Q: What is the difference between a traceable calibration
STANDARD and an ordinary calibration REFERENCE specimen?
A: The traceable calibration standard is a select grade of material
that is individually measured and certified. The quality is higher, with
fewer defects than would be found on the ordinary calibration
specimen. Each piece is individually measured against a transfer
standard previously measured by a national laboratory, which has certified
that the measurement value is traceable to the international meter.
Our certificate of traceable calibration:
- describes the measurement procedure,
- identifies the calibration standard used and its uncertainty as stated by
the national laboratory,
- summarizes the measurement results,
- and provides a statement of uncertainty for both the individual and
average values. The uncertainty statement is given in a form that
meets that international guidelines for expressing uncertainty of
measurements.
Collaborative projects and joint
publications with specialists at several national laboratories show the
accuracy of our measurement procedures. For the 144-nm standard (model
150-2DUTC), ASM's average pitch agreed with PTB's value within 0.033
nm. For the 70-nm standard (model 70-1DUTC), ASM's average pitch
agreed with NMC's and NIST's values within 0.025 nm.
In contrast, the calibration reference specimen is supplied with a
non-traceable certificate which states the average pitch based on a batch
calibration. A rough statement of uncertainty is given, such as
"accurate to +/- X nm" and "standard deviation of single
values is less than Y nm."
Customer comment: As a nano-metrologist, traceability to SI units and measurement uncertainty estimations are part of my daily life. After having carefully read your paper on the corresponding interlaboratory comparison (ILC) study, I am convinced that, in terms of SI traceability, your calibration standard meets our requirements. I do have a few remaining questions concerning the stated uncertainty assigned to the pitch
size.
Q: In the 70-nm pitch study, the measurement uncertainties estimated by the three ILC participants (ASM, NIST and NMC) ranges from 0.017 nm to 0.028 nm. I guess that these uncertainties originate from repeatability data, meaning that the measurements performed in each laboratory have been performed on one day?
A: The stated uncertainties come from uncertainty models devised by each of the three labs and described in the paper. The uncertainty models include repeatability and many other uncertainty components. I cannot speak for NIST or NMC/A-STAR, but the two runs at ASM took place on different days.
Q: On the information sheet for the 70-nm pattern, it is mentioned that the pitch size is accurate within +/- 0.25 nm. However, further down, it is written that the single pitch value has a typical uncertainty of +/- 0.5 nm (95 %
confidence interval). Which one is the certified uncertainty?
A: I apologize for some confusion between the rough statement of accuracy
for the calibration reference ("accurate within +/- 0.25 nm") with
a single pitch uncertainty stated for the standard ("+/-
0.5 nm"). The following remarks may clarify the situation:
Model 70-1D is a non-traceable calibration reference, not a traceable standard. No uncertainties are certified for this model. The calibration sheet we provide with
the reference states:
“Pitch Period: 70.1 nm (+/- 0.25 nm)
The specimens are produced by a series of process steps including interferometric lithography. The period of the line-space pattern is uniform without stitching errors. Standard deviation of individual pitch values < 0.5 nm when measured in a 3 um image, using DiscTrack Plus™. Point to point measurements will be less precise, due to surface and edge roughness.”
Q: The uncertainty values from the ILC study and shown on the information sheet differ by a factor of 10. So I guess, some other uncertainty contributions (e.g., sample homogeneity, stability, etc.) have been included in the overall uncertainty budget?
A: Model 70-1DUTC is a traceable standard. Its certificate of traceable calibration has certified uncertainties both for the mean pitch and for single pitch values. Typical uncertainties (K=2) are better than +/-0.05 nm (mean) and +/-0.50 nm (single values). The traceability path uses the NIST-measured specimen that was the subject of our paper.
For this paper, ASM measured more than twice as many images as we do in our
normal certification process, in order to produce a lower uncertainty for
the average pitch.
Q: Do you have any information on the stability (shelf life) of the standard? How long can you guarantee the certified value and corresponding uncertainty?
A: We’ve been working with the 70-nm grating for more than 2 years and have seen no evidence of pitch change. We’ve been working with other gratings also on silicon substrates since 1996 and also have seen no evidence of pitch change. It is important to use and store the standard carefully and to avoid measuring in a dirty or damaged area.
One SEM user
reported apparent changes in line width when a small area was exposed to a
very high doses of electrons by scanning at very high magnification for a
very long time. It is known that silicon oxide can shrink by
radiolysis and that cumulative dose is more important in this process than
HV settings (D. Joy, personal communication 2011). ASM analyzed the user's
lower magnification image, which showed a "normal" region
surrounding the "altered" region. The DiscTrack Plus™ high
precision measurement software showed that the altered region had a large
increase in contrast (>30%) between the lines and spaces and a small
decrease in apparent line width (< 10%). However, there was no
change in pitch (at the 0.1 nm level, which was limited by the quantity of
data available for analysis), as long as the pitch measurements excluded
features in the boundary between altered and normal regions. Therefore
we recommend: For the most accurate results, always measure pitch in an area
that has uniform contrast. Do not bridge across areas with noticeably
different appearance.
Q:
Will the sample come with a detailed cleaning protocol, if not, do you have any information on that and which you are willing to share with us?
A: The patterned area of the 70-nm grating is fragile. We cannot recommend any cleaning protocol and instead advise “do not clean”.
updated 06/04/2011