KR20070019690A - Method to determine the value of process parameters based on scatterometry data - Google Patents

Abstract

The method according to one embodiment includes obtaining calibration measurement data using an optical detection device from a plurality of sets of marker structures provided on a calibration substrate. Each marker structure set includes one or more calibration marker structures made using different known values of the processing parameters. The method includes obtaining measurement data using an optical detection device from one or more marker structures provided on a substrate and exposed using unknown values of processing parameters; And determining an unknown value of the processing parameter from the obtained measurement data by applying regression coefficients to the model based on the known value of the processing parameter and the calibration measurement data.

Description

METHODS TO DETERMINE THE VALUE OF PROCESS PARAMETERS BASED ON SCATTEROMETRY DATA}

This application claims the benefit of US Provisional Patent Application No. 60 / 546,165, filed February 23, 2004, which is hereby incorporated by reference in its entirety.

FIELD The present invention relates to a lithographic apparatus and method.

BACKGROUND A lithographic apparatus is a machine that applies a desired pattern onto a target portion of a substrate. Lithographic apparatus can be used, for example, in the manufacture of integrated circuits (ICs). In that situation, a patterning structure, such as a mask, can be used to create a corresponding circuit pattern on an individual layer of the IC, which pattern can be used (eg, on one or several dies) on a substrate (eg, silicon wafer). Can be imaged onto a target portion (including a portion). In general, a single substrate will contain a network of adjacent target portions that are subsequently exposed. Known lithographic apparatus scans a pattern in a given direction ("scanning" -direction) through a so-called stepper, and a radiation beam, in which each target portion is irradiated by exposing the entire pattern on the target portion at one time, while A so-called scanner, in which each target portion is irradiated by synchronously scanning the substrate in parallel or anti-parallel to the direction.

Although reference is made herein to specific uses of lithographic apparatus in the manufacture of ICs, the lithographic apparatus described herein includes integrated optical systems, induction and detection patterns for magnetic domain memories, flat panel displays, liquid crystal displays (LCDs), It should be understood that other applications may be present, such as the manufacture of thin film magnetic heads and the like. As those skilled in the art relate to these alternative applications, any use of terms such as "wafer" or "die" herein may be considered synonymous with more general terms such as "substrate" or "target portion", respectively. I will understand that. The substrate referred to herein may be processed before or after exposure, for example in a track (a tool that typically applies a layer of resist to a substrate and develops the exposed resist), a metrology tool, and / or an inspection tool. Where applicable, the disclosure herein may be applied to such and other substrate processing tools. Further, as the substrate may be processed more than once, for example to produce a multilayer IC, the term substrate as used herein may also refer to a substrate that already contains multiple processed layers.

As used herein, the terms “radiation” and “beam” refer to ultraviolet (UV) radiation (eg, wavelengths of about 365, 248, 193, 157, or 126 nm) and (eg, wavelengths of 5 to 20 nm). Encompasses all types of electromagnetic radiation, including extreme ultraviolet (EUV) radiation, as well as particle beams such as ion beams and electron beams.

As used herein, the term “patterning structure” broadly refers to a structure that can be used to impart a predetermined pattern to a cross section of a radiation beam (eg, a projection beam) to create a pattern in a target portion of a substrate. Should be interpreted. It should be noted that the pattern imparted to the beam may not exactly correspond to the desired pattern in the target portion of the substrate. In general, the pattern imparted to the beam will correspond to a particular functional layer in the device to be created in the target portion, such as an integrated circuit.

The patterning structure can be transmissive or reflective. Examples of patterning structures include masks, programmable mirror arrays, and programmable LCD panels. Masks are well known in the lithography art and include mask forms such as binary, alternating phase-shift, and attenuated phase-shift and various hybrid mask forms. One example of a programmable mirror array employs a matrix arrangement of small mirrors, each of which can be individually tilted to reflect the incident radiation beam in a different direction; In this way the reflected beam is patterned.

The support structure supports the patterning structure. That is, bears its weight. This maintains the patterning structure in a manner that depends on the orientation of the patterning structure, the design of the lithographic apparatus, and other conditions, such as for example whether or not the patterning structure is maintained in a vacuum environment. The support may use mechanical clamping, vacuum or other clamping techniques, for example electrostatic clamping under vacuum conditions. The support structure can be a frame or a table that can be fixed or moved as desired and can ensure that the patterning structure is in a desired position relative to the projection system, for example. The use of any term "reticle" or "mask" herein may be considered as synonymous with the more general term "patterning structure".

The term "projection system" as used herein, if appropriate for the exposure radiation used, or for other factors such as the use of immersion liquid or the use of vacuum, for example, refractive optical systems, reflective optical systems, And various types of projection systems, including catadioptric optical systems. Any use of the term "lens" herein may be considered as synonymous with the more general term "projection system."

The illumination system can also encompass various forms of optical components, including refractive, reflective, and catadioptric optical components, configured to direct, shape, or control the projection beam of radiation, which components are also collectively described below. Or alone as a "lens".

The lithographic apparatus may be of a type having two (dual stage) or more substrate tables (and / or two or more mask tables). In such "multiple stage" machines additional tables may be used in parallel, or preparatory steps may be carried out on one or more tables while one or more tables are being used for exposure.

The lithographic apparatus may also be of a type in which the substrate is immersed in a liquid having a relatively high refractive index such as, for example, water to fill the space between the final element of the projection system and the substrate. Immersion liquid may be applied to other spaces in the lithographic apparatus between the mask and the first element of the projection system. Immersion techniques can be used to increase the numerical aperture of the projection system.

Higher-resolution lithography is required due to the continuing trend toward smaller design features and higher device densities. In order to meet this requirement, it may be desirable to control the lithographic process in as many details as possible. Two of the most important process parameters that may require accurate monitoring and control are dose and focus. In general, critical dimension (CD) variations are measured to monitor and control these parameters. However, it can be difficult to distinguish between dose and focus data when measuring CD-variability.

Typically, special or multiple features are used in combination with special or time consuming metrology. The focus can be determined, for example, by a phase shift focus monitor. The focus error can be an overlay error that can be easily detected with an overlay readout tool. In a second technique, the monitoring of focus is performed using the concept of line-end shortening. However, in this technique, the signal of defocus can be extremely difficult to determine. In addition, the most recent techniques are only applicable on test structures.

The need to monitor the quality of the pattern exposed by the lithographic apparatus requires a fast and reliable technique, which is, for example, a chip area or scribe line on all kinds of substrates to be exposed, such as test or product wafers. It can be used at many locations within. Optical metrology techniques called scatterometry can meet these requirements to a certain extent. As used herein, the terms “optical” and “light” refer to a wavelength of 400-1500 nm, ultraviolet (UV) radiation (eg, having a wavelength of about 365, 248, 193, 157, or 126 nm). And all types of electromagnetic radiation, including extreme ultraviolet (EUV) radiation (eg, wavelengths in the range of 5-20 nm), as well as particle beams such as ion beams and electron beams.

In scatterometry, the light beam is directed to a target, typically a specially designed structure such as a diffraction grating. The target then reflects, refracts, and / or diffracts light. Finally, light from the target can be detected by a detector comprising a suitable sensor. Detection by the detector can be made while measuring diffracted and / or non-diffracted light upon reflection or transmission. In the case of incident light, ie light directed at a target, more than one set of properties can be changed at the same time. The terms "scatterometer" and "scatterometer" as used herein encompass all types of measurement techniques and tools in which light is generated and analyzed after interaction with a target. Thus, as used herein, “scatterometer” includes ellipsometers and scanning electron microscopy (SEM). The term "spectrum" as used herein encompasses all types of formats in which light after interaction with a target can be detected. Thus, this includes the image produced by the scattered electrons in the SEM.

Scattermetry is typically used to determine values of processing parameters such as focus and dose. In general, however, some assumptions are made regarding the relationship between processing parameters and scatterometry measurement parameters. Examples of such hypothesized relationships include linear relationship between focus and side wall angle (slope of the side of the linear structure), dose and mid-CD (width at half the height of the linear structure). There is a linear relationship between Indeed, there may be no unique relationship between one single scatterometry measurement parameter and processing parameters such as focus or dose. There may be additional effects other than focus, for example, which contribute to the properties of the sidewall angles. By the assumptions above, these effects will be interpreted abusively as focus.

The detected spectrum (or, if particle beams are used, the detected signal can be an image rather than a spectrum) is analyzed by comparing it with the data stored in the library. The so-called "best match" between the detected spectrum and the spectrum in the library determines the parameter values that best represent the target structure. For lithographic purposes, the identified parameter values, ie focus and dose, can be applied to increase the performance of the lithographic apparatus. The quality of lithographic processing parameter control and monitoring can be highly dependent on the quality of the library. The library is generally used for different scatterometry measurement parameters, such as grating parameters such as grating height, line width and sidewall angle, and for different substrate parameters such as material properties and properties related to the layers in the substrate previously processed. It is filled with theoretical spectra constructed by calculating the values. In particular, it will be readily understood that, in the case where the properties of the substrate to be exposed are regularly changed, making an extremely reliable library can be time consuming and very complex.

In addition, scatterometry measurement parameters such as the thickness of the underlying layers and the optical constants of the materials used can be extremely difficult to determine in the state of manufacture. The use of empirical data, ie experimentally obtained data, has been proposed (eg Allgair et al., Yield Management Solutions, Summer 2002, pp 8-13). In this case, the empirical library is made from a substrate using a number of structures processed by a changing set of processing parameters covering a controlled processing space. However, as mentioned in this reference, the characterization of these structures is based on the required level of control on the processing parameters and of the noise introduced by the "natural variation" (ie, not carefully induced variations). Due to the significant effect, it is not simple (non-trivial).

Embodiments of the present invention include a method of determining one or more processing parameters associated with a lithographic method using empirical data. One embodiment provides a method of determining one or more processing parameters, the method comprising:

Obtaining calibration measurement data from a plurality of calibration marker structure sets provided on a calibration object, wherein each of the plurality of calibration marker structure sets comprises one or more calibration marker structures, and calibration of different calibration marker structure sets. Marker structures are generated using different known values of one or more processing parameters;

Determining a mathematical model comprising a plurality of regression coefficients by using the known values of the one or more processing parameters and by using a regression technique on the calibration measurement data;

Obtaining measurement data from one or more marker structures made using the unknown values of the one or more processing parameters provided on an object; And

Determining an unknown value of the one or more processing parameters for the object from the obtained measurement data by using the regression coefficients of the mathematical model.

In another embodiment of the present invention, a system for determining one or more processing parameters is provided, the system comprising:

A detector arranged to obtain calibration measurement data from a plurality of calibration marker structure sets provided on a calibration object, the plurality of calibration marker structure sets each including one or more calibration marker structures, and calibration markers of different calibration marker structure sets Structures are obtained using different known values of the one or more processing parameters;

A processor unit that uses the known values of the one or more processing parameters and stores a mathematical model including a plurality of regression coefficients, determined by using a regression technique on the calibration measurement data;

The processor unit is further configured to obtain measurement data from one or more marker structures made using an unknown value of the one or more processing parameters provided on an object; And determine the unknown value of the one or more processing parameters for the object from the obtained measurement data by using the regression coefficients of the mathematical model.

In one embodiment of the invention, the system comprises an illumination system configured to provide a radiation projection beam; A support structure configured to support a patterning structure that serves to impart a pattern to a cross section of the radiation beam; A substrate table configured to hold a substrate; And a projection system configured to project the patterned beam onto a target portion of the substrate.

In one embodiment of the present invention, a semiconductor device manufactured by the method of the present invention according to any of the embodiments disclosed herein is provided.

In one embodiment, the system comprises an illumination system configured to provide a radiation beam; A support structure configured to support a patterning structure that serves to impart a pattern to a cross section of the radiation beam; A substrate table configured to hold a substrate having one or more marker structures; And a projection system configured to project the patterned beam onto a target portion of the substrate.

The invention also relates to a semiconductor device fabricated using the system according to any of the above described embodiments.

DESCRIPTION OF THE EMBODIMENTS Hereinafter, embodiments of the present invention will be described, by way of example only, with reference to the accompanying schematic drawings in which corresponding reference numbers indicate corresponding parts.

1 shows a lithographic apparatus according to an embodiment of the invention;

2 shows a scatterometer of the highest technology level of this technique;

3 illustrates the functional flow of a library-based method;

4 shows a functional flow of a library based method using measured calibration spectra;

5A and 5B show functional block diagrams representing two phases in accordance with one embodiment of the present invention;

6 shows a functional block diagram of the concept of regression in accordance with one embodiment of the present invention;

7A and 7B illustrate the concept of partitioning into harmonics and the concept of partitioning into principal components, in accordance with an embodiment of the present invention;

8A and 8B show top views of different combinations of marker structures in accordance with one embodiment of the present invention.

9 illustrates a lithographic system according to an embodiment of the present invention.

1 schematically depicts a lithographic apparatus according to an embodiment of the invention. The apparatus comprises: an illumination system (illuminator) IL configured to provide a radiation projection beam PB (eg, radiation having UV radiation or other wavelengths); And a first support structure configured to support the patterning structure (e.g., mask) MA and connected to the first positioner PM configured to accurately position the patterning structure with respect to the projection system, item PL. (Eg, a mask table) MT. The apparatus is also configured to hold a substrate (eg, a resist-coated wafer) W, and a second positioner configured to accurately position the substrate relative to the projection system ("lens"), PL. A substrate table (e.g., wafer table) WT coupled to PW, and the projection system (e.g., refractive projection lens) ("lens") PL is a target portion (e.g. A pattern imparted to the radiation beam B by the patterning structure MA on, for example, one or more dies).

As shown herein, the device is of a transmissive type (e.g. employing a transmissive mask). Optionally, the apparatus may be reflective (e.g., employing a programmable mirror array of a type as mentioned above).

The illuminator IL receives the radiation beam from the radiation source SO. For example, when the source is an excimer laser, the source and the lithographic apparatus may be separate entities. In this case, the source is not considered to form part of the lithographic apparatus, and the radiation beam is, for example, with the aid of a beam delivery system BD comprising a suitable directing mirror and / or beam expander. ) Is passed through to the illuminator IL. In other cases, for example if the source is a mercury lamp, the source may be an integral part of a lithographic apparatus. The source SO and the illuminator IL may be referred to as a radiation system together with the beam delivery system BD if necessary.

The illuminator IL may comprise an adjuster AM configured to adjust the angular intensity distribution of the beam. In general, at least the outer and / or inner radial extent (commonly referred to as sigma-outer and sigma-inner, respectively) of the intensity distribution in the pupil plane of the illuminator can be adjusted. In addition, the illuminator IL may include various other components such as the integrator IN and the condenser CO. The illuminator provides a beam of conditioned radiation, called projection beam PB, with a desired uniformity and intensity distribution in its cross section.

The projection beam PB is incident on the mask MA held on the mask table MT. When the mask MA is crossed, the projection beam PB passes through the lens PL to focus the beam on the target portion C of the substrate W. With the aid of the second positioner PW and the position sensor IF, for example an interferometer device, the substrate table WT positions different target portions C in the path of the beam PB, for example. Can be precisely moved. Similarly, the first positioner PM and another position sensor (not clearly shown in FIG. 1) may be used for the beam B, for example after mechanical retrieval from the mask library or during scanning. It can be used to accurately position the mask MA with respect to the path. In general, the movement of the object tables MT and WT is realized with the help of a long-stroke module (coarse positioning) and a short-stroke module (fine positioning). Will form part of the positioners PM, PW. However, in the case of a stepper (as opposed to a scanner), the mask table MT may only be connected or fixed to a single-stroke actuator. The mask MA and the substrate W may be aligned using the mask alignment marks M1 and M2 and the substrate alignment marks P1 and P2.

The apparatus described can be used in the following preferred modes:

1. In the step mode, the mask table MT and the substrate table WT are basically kept stationary, while the entire pattern imparted to the projection beam is projected onto the target portion C at once (ie, a single static Single static exposure). Subsequently, the substrate table WT is shifted in the X and / or Y direction so that another target portion C can be exposed. In the step mode, the maximum size of the exposure field limits the size of the target portion C imaged during a single static exposure.

2. In the scan mode, the mask table MT and the substrate table WT are synchronously scanned (ie, single dynamic exposure while the pattern imparted to the projection beam is projected onto the target portion C). ))do. The velocity and direction of the substrate table WT relative to the mask table MT is determined by the (de-) magnification and image reversal characteristics of the projection system PL. In the scan mode, the maximum size of the exposure field limits the width (in the unscanned direction) of the target portion during a single dynamic exposure, while the length of the scanning operation determines the height (in the scanning direction) of the target portion.

3. In another mode, the mask table MT remains essentially stationary by holding a programmable patterning device and while the pattern imparted to the radiation beam is projected onto the target portion C, the substrate table ( WT) is moved or scanned. In this mode, a pulsed radiation source is generally employed, and the programmable patterning structure is updated as needed after each movement of the substrate table WT, or between successive radiation pulses during the scan. do. This mode of operation can be readily applied to maskless lithography using programmable patterning means, such as a programmable mirror array of immersion as mentioned above.

It is also possible to adopt combinations and / or variations of the modes described above, or completely different modes of use.

2 shows a scatterometer of the highest technical level. The scatterometer includes a structure 5 on an exposed substrate W lying on a substrate table WT, a light source 1 that directs the light beam 2 to a generally several types of gratings, and a detector ( 4). The detector 4 is connected to the (micro) processor 9, which is connected to the memory 10. The light beam 2 is reflected and / or diffracted in a suitable structure 5 located on the surface of the substrate W. The light beam 2 may be directed at an angle to the substrate W as shown in FIG. 2, but may also be directed perpendicularly to the substrate W. There are several scatterometry concepts in which more than one set of characteristics of light directed at a suitable structure can be changed at the same time. A set of characteristics is, for example, a set of wavelengths, a set of incident angles, a set of polarization states or a set of phases and / or phase differences. The detector may be arranged to detect one or a combination of the above-described sets, and may include one or more sensors for recording different portions of reflected and / or diffracted light.

3 shows a functional overview using a scatterometry based library method. The library generally includes different scatterometry measurement parameters, such as structure parameters such as line width, line height, sidewall angle, etc. of the lines of the structure 5, of the underlying (non-patterned) layers below these lines. It can be constructed by calculating the spectrum for the thickness and the optical constants of all materials that interact with the light beam 2. Before the measurement on the actual physical structure 5, the above-mentioned parameters relating to this particular kind of structure 5 may need to be defined. For each of these defined parameters in a certain range, the spectrum of light adjusted by the structure 5 is calculated and stored by the processor 9 in a spectrum library of the memory 10 at task 301. do.

Then, as will be appreciated by those skilled in the art, theoretical calculations may be performed on known structures. For example, if the library is filled by the processor 9 with a spectrum sufficient to cover the desired area of the spectral performance of the actual structure 5 to be measured, the measurement is performed on the actual structure 5. The method then proceeds to task 302, where the measured spectra of the actual structure 5 are compared by the processor 9 with a number of stored spectra in the spectral library of the memory 10. Optionally, real time fitting may also be applied.

Next, using an interpolation algorithm, a 'best match' is extracted from the memory 10 in task 303 and corresponds to the parameters that were used to generate the extracted spectrum. The values of the parameters are checked. For example, if the measured spectrum is most similar to the spectrum constructed using the value A1 for parameter A and the value B3 for parameter B, then the processor 9 will eventually output {A1, B3}.

Rigorous diffraction modeling algorithms such as Rigorous Coupled Wave Analysis (RCWA) can be used to calculate the spectra of the spectral library. This complex algorithm used to calculate the spectra stored in the spectral library of memory 10 may first of all require knowledge of the optical properties of the materials used. Indeed, especially for product wipers, only values of some of these properties are known, so approximations are generally used. In addition, in the state of manufacture, the properties of the different structures in the foundation layers are insufficiently known. As a result, methods based on conventional libraries can be complex, which can limit daily use in manufacturing conditions.

In the description below, reference will be made to dose and focus as exemplary processing parameters. However, it should be understood that embodiments of the present invention may be applied in a similar manner when other lithographic processing parameters are used. Other examples of processing parameters that may be used include, for example, track parameters related to dose, variations in line width on the reticle, variations from reticle-to-reticles, projection lens aberrations ), The projection lens flare, and the angular distribution of the reticle illumination light.

4 shows a functional description of a method based on a library of scatterometry, in accordance with an embodiment of the present invention. In this way, the measured calibration spectrum is used directly instead of the theoretical spectral data and compared with the spectrum measured on the actual physical structure, which can be a diffractive structure such as, for example, a grating. In practice, calibration is performed on a calibration substrate prior to making measurements of the physical structure.

 In one embodiment of the invention, a number of calibration structures are provided on a calibration substrate, the calibration structures each having a substantially comparable form as the physical structure being measured. The calibration structures can each have a unique location on the calibration substrate, which is made from a combination of unique values of processing parameters such as focus and exposure (dose). In one embodiment of the invention, the value of the first processing parameter is varied in the first direction across the substrate, while the second processing parameter is changed in the second direction that is substantially perpendicular to the first direction. In one embodiment of the invention, the first and second processing parameters are focus and dose. In this case, the calibration substrate is called a focus-exposure matrix (FEM). In the following description, reference will be made to FEMs to illustrate the concepts of embodiments of the invention. However, it will be appreciated that alternative matrices may be used in other embodiments of the invention.

In one embodiment of the invention, the method begins at task 401, where the calibration spectrum is measured by FEM and then in memory 10 with information about the values of focus and dose that were used to make the spectra. Stored. Next, the spectrum of light acting on the actual physical structure is measured. This measured spectrum is then compared in operation 402 with the spectrum stored in memory 10. The method then proceeds to task 403 where the 'best match' is extracted from memory 10. In this step, the values of dose and focus are derived from the extracted spectrum. For example, in FIG. 4, the 'best harmony' between the measured spectrum and the spectra measured on the structure provided by the FEM corresponds to the value for focus (F2) and the value for exposure (dose) (E2). Is determined to be a spectrum.

It will be appreciated that a potential advantage of at least some embodiments as shown in FIG. 4 is that no prior knowledge of the optical properties of the materials is required to determine the parameters.

However, as in any library based methods as described above, the determined values of the selected processing parameters are discretized. In addition, the noise introduced in the calibration operation by 'natural variation' (ie not carefully induced variations) can have a significant impact on the confirmation of the values of the selected processing parameters. It may be desirable to minimize interference with identification due to these natural variations.

Sources of natural variability may include: In a scanner, natural variation can be associated with random focus and exposure dose errors that will differ from its unique focus & dose setting for each individual exposure. In the track, natural variations can be associated with non-uniform processing across the wafer (they are partly dose related). In a wafer, natural variations can be associated with non-uniform foundation layers across the wafer. In scatterometers, natural variations can be associated with thermal, mechanical and electrical noise.

5A and 5B show functional block diagrams of one embodiment of the present invention. In this embodiment, calibration spectra are used to form a mathematical model by using a regression technique on the calibration phase. At this time, the processing parameters used to make the actual structure to be measured can be derived by using a mathematical model obtained in the operating phase. 5A shows the method used in the calibration phase, in this one embodiment of the invention. The method begins at task 501, where the calibration spectrum is measured using a number of calibration structures and stored in memory 10. These calibration structures consist of a known set of processing parameters, which are different for each calibration structure. For example, if the processing parameters are focus and dose, the method first measures the calibration structures using the FEM and stores the measured spectra in memory 10.

The method then proceeds to task 502 where regression analysis is performed on the stored calibration spectrum using a processor coupled to memory 10. This processor may be the processor 9 of one embodiment of the present invention, and may be a different processor in another embodiment of the present invention. The method then proceeds to task 503, where a mathematical model stored in memory is determined. The mathematical model defines the relationship between calibration spectra and processing parameters used to make a calibration structure in which the calibration spectra are measured. The memory may be memory 10 in one embodiment of the invention, but in another embodiment of the invention, it may be a different memory coupled to the processor.

FIG. 5B illustrates an embodiment of the present invention that can be performed by the processor 9 to derive values of selected processing parameters from measurements performed on a “real” structure on a substrate using the model obtained above. The method according to this is shown. The method begins at operation 511 and the reaction signal is measured on the “real” structure on the substrate. The measured signal, which can be a spectrum, serves as an input to the model. The method then proceeds to task 512, where desired values of the selected processing parameters are determined. The method then proceeds to task 513, wherein the determined processing parameters are for example of a lithographic apparatus such as dose settings, focus settings, positioning settings (e.g., movement of the substrate table WT), and the like. It is used in lithographic processing either manually or automatically to modify external settings.

It will be appreciated that the action of natural variation can be minimized in embodiments of the present invention. Since the natural variation of the selected processing parameter is included in the calibration, the generated model can be independent of the natural variation of this processing parameter. In order to further minimize the effects of natural variation, it may be desirable to use random variation (eg, a calibration wafer may be made to do this). In addition, if natural variations in processing parameters are known, they can be used as individual inputs in the formation of a model in the calibration phase. Here, "individual" means additional input, or that it can replace carefully derived processing offsets.

The regression technique used in the regression method may be linear or non-linear in one embodiment of the present invention. In one embodiment of the invention neural networks may also be used. This technique can be applied to provide interpolation, ie interpolation between calibration points of the model, and / or noise reduction.

A functional block diagram of a regression technique according to another embodiment of the present invention is shown in FIG. This concept is based on iterative processing and calculates the regression coefficients b (which combine X and Y) using the specified response signal (X) and a set of predictor parameters (Y). To form a mathematical model. Prediction parameters Y are parameters related to the processing parameters under inspection. The method starts at operation 601, where a set of prediction parameters Y are provided, and then proceeds to operation 603, where the reaction signal X of the structure on the substrate is measured. The prediction parameters Y and the measured response signal X both serve as inputs to the mathematical model, which calculates the regression coefficients b that are modeled in operation 605. Subsequently, in task 607, the significance of all of the regression coefficients b is confirmed. This control task determines whether the mathematical model is robust or not. Insignificant regression coefficients are removed from the mathematical model in task 609 and the regression is repeated using a reduced number of regression coefficients. Operations 605 and 607 are repeated until all of the regression coefficients in the mathematical model are large. The method then proceeds to task 611, where the regression results are used to determine prediction parameters Y for new response signals X.

In one embodiment of the invention, linear regression (MLR) may be used to convert the data into information. Appropriate states develop when there are few response signals, sometimes referred to as factors. If the factors are not redundant, that is, they are collinear, or if they have a well understood relationship with the prediction parameters Y, the MLR can be very useful. However, if any of these three conditions are not met, the MLR may be ineffective or inadequate. Embodiments of the present invention include methods in which the MLR is applied based on the presence of one or more such conditions.

In one embodiment of the invention, the spectrum measured by scatterometry is used to estimate the values of lithographic processing parameters such as dose and focus. In general, the factors comprising spectra reach hundreds and are highly collinear. The prediction parameters Y are in this case the values of the lithographic processing parameters.

7A, 7B show examples of decomposition techniques that may be used in embodiments of the present invention. The first technique shown in FIG. 7A uses Fourier analysis, which is based on the principle that a signal can be represented as the sum of basic harmonic functions, with each function contributing with a particular weight. For example, the signals S1, S2, S3 and S4 of FIG. 7A are weighted [1, -1], [1, -1 / 2], [1, +1/2] and [1,1], respectively. And H1 and H2.

The second technique is a similar technique based on the principle that a signal can be described as the sum of a plurality of intrinsic components, in which the intrinsic components each contribute with a certain weight. The number of intrinsic components can vary considerably. FIG. 7B illustrates by combining the intrinsic components PC1, PC2 with weights [1, -1], [1, -1 / 2], [1, +1/2] and [1,1], respectively. Four example scatterometry spectra F1, F2, F3 and F4 are shown. Decomposition techniques, such as those described above, but not limited to these examples, can be used in embodiments of the present invention for regression analysis. For example, in the case of principal component regression (PCR), the intrinsic components extracted from the measured response signal (X) will act as input to the mathematical model instead of the X-factors as shown in FIG. Can be.

In addition to the two illustrated decomposition techniques, other decomposition techniques may also be used in embodiments of the present invention. Examples of these techniques include, but are not limited to, partial least squares (PLS) modeling and non- as described, for example, in Wold et al. Chemometrics and Intelligent Laboratory Systems, 7 (1989) 53-65. Decomposition techniques based on the concept of linear PLS modeling are included.

In an embodiment of the invention, several kinds of pretreatment may be applied before the spectral data is supplied to the model. Preprocessing can improve the results of the model. Examples of pretreatment operations applicable to the present invention include subtraction of averages, division by standard deviation, and weighting or selection of scatterometry variables such as angle, wavelength and polarization state. As a result, data at a particular wavelength can be removed before data is supplied to the model.

In embodiments of the invention, more than one type of marker structure may be used for both calibration and measurement processing. Thus, the present invention is applicable to multiple calibration structure sets, each set including one or more (different) calibration structures. Thus, each set may comprise one or more calibration structures, and it is possible for the number of calibration structures per set of calibration structures to vary. In addition, the calibration structures in and / or between the calibration structure sets may be different. Within each set, it may be desirable for different types of marker structures to be positioned in proximity to each other on the calibration substrate. It may also be desirable for calibration measurements and sample measurements to be substantially the same in at least some aspects (eg, the same pretreatment, the same marker or combination of markers, and / or the same wafer type). The spectra obtained on these markers can be appended to each other before they are used by the mathematical model. However, in one embodiment of the present invention, it may also be possible to combine these spectra by several kinds of mathematical operations to obtain one combined "spectrum" used by the model.

8A and 8B show front views of a calibration substrate 801 provided in combination with marker structures, which may be used in one embodiment of the present invention. In FIG. 8A, the first marker structure 802 includes a pattern formed on top of a plurality of non-patterned layers. The second marker structure 803 does not include a pattern and is formed only by non-patterned layers. In FIG. 8A, only one set with first and second marker structures 802, 803 is shown. However, in order to practice the calibration method of the present invention, several such sets are fabricated on the same calibration substrate 801 using different processing parameters for different sets. In scatterometry measurements, the second marker structure 803 only reflects variations in the non-patterned layer, while the pattern of the first marker structure 802 contributes to the contributions of these non-patterned layers. Add The scatterometry measurement results obtained on the second marker structure 803 can now be used to reduce the non-patterned layer contribution in the scatterometry measurement results obtained on the first marker structure 802. Examples of operations that can perform this reduction include subtracting the spectrum of the second marker structure 803 from the spectrum of the first marker structure 802, and subtracting the spectrum of the second marker structure 803 from the first marker structure ( After fitting to the spectrum of 802, using the remainder as input to the mathematical model.

In FIG. 8B, different combinations of a set of two marker structures on a substrate 801 are shown, according to one embodiment of the invention. 8B shows one set, but many such sets will be fabricated on the substrate using different processing parameters for different sets to perform the method of the present invention. The first marker structure 802 includes the same pattern as the first marker structure 802 in FIG. 8A, for example. However, unlike the second marker structure 803 of FIG. 8A, the second marker structure 804 shown in FIG. 8B is patterned. In this embodiment of the present invention, both marker structures 802 and 804 are patterned, but the patterns of each marker structure are different. Since marker structures will have different sensitivity to lithographic processing parameters, it is possible to better separate processing parameters. It will be appreciated that other combinations of patterned marker structures can be used in other embodiments of the present invention. In one embodiment of the invention, two or more marker structures may be used.

If the focus is one of the processing parameters measured using one of the embodiments disclosed herein, further optimization may be possible using one of the following techniques. In one embodiment of the invention, smaller marker structures can be used to make large changes in the spectral form per nanometer of defocus, since these structures have a smaller depth of focus. In another embodiment of the present invention, structures with more sidewalls, for example semi-independent contact holes or semi-independent dots, may be used instead of lines to increase sensitivity to focus changes. In another embodiment of the present invention, it may be possible to use a resist that exhibits a larger spectral change with defocus. However, in the manufacturing state, such an option may not be applicable.

It should be understood that any type of substrate, for example product wafers or test wafers, can be used in the applications of embodiments of the present invention. In addition, it should be understood that the actual physical structure to be measured can be located somewhere on the substrate, for example in a chip region or in a scribe-lane, as may be desirable in such applications. Further, in embodiments of the present invention, the light spot of the scatterometer may be as large as the chip area or the exposure field. Spots of this size can enable fast determination of offsets per chip and field, respectively.

It may also be desirable to arrange lithographic processing parameters such as focus and dose across the FEM in a shuffled manner. Otherwise, processing parameters can be increased from one side of the substrate to the other. As a result, process variations originating from sources outside the lithographic apparatus, which are generally linearly traversed and / or symmetric about the center of the substrate, could have a significant impact on the calibration results. By shuffling the values across the FEM, for example, externally derived processing parameters can be significantly removed.

In one embodiment of the invention, for the purpose of limitation, mini-FEMs can be used to eliminate externally induced process variations. This FEM only covers a small part of the substrate. Thus, externally induced process variations are considered negligible.

In manufacturing processes, many of the same wafers are generally processed one after the other. Once the optimal settings of the lithographic apparatus for a particular lithographic manufacturing process have been determined, these settings must be maintained within tight control limits. These settings may be maintained by automated process control (APC) in one embodiment of the present invention. In this case, regular measurements are performed on fabricated wafers, allowing feedback control.

The invention can be used separately in track and lithographic apparatus. The invention can be used for twist knobs on a track or lithographic apparatus that are not directly related to the controlling process parameter. By then measuring the corresponding processing parameters and using the present invention, the effects of such knob twisting can be deduced and the optimal setting of the knobs can be selected. In contrast to the past, using off-line measurements to obtain desired information using techniques such as scanning electron microscopy (SEM) and electrical line width measurement (FLM) can be avoided.

9 shows a lithographic system according to an embodiment of the present invention. In this embodiment, the substrate exposed with the lithographic apparatus 901 is moved to the scatterometer 902 (after development by the track). The scatterometer 902 is connected to a control unit 903 that includes a processor 9 and a memory 10. The lithographic apparatus 901 first prints a marker structure, using predetermined settings for processing parameters focus and dose, to produce a FEM suitable for scatterometry measurements. The substrate is then transported 910 to the scatterometer 902. The scatterometer 902 measures the calibration spectra and stores the measured spectra in the calibration library 904 of the memory 10 (911).

The lithographic apparatus 901 then patterns the fabrication substrate into the same marker structure. The substrate is then transported 912 to the scatterometer 902. The scatterometer 902 measures the spectrum of light reflected from the marker structure produced by the lithographic apparatus 901. The spectrum is supplied 913 in a mathematical model 905 that can be used by the processor 9. The mathematical model 905 is used by the processor 9 to compare the calibration spectra stored in the calibration library 904 with the spectra measured on the marker structure, the processor 9 using regression techniques such as dose and focus. Derive the values of the controlled parameters.

Finally, the processor 9 supplies the derived values of these parameters to the lithographic apparatus 901. The lithographic apparatus 901 may use the derived values, for example, to monitor drifts in the apparatus with respect to a reference state. The derived values are then used in the feedback signals to correct for these movements. In this case, the lithographic apparatus 901 is provided with a correction control unit, which uses the applied correction signals to offset the movement. The correction control unit 903 may be configured to control the height of the substrate table WT, for example, to improve focus.

In an alternative embodiment of the invention, the derived values of these parameters are not supplied to the lithographic apparatus 901 but to different bodies such as tracks, computer terminals or displays. When supplied to a different body, the operator operating the lithographic apparatus 901 can then confirm, for example, whether the derived values fall within the control limits. In another embodiment of the invention, mathematical model 905 and / or calibration library 904 may be located in a different body than control unit 903. In one embodiment of the invention, both lithographic apparatus 901 and scatterometer 902 can be connected to the same track in order to efficiently control the parameters of lithographic apparatus 901. The derived values may also be used in a feed forward signal, which allows for optimization of the settings of the next processing task. The derived values can be sent to an etching apparatus, for example, which can adapt their settings to the substrate to arrive.

Modifiable actions for focus include, for example, a change in tilt in the exposure field, an offset across the wafer and a change in the offset from the wafer to the wafer. Modifiable actions on dose include, for example, changes in tilt and / or curvature in the exposure field, offset across the wafer and change in offset from wafer to wafer.

According to one embodiment of the method, the spectrum is used directly, without the need for complicated calculations and knowledge of the properties of the substrate, to determine the values of one or more processing parameters. In addition, the regression technique used by the mathematical model reduces the noise contribution to the desired data in the process of extracting that information from the spectra. The optical detection device may be a scatterometer. The scatterometer is configured to measure the spectrum in a fast and reliable manner and can be used at many locations on all kinds of substrates to be exposed.

According to one embodiment of the invention, the measurement may be performed on a device pattern in the chip region or on specially designed targets. In another embodiment of the invention, the one or more processing parameters are selected from the group consisting of focus, exposure dose and overlay errors. 1) track parameters related to the dose (e.g. PEB time / temperature), i. There are also dose-related parameters. These actions can be modified by the exposure dose and can also be interpreted as the exposure dose by the model. Other processing parameters of the group may include projection lens aberration, flare for the projection lens, and angular distribution of light illuminating the mask, for example an ellipticity. In one embodiment of the present invention, values for one or more of these parameters can be determined individually, which are important for controlling critical dimension uniformity in the lithographic process.

In one embodiment of the invention, the regression technique used by the mathematical model is selected from the group consisting of eigen component regression, non-linear eigen component regression, partial least squares modeling and non-linear partial least squares modeling. .

In one embodiment of the invention, substrates that can be used include test wafers or product wafers. Depending on the particular application, the marker structure may be located on any location on the substrate. Thus, the marker structure may be located within the chip region or within the scribe-lane. If the marker structure is located in the chip region, it may be part of the device pattern in the chip region. Free positioning of the diffractive structure or free use of a portion of the device structure increases the flexibility of the method of the present invention.

In one embodiment of the invention, the marker structure comprises a diffraction grating. The diffraction grating is a structure suitable for scatterometry applications.

In another embodiment of the present invention, the method also includes the step of preprocessing the obtained calibration data and the obtained measurement data before using a mathematical model. Pretreatment often increases the performance of mathematical models. Mathematical tasks for pretreatment may include subtraction of mean, division by standard deviation, selection of optical parameters and weighting of optical parameters. Optical parameters are, for example, the wavelength, angle and polarization state of the light beam used by the optical detection device.

In one embodiment of the invention, the one or more substrates and the detection substrate comprise two or more different marker structures. In the case of a product substrate, two or more marker structures may be product marker structures, and in the case of a calibration substrate, two or more marker structures may be calibration marker structures. To simplify the term as much as possible, the term “marker structure” as used herein includes both states. Using more than one marker structure can be extremely advantageous when pretreatment is used. Two or more marker structures may be positioned in proximity to each other such that the distance between the two or more marker structures is about the same as the size of the marker structures.

In one embodiment of the invention, the two or more marker structures comprise a first marker structure comprising a plurality of non-patterned layers; And a second marker structure comprising the same non-patterned layers provided with a pattern on top. In this embodiment, the first marker structure is sensitive only to the variation of the non-patterned layers. Any spectral changes due to variation in non-patterned layers can be detected and used for analysis of the spectrum obtained on the second marker structure.

In another embodiment of the present invention, two or more marker structures comprise: a first marker structure comprising a pattern having separate lines; And a second marker structure comprising a pattern with dense lines or separated spaces. These marker structures can have different sensitivity to processing parameters such as focus and dose. As a result, additional information can be obtained, which can be useful for the determination of the values of the processing parameters by a mathematical model.

In one embodiment, the lithographic apparatus is connected to a track and the optical detection device is a scatterometer, which is connected to the same track. This enables an effective way to monitor and fit parameters for uniform performance of the lithographic apparatus.

At least some embodiments of the present invention combine the advantageous effects of preprocessing for the performance of a mathematical model and the advantages of using one or more marker structures to obtain the desired information. Suitable mathematical tasks include subtraction of mean, division by standard deviation, selection of optical parameters and weighting of optical parameters. Optical parameters are parameters such as wavelength, angle and polarization state of light used by the optical detection device.

In one embodiment of the invention, the two or more calibration marker structures comprise a first calibration marker structure comprising a plurality of non-patterned layers; And a second calibration marker structure comprising the same non-patterned layers provided with a pattern on top. In this embodiment, the first calibration marker structure is sensitive only to the variation of the non-patterned layers. Any spectral changes due to variations in the non-patterned layers can be detected and used for analysis of the spectrum obtained on the first calibration marker structure.

In one embodiment of the invention, the two or more calibration marker structures comprise: a first calibration marker structure comprising a pattern having separate lines; And a second calibration marker structure comprising a pattern with dense lines and separated spaces. These marker structures can have different sensitivity to treatment parameters such as focus and dose. As a result, additional information is obtained, which can be useful for the determination of the values of the processing parameters by a mathematical model.

In applications of embodiments of the present invention, calibration may generally be performed off-line. In one embodiment of the invention, it is not desirable that the process line is interrupted, so the measurement is performed on-line after the end of the calibration. In one embodiment of the invention, the scatterometer may be integrated in the track to enable on-line operation. Optionally, some substrates that have completed the processing operation may be measured by a standalone scatterometer, but processing continues. In the latter case, however, the feedback interval can be significantly increased. If the natural fluctuations present on the substrates for calibration (eg manufacturing wafers) are known in advance, for example by measurement, calibration and measurement can both be performed on-line.

Experimental Results: In the experiment, two types of wafers with a diameter of 30 mm were exposed.

The first type was a flat calibration wafer with FEM exposed. The FEM consisted of 13 focus steps (step size 30 nm) and 9 dose steps (step size 0.5 mJ / cm 2, nominal dose of about 29 mJ / cm 2). For each structure printed with unique focus and dose values, spectra were measured with a scatterometer. These spectra combined with the focus and dose offsets used were used to create the regression model.

The second type of wafer was a sampled wafer, i.e. a wafer to be measured. In the experiment, two sampled wafers were measured. Both sampled wafers included carefully made indentations in the wafer to achieve a better focus effect. Fields covering the entire wafer were exposed with focus and dose of one setting. However, due to natural variations in focus and dose, each exposed pattern will correspond to slightly different focus and dose from those set as described above. The printed structures were then measured by scatterometry. Using a regression model obtained from the FEM, each of the spectra belonging to the printed structures of the sample wafers was interpreted as a focus and dose value.

The regression model described above was applied to the spectra from sampled wafers with recesses, resulting in focus and dose distribution. In order to confirm the accuracy of the scatterometry results, a correlation to another established method was established. In this experiment, this correlation is based on the results of the so-called leveling verification test (LVT) for focus, for example, Valley et al., SPIE USE V.1 5375-132 (2004). Only by comparing with results from the tests discussed in. This test uses a reticle-type substrate with wedged-thickness regions, for example formed by providing a number of small prisms, each prism is fixed on a marker structure suitable for measuring overlay. The reticle-type substrate also includes a number of "normal" marker structures available as a reference. Defocus vs. lateral shift for prismatic “below” marker structures is a nearly linear relationship between defocus and image displacement. As a result, the focus errors are transformed into overlay errors. After measuring the wafer with a scatterometer, the wafer was peeled off, recoated and re-exposed for LVT measurements.

LVT-data were interpolated into a scatterometry measurement grid. As can be seen in Table 1 a very good correlation is observed between the two techniques. Table 1 shows the differences in focus measured with LVT and scatterometry for two sampled wafers and two types of recorded scatterometry spectra called α and β ² . Correlation is represented as 3σ-focus difference (dF), regression slope (tilt) and correlation coefficient R ² between the two techniques. Correlation results are very similar for both wafers and do not depend very much on the type of spectrum used. The upper limit on scatterometry accuracy is given by the difference in focus between the two techniques. Since the LVT was also recoated and reexposed between the two measurements with a certain inaccuracy, the actual accuracy would be better.

Focus differences between scatterometry and LVT dF [nm, 3σ] inclination R ² Wafer 1, α 36 0.88 0.85 Wafer 1, β 36 0.87 0.76 Wafer 2, α 37 0.81 0.81 Wafer 2, β 36 0.81 0.74 Average 36 0.85 0.79

In the above description, the marker structure is illuminated after development. However, it is also possible to use latent marker structures that are not yet developed, i. Latent markers can be imaged immediately after exposure, which is advantageous because the feedback-loop can be faster. Also, since the track processing has not yet been completed, the measurement data can be used for the feed forward signal to the track.

While specific embodiments of the invention have been described above, it will be appreciated that the invention may be practiced otherwise than as described. Embodiments of the invention also provide computer programs (eg, one or more sets or sequences of instructions) for controlling a lithographic apparatus to perform a method as described herein, and such a program in machine-readable form. Storage media (eg, disks, semiconductor memories) for storing one or more of them. The description is not intended to limit the invention. For example, the present invention can be applied to different technical areas, including areas such as lithography, MRI, and radar applications. However, it is particularly advantageous to use the invention in the lithographic domain, because of its complex and high-tech nature, which makes it very difficult if not impossible to control all parameters to the desired level. By using the present invention, parameters that cannot be controlled directly are indirectly controlled.

Claims (42)

In the method of determining one or more processing parameters, Obtaining calibration measurement data from a plurality of calibration marker structure sets provided on a calibration object, wherein each of the plurality of calibration marker structure sets comprises one or more calibration marker structures, and calibration of different calibration marker structure sets. Marker structures are generated using different known values of one or more processing parameters; Determining a mathematical model comprising a plurality of regression coefficients by using the known values of the one or more processing parameters and by using a regression technique on the calibration measurement data; Obtaining measurement data from one or more marker structures made using the unknown values of the one or more processing parameters provided on an object; And Determining an unknown value of the one or more processing parameters for the object from the obtained measurement data by using the regression coefficients of the mathematical model. The method of claim 1, The calibration measurement data and the measurement data are obtained using an optical detector. The method of claim 2, The optical detector is a scatterometer. The method according to any one of claims 1 to 3, The regression technique used by the mathematical model is from a group consisting of principal component regression, non-linear eigen component regression, partial least squares modeling and non-linear partial least squares modeling. Selected method. The method according to any one of claims 1 to 4, And the object is a substrate. The method of claim 5, Wherein said substrate comprises one of a group consisting of a test wafer and a product wafer. The method according to claim 5 or 6, Wherein said at least one marker structure is located on said substrate in one of a group consisting of a chip region and a scribe-lane. The method of claim 7, wherein Wherein said at least one marker structure is part of a device pattern in a chip region. The method according to any one of claims 1 to 8, Wherein said at least one marker structure comprises a diffraction grating. The method according to any one of claims 1 to 9, The method further comprising preprocessing the obtained calibration measurement data and the obtained measurement data before using the regression coefficients. The method of claim 10, The preprocessing includes performing one or more of a group of mathematical operations consisting of subtraction of the mean, division by standard deviation, selection of optical parameters and weighting of optical parameters on the data, Said optical parameters comprise one or more of a group of parameters consisting of wavelength, angle and polarization state. The method according to any one of claims 1 to 11, Wherein the plurality of calibration marker structure sets each comprise at least a first and a different second calibration marker structure. The method of claim 12, Wherein the first calibration marker structure comprises a plurality of non-patterned layers, and the second calibration marker structure comprises the same non-patterned layers provided with a pattern on top. The method according to claim 12 or 13, Wherein the first calibration marker structure comprises a pattern with separate lines, and the second calibration marker structure comprises a pattern with dense lines or independent spaces. The method according to any one of claims 12 to 14, The first and second calibration marker structures are in proximity to each other such that the distance between the first and second calibration marker structures is about the same size as the size of the first and second calibration marker structures. . The method according to any one of claims 1 to 15, At least one calibration structure in the set of calibration marker structures and the marker structure have a substantially similar shape. The method according to any one of claims 1 to 16, Wherein said calibration data and said measurement data comprise spectral data. The method according to any one of claims 1 to 17, The method relates to at least one of a lithographic apparatus and a track. The method of claim 18, The one or more processing parameters may include focus, exposure dose, overlay error, track parameters related to dose, variation in line width on the reticle, variation from the reticle-to-reticle, projection lens aberration, projection lens flare, and reticle. And a angular distribution of the illuminating lights. The method of claim 18 or 19, The lithographic apparatus is: An illumination system configured to provide a radiation beam; A support structure configured to support a patterning structure that serves to impart a pattern to a cross section of the radiation beam; A substrate table configured to hold a substrate; And And a projection system configured to project the patterned beam onto a target portion of the substrate. A semiconductor device manufactured by the method according to any one of claims 1 to 20. A system for determining one or more processing parameters, A detector arranged to obtain calibration measurement data from a plurality of calibration marker structure sets provided on a calibration object, the plurality of calibration marker structure sets each including one or more calibration marker structures, and calibration markers of different calibration marker structure sets Structures are obtained using different known values of the one or more processing parameters; A processor unit that uses the known values of the one or more processing parameters and stores a mathematical model including a plurality of regression coefficients, determined by using a regression technique on the calibration measurement data; The processor unit is further configured to obtain measurement data from one or more marker structures made using an unknown value of the one or more processing parameters provided on an object; And determine the unknown value of the one or more processing parameters for the object from the obtained measurement data by using the regression coefficients of the mathematical model. The method of claim 22, And said detector is an optical detector. The method of claim 23, wherein And the optical detector is a scatterometer. The method according to any one of claims 22 to 24, The regression technique used by the mathematical model is selected from the group consisting of eigen component regression, non-linear eigen component regression, partial least squares modeling and non-linear partial least squares modeling. The method according to any one of claims 22 to 25, And the object is a substrate. The method of claim 26, And the substrate comprises one of a group consisting of a test wafer and a product wafer. The method of claim 26 or 27, Wherein said at least one marker structure is located on said substrate in one of a group consisting of a chip region and a scribe-lane. The method of claim 28, And the one or more marker structures are part of a device pattern in a chip region. The method according to any one of claims 22 to 29, Wherein said at least one marker structure comprises a diffraction grating. The method according to any one of claims 22 to 30, And the processor unit is arranged to preprocess the obtained measurement data before using the regression coefficients. The method of claim 31, wherein The preprocessing performs one or more of a group of mathematical tasks consisting of subtraction of the mean on the data, division by standard deviation, selection of optical parameters and weighting of optical parameters, Wherein the optical parameters comprise one or more of a group of parameters consisting of wavelength, angle and polarization state. The method according to any one of claims 22 to 32, Wherein the plurality of calibration marker structure sets each comprise at least a first and a different second calibration marker structure. The method of claim 33, Wherein the first calibration marker structure includes a plurality of non-patterned layers, and the second calibration marker structure includes the same non-patterned layers provided with a pattern on top. The method of claim 33 or 34, Wherein the first calibration marker structure comprises a pattern having separate lines, and the second calibration marker structure comprises a pattern having dense lines or separated spaces. The method according to any one of claims 33 to 35, And the first and second calibration marker structures are in close proximity to each other such that the distance between the first and second calibration marker structures is about the same size as the size of the first and second calibration marker structures. . The method according to any one of claims 22 to 36, At least one calibration structure in the set of calibration marker structures and the marker structure is substantially similar in shape. The method according to any one of claims 22 to 37, And the calibration data and the measurement data comprise spectral data. The method according to any one of claims 22 to 38, The system comprises at least one of a lithographic apparatus and a track. The method of claim 39, The one or more processing parameters include focus, exposure dose, overlay error, track parameters related to dose, variation in line width on the reticle, variation from the reticle-to-reticle, projection lens aberration, projection lens flare, and light illuminating the reticle And a angular distribution of s. The method of claim 39 or 40, An illumination system configured to provide a radiation beam; A support structure configured to support a patterning structure that serves to impart a pattern to a cross section of the radiation beam; A substrate table configured to hold a substrate; And A projection system configured to project a patterned beam onto a target portion of the substrate. 42. A semiconductor device manufactured with the system according to any one of claims 22-41.

Images