Improved control for semiconductor processing systems

Abstract

The present disclosure provides optical data processing with improved latency, repeatability, stability, signal detectability, and other advantages. The improved processing can be used to monitor and control semiconductor processes more accurately and consistently. In one example, a method for processing spectral data comprises: (1) collecting a sequence of optical emission spectroscopic data arranged in chronological order across one or more wavelengths; (2) extracting one or more attributes from the sequence of optical emission spectroscopic data arranged in chronological order; (3) analyzing the characteristics of one or more attributes; (4) determining the conditions of one or more attributes; (5) processing one or more attributes according to a predetermined set of filters, conditions, and characteristics; and (6) selecting a filter configuration for processing spectral data according to the processing of one or more attributes.

Description

Technology Field

[0001] Cross-reference of related applications

[0002] This application claims the benefit of U.S. provisional application serial number 63 / 389,416, titled “Improved control for semiconductor processing systems,” filed by Philent Chris on July 15, 2022, which was ordinarily assigned together with this application, the entirety of which is incorporated herein by reference.

[0003] Technology field

[0004] The present disclosure generally relates to optical spectroscopic systems and methods of use, and more specifically to improved signal processing for lower latency, increased repeatability, and other control benefits between semiconductor tool controllers and real-time data collected from spectrometers used for optical signal acquisition. Background Technology

[0005] Optical monitoring of semiconductor processes is a well-established method for controlling processes such as etching, deposition, chemical mechanical polishing, and implantation. Optical emission spectroscopy (OES) and interferometric endpoints (IEP) are the two basic modes of operation for data acquisition. In OES applications, light emitted from the process, typically from plasma, is collected and analyzed to identify and track changes in atomic and molecular species that indicate the state or progress of the process being monitored. In IEP applications, light is typically supplied from an external source, such as a flash lamp, and delivered to the workpiece. Upon reflection from the workpiece, the sourced light conveys information in the form of the workpiece's reflectance, which indicates the state of the workpiece. The extraction and modeling of the workpiece's reflectance enables the understanding of film thickness and feature size, depth, and width, among other attributes.

[0006] In one embodiment, the present disclosure provides a method for processing spectral data. In one example, the method comprises: (1) collecting a time sequence of optical emission spectroscopy data over one or more wavelengths; (2) extracting one or more attributes from the time sequence of optical emission spectroscopy data; (3) analyzing the characteristics of one or more attributes; (4) determining the conditioning of one or more attributes; (5) processing one or more attributes according to a predetermined set of filters, conditioning, and characteristics; and (6) selecting a filter configuration for processing spectral data based on processing one or more attributes.

[0007] In another aspect, the present disclosure provides a method for controlling a semiconductor process. In one example, the control method comprises: (1) collecting optical emission spectroscopic data over one or more wavelengths; (2) processing the data using a predetermined method determined to provide a minimum process delay when determining an endpoint indication; and (3) modifying the semiconductor process based on processing the data.

[0008] In another aspect, the present disclosure provides a computing device. In one example, the computing device comprises one or more processors that perform operations including: (1) collecting optical emission spectroscopic data over one or more wavelengths; (2) processing the data using a predetermined method determined to provide a minimum process delay when determining an endpoint indication; and (3) changing a semiconductor process based on processing the data.

[0009] In another aspect, the present disclosure provides a computer program product having a series of operation instructions stored on a non-transient computer-readable medium that directs the operation of one or more processors when initiated to perform operations for processing spectral data. In one example, the operations include: (1) collecting a time-sequence sequence of optical emission spectroscopic data over one or more wavelengths from a semiconductor process; (2) extracting one or more attributes from the time-sequence sequence of optical emission spectroscopic data; (3) analyzing the characteristics of one or more attributes; (4) determining the control of one or more attributes; (5) processing one or more attributes according to a predetermined set of filters, control, and characteristics; and (6) selecting a filter configuration for processing spectral data based on processing one or more attributes using one or more filters from a predetermined set of filters. Brief explanation of the drawing

[0010] Now, refer to the following description taken together with the attached drawings. Figure 1 is a block diagram of a system using OES and / or IEP to monitor and / or control the state of a plasma or non-plasma process within a semiconductor process tool. Figure 2 is a schematic diagram generally illustrating the functional elements of a typical area CCD sensor. FIG. 3 is a plot of a typical OES optical signal (spectrum) resulting from the conversion of collected light according to the present disclosure. FIG. 4 is a plot of unprocessed signal trends extracted from digital signals collected from an optical sensor, such as the OES optical signal of FIG. 3, in accordance with the present disclosure. FIG. 5 is a flowchart of a method for collecting data from an optical sensor and processing the data for lower latency, increased repeatability, and other advantages according to the present disclosure. FIG. 6a is a plot of the temporal evolution of noise associated with the trend of FIG. 4, according to the present disclosure. FIG. 6b is a histogram plot of noise associated with the trend of FIG. 4 according to the present disclosure. FIG. 6c is a power spectral density plot of noise associated with the trend of FIG. 4, according to the present disclosure. FIG. 7 is a plot of estimated signals and features selected from the trends of FIG. 4 according to the present disclosure. FIGS. 8a-8g are plots of various filtering methods applied to the trends of FIG. 4 according to the present disclosure. FIG. 9 is a plot comparing the calculated endpoint latencies of the trends of FIG. 4 when various filters are applied, according to the present disclosure. FIGS. 10a and 10b are plots of the trends of FIG. 4 filtered in various ways with or without control, according to the present disclosure. FIGS. 11a and 11b are plots of representative IEP optical signal data processed in various ways according to the present disclosure. FIG. 12 is a block diagram of a spectrometer and certain related systems according to the present disclosure. FIG. 13 illustrates a block diagram of an example computing device configured to apply spectrum and trend processing to spectrum data according to the present disclosure. Specific details for implementing the invention

[0011] In the following description, reference is made to the accompanying drawings, which illustrate specific embodiments that form part of the invention and in which the invention can be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it should be understood that other embodiments may be utilized. It should also be understood that structural, procedural, and systemic changes may be made without departing from the spirit and scope of the invention. Accordingly, the following description should not be taken in a limiting sense. For clarity of description, similar features illustrated in the accompanying drawings are indicated by similar reference numerals, and similar features illustrated in alternative embodiments in the drawings are indicated by similar reference numerals. Other features of the invention will become apparent from the accompanying drawings and the following detailed description. For clarity of description, certain elements in the drawings may not be drawn to scale.

[0012] The continuous advancement of semiconductor processes toward faster processes, smaller feature sizes, and more complex structures has increased the demand for process monitoring technologies. For example, higher data rates are required to accurately monitor much faster etching rates in very thin layers, such as FINFETs and 3D NAND structures, where changes of the angstrom (a small number of atomic layers) are critical. In many cases, wider optical bandwidths and higher signal-to-noise ratios are required for both OES and IEP methodologies to aid in detecting small changes in reflectance and optical emission. Cost and packaging size are also under continuous pressure as process equipment becomes more complex and expensive in itself. All these requirements aim to improve the optical monitoring performance of semiconductor processes. Regardless of whether OES or IEP methodologies are used, critical components of many optical monitoring systems are spectrometers and the ability to consistently and accurately convert received optical data into electrical data for the control and monitoring of semiconductor processes.

[0013] Accordingly, processes, systems, and devices are disclosed that provide improved processing of optical data for the benefit of lower latency, increased repeatability, improved process stability, increased signal detectability, and other benefits, collectively referred to as attributes, through the characterization of the influence of noise, modulation, and filter selection on optical trend data and / or optical features. Such improved processing can be used to monitor and control semiconductor processes more accurately and consistently.

[0014] In relation to monitoring and evaluating the state of a semiconductor process within a process tool, FIG. 1 illustrates a block diagram of a process system (100) utilizing OES and / or IEP to monitor and / or control the state of a plasma or non-plasma process within a semiconductor process tool (110). The semiconductor process tool (110), or simply the process tool (110), generally surrounds a wafer (120) and possibly a process plasma (130) in a generally partially empty volume of a chamber (135) that may contain various process gases. The process tool (110) may include one or more optical interfaces, or simply interfaces (140, 141, and 142), to allow observation into the chamber (135) at various locations and orientations. The interfaces (140, 141, and 142) may include a number of types of optical elements, such as optical filters, lenses, windows, apertures, optical fibers, etc., but are not limited thereto.

[0015] For IEP applications, the light source (150) may be directly connected to the interface (140) or connected via a fiber optic cable assembly (153). As illustrated in this configuration, the interface (140) is oriented perpendicular to the surface of the wafer (120) and is often centered relative to the wafer (120). Light from the light source (150) may enter the internal volume of the chamber (135) in the form of a collimated beam (155). The beam (155) reflected from the wafer (120) may be received again by the interface (140). In typical applications, the interface (140) may be an optical collimator. Following reception by the interface (140), the light may be transmitted to a spectrometer (160) via a fiber optic cable assembly (157) for detection and conversion into a digital signal. The light may include sourced and detected light, and may include, for example, a wavelength range from deep ultraviolet (DUV) to near infrared (NIR). Wavelengths of interest may be selected from any sub-range of the wavelength range. For larger substrates or when understanding wafer non-uniformity is a concern, additional optical interfaces (not shown in FIG. 1) generally oriented with the wafer (120) may be used. The process tool (110) may also include additional optical interfaces located at different positions for other monitoring options.

[0016] For OES applications, the interface (142) may be oriented to collect light emissions from the plasma (130). The interface (142) may simply be a viewport or may additionally include other optical devices (optics), such as lenses, mirrors, and optical wavelength filters. A fiber optic cable assembly (159) may send the collected light to a spectrometer (160) for detection and conversion into digital signals. The spectrometer (160) may include a CCD sensor and a transducer, such as the CCD sensor (200) and transducer (250) of FIG. 2, for detection and conversion. Multiple interfaces may be used individually or in parallel to collect OES-related optical signals. For example, the interface (141) may be positioned to collect emissions from near the surface of the wafer (120), and the interface (142) may be positioned to view the bulk of the plasma (130) as shown in FIG. 1.

[0017] In many semiconductor processing applications, it is common to collect both OES and IEP optical signals, and this collection presents several problems regarding the use of a spectrometer (160). Generally, OES signals are temporally continuous, whereas IEP signals can be temporally continuous, discrete, or both. The mixing of these signals causes many difficulties because process control often has to detect small changes in both OES and IEP signals, and unique variations in one of the signals can obscure the observation of changes in the other signal. For example, supporting multiple spectrometers for each signal type is not advantageous due to cost, complexity, inconvenience of signal timing synchronization, calibration, and packaging.

[0018] After the optical signals received by the spectrometer (160) are detected and converted into analog electrical signals, the analog electrical signals are generally amplified and digitized within a subsystem of the spectrometer (160) and transmitted to a signal processor (170). The signal processor (170) may be, for example, an industrial PC, a PLC, or another system, and it uses one or more algorithms to generate an output (180), such as an analog or digital control value representing, for example, the ratio of two wavelength bands or the intensity of a specific wavelength. Instead of a separate device, the signal processor (170) may alternatively be integrated with the spectrometer (160). The signal processor (170) may use one or more OES algorithms to analyze emission intensity signals at predetermined wavelength(s), access corresponding states such as endpoint detection or etching depth, and determine trend parameters that are related to the state of the process and represent trends. For IEP applications, the signal processor (170) may use one or more algorithms that analyze broadband portions of the spectrum to determine film thickness. For example, refer to the system and method for field monitoring and control of film thickness and trench depth incorporated herein by reference U.S. Patent 7,049,156. FIGS. 11a and 11b are plots of representative IEP optical signal data processed in various ways according to the present disclosure. The output (180) may be transmitted to the process tool (110) via a communication link (185) to monitor and / or modify the production process occurring within the chamber (135) of the process tool (110).

[0019] The components of FIG. 1 are simplified for convenience and are commonly known. In addition to general functions, the spectrometer (160) or signal processor (170) may also be configured to identify stationary and transient optical and non-optical signals and to process these signals according to the methods and / or features disclosed herein. As such, the spectrometer (160) or signal processor (170) may include one or more algorithms, processing capabilities, and / or logic for identifying and processing optical signals and temporal trends extracted therefrom. The algorithms, processing capabilities, and / or logic may be in the form of hardware, software, firmware, or a combination thereof. The algorithms, processing capabilities, and / or logic may be located within a single computing device or may be distributed across multiple devices, such as the spectrometer (160) and the signal processor (170).

[0020] FIG. 2 is a schematic diagram generally illustrating the functional elements of a conventional area CCD sensor (200). The sensor (200) generally includes an active pixel area (210) that can be divided into an array of individual pixels, such as 1024(H)x122(V), as in the S7031 CCD sensor from Hamamatsu, Japan. As a matter of definition and clarity, it should be noted that when referring to the optical sensor herein, the use of “horizontal” and “vertical” refers to the physical major axis and physical minor axis of the optical sensor under discussion, respectively. In spectroscopic applications, the major axis / horizontal axis of the optical sensor is aligned with the orientation of wavelength dispersion, while the minor axis / vertical axis is associated with the imaging or collection of a defined light source or illuminated aperture, such as a fiber or optical slit.

[0021] The sensor (200) also includes a horizontal shift register (220) adjacent to the pixel area (210). For example, optical signals integrated into the sensor (200) from a fiber optic cable assembly (157 or 159) are typically read by shifting the charge stored in each pixel of the pixel area (210) vertically into the horizontal shift register (220) as indicated by the arrow (230). All or part of the active pixel area (210) may be shifted in such a row-by-row manner. Following the vertical shift, a horizontal shift may be performed as indicated by the arrow (240). As each pixel of the horizontal shift register (220) is shifted (toward the top of FIG. 2), the signal content may be converted from analog to digital signal by a converter (250), for example, from an analog electrical signal to a digital electrical signal. Subsequent handling and processing of the resulting digital data may occur inside or outside the spectrometer and may include averaging, curve fitting, threshold detection, filtering, and / or other mathematical operations, for example, to ensure consistency and reduce latency in detecting one or more attributes while processing spectral data.

[0022] The sensor (200) may also include one or more regions of non-illuminated or partially illuminated elements, such as shift register elements (260 and 261) and pixel region elements (270, 271, and 272). Generally, elements (260 and 261) may be referred to as "blank" pixels, and elements (270, 271, and 272) may be referred to as "bevel" pixels. One or more of these regions of elements may be included within the sensor (200) to provide characterization of the non-optical signal levels unique to the sensor (200). Non-optical signals may generally include signal offsets, signal transients, and other forms of signal fluctuations due to temperature or other non-optical factors.

[0023] FIG. 3 illustrates a plot (300) providing context of a typical OES optical signal (spectrum) (320) that can be collected through a spectrometer such as the spectrometer (160) of FIG. 1, which evolves over time with the monitored semiconductor process, from which a series of trends can be extracted and processed as described herein. The plot (300) has an x-axis in waveform units and a y-axis in signal count units. The spectrum (320) can be derived from incident light on a sensor such as the sensor (200) of FIG. 2. The spectrum (320) shows typical features of molecular (broadband structure around 400 nm) and atomic emission (narrow peaks overall). An example of a narrow peak, narrow feature (330) corresponds to the 656 nm emission line of hydrogen and can be extracted for use in monitoring and endpointing the semiconductor etching process.

[0024] FIG. 4 illustrates a plot (400) of an unprocessed signal trend (410) that can be extracted from a time series of a spectrum such as the OES optical spectrum (320) of FIG. 3. The plot (400) has an x-axis of time (seconds) and a y-axis of signal count. Specifically, the trend (410) can be generated by selecting a range of spectral values ​​that occur in proximity to a spectral feature of interest. For example, to monitor 656 nm hydrogen emission, such as the narrow feature (330) shown in FIG. 3, values ​​corresponding to the spectral region from 655 to 657 can be averaged or summed and stored in a time configuration array to generate the trend (410). Due to optical calibration and resolution limitations, spectral features have a finite width in the collected spectrum, and spectral regions wider than the actual emission line width may be used for processing. The trend (410) is collected over a period of 5 seconds, which generally corresponds to a fast semiconductor process. Individual points of the trend (410) and the original corresponding spectrum can be collected at an adjustable rate suitable for analysis. In this example, the trend (410) is collected at 50 samples per second, but it can be collected at a rate ranging from a few samples per second to hundreds of samples per second. The sampling rate and the number of points in the resulting trend can be adjusted to best suit the processing and control requirements described herein, and the described processes can be performed at one or more sampling rates to determine the preferred result. It should be noted that the trend (410) is non-real-time as it is displayed after collection and may include additional data both before and after a specific endpoint stage or monitored process. Real-time data updates will only include a portion of the trend (410) up to the current processing and / or collection time.Trends applicable to the processing described herein may include, for example, single wavelength trends, multiple wavelength trends, and / or combinations of wavelength trends such as ratios, products, sums, and differences.

[0025] FIG. 5 illustrates a flowchart of an example of a method (500) for reading data from an optical sensor and processing the data for lower latency, increased repeatability, and other benefits. It should be noted that the method (500) may include steps performed in real-time or non-real-time during a controlled process, before a controlled process, and / or after a controlled process. Real-time may be defined as occurring during active control or monitoring of a process. Real-time may be associated with causal processing, as the data includes only current and past times. Non-causal processing after the data is collected includes data at times occurring before, during, at a point in time, and after a monitoring event.

[0026] The method (500) begins with a preparation step (510) in which any preparation action may be taken. These actions may include mechanically connecting the components of the optical measurement system, selecting the sampling rate of the spectrometers, and determining spectral lines or features of interest. Step (510) is an example of a step of the method (500) that may be performed prior to the control process. Following any preparation action, the method (500) proceeds to a step (520) in which spectral data may be collected. Spectral data may be collected using spectrometers and accessories as described according to FIGS. 1 and 2.

[0027] Trend data from one or more trends in step (530) may be extracted from the spectrum data collected during step (520). For real-time analysis and control, the extraction of individual trend values ​​occurs almost simultaneously with the collection of each spectrum included within the spectrum data collected during step (520). For non-real-time analysis and control, trend extraction may occur after the collection of any or all portions of the spectrum data in step (520). Trends such as the trend (410) of FIG. 4 may be extracted from various samples of the collected spectrum data. In the next step (540), one or more characteristics of the trend data are analyzed. Characteristics determined and analyzed from the trend data may include, for example, noise characteristics, signal estimation, endpoint characteristics, endpoint detectability, and / or signal-to-noise evaluation as discussed below in relation to FIG. 6 through 9. In step (550), the trend data is adjusted. Before, after, and / or simultaneously with the analysis of the trend data in step (540), the trend data may be adjusted in step (550). Conditioning can include, for example, scaling, normalization, standardization, ratio adjustment, offset adjustment, or other mathematical operations that aid in trend data processing. In general, conditioning data improves its usefulness and applicability to control applications in which the data is used. For example, offset adjustment can be applied to trend data to remove undesirable DC signal offsets from trend data containing underlying information content encoded within variations within the trend, rather than the general signal values. Additionally, ratio adjustment can be used to remove systematic common-mode noise and / or signal variations that could complicate subsequent trend data processing.

[0028] In step (560), trend data is processed based on the characteristic analysis in step (540) and the adjustment of trend data in step (550). Trend data may be processed in real time or post-processed after collection to apply and evaluate combinations of adjustments and filters as described below in relation to FIGS. 6 through 10. Trend data may be selected and obtained from one or more trends extracted in step (530). Processing of trend data may involve understanding the signal and noise associated with the signal, and then going through various approaches to determine how to process the trend data or to optimize the processing method. Determining the processing method may involve testing and evaluating different filters and / or combinations of filters, e.g. those mentioned herein, for trend data having different values. The desired processing result is the amount of time (latency) between the "true" time of feature occurrence and the actual identification time, and consistency in identifying features. For example, this feature ideally occurred at 5 seconds but was not identified / detected until 5.5 seconds, resulting in a latency of 0.5 seconds. Processing does not necessarily involve identifying a specific trend, but may involve identifying one or more features without identifying a specific trend. Accordingly, the processing of step (560) may occur with a defined metric including, for example, the identification of a specific trend, the identification of a feature (specific process metric), and / or a combination thereof.

[0029] In step (570), one or more semiconductor processes are modified based on the analysis, control, processing, or combination thereof of steps (540 to 560) of the method (500). Under conditions where the method (500) is applied in real time, the semiconductor process may be modified in real time, and the semiconductor process may be the process in which spectrum data is collected in step (520). Another semiconductor process may be modified non-real-time of the current semiconductor process in step (520). As an example of non-real-time processing of trend data, a description of the methodology for processing and analyzing trend data from the method (500) or a part thereof may be stored and programmed in a control system for later use during another subsequent real-time semiconductor process. The description of the processing and analysis of trend data may include a number of mathematical operations, equations, formulas, and processes applied to the data to perform control and processing as described herein, for example. Descriptions for the processing and analysis of trend data may be stored and / or programmed in / on memory (1234) and processor (1236) of, for example, a spectrometer (160) or signal processor (170) of a process system (100), memory / storage (1190) of an optical system (1100), an FPGA (1160), a processor (1170), and / or an external system (1120), and / or a computing device (1200). The memory / storage (1190) and memory (1234) may be non-transient computer-readable media.

[0030] The method (500) proceeds to step (580) and terminates. During real-time processing, step (580) may include terminating the semiconductor process and storing associated data for future analysis. It should be noted that the method (500) may be performed any number of times and may be designed to be updated based on additional characterization, analysis, and processing in real-time or non-real-time.

[0031] Working with non-real-time data such as trends (410) enables signal estimation, noise extraction, and characterization by applying non-causal signal processing, such as Savitzky-Golay filtering, to the collected trends. Other filtering processes, such as Weiner filters and other general "matched filters," as well as Savitzky-Golay filtering, can be used causally (typically in real-time) or non-causally (typically in non-real-time). FIG. 6a illustrates a plot (600) of noise associated with the trends of FIG. 4 and trends (610) extracted through processing using a low-order polynomial Savitzky-Golay filter. FIG. 6a has an x-axis in units of time and a y-axis in units of noise counts. Similarly, FIG. 6b illustrates a histogram plot (650) of noise associated with the trends of FIG. 4. FIG. 6b has an x-axis in units of noise count and a y-axis in units of "occurrence frequency." Additionally, FIG. 6c illustrates a power spectral density plot (670) of the noise associated with the trend of FIG. 4. FIG. 6c has an x-axis in units of frequency and a y-axis in units of power spectral density (dB / Hz). Each method of noise processing and analysis provides insight into the time and frequency variations of the noise amplitude and supports further processing of the trend. For example, the power spectral density plot (670) shows a clear frequency distribution of the noise and variations of less than ~3Hz that are not immediately apparent in the time or histogram plot.

[0032] FIG. 7 illustrates a plot (700) of features (730, 740) selected from the estimated signal (720) and the trend (410) of FIG. 4. FIG. 7 has an x-axis in time units and a y-axis in count units. The vertical lines represent peak and lowest inflection point locations over time based on non-causal Savitzky-Golay first derivative estimation. These inflection points and other features may be useful for characterizing various processing methods and resulting feature detection latencies that may be further associated with the determination of endpoints and other process control events. For example, a control system processing a control trend according to the method described herein may first identify an inflection point (730), then identify an inflection point (740), and signal an endpoint time of a few seconds or a few samples after the identification of the inflection point (740). Therefore, the inflection points (730, 740) are change points (trend features) identified through non-causal analysis, which may require control (control points).

[0033] FIGS. 8a-8g illustrate plots (800, 815, 830, 845, 855, 875, and 890) of various filtering methods applied to the trends of FIGS. 4. Each of FIGS. 8a-8g has an x-axis in time units and a y-axis in count units. Each of the plots (800, 815, 830, 845, 855, 875, and 890) includes inflection points (730, 740). In the examples described below, the following table (Table 1) summarizes the various filters and adjustable parameters corresponding to the plots of FIGS. 8a-8g.

[0034]

[0035] Each of the plots (800, 815, 830, 845, 855, 875, and 890) shows the output trend of the result of applying each filter type (indicated above each plot) over a range of parameter values. For each filter type and each range of filter parameter values, variations in noise reduction, signal offset, signal gain, and trend delay can be observed. For example, in the plot (800) for the IIR filter and the plot (815) for the averaging filter, an increasing trade-off between delay and noise reduction can be observed. Similarly, for the plots (830 and 845) corresponding to the Butterworth filter and the Elliptic filter, respectively, high noise reduction and large delay are observed for specific values ​​of each filter.

[0036] Plot (890) in FIG. 8g shows an enlarged detail of plot (875) in FIG. 8f to more clearly show the noise reduction and other changes to the trends provided by various configurations of the combined Savitzky-Golay and 4-sample averaging filter operation. In particular, for most configurations, the detection times of the inflection point (740) are determined very consistently without the delay observed for certain other filters.

[0037] FIG. 9 illustrates a plot (900) comparing the calculated endpoint latencies of the trend in FIG. 4 when various filters are applied. For example, since the exemplary trend is approximately a second-order polynomial, a causal implementation of Savitzky-Golay filtering ("Smooth" filter) with the polynomial order fixed at '2' is applicable and generally provides low latency results. For other trends approaching different orders of the polynomial, the polynomial order of the filter may be changed. Also, in the case of Savitzky-Golay filtering, inflection points must be appropriately separated into multiple samples according to the filter window length. In the case of Butterworth and elliptic low-pass filters, the noise spectrum (noise power at about 3.5 Hz is ~40 dB higher than the DC noise) suggests that low-pass filters can be effective for processing, but these filters generally result in an overall increase in latency due to increased filter complexity. The filter specified as Smoothing(Avg2(n=4)) achieves low latency and is generally insensitive to the Smoothing() length parameter due to a combination of the benefits of appropriate model-based estimation (quadratic polynomial) and short run averages.

[0038] FIGS. 10a and 10Bb illustrate plots (1000 and 1050) of the trends in FIG. 4 filtered with and without conditioning. The legend of FIG. 9 also applies to FIGS. 10a and 10b. Without conditioning of the signals, multi-filter implementations may be affected by transients and other responses that interfere with the expected performance metrics of the filters (latency, smoothing, gain, ringing, stabilization time, etc.) when applied to the trends. For all trends shown in plot (1000), transients and ringing easily appeared within 1 second. Conditioning may include one or more data manipulations on the trends to mitigate undesirable interferences. Conditioning may include scaling, normalization, standardization, ratio adjustment, offset adjustment, or other mathematical operations. For example, conditioning applied to the trends in plot (1050) includes subtracting the first value of the trend from all subsequent values ​​before applying to the filters. In plot (1050), it can be observed that there is no transient phenomenon or ringing compared to plot (1000). Other adjustments to these same trends may include subtracting the average of several initial values ​​from all subsequent values.

[0039] While the preceding examples relate to the processing and analysis of trend data, such as single values ​​over a time range or scalar trend data; however, the methods and processes can be applied to multi-value data (so-called vector trend data) where multiple values ​​are associated with each time point. This type of data is generally more associated with IEP optical data. FIGS. 11a and 11b are plots of representative IEP optical signal data collected and processed in various ways according to the present disclosure. Both figures have an x-axis in wavelength units and a y-axis in count units. Plot (1100) in FIG. 11a contains samples of IEP spectra collected at two different times. Specifically, data (1110) is from an earlier time than data (1120). A comparison of data (1110 and 1120) shows that there are complex differences in the signal over a wavelength range from ~325 nm to 800 nm. These differences can be made more apparent through processing, filtering, and conditioning as discussed herein. In plot (1150) of FIG. 11b, data (1160) is the result of subtracting data (1110 and 1120) and applying offset adjustments to each data set before subtraction. The complex differences in the signals are more clearly expressed by the oscillating set of features, but strong residual signals (e.g., spikes near the 520 nm peak) appear as a result of variations in the flash ramp used during data acquisition. For control where detecting the peak near 520 nm is important, the residual signals make such detection difficult. Data (1170) is a filtered version of data (1160) to which a Savitzky-Golay filter has been applied. Similar to the filtered trends in FIG. 8a-8g, noise reduction can be observed, but significant phase shifts have been introduced by this filtering process.Again, similar to the trends in Figures 8a to 8g, various filters can be reviewed to determine the filter that has the best desired result, such as minimum latency or maximized attribute detection capability.

[0040] FIG. 12 is a block diagram of an optical system (1200) including a spectrometer (1210) and certain related systems according to one embodiment of the present disclosure. The spectrometer (1210) may incorporate the systems, features, and methods disclosed herein for the benefit of measuring, characterizing, and processing optical signals from semiconductor processes and may be associated with the spectrometer (160) of FIG. 1. The spectrometer (1210) may receive optical signals from an external optical device (1230) via fiber optic cable assemblies (157 or 159) and, after integration and conversion, transmit the data to external systems (1220), such as the output (180) of FIG. 1, which may also be used to control the spectrometer (1210) by selecting an operating mode or controlling the integration timing, for example, as defined herein. The spectrometer (1210) may include an optical interface (1240), such as a micro-assembly (SMA), a ferrule connector (FC), a fiber optic connector, or other opto-mechanical interface. Additional optical components (1245), such as slits, lenses, filters, and gratings, may function to form, guide, and color-separate the received optical signals and send them to a sensor (1250) for integration and conversion. The sensor (1250) may be associated with the sensor (200) of FIG. 2. Low-level functions of the sensor (1250) may be controlled by elements such as an FPGA (1260) and a processor (1270). Following the conversion from light to electricity, the analog signals can be sent to an A / D converter (1280) and converted from an electrical analog signal to an electrical digital signal, which can then be stored in memory (1290) for immediate or subsequent use and transmission to, for example, external systems (1220) (see signal processor (170) in FIG. 1).Specific interfaces and relationships are indicated by arrows, but not all interactions and control relationships are shown in FIG. 12. Spectral data illustrated in FIG. 3 may be collected, stored, and / or operated in / by means of one or more steps of the process (500) of FIG. 5, for example, according to one or more memory / storage (1290), FPGA (1260), processor (1270), and / or external system (1220). As such, the spectrometer (1200) may be configured to process signals by testing and evaluating different filters and / or combinations of filters having different values ​​based on detection consistency and latency (i.e., may be designed, configured, or programmed with logic and / or features necessary to perform tasks or tasks). The spectrometer (1210) also includes a power supply (1295), which may be a conventional AC or DC power supply included in the spectrometer.

[0041] FIG. 13 illustrates a computing device (1300) that can be used in processes disclosed herein, such as identifying signals in spectral data and processing signals. The computing device (1300) may be a spectrometer or part of a spectrometer, such as the spectrometer (160 or 1210) disclosed herein. The computing device (1300) may include at least one interface (1332), a memory (1334), and a processor (1336). The interface (1332) includes hardware, software, or a combination thereof required, for example, to receive raw spectral data and to transmit processed spectral data. Part of the interface (1332) may also include hardware, software, or a combination thereof required to transmit analog or digital electrical signals. The interface (1332) may be an existing interface that communicates through various communication systems, connections, buses, etc., according to protocols such as standard protocols or proprietary protocols (for example, the interface (1332) may support I2C, USB, RS232, SPI, or MODBUS). The memory (1334) is configured to store various software and digital data modalities associated with the computing device (1300). Additionally, the memory (1334) is configured to store an algorithm or a series of operation commands corresponding to algorithms that direct the operation of the processor (1336) when initiated, for example, to identify abnormal signals in spectrum data and process the identified abnormal signals. The process (500) and variations thereof are representative examples of algorithms. Processing may include removing or modifying signal data or different operations. The memory (1334) may be a non-transient computer-readable medium (e.g., flash memory and / or other media).

[0042] The processor (1336) is configured to direct the operation of the computing device (1300). As such, the processor (1336) includes logic necessary to perform the functions described herein to communicate with the interface (1332) and memory (1334) and to identify and process abnormal signals in the spectrum data, as in one or more steps of the method (500). Some of the aforementioned devices, systems, or methods may be implemented or performed in various ways, such as conventional digital data processors or computers, wherein the computers are programmed to perform one or more steps of the methods or store executable programs of software instruction sequences. Software instructions of such programs or code may represent algorithms and may be encoded in a machine-executable form on a non-transient digital data storage medium, such as, for example, a magnetic or optical disk, random access memory (RAM), a magnetic hard disk, flash memory, and / or read-only memory (ROM), which enables various types of digital data processors or computers to perform one, many, or all of the functions, systems, or devices described herein, or one or more steps of the methods described above.

[0043] Some of the disclosed embodiments may relate to computer storage products having a non-transient computer-readable medium having program code for performing the steps of the method described herein or for performing various computer-implemented operations that implement a part of a device or apparatus. As used herein, "non-transient" means all computer-readable media excluding transient propagating signals. Examples of non-transient computer-readable media include, but are not limited to: magnetic media such as hard disks, floppy disks, and magnetic tapes; optical media such as CD-ROM disks; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program code, such as ROM and RAM devices. Examples of program code include both machine code, such as that generated by a compiler, and files containing high-level code that can be executed by a computer using an interpreter. "Configured" means, for example, designed, configured, or programmed using logic, algorithms, processing instructions, and / or features necessary to perform a task or tasks.

[0044] The modifications (changes) and others described above may be made in the optical measurement systems and subsystems described herein without departing from the scope of this specification. For example, while specific examples have been described in relation to semiconductor wafer processing equipment, it should be understood that the optical measurement systems described herein may be applied to other types of processing equipment, such as roll-to-roll thin film processing and solar cell manufacturing, or to any application where high-precision optical measurement may be required. Furthermore, while the specific embodiments discussed herein describe the use of conventional optical analysis devices such as imaging spectrometers, it should be understood that multiple optical analysis devices having known relative sensitivities may be utilized. Additionally, while the term "wafer" has been used herein when describing aspects of the invention, it should be understood that other types of workpieces may be used, such as quartz plates, phase-shift masks, LED substrates, and other non-semiconductor processing-related substrates and workpieces, including solid, gas, and liquid workpieces.

[0045] The exemplary embodiments described herein have been selected and described to best illustrate the principles and practical applications of the invention and to enable those skilled in the art to understand the invention in various embodiments having various modifications suitable for the specific use being considered. The specific embodiments described herein are by no means intended to limit the scope of the invention, as they may be practiced in various modifications and environments without departing from the scope and intent of the invention. Accordingly, the invention is not intended to be limited to the described embodiments but should be understood in the broadest scope consistent with the principles and features described herein.

[0046] The flowcharts and block diagrams in the drawings illustrate the architecture, functions, and operations of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block of a flowchart or block diagram may represent a module, segment, or part of code, which includes one or more executable instructions for implementing a specified logical function(s). Additionally, it should be noted that in some alternative implementations, the functions mentioned in the blocks may occur differently from the order mentioned in the drawings. For example, two blocks depicted consecutively may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order depending on the related functions. Each block of the flowchart and / or block diagram, and combinations of blocks of the flowchart and / or block diagram, may be implemented by a special-purpose hardware-based system or a combination of special-purpose hardware and computer instructions that perform the specified functions or behaviors.

[0047] The terms used herein are for describing specific embodiments and are not intended to limit the scope of the invention. The singular forms used herein (“a,” “an,” and “the”) are intended to include the plural forms unless the context explicitly indicates otherwise. It should be understood that the terms “comprising” and / or “comprising” used herein enumerate the presence of the specified features, integers, steps, actions, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, actions, elements, components, and / or groups thereof.

[0048] As will be understood by those skilled in the art, the present invention may be embodied as a method, a system, or a computer program product. Accordingly, the present invention may take the form of an entire hardware embodiment, an entire software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining both software and hardware aspects generally referred to herein as a “circuit” or “module.” Additionally, the present invention may take the form of a computer program product on a computer-usable storage medium having computer-usable program code embodied on the medium.

[0049] Various aspects of the present disclosure, including the devices, systems, and methods disclosed herein, may be claimed. The aspects disclosed and mentioned herein include the following:

[0050] A. A method for processing spectral data comprises: (1) collecting a time sequence of optical emission spectroscopic data over one or more wavelengths; (2) extracting one or more attributes from the time sequence of optical emission spectroscopic data; (3) analyzing the characteristics of one or more attributes; (4) determining the control of one or more attributes; (5) processing one or more attributes according to a predetermined set of filters, control, and characteristics; and (6) selecting a filter configuration for processing spectral data based on processing one or more attributes.

[0051] B. A method for controlling a semiconductor process comprises: (1) collecting optical emission spectroscopic data across one or more wavelengths; (2) processing the data using a pre-selected method selected to provide a minimum process delay when determining an endpoint indication; and (3) modifying the semiconductor process based on processing the data.

[0052] C. In a computing device comprising one or more processors performing operations, the operations include: (1) collecting optical emission spectroscopic data over one or more wavelengths; (2) processing the data using a pre-selected method selected to provide a minimum process delay when determining an endpoint indication; and (3) modifying a semiconductor process based on processing the data.

[0053] D. Provides a computer program product having a series of operation instructions stored on a non-transient computer-readable medium that direct the operation of one or more processors when initiated to perform operations for processing spectral data. In one example, the operations include: (1) collecting a time-sequence sequence of optical emission spectroscopic data over one or more wavelengths from a semiconductor process; (2) extracting one or more attributes from the time-sequence sequence of optical emission spectroscopic data; (3) analyzing the characteristics of one or more attributes; (4) determining the control of one or more attributes; (5) processing one or more attributes according to a predetermined set of filters, control, and characteristics; and (6) selecting a filter configuration for processing spectral data based on processing one or more attributes using one or more filters from a predetermined set of filters.

[0054] Each of embodiments A, B, C, and D may have a combination of one or more of the following additional elements: Element 1: The set of filters includes a single filter. Element 2: The set of filters includes at least one filter selected from a group of filters consisting of an infinite impulse response filter, an averaging filter, a Butterworth filter, an elliptic filter, a Savitzky-Golay smoothing filter, and a Savitzky-Golay smoothing / averaging filter. Element 3: Processing one or more attributes includes changing the parameter values ​​of at least one filter in the set of filters. Element 4: The collecting, extracting, analyzing, determining, and processing one or more attributes are performed in real time. Element 5: The filter configuration includes filters from a predetermined set of filters, and the processing of the spectral data is performed in real time. Element 6: The selection is based on consistency and latency in detecting one or more attributes while processing one or more attributes. Element 7: One or more attributes include one or more trends, one or more features, or a combination of one or more trends and one or more features. Element 8: Optical emission spectroscopic data is received by a spectrometer from a processing tool. Element 9: The filter configuration includes filters from a predetermined set of filters. Element 10: A predetermined method is determined by extracting one or more attributes from a time-sequence sequence of optical emission spectroscopic data, analyzing the characteristics of one or more attributes, determining the control of one or more attributes, processing one or more attributes according to the predetermined set of filters, characteristics, and control, and selecting a predetermined method based on processing one or more attributes. Element 11: One or more attributes include one or more trends, one or more features, or a combination of one or more trends and one or more features. Element 12: Optical emission spectroscopic data is collected from a semiconductor process.Element 13: A pre-selected method is selected by extracting one or more attributes from a time-sequence sequence of optical emission spectroscopic data, analyzing the characteristics of one or more attributes, determining the control of one or more attributes, processing one or more attributes according to a predetermined set of filters, characteristics, and control, and selecting a pre-selected method based on processing one or more attributes. Element 14: One or more attributes include one or more trends. Element 15: One or more attributes further include one or more features or a combination of one or more trends and one or more features. Element 16: The computing device is a spectrometer.

Claims

Claim 1 A method for processing a real-time time-sequence sequence of optical emission spectroscopic data for semiconductor process control, comprising: collecting a non-real-time time-sequence sequence of optical emission spectroscopic data over one or more wavelengths; extracting one or more attributes from the non-real-time time-sequence sequence of optical emission spectroscopic data; analyzing the characteristics of one or more attributes; determining the conditioning of one or more attributes; processing one or more attributes according to a predetermined set of filters, conditioning, and characteristics, wherein each of the predetermined set of filters has a range of corresponding predetermined parameters; and, based on processing one or more attributes, selecting parameters from a range of predetermined parameters corresponding to one of the filters from the predetermined set of filters to process the real-time time-sequence sequence. Claim 2 A method for processing a real-time time-sequence sequence of optical emission spectroscopic data according to claim 1, wherein the set of filters includes a single filter. Claim 3 A method for processing a real-time time-sequence sequence of optical emission spectroscopic data according to claim 1, wherein the set of filters comprises at least one filter selected from a group of filters consisting of an infinite impulse response filter, an averaging filter, a Butterworth filter, an elliptic filter, a Savitzky-Golay smoothing filter, and a Savitzky-Golay smoothing / averaging filter. Claim 4 A method for processing a real-time time sequence of optical emission spectroscopic data according to claim 1, wherein processing one or more attributes includes changing the parameter values ​​of at least one filter among a set of filters. Claim 5 A method for processing a real-time time-sequence sequence of optical emission spectroscopic data, wherein, in claim 1, the collecting, extracting, analyzing, determining, and processing one or more attributes are performed in real time. Claim 6 A method for processing a real-time time-sequence sequence of optical emission spectroscopic data, wherein, in paragraph 5, processing the real-time time-sequence sequence is performed in real time. Claim 7 A method for processing a real-time time-ordered sequence of optical emission spectroscopic data, wherein the selection is based on consistency and latency of detecting one or more attributes while processing one or more attributes in claim 1. Claim 8 A method for processing a real-time time-ordered sequence of optical emission spectroscopic data according to claim 1, wherein one or more attributes include one or more trends, one or more features, or a combination of one or more trends and one or more features. Claim 9 A method for processing a real-time time sequence of optical emission spectroscopic data, wherein the optical emission spectroscopic data is received by a spectrometer from a processing tool. Claim 10 delete Claim 11 A method for controlling a semiconductor process comprises: collecting a non-real-time time-sequence sequence of optical emission spectroscopic data across one or more wavelengths; processing a real-time sequence of optical emission spectroscopic data using a predetermined method determined to provide a minimum process delay when determining an endpoint indication; and modifying the semiconductor process based on processing the real-time time-sequence sequence, wherein the predetermined method comprises: extracting one or more attributes from the non-real-time time-sequence sequence; analyzing characteristics of one or more attributes; determining a control of one or more attributes; processing one or more attributes according to a predetermined set of filters, control and characteristics, wherein each of the predetermined sets of filters has a range of corresponding predetermined parameters; and, based on processing one or more attributes, selecting parameters from a range of predetermined parameters corresponding to one of the filters from the predetermined set of filters for processing the real-time time-sequence sequence. Claim 12 delete Claim 13 A method for controlling a semiconductor process according to claim 11, wherein one or more attributes include one or more trends, one or more features, or a combination of one or more trends and one or more features. Claim 14 In paragraph 11, a method for controlling a semiconductor process in which optical emission spectroscopic data is collected from the semiconductor process. Claim 15 A computing device comprising: one or more processors, wherein the one or more processors perform operations including: collecting a non-real-time time-sequence sequence of optical emission spectroscopic data over one or more wavelengths; processing a real-time time-sequence sequence of optical emission spectroscopic data using a predetermined method determined to provide a minimum process delay when determining an endpoint indication; and modifying a semiconductor process based on processing the real-time time-sequence sequence, wherein the predetermined method comprises: extracting one or more attributes from the non-real-time time-sequence sequence; analyzing characteristics of one or more attributes; determining a control of one or more attributes; processing one or more attributes according to a predetermined set of filters, control, and characteristics—each of the predetermined set of filters having a corresponding range of predetermined parameters—and, based on processing one or more attributes, selecting parameters from a range of predetermined parameters corresponding to one of the filters from a predetermined set of filters for processing the real-time time-sequence sequence. Claim 16 delete Claim 17 In paragraph 15, a computing device in which one or more attributes include one or more trends. Claim 18 A computing device according to claim 17, wherein one or more attributes further comprise one or more features or a combination of one or more trends and one or more features. Claim 19 In paragraph 15, the computing device is a spectrometer, a computing device. Claim 20 A computer program product having a series of operation instructions stored in a non-transient computer-readable medium that directs the operation of one or more processors when initiated to perform operations for processing a real-time time-sequence sequence of optical emission spectroscopic data for semiconductor process control, wherein the operations include: collecting a non-real-time time-sequence sequence of optical emission spectroscopic data over one or more wavelengths from a semiconductor process; extracting one or more attributes from the non-real-time time-sequence sequence of optical emission spectroscopic data; analyzing the characteristics of one or more attributes; determining the control of one or more attributes; processing one or more attributes according to a predetermined set of filters, control, and characteristics, wherein each of the predetermined set of filters has a range of corresponding predetermined parameters; and, based on processing one or more attributes, selecting one of the filters from the predetermined set of filters and parameters from a range of corresponding predetermined parameters to process the real-time time-sequence sequence.

Images