This disclosure relates to interferometry, and more particularly to low coherence interferometry.
Interferometry (e.g., low coherence interferometry) may be used to determine a spatial property of an object. Typical spatial properties include a surface topography or location of the object with respect to some reference. For objects including a thick film overlying an opaque substrate, the low coherence interferometry data may include two spaced part interference patterns resulting, respectively, from the substrate-film interface and film-air interface. The accuracy of interferometry depends on the quality of the interference signal, which can be degraded by several sources of error, many of these systematic and related to the limitations of the instrument.
An example of low coherence interferometry is scanning white light interferometry (SWLI), which involves using a light source with a large bandwidth (e.g., about 100 nm or more) compared with other forms of interferometry (e.g., displacement measuring interferometry) which utilize a monochromatic light source. Large bandwidths, however, can degrade the quality of a SWLI signal because of optical aberrations, dispersion and other sources of error, reducing the accuracy of SWLI measurements and/or reducing the number of different objects that can be characterized. For example, when using SWLI to characterize thin films, signal degradation makes it increasingly difficult to extract information about a film from overlapping interference patterns due to the two interfaces of the film.
This disclosure features methods and systems that improve the quality of SWLI signals. Embodiments include methods and systems that improve the quality of SWLI signals by a correction (e.g., a software correction) for system errors, including, in particular, source spectrum distortions and nonlinear dispersion in the interferometer. In embodiments, these methods operate on a SWLI signal to produce a corresponding repaired signal in which the effects of system errors are reduced (e.g., eliminated).
In embodiments, in a first step, a system characterization procedure determines the system errors by direct measurement of these errors in the frequency domain after a frequency transform (e.g., a Fourier Transform), in terms of the amplitudes and phases of the transform coefficients. In a second step, SWLI data from a sample to be measured are Fourier transformed into the frequency domain, where the known phase and amplitude system errors are removed by subtraction, division or other data manipulation.
If the data processing proceeds in the frequency domain, the corrected data are transmitted to the next processing stage for determining a characteristic of the object surface. Examples of processing in the frequency domain are discussed in U.S. Publication No. 2004/0189999 A1, entitled “PROFILING COMPLEX SURFACE STRUCTURES USING SCANNING INTERFEROMETRY,” by De Groot et al., the entire contents of which is hereby incorporated by reference.
Alternatively to processing in the frequency domain, the data can be transformed back by means of an inverse Fourier Transform to the original signal domain prior to the next processing stage for determining a characteristic of the object surface. Examples of processing in the original signal domain are discussed in U.S. Publication No. 2005/0078318 A1, entitled “METHODS AND SYSTEMS FOR INTERFEROMETRIC ANALYSIS OF SURFACES AND RELATED APPLICATIONS,” by De Groot, the entire contents of which is hereby incorporated by reference.
Various aspects of the invention are summarized as follows.
In general, in one aspect, a method is disclosed that includes: (i) acquiring a first data set from a low coherence interference signal using an imaging interferometry system, the low coherence interference signal including information about a measurement object; (ii) using the first data set to generate a second data set, where the second data set is related to the first data set by a frequency transform; and (iii) using the second data set to generate a third data set, where the third data set corresponds to the second data set with reduced contributions due to imperfections in the interferometry system.
Embodiments of the method may include one or more of the following features.
The method may further include determining information about the measurement object from the third data set. For example, determining the information about the measurement object from the third data set may include using the third data set to generate a fourth data set, where the fourth data set is related to the third data set by an inverse frequency transform. For example, the fourth data set may be in the same form as the first data set. Furthermore, the first and fourth data sets may be scanning white light interferometry signal and a repaired scanning white light interferometry signal, respectively.
The information about the measurement object may include information about a surface of the measurement object. For example, the method may further include using the information about the surface to determine a profile of the surface.
The measurement object may include a first portion of a first material and a second portion of a second material and the information about the measurement object may include information about an interface between the first and second portions.
The measurement object may include a film of a material having a thickness of about 10 μm or less (e.g., about 1 μm or less, or even 0.5 μm or less, or even 0.25 μm or less) and the method may further include using the information to determine a profile of the film's thickness.
The imperfections in the interferometry system may include imperfections in an output spectrum of a light source of the interferometry system. For example, the imperfections in the output spectrum may include deviations of the spectrum from a model spectrum. Furthermore, the method may further include determining information about the measurement object from the third data set based on an assumption that the low coherence interference signal was acquired using the model spectrum. The model spectrum may include a single carrier frequency modulated by a Gaussian envelope or a raised cosine.
The frequency transform may be a Fourier transform.
The third data set may be determined using predetermined information acquired using the imaging interferometry system. For example, the predetermined information may be determined from a first artifact data set acquired from a low coherence interference signal from an artifact using the imaging interferometry system. Furthermore, determining the predetermined information may include transforming the artifact data set into a second artifact data set, where the second artifact data set is related to the first artifact data set by a frequency transform.
Determining the third data set may include multiplying each element, qv, of the third data set by a corresponding coefficient, Cvfix, to provide a corresponding element, q′v, of the third data set.
In another aspect, a method is disclosed that includes: (i) acquiring a first data set from a low coherence interference signal using an imaging interferometry system, the low coherence interference signal including information about a reference object; (ii) using the first data set to generate a second data set, where the second data set is related to the first data set by a frequency transform; (iii) determining a third data set from the second data set, where the third data set relates the second data set to a model data set, wherein the model data set corresponds to a data set acquired using a model interferometry system having reduced imperfections relative to the interferometry system used to acquire the first data set; and (iv) using the third data set to reduce effects of imperfections in the imaging interferometry system on low coherence interference signals acquired using the imaging interferometer system.
In another aspect, a system is disclosed including: (i) a light source; (ii) an imaging interferometer configured to receive light from the light source, to direct a portion of the light to reflect from a measurement object, to direct another portion o the light to reflect from a reference object, and to combine the light reflected from the measurement and reference objects to produce an output beam; (iii) a detector configured to detect the output beam and to produce an interference signal including information about the measurement object; and (iv) an electronic processor configured to acquire a first data set from the interference signal, to generate a second data set related to the first data set by a frequency transform, and to generate a third data set corresponding to the second data set with reduced contribution due to imperfections in the light source or the interferometer.
Embodiments of the system may include any of the following features.
The electronic processor may be further configured to carry out any of the features corresponding to embodiments of the method described above.
The light source may have an output spectrum having a full width at half maximum (FWHM) of about 50 nm or more, or even 100 nm or more.
The light source may include a light emitting diode, or an arc lamp. For example, the light source may be in the visible portion of the electromagnetic spectrum.
The interferometer may be a Mirau interferometer, or a Linnik interferometer.
The interferometer may have a numerical aperture at the object of about 0.1 or more, or even 0.3 or more.
In general, in another aspect, the invention features a method that includes transforming interferometry data acquired for a test sample using a low coherence imaging interferometry system to a frequency domain and, at a plurality of frequencies in the frequency domain, reducing contributions to the transformed interferometry data due to imperfections in the imaging interferometry system thereby producing compensated interferometry data. The errors are reduced based on variations between interferometry data acquired using the low coherence imaging interferometry system for a calibration sample and model interferometry data corresponding to data acquired for the calibration sample using a model interferometry system.
Implementations of the method can features of other aspects and/or one or more of the following features. For example, the imperfections in the interferometry system can include deviations of an output spectrum of a light source of the interferometry system from a model spectrum of the model interferometry system. The model spectrum can correspond to an analytic function. The analytic function can include a Gaussian function or a raised cosine function.
In some embodiments, the imperfections in the interferometry system include imperfections in optical components in the interferometry system. The imperfections in the optical components can result in non-linear dispersion in the light used by the interferometry system to acquire the interferometry data.
The interferometry data can be transformed to a frequency domain using a Fourier transform. The transformed interferometry data can be characterized by a Fourier coefficient at each of the plurality of frequencies and producing the compensated interferometry data comprises changing the phase of the Fourier coefficient at one or more of the plurality of frequencies. The transformed interferometry data can be characterized by a Fourier coefficient at each of the plurality of frequencies and producing the compensated interferometry data comprises changing the magnitude of the Fourier coefficient at one or more of the plurality of frequencies.
In certain embodiments, the interferometry data can be acquired simultaneously for a plurality of locations of the test sample using a detector array.
The test sample can include a thin film supported by a substrate. The substrate can be a semiconductor substrate.
The calibration sample can be a SiC calibration sample.
In some embodiments, the method includes determining information about the test sample from the compensated interferometry data after reducing the errors. Determining information about the test sample can include transforming the compensated interferometry data into a co-ordinate space domain and analyzing the compensated interferometry data in the co-ordinate space domain. In certain embodiments, determining information about the test sample includes analyzing the compensated interferometry data in the frequency domain. The method can include outputting the information about the test sample. For example, outputting the information can include displaying the information or writing the information to memory. In some embodiments, the information about the test sample includes information about a surface of the test sample. The information about the surface can include a profile of the surface. In certain embodiments, the test sample includes a first portion of a first material and a second portion of a second material and the information about the test sample includes information about an interface between the first and second portions. The test sample can include a film of a material having a thickness of about 10 μm or less (e.g., about 5 μm or less, about 2 μm or less, about 1 μm or less, 0.5 μm or less, 0.25 μm or less) and the information about the test sample comprises a profile of the film's thickness.
In general, in a further aspect, the invention features a method that includes transforming interferometry data acquired for a test sample using a low coherence interferometry system to a frequency domain and, at a plurality of frequencies in the frequency domain, reducing contributions to the transformed interferometry data due to non-linear dispersion in the interferometry system thereby producing compensated interferometry data. The errors are reduced based on variations between interferometry data acquired using the low coherence interferometry system for a calibration sample and model interferometry data corresponding to data acquired for the calibration sample using an interferometry system without the non-linear dispersion. Implementations of the method can include one or more features of other aspects.
In general, in a further aspect, the invention features a system that includes a low coherence light source, an imaging interferometer configured to receive light from the light source, to direct a portion of the light to reflect from a test sample, to direct another portion of the light to reflect from a reference object, and to combine the light reflected from the measurement and reference objects to produce an output beam, a detector configured to detect the output beam and to produce an interference signal comprising information about the test sample, and an electronic processor in communication with the detector, the electronic processor being configured to transform interferometry data derived from the interference signal to a frequency domain and, at a plurality of frequencies in the frequency domain, reduce contributions to the transformed interferometry data due to imperfections in the imaging interferometer or the light source thereby producing compensated interferometry data. The errors are reduced based on variations between interferometry data acquired using the imaging interferometer for a calibration sample and model interferometry data corresponding to data acquired for the calibration sample using a model interferometry system.
Embodiments of the system can include one or more of the following features and/or features of other aspects. For example, the model interferometry system can have reduced imperfections in the imaging interferometer or the light source. The model interferometry system can have reduced imperfections in both the imaging interferometer and the light source. The imperfections in the light source can include deviations of the source spectrum from a model spectrum. The model spectrum can correspond to an analytic function. The analytic function can include a Gaussian function or a raised cosine function.
The imperfections in the imaging interferometer can include nonlinear dispersion in the interferometer.
The light source can have an output spectrum having a full width at half maximum (FWHM) of about 50 nm or more (e.g., about 100 nm or more, about 200 nm or more).
The light source can include a light emitting diode or an arc lamp. Light from the light source can be in the visible portion of the electromagnetic spectrum (e.g., in a range from about 400 nm to about 700 nm).
The interferometer can be a Mirau interferometer, a Linnik interferometer, or a Michelson interferometer.
The interferometer can have a numerical aperture at the object of about 0.1 or more (e.g., about 0.2 or more, about 0.3 or more, about 0.4 or more, about 0.5 or more).
Among other advantages, embodiments can extend the useable bandwidth of a light source in an interferometry system with little or no modification to the interferometry system hardware. For examples, usable spectral bandwidth of a light sauce may be extended (e.g., by up to 100% or more) without any substantial loss in the signal to noise ratio of the SWLI signal.
Broad spectral bandwidth light sources can improve resolution of interferometry measurements. Relative to comparable sources with narrower bandwidths, broad bandwidth sources improve resolution because they provide narrower SWLI signal modulation envelopes and increased ranges of constituent frequencies for frequency analysis. Accordingly, the resolution of surface profile measurements can be improved and signals from different interfaces can be more easily separated when measuring thin films.
We note that while SWLI is an acronym for scanning white light interferometry, as used in the present application, SWLI is meant to include broadband radiation that is in any of the ultraviolet, visible, near-infrared, or infrared regions of the electromagnetic spectrum. Further, the basics of SWLI signal processing extend, despite the nomenclature, to any interferometer in which coherence plays a role in the measurement; for example, a narrow-bandwidth interferometer having a high NA (numerical aperture) objective that has localized interference fringes by virtue of limited spatial coherence.
Implementations can be readily implemented with existing SWLI signal analysis algorithms. For example, algorithms can be implemented to correct the signal for hardware defects (e.g., source defects and/or dispersion) without performing other operations that change the form of the signal data. Accordingly, the algorithm output can be in the same form as the original SWLI signal data, allowing further signal processing of the output using algorithms designed to analyze the original SWLI signal data.
The interferometry systems that use the algorithms discussed herein can be used to characterize thin films, such as dielectric thin films that have a thickness of about 1 μm or less.
The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features and advantages of the invention will be apparent from the description, the drawings, and the claims.
Like reference symbols in the various drawings indicate like elements.
Referring to
In
The width of the coherence envelope 154 that modulates the amplitudes of fringes 152 corresponds generally to the coherence length of the detected light. Among the factors that determine the coherence length are temporal coherence phenomena related to, e.g., the spectral bandwidth of the source, and spatial coherence phenomena related to, e.g., the range of angles of incidence of light illuminating the object. Typically, the coherence length decreases as: (a) the spectral bandwidth of the source increases and/or (b) the range of angles of incidence increases. Depending upon the configuration of an interferometer used to acquire the data, one or the other of these coherence phenomena may dominate or they may both contribute substantially to the overall coherence length. The coherence length of an interferometer can be determined by obtaining an interference signal from an object having a single reflecting surface, e.g., not a thin film structure. The coherence length corresponds to the full width half maximum of the envelope modulating the observed interference pattern.
As can be seen from
A low coherence interferometer can be configured to detect an interference signal is detected over a range of OPD's that is comparable to or greater than the coherence length of the interferometer. For example, the range of detected OPD's may be at least 2 times greater or (e.g., about 3 times or more, about 5 times or more, about 10 times or more greater than the coherence length). In some embodiments, the coherence length of the detected light is on the order of the height variations of features of the object, e.g., on the order of a couple of microns or less but more than a nominal wavelength of the detected light.
In general, instrument related contributions to the interference signal, e.g., to the shape and/or phase of the interference patterns, tend to vary slowly with the topography and position of the object. On the other hand, interference patterns shift along the scan position axis for interference signals obtained from object points having different spatial properties, e.g., different relative heights or different relative positions with respect to the interferometer. Accordingly, interference patterns obtained from different object points may have similar shapes but are shifted along the scan position axis by an amount related to the spatial properties of each point.
Referring to
Interference signal 190 includes a first interference pattern 196 resulting from interface 194 and a second interference pattern 197 resulting from interface 195. First and second interference patterns 196,197 are overlapping. For example, maxima of the interference patterns 196,197 are separated by an OPD less than the coherence length of the interferometer and patterns 196,197 are not separated by a region of zero intensity. Existing methods for determining spatial properties of an object with interfaces that produce overlapping interference patterns can yield erroneous results because the overlapping interference patterns distort one another. Applicants Spatial properties of an object with such interfaces can be determined based upon a portion over the overlapping interference patterns. For example, a spatial property of interface 195, e.g., a topography of the outer surface of object 191, can be determined based upon a subset 200 of interference signal 190. Subset 200 is dominated by contributions from interference pattern 197 (from interface 195) as opposed to contributions from interference pattern 196 (from interface 194). A spatial property of interface 194 can be determined based upon a subset similar to subset 200 but located toward the left of the overlapping patterns.
Examples of interferometers than can be configured as low coherence scanning interferometers include Michelson, Linnik and Mirau interferometers.
For simplicity,
In some embodiments, detector 120 is a multiple element (i.e., multi-pixel) camera to independently measure the interference between the measurement and reference light corresponding to different points on the test sample and reference mirror (i.e., to provide spatial resolution for the interference pattern).
A scanning stage 126 coupled to test sample 112 scans the position of the test sample relative to test objective 110, as denoted by the scan coordinate 4 in
Because the scanning occurs in a region where the measurement light is being focused onto the test sample, the scan varies the optical path length of the measurement light from the source to the detector differently depending on the angle of the measurement light incident on, and emerging from, the test sample. As a result, the optical path difference (OPD) from the source to the detector between interfering portions of the measurement and reference light scale differently with the scan coordinate ζ depending on the angle of the measurement light incident on, and emerging from, the test sample. In other embodiments of the invention, the same result can be achieved by scanning the position of reference mirror 116 relative to reference objective 114 (instead of scanning test sample 112 relative to test objective 110).
This difference in how OPD varies with the scan coordinate ζ introduces a limited coherence length in the interference signal measured at each pixel of the detector. For example, the interference signal (as a function of scan coordinate) is typically modulated by an envelope having a spatial coherence length on the order of λ/2(NA)2, where λ is the nominal wavelength of the illumination light and NA is the numerical aperture of the measurement and reference objectives. The modulation of the interference signal provides angle-dependent information about the reflectivity of the test sample. To increase the limited spatial coherence, the objectives in the scanning interferometer can define a large numerical aperture. For example, the objectives can define a numerical aperture of about 0.5 or more (e.g., about 0.7 or more, about 0.8 or more, about 0.9 or more). The interference signal can also be modulated by a limited temporal coherence length associated with the spectral bandwidth of the illumination source. Depending on the configuration of the interferometer, one or the other of these limited coherence length effects may dominate, or they may both contribute substantially to the overall coherence length.
Another example of a scanning interferometer is a Mirau interferometer shown in
The scanning in the embodiment of
Source module 205 includes a spatially extended source 201, a telescope formed by lenses 202 and 203, and a stop 204 positioned in the front focal plane of lens 202 (which coincides with the back focal plane of lens 203). This arrangement images the spatially extended source onto a pupil plane 245 of Mirau interferometric objective assembly 210, which is an example of Koehler imaging. The size of stop 204 controls the size of the illumination field on test sample 220. In some embodiments, the source module may include an arrangement in which a spatially extended source is imaged directly onto the test sample, which is known as critical imaging. Either type of source module may be used with other types of interferometer, such as the Linnik-type scanning interferometry system of
In certain embodiments, analysis of low coherence signals to determine information about a test sample involves comparing a measured interferometry signal (or information obtained from the measured interferometry signal) to a modeled signal (or information related to a signal obtained using a model). Many models assume that the spectrum of the interferometry system light source has a distribution corresponding to an analytic function, such as a single-frequency carrier modulated by a Gaussian envelope, for example. An example of a simulated Gaussian-modulated carrier is shown in
In general, light source spectra are more complex than the Gaussian-modulated carrier shown above. For example, in some embodiments, a light source spectra can include two or more peaks. Referring to
Furthermore, analysis of low coherence signals often does not fully account for optical imperfections in the interferometer. For example, interferometers generally introduce a phase delay for a given surface height value of a test sample surface as a function of wavelength and of the incident angle in the test sample illumination. This phase delay is referred to as dispersion. Dispersion can result in part from an imbalance in the amount of glass in the measurement and reference legs, a common difficulty with, e.g., Linnik interferometers. If dispersion is nonlinear, the net effect can be a loss of fringe contrast and a broadening of the coherence envelope in the low coherence signal.
As an example of the effect of nonlinear dispersion on an interference signal, consider the effect of dispersion on a simulated signal using the same light source spectrum shown in
The effects of imperfections (e.g., the source distortions and nonlinear dispersion discussed above) in interferometry data can be reduced as outlined below. Without wishing to be bound by theory, a discretely-sampled low coherence interferometry signal, I, can be expressed as an incoherent sum of interference patterns over a range of frequencies K:
The range of frequencies is a consequence of both the spectral bandwidth of the interferometry system source and geometric effects of incoherent illumination at nonzero NA. In Eq. (1), ζ is the interferometer scan coordinate orthogonal to the object surface, N is the number of samples taken during the scan, q is a Fourier coefficient, j is the index of the element of the detector array, and z, v are the indices for the scan position ζ0 and frequency K, respectively. Assuming uniform sampling over a scan range that encompasses the entire envelope of the interference signal I, the inverse relationship
that takes one from the from the original intensity data, Ij, to the corresponding Fourier coefficients.
Modeling predicts the Fourier coefficients by incoherent sums of constituent interference patters over all wavelengths λ, all incident angles Ψ and for both polarizations combined (see, e.g., P. de Groot, X. Colonna de Lega “Signal modeling for low coherence height-scanning interference microscopy,” Appl. Opt., 43(25), 4821 (2004)). For positive nonzero frequencies K>0, the Fourier coefficients can be expressed as
Where h is the location of the test sample surface along the z-axis and the values of the directional cosine β=cos(Ψ) for each angular wavenumber k=2π/λ in Eq. (3) satisfy
β=K/2ku. (4)
The discrete sum in Eq. (3) is over an equally-spaced grid of k>K/2 values indexed by u. Z is the effective object intensity reflectivity, including, e.g., the effects of the beamsplitter, and R is the effective reference reflectivity, including both the beamsplitter and the reference mirror. The phase terms each have a contribution ω for the object path in the interferometer, including thin film effects and such from the object surface, and a contribution υ for the reference path, including the reference mirror and objective optics. For convenience and without loss of generality, we can define υ as the system error and ω as the test sample phase change on reflection (PCOR). The weighting factor Γ accounts for the spectral and pupil plane distributions in the instrument.
The Fourier coefficients may be expressed by their magnitudes and phases:
qj,v=Pj,vexp(iφv). (5)
As is noted above in Eq. (3), the phase values φ are the net result of integral summations of complex exponentials over all wavelengths λ and over all incident angles Ψ that when combined have the same frequency K. Since the resultant phase φ is the argument of this sum of complex numbers, we cannot readily and generally separate the contributions for the system error υ from the part PCOR ω. The same caution applies to the magnitude P.
By making certain approximations, it is possible to separate system errors (for example, imperfections in the spectral distribution and a dispersion characteristics of the system) from characteristics of the test sample. Consider, for example, the limit cases to reduce the sum in Eq. (3) to a product of two complex numbers, ρsysρpart, one representing the system and the other the part characteristics, respectively:
qv=exp(iKvh)ρvsysρvpart (6)
This, in principle, makes the phase contributions additive and therefore separable.
One approximation that can allow the system and part errors to be separated is where the interferometry system uses monochromatic illumination in conjunction with an interferometer having a relatively high NA (e.g., about 0.5 or more). Here, the range of k values centered about a nominal k0 is assumed small enough that the terms of the sum may be considered constant for a given K, and therefore can be brought out in front of the sum. Thus, one can then write
Note that in this case, there is a direct correspondence between the Fourier frequencies K and the directional cosine β=K/2k0. Thus the Fourier coefficient magnitudes follow closely the optical spectrum, as shown by P. de Groot and X. Colonna de Lega in “Signal modeling for low coherence height-scanning interference microscopy,” Appl. Opt., 43(25), 4821 (2004).
Another possible approximation that can allow the system and part errors to be separated is where broadband illumination is used in conjunction with a relatively low NA interferometer. Here, there is a limited range of incident angles Ψ such that once again the terms in Eq. (3) are essentially constant over the limited range of k values consistent with the constraint of Eq. (4). Then
where β0 is the nominal directional cosine. Examples include low NA (e.g., β≈1) and ring sources, corresponding, e.g., to a restricted pupil encompassing the central obscuration of a high-NA Mirau. In this case, there is a direct correspondence between the Fourier frequencies K and the wavenumbers k=K/2β0. Thus the Fourier coefficient magnitudes follow closely the optical spectrum.
These two limit cases establish that it is feasible to some degree of approximation to separate the system and part characteristics as in Eq. (6).
Referring now to
The next step is to measure a low coherence interference signal for a calibration sample using the interferometry system. In some embodiments, the calibration sample is a dielectric that introduces little distortion to the optical spectrum and exhibits low nonlinear angle and wavelength-dependent PCOR. Examples of calibration samples include flat samples of SiC. This step, referred to as system characterization, is discussed in more detail below.
Comparing the ideal with the actual system characteristics provides a list of correction coefficients for the Fourier components. Assuming that the system configuration reasonably approximates a situation that meets the separability requirements for the ρsys and ρpart described above, these coefficients are the same regardless of part type in subsequent measurements, including test sample including thin film structures, so that the algorithm can compensate for the system characteristics as shown in
Once the signal has been repaired, analysis can be continued in the frequency domain (FDA) (such as described in U.S. Publication No. 2004/0189999 A1, entitled “PROFILING COMPLEX SURFACE STRUCTURES USING SCANNING INTERFEROMETRY,” by De Groot et al., for example) or can be continued by inverse transforming back to the original signal or scan domain (such as described in U.S. Publication No. 2005/0078318 A1, entitled “METHODS AND SYSTEMS FOR INTERFEROMETRIC ANALYSIS OF SURFACES AND RELATED APPLICATIONS,” by De Groot, for example). In either case, the subsequent analysis proceeds as it would have without the signal repair step. Another analytical technique that may be used is disclosed in U.S. patent application Ser. No. 11/437,002 by Peter de Groot entitled “METHOD AND SYSTEM FOR ANALYZING LOW-COHERENCE INTERFEROMETRY SIGNALS FOR INFORMATION ABOUT THIN FILM STRUCTURES,” filed May 18, 2006, the contents of which are incorporated herein by reference. In this sense, the signal repair is a black box preprocessor that simply improves signal quality prior to the usual low coherence interferometry analysis.
Turning to the details of system characterization, this procedure involves a measurement of a known sample, usually a clean flat sample of a near dielectric material such as SiC. A convenient simplifying assumption is that the part itself contributes negligibly to the results of the system characterization analysis, so that system errors are measured directly. Expressed another way,
ρpart≈1 (during system characterization). (11)
Alternatively, the calculation can account more accurately for the reflectivity of the part, which can be calculated using the Fresnel formulae. For example, in some embodiments, the measured Fourier coefficients can be divided by theoretically known part characteristics.
Assuming that the scan domain signal Isys acquired during system characterization has uniform sampling, a forward Fourier Transform (FT) provides frequency-domain values qsys for each pixel:
The next step is to extract and average the magnitudes and phases within the region of interest defined by the frequency indices (or bins) vmin<v<vmax:
where Y is the total number of pixels and
Pj,vsys=|qj,vsys| (15)
and the measured phase as a function of angular frequency is
φ″j,vsys=connect(φ′″j,vsys) (16)
for
φ′″j,vsys=arg(qj,vsys). (17)
Here, the function connect connects across angular frequencies for each pixel. Where one is interested only in the nonlinear portion of
φvsys=nonlin(
The spectral contributions outside the region of interest are set to zero. Usually it is sufficient to operate only on positive nonzero frequencies as the negative frequency values for q will all be complex conjugates of the positive ones.
The three primes for the phase data φ′″sys in the frequency domain indicate that there are multiple 2-π uncertainties in the phase information: from angular frequency 0 to angular frequency K, from pixel to pixel, and overall with respect to an absolute reference. As mentioned previously, the connect function in Eq. (16) removes one of these primes by connecting across angular frequencies for each pixel. The field averaging in Eq. (14) removes another prime, leaving only the single prime that indicates that the overall offset value for the phase is unknown. For the present purpose, this last offset value is irrelevant and disappears in Eq. (18).
Finally, for the height-independent portion of
ρvsys=
In some embodiments, ρvsys can be divided by the known theoretical part characteristics to get a more accurate value for ρvsys.
Relying on the concept of separability (discussed above) and assuming that the system characterization procedure introduces no errors in magnitude or phase, the repaired frequency-domain values q of a data acquisition for an arbitrary sample part can be defined as
qvrepair=Cvfixqv (20)
where the coefficients Cfix are such that
ρvtarget≈Cvfixρvsys (21)
and ρtarget represents the ideal system characteristics. The goal here is to establish the complex coefficients Cvfix so that Eq. (21) is satisfied as nearly as possible.
To correct for system dispersion, we subtract the known system error φsys directly from the phase values of all subsequent measurements:
arg(Cvfix)=−φvsys. (22)
To correct the magnitude of Cvfix, a smoothly-apodized, symmetric target function is defined in the frequency domain, such as, for example, a raised cosine:
where v is the index for the frequency, v0 is the peak of the target function and σ is the full width to zero=2× full width to half maximum, in units of frequency index. The value of σ follows deterministically from an analysis of the measured magnitude. For example, to match the full width at the 25% points of the target function Ptarget to the full width Δv of the measured magnitude Psys at the same points, set
σ=1.5 Δv. (24)
An example of an alternative target function is a Gaussian target function:
to match to the 25% of peak value points.
The next step is to determine correction coefficients that repair the magnitude of all subsequent measurements. Mathematically, this step could correspond to the decorrelation operation
but this may be problematic where values of Psys approach zero while Ptarget is still nonzero. Alternatively, one can define a coefficient for magnitude correction that places a limit on how strong the magnitude correction can be. For example, to have a maximum correction of 4× with a graceful transition to no correction, one can construct coefficients
Alternative definitions for the magnitude of Cfix can be established according to one's experience with real data.
The correction coefficients are now
Cvfix=|Cvfix|exp(−iφvsys) (29)
where Eq. (28) or some alternative approximate deconvolution defines the magnitude |Cvfix|.
Note that the acquired doing system characterization and the experimental data should have exactly the same number of frames, so that the frequency-domain bins correspond precisely. In some embodiments, where this requirement is not met by the raw data, interpolation or extrapolation of the data can be performed prior to further analysis in order to ensure the that the system characterization and experimental data have the same number of frames.
As an example,
Referring to
Removing the phase error, the asymmetry and the unwanted noise showed considerable improvement in the real and repaired sample data shown in
Another candidate for an extended bandwidth light source is a light emitting diode (LED) with a blue-block filter removed.
Further, while implementations that utilize Fourier transforms are discussed, more generally other techniques for frequency and phase analysis can be used. For example, in certain embodiments, Hilbert transforms, Wavelet transforms, and/or Lomb-Scargle analysis can be used.
In much of the analysis herein, it is assumed that the polarization state of the light in the pupil plane is random, i.e., comprised of approximately equal amounts of both s polarizations (orthogonal to the plane of incidence) and p (orthogonal to the plane of incidence) polarizations. More generally, however, other polarizations are possible, including pure s polarization, such as may be realized by means of a radial polarizer placed in the pupil plane (e.g., in the back-focal plane of the test sample in the case of a Linnik interferometer and in the back focal plane of the common objective in the Mirau interferometer). Other possible polarizations include radial p polarization, circular polarization, and modulated (e.g., two states, one following the other) polarization for ellipsometric measurements. In other words, optical properties of the test sample can be resolved not only with respect to their angle- or wavelength-dependence, but also with respect to their polarization dependence or with respect to a selected polarization. Such information may also be used to improve the accuracy of thin film structure characterization.
To provide such ellipsometry measurements, the scanning interferometry system may include a fixed or variable polarizer in the pupil plane. Referring again to
Moreover, the polarization optics may provide a variable polarization across the pupil plane (e.g., by including multiple polarizers or a spatial modulator). Thus, one can “tag” the polarization state according to spatial frequency, for example, by providing a different polarization for high angles of incidence than shallow angles.
In certain embodiments, the selectable polarization may be combined with a phase shift as a function of polarization. For example, the polarization optics may include a linear polarizer is positioned in the pupil plane and followed by two waveplates (e.g., eighth-wave plates) in opposing quadrants of the pupil plane. The linear polarization results in a full range of polarization angles with respect to the incident planes of the objective. If the waveplates are aligned so that, for example, the predominately s-polarized light has a fixed phase shift, then both radial s polarized and p polarized light are present simultaneously, but shifted in phase with respect to each other, e.g., by π, so that the interferometer is effectively detecting the difference between these two polarization states as the fundamental signal.
In further embodiments, polarization optics may be positioned elsewhere in the apparatus. For example, linear polarization can be achieved anywhere in the system.
Any of the analysis methods described above can be implemented in computer hardware or software, or a combination of both. The methods can be implemented in computer programs using standard programming techniques following the method and figures described herein. Program code is applied to input data to perform the functions described herein and generate output information. Each program may be implemented in a high level procedural or object oriented programming language to communicate with a computer system. However, the programs can be implemented in assembly or machine language, if desired. In any case, the language can be a compiled or interpreted language. Moreover, the program can run on dedicated integrated circuits preprogrammed for that purpose.
Each such computer program is preferably stored on a storage medium or device (e.g., ROM or magnetic diskette) readable by a general or special purpose programmable computer, for configuring and operating the computer when the storage media or device is read by the computer to perform the procedures described herein. The computer program can also reside in cache or main memory during program execution. The analysis method can also be implemented as a computer-readable storage medium, configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner to perform the functions described herein.
In general, program outputs information generated by the analysis. For example, the information can be output to another computer (e.g., over a network), written to memory, or output to a display or printer where a user can view the information. In some embodiments, the information is output in a feedback or feedforward process (e.g., in a sample analysis application such as those described below).
The low coherence interferometry methods and systems described above may used for any of the following surface analysis problems: simple thin films; multilayer thin films; sharp edges and surface features that diffract or otherwise generate complex interference effects; unresolved surface roughness; unresolved surface features, for example, a sub-wavelength width groove on an otherwise smooth surface; dissimilar materials; polarization-dependent properties of the surface; and deflections, vibrations or motions of the surface or deformable surface features that result in incident-angle dependent perturbations of the interference phenomenon. For the case of thin films, the variable parameter of interest may be the film thickness, the refractive index of the film, the refractive index of the substrate, or some combination thereof. Exemplary applications including objects and devices exhibit such features are discussed next.
Photolithography
In many microelectronics applications, photolithography is used to pattern a layer of photoresist overlying a substrate, e.g., a silicon wafer. Referring to
A photolithography apparatus images a pattern onto the object. For example, the pattern may correspond with elements of an electronic circuit (or the negative of the circuit). After imaging, portions of the photoresist are removed revealing the substrate underlying the removed photoresist. The revealed substrate can be etched, covered with deposited material, or otherwise modified. Remaining photoresist protects other portions of the substrate from such modification.
To increase manufacturing efficiencies, more than one device is sometimes prepared from a single wafer. The devices may be the same or different. Each device requires that a subset of the wafer be imaged with a pattern. In some cases, the pattern is sequentially imaged onto different subsets. Sequential imaging can be performed for several reasons. Optical aberrations can prevent achieving adequate pattern focus quality over larger areas of the wafer. Even in the absence of optical aberrations, the spatial properties of the wafer and photoresist may also prevent achieving adequate pattern focus over large areas of the wafer. Aspects of the relationship between the spatial properties of the wafer/resist and focus quality are discussed next.
Referring to back to
Proper focus can be achieved by determining a spatial property of an object within a subset of the object to be imaged (or otherwise processed). Once the position of the subset has been determined, the object (and/or a portion of the photolithography imager) can be moved, e.g., translated, rotated, and/or tilted, to modify the position of the subset with respect to a reference, e.g., a portion of the photolithography imager. The determination and movement (if necessary) can be repeated for each subset to be imaged.
The determination of the spatial property of the subset can include determining a position and/or height of one or more points of an outer surface of a thin layer of the object, the one or more points lying within the subset of the object to be imaged. For example, the position and orientation of the outer surface 39 of subset 402 (
Detecting an interference signal from each subset of an object to be imaged can include detecting light reflected from the subset and reference light over an OPD range that is at least as large as a coherence length of the detected light. For example, the light may be detected at least over its coherence length. In some embodiments, the interferometer is configured so that the light reflected from the illuminated subset is dominated by light reflected from either an outer interface (such as outer surface 39) or an inner interface (such as interface 36). In some embodiments, a spatial property of an object is determined based on only a portion of the interference signal. For example, if the interference signal includes two or more overlapping interference patterns, a spatial property of the object can be determined based upon a portion of one of the interference patterns that is dominated by contributions from a single interface of the object.
Copper Interconnect Structures And Chemical Mechanical Polishing
It is becoming common among chip makers to use the so-called “dual damascene copper” process to fabricate electrical interconnects between different parts of a chip. This is an example of a process which may be effectively characterized using a suitable surface topography system. The dual damascene process may be considered to have six parts: (1) an interlayer dielectric (ILD) deposition, in which a layer of dielectric material (such as a polymer, or glass) is deposited onto the surface of a wafer (containing a plurality of individual chips); (2) chemical mechanical polishing (CMP), in which the dielectric layer is polished so as to create a smooth surface, suitable for precision optical lithography, (3) a combination of lithographic patterning and reactive ion etching steps, in which a complex network is created comprising narrow trenches running parallel to the wafer surface and small vias running from the bottom of the trenches to a lower (previously defined) electrically conducting layer, (4) a combination of metal deposition steps which result in the deposition of copper trenches and vias, (5) a dielectric deposition step in which a dielectric is applied over the copper trenches and vias, and (6) a final CMP step in which the excess copper is removed, leaving a network of copper filled trenches (and possibly vias) surrounded by dielectric material.
Referring to
Referring to
This position dependent polishing rate is known to give rise to variable surface topography on many lateral length scales. For example, it may mean that chips located closer to the edge of a wafer on aggregate are polished more rapidly than those located close to the center, creating copper regions which are thinner than desired near the edges, and thicker than desired at the center. This is an example of a ‘wafer scale’ process nonuniformity—i.e., one occurring on length scale comparable to the wafer diameter. It is also known that regions which have a high density of copper trenches polish at a higher rate than nearby regions with low copper line densities. This leads to a phenomenon known as ‘CMP induced erosion’ in the high copper density regions. This is an example of a ‘chip scale’ process non-uniformity—i.e., one occurring on a length scale comparable to (and sometimes much less than) the linear dimensions of a single chip. Another type of chip scale nonuniformity, known as ‘dishing’, occurs within single copper filled trench regions (which tend to polish at a higher rate than the surrounding dielectric material). For trenches greater than a few microns in width dishing may become severe with the result that affected lines later exhibit excessive electrical resistance, leading to a chip failure.
CMP induced wafer and chip scale process nonuniformities are inherently difficult to predict, and they are subject to change over time as conditions within the CMP processing system evolve. To effectively monitor, and suitably adjust the process conditions for the purpose of ensuring that any nonuniformities remain within acceptable limits, it is important for process engineers to make frequent non-contact surface topography measurements on chips at a large number and wide variety of locations. This is possible using embodiments of the interferometry methods and systems described above.
In some embodiments one or more spatial properties, e.g., the topography of surface 506 and/or the thickness of dielectric 504, are monitored by obtaining low coherence interference signals from the structure before and/or during CMP. Based on the spatial properties, the polishing conditions can be changed to achieve the desired planar surface 506′. For example, the pad pressure, pad pressure distribution, polishing agent characteristics, solvent composition and flow, and other conditions can be determined based on the spatial properties. After some period of polishing, the spatial property can again be determined and the polishing conditions changed as needed. The topography and/or thickness is also indicative of the end-point at which, e.g., surface 504′ is achieved. Thus, the low coherence interference signals can be used to avoid depressions caused by over polishing different regions of the object. The low coherence interference methods and systems are advantageous in this respect because spatial properties of the device, e.g., the relative heights of the surface of the dielectric (a) over copper elements 502 and (b) over substrate surface 510 but adjacent copper elements 502 can be determined even in the presence of the multiple interfaces.
Solder Bump Processing
Referring to
During processing a mass of solder 504 is positioned in contact with wettable region 503. Upon flowing the solder, the solder forms a secure contact with the wettable region 503. Adjacent non-wettable regions 502 act like a dam preventing the flowed solder from undesirable migration about the structure. It is desirable to know spatial properties of the structure including the relative heights of surfaces 507, 509 and the dimensions of solder 504 relative to surface 502. As can be determined from other discussions herein, structure 550 includes a plurality of interfaces that may each result in an interference pattern. Overlap between the interference patterns prevents accurate determinate of the spatial properties using known interference techniques. Application of the systems and methods discussed herein allow the spatial properties to be determined.
Spatial properties determined from structure 550 can be used to change manufacturing conditions, such as deposition times for layers 502, 503 and the amount of solder 504 used per area of region 503. Additionally, heating conditions used to flow the solder can also be changed based on the spatial properties to achieve adequate flow and or prevent migration of the solder.
Liquid Crystal Displays
Referring to
The cell gap Δg determines to a great extent the optoelectrical properties of the LCD, e.g., the contrast ratio and brightness. Cell gap control during manufacturing is critical to obtaining uniform, quality displays. The actual cell gap may differ from the dimensions of spacers 466 because, during assembly, pressure or vacuum is applied to introduce the liquid crystal medium, seals 454 cure and may change dimensions, and the added liquid crystal medium generates capillary forces between plates 452, 453. Both before and after adding the liquid crystal medium 464, surfaces 480, 482 of plates 452, 453 reflect light that results in an interference pattern indicative of the cell gap Δg. The low coherence nature of the interference signal either itself or in combination with the described interference signal processing techniques can be used to monitor properties of the cell including the cell gap Δg during manufacture even in the presence of interfaces formed by other layers of the cell.
An exemplary method can include obtaining a low coherence interference signal including interference patterns indicative of the cell gap Δg prior to adding layer 464. The cell gap (or other spatial property of the cell) is determined from the interference patterns and can be compared to a specified value. Manufacturing conditions, e.g., a pressure or vacuum applied to plates 452, 453 can be changed to modify the cell gap Δg if a difference between the specified value and the determined cell gap exceeds tolerances. This process can be repeated until achieving the desired cell gap. Liquid crystal medium is then introduced into the cell. The amount of liquid crystal medium to be added can be determined from the measured spatial property of the cell. This can avoid over- or underfilling the cell. The filling process can also be monitored by observing interference signals from the surfaces 480, 482. Once the cell has been filed, additional low coherence interference patterns are obtained to monitor the cell gap Δg (or other spatial property). Again, the manufacturing conditions can be changed so that the cell gap is maintained or brought within tolerances.
Laser Scribing and Cutting
Lasers can be used to scribe objects in preparation for separating different, concurrently manufactured structures, e.g., microelectronics structures. The quality of separation is related to the scribing conditions, e.g., laser focus size, laser power, translation rate of the object, and scribe depth. Because the density of features of the structure may be large, the scribe lines may be adjacent thin film or layers of the structures. Interfaces associated with the thin film or layers may create interference patterns that appear when interferometry is used to determine the scribe depth. The methods and systems described herein can be used to determine the scribe depth even in the presence of such adjacent films or layers.
An exemplary method can include scribing one or more electronic structures and separating the structures along the scribe lines. Before and/or after separation, low coherence interference signals can be used to determine the depth of scribe. Other scribing conditions are known, e.g., laser spot size, laser power, translation rate. The scribe depth can be determined from the interference signals. The quality of separation as a function of the scribing conditions, including the scribe depth, can be determined by evaluating the separated structures. Based on such determinations, the scribing conditions necessary to achieve a desired separation quality can be determined. During continued manufacturing, low coherence interference signals can be obtained from scribed regions to monitor the process. Scribing conditions can be changed to maintain or bring the scribe properties within tolerances.
A number of embodiments of the invention have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the invention. Accordingly, other embodiments are within the scope of the following claims.
This application claims benefit of Provisional Patent Application 60/832,452, entitled “COMPENSATION OF SYSTEMATIC EFFECTS IN SCANNING WHITE LIGHT INTERFEROMETRY,” filed on Jul. 21, 2006, the entire contents of which are incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
60832452 | Jul 2006 | US |