U.S. patent application number 13/457885 was filed with the patent office on 2012-08-23 for method for using real-time apc information for an enhanced lot sampling engine.
This patent application is currently assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION. Invention is credited to Gary W. BEHM, Yue LI, Malek BEN SALEM.
Application Number | 20120215490 13/457885 |
Document ID | / |
Family ID | 41401022 |
Filed Date | 2012-08-23 |
United States Patent
Application |
20120215490 |
Kind Code |
A1 |
BEHM; Gary W. ; et
al. |
August 23, 2012 |
METHOD FOR USING REAL-TIME APC INFORMATION FOR AN ENHANCED LOT
SAMPLING ENGINE
Abstract
A method includes passing a lot through a production process and
evaluating a statistical quality of the production process.
Additionally, the method includes calculating an advanced process
control (APC) recipe parameter adjustment (RPA) distribution value
and determining if sampling is indicated. Furthermore, the method
includes, if sampling is indicated, performing a measurement
process of the lot.
Inventors: |
BEHM; Gary W.; (Hopewell
Junction, NY) ; SALEM; Malek BEN; (Wappingers Falls,
NY) ; LI; Yue; (Hopewell Junction, NY) |
Assignee: |
INTERNATIONAL BUSINESS MACHINES
CORPORATION
Armonk
NY
|
Family ID: |
41401022 |
Appl. No.: |
13/457885 |
Filed: |
April 27, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12135514 |
Jun 9, 2008 |
|
|
|
13457885 |
|
|
|
|
Current U.S.
Class: |
702/182 |
Current CPC
Class: |
Y02P 90/20 20151101;
G05B 2219/31458 20130101; G05B 2219/32294 20130101; G05B 19/41865
20130101; Y02P 90/22 20151101; G05B 2219/32199 20130101; Y02P 90/02
20151101 |
Class at
Publication: |
702/182 |
International
Class: |
G06F 15/00 20060101
G06F015/00 |
Claims
1. A computer program product comprising a computer readable
storage medium having readable program code tangibly embodied in
the storage medium, the computer program product includes at least
one component operable to: pass a lot through a production process;
calculate an advanced process control recipe parameter adjustment
(RPA) distribution value using an advanced process controller
(APC); calculate an APC calculated process capability (Cpk) value
using the APC; communicate the RPA and the Cpk value to a lot
sampling engine (LSE); utilize the LSE to correlate the RPA and the
Cpk value in conjunction with a historical Cpk value calculated by
the LSE to determine if sampling of the lot is indicated; and
perform a measurement process of the lot, if sampling is
indicated.
2. The computer program product of claim 1, wherein the at least
one component is further operable to: determine an APC RPA sampling
rate based on the APC RPA distribution value; determine an APC
calculated Cpk sampling rate based on the APC calculated Cpk value;
and determine a historical Cpk sampling rate based on the
historical Cpk value.
3. The computer program product of claim 2, wherein the at least
one component is further operable to: determine a weighted APC RPA
sampling rate, a weighted APC calculated Cpk sampling rate and a
weighted historical Cpk sampling rate; and determine a lot sampling
rate by summing the weighted APC RPA sampling rate, the weighted
APC calculated Cpk sampling rate and the weighted historical Cpk
sampling rate.
4. The computer program product of claim 2, wherein at least one of
the determining the APC RPA sampling rate, the determining the APC
calculated Cpk sampling rate, and the determining the historical
Cpk sampling rate comprises using configurable intervals.
5. The computer program product of claim 2, wherein at least one of
the determining the APC RPA sampling rate, the determining the APC
calculated Cpk sampling rate, and the determining the historical
Cpk sampling rate comprises using an S-curve.
6. The computer program product of claim 2, further comprising
determining a lot sampling rate based on the APC RPA sampling rate,
the APC calculated Cpk sampling rate and the historical Cpk
sampling rate.
7. The computer program product of claim 2, further comprising
determining a weighted APC RPA sampling rate, a weighted APC
calculated Cpk sampling rate and a weighted historical Cpk sampling
rate.
8. The computer program product of claim 7, further comprising
determining a lot sampling rate based on the weighted APC RPA
sampling rate, the weighted APC calculated Cpk sampling rate and
the weighted historical Cpk sampling rate.
9. The computer program product of claim 8, wherein the determining
the lot sampling rate comprises summing the weighted APC RPA
sampling rate, the weighted APC calculated Cpk sampling rate and
the weighted historical Cpk sampling rate.
10. The computer program product of claim 1, wherein the
calculating the APC RPA distribution value is determined according
to: Distribution (D)=Min((UCL-Mean)/Sigma,(Mean-LCL)/Sigma) wherein
UCL is an upper control limit, LCL is a lower control limit, and
Sigma is a standard deviation based on approximately ten lots.
11. The computer program product of claim 1, wherein the
determining if sampling of the lot is indicated is based on at
least one of: a lot sampling rate; the lot being identified as a
rework lot or a send ahead (SAHD) lot; the lot being measured in a
pre-step; and manual rules.
Description
CROSS REFERENCE TO RELATED APPLICATION
[0001] This application is a continuation application of copending
U.S. patent application Ser. No. 12/135,514 filed on Jun. 9, 2008,
the contents of which are incorporated herein by reference in their
entirety.
FIELD OF THE INVENTION
[0002] The present invention generally relates to a method and
system of semiconductor fabrication, and more specifically, to a
method and system of semiconductor fabrication using real-time
Advanced Process Control (APC) information for an enhanced Lot
Sampling Engine (ELSE).
BACKGROUND
[0003] Currently in the 300 mm semiconductor fabricator, there is
no relationship between the fab-wide Advanced Process Control (APC)
system and the Lot Sampling Engine (LSE). As a separate system, the
APC system relies on pre-measurement, post-measurement and
operational data to calculate the recipe parameter required for the
process tool to perform the wafer process on target. Ideal for
optimizing the recipe parameter adjustment (RPA) is to measure
every wafer. However, there is a tradeoff with the cost of
producing the wafers and wasting cycle time.
[0004] Before utilization of the LSE, sampling was fixed by route,
e.g., determined by a lot attribute assigned at a beginning or
early stage of a manufacturing process, and not linked to
manufacturing process capability and/or performance. The attribute
assured that a certain percentage of work-in-progress (WIP) was
measured at various inspection points before and/or after each
operation of a production process. Unfortunately, the attribute did
not account for performance of the operation. Thus, when the
process was performing poorly, not enough lots were being measured,
and when the process was performing well, too many lots were being
measured, thus wasting cycle time.
[0005] Thus, the LSE was developed, which provides a sampling plan
to optimize the throughput of the process. From manufacturing's
perspective, measurement is an overhead and it has no value if it
is not really necessary to be measured. When the process is
performing well, the manufacturer may be measuring too many lots
and wasting cycle time. However, with the LSE, the sampling rate is
linked to the process capability. With this smart sampling method,
the cycle time is no longer wasted on processes that are performing
well. Instead, the focus is on those processes that need more
measurements for process improvement.
[0006] However, there is a problem when the LSE decides to bypass
the measurement (e.g., for throughput benefit), in that the APC
system is not able to calculate the optimized value for the recipe
parameter adjustment (RPA) due to limited number of measurements
(yield degradation). The LSE uses a process capability index
(Cp/Cpk metrics) to adjust sampling rates and reduce the Mean Time
To Detect (MTTD). Determining process capability involves measuring
a variability of a process and comparing the measured variability
with a proposed specification or product tolerance. However, the
Cp/Cpk metrics are calculated over a 28 day period and are only
updated once a week. So, the LSE does not react to the tool/process
issues quickly enough.
[0007] Another challenge is created by "Send Ahead" (SAHD)
operations. In normal manufacturing operations, there is no need to
use, for example, SAHD wafers. However, when tolerance variation in
a particular process is unacceptable, SAHD is required in order to
prevent scrap. SAHD is also necessary for low volume parts to speed
yield learning. In either case, no SAHD lot should be skipped by a
sampling plan. However, when relying on an attribute to determine
sampling, many SAHD lots are not sampled. Without real time
integrated product and process information, it is difficult, if not
impossible, to establish a sampling rate that can account for a lot
attribute, process performance, and SAHD lots in a manufacturing
process. When relying solely on a lot attribute, or manual
sampling, a lot may be sampled too often, or not often enough.
Sampling, i.e., measurement, is a non-value added operation and
actually slows production. Thus, over sampling can be costly.
However, if too few lots are sampled, defective lots can pass
through production. In this case, final testing costs are increased
and a company's reputation for quality may be at risk.
[0008] However, there is no known system that correlates the
benefits of both the APC system and the LSE system, by providing
real-time LSE information to the APC system, such that both systems
are integrated and optimized for the lot sampling plan without
affecting the APC operations.
[0009] Accordingly, there exists a need in the art to overcome the
deficiencies and limitations described hereinabove.
BRIEF SUMMARY
[0010] In a first aspect of the invention, a method comprises
passing a lot through a production process and evaluating a
statistical quality of the production process. Additionally, the
method comprises calculating an advanced process control (APC)
recipe parameter adjustment (RPA) distribution value and
determining if sampling is indicated. Furthermore, if sampling is
indicated, the method comprises performing a measurement process of
the lot.
[0011] In an additional aspect of the invention, a computer program
product comprises a computer usable medium having readable program
code embodied in the medium. The computer program product includes
at least one component operable to pass a lot through a production
process and evaluate a statistical quality of the production
process. Additionally, the at least one component is operable to
calculate an advanced process control (APC) recipe parameter
adjustment (RPA) distribution value and determine if sampling is
indicated. Furthermore, the at least one component is operable to
perform a measurement process of the lot, if sampling is
indicated.
[0012] In a further aspect of the invention, a system comprises a
wafer processing station and an advanced process control (APC).
Additionally, the system comprises an enhanced lot sampling engine
(ELSE) in communication with the APC and configured to receive APC
recipe parameter adjustment (RPA) information and APC calculated
process capability (Cpk) information from the APC. Further, the
ELSE is configured to determine a sampling rate for a lot based on
historical Cpk information, the APC RPA information from the APC
and the APC calculated Cpk information from the APC.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
[0013] The present invention is described in the detailed
description which follows, in reference to the noted plurality of
drawings by way of non-limiting examples of exemplary embodiments
of the present invention.
[0014] FIG. 1 shows an overall system for use in semiconductor
wafer manufacturing;
[0015] FIG. 2 shows an overall system for use in semiconductor
wafer manufacturing with an enhanced lot sampling engine (ELSE) in
accordance with the invention;
[0016] FIG. 3 shows a timing diagram of a method for advanced
process control (APC) with an ELSE in accordance with the
invention;
[0017] FIG. 4A shows an exemplary plot of lot sampling rate versus
process capability obtained without using a lot sampling
engine;
[0018] FIG. 4B shows an exemplary plot of lot sampling rate versus
process capability obtained using a lot sampling engine;
[0019] FIG. 5 shows an exemplary table of ELSE rules in accordance
with the invention;
[0020] FIG. 6A shows an exemplary table used for determining lot
sampling rates in accordance with the invention;
[0021] FIG. 6B shows an exemplary plot of LSE sampling rate versus
Cpk values with an S-curve in accordance with aspects of the
invention; and
[0022] FIG. 7 shows an exemplary flow for performing aspects of the
invention.
DETAILED DESCRIPTION
[0023] The present invention generally relates to a method and
system of semiconductor fabrication, and more specifically, to a
method and system of semiconductor fabrication using real-time
Advanced Process Control (APC) information for an enhanced Lot
Sampling Engine (ELSE). The present invention provides a method and
system to use APC information by the enhanced LSE system to
determine the optimized sample plan required for processing the
lots through the route. By implementing the present invention,
manufacturing with higher throughput and higher yield may be
obtained, by providing a smart sampling mechanism which enforces
lot measurement and measurement data collection where needed by the
APC. Additionally, higher throughput and higher yield may be
obtained by optimizing the throughput/cycle time by minimizing
sampling, while maintaining a desirable level of APC control.
Furthermore, Mean Time To Detect (MTTD) may be reduced by 50% by
feeding additional information to the ELSE and reaching to
currently available information sooner. Additionally, the yield
learning curve may be improved on low volume products and/or routes
through measurement on SAHD lots.
[0024] FIG. 1 shows an overall system 100 used in semiconductor
fabrication including an APC system 130 and an LSE system 135. As
shown in FIG. 1, a measurement stage 105 is performed in an
external metrology tool 120 on a lot of wafers contained in, e.g.,
a front-open unified pod (FOUP) 122. More specifically, an APC
trigger point signal 140 indicates that the lot of wafers are to be
measured and measurement data is to be collected. The external
metrology tool 120 performs a measuring process, and via an
equipment interface (EI) 125, the measurements are sent to the APC
130. As shown in FIG. 1, the measurement data 145 is indicated as
"pre-measurement" data, as this measuring stage is performed prior
to a process stage. In other words, the measurement data 145 is a
pre-process measurement. At the end of the measurement stage 105,
an LSE trigger point signal 148 is sent to the LSE 135 to indicate
that the measuring stage 105 is complete.
[0025] A manufacturing execution system (MES) 150 is responsible
for coordinating and controlling the movements of the lots of
wafers, e.g., the FOUPs 122 and for the automatic collection of
data, e.g., measurement data and process data. Thus, after
collection of the measurements taken at the measurement stage 105,
the MES 150 forwards the lot of wafers to a process stage 110,
where the wafers are further processed in a process tool 158 in
conjunction with the APC 130. For example, the APC 130 may make
recipe parameter adjustments (RPAs) based on, e.g., feed forward
controls. An APC trigger point signal 155 may be sent to the APC
130 to indicate a calculation of the RPAs. Again, the communication
between the process tool 158 and both the MES 150 and the APC 130
may be facilitated using an equipment interface (EI) 125.
[0026] At the end of the process step 110, an LSE trigger point
signal 160 is sent to the LSE 135 to indicate that the process
stage 110 is complete and request a determination by the LSE 135 as
to whether the next measurement stage, e.g., 115, should be skipped
or not. For example, the LSE 135 may determine, based on, e.g.,
rules, historical process capability and/or lot specific rules, to
perform the next metrology stage 115, or to skip the next metrology
stage 115 and proceed to the next process stage (not shown).
[0027] Thus, if the LSE trigger point signal 160 indicates that the
next measurement stage 115 should be skipped, the lot of wafers 122
is not sent to the next measurement stage 115, and instead the lot
of wafers is sent to a next processing stage (not shown). However,
if the response to LSE trigger point signal 160 indicates that the
next measurement stage 115 should be performed, the lot of wafers
is forwarded to the next measurement stage 115. At the next
measurement stage 115, another external metrology tool 165 performs
a measurement process of the lot of wafers and measurement data 170
is sent from the external metrology tool 165 to the APC 130 via
another EI 125. At the end of the measurement stage 115, an LSE
trigger point signal 180 is sent to the LSE 135 to indicate that
the measurement stage 115 is complete.
[0028] In this way, a lot of wafers may proceed through a
manufacturing process comprising, for example, a series of
measurement steps and a series of process steps. However, as shown
in FIG. 1, there is no direct exchange of data between the APC and
the LSE. Therefore, the LSE is not provided with real-time data
from the APC.
[0029] FIG. 2 shows an overall integrated circuit manufacturing
system 200 which includes an enhanced lot sampling engine (ELSE)
225 in direct communication with an APC 220 in accordance with an
aspect of the invention. According to an aspect of the invention,
the APC 220 may send the ELSE 225 real-time information, e.g.,
feedback of process performance, in order for the ELSE 225 to make
a determination to proceed to the next measurement stage 215 or
skip the next measurement stage 215. As shown in FIG. 2, the APC
220 may send the ELSE 225 additional sampling rate information.
More specifically, the APC 220 may send to the ELSE 225 APC rate of
change data (or recipe parameter adjustment (RPA) data) and APC
calculated Cpk data. According to an aspect of the invention, the
ELSE 225 may use this additional data from the APC in conjunction
with the historical Cpk data (e.g., determined over 28 days and
updated weekly) to make a more useful decision to proceed to the
next measurement stage or skip the next measurement stage, as
explained further below.
[0030] In embodiments, the APC recipe parameter adjustment (RPA)
distribution data may be a statistical distribution of the recipe
parameter adjustments with a standard deviation based on, e.g., 10
lots. Additionally, in embodiments, the APC calculated measurement
Cpk may be a statistical distribution of the process capability
based on, e.g., 7 days of minimal of 8 lots per day, or, in
embodiments, may depend on the volume of the lots that go through
the APC 220. Further, the historical Cpk is a statistical
distribution of the process capability based on, e.g., 28 days and
updated on weekly basis.
[0031] FIG. 3 shows an exemplary timing diagram of communications
between the different elements of the present invention. As shown
in FIG. 3, the MES 230 sends an APC runtime capability request to
the APC 220. In response, the APC 220 sends an APC runtime
capability response to the MES 230. Next, the MES 230 sends a
recipe parameter request to the APC 220. In response, the APC 220
sends a recipe parameter response to the MES 230, the EI 125 and
equipment 190, e.g., the processing tool 158. Next, the EI 125
sends control job information (create) to the MES 230 and the APC
220. Additionally, the equipment 190 sends event data (e.g.,
control job ID and/or process parameters) to the APC 220. The EI
125 sends a control job information (executing) signal, and
subsequently, a control job information (completed) signal to the
MES 230 and the APC 220.
[0032] In response to the control job information (completed)
signal, the MES 230 sends a database (DB) trigger on a lot
operation complete to the ELSE 225. In response, the ELSE 225 sends
a lot information query from a materials manager database, which is
a real-time database. Additionally, the ELSE 225 may query the APC
220 for information, e.g., the APC RPA distribution information and
the APC calculated Cpk. In response, the APC 220 sends information
to the ELSE 225, e.g., the APC RPA distribution information and the
APC calculated Cpk. Next, the MES 230 sends a tool state change
information to the ELSE 225. In response and based on the APC RPA
distribution data, the APC calculated Cpk, and a historical Cpk,
the ELSE 230 issues a gatepass transaction signal indicating either
moving to a subsequent measurement stage or skipping the subsequent
measuring stage.
[0033] FIG. 4A shows an exemplary plot 400 of sample rate versus
process capability (and out of control (OOC) percentage) obtained
without using an LSE. As shown in FIG. 4A, a low process capability
(Cpk) value indicates a process is performing poorly. Conversely, a
high Cpk value indicates a process is performing well. This is also
shown by the OOC % scale, which indicates that a low Cpk value
corresponds to a high OOC %, and vice versa. In embodiments, OOC %
may assume a centered process and no vintages. As shown in FIG. 4A,
the lot sampling is occurring too infrequently when it is needed,
i.e., when the process is performing poorly and sampling too
frequently when sampling is not needed, i.e., when the process is
performing well. As discussed above, an aim of the present
invention is to reduce the sample rate when the process is
performing well (to prevent wasted time and resources) and to
increase the sampling rate when the process is performing
poorly.
[0034] FIG. 4B shows an exemplary plot 450 of sample rate versus
process capability achievable by implementing aspects of the
present invention. As shown in FIG. 4B, when the process is
performing well (high Cpk), the sample rate is lower and when the
process is performing poorly (low Cpk), the sample rate is
higher.
[0035] FIG. 5 shows a table 500 containing a hierarchy of rules the
ELSE 225 may use to determine whether to perform a measurement step
or to skip the measurement step and proceed to the next processing
step. As shown in FIG. 5, the rules are applied differently for
different scenarios. FIG. 5 includes a scenario column 505, an
action column 510, a data source column 515, a sampling rate column
520 and a weighting column 525.
[0036] As shown in FIG. 5, with a first scenario relating to a
rework lot, a measurement stage is always performed subsequent to a
processing stage, and thus has a sampling rate of 100%. For
example, a rework lot may involve a rework of a lot due to wafer
under polishing. The data source for a rework scenario is the
materials manager (MM) of the MES 230. As there is a 100% sampling
rate with this scenario, and there is only a single data source, as
indicated in the weighting column 525, a weighting is not
applicable for this scenario.
[0037] With a second scenario relating to a SAHD lot, a measurement
stage is always performed subsequent to a processing stage, and
thus has a sampling rate of 100%. The data source for a SAHD lot
scenario is the APC 220. As there is a 100% sampling rate with this
scenario, and there is only a single data source, as indicated in
the weighting column 525, a weighting is not applicable for this
scenario.
[0038] With a third scenario relating to a lot that was measured in
a pre-step, the lot will be measured in a post-step. For example,
if a thickness was measured in a pre-processing stage and then the
thickness is altered in a processing stage, in order to determine
the change in thickness, the thickness should be measured in a
post-processing stage. Thus, as indicated in the sampling rate
column 520, for this scenario the sampling rate is 100%. Further,
as indicated in the weighting column 525, as there is a single data
source for this scenario the weighting is not applicable.
[0039] With a fourth scenario relating to manual rules, a sampling
rate may be configurable regardless of determined Cpk values. That
is, with this scenario, manual rules may override any Cpk rules.
For example, for a given process PDID (process definition
identification), a measurement PDID, a logical recipe or a process
tool, a sampling rate percentage may be defined regardless of
determined Cpk values. A sampling rate may be defined using manual
rules, for example, based on business decisions. If time is of the
essence, for example, sampling rates can be reduced to accommodate
delivery of lots. In embodiments, as indicated in the data source
column 515, the data source for this scenario is the ELSE 225.
Further, as indicated in the weighting column 525, as there is a
single data source for this scenario, a weighting is not
applicable.
[0040] With a fifth scenario, a sampling rate may be determined
based on a weighted average of the sampling rates determined from
the APC RPA distribution value, the APC measurement Cpk value and
the historical data Cpk value. More specifically, according to an
aspect of the invention, the ELSE 225 may determine a suggested
sampling rate based on an RPA sampling rate, an APC measurement Cpk
sampling rate and an historical data Cpk sampling rate.
[0041] FIG. 6A shows an exemplary table 600 for determining the APC
RPA sampling rate, the APC measurement Cpk sampling rate and the
historical data Cpk sampling rate based on the APC RPA distribution
value, the APC measurement Cpk value and the historical data Cpk
value, respectively, using configurable intervals according to an
aspect of the invention. As shown in FIG. 6A, the table 600
includes a data type column 605, a Distribution (D)/Cpk (X) column
610 and a sampling rate column 615. While FIG. 6A indicates
particular sampling rates for particular ranges of distribution
values and Cpk values, it should be understood that the invention
contemplates that other sampling rates and other ranges of
distribution values and Cpk values may be used. In other words, the
intervals set forth in FIG. 6A are configurable intervals.
[0042] The APC RPA sampling rate may be determined based on the RPA
distribution value, which, in embodiments, is determined according
to equation (1).
Distribution (D)=Min((UCL-Mean)/Sigma,(Mean-LCL)/Sigma) (1)
wherein UCL is the upper control limit, LCL is the lower control
limit, and Sigma is the standard deviation. The distribution value
varies due to changes in the mean value (due to, e.g., changes
based on the wafer recipe) and changes of the standard deviation.
The UCL and the LCL define measurement limits, that if are
exceeded, indicate a defect. As can be observed in FIG. 6A, with a
lower distribution value (D), a higher sampling rate is indicated,
and with a higher distribution value (D), a lower sampling rate is
indicated. Thus, for example, with a distribution (D) value of 1.4,
an APC RPA sampling rate would be 50%.
[0043] Additionally, as shown in FIG. 6A, the APC measurement Cpk
sampling rate may be determined from the APC measurement Cpk value
and the historical data Cpk sampling rate may be determined from
the historical data Cpk value. In light of FIG. 6A, referring to
FIG. 5, the data source for the APC calculated Cpk is the APC 220
and the data source for the historical data Cpk is a historical
data repository, e.g., a data management information warehouse
(DMIW). As can be observed in FIG. 6A, with a lower Cpk value (X),
a higher sampling rate is indicated, and with a higher Cpk value
(X), a lower sampling rate is indicated. Thus, for example, with an
APC measurement Cpk value (X) of 1.65, an APC measurement Cpk
sampling rate would be 16.66%. Additionally, for example, with a
historical data Cpk value (X) of 1.9, a historical data Cpk
sampling rate would be 12.5%.
[0044] With an understanding of FIG. 6A, referring again to FIG. 5,
according to an aspect of the invention, as shown in the weighting
column 525, with the fifth scenario, in embodiments, relative
weights may be assigned to the RPA sampling rate, the APC
measurement Cpk sampling rate and the historical data Cpk sampling
rate. Thus, with the example shown in FIG. 5, the APC RPA sampling
rate may be assigned a weighting of 20%, the APC measurement Cpk
sampling rate may be assigned a weighting of 40% and the historical
data Cpk sampling rate may be assigned a weighting of 40%. As can
be observed the individual weightings sum to a total of 100%. As
should be understood, the relative weightings shown in FIG. 5 are
exemplary and configurable, and the invention contemplates that
other relative weightings may be used.
[0045] According to a further aspect of the invention, the ELSE 225
may determine a suggested sampling rate by summing the weighted
values of the RPA sampling rate, the APC measurement Cpk sampling
rate and the historical data Cpk sampling rate according to
equation (2).
Suggested Sampling Rate=(Weighted RPA Sampling Rate)+(Weighted APC
Measurement Cpk Sampling Rate)+(Weighted Historical Data Cpk
Sampling Rate) (2)
Thus, for example, using the values discussed in the example above
with equation (2) and with the exemplary relative weightings
indicated in FIG. 5, a suggested sampling
rate=(0.2)(0.5)+(0.4)(0.1666)+(0.4)(0.125)=0.21664=21.664%. In
contrast, using only the historical data Cpk to determine a
sampling rate would result in a sampling rate of 12.5%.
[0046] With further embodiments, the APC RPA sampling rate, the APC
calculated Cpk sampling rate and the historical Cpk sampling rate
may be based on the APC RPA distribution value, the APC measurement
Cpk value and the historical data Cpk value, respectively, using an
S curve instead of the configurable intervals. With configurable
intervals, two process performances may be substantially the same,
and yet be assigned different sampling rates. For example, using
the configurable intervals, as shown in FIG. 6A, a Cpk value of
1.24999 would indicate a sampling rate of 100%, while a Cpk value
of 1.25 would indicate a sampling rate of 50%. Thus, according to
an aspect of the invention, an S-curve may be created, for example,
using a Sigmoid function, to provide more granularity to the Cpk
value-sampling rate correlation.
[0047] FIG. 6B shows an exemplary plot 650 of LSE sampling rate
versus Cpk values. As shown in FIG. 6B, the exemplary plot 650
includes configurable interval line 655 and S-curve 660. As should
be understood, the configurable interval line 655 correlates with
the APC measurement Cpk values of FIG. 6A. That is, for example, a
Cpk value of less than 1.25 indicates a sampling rate of 100% and a
Cpk value greater than or equal to 1.25 and less than 1.5 indicates
a sampling rate of 50%, etc. According to an aspect of the
invention, the S-curve 660 may be generated to "fit" the
configurable interval line 655, such that more granularity is
provided in determining a sampling rate. In embodiments, the
S-curve 660 may be generated, for example, using the following
equation (3):
Y=12.5+87.5*(1-1/(1+exp[-10.5*(X-1.35)]) (3)
wherein X is the Cpk value and Y is the sampling rate. However, it
should be understood that different intervals may necessitate a
different equation to "fit" the S-curve to the different intervals.
Thus, it should be understood that equation (3) is an exemplary
equation, and other equations are contemplated by the invention in
order to fit an S-curve to the configured intervals.
[0048] According to this aspect of the invention, a method may
include creating an equation, e.g., a Sigmoid function, that
creates an S-curve to "fit" the configurable interval line 655. In
embodiments, this may include, for example, experimentally
determining the equation, e.g., through trial and error.
Additionally, different confidence levels for each different
section of the curve may be created. Further, a plot of where the
Cpk is on the curve may be generated using the APC calculated Cpk
and/or the historical Cpk. Then, using the plot 650 including the
S-curve 660, a sampling plan may be optimized to improve cycle
time.
Flow Diagram
[0049] FIG. 7 shows an exemplary flow 700 describing how the
real-time APC information is collected, calculated and passed to
the ELSE for an optimized lot sampling plan in accordance with the
present invention. FIG. 7 may equally represent a high-level block
diagram of components of the invention implementing the steps
thereof. The steps of FIG. 7 may be implemented on computer program
code in combination with the appropriate hardware. This computer
program code may be stored on storage media such as a diskette,
hard disk, CD-ROM, DVD-ROM or tape, as well as a memory storage
device or collection of memory storage devices such as read-only
memory (ROM) or random access memory (RAM). Additionally, the
computer program code can be transferred to a workstation over the
Internet or some other type of network. Furthermore, a computer
program product may include a computer usable medium having
readable program code tangibly embodied in the medium.
[0050] At step 705, a lot is reserved at a measurement tool for a
pre-processing measurement. At step 710, the measuring of the wafer
lot is commenced. At step 715, the APC calculates an Cpk based on
the measurement data obtained in step 710. Additionally, at step
715, the APC calculated Cpk is sent to the ELSE. At step 720, the
measurement tool completes the measuring of the wafer lot. At step
725, the wafer lot is reserved at a process tool. At step 730, the
APC calculates a recipe parameter adjustment (RPA) based, e.g., on
feed-forward (FF) and feed-backward (FB) information. Additionally,
at step 730, the APC sends the calculated RPA to the ELSE. At step
735, the wafer lot is processed by the processing tool. At step
740, the processing tool completes the processing of the wafer lot.
At step 745, the lot is reserved at a measurement tool for a
post-processing measurement. At step 750, the wafer lot
post-processing measurement is commenced. At step 755, the APC
calculates a Cpk based on the measurement data determined at step
750. Additionally, at step 755, the APC sends the calculated Cpk
data (which now includes a further measurement) to the ELSE. At
step 760, the measurement tool is finished with the wafer lot.
[0051] It should be understood, that while the steps have been
described as occurring in a particular order, the invention
contemplates that the steps may be performed in other orders.
Further, as can be observed with the exemplary flow of FIG. 7, the
subsequent measurement step (post-processing measurement) is
performed and is not skipped. However, as described above, it
should be understood that the subsequent measuring stage may be
skipped based on the optimized lot sampling plan determined by the
ELSE.
[0052] The terminology used herein is for the purpose of describing
particular embodiments only and is not intended to be limiting of
the invention. As used herein, the singular forms "a", "an" and
"the" are intended to include the plural forms as well, unless the
context clearly indicates otherwise. It will be further understood
that the terms "comprises" and/or "comprising," when used in this
specification, specify the presence of stated features, integers,
steps, operations, elements, and/or components, but do not preclude
the presence or addition of one or more other features, integers,
steps, operations, elements, components, and/or groups thereof.
[0053] The corresponding structures, materials, acts, and
equivalents of all means or step plus function elements in the
claims, if applicable, are intended to include any structure,
material, or act for performing the function in combination with
other claimed elements as specifically claimed. The description of
the present invention has been presented for purposes of
illustration and description, but is not intended to be exhaustive
or limited to the invention in the form disclosed. Many
modifications and variations will be apparent to those of ordinary
skill in the art without departing from the scope and spirit of the
invention. The embodiment was chosen and described in order to best
explain the principles of the invention and the practical
application, and to enable others of ordinary skill in the art to
understand the invention for various embodiments with various
modifications as are suited to the particular use contemplated.
Accordingly, while the invention has been described in terms of
embodiments, those of skill in the art will recognize that the
invention can be practiced with modifications and in the spirit and
scope of the appended claims.
* * * * *