Extended X-ray absorption fine structure in dispersive mode

2014 It has been possible to disperse the X-ray beam provided by the storage ring DCI at LURE in order to perform fast EXAFS experiments (~ 1 s). The weak divergence of the synchrotron beam opens up the possibility of achieving a very good energy resolution. The first data collected with a solid-state detector array (RETICON) show that kinetic studies with a time scale lower than 1/10 s could be performed. J. Physique LETTRES 43 (1982) L-315 L-319 ler MAI 1982, Classification Physics Abstracts 78.70D The feasibility of measuring EXAFS in a dispersive mode rather than a time dependent mode tuning the photon energy, has been tested at the small angle scattering (S.A.S.) station at LURE. Using a triangle-shaped monochromator curved in order to disperse the energies of the reflected beams over 500 eV, and a resolutive film as X-ray detector, it has been possible to achieve a rapid EXAFS spectrum of copper. Two seconds are required (LURE-DCI operated at 1.72 GeV, 200 mA) for acquiring each spectrum on M-KODAK film which is known to be resolutive at the expense of sensitivity. Such a short exposure time, which can be further reduced using solidstate detector arrays (or even more sensitive film), will open up the possibility of kinetic studies performing time-resolved EXAFS experiments. Hence a considerable reduction in time required for data acquisition can be achieved, even for low concentration binary alloys such as Al-6.8 at. % Zn. This metastable solid solution which decomposes at room temperature within a 10-20 seconds time-scale will provide our first example of a kinetic study using this promising structural technique. Data have been collected whose analysis is in progress. Taking advantage of the short time required for this structural experiment, (*) Laboratoire de Metallurgie Physique, associe au C.N.R.S., Universite de Poitiers, 40, avenue du Recteur Pineau, 86000 Poitiers, France. (**) Laboratoire de Physique des Solides, associe au C.N.R.S. (LA 2), Batiment 510, Universite Paris-Sud, 91405 Orsay Cedex, France. (***) Laboratoire de Physique de la Matiere Condensee, associe au C.N.R.S., College de France, 11, place Marcellin Berthelot, 75231 Paris Cedex, France. Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyslet:01982004309031500 L-316 JOURNAL DE PHYSIQUE LETTRES it will be possible to investigate systems in which phase transitions occur in a very narrow domain of variation of the significant parameter (which is often difficult to stabilize at the right value), by scanning over a large domain and selecting the spectra of interest. 1. Monochromator. The monochromator of the S.A.S. station has provided a useful « test bench » for the first experiments and for working out the improvements required to install a station devoted to EXAFS experiments in a dispersive mode. The Ge 111 crystal suitably oriented with respect to the incident beam was bent by a motor-driven constraint at the apex affecting a regular curvature of the 0.5 mm-thick, 6 cm-long blade. Its asymmetric cut (6°) permits an image point close to the crystal calculated to be at 20 cm from the crystal while tolerating a large source to crystal distance ( ~ 15 m) and dependent on the X-ray range of interest [1]. The small curvature ( ~ 3 m) ensures that strains remain within the elastic domain. Behind the focal point X-ray beams of increasing energies are spatially dispersed since the glancing angles decrease continuously along the crystal surface working as a focusing optical element. In our experiment the change of Bragg angles, 50, is about 15 times the effective aperture of the white beam in the horizontal plane (5s). By measuring the intensity distribution across the beam direction behind the focus, with and without the sample, the absorption cross-section can be measured over a large range of photon energies. 2. Results. The 550 eV window of the copper spectrum (100 eV below the K-edge, 450 eV above) was displayed as a 9 cm trace on X-ray film. In figure 1, the EXAFS spectrum of Cu from data collected using a channel cut crystal to tune the photon energies is compared to the one obtained in a dispersive mode [2]. The very good superposition of the EXAFS oscillations in the whole energy range proves that the energy calibration can be achieved easily. In fact the Bragg angles of the elementary beams of increasing energies decrease linearly along the trace of the X-ray beam on the Ge blade. Since Bragg angles depend exclusively on geometry, the achievement or not of a regular curvature, does not affect the energy calibration in the case of a pin-point source of photons, which can be assumed in case here since 50 ~ 15 bE. Irregularities in the curvature of the crystal produce inhomogeneities in the intensity of the reflected beam as shown in the 10 curve on figure 2 where the high intensity on the high energy side of the spectrum is related to the absence of curvature near the apex of the triangular crystal. Fig. 1. ---EXAFS in dispersive mode compared to the EXAFS in conventional mode (20132013) ) for Cu. One motor step rotates the channel-cut by 1/1800 degree. L-317 EXAFS IN DISPERSIVE MODE Fig. 2. Transmitted intensities with and without Cu sheet with LURE DCI operated at 1.54 GeV, 200 mA, recorded on M-KODAK film : the first two peaks, 9 eV apart, are well resolved when the crystal curvature is reduced. The energy calibration is thus trivial even when the response of the crystal is highly disturbed by incorrect gluing as was the case during our first test of the photodiode array (Fig. 4a). The resolution is affected by the curvature of the crystal. However, we have to recall that the small divergence of the synchrotron beam permits good resolution to be maintained even if the curved crystal collects radiation from an appreciable arc segment of the orbit. The synchrotron source has to be thought as an array of elementary divergent sources, diverging in the plane of the orbit. Thus it is not surprising that the smearing of the first peak with a 9 eV separation in the Cu spectrum can be reproduced using a 5 eV-wide binomial window to convolute the spectrum of reference (Fig. 1). It is possible to improve this resolution by relaxing the curvature of the Ge Fig. 3. The spectra have been recorded for Al-6.8 at. % Zn air-quenched from 308 ~C to room temperature. The « b » spectrum was obtained at room temperature after 40 seconds annealing, the « a » spectrum after 11 minutes annealing. The spectra recorded between these two corroborate the change of the second and the third oscillations. L-318 JOURNAL DE PHYSIQUE LETTRES blade (Fig. 2). When only a 180 eV window is kept the separation of the first two peaks is retained. Even a vague hint of the well-known shoulder on the Cu threshold is observed. Increasing the size of the energy window requires only a longer crystal and a smaller angle of the asymmetrical cut. Hence with a weakly curved crystal, studies of X-ray absorption edges are possible with this experimental set-up. The main problem lies with the detection. This difficulty is not unique to dispersive EXAFS experiments, but depends on the availability of position sensitive counters for high photon flux (108-109 cps). Films are used currently to solve this problem. This procedure has been chosen for the Al-Zn study using a automatic microdensitometer (Fig. 3). But further developments require a faster treatment of the data which can be overcome only by new position-sensitive photon counters. Promising results have been obtained with photodiode arrays. These detectors have been used to record dispersive EXAFS spectra realized with the continuous radiation of X-ray tube operated at 14 kV with a current of 30 mA [5] and for S.A.S. experiments too [6]. Very recently we have made a preliminary experiment to test be RC 256 EC/17 array (with one micron-thick Si02 window, manufactured by RETICON) for the synchrotron flux emerging Fig. 4. a) Profiles of the beam with (1, -) and without (10, ...) a Ni sheet as directly observed by the photodiodes array. The intensity variations observed in 10 are due to irregular curvature obtained with the first test of a newly-assembled monochromator. b) The Ni-edge directly calculated from data of figure 4a, expressed as Log (7o/I) versus the number of the photodiodes. L-319 EXAFS IN DISPERSIVE MODE for an « almost flat » monochromator, « positioned across » the Ni-edge. Even with the highly modulated intensity of the incident beam it has been possible to observe the edge and the peaks separated by 9 eV (analogous to the Cu ones) using a very short exposure time : 1.3 ms to record data with the Ni sample and 0.32 ms to record the photocurrent without the sample (Figs. 4a, 4b). This curve which contains the main features of the Ni-K-edge has been obtained without dark current corrections. Just before the edge the even and odd readout electronics did not combine uniformly, thus producing a saw-toothed profile. We were unable to tune the separate readouts to uniformity since previous damage had affected this part of the detector. The main problem lies with the damage observed after four hours of exposure in the beam. This generates an increasing level of the dark current which reaches 20 % of the saturation at the end of the experiment. Since we did not go on till the break down of the photodiode array, we cannot give precise value of the dose accepted during its whole lifetime or even if deterioration saturates with dose. But, at the end of the experiment, each photodiode has received about 1012-1013 photons since each pixel saturates at 3.2 picocoulomb. The cause for the deterioration is not clear. It could be : 1. irreversible damage to photodiodes or 2. deterioration of the microcircuitry. The microcircuitry was partly protected by tapes of lead forming a slit, but it was difficult to adjust the slit along the photodiode array, which has only a 17 mil. aperture width. Even with this damaged detector it has been possible to acquire a good profile of the Ni-edge, by measuring the dark current for each exposure time and subtracting it from the data recorded with X-rays beam. But the deterioration continued. Nevertheless we believe the photodiode array technology is the way for X-ray detection in the near future. With further developments on intrinsic Ge crystal, and immediately with the SFX series manufactured by RETICON using phosphorus for wavelength conversion the problem of deterioration should be solved. Direct illumination may be safer from damage since the aperture width is 2.5 mm giving each sensor element a 100 : 1 aspect ratio suitable for a better beam localization on to the photodiodes. Within the 9 cm-long trace (narrowed to one millimeter height by the slits), recorded on the M-KODAK film the photo flux is estimated to reach 108-109 photons. Pumping out the photon channel (about 1.5 m) gives about five times as many photons. Thus we can estimate the flux to be a few times 109 photons/eV/s. 3. Conclusion. The importance of this structural tool is intimately related to the rapidity with which data can be collected. The possibility of time-resolved EXAFS experiment on a 20-100 ms time scale will be feasible with photodiode array technology. Further developments are related to the optical geometry, too. Focusing the beam of the new generation of storage rings may permit local analysis of the atomic environment of one given atom at the 10 or 20 ~ scale which should be a new probe of great interest in material engineering. We would like to thank S. Megtert, D. Lefur in charge of the port where the last part of these experiments were made, and Mme J. Nogues for her work to produce data from the films.