Gage Repeatability and Reproducibility in Semiconductor Manufacturing

1
Gage Repeatability and Reproducibility in
Semiconductor Manufacturing

https//yieldwerx.com/
2
In the semiconductor manufacturing industry,
precision, reliability, and consistency are of
utmost importance. Every aspect of production and
quality control relies on accurate and repeatable
measurements. To ensure the effectiveness of
measurement systems and processes, a rigorous
measurement system analysis is required. The Gage
Repeatability and Reproducibility (Gage RR)
test, a key tool for such an analysis, is
commonly used to verify the capabilities of a
measuring instrument. Gage RR is a statistical
tool used to study the amount of variation in the
measurement system arising from the measurement
device and the people taking the measurement. It
gives an understanding of the reliability of the
measurement system by quantifying its variance,
and by comparing this variance with the overall
variability observed. Understanding and applying
the concepts of Gage RR will help in mitigating
errors in the process, and subsequently improve
the quality of the final product. This detailed
technical guide will discuss the intricacies of
Gage RR and how it can be effectively applied in
the semiconductor manufacturing industry.
Further, it will shed light on interpreting the
results and using them for better
decision-making. Gage RR The Concept The Gage
RR test assesses the measurement system as a
whole, including the instrument itself, the
operator using it, and the procedure for its use.
It does not evaluate the product or component.
The test is based on the idea that the variation
in measurement results originating from the
instrument should be significantly smaller than
the variations from the measured parts or
components. Two key elements in Gage RR analysis
are repeatability and reproducibility.
Repeatability refers to the variation in repeated
measurements from the same instrument, operator,
part, and time. Reproducibility refers to the
variation in repeated measurements when there are
changes in conditions like different operators,
or times. For instance, in the context of a
semiconductor manufacturing environment,
repeatability could be how consistently a
Scanning Electron Microscope (SEM) measures a
particular feature size on a wafer lot when used
by the same operator and under the same
conditions. Reproducibility, on the other hand,
might reflect how the SEM's measurements vary
when used by different operators or at different
times. However, Gage RR goes beyond just these
two factors. It also takes into account the
operator's skill and understanding of the
measurement procedure, which can significantly
impact the measurement results.
3
Requirements for a Gage RR Test To perform a
Gage RR test, the following requirements must be
met Parts A sample of 5-10 parts manufactured
with varying dimensions. In the context of
semiconductor manufacturing, these could be
wafers with varying feature sizes or other
parameters of interest. Operators At least two
operators are needed for the study to compare
measurements. These could be quality control
inspectors or machine operators, depending on the
specific use case. Measurement System The
measurement system to be analyzed must be
identified. This could range from an optical
microscope to a more complex system such as an
Atomic Force Microscope (AFM), depending on the
parameters being measured. Measurement
Repetitions Three or more measurement repetitions
for each part and each operator. The number of
repetitions needed may depend on the variability
in the measurements and the precision
required. Statistical Software Gage RR procedure
and analysis requires robust statistical software
capable of handling computations. Popular
software includes Minitab, JMP, and even Excel
with certain add-ins. Gage RR Procedure An
understanding of the mathematical models
underpinning Gage RR analysis is beneficial for
interpreting results, but most calculations can
be performed using statistical software. The test
uses the Analysis of Variance (ANOVA) method,
specifically a balanced two-factor crossed-random
model.
4
The steps in conducting a Gage RR study are as
follows Data Collection Gather measurements
from each operator for each part, repeated as per
the requirement (usually 3 or more times). This
step should follow a randomized order to
eliminate bias. Data Entry Enter the collected
data into the statistical software, ensuring it
is properly structured for the Gage RR
analysis. Analysis Run the Gage RR analysis,
typically found in the Quality Tools or
Measurement System Analysis menu in most
statistical software. Interpretation Interpret
the results based on the percentage of total
variation coming from the measuring instrument
and operator. Interpreting the results depends on
the percentage of total variation coming from the
measuring instrument and operator. An instrument
can be used for measurement tasks at hand if this
percentage is below 10 if it is below 20, the
instrument may still be suitable, but with
reduced efficiency. If the percentage exceeds
20, the instrument is generally deemed unfit for
use and may need repair or replacement. Case
Studies Let's consider two examples of Gage RR
analysis in the context of yield in semiconductor
manufacturing Case 1 A semiconductor
manufacturer conducted a Gage RR analysis on
their high-resolution SEM used for measuring
feature size on a newly developed chip. They
found that the percentage of total variation due
to the measurement system was 25. This indicated
that their SEM was not suitable for the task, and
further investigation found that the SEM needed
recalibration and maintenance. Case 2 In another
scenario, a semiconductor company conducted a
Gage RR study on their ellipsometer used for
measuring the thickness of oxide layers on a
wafer. The study showed that the percentage of
total variation due to the measurement system was
only 8, suggesting that their measurement system
was capable and effective. These examples
illustrate how to interpret Gage RR test results
and underline the importance of this tool in
verifying the capability of measurement processes
in the semiconductor industry.
5

• Conclusion
• Gage RR analysis plays a crucial role in the
  quality control and process improvement
  strategies of semiconductor manufacturing. By
  accurately assessing the measurement system's
  capability, companies can ensure that their
  products meet the stringent requirements of
  precision and repeatability demanded by the
  industry.
• References
• Measurement Systems Analysis (MSA), 4th Edition,
  AIAG (Automotive Industry Action Group) - This
  manual provides a comprehensive understanding of
  MSA, including Gage RR.
• Wheeler, D.J. and Lyday, R.W. (1990). Evaluating
  the Measurement Process. Second Edition, SPC
  Press, Inc.
• Pyzdek, T., Keller, P. (2018). The Six Sigma
  handbook. McGraw-Hill Education.
• "Guidance for Industry and FDA Staff Statistical
  Guidance on Reporting Results from Studies
  Evaluating Diagnostic Tests", U.S. Department of
  Health and Human Services, Food and Drug
  Administration, March 2007.