Open Access
Add to BibSonomy
Abstract
We present a single-shot spectrally-resolved interferometry for simultaneously measuring the film thickness and surface profile of each layer of a patterned multilayer film structure. For this purpose, we implemented an achromatic phase shifting method based on the geometric phase using the polarization characteristics of the light and obtained four phase-shifted interferograms in the spectrally-resolved fringe pattern at the same time by combining a pixelated polarizing camera with an imaging spectrometer. As a result, we could simultaneously measure the reflectance and phase of the sample over a wide wavelength range with a single measurement. To evaluate the validity of the proposed method, we measured a patterned five-layer film specimen and compared our measurement results with those from commercial instruments, an ellipsometer and a stylus profiler, respectively. We confirmed the results matched each other well.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
With the advent of the fourth industrial revolution, smart factories are increasingly using automated technology solutions, which combine ICT, big data, and sensor components in the manufacturing process. As a result, changes in the industrial paradigm are improving the quality of the product as well as reducing the cost of the manufacturing process. This is possible because the production and inspection processes have been organically connected with each other, reducing product defects, and resulting in a more stable production process. Many types of high-tech sensors and devices, including artificial intelligence solutions, sensor technology for recognition and detection, and equipment for production inspection, are being used not only to inspect the final product, but also in the step-by-step processes. Accordingly, the demand for inspection and measurement equipment is expected to increase explosively across the industry.
In particular, demand for semiconductors is increasing due to the emergence of new markets such as the Internet of Things (IoT), artificial intelligence (AI), virtual reality (VR), and augmented reality (AR). And as the use of artificial intelligence and big data increases, higher performance storage media and higher specification semiconductors will be required. Such advances will significantly increase demand for semiconductor metrology and inspection in the manufacturing process, as the semiconductor products are steadily miniaturized and packaged to meet the various functions and performance standards required by the market. Interest in 3D inspection is increasing, not only respond to rapid changes in product paradigms, but also to perform more stable quality inspection. However, conventional techniques [1–3] such as ellipsometry and reflectometry are basically point-by-point film thickness measurements and are not suitable for 3D inspection tools. Furthermore, many technical challenges remain, as it requires the simultaneous measurement of the film thickness and surface profile of multilayered film specimens that contain complex features stacked in layers.
Most existing 3D measurement techniques are based on optical interferometry, which measures variations in the height of a simple structural surface relative to a reference, by interpreting the optical interference pattern between the object wave and the reference wave [4,5]. For opaque surfaces which lack transmission inside the specimen, the measurement is relatively simple, as the surface profile measurement can be performed using the interference of the light reflected from the surface and the light reflected from the reference. However, with transparent specimens that have transmission and internal reflection, the situation is different. The beams reflected from the upper and lower surfaces of the specimen interfere with the reference beam to form an interference fringe pattern, creating a more complex situation. The beams reflected from the upper and lower layers are not separated from each other and are included in one interference fringe pattern. This can cause serious measurement errors because the light reflected from the surface of the specimen and the light transmitted through the specimen and multiple reflections from the inside of the specimen collide with each other, distorting the existing interference pattern.
This challenge has motivated many researchers and considerable research has been conducted to address these problems during last few decades [6–19]. Fourier-based methods analyzing spectral interference signals measure the surface and thickness profile of a transparent film by separating the interference signals in the frequency domain [6,7]. Modulation-based structured-illumination microscope also measures the surface and thickness profile of thin films by projecting a sinusoidal fringe pattern on the sample and separating the reflected signals from the top and bottom surface of the film [8]. However, these approaches have the minimum measurable limit of the film thickness because the interference signals reflected from both the top and bottom surfaces overlap each other and are not separated when the film thickness becomes very thin [6–8]. To overcome these measurement limitations, most approaches are based on optimization techniques which iteratively change the values of the film and surface parameters until convergence, to find the analytical model that best fits the measured data [9–15]. For these methods, either the phase or reflectance information of the specimen is used for analysis. Other approaches generate vast amounts of interference fringe patterns as signal libraries in advance, and compare these signal libraries directly with the measured interference signal to quickly find the best match between them [16,17]. For improved accuracy, and to accommodate multilayer films, both the reflectance and the phase spectra of a sample have been utilized to measure the surface profile, as well as the film thickness [18,19]. Significant improvements in measurement accuracy and precision were reported when both the phase and reflectance measurement data were considered compared to when only the reflectance data were included in the calculations [18]. However, these approaches perform several measurement steps to obtain the reflectance and phase spectra data over a wide band of spectrum. Wu et al took measurements by changing narrow bandpass filters to obtain interference signals according to each wavelength [18]. Ghim et al obtained a series of phase-shifted interference patterns using a piezoelectric actuator (PZT) to measure the phase spectra of the sample [19]. These methods also employed another measurement step which blocks the reference arm, to measure the reflectance spectra of the sample.
Typically, these approaches require a time-consuming, multi-step measurement process to measure both the reflectance and phase data over a wide wavelength range. Therefore, when such measurements are performed under harsh environmental conditions such as industrial sites, serious measurement errors occur.
In an attempt to solve the measurement problems of the aforementioned methods, we suggest single-shot spectrally-resolved interferometry, which can measure the thickness and surface profile of each layer of a patterned multilayer film structure in real-time. Our proposed method enables us to obtain the reflectance and phase data of a sample over a large range of wavelengths in just a single measurement, so it is immune to changes in the external environment, and suitable for in-line inspection equipment.
2. Single-shot spectrally resolved interferometry for 3D thickness profile measurement of multilayer films in real time
A schematic diagram of our proposed technique is shown in Fig. 1. We used a tungsten halogen lamp with a wavelength region of 400 nm to 800 nm as the light source, and Köhler illumination was used to deliver uniformly distributed illuminance on the target and reference mirror, respectively. A linear polarizer (LP) was used to increase the visibility of the interference fringe patterns by adjusting the amount of light beams incident to the target and the reference. After passing through a quarter wave plate (QWP) twice in the measurement arm and reference arm, the state of polarization is converted, from s-polarization and p-polarization into p-polarization and s-polarization, respectively. As a result, the beam reflected by a polarizing beam splitter (PBS) passes through the PBS without a change in light intensity and vice versa. Then, another QWP was used to convert the p-polarization and s-polarization into right-hand circular polarization and left-hand polarization, respectively. The QWP used here is an achromatic quarter wave plate to provide λ/4 phase shift entirely independent of wavelength over 350 nm to 850 nm. Therefore, it was assumed that the phase-shift errors introduced by QWP are negligible. Imaging optics delivered the beams reflected from the target and reference into an imaging spectrometer. The imaging spectrometer produces a spectral image in the wavelength range from 380 nm to 800 nm, where one dimension of the image constitutes a spatial line and the other dimension represents the spectrum for each line pixel. In this way it provides the spectral information over the wavelength range provided by the light source (400 nm to 800 nm) for each pixel in the line. Two spectral line beams with orthogonal circular polarization, which are reflected from the target and reference mirror, are interfered by micropolarizer arrays in a pixelated polarizer camera (PPC), where each micropolarizer of a 2 × 2 unit cell array is oriented to 0°, 45°, 90°, and 135°. According to the direction of each micropolarizer angle, the phase-shifted spectral interference pattern at each measuring point in the line can be expressed as [20,21]
where, Ia and Ib are the reflected beam intensities from the sample and reference mirror, respectively. ϕ is the angle of the micro-polarizer and Φ is the phase difference between the measurement and reference beams.
Fig. 1. Schematic diagram of dynamic one-shot interferometry used for 3D thickness profile measurement of multilayer films in real time; LP: Linear Polarizer, QWP: Quarter Wave Plater (45°), PBS: Polarizing Beam Splitter, RM: Reference Mirror, IS: Imaging Spectrometer, PPC: Pixelated Polarizer Camera
In the unit cell of the PPC, ϕ is set 0°, 45°, 90°, and 135° so four different phase-shifted interferograms are generated simultaneously in one measurement
Fig. 2. Fundamentals of single-shot spectrally-resolved interferometry: (a) spectrally-resolved interference fringe image of the multilayer films obtained by a PPC, (b) four spectrally-resolved phase-shifted sub-array images of (a) corresponding to 90°, 45°, 135°, 0° micro-polarizer pixels, (c) spectrally-resolved interference fringe image of the standard reference obtained by a PPC, (d) four spectrally-resolved phase-shifted sub-array images of (c) corresponding to 90°, 45°, 135°, 0° micro-polarizer pixels, (e) spectral phase map of the multilayer films calculated by Eq. (3), and (g) spectral reflectance map of the multilayer films calculated by Eq. (5).
Figure 2(a) and (c) are one-shot spectral images of the sample and the standard reference obtained by PPC, respectively, and four subarray images of each one-shot spectral image corresponding to the micropolarizer angle are shown in Fig. 2(b) and (d). Figure 2(e) and (g) are the spectral phase and reflectance images of the sample, which can be calculated from the four phase-shifted spectral images in Fig. 2(b) and (d) using Eq. (3) and (5).
To measure the film thickness of each layer in multilayer films, a merit function can be defined by consisting of the nonlinear spectral phase and reflectance information, as follows.
3. Experimental results
We tested our proposed technique by measuring the 3D thickness profile of a five-layer film structure sequentially stacked in the order of Si3N4-SiO2-SiON-SiO2-Si3N4 on a patterned silicon wafer, as shown in Fig. 3(a). This sample specimen was fabricated by repeating several times a sequence of steps of photolithography processes such as etching and deposition. After forming each layer, the film thickness and surface profile of each layer were measured using a commercial stylus profiler (Veeco Dektak 8) and an ellipsometer (KLA-Tencor SFX-100), respectively. These measurement results were used to verify the performance of our proposed method.
Fig. 3. Measurement procedure using our proposed method: (a) internal structure of the patterned five layer films, (b) measured spectral phase distribution ${\Phi ^E}$ and its corresponding linear and nonlinear phase components $\Phi _{^l}^E$ and $\Phi _{^{nl}}^E$, (c) measured spectral reflectance data ${\Re ^E}$, (d) comparison of the measured spectral nonlinear phase data $\Phi _{^{nl}}^E$ (black solid line) with its corresponding analytical model $\psi _{^{nl}}^T$ (red crosses) for the film thickness measurement, (e) comparison of the measured spectral reflectance data ${\Re ^E}$ (black solid line) with its corresponding analytical model ${\Re ^T}$ (red crosses) for the film thickness measurement, and (f) comparison of the measured total spectral phase data ${\Phi ^E}$ (black solid line) with the analytical model of the spectral phase ${\Phi ^T}$ (red crosses) for the top surface profile measurement.
In this investigation, we used a tungsten halogen lamp as a broad-band light source and established Köhler illumination for high contrast and the even distribution of light on the sample and reference mirror over the field of view. We used an imaging spectrometer (V8E, SPECIM) to obtain a line spectral image in the VIS wavelength ranges (380 nm to 800 nm) with spectral resolution of 1.82 nm and the corresponding spectral phase was simultaneously measured by acquiring four spectrally-resolved phase-shifted interferograms using a pixelated polarizing camera (BFS-U3-51S5, FLIR). We used a telecentric lens with 2× magnification as the imaging optics for a line scan up to 4.22 mm in length with 3.45 μm resolution at a time. Before measurement, spectral axis calibration was performed by determining the exact wavelength value corresponding to each pixel of the spectral axis on the 2D detector array. As a calibration light source, we used a HgAr lamp with good peaks of known wavelengths in the visible spectral range. The wavelength information of each pixel on the spectral axis was calibrated by fitting a 3rd order polynomial curve along data points in the reference wavelengths of the HgAr lamp and their corresponding pixels. The standard reference was measured in advance for use as a relative reflectance value, in order to obtain the absolute reflectance of the sample. Here, we used a bare silicon wafer as the standard reference.
After these preliminary steps, we measured the spectral phase and reflectance of the sample specimen, as shown in Fig. 3(b) and (c). This measured total spectral phase can be divided into linear and non-linear components, ${\Phi _l}$ and ${\Phi _{nl}}$, and we used the non-linear phase term and reflectance data for the film thickness measurement. Figure 3(d) shows that our analytical models (red ‘x’ mark) are well fitted to the measured data (black solid line). After identifying each layer's film thickness in the previous step, we determined the top surface height h from the total phase information, as shown in Fig. 3(e).
Our proposed technique is basically a line-by-line scanning method so we needed to move the sample perpendicular to the measurement line for complete 3D measurement. Figure 4 shows the result of the 3D measurement of the patterned five-layer films, obtained by stitching the series of sequential line results together. Figure 4(a) - (e) shows the volumetric thickness distributions of each layer’s film, and the top surface height profile is shown in Fig. 4(f). Based on each layer’s film thickness and the top surface information of the sample, the 3D thickness profile was reconstructed, as Fig. 4(g).
Fig. 4. 3D thickness profile measurement results of patterned five layer films: (a) 3D thickness measurement result of a Si3N4 film layer (5th layer), (b) 3D thickness measurement result of a SiO2 film layer (4th layer), (c) 3D thickness measurement result of a SiON film layer (3rd layer), (d) 3D thickness measurement result of a SiO2 film layer (2nd layer), (e) 3D thickness measurement result of a Si3N4 film layer (1st layer), (f) 3D top surface profile of a Si3N4 film layer (5th layer), (g) reconstructed 3D thickness profile by combining the measurement results of (a) to (f), and (h) the cross-sectional surface profile of each film layer along the A-A′ line
To verify the proposed method, the film thickness of each layer at points S1 and S2, and each layer’s film profile along the A-A′ line was compared with the measurement results from the ellipsometer and stylus profiler, respectively. Table 1 summarizes the measurement results comparisons, between our method and a commercial ellipsometer (KLA-Tencor SpectraFx 100) at points S1 and S2 on the sample. We performed the measurements 10 times, and the maximum standard deviation of each film layer was less than 1.6 nm. It shows there was good agreement between the two measurement results, with a maximum difference of less than 14 nm. For a more detailed comparison, the film profile of each layer at A-A′ line was also measured in advance using a Veeco Dektak 8 stylus profiler during the fabrication process.
Table 1. Summary of the film thickness results from a sample compared with those from a commercial ellipsometer (KLA-Tencor SpectraFx 100)
We plotted two measurement results of each layer's film profile at the same time, as shown in Fig. 5(a)-(f), and the average height of each layer's step profile was measured respectively, and the measurement results were summarized in Table 2. The maximum deviation between two measurements was less than 12 nm and, it confirmed that they agreed well with each other without significant deviation.
Fig. 5. Our measurement results (black-solid line) compared with those from a commercial stylus profiler (red solid line) at A-A′ line: (a) Si substrate surface profile, (b) 1st Si3N4 film layer surface profile, (c) 2nd SiO2 film layer surface profile, (d) 3rd SiON film layer surface profile, (e) 4th SiO2 film layer surface profile, and (f) 5th Si3N4 film layer surface profile
Table 2. Summary of the surface profile results from a sample compared with those from a commercial stylus profiler (Veeco Dektak 8)
4. Conclusion
In this paper, a new concept of spectrally-resolved interferometry was proposed for measuring the film thickness and surface profile of each layer of multilayer films at the same time. We acquired four phase-shifted interferograms in the spectrally-resolved fringe pattern with just one measurement, and simultaneously obtained the variations in the spectral phase and spectral reflectance of the specimen by analyzing these interference signals. Both the reflectance and phase data over a wide wavelength range were utilized, to improve measurement reliability. Our proposed technique is line measurement, so lateral scanning is required for areal measurement. However, since each line result can be easily obtained in one measurement, 3D measurement can be performed by simply stitching the series of sequential line results together. As a preliminary step, we calibrated the imaging spectrometer using a HgAr lamp, and a bare silicon wafer was measured in advance as a standard reference to be used later for calculating the spectral reflectance of the sample. The feasibility of our proposed method was verified by measuring a patterned five-layer film sample and comparing our measurement results with those from other well-established techniques. The test specimen was fabricated via a series of sequential photolithography processes, and the thickness and surface profile of each film layer were measured at each fabrication step using commercial instruments, an ellipsometer and a stylus profiler, respectively. The comparison showed that our measurement results were in good agreement with other results and confirmed that our proposed technique successfully can measure the 3D thickness profile of multilayer films without any damage to the sample. We expect this single-shot interferometry will be widely used as an in-line inspection metrology tool in the semiconductor and display fields.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. R.M.A. Azzam and N. M. Bashra, Ellipsometry and Polarized Light (North Holland, Amsterdam, 1987).
2. H. Tompkins and E. A. Irene, Handbook of ellipsometry (William Andrew, 2005).
3. H. Fujiware, Spectroscopic Ellipsometry: Principles and Applications (John Wiley & Sons, Chichester, 2007).
4. L. Deck and P. de Groot, “High-speed noncontact profiler based on scanning white-light interferometry,” Appl. Opt. 33(31), 7334–7338 (1994). [CrossRef]
5. A. Harasaki, J. Schmit, and J. C. Wyant, “Improved vertical-scanning interferometry,” Appl. Opt. 39(13), 2107–2115 (2000). [CrossRef]
6. Y.-S. Ghim and S.-W. Kim, “Thin-film thickness profile and its refractive index measurements by dispersive white-light interferometry,” Opt. Express 14(24), 11885–11891 (2006). [CrossRef]
7. F. Gao, H. Muhamedsalih, and X. Jiang, “Surface and thickness measurement of a transparent film using wavelength scanning interferometry,” Opt. Express 20(19), 21450–21456 (2012). [CrossRef]
8. Z. Xie, Y. Tang, Y. Zhou, and Q. Deng, “Surface and thickness measurement of transparent thin-film layers utilizing modulation-based structured illumination microscopy,” Opt. Express 26(3), 2944–2953 (2018). [CrossRef]
9. S.-W. Kim and G. -H. Kim, “Thickness-profile measurement of transparent thin-film layers by white-light scanning interferometry,” Appl. Opt. 38(28), 5968–5973 (1999). [CrossRef]
10. D. Kim, S. Kim, H. J. Kong, and Y. Lee, “Measurement of the thickness profile of a transparent thin-film deposited upon a pattern structure with an acousto-optic tunable filter,” Opt. Lett. 27(21), 1893–1895 (2002). [CrossRef]
11. S. K. Debnath, M. P. Kothiyal, J. Schmit, and P. Hariharan, “Spectrally resolved white-light phase-shifting interference microscopy for thickness-profile measurements of transparent thin film layers on patterned substrate,” Opt. Express 14(11), 4662–4667 (2006). [CrossRef]
12. Y.-S. Ghim and S.-W. Kim, “Fast, precise, tomographic measurements of thin fims,” Appl. Phys. Lett. 91(9), 091903 (2007). [CrossRef]
13. P. de Groot and X. C. de Lega, “Angle-resolved three-dimensional analysis of surface films by coherence scanning interferometry,” Opt. Lett. 32(12), 1638–1640 (2007). [CrossRef]
14. D. Mansfield, “Extraction of film interfaces from scanning white-light interferometry,” Proc. SPIE. 7101, 71010U (2008). [CrossRef]
15. Y.-S. Ghim, H.-G. Rhee, H.-S. Yang, and Y.-W. Lee, “Thin-film thickness profile measurement using a Mira-type low-cohrence interferometer,” Meas. Sci. Technol. 24(7), 075002 (2013). [CrossRef]
16. P. d Groot and X. C. de Lega, “Signal modeling for low-coherence height-scanning interference microscopy,” Appl. Opt. 43(25), 4821–4830 (2004). [CrossRef]
17. M. F. Fay and T. Dresel, “Applications of model-based transparent surface films analysis using coherence-scanning interferometry,” Opt. Eng. 56(11), 111709 (2017). [CrossRef]
18. K. Wu, C.-C. Lee, N. J. Brock, and B. Kimbrough, “Multilayer thin-film inspection through measurements of reflection coefficients,” Opt. Lett. 36(16), 3269–3271 (2011). [CrossRef]
19. Y.-S. Ghim, H.-G. Rhee, and A. Davies, “Simultaneous measurements of top surface and its underlying film surfaces in multilayer film structure,” Sci Rept 7(1), 11843 (2017). [CrossRef]
20. N. J. Brock, B. T. Kimbrough, and J. E. Millerd, “A pixelated polarizer-based camera for instantaneous interferometric measurements,” Proc. SPIE 8160, 81600W (2011). [CrossRef]
21. J. E. Millerd, N. J. Brock, J. B. Hayes, M. B. North-Morris, M. Novak, and J. C. Wyant, “Pixelated phase-mask dynamic interferometer,” Proc. SPIE 5531, 304–314 (2004). [CrossRef]
