Stability needs for state-of-the-art user experiments at NSLS-II - PowerPoint Presentation

Slide1

Stability needs for state-of-the-art user experiments at NSLS-II

Garth J WilliamsSlide2

Representative experiments

We consider two microscopy techniques as examples

Differential Phase Contrast (DPC)

Physical limits on image contrast are driven by the index of refraction, so phase contrast is often dominant. Normally, a sample is

scanned

through a focused x-ray beam.

Coherent Diffractive

I

maging (CDI), including full-field, reciprocal-space Bragg CDI and scanning-point, real-space

ptychography

Uses a continuous, normally far-field, diffracted intensity to recover the complex field leaving the sample. The amplitude and phase of the field is then interpreted to discover structure and deformation. This normally requires data sets

formed by collections of 2D images

. The requirements on the coherence of the x-ray field are very stringent.Slide3

Representative experiments

We consider two microscopy techniques as examples

Differential Phase Contrast (DPC)

Physical limits on image contrast are driven by the index of refraction, so phase contrast is often dominant. Normally, a sample is

scanned

through a focused x-ray beam.

Coherent Diffractive

I

maging (CDI), including full-field, reciprocal-space Bragg CDI and scanning-point, real-space

ptychography

Uses a continuous, normally far-field, diffracted intensity to recover the complex field leaving the sample. The amplitude and phase of the field is then interpreted to discover structure and deformation. This normally requires data sets

formed by collections of 2D images

. The requirements on the coherence of the x-ray field are very stringent.Slide4

Scanning transmission DPC in detail

The intensity is attenuated and shifted

The angular shift is due to phase retardation in the sample

The shift is measured by comparing opposing

d

etector elements.

This is governed by the index of refraction in the sampleSlide5

Fluorescence imaging

measured “at the focus”

Horizontal Phase Gradient

Measured “at ~0.5 m away from the focus”

1 um

Stability in DPC at NSLS-II

Courtesy Yong ChuSlide6

doi:10.1038/nphoton.2012.209

Coherent Diffractive Imaging

CDI requires highly coherent x-ray fields

CDI solves an inverse problem to recover the complex amplitude of the x-ray field leaving the sample.

The recovered field is interpreted to gain structural information.

We will discuss

ptychography

—a scanning CDI method—and

Bragg CDI

, which relies on reciprocal-space measurements and recovers material deformation.Slide7

Bragg CDI

Collect the 3D intensity distribution around a Bragg peak

Apply iterative “phase retrieval” magic

Recover shape and deformation information at the

nanoscale

with 10

-4

or better sensitivity

Newton et al.

DOI

: 10.1038/NMAT2607Slide8

How does beam stability affect Bragg CDI?

When the motion is fast compared to the measurement time, typically 1 sec or less, the beam motion “smears” the measured signal—it is effectively partially coherent radiation!

It is possible to accommodate

this, but it is vital that the

angular distribution be known.

Stability should be maintained

at the urad level and measured

to the 10s-of-nrad level

Whitehead et al.,

doi

:

10.1103/PhysRevLett.103.243902Slide9

Section summary

Often,

beamlines

perform spatial filtering at a secondary source formed by the optical system. These geometries allow a good degree of isolation from beam motion in the focal plane of the final optics.

Out of the final focal plane, beam pointing instabilities are a significant source of measurement error

Some techniques (CDI) invoke significant data analysis that can accommodate pointing errors, but more direct techniques (DPC) suffer.Slide10

Instrument stability

C

onsider the typical conditions for x-ray experiments

Present HXN measurements as an example of what can be achieved

Highlight limitationsSlide11

Ground Vibrations

Courtesy, Nick Simos

Nanoprobe

Site, 2009

~15 nm

~50 nmSlide12

Vibrations at the MLL Microscope

~1.6 nm

@ 165 Hz

Vertical MLL

Horizontal MLL

Sample, Horizontal

Sample, Vertical

~0.6 nm

@175 Hz

~0.3 nm

@ 250 Hz

~ 0.25 nm @ 295 Hz

Courtesy Yong ChuSlide13

HXN Optical Layout

X-ray

Source

0 m

FE

XBPM

16 m

Mono

XBPM

H. Coll.

mirror

H.

Foc

.

mirror

H.

mono

X-ray

Camera

V

.

Foc

.

CRLs

98 m

67 m

28.4 m

30.4 m

34.1 m

32.6 m

Secondary

Source

Vertical direction

Horizontal direction

H.

demag

= 2.3

V.

demag

= 1.9

FXBPM: sensitive to source angle and source position

MXBPM: highly sensitive source angle

XCAM: sensitive to source angle and position

Courtesy Yong ChuSlide14

Active Feedback for Beam Positioning

No Feedback

1 Hz

5 Hz

10 Hz

25 Hz

50 Hz

100 Hz

200 Hz

Horizontal Direction

Courtesy Yong ChuSlide15

No Feedback

1 Hz

5 Hz

10 Hz

25 Hz

50 Hz

100 Hz

200 Hz

Vertical

Direction

Active Feedback for Beam Positioning

Courtesy Yong ChuSlide16

Current status of HXN

Summarize current best results from HXN beam stabilization efforts

Present before and after measurementsSlide17

X-ray Angular Stability

with active feedback at 100 Hz

Vertical: 6

nrad

RMS

Horizontal: 17

nrad

RMS

Courtesy Yong ChuSlide18

W

XRF image

Ptycho

: amplitude

Ptycho

: phase

Without Active Feedback

DPC

12

keV

: w/ MLLs

Courtesy Yong ChuSlide19

With Active Feedback @ 100 Hz

XRF image

Ptycho

: amplitude

Ptycho

: phase

DPC

12

keV

: w/ MLLs

Courtesy Yong ChuSlide20

Are these solutions universal?

Beamline

-

local feedback

Optical components need to be specially designed and modeled.

The mechanical systems tend to have resonant frequencies lower than a few hundred Hz.

The feedback is typically achieved with a double-crystal

monochromator

, but not all

beamlines have them and feedback with a device designed to take the white beam may be challenging.Sensors can be problematic. In the case of BCDI, the commonly-used segmented diamond screens will adversely affect the data quality.Slide21

Talking-point requirements

Typically,

beamlines

are relatively long, with the final optics sitting more than 50 m from the source. This likely drives the reduction of pointing variation over positional variation.

1

microradian

FWHM might be regarded as a strong upper limit on pointing stability.

State-of-the art experiments will strongly benefit from 100

nanoradian

or better stability (or tracking).Beamline-local feedbacks can help, but will begin to fail to correct motion above 100 Hz. These feedback loops can also be difficult to control for the wide range of beam conditions that an experiment may require.

Experiments are already conducted with dwell times below 10 ms

and this will decrease to below 1 ms.Slide22

Supporting: Simple calculation for dependence of diffraction/deformation sensitivity from Bragg’s law