A novel ultra-high vacuum instrument for X-ray reflectometry and spectrometry-related techniques for nanoanalytics by means of synchrotron radiation has been constructed and commissioned. This versatile instrument was developed by the Physikalisch-Technische Bundesanstalt, Germany's national metrology institute, and includes a 9-axis manipulator that allows for an independent alignment of the samples with respect to all degrees of freedom. In addition, a rotational and translational movement of several photodiodes as well as a translational movement of an aperture system in and out of the beam is provided. Thus, the new instrument enables various analytical techniques based on energy dispersive X-ray detectors such as reference-free X-ray fluorescence analysis (XRF), total-reflection XRF, grazing-incidence XRF in addition to optional X-ray reflectometry measurements or polarization-dependent X-ray absorption fine structure analyses. With this instrument samples having a size of up to 100 mm × 100 mm can be analyzed with respect to their mass deposition, elemental or spatial composition, or the species in order to probe surface contamination, layer composition and thickness, the depth profile of matrix elements or implants, the species of nanolayers, nanoparticles or buried interfaces as well as the molecular orientation of bonds. Selected applications of this advanced ultra-high vacuum instrument demonstrate both its flexibility and capability.

The geometry of a Si3N4 lamellar grating was investigated experimentally with reference-free grazing-incidence X-ray fluorescence analysis. While simple layered systems are usually treated with the matrix formalism to determine the X-ray standing-wave field, this approach fails for laterally structured surfaces. Maxwell solvers based on finite elements are often used to model electrical field strengths for any 2D or 3D structures in the optical spectral range. We show that this approach can also be applied in the field of X-rays. The electrical field distribution obtained with the Maxwell solver can subsequently be used to calculate the fluorescence intensities in full analogy to the X-ray standing-wave field obtained by the matrix formalism. Only the effective 1D integration for the layer system has to be replaced by a 2D integration of the finite elements, taking into account the local excitation conditions. We will show that this approach is capable of reconstructing the geometric line shape of a structured surface with high elemental sensitivity. This combination of GIXRF and finite-element simulations paves the way for a versatile characterization of nanoscale-structured surfaces.

Photon-in/photon-out experiments at thin specimens have been carried out to determine L-subshell fluorescence yields as well as Coster-Kronig transition probabilities of Au, Pb, Mo, and Pd using radiometrically calibrated instrumentation in the Physikalisch-Technische Bundesanstalt (PTB) laboratory at the electron storage ring BESSY II in Berlin. An advanced approach was developed in order to derive the fluorescence line intensities by means of line sets of each subshell that were corrected for self-absorption and broadened with experimentally determined detector response functions. The respective photoelectric cross sections for each subshell were determined by means of transmission measurements of the same samples without any change in the experimental operating condition. All values derived were compared to those of earlier works. A completely traceable uncertainty budget is provided for the determined values.The complete understanding of atomic excitation and emission processes as well as the exact knowledge of related cross sections are crucial for quantitative investigations of elemental concentrations in unknown specimens by x-ray spectrometry. The emitted spectral response of an unknown sample following the excitation by x rays is based on both the specimen's elemental composition and the related atomic processes, the probability of which depends on a set of atomic fundamental parameters. For the quantitative determination of the respective elemental concentrations, the cross sections for photoionization of the (sub)shell and for both elastic and inelastic scattering, the related fluorescence yield, and transition probabilities associated with the specific inner-shell excitations have to be exactly known. Therefore, the uncertainty of these fundamental parameters is relevant for reference-based as well as reference-free x-ray fluorescence analysis results. Here, a reliable uncertainty budget is needed [1-3].Krause et al.[4] estimated relative uncertainties by comparing of theoretical total photoionization cross sections with those determined experimentally, which have still mostly remained valid since the late 1970's [5]. In addition, one may assume that most of the relative uncertainties are estimated rather from the compilation of available data than from the individual experimental uncertainty budget. An example of a compilation-based uncertainty indication is the compilation of theoretical data by Puri et al. published in 1993 for elements 25 Z 96 based on the relativistic Dirac-Hartree-Slater model [6], which states only a fitting error of 2% for elements with higher Z.The fluorescence yields as well as Coster-Kronig (CK) transition probabilities were investigated in earlier works, e.g., by electron-induced high-resolution x-ray fluorescence spectroscopy, Kα-Lα-coincidence methods [7]. When monochromatic tunable synchrotron radiation became available, photoionization experiments were also carried out, but the respective relative uncertainties could not be reduced significantly [5]. This was partially...

A need for analysis techniques, complementary to secondary ion mass spectrometry (SIMS), for depth profiling dopants in silicon for ultra shallow junction (USJ) applications in CMOS technologies has recently emerged following the difficulties SIMS is facing there. Grazing incidence X-ray fluorescence (GIXRF) analysis in the soft X-ray range is a high-potential tool for this purpose. It provides excellent conditions for the excitation of the B-K and the As-L(iii,ii) shells. The X-ray standing wave (XSW) field associated with GIXRF on flat samples is used here as a tunable sensor to obtain information about the implantation profile because the in-depth changes of the XSW intensity are dependent on the angle of incidence. This technique is very sensitive to near-surface layers and is therefore well suited for the analysis of USJ distributions. Si wafers implanted with either arsenic or boron at different fluences and implantation energies were used to compare SIMS with synchrotron radiation-induced GIXRF analysis. GIXRF measurements were carried out at the laboratory of the Physikalisch-Technische Bundesanstalt (PTB) at the electron storage ring BESSY II using monochromatized undulator radiation of well-known radiant power and spectral purity. The use of an absolutely calibrated energy-dispersive detector for the acquisition of the B-Kalpha and As-Lalpha fluorescence radiation enabled the absolute determination of the total retained dose. The concentration profile was obtained by ab initio calculation and comparison with the angular measurements of the X-ray fluorescence.

The accurate characterization of nanolayered systems is an essential topic for today’s developments in many fields of material research. Thin high-k layers and gate stacks are technologically required for the design of current and future electronic devices and can be deposited, e.g., by Atomic Layer Deposition (ALD). However, the metrological challenges to characterize such systems demand further development of analytical techniques. Reference-free Grazing Incidence X-ray Fluorescence (GIXRF) based on synchrotron radiation can significantly contribute to the characterization of such nanolayered systems. GIXRF takes advantage of the incident angle dependence of XRF, in particular below the substrate’s critical angle where changes in the X-ray Standing Wave field (XSW) intensity influence the angular intensity profile. The reliable modeling of the XSW in conjunction with the radiometrically calibrated instrumentation at the PTB allows for reference-free, fundamental parameter-based quantitative analysis. This approach is very well suited for the characterization of nanoscaled materials, especially when no reference samples with sufficient quality are available. The capabilities of this method are demonstrated by means of two systems for transistor gate stacks, i.e., Al2O3 high-k layers grown on Si or Si/SiO2 and Sc2O3 layers on InGaAs/InP substrates.

An experimental method for the verification of the individually different energy dependencies of L(1)-, L(2)-, and L(3)- subshell photoionization cross sections is described. The results obtained for Pd and Mo are well in line with theory regarding both energy dependency and absolute values, and confirm the theoretically calculated cross sections by Scofield from the early 1970 s and, partially, more recent data by Trzhaskovskaya, Nefedov, and Yarzhemsky. The data also demonstrate the questionability of quantitative x-ray spectroscopical results based on the widely used fixed jump ratio approximated cross sections with energy independent ratios. The experiments are carried out by employing the radiometrically calibrated instrumentation of the Physikalisch-Technische Bundesanstalt at the electron storage ring BESSY II in Berlin; the obtained fluorescent intensities are thereby calibrated at an absolute level in reference to the International System of Units. Experimentally determined fixed fluorescence line ratios for each subshell are used for a reliable deconvolution of overlapping fluorescence lines. The relevant fundamental parameters of Mo and Pd are also determined experimentally in order to calculate the subshell photoionization cross sections independently of any database.

In this work two synchrotron radiation-based depth-sensitive X-ray fluorescence techniques, grazing incidence X-ray fluorescence (GIXRF) and grazing emission X-ray fluorescence (GEXRF), are compared and their potential for non-destructive depth-profiling applications is investigated. The depth-profiling capabilities of the two methods are illustrated for five aluminum-implanted silicon wafers all having the same implantation dose of 10 16 atoms per cm 2 but with different implantation energies ranging from 1 keV up to 50 keV. The work was motivated by the ongoing downscaling effort of the microelectronics industry and the resulting need for more sensitive methods for the impurity and dopant depth-profile control. The principles of GIXRF and GEXRF, both based on the refraction of X-rays at the sample surface to enhance the surface-to-bulk ratio of the detected fluorescence signal, are explained. The complementary experimental setups employed at the Physikalisch-Technische Bundesanstalt (PTB) for GIXRF and the University of Fribourg for GEXRF are presented in detail. In particular, for each technique it is shown how the dopant depth profile can be derived from the angular intensity dependence of the Al Ka fluorescence line. The results are compared to theoretical predictions and, for two samples, crosschecked with values obtained from secondary ion mass spectroscopy (SIMS) measurements. A good agreement between the different approaches is found proving that the GIXRF and GEXRF methods can be efficiently employed to extract the dopant depth distribution of ionimplanted samples with good accuracy and over a wide range of implantation energies.

Nanolayer stacks are technologically very relevant for current and future applications in many fields of research. A non-destructive characterization of such systems is often performed using X-ray reflectometry (XRR). For complex stacks of multiple layers, low electron density contrast materials or very thin layers without any pronounced angular minima, this requires a full modeling of the XRR data. As such modeling is using the thicknesses, the densities and the roughnesses of each layer as parameters, this approach quickly results in a large number of free parameters. In consquence, cross-correlation effects or interparameter dependencies can falsify the modeling results. Here, we present a route for validation of such modeling results which is based on the reference-free grazing incidence X-ray fluorescence (GIXRF) methodology. In conjunction with the radiometrically calibrated instrumentation of the Physikalisch-Technische Bundesanstalt the method allows for reference-free quantification of the elemental mass depositions. In addition, a modeling approach of reference-free GIXRF-XRR data is presented, which takes advantage of the quantifiable elemental mass depositions by distributing them depth dependently. This approach allows for a reduction of the free model parameters. Both the validation capabilities and the combined reference-free GIXRF-XRR modeling are demonstrated using several nanoscale layer stacks consisting of HfO 2 and Al 2 O 3 layers.