1 BROOKHAVEN SCIENCE ASSOCIATES Lonny Berman EFAC May 10 th 2007 ID Beamline Optics and Damping Wigglers

1 BROOKHAVEN SCIENCE ASSOCIATES

Lonny Berman EFAC May 10th 2007

ID Beamline Optics and Damping Wigglers


Outline

• Hard x-ray undulator beamline monochromator thermal modelling (silicon and diamond)

• Hard x-ray undulator beamline mirror modelling• Damping wiggler beamline monochromator

thermal modelling (silicon)• Damping wigglers: to cant or not to cant


Undulator Beamline Silicon MonochromatorSimulation of “hockey-puck” liquid nitrogen cooled crystal design using 2σ beam size (1.2 mm wide x 0.75 mm high) from U14 superconducting undulator at maximum K, 30 m distance from source, 12.7º Bragg angle (corresponds to 8.9 keV for Si(111)).

Total power = 109 W (unfiltered), 92 W (filtered)

Maximum temperature = 98 K (unfiltered), 94 K (filtered)

Maximum thermal slope error in beam footprint = ±5 µrad (unfiltered), ±4 µrad (filtered)

Courtesy of Viswanath Ravindranath

-5-4-3-2-1012345

-20 -15 -10 -5 0 5 10 15 20

Crystal surface dimension in y-direction (mm)

Slop

e er

ror (

mic

ro ra

dian

s)


Performance for Longer Undulators

U14 Undulator Length (m)

Filtered or Unfiltered Total Power at Maximum K

Maximum Temperature (K)

Maximum Thermal Slope Error (µrad)

2 unfiltered 109 98 ±5

2 filtered 92 94 ±4

4 unfiltered 219 144 ±7.4

4 filtered 185 125 ±1

6 unfiltered 326 232 ±100

6 filtered 275 180 ±32


4 m Long Undulator Case with Filter

Conclusion: for silicon, there is no better temperature for the illuminated area of the crystal to be at, than 125 K.

Action Item: investigate the use of controls and diagnostics to make sure that this is the temperature of the illuminated area of the crystal.

-8-6-4-202468

-20 -15 -10 -5 0 5 10 15 20


Slop

e er

ror (

mic

ro ra

dian

s)


Using a Water-Cooled Diamond

The diamond wafer is 0.5 mm thick and the (111) Bragg reflection is used (Bragg angle is 19.7º at 8.9 keV). Beam size 2.4 mm wide x 1.5 mm high (4σ).

Here, 60% of the incident power is transmitted through the crystal, and the depth dependence of the 40% absorbed power has to be taken into consideration. The transmitted power propagates through a hole in the copper substrate. There is 1 mm of overlap of the diamond wafer and copper substrate all around. A crucial consideration, in such a design, is the thermal contact between the diamond and the substrate.

Courtesy of Paul Montanez


Diamond Crystal PeformanceMeans of Thermal Contact of Diamond with Copper (Presumed Heat Transfer Coefficient in W/mm2-°C)

Power Absorbed by Diamond (W)

Maximum Temperature of Diamond (°C)

Maximum Thermal Slope Error (µrad)

Ga-In Eutectic (0.04)

166 267 ±30

Brazing (0.4)

166 103 ±9

Conclusion: pathways to improvement involve better thermal contact, larger contact area, thinner diamond to absorb less power.

Action Item: watch developments in the field, as the main challenge with diamonds is in obtaining an assured supply of large, good quality crystals.


NSLS X25 1 m Long Vertical Focusing Mirror

X25 Vertical Mirror Parameters

Source to Optic = 23 m Optic to Focus = 3.5 m

Demag = 6.6:1measured rms slope = 1.4 μrad

NSLS-Iperfect image size fwhm = 3.25 μm

calc. image size incl. fig. err. fwhm = 23 μm

NSLS-IIperfect image size fwhm = 1.1 μm

calc. image size incl. fig. err. fwhm = 23 μm

1 m long dynamically bent palladium coated mirror, fused silica substrate.

electron source size, σ

VFM

ellipsedemag. source, σ’

23 m3.5 m

X25 Mirror Bender System

Courtesy of James Ablett


X25 SHADOW Ray-Tracing

Ideal Mirror

X25 Mirror

Source

X25 Mirror, ‘width’ ~ 16 μm Ideal Mirror, fwhm = 1.03 μm

x [mm] Normalized Intensity

X25 Mirror at NSLS-IISource

Ideal Mirror

X25 Mirror

X25 Mirror, fwhm = 16.6 μm Ideal Mirror, fwhm = 3.2 μm

y [μ

m]

X25 Mirror at NSLS-IIx [mm] Normalized Intensity

y [μ

m]

rms slope=1.4 μrad

distance along mirror [cm]

vertical [microns]

vertical [microns]

X25 Mirror at NSLS-I

distance along mirror [cm]


Damping Wiggler Silicon MonochromatorWe looked at the same crystal design as we used for undulator beams, and studied its performance using different size wiggler beams.

The case shown here is based on a beam size of 3.6 mm wide x 2.25 mm high at 30 m from the 7 m long damping wiggler source (original design), Si(111) at 8.9 keV (Bragg angle 12.7º).

Total power = 475 W

Maximum temperature = 125 K

Maximum thermal slope error in beam footprint = ±6.6 µrad, not bad!

-7

-5

-3

-1

1

3

5

7

-20 -15 -10 -5 0 5 10 15 20


Slop

e er

ror (

mic

ro ra

dian

s)



Damping Wiggler Mono Using Bigger BeamWe also looked at the case of a beam size twice as large in the horizontal and vertical directions, i.e. 7.2 mm wide x 4.5 mm high.

Total power = 1.77 kW

Maximum temperature = 355 K

Maximum thermal slope error in beam footprint = ±300 µrad

Notice that the temperature rise in this case is no longer concentrated right at the beam footprint. The crystal is too small in size to handle the power.

Conclusion: bigger beams require bigger crystals with bigger cooling interfaces.


-300

-200

-100

0

100

200

300

-20 -15 -10 -5 0 5 10 15 20


Slop

e er

ror (

mic

ro ra

dian

s)


• Multiple yet independent damping wiggler beamlines may be accommodated in individual straight sections, either by viewing different off-axis portions of the same damping wiggler fan, or by viewing on-axis radiation emissions from separate wigglers which are canted by a few milliradians with respect to each other.

• The larger the canting angle, the larger the impact on the emittance of the ring. E.g. if two 3.5 m long wigglers canted by 2 mrad with respect to each other are installed in each of these 8 straight sections, then the emittance increases by 8%.

2 mrad

Critical Energy Dependence Across Each Canted Wiggler Radiation Fan

For each fan:

Ec = Ec,max(1-[ө/өmax]2)1/2

өmax = K/γ10.8 keV

Ec (keV)

ө (mrad)

10

5

00 5-5

Damping Wigglers: To Cant or Not to Cant


Flux Variation Across Wiggler Fan

Flux Conclusion: canted damping wigglers offer no advantages for flux-dependent applications except at the highest photon energies, as compared with viewing the fan at even 1.5 mrad off-axis horizontally.

Courtesy of Steve Hulbert


Brightness is a Different Matter

Insertion device codes do not handle this effect correctly, but it has been empirically observed at the X25 wiggler:

Lonny Berman and Zhijian Yin (1997)

Brightness Conclusion: when viewed at 1 mrad off-axis horizontally, the horizontal source size of a 7 m long damping wiggler will appear to be 7 mm wide, diminishing the brightness by a factor of ~20, relative to on-axis viewing.

If brightness is an important consideration for a damping wiggler source, it must be viewed on-axis. The only means to achieve this, for two completely independent beamlines, is to implement canted damping wigglers.


Conclusions• Existing liquid nitrogen cooled silicon crystal design will handle

NSLS-II undulator beams, with attention to temperature control• For water-cooled diamond crystals to be effective in NSLS-II

undulator beams, attention to crystal size and thermal contact is necessary

• Small mirror figure errors can introduce structure into an NSLS-II undulator focused beam as well as blur image

• Liquid nitrogen cooled silicon crystal designed for an NSLS-II undulator beam could handle a damping wiggler beam with dimensions of ~4-5 mm; larger size beams will need to be handled by larger size crystals

• Canted damping wigglers pose no advantage for flux-dependent applications (compared with off-axis views of a single wiggler), but will be necessary for brightness-dependent applications if more than one independent beamline per damping wiggler straight section is desired

Documents

1 BROOKHAVEN SCIENCE ASSOCIATES Lonny Berman EFAC May 10 th 2007 ID Beamline Optics and Damping Wigglers