Layer transfer of film utilizing controlled propagation

Originally, humans rely on the "sun" to obtain almost all useful energy. Such energy comes from oil, light, wood and various forms of thermal energy. As just one example, humans rely heavily on petroleum resources, including coal and gas, for most of their demand. Unfortunately, such oil resources have plummeted, causing other problems. As an alternative, in part, solar energy has been proposed to reduce reliance on petroleum resources. As just one example, solar energy can be obtained from "solar cells", which are commonly made from silicon.

Silicon solar cells generate electricity when exposed to sunlight. Solar radiation interacts with the atoms of silicon, moving to the p and n-doped regions of silicon and forming electrons and holes that generate potential differences and currents between the doped regions. A condensing element is coupled to the solar cell to improve efficiency. As an example, sunlight is collected and focused using a condensing element that directs sunlight to one or more parts of the activated photovoltaic material. Although efficient, their solar cells still have many limitations.

As just one example, solar cells rely on starting materials such as silicon. Such silicon is usually made using polysilicon and / or single crystal silicon materials. Polysilicon is further referred to as polycrystalline, microcrystalline or nanocrystalline, depending on the size and crystallinity of a single crystalline particle. For single crystal (single crystal) materials that do not have many irregular crystal orientations or many grain boundaries, these materials are hereinafter referred to as "polysilicon". Amorphous silicon is not in the form of silicon commonly used in wafer-like solar cells because it is thinner than a few microns and has a short carrier life.

Materials for solar cells are usually difficult to manufacture. Polysilicon batteries are usually formed by manufacturing polysilicon plates. Although those plates may be cost-effectively formed using a crystal furnace, they do not have optimal properties suitable for high-grade solar cells. In particular, polysilicon plates do not provide the highest possible efficiency for capturing solar energy and converting the captured solar energy into useful power.

In contrast, single crystal silicon (c-Si) has properties suitable for high quality solar cells. However, such single crystal silicon is expensive to manufacture and difficult to use in sunlight-utilizing applications in an efficient and cost-effective manner.

In addition, polysilicon and single crystal silicon materials suffer material loss during the conventional manufacture of single crystal silicon substrates. A slicing process is usually used to physically desorb the thin single crystal silicon layer from the initially grown single crystal silicon ingot. For example, an inner peripheral blade (ID) slicing process or wire slicing process removes 40% or even 60% of the starting material from the casting or growth boules and slices the material into wafer shape elements. This is a highly inefficient method for processing thin polysilicon or single crystal silicon plates for the use of solar cells.

<p> Thin-film solar cells have been proposed to overcome the drawbacks of using silicon materials. Thin-film solar cells are often less costly due to the use of less silicon material or alternative materials, but their amorphous or polycrystalline structures are less efficient than the more expensive bulk silicon batteries made from single crystal silicon substrates. ..</p><p> From the above, it can be seen that there is a strong demand for a technique for producing a high-quality single crystal silicon thin plate suitable for low cost and high productivity.</p>

<p> The material film can be formed by providing a semiconductor substrate having a surface region and a cleavage region located below the surface region at a predetermined depth. During the process of cleavage of the membrane from the substrate, the shear in the cleavage region is carefully controlled. According to one embodiment, the in-plane shear (KII) configuration is maintained at approximately zero at a given film thickness to maintain the propagating cleavage plane within the desired depth. In one embodiment, a near zero KII configuration is achieved by adiabatic heating of silicon through exposure to electron beam irradiation that can transfer thermoelectromotive forces and moments to achieve the desired KII conditions. In another embodiment, the KII configuration is deliberately maintained at a high level by any external energy injected to guide fracture propagation using energy control methods. In such an embodiment, exposure to irradiation such as a laser or electron beam that provides a temperature gradient to achieve controlled propagation at a precisely set depth within silicon with depth-dependent cleavage energy control. A high KII configuration is achieved by heating the silicon through.</p><p> Embodiments of the present invention generally relate to layer transfer techniques for forming thick films. More specifically, the present invention provides methods and devices for low-thickness or non-injection layer transfers for solar cells. As just one example, it is applied to cleave a thick film along the crystal orientation plane of a single crystal silicon substrate. However, it will be recognized that the present invention has a wider range of applicability.</p><p> Many advantages can be obtained by practicing the present invention. In a preferred embodiment, a single crystal silicon ingot substrate having a surface in the crystal orientation of {111} or {110} can be selected. Therefore, the ion implantation process for forming the cleavage region can be omitted or limited to forming the cleavage initiation region in a part of the outer peripheral region. This substantially simplifies the layer transfer process, reducing system energy costs and increasing productivity.</p><p> In certain embodiments, the process is (i) a initiation process that uses a higher injection donor amount (starter donor amount) to initiate fracture in a relatively small region (starter region) of the silicon ingot substrate, and (ii). It is divided into a propagation process that expands the tips of cracks that occur in the rest of the silicon ingot substrate to release the thick film (propagation region). Since the propagation process must extend the starting region to most of the surface area of the thick film to be released, this process should be reliably processed with little or no donation. This is called the propagation donation, and therefore the total donation required for the process will be the weighted area average of the initiation and propagation donations. For example, the starting area is 1% of the total area and 6x10¹⁶cm^-2Hydrogen is used, while the propagation region is 1 × 10¹⁶cm^-2If hydrogen is used, the total effective amount is 0.01 * 6 × 10¹⁶cm^-2+0.99*1×10¹⁶cm^-2That is, 1.05 × 10¹⁶cm^-2Is. Reducing or omitting the propagation donation will have a primary effect on the total donation requirement for this cleavage process.</p><p> Reducing the donation is also desired (i) optimizing the effect of the injected hydrogen to maximize its cleavage introduction operation, and (ii) without breaking or breaking the membrane. It relies on optimizing the cleavage process by generating appropriate stress strength conditions at the crack tips to keep the cleavage front in depth. The embodiments according to the present invention are aimed at optimizing the mechanical configuration during the cleavage propagation process.</p><p> In certain embodiments, the cleavage initiation region can be formed by patterned injection by irradiating high energy ion particles at a controlled dose produced using a linear accelerator. In certain embodiments, the present invention is to generate a stress gradient (appropriate) so that a substantially zero shear region can be formed despite the natural tendency of the mechanical configuration to generate a mixed mode load. It provides a mechanical moment load with a load of force (generated through a heat treatment process and / or mechanical force).</p><p> To generate the desired conditions for controlling propagation, another embodiment uses pure heat loading (which may be heating alone or in combination with differential cooling).</p><p> Using embodiments of the invention, high quality thick film crystalline materials with thickness ranges from tens of microns to hundreds of microns are produced with substantially lower grooving loss than available prior art. sell. The resulting single crystal silicon thick film is particularly suitable for high efficiency (20% or higher) photovoltaic batteries. Some embodiments can use existing manufacturing processing systems and techniques and enjoy the benefits of new and improved techniques for manufacturing thin wafers / substrates for various semiconductor device applications. Further details regarding the various embodiments of the present invention can be found in the detailed description below.</p>

<figref num="1">It is a schematic diagram which shows the formation of the cleavage start region by patterned injection of high energy particles which concerns on embodiment of this invention.</figref><figref num="2">It is a schematic diagram of a general case of propagation.</figref><figref num="3">It is a figure which shows the hydrogen layer stress distribution model.</figref><figref num="3A" /><figref num="4">It is a figure which shows the simplified vertical-horizontal arrangement of an effective heat load composition.</figref><figref num="5">It is a graph which shows the type of cleavage obtained according to the time coefficient of heat load application.</figref><figref num="6A">It is a figure which shows the concept of heat immersion for raising an absolute temperature in a cleavage plane region.</figref><figref num="6B">It is a figure which shows the concept of a thermal pulse for generating a film substrate temperature difference.</figref><figref num="7">It is a figure which shows the plot (CSDA range) of the maximum electron depth with respect to energy.</figref><figref num="8">It is a figure which shows the Monte Carlo simulation of the volumetric energy deposition distribution of an electron beam in silicon.</figref><figref num="9">It is a simulated electron beam distribution map for realizing KII controlled propagation cleavage.</figref><figref num="10">It is a figure which shows the stress strength and energy result using the adiabatic electron beam KII controlled propagation.</figref><figref num="11" />

Individual embodiments of the present invention provide methods and devices for thick film layer transfer for solar cells. As just one example, it is applied to cleave a thick film along the crystal orientation plane of a single crystal silicon substrate. However, it will be recognized that the present invention has a wider range of applications. For example, other materials such as germanium, gallium arsenide (GaAs), gallium nitride (GaN) or silicon carbide (SiC) are used in the cleavage process to release material films for solar energy, photoelectron or semiconductor applications. Can be done.

As described in the background technology, the development of silicon-based solar cells depends on reducing bottlenecks for the cost of grooving loss on wafers. Adopting traditional slicing to thicken the film suitable for solar cells, or recently reported wafer techniques (eg, multi-wire slicing, spark cutting, laser cutting or plasma cutting), has one or more challenges: high cutting. It may exhibit limited usefulness due to groove loss, slow cutting speed and lack of productivity.

The solution is to use a high energy ion beam to form a cleavage region under the substrate surface at the desired thickness and then perform a layer transfer process to release the thick film from the residual substrate. Is. However, the use of injection ions alone to form cleavage-prone cleavage regions may require high ion donations and extensive injection regions. Moreover, such dependence on injected ions can result in higher surface roughness, higher ion donations and additional costs due to non-productivity, and potentially lower yields and film quality. Depending on the embodiment, these and other constraints are overcome using the methods and structures according to the invention.

According to the individual embodiments of the present invention, cleavage of the material membrane is achieved using a cleavage process that carefully controls the shear conditions during the cleavage, with substantially reduced or no injection of ions. To. In certain embodiments, the material film is formed by providing a semiconductor substrate having a surface region, an outer peripheral region and a cleavage region at a predetermined depth below the surface region. For the purposes of this patent application, the term "open region" does not necessarily mean a region that has received radiant energy or injected ions, but rather separates from the substrate following the use of radiant energy and / or injected ions. Means the area that will be.

The cleavage start region may be defined in the outer peripheral region portion and the vicinity of the cleavage region. The cleavage initiation region may be formed by exposing the region to thermal, chemical, electrical and / or mechanical processes to crush or release the cross section of the membrane within the initiation region.

In certain embodiments, the initiation of cleavage exposes the cleavage initiation region to a local heat treatment so that the cleavage front is initiated within this region and propagates to the perimeter of the low donor initiation region and does not facilitate further propagation. Is realized by. The propagation region, on the other hand, is controlled by the local energy itself. The general membrane release process can then continue to propagate the initiated membrane from the existing cleavage front through the rest of the substrate.

FIG. 1 is a schematic view showing a side surface of a semiconductor substrate having a cleavage region at a predetermined depth below a surface region according to an embodiment of the present invention. This figure is merely an example, and the scope of this claim must not be unreasonably limited. Those skilled in the art will understand many variants, modifications, and alternatives.

As shown in FIG. 1, the semiconductor substrate 1700 is prepared to have a surface region 1702 and an outer peripheral region 1704. In addition, a cleavage region 1706 is provided. This cleavage region is essentially a virtual surface or layer located at a predetermined depth d below the surface region 1702 and is used to determine the thickness of the thick film 1720 separated from the semiconductor substrate 1700. In certain embodiments, the semiconductor substrate has a thickness approximately greater than the depth h. In certain embodiments, the semiconductor substrate is a single crystal silicon material for use in photovoltaic solar cells. In certain embodiments, the silicon substrate is generally selected to approximate the crystal orientation plane of {111} or {110} (there may be any small miscut angle less than 1 degree). Has a surface. In certain embodiments, the defined cleavage region is approximately parallel to the surface region. The cleavage action is energetically easier along the {110} plane next to the {111} plane than in the conventional {100} plane, so the surface that is cleaved by the lower surface energy crystal orientation cleavage plane. It would be desirable to orient the materials so that they are cleaved in line with. A more detailed description of the technique for selecting any direction of silicon boules for slicing or cleavage is incorporated herein by reference in its entirety, USA, filed May 7, 2008 by Francois J. Henley. It can be found in provisional application No. 61 / 051,344 METHOD AND DEVICE FOR SLICING A SHAPED SILICON INGOT USING LAYER TRANSFER (agent case number: 018419-025600 US).

Further, FIG. 1 shows the formation of a cleavage start region by injecting a pattern of high-energy particles according to the embodiment of the present invention. This figure is merely an example, and the scope of this claim must not be unreasonably limited. Those skilled in the art will understand many variants, modifications, and alternatives.

As shown in FIG. 1, a portion of the outer peripheral region 1704 may be associated with a predetermined pattern region (not directly shown in the cross section of FIG. 1) within the surface region 1702. In certain embodiments, the selected portion of the outer peripheral region is within the margin of the cleavage region 1706.

The pattern region of the surface region 1702 is then exposed to a high energy ion beam 1740, for example using a high energy particle accelerator, with H + ions having an energy level of 1 MeV or higher. The accelerator selection can range to linear accelerators (electrostatic or RF driven) or non-linear accelerators such as cyclotron accelerators. Since pattern injection is desirable, scanning a particle beam with a limited beam size is preferred. The scan may be electromagnetic, electrostatic, or mechanical, depending on thermal, mechanical, cost, and other considerations. In one embodiment, the area of the pattern start area is 1 to 3% of the total substrate surface area (eg, 2 to 5 cm for a substrate of size 125 mm x 125 mm).²Being limited to (below), the amount of ion particles donated well controls the minimum energy cost of the system and increases the productivity of the thick film cleavage process.

High-energy ions are injected below the surface region to reach a region within the vicinity of cleavage region 1706. The penetration depth of ions depends on the energy level and can be controlled to a desired value that determines the depth h of the cleavage region. The injected ions slow down in the crystal lattice by transferring kinetic energy to the lattice in the form of ionization (electron braking) and a small amount of atomic damage (nuclear arrest) due to the movement of atoms.

During the final phase (within about 2-5% of the total range), the ions interact more fully with the crystal lattice in the nuclear arrest state, forming relatively stressful and damaged thin regions of lattice coupling. Then, the cleavage start area 1708 is defined. As shown, the formed cleavage start region 1708 is a small planar region extending inward of the cleavage region 1706 from a portion of the outer peripheral region 1704. Since the pattern injection is performed by irradiating the ion particles within an area of less than 1 to 3% of the total area of the surface area, this starting donation can be higher than the propagation donation. This allows the average area donation to be kept low due to high productivity. Of course, there can be many variants, modifications, and alternatives.

Control propagation method

One embodiment of the present invention attempts to modify the cleavage structure in order to reduce, eliminate or control the shear mode II stress strength factor (KII) at the tip of the propagation cleavage fissure. The process is generally referred to as controlled propagation when the cleavage action proceeds reliably along the desired depth below the surface 1702.

Two forms of controlled propagation can be utilized to reliably propagate cleavage in material 1700 in order to release a film of thickness h from the substrate and avoid cleavage failure. The first form of controlled propagation is called "KII controlled propagation", while the second form of controlled propagation is called "energy controlled propagation". Both forms can control the cleavage plane propagating at the desired depth to release the h-thick film. However, those forms of controlled propagation achieve this goal using different methods and techniques.

A key consideration in designing a cleavage technique is the avoidance of unwanted cleavage failures such as cracking or breaking of the membrane. Cleavage failure is basically defined as an unfavorable branch of the cleavage surface, which usually results in a cracked membrane. Thus, avoiding unwanted bifurcation of the cleavage plane is a consideration in the cleavage technique. Other factors known to affect unfavorable branching include, for example, crystal orientation, injection donation, depth and temperature distribution.

Other effects have proven to be important. For example, for single crystal silicon, the thermal energy injected into the vicinity of the cleavage plane and used to generate cleavage stress is the fact that the injected cleavage plane has considerably lower thermal conductivity over uninjected single crystal silicon. Will be affected by. This low thermal conductivity is the result of hydrogen injection damage and tends to alter the temperature distribution and will also alter the cleavage stress.

KII control propagation

The first form of controlled propagation is KII controlled propagation, defined as the threshold cleavage energy G'(defined below) with 0 or almost 0 KII realized at the same time. Based on the basic principle of linear elastic fracture mechanics (LEFM), which redirects the propagation cleavage front to achieve and maintain KII = 0, this condition allows the propagation cleavage plane to stay at the desired depth. To do. Subtle changes in depth result in stable KII control in restoring and generating KII shear levels and signs to direct the cleavage plane in the desired depth direction.

While the condition KII = 0 is usually considered to be the ideal condition for stable cleavage propagation, within that range there is a specific range of KII such that cleavage follows a crystallographic cleavage plane. .. Therefore, when considering the maintenance of KII within a limited range of around 0, it is preferable that the cleavage action continues to follow the cleavage plane.

Energy controlled propagation

The second form of controlled propagation is energy controlled propagation, in order to propagate the cleavage plane below the desired depth so that the propagation is facilitated by a high KII coefficient but not energetically supported. It is characterized as depth control, where the required threshold energy changes. Cleavage with an injected hydrogen layer is an example of energy controlled cleavage.

In the case of pure cleavage due to the hydrogen cleavage plane, the displacement of the cleavage plane toward the surface reduces the energy effective for propagation, so if the cleavage progresses away from the injection depth, propagation will stop. As a result, the strong tendency to propagate in the surface direction is canceled by the strong energy-depth dependence.

Derivation of control propagation equations

The discussion below deals with the general linear elastic fracture mechanics (LEFM) equations governing cleavage according to embodiments of the present invention. This analysis assumes large area tiles where the membrane is thinner than the rest of the silicon tiles.

The mechanical configuration that follows the initiation process is shown in Figure 2. In particular, the thick film of thickness h is partially released from the rest of the silicon ingot substrate having thickness H. Due to the larger dimensions of the substrate with respect to the cleavage material, h H and the tile moment and force M₂, P₂, M₃, P₃Is understood to be very small. Thus, the source of stress strength found at the tip of the crack is the moment and force M bound to the partially released thick film.₁And P₁Is dominated by.

The basic concept of the embodiment of the present invention is a hydrogen cleavage plane. The hydrogen cleavage plane (H-plane) affects the cleavage action due to stress and layer damping effects. In particular, the H-plane can be used as a starting layer at higher donations and as both an inducing or propagating layer at lower donations. The action of the H-layer differs within the morphology of these two donors.

One effect on the cleavage action of the hydrogen injection layer at the end of the range (EOR) is the reduction of fracture strength around the cleavage plane. Such a reduction in fracture strength can occur due to stress induced by bond damage and by the presence of hydrogen itself. Both of these factors can reduce the energy required to form a new surface during cleavage.

The energy required to form a new surface during cleavage is hereinafter referred to as surface energy (γ). Fracture toughness in silicon is sometimes 4-6 J / m²In uninjected single crystal silicon, the surface energy is approximately 1.2 for the {111} orientation, although it includes a lattice trap-like effect to provide effective energy to create a new surface. J / m²Is. In the analysis shown below for silicon in {111} orientation, 1.2 J / m per surface²Surface energy (total 2.4J / m)²) Is assumed.

In contrast, the modified surface energy (γ') value along the injected cleavage plane can be substantially reduced, perhaps by a factor of 5 or more. The value γ'of the surface energy of the effective cleavage plane relates to the uninjected surface energy (γ) according to the following relational expression. (1) γ'= α_H²* γ Here, α_HIs a coefficient between 0 and 1 that works to quantify the decay of cleavage energy due to hydrogen embrittlement. α_HThe term deserves all the effects necessary to provide an accurate representation of all stress and bond damage effects. α_HIs experimentally calculated as a function of cleavage energy, donation amount, injection heating conditions and post-injection heating conditions.

Since cleavage energy is a fundamental parameter in linear elastic fracture mechanics, the correlation between injection donation and energy allows accurate prediction and modeling of cleavage action.

Two surfaces are formed during the cleavage process. Therefore, the cleavage energy release rate (G') is related to the surface energy as follows. (2) G'= 2 * γ'= 2 * α_H²* γ

Embrittlement coefficient (α_H) Depending, G'is about 2.4J / m in the cleavage plane²Can be changed considerably less from. For example, in an experiment with a 50 μm thick single crystal silicon film, 2-8 × 10¹⁶cm^-2The cracks on the cleavage surface formed by the amount of H injection donated are about 4 to 6 × 10.¹⁶cm^-2A measurable reduction in cleavage energy measured by the dual cantilever beam mechanical configuration for higher H donations (α)_H²Indicates that there is &lt;1).

(4×10¹⁶cm^-2For cleavage planes with lower (less) hydrogen donations, α_H²~ 1 is calculated experimentally. Thus, cleavage inducing a lower donation by cleavage energy, which basically has a intrinsic cleavage energy value, is an injected hydrogen compression that holds the propagation in front of the cleavage along the cleavage plane by the principle of energy control. It is mainly dominated by the stress distribution.

A series of closed-form equations have been developed by studying the behavior of cleavage. The development of these equations assumes that a starting crack exists at the depth (h) of the substrate, thereby determining the thickness of the cleavage membrane, resulting from the cleavage initiation. In many examples h = 50 μm, but any film thickness is acceptable in the equation. The model assumes separate membrane lengths (c or sometimes referred to as L) of a few millimeters to a few centimeters. The shape is two-dimensional, meaning that the width w does not change so that cleavage occurs along the line.

Cleavage force mode (KI) and in-plane shear force mode (KII) are important parameters for modeling the cleavage process. If known fracture energies are used for the cleavage configuration, it is expected that the resulting conditions leading to crack expansion and propagation will be compatible with silicon or other materials. For example, G'is 2 * γ', about 2.4J / m for uninjected layer², Is known to occur in single crystal silicon under fracture conditions.

The threshold for fracture propagation is defined as follows.<maths num="1"><img file="JP2010103488A_D0001.tif" /></maths>(3) Where E'is E'= E / (1-v²), Which in the present specification is G = G'and generally relates to the cleavage plane in equation (2). (2) G'= 2 * γ'= 2 * α_H²* γ

FIG. 2 shows a general example of propagation problems, assuming an isotropic, uniform, and linear elastic material. Existing cracks are evenly subjected to axial forces and moments along the three ends. Solving the general solution for the stress strength coefficient using equation (2) yields the following solution for the energy release rate at the crack tip.<maths num="2"><img file="JP2010103488A_D0002.tif" /></maths>(Four) Where P_nAnd M_nIs the respective force and moment acting in each term.

The coefficients of KI and KII are expressed by the following equations.<maths num="3"><img file="JP2010103488A_D0003.tif" /></maths>(5, 6) Here, the parameters are derived from the forces, moments and shapes detailed in "Mixed Mode Cracking in Layered Materials", Advances in Applied Mechanics, Vol. 29 (1992) by Hutchinson and Suo, for all purposes. The whole is used here as a reference.

If the thickness (H) of the residual substrate is assumed to be very thick compared to the film thickness (h), the above equations (5, 6) are simplified to equations (7, 8) by: And summarized. γ (horn) = 0 P = P1 M = M1 U = 1 V = 1/12 ω = 52.07 degrees η = h / H ~ 0 When all constants C are 0<maths num="4"><img file="JP2010103488A_D0004.tif" /></maths>(7, 8) Is.

Equations (7, 8) are used to derive various constructs of the cleavage method. The only change in the constitutive equation is to include the appropriate forces (P) and moments (M) for the individual loading conditions.

Hydrogen cleavage model

The hydrogen cleavage plane has a peak stress σ centered on the depth h, and h +/- h_cIt is understood that it is reasonably well modeled by using a triangular stress distribution in which the stress linearly decreases to zero. Figure 3 shows h = 50um and h_cThe distribution at = 5um is shown. The functions of M and P are P = -σh_c/ 2 M = + σ (hh)_c/ 4-h_c2/6) Is. The KI and KII functions (ignoring small intersections),<maths num="5"><img file="JP2010103488A_D0005.tif" /></maths>(9, 10) Is.

These equations show that the stress-strength mixed mode ratio is approximately -2.45 and tends to propagate in the cleavage plane towards the surface. However, when this happens, the stress coefficient drops rapidly below the threshold energy required to support propagation.

In "Hydrogen-Induced Silicon Wafer Splitting", J.App.Phys., Vol.94, No.3, pp.1454 (2003), F.Yang states that hydrogen is a pressure proportional to the root of the donation amount and temperature. Developed the hydrogen wafer open surface theory to produce. This article is incorporated herein by reference in its entirety for all purposes.

According to this theory, the phenomenological stress function can be described as follows. σ = sqrt (kTφ) * α_H* K₀ (11) here,<maths num="6"><img file="JP2010103488A_D0006.tif" /></maths>And kT = Boltzmann constant * absolute cleavage surface temperature, and φ = hydrogen donation amount Is.

Ratio h / h_cEquations 8-11 show depth independence of the stress strength factor when is found to be invariant over the cleavage depth of interest, or about 20-150 um for hydrogen injection into silicon. ..

General external energy model

LEFM makes it possible to superimpose external energy on the existing hydrogen cleavage surface by adding the following KI and KII stress strength functions.<maths num="7"><img file="JP2010103488A_D0007.tif" /></maths>(12, 13) Here, the cleavage plane has a peak stress σ and a characteristic width h._cIt is located in the hydrogen cleavage plane and has external energy applied to the system in the form of force P and moment M. They work to correct cleavage energy and stress strength.

The solution to the above general cleavage propagation equation (12, 13) using the energy and KII controlled propagation method is derived in the next section. In all cases, the stress strength equations (12, 13) are solved by the cleavage energy threshold equations (2, 3) to calculate the energy and cleavage direction conditions.

Coordinate system used for the model

To quantify the various configurations, the coordinate system is determined by the cleavage plane along the X direction (positive X to the right) and the film thickness along the Y direction (positive upward). The Z direction is positive from the paper to the front. This coordinate system is used in the code for finite element analysis (FEA). Software available from AnSys, Inc. in Canonsburg, Pennsylvania (Ansys Software) is used to model crack propagation systems and interacts with the interface coupling zone model (CZM). Includes thermal and mechanical parts. The CZM code made it possible to determine the conditions under which the applied stresses and moments propagate cracks.

The configuration in Figure 2 with the detached initiation membrane on the left was entered into Ansys software. The results are summarized in the next section and show the counteracting effect of KII resulting from different load (P and M) configurations.

Ansys software was also used here to extract the cleavage energies of Mode I (crack opening) and Mode II (in-plane shear) represented by D1 and D2, respectively. In particular, their cleavage energies are extracted during CZM elimination (cleavage). This extracted data demonstrates the effectiveness of energy and KII controlled propagation techniques.

To test the ansys software, displacement and shear cleavage loads were often compared to closed-form solutions of M-only and P-only cleavage configurations. Cleavage energy is 2.4 J / m for all models², Young's modulus is 187 GPa, coefficient of thermal expansion (α)_CTE) Is 2.5 ppm and Poisson's ratio is 0.272.

External energy composition

A common configuration that guides the pressurization of M and P, which can propagate through the cleavage plane through the desired thickness, is a complex combination of chemical, physical, thermal and mechanical elements. .. They can be summarized as follows.

A. Mechanical forces and moments

Forces and / or moments mechanically exerted on the membrane can be applied to cleavage energy and assist or enable controlled propagation. Due to the limited thickness of the membrane and its relative fragility, the mechanical energy input is preferably a positive moment in the form of bending of the membrane in a controlled manner.

Moment application conditions

The moment applied to the membrane was the bending of the membrane away from the rest of the substrate. Such bending will create KI and KII conditions that tend to break the membrane. If pure moments (P = 0) are used, KI and KII change with the moment value, but ω does not change with a value of 52.07 degrees, so -cos (ω) / sin (ω) or- It has a constant ratio (KII / KI) of 0.78. This ratio is defined as the mixing ratio of stress strength, the value of which determines the tendency towards the propagation front to change the thickness (in the Y direction). Therefore, the application of the moment is negative in the mixing ratio (breaks the membrane) and is proportional to the bending of the membrane as experienced by the tip of the crack.

There are stable, metastable and unstable methods for applying moments to the membrane. The main configurations are summarized below.

Constant displacement load

Since all crack expansion reduces the moment applied, a constant displacement applied to the membrane at a given distance away from the crack tip is considered a stable load configuration. The displacement increases to reload the crack tip and the propagation process can resume. Other means of producing a constant distance to both X and Y at any point below the blade or pre-cleavage membrane will produce a constant displacement load.

Constant force load

Since all crack expansion increases the applied moment, a constant force applied to the membrane at a given distance away from the crack tip is considered an unstable load configuration. This usually results in uncontrolled crack expansion and is therefore not a suitable load configuration.

Constant moment load

A constant moment load on the membrane can be formed using a given configuration as described herein. One advantage of this configuration is the ability of the system to stay on the load of the desired moment without active assistance. For example, if crack expansion is caused by applying a second load, the system will continue to be loaded by the same moment. If the moment is chosen to exceed the fracture strength of the material, this configuration will be eliminated as uncontrolled crack expansion can occur.

The membrane bending moment can be applied statically or quasi-statically, or in a dynamic manner such as ultrasonically excited in or out of resonance. In all cases, the moment load M transmitted to the crack tip according to certain embodiments of the present invention is a design point that allows ultrasonic vibration, displacement, and controlled crack propagation along the desired cleavage plane. It is designed to increase the stress strength coefficients of KI and KII using the force to.

If the cleavage plane is sufficiently guided (high Δ)_KII-And Δ_{KII +}), There is no reason to add P to the configuration, and threshold crack expansion along the desired cleavage plane is possible. The more the original KII / KI mixing ratio of -0.78 in this moment load configuration is invalidated, the more Δ_KII-And Δ_{KII +}This happens if is high enough. This does not occur spontaneously in silicon, but on the optimal cleavage plane, which can be a sufficient load configuration to achieve membrane detachment.

BH-Layer Cleavage Layer

The presence of an injection layer that drives and guides the propagation cleavage front is another element of a successful cleavage configuration. The use of low-donation cleavage layers in addition to external energy (mechanical, thermal, etc.) can be simulated using equations (12, 13) to produce stable and desired cleavage conditions. As the equation shows, in order to achieve energy-controlled propagation cleavage, the compression cleavage layer applies a negative force (-P) and a positive moment (+ M) to the equation, or has an appropriate external M, P input. In addition, KII controlled propagation can occur.

C. External thermal energy

The use of external heat input energies can produce a wide variety of M and P inputs in equations 12 and 13. Figure 4 shows a heating and cooling matrix that can be used as a general guide in choosing the appropriate thermal energy arrangement and format. This matrix assumes cooling or heating of the surface and is therefore partially filled. For example, volumetric heating on the side of the substrate below h / 2 reverses the moment to M +.

With the use of constant temperature across the membrane, up to the thermal M and P loads to achieve KII cancellation, thermal energy can predominantly become P loads. In addition, they will be expanded in the next section.

The time factor of the load is important. For example, membrane cooling and heating must occur within a predetermined time to form an effective P-load. Therefore, the use of membranes as part of the heat loading configuration is effective only under certain conditions.

FIG. 5 shows the time coefficient of heat treatment according to the two load examples shown in the graph. The first example is a surface substrate heating method applied to an existing injection cleavage layer in order to achieve stable cleavage propagation via energy control. In this example, a laser heating source is described, but other surface and subsurface heating sources may be effective as well.

The second load example is the use of an electron beam pulse volumetric heating source. This allows the membrane side near the cleavage anterior to be adiabatically heated to generate M and P loads to achieve KII controlled propagation. This method can use the injection layer, but can provide cleavage depth control without any propagation injection donation amount.

Cleavage configuration leading to energy controlled propagation

General equations 12 and 13 show that the cleavage plane can act as an energy barrier to maintain propagation within the desired depth. If a varying amount of propagation donation is used, the cleavage depth tends to vary to maintain the cleavage threshold energy condition. For example, lower propagation donations tend to increase cleavage depth so that more injected stress layers are obtained. This is h + h on the cleavage plane_cIf the integral stress on the entire surface of the film should be below the threshold energy cleavage condition, cleavage can no longer occur. External Ms and Ps can inject sufficient energy to sustain propagation and tend to lower cleavage depths. If the injection dose is too small, its depth control function is lost and uncontrolled cracking can occur. Therefore, if a reliable energy controlled propagation cleavage is used, it is an important consideration that a sufficiently high donor amount of hydrogen exists to serve as a depth control layer.

Hydrogen limited cleavage (50um cleavage depth) is about 2.5 × 10 at 650-700 ° C cleavage temperature¹⁶cm^-2It has been experimentally confirmed that.

The conditions for thermally induced membrane stress are summarized in the KI and KII equations as follows. M = 0 and P = -σ_th* h Size α_CTEThe thermal stress function of EΔT is the coefficient of thermal expansion α_CTEThe equation is that it is found to be caused by temperature changes in materials with (14) P = -α_CTEEhΔT Is. Here, ΔT is the temperature difference between the membrane exposed to the temperature treatment and the material to be bonded.

Substituting into equations (12, 13)<maths num="8"><img file="JP2010103488A_D0008.tif" /></maths>(15, 16) Is. ΔT is the temperature difference given by the material above the cleavage front and the residue of the substrate. This is mostly KII shear cleavage as the heating source is added to KII.

Since the first term is proportional to the absolute temperature and the second term is proportional to the temperature difference, thermal immersion following a dynamic heating pulse can form a temperature distribution that is adjusted to optimize cleavage conditions. For example, if a higher absolute cleavage surface temperature is used, a lower propagating cleavage donation is effective. At that time, the temperature differences can be superimposed.

This two-step process can be easily realized by changing the time intensity distribution of the heating source. For example, for a 50um silicon film, the film thermal time constant is about 30 to 40 useconds.

Heat immersion can occur by heating in the range of hundreds of microseconds to milliseconds, while heating pulse ΔT can occur efficiently by heating with higher intensity in the tens of microseconds time factor. For example, a 500 ° C absolute temperature can be formed using a millisecond laser beam heat treatment in addition to a film-substrate temperature difference of 100 to 300 ° C to satisfy the threshold temperature.

If the substrate does not change with respect to the applied external energy, a thermal immersion / pulse approach can be performed. In another embodiment, the substrate or the energy applied can vary in correlation. In certain embodiments, a single source of energy applied may have intensities that change over time in order to achieve immersion and pulsing. According to another embodiment, multiple sources of energy applied can be scanned with one source (immersion) that raises absolute temperature and a second source (pulse) that forms a temperature difference.

Figures 6A and 6B show a two-step thermal immersion / pulse process as a thermal beam colliding on a substrate surface just in front of the cleavage front. If shock waves (see next section) are not used, all heating of the cut membrane simply spreads over the membrane.

Cleavage configuration leading to KII control propagation

General equations 7 and 8 show that the forces and moments applied to the membrane produce the KI and KII stress strength coefficients at the crack tip as described in the equation. Membrane force P is the force per width unit applied to the membrane and is related to the membrane stress at the crack tip such as P = σ * h. The moment is the moment transmitted on the membrane that produces M at the crack tip.

The fact that the relationship between M and P is addition in equation 7 but subtraction in equation 8 suggests that the pressurization of P and M can be applied in KI and decreased in KII. Therefore, if the moment M and the membrane force P are properly selected, the condition that KII is canceled (zeroed) and crack expansion is achieved is that the propagating crack propagates along the cleavage plane. Will ensure that it has the greatest tendency for. This optimal set of conditions allows low or zero donation cleavage plane operation and is further controlled via active control of M and P as a function of the actual cleavage depth achieved during crack expansion. Allows depth guidance. According to embodiments of the present invention, control over such reduced or no-donation conditions and cleavage depths is particularly pronounced among the effective benefits realized.

Cleavage using thermal shock (dynamic cleavage)

According to certain embodiments, cleavage can be achieved through the action of stress within the membrane layer caused by rapid temperature differences. In particular, exposure of the substrate to cooling (cooling impact) or heating impact can result in membrane cleavage. On the other hand, the impact heating of the membrane behind the crack tip generates the stress required for the crack to proceed under predetermined conditions. Both P-load configurations are summarized in more detail later in this section.

The heat-only cleavage energy equation (without H cleavage plane) can be derived by solving equations (2, 3) (7, 8) with M = 0 and P defined in equation (14). (17) G'= 2 * γ'= 1/2 * α_CTE²* E * h * ΔT²* (1-v²)

According to a particular embodiment of the present invention, the operation of the positive thermal KII and the negative moment KII simultaneously realizes the conditions of G'= 2 * γ'(arbitrary appropriate threshold cleavage energy condition) and KII = 0. Can be combined to give rise to a cleavage propagation technique that allows.

Thermal heating shock cleavage

If the heating impact is deployed fast enough to transfer the inertial force to the cleavage front, the heating impact can be applied onto the membrane behind the crack tip. To allow the heating shock to deploy P efficiently in this configuration, the heating produces a shock wave before there is mechanical relaxation of expansion through the movement of the membrane separating from the crack tip region.

Compared to the prior membrane substrate heating configuration when the heating time is associated with the thermal time constant of the membrane, the membrane heating configuration is applied to a time coefficient of the same order as the acoustic time constant of the membrane material. Therefore, high speed heating is applied at very short time intervals in the range of less than 100 nanoseconds to a few microseconds, depending on the width of the heating impact and its proximity to the crack tip.

To achieve the threshold crack propagation condition (ΔT = 183 ° C), for example, with 50 um silicon, 6 MW / cm within 250 nanoseconds.²Power density is required. This is a very high power density that makes silicon an adiabatic heating type.

Volumetric heating is preferred over surface heating to avoid creating thermal stresses along the silicon film depth, which can lead to damage in the form of spotting, surface melting, film breakage and cracking. Electron beam (E-beam) heating is a suitable alternative technique applied to this heating because it can volumetrically heat a portion of the membrane (open or uncleavage).

In addition, the electron beam can be highly controlled by the following general characteristics. The electron beam allows the beam to be scanned over a large area due to its high degree of control and speed. The electron beam allows excellent control of beam strength and diameter (from microns to centimeters in size). The electron beam allows control of pulse processing from nanoseconds to CW. The electron beam allows control of the heating depth by converting the electron beam energy (from KeV to MeV). As shown in FIG. 7, their electron beam energies correspond to a beam transmission range from a few microns to a few centimeters.

Therefore, the power flow rate of the beam is easily controlled by selecting the diameter of the beam, the energy of the beam and the intensity of the beam, while the transmission range is selected by the beam energy. For example, a 50um electron beam transmission in silicon is about 80KeV, and a 0.5mm beam diameter requires a beam intensity of 200mA for a 250nsec pulse. On the other hand, pulsed beam processing can be done with a CW beam that is scanned fast enough. For the above example, the required beam scanning speed would be 0.5 mm / 250 nsec or 200,000 cm / sec.

Most, if not all, of the above properties can be useful in existing systems built for precision vacuum welding and material modification applications. For example, electron beam systems from Pavac Industries, Inc. (Richmond, BC Canada, www.pavac.com) and others can be used as an adiabatic heating source.

The electron beam can control the propagation of the crack tip under the KII canceling configuration. Also, electron beam technology can be used to initiate the first region of the membrane as well as possible cleavage under pure shear conditions. Those methods are described below.

Cleavage depth control using KII cancellation

Maintaining precise control of film thickness is important not only to achieve a highly uniform film, but also to avoid cleavage failure. The KII cancellation scheme provides a unique function of controlling the propagation cleavage depth by adjusting the moment applied as a function of active depth control feedback.

Equations 12 and 13 show how the KI and KII stress strength coefficients change with the applied load. Using an adiabatic heating source, KII cancellation is possible with positive P and M. With reference to FIG. 4, this combination is possible by membrane impact heating.

The electron beam is effective in meeting the required specifications. Figure 8 shows a Monte Carlo simulation of the energy deposition heating distribution for a 100 KeV electron beam in silicon. If the energy is deposited faster than the thermal diffusion time constant of the system, this distribution also shows the instantaneous temperature distribution and thermoelastic stress distribution transmitted to the system.

By varying the beam intensity and energy, M and P can be simultaneously adjusted to the desired combination that can result in energy or KII controlled propagation. With electron energy of 1 MeV or less, it is recognized that the general shape of the volumetric heating distribution is energy independent and only a measure of depth.

For example, the solutions of equations 12 and 13 with simulated electron energy deposition distributions may demonstrate the effectiveness of the method of the invention for generating KII controlled propagation. FIG. 9 shows a simulated electron beam distribution with a peak temperature of 635 ° C and a peak position of about 5.35 um within the membrane. Assuming that M and P are completely bonded to the crack tip, this is a moment M ~ 8 × 10^-2Nm / m and force P ~ 4.3 × 10³Will produce N / m. The KI of the calculation result is 2.4J / m²Just beyond, KII is countered.

Figure 10 shows the stress strength components (KI as D1).²/ E'and KII as D2²The contribution of cleavage energy from each of / E') is shown. The condition of KII = 0 corresponds to D2 = 0, while the D1 energy is just 2.4J / m.²It is in a state that exceeds. α_CTENote that we assume a more accurate 3.5 ppm over the temperature range of 600-700 ° C. This KII = 0 condition will propagate cracks horizontally along the desired depth h = 50um.

This condition should allow the fissure to propagate to a given distance until it has stopped progressing. Since the cleavage rate is faster than the system's thermal time constant, this temperature distribution is persistent and does not dissipate heat through the cleavage process.

The size of the electron beam from about 500um to several millimeters should be about the same length and expand the fissure. After thermal relaxation of the thermal pulse, the process can repeat moving the crack in another incremental manner.

The above describes the use of a pulsed beam over the membrane region adjacent to the crack tip. On the other hand, the CW or pulsed beam could achieve the same function by scanning along the cleavage tip (in the figure). By adjusting the beam intensity to provide the same temperature-time distribution using the beam spot displacement velocity, the scanning beam can dynamically extend the crack in synchronicity with the scanning beam. Therefore, depth control can be achieved by adjusting the energy and intensity of the beam, during which the crack propagates to achieve the desired depth. This indicates that even reduced or no donation propagating is possible using the techniques of the invention.

Utilization of thermal shock technology for membrane initiation reaction

The use of patterned injections has been described above in relation to enabling the lowest total donation membrane cleavage process. In the proposed process sequence, the membrane initiation reaction is achieved with a higher donation to partially release the membrane several millimeter wide. Under another method of KII cancellation or membrane propagation, this partially released membrane can be used to propagate the cleavage balance in turn.

Electron beam technology, laser or flash lamp technology can be used to strip the film from the substrate. Electron beam technology may be particularly suitable for this purpose, as the energy range can be adjusted to allow the film temperature to increase volumetrically within the starting region near the edge of the substrate. The pulsed energy accumulates a nearly constant temperature rise ΔT within the film thickness over time, with rapid changes in the temperature distribution adjusted at the cleavage plane or where it is adjacent.

Such adiabatic heating of silicon by electron beam exposure is shown in Figure 9, which plots the energy density vs. depth in silicon for electron beam irradiation over a period of time. The beam intensity and energy are selected to correspond to KII controlled propagation. (111) In this example of 50um membrane cleavage on silicon, KII is counteracted at a peak thermal depth of 5.35um.

For example, a 250 nsec pulse basically only diffuses the temperature distribution at about 5 um, which is thinner than the film thickness. This causes a cleavage motion that limits the presence of thermal shock in the thickness direction of the membrane, producing KI and KII components that can cleave the membrane in a controllable manner. The initiation cleavage begins at an edge and can be stretched over the desired membrane width to support the propagation cleavage method. Alternatively, the starting cleavage can begin in the medial position and then extend to the outer circumference. If an inner starting cleavage is used, the vacuum environment does not have enough pressure to generate a restoring force against the upward movement of the membrane, so the cleavage is done by reducing the energy required for buckling of the membrane. The process is convenient.

Use of insulation technology for membrane propagation

Also, concentrated heat pulses combined with volumetric heating properties can be used to propagate the membrane from the starting region. Applications of fast pulses that can be uniformly heated to film thickness, especially using electron beam technology, can allow controlled propagation of the film, either by pure shear cleavage or with the help of moments. The use of electron beam technology for membrane cleavage is described in US Pat. No. 6,013,563, which is hereby incorporated by reference in its entirety for all purposes. According to the individual embodiments of the present invention, the irradiation of the electron beam can be applied perpendicular to the front surface of the substrate.

Matching of electron beam volume heating (volumetric heating shape and peak heat temperature Rp) can be achieved by selecting a higher or lower energy electron beam. Its R within the thickness of the membrane cleaved to produce the desired moment M_pAs there is, the electron beam energy can be selected.

The use of electron beam irradiation may be particularly suitable here, as the range of energy can be adjusted to allow a positive increase in membrane temperature within the membrane near the crack tip. The pulse energy accumulates abrupt changes in the temperature distribution adjusted to produce the desired M and P loads within the film thickness.

For example, a 250 nsec pulse basically only diffuses the temperature distribution at about 5 um, which is thinner than the film thickness. This causes a shear cleavage that limits the presence of thermal shock in the thickness direction of the membrane, resulting in a large shear that can control the membrane.

The use of electron beam heating methods is convenient in this mode, as rapid adiabatic heating provides a rapid compressive stress distribution that can facilitate the induction of propagation cleavage. Cleavage depth can also be controlled by changing the thickness of the membrane exposed to volumetric heating, for example, in exchange for electron beam energy.

It was also confirmed that the choice of heating area width results in control over cleavage and extension. The heating zone width first determines the amount of crack elongation that occurs during the heating cycle. This is because after the anterior surface of the cleavage reaches the edge of the heated cleavage region, G'falls and the propagation cleavage stops.

Finally, all of the heating techniques used with the injection cleavage layer can facilitate the more efficient use of injection-induced stresses. This is caused by higher temperatures and at the same time increases the injection stress (they are usually proportional to kT). Therefore, the heating pulse has a secondary favorable effect on the cleavage plane and increases the stress on the cleavage plane for better induction of cleavage propagation.

Continuation of the cleavage process ultimately results in complete exfoliation of the thick film along the cleavage region from the residue of the semiconductor substrate. In certain embodiments, this residue with the newly exposed surface area is in a state for repeating the cleavage process by re-implementing the cleavage initiation and propagation method.

Cleavage surfaces with lower surface energy may be selected as the surface region for the semiconductor substrate. Thus, in certain embodiments, after any thick film has been exfoliated from the substrate, the new surface is essentially in good condition in the original crystal orientation plane and is relatively small for additional layer transfer products. It has roughness and does not require complicated surface treatment such as corrosion or polishing.

Of course, there may be other modified forms, modified forms, and alternative forms. Depending on the embodiment, the process can be for the composition of solar cells, integrated circuits, optical devices, all combinations thereof, and the like.

The above is all descriptions of the specific embodiments, but various modifications, alternative sequences and devices may be used. The above is described using the selected order of steps, but any combination of elements of the described steps or anything else can be used. Also, any step can be combined and / or removed depending on the embodiment.

In addition, the types of ion particles can vary from hydrogen ions to helium ions, deuterium ions or any combination to allow the formation of open regions, according to alternative embodiments. Furthermore, the cleavage process may include temperature control / assist cleavage using a vacuum chuck or electrostatic chuck process. Of course, there may be other modified forms, modified forms, and alternative forms.

Added here as Attachment A and incorporated herein by reference in its entirety for all purposes is the "Direct Film Transfer (DFT) Technology for Kerf-Free Silicon Wafering" by Henley et al., The following treatise. Added here as Attachment B and incorporated herein by reference in its entirety for all purposes is a set of presentation slides.

Therefore, the above embodiments and embodiments are for explanatory purposes only, and from these points of view various modifications and changes are suggested to those skilled in the art and are included within the spirit and scope of the present application and the appended claims. It is also understood that