MICRO XRF OF LIGHT ELEMENTS USING A POLYCAPILLARY LENS AND AN ULTRA THIN WINDOW SILICON DRIFT DETECTOR INSIDE A VACUUM CHAMBER

Transcription

1 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume MICRO XRF OF LIGHT ELEMENTS USING A POLYCAPILLARY LENS AND AN ULTRA THIN WINDOW SILICON DRIFT DETECTOR INSIDE A VACUUM CHAMBER P. Wobrauschek, B. Frank, N. Zoeger, C. Streli, N. Cernohlawek, C. Jokubonis, and H. Hoefler ABSTRACT Atominstitut, TU-Wien, A-1020 Vienna, Austria Since most micro-spectrometers are operated in air, up to now only mapping of medium and higher Z elements was possible with micro X-ray fluorescence analysis (µ-xrf). Though for some applications (e.g.: biological- and geological samples, samples from art and archeology) the extension to light elements is of great interest. Therefore µ-xrf measurements using a polycapillary lens mounted inside a vacuum chamber have been performed. X-rays from a 2kW high power diffraction tubes (Mo and Cr anode) as well as from a low power 50 W Pd-anode tube were focused by means of the polycapillary X-ray optics. The polycapillary adjustment was performed with a combination of an X/Y stage and a gimbal allowing 3 translation and 2 rotation movements. With this setup it was possible to produce an X-ray spot of about 40µm in diameter at the sample surface. Sample motion is performed by an x-y-z-micrometer driven stage PC controlled. The position dependent fluorescence signal was collected by a Peltier cooled energy dispersive detector (10 mm2 Silicon Drift type) with an ultrathin window (300nm) allowing the detection down to oxygen. Scanning measurements have been carried out on various samples of human femural bone. Results from an area scan (500 x 1100 µm 2 ) across the transition zone of trabecular- and cortical bone are presented, including the distribution of O, P, Ca, and Zn. To compare the performance of the different anode X-ray tubes single point measurements on NIST glass (soda lime glass) have been performed with regard to light element detection. A comparison of results using the different excitation conditions of Mo-L and Pd-L on various samples will be presented with excellent conditions regarding excitation, detection, and background using a vacuum chamber for µ-xrf. INTRODUCTION Micro X-ray fluorescence analysis (µ-xrf) has nowadays become a well established method to nondestructively determine elemental distributions in various types of samples with a spatial resolution of about 10-30µm [1-4]. Many laboratory and commercial systems have been developed, using either monocapillary or polycapillary X-ray optics to produce an intense microbeam at the sample surface. Since most of the micro-spectrometers are operated in air the fluorescence radiation with lower energy is absorbed on the way to the detector and the detectable elemental range is therefore practically limited to elements above S (E Kα = 2.3keV). To overcome this problem and to have the possibility to measure the distribution of low-z elements in the sample a microbeam vacuum chamber has been designed [5]. The main goal was to reduce absorption of the fluorescence radiation between sample and detector as well as to reduce the scattering in air. The spectral background is minimized, the Ar- K line practically removed, and optimized conditions for the detection of light elements are achieved. A further

2 This document was presented at the Denver X-ray Conference (DXC) on Applications of X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD). This document is provided by ICDD in cooperation with the authors and presenters of the DXC for the express purpose of educating the scientific community. All copyrights for the document are retained by ICDD. Usage is restricted for the purposes of education and scientific research. DXC Website ICDD Website -

3 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume advantage arises when using capillary optics as focusing devices, which is based on total reflection of X-rays. When X-rays propagate through a glass channel, total external reflection on the capillary walls occurs. Due to the energy dependence of the critical angle for total reflection, higher energy photons are absorbed in the capillary. Only X-rays with lower energy than high energy cut off energy are reflected towards the exit of the device (cut-off effect), leading to efficient excitation of light elements. In addition the background in the low energy region is improved as the Compton edge resulting from backscattered high energy photons from the detector itself is missing as the high energy photons are not present. EXPERIMENTAL The existing vacuum chamber described earlier has been improved [5]. A capillary adjustment unit has been installed inside the chamber, where the combination of an X/Y stage and a gimbal mount allows to perform a 5-axis movement (3 translations, 2 rotations). By iteratively changing the five degrees of freedom and controlling the focused beam either optically (ZnS screen + microscope) or by successively scanning a Cu or W wire of 10µm thickness across the beam an optimal alignment of the capillary was reached. For beam focusing a monolithic polycapillary optics ( ) from XOS [6] with an entrance focal lengths of 53mm, an exit focal length of 13.5mm and a length of 62mm was used. For optically controlling the measurement position a microscope with long working distance of 165mmm (Olympus) and a CCD camera attached to it is placed outside the vacuum chamber looking through an adequate sized borehole into the chamber on the sample. To perform automatic scanning of the sample relative to the incoming beam the samples were mounted on an XYZ sample stage with 3 motorized, micrometer driven, translation stages. An ORIEL controller and a self developed software allows communication between MCA, motion control, and micrometers. A stepsize from 1µm upwards can be choosen and 25mm x 25 mm horizontal x vertical travel is possible. The fluorescence photons were detected by a KETEK silicon drift detector [7] with an active area of 10mm 2, a thickness of 300 µm thickness, and a resolution of 140eV kev, equipped with an ultrathin window of 300nm thickness (Moxtek, AP1.3), which offers a transmission of 60% for the Oxygen-Kα line (523 ev). The angles between beam-sample and sample-detector were fixed to 45 each. Two different tubes were used: an OXFORD [8] low power tube with Pd anode (50 W) with a focal spot size of 150 x 70 µm and a 2kW diffraction tube with Mo target, focal spot 8 x 0.4 µm (point focus). The setup with the low power tube attached is shown in Figure1.

4 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume Figure 1. µ-xrf vacuum chamber: schematic view (a), overview (b), polycapillary, adjustment- and sample stages (c) and detail view (d). RESULTS The respective focal spots obtained after careful adjustment of the polycapillary with the 2 different tubes have been determined by scanning a 10 µm Cu wire in the focal plane of the capillary. The beam size was determined to be 42 µm for the low power Pd tube and 41 µm for the Mo diffraction tube (Figure 2). The calculation of the beam size respects all correction effects due to the measuring geometry under 45 degree motion versus the incident beam direction.

5 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume Cu 10µm wire Intensity [cts] Pd FWHM 43µm Beamsize 42µm Mo FWHM 42µm Beamsize 41µm Pd low Power Mo Intensity [cts] motor position [µm] 0 Figure 2. Scan of a 10 µm Cu wire. The beamsize was determined to be 42 µm for the Pd low power tube (blue) and 41 µm for the Mo anode diffraction tube (red). For better representation the 0 position has been slightly changed for the ordinates left (Pd) and right (Mo). A performance comparison of the two X-ray tubes was done by single point measurements on NIST SRM621, soda lime glass. Figure 3 shows the spectrum obtained with the low power tube. A comparison of the Ca net counts and the operating conditions for each experiment are shown in Table 1. Table 1. NIST SRM621, soda lime glass, measured with the 2 X-ray tubes. tube kv ma Ca net counts in 1000s Pd Mo The obtained result is about the same, showing that the low power tube provides practically the same photon flux through the polycapillary as the high power tube. Although a higher electron current can be applied to the diffraction tube (due to better heat dissipation of the larger focal area), the point to point X-ray optical properties of the capillary accepts the same fraction (40µm) of the focal spot on the anode for both tubes. Figure 4 shows a single spectrum from the scan across the transition between trabecular and cortical bone in a slice of human hip head performed with the Pd low power tube. Since the Ca- Ka escape peak of the Si detector has an energy of (E( Ca-escape) = 1.95 kev) overlaps partly with the P-K line (E (P-Kα) = 2.01 kev) special excitation conditions had to be chosen for this scan. The tube voltage was set to 15 kv to excite the Pd-L lines (E(Lα) = 2.84 kev, E(Lβ) = 2.99 kev) efficiently from the anode material. The Pd-K lines (E(Kα) = kev, E(Kβ) = kev) are not present in the exciting X-ray spectrum. Thus the excitation of Ca (only excited by the bremsstrahlung) was reduced, while P was efficiently excited by the Pd-L lines. By operating

6 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume the system at this so called selective excitation conditions the Ca escape peak could be almost suppressed (5 per mill of the Ca intensity) and it was possible to obtain reliable data for P. Figure 3. Spectrum of NIST SRM 621 soda lime glass, excited with the low power tube. Figure 4. spectrum of trabecular bone, excited with the Pd tube at 15 kv. Results from the whole scan of a 200µm thick bone slice across the transition zone articular cartilage-cortical/trabecular bone are shown in Figure 5. The operation conditions of the tube were 15kV/0.9mA and a measuring time of 80s per pixel was chosen to achieve data with good counting statistics. Element maps of O, P, S, Cl, and Ca as well as the calculated intensity ratio between Ca and P are displayed. From the maps of Ca and P one can see the structure of the bone analyzed, showing clearly the region of cortical bone (horizontal area of high intensities in the middle of the map) with adjacent trabecular bone. S can be regarded as a marker for articular cartilage and was therefore only detected outside of the bone substance. Although Oxygen should be more or less equally distributed over the entire area higher O intensities were detected in the

7 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume region of articular cartilage. This is due to the fact that O is much more absorbed in the bone matrix (mostly Ca) than it is in the matrix of cartilage (no Ca). Figure 5. Area scan of 200µm bone slice. In the transition zone cortical/trabecular bone Pd 15kV 0.9mA 80s per point, Scanned area: 23 x 27 pixels, 1150µm x 1350µm, Distance between pixels: 50µm, measuring time per pixel: 80s. CONCLUSIONS Inside the chamber technical improvements were made. The adjustment unit for the monolithic polycapillary has been installed successfully and reduced the focal spot from 69 µm to 42 µm. A comparison of a low power Pd anode tube with the Mo anode diffraction tube showed almost same intensity of Ca, though the power was a factor of 10 higher for the diffraction tube. Microanalysis of low-z elements is possible when using an SDD with ultrathin window. Using selective excitation, the Ca/P ratio in bone can be determined. ACKNOWLEDGEMENTS The authors would like to acknowledge the Physics, Chemistry, Instrumentation Laboratory, the IAEA Laboratories Seibersdorf for providing the low power X-ray tube. REFERENCES: [1] Kumakhov, SPIE 2859, 116 (1996) [2] Hoffman, Thiel, Bilderback, Nucl. Instr. Meth. A347,384 (1994) [3] B. Vekemans, K. Janssens,G. Vittiglio, F. Adams, L. Andong, Y. Yiming, JCPDS, (1999), [4] G.Havrilla, Adv.Ax-ray Anal.41, 234 (1999)

8 Copyright JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume [5] C. Streli, N. Marosi, P. Wobrauschek, B. Frank, The Rigaku Journal 20(2), (2003) [6] [7] [8]