WO2011117872A1

WO2011117872A1 - System and method for controlling a lithography process

Info

Publication number: WO2011117872A1
Application number: PCT/IL2011/000274
Authority: WO
Inventors: Boaz Brill
Original assignee: Nova Measuring Instruments Ltd.
Priority date: 2010-03-24
Filing date: 2011-03-24
Publication date: 2011-09-29

Abstract

A measurement system and method is presented for measuring properties of a structure having a pattern of spaced-apart features arranged along a pattern axis. The measurement system comprises: a structure support unit defining a support plane for supporting the structure, an optical system comprising an illumination system defining an illumination path, and at least one detection system defining one or more detection paths, and a control system. The optical system has a predetermined numerical aperture, and is configured to define an incidence plane and the detection corresponding to dark-field detection mode for collecting light propagating from an illuminated region on the structure with a solid angle outside that of specular reflection, said incidence plane being oriented with respect to said support plane such as to form a selected angle other than 90 degrees with said pattern axis. The control system is configured and operable for receiving from the detection system data indicative of light detected with said dark-field mode and processing the received data by applying thereto predetermined modeled data based on a predetermined unit cell having a dimension along the patterned axis selected in accordance with the numerical aperture of the optical system.

Description

SYSTEM AND METHOD FOR CONTROLLING A

LITHOGRAPHY PROCESS

FIELD OF THE INVENTION

This invention is generally in the field of process control techniques, and relates to a method and system for controlling a lithography process used in the manufacture of patterned structures. The invention is particularly pseful for patterning of reticles or semiconductor wafers used in the manufacture of semiconductor devices.

BACKGROUND OF THE INVENTION

Lithography is widely used in various industrial applications, such as manufacture of integrated circuits, flat panel displays, micro-mechanical systems, micro-optical systems, etc. In some applications, for example the manufacture of integrated circuits, a semiconductor wafer undergoes a sequence of lithography-etching steps to produce a plurality of spaced-apart stacks, each formed by a plurality of different layers having different optical properties. Each lithography procedure applied to the wafer results in the pattern on the uppermost layer formed by a plurality of spaced-apart photoresist regions. One of the important parameters for lithography is a critical dimension (CD) of the pattern, i.e. the smallest lateral dimension of the developed photoresist, usually being the width of the finest lines or the width of the smallest spaces between lines.

Lithography quality is a key issue in semiconductor manufacturing as continuous process design rule is shrinking. This is because it is often the case with lithography that the edge of the pattern feature has certain deviation from the average CD value, which is more critical for structures with smaller CD values, as the deviation value approaches the required CD. Current methodology for quality control during production is based on existing metrology tools which are focused mainly on measuring average CD values of feature (line) width, and measuring additional profile parameters such as side wall angle. These average geometrical values, measured either by CD-SEM, optical techniques (Scatterometry), probe techniques (AFM) or otherwise are then used mainly for controlling dose and focus of the lithographical exposure step.

As lithographical design rules further shrink it becomes more important to control not only the average values of the features but also the deviations from such average features. Common types of such deviations are known as Line Edge Roughness (LER) or Line Width Roughness (LWR), relating to the short-range random variations that are found in actual printed features (unlike the design intent, usually designed for smooth feature boundaries). Additional types of irregularities occur whenever the lithographical process is sub-optimal for some reason and may be manifested by either enhanced bottom footings, remaining photoresist layers ('residue') in intended space areas, falling lines and various defect types. Fig. 1A schematically illustrates the LER effect resulting from a lithography process. As shown the CD_average value varies along the feature (line) of the pattern.

U.S. Patent No. 7,184,152, assigned to the assignee of the present application, describes a method and system for optical measurements of line edge roughness (LER) of patterned structures. According to this technique, the structure is illuminated with incident radiation and a spectral response of the structure is detected. Then, data about the line edge roughness parameter(s) from the spectral response.

GENERAL DESCRIPTION

The invention provides a novel approach for process and quality control of the lithography process suitable to be used in the manufacture of patterned structures with small CD values. The invention utilizes dark-field detection of spectral response of a patterned structure with a certain azimuthal angle.

Dark-field inspection is known as that where an optical response of an illuminated region is collected at azimuth and elevation outside those where specular reflection occurs. This can for example be achieved using normal incidence of illuminated light and detection of light returned from the illuminated region with a certain grazing angle, or using oblique illumination and detection of light propagating normally from the illuminated surface. On the contrary, bright-field detection utilizes collection of light at azimuth and elevation where most of specular reflection occurs.

The above is schematically illustrated in Figs. IB and 1C showing in a self- explanatory manner two examples, respectively, of bright-field and dark-field detection schemes.

The need for the dark-field detection in the present invention is associated with the following. The invention is aimed at detecting irregularities in patterned structures fabricated using lithography with shining wide-band light on a substantially periodic structure (grating). With the bright-field detection, collected light is formed by a signal from the periodic (regular) structure and also a signal from irregularities, while with the dark-field collected light practically does not include signal associated with the regular structure but mainly includes a signal due to irregularities.

The present invention is based on the inventors' understanding of the fact that while light returned from flat surfaces includes solely specular reflection propagating with a well-defined angle, light returned from surfaces having small irregularities, much smaller than the wavelength utilized, is scattered light propagating to some extent with any solid angle. In the general situation, dark field detection systems are used with flat or general surfaces and structure and are configured to solely eliminate collection of specular reflection. The present invention deals with structures having irregularities in periodic structures (gratings). Accordingly, in order to detect those irregularities it is insufficient to avoid collection of specular reflection, but also it is required to eliminate or at least significantly reduce collection of non-zero diffraction orders that may arise from the interaction of an illuminating beam and the grating. To this end, the invention utilizes light collection with a certain azimuth angle that prevents any diffraction order from being collected. For example, for a system utilizing oblique illumination channel and a normal (to the sample) collection channel, and applied to a structure having a pattern in the form of spaced-apart lines, it would be preferably to operate with the azimuth satisfying a condition that angular orientation of the incidence plane with respect to the sample's plane is such that the lines of the pattern are not perpendicular to the incidence plane, and preferably substantially parallel thereto. The simplest example would be azimuth in which the lines are completely in the incidence plane. Also, the invention utilizes a wide spectral range optical system in which illumination is implemented using a multi-spectral illumination source (e.g. a discharge lamp) and light detection utilizes a multi-spectral detection unit, e.g. a spectrometer, in which case measured data is spectral data (the returned signal as function of wavelength). Since scattering by small objects is highly wavelength dependent (Rayleigh scattering model predicts scattering intensity to scale as λ^"4 ), using multi- spectral optical system allows for obtaining additional information enabling measurement of several statistical properties of the grating irregularities. The invention may thus utilize UV spectrum or part thereof for the detection of imperfections in the structures under study.

Thus, according to one broad aspect of the present invention, there is provided a measurement system for measuring properties of a structure having a pattern of spaced- apart features arranged along a pattern axis, the measurement system comprising: a structure support unit defining a support plane for supporting the structure, an optical system comprising an illumination system defining an illumination path, and at least one detection system defining one or more detection paths, and a control system, wherein:

the optical system has a predetermined numerical aperture, and is configured to define an incidence plane and the detection corresponding to dark-field detection mode for collecting light propagating from an illuminated region on the structure with a solid angle outside that of specular reflection, said incidence plane being oriented with respect to said support plane such as to form a selected angle other than 90 degrees with said pattern axis;

the control system is configured and operable for receiving from the detection system data indicative of light detected with said dark-field mode and processing the received data by applying thereto predetermined modeled data based on a predetermined unit cell having a dimension along the patterned axis selected in accordance with the numerical aperture of the optical system.

Preferably, the incidence plane is substantially parallel to the pattern axis. Generally, this angle is selected so as to reduce non-zero diffraction components in the dark-field detected light. Preferably, said dimension of the unit cell along the patterned axis is selected such that multiple diffraction orders from the modeled unit cell are collectable with the predetermined numerical aperture.

The optical system is configured and operable as a wide band system. For example, the system comprises a spectrometer, and data indicative of light detected with said dark-field mode is spectral data. The wide band may comprise at least a part of UV spectral band.

In some embodiments, the optical system is further configured and operable in bright-field detection mode.

The control system may comprise a library of modeled data, and/or may be configured and operable for creating the modeled data.

Preferably, the structure support unit is configured such that its upper surface is tiltable with respect to a horizontal plane, thereby enabling calibration measurements.

According to another broad aspect of the invention, there is provided a method for measuring on patterned structures having a pattern of spaced-apart features arranged along a pattern axis, the method comprising:

Illuminating a region on said pattern and detecting light returned from the illuminated region, the detected light comprising at least light propagating with a solid angle outside that of specular reflection, wherein said illuminating comprising orienting incidence plane with respect to a structure's surface such as to form a selected angle other than 90 degrees with said pattern axis;

processing data indicative of said detected light with the solid angle outside that of specular reflection, said processing comprising applying predetermined modeled data based on a predetermined unit cell smaller than the illuminated region and having a dimension along the patterned axis selected in accordance with a numerical aperture of the optical system.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to understand the invention and to see how it may be carried out in practice, embodiments will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which: Fig. 1A schematically illustrates an LER effect in a patterned structure resulting from a lithography process creating the pattern;

Figs. IB and 1C show two examples respectively of light propagation scheme, each using bright- and dark-field detection, where Fig. IB shows oblique illumination path and a detection path along an axis normal to the surface of the article being examined, and Fig. 1C shows a normal-incidence illumination path and grazing-angle detection path;

Fig. 2 is a block diagram of a measurement system of the present invention;

Fig. 3 shows an optical scheme of light propagation in the system of Fig. 2; Fig. 4 illustrates the principles of a calibration procedure;

Figs. 5A to 5C exemplify spectral responses of a patterned structure for different structure parameters: Fig. 5 A shows a measured scattered spectrum for oxide grating on Si, with a film thickness of 600nm; Fig. 5B shows similar graphs as in Fig. 5 A for gratings with different pitches, and Fig. 5C shows similar graphs as in Fig. 5 A for gratings with different film (line) heights; and

Figs. 6A to 6C illustrate the principles of the invention for selecting a unit cell for modeling the optical response of the patterned structure.

DETAILED DESCRIPTION OF EMBODIMENTS

As described above and exemplified in Fig. 1A, one of the problems with a lithography process is an LER effect consisting of a varying value of CD_average along the feature (line) of the pattern produced by said lithography process. Figs. IB and 1C show two examples of optical schemes in an inspection system utilizing dark field detection.

Reference is made to Fig. 2 showing a bock diagram of a measurement system

10 of the present invention for process and quality control of the lithography procedure applied to an article, typically a semiconductor wafer W located on a structure supporting unit (e.g. stage) 21. The measurement system 10 includes an optical system

11 and a control system 20. The optical system 11 includes illumination system 12 appropriately accommodated with respect to the wafer to define an illumination path 14, and a detection system 16 appropriately accommodated with respect to the wafer to define a detection path 18. The optical system is configured to enable operation with a predetermined dark- field detection mode. Generally, the dark-field detection utilizes collection of light from the wafer propagating with a solid angle outside that of the specular reflection. The dark-field detection mode used in the invention utilizes this general principle. As shown schematically in the example of Fig. 2, the illumination system is oriented/accommodated to define the illumination channel along an axis perpendicular to the wafer's surface (normal incidence) and the detection system is accommodated/oriented to define oblique light collection channel.

It should be understood that the illumination system typically includes a light source unit (including one or more light sources) and may also include a light directing optics, in which case the illumination path orientation is defined by the light directing optics. Similarly, the detection system typically includes a light detector unit (including one or more light detectors) and light collecting optics, the detection path orientation being actually defined by the light collecting optics.

The control system 20 is typically a computer system connectable (via wires or wireless signal transmission, as the case may be) to the output of the detection system 16 for receiving therefrom measured data indicative of the light response, and process and analyze this data to determine properties of the structure, such as CD_aVerage and its fluctuations being indicative of LER or LWR. The control system 20 thus includes data input and output utilities, generally at 26, a memory utility 22 for storing inter alia predetermined modeled data, and a processor utility 24 for processing and analyzing the measure data by applying thereto the modeled data, and generating output data indicative of one or more desired property of the structure. The operation of the processor utility will be described more specifically further below.

According to the invention, the optical system is configured as a wide spectral range system. This is associated with the following: The optical system of the invention is aimed at detecting light scattered (not specularly reflected) from the patterned surface. Scattering by small objects (a pattern of small features in this specific case) is highly wavelength dependent (Rayleigh scattering model predicts scattering intensity to scale as λ^"4). Hence, using multi-spectral optical system allows for obtaining additional information about the patterned structure enabling measurement of several statistical properties of the grating irregularities. To this end, the illumination system utilizes a broadband illumination, which may be implemented using a broad band or multi- spectral light source assembly (e.g. a discharge lamp), and the light detection system utilizes a multi-spectral detection unit, e.g. a spectrometer. Thus, measured data includes spectral response of the illuminated patterned region. This technique allows for using UV spectrum for detection of imperfections in the patterned structure such as wafer.

As indicated above, the present invention utilizes a predetermined dark-field detection mode, rather than just collecting light propagating with solid angle(s) outside that of the specular reflection. The technique of the present invention utilizes modeling of spectral light response of substantially periodic structures (gratings). Light interacting with such a structure undergoes diffraction, and accordingly light scattered from the structure includes non-zero diffraction orders. The present invention utilizes such dark- field detection mode which, in addition to avoidance of collection of peculiarly reflected light from the wafer, also enables to eliminate or at least significantly reduce collection of the non-zero diffraction orders.

In this connection, reference is made to Fig. 3 illustrating the light propagation scheme in the optical system of the invention. The optical system includes the illumination system including a light source unit (not shown) and possibly also includes light directing optics, and defines the illumination path 14. In the present example, oblique illumination is used, i.e. the illumination path forms a certain angle a being 90< a<0. The detection system includes a detector unit (not shown) and light collecting optics (lens assembly) 30. In the present example, the collection lens 30 is accommodated to define the detection path 18 along an axis substantially normal to the wafer's surface. As shown in the figure, light returned from the patterned wafer has multiple diffraction orders. According to the invention, the optical system is configured for light collection with a certain azimuth angle that substantially prevents diffraction orders from being collected. The structures with which the invention is used are of the type where a pattern is in the form of spaced-apart features extending along a certain pattern axis PA (e.g. parallel regions or lines or parallel arrays of spaced-apart discrete elements). In the present example, where the optical system utilizes oblique illumination channel 14 and normal (to the sample) collection channel 18, the selected azimuth is such that angular orientation of the incidence plane with respect to the sample's plane is not perpendicular to the lines (features) of the pattern), and preferably is substantially parallel thereto. The simplest example would be azimuth in which the lines are completely in the incidence plane. Generally speaking, the optical system 11 is configured such that the incidence plane defined by the system is oriented with respect to the structure supporting plane P such as to define a certain non-right angle with respect to the pattern axis PA, preferably substantially zero angle.

5 As indicated above, the present invention utilizes accurate, from basic principles modeling of LER. To this end, a calibrated spectrum and an accurate model are provided. The principles of the calibration procedure used in the invention are schematically illustrated in Fig. 4 showing a calibration setup 32. The calibration setup 32 includes an illumination system (not shown) defining an oblique illumination path

10 14, a detection system including a detector unit (not shown and light collecting optics (lens) 30 and a surface 34 of known reflectivity (e.g. Silicon). The surface 34 is tilted with respect to a measurement plane (wafer's surface) a certain angle β such as to reflect light along the detection path directly into the collection lens 30 of certain known numerical aperture NA. Preferably, the surface 34 is tilted such that the normal 36 to the

15 surface 34 is a bisektris of an angle φ between the illumination path 14 (illumination beam) and the detection path 18 (an axis passing though the center of the collection NA); in other words the normal 36 to the surface divides the angle φ into two equal parts.

Resulting calibrated spectra (detected by the detector unit) are shown in Figs. 0 5A-5C. Fig. 5A shows a graph Gi corresponding to the measured scattered spectrum, from oxide grating of lOOnm CD_average (line width) and 250nm pitch on Si film of 600nm thickness. Fig. 5B shows measured scattered spectra for similar structures but with different pitch values: 250 nm pitch (graph Gi), 500 nm pitch (graph G₂), and 1000 nm pitch (graph G3). Fig. 5C shows measured scattered spectra Hi, ¾, ¾ and H₄ for 5 similar structures but with different film (line) heights respectively. As can be seen, faster oscillations occur for larger line height (increased LER signal at lower thickness values is due to previous etching).

Accurate modeling can be applied to the problem of LER by approximating this situation to diffraction from a grating. Reference is made to Figs. 6A - 6C exemplifying0 such approximation technique. In order to make use of calculation methods that where developed for repetitive structures, an effective unit cell is defined. According to the invention, the unit cell is selected to be sufficiently long along an axis in the incidence plane (Fig. 6B, dimension a), such that an angular difference between diffraction orders is sufficiently smaller than the collecting NA for the shortest measurement wavelength (Fig. 6A). In this way, under any condition (any wavelength), the number of collected diffraction orders is sufficiently large to avoid artificial jumps whenever the number of collected diffraction modes is changing during the actual measurements. In the other direction, a size of the unit cell is defined by the actual structure on the wafer and is typically much smaller (Fig. 6B, dimension b). Then, LER is applied by randomly sampling small deviations around relevant side walls (Fig. 6B). With regard to selection of the long dimension a, the above condition thereof is based on the principles of diffraction. More specifically, the diffraction equation is:

d^sinBi-sinB^n ,

wherein d is the dimension of the diffractive element (corresponding to dimension a in Fig. 6B), n in the integer corresponding to the diffraction order number, Q_\ and θ₂ being angles of propagation for incident and diffracted rays. In the simplest case of normal incidence or collection, sin9₂=0, and therefore we have'

The condition that the numeral aperture NA of the system used in modeling procedure is much larger than the difference between the diffraction orders (i.e. the NA allows collection of multiple diffraction orders) can actually be expressed as:

n>l/NA

For example, for NA=0.1 and the shortest wavelength (or that closer to the low limit of the range used)

we obtain the unit cell dimension a to be about ΙΟμπι.

It should be noted that using a long unit cell also enables significant sampling of the LER within the single artificial unit cell. The sampling statistics could be in the simplest case white noise, however more complex statistics can also be used, e.g. allowing a predefined amount of spatial correlations as a function of the spatial frequency, approximating real LER behavior as analyzed by power spectrum density (PSD). The effective PSD can be represented by one or more parameters that can then be treated as variables in a fitting procedure. LER variations can also be simulated to have a dependence on the height (z) to allow closer representation of a real process.

Once the structure is effectively defined, calculation can be done for example by using RCWA, and the signal entering the NA through the suitable diffraction modes can be integrated over either coherent or non-coherent summation. It should be noted that in order to better represent the statistical distribution of the LER that can be done in a single unit cell, a multiplicity of specific realizations (different runs of the random generator) can be sampled to create several structures. The LER signal calculated for different realizations can then be used for example by way of averaging, to create a more accurate modeling of the LER process.

Fig. 6C shows results of the calculation (modeling) done using the unit cell of Fig. 6B for a line grating of lOOnm width, 300nm pitch and 600nm height. As predicted, the collected spectrum has strong wavelength dependence, and in addition it shows characteristic oscillations with wavelength typical for thin film spectra. As can be seen, the resulting spectrum bears resemblance to the measured spectrum Gi (Fig. 5A). Hence, it is likely that by correctly changing the parameters of the model in a fitting procedure a good match can be obtained between measured and modeled data (spectra), enabling the measurement of the magnitude (standard deviation) and statistical properties (PSD parameters) of the structure under measurement using from-basic- principles modeling.

Turning back to Fig. 2, the modeled data stored in the memory utility 22 of the control system 20 includes a library of spectra corresponding to patterned structures of different parameters of the pattern, in particular the line width and height. The processor utility 24 performs a fitting procedure to find the best fit between the measured and modeled data and thus perform LER measurements enabling to control the process and quality of lithography. It should be understood that modeled data can be provided off line or can be provided or updated during actual measurements (online) using real time calculation.

It should also be noted that the same optical system can be used for accurate determination of the desired parameters of the structure such as CD profile, side wall angle and other parameters relating to the pattern on the structure (e.g. using so-called scaterrometry), and for LER and/or LWR. This can be implemented using both the conventional bright-field mode and the above described dark-field mode of the present invention. In this connection, reference is made back to Fig. 2, showing by dashed lines that the system 10 may include also the bright-field detection channel. As shown, in the present not limiting example, where the illumination channel is normal to the structure's surface, the bright field detection channel coincides with the illumination channel. To this end, appropriate optics is typically provided in the bright-field detection channel for collection of the returned light (reflected / diffracted) and directing it to the detection system. In the present example of Fig. 2, the detection system 16 may utilizes the same detector for timely modulated operation (e.g. sequentially) in the bright-field and dark- field detection modes or may utilize separate detectors with their associated optics. Accordingly, as shown in the figure in dashed lines, the system includes a partially transparent element 19 accommodated in the illumination/detection channel 14 for transmitting illuminating beam and reflecting the returned beam.

It should be noted, although not specifically illustrated, that another possible implementation of a combination of the bright- and dark-field modes is by using an oblique illumination channel, an oblique bright-field collection channel and a normal dark-field collection channel. While the oblique bright-field collection channel measures zero order diffraction enabling the measurement of CD and other profile parameters of the structure used for process control, the normal dark-field collection channel, or alternatively another oblique dark-field collection channel at a different angle, collects scattered light associated with LER effect and enables monitoring (detection) or a full model-based measurement of LER. Thus, the invention may provide a single measurement system enabling concurrent or sequential measurements of various parameters / conditions of the structure using combination of the bright-field and dark-field modes, while utilizing a predetermined azimuth of light collection in the dark mode with respect to the patterned features.

The technique of the present invention provides for monitoring the quality of lithography in production. Detection of deviations of the pattern from that of an ideal periodic structure using the principles of the dark-field technique modified according to the invention as described above can be used for monitoring process imperfections during the manufacturing of semiconductor devices. By registering the intensity or the full spectrum of the dark field scattering and monitoring it over time the technique enables detection of any deviation in the lithographic process. It should be understood that the invention can provide useful monitoring of the quality of the lithographic process, and in some cases even without modeling and exact translation to LER statistical size (but using the integrated intensity, the maximal intensity, intensity at a specific wavelength or any other characteristic of the measured spectrum). It should also be noted that while the above description is focused mainly on LER measurements which is typically a characteristic of the side walls, the same technique can be used for detecting additional effects that are typical for imperfect lithography process and can be a good cause for invoking additional tests. These additional effects may include one or more of the following: resist loss which is typically accompanied by roughness on the top of the lines, collapsing features, excessive bottom rounding (e.g. typical for under-exposure condition) which in extreme cases can lead to residues, significant defocus increasing the amount of LER, material changing, etc.

It should also be understood that the technique of the present invention is applicable to periodic structures. While in the general case patterns used in semiconductor devices might include regions with non-periodic or not exactly periodic patterns. In such cases, where suitable repetitive structures having sufficient area to accommodate the optical measurement spot cannot be found in the patterned structure under measurements (in-die), special test sites having repetitive (periodic) structures over sufficient area can be used. Such test structures may for specifically designed in order to enhance the sensitivity of the above described LER/LWR measurement technique. More specifically, the test site (structure) may be designed to have higher sensitivity to process variations than the regular structures of the real devices. Some examples of suitable test structures are described in WO 2009/107143, assigned to the assignee of the present application, which is incorporated herein by reference with respect to this specific example.

Claims

CLAIMS:

1. A measurement system for measuring properties of a structure having a pattern of spaced-apart features arranged along a pattern axis, the measurement system comprising: a structure support unit defining a support plane for supporting the structure, an optical system comprising an illumination system defining an illumination path, and at least one detection system defining one or more detection paths, and a control system, wherein:

2. The measurement system of Claim 1, wherein said incidence plane is substantially parallel to said pattern axis.

3. The measurement system of Claim 1 or 2, wherein said angle other than 90 degrees is selected so as to reduce non-zero diffraction components in the dark-field detected light.

4. The measurement system of any one of Claims 1 to 3, wherein said dimension along the patterned axis is selected such that multiple diffraction orders from said modeled unit cell are collectable with said numerical aperture.

5. The measurement system of any one of Claims 1 to 4, wherein said optical system is configured and operable as a wide band system.

6. The measurement system of Claim 5, wherein said optical system comprises a spectrometer, data indicative of light detected with said dark-field mode being spectral data.

7. The measurement system of Claim 5 or 6, wherein said wide band comprises at least a part of UV spectral band.

8. The measurement system of any one of Claims 1 to 7, wherein said optical system is further configured and operable in bright-field detection mode.

9. The measurement system of any one of Claims 1 to 8, wherein said control system comprises a memory utility for storing a library of modeled data.

10. The measurement system of any one of Claims 1 to 9, wherein said control system is configured and operable for creating the modeled data.

11. The measurement system of any one of Claims 1 to 10, wherein said structure support unit is configured such that its upper surface is tiltable with respect to a horizontal plane, thereby enabling calibration measurements.

12. A method for measuring on patterned structures having a pattern of spaced-apart features arranged along a pattern axis, the method comprising:

illuminating a region on said pattern and detecting light returned from the illuminated region, the detected light comprising at least light propagating along a detection path of a solid angle outside that of specular reflection, wherein said illuminating comprising orienting incidence plane with respect to a structure's surface such as to form a selected angle other than 90 degrees with said pattern axis;

13. The method of Claim 12, wherein said incidence plane is substantially parallel to said pattern axis.

14. The method of Claim 12 or 13, wherein said angle other than 90 degrees is selected so as to reduce non-zero diffraction components in the dark-field detected light.

15. The method of any one of Claims 12 to 14, comprising selecting said dimension along the patterned axis such that multiple diffraction orders from said modeled unit cell are collectable with said numerical aperture.

16. The method of any one of Claims 12 to 15, wherein said illumination comprises wide band ilumination.

17. The method of Claim 16, wherein said data indicative of the detected light comprises spectral data.

18. The method of Claim 16 or 17, wherein said wide band comprises at least a part of UV spectral band.

19. The method of any one of Claims 12 to 18, further comprising detecting light specularly reflected from the illuminated region.

20. The method of any one of Claims 12 to 19, comprising creating a library of modeled data.

21. The method of any one of Claims 12 to 20, further comprising performing a calibration procedure by determining reflection from a calibrated reflective surface oriented such that collected reflected light propagate along said detection path.