Materials characterization by synchrotron x-ray microprobes and nanoprobes

In recent years synchrotron x-ray microprobes and nanoprobes have emerged as key characterization tools with a remarkable impact for different scientific fields including solid-state, applied, high-pressure, and nuclear physics, chemistry, catalysis, biology, and cultural heritage. This review provides a comparison of the different probes available for the space-resolved characterization of materials (i.e., photons, electrons, ions, neutrons) with particular emphasis on x rays. Subsequently, an overview of the optics employed to focus x rays and the most relevant characterization techniques using x rays (i.e., x-ray diffraction, wide-angle x-ray scattering, small-angle x-ray scattering, x-ray absorption spectroscopy, x-ray fluorescence, x-ray-excited optical luminescence, and photoelectron spectroscopy) is reported. Strategies suitable to minimize possible radiation damage induced by brilliant focused x-ray beams are briefly discussed. The general concepts are then exemplified by a selection of significant applications of x-ray microbeams and nanobeams to materials science. Finally, the future perspectives for the development of nanoprobe science at synchrotron sources and free-electron lasers are discussed.

Figure 1

Schematic representation of the key factors leading to x-ray microbeam and nanobeam science development: (i) the increasing demand for space-resolved probes to characterize heterogeneous samples down to the nanoscale, coming from very different research fields; (ii) the high brilliance currently achievable exploiting third- and fourth-generation synchrotron sources; (iii) the impressive progresses in x-ray-focusing optics; and (iv) the rich interaction of x rays with matter, providing unique characterization tools. The top left panel reports the time evolution for x-ray brilliance, distinguishing between x-ray tubes, different generations of synchrotron sources [dedicated second-generation sources after the early parasitic use, third-generation sources using insertion devices, fourth generation diffraction-limited storage rings (DLSRs)], and linear XFEL sources. For DLSRs, both current and foreseen values are reported ([689]). For XFELs, peak-brilliance values are indicated as also reported in Table 3 to which the reader is referred for information about pulse duration and repetition rate.

Figure 2

Overview of the wide range of multidisciplinary applications which exploit x-ray microbeams and nanobeams. The reported illustrations representative of each application field are adapted with permission from the following sources, starting from top left, in clockwise order: nuclear science. From [301]. High-pressure science: previously unpublished. Environmental science. From [578]. Earth science. From [513]. Cultural heritage. From [674]. Chemistry. From [423]. Materials science. From [603], [269], and [353]. Solid state physics. From [553]. Life science. From [506]. Macromolecular crystallography. From [548].

Figure 3

Schematic representation of the teardrop-shaped interaction volume of the electron beam inside the sample. The different signals that can be measured in a (a) TEM or (b) SEM microscope are also indicated.

Figure 4

Schematic representation of the different scattering, excitation, and decay phenomena that occur when an ion beam interacts with matter and the consequent IBA techniques: photon emissions [ionoluminescence, PIXE, and proton-induced or particle-induced $\gamma$-ray emission (PIGE), depending on the photon energy]; backscattered ions and nuclear reaction products in Rutherford backscattering spectroscopy (RBS) and nuclear reaction analysis (NRA), respectively; sputtered ions in secondary ions mass spectrometry (SIMS); and nuclei recoils in elastic recoil detection analysis (ERDA).

Figure 5

X-ray focusing optics classification. The three principal categories (refractive, reflective, and diffractive optics) and the most common typologies for each category are indicated.

Figure 6

(a) Schematic of the geometry for optical demagnification, described for a thin lens by the Gaussian lens formula (3). The object (or source $S$) and image ($s$) distances are indicated with $p$ and $q$, respectively, while the lens focal length is $f$ ($F$ being the focal point). The lens opening angle is referred to as $\theta$. (b) Definition of numerical aperture in the focus-to-infinity ($S\to \infty$) case. The geometrical aperture of the lens is indicated with $D$.

Figure 7

Time evolution of the minimum hard-x-ray ($E>4\mathrm{keV}$) spot size during the last two decades and extrapolation for the near future, distinguishing among refractive (orange circles and line), reflective (magenta triangles and line), and diffractive (blue squares and line) optics. Data were sorted from the values published in a suitable series of reports, while the trend lines are obtained upon monoexponential fits (notice the logarithmic scale for the ordinate axis). The dashed parts of the trend lines indicate the extrapolations for the very next future. The ultimate resolution limits ($S^L$) for the three categories are also indicated as dotted horizontal lines.

Figure 8

(a) Schematic representation of the principal x rays and matter interactions. (b) Classification of most popular x-ray-based space-resolved analytical tools with micrometric and submicrometric resolution; four main categories are represented: microspectroscopy, microdiffraction, (spectro)microscopy, and imaging. For graphical reasons only the “micro-” prefix is reported although nowadays the majority of these analytical techniques can provide information at the submicrometric scale. The most common variants, methods, and operation modalities per each category are reported as well.

Figure 9

(a) Incoherently illuminated assembly of nano-objects, relative 2D and 1D scattering patterns. (b) Coherently illuminated assembly of nano-objects, relative 2D and 1D scattering patterns.

Figure 10

Schemes of the different processes and related spectroscopies resulting after the interaction of a high-energy beam with matter: (a) excitation process, core hole creation; (b) radiative and (c) Auger decay channels.

Figure 11

(a) Compilation of IMFP measurements for different materials as a function of the electron kinetic energy $T$ (scattered dots). The curve represents the empirical least squares best fit to Eq. (18) over all the reported data. Adapted from [600]. (b) IMFP for $\mathrm{Si}(2p)$ photoelectrons in silicon as a function of photon energy. Adapted from [690].

Figure 12

Optical scheme of the (a) full-field x-ray microscope and (b) the scanning x-ray microscope.

Figure 13

Schematic model of a prototype scanning nanoprobe or microprobe beam lines dedicated to multitechnique space-resolved characterization.

Figure 14

Schematic representation of the cascade channels of energy transfer involved in the process of interaction between an intense hard-x-ray beam and condensed matter. (a) Classifies the different channels in a qualitative (energy, time) plane. (b) Depicts the physical processes involved in the main channels. In (a) also represented are the typical FEL (blue arrow, 2–200 fs, see Table 3) and synchrotron (red arrow 20 ps—1 ns) pulse duration. From [394].

Figure 15

Finite-element modeling of the temperature increase inside an InP NW supported on a ${\mathrm{Si}}_3{\mathrm{N}}_4$ membrane (a)–(c) after exposure to x-ray pulses of length 0.1 ns (photon $\text{energy}=10\mathrm{keV}$; $\mathrm{flux}={10}^{12}\mathrm{photons}{\mathrm{s}}^{-1}$; $\text{power}=1.6\mathrm{mW}$). (a) 3D representation of the model; each node in the finite-element mesh is indicated with a dot. (b) Simulated temperature increase $dT$ (relative to RT) after absorption of a single x-ray pulse (3000 photons) in the 0.1–5.0 ns interval after beam exposure. The color scale is different for each time. (c) Effect of a pulse repetition with period 3 ns on the local NW temperature at the beam position (black); the average NW temperature [orange (light gray), the arrow indicates the value obtained in a steady-state simulation]; the substrate temperature [blue (dark gray)]. (d)–(f) Simulated temperature in the InP NW contacted with 100-nm-thick gold contacts, as viewed along the $z$ axis (beam direction). (d) Drawing of the sample. Temperature gradients obtained at a steady-state condition by impinging the x-ray beam at the center of the NW (e) or at the left interface between InP and Au (same color scale). Adapted from [683].

Figure 16

(a) Top: SXD map showing the diffracted intensity at the $\mathbf{Q}$ value of (004) Bragg reflection for relaxed SiGe/Si(001). Bottom left: Optical microscopy image showing a portion of the SXD map. Bottom right: Cartoon schematizing the method. When the $\mu \mathrm{m}$-focused x-ray spot illuminates a single SiGe island a diffused peak is observed, due to lattice spacing distribution inside the SiGe islands (blue halo) together with the sharp (004) peak of the Si substrate (green peak), appearing at a different $\mathbf{Q}$ value. (b) RSM of SiGe islands around (004) reflection. The left panel was obtained with a defocused beam averaging the scattering from about ${10}^3$ islands. 1: linear detector saturation streak, 2: monochromator streak, 3: Si surface crystal truncation rod, and 4: facet streaks originating from the (111) island side facets. Right panels: RSM from two individual islands IL1 (fully developed truncated pyramid) and IL2 (flat island). The top insets report the corresponding SEM micrographs. Adapted from [459]. (c) Top panels: $3\times 3\mu {\mathrm{m}}^2$ SEM micrograph of a single SiGe island and corresponding SXD map obtained using the diffusely scattered x-ray intensity in proximity of the symmetric Si(004) reflection. Bottom panel: Reassembled diffraction pattern of a single island, obtained using frames 12–14 of the SXD map. Adapted from [269].

Figure 17

(a) Top panels: SEM images of the SiGe/Si MOSFET device at increasing magnifications. Bottom panels: SDX maps collected at similar scales; left panel: SDX map monitoring diffracted intensities around the Si(004) reflection for finding the transistor junction; middle and right panels: SDX map monitoring diffracted intensities around the SiGe(224) reflection for locating the dot islands. (b) RSM map around the (224) reflection measured on the SiGe dot buried under the transistor gate; the map includes the Si(224) bulk peak and the SiGe(224) dot signal in its lower left section (highlighted by the dashed circle). (c) Top: FEM-based simulated data, calculated for the region around the diffuse SiGe(224) signal, defined with the orange rectangular box in (b); bottom panels: resulting 2D maps of the Ge concentration and the in-plane strain ${ϵ}_{xx}$. Adapted from [293].

Figure 18

(a) SEM micrographs of the three pseudomorphic GaAs/InGaAs RUNTs exhibiting different GaAs thickness: (I), (II), and (III). (b) Correlation between the optical microscopy image and the x-ray measurements determining the RUNT position: the selected tube is optically prealigned with its axis perpendicular to the incoming x-ray beam. Because of the rolling up, the crystalline lattice of the RUNT is oriented isotropically perpendicular to the tube axis, and the isotropic scattering intensity in the azimuthal direction is independent of $\alpha$, geometry (1) By detuning the incident angle, geometry (2), the thermal diffuse background scattering of the substrate is significantly reduced, allowing the scattering of the RUNTs to be discriminated from the background signal. (c) Experimental intensity distribution, normalized to the GaAs (004) reflection for RUNTs (I)–(III) (black dots) and corresponding simulations (pink curves). (d) Lattice parameter distribution in tangential and radial directions ($a_t$, full dots, and $a_r$, empty dots, respectively) as used for the simulations. Adapted from [347].

Figure 19

(a) Scheme of the multimodal nanocharacterization setup available at the ID22 (successively implemented and moved to ID16B) beam line of the ESRF. (b) XEOL spectra collected on the corner (blue circles) and edge (red triangles) points evidenced in (c) with the same symbols. (c) $800\times 800{\mathrm{nm}}^2$ XEOL intensity map (acquired at the MQW emission energy of 2.86 eV) of a hexagonal core-shell MQW structure. (d) Electron probability density calculated for the same structure. (e) Decay of the time-resolved XEOL signal collected for the MQW emission at the hexagon corner. Adapted from [407].

Figure 20

PES intensity maps for the $\mathrm{Ga}(3d)$ peak in a sequence of $p\text{−}n$ GaAs homojunctions. (a) Map for the $\mathrm{Ga}(3d)$ peak in $n\text{−}\mathrm{GaAs}$; (b) map for the $\mathrm{Ga}(3d)$ peak in $p\text{−}\mathrm{GaAs}$; (c) digit-by-digit subtraction image of (b) from (a). Note the reversed contrast between (a) and (b), further enhanced in the subtraction image. From [39].

Figure 21

(a) Data collection scheme of a PCDI experiment realized at the cSAXS PSI beam line. (b) SEM image of CdSe/CdS octapod-shaped nanocrystals grouped in linear chains. (c) Typical speckled diffraction pattern as collected for each of the overlapping area of the explored region. (d) Reconstructed $2.5\times 2.5\mu {\mathrm{m}}^2$ image. (e) Detail of (d). (a), (c) Previously unpublished. (b), (d), and (e) Adapted from [139].

Figure 22

(a) Average XRF spectrum (collected exciting at 12 keV) of the crossed multiwire structure. (b) SEM micrograph of the structure. (c) Color map of XRF intensities for Ga, Sn, and Cr [red, green, and blue, respectively(light gray, white, and dark gray)], reported using a brightness scale (dark: low counts; light: high counts). A SEM magnification of the region marked with the white circle is reported in (d). (d) SEM micrograph of the nanowire junction. (e) 12-keV x-ray transmission map of the junction region. (f) Detail of the RGB XRF map in the junction region, same color code as in (c). (g) Normalized XRF profiles for Ga (red circles), Sn (green squares), and Cr (blue triangles), acquired along the white dotted line visible in (f). (h) Ga $K$-edge XANES spectra (vertically shifted for clarity) collected in the junction region and outside. (i) Experimental (white circles) Ga $K$-edge FT of the EXAFS spectra and respective best fits (solid lines) collected at the junction and outside. Adapted from [408].

Figure 23

(a) SEM micrograph of a single ZnO nanowire implanted with Co. (b) Map showing the Co $K\alpha$ and $K\beta$ fluorescence collected using a 12 keV beam. The Co content was determined from the XRF quantification. (c), (d) Zn $K$-edge XANES spectra, collected at points 1–3 of (a), with the $c$ axis oriented perpendicular (c), or parallel (d), to the electric field vector of the x-ray beam. (e) Zn (black dots) and Co (blue dots) $K$-edge XANES spectra collected at points 1 and 2. For both edges, the abscissa axis has been obtained rescaling the photon energy with respect to the ionization energy, allowing a direct comparison. (f) Magnitude of the experimental Zn $K$-edge FT of the EXAFS functions collected at points 1–3 of (a) (scattered dots) and corresponding best fits (solid lines). Adapted from [603].

Figure 24

(a) Average XRF spectra acquired in different areas of a GaN:Mn sample. (b) Color map: pink, blue, and green (dark gray, black, and light gray) correspond to the Mn $K\alpha$, Ga $K\alpha$ fluorescence lines, and Compton scattering signal, respectively. Ga (black) and Mn (pink) fluorescence profiles along the white scan line are also shown. (c) Mn oxidation state maps showing ${\mathrm{Mn}}^0$, ${\mathrm{Mn}}^{2+}$, and ${\mathrm{Mn}}^{3+}$ centers. (a), (b) Adapted from [410]. (c) Previously unpublished figure, reporting data published in [409].

Figure 25

(a) XRF maps obtained on AlGaN:Si selecting the $\mathrm{Al}\text{−}K\alpha$, $\mathrm{Si}\text{−}K\alpha$ and $\mathrm{Ga}\text{−}L\alpha$ fluorescence lines. The pixel size is $5\times 20\mu {\mathrm{m}}^2$ ($H\times V$) in the larger and $2\times 1\mu {\mathrm{m}}^2$ in the smaller maps. (b), (c) Si $K$-edge XANES spectra acquired at points with different Si concentrations in both (b) fluorescence yield (FLY) and (c) TEY modes. Adapted from [622].

Figure 26

(a) Sketch of the experimental setup at the ID10 beam line of the ESRF reporting the distance $d$ from the x-ray source of the different elements. (b) Example of some CDI diffraction patterns among the 73 frames taken for sample tilts between $-72°$ and $+72°$ with a step of 2°. (c) SEM image of the investigated ${\mathrm{Fe}}_2\mathrm{P}$ cluster. (d) Reconstructed 3D image of the sample (resolution of 59 nm). Continuous arrows show several voids; dashed arrows point to the high-density aggregates. (e) Averaged SAXS profile (experimental, black circles) and computed profile (red line). (a)–(d) From [117]. (e) From Dr. Chushkin (ESRF).

Figure 27

(a) Sketch of the dynamic high-pressure setup available at the ESCA microscopy beam line at Elettra. (b) Time profiles of the pressure values at the sample (left axis) and inside the electron analyzer (right axis) as a function of pulse duration of a pulsed valve fed by an ${\mathrm{O}}_2$ gas pressure of 3.5 bar. (c) Standard SEM micrograph of a polycrystalline PtRh particle. (d) SPEM $\mathrm{Rh}({3d}_{5/2})$ map of a PtRh particle with a similar shape. (e) $\mathrm{Rh}({3d}_{5/2})$ spectra acquired on a metal Rh(100) reference sample and on the reduced PtRh particle. (f), (g) $\mathrm{Rh}({3d}_{5/2})$ spectra acquired from the two points of the PtRh particle, labeled as $A$ and $B$ in (d), after an oxidation treatment of increasing duration (195 and 715 min, respectively). All XPS spectra were acquired with a photon energy of 650 eV. Adapted from [607].

Figure 28

(a) General scheme of a typical fixed-bed quartz capillary setup with reactor, gas blower, and thermocouple. The red and blue (dark and light gray) represent the parts of the catalytic bed where Pt particles are in the oxidized and reduced phases, respectively. (b) $\mathrm{Pt}{L}_3$-edge XANES spectra of oxidized and reduced Pt phases (measured at the inlet and the end zone of the catalytic bed, respectively). The highest contrast between the spectra of the two species occurs at 11 596 eV. (c)–(g) Time evolution of the front of reduction of the catalyst moving through the middle part of the fixed bed monitored by measuring the absorption coefficient at 11 596 eV. The size of the reported region is $1\times 1{\mathrm{mm}}^2$: reduced species (red-violet, dark gray); oxidized species (yellow-orange, light gray). (h)–(l) Similar to (c)–(g) for the reduction (ignition) of single particles. (m) Pure x-ray absorption image of the region monitored in (h)–(l). (n) Model for the reduction of single particles at the end of the catalyst bed during ignition of the catalytic partial oxidation of methane. ${\mathrm{O}}_2$ is fully consumed in an outer layer and reforming reactions occur on the inside, which causes the noble metals to be reduced. Adapted from [332].

Figure 29

Spatial distribution of different ${\mathrm{K}}_x{\mathrm{Fe}}_{2-y}{\mathrm{Se}}_2$ phases at 15 K. The intensity distribution is obtained by integrating intensities of (004) diffraction peaks corresponding to different phases. The interface phase (green, white) is clearly visible between the majority AFM phase (red, dark gray) and the minority metallic filamentary phase (blue, black). The lower panels show the spatial distributions of different phases at several temperatures, revealing the evolution of percolative paths (blue) below the phase separation temperature, getting protection of the interface phase (green) at low temperature. Adapted from [553].

Figure 30

(a) Typical layout of the chip used for device writing. Current and voltage electrodes are labeled as $I^{+}$, $I^{-}$, $V^{+}$, and $V^{-}$. (b) Irradiation geometry used at the ESRF ID16B beam line to pattern the device in the central part of the chip. Pt pillars are used for alignment purposes only. Superconducting Cu planes are parallel to the substrate. (c) Sketch of trench 1, along with the supercurrent path induced by the trench. The green arrow represents the x-ray nanobeam. (d) $I\text{−}V$ characteristics of a patterned device measured at constant temperatures between 52 and 80 K. The purple arrows exemplify the typical hysteretic pattern in the case of $T=56\mathrm{K}$. The inset shows the temperature behavior of the corresponding critical current values $I_c$. Void circles refer to curves that have not been shown in the main panel for clarity. Unpublished figure reporting data published in [658].

Figure 31

(a) Variation of the electrical resistance of the Bi-2212 sample measured online during the x-ray exposure to write the first trench (red curve, dark gray). The black curve represents the initial sample resistance immediately before the x-ray nanopatterning procedure. (b) Temporal evolution of the sample temperature calculated with the finite-element method at the point of maximum temperature considering the average continuous photon flux (black curve) and the pulsed nature of SR (red curve, dark gray). The initial temperature of the system was set to 300 K. Previously unpublished figure reporting data published in [446].

Figure 32

(a) Optical micrograph of the InP substrate patterned with ${\mathrm{SiO}}_2$ stripes used for the SAG technique. In the inset a magnification of the interface between SAG and FIELD regions clarifying the growth mechanism is shown. All the measurements have been performed along the white line parallel to the ${\mathrm{SiO}}_2$ stripes and equidistant from them. This mask had $20\mu \mathrm{m}$ wide ${\mathrm{SiO}}_2$ stripes with a $30\mu \mathrm{m}$ gap between them. (b) XRD patterns acquired along the line reported in (a), starting $30\mu \mathrm{m}$ before the end of the stripes in the SAG region. (c) Barrier and well widths and period as a function of the position, obtained by simulation of the experimental XRD patterns reported in (b). (d) Similar to (c) for the well, barrier, and overall mismatches. Adapted from [449].

Figure 33

(a) Wavelength corresponding to the maximum of emission for each PL spectrum acquired on samples with different ${\mathrm{SiO}}_2$ stripes’ width ($W_S$) along the white line shown in Fig. 32. (b) Spatial maps of the ratio between Ga ($K\alpha$) and As ($K\beta$) fluorescence counts which point out the variation in the average (between well and barrier) ${\mathrm{Al}}_x{\mathrm{Ga}}_y{\mathrm{In}}_{1-x-y}\mathrm{As}$ alloy composition between SAG and FIELD regions for the different masks. The horizontal $\times$ vertical dimensions of the maps are $55\times 230$, $50\times 290$, and $50\times 205\mu {\mathrm{m}}^2$ for the samples with $W_S=30$, 20 and 10 μm, respectively. Adapted from [445].

Figure 34

(a), (c), (e), (g) Photoelectron emission microscopy images of the switching filament in the LRS, HRS, LRS II, and HRS II, respectively (indicated by the black arrow) for a photon energy of 531.6 eV. Scale bars, $1\mu \mathrm{m}$. (b), (d), (f), (h) O $K$-edge spectra for the switching filament in (a), (c), (e), (g) [green and blue lines for the LRS and HRS, respectively (light and dark gray)] and the surrounding device area (black lines). (i) Device resistance as a function of the device state. Adapted from [34].

Figure 35

(a) Forward bias electroluminescence image of part of the solar cell showing recombination-active defects. Breakdown light emission was observed at the sites marked by the white circles. (b) SEM image of the part of the solar cell marked by the dotted rectangle in (a); the micro-XRF maps of metal Fe precipitates, labeled as 1 and 2 in the SEM micrograph and highlighted by white circles, are also reported. Adapted from [353].

Figure 36

XBIC image and XRF maps for $\mathrm{Cu}\text{−}K\alpha$ and $\mathrm{F}\mathrm{e}\text{−}K\alpha$ in the polycrystalline-Si solar cell. Adapted from [659].

Figure 37

(a) Plan-view SEM image of the $e\text{−}\mathrm{SiC}/\mathrm{SOI}$ channel device. (b) Cross-sectional TEM image of the $e\text{−}\mathrm{SiC}/\mathrm{SOI}$ channel. (c) Micro-XRD patterns acquired in the SOI channel (red dots) and $0.8\mu \mathrm{m}$ away from the channel (black squares). (d) Schematic cross-sectional geometry of the device, indicating the out-of-plane strain ${ϵ}_{33}$ calculated using the Eshelby inclusion model in the SOI channel (red region, dark gray) and in the $e\text{−}\mathrm{SiC}$ feature $0.8\mu \mathrm{m}$ away from the channel (blue region, black) corresponding to the XRD measurement locations. Adapted from [472].