banner image
No data available.
Please log in to see this content.
You have no subscription access to this content.
No metrics data to plot.
The attempt to load metrics for this article has failed.
The attempt to plot a graph for these metrics has failed.
Very low angle annular dark field imaging in the scanning transmission electron microscope: A versatile tool for micro- and nano-characterization
Rent this article for
View: Figures


Image of FIG. 1.
FIG. 1.

(Color) Comparison of HAADF and VLAADF. In HAADF imaging (a), the camera length (i.e., the magnification of the diffraction pattern) is chosen so that only diffracted beams with large Bragg angles θB > 50 mrad are collected by the collector. In VLAADF imaging (b), the camera length (and thus the magnification of the diffraction pattern) is increased until the zero beam fills the detector opening almost entirely. That will allow only beams with the smallest Bragg angles (open circles) to strike the detector, thus forming the image. An (optional) objective aperture can be used to choose an upper limit for the contributing beams, typically θB > 25 mrad. Both HAADF and VLAADF settings can be realized with the same detector.

Image of FIG. 2.
FIG. 2.

(Color online) Typical setting for VLAADF imaging in the Tecnai F20. The figure shows the electron diffraction pattern of a silicon sample at the 110 pole, with camera length set to 771 mm, and a nominal probe size of 0.7 nm. The bright ring in the center of the inner orange ring is the zero beam with a convergence semiangle of 6 mrad. The inner orange ring depicts the inner radius of the annular dark field detector, the outer ring a 100 μm objective aperture. Thus, only the beams between the two orange rings contribute to the image, namely the {111}, {220}, {311}, and {400} reflections in silicon.

Image of FIG. 3.
FIG. 3.

Direct comparison between HAADF (a) and VLAADF (b) imaging of a modern Cu interconnect structure. Both images show the same area of the sample, which contains Cu lines, pads, and vias. While different grains can hardly be distinguished in the conventional HAADF image, the VLAADF image depicts every grain with a different gray level. It should be pointed out that the left two metal lines are each filled with a single grain; thus they do not show any contrast variation in the VLAADF image.

Image of FIG. 4.
FIG. 4.

VLAADF image of experimental three level Cu interconnect structure in 32 nm technology. Metal1 and metal 3 lines are viewed end-on, while metal 2 is imaged parallel to the line. The sample was prepared by focused ion beam (FIB) lift-out. Imaging conditions are the same as for Fig. 2. Therefore, only the {111}, {200}, and {022} reflections in Cu contribute to the grain contrast.

Image of FIG. 5.
FIG. 5.

(a) VLAADF composite image of 4 μm Cu line, viewed from the side in a sample prepared by FIB lift-out. The image is stitched together from several separately acquired images in order to retain adequate resolution. (b) Detail of the Cu line imaged in (a). Grain size distributions are obtained by marking the intercepts of a horizontal line with the grain boundaries found in the Cu microstructure (Heyn’s technique). Depending on the position of the line, one can obtain grain size information from the bottom, the middle, or the top of the metal line.

Image of FIG. 6.
FIG. 6.

(Color) Since VLAADF captures virtually every grain in the Cu line, statistics become good enough to capture even subtle effects. Panel (a) depicts the difference in grain size between the bottom of the Cu line and the top of the Cu line. While small grains dominate at the bottom of the line, more medium sized grains are found at the top of the structure. Panel (b) shows the difference in grain size distribution for two different Ta/TaN liner processes before Cu seed layer deposition. Liner process A clearly produces more small grains in the 20–30 nm range than liner B, while liner B renders more grains in the 70–90 nm range than liner A.

Image of FIG. 7.
FIG. 7.

(Color online) Comparison of (a) HAADF and (b) VLAADF imaging of a PCM cell after a programming pulse. The programming pulse transforms phase-change material near the heater into the amorphous state. (a) In the HAADF image, the Ge2Sb2Te5 (GST) above the heater does not look any different compared to the rest of the GST. (b) In the VLAADF image, however, the amorphous dome over the heater is revealed as an area with completely homogeneous, constant gray level contrast, embedded in poly-crystalline GST. The dashed line around the dome is a guide for the eyes. The white specks in the GST are nano-crystallites of GST, which happen to have a set of lattice planes in Bragg reflection toward the detector. The VLAADF image also highlights the bamboo structure of the material of the top electrode.

Image of FIG. 8.
FIG. 8.

(Color online) Root-mean-square contrast of Cu grains for different settings of the STEM collection angle. The rms-contrast was determined from ∼40 000 pixels in identical areas of the same Cu line, delineated by the dashed lines. The lower cut-off angle is defined by the inner radius of the annular dark field detector. The two insets show the STEM images for VLAADF and for HAADF imaging, respectively. The grain contrast increases about twofold as soon as the VLAADF regime (θB < 20 mrad) is entered.

Image of FIG. 9.
FIG. 9.

VLAADF images for five different settings of the beam convergence semiangle α. For α < 6 mrad, the diffraction disks of the Cu grains do not overlap, (a) and (b). For α = 6.01 mrad, the {111} diffraction disks are just touching the zero-beam, (c). For (d) and (e), the diffraction disks overlap to a considerable degree. Doubling the convergence angle from (a) to (e), and thus moving from nonoverlap to overlapping disks on the detector, hardly changes the appearance of the VLAADF image. The rms-contrast varies less than 3% between the images.


Article metrics loading...


Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: Very low angle annular dark field imaging in the scanning transmission electron microscope: A versatile tool for micro- and nano-characterization