JP4854118B2 - Optical monitoring method in a two-stage chemical mechanical polishing process - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
Cross-reference for related applications
This application claims priority from US Patent Application No. 60 / 176,645, filed Jan. 18, 2000.
[0002]
The present invention relates generally to chemical mechanical polishing of substrates, and more particularly to a method and apparatus for detecting the end of a metal layer during a chemical mechanical polishing operation.
[0003]
[Background]
Integrated circuits are typically formed on a substrate by sequential deposition of conductive, semiconducting, or insulating layers on a silicon wafer. After each layer is deposited, the layer is etched to create circuit features. As a series of layers are sequentially deposited and etched, the outer or top surface of the substrate, i.e., the exposed surface of the substrate, becomes increasingly non-planar. This non-planar surface presents problems during the photolithographic stage of the integrated circuit manufacturing process. Therefore, there is a need to periodically planarize the substrate surface.
[0004]
Chemical mechanical polishing (CMP) is one accepted method of planarization. This planarization method usually requires that the substrate be mounted on a carrier head or polishing head. The exposed surface of the substrate is placed against a rotating polishing pad. The polishing pad can be either a “standard” pad or a fixed abrasive pad. Standard pads have a durable rough surface, whereas fixed abrasive pads have abrasive grains held in a containment medium. The carrier head provides a controllable load, i.e. pressure, that presses the substrate against the polishing pad. A polishing slurry comprising at least one chemically reactive agent and, if a standard pad is used, abrasive grains is supplied to the surface of the polishing pad.
[0005]
One problem in CMP is determining whether the polishing process is complete, i.e., whether the layer of the substrate has been planarized to the desired flatness or thickness. Variations in the initial thickness of the substrate layer, slurry composition, polishing pad condition, relative speed between the polishing pad and the substrate, and the load on the substrate can vary the material removal rate. This variation varies the time required to reach the polishing end point. Therefore, the polishing end point cannot simply be determined as a function of the polishing time.
[0006]
One way to determine the polishing end point is to remove the substrate from the polishing surface and inspect it. For example, the substrate can be transferred to a measuring station where the thickness of the substrate layer is measured, for example, by a profilometer or resistivity measurement. If the desired specifications are not met, the substrate is reloaded into the CMP apparatus for further processing. This is a time consuming procedure that reduces the throughput of the CMP apparatus. Instead, the inspection may reveal that an excessive amount of material has been removed, rendering the substrate unusable.
[0007]
Several methods have been developed for in-situ polishing end point detection. Most of these methods involve monitoring parameters associated with the substrate surface and indicating an end point when the parameters change suddenly. For example, if the insulating or dielectric layer is polished and the underlying metal layer is exposed, the coefficient of friction and reflectivity of the substrate will change rapidly as the metal layer is exposed.
[0008]
If the monitored parameter changes abruptly at the polishing end point, such an end point detection method is acceptable. However, as the substrate is being polished, the polishing pad condition and slurry composition at the pad-substrate interface can change. Such a change may mask the underlayer exposure, or the change may simulate an end condition. In addition, such end point detection methods are useful if only planarization is being performed, if the underlying layer is to be overpolished, or if the underlying layer and the top layer have similar physical properties. There will be no.
[0009]
SUMMARY OF THE INVENTION
In one aspect, the present invention is directed to a method for polishing a substrate. The first layer of the substrate is chemically mechanically polished with a first polishing fluid. The substrate has a second layer disposed below the first layer, and the first and second layers have different reflectivities. The substrate is optically monitored during polishing with the first polishing slurry to produce a plurality of intensity traces. Each intensity trace includes intensity measurements from different radial ranges of the substrate. After any of the intensity traces indicate an initial sweep of the first layer, the substrate is chemically mechanically polished with a second polishing fluid that has different polishing characteristics than the first polishing fluid. Optical monitoring of the substrate continues during polishing with the second polishing slurry, and polishing is stopped after all traces indicate that the second layer is fully exposed.
[0010]
Embodiments of the invention may include one or more of the following features. Optical monitoring directs the light beam through a window at the polishing surface, moves the light beam in a path across the substrate, monitors the reflectance signal created by the light beam reflected from the substrate, and reflectivity Extracting a plurality of intensity measurements from the signal may be included. Generating multiple intensity traces sorts each intensity measurement into one of the radial ranges according to the position of the light beam being measured, and determines the intensity trace from the intensity measurements associated with the radial range May include. The first slurry may be a high selectivity slurry and the second slurry may be a low selectivity slurry. The first layer may be more reflective than the second layer. The first layer can be a metal layer such as copper. The second layer may be an oxide layer such as silicon dioxide or a barrier layer such as tantalum or tantalum nitride.
[0011]
In another aspect, the present invention is directed to a method of polishing a substrate, wherein the surface of the substrate is contacted with a polishing surface having a window. The substrate has a first layer disposed over the second layer, and the first and second layers have different reflectivities. A first slurry is supplied to the substrate for the first polishing stage and a relative motion is caused between the substrate and the polishing surface. A light beam is directed through the window and movement of the polishing surface relative to the substrate moves the light beam in a path across the substrate. The reflectivity signal created by the light beam reflecting from the substrate is monitored, multiple intensity measurements are extracted from the reflectivity signal, and multiple intensity traces are obtained from different radial ranges on the substrate. Generated by each intensity trace that it contains. A second slurry is supplied to the substrate for polishing the second polishing stage when any of the intensity traces indicates an initial cleanup of the first layer. The second slurry has polishing characteristics that are different from the first slurry. Polishing is stopped after all intensity traces indicate that the second layer is fully exposed.
[0012]
Embodiments of the invention may include one or more of the following features. The first slurry may be a high selectivity slurry and the second slurry may be a low selectivity slurry. The second layer may be disposed on the third layer with the substrate.
[0013]
In another aspect, the present invention is directed to a method of polishing a substrate having a metal layer disposed on an oxide layer. In the method, the surface of the substrate is brought into contact with a polishing surface having a window. A high selectivity slurry is supplied to the polishing surface. Relative motion is caused between the substrate and the polishing surface. A light beam is directed through the window and movement of the polishing surface relative to the substrate moves the light beam in a path across the substrate. A reflectivity signal created with a light beam reflected from the substrate is monitored and a plurality of intensity measurements are extracted from the reflectivity signal. A radial position is determined for each intensity measurement, and the plurality of intensity measurements are classified into a plurality of radial ranges according to the radial position. A plurality of intensity traces are generated, each intensity trace including an intensity measurement from one of a plurality of radial ranges. When any of the intensity traces directs the initial cleanup of the metal layer, a low selectivity slurry is applied to the polishing surface, and polishing is performed when all intensity traces indicate that the oxide layer is fully exposed. Stopped.
[0014]
Embodiments of the invention may include one or more of the following features. A sudden drop in the reflectivity trace may indicate an initial sweep of the metal layer in the radial range associated with the reflectivity trace. The planarization of the reflectivity trace may indicate the exposure of the oxide layer in the radial range associated with the reflectivity trace. The substrate may include a barrier layer, such as tantalum or tantalum nitride, between a metal layer, such as copper, and an oxide layer, such as silicon dioxide.
[0015]
Advantages of the invention include one or more of the following. The copper coated substrate may be polished with reduced dishing and erosion. Both the polishing end point and the point at which the polishing apparatus should switch from a high selectivity slurry to a low selectivity slurry may be precisely determined. Detailed data is available on the progress of the metal polishing operation at different locations on the wafer.
[0016]
Other features and advantages of the invention will be apparent from the following description, including the drawings and the claims.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
With reference to FIGS. 1 and 2, one or more substrates 10 may be polished by a CMP apparatus 20. A description of a similar polishing apparatus 20 can be found in US Pat. No. 5,738,574, the entire disclosure of which is incorporated herein by reference. The polishing apparatus 20 includes a series of polishing stations 22 and a transfer station 23. The transfer station 23 performs a number of functions, receiving individual substrates 10 from a loading device (not shown), cleaning the substrates, loading substrates onto the carrier head, receiving substrates from the carrier head, substrates Cleaning the substrate and finally transferring the substrate back to the loading device.
[0018]
Each polishing station includes a rotatable platen 24 on which a polishing pad 30 is mounted. The first and second stations can include a two-layer polishing pad with a hard durable outer surface, whereas the final polishing station can include a relatively soft pad. If the substrate 10 is an "8 inch" (200 millimeter) or "12 inch" (300 millimeter) diameter disk, the platen and polishing pad will be about 20 inches or 30 inches in diameter, respectively. Each platen 24 may be connected to a platen drive motor (not shown). For most polishing processes, the platen drive motor rotates the platen 24 at about 30 to 200 revolutions per minute, although lower or higher rotational speeds may be used. Each polishing station can also include a pad conditioner device 28 that maintains the state of the polishing pad to effectively polish the substrate.
[0019]
The polishing pad 30 has a backing layer 32 that normally contacts the surface of the platen 24 and an overcoat layer 34 that is used to polish the substrate 10. The overcoat layer 34 is usually harder than the backing layer 32. However, some pads only have a backing layer and no backing layer. The overcoat layer 34 may be made of foamed polyurethane with open cells or polyurethane sheet having a grooved surface. The backing layer 32 may be composed of felt fibers leached with urethane and compressed. A two-layer polishing pad with an overlay layer composed of IC-1000 and a backing layer composed of SUBA-4 is available from Rodel, Inc., Newark, Delaware. (IC-1000 and SUBA-4 are product names of Rodel, Inc.)
A rotatable multi-head carousel 60 is supported on a central post 62 and rotated about a carousel shaft 64 by a carousel motor assembly (not shown). The center column 62 supports the carousel support plate 66 and the cover 68. The carousel 60 includes four carrier head systems 70. The center post 62 allows the carousel motor to rotate the carousel support plate 66 about the carousel shaft 64 to pivot the carrier head system and the substrate attached thereto. Three of the carrier head systems receive and hold the substrate and polish the substrate by pressing it against the polishing pad. On the other hand, one of the carrier head systems receives the substrate from the transfer station 23 and delivers the substrate to it.
[0020]
Each carrier head system includes a carrier or carrier head 80. A carrier drive shaft 74 connects a carrier head rotation motor 76 (shown with a quarter removal of the cover 68) to each carrier head 80 so that each carrier head 80 is independent about its own axis. And can be rotated. There is one carrier drive shaft and motor for each head. In addition, each carrier head 80 swings independently laterally within a radial slot 72 formed in the carousel support plate 66. A slider (not shown) supports each drive shaft in its associated radial slot. A radial drive motor (not shown) can swing the slider to swing the carrier head laterally.
[0021]
The carrier head 80 performs several mechanical functions. In general, the carrier head holds the substrate against the polishing pad during a polishing operation, distributes downward pressure evenly across the back of the substrate, transfers torque from the drive shaft to the substrate, and the substrate Ensure that the carrier head does not slide underneath.
[0022]
The carrier head 80 can include a flexible membrane 82 that provides a mounting surface for the substrate 10 and a retaining ring 84 that holds the substrate directly below the mounting surface. Pressurization of the chamber 86 defined by the flexible membrane 82 presses the substrate against the polishing pad. The retaining ring 84 may be formed of a highly reflective material or may be coated with a reflective layer to provide a reflective lower surface 88. A description of a similar carrier head 80 is given in US patent application Ser. No. 08 / 745,679, filed Nov. 8, 1996, by Steven M. Zuniga et al., Assigned to the assignee of the present invention. A carrier head with a flexible membrane for a chemical mechanical polishing system can be found in "A CARRIER HEAD WITH A FLEXIBLE MEMBRANE FOR A CHEMICAL MECHANICAL POLISHING SYSTEM", the entire disclosure of which is incorporated herein by reference.
[0023]
A slurry 38 containing a reactive agent (eg, deionized water for oxide polishing) and a chemically reactive catalyst (eg, potassium hydroxide for oxide polishing) is supplied to the surface of the polishing pad 30 as a slurry supply port or It may be supplied by a dual-purpose slurry / rinse arm 39. If the polishing pad 30 is a standard pad, the slurry 38 may also contain abrasive grains (eg, silicon dioxide for oxide polishing).
[0024]
In operation, the platen is rotated about its central axis 25 and the carrier head is rotated about its central axis 81 and is moved laterally across the surface of the polishing pad.
[0025]
A hole 26 is formed in the platen 24 and a transparent window 36 is formed in the portion of the polishing pad 30 above the hole. A transparent window 36 is disclosed in US patent application Ser. No. 08 / 689,930 filed Aug. 26, 1996 by Manoocher Birang et al., Assigned to the assignee of the present invention, entitled “Chemical Mechanical Polishing Device”. Can be constructed as described in `` METHOD OF FORMING A TRANSPARENT WINDOW IN A POLISHING PAD FOR A CHEMICAL MECHANICAL POLISHING APPARATUS ''. Incorporated herein. The holes 26 and the transparent window 36 are positioned to have a view of the substrate 10 during a portion of the platen rotation, regardless of the carrier head transition position.
[0026]
A reflectometer 40 is secured to the platen 24, generally beneath the hole 26, and rotates with the platen. The reflectometer includes a light source 44 and a detector 46. The light source generates a light beam 42 that propagates through the transparent window 36 and slurry 38 (see FIG. 3) and strikes the exposed surface of the substrate 10. For example, the light source 44 can be a laser and the light beam 42 can be a collimated laser beam. The optical laser beam 42 is angled from an axis perpendicular to the surface of the substrate 10, i. In addition, if the aperture 26 and window 36 are elongated, a beam expander (not shown) can be positioned in the path of the light beam, extending the light beam along the axis of the elongated window. . The laser 44 can operate continuously. Alternatively, the laser may be activated to generate the laser beam 42 during the time that the holes 26 are generally adjacent to the substrate 10.
[0027]
2 and 5A-5E, the CMP apparatus 20 may include a position sensor 160, such as an optical interrupter, to detect when the window 36 is near the substrate. For example, the optical breaker could be mounted at a fixed point opposite the carrier head 80. A flag 162 is attached to the outer periphery of the platen. The mounting point and length of the flag 162 is selected so that the flag blocks the optical signal of the sensor 160 from a time point just before the window 36 sweeps directly under the carrier head 80 to a time point slightly later. While the optical signal of sensor 160 is interrupted, the output signal from detector 46 may be measured and stored.
[0028]
In operation, the CMP apparatus 20 uses a reflectometer 40 to determine the amount of material removed from the surface of the substrate, or to determine when the surface has been planarized. A general purpose programmable digital computer 48 may be connected to the laser 44, the detector 46, and the sensor 160. The computer 48 activates the laser when the substrate is generally over the window, stores the intensity measurement from the detector, displays the intensity measurement on the output device 49, stores the intensity measurement, It may be programmed to sort the measured values into a radial range and detect the polishing end point.
[0029]
Referring to FIG. 3, the substrate 10 includes a silicon wafer 12 and an overlying metal layer 16 disposed over an oxide or nitride layer 14. The metal layer may be copper, tungsten, aluminum, among others. As different portions of the substrate with different reflectivities are polished, the signal output from detector 46 varies with time. Specifically, when the metal layer 16 is completely polished to expose the oxide or nitride layer 14, the reflectivity of the substrate decreases. The time-varying output of detector 46 may be referred to as an in situ reflectance measurement trace (or more simply, a reflectance trace). As will be discussed later, this reflectance trace may be used to determine the end point of the metal layer polishing operation.
[0030]
4 and 5A-5E, a measured reflectance trace with a transient intensity waveform 90 generated by polishing a metallized wafer is shown. The intensity waveform 90 is generated over a relatively long time scale (measured in seconds). The characteristic features of the waveform include an upper level plateau 97, each of which is surrounded by a left and right middle plateau 98. One period of waveform 90 includes left and right middle level plateau 98, one of upper level plateau 97, and background level 94.
[0031]
The middle plateau 98 represents reflection from the retaining ring 84, while the upper level plateau 97 represents reflection from the substrate 10. The background level represents scattered reflections from windows and slurries. The reflection from the retaining ring 84 is higher than the background level. As the substrate 10 is polished, the metal layer 16 is removed, and the underlying layer 14 is exposed, the end point waveform 90 falls toward or below the level of the intermediate plateau 98.
[0032]
With reference to FIG. 4 and FIGS. 5A-5E, the large scale configuration of the reflectance trace 90 can be described with reference to the angular position of the platen 24. Initially, the window 36 does not have a view of the substrate (see FIG. 5A). As a result, the laser beam 42 is not reflected and the intensity measured at the detector 46 is a result of background intensity including reflection from the slurry 38 and the transparent window 36. This low intensity corresponds to the background level 94. As the platen 24 rotates, the window 36 first sweeps under the retaining ring 84 of the carrier head 80 (see FIG. 5B). The lower surface 88 of the retaining ring 84 reflects a portion of the laser beam 42 into the detector 46 and creates an intermediate intensity measurement corresponding to the intermediate plateau 98. As the window 36 sweeps directly under the substrate 10 (see FIG. 5C), a portion of the laser beam 42 is reflected off the substrate. In general, the metal layer of the substrate 10 has a high reflectivity, resulting in an upper level plateau 97 on the reflectivity trace 90. As the platen continues to rotate, the window 36 again passes directly under the retaining ring 84 (see FIG. 5D). Finally, the window 36 sweeps off just below the carrier head 80 (see FIG. 5E) and the detector measures a low intensity corresponding to the background 94.
[0033]
The computer 48 of the CMP apparatus 20 can use the reflectance trace generated by the reflectometer 40 to determine the end point of the metal layer polishing operation. Each measurement may be performed at a plurality of radial positions. In addition, the computer 48 can use intensity measurements to determine substrate flatness and polishing uniformity for CMP tool and process qualification, as described below.
[0034]
Referring now to FIG. 6, the end point determination process is shown. Initially, several polishing parameters used during end point determination are stored in the memory of computer 48 (step 101). The polishing parameters of interest include the platen rotation rate and the carrier head sweep profile.
[0035]
The metal layer on the surface of the substrate 12 is polished by bringing the surface of the substrate into contact with the polishing pad 30 (FIG. 2) (step 102). The polishing pad 30 is rotated and causes a relative movement between the substrate and the polishing pad.
[0036]
Transient intensity data is monitored and collected for a plurality of sampled areas (step 104). This is done by directing the light beam generated by reflectometer 40 through the window. Movement of the polishing pad 30 relative to the substrate 12 moves the light beam in a path across the substrate surface. The reflection of the light beam from the substrate 10 and the retaining ring 84 is detected by a sensor, which generates reflection data related to the reflection of the light beam.
[0037]
The transient intensity data is displayed on the monitor (step 106), and the operator monitors the progress of the polishing operation. A pattern recognizer is applied to the transient intensity data to detect signal changes (step 108). The pattern recognizer can simply be a threshold detector that checks whether the intensity data has fallen below a predetermined threshold. Alternatively, in another embodiment, window logic can be applied to the data to detect a sequence of signal changes. Three types of window logic are used to detect local maxima and minima: window logic with a downward tip (cusp) that detects a downward trend in reflection data; with a forward tip that detects an upward trend in reflection data Window logic; and window logic with substantially flat lines that detect that the reflection data is relatively stationary. The signal changes may be averaged. Further discussion of pattern recognition algorithms for endpoint detection can be found in US patent application Ser. No. 08 / 689,930 referred to above.
[0038]
The output of the pattern recognizer is a stop signal, which is supplied to the polisher controller along with additional feedback data (step 110). The polisher controller uses feedback data to adjust various variables and parameters to minimize surface layer erosion and dishing. For example, polishing pressure, polishing rate, chemicals, and slurry composition may be developed to optimize overall polishing performance and / or polishing quality. The stop signal causes the polisher controller to stop the current metal layer polishing operation (step 112).
[0039]
In parallel with steps 106-112, the process of FIG. 6 stores transient intensity data on a data storage device, such as a computer disk, for subsequent processing (step 114). Briefly, the intensity for each sampling area is determined (step 116), the radial position of each sampling area is calculated (step 118), and the intensity measurements are sorted into the radial range (step 150). ). The sorted intensity measurements are used to measure polishing uniformity and removal rates at different radial ranges of the substrate (step 152). Each of these steps will be discussed in more detail later.
[0040]
In general, the reflected intensity varies during polishing for different radial locations on the substrate. The metal layer may be removed at different rates for different parts of the substrate. For example, the metal layer near the center of the substrate may be removed last, while the metal layer near the periphery or edge of the substrate may be removed first, or vice versa. Reflection data from the entire wafer is captured on the order of milliseconds on a relatively fine time scale and is available for experiments that improve the deposition process. By analyzing the recorded data, the process can be modified to make it faster, shorter and smoother. As can be appreciated, the stored data is useful for process research and development to optimize process performance.
[0041]
Referring to FIGS. 7A and 7B, the combination of platen rotation and carrier head linear sweep sweeps window 36 (and hence laser beam 42) across carrier head 80 and the bottom surface of substrate 10 in a sweep path 120. Let As the laser beam sweeps across the substrate, the reflectometer 40 integrates the measured intensity over a sampling period Tsample to produce a series of individual intensity measurements Ia, Ib,. The sampling rate F of the reflectometer 40 (the rate at which intensity measurements are generated) is given by F = 1 / Tsample. The reflectometer 40 can have a sampling rate between about 10 and 400 hertz (Hz), corresponding to a sampling period between about 2.5 and 100 milliseconds. In particular, reflectometer 40 may have a sampling rate of about 40 Hz and a sampling period of about 25 milliseconds.
[0042]
Thus, each time the laser 44 is activated, the reflectometer 40 measures the intensity from the plurality of sampling areas 122a-122j. Each sampling area corresponds to the area of the substrate over which the laser beam sweeps during the corresponding sampling period. In summary, at step 106, reflectometer 40 generates a series of intensity measurements Ia, Ib,..., Ij corresponding to sampled areas 122a, 122b,.
[0043]
Although FIG. 7A illustrates ten sampling zones, there may be more or fewer zones depending on the platen rotation rate and sampling rate. Specifically, a low sampling rate will result in a small and wide sampling area, whereas a high sampling rate will result in a large number of narrow sampling areas. Similarly, a low rotation rate will result in a large number of narrow sampling areas, whereas a high rotation rate will result in a small number of wide sampling areas. In addition, multiple detectors could be used to provide more sampling areas.
[0044]
As shown in FIG. 7B, the intensity measurements Ia and Ij for the sampling areas 122a and 122j, respectively, are low because the window 36 does not have a carrier head view and as a result the laser beam 42 is not reflected. Sampling areas 122b and 122i are located directly below retaining ring 84, and therefore intensity measurements Ib and Ii will be of intermediate intensity. Sampling areas 122c, 122d,..., 122h are located directly below the substrate, resulting in relatively large intensity measurements Ic, Id,..., Ih at various different radial locations across the substrate.
[0045]
FIG. 12 is an overlap of several transient signal diagrams 300-320. Each of the transient signal diagrams 300-320 represents intensity data over the interval associated with the sweep of the window directly below the carrier head. For example, diagram 300 shows end point data between about 1.7 and about 2.7 seconds, and diagram 320 shows end point data between about 350.8 seconds and about 351.8 seconds. Indicates. Of course, the transient signal diagram can be stored in the computer 48 for later reference.
[0046]
FIG. 12 shows how the reflected intensity signal at the end point changes during the polishing operation. Initially, in period 300, the metal layer on the surface of substrate 10 is jagged. The metal layer 16 has a fairly initial topography due to the topology of the underlying patterned layer 14. Because of this unevenness, the light beam is scattered when it strikes the metal layer. As the polishing operation proceeds, the metal layer becomes more planar and the reflectivity of the polished metal layer increases during periods 302-308. Thus, the signal strength increases steadily to a stable level. From period 310-320, as the metal layer 16 is increasingly removed to expose the oxide layer 14, the overall signal strength decreases and the polishing operation is finally complete. Therefore, in the period 320, only a minute amount of metal remains in the center of the substrate 10.
[0047]
When the entire surface of the substrate is covered with a metal layer such as copper, the reflection from the substrate 10 has a square profile. As the metal layer is removed from the edge of the substrate 10, the profile of reflection from the substrate assumes a trapezoid. When the metal layer is finally almost removed by the polishing operation, the reflection profile from the substrate 10 takes a triangular shape.
[0048]
Transient signal diagrams 300-320 may be observed by the operator on display 49 during or after the polishing operation. The operator can use the displayed transient signal diagram for various diagnostic and process control decisions (which may be applicable to both reflectivity measurements in metal polishing and interferometry measurements in oxide polishing). Can be used for. The transient signal diagram can be used to select process parameters to optimize polishing uniformity. For example, if process parameters such as plate rotation rate, carrier head pressure, carrier head rotation rate, carrier head sweep profile, and slurry composition are selected for the first time, the test wafer can be polished. The high reflectivity region represents a portion where the metal remains on the substrate, and the low reflectivity region represents a portion where the metal has been removed from the substrate. A noisy transient signal diagram indicates that the metal was not removed equally from the substrate, while a relatively flat transient signal diagram indicates uniform polishing. As a result, the operator can draw a direct conclusion on the effectiveness of the selected process parameters without resorting to measuring the substrate layer thickness with a metrology tool. The operator then adjusts the polishing parameters, polishes another test wafer, and determines whether the new polishing parameters have improved polishing uniformity.
[0049]
The operator can also examine the transient signal diagram to determine whether the substrate has been polished to a flat surface and whether polishing should be stopped. In addition, if the operator notices that a portion of the substrate is being polished too slowly or too quickly during actual device wafer polishing, the process parameters adjust the polishing rate profile while polishing is in progress. Can be changed to
[0050]
The transient signal diagram can also be used as a measure of process repeatability. For example, if the transient signal diagram deviates significantly from the expected shape, this indicates that there is some problem with the polishing machine or process.
[0051]
In addition, transient signal diagrams can be used to “qualify” the process. In particular, if the polishing machine accepts a new set of consumables, for example, if the polishing pad or slurry is replaced, the operator may wish to verify that the polishing uniformity is not affected. I will. The operator can compare the transient signal diagrams for the polished substrate before and after the consumable change and determine if polishing uniformity has been affected.
[0052]
Returning now to FIG. 8, at step 108, the radial positions Ra, Rb,..., Rj corresponding to the sampling zones 122a, 122b,. One way to determine the radial position of the sampling area is to calculate the position of the laser directly under the substrate based on the measurement time Tmeasure, the platen rotation rate and the carrier head sweep profile. Unfortunately, the actual platen rotation rate and carrier head sweep profile may not exactly match the polishing parameters. Therefore, a preferred method 130 for determining the radial position of the sampling area is shown in FIG. 9A. Initially, a time T sym for the laser beam 42 to pass directly below the substrate centerline 124 (see FIG. 5C) is determined (step 132). Next, the radial position of the sampling area is determined from the time difference between the measurement time Tmeasure and the symmetry time Tsym (step 134).
[0053]
One way to determine the symmetry time Tsym is to average the time of the first and last high intensity measurements from each sweep, since these intensity measurements should correspond to the substrate edge. However, this results in some uncertainty in Tsym since the location of the sampling area directly under the substrate is unknown.
[0054]
Referring to FIG. 9B, to calculate the symmetry time T sym at step 132, the computer 48 determines the first and last high intensity measurements from the sweep path 120, ie, intensity measurements Ic and Ih, and the corresponding measurements. Store times Tlead and Ttrail. These leading and trailing edge times Tlead and Ttrail are accumulated for each sweep, producing a series of leading edge times Tlead1, Tlead2, ..., TleadN and trailing edge times Ttrail1, Ttrail2, ..., TtrailN. The computer 48 stores for each leading edge spike 96 the leading edge times Tlead1, Tlead2,..., TleadN and the number of platen rotations 1, 2,. Similarly, the computer 48 stores the trailing edge time Ttrail1, Ttrail2,..., TtrailN and the associated number of rotations 1, 2,. Assuming that platen 24 rotates at a substantially constant rate, times Tlead1, Tlead2,..., TleadN form a substantially linearly increasing function (shown by line 136). Similarly, times Ttrail1, Ttrail2,..., TtrailN also form a function that increases substantially linearly (indicated by line 137). Computer 48 performs two least squares fits and generates two linear functions Tlead (n) and Ttrail (n) as follows:
Tlead (n) = a1 + (a2^*n)
T trail (n) = a3 + (a4^*n)
Here, n is the number of platen rotations, and a1, a2, a3 and a4 are fitness coefficients calculated during the application of the least square method. After the fit factor is calculated, the symmetry time T sym that the laser beam 42 traverses the center line 124 (indicated by phantom line 138) can be calculated as follows:
Tsym = {(a₁+ A_Three) / 2} + {(a₂+ A_Four) / 2} n
By calculating the symmetric time Tsym using the least squares method over several platen rotations, the uncertainty due to the difference in the relative position of the sampling area directly below the retaining ring is substantially reduced, and This significantly reduces the uncertainty at the symmetric time Tsym.
[0055]
After the computer 48 calculates the time T sym that the laser beam 42 traverses the center line 124, in step 132 the radial distance Ra, Rb,... Of each sampling area 122a, 122b,. , Rj is calculated. Referring to FIG. 10, the radial position can be calculated as follows:
R = (d²+ L²+2 dL cos θ)^1/2
Where d is the distance between the center of the polishing pad and the center of the window 36, L is the distance from the center of the polishing pad to the center of the substrate 10, and θ is the angular position of the window. The angular position θ of the window can be calculated as follows:
θ = f_platen・ 2π (T_measure-T_sym)
Where f_platenIs the platen rotation rate (in rpm). Assuming that the carrier head moves in a sinusoidal pattern, the linear position L of the carrier head can be calculated as follows:
L = L₀+ A ・ cos (ω ・ T_measure)
Where ω is the sweep frequency, A is the sweep amplitude, and L₀Is the center position of the carrier sweep.
[0056]
In another embodiment, the position sensor 160 could be used to calculate the time T sym that the window crosses the center line 124. Assuming that the sensor 160 is positioned opposite the carrier head 80, the flag 162 will be positioned symmetrically across the transparent window 36. The computer 48 stores both the start time Tstart when the flag blocks the light beam of the sensor and the start time Tend when the flag does not block the light beam. The time Tsym may be calculated as the average of Tstart and Tend. In yet another embodiment, the position of the platen and carrier head can be determined at each sampling time Ta, Tb,..., Th from optical encoders connected to the platen drive motor and radial drive motor, respectively. Like.
[0057]
After the radial position Ra, Rb,..., Rm of the sampling area has been calculated, some intensity measurements can be neglected. If the radial position R of the sampling area is larger than the radius of the substrate, the intensity measurements for that sampling area mainly include radiation reflected by the retaining ring, or background reflection from the window or slurry. Therefore, intensity measurements for any sampled area that is primarily beneath the retaining ring are ignored. This ensures that false intensity measurements are not used in the calculation of the reflected intensity of the thin film layer.
[0058]
After several sweeps of the laser beam 42 directly under the substrate, the computer 48 sets a set of intensity measurements I1, I2,..., IN, and radial directions associated with the measurement times T1, T2,. The positions R1, R2, ..., RN are accumulated. Referring to FIG. 11, when intensity, time, and radial position measurements are accumulated in steps 106 and 108, time and intensity measurements are binned in data structure 140 in step 110. Sorted in. Each bin is associated with a radial extent of the sampling area. For example, intensity measurements for a sampling area located up to 20 mm from the center of the substrate may be placed in the first bin 142 (see FIG. 13A), which will be discussed later as 20 mm from the center of the substrate. Intensity measurements made for sampling areas located between 30 mm may be placed in the second bin 144 (see FIG. 13B) and placed between 30 mm and 40 mm from the center of the substrate. Intensity measurements made for the sampled area to be made may be placed in the third bin 146 (see FIG. 13C), and so on. The exact number of bins and the radial range of bins depends on the information that the user wishes to retrieve. In general, the radial extent of each bin may be selected such that a sufficient number of intensity measurements are accumulated to provide visually meaningful information.
[0059]
The calculations discussed above are performed for each bin, thereby providing reflected intensity measurements at multiple radial locations across the surface of the substrate. A diagram of the initial and final reflected intensity of the thin film layer as a function of radius is shown in FIGS. 13A-13H as well as FIG. 12 discussed above.
[0060]
Returning now to FIGS. 13A-13H, a number of traces are shown, which show how the reflected intensity changes during polishing for different radial locations on the substrate 10. The diagrams in FIGS. 13A-13H illustrate that the metal layer is removed at different rates for different portions of the substrate. In general, FIGS. 13A-13H show that the metal layer near the center of the substrate is removed last, while the metal layer near the periphery or edge of the substrate is removed first. For example, FIG. 13A shows that a metal layer within a radius range of 0-20 mm is removed in about 330 seconds. FIG. 13B shows that a metal layer within a radius range of 20-30 mm is removed in about 325 seconds. FIG. 13C shows that a metal layer within a radius range of 30-40 mm is removed in about 318 seconds. FIG. 13D shows that a metal layer within a radius range of 40-50 mm is removed in about 310 seconds. FIG. 13E shows that a metal layer within a radius range of 50-60 mm is removed in about 295 seconds. FIG. 13F shows that a metal layer within a radius range of 60-70 mm is removed in about 290 seconds. FIG. 13G shows that the metal layer within the radius range of 70-80 mm is removed in about 290 seconds, and FIG. 13H shows that the metal layer within the radius range of 80-90 mm is removed as early as about 260 seconds. It shows that.
[0061]
As shown here, the reflectance trace for some of the radial ranges presents two intensity levels (shown by lines 160 and 162). The distance between the two intensity levels increases with the radius of the substrate. Without being limited to any particular theory, the two intensity levels can be attributed to the asymmetric distribution of the product of the slurry or reaction of the slurry and the metal layer on the substrate. Specifically, for each sweep of the laser beam across the substrate, two data points are usually entered into the bin: one data point near the front edge of the substrate and one data point near the back edge of the substrate. is there. However, because of the asymmetric distribution of slurry and reaction products directly under the substrate, the laser beam may be attenuated more as it passes through the slurry layer adjacent to different parts of the substrate. Thus, the reflectance trace may also be used as a measure of the uniformity of the slurry distribution just below the substrate.
[0062]
In another embodiment, the operator may decide to use only a single bin. In this case, all of the intensity measurements for a particular radial range are used to determine a single intensity trace, which is used for determining the polishing end point in a conventional manner. The operator can specify this radial range based on inspection of the transient signal diagram. For example, if the transient signal diagram indicates that the center of the substrate is the last part to be polished, the operator can select a radial range near the center of the substrate and finish until all of the metal has been polished. Ensure that is not started.
[0063]
Reflection intensity changes during polishing are thus captured for different radial positions on the substrate. Acquisition of high resolution data allows precise time control of each process stage in multi-stage operation. A large number of parameters are captured, such as overall wafer uniformity and removal rates for different radial portions of the wafer. The acquired high resolution data can be processed online or offline to adjust various variables and parameters to minimize surface layer erosion and dishing. If the data is processed in real time, the real time feedback data allows strict closed loop control with process parameters. In addition, the reflection data is available for process engineers to experiment with their process parameters to improve the polishing process.
[0064]
The reflectance trace generated by the optical monitoring system is particularly useful in multi-step polishing processes, such as copper polishing. Referring to FIG. 14, a substrate 10 'includes a silicon wafer 12', a patterned oxide layer 14 ', and a tantalum (Ta) or tantalum nitride (TaN) barrier disposed on the oxide layer 14'. A layer 18 and a copper layer 16 ′ disposed on the barrier layer 18 are included. Referring to FIGS. 15A and 15B, as the substrate 10 ′ is polished, a reflectance trace 200 is generated for each radial area. Each reflectance trace includes a bulk removal of the metal layer (flat area 202), an initial sweep of the metal layer and a transition to the barrier layer (fall point 204), an initial sweep of the barrier layer and a transition to the oxide layer ( Illustrates a first decrease point 206 at the slope, and a complete sweep of the barrier layer and exposure of the oxide layer (second decrease at the slope 208 and planarization of the trace).
[0065]
Initially, the substrate is polished with a high selectivity slurry, such as Cabot 5001. Mass polishing of the metal layer 16 ′ proceeds until the first penetration into the barrier layer 18. In this regard, the metal may still be on a portion of the barrier layer 18, but the barrier layer will be exposed at at least one site. The optical monitoring system can detect the initial exposure of this barrier layer. Specifically, the first reflectance trace that begins to fall may be used as an indication that the barrier layer 18 is exposed in the associated radial area. When the optical monitoring system detects the first exposure of the barrier layer, polishing with the high selectivity slurry is stopped and polishing with the low selectivity slurry, eg, Arch Cu10k, is started. Polishing with the low selectivity slurry continues until all of the metal layer 16 'and the barrier layer 18 are removed. Complete removal of all metal layer 16 'and barrier layer 18 may be indicated when all of the reflectance traces are leveled off. At this point, the polishing operation is complete. With a reliable switch from a high selectivity slurry to a low selectivity slurry with the first sweep of the barrier layer, dishing and erosion are significantly reduced. In addition, complete removal of the barrier layer may be ensured by stopping polishing only after all reflectance traces indicate.
[0066]
The invention has been described with reference to the preferred embodiment. However, the invention is not limited to the illustrated and described embodiments. Rather, the scope of the present invention is defined by the appended claims.
[Brief description of the drawings]
FIG. 1 is an exploded perspective view of a chemical mechanical polishing apparatus.
FIG. 2 is a side view of a chemical mechanical polishing apparatus including an optical reflectometer.
FIG. 3 is a simplified cross-sectional view of a substrate being processed, schematically showing a laser beam impinging on and reflecting from the substrate.
FIG. 4 is a diagram showing measured reflectance traces in arbitrary intensity units (au).
FIG. 5A is a simplified plan view showing the position of a window in a polishing pad as the platen rotates.
FIG. 5B is a simplified plan view showing the position of the window in the polishing pad as the platen rotates.
FIG. 5C is a simplified plan view showing the position of the window in the polishing pad as the platen rotates.
FIG. 5D is a simplified plan view showing the position of the window in the polishing pad as the platen rotates.
FIG. 5E is a simplified plan view showing the position of the window in the polishing pad as the platen rotates.
FIG. 6 is a flowchart of a method for determining an end point of polishing of a metal layer during CMP.
FIG. 7A is a schematic diagram showing the path of a laser directly under a carrier head.
FIG. 7B is a diagram illustrating a hypothetical portion of a reflectance trace generated by a single sweep of a window directly below the carrier head.
FIG. 8 is a schematic diagram showing the radial position of the sampling area from the path of the laser.
FIG. 9A is a flowchart of a method for determining a radial position of a sampling area.
FIG. 9B is a diagram showing the time for the laser beam to pass directly under the leading and trailing edges of the substrate as a function of the number of platen rotations.
FIG. 10 is a schematic diagram illustrating the calculation of the radial position of the sampling area.
FIG. 11 is a schematic diagram of a data structure for storing intensity measurements.
FIG. 12 is a diagram showing the overlap of several reflectance traces taken at different times.
FIG. 13A is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13B is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13C is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13D is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13E is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13F is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13G is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 13H is a diagram illustrating the reflection intensity of a metal layer over the polishing period as a function of distance from the center of the substrate.
FIG. 14 is a simplified cross-sectional view of a substrate having a barrier layer.
15A is a diagram showing a reflectance trace during polishing of the substrate shown in FIG.
15B is a diagram showing a reflectance trace during polishing of the substrate shown in FIG.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Board | substrate, 12 ... Silicon wafer, board | substrate 14 ... Underlayer patterned layer, foundation layer, oxide layer, 16 ... Metal layer, copper layer, 18 ... Barrier layer, 20 ... Polishing apparatus, 22 ... Polishing station, 23 ... Transfer station, 24 ... Platen, 25 ... Central axis, shaft, 26 ... Hole, 28 ... Pad conditioner device, 30 ... Polishing pad, 32 ... Backing layer, 34 ... Overcoat layer, 36 ... Window, 38 ... Slurry, 39 ... rinse arm, 40 ... reflectometer, 42 ... laser beam, 42 ... light beam, optical laser beam, 44 ... laser, light source, 46 ... detector, 48 ... computer, digital computer, 49 ... display, 49 ... output device , 60 ... carousel, 62 ... central support, 64 ... carousel shaft, 66 ... carousel support plate, 68 ... cover, 70 ... carrier head system, 72 ... radial slot, 74 ... carrier drive shaft, 76 ... carrier head rotary motor, 80 ... carrier head, 81 ... central axis, 82 ... flexible membrane, 84 ... retaining ring, 86 ... chamber, 88 ... lower surface, 90 ... Reflectance trace, 90 ... Intensity waveform, 90 ... Waveform, end point waveform, transient intensity waveform, 94 ... Background, background level, 97 ... Upper level flat region, 98 ... Middle level flat region, Intermediate flat region, after each Edge spike, 120 ... sweep path, 122a-122j ... sampling area, 124 ... center line, 126 ... center, 136, 137, 160 ... line, 138 ... virtual line, 140 ... data structure, 160 ... sensor, position sensor, 162: Flag, 200: Reflectance trace, 202 ... Site, 204 ... Decline, 206 ... Decrease point, 300, 320 ... Diagram, 300 ... Period, Transient signal diagram, 02,310 ... period.

Claims (13)

A method of polishing a substrate, the method comprising: chemically mechanically polishing a first layer of a substrate with a first polishing slurry , wherein the substrate is disposed under the first layer. Having a layer, wherein the first and second layers have different reflectivities;
A substrate by optically monitored during polishing by the first polishing slurry, comprising the steps of: generating a plurality of intensity traces, each intensity trace is a measure of the strength of different radial range on the said substrate Including steps, and
If any of the intensity traces indicates that the second layer has been exposed at at least one location, the substrate is removed from the chemical machine by a second polishing slurry having a polishing characteristic different from that of the first polishing slurry. Polishing step,
Continuing to optically monitor the substrate during polishing with the second polishing slurry;
Stopping polishing after all the intensity traces indicate that the second layer is fully exposed;
Including methods. Step of monitoring optically directs to transmit a window portion with the light beam within the polishing surface, moving said light beam in a path across said substrate is generated by the light beam reflected from the substrate The method of claim 1, comprising monitoring a reflected signal and extracting a plurality of intensity measurements from the reflected signal.   The step of generating the plurality of intensity traces sorts each intensity measurement into one of the radial ranges according to the position of the light beam during the previous intensity measurement, and from the intensity measurements associated with the radial range, the The method of claim 1, comprising determining an intensity trace.   The method of claim 1, wherein the first slurry is a high selectivity slurry and the second slurry is a low selectivity slurry.   The method of claim 1, wherein the first layer is a metal layer.   The method of claim 5, wherein the metal layer comprises copper.   The method of claim 5, wherein the second layer is an oxide layer.   The method of claim 7, wherein the oxide comprises silicon dioxide.   The method of claim 5, wherein the second layer is a barrier layer.   The method of claim 9, wherein the barrier layer comprises tantalum or tantalum nitride.   The method of claim 1, wherein the first layer is more reflective than the second layer. A method for polishing a substrate, comprising:
A step of contacting the polishing surface and the surface of the substrate having a window portion, said substrate having a first layer disposed on the second layer, said first and second layers Steps having different reflectivities; and
Supplying a first slurry to the substrate for a first polishing stage;
Causing a relative movement between the substrate and the polishing surface;
Directing a light beam through the window , wherein the movement of the polishing surface relative to the substrate moves the light beam in a path across the substrate;
Monitoring a reflectivity signal generated by the light beam reflected from the substrate;
Extracting a plurality of intensity measurements from the reflectance signal;
A step of generating a plurality of intensity traces, each intensity trace comprising an intensity measurements different radial range on the said substrate, the steps,
Said intensity either trace, indicating that the second layer is exposed at at least one site, for the second polishing step, the second slurry comprising the steps supplying to said substrate, The second slurry has a different polishing characteristic than the first slurry; and
Stopping polishing after all the intensity traces indicate that the second layer is fully exposed;
Including methods.   The method of claim 12, wherein the first slurry is a high selectivity slurry and the second slurry is a low selectivity slurry.

Images