Method and system for distinguishing defects present on the front side of a transparent substrate from defects present on the back side - Patents.com

Abstract

The present invention relates to a device and method for distinguishing defects present on the front side of a substrate from defects present on the back side of the substrate, the substrate being made of a material transparent at an inspection wavelength, comprising: disposing the substrate in an inspection system comprising at least one light source coupled to an optical system, the substrate being disposed on a support and positioned and disposed relative to the optical system such that a first light beam and a second light beam intersect at a measurement spot on the front side of the substrate, and controlling a relative movement of the support and the optical system to scan the measurement spot along a measurement path on the front side of the substrate, the relative movement being controlled such that a reference plane is kept tangential to the measurement path. The method further comprises identifying in the signal a first pattern that corresponds to light scattered by a particle on the back side of the substrate and presents two intensity peaks separated from each other by a determined separation interval corresponding to the time required for the defect (P) to be moved over a distance (d) that separates the two illumination spots (S1, S2).

Description

[Technical field]

[0001] The present invention relates to a system and a method for the optical inspection of a substrate surface to detect the possible presence of defects such as particles or more generally to characterize the surface state of the substrate. The substrate may for example be a wafer for the manufacture of devices in the fields of microelectronics, optics, microsystems or optoelectronics. [Background technology]

[0002] Substrates such as wafers for electronic, optical, or optoelectronic applications need to be inspected to detect, identify, and / or characterize defects that may be present on their surface. The defects may be particles, crystal defects, scratches, or surface roughness.

[0003] This inspection is generally intended to provide qualitative or quantitative information, such as, for example, the location, size, and / or nature of the defects. This information regarding the surface condition of the substrate may be indicative of the quality of the substrate manufacturing process or production step in which the substrate is used.

[0004] The documents WO 2017 / 167573, US 2021 / 215617(A1), US 2018 / 0231370 and WO 0239099 disclose dark field systems for inspecting the front side of such substrates, i.e. the surface of the substrate prepared to be processed in a subsequent manufacturing step. Those inspection systems are based on Laser Doppler Velocimetry (LDV) and comprise a light source emitting a first light beam and a second light beam oriented relative to each other to form at their intersection a measurement volume containing a plurality of parallel interference fringes. The inspection system comprises a support that receives the substrate and positions it relative to the light source such that the measurement volume intersects the front side of the substrate and defines a measurement spot on this front side surface. The support is rotated such that the measurement spot is scanned along a concentric measurement path on the front side of the wafer. The scattered light is collected by a collecting mirror and directed to a detector which receives the collected light and provides an electrical signal representative of the change in intensity of the collected light as the measurement spot is scanned along the measurement path.

[0005] The presence of a defect on the front side of the substrate results in the scattering of a Doppler pulse measured by the detector when this defect crosses the measurement spot and has a size smaller than the period of the interference fringes. The Doppler pulse is a signal with dual frequency components: a low frequency component forming the envelope of the signal, corresponding to the average light intensity scattered by the defect, and a high frequency component corresponding to the Doppler frequency, which contains information about the velocity of the defect through the measurement spot.

[0006] The presence of a relatively large defect, having a size comparable to or larger than the period of the interference fringes, results in scattering of the pulse with no high frequency components or with high frequency components of extremely reduced intensity as such a defect crosses the measurement spot.

[0007] For wafers made of transparent materials (at the inspection wavelengths of the two interfering beams), the two inspection beams also illuminate the backside of the substrate at two respective illumination spots. When a defect present on the backside crosses one of these illumination spots, the light backscattered towards the measurement spot is also collected by the mirror and the detector produces a pulse similar to that produced by a larger defect on the frontside of the wafer, but of smaller intensity.

[0008] Prior art inspection systems are used to inspect wafers that are transparent at the inspection wavelength, where light scattered by backside defects contaminates light scattered by frontside defects, and therefore the frontside condition of the substrate cannot be unambiguously determined.

[0009] The presence of high frequency components in a pulse may help associate the pulse with defects on the front side of the wafer, while a pulse with the high frequency components removed may be associated with either front side or back side defects.

[0010] (Objective of the Invention) The object of the present invention is to at least partially solve the above-mentioned problems and to propose a system and a method adapted for the inspection of a substrate made of a material transparent at the inspection wavelength. More specifically, the object of the present invention is to propose an inspection system and a method for distinguishing defects present on the front side of a substrate from defects present on the back side of the substrate, the substrate being made of a material transparent at the inspection wavelength. Summary of the Invention

[0011] To this effect, the invention relates to a method for distinguishing defects present on a front side of a substrate from defects present on a back side of the substrate, the substrate being made of a material transparent at an inspection wavelength, the method comprising: - disposing the substrate in an inspection system comprising at least one light source coupled to an optical system for emitting a first light beam and a second light beam, the first light beam and the second light beam having an inspection wavelength and contained in a reference plane and intersecting thereat, the substrate being disposed on a support and positioned and disposed with respect to the optical system such that the first light beam and the second light beam intersect at a measurement spot on the front side of the substrate and also illuminate the back side of the substrate with two respective illumination spots separated by a distance; - controlling a relative movement between the support and the optical system to scan a measurement spot along a measurement path on the front side of the substrate, the relative movement being controlled such that a reference plane is kept tangential to the measurement path; and - collecting at least a portion of the light scattered by defects present on the front side and / or back side of the substrate, and establishing a signal representative of the variation in intensity of the collected light as the measurement spot is scanned along a measurement path.

[0012] The method further includes identifying a first pattern in the signal that corresponds to light scattered by a defect on the backside of the substrate and that presents two intensity peaks separated from each other by a determined separation interval that corresponds to the time required for the defect to be moved over a distance separating the two illumination spots.

[0013] According to a further non-limiting feature of the present invention, taken alone or in any technically feasible combination, the method further comprises filtering out the first pattern from the signal to prepare a front side signal representative of defects present only on the front side of the substrate; the method further comprises providing a backside signal representative of defects present only on the backside of the substrate; - identifying the first pattern includes determining a separation distance using at least one of the following information: an angle of intersection of the first and second light beams at the substrate, a thickness of the substrate, and a speed of movement of the substrate relative to the first and second light beams; - identifying the first pattern includes generating a reference signal or a mask signal using information about the separation interval; -The method is Sampling the signal at a given sampling rate to prepare raw measurement samples; establishing the separation distance as a determined number of samples based on an intersection angle of the first and second light beams, a thickness of the substrate, a relative movement speed, and a sampling rate; identifying a first pattern in the signal includes identifying, in at least some of the raw measurement samples, two intensity peaks separated by a determined number of samples; - The first pattern is identified by a first step of applying a transformation function to at least some of the raw measurement samples to provide transformed samples, the transformation function being configured to map a first pattern in the raw measurement samples to a second pattern of greater intensity in the transformed samples; A second step of detecting a second pattern in the transformed samples. - the second pattern includes peaks providing the location of the first pattern in the raw measurement sample, The detection step further provides an average intensity of the two peaks of the first pattern and a background intensity. the substrate is disc-shaped and presents an axis of symmetry, the relative movement between the support and the optical system being a rotation about the axis of symmetry of the substrate.

[0014] According to another aspect, the present invention relates to an inspection system for distinguishing defects present on a front side of a substrate from defects present on a back side of the substrate, the substrate being made of a material that is transparent at an inspection wavelength. The system comprises: at least one light source coupled to an optical system for emitting a first light beam and a second light beam, the first light beam and the second light beam having an inspection wavelength and included in and intersecting at a reference plane; - a support for receiving the substrate and for positioning the substrate relative to the optical system such that the first light beam and the second light beam intersect at a measurement spot on a front side of the substrate, the support and the light source being movable relative to each other; a controller for controlling a relative movement between the support and the optical system to scan a measurement spot along a measurement path on the front side of the substrate, the controller being configured to control the relative movement such that the reference plane is kept tangential to the measurement path; and a collector for collecting at least a portion of the light scattered by defects present on the front side and / or the back side of the substrate; a detector associated with the collector to receive the collected light and to provide a signal representative of the variation in intensity of the collected light as the measurement spot is scanned along the measurement path.

[0015] The inspection system also includes a processing device connected to the detector, the processing device being configured to process the signal to identify a first that corresponds to light scattered by a particle on the backside of the substrate and presents two intensity peaks separated from each other by a determined separation interval that corresponds to the time required for the defect to be moved over the distance separating the two illumination spots.

[0016] According to further non-limiting features of this aspect of the invention, alone or in any technically feasible combination, The processing device also - filtering out the first pattern from the signal to provide a front side signal representative of defects present only on the front side of the substrate; and / or o It is configured to provide a backside signal that is representative of defects present only on the backside of the substrate. the processing device comprises a converter for sampling the signal at a given sampling rate and preparing raw measurement samples; The processing device comprises an FPGA and / or a processor for identifying the first pattern at a given sampling rate. [Brief description of the drawings]

[0017] Numerous other features and advantages of the invention will become apparent from the following detailed description when considered in conjunction with the accompanying drawings.

[0018] [Figure 1] FIG. 1 illustrates an inspection system according to one embodiment. [Diagram 2] 1 is a diagram illustrating the phenomena occurring on the front and back sides of a transparent substrate positioned on a support of an inspection system. [Figure 3a] 1A-1C are diagrams illustrating advantages of the inspection system when used to inspect a transparent substrate. [Figure 3b] 1A-1C are diagrams illustrating advantages of the inspection system when used to inspect a transparent substrate. [Figure 3c] 1A-1C are diagrams illustrating advantages of the inspection system when used to inspect a transparent substrate. [Figure 4a] 1 illustrates the application of a transformation function to a signal provided by a detector of an inspection system. [Figure 4b] 1 illustrates the application of a transformation function to a signal provided by a detector of an inspection system. [Diagram 5] FIG. 1 illustrates a process flow of one embodiment of the present invention. [Figure 6a] FIG. 13 illustrates the arrangement in which signals are filtered out from the backside particle signature. [Figure 6b] FIG. 13 illustrates the arrangement in which signals are filtered out from the backside particle signature. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] (Inspection System) FIG. 1 represents an inspection system 1 according to an embodiment. The inspection system 1 comprises at least one light source 2, for example a laser source injected into an optical fiber. The light source 2 generates a main light beam 3, which is preferably monochromatic and exhibits a wavelength in the UV, visible or infrared range, referred to in the remainder of the disclosure as the "inspection wavelength". The light source 2 in this embodiment is coupled to an optical system 2' comprising a beam splitter 4 that splits the light into two fiber paths 5, 5', so that the main light beam 3 from the light source 2 generates two light beams 6, 6'. The optical system may be an interferometer system. The beam splitter 4 and the fiber paths 5, 5' are configured such that the first light beam 6 and the second light beam 6' are contained in a single plane, referred to as the "reference plane", or together with their optical axes define such a reference plane. Two converging optical elements 7, 7', such as collimating lenses, are arranged in the propagation paths of the first light beam 6 and the second light beam 6', respectively, such that the two light beams 6, 6' are oriented relative to each other to form a measurement volume at their intersection. In a preferred embodiment, the two light beams 6, 6' are coherent and can interfere with each other, so that the measurement volume contains a number of parallel interference fringes.

[0020] Other configurations of the light source 2 and the optical system 2' are of course possible other than the one presented in the embodiment of Fig. 1. Generally speaking, in an inspection system 1 according to the invention, the at least one light source 2 and the optical system 2' are configured such that a first light beam 6 and a second light beam 6', preferably of coherent light, are contained in a reference plane and intersect therein.

[0021] Continuing with the description of the embodiment presented in Fig. 1, the inspection system 1 also comprises a support 8 for receiving a substrate S. The support 8 is accurately positioned with respect to the optical system 2' such that the first and second light beams intersect and the measurement volume intersects the substrate front-side and defines a measurement spot M on said front-side.

[0022] The substrate S may be a wafer, i.e. a circular plate or disk of material exhibiting two opposing parallel surfaces, called back side and front side, respectively. By convention, the back side of the substrate corresponds to the surface that is brought into contact with or oriented towards the support 8 of the inspection device 1. The wafer may be made of any material that is transparent at the inspection wavelength, such as glass or synthetic sapphire, i.e. single crystal aluminium oxide for visible wavelengths, or a semiconducting material such as silicon for infrared inspection wavelengths. However, the invention is not limited to substrates in the form of wafers, and its principles may be applied to any substrate, in particular any transparent substrate (at the inspection wavelength) of any shape and configuration, as detailed below.

[0023] The support 8 and the optical system 2' are movable relative to each other and the inspection system 1 is designed such that the front side of the substrate S passes through this measurement volume during its movement. As is well known in the art, the support 8 and / or the optical system may be associated with actuators capable of moving one relative to the other to achieve this relative movement.

[0024] The inspection system 1 comprises a controller 12 for controlling the relative movement of the support 8 and the optical system 2′ to scan the measurement spot M along a desired measurement path on the front side of the substrate S. The controller may comprise a microcontroller, a data storage device, input / output ports connected to the actuators and other elements of the inspection system 1, and further computing resources configured in hardware or software to precisely control the displacement of the measurement spot M on the desired measurement path. Advantageously, the controller 12 also controls the speed of the measurement spot M along this path. Once the substrate S is precisely positioned on the support 8, the controller 12 stores and controls the position (linear or polar coordinates) and the speed (linear or rotational speed) of the measurement spot M in a reference linked to the substrate S.

[0025] The inspection system 1 also comprises a collector for collecting at least a portion of the light scattered by defects present on the front side of the substrate S. In the embodiment of Fig. 1, the collector comprises a lens 9 for collecting a portion of the diffuse light radiation. The lens 9 in this embodiment is centered on an axis passing through the measurement spot M and perpendicular to the substrate S. A portion of the scattered light 12 reflected from the substrate S is collected by the lens 9 and directed to a point optically conjugate to the spot M. The collector also comprises a complementary device 10, such as an optical fiber or a bundle of optical fibers, arranged at or near a conjugate point of the lens 9 for collecting the scattered light. The inspection device 1 also comprises a detector 11, e.g. a photodetector or an array of photodetectors, associated with the collector for receiving the collected light. The detector 11 provides a signal U(t), e.g. an electrical signal, representative of the variation in intensity of the collected light as the measurement spot M is scanned along the measurement path.

[0026] 1 comprises a processing device 13 connected to the detector 11 for receiving a signal U(t) as the measurement progresses. The processing device 13 is also connected to a controller 12 for recording the position X of the measurement spot M as the measurement progresses. M (t) from the controller and, in some cases, the speed V M (t) is also received if this information is measured or determined by the controller 12. M (t) is also the position X M (t) or may take a constant value such that this information does not have to be provided by the controller 12 to the processing device 13. The processing device may correspond to or include a microprocessor, microcontroller, FPGA, or any other computing resource (input / output ports, data storage, etc.) for processing the received information, and in particular the signal U(t), in order to detect the presence of particles on the backside and / or the frontside of the substrate and to distinguish between the two situations. This will be exemplified in further parts of this specification. This processing may be performed "on the fly" simultaneously with the measurement or may be deferred after the measurement has been achieved.

[0027] The processing device 13 (or detector 11) samples the signal U(t) at a given sampling rate and produces raw measurement samples U k The device may include a transducer for providing the

[0028] FIG. 2 illustrates the phenomenon occurring at the front side FS and the back side BS of a substrate S positioned on a support 8 of the inspection system 1, when this substrate S is made of a material transparent at the inspection wavelength. FIG. 2 shows a cross-section of the substrate S in a reference plane containing a first light beam 6 and a second light beam 6'. As already mentioned, the first light beam 6 and the second light beam 6' interfere in a measurement volume V intersecting the front side FS of the substrate S to form a measurement spot M on this front side FS. The substrate S is made of a material transparent at the inspection wavelength of the light beams 6, 6', the beams 6, 6' propagate in the thickness Th of the substrate and also illuminate the back side BS of the substrate S with two respective illumination spots S1, S2. The two illumination spots S1, S2 are separated by a distance d that depends on the thickness of the substrate Th and the angle f of the beams propagating in the substrate S. The two light beams 6, 6' are thus separated from each other at the back side BS and illuminate it with two respective illumination spots S1, S2 that are different from each other. The angle f depends on the angle of incidence, i.e. the angle α between the first light beam 6 and the second beam 6' and the surface of the substrate, and on the refractive index of the substrate according to the Snell-Descartes law. If the two beams 6, 6' are relatively angled in the substrate S by an angle f, the distance d is d=2 * Th * It can be determined geometrically as tan(f / 2).

[0029] According to an important aspect, the controller 12 is configured to control the relative movement of the support 8 and the optical system 2' such that the reference plane is kept tangential to the measurement path. This condition means that the velocity vector of the measurement spot M with respect to the substrate is contained in the reference plane or is parallel to the reference plane when this spot M is scanned along the measurement path. The velocity vector therefore extends in the longitudinal direction x in Figs. 1 and 2 and has no component along the transverse direction y. It is understood that in order to ensure that the measurement volume V continues to intersect the front side FS of the substrate S, the velocity vector should also be parallel to the surface of the substrate or at least should not have a significant component along the height direction z. Thus, a velocity vector that is in the reference plane or is parallel to the reference plane, the velocity V of the measurement spot M, M (t) corresponds to the velocity of this spot along the longitudinal direction x.

[0030] The relative movement tangential to the measurement path and manipulated by the controller 12 may be a rotation or a rotation of the support 8 (and of the substrate S), for example along a rotation axis r perpendicular to the plane of the support 8 / substrate S and arranged in the center of the substrate S. The measurement path described by the measurement spot M then forms a concentric circle on the front side of the substrate S. Alternatively, the relative movement may be a straight line along the longitudinal direction x, and the measurement path then forms a straight line on the front side of the substrate S. The inspection system may be manipulated iteratively after the relative displacement of the support 8 with respect to the optical system 2′ to provide multiple measurement paths along parallel, radial or concentric measurement paths to provide a complete inspection of the front side of the substrate S.

[0031] Figures 3a-3c illustrate the advantages of the inspection system 1 when it is used to inspect a substrate S made of a transparent material at the inspection wavelength. The substrate S has a particle P on its back side. In Figure 3a, at the instant t1 when the measurement spot M is scanned on the front side of the substrate S, the particle P is precisely arranged in a first illumination spot S1 formed on the back side by the propagation of a first light beam 6. This light beam is scattered by the particle P, and a part of the scattered light is diffused towards the measurement spot M. This part of the scattered light is collected and detected by the inspection system as if it was generated at the measurement spot M itself. The signal U(t) prepared by the detector 11 therefore shows a first peak of intensity at a first instant t1, as this is represented in Figure 3c.

[0032] Figure 3b represents the substrate S at a moment t2 immediately after the moment t1 represented in Figure 3a. At this moment t2, the particle P is precisely arranged in a second illumination spot S2 formed on the back side by the propagation of a second light beam 6'. This light beam is scattered by the particle P, and a part of the scattered light is diffused towards the measurement spot M. This part of the scattered light is collected and detected by the inspection system as if it was generated at the measurement spot M itself. The signal U(t) prepared by the detector 11 therefore shows a second peak of intensity at the second moment t2, as this is represented in Figure 3c.

[0033] The relative movement is controlled such that the reference plane is tangential to the measurement path, so that a particle P present on the back side that crosses one of the illumination spots S1, S2 will necessarily cross the other illumination spot S2, S1 as well. As a result, the light scattered by the particle P on the back side of the substrate will generate a first pattern in the signal U(t) showing two intensity peaks U(t1), U(t2) separated from each other by a separation interval. This interval can be related to a determined period T=(t2-t1), which corresponds to the time required for the particle P to be moved over a distance d that separates the two illumination spots S1, S2, and this distance d can be expressed as the velocity V of the measurement spot. M By dividing by , i.e., T=d / V MIt can be obtained by

[0034] When the signal provided by the detector 11 is expressed as a function of the position of the measurement spot on the substrate or on the measurement path, the separation corresponds to a distance d.

[0035] (Methods for distinguishing defects) The present invention exploits this observation, which results from the inspection system configuration described above, to propose a method for distinguishing defects present on the front side of the substrate S from defects present on the back side of the substrate S, when the substrate S is made of a material that is transparent at the inspection wavelength.

[0036] The method comprises arranging the substrate S on a support 8 of the inspection system 1, which is positioned relative to the optical system 2' such that the first light beam 6 and the second light beam 6' intersect at the front side of the substrate S. The method then comprises moving the support and the optical system 2' relative to each other to scan a measurement spot M along a desired measurement path on the front side of the substrate S. The relative movement is controlled by the controller 12 such that a reference plane is kept tangential to the measurement path. As the measurement spot M is scanned along the front side of the substrate, at least a portion of the light scattered by the surface of the substrate S, in particular by defects present on the front side and / or back side of the substrate S, is collected on the detector 11 to establish a signal U(t) representative of the variations in the intensity of the collected light.

[0037] According to the invention, the method further comprises identifying in the signal U(t) a first pattern exhibiting two intensity peaks separated from each other by a determined separation distance. As this has been explained in connection with the description of figures 3a-3c, the first pattern corresponds to light scattered by particles on the back side of the substrate S. If such a first pattern is present in the signal U(t), it can therefore be unambiguously established that a particle P is present on the back side of the substrate. The position of the particle P (its coordinates in a plane parallel to the back side of the substrate) corresponds to the position of the measurement spot M at the mean time (t1+t2) / 2, i.e. at half the separation distance between the two peaks.

[0038] Conversely, intensity peaks in the signal U(t) that do not fit the first pattern correspond to particles present on the front side of the substrate FS.

[0039] Since it is possible to distinguish defects present on the front side of the substrate from defects present on the back side of the substrate, the inspection device 1 generates a front side signal U(t), which is representative of defects present only on the front side of the substrate, for example by filtering out the first pattern from the electrical signal U(t). fS Additionally or alternatively, the inspection device 1 may be configured to provide a backside signal U(t) representative of defects present only on the backside of the substrate S. bs (t) may be configured to provide

[0040] Identification of the first pattern in the signal may require calibrating this pattern to the substrate under inspection, in particular to determine the separation distance that separates the two intensity peaks. As indicated above, this separation distance may be pre-calculated from knowledge of the thickness of the substrate and from the angular configuration of the first and second light beams if the scanning speed is kept constant or known. More generally, the separation distance may be determined using at least one of the following information: the angle of incidence of the first and second light beams on the substrate, the refractive index of the substrate, the intersection angle of the first and second light beams at the substrate, the thickness of the substrate, the speed of the substrate relative to the light beams.

[0041] The calibrated first pattern may be stored in the inspection device 1 , for example in the processing device 13 .

[0042] Alternatively, some characteristics of the first pattern, such as the focus, or distance or time interval, of the first pattern, each corresponding to a type of substrate, may be stored in the inspection device and retrieved according to the properties of the substrate under test before inspection begins.

[0043] If the velocity of the measurement spot is not kept constant during the measurement path, the separation interval of the first pattern may be calculated or adjusted to the actual velocity at each given instant, and the first pattern is dynamically calibrated.

[0044] Raw measurement sample U k When the signal is sampled by a transducer of the processing device 13 to establish, the separation interval may be expressed as a determined number of samples 2s, and identifying the first pattern in the signal involves identifying two intensity peaks separated by a determined number of samples 2s in successive raw measurement samples.

[0045] This identification of the first pattern can be performed by the processing device 13 according to many different schemes. For example, the processing device may implement any kind of pattern recognition algorithm. It may for example implement a correlation method that correlates the signal with a template pattern signal. It may also implement a pattern recognition algorithm trained by machine learning, which sets an indicator signal if the first pattern is recognized in the signal as the measurement progresses.

[0046] Alternatively, the intensity peaks can be calculated from the raw measurement sample U k It can be established (for example, by numerical derivation) that the first pattern is the same as the first pattern in the raw measurement sample Ui, Uj if the indices i, j of the samples Ui, Uj corresponding to two successive peaks differ by a predetermined number of samples, i.e., ji = 2s. k This operation may be performed "on the fly" as the raw measurement sample is established by the processing device 13. The raw information sample may be stored in a data store included in (or associated with) the processing device 13, in which case the identification of the first pattern may be deferred until after the test is completed.

[0047] Regardless of whether the identification is performed "on the fly" or in a deferred manner, it is advantageous to also provide the position of the first pattern in the raw measurement sample, i.e., the index i+s or js. This index information is then used to calculate the position X of the measurement spot M as the measurement progresses. M(received from controller 12), it is possible to precisely place particle P at the origin of the identified first pattern on the backside of the wafer. The particle locations can be stored as the inspection progresses, displayed as a substrate map, or otherwise processed at the end of the inspection sequence.

[0048] (Identification by applying a transformation function) The intensity peaks of the electrical signal corresponding to the backside particles P may be of relatively low amplitude because only a portion of the light scattered by such particles P is directed to the measurement spot M, collected and detected.

[0049] Advantageously, therefore, the identification of the first pattern comprises a first step of applying a transformation function to the signal U provided by the detector 11. The transformation function is chosen so as to map this signal U into a transformed signal V, in which a first pattern present in the original signal U is transformed into a second pattern of greater intensity, or more generally of greater detectability, in the transformed space.

[0050] The identification then comprises a second step of detecting a second pattern in the transformed signal V.

[0051] To illustrate this principle, according to a particular embodiment, the transformation may be implemented by the following function (when processing digital raw measurement samples):

[0052]

number

[0053] For ease of implementation by the processing device 13, the function can also be expressed as: W k represents a mask consisting of 6s+1 elements, only four of which are non-zero, while the remaining 6s-3 entries are zero.

[0054]

number

[0055] [Table 1]

[0056] Original sample U k contains a first pattern corresponding to the backside grain. k The behavior of is shown in Figure 4a. It includes a second pattern that shows a single peak at the midpoint between the two peaks of the first pattern, and the only peak of the second pattern is precisely located at the backside defect location. The amplitude A of this peak is measured by the raw measurement sample U k The amplitude A / 2 of the first pattern peak at is twice the amplitude A / 2 of the first pattern peak at . The transformed samples also include negative values, which are not shown in the figure and are irrelevant; they are simply set to zero.

[0057]

number

[0058] Converted sample V k features a very different behavior for a single peak in the signal generated by the nearside particle, as shown in Fig. 4b. In this case, the transformed sample shows two peaks separated by an interval of length 2s. Specifically, if the nearside particle is generated with a peak centered at instant k, the transformed sample V k The corresponding profile of V k+S and V k-S The amplitudes of the two peaks on the transformed signal are the same as they would have been in the original signal.

[0059] Note that the signal illustrated in Figure 4b is for a particle that is large enough to average out the high frequency signal, which would otherwise result in a smaller particle crossing the interference fringes of the two interfering beams, or alternatively, this signal could result from any particle crossing the two beams that are mutually incoherent and therefore do not interfere.

[0060] In summary, the converted signal V or sample V according to this embodiment k has the following properties: • It is zero in the region of constant intensity. • It is independent of background signal levels, i.e. the average signal value in the absence of particles (sometimes referred to in the art as "haze"). ●Has one pulse at the backside particle location. • Has two separate pulses around the front particle.

[0061] A transformation function with such properties is a function of the transformed signal or sample V k This transformed signal can also be used alone. It can also be used to better differentiate between back and front particles by detecting, for example, a single peak in the raw measurement signal that coincides with a bimodal peak in the transformed signal and corresponds to a front particle, and a bimodal peak in the raw measurement signal that coincides with a single peak in the transformed signal and corresponds to a front particle. k may be used in combination with

[0062] Referring to FIG. 5, according to this preferred embodiment, in which the identification of the first pattern involves the application of a transformation function, the identification may include the following steps performed by the processing device:

[0063] First, in step 20, the transformed sample V k But for all the original samples U kThese samples may be subject to noise, and optionally, to eliminate the noise, the transformed samples V that exceed a first specified threshold are calculated on the fly or in a deferred manner as described above. k Only those samples Vs are retained for further analysis. Then, in step 22, maxima in the transformed samples are identified. This can be done, for example, by derivation or by selecting samples that exceed a second specified threshold. In step 24, all maxima that have similar amplitudes and are separated by an interval 2s are ignored, since they correspond to front side particles. This detection can be enhanced by verifying that pairs of these maxima correspond to peaks in the raw measurement signal. The remaining transformed samples Vs are then k contains only a single peak each due to a backside grain so that the positions of these grains can be obtained.

[0064] In step 26, the original signal or sample can be processed using the positions of the backside particles to either prepare a frontside signal representative of defects present only on the front side of the substrate and / or to prepare a backside signal representative of defects present only on the backside of the substrate.

[0065] (Removal of the contribution of backside particles in the signal provided by the detector) According to one embodiment, the positions of the backside particles may be used to process the original signal or sample to prepare a frontside signal representative of defects present only on the front side of the substrate, for the purposes of this section of the specification, it being understood that a similar method may be used to prepare a backside signal representative of defects present only on the backside of the substrate.

[0066] As a preamble, the signal provided by the detector 11 (or the raw measurement sample U k) is the sum of the background portion b and the particle scattering portion. As mentioned above, the amplitude (A) of the second pattern peak is twice the amplitude (A / 2) of the corresponding peak in the first pattern. The intensity profile of the laser beam is well approximated by a Gaussian profile with a known standard deviation a expressed in number of samples: Thus, the double peak signal generated for sample number i by a backside particle located at sample number m is modeled as: where s is half the separation interval expressed in number of samples.

[0067]

number

[0068] According to this embodiment, the background level of the original signal or raw measurement sample is determined. The method comprises determining the amplitude A of the peak in the second pattern identified in the transformed sample and subtracting half this amplitude, A / 2, from the peak amplitude in the raw measurement sample to establish a background level b. This in effect provides two background level estimates, one for each of the two peaks. The estimated background level may be chosen to correspond to the average value, where m is the sample number corresponding to the detected position of the background particle. Then, using the above model U BS Based on the determined amplitude level A / 2, the position of the particle m, and the background level b, a reference signal U'(i) of light scattered by the backside particle is prepared.

[0069]

number

[0070] The final step is to subtract the reference signal from the signal provided by the detector, which removes the backside particle signature and leaves the frontside particle contribution intact in the signal.

[0071] Figures 6a and 6b illustrate this embodiment: Figure 6a shows a noise signal from the detector with one single peak (front grain) and one double peak (back grain).

[0072] FIG. 6b demonstrates the signal filtered from the backside particle signature, while the contribution from the nearside particles is kept unchanged.

[0073] Alternatively, once the double peak corresponding to the backside particles is identified, a model function such as that described above can be fitted to the raw measured signal using any curve fitting or least squares method and subtracted from this signal to retain only the background noise.

[0074] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

Claims

1. 1. A method for distinguishing defects present on a front side of a substrate from defects present on a back side of the substrate, the substrate being made of a material that is transparent at an inspection wavelength; The method comprises: - placing the substrate in an inspection system comprising at least one light source coupled to an optical system for emitting a first light beam and a second light beam, the first light beam and the second light beam having the inspection wavelength and contained in a reference plane and intersecting thereat, the substrate being arranged on a support and positioned and arranged with respect to the optical system such that the first light beam and the second light beam intersect at a measurement spot on the front side of the substrate and also illuminate the back side of the substrate with two respective illumination spots separated by a given distance; - controlling a relative movement between the support and the optical system to scan the measurement spot along a measurement path on the front side of the substrate, the relative movement being controlled such that the reference plane is kept tangential to the measurement path; - collecting at least a portion of light scattered by defects present on the front side and / or the back side of the substrate and establishing a signal representative of the variation in intensity of the collected light as the measurement spot is scanned along the measurement path, The method further includes identifying a first pattern in the signal that corresponds to the light scattered by a defect on the backside of the substrate and that presents two intensity peaks separated from each other by a determined separation interval that corresponds to the time required for the defect to be moved over the distance separating two of the illumination spots.

2. The method of claim 1 , further comprising filtering out the first pattern from the signal to provide a front-side signal representative of defects present only on the front side of the substrate.

3. The method of claim 1 , further comprising providing a backside signal representative of defects present only on the backside of the substrate.

4. 2. The method of claim 1, wherein identifying the first pattern includes determining the separation distance using at least one of information about an intersection angle of the first light beam and the second light beam at the substrate, a thickness of the substrate, and a moving speed of the substrate relative to the first light beam and the second light beam.

5. The method of claim 1 , wherein the identifying the first pattern includes generating a reference signal or a mask signal using information about the separation interval.

6. - sampling said signal at a given sampling rate to prepare raw measurement samples; establishing the separation distance as a number of samples determined based on an intersection angle of the first light beam and the second light beam, a thickness of the substrate, a relative movement speed, and the sampling rate; 2. The method of claim 1, wherein the identifying the first pattern in the signal comprises identifying two intensity peaks in at least some of the raw measurement samples, the peaks being separated by the determined number of samples.

7. The identification of the first pattern includes: a first step of applying a transformation function to at least some of the raw measurement samples to provide transformed samples, the transformation function being configured to map a first pattern in the raw measurement samples to a second pattern of greater intensity in the transformed samples; a second step of detecting said second pattern in said transformed samples.

8. The method of claim 7 , wherein the second pattern includes peaks that provide the location of the first pattern in the raw measurement sample.

9. The method of claim 7 , wherein the second step further provides an average intensity of two peaks of the first pattern and a background intensity.

10. 2. The method of claim 1, wherein the substrate is disk-shaped and presents an axis of symmetry, and the relative movement between the support and the optical system is a rotation about the axis of symmetry of the substrate.

11. 1. An inspection system for distinguishing defects present on a front side of a substrate from defects present on a back side of the substrate, the substrate being made from a material that is transparent at an inspection wavelength, the inspection system comprising: at least one light source coupled to an optical system for emitting a first light beam and a second light beam, said first light beam and said second light beam having said test wavelength and contained in a reference plane where they intersect; a support for receiving the substrate and for positioning the substrate relative to the optical system such that the first and second light beams intersect at a measurement spot on the front side of the substrate and also illuminate the back side of the substrate with two respective illumination spots separated by a given distance, the support and the optical system being movable relative to each other; a controller for controlling the relative movement of the support and the optical system to scan the measurement spot along a measurement path on the front side of the substrate, the controller being configured to control the relative movement such that the reference plane is kept tangential to the measurement path; and a collector for collecting at least a portion of the light scattered by defects present on the front side and / or the back side of the substrate; a detector associated with said collector for receiving said collected light and for providing a signal representative of the variation in intensity of said collected light as said measurement spot is scanned along said measurement path; The inspection system also includes a processing device connected to the detector, the processing device being configured to process the signal to identify a first pattern corresponding to the light scattered by a defect on the back side of the substrate and presenting two intensity peaks separated from each other by a determined separation interval corresponding to the time required for the defect to be moved over the distance separating two of the illumination spots.

12. The processing device also - filtering out the first pattern from the signal to provide a front-side signal representative of defects present only on the front side of the substrate; and / or The inspection system of claim 11, configured to provide a backside signal representative of defects present only on the backside of the substrate.

13. 12. The inspection system of claim 11, wherein the processing device comprises a converter for sampling the signal at a given sampling rate and preparing a raw measurement sample.

14. The inspection system of claim 13 , wherein the processing device comprises an FPGA and / or a processor for identifying the first pattern at the given sampling rate.

Images