U.S. patent application number 10/768559 was filed with the patent office on 2004-08-26 for method and apparatus for using dfss to manage a research project.
Invention is credited to Cawse, James Norman.
Application Number | 20040168135 10/768559 |
Document ID | / |
Family ID | 23529420 |
Filed Date | 2004-08-26 |
United States Patent
Application |
20040168135 |
Kind Code |
A1 |
Cawse, James Norman |
August 26, 2004 |
Method and apparatus for using DFSS to manage a research
project
Abstract
In an exemplary embodiment, an application of combinatorial
materials development with minimum materials development, minimum
variance, and maximum integration is provided. The embodiment is
directed to a method of project development of a combinatorial
materials development process using DFSS techniques having four
major elements. The first element is the use of a design for six
sigma (DFSS) process mapping to convert a complex and disorganized
process structure to an organized structure that can be further
analyzed. The second element comprises the use of quality function
deployment houses as a method of flowing critical to quality
characteristics (CTQ) through a research project. The third element
comprises a transfer function that connects the overall steps of
the project to the output which is measured as variability not as
mean. Score cards are used as the "function" to total the
variabilities of each process step. The final element comprises an
extension of design of experiment (DOE) techniques.
Inventors: |
Cawse, James Norman;
(Pittsfield, MA) |
Correspondence
Address: |
Philip D. Freedman
Philip D. Freedman PC
P.O. Box 19076
Alexandria
VA
22320
US
|
Family ID: |
23529420 |
Appl. No.: |
10/768559 |
Filed: |
February 2, 2004 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10768559 |
Feb 2, 2004 |
|
|
|
09387332 |
Aug 31, 1999 |
|
|
|
6725183 |
|
|
|
|
Current U.S.
Class: |
716/132 |
Current CPC
Class: |
B01J 2219/00747
20130101; B01J 2219/00689 20130101; G06F 2111/08 20200101; G06Q
10/06 20130101; B01J 2219/00745 20130101; G06F 30/00 20200101; C40B
40/18 20130101; B01J 2219/00695 20130101; B01J 2219/007 20130101;
C40B 30/08 20130101 |
Class at
Publication: |
716/002 |
International
Class: |
G06F 017/50 |
Claims
What is claimed is:
1. A method for using design for six sigma techniques for managing
a research project, said method comprising: using at least one six
sigma technique to organize the research project; using quality
function deployment houses to flow critical to quality
characteristics through said research project; measuring a
variability of each portion of said research project; and using
design of experiment techniques with combinatorial designs to
minimize said variability and maximize research results.
2. The method of claim 1, wherein said houses comprise a first
house for developing and measuring critical to quality
characteristics of the said research project.
3. The method of claim 2, wherein said houses comprise a second
house wherein said research project is analyzed.
4. The method of claim 1, wherein said research project comprises a
combinatorial material development project.
5. The method of claim 1, wherein said project comprises a search
for catalysts.
6. The method of claim 5, wherein said catalysts are for the
production of diphenyl carbonate.
7. The method of claim 1, wherein said at least one six sigma
technique include a decision on a system to study, a prework
process, a setup, react and evaluate process, and a data analysis
process.
8. The method of claim 1, wherein said portions of said research
project include sample preparation, chemical reaction, analytical
preparation, chemical analysis, data analysis and database
system.
9. The method of claim 1, wherein said design of experiment
techniques include methods to reduce variability components of each
portion of said research project.
10. A computer usable medium having computer readable program code
means embodied therein for using design for six sigma techniques
for managing a research project, said medium comprising: computer
readable program code means for causing a computer to use six sigma
techniques for converting a complex, disorganized process structure
to an organized structure that can be further analyzed; computer
readable program code means for causing a computer to use quality
function deployment houses to flow critical to quality
characteristics through said research project; computer readable
program code means for causing a computer to measure a variability
of each portion of said research project; and computer readable
program code means for causing a computer to use design of
experiment techniques with combinatorial designs to minimize said
variability and maximize research results.
11. The medium of claim 10, wherein said houses comprise a first
house develops and measures critical to quality characteristics of
the said research project.
12. The medium of claim 11, wherein said houses comprise a second
house wherein said research project is analyzed.
13. The medium of claim 10, wherein said research project comprises
a combinatorial material development project.
14. The medium of claim 13, wherein said project comprises a search
for catalysts.
15. The medium of claim 14, wherein said catalysts are for the
production of diphenyl carbonate.
16. The medium of claim 10, wherein said six sigma techniques
include a decision on a system to study, a prework process, a
setup, react and evaluate process; and a data analysis process.
17. The medium of claim 10, wherein said portions of said research
project include sample preparation, chemical reaction, analytical
preparation, chemical analysis, data analysis and a database
system.
18. The medium of claim 10, wherein said design of experiment
techniques include methods to reduce variability components of each
portion of said research project.
19. A program storage device readable by a machine, tangibly
embodying a program of instructions executable by said machine to
perform a method for using design for six sigma techniques for
managing a research project, said method comprising: using six
sigma techniques for converting a complex, disorganized process
structure to an organized structure that can be further analyzed;
using quality function deployment houses to flow critical to
quality characteristics through said research project; measuring a
variability of each portion of said research project; and using
design of experiment techniques with combinatorial designs to
minimize said variability and maximize research results.
20. The device of claim 19, wherein said houses comprise a first
house for developing and measuring critical to quality
characteristics of the said research project.
21. The device of claim 19, wherein said houses comprise a second
house wherein said research project is analyzed.
22. The device of claim 19, wherein said research project comprises
a combinatorial material development project.
23. The device of claim 19, wherein said project comprises a search
for catalysts.
24. The device of claim 23, wherein said catalysts are for the
production of diphenyl carbonate.
25. The device of claim 19, wherein said six sigma techniques
include a decision on a system to study, a prework process, a
setup, react and evaluate process, a data analysis process.
26. The device of claim 19, wherein said portions of said research
project include sample preparation, chemical reaction, analytical
preparation, chemical analysis, data analysis and a database
system.
27. The device of claim 19, wherein said design of experiment
techniques include methods to reduce variability components of each
portion of said research project.
28. A computer program product, said computer program product
comprising: a computer usable medium having computer readable
program embodied in said program for using design for six sigma
techniques for managing a research project, said computer program
including: computer readable program means for using six sigma
techniques for converting a complex, disorganized process structure
to an organized structure that can be further analyzed; computer
readable program means for using quality function deployment houses
to flow critical to quality characteristics through said research
project; computer readable program means for measuring a
variability of each portion of said research project; and computer
readable program means for using design of experiment techniques
with combinatorial designs to minimize said variability and
maximize research results.
29. The computer program of claim 28, wherein said houses comprise
a first house for developing and measuring critical to quality
characteristics of the said research project.
30. The computer program of claim 29, wherein said houses comprise
a second house wherein said research project is analyzed.
31. The computer program of claim 28, wherein said research project
comprises a combinatorial material development project.
32. The computer program of claim 28, wherein said project
comprises a search for catalysts.
33. The computer program of claim 32, wherein said catalysts are
for the production of diphenyl carbonate.
34. The computer program of claim 28, wherein said six sigma
techniques include a decision on a system to study, a prework
process, a setup, react and evaluate process and a data analysis
process.
35. The computer program of claim 28, wherein said portions of said
research project include sample preparation, chemical reaction,
analytical preparation, chemical analysis, data analysis and a
database system.
36. The computer program of claim 28, wherein said design of
experiment techniques include methods to reduce variability
components of each portion of said research project.
37. The computer product of claim 28, wherein said product includes
an internet or intranet interconnection for access to six sigma
flowdown programs and to provide data on the various variabilities.
Description
BACKGROUND OF THE INVENTION
[0001] The invention relates to a novel application of a
combinatorial materials development with minimum variance and
maximum integration. In particular, the invention is a system and
method of project development for a combinatorial materials
development process using DFSS techniques.
[0002] As illustrated in FIG. 1 Combinatorial materials development
(CMD) is an experimental approach to rapidly identify or optimize
new material compositions or processes. CMD uses a parallel
approach to generate thousands of target materials. The targets are
evaluated quickly and reliably using automated analytical systems.
The final step is to use statistical data analysis and
visualization to identify promising leads.
[0003] FIG. 2 illustrates the transition from traditional chemical
research to Combinatorial technology. From the 1890s to the 1990s
chemists as individuals might perform one or two experiments per
day with experimental sizes limited to 1 to 1000 grams per
experiment. These 100 to 500 experiments per year might lead to 1
or 2 new leads per year.
[0004] Combinatorial technology 10 typically comprises several
steps including: Experimental Planning 12; Sample Preparation 14;
Chemical Reaction 16; Analytical Preparation 18; Chemical Analysis
20; and Data Analysis 22. These steps will be discussed further in
relation to the use of six sigma techniques. In the 1990s,
development of Combinatorial Technology permitted a team approach
using experimental sizes 1 to 100 milligrams per experiment with 10
to 200 or more experiments per day. Depending on the chemistry
involved, the 1000 to 10,000 or more experiments per year are
likely to generate 10 or more new leads per year. The Combinatorial
Technology approach can be used for discoveries of new materials
when there are many possible components and small changes in
components cause big changes in material properties. The CMD
process may not be as effective for minimizing material problems
where components are few and well known. The combinatorial approach
was developed to overcome competitive threats, address the need for
speed, reduced cost, and broad patent coverage, and to deal with
increasing system complexity and expectations. The advantages of
the CMD approach are high-speed innovation with the possibility of
broad patent protection. The hardware and software that make CMD
possible are now available.
[0005] For any process (business, manufacturing, service, research,
etc.), the sigma value is a metric that indicates how well that
process is performing. The higher the sigma value, the better the
output. Sigma measures the capability of the process to perform
defect-free-work, where a defect is synonymous with customer
dissatisfaction. With six sigma, the common measurement index is
defects-per-unit where a unit can be virtually anything. Examples
include a component, a piece part of a jet engine, and an
administrative procedure. The sigma value indicates how often
defects are likely to occur. As sigma increases, customer
satisfaction goes up along with improvement of other metrics (e.g.,
cost and cycle time).
[0006] The six sigma methodology has been used by a number of
companies such as Motorola Semiconductors, Texas Instruments,
Allied Signal and Digital Corporation. All of these companies use
this process for a specific application such as semiconductor
manufacturing in the case of Motorola and Texas Instruments.
General Electric Company, the assignee of this application, has
used six sigma technology in a wide number of areas.
[0007] FIG. 3 is a flowchart of a design for six sigma (DFSS)
process in new product development. The overall DFSS process of
FIG. 3 is divided into four sub-processes labeled Identify, Design,
Optimize and Validate. Each sub-process includes sub-steps. The
Identify sub-process includes sub-steps 102 and 104. The Design
sub-process includes sub-steps 106-112. The Optimize sub-process
includes sub-steps 114-126. The Validate sub-process includes
sub-steps 128-134. The DFSS process shown in FIG. 3 is useful for
improving the process of designing a product or procedure. The
invention can also be applied to other six sigma processes such as
the Measure, Analyze, Improve and Control (MAIC) process used for
improving processes (such as manufacturing processes or business
processes).
[0008] The six sigma process includes a method for identifying
critical to quality (CTQ) dependencies in quality function
deployment. Quality function deployment (QFD) is a methodology for
documenting and breaking down customer requirements into manageable
and actionable details. The concept of "houses of quality" has been
used to represent the decomposition of higher level requirements
such as critical to quality characteristics or CTQ's (also referred
to as Y's) into lower level characteristics such as key control
parameters or KCP's (also referred to as X's). FIG. 4 depicts a
conventional house of quality hierarchy in which high level
requirements such as customer requirements are decomposed into
lower level characteristics such as key manufacturing processes and
key process variables within the manufacturing processes.
[0009] Each house of quality has previously corresponded to a stage
or level of the process of designing a product. At the highest
level, represented as house of quality #1 152, customer
requirements are associated with functional characteristics of a
product. At the next level of the design process, represented as
house of quality #2 154, the functional characteristics of the new
product are associated with new product characteristics. At the
next level of the design process, represented as house of quality
#3 206, the part characteristics are associated with manufacturing
processes. At the next level of the design process, represented as
house of quality #4 208, the manufacturing processes are associated
with manufacturing process variables.
[0010] Conventionally, new chemical entities with useful properties
are generated by identifying a chemical compound (called a "lead
compound") with some desirable property or activity, creating
variants of the lead compound, and evaluating the property and
activity of those variant compounds. Examples of chemical entities
with useful properties include paints, finishes, plasticizers,
surfactants, scents, flavorings, and bioactive compounds, but can
also include chemical compounds with any other useful property that
depends upon chemical structure, composition, or physical state.
Chemical entities with desirable biological activities include
drugs, herbicides, pesticides, veterinary products, and the
like.
[0011] One deficiency in traditional chemical research pertains to
the first step of the conventional approach, i.e., the
identification of lead entities. As stated by Claudia M. Caruana,
"Combinatorial Chemistry Promises Better Catalysts and Materials",
Chem. Eng. Prog., October 1998, p 11-14, "Typically, catalyst
discovery involves inefficient trial-and-error, because catalytic
activity is difficult to screen." Consequently, a fundamental
limitation of the conventional approach is the ability to generate
and analyze large numbers of catalyst candidates. The generation of
such candidates is very labor intensive and time consuming. For
example, it takes many chemist years to produce and evaluate even a
small subset of the variants in a single catalyst system. Caruana,
in the article referenced above, states that "conventional
discovery strategies usually are based on the time-consuming
"one-sample-at-a-time" approach, which can take months or years to
determine suitable candidate materials."
[0012] Recently, attention has been focused on the use of
combinatorial chemical methods to assist in the generation of new
materials development leads. "Combinatorial chemistry uses a
parallel approach to discover thousands of target materials and
then produces "libraries" of these substances quickly." (Ref.:
Caruana) A combinatorial materials library is a collection of
diverse materials generated by either chemical synthesis or by a
combination of formulation and process steps, combining chemical
"building blocks" such as reagents. For example, a combinatorial
catalyst library is formed by combining precursor solutions to
generate an array of formulations, subjecting it to appropriate
processing conditions to produce (possibly) active catalysts, and
evaluating the activity of each formulation. Millions of potential
catalysts or other materials theoretically can be generated through
such combinatorial mixing of chemical building blocks and multiple
process steps. For example, Peter Schultz in "Generating New
Molecular Function: a Lesson from Nature", Angew. Chem. Int. Ed,
1999, 38, 36-54, has observed that "Given approximately 60 elements
in the periodic table that can be used to make compositions
consisting of three, four, five, or even six elements, the universe
of possible new compounds remains largely uncharted."
[0013] However there is a need for a system and method for
efficiently and effectively generating new leads designed for
specific utilities. Combinatorial technology has the ability to
develop many leads, but the variability in the results develops
many inefficiencies. There is a need to reduce the variability of
the results in combinatorial technology processes.
SUMMARY OF THE INVENTION
[0014] Acordingly, the invention relates to a system for
implementing a combinatorial chemistry research project using
Design for Six Sigma (DFSS) techniques. In an exemplary embodiment,
a novel application of a combinatorial materials development with
minimum variance and maximum integration is provided. In
particular, the present invention provides a system and method of
project development of a combinatorial materials development
process, such as catalyst development, using DFSS techniques.
[0015] The process has four major elements. The first element is
the use of a Design for Six Sigma (DFSS) process mapping to convert
a complex and disorganized process structure to an organized
structure that can be further analyzed.
[0016] The second element comprises the use of quality function
deployment (QFD) houses as a method of flowing critical to quality
characteristics (CTQ's) through a research project. After the
customer CTQ's and the measures are developed conventionally in
House 1, a novel usage of QFD is done by making the entire project
the "product" which is analyzed by House 2. Individual process
elements of the project are analyzed for CTQ's, and those CTQ's
become the "how's" in the QFD House 2. Doing this allows effective
prioritization of all the Measure, Analyze, Improve and Control
(MAIC) projects applied to each process element.
[0017] The third element comprises a transfer function that
connects the overall steps of the project to the output which is
measured as variability not as mean. Score cards are used as the
"function" to total the variabilities of each process step.
[0018] The fourth and final element comprises an extension of
design of experiment (DOE) techniques. Conventional DOE's,
discussed above, are ineffective for combinatorial chemistry
because of the large size and complexity of the experiments.
Additionally, in the context of these particular experiments for
catalyst development, generally only synergistic effects of
co-catalyst combinations are meaningful. This requires novel DOE
approaches such as full combinatorial designs.
[0019] The combined use of combinatorial techniques and six sigma
techniques as disclosed herein reduces the variabilities in the
results, thereby producing a better result from the development
program. These and other features and advantages of the present
invention will be apparent from the following brief description of
the drawings, detailed description, and appended claims.
BRIEF DESCRIPTION OF THE DRAWINGS
[0020] The invention will be further described in connection with
the accompanying drawings in which:
[0021] FIG. 1 illustrates a combinatorial materials development
program;
[0022] FIG. 2 illustrates the development of combinatorial
technology;
[0023] FIG. 3 illustrates a six sigma flow stet;
[0024] FIG. 4 illustrates the house of quality concept;
[0025] FIG. 5 illustrates the combination of combinatorial
technology and DFSS;
[0026] FIG. 6 illustrates a catalyst goal in a combinatorial
process;
[0027] FIG. 7 illustrates the combinatorial discovery cycle;
[0028] FIG. 8 illustrates the DFSS flowdown in a combinatorial DPC
program;
[0029] FIG. 9 illustrates the factors in Combinatorial QFD: House
1;
[0030] FIG. 10 illustrates the factors in House 2;
[0031] FIG. 11 illustrates a combinatorial transfer function;
[0032] FIG. 12 illustrates the use of a scorecard to predict an
initial process capability;
[0033] FIG. 13 illustrates a first specific MAIC project;
[0034] FIG. 14 illustrates a second specific MAIC project;
[0035] FIG. 15 illustrates the use of highly nested DOE's to find
little x's;
[0036] FIG. 16 illustrates the results of the DFSS analysis;
[0037] FIG. 17 illustrates using DOE techniques; and
[0038] FIG. 18 illustrates the number of catalysts produced using
the combinatorial technology with a six sigma analysis.
DETAILED DESCRIPTION
[0039] An exemplary embodiment provides a system for project
development of a combinatorial materials development process using
DFSS techniques. As illustrated in FIG. 5, using DFSS techniques
with CMD involves four major steps for reaching Quality by Design
200: Identification 202; Design 204; Optimization 206; and
Validation 208.
[0040] In the Identification 202 phase, the process identifies
customer critical to quality characteristics (CTQ) for needed
materials 204 and performs a CTQ flowdown 206. The identification
phase additionally analyzes the measurement system capability 208
and generates and validates system and sub-system modules 210. As a
result, the identification stage builds scorecard predictions based
on parts, process and performance factors 212.
[0041] In Design stage 204 there is a roll-up of Scorecards for all
sub-systems, including a capability flow-up 214. Design stage
204also identifies the gaps, including a low defects per unit (DPU)
on a scorecard 220.
[0042] In Optimization stage 206, there is use of Design of
Experiment (DOE) on prototypes to find the critical few X's 222.
Optimization stage 206uses an Analysis Spreadsheet for parameter
and tolerance design 224 and generates purchase and manufacturing
specifications 226.
[0043] Validation stage 208 confirms that pilot builds have matched
predictions 228. Validation stage 208 attempts to mistake proof the
process 230. Additionally, Validation stage 208 refines the models,
scorecards and product characterization database 232. The
Validation stage documents the effort and the results 234.
[0044] The number of patents issued to pharmaceutical companies has
grown significantly since the late 1980s when the combinatorial
approach was embraced by the pharmaceutical industry. In addition
to pharmaceutical technology, the combinatorial approach has
significant application in: plastics, including catalysts, carbon
fibrils and blends; lightning, including fluorescent lamp cathodes,
phosphors and LED; medical systems including scintillators and
superconductors; and aircraft engines and turbines, including
coatings and alloys. The assignee has experimentally testing
combinatorial technology using a DFSS approach to develop a highly
productive process for synthesis and screening of chemicals.
Additionally, the combined approach has been used to identify
longer term opportunities for use of combinatorial methodology for
new materials discovery.
[0045] As shown in FIG. 6, an example of a research project is a
Combinatorial Process Pilot System for Diphenyl Carbonate (DPC)
reaction. The goal is to search for a new catalyst that improves
the output of DPC. Diphenyl carbonate is an ingredient used in the
manufacture of polycarbonate. Two hundred fifty million pounds per
year of DPC are needed in each new polycarbonate plant.
Polycarbonate is currently made with a 2 step technology. With a
new catalyst, a one step process is possible resulting in
significant capital cost and variable cost reduction.
[0046] As seen in FIG. 7, a combinatorial discovery cycle 250 is a
complex disorganized process structure with multiple interconnected
elements 252-276. The exemplary embodiment converts this complex,
disorganized process structure into an organized structure that can
be further analyzed. Even though the efficiency of a combinatorial
process is much greater than traditional techniques discussed
above, the variation possible in each step of the combinatorial
process make the combinatorial discovery cycle a complex process
that requires structure and discipline.
[0047] FIG. 8 illustrates use of a DFSS flowdown in a Combinatorial
DPC program and is exemplary of the "Formulate Concept Design" 106
and "For Each CTQ, Identify Design Parameters portions" 110 in FIG.
3. The main categories have equivalencies to the DFSS structure
described in FIG. 3. In a DFSS or MAIC analysis most sub-systems
have their own set of CTQ flowdowns.
[0048] The first major portion is a "Decide on System to Study"
phase 302 that is equivalent to the Identify stage of DFSS and
Measure stage of MAIC. Two of the elements of the flowdown in this
phase are "Plan and prioritize the Chemical `Universe`" 310 and
"Plan Combinatorial Strategy for Exploration" 312.
[0049] A second major portion of a DFSS flowdown in a Combinatorial
DPC program is called "Prework Process" 304 (similar to the Design
stage in FIG. 3) and contains at least four elements including
"Purchase Chemicals" 314, "Combinatorial Synthesis" 316, "Perform
safety, stock & compatibility tests" 318, and "Define Stock
Solutions & Vial Mixtures" 320.
[0050] A third major portion of a DFSS flowdown in a Combinatorial
DPC program is called "Setup, React, & Evaluate" 306 (similar
to the Optimize stage in FIG. 3) and contains at least three
elements including "Vial Setup" 322, Reactor 324, and "Sample
Preparation & Analysis" 326.
[0051] A fourth major portion of the DFSS flowdown in a
Combinatorial DPC program is called "Data Analysis" 308 (equivalent
to the Validate Stage in FIG. 3) and contains at least three
elements including "System Database" 328, "Data Analysis" 330, and
"Inferential Engine" 332.
[0052] The CTQ's for the Combinatorial DPC Program come under a
general goal of having a highly productive catalyst for DPC with a
generalizable process. FIG. 9 illustrates a Combinatorial QFD:
House 1 350. The Customer Requirements are the What's of the House
1. The Design requirements are the How's of House 1 and are also
listed in FIG. 9. An importance factor is assigned to each Customer
requirement and to each Design requirement listed in House 1. The
importance factors are normalized to a 1-5 in importance. As shown
in FIG. 9, the specific CTQ's come in two general categories,
Optimize Catalyst 352 and Innovate Process 354. The "Optimize
Catalyst" CTQ's include: (1) High throughput system: reactions per
week; (2) Scalability from micro reactions to macro reactions; (3)
Combinatorial synthesis of more than 100 co-catalysts; (4) Catalyst
descriptors for prediction; and Identify at least one lead per
month for the DPC chemical team.
[0053] The Management category of CTQ's includes: (1) the results
placing manufacturer of DPC in a strong position; (2) the results
permit the organization freedom to practice combinatorial
methodology with out licensing of other technology; and (3) an
ability to translate the combinatorial expertise to other
technologies. Other research projects may have other CTQ's related
to the specific project.
[0054] FIG. 9 illustrates the use of quality function deployment
(QFD) houses discussed above as a method of flowing the CTQ's
listed above through a research project such as the DPC Catalyst
Project. After the CTQ's and measures are developed conventionally
in House 1 as shown in FIG. 9, a novel usage of QFD is making the
entire project the "product" which is analyzed by House 2 as
illustrated in FIG. 10.
[0055] Individual process elements of the project are analyzed for
CTQ's and those CTQ's become the "hows" in QFD 2.sup.nd House.
Doing this allows an effective prioritization of all the projects
applied to each process element. The Design requirements of House 1
become the What's in House 2 as shown in FIG. 10. The CTQ's of the
various subsystems become the "How's of the 2.sup.nd House. The
Hows of House 2 include "Decide on Chemical system to Study" 302,
"Prework Process" 304, Setup, React and Evaluate 306, and Data
Analysis 308. The main subsystems of a Combinatorial process Vial
(Sample) Preparation 402, Chemical reaction 404, Analytical
Preparation 406, Chemical Analysis 408, Data Analysis and Database
System 412 are each analyzed and MAIC and DPC priorities flow from
the importance score. The more important a combined what and how
are to the overall process, the more attention it receives under a
MAIC analysis. A DPC Project has a Vial Preparation 402 subsystem.
In other Combinatorial projects this subsystem is more generally
called Sample Preparation.
[0056] The transfer function connects the overall steps of the
project to the output specification which is measured as
variability not the mean. FIG. 11 illustrates how the combinatorial
transfer function is developed. The variability of each stage, vial
(sample) preparation 402, chemical reaction 404, analytical
preparation 406, chemical analysis 408, data analysis 410 and
database system 412 are shown. The little y's (CTQ's) such as
Accurate Lead Preparation, Rapid Lead Identification and Pd
Turnover (TON) greater than 1000 are all elements towards the
specification of achieving 95% probability of detecting 250 TON
Catalyst Activity increase. The key to variability reduction is the
signal to noise ratio. The fewer noise elements in each step of the
process, the more predictable the overall result.
[0057] A top level scorecard, as discussed in above is used to
determine the overall DPU's and Zst (value of short term sigma) for
the major elements of any project, the parts, the process, the
performance and software (if any) of each step. In the DPC
combinatorial technology program this includes elements 402 through
412 shown in FIGS. 10 and 11. FIG. 12 illustrates the use of a
scorecard to predict process capability with wide variability. As
can be seen the variability's are very broad and produce Defects
per Unit (DPU) and Zst levels are much less than satisfactory for
meeting the CTQ's outlined above. In the illustrative DPC project,
the unit is a Reactor Block containing 58 vials. The goal is to
reduce the DPU from the 22.73 defects per 58 vials to a much lower
number. The specification of the project is to develop a process
that has a 95% probability of detecting 250 TON catalyst activity
increase. The defects in the Process portion contain several
portions with unsatisfactory levels. These include the manual vial
preparation with a DPU of 5.58, manual analytical preparation with
a DPU of 2.44 and the use of a manual database with 9.40 DPU. Those
skilled in the art will readily realize that the goals and CTQs are
exemplary only and that different goals and CTQs will be developed
in various applications.
[0058] After the above analysis using DFSS techniques, specific
Measure, Analyze, Improve, Control (MAIC) projects are assigned to
specific portions of the overall combinatorial process. Two
illustrations are outlined in FIGS. 13 and 14. FIG. 13 illustrates
a MAIC project from the chemical analysis portion of the
combinatorial process for the DPC catalysts study. The goal in this
MAIC project is to reduce the variability in the Chemical Analysis
portion of the combinatorial process. In the combinatorial process,
there is typically extensive use of automation which requires
process capability above laboratory norms. It was found in the DPC
catalyst program that day to day drift created long term loss of
capability that required manual peak picking. The MAIC project
identified and used multiple internal standards which eliminate the
effect of day to day variability, thus reducing the overall
DPU's.
[0059] FIG. 14 illustrates a MAIC project for the Analytical
Preparation portion of the combinatorial process for the DPC
catalysts study. The goal in this MAIC is to reduce the variability
in the Analytical Preparation portion. The MAIC project first
mapped the Analytical Preparation process. It then identified the
key steps contributing to the overall variability. An example of
the variability of capability before the MAIC shows that the
variability is very wide and the peak for Internal Standard (IS)
delivery was below the Lower Specification Limit (LSL). After the
use of DOE techniques within the MAIC project, the variability of
IS delivery was significantly inside the LSL and Upper
Specification Limit (USL), and the peak was also within the LSL and
USL. The MAIC project for Analytical Preparation also compared a
Robot approach for preparation as compared to a manual approach for
preparation. The Robot approach significantly reduced the
variability in the DPC Analysis. All of these improvements led to a
DPC:IS ratio which was centered between its LSL and USL and showing
very small variation.
[0060] An additional DFSS tool for determining the variability in a
complex process is the use of highly nested Design of Experiments
(DOE's) to find the "little x's" as shown in FIG. 15. The
fundamental structure of the combinatorial chemistry process leads
to sources of variation from different experimental units (shown as
"Levels" in FIG. 15). The variation arising from these Levels
should be analyzed statistically in an appropriate manner (as
indicated by the "Degrees of Freedom" table in FIG. 15). If this is
not done properly, important sources of variation (little x's) may
not be correctly identified. As shown in FIG. 16, using these DOE
techniques resulted in focusing attention on the autoclave
temperature capability.
[0061] FIG. 17 illustrates a scorecard that demonstrates the
improvement in, variability after the study using DFSS/MAIC
techniques with combinatorial technology. For example, the
reduction in defects in the Chemical reaction portion went from
5.58 DPU to 0.27 DPU. The DPU improvement in the Analytical
preparation portion went from 2.44 DPU when using a manual system
to 0.64 DPU when a robot method was used. The most dramatic
improvement was seen from using a manual database to a computerized
database. The DPU change was from 9.40 to 0.49 DPU. As can be seen
from the Scorecard illustrated in FIG. 17, the DPU went from 22.73
shown in FIG. 12 to 5.71. The improvement was significant enough
that capability of developing catalysts in the DPC project was
deemed adequate for a pilot run.
[0062] FIG. 18 illustrates a further use of the DOE techniques. The
DFSS process identified the need for new DOE arrays, such as this
combinatorial design that allowed complete examination of all two
metal systems while reducing the time for finding new catalysts
from 1.5 years to approximately 1 month. In addition, this DOE
revealed two metal synergies for the DPC system, which became new
catalysts.
[0063] The use of the six sigma (DFSS) tools in a research project
focuses the project team on the variance of subsystems to reduce
the overall variance. It also permits the research team to link and
prioritize multiple complex steps with Quality Function Deployment
(QFD). The research team can use scorecard techniques to roll up
subsystem variance. Then the team can attack the variance using DOE
tools. The use of the DFSS changes behavior and thinking by
focusing on the variations in all steps of the process.
[0064] The present invention can be embodied in the form of
computer-implemented processes and apparatuses for practicing those
processes. The present invention can also be embodied in the form
of computer program code containing instructions embodied in
tangible media, such as floppy diskettes, CD-ROMs, hard drives, or
any other computer-readable storage medium, wherein, when the
computer program code is loaded into and executed by a computer,
the computer becomes an apparatus for practicing the invention. The
present invention can also be embodied in the form of computer
program code, for example, whether stored in a storage medium,
loaded into and/or executed by a computer, or transmitted over some
transmission medium, such as over electrical wiring or cabling,
through fiber optics, or via electromagnetic radiation, wherein,
when the computer program code is loaded into and executed by a
computer, the computer becomes an apparatus for practicing the
invention. When implemented on a general-purpose microprocessor,
the computer program code segments configure the microprocessor to
create specific logic circuits.
[0065] The computer apparatus may include an interconnection with
the internet or an intranet allowing scientists in multiple
locations to access six sigma flowdown programs and provide data on
the various variabilities.
[0066] While the embodiment described supra is related to a
catalyst research project using combinatorial techniques, the use
of DFSS techniques in a research project is not limited to
combinatorial chemistry or chemical research projects in general.
The use of DFSS techniques in any research project containing
variables that need to be focused upon is relevant to the
invention.
[0067] While the invention has been described with reference to
preferred embodiments, it will be understood by those skilled in
the art that various changes may be made and equivalents may be
substituted for elements thereof without departing from the spirit
and scope of the invention. In addition, many modifications may be
made to adapt a particular situation or material to the teachings
of the invention without departing from the essential scope
thereof. Therefore, it is intended that the invention not be
limited to the particular embodiments disclosed, but that the
invention will include all embodiments falling within the scope of
the appended claims.
* * * * *