Method for controlling photoetching overlay precision in grouping manner
By classifying and scoring the photolithography overlay map, and tracking the previous process equipment, online correction of photolithography overlay accuracy is achieved, which solves the problem of the impact of the previous process on photolithography overlay accuracy and improves overlay accuracy and device yield.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-20
- Publication Date
- 2026-04-03
AI Technical Summary
Existing technologies cannot effectively correct the impact of each preceding process on the photolithography overlay accuracy, resulting in insufficient overlay accuracy. Especially at advanced technology nodes, the alignment correction of the photolithography process cannot completely eliminate the influence of etching and polishing processes, leading to a decrease in device yield.
By collecting lithography overlay maps, the first classification is performed and the equipment of the previous process is tracked. Scores are calculated to achieve a one-to-one correspondence between equipment and classification groups. Subsequent batches are classified a second time and runtime overlay accuracy compensation values are set. Online correction is performed using the APC system.
It improves the precision of photolithography overlay, reduces overlay defects, increases device yield, reduces online rework rate, and saves costs.
Smart Images
Figure CN121785068A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a method for manufacturing semiconductor integrated circuits, and more particularly to a method for controlling the precision of photolithography overlay in groups. Background Technology
[0002] The pattern remaining on the photoresist after exposure and development is called the current layer, and it must be aligned with the existing pattern (reference layer) on the wafer substrate. This ensures correct connections between different parts of the device. Excessive alignment error is one of the main causes of short circuits and open circuits in devices, significantly impacting device yield.
[0003] The overlay error data measured on the wafer is processed by a model and then fed back to the lithography machine, allowing for further corrections to the machine's alignment system. For example... Figure 1 The diagram shows a schematic of the alignment system correction in a conventional lithography machine. A batch of wafers is placed in wafer cassette 101. During lithography, a wafer 105 to be lithographically processed is placed on the wafer placement platform of the lithography machine 102, while the mask 103 is placed on the mask placement platform. After the laser beam passes through the mask 103, it is scaled by the lens group 104 and projects the pattern in the mask 103 onto the wafer 105, thus transferring the pattern onto the wafer 105. Before exposure, alignment is required, i.e., alignment between the wafer 105 and the mask 103. Alignment requires an alignment laser beam. As shown in the figure corresponding to mark 106, the alignment accuracy is determined by measuring the overlay error between the current layer pattern and the reference pattern after lithography. Higher alignment accuracy results in a smaller overlay error. Based on the overlay error, the model shown in mark 107 is calculated, and the correctable amount is fed back to the lithography machine 102. In advanced process control (APC), overlay error feedback correction can be set up. When setting up the feedback system, the correction parameters calculated by the model are typically averaged over a certain period before being fed back to the lithography machine. This reduces fluctuations in the feedback parameters and avoids instability in the feedback system's operation.
[0004] Online correction models include linear models, higher-order models, and correction per exposure (CPE). Generally, overlay accuracy control uses feedback (FB), while feedforward (FF) is used for higher accuracy requirements to reduce overlay differences caused by lot-to-lot variations. However, even with FF, it can only predict part of the impact of the previous layer process by observing the wafer overlay behavior of the previous lithography station, making it difficult to fully meet the overlay accuracy control standards of the current layer.
[0005] As technology nodes become more advanced, the requirements for overlay accuracy in photolithography are becoming increasingly stringent. Meanwhile, the impact of various front-end processes, including those other than photolithography, on photolithography is becoming increasingly significant. For example, etching and polishing processes severely damage photolithographic alignment and overlay patterns. Not only are there differences in the impact on the center and edges, but the degree of impact on photolithography also varies between different equipment in the same process step. The alignment correction in existing photolithography processes cannot eliminate these effects. Summary of the Invention
[0006] The technical problem to be solved by the present invention is to provide a method for grouping and controlling the overlay accuracy of photolithography, which can correct the influence of each previous process on the overlay accuracy of the current layer pattern photolithography, thereby improving the overlay accuracy.
[0007] To solve the above-mentioned technical problems, the method for controlling the photolithography overlay accuracy by grouping provided by the present invention includes the following steps: Step 1: Obtain multiple overlay maps of the completed lithography pattern for the current layer.
[0008] Step 2: Perform the first classification of the overlay map images by batch.
[0009] Step 3: According to the first classification results, track and calculate the scores of each machine that will affect the overlay accuracy of the current layer pattern lithography corresponding to each first classification. The score is the percentage of wafers that have undergone the previous layer process on the corresponding machine and belong to the classification group corresponding to the first classification. Determine the correspondence between the machine and the corresponding classification group in the first classification through the scores.
[0010] Step 4: When subsequent batches of wafers complete the previous layer process on the corresponding machine, the batches are classified a second time according to the score of the machine.
[0011] Step 5: Before performing the in-layer pattern lithography on the wafers of each batch that have completed the second classification, control the overlay accuracy of the in-layer pattern lithography according to the classification group of the second classification corresponding to the batch.
[0012] A further improvement is that step one also includes performing a machine averaging process on the overlay map to eliminate the influence of the lithography machine corresponding to the current layer pattern lithography.
[0013] A further improvement is that, in step two, the first classification is achieved by performing similarity analysis on each of the overlay map images.
[0014] A further improvement is that the similarity analysis algorithm for the first classification includes either principal component analysis (PCA) or random forest.
[0015] A further improvement is that the closer the score is to 1, the more accurate the one-to-one correspondence between the machine and the corresponding classification group in the first classification.
[0016] A further improvement is that, in step three, the score is the product of the first score and the second score, expressed by the formula: KPI = KPI1 * KPI2; KPI1=(wafer number of each group by tool by step) / (wafer sum by toolby step); KPI2=(wafer number of each group by tool by step) / (wafer sum by groupby step); Where Group represents a classification group in the first classification, step represents a step of the previous process, tool represents the machine corresponding to the completion of the previous process; wafer number represents the number of wafers, and wafer sum represents the total number of wafers. The wafer number of each group by tool by step represents the number of wafers in the corresponding classification group that have undergone the corresponding front-end process and the corresponding machine. The wafer sum by tool by step represents the total number of wafers that have undergone the corresponding previous process and the corresponding machine. "Wafer sum by group by step" represents the total number of wafers in the corresponding classification group that have undergone the corresponding previous process. KPI represents the score; KPI1 represents the first score, which is the ratio of the number of wafers belonging to the corresponding category group of the first category that use the machine to perform the front-end process to the total number of wafers that use the machine to perform the front-end process. KPI2 represents the second score, which is the ratio of the number of wafers in the corresponding category group of the first category that have undergone the previous process using the machine to the total number of wafers in the corresponding category group of the first category that have undergone the previous process.
[0017] A further improvement is that, in step four, the machine can be further unfolded into components corresponding to the machine, and the score corresponds to the components of the machine.
[0018] A further improvement is that step five includes tagging the wafers of each batch that have completed the second classification; then, the overlay accuracy of the current layer pattern lithography of each batch that has completed the second classification is controlled by the R2R system.
[0019] A further improvement is that, in step five, the control includes: setting a Runtime overlay accuracy compensation value corresponding to the classification group corresponding to the second classification of the batch.
[0020] When performing the in-layer pattern photolithography, the overlay accuracy is corrected online using the Runtime overlay accuracy compensation value.
[0021] A further improvement is that the Runtime overlay accuracy compensation value corresponding to each classification group in the second classification is set in the APC system.
[0022] A further improvement is that the Runtime overlay accuracy compensation value is determined by the overlay map corresponding to the machine corresponding to the second classification group of the batch and the first classification group corresponding to the machine, in order to eliminate overlay defects in the corresponding overlay map.
[0023] This invention collects overlay maps of the current layer pattern lithography and performs an initial classification of these maps. Based on this initial classification, reverse tracing and score calculations establish a one-to-one correspondence between the equipment and the corresponding classification groups. This allows for a second classification of wafers in subsequent batches during the corresponding front-end processes, based on the equipment they pass through. Since the relationship between the equipment and the overlay maps corresponding to the first classification groups is one-to-one, this second classification enables pre-control based on the equipment used by the wafer, thus avoiding overlay defects appearing in the overlay maps. Because front-end processes involve multiple processes and are implemented through multiple equipment, this invention overcomes the overlay problems caused by each equipment, ultimately correcting the impact of each front-end process on the overlay accuracy of the current layer pattern lithography, thereby improving overlay accuracy. Attached Figure Description
[0024] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: Figure 1 This is a schematic diagram of the alignment system correction in an existing lithography machine; Figure 2 This is a flowchart of a method for controlling the precision of photolithography overlay in groups according to an embodiment of the present invention; Figure 3 This is the overlay map of each classification group after the first classification in the method for controlling the overlay accuracy of photolithography in the embodiment of the present invention; Figure 4 This is a bar chart showing the scores of each machine corresponding to the classification group in the first classification obtained by the method of grouping and controlling the photolithography overlay accuracy in the embodiments of the present invention. Detailed Implementation
[0025] like Figure 2 The diagram shows a flowchart of a method for controlling photolithography overlay accuracy in groups according to an embodiment of the present invention. The method includes the following steps: Step 1: Obtain multiple overlay maps of the completed lithography pattern for the current layer.
[0026] In this embodiment of the invention, the method further includes performing a de-equipment averaging process on the overlay map to eliminate the influence of the lithography machine corresponding to the current layer pattern. The de-equipment averaging process involves averaging multiple corresponding overlay map images that have undergone lithography. Each overlay map image undergoes the same process and uses the same lithography machine as the previous layer, thus eliminating the influence of fluctuations in lithography machine parameters.
[0027] Step 2: Perform the first classification of the overlay map images by batch.
[0028] In this embodiment of the invention, the first classification is achieved by performing similarity analysis on each of the overlay map images. Preferably, the similarity analysis algorithm for the first classification includes the PCA algorithm or the random forest algorithm.
[0029] like Figure 3 The image shown is an overlay map of each group after the first classification in the method for controlling the overlay accuracy of photolithography in an embodiment of the present invention; Figure 3 It can be seen that the first classification divides the overlay map into three classification groups, namely G1, G2 and G3.
[0030] Step 3: According to the first classification results, track and calculate the scores of each machine that will affect the overlay accuracy of the current layer pattern lithography corresponding to each first classification. The score is the percentage of wafers that have undergone the previous layer process on the corresponding machine and belong to the classification group corresponding to the first classification. Determine the correspondence between the machine and the corresponding classification group in the first classification through the scores.
[0031] In this embodiment of the invention, the closer the score is to 1, the more accurate the one-to-one correspondence between the machine and the corresponding classification group in the first classification.
[0032] In some embodiments, the score is the product of the first score and the second score, expressed by the formula: KPI = KPI1 * KPI2; KPI1=(wafer number of each group by tool by step) / (wafer sum by toolby step); KPI2=(wafer number of each group by tool by step) / (wafer sum by groupby step); Where Group represents a classification group in the first classification, step represents a step of the previous process, tool represents the machine corresponding to the completion of the previous process; wafer number represents the number of wafers, and wafer sum represents the total number of wafers. The wafer number of each group by tool by step represents the number of wafers in the corresponding classification group that have undergone the corresponding front-end process and the corresponding machine. The wafer sum by tool by step represents the total number of wafers that have undergone the corresponding previous process and the corresponding machine. "Wafer sum by group by step" represents the total number of wafers in the corresponding classification group that have undergone the corresponding previous process. KPI represents the score; KPI1 represents the first score, which is the ratio of the number of wafers belonging to the corresponding category group of the first category that use the machine to perform the front-end process to the total number of wafers that use the machine to perform the front-end process. KPI2 represents the second score, which is the ratio of the number of wafers in the corresponding category group of the first category that have undergone the previous process using the machine to the total number of wafers in the corresponding category group of the first category that have undergone the previous process.
[0033] like Figure 4 The figure shown is a bar chart of the scores of each machine corresponding to the classification group in the first classification obtained by the method of group control of photolithography overlay accuracy in an embodiment of the present invention. Figure 4In the diagram, dashed boxes 201a and 200 each correspond to a classification group in the first classification. The bars in the histogram represent the scores of the corresponding machines. It can be seen that in dashed box 201a, the bars corresponding to the two machines in box 202 are both relatively high, meaning their scores are large. Therefore, the two machines in box 202 correspond to the classification group corresponding to dashed box 201a. In dashed box 201b, the bar corresponding to one machine in box 203 is relatively high, and its score is close to 1. Therefore, one machine in box 203 corresponds to the classification group corresponding to dashed box 201b. The correspondence between the machines in dashed box 201b and the classification bars is more precise.
[0034] Step 4: When subsequent batches of wafers complete the previous layer process on the corresponding machine, the batches are classified a second time according to the score of the machine.
[0035] In some embodiments, the machine tool can be further unfolded into components corresponding to the machine tool, and the score corresponds to the component of the machine tool.
[0036] Step 5: Before performing the in-layer pattern lithography on the wafers of each batch that have completed the second classification, control the overlay accuracy of the in-layer pattern lithography according to the classification group of the second classification corresponding to the batch.
[0037] In this embodiment of the invention, step five further includes tagging the wafers of each batch that have completed the second classification; then, the overlay accuracy of the current layer pattern lithography of each batch that has completed the second classification is controlled by an R2R system.
[0038] The control includes setting a Runtime overlay accuracy compensation value corresponding to the classification group corresponding to the second classification of the batch.
[0039] The Runtime overlay accuracy compensation value is determined by the machine corresponding to the second classification group of the batch and the overlay map corresponding to the first classification group of the machine, in order to eliminate overlay defects in the corresponding overlay map.
[0040] In some embodiments, the Runtime overlay accuracy compensation value corresponding to each classification group of the second classification is set in the APC system.
[0041] When performing the in-layer pattern photolithography, the overlay accuracy is corrected online using the Runtime overlay accuracy compensation value.
[0042] This invention collects overlay maps of the current layer pattern lithography and performs a first classification of these maps. Based on this first classification, reverse tracing and score calculations reveal a one-to-one correspondence between the equipment and the corresponding classification groups. This allows for a second classification of wafers in subsequent batches during the corresponding front-end processes, based on the equipment they pass through. Since the relationship between the equipment and the overlay maps corresponding to the first classification groups is one-to-one, this second classification enables pre-control based on the equipment used by the wafer, thus avoiding overlay defects appearing in the overlay maps. Because front-end processes involve multiple processes and are implemented through multiple equipment, this invention overcomes the overlay problems caused by each equipment. Therefore, it can ultimately correct the impact of each front-end process on the overlay accuracy of the current layer pattern lithography, thereby improving overlay accuracy.
[0043] This invention enables front-end process tracking and grouping of different batches of lots. After processing the incoming overlay map of the current lithography layer, the similarity of the incoming overlay map is analyzed and classified using PCA or Random Forest algorithms. Based on the KPIs of a custom issue wafer, various lots are grouped to identify the front-end process steps of the issue, thereby achieving targeted grouping and control of the overlay accuracy of the current layer. Furthermore, it enables online runtime overlay accuracy feedback, reducing the probability of online rework, saving costs, and increasing production capacity.
[0044] Taking the photolithography process of the active metal zero layer (MOA) as an example, after removing the influence of the photolithography equipment on the incoming overlay map of the MOA, the similarity of the incoming overlay map is analyzed and classified using the PCA algorithm. Based on the custom issue wafer KPI for polysilicon etching equipment, and further considering the lifetime of etching equipment components such as edge rings, various lots are grouped to achieve targeted control of the overlay accuracy of the current layer. The steps are as follows: 1) First, process the incoming overlay map by removing the average overlay map from the machine.
[0045] 2) Since the wafer overlay differences mainly originate from the edge region, the processed edge overlay map is classified into lots using the PCA algorithm, which is the first classification.
[0046] 3) Calculate the machine score for the Poly etching process step using a custom KPI. The formula is: KPI = KPI1 * KPI2; KPI1=(Wafer num of each group by tool by step) / (Wafer sum by tool bystep), KPI2=(Wafer num of each group by tool by step) / (Wafer sum by group bystep).
[0047] 4) Based on the calculated scores of each etching process machine, the lot is pre-classified, i.e., the second classification.
[0048] 5) Based on the grouping results of the second classification, set the supplementary values in advance in the APC system.
[0049] The present invention has been described in detail above through specific embodiments, but these are not intended to limit the invention. Many modifications and improvements can be made by those skilled in the art without departing from the principles of the invention, and these should also be considered within the scope of protection of the present invention.
Claims
1. A method for controlling the precision of photolithographic overlay in groups, characterized in that, Includes the following steps: Step 1: Obtain multiple overlay maps of the completed lithography pattern for the current layer; Step 2: Perform the first classification of the overlay map images by batch; Step 3: According to the first classification results, track and calculate the scores of each machine that affects the overlay accuracy of the current layer pattern lithography corresponding to each first classification. The score is the percentage of wafers that have undergone the previous layer process on the corresponding machine and belong to the classification group corresponding to the first classification. Determine the correspondence between the machine and the corresponding classification group in the first classification based on the scores. Step 4: When subsequent batches of wafers complete the previous layer process on the corresponding machine, the batches are classified a second time according to the score of the machine. Step 5: Before performing the in-layer pattern lithography on the wafers of each batch that have completed the second classification, control the overlay accuracy of the in-layer pattern lithography according to the classification group of the second classification corresponding to the batch.
2. The method for controlling the precision of photolithography overlay in groups as described in claim 1, characterized in that: Step one also includes performing machine averaging processing on the overlay map to eliminate the influence of the lithography machine corresponding to the current layer pattern lithography.
3. The method for controlling the photolithography overlay accuracy by grouping as described in claim 2, characterized in that: In step two, the first classification is achieved by performing similarity analysis on each of the overlay map images.
4. The method for controlling the photolithography overlay accuracy by grouping as described in claim 3, characterized in that: The similarity analysis algorithm for the first classification includes either the PCA algorithm or the random forest algorithm.
5. The method for controlling the photolithography overlay accuracy by grouping as described in claim 2, characterized in that: The closer the score is to 1, the more accurate the one-to-one correspondence between the machine and the corresponding classification group in the first classification.
6. The method for controlling the precision of photolithography overlay in groups as described in claim 5, characterized in that: In step three, the score is the product of the first score and the second score, expressed by the formula: KPI = KPI1 * KPI2; KPI1=(wafer number of each group by tool by step) / (wafer sum by tool bystep); KPI2=(wafer number of each group by tool by step) / (wafer sum by group bystep); Where Group represents a classification group in the first classification, step represents a step of the previous process, tool represents the machine corresponding to the completion of the previous process; wafer number represents the number of wafers, and wafersum represents the total number of wafers. The wafer number of each group by tool by step represents the number of wafers in the corresponding classification group that have undergone the corresponding front-end process and the corresponding machine. The wafer sum by tool by step represents the total number of wafers that have undergone the corresponding previous process and the corresponding machine. "Wafer sum by group by step" represents the total number of wafers in the corresponding classification group that have undergone the corresponding previous process. KPI represents the score; KPI1 represents the first score, which is the ratio of the number of wafers belonging to the corresponding category group of the first category that use the machine to perform the front-end process to the total number of wafers that use the machine to perform the front-end process. KPI2 represents the second score, which is the ratio of the number of wafers in the corresponding category group of the first category that have undergone the previous process using the machine to the total number of wafers in the corresponding category group of the first category that have undergone the previous process.
7. The method for controlling the photolithography overlay accuracy by grouping as described in claim 2, characterized in that: In step four, the machine can be further unfolded into components corresponding to the machine, and the score corresponds to the component of the machine.
8. The method for controlling the precision of photolithography overlay in groups as described in claim 2, characterized in that: Step five also includes tagging the wafers of each batch that have completed the second classification; then, the overlay accuracy of the current layer pattern lithography of each batch that has completed the second classification is controlled by the R2R system.
9. The method for controlling the precision of photolithography overlay in groups as described in claim 8, characterized in that: In step five, the control includes: setting the Runtime overlay accuracy compensation value corresponding to the classification group corresponding to the second classification of the batch; When performing the in-layer pattern photolithography, the overlay accuracy is corrected online using the Runtime overlay accuracy compensation value.
10. The method for controlling the precision of photolithography overlay in groups as described in claim 9, characterized in that: The Runtime overlay accuracy compensation value corresponding to each classification group in the second classification is set in the APC system.
11. The method for controlling the precision of photolithography overlay in groups as described in claim 9, characterized in that: The Runtime overlay accuracy compensation value is determined by the machine corresponding to the second classification group of the batch and the overlay map corresponding to the first classification group of the machine, in order to eliminate overlay defects in the corresponding overlay map.