TWI496661B - Automatic generation of reference spectra for optical monitoring - Google Patents

Description

Automatic generation of reference spectra for optical monitoring

The present invention is generally directed to a reference spectrum for establishing optical monitoring, for example, during chemical mechanical polishing.

An integrated circuit is usually formed on a substrate by continuously depositing a conductive layer, a semiconductive layer or an insulating layer on a germanium wafer. One manufacturing step involves depositing a filler layer on a non-planar surface and planarizing the filler layer. For some applications, the filler layer is planarized until the top surface of the patterned layer is exposed. For example, a layer of conductive filler can be deposited over the patterned insulating layer to fill the trenches or holes in the insulating layer. After planarization, portions of the remaining conductive layer between the floating patterns of the insulating layer form vias, plugs, and lines, such vias, plugs, and wires, and provide conductive paths between the thin film circuits on the substrate. . For other applications, such as oxide milling, the filler layer is planarized until a predetermined thickness remains on the non-planar surface. In addition, lithography generally requires planarization of the substrate surface.

A well-established planarization method is chemical mechanical polishing (CMP). This planarization method typically requires mounting the substrate on a carrier head. Typically the exposed surface of the substrate is placed against a rotating polishing pad having a durable rough surface. The carrier head provides a controllable load on the substrate to push the substrate against the polishing pad. Typically, a slurry is provided to the surface of the polishing pad, such as a slurry having abrasive particles.

One problem in CMP is the use of a suitable polishing speed to achieve the desired profile (e.g., a substrate layer is planarized to a desired flatness or thickness), or the desired amount of material removed. The initial thickness of the substrate layer, the mud composition, the condition of the polishing pad, the relative velocity between the polishing pad and the substrate, and variations in the load on the substrate can cause variations in the rate of material removal across the substrate and the substrate. These changes result in changes in the time required to reach the end of the polishing and the amount of removal. Therefore, it may not be possible to determine the polishing end point based solely on the grinding time, or to achieve the desired appearance only by applying a constant pressure.

In some systems, the substrate is optically monitored in situ during milling, for example, via a window in the polishing pad. However, existing optical monitoring technologies may not meet the increasing demands of semiconductor device manufacturers.

In one aspect, a computer implemented method of generating a reference spectrum includes the steps of: grinding a first substrate in a polishing apparatus having a rotatable platform; and measuring an amount from the substrate by an in-situ monitoring system during grinding a series of spectra; correlating each of the spectra in the series of spectra with an index value equal to one of the number of revolutions of the platform, wherein the respective spectra are measured at the rotation of the platform; and storing the series of spectra as a reference spectrum.

Embodiments may include one or more of the following features. A target index value can be determined. The first substrate may be ground for a predetermined time, and the target index value may be the number of times the platform is rotated at the predetermined time. The first substrate can be monitored by a second in-situ monitoring system, and the polishing end point of one of the first substrates can be monitored by the second in-situ monitoring system. The target index value may be the number of times the platform rotates when the second in-situ monitoring system detects the polishing end point of the first substrate. Determining the target index value can include combining a plurality of endpoint times, and the target index value can be one of a number of platform rotations at the combined plurality of endpoint times. A post-grinding thickness measurement of one of the first substrates can be performed. An initial index value can be determined and the initial index value can be adjusted based on the post-grind thickness measurement. A second substrate can be ground in the polishing apparatus. During the grinding, a second series of spectra from one of the second substrates can be measured using an in-situ monitoring system. For each of the measured spectra in the second series of spectra, a best matching reference spectrum can be determined based on the reference spectrum. For each of the best matching reference spectra, an index value can be determined to produce a series of index values. A linear function can be fitted to the series of index values. Performing the steps of measuring each region in the second substrate: measuring a second series of spectra, determining a best matching reference spectrum based on the reference spectrum, determining an index value, and fitting a linearity to the series of index values function. An estimated time may be determined based on the linear function at which at least one of the regions in the second substrate will reach the target index value. Adjusting a grinding parameter of at least one region of the one substrate to adjust the grinding speed of the at least one region such that the at least one region is closer to the target index than in the absence of such adjustment at the estimated time . An end point may be detected based on a time that is a linear function of one of the reference areas of the at least one region to reach the target index value. An end point can be detected based on a second in situ monitoring system. The second in-situ monitoring system can include one or more of a motor torque monitoring system, an eddy current monitoring system, a friction monitoring system, or a monochrome optical monitoring system.

In another aspect, a computer implementation for controlling polishing a substrate includes the steps of: grinding a substrate; during the grinding, monitoring an area of a substrate with an in-situ spectroscopy system; during the polishing, in addition to the original In addition to the bit spectrum monitoring system, the substrate is monitored by an endpoint detection system; an estimated endpoint time is determined based on a plurality of spectra collected by the in situ spectral monitoring system; and one of the at least one region on the substrate is adjusted to be ground. a parameter to adjust the polishing speed of the at least one region such that at the predicted end time, the at least one region is closer to a target thickness than in the absence of such adjustment; and when the endpoint detection system detects one At the end of the grinding, the grinding is stopped.

Embodiments may include one or more of the following features. The endpoint detection system can include one or more of a motor torque monitoring system, an eddy current monitoring system, a friction monitoring system, or a monochrome optical monitoring system. The in-situ spectroscopy monitoring system can be used during grinding to measure a first series of spectra from a first region of the substrate. For each of the measured spectra in the first series of spectra, a best matching reference spectrum can be found from the first plurality of reference spectra to produce a first series of best matched spectra. For each of the best matching reference spectra in the first series of best matched spectra, one of the best matching reference spectra may be indexed to produce a first series of index values. The in-situ spectroscopy monitoring system can be used to measure the spectrum measured from the second series of one of the second regions of the substrate during milling. For each of the measured spectra in the second series of spectra, a best matching reference spectrum can be found from the second plurality of reference spectra to produce a second series of best matched spectra. For each of the best matching reference spectra in the second series of best matched spectra, one of the best matching reference spectra may be indexed to produce a second series of index values. An estimated time may be determined based on the first series of index values at which the first region of the substrate will reach a target index value. The grinding parameter of one of the second regions can be adjusted such that at the predicted time, the second region is closer to the target index than in the absence of such adjustment.

In one aspect, a computer implemented method of generating a reference spectrum includes the steps of: grinding a first substrate in a polishing apparatus; and measuring a series from the first substrate using an in-situ optical monitoring system during grinding a spectrum; for each of the spectra of the series, determining a best matching reference spectrum from the first plurality of reference spectra to generate a series of reference spectra; calculating the series of spectra to fit one of the series of reference spectra a value; comparing the value of the fitting metric with a threshold, and determining whether to generate a second database based on the comparison; and if the second database is determined to be generated, storing the series of spectra as a second plurality of references spectrum.

Embodiments may include one or more of the following features. A second substrate can be ground in the grinding apparatus; the in-situ monitoring system can be used to measure a second series of spectra from the second substrate during grinding; and for each of the spectra in the second series of spectra, A reference set of spectra of the second plurality of reference spectra is included to determine a best matched reference spectrum to produce a second series of reference spectra. The set of reference spectra can include the first plurality of reference spectra. The set of reference spectra may also not include the first plurality of reference spectra. Each of the second plurality of reference spectra can be associated with an index value that is proportional to one of the number of revolutions of the platform, wherein the spectrum is measured as the platform rotates. For each of the spectra in the second series of reference spectra, a correlation index value can be determined to produce a series of index values, and a linear function can be fitted to the series of index values. A target index value can be determined. The plurality of regions of the second substrate can perform the steps of measuring a second series of spectra, determining a best matching reference spectrum, determining a correlation index value, and fitting a linear function. An estimated time may be determined based on the linear function at which at least one of the regions in the second substrate will reach the target index value. Adjusting a grinding parameter of the at least one region of the second substrate to adjust a grinding speed of the at least one region such that at the predicted time, the at least one region is closer to the condition than in the absence of such adjustment Target index value. A polishing end point of the second substrate can be determined based on the time at which the linear function reaches a target index value. Determining the target index value can include calculating the target index value using the at least one first value. The first value may be one of the number of revolutions of the platform, wherein an end of a third substrate that is ground prior to the first substrate is detected where the platform is rotated. Calculating the target index value may include calculating an average of a plurality of values including the first value and a second value. The second value may be one of the number of revolutions of the platform, wherein an end of a fourth substrate that is ground prior to the third substrate is detected where the platform is rotated. The fourth substrate, the third substrate, and the first substrate may be continuously polished. In addition to the optical monitoring system, the first substrate is monitored by a second in-situ monitoring system. The second in-situ monitoring system can be used to detect a grinding end point of the first substrate. Determining the target index value can include determining a number of times the platform is rotated when the second in-situ monitoring system detects the polishing end of the first substrate. The second in-situ monitoring system can include one or more of a motor torque monitoring system, an eddy current monitoring system, a friction monitoring system, or a monochrome optical monitoring system. A grinding thickness measurement can be performed after the first substrate can be performed. An initial index value can be determined, and determining the target index value can include adjusting the initial index value based on the post-grind thickness measurement. For each of the spectra in the first series of reference spectra, a correlation index value can be determined to produce a series of index values. A linear function can be fitted to the series of index values. The fit metric can include the linear function fitting a good fit to one of the series of index values. The fit can include a sum of one of the squared differences between the series of index values and the linear function. Calculating the value of the fit metric can include determining, for each of the series of spectra, a difference between the spectrum and the best matched reference spectrum to provide a series of differences and accumulating the differences . Determining a difference between the spectrum and the best matched reference spectrum can include calculating a sum of one of the absolute values of the intensity differences over a range of wavelengths or calculating a sum of one of the intensity squared differences over a range of wavelengths.

In other aspects, a grinding system and a computer program product specifically mounted on a computer readable medium are provided to perform the methods.

Some embodiments may have one or more of the following advantages. The reference spectrum and a target index value can be automatically generated, thus significantly reducing the time required for the semiconductor manufacturer to begin polishing a new device substrate (eg, a substrate based on a new mask pattern). The reference spectrum and a target index value can be automatically triggered, thus allowing the creation of a new reference library that can be used to polish a subsequent substrate (eg, a subsequent substrate of the same batch). Adding or switching to a new database can improve the reliability and accuracy of the endpoint determination and/or reduce in-wafer non-uniformity when a previous database does not provide a better fit. Eliminates the need for different preset algorithms for each device/mask pattern.

The details of one or more embodiments are set forth in the drawings and the description below. Other features, aspects, and advantages will become apparent from the description, drawings and claims.

For an optical monitoring system, establishing a reference spectrum and a target can be time consuming, wherein the optical monitoring system is used to monitor the spectrum of reflected light from the substrate undergoing grinding. However, for example, the reference spectrum is automatically established by measuring the spectrum of the first substrate from a batch and using the measured spectrum as a reference spectrum. For example, the target can be automatically established by using the second endpoint detection system to identify the grinding endpoint time and then determining the time-related index. Thereafter, optical monitoring of subsequent substrates can be performed using established reference spectra and targets. In addition, a new reference spectral library can be automatically triggered. For example, if the reference spectrum of the database does not provide a good fit during the polishing of the substrate (eg, the fit exceeds a certain threshold), the measured spectrum from the substrate can be stored as a reference for the new database. spectrum. Thereafter, a new database can be used for optical monitoring of subsequent substrates. Therefore, the time required for the semiconductor manufacturer to start grinding the substrate having the new pattern can be significantly reduced.

FIG. 1 illustrates an example of a grinding apparatus 100. The grinding apparatus 100 includes a rotatable disk-shaped platform 120 in which the polishing pad 110 is located on the platform 120. The platform is operable to rotate about the axis 125. For example, the motor 121 can rotate the drive shaft 124 to rotate the platform 120. For example, the polishing pad 110 can be detachably secured to the platform 120 by an adhesive layer. The polishing pad 110 can be a two-layer polishing pad having an outer polishing layer 112 and a softer back layer 114.

The grinding apparatus 100 can include a combined mud/washing arm 130. During grinding, the arm 130 is operable to dispense a slurry 132, such as mud, on the polishing pad 110. Although only one mud/wash arm 130 is shown, additional nozzles may be used, such as one or more dedicated mud arms per carrier head. The polishing apparatus can also include a polishing pad conditioner for abrading the polishing pad 110 to maintain the polishing pad 110 in a consistent abrasive state.

In this embodiment, the grinding apparatus 100 includes two (or two or more) carrier heads 140. Each carrier head 140 is operable to hold the substrate 10 (e.g., the first substrate 10a on one carrier head and the second substrate 10b on the other carrier head) against the polishing pad 110, i.e., the same polishing pad. Each carrier head 140 can independently control the grinding parameters associated with each substrate, such as pressure. In some embodiments, the grinding apparatus 100 includes a plurality of carrier heads, but the carrier heads (and the holding substrates) are located on different polishing pads rather than the same polishing pad. For such embodiments, the following discussion of obtaining the simultaneous endpoints of multiple substrates on the same platform is not applicable, but the discussion of simultaneous endpoints for obtaining multiple regions (although on a single substrate) will still apply.

In detail, each of the carrier heads 140 may include a positioning ring 142 to position the substrate 10 below the elastic film 144. Each of the carrier heads 140 also includes a plurality of independently controllable pressurizable chambers (e.g., three chambers 146a-146c) defined by a membrane so as to be independent of the elastic membrane 144 and associated regions 148a-148c on the substrate 10. Apply pressure to the ground (see Figure 2). Referring to Figure 2, the central region 148a can be generally circular, and the remaining regions 148b-148c can be concentric annular regions surrounding the central region 148a. Although only three chambers are illustrated for illustration in Figures 1 and 2, there may be two chambers or four or more chambers, such as five chambers.

Returning to Figure 1, each carrier head 140 is suspended from a support structure 150 (e.g., a swivel rack) and coupled to a carrier head rotation motor 154 by a drive shaft 152 such that the carrier head can rotate about the shaft 155. Alternatively, for example, each of the carrier heads 140 may be laterally vibrated by a sliding member on the rotating rack 150 or by rotational vibration of the rotating rack itself. In operation, the platform rotates about the center axis 125 of the platform, and each carrier head rotates about the center axis 155 of each carrier head and translates laterally across the top surface of the polishing pad.

Although only two carrier heads 140 are shown, more carrier heads may be provided to hold additional substrates so that the surface area of the polishing pad 110 can be effectively utilized. Thus, for a simultaneous polishing process, the number of carrier head elements suitable for holding the substrate can be based, at least in part, on the surface area of the polishing pad 110.

The grinding apparatus also includes an in-situ monitoring system 160 that can be used to determine whether to adjust the grinding speed or to adjust the grinding speed as described below. The in situ monitoring system 160 can include an optical monitoring system, such as a spectral monitoring system or an eddy current monitoring system.

In one embodiment, the monitoring system 160 is an optical monitoring system. The optical entrance through the polishing pad is provided by including an aperture (i.e., a hole through the pad) or a solid window 118. The solid window 118 can be secured to the polishing pad 110 (eg, as a plug that fills the aperture in the polishing pad, for example, molded or adhesively secured to the polishing pad), but in some embodiments, the solid window can be supported on the platform 120. Up and raised into the aperture of the polishing pad.

The optical monitoring system 160 can include a light source 162, a light detector 164, and circuitry 166 to transmit and receive signals between the remote controller 190 (eg, a computer) and the light source 162 and the light detector 164. One or more optical fibers can be used to transfer light from the light source 162 to the optical inlet in the polishing pad and to transmit the reflected light from the substrate 10 to the detector 164. For example, the bifurcated fiber 170 can be used to transfer light from the light source 162 to the substrate 10 and back to the detector 164. The bifurcated fiber can include a trunk 172 positioned adjacent to the optical inlet and two branches 174 and 176 coupled to the light source 162 and the detector 164, respectively.

In some embodiments, the top surface of the platform can include a recess 128 in which the optical head 168 holding one end of the main line 172 of the bifurcated fiber is secured in the recess 128. The optical head 168 can include a mechanism for adjusting the vertical distance between the top of the trunk 172 and the solid window 118.

The output of circuit 166 can be a digital electronic signal that is transmitted along a drive shaft 124 to a controller 190 of the optical monitoring system via a rotary coupler 129 (eg, a slip ring). Similarly, the light sources can be turned on or off in response to control commands that are communicated from the controller 190 to the digital electronic signals of the optical monitoring system 160 via the rotary coupler 129. Alternatively, circuit 166 can be communicated to controller 190 by wireless signals.

Light source 162 can be operated to emit white light. In one embodiment, the emitted white light comprises light having a wavelength of from 200 nanometers to 800 nanometers. A suitable source of light is a xenon lamp or a mercury lamp.

The photodetector 164 can be a spectrometer. A spectrometer is an optical instrument used to measure the intensity of light on a portion of the electromagnetic spectrum. A suitable spectrometer is a grating spectrometer. A typical spectrometer output is the light intensity as a function of wavelength (or frequency).

As described above, the light source 162 and the photodetector 164 can be coupled to a computing device, such as a controller 190 that is operable to control the operation of the light source 162 and the photodetector 164 and to receive signals from the light source 162 and the photodetector 164. . The computing device can include a microprocessor located adjacent to the grinding device, such as a programmable computer. In terms of control, for example, the computing device can synchronize the opening of the light source with the rotation of the platform 120.

In some embodiments, the light source 162 and detector 164 of the in-situ monitoring system 160 are mounted in the platform 120 and rotate with the platform 120. In this case, the motion of the platform will cause the sensor to scan across the various substrates. In particular, when the platform 120 is rotated, the controller 190 can cause the light source 162 to emit a series of flashes that only begin/end before/after the respective substrate 10 passes through the optical inlet. Alternatively, the computing device can cause the light source 162 to continuously emit light that only begins/ends before/after the respective substrate 10 passes through the optical inlet. In either case, the signal from the detector can be integrated during the sampling period to produce a spectral measurement at the sampling frequency.

In operation, for example, controller 190 can receive a signal carrying spectral information describing the light received by the photodetector for a particular source flash or detector period. Therefore, the spectrum is the spectrum measured in situ during grinding.

As shown in FIG. 3A, if the detector is mounted in the platform, due to the rotation of the platform (shown by arrow 204), when the window 108 is moved under a carrier head (eg, a carrier that holds the first substrate 10a), An optical monitoring system that performs spectral measurements at the sampling frequency will cause spectral measurements to be taken at a location 201 across the arc of the first substrate 10a. For example, each of the points 201a-201k represents a spectral measurement position of the monitoring system of the first substrate 10a (the number of points is illustrative; depending on the sampling frequency, more or less than the stated measurement can be performed Measurement). As shown, within one revolution of the platform, the spectra are obtained at different radii on the substrate 10a. That is, certain spectra are obtained from locations near the center of the substrate 10a and some are close to the edges. Similarly, as shown in FIG. 3B, due to the rotation of the platform, when the window is moved under another carrier head (for example, the carrier head holding the second substrate 10b), the optical monitoring system for performing spectral measurement at the sampling frequency will The resulting spectral measurements are taken at a location 202 across the arc of the second substrate 10b.

Thus, for any given rotation of the platform, based on timing and motor encoder information, the controller can determine which substrate (eg, substrate 10a or 10b) is the source of the measured spectrum. In addition, for any given scan of the optical monitoring system across the substrate (eg, substrate 10a or 10b), controller 190 may be based on timing, motor encoder information, and edge optical detection of the substrate and/or positioning ring. The radius position from each of the measured spectra of the scan (relative to the center of the particular substrate 10a or 10b being scanned) is calculated. The grinding system can also include a rotational position sensor (e.g., attached to the flange of the edge of the platform that will pass through the fixed optical interrupter) to provide additional information to determine the substrate position of the substrate and the measured spectrum. Thus, the controller can associate various measured spectra with controllable regions 148b-148c (see Figure 2) on substrates 10a and 10b. In some embodiments, the spectral measurement time can be used as an alternative to accurately calculating the radius position.

Within a plurality of rotations of the platform, a series of spectra can be obtained over time for each region of each substrate. Without any particular theoretical limitation, since the thickness of the outermost layer changes, the spectrum of the reflected light from the substrate 10 progresses with the grinding (eg, during multiple rotations of the platform, rather than a single flaw across the substrate). Gradually formed during the sweep, resulting in a series of time-varying spectra. In addition, a particular spectrum is represented by a particular thickness of the layer stack.

In some embodiments, a controller (eg, a computing device) can be programmed to compare the measured spectrum to a plurality of reference spectra and determine which reference spectrum provides the best match. In particular, the controller can be programmed to compare each of the spectra and the plurality of reference spectra in a series of spectra from each of the various regions of the respective substrate to produce a series of best matching references for each of the regions of the respective substrate. spectrum.

As used herein, the reference spectrum is a predefined spectrum that is produced prior to polishing the substrate. The reference spectrum can have a predefined association (i.e., prior to the grinding operation) with a time value indicative of the occurrence of the desired spectrum in the polishing process, while assuming that the actual polishing rate follows the expected polishing rate. Alternatively or additionally, the reference spectrum may have a predefined association with substrate property values such as the outermost layer thickness.

For example, a reference spectrum can be empirically generated by measuring the spectrum from a test substrate (eg, a test substrate having a thickness of a known initial layer). For example, to generate a plurality of reference spectra, when a series of spectra is collected, the set substrate is ground using the same grinding parameters that will be used during the polishing of the wafer. For each spectrum, a value is recorded to indicate the time at which the spectrum was collected during the polishing process. For example, the value can be the number of times the time or platform has been rotated. The substrate can be over-polished (i.e., ground beyond the desired thickness) such that the spectrum of reflected light from the substrate can be obtained when the target thickness is achieved.

To correlate the individual spectra to substrate property values such as the thickness of the outermost layer, the initial spectra and properties of the "set" substrate having the same pattern as the product substrate pattern can be pre-polished at the metrology station. The final spectrum and properties of the post-milling can also be measured using the same metrology station or different metrology stations. The spectral properties between the initial spectrum and the final spectrum can be determined by interpolation, for example based on linear interpolation of the elapsed time of the spectrum of the test substrate.

In addition to being empirically determined, some or all of the reference spectra may be calculated theoretically, such as using an optical model of the substrate layer. Moreover, for example, an optical model can be used to calculate a reference spectrum for a given outer layer thickness D. For example, by assuming that the outer layer is removed at a uniform polishing rate, a value indicative of the time of the reference spectrum to be collected in the polishing process can be calculated. For example, by assuming that the initial thickness is D0 and the uniform polishing speed is R, the time Ts (Ts = (D0 - D) / R) for a specific reference spectrum can be simply calculated. As another example, the pre-grinding and post-grinding thicknesses D1, D2 (or other thicknesses measured at the metrology station) based on the thickness D used by the optical model may be performed between the measurement times T1, T2 Linear interpolation (Ts=T2-T1*(D1-D)/(D1-D2)).

Referring to Figures 4 and 5, during the grinding, the measurement spectrum 300 (see Figure 4) can be compared to a reference spectrum 320 from one or more databases 310 (see Figure 5). As used herein, a reference spectral library is a collection of reference spectra representing substrates that share a common property. However, the nature of common sharing in a single database can vary across multiple databases of reference spectra. For example, two different repositories may include a reference spectrum representing a substrate having two different underlying thicknesses. For a given library of reference spectra, the variation in thickness of the upper layer can be the primary cause of the difference in spectral intensity, rather than other factors (such as wafer pattern, thickness of the underlying layer, or difference in layer composition).

A reference spectrum 320 for a different database 310 can be generated by grinding a plurality of "set" substrates having different substrate properties (eg, a lower layer thickness or layer composition) and collecting spectra as described above; from a setting The spectrum of the substrate provides a first library, and a spectrum from another substrate having a different underlying thickness provides a second library. Alternatively or additionally, the reference spectra of the different databases may be calculated theoretically, for example, an optical model of the lower layer having a first thickness may be used to calculate the spectrum of the first database, and an optical model of the lower layer having a different thickness may be used. Calculate the spectrum of the second database.

In some embodiments, each reference spectrum 320 is assigned an index value 330. In general, each database 310 can include a plurality of reference spectra 320, such as one or more (eg, only one) reference spectra for each platform rotation during the expected polishing time of the substrate. The index 330 can be a value (eg, a number) indicative of the time at which the reference spectrum 320 is desired to be viewed in the polishing process. The spectrum can be indexed such that each spectrum in a particular library has a unique index value. Indexing can be performed to arrange the index values in the order in which the spectra are measured. The index value can be selected to vary monotonically with the progress of the grinding, such as increasing or decreasing. In particular, the index values of the reference spectra can be selected such that the index values form a linear function of the time or number of plate rotations (assuming the grinding speed follows the speed of the model or test substrate used to generate the reference spectrum in the database). For example, the index value can be proportional to (eg, equal to) the number of times the platform is rotated, where the test substrate is measured at the reference point of rotation or where the reference spectrum will appear as an optical model. Therefore, each index value can be an integer. The number of indices can represent the expected platform rotation that will occur in the relevant spectrum.

The reference spectrum and its associated index values can be stored in a reference library. For example, each reference spectrum 320 and its associated index value 330 can be stored in a record 340 of a database 350. A database 350 of reference libraries of reference spectra can be implemented in the memory of the computing device of the polishing apparatus.

In some embodiments, a reference spectrum of a given batch of substrates can be automatically generated. When the optical monitoring system measures the spectrum, a certain batch of the first substrate or the first substrate having the new device/mask pattern is ground, but the polishing speed is not controlled (as described in Figures 8 to 10 below). This condition produces a series of spectra of the first substrate, each region of the window below the substrate, each sweep (eg, per platform rotation) having at least one spectrum.

For example, a set of reference spectra for each region is automatically generated from a series of spectra of the first substrate. In short, the spectrum measured from the first substrate becomes a reference spectrum. More specifically, the measured spectrum from each region of the first substrate becomes the reference spectrum for that region. Each reference spectrum is associated with a number of revolutions of the platform, wherein the reference spectrum is measured from the first substrate at the rotation of the platform. If there are multiple measurements of a particular region of the first substrate at a particular platform rotation, the spectra can be combined, for example, averaged to produce an average spectrum of platform rotation. Alternatively, the reference library can simply save each spectrum as a separate reference spectrum and, as described below, compare the measured spectra of the subsequent substrates to the respective reference spectra to find the best match. Optionally, a database can store a set of predetermined reference spectra, and then replace a set of predetermined reference spectra by a set of reference spectra generated from a series of spectra in the first substrate.

As mentioned above, the target index value can also be generated automatically. In some embodiments, the first substrate is ground for a fixed polishing time, and the number of platform rotations at the end of the fixed polishing time can be set to a target index value. In some embodiments, wafer-to-wafer feedforward or feedback control may be used in some form from a factory host or CMP tool (eg, as described in U.S. Application Serial No. 12/625,480, incorporated herein by reference). To adjust the grinding time of the first wafer instead of the fixed grinding time. The number of platform rotations at the end of the adjustment grinding time can be set as the target index value.

In some embodiments, as shown in FIG. 1, the polishing system can include another endpoint detection system 180 (in addition to the spectral optical monitoring system 160), such as using friction measurements (eg, as incorporated by reference) U.S. Patent No. 7,513,818; the eddy currents (for example, as described in U.S. Patent No. 6,924,641, incorporated by reference); Or a monochromatic light, such as a laser (for example, as described in U.S. Patent No. 6,719,818, incorporated herein by reference). Another endpoint detection system 180 can be in the separation recess 129 in the platform or in the same recess 128 as the optical monitoring system 160. In addition, although another endpoint detection system is located on the opposite side of the axis of rotation 125 of the platform as shown in FIG. 1, this condition is not required; however, the sensor of the endpoint detection system 180 can be coupled to the optical monitoring system 160. The same radial moment from the axis 125 is the same. The other end point detection system 180 can be used to detect the polishing end point of the first substrate, and can set the number of platform rotations when the other end point detection system detects the end point as the target index value. In some embodiments, the first substrate post-grinding thickness measurement can be performed, and for example, the initial target index value determined by one of the above techniques can be adjusted by linear scaling, for example, by The thickness is measured by multiplying the ratio of the target thickness by the subsequent grinding amount.

In addition, the target index value can be further improved based on the new processing substrate and the new desired end time. In some embodiments, the target index can be dynamically determined based on a plurality of previously polished substrates, for example, by combining (eg, weighted averaging) feedforward or feedback control of the wafer to another endpoint detection system. Instead of using only the first substrate to set the target index value, the indicated end time. A predefined number of previously ground substrates (eg, four or less) that are ground prior to the existing substrate can be used in the calculation.

In any case, once the target index value is determined, the technique described below can be used to grind one or more subsequent substrates to adjust the pressure applied to one or more regions such that the region is less than the adjusted condition. The target index is closer to the same time (or at the expected end time, closer to its target index).

As described above, for each region of each substrate, based on a series of measured spectra or regions and substrates, controller 190 can be programmed to produce a series of best matched spectra. The best matching reference spectrum can be determined by comparing the measured spectra with a reference spectrum from a particular database.

In some embodiments, for each reference spectrum, the best matching reference spectrum can be determined by calculating the sum of the squared differences between the measured spectrum and the reference spectrum. The reference spectrum with the lowest sum of squared differences has the best fit. There may be other techniques for finding the best matching reference spectrum.

The method that can be applied to reduce computer processing is to limit the portion of the database used to search for matching spectra. The database typically includes a broader spectral range than the spectral range that will be obtained when polishing the substrate. During substrate polishing, database searches are limited to a predetermined range of library spectra. In some embodiments, the current rotation index N of the substrate being ground is determined. For example, in the initial platform rotation, N can be determined by searching the reference spectra of all databases. For the spectra obtained during subsequent rotations, the database is searched for within the range of degrees of freedom of N. That is, if the index number is found to be N during one rotation, and the degree of freedom is Y during the subsequent rotation after X rotations, the search range will be from (N+X)-Y to (N+X)+Y. .

Referring to Figure 6, this Figure 6 illustrates only the results of a single region of a single substrate, and an index value for each of a series of best matching spectra can be determined to produce an index value 212 for the time varying series. This series of index values can be referred to as an index trajectory 210. In some embodiments, the index trajectory is generated by comparing the respective measured spectra to a reference spectrum from only one database. In general, the index trajectory 210 can include one (eg, exactly one) index of each swept index of the optical monitoring system beneath the substrate.

For a given index trajectory 210 in which there are multiple measurement spectra (referred to as "current spectra") for a particular substrate and region of a single sweep of the optical monitoring system, each of the current spectra can be The best match is determined between the reference spectra of one or more (eg, exactly one) databases. In some embodiments, each of the selected current spectra is compared to a respective reference spectrum of one or more selected databases. Given the current spectra e, f and g, and the reference spectra E, F and G, for example, the matching coefficients can be calculated for each of the following combinations of the current spectrum and the reference spectrum: e and E, e and F, e and G, f and E, f and F, f and G, g and E, g and F, and g and G. Any matching coefficient indicates the best match (eg, minimum), which determines the best matching reference spectrum and index value. Alternatively, in some embodiments, the current spectrum can be combined (eg, averaged) and the resulting combined spectrum compared to the reference spectrum to determine the best match and index value.

In some embodiments, a plurality of index trajectories may be generated for at least some regions of some of the substrates. For a given area of a given substrate, an index trajectory can be generated for each reference library of interest. That is, for each reference library of interest for a given area of a given substrate, comparing each of the measured spectra in a series of measured spectra with a reference spectrum from a given database determines a series of best matches. The reference spectrum, and the index values of a series of best matching reference spectra provide an index trajectory for a given database.

In summary, each index trajectory includes a series 210 of index values 212, wherein each particular index value 212 of the series is generated by selecting an index of the reference spectrum from a given library that best fits the measurement spectrum. The time value of each index in the index trajectory 210 can be the same as the time at which the measurement spectrum is measured.

Referring to Figure 7, a plurality of index trajectories are illustrated. As described above, an index trajectory can be generated for each region of each substrate. For example, for the first region of the first substrate, a first series 210 of index values 212 (shown by hollow circles) can be generated; for a second region of the first substrate, an index value 222 can be generated (by solid squares) a second series 220 of the second substrate; for the first region of the second substrate, a third series 230 of index values 232 (shown by solid circles) can be generated; and for the second region of the second substrate, The fourth series 240 of index values 242 (shown by the squares of the holes).

As shown in FIG. 7, for each substrate index trajectory, a polynomial function of a known order (eg, a first order function (eg, line)) is fitted to the relevant region and, for example, using a solid line fit. One of the series of index values for the wafer. For example, the first line 214 can be fitted to the index value 212 of the first region of the first substrate; the second line 224 can be fitted to the index value 222 of the second region of the first substrate; the third line 234 can be fitted to the first The index value 232 of the first region of the second substrate; and the fourth line 244 can be fitted to the index value 242 of the second region of the second substrate. The fitting of the index value may include calculating the slope S of the line and the x-axis intersection time T of the line and the starting index value (eg, 0). The function can be expressed in the following form: I(t) = S‧(t-T), where t is time. The x-axis intersection time T can have a negative value indicating that the starting thickness of the substrate layer is less than the expected thickness. Therefore, the first line 214 can have a first slope S1 and a first x-axis intersection time T1; the second line 224 can have a second slope S2 and a second x-axis intersection time T2; the third line 234 can have a third slope S3 and The third x-axis intersects time T3; and the fourth line 244 can have a fourth slope S4 and a fourth x-axis intersection time T4.

For example, in the case where a plurality of substrates are simultaneously polished on the same polishing pad, variations in the polishing speed between the substrates may cause the substrates to reach their target thicknesses at different times. On the one hand, if the substrate is stopped at the same time, some of the substrates will not reach the desired thickness. On the other hand, if the substrate is ground at different times, some of the substrates may have defects and operate the polishing apparatus at a lower throughput.

By determining the polishing speed of each region of each substrate from the in-situ measurement, the estimated end time of the target thickness of each region of each substrate or the expected thickness of the target end time can be determined, and at least one of the at least one substrate can be adjusted. The grinding speed of the area allows the substrate to reach a nearer end condition. By "closer end condition", the term means that the substrate area will reach the target thickness of the substrate area at the same time closer than in the absence of such adjustment, or if the substrate is stopped at the same time, The substrate area will be closer to the same thickness than without such adjustment.

Adjusting the polishing parameters of at least one region of at least one substrate (eg, at least one region per substrate) at a time during the polishing process, such as at time T0, to adjust the polishing rate of the substrate region such that the polishing endpoint Time, the plurality of regions of the plurality of substrates are closer to their target thickness than without such adjustments. In some embodiments, each of the plurality of substrates may have approximately the same thickness at the end time.

Referring to Figure 8, in some embodiments, an area of a substrate is selected as the reference area and the reference area is determined to reach the predicted end time TE of the target index IT. For example, as shown in FIG. 8, the first region of the first substrate is selected as the reference region, but different regions and/or different substrates may be selected. The target thickness IT is set by the user before the grinding operation and storage.

To determine the estimated time that the reference region will reach the target index, a line of the reference region (eg, line 214) can be calculated to intersect the target index IT. Assuming that the grinding speed in the remaining grinding process does not deviate from the expected grinding speed, a series of index values should retain a substantially linear progression. Therefore, the expected end time TE can be calculated as a simple linear interpolation of the line-to-target index IT, such as IT = S ̇ (TE-T). Therefore, in the example of Fig. 8, the first region of the first substrate is selected as the reference region, wherein the associated first line 214, IT = S1 ̇ (TE - T1), that is, TE = IT / S1 - T1.

One or more regions (eg, all regions) other than the reference region (including regions on other substrates) may be defined as adjustable regions. The position at which the line of the adjustable region intersects the expected end time TE defines the predicted end point of the adjustable region. Thus, a linear function of each adjustable region (eg, lines 224, 234, and 244 in FIG. 8) can be used to extrapolate the index that will be achieved at the predicted end time ET of the relevant region, such as EI2, EI3, and EI4. . For example, the second line 224 can be used to extrapolate the predicted index EI2 at the predicted end time ET of the second region of the first substrate; the third line 234 can be used to extrapolate the prediction of the first region of the second substrate The predicted index EI3 at the end time ET; and a fourth line can be used to extrapolate the predicted index EI4 at the predicted end time ET of the second region of the second substrate.

As shown in FIG. 8, if the polishing speed of any region in the region of any substrate is not adjusted after time T0, then if the end points of all the substrates are forced at the same time, each substrate may have a different thickness or each substrate may have Different endpoint times (this condition is undesired because this condition can lead to defects and loss of yield). Here, for example, the end of the second region of the first substrate (shown by line 224) will be at a larger index (and less thicker) than the predicted index of the first region of the first substrate. Likewise, the end of the first region of the second substrate will be at a smaller index (and greater thickness) than the predicted index of the first region of the first substrate.

As shown in Figure 8, if the target index is reached at different times for different substrates (or, equally, the adjustable region will have different predicted indices at the expected end time of the reference region), then up or down Lower the grinding speed so that the substrate will reach the target index (and target thickness) at the same time closer than in the absence of such adjustment, for example at approximately the same time; or the substrate will be in the absence of such adjustment There are closer identical index values (and the same thickness) at the target time, such as approximately the same index value (and approximately the same thickness).

Thus, in the example of Figure 8, at least one of the grinding parameters of the second region of the first substrate is modified starting at time T0 such that the grinding speed of the region is reduced (and thus the slope of index trajectory 220 is reduced). Likewise, in this example, at least one of the grinding parameters of the first region of the second substrate is modified such that the grinding speed of the region increases (and thus the slope of the index trajectory 230 increases). Similarly, in this example, at least one of the grinding parameters of the second region of the second substrate is modified such that the grinding speed of the region increases (and thus the slope of the index trajectory 240 increases). Therefore, at approximately the same time, both regions of the two substrates will reach the target index (and target thickness) (or if the two substrates are stopped at the same time, the two regions of the two substrates will be approximately the same End of thickness).

In some embodiments, if the predicted index at the predicted end time ET indicates that the substrate area is within a predefined range of target thicknesses, the adjustment area may not be needed. The range can be 2% of the target index, for example within 1%.

The grinding speed of the adjustable area can be adjusted such that all areas are closer to the target index than the expected end time in the absence of such adjustments. For example, the reference area of the reference substrate can be selected and the processing parameters of all other areas adjusted so that the end points of all areas will approximate the expected time of the reference substrate. The reference area can be, for example, a predetermined area, such as a central area 148a or a region 148b directly surrounding the central area, an area having the earliest or latest predicted end time of any area of any substrate, or an area of the substrate having the desired end point. If the polishing is stopped at the same time, the earliest time corresponds to the thinnest substrate. Similarly, if the polishing is stopped at the same time, the latest time corresponds to the thickest substrate. The reference substrate can be, for example, a predetermined substrate, i.e., a substrate having a region having the earliest or latest predicted end time of the substrate. If the grinding is stopped at the same time, the earliest time corresponds to the thinnest area. Similarly, if the polishing is stopped at the same time, the latest time corresponds to the thickest region.

For each of the adjustable regions, the desired slope of the index trajectory can be calculated such that the adjustable region reaches the target index at the same time as the reference region. For example, the desired slope SD can be calculated from (IT-I)=SD*(TE-T0), where I is the index value at the time T0 at which the grinding parameter will change (calculated from a linear function fitted to a series of index values) , IT is the target index, and TE is the estimated end time. In the example of FIG. 8, for the second region of the first substrate, the desired slope SD2 can be calculated from (IT-I2)=SD2*(TE-T0); for the first region of the second substrate, The desired slope SD3 can be calculated from (IT-I3)=SD3*(TE-T0); and for the second region of the second substrate, the desired slope can be calculated from (IT-I4)=SD4*(TE-T0) SD4.

Referring to Figure 9, in some embodiments, there is no reference area. For example, the predicted endpoint time TE' can be, for example, a predetermined time set by the user prior to the polishing process, or an average of the expected endpoint times from two or more regions of the one or more substrates. Calculated by other combinations (eg, by projecting the lines of various regions to the target index). In this embodiment, the desired slope is calculated substantially as described above (using the predicted end time TE' instead of TE), but the desired slope of the first region of the first substrate must also be calculated, for example, from (IT-I1) =SD1*(TE'-T0) calculates the desired slope SD1.

Referring to Figure 10, in some embodiments, (which may also be combined with the embodiment shown in Figure 9), there are different target indices for different regions. This condition allows for a prudent but controllable uneven thickness profile on the substrate. The user can enter the target index using, for example, an input device on the controller. For example, the first region of the first substrate may have a first target index IT1; the second region of the first substrate may have a second target index IT2; the first region of the second substrate may have a third target index IT3; The second region of the substrate may have a fourth target index IT4.

For each adjustable region, the desired slope SD can be calculated from (IT-I)=SD*(TE-T0), where I is the region index value at the time T0 at which the grinding parameter will change (self-fitting to the region) A linear function of a series of index values), IT is the target index for a particular region, and TE is the calculated end time (from the reference region associated with Figure 8 as described above, or from a preset endpoint time or as described above) The combination of the expected end time associated with Figure 9). In the example of FIG. 10, for the second region of the first substrate, the desired slope SD2 can be calculated from (IT2-I2)=SD2*(TE-T0); for the first region of the second substrate, The desired slope SD3 can be calculated from (IT3-I3)=SD3*(TE-T0); and for the second region of the second substrate, the desired slope can be calculated from (IT4-I4)=SD4*(TE-T0) SD4.

For any of the methods described above with respect to Figures 8 through 10, the grinding speed is adjusted to bring the index trajectory slope closer to the desired slope. For example, the grinding speed can be adjusted by increasing or decreasing the pressure in the corresponding chamber of the carrier head. It can be assumed that the change in the grinding speed is proportional to the pressure change, for example, a simple Prestonian model. For example, for each region of each substrate, in the case where the region is ground with the pressure Pol before the time T0, the new pressure Pnew applied after the time T0 can be calculated as Pnew=Pold*(SD/S), where S Is the line slope before time T0 and SD is the desired slope.

For example, assume that the pressure Pold1 is applied to the first region of the first substrate, the pressure Pol2 is applied to the second region of the first substrate, the pressure Pol3 is applied to the first region of the second substrate, and the pressure Pold4 is applied to the second region of the second substrate. a region; the new pressure Pnew1 of the first region of the first substrate can be calculated as Pnaw1=Pold1*(SD1/S1); the new pressure Pnew2 of the second region of the first substrate can be calculated as Pnew2=Pold2*(SD2/ S2); calculating a new pressure Pnew3 of the first region of the second substrate as Pnew3=Pold3*(SD3/S3); and calculating a new pressure Pnew4 of the second region of the second substrate as Pnew4=Pold4*(SD4) /S4).

The process of determining the expected time at which the substrate will reach the target thickness and adjusting the polishing rate may only be performed once during the polishing process, such as at a specified time (eg, at 40% to 60% of the expected polishing time); or in a grinding process It is executed multiple times during the period, for example every thirty to sixty seconds. The subsequent time during the polishing process, if appropriate, the speed can be adjusted again. During the grinding process, the change in the grinding speed can only be performed several times, such as four, three, two or only once. Adjustments can be made near the beginning of the polishing process, in the middle of the polishing process, or near the end of the polishing process.

The grinding is continued after the grinding speed is adjusted (eg, after time T0), and the optical monitoring system continues to collect the spectra and determines the index values for the various regions of the respective substrates. Once the index trajectory of the reference region reaches the target index (for example, by fitting a new linear function to a series of index values after time T0 and determining the time at which the new linear function reaches the target index), the end point is determined and the two are terminated. The grinding operation of the substrates. The reference area for determining the end point may be the same reference area used to calculate the predicted end time as described above, or a different area (or if all areas are adjusted as described in Figure 8, the reference area may be selected to reach the end point) The purpose of the decision).

In some embodiments, for example, for copper polishing, after detecting the end of the substrate, the substrate is immediately subjected to an over-grinding process, such as to remove copper residues. Excessive grinding of all areas of the substrate under uniform pressure, for example, from 1 psi to 1.5 psi. The overgrinding process can have a preset duration, such as 10 seconds to 15 seconds.

In some embodiments, the polishing of the substrate is not stopped at the same time. In such embodiments, a reference area for each substrate may be present for the purpose of endpoint determination. Once the index trajectory of the reference area of the specific substrate reaches the target index (for example, by fitting the time when the linear function of one of the series of index values after time T0 reaches the target index), the end point of the specific substrate is determined and simultaneously stopped. Pressure is applied to all areas of a particular substrate. However, one or more other substrates can be continuously ground. Cleaning of the polishing pad begins based on the reference area of the remaining substrate only after the end of all remaining substrates is identified (or after all substrates have been over-ground). In addition, all of the carrier heads can simultaneously lift the substrate away from the polishing pad.

In the case of generating a plurality of index trajectories of a specific area and a substrate (for example, an index trajectory of each database interested in a specific area and a substrate), one of the index trajectories may be selected for the specific area and the substrate. The end point or pressure control algorithm. For example, by generating the same region and each index trajectory of the substrate, the controller 190 can fit the linear function of the index value of the index trajectory and determine the fitting goodness of the linear function to a series of index values. The resulting index trajectory with lines can be selected as the index trajectory for a particular region and substrate, and the line has the best fit to its own index value. For example, when deciding how to adjust the grinding speed of the adjustable region (eg, at time T0), a linear function with the best fit can be used in the calculation. As another example, an endpoint may be identified when a computed index with a best fit to the line (calculated from a linear function fitted to a series of index values) matches or exceeds the target index. Similarly, the index value itself and the target index can be compared to determine the end point, rather than calculating the index value from a linear function.

Determining whether the index trajectory associated with the spectral database has a best fit to the linear function associated with the database may include: the magnitude of the difference between the associated solid line and the index trajectory associated with another database To determine whether the amount of difference between the index trajectory of the relevant spectral database and the associated solid line is relatively minimal, such as the lowest standard deviation, the maximum correlation, or other measurements of variance. In one embodiment, the fit is determined by calculating the sum of the squared differences between the index data points and the linear function; the database with the smallest sum of squared differences has the best fit.

In some embodiments, a new library of reference spectra can be automatically triggered. Controller 190 can store a threshold value that will be used to decide whether to trigger the generation of a new database. For example, the threshold can be set by the operator prior to grinding, or the threshold can be a pre-programmed value.

During milling, for example, a series of measurement spectra of a substrate (eg, various regions of the substrate) is obtained using techniques as described above. The measurement spectra can be used to control the polishing rate (as described above in Figures 8 through 10) and/or to control the endpoint. For each measurement spectrum from the series, for example, the technique described above is used to select the reference spectrum from the pre-existing database of the reference spectrum as the best matched spectrum, thus producing a series of best matched spectra.

The controller can be programmed to calculate a fitted measure of a series of best-matched spectra versus a series of measured spectra. Compare the fitted metrics to the critical values. If the fitted metric exceeds a critical value (eg, above or below, depending on whether the low or high value of the fit metric indicates a better fit, respectively), then a new database of reference spectra is generated.

In some embodiments, the fit metric is the fit of the linear function to the index trajectory, for example, other measures of the standard deviation, correlation, or variance between the linear function of the measured spectrum and a series of index values. For example, the fit metric can be the sum of the squared differences between the index data points and the linear function. In some embodiments, the fit metric is based on a cumulative difference between a series of measured spectra and a series of best matched reference spectra. For example, the fit metric can be determined by calculating the difference between the reference spectra of the best matching spectra of the respective measured spectra and the measured spectra and calculating the sum of the differences. Calculating the difference between the measured spectrum and the best matched spectrum may include calculating a difference as a total of absolute values of the intensity differences over the range of wavelengths. that is:

Difference=Σ ^B _λ=A abs(I _measured (λ)-I _reference (λ))

Wherein, A and B are the lower and upper limits of the spectral wavelength range, respectively, and Imeasured (λ) and Ireference (λ) are respectively measured spectral intensity and reference spectral intensity of a given wavelength. Alternatively, calculating the difference between the measured spectrum and the best matched spectrum may include calculating a difference as the total number of intensity squared differences in the wavelength range. that is:

Difference=Σ ^B _λ=A (I _measured (λ)-I _reference (λ)) ²

In some embodiments, the fit metric is based on measuring the difference between the endpoint time TE of the substrate and the average endpoint time of the previously measured substrate.

If a new library of reference spectra is triggered, a set of spectra from the substrate becomes a new library of reference spectra. This condition may use a technique similar to that described above for generating a reference library from a batch of first substrates. In short, an index value is generated for each reference spectrum and associated therewith. The index value can be proportional to (eg, equal to) the number of times the platform is rotated, and each reference spectrum is measured from the substrate as the platform rotates. In addition, new target index values can be automatically generated. A new target index value can be calculated from the number of platform rotations, and the end of one or a previous substrate is detected at the rotation of the platform, such as one or more substrates, rather than triggering a substrate that creates a new database. For example, a new target index value can be calculated from the average of the number of platform rotations (eg, a weighted average), and two or more consecutive substrates before triggering the substrate that generates the new database are detected at the platform rotation. end. Alternatively, if the polishing system includes another endpoint detection system 180 (in addition to the spectral optical monitoring system 160), the number of platform rotations at which the other endpoint detection system 180 detects the endpoint may be set to the target index value. In some embodiments, a first substrate post-grinding thickness measurement can be performed, and an initial target index value determined by one of the above techniques can be adjusted by, for example, linear scaling, such as using a target thickness. The ratio is multiplied by the post-grinding measurement thickness.

New repositories can be added to existing groups of one or more repositories, or old repositories can be replaced.

Referring to Figure 11, a simplified flow chart 600 is illustrated. As described above, a plurality of regions of a plurality of substrates are simultaneously polished in a polishing apparatus having the same polishing pad (step 602). During the lapping operation, each region of each substrate has a polishing speed that is controllable by independent variable grinding parameters independent of other substrates, such as pressure applied to a particular region by a chamber in the carrier head above a particular region. During the lapping operation, the substrate is monitored (e.g., as described above) with measurement spectra obtained from various regions of each substrate (step 604). A reference spectrum that best matches is determined (step 606). The index values of the most fitted reference spectra are determined to produce a series of index values (step 608). For each region of each substrate, a linear function is fitted to a series of index values (step 610). In one embodiment, the linear interpolation of the linear function is used, for example, to determine the expected end time of the linear function of the reference region to reach the target index value (step 612). In other embodiments, the predicted endpoint time may be predetermined or calculated as a combination of predicted endpoint times for multiple regions. If necessary, adjusting the polishing parameters of other regions of the other substrate to adjust the polishing speed of the substrate, so that a plurality of regions of the plurality of substrates reach the target thickness at approximately the same time, or a plurality of regions of the plurality of substrates have approximation at the target time The same thickness (or target thickness) (step 614). Continuous grinding after adjusting the parameters, and for each region of each substrate, continuously measuring the spectrum during the time period after adjusting the grinding parameters, determining the best matching reference spectrum from the database, and determining the index value of the best matching spectrum A new series of index values is generated and a linear function is fitted to the index values (step 616). Once the index value of the reference region (eg, the computed index value resulting from a linear function fitted to the new series index value) reaches the target index, the grinding may be stopped (step 630).

The techniques described above can also be applied to monitoring metal layers using eddy current systems. In this case, the layer thickness (or its representative value) is directly measured by the eddy current monitoring system instead of performing spectral matching, and the layer thickness is used instead of the index value for calculation.

The method used to adjust the end point can vary based on the type of grinding performed. For copper block grinding, a single eddy current monitoring system can be used. For copper-clear CMP with multiple wafers on a single platform, a single eddy current monitoring system can be used first, allowing all substrates to achieve a first breakthrough at the same time. The eddy current monitoring system can then be switched to the laser monitoring system to remove and over-grind the wafer. An optical monitoring system can be used for barriers with multiple wafers and dielectric CMP on a single platform.

In some embodiments where the grinding system includes another endpoint detection system (other than the spectroscopy system), the technique described above can be used to adjust the zone pressure, but the other endpoint detection system can be used to detect the actual endpoint. . For example, for copper grinding, this condition allows the spectral monitoring system to reduce residue and over-grinding, but allows another system that can more reliably determine the end of grinding (eg, motor torque sensor or friction-based sensor) Determine the end point of the grinding.

As used herein, the term substrate may include, for example, a product substrate (eg, a product substrate including a plurality of memory or processor dies), a test substrate, a bare substrate, and a grating substrate. The substrate can be in various stages of integrated circuit fabrication, for example, the substrate can be a bare wafer, or the substrate can include one or more deposited and/or patterned layers. The term substrate can include circular disks and rectangular sheets.

The controller 190 may include a central processing unit (CPU) 192, a memory 194, and a support circuit 196 such as an input/output circuit, a power supply, a clock circuit, a cache memory, and the like. In addition to receiving signals from optical monitoring system 160 (and any other endpoint detection system 180), controller 190 can be coupled to polishing apparatus 100 to control grinding parameters, such as one or more platforms and various rotational speeds of the carrier head. And one or more pressures applied by the carrier head. The memory is connected to the CPU 192. The memory or computable readable medium can be one or more removable memories, such as random access memory (RAM), read only memory (ROM), floppy disk, hard Disc or other form of digital storage. Additionally, although illustrated as a single computer, controller 190 can be, for example, a distributed system including a plurality of independently operating processors and memory.

Embodiments of the invention and all functional operations described in this specification can be in digital electronic circuitry, or in computer software, firmware or hardware (including structural components and structural equivalents thereof disclosed in this specification) Or implemented in a combination thereof. Embodiments of the invention may be implemented as one or more computer program products (ie, one or more computer programs tangibly embodied in a machine readable storage medium) for execution by a data processing device or for control operations A data processing device, such as a programmable processor, a computer, or a plurality of processors or computers. Computer programs (also known as programs, software, software applications or code) can be written in any form of programming language, including compiling or interpreting languages; and computer programs can be deployed in any form, including as separate programs or as modules, components Subnormal or other unit suitable for the computing environment. The computer program does not necessarily correspond to the file. The program may be stored in a portion of the file in which the other program or data is stored, in a dedicated single file of the program or in multiple coordinated files (for example, a file storing one or more modules, subprograms or portions of code) in. A computer program can be deployed to be implemented or distributed in multiple places on one computer or on multiple computers and interconnected by a communication network.

The process and logic flow described in this specification can be performed by one or more programmable processors that execute one or more computer programs to perform functions by computing input data and generating output. The process and logic flow can also be performed by dedicated logic circuits and the device can also be implemented as dedicated logic circuits, such as field programmable gate arrays (FPGAs) or application-specific integrated circuits (application-specific Integrated circuit; ASIC).

The grinding apparatus and method described above can be applied to a variety of grinding systems. The polishing pad or carrier head, or both, can be moved to provide relative motion between the abrasive surface and the substrate. For example, the platform can loop rather than rotate. The polishing pad can be a circular (or some other shape) pad that is secured to the platform. Some aspects of the endpoint detection system can be applied to linear abrasive systems, for example, where the polishing pad is a linearly moving continuous or reel-to-reel belt. The abrasive layer can be a standard (eg, polyurethane with or without filler) abrasive material, a soft material, or a fixed abrasive material. The term relative positioning is used; it should be understood that the abrasive surface and substrate can be held in a vertical or some other direction.

Specific embodiments of the invention have been described. Other embodiments are within the scope of the following patent claims.

10. . . Substrate

10a. . . First substrate

10b. . . Second substrate

100. . . Grinding equipment

108. . . Windows

110. . . Abrasive pad

112. . . Outer abrasive layer

114. . . Back layer

118. . . Windows

120. . . platform

121. . . motor

124. . . axis

125. . . axis

128. . . Groove

129. . . Rotary coupler

130. . . arm

132. . . Slurry

140. . . Carrier head

142. . . Locating ring

144. . . film

146a. . . Chamber

146b. . . Chamber

146c. . . Chamber

148a. . . region

148b. . . region

148c. . . region

150. . . Support structure / rotary rack

152. . . Drive shaft

155. . . axis

154. . . motor

162. . . light source

160. . . Optical monitoring system

166. . . Circuit

164. . . Light detector

170. . . Bifurcated fiber

168. . . Optical head

174. . . Branch

172. . . Trunk

180. . . End point detection system

176. . . Branch

192. . . Central processing unit

190. . . Controller

196. . . Support circuit

194. . . Memory

201a. . . point

201. . . position

201c. . . point

201b. . . point

201e. . . point

201d. . . point

201g. . . point

201f. . . point

201i. . . point

201h. . . point

201k. . . point

201j. . . point

204. . . arrow

202. . . position

212. . . Index value

210. . . Index track / first series

S1. . . First slope

222. . . Index value

214. . . First line

S2. . . Second slope

220. . . Index track / second series

232. . . Index value

S3. . . Third slope

224. . . Second line

242. . . Index value

230. . . Index track / third series

S4. . . Fourth slope

310. . . database

234. . . Third line

330. . . Index value

240. . . Index track / fourth series

350. . . database

602. . . step

244. . . Fourth line

606. . . step

300. . . Measurement spectrum

610. . . step

320. . . Reference spectrum

614. . . step

340. . . recording

630. . . step

600. . . flow chart

SD1. . . Desired slope

604. . . step

SD3. . . Desired slope

608. . . step

TE’. . . Estimated end time

612. . . step

616. . . step

SD. . . Desired slope

SD2. . . Desired slope

SD4. . . Desired slope

IT. . . Target index

IT2. . . Second target index

IT4. . . Fourth target index

TE. . . End time

Pold. . . pressure

Pold2. . . pressure

Pold4. . . pressure

Pnew2. . . New pressure

Pnew4. . . New pressure

EI2/EI3/EI4. . . index

D2. . . Grinding thickness

T2. . . time

T4. . . time

S. . . Slope

IT1. . . First target index

IT3. . . Third target index

T0. . . time

Pnew. . . New pressure

Pold1. . . pressure

Pold3. . . pressure

Pnew1. . . New pressure

Pnew3. . . New pressure

Ts. . . time

D1. . . Grinding thickness

T1. . . time

T3. . . time

T. . . time

Figure 1 is a schematic cross-sectional view showing an example of a grinding apparatus having two polishing heads;

Figure 2 illustrates a top plan view of a substrate having a plurality of regions;

Figure 3A illustrates a top view of the polishing pad and shows the location of the in situ measurement on the first substrate;

Figure 3B illustrates a top view of the polishing pad and shows the location for in situ measurement on the second substrate;

Figure 4 illustrates the measured spectra of the in-situ optical monitoring system;

Figure 5 illustrates a library of reference spectra;

Figure 6 illustrates the index trajectory;

Figure 7 illustrates a plurality of index traces for different regions of different substrates;

Figure 8 illustrates the calculation of the plurality of desired slopes of the plurality of adjustable regions based on the time at which the index trajectory of the reference region reaches the target index;

Figure 9 illustrates the calculation of the plurality of desired slopes of the plurality of adjustable regions based on the time at which the index trajectory of the reference region reaches the target index;

Figure 10 illustrates a plurality of index trajectories for different regions of different substrates, wherein different regions have different target indices;

Figure 11 is a flow diagram of an exemplary process for adjusting the polishing rate of a plurality of regions in a plurality of substrates such that the plurality of regions have approximately the same thickness at the target time.

Similar component symbols and designations in the various figures indicate similar components.

600. . . flow chart

602. . . step

604. . . step

606. . . step

608. . . step

610. . . step

612. . . step

614. . . step

616. . . step

630. . . step

Claims (42)

A computer implemented method for generating a reference spectrum, the computer implementation method comprising the steps of: grinding a first substrate in a polishing apparatus having a rotatable platform; and measuring from the first in situ monitoring system during the grinding a first series of spectra of the first substrate; associating each of the spectra in the first series of spectra with an index value equal to one of the number of revolutions of the platform, wherein the respective spectra are measured at the rotation of the platforms Storing the first series of spectra as a reference spectrum; receiving data from a second in situ monitoring system, wherein the second in situ monitoring system monitors the first substrate during the first substrate grinding; based on the second in situ Monitoring the data of the system, detecting a polishing end point of the first substrate during the polishing of the first substrate, and not using the first in-situ monitoring system to detect the polishing end point of the first substrate; When the second in-situ monitoring system detects the polishing end point of the first substrate, determining a target index value based on the number of rotations of the platform; grinding the first in the grinding device Substrate; during the polishing of the second substrate, measuring the second series of spectra from the second substrate using the first in-situ monitoring system and without using the second in-situ detection system; and based on the reference spectrum The second series of spectra and the target index And performing at least one of determining a polishing end point for the second substrate or adjusting a polishing parameter for at least one region on the second substrate. The method of claim 1, wherein the step of determining the target index value comprises the step of combining a plurality of endpoint times, and the target index value is the number of times the platform is rotated when the plurality of endpoints are combined. The method of claim 1, further comprising the step of performing a post-grinding thickness measurement of one of the first substrates. The method of claim 3, further comprising the steps of: determining an initial index value, and adjusting the initial index value based on the post-grind thickness measurement. The method of claim 1, further comprising the steps of: determining, for each of the measurement spectra in the second series of spectra, a best matching reference spectrum from the reference spectrum to produce a series of best matching reference spectra; A series of index values are generated by: determining, for a reference spectrum from each of the series of best matching reference spectra, an index value associated with the best matching reference spectrum to generate the series of indices Value; and fitting a linear function to the series of index values. The method of claim 5, wherein the following steps are performed for each of the regions in the second substrate: measuring a second series of spectra; determining a best matching reference spectrum from the reference spectra; generating a series of index values; And fitting a linear function to the series of index values. The method of claim 6, further comprising the step of: determining an estimated time based on the linear function, wherein at the estimated time, at least one of the second substrates will reach the target index value; and adjusting the One of the at least one region on the substrate grinds the parameter to adjust the polishing rate of the at least one region such that at the predicted time, the at least one region is closer to the target index value than in the absence of such adjustment. The method of claim 7, further comprising the step of: detecting the polishing end point of the second substrate based on a time, the time being the time when the linear function of the reference region of the at least one region reaches the target index value . The method of claim 1, wherein the second in-situ monitoring system comprises one or more of a motor torque monitoring system, an eddy current monitoring system, a friction monitoring system, or a monochrome optical monitoring system. A computer program product for generating a reference spectrum, the computer program Having instructions to cause a processor to perform the following steps: receiving a first series of spectra from a first substrate during a grinding of the first substrate; the first series of spectra Each of the spectra is associated with an index value equal to a number of revolutions of the platform of the platform, wherein the respective spectra are measured at the platform; the first series of spectra is stored as a reference spectrum; Receiving data by the in-situ monitoring system, wherein the second in-situ monitoring system monitors the first substrate during the first substrate grinding; detecting the grinding during the first substrate based on the data from the second in-situ monitoring system An end point of the first substrate, and the first in-situ monitoring system is not used to detect the polishing end point of the first substrate; when the first substrate is detected by the second in-situ monitoring system At the end point, a target index value is determined based on the number of times the platform is rotated; during the grinding of the second substrate, the first in-situ monitoring system is used and the second in-situ detection system is not used to receive Measuring a second series of spectra from a second substrate; and based on the reference spectrum, the second series of spectra, and the target index value, performing a polishing end point on the second substrate or on the second substrate At least one of the regions adjusts at least one of the grinding parameters. The computer program product of claim 10, further comprising instructions to perform the following steps: for each of the measured spectra in the second series of spectra, The reference spectrum determines a best matching reference spectrum to produce a series of best matching reference spectra; the following steps are used to generate a series of index values: for each of the best matching reference spectra from the series of best matching reference spectra, Determining an index value associated with the best matching reference spectrum to generate the series of index values; and fitting a linear function to the series of index values The computer program product of claim 10, comprising instructions for performing, for each region in the second substrate, the step of: receiving a measurement of the second series of spectra; determining a best matching reference spectrum from the reference spectrum; Generating a series of index values; and fitting a linear function to the series of index values. The computer program product of claim 12, comprising instructions to: determine an estimated time based on the linear function, wherein at least one of the regions in the second substrate will reach the target index value at the estimated time And adjusting a grinding parameter of at least one region of the one substrate to adjust a grinding speed of the at least one region such that at the predicted time, the at least one region is closer to the target than in the absence of such adjustment Index value. The computer program product of claim 13, comprising instructions for performing the step of detecting the polishing end point of the second substrate based on a time that the linear function of the reference region of the at least one region reaches the target The time of the index value. A computer-implemented method of producing a reference spectrum, the method comprising the steps of: grinding a first substrate in a polishing apparatus; measuring a series of spectra from the first substrate by an in-situ optical monitoring system during grinding; Each of the series of spectra determines a best matching reference spectrum from a first plurality of reference spectra to produce a series of reference spectra; calculating a value of the series of spectra for one of the series of reference spectra; Comparing the value of the fitting metric to a threshold value, and determining whether to generate a second database based on the comparison; and if determining to generate the second database, storing the series of spectra as a second plurality Reference spectrum. The method of claim 15, further comprising the steps of: grinding a second substrate in the polishing apparatus; and measuring a second series of spectra from the second substrate by the in-situ optical monitoring system during grinding; as well as For each of the spectra in the second series of spectra, a reference spectrum of a best match is determined from a set of reference spectra including the second plurality of reference spectra to produce a second series of reference spectra. The method of claim 16, wherein the set of reference spectra comprises the first plurality of reference spectra. The method of claim 16, wherein the set of reference spectra does not include the first plurality of reference spectra. The method of claim 16, further comprising the step of associating each of the second plurality of reference spectra with an index value that is proportional to a number of times of the platform rotation, wherein the platform is rotated The spectrum is measured. The method of claim 19, further comprising the steps of: determining a correlation index value for each of the spectra in the second series of reference spectra to generate a series of index values; and fitting a linear function to the series of index values . The method of claim 20, further comprising the step of determining a target index value. The method of claim 21, wherein the plurality of second substrates are The region performs the following steps: measuring a second series of spectra; determining a best matching reference spectrum; determining a correlation index value; and fitting a linear function. The method of claim 22, further comprising the step of: determining an estimated time based on the linear function, at which the at least one region of the second substrate will reach the target index value; and adjusting the One of the at least one of the two substrates grinding parameters to adjust the polishing rate of the at least one region such that at the predicted time, the at least one region is closer to the target index than in the absence of such adjustment . The method of claim 21, further comprising the step of detecting a polishing end of the second substrate based on the linear function reaching one of the target index values. The method of claim 21, wherein the step of determining the target index value comprises the step of calculating the target index value using at least a first value, the first value being one of a number of platform rotations, wherein the platform is rotated Detecting an end point of a third substrate that is ground before the first substrate. The method of claim 25, wherein the step of calculating the target index value comprises the step of calculating a complex number including the first value and a second value One of the values is an average value, which is one of the number of rotations of the platform, wherein an end of a fourth substrate that is ground prior to the third substrate is detected where the platform is rotated. The method of claim 26, wherein the fourth substrate, the third substrate, and the first substrate are continuously polished. The method of claim 21, further comprising the steps of: monitoring the first substrate with a second in-situ monitoring system other than the in-situ optical monitoring system, and detecting by the second in-situ monitoring system One of the first substrates is grounded. The method of claim 28, wherein the step of determining the target index value comprises the step of determining the number of times the platform is rotated when the second in-situ monitoring system detects the polishing end of the first substrate. The method of claim 29, wherein the second in-situ monitoring system comprises one or more of a motor torque monitoring system, an eddy current monitoring system, a friction monitoring system, or a monochrome optical monitoring system. The method of claim 21, further comprising the step of performing a post-grinding thickness measurement of one of the first substrates. The method of claim 31, further comprising the steps of: determining An initial index value, and wherein the step of determining the target index value comprises the step of adjusting the initial index value based on the post-grind thickness measurement. The method of claim 15, further comprising the steps of: determining a correlation index value for each of the first series of reference spectra to generate a series of index values; and fitting a linear function to the series of index values . The method of claim 33, wherein the fitting metric comprises the linear function fitting a good fit to one of the series of index values. The method of claim 34, wherein the fit is a sum of one of a squared difference between the series of index values and the linear function. The method of claim 15, wherein the step of calculating the value of the fitting metric comprises the step of determining a difference between the spectrum and the best matching reference spectrum for each of the spectra in the series of spectra To provide a series of differences and accumulate the differences. The method of claim 36, wherein the step of determining a difference between the spectrum and the best matching reference spectrum comprises the step of: calculating a sum of absolute values of intensity differences over a range of wavelengths, or calculating The sum of the squared differences in intensity over a range of wavelengths. A computer program product for generating a reference spectrum, the computer program product having instructions to cause a processor to perform the following steps: receiving a series of in-situ optical monitoring systems from the first substrate during polishing of the first substrate Measure the spectrum; for each spectrum in the series of spectra, determine a best matching reference spectrum from a first plurality of first reference spectra; calculate a value of the series of spectra for one of the series of reference spectra And comparing the value of the fitting metric to a threshold value, and determining whether to generate a second database based on the comparison; and if determining to generate the second database, storing the series of spectra as a second plurality Reference spectra. The computer program product of claim 38, comprising instructions for performing the following steps: receiving, during the polishing of the second substrate, the second series of spectra from the second substrate during the grinding of the second substrate; and For each of the spectra in the second series of spectra, a reference spectrum of a best match is determined from a set of reference spectra comprising the second plurality of reference spectra to produce a second series of reference spectra. The computer program product of claim 39, wherein the set of reference spectra comprises the first plurality of reference spectra. The computer program product of claim 39, wherein the set of reference spectra does not include the first plurality of reference spectra. The computer program product of claim 39, wherein the step of calculating the value of the fit metric comprises instructions to perform the step of determining, for each of the spectra in the series of spectra, the reference spectrum of the spectrum that matches the best match A difference between them to provide a series of differences and to accumulate the differences.