DETAILED DESCRIPTION OF THE INVENTION

[0027] The present invention relates to planarizing machines and methods for monitoring and controlling planarizing processes in mechanical or chemical-mechanical planarization of microelectronic substrate assemblies. Many specific details of the invention are described below with reference to planarizing semiconductor wafers using web-format and rotary planarizing machines to provide a thorough understanding of such embodiments. For example, general aspects of a representative web-format planarizing machine and illustrative methods for controlling CMP processing using this machine are initially described below with reference to FIGS. 2 - 6 . Several detailed embodiments of planarizing machines for practicing methods in accordance with the invention are then described with reference to FIGS. 7 - 14 . The present invention, however, may have additional embodiments and/or can be practiced without several of the details described in the following description.

[0028] A. Representative Planarizing Machines and Monitoring Systems

[0029] FIG. 2 is a schematic cross-sectional view of a web-format planarizing machine 100 having a monitoring system for monitoring and controlling planarization of a microelectronic substrate assembly 12 in accordance with the invention. The planarizing machine 100 includes a table 110 , a carrier assembly 130 over the table 110 , and a polishing pad 140 on the table 110 . The carrier assembly 130 and the polishing pad 140 can be substantially the same as those describe above with reference to FIG. 1 . The polishing pad 140 is accordingly coupled to a pad-advancing mechanism having a plurality of rollers 120 - 123 . The pad-advancing mechanism can also be the same as that described above with reference to FIG. 1 .

[0030] The planarizing machine 100 also includes a monitoring system that measures the drag force between the substrate assembly 12 and the polishing pad 140 during planarization to endpoint or control other aspects of the CMP process. The monitoring system can include a drag force measuring assembly 160 (identified by reference numbers 160 a and 160 b ) to measure a parameter that accurately indicates the drag force between the pad 140 and the substrate assembly 12 along a lateral axis. The lateral axis can extend generally parallel to a plane defined by the interface of the pad 140 and the substrate assembly 12 . The drag force measuring assembly 160 generates a waveform having minimum and maximum drag force peaks along the lateral axis. The waveform, for example, can be a generally sinusoidal wave with varying amplitudes corresponding to the minimum and maximum drag forces along the lateral axis.

[0031] The drag force measuring assembly 160 can be coupled to the table 110 and/or a carrier head 132 of the carrier assembly 130 . The drag force measuring assembly 160 is generally configured to isolate a drag force parameter indicative of drag forces between the substrate assembly 12 and the polishing pad 140 that is not influenced by energy losses in motors, gears or other components that drive either the polishing pad 140 or the carrier head 132 . For example, the drag force measuring assembly 160 a in the table 110 can have a first component 162 a coupled to the polishing pad 140 , a second component 164 a that is fixedly attached to or integral with the table 110 , and a force detector 190 to detect lateral forces or lateral displacement between the first component 162 a and a second component 164 a . The drag force measuring assembly 160 b in the carrier assembly 130 can include a first component 162 b coupled to the substrate assembly 12 and a second component 164 b coupled to either the carrier head 132 , the shaft 139 or another component of the carrier assembly 130 . The drag force measuring assembly 160 b also has a force detector 190 to detect lateral forces or lateral displacement between the first and second components 162 b and 164 b . The drag force parameter can accordingly be lateral displacement or lateral forces between the first component 162 (identified by reference numbers 162 a and 162 b ) and the second component 164 (identified by reference numbers 164 a and 164 b ) corresponding to a shear force between the substrate assembly 12 and the planarizing surface 142 of the polishing pad 140 . Several embodiments of force detectors 190 and drag force measuring assemblies that isolate the drag force parameter from power losses are described in detail below with reference to FIGS. 7 - 14 .

[0032] The drag force monitoring system can also include a processor 199 coupled to the drag force measuring assembly 160 . The processor 199 receives signals from the drag force measuring assembly 160 corresponding to the waveform of the measured drag forces. The processor generates a force-time relationship between the peak drag forces of the waveform and time. In one particular application of the present invention for planarizing a substrate assembly having a shallow-trench-isolation structure (“STI”), the force-time relationship generally has an initial section with an increasing slope, an intermediate section with a decreasing slope, and a final section with a relatively flat slope. The processor 199 can further perform a first regression on the intermediate section to generate a first line and a second regression on the final section to generate a second line. The processor determines a reference time indicating the exposure of an endpointing layer by determining the time corresponding to the intersection between the first and second lines. The processor 199 then endpoints CMP processing of the STI substrate assembly at an estimated endpoint time equal to approximately the reference time plus a predetermined over-polish time. Several methods for controlling or endpointing CMP processing using the force-time relationship between the peak drag forces and time are set forth below.

[0033] B. Illustrative Methods for Endpointing and Controlling CMP Processing

[0034] FIG. 3 is a flowchart of a method for endpointing CMP processing of the substrate assembly 12 in accordance with one embodiment of the invention. The method illustrated in FIG. 3 is a comprehensive method that includes several procedures that may be combined with each other or completely excluded in other embodiments of the invention. Accordingly, other embodiments of the invention may also include different procedures or a different order of procedures. The CMP process of FIG. 3 is also applicable to several different types of semiconductor wafers, field emission displays and other microelectronic substrate assemblies.

[0035] The planarizing process illustrated in FIG. 3 includes removing material from a substrate assembly by pressing the substrate assembly against a planarizing surface (procedure 410 ) and moving the substrate assembly and/or the polishing pad with respect to the other (procedure 412 ). The polishing pad can be the web-format pad 140 shown in FIG. 2 or a rotary pad. The polishing pad can also be a fixed-abrasive pad with abrasive particles fixedly attached to a suspension medium or a non-abrasive pad without abrasive particles. The substrate assembly is generally pressed against the polishing pad in the presence of a planarizing solution.

[0036] FIG. 4 is a schematic cross-sectional view illustrating the removal of material from an STI substrate assembly 12 . The STI substrate assembly 12 has a substrate 13 with a plurality of trenches 14 , an endpointing layer 15 composed of a first material with a first polishing rate, and a fill layer 16 or cover layer composed of a second material having a second polishing rate different than the endpointing layer 15 . The endpointing layer 15 is generally a polish-stop layer that has a lower polishing rate than the cover layer 16 to inhibit planarization below lands 17 at a desired endpoint elevation in the substrate assembly 12 . In an alternate embodiment, the endpoint layer 15 can have a higher polishing rate than the cover layer 16 . The endpointing layer 15 , for example, can be a silicon nitride or carbon polish-stop layer, and the fill layer 16 can be a doped or undoped silicon dioxide layer. The substrate assembly 12 contacts the planarizing surface 142 of the polishing pad 140 at a pad/substrate interface 143 defined by the surface area “SA” in contact with the planarizing surface 142 . The carrier assembly 130 presses the substrate assembly 12 against the planarizing surface 142 at a downforce F d . The carrier assembly 130 also moves the substrate assembly 12 with respect to the polishing pad 140 to rub the substrate assembly 12 against the planarizing surface 142 at a relative velocity V r . The friction between the substrate assembly 12 and the planarizing surface 142 creates a drag force F D that acts against the polishing pad 140 .

[0037] Referring again to FIG. 3 , the planarizing process 400 continues with a measuring procedure 420 that includes measuring a drag force parameter indicative of the drag force F D between the substrate assembly 12 and the polishing pad 140 . The drag force parameter is generally isolated from energy losses in components that drive either the polishing pad or the substrate assembly to provide a more accurate indication of the drag force F D at the pad/substrate interface 143 ( FIG. 4 ). The drag force parameter can be measured along a lateral axis that is generally parallel to a plane defined by the pad/substrate interface to generate the sinusoidal waveform (examples of lateral axes are identified by lines L 1 or L 2 of FIG. 7 ). Several devices for isolating and measuring the drag force parameter are described below with reference to FIGS. 7 - 14 .

[0038] The planarizing process continues with a data processing procedure 430 in which a waveform of the measured drag force is generated. FIG. 5 is a graph of a waveform 432 of the measured drag force F D along a lateral axis at the pad/substrate interface for an STI substrate assembly over time. The measured waveform 432 for the STI substrate assembly has a plurality of maximum peaks 434 (identified by reference numbers 434 a and 434 b ) and a plurality of minimum peaks 436 (identified by reference numbers 436 a and 436 b ). The maximum and minimum peaks 434 and 436 correspond to the maximum and minimum drag forces between the pad 140 and a substrate assembly 12 along the lateral axis. The waveform 432 is generally a sinusoidal waveform in which the amplitude between the maximum peaks 434 and the minimum peaks 436 indicates increases or decreases in the drag force at the pad/substrate interface. For an STI substrate assembly, the waveform 432 has a first section 437 in which the peak-to-peak amplitude increases, a second section 438 in which the peak-to-peak amplitude decreases, and a third section 439 in which the peak-to-peak amplitude remains substantially constant.

[0039] The method 400 shown in FIG. 3 further continues with a correlating procedure 440 in which the maximum peak drag forces 434 , or the differences between the maximum and minimum peak drag forces 434 and 436 ( FIG. 5 ), are correlated with time to generate a force-time relationship. In another embodiment, the correlating procedure 440 can correlate the minimum peak drag forces 436 with time The correlating procedure 440 produces a peak drag force curve corresponding to the peak drag forces along the lateral axis.

[0040] FIG. 6 is a graph illustrating a peak drag force curve 442 corresponding to either the maximum peak drag forces 434 or the differences between the maximum and minimum peak drag forces 434 and 436 of the waveform 432 shown in FIG. 5 . The times t 1 and t 2 in FIG. 5 correspond to the times t 1 and t 2 in FIG. 6 . The peak drag force curve 442 has a first or initial section 447 with an increasing slope corresponding to the increasing peak drag forces 434 in the first section 437 of the waveform 432 shown in FIG. 5 . The peak drag force curve 442 has a second or intermediate section 448 with a generally downward slope corresponding to the decreasing peak drag forces 434 of the waveform 432 shown in FIG. 5 . The peak drag force curve 442 also has a third or end section 449 with a relatively flat slope corresponding to the substantially constant peak drag forces 434 in the third section 439 of the waveform 432 shown in FIG. 5 . The peak drag force curve 442 is used to estimate the endpoint of the planarizing cycle or to estimate the status of another parameter of the CMP process.

[0041] The process 400 of FIG. 3 further includes an estimating procedure 450 for estimating a reference time t r corresponding to an exposure time that the lands 17 of the endpoint layer 15 ( FIG. 4 ) are exposed during planarization. The estimating procedure includes performing a mathematical regression of the intermediate section 448 of the peak drag force curve 442 to create a downwardly sloping first line 444 , and performing a regression of the end section 449 of the peak drag force curve 442 to determine a second line 446 . Suitable software or hardware for performing the regressions of the peak drag force curve 442 are commercially available and known to those skilled in the semiconductor manufacturing arts. The reference time is estimated by determining the time corresponding to the intersection between the first line 444 and the second line 446 .

[0042] The method 400 of FIG. 3 continues with a terminating procedure 460 that terminates removal of the material from the substrate assembly 12 at an estimated endpoint time. The terminating procedure 460 calculates the estimated endpoint time by adding a predetermined over-polish time to the reference time t r . In a typical STI application, the over-polish time is approximately 10-50 seconds, and more specifically approximately 25-35 seconds. The estimating procedure 450 actually occurs during the initial portion of the third section 449 of the peak-drag force curve 442 ( FIG. 6 ) because a sufficient number of data points indicating that the planarizing process has entered the end section 449 must be obtained. In several STI applications, the reference time t r corresponding to the exposure of the endpoint layer 15 ( FIG. 4 ) can be calculated approximately 7 seconds after the peak drag force curve 442 enters the third section 449 . Therefore, because the over-polish time is approximately 10-50 seconds after the reference time t r occurs, the terminating procedure 460 can estimate the endpoint of the planarizing process in situ and in real time.

[0043] C. Embodiments of Endpointing and Drag Force Measuring Assemblies

[0044] FIGS. 7 - 14 illustrate several embodiments of endpointing apparatuses that execute the measuring procedure 420 ( FIG. 3 ) by isolating a drag force parameter related to the drag force between the substrate assembly 12 and the polishing pad 140 from other energy losses, and measuring the isolated drag force during planarization. For the following description, the endpointing apparatuses described in FIGS. 7 - 14 define one type of drag force measuring assembly 160 shown in FIG. 2 . Therefore, it will be understood that the drag force measured with the endpointing apparatuses shown in FIGS. 7 - 14 can also be used to diagnose or control other aspects of the CMP processes described above with reference to FIGS. 2 - 6 .

[0045] FIG. 7 is a schematic isometric view of the web-format planarizing machine 100 including an endpointing apparatus for measuring the drag force between the substrate assembly 12 and the polishing pad 140 during planarization. The endpointing apparatus generally includes a secondary support member defined by a sub-platen 150 , a primary support member defined by a platen 170 , and at least one force detector 190 between the sub-platen 150 and the platen 170 . The platen 170 and the sub-platen 150 can be separate components of the table 10 . The polishing pad 140 is releasably coupled to the platen 170 so that the drag forces FD between the substrate assembly 12 and the pad 140 exert lateral forces against the platen 170 independent of friction losses or power losses in the carrier assembly 130 . The lateral force exerted by the pad 140 against the platen 170 is thus an isolated parameter indicative of the drag FD between the substrate assembly 12 and the pad 140 .

[0046] FIG. 8 is a schematic cross-sectional view of the planarizing machine 100 illustrating the endpointing apparatus in greater detail. Referring to FIGS. 7 and 8 together, the sub-platen 150 can be a base supporting the platen 170 . The sub-platen 150 has a recess 152 defined by a base surface 153 and a plurality of walls (identified by reference numbers 154 a , 154 b , 156 a and 156 b ) projecting upwardly from the base surface 153 transversely with respect to a planarizing plane P-P ( FIG. 8 ). For the purposes of the present disclosure, the term “transverse” means any non-parallel arrangement and is not limited to a perpendicular arrangement. The walls can include a first side-wall 154 a , a second side-wall 154 b opposite the first side-wall 154 a , a first end-wall 156 a at one end of the side-walls 154 a and 154 b , and a second end-wall 156 b at the other end of the side-walls 154 a and 154 b . The walls can be configured in a rectilinear pattern or other suitable patterns to receive the platen 170 .

[0047] The platen 170 is positioned in the recess 152 of the sub-platen 150 . The platen 170 can be a plate having a first side-face 172 a , a second side-face 172 b opposite the first side-face 172 a , a first end-face 174 a between one end of the side-faces 172 a and 172 b , and a second end-face 174 b between the other end of the side-faces 172 a and 172 b . In the embodiment shown in FIG. 3 , the first side-face 172 a is adjacent to the first side-wall 154 a , the second side-face 172 b is adjacent to the second side-wall 154 b , the first end-face 174 a is adjacent to the first end-wall 156 a , and the second end-face 174 b is adjacent to the second end-wall 156 b . The platen 170 also includes a bearing surface 176 facing the backside of the polishing pad 140 to support at least a portion of the polishing pad 140 in a planarizing zone under the head 132 . The platen 170 further includes a back surface 178 facing the base surface 153 of the sub-platen 150 . The polishing pad 140 is coupled to the bearing surface 176 during planarization so that the pad transmits lateral forces to the platen 170 . Suitable devices and methods for coupling the polishing pad 140 to the bearing surface 176 are disclosed in U.S. patent application Ser. Nos. 09/285,319 filed on Apr. 2, 1999, and 09/181,578 filed on Oct. 28, 1998, both of which are herein incorporated by reference.

[0048] The platen 170 can move with respect to the sub-platen 150 in a lateral motion at least generally parallel to a planarizing plane P-P ( FIG. 8 ). In this embodiment, the endpointing apparatus also includes a bearing mechanism 180 ( FIG. 8 ) to reduce the friction between the base surface 153 of the sub-platen 150 and the back surface 178 of the platen 170 . The bearing assembly 180 can be a roller mechanism having a plurality of rollers attached to either the sub-platen 150 or the platen 170 to allow the platen 170 to freely roll across the sub-platen 150 . The bearing assembly 180 can also be a low-friction coating or lubricant between the base surface 153 and the back surface 178 , or a flexible bladder (not shown) between the sub-platen 150 and the platen 170 . In still another embodiment, the bearing assembly 180 can be a frictionless device having a number of air bearings defined by air holes through the sub-platen 150 that are connected to a pressurized air source that provides a continuous layer of air between the sub-platen 150 and the platen 170 . In still another embodiment, the bearing assembly 180 can be a magnetic device including magnetic bearings that prevent the back surface 178 from contacting the base surface 153 by positioning magnetic fields of a like polarity adjacent to one another. In operation, the bearing assembly 180 can frictionally isolate the platen 170 from the sub-platen 150 so that the drag forces between the substrate assembly 12 and the pad 140 drive the platen 170 laterally with respect to the sub-platen 150 without substantial friction losses.

[0049] The force detectors 190 (identified by reference numbers 190 a - 190 d ) can be positioned between the walls of the recess 152 in the sub-platen 150 and the faces of the platen 170 . Each force detector 190 can be a contact sensor that contacts both the sub-platen 150 and the platen 170 to sense the lateral forces exerted by the platen 170 against the sub-platen 150 in correlation to the lateral forces exerted by the substrate assembly 12 against the polishing pad 140 during planarization. Suitable contact force detectors are strain gauges, piezoelectric elements or other transducers that generate signals corresponding to the force exerted by the platen 170 against the sub-platen 150 . The force detectors 190 can be other sensors that generate electrical signals corresponding to the lateral forces or displacement between the sub-platen 150 and the platen 170 . For example, in other embodiments in which the force detectors 190 do not contact the platen 170 and the sub-platen 150 does not have dead stops so that the platen 170 can move relative to the sub-platen 150 , the force detectors 190 can be lasers, accelerometers, capacitance displacement sensors, linear variable differential transformers or other displacement sensors.

[0050] In the particular embodiment of the planarizing machine 100 illustrated in FIGS. 7 and 8 , four force detectors 190 a - 190 d are configured along two orthogonal lateral axes L 1 and L 2 . Each lateral axis L 1 and L 2 defines a lateral axis along which the drag forces between the pad 140 and the substrate assembly 12 can be measured to generate the drag force waveform 432 shown in FIG. 5 . In other embodiments, the planarizing machine 100 can have only one force detector positioned along one axis, or two force detectors positioned along two orthogonal axes, or any number of force detectors positioned between the walls of the sub-platen 150 and the faces of the platen 170 . For example, in an embodiment having two force detectors 190 positioned along orthogonal axes, a first force detector 190 a can contact the first end-wall 156 a and the first end-face 174 a at a first force detector site, a second force detector 190 b can contact the first side-wall 154 a and the first side-face 172 a at a second force detector site, and dead stops can be substituted for the force detectors 190 c and 190 d . The first end-wall 156 a and the first side-wall 154 a of the sub-platen 150 accordingly define first and second stop surfaces, and the first end-face 174 a and the first side-face 172 a of the platen 170 accordingly define first and second contact surfaces. In still another embodiment, the first and second force detectors 190 a and 190 b can be positioned as explained above, and the dead stops or force detectors 190 c and 190 d can be eliminated by sizing the platen 170 such that the second end-face 174 b abuts the second end-wall 156 b and the second side-face 172 b abuts the second side-wall 154 b.

[0051] FIG. 9 is a schematic cross-sectional view of the planarizing machine 100 in accordance with another embodiment of the invention. In this embodiment, the sub-platen 150 has a post 155 projecting upwardly from the base surface 153 , and the platen 170 is fixedly attached to the post 155 . The walls 172 / 174 of the platen 170 do not contact either dead stops, the faces 154 / 156 of the sub-platen 150 , or other devices that inhibit the platen 170 from moving with respect to the sub-platen 150 . The movement of the substrate assembly 12 across the polishing pad 140 accordingly displaces the platen 170 relative to the sub-platen 150 and generates torsional forces in the post 155 that are expected to be proportionate to the drag force between the substrate assembly 12 and the polishing pad 140 . The force detector 190 can be a strain gauge attached to the post 155 to measure the torsional displacement of the post 155 , a laser, or another type of displacement sensor. The force detector 190 accordingly senses the change in the displacement or the torsional forces exerted on the platen 170 and sends a corresponding signal to the processor 199 a.

[0052] FIG. 10 is a schematic cross-sectional view of the planarizing machine 100 in accordance with another embodiment of the invention in which a number of small posts 155 attach the platen 170 to the sub-platen 150 . As with the embodiment of the planarizing machine shown in FIG. 10 , the walls 172 / 174 of the platen 170 do not contact either dead stops, the faces 154 / 156 of the sub-platen 150 , or other devices that inhibit the platen 170 from moving with respect to the sub-platen 150 . The posts 155 can be threaded studs having a diameter of approximately 1.0 inch and a length of 3.0 inches made from metal, high density polymers or other suitable materials. The posts 155 can also be other supports that can flex more in one direction than others, and the posts 155 can be made from other materials. The posts 155 of this embodiment accordingly do not frictionally isolate the platen 170 from the sub-platen 150 , but rather they deflect to control the motion between the platen 170 and the sub-platen 150 in correspondence to the drag forces between the substrate assembly 12 and the polishing pad 140 . The force detectors 190 accordingly measure the displacement between the platen 170 and the sub-platen 150 to determine the drag forces between the substrate assembly 12 and the polishing pad 140 .

[0053] FIG. 11 is a schematic isometric view of a planarizing machine 100 in accordance with still another embodiment of the invention. In this embodiment, the planarizing machine 100 has a circular platen 170 and the recess 152 in the sub-platen 150 has a single circular wall 154 . The platen 170 accordingly has a single, circular side-face 174 . The platen 170 can be coupled to the sub-platen 150 by any of the bearings 180 or posts 155 described above with reference to FIGS. 7 - 10 .

[0054] FIG. 12 is a schematic isometric view of a planarizing machine 200 in accordance with another embodiment of the invention, and FIG. 13 is a schematic cross-sectional view of the planarizing machine 200 shown in FIG. 12 taken along line 13 - 13 . The planarizing machine 200 has a sub-platen 250 coupled to a rotary drive mechanism 251 to rotate the sub-platen 250 (arrow R), a platen 270 movably coupled to the sub-platen 250 , and a polishing pad 240 attached to the platen 270 . The sub-platen 250 has a base surface 253 facing the polishing pad 240 and a tab 254 projecting upwardly from the base surface 253 . The tab 254 has a stop surface 256 facing in the direction of the rotation of the sub-platen 250 . The platen 270 includes an opening 271 having a contact surface 272 facing the stop surface 256 of the tab 254 . The planarizing machine 200 further includes a bearing assembly 280 that can be the same as the bearing assembly 180 described above with reference to FIG. 8 . The planarizing machine 200 also includes a force detector 290 contacting the stop surface 256 of the tab 254 and the contact surface 272 of the platen 270 .

[0055] The planarizing machine 200 is expected to enhance the accuracy of detecting the endpoint of planarizing a substrate assembly in rotary planarizing applications. In operation, a carrier assembly 230 ( FIG. 13 ) moves a carrier head 232 to press the substrate assembly 12 against a planarizing surface 242 of the polishing pad 240 . The rotary drive assembly 251 also rotates the sub-platen 250 causing the tab 254 to press the force detector 290 against the contact surface 272 . The sub-platen 250 accordingly rotates the platen 270 in the direction R, but the drag force between the substrate assembly 12 and the polishing pad 240 resists rotation in the direction R. The bearing assembly 280 allows the drag forces between the substrate assembly 12 and the planarizing surface 242 to drive the contact surface 272 of the platen 270 against the force detector 290 in correlation to the drag forces. As the drag force increases between the substrate assembly 12 and the planarizing surface 242 , the force detector 290 accordingly detects an increase in the lateral force that the platen 270 exerts against the tab 254 . The force detector 290 is coupled to a processor 299 to convert the signals from the force detector 290 into data that can be analyzed to determine the endpoint of the planarizing process as described above with reference to FIGS. 2 - 6 .

[0056] FIG. 14 is a schematic cross-sectional view of a carrier system 330 for a planarizing machine in accordance with another embodiment of the invention. The carrier assembly 330 can include a carrier head 332 having a lower portion 333 with a lower cavity 334 to receive a substrate assembly 12 and an upper portion 336 with an upper cavity 338 . A pivoting joint 350 is attached to the head 332 in the cavity 338 , and a drive-shaft 339 is pivotally attached to the joint 350 . In this embodiment, the endpointing apparatus includes a primary support member defined by the head 332 , a secondary support member defined by the drive-shaft 339 , and a first contact surface defined by the side-wall of the upper cavity 338 . In one embodiment, the joint 350 is a gimbal joint or other bearing assembly that allows universal pivoting between the head 332 and the shaft 339 . The carrier head 332 also includes a force detector 390 attached to an interior wall of the cavity 338 . The force detector 390 , for example, can be an annular piezoelectric ring.

[0057] In operation, the drag forces between the substrate assembly 12 and the polishing pad 140 cause the shaft 339 to pivot about the joint 350 such that the lower end of the shaft 339 contacts the force detector 390 . The force exerted by the driveshaft 339 against the force detector 390 will be proportional to the drag forces between the substrate assembly 12 and the polishing pad 140 . Accordingly, the isolated drag force parameter of this embodiment is the displacement between the shaft 339 and the carrier head 332 . The force detector 390 is coupled to a processor 199 a ( FIG. 2 ) to detect the endpoint of the planarizing process in a manner similar to that described above with respect to FIGS. 2 - 6 .

[0058] D. Monitoring and Controlling Applications

[0059] The planarizing machines and methods described above with reference to FIGS. 2 - 14 are expected to enhance the accuracy of endpointing CMP processing compared to processes and devices that monitor changes in the current of the drive motors. The methods described above with reference to FIGS. 2 - 6 , for example, accurately endpoint CMP processing because they accurately estimate the exposure time of the endpoint layer by measuring the peak drag forces between the substrate assembly and the polishing pad along a lateral axis to obtain a peak drag force waveform, developing a peak drag force curve from the peak drag force waveform, and determining a reference time at the intersection of a first line corresponding to a downwardly sloping section of the drag force curve and a second line corresponding to a relatively flat section of the drag force curve. One aspect of several embodiments of methods set forth above with respect to FIGS. 2 - 6 is that the drag force monitoring systems accurately measure the minimum and maximum peak drag forces along a lateral axis generally parallel to a plane defined by the pad/substrate interface. Another aspect of several embodiments of these methods is that the peak drag forces, or the differences between the maximum peak drag forces and the minimum peak drag forces, can be correlated with time in a peak drag force curve that accurately indicates an estimated exposure time for the endpoint layer. Compared to conventional endpointing methods that may or may not provide a significant signal change as the endpoint layer is exposed, the peak-to-peak processing of the drag force waveform provides a more significant change to identify the exposure of the endpoint layer. As such, several embodiments of the methods described above with reference to FIGS. 2 - 6 can accurately endpoint CMP processing.

[0060] The planarizing machines described above with reference to FIGS. 2 and 7 - 14 are further expected to enhance the accuracy of endpointing CMP processing because they isolate a drag force parameter that is not influenced by energy losses unrelated to the drag force at the pad/substrate interface. In contrast to conventional planarizing processes that endpoint CMP processing using the current of the drive motors, several embodiments of the planarizing machines described above with reference to FIGS. 7 - 14 measure the drag force between the substrate assembly and the polishing pad by isolating the displacement or the lateral forces between either a platen and sub-platen, or a carrier head and a drive shaft. The isolated drag force parameter provides a much more accurate indication of the actual drag force at the pad/substrate interface than measuring motor current because energy losses and other factors associated with moving the carrier head or the polishing pad do not influence or otherwise overshadow the changes in drag force between the pad and the substrate assembly. The endpointing apparatuses and monitoring systems described above with reference to FIGS. 7 - 14 , therefore, are expected to enhance the accuracy of detecting the endpoint in CMP processing.

[0061] From the foregoing it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the invention. Accordingly, the invention is not limited except as by the appended claims.