Patent Application
20190137265
The present invention relates to a method for the optical inspection and measurement of a face of an object, in particular for imaging patterns present on said face. It also relates to a system for the inspection of a face of an object implementing such a method.
The field of the invention is more particularly, but non-limitatively, that of optical profilometry.
Optical profilometry makes it possible to inspect a face of an object in order in particular to detect and to image patterns, such as for example steps or trenches present on said face.
It is based on the measurement, then the study, of an interferometric signal obtained between a reference optical radiation and an inspection optical radiation originating from one and the same source, sent to the inspected face and reflected by said face. By relatively varying the optical path of the reference or inspection radiations, it is possible to determine from analysis of interference fringes the difference in length of the optical path travelled by the reflected inspection optical radiation relative to that of the optical path of the reference radiation, and deducing from it the depth or the height of the inspected face at each measurement point in order to detect and image the patterns present on said face.
However, the current optical profilometry techniques are limited in lateral resolution by the resolution of the optical imaging system which makes it possible to image the surface of the object and the interference fringes. In particular, they do not make it possible to inspect or carry out measurements on narrow patterns of characteristic dimensions close to the resolution limit of the optical system, even with a lens with a high enlargement factor such as a 50× lens. In this case,
the interferometric signal obtained cannot be suitably exploited as the items of information relating to two adjacent surfaces of different depths, such as for example the two surfaces of a trench or a step, are invariably mixed together.
An objective of the present invention is to propose a method and a system for the inspection and measurement of a face of an object more accurately.
Another objective of the present invention is to propose a method and a system for the inspection and measurement of a face of an object making it possible to detect and image narrow patterns with accuracy without the use of expensive optical means.
Another objective of the present invention is to propose a method and a system for the inspection and measurement of a face of an object making it possible to detect and image with accuracy narrow patterns the characteristic dimensions of which are of the order of or less than the resolution limit of the optical imaging system.
It is also an objective of the present invention to propose a method and a system for the inspection and measurement of a face of an object making it possible to image narrow patterns with current inspection instruments with very little or even with no modification of their hardware architecture.
At least one of these objectives is achieved with a method for the inspection and measurement of a face of an object comprising at least two surfaces staggered depthwise with respect to one another, said surfaces forming in particular a step or a trench on or in said face, said method comprising the following steps:
Such an extraction step, proposed by the method according to the invention, is particularly useful when the measurement point is located at the level of an interface between two adjacent surfaces of different depths. In fact, in this case, the measured interferometric signal contains, mixed together, items of information relating to each of the adjacent surfaces.
The method according to the invention thus proposes to isolate, individually, the item of information relating to each surface constituting the inspected face, by selecting from the measured interferometric signal, the part of said measured signal corresponding to each surface, before the step of profilometric analysis. Once isolated, each individual interferometric signal can be analysed according to a known profilometry technique.
Thus, the method according to the invention makes it possible to reduce or even undo the mixtures of items of information relating to two adjacent surfaces of different depths, in particular at the interface of these two surfaces, which makes it possible to carry out a more accurate inspection of a face of an object.
In addition, by reducing the mixtures of items of information relating to two adjacent surfaces of different depths, in particular at the level of the interface of these two surfaces, the method according to the invention makes it possible, with a given optical imaging system and a sensor, to image patterns of smaller dimensions, in particular narrower, compared with the methods of the state of the art.
Moreover, the extraction step can be carried out by digital or analogue processing, and has little or no impact on the signal measurement steps. Consequently, the method according to the invention can be implemented by the current inspection or measurement devices with very little or no architectural modification, in particular of the optical part of these devices.
In other words, the method according to the invention makes it possible to push back the lateral resolution limit imposed by the optical imaging system and the sensor, by allowing the inspection and the dimensional measurement of patterns with characteristic dimensions of the order of or less than this resolution limit,
which could not be inspected or measured by this optical system otherwise.
The extraction step can be implemented for all the measurement points.
Alternatively, the extraction step can be implemented for the measurement points only, located at the level of an interface between two adjacent surfaces of different depths.
Advantageously, the step of measuring the interferometric signal can carry out a measurement of an interferometric signal for each pixel of a sensor carrying out a full-field measurement.
In this case, each pixel can correspond to a measurement point for which an interferometric signal is measured.
Advantageously, the method according to the invention can moreover comprise a step of, independently constructing each surface as a function of the profilometric analysis of the individual signals from said surface.
In fact, by exploiting the depth detected in each individual interferometric signal for each measurement point, it is possible to merge the individual signals relating to this surface and construct this surface independently.
In order to do this, each measurement point is positioned at the depth detected for said measurement point in the individual interferometric signal relating to said measurement point.
Moreover, the method according to the invention can furthermore comprise a step of constructing a representation of the inspected face, and in particular a three-dimensional representation of said face, comprising the patterns found on said inspected face.
Such a construction, in particular three-dimensional, can be produced by concatenation of the measured depth values in each individual signal at the level of each measurement point, and for the surfaces in their entirety.
In other words, such a construction can be produced by concatenation of the measurement points, at the depths detected in the individual interferometric signals for the surfaces in their entirety
In a particularly advantageous version, the step of constructing a representation of the inspected face can be produced from the individual representations of the surfaces.
In this case, the construction step can comprise, for at least one measurement point, an iteration of the following steps:
In particular, these steps can be carried out for at least one, in particular each, measurement point allocated to two adjacent surfaces of different depths. In this case, the individual representations considered are those of said adjacent surfaces.
Thus, the method according to the invention makes it possible to determine a three-dimensional representation of the inspected face with greater accuracy.
In fact, it can happen that a measurement point which is located at the level of the interface between two adjacent surfaces of different depths is allocated to each of these two surfaces. In this case, a standard three-dimensional representation would show said measurement point in each of these adjacent surfaces. This would be incorrect. The method according to the invention makes it possible to avoid such a double allocation, by distinguishing on the basis of the quality of the signal obtained for each of the adjacent surfaces, and allocating the measurement point to one of the adjacent surfaces only.
In a non-limitative embodiment, the allocation step can be carried out by a predetermined comparison relationship taking into account:
Thus, the method according to the invention makes it possible to allocate a measurement point to a surface when, for said measurement point, the quality of the signal in the individual representation of one of the surfaces is greater, optionally by a multiplier coefficient, than the signal quality value in the individual representation of the other one of the surfaces.
According to a non-limitative embodiment example, the multiplier coefficient can be determined empirically or experimentally.
According to another non-limitative embodiment, the multiplier coefficient can be determined by learning, for example from reference measurements on objects of known characteristics.
Alternatively or in addition, the multiplier coefficient can be a function of at least one parameter of a measurement sensor used during the measurement step. Such a parameter of the measurement sensor can for example be a sensitivity of the sensor or a measurement uncertainty value of said sensor, for example given by the manufacturer or measured during prior tests.
Alternatively or in addition, the multiplier coefficient can be determined as a function of at least one parameter of the inspected face. Such a parameter of the inspected face can for example be a reflection/refraction value of the material used, a value of difference of theoretical depth between two adjacent surfaces, a characteristic dimension of the pattern, etc.
In a particularly preferred embodiment, for at least one surface, the profilometric analysis step can comprise for each individual signal:
In fact, the phase of the Fourier transform of the individual interferometric signal of a simple surface is linear, and analysis of this phase makes it possible to accurately deduce an item of topographical information. On the other hand, it should be noted that this simple method does not work with an interferometric signal which comprises a mixture of items of information relating to two or more adjacent surfaces, as in this case the phase of the
Fourier transform of the interferometric signal measurement shows no such linearity.
In a version of the method according to the invention, for at least one surface, the step of extracting a measured signal relating to said surface can comprise selecting a portion of said measured interferometric signal comprising an envelope corresponding to said surface in said measured interferometric signal. This envelope can correspond to a significant local amplitude of the fringes or of the interference signals.
Such an extraction step is not very complicated to implement, requires few resources and a very short processing time.
In particular, for two adjacent surfaces of different depths, the selection step can advantageously produce a splitting of the measured interferometric signal into two portions each comprising an envelope corresponding to one of said surfaces in said measured signal, the individual signal for each surface corresponding to one of said portions.
Of course, when more than two adjacent surfaces of different depths exist for one measurement point, the splitting can be carried out by considering the adjacent envelopes in the measured signal in pairs.
Such splitting is not very complicated to implement, requires few resources and a very short processing time for the measured interferometric signal.
According to a particularly preferred embodiment, the splitting of the measured signal, for two adjacent envelopes, can be carried out at a position of said measured signal:
The positions of the envelopes can for example correspond to their respective peaks.
For example, if the envelopes corresponding to two adjacent surfaces are separated, in the measured interferometric signal, by a distance of depth Δ, then the measured interferometric signal is
split in two at a position located between the two envelopes, at a distance Δ/2 from the position of each envelope.
In a version, the depth of each surface, and therefore the position of each envelope in the measured interferometric signal, can be provided beforehand, in particular approximately or theoretically, prior to the inspection, for example by a designer or a manufacturer of the object the face of which is inspected.
The depth of at least one surface of the inspected face can be provided relative to another surface of said surface.
Alternatively or in addition, the method according to the invention can comprise a step of estimating the position, in the measured interferometric signal, of at least one envelope corresponding to a surface, prior to the extraction step.
Such a step of estimating the position of an envelope can be carried out in different ways, by analysis of the measured interferometric signal.
In particular, the step of estimating the position of an envelope in the measured interferometric signal can comprise a step of:
For example, in the context of an energy analysis, the position of an envelope in the measured signal can be detected by detecting the position of a local maximum of the energy of the measured interferometric signal.
In the context of an analysis by demodulation, the position of an envelope in the measurement signal can be detected by applying a low-pass filter to the rectified signal. This low-pass filter makes it possible to eliminate the high frequency component of the rectified signal, i.e. the fringes, while retaining the low frequency component, i.e. the envelope of the signal. The rectified signal can be obtained for example with an absolute value operator, average value thresholding, squaring, or
multiplication by a carrier of the same frequency (synchronous demodulation).
In the context of an analysis of fringe contrast, the position of an envelope in the measured signal can be detected by looking for the amplitude and/or the peaks of the interference fringes, for example with comparison operators or by derivation.
According to another aspect of the same invention, a system is proposed for the inspection and measurement of a face of an object comprising at least two surfaces staggered depthwise with respect to one another, said surfaces forming in particular a step or a trench on or in said face, said system comprising:
The configuration of the processing module can be carried out in an electronically and/or by computer, in particular with instructions executable by a processor or an electronic chip, of EEPROM type for example.
The processing module can be incorporated in the measurement device, or be external to the measurement device and connected to said measurement device in a wired or wireless manner.
In an advantageous version, the measurement device can comprise a full-field interferometric sensor.
In this case, a measurement point can correspond to a pixel of said sensor.
The method and the system according to the invention can each be used for the inspection of a face of a semiconductor or wafer element, in particular for measuring the depth(s) of trench(es) and/or the height(s) of step(s) present on said face, or also for imaging said face.
More generally, the method and the system according to the invention can each be used for the inspection of a face of an object, in particular for the detection and/or the characterization and/or the imaging of at least one pattern of said face.
Other advantages and characteristics of the invention will become apparent on examination of the detailed description of examples which are in no way limitative, and the attached drawings, in which:
It is well understood that the embodiments which will be described hereinafter are in no way limitative. Variants of the invention can be considered comprising only a selection of the characteristics described hereinafter, in isolation from the other characteristics described, if this selection of characteristics is sufficient to confer a technical advantage or to differentiate the invention with respect to the state of the prior art. This selection comprises at least one, preferably functional, characteristic without structural details, or with only a part of the structural details if this part alone is sufficient to confer a technical advantage or to differentiate the invention with respect to the state of the prior art.
In particular, all the variants and all the embodiments described can be combined together if there is no objection to this combination from a technical point of view.
In the figures, the elements that are common to several figures retain the same reference.
The method 100, represented in
The method 100 comprises moreover a processing phase 104, carried out for each measurement point, i.e. for each measured interferometric signal.
The processing phase 104 comprises a step 106 carrying out an estimation of the position of each envelope in the measured interferometric signal. This estimation step 106 is carried out by demodulation of the measured interferometric signal by applying a low-pass filter on the rectified signal after background subtraction (or by carrying out a synchronous demodulation). The background is calculated by smoothing the signal with a sufficiently broad averaging sliding window. The low-pass filter makes it possible to remove the high-frequency components of the rectified signal while retaining the low-frequency component, i.e. the envelope. The local maxima of the demodulated signal which exceed a predetermined amplitude threshold are detected and their position stored. The amplitude threshold can be chosen, for example, in order to find a good compromise between a number of false detections and a number of non detection of envelopes, the challenge being to detect the weak signals without the noise in the interferometric signals leading to too many false detections. This detection threshold can be set or adaptive as a function, for example:
The interferometric signal is then processed, during step 108, by considering that each envelope detected during step 106 corresponds to a surface of different depth. In particular, the processing consists of splitting the interferometric signal into as many portions as there are envelopes in said measured interferometric signal. Splitting the interferometric signal is carried out between the adjacent envelopes, taken in pairs, in a position located substantially at an equal distance from the position of each of said two adjacent envelopes. For example, when the interferometric signal comprises N envelopes Ek, with 1≤k≤N and Dk the position of the envelope k in said measured interferometric signal, a first portion P1 comprising the envelope E1 is firstly split in a splitting position DD1 located between the positions D1 and D2, and at an equal distance from the positions D1 and D2. Then, a second portion P2 comprising the envelope E2 is split: this second portion corresponds to the portion of the measured interferometric signal located between the first splitting position DD1 and a second splitting position DD2 located between the positions D2 and D3, at an equal distance from the positions D2 and D3, and so on. The last portion PN corresponds to the portion of the measured interferometric signal located between the penultimate splitting position DDN-1 and the end of the measured interferometric signal.
When the measured interferometric signal comprises only two envelopes E1 and E2, then it is split into two portions in a splitting position DD1 located between the positions D1 and D2, and at an equal distance from the positions D1 and D2. The first portion P1 comprises the start of the measured signal up to the splitting position DD1 and the second portion P2 comprises the end of the measured signal starting from the splitting position DD1.
Each portion obtained during the splitting step forms an individual signal for each surface of the inspected face.
During step 110, a profilometric analysis of each individual signal is carried out in order to detect the position of the surface to which the single envelope contained in the individual signal corresponds. During this step 110 each signal individual undergoes:
The frequency domain where the phase of the Fourier transform is linear corresponds to the frequency domain of the light source of the profilometer.
In addition, the depth of the surface at the corresponding measurement point can be deduced from the slope of the phase in this frequency domain or from the value of the phase at the central frequency of the light source of the profilometer.
The processing phase 104 ends at step 110.
During step 112, as a function of the profilometric analysis of the individual signals, an individual construction of each surface of a given depth is carried out by concatenation of the measurement points detected at said depth.
During the construction of the surfaces individually, it is possible and frequent, in particular in the case where a lens with a high enlargement factor, such as a 50×, is used, that at one and the same measurement point, two different depths are detected, and that consequently this measurement point is allocated to two surfaces of different depths. Such a situation occurs in particular when the measurement point is at the limit between two adjacent surfaces of different depths.
During phase 114, a three-dimensional representation of the inspected face is carried out.
During this phase 114, step 116 carries out a concatenation/merging of the individual representations obtained during step 112, for all the measurement points.
When a contentious measurement point, denoted (i,j), is detected as belonging to two different surfaces, a step 118 determines the quality Q1(i,j) and Q2(i,j) of the individual measurement signal corresponding to surface 1, respectively to surface 2. This quality measurement is obtained from the maxima observed on the demodulated signal during the step of detecting the interfaces/envelopes (step 106). It corresponds for example to the maximum amplitude of the envelope of the surface considered.
Step 120 carries out an allocation of said contentious measurement point to one of the two surfaces by comparing the qualities Q1(i,j) and Q2(i,j). For example:
A weighting coefficient, or multiplier coefficient β is applied to the quality measurements in order to carry out the comparison. In the embodiment implemented, this multiplier coefficient β is determined experimentally so as to substantially compensate for the difference in light energy reflected by the different surfaces of the patterns. In fact, the base of the patterns (surface 2 in the examples presented) in general naturally reflects light less than the upper surfaces (surface 1). Thus a multiplier coefficient β>1 is chosen, such as for example β=5.
During a step 122, a graphical representation of the inspected face is produced.
The method 100 can moreover comprise analysis and statistical steps relating to the widths, heights, depths of patterns, such as steps or trenches.
In particular, the face 200, shown on
two envelopes: envelope 2101 corresponds to a step 202 and envelope 2102 corresponds to a trench 204.
for example in step 122 of the method 100 of
It can be noted in particular that an in-plane representation of the pattern is obtained with a better localization of the transitions than in the original image in
The system 300, shown in
In the full-field interferometer 306, the light beam 304 is separated into a reference beam which illuminates a reference mirror and a measurement beam which illuminates a surface to be inspected, for example the surface 200 in
The system 300 comprises optics and lenses, including an imaging lens, arranged so as to image the surface 200 on the detector
array 310. When the difference in optical paths between the measurement beam and the reference beam is less than the coherence length of the light source 302, interference fringes due to the interferences between the measurement beam and the reference beam are also visible.
Different types of full-field interferometers 306 exist that can be used within the context of the invention, which are well known to a person skilled in the art and will not be detailed here.
The system 300 comprises moreover an electronic/computer module 312, such as a processor or an electronic chip or also a personal computer for example, connected to the detector array 310, and configured in order to implement all the steps of the method according to the invention, such as for example steps 104-122 of the method 100 of
Of course, the invention is not limited to the examples which have just been described and numerous adjustments can be made to these examples without exceeding the scope of the invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 1653158 | Apr 2016 | FR | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2017/058145 | 4/5/2017 | WO | 00 |
