Centrosymmetric grating assembly, tilt angle measurement method and tilt elimination method

Abstract

This invention relates to a centrosymmetric grating assembly, a tilt angle measurement method, and a tilt elimination method. The tilt angle measurement method based on a centrosymmetric grating assembly acquires moiré fringes in at least one quadrant formed by a first grating and a second grating. It acquires the frequency components of the moiré fringes along the two diagonal directions of the quadrant, calculates the angle between the fringes and the original grating direction using the frequency components, and calculates the tilt angle between the first and second gratings using the angle. In this scheme, the orthogonal components of the moiré fringes obtained after tilting the gratings can be extracted by decomposing the frequency or its equivalent phase change. The corresponding angle of the moiré fringes is then obtained based on these orthogonal components, and the final tilt angle between the two gratings can be obtained from the tilt angle of the moiré fringes. Therefore, the angle between the mask and the wafer during photolithography can be corrected based on the obtained tilt angle.

Description

Technical Field

[0001] This invention belongs to the field of photolithography alignment technology, specifically relating to a centrally symmetric grating assembly, a tilt angle measurement method, and a tilt elimination method. Background Technology

[0002] Photolithography is a crucial precision machining technology in semiconductor manufacturing, and its accuracy directly determines the alignment quality of wafer patterns and the final yield. During photolithography alignment, the tilt error between the wafer and the mask is one of the key factors affecting alignment accuracy. Traditional tilt error correction methods typically rely on linear moiré fringes generated by unidirectional gratings, resolving the tilt angle through phase calculation. However, due to the insufficient symmetry of unidirectional gratings, the generated moiré fringes can only provide tilt angle information in a single direction, making them susceptible to noise interference in practical applications and making it difficult to ensure the stability and accuracy of the calculation results. Summary of the Invention

[0003] In view of the limited stability and accuracy of existing technologies for photolithography tilt detection based on moiré fringes generated by unidirectional gratings, this invention provides a centrally symmetric grating assembly, a tilt angle measurement method, and a tilt elimination method.

[0004] This application provides a centrally symmetric grating assembly, including a first grating and a second grating, wherein the grating planes of the first grating and the second grating are perpendicular to the illumination light path, and the first grating and the second grating are arranged sequentially along the illumination light path;

[0005] The first and second gratings are divided into four quadrants (first, second, third, and fourth) around the center of the grating by two orthogonal axes. An L-shaped grating is formed in each sub-quadrant, and each L-shaped grating has a first grating segment and a second grating segment parallel to the two orthogonal axes. The first grating segment and the second grating segment are connected at one end to form a bend, which extends outward along the diagonal of the quadrant.

[0006] The grating period of the L-shaped grating repeats alternately with the first period P1 and the second period P2 in different quadrants, and P1≠P2. The grating period of the L-shaped grating in the corresponding quadrant of the second grating is complementary to the grating period of the L-shaped grating in the corresponding quadrant of the first grating.

[0007] Preferably, the relationship between the first period P1 and the second period P2 is P1 = 1.1P2.

[0008] This application provides a tilt angle measurement method based on the above-mentioned centrosymmetric grating assembly. The method involves acquiring moiré fringes in at least one quadrant formed by a first grating and a second grating, acquiring the frequency components of the moiré fringes along the two diagonal directions of the quadrant, calculating the angle θ1 between the fringes and the original grating direction using the frequency components, and then calculating the tilt angle ε between the first grating and the second grating using the angle θ1.z1 .

[0009] Preferably, in the calculation of the angle θ1 between the stripe and the original grating direction obtained by frequency component calculation, the angle θ1 is calculated by the following formula:

[0010] θ1=cot -1 (f x45° / f y135° -45°

[0011] Among them, f x45° and f y13° These are frequency components in different diagonal directions.

[0012] Preferably, the tilt angle ε between the first grating and the second grating is calculated using the included angle θ1. z1 By substituting the included angle θ1 into the following formula, we obtain:

[0013]

[0014] Preferably, in the calculation of the angle θ1 between the stripe and the original grating direction obtained by frequency component calculation, the angle θ1 is calculated by the following formula:

[0015] in, and This represents the phase difference along different diagonal directions.

[0016] Preferably, the phase difference is obtained through continuous phase estimation. In the continuous phase estimation process, the stripe image is converted from the spatial domain to the frequency domain by two-dimensional Fourier transform to separate the positive first-order spectrum. The positive first-order spectrum is restored to the time domain by inverse Fourier transform to generate a phase distribution map.

[0017] Preferably, the phase unwrapping is performed using the quality map-guided method during the continuous phase acquisition process; during the unwrapping process, the most reliable integral path is selected through dynamic path planning, and the integral result of the path is used as the final unwrapping result.

[0018] This application provides a tilt elimination method, which involves cyclically obtaining the real-time tilt angle between the wafer and the mask using any of the tilt angle measurement methods described above during the alignment process. Then, the selected real-time tilt angle is input into a high-precision rotary stage, and the relative tilt angle between the wafer and the mask is adjusted in real time. The above steps are repeated until the tilt error is completely eliminated or the preset accuracy condition is met.

[0019] Preferably, in the step of obtaining the real-time tilt angle between the wafer and the mask, the real-time tilt angle is taken from one or more regions of the centrosymmetric grating assembly; when using multiple regions of the centrosymmetric grating, the real-time tilt angle is taken from the average of the tilt angles of the multiple regions or the value of the tilt angle with the highest accuracy among them.

[0020] This application proposes a method for measuring the tilt angle of a centrosymmetric grating assembly and a corresponding method for eliminating lithographic tilt. In this method, the orthogonal components of the moiré fringes obtained after grating tilting can be extracted by decomposing the frequency or equivalent phase change of the fringes. Based on these orthogonal components, the included angle of the moiré fringes is obtained. Furthermore, the final tilt angle between the two gratings can be obtained from the tilt angle of the moiré fringes. Therefore, the angle between the mask and the wafer during lithography can be corrected based on the obtained tilt angle. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the structure of the centrally symmetric grating assembly of the present invention;

[0022] Figure 2 This is a schematic diagram of the structure of the first grating 11 of the present invention;

[0023] Figure 3 , Figure 4 ε z1 =0.5° and ε z1 Moiré fringe pattern at 1°;

[0024] Figure 5 This is a schematic diagram illustrating the calculation of the tilt angle according to the present invention.

[0025] In the picture:

[0026] 1: Centrally symmetric grating assembly; 11: First grating; 111: First grating segment; 112: Second grating segment; 113: Folded corner; 12: Second grating; 13: L-shaped grating; D: Straight line; O: Grating center. Detailed Implementation

[0027] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. In this specification, the dimensions of the drawings do not represent the actual dimensions. They are only used to illustrate the relative positional and connection relationships between the components. Components with the same name or the same reference numeral represent similar or identical structures and are limited to illustrative purposes.

[0028] In the description of this embodiment, the terms "upper," "lower," "left," and "right," etc., refer to the orientation or positional relationship shown in the accompanying drawings. They are used only for ease of description and simplification of operation, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. In addition, the terms "first" and "second" are used only for distinction in description and have no special meaning.

[0029] This invention first provides a centrally symmetric grating assembly 1. Its structure is as follows: Figure 1 As shown, there are two sets of centrally symmetrical gratings arranged in parallel, namely the first grating 11 and the second grating 12 in the figure. The grating planes of the first grating 11 and the second grating 12 are perpendicular to the illumination light path, and the first grating 11 and the second grating 12 are arranged sequentially along the illumination light path.

[0030] For the first grating 11 and the second grating 12, they are divided into four quadrants—the first, second, third, and fourth—circumferentially distributed around the grating center O by two orthogonal axes. For example... Figure 2 As shown, an L-shaped grating 13 is formed in each sub-quadrant. Each L-shaped grating 13 has a first grating segment 111 and a second grating segment 112, each parallel to two orthogonal axes. The first grating segment 111 and the second grating segment 112 are connected at one end to form a bend 113. In the L-shaped grating, the bend 113 extends outward along a straight line D within that quadrant. In some embodiments, the straight line D is the angle bisector of that quadrant.

[0031] For the L-shaped grating 13, its grating period repeats alternately with a first period P1 and a second period P2 in different quadrants. To achieve moiré fringe generation, the grating period of the L-shaped grating 13 in the corresponding quadrant of the second grating 12 must be complementary to the grating period of the L-shaped grating 13 in the corresponding quadrant of the first grating 11. For example, for the first grating 11, its grating period in the first, second, third, and fourth quadrants is P1, P2, P1, P2, respectively. Correspondingly, on the second grating 12, its grating period in the first, second, third, and fourth quadrants is P2, P1, P2, P1, respectively.

[0032] In the actual photolithography alignment process, the first grating 11 and the second grating 12 can be formed on the wafer and the mask surface, respectively. Together, they form a centrally symmetric grating assembly 1, ensuring that the period design of the grating meets the requirement of central symmetry. Specifically, for the first grating 11, the grating period in the first and third quadrants is set to P1, and the grating period in the second and fourth quadrants is set to P2. P1 ≠ P2, specifically set to P1 = 1.1P2. This period design allows for a more significant moiré fringe amplification effect during interference.

[0033] The alignment marks on the mask are consistent with the alignment mark structure on the wafer, and their periodic parameters and centrosymmetry design are complementary to those of the wafer marks. The design of the centrosymmetric grating ensures that the grating periods of each quadrant are matched to generate square moiré fringes during subsequent photolithography alignment, laying the foundation for tilt angle calculation.

[0034] like Figure 1 As shown, the optical path is constructed using coaxial illumination. A high-intensity, stable collimated light source (such as a laser or LED) is selected, and the light source is uniformly expanded through a collimating beam expander to cover the entire alignment mark area on the wafer and mask. An aperture filter is placed in the optical path to filter out stray light and unwanted interference light, improving the signal-to-noise ratio of the optical path. The collimated light source is guided sequentially through the alignment marks on the mask and wafer. The structure of the centrally symmetric grating ensures the generation of a four-way symmetrical diffracted beam when the light source penetrates. By adjusting the relative position and angle of optical elements (such as mirrors and collimating lenses), the light source is ensured to be perpendicularly incident on the mask and wafer surfaces, guaranteeing the formation of stable square moiré fringes by the diffracted light.

[0035] Corresponding to the centrally symmetric grating assembly 1 above, the relative rotational tilt between the first grating 11 and the second grating 12 can be corrected.

[0036] Specifically, the collimated light undergoes diffraction when passing through the first grating 11 and the second grating 12. The diffracted beam is generated with a period of P through superposition and interference modulation. m Square moiré fringes. Specific description follows:

[0037] Moiré fringe period calculation: When the wafer rotates between itself and the mask, the resulting tilt angle is ε. z1 At that time, the period of the moiré fringe is P m With tilt angle ε z1 The relationship is:

[0038]

[0039] It is obvious that different tilt angles will lead to changes in the period of the moiré fringes.

[0040] Stripe variations caused by tilting: The size of the tilt angle not only affects the stripe period but also causes the stripes to tilt in direction. For example... Figure 3 , Figure 4 As shown, ε 21 =0.5° and ε z1 Moiré fringe pattern at 1°, different tilt angles lead to significant changes in fringe direction and density.

[0041] like Figure 5As shown, the square moiré fringes are divided into four regions—left I1, right I2, upper I3, and lower I4—by a straight line D. The fringes in each region represent diffraction information in different directions. To improve calculation accuracy, the fringes are further cut along the junction of the two L-shaped gratings 13 in each region. Within any region, local fringe features can be extracted for phase calculation. The moiré fringes of any first grating segment 111 or any second grating segment 112 are extracted to calculate the tilt angle after rotation between the first grating 11 and the second grating 12.

[0042] Taking the region containing one of the first grid segments 111 as an example, its fringe intensity can be expressed as:

[0043] I(x,y)=a(x,y)+b(x,y)cos[2πxf x +2πyf y ]

[0044] Where (x,y) represents pixel coordinates; a(x,y) and b(x,y) represent background intensity and amplitude intensity, respectively; f x =f m cos(θ1) and f y =f m sin(θ1) represents the fringe frequency f m =1 / P m Decompose the frequency components on the X and Y axes; θ1 represents the angle between the stripe and the original grating direction.

[0045] The stripe frequency f m Along X 45° (Quadrant diagonal) direction and Y 135° The two orthogonal axes along the (quadrant diagonal) direction are decomposed into f. x45° and f y135° The frequency component can be represented as:

[0046]

[0047] Therefore, the angle θ1 between the stripes and the original grating direction can be calculated:

[0048] θ1=cot -1 (f x45° / f y135° -45°.

[0049] Because the angle θ1 between the stripes and the original grating direction is equal to the tilt angle ε... z1 The numerical relationship between them can be expressed as:

[0050]

[0051] Substituting the calculated included angle θ1 into the above formula, the tilt angle ε can be calculated.z The value of .

[0052] For example, the square moiré fringes are divided diagonally into four regions: left I1, right I2, upper I3, and lower I4. The fringes in each region represent diffraction information in different directions. To improve computational accuracy, the fringes in each region are further subdivided along the midline, and local fringe features are extracted for phase calculation. For instance, the intensity of the upper fringe in left I1 can be expressed as:

[0053]

[0054] (x,y) represents pixel coordinates; a(x,y) and b(x,y) represent background intensity and amplitude intensity, respectively; f x =f m cos(θ1) and f y =f m sin(θ1) represents the fringe frequency f m =1 / P m Decompose the frequency components on the X and Y axes; θ1 represents the angle between the upper stripe of the left I1 region and the Y axis.

[0055] The angle θ1 between the upper stripe of the left I1 region and the Y-axis and the tilt angle ε z The numerical relationship between them can be expressed as:

[0056]

[0057] The stripe frequency f m Along X 45° Direction and Y 135° The two orthogonal axes of direction are decomposed into f x45° and f y135° The frequency component can be represented as:

[0058]

[0059] Therefore, angle θ1 can be calculated:

[0060] θ1=cot -1 (f x45° / f y135° -45°

[0061] Substituting the calculated angle θ1 into the previous equation, the tilt angle ε can be determined. z The value of .

[0062] In a further scheme, the frequency components can be calculated through continuous phase decomposition. The continuous phase is decomposed along the orthogonal axis X. 45° and Y 135° Directional decomposition, two frequency components f x45° and f y135°Calculate using the following formulas respectively:

[0063]

[0064] and For two orthogonal axes X 45° and Y 135° Phase difference on; Δn x45° and Δn y135° Let Δn be the number of pixels in the corresponding direction. Typically, for ease of calculation, let Δn be... x45° =Δn y135° The tilt angle θ1 is calculated using the following formula:

[0065]

[0066] Obviously, the corresponding included angle θ1 can be obtained from this, and the tilt angle ε can be further obtained. z1 .

[0067] To obtain continuous phase estimation, the technical solution of this application first employs a two-dimensional fast Fourier transform (2D-FFT) algorithm to process the moiré fringe pattern, achieving high-precision calculation of phase extraction and phase unwrapping. First, the fringe image is transformed from the spatial domain to the frequency domain using a two-dimensional Fourier transform to separate the positive first-order spectrum containing phase information. During frequency domain processing, a two-dimensional Hamming bandpass filter is designed and applied to filter out zero-frequency and high-frequency noise components while reducing spectral leakage. The main lobe width and side lobe attenuation rate of the filter are optimized to ensure the integrity of the positive first-order spectrum while effectively suppressing noise interference. Subsequently, the filtered positive first-order spectrum is restored to the time domain using an inverse Fourier transform (IFT) to generate a phase distribution map.

[0068] To further obtain continuous phase, this invention employs a quality map-guided method from path tracing algorithms for phase unwrapping. This method generates a quality map to evaluate the reliability of each pixel in the wrapped phase, prioritizing unwrapping pixels in high-quality regions and avoiding low-quality regions caused by noise or distortion, thereby significantly improving the robustness and accuracy of unwrapping. During unwrapping, the most reliable integral path is selected through dynamic path planning, and the integral result of the path is used as the final unwrapping result. To further optimize the unwrapping effect, this invention preprocesses the stripe image with Gaussian filtering before unwrapping to reduce the influence of random noise, and uses a region-weighted averaging method to reduce errors when processing boundary points.

[0069] Based on the aforementioned centrosymmetric grating assembly, we provide a method for eliminating photolithographic tilt. During the alignment process, the method cyclically acquires the real-time tilt angle between the wafer and the mask using the centrosymmetric grating assembly. Then, the selected real-time tilt angle value is input into a high-precision rotary stage, and the relative tilt angle between the wafer and the mask is adjusted in real time. This process is repeated until the tilt error is completely eliminated or a preset accuracy condition is met. In the step of acquiring the real-time tilt angle between the wafer and the mask using the centrosymmetric grating assembly, the real-time tilt angle is taken from one or more regions of the centrosymmetric grating assembly. When using multiple regions of the centrosymmetric grating, the real-time tilt angle can be the average of the tilt angles of the multiple regions or the value of the tilt angle with the highest accuracy.

[0070] The combination of a centrosymmetric grating and moiré fringe technology provides a novel solution for high-precision correction of lithography tilt errors. The unique structure of the centrosymmetric grating generates square moiré fringes, a characteristic significantly different from traditional linear moiré fringes. These square moiré fringes have independent phase distributions in four directions, allowing tilt angle information to be extracted separately from each direction, thus greatly enhancing anti-interference capabilities. By comparing the calculated tilt angles in the four directions, the tilt angle value with the smallest error and highest accuracy is selected as the feedback input for adjustment of the high-precision rotary stage. This multi-directional calculation and selection mechanism significantly improves the robustness and accuracy of tilt correction, providing a more stable and reliable technical guarantee for lithography alignment in complex application scenarios.

[0071] Despite the significant theoretical advantages of centrosymmetric gratings, research on lithographic tilt correction techniques based on square moiré fringes remains incomplete. Numerous technical challenges persist in the practical implementation of rapidly extracting the moiré fringe phase, decomposing orthogonal phase components, and accurately calculating the tilt angle. Therefore, an innovative method is urgently needed to fully leverage the four-directional characteristics of centrosymmetric gratings, efficiently extract tilt angle information, and achieve high-precision tilt correction between the wafer and the mask.

[0072] The beneficial technical effects of this invention are:

[0073] 1. Significantly Improved Correction Accuracy: This invention achieves multi-directional tilt angle information calculation through independent phase extraction and decomposition of the four directions of a square moiré fringe. Compared to the linear moiré fringes generated by traditional unidirectional gratings, this invention can select the tilt angle with the smallest error and highest accuracy for feedback adjustment, greatly improving the accuracy of correction.

[0074] 2. Enhanced anti-interference capability: Traditional linear moiré fringes are susceptible to noise and optical path interference, leading to instability in tilt angle calculation results. This invention utilizes the structural characteristics of a centrally symmetric grating to generate square moiré fringes, exhibiting excellent symmetry and anti-interference capability, ensuring the reliability of correction even under complex optical path conditions.

[0075] 3. Higher computational efficiency: By extracting the fringe phase through 2D-FFT and combining orthogonal axis decomposition and fast phase unwrapping techniques, this invention significantly simplifies the complexity of tilt angle calculation and achieves higher computational efficiency while maintaining accuracy compared with existing methods.

[0076] 4. Wider range of applications: This invention is applicable to a variety of photolithography alignment scenarios, especially in semiconductor manufacturing applications where tilt error correction is required to a high degree. It has good applicability and can effectively meet the needs of high-precision photolithography alignment.

[0077] In summary, this invention overcomes the shortcomings of insufficient correction accuracy, poor anti-interference ability and high computational complexity in the prior art, and provides a tilt error correction method with high accuracy, high efficiency and strong stability.

[0078] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for measuring the tilt angle of a centrally symmetric grating assembly, characterized in that, The centrally symmetrical grating assembly includes a first grating (11) and a second grating (12). The grating planes of the first grating (11) and the second grating (12) are perpendicular to the illumination light path. The first grating (11) and the second grating (12) are arranged sequentially along the illumination light path. The first grating (11) and the second grating (12) are divided into four quadrants, namely the first, second, third and fourth quadrants, by two orthogonal axes around the grating center (O). An L-shaped grating (13) is formed in each sub-quadrant. The L-shaped grating (13) has a first grating segment (111) and a second grating segment (112) that are parallel to the two orthogonal axes. The first grating segment (111) and the second grating segment (112) are connected at one end to form a bend (113), and the bend (113) extends outward along the diagonal of the quadrant. The grating period of the L-shaped grating (13) is repeated alternately with the first period P1 and the second period P2 in different quadrants, and P1≠P2. The grating period of the L-shaped grating (13) in the corresponding quadrant of the second grating (12) is complementary to the grating period of the L-shaped grating (13) in the corresponding quadrant of the first grating (11). The tilt angle measurement method obtains moiré fringes in at least one quadrant formed by the first grating (11) and the second grating (12), obtains the frequency components of the moiré fringes in the two diagonal directions of the quadrant, calculates the angle θ1 between the fringes and the original grating direction using the frequency components, and calculates the tilt angle ε between the first grating (11) and the second grating (12) using the angle θ1. z1 .

2. The tilt angle measurement method as described in claim 1, characterized in that, The angle θ1 between the stripe and the original grating direction is calculated using the frequency components. The angle θ1 is calculated by the following formula: θ1=cot -1 (f x45° / f y135° )-45°, Among them, f x4° and f y135° These are frequency components in different diagonal directions.

3. The tilt angle measurement method as described in claim 1, characterized in that, The tilt angle ε between the first grating (11) and the second grating (12) is calculated using the included angle θ1. z1 By substituting the included angle θ1 into the following formula, we obtain:

4. The tilt angle measurement method as described in claim 1, characterized in that, The angle θ1 between the stripe and the original grating direction is calculated using the frequency components. The angle θ1 is calculated by the following formula: in, and This represents the phase difference along different diagonal directions.

5. The tilt angle measurement method as described in claim 4, characterized in that, The phase difference is obtained through continuous phase estimation. During the continuous phase estimation process, the stripe image is converted from the spatial domain to the frequency domain by two-dimensional Fourier transform to separate the positive first-order spectrum. The positive first-order spectrum is restored to the time domain by inverse Fourier transform to generate a phase distribution map.

6. The tilt angle measurement method as described in claim 5, characterized in that, During the process of obtaining continuous phase, the phase unwrapping method is performed using the quality map-guided method. During the unwrapping process, the most reliable integral path is selected through dynamic path planning, and the integral result of the path is used as the final unwrapping result.

7. The tilt angle measurement method as described in claim 1, characterized in that, The relationship between the first period P1 and the second period P2 is P1 = 1.1P2.

8. A method for eliminating tilt, characterized in that, During the alignment process, the real-time tilt angle between the wafer and the mask is obtained cyclically using the tilt angle measurement method described in any one of claims 1-7. The selected real-time tilt angle is then input into the rotary table, and the relative tilt angle between the wafer and the mask is adjusted in real time. The above steps are repeated until the tilt error is completely eliminated or the preset accuracy condition is met.

9. The tilt elimination method as described in claim 8, characterized in that, In the step of obtaining the real-time tilt angle between the wafer and the mask, the real-time tilt angle is taken from one or more regions of the centrosymmetric grating assembly; when using multiple regions of the centrosymmetric grating, the real-time tilt angle is taken from the average of the tilt angles of the multiple regions or the value of the tilt angle with the highest accuracy among them.

Images