Useful Information on CARS Style KB mirror system for Microfocusing
Started February 1, 2002
Table of Content:
New Source of Highly Polished Flats.
Peter Eng and I have found an additional source of mirror flats for his
KB bender microfocusing set up. The company formerly known as General Optics
was purchased by WavePrecision GSI Lumunics. The new company has kept the
superpolishing expertise of General Optics and provides high quality flats
from 50 mm to 300 mm.
I needed to get blanks for MHATT-CAT's bender so we tested them. To get some
mirrors more quickly, Peter provided unpolished Si blanks made by the APS
Optics Shop and we asked WavePrecision to polish them. I received the blanks
last week and their optical quality has been measured by Lahsen Assoufid's
Metrology group. Peter Eng was quite impressed by the metrology data provided
by WavePrecision. Lahsen's group results agreed quite well with WavePrecision,
namely a rms roughness of 0.7 Ang., and the maximum slope errors around 1.5
microrads. The flats were slighly convex (> a few km) and once bent, Lahsen
expects the slope error should be well below the above value. Also, the vertical
mirror error was smaller than the horizontal mirror slope errors.
The mirrors will be coated this month and shall be tested this run by Peter.
The results will be reported at TWG at a later time. Find below some links
such as a recent quote, and the mirror specifications from Peter Eng.
The quote is about 1650$ per flats which is about half price from the quote
that were circulated earlier at TWG by E. Alp. The quote for superpolishing
Peter' blanks was about 1350$ each. I am pretty confident it is worth getting
the whole job done at WavePrecision, since this will allow them the maximum
flexibility in fabrication. They do etch the substrate before polishing. The
blanks from WP are uncoated, but this work is easily done on-site. Note that the
company purchasing dept. is in Canada near Ottawa, but their fabrication
facility is still in Moorpark CA. Ken Scribner no longer works there. I've
dealt with John Tardif.
Fig 03-06.1. Vertical beam profiles with the mono set to 8.954 keV
for different ID gaps.
MHATT-CAT Alignment Notes.
In order to get the smallest beam size possible, it is very important to make
sure that white beam slits before the monochromator are positioned right on the
Undulator axis. Peter Eng recommends the following trick to define the undulator
axis easily (also published in JSR by Sandy et al. JSR vol 6 p1174-84 (1999)).
One first sets the mono a fraction of a keV below the undulator fundamental.
Then one can scan the beam profile and will find a valley at the on-axis
position. Fig03-06.1 shows the effects for 5 different
ID gaps on 7ID in the vertical direction. Fig03-06.2
shows the horizontal beam profile for 4 different ID gap with the monochromator
set to the Cu K-edge. The slit position in both direction were set to the bottom
of the valley. Note that our slit position at 7ID is typically right on-axis so
this test is time consuming and not necessary. A simple centering scan is
adequate enough so vertical and horizontal center alignment scans are all that
Fig 03-06.2. Horizontal beam profiles with the mono set to 8.954 keV
for different ID gaps.
On March 6, 2002, the scan above were performed and the beamline L5-20 slits
were centered. The L5-20 slit opening was set to 0.2 mm by 0.2 mm. Once the
monochromator energy was decided upon (on 03/06/02 we set it to 19.5 keV), one
needs to estimate the mirror grazing angle of incidence. Using this angle theta,
the active length (AL) of the mirror (about 70 mm long), the beam before the
mirror should be limited to dimensions less than AL*sin(theta) using a slit
positioned right before the vertical mirror. Typically, we use one of the XPCS
set up slit for this aperture. The best mirrors we have at the moment are Si
flats coated with Rh, thus from the critical energy needed in an experiment, one
can determine the critical angle theta. The alignment procedure for the KB setup
can be summarized below:
- Line up the beamline, L5-20 and set up the monochromator energy.
- With the KB mirrors unbent, make sure the beam can pass through undefelected
then calibrate the opening and center the in-hutch slit to define the incident
beam size on the mirror. The set up normally includes a slit, followed by an
ion chamber, then the mirror assembly.
- Next level the mirror parallel to the beam and determine the positions where
the mirrors bissect the beam. One can do that with an ion chamber positioned
downstream of the two mirrors or by using the YAG 1:1 imaging system to view
the straight through and the deflected beams. With the YAG screen position about
a meter away from the mirrors, one can easily find the level positions well
within 0.1 mrad accuracy.
- One then needs to set up the working angles of the mirrors and it is best
to measure the angles chosen with the YAG screen imaging system, with the
imaging system placed in the focal plane or not too far away since the field of
view of this system is only about 4 mm. One should not trust the EPICS motor
calibrations since the rotation stage is a sine drive and has thus a non-linear
- Once the mirror are set to their working angle, the next step is to find
the bender positions where the bending just starts and set the zeros of the
benders. Each mirror has two benders one upstream, one downstream. It is good
to determine these positions within 10 micron or so.
- Center and set up the high resolution YAG screen (10:1 magnification) to
view the doubly deflected beam. Make sure it is focussed on the front YAG
- One can finally start to bend the mirrors uniformly in steps of say 100
microns. Moving both of them simultaneously gives more curvature. Once no more
changes occur, one need to take small step and optimize each bender separately.
- Profile the focussed beam with an edge (GaAs cleaved blade or wire).
New Mirrors received on 08/22/02.
On 8/22/02 we received 4 new flats with specs from WavePrecision. For our order
this time, we did not send unpolished blanks from ANL to WP but asked them to
fabricate the whole thing. We received ours yesterday and 4ID received theirs a
few days ago.
The metrology that they provided is far from the first set we received from
them. The roughness quoted are excellent 0.5 Ang. rms but the slope errors
quoted are ranging from 7-10 urad rms. WP said that they did not remove the
best cylinder from their data. 4ID has requested the optics lab to measure its
set and on 8/23/02 Lahsen Assoufid provided me with results from the 4ID flats.
The slope errors are good when best cylinders are removed with figure errors
of 0.9 and 1.33 urad rms for their two mirrors. The best fitted cylinders have
radius of -13km and 3.67 km, consistent with our specs, and the roughness is
below 1 Ang (between 0.6-0.8 Ang.). Thus WP seems to have delivered 4ID with
mirrors consistent with our specifications. We will have to get ours profiled
as well soon to see.
Back to my homepage (retour vers ma page maison).
Récemment modifié le 03/18/2004
par Eric Dufresne,
Bldg 432 Rm D007 630-252-0274