Improved scanner matching using Scanner Fleet Manager (SFM)

Transcription

1 Improved scanner matching using Scanner Fleet Manager (SFM) Shian-Huan Cooper Chiu a, Chin-Lung Lee a, Sheng-Hsiung Yu a, Kai-Lin Fu a, Min-Hin Tung a, Po-Chih Chen a ; Chao-Tien Huang b, Chien-Chun Elsie Yu b, Chin-Chou K. Huang c, John C. Robinson* c, David Tien c ; a Rexchip Electronics Corp., No.429-1, Sanfong Rd.,Houli Township, Taichung County, Central Taiwan Science Park, Taiwan; b KLA-Tencor Taiwan, 1F, 22, TaiYuan Street, ChuPei City, HsinChu County 302, Taiwan ; c KLA-Tencor Corp., One Technology Drive, Milpitas, CA 95035; ABSTRACT This project is the continuation of work reported previously at this conference (Yu, et. al., SPIE 2009). A new software tool for developing a static scanner fleet matching (SFM) matrix is tested including fleet snapshot and scanner pair drilldown. In addition the latest scanner models can adjust the distortion performance dynamically, at run-time, improving effective overlay performance of the scanner fleet, and allowing more flexibility for mix-and match exposure. The goal is to improve overlay mean +3s significantly between scanners for critical layer pairs. Keywords: scanner matching, overlay, metrology 1. INTRODUCTION Overlay performance has been a critical factor for advanced semiconductor manufacturing for many years. Over time these requirements become more stringent as design rules shrink. The transition to advanced nodes is requiring significant innovations such as the transition to high order control. A dominant component of the overlay error budget is a result of scanner matching errors, either machine to machine or illumination mode to illumination mode. For most critical layers, scanners are lot-to-lens dedicated, so no mix and match is allowed. For non-critical layers, however, mix and match enables better overall equipment effectiveness. Material can be routed to available machines, and multiple generations of scanners can be utilized. Current methods for fleet management are typically very time consuming. Manually checking scanner mix and match results by the golden tool method is the traditional mode. Once established, a monitor wafer is used to put a scanner back into production. For this process it is desirable to have a software tool to provide improved time to results and more powerful analysis capability. KLA-Tencor Corp. has developed a module in KT Analyzer to provide Scanner Fleet Matching (SFM) capability. The goal is to allow the user to calculate distortion matching across the scanner fleet, and to calculate distortion fingerprints for specific combinations of scanner and illumination modes. The user can group matched scanners together based on the user s spec. Scanners can be from different generations and from different vendors. The steps involved with the analysis are as follows, as shown in Figure 1. First it is necessary to plan out the mix and match scheme. Typically an existing fab scheme is utilized for this analysis. Second, it is necessary to create the PM, or preventative maintenance, lots per plan. Then it is necessary to measure the lots and approve them for use in this analysis. Finally, the data is available for the mix and match analysis. The analysis tool enables a flexible scanner and illumination condition snapshot for fleet matching, and in addition, the ability to look at scanner pair deltas or fingerprints. Metrology, Inspection, and Process Control for Microlithography XXIV, edited by Christopher J. Raymond, Proc. of SPIE Vol. 7638, 76382A 2010 SPIE CCC code: X/10/$18 doi: / Proc. of SPIE Vol A-1

2 Figure 1. SFM work flow: (1) planning, (2) make PM lots, (3) measure & approve PM lots, and (4) group scanners. 2. EXPERIMENT The experiments done in this work involve mix and match between ArF scanners, KrF scanners, and i-line scanners from multiple vendors. The metrology was done with an Archer 100 from KLA-Tencor Corp. and the analysis was done using KT Analyzer s Scanner Fleet Manager (SFM) option. The current method for scanner fleet management involves the golden scanner approach, as shown in Figure 2. For a given group of scanners, matching is established between each tool and the golden scanner only. If the matching is out of spec, then the scanner is shut down until adjustments can be made and it can be established that the scanner is back in spec. Proc. of SPIE Vol A-2

3 Figure 2. Schematic of Golden-tool based scanner fleet management. Each tool for a given group of scanners, such as the green group or the red group, must individually match the golden scanner. The method used for this work involve a more flexible and powerful approach. The SFM method provides a matching matrix to all tools, as shown in Figure 3, which is not the current fab practice. Further, there is no need for additional wafers or metrology steps than are currently being used. If a scanner is found to be out of spec, the Sources of Variation (SOV) analysis tool can be used to quickly troubleshoot the issue and effectively fix the scanner. Figure 3. Matching matrix approach to scanner fleet management. Analysis of all scanner/illumination combinations are possible. Proc. of SPIE Vol A-3

4 The Sources of Variation (SOV) analysis allows the user the ability to quantify wafer and field effects, including linear and high order, as well as un-modeled random contributions, as shown in Figure 4. From this it can quickly be determined which machine and/or illuminations have problems, and what needs to be done to resolve those issues. Figure 4. Sources of Variation (SOV) analysis quantifying wafer and field contributions of linear, high order, and un-modeled random effects. Further, it is then instructive to visualize the wafer and/or field level signatures on a pair-wise basis, as shown for example, in Figure 5. From this the engineer can then decide which issues can and should be addressed. Figure 5. Wafer level signatures for S1 to S2. (1) 2 nd and 3 rd order non-linear contribution, (b) wafer signature. Proc. of SPIE Vol A-4

5 3. SIMULATION RESULTS In this section we explore the simulated results for the improved mix and match method based on existing fab matching data. In this case we find that the scanner field-level matching is worse in the X direction than in Y. Further, the field level matching for the same scanner with different illumination IDs is very good, as shown in Figure 6. Figure 6. Scanner mix and match matrix shown with uncorrected mean + 3 sigma in X-dir. as the metric. The scanner field signature delta is shown for the combination shown in yellow. The circles show same scanner with different illumination IDs. As shown in Figure 7, the red square shows acceptable matching for scanners S2, S3, S4, and S5, however, it is also evident that matching performance between S1 and S6 requires some improvement. Figure 7. Scanner mix and match matrix shown with uncorrected mean + 3 sigma in Y-dir. as the metric. The scanner field signature delta is shown for the combination shown in yellow. The red square shows a suggested matching group. The black squares show potential matching issues. As shown in Figure 8, the yellow shows the scanner matching between S6-ID1 and S6-ID3. Figure 8. Scanner mix and match matrix shown with uncorrected mean + 3 sigma in X-dir as the metric. SFM can provide both Vector and Distortion Frame map of field signature delta. Proc. of SPIE Vol A-5

6 As shown in Figure 9, the yellow shows the scanner matching between S5-ID2 and S6-ID3. Figure 9. Scanner mix and match matrix shown with uncorrected mean + 3 sigma in Y-dir as the metric. The scanner field signature delta is shown for the combination shown in yellow. As shown in Figure 10, the yellow shows the scanner matching between S4-ID1 and S5-ID2. Figure 10. Scanner mix and match matrix shown with uncorrected absolute max X or Y as the metric. The scanner field signature delta is shown for the combination shown in yellow. As shown in Figure 11, the yellow shows the scanner matching between S3-ID1 and S4-ID1. Figure 11. Scanner mix and match matrix shown with uncorrected absolute max X or Y as the metric. The scanner field signature delta is shown for the combination shown in yellow. SFM can provide various and flexible scanner matching results for scanner grouping and scanner management purposes. Proc. of SPIE Vol A-6

7 4. CONCLUSIONS In this study we have seen the advantages of a systematic methodology for scanner fleet management using KLA-Tencor Corp. s KT Analyzer and the Scanner Fleet Manager (SFM) capability. Multiple metrics can be used, enhancing the ability to diagnose scanner issues, and the system supports any scanner vendor or generation. No extra wafers are required beyond current practices, and yet a more comprehensive matching matrix between scanner and scanner illumination combinations can be realized. The benefits include a significant saving of engineering time, the ability to quickly identify scanner issues regarding fleet management, and the potential to increase fab mix and match capability thereby improving overall equipment effectiveness. REFERENCES [1] Shian-Huan Cooper Chiu a ; Sheng-Hsiung Yu a ; Min-Hin Tung a,e ; Lei-Ken Wu a ; Ya-Tsz Yeh a ; James Manka b ; Chao- Tien Healthy Huang b ; John C. Robinson b, Chin-Chou Kevin Huang b ; David Tien b ; YuYu Chen c ; Katsushi Makino d ; Jium-Ming Lin e Improve scanner matching using automated real-time feedback control via scanner match maker (SMM) Proc. SPIE (2009) Proc. of SPIE Vol A-7