Invention and Engineering of Sample Delivery Techniques for Advanced X-ray Sources: R. Bruce Doak
ACTIVITIES. Bruce Doak and his group invent and develop novel methods of sample delivery for use at advanced X-ray sources, including X-ray Free-Electron Lasers (XFEL) and fourth generation synchrotrons. Based on their research and development, they design and fabricate well-engineered sample injectors for X-ray scattering facilities worldwide. Doak has built liquid jet sample injectors for the Linac Coherent Light Source (LCLS), the SPring-8 Angstrom Compact free electron LAser (SACLA), and the European XFEL. His group and their collaborators at the Max-Planck-Institut für medizinische Forschung have ported these same XFEL injection and sample collection techniques to biological X-ray studies at leading synchrotrons, notably the Swiss Light Source (SLS) and the European Synchrotron Radiation Facility (ESRF). His has been the leading international group in the field of microscopic liquid free-jets since the mid-2000’s, and in XFEL sample injection based on such jets since the beginning of the hard X-ray bio-XFEL measurements in 2009. The Doak group also develops complementary sample delivery methods, such as High Viscosity Extrusion (HVE) injectors for high viscosity samples as well as microscopic sample support arrays fabricated with semiconductor lithographic techniques. They recently fabricated a helium ambient scattering chamber and injector for the LCLS and demonstrated that XFEL operation is not only possible at one atmosphere pressure, but useful. Provided the surrounding gas is helium, which has a much smaller X-ray scattering cross-section than air, background scattering is sufficiently low that useful diffraction measurements result.
BACKGROUND. The last few years have seen dramatic advances in the technology of X-ray generation, notably the advent of linear-accelerator-based XFEL sources. Biologists immediately recognized the tremendous experimental potential unleashed by these new sources. The very intense photon flux enables collection of high resolution undamaged diffraction data from very small crystals. The femtosecond pulse duration makes possible room temperature measurements of the structure and dynamics of biological macromolecules, and with unprecedented temporal resolution. Nonetheless, the microscopic size and enormous photon flux of the XFEL beam impose significant experimental hurdles for biological studies, as does the vacuum environment in which XFEL beams are generated and delivered. The biological species of interest must be delivered into a micron-sized region in the center of a typically very large high vacuum chamber, without desiccating or otherwise denaturing the species and without compromising the vacuum. This requires, in general, that the sample species arrive fully solvated in a specific liquid solution dictated by the biology and not by the liquid flow characteristics. Moreover, since the incredibly intense X-ray beam annihilates any material object it encounters, the sample within the X-ray focus must be replenished at the repetition rate of the XFEL X-ray pulses.
MICROSCOPIC LIQUID FREE-STREAMS. Recalling experimental work in Göttingen decades ago by Faubel and Toennies, who first demonstrated that miniscule free-jet streams of water could be delivered into vacuum, Bruce Doak proposed in the mid 2000’s to use microscopic liquid streams to surmount the experimental XFEL hurdles delineated above. A liquid jet immediately solves the sample replenishment problem, yet raises concerns of its own. The liquid stream can be no more than a few micrometers diameter in order to avoid swamping X-ray diffraction from the biospecies of interest with that from the carrier liquid. Such very small jets are not possible with simple solid-walled convergent nozzles, since the nozzle diameter of the nozzle equals that of the desired free-stream and solid-walled nozzles this small immediately clog (particularly when exposed to problematic biological solutions). Doak accordingly developed the Gas Dynamic Virtual Nozzle (GDVN) for biological sample injection. In this device, a coaxially co-flowing gas effectively replaces the solid nozzle wall. By arranging the gas/liquid flow geometry appropriately, gas dynamic forces exerted by the gas on the liquid accelerate and elongate the liquid stream, forcing it to a much smaller diameter than the capillary from which it emerges. A micron-sized liquid free-stream results, even though all constrictions along the flow path are many tens of micrometer in diameter. Orifices this large do not clog, even with problematic biological solutions. The MPI FEL group now typically employs nozzle capillaries of 75 or even 100 micrometer diameter, with GDVN nozzle throats of roughly the same size. At a typical sample flow rate of 25 microliter/minute, this yields a jet of 3 to 4 micrometer diameter (a factor of 20 smaller in diameter than a typical human hair!). Particles or sample aggregates that would clog a 5 or 10 micrometer diameter nozzle simply pass through these larger orifices. The gas flow is momentarily disrupted but immediately re-establishes both itself and the liquid stream. Not surprisingly, the GDVN sample injector has become the workhorse injector for XFEL measurements in spite of its primary drawback, which is its high sample consumption.
ONGOING. The Doak group is developing several possible methods of reducing GDVN sample consumption by operating in triggered intermittent rather than continuous jet flow. This is of particular interest for the European XFEL, which will generate 600 microsecond bursts spaced 1/10 second apart. Of course a standard continuous-flow GDVN will be the ideal injector for the quasi-continuous XFEL’s such as the LCLS II – provided the speed of the jet is sufficiently high to remove the impacted section of the jet betewwn XFEL pulse arrives. Towards this end the Doak group has devoted considerable effort to the generation and speed characterization of GDVN jets having speeds in excess of 100 m/s. Accurate speed characterization of the jet is crucial in any case for setting pump-probe timing in time-resolved measurements. Crucial to all efforts of the Doak group is detailed experimental measurement to fully characterize the behavior of their injectors. Towards this end Doak introduced the use of very high speed microphotography for GDVN jets and, for very high speed liquid jets, nanosecond laser flash microphotography. Basic science and engineering remains an integral part of these developments. For example, the laser flash photography required a thorough of investigation of laser speckle, either how to avoid it or, when unavoidable, how to mitigate it by appropriate decohering of the laser beam.