Ion implantation method, silicon-on-insulator substrate and method for manufacturing the same

Abstract

The present application provides ion implantation method, silicon-on-insulator substrate and its preparation method. The wafer to be treated is carried on the electrostatic chuck of the ion implanter, and ion implantation is carried out using ion beam in at least two implantation steps. Between two adjacent implantation steps, the wafer to be treated is rotated circumferentially relative to the electrostatic chuck, so that the same wafer area is sequentially cooled by different cooling areas in different implantation steps, reducing the continuous influence of fixed cooling space difference. The number of circumferential positions used during implantation can be determined by angular harmonic analysis of the thermal wave data of the reference wafer. After the donor wafer completes hydrogen ion implantation, it is bonded with the acceptor wafer and is subjected to heat treatment to peel off along the implanted layer, forming a top layer of silicon on the acceptor wafer to obtain a silicon-on-insulator substrate.

Description

Technical Field

[0001] This application relates to the field of semiconductor fabrication technology, specifically to an ion implantation method, a silicon-on-insulator substrate, and a method for fabricating the same. Background Technology

[0002] Silicon on Insulator (SOI) substrate is a semiconductor substrate structure that includes a supporting substrate, an insulating layer, and a top silicon layer. The insulating layer is located between the supporting substrate and the top silicon layer.

[0003] SOI substrates can be fabricated using layer transfer processes. For example, ion implantation is performed on a donor wafer, the donor wafer is bonded to a acceptor wafer, and then the bonded wafer structure is peeled off to form a top layer of silicon on the acceptor wafer side.

[0004] In related ion implantation processes, there is an issue of uneven implantation conditions within the wafer surface. These differences in implantation within the wafer surface can affect the thickness distribution of the top silicon layer formed by subsequent layer transfer processes. Summary of the Invention

[0005] This application provides an ion implantation method. A wafer to be processed is placed on an electrostatic chuck of an ion implanter, and an ion beam is used to implant ions into the wafer in at least two implantation steps. Between adjacent implantation steps, the wafer to be processed is rotated circumferentially relative to the electrostatic chuck, so that the same area of ​​the wafer to be processed is sequentially cooled by different cooling regions of the electrostatic chuck in different implantation steps.

[0006] This application also provides a method for fabricating a silicon-on-insulator substrate. A donor wafer is provided as the wafer to be processed. An ion implantation method is used, employing at least two implantation steps and rotating the donor wafer circumferentially relative to an electrostatic chuck between adjacent implantation steps, to implant hydrogen ions into the donor wafer, forming an implantation layer. The donor wafer is then bonded to a acceptor wafer. The bonded structure is then heat-treated to peel off the donor wafer along the implantation layer, forming a top silicon layer on the acceptor wafer, thus obtaining a silicon-on-insulator substrate.

[0007] This application also provides a silicon-on-insulator (SCI) substrate. The SCI substrate includes a support substrate, an insulating layer, and a top silicon layer, with the insulating layer located between the support substrate and the top silicon layer. The top silicon layer is formed from a donor wafer through hydrogen ion implantation, bonding, and stripping. Hydrogen ion implantation is performed in at least two implantation steps, during which the donor wafer is supported on an electrostatic chuck. Between adjacent implantation steps, the donor wafer is rotated circumferentially relative to the electrostatic chuck, so that the same region of the donor wafer is sequentially cooled by different cooling regions of the electrostatic chuck in different implantation steps. The number of circumferential positions of the donor wafer relative to the electrostatic chuck during the implantation steps is determined based on the amplitudes of multiple orders of angular harmonics, which are obtained from position data of multiple measured positions on a reference donor wafer and the corresponding thermal wave values. The number of circumferential positions is greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold. The thermal wave uniformity of the donor wafer after hydrogen ion implantation is less than or equal to... Thermal uniformity is defined as the ratio of the sample standard deviation to the mean of thermal wave values ​​at multiple measurement locations on the donor wafer after implantation, expressed as a percentage. The thickness uniformity of the top silicon layer is less than... Thickness uniformity is the difference between the maximum and minimum thickness values ​​at multiple thickness measurement locations of the top silicon layer.

[0008] In this embodiment, the wafer to be processed is rotated circumferentially relative to the electrostatic chuck between two adjacent implantation steps. This allows the same wafer region to be cooled sequentially by different cooling regions of the electrostatic chuck in different implantation steps. This reduces the impact of continuous cooling space differences acting on the same region, thereby improving the uniformity of the implantation state within the wafer plane and the thickness distribution of the top silicon layer after layer transfer. Furthermore, multiple orders of angular harmonics are obtained based on the position data of multiple measurement locations on the reference wafer and the corresponding thermal wave values, reflecting the circumferential variation characteristics of the thermal wave spatial distribution. The number of circumferential positions is determined based on the amplitude of the angular harmonics, and the number of circumferential positions is greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold. This ensures that the number of circumferential positions used in the step-by-step implantation corresponds to the obtained angular spatial variation characteristics. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the ion implantation device according to one embodiment of this application.

[0010] Figure 2 This is a flowchart illustrating the process of determining the number of circumferential positions according to one embodiment of this application.

[0011] Figure 3 This is a schematic diagram of the reference wafer measurement position and angular sampling relationship according to one embodiment of this application.

[0012] Figure 4This is a schematic diagram of multiple circumferential positions according to one embodiment of this application.

[0013] Figure 5 This is a flowchart illustrating a method for fabricating a silicon-on-insulator substrate according to one embodiment of this application.

[0014] Figure 6 This is a comparative spatial distribution map of thermal waves.

[0015] Figure 7 This is a spatial distribution diagram of the heat wave in Example 1.

[0016] Figure 8 This is a spatial distribution diagram of the heat wave in Example 2.

[0017] Figure 9 This is a comparative example of the spatial distribution of the top silicon thickness on a silicon-on-insulator substrate.

[0018] Figure 10 This is a spatial distribution diagram of the top silicon thickness of the silicon-on-insulator substrate in Example 1.

[0019] Figure 11 This is a spatial distribution diagram of the top silicon thickness of the silicon-on-insulator substrate in Example 2. Detailed Implementation

[0020] The technical solutions in this application will now be clearly and completely described with reference to the accompanying drawings.

[0021] In the fabrication of silicon-on-insulator (SiS) substrates using a layer transfer process, a donor wafer is first ion-implanted, then the donor and acceptor wafers are bonded together, and finally, after heat treatment, the top silicon layer is peeled off along the implanted layer. In-plane variations in the ion implantation state can affect the thickness distribution of the top silicon layer after layer transfer.

[0022] The wafer to be ion-implanted can be a bare silicon wafer or a silicon wafer with an oxide layer on its surface, which can be a silicon dioxide layer. In a specific embodiment of fabricating a silicon-on-insulator substrate, a donor wafer with a silicon dioxide layer on its surface is implanted. After bonding and layer transfer, the silicon dioxide layer forms an insulating layer between the top silicon layer and the supporting substrate.

[0023] During hydrogen ion implantation, the donor wafer is supported on an electrostatic chuck (ESC), which provides cooling. To mitigate implantation inhomogeneities caused by differences in ion beam spatial distribution, conventional corner implantation divides ion implantation into... Between two consecutive implantations, the electrostatic chuck rotates the donor wafer relative to the ion beam. , This indicates the number of conventional corner implantations. However, in some hydrogen ion implantation experiments, even with conventional corner implantation, the thermal wave (TWW) after implantation is still observed. The test results still showed an abnormal spatial distribution, with one side being higher and the other side lower. Thermal wave testing is used to evaluate the wafer state after implantation, and the distribution of thermal wave values ​​on the wafer surface is called thermal wave spatial distribution.

[0024] Analysis of the conventional corner implantation process reveals that when the electrostatic chuck rotates the donor wafer, the relative position between the donor wafer and the electrostatic chuck remains unchanged, and the same region of the donor wafer always corresponds to the same bearing region of the electrostatic chuck. When the cooling capacity of some bearing regions of the electrostatic chuck is stronger than that of others, the cooling effect on different regions of the donor wafer varies, potentially leading to localized temperature differences. The cooling region refers to the spatial area where the electrostatic chuck provides cooling to the supported wafer. The high-dose hydrogen ion implantation used in the Smart Cut process is highly sensitive to temperature conditions during implantation. Localized temperature differences may affect the implantation results in different regions of the donor wafer, manifesting as uneven thermal wave spatial distribution after implantation. Therefore, the abnormal thermal wave spatial distribution observed in the experiment may be related to the differences in cooling capacity between different bearing regions of the electrostatic chuck.

[0025] Conventional corner implantation alters the orientation of the donor wafer relative to the ion beam, but does not change the regional correspondence between the donor wafer and the electrostatic chuck. The same region of the donor wafer is continuously cooled by the same bearing region of the electrostatic chuck; therefore, conventional corner implantation cannot eliminate implantation non-uniformity caused by differences in the cooling capacity of the electrostatic chuck.

[0026] When the donor wafer is rotated circumferentially relative to the electrostatic chuck between two adjacent implantation steps, the same wafer region corresponds to different cooling regions in different implantation steps. The superposition of the step-by-step implantation results reduces the continuous impact of differences in the fixed cooling space, thereby improving the uniformity of the implantation state within the wafer plane and the thickness distribution of the top silicon layer after layer transfer. This effect corresponds to at least two implantation steps and the relative rotation between steps; implantation at a single circumferential position is used as a comparative example to evaluate the improvement effect of step-by-step implantation.

[0027] When the spatial distribution of heat waves changes with process conditions, the number of circumferential positions selected empirically may need to be adjusted through experiments. For example, the number of circumferential positions can be empirically set to 2 or 4; a further implementation method determines the number of circumferential positions based on the angular changes in heat wave data, so that the selection criteria correspond to the actual spatial distribution.

[0028] The wafer undergoing ion implantation is called the wafer to be processed. In a silicon-on-insulator (SiI) fabrication scenario, the wafer to be processed is the donor wafer, used for bonding and stripping after hydrogen ion implantation. The wafer used to acquire step-by-step data is called the reference wafer, and the corresponding donor wafer is called the reference donor wafer. The reference donor wafer is used to acquire reference thermal wave data, and the wafer using the selected step-by-step implantation process and used for subsequent layer transfer is called the donor wafer.

[0029] The reference wafer and the wafer to be processed can be different wafers, but they use the same wafer specifications and the same ion implantation conditions except for the relative rotation angle between steps. After the reference wafer completes the baseline ion implantation, it undergoes thermal imaging testing, and the wafer to be processed uses a step-by-step implantation process determined based on the thermal imaging data of the reference wafer.

[0030] Step-by-step injection alters the correspondence between the same wafer region and the electrostatic chuck cooling region by rotating the wafer to be processed circumferentially relative to the electrostatic chuck. When using thermal wave data for step selection, the number of circumferential positions is determined based on the angular harmonic amplitude of the reference wafer. Then, the wafer to be processed is injected using circumferential positions with a number greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold.

[0031] like Figure 1 As shown, the ion implantation apparatus 100 includes an electrostatic chuck 120 and an ion implantation assembly 140. A wafer 220 to be processed is supported on the electrostatic chuck 120, and the beam exit side of the ion implantation assembly 140 faces the implantation surface of the wafer 220. The ion implantation assembly 140 provides an ion beam, and the electrostatic chuck 120 supports and cools the wafer 220.

[0032] An ion implantation is performed in at least two implantation steps. Between two adjacent implantation steps, the wafer to be processed 220 is rotated circumferentially relative to the electrostatic chuck 120. The circumferential position of the wafer to be processed 220 relative to the electrostatic chuck 120 during the implantation step refers to the circumferential orientation of the wafer relative to the chuck about an axis perpendicular to the wafer surface. The number of circumferential positions is included in the orientation at the initial implantation, and the measurement position refers to the sampling point on the wafer surface used to obtain thermal wave values.

[0033] After the wafer 220 to be processed rotates circumferentially relative to the electrostatic chuck 120, the correspondence between the same wafer region and the cooling region of the electrostatic chuck 120 changes. For example, in the first implantation step, the first wafer region of the wafer 220 corresponds to the first cooling region of the electrostatic chuck 120. After the first implantation step is completed, the wafer 220 to be processed rotates circumferentially relative to the electrostatic chuck 120, so that the first wafer region corresponds to the second cooling region in the second implantation step. Thus, the first wafer region is sequentially cooled by different cooling regions in different implantation steps.

[0034] Within a single implantation step, the wafer to be processed 220 and the electrostatic chuck 120 maintain their relative positions. In one specific implementation, the ion beam is a horizontal band with a width greater than [missing information]. The electrostatic chuck 120 moves the wafer 220 up and down, allowing the ion beam to scan the wafer surface. A single implantation step can also employ conventional corner implantation, typically divided into four stages, with the electrostatic chuck 120 rotating the wafer 220 together. Adjacent corners are... The common motion within a single step and the wafer rotation relative to the chuck between adjacent injection steps are performed separately. Figure 1 This indicates the relative configuration of the ion implantation components, wafer, and chuck. The orientation of the drawing does not limit the mounting orientation of the device relative to the direction of gravity.

[0035] When using thermal wave data for step selection, multiple measurement locations and thermal wave values ​​of the reference wafer are first acquired, and then the angular harmonics and amplitudes are determined. The number of circumferential locations is greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold, thereby selecting a step-by-step injection method for the corresponding angular changes.

[0036] like Figure 2 As shown, the number of circumferential positions of the wafer to be processed relative to the electrostatic chuck during the implantation step is determined based on the thermal wave data of the reference wafer. To represent the step selection formula, this number is denoted as... , It is an integer greater than or equal to 2. In one implementation, one circumferential position corresponds to one injection step, and the number of injection steps is the same as the number of circumferential positions, therefore, it adopts... Each circumferential position One injection step.

[0037] In step S100, position data and corresponding thermal wave values ​​of multiple measurement locations on the reference wafer are acquired.

[0038] Thermal wave testing was performed on a reference wafer, and thermal wave values ​​were obtained at multiple measurement locations. With the center of the reference wafer as the origin, the... The data at each location is denoted as ,in , Rectangular coordinates, units are 1. ; The original thermal wave measurement value, The thermal wave value is measured using the same equipment and the same measuring aperture, and the thermal wave value is not equivalent to the thickness of the top silicon layer.

[0039] In one specific implementation, 52 measurement locations are set on the reference wafer, which is the number of measurement locations involved in the calculation. In other implementations, the number of measurement locations can be more than 52.

[0040] In step S120, the rectangular coordinates of multiple measurement locations are converted into polar coordinates.

[0041] The measurement positions on the reference wafer surface provide both distance information to the wafer center and angular information along the wafer circumference. To analyze the spatial variation of thermal wave values ​​along the circumference of the reference wafer, multiple measurement positions are converted from Cartesian coordinates to polar coordinates. The polar angle of the measurement position describes the sampling point and has a different meaning than the circumferential position of the wafer to be processed relative to the electrostatic chuck.

[0042] No. Polar diameter corresponding to each measurement location Determined according to formula (1.1): (1.1) No. Polar angles corresponding to each measurement position Determine according to formula (1.2): (1.2) Polar diameter To measure the distance from the location to the center of the wafer, the unit is . Polar angle Indicates relative measurement position The orientation of the positive axis, in units of . To perform the arctangent operation to determine the quadrant based on two coordinates, this implementation unifies the polar angle to... Interval. Interval endpoints are set by... The periodic connection means that the measurement position involved in the angular analysis is not located at the origin of the coordinate system.

[0043] In step S140, the radial trend in the thermal wave data is removed.

[0044] The thermal wave values ​​in the spatial distribution of thermal waves may vary simultaneously with both the polar radius and polar angle. When the thermal wave component varying with the polar radius is superimposed on the angular spatial variation, the thermal wave values ​​at different polar angle positions will be affected by the distance between the measurement position and the center of the reference wafer. To obtain data for angular analysis, the radial variation in the thermal wave data can be separated before performing angular analysis.

[0045] In one specific implementation, a radial linear fitting relationship is established based on the polar radius and thermal wave values ​​at multiple measurement locations: (2.1) Among them, in equations (2.1) to (2.4) , These are the radial fitting parameters, and the corresponding letter coefficients of the angular Fourier expansion are distinguished according to the formula they belong to. Polar radius The thermal wave fitting value at that location, The intercept is... The slope is the fitted slope. Use the same units as the heat wave value. The unit is the heat wave value unit / ; The positive or negative sign indicates that the fitted thermal wave value increases or decreases as the polar diameter increases.

[0046] Fitting slope Determine according to formula (2.2): (2.2) Fit intercept Determine according to formula (2.3): (2.3) For the Each measurement location, from the thermal wave value After deducting the polar diameter The corresponding heat wave fitting value is used to obtain the residual heat wave value. : (2.4) Radial detrending reduces the impact of the overall variation of thermal wave value with polar radius on angular analysis, avoiding direct attribution of radial variations to angular harmonics. The summation in equations (2.2) and (2.3) both affect... to The denominator of equation (2.2) is non-zero, requiring that the measurement positions involved in the fitting have different polar diameters.

[0047] In step S160, the thermal wave data are sorted according to the polar angles of multiple measurement locations, and the corresponding weights are determined based on the polar angle spacing.

[0048] Arrange the polar angles in non-decreasing order to obtain When the polar angles are the same, maintain the original measurement order and do not exchange positions with the same angle. Let... For the sorted number The original index corresponding to each position, and the polar angle after sorting. Residual heat wave values ​​after sorting Location, polar angle, and residual thermal wave value correspond to the same index.

[0049] Multiple measurement locations can be distributed non-uniformly along the polar angle direction. When the polar angle distribution is uneven, the circumferential angle range represented by different measurement locations will differ. To correlate the contribution of different measurement locations to the angular analysis with their corresponding angle ranges, the weights corresponding to the measurement locations are determined based on the polar angle distance between adjacent measurement locations.

[0050] In the implementation method where the 52-point measurement location covers a full circle, the angular distance between the first and last cycles is denoted as... ,according to The period is taken as the shorter distance between the first and last polar angles. In this embodiment, the distance between the first and last boundaries is less than... ,satisfy .

[0051] In a specific implementation using 52 measurement locations, the first... angular spacing corresponding to each measurement position Determine according to formula (3.1): (3.1) Determine the first according to the angular distance. Normalized weights corresponding to each measurement location : (3.2) angular spacing Indicates the first Each sorting position represents a range of angles, in units of... Normalized weights Dimensionless, the sum of weights is 1. Positions with the same polar angle are retained according to the original measurement order, and their weights are calculated separately according to equation (3.1). If two adjacent positions of a certain position have the same polar angle, the angular distance and weight of that position are 0.

[0052] In step S180, according to the corresponding weights, the thermal wave data after removing the radial trend is decomposed into angular Fourier transform to obtain multiple orders of angular harmonics and their corresponding amplitudes.

[0053] Angular Fourier decomposition can be expressed as: (4.1) in, To determine the highest order of analysis, this implementation method takes... In equation (4.1) Take from 0 to , The corresponding constant term represents the average component. , These are the cosine and sine coefficients, respectively, with the same units as the residual heat wave values; the angles in the cosine and sine functions are expressed using... Finite-order expressions are used to approximate angular distributions. Equations (4.2) and (4.3) are only used for... to The coefficients are calculated; the zero-order term does not include the total power of the angular non-uniform components.

[0054] No. Cosine coefficients corresponding to the angular harmonics Determine according to formula (4.2): (4.2) No. Sine coefficients corresponding to the angular harmonics Determine according to formula (4.3): (4.3) in, The residual heat wave values ​​after sorting. For the corresponding polar angle, Equations (4.2) and (4.3) approximate the coefficients using a weighted integral with a sum of weights equal to 1, without dividing by the number of measurement locations. . No. Amplitude of the angular harmonic Determine according to formula (4.4): (4.4) Different orders represent different numbers of periodic changes within a week. The first order corresponds to a unidirectional high-low change, the second order corresponds to two periodic changes and can be elliptical in distribution, the third order can be trilobed, and the fourth and fifth orders can be tetralobed or cross-shaped and pentlobed, respectively. Higher-order components may be related to local variations or measurement noise, and cannot be entirely identified as noise based solely on their order.

[0055] The multiple orders of angular harmonics obtained from angular Fourier decomposition and their corresponding amplitudes together characterize the angular spatial features in the spatial distribution of thermal waves. The number of circumferential positions can be determined based on the amplitude of the angular harmonics, rather than using the thermal wave value at a single measurement location as a direct basis for determining the number of circumferential positions.

[0056] In step S200, the number of first candidate circumferential positions is determined based on the maximum order of the angular harmonics whose amplitude exceeds a preset threshold. The preset threshold is determined based on the empirical noise floor of the thermal wave data.

[0057] In one specific implementation, the average of the fourth-order angular harmonic amplitude and the fifth-order angular harmonic amplitude is used as the empirical noise floor estimate. Determine according to formula (5.1): (5.1) The preset threshold is denoted as In this implementation, the empirical noise floor estimate is taken as three times the empirical noise floor estimate. The empirical noise floor is not the standard deviation obtained from independent repeated measurements, therefore this threshold is used for amplitude screening.

[0058] (5.2) The amplitudes of angular harmonics of different orders are compared with preset thresholds. Angular harmonics with amplitudes exceeding the preset thresholds are used to determine the number of circumferential positions.

[0059] The highest order of angular harmonics whose amplitude exceeds a preset threshold Determine according to formula (5.3): (5.3) The number of first candidate circumferential positions is determined based on the maximum order of the angular harmonics whose amplitude exceeds a preset threshold. The number of first candidate circumferential positions is 1 greater than the maximum order.

[0060] In equation (5.3) It is a set of positive integers. This indicates taking the maximum value in the set, applicable when at least one non-zero magnitude exceeds the threshold. To analyze the upper limit, The two terms, "maximum order" and "magnitude", have different meanings.

[0061] (5.4) For an equal-weighted rotational stacking model with the same chuck space distribution, step-by-step injection uses circumferential positions distributed at equal angles along the circumference. The total injection dose is evenly distributed across each injection step, and the chuck space state between steps is considered to remain comparable. Let... This represents a periodic distribution at a fixed radial position that varies with the polar angle, and this distribution can be expanded into an angular Fourier series:

[0062] In the formula This is the average value of the angular distribution. and These are the cosine and sine coefficients for the corresponding orders. The distributions corresponding to the circumferential positions at equal angles are averaged and superimposed with equal weights, and the average level represented by the zero-order term remains unchanged.

[0063] Therefore, in the equal-angle, equal-weight rotation superposition model, when At that time, 1 to Rank is not Integer multiples of these orders, the angular components of these orders are eliminated by average. It can eliminate first-order components. It can eliminate first- and second-order components. It can eliminate first to third order components. It can eliminate first to fourth order components; for higher orders, it is still necessary to check whether they can be eliminated. Divisible components. The summation of sets with the same position and the same weight does not change with the order of arrangement, but this mathematical property cannot be used to assume that the thermal history in actual processes is unaffected by the injection order. This elimination relationship is premised on the validity of the model conditions and cannot be derived from the fact that the number of circumferential positions is greater than the maximum order, thus failing to derive the elimination results under arbitrary angles and dose distributions.

[0064] In step S210, the number of second candidates is determined based on the residual non-uniformity corresponding to the number of candidate circumferential positions. The residual rate is evaluated using an equal-angle, equal-weight rotation superposition model, which evaluates the theoretical residual of the angular component. Actual thermal wave results are verified by measurements after implantation. The theoretical residual rate is not equivalent to the thermal wave uniformity after donor wafer implantation, nor is it equivalent to the thickness uniformity of the top silicon layer.

[0065] No. Power of the angular harmonic Determine according to formula (5.5): (5.5) Total harmonic power of all angular harmonics Determine according to formula (5.6): (5.6) Indicates the first The square of the amplitude, From 1 to The sum of the squares of the amplitude values, both in units of the square of the thermal wave value. Power here represents the square of the angular component, not the physical power of the ion beam.

[0066] For the number of candidate circumferential positions The corresponding residual rate Determine according to formula (5.7): (5.7) In the formula, For the modulo operation, the summation covers 1 to 1. Zhongneng was The order of divisibility. for The proportion of non-zero order harmonic power retained after averaging at equal angular positions to the total harmonic power. When calculating this proportion... If the value is greater than 0, the zero-order term and the radial trend that has been subtracted are not included in the numerator or denominator.

[0067] In one specific implementation, Set as the residual rate threshold. From multiple candidate circumferential positions, select those with a residual rate less than [a certain threshold]. The minimum number of candidate circumferential positions is used as the number of second candidate circumferential positions. : (5.8) The residual rate is used to characterize the residual angular spatial components corresponding to different numbers of candidate circumferential positions, thereby obtaining the number of second candidate circumferential positions.

[0068] In equation (5.8) This indicates selecting the smallest candidate value that meets the conditions, with the number of candidates ranging from 2 to... Integers between 0 and 1, applicable to candidate sets containing a number of candidates that satisfy the residual rate condition. This implementation method... The corresponding candidates are 2, 3, 4, 5, and 6.

[0069] In step S220, based on the number of first candidate circumferential positions Number of second candidate circumferential positions Determine the final number of circumferential positions.

[0070] In one specific implementation, the final number of circumferential positions Determine according to formula (5.9): (5.9) by After being selected as candidate results, calculation And check if it is less than .different For different sets of divisibility orders, the residual rate does not necessarily change. Increase and decrease, take and Even after reaching a large value, further testing is still required. The step selection formula is applicable when there are threshold-crossing orders. And if the candidate set is not empty, only those that satisfy... and The candidate results are used as the output of this implementation.

[0071] Steps S100 to S200 obtain the angular harmonics based on the measurement position and thermal wave value of the reference wafer, and determine the first candidate quantity according to the maximum order exceeding the threshold. Steps S210 and S220 further filter the quantity using the residual rate, so that the final quantity satisfies the maximum order relationship and the residual rate condition.

[0072] The implementation method using the first candidate number The number of circumferential positions is one greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold. When further employing the residual rate criterion, the final number can be greater than... Using only the first criterion for step selection satisfies the relationship that the number of positions is one greater than the maximum order; using both criteria simultaneously satisfies the relationship that the number of positions is greater than the maximum order, and the final number can be two or more greater than the maximum order.

[0073] The reference wafer provides the location data and thermal values ​​required for step selection, and the wafer to be processed accepts the determined step-by-step implantation process. The first criterion controls the maximum order of significant harmonics, while the residual rate criterion further evaluates the angular components that have not been averaged out; the two criteria have different functions.

[0074] like Figure 3 As shown, reference wafer 200 has a wafer center. Multiple measurement locations on the reference wafer 200 are distributed at different radial distances and polar angles to obtain the corresponding thermal wave values.

[0075] Figure 3 In This indicates the number of segments after sorting by polar angle. Each measurement location corresponds to an original measurement index. The coordinates of the sorting position are... , The polar radius is The polar angle is .

[0076] Figure 3 Shown in the direction of increasing polar angle , and . It is half the sum of the spans of the two adjacent polar angles, that is, the angle width between the midpoints of the adjacent polar angles; the dashed line represents the angle width boundary and does not represent the solid partition on the wafer.

[0077] Figure 3 In , Represents the rectangular coordinate axes, Indicates the center of the wafer. and These represent the polar radius and polar angle of the sorting position, respectively. The measurement positions in the figure are only for illustrating the sampling relationship; the specific coordinates of the 52 points are listed in Table 1.

[0078] When the number of measurement locations exceeds 52, the same polar coordinate transformation, detrending, and angular analysis process can be used. The specific expression for the first and last weights must correspond to the sampling distribution and the cyclic angular interval. Equation (3.1) gives the calculation rules for the 52-point distribution in this embodiment.

[0079] like Figure 4 As shown, determine the number of circumferential positions. Then, the wafer to be processed 220 can be sequentially positioned relative to the electrostatic chuck 120. There are several circumferential positions, and ion implantation is performed at each of these circumferential positions. In one embodiment, one circumferential position corresponds to one implantation step, and one ion implantation is performed at... The injection step is complete.

[0080] In the implementation of the equiangular distribution, the angular interval between two adjacent circumferential positions is... The ratio to the number of circumferential positions. This angular interval is denoted as... , In degrees, the polar angle relative to the measurement position distinguish:

[0081] Using the initial circumferential position as a reference, arrange the pairs according to the circumferential direction. The circumferential position number, the first The angles of each circumferential position relative to the initial circumferential position are:

[0082] , Indicates the index of the circumferential position. The initial relative angle of the circumferential position is 0. This indicates the angular interval between positions; when injection is performed in a circumferentially adjacent order, the angle of rotation in the same direction between steps is also . .

[0083] In the embodiment where ion implantation is performed sequentially at adjacent circumferential positions, after ion implantation is completed at the first circumferential position, the wafer to be processed 220 rotates relative to the electrostatic chuck 120. Upon reaching the second circumferential position, the second implantation step is performed. The wafer 220 to be processed is sequentially rotated to subsequent circumferential positions until completion. One injection step.

[0084] The sum of the implantation doses at each circumferential position constitutes the total dose of one ion implantation. In an implementation where one implantation step corresponds to one circumferential position, Between each injection step After one relative rotation, the injection process ends after completing the last injection step.

[0085] In one specific embodiment, the number of circumferential positions is 3, and the angular interval between two adjacent circumferential positions is:

[0086] After ion implantation is completed at the first circumferential position, the wafer 220 to be processed rotates relative to the electrostatic chuck 120. Move to the second circumferential position and perform ion implantation; then rotate again relative to the electrostatic chuck 120 degrees. Ion implantation is then performed at the third circumferential position. The three implantation steps corresponding to the three circumferential positions together complete one ion implantation cycle.

[0087] like Figure 5 As shown, this embodiment also provides a method for preparing a silicon-on-insulator substrate.

[0088] In step S300, a reference donor wafer, a donor wafer, and a acceptor wafer are provided. The reference donor wafer is used to obtain the thermal data required for the selected step, the donor wafer accepts the selected step implantation process and is used for subsequent layer transfer, and the acceptor wafer is used to receive the top layer silicon and form a support substrate.

[0089] In step S320, multiple orders of angular harmonics are obtained based on the position data of multiple measurement locations on the reference donor wafer and the corresponding thermal wave values. The number of circumferential positions of the donor wafer relative to the electrostatic chuck during the implantation step is determined based on the amplitude of the angular harmonics. The number of circumferential positions is equal to the number of implantation steps and is greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold. In an embodiment using only the first criterion, the number of circumferential positions is 1 greater than the maximum order. In another embodiment using both criteria simultaneously, a final number is also determined based on the residual rate, and the final number may be greater than the first candidate number.

[0090] In step S340, the donor wafer is placed on an electrostatic chuck and hydrogen ion implantation is performed in at least two implantation steps to form an implantation layer. Between two adjacent implantation steps, the donor wafer is rotated circumferentially relative to the electrostatic chuck, so that the same wafer region is sequentially cooled by different cooling regions in different implantation steps. When a three-step implantation is used, the total dose is evenly distributed across three steps, and the donor wafer is implanted sequentially at three adjacent circumferential positions, with rotation relative to the electrostatic chuck between adjacent steps. .

[0091] In step S360, the donor wafer that has undergone hydrogen ion implantation is subjected to low-temperature hydrophilic bonding with the acceptor wafer, and the bonding temperature does not exceed [a certain value]. In one specific embodiment, the donor wafer and the acceptor wafer are bonded at room temperature, followed by a peeling heat treatment of the bonded wafer structure.

[0092] In step S380, the bonded wafer structure is subjected to a peeling heat treatment to peel the donor wafer along the implantation layer. In step S400, an insulator-on-silicon substrate with a top silicon layer is obtained on one side of the acceptor wafer.

[0093] Figure 5 This illustrates an implementation method for fabricating silicon-on-insulator substrates using thermal wave data selection. Between adjacent implantation steps, the donor wafer is rotated circumferentially relative to the electrostatic chuck, altering the correspondence between the same wafer region and the electrostatic chuck cooling region. The implantation layer formed through step-by-step implantation is used for subsequent lift-off.

[0094] The silicon-on-insulator substrate includes a supporting substrate, an insulating layer, and a top silicon layer, with the insulating layer located between the supporting substrate and the top silicon layer. In an embodiment where the donor wafer has a silicon dioxide layer on its implanted surface, the donor wafer is bonded to the acceptor wafer via the side containing the oxide layer. After layer transfer, the silicon dioxide layer constitutes the insulating layer, the acceptor wafer constitutes the supporting substrate, and the silicon layer retained by the transferred donor wafer constitutes the top silicon layer.

[0095] A step-selective SOI fabrication process using thermal wave data ensures that the number of implantations in the donor wafer corresponds to the angular spatial variation of the reference donor wafer. The donor wafer is bonded and then peeled off to form the top silicon layer. The thermal wave distribution after donor wafer implantation and the thickness distribution after layer transfer are evaluated through corresponding tests.

[0096] In a specific implementation of the three-step equal-dose implantation method, the thermal uniformity of the donor wafer after implantation is less than or equal to [a certain value]. The uniformity of the top silicon thickness after layer transfer is less than Thermal uniformity is the ratio of the sample standard deviation to the mean of thermal values ​​measured at multiple thermal measurement locations on the donor wafer after implantation, expressed as a percentage; top silicon thickness uniformity is the difference between the maximum and minimum thickness measured at multiple thickness measurement locations. These two metrics evaluate the donor wafer after implantation and the top silicon after layer transfer, respectively.

[0097] Comparative examples and embodiments To verify the effect of stepwise ion implantation on the spatial distribution of thermal waves on the donor wafer and the thickness distribution of the top silicon layer on the silicon-on-insulator substrate, comparative examples, Example 1, and Example 2 were set up.

[0098] Comparative thermal wave data were used as reference donor wafer data for selective step analysis. Both Examples 1 and 2 employed a three-step implantation process to evaluate the step-by-step implantation results and repeatability.

[0099] Comparative examples, Example 1, and Example 2 used samples with a diameter of 12 inches, a crystal orientation of (100), and a resistivity of [missing information]. A silicon wafer is used as a donor wafer, and a layer with a thickness of approximately [thickness missing] is formed on the surface of the donor wafer. The silicon dioxide layer. Ion implantation uses hydrogen ions (… ), injecting energy as The total injected dose is During ion implantation, the temperature of the donor wafer is controlled at room temperature. When multiple implantation steps are used, the total implantation dose is evenly distributed across each step. After ion implantation, thermal wave testing is performed at 52 measurement sites on the donor wafer.

[0100] In the three-step injection, the hydrogen ion injection dose at each step is: The three steps totaled as follows: The turning point between steps is... The scanning or regular turning performed within each step does not increase the number of relative circumferential positions between steps. .

[0101] Comparative Example The comparative example did not change the circumferential position of the reference donor wafer relative to the electrostatic chuck between implantation steps. The reference donor wafer maintained the same circumferential position relative to the electrostatic chuck during ion implantation. The number of circumferential positions for the comparative example was recorded as 1, representing a comparison process without inter-step relative rotation.

[0102] After ion implantation, a 52-point thermal imaging test was performed on the reference donor wafer. The comparative thermal imaging spatial distribution is shown below. Figure 6 As shown.

[0103] Substituting the 52 thermal wave data points from the comparative example in Table 1 into equations (1.1) to (5.9), the results are calculated. Approximately 0.205664, It is approximately 4107.080705. Use thermal wave reading units. Using thermal wave reading units / .

[0104] After sorting by polar angle stability and calculating the cyclic weights, the amplitudes of the 1st to 6th orders are approximately 68.247606, 16.317276, 12.423269, 12.464844, 7.202539, and 5.363338, respectively, all in thermal wave reading units. The empirical noise floor is approximately 9.833692, and the threshold... The value is approximately 29.501075, with the maximum order exceeding the threshold being 1, representing the number of first-choice candidates. The value is 2.

[0105] When the number of candidates is 2, 3, 4, 5, and 6, Approximately , , , and .satisfy The minimum number of candidates is 3, therefore , ,and Ultimately, a three-step injection method is adopted. This example uses a combined step selection based on two criteria, resulting in a number of three circumferential positions greater than the maximum order of 1; when only the first criterion is used, the number of positions corresponding to the same set of reference data is 2. Examples 1 and 2 use the three circumferential positions obtained through combined step selection.

[0106] The threshold comparisons and candidate selections in the calculations use unrounded values, and the intermediate results listed are retained to six decimal places. When the same dataset is converted to thermal wave reading units with the same direct ratio, the amplitude and empirical noise floor change proportionally, while the threshold order and residual rate remain unchanged.

[0107] Example 1 In Example 1, the donor wafer undergoes three implantation steps at three circumferential positions, which correspond to the initial orientation and rotation relative to the initial orientation, respectively. Orientation and rotation relative to the initial orientation The location.

[0108] A single ion implantation process consists of three implantation steps. After the first implantation step is completed, the donor wafer rotates relative to the electrostatic chuck. Then, the second implantation step is performed. After the second implantation step is completed, the donor wafer rotates again relative to the electrostatic chuck. Then, the third implantation step is performed. After the third implantation step is completed, the ion implantation is finished.

[0109] After ion implantation, thermal wave testing was performed at 52 measurement locations on the donor wafer. The spatial distribution of the thermal waves in Example 1 is as follows. Figure 7 As shown.

[0110] Example 2 Example 2 employed the same three-step ion implantation method as Example 1, serving as a repeatability experiment, with three implantation steps. After ion implantation, thermal wave testing was performed at 52 measurement locations on the donor wafer. The spatial distribution of the thermal waves in Example 2 is shown below. Figure 8 As shown.

[0111] The 52-point thermal wave test data for the comparative example, Example 1, and Example 2 are listed in Table 1.

[0112] Table 1. Thermal wave test data at 52 points for comparative examples, Example 1, and Example 2.

[0113] The thermal wave range is the difference between the maximum and minimum thermal wave values ​​at the measurement location, and the thermal wave standard deviation is the sample standard deviation. Thermal wave homogeneity is the ratio of the thermal wave standard deviation to the thermal wave mean multiplied by [missing value]. Top-layer silicon thickness uniformity is determined using the total thickness variation. The thickness is represented as the difference between the maximum and minimum thickness at the measurement location, expressed in units of... .

[0114] (6.1) (6.2) (6.3) (6.4) in, For extreme thermal wave, , These are the maximum and minimum values ​​of the thermal wave measurement. The standard deviation of the sample is 1. For the first One thermal wave measurement value, When calculating the sample standard deviation for the number of thermal wave measurement locations. Greater than 1. The average value of the heat wave is denoted by the line above it. ,equal The sum of all thermal wave measurements divided by The average value in this embodiment is positive; The uniformity of thermal waves is expressed as a percentage. and These are the maximum and minimum values ​​of the top silicon thickness measurement, in units of... The thermal range, standard deviation, and thermal value are measured in the same unit of measurement as the equipment readings.

[0115] After statistical analysis of the 52-point heat wave test results, the comparative heat wave range was 244.9, the standard deviation was 58.9, and the heat wave uniformity was [missing value]. In Example 1, the thermal wave range was 69.4, the thermal wave standard deviation was 19.8, and the thermal wave uniformity was [missing value]. In Example 2, the thermal wave range was 57.5, the thermal wave standard deviation was 15.5, and the thermal wave uniformity was [missing value]. .

[0116] Table 2. Top silicon thickness test data at 49 points for comparative examples, Example 1, and Example 2.

[0117] The thermal wave range, thermal wave standard deviation, and thermal wave uniformity of Examples 1 and 2 are all lower than those of the comparative example. The thermal wave test results of the comparative example and examples show that the spatial distribution of the thermal wave of the donor wafer changes after changing the circumferential position of the donor wafer relative to the electrostatic chuck between adjacent implantation steps.

[0118] After completing the thermal wave test, the donor wafers corresponding to the comparative example, Example 1, and Example 2 were subjected to low-temperature hydrophilic bonding, with the bonding temperature not exceeding [a certain value]. For example, room temperature bonding is used. This is followed by a peeling heat treatment. In one specific embodiment, the peeling heat treatment is performed in an argon atmosphere at a temperature of [temperature missing]. The heat preservation time is After the stripping is completed, a silicon-on-insulator substrate is obtained.

[0119] The top silicon thickness of a silicon-on-insulator substrate was measured at 49 points using an ellipsometry. A comparative spatial distribution of the top silicon thickness is shown below. Figure 9 As shown, the spatial distribution of the top silicon thickness in Example 1 is as follows: Figure 10 As shown, the spatial distribution of the top silicon thickness in Example 2 is as follows: Figure 11 As shown.

[0120] Figures 6 to 8 The results of thermal wave spatial distribution for the comparative example, Example 1, and Example 2 are shown respectively. Figures 9 to 11 The spatial distribution of top silicon thickness is shown for the comparative example, Example 1, and Example 2, respectively. (Figures are incomplete and require further context.) , The plotting coordinates are used to represent positions within the wafer plane. Grayscale areas and isopleths are used to represent the spatial variation trends of the corresponding test results. The color scale ranges of different plots can be set according to the display needs of the corresponding test data; therefore, quantitative comparisons between different plots are not made directly based on grayscale depth.

[0121] Figures 6 to 11 The isopleth regions in the table are interpolated based on the test data and are only used to assist in representing spatial distribution trends. They do not represent the physical structure on the wafer and do not constitute new measurement results. The specific coordinates and original measurement values ​​of each measurement location are based on the data listed in Tables 1 and 2. The thermal wave range, thermal wave standard deviation, thermal wave uniformity, and top silicon thickness uniformity are all calculated directly from the original measurement values ​​in Tables 1 and 2.

[0122] The thickness range of the top silicon layer in the comparative example is: Table 3 is presented in two decimal places. Example 1 and Example 2 are respectively and All are lower than the comparative ratio and meet the condition of being less than [the specified value]. The indicators.

[0123] The main process conditions and test results of the comparative examples, Example 1 and Example 2 are summarized in Table 3.

[0124] Table 3. Main process conditions and test results for comparative examples, Example 1, and Example 2

[0125] Table 3 shows that, under the process conditions of this embodiment, the thermal wave range, thermal wave sample standard deviation, thermal wave uniformity, and top silicon thickness range of Examples 1 and 2 using the three-step implantation method are all lower than those of the comparative example, and the results of the two groups show a consistent improvement trend. The test results indicate that changing the circumferential position of the donor wafer relative to the electrostatic chuck between adjacent implantation steps helps to improve the thermal wave uniformity after implantation and the top silicon thickness uniformity after layer transfer.

[0126] This application illustrates the technical solution through specific embodiments. The specific process conditions, parameter values, and test results in the embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art can make corresponding adjustments to the implementation methods without departing from the technical concept of this application, or combine technical features from different implementation methods where they are technically compatible. The scope of protection of this application is determined by the content of the claims.

Claims

1. An ion implantation method, characterized in that, include: The wafer to be processed is placed on the electrostatic chuck of the ion implanter; The wafer to be processed is ion implanted using an ion beam in at least two implantation steps; Between two adjacent injection steps, the wafer to be processed is rotated circumferentially relative to the electrostatic chuck, so that the same area of ​​the wafer to be processed is successively cooled by different cooling areas of the electrostatic chuck in different injection steps.

2. The ion implantation method according to claim 1, characterized in that, The ion implantation method further includes: Acquire position data for multiple measurement locations on a reference wafer and the corresponding thermal wave values ​​for those multiple measurement locations; Multiple orders of angular harmonics are obtained based on the location data and the thermal wave value; Based on the amplitude of the multiple orders of angular harmonics, the number of circumferential positions of the wafer to be processed relative to the electrostatic chuck in the implantation step is determined. The number of circumferential positions is equal to the number of implantation steps and is greater than the largest order of the angular harmonics whose amplitude exceeds a preset threshold among the multiple orders of angular harmonics.

3. The ion implantation method according to claim 2, characterized in that, The number of circumferential positions is 1 greater than the largest order of the angular harmonics whose amplitude exceeds the preset threshold among the multiple orders.

4. The ion implantation method according to claim 2, characterized in that, The angular interval between two adjacent circumferential positions is The ratio of the number of circumferential positions to the number of positions.

5. The ion implantation method according to claim 1, characterized in that, During the injection step, the wafer to be processed and the electrostatic chuck maintain a constant relative position.

6. The ion implantation method according to claim 1, characterized in that, The number of circumferential positions of the wafer to be processed relative to the electrostatic chuck during the implantation step is 3.

7. A method for preparing a silicon-on-insulator substrate, characterized in that, include: Provide donor wafers; Using the donor wafer as the wafer to be processed, hydrogen ion implantation is performed on the donor wafer using the ion implantation method described in claim 1 to form an implantation layer; the donor wafer is then bonded to the acceptor wafer. as well as The bonded structure is heat-treated to peel off the donor wafer along the implantation layer and form a top silicon layer on the acceptor wafer to obtain a silicon-on-insulator substrate.

8. The method for preparing a silicon-on-insulator substrate according to claim 7, characterized in that, The preparation method further includes: Acquire position data of multiple measurement locations on the reference donor wafer and the thermal wave values ​​corresponding to the multiple measurement locations; Multiple orders of angular harmonics are obtained based on the location data and the thermal wave value; Based on the amplitude of the multiple orders of angular harmonics, the number of circumferential positions of the donor wafer relative to the electrostatic chuck in the implantation step is determined. The number of circumferential positions is equal to the number of implantation steps and is 1 greater than the largest order of the angular harmonics whose amplitude exceeds a preset threshold among the multiple orders of angular harmonics.

9. The method for preparing a silicon-on-insulator substrate according to claim 8, characterized in that, The donor wafer, during the implantation step, comprises at least two circumferential positions relative to the electrostatic chuck, with an angular interval between two adjacent circumferential positions. The ratio of the number of circumferential positions to the number of positions.

10. The method for preparing a silicon-on-insulator substrate according to claim 7, characterized in that, The bonding between the donor wafer and the acceptor wafer is a low-temperature hydrophilic bonding.

11. A silicon-on-insulator substrate, characterized in that, It includes a supporting substrate, an insulating layer, and a top silicon layer, wherein the insulating layer is located between the supporting substrate and the top silicon layer, wherein, The top silicon layer is formed from a donor wafer through hydrogen ion implantation, bonding, and stripping. The hydrogen ion implantation is performed in at least two implantation steps, during which the donor wafer is supported on an electrostatic chuck. Between two adjacent implantation steps, the donor wafer is rotated circumferentially relative to the electrostatic chuck, so that the same area of ​​the donor wafer is successively cooled by different cooling areas of the electrostatic chuck in different implantation steps. The number of circumferential positions of the donor wafer relative to the electrostatic chuck in the implantation step is determined based on the amplitude of multiple orders of angular harmonics, which are obtained based on position data of multiple measurement positions of the reference donor wafer and the thermal wave values ​​corresponding to the multiple measurement positions. The number of circumferential positions is greater than the maximum order of the angular harmonics whose amplitude exceeds a preset threshold among the multiple orders of angular harmonics; The thermal uniformity of the donor wafer after hydrogen ion implantation is less than or equal to The thermal wave uniformity is the ratio, expressed as a percentage, of the sample standard deviation to the mean of the thermal wave values ​​measured at multiple thermal wave measurement locations after the donor wafer has completed the hydrogen ion implantation; and The thickness uniformity of the top silicon layer is less than The thickness uniformity is the difference between the maximum and minimum thicknesses measured at multiple thickness measurement locations on the top silicon layer.

Images