WO2001050496A1

WO2001050496A1 - Automated high-density plasma (hdp) workpiece temperature control

Info

Publication number: WO2001050496A1
Application number: PCT/US2000/025725
Authority: WO
Inventors: Allen Lewis Evans
Original assignee: Advanced Micro Devices, Inc.
Priority date: 2000-01-04
Filing date: 2000-09-20
Publication date: 2001-07-12
Also published as: JP2003519906A; KR20020063616A; EP1245037A1

Abstract

A method is provided for manufacturing, the method including processing a workpiece (305) in a high-density plasma (HDP) processing step (310), measuring a parameter (317) characteristic of the HDP processing performed on the workpiece in the HDP processing step (310), and modeling the characteristic parameter measured (330, 335). The method also includes applying the model to modify at least one HDP control input parameter (340, 360, 320).

Description

AUTOMATED HIGH-DENSITY PLASMA (HDP) WORKPIECE TEMPERATURE CONTROL

TECHNICAL FIELD

This invention relates generally to semiconductor fabrication technology, and more particularly, to a method for automating workpiece temperature control during high-density plasma (HDP) etch and/or deposition processes

BACKGROUND ART There is a constant drive within the semiconductor industry to increase the quality, reliability and throughput of integrated circuit devices, e g . microprocessors, memory devices, and the like This drive is fueled by consumer demands for higher quality computers and electronic devices that operate more reliably These demands have resulted in a continual improvement in the manufacture of semiconductor devices, e g , transistors, as well as in the manufacture of integrated circuit devices incorporating such transistors Additionally, reducing defects in the manufacture of the components of a typical transistor also lowers the overall cost per transistor as well as the cost of integrated circuit devices incorporating such transistors The technologies underlying semiconductor processing tools have attracted increased attention over the last several years, resulting in substantial refinements However, despite the advances made in this area, many of the processing tools that are currently commercially available suffer certain deficiencies In particular, such tools often lack advanced process data monitoring capabilities, such as the ability to provide historical parametric data in a user-friendly format, as well as event logging, real-time graphical display of both current processing parameters and the processing parameters of the entire run, and remote, / e , local site and worldwide, monitoring These deficiencies can engender nonoptimal control of critical processing parameters, such as throughput accuracy, stability and repeatability, processing temperatures, mechanical tool parameters, and the like This variability manifests itself as within-run disparities, run-to-run disparities and tool-to-tool disparities that can propagate into deviations in product quality and performance, whereas an ideal monitoring and diagnostics system for such tools would provide a means of monitoring this variability, as well as providing means for optimizing control of critical parameters

Among the parameters it would be useful to monitor and control are the thickness of a layer formed on the workpiece, the bulk physical properties of a material formed on the workpiece, and the overall uniformity For example, it would be useful to monitor and control the thickness of a dielectric film deposited on the workpiece during an HDP etch and/or deposition process, the refractive index of the dielectric film, and thickness variations across the workpiece of the dielectric film

However, traditional statistical process control (SPC) techniques are often inadequate to control precisely such parameters in semiconductor and microelectronic device manufacturing so as to optimize device performance and yield Typically, SPC techniques set a target value, and a spread about the target value, for such parameters The SPC techniques then attempt to minimize the deviation from the target value without automatically adjusting and adapting the respective target values to optimize the semiconductor device performance, and/or to optimize the semiconductor device yield and throughput Furthermore, blindly minimizing non-adaptive processing spreads about target values may not increase processing yield and throughput

In particular, variations in the parameters for films deposited and/or etched by HDP processes typically result from poor control of the temperature of the workpiece during HDP processes The HDP process typically uses an electromagnetic chuck to clamp, hold, and cool the workpiece during the HDP processing The workpiece temperature T is a complicated function of the argon (Ar) sputtering parameters, the workpiece contact w ith the chuck, and the cooling of the chuck by water flow cooling and/or by helium (He) flow cooling, so the co entional approach uses manual control based on a process engineer s experience and potentially fallible judgment

The present invention is directed to overcoming, or at least reducing the effects of, one or more of the problems set forth above

DISCLOSURE OF INVENTION

In one aspect of the present invention, a method is provided for manufacturing, the method including processing a workpiece in a high-density plasma (HDP) processing step, measuring a parameter characteristic of the HDP processing performed on the workpiece in the HDP processing step, and modeling the characteristic parameter measured The method also includes applying the model to modify at least one HDP control input parameter

In another aspect of the present invention, a computer-readable, program storage device is provided, encoded with instructions that, when executed by a computer, perform a method for manufacturing a workpiece, the method including processing a workpiece in a high-density plasma (HDP) processing step, measuring a parameter characteristic of the HDP processing performed on the workpiece in the HDP processing step and modeling the characteristic parameter measured The method also includes applying the model to modιf\ at least one HDP control input parameter

In yet another aspect of the present invention, a computer programmed to perform a method of manufacturing is provided, the method including processing a workpiece in a high-density plasma (HDP) processing step, measuring a parameter characteristic of the HDP processing performed on the workpiece in the

HDP processing step, and modeling the characteristic parameter measured The method also includes apphing the model to modify at least one HDP control input parameter

BRIEF DESCRIPTION OF THE DRAWINGS The invention may be understood by reference to the following description taken in conjunction w ith the accompanying drawings, in which the leftmost significant dιgιt(s) in the reference numerals denote(s) the first figure in which the respective reference numerals appear, and in which

Figures 1-7 schematically illustrate various embodiments of a method for manufacturing according to the present invention, and, more particularly

Figure 1 schematically illustrates a method for fabricating a semiconductor device practiced in accordance with the present invention,

Figure 2 schematically illustrates workpieces being processed using a high-density plasma (HDP) processing tool, using a plurality of control input signals, in accordance with the present invention,

Figures 3-4 schematically illustrate one particular embodiment of the process and tool in Figure 2, Figure 5 schematically illustrates one particular embodiment of the method of Figure 1 as may be practiced with the process and tool of Figures 3-4,

Figure 6 schematically illustrates conventional HDP deposition of a film on high aspect ratio metal lines, and

Figure 7 schematically illustrates an HDP deposition of a film on high aspect ratio metal lines in accordance with various embodiments of the present invention While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof have been shown by way of example in the drawings and are herein described in detail It should be understood, however, that the description herein of specific embodiments is not intended to limit the invention to the particular forms disclosed, but on the contrary the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims

MODE(S) FOR CARRYING OUT THE INVENTION Illustrative embodiments of the invention are described below In the interest of clarity, not all features of an actual implementation are described in this specification It will of course be appreciated that in the development of any such actual embodiment, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance w ith system-related and business-related constraints, which will vary from one implementation to another Moreover it will be appreciated that such a development effort might be complex and time-consuming, but would nev ertheless be a routine undertaking for those of ordinary skill in the art having the benefit of this disclosure

Figure 1 illustrates one particular embodiment of a method 100 practiced in accordance with the present invention Figure 2 illustrates one particular apparatus 200 with which the method 100 ma be practiced For the sake of clarity, and to further an understanding of the invention the method 100 shall be disclosed in the context of the apparatus 200 However, the invention is not so limited and admits wide variation, as is discussed further below

Referring now to both Figures 1 and 2, a batch or lot of workpieces or wafers 205 is being processed through a high-density plasma (HDP) processing tool 210 The HDP processing tool 210 may be any HDP processing tool known to the art, such as a Novellus HDP tool, provided it includes the requisite control capabilities The HDP processing tool 210 includes an HDP processing tool controller 215 for this purpose The nature and function of the HDP processing tool controller 215 will be implementation specific For instance, the HDP processing tool controller 215 may control HDP control input parameters such as HDP recipe control input parameters The HDP recipe control input parameters may include HDP control input parameters for water flow cooling, helium flow cooling, argon sputtering, electrostatic chuck clamping voltage, and the like Four workpieces 205 are shown in Figure 2, but the lot of workpieces or wafers, i e , the "wafer lot," may be any practicable number of wafers from one to any finite number

The method 100 begins, as set forth in box 120, by measuring a parameter characteristic of the HDP processing performed on the workpiece 205 in the HDP processing tool 210 The nature, identity, and measurement of characteristic parameters will be largely implementation specific and even tool specific For instance, capabilities for monitoring process parameters vary, to some degree, from tool to tool Greater sensing capabilities may permit wider latitude in the characteristic parameters that are identified and measured and the manner in which this is done Conversely, lesser sensing capabilities may restrict this latitude For example, the Novellus HDP tool reads the temperature of a workpiece 205, and/or an average of the temperatures of the workpieces 205 in a lot, using a pyrometer, and the pyrometer needs to know the emissivity of the workpiece 205 and/or workpieces 205 being read, but this emissivity may vary from wafer to wafer The Novellus HDP tool pyrometer typically does not feedback the temperature information to the Novellus HDP tool The temperature of a workpiece 205, and/or an average of the temperatures of the workpieces 205 in a lot, is an illustrative example of a parameter characteπstic of the HDP processing performed on the workpiece in the HDP processing tool 210

Turning to Figure 2, in this particular embodiment, the HDP process characteristic parameters are measured and/or monitored by tool sensors (not shown) The outputs of these tool sensors are transmitted to a computer system 230 over a line 220 The computer system 230 analyzes these sensor outputs to identify the characteristic parameters

Returning, to Figure 1 , once the characteristic parameter is identified and measured the method 100 proceeds by modeling the measured and identified characteπstic parameter, as set forth in box 130 The computer system 230 in Figure 2 is, in this particular embodiment programmed to model the characteristic parameter The manner in which this modeling occurs will be implementation specific

In the embodiment of Figure 2, a database 235 stores a plurality of models that might potentially be applied, depending upon which characteristic parameter is identified This particular embodiment therefore, requires some a priori knowledge of the characteπstic parameters that might be measured The computer system 230 then extracts an appropriate model from the database 235 of potential models to apply to the identified characteristic parameters If the database 235 does not include an appropriate model, then the characteristic parameter may be ignored, or the computer system 230 may attempt to develop one, if so programmed The database 235 may be stored on any kind of computer-readable, program storage medium, such as an optical disk 240, a floppy disk 245, or a hard disk drive (not shown) of the computer system 230 The database 235 may also be stored on a separate computer system (not shown) that interfaces with the computer system 230

Modeling of the identified characteristic parameter may be implemented differently in alternative embodiments For instance, the computer system 230 may be programmed using some form of artificial intelligence to analyze the sensor outputs and controller inputs to develop a model on-the-fly in a real-time implementation This approach might be a useful adjunct to the embodiment illustrated in Figure 2, and discussed above, where characteristic parameters are measured and identified for which the database 235 has no appropriate model

The method 100 of Figure 1 then proceeds by applying the model to modify an HDP control input parameter, as set forth in box 140 Depending on the implementation, applying the model may yield either a new value for the HDP control input parameter or a correction to the existing HDP control input parameter The new HDP control input is then formulated from the value yielded by the model and is transmitted to the HDP processing tool controller 215 over the line 220 The HDP processing tool controller 215 then controls subsequent HDP process operations in accordance with the new HDP control inputs

Some alternative embodiments may employ a form of feedback to improve the modelmg of characteristic parameters The implementation of this feedback is dependent on several disparate facts, including the tool's sensing capabilities and economics One technique for doing this would be to monitor at least one effect of the model's implementation and update the model based on the effect(s) monitored The update may also depend on the model For instance, a linear model may require a different update than would a non-linear model, all other factors being the same

As is evident from the discussion above, some features of the present invention are implemented in software For instance, the acts set forth in the boxes 120-140 in Figure 1 are, in the illustrated embodiment, software-implemented, m whole or in part Thus, some features of the present invention are implemented as instructions encoded on a computer-readable, program storage medium The program storage medium may be of any type suitable to the particular implementation However, the program storage medium will typically be magnetic, such as the floppy disk 245 or the computer 230 hard disk drive (not shown), or optical, such as the optical disk 240 When these instructions are executed by a computer, they perform the disclosed functions The computer may be a desktop computer, such as the computer 230 However, the computer might alternatively be a processor embedded in the HDP processing tool 210 The computer might also be a laptop a workstation, or a mainframe in various other embodiments The scope of the invention is not limited by the type or nature of the program storage medium or computer with which embodiments of the invention might be implemented

Thus, some portions of the detailed descriptions herein are, or may be, presented in terms of algorithms, functions, techniques, and/or processes These terms enable those skilled in the art most effectively to convey the substance of their work to others skilled m the art These terms are here, and are generally, conceived to be a self-consistent sequence of steps leading to a desired result The steps are those requiring physical manipulations of physical quantities Usually, though not necessarily, these quantities take the form of electromagnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated It has proven convenient at times principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters terms, numbers, and the like All of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities and actions Unless specifically stated otherwise, or as may be apparent from the discussion terms such as

"processing," "computing," "calculating," "determining," "displaying," and the like, used herein refer to the actιon(s) and processes of a computer system, or similar electronic and/or mechanical computing device, that manipulates and transforms data, represented as physical (electromagnetic) quantities within the computer system's registers and/or memories, into other data similarly represented as physical quantities within the computer system's memories and/or registers and/or other such information storage, transmission and/or display devices

Construction of an Illustrative Apparatus. An exemplary embodiment 300 of the apparatus 200 in Figure 2 is illustrated in Figures 3-4, in which the apparatus 300 comprises a portion of an Advanced Process Control ("APC") system Figures 3-4 are conceptualized, structural and functional block diagrams respectively, of the apparatus 300 A set of processing steps is performed on a lot of wafers 305 on an HDP processing tool 310 Because the apparatus 300 is part of an APC system, the wafers 305 are processed on a run-to-run basis Thus, process adjustments are made and held constant for the duration of a run, based on run-level measurements or averages A "run" may be a lot, a batch of lots, or even an individual wafer

In this particular embodiment, the wafers 305 are processed by the HDP processing tool 310 and various operations in the process are controlled by a plurality of HDP control input signals on a line 320 between the HDP processing tool 310 and a workstation 330 Exemplary HDP control inputs for this embodiment might include a water flow cooling signal, a helium flow cooling signal, and an argon sputtering signal, an electrostatic chuck clamping voltage signal, and the like

When a process step in the HDP processing tool 310 is concluded, the semiconductor wafers 305 being processed in the HDP processing tool 310 is examined in a review station 317 The HDP control inputs generally affect the temperature of the semiconductor wafers 305 and, hence, the variability and properties of the dielectric film etched/deposited by the HDP processing tool 310 on the wafers 305 Once errors are determined from the examination after the run of a lot of wafers 305, the HDP control inputs on the line 320 are modified for a subsequent run of a lot of wafers 305 Modifying the control signals on the line 320 is designed to improve the next process step m the HDP processing tool 310 The modification is performed in accordance with one particular embodiment of the method 100 set forth in Figure 1, as described more fully below Once the relevant HDP control input signals for the HDP processing tool 310 are updated, the HDP control input signals with new settings are used for a subsequent run of semiconductor devices Referring now to both Figures 3 and 4, the HDP processing tool 310 communicates with a manufacturing framework comprising a network of processing modules One such module is an APC system manager 440 resident on the computer 340 This network of processing modules constitutes the APC system The HDP processing tool 3 10 generally includes an equipment interface 410 and a sensor interface 415 A machine interface 430 resides on the workstation 330 The machine interface 430 bridges the gap between the APC framework, e g the APC system manager 440, and the equipment interface 410 Thus the machine interface 430 interfaces the HDP processing tool 310 with the APC framework and supports machine setup, activation, monitoring, and data collection The sensor interface 415 provides the appropriate interface environment to communicate ith external sensors such as LabView^® or other sensor bus-based data acquisition software Both the machine interface 430 and the sensor interface 415 use a set of functionalities such as an RS232 communication standard, to collect data to be used The equipment interface 410 and the sensor interface 415 communicate over the line 320 with the machine interface 430 resident on the workstation 330

More particularly, the machine interface 430 receives commands, status events, and collected data from the equipment interface 410 and forwards these as needed to other APC components and event channels In tum, responses from APC components are received by the machine interface 430 and rerouted to the equipment interface 410 The machine interface 430 also reformats and restructures messages and data as necessary The machine interface 430 supports the startup/shutdown procedures within the APC System Manager 440 It also serves as an APC data collector, buffering data collected by the equipment interface 410, and emitting appropriate data collection signals In the particular embodiment illustrated, the APC system is a factory-wide software system, but this is not necessary to the practice of the invention The control strategies taught by the present invention can be applied to virtually any semiconductor HDP processing tool on a factory floor Indeed, the present invention may be simultaneously employed on multiple HDP processing tools in the same factory or in the same fabrication process The APC framework permits remote access and monitoring of the process performance Furthermore by utilizing the APC framework, data storage can be more convenient, more flexible, and less expensive than data storage on local drives However, the present invention may be employed, in some alternative embodiments, on local drives

The illustrated embodiment deploys the present invention onto the APC framework utilizing a number of software components In addition to components within the APC framework, a computer script is written for each of the semiconductor HDP processing tools involved in the control system When a semiconductor HDP processing tool m the control system is started in the semiconductor manufacturing fab, the semiconductor HDP processing tool generally calls upon a script to initiate the action that is required by the HDP processing tool controller The control methods are generally defined and performed using these scripts The development of these scripts can comprise a significant portion of the development of a control system

In this particular embodiment, there are several separate software scripts that perform the tasks involved in controlling the HDP processing operation There is one script for the HDP processing tool 310, including the review station 317 and the HDP processing tool controller 315 There is also a script to handle the actual data capture from the review station 317 and another script that contains common procedures that can be referenced by any of the other scripts There is also a script for the APC system manager 440 The precise number of scripts, however, is implementation specific and alternative embodiments may use other numbers of scripts Operation of an Illustrative Apparatus. Figure 5 illustrates one particular embodiment 500 of the method 100 in Figure 1 The method 500 may be practiced with the apparatus 300 illustrated in Figures 3-4 but the invention is not so limited The method 500 may be practiced with any apparatus that may perform the functions set forth in Figure 5 Furthermore, the method 100 in Figure 1 be practiced in embodiments alternative to the method 500 in Figure 5

Referring now to all of Figures 3-5, the method 500 begins with processing a lot of w afers 305 through an HDP processing tool 310, as set forth in box 510 In this particular embodiment the HDP processing tool 310 has been initialized for processing by the APC system manager 440 through the machine interface 430 and the equipment interface 410 In this particular embodiment before the HDP processing tool 310 is run the APC system manager script is called to initialize the HDP processing tool 310 At this step, the script records the identification number of the HDP processing tool 310 and the lot number of the wafers 305 The identification number is then stored against the lot number in a data store 360 The rest of the script, such as the APCData call and the Setup and StartMachine calls, are formulated with blank or dummy data in order to force the machine to use default settings

As part of this initialization, the initial setpoints for HDP control are provided to the HDP processing tool controller 315 over the line 320 These initial setpoints may be determined and implemented in any suitable manner known to the art In the particular embodiment illustrated, HDP controls are implemented by control threads Each control thread acts like a separate controller and is differentiated by various process conditions For HDP control, the control threads are separated by a combination of different conditions These conditions may include, for example, the semiconductor HDP processing tool 310 currently processing the wafer lot, the semiconductor product, the semiconductor manufacturing operation, and one or more of the semiconductor processing tools (not shown) that previously processed the semiconductor wafer lot

Control threads are separated because different process conditions affect the HDP error differently By isolating each of the process conditions into its own corresponding control thread, the HDP error can become a more accurate portrayal of the conditions in which a subsequent semiconductor wafer lot in the control thread will be processed Since the error measurement is more relevant changes to the HDP control input signals based upon the error will be more appropriate

The control thread for the HDP control scheme depends upon the current HDP processing tool, current operation, the product code for the current lot, and the identification number at a previous processing step The first three parameters are generally found in the context information that is passed to the script from the HDP processing tool 310 The fourth parameter is generally stored when the lot is previously processed Once all four parameters are defined, they are combined to form the control thread name, HDP02 OPER01 PROD01 HDP01 is an example of a control thread name The control thread name is also stored in correspondence to the wafer lot number in the data store 360

Once the lot is associated with a control thread name, the initial settings for that control thread are generally retrieved from the data store 360 There are at least two possibilities when the call is made for the information One possibility is that there are no settings stored under the current control thread name This can happen when the control thread is new, or if the information was lost or deleted In these cases, the script initializes the control thread assuming that there is no error associated with it and uses the target values of the HDP errors as the HDP control input settings It is preferred that the controllers use the default machine settings as the initial settings By assuming some settings, the HDP errors can be related back to the control settings in order to facilitate feedback control Another possibility is that the initial settings are stored under the control thread name In this case one or more wafer lots have been processed under the same control thread name as the current wafer lot and have also been measured for HDP error using the review station 317 When this information exists the HDP control input signal settings are retrieved from the data store 360 These settings are then downloaded to the HDP processing tool 310

The wafers 305 are processed through the HDP processing tool 310 This includes in the embodiment illustrated dielectric film or layer etch and/or deposition and/or etch/deposition The wafers 305 are measured on the review station 317 after their HDP processing on the HDP processing tool 310 The review station 317 examines the wafers 305 after they are processed for a number of errors The data generated by the instruments of the review station 317 is passed to the machine interface 430 via sensor interface 415 and the line 320 The review station script begins with a number of APC commands for the collection of data The review station script then locks itself in place and activates a data available script This script facilitates the actual transfer of the data from the review station 317 to the APC framework Once the transfer is completed the script exits and unlocks the review station script The interaction with the review station 317 is then generally complete As will be appreciated by those skilled in the art having the benefit of this disclosure the data generated by the review station 317 should be preprocessed for use Review stations, such as KLA review stations, provide the control algorithms for measuring the control error Each of the error measurements in this particular embodiment, corresponds to one of the HDP control input signals on the line 320 in a direct manner Before the error can be utilized to correct the HDP control input signal, a certain amount of preprocessing is generally completed

For example, preprocessing may include outlier rejection Outlier rejection is a gross error check ensuring that the received data is reasonable in light of the historical performance of the process This procedure involves comparing each of the HDP errors to its corresponding predetermined boundary parameter In one embodiment, even if one of the predetermined boundaries is exceeded, the error data from the entire semiconductor wafer lot is generally rejected

To determine the limits of the outlier rejection, thousands of actual semiconductor manufacturing fabrication ("fab") data points are collected The standard deviation for each error parameter in this collection of data is then calculated In one embodiment, for outlier rejection, nine times the standard deviation (both positive and negative) is generally chosen as the predetermined boundary This was done primarily to ensure that only the points that are significantly outside the normal operating conditions of the process are rejected

Preprocessing may also smooth the data, which is also known as filtering Filtering is important because the error measurements are subject to a certain amount of randomness, such that the error significantly deviates in value Filtering the review station data results in a more accurate assessment of the error in the HDP control input signal settings In one embodiment, the HDP control scheme utilizes a filtering procedure known as an Exponentially-Weighted Moving Average ("EWMA") filter, although other filtering procedures can be utilized in this context

One embodiment for the EWMA filter is represented by Equation (1)

AVG_N = W * M_C + (1-W) * AVGp (1) where AVG_N ≡ the new EWMA average,

W ≡ a weight for the new average (AVG_\), M_c ≡ the current measurement; and AVGp ≡ the previous EWMA average.

The weight is an adjustable parameter that can be used to control the amount of filtering and is generally between zero and one. The weight represents the confidence in the accuracy of the current data point. If the measurement is considered accurate, the weight should be close to one. If there were a significant amount of fluctuations in the process, then a number closer to zero would be appropriate.

In one embodiment, there are at least two techniques for utilizing the EWMA filtering process. The first technique uses the previous average, the weight, and the current measurement as described above. Among the advantages of utilizing the first implementation are ease of use and minimal data storage. One of the disadvantages of utilizing the first implementation is that this method generally does not retain much process information. Furthermore, the previous average calculated in this manner would be made up of every data point that preceded it, which may be undesirable. The second technique retains only some of the data and calculates the average from the raw data each time.

The manufacturing environment in the semiconductor manufacturing fab presents some unique challenges. The order that the semiconductor wafer lots are processed through an HDP processing tool may not correspond to the order in which they are read on the review station. This could lead to the data points being added to the EWMA average out of sequence. Semiconductor wafer lots may be analyzed more than once to verify the error measurements. With no data retention, both readings would contribute to the EWMA average, which may be an undesirable characteristic. Furthermore, some of the control threads may have low volume, which may cause the previous average to be outdated such that it may not be able to accurately represent the error in the HDP control input signal settings.

The HDP processing tool controller 315, in this particular embodiment, uses limited storage of data to calculate the EWMA filtered error, i.e., the first technique. Wafer lot data, including the lot number, the time the lot was processed, and the multiple error estimates, are stored in the data store 360 under the control thread name. When a new set of data is collected, the stack of data is retrieved from data store 360 and analyzed. The lot number of the current lot being processed is compared to those in the stack. If the lot number matches any of the data present there, the error measurements are replaced. Otherwise, the data point is added to the current stack in chronological order, according to the time periods when the lots were processed. In one embodiment, any data point within the stack that is over 48 hours old is removed. Once the aforementioned steps are complete, the new filter average is calculated and stored to data store 360.

Thus, the data is collected and preprocessed, and then processed to generate an estimate of the current errors in the HDP control input signal settings. First, the data is passed to a compiled Matlab® plug-in that performs the outlier rejection criteria described above. The inputs to a plug-in interface are the multiple error measurements and an array containing boundary values. The return from the plug-in interface is a single toggle variable. A nonzero return denotes that it has failed the rejection criteria, otherwise the variable returns the default value of zero and the script continues to process.

After the outlier rejection is completed, the data is passed to the EWMA filtering procedure. The controller data for the control thread name associated with the lot is retrieved, and all of the relevant operation upon the stack of lot data is carried out. This includes replacing redundant data or removing older data. Once the data stack is adequately prepared, it is parsed into ascending time-ordered arrays that correspond to the error values These arrays are fed into the EWMA plug-in along with an array of the parameter required for its execution In one embodiment the return from the plug-in is comprised of the six filtered error values

Returning to Figure 5, data preprocessing includes measuring a characteristic parameter in an HDP operation, such as workpiece 305 temperature, arising from HDP processing control of the HDP processing tool 310, as set forth in box 520 Known, potential characteristic parameters may be identified by characteristic data patterns or may be identified as known consequences to modifications to critical dimension control The example of how changes in temperature affect deposition variability of the etch/deposited dielectric film given above falls into this latter category

The next step in the control process is to calculate the new settings for the HDP processing tool controller 315 of the HDP processing tool 310 The previous settings for the control thread corresponding to the current wafer lot are retrieved from the data store 360 This data is paired along with the current set of HDP errors The new settings are calculated by calling a compiled Matlab® plug-in This application incorporates a number of inputs, performs calculations in a separate execution component, and returns a number of outputs to the mam script Generally, the inputs of the Matlab® plug-in are the HDP control input signal settings, the review station errors, an array of parameters that are necessary for the control algorithm, and a currently unused flag error The outputs of the Matlab® plug-in are the new controller settings calculated in the plug-in according to the controller algorithm described above

A HDP process engineer or a control engineer, who generally determines the actual form and extent of the control action, can set the parameters They include the threshold values, maximum step sizes, controller weights, and target values Once the new parameter settings are calculated, the script stores the setting in the data store 360 such that the HDP processing tool 310 can retrieve them for the next wafer lot to be processed The principles taught by the present invention can be implemented into other types of manufacturing frameworks

Returning again to Figure 5, the calculation of new settings includes, as set forth in box 530, modeling the identified characteristic parameter This modeling may be performed by the Matlab® plug-m In this particular embodiment, only known, potential characteristic parameters are modeled and the models are stored in a database 335 accessed by a machine interface 430 The database 335 may reside on the workstation 330, as shown, or some other part of the APC framework For instance, the models might be stored in the data store 360 managed by the APC system manager 440 in alternative embodiments The model will generally be a mathematical model, i e , an equation describing how the change(s) in HDP recipe control(s) affects the HDP performance and the dielectric film properties such as deposition uniformity, film thickness variation across the wafer, film refractive index, and the like For example, an HDP sputter rate f(x) may be given by the equation f(x) = a(HFRFpower) + b(Argonflow) + c(Oxygenflow) + d (Temperature) , where the coefficients a, b, c and d have the appropriate units and dimensions

The particular model used will be implementation specific, depending upon the particular HDP processing tool 310 and the particular characteristic parameter being modeled Whether the relationship in the model is linear or non-linear will be dependent on the particular parameters involved

The new settings are then transmitted to and applied by the HDP processing tool controller 315 Thus, returning now to Figure 5, once the identified characteristic parameter is modeled, the model is applied to modify at least one HDP recipe control input parameter, as set forth in box 540 In this particular embodiment, the machine interface 430 retrieves the model from the database 335, plugs in the respective value(s), and determines the necessary change(s) in the HDP recipe control input parameter(s) The change is then communicated by the machine interface 430 to the equipment interface 410 ov er the line 320 The equipment interface 410 then implements the change

The present embodiment furthermore provides that the models be updated This includes as set forth in boxes 550-560 of Figure 5, monitoring at least one effect of modifving the HDP recipe control input parameters (box 550) and updating the applied model (box 560) based on the effect(s) monitored For instance various aspects of the HDP processing tool 310's operation will change as the HDP processing tool 310 ages By monitoring the effect of the HDP recipe change(s) implemented as a result of the characteristic parameter (e g , workpiece 305 temperature) measurement, the necessary value could be updated to yield superior performance

As noted above, this particular embodiment implements an APC system Thus changes are implemented "between" lots The actions set forth in the boxes 520-560 are implemented after the current lot is processed and before the second lot is processed, as set forth in box 570 of Figure 5 However, the invention is not so limited Furthermore, as noted above, a lot may constitute any practicable number of wafers from one to several thousand (or practically any finite number) What constitutes a "lot is implementation specific, and so the point of the fabrication process in which the updates occur will vary from implementation to implementation Hierarchical Ordering of Conditions. One particular variation of this implementation includes a hierarchical ordering of conditions that constitute a control thread Hierarchical ordering is taught generally, without reference to the present invention, in Application Serial No 09/371,665, filed 8/10/1999, entitled "Method and Apparatus for Performing Run-to-Run Control in a Batch Manufacturing Environment, ' by Anthony J Toprac, William J Campbell, and Christopher A Bode (Attorney Docket No 2000 015200/TT3078) In one embodiment, the hierarchical ordering of control thread data is related to the strength of the effects that these conditions exert on the control of a manufacturing process

Generally, a number of discrete process factors will affect the performance of a given manufacturing process These factors can be arranged in a hierarchy The different levels in the hierarchy can be arranged in order of the relative impact each of the previous process factors will have on the variance of the present manufacturing process Each wafer lot processed will generally involve a discrete value from each hierarchy, but will contribute process information to multiple hierarchical levels Control of each wafer lot will use the lowest hierarchical level (the second level being lower than the first level) for which there is previous process metrology information Generally, the processing of each wafer lot will add information to each hierarchical level

More particularly, the HDP control inputs of the HDP processing tool used in a process may be defined as the first level of a hierarchical ordering of control thread data The HDP control inputs of a previous operation may be the second most influential factor on the HDP control inputs of the present process, which can be defined as the second level of the hierarchical ordering of control thread data The HDP control inputs relating to a similar product type may be the third most influential factor on the HDP control inputs of the present process, which can be defined as the third level of the hierarchical ordering of control thread data Further levels of the hierarchical ordering of control thread data can be defined using other similarities between previous processes and a current process

Using the aforementioned hierarchical ordering of control thread data, an automatic spawning of control threads can be implemented In one embodiment, initially, the HDP control inputs for a plurality of semiconductor lots are placed into a single control thread When sufficient data is present to prove a statistically significant difference between lots belonging to different hierarchical levels, a process controller splits the initial control thread into two control threads As more data is obtained, new control threads that represent different hierarchical levels are generated The modeled characteristic parameter may be one factor defining or ordering the hierarchy as described above

Generally speaking, one control thread is used to run a manufacturing process on a lot of semiconductor devices, much as is described above for Figures 3-5 Metrology process data is acquired and HDP control input errors are calculated The HDP control inputs for the next process run are modified on a run-to-run basis, based upon errors detected from a previous manufacturing run Based upon the errors detected from the previous manufacturing run new HDP control input settings are determined and an appropriate bin in a hierarchical level is filled with data, effectively creating new control threads

Figure 6 schematically illustrates conventional HDP deposition of a film on high aspect ratio metal lines As shown in Figure 6, a structure layer 600 of a workpiece (such as a semiconducting wafer) has two conducting lines 605 (such as metal lines) formed thereon For example, the conducting lines 605 may be parallel bit and/or digit lines on a DRAM chip The conducting lines 605 may have a thickness t_meta) in a range of approximately 3000 A to 7000 A, for example An opening and/or space 610 may be formed between the conducting lines 605, the space 610 having a width w in a range of approximately 2000 A to 4000 A, for example The space 610 may have sides 61 OS and a bottom 610B The aspect ratio r of the space 610 may be give by the thickness t_raeta| divided by the width w r = t_meta|/w

The conducting lines 605 may be electrically isolated from each other by a dielectric film 615, formed of silicon dioxide (Sι0₂), for example The dielectric film 615 may have a thickness t„ι_m in a range of approximately 5000 A to 10000 A, for example Conventional HDP etch/deposition of S1O7, using silane (SιH₄) in the presence of molecular oxygen (0₂) and an argon (Ar) sputter (so that the Sι0₂ is directlv deposited and then sputtered by the Ar), without the benefit of automated HDP workpiece temperature control, such as that provided by various illustrative embodiments of the present invention, can lead to irregularities, as shown in Figure 6, and to inadequate step coverage of spaces 610 having high aspect ratios r For example, sidewall coverage is non-uniform, with protuberances 620 and scant or no coverage on portions of the sides 61 OS and on the bottom 610B, with voids between the sides 61 OS

By way of contrast, Figure 7 schematically illustrates an HDP etch/deposition of an Sι0₂ film on high aspect ratio metal lines, in accordance with various embodiments of the present invention As shown in Figure 7, a structure layer 700 of a workpiece (such as a semiconducting wafer) has two conducting lines 705 (such as metal lines) formed thereon For example, the conducting lines 705 may be parallel bit and/or digit lines on a DRAM chip The conducting lines 705 may have a thickness t_metaι in a range of approximately 3000 A to 7000 A, for example An opening and/or space 710 may be formed between the conducting lines 705, the space 710 having a width w in a range of approximately 2000 A to 4000 A, for example The space 710 may hav e sides 710S and a bottom 710B The aspect ratio r of the space 710 may be give by the thickness t_raetal divided by the width w

The conducting lines 705 may be electrically isolated from each other by a dielectric film 715, formed of silicon dioxide (Sι0₂), for example The dielectric film 715 may have a thickness t_fiι_m in a range of approximately 5000 A to 10000 A, for example As shown in Figure 7, silane (SιH₄) used in the presence of molecular oxygen (0₂) and an argon (Ar) sputter (so that the Sι0₂ is directly deposited and then sputtered by the Ar), having the benefit of automated HDP workpiece temperature control, such as that provided by various illustrative embodiments of the present invention, as shown in Figures 1-5 and as described above, can lead to substantial uniformities, and to adequate step coverage of spaces 710 having high aspect ratios r For example, overall coverage is substantially uniform covering sides 71 OS and bottom 71 OB, with no protuberances such as the protuberances 620 in Figure 6, and w ith no voids between the sides 71 OS

In various illustrative embodiments the engineer may be provided with advanced process data monitoring capabilities, such as the ability to provide historical parametric data in a user-friendly format as well as event logging, real-time graphical display of both current processing parameters and the processmg parameters of the entire run, and remote, / e , local site and worldwide, monitoring These capabilities may engender more optimal control of critical processmg parameters, such as throughput accuracy, stability and repeatability, processing temperatures, mechanical tool parameters, and the like This more optimal control of critical processing parameters reduces this variability This reduction in variability manifests itself as fewer within-run disparities, fewer run-to-run disparities and fewer tool-to-tool disparities This reduction in the number of these disparities that can propagate means fewer deviations in product quality and performance In such an illustrative embodiment of a method of manufacturing according to the present invention, a monitoring and diagnostics system may be provided that monitors this variability and optimizes control of critical parameters

Any of the above-disclosed embodiments of a method of manufacturing according to the present invention enables the monitor and control the thickness of a dielectric film deposited on the workpiece during an HDP etch and/or deposition process, the refractive index of the dielectric film, and thickness variations across the workpiece of the dielectric film, and enables the making of supervisory processing adjustments, either manually and/or automatically, to improve and/or better control the yield Additionally, any of the above-disclosed embodiments of a method of manufacturing according to the present invention enables semiconductor device fabrication with increased device accuracy and precision, increased efficiency and increased device yield, enabling a streamlined and simplified process flow, thereby decreasing the complexity and lowering the costs of the manufacturing process and increasing throughput

In particular, any of the above-disclosed embodiments of a method of manufacturing according to the present invention enables reduced variations in the parameters for films deposited and/or etched by HDP processes, resulting, in part, from better control of the temperature of the workpiece during HDP processes Even though an HDP process may use an electromagnetic chuck to clamp, hold, and cool the workpiece during the HDP processing, and even though the workpiece temperature T may be a complicated function of the argon (Ar) sputtering parameters, the workpiece contact with the chuck, and the cooling of the chuck by water flow cooling and or by helium (He) flow cooling, any of the above-disclosed embodiments of a method of manufacturing according to the present invention enables automatic control based on modeling of parameters characteπstic of HDP processing, thereby improving over the conventional approach that uses manual control based on a process engineer's experience and potentially fallible judgment

The particular embodiments disclosed above are illustrative only, as the invention may be modified and practiced in different but equivalent manners apparent to those skilled m the art having the benefit of the teachings herein Furthermore, no limitations are intended to the details of construction or design herein shown, other than as described in the claims below It is therefore evident that the particular embodiments disclosed above may be altered or modified and all such variations are considered within the scope and spirit of the invention In particular, every range of values (of the form, "from about a to about b,ⁿ or, equivalently, "from approximately a to b," or, equivalently, "from approximately a-b") disclosed herein is to be understood as referring to the power set (the set of all subsets) of the respective range of values, in the sense of Georg Cantor Accordingly, the protection sought herein is as set forth in the claims below

Claims

1 A method of manufacturing, the method comprising processing a workpiece (305) in a high-density plasma (HDP) processing step (310) measuring a parameter (317) characteristic of the HDP processing performed on the workpiece (305) in the HDP processing step (310), modeling the characteristic parameter measured (330, 335), and applying the model to modify at least one HDP control input parameter (340, 360, 320)

2 The method of claim 1 , wherein measuring the characteristic parameter includes measuring the workpiece (305) temperature

3 The method of claim 1 , wherein modeling the characteristic parameter measured (330, 335) includes extracting an appropriate model from a store (335, 360) of potential models

4 The method of claim 1 , wherein modeling the characteristic parameter measured (330, 335) includes correlating previously measured characteristic parameters with respective HDP recipe control input parameters and building the model using the correlations (340, 360)

5 The method of claim 1 , wherein applying the model to modify at least one HDP control input parameter (340, 360, 320) includes adjusting at least one of a first HDP recipe control input parameter (315) for water flow cooling, a second HDP recipe control input parameter (315) for helium flow cooling, a third HDP recipe control input parameter (315) for argon sputtering, and a fourth HDP recipe control input parameter (315) for electrostatic chuck clamping voltage

6 A method of manufacturing, the method comprising processing a workpiece (305) m a high-density plasma (HDP) processing step (310), measuring a parameter (317) characteristic of the HDP processing performed on the workpiece (305) m the HDP processing step (310), modeling the characteristic parameter measured (330, 335), applying the model to modify at least one HDP control input parameter (340, 360, 320), monitoring (317) at least one effect of modifying the at least one HDP control input parameter (340, 360, 320), and updating the model (330, 335) based on the at least one effect monitored (317)

7 The method of claim 6, wherein modelmg the characteristic parameter measured (330, 335) includes extracting an appropriate model from a store (335, 360) of potential models

8 The method of claim 6, wherein modeling the characteristic parameter measured (330, 335) includes correlating previously measured characteπstic parameters with respective HDP recipe control input parameters and building the model using the correlations (340, 360) 9 The method of claim 6, wherein applying the model to modifv at least one HDP control input parameter (340 360, 320) includes adjusting at least one of a first HDP recipe control input parameter (315) for water flow cooling, a second HDP recipe control input parameter (315) for helium flow cooling a third HDP recipe control input parameter (315) for argon sputtering, and a fourth HDP recipe control input parameter (315) for electrostatic chuck clamping voltage

10 A computer-readable, program storage device (235, 240, 245), encoded with instructions that, when executed by a computer (230), perform a method for manufacturing a workpiece (305), the method comprising processing a workpiece (205) in a high-densitv plasma (HDP) processing step (210), measuring a parameter (317) characteristic of the HDP processing performed on the workpiece (205) in the HDP processing step (210), modeling the characteristic parameter measured (230, 235), and applying the model to modify at least one HDP control input parameter (230. 220, 215)