TW201440933A - Method of performing beam characterization in a laser scribing apparatus, and laser scribing apparatus capable of performing such a method - Google Patents

Abstract

A method of performing beam characterization in a laser scribing apparatus that is used to scribe a substantially planar semiconductor substrate, which apparatus comprises: A movable substrate holder on which said substrate can be clamped, whereby the substrate is clamped in a clamping zone of the holder so as to present a target surface to be scribed along at least one scribelane; A laser source that can be used to direct at least one laser beam along an optical axis toward the target surface, which method comprises the following steps: Providing an image recording device that can be used to view and control the scribing process, said device having a field of view with a viewing axis normal thereto; Providing a reference plate on the substrate holder outside the clamping zone, the reference plate comprising a material that produces stimulated radiation in response to irradiation by said laser beam; Positioning the substrate holder such that the laser beam impinges upon a point of the reference plate and excites said material to produce a localized glow at said point of impingement; Using the image recording device to detect a physical parameter of said glow. Examples of such a physical parameter include: A location of said glow relative to a reference point in said field of view; A diameter of said glow; An integrated intensity of said glow; A shape of said glow, and combinations hereof.

Description

Method for performing beam characterization in a laser scribe device, and laser scribe device capable of performing the same

The present invention relates to a method of performing beam characterization in a laser scribe device for scribing a substantially planar semiconductor substrate, the device comprising: - a movable substrate holder, the substrate being closable Removing the substrate holder, whereby the substrate is clamped in a clamping region of the holder to present a target surface to be scribed along at least one scribe lane; a laser source that can be used along An optical axis directs at least one laser beam toward the target surface.

The invention also relates to a laser scribing apparatus in which such a method can be performed.

For the purposes of clarity and consistency, the following terms used throughout this text and the scope of the accompanying claims are to be accorded to the following description: - the word "beam characterization" refers to the determination of the accuracy of the laser scribe method, Quality, reliability, and/or reproducibility have a program that affects one or more beam characteristics. The wording encompasses features such as beam alignment and focus, as well as parameters such as size, shape, and (integral) intensity of the beam profile (on the substrate surface). These points are detailed below.

- The wording "substantially flat" should be interpreted to mean a substrate in the form of (approximately) a sheet, plate, blade, sheet, plate or the like. This substrate will form a large system (substance Upper) is flat and presents two opposing major surfaces separated by a relatively thin intervening "sidewall".

- The word "semiconductor substrate" is to be interpreted broadly to encompass any substrate on which a semiconductor device or other integrated device is fabricated. Such substrates may, for example, comprise wafers of (of various diameters) germanium or germanium wafers and/or compound species such as InAs, InSb, InP, GaSb, GaP or GaAs. The term also encompasses non-semiconductor materials (such as sapphire) on which one or more layers of semiconductor material have been deposited (eg, as in the manufacture of LEDs). Semiconductor devices or other integrated devices of interest may, for example, be an integrated circuit, (passive) electronic components, optoelectronic components, bio-wafers, MEMS devices, and the like. Such devices will typically be manufactured in large quantities on a given substrate and will typically be arranged in a matrix configuration on at least one of the major surfaces.

- the term "scratching" (also sometimes referred to as "cutting a deep etch") is understood to mean a path (track, deep eclipse, strip) extending along one of the major surfaces of a substrate. The substrate is scribed along the path. The term "scratching" as used in this context covers processes such as scoring, cutting, cutting, cutting, cutting, denting and planing; in particular, it covers so-called "groove" It involves radiation (eg, ablation) removal of one of the substrate materials (eg, including so-called "low-k" materials) from a wide strip to a depth less than one of the substrate thicknesses. In the particular case of a semiconductor substrate, a scribe lane typically extends between adjacent columns of the integrated device of the substrate and defines a path along which the substrate will be "cut" to allow for the device in question. (final) separation. This procedure is often referred to as "single granulation." It should be expressly noted that this singulation procedure can be single-step (where the substrate is cut/cut through the full depth of the substrate in a single operation) or multi-step (where a first scribe does not pass through) The substrate is cut to the full depth of the substrate and one or more subsequent procedures are used to perform the cutting process, such as additional radiation scoring, mechanical scoring/cutting, and the like. It should also be noted that the scribe lanes on the target surface can be configured in a regular and/or irregular (repetitive) configuration. For example, some wafers may include the same product separated from one another by forming a regular orthogonal grid. A rule matrix for one of the body devices. Alternatively, other wafers may include devices having different sizes and/or positioned at irregular intervals relative to one another, implying a corresponding irregular scribe configuration. The arrangement of such scribes is not necessarily orthogonal, for example, it may be (partially) triangular/hexagonal.

These points are discussed in more detail below.

One of the scoring devices, such as described in detail in the above paragraph, is discussed, for example, in WO 2002/076699 A1. Here, a single laser beam is used to form a etched zone along a scribe line configured as an orthogonal grid structure on one of the major surfaces of a coated semiconductor wafer. Once the grid has been formed, a mechanical saw is used to singulate the substrate into grains by cutting along the zones.

Another scribing device of the type set forth in the opening paragraph is discussed, for example, in U.S. Patent No. 5,922,224. In this document, a linear cluster of laser beams is used to ablate the substrate material along a scribe lane, thus causing the substrate geese ablation line to be "radiated". Using multiple beams instead of a single (stronger) beam in this manner can help create a narrower ablation zone on the substrate. This may have certain advantages, for example, especially when the scribing track in question is in close proximity to a (fragile and expensive) device on a semiconductor substrate.

The laser beam used in the current context (one or more) can be focused onto the target surface of the substrate, or can be focused below the surface as desired (so-called "invisible cutting"). This primary surface method can be used (for example In other words, it helps to reduce the contamination problem by causing the ablation debris to be confined within the body/lattice of the substrate material rather than falling onto the surface of the substrate. Even if a laser beam has a focus below a surface, it will exhibit a spot of light at the point where it illuminates the surface.

In a laser scribing procedure, such as the laser scribing procedure mentioned in the open paragraph, the laser beam(s) used to scribe the semiconductor substrate have a scribe relative to the scribe One of the well-defined locations of the road is important. As discussed above, a scribe will generally be relatively narrow and will typically be widened by relatively expensive and delicate integrated devices (in one Or a plurality of sides; if a scoring laser beam does not have a carefully controlled position within a given scribe lane, it may result in damage to the adjacent integrated device and/or an unacceptable Scratch the form and/or produce an incorrect single granulated grain size. Despite all the care that should be applied when designing, constructing, and using a laser scribe device, drift effects (eg, mechanical vibrations, thermal phenomena, etc.) can cause the position/angle of the optical axis of the laser source to be unpredictable Mode (gradual) shifting, if it is not corrected, may result in unacceptable scoring results. It is therefore important to have a method of mitigating these drift effects.

Similarly, the focal position of the laser beam relative to the target surface of the substrate is important because it will determine the size and intensity of the spot, which in turn will affect the laser beam. The width and depth of the scribe and the size of the so-called heat affected zone (HAZ) outside the deep etched road (and the temperature therein). As mentioned above (and below), there are different methods of radiation scribing, some using a laser beam focused onto the target surface and other methods using a laser beam deliberately focused below the target surface; for selection and assurance The reliable performance of these different technologies is therefore important to have accurate information about the focus state/axial alignment of the laser beam. Again, the drift effect can cause unfavorable shifts in this regard.

Additionally, drift effects, such as those mentioned above, may, for example, result in changes in optical distortion or other aberrations and/or relative position along optical components (such as diffractive optical elements, lenses, etc.) / One of the orientation shifts has an adverse effect on the optical components used to focus/point (one or more) of the laser beams employed. This in turn can adversely affect the generation of the target surface of the substrate, for example, due to undesired elongation/convexity of the spot, appearance of the satellite spot, etc., and/or undesired changes in the intensity distribution within the spot. The shape of one of the laser spots.

One method of performing beam characterization is to perform regular manual inspection/correction of laser beam alignment, focus, and/or other beam characteristics in the scoring apparatus. This will typically involve the intervention of the user of the scoring device, which requires him to perform laser beam alignment, spot size, etc. Visual inspection. However, this manual method has a similar disadvantage. For example: - because it involves human judgment, it will be subjective in nature and subject to a relatively large error; - it will require immediate attention from one person, in the field or via a remote link, thus requiring personnel Level and cost; - it inevitably involves the use of a sacrificial consumable (such as an open wafer) on which one or more test marks and/or lines will be scribed/burned as part of a manual inspection process . This translates into additional costs and the need to keep a stock of one of the consumables on hand.

The need to use a consumable can also result in an extra time loss. For example, if someone wants to temporarily interrupt the scrimming of a batch of substrates in order to perform a "in-process" beam inspection, then: - the normal workflow of the substrate will have to be interrupted to allow manual loading of an empty consumable And then clamped to the substrate holder; - after the inspection has been performed, the empty consumables will have to be loosened and manually removed before the normal workflow of the substrate can be resumed.

This procedure is time consuming and cumbersome.

One of the objects of the present invention is to solve such problems. More specifically, it is an object of the present invention to provide a method for performing beam characterization in a laser scribe device that allows for more efficient and objective inspection and correction of alignment such as a scribed laser beam Beam parameters such as size, strength and shape. In particular, it is an object of the present invention to provide a method that provides the risk of causing a semiconductor substrate to be reduced in one of the undesired locations/in an undesired manner.

Achieving these and other objects in one of the methods as detailed in the open paragraph, the method is characterized by the steps of: providing an image recording device that can be used to view and control a scribing process, the device The member has a field of view having a normal viewing axis; a reference plate is provided on the substrate holder outside the clamping region, the reference plate comprising a response in response to the irradiation of the laser beam One of the stimulated radiation; positioning the substrate holder such that the laser beam illuminates a point on the reference plate and excites the material to produce a localized (regional) glow at the illumination point; - using the image The device is recorded to detect one of the physical parameters of the glow.

Examples of such a physical parameter include, for example: - a position of the glow relative to one of the reference points in the field of view; - one of the diameters of the glow; - one of the intensity of the glow; - one of the glows Shape, and combinations thereof.

As used herein, a reference plate containing "a material that produces one of the stimulated radiation in response to the irradiation of the laser beam" is understood to mean the mechanism by which the material produces light (the "localized glow") rather than By a simple reflection of a portion of the incoming laser light. The intent here is the physical process from the "raw" photon stimulating material of the laser beam that is illuminated by the material, which causes it to emit "secondary" (excited) photons; in general, this can This is called bandgap excitation/absorption, followed by radiation recombination. In this regard, examples of such mechanisms include the following: - the so-called "wavelength conversion." Here, incoming (primary) photons from the laser beam produce electron-hole pairs in the material. When these electron-hole pairs are subsequently recombined, there is one release of energy in the form of a localized glow (luminescence) of the secondary (excited) photons. Examples of wavelength converting materials include, for example, SiC (cerium carbide) and GaN (gallium nitride).

- The so-called "two-photon absorption". Here, one of the molecules of the material simultaneously absorbs two original photons from the irradiated laser beam. When the molecule is de-excited, it emits a secondary (excited) photon that is one of the wavelengths of the wavelength of the original photon. Examples of two-photon absorbing materials are, for example, GaAs (gallium arsenide) or ZnTe (zinc telluride) (which can be used with infrared laser wavelengths). For more information, see: http://en.wikipedia .org/wiki/Two-photon_absorption

- Incandescent light. Here, the incoming laser beam thermally excites the atoms of the material, causing it to produce a hot glow. For example, see: http://en.wikipedia.org/wiki/Incandescence

- Luminescence (for example, this general term covers more specific phenomena such as photoluminescence, fluorescence, and phosphorescence). Here, the mechanism by which photons are emitted is essentially non-thermal. For example, see: http://en.wikipedia.org/wiki/Luminescence http://en.wikipedia.org/wiki/Photoluminescence http://en.wikipedia.org/wiki/Fluorescence http://en .wikipedia.org/wiki/Phosphorescence

By their very nature, these mechanisms will generally tend to produce a glow that is multi-colored.

This method according to the invention is highly advantageous for several reasons. For example: - It does not use a consumable part (which must be loaded and unloaded each time) as part of the beam characterization program. Rather, it uses a dedicated reference plate that is permanently present on the substrate holder, for example, in/near a corner of its upper surface. Thus, performing a beam characterization in accordance with the present invention involves only temporarily moving the substrate holder to position the reference plate in the path of the laser beam without the need for (loading/loading) substrate/consumables (time consuming and cumbersome). .

- It does not require a permanent scribing/burning mark as part of the beam characterization process. Rather, the beam parameters are evaluated by examining one of the instantaneous glows at the illumination point of the laser on the reference plate. This glow can be produced by a much more intense laser power/flux than the laser power/flux required to scribe/burn a sacrificial wafer, so there is no need to have a reference Any substantial radiation damage to the board. Additionally, the reference plate can comprise a material (such as SiC) that is much rougher than a typical tantalum wafer.

- When a laser beam is directed at a generally solid surface, it produces a spot of scattered light. This spot typically exhibits a phenomenon known as "spot", which is caused by the coherence effect in the laser beam, thus causing the laser spot to have a mottled "shape shift" appearance. This speckled/granular appearance often confuses the human eye and the optical detector, making it difficult to accurately determine the position/size of the spotted spot, and thus, for example, hindering the beam that may use this spot Align efforts. In contrast, in the present invention, the material of the reference plate produces a localized glow of one of the stimulated radiation. This glow is irrelevant (this is due to the non-emission of lasers generated in the reference plate) and therefore does not suffer from speckle.

- The method of the invention facilitates complete automation since it uses an image recording device such as a camera and because it mathematically determines quantifiable variables (position, diameter, integral intensity, etc.). When used in conjunction with, for example, an image recognition software, the method of the present invention can perform beam characterization in a fully automated/autonomous manner. The problems mentioned above (subjectivity, human existence) associated with artificial characterization are eliminated.

Advantageously, the reference plate will have a substantially flat upper surface with a plane substantially parallel to the plane of one of the clamping zones through which the substrate is clamped; however, this is not a critical requirement (see below).

In an embodiment of the method according to the invention, data from the image recording device is used to calculate a positional deviation between the optical axis and the viewing axis (e.g., based on the location of the intersection of each axis with a predetermined reference plane). More specifically, the calculation is performed using the detected position of the glow relative to the reference point mentioned above in the field of view, which will result in any amount of planar beam misalignment (eg, in the reference plane) Value / direction. The selected reference point is a selection problem: it can be, for example, a geometric center point (abstract or real) or one of the "crosshair" marks in the field of view.

Once the optical axis of the laser source and the image recording device have been calculated using the method of the present invention Looking at a positional deviation between the axes, it can be used in different ways. For example, in one method, someone may attempt to adjust, for example, by using an (automated) correcting member such as one or more optical deflectors or one or more mechanical actuators or a combination of the two The path of the laser beam removes (or at least reduces) the measured deviation and/or moves the laser source itself back (or toward) the optimal alignment with the viewing axis. In another method, a person accepts the measured positional deviation as it is, but seeks to make a corrective adjustment when performing a subsequent scoring procedure, for example, by appropriately adjusting the position of the substrate holder. In the context of this latter method, an embodiment of the method according to the invention has the additional steps of: providing a clamped substrate on the substrate holder; positioning the substrate holder to position one of the substrates Within the field of view; using an image recording device to detect a feature of the scribe track (such as, for example, an edge, a center axis, a corner, or a mark) relative to the reference point (in the field of view) a position; - applying the positional deviation to determine that the optical axis infers a position relative to one of the features.

In other words, this embodiment carries out the following: - at the reference plate: measuring the position P _{1 of} the glow G with respect to one of the reference points R mentioned above (vector); - at the clamped substrate: amount The position of the scribe track feature F relative to the reference point R (vector) P _{2 is} measured; at the clamped substrate: the position of the laser optical axis O relative to F is inferred to be the phylogenetic difference P ₁ - P ₂ .

This information can then be used to make a vector correction to the position of the substrate holder, namely -( P ₁ - P ₂ ), in order to accurately position O on F.

In another embodiment of the method according to the invention, data from the image recording device is used to calculate an axial distance (along O) between the illumination point and one of the positions of the best focus of the laser beam. More specifically, the calculation is performed using the diameter and/or integrated intensity detected by one of the glows, which will produce a magnitude of any axial beam misalignment. This embodiment takes advantage of the fact that one considers a section of a laser beam at different points along the optical axis, and the profile will (notoriously) exhibit the smallest diameter (or "waist size" in the focal plane of the beam; D _min ) and maximum integrated intensity (I _max ). Such "best focus" values can be determined, for example, from a single calibration run or an in-line measurement, thereby stepping the reference plate over a range of different axial positions along the beam, and the glow Diameter (D) and/or integral intensity (I) represents a function of one of these axial positions, allowing for measurement/interpolation (by traversing the best focus) or estimation (by approaching the best focus) D _Min /I _max . Once the "best focus" value is known, the best focus of the illumination point and the laser beam can be calculated (for example, by automatically looking up a lookup table) using the instantaneous values of D and/or I measured at the reference plate. The distance of the position, or equivalently, the distance of the top surface of the reference plate relative to the focal plane of the laser. The "stepping" referred to herein may be by adjusting the axial position of the substrate holder relative to the laser source and/or by using an adjustable optical element to position the laser relative to one of the substrate holders. Shifting is achieved; either or both of these mechanisms can be used to perform subsequent corrections to one of the calculated axial beam misalignments (focal displacements).

If the topmost (irradiated) surface of the reference plate has exactly the same height as the clamped substrate when viewed along the viewing axis (eg, because it is substantially coplanar), the reference plate can be used directly The positional deviation (planar and/or axial) calculated by the location is used to perform a calibrated position adjustment at the clamped substrate. However, if there is a small difference in height (eg, due to a relative step or tilt between the substrate and the reference plate), then it can be determined that when the position output measured at the reference plate is translated to be at the clamped substrate A small additional correction is necessary when using the position input. This is due to: - For a given angular deviation between the optical axis and the viewing axis, a small difference in one of the (vertical) heights will be translated as a small (horizontal/planar) positional shift perpendicular to the viewing axis; A height difference mentioned in the table plays a direct role in the (vertical/axial) positioning of the laser focus relative to the target surface of the substrate.

In this context, a particular embodiment of the method according to the invention comprises the following steps Step: - determining a height difference between the reference plate and the clamped substrate at a plane of the substrate; - using at least one of the positional deviation and the axial distance to adjust the beam alignment.

The height difference referred to herein may be, for example, calculated (based on the known thickness of the reference plate and the clamped substrate and its positioning on the substrate holder) and/or measured (in an accidental calibration example) Cheng Zhong) to judge. This measurement can be performed either automatically or manually. As an example of the former, for example, a (commercially available) autofocus sensor can be provided for an image recording device, which automatically indicates (focal length) at the optimal focus setting of the reference plate and the substrate. The detector outputs a reading and then subtracts these readings. Alternatively or in addition, an optical sensor, such as an interferometer, can be used to perform the desired measurements.

In addition to measuring beam parameters such as lateral/planar beam position, axial beam position (focal position), beam diameter and/or integrated beam intensity (in a given plane), the invention may also be used (eg, by use) The image recognition software as mentioned above determines whether, for example, one of the laser spots (glow) formed on the reference plate is circular or centrifugal, regular or irregular, single or satellite light The shape of the beam is checked, judged, and accepted/rejected (profiled) at points (eg, undesired higher diffraction orders and/or parasitic reflections). In the case where a measured spot shape is judged (by software) to be outside the acceptable boundary, a predefined action may be taken, for example, a warning signal may be given via a user interface (such as a display) and The scribe procedure is paused until a device user intervenes.

Until now, the discussion of the present invention has generally referred to "a" laser beam. However, it is important to recognize that the laser source employed can actually generate a plurality of laser beams, for example, because it uses a beam splitting configuration, such as one or more diffractive optical elements (DOE), to subdivide A single laser beam of a single laser, or because it uses several lasers. In this context, a particular embodiment of the method according to the invention has the following A: the laser source produces a cluster of laser beams; the at least one laser beam is a member of the cluster and serves as a basis for performing beam alignment of the entire cluster.

Here, one (at least) one particular laser beam is selected as the basis for performing beam alignment of the entire cluster of one of the associated laser beams. The particular laser beam selected for this purpose may be a matter of choice: for example, one of the beams at one (or near) the center of mass, or one of the beams at or near one end of the cluster, may be selected. . An example of such a cluster of laser beams is provided, for example, in the prior art document US 5,922,224 mentioned above, but many other possible cluster configurations are contemplated. In all cases involving a cluster of laser beams, those skilled in the art will be well able to select an appropriate beam (or group of beams) to use as a basis for performing the beam alignment method of the present invention.

It should be noted that there may be instances in which an image recording device of the present invention can be advantageously provided for use with a particular type of optical filter. For example, a glow that is generated (eg, by wavelength conversion) in a reference plate of the present invention is typically multi-colored, and a wavelength filter can be used to select one (quasi) monochromatic portion of the glow, thereby Relatively inexpensive imaging optics with relatively simple chromaticity correction are allowed. Alternatively or additionally, a filter may be optionally used to reduce the intensity of the glow for adjustment according to the sensitivity of the image recording device employed, but alternatively only to reduce the laser from the selection The intensity of the source's radiation output. Those skilled in the art will understand these points and can implement these points as needed for a particular setting (or choose to ignore such points).

Regarding the image recognition software that can be used to perform the present invention, those skilled in the art will be familiar with the software and its use. The image recognition algorithm can operate, for example, by seeking cross-correlation with a given artificial "stylized" reference image or by performing a template match with a previously selected template. There are various commercially available software packages for this purpose.

It should be noted that the image recording device used in the present invention may be, for example, a photograph. A camera or still camera, and ideally directly produces a digital image that facilitates processing with the image recognition software; another option, a digital analog image can also be contemplated.

While the present invention yields many advantages over the prior art, one advantage is particularly worth mentioning here. Better than the present invention allows for more accurate characterization/measurement and correction of parameters such as beam position, beam size and beam shape, thus allowing one of the lateral positional accuracy of a scribe/groove to be formed in a substrate Significantly improved. It is of particular importance in a "multi-pass" scribing process in which successive scores/grooves of depth ΔZ _n overlap in order to produce one of the cumulative depths Σ ΔZ _n The resulting score/groove. In this scenario, the present invention allows for greatly increased overlap accuracy, resulting in a more satisfactory end result.

It is worth mentioning that although the "optical head" (ie, the focusing optics (beam splitter or DOE, etc. for the laser beam(s)) is fixed in the scribing of the semiconductor substrate and is moved as needed Substrate holders are conventional, but it is conceivable to reverse these moving characters, i.e., to keep the substrate fixed and to move the optical head (e.g., by attaching the optical head to one of the fixed lasers). This mechanical reversal does not have any material impact on the key to the present invention and can be considered to fall within the scope of the accompanying claims.

In an exemplary method according to the invention, which is not intended to limit the scope of the invention, but is presented here for the purpose of providing a specific practical example, the following aspects apply: - an input laser The beam is selected to have one of the wavelengths in the range of 200 nm to 3000 nm and one of the output in the range of 1 mW to 100 W. The laser selected will greatly depend on the material of the substrate being scribed. The wavelengths within this range can be generated by a variety of lasers. For example, a solid state Nd:YAG laser produces one wavelength of 1064 nm, with the harmonics being 532 nm and 355 nm. Alternatively, a doped fiber laser having one of the wavelengths of 1062 nm can be used, for example. The 355 nm wavelength is particularly attractive because it is often strongly absorbed by semiconductor materials; It is usually relatively easy to focus on a relatively small spot size.

However, this is entirely a matter of choice, and other wavelengths may alternatively be employed.

- Using a laser source capable of delivering a pulsed laser beam, wherein one pulse duration is in the range of about 1 microsecond to 100 femtoseconds. In this regard, it should be noted that when performing actual laser scoring, the laser is used in a pulsed mode; when performing beam alignment in accordance with the present invention, the beam is not necessarily pulsed and can be used in continuous wave mode as needed. .

- Using a diffractive optical element (DOE) (or, for example, a cluster of polarizing beam splitters), the input laser is divided to form a beam cluster having one of 2 to 20 members. This step is optional and can instead be selected using a single/divided beam. When the input laser is selected to be split, the separation of adjacent focal points may, for example, be selected to be within the range of about 5 [mu]m to 500 [mu]m, for example, about 50 [mu]m to 70 [mu]m. It should be noted that the beam cluster form thus generated does not have to be linear (one-dimensional); instead one of the forms of a rectangle that produces, for example, n spots x m spots ( n and m-type integers) can be selected. Two-dimensional clustering.

- if the laser beam(s) are selected to be focused into the substrate body rather than being focused onto its target surface (invisible cut), the (one or more) focus of the (equal) laser beam is lower than the target The depth of the surface(s) is selected to be in the range of 20 μm to 700 μm.

For example, the reference plate (the top surface) comprises a wavelength converting material (such as GaN or SiC) or a two-photon absorbing material (such as GaAs). For example, it has a lateral dimension ca. 1 x 1 cm ² .

Example 1

1 presents an end view of a portion of a portion of a device A suitable for performing one of the methods of the present invention, the device being operable for at least one of the target surfaces 3 of a substantially planar semiconductor substrate 1 The substrate 1 is scribed 2 (not shown; see Fig. 2). On the other hand, Fig. 2 shows a plan view of one of the lower portions of Fig. 1. Figure 3 shows the same apparatus as in Figure 1, but during the execution of one aspect of one of the embodiments of the method of the present invention. Note the Cartesian coordinate systems X, Y, Z shown in the figures.

In particular, Figures 1, 2 and 3 together show the following: a laser source 4 which outputs at least one (pulsed) laser beam L along an optical axis 7. The laser source 4 is coupled to a controller 14 that can be used to control parameters such as pulse duration and/or power/flux of the beam and others. The laser source 4 can be caused to operate in a continuous wave (CW) mode as needed.

▪ A movable substrate holder (workbench, chuck) 9 having one (center-positioned) clamping area to which the substrate 1 is mounted to present the target surface 3 to the laser source 4. For example, this installation has traditionally occurred via perimeter clamping.

▪ An assembly 17 that positions the substrate holder 9 in the XY plane relative to the optical axis 7.

A projection (i.e. imaging) system 11 for projecting a laser beam L onto a substrate 1. At the point of illumination of the beam L on the substrate 1, a spot S is formed. The projection system 11 can be used to focus the light beam L onto or into the substrate 1 as desired, and can also perform aberration/distortion correction (for example).

Figure 2 shows the substrate 1 viewed from above, resting on the clamping area of the substrate holder 9 (the clamping area is not explicitly shown here, but it will be considered as an area of the holder 9, which is intended (e.g. Holding one of the substrates to be scribed by means of a suitable clamping member in/around the area. On the target surface 3, various scribe lanes 2 are illustrated. These scribe lanes 2 are stretched in an X/Y grid pattern between the integrated devices 23 (which are distributed in a matrix arrangement on the surface 3); there will typically be a very large number of such devices on a typical semiconductor substrate 1. 23, but only a few are illustrated here so as not to confuse the schema. The figure illustrates a method of "longitudinal scanning and lateral stepping" of one of the substrates 1 along a plurality of consecutive scribe lanes 2 along a particular direction (in this case, ±X). For example: ▪ The substrate 1 is scribed along the scribe lane 2a by scanning the laser beam L in the -X direction; in fact, this relative motion can actually be achieved by using the stage assembly 17 (see Figure 1). It is achieved to scan the substrate holder 9 in the +X direction.

▪ After completing the scribe along the scribe lane 2a, the stage holder 17 will be used to step the substrate holder 9 in the +Y direction by an amount ΔY; therefore, the laser beam L will actually be relative to the target surface 3 step by one amount - △ Y.

▪ The substrate 1 is now scribed along the scribe lane 2b by scanning the laser beam L in the +X direction; in fact, this relative motion can be achieved by scanning the substrate holder 9 in the -X direction using the stage assembly 17. .

▪ Wait.

It should be noted that there are various ways of embodying the station assembly 17, and those skilled in the art will Many alternatives can be implemented in this regard. One particular embodiment, schematically illustrated in Figure 2, uses two separate linear motors (not shown) to independently drive the substrate holder 9 along axes A1 and A2, with axes A1 and A2 being aligned with the X and Y axes. 45° angle; movement along X or Y involves simultaneous driving along the A1 and A2 axes. Typically, the substrate holder 9 will be smoothly floated over a reference surface (such as a polished stone surface) in the XY plane, for example, by means of an air bearing or magnetic bearing (not shown). For example, the exact position of the substrate holder 9 can be monitored and controlled by means of a positioning instrument (not shown) such as an interferometer or linear encoder. In addition, focus control/level sensing (not shown) will also typically be employed to ensure that the target surface 3 of the substrate 1 is maintained at a desired level relative to one of the projection systems 11. All such conventional positioning and control aspects will be familiar to those skilled in the art and do not require any further explanation here.

Those skilled in the art will also appreciate that, conventionally, one of the substrates 1 to be scribed will be first mounted over a foil that spans a circumferential frame and will therefore have to be mounted to the substrate holder 9 On the area is a composite structure of a substrate, a foil and a circumferential frame. Similarly, those skilled in the art will appreciate that after scribing an entire substrate 1, the substrate can be separated along various scribe lanes by laterally stretching the foil. These are the nature of the field of semiconductor substrate scribing, and therefore need not be further explained here; for further information, reference is made to the following publications (for example): US 2008/0196229 A1 and US 5,979,728.

▫ http://en.wikipedia.org/wiki/Dicing_tape

▫ http://www.lintec-usa.com/di_t.cfm#anc01 .

Returning now to Figure 1, device A further includes: an image recording device 8 (e.g., a digital camera) that can be used to view a portion of the target surface 3 near the spot S, the size of the portion being viewed by device 8 The size of the domain is determined. Vertically to the field of view, device 8 has an viewing axis (normal line of sight) 7" and is coupled to a controller 12 that can, for example, be used to perform image recognition and various calculations in the context of the present invention. (as follows).

A light 16 that produces light in response to commands from the controller 12 and which assists the image recording device 8 in viewing features of interest on the target surface 3, such as the scribe track 2. Lamp 16 can produce a (strobe) flash or continuous illumination as desired for a given situation. For example, it can include a light emitting diode (LED).

▪ A beam splitter 6 (e.g., a dichroic mirror) that can be used to form a so-called "on-axis" configuration whereby the optical axis 7 of the laser source 4 and the viewing axis 7" of the image recording device 8 follow A portion of the common axis 7' is nominally coaxial with respect to one of the substrate holders 9. A portion of the light emitted from the view portion of the target surface 3 will travel along the axis 7', striking the beam splitter 6 and along the axis 7" It is reflected to the image recording device 8. An alternative to this on-axis configuration (using a so-called "off-axis" setting) will be discussed below with respect to Figure 5.

▪ A filter 10, such as a suitable interference filter. For example, the filter 10 can be used to substantially reduce the wavelength of the laser light employed from the laser source 4, but substantially transmissive (at least some) of the light produced by the lamp 16. In this manner, the digital camera 8 can capture an image in which the position of the spot S is clearly visible against the target surface 3 without being oversaturated by excessive intensity from the spot S.

Controller 12 is coupled to laser controller 14 as shown herein. The controller 12 is also coupled to the stage assembly 17 such that the relative positioning of the substrate 1 and the shaft 7' can be adjusted in response to automated analysis performed by the controller 12 in accordance with the image captured by the image recording device 8. This aspect of the invention will be explained in more detail below.

As discussed above, the optical axis 7 of the laser beam L from the laser source 4 and the viewing axis 7" of the image recording device 8 are nominally coaxial along the common axis 7' of the substrate holder 9. However, in practice "parasitic" thermal and/or mechanical effects (for example) may cause a slight (regular) drift of the laser optical axis away from the nominal position such that the laser optical axis and the viewing axis of the image recording device 8 are in The intersection of the target surfaces 3 is no longer acceptable for coaxial/coincidence. This is unfavorable because of the nominality of the image recording device 8 illustrated in FIGS. 1 and 3. Align with the appropriate laser beam assumed to be with its viewing axis.

To address this problem, the present invention provides a method (and others) for examining and correcting beam alignment (and other beam parameters) in a device, such as the device discussed above. To this end, a relatively small reference plate 5 is provided on the substrate holder 9 in a position outside the periphery of the clamped substrate 1 (i.e., outside the clamping region of the substrate holder 9). As discussed above, the reference plate (at least a portion of an upper surface) includes a material that produces stimulated radiation in response to irradiation of the laser beam, such as a wavelength converting material such as SiC or GaN. Ideally, the top surface of this reference plate 5 is coplanar with the target surface 3 of the substrate 1 such that there is no "height difference" between the two when viewed along the Z direction; however, this is not necessarily the case. The reference plate 5 can be attached to the substrate holder 9 by various means such as glue, magnetic clamping, screws, and the like. By giving the controller 17 an appropriate command, the substrate holder 9 can be moved to position the reference plate 5 in the path of the laser beam L, i.e. such that the beam L from the laser source 4 (which can be used for this purpose) The choice is to illuminate the top surface of the reference plate 5 by operating in the CW mode instead of the pulse mode, if so desired. The irradiated laser light excites the material of the reference plate 5 such that it produces a localized (regional) glow S' at the illumination/irradiation area. Further use of this reference plate 5 will now be explained with the aid of Figures 4A and 4B.

4A and 4B illustrate a field of view V of an image recording device 8 during execution of an embodiment (one aspect) of a method in accordance with the present invention. More specifically: - Figure 4A shows the field of view V of the situation of Figure 3, whereby the reference plate 5 and the glow S' are viewed; - Figure 4B shows the field of view V of the situation of Figure 1, whereby the target surface 3 is viewed.

In both cases, the field of view V is illustrated as being circular; however, this should not be considered a limitation, and those skilled in the art will recognize that such a field of view may have other shapes, such as a rectangle.

Beginning with Figure 4A, one of the reference points in the field of view V is indicated by R and is here considered to be the geometric center point of V (although other options are also possible as mentioned above). The localized glow produced by the illumination of the laser beam L on the reference plate 5 is illustrated here as being a small circle S'. Ideally, the glow S' will (in this case) exactly coincide with the reference point R; however, due to the drift effect discussed above, there is now one (non-zero) vector displacement (positional deviation) of S' versus R Where it is represented by the vector P. Using the image recognition software, the position of S' can be automatically recognized and the vector P can be automatically determined. Turning now to Figure 4B, the situation above the target surface 3 of the semiconductor substrate 1 is shown. For ease of reference, items S' and P from Figure 4A have been illustrated herein as phantom features using dashed lines. A set of intersecting scribe lanes 2 are shown in the field of view V, one extending in the Y direction and the other extending in the X direction. The dashed line 2' indicates the central axis of the scribe lanes, and the item F indicates a particular feature of interest, which in this case is the intersection of the two axes 2'. Once again, the image recognition software can be used to automatically detect features 2, 2' and F. As illustrated herein, the laser beam L is turned off and is intended to be turned on only when its optical axis is properly aligned with the feature F. To this end, the image recognition software can be used to automatically measure the vector position P' of the feature F relative to the reference point R. According to the invention, the position vector P - P ' of the optical axis relative to the feature F can then be calculated. If the substrate holder 9 moves (by the controller 17) a vector - ( P - P ') before the laser beam L is turned on, the laser will accurately illuminate the feature F.

As an alternative (or supplemental) to one of the set points of the stage assembly 17, the projection system 11 can be embodied, for example, to include at least a laser beam L that can be configured to direct the laser beam L in the X/Y direction. An actuator drives the adjustable optical element and the set point of the element can be adjusted, for example.

Example 2

How the method of the present invention can be used to probe/correct axial beam alignment, i.e., the focus state of the laser beam L employed, will now be explained in more detail.

Turning first to Figure 3, as discussed above, when the laser beam L is illuminated on the reference plate 5, it produces a substantially circular glow S', as illustrated in Figure 4A. Optical science stipulates that the diameter (D) and integral intensity (I) of this spot S' will vary depending on the state of focus of the beam L relative to the top surface of the reference plate 5. This focus state can be adjusted in various ways, for example, by moving the holder 9 up/down along the Z-axis, and/or by using one of the projection systems 11 to adjust the lens element (ALE), as is familiar with this item. The technician will fully understand. More specifically, if the beam L moves over the reference plate 5 past its focus (from above the focus to below the focus), then: - the diameter D will proceed from (relatively) large to small and back to large with the best focus One of the minimum values D _min ; - the integrated intensity I will go from (relatively) low to high and back to low with one of the best focus points, I _max .

A calibration routine can now be performed whereby the values of D and/or I are represented for different focus states of the beam L (appropriately indicating the Z position of the holder 9 and/or the position of the ALE), from which a lookup table can be compiled .

This lookup table (along with possible interpolation/estimation) can be used in later cases to translate one of the measured values of D and/or I into a pair of possible focus states. If two values of D and/or I are measured (close to each other with a small intervention focus state adjustment), a unique focus state can be derived from the lookup table. Knowing the focus state at the reference plate 5 allows the determination of the focus state at the substrate 1 (Fig. 1) - taking into account the possible height difference (along Z) between one of the top surface of the reference plate 5 and the surface 3 of the substrate 1. .

In addition to testing parameters such as the diameter and/or intensity of the glow zone S', the shape can also be (automatically) analyzed. In the present example, this shape is ideally/nominally (substantially) circular; however, as discussed above, it is contemplated that drift effects may result in deformation of this shape, for example, resulting in undesired elongation, an irregular shape. Presentation, a messy perimeter, etc. The image recognition software can recognize such deviations that may exist and alert the device user to pay attention to such deviations before the start of the scribing.

Example 3

As is known from the above, the arrangement illustrated in Figures 1 and 3 employs an on-axis viewing configuration of the image recording device 8. However, for example, it is also conceivable to adopt an off-axis configuration ( One embodiment of the configuration as shown in FIG. Here, the image recording device (camera) 8 is independently directed to the substrate holder 9 and its viewing axis 7" is not coaxial with the optical axis 7 along which the laser beam L propagates. The camera 8 can be pointed so that it is viewed The shaft 7" intersects the optical axis 7 at or near the plane of the target surface 3; thus, the spot S will nominally be at or near the center of the field of view of the camera 8 (if the top surface of the reference plate 5 and the target surface 3 are at least This is roughly coplanar, as will the glow S' (when the reference plate 5 is moved into the field of view of the camera 8)).

This off-axis setting can be used in much the same way as the on-axis configuration of Figures 1 and 3, with the following differences: - Due to the viewing angle of the camera 8, the spot S and the glow S' will typically be somewhat oval/elliptical . For example, the "diameter" D mentioned in Example 2 can then be selected as the lateral/minimum width of the oval/ellipse of interest.

- A height difference between the top surface of the reference plate 5 and the target surface 3 will now tend to play a more significant role in the planar (lateral) alignment of the beam L.

Those skilled in the art will understand these problems and can process them by performing appropriate (smaller) mathematical corrections, for example, using appropriate coordinate transformations.

Example 4

In the case discussed in the above embodiment 1, a single laser beam is irradiated on the substrate 1. As an alternative, Figure 6 illustrates a situation in which multiple laser beams are illuminated on a substrate.

To achieve the scenario depicted in Figure 6, a beam splitting element, such as a diffractive optical element, can be used to divide an original laser L beam into a laser beam cluster {71 to 75}, which in this case The X directions (perpendicular to the plane of Figure 1; parallel to the plane of Figure 2) are separated from one another. It should be noted that for the sake of clarity, only the optical axes of each beam {71 to 75} are shown, otherwise Figure 6 will be extremely confusing; however, this (common) simplification will not prevent the person skilled in the art from understanding the depiction. The structure and operation of the setup.

As illustrated herein, each of the light beams {71 to 75} is focused into one of the substrates 1 The primary surface focus {F1 to F5} (invisible cut), the focus {F1 to F5} adjacent to the beam {71 to 75} is located at different depths {d1 to d5} below the target surface 3. Each of the beams {71 to 75} also produces a corresponding spot {S1 to S5} on the surface 3. The spots {S1 to S5} are placed along a line 15, and the points {F1 to F5} are placed along a line 15. As illustrated herein, the line 15 is inclined at a straight line below the line 13 at a downward angle θ; however, this is not limiting and the line 15 may instead be curved or waved (for example). In particular, the following values apply: The width of the beam cluster {71 to 75} in the X direction is 70 μm.

▪ Depth difference d1 to d5 is 40 μm.

▪ Therefore, the tilt angle θ is tangent (40/70) = 29.75°.

▪ The shallowest focus F5 is located 20 μm below the exposed surface 3.

It should be noted that the line 15 need not be tilted relative to the line 13; in an alternative arrangement, the lines 15 and 13 may be substantially parallel (ie, θ 0). It should also be noted that the line 15 may alternatively be located on the surface 3 rather than below it.

More information on such beam clusters can be gathered from, for example, US 7,968,432.

In the context of the present invention, at least one of the beams in the beam cluster {71 to 75} can be selected and used to perform beam alignment of the entire beam cluster {71 to 75} (and other types of beam characterization). ) One of the foundations.

1‧‧‧Substantially flat semiconductor substrate/substrate

2‧‧‧ scratching the road

2'‧‧‧Dash/axis/axis of the dotted line

2a‧‧‧ scribed

2b‧‧‧ scribed

3‧‧‧ Target surface/surface

4‧‧‧Laser source

5‧‧‧ Relatively small reference board/reference board

6‧‧‧beam splitter

7‧‧‧ Optical axis

7'‧‧‧Common shaft/axis

7”‧‧‧View axis (normal line of sight)/axis

8‧‧‧Image Recording Device/Device/Digital Camera/Image Recording Device (Camera)/Camera

9‧‧‧Removable substrate holder (workbench, chuck) / substrate holder / holder / substrate holder

10‧‧‧ Filter

11‧‧‧projection (ie imaging) system / projection system

12‧‧‧ Controller

13‧‧‧ line

14‧‧‧Controller/Laser Controller

Line 15‧‧‧

16‧‧‧ lights

17‧‧‧Taiwan Assembly/Controller

23‧‧‧Integrated devices/devices

71-75‧‧‧Laser beam cluster/beam/beam cluster

A‧‧‧ device

A1‧‧‧Axis

A2‧‧‧Axis

D1‧‧‧depth/depth difference

D3‧‧‧depth/depth difference

D5‧‧‧Depth/depth difference

F‧‧‧ scribed features/projects/specific features/features of interest

F1‧‧‧ surface focus/focus

F3‧‧‧ surface focus/focus

F5‧‧‧ surface focus/focus

L‧‧‧(pulse) laser beam/laser beam/beam

P‧‧‧Vector/Project

P’‧‧·Vector position

R‧‧‧ reference point

S‧‧‧ spot

S’‧‧‧Localized (Zone) Glow/Glow/Localized Glow/Small Circle/Item/Substantial Round Glow/Light/Glow Zone

S1‧‧‧ corresponding spot/light spot

S3‧‧‧ corresponding spot/spot

S5‧‧‧ corresponding spot/light spot

V‧‧ Sight

X‧‧‧axis

+x‧‧‧direction

-x‧‧‧ directions

Y‧‧‧Axis

-△y‧‧‧

Z‧‧‧Axis

Θ‧‧‧ downward tilt/tilt angle

The invention will now be explained in more detail on the basis of exemplary embodiments and the accompanying schematic drawings in which: FIG. 1 presents a device suitable for performing one of the methods according to the invention. An end view (along an X axis) of a portion of a particular embodiment.

Figure 2 shows a portion of the subject matter of Figure 1 (specifically, a lower portion of Figure 1) Surface (along a Z axis).

3 shows the subject matter of FIG. 1 during execution of an embodiment of a method in accordance with the present invention.

4A and 4B illustrate a field of view of an image recording device during execution of an embodiment of a method in accordance with the present invention.

Figure 5 shows a variation of the device illustrated in Figure 1 (having an "off-axis" image recording device).

Figure 6 presents an enlarged elevational view (along a Y-axis) of one of the elements of Figure 1 (specifically, a lower portion of Figure 1) according to a particular embodiment of the invention employing a laser beam cluster. .

In the figures, corresponding elements are used to indicate corresponding parts.

1‧‧‧Substantially flat semiconductor substrate/substrate

2‧‧‧ scratching the road

2'‧‧‧Dash/axis/axis of the dotted line

3‧‧‧ Target surface/surface

4‧‧‧Laser source

5‧‧‧ Relatively small reference board/reference board

6‧‧‧beam splitter

7‧‧‧ Optical axis

7'‧‧‧Common shaft/axis

7”‧‧‧View axis (normal line of sight)/axis

8‧‧‧Image Recording Device/Device/Digital Camera/Image Recording Device (Camera)/Camera

9‧‧‧Removable substrate holder (workbench, chuck) / substrate holder / holder / substrate holder

10‧‧‧ Filter

11‧‧‧projection (ie imaging) system / projection system

12‧‧‧ Controller

14‧‧‧Controller/Laser Controller

16‧‧‧ lights

17‧‧‧Taiwan Assembly/Controller

A‧‧‧ device

F‧‧‧ scribed features/projects/specific features/features of interest

L‧‧‧(pulse) laser beam/laser beam/beam

P‧‧‧Vector/Project

P’‧‧·Vector position

R‧‧‧ reference point

S’‧‧‧Localized (Zone) Glow/Glow/Localized Glow/Small Circle/Item/Substantial Round Glow/Light/Glow Zone

V‧‧ Sight

X‧‧‧axis

Y‧‧‧Axis

Z‧‧‧Axis

Claims (13)

A method of performing beam characterization in a laser scribe device for scribing a substantially planar semiconductor substrate, the device comprising: a movable substrate holder slidable to the movable substrate holder Thereby, the substrate is clamped in a clamping region of the holder to present a target surface to be scribed along at least one scribe line; a laser source operable to face the optical axis The target surface directs at least one laser beam, the method being characterized by the steps of providing an image recording device that can be used to view and control a scribing process, the device having a field of view having a normal view of one of the views a reference plate on the substrate holder outside the clamping region, the reference plate comprising a material that generates stimulated radiation in response to the irradiation of the laser beam; positioning the substrate holder such that the laser A beam of light is incident on a point of the reference plate and the material is excited to produce a localized glow at the point of illumination; the image recording device is used to detect a physical parameter of the glow. The method of claim 1, wherein the physical parameter is selected from the group consisting of: a position of the glow relative to one of the reference points in the field of view; a diameter of the glow; an integral intensity of the glow One of the glow shapes, and combinations thereof. The method of claim 1 or 2, wherein the information from the image recording device is used to calculate a positional deviation between the optical axis and the viewing axis. The method of claim 1 or 2, wherein the data from the image recording device is used to calculate an axial distance between the illumination point and a position of the best focus of the laser beam. The method of claim 1 or 2, wherein the stimulated radiation is generated via a mechanism selected from the group consisting of: wavelength conversion, two-photon absorption, incandescent light, fluorescence, and combinations thereof. The method of claim 3, comprising the additional step of: providing a clamped substrate on the substrate holder; positioning the substrate holder to position one of the substrates in the field of view; using the image recording The device detects a position of one of the scribe lanes relative to the one of the reference points; applying the positional deviation to determine that the optical axis infers a position relative to one of the features. The method of claim 4, comprising the additional step of: providing a clamped substrate on the substrate holder; positioning the substrate holder to position one of the substrates in the field of view; applying the axial direction The distance is used to determine a focus position of the laser beam relative to the target surface. The method of claim 6, comprising the additional step of: determining a height difference between the reference plate and the clamped substrate at a plane of the substrate; combining at least one of the positional deviation and the axial distance This height difference is used to adjust the beam alignment. The method of claim 7, which includes the following additional steps: Determining a height difference between the reference plate and the clamped substrate at a plane of the substrate; combining the positional deviation and the axial distance uses the height difference to adjust the beam alignment. The method of claim 1 or 2, wherein the material from the image recording device is used to determine a shape of the glow. The method of claim 1 or 2, wherein: the laser source produces a cluster of laser beams; the at least one laser beam is a member of the cluster and serves as a basis for performing beam alignment of the entire cluster. A device for performing laser scribing of a substantially planar semiconductor substrate, the device comprising: a substrate holder having the substrate clampable in one of the clamping regions for presenting along at least one scribe Channel marking a target surface; a laser source for guiding at least one laser beam along an optical axis toward the target surface; for achieving relative motion between the substrate holder and the laser beam A member, characterized in that the device further comprises: an image recording device operable to view and control the scribing process; a reference plate on the substrate holder external to the clamping region, the reference plate comprising Irradiating the laser beam to produce a material of the stimulated radiation, thereby generating a localized glow at an illumination point of the laser beam; a controller for performing the generation by the image recording device An analysis of one of the glow images and from which at least one physical parameter of the glow is derived. The device of claim 12, wherein the physical parameter is selected from the group consisting of Group: a position of the glow relative to one of the reference points in one of the fields of view of the image recording device; a diameter of the glow; an integrated intensity of the glow; a shape of the glow, and combinations thereof.