This is especially important when illuminating very small samples, as in protein microcrystallography where crystals can be on the order of a micron across and diffract weakly compared to larger crystals. Any excess scatter in these conditions will contribute unwanted noise and decrease the overall signal-to-noise ratio – an important measure of data quality. Consider an experiment where you first must take the water from a firehose and somehow get a water thread thinner than a human hair without any mist! That is akin to the scale of creating x-ray microbeam at CHESS.
One solution would be to simply block the x-rays down to the size desired, but this has the unfortunate side effect of throwing away vast numbers of photons. Fortunately, x-rays can be manipulated similar to visual light and therefore focused using optical components such as mirrors and lenses. Recently, an optical design of interest at CHESS incorporates the focusing power of x-ray compound refractive lenses (CRLs) to create an x-ray beam on the order of microns across – effectively, a microbeam.
CRLs are a linear array of x-ray lenses (Figure 1) which can focus x-rays in the energy range of 5-40keV. Typical CRLs are made of a low Z material, such as beryllium or aluminum as normal optical material (usually made of glass) readily absorbs too many x-rays to be a feasible x-ray lens material. Additionally, the x-ray refractive index for materials is slightly less than one, which has the effect of producing very long focal lengths. Thus, lenses are designed with a small radius of curvature and stacked in series to shorten the focal length to less than a meter (a very desirable specification in the sparse real estate of beamline end stations).


This past spring, MacCHESS purchased a set of CRLs for both the serial microcrystallography and BioSAXS projects to share in order to provide both projects with increased flux at very small sample footprints. The goal of these lenses is to provide a much better vertical focus than the one obtained with upstream toroidal mirrors. At 12keV (a suitable energy for serial crystallography), the CRL design specification would yield a focal length of 250mm, a gain of ~6800 over CRL-unfocused beam, and focal spot size of 9 µm by 0.8 µm – certainly within the regime of microbeam! For BioSAXS, an ideal setup at 10keV would theoretically give a CRL a focal length of 350mm, a gain of ~4000 over CRL-unfocused beam, and with spot size of 14 µm x 1 µm.

With days before CHESS user operations halted in June for the summer down, a team consisting of Richard Gillilan, Howie Joress, Arthur Woll and Jeney Wierman, assembled and briefly installed the CRLs at CHESS beamline G1 to perform an initial characterization of the lenses. Luckily, the lens stack assembly was trivial, as the alignment is provided by a v-groove within the casing with sub-micrometer precision.
Once assembled, the casing was installed on both rotational and translational stages for alignment with the beam. Four sets of large pinholes upstream and downstream from the stack of lenses aided in very quick alignment, which took only minutes to optimize. Once aligned, the beam was trimmed and shaped upstream from the stack by adjusting the beam-defining and scatter-guard slits, and the upstream mirror.

Scans of the beam were performed with cross wires at the sample position and recorded by counts with a downstream ion chamber. The team measured a 2D array of beam-perpendicular scans at the sample position over a series of focal depths, to measure where the beam waist along the beam was located. Without much optimization of the setup, the final measurement of the focal waist of the CRL-focused beam yielded an impressive 1.6 x 1011 photons per second in an 8 x 13 µm2 beam at 10keV.
With further modification of the sample environment (helium sample chamber, for example, to eliminate excess scatter from the beam), improvements to the upstream beamline, and a shift to CHESS beamline G3 at 12keV for serial crystallography, the team believes this number will improve and more closely reflect the theoretical design parameters.
Regardless, this “new” optic design holds great promise for microbeam experiments at CHESS. Both the BioSAXS and serial crystallography teams are currently designing experiments to capitalize on the increased gain and microbeam capabilities offered with the use of the CRLs.
Many thanks to Peter Beaucage who graciously gave hours from his own beamtime to allow time for this initial installation!
References:
"Compound refractive X-ray optics" http://www.rxoptics.de, retrieved August 10 2017.
A. Snigirev, V. Kohn, I. Snigireva, A. Souvorov, and B. Lengeler, "Focusing High-Energy X Rays by Compound Refractive Lenses," Applied Optics 37, 653-662 (1998).