METHOD AND SYSTEM FOR OPTICAL CHARACTERIZATION OF PATTERNED SAMPLES

Abstract

A method and system are presented for use in measuring on patterned samples, aimed at determining asymmetry in the pattern. A set of at least first and second measurements on a patterned region of a sample is performed, where each of the measurements comprises: directing illuminating light onto the patterned region along an illumination channel and collecting light reflected from the illuminated region propagating along a collection channel to be detected, such that detected light from the same patterned region has different polarization states which are different from polarization of the illuminating light, and generating a measured data piece indicative of the light detected in the measurement. Thus, at least first and second measured data pieces are generated for the at least first and second measurements on the same patterned region. The at least first and second measured data pieces are analyzed and output data is generated being indicative of a condition of asymmetry in the patterned region.

Claims

1. A method for use in measuring on patterned samples, the method comprising: performing first and second measurements on a patterned region of a sample and generating measured data comprising first and second data pieces for said first and second measurements, respectively, performed on the same patterned region, said performing of each of the first and second measurements comprising: directing illuminating light onto said patterned region along an illumination channel and controlling polarization state of the illuminating light, such that the illuminating light in each of the first and second measurement sessions has a predetermined polarization state different from that of the other of the first and second measurements, and collecting light reflected from said patterned region propagating along a collection channel to be detected while affecting polarization state of the reflected light propagating along the collection channel, such that light being detected in each of the first and second measurements has a predetermined polarization state which is different from a polarization state of the illuminating light in said measurement, each of the first and second measured data pieces thereby corresponding to the detected light corresponding to a reflection response of the patterned region to the illuminating light having different polarization state; and analyzing the first and second measured data pieces to determine a relation between them, providing a direct measure of a condition of asymmetry of the pattern in said patterned region, and generating output data indicative of the condition of asymmetry in said patterned region.

2. A system configured and operable to carry out the method of claim 1, for measurements in patterned samples, the system comprising: (a) an illumination unit and a light detection unit which are configured to define, respectively, an illumination channel for illuminating a region of a sample with an illuminating beam and a collection channel for detecting collected light by a detector, which generates data indicative of detected light; (b) a polarization unit comprising at least one polarization filter located in the illumination channel configured and operable to control polarization state of the illuminating light to be of first and second different polarization states in first and second measurements, and at least one polarization filter located in the collection channel and configured and operable during each of the first and second measurements to allow passage of light of polarization state different from the polarization state of the illuminating light beam used in said measurement to the detector for generating data indicative of light detected in said measurement, and (c) a control unit configured and operable to receive the data indicative of the detected light from same patterned region of a sample and generating and analyzing first and second measured data pieces corresponding to the first and second measurements each being indicative of detected light having polarization state different from the polarization state of the illuminating light beam in the respective measurement, said analyzing comprising determining a relation between the first and second measured data pieces, and generating output indicative of a condition of symmetry of the pattern in said region of the sample.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to better understand the subject matter that is disclosed herein and to exemplify how it may be carried out in practice, embodiments will now be described, by way of non-limiting examples only, with reference to the accompanying drawings, in which:

[0035] FIG. 1A exemplifies a patterned structure wherein a pattern is asymmetric in that the features of the pattern are tilted;

[0036] FIG. 1B and FIG. 1C exemplify test structures with, respectively, matched patterns in two adjacent patterned layers (symmetric configuration) and a mismatched overlay pattern (asymmetric due to a shift between the patterns);

[0037] FIG. 2A illustrates schematically a measurement system suitable for implementing the technique of the invention;

[0038] FIG. 2B exemplifies a measurement scheme used in the system of the invention;

[0039] FIG. 3 shows a flow diagram exemplifying a method according to the present invention;

[0040] FIGS. 4A to 4D illustrate four examples of the technique of the invention, wherein FIG. 4A exemplifies light collection with linear polarization, FIG. 4B exemplifies illumination and collection with linear polarized light; FIG. 4C exemplifies illumination and collection with circular polarized light; and FIG. 4D exemplifies the use of s and p polarized light;

[0041] FIG. 5 illustrates a patterned sample for overlay mismatch measurement; and

[0042] FIG. 6 shows measurement results obtained on the sample of FIG. 5 using the technique(s) of the invention and the conventional technique for overlay mismatch measurements.

DETAILED DESCRIPTION

[0043] As indicated above, the technique of the present invention allows for simple and robust characterization of patterned sample to identify asymmetry of a pattern, as well as determine a level of asymmetry.

[0044] Reference is made to FIGS. 1A to 1C schematically exemplifying asymmetry of patterns in patterned samples, which might occur during the manufacturing process.

[0045] FIG. 1A illustrates a sample 10 including a substrate structure 12 (single or multi-layer structure) with a patterned layer 14 on surface 12A of the substrate, where a pattern is in the form of a tilted grating 14 (constituting an asymmetric pattern, which in this particular example is breaking the mirror symmetry). Such a tilted grating might result from imperfections in the fabrication (patterning) process. As shown, side wall angles SWA.sub.1 and SWA.sub.2 of the grating 14 are unequal.

[0046] The example of FIGS. 1B and 1C illustrate a sample 100 including a substrate structure 12 (single or multi-layer structure) having a surface 12A on which two patterned layers with patterns 14A and 14B respectively are formed, one above the other (with or without intermediate layer(s) between them). As shown in FIG. 1B, the pattern features 16 are properly aligned corresponding to matched overlay between the patterns 14A and 14B resulting in mirror symmetry along an axis A, while FIG. 1C illustrates a similar sample having a mismatch overlay between patterns 14A and 14B due to asymmetry in the pattern features 16.

[0047] Common metrology techniques are generally used to identify specific attributes of patterned samples. Such attributes include various geometric parameters as well as material characteristics. This is while the symmetry, or asymmetry, of patterns on the sample may be an important parameter, either in cases where asymmetry is considered as defect or when a certain level of asymmetry is desired.

[0048] As indicated above, the technique of the present invention provides for detection of asymmetry and preferably also measuring a level of asymmetry, and is exemplified herein with reference to reflection symmetry (also termed mirror-symmetry), i.e. symmetry of a pattern with respect to certain plane. More specifically, the technique of the invention allows for measuring the asymmetry level (shift value and direction of shift, e.g. for overlay measurements) e.g. with respect to a plane determined by a major axis of the arrangement of the pattern features. For example, in order to identify misalignment along one axis, a pattern or grating on the sample along said axis is measured. Determining misalignment along both the x and y axes requires measurements on gratings, or grating-like structures along both axes. As indicated above, one or more intermediate layers may exist between the two patterned layers, while the intermediate layer(s) is/are of less interest.

[0049] Thus, the technique of the present invention is based on measurements of optical response (reflection and/or diffraction) of a patterned sample to incident light, being monochromatic or polychromatic light. In some embodiments, the technique utilizes a set of two or more measurements on the same region of a sample, for example where light of different polarization states, with respect to that of the incident light, is detected. In some embodiments, the measurements are applied to several (generally, at least two) predetermined locations on the sample, for example within test sites.

[0050] Reference is made to FIG. 2A schematically illustrating, by way of a block diagram, a measurement system 100 of the invention configured and operable for identifying asymmetry of a patterned sample and preferably determining a level of asymmetry. The measurement system 100 includes an optical unit 102 accommodated with respect to a sample holder 115, and connectable, via wires or wireless signal transmission, with a control unit 140. The optical unit 102 defines an illumination channel 12 and a collection channel 13, and includes a light source unit 120 configured to illuminate at least a region on the surface of a sample through the illumination channel 12, a detection unit 130 configured to receive light returned (reflected) from the sample and propagating along the collection channel 13. Further provided in the optical unit 102 is a polarization unit 132 associated with at least the collection channel 13, i.e. includes one or more polarizer filters located at least in the collection channel 13.

[0051] The control unit 140 is typically an electronic device including inter alia such software/hardware utilities/modules as data input (or data readout) and output 144 and 143, memory 145, and data processor 148, and also includes a polarization filter controller 142. The control unit 140 may also include an illumination controller 147, as well as a position controller 149. The control unit 140 is thus configured to operate the optical unit (i.e. its polarization unit 132 and possibly also the light source unit 120), and receive and process measured data provided by the detection unit 130 as will be described further below.

[0052] The optical unit 102 is preferably configured (i.e. its elements are accommodated with respect to the sample holder and with respect to one another) to orient the illumination and collection channels 12 and 13 such that the collection channel collects light including specular reflection from the sample. Additionally, the optical unit 102 may be configured to orient the illumination and collection channels with an oblique light propagation arrangement with respect to the sample plane such that s and p polarization states can be well defined.

[0053] The polarization unit 132 includes at least one polarization filter 135, located in the collection channel 13 which may provide a linear or circular polarization. In this connection, a circular polarization filter is an optical assembly configured to allow transmission of either clockwise or counterclockwise polarization component of light. Such circular polarization filter may include a linear polarizer located in between two quarter-wave plates, and may be oriented such that only clockwise (or counterclockwise) polarization component pass through the polarizer assembly. The control unit 140 includes the polarization controller 149 connectable to the polarization filter 135 (generally, to the polarization unit 132) and configured to controllably vary the orientation of its preferred plane of polarization between different polarization states (at least a first polarization state and a second polarization state).

[0054] Additionally, in some embodiments, the polarization unit 132 also includes one or more polarization filters 125 located in the illumination channel 12. The illumination-channel polarization filter 125 may be generally similar to the collection-channel polarization filter 135 but differing in the orientation of its preferred plane of polarization and thus differing in the polarization component it transmits. The control unit 140 (its polarization controller 149) may also be connectable to the polarization filter 125 and configured to cause the polarization filter 125 to rotate or shift between at least first and second orientations, thereby transmitting light of a first or second polarization state.

[0055] The control unit may include the illumination controller 147 to operate the light source unit 120 to provide illumination of a selected type (wavelengths, illumination pattern, and/or operation mode, i.e. CW or pulse mode), and may include the position controller 149 for operating a relative accommodation between the sample holder 115 and at least some elements of the optical unit 102, e.g. for providing measurements on preset or desired locations on the sample.

[0056] For example, the control unit 140 may operate the light source unit 120 to sequentially illuminate the sample 110 with predetermined first and second different polarization states of light, or unpolarized and polarized light. The polarization filter controller 142 may accordingly operate the collection-channel polarizer 135 and possibly also the illumination-channel polarizer 125 to allow sequential detection of predetermined polarization components of reflected light, and generate first and second measured data indicative thereof (i.e. indicative of intensity of the corresponding polarization component in the detected light).

[0057] The data processor utility 148 of the control unit includes a symmetry detection module 146 (software utility) which is adapted (preprogrammed) to analyze the first and second measured data pieces, and generate data indicative of existence of asymmetry in the pattern being measured. In some embodiments, the symmetry detection module also operates to determine a relation between the first and second measured data pieces (e.g. difference), and estimate a level of asymmetry of the pattern on the sample being measured.

[0058] The invention provides a measurement scheme which is preferably based on spectral reflectometry, where the light response of the sample over a broad spectral range is measured. Preferably, illumination is at oblique angle to the sample, and collection is correspondingly aligned to collect specularly reflected light (i.e. at the same angle as illuminating channel angle). Two (or more) measurements are applied, in one or more orientations of the analyzer (polarizer filter 135 in the collection channel) and possibly also the polarizer 125 in the illumination channel. The data processing scheme is aimed at highlighting the information on the sample's asymmetry. For example, the illumination channel may be operated to sequentially provide illumination of the same region with two or more different polarization states of incident light, while the collection channel is operated to sequentially collect one or more different polarization components of the reflected light, selected in accordance with the polarization state of the incident light.

[0059] Let us consider a patterned sample including at least one grating structure (pattern) along an axis being a longitudinal grating axis. This is illustrated schematically in FIG. 2B. In this specific but not limiting example, the unpolarized light source 120 is used, and the polarizing unit 132 includes the illumination-channel polarizer 125 and the collection-channel analyzer 135. The use of the technique of the invention with this measurement scheme, utilizing various (at least first and second) orientations of the preferred plane of polarization of the illuminated and detected light components, provides for determining data indicative of a level of asymmetry of the pattern with respect to reflection about a plane perpendicular to the longitudinal grating axis.

[0060] The first and second polarization states of the illumination may be controlled by orientation of the polarization filter 125 located in the illumination channel 12. As will be described further below, according to some embodiments, the first and second different polarization states of illumination may be obtained by circular or elliptical polarization of light. The illumination channel polarization filter 125 may be generally similar to the collection-channel polarization filter 135 and may be controlled by the polarization filter controller 142 to sequentially allow collection of light components of first or second selected polarization states.

[0061] Reference is made to FIG. 3 exemplifying a method of the invention for detection of the condition of asymmetry and possibly also determining a degree of symmetry/asymmetry of the surface pattern/grating on the sample. As described above, in some embodiments, the invention utilizes a plurality, generally at least two, reflectivity measurements on the same region of a sample. The two measurements utilizes (at least) detection of light of two different polarization states, respectively, which are reflected from the sample in response to illumination that could be unpolarized or having different polarization states for two measurements.

[0062] As shown in the figure in a self-explanatory manner, first and second measurement sessions 1010 and 1110 are applied to the same region on the sample. It should be understood that the measurement sessions are different from one another in the detection of light of different polarizations states, both different from that of the illumination, while the two measurement session may be performed concurrently in case the illumination is unpolarized in both measurements, or sequentially, in case the illumination is performed sequentially 1010a and 1110b with first and second different polarization states respectively. Thus, during the measurement sessions 1010 and 1110 light with different polarization states is detected being modified by a sample under measurements in response to unpolarized or polarized illumination; and first and second measured data pieces 1020 and 1120 indicative of the intensity for the first and second detected specular reflection light components of different polarization states are obtained, for the two measurement sessions respectively. In case of unpolarized illumination, detection of light having different polarization states could be performed in parallel, using static polarizers and concurrently collecting ordinary and non-ordinary beams, or sequentially, using rotatable/movable polarization unit. The first and second measured data pieces are received, 1030 and 1130 (in parallel or sequentially), at the control unit and analyzed to identify the asymmetry condition: a change in the polarization state of the detected light with respect to that of the incident light; and preferably, a difference (generally, a relation, being in the form of a predetermined function of the measured data pieces) between the two measured data pieces measurements is determined (step 1210), and information indicative of a level of asymmetry of the pattern in the inspected region is provided (step 1220). It should be noted that an asymmetry of the pattern on a sample can be identified if the difference between the first and second measured data pieces is above a predetermined threshold.

[0063] As indicated above, the principles of the invention are based on the inventors' understanding of the reflective properties of patterned surfaces. Generally, reflection of polarized light can be described by the Jones formalism: when polarized light (fully polarized light), represented by Jones vector, crosses a linear optical element represented by Jones matrix the resulting polarization of the emerging light is found by taking the product of the Jones matrix of the optical element and the Jones vector of the incident light. The transverse electromagnetic (EM) field can be specified by the Jones vector

[00001]$\begin{array}{cc}\stackrel{.fwdarw.}{p}=\left(\begin{array}{c}{E}_s.Math.e^{i\left(\omega .Math.t+{\phi }_s\right)}\\ E_p.Math.e^{i\left(\omega .Math.t+{\phi }_p\right)}\end{array}\right)& \left(\mathrm{equation}.Math..Math.1\right)\end{array}$

where E.sub.S, E.sub.p are the field amplitudes in the s (perpendicular to the incidence plane) and p (parallel to the incidence plane) directions and ω, ϕ.sub.S, ϕ.sub.p are the angular frequency and phases of the EM wave. Additionally, scattering properties of a patterned surface (e.g. in the inspected region of the sample) can be described by the Jones matrix

[00002]$\begin{array}{cc}\hat{j}=\left(\begin{array}{cc}{r}_{\mathrm{ss}}& r_{\mathrm{sp}}\\ r_{\mathrm{ps}}& r_{\mathrm{pp}}\end{array}\right)& \left(\mathrm{equation}.Math..Math.2\right)\end{array}$

The Jones matrix is a 2×2 complex matrix, holding reflection amplitude and phase for each element. Here, r.sub.ij is the complex reflectance component for incident polarization j and outgoing polarization i. For example, r.sub.ss is the complex reflected field in the s direction (perpendicular to the plane of incidence) when the incident field of linear polarization s. Similarly, r.sub.sp represents the reflected s polarized field for incident field linearly polarized in the p direction (parallel to the plane of incidence). The Jones matrix is thus related to the degree of polarization conversion created by the sample.

[0064] As described above, the measurement system 100 according to some embodiments of the present invention exemplified in FIGS. 2A and 2B, includes a light source 120 followed by a polarization filter 125, for illuminating a sample with a specifically polarized light. Light specularly reflected from the sample is collected through an analyzer (linear polarization filter 135) onto a detector 130. This setup including Polarizer.fwdarw.Sample.fwdarw.Analyzer can be mathematically formulated by the expression:

[00003]$\begin{array}{cc}{E}_{\mathrm{out}}=\left(\mathrm{cos\beta }.Math.-\sin .Math.\beta \right).Math.\left(\begin{array}{cc}\left|r_{\mathrm{ss}}\right|& \left|r_{\mathrm{sp}}\right|e^{{i.Math.\varphi }_1}\\ \left|r_{\mathrm{ps}}\right|e^{{i.Math.\varphi }_2}& \left|r_{\mathrm{pp}}\right|e^{{i.Math.\varphi }_3}\end{array}\right).Math.\left(\begin{array}{c}\mathrm{cos\alpha }\\ \mathrm{sin\alpha }\end{array}\right)& \left(\mathrm{equation}.Math..Math.3\right)\end{array}$

where the measured intensity is |E.sub.out|.sup.2, α and β are the polarizer and analyzer angles with respect to the incidence plane, respectively, and

[00004]$\hat{j}=\left(\begin{array}{cc}\left|r_{\mathrm{ss}}\right|& \left|r_{\mathrm{sp}}\right|e^{{i.Math.\varphi }_1}\\ \left|r_{\mathrm{ps}}\right|e^{{i.Math.\varphi }_2}& \left|r_{\mathrm{pp}}\right|e^{{i.Math.\varphi }_3}\end{array}\right)$

is the Jones matrix of the inspected region of the sample.

[0065] It should be noted that the phase components of the complex coefficients of the Jones matrix are explicitly shown here for simplicity. It should be noted that the phase may be arbitrarily selected to be zero for the r.sub.ss coefficient while the relative phase between different components of the EM radiation has a specific measurable meaning.

[0066] A sample having a symmetric pattern/structure has reflection symmetry with respect to at least one plane. For example, grating lines are symmetric to reflection, but tilted grating lines, i.e. having different left and right SWA as exemplified in FIG. 1A, are not symmetric to reflection. A patterned surface that is symmetric to reflection can be described by polarization rotation coefficients of the Jones matrix satisfying the condition r.sub.ps+r.sub.sp=0. Accordingly, determination of the quantity r generally known as the asymmetry coefficient, provides a sensitive indication on asymmetry level of the inspection region/sample [5].

[0067] Reference is now made to FIGS. 4A to 4D illustrating four examples of the measurement technique of the present invention. These embodiments utilize different orientations of the polarization states of incident and detected light components to determine the level of asymmetry of a pattern in a region of the sample being measured.

[0068] FIG. 4A illustrates an example of first and second measurement sessions on a patterned region of a sample. As shown in the figure, in a first measurement session, denoted 100A, a spot 32A on the sample's surface is illuminated. Light reflected from this region 32A is collected through the polarization filter (analyzer 135 shown in FIGS. 2A and 2B) having certain first orientation of its preferred plane of polarization (first polarization state) 34A. In a second measurement session, denoted 100B, the same region 32A of the sample is illuminated and the reflected light is collected through the analyzer at a different second orientation of its preferred plane of polarization (second polarization state) 34B.

[0069] In some embodiments of the present invention, the illumination is unpolarized in both the first and second measurement sessions. Thus, the illumination region 32A is substantially the same in the first and second measurement sessions. In this configuration, the collection channel utilizes a linear polarization filter (analyzer) configured and operable to selectively pass light of the first and second polarization states. In the first measurement, the analyzer is oriented to allow passage of light at 45 degrees to the p polarization (generally at 1/√{square root over (2)}[p+s]), while in the second measurement, the analyzer is oriented to allow passage of light at −45 degree (generally at 1/√{square root over (2)}[p−s]). Utilizing the above described equation 3 and the Jones matrix, the detected intensity is provided by:

[00005]$\begin{array}{cc}{E}_{\mathrm{out}}=\left(\mathrm{cos\beta }.Math.-\sin .Math.\beta \right).Math.\left(\begin{array}{cc}\left|r_{\mathrm{ss}}\right|& \left|r_{\mathrm{sp}}\right|e^{{i.Math.\varphi }_1}\\ \left|r_{\mathrm{ps}}\right|e^{{i.Math.\varphi }_2}& \left|r_{\mathrm{pp}}\right|e^{{i.Math.\varphi }_3}\end{array}\right).Math.\left(\begin{array}{c}1/2\\ 1/2\end{array}\right)& \left(\mathrm{equation}.Math..Math.4\right)\end{array}$

[0070] The measured intensity in the first and second measurements is shown in Table 1 below. As can be seen, the measured intensities provide indication on the level of asymmetry of the inspected region as these intensities relate to the above described quantity δ.

TABLE-US-00001 TABLE 1 Measurement # β |Ε.sub.out|.sup.2 1 π/4 |r.sub.ss + r.sub.pp + r.sub.sp + r.sub.ps|.sup.2 = |R + δ|.sup.2 2 −π/4 |r.sub.ss + r.sub.pp − r.sub.sp − r.sub.ps|.sup.2 = |R −δ|.sup.2

[0071] As described above, δ≡r.sub.sp+r.sub.ps and R≡r.sub.ss+r.sub.pp. Determining the difference between the first and second measured intensities provides:

|R+δ|.sup.2−|−δ|r.sup.24{δR°}32 2(r.sub.sp+r.sub.ps)(r.sub.ss+r.sub.pp)°+c.c.  (equation 5)

Here, and along the description, the term c.c. describes the use of complex conjugate of the previous term.

[0072] Generally, this embodiment may utilize simpler optical setup (including less optical elements, as no polarizer is used in the illumination channel), and may be beneficial in terms of light budget (lack of polarizer in the illumination channel increases the amount of light incident onto the sample). It should, however, be noted that the illumination light is preferably unpolarized, and the use of elliptical polarized light may cause variations in the measured data.

[0073] FIG. 4B illustrates another embodiment of the technique of the invention. In this example, assessment of asymmetry is accomplished as follows: The sample's reflectivity is first measured with a polarizer in the illumination path and an analyzer in the collection path aligned at 45 and −45 degrees, respectively, to the p polarization (parallel to the plane of incidence). The second measurement is then performed with the polarizer and analyzer aligned at −45 and 45 degrees respectively to the p polarization. More specifically, in this example, the illumination spots 42A and 42B in the first and second measurements are formed on the same region of the sample by light of linear polarizations, oriented at ±45 degree respectively with respect to the p polarization. The collection channel utilizes the analyzer at the orthogonal polarization orientations 44A and 44B, i.e. if the sample is illuminated at 1/√{square root over (2)}[p+s] the analyzer is configured to pass light at and vice versa. Utilizing the above equation 3, the level of asymmetry of the inspected region can be determined in accordance with the following Table 2, which projects the measurement configuration on Jones matrix as described above.

TABLE-US-00002 TABLE 2 Measurement # α β |Ε.sub.out|.sup.2 1 π/4 −π/4 |r.sub.ss + r.sub.pp + r.sub.sp + r.sub.ps|.sup.2 = |R + δ|.sup.2 2 −π/4 π/4 |r.sub.ss + r.sub.pp − r.sub.sp − r.sub.ps|.sup.2 = |R − δ|.sup.2

[0074] The difference in measured intensity between the first and second measurements in this embodiment is similar to the above equation 5. Similarly, the resulting term of equation 5 is directly proportional to δ and thus provides a direct measure to estimate the asymmetry of the inspected region of the sample. As mentioned above, symmetrical pattern on a sample can be identified if the difference between the first and second measurements is below a predetermined threshold. Thus, if a certain level of symmetry is desired; the sample is considered appropriate if the measured quantity has the desired value within certain error limits.

[0075] In the example of FIG. 4C, the first and second measurements 100A and 100B are performed using circular polarization for the illumination and collection channels. In this embodiment, the first and second measurements are done utilizing a circular polarization filter in the illumination channel providing counterclockwise and clockwise polarized illumination 52A and 52B, and analyzer in the collection channel is oriented to allow passage/collection of light at clockwise and counterclockwise circular polarizations 54A and 54B.

[0076] In a similar manner to the above, a difference between the first and second measured data pieces (intensities) is directly indicative of a level of reflection symmetry of the inspected region. Moreover, the determined data provides sensitivity to a direction of asymmetry indicating a direction to which the pattern deviates from symmetry. This configuration provides lower sensitivity to actual orientation of the polarization filters in the collection and illumination channels.

[0077] FIG. 4D exemplifies the principles of some other embodiments of the present invention. In this example, the measured patterned region is illuminated with linearly polarized light 62 having either one of p and s polarization states, reflected light is collected through the analyzer, which is configured to allow collection of the orthogonal linear polarization 64, i.e. s or p states respectively. It should be understood, however, that the polarization states of the illuminating and collected light may be interchangeable. Generally, one polarizer is aligned along the s polarization axis and the other is aligned along the p polarization axis. In this embodiment, the illumination-detection plane, defined by the illumination and collection channels, is either parallel or perpendicular to the gating lines of the pattern on the sample; and the sample is illuminated with a polarization parallel or perpendicular to the grating direction, and the collection-channel analyzer is oriented perpendicular to the illumination polarizer.

[0078] If the pattern on the measured region of the sample is symmetric to reflection, the off diagonal Jones elements r.sub.sp, r.sub.ps vanish due to symmetry, and the reflected light will be blocked on the return path. However, variations from symmetry (asymmetric patterns) cause rotation in polarization of the reflected light, and give rise to nonzero signal at the detector. Thus, a single measurement may provide data indicative of the symmetry of the pattern.

[0079] Table 3 illustrates the measurement configuration based on the above described Jones matrix terms. Table 3 shows the resulting data for a measurement (measurement 1) where the illumination and collection channels are configured to allow illumination at p polarization state and collection of the reflection of only s polarization state, as well as data for the orthogonal measurement (measurement 2) where the illumination- and collection-channels' polarization filters are oriented to illuminate with s polarized light and collect the p polarized light. As shown, each of these measurements provides data about the magnitude of one of the off-diagonal elements of the Jones matrix. It should be noted that although this embodiment has an advantage of providing symmetry data in a single measurement, the resulting information does not inherently include data about direction of deviation from symmetry, as provided with the use of differential measurements.

TABLE-US-00003 TABLE 3 Measurement # α β |Ε.sub.out|.sup.2 1 0 π/2 |r.sub.ps|.sup.2 2 π/2 0 |r.sub.sp|.sup.2

[0080] Additionally, the sample may be inspected utilizing several, different, incident angles to provide higher sensitivity by providing data about the level of symmetry for different directions. In this connection, variation of the incident angle may provide additional information about the level of symmetry for reflection about a specific plane, while variation of the incident azimuth provides additional data about the level of symmetry for reflection about additional planes. Thus, a set of consecutive measurements according to the above described technique can provide data about the level of symmetry of a sample having complex patterns including various layers having patterns along one or more axes, thereby enabling characterization of symmetry properties with respect to one or more planes.

[0081] It should also be noted that the technique of the invention may be implemented utilizing reflected light at the above defined polarization configurations in combination with a dome spectrometer system, for example described in WO14016839 assigned to the assignee of the present application, which publication is incorporated herein by reference with respect to non-limiting examples of the use of the present invention. This can provide broad angular information, i.e. indicate level of symmetry about different planes and/or angles.

[0082] As indicated above, the technique of the present invention provides efficient and accurate information about the level of symmetry of a patterned sample. The advantages of the technique can be demonstrated utilizing measurement on a known sample, which is schematically illustrated in FIG. 5. The figure illustrates a litho-etch-litho-etch manufactured sample 500 including two printed patterns with a predetermined overlay. The first pattern 510 is printed on a wafer, the wafer is then transferred into silicon and then a second pattern 520 is printed with half-period shift to allow a second etch process to transfer it into silicon for a pitch-doubling effect. The overall pattern has a pitch size of 2400 Å. The technique of the invention was used to detect overlay shifts between the first litho-etch step (layer 510) and the second lithography pattern 520. As indicated, the parameter for characterization in such patterned samples is the degree of alignment in between the layers 510 and 520. When aligned perfectly, the structure is mirror-symmetric, while misalignment would break this symmetry. It has been shown that measurement using a standard polarized reflectance is very difficult due to the low spectral sensitivity to the overlay shift.

[0083] FIG. 6 shows measurement results of misalignment (A) between the patterned layers 510 and 520 of the sample of FIG. 5 as measured in different reflected spectra. Graph A corresponds to the measurement results for the conventional technique of spectral reflectometry (SR) using p-polarized illumination. Graph B corresponds to the measurement results of spectral differences metrology (SDM) according to the present invention where the illumination and collection channels utilize linear polarizer and analyzer as described above with reference to FIG. 4B. Graph C shows the results provided by the technique of the present inventions utilizing circular polarization as exemplified above with reference to FIG. 4C where single measurement provides for detecting the asymmetry condition of the pattern.

[0084] As shown, the effects of misalignment on the measured spectra using the conventional SR technique (graph A) are weak, and may generally be completely overwhelmed by noise, the measurement had signal to noise ratio (SNR) of about 100. As for the results for the SDM technique of the invention, graphs B and C, they show clear sensitivity to misalignment, well above the noise threshold.

[0085] It should also be noted that the technique of the present invention as described above provides for determination of deviations from symmetry of the pattern, while having low sensitivity to uncertainties unrelated to symmetry. It is known that Optical Critical Dimensions (OCD) measurements commonly suffer from difficulty in separating between variations of the parameters of the structure (in this case, the misalignment of the patterned layers) and other parameters characterizing the application (critical dimensions, thicknesses), which may have significant effect on the measured spectrum but are of no practical interest. The spectra obtained using the technique of the invention specifically highlights sensitivity to misalignment, while being insensitive to the other parameters characterizing the application.

[0086] The signal obtained using any of these techniques can be analyzed in one of several ways. One example would be to follow the standard approach utilized in OCD metrology, namely to compare the measured signal with a calculated (model-based) signal. The calculated signal corresponds to some structure with known geometrical attributes. Once a good fit is obtained between the calculated and measured signals, it is assumed that the geometrical and material properties of the best-fit calculated structure correspond to those characterizing the measured structure. Calculation of the reflected signal can be based on one of many optical modeling techniques (e.g. Rigorous Coupled Wave Analysis, Finite Element Method, Finite Difference Time Domain etc.). It is furthermore possible that the calculation effort involved in interpretation of measurements taken using the approach described in this invention will be significantly simplified, compared to standard OCD. Whereas for most OCD applications, the reflected signal depends on multiple parameters characterizing the measured target, in this invention the dependence on most parameters will be weak, and the dependence on some geometrical \ material properties could be neglected.

[0087] An alternative approach would be based on an empirical approach. In this method, in a first step (termed the ‘learning’ stage), signal will be measured for few samples of known characteristics (characterized by some other reliable metrology method). A correlation will be established between the measured signal and the geometrical property of interest. For example, the analysis can establish a relationship between the amplitude or position of some specific spectral peak and the parameter of interest. Samples measured following this step will be analyzed based on the found correlation. The fact that the method presented in this invention is highly effective in highlighting the dependence to the parameters of interest makes such empirical approach realizable, using a small set of samples for the learning stage.

[0088] Thus, the present invention provides for a novel technique enabling simple and robust analysis of symmetry properties of patterned substrates. Those skilled in the art will readily appreciate that various modifications and changes can be applied to the embodiments of the invention as hereinbefore described without departing from its scope defined in and by the appended claims.