Synchrotron beamlines typically use macroscopic, quasi-static optics to manipulate x-ray beams. We present the use of dynamic microelectromechanical systems-based optics (MEMS) to temporally modulate synchrotron x-ray beams. We demonstrate this concept using single-crystal torsional MEMS micromirrors oscillating at frequencies of 75 kHz. Such a MEMS micromirror, with lateral dimensions of a few hundred micrometers, can interact with x rays by operating in grazing-incidence reflection geometry; x rays are deflected only when an x-ray pulse is incident on the rotating micromirror under appropriate conditions, i.e., at an angle less than the critical angle for reflectivity. The time window for such deflections depends on the frequency and amplitude of the MEMS rotation. We demonstrate that reflection geometry can produce a time window of a few microseconds. We further demonstrate that MEMS optics can isolate x rays from a selected synchrotron bunch or group of bunches. With ray-trace simulations we explain the currently achievable time windows and suggest a path toward improvements.
Time-resolved synchrotron x-ray measurements often rely on using a mechanical chopper to isolate a set of x-ray pulses. We have started the development of micro electromechanical systems (MEMS)-based x-ray optics, as an alternate method to manipulate x-ray beams. In the application of x-ray pulse isolation, we recently achieved a pulse-picking time window of half a nanosecond, which is more than 100 times faster than mechanical choppers can achieve. The MEMS device consists of a comb-drive silicon micromirror, designed for efficiently diffracting an x-ray beam during oscillation. The MEMS devices were operated in Bragg geometry and their oscillation was synchronized to x-ray pulses, with a frequency matching subharmonics of the cycling frequency of x-ray pulses. The microscale structure of the silicon mirror in terms of the curvature and the quality of crystallinity ensures a narrow angular spread of the Bragg reflection. With the discussion of factors determining the diffractive time window, this report showed our approaches to narrow down the time window to half a nanosecond. The short diffractive time window will allow us to select single x-ray pulse out of a train of pulses from synchrotron radiation facilities.
We demonstrate the use of electrostatically driven micro-electromechanical systems (MEMS) devices to control and deliver synchrotron x-ray pulses at high repetition rates. Torsional MEMS micromirrors, rotating at duty cycles of 2 kHz and higher, were used to modulate grazing-incidence x rays, producing x-ray bunches shorter than 10 μs. We find that dynamic deformation of the oscillating micromirror is a limiting factor in the duration of the x-ray pulses produced, and we describe plans for reaching higher operating frequencies using mirrors designed for minimal deformation.
Epitaxial heterostructures constitute a wide variety of modern microelectronics devices. In the limit of ever decreasing feature dimensions, now entering the nanoscale in some cases, the interfaces of such devices are crucial to their operation and performance. In general the properties of the interfaces will differ significantly from those of the bulk structure of either the substrate or the heteroepitaxial film. To date, direct, non-destructive characterizations of the atomic-level structure of films and interfaces have not been readily available and this has hampered the design and optimization of heteroepitaxial devices. We describe here a novel x-ray interference method which is useful for imaging such structures with sub-Ångstrom spatial resolution while also providing chemical composition information from a map of the electron density. We illustrate the method, known as Coherent Bragg Rod Analysis (COBRA), with recent results on GaSb-InAs heterostructures of interest as infrared sources and detectors. We show that, with detailed knowledge of the interfaces from COBRA, it is now feasible to correlate specific molecular beam epitaxy growth conditions with desired electronic characteristics associated with the interface bonding. The COBRA method is quite general and only requires an epitaxial relationship between the substrate and the nanostructure that is deposited on it.