Draft Minutes
ASC OP1 Optics and Electro-Optical Instruments Optical Elements and Assemblies Wavefront Standard
Sunday, January 20, 2008, 2:00 p.m. 5:00 p.m.
Fairmont Hotel, California Room
170 South Market Street
San Jose, CA 95113
Attending
Committee Members (14 of 24 w/1 alt.)
Representing
David Aikens
Savvy Optics Corp.
Gordon Boultbee (Alternate)
JDSU
Andre Brunfeld
Xyratex
Benjamin Catching
JDSU
Walter Czajkowski
APOMA (Edmund Optics)
Marla Dowell
IEEE/LEOS (NIST)
Lincoln Endelman
SPIE (Endelman Enterprises)
Chris Evans
Zygo Corporation
Krishna Gupta
Zygo Corporation
John M. Hamilton
Northrop Grumman Corporation
James E. Harvey
CREOL
Hal Johnson
Harold Johnson Optical Lab
Allen Krisiloff
Triptar Lens Company, Inc.
Gene Martin
Jonathan McGuire
Northrop Grumman Laser Systems
Wayne McKinney
Lawrence Berkeley National Lab
Michael Morrill
Lockheed Martin Space Systems Co.
Bruce Netherton
Lockheed Martin Coherent Technologies
William Royall
Eastman Kodak Company (Retired)
Peter Z. Takacs (by phone)
Brookhaven National Lab
Trey Turner
Research Electro-Optics, Inc.
Steven VanKerkhove
Corning Tropel
Ray Williamson
Ray Williamson Consulting
Dr. Valeriy V. Yashchuk
Lawrence Berkeley National Lab
Observers (2)
Lahsen Assoufid
Argonne National Laboratory
Gene Kohlenberg
OEOSC
Stephen Martinek
4D Technology
Auditor's Summary of Meeting
A representative from Zygo Corp. described the robust peak-to-valley calculation that is being proposed as a
replacement for the traditional visual peak-to-valley number that has been used to describe how closely a manufactured
optical surface conforms to the design.
A representative from Lawrence Berkeley National Laboratory prepared an outline of the draft BSR/OEOSC OP1.004,
"Statistical methods for measuring wavefronts and surfaces." A representative from Brookhaven National Laboratory
had begun fleshing out the outline using some material from the ISO 10110-8 standard, "Optics and optical instruments
Preparation of drawings for optical elements and systems Part 8: Surface texture."
06/06/08 11:32:37 AM
File: TF 3 Wavefront Draft Minutes, 1-20-08.odt
ASC OP1 Optics and Electro-Optical Instruments Optical Elements and Assemblies Wavefront Standard, Continued
A representative from Corning Tropel reviewed his draft outline of BSR/OEOSC OP1.005, "Deterministic methods for
measuring wavefronts and surfaces."
The Task Force agreed to hold a teleconference to examine both drafts after suggestions provided by the Task Force had been
incorporated. A second in-person meeting was also scheduled to be held in Rochester, NY during the OSA Annual Meeting in
October.
1. Welcome and Introductions
S. VanKerkhove opened the meeting at 2:10 p.m. with a round of introductions.
2. Adoption of Agenda
C. Evans moved to adopt the agenda. B. Catching seconded the motion, which carried unanimously.
3. Approval of draft minutes of the Friday November 30, 2007 meeting
A. Brunfeld moved that the minutes be approved. M. Dowell seconded the motion, which carried unanimously.
4. Review of PVr
C. Evans presented a talk on the robust peak-to-valley term, PVr. He said that the first question is, "why does the optics
community need a new parameter to describe surfaces and wavefronts?" His first answer is, "one doesn't need a new term because
many use rms irregularity along with PSD and slope." But many have the historical custom of describing surfaces in peak-to-valley
(PV) terms of wavelength, e.g. 1/10 . PV is a badly biased parameter for interferometers. However, it is difficult to move the
community away from an amplitude term.
If one looks at the bias for a real surface, one can show that it is a function of the higher moments of the surface. Zygo has
defined an amplitude parameter called PVr that is fit to 36 terms of the Zernike polynomial + 3 of residuals.
For circular apertures
PVr =PV
36Zernikes *
N×
36 ZernikeResid
, default
N =3
,
PVr PV
His first example was a 12 in. optic. He showed that PV changes greatly depending upon the number of pixels used to sample the
surface wavefront, while the PVr yields a constant result for different detector resolutions. PVr is insensitive to shop floor noise.
His second example shows obvious fliers and noise in the data. He showed that if 2% of the data were discarded or clipped, the
PVr changed very little.
There could be options specified, such as, 1<N<6, limit clip total area, and limit maximum contiguous clip area. However,
C. Evans sees no reason why the optics industry should not agree upon using N=3.
D. Aikens said that he "fell in love" with PVr because PV has been used for over one-hundred years as an easy single value to
describe and optical surface. It was never intended to be applied to statistical determination of a surface. PV was intended for
smoothly changing surfaces without tiny spikes along the way. An optician was interested in the shape of the irregularity and how
large it was. PVr gives a vendor a tool to communicate with a customer, who probably has less understanding than the vendor does
about surface variation. The vendor is able to express his test results in terms of peak-to-valley errors used by the customer, but he
can explain to that customer that he is interpreting the errors calculated by the PVr method. Ninety-nine per cent of the time the
customer will not care if the errors are computed by the PVr method. Those customers who want a more rigorous interpretation of the
test optical surface will not specify PV on their drawings.
W. McKinney said that he could see how the PVr method could cause less grief, but will the manufacturers add the function to
their interferometers? D. Aikens said that Zygo has already included it, and he thought that Zygo would not object to other
interferometer manufacturers also adopting the technique. C. Evans said that he would be delighted if others would adopt the PVr
technique.
S. VanKerkhove asked if the optics industry would be setting itself up for a confusing situation where customers would become
confused whether PV or PVr was being used. There should be a note on the drawing specifying which calculation method was being
used.
D. Aikens said that he has met no one who is opposed to the PVr concept. He thinks that the idea will catch-on and
manufacturers will begin using it without hesitation.
A. Brunfeld suggested that the PVr should be quoted over a scale length to remove the problem of different numbers of pixels
being used for the calculation. D. Aikens replied that as soon as scale length is incorporated to describe the surface, the measurement
becomes a statistical method and RMS should be used. The auto industry has abandoned PV measurements for their precision
machined parts and are making specifications using RMS.
06/06/08 11:36:12 AM
2 of 6
File: TF 3 Wavefront Draft Minutes, 1-20-08.odt
ASC OP1 Optics and Electro-Optical Instruments Optical Elements and Assemblies Wavefront Standard, Continued
5. Wavefront Standards
a) OP1.004 statistical methods for measuring wavefronts and surfaces
·
Reconciliation of definitions with ISO/DIS 25178-2 and ASME B46.1
P. Takacs said that the first page of the draft 1.004 contains an outline of what should be covered in the specification.
V. Yashchuk prepared the outline, and P. Takacs made some modifications. P. Takacs started with ISO 10110-8; he tossed
out some things, modified others and added new material as he worked through the document.
There is a problem in obtaining other standards that may contain relevant material, e.g. terminology and symbols that
could be incorporated into this draft. He is working with ANSI to see if he can get courtesy copies for his review. There are
25 standards that may have relevant terminology. The SEMI MF1811-0704 Guide for Estimating the Power Spectral
Density Function and Related Finish Parameters from Surface Profile Data standard, contains a lot of the work that
Gene Church and John Stover have completed during the past two decades to express surface roughness using the power-
spectral-density function. P. Takacs is adapting those sections that are pertinent to optics applications.
P. Takacs said that he has included symbols in the definitions of some of the terms. He does not know if these symbols
are those customarily used. He asked the members of the Task Force for input concerning the appropriate symbology. As an
example he is representing slope with "m" rather than "z" for difference in height.
C. Evans said that there is additional work by the ISO TC 213, "Dimensional and geometrical product specifications and
verification" committee that does not, yet, appear on the ISO web site as a published standard. Paul Scott from Taylor
Hobson and Professor Jane Xiang from the University of Huddersfield are involved in this work. D. Aikens agreed to contact
E. Leitner at ISO to see if the pertinent drafts of TC 213 could be made available for this project. D. Aikens asked C. Evans
to give him the list of TC 213 documents that are of interest.
·
Address micro-defects (defined both in 3.4 and 3.14)
S. Martinek was not able to attend the meeting, so this item was postponed until a future meeting.
·
Expanded 2D PSD definitions
P. Takacs said that the Task Force needs to decide how to approach this topic. The SEMI document does not cover this
situation. The standard should indicate how to handle the higher spatial-frequency data derived from measurement machines
such as optical-surface profilometers or full-aperture interferometers. S. VanKerkhove asked if P. Takacs wanted to tackle
this topic alone, or if he wanted a small sub-group to work with him. P. Takacs said that he had an idea that he would like to
document and circulate to the members of the Task Force for feedback before the next meeting.
·
A Higher Level Look at this Issue
D. Aikens asked to talk about a higher-level issue. He reviewed the work done while struggling with this subject,
and noted that the Task Force decided to separate the topic into the two drafts OP1.004 and OP1.005. The original
concept was to expand ISO 10110-8 (which covers surface texture with definitions of the parameters and specifies a
notation) to make it more robust. P. Takacs has done a lot of good work which is more of a test-and-measurement
standard, not a notation standard. So the Task Force should consider creating a third document, which references the test-
and-measurement document, to address notation a la ISO 10110-8.
P. Takacs said that he started with the original ISO 10110-8 outline, but realized that he had to explain what PSD
and slope PSD are. He was not sure how to reconcile his resulting outline with ISO 10110-8. D. Aikens replied that in
this case, the ISO approach to separate the definitions, terms, and mathematics from the notation document may be
preferred.
S. VanKerkhove said that P. Takacs document should be finished before completing the ISO 10110-8 update.
D. Aikens said that the two documents could be synchronized. He also said that he could draft and update for ISO
10110-8 in a day to be presented to TC 172/SC 1 meeting in Bucharest, Romania in May.
D. Aikens continued by saying that there has to be a US national standard covering notation. Perhaps notation
should be incorporated into OP1.005, and the Task Force should move the Zernike polynomial decomposition process to
OP1.004. P. Takacs said that there needs to be a lot of background information documented before moving into a
notation. S. VanKerkhove assumed that notation would be included in each of the current drafts. M. Dowell proposed
that the Task Force continue with the two drafts as planned, and put notations in each of them.
P. Takacs said that his approach was to generate a document that defines how to calculate the desired parameters
from the machine data so that it can be recorded on a drawing. What he has written is background information that is
necessary to understand the number that is included in the drawing notation specified by a US national standard. His
work is analogous to ISO 14999-1 through 4.
06/06/08 11:36:12 AM
3 of 6
File: TF 3 Wavefront Draft Minutes, 1-20-08.odt
ASC OP1 Optics and Electro-Optical Instruments Optical Elements and Assemblies Wavefront Standard, Continued
D. Aikens concluded that there will need to be another sub-task force to update ISO 10110-8 for the upcoming
meeting in Romania so that there is a good notation for roughness. The international committee is going to want a
simple update of ISO 10110-8, which is being used in 13 countries. D. Aikens will look at the update Of ISO 10110-8 so
that this Task Force can continue working with OP1.004. Sometime in the future the notations can be reconciled.
W. McKinney said that he does not know how to do 2D PSD fitting for non-isotropic surfaces. The way the data is
detrended effects the results. Settling the problem will take some work.
·
rms, rms slope, PSD in the foreword
P. Takacs said that once the power-spectral-density function is calculated, then rms and rms slope can be obtained by
summing over the appropriate band widths. He will work up some examples.
·
Slope
P. Takacs said that slope is more easily done in the frequency domain. He will add mathematics for filtering the slope.
C. Evans asked if P. Takacs had considered the structure function used by astronomers. It is the specification of a
surface that is the expectation value as a function of separation of points on the surface. W. McKinney said that it would be
like Church's D function.
S. VanKerkhove asked if there was an ISO document that defines slope. D. Aikens said that he knew of none; however,
ISO 10110-12 references the specification of a local slope on an asphere, but there is no description about how to calculate
the slope. S. VanKerkhove suggested that the Task Force find an industrial user who understands the use of slope and ask
that person to propose how the standard should be written. D. Aikens suggested that P. Takacs would be that person.
P. Takacs agreed that he should write the slope section.
S. VanKerkhove asked if a user specified either a 1D or 2D PSD, would the slope also need to be specified.
W. McKinney replied that the slope would not need to be specified. He thought that metrology labs would like to see slope
eliminated. D. Aikens said that slope is a direct predictor of focus-ability. The rms slope values tells the designer how well a
spot can be imaged. S. VanKerkhove said that slope should be treated similarly to how peak-to-valley is handled. The
standard should direct that designers specify PSD. D. Aikens agreed with that approach: do not give a good method for
specifying slope because it should not be used. The Task Force should provide a definition that the manufacturers can
reference to show how they calculate slope, if they insist upon using it.
P. Takacs reviewed the discussion by asking if the Task Force wanted to put emphasis on PSD rather than slope.
S. VanKerkhove replied that the standard will need a provision similar to PVr for those who insist upon using slope. C. Evans
said that Code V types of programs do not have a PSD optimizer. D. Aikens said that the stray-light programs are the ones
that deal with surface scatter. He presumed that the computer program ASAP would already allow an engineer to enter a
PSD value for a surface. One would never design a lens using PSD, but one would evaluate the design using PSD.
P. Takacs said that slope can be avoided when photons are considered to be bullets, but when one deals with coherent
wavefronts you need detailed height or phase information for the wavefront. So the standard would have to provide for both
incoherent and coherent optical systems.
S. VanKerkhove suggested that the topics of 2D PSD and Slope be combined for future meetings.
·
Structure of Measurement Section
D. Aikens thought that it is too early to talk about the measurement section.
·
Bandwidth limits
D. Aikens said that the Task Force should clarify bandwidth over the statistical property. The measurer needs to make
sure that their measurements accommodate that specification.
S. VanKerkhove thanked P. Takacs for the work that he has done.
b) OP1.005 deterministic methods for measuring wavefronts and surfaces
·
Background
S. VanKerkhove reviewed the past efforts concerning the draft of this standard. He distributed copies of the draft outline
of BSR/OEOSC OP1.005 that was posted on the web site for this meeting. He said that the new draft can look similar to
ISO 10110-5 and 10110-14. He prefers the 1996 version as opposed to the current version. He proposed that the Task Force
examine his draft outline at this meeting. The areas addressed are
Normative Reference
Definitions
Terms and Functions
06/06/08 11:36:12 AM
4 of 6
File: TF 3 Wavefront Draft Minutes, 1-20-08.odt
ASC OP1 Optics and Electro-Optical Instruments Optical Elements and Assemblies Wavefront Standard, Continued
Specification of Tolerances for Surface Form Error
Indication on Drawings
Examples of Tolerance Indications
S. VanKerkhove said that he would like to incorporate the Zernike Polynomial definitions that were included in the
ophthalmic draft standard ISO/DIS 24157.
D. Aikens said that after standardized pixel-filtering is defined, then it could be discussed in the notations section or the
one following.
C. Evans asked how those who use test plates would use this. S. VanKerkhove said that this requirement would be left
off of the drawing. D. Aikens said that the "quarter-wave" type specification would be included in the section on notations.
A. Brunfeld asked if the user is assumed to be measuring the surface or the wavefront deviation. S VanKerkhove said
that ISO decided to split wavefront measurement from surface-form measurement. For the US standard, the two
measurements are being combined in one document. A. Brunfeld said that he wanted to point out that the wavefront
deformation depends upon the propagation, and there is a problem, because the standard has to define where the wavefront is
being measured. S. VanKerkhove replied that the standard is created to provide the user with a means for noting the
specification, not where the value of the specification should be measured. The user of the standard should determine where
the value is measured. D. Aikens pointed out that A. Brunfeld's comment is valid, but for the specification of long scale-
length items, then the location of the measurement plane for a transmitted optical wavefront error of an optical system is not
that critical. Short scale-length errors will cause problems, and for that instance a note on the drawing should point out the
concern. D. Aikens cautioned the Task Force to avoid creating a notation for the location of the measurement of a
transmitted wavefront error. He thought such a notation would confuse the optical manufacturing community. Transmitted
wavefront specifications are only used when an optical surface cannot be directly measured (e.g. a buried surface).
A. Brunfeld stressed that the specification should be complete, and at a minimum a note should be required for this
instance. M. Dowell pointed out that the Task Force is now looking at the outline, and this problem can be addressed when
the document is drafted.
M. Dowell suggested that PVr should be added to the "Terms and Functions" section.
R. Williamson pointed out that section 6) should state, "Indication on Drawings: The indication on drawings
will
may
contain the following:" D. Aikens suggested that the phrase "(But not limited to:)" should be removed because this is the list
of standard notation. Notes may be written for other parameters. S. VanKerkhove said that this was a phrase intended for the
Task Force, not for the final draft.
C. Evans asked if the clear aperture is an inscribed or circumscribed circle. S. VanKerkhove said that he would add it to
the list of items to be addressed.
·
Annex
S. VanKerkhove said that these are some of the topics that have been discussed in the past. The Task Force should
decide if they are to be included here.
·
Measurement Methods
S. VanKerkhove asked if the standard should tell the user how to do interferometric metrology. What happens when the
reader is not using an interferometer?
·
Visual Interpretation
S. VanKerkhove asked if the Task Force wanted to include topics such as "Visual Interpretation" in the Annex.
D. Aikens said that this topic is included in ISO 14999-4, so that the ISO document could be referenced in the
informative section of this OP1.005. He continued that a note could be placed somewhere directing the user to ISO
1499-4. He added that wedge is covered by ISO 10110-6.
·
Non-Circular Apertures
S. VanKerkhove said that ISO 10110-5 described "non-circular test areas." He asked C. Evans if he had given any
thought to employing PVr over a non-circular aperture. C. Evans said that polynomials can be defined for most common
non-circular apertures. S. VanKerkhove concluded that this topic should be included. D. Aikens recommended that the
Task Force not develop polynomials for the other shapes, but rather include a note describing what must be considered
for non-circular apertures.
·
Power Spectral Density
S. VanKerkhove said that this topic would be covered in OP1.004.
·
Validation of Standard
06/06/08 11:36:12 AM
5 of 6
File: TF 3 Wavefront Draft Minutes, 1-20-08.odt
ASC OP1 Optics and Electro-Optical Instruments Optical Elements and Assemblies Wavefront Standard, Continued
S. VanKerkhove said that D. Aikens had suggested this topic. D. Aikens said that it should be "Validation of the
Specification."
·
Uncertainty (K-value)
D. Aikens suggested that this could be covered by a normative reference to the document that defines uncertainty.
C. Evans said that the NIST Technical Note 1297, is "Guidelines for Evaluating and Expressing the Uncertainty of
NIST Measurement Results." The combined standard uncertainty
u
c
is a widely employed measure of uncertainty.
The NIST document states, "what is often required is a measure of uncertainty that defines an interval about the
measurement result y within which the value of the measurand Y is confidently believed to lie. The measure of
uncertainty intended to meet this requirement is termed expanded uncertainty, suggested symbol U, The value k is
known as the coverage factor in the equation
U =ku
c
y
. C. Evans said that the standard should say that
k=2
unless
otherwise specified. D. Aikens said that he prefers that
k=1
. M. Dowell said that
k=2
is like saying 2. Depending
upon the distribution of the data, it could be a 95% confidence interval.
6. Time and Place for next OP/TF 2 Wavefront Meeting
After looking at the options, M. Dowell suggested that the Task Force hold a teleconference in the spring, and then meet in
Rochester in October. S. VanKerkhove replied that if there is a teleconference, then he wants specific action items on the agenda.
Drafts should be made available before the call. G. Kohlenberg will not be available from May 1 through May 19, D. Aikens will be
unavailable May 19 23, and P. Takacs will be unavailable the last week of May through the first week of June. S. VanKerkhove will
be in Germany the second week of June. D. Aikens said the main reason for a teleconference is to encourage P. Takacs and
S. VanKerkhove to get the documents ready. D. Aikens moved that the Task Force meet in person in Rochester during the October
OSA Annual meeting and that it hold a teleconference, sometime before the October meeting, at a time chosen by S. VanKerkhove
and P. Takacs. C. Evans seconded the motion. The motion carried unanimously.
7. Adjournment
D. Aikens moved that the meeting be adjourned. M. Dowell seconded the motion. The meeting was adjourned at 5:10 p.m.
06/06/08 11:36:12 AM
6 of 6
File: TF 3 Wavefront Draft Minutes, 1-20-08.odt