The MX endstation shares all X-ray optics with the SAXS endstation. When the SAXS station is retracted a straight section of vacuum pipe sealed at each end with 50 µm beryllium windows is inserted to allow the X-rays to reach the MX station (see §5
). All hardware that requires movement has been motorized and is controlled via
system. Equipment available at the MX station includes a fast experimental shutter (capable of 50 ms exposures), an active video feedback system for beam stabilization, an on-axis sample-visualization system, a Huber XYZ sample stage (Huber 5102.05 XY-Stage, Huber 5104.B10 Z-Stage) and an air bearing (Fox Instrument and Air Bearing, Livermore, CA, USA) for rapid ϕ-axis rotation. The air bearing enables ‘round robin’ MAD data collections where the crystal is flipped 180° after every image and the photon energy is changed every other image to measure both anomalous and dispersive differences as close in time as possible. Additional equipment includes a retractable Evex silicon drift diode fluorescence detector (Evex, Princeton, NJ, USA) with 128 eV energy resolution connected to a DSA-1000 multi-channel analyzer (Canberra Industries Inc., Meriden, CT, USA), user adjustable scatterless slits to define beam size on the sample, a cryogenic cold stream system, an automated sample-mounting system adapted from the SSRL-style SAM automounter, and an ADSC Quantum 315r CCD detector (mounted on a robust gantry system capable of 2θ offsets from −5° up to +45° and distances from 120 to 1600 mm). Details of some of these key features are described below.
4.1. Fast experimental shutter
The shutter is a model LS055 from NM Laser (San Jose, CA, USA). The same model is used for both the SAXS and the MX endstations. We have measured the jitter of these shutters, and it is ~0.6 ms r.m.s. and dominated by the electronics, not the mechanics. The shutter and spindle are synchronized by a PMAC (Delta Tau Inc., Chatsworth, CA, USA) motion controller running a control program every 2 ms. This program actuates the shutter when the spindle encoder has passed the desired opening or closing positions, generating a sawtooth-shaped distribution of opening-time errors. The r.m.s. variation of a sawtooth with 2 ms period is 0.577 ms, and the average error of two shutter events is therefore expected to be 0.41 ms. This is remarkably consistent with the 0.47 ms r.m.s. timing error inferred from the variation in the refined miss-setting angle about the spindle axis observed in MX data taken with 0.1 s exposure times. In the absence of sample drift, the miss-setting angle in this direction divided by the rotation speed is the average error of two shutter events, indicating that the 2 ms execution time of the programmable logic controller is the dominant source of error in shutter timing. For this reason, users are advised to keep their exposures above 50 ms so that the shutter does not introduce more than 1% error into partially recorded reflections.
4.2. Beam positioning
The focused uncollimated beam size at the sample is 165 × 130 µm FWHM (H × V). The user-adjustable slits (§4.4
) enable the horizontal and vertical dimensions to be independently adjusted from fully open to fully closed. During an experiment the beam must remain fixed on the sample for long periods of time, over which we observe slow thermal variations of optical supports and other environmental changes that cause fluctuations in beam position. When the user defines a small beam with the adjustable slits the problem of drift is less severe, because the beam overfills the limiting aperture directly before the sample, but it is still important to maintain a consistent beam position. Similar to other MX beamlines at the ALS, we have implemented a video feedback system for maintaining a stable beam position (MacDowell et al.
). A 50 µm-thick cerium-doped yttrium aluminium garnet (YAG; Startech Instruments, New Fairfield, CT, USA) is glued to the upstream side of the shutter blade located 170 mm before the sample in a helium-filled aluminium box that contains the experimental shutter and an ion chamber to measure beam flux. The shutter blade is inclined by 7° relative to the X-ray beam, and operates by flipping the blade in the vertical: like a diving board. X-rays cause the YAG to luminesce and the image is monitored by a CCD camera positioned on top of the helium box looking down through a viewport at the shutter. After a full tune-up of the beamline optics the centroid of the luminescing YAG image is recorded and the video feedback system makes small adjustments to the pitch of the second mirror (M2 tilt) and the roll of the second monochromator crystal (Chi2) to move the beam in the vertical and horizontal directions, respectively. The feedback system maintains a stable beam for hours, even with frequent monochromator energy changes. An identical feedback system also improves the beam quality for the SAXS experiments by maintaining a steady beam position and preventing the beam from drifting into the guard slits, which would generate undesirable scatter.
4.3. DOMO (dynamic offsite MX operator)
Our automated sample-mounting system, DOMO, was adapted from the SSRL SAM design with modifications to fit within the constraints of our MX station. The basic SAM system (Cohen et al.
) is based on a commercial Epson SCARA robot arm (model E2S453SM, EPSON Robots, Carson, CA, USA). A further modification to the SAM design was the addition of a sliding-lid sample-dispensing dewar (Fig. 5). The sliding lid minimizes vibration associated with a hinged clam-like lid as well as the production and precipitation into the dewar of the ice that forms when warm humid air mixes with cold nitrogen gas. The sample dewar holds two 96-port SSRL-style sample cassettes (Crystal Positioning Systems, Jamestown, NY, USA) for a total capacity of 192. DOMO is fully compatible with SSRL cassettes already in circulation. The sample-dispensing dewar can also hold two uni-puck adapters which accommodate eight uni-pucks, for a total capacity of 128 samples. In addition to mounting and dismounting crystals, users can command DOMO to wash adherent ice off of their samples from the Blu-Ice
interface. DOMO is equipped with a multi-axis force sensor to facilitate robot alignment tasks. At the beginning of each MX session automated procedures calibrate the magnetic picker/placer tool, the sample cassettes and the goniometer. The force sensor is also used to probe individual samples in the cassette to determine if they are loaded improperly or if they have ice on the bottom that will interfere with proper robot operation. Implementation of DOMO at the SIBYLS MX station has enabled higher throughput and greater user accessibility and is a key feature of our remote MX data collection program (§7.2
Figure 5 MX endstation showing key features of the sample-positioning system. The DOMO automated MX sample-loading robot is colored white. The supporting gantry has been omitted for clarity. The lid of the sample-dispensing dewar is shown in the open position. (more ...)
4.4. Scatterless slits
A recent upgrade to the MX endstation involved replacement of the fixed-diameter tantalum pinhole system (consisting of interchangeable 30, 50 and 100 µm pinholes) with a user-adjustable piezo-actuated hybrid tantalum metal/single-crystal slit system. Our design was inspired by similar scatterless slits developed for SAXS experiments that use silicon or germanium crystals for the slit edges (Li et al.
). One significant difference is our choice of single tantalum crystals for the slit edges over silicon or germanium, as the former provide superior attenuation of X-rays in the energy range used for MX. The slits allow the user to independently adjust the size of the X-ray beam on the sample from 10 to 130 µm in both the horizontal and vertical directions. This is useful because matching the beam size to the crystal size results in an improved diffraction limit, lower mosaicity, a lower R
and a better signal-to-noise ratio for the data (Sanishvili et al.
). Sanishvili and co-workers also showed that a small beam allows users to selectively irradiate smaller better-diffracting regions of a larger imperfectly diffracting crystal. Note that this system provides a ‘mini beam’ where small beam sizes are achieved not by focusing but rather by using slits to reduce the focused beam profile down to the desired size. For beam sizes down to 10 µm the slits may change the flux (photons per second) but do not change the flux density (photons per area per second) and, therefore, also do not change the crystals’ useful lifetime in the beam. They will still reach the Owen et al.
) damage limit (30 MGy) in about 30 min at the SIBYLS beamline (Holton, 2009
; Holton & Frankel, 2010
4.5. Sample visualization
The MX endstation is equipped with both low- and high-magnification on-axis sample-viewing systems. Initial sample alignment is done at low magnification. This has the advantage of making high-magnification alignment fairly easy, and the sample can be quickly positioned in the cryostream. The on-axis sample viewing is particularly critical for alignment of very small crystals. A 5 × 5 mm 45° mirror, with a 0.8 mm-diameter hole drilled through to allow X-rays to pass, is positioned 8 mm before the sample and reflects visible light from the sample to a 10× long-working-distance microscope objective (Mitutoyo M Plan Apo 378-803-3), with a field of view on the CCD camera of 580 × 460 µm. Diffuse low-coherence back illumination is provided by focusing a fiber-optic illuminator at the strip of white polyethylene foam that supports the back stop. The high-magnification microscope, the back stop and the adjustable slits are mounted on a large XY stage (see ‘vertical collimator’ in Fig. 5) that lowers them ~180 mm out of the way prior to sample mounting. This feature allows the goniometer spindle to be completely accessible to the DOMO sample-mounting robot without disturbing the delicate components of the optical system, the back stop or the adjustable slits. With the stage lowered, a large 45° mirror, located on the front of the helium-filled shutter box, captures the sample image on-axis and reflects it to the low-magnification microscope, an Infinity Optics K2/SC long-working-distance microscope (Infinity Photo-Optical, Boulder, CO, USA) with a field of view on the 1/2 inch CCD camera of 4.36 × 3.28 mm. This view allows for coarse alignment of the sample after initial mounting. When the vertical collimator stage is raised, the sample is viewed with the higher-magnification system for more precise final alignment.
4.6. Control system
The MX endstation uses the Blu-Ice
control system originally developed at SSRL (McPhillips et al.
; Soltis et al.
) because of the modular design of DCS
(distributed control system), the ease of customizing the Tcl
graphical user interface (GUI), and the ability to quickly and easily write new DHS (distributed hardware server) modules for new hardware. Briefly, DCS
is the central hub through which all commands are sent and complex operations are coordinated, DHSs allow specific pieces of hardware to communicate with DCS
, and Blu-Ice
is the GUI for users and beamline scientists to control the entire beamline. The SAXS endstation also uses Blu-Ice
, although it has been heavily modified for the specific requirements of collecting SAXS data (Classen et al.
). It is a testament to the flexibility and robustness of the SSRL Blu-Ice
system that it has been so easily integrated into an independent beamline design. The details of the Blu-Ice
GUI have been thoroughly described elsewhere (McPhillips et al.
; Soltis et al.
). For the MX endstation the most significant changes to the Blu-Ice
code have been to modify features unique to SIBYLS (e.g.
our 2θ offset versus
a simple detector translation at SSRL and our two-cassette sample-dispensing dewar versus
the SSRL three-cassette dewar). Additionally, many motors are controlled by the underlying LabVIEW
system, which required the development of LabVIEW
-specific ‘thin’ DHSs to translate commands back and forth between DCS