Project

Development of X-ray absorption and fluorescence microspectroscopic imaging using novel detection systems

Code
178WE1113
Duration
01 January 2013 → 31 December 2016
Funding
Regional and community funding: IWT/VLAIO
Research disciplines
  • Natural sciences
    • Analytical chemistry
  • Medical and health sciences
    • Pharmaceutical analysis and quality assurance
Keywords
x-ray absorption oxidation state atoms synchrotron
 
Project description

X-ray fluorescence (XRF) spectroscopy provides qualitative and quantitative information on the elemental composition of a sample by illuminating it with X-rays and monitoring the emitted characteristic radiation.1, 2 X-ray absorption spectroscopy (XAS) provides information on the local chemical structure of an element of interest in a sample by tuning the primary X-ray beam energy over an absorption edge of the element of interest, and monitoring the amount of absorbed X-rays.3, 4 Based on the energy position of the absorption edge, information is retrieved on the oxidation state whereas the energy dependent fluctuations provide information on the type, amount and distance of neighbouring atoms. Due to the requirement of beam energy tunability and high X-ray photon fluxes, most XAS experiments are performed at synchrotron radiation (SR) facilities. The most basic approach towards XAS is the bulk mode approach, making use of a wide X-ray beam. This approach only provides averaged information on the relatively large probed sample volume of several mm3. In order to differentiate between sample heterogeneities, a scanning approach can be performed in which the primary X-ray beam is focussed down to microscopic dimensions. The sample is then raster scanned through this beam, providing spatially resolved information. 3D resolved information can be achieved by applying a tomographic detection scheme. However, this approach demands much time, rendering it less ideal for most SR experiments due to the limited access and measurement time constraints. The confocal detection scheme partially resolves this issue as it allows for direct 3D information extraction from subvolumes in a sample. However, the investigation of large volumes is still limited due to the inherently long measurement times of XAS experiments on diluted samples. For this reason, full-field XAS was developed in which spatially resolved information is retrieved directly and simultaneously from a large sample volume. To achieve this, the SLcam was used providing information from a 2.1×2.1mm2 sample area with 8 μm spatial resolution in measurement times equivalent to the acquisition of a XANES spectrum each 620 ms following the default point-by-point scanning method. The several approaches towards emission mode XAS presented above are discussed, along with several applications displaying the strengths of each of the separate measurement approaches.

A.1 Materials and Methods

A.1.1 BM26A

Beamline BM26A or the DUBBLE beamline (Dutch-Belgian beamline) at the ESRF (Grenoble, France) provides an X-ray beam originating from a 0.4T bending magnet, rendered monochromatic using a Si(111) monochromator. Higher harmonics are rejected and vertically focussed by a Si and Pt coated mirror behind the monochromator.5 The primary and transmitted beam intensities are monitored using ionization chambers, filled with gasses to provide approximately 10% and 70% absorption before and after the sample position respectively.

A.1.2 P-06

The PETRA-III P-06 beamline at DESY (Hamburg, Germany) operates an undulator generating primary X-ray beam was rendered monochromatic using a Si(111) double crystal monochromator, providing an energy resolution dE/E of approximately 1.4·10−4. A Kirkpatrick-Baez (KB) mirror system was used to focus the beam to a vertically oriented sheet beam of approximately 1.8×0.007mm2 (V×H), with a horizontal divergence of approximately 2 mrad. No change in sheet beam position was perceived by scanning through the primary X-ray beam energy. A polycapillary X-ray optic based confocal setup was built at the BM26A beamline. Primary X-ray beam focussing occurs with an XOS (X-ray Optical Systems, NY, USA) designed focussing optic, characterised by a large acceptance window (7.5×3mm2), a focal distance of approximately 2.5mm and an estimated focal spot size of 10 μm at 6.4 keV. The 6 cm long polycapillary optic is contained in a housing with entrance window diameter of 10mm and exit window diameter of 2.8 mm. Due to the large acceptance window, a large fraction of the BM26A X-ray beam is guided through the optic, resulting in a flux as high as possible impinging the microscopic beam spot on the sample. By applying a second polycapillary optic to the Vortex-EM SDD detector (Hitachi High-Technologies Science America Inc.) and coinciding the focii of both optics, a confocal detection scheme is achieved. As confocal optic a 4 cm long XOS manufactured collimating optic with a 2mm diameter entrance window, 10mm diameter exit window and 2mm focal spot distance was used. Purging holes to flow He through the polycapillary optics, decreasing absorption of X-ray photons by air, are foreseen in the optic’ housing design. An image of the confocal setup is displayed in Figure A.1.

A.1.4 SLcam

The SLcam (Strüder-Langhoff camera, or Colour X-ray Camera) is a Si based pnCCD type energy dispersive X-ray detector, consisting of a 450 μm thick Si layer with 264 by 528 48×48 μm2 pixels, mutually developed by PNSensor GmbH (Munich, Germany), the Institute for Scientific Instruments GmbH (IFG, Berlin, Germany), the BAM Federal Institute for Materials Research and Testing (Berlin, Germany), and the Institut für Angewandte Photonik e.V. (IAP, Berlin, Germany).6 The outer 264 by 132 pixels of the chip are covered to prevent illumination by X-rays. The chip is read out with a 400 Hz frequency, in which a fast sideways transfer of the central 264 by 264 pixels to the sides, the so-called ‘ark frame storage mode’ is combined with the slower read out of the dark frames by two CAMEX chips each, effectively splitting the SLcam readout window in four quarters. Two polycapillary optics are available: a straight 1:1 magnification optic resulting in 48×48 μm2 spatial resolution and a conical 6:1 magnification optic for 8×8 μm2 spatial resolution (Figure A.2). The detector chip was cooled to -20 °C to reduce electronic noise and provide an energy resolution of 150 eV for the Mn-K fluorescence line energy.

A.2 Results and Discussion

In the bulk mode approach, the sample is illuminated by a ‘arge’X-ray beam with a size in the order of a few millimetre. The resulting absorption and fluorescence spectrum is then the average of the entire illuminated sample volume. Despite the current synchrotron radiation development trend towards smaller beam sizes, currently achieving sizes down to only a few nanometre,7 bulk XAS experiments remain very popular as they are the ideal approach when studying samples with millimetre sized homogeneity length scales. Due to the high flux attainable at synchrotron radiation facilities, bulk XANES experiments can be performed at a rate of less than a minute per spectrum, even down to a subsecond frequency, creating the opportunity of in-situ experiments such as following catalytic processes.8–1 Vanadium containing MOF catalysts were measured to investigate the chemical and structural changes around V during the breathing process of this MOF as well as the influence of the Al/V content of the catalyst on the V state.12 Additionally, a Au/TiO2 catalyst particle was investigated to determine the effects of different reduction synthesis methods on the Au oxidation state.13 As spatially resolved information was not required for these experiments, bulk mode XAS presented the ideal method due to the higher photon flux impending the samples. This resulted in shorter measurement times per sample, thus allowing for the investigation of more samples. Additionally, the fairly straightforward measurement setup allows for the implementation of specialized sample environments such as cryogenic conditions and He atmosphere. Micro-focal emission mode XAS experiments were performed at the ID21 and BM26A beamlines, investigating carbonized Herculaneum papyrus scrolls.14, 15 Here, spatially resolved information is of the essence as elemental composition images displayed the correlation between several elements, as well as showed the elemental constituents of the writing on the papyrus. Micro-XAS provided additional information on the compounds present in the papyrus writing, providing insights on the ink or pigment that was potentially used when drafting the manuscript. Due to the implementation of an X-ray optic in the primary beam path a slight decrease in sensitivity is observed due to the slightly reduced X-ray beam flux: 1-20 ppm for bulk mode and 2-35 ppm for micro-focal measurements with 1 s exposure time and atomic number Z from 19 to 29. However, the brilliance increases significantly, allowing for fast data acquisition. XAS images in which a XAS spectrum is recorded for each pixel of a predetermined array are however rarely applied due to the long measurement times involved in this method. It should be noted that due to the recent developments in large solid angle detectors this scanning method does become feasible, even when considering 3D tomographic detection schemes.16–1 To obtain 3D spatially resolved information in manageable measurement times, a confocal detection scheme can be applied. At the BM26A beamline information was extracted selectively from a 8×8×10 μm3 volume, with detection limits that are only a fraction higher than those obtained for micro-focal experiments (4-40 ppm). The confocal setup was applied for the investigation of Fe in chemically strengthened boroaluminosilicate glasses to monitor the change of Fe with sample depth and Al content.22 Additionally, natural diamonds with deep Earth inclusions were investigated to determine the Fe minerals contained in these inclusions as they provide a unique way in which deep Earth can be sampled directly. Nevertheless, scanning a 3D sample volume rapidly increases in measurement time with increasing array size using the confocal point-by-point measurement scheme. As a result, full-field XAS was applied making use of the SLcam, providing XRF and XAS spectra of 69696 pixels simultaneously. Although a single full-field XAS measurement easily takes over 10 hours of measurement time, it should be noted this corresponds to measuring a XAS spectrum every 620 ms for each point, including sample movement, in order to cover the same sample area. It is clear the SLcam presents a vast increase in 2D and 3D XAS data acquisition. An initial 2D full-field XAS experiment with the SLcam was performed at BM26A, investigating geological Nitisol soil samples and reference structures.23 A data reduction method based on a combination of PCA and K-means clustering was applied to combine similar XANES spectra, rendering the evaluation of nearly 70 thousand XAS spectra more manageable (Figure A.3). Due to the illumination of the sample with a broad beam which was spread out over the sample surface to illuminate an area nearly as large as the SLcam field of view, the detected information originates from varying depths with respect to the sample surface. As such, the information is spatially resolved in two dimensions, however the third dimension remains unknown. By illuminating the sample with a sheet beam with known width, a confocal detection scheme can be obtained using the SLcam, providing full-field confocal XAS measurements. This was displayed with a Au/MgO catalyst particle, measured at beamline P-06 (PETRA-III, DESY, Hamburg). A virtual 3D slice of the sample was made with 8×8×8 μm3 spatial resolution, which could be overlapped with transmission tomographic data acquired at a lab source (Figure A.4). Based on the extracted XANES curves the gold clusters in the sample could be identified as metallic.

A.3 Conclusions

Several approaches towards emission mode XAS data acquisition were discussed: bulk mode, focal, confocal and ultimately full-field and confocal full-field XAS. The methods were compared based on the investigation of different types of samples, including geological, catalytic and industrial applications. The applications presented in this work present insights in the respective fields of research providing advances in the synthesis or interpretation of previously obtained data, displaying XAS is a powerful tool to study the local chemical environment in a sample. Furthermore, it can be concluded that each XAS method is characterised by its own strengths and weaknesses. The discussed applications present samples and research questions appropriate for the corresponding methodologies. It is advised to consider the type of information one hopes to obtain from a sample before deciding on a XAS methodology: in-situ reactions may prove difficult to monitor with sufficient time resolution if one opts to perform full-field experiments, whilst investigating local hot spots in a larger sample matrix is not ideal with bulk mode XAS. This comparison of methodologies helps in determining the appropriate detection method for a given research question. Full-field emission mode XAS is shown to provide microscopic spatially resolved information of large sample volumes in previously unattainable measurement times, rendering this approach feasible for thick and diluted samples at synchrotron radiation facilities.