TW201205703A - Dynamically or adaptively tracking spectrum features for endpoint detection - Google Patents

Abstract

A method of controlling polishing includes polishing a substrate and receiving an identification of a selected spectral feature, a wavelength range having a width, and a characteristic of the selected spectral feature to monitor during polishing. A sequence of spectra of light from the substrate is measured while the substrate is being polished. A sequence of values of the characteristic of the selected spectral feature is generated from the sequence of spectra. For at least some spectra from the sequence of spectra, a modified wavelength range is generated based on a position of the spectral feature within a previous wavelength range used for a previous spectrum in the sequence of spectra, the modified wavelength range is searched for the selected spectral feature, and a value of a characteristic of the selected spectral feature is determined.

Description

201205703 f. Description of the invention: [Technical field to which the invention pertains] The present disclosure relates to optical monitoring of a difficult lamp during a chemical mechanical polishing period of a substrate. [Prior Art] An integrated circuit is usually formed on a substrate by sequentially 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 conductive fill layer can be deposited over the patterned insulating layer to fill the trenches or holes in the insulating layer. After planarization, vias, plugs, and wires are formed in portions of the remaining conductive layer between the raised patterns of the insulating layer, and the vias, plugs, and wires provide a conductive path between the thin film circuits on the substrate. For other applications, such as oxide polishing, the filler layer is planarized until a predetermined thickness is left on the non-planar surface. In addition, photolithography generally requires planarizing the surface of the substrate. Chemical mechanical polishing (CMP) is an acceptable planarization method. This method of planarization typically requires the inclusion of a soil plate on the carrier head or the grinding head. The exposed surface of the substrate is typically placed against the red-turn polishing pad. The carrier head provides a controllable load on the substrate to push the substrate against the polishing pad. Abrasive abrasives are typically supplied to the surface of the polishing pad. 201205703 A problem with CMP is to determine if the polishing process is complete (ie, whether the substrate layer has been flattened to the desired flatness or thickness), or to remove the desired amount of material. (4) Distribution, grinding conditions, relative speed between the grinding plate and the substrate, and changes in the load on the substrate can all cause changes in the material removal rate. These changes and changes in the initial thickness of the substrate layer cause a change in the time required to reach the end of the polishing. ° Therefore, the end of the grinding can not be determined only as a function of the grinding time. * In some systems, the substrate is optically monitored in-situ during polishing, for example, via a window in a grinding raft. However, existing opticals may not meet the increased needs of semiconductor device manufacturers. Content] - Some optical final guessing systems track selected spectral features in spectral measurements to determine endpoints or to change a polishing rate. In a pupil, spectral features similar to the selected spectral features make it difficult to track the selected spectral features. Identifying the wavelength range of the optical endpoint detection system to search for the selected spectral signature allows the optical endpoint to correctly identify the selected spectral signature and use reduced processing resources. In some of the polishing processes, removing from a substrate—a second layer (eg, a barrier layer) of the second material (eg, nitride, eg, nitride or titanium nitride) is exposed to include A first layer or layer structure of a different first material (eg, a dielectric material, a low dielectric material, and/or a low dielectric cap material). It is often desirable to remove the first material until the remaining - target thickness. Tracking in Spectral Measurements - Selecting Frequency Offsets 201205703 Signature Characteristics to Determine Endpoints or Change Some Abrasive Rate Some optical endpoint detection techniques can be problematic in this grinding process because of the initial thickness of the second material Not known. However, if the spectral signature tracking is triggered by another monitoring technique (eg, motor torque, eddy current, or optical intensity), another monitoring technique can reliably detect the removal of the second material and the underlying layer structure. Exposure can avoid these problems. Additionally, the thickness of the layer or layer structure may vary between substrates. To increase the uniformity between the substrates of the final thickness of the layer or layer structure, the initial thickness of the layer or layer structure can be measured prior to grinding, and a target eigenvalue can be calculated from the initial thickness and the target thickness. You 1 grind a substrate due to sputum; and receive one of the selected spectral features to identify, have a wavelength range of one width, and one of the selected spectral features to monitor during the grinding. The series spectrum of light from the substrate is measured while the substrate is being ground. The u value of the characteristic of the chirp spectrum i is generated from the series of spectra. The step of generating includes the following steps: for at least some of the spectrum from the series of frequencies, based on the frequency level feature, one of the positions in the previous wavelength range is produced, and the tampering Searching for the selected spectrum within the wavelength range, the selected frequency is better than the one of the selected characteristics. One of the characteristics of one of the features is the one that precedes the series. First spectrum. Based on the series of values # $ grinding the end point or the & Ning Xi J value decides the grinding rate - at least one of the adjustment towels. The example of the Court can be 4 - fixed wide or multiple features. The wavelength range can have a step of generating the modified wavelength range. The method can include the following steps: 201205703. The fixed width is centered at the location of the characteristic in the previous wavelength range. The step of generating the modified wavelength range may include the steps of determining a position of the characteristic in the previous wavelength range and adjusting the wavelength range such that in the modified wavelength range, the characteristic is positioned closer to the modified wavelength range - at the center. The step of generating the modified wavelength range may include the steps of: determining, for the spectrum of the series of frequencies, a wavelength value of the selected spectral feature to generate a series of wavelength values; fitting one of the series of wavelength values Function; and according to the function. The calculation is for an expected wavelength value of the selected spectral characteristic for subsequent spectral measurements. This function can be a linear function. The step of generating the modified wavelength range can include the step of centering the width of the wavelength range to the expected wavelength value. The method can include the steps of fitting a function to the series of values and determining at least one of a grinding end point or a - grinding rate adjustment based on the function. Determining a polishing end point may include the steps of: calculating an initial value of the characteristic according to the function 'calculating the secret-current value according to the function, and calculating a difference between the initial value and the current value, and when When the difference reaches a target difference, the grinding is interrupted. This function can be a linear function. The selected spectral signature can include: - spectral peaks, - spectral valleys or - spectral zero crossings. This characteristic can include: a wavelength, a width, or a - intensity. The selected frequency read feature can include a spectral peak and the characteristic can include a peak width. The spectrum of visible light can be measured and the wavelength range can have a width between 5 〇 and 200 nm. In another aspect, a method of controlling polishing includes the steps of: 201205703 receiving a selection - a user input of a fixed wavelength range that is a subset of wavelengths measured by an in situ monitoring system; receiving one Identifying one of the selected spectral features and characteristic of the selected spectral feature for monitoring during polishing; grinding a substrate; for each spectrum in the series of spectra, measuring one of the light from the substrate while grinding the substrate a series of spectra; searching for the selected spectral characteristics of the fixed wavelength range of the respective spectra, determining a value of one of the characteristics of the selected spectral feature, generating a series of values, and determining a polishing endpoint or a polishing rate based on the series of values At least one of the adjustments. The embodiment may include one or more of the following features. The in-situ monitoring system can measure the intensity of at least the wavelength of visible light and the fixed wavelength range can have a width between 5G and 2GG nanometers. The selected spectral characteristics can be - spectral peaks, - spectral valleys, or - spectral zero crossings. This characteristic can be a wavelength, a width or an intensity. In another aspect, a method of controlling polishing includes: grinding a substrate having H; receiving - 敎 spectral characteristics - two and the selected (four) characteristic H for performing during polishing. The substrate simultaneously measures a series of light from the substrate: a frequency j 'determines the -th value of the characteristic of the feature at the time of the first layer exposure; and adds the -offset to the _th value, To generate - and monitor the characteristic of the feature, and to suspend grinding when the material of the (4) is determined to reach the second value. This embodiment may include one or more of the following features. Set, width or intensity. This feature can be - in the entire spectrum of the series, and the selected feature can be 201205703 continuous-evolving site, width or intensity. This feature can be a wave of bees or troughs of the spectrum. The substrate can include a ninth layer covering the first layer, and the step of grinding can include the steps of: grinding the second layer and detecting the first layer of exposure using an in situ monitoring system. The step of determining the 兮-value and detecting the exposure of the first layer may be one of the steps of monitoring the characteristic of the feature during the first-in-situ monitoring technique. Process. The step of detecting the exposure of the first layer can include the step of monitoring the total reflected intensity from the substrate. The step of monitoring the total reflected intensity can include the step of integrating the spectrum over a range of wavelengths for each of the spectra in the series to produce a 'total reflected intensity. The in-situ monitoring system can include a motor torque or friction monitoring system. The first value can be determined during the grinding of the first layer (e.g., immediately after the grinding of the first layer is initiated). The first layer can be exposed before the substrate: grinding begins. The step of monitoring the characteristic of the feature can include the step of: for each spectrum from the series of spectra, relying on the value of the characteristic, the value of > 4*, to produce a series of values. The characteristic of the feature can be determined to reach the second value by fitting a linear function to the series of values and determining that the linear function is equal to the -end time at the second value.

The thickness of the first layer can be received by -sV, and the offset can be calculated based on the thickness before the polishing. The step of calculating the offset value Δν may include the steps of: calculating .dT)/(dD/dv), where & is - target thickness, 1 is from the substrate - the first layer - the thickness before grinding, and D2 is From the thickness of the first layer of the mounting substrate - after polishing, and /dV is the rate of change of the thickness as a function of the characteristics of the sea. The calculation of the 201205703 step of calculating the offset value Δv may include the following steps. + (4) is called / (10) (four) + (9) such as / (four) wide), the towel is the thickness: the thickness before grinding, 〇 1 is the thickness from the first layer of one of the mounting substrates, and Δν0 is the mounting substrate A difference between the pre-polished thickness of the first layer and the post-grinding thickness at a value of the characteristic of the feature. The pre-polishing thickness d丨 can be measured at a separate measuring station as one of the characteristics The rate of change of the thickness of the function, #dD/dv, may be the rate of change of thickness near one of the ends of the polishing. The first layer may comprise polycrystalline germanium and/or dielectric material 'e.g., consisting of substantially pure polycrystalline germanium. , composed of a dielectric material or a combination of polycrystalline germanium and dielectric material Palladium cases may include one or more of the following advantages as needed. Identifying a wavelength and searching for selected spectral features allows for greater accuracy in detecting endpoints or determining-grinding rate variations, for example, the system is It is unlikely to be selected during spectral metrology—incorrect spectral features. Tracking spectral features in the f-wavelength range, rather than on the entire spectrum, allows for easier and faster identification of such spectral features. The processing resources required to select the spectral features may be reduced; a semiconductor manufacturer develops the time to detect the end of a particular product substrate. The spectral feature tracking can be applied to a grinding that begins with the grinding of a reflective layer. Operation, and can improve the thickness uniformity between wafers (wafer_t0_wafer thickness unif〇rmity; wtwu). The initial thickness of the layer can be measured after grinding, and can be calculated according to the initial thickness and the target thickness - target characteristics Values, thereby providing - more accurate: 201205703 endpoint decision. Details of one or more embodiments are set forth in the accompanying drawings and the description below. Other aspects, features, and advantages will be apparent from the description and drawings, and in accordance with the scope of the claims. [Embodiment] An optical monitoring technique is to measure the frequency of light reflected from a substrate during grinding and to identify from the library. A matching reference spectrum. One potential problem with the spectral matching method is that for some types of substrates, there are significant inter-substrate differences in the underlying grain features, resulting in a spectrum of substrate reflections from the surface having the same outer thickness. These changes increase the difficulty of proper spectral matching and reduce the reliability of optical monitoring. A technique to counteract this problem is to measure the spectrum of light reflected from the substrate being polished and to identify changes in spectral characteristics. The characteristic characteristics (such as the 'wavelength of the spectral peaks') allow for better uniformity of polishing between the substrates within the batch. By determining the target difference of the spectral characteristic, when the characteristic has changed the target amount, the endpoint can be called. a single dielectric layer, or

From Applied Materials, Inc.) the substrate can be stacked in a significantly more complex layer only for placement in the semiconductor layer. For example, the second layer on the first floor. 'or low dielectric value (l〇w-k) sealant, for example, Black DiamondTM (to or CoralTM (from Novellus 201205703)

SystemMne.). The second layer may be a barrier layer, and the composition of the barrier layer is the same as the first layer: the same layer. For example, the barrier layer can be a metal or metal nitride such as a nitride button or titanium nitride. If necessary, between the first layer and the second layer - or a plurality of additional layers, for example, a low dielectric value covering material, for example, tetraethoxy cerium tetraethyl 〇rth〇s; material. Both the first layer and the second layer are at least translucent. The first layer, together with one or more additional layers, if present, provides a layer stack below the second layer. In the absence of these, only a single layer containing polycrystalline spine and/or dielectric is polished, for example, although an outer layer may be present beneath the layer being ground. Chemical mechanical polishing can be used to planarize the substrate until the second layer is exposed. For example, if an opaque conductive material is present, the moon-transparent conductive material can be ground until the second layer (e.g., ' barrier layer) is exposed. Thereafter, the portion of the second layer remaining on the first layer is removed, and the substrate ' is ground until the first layer (e.g., dielectric layer) is exposed. In the meantime, it is desirable to grind the first layer (e.g., the dielectric layer) until the target thickness is left or the amount of target material has been removed. The method of polishing is to polish the conductive layer on the first polishing pad until at least the second layer (e.g., the barrier layer) is exposed. In addition, the second layer can be removed, for example, at the first polishing pad in an excessive grinding step. Thereafter, the substrate is transferred to the second abrasive crucible, wherein the first layer: e.g., the barrier layer is completely removed, and a portion of the thickness of the lower first layer (e.g., "electrical dielectric) is also removed. Alternatively, if 12 201205703 is in the additional layer or layers between the first layer and the second layer, it may be removed in the same polishing operation at the first polishing pad. However, the initial thickness of the second layer may not be known when the substrate is transferred to the second polishing pad. As noted above, this condition can cause problems for optical endpoint detection techniques. These optical endpoint detection techniques track selected spectral feature characteristics in spectral measurements to determine the endpoint at the target thickness. However, this problem can be alleviated by triggering spectral feature tracking by another monitoring technique that reliably detects the removal of the second layer and the exposure of the underlying layer or layer structure. Further, by measuring the initial thickness of the first layer and calculating the target characteristic value based on the initial thickness of the first layer and the target thickness, the inter-substrate uniformity of the thickness of the first layer can be improved. Spectral features may include spectral peaks, spectral troughs, spectral inflection points, or frequency zero crossings. Characteristics of the features may include wavelength, width or intensity. FIG. 1 illustrates a polishing apparatus 2 that is operable to polish the substrate 10. The grinding apparatus 2G includes a rotatable disc-shaped platform 24 on which the grinding crucible 3 is positioned. The platform is operable to rotate about the axis 25. For example, the motor can rotate the drive shaft 22 to rotate the platform 24. For example, the abrasive raft 3G can be detachably secured to the platform 可由 by an adhesive layer. The grinding pad 30 can be removed and replaced when worn. The polishing pad % can be a two-layer polishing pad having an outer polishing layer 32 and a softer backing layer 34. The optical access point 36 is provided through the polishing pad in a manner that includes an aperture (i.e., a hole through the aperture) or a solid window. The glazing may be fixed to the polishing pad. However, in some embodiments the solid window may be supported on the platform ’ and protrude into the aperture in the polishing crucible. The polishing pad 3 is usually placed on the 13 201205703 platform 24 such that the aperture or window covers the optical head 53 of the v, ten, and 24 recesses 20. The optical head 53 can be optically accessed through the aperture or the window because it can be taken seven times. For example, 'the window can be rigidly crystallized. A opaque material (for example, stone or glass), or a softer plastic material I I, for example, a silicone resin, a formate or a toothed polymer (for example, containing ^ Polymer)), or mention of the singularity of the horse. If the top surface of the solid window is rigid crystal or glass, the top surface should be recessed from the surface to prevent scratches. If the top surface is close and accessible to the surface, the top surface of the f should be a softer material. In some embodiments, the solid window is secured to the mat and is a polyurethane or a window having a combination of quartz and polyurethane. Window =, monochromatic light of a particular color (e.g., 'blue light or red light') may have a transmissive (four), for example, approximately 80% transmittance. The 9 window is sealed for the polishing pad σ so that the liquid does not leak through the interface of the window and the polishing pad π. In the case of a solid yoke, the window comprises a rigid crystal or glass poem covered by an outer layer of a softer plastic material. The top surface of the softer material can be ugly and flat. The bottom surface of the rigid material may be coplanar with the bottom surface of the polishing pad or recessed relative to the bottom surface of the polishing pad. In particular, a solid window can be integrated into the abrasive layer and the bottom layer can have an aperture aligned with the solid window. The bottom of the window && ' can be visually included with one or more grooves. The recess can be shaped to accommodate the end of the fiber such as, for example, V, for example, the end of the fiber optic cable or the end of the eddy current sensor 14 201205703. The recess allows the end of the fiber optic cable or the end of the eddy current sensor to be positioned at a distance from the surface of the substrate being ground that is less than the thickness of the window. In the case where the window comprises a rigid crystalline portion or a glassy portion, and the recess is formed by mechanical processing into the embodiment in this portion, the recess is ground to remove the trace caused by machining. Alternatively, a solvent and/or liquid polymer may be applied to the surface of the recess to remove traces caused by machining. The removal of scratches, usually caused by machining, reduces the scattering and allows the transmission of south light through the window. The backing layer 34 of the polishing pad can be attached to the polishing layer 32 outside of the polishing pad, for example, by an adhesive. The aperture providing the optical access point 36 can be formed in the pad 30 (e.g., by cutting or by modeling the pad 3 to include apertures), and the window can be inserted into the aperture and secured to the pad 3, for example With adhesive. Alternatively, the liquid precursor of the window can be dispensed into the apertures in the pad' and allowed to cure to form a window. Alternatively, the solid transparent material (e.g., the crystalline or glazed portion described above) can be positioned in the liquid coffin material and the liquid pad material can be cured to form the pad 30 around the transparent member. In either of the latter two conditions, a pad of material can be formed and the layer of the pad containing the modeling window can be cut from the block. The grinding apparatus 20 includes a combined slurry/flushing arm 39. During milling, the arm 39 is operable to dispense a slurry 38 containing a liquid and a pH adjuster. Alternatively, the grinding apparatus includes a slurry crucible that is operable to dispense the slurry onto the abrasive crucible 30. The grinding apparatus 20 includes a carrier head 7G that is operable to hold the substrate 1A against the polishing pad 3''. The carrier head 7G is suspended from the fulcrum structure 72 (e.g., the rotating material 15 201205703 frame) and is coupled to the carrier head rotation motor 76 by the carrier drive shaft 74 such that the carrier head is rotatable about the shaft 71. In addition, the carrier head can be laterally vibrated in a radial groove formed in the support structure 72. In operation, the platform rotates about the platform center axis 25 and the carrier head rotates about the carrier head central axis 71 and translates laterally across the top surface of the polishing pad. The grinding apparatus also includes an optical monitoring system that can determine the end of the grinding as follows. The optical monitoring system includes a light source 51 and a photodetector 52. The light from the light source 51 # passes through the optical access point 36 of the #3 wipe, collides and passes through the optical access point 36 to reflect back from the substrate 1 and travels to the photodetector 52. The bifurcated fiber optic cable 54 can be used to transmit light from the source 51 to the optical access 236 and back to the photodetector $2 from the optical access point 36. The split-fork fiber optic cable 54 can include a "trunk" 55 and two "spurs" 56 and 58. As mentioned above, the platform 24 includes a recess 26 in which the optical head 53 is positioned. The optical head 53 holds the mains of the bifurcated fiber electrical retardation 54. The end is a forked fiber electrical retarder 5 4 arranged to conduct light to the surface of the soil to be ground and to conduct light from the surface of the substrate to be polished. Optical " may include one or more lenses or windows covering the ends of the split-fork fiber cable 54. + soil, or, the optical head 53 can only hold the solid window adjacent to the solid window in the polishing pad <End of the main line 55. The optical head 5 can be removed from the recess 26 as needed to achieve preventive maintenance or corrective maintenance. The platform includes a removable in-situ monitoring module 50. The in-situ monitoring module 50 - or more The light source 51, the light detector 52, and the transmitting and receiving 16 to and from the light source 51 and the light detecting example, the output of the detector 52 can be an orbital) and transmitted to the optical monitoring system. Controlling the benefits of the loss (eg, the digital signal of the device. Similarly, the circuit of the signal of the device 52. The rotation through the drive shaft 22 can be transmitted from the controller to the die via the rotary coupler and the slave A; k惕, the control command in the digital signal of 5〇 turns the light source on or off. The in-situ monitoring module 50 can also hold the respective ends of the branch portions 56 and 58 of the split optical fiber 54. In transmission &, the light system is lightly conducted by the branch line S6 and is conducted out from the end of the trunk line 55 located in the optical head s3 and impinges on the substrate to be polished. The light reflected from the substrate is located in the optical head 53. Received at the end of the main line 55 and via The branch line 58 is conducted to the photodetector 52. In one embodiment, the bifurcated fiber cable 54 is a bundle of fibers. The bundle includes a first set of fibers and a second set of fibers. The fibers in the first set are connected to Light from the source 5 is conducted to the surface of the substrate being polished. The fibers in the first group are connected to receive light reflected from the surface of the substrate being polished, and the received light is conducted to the photodetector 52. The optical fibers are such that the fibers in the second group form an X-shape that is centered on the longitudinal axis of the bifurcated fiber 54 (when viewed in the cross-section of the split fiber optic 54). Alternatively, other arrangements can be implemented. For example, the fibers in the second group can form a v-shaped shape that mirrors each other. A suitable bifurcated fiber "T is purchased from Verity Instruments, Inc. of Carrollton, Texas. 17 201205703 There is usually an optimum distance between the pad window and the end of the main line 55 of the bifurcated fiber cable 54 closest to the window of the polishing pad. This distance can be determined empirically and subject to (for example) window reflexivity, self-forking Fiber cable The shape of the beam being beamed and the distance from the substrate being monitored. In one embodiment, the fiber-optic cable is positioned such that the end closest to the window is as close as possible to the bottom of the window without actually contacting the In the case of this embodiment, the grinding apparatus 2 can include a mechanism (e.g., as part of the optical head 53) that is operable to adjust the end of the bifurcated fiber optic cable 54 and the bottom surface of the polishing pad window Alternatively, the closest end of the split fiber cable 54 is embedded in the window. The light source 51 4 is operable to emit white light. In one embodiment, the emitted white light comprises 200-800 nm. Wavelength of light. Suitable light source for xenon or xenon mercury lamps. The photodetector 52 can be a spectrometer. A spectrometer is basically an optical instrument for measuring the properties (e.g., intensity) of light over a portion of the electromagnetic spectrum. A suitable spectrometer is a grating spectrometer. The typical output of 1770 is the intensity of light. The intensity of the light is a function of wavelength. The light source 5 1 and the optical debt detector 5 2 are connected to 5 - Γ 4 〇 L # counter money to an operable computing device to control the operation of the light source 5 1 and the photodetector 5 2 : ^ ^ JW 'and receive The signal of the light source 5 and the photodetector 52. The computing device can include a microprocessor positioned adjacent to the grinding device, such as a personal detachment knee. With regard to control, the computing device can, for example, synchronize the rotation of the starter 10 of the light source 51. As shown in Figure 2, the computer can make a series of flashes from the 灾 灾 尤 尤 原 , 51, which is just before the substrate 10 passes over the in-situ monitoring module 50, and just before the base fe 1G crosses the original The bit monitoring module 5 () ends. Each of the points 2 〇 1 211 indicates that the light from the in-situ monitoring module 50 impinges on the substrate 10 and is reflected from the substrate 1 。. Alternatively, the computer can cause the light source 51 to continuously emit light that begins just before the substrate 10 passes over the in-situ monitoring module 50 and ends just after the substrate 1 has passed the in-situ monitoring module 50. The series of spectra is provided during the grinding process, for example, from the spectrum obtained by continuous pulsing of the inductors on the substrate from the platform. In some embodiments, light source 51 emits a series of flashes of light onto a plurality of sections/knife of substrate 10. For example, the light source can emit a flash of light onto portions of the substrate 10 and portions other than the substrate 1G. τ is received by light detector 52 from light reflected from substrate 10 to determine a plurality of series of spectra from portions of substrate 1G. These features can be identified in the frequency 4 in which each feature is associated with a portion of the substrate 10. For example, features can be used to determine the endpoint conditions for the polishing of substrate 10. In some embodiments, monitoring of portions of substrate 10 allows for varying the polishing rate on one or more portions of substrate 10. The receiving signal' computing device can receive, for example, a signal carrying information describing the spectrum of light received by the optical detection benefit 52. Figure 3A ® shows an example of a spectrum measured from the amount of light emitted from a single flash from a source. The spectrum 3〇2 is measured based on the light reflected from the product substrate. The spectrum 304 is measured from light reflected from a substrate 矽 substrate which is a wafer having only a ruthenium layer. The spectrum 3〇6 is from 19 201205703 The light received by the optical head 存在 in the case where there is a substrate positioned on the optical head 53. Under this condition (referred to as dark conditions in this specification), the received light is typically ambient light. The device can process the signal or a portion of the signal to determine the end of the grinding step. Without being limited to any particular theory, the spectrum of light reflected from the substrate ίο evolves as the grinding progresses. Figure 3b provides an example of the evolution of the spectrum as it progresses to the film of interest. Different spectral lines represent different points in the polishing process. As can be seen, as the thickness of the film changes, the nature of the spectrum of the reflected light changes, and the inter-spectral spectrum is exhibited by the thickness of the film. When the polishing of the film is performed, the peak in the spectrum of the reflected light is observed (that is, the local maximum illusion, the height of the peak generally changes, and as the material is removed, the peak tends to: widen. In addition to widening, The wavelength at which a particular peak is located typically increases with the development. In some embodiments, the wavelength at which a particular peak is located = the hang decreases as the milling proceeds. For example, the peak 310(1) is shown at a specific time during the period. The peak in the spectrum, while the peak 310(7) is not at the same peak after the grinding period. The peak is at a longer wavelength and is wider than the peak 310(1). According to the empirical formula 'use the wavelength and/or width of the peak Relative change (for example, measure the 丫::_ and / or width, or both to determine the degree of grinding at the intermediate height between the centroid and the nearest trough at the distance: 佳波峰 (or more) The peaks vary according to the pattern of the two materials and their materials. 20 201205703 In the second example, the change of the peak wavelength can be used to determine the political point. For example, when the peak wavelength and the current wavelength of the peak between When the target difference is reached, the polishing apparatus 20 can stop the polishing base. The characteristics other than the peaks can be used to determine the difference in wavelength of the light reflected from the substrate 1(). For example, the wave detector can be monitored by the photodetector 52. The wavelength, the inflection point or the X-axis or the y-axis wear distance, and when the wavelength has changed by a predetermined amount, the grinding apparatus 2 can stop grinding the substrate 1 〇. In the second embodiment, in addition to the wavelength, the monitored characteristic can be For the width or intensity of the feature, the wavelength may not be monitored. The feature may be offset by a sequence of approximately 40 nm to 120 nm, although other offsets are also possible. For example, the upper limit may be much larger, especially in dielectric In the case of quality grinding, Fig. 4A provides a spectrum 400a of the amount of light reflected from the substrate 10. The optical dry monitoring system can pass the spectrum 400a through a high-pass filter to reduce the overall slope of the spectrum, thereby producing the 4B. In the figure, the frequency is shown in Fig. 400b. For example, during processing of multiple substrates in a batch, there may be a large spectral difference between the wafers. A high pass filter can be used to normalize the frequency 4 ' To reduce the substrate on the same batch Spectral variation. An exemplary high «wave H can have a cutoff frequency of (10) 5 Hz and a filter order. 4. The pass filter is not only used to help filter out the sensitivity to the underlying changes but also to flatten "Legal signal to make feature tracking easier. In order for the user to select which feature of the end point will be tracked to determine the ', point, a contour map can be generated and displayed to the user. Figure 5B provides a basis for grinding A plurality of frequencies 21 during the period of reflection of the light reflected from the substrate 1 201205703 - the example of the contour line generated in Figure 5B, and the 5A diagram provides the measurement spectrum from the specific transient in the contour map Example. Contours Figure 5_ includes features such as peak regions 5() 2 and trough regions 504 generated by correlation peaks 502 and troughs 5G4 on spectrum 5〇〇a. As time passes, the substrate is polished and the light reflected from the substrate changes, as illustrated by the change in spectral characteristics in the contour map Figure 5b. A test substrate can be ground to produce a contour map 500b', and light reflected from the test substrate can be measured by photodetector 52 during polishing to produce a series of spectra of light reflected from substrate 10. The series of spectrum can be stored, for example, in a computer system that can be an optical monitoring system as needed. Grinding of the mounting substrate can begin at time η and continue beyond the estimated endpoint time. When the polishing of the test substrate is completed, the computer (e.g., on the computer screen) presents a contour plot to the operator of the polishing apparatus 20. In some embodiments, for example, 'by assigning red to a higher intensity value in the spectrum, blue ϋ to the lower intensity value in the spectrum, and intermediate color (orange to green) to the intermediate intensity in the spectrum Value, computer color mark contour map. In other embodiments, the darkest shades of gray are assigned to the lower intensity values in the spectrum and the brightest shades of gray are assigned to the higher intensity values in the spectrum, and the intermediate shadows are assigned to the inter-matrix in the spectrum. The intensity value causes the computer to produce a grayscale contour map. Alternatively, the computer can generate a three-high line® with a maximum ζ value representing the higher intensity value in the spectrum' and a minimum ζ value representing the lower intensity value in the spectrum, with 22 201205703: 'Z value for spectrum The middle value in the middle. For example, the three-dimensional contour line .° is displayed by heart color, gray scale or black and white. In some embodiments, the grinding apparatus 20 is smashed. The % can be interacted with the 3D contour map to observe the different characteristics of the spectrum. ^ 'The contour of the reflected light produced by the monitoring of the test substrate during the grinding process. Figure 5嶋, may contain spectral features such as peaks, troughs, spectral zero crossing points, and & Features may have characteristics such as wavelength, width, and/or degree. As shown by the contour map 500b, when the top surface of the polishing pad 3 〇 1 is moved (four), the light reflected from the wire can be changed by the time of Ik, so that the characteristic characteristics change with the passage of time. Before the grinding of the device substrate, the operator of the grinding device 2 can observe the isometric drawing 500b and the characteristic characteristics of the substrate to have a batch of substrates similar to the mounting substrate. Tracking during processing. For example, the operator of the grinding apparatus 20 can select the wavelength of the companion 506 to track. The contour map 500b (especially JL is a color ώ 435 ^ 々〇 which is a color mark or a three-dimensional contour map) has the advantage that such a graphical display user can more easily select the 掐δ feature due to features (for example, and Eight features that have a linear change in Pic time) are visually distinguishable. · In order to select the final material, the thickness of the layer on the m-type substrate D]n can be calculated by linear interpolation based on the thickness of the pre-polished thickness of the (4) substrate. Before the eight-grinding (for example, the time at which the grinding starts, Τι & can be measured), and after the grinding (for example, the thickness of the substrate of the 201205703 type after the thickness of the substrate is 2), and The value of the characteristic can be measured at the time D of the target thickness D. T, which can be calculated from 1 (T2 T1)*(D2-D')/(D2-D1), and the value of the characteristic v, can be based on the spectrum measured at time T|. The target difference δν of the characteristic of the selected feature (such as a particular change in the wavelength of the peak 5G6) can be determined according to V, _VU, where V1 is the initial characteristic value (at time T1). Since &, the target difference sv may be a change in the value V' of the characteristic from the initial tee value VI before the grinding at the time port to the time τ at which the grinding is expected to be completed. The player of the grinding apparatus 侑20 can wheel the target difference 604 (e.g., a λ rt: i 5V) of the characteristic characteristic to be changed into the computer associated with the grinding apparatus 20. In order to determine the value of v and to determine the value of point 6〇2, the data line can be used to estimate the data. Line 5〇8. The value of line 508 at time | | can be subtracted from 6 〇 2 at T1. The value of the final value 508 to determine the point can be based on the correlation between the target difference of the characteristic characteristics and the amount of material removed, and the substrate shift _ to select characteristics such as the spectral peak 506. The grinding equipment 2 can be selected to have different characteristics and/or characteristic characteristics to find out the characteristics of the characteristic difference between the target and the self-installed substrate. In other embodiments, the endpoint decision logic determines the criteria to be pursued and the endpoint. 5曰The grinding of the steering device substrate, Figure 6A shows the characteristic characteristics of the tracking during the grinding, the difference 602a of the soil plate 10 Exemplary graph of -d 24 201205703 0〇〇a. Substrate 10 may select a characteristic characteristic (such as a peak or trough wavelength) of an operator of the grinding apparatus 20 that is being ground in a batch of substrates to be installed according to The contour of the substrate is tracked in Figure 5〇〇b. When the substrate is polished, the photodetector 52 measures the spectrum of the light reflected from the substrate. The end point determines the spectrum of the logical dimming to determine the series of characteristic characteristics. The material is removed from the surface of the substrate 10, and the value of the selected characteristic can be changed. The difference between the series of characteristic characteristics and the initial value VI of the characteristic characteristic is used to determine the difference 6〇2a d. When the substrate 10 is polished, the end point The decision logic may determine the current value of the feature being tracked. In some embodiments, when the current value of the feature has been: The initial value changes target 1 604, the endpoint may be called. In some embodiments' (eg Using a robust line fit, the line 606 is fitted to the difference 602a_d. The grind end time can be determined based on the difference 6〇2a_d to determine the function of the line 6〇6. In some embodiments, the function is time-paired. The linear function of the difference. When calculating the new difference, the function of line 606 (e.g., slope and intercept) may vary during the grinding of substrate 1 。. In some embodiments 'line 606 reaches the target difference of 6 〇 4 The time provides an estimated end time _ ° when the function of line 606 changes to accommodate the new difference, the estimated end time 680 can vary. In some embodiments, the function of line 606 is used to determine the material removed from substrate 10. Quantity, and use the change in the current value determined by the function to determine when the target difference is reached and when the end point needs to be called. Line 6〇6 tracks the amount of material removed. Or, when removing a specific thickness from the substrate 1〇 When you are able to The change of the current value determined by the function is used to determine the target of the removal of the top surface of the substrate from the 25 201205703 substrate, and the operator can set the target difference... + and then, The characteristic nanometer is selected. For example, the wavelength of the selected peak can be selected to determine how much material is removed from the top layer of the substrate 1 and when the end point is called at time T1, on the substrate. Before the grinding of 1〇, the characteristic value difference of the selected features is 0. When the grinding «3G starts to grind the substrate _, the characteristic value of the identified feature can be grounded from the top surface of the substrate 10 with the material from the °' For example, the wavelength of the selected characteristic during 'during grinding' may be a relatively low or low wavelength. Excluding the noise effect, the wavelength of the feature (and therefore the difference in wavelength) tends to change monotonically and often changes linearly. At time ,·, the endpoint decision logic determines that the identified characteristic has changed the target difference δν and can be called to the end point. For example, when the wavelength of the feature has changed the target difference by 50 nm, the end point is called and the polishing 塾3 〇 stops grinding the substrate 10. When processing a batch of substrates, the optical monitoring system 50 can, for example, track the same spectral features on all of the substrates. The spectral features can be associated with the same grain features on the substrate. Based on the underlying changes in the substrate, the starting wavelength of the spectral features can vary between substrates in the batch. In some embodiments, in order to minimize variability on a plurality of substrates, the endpoint decision logic may be used when a function that selects a characteristic property value or a value that fits to the characteristic property changes the endpoint metric (rather than the target deviation) Call the end point. The endpoint decision logic can use the expected initial value EIV determined by the mounting substrate. At time Τ1 identifying the characteristic characteristics being tracked on the substrate 10, the endpoint decision logic determines the actual initial value AIV of the substrate being processed. Endpoint Decision Logic 26 201205703 Use the initial value weight IVW to reduce the actual initial value to the end point decision, and consider the change in the substrate on a batch. For example, the variation of the slab may include the thickness of the substrate or the thickness of the underlying structure. Initial value weight 2 can be correlated with substrate variations to increase uniformity between processes between substrates. For example, the endpoint metric can be determined by multiplying the initial value weight by the 之间 between the actual initial value and the expected initial value, plus the target difference, for example, EM = IVW * (AIV - EIV) + 5V. In some embodiments, a weighted combination is used to determine the endpoint. For example, the endpoint decision logic may calculate the initial value of the property based on the function and calculate the current value of the property based on the function and calculate the first: difference between the initial value and the current value. The endpoint decision logic may calculate a first difference between the initial value and the target value and generate a weighted combination of the first difference and the second difference. Fig. 6B is an exemplary graph 特性 of the characteristic amount difference versus time taken at the two portions of the substrate 1 (). For example, the 'optical monitoring system 5' can track the features positioned toward the edge portion of the substrate 1G and the center toward the substrate 10; >&& 0 τ. Another feature of the 疋 position is to determine how much material has been removed from the substrate. When testing the mounting substrate, the grinding device 2 can, for example, identify two features corresponding to different portions of the mounting substrate for tracking. In some examples, the spectral characteristics correspond to the grain characteristics of the type on the plate. In the ': spectral features associated with different types of grain features on the mounting substrate. When the substrate ίο is ground, - the beauty of the eve-den~ is measured from the spectrum of the two columns of light corresponding to the substrate. The two characteristics can be determined by the endpoint decision logic: 27 201205703 Pre-characteristic value series values. The first difference 61 〇 a-b of the first portion of the substrate 10 can be calculated by calculating the feature as it proceeds during the grinding time. The base 612 ".S - the characteristic characteristic of the S - part of the series - the second difference ° ° difference 610a-b fits the first line 614 and can be anthropomorphic to the second difference 6l2a-b The first time and the second function 610 may be determined according to the first function and the second function, respectively, to determine the adjustment of the final polishing or the polishing rate 620 of the substrate 10. I Λ 第一 The first function of the first part of the board 1 ,, and with the second function of the first 邛 基板 八 八 基板 基板 基板 在 在 基板 基板 基板 第二 第二 第二 第二 第二 第二 第二 第二 622 622 622 622 〜 〜 622 622 622 The first part of the right substrate is different from the estimated end time of the first/minute of the substrate (for example, the first part will reach the target thickness at the second step), then the polishing rate 620 can be adjusted. Thus, the first function and the second function will have the same end time 18. In some embodiments, the polishing rate of the first and second portions of the substrate is adjusted such that the end point is reached simultaneously at both portions. Alternatively, it can be adjusted The polishing rate of the first part or the second part. For example, The grinding rate is adjusted by increasing or decreasing the pressure in the corresponding region of the carrier head 70. The change in the grinding rate can be assumed to be proportional to the change in pressure 'for example' a simple Prestonian model For example, when the first region of the substrate 10 protrudes at the time 以 to reach the target thickness, and the system has established the target time Η, the carrier head pressure in the corresponding region of time Τ3 can be multiplied by ΤΤ/ΤΑ. To provide load bearing head pressure after 28 201205703 time T3. In addition, 彳 develop a control model for grinding the substrate, which takes into account the influence of the platform or head rotation speed, the second-order effect of different head pressure combinations, the grinding temperature, the slurry The flow or other parameters affecting the rate of polishing. The subsequent time during the polishing process 'if appropriate, the rate can be adjusted again. In some embodiments, the computing device uses a range of wavelengths to easily identify the substrate from the device 1 The selected frequency in the spectrum measured by the light. The computing device searches for the selected spectral features in the wavelength range to distinguish the selected ones. The spectral features are, for example, similar in intensity, width or wavelength to other spectral features of the selected spectral features in the measured spectrum. Figure 7 illustrates the spectrum based on the light received by photodetector 52. A case of 700a. The spectrum 7〇〇a includes selected spectral features, such as spectral peaks. The selected spectral features 7〇2 can be selected by the endpoint decision logic to be tracked during the CMP of the substrate 10. Selected spectral features 7 02 The characteristic 704 (eg, wavelength) can be logically identified by the endpoint decision. When the characteristic 704 has changed the target difference, the endpoint determines the logical recall endpoint. In some embodiments, the endpoint decision logic determines the wavelength range 7〇6 to be in the wavelength range. Search for the selected spectral feature 7〇2. The wavelength range 7〇6 can have a width between about 50 and about 200 nm. In some embodiments, the 'wavelength range 706 is predetermined, for example, by an operator (eg, by receiving user input of a selected wavelength range), or as a process parameter for a batch of substrates (by making the wavelength range from The memory associated with the batch of substrates retrieves the wavelength range). In some embodiments, the wavelength range 201205703 is based on historical data', e.g., the average or maximum distance between successive spectral measurements. In some embodiments, the wavelength range 7 〇 6 is based on information about the test substrate, for example, twice the target difference δν. The seventh diagram is an example of the spectrum 700b measured from the amount of light received by the photodetector 52. For example, after the spectrum 700a is taken during the rotation of the platform 24, the spectrum 700b is measured. In some embodiments, the endpoint decision logic determines the value of the characteristic 704 in the previous spectrum 700a (e.g., 520 nm) and adjusts the wavelength range 7〇6 such that the center of the wavelength range 7〇8 is positioned closer to the characteristic 704. In some embodiments, the endpoint decision logic uses a function of line 606 to determine the expected current value of characteristic 704. For example, the endpoint decision logic may use the current grind time to determine the expected difference and determine the expected current value of the characteristic 7〇4 by adding the expected difference to the initial value V1 of the characteristic 704. The endpoint decision logic centers the wavelength range 7〇8 on the expected current value of the characteristic 7〇4. Fig. 7C is another example of the spectrum 7 〇〇 C measured from the amount of light received by the photodetector 52. For example, after the spectrum 700a is taken up during the rotation of the platform 24, the spectrum 7〇〇c is measured. In some embodiments the 'end point decision logic uses the previous value of the characteristic 7G4 for the center of the wavelength range 7 1 〇. 5 For example, the endpoint decision logic determines the average average between the values of the characteristics 704 determined during the two sequential transmissions of the optical head below the substrate 1 , and the decision logic can set the width of the wavelength range 710 to be flat. Two times twice. In some embodiments, the endpoint decision logic uses the variance between the values of the characteristics 7〇4 when determining the width of 201205703. In some embodiments, the width of the wavelength range 706 is the same for all spectra. For example, the width of the 'wavelength range 7〇6, the wavelength range· and the wavelength range @7H) (4). In some implementations, the wavelength range is different. For example, when the estimated characteristic 7〇4 is changed from the first characteristic of the characteristic to the second measurement, the width of the wavelength range 7〇8 is 6〇N. Tian estimates. The ten characteristic 704 has a wavelength range 708 of 8 nanometers from the previous measurement of the characteristic change of 5 nm, and the wavelength range of 8 nanometers is larger than the wavelength range having a small characteristic change of the vehicle. In some embodiments, the wavelength range 706 is the same for all spectral measurements during the grinding of the substrate 1〇. For example, 'the wavelength range is 7〇6^475 nm to 555 nm, and for all spectral measurements made during the polishing of the substrate, the endpoint decision logic searches for wavelengths between 475 nm and 555 nm. It is also possible to select other spectral ranges for the spectral features 7G2. The wavelength range 706 can be selected by the user input as a subset of the full spectrum range measured by the in situ monitoring system. In some embodiments, the endpoint decision logic searches for selected spectral features 702 in the modified wavelength range of the spectral measurements and in the wavelength range of the previous spectrum used in the remainder of the _. For example, the final logic 1 searches for the selected spectral feature 702 in the wavelength range 706 of the spectrum measured during the first rotation of the platform 24 and the wavelength range 708 of the spectrum measured during the sequential rotation of the platform 24. The two measurements of Lishi 2-7 A are performed in the first region of the substrate H). Continuing with the example, the endpoint decision 31 201205703 logic in the same platform rotation period 710, searching for the wavelength range of the two spectra of the other ^, ] 1 to search for another selected spectral feature, 1 10 is different from the first region of the first region _... The two measurements are performed in the substrate ~ tooth - Ue domain. In some embodiments, the spectral feature 7〇2 is a frequency-level 七 seven-phase zero crossing point. In a consistent embodiment of the method, the characteristic 704 is the intensity or width of the peak or wave (for example, measured at a fixed distance below the peak ^ w , anti ', or The width measured at the middle of the peak between the peak and the nearest trough 97). Figure 8 illustrates a method 800 for selecting a target difference M 乂 for use in determining the ··, , and point of the polishing process. Measurement and I, Yu Xing production. The nature of the substrate of the same pattern of σ substrates (step 802). The substrate to be measured is referred to as a "mounting" substrate in this specification. The mounting substrate can be simply a substrate that is similar or identical to the product substrate, or the mounting substrate can be a substrate from a batch of product substrates. The properties of the measurement can include the pre-grind thickness of the film of interest at a particular site of interest on the substrate. Typically, the thickness at multiple sites is measured. Sites are typically selected to measure the same type of grain characteristics for each site. Measurements can be performed at the measurement station. Prior to grinding, the in situ optical monitoring system measures the spectrum of light reflected from the substrate. The mounting substrate is ground according to the grinding step of interest, and the frequency stipulated during the grinding is collected (step 804). Grinding and spectrum collection can be performed at the above grinding equipment. The spectrum is collected by the in situ monitoring system during grinding. Excessive grinding of the substrate (i.e., grinding beyond the estimated end point) results in a spectrum of light reflected from the substrate when the target thickness is achieved. 32 201205703 Measure the properties of the overgrinded substrate (step 806). The properties include the post-grinding thickness of the thin layer of interest at the particular site or at the site for measurement prior to grinding. The measured thickness and the collected spectrum are selected (by examining the collected spectrum) for specific features (such as peaks or troughs) for viewing during the polishing (step 808). The features may be selected by the operator of the grinding apparatus, or the selection of features may be automatic (e.g., based on a conventional peak-finding algorithm and a Peak_SeleCti〇n formula). By way of example, as described above with reference to Figure 5B, a contour map 5〇〇b can be presented to the operator of the grinding apparatus 2, and the operator can select features from the contour map 5_ for tracking. It is only necessary to consider the features in the region if it is expected that the region contains features that are expected to be monitored (eg, from past experience, or based on theoretical characteristics). A feature is typically selected that exhibits a correlation between the amount of material removed from the top of the mounting substrate as the substrate is being polished. Linear interpolation & can be performed using *measured pre-t-thick film thickness and post-grinding substrate thickness to determine the approximate time to achieve the target film thickness. This approximate time can be compared to the spectral contour map to determine the endpoint value of the selected characteristic. The difference between the end point value of the characteristic characteristic and the initial value can be used as the target difference. In some embodiments, a function is fitted to the value of the characteristic characteristic to normalize the value of the characteristic characteristic. The difference between the endpoint value of the function and the initial value of the function can be used as the target difference. The same features are monitored during the grinding of the remaining substrates of the batch of substrates. 33 201205703 As needed to 'process the spectrum to improve accuracy, the spectrum can be processed' to normalize the spectrum to two, for example, and to or from the spectrum. In the case where the pass filter is applied to the spectrum, in a case of a uniform, the low collar is reduced to reduce or eliminate sudden spikes. Two: with = select the logic for the specific end point decision, the frequency to be monitored: < passed on the computer device to achieve the target thickness by applying points based. The endpoint decision logic uses: : The target difference in the feature to determine when the endpoint should be called. When the grinding begins, the change in characteristics can be measured relative to the initial characteristic value of the feature. Or 'except the target difference δν' can be used to call the end point 'end point logic can multiply the difference between the actual initial value and the initial value of the page' by the initial value MV, To compensate for the underlying changes between the substrates. For example, when you want to EM=SV "AIV-Talk", the end point determines 'can end; in the second embodiment, the weighted combination is used to determine the end point. For example, the end point decision logic can be based on The function calculates the initial value of the characteristic, and calculates the current value of the characteristic according to the function, and calculates the difference between the initial value and the current value. The end point decision logic can calculate the difference between the initial value and the target value' and generate the first difference Weighted combination with the second difference. The end point can be called when the weighted value reaches the target value. The end point decision logic can determine the time when the difference (or difference) of the rutting characteristics is monitored and the target difference. Call the end point. If the monitored difference matches the target difference or exceeds the target blue' then the end point is called. In one embodiment, the 'monitoring difference must match the target of the 201205703 standard deviation or exceed the target difference for a certain period of time (eg, The platform is rotated twice) to call the endpoint. Figure 9 illustrates a method 901 for selecting a target value for a characteristic associated with a particular target thickness and a selected spectral feature of a particular endpoint decision logic. The substrate is measured and ground (step 903), as described in step 8〇2_8〇6. In detail, the spectrum is collected and the time of each collected spectrum is stored and measured. Grinding rate of the apparatus (step 9〇5) The average grinding rate pR can be calculated by using the pre-grinding thickness D1 and the post-grinding thickness D2 and the actual grinding time PT, for example, PR=(D2-D1)/PT. The end time of the substrate is mounted (steps 9〇7) to provide a calibration point to determine the target value of the characteristic of the selected feature, as discussed below. Based on the calculated polishing rate PR, the pre-grinding start of the film of interest The endpoint ST is calculated from the thickness ST and the target thickness ττ of the film of interest of interest. Assuming that the polishing rate is constant throughout the polishing process, the final time α is calculated as a simple linear interpolation, for example, et = (st_tt) / pr. If necessary, the calculated end time can be estimated by grinding another substrate of the batch of patterned substrates, stopping the grinding at the calculated end time, and measuring the thickness f of the film of interest. If the thickness is at the target thickness Full circumference The calculated end time is satisfactory. Otherwise, the calculated end time can be recalculated. At the calculated end time, the target characteristic value of the selected feature is recorded from the spectrum collected from the self-installed substrate (step 9G9) If the parameter of interest 35 201205703 relates to the change of the location or width of the selected feature, then the information collected by the period before the calculated endpoint time can be used to determine the information. The initial value of the characteristic and the target The difference between the values is recorded as the difference in the characteristics of the feature. In some embodiments, a single target difference is recorded. Figure 00 illustrates a method 1000 for determining the end of a grinding step using peak-based endpoint decision logic. The other substrate in the batch of patterned substrates is ground using the above-described polishing apparatus (step 1002). The identification 'wavelength range of the selected spectral signature and the characteristics of the selected spectral signature are received (step 1004). For example, the endpoint decision logic is self-receiving recognition, and the computer has processing parameters for the substrate. In one embodiment, the processing parameters are based on information determined during processing of the mounting substrate. The substrate is initially polished' to measure the light reflected from the substrate to produce a spectrum, and the characteristic values of the selected spectral features are determined in the wavelength range of the measured spectrum. During each rotation of the platform, the following steps are performed. One or more spectra of light reflected from the surface of the substrate being ground are measured to obtain one or more current spectra of the current platform (steps 1 - 6). The one or more spectra measured by the current platform are processed as needed to improve accuracy and/or precision as described above with reference to Figure 8. If only one spectrum is measured, the one spectrum is used as the current spectrum. If more than one current spectrum is measured for platform rotation, the current spectrum is grouped, averaged within each group' and the average represents the current spectrum. The spectrum can be grouped by the radial distance from the center of the substrate. 36 201205703 For example, the first-current spectrum can be obtained from the spectrum measured from points 202 and 210 (Fig. 2), and the second current spectrum can be obtained from the spectrum measured from point 2〇3 and Xie. The third current spectrum is obtained from the spectrum measured from the point 2〇4& observation, and so on. The characteristic values of the selected spectral peaks for each current frequency error can be determined, and the grinding can be monitored separately in each region of the substrate. Or the worst case value of the characteristic of the selected spectral peak can be based on the current spectrum and can be used by the endpoint decision logic.额外 Additional _ or more spectra can be added to the series of spectra of the current substrate during each rotation of the platform. At least some of the spectra in the sequence differ when the milling is performed' due to material being removed from the substrate during milling. As described above with reference to Figures 7A-C, the repair wavelength range of the current spectrum is generated (step i008). For example, the endpoint logic determines the modified wavelength range of the current spectrum based on the prior eigenvalues. It is possible to modify the wavelength range U center to the U characteristic value. In some embodiments, the modified wavelength range is determined based on the expected characteristic value, e.g., the wavelength range center coincides with the expected characteristic value. In some embodiments, different methods are used to determine some of the wavelength ranges of the current spectrum. For example, by centering the wavelength range on the characteristic values from the previous spectrum measured in the same edge region of the substrate, the wavelength range of the spectrum measured from the amount of light reflected in the edge regions of the substrate. Continuing with the example, the wavelength range of the spectrum measured from the amount of light reflected in the central region of the substrate is determined by centering the wavelength range on the expected characteristic value of the central region. 37 201205703 In an embodiment, the wavelength range of the current spectrum is in some embodiments, the width of the wavelength range of the current spectrum: the different identification wavelength fc is used to search for the selected spectral characteristic and the polishing rate for the end point. The decision to change is more accurate, and the system is characterized during subsequent spectral measurements. In the wavelength range rather than the optional: incorrect frequency error, it is easier and faster to track the processing resources required for spectrum-specific spectral features. The salt identification identifies the current 1〇1〇 of the selected peak from the modified wavelength range, and uses the above content in Fig. 8 (step series to compare the current characteristic value* end point to determine the logical/target characteristic value (step 1012). For example, 'determine a series of values of the current characteristic characteristics according to the series, and fit the function to the series of values. For example, the function may be a linear function, which may be based on the current characteristic value and the initial characteristic value. Poor, to approximate the amount of material removed from the substrate during the grinding process. ^To determine the end point condition (the "no" branch of step 1014), the cutting continues, and step 1006 is repeated as appropriate. '1008 ' 1〇1〇> ιηΐΛ -η 012 and 1014. For example, the endpoint decision logic is based on (4) determining the target difference of the unresolved (4). In some embodiments, the equivalent is measured from multiple substrates. The frequency of the partially reflected light's time-of-day determination logic can determine the need to adjust the substrate- or eve. The grinding rate of the P-knife allows the grinding of multiple parts at the same time or near the same time. 38 20 1205703 When the endpoint decision logic determines that the endpoint condition has been met ("Yes" in the branch ι〇ΐ4), the endpoint is called and the grinding is stopped (step ι〇ΐ6). The spectrum can be normalized to remove or reduce unwanted Effects of light reflections. Light reflections from media other than one or more films of interest, including light reflections from the polishing pad window and the substrate layer from the substrate. The light spectrum received from the window is estimated in dark conditions (ie, when the substrate is placed on the in-situ monitoring system) to estimate the light reflection from the window. The spectrum of the light reflected by the bare germanium substrate can be measured, To estimate the light reflection from the enamel layer. These light reflections are usually obtained before the start of the grinding step. The original spectrum is normalized as follows: Normalized spectrum = (A-Dark) / (Si-Dark) where d is The original spectrum, moxibustion is the frequency obtained in dark conditions and is the spectrum obtained from the bare stone substrate. In the described embodiment, the change in the wavelength peak in the spectrum is used to perform the endpoint detection. generation Wave peaks or combined peaks use changes in the wavelength troughs (ie, local minima) in the spectrum. You can also use multiple peaks (or troughs) to change at the end of the detection. For example, individual peaks can be monitored individually, and The endpoint can be called when most of the peak changes meet the endpoint condition. In other embodiments, the endpoint or the spectral zero-father change can be used to determine endpoint detection. In some embodiments, the algorithmic setup process 1100 ( Figure 11) is followed by the use of a triggered feature tracking technique 丨 2 〇〇 (Fig. 2) of one or more substrates. 39 201205703 $初' (for example) using one of the above techniques The characteristics of the interest feature in the medium for the pursuit of the frequency of the boat 腓 η goods in the first layer of the grinding in the use of rumors can be rumored '(four) can ride the peak or trough, and the location of the wavelength or frequency of the valley or Width, or wave a / leather strength. If you have a different taste, you can apply it to the characteristics. The equipment manufacturer may pre-select the grinding rate dD/dt that is determined to be close to the grinding end point. (Step: For example, according to the grinding to be used for the grinding of the product substrate, the grinding time is different from the expected end grinding time) , the plurality of mounting substrates are ground, and the mounting substrate may have the same pattern as the product substrate. For each mounting substrate, the thickness of the pre-grinding layer before the grinding may be measured and the amount of removal may be calculated according to the difference. And storing the removal of the substrate and the associated grinding time to provide a data set. A linear function of the amount of removal can be fitted to the data set, the linear function of the removal being a function of time; The slope of the linear function provides the polishing rate. The algorithm mounting process includes measuring the initial thickness h of the first layer of the mounting substrate (step 1106). The mounting substrate can have the same pattern as the product substrate. The first layer can be a dielectric. For example, a low dielectric value material, for example, a carbon-defining two-gas fossil, for example, Black DiamondTM (from Applied Materials, Inc) or c〇ralTM (from N〇veUus Systems, Inc.) Depending on the composition of the first material, different from the first and second materials (eg, low dielectric value covering material, eg 201205703, for example, tetraethoxy decane (TE0S)) One or more additional layers of another material (eg, a dielectric material) (step 1107). The first layer provides a layer stack with the one or more additional layers. Next, in the first layer or layer Depositing a second layer (eg, a barrier layer) of a second material (eg, a 'nitride' such as tantalum nitride or titanium nitride) on the stack (step (4) 08). Additionally, on the second layer (and by A conductive layer is deposited in the trench provided by the pattern of the first layer, for example, a metal layer, for example, copper (step 1109). ° Grinding 'measurement at the measurement system other than the optical monitoring system used during the month 'for example' In-line or separate measurement stations, such as profilers or optical measurement stations using ellipsometers. For some measurement techniques (eg, a temple), the initial thickness of the first layer is measured before the second layer is deposited. 'But for other measurement techniques (eg, ellipses) For the circular polarimeter, the measurement can be performed before or after the deposition of the second layer. Thereafter, the first step is to use the first polishing station according to the grinding process substrate (step 〇) for the polishing process of interest. Polishing the pad to polish and remove the conductive layer and a portion of the second layer (step ui〇a). Thereafter, the second polishing pad can be used to polish and remove the second layer and a portion of the first layer at the second polishing station ( Step (1) 〇 b). However, it should be noted that for some A targets there is no conductive layer 5 ', for example, the second layer is the outermost layer at the beginning of the grinding. At least during the removal of the second layer 'and possibly The entire polishing operation at the second polishing station fq. /, J. 'Use the above technique to collect the spectrum (step 1112). In addition, the separation detection technique is used to detect the removal of the second layer and the exposure of the first layer of the 201205703 (steps) M4). For example, exposure of the first layer can be detected by a sudden change in the total torque of the motor torque or light reflected from the substrate. At the time when the second layer is cleared, the value of the characteristic of the feature of interest of the spectrum stored at time v is calculated. It is also possible to store the time T at which the detection was cleared. After the detection of the clearing, the grinding may be paused at a preset time (step 1118). The preset time is large enough for the grinding to pause after exposure to the first layer. The preset time is selected such that the thickness after grinding is sufficiently close to the target thickness that the polishing rate can be assumed to be linear between the thickness after grinding and the target thickness. At the time of the grinding pause, the value V2 of the characteristic of the feature of interest can be detected and stored (4), and the time T2 of the grinding pause can also be stored. For example, the post-polation thickness 02 of the first layer is measured using a measurement system that is the same as the initial thickness used to measure (step 1120). The preset target change of the value of the calculated characteristic is calculated (step 1122). This preset target change will be made in the algorithm. (4) η 丨 at the end of the production-substrate ΙΜ贞 愈 据 据 据 据 据 据 据 据 据 在 Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Ιΐ Near the end of the grinding operation, the change in thickness as a function of the monitored characteristics (4) dD/dv (the step function assumes that the peak is being monitored, the block... (for example, the rate of change at the wavelength of each E-peak can be expressed as ... long position offset, the number of angstroms of the removed material. As another example, the τ Π 〜 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The number of angstroms of the material, the offset of the frequency of the width of the guide, is shifted in one embodiment, the value at the end of the grindstone can be twisted, the exposure time of the layer, and the simple calculation as a function of time The value, for example, using data from the end of the grinding close to the mounting substrate (eg, the smaller of the last 25% or time τ^τ2), can fit the measured value as a function of time The slope of the line is provided as a function of time The rate of change dv/dt of the value. In any case, thereafter, by changing the rate of change of the polishing rate by the value, the rate of change of thickness _dD/dv as a function of the monitored characteristic is calculated, that is, dD /dV=(dD/dt)/(dV/dt). Once the rate of change dD/dv is calculated, the rate of change should be fixed for the product; it is not necessary to recalculate dD/dv for the same product in different batches. Once the mounting process has been completed, the product substrate can be ground. The initial thickness dl of the first layer of at least one substrate from the substrate of the batch product is measured as needed (step 12〇2). The product substrate has at least a mounting substrate The same layer structure, and if desired, has the same pattern as the mounting substrate. In some embodiments, each product substrate is not measured. For example, one substrate from a batch can be measured and the initial thickness is available For all other substrates from the batch. As another example, one substrate from the cartridge can be measured and the initial thickness can be used for all other substrates from the cartridge. In other embodiments, each product substrate is measured. Can 4 3 201205703 Performing a measurement of the thickness of the first layer of the product substrate before or after the installation process is completed. As described above, the first layer may be a dielectric, for example, a low dielectric material, for example, doped. Carbon dioxide, for example, Black Diam〇ndTM (from Applied Materials, Inc.) or c〇ralTM (from N〇veUus

Systems, Inc.). Measurements can be performed at the measurement system other than the optical monitoring system that can be used during grinding, for example, in-line or separate measurement stations, such as a profiler or optical measurement station using an ellipsometer. Depending on the composition of the first material, deposition on the first layer on the product substrate is different from the first and second materials (eg, a low dielectric value covering material, such as tetraethoxy decane (eg, tetraethoxy decane) One of the other materials or a plurality of additional layers (step 1203) the first layer together with the one or more additional layers provides a layer stack. Next, the first layer or layer stack on the product substrate Depositing a second layer of a different first material (eg, a nitride, such as a nitride button or titanium nitride), for example, a barrier layer (step 12〇4). Additionally, on the first layer of the product substrate (and in the trench provided by the pattern of the first layer) depositing a conductive layer, such as a metal layer, such as copper (step 〇2〇5). However, it should be noted that for some embodiments, there is no conductive layer For example, the second layer is the outermost layer at the beginning of the grinding. For some measurement techniques (eg, profilers), the initial thickness of the first layer is measured before the deposition of the layer, but for other measurement techniques (eg , ellipsometer, for example, can be deposited in the second Perform measurement before or after 2012 201203. The deposition of the second layer and the conductive layer may be performed before or after the installation process is completed. For each product substrate to be ground, the target characteristic difference is calculated based on the initial thickness of the first layer. Δν (step 1206). Typically, this occurs before the start of the grind, but the calculation may occur after the start of the grind but before the start of spectral feature tracking (in step 12). In detail, for example, autonomy The computer receives the initial thickness of the stored product substrate and the target thickness dT. In addition, it can receive the initial thickness and the end thickness D2, the change rate dD/dV of the thickness as a function of the monitored characteristic, and the value of the determined mounting substrate. Preset target change Δ. In one embodiment, the target characteristic difference Δν is calculated as follows: AV = AVD + (d1 - D1) / (dD / dV) + (D2 - dT) / (dD / dV) In some embodiments , the front thickness will not be available. In this case, "(d丨-D丨) / (dD / dV)" will be omitted from the above equation, that is, AV = AVD + (D2-dT) / (dD / dV Grinding the product substrate (step 1208). For example, A first polishing pad is used at the first polishing station to polish and remove the conductive layer and a portion of the second layer (step 1208a). Thereafter, the second polishing pad can be used to polish and remove the second polishing station. The second layer and a portion of the first layer (step 1208b). However, it should be noted that for some embodiments, there is no conductive layer 'for example, the second layer is the outermost layer at the beginning of the grinding. Using in situ monitoring. Detecting the removal of the second layer and the exposure of the first layer (step 12 10 ). For example, the mutation may be detected by a sudden change in the total intensity of the light reflected from the substrate or by the light reflected from the substrate. a left spit on the 45th of the 11 between the 0f 201205703 exposure. For example, Figure 13 illustrates a graph of the total intensity of light received from the substrate as a function of time during the grinding of the metal layer to expose the underlying barrier layer. This total intensity can be generated based on the spectral signal acquired by the spectrum monitoring system by integrating the spectral intensities over all wavelengths measured or over a preset wavelength range. Alternatively, the intensity at a particular monochromatic wavelength can be used instead of the total intensity. As shown in Figure 13, when the copper layer is removed, the total strength decreases, and when the barrier layer is completely exposed, the total strength is stable. The plateau of intensity can be detected and the plateau of intensity can be used as a trigger to initiate spectral feature tracking. Detection of at least the cleaning of the second layer begins (and possibly earlier, for example, from the beginning of polishing the product substrate using the second abrasive raft), and the spectrum is obtained during the grinding using the in-situ monitoring technique described above (step 1212). The above technique is used to analyze the spectrum to determine the value of the characteristic of the feature being traced. By way of example, Figure 14 illustrates a graph of the wavelength position of the frequency 1 peak as a function of time during polishing. Decide on the second layer of the debt test: the value of the characteristic of the feature being traced in the spectrum at t, v丨. The target value ντ of the different characteristic can now be counted (step 1214). The target value can be calculated by adding the target characteristic difference Δν to the value at the time from the first spread to the clearing of the first layer, that is, τ servant, ντ=ν 丨+ Δ▽. When characterizing the characteristics of the tracking), pause the grinding (step ϋ 'for each measurement spectrum (for example, in each platform rotation), determine the specific value of the tracking at each 6A value to generate The number of series (for example, the n(four)-column value fits the linear function between the functions). In some embodiments, the function can be fitted to the value in the window between 2012 and 201205. A pause is provided if the function satisfies the target value. The end time of the grinding. The value of the characteristic at the time t1 of the second layer can also be removed by fitting a function close to the column value (for example, a linear function), although by the \2: and The method illustrated in Figure 13 includes the deposition and removal of the second layer 'but for the 眚fy | two-pass target case, there is no second layer, for example, the first layer is the sharpest at the beginning of the grinding Attack the outer layer. For example, measure the initial thickness of the first layer before grinding And the process of calculating the target eigenvalue according to the initial thickness and the target thickness, which may be applicable with or without covering the second layer; the second layer is optional m omitting the deposition of the second layer And detecting the exposure of the first layer (4). The first layer may comprise a polycrystalline stone and/or a dielectric material, for example, consisting of substantially pure polycrystalline germanium, composed of a dielectric material or polycrystalline germanium. A combination of dielectric materials. The dielectric material can be an oxide (eg, oxidized oxide), or a combination of nitride (eg, 'nitrogen cut) or dielectric f. For example, 'measurement from-batch products An initial thickness of the first layer of at least one of the substrates (eg, as discussed for step 12-2). Calculating the target characteristic difference based on the initial thickness of the first layer (eg, as discussed for step 1206) Initiating the grinding of the first layer of the product substrate and obtaining the spectrum during the grinding of the first layer using the in situ monitoring technique described above. During the grinding of the first layer (eg, immediately after the first layer of grinding is initiated, or At startup The value of the characteristic V1 is measured shortly after the grinding of the first layer (for example, after a few seconds). Waiting for a few seconds can stabilize the signal from the monitoring system, so that the value of the value V丨 is more accurate 47 201205703 The target value of the computable characteristic r

τ (for example, as discussed for step 12U). For example, the J liver can be added to the characteristic value Vl, i.e., vT = v + AV, as the characteristic difference Δν. When the characteristics of the field break tracking feature reach the target value, the grinding is paused (for example, as discussed for the step coffee). This method valley (4) is divided to the target thickness while compensating for the inter-substrate variation of the absolute peak sites caused by the difference between the substrates in the underlying structure. There are many techniques for removing chowder from the series of values. Although the above describes the fitting of the line to the sequence, but also 广 a wide τ, , Α 疋 can also fit the nonlinear function to the sequence, or a low-pass median waver can be used to smooth the sequence (in this case, According to the value of the wave, it is directly compared with the target value to determine the end point). As used in this specification, the term substrate may include, for example, a product substrate (e.g., a product substrate including a plurality of memory or processor dies), a substrate-based substrate, and a gate substrate. The substrate can be at various stages of the fabrication of the integrated circuit, for example, the substrate can be a bare wafer, or the earth plate can include a deposition and/or patterned layer. The term substrate can include circular disks and rectangular sheets. The embodiments of the invention and all of the functional operations described in this specification can be implemented in digital electronic circuitry, or in computer software, firmware or hardware (including structural components and structural equivalents thereof disclosed in this specification). Medium or solid yoke in its combination'. Embodiments of the present invention may be implemented as one or a computer program product (ie, one of tangiblely implemented in an information carrier (eg, in a machine H readable/missing device or in a transmitted signal) The computer program is 'executed by a data processing device (for example, a stylized computer, a computer or a plurality of processors or computers), or a control device (for example, a programmable computer). Can be written by: 11, computer or multiple processors or translation or interpretation language): = design language (including software application or code), =1 also known as program, software, program or as Modules, components, and forms (including the use of the #__* as a stand-alone medium or suitable for deploying the computer program in a computing environment. The computer program is not required to be stored in the file. Program or resource λ sub; l privately used in a single file or: stored in multiple coordinating slots (for example, storage - or multiple modules, subprograms or partial code files) Executed on one computer or multiple computers at one location, or distributed across multiple locations and interconnected by a communication network. The processing and logic flows described in this specification can be performed by - or multiple computer programs Executing by one or more programmable processors to perform functions by operating on the input and generating outputs, or by dedicated logic circuits (eg, field programmable gate arrays; An FPGA or an application-specific integrated circuit (ASIC) is used to perform processing and logic flow, and the device can also be implemented as the dedicated logic circuit. The above grinding apparatus and method can be applied to various grinding systems. The polishing pad or carrier head or both can be moved to provide relative movement between the abrasive surface and the substrate. For example, the platform can be orbited rather than rotated. The polishing pad can be fixed to the circular shape of the platform. (or some other shape) pad. Some aspects of the endpoint detection system can be applied to linear grinding systems, 49 201205703 For example, 'the polishing pad is a linear moving continuous or reel-to-reel belt system. The abrasive layer can be Abrasive materials, soft materials, or fixed abrasive materials are standard (for example, polyurethane urethane with or without fillers.) The term relative positioning is used; it should be understood that the abrasive surface and substrate can be held in a vertical orientation or other orientation. Specific embodiments of the invention are described. Other embodiments are within the scope of the following claims, for example, may be performed in a different order The behavior described in the scope of application for patents, and still achieve the desired results. [Simplified description of the drawings] Figure 1 shows the chemical mechanical polishing equipment. Figure 2 is a top view of the polishing pad, and shows the position of the in-situ measurement. Figure 3A illustrates the spectrum obtained by in-situ measurement. Figure 3B illustrates the evolution of the spectrum obtained by in-situ measurement as the grinding progresses. Figure 4A illustrates the spectrum of light reflected from the substrate. An exemplary chart. Figure 4B illustrates a graph through Figure 4A of a high pass filter. Figure 5A illustrates the spectrum of light reflected from the substrate. Figure 5B illustrates the in situ measurement of light reflected from the substrate. Contour map of the spectrum. Fig. 6A illustrates an exemplary graph of the progress of the grinding, which is measured in terms of characteristic difference versus time. The figure illustrates an exemplary graph of the progress of the grinding, which is measured by the difference in characteristics versus time, wherein the characteristics of two different features are measured to adjust the polishing rate of the substrate. Figure 7A illustrates another spectrum of light obtained by in situ measurement. Figure 7B illustrates the spectrum of light obtained after the spectrum of Figure 7A. Fig. 7C is a diagram showing another frequency of light obtained after the spectrum of Fig. 7A. Fig. 8 illustrates a method of selecting a peak for monitoring. Figure 9 is a method for not obtaining the target parameters of the selected peak. The figure illustrates the method used for the endpoint decision. Figure 11 shows the method of setting the endpoint detection. Figure 12 illustrates a chart for another 13th dimension of the endpoint decision. A graph showing the total reflection intensity as a function of time during grinding. Figure 1 Spectrum of the wavelength of a function. [Main component symbol description] The same component symbols and tables in each drawing indicate the same component 51 201205703 10 Substrate 20 Grinding equipment 22 Drive shaft 24 Platform 25 Shaft/center shaft 26 Groove 30 Abrasive pad/pad 32 Outer grinding layer 34 Backing layer 36 Optical access point 38 Slurry 39 Slurry arm / Flushing arm / arm 50 In-situ monitoring module 51 Light source 52 Photodetector 53 Optical head 54 Bifurcated fiber optic cable / Bifurcated 55 Trunk fiber cable / Bifurcated fiber 58 branch line 56 branch line 71 shaft/center shaft 70. carrier head 74 drive shaft 72 support structure 201 point 76 carrier head rotation motor 203 point 202 point 205 point 204 point 207 point 206 point 209 point 208 point 211 point 210 point 304 Spectrum 3 02 Spectrum 310(1) Peak 306 Spectrum 400a Spectrum 310(2) Peak 500a Spectrum 400b Spectrum 502 Peak Area/Crest 500b Contour Figure 506 Crest 52 201205703 504 Valley Area/Valley 602b Difference 508 Line 602d Difference 602c Difference Value 700b Spectrum 700a Spectrum 702 Selected Spectrum Characteristics 700c Spectrum 706 Wavelength Range 704 Characteristics 710 Wavelength Range 708 Wavelength Range 802 Step 800 Method 806 Step 804 Step 901 Method 808 Step 905 Step 903 Step 909 Step 907 Step 1002 Step 1000 Method 1006 Step 1004 Step 1010 Step 1008 Step 1014 Step 1012 Step 1100 Algorithm Setup Process 1016 Step 1104 Step 1102 Step 1107 Step 1106 Step 1109 Step 1108 Step 1110a Step 1110 Step 1112 Step 1110b Step 1118 Step 1114 Step 1122 Step 1120 Step 1200 Trigger Feature Tracking Technique 53 201205703 1124 Step 1202 Step 1203 Step 1204 Step 1205 Step 1206 Step 1208 Step 1208a Step 1208b Step 1210 Step 1212 Step 1214 Step 1216 Step ΤΙ Time T2 Time Τ 'Time 5V Target difference tl Time 54

Claims (1)

201205703 - VII. Patent application scope: 1. A method for controlling grinding, comprising the steps of: grinding a substrate; receiving one of the selected spectral features to identify 'one wavelength range having a taste degree and one of the selected spectral characteristics To monitor during polishing; to measure the light from the substrate while grinding the substrate. The series of spectra from the series of spectra produces the characteristic ~-column value of the selected spectral feature, the step of generating The method includes the following steps: for the spectrum from the series; some spectrums, based on the spectral characteristics, generating a modified wavelength range at a position within a previous wavelength range, searching for the selected spectral feature within the modified wavelength range, and determining the selected spectrum a value-of-characteristic value for the previous spectrum of the series of spectra; and determining at least one of a polishing endpoint or an adjustment for a polishing rate based on the series of values. 2. Such as - month project! The method wherein the wavelength range has a fixed width. The method of claim 2, wherein the step of generating the modified wavelength range is the following step: centering the fixed width at the position of the characteristic in the previous wavelength range. The method of claim 1, wherein the step of generating the modified wavelength range comprises the steps of: determining a position of the one of the characteristics in the previous wavelength range and adjusting the wavelength range such that the modified wavelength range The characteristic is located closer to the center of one of the modified wavelength ranges. 5. The method of claim 1 wherein the step of generating the modified wavelength range comprises the step of: for at least some of the spectrum of the series, determining a wavelength value of the selected spectral signature to generate a series of wavelengths a value that fits a function of the series of wavelength values; and calculates an expected wavelength value for the selected spectral feature for a subsequent spectral measurement based on the function. 6. The method of claim 5, wherein the function is a linear function. 7. The method of claim 5 wherein the step of generating the modified wavelength range comprises the step of centering the width of the wavelength range to the expected wavelength value. 8. The method of claimant, further comprising the steps of: fitting a column value to a function' and determining at least one of a polishing endpoint or an adjustment to a polishing rate based on the function. 9 = The method of claim 8, wherein the determining - the step of grinding the end point is as follows: The following step: calculating an initial value of the characteristic according to the function, root ^ 56 201205703 The function calculates one of the current values of the characteristic 'and calculates the initial A difference between the value and the just value, and the grinding is interrupted when the difference reaches a target difference. 10. The method of claim 8, wherein the function is a linear function. 11_ The method of claim 1, wherein the selected spectral feature comprises: a spectral peak, a spectral trough, or a spectral zero crossing. 12. The method of claim 1, wherein the characteristic comprises: a wavelength, a width, or an intensity. The method of claim 12, wherein the selected spectral feature comprises - a spectral peak, and the characteristic Contains a peak width. The method of claim i, wherein the spectrum of visible light is measured, and the wavelength range has a width between 5 〇 and 2 〇〇 nanometers. 15. A method of controlling polishing comprising the steps of: receiving a user input selecting a fixed wavelength range that is a subset of wavelengths measured by an in situ monitoring system; receiving one of selected selected spectral features Identifying and characterizing one of the selected spectral features for monitoring during grinding; grinding a substrate; measuring a series of spectra of light from the substrate while grinding the substrate; 57 201205703 for each spectral wavelength in the series of spectra The selected spectral characteristic of the range - the value of the characteristic to be generated - determines at least one of the polishing endpoints based on the series of values. 'Searching for the fixedness of the respective spectra' and determining the value of the selected spectral feature; and or adjusting for one of the polishing rates. 16. wherein the in-situ monitoring system measures to , and the fixed wavelength range has a dielectric width. The method of claim 15 wherein the intensity of the wavelength of visible light is less than between 50 and 2 nanometers - 17. The method of claim 1 - spectral peak 15 wherein the selected spectral feature, Spectral troughs or – spectrum zero crossings. The method of claim 15, wherein the characteristic comprises a width or an intensity. 19. A method of controlling polishing comprising the steps of: grinding a substrate having a first layer; receiving a characteristic of one of the selected spectral features and characteristic of the selected spectral feature to monitor during grinding; Measuring a series of spectra of light from the substrate when the substrate is ground; determining a first value of the characteristic of the feature at a time of exposure of the first layer; adding an offset to the first value to Generating a second value; and 58 201205703 monitoring the characteristic of the feature and pausing the grinding when determining that the characteristic of the feature reaches the first value. The method of claim 9, wherein the characteristic comprises: a position, a width or an intensity. The method of claim 20, wherein the selected feature continues for an evolutionary site, width or intensity in the entirety of the series of spectra. 2. The method of claim 21, wherein the feature comprises a peak or trough of the spectrum. The method of claim 19, wherein the substrate comprises a second layer covering the first layer, wherein the step of grinding comprises the steps of: grinding the second layer, and further comprising the step of: using an in situ Surveillance Department (4) 'The exposure of the first layer. The method of claim 23, wherein the first value is determined at a time when the first in-situ monitoring technique is called the first layer of exposure. 25 = The method described in the spring item 23 wherein the exposure of the first layer is a process separate from the step of monitoring the characteristic of the feature. The method of claim 25, wherein the step of detecting the exposure of the first layer comprises the step of monitoring a total reflected intensity from the substrate. 27. The method of claim 25, wherein the step of monitoring the total reflected intensity comprises the step of integrating the spectrum over a range of wavelengths for each of the series of spectra to produce the total reflected intensity. The method of claim 25, wherein the in-situ monitoring system comprises a motor torque or friction monitoring system. The method of claim 19, wherein the first value is determined during the grinding of the first layer. The method of claim 29, wherein the first value is determined immediately after initiation of the grinding of the first layer. The method of claim 29, wherein the first layer is exposed at the beginning of the polishing of the substrate. The method of claim 19, wherein the step of monitoring the characteristic of the feature comprises the dream that for each spectrum from the series of spectra, the value/value' is determined to produce a series of values. The method of claim 32, wherein the feature is determined by fitting a linear function to the series of values and determining that the linear function is equal to an end time of the second value. This characteristic reaches this second value. The method of claim 19, further comprising the steps of: receiving a pre-polished thickness of the first layer, and calculating the offset value from the pre-polishing thickness. 35. The method of claim 34, wherein the step of calculating the offset value comprises the step of: calculating (D2-dT) / (dD/dV), wherein dT is - the thickness 'D' is from a package One of the first layers of the substrate is provided with a pre-polished thickness, 〇2 is the thickness of one of the first layers from a mounting substrate, and dD/dV is the rate of change of thickness as a function of the characteristic. The method of claim 34, wherein the step of calculating the offset value comprises the steps of: AV=AVD+(d1-Dl)/(dD/dV)+(D2-dT)/(dD/dV) The pre-polishing thickness, dT is a target thickness, ... is one of the first layers from the mounting substrate, and the thickness 'A before polishing is a post-grinding thickness of the first layer from a mounting substrate, Δν〇 is A difference between the pre-polished thickness of the first layer of the substrate and the post-polishing thickness at the value of the characteristic of the feature, and dD/dV is the rate of change of thickness as a function of the characteristic. The method of claim 36, further comprising the step of measuring the pre-grind thickness d i at a separate measurement station. 38. The method of claim 35, wherein dD/dV is a rate of change of thickness near the end of the grinding. The method of claim 19, wherein the first layer comprises polysilicon and/or a dielectric material. 40. The method of claim 39, wherein the first layer consists of substantially pure polycrystalline germanium. The method of claim 39, wherein the first layer is comprised of a dielectric material. The method of claim 39, wherein the first layer is a combination of polycrystalline germanium and one of a dielectric material. 62