Title: FEL Offset Mirrors
1FEL Offset Mirrors
- LCLS Week FAC Meeting
- 2627 October 2005
2Overview
- FEL offset mirrors will serve as a low-pass
filter, eliminating the high-energy spontaneous
spectrum - Basic concept is to rely on the relatively sharp,
step-function nature of grazing-incidence
reflectivity curves - Original plan calls for two sets of mirrors
- Be operates below 2 keV _at_13 mrad
- SiC operates above 2 keV _at_1.5 mrad
3Approach
- Actual implementation must account for
- Safety
- Cost
- Complexity
- Guided by ISM Core Functions
- Define work scope
- Analyze work for hazards (e.g., use of beryllium)
- Must start with physics requirements
4Original concept
- Mirrors made from monolithic blanks
- Graze angles of 1.5 and 13 mrad cut off
reflectance at 2 and 24 keV - Mirror pairs spaced to give FEL 24 mm of offset
- Mirrors must not increase divergence (i.e.,
decrease brilliance) ? Very stringent specs
5Length of the mirror
Projected Footprint on Mirror
E (keV) Beam Size (mm) Beam Size (mm)
E (keV) FWHM projected
Be mirror, q 13 mrad Be mirror, q 13 mrad Be mirror, q 13 mrad
0.83 1.118 86
1.00 0.990 76
2.00 0.551 42
SiC mirror, q 1.5 mrad SiC mirror, q 1.5 mrad SiC mirror, q 1.5 mrad
2.0 0.551 367
3.2 0.367 245
4.8 0.272 181
6.7 0.216 144
8.3 0.190 127
- Mirror length L must be larger than the
projection of FEL beam of width w - Driven by lowest energy to be focused
- Be mirror must be 200 mm long
- SiC mirror must be 600 mm long
6General approach to mirror specifications
- Need to worry about three regimes
- Low-spatial frequency (i.e., figure)
- Length scales gt 1 mm
- High-spatial frequency (i.e., finish)
- Length scales lt 1 micron
- Mid-spatial frequency (typically called mids)
- 1 micron lt length scales lt 1 mm
- Overall goal is to increase divergence less than
10
7Impact of errors
- High- and low-spatial frequency errors mainly
impact throughput - Mids will broaden PSF
Figure 7 JE Harvey, Applied Optics, 34, 3715,
1996
8Figure specifications
- Mirrors should not increase divergence more than
10 - At 8 keV, divergence Ds 1 mrad
- Two independent reflections, Dsmirror 71 nrad
(0.015?) - 7.1 nm PV over 100 mm
- Very challenging!
- 600 mm long CVD-SiC mirror made for ESRF had
slope error of 3 µrad (0.3?)
9Mids specifications
- Past experience with grazing incidence optics
indicates mids are often the limiting factor in
performance. - Specs will be determined through analytic methods
and Monte Carlo simulations.
10Finish specifications
Normalized SiC throughput versus s
- To first order, finish or micro-roughness will
only impact reflectivity (throughput). - Must consider how R2 (two mirrors) degrades as
micro-roughness s increases. - Essentially little impact as long as s 6 Å.
11Concerns with current baseline plan
- Material choices
- Use of beryllium requires additional safety
measures - Be is incredibly difficult to polish
- Current work indicates best finish achievable is
the range s 1525 Å - Fabrication method
- Monolithic mirrors are very expensive
- State-of-the-art mirrors may not even meet spec
Is there another approach that can reduce cost
and risk?
12Thin coatings on silicon substrates
- Deposit SiC and B4C (instead of Be) on
super-polished, figured Si substrates - Advantages
- Lower cost
- Eliminates beryllium-related safety issues
- Leverages expertise and infrastructure developed
at LLNL for the EUVL project - SiC and B4C properties currently being optimized
for their use in multilayer applications for LCLS
13Sputtered SiC
41 nm of a-SiC has performance similar to a thick
mirror
14Sputtered B4C
47 nm of B4C has performance similar to a thick
Be mirror
15Conceptual approach
- Fabricate Si substrates with appropriate figure
and finish - Deposit thin films on substrates
- Substrate length limited to lt 150200 mm with
current deposition chambers - Tile pieces into long mirrors
- Mount coated-segments into mirror fixtures
coated Si substrate
fixture
16Outstanding issues
- Need to perform FEA to determine optimal
substrate shapes - Requires mirror specifications and detailed
design of fixtures (e.g., gravitational sag) - Determine if gaps will impact performance
- Verify that silicon substrates will not be
affected by high-energy spontaneous beam