Preparation of Particles for Full-Range Tomography 1155
between Au and STO. Before segmentation, a 1.5 pixel Gaussian blur is applied to reduce the impact of Poisson noise. This noise reduction narrows the three peaks in the histogram. Missing wedge artifacts introduce spurious intensity
variations that impair the accuracy of threshold segmenta- tion, as illustrated in Figure 6a. In the Au/STO tomogram reconstructed with a ±75° tilt range (considered a good range for conventional electron tomography), the back- ground and STO peaks are broad and overlapping, pre- cluding accurate segmentation. The consequence is a mischaracterization of the specimen structure: edges are distorted so that flat faces appear curved and the STO wall separating the interior void is not resolved. These effects are especially severe for edges oriented in the direction of the missing wedge because essential information in that direc- tion is missing in Fourier space. Furthermore, shadow arti- facts that appear adjacent to highly scattering objects, such as the Au particles, could be mistaken as voids. For the full range tomogram, peaks for each material are relatively sharp and the STO peak is well separated from the background so segmentation will be accurate and robust to the choice of threshold. Peaks for the vacuum and weakly scattering car- bon are also resolved. The resulting segmentation accurately captures the structure of the Au/STO particle, including the surface geometry and void connectivity. Although the Au particles occupy too few voxels to
Figure 5. Visualization of Au/STO reconstruction. a: Three- dimensional volume rendering with orange indicating high inten- sity (gold nanoparticles) and purple indicating low intensity (STO). b–d: Two-dimensional slices through reconstruction that are perpendicular to specimen tilt axis selected from the left, mid- dle, and right of the reconstruction. Noise was reduced using a 1.25 pixel radius Gaussian blur. Dimensions of rendered volume are 158×203×215 nm.
materials in a tomogram, as this is an essential prerequisite for quantitative analysis. When possible, it is preferable to segment tomograms automatically rather than manually, as this removes a large amount of subjective judgment and improves the efficiency of analysis to provide better throughput and statistical sampling. The simplest automatic segmentation approach is to separate materials based on intensity thresholds, chosen either manually or with an algorithm such as Otsu’s method. This is particularly well suited to ADF STEM tomography, which provides atomic- number sensitivity. Accurate threshold segmentation requires that the histogram have distinct, well-separated peaks for each material of interest, and overlaps between the peaks will result in categorization errors for some voxels. We illustrate the impact of the missing wedge on seg-
mentation for the Au/STO tomogram in Figure 6. For this tomogram, we seek to identify three distinct materials: Au, STO, and background (including both void and carbon fiber support).We proceed by first separating all specimen material (Au and STO) from the background, and then distinguishing
tomography enable accurate quantification of awide variety of structural properties relevant to powder samples. For instance, surface geometry plays a central role in the performance of catalytic nanomaterials, as different facets and step edges provide different catalytic activity. Electron tomography has been used for quantitative analysis of surface geometry in heterogeneous catalysts and other nanomaterials (Jinnai et al., 2000; Ward et al., 2007). While missing wedge artifacts in conventional electron tomography introduce significant dis- tortions in the apparent surface geometry, full-range tomo- graphy allows the surface of a powder specimen to be more accurately identified and analyzed. To illustrate, we calculated the mean curvature of the STO surface in the Au/STO tomogram (Fig. 6c). An isosurface mesh was extracted from
appear as a distinct peak in the histogram, they have much higher intensity than the STO and are easily selected by a threshold. However, because STO falls at an intermediate intensity, “bubbles” of blurred intensity around the bright Au particles will be selected by the STO thresholds, and must be removed. One approach to correct this is the application of morphological filters to the binarized volumes. Using a 3.5-nm diameter spherical structure element, we applied a morphological close operation on the combined Au+STO component, followed by a morphological open on the STO component. This removes the “bubble” artifact around the Au particles, as well as some noise in binarized volumes. A 3D rendering of the resulting segmentation, shown in Figure 6b, demonstrates clean separation of the materials and accurate representation of the specimen morphology and void structure across the tomogram. The undistorted results of segmentation for full-range
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132 |
Page 133 |
Page 134 |
Page 135 |
Page 136 |
Page 137 |
Page 138 |
Page 139 |
Page 140 |
Page 141 |
Page 142 |
Page 143 |
Page 144 |
Page 145 |
Page 146 |
Page 147 |
Page 148 |
Page 149 |
Page 150 |
Page 151 |
Page 152 |
Page 153 |
Page 154 |
Page 155 |
Page 156 |
Page 157 |
Page 158 |
Page 159 |
Page 160 |
Page 161 |
Page 162 |
Page 163 |
Page 164 |
Page 165 |
Page 166 |
Page 167 |
Page 168 |
Page 169 |
Page 170 |
Page 171 |
Page 172 |
Page 173 |
Page 174 |
Page 175 |
Page 176 |
Page 177 |
Page 178 |
Page 179 |
Page 180
