Compensation for substrate doping in edge reconstruction for in situ electromagnetic induction monitoring
Through in-situ electromagnetic induction monitoring system and neural network technology, the problem of the degree of planarization of substrate layer and the accuracy of end-point detection in chemical mechanical grinding is solved, and the impact on substrate doping is achieved is achieved, and the control and end-point detection of grinding process are improved. reliability.
Patent Information
- Application Number
- CN202410998518.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2018-09-26
- Publication Date
- 2024-11-12
AI Technical Summary
During the chemical mechanical grinding process, it is difficult to accurately determine the degree of planarization of the substrate layer and the end point of material removal, resulting in problems of inhomogeneity within the wafer and uncertainty of the grinding end point.
The in-situ electromagnetic induction monitoring system combined with neural network technology is used to monitor the thickness and eddy current signals of the conductive layer to generate adjusted traces to compensate for the impact of substrate doping on the measurement signal, and improve the reliability of control parameters and the accuracy of end point detection. sex.
It effectively reduces the influence of signal distortion at the edge of the substrate, improves the reliability of the control parameter determination and end point detection of the grinding process, and avoids insufficient and uneven wafer grinding.
Smart Images

Figure CN118943037A_ABST
Abstract
Description
[0001] This application is a divisional application of an invention patent application with an application date of September 26, 2018, application number "201880083574.6", and invention name "Compensation for substrate doping in edge reconstruction for in-situ electromagnetic induction monitoring". Technical Field
[0002] The present disclosure relates to chemical mechanical polishing, and more particularly to monitoring a conductive layer during chemical mechanical polishing. Background Art
[0003] Integrated circuits are typically formed on a substrate by depositing a conductive layer, a semiconductor layer, or an insulating layer in sequence on a silicon wafer. Various manufacturing processes require planarization of the layers on the substrate. For example, one manufacturing step includes 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 may be deposited on a patterned insulating layer to fill the grooves and holes in the insulating layer. After planarization, the remaining portions of the metal in the grooves and holes of the patterned layer form vias, plugs, and wiring to provide conductive paths between thin film circuits on the substrate.
[0004] Chemical mechanical polishing (CMP) is a well-known planarization method. This planarization method generally requires mounting the substrate on a carrier head. The exposed surface of the substrate is generally placed against a rotating polishing pad. The carrier head provides a controllable load on the substrate to push the substrate against the polishing pad. A polishing slurry having abrasive particles is generally supplied to the surface of the polishing pad.
[0005] One problem in CMP is determining whether the grinding process is complete, i.e., whether the substrate layer has been planarized to a desired flatness or thickness, or when a desired amount of material has been removed. Variations in slurry composition, polishing pad conditions, relative speed between the polishing pad and the substrate, initial thickness of the substrate layer, and load on the substrate can result in variations in the material removal rate. These variations result in variations in the time required to reach the grinding endpoint. Therefore, determining the grinding endpoint based solely on the grinding time can result in non-uniformity within a wafer or between wafers.
[0006] In some systems, the substrate is monitored in situ during polishing, for example via a polishing pad.One monitoring technique is to sense eddy currents in a conductive layer and detect changes in the eddy currents as the conductive layer is removed. Summary of the invention
[0007] In one aspect, a method for compensating for a contribution of conductivity of a semiconductor wafer to a measured trace by an in-situ electromagnetic induction monitoring system includes: storing or generating a modified reference trace representing a measurement of a bare doped reference semiconductor wafer by the in-situ electromagnetic induction monitoring system modified by a neural network; monitoring the substrate using the in-situ electromagnetic induction monitoring system while polishing a conductive layer to generate a measured trace that depends on the thickness of the conductive layer; applying at least a portion of the measured trace to the neural network to generate a modified measured trace; and generating an adjusted trace, including subtracting the modified reference trace from the modified measured trace.
[0008] In one aspect, a method for polishing a substrate includes: storing or generating a modified reference trace, the modified reference trace representing a measurement of a bare doped reference semiconductor wafer by an in-situ electromagnetic induction monitoring system modified by a neural network; contacting a substrate having a conductive layer disposed above the semiconductor wafer with a polishing pad; generating relative motion between the substrate and the polishing pad; monitoring the substrate using the in-situ electromagnetic induction monitoring system while polishing the conductive layer to generate a measured trace that depends on the thickness of the conductive layer; applying at least a portion of the measured trace to the neural network to generate a modified measured trace; generating an adjusted trace to at least partially compensate for a contribution of the conductivity of the semiconductor wafer to the measured trace, including subtracting the modified reference trace from the modified measured trace; and performing at least one of pausing polishing or modifying polishing parameters based on the adjusted trace.
[0009] Each of these aspects may also be implemented as a computer program product tangibly encoded on a computer-readable medium, the computer program product including instructions to cause a computer system to perform appropriate operations (e.g., store or generate a modified reference trace, apply a measured trace, and generate an adjusted trace); or as a polishing system including a controller configured to perform the appropriate operations.
[0010] Implementations of the methods, computer program products, and / or systems may include one or more of the following features.
[0011] The modified reference trace may include a series of equivalent thickness values, and the modified measured trace may include a series of actual thickness values. At least a portion of the initial reference trace may be applied to a neural network to generate a modified reference trace. Raw signal values in the preliminary reference trace may be converted to thickness values to generate the initial reference trace. User input may be received to select a reference trace from a plurality of reference traces. Generating the modified reference trace may include scanning a sensor of the in-situ electromagnetic induction monitoring system across a bare doped reference semiconductor wafer.
[0012] Generating the adjusted trace may include scaling the difference between the modified reference trace and the modified measured trace. The adjusted trace A(x) may be calculated such that A(x)=(T(x)-S(x)-b) / k, where T(x) is the modified measured trace, S(x) is the modified reference trace, and b and k are constants. The constants b and k are based on the configuration of the sensors of the in-situ monitoring system.
[0013] The at least a portion of the measured trace applied to the neural network includes a portion corresponding to an edge region of the substrate. The at least a portion of the measured trace applied to the neural network need not include a portion corresponding to a center region of the substrate. The neural network may be trained with a plurality of training traces representing measurements of one or more training substrates having a conductive layer on an undoped semiconductor wafer, wherein different training traces correspond to different thicknesses of the conductive layer and different edge profiles.
[0014] Implementations may include one or more of the following advantages. During monitoring of processing (e.g., grinding) of a substrate, possible inaccuracies in the correlation between a measured eddy current signal and the thickness of a conductive layer caused by doping of an underlying semiconductor wafer may be mitigated, especially at the edge of the substrate. Adjusting the eddy current signal or adjusting the thickness of the conductive layer using a compensation process may be more accurate. The system may compensate for distortion in a portion of the signal corresponding to an edge of the substrate. The adjusted eddy current signal and / or the adjusted conductive layer may be used to determine control parameters during a grinding process and / or determine an endpoint of a grinding process. The reliability of control parameter determination and endpoint detection may be improved, wafer under-grinding may be avoided, and non-uniformity within a wafer may be reduced.
[0015] The details of one or more implementations are set forth in the accompanying drawings and the following detailed description.Other aspects, features, and advantages will be apparent from the description and drawings, and from the claims. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 A schematic cross-sectional view of an example of a grinding station including an electromagnetic induction monitoring system is depicted.
[0017] Figure 2 A schematic top view of an exemplary chemical mechanical polishing station is depicted.
[0018] FIG. 3A to FIG. 3C is a schematic cross-sectional view of a substrate illustrating a polishing process.
[0019] Figure 4 A schematic cross-sectional diagram illustrating an exemplary magnetic field generated by an electromagnetic induction sensor.
[0020] Figure 5Schematic top view of a substrate being scanned by a sensor head of a grinding device.
[0021] Figure 6 is a schematic graph of a measured signal obtained while monitoring the position on the substrate.
[0022] Figure 7 is an exemplary neural network.
[0023] Figure 8 Schematic illustration of a process for compensating substrate doping while also applying edge reconstruction.
[0024] Fig. 9 Plotted is a graph of an exemplary eddy current phase signal as a function of conductive layer thickness.
[0025] Fig.10 is a schematic depiction of producing an adjusted profile of a substrate being polished. DETAILED DESCRIPTION
[0026] One monitoring technique for the grinding operation is to sense the current in the conductive layer on the substrate. The induced current can be measured in situ during grinding by an inductive monitoring system to generate a signal. Assuming that the outermost layer undergoing grinding is a conductive layer, the signal from the sensor should depend on the thickness of the conductive layer. Based on the monitoring, the control parameters for grinding can be adjusted, for example, so that the positions of the layer have substantially the same thickness after grinding, or so that the grinding of the positions of the layer is completed at about the same time. This profile control can be referred to as real-time profile control (RTPC). In addition, the grinding operation can be terminated based on an indication that the monitored thickness has reached a desired endpoint thickness.
[0027] In-situ monitoring systems may suffer from signal distortion for measurements taken at locations close to the edge of a substrate. For example, an inductive monitoring system may generate a magnetic field. Close to the edge of the substrate, the signal may be artificially reduced because the magnetic field only partially overlaps the conductive layer of the substrate. Various techniques may be used to compensate for the distortion. For example, the signal may be fed into an artificial neural network to generate a modified signal.
[0028] In practice, the magnetic field generated by the eddy current sensor does not stop within the conductive layer, but may extend into the underlying substrate. Without being bound by any particular theory, for the electromagnetic frequencies employed in the eddy current sensor, the skin depth in these magnetically permeable materials may be greater than the thickness of the conductive layer and the underlying semiconductor wafer. Therefore, the signal generated by the eddy current sensor may depend on the conductivity of the semiconductor wafer.
[0029] If the semiconductor wafer is undoped, for example, as is often the case in "blank" wafers and base substrate wafers used for system calibration, the resistance of the wafer may be high enough so that the presence of the wafer has no detectable effect on the eddy current signal. However, for actual device manufacturing, the wafer will typically be doped for various purposes, for example, highly doped. In this case, depending on the conductivity of the semiconductor wafer, the signal generated by the eddy current sensor may have a significant contribution from the wafer. As such, the thickness measurement based on the signal collected by the eddy current sensor may be inaccurate. Techniques can be used to compensate for this inaccuracy, for example, by taking into account the contribution to the signal from the semiconductor wafer. However, when edge reconstruction techniques are utilized, these compensations may introduce additional errors at the substrate edge.
[0030] However, via an edge reconstruction algorithm, the trace from the substrate and the trace from the doped wafer can be run separately. The modified doped wafer trace can be subtracted from the modified measured substrate trace; the resulting difference will be closer to the actual thickness of the layer being ground. Additionally, the difference can be scaled to compensate for the sensor configuration.
[0031] Figure 1 and Figure 2 An example of a polishing station 20 of a chemical mechanical polishing system is depicted. The polishing station 20 includes a rotatable disc-shaped platform 24 on which a polishing pad 30 is located. The platform 24 is operable to rotate 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 two-layer polishing pad having an outer polishing layer 34 and a softer backing layer 32.
[0032] The polishing station 30 may include a supply port or combined supply rinse arm 39 to dispense a polishing liquid 38, such as abrasive slurry, onto the polishing pad 30. The polishing station 20 may include a pad conditioning device having a conditioning disk to maintain the surface roughness of the polishing pad.
[0033] The carrier head 70 is operable to hold the substrate 10 against the polishing pad 30. The carrier head 70 is suspended from a support structure 72 (e.g., a turntable or track) and is connected to a carrier head rotation motor 76 by a drive shaft 74 so that the carrier head can rotate about an axis 71. Optionally, the carrier head 70 can oscillate laterally by moving along a track or by rotational oscillation of the turntable itself, for example, on a slider on the turntable.
[0034] The carrier head 70 can include a retaining ring 84 to hold the substrate. In some implementations, the retaining ring 84 can include a highly conductive portion, for example, the carrier head can include a thin lower plastic portion 86 that contacts the polishing pad, and a thick upper conductive portion 88. In some implementations, the highly conductive portion is metal, for example, the same metal as the layer being polished, such as copper.
[0035] In operation, the platform rotates about its central axis 25, and the carrier head rotates about its central axis 71 and translates across the top surface of the polishing pad 30. Where multiple carrier heads are present, each carrier head 70 can have independent control over its polishing parameters, e.g., each carrier head can independently control the pressure applied to each respective substrate.
[0036] The carrier head 70 may include a flexible membrane 80 having a substrate mounting surface to contact the backside of the substrate 10 , and a plurality of pressurizable chambers 82 for applying different pressures to different regions (eg, different radial regions) on the substrate 10 .
[0037] In some implementations, the grinding station 20 includes a temperature sensor 64 to monitor the temperature of the grinding station or a component of the grinding station / components in the grinding station. Figure 1 1 and 2. The temperature sensor 64 is shown as being positioned to monitor the temperature of the polishing pad 30 and / or the slurry 38 on the pad 30, but the temperature sensor 64 can be positioned inside the carrier head 70 to measure the temperature of the substrate 10. The temperature sensor 64 can be in direct contact (i.e., a contact sensor) with the outermost layer of the polishing pad or substrate 10 (which can be a conductive layer) to accurately monitor the temperature of the outermost layer of the polishing pad or substrate. The temperature sensor can also be a non-contact sensor (e.g., an infrared sensor). In some implementations, multiple temperature sensors are included in the polishing station 22 to, for example, measure the temperature of different components of / in the polishing station. The temperature(s) can be measured in real time, for example, periodically and / or in association with real-time measurements made by an eddy current system. The monitored temperature(s) can be used to adjust the eddy current measurements in situ.
[0038] refer to Figure 3A , the grinding system can be used to grind a substrate 10 including a conductive material covering a patterned dielectric layer and / or embedded in a patterned dielectric layer. For example, the substrate 10 may include a layer of conductive material 16 (e.g., a metal such as copper, aluminum, cobalt, or titanium) covering and filling a trench in a dielectric layer 14 (e.g., silicon oxide or a high-k dielectric). Optionally, a barrier layer 18 (e.g., tantalum or tantalum nitride) may line the trench and separate the conductive material 16 from the dielectric layer 14. The conductive material 16 in the trench may provide a via, pad, and / or interconnect in a completed integrated circuit. Although the dielectric layer 14 is depicted as being deposited directly on the semiconductor wafer 12, one or more other layers may be interposed between the dielectric layer 14 and the wafer 12.
[0039] Semiconductor wafer 12 may be a silicon wafer, e.g., single crystal silicon, although other semiconductor materials are possible. Additionally, semiconductor wafer 12 may be doped, e.g., with p-type or n-type doping. Doping may be uniform across the wafer, or the wafer may be selectively doped, e.g., suitable for use in fabricating transistors in integrated circuits using the semiconductor wafer.
[0040] Initially, the conductive material 16 covers the entire dielectric layer 14. As the grinding progresses, the bulk of the conductive material 16 is removed, thereby exposing the barrier layer 18 (see Figure 3B ). Continued polishing then exposes the patterned top surface of the dielectric layer 14 (see Figure 3C ). Additional grinding may then be used to control the depth of the trench containing the conductive material 16.
[0041] In some implementations, the grinding system includes additional grinding stations. For example, the grinding system may include two or three grinding stations. For example, the grinding system may include a first grinding station having a first electromagnetic induction monitoring system and a second grinding station having a second electromagnetic induction current monitoring system.
[0042] For example, in operation, bulk grinding of a conductive layer on a substrate may be performed at a first grinding station, and grinding may be paused when a target thickness of the conductive layer remains on the substrate. The substrate is then transferred to a second grinding station, and the substrate may be ground down to an underlying layer (e.g., a patterned dielectric layer).
[0043] return Figure 1 , the grinding system may include an in-situ electromagnetic induction monitoring system 100, which may be coupled to the controller 90 or be considered to include the controller 90. The rotation coupler 29 can be used to electrically connect components in the rotatable platform 24 (e.g., sensors of the in-situ monitoring system) to components outside the platform, such as the drive and sensing circuit system or the controller 90.
[0044] The in-situ EM monitoring system 100 is configured to generate a signal dependent on the depth of the conductive material 16 (eg, metal). The EM monitoring system may operate by generating eddy currents in a sheet of conductive material overlying a dielectric layer or by generating currents in conductive loops in trenches formed in a dielectric layer on a substrate.
[0045] As an eddy current monitoring system, the electromagnetic induction monitoring system 100 can be used to monitor the thickness of a conductive layer by inducing eddy currents in a conductive sheet. Alternatively, as an induction monitoring system, the electromagnetic induction monitoring system can operate by generating a current in a conductive loop formed in a dielectric layer 14 of the substrate 10 for monitoring purposes (e.g., as described in U.S. Patent Publication No. 2015-0371907).
[0046] In operation, the grinding system can use the in-situ monitoring system 100 to determine when the conductive layer has reached a target thickness (e.g., a target depth of metal in a trench or a target thickness of a metal layer covering a dielectric layer) and then pause grinding. Alternatively or additionally, the grinding system can use the in-situ monitoring system 100 to determine differences in the thickness of the conductive material across the substrate 10 and use this information to adjust the pressure in one or more chambers 82 in the carrier head 80 during grinding in order to reduce grinding non-uniformities.
[0047] A groove 26 may be formed in the platform 24, and optionally a thin section 36 may be formed in the polishing pad 30 covering the groove 26. The groove 26 and thin section 36 may be positioned so that the groove 26 and thin section 36 pass under the substrate 10 during a portion of the platform's rotation regardless of the translational position of the carrier head. Assuming the polishing pad 30 is a two-layer pad, the thin section 36 may be constructed by removing a portion of the backing layer 32 and optionally by forming a groove in the bottom of the polishing layer 34. The thin section may optionally be light transmissive, for example, when an in-situ optical monitoring system is integrated into the platform 24.
[0048] The in-situ monitoring system 100 may include a sensor 102 mounted in the recess 26. The sensor 102 may include a magnetic core 104 positioned at least partially in the recess 26, and at least one coil 106 wound around a portion of the magnetic core 104. A drive and sense circuit system 108 is electrically connected to the coil 106. The drive and sense circuit system 108 generates signals that may be sent to the controller 90. Although depicted as being external to the platform 24, some or all of the drive and sense circuit system 108 may be mounted in the platform 24.
[0049] refer to Figure 1 and Figure 4 , the drive and sense circuit system 108 applies an AC current to the coil 106, which generates a magnetic field 150 between the two poles 152a and 152b of the coil 104. Figure 4 A C-shaped core is depicted, but other cores are possible, for example, E-shaped, I-shaped, etc. In operation, when the substrate 10 intermittently covers the sensor 102 , a portion of the magnetic field 150 extends into the substrate 10 .
[0050] The circuit system 108 may include a capacitor connected in parallel with the coil 106. The coil 106 and the capacitor together may form an LC resonant tank.
[0051] If the thickness of a conductive layer on a substrate needs to be monitored, when the magnetic field 150 reaches the conductive layer, it can pass through and generate a current (if the target is a loop) or induce eddy currents (if the target is a sheet). This modifies the effective impedance of the LC circuit.
[0052] However, the magnetic field 150 may also penetrate into the semiconductor substrate 12. As such, the effective impedance of the LC circuit, and therefore the signal from the drive and sense circuitry 108, may also depend on the doping of the semiconductor substrate 12 and the resulting conductivity.
[0053] The drive and sense circuit system 108 may include a marginal oscillator coupled to the combined drive / sense 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, for example, as described in U.S. Pat. No. 7,112,960. Other configurations are possible for the drive and sense circuit system 108. For example, separate drive and sense coils may be wound around a magnetic core. The drive and sense circuit system 108 may apply current at a fixed frequency, and the signal from the drive and sense circuit system 108 may be a phase shift of the current in the sense coil relative to the drive coil, or the amplitude of the sensed current, for example, as described in U.S. Pat. No. 6,975,107.
[0054] refer to Figure 2 , as the platform 24 rotates, the sensor 102 is swept under the substrate 10. By sampling the signal from the circuit system 108 at a particular frequency, the circuit system 108 generates a series of measurements at the sampling regions 94 across the substrate 10. For each sweep, the measurements at one or more of the sampling regions 94 may be selected or combined. Thus, over multiple sweeps, the selected or combined measurements provide a time-varying sequence of values.
[0055] The polishing station 20 may also include a position sensor 96, such as an optical interrupter, to sense when the sensor 102 is under the substrate 10 and when the sensor 102 is away from the substrate. For example, the position sensor 96 may be mounted at a fixed position relative to the carrier head 70. Markers 98 may be attached to the periphery of the platform 24. The attachment point and length of the mark 98 are selected so that the mark 98 can signal the position sensor 96 when the sensor 102 is swept under the substrate 10.
[0056] Alternatively or additionally, the grinding station 20 may include an encoder to determine the angular position of the platform 24 .
[0057] return Figure 1 , a controller 90 (e.g., a general purpose programmable digital computer) receives a signal from a sensor 102 of an in-situ monitoring system 100. As the sensor 102 sweeps under the substrate 10 with each rotation of the platform 24, information about the depth of the conductive layer (e.g., a bulk layer or conductive material in a trench) is accumulated in-situ (once per platform rotation). The controller 90 can be programmed to sample measurements from the in-situ monitoring system 100 when the substrate 10 substantially covers the sensor 102.
[0058] In addition, the controller 90 can be programmed to calculate the radial position of each measurement and to sort the measurements into multiple radial ranges. By arranging the measurements into multiple radial ranges, data about the thickness of the conductive film in each radial range can be fed into a controller (e.g., the controller 90) to adjust the grinding pressure profile applied by the carrier head. The controller 90 can also be programmed to apply endpoint detection logic to the measurement sequence generated by the in-situ monitoring system 100 signal and detect the grinding endpoint.
[0059] Information about the thickness of the conductive layer is accumulated in situ and continuously in real time as the sensor 102 sweeps beneath the substrate 10 with each rotation of the platform 24. During polishing, measurements from the sensor 102 can be displayed on an output device to permit an operator of the polishing station to visually monitor the progress of the polishing operation.
[0060] refer to Figure 2 and Figure 5 , changes in the position of the sensor head relative to the substrate 10 may result in changes in the signal from the in-situ monitoring system 100. That is, as the sensor head scans across the entire substrate 10, the in-situ monitoring system 100 will measure multiple regions 94 (e.g., measurement points 211) at different locations on the substrate 10. The regions 94 may partially overlap.
[0061] Figure 6 A graph is depicted showing a signal 220 from the in-situ monitoring system 100 during a single pass of the sensor 102 under the substrate 10. This signal 220 may be referred to as a "trace" across the substrate. The signal 220 is composed of a series of individual measurements from the sensor head as the sensor head is swept under the substrate. The graph may be a function of measurement time or measurement position (e.g., radial position) on the substrate. In either case, different portions of the signal 220 correspond to measurement points 211 at different locations on the substrate 10 scanned by the sensor 102. Thus, the graph depicts corresponding measured signal values from the signal 220 for a given location of the substrate scanned by the sensor head.
[0062] refer to Figure 5 and Figure 6 , the signal 220 includes a first portion 222 corresponding to a position in the edge region 203 of the substrate 10 when the sensor 102 traverses the leading edge of the substrate 10; a second portion 224 corresponding to a position in the central region 201 of the substrate 10; and a third portion 226 corresponding to a position in the edge region 203 when the sensor 102 traverses the trailing edge of the substrate 10. The signal may also include a portion 228 corresponding to an off-substrate measurement, i.e., a position measured while the sensor head is scanning. Figure 5The signal generated when the area other than the edge 204 of the substrate 10 is in the substrate 10.
[0063] The edge region 203 may correspond to a portion of the substrate where a measurement point 211 of the sensor head overlaps the substrate edge 204. The center region 201 may include an annular anchor region 202 adjacent to the edge region 203, and an inner region 205 surrounded by the anchor region 202. The sensor head may scan such regions on its path 210 and generate a sequence of measurements corresponding to a series of positions along the path 210.
[0064] In the first portion 222, the signal strength gradually rises from an initial strength (typically the signal generated when there is no substrate and no carrier head) to a higher strength. This is due to the transition of the monitoring position from initially only slightly overlapping the substrate at the edge 204 of the substrate (producing an initial lower value) to a monitoring position that almost completely overlaps the substrate (producing a higher value). Similarly, in the third portion 226, the signal strength gradually decreases as the monitoring position transitions to the edge 204 of the substrate.
[0065] Although the second portion 224 is depicted as flat, this is for simplicity, and the real signal in the second portion 224 will likely include fluctuations due to noise and variations in layer thickness. The second portion 234 corresponds to the monitoring position of the scan center region 201. The second portion 224 includes two sub-portions 230 and 232, which are caused by the monitoring position of the anchor region 202 of the scan center region 201; and a sub-portion 234, which is caused by the monitoring position of the inner region 205 of the scan center region 201.
[0066] As described above, the variation in signal strength in regions 222, 226 is caused in part by the overlap of the measurement region of sensor 106 with the substrate edge, rather than by inherent variations in the thickness or conductivity of the layer being monitored. Thus, such distortion in signal 220 may cause errors in calculating characteristic values of the substrate (e.g., the thickness of a layer near the substrate edge). To address this problem, controller 90 may include a neural network (e.g., Figure 7 A neural network 300 is provided to generate modified signals corresponding to one or more locations of the substrate 10 based on measured signals corresponding to those locations.
[0067] Reference Figure 7, the neural network 300 is configured to generate a modified signal when properly trained that reduces and / or removes distortion of the calculated signal values near the edge of the substrate. The neural network 300 receives a set of inputs 304 and processes the inputs 304 via one or more neural network layers to generate a set of outputs 350. The layers of the neural network 300 include an input layer 310, an output layer 330, and one or more hidden layers 320.
[0068] Each layer of neural network 300 includes one or more neural network nodes. Each neural network node in a neural network layer receives one or more node input values (either from input 304 to neural network 300 or from outputs of one or more nodes of a previous neural network layer), processes the node input values according to one or more parameter values to produce activation values, and optionally applies a nonlinear transformation function (e.g., a sigmoid or tanh function) to the activation values to produce an output of the neural network node.
[0069] Each node in the input layer 310 receives one of the inputs 304 to the neural network 300 as a node input value.
[0070] The input 304 to the neural network includes measured signal values from the in-situ monitoring system 100 for a plurality of different points 211 on the substrate 10, such as a first measured signal value 301, a second measured signal value 302, through an nth measured signal value 303. The measured signal values may be individual values in a sequence of values in the signal 220.
[0071] In general, the plurality of different locations includes locations in edge region 203 and (optionally) anchor region 202 of substrate 10. In some implementations, the plurality of different locations is only in edge region 203 and anchor region 202. In other implementations, the plurality of different locations spans all regions of the substrate.
[0072] Such measured signal values are received at signal input nodes 344. Optionally, the input nodes 304 of the neural network 300 may also include one or more state input nodes 316 that receive one or more process state signals 304, such as a wear metric of the pad 30 of the grinding device 20.
[0073] The nodes of hidden layer 320 and output layer 330 are depicted as receiving input from each node of the previous layer. This is the case in a fully connected feedforward neural network. However, neural network 300 may be a non-fully connected feedforward neural network or a non-feedforward neural network. Furthermore, neural network 300 may include at least one of one or more fully connected feedforward layers; one or more non-fully connected feedforward layers; and one or more non-feedforward layers.
[0074] The neural network generates a set of modified signal values 350 at the nodes of the output layer 330 (i.e., “output nodes” 350). In some implementations, there is an output node 350 for each measured signal from the in-situ monitoring system, which is fed to the neural network 300. In this case, the number of output nodes 350 may correspond to the number of signal input nodes 304 of the input layer 310.
[0075] For example, the number of signal input nodes 344 may be equal to the number of measurements in the edge region 203 and the anchor region 202, and there may be an equal number of output nodes 350. Thus, each output node 350 generates a modified signal corresponding to a respective measured signal supplied as input to the signal input node 344, e.g., a first modified signal 351 for the first measured signal 301, a second modified signal 352 for the second measured signal 302, and an nth modified signal 353 for the nth measured signal 303.
[0076] In some implementations, the number of output nodes 350 is less than the number of input nodes 304. In some implementations, the number of output nodes 350 is less than the number of signal input nodes 344. For example, the number of signal input nodes 344 may be equal to the number of measurements in the edge region 203, or equal to the number of measurements in the edge region 203 and the anchor region 202. Again, each output node 350 of the output layer 330 generates a modified signal corresponding to a corresponding measured signal supplied as a signal input node 304, e.g., a first modified signal 351 for the first measured signal 301, but only for a signal input node 354 that receives a signal from the edge region 203.
[0077] The polishing device 100 can generate a modified signal using the neural network 300. The modified signal can then be used to determine the thickness of each location in the first set of locations of the substrate, such as locations in the edge region (and possibly in the anchor region). For example, referring back to Figure 6 , the modified signal value of the edge region may provide a modified portion 230 of the signal 220 .
[0078] In some implementations, for a modified signal value corresponding to a given measured location, neural network 500 can be configured such that only input signal values from measured locations within a predetermined distance of the given location are used to determine the modified signal value.
[0079] To train the neural network, a profile of a reference substrate may be generated using the sensors 102 of the in-situ monitoring system 100. Additionally, ground truth measurements of the thickness of the reference substrate may be obtained; such measurements may be performed for locations to be processed by the neural network. The system may generate ground truth measurements of thickness using an electrical impedance measurement method such as a four-point probing method. Signal values from the reference substrate are applied to input 304, while the ground truth measurements are applied to output 350, and the system is run in a training mode (e.g., gradient descent with back-propagation).
[0080] The reference substrate may include a blank undoped wafer on which a uniform thickness of conductive material is deposited. The amount of conductive material may be selected to simulate the presence of a doped wafer.
[0081] The reference substrate may also include a sample device substrate at an equivalent processing stage as the device substrate for which the in-situ monitoring system is to be used to control the polishing of, for example, a substrate having a layer with a different edge profile.
[0082] As mentioned above, the signal generated by the in-situ monitoring system also includes contributions from the doped wafer. If not handled properly, attempting to compensate for the contribution to the signal from the doped wafer may introduce additional errors, for example, at the substrate edge when edge reconstruction techniques are used.
[0083] refer to Figure 8 , generating a reference trace 420 across a blank undoped wafer. This reference trace 420 is generated before grinding the substrate. The blank undoped wafer has the same doping profile as the wafer that will be used in the device substrate to be ground. In some implementations, the reference trace is generated by scanning a sample blank undoped wafer (e.g., a sacrificial wafer) with the sensor 102 of the in-situ monitoring system 100. For example, the reference trace can be generated by a wafer factory operator. Alternatively, the system manufacturer can generate reference traces for wafers with a variety of different doping (e.g., concentrations and / or doping materials), and these traces can be stored in a library. The operator can then select one of the reference traces from the library (e.g., from a drop-down menu or similar user interface) that most closely corresponds to the doping of the wafer in the device substrate to be ground.
[0084] A correlation curve may be used to convert raw signal values in reference trace 420 from sensor 102 into thickness values (represented by reference trace 420').
[0085] Fig. 9 A correlation curve 510 is shown between the thickness of a conductive layer of a given resistivity and the signal from the electromagnetic induction monitoring system 100 for a given resistivity. 开始 represents the initial thickness of the conductive layer, S 开始 Corresponding to the initial thickness D 开始The required signal value; D 最终 represents the final thickness of the conductive layer, and S 最终 is the desired signal value corresponding to the final thickness; and K is a constant representing the signal value for zero conductive layer thickness.
[0086] The relationship curve 510 can be represented by a function in the controller 90, such as a polynomial function, such as a second-order function, a third-order function, or a higher-order function. The correlation between the signal X(x) and the thickness D(x) can be represented by the following equation:
[0087] X(x)=W1·D(x) 2 +W2·D(x)+W3 (Equation 1)
[0088] Where W1, W2 and W3 are real coefficients. Therefore, the controller can store the values of the coefficients of the function, such as the values of W1, W2 and W3, and the value of the resistivity ρ0 to which the relationship curve 510 applies. In addition, the relationship can be represented by a linear function (Bezier curve) or a non-polynomial function (such as exponential or logarithmic).
[0089] The relationship curve 510 can be used to convert the signal values in the raw signal 420 from the reference wafer into "equivalent" thickness measurements. That is, the measurements can be expressed as thickness values despite the fact that there is no conductive layer on top of the doped reference wafer. These are "equivalent" thickness values because each is the thickness of an equivalent conductive layer on an undoped wafer that would produce the same signal as the doped reference wafer.
[0090] refer to Figure 8 , the reference trace 420 ′ is then processed by a neural network as if it were a normal signal to perform an edge reconstruction algorithm on the reference trace. This operation produces a portion of a modified reference trace 450 having a modified signal value 430 .
[0091] In some implementations, the conversion to thickness is performed beforehand, and stored in the library (and selected by the operator) is the reference trace 420' with the thickness value. In some implementations, the thickness conversion and edge reconstruction are performed beforehand, and stored in the library (and selected by the operator) is the modified reference trace 450.
[0092] During the polishing operation, the substrate 10 is monitored by the in-situ monitoring system and a measured trace 220 of the substrate 10 is generated for each sweep of the sensor 102 across the substrate 10. This measured trace 220 may also be referred to as a "total" trace or signal because it includes contributions from the conductive layer being polished as well as the underlying doped wafer.
[0093] Relationship curve 510 (see Fig. 9) may be used to convert the signal values in the signal 220 from the polished substrate into a thickness measurement (represented by the measured trace 220').
[0094] Each measured trace 220 ′ is processed by a neural network as described above to produce a portion of a modified measured trace 250 having a modified value 230 .
[0095] In some implementations, the conversion from raw signal to thickness can be performed for both the reference wafer and the substrate being polished after edge reconstruction is performed.
[0096] To compensate for wafer doping, the controller 190 may generate an adjusted trace 480. Generating the adjusted trace includes subtracting the modified reference trace 450 from the modified measured trace 250. Assuming that the modified reference trace 450 is represented by S(x) and the modified measured trace 250 is represented by T(x), where x is the radial position, then T(s)-S(x) provides an apparent thickness trace.
[0097] For some configurations of sensor 102, the contributions to the traces from the doped wafer and substrate are not simply additive. Rather, the apparent thickness of the conductive layer may be slightly less than the actual thickness. This problem may become more pronounced at higher drive frequencies.
[0098] However, any particular sensor configuration (e.g., drive frequency, core shape and size, coil winding location and number, etc.) does have a generally linear relationship between actual thickness and apparent thickness. Fig.10 This relationship is plotted in . Function 520 relating apparent thickness to actual thickness can be represented as a linear function with a slope of k and a y-intercept of b (where thickness should be zero). These values of k and b can be determined empirically through testing and will vary between different sensor configurations. Values of k tend to be less than or equal to 1, for example, values from 0.7 to 0.
[0099] Therefore, the adjusted film thickness curve A(x) of the conductive layer on the substrate may be calculated according to A(x)=(T(x)-S(x)-b) / k.
[0100] When the adjusted thickness value A(x) reaches the target thickness value D 目标 Similarly, the adjusted thickness value A'(x) can be used to control grinding parameters, for example, to calculate grinding pressure to reduce non-uniformity.
[0101] In some cases, the relationship between the apparent thickness and the actual thickness of a particular sensor configuration may not be linear. In this case, a more complex equation (e.g., a polynomial) may be used to calculate the actual thickness.
[0102] In some implementations, the raw signal is normalized before conversion to a thickness value. This technique is applicable to reference trace 420 as well as substrate trace 220. For example, the calibration signal X'(x) may be generated according to the following equation:
[0103] X'(x) = G*X(x) - ΔK (Equation 2)
[0104] Where G is the gain, and ΔK is the offset, but it is determined experimentally for an in-situ monitoring system using a blank wafer with a conductive layer of known thickness and conductivity. X(x) represents the raw signal value, e.g., from the reference trace 420 or substrate trace 220, suitable for processing the corresponding trace. The calibration signal X'(x) is then used in the correlation curve, e.g., in place of X(x) in Equation 1 above, to determine the thickness value.
[0105] Additionally, the resistivity of the layer may be taken into account during the conversion of the raw signal value to the thickness value. For example, the thickness value calculated using the correlation curve (e.g., Equation 1 above) may be adjusted based on the resistivity of the layer to provide a corrected thickness value. This technique may be used for both the reference trace 420 as well as the substrate trace 220.
[0106] The corrected thickness value D'(x) can be calculated as follows:
[0107] D'(x)=D(x)*(ρ X / ρ0)(Equation 3)
[0108] where ρ X is the resistivity of the conductive layer, and ρ0 is the resistivity applying the correlation curve 410 (and the values W1, W2, W3), and where D(x) represents the initial thickness value calculated using the correlation curve (appropriately from the reference trace 420 or the substrate trace 220). An edge reconstruction algorithm may be applied to the corrected thickness value D'(x) instead of the initial thickness value D(x).
[0109] In addition to resistivity variations between substrates, changes in layer temperature can result in changes in the resistance of the conductive layer. For example, the conductive layer may become hotter and therefore more conductive (lower resistivity) as grinding proceeds. Specifically, the controller performing the process may also calculate the resistivity p of the conductive layer at the real-time temperature T(t) T The real-time temperature T(t) may be determined from the temperature sensor 64. In some implementations, the adjusted resistivity p is calculated based on the following equation: T :
[0110] ρ T =ρ X [1+α(T(t)-T ini )](Equation 4)
[0111] Where T ini is the initial temperature of the conductive layer at the beginning of the grinding process. Then replace the resistivity ρ X (For example, in Equation 3 above (or in calculating gain and offset in Equation 2)), using the adjusted resistivity p T .
[0112] In the case where the grinding process is carried out at room temperature, T ini A value of about 20°C can be used, ρ X For the conductive layer at T ini Under the resistivity, the T ini It can be room temperature. Typically, α is a known value that can be found in the literature or obtained experimentally. Although the original signal 220 includes contributions from the underlying doped wafer, the value of α for the conductive layer can be used as a first approximation when calculating the thickness value of the trace 220 ′.
[0113] In some implementations, the temperature T and T ini is the temperature of the conductive layer, for example, as measured by a temperature sensor in the carrier head. ini It may be the temperature of the polishing pad or the temperature of the slurry instead of the temperature of the conductive layer.
[0114] The above-described polishing apparatus and method can be applied to a variety of polishing systems. The polishing pad or the carrier head or both can be moved to provide relative motion between the polishing surface and the substrate. For example, the platform can move around a track instead of rotating. The polishing pad can be a circular (or some other shape) pad fastened to the platform. Some aspects of the endpoint detection system can be applicable to linear polishing systems, for example, where the polishing pad is a roll-to-roll conveyor belt that moves continuously or linearly. The polishing layer can be a standard (e.g., polyurethane with or without fillers) abrasive material, a soft material, or a fixed abrasive material. The term relative positioning is used to represent relative positioning within a system or substrate; it should be understood that during the polishing operation, the polishing surface and the substrate can be held in a vertical orientation or some other orientation.
[0115] The functional operation of the controller 90 may be implemented in a digital electronic circuit system, or in computer software, firmware or hardware and their structural equivalents or in a combination thereof. Computer software may be implemented as one or more computer program products, i.e., one or more computer programs tangibly embodied in a non-transitory computer-readable storage medium, or used to control the operation of a data processing device (e.g., a programmable processor, a computer, or multiple processors or computers). Computer programs (also referred to as programs, software, software applications, or codes) may be written in any form of programming language including compiled or interpreted languages, and the computer program may be adopted in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program does not necessarily correspond to a file. A program may be stored in a portion of a file that stores other programs or data, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or portions of code). A computer program may be deployed to be executed on one or more computers at one site, or distributed on multiple sites and interconnected by a communication network.
[0116] The processes and logic flows described in this specification may be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows may also be performed by, and the apparatus may also be implemented as, a special purpose logic circuit system, such as an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).
[0117] Many embodiments of the present invention have been described. However, it will be appreciated that various modifications may be made without departing from the spirit and scope of the present invention. For example, although the above description focuses on chemical mechanical polishing, the control system may be applicable to other semiconductor processing techniques, such as etching or deposition, such as chemical vapor deposition. In addition, the technology may be applied to in-line or stand-alone metrology systems rather than in-situ monitoring. Therefore, other embodiments are within the scope of the appended claims.
Claims
1. A method for grinding a substrate, comprising: storing or generating a first trace representing measurements of a bare doped reference semiconductor wafer by an in-situ electromagnetic induction monitoring system; contacting a substrate having a conductive layer disposed over a semiconductor wafer with a polishing pad; generating a relative motion between the substrate and the polishing pad; monitoring the substrate using the in-situ electromagnetic induction monitoring system while polishing the conductive layer to produce a measured trace dependent on the thickness of the conductive layer; generating an adjusted trace using the measured trace, the first trace, and a neural network configured to reduce distortion of calculated signal values near an edge of the substrate, the adjusted trace at least partially compensating for a contribution of conductivity of the semiconductor wafer to the measured trace and distortion of calculated signal values near an edge of the substrate; as well as At least one of pausing grinding or modifying grinding parameters based on the adjusted trace is performed. The method of claim 1 , wherein generating comprises subtracting the reference trace.
3. The method of claim 2, wherein the first trace represents a measurement of a bare doped reference semiconductor wafer modified by the neural network.
4. The method of claim 3, wherein generating comprises applying at least a portion of the measured trace to a neural network to generate a modified measured trace, and subtracting the first trace from the modified measured trace.
5. The method of claim 1, comprising: The bare doped reference semiconductor wafer is scanned using a sensor of an in-situ electromagnetic induction monitoring system to generate a preliminary reference trace having raw signal values, the raw signal values in the preliminary reference trace are converted into thickness values to generate an initial reference trace, and at least a portion of the initial reference trace is applied to the neural network to generate the first reference trace.
6. The method of claim 1, wherein the at least a portion of the trace applied to the neural network includes a portion corresponding to an edge region of the substrate.
7. The method of claim 1, wherein the step of generating the first trace comprises: A sensor of an in-situ electromagnetic induction monitoring system is scanned across the bare doped reference semiconductor wafer.
8. A computer program product tangibly embodied in a non-transitory computer readable medium, comprising instructions to cause one or more computers to: storing or generating a modified reference trace representing measurements of a bare doped reference semiconductor wafer by an in-situ electromagnetic induction monitoring system; receiving, from an in-situ electromagnetic induction monitoring system, a measured trace dependent on a thickness of a conductive layer on a substrate while the conductive layer is being polished; generating an adjusted trace using the measured trace, the first trace, and a neural network configured to reduce distortion of calculated signal values near an edge of the substrate, the adjusted trace at least partially compensating for a contribution of conductivity of the semiconductor wafer to the measured trace and distortion of calculated signal values near an edge of the substrate; as well as At least one of pausing grinding or modifying grinding parameters based on the adjusted trace is performed.
9. The computer program product of claim 8, wherein the instructions to generate the adjusted trace include instructions to subtract the reference trace.
10. The computer program product of claim 9, wherein the first trace represents a measurement of a bare doped reference semiconductor wafer modified by the neural network.
11. The computer program product of claim 10, wherein the instructions to generate the adjusted trace include instructions to apply at least a portion of the measured trace to a neural network to generate a modified measured trace, and to subtract the first trace from the modified measured trace.
12. The computer program product of claim 11, wherein the instructions to generate the adjusted trace include instructions for scaling a difference between the first trace and the modified measured trace.
13. The computer program product of claim 12, wherein the adjusted trace A(x) is calculated such that A(x)=(T(x)-S(x)-b) / k Wherein T(x) is the modified measured trace, S(x) is the modified reference trace, and b and k are constants.
14. The computer program product of claim 11, wherein the at least a portion of the measured trace applied to the neural network includes a portion corresponding to an edge region of the substrate.
15. A grinding system comprising: A support member, the support member is used to hold the polishing pad; a carrier head, the carrier head being used to hold a substrate in contact with the polishing pad; an in-situ electromagnetic induction monitoring system for monitoring the substrate while grinding a conductive layer on the substrate to produce a measured trace dependent on a thickness of the conductive layer; as well as A controller configured to storing or generating a first trace representing measurements of a bare doped reference semiconductor wafer by an in-situ electromagnetic induction monitoring system; receiving the measured trace from the in-situ electromagnetic induction monitoring system, generating an adjusted trace using the measured trace, the first trace, and a neural network configured to reduce distortion of calculated signal values near an edge of the substrate, the adjusted trace at least partially compensating for a contribution of conductivity of the semiconductor wafer to the measured trace and distortion of calculated signal values near an edge of the substrate; as well as At least one of pausing grinding or modifying grinding parameters based on the adjusted trace is performed.
16. The system of claim 15, wherein the controller is configured to generate the adjusted trace comprises subtracting the reference trace.
17. The system of claim 16, wherein the first trace represents a measurement of a bare doped reference semiconductor wafer modified by the neural network.
18. The system of claim 17, wherein the controller is configured to generate the adjusted trace by applying at least a portion of the measured trace to a neural network to generate a modified measured trace, and subtracting the first trace from the modified measured trace.
19. The system of claim 18, wherein the controller is configured to generate the adjusted trace by scaling a difference between the modified reference trace and the modified measured trace.
20. The system of claim 19, wherein the adjusted trace A(x) is calculated such that A(x)=(T(x)-S(x)-b) / k Wherein T(x) is the modified measured trace, S(x) is the modified reference trace, and b and k are constants.
