In chemical mechanical planarization, one area that can potentially impact inter-layer dielectric removal rates is the coefficient of friction associated with the pad, wafer, and slurry abrasive particles. To understand the dominant tribological mechanism present when polishing interlayer dielectrics, Stribeck curves are presented using a dimensionless grouping of CMP specific parameters, called the Sommerfeld number.
So=μU/pδ
eff
where μ is the slurry viscosity, U is the relative pad-wafer velocity, p is the applied wafer pressure, and δeff is the effective fluid thickness. Determination of μ and U are fairly straightforward as can be measured experimentally for a given slurry and U depends upon tool geometry and the relative angular velocity of the platen and wafer.
Wafer pressure is the applied downforce divided by the contact area between the wafer and the pad. Each pad type consist of a different surface area due to different grooving configurations, so each pad will experience a different pressure when subjected to the same normal force. To account for this a dimensionless parameter, α is determined for each pad type equal to the ratio of the total up (raised) area of the pad and the total flat surface area of the pad. The determination of actual pressure experienced by the wafer is 1/α multiplied by the applied pressure. The δeff may moreover be calculated using this α in the following formula1
δeff=α×Ra+(1−α)×δgroove
The coefficient of friction is defined as the quotient of the shear force and the normal force. The plot of the coefficient of friction versus the Sommerfeld number is known as either the Stribeck-Gumbel curve or the McKee Petroff Curve and will be referred to herein as the “Stribeck curve”.2 Recent research has indicated, however, that δeff may be closely quantified as Ra.3 1 Ara Philipossian and Scott Olsen, Fundamental Tribological and Removal Rate Studies of Inter-layer Dielectric Chemical Mechanical Planarization. Jpn. J. Appl. Phys. Vol. 42 (2003) pp. 6371-6379 hereinafter totally incorporated by reference.2 Ibid.3 “Determining the Effects of Slurry Surfactant, Abrasive Size, and Abrasive Content on the Tribology and Kinetics of Copper CMP” by Z. Li, K. Ina, P. Lefevre, I. Koshiyama and A. Philipossian. J. Electrochem. Soc., 152 (2005) G299-G304 hereinafter totally incorporated by reference.
The Stribeck curve is useful in CMP for providing evidence of the extent of contact between the rotating wafer, the rotating pad and the abrasive particles. Three major areas of the Stribeck curve can be distinguished: boundary lubrication at small Sommerfeld numbers where all solid bodies are in intimate contact with each other, mixed lubrication where the wafer and pad are not in intimate contact but some abrasives remain in contact and a thin fluid film about the thickness of the roughness of the pad forms, and hydrodynamic lubrication at large Sommerfeld numbers wherein the fluid film between pad and wafer is larger than the roughness of the pad.4 4 Ibid.
Use of the Stribeck curve can demonstrate what lubrication conditions correspond to what pressure and velocity conditions for a particular pad. Not only is this very useful for characterising the properties of CMP consumables (i.e., polishing pads, slurries and conditioner discs) but also for studying the tribological mechanism for specific pads and slurries and for determining or comparing optimal or other CMP process conditions.
Determination of the Stribeck curve heretofore has involved the determination of COF by measuring shear force in real time with a set normal force that could only be changed by stopping the operation and resetting the load on the wafer and then starting again. In addition to being time consuming, there was also the problem that having stopped a run, it would not be a straightforward matter of starting a subsequent run and obtaining consistent conditions. The necessity of stopping the CMP process to adjust the load was a major problem in the prior art.
In Fundamental Tribological and Removal Rate Studies of Inter-Layer Dielectric Chemical Mechanical Planarization by Philipossian and Olsen5 hereinafter totally incorporated by reference, plotting of Stribeck Curves for specific wafer, pad and slurry combinations was accomplished by use of a scaled polisher capable of measurement of the shear force in real and capable of alteration of the speed of rotation of the wafer and pad during operation. However, there was no direct measurement of the normal force and change in pressure required the process to be stopped and the pressure changed. Although it was possible to construct Stribeck Curves for a number of points based on change in relative velocity alone, it was not possible to do so easily or consistently for change in pressure and both pressure and relative velocity could not be altered at the same time or within the same operational sequence. From the Stribeck Curve it is possible to determine accurately the mechanism of lubrication consequently of wear in CMP polishing procedures for particular wafers, pads and under particular conditions. 5 Ibid. Philipossian Olsen, (2003) pp. 6371-6379
The present invention relates to a method for obtaining data to generate the Stribeck curve and consequently the lubrication mechanism in CMP polishing processes in which both pressure between the polished substrate and polishing pad and relative velocity of the surfaces may be altered in real time during the same CMP processing operation and the normal force component of the coefficient may be measured directly and accurately during the CMP polishing process. More specifically the method of the present invention is accomplished by use of a CMP polishing tool according to U.S. patent application Ser. No. 12/254,291, and more specifically a CMP polishing tool that determines and reports the shear force between the polished substrate and the polishing pad wherein the said determination is accomplished by means of a plate positioned above the wafer head and from which the wafer head and its supporting apparatus hang, or upon which they rest, said plate connected to the framework of the CMP polishing tool by low friction motion means capable of sliding in a direction perpendicular to the line between the center of the polishing pad and the center of the wafer head, and a load cell sensor firmly fixed to the framework of the CMP polishing tool or other immovable structure and positioned to contact and determine the force from the leading edge of said plate when the CMP polishing tool is in operation and the wafer head is in contact with the polishing pad, the signals thus obtained reporting the shear force and additionally in which load cells are attached to and support the bottom of the module of the polishing pad of the CMP polishing tool to determine the normal force, and the normal force and shear force thus obtained may be used to calculate the coefficient of friction between the wafer head and the polishing pad. Further, the said CMP tool is capable of alteration of the pressure and relative velocity during operation by means of stepping motors and a device for the variation of load on the wafer during the CMP polishing procedure. Using the said CMP tool, the pressure and relative velocity may be varied separately or together for the desired length of time according to the needs of the operator so that within one CMP process several measurements may be taken. However, change in load and consequently changes of pressure in the apparatuses and methods of the prior art could only be accomplished by stopping the CMP polishing process and manually changing the load on the wafer or the apparatus and there was no means by which to directly measure the normal force which then had to be estimated.
The inventor(s) of the present invention, in order to solve the problem of obtaining data for the determination of a Stribeck curve that describes the lubrication mechanism for a CMP process using a certain polishing pad, wafer and slurry without having to stop and restart the process to change the pressure and which accurately provides data on the normal force, have with considerable effort and investigation developed the method of the present invention.
More particularly they have devised a method of determination of a Stribeck curve using a CMP polishing tool according to U.S. patent application Ser. No. 12/254,291, and more specifically a CMP polishing tool that determines and reports the shear force between the wafer head and the polishing pad wherein the said determination is accomplished by means of a plate positioned above the wafer head and from which the wafer head and its supporting apparatus hang, or upon which they rest, said plate connected to the framework of the CMP polishing tool by low friction motion means capable of sliding in a direction perpendicular to the line between the center of the polishing pad and the center of the wafer head, and a load cell sensor firmly fixed to the framework of the CMP polishing tool or other immovable structure and positioned to contact and determine the force from the leading of said plate when the CMP polishing tool is in operation and the wafer head is in contact with the polishing pad, the signals thus obtained reporting the shear force and additionally in which load cells are attached to and support the bottom of the module of the polishing pad of the CMP polishing tool to determine the normal force, and the normal force and shear force thus obtained may be used to calculate the coefficient of friction between the wafer head and the polishing pad. Further, the said CMP tool is capable of alteration of the pressure and relative velocity during operation by means of stepping down devices and a device for the variation of load on the wafer during the CMP polishing procedure. Using the said CMP tool, the pressure and relative velocity may be varied separately or together for the desired length of time according to the needs of the operator so that within one CMP process several measurements may be taken.
The method of the present invention has been developed in response to the present state of the art, and in particular, in response to the problems and needs in the art that have not yet been fully solved by currently available CMP methods for measurement of data for determination of lubrication mechanism in CMP. Thus, it is an overall objective of the present invention to provide a method for determination of the lubrication mechanism in CMP and more particularly in the effective acquisition of data for the plotting of Stribeck curves that remedy the shortcomings of the prior art.
The purpose of the method of the present invention is to allow significantly more accurate and more convenient determination of the Stribeck curve for CMP polishing between the wafer and related structures and the CMP polishing pad and by extension to permit significantly more effective, accurate and convenient determination of the lubrication mechanism of CMP.
Through the use of the method for determining the lubricating mechanism in CMP of the present invention, a more accurate, effective and convenient observation of the COF relationship to variation in pressure and relative velocity between the wafer and the CMP polishing pad may be rapidly obtained with a minimum of error. Furthermore, this method may be easily and cost-effectively applied to facilitate and enhance the use of CMP polishing tools and the polishing pads used in them. These and other features and advantages of the present invention will become more fully apparent from the following description and appended claims, or may be learned by the practice of the invention as set forth hereinafter.
The method of the present invention comprises measuring the shear force and normal force between a wafer and a CMP polishing pad on a CMP polishing tool in which the pressure between the wafer and the relative velocity of the wafer and CMP polishing pad may be independently and precisely varied, and to sequentially vary the said pressure, relative velocity or both at set intervals sufficient to obtain a reading for shear force, normal force and COF. This data may be analyzed and the COF values and Stribeck curve determined and output by data processing means such as a computer.
The CMP polishing tool used in the method of the present invention is not particularly limited, provided however, that it should be possible to measure at the shear force in at least one direction and to measure the normal force during polishing and to change precisely during operation the load on the wafer and the rate of rotation of at least the polishing pad and preferably the rate of rotation wafer as well. It is preferred that the CMP polishing tool of the present invention is capable of measuring shear force in two component directions. More preferably, the CMP polishing tool of the present invention possesses a plate positioned above the wafer head from which the wafer head and its supporting apparatus hang, or upon which they rest, said plate connected to the framework of said CMP polishing tool by low friction motion means capable of sliding in a direction perpendicular to the line between the center of the polishing pad and the center of the wafer head, and a load cell sensor firmly fixed to the framework of the CMP polishing tool or other immovable structure and positioned to determine contact and determine force from the leading edge of said plate when the CMP polishing tool is in operation and the wafer head is in contact with the CMP polishing pad, the signals thus obtained reporting the shear force and in which load cells are attached to and support the bottom of the CMP polishing pad module of the CMP polishing tool to determine the normal force, and the shear force and the normal force thus obtained may be used to calculate the coefficient of friction in accordance with U.S. patent application Ser. No. 12/254,291 hereby totally incorporated by reference.
Even more preferably, the CMP polishing tool of the present invention determines two perpendicular force components of shear force and the resulting shear force by means of two plates positioned above the wafer head from which the wafer head and its supporting apparatus hang, or upon which they rest, one plate connected to the framework of said CMP polishing tool by low friction motion means, the other said plate connected to the first said plate by low friction motion means allowing the said other plate to slide in a direction perpendicular to the motion of the first said plate and load sensor cells firmly fixed to the framework of the CMP polishing tool or other immovable structure and positioned to determine contact and determine force from the leading edge of said plate when the CMP polishing tool is in operation and the wafer head is in contact with the CMP polishing pad, and reports the said perpendicular force components as signals and in which load cells are attached to and support the bottom of the CMP polishing pad module of the CMP polishing tool to determine the normal force, and the shear force and the normal force thus obtained are used to calculate the coefficient of friction in accordance with U.S. patent application Ser. No. 12/254,291.
Yet even more preferably, the CMP polishing tool of the present invention determines two perpendicular force components of shear force and the resulting shear force by means of two plates positioned above the wafer head from which the wafer head and its supporting apparatus hang, or upon which they rest, one plate connected to the framework of said CMP polishing tool by low friction rails, the other said plate connected to the first plate by low friction rails allowing the other said plate to slide in a direction perpendicular to the motion of the first said plate, load cells sensor firmly fixed to the framework of the CMP polishing tool or other immovable structure and positioned to contact and determine force from the leading edge of said plate when the CMP polishing tool is in operation and the wafer head is in contact with the polishing pad, the signals of the load cell sensors thus obtained reporting the components of shear force, and in which load cell sensors are attached to and support the bottom of the CMP polishing pad module of the CMP polishing tool to determine the normal force, and the shear force and the normal force thus obtained are delivered to a data processing means to calculate the coefficient of friction in accordance with U.S. patent application Ser. No. 12/254,291.
The means of alteration of the relative velocity between the CMP wafer and the CMP polishing pad during CMP polishing are not particularly limited but stepping down and variable velocity apparatus incorporated into the polishing tool for that purpose either as commercially available or by subsequent modification are preferred. The means of alteration of the pressure between the CMP wafer and the CMP polishing pad during CMP polishing are not particularly limited and any suitable means such as mechanical screws, solenoids, hydraulic systems and the like may be employed either in the system as purchase commercially or as subsequently modified.
Using the aforementioned apparatuses, the relative motion of the wafer may be calculated from the rotation speed of the CMP polishing pad and wafer carrier. If the rotation rate of pad and wafer are significantly different, than the relative motion of wafer has to take into account both rotation rates. The pressure between the wafer and pad is calculated by dividing the load applied to the wafer by the area of the wafer in contact with the CMP polishing pad. The components of shear force and the normal force are obtained directly from the sensors. Calculation may be accomplished manually but use of signal data in calculation by a data processor such as a computer with software prepared for that purpose is preferred.
Any wafer that may be used in CMP polishing may be used in the present invention including without limitation wafers with surfaces of copper, aluminium, tungsten, silicon dioxide and the like. Patterned wafers may also be used. Likewise, any polishing pad that may be used in CMP polishing may be used in the present invention. The conditions of the CMP polishing operation that may be employed are not particularly limited and conditions that are normally used in CMP polishing of various types of CMP wafers may be used. The length of the CMP polishing run is not particularly limited but between 30 seconds and 3 minutes is preferred. Conditioning pads may be used to roughen the surface of the CMP polishing pad during CMP polishing. The temperature at which the CMP polishing process is carried out is not limited but between 10 degrees and 45 degrees centigrade is preferred and room temperature is more preferred.
The type of slurry used in the present invention is not particularly limited and any slurry applicable to CMP polishing may be used including without limitation any commercially available alumina, silica or ceria slurry. Since the viscosity of the slurry has not been regarded as changing significantly under conditions of the present invention, it may be regarded as constant for purposes of the present invention; however, it may be desirable to determine the viscosity of the slurry for the purpose of more detailed reporting of the relationship between the resulting COF data and the relative velocity of the wafer and CMP polishing pad and the pressure between them. There is no one tool that can provide this information but a combination of tools and techniques that can render the determination of local viscosity as accurately as possible is preferred. The viscosity of commercial slurries, if not already provided as part of the specifications of the product when sold may easily be determined by conventional methods.
The number of measurements taken using the method of the present invention during a single CMP polishing procedure is not limited. However, a number of between 3 and 15 is preferred and between 5 and 10 is more preferred. If the number is too large, the time period for measurement between changes becomes too short and since force and COF measurements tend to show a short period of adjustment of a second or less, it is desirable to have a long enough test period to yield reliable data. A time period that is too long is wasteful in that the advantage of the present invention is to allow multiple measurements of the same system under the same conditions except for pressure and relative velocity and a long period means fewer runs. The length of the period for each measurement may be the same or may vary, but equal periods of measurement for each set of process conditions is preferred. Time periods of between 5 and 70 seconds are preferred and between 10 and 40 seconds are more preferred. Measurement of shear and down forces itself may be done at high frequency for example, cases, above 500 hz. And more preferably 1000 Hz or
Sequences of changes in measurements are not particularly limited. Measurements may proceed by ascending or descending value of either relative velocity or pressure with the other variable maintained constant, or they may both be increased or decreased at the same time. One may be increased and the other decreased at the same time or one may be increased or decreased while the other is kept constant and then the other increased or decreased while the former is kept constant. The change may also be entirely random. Specific measurement conditions may also be repeated within a single CMP polishing procedure as a way of verifying whether there is any difference at one point in a CMP polishing procedure as compared with another point later on.
The minimum difference in pressure conditions between measurements is not particularly limited but variations of greater than ¼th psi are preferred. The minimum difference in relative velocity conditions between measurements is not particularly limited but differences greater than 0.2 meters per second are preferred.
The practice of the present invention is demonstrated without being limited by reference to the following practice example:
An Araca APD-800 polisher equipped with a single plate shear force measuring device, load sensors beneath the CMP polishing pad module and stepping motors and fixtures for variation of the pressure applied to the wafer CMP polishing pad contact was used as the polishing tool and it was equipped with a MMC TRD 100 grit diamond conditioner disc used in situ under a conditioning force of 5.8 lbf with 30 RPM rotation and 10 per minute sweep frequency. The polishing pad used was a Rohm and Haas IC1000-k-groove pad. The silica based slurry was introduced at a flow rate of 300 cc/minute. The wafer used was a 200 mm blanket copper wafer. The wafer was polished under these conditions for a total of 150 seconds during which the pressure on the wafer, the relative velocity of the wafer and CMP polishing pad or both were changed as indicated in Table 1 approximately forty seconds after starting and approximately every thirty seconds thereafter. The results of shear force, normal force and COF are shown in
Except that the sequence of pressure and relative velocity were as shown in Table 1, and a different 200 mm blanket copper wafer used, a CMP polishing procedure was carried out in the same way as in Example 1. The results of shear force, normal force and COF are shown in
As can be observed clearly from the results in