WO2022129721A1

WO2022129721A1 - Method for inspecting a surface of an object

Info

Publication number: WO2022129721A1
Application number: PCT/FR2021/052091
Authority: WO
Inventors: Tristan COMBIER
Original assignee: Unity Semiconductor
Priority date: 2020-12-18
Filing date: 2021-11-25
Publication date: 2022-06-23
Also published as: FR3118175A1

Abstract

The invention relates to a method employing a device (1) for inspecting a surface (S) of an object. The device comprises a polychromatic light source (20) for projecting an inspection beam onto the surface (S); a confocal mask (6a, 6b) that intercepts the inspection beam and a beam reflected by the surface, and that comprises a plurality of color-filtering apertures; a chromatic system (3) that spatially spreads the focus of the inspection beam over focal planes spread out along an optical axis (AO) in a depth of field, and that intercepts the reflected beam so as to project it onto a detection plane (P) conjugated with the focal planes; a movable holder (4) for positioning the surface (S) in the depth of field; a time-delay and integration image sensor (5) synchronised with the movement of the surface, the image sensor (5) comprising a matrix array of photodetectors placed in the detection plane (P), the color-filtering apertures of the confocal mask illuminating some of the photodetectors.

Description

DESCRIPTION

TITLE: METHOD FOR INSPECTION OF A SURFACE OF AN OBJECT

FIELD OF THE INVENTION

The present invention relates to a device and a method for inspecting a surface of an object. A device and a method in accordance with the invention allow the inspection of the surface of the object when the latter comprises strong variations in height and makes it possible to generate high-speed intensity images representative of the reflectivity of this surface. with extended depth of field and great lateral resolution. In general, the invention relates to the field of imaging for the inspection and control of objects, such as substrates for the semiconductor industry or integrated optics (semiconductor wafers or substrates in all forms, in progress or at the end of manufacture of semiconductor devices), or other objects in the fields of mechanics or manufacturing industry for example (micro-mechanisms, parts or assembly of watchmaking, assembled systems) .

TECHNOLOGICAL BACKGROUND OF THE INVENTION

In the field of the semiconductor industry, the increase in the production rates of semiconductor device wafers and the reduction in the size of the patterns create a growing need for high-speed imaging inspection. The methods for imaging these wafers must therefore allow the rapid localization of defects, structures and patterns over the entire surface of the wafers. Furthermore, due to the curvature of the wafers in particular, or due to the height of the structures to be imaged (edges of the wafer, etc.), it is necessary to be able to produce sharp images in an area of extended depth of field. In other fields, such as those of micromechanics or watchmaking, the objects to be inspected may have significant relief or height differences, also requiring extended depths of field.

Optical imaging systems such as conventional microscopes are known. These systems are generally limited in depth of field: the optical diffraction limit makes it very difficult to have both high magnification and a large depth of field. This is particularly penalizing when inspecting distant surfaces or structures with significant heights or depths. This problem is generally solved by maintaining the objective at the desired distance from the measured surface, by an autofocus type system. This translates into lower travel speeds and more complex systems, or even the need to perform multiple passes to acquire images.

Document US 8,654,324 implements a method for increasing the depth of field by exploiting the chromatic dispersion of a chromatic lens. Light transmitted through a slit is projected onto an object through a chromatic lens. The light reflected by the inspected surface is returned to the chromatic objective then transmitted through a slit placed in front of a linear detector. The chromatic objective used in this configuration makes it possible to increase the measurement depth of field by exploiting the fact that different wavelengths are focused at different distances on the object. However, this system is not suitable for forming intensity images in two dimensions and with high lateral resolution. This is due to its slit configuration which does not meet the criteria for confocal detection in one dimension and thus generates crosstalk or cross-talk between measurement channels.

Document US 9,739, 600 B1 describes a system allowing the location of patterns in two dimensions with an extended depth of field obtained by a chromatic lens. The profile or the distance of the identified structures can be measured by the same instrument respecting the criteria of a detection confocal, with a minimum of cross-talk between measurement channels. The acquisition speed of the system is however directly limited by the quantity of light collected by the photodetector and especially by the number of measurement channels that it is possible to implement.

More particularly, the inspection rates depend on the optimization of the filling factor of the measurement points on the inspected wafers and defining the density of points measured at each instant. The fill factor is defined as the ratio between the diameter of the measurement points (diameter of the beam of light projected on the wafer at the point of focus) and the separation between two measurement points. Thus, the use of optical fibers or integrated optical circuits poses a physical limitation on the possibility of bringing the measurement channels closer together.

One of the objectives of the invention is to propose a device and a method for rapidly imaging over an extended depth of field a surface of an object.

Another object of the invention is to propose such a device and a method allowing a high density of measurement points (high spatial resolution) while minimizing the coupling of light (cross-talk) between the detection channels.

Another object of the invention is to allow localization and inspection of structures and patterns of a surface of an object such as a semiconductor substrate. The patterns can in particular be solder bumps placed on the conductive pads of the electronic chips of a wafer.

BRIEF DESCRIPTION OF THE INVENTION

With a view to achieving one of these aims, the invention proposes a method for inspecting a surface of an object, the method including:

- a first step of setting up during which the object is placed on a mobile support of an inspection device so as to position the surface in a depth of field of a chromatic system of this device, the system chromatic spatially spreading the focusing of a polychromatic inspection beam according to focusing planes in said depth of field arranged along an optical axis, and intercepting a beam reflected by the surface in order to project it onto a matrix of photodetectors of a time-delayed integration image sensor arranged in a detection plane, conjugate to the focusing planes;

- a second step of activating at least one polychromatic light source to project the inspection beam onto the surface and thus form the reflected beam, at least one confocal mask having a plurality of chromatic filter openings intercepting the beam inspection and the reflected beam, the chromatic filtering openings of the confocal mask making it possible to illuminate part of the photodetectors;

- a third survey step during which the mobile support is controlled to move the surface relative to the chromatic system at least along a direction of inspection perpendicular to the optical axis and to travel a measurement trajectory, and during which controls the image sensor to synchronize it to the movement of the surface;

- a fourth step of acquiring the data established by the image sensor and preparing an image formed of pixels respectively associated with measurement points on the surface of the object, each pixel reflecting the intensity of the reflected radiation by the measuring point of the surface of the object.

According to other advantageous and non-limiting characteristics of the invention, taken alone or according to any technically feasible combination: the object is a chip wafer, the wafer being provided with a plurality of solder bumps; - The weld bumps are arranged in a line, and the traj ectory is at least partly parallel or perpendicular to the lines of weld bumps;

- The traj ectory is formed of a plurality of passages having different altitudes;

- The third and/or the fourth step are implemented by a control unit of an inspection device;

the inspection method comprises a step of analyzing the image to locate the arrangement of protruding elements of the surface;

the analysis step leads to determining whether the arrangement of salient elements conforms to an expected arrangement;

- the image sensor is a color or multispectral image sensor providing color or spectral information, and the method comprises a step of converting the color or spectral information provided into distance information to form a 3D image the surface of the object;

- the array of photodetectors is arranged in the detection plane and arranged in a plurality of rows and columns, the charges generated by the photodetectors being able to be moved from one row to another, in a column direction synchronized with the movement according to the direction of inspection;

- the chromatic filtering apertures are arranged on the confocal mask to illuminate a plurality of photodetectors arranged on a column of the matrix corresponding to a plurality of measurement points arranged along the direction of inspection; - the chromatic filtering openings are distributed in a complementary manner over at least two lines of the confocal mask to illuminate a plurality of photodetectors corresponding to a plurality of measurement points arranged along at least two lines arranged in a direction perpendicular to the direction of inspection ;

- the confocal mask comprises a plurality of groups of at least two lines comprising chromatic filtering apertures distributed in a complementary manner over said at least two lines;

- the inspection method comprises a confocal lighting mask placed in the inspection beam and a confocal detection mask placed in the reflected beam, the confocal lighting mask and the confocal detection mask being respectively placed in a conjugate plane of focal planes with respect to the chromatic system;

- The inspection method comprises a single confocal mask, placed in a conjunct conjugated plane of the detection plane and the focusing planes vis-à-vis the chromatic system;

- The polychromatic light source comprises a halogen lamp or a strip of light-emitting diodes, associated with a diffusion device to homogenize the intensity of the inspection beam;

the polychromatic light source has a line spectrum, having a determined number of emission wavelengths or bands of wavelengths;

- the chromatic system includes a chromatic lens; the color system includes a separator element; - the photodetectors are equipped with a color filter or a spectral filter;

- the chromatic filtering openings are arranged to each illuminate a plurality of photodetectors provided with filters of different colors or spectral bands.

BRIEF DESCRIPTION OF THE DRAWINGS

Other characteristics and advantages of the invention will emerge from the detailed description of the invention which will follow with reference to the appended figures in which:

[Fig. 1] Figure 1 shows an inspection device according to one embodiment of the invention;

[Fig. 2] FIG. 2 represents a block diagram of the confocal chromatic operation implemented by the device of FIG. 1;

[Fig. 3] Figure 3 shows a front view of a detection mask illuminated by a reflected beam;

[Fig. 4] Figure 4 shows an aperture arrangement on a confocal mask in alignment with a photodetection array;

[Fig. 5] Figure 5 shows two inspection fields on the surface of a surface to be inspected.

DETAILED DESCRIPTION OF THE INVENTION

For the sake of simplifying the description to come, the same references are used for identical elements or providing the same function in the state of the art or in the various exposed embodiments of the device. FIG. 1 represents an embodiment of an inspection device according to the present invention.

In general, the inspection device 1 projects an inspection light beam onto a surface S of an inspected object. The beam is reflected on this surface S to project onto a detection plane P of an image sensor 5.

The inspected object, such as a semiconductor substrate, is placed on a mobile support 4 of the inspection device 1, for example a table whose movement can be controlled. The mobile support 4 makes it possible to move the surface along at least one inspection direction I under the inspection beam.

The inspection device 1 comprises a lighting system 2 comprising a polychromatic light source 20 to form the inspection beam. The source 20 therefore emits several wavelengths which may for example be included in the visible spectrum, in an interval of wavelengths included between 380 nm to 700 nm, or in the infrared beyond 700 nm.

In general, the wavelengths making up the emission spectrum of the source 20 are chosen according to the nature of the object inspected and the layers which are likely to cover it. The wavelengths of the polychromatic source 20 will be chosen so that the inspection beam is reflected by the surface S of the object to be inspected. In addition, when a covering layer is present, it is possible to choose the wavelengths of the polychromatic source 20 so that they are included in the transparency spectrum of this covering layer.

By way of example, the polychromatic light source 20 can be formed by a halogen lamp or a strip of light-emitting diodes. It may be a source emitting a broad spectrum to emit so-called "white" light, or a source exhibiting a line spectrum, and therefore exhibiting a determined number of inspection wavelengths or bands. inspection wavelengths. Each line can be provided by a laser source or by a source covering a narrow band of wavelengths.

It is generally sought to form an inspection beam of large transverse dimension so as to illuminate a large portion of the inspected surface S, referred to as the “inspection field” in the remainder of this description. In an application in the semiconductor field, the inspection field may have a dimension of between a few hundred microns and a few millimeters on a side. To this end, it is possible to provide, as is the case in the mode of implementation represented in FIG. 1, a diffusion device 21 to homogenize the intensity of the inspection light beam produced by the lighting system 2. The diffusion device 21 can for example be made of frosted glass or of an optical guide element ensuring a mixture of the modes of propagation of the radiation produced by the polychromatic source 20 so as to make the intensity of the beam of inspection.

It is also possible to provide optical fibers 22 or any other form of guide for light radiation, to conduct the light beam produced by the source 20 to the injection zone of the inspection beam in the device 1, here occupied by the diffusion device 21.

Of course, the lighting system 20 is not limited to that described in this mode of implementation. One can for example consider that it is composed of a supercontinuum polychromatic light source whose radiation is guided by at least one optical guide as far as the injection zone.

Continuing the description of FIG. 1, the inspection device 1 also comprises a chromatic system 3 placed on the optical path of the inspection beam and the reflected beam. The chromatic system 3 generates a chromatic dispersion of the light which passes through it and thus spatially spreads the focusing of the polychromatic inspection beam, according to the wavelengths present, according to focusing planes distributed along an optical axis AO in a depth of field of the color system 3 extended compared to the depth of field obtained for an individual wavelength.

For this purpose, the chromatic system 3 comprises a chromatic objective 3a resulting in focusing the different wavelengths of the inspection beam, over its entire transverse extent, at different distances from the objective. The chromatic objective 3 can be produced in multiple ways, for example by means of a lens or a combination of lenses. These lenses can intercept the entire transverse extent of the inspection beam, or be respectively dedicated to portions of this beam constituting measurement channels. They can also be diffraction lenses or holographic devices. They can also be metallenses, that is to say reflective or transmissive optical elements having structured reflection or transmission surfaces. The patterns constituting this surface structuring have smaller dimensions than the wavelengths of the inspection light beam and affect the phase of the radiation of this beam and therefore its shape and its propagation. The surface patterns of the optical elements can be produced by known methods of photolithography and etching.

Whatever the nature of the chromatic objective 3a, the latter is chosen to present a significant chromatic aberration, leading to the spatial spread of the inspection beam, according to the wavelengths which compose it, according to focal planes arranged in the depth of field of the chromatic system 3. This depth of field can for example extend over a distance of 100 to 200 microns. During a measurement, the surface S of the inspected object is placed in the depth of field of the chromatic system 3. For this purpose, provision can be made for the mobile support 4 to be able to adjust the position of the surface S by moving the inspected object. along the direction of the axis AO of the chromatic system 3, so as to maintain the surface S of the object in the depth of field. This characteristic is particularly advantageous when the surface S has significant reliefs in elevation or in depression.

Consequently, and advantageously, the mobile support 4 may consist of a table provided with means that can control its movement in a plane perpendicular to the inspection beam in order to move the surface in the direction of inspection I, and complementary means ensuring a displacement in a direction perpendicular to this plane in order to maintain the surface S in the depth of field of the chromatic system 3. However, insofar as the depth of field of the chromatic system 3 is extended, the constraints for maintaining there surface S are relaxed with respect to an achromatic system.

The displacement in space of the surface S during a measurement sequence can be controlled by an electronic control unit 7. This can also collect the trajectory carried out during the measurement sequence in order to be able to reconstitute a faithful image of the inspected surface S.

Continuing the description of the mode of implementation represented in FIG. 1, the chromatic system 3 is also provided with a separating element 3b, for example a separating cube or a reflecting plate, to intercept the beam reflected by the surface S and the project onto a detection plane P of an image sensor 5. The splitter element 3b is arranged on the optical path of the reflected beam, downstream of the chromatic optics 3a, so that the reflected beam passes through this lens 3a before projecting onto the detection plane P. Other optical parts of the chromatic system 3 can be placed on the optical path between the separator element 3b and the detection plane of the image sensor 5, but in any case the detection plane P is arranged in a plane conjugate, vis-à-vis the chromatic system 3, of the focal planes.

According to the invention, the image sensor 5 is a time-delayed integration image sensor (referred to as “image sensor” in the rest of this presentation for simplicity). The detailed operation of such a sensor is for example described in the document EP2088763. It comprises a matrix of photodetectors arranged in a plurality of rows and columns. The charges generated by the photodetectors are moved from one row to another, along a column direction synchronously with the movement of the imaged object. The last line of the matrix is moved in a shift register, or any other form of electronic component making it possible to extract the charges from this line, in order to constitute an image line of the object.

In the inspection device 1, the matrix of photodetectors of the image sensor 5 is arranged in the detection plane P onto which the reflected beam is projected. In other words, the image sensor 5 is placed in the inspection device 1 to image the inspection field. Such a matrix can have a very large number of rows and columns, for example more than 10,000 columns and several hundred rows.

The image sensor 5 is connected to the electronic control unit 7 in order to synchronize it with the movement of the surface S along the direction of inspection I, and to repeatedly recover the lines of charges extracted from the matrix of photodetectors at the end of columns. In this configuration, and in the absence of any masking, a point on the surface S which moves along the direction of inspection I is imaged successively on the photodetectors of a column of the matrix, that is to say on successive rows of the column. Simultaneously, the electrical charge generated by a photodetector is moved synchronously from one line to another, along a column direction, so that the charge accumulated at the end of this scan, at the end of the column, is representative of the light reflected by the point on the surface S throughout the duration of its displacement, in the conjugate plane of the image sensor 5, along the corresponding column of the sensor in the inspection direction I.

The inspection device represented in FIG. 1 also comprises two confocal masks 6a, 6b arranged in conjugate planes of the surface S of the object with respect to the chromatic system 3, and respectively placed in the optical path of the beam inspection and the reflected beam.

A first confocal lighting mask 6a is here arranged on the diffusion device 21. The mask has a plurality of lighting apertures. These openings can be crossed by the inspection beam from the lighting system, and the confocal lighting mask blocks the beam outside these openings. The inspection beam therefore consists of a plurality of lighting paths defined by the lighting apertures, projecting onto the surface S at the level of a plurality of measurement points distributed in the inspection field of the device 1, according to planes of focus according to the wavelengths in the depth of field of the chromatic system 3.

A second confocal detection mask 6b is here arranged on the detection plane of the image sensor 5. Similar to the first lighting confocal mask 6a, the confocal detection mask 6b comprises a plurality of chromatic filtering apertures (or confocal chromatic), combined with the lighting apertures of the confocal lighting mask 6a. More precisely, the lighting apertures and the chromatic filtering apertures are conjugated from the same measurement points by the chromatic system 3. The confocal detection mask 6b and the image sensor 5 are arranged relative to each other so that the chromatic filtering openings of the confocal mask 6b are respectively placed opposite a photodetector or a group of photodetectors and illuminate part of the photodetectors of the array.

For this, the confocal detection mask 6b can be plated, glued, or produced directly on the surface of the image sensor 5. It can in particular be produced by selective deposits of layers of dielectric or metallic materials on the surface of the sensor 5 .

The confocal detection mask 6b can also be produced in the form of an element distinct from the sensor 5, positioned in a conjugate optical plane of the sensor by a lens or an imaging system.

In the inspection device 1 thus configured, the illumination paths of the inspection beam pass through the chromatic objective 3a. The beam is focused, according to the different wavelengths which constitute it, in focusing planes distributed over the very wide depth of field by the effect of the chromatic lens 3a. The radiation from each illumination channel is reflected by the inspected surface S, arranged in the depth of field, at the measurement points, again passes through the chromatic objective 3a, and is projected onto the confocal detection mask 6b by the intermediary of the separating element 3b.

There is thus shown in Figure 2 a block diagram of the confocal chromatic operation implemented by the device of Figure 1. In this figure, the lighting mask 6a has a lighting aperture defining an inspection beam comprising here a single lighting channel. By the effect of the chromatic objective 3a disposed on its optical path, the inspection beam is focused, depending on the optical wavelengths present, on a extended depth of field. In FIG. 2, there is shown a first wavelength Xo of the beam which is focused in a focusing plane PfO and a second wavelength Xi of the beam which is focused on a second focusing plane Pfl, quite distinct from the foreground.

The surface S to be inspected has been placed at the level of the first focusing plane PfO, so that the illumination aperture of the illumination mask is imaged there for the first wavelength. The light at the first wavelength Xo coming from the illumination aperture is therefore reflected on the surface S to be focused on the detection plane P of the image sensor 5, in the chromatic filtering aperture of the mask detection 6b. Thus, this light at the first wavelength Xo reaches in all its intensity or almost the detection plane P through the opening of the mask.

On the contrary, the light at the second wavelength Xi is not perfectly focused on the first focusing plane PfO in which the surface S resides. As a result, this light is diffusely reflected, up to the detection mask 6b which largely blocks it, so that little of its intensity reaches the detection plane P.

There is shown in Figure 3, a front view of the detection mask 6b, which here has 12 chromatic filtering apertures O arranged in 4 columns and 3 rows. The mask 6b is illuminated by the reflected beam in an assembly similar to that shown in FIG. 2. Also shown in FIG. 3 is the outline of the luminous halo H corresponding to the projection onto the mask of radiation at wavelengths reflected outside their focal planes on the inspected surface S.

To limit the phenomenon of coupling between different detection channels, it is important that the openings of the masks 6a, 6b are sufficiently distant from each other, so that the diffuse halo associated with an opening does not cover a neighboring opening with too much intensity.

For example, the centers of two adjacent openings on the illumination mask 6a or on the detection mask 6b can be separated by a distance greater than twice the dimension of these openings (their diameters if these openings are circular).

However, the fact of moving the openings of the masks 6a, 6b away from each other leads to a reduction in the filling factor of the measurement points in the inspection field. As noted in the introduction, a reduction in the fill factor affects the inspection rate and the density of the measurement points.

To solve this problem, the openings of the lighting and detection masks are cleverly arranged vis-à-vis the photodetectors of the image sensor 5.

There is thus represented in FIG. 4 a part of the photodetector matrix in the form of a grid G. Each box of the grid symbolizes a photodetector or a group of photodetectors. The part of the matrix shown is composed of 5 columns (referenced C1 to C5) and 5 rows (referenced L1 to L5). The openings of the masks 6a, 6b are represented by circles, arranged here in alignment with the photodetectors of the camera. These openings are capable of passing the reflected beam which is then projected onto the exposed photodetectors to be measured by the image sensor 5. It is noted that these are indifferently the openings of the lighting mask 6a or of the detection 6b, these two masks being optically conjugated to each other when the photodetectors are exposed.

In this arrangement, the chromatic filtering openings are arranged on the confocal mask (6a, 6b) to respectively illuminate a plurality of photodetectors, or a plurality of groups of photodetectors, arranged on a column (C1 -C5) of the matrix. The image sensor 5 is synchronized with the movement of the surface to be inspected so that the plurality of photodetectors of a column under the openings is in correspondence with a plurality of measurement points arranged along the direction of inspection (I) . When the surface S is moved along the inspection direction I, the photodetectors of the column under the openings of the mask are successively exposed to the light coming from a measurement point. At the end of the column, a charge is recovered corresponding to the quantity of light reflected by the measurement point during an extended exposure time, corresponding to the exposure time of each photodetector multiplied by the number of openings in the column. It is thus possible to produce a high-intensity image of the inspected surface and/or to increase the rate of inspection.

In this arrangement also, the chromatic filtering openings are distributed over the confocal mask 6b in a complementary manner over at least two lines, that is to say for example that for a given column, if a line does not include an opening chromatic filtering, another line includes one. In particular, such a complementary distribution makes it possible to ensure that, for a group of rows which each include a chromatic filtering opening, for a subset of the columns, each column includes the same number of chromatic filtering openings (for example a ) for the row group . In this way, a plurality of photodetectors are also distributed in a complementary manner over at least two rows of the matrix. In the example of FIG. 4, the photodetectors of the lines L1 and L2, and of the lines L3 and L4 are thus arranged in a complementary manner on these two lines. In other words, two contiguous photodetectors in a row or in a column are not simultaneously exposed under an aperture of the mask. In this way, the openings are spaced sufficiently apart from each other to limit the parasitic coupling between several detection channels. The image sensor 5 is synchronized with the movement of the surface S to be inspected such that the photodetectors distributed over the lines correspond to a plurality of measurement points arranged similarly along lines arranged in a direction perpendicular to the direction of inspection (I). By virtue of the complementary distribution of the chromatic filtering openings between the lines, when the surface S is moved along the inspection direction I, the measurement points imaged on the detector combine very densely in the inspection field.

There is thus represented in FIG. 5 a first inspection field Ci 1 as well as the measurement points (in dotted lines) imaged on the image sensor 5 of this inspection field. The surface has been moved along the direction of inspection and a second inspection field CI2 has been shown as well as the measurement points in this field (in solid lines). The distribution over several complementary lines of the openings makes it possible, as the object moves, to carry out dense measurements in the field of inspection.

Of course, the openings can be distributed in a complementary manner over a larger number of lines, which makes it possible to space these openings even further from each other and further limit the coupling phenomenon. For a given dimension of the photodetector array, this amounts to placing fewer photodetectors on a column, and therefore affecting the intensity quality of the image.

Advantageously, the confocal mask (6a, 6b) can comprise a plurality of groups of lines with a complementary distribution of the chromatic filtering openings. These groups of lines can for example comprise an identical distribution of chromatic filtering openings. Thus each measurement point of the object is imaged on the detector in a plurality of chromatic filtering apertures , sequentially , which makes it possible to accumulate more intensity or charges and to fully exploit the number of lines of color . the array of photodetectors. The device as described makes it possible to obtain images in intensity, or in gray level, of the surface of the object, or in other words of its reflectivity.

Insofar as the chromatic system encodes the height of the object in wavelength, it is also possible to obtain depth information. This requires determining the wavelength of the radiation detected by the photodetectors. This can be implemented by employing a color image sensor, thus returning color information in addition to the intensity information of the detected radiation. By “color image sensor” is meant any image sensor capable of associating wavelength information with the intensity of the radiation received at the level of a photodetector. These sensors are therefore able to form color images, for example by coding the color information in the form of RGB pixels, as is very conventional, in which each pixel is associated with the light intensity detected in the red, the green and blue. But more generally, the color image sensors according to the present description are capable of providing hyperspectral images, a pixel of such an image being associated with the light intensity detected in a band among a plurality of bands of wavelengths , forming bands of spectral decomposition of the radiation received. This plurality of bands can be large, for example greater than 50, 100, and generally between 50 and 200. Thus, pixels or groups of pixels of the detector are respectively sensitive to light in a given spectral band. It is thus possible to cover and break down a wide useful spectrum, this useful spectrum corresponding of course to the emission spectrum of the polychromatic source 20.

This approach can be implemented in different ways.

According to a first approach, a plurality of photodetectors are placed under the openings of the mask, respectively provided with color filters, for example RGB, or provided with filters interference allowing light to pass only in a determined spectral band. Each measurement point is then characterized by the intensity measurement in each of the spectra defined by these filters.

According to another approach, the photodetector (or photodetectors) under an opening is provided with a single color or interference filter, but the nature of this filter varies from one column to another, along the line. The radiation emitted by adjacent measurement points of the surface S, corresponding to adjacent columns of the detector, is therefore detected by photodetectors respectively dedicated to a color.

According to yet another variant, the image sensor has a plurality of matrices of TDI photodetectors as described above, and each matrix is associated with a single color (associated with a single color or interference filter, to use the terminology of the paragraphs previous ones). Advantageously, the columns of the matrices are aligned with each other, so that the same measurement point of the surface S is successively imaged by the photodetectors of each of the matrices. Each matrix is associated with a confocal mask 6b as described previously. In this way, it is possible to very precisely associate a single measurement point of the surface S with a plurality of color information.

The color information, even coarse, obtained for example using an RGB or interference type filter, can advantageously be used by the electronic control unit 7 to estimate the elevation of the measurement points, and more specifically the distance separating the measurement point from a reference plane of the device. The control unit 7 can therefore estimate an average distance separating the surface S from this reference plane. It can slave the movement along the optical axis AO of the mobile support 4 to this average distance estimate to maintain the inspection surface S in the depth of field of the chromatic system 3 while moving this surface. S along inspection direction I, during a measurement sequence.

The speed of displacement of the surface S, according to the direction of inspection can reach for example a value of the order of 100 mm/second.

An example of application of the inspection device 1 which has just been described will now be described, this example forming a particularly advantageous mode of implementation of the invention.

The object to be inspected in this application is for example a semiconductor wafer, for example based on silicon, or any other material or assembly of materials whose properties can be exploited to form integrated electronic components. The wafer has undergone prior treatment with a conventional manufacturing process in the field of microelectronics and/or microelectromechanical systems, so as to form finished or semi-finished integrated chips on the surface of this wafer. The chips have in particular conductive pads ("conductive pads" according to the Anglo-Saxon expression of the field) intended to be electrically connected to interconnection elements (pins of a package, matrix of balls of an interconnection substrate “interpose”) constituting a first level of interconnection between the chip and the rest of the system in which it is intended to be integrated. To facilitate this first level of assembly, it is usual to have solder bumps (“solder bump” according to English terminology) on the conductive pads, these bumps possibly having a lateral dimension in the plane of the wafer (a width) and a perpendicular dimension (a height) to this plane typically between 15 microns or 20 microns to 250 microns.

The wafer, once provided with these elements, therefore has a surface topology (that is to say a 3D elevation profile) conforming to the arrangement of the weld bumps on the face. main part of the wafer, on the conductive pads of the chips of this wafer.

The placement of solder bumps on the chips of the wafer is an operation that must be controlled, because a bad arrangement of these bumps (an incorrect positioning, the absence or an insufficient or excessive height of a bump or a plurality of bumps) can lead to forming a non-functional component. It is noted that this inspection can implement a 2-dimensional reading, in the main plane of the wafer, of the surface state S of this wafer, for example to identify a bad positioning of a bump, or the presence of a parasitic deposit or particles on the surface of the wafer. The inspection can also record height information, for example during a 3-dimensional survey, to identify that the respective maximum elevations of the bumps, or at least the respective maximum elevations of the bumps of the same chip, resides well in the same plane or has a variation of elevation controlled around the same plane.

This mode of implementation therefore proposes to use the inspection device 1 which has just been described in order to inspect the surface S of the wafer and to ensure that its topology is indeed in conformity with an elevation profile determined.

Thus, and according to a first step of setting up the inspection method which is the subject of this mode of implementation, the wafer is placed on the mobile support 4 of the inspection device 1. The surface S of this wafer, the one having the topology to be inspected, is of course oriented so as to be able to be exposed to the inspection light beam. The mobile support 4 is also adjusted in position along the direction of the axis AO of the chromatic system 3, so as to position the surface S of the object at least partly in the depth of field. The inspection device 1 is also operated via, for example, the control unit 7, during a second step of activating the polychromatic light source 20, in order to project onto the surface S of the wafers the inspection beam and thus forms the reflected beam.

During a third measurement step of the method, the control unit 7 moves the surface S of the wafer relative to the inspection beam to travel a measurement path, for example by moving the mobile support 4.

The control unit 7 simultaneously synchronizes the image sensor 5 and in particular the accumulation and displacement of the charges, from line to line, of the matrix of photodetectors, these charges being those generated by the projection of the light beam reflected on some of the photodetectors through the masks 6a, 6b.

The image sensor 5 repeatedly prepares the lines of charges extracted from the matrix of photodetectors at the end of the columns, and these measurements are made available to the control unit 7 which acquires them and prepares an image of intensity during a fourth stage called "acquisition".

Thanks to the inspection device according to the invention, the speed of movement of the beam with respect to the surface S can be rapid and reach for example a value of the order of 10 Omm/second or more, while collecting enough charges to form an image representative of the state of the surface S.

At the end of this reading step, there is therefore an image of at least part of the surface S of the wafer, this part corresponding to the portion of the surface S of the wafer illuminated by the inspection beam after it has traveled the measurement path. This image is composed of pixels respectively associated with measurement points of the surface S of the wafer. The value of each pixel reflects the intensity of the radiation reflected by the associated measurement point of the surface S, i.e. the degree of presence of this measurement point in the very wide depth of field of the chromatic system 3 (which can for example extend over a distance of 100 to 200 microns, as already mentioned). Assuming that the mean plane of elevation of the surface S is placed, during the placement step, at the level of one of the focal planes farthest from the depth of field, the image makes it possible to locate the arrangement on the surface of the wafer of the salient patterns, and in particular of the weld bumps. The analysis of the image makes it possible in particular to determine whether this arrangement of protruding elements conforms to the expected arrangement, or if elements are missing (unformed or poorly formed solder bump) or if elements are too many (particles , parasitic deposit, etc.).

In the application taken as an example, the solder bumps forming the projecting elements which it is sought to identify are arranged on the surface of the wafer in an ordered manner, for example in lines. In such a configuration, it is advantageous to prepare the measurement trajectory so that it is orthogonal or parallel to these lines.

Provision can also be made to illuminate a portion of the wafer a plurality of times using the inspection beam, for example during a plurality of passages forming a measurement trajectory or a plurality of such trajectories, each passage being made with a different altitude. It is possible to compensate in this way for any variation in spectral reflectivity of the surface S at a measurement point, by placing this point at a focal plane different from the depth of field of the chromatic system 3, and therefore in the illuminating with light rays having different wavelengths.

For example, the different passages can be offset by a distance of a few microns to a few tens of microns. As indicated in a previous passage of this description, the polychromatic source 20 can be chosen very freely, in particular to emit white light or to present a line spectrum.

According to a very advantageous variant of the method, the source 20 is chosen to emit according to a line spectrum, and the number of lines of this spectrum is limited for example to 2, 35 or 10 wavelengths. The depth of field of the chromatic system 3 is then defined by a restricted number (2, 35 or 10) of focusing planes. When a salient element of the wafer (a weld bump, a particle, a parasitic deposit) intercepts one of these planes, it is identified on the measurement image.

By way of illustration of this variant, the polychromatic source 20 provides radiation composed of two wavelengths. The first wavelength is associated with a base plane of focus in the depth of field, and this base plane is intended to be aligned with the main plane of the wafer (i.e. the plane on which reside the salient elements) . This alignment can be carried out during the first stage of setting up the method.

The second wavelength is associated with a second focal plane distant from the first plane by the expected height of the salient elements which it is sought to locate, or by a little less than this expected distance. By carrying out the inspection process using equipment in accordance with the description and equipped with such a source, an image is produced of at least part of the surface S of the wafer, this image having peaks intensity which correspond to salient elements having at least one height corresponding to the expected height. It is possible to identify on this image an absence of a protruding element or a protruding element presenting an insufficient height (for example a weld bump) if the zone of the image where this element should have been located does not present any light intensity.

The image produced using the method which has just been described is essentially a 2D image of the surface of the wafer which contains little or no information on the elevation profile of the surface S inspected. This image is nevertheless very useful for identifying very quickly and over an extended surface of the wafer, certain types of defects.

To further improve the inspection capacity, a method in accordance with the invention can also take advantage of the color or spectral band distribution information which can be supplied by the timed integration image sensor 5 of a inspection 1 in accordance with this description. As mentioned in a previous passage, the determination of the wavelength of the radiation detected by the photodetectors of the image sensor 5 makes it possible to determine the depth of a measurement point in the depth of field of the chromatic system. 3, ie the height of this measurement point relative to the main plane of the wafer.

The color information (RGB or more spectrally detailed) returned by the image sensor 5, in addition to the information on the intensity of the radiation detected, can be converted, for example by the control device 7, into information height. The image provided in this case, which can also materialize in the form of a point cloud, is a 3D image of the surface of the wafer and represents the topology of this wafer, i.e. its elevation profile over its entire length. extent .

This image can be used very directly to determine the maximum height of weld bumps, and to check whether this height is compliant.

The color information supplied by the image sensor 5 can be exploited in all the variations of the method. inspection which have just been described, in particular that according to which the polychromatic source 20 is chosen to emit according to a spectrum having a limited number of lines.

Of course, the invention is not limited to the embodiments described and variant embodiments can be added thereto without departing from the scope of the invention as defined by the claims.

Thus, the two opaque lighting and detection masks 6a, 6b can be replaced by a single mask, arranged in this case under the separating element 3b. The mask can for example be fixed directly under a separator cube or a reflecting plate, for example by depositing an opaque material on this optical part. In this alternative configuration, the openings of the single mask perform both the function of the lighting and chromatic filtering openings. The openings of this mask are then optically conjugated from the surface S of the object by the chromatic system 3, and optically conjugated from the image sensor 5.

Claims

CLAIMS Method for inspecting a surface (S) of an object, the method comprising:

- a first positioning step during which the object is placed on a mobile support (4) of an inspection device (1) so as to position the surface (S) in a depth of field of a chromatic system (3) of this device (1), the chromatic system (3) spatially spreading the focusing of a polychromatic inspection beam according to focusing planes in said depth of field arranged along an optical axis ( AO), and intercepting a beam reflected by the surface (S) to project it onto a matrix of photodetectors of a time-delayed integration image sensor (5) arranged in a detection plane (P), conjugate of the focusing planes ;

- a second step of activating at least one polychromatic light source (20) to project the inspection beam onto the surface (S) and thus form the reflected beam, at least one confocal mask (6a, 6b) having a plurality of chromatic filtering openings intercepting the inspection beam and the reflected beam, the chromatic filtering openings of the confocal mask (6a, 6b) making it possible to illuminate part of the photodetectors;

- a third survey step during which the mobile support (4) is controlled to move the surface (S) relative to the chromatic system (3) at least along an inspection direction (I) perpendicular to the optical axis ( AO) and traverse a measurement trajectory, and during which the image sensor (5) is controlled to synchronize it with the movement of the surface (S);

- a fourth step of acquiring the data established by the image sensor (5) and of preparing an image formed of pixels respectively associated with measurement points of the surface (S) of the object, each pixel reflecting the intensity of the radiation reflected by the measuring point.

28 Inspection method according to the preceding claim, in which the object is a chip wafer, the wafer being provided with a plurality of solder bumps. Inspection method according to the preceding claim, in which the solder bumps are arranged in a line, and the trajectory is at least partly parallel or perpendicular to the lines of solder bumps. Inspection method according to one of the preceding claims, in which the trajectory is formed of a plurality of passages having different altitudes. Inspection method according to one of the preceding claims, in which the third and/or the fourth step are implemented by a control unit (7) of an inspection device (1). Inspection method according to one of the preceding claims comprising a step of analyzing the image to locate the arrangement of protruding elements of the surface (S). Inspection method according to the preceding claim, in which the analysis step results in determining whether the arrangement of projecting elements conforms to an expected arrangement. Inspection method according to one of the preceding claims, in which the image sensor is a color or multispectral image sensor supplying color or spectral information, and the method comprises a step of converting the color information or spectrum provided in distance information to form a 3D image of the surface (S) of the object. Inspection method according to one of the preceding claims, in which the matrix of photodetectors is disposed in the detection plane (P) and arranged in a plurality of rows and columns, the charges generated by the photodetectors being able to be moved from one row to another, in a column direction synchronized with the movement in the direction inspection (I) . Inspection method according to one of the preceding claims, in which the chromatic filtering openings are arranged on the confocal mask (6a, 6b) to illuminate a plurality of photodetectors arranged on a column of the matrix corresponding to a plurality of measurement points arranged along the direction of inspection (I). Inspection method according to one of the preceding claims, in which the chromatic filtering apertures are distributed in a complementary manner over at least two lines of the confocal mask (6a, 6b) to illuminate a plurality of photodetectors corresponding to a plurality of measurement points arranged along at least two lines arranged along a direction perpendicular to the direction of inspection (I). Inspection method according to the preceding claim, in which the confocal mask (6a, 6b) comprises a plurality of groups of at least two lines comprising chromatic filtering openings distributed in a complementary manner over the said at least two lines. Inspection method according to one of the preceding claims comprising a confocal lighting mask (6a) placed in the inspection beam and a confocal detection mask (6b) placed in the reflected beam, the confocal lighting mask ( 6a) and the confocal detection mask (6b) being respectively placed in a conjugate plane of the focusing planes with respect to the chromatic system (3). Inspection method according to one of Claims 1 to 12, comprising a single confocal mask, arranged in a plane jointly conjugate of the detection plane (P) and of the focusing planes vis-à-vis the chromatic system (3).

15. Inspection method according to one of the preceding claims, in which the polychromatic light source

(20) comprises a halogen lamp or a strip of light-emitting diodes, associated with a diffusion device

(21) to homogenize the intensity of the inspection beam.

16. Inspection method according to one of the preceding claims, in which the polychromatic light source (20) exhibits a line spectrum, exhibiting a determined number of emission wavelengths or bands of wavelengths.

17. Inspection method according to one of the preceding claims, in which the chromatic system (3) comprises a chromatic lens (3a).

18. Inspection method according to one of the preceding claims wherein the color system (3) comprises a separator element (3b).

19. Inspection method according to one of the preceding claims, in which the photodetectors are equipped with a color filter or a spectral filter.

20. Inspection method according to the preceding claim, in which the chromatic filtering openings are arranged to each illuminate a plurality of photodetectors provided with filters of different colors or spectral bands.