Depth-Resolved X-ray Absorption Spectroscopy by Means of Grazing

Article pubs.acs.org/ac

Depth-Resolved X‑ray Absorption Spectroscopy by Means of Grazing Emission X‑ray Fluorescence Yves Kayser,*,† Jacinto Sá,‡,¶ and Jakub Szlachetko†,§ †

Paul Scherrer Institut, 5232 Villigen-PSI, Switzerland University of Uppsala, Ångström Laboratory, Department of Chemistry, 751 05 Uppsala, Sweden ¶ Polish Academy of Sciences, Institute of Physical Chemistry, 01-224 Warsaw, Poland § Jan Kochanowski University, Institute of Physics, 25-406 Kielce, Poland ‡

ABSTRACT: Grazing emission X-ray fluorescence (GEXRF) is well suited for nondestructive elemental-sensitive depthprofiling measurements on samples with nanometer-sized features. By varying the grazing emission angle under which the X-ray fluorescence signal is detected, the probed depth range can be tuned from a few to several hundred nanometers. The dependence of the XRF intensity on the grazing emission angle can be assessed in a sequence of measurements or in a scanning-free approach using a position-sensitive area detector. Hereafter, we will show that the combination of scanning-free GEXRF and fluorescence detected X-ray absorption spectroscopy (XAS) allows for depth-resolved chemical speciation measurements with nanometer-scale accuracy. While the conventional grazing emission geometry is advantageous to minimize self-absorption effects, the use of a scanning-free setup makes the sequential scanning of the grazing emission angles obsolete and paves the way toward time-resolved depth-sensitive XAS measurements. The presented experimental approach was applied to study the surface oxidation of an Fe layer on the top of bulk Si and of a Ge bulk sample. Thanks to the penetrating properties and the insensitivity toward the electric conduction properties of the incident and emitted X-rays, the presented experimental approach is well suited for in situ sample surface studies in the nanometer regime. razing X-ray fluorescence (XRF) techniques allow for nondestructive surface-sensitive measurements on length scales ranging from few to several hundred nanometers in the direction normal to the sample surface. The probed depth range, which extends from the surface interface of the sample into the sample, is usually defined by the extinction depth (i.e., the depth after which the intensity of the X-rays propagating at a grazing angle is attenuated by a factor e−1)1 and can be tuned by 2−3 orders of magnitude either by varying the grazing incidence angle of the X-rays used to excite the XRF radiation or by varying the grazing emission angle under which the emitted photons are detected. The respective two geometrical arrangements are called grazing incidence XRF (GIXRF)2−6 and grazing emission XRF (GEXRF).7−10 The extinction depth is restricted in the angular range below the critical angle of total external reflection to the first few nanometers below the sample surface by the creation of an evanescent wave.11 Any contribution from the bulk volume is suppressed for grazing angles smaller than the critical angle. In the angular range above the critical angle, the surface sensitive character of grazing XRF measurements is due to the inherently long incidence (GIXRF) or emission paths (GEXRF). Depth ranges of a few tens of nanometers and more can be probed. Both, GEXRF and GIXRF, are equivalent from a physical point of view through the principle of microscopic reversibility and reciprocity.11

G

© 2015 American Chemical Society

From an experimental point of view, in GEXRF, only the XRF signal originating from the near-surface region can be detected whereas in GIXRF the excitation of the XRF radiation is limited to the near-surface region resulting in general in a more efficient use of the incident X-ray photons. For simple elemental studies a measurement at a single grazing angle is sufficient and the grazing geometries allow one to probe minute sample amounts at improved signal-to-background ratios. The study of the elemental distribution in the direction perpendicular to the sample surface requires, however, an angle-resolved measurement of the XRF intensity. Both, GEXRF and GIXRF, have found applications in trace-element detection and micro analysis,12−17 thin-film characterization,7,18−21 the analysis of nanoparticulate systems on the top of a substrate,22−29 and depth profiling of ion-implanted dopant distributions.30−35 The probed angle-dependent depth ranges are suitable to bridge the gap between probing techniques based on electrons or soft X-rays, which cover the first few nanometers below the surface, and standard X-ray-based techniques, which explore micrometer ranges. Thus, GEXRF and GIXRF are proven to be Received: May 13, 2015 Accepted: October 12, 2015 Published: October 12, 2015 10815

DOI: 10.1021/acs.analchem.5b03346 Anal. Chem. 2015, 87, 10815−10821

Article

Analytical Chemistry

Figure 1. The dependence of the emitted XRF intensity on the grazing emission angle was measured in a scanning-free approach at a single sample− detector orientation by dispersing the angular scale along a position-sensitive detector positioned at 1 m from the sample (a). The grazing emission angles are discriminated by the individual pixel rows in the direction perpendicular to the substrate surface while the XRF signal was excited by means of a synchrotron radiation beam whose energy was scanned around an absorption edge. This allowed one to measure in a one-dimensional sequence of measurements the angle-dependent and incident photon-energy dependent XRF intensity, e.g., for a Ge bulk sample (b). From this measured map, the GEXRF profiles for different incident energies can be extracted with vertical cuts (c) as well as angle-dependent, and thus depthdependent, XAS spectra by considering horizontal cuts (d).

contrast to the grazing incidence and conventional XRF detection geometries, the detected XRF yield depends in GEXRF directly on the absorption coefficient of the studied element.36 Hence, in the grazing emission geometry, the probed sample volume can be, independent of the incident Xray photon energy and grazing emission angle, considered as thin for the incident X-ray photons and the number of atoms contributing to the detected XRF yield does not vary with the incidence energy. In GIXRF and standard XRF measurements, self-absorption effects have to be expected in the measured XAS spectra38−41 although these might be corrected for (see for example, refs 42 and 43). The insensitivity of fluorescence detected XAS measurements realized in the grazing emission geometry to self-absorption effects was experimentally confirmed by transmission XAS measurements on thin samples36 and in comparative XAS measurements in the GIXRF and GEXRF geometries on dried residues.44 With increasing sample mass within the dried residue, the white line intensity in the XAS spectra was attenuated in the grazing incidence geometry but not in the grazing emission geometry.44 Observed changes in the features of the XAS spectra recorded in the GEXRF configuration can be rather attributed to the sample inhomogeneity and the comparison between unfocused and focused incident X-rays (enhancing the sensitivity to the sample

complementary to the existing nanoanalytical techniques for the investigation of the elemental distribution within a sample, i.e., the sample structure. Since GEXRF and GIXRF are X-ray probe-based techniques, an ultrahigh-vacuum environment is not necessarily required and the samples can be investigated under in situ conditions. Insulating materials can be explored as well since charging effects do not occur. Moreover, GEXRF and GIXRF can be combined with routes offering chemical sensitivity, e.g., fluorescence detected X-ray absorption spectroscopy (XAS), which allows one to probe the local and electronic environment of specific elements. Together with the angle-dependent probed depth range in the near-surface sample region, GEXRF and GIXRF allow for nondestructive depthresolved absorption measurements.36,37 Hereafter, we will discuss the advantages offered by combining fluorescence detected XAS with scanning-free GEXRF and present depthresolved chemical speciation measurements in the nanometer range.

■

MOTIVATION In the conventional GEXRF geometry, the probing X-rays are perpendicularly incident on the sample surface which implies that the extinction depth of the XRF photons is even for grazing emission angles above the critical angle smaller than the attenuation length of the incident X-ray beam. Consequently, in 10816


Article


(1:1 focusing). The GEXRF intensity profiles were acquired by dispersing the grazing emission angular scale along one of the dimensions of a two-dimensional position-sensitive detector, a PILATUS 100 K area detector51 positioned at 1 m from the sample (Figure 1a). Regarding, the detector pixel size Δh of 172 μm for the PILATUS detector, this sample-to-detectordistance D fulfills the requirement that the opening angle per pixel (φ ≈ Δh/D), and thus the accuracy with which the grazing emission angles in a detector column can be discriminated, is better than 0.01°. The position-dependent detection of the XRF yield in the vertical direction to the sample surface allowed one to measure the GEXRF intensity profile in a single acquisition on an angular range only limited by the detector extension. Moreover, for a sufficiently large D, the variation of the grazing emission angle over the detector dimension perpendicular to the angular dispersion direction becomes smaller than the difference between two neighboring pixels in the dispersion direction. For the used setup, this is the case for D > 0.8 m.49 Finally, this allowed one to sum together the XRF intensities recorded in the pixels of a horizontal row in order to significantly improve the detection statistics in the GEXRF and XAS intensity profiles at the expense of a slightly broader angular resolution than what can be discriminated by the individual pixels. For the present setup, the experimentally retrieved angular resolution for the setup was 0.015°.49 This projection of all detector columns on the angular-dispersion axis allowed one to acquire the GEXRF intensity profile in a single acquisition at a fixed incident photon energy above the relevant elemental absorption edge and thus to characterize the elemental distribution in the depth direction of the sample. The XAS measurements were then straightforwardly realized by scanning the incident photon energy around an absorption edge of interest and monitoring at each incident energy the corresponding XRF yield for the different grazing emission angles in a single acquisition. This allowed one to assess from a one-dimensional scan the dependence of the detected XRF intensity on the incident photon energy and the grazing emission angle simultaneously as shown in Figure 1b for the example of bulk Ge. From the recorded intensity map, the GEXRF intensity profiles for different incidence energies or the fluorescence detected XAS spectra for different grazing emission angles, thus different probed depth ranges, could readily be obtained by extracting the corresponding columns (Figure 1c) and rows (Figure 1d). Unlike to conventional GEXRF setups where the probing Xray beam is incident at 90°, a shallow angle of about 2.8° was selected for the incident photon beam. This choice was motivated by a more efficient excitation of the XRF signal in the near-surface region (about a factor 20), thus contributing to the overall efficiency of the setup and compensating for the transmission losses in air between the sample and the detector. At the same time, the advantages offered by the grazing emission geometry for XAS measurements were preserved since only the near-surface region was investigated (sample depths of a hundred nanometer maximum). The information for grazing emission angles with larger extinction depths was not considered in order to ensure that self-absorption could still be safely neglected. The penetration depth of the incident photon beam was significantly larger than the investigated depth region, and the probed depth region was not affected by the change in the incident photon beam energy. In the following, the potential offered by combining the scanning-free GEXRF and fluorescence detected XAS techni-

inhomogeneity). These factors can be neglected hereafter since homogeneous samples will be investigated. Furthermore, the probed depth range in a conventional GEXRF setup with a large incidence angle for the probing Xrays depends solely on the energy of the monitored X-ray emission signal and not on the energy and incidence angle of the probing X-ray photons. Thus, the probed depth range for each grazing emission angle is constant during an incident photon energy scan, and depth-resolved fluorescence detected XAS measurements can be straightforwardly realized under grazing emission conditions by varying the incident photon energy and grazing emission angle.36 In contrast, the extinction depth of the incident X-rays in the GIXRF geometry varies pronouncedly while the photon energy is tuned around an elemental absorption edge. In order to perform XAS measurements under grazing incidence conditions with an approximately constant penetration depth for the incident X-ray beam, the incidence angle has thus to be varied together with the incident photon energy.37,45 The appropriate tuning of the grazing incidence angle, however, deteriorates the depth resolution and requires moreover an accurate knowledge of the energy-dependent optical constants of the sample material. In this view, resonant Raman spectroscopy, which uses a constant incident energy, has been proposed in combination with GIXRF for depth-resolved chemical speciation.46 If only the first few nanometers below the surface are of interest, another possibility is to select a grazing incidence angle below the smallest critical angle in the course of the energy scan in order to fulfill for each incident energy the total external reflection condition and thus to confine the sample volume in which the XRF signal is excited to the surface itself.47,48 Finally, while in conventional GIXRF and GEXRF setups the XRF intensity dependence on the grazing angle is acquired in a sequential manner, only the grazing emission geometry allows for a scanning-free acquisition of the angular XRF intensity profile by using a position-sensitive detector: the grazing emission angle scale is dispersed along one of the detector dimensions and the different grazing emission angles are discriminated by the individual detector pixels.49 This allows one to alleviate the most common drawback of GEXRF setups which is the small solid angle of detection and paves the way toward time-resolved studies. Indeed, in combination with XAS measurements, the simultaneous detection of the XRF yield for different grazing emission angles allows one to decrease the dimensions which need to be scanned sequentially to the incident energy solely. The XAS measurement is then performed at once for all the grazing emission angles covered by the setup, and by discriminating the fluorescence yield for the different emission angles, i.e., extinction depths, a depthresolved XAS experiment is realized. This can be performed without having a priori knowledge on the sample structure since this information can be deduced from the shape of the XRF dependence on the grazing emission angle for incidence energies above the elemental absorption edge.

■

EXPERIMENTAL SECTION AND ANALYSIS The scanning-free GEXRF setup used for the fluorescence detected XAS measurements under grazing emission conditions is described in ref 49 and was realized at the X05DA Optics beamline50 of the Swiss Light Source (SLS). The incident photon beam was delivered by a bending magnet, monochromatized by a cryogenically cooled Si (111) channel cut monochromator, and focused by a bendable toroidal mirror 10817


Article


Figure 2. Experimental data for the grazing emission angle-dependent XAS spectra at the K edge of an Fe layer on the top of Si (left panel). From the angular dependence of the X-ray emission signal above the Fe K edge, the layer thickness was fitted to 28.6 ± 0.6 nm. The different emission angles correspond to different probed sample depths (right panel). The black solid line corresponds to the extinction depth, i.e., the measure for the probed depth region.

Figure 3. Normalized XAS spectra around the Fe K edge for the Fe layer on the Si sample (left panel) and around the Ge K edge for the Ge wafer (middle panel) obtained from Figures 2 and 1b, respectively. For the Fe on the Si sample, the XAS spectra of grazing emission angle intervals of comparable extinction depths had to be binned together to reduce the statistical noise. For the Ge sample, this proved to be unnecessary and the XAS spectra measured for each monitored grazing emission angle could be used in the analysis. Only selected Ge XAS spectra are displayed. Exemplary fits for one probed depth range restricted to the vicinity of the surface are shown on the right upper and lower panels for the Fe layer on the top of Si (top right panel) and the Ge wafer sample (bottom right panel), together with the relative contribution of the two components used in the fits.

ques is demonstrated using a Ge ⟨100⟩ bulk substrate (Figure 1b) and a Fe layer on the top of a Si substrate (Figure 2). For an increased detection efficiency and improved data statistics, the XAS spectra recorded at grazing emission angles for which the extinction depth was identical within a few nanometers were binned together at the expense of the depth resolution for the Fe layer on the top of Si (Figure 3, left panel). For the Ge bulk sample, this proved to be unnecessary and the XAS spectra of the individual monitored grazing emission angles could be used. Indeed, the acquisition time per energy point was 70 s for the Fe on the Si sample and 240 s for the Ge bulk sample. Both samples were stored in ambient conditions and not subject to any chemical treatment before the measurements. From the GEXRF intensity profile, the Fe layer thickness was found to be 28.6 ± 0.6 nm by following the approach described in ref 49. The scanned photon energy range was 7098−7148 eV for the Fe layer sample (K absorption edge at 7112 eV) and 11 057− 11 137 eV for the Ge bulk sample (K absorption edge at 11 103 eV). The step size in the energy scans was 1 eV. For both samples, the different characteristic emission lines accompany-

ing an electron decay to the created K hole were monitored. Indeed, the PILATUS area detector is a single photon counting detector which allows one only to set a threshold with respect to lower photon energies, e.g., to reject the XRF signal from Si, but does not allow one to discriminate the individual contributions of the different XRF lines. The relative contributions of the different characteristic emission lines of an element are, however, known from tabulated values.52

■

RESULTS For the Fe layer on Si and the Ge bulk, the surface oxidation, meaning the natural oxidation of the samples stored and measured under ambient conditions, as a function of the depth will be extracted from the recorded XAS spectra. Understanding the factors underpinning surface oxidation chemistry and the affected depth region is important in fields like interface science, energy generation, environmental sciences, catalysis, metallurgy, and semiconductor applications since these provide valuable clues about catalyst activation/deactivation, metal corrosion, or surface passivation processes necessary to improve 10818


Article

Analytical Chemistry material performance and lifetime. In particular, the Ge surface oxidation is of interest53 in view of using Ge as a substitute for Si in industrial semiconductor applications in order to profit from its attractive electrical and chemical properties (ability to deposit high-k gate dielectrics, high mobility of electrons and holes).54−56 The full characterization of a sample requires measuring the depth profiles of the chemical composition and state with an enhanced sensitivity to the narrow surface interfacial region compared to the bulk phase below it. The normalized experimental XAS spectra from the Fe layer on Si and the Ge bulk (Figure 3, left and middle panel) were fitted with normalized reference spectra provided in the literature of Fe and Fe2O357 and of bulk crystalline Ge and GeO2 foils58 in order to extract the depth-dependent surface oxidation of the respective samples. Only the native oxide compounds of Fe and Ge were considered while the experimental XAS spectra for which a minimal contribution from the oxide was expected were used to adjust the incidence energy scale of the reference spectra. A linear combination of the reference spectra (Iref 1(E) and Iref 2(E) for the pure element and for the corresponding native oxide compound) was used to fit the experimentally acquired XAS profiles (Iexp(E) = a1 × Iref 1(E) + a2 × Iref 2(E), a1 and a2 being the fitting parameters which represent the relative concentration) in order to estimate the contribution of each component as a function of the probed sample depth (see Figure 3, right panels) and thus to deduce the oxygen concentration as a function of the sample depth (Figure 4). The only condition in the fitting procedure was to

z = λ /(4π Im( n2 − cos2φ ))

(1)

with λ being the fluorescence wavelength, n the complex refractive index of the investigated sample matrix for the fluorescence wavelength, Im the imaginary part, and φ the grazing emission angle. For the presented measurements, the attenuation of the incident beam had to be considered as well, given the shallow incidence angle of 2.8° used in the experiment. In order to estimate the largest depth for which self-absorption effects could be safely neglected, an incident photon energy just above the absorption edge was considered for the calculation until which depth a factor of 1 − e−1 of the incident photons was transmitted (in analogy to the definition of the extinction depth fro the XRF signal). This approach is intended to ensure that the probed depth range is defined essentially by the extinction depth of the monitored XRF signal and that the variation of the excitation rate of the XRF signal within the extinction depth is, if not constant, at least not to large. While for the Fe layer this aspect was not of importance given the small thickness of the layer, for the Ge layer, only extinction depths below 200 nm could be safely considered. The XAS spectra for grazing emission angles with a larger extinction depth were not considered. For the near-surface region affected by the oxidation, this estimation did furthermore not insert any error since the probed depth range is confined by the grazing emission geometry to much smaller values. Moreover, the differences in the extinction depth between the pure element and the respective native oxide could be neglected for Fe and Ge given the fact that the oxidation only affected the near-surface region. Indeed, for grazing angles below the critical angle, the extinction depths did not change with the oxidation for the investigated elements. If necessary, the information from the analysis of the XAS spectra can be used to calculate more accurate extinction depths for each grazing emission by taking into account the chemical composition of the sample. For both samples, the investigation of the GEXRF intensity profile for incidence energies above the absorption edge did not reveal a noticeable surface roughness. The surface roughness can be taken into account in the fits of the GEXRF intensity profiles using the approach described in ref 49 by modifying the reflection coefficients59 as it was realized in the investigation of Al layers on the top of Si.7 The presence of surface roughness is of importance for specifying the investigated depth range in the angular range below the critical angle where the refraction at the surface interface of the XRF photons (caused by a sudden change in the complex refractive index, which is itself determined by the energy of the XRF photons and the sample matrix) and the resulting evanescent wave determine the probed depth range as specified by eq 1. For grazing emission angles above the critical angle, the probed depth range is essentially determined by the attenuation of the XRF signal on the grazing emission path and increases roughly proportionally with the grazing emission angle (eq 1). In the case of significant surface roughness, i.e., of the order of the X-ray wavelength, the attenuation factor will be the determining factor for all the grazing emission angles and the refraction processes can be neglected in the calculation of the extinction depth. The probed depth range can then be calculated through a geometric consideration using only the absorption cross-section of the sample matrix for the XRF signal. For the Fe layer on the top of Si, the oxide is mostly confined to the surface, namely, the first 4 nm, with a relative

Figure 4. Depth-dependent oxide concentrations retrieved from the spectra shown in Figure 3 using reference absorption spectra for Fe and Fe2O3, respectively, and Ge and GeO2 (bottom panel). A relative concentration of 1 corresponds to times 2.5 × 1022 Fe atoms/cm3 for Fe, 6.0 × 1022 O atoms/cm3 in the case of Fe2O3, 4.4 × 1022 Ge atoms/cm3 for Ge, and 4.8 × 1022 O atoms/cm3 in the case of GeO2.

require a positive contribution of each component to the fit (a1 ≥ 0 and a2 ≥ 0). The standard error returned by the nonlinear least-squares fitting algorithm for each parameter was used as an estimate for the uncertainty of the contribution of each component (error bars in Figure 4). The measured XAS spectra could be thoroughly reproduced with the two components used for each sample. Indeed, the sum of the two components was for each point equal to 1 within the errors (see Figure 4, left panels for selected examples). The result of the fitting procedure confirmed that the presence of an additional oxidation state was unlikely. In GEXRF, the probed depth range is determined by the angle-dependent extinction depth of the emitted XRF signal (Figure 2, right panel, black solid line)1 10819


Article

Analytical Chemistry concentration of about 60% (O concentration of 3.6 × 1022 atoms/cm3). For larger probed depth ranges, the oxide concentration drops sharply to about 20% (O concentration of 1.2 × 10 22 atoms/cm 3 ) indicating that the oxide concentration decreases with increasing distance to the surface interface (Figure 4). Note that the probed depth range stands for the relative weight with which the different depths are investigated and increasing the probed depth range implies an increase in the studied sample volume (Figure 2, right panel). The width of the oxidized surface region as well as the fact that Fe2O3 is the dominant oxide are in agreement with experiments on iron nanoparticles using different techniques.60 This shows that the XAS measurements under grazing emission conditions can contribute to the understanding of oxidation of thin metal films and layers. For the Ge bulk sample, the relative surface oxidation concentration is significantly lower (18% at a probed depth range of 3 nm corresponding to an O concentration of 8.6 × 1021 atoms/cm3) and no oxide is present at depth ranges larger than 9 nm (Figure 4). Moreover, it can be observed that the estimate up to which depth self-absorption effects due to the shallow incidence angle can be neglected was too optimistic. Indeed, comparing the spectra at 31 and 98 nm in Figure 3 (middle panel), it can be seen that the white line is already damped in the XAS spectrum acquired for a larger probed depth. The agreement with the Ge reference spectrum is also slightly less. The damping of the white line gets even more pronounced with increasing probed depths. If larger depth ranges are to be explored without suffering from selfabsorption effects, a larger incidence angle for the experiment should be selected. From the evolution of the XAS spectra, it appears that for the presented measurements self-absorption effects can be safely neglected up to depths of 70 nm for the investigated Ge sample (still clearly outside the region affected by surface oxidation) rather than 200 nm as estimated. This finding about the surface oxide concentration for Ge shows that the used combination of fluorescence detected XAS and scanning-free GEXRF presents a promising tool (which is complementary to ellipsometry or electron spectroscopy methods) for investigating the GeO2/Ge interface under different conditions, e.g., thermal oxidation studies of the Ge surface. Unlike in the present study, these investigations could also be performed using the characteristic L-lines to profit from a less sharp evolution of the probed depth with the grazing emission angle. Indeed, for the present measurement performed at the K edge, the extinction depth increased from 4.5 to 130 nm in an angular range of 0.1° centered on the critical angle at 0.248° while for the L lines the critical angle is 1.35° and the extinction depth only varies from 4 to 50 nm over an angular range of 1.2° centered on the critical angle. However, measurements at the L-edge would require a vacuum environment because of the lower X-ray energy.

GEXRF data allowed one to assess the elemental depth distribution with nanometer-scale accuracy and thus to obtain a structural characterization of the sample, the angle dependent fluorescence detected XAS data allowed for a nondestructive depth-resolved characterization of the chemical state of the probed element. Indeed, characterizing the surface oxidation requires the enhancement of the signal from the surface region with respect to the one from the bulk volume underneath. At the same time, the presented scanning-free GEXRF-XAS approach allows for experiments at a good quality signal-tonoise ratio as compared to electron-based techniques or to the transmission XAS technique. Moreover, the method is not subject to the same sample limitations in terms of accessible depth, sample thickness, or concentration. Intrinsic to the conventional grazing emission geometry, self-absorption effects are, if not absent, minimal which ensures a reliable interpretation of the XAS data, independent of the sample thickness, the elemental concentration within the sample, or the sample density or uniformity. Therefore, the presented GEXRF-XAS setup presents the potential to contribute to the understanding of the physical and chemical properties at the surface or sharp intrasample interfaces and to be used for the study of dynamical processes. Compared to electron-probe methods, the presented technique can be applied to crystalline and bulk samples regardless of their electric and magnetic properties and the presence of external magnetic fields. Finally, by reducing the transmission losses in air, the described GEXRF-XAS technique presents the potential to be used as a reliable and sensitive tool for depth-resolved in situ studies of the elemental chemical state in novel and technologically relevant samples. In view of actual in situ studies with the presented setup, a reduction of transmission losses in air (1 m air paths upstream of the sample and from the sample to the detector due to spatial constraints and angular resolution requirements) would allow one to acquire data with good statistics in significantly shorter overall acquisition times. The feasibility of in situ studies was already demonstrated experimentally although, compared to the presented measurements, a shorter sample-to-dector distance was used to profit from a larger solid angle of detection at the expense of the angular resolution and only the XAS spectra obtained for grazing emission angles above the critical angle were investigated because of sample surface roughness.61,62

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +41 56 310 3555. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

■

ACKNOWLEDGMENTS The authors thank the Swiss Light Source for provision of beam time. The experiments were performed at the X05DA Optics beamline.

CONCLUSIONS AND PERSPECTIVES The presented GEXRF-XAS measurements demonstrate the possibility to acquire chemical state resolved depth-profiles by combining and exploiting the chemical sensitivity of fluorescence detected XAS and the element- and depthsensitive character of GEXRF. The experiment was realized with a scanning-free GEXRF setup using a two-dimensional position-sensitive area detector to record the XRF intensity dependence on the grazing emission angle in a single acquisition at a fixed sample−detector orientation for each point in the incident photon beam energy scan. While the

■

REFERENCES

(1) Klockenkämper, R. Total Reflection X-ray Fluorescence Analysis; Chemical analysis; Wiley-Interscience: New York, 1996; Vol. 140. (2) Alov, N. Inorg. Mater. 2011, 47, 1487−1499. (3) Meirer, F.; Singh, A.; Pianetta, P.; Pepponi, G.; Meirer, F.; Streli, C.; Homma, T. TrAC, Trends Anal. Chem. 2010, 29, 479−496. (4) von Bohlen, A. Spectrochim. Acta, Part B 2009, 64, 821−832. (5) Wobrauschek, P. X-Ray Spectrom. 2007, 36, 289−300. 10820


Article

Analytical Chemistry (6) Iida, A.; Yoshinaga, A.; Sakurai, K.; Gohshi, Y. Anal. Chem. 1986, 58, 394−397. (7) Kayser, Y.; Szlachetko, J.; Banaś, D.; Cao, W.; Dousse, J.-Cl.; Hoszowska, J.; Kubala-Kukuś, A.; Pajek, M. Spectrochim. Acta, Part B 2013, 88, 136−149. (8) Claes, M.; de Bokx, P.; Van Grieken, R. X-Ray Spectrom. 1999, 28, 224−229. (9) de Bokx, P.; Kok, C.; Bailleul, A.; Wiener, G.; Urbach, H. Spectrochim. Acta, Part B 1997, 52, 829−840. (10) Urbach, H. P.; de Bokx, P. K. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 53, 3752−3763. (11) Becker, R. S.; Golovchenko, J. A.; Patel, J. R. Phys. Rev. Lett. 1983, 50, 153−156. (12) Ashida, T.; Tsuji, K. Spectrochim. Acta, Part B 2014, 101, 200− 203. (13) Kubala-Kukuś, A.; Banaś, D.; Cao, W.; Dousse, J.-Cl.; Hoszowska, J.; Kayser, Y.; Pajek, M.; Salomé, M.; Susini, J.; Szlachetko, J.; Szlachetko, M. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80, 113305. (14) Szlachetko, J.; Banaś, D.; Kubala-Kukuś, A.; Pajek, M.; Cao, W.; Dousse, J.-Cl.; Hoszowska, J.; Kayser, Y.; Szlachetko, M.; Kavčič, M.; Salome, M.; Susini, J. J. Appl. Phys. 2009, 105, 086101. (15) Pahlke, S. Spectrochim. Acta, Part B 2003, 58, 2025−2038. (16) Streli, C.; Pepponi, G.; Wobrauschek, P.; Zöger, N.; Pianetta, P.; Baur, K.; Pahlke, S.; Fabry, L.; Mantler, C.; Kanngieer, B.; Malzer, W. Spectrochim. Acta, Part B 2003, 58, 2105−2112. (17) Pianetta, P.; Baur, K.; Singh, A.; Brennan, S.; Kerner, J.; Werho, D.; Wang, J. Thin Solid Films 2000, 373, 222−226 Proceedings of the 11th International Conference on Thin Films.. (18) Koshelev, I.; Paulikas, A.; Beno, M.; Jennings, G.; Linton, J.; Grimsditch, M.; Uran, S.; Veal, B. Oxid. Met. 2007, 68, 37−51. (19) Monaghan, M. L.; Nigam, T.; Houssa, M.; De Gendt, S.; Urbach, H. P.; de Bokx, P. K. Thin Solid Films 2000, 359, 197−202. (20) Müller, M.; Hönicke, P.; Detlefs, B.; Fleischmann, C. Materials 2014, 7, 3147−3159. (21) Unterumsberger, R.; Pollakowski, B.; Müller, M.; Beckhoff, B. Anal. Chem. 2011, 83, 8623−8628. (22) Kayser, Y.; Sá, J.; Szlachetko, J. Nanoscale 2015, 7, 9320−9330. (23) Nowak, S. H.; Banaś, D.; Błchucki, W.; Cao, W.; Hönicke, P.; Hoszowska, J.; Jabłoński, Ł.; Kayser, Y.; Kubala-Kukuś, A.; Pajek, M.; Reinhardt, F.; Savu, A. V.; Szlachetko, J. Spectrochim. Acta, Part B 2014, 98, 65−75. (24) Brücher, M.; Chakif, M.; Gurevich, E. L.; Hergenröder, R. Spectrochim. Acta, Part B 2014, 98, 60−64. (25) Menzel, M.; Fittschen, U. E. A. Anal. Chem. 2014, 86, 3053− 3059. (26) Motellier, S.; Derrough, S.; Locatelli, D.; Amdaoud, M.; Lhaute, K. Spectrochim. Acta, Part B 2013, 88, 1−9. (27) Reinhardt, F.; Osan, J.; Torok, S.; Pap, A. E.; Kolbe, M.; Beckhoff, B. J. Anal. At. Spectrom. 2012, 27, 248−255. (28) Tiwari, M. K.; Sawhney, K. J. S.; Lee, T.-L.; Alcock, S. G.; Lodha, G. S. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80, 035434. (29) von Bohlen, A.; Brücher, M.; Holland, B.; Wagner, R.; Hergenröder, R. Spectrochim. Acta, Part B 2010, 65, 409−414. (30) Kayser, Y.; Hönicke, P.; Banaś, D.; Dousse, J.-Cl.; Hoszowska, J.; Jagodziński, P.; Kubala-Kukuś, A.; Nowak, S. H.; Pajek, M. J. Anal. At. Spectrom. 2015, 30, 1086−1099. (31) Hönicke, P.; Kayser, Y.; Beckhoff, B.; Muller, M.; Dousse, J.-Cl.; Hoszowska, J.; Nowak, S. H. J. Anal. At. Spectrom. 2012, 27, 1432− 1438. (32) Kayser, Y.; Banaś, D.; Cao, W.; Dousse, J.-Cl.; Hoszowska, J.; Jagodziński, P.; Kavčič, M.; Kubala-Kukuś, A.; Nowak, S.; Pajek, M.; Szlachetko, J. X-Ray Spectrom. 2012, 41, 98−104. (33) Pepponi, G.; Giubertoni, D.; Bersani, M.; Meirer, F.; Ingerle, D.; Steinhauser, G.; Streli, C.; Hönicke, P.; Beckhoff, B. J. Vac. Sci. Technol. B 2010, 28, C1C59−C1C64. (34) Hönicke, P.; Beckhoff, B.; Kolbe, M.; Giubertoni, D.; van den Berg, J.; Pepponi, G. Anal. Bioanal. Chem. 2010, 396, 2825−2832.

(35) Giubertoni, D.; Iacob, E.; Hönicke, P.; Beckhoff, B.; Pepponi, G.; Gennaro, S.; Bersani, M. J. Vac. Sci. Technol. B 2010, 28, C1C84− C1C89. (36) Suzuki, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 1989, 39, 3393−3395. (37) Pollakowski, B.; Beckhoff, B.; Reinhardt, F.; Braun, S.; Gawlitza, P. Phys. Rev. B: Condens. Matter Mater. Phys. 2008, 77, 235408. (38) Brewe, D. L.; Pease, D. M.; Budnick, J. I. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 9025−9030. (39) Pfalzer, P.; Urbach, J.-P.; Klemm, M.; Horn, S.; denBoer, M. L.; Frenkel, A. I.; Kirkland, J. P. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 60, 9335−9339. (40) Tröger, L.; Arvanitis, D.; Baberschke, K.; Michaelis, H.; Grimm, U.; Zschech, E. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 46, 3283−3289. (41) Pease, D.; Brewe, D.; Tan, Z.; Budnick, J.; Law, C. Phys. Lett. A 1989, 138, 230−234. (42) Eisebitt, S.; Böske, T.; Rubensson, J.-E.; Eberhardt, W. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 47, 14103−14109. (43) Booth, C. H.; Bridges, F. Phys. Scr. 2005, 2005, 202. (44) Meirer, F.; Pepponi, G.; Streli, C.; Wobrauschek, P.; Zoeger, N. J. Appl. Phys. 2009, 105, 074906. (45) Pollakowski, B.; Hoffmann, P.; Kosinova, M.; Baake, O.; Trunova, V.; Unterumsberger, R.; Ensinger, W.; Beckhoff, B. Anal. Chem. 2013, 85, 193−200. (46) Leani, J. J.; Sánchez, H. J.; Pérez, R. D.; Pérez, C. Anal. Chem. 2013, 85, 7069−7075. (47) Barrett, N. T.; Gibson, P. N.; Greaves, G. N.; Mackle, P.; Roberts, K. J.; Sacchi, M. J. Phys. D: Appl. Phys. 1989, 22, 542. (48) Fox, R.; Gurman, S. J. J. Phys. C: Solid State Phys. 1980, 13, L249. (49) Kayser, Y.; Szlachetko, J.; Sá, J. Rev. Sci. Instrum. 2013, 84, 123102. (50) Flechsig, U.; Jaggi, A.; Spielmann, S.; Padmore, H.; MacDowell, A. Nucl. Instrum. Methods Phys. Res., Sect. A 2009, 609, 281−285. (51) Broennimann, C.; Eikenberry, E. F.; Henrich, B.; Horisberger, R.; Huelsen, G.; Pohl, E.; Schmitt, B.; Schulze-Briese, C.; Suzuki, M.; Tomizaki, T.; Toyokawa, H.; Wagner, A. J. Synchrotron Radiat. 2006, 13, 120−130. (52) Storm, L.; Israel, H. I. At. Data Nucl. Data Tables 1970, 7, 565− 681. (53) Shibayama, S.; Kato, K.; Sakashita, M.; Takeuchi, W.; Taoka, N.; Nakatsuka, O.; Zaima, S. Jpn. J. Appl. Phys. 2014, 53, 08LD02. (54) Scappucci, G.; Capellini, G.; Klesse, W. M.; Simmons, M. Y. Nanoscale 2013, 5, 2600−2615. (55) Xie, Q.; Deng, S.; Schaekers, M.; Lin, D.; Caymax, M.; Delabie, A.; Qu, X.-P.; Jiang, Y.-L.; Deduytsche, D.; Detavernier, C. Semicond. Sci. Technol. 2012, 27, 074012. (56) Pillarisetty, R. Nature 2011, 479, 324−328. (57) Yang, D.; An, Y.; Wang, S.; Wu, Z.; Liu, J. RSC Adv. 2014, 4, 33680−33686. (58) Araujo, L. L.; Giulian, R.; Sprouster, D. J.; Schnohr, C. S.; Llewellyn, D. J.; Kluth, P.; Cookson, D. J.; Foran, G. J.; Ridgway, M. C. Phys. Rev. B: Condens. Matter Mater. Phys. 2008, 78, 094112. (59) de Boer, D. K. G. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 53, 6048−6064. (60) Wang, C. M.; Baer, D. R.; Thomas, L. E.; Amonette, J. E.; Antony, J.; Qiang, Y.; Duscher, G. J. Appl. Phys. 2005, 98, 094308. (61) Shinoda, K.; Suzuki, S.; Yashiro, K.; Mizusaki, J.; Uruga, T.; Tanida, H.; Toyokawa, H.; Terada, Y.; Takagaki, M. Surf. Interface Anal. 2010, 42, 1650−1654. (62) Okumura, T.; Nakatsutsumi, T.; Ina, T.; Orikasa, Y.; Arai, H.; Fukutsuka, T.; Iriyama, Y.; Uruga, T.; Tanida, H.; Uchimoto, Y.; Ogumi, Z. J. Mater. Chem. 2011, 21, 10051−10060.

10821


Depth-Resolved X-ray Absorption Spectroscopy by Means of Grazing

Recommend Documents