Computer program product, method, and polishing system for endpoint control for inconsistent underlayer

Abstract

During polishing, the substrate is monitored using an in-situ monitoring system to generate a sequence of signal values. This sequence of signal values ​​from the region is converted into a sequence of effective thickness values ​​for that region, where each effective thickness value includes contributions from the outer layer and one or more bottom layers. A function is fitted to this sequence of effective thickness values, and the fitted function is used to determine the effective initial thickness value of the layer at the start of polishing. An adjusted target thickness value is calculated based on the initial target value, the initial thickness value, and the effective initial thickness value. The polishing endpoint is detected or polishing parameters are modified based on these effective thickness value sequences and the adjusted target thickness value.

Description

[Technical Field]

[0001] This disclosure relates to chemical mechanical polishing, and more specifically, to the monitoring of the conductive layer during chemical mechanical polishing. [Previous Technology]

[0002] Integrated circuits are typically formed on a substrate by sequentially depositing conductive, semiconductive, or insulating layers on a silicon wafer. Various manufacturing processes require planarization of the layers on the substrate. For example, 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 metal layer can be deposited on a patterned insulating layer to fill trenches and holes in the insulating layer. After planarization, the remaining metal in the trenches and holes of the patterned layer forms vias, plugs, and lines to provide conductive paths between thin-film circuits on the substrate.

[0003] Chemical mechanical polishing (CMP) is a recognized planarization method. This planarization method typically requires mounting the substrate on a carrier head. The exposed surface of the substrate is usually placed on a rotating polishing pad. The carrier head provides a controlled load on the substrate to push the substrate against the polishing pad. A polishing slurry containing abrasive particles is typically supplied to the surface of the polishing pad.

[0004] One problem with CMP is determining whether the polishing process is complete, i.e., whether the substrate layer has been planarized to the desired flatness or thickness, or when the required amount of material has been removed. Variations in slurry composition, polishing pad conditions, the relative speed between the polishing pad and the substrate, the initial thickness of the substrate layer, and the load on the substrate can cause variations in the material removal rate. These variations lead to variations in the time required to reach the polishing endpoint. Therefore, determining the polishing endpoint solely as a function of polishing time can result in inhomogeneities within or between wafers.

[0005] In some systems, the substrate is monitored in situ during polishing, for example, using a polishing pad. One monitoring technique involves inducing eddy currents in the conductive layer and detecting changes in the eddy currents when the conductive layer is removed. [Summary of the Invention]

[0006] In one state, an initial thickness value of a conductive outer layer in a region on a substrate is received, and an initial target thickness value of the conductive outer layer in that region on the substrate is received. The initial thickness value is generated before polishing and represents the thickness of the outer layer in that region before polishing, and the initial target thickness profile is generated before polishing and represents the desired thickness of the outer layer in that region after polishing. The conductive outer layer on the substrate is polished, and the substrate is monitored during polishing using an in-situ monitoring system to generate a sequence of signal values, which depend on the thickness of the conductive outer layer and the thickness and / or conductivity of one or more underlying layers beneath the outer layer. The sequence of signal values ​​from the region is converted into a sequence of effective thickness values ​​for the region, where each effective thickness value includes contributions from the outer layer and the one or more underlying layers. A function is fitted to the sequence of effective thickness values, and the fitted function is used to determine the effective initial thickness value of the layer at the start of polishing, and an adjusted target thickness value is calculated based on the initial target value, the effective initial thickness value, and the initial thickness value. The polishing endpoint can be detected or polishing parameters can be modified based on the effective thickness value sequence and the adjusted target thickness value.

[0007] In another embodiment, prior to polishing, an initial thickness profile of a conductive outer layer on a substrate is received, and an initial target thickness profile of the conductive outer layer on the substrate is received. The initial thickness profile represents the thickness of the outer layer at a plurality of regions on the substrate before polishing, and the initial target thickness profile represents the desired thickness of the outer layer at the plurality of regions after polishing. During polishing of the substrate, a sequence of signal values ​​is received from an in-situ monitoring system, the signal values ​​depending on the thickness of the conductive outer layer being polished and the thickness and / or conductivity of one or more underlying layers beneath the outer layer. For each corresponding region among the plurality of regions, the sequence of signal values ​​from the corresponding region is converted into a sequence of effective thickness values ​​for the corresponding region, wherein each effective thickness value includes a contribution from the conductive outer layer and the one or more underlying layers, thereby providing a plurality of effective thickness value sequences, wherein each corresponding sequence among the plurality of sequences corresponds to a corresponding region. For each corresponding region among the plurality of regions, a function is fitted to the effective thickness value sequence, thereby providing a plurality of fitting functions, where each corresponding fitting function corresponds to a corresponding region. These plurality of fitting functions are used to determine the effective initial thickness profile of the layer at the start of polishing, and an adjusted target thickness profile is calculated based on the initial target thickness profile, the effective initial thickness profile, and the initial thickness profile. Polishing parameters are then modified based on the plurality of effective thickness value sequences and the adjusted target thickness profile.

[0008] In another embodiment, a method for determining the initial thickness profile of a layer on a substrate includes the following steps: polishing a conductive outer layer on a calibration substrate; monitoring the calibration substrate during polishing by repeatedly sweeping a sensor of an in-situ monitoring system across the calibration substrate to generate a first trajectory sequence, wherein each first trajectory in the first trajectory sequence corresponds to a sweep performed by the sensor, and each first trajectory includes a sequence of effective thickness values ​​depending on the thickness of the conductive outer layer and the thickness and / or conductivity of one or more underlying layers beneath the outer layer; detecting exposure of the one or more underlying layers beneath the outer layer; and after exposure of the one or more underlying layers, continuing to monitor the calibration substrate by sweeping the sensor of the in-situ monitoring system across the calibration substrate and generating a second trajectory including a sequence of underlying layer thickness values ​​depending on the one or more underlying layers. The thickness and / or conductivity; for each corresponding first trajectory from the first trajectory sequence, subtracting the second trajectory from the first trajectory to generate a modified first trajectory, thereby generating a modified trajectory sequence; for each corresponding region in a plurality of regions on the substrate and for each corresponding modified trajectory from the modified trajectory sequence, converting a portion of the signal value sequence of the corresponding region from the corresponding modified trajectory into a thickness value of the corresponding region, thereby providing a plurality of thickness value sequences, wherein each corresponding sequence in the plurality of sequences corresponds to a corresponding region; for each corresponding region in the plurality of regions, fitting a function to the thickness value sequence of the corresponding region, thereby providing a plurality of fitting functions, wherein each corresponding fitting function in the plurality of fitting functions corresponds to a corresponding region; and using the plurality of fitting functions to determine the initial thickness profile of the conductive layer at the start of polishing.

[0009] Implementations of methods, computer programs, and / or systems for achieving this state may include one or more of the following features.

[0010] Calculating the adjusted target thickness value may include calculating an adjustment based on the difference between the initial thickness value and the initial target value. Calculating the adjusted target thickness value may include subtracting the adjustment from the effective initial thickness value. Calculating the adjustment includes multiplying the difference between the initial thickness value and the initial target value by a constant. This constant may be between 1.0 and 1.2.

[0011] The function can be a linear function. Determining the effective starting thickness value can include extrapolating the function backward to the start time of the polishing. The in-situ monitoring system can be an eddy current monitoring system. The polishing endpoint can be detected by determining when the function equals the adjusted target thickness value. Fitting the function can include repeatedly fitting the function to a moving window of effective thickness values ​​from the sequence of effective thickness values.

[0012] Converting the sequence of signal values ​​from the region into the sequence of effective thickness values ​​may include using a conversion algorithm to calculate a preliminary sequence of effective thickness values ​​from the effective thickness values, the conversion algorithm taking signal values ​​as input and producing thickness values ​​as output. The sensor of the in-situ monitoring system may perform multiple sweeps across the substrate to generate a sequence of tracks, and each corresponding track in the sequence of tracks may be defined by thickness values ​​from the sequence of thickness values ​​for the corresponding sweep in the multiple sweeps. For each corresponding track, edge reconstruction may be performed on the corresponding track to generate a modified track with modified thickness values, thereby generating a modified track sequence. For each corresponding track, an effective thickness value for the region may be calculated based on the modified thickness values ​​of the region from the corresponding track, thereby generating the sequence of effective thickness values. Calculating the effective thickness value of the region based on the modified thickness values ​​may include averaging the modified thickness values ​​from the region.

[0013] The implementation may include one or more of the following advantages. When calculating the thickness of the layer being polished based on the monitored signal, this technique can compensate for the contribution of the conductive underlayer (e.g., a doped substrate or metal layer) to the signal, even if the contribution of the underlayer is inconsistent on a wafer-to-wafer basis. Therefore, the thickness of the layer being polished can be calculated with higher accuracy or reliability. The calculated thickness can be used to determine control parameters during the polishing process and / or to determine the endpoint of the polishing process. This can improve the reliability of control parameter determination and endpoint detection, avoid under-polishing of wafers, and reduce wafer-to-wafer non-uniformity (WTWNU) and intra-wafer non-uniformity (WIWNU).

[0014] Details of one or more embodiments will be set forth in the accompanying drawings and the description below. Other features, characteristics, and advantages will become apparent from the description, the drawings, and the claims.

Implementation Method

[0030] One monitoring technique used in polishing operations is to induce eddy currents in a conductive layer on a substrate, for example, using an alternating current (AC) drive signal. During polishing, an eddy current sensor can measure the induced eddy currents in situ to generate a signal. Assuming the outermost layer undergoing polishing is a conductive layer, the signal from the sensor should depend on the thickness of the conductive layer. Based on this monitoring, control parameters (such as the polishing rate) used in the polishing operation can be adjusted in situ. Furthermore, the polishing operation can be terminated based on an indication that the monitored thickness has reached the desired termination thickness.

[0031] In reality, the magnetic field generated by the eddy current sensor does not end within the outermost conductive layer but can extend into the underlying conductive layer or the underlying substrate. Therefore, the signal generated by the eddy current sensor can depend on the conductivity of the semiconductor wafer and the presence of the underlying conductive layer. If the semiconductor wafer is undoped, such as in "blank" wafers used for system calibration and as commonly used in base substrate wafers, then the resistance of the wafer may be high enough that its presence will not have a detectable effect on the eddy current signal. However, for actual device manufacturing, wafers are often doped, for example, heavily doped, for various purposes. In this case, depending on the conductivity of the semiconductor wafer, the substrate may contribute significantly to the signal generated by the eddy current sensor. Therefore, thickness measurements based on the signal captured by the eddy current sensor may be inaccurate.

[0032] If the contribution of the substrate due to doping is known, then the calculation of the thickness of the layer being polished can take into account the contribution of the semiconductor substrate to the signal. However, during semiconductor device manufacturing, doping is often poorly controlled, whether across a given wafer or between wafers within a given production batch.

[0033] This disclosure relates to a technique for performing endpoint and profile control when the underlying layer (such as a doped semiconductor wafer) makes inconsistent contributions to signals from an eddy current sensor. An initial effective thickness profile can be calculated by calculating the polishing rate for each of a plurality of regions on a substrate based on signals from the sensor, and by pushing the effective thickness backwards for each region to the start of the polishing operation. A deposition layer profile representing the actual initial thickness of the deposited film can then be subtracted from the effective initial thickness profile to provide an underlying layer profile, i.e., the underlying layer contribution in the form of the effective thickness. The underlying layer profile can then be added to a target thickness profile to produce an adjusted target thickness profile, which can be used for endpoint detection.

[0034] Figures 1 and 2 show an example of a polishing station 20 in a chemical mechanical polishing system. The polishing station 20 includes a rotatable disc-shaped platform 24 on which a polishing pad 30 is located. The platform 24 can be rotated about an axis 25. For example, a motor 22 can rotate a drive shaft 28 to rotate the platform 24. The polishing pad 30 can be a double-layered polishing pad with an outer polishing layer 34 and a softer backing layer 32.

[0035] Polishing station 20 may include a supply port or a combined supply-rinse arm 39 to dispense polishing fluid 38 (such as abrasive slurry) onto polishing pad 30. Polishing station 20 may include a pad adjustment device with an adjustment disc to maintain the surface roughness of the polishing pad.

[0036] The carrier head 70 can be used to hold the substrate 10 against the polishing pad 30. The carrier head 70 is suspended on the support structure 72 (e.g., a rotating rack or track) and connected to the carrier head rotation motor 76 via the drive shaft 74, so that the carrier head can rotate about the axis 71. Optionally, the carrier head 70 can swing laterally, for example, on a slider on the rotating rack, by moving along the track, or by the rotational swing of the rotating rack itself.

[0037] The carrier head 70 may include a flexible diaphragm 80 having a substrate mounting surface for contacting the back side of the substrate 10. The diaphragm 80 may form a plurality of pressurizable chambers 82 to apply different pressures to different areas (e.g., different radial areas) on the substrate 10. The carrier head may also include a retaining ring 84 to hold the substrate below the diaphragm 80.

[0038] During operation, the platform rotates about its central axis 25, and the carrier head rotates about its central axis 71 and laterally translates across the top surface of the polishing pad 30. In the case of multiple carrier heads, each carrier head 70 may have independent control over its polishing parameters; for example, each carrier head may independently control the pressure applied to each corresponding substrate.

[0039] In some embodiments, polishing station 20 includes a temperature sensor 64 to monitor the temperature of the polishing station or components of the polishing station / components in the polishing station.

[0040] Referring to FIG. 3A, a polishing system can be used to polish a substrate 10, which includes conductive material covering and / or embedded in a patterned dielectric layer. For example, substrate 10 may include a layer of conductive material 16 (e.g., a metal such as copper, aluminum, cobalt, or titanium) that covers and fills trenches in dielectric layer 14, such as silicon oxide or a high-k dielectric. Optionally, a barrier layer 18 (e.g., tantalum or tantalum nitride) may be lining the trenches and separating the conductive material 16 from dielectric layer 14. The conductive material 16 in the trenches may provide vias, pads, and / or interconnects in the completed integrated circuit. Although dielectric layer 14 is illustrated as being deposited directly on semiconductor wafer 12, one or more other layers may also be inserted between dielectric layer 14 and wafer 12.

[0041] The semiconductor wafer 12 can be a silicon wafer, such as single-crystal silicon; however, other semiconductor materials are also possible, such as gallium arsenide or gallium nitride. Furthermore, the semiconductor wafer 12 can be doped, for example, with p-type or n-type doping. Doping can be uniform across the wafer, or the wafer can be selectively doped, for example, selectively doped according to the needs of manufacturing transistors in integrated circuits using the semiconductor wafer.

[0042] Initially, conductive material 16 covers the entire dielectric layer 14. As polishing proceeds, most of the conductive material 16 is removed, thereby exposing the barrier layer 18 (see FIG. 3B). Polishing then continues to expose the patterned top surface of the dielectric layer 14 (see FIG. 3C). Additional polishing can then be used to control the depth of the trenches containing the conductive material 16.

[0043] In some embodiments, the polishing system includes additional polishing stations. For example, the polishing system may include two or three polishing stations. For example, the polishing system may include a first polishing station having a first electromagnetic induction monitoring system and a second polishing station having a second electromagnetic induction current monitoring system.

[0044] For example, in operation, batch polishing of the conductive layer on the substrate can be performed at the first polishing station, and polishing can be stopped when the target thickness of the conductive layer on the substrate is still present. Then, the substrate is transferred to the second polishing station and polished until the bottom layer (e.g., the patterned dielectric layer) is reached.

[0045] Referring back to Figure 1, the polishing system includes an in-situ eddy current monitoring system 100, which may be coupled to or considered to include the controller 90. The in-situ eddy current monitoring system 100 is configured to generate a signal that depends on the thickness of the conductive material 16 (e.g., metal) of the layer being polished.

[0046] During operation, the polishing system may use the in-situ eddy current monitoring system 100 to determine when the conductive layer has reached a target thickness (e.g., the target thickness of a metal layer covering the dielectric layer) and then stop polishing. Alternatively or additionally, the polishing system may use the in-situ eddy current monitoring system 100 to determine the thickness difference of the conductive material 16 across the substrate 10 and use this information during polishing to adjust the pressure in one or more chambers 82 in the carrier head 80 to reduce polishing non-uniformity. During polishing, measurements from the sensor 102 may be displayed on an output device to allow the polishing station operator to visually monitor the progress of the polishing operation; however, this is not required.

[0047] The sensor 102 of the in-situ monitoring system 100 can be mounted in a recess 26 in the platform 20. The sensor 102 may include a magnetic core 104 at least partially positioned in the recess 26 and at least one coil 106 wound around a portion of the magnetic core 104. A drive and sensing circuitry 108 is electrically connected to the coil 106. The drive and sensing circuitry 108 generates a signal that can be sent to the controller 90. The circuitry 108 may include a capacitor connected in parallel with the coil 106. The coil 106 and the capacitor may together form an LC resonant groove. Although illustrated outside the platform 24, some or all of the drive and sensing circuitry 108 may also be mounted inside the platform 24. A rotary coupler 29 may be used to electrically connect components in the rotatable platform 24 (e.g., the sensor 102 of the eddy current monitoring system 100) to components outside the platform (e.g., the drive and sensing circuitry or the controller 90). Alternatively, the sensor 102 may communicate wirelessly with the controller 90.

[0048] Referring to Figures 1 and 4, the drive and sensing circuitry 108 applies an AC current to the coil 106, which generates a magnetic field 110 between the two poles 112a and 112b of the magnetic core 104. In operation, a portion of the magnetic field 110 extends into the substrate 10 as the substrate 10 intermittently covers the sensor 102.

[0049] If it is necessary to monitor the thickness of the conductive layer on the substrate, then when the magnetic field 110 reaches the conductive layer 16, the magnetic field 110 can generate eddy currents in the conductive layer. This modifies the effective impedance of the LC circuit.

[0050] However, the magnetic field 110 may also penetrate into the various conductive layers 18 "below" the layer being polished (above in the view of FIG. 4) and into the semiconductor substrate 12. Therefore, the effective impedance of the LC circuit (and thus the signal from the driving and sensing circuit system 108) can also depend on the conductivity of the underlying conductive layer and the doping and resulting conductivity of the semiconductor substrate 12. For ease of illustration, the patterning of the layers is not shown in FIG. 4.

[0051] The drive and sensing circuitry 108 may include a marginal oscillator coupled to the combined drive / sensing coil 106, and the output signal may be the current required to maintain the peak-to-peak amplitude of the sinusoidal oscillation at a constant value, as described in, for example, U.S. Patent No. 7,112,960. The drive and sensing circuitry 108 may also have other configurations. For example, separate drive and sensing coils may be wound around a magnetic core. The drive and sensing circuitry 108 may apply current at a fixed frequency, and the signal from the drive and sensing circuitry 108 may be a phase shift of the current in the sensing coil relative to the drive coil, or the amplitude of the sensed current, as described in, for example, U.S. Patent No. 6,975,107.

[0052] Referring to FIG. 2, as the platform 24 rotates, the sensor 102 sweeps along path 120, which passes under the substrate 10. By sampling signals from the circuit system 108 at a specific frequency, the circuit system 108 generates measurement results at a series of sampling regions 94 across the substrate 10. For each sweep, the measurement results of one or more sampling regions 94 can be selected or combined. Therefore, after multiple sweeps, the selected or combined measurement results provide a time-varying sequence.

[0053] Polishing station 20 may also include a position sensor 96, such as an optical interrupter, to sense when sensor 102 is under substrate 10 and when sensor 102 leaves substrate. For example, position sensor 96 may be mounted in a fixed position opposite carrier head 70. Flag 98 may be attached to the periphery of platform 24. The attachment point and length of flag 98 are selected so that it can signal to position sensor 96 as sensor 102 sweeps under substrate 10.

[0054] Alternatively or additionally, the polishing station 20 may include an encoder to determine the angular position of the platform 24.

[0055] Referring back to Figure 1, the controller 90 (e.g., a general-purpose programmable digital computer) receives signals from the sensor 102 of the in-situ monitoring system 100. Since the sensor 102 sweeps beneath the substrate 10 with each rotation of the platform 24, information about the depth of the conductive layer (e.g., the body layer or conductive material in the trench) is accumulated in situ (once per platform rotation). The controller 90 can be programmed to sample the signals from the in-situ monitoring system 100 when the substrate 10 substantially covers the sensor 102.

[0056] In addition, the controller 90 can be programmed to calculate the radial position of each measurement result and classify the measurement results according to the radial range, as discussed in U.S. Patent No. 6,399,501.

[0057] Since the sensor 102 sweeps under the substrate 10 each time the platform 24 rotates, information about the thickness of the conductive layer is accumulated in situ and on a continuous real-time basis for each of the multiple different radial regions on the substrate.

[0058] Figure 5 shows a graph 130 illustrating the relationship between the thickness of the conductive layer and the signal from the eddy current monitoring system 100 at a given resistivity. As shown, the given thickness D can be calculated from the signal S; K is a constant representing the signal value when the thickness of the conductive layer is zero.

[0059] The relationship curve 130 can be represented by a function in the controller 90, such as a polynomial function, like a second-order function, a third-order function, or a higher-order function. In the case of an undoped semiconductor wafer, the correlation between the signal S and the thickness D can be represented by the following equation: S = W1·D² + W2·D + W3 (Equation 1) where W1, W2, and W3 are real coefficients. Therefore, the controller can store the values ​​of the coefficients of the function, such as W1, W2, and W3, and the resistivity ρ0 to which the relationship curve 410 applies. Furthermore, this relationship can be represented by a linear function, a Bezier curve, or a non-polynomial function (such as an exponential or logarithmic function).

[0060] In addition, the processing temperature, i.e. the temperature of the polishing pad, polishing fluid, substrate, or some combination thereof, can be taken into account when calculating the relationship curve or thickness for the layer. For example, the corrected thickness value D' can be calculated as follows: D' = D*(ρT / ρ0) ρT =ρX[1 + α (T-Tini)] where ρX is the resistivity of the conductive layer, ρ0 is the resistivity used to generate the relationship curve 130, Tini is the temperature used to generate the relationship curve 130, T is the measured temperature of the process (e.g., from temperature sensor 64), and α is a constant that can be derived from experience or scientific textbooks.

[0061] Some variations in the signal intensity from sensor 102 may be due to the overlap between the measurement area of ​​sensor 102 and the edge of the substrate, rather than inherent variations in the thickness or conductivity of the layer being monitored. Thus, this signal distortion can lead to errors in calculating the thickness value of the substrate, particularly near the substrate edge. To address this issue, controller 90 can feed each thickness trajectory into a neural network configured to generate a modified thickness trajectory that compensates for such signal distortion. This process is described in U.S. Patent Publication No. 2021-0379723. The resulting thickness measurements can be categorized according to radial range.

[0062] As described above, the signal generated by the eddy current sensor also includes contributions from the underlying conductive layer and the doped semiconductor wafer. Because the eddy currents excited in the doped semiconductor wafer and the conductive layer are independent and separated by the insulating layer, the power losses in the doped silicon substrate and the insulating conductive film layer are additive and should generally follow the superposition principle of linear systems.

[0063] Unfortunately, in many cases, the doping of semiconductor wafers is not strictly controlled, and therefore can vary greatly across a single wafer and between wafers, and even within a given production batch. On the other hand, in some cases, the initial thickness of the conductive layer being polished is known with high accuracy and reliability. This may be the case, for example, when the deposition process used to form conductivity is reliable, and the thickness of the conductive layer on the calibration substrate or on the sample substrate from the production batch can be measured using other techniques (e.g., a four-point probe).

[0064] Figure 6 shows a flowchart illustrating the process 200 for polishing a substrate and detecting the endpoint or adjusting polishing parameters during the polishing operation.

[0065] The controller 90 can receive and store the initial film thickness profile and the target thickness profile (202). This initial thickness profile can be a set of initial thickness values ​​representing the initial (i.e., pre-polishing) thickness of the conductive layer to be polished at each of a plurality of radial regions. The number of radial regions (and therefore the number of initial thickness values) can be equal to the number of independently controllable regions provided by the carrier head 140, such as the three regions corresponding to the three concentric chambers 82a-82c shown in FIG. 1. These values ​​can be obtained and entered by the plant facility operator. For example, the thickness of the conductive layer on a calibration substrate or a sample substrate from a production batch can be measured (e.g., using a four-point probe technique) at each radial range. For N radial ranges, the initial thickness values ​​can be represented as ST1, ST2...STN. In the case of using a reference radial range and a dynamically controlled radial range, the target thickness values ​​can be represented as STR and STD.

[0066] Similarly, the target thickness profile can be a set of target thickness values ​​representing the desired thickness to be achieved (i.e., after polishing) at each of the plurality of radial zones. Likewise, the number of target thickness values ​​can be equal to the number of initial thickness values ​​and equal to the number of independently controllable zones provided by the carrier head 140. For N radial ranges, the target thickness values ​​can be represented as TT1, TT2…TTN. These values ​​can be selected by the factory facility operator based on the specifications of the integrated circuit being manufactured. When using reference radial ranges and dynamically controlled radial ranges, the target thickness values ​​can be represented as TTR and TTD.

[0067] Referring to Figures 2 and 6, each time the sensor 102 sweeps under the substrate 10, the eddy current monitoring system generates a series of measurement results in the form of raw signal values, corresponding to different sampling areas 94. The controller 90 classifies these signal values ​​according to different radial ranges. These different radial ranges can be matched with the radial range of the initial thickness value, i.e., in a one-to-one relationship.

[0068] For each radial range, raw signal values ​​from one or more sampling areas 94 within that radial range can be selected or combined to provide a measurement result for that range, i.e., a signal value. Therefore, for each sweep, the eddy current monitoring system 100 generates one or more signal values ​​for each radial range. Thus, as polishing proceeds over time and the sensor 12 performs multiple sweeps, the eddy current monitoring system 100 generates a sequence of signal values ​​for each radial range (step 204).

[0069] Referring to Figures 2 and 5, each signal value (S) from the sequence is converted into an effective thickness value (D) (step 206), for example, using the relationship curve 130 conversion, thereby generating an effective thickness value sequence for each radial range. The term "effective thickness value" is used because the value is calculated assuming that the signal is entirely caused by the conductivity of the layer being polished. However, in reality, the signal includes contributions from the underlying layer and the doped substrate, whose conductivity may differ from that of the layer being polished, but is included in the measurement results as additional thickness of the material being polished.

[0070] Figure 7 shows a diagram (shown for illustrative purposes; no diagram needs to be generated or displayed during operation) showing a sequence 160 of effective thickness values ​​150 generated for a specific radial range, wherein the sensor 102 generates a signal value 200 each time it sweeps under the substrate 10.

[0071] For each radial range, a polynomial function (e.g., a first-order function, i.e., line 170) is fitted to a sequence 160 of effective thickness values ​​150 for that radial range (step 208). For example, robust regression can be used to perform the fitting. In particular, line 170 can be fitted to the effective thickness values ​​accumulated after polishing begins but before the first parameter change (e.g., the first change in carrier head pressure or other variables to improve polishing uniformity). Effective thickness values ​​from the first one to three rotations of the platform can be excluded because the polishing rate may be unstable during this period. And, depending on when the carrier head pressure begins to change, effective thickness values ​​acquired after the pressure change should be excluded, so that the slope of line 170 thus represents the polishing rate for the associated radial range. Data from four to ten effective thickness values ​​may be sufficient to fit line 170; however, fitting of the line can also continue as polishing progresses.

[0072] Calculate the effective thickness value at the start of the polishing operation (step 208). For example, based on the fitted polynomial function, the controller 90 extrapolates backward to determine the effective initial thickness value (PRE) of the substrate at the start of the polishing process. For N radial ranges, the effective initial thickness value can be represented as PRE(F+S),1, PRE(F+S),2…PRE(F+S),N. The subscript F+S indicates that the thickness value includes contributions from both the layer being polished and the underlying substrate. For the reference radial range, the effective initial thickness value can be represented as PRE(F+S),R, and for the dynamically controlled radial range, the effective initial thickness value can be represented as PRE(F+S),D.

[0073] With the target thickness value, the initial thickness value, and the effective thickness value, the adjusted target thickness value can be calculated (step 210). In particular, the adjusted target thickness value (AT) can be calculated as: AT = PRE(F+S)- (ST-TT)*k where k is a constant.

[0074] In some embodiments, this calculation is performed only for a reference radial range from the set of radial ranges, for example, ATR = PRE(F+S), R - (STR-TTR)*k. In some embodiments, this calculation is performed for each radial range, for example, AT1 = PRE(F+S), 1 - (ST1-TT1)*k, AT2 = PRE(F+S), 2 - (ST2-TT2)*k, and so on. If the calculation is performed for each radial range, then the same k value can be used for each range.

[0075] Without being limited to any particular theory, k can be one or both of the following: 1) compensation for the influence of the nonlinear relationship curve; and 2) crosstalk effect between the conductive layer being polished and the underlying layer, which causes the measurement results to not follow the superposition principle of linear systems. The constant k can be determined empirically and can be in the range of 1.0 to 1.2. For example, for a highly doped P++ silicon wafer, k can be about 1.04.

[0076] The controller 90 can use conventional algorithms to perform endpoint detection or contour control, i.e., calculate the adjustment of the carrier head pressure, but in the endpoint or contour control algorithm, the adjusted target thickness value is used instead of the target thickness value (step 212).

[0077] For example, still referring to Figure 7, as polishing proceeds, the controller 90 can be pushed forward to determine the time when line 170 equals (at 180) the adjusted target thickness value. This time provides the endpoint time (ET).

[0078] As another example, by replacing the target thickness value with an adjusted target thickness value in the contour control algorithm, the algorithm drives the polishing process to provide these adjusted target thickness values. For example, Figure 8 shows a graph (shown for illustrative purposes only; no graph needs to be generated or displayed during operation) showing two sequences of thickness values ​​generated for two different radial ranges. In this example, a first effective thickness value 152a is obtained, and a first line 172a is fitted to the first effective thickness value 152a for the first radial range, and a second effective thickness value 152b is obtained, and a second line 172b is fitted to the second effective thickness value 152b for the second radial range. Assuming the first radial range is used as the reference radial range, the endpoint time ETR can be calculated as the time when the first line 172a reaches the first adjusted target thickness value ATR (at 182a). Assuming the second radial range is used as the dynamic range, the endpoint time ETD can be calculated as the time when the second line 172b reaches the first adjusted target thickness value ATD (at 182b). The controller 90 can calculate a target polishing rate 190 for the dynamic radial range starting at adjustment time T0, such that both the controlled radial range and the dynamic radial range substantially simultaneously reach their corresponding adjusted target thickness values ​​ATR and ATD. To achieve the target polishing rate 190, the controller 90 can calculate adjustments to one or more polishing parameters, such as adjustments to the pressure in the region of the carrier head corresponding to the dynamic radial range. This calculation can be based on the Preston Equation. Therefore, even if the contribution of the underlying layer to the eddy current signal is inconsistent on a wafer-to-wafer basis, the resulting polished profile should more closely match the target profile.

[0079] As an alternative to measuring the initial thickness value on the layer using a four-point probe, an in-situ monitoring system can be used to determine the effective initial thickness value for each radial region. Furthermore, the thickness of some regions can be represented using an offset relative to a reference region, rather than using absolute values. The procedures for using these two techniques are described below, but each technique can be used independently.

[0080] Figures 9A-9C illustrate a method 300 for performing endpoint and profile control in chemical mechanical polishing, the method including a process 302 for determining an initial thickness value for each radial region using an in-situ monitoring system. This process includes generating a thickness trajectory (304) during bulk polishing of a conductive layer on a calibration substrate, generating a thickness trajectory (306) once the conductive layer has been removed to expose the underlying layer, and processing the thickness trajectory to generate an initial thickness value for each radial range (308). Each of these steps will be described in more detail below.

[0081] Initially, referring to FIG. 9A, a calibration substrate having the same layers and pattern as the product substrate is polished at polishing station 20. During the polishing of the outer conductive layer (step 310), the substrate is monitored using an in-situ eddy current monitoring system 160. Referring to FIG. 2, FIG. 9A, and FIG. 12, the sensor 120 generates a raw signal trajectory each time it sweeps under the substrate 10 (step 312). Each signal value in the raw signal trajectory can be converted into a thickness value (step 314), for example, using the calibration relationship discussed above with reference to FIG. 5. This generates an effective thickness trajectory 300 for each sweep of the sensor 102 under the calibration substrate.

[0082] As shown in FIG12, the effective thickness trajectory 250 includes a first portion 252, a second portion 254, and a third portion 256. The first portion corresponds to the position in the edge region of the substrate 10 when the sensor passes the leading edge of the substrate 10, the second portion corresponds to the position in the central region of the substrate 10, and the third portion corresponds to the position in the edge region when the sensor 102 passes the trailing edge of the substrate 10. The signal may also include a portion 258 corresponding to the measurement result outside the substrate, i.e., the signal generated when the sensor head scans the region outside the edge of the substrate 10.

[0083] Although the first portion 252 and the third portion 256 are shown as smooth, and the second portion 254 is shown as flat, this is only for simplicity. The actual signal will include fluctuations caused by both noise and layer thickness variations. Furthermore, although the effective thickness trajectory 250 is shown as continuous, it is actually provided by a sequence of individual effective thickness values.

[0084] Continuing to refer to Figures 9A and 12, as described above, the thickness measurement results at the substrate edge may be inaccurate. Therefore, an edge reconstruction technique can be performed on each effective thickness trajectory 250 to generate a modified effective thickness trajectory 260 for each scan (step 316). In particular, each effective thickness trajectory 250 can be fed into a neural network configured to compensate for signal distortion at the substrate edge. This process is described in U.S. Patent Publication No. 2021-0379723. Each modified effective thickness trajectory 260 can be stored for subsequent processing.

[0085] Referring to Figures 9A and 10, the thickness values ​​from each modified effective thickness trajectory 260 can be classified according to radial range, and the effective thickness value can be calculated for each radial range for each sensor sweep (step 320). For example, a sequence 164a of effective thickness values ​​154a is generated for a reference radial range, and a sequence 164b of effective thickness values ​​154b is generated for each dynamic radial range (for ease of illustration, only two sequences are shown in Figure 10). The effective thickness value can be generated by averaging the effective thickness values ​​from the trajectory 260 over a specific radial range.

[0086] The controller 90 can detect when the conductive layer is polished away, exposing the underlying layer. In this state, the top surface of the underlying layer is exposed, but conductive material may remain in the trenches, forming wires, vias, etc. In one embodiment, the controller 90 or operator can detect the slope change (at 184) of a line fitted to a sequence 164a of effective thickness values ​​154a in a reference area. For example, the controller 90 can repeatedly fit the line to the effective thickness value 154a from a moving window of three to ten effective thickness values ​​and calculate the polishing rate based on the slope of the line (step 322). If the slope change of the line exceeds a threshold, the controller 90 can use this as an indication that the underlying layer has been exposed (step 324). In some embodiments, the slope change in the radial range alone is sufficient to indicate exposure. In some embodiments, it is necessary to refer to the slope changes in both the radial range and the dynamic radial range. Furthermore, other techniques, such as separate optical monitoring signals, can be used to locate the time of underlying layer exposure.

[0087] With the substrate exposed, referring to FIG9B, the calibration substrate can now be scanned during the "post-polishing" process. Specifically, after the substrate is exposed, the substrate is held on the polishing pad (step 330), and the platform 24 continues to rotate to scan the sensor 102 under the substrate 10, thereby generating a sequence of raw signal trajectories (step 332). To prevent material removal during this post-polishing process, the calibration substrate can be scanned with the polishing pad surface rinsed with deionized water (DIW) and / or with the chamber in the carrier head under low pressure (e.g., evacuated to atmospheric pressure or less than 0.1 psi). Polishing may be substantially not performed during this post-polishing process; for example, the polishing rate may be less than 5%, such as less than 1% of the polishing rate during large-area polishing (in step 310).

[0088] Referring to Figures 9B and 13, each signal value in the original signal trajectory can be converted into a bottom layer thickness value (step 334), for example, using the relationship curve discussed above with reference to Figure 5. For the post-polishing process, this will generate a bottom layer thickness trajectory 270 for each sweep of sensor 102 under the calibration substrate.

[0089] As shown in FIG13, the bottom layer thickness trajectory 270 also includes a first portion 272, a second portion 274, and a third portion 276. The first portion corresponds to the position in the edge region of the substrate 10 when the sensor passes the leading edge of the calibration substrate, the second portion corresponds to the position in the central region of the substrate 10, and the third portion corresponds to the position in the edge region when the sensor 102 passes the trailing edge of the substrate 10. The signal may also include a portion 278 corresponding to the measurement result outside the substrate, i.e., the signal generated when the sensor head scans the region outside the edge of the substrate 10.

[0090] Continuing to refer to Figures 9B and 13, as described above, the thickness measurement results at the substrate edge may be inaccurate. Therefore, an edge reconstruction technique can also be performed on each bottom layer thickness trajectory 270 to generate a modified bottom layer thickness trajectory 280 for each scan (step 336). In particular, each bottom layer thickness trajectory 270 can be fed into a neural network configured to compensate for signal distortion at the substrate edge. This process is described in U.S. Patent Publication No. 2021-0379723. Each modified bottom layer thickness trajectory 280 can be stored for subsequent processing.

[0091] Using the adjusted effective thickness trajectory 260 (step 304) obtained from the large-area polishing of the conductive layer on the substrate and the adjusted bottom layer thickness trajectory 280 (step 306) obtained from the post-polishing scan of the substrate, the starting thickness value for each radial range can now be determined (step 308).

[0092] In particular, referring to Figures 9B and 14, a representative underlayer thickness trajectory 282 (340) is generated from the adjusted underlayer thickness trajectory 280. For example, one of the adjusted underlayer thickness trajectories can be selected as the representative underlayer thickness trajectory. Alternatively, the representative underlayer thickness trajectory can be generated by averaging two or more adjusted underlayer thickness trajectories. For example, the average underlayer thickness trajectory DS(t) can be calculated as follows:

[0093] , where Di(t) represents an individual adjusted bottom layer thickness trajectory, and N is the number of bottom layer thickness trajectories.

[0094] Next, referring to Figures 9B and 14, the representative bottom layer trajectory 282 is subtracted from each effective thickness trajectory 260 to generate a thickness trajectory 290 for each scan of the substrate obtained in step 304 (step 342), i.e.

[0095] This trajectory can be called the "thickness trajectory" because by subtracting the contribution of the underlying layer from the effective thickness value including both the substrate and the film thickness, only the contribution from the film thickness should remain.

[0096] Referring to Figures 9B, 11, and 14, the thickness values ​​from each thickness trajectory 290 can be classified according to radial range, and the thickness value can be calculated for each radial range for each sensor sweep (step 344). For example, the thickness value for a specific radial range for a specific sweep. Thus, a sequence 166a of thickness values ​​156a is generated for a reference radial range, and a sequence 166b of thickness values ​​156b is generated for each dynamic radial range (for ease of illustration, only two sequences are shown in Figure 11). For example, the thickness values ​​for radial ranges R1 to R2 can be generated by averaging the thickness values ​​from corresponding portions 292 of the thickness trajectory 290.

[0097] For each radial range, a polynomial function (e.g., a first-order function, i.e., line 176) is fitted to a sequence 166a, 166b of thickness values ​​156a, 156b for that radial range (step 346). For example, robust regression can be used to perform the fitting. In particular, line 176 can be fitted to thickness values ​​accumulated after polishing begins but before the underlying layer is exposed. Thickness values ​​from the first one to three rotations of the platform can be excluded because the polishing rate may be unstable during this period. In some implementations, data from four to thirty thickness values ​​are sufficient to fit line 176.

[0098] Calculate the thickness value at the start of the polishing operation for each radial range (step 348). For example, based on the fitted polynomial function, the controller 90 extrapolates backward to determine the initial thickness value (PRE) of the substrate at the start of the polishing process for the calibrated substrate. For the reference radial range, the initial thickness value can be represented as PRE(F),R, and for the dynamically controlled radial range, the effective initial thickness value can be represented as PRE(F),D. The subscript F indicates that the thickness value should essentially only represent the contribution of the conductive layer, i.e., not the contribution of the underlying layer or substrate.

[0099] Referring to FIG9C, for some implementations, the initial thickness value of the dynamic radial range can be stored as an offset (349). For example, the offset ΔPRED can be calculated for each dynamic radial range as: ΔPRE(F),D = PRE(F),R - PRE(F),D.

[0100] The system should now be ready to polish the product substrate. As mentioned above, the target thickness values ​​TTR and TTD for the reference radial range and dynamic radial range are typically provided by the operator depending on the application and the integrated circuitry fabricated on the product substrate. The target thickness value for the dynamic radial range can also be stored as an offset. For example, the offset ΔTTD can be calculated for each dynamic radial range as: ΔTTD = TTR - TTD.

[0101] Continuing with reference to Figures 9C and 15, the product substrate (350) is polished at polishing station 20, and the substrate is monitored using an in-situ eddy current monitoring system 160 during polishing of the exposed conductive layer. Each time the sensor 102 sweeps under the substrate 10, it generates a raw signal trajectory (step 352). Each signal value in the raw signal trajectory can be converted into an effective thickness value (step 354), for example, using the relationship curve discussed above with reference to Figure 5. This generates an effective thickness trajectory 250' for each sweep of the sensor 102 under the product substrate. As with the calibration substrate, thickness measurements at the edges of the product substrate may be inaccurate. Therefore, an edge reconstruction technique can be performed on each effective thickness trajectory 250' to generate a modified effective thickness trajectory 260' for each scan (step 356). In particular, each effective thickness trajectory 250' can be fed into a neural network configured to compensate for signal distortion at the edges of the substrate. This process is described in U.S. Patent Publication No. 2021-0379723.

[0102] Continuing with reference to FIG9C and returning to FIG8, the thickness values ​​from each modified effective thickness trajectory 260' can be classified according to radial range, and the effective thickness value can be calculated for each radial range for each sensor sweep (step 360). For example, a sequence 162a of effective thickness values ​​152a is generated for the reference radial range, and a sequence 162b of effective thickness values ​​152b is generated for each dynamic radial range (for ease of explanation, FIG8 only shows two sequences). Each effective thickness value 152a, 152b can be generated by averaging the effective thickness values ​​from corresponding portions of the corresponding effective thickness trajectory 260'. For example, the effective thickness values ​​for radial ranges R1 to R2 can be generated by averaging the effective thickness values ​​from corresponding portions 262 of the effective thickness trajectory 260'.

[0103] Once a sufficient number of effective thickness values ​​have been collected, a polynomial function (e.g., a first-order function, i.e., a line) is fitted to each sequence of effective thickness values ​​for that radial range (step 362). For example, the first line 172a is fitted to a sequence 162a of thickness values ​​152a for the reference radial range. Furthermore, for each dynamic radial range, line 172b is fitted to a sequence 162b of effective thickness values ​​152b for that dynamic radial range. Robust regression can be used to perform the fitting. Effective thickness values ​​from the first one to three rotations of the platform can be excluded, as the polishing rate may be unstable during this period. Also, depending on when the carrier head pressure begins to change, effective thickness values ​​acquired after any pressure change should be excluded, such that the slopes of lines 172a and 172b thus represent the polishing rate of the associated radial range before the pressure adjustment. Data from four to ten effective thickness values ​​may be sufficient to fit lines 172a and 172b.

[0104] Calculate the effective thickness value at the start of the polishing operation (step 364). For example, based on the fitted polynomial function, the controller 90 pushes backward to determine the effective initial thickness value (PRE) of the substrate at the start of the polishing process. For the reference radial range, the effective initial thickness value can be expressed as PRE(F+S),R, and for the dynamically controlled radial range, the effective initial thickness value can be expressed as PRE(F+S),D.

[0105] In some implementations, the effective starting thickness value of the dynamic radial range can be stored as an offset. For example, the offset ΔPRED can be calculated for each dynamic radial range as: ΔPRE(F+S),D = PRE(F+S),R - PRE(F+S),D.

[0106] With the target thickness value, the initial thickness value, and the effective thickness value, the adjusted target thickness value can be calculated for each radial range (step 366).

[0107] In particular, the adjusted target thickness value (ATR) for the reference radial range can be calculated as: ATR = PRE(F+S),R - (PRE(F),R-TTR)*k where k is the constant discussed above.

[0108] In addition, the adjusted target thickness value (ATD) for each dynamic radial range can be calculated as: ATD = ATRP + ΔPRE(F+S),D - ΔPRE(F),D + ΔTTR.

[0109] The controller 90 can use conventional algorithms to perform endpoint detection or contour control, i.e., calculate the adjustment of the carrier head pressure, but in the endpoint or contour control algorithm, the adjusted target thickness value is used instead of the target thickness value (step 370).

[0110] Continuing to refer to Figure 8, a first effective thickness value 152a is obtained for a reference radial range, and a first line 172a is fitted to the first effective thickness value 152a. A second effective thickness value 152b is obtained for a dynamic radial range, and a second line 172b is fitted to the second effective thickness value 152b. The endpoint time ETR is calculated as the time it takes for the first line 172a to reach the first adjusted target thickness value ATR (at 182a). The second radial range is used as the dynamic range, and the endpoint time ETD can be calculated as the time it takes for the second line 172b to reach the adjusted target thickness value ATD (at 182b) of the second radial range.

[0111] The controller 90 can calculate a target polishing rate 190 for the dynamic radial range starting at adjustment time T0, such that both the controlled radial range and the dynamic radial range substantially simultaneously reach their corresponding adjusted target thickness values ​​ATR and ATD. To achieve the target polishing rate 190, the controller 90 can calculate adjustments to one or more polishing parameters, such as adjustments to the pressure in the region of the carrier head corresponding to the dynamic radial range. This calculation can be based on the Preston Equation. Therefore, even if the contribution of the underlying layer to the eddy current signal is inconsistent on a wafer-to-wafer basis, the resulting polished profile should more closely match the target profile.

[0112] The polishing apparatus and method described above can be applied to various polishing systems. Either or both of the polishing pad or carrier head can be moved to provide relative movement between the polishing surface and the substrate. For example, the platform can be circumferential rather than rotating. The polishing pad can be a circular (or some other shape) pad fixed to the platform. Some configurations of the endpoint detection system can be applied to linear polishing systems, such as when the polishing pad is a linearly moving continuous belt or roll-to-roll belt. The polishing layer can be a standard (e.g., filled or unfilled polyurethane) polishing material, a soft material, or a fixed abrasive material. The term relative positioning is used to refer to relative positioning within the system or substrate; it should be understood that during the polishing operation, the polishing surface and the substrate can remain vertically oriented or in some other orientation.

[0113] The functional operation of the controller 90 can be implemented using one or more computer program products, that is, one or more computer programs tangibly implemented in a non-transitory computer-readable storage medium for execution or control of its operation by a data processing device (e.g., a programmable processor, a computer or multiple processors or computers).

[0114] Some embodiments of the invention have been described. Nevertheless, it will be understood that various modifications can be made without departing from the spirit and scope of the invention. Therefore, other embodiments are also within the scope of the following claims. [Simplified Explanation of the Diagram]

[0015] Figure 1 shows a schematic cross-sectional view of an example polishing station including an electromagnetic induction monitoring system.

[0016] Figure 2 shows a schematic top view of an example chemical mechanical polishing station, showing the path of sensor scanning across the substrate.

[0017] Figures 3A-3C are schematic cross-sectional views of the substrate, showing the polishing process.

[0018] Figure 4 is a schematic cross-sectional view showing an example magnetic field generated by an electromagnetic induction sensor.

[0019] Figure 5 shows an example signal diagram from an eddy current sensor as a function of the conductive layer thickness.

[0020] Figure 6 shows a flowchart of a method for performing endpoint and profile control in chemical mechanical polishing.

[0021] Figure 7 is a sequence diagram of thickness values.

[0022] Figure 8 is a graph of two thickness value sequences from two different radial ranges on the product substrate.

[0023] Figures 9A-9C show flowcharts of another embodiment of a method for performing endpoint and contour control in chemical mechanical polishing.

[0024] Figure 10 is a graph of two effective thickness value sequences from two different radial ranges during the polishing of the calibration substrate.

[0025] Figure 11 is a graph of two thickness value sequences from two different radial ranges derived from the polishing of the calibration substrate.

[0026] Figure 12 is a diagram of the effective thickness trajectory of the polished material from the calibration substrate.

[0027] Figure 13 is a diagram of the effective thickness trajectory of over-polishing from the calibration substrate.

[0028] Figure 14 is a thickness trajectory diagram derived from the trajectories in Figures 12 and 13.

[0029] Figure 15 is a diagram of the effective thickness trajectory of the polished substrate. [Biomaterial Storage]

[0116] Domestic Deposit Information (Please note in order of deposit institution, date, and number) None

[0117] Overseas Deposit Information (Please note in the order of deposit country, institution, date, and number) None

Claims

1. A computer program product for controlling a polishing system, the computer program product residing on a non-transitory computer-readable medium and including instructions for causing one or more computers to: receive an initial thickness value of a conductive outer layer in a region on a substrate, the initial thickness value being generated before polishing and representing a thickness of the outer layer in the region before polishing; receive an initial target thickness value of the conductive outer layer in the region on the substrate, the initial target thickness value being generated before polishing and representing a desired thickness of the outer layer in the region after polishing; receive a sequence of signal values ​​from an in-situ monitoring system during polishing of the substrate, the signal values ​​depending on a thickness of the conductive outer layer being polished and the thickness and / or conductivity of one or more underlying layers beneath the outer layer; convert the sequence of signal values ​​from the region into a sequence of effective thickness values ​​for the region, wherein each effective thickness value includes a contribution from the outer layer and the one or more underlying layers; fit a function to the sequence of effective thickness values; use the fitted function to determine an effective initial thickness value of the conductive outer layer at the start of polishing; An adjusted target thickness value is calculated based on the initial target value, the effective starting thickness value, and the starting thickness value; and a polishing endpoint is detected based on the sequence of effective thickness values ​​and the adjusted target thickness value.

2. The computer program product as claimed in claim 1, wherein the instructions for calculating the adjusted target thickness value include instructions for calculating an adjustment based on a difference between the initial thickness value and the initial target value.

3. The computer program product as claimed in claim 2, wherein the instructions for calculating the adjusted target thickness value include instructions for subtracting the adjustment from the effective starting thickness value.

4. The computer program product as claimed in claim 3, wherein the instructions for calculating the adjustment include instructions for multiplying the difference between the initial thickness value and the initial target value by a constant.

5. The computer program product as described in claim 4, wherein the constant is between 1.0 and 1.

2.

6. The computer program product as described in claim 1, wherein the function is a linear function.

7. The computer program product as claimed in claim 6, wherein the instructions for determining the effective starting thickness value include instructions for pushing the function backward to a starting time of the polishing.

8. The computer program product as claimed in claim 1, including instructions for detecting the polishing endpoint by determining a time when the function is equal to the adjusted target thickness value.

9. The computer program product as claimed in claim 8, wherein the instructions for fitting the function include instructions for repeatedly fitting the function to a moving window of effective thickness values ​​from the sequence of effective thickness values.

10. The computer program product as claimed in claim 1, wherein the instructions for converting the sequence of signal values ​​from the region into the sequence of effective thickness values ​​include instructions for using a conversion algorithm to calculate a preliminary sequence of effective thickness values ​​from the effective thickness values, the conversion algorithm accepting a signal value as an input and producing a thickness value as an output.

11. The computer program product as claimed in claim 10, wherein the sequence of thickness values ​​defines a sequence of trajectories, wherein each corresponding trajectory in the sequence of trajectories corresponds to a corresponding sweep in a plurality of sweeps of a sensor of the in-situ monitoring system across the substrate.

12. The computer program product as claimed in claim 11, including instructions for performing edge reconstruction on each corresponding trajectory to generate a modified trajectory having a modified thickness value, thereby generating a sequence of modified trajectories.

13. The computer program product as claimed in claim 12, including instructions for calculating an effective thickness value for a region based on the modified thickness values ​​of that region from the corresponding trajectory for each corresponding trajectory, thereby generating a sequence of effective thickness values.

14. The computer program product as claimed in claim 13, wherein the instructions for calculating the effective thickness value of the region based on the modified thickness values ​​include instructions for averaging the modified thickness values ​​from the region.

15. A chemical mechanical polishing method, the method comprising the steps of: receiving an initial thickness value of a conductive outer layer in a region on a substrate, the initial thickness value being generated prior to polishing and representing a thickness of the outer layer in the region prior to polishing; receiving an initial target thickness value of the conductive outer layer in the region on the substrate, the initial target thickness value being generated prior to polishing and representing a desired thickness of the outer layer in the region after polishing; polishing the outer layer on the substrate; monitoring the substrate during polishing with an in-situ monitoring system to generate a sequence of signal values, the signal values ​​depending on a thickness of the conductive outer layer and the thickness and / or conductivity of one or more underlying layers beneath the outer layer; converting the sequence of signal values ​​from the region into a sequence of effective thickness values ​​for the region, wherein each effective thickness value includes a contribution from the outer layer and the one or more underlying layers; fitting a function to the sequence of effective thickness values; and using the fitted function to determine an effective initial thickness value of the conductive outer layer at the start of polishing. Calculate an adjusted target thickness value based on the initial target value, the effective starting thickness value, and the starting thickness value; and detect at least one of a polishing endpoint or modify a polishing parameter based on the sequence of effective thickness values ​​and the adjusted target thickness value.

16. The method of claim 15, wherein the in-situ monitoring system includes an eddy current monitoring system.

17. A polishing system, comprising: A platform for supporting a polishing pad; A carrier head for holding a substrate in contact with the polishing pad; a motor for generating relative motion between the carrier head and the platform; an eddy current monitoring system for monitoring the substrate during polishing and generating a sequence of signal values, the signal values ​​depending on a thickness of a conductive outer layer being polished and the thickness and / or conductivity of one or more underlying layers beneath the outer layer; and a controller configured to: receive an initial thickness value of the conductive outer layer in a region on the substrate, the initial thickness value being generated before polishing and representing a thickness of the outer layer in the region before polishing; receive an initial target thickness value of the conductive outer layer in the region on the substrate, the initial target thickness value being generated before polishing and representing a desired thickness of the outer layer in the region after polishing; receive the sequence of signal values ​​from the eddy current monitoring system; convert the sequence of signal values ​​from the region into a sequence of effective thickness values ​​for the region, wherein each effective thickness value includes contributions from the outer layer and the one or more underlying layers; and fit a function to the sequence of effective thickness values. The fitted function is used to determine an effective initial thickness value of the conductive outer layer at the start of polishing; an adjusted target thickness value is calculated based on the initial target value, the effective initial thickness value, and the initial thickness value; and at least one of detecting a polishing endpoint or modifying a polishing parameter is performed based on the sequence of effective thickness values ​​and the adjusted target thickness value.

18. The system of claim 17, wherein the controller is configured to calculate the adjusted target thickness value including instructions for calculating an adjustment based on a difference between the initial thickness value and the initial target value.

19. The system of claim 18, wherein the controller is configured to calculate the adjusted target thickness value including instructions for subtracting the adjustment from the effective starting thickness value.

20. The system of claim 19, wherein the controller is configured to calculate the adjustment by including instructions for multiplying the difference between the initial thickness value and the initial target value by a constant.

21. The system as described in claim 20, wherein the constant is between 1.0 and 1.2.

Images