System for measuring three-dimensional shape of wafer and method therefor

Abstract

Disclosed are a system for measuring a three-dimensional shape of a wafer and a method therefor. Unlike the prior art, the present invention allows a continuous phase generation unit to quickly change a wrapped phase map to a continuous phase map by applying a compensation algorithm. In addition, the shape of the wafer is calculated through such a continuous phase map. Therefore, the present invention can quickly and accurately calculate the shape of the wafer.

Description

Wafer 3D Shape Measurement System and Method

[0001] The present invention relates to a system and method for measuring the shape of a wafer.

[0002] In particular, the present invention relates to a wafer shape measurement system and method capable of quickly and accurately calculating the wafer shape by performing an algorithm quickly and accurately when measuring the wafer shape.

[0003] The semiconductor industry is a culmination of advanced technologies, manufacturing Integrated Circuits (ICs) through precise and efficient processes. As the miniaturization of semiconductor devices continues, technology capable of accurately measuring wafer flatness and fine surface features has become essential. In particular, measuring wafer surface features has established itself as a critical factor in ensuring the quality of microcircuit patterns, process yield, and reliability.

[0004] A representative optical method for measuring the shape of a wafer surface is the deflectometry measurement technique, which reconstructs the shape through phase changes of a pattern. This method calculates the shape of the wafer surface in the form of phase values ​​using patterns in the form of trigonometric functions (Sin or Cos). In deflectometry, multiple interference pattern images are captured, and phase information is also transformed using a phase shifting algorithm; based on this, the measured shape of the surface is reconstructed.

[0005] However, the phase data calculated through the phase transition algorithm is inevitably obtained as wrapped phase data that repeats within the range of ??pi to +pi. This wrapped phase data has limitations in that the phase values ​​appear periodically repeating within a certain range and cannot directly represent the continuous shape changes of the actual wafer surface.

[0006] Therefore, an unwrapping process is essential to convert this truncated phase data into continuous phase data proportional to the shape change of the actual wafer. However, when converting truncated data into continuous data, compensation operations for discontinuous phase values ​​must be performed at the pixel level, leading to a problem where processing time increases as the amount of data grows. Additionally, filtering or judgment algorithms may be added to remove noise that can act as a disruptive factor during the unwrapping process. Consequently, there is a limitation in that the entire phase unwrapping process becomes complex and computation time inevitably increases.

[0007] Due to these problems, there is a growing need for efficient and improved systems and methods to accurately and rapidly measure wafer shapes without original loss.

[0008] The present invention according to one embodiment aims to provide a wafer shape measurement system and method capable of quickly and accurately calculating the shape of a wafer without loss of the original, thereby solving the aforementioned problems.

[0009] A wafer shape measurement system according to one embodiment may include a cutting phase generation unit that generates a cutting phase map by combining images of a plurality of wafers in which a set pattern is investigated, a detection unit that calculates the difference in phase values ​​of pixels of the cutting phase map and calculates the part where a difference in phase values ​​of a preset phase value occurs, and a continuous phase generation unit that generates a continuous phase map by calculating a preset compensation algorithm on pixels where a difference in phase values ​​of the cutting phase map exists.

[0010] It may be characterized by further including a pattern irradiation unit that changes the image phase of the wafer and irradiates a plurality of patterns on the wafer.

[0011] The above detection unit may be characterized by calculating the portion where the phase value between pixels of the above-described cutting phase map changes discontinuously.

[0012] The above detection unit may be characterized by grouping pixels having positive phase values ​​into a positive group and pixels having negative phase values ​​into a negative group in the clipping phase map.

[0013] The detection unit may be characterized by distinguishing between pixels of the positive group and pixels not included in the positive group, and performing an operation to include the pixels in the positive group.

[0014] The detection unit may be characterized by distinguishing between pixels of the negative group and pixels not included in the negative group, and performing an operation to include the pixels in the negative group.

[0015] The detection unit may be characterized by performing calculations on pixels included in the positive group up to a preset range of pixels at the boundary of the positive group, and performing calculations on pixels not included in the positive group as being included in the positive group.

[0016] The detection unit may be characterized by performing calculations on pixels included in the negative group up to a preset range of pixels at the boundary of the negative group, and performing calculations to include pixels not included in the negative group as part of the negative group.

[0017] The above continuous phase generation unit may be characterized by generating a continuous phase map by applying a compensation algorithm that collectively compensates phase values ​​to pixels between the boundaries of the clipped phase maps grouped by the detection unit.

[0018] The above compensation algorithm may be characterized by an operation of adding 2π to the phase value of the pixel.

[0019] The above continuous phase map may be further characterized by including a gradient vector field calculation unit that calculates gradient vector field values.

[0020] It may further include a shape restoration calculation unit that calculates the shape of the wafer by applying a shape restoration algorithm to the above gradient vector field value.

[0021] The above shape restoration algorithm may be characterized by performing an operation to integrate the gradient vector field value.

[0022] A wafer shape measurement method according to one embodiment may include a pattern light irradiation step in which a pattern irradiation unit irradiates a set pattern onto a wafer; a cut phase map generation step in which a cut phase generation unit combines images of the wafer irradiated with the set pattern to generate a cut phase map of a single wafer; a detection step in which a detection unit calculates the phase value difference of each pixel of the cut phase map and calculates the part where a pre-set phase value difference occurs; and a compensation algorithm application step in which a continuous phase generation unit calculates a pre-set compensation algorithm on the pixel where the phase value difference exists in the cut phase map.

[0023] In the detection step above, the detection unit may be characterized by grouping pixels with positive phase values ​​into a positive group and pixels with negative phase values ​​into a negative group in the clipping phase map.

[0024] In the above detection step, the detection unit may be characterized by distinguishing parts where the phase value between pixels of the clipping phase map changes discontinuously.

[0025] After the compensation algorithm application step, the continuous phase generation unit may further include a compensation calculation step that calculates the clipped phase map into a continuous phase map in which the phase value is compensated.

[0026] After the compensation calculation step, the gradient vector field calculation unit may further include a gradient vector field calculation step for calculating the gradient vector field value of the continuous phase map.

[0027] After the gradient vector field calculation step, the shape restoration calculation unit may further include a shape restoration step in which the shape of the wafer is calculated by applying a shape restoration algorithm to the gradient vector field value.

[0028] According to one embodiment, the present invention allows the detection unit to detect parts where discontinuous changes in phase values ​​occur, and the continuous phase generation unit to collectively compensate for phase values, thereby rapidly performing calculations that occur during wafer shape calculation and enabling the wafer shape to be restored quickly and accurately.

[0029] According to one embodiment, the present invention can quickly and accurately calculate the shape of a wafer by restoring the shape of the wafer through a process of rapidly generating a continuous phase map modified from a cut phase map, obtaining a gradient vector field, and then restoring the shape.

[0030] FIG. 1 is a configuration diagram of a wafer shape measurement system according to one embodiment.

[0031] FIG. 2 illustrates the operation of a detection unit of a wafer shape measurement system according to one embodiment.

[0032] FIG. 3 illustrates the operation of a detection unit of a wafer shape measurement system according to one embodiment.

[0033] FIG. 4 illustrates the operation of a continuous phase generation unit of a wafer shape measurement system according to one embodiment.

[0034] FIG. 5 is a flowchart of a wafer shape measurement method according to one embodiment.

[0035] Hereinafter, an embodiment of the present invention will be described in detail with reference to exemplary drawings. However, this is not intended to limit the scope of the present invention.

[0036] It should be noted that when assigning reference numerals to the components of each drawing, the same components are assigned the same reference numeral whenever possible, even if they are shown in different drawings. Furthermore, in describing the present invention, if it is determined that a detailed description of related known components or functions could obscure the essence of the invention, such detailed description is omitted.

[0037] Furthermore, the size or shape of components depicted in the drawings may be exaggerated for clarity and convenience of explanation. Additionally, terms specifically defined in consideration of the structure and operation of the present invention are intended only to describe embodiments of the present invention and do not limit the scope of the present invention.

[0038] FIG. 1 is a configuration diagram of a wafer shape measurement system according to one embodiment.

[0039] A wafer shape measurement system according to one embodiment may include a pattern investigation unit (100), a cut phase generation unit (200), a detection unit (300), a continuous phase generation unit (400), a gradient vector field calculation unit (500), and a shape restoration calculation unit (600).

[0040] The pattern irradiation unit (100) can irradiate pattern light having a set phase. The pattern irradiated by the pattern irradiation unit (100) can be irradiated onto a wafer placed on a stage, reflected, and transmitted to the cutting phase generation unit (200). The pattern light irradiated by the pattern irradiation unit (100) can be irradiated while moving the phase at a constant interval (2pi / n).

[0041] The pattern irradiation unit (100) can irradiate the wafer with pattern light that has a phase shifted according to a constant interval. For example, the pattern irradiation unit (100) can irradiate pattern light having a constant phase difference about 3 times or more. Accordingly, the cutting phase generation unit (200) can acquire multiple images of the wafer that are reflected by irradiating multiple pattern lights.

[0042] Here, the pattern irradiation unit (100) can irradiate pattern light in the X direction and pattern light in the Y direction, respectively. Therefore, through the pattern irradiation unit (100), wafer images irradiated with pattern light in the X direction and Y direction can be obtained, respectively.

[0043] The cut phase generation unit (200) can convert a plurality of wafer images (four images each of X and Y) obtained through the pattern inspection unit (100) into a phase map through a phase transition algorithm. Therefore, the cut phase generation unit (200) can generate a cut phase map by combining a plurality of wafer images. Here, in the cut phase map, a part where the phase value changes rapidly (the phase value changes rapidly with a difference of 2π) can be identified between the pattern boundaries of the wafer. This part may occur because it has a period of 2π due to the characteristics of the calculation of the phase value based on trigonometric functions.

[0044] Generally, the clipping phase generation unit (200) can perform noise filtering using various types of filters after generating a clipping phase map. The filtering process can be performed to the extent that the original phase information is not distorted.

[0045] Meanwhile, the clipped phase map exhibits discontinuities in phase values ​​between pixels, which can be indirectly observed through changes in image brightness (e.g., bright → dark or dark → bright).

[0046] The detection unit (300) can detect parts where the phase changes rapidly in the cut phase map. The detection unit (300) can calculate parts where a difference in phase value occurs. For example, as can be seen in detail in FIG. 2 as described later, parts where the bright area changes rapidly can be identified when observing from left to right.

[0047] The detection unit (300) can distinguish the parts where the phase value changes rapidly into a first compensation pixel, a second compensation pixel, ..., an nth compensation pixel, etc. These nth compensation pixels can be divided into a group having a positive phase value and pixels having a negative phase value.

[0048] The detection unit (300) separates all pixels of the clipping phase map into positive pixels and negative pixels, and can distinguish and detect pixels at the boundary between the n-th reward pixel and the n+1-th reward pixel. The detection unit (300) can also perform an operation to include pixels that are not included as positive pixels (or negative pixels) of the n-th reward pixel as positive pixels (or negative pixels). This will be explained in detail later.

[0049] The detection unit (300) can distinguish pixels based on a boundary region (generally a contour) having discontinuity. Here, the distinguished pixels can be broadly divided into compensated pixels and non-compensated pixels (especially pixels where a phase jump occurs and overlaps with a boundary region such as a pattern or chip). This allows for distinguishing whether compensation is provided based on the unique characteristics of each pixel (positive / negative, nth compensated pixel, etc.). The detection unit (300) can recalculate the nth compensated pixel, which is divided into positive and negative pixels, into a single nth compensated pixel.

[0050] The continuous phase generation unit (400) can regenerate the clipped phase map calculated by the detection unit (300) into a continuous phase map. The continuous phase generation unit (400) can apply a compensation algorithm to the first to n compensation pixels distinguished by the detection unit (300). The compensation algorithm can perform an operation of adding a pre-set phase value to each group distinguished by the detection unit.

[0051] More precisely, the continuous phase generation unit (400) can perform a compensation operation suitable for one or more pixels. Here, the continuous phase generation unit (400) can distinguish pixels where the phase value changes rapidly and perform a batch operation in the same area. The continuous phase generation unit (400) can generate a modified continuous phase map by performing such a compensation operation on the clipped phase map. Through this, unclear phase jump regions of the phase map caused by metal patterns, chips, holes, etc. can be precisely detected, and unnecessary compensation can be avoided.

[0052] Therefore, the continuous phase generation unit (400) can generate an image of a wafer that reflects the actual wafer surface shape with greater accuracy. Here, since the continuous phase generation unit (400) performs a collective compensation operation on the phase value for each pixel group of the wafer separated by the detection unit, the compensation of the phase value can be calculated very quickly.

[0053] Meanwhile, this operation can be performed on images of a wafer irradiated with pattern light in both the X and Y directions.

[0054] The gradient vector field calculation unit (500) can calculate gradient vector field values ​​based on a continuous phase map. The continuous phase values ​​obtained through the continuous phase generation unit (400) are related to the surface shape information of the wafer. Through this, the gradient vector field calculation unit (500) can calculate the gradient of the wafer by calculating a gradient value for each pixel of the continuous phase map. The shape restoration calculation unit (600) can restore the height value of each pixel using a shape restoration algorithm based on the gradient vector field values ​​calculated by the gradient vector field calculation unit (500), and finally restore the shape of the wafer by expressing the value in the pixel.

[0055] Thus, the present invention can calculate the shape of a wafer. Here, a rapidly modified continuous phase map is generated through a continuous phase generation unit (400), and the shape of the wafer is calculated through a gradient vector field calculation unit (500) and a shape restoration calculation unit (600), thereby enabling the shape of the wafer to be calculated quickly and accurately.

[0056] That is, the shape restoration calculation unit (600) can restore a three-dimensional shape (Z map) from the gradient vector field of the surface and derive Peak-Valley, Warpage values, etc. of the shape. In addition, the shape restoration calculation unit (600) can distinguish each chip area from the calculated Z map and derive Peak-Valley, Warpage values, etc. within each chip area.

[0057] FIG. 2 illustrates the operation of a detection unit of a wafer shape measurement system according to one embodiment.

[0058] As described above, the cut phase generation unit (200) can generate a cut phase map. However, the phase values ​​of some pixels not included in the cut phase map may be subject to the compensation algorithm of the continuous phase generation unit (400). Therefore, if compensation is performed on a pixel-by-pixel basis, the computation time may increase when computing the wafer surface. To resolve this, the detection unit (300) may perform an operation of grouping pixels into a positive group or a negative group.

[0059] Let us examine this with reference to Fig. 2. When the detection unit (300) groups the pixels of the clipping phase map into a negative group (phase value greater than or equal to +π), some pixels may not be included in the negative group. That is, some pixels may not be classified as compensation pixels.

[0060] The detection unit (300) can expand the boundary area of ​​the negative group by the amount of the set area. Here, the range of the set area may be set differently depending on the user. The detection unit (300) can perform an operation to include some pixels not included in the negative group into the negative group. Therefore, the clipping phase map can be grouped into the negative group in a batch, and a batch compensation operation can be performed.

[0061] This operation can be performed by the detection unit (300) in the same way for the positive group. That is, the detection unit (300) groups the clipping phase map into a positive group (0 or greater, less than π) and performs an operation that includes ungrouped pixels into the positive group. Therefore, the clipping phase map modified by the detection unit (300) can be grouped into a positive group and a negative group.

[0062] Therefore, the present invention can group all pixels into a positive group or a negative group.

[0063] FIG. 3 illustrates the operation of a detection unit of a wafer shape measurement system according to one embodiment, and FIG. 4 illustrates the operation of a continuous phase generation unit of a wafer shape measurement system according to one embodiment.

[0064] As can be seen through FIG. 3, the detection unit (300) can detect parts where the phase value of the cut phase map changes discontinuously. Due to the detection unit (300), the phase value of the cut phase map can be distinguished as positive or negative. Therefore, as shown in FIG. 3, the phase value for the wafer surface may only have parts that undergo sudden, abrupt changes at specific pixels. (This is because, as previously explained, the phase value calculated based on trigonometric functions is repeated within a certain range (-π to +π).)

[0065] The detection unit (300) can detect parts where the phase value changes discontinuously between pixels, that is, parts where the phase changes abruptly, through the above operation. Therefore, this part can also be grouped.

[0066] As illustrated in FIG. 4, the continuous phase generation unit (400) applies a preset compensation algorithm to phase discontinuous pixels detected by the detection unit (300). According to an embodiment of the present invention, by performing a batch phase compensation (e.g., +2π) for a point where the phase decreases rapidly (e.g., a part that changes from π to -π), the phase value can be restored to a continuous state.

[0067] The continuous phase generation unit (400) can generate a continuous phase map with continuous phases quickly and accurately by performing the above compensation operation collectively on the area of ​​the cut phase map where compensation is required. As a result, the actual shape information of the wafer surface can be represented more accurately, and the speed, accuracy, and reliability of the subsequent gradient vector field operation and shape restoration (e.g., integration operation) can be significantly improved.

[0068] Meanwhile, it is natural that the compensation of the continuous phase generation unit (400) provides different compensation to the parts grouped by the detection unit (300) in a collective manner.

[0069] FIG. 5 is a flowchart of a wafer shape measurement method according to one embodiment.

[0070] A wafer shape measurement method according to one embodiment may include a pattern light irradiation step (S100), a cut phase map generation step (S200), a detection step (S300), a compensation algorithm application step (S400), a gradient vector field calculation step (S500), and a shape restoration step (S600).

[0071] The pattern light irradiation step (S100) is a step in which the pattern irradiation unit (100) irradiates a plurality of pattern lights having preset phase values ​​onto a wafer. In the pattern light irradiation step (S100), the pattern irradiation unit (100) irradiates about three or more pattern lights having different phase values ​​onto the wafer surface, and in particular, can irradiate pattern lights in the X direction and the Y direction, respectively.

[0072] The cut phase map generation step (S200) is a step in which the cut phase generation unit (200) acquires a plurality of pattern light images reflected from a wafer and combines them to generate a cut phase map. The cut phase generation unit (200) can generate a cut phase map in the form of a phase map by combining a plurality of wafer images. (At this time, the cut phase generation unit (200) can generate a cut phase map with noise removed by removing noise from the generated cut phase map.)

[0073] The detection step (S300) may mean a step in which the detection unit (300) groups the clipping phase map into a positive group and a negative group, identifies pixels in the boundary area that are not included in these two groups, and performs an operation to include them in the positive group or the negative group. Here, the details regarding the pixels in the boundary area are as described through FIG. 2.

[0074] The compensation algorithm application step (S400) is a step in which the continuous phase generation unit (400) applies a pre-set compensation algorithm to phase discontinuous pixels detected by the detection unit (300) to compensate for the phase value. In an embodiment of the present invention, the detection unit (300) can perform collective compensation on each of the grouped boundary regions.

[0075] Subsequently, the continuous phase generation unit (400) can generate a continuous phase map in which the phase is continuously adjusted, including pixels to which a compensation algorithm is applied. This eliminates discontinuities in the phase between pixels and secures a continuous phase distribution, thereby enabling the generation of a continuous phase map capable of expressing accurate shape information.

[0076] The gradient vector field calculation step (S500) is a step in which the gradient vector field calculation unit (500) calculates the gradient vector field based on a continuous phase map. In this step, the gradient vector field calculation unit (500) extracts gradient information of the wafer surface from the continuous phase change between pixels and calculates an accurate gradient value.

[0077] Finally, the shape restoration step (S600) is a step in which the shape restoration calculation unit (600) restores a 3D shape (Z map) from the gradient vector field calculated by the gradient vector field calculation unit (500) and derives the Peak-Valley, Warpage values, etc. of the shape.

[0078] The shape restoration operation unit (600) calculates absolute shape information per pixel by applying a shape restoration algorithm (e.g., integration) to the gradient vector field, thereby obtaining accurate and reliable shape information of the wafer surface.

[0079] As such, the wafer shape measurement method according to the present invention is configured to enable rapid and accurate wafer shape measurement through the organic linkage of each step.

[0080] Although the present invention has been illustrated and described in relation to specific embodiments, it will be obvious to those skilled in the art that the present invention can be modified and changed in various ways without departing from the technical spirit of the invention as provided by the following claims.

[0081] [Explanation of the symbol]

[0082] 100 : Pattern Investigation Department

[0083] 200 : Cutting phase generation unit

[0084] 300 : Detection unit

[0085] 400 : Continuous phase generation unit

[0086] 500 : Gradient vector field operation unit

[0087] 600 : Shape restoration operation unit

Claims

1. A cutting phase generation unit that generates a cutting phase map by combining images of multiple wafers in which a set pattern has been investigated; A detection unit that calculates the difference in phase values ​​of pixels of the above-mentioned clipping phase map and calculates the part where the difference in phase values ​​of a preset phase value occurs; and A continuous phase generation unit that generates a continuous phase map by applying a preset compensation algorithm to pixels where a difference in phase values ​​exists in the above-mentioned clipping phase map. A wafer shape measurement system including 2. In Paragraph 1, Further including a pattern irradiation unit that changes the image phase of the wafer and irradiates a plurality of patterns on the wafer. A wafer shape measurement system characterized by 3. In Paragraph 1, The above detection unit Calculating the part where the phase value between pixels of the above-mentioned clipping phase map changes discontinuously A wafer shape measurement system characterized by 4. In Paragraph 3, A wafer shape measurement system characterized by the above detection unit grouping pixels having positive phase values ​​into a positive group and pixels having negative phase values ​​into a negative group in a cut phase map.

5. In Paragraph 4, The above detection unit A wafer shape measurement system characterized by distinguishing between pixels of the above-mentioned positive group and pixels not included in the above-mentioned positive group, and performing an operation to include the pixels in the positive group.

6. In Paragraph 5, The above detection unit A wafer shape measurement system characterized by distinguishing between pixels of the negative group and pixels not included in the negative group, and performing an operation to include the pixels in the negative group.

7. In Paragraph 5, The above detection unit A wafer shape measurement system characterized by performing calculations on pixels included in the positive group up to a preset range of pixels at the boundary of the positive group, and performing calculations on pixels not included in the positive group as positive group.

8. In Paragraph 6, The above detection unit A wafer shape measurement system characterized by performing calculations on pixels included in the negative group up to a preset range of pixels at the boundary of the negative group, and performing calculations that include pixels not included in the negative group as the negative group.

9. In Paragraph 8, The above continuous phase generation unit Generating a continuous phase map by applying a compensation algorithm that collectively compensates phase values ​​to pixels between the boundaries of the clipped phase maps grouped by the detection unit. A wafer shape measurement system characterized by 10. In Paragraph 9, The above reward algorithm is It is an operation that adds 2π to the phase value of the above pixel. A wafer shape measurement system characterized by 11. In Paragraph 10, Further including a gradient vector field calculation unit that calculates gradient vector field values ​​in the above continuous phase map. A wafer shape measurement system characterized by 12. In Paragraph 11, A shape restoration calculation unit that calculates the shape of the wafer by applying a shape restoration algorithm to the above gradient vector field value. A wafer shape measurement system including further 13. In Paragraph 12, The above shape restoration algorithm performs an operation of integrating the above gradient vector field value. A wafer shape measurement system characterized by 14. A pattern light irradiation step in which a pattern irradiation unit irradiates a set pattern onto a wafer; A cutting phase map generation step in which a cutting phase generation unit combines images of wafers in which the set pattern is investigated to generate a cutting phase map of a single wafer; A detection step in which a detection unit calculates the phase value difference of each pixel of the above-mentioned cutting phase map and calculates the part where a pre-set phase value difference occurs; and A compensation algorithm application step in which a continuous phase generation unit computes a preset compensation algorithm on pixels in the clipped phase map where the phase value difference exists. A wafer shape measurement method including 15. In Paragraph 14, In the above detection step The detection unit groups pixels with positive phase values ​​in the clipping phase map into a positive group, and groups pixels with negative phase values ​​into a negative group. A wafer shape measurement method characterized by 16. In Paragraph 14, In the above detection step, the detection unit distinguishes the parts where the phase values ​​between pixels of the clipping phase map change discontinuously. A wafer shape measurement method characterized by 17. In Paragraph 16, After the above compensation algorithm application step The above continuous phase generation unit A compensation calculation step that calculates the above clipped phase map into a continuous phase map in which the phase values ​​are compensated. A wafer shape measurement method further comprising 18. In Paragraph 17, After the compensation calculation step A gradient vector field calculation step in which a gradient vector field calculation unit calculates the gradient vector field value of the continuous phase map. A wafer shape measurement method further comprising 19. In Paragraph 18, A wafer shape measurement method further comprising a shape restoration step in which, after the above-mentioned gradient vector field calculation step, a shape restoration calculation unit applies a shape restoration algorithm to the above-mentioned gradient vector field value to calculate the shape of the wafer.

Images