The intermediate-energy scanning x-ray microscope at beamline 2-ID-B at the Advanced Photon Source is a dedicated instrument for materials and biological research. The microscope uses a zone plate lens to focus coherent I-4 keV x-rays to a 60 nm focal spot of 10 9 photons/ s onto the sample. It records simultaneous transmission and energy-resolved fluorescence images. We have used the microscope for nanotomography of chips and microspectroscopy of cells.

The dynamics of a thixotropic silica aerosil gel dispersed in an octylcyanobiphenyl liquid crystal were directly probed by x-ray intensity fluctuation spectroscopy. For all samples, the time-autocorrelation function of the gel was well described by a modified-exponential function over the q range studied. Compared to a pure gel sample, a dilute (0.06 g cm(-3)) gel embedded within the liquid crystal displayed more complex and temperature dependent dynamics. Near the second-order smectic-A-to-nematic phase transition of the liquid crystal the gel relaxation became significantly more complex and slower (tau approximately 2150 s) compared to relaxations observed well within either phase. This clearly demonstrates coupling between the dynamics of the gel and the host liquid crystal, consistent with critical slowing down of smectic and director fluctuations. A random dampening field, elastically coupled to the liquid crystal, would explain the earlier observed crossover of this transition towards 3d-XY behavior.

We measure the spatial coherence function of a quasi-monochromatic 1.1 keV X-ray beam from an undulator at a third-generation synchrotron. We use a Young's slit apparatus to measure the coherence function and find that the coherence measured is poorer than expected. We show that this difference may be attributed to the effects of speckle due to the beamline optics. The conditions for successful coherence transport are considered.

As part of an effort to develop high resolution microtomography for engineered structures, a two-level copper integrated circuit interconnect was imaged using 1.83 keV x rays at 14 angles employing a full-field Fresnel zone plate microscope. A major requirement for high resolution microtomography is the accurate registration of the reference axes in each of the many views needed for a reconstruction. A reconstruction with 100 nm resolution would require registration accuracy of 30 nm or better. This work demonstrates that even images that have strong interference fringes can be used to obtain accurate fiducials through the use of Radon transforms. We show that we are able to locate the coordinates of the rectilinear circuit patterns to 28 nm. The procedure is validated by agreement between an x-ray parallax measurement of 1.41±0.17 μm and a measurement of 1.58±0.08 μm from a scanning electron microscope image of a cross section.

Speckles are produced by scattering of coherent light from a disordered system. Until recently, mainly pinhole setups have been used in x-ray speckle experiments. The full-width at half-maximum (FWHM) size of a speckle is inversely proportional to the diameter of the illuminated spot D on the sample [1]. Using focusing optics (e.g., a zone plate), D can be decreased to submicron size. In this work, we compared the dependence of the speckle size and contrast on D for the common pinhole setup [2, 3–11] with a focusing setup using a zone plate [12].

We measured static x-ray speckle contrast variation with the incident photon energy across a sample-specific absorption edge. In this paper, we present a theoretical description of this energy dependency consistent with our data. We found that the contrast depends mainly on the imaginary part of the complex index of refraction in the sample, as well as on the instrumental resolution. The speckle contrast decreases as the absorption cross section in the sample increases at the absorption edge. This result is not predicted by commonly used theory.

An integrated circuit interconnect was subject to accelerated-life test conditions to induce an electromigration void. The silicon substrate was removed, leaving only the interconnect test structure encased in silica. We imaged the sample with 1750 eV photons using the 2-ID-B scanning transmission x-ray microscope at the Advanced Photon Source, a third-generation synchrotron facility. Fourteen views through the sample were obtained over a 170° range of angles (with a 40° gap) about a single rotation axis. Two sampled regions were selected for three-dimensional reconstruction: one of the ragged end of a wire depleted by the void, the other of the adjacent interlevel connection (or “via”). We applied two reconstruction techniques: the simultaneous iterative reconstruction technique and a Bayesian reconstruction technique, the generalized Gaussian Markov random field method. The stated uncertainties are total, with one standard deviation, which resolved the sample to 200±70 and 140±30 nm, respectively. The tungsten via is distinguished from the aluminum wire by higher absorption. Within the void, the aluminum is entirely depleted from under the tungsten via. The reconstructed data show the applicability of this technique to three-dimensional imaging of buried defects in submicrometer structures relevant to the microelectronics industry.

We investigated the behavior of speckle contrast and size under various experimental conditions using 1.82 keV x-rays. In this paper, we report the comparison of two different setups for x-ray speckle experiments: one employing a focusing zone plate and one in which a pinhole selects the size of the coherent x-ray beam. We found a strong dependence of the speckle contrast and size on the type of setup. In general, the pinhole setup results in higher contrast but smaller speckle size. On the other hand, the zone plate setup allows one to target much smaller areas of interest in the sample, down to submicron dimensions, and also to adjust the speckle size. We anticipate that these results will be useful in future time-correlation spectroscopy experiments.

The authors measured static x-ray speckle contrast variation with the incident photon energy across sample-specific absorption edges. They propose that the variation depends strongly on the spectral response function of the monochromator. Speckle techniques have been introduced to the x-ray regime during recent years. Most of these experiments, however, were done at photon energies above 5 keV. They are working on this technique in the 1 to 4 keV range, an energy range that includes many important x-ray absorption edges, e.g., in Al, Si, P, S, the rare-earths, and others. To their knowledge, the effect of absorption edges on speckle contrast has not yet been studied. In this paper, they present their initial measurements and understanding of the observed phenomena.

We demonstrate quantitative noninterferometric x-ray phase-amplitude measurement. We present results from two experimental geometries. The first geometry uses x rays diverging from a point source to produce high-resolution holograms of submicrometer-sized objects. The measured phase of the projected image agrees with the geometrically determined phase to within +/-7%. The second geometry uses a direct imaging microscope setup that allows the formation of a magnified image with a zone-plate lens. Here a direct measure of the object phase is made and agrees with that of the magnified object to better than +/-10%. In both cases the accuracy of the phase is limited by the pixel resolution.

This letter describes quantitative nondestructive measurements of multilayer submicron Cu/SiO2 interconnect structures such as Cu lines, vias, and W lines with lateral dimensions down to 300 nm and electromigration defect structures using scanning transmission x-ray microscopy employing a 0.2 μm x-ray beam. Typical measurement accuracies are ⩽60 nm for widths and lengths and ⩽10% in height. The high-resolution and nondestructive nature of this technique provide a very powerful probe of physical properties of nanoscale and submicron materials and structures.

Artifacts induced by distortions which sometimes occur in two- dimensional projection images can appear in the resulting tomographic reconstructions. We describe a procedure for analyzing, correcting and removing experimental artifacts, and hence reducing reconstruction artifacts. Two-dimensional and three-dimensional images acquired with scanning transmission x-ray microscopy of a sample containing an integrated circuit interconnect show how these procedures can be successfully applied.

We have constructed a high resolution scanning x-ray microscope at the 2-ID-B beamline at the Advanced Photon Source for 1-4 keV x-ray imaging and microspectroscopy experiments. The microscope uses a Fresnel zone plate to focus coherent x-ray undulator radiation to a 150 nm focal spot on a sample. The spectral flux in the focus is 10(8) ph/s/0.1% BW. X-ray photons transmitted by the sample are detected by an avalanche photodiode as the sample is scanned to form an absorption image. The sample stage has both coarse and fine translation axes for raster scanning and a rotation axis for microtomography experiments. The incident x-ray beam energy can also be scanned via the 2-ID-B monochromator while the sample is kept in focus to record spatially resolved absorption spectra. We have measured the performance of the instrument with various test objects. The microscope hardware, software, and performance are discussed in this paper.