the corresponding reference numerals of fig. 3 plus "100" (e.g., component 15 becomes component 115). As shown in fig. 5, the top plate 113 comprises a protrusion 117 for contacting the front surface 8 of the structure 9. The protrusion 117 forms a seal with the structure 9 to allow for an increase in pressure in the peripheral cavity 118.

In some embodiments and as shown in fig. 6-7, the structure 9 has a coating 39 (fig. 6) or a coating 40 (fig. 7) on at least a portion of the structure surface. As shown in fig. 6, the coating 39 extends across the circumferential edge 5 of the structure 9 and the front and back surfaces 8, 4 and the portion adjacent the peripheral edge 5. As shown in fig. 7, the coating 40 also extends across the entire back surface 4 of the structure. Alternatively or additionally, the coating extends across one or more surfaces of the structural grip. Coating 39 or coating 40 (or coating extending beyond the structural grip) is made of a material such as graphite, hexagonal boron nitride, MS ₂ 、WS ₂ SiCN、AlCr(V)N、TiAl(Y)N、CaF ₂ 、BaF ₂ 、SrF ₂ Or BaCrO ₄ Is composed of a low friction material. In some embodiments, the structure 9 has a coating on the front surface of the structure that reduces or even prevents evaporation of the volatile film components of the structure. Suitable coatings for reducing evaporation include amorphous silicon.

Another embodiment of a structural grip 211 is shown in fig. 9. It should be noted that the holder components shown in fig. 9 are similar to those in fig. 3, designated by the corresponding reference numerals of fig. 3 plus "200" (e.g., component 15 becomes component 215). The top plate 213 of the structure clamp 211 comprises indentations adapted to form a central cavity 240 between the top plate 213 and the front surface 8 of the structure 9 during use of the structure clamp 211. A central cavity 240 is formed by lowering the top plate 213 onto the semiconductor structure 9. The indentations are defined by annular walls 242. The indentations have a radius less than the strain radius of the structure. As used herein, "strain radius" refers to the radius of the structure prior to radial compression (deformation) of the structure 9 using the structure clamp 211.

The top plate 213 includes an outlet 246, the outlet 246 being in fluid communication with a pump (not shown) to maintain the pressure P in the central chamber 240 ₁ . In this way, a pressure differential is maintained between the central chamber 240 and the peripheral chamber 218 such that the structure 9 is radially compressed. By the pressure P to be in the central chamber 240 ₁ Maintained at less than the pressure P in the peripheral cavity 218 ₂ The structure 9 is compressed. Pressure P ₁ And/or P ₂ Is within the scope described above.

In some embodiments, the structure 9 is radially compressed until the radius of the compressed structure is substantially the same as (or slightly less than) the radius of the recess in the top plate 213. When the structure 9 is compressed to the radius of the recess, the central cavity 240 is in fluid communication with the peripheral cavity 218, which allows pressure equalization between the cavities, thereby limiting the compression of the structure 9. Accordingly, as the top plate 213 is recessed, the structural grip 211 is self-limiting, limiting radial compression of the structure 9.

As shown in fig. 10, the structure holder for the radially compressed structure is adapted to compress multiple structures simultaneously. The clip members shown in fig. 10 are similar to those in fig. 3, designated by the corresponding reference numerals of fig. 3 plus "700". The grip 711 comprises a back plate 715, which back plate 715 is adapted to contact the structures 9a, 9b, 9c, 9d adjacent to the peripheral edges of the structures. The grip 711 includes a top plate 713, which top plate 713 contacts the structures 9a, 9b, 9c, 9d adjacent to the peripheral edges of the structures. In the circumferential direction of the back plate 715, top plate 713 and structures 9a, 9b, 9c, 9dA peripheral cavity 718 is formed between the edges. The top plate 713 contains a cavity 760 extending to the front surface of the structures 9a, 9b, 9c, 9d to allow the structures to be exposed to pressure P in a main cavity (not shown) ₁ . Maintaining the peripheral cavity 718 at a pressure P ₂ 。

By combining P ₁ Maintained at less than P ₂ Radial compression structures 9a, 9b, 9c, 9d. At P ₁ And P ₂ The difference between them is at least about 10MPa and is within any of the ranges described above. In such an embodiment, P ₁ Is at atmospheric pressure and the top plate 713 is a continuous section that does not contain the divided chamber 760. While the substrate holder shown in fig. 10 is described and shown as having only one back plate and one top plate, it should be understood that the holder has multiple separate back or top plates sealing a single structure or combination of structures. Further, while the structure grip 711 shown in fig. 10 is capable of radially compressing four structures, it should be noted that the grip is configured to simultaneously compress more or less structures without limitation.

In addition to the devices described above, devices that grip the structure (such as through the use of clamps or other gripping elements at the peripheral edges) and allow the structure to be compressed are used to relax heterostructures in the device such as described below. Referring now to fig. 11-19, compression of the structure is achieved through the use of a structural grip that is radially movable relative to the structure. In such embodiments, the structural grip is part of the device for compressing the structure. Such an arrangement is similar to the arrangement 36 shown in fig. 8, and the arrangement includes a housing 35, the housing 35 defining a main cavity 27 in which the clip 11 is disposed. The device includes a heating element 30 to heat the structure 9 to activate the dislocation source layer during compression by any of the uses of the structures of figures 11-19.

Referring now to fig. 11, a structural clamp 1720 includes a plurality of triangular shaped segments 1785 that are directed inwardly toward a central axis a of the clamp. Each segment has at least one fluid passage 1787 formed therein to pull a vacuum on the substrate. Section 1785 is configured to move inwardly from center axis a such that the substrate is compressed.

Referring now to fig. 12, device 911 is a fixture that includes a front plate 931 and a back plate 932 that apply a clamping force to substrate 9. As shown in fig. 12, the top plate 931 and the back plate 932 are rings. The top plate 931 may have other shapes and may extend entirely across the substrate 9 without limitation. The front plate 931 and back plate 932 are radially movable inward from the center of the device by any mechanical means including the use of pneumatic, hydraulic, motor, and the like.

Referring now to fig. 13, in another embodiment, the structure clamp 1011 includes a generally planar back plate 1146 including an annular boss 1147, the annular boss 1147 being modified and shaped to be received in a slot 1148 in the back of the structure 9. The boss 1147 is movable so that it compresses the structure 9.

In some embodiments and as shown in fig. 14, the structural clip 1220 also includes a front plate 1250 with an annular ring 1252 extending from the front plate 1250. During heating, ring 1252 applies a downward force to structure 9 to prevent the structure from loosening from boss 1247 during compression of the structure. Other structures for accomplishing this function are contemplated within the scope of the present disclosure.

In other embodiments and as shown in fig. 15, the structural clamp 1320 includes a back plate 1346 and a boss 1347 similar or identical to those shown in fig. 13 and 14. Substrate holder 1320 also includes a front plate 1351 and a front boss 1355 that are modified and shaped to be received in a slot 1357 in the front surface of structure 9.

Referring to fig. 16-18, the structural clip 1420 of this embodiment includes a flat back plate 1461 for supporting the structure 9 and generally uses circular opening cyclic presses 1460 for receiving and compressing the structure. As in fig. 16-18, the flat plate may extend only partially toward the center of the structure, or may extend continuously under the structure 9. The pusher 1460 may continuously encircle the structure, or as shown in fig. 18, may include a plurality of arcuate segments 1563 that form openings for receiving the structure 9. Press 1460 and/or segment 1563 are movable inwardly relative to structure 9 to compress the structure.

Referring to fig. 19, the structural clip 1620 includes a generally planar back plate 1681 and a flange 1683. The structure 9 includes a loop 1680 attached to the back surface of the structure adjacent the peripheral edge of the structure. The flange 1683 is adapted to engage the ring 1680. The support 1681 and flange 1683 are movable relative to the structure to compress the structure.

In some embodiments, the stress applied by the device described above is cyclic, such as by reducing the pressure differential across the structure (e.g., by reducing or increasing the pressure in the periphery or the main cavity) or by reducing the stress applied in the device gripping the substrate. Such cycling relieves any elastic stress that develops in the structure.

As used herein, the terms "about," "substantially," and "approximately" when used in connection with a range of sizes, concentrations, temperatures, or other physical or chemical properties or characteristics are intended to cover variations that occur at the upper and/or lower limits of the range of properties or characteristics, including, for example, variations resulting from rounding, measurement methodologies, or other statistical variations.

When introducing elements of the present disclosure or the preferred embodiments thereof, the articles "a," "an," "the," and "said" are intended to mean that there are one or more of the elements. The terms "comprising," "including," and "having" are intended to be inclusive and mean that there may be additional elements other than the listed elements.

As various changes could be made in the above apparatus and methods without departing from the scope of the disclosure, it is intended that all matter contained in the above description and shown in the accompanying drawings shall be interpreted as illustrative and not in a limiting sense.

Claims (16)

1. An apparatus for applying radial stress to a generally circular semiconductor substrate, the substrate having a central axis, front and back surfaces generally perpendicular to the central axis, a peripheral edge extending from the front surface to the back surface, and a circumferential groove in the back surface adjacent the peripheral edge, the apparatus comprising:

a structural grip comprising:

a generally planar back plate having an annular boss adapted to be received in the groove in the back surface of the substrate, the boss being movable to exert a stress on the substrate.

2. The apparatus of claim 1, in combination with the substrate, the substrate comprising a slot in the back surface of the substrate, the boss being received in the slot.

3. The device of claim 1, further comprising a front plate that exerts a downward force on the structure to prevent the structure from backing out of the boss.

4. The apparatus of claim 1, wherein the boss is a first boss, the apparatus further comprising a front plate having a second boss adapted to be received in a slot in the front surface of the substrate.

5. The apparatus of claim 1, further comprising a heating element for heating the structure.

6. An apparatus for applying radial stress to a generally circular semiconductor substrate, the substrate having a central axis, front and back surfaces generally perpendicular to the central axis, a peripheral edge extending from the front surface to the back surface, and a ring bonded to the back surface adjacent the peripheral edge, the apparatus comprising:

a structural grip comprising:

a generally planar back support having a flange adapted to engage the ring on the back surface of the substrate, the support being movable to exert a stress on the substrate.

7. The apparatus of claim 6, in combination with a substrate comprising a ring bonded to the back surface of the substrate.

8. The apparatus of claim 6, further comprising a heating element for heating the structure.

9. An apparatus for applying radial stress to a semiconductor substrate, the substrate having a central axis, front and back surfaces generally perpendicular to the central axis, and a peripheral edge extending from the front surface to the back surface, the apparatus comprising:

A structural clip mounted in the cavity, comprising:

a generally planar back plate; and

a presser for receiving and compressing the substrate, the presser being adapted to compress the substrate radially inward toward a central axis of the substrate at a peripheral edge of the substrate.

10. The device of claim 9, wherein the depressor comprises a plurality of arcuate segments forming an opening for receiving the structure.

11. The device of claim 9, wherein the pushers continuously encircle the structure.

12. The device of claim 9, wherein the back plate extends only partially toward the central axis of the substrate.

13. The device of claim 9, wherein the back plate extends continuously under the structure.

14. The apparatus of claim 9, further comprising a heating element for heating the structure.

15. An apparatus for applying radial stress to a semiconductor structure having a front surface, a back surface, a circumferential edge, and a central axis, the apparatus comprising:

a triangular-shaped segment directed inwardly toward a central axis, the segment configured to move from the central axis to apply stress to the structure; and

A fluid channel is formed in each segment for creating a vacuum between the segment and the structure.

16. The apparatus of claim 15, further comprising a heating element for heating the structure.