ALBERT

Notes: The synchrotron x-ray fluorescence (SXRF) microprobe has proven to be a valuable tool for trace element research. It permits analysis down to a few parts per million of many elements in a spot size of less than 10 μm. Existing SXRF microprobes are using energy dispersive detectors (EDS), either Si(Li) or intrinsic Ge diodes. Such detectors have the advantage of collecting the entire fluorescence spectrum at once. They can also be positioned to collect a relatively large solid angle. However, EDS detectors suffer from several significant problems: resolution at Fe Kα is about 150 eV, which is roughly 60 times the natural linewidth; the maximum count rate is less than 20 000 counts/s in the entire spectrum; there is significant low-energy background due to scattering and incomplete charge collection in the device. For geochemical analyses these limitations preclude trace element analyses in the presence of a large amount of a high atomic number element: for example, trace element studies of galena (PbS) and zircon (ZrSiO4), or measurements of Cr or Ti in minerals with more than a few percent Fe or Mn. The poor energy resolution prevents the measurement of small amounts of rare-earth elements in samples with significant concentrations of first-row transition elements. Wavelength dispersive spectrometers, based upon Bragg diffraction from a bent crystal, have several distinct advantages over EDS detectors. The resolution at Fe Kα is about 10 eV, or only 4 times the natural linewidth. This permits the analysis of rare-earth elements and also lowers the background which improves detection limits to the 0.1 ppm range.The WDS spectrometer only detects a single energy at once, so it is possible to measure trace elements in the presence of intense fluorescence of a major element. We have installed a commercial wavelength dispersive spectrometer (model WDX-3PC from Microspec Corp., Fremont, CA) on the X-26A microprobe beamline at the NSLS. The spectrometer can scan the range from 33° to 135° 2θ. It contains four analyzing crystals (TAP, PET, LiF200, LiF220) mounted on a motor-driven turret, which cover the energy range from 1 to 17 keV. The detector is equipped with tandem proportional counters: a thin-window flow counter (P-10 gas) followed by a Be-windowed sealed Xe counter. A remotely adjustable exit slit is located just before the flow counter. This slit can be used to trade off count rate for energy resolution. Measured resolution at Fe Kα is 11 eV. The peak/background ratio on Fe metal is 105, which is roughly 100 times better than with a Si(Li) detector. The measured collection efficiency varies from roughly 10−3 to 10−4, which is a factor of 3–10 lower than that for the Si(Li) detector as it is normally used at X-26A. The X-26A microprobe has been configured to allow simultaneous use of both the WDS and Si(Li) detector. The detectors complement each other nicely, with the Si(Li) providing an overview of the entire spectrum and the WDS available to study selected peaks with significantly better energy resolution and sensitivity. © 1995 American Institute of Physics.

Notes: Multielement Ge and Si(Li) detectors have been used in recent years to improve the increase count rate capability and to improve the solid-angle efficiency in fluorescence x-ray absorption spectroscopy (XAS). Such systems have typically been equipped with one or more single-channel analyzers (SCAs) for each detector element. Such SCA-based electronics are sufficient when only the counts in one or two well-resolved peaks are of interest. For the fluorescence (XRF) microprobe at beamline X-26A at the NSLS, SCA-based electronics were not a satisfactory solution for two reasons: (1) for XRF experiments, the entire fluorescence spectrum is required; (2) for micro-XAS studies of trace elements in complex systems, the fluorescence peak often sits on a significant background or partially overlaps another fluorescence peak, requiring software background subtraction or peak deconvolution. An electronics system which permits collection of the entire fluorescence spectrum from each detector element has been designed. The system is made cost-effective by the use of analog multiplexors, reducing the number of analog-to-digital converters (ADCs) and multichannel analyzers (MCAs) required. The system was manufactured by Canberra Industries and consists of: (1) a 13 element Ge detector (11 mm diameter detector elements), (2) 13 NIM spectroscopy amplifiers with programmable gains, (3) four analog multiplexors with maximum of eight inputs each, (4) four ADCs with programmable offsets and gains and 800 ns conversion time, and (5) two MCAs with Ethernet communications ports and two ADC inputs each.The amplifiers have shaping times which are adjustable from 0.5 to 12 μs. The analog multiplexors were modified to perform pileup rejection. The analog multiplexing does not significantly reduce the count rate capability of the system, even at the shortest amplifier shaping times. The average detector resolution is 170 eV at 12 μs shaping time and 200 eV at 4 μs shaping time. The maximum aggregate count rate is 400 kHz with 0.5 μs shaping time. The system is controlled by software based upon a package from Canberra and another commercial package (IDL), both running on a VAXstation 4000/90. The software automatically adjusts the gains of the amplifiers and offsets of the ADCs so that the spectra from each detector have identical calibrations and can be added channel for channel. The overhead to read a 1024 channel spectrum from each of the 13 elements and sum them is about 2 s. The software allows a range of options for data storage, from saving the complete spectrum for each of the 13 detectors elements ((approximately-greater-than)50 000 bytes/point) to saving only the net counts under a single fluorescence peak summed over all the detector elements (4 bytes/point). These data can be stored at each pixel in an elemental map or at each point in a monochromator scan. The system has been commissioned and is being used for XRF and micro-XAS studies. © 1995 American Institute of Physics.

Notes: The melting behavior of indium at high pressure has been studied in an externally heated diamond anvil cell (DAC) using x-ray diffraction measurements. Melting at high pressure was identified by the appearance of diffuse scattering from the melt with the simultaneous disappearance of crystalline diffraction signals. The observed melting curve is in good agreement with previous determinations based on resistivity measurements in a piston cylinder apparatus. These results demonstrate the successful melting experiments in a DAC using the x-ray diffuse scattering as the melting criterion. © 2001 American Institute of Physics.

Notes: We describe a laser heated diamond anvil cell system at the GeoSoilEnviroCARS sector at the Advanced Photon Source. The system can be used for in situ x-ray measurements at simultaneously ultrahigh pressures (to 〉150 GPa) and ultrahigh temperatures (to 〉4000 K). Design goals of the laser heating system include generation of a large heating volume compared to the x-ray beam size, minimization of the sample temperature gradients both radially and axially in the diamond anvil cell, and maximization of heating stability. The system is based on double-sided laser heating technique and consists of two Nd:YLF lasers with one operating in TEM00 mode and the other in TEM01* mode, optics to heat the sample from both sides, and two spectroradiometric systems for temperature measurements on both sides. When combined with an x-ray microbeam (3–10 μm) technique, a temperature variation of less than 50 K can be achieved within an x-ray sampled region for longer than 10 min. The system has been used to obtain in situ structural data and high temperature equations of state on metals, oxides, and silicates to 3500 K and 160 GPa. © 2001 American Institute of Physics.