Model-based scanner tuning systems and methods

0001<u style="single">Cross-reference to related applications</u> [0001] The present application is US Provisional Patent Application No. 61 / 141,578 filed on December 30, 2008, US Provisional Patent Application No. 61 / 142,305 filed on January 2, 2009, June 3, 2008. Claim the priority of US Provisional Patent Application No. 61 / 058,511 filed in, and US Provisional Patent Application No. 61 / 058,520 filed on June 3, 2008, and explicitly incorporate them in the present application as a reference. ..

0002[0002] The present invention relates generally to systems and methods for performing model-based scanner tuning and optimization, and more specifically to optimizing the performance of a plurality of lithography systems.

0003[0003] Lithographic equipment can be used in the manufacture of integrated circuits (ICs). The mask contains a circuit pattern corresponding to each layer of the IC, which pattern is placed on a target portion containing one or more dies on a substrate consisting of a silicon wafer coated with a layer of radiation sensitive resist material. Image image. Generally, a single wafer contains a network of adjacent target portions that are continuously irradiated one at a time through a projection system. In one type of lithographic projection apparatus, each target portion is irradiated by exposing the entire mask pattern onto the target portion in a single time, and such a device is usually called a wafer stepper. The step-and-scan device progressively scans the mask pattern under a projected beam in a particular reference or "scan direction" while illuminating each target area by scanning the substrate parallel or antiparallel in this direction. To do. In a projection system with a magnification factor of M (usually &lt;1), the speed V of scanning the substrate table is M multiplied by the speed of scanning the mask table. Further information about lithographic devices as described herein can be obtained, for example, from US Pat. No. 6,046,792, which is incorporated herein by reference.

0004[0004] In a manufacturing process using a lithographic projection apparatus, a mask pattern is imaged on a substrate that is at least partially covered by a layer of radiation sensitive resist material. Prior to this imaging step, the substrate can undergo various procedures such as priming, resist coating, and soft baking. After exposure, the substrate can perform other procedures such as post-exposure bake (PEB), development, hard bake and measurement / inspection of the imaged features. This sequence of steps is used as a basis for patterning individual layers of devices such as ICs. Such patterned layers can then undergo various processes such as etching, ion implantation or doping, metallization, oxidation, chemical mechanical polishing, etc., followed by the finishing of the individual layers. If several layers are required, the entire procedure, or its transformation, must be repeated for each new layer. Finally, the device array is provided on the substrate wafer. These devices can then be separated from each other by techniques such as dicing or sewing, allowing individual devices to be mounted on carriers, connected to pins, and so on.

0005[0005] Projection systems (lenses) include, for example, various types of projection systems, including catadioptric systems, catadioptric systems, and catadioptric systems, and may include one or more lenses. The lens can also include components of the radiation system used to guide, shape, or control the radiation projection beam. Further, the lithographic apparatus may be of a type having two or more substrate tables and / or two or more mask tables. In such multistage devices, additional tables can be used in parallel, and / or preliminary steps can be performed on one table while another table can be used for exposure. Twin-stage lithography equipment is described, for example, in US Pat. No. 5,969,441, which is incorporated herein by reference.

0006[0006] The photolithography mask described above includes a geometric pattern corresponding to a circuit component integrated on a silicon wafer. The patterns used to generate such masks are generated using computer-aided design (CAD) programs, a process often referred to as electronic design automation (EDA). Most CAD programs generate functional masks according to a set of predetermined design rules. These rules are set by processing constraints and design constraints. For example, design rules define spatial tolerances between circuit devices or interconnect lines, such as gates and capacitors, to ensure that circuit devices or wires do not interact in an undesired way. The limits of design rules are usually referred to as critical dimensions (CD). The CD of a circuit can be defined as the minimum width of a wire or hole, or the minimum space between two wires or two holes. Therefore, the CD determines the overall size and density of the circuit being designed. Of course, one of the goals of integrated circuit fabrication is to faithfully reproduce the original circuit design on the wafer through a mask.

0007[0007] In general, different types of lithographic systems (eg, scanners) do not require a great deal of time and resources to determine the required settings for each lithographic system to achieve optimal / acceptable imaging performance. ) Can benefit from using a common process to image a given pattern. Numerical aperture (NA), σ, when designers and engineers initially set the process for a particular scanner to obtain an image that meets the predefined design conditions._in, Σ_{out out}It spends a great deal of time and money determining the optimal settings for the lithography system, including. In many cases, a trial-and-error process is employed, where the scanner settings are selected, the desired pattern is imaged and then measured to determine if the output image is within specified tolerances. If the output image is not within tolerance, adjust the scanner settings and re-image the pattern for measurement. Repeat this process until the resulting image is within the specified tolerances.

0008[0008] However, even if the scanners are of the same model type, the actual pattern imaged on the substrate will be due to the different light proximity effects (OPE) presented by the different scanners when the pattern is imaged. Can vary from one to another. For example, different OPEs associated with a particular scanner may introduce significant CD variation throughout the pitch. As a result, it is often impossible to replace the scanners to obtain the same imaging pattern. Therefore, if the scanner is new or different, and if it is used to print a pattern with the expectation of obtaining an image of the result that meets the design criteria, the technician needs to optimize or adjust the scanner. Currently, this expensive and time-consuming trial-and-error process is used to tune processors and scanners.

0009[0009] In current state-of-the-art technology, a common form of scanner adjustment is proximity matching. The purpose is to align the printed wafer CDs to a set of predefined patterns between the adjustable scanner and the reference scanner. In general, attention is being paid to a one-dimensional pattern (1D pattern) that passes through a pitch. This is because the critical dimension uniformity for that pattern is of paramount importance to semiconductor device performance. The predefined pattern is exposed on the wafer using a reference scanner and an adjustable scanner and the wafer CD value is measured. Differences in the CD are used to propel the adjustment offset on the adjustable scanner to match the CD values after adjusting to those from the reference scanner. The optimization is done in a linear fashion, assuming a linear dependence of the CD value on the adjustment offset. The linear dependence is characterized by the sensitivity defined as the partial derivative of the CD value with respect to the knob offset. Sensitivity may be measured or simulated from a lithography model such as that provided by US Pat. No. 7,003,758.

0010[0010] The existing method has some drawbacks that the present invention seeks to overcome. First, every pattern to be matched needs to be measured, which is not the most efficient use of wafer metrology time in manufacturing (usually in high demand). Conversely, there are no requirements for the level of matching or imaging behavior for patterns other than the one being measured. This is generated such that a set of 1D patterns are well matched, but some two-dimensional ("2D") real device patterns have obvious inconsistencies in the adjusted wafer imaging. Known to cause problems in the environment. "Accurate Model Base Verification Scheme To Eliminate Hotspots And Manage Warmspots", Proc.SPIE, Vol.6925, 69250Z (2008) and "Scanner Fleet Management Utilizing Programmed Hotspots" Please refer to "Patterns", Proc.SPIE, Vol.7028,70280W (2008).

0011[0011] One embodiment of the present invention includes a system and method for coordinating a photolithography process. Scanner adjustments can be categorized as scanner adjustments, scanner adjustments for process alignment, and scanner adjustments for performance optimization. Hereinafter, the adjustable scanner to be adjusted is referred to as a target scanner, and the desired result of the adjustment movement is referred to as a reference. In certain embodiments, the adjustment criterion may be a measured wafer contour or CD, a simulated wafer contour or CD, or a design target polygon.

0012[0012] In one embodiment of the invention, the model of the target scanner is maintained, which defines the sensitivity of the target scanner and the components of the target scanner into a set of adjustable parameters. The difference model may be generated to represent the deviation of the target scanner with respect to the reference. The target scanner may be tuned based on the reference scanner and diff model settings.

0013[0013] One embodiment provides a system and method that characterizes the performance of a family of related scanners relative to the performance of a reference scanner. A family of scanners may include scanners manufactured by one vendor, and the scanners within the family may belong to the same model type or different model types. The family of scanners may include scanners manufactured by different vendors, including at least some functionally similar elements. For example, a scanner using a particular wavelength laser may be modeled by a common base model. Where the family of scanners is modeled by a common base model, additional delta models can be used to maintain calibration information and specific adjustment information that accommodate changes for individual family members relative to the common base model. .. The difference model may include information such as parametric offsets and other differences that may be used to simulate differences in imaging behavior.

0014[0014] One embodiment of the present invention includes model-based simulations down to the full chip level to determine the deviation between the achieved wafer contour and the reference. Combining such simulated deviations with measured deviations facilitates optimization of target scanner settings. In some embodiments, this optimization involves one or more iterations.

0015[0015] In certain embodiments, changes in the critical dimension (CD) or wafer contour resulting from changes in the scanner knob are simulated via a sensitivity model for the target scanner.

0016[0016] The present invention itself may be better understood by reference to the following detailed description and accompanying schematics, along with additional objectives and advantages.

0017<figref num="1">[0017] FIG. 1 shows a lithography model according to an aspect of the present invention.</figref><figref num="2">[0018] FIG. 2 shows a general procedure for calibrating a lithography model according to certain aspects of the invention.</figref><figref num="3">[0019] FIG. 3 shows a process for generating, adjusting and optimizing a differential lithography model according to certain aspects of the invention.</figref><figref num="4">[0020] FIG. 4 shows an example of a process for simulating and predicting optical parameters from a scanner model complemented by scanner metrology, according to an embodiment of the present invention.</figref><figref num="5">[0021] FIG. 5 shows sensitivity modeling according to certain aspects of the invention.</figref><figref num="6">[0022] FIG. 6 shows a process for calibrating a differential model for multiple scanners according to certain aspects of the invention.</figref><figref num="7">[0023] FIG. 7 illustrates the relationship between the base model parameters and the derived model in certain embodiments of the present invention.</figref><figref num="8">[0024] FIG. 8 shows the generation of simulated contours from a differential model according to certain aspects of the invention.</figref><figref num="9">[0025] FIG. 9 is a block diagram showing a computer system according to an aspect of the present invention.</figref><figref num="10">[0026] FIG. 10 schematically shows a lithography projection apparatus according to an aspect of the present invention.</figref><figref num="11">[0027] FIG. 11 is a flowchart showing an adjustment method using full-chip simulation according to an aspect of the present invention.</figref><figref num="12">[0028] FIG. 12 is a flowchart showing an adjustment process adopting a mini-layout method according to an aspect of the present invention.</figref>

0018[0029] Embodiments of the present invention will be described in detail below with reference to the drawings. The drawings are provided as illustrated examples so that those skilled in the art can practice the present invention. In particular, the figures and examples below are not intended to limit the scope of the invention to a single embodiment, but to allow other embodiments by exchanging some or all of the elements described or illustrated. Become. Wherever convenient, the same reference numbers are used throughout the drawings to indicate the same or equivalent parts. Where certain elements of these embodiments can be partially or fully realized using known components, only the parts of such known components that are necessary for the understanding of the present invention will be described. In order not to obscure the present invention, detailed description of other parts of such known components will be omitted. In the present specification, one embodiment showing a singular component should not be regarded as limited, but rather, the present invention includes a plurality of the same components unless otherwise expressly stated in the present specification. And vice versa. Moreover, the Applicant shall not assign any unusual or special meaning to any term within this specification or claims unless such explicit provisions are made. In addition, the invention includes current and future known equivalents to the components referred to herein by way of example.

0019[0030] In one embodiment of the invention, full chip wafer simulation and verification is employed as an alternative or complement to full chip wafer measurements for scanner adjustment. The model used during the simulation may include a sensitivity model and a difference model. The sensitivity model describes changes in the imaging behavior of the scanner in response to adjustment inputs (ie, when the knob is turned). The difference model describes and parameterizes the differences in the behavior of the lithography process under known settings. The difference model is calibrated using scanner sensor data such as Jones pupil, illuminator map, and wafer metrology data.

0020[0031] FIG. 1 shows a lithography model 10 according to an aspect of the present invention. The lithography model includes a mask model 100, an optical model 102 and a resist model 104. In some embodiments, the lithography model also includes an etch model and is not shown in the drawings for brevity. The mask model may reflect the volatility introduced by changes in multiple mask parameters 120. The optical model 102 may be affected by changes in the optical parameter 122 and the resist model 104 may be controlled by the setting of the resist parameter 124. Model 10 may be used to predict the resist contour 164, or if an etch model component is included, to predict the post-etch contour generated from the mask design 140. The mask model 100 configured by the mask parameter 120 produces the predicted mask image 160, and if this mask image 160 is provided to the optical model 102, the simulated optical image 162 based on the optical parameter 122. Generate. A resist model 104 composed of resist parameters 124 may be used to predict a resist contour 164 from a simulated optical image 162. If an etch model composed of etching parameters is included, the etch model may be used to predict the post-etch contour from the resist contour 164.

0021[0032] Optical parameter 122 includes adjustable and non-adjustable parameters. Here, the "adjustable parameter" refers to a knob (knob) that can be adjusted on the scanner such as NA (numerical aperture), and the "non-adjustable parameter" is Jones for general scanner design. Refers to scanner parameters that cannot be adjusted, such as the pupil. The method of the present invention does not depend on which parameters are adjustable or non-adjustable on the scanner. For model calibration purposes, both non-adjustable and adjustable parameters may be adjusted until the image produced by the model matches the actual imaging result produced by the reference scanner. The adjustment of a parameter in model calibration depends on the degree of knowledge of this parameter, not on the adjustability. For example, if accurate measurements of the illuminated pupil are possible via scanner metrology, such measurements may be used directly in model calibration without further adjustment. On the other hand, parameters that do not involve direct measurement via scanner metrology are optimized to fit the wafer data. Scanner metrology measurements may be made using an integrated lens interferometer. In one embodiment, the integrated lens interferometer is a wavefront sensor and is used to measure lens aberrations at each field point. The wavefront sensor is based on the principle of shear interferometry and includes a source module and a sensor module. The source module has a patterned chrome layer disposed on the objective plane of the projection system, on which additional optics are provided. This combination provides a wave front of radiation to the entire pupil of the projection system. The sensor module has a patterned chrome layer arranged on the image plane of the projection system and a camera located rearward at some distance from such chrome layer. The patterned chromium layers on the sensor module diffract the radiation into several diffraction orders that interfere with each other and raise the interferogram. The interferogram is measured by a camera. projection

0022[0033] FIG. 2 shows a general procedure for calibrating the lithography model 222. One or more mask designs 200 may be used for calibration. In some embodiments the mask design 200 may be generated specifically for calibration, while in other embodiments the mask design generated for production use is used for calibration. The modeled mask, optics and resist parameters 220 used in the lithography model 222 are selected to reflect the mask, optics and resist effect 240 used in the lithography process 242. The resulting simulated resist contour 224 and measured resist contour 244 may be compared and analyzed, with parameter 220 being optimal for minimizing the difference between the simulated contour and the measured contour. It may be converted. An analysis can be performed using the cost function 260, which will be described in more detail below.

0023[0034] In certain embodiments, the model calibration process involves all measurements, including both wafer metrology (CD-SEM measurements and contours, scatterometry, etc.) and scanner data (either designed or measured). And each uncertainty is taken into account and balanced and formulated as a maximum likelihood problem. In one embodiment, the calibration process is repetitive and the model parameters are repeatedly adjusted to obtain calibration that provides the imaging results produced by the model measured to be close enough to the actual wafer data. Predefined error criteria can be established and / or criteria for "best possible match" can be defined or quantified. In certain embodiments, any suitable model that simulates the imaging performance of the scanner can be used, which is provided by the system and method of US Pat. No. 7,003,758.

0024<u style="single">Absolute accuracy vs. difference accuracy</u> [0035] In contrast to traditional model-based OPC applications, great emphasis has been placed on absolute prediction accuracy for CD-SEM measurements, generally under nominal exposure conditions. The advent of OPC validation across process windows and process window sensing OPCs has expanded its importance to include predictive accuracy across process windows (US Patent Application No. 11 / 461,994, System and Method For Creating a Focus-Exposure). Model of a Lithography Process "). However, the figure of merit remains the difference between the measured CD and the predicted CD.

0025[0036] The importance is necessarily different for model-based scanner tuning, including matching and performance optimization. The amount of interest arises from changes in scanner settings, differences between scanners and / or differences between processes. The quantity is generally measurable below about a few nanometers, which is comparable to the absolute accuracy of a typical OPC model. Modeling, simulating and predicting such differences imposes different conditions on model accuracy compared to those required for OPC modeling. One embodiment of the present invention employs a new algorithm that handles and satisfies these different conditions.

0026[0037] Figure 3 shows the process of generating, adjusting, and optimizing the differential lithography model 322. The mask design 300 undergoes processing by multiple scanners 342 and simulations using the scanner 322 models under a set of process conditions 340. The simulated resist contour 324 may be analyzed against the physically generated resist contour 344. Model parameters can be adjusted using the cost function 360 (discussed below) to obtain a model that can accurately characterize one or more differential models associated with multiple scanners.

0027The exact difference model still formally simulates the pattern contour on the wafer either after resist development or after etching. However, the purpose of such a model is not necessarily absolute CD accuracy, but the accuracy of prediction for CD or contour changes when one or more model parameters are perturbed. It describes the differences between scanners or simulates the effects of active scanner adjustments. Therefore, the simulation can require two paths, one with parameter perturbation and one without parameter perturbation. The amount of interest in a given pattern i is as follows.<maths num="1"><img id="000002" he="6" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>

0028<u style="single">Derived model generation</u> [0039] Assuming that a model with sufficient difference accuracy (difference model) is available, certain aspects of the invention facilitate the generation of derived models based on the difference model and the base model. .. In one embodiment, the base model is the same as the pre-perturbation model, in which case the derived model is the same as the post-perturbation model. In these embodiments, the derived model requires only one imaging simulation using a perturbation model. In other embodiments, the base model differs from the pre-perturbation model, in which case the derived model requires three imaging simulations, each imaging simulation using a base model, a non-perturbative model and a perturbation model. In an example of a later embodiment, the base model may be an OPC model.

0029<u style="single">Sensitivity modeling</u> [0040] FIG. 4 shows the effect of knob setting 400 on optical parameters via scanner model 402 and scanner metrology 404. Certain optical parameters are unaffected by changes in the available or in-use scanner knobs and are thus completely fixed by the scanner metrology. This example includes a laser spectrum for a scanner adapted by a laser that does not have bandwidth control. In other cases, the optical parameters are affected by knob changes and can be derived from the combination of the scanner model 402 and the scanner metrology 404. For example, the illuminated pupil is affected by NA and sigma changes for a particular type of scanner, as well as other changes, including ellipticity settings. Therefore, the illuminated pupil can be predicted using pupil measurements combined with a scanner model.

0030[0041] Figure 5 predicts imaging changes (ie, critical dimension changes and contour changes) for any pattern in response to configuration changes for one scanner, while keeping all other aspects of the lithography process unchanged. The basic aspect of the present invention including. In the illustrated example, a series of N simulations are performed, where each simulation is available under the simulated conditions or the measured contours 560-562 generated under those conditions (respectively). ) Corresponds to the simulated contours 540 ~ 542. Each simulation may be distinguished by a different set of knob settings 500-502 used in the scanner model 510. Because the scanner model 510 produces optical parameters 520-522 that can be arbitrarily generated using the input from the scanner metrology 512, and the optical parameters 520-522 generate simulated contours 540-542, respectively. Used for. Simulated contours 540-542 and measured contours 560-562 may be analyzed to generate, calibrate and optimize model parameters 572. In one example, the simulated and measured contours may be mathematically processed using the cost function 570.

0031[0042] Symbolically, the purpose of sensitivity modeling is knob change Δk_jCD change ΔCD with respect to pattern i in response to_iIs to predict. For typical scanner adjustment applications, the linear model can work reasonably well due to the small amount of adjustment, but the invention is not limited to linear model scenarios. Therefore, wherever a linear model can be applied<maths num="2"><img id="000003" he="14" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>And the purpose of the sensitivity model is the partial derivative<maths num="3"><img id="000004" he="13" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Is calculated, and here the mask pattern is i. According to the chain rule of derivatives<maths num="4"><img id="000005" he="13" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>And here p_mRefers to the physical parameters in the scanner model. Therefore, the first factor<maths num="5"><img id="000006" he="12" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Is related to the lithography imaging model, while the second factor<maths num="6"><img id="000007" he="13" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Is related to the scanner model.

0032[0043] In a more general non-linear example, the physical properties and model can be represented as follows.<maths num="7"><img id="000008" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Resist, optical and scanner physics can be represented as separate modeling components. The accuracy of the sensitivity model depends on the accuracy of both the litho model (optical and resist) and the scanner model 510.

0033The resist model may be empirical or based on the physical and chemical properties of the resist process. Optical models are usually physical models and are based on first principles and have the potential for approximation processing of certain effects, such as 3D scattering of EM radiation by masks, to reduce simulation time. Other approximation methods, such as truncation of the optical interaction range (also known as the finite region) or truncation of the TCC eigenseries in the Hopkins method, are also possible. The scanner model 510 may be based on the physical considerations and design knowledge of the scanner. Different levels of rigor may exist for the scanner model. For example, ray tracing-based models can generate very accurate predictions of the pupil, but tend to be very computationally expensive. Approximate and more empirical models may be constructed by calibrating the exact model or measurement.

0034The concept of sensitivity model accuracy is closely related to the concept of model separability, both of which relate to imaging predictions for different scanner settings. See, for example, US Patent Applications 11 / 461,929 and 11 / 530,402. For OPC-type applications, separable models are desirable because of the predictive system (typically focus and exposure) across the process window and the reduction of model calibration inversion time when exposure settings change. Litho models generally include optical models, resist models, and sometimes etch models, emphasizing separability between different model steps.

0035[0046] One difference element of the sensitivity model for the purpose of scanner adjustment is the incorporation of a predictive scanner model that requires detailed knowledge of scanner design. An exemplary component of the scanner model 510 is an illuminator prediction model that simulates illumination optics to predict illumination on the reticle surface. For sensitivity modeling, this model predicts changes in the illuminator under changes in exposure settings such as NA, sigma and PUPICOM settings.

0036[0047] If the resist process is the same or close enough to multiple scanners, and the calibrated resist model is part of an accurately calibrated sensitivity model from a single scanner lithographic process. The separability of the model form also allows a precisely calibrated resist model to move between multiple scanners. This flexibility can be important in practice, as resist models tend to be more empirical than optical and scanner models, and thus require more constraints from wafer-based calibration. Therefore, moving the resist model allows for efficient use of wafer metrology. The scanner model 510 and the optical model are based on first principles and known physics and are less dependent on wafer measurements.

0037[0048] In other embodiments, the lithography process is quite different at the resist portion. For example, one process employs immersion lithography and another process does not. The two processes typically use completely different resist materials and film stacks. In the example, the resist model cannot be moved between the two processes and the resist effect is quite different, so the sensitivity models need to be created separately.

0038[0049] For calibration of the sensitivity model, one embodiment includes detailed scanner data such as Jones pupil, stage vibration, chromatic aberration and defocus due to the laser spectrum. In some embodiments, calibrating the sensitivity model allows the metrology data to be set up in multiple scanners, or perturbed (k)._j+ Δk_j) Nominal state to add k_jNeed to take. One or more knobs may change for each perturbation state. The cost function for sensitivity model calibration is as follows.<maths num="8"><img id="000009" he="27" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>The first term quantifies the absolute accuracy through the weighted RMS difference between the model and the wafer, and the second term quantifies the sensitivity accuracy by comparing the model prediction CD changes with those measured on the wafer. The relative weighting of absolute and sensitivity accuracy can be adjusted. In addition, other metrics can be used instead of RMS, such as range (maximum to minimum) or LP norm. Then calibration can go into optimization problems, which are often constrained.

0039[0050] The calibrated sensitivity model may be applied at full chip level to predict imaging differences for all patterns that occur in the chip design.

0040Note that the sensitivity model may be the same as or different from the lithography model used for OPC or OPC verification. In certain embodiments, the sensitivity model uses more lithographic process knowledge than the OPC model associated with masks, scanner optics and resists. For example, in one embodiment, the OPC model has nominal mask or Kirchhoff boundary conditions for mask diffraction, a small optical interaction range and / or a few terms from the TCC eigen series expansion. Or use only the ideal optical system. These modeling methods may be inadequate for the accuracy requirements of sensitivity modeling. Therefore, in certain embodiments, the sensitivity model uses scanner optics, 3D mask diffraction, a larger optical interaction range, and / or more TCC terms. The test pattern used to calibrate the sensitivity model may be the same as or different from that used for the OPC or OPC validation model.

0041[0052] In certain embodiments, the sensitivity model may be combined with different base models, such as the OPC model, to form a new derived model. This new derived model can be formed by applying the delta to the model parameters, simulated spatial image or simulated resist image, but simulating the delta CD from the differential model or the contour edge position from the base model. It may be formed by applying it to a simulated CD or contour edge position. Applying deltas to model parameters can only be done if the base model contains perturbed parameters and such parameters are used in the correct way. In one embodiment, the base model is a calibration model with different forms, different vendors of modeling software or different combinations of model components, which poses difficulties for applying parameter deltas directly. In particular, the base model is top-hat An illumination) shape may have been used, in which case applying the delta-sigma value to crown illumination does not give accurate results. The resist model in the base OPC model also tends to be inadequate in terms of difference accuracy. Under such circumstances, it is possible to combine the base OPC model and the sensitivity model at the simulated CD or contour level.

0042Combining the sensitivity model with the base OPC model offers at least two advantages. First, the OPC model serves to ensure absolute CD prediction accuracy for certain conditions, typically calibrated by a large pattern set. Therefore, combining the sensitivity model with the OPC model can provide an accurate prediction of absolute CD in the presence of scanner knobs or parameter changes. Second, OPC correction is performed by the OPC model, which is expected to have simulated contours from the OPC model very close to the pre-OPC target pattern. Therefore, combining the sensitivity model with the OPC model allows simulation-based validation against pre-OPC targets in the presence of scanner knobs or parameter changes.

0043<u style="single">Difference modeling</u> [0054] In certain embodiments, system-level simulation comprises defining the performance of a family of related scanners relative to the performance of a reference scanner. A family of scanners may include scanners manufactured by one vendor and which may belong to the same model type. The family of scanners may include scanners manufactured by different vendors, where the scanner contains at least some functionally similar elements. The family of scanners is modeled by an additional difference model in addition to the common base model to maintain calibration information that accommodates the differences of individual family members from the common base model.

0044FIG. 6 shows a process for calibrating a difference model for multiple scanners according to certain aspects of the invention. In the illustrated example, a set of N scanners 600-602 is simulated. The scanner model 610 uses the input from the scanner metrology 612 to generate optical parameters 620-622 for each of the scanners 600-602. Simulated contours 640-642 are generated using optical parameters 620-622, respectively, which are then processed with measured contours 660-662 to calibrate and optimize model parameters 672. The simulated and measured contours may be processed mathematically using the cost function 670.

0045[0056] For the purposes of differential model calibration, both non-adjustable and adjustable scanner parameters may be adjusted until the simulated differences produced by the model match the actual wafer differences. The adjustment of the parameter in the difference model depends on the recognition level of this parameter, not on the adjustability. For example, if accurate measurements of the illuminated pupil are available for multiple scanners 600-602 via the scanner metrology 612, such measurements are directly in the model calibration without further adjustment. May be used. On the other hand, the parameters without direct measurement via the scanner Metrology 612 are optimized to fit the wafer data. In certain embodiments, the model calibration process involves all measurements, including both wafer metrology (CD-SEM measurements and contours, scatterometry, etc.) and scanner data (either designed or measured) and each. It is formulated as the maximum likelihood problem, taking into account the uncertainty of and balancing.

0046In some embodiments, differential modeling is applied to a number of different lithographic processes, such as mask differences (spatial bias distribution, proximity effect due to mask formation, corner rounding), resist material differences (quenching agent concentration,). Includes differences in lithography steps excluding scanners, including diffusion), track differences (baking temperature) and etch differences.

0047One important issue associated with differential model calibration is the possibility of degeneracy between different process parameters with respect to the effect on imaging of the selected set of calibration patterns. This is because certain parameters can have a correlation or degenerate effect on the imaging of a set of semi-optimally selected calibration patterns, so the imaging difference for the calibration pattern is significantly different from the true difference depending on the calibration result. It means that it is erroneously caused by the deviation of the parameter difference. For example, the difference in exposure may degenerate with mask bias, both of which result in the feature CD changing in one direction (larger or smaller). This problem is exacerbated by the presence of disturbance noise in wafer measurements. For this reason, one embodiment selects a pattern that is sensitive to parameter differences in an "orthogonal" manner. Otherwise, erroneously calibrated parameter differences can result in erroneous predictions of imaging differences, especially for patterns not covered by the calibration set.

0048[0059] Physical Target Scanner<maths num="9"><img id="000010" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>And physical reference scanner<maths num="10"><img id="000011" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Simulations can be used to predict the differences in the physical results obtained from, which are represented as follows.<maths num="11"><img id="000012" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Modeled target scanner<maths num="12"><img id="000013" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>And a modeled reference scanner<maths num="13"><img id="000014" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>The difference model that identifies the difference in the results of is expressed as follows.<maths num="14"><img id="000015" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Therefore, the accuracy of the difference model can be expressed as follows.<maths num="15"><img id="000016" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>

0049[0060] The RMS or other metric (range, LP-norm, etc.) calculated for a set of test patterns based on the above quantities is used as a cost function for calibrating the difference model.

0050[0061] One embodiment uses the calibration procedure used when wafer data is available for both current process conditions and adjustment target process conditions. For example, if two physical scanners are modeled under the same resist process, composite calibration may be performed on the wafer data using the conditions of both the current scanner and the target scanner. This generally forces the resist model parameters to change but be the same in both current and target scanner conditions, allowing the scanner parameters to change independently under both conditions. Requires a composite model calibration process to be performed. After the combined calibration, the sensitivity model and the difference model are obtained at the same time.

0051[0062] In order to utilize the results of the differential calibration, a new model is formed from the base model and the calibrated parameter differences. The simulated CD difference between this derived model and the base model is taken as a prediction of the actual difference from the wafer measurements. FIG. 7 illustrates the relationship between the base model parameter 70 and the derived model parameter 72. The mask parameter 720 in the derived model 72 can be calculated using the mask parameter 720 and the difference 710 in the base model 70. The optical parameter 722 in the derived model 72 can be calculated using the optical parameter 702 and the difference 712 of the base model 70. The resist parameter 724 in the derived model 72 can be calculated using the resist parameter 704 and the difference 714 in the base model 70.

0052[0063] In certain embodiments, the difference model may be combined with different base models, such as the OPC model, to form a new derived model. This new derived model can be formed by applying the delta to the model parameters, simulated spatial image or simulated resist image, but simulating the delta CD or contour position from the differential model from the base model. It may be arbitrarily formed by applying it to a simulated CD or contour position. Applying deltas to model parameters can only be done if the base model contains perturbed parameters and such parameters are used in the correct way. In one embodiment, the base model is a calibration model with different forms, different vendors of modeling software or different combinations of model components, which poses difficulties for applying parameter deltas directly. In particular, the base model may have used the crown illumination shape, in which case applying the delta-sigma value to the crown illumination does not give accurate results. The resist model in the base OPC model also tends to be inadequate in terms of difference accuracy. Under such circumstances, it is possible to combine the base OPC model and the difference model at the simulated CD or contour level.

0053[0064] As shown in Figure 8, Mask Design 800 is used as an input for lithography simulation. The simulated contour A 840 is generated from the lithography model A 820 (base model). From the difference model, the simulated contours 841 and 842 are generated from the models 821 and 822. The delta between the contour 821 and the contour 822 is added to the contour 840 to form the final simulated contour 880. In some embodiments, arithmetic operations (+ and-) are applied in the sense of edge motion along the vertical direction of the contour.

0054Combining the difference model with the base OPC model offers at least two advantages. First, the OPC model serves to ensure absolute CD prediction accuracy for certain conditions, typically calibrated by a large pattern set. Therefore, combining the difference model with the OPC model can provide an accurate prediction of absolute CD in the presence of lithography process differences, including scanner differences. Second, OPC correction is performed by the OPC model, which is expected to have simulated contours from the OPC model very close to the pre-OPC target pattern. Therefore, combining the difference model with the OPC model allows simulation-based validation against pre-OPC targets in the presence of lithography process differences.

0055<u style="single">Scanner adjustment and simulation using adjustment model</u> For scanner matching and performance optimization, adjustment models are generated based on sensitivity and base models, as well as knob offsets. This includes the use of the resist model part of the sensitivity model, and changing a parameter representing the scanner knobs to include knob offsets, the base model a set and a be matched seen.

0056[0067] In certain embodiments of the invention, full-chip wafer simulation and verification is employed as an alternative to full-chip wafer measurement for scanner adjustment. The difference between the desired contour target and the actual contour (measured or simulated) may be used to facilitate the calculation of the required knob offset so that the printed contour is within the tolerance of the target. To be consistent. Details on how to adjust offset generation, simulation and validation are described below.

0057[0068] Aspects of the invention can allow the scanner to be tuned to a known model or known wafer contour or other target pattern. The processes provided by aspects of the invention include lithography process drift correction, scanner optimization for a given OPC process, scanner optimization for a particular device mask to optimize the CDU, and scanner optimization for known mask errors. to enable.

0058[0069] If desired, the effect of adjustments on the pattern can be analyzed using OPC verification tools, since the model can quantitatively analyze the effects of adjustment-related changes on the model on the full-chip pattern. In one example according to one aspect of the invention, a suitable method is to use an OPC verification tool to simulate a full-chip on-wafer contour using pre- and post-adjusted models, and the difference between the two contours. There is a step to compare and analyze the difference between the two models.

0059<u style="single">Lithography equipment and process adjustment</u> [0070] The methods of the invention that coordinate offset generation, simulation and validation according to further aspects of the invention are described below.

0060[0071] In one embodiment, the adjustment reference includes a measured wafer contour. In an alternative embodiment, the adjustment criteria include a CD, a simulated wafer contour or CD, a design target polygon, or a combination of any type of adjustment criteria described above. In one embodiment, different types of criteria (eg, wafer measurements, wafer simulations and design polygons) apply to a subset of all patterns on the chip. Measured and / or simulated wafer contours may be used as adjustment criteria to match the performance of two or more scanners and reduce variability in the manufacturing process. Design target polygons may be used as adjustment criteria to improve pattern fidelity on wafers, including, for example, CD uniformity. It is understood that the ultimate purpose of scanner adjustment is to increase the yield when producing integrated circuit chips, or to enhance the electrical performance of working integrated circuit chips produced by the lithographic apparatus adjusted by the present invention. Will be done.

0061[0072] Scanner adjustments can be categorized into scanner adjustments for process alignment, scanner adjustments for process alignment, and scanner adjustments for performance optimization, based on the type of process difference or defect to be compensated for. In the simplest case, scanner matching is employed to compensate for the differences between scanners and match the imaging performance of multiple scanners in the absence of mask, resist or etch differences. In addition, process differences may be present in the mask, resist or etch (in addition to potential scanner differences) and the scanner may be adjusted to compensate for all differences to obtain process matching. In another example, scanner adjustment can compensate for discontinuous events or yield limiting shortcomings resulting from shortcomings in the OPC correction and / or mask forming process. The adjustment may be used to improve the CD uniformity of the device layer.

0062[0073] In certain embodiments, the scanner with adjustment may be of the same type provided by the same manufacturer (eg, both ASML XT: 1900i scanners), but of the same manufacturer but of a different type (eg, for example). One ASML XT: 1900i scanner, one ASML XT: 1700i scanner), or the scanner may be manufactured by different manufacturers.

0063<u style="single">Model generation and simulation</u> [0074] One embodiment of the present invention evaluates and optimizes the effect of imaging on a large pattern set, including a full chip, as a result of scanner adjustment. Current technology for wafer metrology does not provide an economical way to achieve this goal. In one embodiment, the sensitivity model is used to derive a set of desired parameter values (knob offsets), and the desired parameter values (knob offsets) are used to obtain a simulated wafer contour. .. In a further embodiment, the difference model is used to predict the contour and / or CD difference between the reference scanner and the target scanner. Details on model generation and simulation are provided above. Suitable models for simulating image performance include, for example, the systems and methods described in US Pat. No. 7,003,758.

0064<u style="single">Adjustment flow</u> [0075] Certain embodiments of the present invention include systems and methods for coordinating photolithography processes. According to certain aspects of the invention, the adjustable and non-adjustable properties of the scanner may be modeled and used to facilitate adjustment. The target scanner may be tuned towards a reference using the target scanner's sensitivity model, where the sensitivity model defines the target scanner's imaging sensitivity to a set of adjustable parameters. The target scanner difference model may be generated to represent the deviation of the target scanner with respect to the reference in terms of coupling performance. The difference model may include non-adjustable differences in performance characteristics between scanners, which in some cases can be adapted by adjusting other adjustable parameters.

0065[0076] In certain embodiments, the tuning method involves one or more iterations, and each iteration requires full-chip simulation and validation with full-chip simulation data. In one embodiment (FIG. 11), a trial adjustment recipe (ie, a trial set of parameter values) is first generated based on a limited set of adjustment target patterns (eg, 1D patterns through the pitch). The trial adjustment recipe is generated using a linear or non-linear optimization procedure that resolves the combination of knob offsets (parameter values) and minimizes the cost function that quantifies the deviation from the desired reference. A trial recipe can be used to generate a new lithographic model by the steps described above in relation to Figures 1-8, and this new lithographic model is a full-chip simulation step using the full-chip layout 1100 to which the adjustment recipe is applied. Provided to 1120. Validation step 1160 identifies the simulated contour 122 generated by simulation 1120 by comparing it to the reference contour 1142 generated separately by the simulation step that applies the model for reference on the same full-chip layout 1100. Detect hotspots according to user-defined rules and tolerances. If one or more hotspots were identified in step 1162, in step 1180 the hotspots were added to the set of adjustment target patterns to drive a new round of optimization 1182, from which the updated adjustment lithography model 1184 Will be generated. The updated model 1184 returns to simulation step 1120 and begins a new iteration. Convergence yields the optimal adjustment recipe 1164, which compromises and balances imaging performance for all patterns on a full-chip layout.

0066[0077] In some embodiments, the tuning process involves one or more iterations, with each iteration performing simulation and validation on a reduced pattern set, which is hereinafter referred to as the "mini-layout". Referring to FIG. 12, mini-layout 1206 contains a set of "worm spots" selected from the actual full-chip layout 1200 via simulation and verification step 1202, and step 1202 is a step to include in mini-layout 1206. You can identify critical and frail areas in the layout that can be selected in 1204. Worm spot selection 1204 is typically based on consideration of pattern importance and sensitivity to optical changes. Selection criteria are contour CDs that exceed the lower limit in the nominal or perturbed state (indicating the risk of bridging or necking), the difference between the target CD and the contour CD that exceed the upper limit in the nominal or perturbed state, and the nominal and perturbed states that exceed the upper limit. Difference in contour CD between and (showing the pattern is too sensitive to optical fluctuations), difference in contour CD between the bias mask pattern exceeding the upper limit and the nominal mask pattern (high sensitivity to mask error) (Showing), and spatial or resist image tilt above the lower limit (showing high sensitivity to exposure and other process effects).

0067[0078] The cost function optimized by the adjustment reflects the purpose and criteria of the adjustment. In one embodiment, the cost function comprises a plurality of terms corresponding to a plurality of patterns, each term quantifying the deviation of the achieved contour from a reference contour on one or more patterns. In one embodiment, the terms are summed with the predefined weights to calculate the entire cost function. In some embodiments, the cost function term takes the form of a squared error between the achieved contour and the reference contour or other well-defined error norm. In certain embodiments, the cost function term consists of different forms for different pattern types and includes metric such as minimum to maximum range for a particular pattern type, such as a 1D pattern through a pitch. In some embodiments, the cost function term is asymmetric to the positive and negative errors around the reference CD. For example, if the pattern indicates a potential risk of bridging or necking, erroring the smaller one is not very detrimental to the tuned CD. Therefore, the cost function should make smaller CDs even more disadvantageous than larger CDs. In certain embodiments, the deviations for a particular pattern are constrained, representing the user's emphasis on the imaging performance of such patterns.

0068[0079] The cost function for scanner adjustment is as follows.<maths num="16"><img id="000017" he="12" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Where Δk_jIndicates the knob offset on the target scanner, the subscript j points to the adjustable knob,<maths num="17"><img id="000018" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Indicates the deviation of a particular imaging metric between the achieved contour and the reference contour, the subscript i points to a different pattern in the adjustment target set, α_iIndicates the weight of the cost term for that metric for the relevant pattern. The term "CD" symbolically refers to one or more imaging metrics such as critical dimensions, edge placement, overlay differences, process windows including focus and exposure latitude, and the choice of imaging metrics depends on the pattern. May be good. The norm may include one or more of the sum of squares, ie Euclidean, LP norm, minimum to maximum range, and so on. The norm can be asymmetric with respect to positive and negative differences in imaging metrics. Criteria may be selected as measured wafer contours, simulated wafer contours or design target polygons. The purpose of the scanner adjustment is to minimize this cost function by selecting the knob offset.

0069[0080] Various linear and non-linear optimization techniques and algorithms for calculating knob offsets, including least squares, quadratic design, gradient methods such as Gauss-Newton, Levenberg-Marquardt and BFGS algorithms, and simplex methods. May be used. In general, scanner knobs are subject to mechanical constraints that can be incorporated into optimization criteria.

0070[0081] To illustrate the concept, a linear case is used below as an example. In this case, the above cost function can be described as follows.<maths num="18"><img id="000019" he="15" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Here, the CD vs. knob partial derivative is generated from the sensitivity model. In some embodiments, the norm is Euclidean and the knob offset can be solved by the method of least squares. Further, for scanner matching or process matching when the reference is a contour from a reference scanner or reference process, the above cost function can be described as follows.<maths num="19"><img id="000020" he="29" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>here,<maths num="20"><img id="000021" he="7" wi="159" file="JP5750417B2_D0001.tif" img-format="tif" img-content="drawing" /></maths>Is the CD difference between the untuned scanner and the reference, as predicted by the difference model.

0071[0082] Model-based scanner tuning offers many advantages over traditional methods. One aspect of the invention provides a systematic and cost-effective method for optimizing imaging performance and OPE matching between various lithography systems, including scanners used to image common target patterns. ..

0072[0083] If desired, the effect of adjustment on the pattern can be analyzed using OPC verification tools such as Brion's Tachyon lithography manufacturability check (LMC). This is because the model can quantitatively analyze the effect of model adjustment-related changes on the full-chip pattern. In one example according to one aspect of the invention, a suitable method would be to use an LMC and compare the difference between the two contours with the step of simulating a full-chip on-wafer contour using pre- and post-adjustment models. Analyze the differences between the two models.

0073[0084] With reference to FIG. 9, the computer system 900 may be arranged to assist the model-based process simulation method according to certain embodiments of the present invention. The computer system 900 may include a bus 902 or other communication mechanism for information communication and a processor 904 coupled to the bus 902 for information processing. Computer system 900 may include main memory 906, such as random access memory (RAM) or other suitable dynamic storage device coupled to bus 902 that stores information and instructions executed by processor 902. The main memory 906 can also be used to store temporary variables or other intermediate information during the execution of instructions executed by processor 904. Computer system 900 further includes read-only memory (ROM) 908 or other static storage device coupled to bus 902 that stores static information and instructions for processor 904. A storage device 901, such as a magnetic disk or optical disk, is provided to store information and instructions and is coupled to bus 902.

0074[0085] Computer system 900 is a bus 902 or other to display system 912, such as a cathode ray tube (CRT), flat panel display or touch panel display, configured and adapted to display information to users of computer system 900. May be combined via the connection of. The input device 914, which contains alphanumeric characters and other keys, is coupled to bus 902 to convey information and command selection to processor 904. Another type of user input device may be used, such as a mouse, trackball or cursor direction key to convey direction information and command selection to processor 904 and control cursor movement on the display 912. Includes cursor control 916. This input device typically has two degrees of freedom on two axes, which allows the device to locate in a plane. A touch panel display can also be used as an input device. User inputs and outputs, whether wired or wireless, may be provided remotely using a network.

0075[0086] According to one embodiment of the invention, a part of the scanner tuning process, eg, a simulation operation, responds to processor 904 executing one or more sequences of one or more instructions contained in main memory 906. And may be done by computer system 900. Such instructions can be read into main memory 906 from another computer-readable medium, such as storage device 910. Execution of a sequence of instructions contained in main memory 906 causes processor 904 to perform the process steps described herein. One or more processors in a multiprocessing configuration can also be used to execute a sequence of instructions contained in main memory 106. In alternative embodiments, hard-wired circuits can be used in place of or in combination with software instructions to realize the present invention. Therefore, embodiments of the present invention are not limited to specific combinations of hardware circuits and software.

0076[0087] As used herein, the term "computer-readable medium" refers to any medium involved in providing instructions to processor 904 for execution. Such media take many forms, such as, but not limited to, non-volatile media, volatile media, and transfer media. The non-volatile medium may include, for example, an optical or magnetic disk such as a storage device 910 and may be provided locally to the processor 904 or remotely connected by a network. The non-volatile storage device is of a memory card or stick that can be easily connected to or disconnected from the computer using, for example, a Blu-ray, DVD or CD storage device, or a standard interface including USB, etc. As such, it may be removable from the computing system 904.

0077[0088] The volatile medium includes dynamic memory such as main memory 906. Transfer media include coaxial cable, copper wire and optical fiber, which includes wire with bus 902. The transfer medium can also take the form of sound waves or light waves as produced during radio frequency (RF) and infrared (IR) data communications. Common forms of computer-readable media include, for example, floppy disks, flexible disks, hard disks, magnetic tape, any other magnetic medium, CD-ROM, DVD, Blu-ray, any other optical medium, punch cards, etc. Paper tape, any other physical medium with a pattern of holes, RAM, PROM, and EPROM, FLASH-EPROM, any other memory chip or cartridge, carrier as described below, or computer readable. Any other medium can be mentioned.

0078[0089] Various forms of computer-readable media can be involved in delivering one or more sequences of one or more instructions to processor 904 for execution. For example, the instruction may first be recorded on the magnetic disk of the remote computer. The remote computer can load the instructions into its dynamic memory and use a modem to send the instructions over the telephone line. A modem local to computer system 900 can receive data over the telephone line and use an infrared transmitter to convert the data into an infrared signal. An infrared detector coupled to bus 902 can receive the data carried by the infrared signal and place that data on bus 902. The bus 902 transports data to the main memory 906, from which the processor 904 retrieves and executes instructions. Instructions received by the main memory 906 can optionally be stored in the storage device 910 before or after the processor 904 executes.

0079[0090] Computer system 900 also preferably includes a communication interface 918 coupled to bus 902. Communication interface 918 provides a bidirectional data communication coupling to network link 920 connected to local network 922. For example, the communication interface 918 may be an Integrated Services Digital Network (ISDN) card or modem that provides a data communication connection to the corresponding type of telephone line. As another example, the communication interface 918 may be a local area network (LAN) card that provides a data communication connection to a compatible LAN. Wireless links can also be realized. In any of these embodiments, the communication interface 918 transmits and receives electrical, electromagnetic or optical signals carrying digital data streams representing various types of information.

0080[0091] The network link 920 typically provides data communication to other data devices through one or more networks. For example, network link 920 can provide a connection through a local network 922 to a host computer 924 or to a data device operated by an Internet service provider (ISP) 926. ISP926 provides data communication services through the global packet data communication network, now commonly referred to as the "Internet" 928. Both local networks 922 and Internet 928 use electrical, electromagnetic or optical signals to carry digital data streams. Signals that pass through various networks, and signals that pass through communication interface 918 over network link 920, that carry digital data to and from computer system 900, are exemplary forms of carrier waves that carry information.

0081[0092] Computer system 900 can send messages through (one or more) networks, network links 920, and communication interface 918 and receive data, including program code. In the Internet example, the server 930 can transfer the code requested for the application program through the Internet 928, ISP926, local network 922 and communication interface 918. According to the present invention, one such downloaded application provides, for example, a scanner simulation of an embodiment. The received code can be executed by the processor 904 as received and / or stored in storage device 910 or other non-volatile storage device for later execution. In this way, the computer system 900 can obtain the application code in the form of a carrier wave.

0082FIG. 10 schematically illustrates an example of a lithographic projection apparatus that can benefit from process adjustments provided by certain aspects of the invention. The device is -Providing a projected beam PB of radiation, in this example the radiation systems Ex, IL, which also have a source LA, -A first object table or mask table MT that is provided with a mask holder to hold the mask MA such as a reticle and is connected to the first positioning means that accurately positions the mask with respect to the item PL. -A second object table or board table WT provided with a board holder for holding the board W such as a resist-coated silicon wafer and connected to a second positioning means for accurately positioning the board with respect to the item PL. -With a projection system such as a refractive, reflective or reflective refractive optical system or a "lens" PL that images the illuminated portion of the mask MA onto the target portion C (eg, including one or more dies) of the substrate W. Be prepared.

0083[0094] As illustrated, the device is of a transmissive type with a transmissive mask. Further, the device may be of a reflective type having, for example, a reflective mask. Alternatively, the device can use another type of patterning means instead of using a mask, examples of which include programmable mirror arrays or LCD matrices.

0084[0095] The source LA may be, for example, a mercury lamp or excimer laser, or other device that produces a radiation beam. This beam may be fed directly to the lighting system or illuminator (IL), either directly or after passing through a conditioning instrument such as the beam expander EX. The illuminator IL can be provided with an adjusting means "AM" that sets the outer and / or inner radial range of the beam intensity distribution (referred to as σ-outer and σ-inner, respectively). In addition, the illuminator IL can include various other components such as integrator IN and capacitor CO, and the resulting beam PB can be incident on the mask MA with the desired uniformity and intensity distribution in its cross section. it can.

0085[0096] With respect to FIG. 10, the source LA may be provided within the housing of the lithographic projection apparatus, especially if the source LA includes, for example, a mercury lamp. The source LA may be provided away from the lithographic projection apparatus, and the radiation beam produced by the source LA may be guided into the apparatus by an optical conductor using a suitable induction mirror and / or lens or the like. .. In one example, the source LA, including an excimer laser based on KrF, ArF or F2 lasing, may be located, for example, at some distance from the projector.

0086[0097] In the illustrated example, the beam PB then collides with the mask MA held on the mask table MT. After passing through the mask MA, the beam PB passes through the lens PL that concentrates the beam PB on the target portion C of the substrate W. The second positioning means and / or the interferometer measuring means IF can be used to accurately move the substrate table WT to position, for example, various target portions C in the path of the beam PB. Similarly, the first positioning means can be used to accurately position the mask MA with respect to the path of the beam PB, for example, after mechanically removing the mask MA from the mask library or during scanning. In general, the movement of the object tables MT and WT can be achieved by a long stroke module or coarse motion positioning system and a short stroke module or fine motion positioning system not explicitly shown in FIG. However, in the case of a wafer stepper, the mask table MT may be connected or fixed only to the short stroke actuator.

0087[0098] The illustrated system can be used in different modes. -In step mode, the mask table MT is kept virtually stationary and the entire mask image is projected onto target portion C in one step, a single flash. The substrate table WT is then moved in the x and / or y directions so that another target portion C can be illuminated by the beam PB. -In scan mode, basically the same scenario applies, but given target portion C is not exposed with a single flash. Instead, the mask table MT can move at a velocity v in a given direction, the so-called scanning direction (eg, y direction), so that the projected beam PB scans the mask image, while at the same time the substrate table WT, Simultaneously move in the same or opposite directions at velocity V = Mv, where M is the magnification of the lens PL. Usually M = 1/4 or 1/5. In this way, a relatively large target portion C can be exposed while maintaining the system resolution.

0088[0099] The systems and methods provided by certain aspects of the invention can simulate or mathematically model any common imaging system that images subwavelength features, with increasingly shrinking wavelengths. It is believed that such systems and methods can be advantageously used with emerging imaging techniques capable of producing. The emerging technology already in use is extreme ultraviolet (EUV) lithography that can generate a wavelength of 193 nm using an ArF laser and even a wavelength of 157 nm using a fluorine laser. including. In addition, EUV lithography uses synchrotrons to generate photons in the range of 20 nm to 5 nm, or produces wavelengths in this range by colliding high-energy electrons with solid or plasma materials. be able to. Most materials have absorbency within this range, so reflection mirrors with a multi-stack of molybdenum and silicon can produce illumination. The multi-stack mirror has 40 layers of molybdenum and silicon pairs, each layer having a thickness of 1/4 wavelength. X-ray lithography can generate even smaller wavelengths. A synchrotron is typically used to generate x-ray wavelengths. Most materials are absorbent at x-ray wavelengths, so flakes of absorbent material define where features are printed or not printed, depending on whether positive or negative resists are used, respectively. To do.

0089[00100] The concepts disclosed herein can be used to image on a substrate such as a silicon wafer, while the disclosed concepts are for, for example, to image on a substrate other than a silicon wafer. It should be understood that it can be used in any type of lithography imaging system, such as that used for.

0090<u style="single">Additional Description of Some Aspects of the Invention</u> [00101] Certain embodiments of the present invention provide systems and methods for system level matching of scanners. Some of these embodiments are a step of maintaining a reference model that identifies the sensitivity of the reference scanner with a set of adjustable parameters and a step of generating a difference model for the target scanner. Includes a step that provides a mapping between the reference model and the target model that identifies the sensitivity of the target scanner, and a step of adjusting the target scanner based on the differential model and the reference model.

0091[00102] In some embodiments, a combination of adjustment and calibration information is used during the simulation. In certain embodiments, the adjustment and calibration information is presented as a differential model that characterizes the difference in imaging performance between the selected scanner and the reference scanner, where the reference scanner models the performance of an ideal or typical scanner. To become. In certain embodiments, an ideal scanner is generated. In certain embodiments, the ideal scanner is initially based on the scanner's design requirements. In certain embodiments, the ideal scanner is performed with the specified nominal value.

0092[00103] In certain embodiments, the reference model is modified to reflect the actual performance of one or more scanners. In some embodiments, the observed deviation from the nominal value is added to the model. In one embodiment, the operating environment of the scanner, the type of material used to make the chip and other factors are characterized for the scanner. In certain embodiments, the reference model is tuned based on deviations due to the operating environment.

0093Although the present invention has been described with respect to certain exemplary embodiments, it will be appreciated by those skilled in the art that various modifications and modifications may be made to these embodiments without departing from the broad spirit and scope of the invention. It will be clear. Therefore, the specification and drawings should be considered as exemplary rather than in a restrictive sense.