Figure 1

(Color online) The escape depth of secondary photoelectrons at the $\mathrm{C}K$-edge is $\sim 3\mathrm{nm}$, therefore only one layer of nacre tablets or prismatic crystals at a time is probed. The curve is an exponential fit to the experimental data points. Data adapted from Frazer et al. (Ref. 22).Reuse & Permissions

Figure 2

(Color online) The geometry of the X-PEEM spectromicroscopy experiment at SRC in top view is reported on a gray background. Notice that the x rays illuminate the magenta sample at 16° grazing incidence and the electron optics axis is parallel to the sample normal. The geometry of the XANES experiment at the ALS is reported on a white background. In this case, we present a front view of the single-crystal sample, with its $c$-axis in plane, and the azimuthal rotation afforded by the manipulator. Notice that the x-ray beam illuminates the sample surface at normal incidence. In both experiments, the polarization is linear and horizontal (in the orbit plane), and represented by a purple vector.Reuse & Permissions

Figure 3

(Color online) (a) Carbon ${\pi }^{*}∕{\sigma }^{*}$ map of the polished surface of red abalone nacre, showing polarization-dependent imaging contrast (PIC). The map was obtained by digital ratio of two SPHINX micrographs, acquired at 290.3 and $302\mathrm{eV}$, respectively, the ${\pi }^{*}$ and ${\sigma }^{*}$ peak energies shown in (b). The immediately adjacent tablets labeled 1 and 5 exhibit the highest contrast. Different gray levels correspond to different crystal orientations. (b) Carbon spectra extracted from the correspondingly labeled regions in (a). The intensities of the ${\pi }^{*}$ and ${\sigma }^{*}$ peaks are anticorrelated and vary dramatically from tablet to tablet, generating the contrast seen in (a), which in turn depends on crystal orientation and polarization. These and all other C spectra were normalized to a beamline background, and then to a linear fit above $310\mathrm{eV}$, so that the spectral intensities below $283\mathrm{eV}$ and above $310\mathrm{eV}$ are 0 and 1, respectively. (c) Oxygen map obtained by digital ratio of images at 534 and $531.7\mathrm{eV}$, the ${\pi }^{*}$ and pre-edge energies, respectively. (d) Oxygen spectra from the same regions, reproducing the anticorrelation of the ${\pi }^{*}$ and ${\sigma }^{*}$ peaks seen for carbon. All O spectra were normalized to a beamline background, and then to a linear fit above $555\mathrm{eV}$ so that spectral intensities below $530\mathrm{eV}$ and above $555\mathrm{eV}$ are 0 and 1, respectively. (e) Calcium map of the same nacre area, obtained by digital ratio of images at 352.7 and $343.7\mathrm{eV}$. This map represents only a distribution map, showing much greater Ca concentration in the tablets compared to the organic matrix layers separating them. The polarization dependence observed in (a) and (c) is not detected in this map. (f) Calcium spectra extracted from the regions labeled in (a), showing that all regions examined are aragonite. All Ca spectra were normalized to a linear background, representative of the beamline background, and then to a linear fit above $355\mathrm{eV}$ so that spectral intensities below $346\mathrm{eV}$ and above $355\mathrm{eV}$ are 0 and 1, respectively. The nacre growth direction of this sample is perpendicular to the sample surface, therefore the organic matrix layers are parallel to the sample surface. As they are not perfectly flat, they appear as contour lines.Reuse & Permissions

Figure 4

(Color online) (a) Aragonite and (b) calcite $\mathrm{Ca}L$-edge XANES spectra acquired on geologic aragonite and calcite crystals (red curves). These spectra are dramatically different in the crystal field resonance regions highlighted by the gray boxes. They are, however, very similar to the spectra of nacre and prismatic crystals, respectively (black curves).Reuse & Permissions

Figure 5

(Color online) (a) Carbon ${\pi }^{*}∕{\sigma }^{*}$ map from the surface of a polished cross section of red abalone prismatic layer. (b) Carbon $K$-edge XANES spectra extracted from the correspondingly labeled regions in (a) exhibit anticorrelated intensity of the ${\pi }^{*}$ and ${\sigma }^{*}$ peaks. (c) Oxygen map of the ${\pi }^{*}$/pre-edge signal at 534 and $528\mathrm{eV}$, respectively. (d) Oxygen $K$-edge XANES spectra. (e) Calcium distribution map of the same prismatic region, obtained by a digital ratio of images acquired at $351.4\mathrm{eV}$ (on the calcite peak) and $344.4\mathrm{eV}$. (f) Calcium $L$-edge XANES spectra extracted from regions labeled in (a), showing that all regions examined are calcite.Reuse & Permissions

Figure 6

(Color online) (a) Carbon spectra acquired from a single geologic aragonite crystal as the polar angle $\theta$ was varied. Only a subset of the spectra was displayed and offset for clarity. (b) FEFF simulated C spectra for aragonite at different polar angles. While the lineshape does not accurately portray the experimental data, the anticorrelation between the ${\pi }^{*}$ and ${\sigma }^{*}$ peaks is consistent in the experimental and simulated spectra. (c) Oxygen spectra from another single crystal of aragonite. (d) FEFF simulated O spectra. (e) Polar dependence (black dots) of the ${\pi }^{*}$ peak, obtained by measuring, in the raw data of Fig. 7 and others, the intensities at 290.3 and $289.5\mathrm{eV}$. These are the ${\pi }^{*}$ peak and the immediate pre-peak intensity, respectively, shown in the inset. Due to intensity variations in the unnormalized data, each of the ${\pi }^{*}$ peak intensities had to be scaled to its pre-peak intensity. The parameter displayed, therefore, is not $I\left({\pi }^{*}\right)-I\left(\text{pre-peak}\right)$ but $[I\left({\pi }^{*}\right)-I\left(\text{pre-peak}\right)]∕I\left(\text{pre-peak}\right)$. A direct measurement of the error bars was obtained from three repeated acquisitions returning to $\theta =-11°$ and $\theta =22°$. Theory predicts a variation of the ${\pi }^{*}$ intensity with a ${\cos }^2\theta$ dependence (Refs. 14, 15). A fit of the experimental data (red curve) reveals that the dependence is indeed $[I\left({\pi }^{*}\right)-I\left(\text{pre-peak}\right)]∕I\left(\text{pre-peak}\right)=\alpha {\cos }^2\theta$. The fit parameter $\alpha$ is 0.9 for carbon in aragonite. (f) The polar dependence for the normalized oxygen spectra in (c), and $I\left({\pi }^{*}\right)-I\left(\text{pre-edge}\right)=\alpha {\cos }^2\theta$ fit (red curve), with $\alpha =10$ for O in aragonite.Reuse & Permissions

Figure 7

(Color online) Selected raw data XANES spectra acquired at different polar angles. Even in these raw data the decrease of ${\pi }^{*}$ peak intensity with $\theta$ is evident. Again, only selected spectra are shown for clarity. The only processing here was vertical displacement and scaling, necessary because the spectra had wide variations in intensity. The black curve, labeled $I_o$, is the beamline background curve, acquired simultaneously with the $\theta =0$ spectrum. Each of the spectra in Fig. 6a was normalized to a different $I_o$, simultaneously acquired. The polar variation of Fig. 6e was obtained from these and many other spectra.Reuse & Permissions

Figure 8

(Color online) (a) Carbon spectra of a single crystal of geologic calcite at different polar angles. (b) FEFF simulated carbon spectra of calcite. Oxygen (c) experimental and (d) simulated spectra. (e) Carbon and (f) oxygen polar dependence data (black dots) and fit (red curves). Again, the two data points at $\theta =90°$ for C and $\theta =-10°$ for O indicate the experimental uncertainty. As in Fig. 6, we report $[I\left({\pi }^{*}\right)-I\left(\text{pre-peak}\right)]∕I\left(\text{pre-peak}\right)$ for C and $[I\left({\pi }^{*}\right)-I\left(\text{pre-edge}\right)]$ for O spectra. As for aragonite, also for calcite an acceptable fit is obtained by $\alpha {\cos }^2\theta$ with $\alpha =1$ for carbon and $\alpha =10$ for oxygen.Reuse & Permissions

Figure 9

(Color online) (a) Composite VLM images of a polished cross section of red abalone, acquired in differential interference contrast mode. The green circles indicate prismatic areas examined with SPHINX. (b) A series of ${\pi }^{*}$/pre-edge O maps of the areas shown in (a). (c) VLM image taken with crossed polarizers on the same red abalone shell. The prismatic crystals appear on top in various shades of brown, the nacre is cyan, at the bottom.Reuse & Permissions

Figure 10

Single-energy SPHINX image at $344\mathrm{eV}$, carbon and calcium maps, taken near a growth layer in polished red abalone nacre. The carbon ${\pi }^{*}∕{\sigma }^{*}$ map is the ratio of two images taken at 290.3 and $302\mathrm{eV}$, respectively, and the calcium map is the ratio of 352 and $344\mathrm{eV}$ images. In this 90° cross section, the growth direction was in plane, at 74° from the polarization vector. Notice the PIC imaging contrast in the carbon map, while the direct image and the Ca map do not show any contrast. Dark Ca-poor spots in the Ca map appear at the sides of each tablet, where organics were located.Reuse & Permissions

Figure 11

(Color online) (a) Visible light microscopy composite micrographs of a polished section of red abalone shell, acquired in differential interference contrast mode. The nacre growth direction was 30° from the sample surface normal. The line between two magenta arrowheads is the prismatic-nacre boundary. The black rectangles outline the areas in which the O maps in (b) were acquired with SPHINX. (b) Series of oxygen PIC maps, obtained by digital ratio of images acquired at 534 and $530\mathrm{eV}$, ${\pi }^{*}$ and pre-edge, respectively. PIC is clearly detectable across the entire thickness of the nacre layer. The central PIC map was acquired across one of the vertical lines evident in the optical image, which do not appear to affect tablet order or contrast. To enhance contrast in bulk nacre, all raw grayscale maps were leveled and converted to a blue quadtone.Reuse & Permissions

Figure 12

(Color online) (a) X-ray diffraction pole figure indicates random orientation of nacre tablets in the $ab$-plane. (b) Rocking curve taken on the (002) Bragg peak on a piece of red abalone nacre. The 22° footprint of the rocking curve indicates that the $c$ axis variation in bulk nacre varies by $\pm 11°$, with instrument resolution better than 0.1°.Reuse & Permissions