Open Access
Add to BibSonomyAbstract
Angular scatterometry, which has the advantage of good measurement precision, is an optical measurement technology based on the analysis of light scattered from periodic features, such as a linear grating, and is proposed as an alternative solution for overlay metrology. We present overlay measurements using an angular scatterometer and a bright-field microscope. A theoretical library based on rigorous coupled wave theory was created, and the reflected signatures measured by angular scatterometer were matched to the library to obtain structure parameters, including overlay, critical dimension (CD), and sidewall angle at the same time. The results reveal that angular scatterometer has a good precision and low tool-induced shift for overlay measurements, and has the potential for integrated metrology.
©2006 Optical Society of America
1. Introduction
Integrated metrology (IM) has been a rising trend for advanced process control (APC) in semiconductor manufacturing in recent years. Integrated metrology is basically defined as the coupling of a measurement tool to a process tool for the measurement of multiple parameters. In an ideal integration scheme, different types of measurements, like overlay and critical dimension (CD), can be obtained from the same measurement tool. Ausschnitt [1] proposed a MOXIE (Metrology Of eXtremely Irrational Exuberance) target, which consisted of one set of grating targets for CD and thin film measurements and another set of differential CD targets for overlay measurements. Lensing et al. [2] integrated an optical reflectometer into a clean track system to meet the requirements for lithography cell integration and process CD control of Flash memory applications. Kota et al. [3] and Mui et al. [4] proposed the integration of an optical CD measurement system and etch system for real-time feedforward and feedback measurements on etch processes. Advanced process control benefits from the development of integrated metrology, which involves creating multi-parameters metrology systems or incorporating high-throughput metrology systems into the standard flow of semiconductor process equipment, thus realizing a significant improvement over run-to-run process control.
This is the first time that angular scatterometer was applied in an overlay measurement. Angular scatterometer, which is an optical metrology based on the analysis of light scattered from periodic features, takes advantage of good repeatability. When a series of periodic features are illuminated by a light source, the reflectance properties of the scattered/diffracted light depend on the structure and composition of the features. Analyzing the scatter signatures allows the shape and dimensions of periodic features to be determined [5,6]. For a two-layer linear grating, the overlay error between the two layers can also be determined.
This work investigates the use of overlay measurement by using an angular scatterometer, which employed a two-layer linear grating as overlay target. In order to make a comparison with current mainstream method of overlay measurement - bright-field microscope - a bar-in-bar overlay target, which is commonly used in that technology, was also manufactured. The results reveal that the precision of overlay measurement using angular scatterometer was only about 0.2 nm, which is superior to the approximate 0.5 nm precision results that were obtained by the bright-field microscope. According to ITRS 2005 report, the requirement of overlay metrology precision is 0.9 nm for 50nm node in 2009. Moreover, the angular scatterometer was composed of less optical components, resulting in an obvious improvement of tool induce shift (TIS). With these high precision and low TIS, angular scatterometer has the potential for integrated metrology of CD and overlay measurement when a two-layer grating target is used.
2. Theoretical model
Rigorous coupled wave theory [7–9] is the most popular theory to analyze the electromagnetic field scattering from a periodic structure. In this work we summarize Chateau’s algorithm, which describes the system in the matrix form
where fF, bF, and fL are the complex amplitude of the incident beam, reflected beam, and transmitted beams, respectively, and bL is the complex amplitude of the reflected beam from the transmitted region. Generally, bL is set to be zero. [C(z0)] and [C(zm)] are the interface matrices in the incident region and transmitted region. [Gl] is the characteristic matrix of the lth layer in the grating region, and is defined as
where [Pl] and [Dl] are the eigenvectors and the diagonal matrix of the eigenvalues, respectively, in the lth layer of the grating region. The periodic modulation of the grating layer is represented by the Fourier expansion
For a non-offset grating layer, ñ is defined as
where n1 and n2 are the refractive index of line and space of grating, respectively, and f is the line-to-space ratio. But for offset grating layer, this definition should be amended as
where ΔOL is the offset value, and Λ is the grating period.
3. Numerical results
3.1 Description of the stacked grating target
Table 1. Stacked Grating Parameters
Figure 1 shows the cross section of the overlay target for the angular scatterometer. The stacked structure was manufactured as: photo-resist (top layer)/poly-silicon (interlayer)/SiO2 (bottom layer)/silicon (substrate). An intentional offset was included between the top grating and bottom grating layers, as noted in section 3.2. Table 1 lists the stacked grating and the refractive index of each material, which were measured using a commercial n&k measurement instrument.
3.2 Target design for overlay metrology
Fig. 2. Reticle design (a) Linear grating targets in both the X and Y directions (b) Configuration of reticle.
Figure 2(a) illustrates the linear grating targets in both the X and Y directions. The target size was 85µm×60µm and the distance between these two targets were 20µm. Figure 2(b) illustrates the configuration of the reticle. Each module, which is enlarged to the right side of the figure, contains numerous linear grating targets with various line-to-space ratios ranging from 1:5 to 5:1, including 1:1, and one bar-in-bar target placed in the center. An intentional overlay offset Δ was designed, and ranged from 0 nm to 800 nm with an increment of 50 nm.
3.3 Simulation results
Figure 3 shows the simulation results based on our rigorous coupled wave theory (RCWT) model. RCWT, which is derived from Maxwell’s equations, considers the propagating vector and the polarization state. The stacked grating used in this simulation result is photo-resist/poly-silicon/SiO2/silicon with a given pitch of 800 nm and line-to-space ratio of 1:1. In this work, only the zeroth reflected order is considered. The diffraction efficiency of the reflected light, also known as reflectance, changes as the incident angle varies, and is symmetric to positive and negative incident angles. Moreover, according to the symmetric characteristic of the stacked structure, the reflectance signatures are also symmetric to overlay offset centered at the 1/2 pitch (400nm). As Fig. 3 illustrates, as the overlay offset “OL” between the top and bottom grating layer changes, the diffraction efficiency of the zeroth order reflected light also changes.
4. Experimental results
4.1 Measurement instrument
The angular scatterometer system, depicted as Fig. 4, uses a focused He-Ne laser source incident on the overlay grating target, and measured the zeroth order reflected light by varying the incident angle and reflected angle simultaneously. The reflected light was separated into S-polarized and P-polarized light using a polarizing beam splitter, which were then measured by two different detectors. This angular scatterometer is also known as a 2θ system for the equality of incident and detection angles. The scan angle was from -30° to +30°. The image-based overlay measurements were performed using a commercial bright field overlay instrument with a numerical aperture of 0.5 and a green color filter.
4.2 Experimental signature
Previous studies [10,11] indicated scatterometry overlay measurements have higher sensitivity when a nominal overlay offset of 1/4 pitch is introduced. In this study, five targets with different intentional overlay offsets were selected. The intentional overlay offsets bracketed the optimal 1/4 pitch condition and were 100 nm, 150 nm, 200 nm, 250 nm, and 300 nm, and these five targets were denoted as Target1, Target2, Target3, Target4, and Target5, respectively. In order to simplify the parameters of the grating target while creating the theoretical library, the thicknesses of the photo-resist, poly-silicon, and SiO2 were fixed to the results obtained from a thin film measurement instrument and AFM. Figure 5 shows the measured signatures and theoretical signatures when matched to the theoretical library for both measured polarizations. The %RMSE between the measured and matched signatures is below 8%, which is good for a two-layer grating target. Table 2 lists the library match results of the selected overlay targets, including the overlay offsets between the bottom and top layer gratings, the CD and sidewall angles of both the bottom and top layer gratings, and the thickness of each layer. The overlay offsets of these five targets were 106.2nm, 161.8nm, 208.7nm, 254.3nm, and 302.8nm, respectively, which reasonably approaches to the designed values. The small variations of CD and sidewall angle are considered to be within the tolerances of manufacturing the samples.
Table 2. Library Match Results
4.3 Comparisons between angular scatteroemeter and bright-field microscope
The bright-field microscope was used to measure the bar-in-bar target, which was located approximately 350 µm from the grating target. Ten static repeats were performed at each grating and bar-in-bar target at stage orientations of both 0° and 180°. The 3σ overlay precision, , and TIS 3σ of the measurement systems were calculated as
where OLn(0o) and OLn(180o) are the nth measured overlay offset at stage orientations of 0° and 180°, respectively. is the average of OLn,n=1~10, and is the average of TISn,n=1~10. Table 3 lists the overlay offset, precision 3σ, , and TIS 3σ of the angular scatterometer and bright field microscope. The average overlay offset of both measurement instruments has good agreement with a difference of within 2 nm. Figure 6 shows the two systems have a linear relation with the slope of 1.02, the offset of 5.02 nm, and the correlation of 1.00. Figure 7 shows the comparisons of histograms of the measurement results at different intentional overlay targets. When averaged across all five targets, the 3σ overlay precision of the angular scatterometer, is 0.2 nm, while that of the bright-field microscope is 0.5 nm. The average TIS and TIS 3σ of the angular scatterometer was 0.20 nm and 0.05 nm, respectively, which for both metrics is approximately an order of magnitude less than that of the bright-field microscope.
Table 3. Measurement Results
Fig. 7. Precision 3σ, and TIS 3σ of overlay measurement using angular scatterometer and bright-field microscope
5. Summary
Integrated Metrology is in its infancy, and many related technologies are still under development. In the predictable future, integrated metrology will have a significant impact on process monitoring, control and yield. This study demonstrates an approach using an angular scatterometer, which has previously been applied for CD measurement. For practical overlay metrology, the results reveal that angular scatterometer had better precision and TIS performance than a conventional bright-field microscope. Based on these findings, angular scatterometry has the potential for integrated metrology of both CD and overlay measurements in the semiconductor fab. New algorithms are now being developed for enhancing this multi-parameter metrology, and will be expected to further improve the measurement capabilities.
Acknowledgments
The authors would like to thank Dr. Nigel Smith of Accent Optical Technologies for assistance in collecting the bright-field microscope overlay measurement data.
Reference and links
1. C. P. Ausschnitt, “A New Approach to Pattern Metrology,” Proc. SPIE 5375, 51–65 (2004). [CrossRef]
2. K. Lensing, C. Miller, G. Chudleigh, B. Swain, M. Laughery, and A. Viswanathan, “Scatteromtry Feasibility Studies for 0.13-Micron Flash Memory Lithography Applications; Enabling Integrated Metrology,” Proc. SPIE 5375, 307–316 (2004). [CrossRef]
3. G. P. Kota, J. Luque, V. Vahedi, A. Khathuria, T. Dziura, and A. Levy, “Advanced process control for polysilicon gate teching using integrated optical CD metrology,” Proc. SPIE 5044, 90–96 (2003). [CrossRef]
4. D. S. L. Mui, H. Sasano, W. Liu, J. Yamartino, and A. Skumanich, “In-tool process control for advanced patterning based on integrated metrology,” Proc. SPIE 5378, 10–17 (2004). [CrossRef]
5. C. J. Raymond, M. R. Murnance, S. L. Prins, S. S. H. Naqvi, John R. McNeil, and J. W. Hosch, “Multi-parameter CD measurement using scatterometry,” Proc. SPIE 2725, 698–709 (1996). [CrossRef]
6. C. J. Raymond, M Littau, B Youn, C. J. Sohn, J. A. Kim, and Y. S. Kang, “Applications of angular scatterometry for the measurement for multiply-periodic features,” Proc. SPIE 5038, 577–584 (2003). [CrossRef]
7. M. G. Moharam and T. K. Gaylord, “Rigorous coupled-wave analysis of planar-grating diffraction,” J. Opt. Soc. Am. , 71, 811–818 (1981). [CrossRef]
8. M. G. Moharam, E. B. Grann, and D. A. Pomment, “Formualtion for stable and efficient implementation of the rigorous coupled-wave analysis of binary gratings,” J. Opt. Soc. Am. A , 12, 1068–1076 (1995). [CrossRef]
9. N. Chateau and J. P. Hugonin, “Algorithm for the rigorous coupled-wave analysis of grating diffraction,” J. Opt. Soc. Am. A , 11, 1321–1331 (1994). [CrossRef]
10. W. Yang, R. L. Webb, S. Rabello, J. Hu, J. Y. Lin, J. Heaton, M. Dusa, A. Boef, M. Schaar, and A. Hunter,” A novel diffraction based spectroscopic method for overlay metrology,” Proc. SPIE 5038, 200–207 (2003). [CrossRef]
11. H. T. Huang, G. Rafhavendra, A. Sezginer, K. Johnson, F. Stanke, M. Zimmerman, C. Cheung, M. Miyagi, and B. Singh, “Scatterometry-Based Overlay Metrology,” Proc. SPIE 5038, 126–137 (2003). [CrossRef]
