Predictive calibration scheduling apparatus and method
The predictive calibration system for optical devices in light sources optimizes calibration timing based on degradation metrics, reducing downtime and maintaining accuracy by scheduling calibrations only when required.
Patent Information
- Application Number
- JP2024105902
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-02-19
- Filing Date
- 2024-07-01
- Publication Date
- 2024-11-07
- Estimated Expiration
- 2041-07-28
AI Technical Summary
Conventional calibration methods for optical devices in light sources, such as excimer lasers, are inefficient and often result in unnecessary downtime due to calibrations performed at predetermined intervals, leading to either excessive downtime or operational inaccuracies when calibrations are missed.
A predictive calibration system that uses a predictive controller to analyze degradation metrics based on past performance data to determine the optimal timing for calibrations, adjusting operating parameters to maintain accuracy without unnecessary interruptions.
The system reduces downtime by scheduling calibrations only when necessary, ensuring consistent performance and accuracy of optical devices by anticipating degradation thresholds.
Smart Images

Figure 2024157560000001_ABST
Abstract
Description
[Technical field]
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority to U.S. patent application Ser. No. 63 / 066,888, entitled "PREDICTIVE CALIBRATION SCHEDULING APPARATUS AND METHOD," filed on August 18, 2020, and U.S. patent application Ser. No. 63 / 151,411, entitled "PREDICTIVE CALIBRATION SCHEDULING APPARATUS AND METHOD," filed on February 19, 2021, both of which are incorporated by reference in their entireties herein.
[0002] The disclosed subject matter relates to an apparatus for predicting when calibration of an optical device in a light source should be performed. [Background technology]
[0003]
[0003] One type of gas discharge light source used in photolithography is called an excimer light source or laser. Typically, an excimer laser uses a combination of one or more noble gases, which may include argon, krypton, or xenon, and a reactive gas, which may include fluorine or chlorine. Excimer lasers can form excimers, or pseudo-molecules, under the appropriate conditions of electrical simulation (energized) and high pressure (of the gas mixture), which only exist in the energized state. The energized excimers generate amplified light in the ultraviolet range. An excimer light source can use a single gas discharge chamber or multiple gas discharge chambers. When an excimer light source is operating, it generates a deep ultraviolet (DUV) light beam. DUV light can include wavelengths, for example, from about 100 nanometers (nm) to about 400 nm.
[0004]
[0004] The DUV light beam can be directed to a photolithography exposure tool or scanner, which is a machine that applies a desired pattern to a target portion of a substrate (e.g., a silicon wafer). The DUV light beam interacts with a projection optical system, which projects the DUV light beam through a mask onto the photoresist of the wafer. In this way, one or more layers of chip designs are patterned onto the photoresist, and the wafer is then etched and cleaned. Summary of the Invention
[0005] In some general aspects, a method for scheduling calibration for an optical device in a light source is implemented, the method including receiving characteristics associated with the optical device while the optical device is being calibrated, calculating a current degradation metric that models a behavior of the optical device based at least on the optical device characteristics, estimating when degradation of the optical device will exceed a threshold based on the current degradation metric, and scheduling a calibration of the optical device based at least in part on the estimation of degradation of the optical device.
[0006] Implementations may include one or more of the following features: For example, a calibration of a light device may be scheduled by determining when a calibration of the light device should be performed.
[0007]
[0007] The current degradation metric may be calculated based on a calibration characteristic.
[0008]
[0008] The optical device may be configured to measure a spectral characteristic of a light beam generated by the light source. The measured spectral characteristic of the light beam may be the wavelength of the light beam.
[0009]
[0009] During the calibration of an optical device, an optical device characteristic can be received by receiving an error measurement associated with a spectral characteristic of a light beam generated by a light source, the error measurement being a difference between the measured spectral characteristic of the light beam and a reference spectral characteristic.
[0010]
[0010] The current degradation metric may be calculated by estimating how much the optical device characteristic changes in response to changes in usage of the optical device. Estimating how much the optical device characteristic changes in response to changes in usage of the optical device may include estimating a data-based error for a previous value of the optical device characteristic and a previous degradation metric.
[0011]
[0011] The current degradation metric may be calculated by assuming that the optical device characteristic changes linearly with respect to the change in the optical device usage. The current degradation metric may be calculated by estimating the error in the optical device characteristic based on a linear function, where the previous degradation metric is the slope of the linear function, and the linear function includes the previous value of the optical device usage and the cumulative sum of the error in the optical device characteristic. The current degradation metric may be calculated by estimating the error in the optical device characteristic. The current degradation metric may be calculated by analyzing how the optical device characteristic has a trend.
[0012]
[0012] The degradation metric can model the trend behavior of an optical device.
[0013]
[0013] The calibration of the optical device may be scheduled by determining whether the optical device characteristics change nonlinearly with respect to the usage of the optical device. Determining whether the optical device characteristics change nonlinearly with respect to the usage of the optical device includes analyzing a confidence in a current degradation metric.
[0014]
[0014] Estimating when degradation of the optical device will exceed a threshold may include estimating when an error in an optical device characteristic associated with the optical device will exceed a threshold. Scheduling a calibration of the optical device based on the estimation may include estimating an amount of additional usage of the optical device that would be tolerable before requiring calibration.
[0015]
[0015] The method may include, after scheduling a calibration of the optical device, updating an optical device characteristic associated with the optical device when the next scheduled calibration is performed and assigning a current degradation metric to a previous degradation metric.
[0016]
[0016] The degradation metric may correspond to a locally linear approximation of the slope of the optical device characteristic.
[0017]
[0017] The optical device may be configured to estimate the energy of the light beam generated by the light source. The optical device characteristic may be received while the optical device is being calibrated by receiving a measurement scale associated with the optical device that performs a conversion from a direct measurement value to an indirect value, the direct measurement value relating to the light beam generated by the light source while the optical device is being calibrated. By estimating when a drift in the optical device characteristic will exceed a drift threshold, a degradation of the optical device may be estimated to exceed a threshold. By calculating a drift threshold based on an error threshold associated with the indirect value, a drift in the optical device characteristic may be estimated to exceed a drift threshold.
[0018]
[0018] In another general aspect, a calibration system includes a calibration device configured to calibrate an optical device configured to measure an aspect related to a light beam generated by a light source, and a predictive controller in communication with the calibration device. The predictive controller includes an input module, an adaptation module, and an output module. The input module is configured to receive measured characteristics from the optical device associated with the measured aspect of the light beam while the optical device is being calibrated. The adaptation module is configured to receive the measured characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the measured characteristics, and estimate when degradation of the optical device will exceed a threshold based on the current degradation metric. The output module is configured to schedule a calibration of the optical device based on the estimate from the adaptation module, and instruct the calibration device to calibrate the optical device based on the schedule.
[0019]
[0019] Implementations may include one or more of the following features: For example, the measured characteristic associated with the optical device includes a measurement of a spectral characteristic of a light beam generated by a light source.
[0020]
[0020] The calibration apparatus may include a calibration material having an optical transition profile with known energy transitions. The optical device may include a spectral analysis module configured to sense a spectral profile of a light beam generated by the light source, and the calibration apparatus may be configured to calibrate the spectral analysis module using the calibration material. The estimation of the degradation of the optical device may be an indication of a drift in the calibration of the spectral analysis module. Configuring the calibration apparatus to calibrate the optical device may include adjusting one or more operating parameters of the spectral analysis module.
[0021]
[0021] The degradation metric may correspond to a local linear approximation of the slope of the trend of the measured characteristic.
[0022]
[0022] In another general aspect, a predictor is in communication with a calibration device configured to calibrate an optical device configured to measure aspects related to a light beam generated by a light source. The predictor includes an input module configured to receive measured characteristics from the optical device associated with the measured aspects of the light beam while the optical device is being calibrated, an adaptation module configured to receive the measured characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the measured characteristics, and estimate when degradation of the optical device will exceed a threshold based on the current degradation metric, and an output module configured to schedule a calibration of the optical device based on the estimation from the adaptation module and instruct the calibration device to calibrate the optical device based on the schedule.
[0023]
[0023] In another general aspect, a calibration system includes a calibration apparatus configured to calibrate an optical device configured to measure an aspect related to a light beam generated by a light source, and a predictive controller in communication with the calibration apparatus. The predictive controller includes an input module configured to receive characteristics associated with the optical device from the optical device while the optical device is being calibrated, an adaptation module configured to receive the optical device characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics, and estimate when degradation of the optical device will exceed a threshold based on the current degradation metric, and an output module configured to schedule a calibration of the optical device based on the estimation from the adaptation module, and instruct the calibration apparatus to calibrate the optical device based on the schedule.
[0024]
[0024] Implementations may include one or more of the following features. For example, the optical device may be configured to provide an estimate of the energy of the optical beam based on a measurement scale. The optical device characteristic may include the measurement scale. The measurement scale may enable conversion of a direct measurement of a characteristic of the optical beam to an indirect value of the energy of the optical beam.
[0025]
[0025] The calibration apparatus may include a power meter that receives at least a portion of the light beam directed to the optical device, the power meter outputting a measured power, and a processor. The processor is configured to calculate the energy in the pulses of the light beam based on the pulse repetition rate of the light source and the measured power output from the power meter, and compare the calculated energy to the estimated energy of the light beam from the optical device. The processor may be configured to estimate a drift in the measurement scale of the optical device based on the comparison. The calibration apparatus configured to calibrate the optical device may include adjusting the measurement scale of the optical device based on the comparison. Degradation of the optical device may be indicated by an estimate of the drift in the measurement scale of the optical device. The degradation metric may correspond to a local linear approximation of the slope of the trend of the measurement scale of the optical device.
[0026]
[0026] In another general aspect, a prediction device in communication with the calibration device is configured to calibrate an optical device configured to measure an aspect related to a light beam generated by a light source. The prediction device includes an input module configured to receive characteristics associated with the optical device from the optical device while the optical device is being calibrated, an adaptation module configured to receive the optical device characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics, and estimate when degradation of the optical device will exceed a threshold based on the current degradation metric, and an output module configured to schedule a calibration of the optical device based on the estimation from the adaptation module and instruct the calibration device to calibrate the optical device based on the schedule. [Brief description of the drawings]
[0027] [Figure 1] FIG. 1 is a block diagram of a calibration system including an optical device, a light source generating a light beam, and a calibration apparatus and a predictive controller. [Diagram 2]
[0028] 2 is a block diagram of an implementation of the calibration system of FIG. 1 including a calibration device and an implementation of a predictive controller including an input module, an adaptation module, and an output module. [Diagram 3]
[0029] 2 is a block diagram of an implementation of the optical device of FIG. 1 that is a spectral analysis module that includes a light separation device and is configured to sense the spectral characteristics of the light beam generated by the light source of FIG. 1. [Figure 4]
[0030] 4 is a block diagram of an implementation of the spectral analysis module of FIG. 3, including an implementation of the optical isolation device of FIG. [Diagram 5]
[0031] 2 is a block diagram of an implementation of a light device, a light source, and the calibration system of FIG. 1, the implementation of the calibration system including a predictive controller and an implementation of the calibration apparatus of FIG. [Figure 6]
[0032] This is an implementation form of the calibration device in FIG. 5, including a detector and a cell for holding a calibration material. [Figure 7]
[0033] It is a flowchart of a procedure for scheduling calibration regarding the optical device in FIG. 1, which is implemented according to the implementation form of the calibration system in FIG. 1 or the calibration system in FIG. 5. [Figure 8A]
[0034] A first graph showing the behavior of the optical path length of the first spectral analysis module associated with the first light source with respect to the usage amount of the first spectral analysis module, and a second spectral analysis module different from the first spectral analysis module, which is associated with a second light source different from the first light source, showing the behavior of the optical path length of the second spectral analysis module with respect to the usage amount of the second spectral analysis module. [Figure 8B]
[0035] A first graph showing the behavior of the cumulative sum of the characteristics measured by the first spectral analysis module in FIG. 8A with respect to the same usage amount of the first spectral analysis module in FIG. 8A, and a second graph showing the behavior of the cumulative sum of the characteristics measured by the second spectral analysis module in FIG. 8A with respect to the same usage amount of the second spectral analysis module in FIG. 8A. [Figure 9]
[0036] It is a graph showing the operation of the procedure in FIG. 7 as the cumulative sum value of the spectral analysis module in FIG. 3 with respect to the usage amount of the spectral analysis module in the local window of the usage amount. [Figure 10]
[0037] It is a flowchart of a procedure implemented after scheduling calibration regarding the optical device using the procedure in FIG. 7. [Figure 11]
[0038] A photolithography system including an illumination system for generating an optical beam guided to a photolithography exposure device, the optical device in FIG. 1, and the calibration system in FIG. 1. [Figure 12]
[0039] It is a block diagram of an implementation form of the optical device and the calibration system in FIG. 1. [Figure 13]
[0040] 13 is a flow chart of a procedure for scheduling a calibration using the optical device and calibration system of FIG. 12, and optionally using the optical device and calibration system of FIGS. 1 and 5. [Figure 14]
[0041] 14 is a graph showing the operation of the procedure of FIG. 13, and in particular the measurement scale values of the optical device of FIG. 12, versus the usage of the optical device of FIG. 12 in a local window of usage. [Figure 15]
[0042] 14 is a flow chart of a procedure performed after scheduling a calibration for the optical device of FIG. 12 using the procedure of FIG. 13. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0028]
[0043] 1, a calibration system 100 is disposed relative to an optical device 105. The optical device 105 is configured to measure or sense an aspect L105 of a light beam 101 generated by a light source 102, which is provided to and used by an output device 104. The measured or sensed aspect L105 of the light beam 101 may be provided to a controller associated with the light source 102, the output device 104, or one or more of the light source 102 and the output device 104. The optical device 105 receives a diagnostic light beam 103. The diagnostic light beam may be the light beam 101 or a portion of the light beam 101 split off from the light beam 101.
[0029]
[0044] During operation, various undesirable effects may degrade the accuracy of measurements performed by the optical device 105 as the optical device 105 is used. For example, the optical device 105 may rely on the optical path length that the diagnostic optical beam 103 experiences within the optical device 105 to measure the spectral features (e.g., central wavelength) of the optical spectrum of the optical beam 101. The measured aspect L105 may correspond to this measured spectral feature. The optical path length is related to the geometric length of the path that the diagnostic optical beam 103 follows through the optical device 105 as well as the refractive index of the medium through which the diagnostic optical beam 103 propagates. With usage, the optical path length within the optical device 105 may change, and these changes may be due to one or more changes in the refractive index and the geometric path length. The changes in the optical path length experienced by the diagnostic optical beam 103 cause errors in the measurements of the spectral features of the diagnostic optical beam 103 performed by the optical device 105. As a result, it becomes difficult to control the spectral features and optical spectrum of the optical beam 101. The output device 104 may require such control.
[0030]
[0045] In order to maintain the accuracy of the measurements performed by the optical device 105, the optical device 105 must be properly calibrated. To this end, the calibration system 100 includes a calibration apparatus 110 configured to calibrate the optical device 105. During calibration, one or more operating parameters of the optical device 105 are adjusted to compensate for this degradation in the optical device 105 due to use. Calibration of the optical device 105 may take several seconds or more than a minute. For example, calibration may take one minute, two minutes or even five minutes. Furthermore, during calibration, the light beam 101 is not available for use by the output device 104. Therefore, the operating time of the output device 104 is stopped to perform the calibration.
[0031]
[0046] Previously, the calibration device 110 would initiate a calibration of the optical device 105 based solely on a preset value of the usage of the optical device 105. For example, if the light source 102 is a pulsed light source that generates pulses of the optical beam 101, the usage of the optical device 105 can be determined by the total number of pulses of the optical beam 101, and thus the total number of pulses of the diagnostic optical beam 103 sensed by the optical device 105. Thus, in conventional calibration systems, the calibration device 110 could perform a calibration every X pulses of the optical beam 101. In some cases, the value of X could be on the order of hundreds of millions. For example, for some light sources 102, the calibration device would perform a calibration of the optical device 105 every 300 million pulses of the optical beam 101.
[0032]
[0047] Initiating a calibration based solely on a pre-configured or predetermined value of the usage of the light device 105 may result in calibrating the light device 105 when calibration is unnecessary or not calibrating the light device 105 when calibration is necessary. Calibrating the light device 105 unnecessary means excessive and unnecessary downtime for the output device 104. In contrast, not calibrating the light device 105 when calibration is necessary means that the light beam 101 may be operating outside of the specifications required by the output device 104.
[0033]
[0048] Accordingly, the calibration system 100 includes a predictive controller 130 in communication with the calibration apparatus 110 and the light device 105. The predictive controller 130 is configured to schedule calibration of the light device 105 when calibration is needed and not schedule calibration when calibration is not needed. Furthermore, the predictive controller 130 determines the calibration schedule during operation of the light device 105 and not based solely on a pre-set value of usage of the light device 105. The predictive controller 130 makes decisions regarding whether and when to calibrate based on previous or past performance of the light device 105. The predictive controller 130 determines when to schedule a subsequent or next calibration of the light device 105 between calibrations of the light device 105 and in real-time during operation of the light source 102.
[0034]
[0049] In the drawings, solid lines connecting two elements indicate data paths along which data (e.g., information and / or command signals) may flow. Data may flow via wireless or wired connections. Thus, data flowing between the predictive controller 130, the calibration apparatus 110, and the optical device 105, or within each of these components, may be via wireless or wired connections. Additionally, dotted lines indicate optical paths along which light may propagate.
[0035]
[0050] To this end, and with further reference to FIG. 2, the predictive controller 130 includes an input module 232, an adaptation module 234, and an output module 236. In a further implementation, the predictive controller 130 may include or have access to a memory 238. Each time a measured characteristic is received at the input module 232, the memory 238 may be configured to store the value of the measured characteristic. The memory 238 may be configured to store output instructions with a calibration schedule from the output module 236, other information from the optical device 105, or even degradation metrics from the adaptation module 234, such information being available for various uses by the predictive controller 130 during operation. The memory 238 may be a read-only memory and / or a random access memory and may provide a suitable storage device for tangibly embodying computer program instructions and data.
[0036]
[0051] The input module 232 communicates with the optical device 105 and receives information regarding previous or past performance of the optical device 105. In general, as discussed in more detail below, the input module 232 is configured to receive from the optical device 105 measured characteristics associated with measured aspects of the diagnostic light beam 103. The adaptation module 234 is configured to receive the measured characteristics from the input module 232 and estimate when degradation of the optical device 105 will exceed a threshold based on the measured characteristics. In addition, the adaptation module 234 can estimate when degradation of the optical device 105 will exceed a threshold based not only on the current measured characteristics but also on previous measured characteristics. For example, the adaptation module 234 can estimate when degradation of the optical device 105 will exceed a threshold based on a cumulative sum of the current measured characteristics and all or some previous measured characteristics, such cumulative sum being stored in the memory 238. The output module 236 is configured to schedule a calibration of the optical device 105 based on the estimation from the adaptation module 234.
[0037]
[0052] The predictive controller 130 may include or have access to one or more programmable processors, each capable of executing a program of instructions to perform desired functions including receiving measured characteristics from the optical device 105 and generating output instructions, including a calibration schedule, to the calibration apparatus 110. The predictive controller 130 (and any one of the modules 232, 234, 236) may be implemented in any of digital electronic circuitry, computer hardware, firmware, or software. The calibration apparatus 110 may also include one or more input devices and one or more output devices.
[0038]
[0053] 3, in some implementations, the optical device 105 is a spectral analysis module 305. The spectral analysis module 305 includes an optical isolation device 342 and a sensor 344. In other implementations, the spectral analysis module 305 includes other components not shown, such as one or more diffusers, windows, and reflective or refractive optics. Implementations of the optical isolation device 342 are described below with reference to FIG.
[0039]
[0054] 3, the measured characteristics associated with the spectral analysis module 305 include measurements of the spectral characteristics of the diagnostic light beam 103 generated by the light source 102. The spectral analysis module 305 is configured to sense the spectral characteristics of the diagnostic light beam 103 by sensing and analyzing a spectral profile 308 of the diagnostic light beam 103. The spectral profile 308, also referred to as the optical spectrum, is the spectral shape or intensity spectrum of the diagnostic light beam 103 (and light beam 101) as a function of the wavelength (or frequency) of the diagnostic light beam 103. Thus, the spectral profile 308 contains information about how the optical energy or power of the light beam 101 or 103 is distributed across various wavelengths (or frequencies).
[0040]
[0055] The diagnostic light beam 103 is directed through an aperture 341 of the spectral analysis module 305 to the optical isolation device 342. The diagnostic light beam 103 travels along an optical path defined by an optical path length within the optical isolation device 342. The optical isolation device 342 interacts with the diagnostic light beam 103 and outputs a number of spatial components 306 that correspond to the spectral components of the diagnostic light beam 103. The spectral components reside in a spectral profile 308 of the light beam 103. In operation, the spectral analysis module 305 can generate the spectral profile 308 by varying the wavelength of the light beam 103 and scanning it across a particular wavelength range.
[0041]
[0056] The spectral components correspond to how the optical energy or power values of light beam 103, shown as spectral intensity 303i, are distributed across various wavelengths 303w. The spatial components 306 correspond to the values of spectral intensity 303i mapped into two-dimensional space. In this manner, optical separator 342 converts the spectral information, e.g., wavelength, of light beam 103 into spatial information that can be sensed or detected by sensor 344. The conversion maps the spectral information to different locations in space such that the spectral information can be observed by sensor 344. The transmission of optical separator 342 is shown as interference fringe pattern 307, which produces spectral profile 308.
[0042]
[0057] The sensor 344 receives and senses the output spatial components 306. The sensor may be defined by a plane that generally indicates the active area of its sensing region. The plane of the sensing region may be perpendicular to the propagation direction of the spatial components 306. The sensor 344 may be a detector that receives and senses the output spatial components 306. For example, one type of suitable detector that can be used to measure along one dimension is a linear photodiode array. A linear photodiode array consists of multiple elements of the same size formed in a linear arrangement with equal spacing in one package. The photodiode array is sensitive to the wavelength of the light beam 103, and when the light beam 103 has a wavelength in the deep ultraviolet range, the photodiode array is sensitive to light having a wavelength in the deep ultraviolet range. As another example, the sensor 344 may be a two-dimensional sensor, such as a two-dimensional charge-coupled device (CCD) or a two-dimensional complementary metal-oxide semiconductor (CMOS) sensor. The sensor 344 is capable of reading data at a sufficiently fast speed, for example, about 6 kHz.
[0043]
[0058] The output of the sensor 344 is connected to a control module 350, which measures properties of the spatial components 306 and analyzes the measured properties to calculate an estimate of the spectral characteristics of the light beam 103. As discussed above, the estimates of the spectral characteristics of the diagnostic light beam 103 and the light beam 101 are provided to a controller associated with the light source 102, the output device 104, or one or more of the light source 102 and the output device 104. The control module 350 can be connected to the sensor 344 through a data connection and can further communicate with the light source 102 and / or the output device 104. The control module 350 can perform measurements, analyses, and calculations for each pulse of the light beam 103 or a set of pulses of the light beam. The control module 350 can also be connected to the calibration device 110 and the predictive controller 130 (FIGS. 1 and 2) of the calibration system 100. The control module 350 can send data of the measured characteristics of the light beam 103 (including analysis of the measured characteristics) to the calibration device 110 and the predictive controller 130 via data connections.
[0044]
[0059] In the example of FIG. 3, the calculated estimate of the spectral characteristic is the wavelength λ 305 of the light beam 103. The estimated wavelength λ 305 may correspond to or be based on a central wavelength 309 of the spectral profile 308 of the light beam 103. During operation, the optical separator 342 of the spectral analysis module 305 senses or measures aspects of the spectral profile 308, including the central wavelength 309. The value of the estimated wavelength λ 305 of the light beam 103 may be based on this measured central wavelength 309, which depends on measurements of the spatial components 306 sensed by the sensor 344 and a calibrated model of the optical path length of the optical separator 342. For example, the central wavelength 309 depends on the diameter of the ring visible in the spatial components 306 and the optical path length of the optical separator 342. With usage, i.e., during operation of the spectral analysis module 305, the optical path length in the optical separator 342 may change due to changes in one or more of the refractive index and the geometric path length. Because measurements of the spectral characteristics of the diagnostic light beam 103 depend on a precise model of the optical path length of the optical isolation device 342, changes in the optical path length experienced by the diagnostic light beam 103 introduce errors into the measurements of the spectral characteristics, including the central wavelength 309, of the spectral profile 308 of the diagnostic light beam 103 performed by the spectral analysis module 305.
[0045]
[0060] 2, the calibration system 100 can communicate with the spectral analysis module 305 during operation to calibrate the spectral analysis module 305, thereby compensating for degradation of the spectral analysis module 305 due to use. Specifically, the calibration device 110 calibrates the spectral analysis module 305 (i.e., the optical device in the example of FIG. 3) by adjusting one or more operational parameters of the spectral analysis module 305. For example, the calibration device 110 can adjust any one or more of software parameters, control parameters, and physical parameters of the spectral analysis module 305. In some implementations, the calibration device 110 calibrates the spectral analysis module 305 by modifying or updating a model of the optical path length that the control module 350 in the spectral analysis module 305 uses to estimate the wavelength λ 305 of the diagnostic light beam 103.
[0046]
[0061] The predictive controller 130 schedules the calibration of the spectral analysis module 305. Thus, the predictive controller 130 sends instructions to the calibration device 110 (via the output module 236) as to whether and when to perform the next or subsequent calibration of the spectral analysis module 305. The predictive controller 130 determines when calibration is necessary and also determines when calibration is not necessary.
[0047]
[0062] 4, the implementation 405 of the spectral analysis module 305 includes an optical separator 442 and a sensor 444. The optical separator 442 is an implementation of the optical separator 342 (FIG. 3), and the sensor 444 is an implementation of the sensor 344 (FIG. 3). In the illustrated implementation, the optical separator 442 includes an input lens 461, an optical frequency separator that is an etalon 462, and an output lens 463. In other implementations, other configurations are used. The aperture 441 of the spectral analysis module 405 provides a mechanism for inputting the diagnostic light beam 103. The aperture 441 can be located at the focal plane of the input lens 461. By placing the aperture 441 of the spectral analysis module 440 at the focal plane of the input lens 461, each point from the focal plane acts as a point source, and accordingly, the input lens 461 acts to collimate the diagnostic light beam 103 before entering the etalon 462. An output lens 463 is positioned at the exit of the etalon 462 such that the focal plane of the output lens 463 overlaps with the active area of the sensor 444 .
[0048]
[0063] The etalon 462 includes a pair of partially reflective glass or optical plates 462A, 462B, which can be spaced apart by a short distance (e.g., millimeters to centimeters) with the reflective surfaces facing each other. In other implementations, the etalon 462 includes a single plate with two parallel reflective surfaces. The plates 462A, 462B can be made in a wedge shape to prevent the back surface from producing interference fringes. The back surface can also have an anti-reflective coating. When the light beam 103 passes through the pair of plates 462A, 462B, the light beam 103 is multiple-reflected, generating multiple transmitted beams, which are focused by the output lens 463 and lead to the active area of the sensor 444. The spectral analysis module 405 can include an optical relay between the output lens 463 and the sensor 444 to ensure that the sensor 444 is in the focal plane of the output lens 463.
[0049]
[0064] The etalon 462 interacts with the light beam 103 and outputs a number of spatial components (e.g., spatial components 306 in FIG. 3) corresponding to the spectral components of the light beam 103. The etalon 462 converts the spectral information, including the wavelength of the light beam 103, into spatial information that can be sensed or detected by the sensor 444. The spatial components produced by the etalon 462 are an interference pattern that appears as a set of concentric rings (e.g., spatial components 306). If the intensity distribution of the light beam 103 on the aperture 441 is more uniform, the interference pattern will contain a more uniform intensity distribution. Specifically, the clarity of the rings depends on the reflectivity of the plates 462A, 462B of the etalon 462. Thus, if the reflectivity of the plates 462A, 462B is high and the light beam 103 is a monochromatic light beam, the etalon 462 will produce a set of narrow, bright rings against a dark background. The transmission of etalon 462 as a function of wavelength is an interference fringe pattern (eg, interference fringe pattern 307 in FIG. 3) that produces a spectral profile (eg, spectral profile 308 in FIG. 3) of light beam 103.
[0050]
[0065] The sensor 444 receives and senses the output spatial components. The control module 350 receives the sensed spatial components from the sensor 444 and measures and / or analyzes the received data. The control module 350 sends the measurement and / or analysis data (e.g., the estimated wavelength λ 305) to the calibration system 100 (FIG. 1). The calibration system 100 calibrates the spectral analysis module 405 at scheduled times and predicts the next or subsequent calibration of the spectral analysis module 405 based at least in part on the measurement and / or analysis data. The calibration system can calibrate the spectral analysis module 405 by adjusting one or more operating parameters, such as software parameters, control parameters, and physical parameters, of the spectral analysis module 305. As an example, a model of the optical path length of the etalon 462 can be adjusted. As another example, the position of each plate 462A, 462B of the etalon 462 can be adjusted to adjust the optical path length of the light beam 103 in the etalon 462. In this manner, calibration system 100 modifies the optical path length experienced by diagnostic light beam 103 in spectral analysis module 405. Additionally, because calibration system 100 also predicts the next or subsequent calibrations (thereby reducing the number of calibrations performed compared to conventional methods), fewer calibrations are performed on spectral analysis module 405, thereby increasing the duration over which spectral analysis module 405 may operate.
[0051]
[0066] 5, an implementation 500 of the calibration system 100 includes an implementation 510 of the predictive controller 130 (FIG. 1) and the calibration apparatus 110 (FIG. 1). The optical device 105 (FIG. 1) is positioned relative to the calibration system 500 and includes an aperture 541o through which the optical device 105 receives the diagnostic light beam 103 from the light source 102. The calibration apparatus 510 includes an aperture 541c, a calibration material 512, at least one detector 514, and a calibration control module 516. The calibration material 512 is a known calibration material having an optical transition profile with known energy transitions. In other implementations, the calibration apparatus 510 may include more components than shown in FIG. 5. The calibration control module 516 communicates with the detector 514, the optical device 105, and the predictive controller 130 via one or more data connections 552, each of which may be wired or wireless.
[0052]
[0067] The calibration device 510 receives the diagnostic light beam 103 from the light source 102 through the aperture 541c. The diagnostic light beam 103 interacts with the calibration material 512 to generate a transition (e.g., absorption) profile 515 of the energy transitions of the material 512, and the detector 514 detects the transition profile 515. The transition profile 515 shows a relationship between a feature associated with the interaction between the diagnostic light beam 103 and the material 512 and the wavelength of the diagnostic light beam 103. Thus, the transition profile 515 is shown as a graph of the feature (measured by the detector 514) versus the wavelength of the diagnostic light beam 103. To detect the transition profile 515, the wavelength of the diagnostic light beam 103 is scanned across the energy transitions.
[0053]
[0068] The calibration material 512 can be made of any material with known energy transitions that match the wavelength range of the diagnostic light beam 103 (and also the light beam 101). Thus, if the light beam 103 has a wavelength in the deep ultraviolet range (DUV), the material 512 is selected and the known energy transitions are selected to be in the same DUV range. In some implementations where the light beam 101 and the diagnostic light beam 103 have wavelengths in the DUV range, the material 512 is platinum, which has a known transition with a peak absorption value λref of 193.4 nanometers (nm). The known energy transitions monitored in the material 512 can be electronic, vibrational or rotational transitions. These energy transitions result from the fact that particles (e.g., atoms or molecules) in the material 512 take on specific discrete values of energy called energy levels. The term "energy levels" is used for the energy levels of electrons in atoms, ions or molecules that are bound by the electric field of the atomic nuclei. Alternatively, the term "energy level" may refer to a nuclear energy level or a vibrational or rotational energy level in a molecule.
[0054]
[0069] The material 512 can be in a vapor cell or enclosure. In some implementations, the material 512 is generated as an emitting plasma and is part of a laser galvatron, an optogalvanic sensor that utilizes the phenomenon of resonance between the discharge plasma of the material 512 and the light beam 103. This is discussed in more detail below with reference to FIG.
[0055]
[0070] 6, the implementation 610 of the calibration device 510 includes an aperture 641, a cell 618 that holds a calibration material 612, and a detector 614. The light beam 103 is directed into the calibration device 510 through the aperture 641. The cell 618 is defined and closed by walls 611 made of a rigid and chemically inert material, thereby defining a cavity in which the calibration material 612 is housed and held. The material of the walls 611 does not interact with the material 612 or the light beam 103. Furthermore, the cell 618 includes at least an input window 611i and an output window 611o through which the diagnostic light beam 103 can pass. The cell 618 is sealed from the outside environment (by the design of the walls 611) to contain the material 612 within the cavity defined by the walls 611. Detector 614 is configured to detect characteristics associated with the interaction between material 612 and diagnostic light beam 103 as the diagnostic light beam 103 is directed through material 612 and while the wavelength of diagnostic light beam 103 is scanned across a known energy transition.
[0056]
[0071] In some implementations, the calibration device 610 operates using the principles of optical spectroscopy. In these implementations, the detector 614 is an optical detector, such as a photodiode detector or a photomultiplier tube, that measures the intensity of the diagnostic light beam 103 that has passed through the material 612. This measurement corresponds to a transition profile 515 (FIG. 5).
[0057]
[0072] In other implementations, the calibration device 610 can operate as a galvatron, where the material 612 is a gas or a plasma discharge (generated from electrodes) made of the material 612 and placed in a cell 618. In some implementations, optogalvanic spectroscopy can be used. In optogalvanic spectroscopy, the current passing through the gas discharge (which is the material 612) is monitored by the detector 614 as the light beam 103 is tuned over the frequencies of allowed transitions for excited atoms or molecules in the gas discharge. When the light beam 103 resonantly excites the atoms or molecules, for example, from a lower state to a higher excited state, the atoms or molecules are excited to a less bound state, thereby increasing the probability that the atoms or molecules will be ionized by discharge collisions and contribute to an increase in the discharge current. Small changes in the discharge current can be detected with high sensitivity by the detector 614. In contrast to other spectroscopic methods that can be used to detect features associated with the interaction between the light beam 103 and the material 612, optogalvanic spectroscopy does not require an optical detector (e.g., a photomultiplier or photodiode detector) to obtain an atomic transition spectrum because the gas discharge acts as a resonant optical detector.
[0058]
[0073] Referring again to FIG. 5, the calibration apparatus 510 calibrates the optical device 510 using the calibration material 512. Specifically, to calibrate the optical device 105, the light source 102 scans the wavelength of the light beam 101 (and thus the diagnostic light beam 103 formed from the light beam 101) across the energy transitions of the calibration material 512, and the detector 514 detects the transition profile 515. As the wavelength of the light beam 101 is scanned, the optical device 105 (which may be the spectral analysis module 305 of FIG. 3) measures the wavelength of the diagnostic light beam 103. As an example, the spectral analysis module 305 measures the central wavelength 309 of the spectral profile 308 of the light beam 103, as shown in FIG. 3. In addition, as the wavelength of the light beam 101 is scanned by the light source 102, the spectral analysis module 305 outputs an estimated wavelength λ 305 of the diagnostic light beam 103 (based on the central wavelength 309) to the calibration control module 516. As discussed above, the spectral analysis module 305 estimates the wavelength λ 305 of the light beam 101 based on measurements of the spatial components 306 sensed by the sensor 344 and a calibrated model of the optical path length of the optical isolation device 342. Additionally, as the wavelength of the light beam 101 is scanned by the light source 102, the detector 514 outputs measured characteristics of the diagnostic light beam 103 to the calibration control module 516. Specifically, the detector 514 outputs a transition profile 515 to the calibration control module 516.
[0059]
[0074] The calibration control module 516 determines the moment when the transition profile 515 is at a peak value (or known energy transition) λref. The calibration control module 516 compares the wavelength λ305p of the light beam 103 output from the spectral analysis module 305 with the moment when the calibration control module 516 determines that the transition profile 515 is at the peak value λref. In other words, the wavelength λ305p is the wavelength calculated by and output from the control module 350 (FIG. 3) in the spectral analysis module 305 at the moment when the calibration control module 516 determines that the transition profile 515 is at the peak value λref. The calibration control module 516 adjusts one or more operating parameters of the spectral analysis module 305 until it determines that the wavelength λ305p is equal to the peak value λref. When these two values (λ305p and λref) are equal, the spectral analysis module 305 is calibrated.
[0060]
[0075] In general, if these two values (i.e., λref and λ305p) are equal to each other or within an acceptable range of each other, the spectral analysis module 305 outputs an accurate measurement of the wavelength λ305 of the diagnostic light beam 103. However, if these two values (λref and λ305p) are not equal and are not within an acceptable range of each other, it means that the accuracy of the spectral analysis module 305 is not acceptable. Before the spectral analysis module 305 enters a state where the two values (λref and λ305p) are not within their respective acceptable ranges, a calibration must be performed. On the other hand, if the two values λref and λ305p are equal or very close, a calibration is not necessary.
[0061]
[0076] The acceptable range can quantify the absolute wavelength error and be defined by the absolute difference Δλ between λref and λ305p. Specifically, the acceptable range can be an upper or maximum value of the absolute wavelength error Δλ that would be tolerated during operation. In other words, for example, if the absolute difference Δλ between the two values, i.e., λref and λ305p, is less than 0.2 pm, less than 0.3 pm, or less than 0.4 pm, the two values, i.e., λref and λ305p, are within their respective acceptable ranges. Thus, even if the absolute wavelength difference Δλ was substantially less than 0.2 pm until then, the calibration device 510 may have performed a calibration. For example, even in a situation where the absolute wavelength difference Δλ was less than 0.007 pm until then, the calibration device 510 may have performed an unnecessary calibration. However, in such a situation, the calibration device 510 would not need to perform a calibration. Thus, in the existing situation, there are many unnecessary calibrations, which increases the downtime of the system and reduces the availability of the system. On the other hand, if the absolute difference Δλ between these two values λref and λ305p is close to but does not exceed the limit (0.2 pm, 0.3 pm or 0.4 pm), then the calibration device 510 must have performed a calibration on the spectral analysis module 305.
[0062]
[0077] The calibration of the spectral analysis module 305 performed by the calibration device 510 may take several minutes to perform and complete. For example, the calibration may take 1-5 minutes to perform and complete. During the calibration of the spectral analysis module 305, the output device 104 is not receiving the light beam 101 within the required specifications, and therefore the output device 104 is not operating. This downtime in the operation of the output device 104 due to the calibration reduces the output and production of the output device 104. For this reason, as discussed next, the calibration system 100 and the predictive controller 130 implement procedures to schedule the calibration of the optical device 105 and the spectral analysis module 305 when calibration is needed, rather than when calibration is not needed. The predictive controller 130 looks at the data from the last few (e.g., 3, 4, 5 or more) calibrations and predicts when the two values (λref and λ305p) will fall outside their respective acceptable ranges in the future, and therefore predicts when the next calibration should be performed to ensure that the spectral analysis module 305 operates with acceptable accuracy and avoids situations where the absolute difference Δλ exceeds the acceptable range. Referring to FIG. 7, a procedure 770 is performed to schedule a calibration for the optical device 104. The procedure 770 is performed by the calibration system 100 (FIGS. 1 and 2) or the calibration system 500 (FIG. 5).
[0063]
[0078] The procedure 770 includes receiving 772 a measured characteristic associated with the optical device 105 while the optical device 105 is being calibrated. The measured characteristic associated with the optical device 105 can be received by the input module 232 of the predictive controller 130 (FIG. 2). In some implementations, the calibration control module 516 provides the measured characteristic to the input module 232 of the predictive controller 130. In one example, the calibration control module 516 provides an error measurement or wavelength error, i.e., an absolute difference Δλ, which is the difference between the measured wavelength λ 305p of the diagnostic light beam 103 and a reference central wavelength λ ref associated with a known energy transition of the material 512 in the calibration apparatus 510. The error measurement is given by Equation 1: Δλ=λ305p-λref Equation (1) λref is a reference wavelength of a known energy transition of the calibration material 512, and λ305p is an estimate of the wavelength of the diagnostic light beam 103 output from the spectral analysis module 305 when the light source 102 outputs the light beam 101 having the reference wavelength λref. The wavelength error Δλ quantifies how much the accuracy of the spectral analysis module 305 drifts when one or more parameters in the spectral analysis module 305 change due to use. While the spectral analysis module 305 is being calibrated, the calibration control module 516 can calculate the wavelength error Δλ, and the input module 232 of the predictive controller 130 receives the wavelength error Δλ from the control module 516.
[0064]
[0079] Next, referring back to FIG. 7, a current degradation metric α is calculated (774) based on at least the measured characteristic. For example, the adaptive module 234 of the predictive controller 130 receives the measured characteristic from the input module 232, and the adaptive module 234 calculates a current degradation metric α based on at least the measured characteristic (in this example, the wavelength error Δλ). The degradation metric α models the behavior of the optical device 105. Specifically, the adaptive module 234 can estimate how much the measured characteristic changes with respect to changes in the usage of the optical device 105. In other words, as the optical device 105 is used over time during operation of the light source 102, the measured characteristic may change in value. In one example, referring to FIG. 3 and FIG. 4, the optical path length of the light beam 103 in the spectral analysis module 305 may change due to changes in any one or more of the refractive index and the geometric path length that occur during operation of the calibration system 500, which may cause errors in the measurement of the wavelength λ 305 by the spectral analysis module 305. This measurement error, manifested in the wavelength error Δλ, causes a change in the value of the measured property, and this measurement error can be modeled by the adaptation module 234 as a degradation metric.
[0065]
[0080] The degradation metric α may be based on a cumulative sum Σ of the current wavelength error Δλ plus one or more wavelength errors Δλp output from the control module 516. The current degradation metric α may correspond to a local linear approximation of the slope of the trend of the measured characteristic. For example, the trend may be a linear trend or a non-linear trend. The local linear approximation of the slope may be positive or negative if the slope is positive or negative, respectively. In this way, the measured characteristic is assumed to change linearly with respect to changes in the usage of the optical device 105. In particular, the cumulative sum Σ of the current wavelength error Δλ and one or more wavelength errors Δλp output from the control module 516 provides information about the trend of the wavelength error Δλ.
[0066]
[0081] The predictive controller 130 then estimates (776) when the degradation of the optical device 105 will exceed a predefined threshold MaxErr based on the current degradation metric α(i), where i corresponds to the most recent (and current) calibration event performed by the spectrum analysis module 305. Thus, i-1 is the calibration event performed by the spectrum analysis module 305 that occurred immediately before calibration event i, and i+1 is the next calibration event to be scheduled by the predictive controller 130. The adaptation module 234 of the predictive controller 130 performs an analysis based on the calculated current degradation metric α (from 774) to estimate when the accuracy of the optical device 105 will become unacceptable. The predictive controller 130 schedules (778) a calibration of the optical device 105 based at least in part on the estimate of the degradation of the optical device 105. For example, the adaptation module 234 can schedule a calibration based on the degradation estimate (from 776) and provide the scheduled event to the output module 236. The adaptation module 234 estimates, UsePredict(i+1), how many pulses of the light beam 101 can be generated until the degradation of the optical device 105 exceeds a threshold MaxErr. This estimate can be expressed as:
number
[0067]
[0082] As mentioned, in step 774, the adaptation module calculates a degradation metric α that models the behavior of the optical device 105. This estimation relies on the assumption that the current degradation metric α corresponds to a local linear approximation of the slope of the trend of the measured characteristic. If the local behavior of the trend of the measured characteristic is not linear, the calculated degradation metric α may not adequately capture the local behavior of the trend of the measured characteristic. In this case, the adaptation module 234 can adjust how it schedules the calibration of the optical device 105 so that it takes into account the reliability of the model of the behavior of the optical device 105. Details regarding the steps of procedure 770 are discussed next with reference to an example.
[0068]
[0083] 8A, graphs 856_1, 856_2 show the behavior of the optical path length PL of a corresponding distinct spectral analysis module 805_1, 805_2 associated with a corresponding light source 802_1, 802_2 versus the usage 855u of that spectral analysis module 805_1, 805_2. Each data point of the graphs 856_1, 856_2 corresponds to a calibration event where the spectral analysis module 805_1 or 805_2 is calibrated by the calibration device 110. The most recent and current calibration events are labeled i, the most recent previous calibration event is i-1, and the future or planned next calibration event is i+1. The usage 855u of the spectral analysis modules 805_1, 805_2 may correspond to the number of pulses of the light beams 801_1, 801_2 generated by the light sources 802_1, 802_2 and also the number of pulses of the diagnostic light beams 803_1, 803_2 analyzed by the corresponding spectral analysis modules 805_1, 805_2. As is evident from these graphs 856_1, 856_2, the optical path length PL of the spectral analysis modules 805_1, 805_2 drifts with usage. Moreover, the optical path length PL tends to drift continuously and smoothly. In these examples, this drift occurs slowly over many pulses of the light beams 801_1, 801_2. In particular, the scale of usage Δu shown in the graphs 856_1, 856_2 is on the order of millions or billions of pulses of the light beams 801_1, 801_2.
[0069]
[0084] The behavior of the optical path length PL of the spectral analysis module 805_1 tends to decrease and level off with the usage amount 855u, and the behavior lacks an inflection point during the usage period shown in the graph 856_1. This means that the optical path length PL of the spectral analysis module 805_1 tends to change in one direction. In contrast, the behavior of the optical path length PL of the spectral analysis module 805_2 initially decreases during the usage period shown in the graph 856_2, then passes through an inflection point PL_IP, and then increases. This means that the optical path length PL of the spectral analysis module 805_2 changes in two directions at different moments of the usage amount 855u.
[0070]
[0085] The nonlinearity in the form of graphs 856_1 and 856_2 may be caused by changes in the properties of the respective spectral analysis modules 805_1, 805_2. For example, the spectral analysis module 405 may undergo oxidation of the organic material, which may change the refractive index within the module 405, which may affect the optical path length within the etalon. For the spectral analysis module 805_2, for which graph 856_2 has an inflection point PL_IP, the downward PL trajectory may reflect a "getter phase", during which oxygen oxidizes the organic material in the module 805_2, decreasing the gas index. Also, as oxygen is consumed due to continuous outgassing of methanol, organic material begins to deposit on the etalon cavity, which reverses the PL trajectory, and the trajectory begins to rise.
[0071]
[0086] Referring to FIG. 8B, graphs 850_1, 850_2 show the behavior of the corresponding cumulative sum Σ855p of the measured characteristics for the corresponding spectral analysis modules 805_1, 805_2 against the usage 855u over the same usage 855u for which graphs 856_1, 856_2 are generated. As mentioned above, each data point of graphs 850_1, 850_2 corresponds to a calibration event where the spectral analysis module 805_1 or 805_2 is calibrated by the calibration device 110. The most recent and current calibration events are labeled i, the most recent previous calibration event is i-1, and the future or planned next calibration event is i+1. Graphs 850_1, 850_2 drift continuously and smoothly, as do graphs 856_1, 856_2. Moreover, the drift occurs slowly over many pulses of the light beams 801_1, 801_2. The trend of graph 850_1 is similar in shape to the trend of graph 856_1, but is inverted, and the trend of graph 850_2 is similar in shape to the trend of graph 856_2, but is inverted.
[0072]
[0087] Since the behavior shown in the graphs 850_1 / 850_2 is related to the behavior shown in the corresponding graphs 856_1 / 856_2, the shape and trend of the cumulative sum of the measured properties Σ855p can be used to understand the drift in the optical path length PL. Also, since the drift of the optical path length PL is directly related to the degradation of the spectral analysis modules 805_1, 805_2, it is possible to predict the degradation of the spectral analysis modules 805_1, 805_2 based on the cumulative sum of the measured properties Σ855p. This is useful because the cumulative sum of the measured properties Σ855p does not require unit conversion to develop a model of the degradation of the spectral analysis module 305 (e.g., in step 774 of procedure 770), and further provides a more intuitive understanding of the performance of the spectral analysis module 305. In particular, the difference between each of two data points in the cumulative sum of the measured properties Σ855p indicates how much wavelength error would be experienced if no calibration was performed between those two points.
[0073]
[0088] Similar to the behavior of the optical path length PL of the spectral analysis module 805_1, the cumulative sum Σ855p of the measured properties of the spectral analysis module 805_1 tends to rise and then plateau with usage 855u, lacking an inflection point in the period of usage shown in graph 850_1. This means that the cumulative sum Σ855p of the measured properties of the spectral analysis module 805_1 tends to change in one direction during usage. The behavior of the cumulative sum Σ855p of the measured properties of the spectral analysis module 805_1 can always be approximated linearly during usage when observed in a small usage window, for example from one calibration event to an adjacent calibration event.
[0074]
[0089] In contrast, the cumulative sum Σ855p of the measured characteristics of the spectrum analysis module 805_2 rises initially, then passes through the inflection point Σ_IP, and then falls. This means that the cumulative sum Σ855p of the measured characteristics of the spectrum analysis module 805_2 changes in two directions at different moments of usage 855u. The behavior of the cumulative sum Σ855p of the measured characteristics of the spectrum analysis module 805_2 can be approximated linearly when observed in a specific small usage window that is not close to the inflection point Σ_IP, where the small usage window includes from one calibration event to the adjacent calibration event. The cumulative sum Σ855p of the measured characteristics of the spectrum analysis module 805_2 may need to be approximated nonlinearly when observed in a specific small usage window that is close to or overlaps with the inflection point Σ_IP, where the small usage window includes from one calibration event to the adjacent calibration event.
[0075]
[0090] 9, to illustrate the operation of procedure 770, values of the cumulative sum Σ of the spectral analysis module 305 are shown for usage 955u in a local window 950W of usage. In this figure, the current and most recent calibration events are i and occur at usage Ui. Furthermore, i-1 is the calibration event performed by the spectral analysis module 305 that occurred immediately before calibration event i, and i+1 is the next calibration event to be scheduled by the predictive controller 130. The current degradation metric α(i) provides an estimate of the slope of the cumulative sum Σ in a small usage window bounded by the current and most recent calibration events occurring at usage Ui. The current degradation metric α(i) forms the hypotenuse of a right triangle, where it is assumed that the cumulative sum Σ changes linearly in this small usage window from value (i). Expressed in graphical form, the usage UsePredict(i+1) is estimated from the value of the cumulative sum Σ955p at the last calibration event i by forming a right triangle with the hypotenuse extended along the calculated slope α(i) by the usage corresponding to a predefined threshold MaxErr.
[0076]
[0091] Next, an adaptive method is described for calculating a current degradation metric α(i) based on the measured characteristic (step 774). In the following discussion, the measured characteristic is the previous wavelength error Δλ(i-1), which corresponds to the difference between the measured wavelength λ 305p(i-1) of the diagnostic light beam 103 (output from the spectrum analysis module 305 at the previous calibration event i-1) and the reference central wavelength λ ref associated with a known energy transition of the material 512 in the calibration device 510. The final value of the cumulative sum Σ is given by Σ(i-1), which corresponds to the summation up to this value of the previous wavelength error, including the wavelength error Δλ(i-1).
[0077]
[0092] At the initialization of the adaptive method, such as when the spectrum analysis module 305 is first started and the first calibration corresponds to i=1, the following values are set: α(1)=0; EstErr(1)=0; Σ(1)=Δλ(1); EstΣ(1)=Σ(1); and MaxErr=0.1 pm. MaxErr can be set, for example, to be half the desired requirement from the customer, thus allowing a safety margin of half the desired requirement value. EstErr is a measure of the reliability of the adaptive method implemented for the current calibration event i. Est is an estimate of the accumulated wavelength error for the current calibration event i. For each iteration and the current and most recent calibration events i, the estimated accumulated wavelength error Est(i) is given by Equation 3: EstΣ(i)=Σ(i-1)+α(i-1)*U(i) Equation (3) The value of the reliability estimate EstErr(i) for the current and most recent calibration event i is then given by Equation 4: EstErr(i)=Σ(i)-EstΣ(i) Equation (4) where Σ(i) corresponds to the summation of the previous wavelength error up to this value, including the current wavelength error Δλ(i). The reliability estimate EstErr(i) provides an indication of how reliable the local linear approximation is, and thus may provide an indication of whether the local trend of the cumulative sum Σ can be approximated linearly, or whether the linear approximation of the cumulative sum Σ is inappropriate (such as the inflection point Σ_IP (FIG. 8B)). The current degradation metric α(i) is calculated (774) based on this reliability estimate as follows: α(i)=α(i-1)+ηEstErr(i) Equation (5) where η is a constant that acts as the gain of the low pass filter. The constant η can be chosen to be 0<η<1 and can be a configurable parameter.
[0078]
[0093] As discussed above, the adaptation module 234 schedules the calibration based on this degradation metric α(i) (778), for example, by first estimating how many pulses of the light beam 101 can be generated. UsePredict1(i+1) (776) from the above equation is then regenerated as Equation 6.
number
[0079]
[0094] In some implementations, the next calibration is scheduled 778 based on the actual value of UsePredict1(i+1), i.e., the next calibration is scheduled to occur at the number of pulses determined in UsePredict1(i+1).
[0080]
[0095] In other implementation forms, the following calibration schedule setting (778) takes into account not only the value of UsePredict1(i + 1) but also the reliability estimation value EstErr(i). For example, the second value of UsePredict2(i + 1) can be determined by comparing the reliability estimation value EstErr(i) with one or more thresholds Thr (e.g., Thr1, Thr2, Thr3, Thr4) given in units of distance (wavelength). The values for the threshold Thr are configurable and can be modified. In one example, Thr1 = 0.4 pm, Thr2 = 0.2 pm, Thr3 = 0.01, and Thr4 = 0.005. When the absolute value of EstErr(i) > Thr1, UsePredict2(i + 1) = UseValue1, which can be 300 million pulses of the optical beam 101. When Thr2 < the absolute value of EstErr(i) < Thr1, UsePredict2(i + 1) = UseValue2, which can be 600 million pulses of the optical beam 101. When Thr3 < the absolute value of EstErr(i) < Thr2, UsePredict2(i + 1) = UseValue3, which can be 1.2 billion pulses of the optical beam 101. Generally, when Thr(n - 1) < the absolute value of EstErr(i) < Thr(n) and n is 1, 2, 3, 4 or greater, UsePredict2(i + 1) = UseValue(n), which can be 300 × 2 n and can be. The next calibration is scheduled by determining the minimum value between UsePredict1(i + 1) and UsePredict2(i + 1).
[0081]
[0096] In this manner, the output module 236 determines when the next or subsequent calibration of the spectral analysis module 305 should be performed. The output module 236 can instruct the calibration device 510 to calibrate the spectral analysis module 305 based on this schedule. In this manner, the predictive controller 130 determines when the next calibration of the spectral analysis module 305 should be advantageously scheduled, and communicates with the calibration device 510 to calibrate the spectral analysis module 305 if calibration is required. Furthermore, if the spectral analysis module 305 does not need to be calibrated, the spectral analysis module 305 is not calibrated, thereby reducing the time that the spectral analysis module 305 is out of service (being calibrated).
[0082]
[0097] 10, after scheduling a calibration of the optical device 105 (which may be the spectral analysis module 305) in step 778 of procedure 770, procedure 1071 is performed. After the next calibration of the spectral analysis module 305 is scheduled (778), the output module 236 of the predictive controller 130 sends instructions to the calibration device 510 regarding when the next calibration should be performed (1073). Furthermore, the current degradation metric α(i) is assigned to the previous degradation metric α(i-1), and other variables in the adaptive method discussed above may be updated similarly (1077). The current degradation metric α(i) associated with this next calibration may then be calculated as the updated metric, for example, using procedure 770.
[0083]
[0098] 11, photolithography system 1180 includes an illumination system 1102 that generates a light beam 1101, nominally having a wavelength at a central wavelength, that is directed to a photolithography exposure tool or scanner 1104 by a beam preparation system 1183, which may include beam introduction and beam modifying optics. Illumination system 1102 is an implementation of light source 102, and exposure tool 1104 is an implementation of optical device 104 (FIG. 1). Photolithography system 1180 also includes optical device 105 and calibration system 100 that is positioned relative to optical device 105. Optical device 105 receives diagnostic light beam 103. The diagnostic light beam is a portion of optical beam 101 that is split from optical beam 101 by beam splitter 1182.
[0084]
[0099] A light beam 1101 is generated by a light source 1184 and directed to a wafer 1186 in an exposure tool 1104, whereby microelectronic features are patterned on the wafer 1186. The light beam 1101 has a wavelength in the deep ultraviolet (DUV) range, for example about 248 nanometers (nm) or about 193 nm. The size of the microelectronic features patterned on the wafer 1186 depends on the wavelength of the light beam 1101, with shorter wavelengths resulting in smaller minimum sizes. When the wavelength of the light beam 1101 is 248 nm or 193 nm, the minimum size of the microelectronic features can be, for example, 50 nm or less.
[0085]
[0100] Various disturbances (e.g., temperature gradients, pressure gradients, optical distortions) act on the light source 1184 and the light beam 1101 to alter the spectral properties or characteristics (e.g., wavelength) of the light beam 1101. Thus, the lithography system 1180 includes other components, such as, for example, a spectral feature selection system 1181, an optical device 105 configured to measure aspects (e.g., wavelength) related to the light beam 1101, and a control system 1188, which are used in combination to determine the effect of the disturbances on the light beam 1101 and to compensate for the effect of the disturbances on the light beam 1101.
[0086]
[0101] In some implementations, the light source 1184 is an excimer light source that outputs a pulsed light beam 1101 (which may be a laser beam). Once the pulsed light beam 1101 enters the photolithography exposure apparatus 1104, it is directed through a projection optical system 1185 and projected onto a wafer 1186 to form one or more microelectronic features on a photoresist on the wafer 1186. The photolithography exposure apparatus 1104 also includes a controller 1187 that can communicate with a control system 1188 of the illumination system 1102.
[0087]
[0102] In the illustrated implementation, the light source 1184 is a two-stage system that includes a master oscillator (MO) 1184MO that provides a seed light beam to a power amplifier (PA) 1184PA. The MO 1184MO and the PA 1184PA can be considered subsystems of the light source 1184 or systems that are part of the light source 1184. The power amplifier 1184PA receives the seed light beam from the master oscillator 1184MO and amplifies the seed light beam to generate the light beam 1101 for use in the photolithography exposure tool 1104. For example, the master oscillator 1184MO can emit a pulsed seed light beam with a seed pulse energy of about 1 millijoule (mJ) per pulse, and these seed pulses can be amplified by the power amplifier 1184PA to about 10-15 mJ.
[0088]
[0103] The master oscillator 1184MO includes a MO discharge chamber having two elongated electrodes (forming an excitation mechanism), a gain medium that is a gas mixture, and a fan for circulating the gas between the electrodes. A resonator is formed between the spectral feature selection system 1181 on one side of the master oscillator 1184MO and an output coupler on the second side of the discharge chamber master oscillator 1184MO. The power amplifier 1184PA includes a beam combining optical system that receives a seed light beam from the master oscillator 1184MO and directs the seed light beam through the PA discharge chamber to a beam rotation optical element that redirects or changes the seed light beam so that it is sent back into the PA discharge chamber. The PA discharge chamber includes a pair of elongated electrodes, a gain medium that is a gas mixture, and a fan for circulating the gas mixture between the electrodes.
[0089]
[0104] The gas mixture of the gain medium used in each discharge chamber can be any gas suitable for generating a light beam at the wavelength and bandwidth required for application in the output device (photolithography exposure apparatus 1104). In the case of an excimer source, the gas mixture can include, for example, a rare gas (noble gas) such as argon or krypton, a halogen such as fluorine or chlorine, and a trace amount of xenon other than helium and / or neon as a buffer gas. Specific examples of gas mixtures include argon fluoride (ArF), which emits light at a wavelength of about 193 nm, krypton fluoride (KrF), which emits light at a wavelength of about 248 nm, or xenon chloride (XeCl), which emits light at a wavelength of 351 nm. The excimer gain medium (gas mixture) is pumped with short (e.g., nanosecond) current pulses in a high-voltage discharge by applying a voltage to the corresponding elongated electrodes.
[0090]
[0105] The light beam 1101 is a pulsed light beam and may include one or more bursts of pulses separated from one another in time. Each burst may include one or more light pulses. In some implementations, a burst may include hundreds of pulses, e.g., 100-400 pulses.
[0091]
[0106] As discussed above, when the gain medium in the MO and PA is pumped by applying a voltage to the electrodes, the gain medium of the MO or PA emits light. When a voltage is applied to the electrodes in a pulsed manner, the light emitted from the gain medium is also pulsed. Thus, the repetition rate of the pulsed light beam 1101 is determined by the frequency at which a voltage is applied to the electrodes, and each application of the voltage produces a light pulse. The light pulses propagate through the gain medium and exit the chamber through the output coupler. Thus, a train of pulses is formed by periodically repeating the application of a voltage to the electrodes. The repetition rate of the pulses can be in the range of about 500 Hz to 6,000 Hz. In some implementations, the repetition rate can be greater than 6,000 Hz, for example, 12,000 Hz or greater. The pulsed light beam 1101 can have an average output power in the range of tens of watts, for example, from about 50 W to about 130 W. The irradiance (i.e., average power per unit area) of the light beam 1101 at the output can be greater than 60 W / cm. 2 ~80W / cm 2 The range may be:
[0092]
[0107] 12, in some implementations, the optical device 105 is an energy analysis module 1205 and the calibration system 100 is configured as a calibration system 1200 for interfacing with the energy analysis module 1205. The energy analysis module 1205 is configured to receive the diagnostic light beam 103 from the light source 102, the diagnostic light beam 103 being split off from the main light beam 101. The energy analysis module 1205 includes a light detector 1205A and a conversion module 1205B in communication with the light detector 1205A. An output from the conversion module 1205B is provided to an energy control component 1206 associated with or within the light source 102 to control the energy of the light beam 101.
[0093]
[0108] Photodetector 1205A may include a photodiode that outputs a current or voltage value, which is converted to a digital value M. Because diagnostic light beam 103 is a split-off portion of light beam 101, digital value M is proportional to the energy of light beam 101. Digital value M is output to conversion module 1205B. Conversion module 1205B converts this digital value M to an energy value E, which is an estimate of the energy of the primary light beam 101, and is used by energy control component 1206 to control the energy of light beam 101.
[0094]
[0109] The calibration system 1200 includes an implementation 1210 of the calibration apparatus 110 and an implementation 1230 of the predictive controller 130. The predictive controller 1230 can be designed like the predictive controller 130 of FIG. 2. The calibration apparatus 1210 includes a measurement device 1214, such as an optical power meter, and a calibration control module 1216 in communication with the optical power meter 1214. The calibration apparatus 1210 communicates with the predictive controller 1230 and further with the optical device 1205. Although the calibration control module 1216 is shown as being co-located with the power meter 1214, it is also possible for the calibration control module 1216 to instead be co-located with or integrated within the optical device 1205.
[0095]
[0110] The power meter 1214 is a device or instrument arranged to directly measure the power (energy delivered per unit time) of the primary optical beam 101. In some implementations, the power meter 1214 includes a calibrated sensor that measures the average power of a series of pulses of the primary optical beam 101 (when the pulse repetition rate is known). The power meter 1214 includes a sensor head including a power sensor that moves into and out of the path of the primary optical beam 101. The power sensor may be a photodiode or other suitable photodetector with a wavelength range that encompasses the wavelength of the primary optical beam 101. Additionally, the power meter 1214 may also include additional electronics to convert an analog signal into a digital signal that estimates or measures the power Pm of the primary optical beam 101. The measured power Pm is provided to a calibration control module 1216. The calibration control module 1216 communicates with the power meter 1214 and the optical device 1205 to perform a calibration of the optical device 1205 as described in more detail below.
[0096]
[0111] As discussed above, the conversion module 1205B converts the digital value M (based on the direct measurement of the diagnostic light beam 103 by the photodetector 1205A) into an energy value E, which is an indirect estimate of the energy of the main light beam 101. This conversion can be performed according to the following equation: E=MS+O formula (7) where S is the measurement scale and O is the offset. The measurement scale S is a scale that serves several purposes. First, the measurement scale S is used to convert the digital values M output from the photodetector 1205A to energy values. Second, the measurement scale S is used to extend the scale of energy values from values associated with the diagnostic light beam 103 to estimated values associated with the main light beam 101. As discussed above, the measurement scale S is necessary because the energy of the diagnostic light beam 103 is lower than the energy of the main light beam 101 since the diagnostic light beam 103 is a portion or fraction of the optical energy of the main light beam 101.
[0097]
[0112] The measurement scale S may drift over time. Accordingly, the optical device 1205 needs to be periodically calibrated to address or compensate for this drift in the sensor scale S. During calibration, the power meter 1214 is placed in the path of the primary optical beam 101, thereby measuring the power Pm of the primary optical beam 101. This measured power Pm is provided to the calibration control module 1216, which converts this value to an energy value Em by dividing it by the repetition rate of the primary optical beam 101. The calibration control module 1216 compares the measured energy value Em from the power meter 1214 with the estimated energy value E from the optical device 1205 to determine how to adjust or modify the measurement scale S so that the two values are equal. The calibration control module 1216 may further perform a final verification step after the measurement scale S has been adjusted to verify that the adjustment is correct.
[0098]
[0113] The calibration of the light device 1205 performed by the calibration apparatus 1200 can take several minutes to perform and complete. Furthermore, while the light device 1205 is being calibrated, the output device 104 is not receiving the light beam 101 within the required specifications, and therefore the output device 104 is not operational. This downtime in the operation of the output device 104 due to calibration reduces the output and production of the output device 104. For this reason, as discussed next, the calibration system 1200 and predictive controller 1230 implement procedures to schedule the calibration of the light device 1205 when calibration is needed, rather than when calibration is not needed. The predictive controller 1230 looks at data from the last few calibrations (e.g., 3, 4, 5 or more calibrations) to predict when the measurement scale S will go out of the acceptable drift range in the future and therefore predicts when the next calibration of the optical device 1205 should be performed to ensure that the optical device 1205 operates with acceptable accuracy and avoids a situation where the drift of the measurement scale S exceeds the acceptable drift range. The maximum acceptable drift in the measurement scale S is given by MaxErr(S). The maximum acceptable drift MaxErr(S) in the measurement scale S can be calculated from the maximum acceptable error in the energy measurement MaxErr(E), which is a measurement of the energy of the primary optical beam 101, as follows:
number
[0099]
[0114] 13, procedure 1370 is performed to schedule a calibration of optical device 1205. Procedure 1370 is similar to procedure 770 in that procedure 770 is one way of performing procedure 1370. Procedure 1370 is performed by calibration system 1200. Although not discussed, procedure 1370 can be performed by calibration system 100 (FIGS. 1 and 2) or calibration system 500 (FIG. 5).
[0100]
[0115] The procedure 1370 includes receiving 1372 characteristics associated with the light device 1205 while the light device 1205 is being calibrated. The light device characteristics can be received by the input module 232 of the predictive controller 1230 (FIGS. 2 and 12). In some implementations, the calibration control module 1216 provides the light device characteristics to the input module 232 of the predictive controller 1230. In other implementations, the predictive controller 1230 accesses the light device characteristics directly from the light device 1205. In one example, when the predictive controller 1230 detects that a calibration has been performed, the predictive controller 1230 requests data associated with the calibration from the light device 1205 [data(cal)]. The light device 1205 provides the data(cal) to the predictive controller 1230. The data(cal) includes the current value of the measurement scale S(i), the value of the measurement offset O, and the number of pulses of the light beam 101 generated since the last calibration of the light device 1205 [UsePredict(i)].
[0101]
[0116] A current degradation metric α(i) is then calculated (1374) based on at least the optical device characteristics. For example, the adaptive module 234 of the predictive controller 1230 receives the optical device characteristics (including the measurement scale S(i)) from the input module 232, and the adaptive module 234 calculates a current degradation metric α(i) based on at least the optical device characteristics. The degradation metric α models the behavior of the optical device 1205. Specifically, the adaptive module 234 can estimate how much the optical device characteristics (the measurement scale S) change with respect to changes in the usage of the optical device 1205. In other words, as the optical device 1205 is used over time during operation of the light source 102, the optical device characteristic S can vary (or drift) in value. This drift in S causes a change in the value of the estimated energy E, and this drift can be modeled by the adaptive module 234 as a degradation metric.
[0102]
[0117] The degradation metric α may correspond to a local linear approximation of the slope of the trend of the measurement scale S. For example, the trend may be a linear trend or a non-linear trend. The local linear approximation of the slope may be positive or negative, respectively, when the slope is positive or negative. In this way, the optical device characteristics are assumed to change linearly with respect to changes in the usage of the optical device 1205.
[0103]
[0118] The predictive controller 1230 then estimates (1376) when the degradation of the optical device 1205 will exceed a predefined threshold MaxErr(S) based on the current degradation metric α(i), where i corresponds to the most recent (and current) calibration event performed on the optical device 1205. Thus, i-1 is the calibration event performed on the optical device 1205 that occurred immediately before calibration event i, and i+1 is the next calibration event to be scheduled by the predictive controller 1230. The adaptive module 234 of the predictive controller 1230 performs an analysis based on the calculated current degradation metric α (from 1374) to estimate when the accuracy of the optical device 1205 will become unacceptable. The predictive controller 1230 schedules (1378) a calibration of the optical device 1205 based at least in part on the estimate of the degradation of the optical device 1205. For example, the adaptation module 234 may schedule a calibration based on the degradation estimate (from 1376) and provide this scheduled event to the output module 236. The adaptation module 234 estimates, UsePredict(i+1), how many pulses of the light beam 101 can be generated until the degradation of the optical device 1205 exceeds a threshold MaxErr(S). This estimate may be expressed as the following equation:
number
[0104]
[0119] As mentioned, in step 1374, the adaptation module calculates a degradation metric α that models the behavior of the light device 1205. This estimation relies on the assumption that the current degradation metric α corresponds to a local linear approximation of the slope of the trend of the measurement scale S. If the local behavior of the trend of the measurement scale S is not linear, the calculated degradation metric α may not adequately capture the local behavior of the trend of the measurement scale S. In this case, the adaptation module 234 can adjust how it schedules calibrations of the light device 1205 so that it takes into account the reliability of the model of the behavior of the light device 1205.
[0105]
[0120] 14, to illustrate the operation of procedure 1370, values of the measurement scale S of the light device 1205 are shown versus usage 1455u in a local window 1450W of usage. In this figure, the current and most recent calibration events are i and occur at usage Ui. Furthermore, i-1 is the calibration event performed on the light device 1205 that occurred immediately before calibration event i, and i+1 is the next calibration event to be scheduled by the predictive controller 1230. The current degradation metric α(i) provides an estimate of the slope of the measurement scale S in a small usage window bounded by the current and most recent calibration events occurring at usage Ui. The current degradation metric α(i) forms the hypotenuse of a right triangle, where we assume that the measurement scale S changes linearly in this small usage window from value (i). Expressed in graphical form, the usage UsePredict(i+1) is estimated from the value 1455p of the measurement scale S at the last calibration event i by forming a right triangle with the hypotenuse extended along the calculated slope α(i) by the usage corresponding to a predefined threshold MaxErr(S).
[0106]
[0121] Next, an adaptive method (step 1374) is described for calculating a current degradation metric α(i) based on the optical device characteristic. In the following discussion, the optical device characteristic is a measurement scale S. The final value of the measurement scale S is given by S(i-1).
[0107]
[0122] At the initialization of the adaptive method, such as when the Spectral Analysis Module 305 is first started and the first calibration corresponds to i=1, the value of the degradation metric is set as follows:
number
[0108]
[0123] As discussed above, the adaptation module 234 schedules 1378 the calibration based on this degradation metric α(i), for example, by first estimating how many pulses of the light beam 101 can be generated. UsePredict1(i+1) (1376) from the above equation is then regenerated as Equation 14.
number
[0109]
[0124] In some implementations, the next calibration is scheduled 1378 based on the actual value of UsePredict1(i+1), i.e., the next calibration is scheduled to occur at the number of pulses determined in UsePredict1(i+1).
[0110]
[0125] In other implementations, scheduling the next calibration (1378) takes into account not only the value of UsePredict1(i+1) but also the reliability estimate EstErr(i). For example, the second value of UsePredict2(i+1) can be determined by comparing the reliability estimate EstErr(i) with one or more thresholds Thr (e.g., Thr1, Thr2, Thr3, Thr4) given in units of absolute energy error (in percent). The values for the thresholds Thr are configurable and can be modified. In one example, Thr1=2.1%, Thr2=1.4%, Thr3=0.93% and Thr4=0.6%.
[0111]
[0126] The second value of UsePredict2(i+1) is a decreasing function based on the value of EstErr(i). In a specific implementation, when the absolute value of EstErr(i) > Thr1, UsePredict2(i+1) = UseValue1, which can be 500 million pulses of the optical beam 101. When Thr2 < the absolute value of EstErr(i) < Thr1, UsePredict2(i+1) = UseValue2, which can be 1 billion pulses of the optical beam 101. When Thr(3) < the absolute value of EstErr(i) < Thr(2), UsePredict3(i+1) = UseValue3, which can be 2 billion pulses of the optical beam 101. When Thr(4) < the absolute value of EstErr(i) < Thr(3), UsePredict4(i+1) = UseValue4, which can be 4 billion pulses of the optical beam 101. When Thr(5) < the absolute value of EstErr(i) < Thr(4), UsePredict5(i+1) = UseValue5, which can be 6 billion pulses of the optical beam 101. The next calibration is scheduled by determining the minimum value between UsePredict1(i+1) and UsePredict2(i+1).
[0112]
[0127] In this way, the output module 236 determines when the next or subsequent calibration of the optical device 1205 should be performed. The output module 236 can instruct the calibration device 1210 to calibrate the optical device 1205 based on this schedule. Thus, the prediction controller 1230 determines when the next calibration of the optical device 1205 should be advantageously scheduled, and if calibration is necessary, communicates with the calibration device 1210 to calibrate the optical device 1205. Further, if the optical device 1205 does not need to be calibrated, the optical device 1205 is not calibrated, thereby reducing the time when the optical device 1205 is inoperable (under calibration).
[0113]
[0128] 15, after scheduling a calibration of the optical device 1205 in step 1378 of procedure 1370, procedure 1571 is performed. After the next calibration of the optical device 1205 is scheduled (1378), the output module 236 of the predictive controller 1230 sends instructions to the calibration device 1210 regarding when the next calibration should be performed (1573). Furthermore, the current degradation metric α(i) is assigned to the previous degradation metric α(i-1), and other variables in the adaptive method discussed above may be updated similarly (1577). The current degradation metric α(i) associated with this next calibration may then be calculated as the updated metric, for example, using procedure 1370.
[0114]
[0129] In other implementations, some of the components in the calibration system 1200 may be incorporated into some of the components of the other calibration systems described above. For example, some of the components are dedicated for use with the energy analysis module 1205, and others of the components are dedicated for use with optical devices such as the spectral analysis module 305.
[0115]
[0130] In other implementations, the calibration device 1210 may include an optical energy meter instead of the power meter 1214, which measures the actual energy of the pulses of the light beam 101.
[0116]
[0131] Implementations and / or embodiments may be further described using the following clauses. 1. A method for scheduling calibration for a light device in a light source, comprising: receiving a characteristic associated with the optical device while the optical device is being calibrated; calculating a current degradation metric that models a behavior of the optical device based at least on the optical device characteristics; estimating when degradation of the optical device will exceed a threshold based on a current degradation metric; scheduling a calibration of the optical device based at least in part on the estimate of the optical device degradation; A method comprising: 2. The method of claim 1, wherein scheduling calibration of the optical device based on the estimate includes determining when calibration of the optical device should be performed. 3. The method of claim 1, wherein calculating the current degradation metric is based on a calibration characteristic. 4. The method of claim 1, wherein the optical device is configured to measure the spectral characteristics of the light beam generated by the light source. 5. The method of claim 4, wherein the measured spectral characteristic of the light beam is the wavelength of the light beam. 6. The method of clause 1, wherein receiving the optical device characteristics while the optical device is being calibrated includes receiving error measurements associated with spectral characteristics of the light beam generated by the light source while the optical device is being calibrated. 7. The method of claim 6, wherein the error measure is the difference between the measured spectral characteristic of the light beam and a reference spectral characteristic. 8. The method of clause 1, wherein calculating a current degradation metric based on at least the optical device characteristic includes estimating how much the optical device characteristic changes in response to changes in usage of the optical device. 9. The method of claim 8, wherein estimating how much the optical device characteristic will change in response to changes in usage of the optical device includes estimating an error based on data relating to previous values of the optical device characteristic and previous degradation metrics. 10. The method of clause 1, wherein calculating a current degradation metric based on at least the optical device characteristic includes assuming that the optical device characteristic changes linearly with changes in usage of the optical device. 11. Calculating a current degradation metric based on the optical device characteristic includes estimating an error in the optical device characteristic based on a linear function, where the previous degradation metric is a slope of the linear function; 2. The method of claim 1, wherein the linear function comprises a previous value of the usage of the optical device and a cumulative sum of the error in the optical device characteristic. 12. The method of claim 11, wherein calculating a current degradation metric based on the optical device characteristic includes estimating an error in the optical device characteristic. 13. The method of claim 1, wherein calculating a current degradation metric based on at least the optical device characteristic includes analyzing how the optical device characteristic is trending. 14. The method of claim 1, wherein the degradation metric models trend behavior of the optical device. 15. The method of clause 1, wherein scheduling calibration of the optical device based on the estimation includes determining whether the optical device characteristics vary nonlinearly with usage of the optical device. 16. The method of clause 15, wherein determining whether the optical device characteristics change nonlinearly with respect to the amount of use of the optical device includes analyzing a confidence level in a current degradation metric. 17. The method of claim 1, wherein estimating when degradation of the optical device will exceed a threshold value includes estimating when an error in an optical device characteristic exceeds a threshold value. 18. The method of clause 1, wherein scheduling calibration of the optical device based on the estimation includes estimating an amount of additional usage of the optical device that would be acceptable before requiring calibration. 19. The method of claim 1, further comprising, after scheduling a calibration of the optical device, updating the optical device characteristics when the next scheduled calibration is performed, and assigning a current degradation metric to a previous degradation metric. 20. The method of claim 1, wherein the degradation metric corresponds to a locally linear approximation of the slope of the optical device characteristic. 21. The method of claim 1, wherein the optical device is configured to estimate the energy of the light beam generated by the light source. 22. Receiving the optical device characteristic while the optical device is being calibrated includes performing a conversion from the direct measurement to an indirect value; 2. The method according to claim 1, wherein the direct measurement relates to a light beam generated by the light source while the optical device is being calibrated. 23. The method of claim 22, wherein estimating when degradation of the optical device will exceed a threshold comprises estimating when drift in an optical device characteristic will exceed a drift threshold. 24. The method of clause 23, wherein estimating when the drift in the optical device characteristic will exceed a drift threshold includes calculating the drift threshold based on an error threshold associated with the indirect value. 25. A calibration apparatus configured to calibrate an optical device configured to measure an aspect of a light beam generated by a light source; a predictive controller in communication with the calibration device, an input module configured to receive from the optical device a measured characteristic associated with a measured aspect of the light beam while the optical device is being calibrated; an adaptation module configured to receive the measured characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the measured characteristics, and estimate when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module configured to schedule a calibration of the optical device based on the estimates from the adaptation module and to instruct a calibration device to calibrate the optical device based on the schedule; A predictive controller including: A calibration system including: 26. A calibration system as described in clause 25, wherein the measured characteristic associated with the optical device includes a measurement of a spectral characteristic of a light beam generated by the light source. 27. A calibration system as described in clause 25, wherein the calibration device includes a calibration material having an optical transition profile with known energy transitions. 28. An optical device includes a spectral analysis module configured to sense a spectral profile of a light beam generated by a light source; 28. The calibration system of claim 27, wherein the calibration device is configured to calibrate the spectral analysis module using the calibration material. 29. The calibration system of claim 28, wherein the estimate of degradation of the optical device is an indication of drift in the calibration of the spectral analysis module. 30. The calibration system of clause 28, wherein the calibration apparatus being configured to calibrate the optical device includes adjusting one or more operating parameters of the spectral analysis module. 31. A calibration system as described in clause 25, wherein the degradation metric corresponds to a local linear approximation of the slope of the trend of the measured characteristic. 32. A prediction apparatus in communication with a calibration apparatus configured to calibrate an optical device configured to measure an aspect of a light beam generated by a light source, the prediction apparatus comprising: an input module configured to receive from the optical device a measured characteristic associated with a measured aspect of the light beam while the optical device is being calibrated; an adaptation module configured to receive the measured characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the measured characteristics, and estimate when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module configured to schedule a calibration of the optical device based on the estimates from the adaptation module and to instruct a calibration device to calibrate the optical device based on the schedule; A prediction device comprising: 33. A calibration apparatus configured to calibrate an optical device configured to measure an aspect of a light beam generated by a light source; a predictive controller in communication with the calibration device, an input module configured to receive from the optical device a characteristic associated with the optical device while the optical device is being calibrated; an adaptation module configured to receive the optical device characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics, and estimate when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module configured to schedule a calibration of the optical device based on the estimates from the adaptation module and to instruct a calibration device to calibrate the optical device based on the schedule; A predictive controller including: A calibration system including: 34. A calibration system as described in clause 33, wherein the optical device is configured to provide an estimate of the energy of the light beam based on the measurement scale. 35. A calibration system as described in clause 34, wherein the optical device characteristics include a measurement scale. 36. A calibration system as described in clause 34, in which the measurement scale allows conversion of a direct measurement of a characteristic of the light beam into an indirect value of the energy of the light beam. 37. Calibration equipment is a power meter that receives at least a portion of the optical beam directed to the optical device and outputs a measured power; a processor configured to calculate the energy in the pulses of the light beam based on the pulse repetition rate of the light source and the measured power output from the power meter, and to compare the calculated energy with an estimated energy of the light beam from the optical device; 35. A calibration system as described in clause 34, comprising: 38. The calibration system of clause 37, wherein the processor is configured to estimate drift in a measurement scale of the optical device based on the comparison. 39. The calibration system of clause 38, wherein the calibration apparatus being configured to calibrate the optical device includes adjusting a measurement scale of the optical device based on the comparison. 40. A calibration system as described in clause 38, wherein degradation of the optical device is indicated by an estimation of drift in a measurement scale of the optical device. 41. A calibration system as described in clause 34, wherein the degradation metric corresponds to a local linear approximation of a slope of a measurement scale trend of the optical device. 42. A prediction apparatus in communication with a calibration apparatus configured to calibrate an optical device configured to measure an aspect of a light beam generated by a light source, the prediction apparatus comprising: an input module configured to receive from the optical device a characteristic associated with the optical device while the optical device is being calibrated; an adaptation module configured to receive the optical device characteristics from the input module, calculate a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics, and estimate when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module configured to schedule a calibration of the optical device based on the estimates from the adaptation module and to instruct a calibration device to calibrate the optical device based on the schedule; A prediction device comprising:
Claims
1. 1. A method for scheduling calibration for a light device in a light source, comprising: receiving a characteristic associated with the optical device while the optical device is being calibrated; calculating a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics; estimating when an impairment of the optical device will exceed a threshold based on the current impairment metric; scheduling a calibration of the optical device based at least in part on the estimate of optical device degradation; A method comprising:
2. The method of claim 1 , wherein the scheduling the calibration of the light device based on the estimation comprises determining when a calibration of the light device should be performed.
3. The method of claim 1 , wherein the calculating of the current degradation metric is based on a calibration characteristic.
4. The method of claim 1 , wherein the optical device measures a spectral characteristic of a light beam produced by the light source.
5. The method of claim 4 , wherein the measured spectral characteristic of the light beam is the wavelength of the light beam.
6. 2. The method of claim 1, wherein receiving the optical device characteristics while the optical device is being calibrated comprises receiving an error measurement associated with a spectral characteristic of the light beam generated by the light source while the optical device is being calibrated.
7. The method of claim 6 , wherein the error measure is a difference between a measured spectral characteristic of the light beam and a reference spectral characteristic.
8. The method of claim 1 , wherein the calculating the current degradation metric based on at least the optical device characteristic comprises estimating how much the optical device characteristic changes in response to changes in usage of the optical device.
9. 9. The method of claim 8, wherein estimating how much the optical device characteristic will change in response to the change in usage of the optical device comprises estimating an error based on data relating to previous values of the optical device characteristic and previous degradation metrics.
10. The method of claim 1 , wherein the calculating the current degradation metric based on at least the optical device characteristic comprises assuming that the optical device characteristic changes linearly with changes in usage of the optical device.
11. the calculating the current degradation metric based on the optical device characteristic includes estimating an error in the optical device characteristic based on a linear function, where a previous degradation metric is a slope of the linear function; The method of claim 1 , wherein the linear function comprises a previous value of the usage of the optical device and a cumulative sum of the error in the optical device characteristic.
12. The method of claim 11 , wherein the calculating the current degradation metric based on the optical device characteristic comprises estimating an error in the optical device characteristic.
13. The method of claim 1 , wherein the calculating the current degradation metric based on at least the optical device characteristic comprises analyzing how the optical device characteristic is trending.
14. The method of claim 1 , wherein the degradation metric models a trend behavior of the optical device.
15. The method of claim 1 , wherein the scheduling the calibration of the optical device based on the estimation comprises determining whether the optical device characteristics vary non-linearly with usage of the optical device.
16. The method of claim 15 , wherein the determining whether the optical device characteristic varies non-linearly with the usage of the optical device comprises analyzing a confidence in the current degradation metric.
17. The method of claim 1 , wherein the estimating when the degradation of the optical device will exceed a threshold comprises estimating when an error in the optical device characteristic exceeds a threshold.
18. The method of claim 1 , wherein the scheduling the calibration of the light device based on the estimation comprises estimating an amount of additional usage of the light device that would be acceptable before requiring the calibration.
19. 2. The method of claim 1, further comprising: after scheduling the calibration of the optical device, updating the optical device characteristics when the next scheduled calibration is performed; and assigning the current degradation metric to a previous degradation metric.
20. The method of claim 1 , wherein the degradation metric corresponds to a locally linear approximation of the slope of the optical device characteristic.
21. The method of claim 1 , wherein the optical device estimates the energy of a light beam generated by the light source.
22. receiving the optical device characteristic while the optical device is being calibrated includes receiving a measurement scale associated with the optical device that performs a conversion from a direct measurement to an indirect value; The method of claim 1 , wherein the direct measurements are of the light beam generated by the light source while the optical device is being calibrated.
23. 23. The method of claim 22, wherein estimating when the degradation of the optical device will exceed the threshold comprises estimating when a drift in the optical device characteristic will exceed a drift threshold.
24. 24. The method of claim 23, wherein estimating when the drift in the optical device characteristic will exceed the drift threshold comprises calculating the drift threshold based on an error threshold associated with an indirect value.
25. a calibration apparatus for calibrating an optical device that measures an aspect of a light beam generated by a light source; a predictive controller in communication with the calibration device, an input module that receives from the optical device a measured characteristic associated with the measured aspect of the light beam while the optical device is being calibrated; an adaptation module that receives the measured characteristics from the input module, calculates a current degradation metric that models a behavior of the optical device based on at least the measured characteristics, and estimates when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module for scheduling a calibration of the optical device based on the estimates from the adaptation module and instructing the calibration unit to calibrate the optical device based on the schedule; a predictive controller having A calibration system comprising:
26. 26. The calibration system of claim 25, wherein the measured characteristic associated with the light device comprises a measurement of a spectral characteristic of a light beam produced by the light source.
27. 26. The calibration system of claim 25, wherein the calibration device includes a calibration material having an optical transition profile with known energy transitions.
28. the optical device includes a spectral analysis module that senses a spectral profile of a light beam generated by a light source; 28. The calibration system of claim 27, wherein the calibration device calibrates the spectral analysis module using the calibration material.
29. 30. The calibration system of claim 28, wherein the estimate of the degradation of the optical device is an indication of drift in the calibration of the spectral analysis module.
30. 30. The calibration system of claim 28, wherein the calibration apparatus calibrating the optical device comprises adjusting one or more operating parameters of the spectral analysis module.
31. 26. The calibration system of claim 25, wherein the degradation metric corresponds to a local linear approximation of the slope of the trend of the measured characteristic.
32. 1. A prediction apparatus in communication with a calibration apparatus that calibrates an optical device that measures aspects related to a light beam generated by a light source, the prediction apparatus comprising: an input module that receives from the optical device a measured characteristic associated with the measured aspect of the light beam while the optical device is being calibrated; an adaptation module that receives the measured characteristics from the input module, calculates a current degradation metric that models a behavior of the optical device based on at least the measured characteristics, and estimates when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module for scheduling a calibration of the optical device based on the estimates from the adaptation module and instructing the calibration unit to calibrate the optical device based on the schedule; A prediction device comprising:
33. a calibration apparatus for calibrating an optical device that measures an aspect of a light beam generated by a light source; a predictive controller in communication with the calibration device, an input module for receiving characteristics associated with the optical device from the optical device while the optical device is being calibrated; an adaptation module that receives the optical device characteristics from the input module, calculates a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics, and estimates when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module for scheduling a calibration of the optical device based on the estimates from the adaptation module and instructing the calibration unit to calibrate the optical device based on the schedule; a predictive controller having A calibration system comprising:
34. 34. The calibration system of claim 33, wherein the optical device provides an estimate of the energy of the optical beam based on a measurement scale.
35. 35. The calibration system of claim 34, wherein the optical device characteristic includes the measurement scale.
36. 35. The calibration system of claim 34, wherein the measurement scale enables conversion of a direct measurement of a property of the light beam to an indirect value of the energy of the light beam.
37. The calibration device comprises: a power meter receiving at least a portion of the optical beam directed to the optical device, the power meter outputting a measured power; a processor for calculating the energy in the pulses of the light beam based on the pulse repetition rate of the light source and the measured power output from the power meter, and comparing the calculated energy to the estimated energy of the light beam from the optical device; 35. The calibration system of claim 34, comprising:
38. 38. The calibration system of claim 37, wherein the processor estimates drift in the measurement scale of the optical device based on the comparison.
39. 40. The calibration system of claim 38, wherein the calibration apparatus calibrating the optical device comprises adjusting the measurement scale of the optical device based on the comparison.
40. 39. The calibration system of claim 38, wherein the degradation of the optical device is indicated by the estimate of the drift in the measurement scale of the optical device.
41. 35. The calibration system of claim 34, wherein the degradation metric corresponds to a local linear approximation of the slope of the trend of the measurement scale of the optical device.
42. 1. A prediction apparatus in communication with a calibration apparatus that calibrates an optical device that measures aspects related to a light beam generated by a light source, the prediction apparatus comprising: an input module for receiving characteristics associated with the optical device from the optical device while the optical device is being calibrated; an adaptation module that receives the optical device characteristics from the input module, calculates a current degradation metric that models a behavior of the optical device based on at least the optical device characteristics, and estimates when an degradation of the optical device will exceed a threshold based on the current degradation metric; an output module for scheduling a calibration of the optical device based on the estimates from the adaptation module and instructing the calibration unit to calibrate the optical device based on the schedule; A prediction device comprising:
Citation Information
Patent Citations
Laser diode array assembly and its operating method
JP1998190134A
Absolute wavelength stabilizing method of excimer laser, stabilizer and aligner
JP2001257406A
Monitoring device and ultraviolet laser device
JP2003214949A
Laser light source device
JP2009009955A
Wavelength detector and wavelength calibration system
JP2013033932A
