In the wake of continuous miniaturization, optical metrology has solidified its role as an essential instrument in semiconductor manufacturing, chiefly due to its non-invasive, high-precision, and rapid measurement capacities. Particularly crucial in advanced techniques such as automated optical inspection (AOI), its significance grows with the rising demand for accurate 3D integrated circuit packaging characterization. This introduces notable challenges for optical metrology, which need addressing to meet rigorous in-line process demands. This talk aims to present an overview of key measurement methods in semiconductor fabrication, encompassing CD-SEM, X-ray or EUV scatterometry, reflectometry for optical critical dimension (OCD) metrology, AFM, advanced optics, and white light interferometry. The talk will share insights into the intrinsic limitations of these techniques, potential future innovations, and recent advancements in optical metrology. The session will conclude with a perspective on the evolution of these techniques amidst the prevailing technological constraints.
This study develops a parametric system transfer function (STF) model using scalar diffraction theory and Fourier optics to address the loss of precision in image-based positioning caused by the diffraction limit on marker scale. By fitting the model to observed STFs and employing deconvolution and a deep convolutional neural network, the method enhances image quality, overcoming traditional deconvolution limitations. Applied to critical dimension measurements, it improved radius accuracy for vias and pillars by 54.8% and reduced displacement measurement bias by 36.4%. The development particularly benefits automatic optical inspection (AOI) for quality control in semiconductor manufacturing.
This paper introduces a new AI-enpowered method for accurately measuring submicron structures with high aspect ratios (HAR) in semiconductor packaging using spectral scatterometry across DUV, visible, and SWIR wavelengths. By optimizing polarization and spectrometer calibration, the method improves spectral signal contrast for precise critical dimension (CD) metrology. An Artificial Neural Network (ANN) tackles phase shift problems for trench spacings near light wavelengths, enabling precise CD measurement. Experiments demonstrate DUV light's proficiency in measuring small CD differences and VIS and SWIR's effectiveness for larger, deeper structures. The DUV system measures HARs up to 10:1 and apertures down to 0.46 μm with accuracy within 3% of Focused Ion Beam/Scanning Electron Microscope (FIB/SEM) comparisons.
This research means to solve the challenges in measuring deep high-aspect-ratio (HAR) and thin-film structures in 3D integrated circuits. As the semiconductor industry reaches its physical limitations in device scaling, advanced technologies such as advanced lithography and packaging have become crucial in extending Moore's law. This has led to the use of denser nano-to-sub-micron structures in three-dimensional integrated circuits (3D-IC), resulting in smaller, more functional devices. However, measuring these complex and deep HAR and thin-film structures with a large depth range from a few nanometers to a few hundred micrometers using a single optical system is challenging. To address this need, this article presents an AI-guided scatterometry method using numerical aperture control to achieve a large measurement range. The system uses broadband light to generate multi-wavelength reflection responses from the samples. With the help of an electromagnetic simulation tool and an artificial neural network model, the depth resolution can be improved through inverse modeling. The results demonstrate the ability to measure a wide range of samples with depths ranging from nanometers to micrometers scale, including sub-micron HAR openings and ultra-thin films, as long as the measurement bias is controlled within acceptable limits.
A new non-integral optical scatterometry technique has been introduced to circumvent issues with traditional methods in the critical dimension (CD) characterization of micro and nano-structures in semiconductor inspections. This method uses the high spatial coherence of the laser source, and an adjustable numerical aperture (NA) for effective beam shaping, enabling precise measurement of high-aspect-ratio structures. It incorporates a model-based approach with a virtual optical system and the Finite- Difference Time-Domain (FDTD) method for multiple CD characterizations, improving measurement precision. Early tests indicate a minimal average bias of 1.74% from calibrated references and standard deviations within 7 nm.
KEYWORDS: Point spread functions, Deconvolution, Databases, Imaging systems, Image deconvolution, Super resolution, Image processing, Diffraction limit, Convolution, Education and training
An AI-assisted computational method is developed to achieve superresolution imaging by overcoming the limitations of optical imaging caused by the diffraction limit. The method builds and fits a parametric optical imaging model by resolving an inverse problem. Highresolution imaging often requires a high-magnification lens, but this reduces the field of view. The transfer function of the imaging system is parameterized using diffraction theory and simulations of real imaging disturbances. "Known sample" and corresponding "Measured image" pairs are used to train the model and fit the real transfer function of the system. A deconvolution algorithm is applied to resolve the reverse problem and maintain high resolution under an enlarged field of view. The spatial resolution can be improved by 2.33 times compared to the diffraction limit. This method is useful for semiconductor critical dimension metrology in automated optical inspection.
Global semiconductor packaging manufacturers are developing advanced process technologies with the rapid rise of heterogeneous packaging and 3D packaging. High-aspect-ratio (HAR) structures like through silicon vias (TSV) or redistribution layers (RDL) that come with the prevalence of 3D packaging technology have further significantly increased the difficulties in optical critical dimension (OCD) metrology. Due to emerging technical challenges, effective sub-micron HAR OCD solutions are highly demanded to resolve the technical bottleneck. Thus, This article presents an AI-guided method for simulating and building a training dataset using the finite difference time domain (FDTD), then forming a DNN model for reconstructing CDs. At the same time, an optical scatterometry-based microscope was developed to adopt optical light capable of penetrating a sub-micron opening size structure and characterizing critical dimensions such as top critical dimension (TCD) and depth. In the optical system design, the optical FOV can be narrowed down only to cover a single sub-micron structure for OCD metrology. A preliminary test verified that a single sub-micron structure with an aspect ratio of 1:3.3 and the maximum bias between the measured data and the SEM references could be kept within a few tens of nanometers for its depth, TCD, and BCD measurement.
A neural network-assisted spectral scatterometry method is presented to measure multi-dimensional critical dimensions (CDs) on high aspect ratio (HAR) structures with micron or submicron scales. With the rise of 3D integrated circuit packaging, there is a need for accurate characterization of HAR sub-micron structures. This method uses DUV scatterometry and a broadband light source from DUV to visible light to gather multi-channel reflection data. The inverse modeling method and artificial neural network model enable accurate measurement of multiple CDs of test structures. The results showed accurate measurement of deep trench critical dimensions with a nominal line width of 0.6 μm and aspect ratio up to 5:1, with accuracy within a few nanometers comparable to SEM results using the same sample.
A new optical surface measuring method based on correlation-based diffractive image profilometry (DIP) is developed for accuracy enhancement by introducing external optical aberration to the microscope. According to the diffraction theory, the diffractive images formed in the microscope mainly depend on the microscopic optical system and the surface features of the tested object. The most critical issue affecting the measurement accuracy of the DIP is that the uniqueness of the diffractive images corresponding to various surface geometric parameters such as different heights and orientations cannot be always guaranteed. This situation can bring undesired uncertainties in surface measurement since undesired ambiguity in image correlation or model estimation may be introduced. To resolve this, a designed foreign aberration is introduced into the microscopic optical system to develop the feature variance of diffractive images for significantly increasing the degree of the image variance, therefore the risk of ambiguity is effectively avoided. Proved by some experimental tests, with this method, the accuracy in measuring height, tilting angle, and tilting direction can be achieved to a level of sub-micrometer and less than 0.01 degrees, respectively.
This work presents a novel microscopic profilometry using diffraction image correlation, which is appropriate for in-situ automated optical inspection (AOI). In contrast to the traditional confocal microscopy, the developed technique replaces the detector pinhole by observing and matching the diffraction patterns. Thus accurate surface depth detection can be achieved by eliminating time-consuming vertical scanning operation. The development is the first attempt in optics to use image correlation between the pre-calibrated database of diffraction images and the measured one of the detecting object surface. The feasibility of the method has been theoretically verified by scalar diffraction theory, then verified by experimental testing. Also, a depth response curve with the physical meaning of similarity is introduced and interpreted. Meanwhile, multi-point lateral scanning in one field of view (FOV) is achieved by quickly switching micromirrors on the digital micromirror device (DMD), thus the quasi full-field 3-D reconstruction can be acquired by combining tens of captured images. Verified experimentally, a 3-D reconstruction with sub-micrometer vertical resolution can be realized, with a tunable lateral resolution.
A new full-field profilometry based on diffraction image correlation (DIC) was developed in which the technique bases on calibrated database of reference diffractive images (RDIs) to estimate surface depth information. This technique has the advantage of removing vertical scanning for achieving high measurement efficiency in microscopic surface profilometry. However, as the diffractive images not only depend on the tested height but also the local surface tilt. Thus, the pre-built image database may not match the measured diffractive image and lead to a systematic measured error incurred by the surface tilting condition. Thus, in the article, the influence of surface tilt to the diffraction images is investigated and analyzed to understand the relationship between the surface tilt and the image variation. A lateral drifting phenomenon depends on the tilt angle and tilt direction was also quantified to estimate the impact of the tested surface height. Meanwhile, a detection algorithm to determine the center of diffractive image was developed to position the shifting quantity of the image. Moreover, the information coupling problem between height and tilting parameters, such as the tilt angle and tilt direction was studied to decouple these parameters from the tested height, so the tested surface can be reconstructed accurately. To realize the proposed method, some data processing strategies were also proposed to decouple the depth information from multi-surface parameters such as surface tilt angles (pitch and yaw) and direction effectively.
By combining classic differential interference contrast (DIC) with the chromatic confocal principle, we show that phaseshifting calibration can be avoided in DIC by using spectral information induced by the investigated sample. The created spectral fringe can be further used to unwrap the phase. This unwrapping is limited by the spectral resolution of the spectrometer. Therefore, the depth-difference around a single measurement point can be determined instantaneously. To reconstruct the depth profile, the integration of a depth-gradient is necessary. By combining the depth information of the chromatic confocal carrier signal with the differential depth information of the carried DIC signal, the accumulation of measurement uncertainty can be reduced. To our best knowledge, the proposed chromatic confocal differential interference contrast (CCDIC) is a novel profile reconstruction principle. To verify the feasibility of the CCDIC, a prototype probe with an adjustable shear and phase has been developed. Preliminary experiments achieve sub-micrometer depth resolution. A current challenge requiring further work is the stable unwrapping of the phase-difference by spectral frequencies.
Publisher’s Note: This paper, originally published on 17 July 2015, was withdrawn per author request, if you have any questions please contact SPIE Digital Library Customer Service for assistance.
In this research, new nano-scale measurement methodology based on spectrally-resolved chromatic confocal interferometry (SRCCI) was successfully developed by employing integration of chromatic confocal sectioning and spectrally-resolve white light interferometry (SRWLI) for microscopic three dimensional surface profilometry. The proposed chromatic confocal method (CCM) using a broad band while light in combination with a specially designed chromatic dispersion objective is capable of simultaneously acquiring multiple images at a large range of object depths to perform surface 3-D reconstruction by single image shot without vertical scanning and correspondingly achieving a high measurement depth range up to hundreds of micrometers. A Linnik-type interferometric configuration based on spectrally resolved white light interferometry is developed and integrated with the CCM to simultaneously achieve nanoscale axis resolution for the detection point. The white-light interferograms acquired at the exit plane of the spectrometer possess a continuous variation of wavelength along the chromaticity axis, in which the light intensity reaches to its peak when the optical path difference equals to zero between two optical arms. To examine the measurement accuracy of the developed system, a pre-calibrated accurate step height target with a total step height of 10.10 μm was measured. The experimental result shows that the maximum measurement error was verified to be less than 0.3% of the overall measuring height.
KEYWORDS: Confocal microscopy, Point spread functions, Colorimetry, Deconvolution, Calibration, Signal processing, Near field optics, Sensors, Convolution, Spectroscopy
In this research, novel deconvolution methodology is proposed to resolve the lateral and axial cross-talk problems encountered in line-scanning chromatic confocal surface profilometry. The strategy integrates chromatic confocal principle, infinitive microscopic optics and deconvolution theory to resolve the entangled cross-talk problem in microscopic confocal measurement, so the measuring resolution can be greatly enhanced from the level of the traditional line-scanning up to the one achieved by generally traditional point-type confocal measurement. To overcome the problem, this research analyzes the physical phenomenon of optical near field using photonic spectrum analyses for establishing relationship between the light expansion and propagation depth, as well as light wavelength. In the confocal image, acquired spectrum intensity can be regarded as the convolution between the ideal signal from objects and the point spread function (PSF) of incident light. By employing spectrum analyses, important calibrated characteristics of the PSF along both of the lateral and depth directions can be carefully established. By using the individual PSF for its corresponding wavelength detected at its matching focal depth, the proposed deconvolution method has been proved effective theoretically and experimentally in greatly minimizing the full width half maximum (FWHM) of the depth response curve by more than 25 times, thus significantly improving the accuracy and repeatability of microscopic surface profilometry.
Full-field chromatic confocal surface profilometry employing a digital micromirror device (DMD) for spatial correspondence is proposed to minimize lateral cross-talks between individual detection sensors. Although full-field chromatic confocal profilometry is capable of enhancing measurement efficiency by completely removing time-consuming vertical scanning operation, its vertical measurement resolution and accuracy are still severely affected by the potential sensor lateral cross-talk problem. To overcome this critical bottleneck, a DMD-based chromatic confocal method is developed by employing a specially-designed objective for chromatic light dispersion, and a DMD for lateral pixel correspondence and scanning, thereby reducing the lateral cross-talk influence. Using the chromatic objective, the incident light is dispersed according to a pre-designed detection range of several hundred micrometers, and a full-field reflected light is captured by a three-chip color camera for multi color detection. Using this method, the full width half maximum of the depth response curve can be significantly sharpened, thus improving the vertical measurement resolution and repeatability of the depth detection. From our preliminary experimental evaluation, it is verified that the ±3σ repeatability of the height measurement can be kept within 2% of the overall measurement range.
The active interferometer has been approved as an effective architecture to increase the precision of surface profile
measurement by detecting the environmental vibration and then stabilized by a piezoelectric transducer. The
performance of such system can improve the accuracy within nano-meters. Most of the active interferometers have being
implemented on the PC based platform because it is easy to be setup and modulized, However, the latency of image
grabbing processing, mathematic algorithm and I/O trigger delay is around 10~15ms. Moreover, the latency is varying if
different PCs are used.
In this study, an embedded real-time controller based on the ADI Blackfin BF561 DSP microprocessor was developed to
improve the performance of a vibration-resistance system using white-light interferometer. The purpose of this
embedded control system is to detect the vibration and compensate it for White Light Interferometer when processing
stepping photograph grabbing for surface profile measurement.
A customized embedded control system with a 600 MHz high performance microprocessor, Camera Link interface,
DAC circuit and UART command port has been developed. A complete command set has also been defined to
communicate with PC based interferometer user interface for close-loop control. The latency has been reduced from tens
of millimeters to 250 μs as tested for all the procedures from CCD trigger, image processing, to PZT control signal
output. It has been proved to be useful of fixing the latency between phase detecting and compensation output. With this
advantage, the response of the interferometer can accelerated so that the vibration-resistance is improved and the overall
accuracy can be in the range of a few nano-meters.
In the research, full-field chromatic confocal surface profilometry employing digital micro-mirror device (DMD) for
spatial correspondence is proposed to minimize lateral cross talks between individual detection sensors. Although fullfield
chromatic confocal profilometry is capable of enhancing measurement efficiency by completely removing timeconsuming
vertical scanning operation, its vertical measurement resolution and accuracy are still severely affected by the
potential sensor cross talk problem. To overcome this critical bottleneck, a DMD-based chromatic confocal method is
developed by employing a specially-designed objective for chromatic light dispersion and a DMD for lateral pixel
correspondence and scanning. Using the chromatic objective, the incident light is dispersed according to a pre-designed
detection range from a few micrometers to several millimeters and a full-field reflected light is captured by a three-chip
color camera for multi color detection. Using this method, the full width half maximum (FWHM) of the depth response
curve can be significantly sharpened, thus improving the vertical measurement resolution and repeatability of the depth
detection. From our preliminary experimental evaluation, it is verified that the ±3σ repeatability of the height
measurement can be kept within 2% of the overall measurement range.
A dynamic 3-D nano-scale surface profilometer using stroboscopic white light interferometry with novel image
deconvolution and automatic identification of structure resonant modes was successfully developed. As micro
electromechanical systems (MEMS) increase rapidly towards industrial application, the needs of accurate dynamic
characterization are extremely important to optimal design and fabrication. To meet the demands, an optical microscopy
based on stroboscopic interferometry was developed to achieve full-field vibratory out-of-plane surface profilometry
and system characterization. A novel deconvolution strategy with correction of the light response function was
established to remove the potential image blurs caused by the unavoidable vibration of the tested parts. With this
technical advance, the bandwidth of dynamic measurement can be significantly increased up to 10 MHz without
sacrificing measurement accuracy. Meanwhile, an innovative detection algorithm based on image contrast measure was
developed for automatic identification of accurate resonant modes. The detection method provides the simplest and
most economic way to detect accurate resonant peaks without adding any significant hardware in a stroboscopic
interferometric framework. To verify the effectiveness of the developed methodology, AFM cantilever beams were
measured to analyze the full-field resonant vibratory modes and dynamic characteristics. The experimental results
confirm that the resonant vibration behavior of the tested microcantilever beams can be accurately characterized and 5
nm of vertical measurement accuracy as well as tens micrometers of vertical measurement range can be achieved. The
measured results were satisfactorily consistent with the theoretical simulation outcomes from ANSYS.
A static and dynamic 3-D surface profilometer with nano-scale measurement resolution was successfully developed using stroboscopic illumination and white-light vertical scanning techniques. Microscopic interferometry is a powerful technique for static and dynamic characterization of micro electromechanical systems (MEMS). As MEMS devices move rapidly towards commercialization, the issue of accurate dynamic characterization has emerged as a major challenge in design and fabrication. In view of this need, an interferometric microscopy based on white-light stroboscopic interferometry using vertical scanning principle was developed to achieve static and dynamic full-field profilometry and characterization of MEMS devices. A micro cantilever beam used in AFM was characterized using the developed instrument to analyze its full-field resonant vibratory behavior. The first five mode resonant vibration can be fully characterized and 3-5 nm of vertical measurement accuracy as well as tens micrometers of vertical measurement range can be achieved. The experimental results were consistent with the theoretical simulation outcomes from ANSYS. Using white-light stroboscopic illumination and white-light vertical scanning techniques, our approach has demonstrated that static and dynamic 3-D nano-scale surface profilometry of MEMS devices with tens-micrometer measurement range and a dynamic bandwidth up to 1MHz resonance frequency can be achieved.
KEYWORDS: 3D modeling, Reverse engineering, Reverse modeling, Sensors, CCD cameras, Visual process modeling, Process modeling, 3D metrology, Optical filters, Digital filtering
A vision-drive automatic digitization process for free-form surface reconstruction has been developed, with a coordinate measurement machine (CMM) equipped with a touch-triggered probe and a CCD camera, in reverse engineering physical models. The process integrates 3D stereo detection, data filtering, Delaunay triangulation, adaptive surface digitization into a single process of surface reconstruction. By using this innovative approach, surface reconstruction can be implemented automatically and accurately. Least-squares B- spline surface models with the controlled accuracy of digitization can be generated for further application in product design and manufacturing processes. One industrial application indicates that this approach is feasible, and the processing time required in reverse engineering process can be significantly reduced up to more than 85%.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.