Potential Large-area Imaging of Butterfly Wing Scales with Transmission Electron Microscopy

The complexity of self-assembled photonic structures like butterfly wing scales spans 3–4 orders of magnitude in length scales. These monolithic structures produced by single cells on a butterfly wing typically measure a few hundred micrometres across, with intricate features going well below the nanometre scale. While this complexity makes the self-assembly of these structures fascinating, imaging and capturing this gamut of complexity is extremely challenging. The available imaging methods can either capture the overall field of view at a lower resolution (scanning electron microscopy (SEM)) or can capture high-resolution details of the structure at a specific region (resin section transmission electron microscopy) [1]. Further, three-dimensional (3D) characterization techniques either can only image the topology of a single surface (atomic force microscopy) [1] or are limited to fewer synchrotron facilities (ptychographic x-ray tomography (PXCT)) [2], rendering them impractical to promptly test a large number of scales for gene expression studies. Automation is key to imaging large complex structures like the butterfly scale; the advent of cryo-electron microscopy (cryo-EM) aided the development of automation tools for the transmission electron microscopy (TEM) [3]. Here we propose an imaging framework which utilizes the new possibilities of TEM provided by such automation tools, energy filters and direct electron detectors to rapidly image an extended field of view of photonic crystals in butterfly wing scales at nanoscale resolution. In addition, we demonstrate that the application of pop-out 3D metrology in this framework provides 3D morphological insights [4]. We imaged butterfly scales by placing them on a cryo-EM grid, as their extended laminar structure is suitable for imaging in a TEM. The experimental micrographs in Fig. 1 show that even at higher magnification (high-mag), the scales are stable and provide high-resolution images revealing vein-like features that have not been recorded before. Further, using automation tools in TEM, a larger scale area can be scanned by continuously capturing a TEM movie while the field of view moves across a programmed path. The coordinates of the shift position and the time point data are used to stitch the frames together to obtain a large-area high-resolution TEM map. In our previous work, we showed that 3D imaging is possible with a single energy-filtered TEM micrograph (Pop-out 3D metrology) [4]. Applying pop-out on the stitched large-area map provides 3D insights such as the local tilt map of the lower lamina, the height distributions of ridges and cross-ribs, and the separation distances between the upper and lower lamina. We show a numerical implementation of the proposed framework in Fig. 2. A series of TEM frames are simulated using a PXCT reconstruction of a pansy butterfly’s (Junonia orythia) male blue scale followed by a pop-out reconstruction on the stitched TEM map. Though the reconstruction is affected by artefacts, the cross-section of the reconstruction (Fig. 2f), when compared with the ground truth (Fig. 2e) and a similar experiment resin section TEM cross-section (Fig. 2d), shows that the upper and lower lamina separation is captured well by pop-out. The proposed framework is non-destructive, does not require extensive sample preparation, and images the scales in their native state. The results from numerical experiments show that the pop-out reconstruction can provide 3D insights useful for further modelling to study these nanostructures' optical and structural characteristics. a) Junonia butterfly scales placed on a cryo-TEM grid. The inset shows a male blue scale on the copper mesh; b) a low-mag TEM image of a blue scale and the copper mesh; c) a medium-mag TEM image of the scale on the carbon grid; d) and e) high-mag TEM images of the scale, which show vein-like features inside cross-ribs. a) TEM images simulated as if they are scanned over a 4x4 µm2 area of an experimental PXCT reconstruction and stitched together; b) the ground truth 3D morphology of the scale obtained from the PXCT reconstruction; c) the pop-out 3D reconstruction from the stitched TEM images in (a); d) a resin section TEM image (Experiment); e) and f) cross-sections from the ground truth (e) and the pop-out 3D reconstruction (f).