Beam stability requirements experience from the NSLS X

Beam stability requirements: experience from the NSLS X 1 A STXMs Chris Jacobsen Department of Physics & Astronomy Stony Brook University, NY, USA Chris. Jacobsen@stonybrook. edu

Coherence requirements for microscopy • General property of lenses: preserve size • angle product (Liouville’s Theorem), so Scanning • Microprobe/STXM: want so that diffraction limit of optic rather than geometrical image of source dominates spot size • Rayleigh resolution of thus gives phase space of • Full-field: each pixel in object is diffraction limit of optic, but can have many pixels simultaneously! • Fluctuations of illumination from one pixel to its neighbor are therefore unimportant. Full-field

Effects of spatial incoherence on STXM How close must p=hθ be to λ ? Effect on modulation transfer function MTF (50% central stop) Effect on point spread function PSF (50% central stop) Jacobsen et al. , Ultramicroscopy 47, 55 (1992); Winn et al. , J. Synch. Rad. 7, 395 (2000).

Controlling spatial coherence • Coherent flux: phase space area of λ in each dimension. Coherent flux is B • λ 2. Green, BNL-50522 (1976); Kondratenko and Skrinsky, Opt. Spectr. 42, 189 (1977). • Spatial filter: pinhole at the focus of a lens. Passes only the spatially coherent fraction of an incident beam. Diagram, photo from Newport catalog

Beam sharing at X 1 A • Electron beam emittance determines photon source emittance. • Horizontal emittance is about 100λ; vertical emittance is about 3λ. Flux losses: • Because we cannot scan the gap, we often work 2 -4 x off of the spectral peak. • Because of sharing with X 1 B we often lose around 2 x from center of horizontal distribution, and our sensitivity to noise is magnified. Incoherent flux One-mode slices for spectroscopy spatial coherence (X 1 B) (X 1 A 1 and X 1 A 2)

Optical configuration of X 1 A See Winn et al. , J. Synch. Rad. 7, 395 (2000).

Beam stability has been inconsistent Aug. 27, 2004: discrete flux changes, ~100 Hz oscillations. • Problem is intermittent; there are good beam days and bad beam days. • Beam stability was much better in late 1990 s/early 2000 than in the last few years. • Accelerator physics staff has been unable to correlate any of their diagnostics with our problems, but that may reflect limitations in beam diagnostics.

What is needed for beam stability? • Month-to-month, fill-to-fill, and hour-to-hour for steady flux levels. • For present-day NSLS, STXM experiments work with 1 -10 msec dwell time so need stability at much better than 0. 1 -1 k. Hz. For NSLS II, scale the bandwidth requirements up by the brightness gain! 8

Requirements on fluctuations • Fluctuations from pixel to pixel must be small compared to contrast variations. • How many photons do you need to see something? Signalto-noise goes like • Contrast parameter is • Number of photons required is • Rose criterion: SNR=5 (but SNR=2 still recognizable) 9

Fluorescence or scattering • Fluorescence or large-angle scattering detection with zero background has so. • Thus exposure equals number of incident photons required to produce the fluorescent or scattered photons. • Normalized fluctuation in incident beam equals normalized fluctuation in final image (few percent probably OK!) 10

Absorption contrast: 10 nm protein in ice 11

Absorption: 1 nm Au in 100 nm Si 12

Demands are tighter at high resolution • As lateral sizes get smaller, longitudinal sizes are likely to as well. • Thinner things require more photons to deliver a certain SNR, roughly as the square of making things thinner. 13

General comments • If you have a multimode beam, and need only a single mode for your experiment, you gain tolerance by overfilling your beamline and selecting the single mode at the end. • If you have a single mode beam, you have no such luxury of overfilling! NSLS II experiments will require more stability than is now needed at NSLS and even APS, ALS. 14

Angle stability requirements • Image position does not shift, but flux accepted by beamline does. 15

Position stability requirements • Shifts in the beam position produce both intensity fluctuations and image position fluctuations. If multimode, prefer higher beta straight? 16

Fast phase contrast detector M. Feser, B. Hornberger, C. Jacobsen (Stony Brook); P. Rehak, G. de Geronimo (BNL Instrumentation); L. Strüder, P. Holl (MPI München) • Silicon drift detector • Simultaneous recording of bright field, dark field, differential contrast at msec pixel times • No significant upper limit to signal rate. Acceptable dark noise (~8 photons/msec equivalent; room temperature) • High quantum efficiency (>90%) Assembly: 40 mm across Active area: 600 μm

Simultaneous Availability Of Contrast Modes • Silica spheres 1 μm diameter or less • Differential phase contrast filters out intensity fluctuations of the source!

Closing comments • STXMs/nanoprobes are very demanding on beam stability. • Overfilling the optical system is great if the beam is fat enough. • Must have very stable flux on timescales of 0. 01 -100, 000 Hz (? ) • On that timescale, must have stability in flux 4 to 1: 10 or better for transmission experiments. 19