Apply multiple target for advanced gate ADI critical dimension measurement by scatterometry technology

Transcription

1 Apply multiple target for advanced gate ADI critical dimension measurement by scatterometry technology Wei-Jhe Tzai a ; Howard Chen a ; Yu-Hao Huang a ; Chun-Chi Yu a ; Ching-Hung Bert Lin b ; Shi-Ming Jeremy Wei b ; Zhi-Qing James Xu b ; Sungchul Yoo b ; Chien-Jen Eros Huang b ; Chao-Yu Harvey Cheng b ; Juli Cheng b ; Lanny Mihardja b ; Houssam Chouaib b a United Microelectronics Corp. (UMC); Tainan, Taiwan, R.O.C. b KLA-Tencor Corporation; one Technology drive, Milpitas, CA 95035, U.S.A. ABSTRACT Scatterometry-based metrology measurements for advanced gate after-develop inspection (ADI) and after-etch inspection (AEI) structures have been well proven 1. This paper discusses the metrology challenges encountered in implementing a production-worthy methodology for accurately measuring gate ADI middle CD (MCD) and sidewall angle (SWA) to monitor focus and exposure dose. A Multi-Target Measurement (MTM) methodology on KLA-Tencor s SpectraShape 8810 was evaluated on its ability to characterize and measure FEM (Focus Exposure Matrix) and EM (Exposure Matrix) wafers. The correlation of MCD and SWA to the focus and exposure dose was explored. CD-SEM measurements were used as a reference to compare the accuracy of scatterometry MCD measurements. While there was no reference tool available to compare scatterometry SWA measurements, the SWA and focus tracking on the FEM wafer were verified. In addition to the MTM methodology evaluation, a fleet of four SpectraShape 8810 tools was evaluated to measure the fleet s capability for in-line monitoring in high volume manufacturing. The final results confirmed that the Multi-Target Measurement approach on SpectraShape 8810 is an effective solution for gate ADI metrology and the robust fleet matching performance would enable in-line monitoring use. Keywords: Scatterometry, MTM, Multi-Target Measurement, Sensitivity, FEM, Focus Exposure Matrix, EM, Gate ADI, Parameter Correlation 1. INTRODUCTION Scatterometry Critical Dimension (SCD) technology has been widely adopted as a CD and shape process control solution for today s advanced design nodes. In this study, SCD applications for front end of line (FEOL) 28nm gate ADI FEM and EM wafers were explored. One of the currently existing SCD methods to measure and monitor gate ADI is to separately measure MCD on the dense grating target to monitor the exposure modulation, and to perform another measurement for SWA on the isolated grating target to monitor the focus modulation. Simultaneous MCD and SWA measurement usually yields inaccurate results due to the limited sensitivity of and the high parameter correlation between the parameters of interest. In this paper we first evaluated the Multi-Target Measurement (MTM) method available on the SpectraShape 8810 to address the challenges associated with simultaneous MCD and SWA measurements. Our data analyses provided solid evidence that this method is highly effective in breaking the strong parameter correlation and in providing the unique sensitivity needed for gate ADI applications. Secondly, we verified the accuracy of the MTM results with CD-SEM reference data. The analysis showed that the SCD result has high correlation to CD-SEM reference data and less error in the measurement data than CD-SEM. Finally, tool fleet performance was evaluated to assess SpectraShape 8810 fleet capability for gate ADI in-line monitoring. The data showed that MTM and superior tool-to-tool matching performance provides the robustness needed in high volume manufacturing. Metrology, Inspection, and Process Control for Microlithography XXVI, edited by Alexander Starikov, Proc. of SPIE Vol. 8324, SPIE CCC code: X/12/$18 doi: / Proc. of SPIE Vol

2 In summary, as modern semiconductor device feature dimensions become more and more critical, they become increasingly more challenging to control. Effectively monitoring and controlling the gate ADI process reduces gate AEI process variation and improves device performance. KLA-Tencor s SpectraShape 8810 scatterometry-based CD and shape metrology system with MTM capability provides an in-line monitoring solution for gate ADI high volume manufacturing because of its sensitivity, robust parameter de-correlation, and the non-destructive nature of the measurements. 1.1 Scatterometry Based Metrology Scatterometry Critical Dimension (SCD) metrology tools are a widely adopted metrology solution for 90, 65, 45, and 40nm poly gate devices at the gate ADI and AEI process levels. SCD is utilized as an advanced process control (APC) system to monitor process variations with higher accuracy compared to a traditional CD-SEM tool. A new-generation SCD tool, the SpectraShape 8810, uses broadband light (down to wavelengths in the extreme UV portion of the spectrum 2 ),multi-azimuth angles (AZ), spectroscopic ellipsometer (SE) technology, and enhanced ultra-violet reflectivity (euvr) to enhance critical parameters sensitivities and reduce the correlation of the parameters. With the advancement of semiconductor processes and the shrinkage of node size below 28nm, metrology measurements for poly gate levels are becoming more and more challenging. The smaller physical size of the structures lowers the sensitivity signals and the more complex underlayer film stacks interfere with the already limited sensitivity to cause high parameter correlation. One of the measurement methods available on the SpectraShape 8810 is Multi-Target Measurement (MTM). The concept of MTM is to utilize the unique sensitivity of different measurement targets at different wavelength to break the strong parameter correlation. The spectral response collected at the different targets are analyzed simultaneously to extract the unique sensitivities of each parameter of interest, while at the same time constraining them based on the known parameter constraints on each target. In this study, the capability of MTM to improve and optimize gate ADI metrology applications was explored. Figure 1 shows the schematic overview of the SCD MTM measurement flow. Spectral responses from three targets dense grating, iso grating, and film pad are collected and then analyzed simultaneously. Figure 2 shows the unique sensitivity signals obtained at different wavelengths from different targets. Proc. of SPIE Vol

3 Fig. 1: Schematic overview of SCD MTM methodology flow. Figure 2: The overview of the SCD MTM concept. Proc. of SPIE Vol

4 2. GATE ADI PARAMETERS AND MEASUREMENT RESULTS 2.1 Gate ADI MCD and SWA Analyses on FEM Wafer In-line monitoring of gate poly critical dimensions at the after-develop inspection (ADI) process layer is critical to device performance 3. To study the sensitivity of scatterometry measurements for the gate ADI layer, two design-ofexperiment (DOE) wafers were prepared and measured after the ADI process step on the SpectraShape One wafer was exposed to Focus Exposure Matrix (FEM) lithography conditions and the second wafer was only exposed to Exposure Matrix (EM) lithography conditions. For the same DOE conditions the PMOS and NMOS structure targets were created. The detailed layer structures of the CD gratings and film pad are shown in figure 3. The parameters of interest are listed in table 1. L8 L7 L11 L6 L5 L4 L3 L2 Dense and Iso Grating pad Film pad Figure 3. Gate ADI grating and film pad model and stack information Table 1. Parameters of interest and technology used for this model. Proc. of SPIE Vol

5 As shown in figure 3, there are seven different film layers under the resist in this structure. This complex underlayer scheme adds another dimension to the challenge of SCD measurement as the signals of the individual layers interact with each other, resulting in strong parameter correlation between two or more layers in the model. For the FEM wafer, the middle CD (MCD) parameter is an effective parameter to track and monitor the exposure or energy modulation, while the sidewall angle (SWA) tracks the focus modulation. In our test, we first used a single-target SCD measurement approach to measure the critical parameters for gate ADI. Two separate SCD measurements were performed to solve for MCD and SWA: one on the dense grating target and the other on the iso grating target. From the dense grating measurement, the MCD measurement results correctly tracked the exposure level, but the SWA measurement results provided an incorrect focus trend. From the iso grating measurement, the SWA measurement results correctly tracked the focus level, but the MCD results did not correctly track the energy modulation. From these results, it was clear that a single-target measurement approach only provided selective sensitivity for either MCD or SWA and not both at the same time. Figure 4 shows the SCD measurement results from the FEM wafer using single-target measurements. Figure 4. A single-target measurement approach provides only selective sensitivity for MCD or SWA on a FEM wafer. Understanding this selective sensitivity for MCD and SWA from different targets provided the key information that all the sensitivity signals are available for all parameters of interest if the inputs from both the dense and the iso gratings are taken into account and analyzed together. Therefore, the use of a Multi-Target Measurement and analysis approach for gate ADI applications is imperative. The methodology explored in this study combined the advanced SCD modeling software AcuShape2 with the analysis of MTM and floating n,k. The multi targets utilized included: dense grating, iso grating, and resist film pad. The critical parameters of interest being solved for were MCD, SWA, and the resist height (HT). The final CD measurement using this methodology confirmed that the simultaneous sensitivity of MCD and SWA was achieved with no parameter correlation. Figure 5 shows wafer maps of MCD and SWA measurements on which the focus and exposure dose trends across the FEM wafer are correctly tracked. Proc. of SPIE Vol

6 Figure 5. Simultaneous accurate MCD and SWA measurements on a FEM wafer using MTM and AcuShape2 on the SpectraShape Further verification of the measurement accuracy was done by measuring the gate structure CD on a reference CD-SEM tool, and generating an SCD and CD-SEM correlation plot. The correlation plot confirmed high correlation of the SCD results to the CD-SEM measurement, meaning the SCD results are accurate. Furthermore, Bossung curves of SCD and CD-SEM measurements were generated and analyzed. Bossung curves from SCD measurements track both focus and exposure dose modulation with less error than the CD-SEM measurements. Figure 6 displays the correlation plot of SCD and CD-SEM measurements and the Bossung curve comparison between the two measurement tools. Figure 6. SCD and CD-SEM correlation plot and Bossung curves from measurements made on the FEM wafer. Proc. of SPIE Vol

7 In the next verification step, another FEM wafer was prepared with a larger focus and exposure modulation step across the wafer. In the previous FEM wafer test, the wafer was prepared with a relatively small focus and exposure modulation step across the wafer. In this new FEM wafer, the expected Bossung curve trend for the dense grating is U-shaped, while the expected Bossung curve trend for the iso grating is upside-down U-shaped. This trend curve is sometimes referred to as a rainbow curve. As part of this verification step, the measurements of SCD and CD-SEM were again compared and analyzed. The results (figure 7) show that SCD tracked the exposure modulation correctly for both the dense and iso gratings on NMOS and PMOS targets. Figure 7. SCD tracks the expected Bossung curves for a different FEM wafer with larger modulation steps across the wafer. 2.2 Gate ADI CD and SWA Analysis on EM Wafer A similar test suite was performed on an EM wafer to further verify the SCD measurement robustness and accuracy. On an EM wafer, the parameter of interest MCD tracks the exposure modulation trend across the wafer. There is no focus modulation on an EM wafer, therefore a SWA measurement should have negligible variation across the wafer, showing no clear trends other than those related to process noise. The test result supported these expectations; the statistical distribution of the SWA measurement across the wafer was negligible (SWA mean across the wafer is degree, 3-sigma 0.5%, and range 0.3%) and within the expected process variation tolerance. The MCD measurement clearly indicated the expected exposure modulation across the wafer. Similar to FEM wafer test, the SCD and CD-SEM correlation plot and Bossung curves of the EM wafer were also generated to verify SCD measurement accuracy. Figures 8 shows the SCD wafer maps from measurements of the EM wafer. The SCD/CD-SEM correlation plot and Bossung curves are shown in figure 9. Finally, the SCD and CD-SEM correlation values for the exposure or energy trend were analyzed and summarized for both the FEM and EM wafers (figure 10). The correlation values R 2 for the SCD measurements were higher than that of the CD-SEM measurements, confirming that the SCD measurements contained less error than CD-SEM measurements. Proc. of SPIE Vol

8 Figure 8. SCD wafer maps of MCD and SWA for an EM wafer. Figure 9. SCD and CD-SEM correlation plot and Bossung curves of an EM wafer. Proc. of SPIE Vol

9 Figure 10. SCD and CD-SEM correlation with exposure (or energy). 2.3 SCD Tool Fleet Matching Capability As advanced node processes progress from R&D to pilot and eventually high volume manufacturing, the need for a robust metrology tool fleet is essential. Tool-to-tool matching performance is required to be within the process control specification threshold to avoid negative impact to the fleet data integrity and APC monitoring. In order to evaluate the tool fleet performance for gate ADI SCD applications, four SpectraShape 8810 tools were evaluated in this study. The same set of FEM and EM wafers were used to analyze the tool-to-tool matching performance. The first analysis was done by performing MCD and SWA site-by-site measurements and comparing the site-by-site results across the four tools. The results showed that the four tools highly matched with each other on both FEM and EM wafer measurements with a range (or bias) lower than 0.09 nm for MCD and 0.09 degree for SWA. Figures 11 and 12 display the FEM and EM wafers site-by-site measurement results for all four tools, and table 2 presents the measurement bias for the tool fleet. Figure 11. FEM wafer site-by-site measurement results from four SpectraShape 8810 tools. Proc. of SPIE Vol

10 Figure 12. EM wafer site-by-site measurement results from four SpectraShape 8810 tools. SpectraShape 8810 Tool Fleet Matching Analysis Fleet FEM wafer EM wafer MCD (nm) SWA (degree) MCD (nm) SWA (degree) Tool Tool Tool Tool Range Table 2. Tool fleet matching analysis. In order to complete the tool fleet matching analysis, the MCD and SWA were measured across both FEM and EM wafers to generate the trend wafer maps. Once again, the fleet demonstrated superior tool-to-tool matching performance. Figures 13 and 14 display the FEM trend wafer maps and EM trend wafer maps respectively. From these studies, we can conclude that the SpectraShape 8810 fleet has the performance robustness required for in-line gate ADI monitoring in high volume manufacturing. Figure 13. FEM wafer maps of MCD and SWA measured by four SpectraShape 8810 tools. Proc. of SPIE Vol

11 Figure 14. EM wafer maps of MCD and SWA measured by four SpectraShape 8810 tools. 3. SUMMARY As semiconductor structures become increasingly complex, they require ever more sophisticated metrology tools for characterization and process control. Highly flexible SCD measurement methods target the sensitivity of critical parameters of interest and achieve the required precision and accuracy for tracking process trends. Robust fleet matching performance is crucial for high volume manufacturing. For advanced node gate ADI metrology measurements and process monitoring, the KLA-Tencor SpectraShape 8810 demonstrated superior performance compared to the CD-SEM reference tool, and met the required capabilities through Multi-Target Measurement (MTM), AcuShape2 modeling software, and superior tool fleet matching performance. REFERENCES [1] Yu-Hao Huang, Howard Chen, Kyle Shen, H H Chen, Chun Chi Yu, J H Liao, Xiafang(Michelle) Zhang, Russell Teo, Zhi-Qing (James) Xu, Sungchul Yoo, Ching-Hung Bert Lin,Chao-Yu Harvey Cheng, Jason Lin, Scatterometry Measurement for Gate ADI and AEI Critical Dimension of 28nm Metal Gate Technology, Proc. SPIE Vol. 7971, (2011). [2] Matthew Sendelbach, Alok Vaid, Pedro Herrera, Ted Dziura, Xiafang Zhang, Arun Srivatsa, Use of multiple azimuthal angles to enable advanced scatterometry applications, Proc. SPIE Vol. 7638, (2010). [3] Thaddeus G. Dziura, Benjamin Bunday, Casey Smith, Muhammad M. Hussain, Rusty Harris, Xiafang Zhang, Jimmy M. Price, Measurement of high-k and metal film thickness on FinFET sidewalls using Scatterometry, Proc. SPIE Vol. 6922, (2008). [4] Sendelbach, M., Natzle, W., Archie, C. N., Banke, B., Prager, D., Engelhard, D., Ferns, J., Yamashita, A., Funk, M., Higuchi, F., Tomoyasu, M., "Feedforward of mask open measurements on an integrated scatterometer to improve gate linewidth control" in Metrology, Inspection, and Process Control for Microlithography XVIII, edited by Richard M. Silver, Proceedings of SPIE Vol (SPIE, Bellingham, WA 2004) pp , (2004). Proc. of SPIE Vol