This invention relates to the field of manufacturing processes. More particularly, this invention relates to controlling tool settings during manufacturing processes.
It is common for tools used in manufacturing processes to have adjustable settings which affect parameters of articles produced by the manufacturing processes. It is further common for a manufacturing process to use a set of one or more tools to perform a particular fabrication step, each tool in the set being capable of performing the fabrication step, to provide a desired level of capacity for the manufacturing process. Moreover, it is common for more than one model of article to be produced by such a manufacturing process in which instances of each model are processed through more than one tool in the set, and in which each tool in the set processes instances of more than one model. Adjusting the settings of the tools to produce the articles with parameters close to desired values may be problematic when models have different desired parameter values and/or performances of the tools drift.
The instant invention provides a manufacturing process which uses a controller method of generating offsets for tool parameter settings for multiple devices so as to reduce differences between measured device parameter values and corresponding parameter targets. The tool parameter settings generated by the controller method include tool offsets and device offsets. After device parameters are measured, the controller method updates values of the appropriate tool offset and device offset, such that only one of the tool offset and device offset is significantly changed.
The controller method includes three separate modes of operation: generating tool parameter settings for a characterized tool which is processing a characterized device, generating tool parameter settings for an uncharacterized tool, and generating tool parameter settings for a characterized tool which is processing an uncharacterized lot.
The present invention is described with reference to the attached figures, wherein like reference numerals are used throughout the figures to designate similar or equivalent elements. The figures are not drawn to scale and they are provided merely to illustrate the invention. Several aspects of the invention are described below with reference to example applications for illustration. It should be understood that numerous specific details, relationships, and methods are set forth to provide a full understanding of the invention. One skilled in the relevant art, however, will readily recognize that the invention can be practiced without one or more of the specific details or with other methods. In other instances, well-known structures or operations are not shown in detail to avoid obscuring the invention. The present invention is not limited by the illustrated ordering of acts or events, as some acts may occur in different orders and/or concurrently with other acts or events. Furthermore, not all illustrated acts or events are required to implement a methodology in accordance with the present invention.
For the purposes of this specification, the term “tool” will refer to items of manufacturing equipment which perform a fabrication step on a workpiece. For example, the term “tool” may refer to a photolithographic wafer stepper which exposes a pattern in a layer of photoresist on a semiconductor wafer, the workpiece being the semiconductor wafer including the photoresist layer.
The instant invention provides a controller method of generating offsets for tool parameter settings for one or more tools performing a fabrication step in a manufacturing process which fabricates multiple instances, referred to herein as lots, of possibly more than one model, referred to herein as a device, of a manufactured article so as to reduce differences between measured device parameter values and corresponding desired device parameter values, referred to herein as parameter target values. Differences between the measured device parameter values and the corresponding desired device parameter values are referred to herein as error values. In this context, a lot may include more than one physical article of manufacture.
The tool parameter settings take into account tool offsets which are specific to each tool and device offsets which are specific to each device. After device parameters of a lot are measured, the controller method computes updated values of the appropriate tool offset and device offset. Tool weights and devices weights are applied to the updating process so that only one of the tool offset and device offset is significantly changed. Manufacturing process flows in which device parameters are measured on a subset of the processed lots, and in which device parameters are measured on lots in a different sequence than a sequence of performing the fabrication step, are accommodated by the inventive method.
Devices which have had a prescribed number of lots fabricated using the inventive method are referred to herein as characterized devices. Uncharacterized devices, for example new devices that are being introduced into the manufacturing process, have not met this criterion. Similarly, characterized tools have fabricated a prescribed number of lots; uncharacterized tools, for example tools newly installed in the manufacturing process or refurbished tools, have not processed the prescribed number of lots. The controller method includes three separate modes of operation: generating tool parameter settings for a characterized tool which is processing a characterized device, generating tool parameter settings for an uncharacterized tool, and generating tool parameter settings for a characterized tool which is processing an uncharacterized lot.
During operation on a characterized tool and a characterized device, the controller method updates the tool offset while maintaining cumulative sums of error values for the tool and the device. If the tool cumulative error sum exceeds a prescribed limit, action such as repair or maintenance of the tool may be performed. If the device cumulative error sum exceeds a prescribed limit, the appropriate device offset is adjusted and updated.
During operation on an uncharacterized tool, the controller method updates the tool offset, but does not update the cumulative error sums. During operation on an uncharacterized device, the controller method updates the device offset, but not the cumulative error sums. The device offset may be updated by making relatively large corrections on initial lots followed by smaller corrections on subsequent lots.
After processing through the tools (1004), each lot (1002) may have a device parameter measured by a measurement tool (1008) or may bypass measurement, as denoted by the bypass operation (1010). The lots (1002) may be measured in a second sequence (1012) which is different from the first sequence (1006).
Next is step (2006) which is a decision of whether or not to measure Lot N. If the decision is yes, step (2008) is executed, which is to measure a value of a device parameter of Lot N. If step (2008) is executed, step (2010) is subsequently executed, which is to run the controller method. The step of running the controller method includes updating the current value of the Tool K offset and the current value of the Device J offset. The details of step (2010) will be discussed in reference to
After step (2010) is executed, or if the result of step (2006) is no so that steps (2008) and (2010) are bypassed, step (2012) is to process a second lot. The second lot is either a Device J lot or is process on Tool K, or both, so that at least one of the updated Device J offset and the updated Tool K offset is used to generate an updated tool setting to process the second lot. After step (2102) is executed, the fabrication process is ended (2014).
Next, step (3020) is to compute a value of a setting for Tool K. In the instant embodiment, the Device J parameter is taken to be a linear function of the Tool K setting, offset by the Tool K offset and the Device J offset. A slope of the Device J parameter with respect to the Tool K setting is a Tool K process factor. The Tool K setting is computed, for example using EQN. 1.
Subsequent to computing the Tool K setting, step (3022) is to process Lot N in Tool K set at the Tool K setting value computed in step (3020). Step (3024) is to store the value of the Device J offset used to process Lot N in a storage medium (3026). Step (3028) is to store the value of the Tool K offset used to process Lot N in a storage medium (3030). In an alternate embodiment, steps (3022), (3024) and (3028) may be performed in another order. After completion of the above steps, processing Lot N of Device J in Tool K is ended (3032).
Subsequently, step (4010) determines if the Tool K cumulative error sum is above a prescribed limit. If a result of step (4010) is yes, step (4012) is executed, which is to take an appropriate action for Tool K. In one embodiment, Tool K may be cleaned, repaired, refurbished or otherwise modified. In another embodiment, the Tool K offset and/or Tool K process factor may be modified.
Following execution of step (4012), or step (4010) if the result of step (4010) is no, step (4014) determines if the Device J cumulative error sum is above a prescribed limit. If a result of step (4014) is yes, step (4016) is executed, which is to update the Device J offset. Details involved in step (4016) are discussed in reference to
If a result of step (5004) is yes, step (5006) is executed. A first sub-step of step (5006) is to assign the tool weight a value for an uncharacterized tool, referred to as an untuned tool weight. In one embodiment, the untuned tool weight may be a fixed value, for example 0.25. In another embodiment, the untuned tool weight may be less than a previous value of the untuned tool weight applied to Tool K. In a further embodiment, the untuned tool weight may be dependent on the measured parameter value of Lot N.
A second sub-step of step (5006) is to assign a non-interference value to the device weight, so that parameter error values resulting from Tool K being uncharacterized do not interfere with subsequent adjustments to the Device J offset. In one embodiment, the device weight may be assigned a number with an absolute value less than a prescribed fraction, for example one percent, of the tool weight value assigned in the first sub-step of step (5006). In another embodiment, the device weight may be assigned a value of zero. In a further embodiment, the device weight may be assigned a low fixed value, for example 0.00015.
A third sub-step of step (5006) is to assign the device cumulative error sum factor a low incrementing value, for example 0.001 or zero. A fourth sub-step of step (5006) is to assign the lot cumulative error sum factor a low incrementing value, for example 0.001 or zero. One purpose of assigning low incrementing values to the cumulative error sum factors is to prevent error values resulting from Tool K being uncharacterized leading to undesired increases in a Device J cumulative error sum and/or a Tool K cumulative error sum.
If the result of step (5004) is no, step (5008) is executed, which is to determine if Device J is designated as uncharacterized. In one embodiment, Device J may be designated as uncharacterized if a prescribed number of lots of Device J, for example ten lots, have not been run and subsequently measured, since Device J began its uncharacterized status. In another embodiment, Device J may be designated as uncharacterized if measured parameters of lots of Device J are not within a prescribed range. In a further embodiment, Device J may be designated as uncharacterized based on other criteria.
If a result of step (5008) is yes, step (5010) is executed. A first sub-step of step (5006) is to assign the device weight a value for an uncharacterized device, referred to as an untuned device weight. In one embodiment, the untuned device weight may be a fixed value, for example 0.20. In another embodiment, the untuned tool weight may be dependent on the measured parameter value of Lot N. In a further embodiment, the untuned device weight may be less than a previous value of the untuned device weight applied to Device J. In a particular embodiment, the untuned device weight may be computed using EQN 2.
Where: initial weight is a fixed value, for example 0.75
The values for the initial weight and a in EQN. 2 may be selected such that all the untuned device weight values assigned while Device J is designated as uncharacterized sum to a prescribed number, for example a value between 2.75 and 3.5, which provides a desired rate of tuning the device J offset.
A second sub-step of step (5010) is to assign a non-interference value to the tool weight, so that parameter error values resulting from Device J being uncharacterized do not interfere with subsequent adjustments to the Tool K offset. In one embodiment, the tool weight may be assigned a number with an absolute value less than a prescribed fraction, for example one percent, of the device weight value assigned in the first sub-step of step (5010). In another embodiment, the tool weight may be assigned a value of zero. In a further embodiment, the tool weight may be assigned a low fixed value, for example 0.00015.
A third sub-step of step (5010) is to assign the device cumulative error sum factor a low incrementing value, for example 0.001 or zero. A fourth sub-step of step (5010) is to assign the lot cumulative error sum factor a low incrementing value, for example 0.001 or zero. One purpose of assigning low incrementing values to the cumulative error sum factors is to prevent error values resulting from Device J being uncharacterized leading to undesired increases in a Device J cumulative error sum and/or a Tool K cumulative error sum.
If the result of step (5008) is no, step (5012) is executed. A first sub-step of step (5012) is to assign the tool weight a value for a characterized tool, referred to as a tuned tool weight. In one embodiment, the tuned tool weight may be a fixed value, for example 0.15. In another embodiment, the tuned tool weight may be less than a previous value of the tuned tool weight applied to Tool K. In a further embodiment, the tuned tool weight may be dependent on the measured parameter value of Lot N.
A second sub-step of step (5012) is to assign a non-interference value to the device weight, so that the Device J offset is not significantly altered with each measured lot. In one embodiment, the device weight may be assigned a number with an absolute value less than a prescribed fraction, for example one percent, of the tool weight value assigned in the first sub-step of step (5012). In another embodiment, the device weight may be assigned a value of zero. In a further embodiment, the device weight may be assigned a low fixed value, for example 0.00015.
A third sub-step of step (5012) is to assign the device cumulative error sum factor a value which causes the Device J cumulative error sum to be updated with information from the Lot N measurement. In one embodiment, the device cumulative error sum factor may be assigned a value of 1. In another embodiment, the device cumulative error sum factor may be assigned a value dependent on the measured parameter value of Lot N.
A fourth sub-step of step (5006) is to assign the tool cumulative error sum factor a value which causes the Tool K cumulative error sum to be updated with information from the Lot N measurement. In one embodiment, the tool cumulative error sum factor may be assigned a value of 1. In another embodiment, the tool cumulative error sum factor may be assigned a value dependent on the measured parameter value of Lot N.
After step (5006), step (5010) or step (5012) is completed, the process of generating a device weight, a tool weight, a device cumulative error sum factor and a tool cumulative error sum factor is ended (5014).
Subsequently, step (6024) is executed, which is to compute a new value for the Tool K offset, using a formula which contains a relationship as expressed in EQN 3.
New ToolK offset=Current Tool K offset−(tool weight×(Lot N tool offset−Current Tool K offset+Lot N measured parameter value−Device J parameter target)) EQN. 3
Computing a new value for the Tool K offset as described in reference to step (6024) may advantageously account for parameter measurements of lots which are performed in a different sequence than processing of the lots.
Next, step (6026) is to store the new Tool K offset value as the current Tool K offset in a storage medium (6028).
Step (6030) is to compute a new value for the Device J offset, using a formula which contains a relationship as expressed in EQN 4.
New Device J offset=Current Device J offset+(device weight×(Lot N device offset−Current Device J offset+Lot N measured parameter value−Device J parameter target)) EQN. 4
Computing a new value for the Device J offset as described in reference to step (6030) may advantageously account for parameter measurements of lots which are performed in a different sequence than processing of the lots.
Next, step (6032) is to store the new Device J offset value as the current Device J offset in a storage medium (6034).
In an alternate embodiment, step (6024) and step (6030) may be performed in another order. After the new value for the Tool K offset and the new value for the Device J offset are computed and stored, the process for updating the Tool K offset and the Device J offset is ended (6036).
Subsequently, step (7016) is executed, which is to compute a new Tool K cumulative error sum, using a formula which contains a relationship as expressed in EQN. 5.
New Tool K cumulative error sum=Current Tool K cumulative error sum+(tool cumulative error sum factor×(Lot N measured parameter value−Device J parameter target)) EQN. 5
Computing the Tool K cumulative error sum as described in reference to step (7016) may advantageously prevent errors due to untuned device offsets from affecting the Tool K offset.
Next, step (7018) is to store the new Tool K cumulative error sum as the current Tool K cumulative error sum in a storage medium (7020).
Step (7022) to compute a new Device J cumulative error sum, using a formula which contains a relationship as expressed in EQN. 6.
Computing the Device J cumulative error sum as described in reference to step (7022) may advantageously prevent errors due to untuned tool offsets from affecting the Device J offset.
Next, step (7024) is to store the new Device J cumulative error sum as the current Device J cumulative error sum in a storage medium (7026).
Step (7028) to compute a new Device J cumulative run sum, using a formula which contains a relationship as expressed in EQN. 7.
Next, step (7030) is to store the new Device J cumulative run sum as the current Device J cumulative run sum in a storage medium (7032).
In an alternate embodiment, steps (7018), (7024) and (7030) may be performed in another order. After the new value for the Tool K cumulative error sum, the new value for the Device J cumulative error sum and the new value for the Device J cumulative run sum are computed and stored, the process for updating the Tool K cumulative error sum and the Device J cumulative error sum is ended (7034).
Subsequently, step (8016) is executed, which is to compute a new value of the Device J offset, for example using EQN. 8.
Next, step (8018) is to store the new Device J offset as the current Device J offset in a storage medium (8020). After the new value for the Device J offset is computed and stored, the process for updating the Device J offset is ended (8022).