TW201316389A - Method for correcting positional deviation on a first wafer bonded to a second wafer - Google Patents

Abstract

The invention relates to a method for correcting misalignment of positions on a first wafer bonded to a second wafer, the method comprising the application (F4), to coordinates of each position, of a predetermined correction function for the said first wafer, the correction applied by the correction function being a function only of the distance of the position relative to the centre of the first wafer, characterized in that the applied correction varies, over the whole of the first wafer, in a non-linear manner with respect to the distance of the position relative to the centre.

Description

Method for correcting positional deviation on a first wafer bonded to a second wafer

The present invention relates to the field of producing semiconductor wafers or substrates comprising at least two overlapping microcomponent layers.

The stacking of several consecutive component layers (" 3D IC stack") and the resulting multilayer wafers are particularly useful in the fabrication of sensors for backside illuminated image sensors (often referred to as BSI sensors).

The wafers are produced by bonding a donor substrate to a receptor substrate that allows a portion of the wafer forming one of the donor substrates to be transferred to the acceptor substrate. The transferred layer can include electronic or optoelectronic microcomponents (which can be active or passive). The acceptor substrate can also include such components.

It is often necessary to process (e.g., to form) other micro-components, provide surface access to existing micro-components, or achieve interconnections after the transfer of layers. These processes are basically implemented by photolithography.

In this context, the present invention addresses the problems caused by uneven deformation occurring during the transfer of the layer during this subsequent processing step.

These non-uniform deformations have actually been observed in a general manner during the process of three-dimensional integration of components and especially during installation of components by photolithography.

In particular, it has been found that after the transfer of one layer, it is difficult to form additional micro-components that are satisfactorily aligned, parallel to the plane of the wafer, wherein the micro-components are present in the layer or in a lower layer.

This positional deviation (or "stacked") phenomenon with respect to Figures 1A through 1E is illustrated, which illustrates an exemplary embodiment of a three dimensional structure. This example includes transferring a microcomponent layer formed on a donor substrate to an acceptor substrate; and forming an additional microcomponent layer on the exposed side of the layer.

FIG. 1A illustrates a donor substrate 10 on which a first series of microcomponents 11 are formed. The micro-assembly 11 is formed, for example, by using light lithography that is exposed through one of the first masks. 1B (which is a cross-section of the substrate of FIG. 1A in plane P) shows that the micro-assembly 11 is formed in a thin surface layer of the substrate 10 and is flush with the surface of the latter.

As illustrated in FIG. 1C, the face of the donor substrate 10 including the micro-assembly 11 is then placed in face contact with one of the receptor substrates 20. The bonding between the substrate 10 and the substrate 20 is, for example, carried out by molecular attachment. This gives a layer of buried micro-assembly 11 at the bonding interface between the substrates. After the bonding and as shown in FIG. 1D, the donor substrate 10 is thinned to remove a portion of the material present on the layer of the micro-assembly 11 such that only a thin layer of the donor substrate 10 remains on the substrate 20. 10a.

As shown in FIG. 1E, a subsequent step is to form a second layer of micro-components 12 at the exposed surface of layer 10a in alignment with the components contained in layer 10a, or to perform additional on the exposed surface. Technical steps (replacement of contacts, replacement of interconnects, etc.).

To form the micro-assembly 12 aligned with the embedded micro-assembly 11 (or to perform additional technical steps), a second photolithographic mask obtained from the mask used to form the micro-assembly 11 is used.

However, if the second mask has been directly defined based on the first mask, an offset in the plane of the substrate 20 and the layer 10a is found between some of the micro-components 11 and 12, as indicated in FIG. 1E. △11, △22, △33, △44 (corresponding to several micro-component pairs 11 ₁ /12 ₁ , 11 ₂ /12 ₂ , 11 ₃ /12 ₃ and 11 ₄ /12 ₄ respectively in a plane parallel to the layer The observed offset between). This positional deviation ("stacking") phenomenon between the two layers of micro-components 11 and 12 can be a cause of a short circuit, loss of contact, distortion in the stack, or connection failure between the micro-components of the two layers.

The overall positional deviation of components 11 and 12 is not due to a simple transformation (obtained by a combination of translation and rotation) which may be the result of an inaccurate assembly of one of substrates 10 and 20 during the bonding step. This overall positional deviation is caused by localized uneven deformation occurring in the layer 10a during assembly of the donor substrate 10 to the acceptor substrate 20.

In particular, once bonded to the acceptor substrate 20, the donor substrate 10 has a geometry (curvature and warpage) that is different from the geometry it has at the beginning. This new geometry is specifically due to the fact that the acceptor substrate 20 also has a particular geometry (including curvature and warpage) that is different from the geometry of the donor substrate 10. Thus, when the substrate 10 is placed in contact with the receptor substrate 20, the two substrates must be adapted to each other's geometry, which creates a stress region in each of the substrates.

Therefore, if no corrective action is taken, there is a considerable risk that certain components (as shown in FIG. 1E) of the micro-assembly 12 formed on the exposed surface of the substrate after transfer are presented with the corresponding micro-assembly 11 It is offset by a position of about several hundred nanometers or even one micrometer.

Therefore, when the transferred micro-component is formed by a pixel imager and transferred When the processing steps are designed to form a color filter on each of these pixels, one of the coloring functions of some of the pixels has been observed to be lost.

One solution for overcoming this difficulty is to use a positional offset correction algorithm that makes it possible to correct the exposure of the photolithographic mask to effect proper etching of one of the series of components 12.

This positional offset correction algorithm is based on the observation of the movement of the markers present thereon prior to the transition of layer 10a. This practice selects such markers around a given circle or contour of one of the centers of the wafer. These marks are elements of the first layer of micro-components 11 that are readily identifiable, or indicators that are specifically marked on the wafer for alignment purposes.

Prior to performing photolithographic exposure and etching of the series of components 12, such marks are identified and their maintained movement is determined. This set of movements is then transferred to a single mathematical model for the entire wafer.

The single mathematical model is then used to determine the aligned position of the photolithography of the entire wafer, applying an alignment to each component 12 in a systematic and identical manner. This makes it possible to locate and calibrate the photolithography tool more than once per wafer to be processed. In particular, a solution to realign the lithography tools for different areas of the wafer to be processed will not be satisfactory from an industry perspective, as it will impose a heavy emphasis on the user. Constraint (time, manipulation).

The single mathematical model specifically includes a correction function applied to one of the coordinates (X, Y) of one of the positions to be aligned, which coordinates have typically been associated with a previously transformed object in the context of the mathematical model. The correction determined by the function of one position to be aligned takes a motion vector (Xc, Yc) added to one of the coordinates (X, Y) form. This correction is usually the last step by alignment of a single mathematical model.

This practice uses this correction function, which depends only on the distance from the center of the position to be calibrated and is linear with respect to this distance. In addition, the correction value is selected to be zero at the center of the wafer.

The components of the correction function on the two axes are caused by the residual positional deviation of the series of markers selected on a given circle around one of the centers on each of the equiaxions (i.e., after the application of the correction function) The average of the values is minimized and predetermined.

This empirical approach has yielded satisfactory results, but as technology advances, components such as components 11 and 12 are reduced in size, which requires finer alignment. In addition, the diameter of the proposed wafer is increased. The result of the above is that it is difficult to increase the risk by one of the different residual positional deviations on various regions of the wafer (especially the perimeter) corrected for a single mathematical model according to one of the prior art.

Finally, certain bonding methods (specifically, combinations of molecular bonding types) cause specific distortions that are only partially corrected by existing positional deviation correction algorithms, as will be explained in the remainder of the description. It is an object of the present invention to address such problems and thus improve the alignment of the microcomponents.

In this context, a subject of the present invention is a method for correcting a positional deviation coupled to a first wafer on a second wafer, the method comprising: using one of the first wafers The predetermined correction function is applied to the coordinates of each position, and the correction applied by the correction function only follows the position relative to The distance from the center of the first wafer varies, the method being characterized in that the applied correction varies in a non-linear manner over the entire first wafer relative to the distance of the location relative to the center.

With this method, the corrections performed are of better quality than the corrections obtained with the previous method, and allow for more precise alignment for one of a larger number of micro-components while providing greater speed to the manufacturer in the alignment procedure. This is because no recalibration is required and the correction function is predetermined for the wafer.

According to a particularly advantageous embodiment, the positional deviation correction method additionally comprises determining that at least one of the first series of marks present at a first distance from the center and one of the second from the center The previous step of the correction function is changed by the positional deviation value observed on the first wafer of the second series of the distance.

This feature makes it possible to correlate the correction function with the positional deviation values of the marks at two different distances from the center of the wafer and thus obtain a good model of one of these positional deviation values by a non-linear function.

In one embodiment, the correction function is linear with respect to the distance of the position relative to the center, a central region defining a first segment and a peripheral region defining a second segment. In this embodiment, two different linear functions on the two segments are selected to faithfully model the positional deviation values to finely correct them.

In another embodiment, the correction function is defined in a non-linear manner over the entire first metal plate in a single manner with respect to the distance of the position relative to the center. This choice of unequally defining a function in a single way provides an added simplification of the information implementation. Perhaps an at least two polynomial As a correction function, the position deviation value is modeled as faithfully to finely correct it.

According to one feature of the invention, the applied correction increases with distance from the center, which makes it possible to consider the major positional deviations occurring on the periphery of the wafer.

According to another feature of the invention, the applied correction is increased to a greater extent with respect to the distance from the center over a peripheral region of one of the first wafers, which makes it possible Also consider the main positional deviations that occur around the wafer.

As used in the introduction, it is meant that the positional deviation in the periphery of the wafer of this size is particularly large, when the method is applied to a wafer having a diameter of more than 200 mm or a first crystal. It is especially advantageous when it is round.

This method is also particularly advantageous if the bond is a molecular bond, as it makes it possible to correct for specific positional deviations that occur in the structure that has been combined in this way.

It should be noted that, in accordance with a feature of the invention, the correction function is added to the coordinates of the position, and furthermore applied to the coordinates of the position obtained after linear transformation of one of the coordinates is applied to the initial coordinates of the position.

According to a second subject matter, the present invention is also directed to a method for producing a three-dimensional composite structure comprising: a step of producing a first micro-component layer on one side of a first substrate; the micro-component will be included a step of bonding the face of the first substrate to a second substrate; and generating a second component on the face of the first substrate facing the first micro-component layer One step of the layer, the method is characterized by implementing the second The steps of the layer simultaneously correct the positional deviation using a method as discussed above.

Another subject of the invention is a three-dimensional composite structure obtained according to one of the embodiments just described above. The invention is also directed to a backside illuminated image sensor comprising such a three dimensional composite structure.

Finally, the present invention is also directed to an apparatus for correcting a positional deviation coupled to a first wafer of a second wafer, the apparatus comprising for applying a predetermined correction function for the first wafer The component of the coordinate at each location, the correction applied by the correction function varies only with the distance of the location relative to the center of the first wafer, the device being characterized in that the applied correction is over the entire first wafer The distance from the center relative to the center varies in a non-linear manner.

This device provides the advantages specified by the positional deviation correction method that is the subject of the present invention.

The invention will now be described in relation to the figures.

The invention will be elucidated below with respect to one of the embodiments given as an illustration.

Referring to Figure 2, a semiconductor wafer 200 is shown in a top view. This wafer is produced by a stacking and transfer procedure as described in the introduction. It has a circular geometry centered on a point X.

Figures 2 through 6 present information relating to this wafer and to positional offset corrections obtained by a known algorithm. Such information will make it possible to discuss the positional deviation correction method of the present invention as shown in Figures 7-12 advantage.

Alignment marks 201 to 204 exist on the surface of the wafer 200. In this example, as an illustration, there are four markers. They are all at the same distance from the center X. Other tags can already be displayed. In practice, such indicia may have been specifically placed or may be components of the microcomponents that are individually viewable on the wafer for the purpose of acting as an alignment indicator. The positional offset correction algorithm according to the prior art uses such marks placed at one and the same distance relative to the center of the wafer.

To determine the quality of the alignment obtained by a known positional deviation correction algorithm, the residual movement has been determined point by point for the wafer 200. "Residual movement" in this example means the movement that is not considered by this algorithm, which is determined between the position established by a given position and the position of the marker estimated by the alignment algorithm for a given position. difference.

It can be seen that a central region 210 extends over two-thirds of the diameter of the wafer 200 and is low for its residual movement value. A circular area 220 is visible around the central area 210, and conversely, the residual movement is extremely large on the circular area.

To understand this distribution of residual positional deviation values, a detailed study has been carried out.

The experimentally measured, uncorrected radial movement values along a radius similar to one of the wafers of wafer 200 are shown in FIG. In this example, the wafer has a wafer of 100 mm radius (or diameter of 200 mm).

It should be noted that in a central region, the value of the movement is relatively constant and close to zero (in any case less than 50 nm or even 25 nm), but beyond the center At a distance of 80 mm, these movement values increase rapidly to reach a value of approximately 100 nm.

Referring to Figure 4, the figure is shown in the form of a vector experimentally measured movement on a surface of a wafer 400 having a diameter of one of 300 mm. Note that in this figure, there is a large intensity motion vector in one of the regions 420 near the periphery of the wafer. Conversely, the motion vector in one of the central regions 410 of the wafer has a significantly shorter length.

Referring to FIG. 5, there is a change in the gravity effect during the bonding and transfer of a thin layer (as shown in FIGS. 1B to 1D) along the radius of one of the wafers undergoing a bonding. One of the absolute values of the movement in the plane of the point of the wafer is simulated. The two wafers are horizontally positioned during assembly, which is important due to decisive gravity effects.

This simulation of the value of the movement in the plane is shown in the form of curve 510, which shows the position along the diameter of one of the wafers on the X-axis and the intensity of the movement on the Y-axis. Note that there is a sudden increase in the position of moving away from the center of the wafer compared to the position at the center of the wafer, after a turning point. Curve 520 also does not allow for the simulation of the phenomenon by simply exhibiting a linear regression with respect to the distance from the center of the wafer.

6 shows an analog wafer (element symbol 540) as specified in FIG. 5 in accordance with the method of the prior art, indicating that the value for its residual movement is first linearly transformed with respect to one of the coordinates and then relative to the wafer. One of the distances between the centers is the point at which a linear correction C is applied after a continuous application exceeds a certain threshold. Note that for the value of its residual movement, one of the tiny central areas occupies the wafer. The center of the two-thirds and the larger of the values of the residual movement form a peripheral halo having a radius of about 12 cm.

Also shown in this figure is the value of the residual motion simulated as a function of the distance from the center of the wafer in the form of a curve 550. It is possible to identify a central region that is small for its residual movement and a peripheral radial region with a high value.

Note that this distribution is identical to the one found experimentally and shown in Figure 2. Therefore, it seems that the experimental data is excellently reproduced to simulate such movements and in particular based on the gravity effect during molecular bonding.

It should be noted that this result was obtained by means of two different simulations, one of which was related to the situation in which the donor substrate was below the receptor substrate during the bonding step and the other was related to the opposite situation.

Based on these findings, an alignment algorithm in accordance with the present invention has been developed to achieve high quality alignment for large size wafers (having one layer transferred by a molecular bond).

Referring to Figure 7, the determination of an alignment algorithm is implemented in the following manner. This method is applied to a circular semiconductor wafer having one of the objects on its surface and having to be transferred.

First, two contours are selected, for example, a circle centered on the center X of the wafer and having respective diameters D1 and D2. Six to ten markers are selected along each of the two circles during a step E1. If the microcomponents present on the surface of the wafer are sufficiently visible, they can act as markers. It is also possible to place markers specifically for alignment purposes.

During a step E2 (which may be similar to the step of Figure 1C), the implementation layer Transfer. This step E2 may comprise a binding step (especially molecular bonding) and a thinning step.

The following steps (E3 to E5) may be carried out after the steps E1 and E2 in the same working unit or conversely after the wafer is delivered to another working unit. These steps are therefore shown in Figure 7 and in Figures 8 and 9, and are depicted in dashed lines to indicate alternatives.

During a step E3, the movement of each of the markers is measured along a two axis in the plane in the form of a vector (X, Y).

A minimum algorithm is then used during a step E4 to calculate each of the markers to be applied to minimize the positional deviation on all of the markers, including at least one translation and one rotation transformation T. Marks having a circle of diameters D1 and D2, for example, are treated with minimization of the same importance (same weight).

It is also possible at this stage to compute a one-bit transform and an orthogonality vector, the equidistant transform and the orthogonality vector combined with the translation and rotation making it possible to use a single transformation of all the movements for the marking of the light lithography. T gets the most likely approximation.

Thus a transform T consisting of a translation, a rotation, an amplification, and an application of an orthogonality vector acts in a linear manner on one of the coordinates of the positions. It is determined to be system one at all locations of the wafer.

During a step E5, for each of the markers having the circle of diameter D1, an intermediate residual movement is calculated based on the coordinates transformed by the transformation T determined during step E4. This intermediate residual movement is thus defined as the difference between the true movement and the correction by the transformation T. It consists of a vector (x, y) composition. Similarly, during step E5, an intermediate residual movement (x, y) is calculated for each of the markers having a circle of diameter D2.

An average of one of the intermediate residual movements on the circle having the diameter D1 is then calculated for each of the marks along the circle having the diameter D1 in the form of one of the average residual movement vectors (Xa, Ya). An average of the intermediate residual alignment is also calculated for all marks along a circle having a diameter D2.

A correction function C is then determined to apply to the coordinates originating from the position of the transform T. This correction function C depends only on the distance of the position to be corrected relative to the center X. For a circle having a diameter D1 and for a circle having a diameter D2, it is determined to minimize the average of the residual movements obtained after its application.

The correction value forms a vector (Xc, Yc) added to the coordinates derived from the previous step.

In a first variation, the correction function is a linear function of one of the distances from the center piece by piece, a first segment consisting of a central region of the wafer, and a second segment consisting of one of the wafer perimeters Regional composition. In this variation, the central region includes indicia placed on a circle having a diameter D1, and the peripheral region includes indicia placed on a circle having a diameter D2.

The correction function is then determined by selecting two affine functions (one affine function per region) that make it possible to eliminate the mean of the intermediate residual movement along the diameters of the diameters D1 and D2. . In addition, the function applied to the central region of the wafer is selected to be zero at the center of the wafer, while the second function is selected at the junction between the two regions to be equal to the first function. In most cases the correction function is added on two segments, where There is a steeper slope on the second segment.

In another variation, the correction function is uniquely defined over the entire length of the radius of the wafer (with or without corners). Therefore, it can be a second order or a higher order one polynomial, or a more complex function. It is then determined by selecting at least two parameters to minimize the average of the intermediate residual movements calculated in the previous step.

In one embodiment, the circles having diameters D1 and D2 are treated with the same weight but it is also possible to assign different weights to the circles.

The characteristics of the conversion T determined in the step E4 and the correction C determined in the step E5 are stored, for example, by being stored on a computer medium.

Referring to Figure 8, a photolithography method is implemented in a work cell in which a wafer has been delivered to one of its specialized work cells or in which wafer fabrication takes place.

At this stage it is noted that it is advantageous to implement steps E3 to E5 just prior to exposing the wafer by photolithography, as shown in Figure 8, although as mentioned above, it can be implemented at the time of wafer fabrication.

During a step F1, the parameters of the transformation T determined in step E4 and the correction C determined in step E5 are placed in the memory of a photolithography apparatus.

Then, transform T and correction C are applied to each position of the object to be processed to form a component on the surface of the wafer. According to a repetitive process F2-F6, step F2 begins processing a position, and step F5 The process ends with processing a location. Step F3 applies the transformation T to the coordinates of the processed position. Step F4 will be corrected for the distance C with respect to the center and then applied to the coordinates derived from step F3. Then step F6 is followed by the previous step The actual photolithography is implemented in the position where the corrected coordinates have been obtained. Step F6 can also be included as a variant in the repetitive process.

The function of transform T and correction C is determined in a general manner to perform a single initial alignment prior to the step of exposing the wafer by photolithography. Therefore, the photolithography apparatus does not need to move for each area of the slice to be processed. In addition, the alignment obtained is particularly good, including a 300 mm wafer for carrying one of the layers transferred by a molecular bonding method.

Figure 9 shows an alternative to the method of Figure 7. After setting the conversion determined in step E4 and one of the parameters of the correction determined in step E5, step F'1, in the memory of a photolithography apparatus, the transformation T is applied to the desired one during step F'3. All positions are processed, and then, during a step F'4, the correction C is applied to the coordinates derived from all positions of step F'3. Step F'5 of photolithography is then performed for all locations.

10 shows a wafer 600 on which marks 651, 652 and 661 through 664 have been shown. A first group of marks (651 and 652) is located on a circle having a radius D1 centered on the center X while A second group of indicia (661 to 664) is located on a circle having a radius D2 centered on the center X. In the example shown, the distance D2 is approximately one-half the distance D1. The mark of the first group is located in a peripheral area 620 and the mark of the second group is located in a central area 610.

Referring to Figure 11, the correction C consists of one of linear corrections on the central region 610 of the wafer and one of the different linear corrections on the peripheral region 620 of the wafer. In the illustrated embodiment, each of the corrections on the two regions is linear over the area in question, but the entire correction C is on the entire wafer. Nonlinear.

Thus, the correction function is linear with respect to the distance from the center, the central region 610 defines a first segment and the peripheral region 620 defines a second segment. This correction is larger on the peripheral area. Reference 800 shows the difference between the corrections in the peripheral region and the central region, and also shows the curve 550 of the residual movement previously explained with reference to FIG. All piecewise linear corrections as a function of distance from the center make it possible to satisfactorily approach the true movement and thus provide one of the best quality alignments.

Alternatively, the correction function C can be defined in a non-linear manner on all wafers in a single manner but with respect to the distance from the center. As mentioned previously, it may involve an at least quadratic polynomial, or a more complex function.

Furthermore, the correction function according to a variant of the invention can be determined without the use of indicia placed on two different circles. Regarding one of its parameters being a non-linear function (for example a second-order polynomial) determined by another criterion or selected as fixed, only a series of marks at one and the same distance can be used. Thus, according to this variation, the correction is less closely adapted to the particular positional deviation of the wafer, but as such it makes it possible to consider the non-linear characteristic of the residual positional deviation value.

Finally, Figure 12 shows a method for producing a three-dimensional composite structure using a positional offset correction method in accordance with the present invention.

The method begins by forming a first series of micro-components on a starting substrate or donor substrate, step G1 (as shown in Figures 1A and 1B). The micro-components of the first series form a micro-component layer. This formation step is followed by The donor substrate is bonded to one of the receptor substrates, step G2 (as shown in Figure 1C), and in some variations, followed by a thinning step G3 (Fig. 1D).

The method continues with one of the steps of determining alignment parameters as shown in steps E3 through E5 of Figures 7-9. Therefore, the parameters to be applied to determine the photolithographic mask are known at this stage. These parameters specifically define a linear transformation T for one of the coordinates and a correction C along with the distance from the center, the correction C varying in a non-linear manner over the wafer relative to the distance from the center.

Once such parameters are determined, the method continues with a second series of micro-components forming a photolithography step of a component layer. By having determined that one of the entire wafers has been calibrated before starting the exposure (as shown in Figure 9 (steps F'3 to F'5)), or by individually correcting the exposure just prior to performing the exposure. This photo lithography is implemented in position (as shown in Figure 8 (steps F3 to F5)). In both cases, the correction is made by means of the parameters determined during steps E3 to E5.

The formation of an application or other component layer that can be processed by a method for making a three-dimensional composite structure continues. The finished product can be specifically a BSI sensor.

In summary, the present invention greatly reduces residual positional deviation and improves the performance of the resulting three-dimensional structure. The invention is not limited to the embodiments described and extends to the alternatives conceivable to those skilled in the art in the context of the scope of the claims.

10‧‧‧body substrate/substrate/first substrate

10a‧‧‧thin/layer/first wafer/first metal plate

11‧‧‧Microcomponents/components/microcomponent layers/first microcomponent layer

11 ₁ ‧‧‧Microcomponents

11 ₂ ‧‧‧Microcomponents

11 ₃ ‧‧‧Microcomponents

11 ₄ ‧‧‧Microcomponents

12 ₁ ‧‧‧Microcomponents

12 ₂ ‧‧‧Microcomponents

12 ₃ ‧‧‧Microcomponents

12 ₄ ‧‧‧Microcomponents

20‧‧‧Receiver substrate/substrate/second wafer/second substrate/second micro-component layer

200‧‧‧Semiconductor Wafer/Wafer

201‧‧‧ alignment mark

202‧‧‧ alignment mark

203‧‧‧ alignment mark

204‧‧‧Alignment marks

210‧‧‧Central area

220‧‧‧Circular area

510‧‧‧ Curve

520‧‧‧ Curve

540‧‧‧Simulated Wafer

550‧‧‧ Curve

600‧‧‧ wafer

610‧‧‧Central area

620‧‧‧ surrounding area

651‧‧‧ mark

652‧‧‧ mark

661‧‧‧ mark

662‧‧‧ mark

663‧‧‧ mark

664‧‧‧ mark

800‧‧‧Difference between corrections in the surrounding area and the central area

P‧‧‧ plane

△11‧‧‧Offset

△22‧‧‧Offset

△33‧‧‧Offset

△44‧‧‧Offset

1A-1E show a third of a semiconductor component according to the prior art The generation of dimensional structure.

2 shows experimental values of residual positional deviations on a wafer after application of an alignment algorithm according to one of the prior art.

3 and 4 respectively show the average value and position deviation vector of the radial position deviation of a wafer after the bonding and transfer of one layer.

Figure 5 shows a simulation of one of the radial positional deviation values modeled according to one of the effects of gravity.

Figure 6 shows a simulation based on one of the final residual position deviation values modeled by the alignment algorithm according to one of the prior art.

Figure 7 shows a method for preparing a wafer in accordance with the present invention.

Figures 8 and 9 show two photolithography methods using a positional offset correction method in accordance with the present invention.

Figure 10 shows experimental values of residual positional deviations on a wafer after application of a positional deviation correction algorithm in accordance with the present invention.

Figure 11 illustrates a positional deviation correction algorithm in accordance with a particular embodiment of the present invention.

Figure 12 shows a method for producing a three-dimensional composite structure using a positional deviation correction method in accordance with the present invention.

Claims (16)

A method for correcting a positional deviation coupled to a first wafer (10a) on a second wafer (20), the method comprising: using a predetermined correction for the first wafer (10a) The function (C) applies (F4, F'4) coordinates at each position, and the correction applied by the correction function varies only with the distance of the position relative to the center (X) of the first wafer (10a). The method is characterized in that the applied correction varies over the entire first wafer (10a) with respect to the distance of the position relative to the center (X) in a non-linear manner. The method of claim 1 for correcting a positional deviation, further comprising determining (E5) at least with respect to the first series of marks (651, 652) respectively present at a first distance from one of the centers (X) And a correction function that is observed at a position deviation value observed on the first wafer (10a) of the second series of marks (661 to 664) at a second distance from the center (X) ( C) One of the previous steps. A method for correcting a positional deviation according to claim 1 or claim 2, wherein the correction function (C) is linear with respect to the distance of the position relative to the center (X), and a central region (610) is defined A first segment and a peripheral region (620) define a second segment. A method for requesting a positional deviation according to claim 1 or claim 2, wherein the correction function (C) is in a single manner over the entire first metal plate (10a) with respect to the position relative to the center (X) This distance is defined in a non-linear manner. As with the method of claim 1 or claim 2 for correcting the positional deviation, the applied correction is increased with respect to the distance relative to the center. If the method of claim 1 or claim 2 is used to correct the positional deviation, the correction applied thereto is on the peripheral area (620) of the first wafer (10a) than on the first wafer (10a). One of the central regions (610) increases to a greater extent with respect to the distance from the center. A method for correcting a positional deviation as claimed in claim 1 or claim 2, which is applied to a first wafer (10a) having a diameter equal to one of 200 mm. A method for correcting a positional deviation as claimed in claim 1 or claim 2, which is applied to a first wafer (10a) having a diameter larger than one of 200 mm. A method for correcting a positional deviation as claimed in claim 8, which is applied to a first wafer (10a) having a diameter equal to one of 300 mm. The method of claim 1 or claim 2 for correcting the positional deviation, the bonding of the first wafer (10a) to the second wafer (20) is a molecular bond. A method for correcting a positional deviation as claimed in claim 1 or claim 2, wherein the correction function (C) is added to a coordinate of the position. A method for correcting a positional deviation as claimed in claim 1 or claim 2, wherein the correction function (C) is applied to the one obtained after applying a linear transformation (T) about one of the coordinates to an initial coordinate of the position The coordinates of the location. A method for producing a three-dimensional composite structure comprising: step (G1) generating a first micro-component layer (11) on a face of a first substrate (10); including the micro-component layer ( 11) a step of bonding the surface of the first substrate (10) to (G2) to a second substrate (20); and facing the surface including the first micro-component layer (11) a step of generating (F3 to F5, F'3 to F'5) a second micro-component layer (20) on the face of the first substrate (10), The method is characterized in that the step of generating (F3 to F5, F'3 to F'5) the second layer (12) is performed while correcting the positional deviation using the method of any one of claims 1 to 12. A three-dimensional composite structure obtained according to a method of generating one of the claims 13. A backside illuminated image sensor comprising a three dimensional composite structure as claimed in claim 14. A device for correcting a positional deviation coupled to a first wafer (10a) of a second wafer (20), the device comprising a predetermined correction for use in one of the first wafers (10a) The function (C) is applied to the member of the coordinates of each position, and the correction applied by the correction function varies only with the distance of the position relative to the center (X) of the first wafer (10a), the characteristics of the device The applied correction varies in a non-linear manner over the entire first wafer (10a) with respect to the distance of the location relative to the center (X).