Editing Scanning electron microscope (section)

==3D in SEM==
SEMs do not naturally provide 3D images contrary to [[Scanning probe microscope|SPMs]]. However 3D data can be obtained using an SEM with different methods as follows.

===3D SEM reconstruction from a stereo pair===
* [[photogrammetry]] is the most metrologically accurate method to bring the third dimension to SEM images.<ref name="Mignot-2018"/> Contrary to photometric methods (next paragraph), photogrammetry calculates absolute heights using [[triangulation (computer vision)|triangulation]] methods. The drawbacks are that it works only if there is a minimum texture, and it requires two images to be acquired from two different angles, which implies the use of a tilt stage. ([[Photogrammetry]] is a software operation that calculates the shift (or "disparity") for each pixel, between the left image and the right image of the same pair. Such disparity reflects the local height).

<gallery widths="340px" heights="220px">
File:SEM Stereo Pair.jpg|An SEM stereo pair of [[microfossils]] of less than 1&nbsp;mm in size ([[Ostracoda]]) produced by tilting along the longitudinal axis
File:SEM Stereo pair of micro-fossil (Juxilyocypris schwarzbachi Ostracoda).gif|From this pair of SEM images, the third dimension has been reconstructed by photogrammetry (using [[MountainsMap|MountainsSEM]] software, see next image); then a series of 3D representations with different angles have been made and assembled into a GIF file to produce this animation.
File:3D surface reconstruction from 2 scanning electron microscope images.gif|3D surface reconstruction of a (Ra = 3&nbsp;μm) [[Surface roughness|roughness]] calibration sample (as used to calibrate profilometers), from 2 scanning electron microscope images tilted by 15° (top left). The calculation of the 3D model (bottom right) takes about 1.5 second<ref>Stereo SEM reconstruction using MountainsMap SEM version 7.4 on i7 2600 CPU at 3.4 GHz</ref> and the error on the Ra roughness value calculated is less than 0.5%.
</gallery>

===Photometric 3D SEM reconstruction from a four-quadrant detector by "shape from shading"===
This method typically uses a four-quadrant BSE detector (alternatively for one manufacturer, a 3-segment detector). The microscope produces four images of the same specimen at the same time, so no tilt of the sample is required. The method gives metrological 3D dimensions as far as the slope of the specimen remains reasonable.<ref name="Mignot-2018"/> Most SEM manufacturers now (2018) offer such a built-in or optional four-quadrant BSE detector, together with proprietary software to calculate a 3D image in real time.<ref>{{cite journal|last1=Butterfield|first1=Nicholas|last2=Rowe|first2=Penny M.|last3=Stewart|first3=Emily|last4=Roesel|first4=David|last5=Neshyba|first5=Steven|title=Quantitative three-dimensional ice roughness from scanning electron microscopy|journal=Journal of Geophysical Research: Atmospheres|date=16 March 2017|volume=122|issue=5|pages=3023–3025|doi=10.1002/2016JD026094|ref=Butterfield2017|bibcode=2017JGRD..122.3023B|doi-access=free}}</ref>

Other approaches use more sophisticated (and sometimes GPU-intensive) methods like the [[optimal estimation]] algorithm and offer much better results<ref>{{cite journal|last1=Butterfield|first1=Nicholas|last2=Rowe|first2=Penny M.|last3=Stewart|first3=Emily|last4=Roesel|first4=David|last5=Neshyba|first5=Steven|title=Quantitative three-dimensional ice roughness from scanning electron microscopy|journal=Journal of Geophysical Research: Atmospheres|date=16 March 2017|volume=122|issue=5|pages=3025–3041|doi=10.1002/2016JD026094|ref=Butterfield2017|bibcode=2017JGRD..122.3023B|doi-access=free}}</ref> at the cost of high demands on computing power.

In all instances, this approach works by integration of the slope, so vertical slopes and overhangs are ignored; for instance, if an entire sphere lies on a flat, little more than the upper hemisphere is seen emerging above the flat, resulting in wrong altitude of the sphere apex. The prominence of this effect depends on the angle of the BSE detectors with respect to the sample, but these detectors are usually situated around (and close to) the electron beam, so this effect is very common.

===Photometric 3D rendering from a single SEM image===
This method requires an SEM image obtained in oblique low angle lighting. The grey-level is then interpreted as the slope, and the slope integrated to restore the specimen topography. This method is interesting for visual enhancement and the detection of the shape and position of objects; however the vertical heights cannot usually be calibrated, contrary to other methods such as photogrammetry.<ref name="Mignot-2018"/>
<gallery widths="220px" heights="160px">
File:FLY EYE.jpg|SEM image of a house fly compound eye surface at 450× magnification
File:Fly eye detail.jpg|Detail of the previous image
File:Fly Eye 3D SEM Image with form.jpg|SEM 3D reconstruction from the previous using [[shape from shading]] algorithms
File:Fly Eye 3D SEM Image without form.jpg|Same as the previous, but with lighting homogenized before applying the shape from shading algorithms
</gallery>

===Other types of 3D SEM reconstruction===
* Inverse reconstruction using electron-material interactive models<ref>{{cite conference
|last = Baghaei Rad
|first = Leili
|year = 2007
|title = Computational Scanning Electron Microscopy
|conference = International Conference on Frontiers of Characterization and Metrology
|volume = 931
|page = 512
|doi = 10.1063/1.2799427
|bibcode = 2007AIPC..931..512R
|doi-access = free
}}</ref><ref>{{cite journal |last1=Baghaei Rad |first1=Leili |last2=Downes |first2=Ian |last3=Ye |first3=Jun |last4=Adler |first4=David |last5=Pease |first5=R. Fabian W. |author-link5=R. Fabian Pease |year=2007 |title=Economic approximate models for backscattered electrons |journal=Journal of Vacuum Science and Technology |volume=25 |issue=6 |pages=2425 |bibcode=2007JVSTB..25.2425B |doi=10.1116/1.2794068}}</ref>
* Multi-Resolution reconstruction using single 2D File: High-quality 3D imaging may be an ultimate solution for revealing the complexities of any porous media, but acquiring them is costly and time-consuming. High-quality 2D SEM images, on the other hand, are widely available. Recently, a novel three-step, multiscale, multiresolution reconstruction method is presented that directly uses 2D images in order to develop 3D models. This method, based on a Shannon Entropy and conditional simulation, can be used for most of the available stationary materials and can build various stochastic 3D models just using a few thin sections.<ref>{{cite journal|last1=Tahmasebi|first1=Pejman|last2=Javadpour|first2=Farzam|last3=Sahimi|first3=Muhammad| title=Multiscale and multiresolution modeling of shales and their flow and morphological properties|journal=Scientific Reports|volume=5|pages=16373|doi=10.1038/srep16373|pmid=26560178|pmc=4642334|year=2015|bibcode=2015NatSR...516373T}}</ref><ref>{{cite journal|last1=Tahmasebi|first1=Pejman|last2=Javadpour|first2=Farzam|last3=Sahimi|first3=Muhammad| title=Three-Dimensional Stochastic Characterization of Shale SEM Images|journal=Transport in Porous Media|volume=110|issue=3|pages=521–531|doi=10.1007/s11242-015-0570-1|year=2015|bibcode=2015TPMed.110..521T |s2cid=20274015 }}</ref><ref>{{cite journal|last1=Tahmasebi|first1=Pejman|last2=Sahimi|first2=Muhammad| title=Reconstruction of three-dimensional porous media using a single thin section|journal=Physical Review E|volume=85|issue=6|pages=066709|doi=10.1103/PhysRevE.85.066709|pmid=23005245|year=2012|bibcode=2012PhRvE..85f6709T|s2cid=24307267 }}</ref>
* Ion-abrasion SEM (IA-SEM) is a method of nanoscale 3D imaging that uses a focused beam of [[gallium]] to repeatedly abrade the specimen surface 20 nanometres at a time. Each exposed surface is then scanned to compile a 3D image.<ref name="Murphy-2010">{{cite journal |last1=Murphy |first1=GE |last2=Lowekamp |first2=BC |last3=Zerfas |first3=PM |title=Ion-abrasion scanning electron microscopy reveals distorted liver mitochondrial morphology in murine methylmalonic acidemia. |journal=Journal of Structural Biology |date=August 2010 |volume=171 |issue=2 |pages=125–32 |doi=10.1016/j.jsb.2010.04.005 |pmid=20399866|pmc=2885563 }}</ref><ref name="nsf.gov">{{cite web |title=Multimedia Gallery - 3-D Imaging of Mammalian Cells With Ion-Abrasion SEM {{!}} NSF - National Science Foundation |url=https://www.nsf.gov/news/mmg/mmg_disp.jsp?med_id=64543 |website=www.nsf.gov}}</ref>

===Applications of 3D SEM===
One possible application is measuring the roughness of ice crystals. This method can combine variable-pressure environmental SEM and the 3D capabilities of the SEM to measure roughness on individual ice crystal facets, convert it into a computer model and run further statistical analysis on the model.<ref>{{cite journal|last1=Butterfield|first1=Nicholas|last2=Rowe|first2=Penny M.|last3=Stewart|first3=Emily|last4=Roesel|first4=David|last5=Neshyba|first5=Steven|title=Quantitative three-dimensional ice roughness from scanning electron microscopy|journal=Journal of Geophysical Research: Atmospheres|date=16 March 2017|volume=122|issue=5|pages=3023–3041|doi=10.1002/2016JD026094|ref=Butterfield2017|bibcode=2017JGRD..122.3023B|doi-access=free}}</ref>  Other measurements include fractal dimension, examining fracture surface of metals, characterization of materials, corrosion measurement, and dimensional measurements at the nano scale (step height, volume, angle, flatness, bearing ratio, coplanarity, etc.).{{Citation needed|date = February 2016}}

SEM is also used by [[art conservation]]ists to discern threats to paintings' surface stability due to aging, such as the formations of complexes of [[zinc]] ions with [[fatty acid]]s.<ref>{{Cite journal |last1=Hermans |first1=Joen |last2=Osmond |first2=Gillian |last3=Loon |first3=Annelies van |last4=Iedema |first4=Piet |last5=Chapman |first5=Robyn |last6=Drennan |first6=John |last7=Jack |first7=Kevin |last8=Rasch |first8=Ronald |last9=Morgan |first9=Garry |last10=Zhang |first10=Zhi |last11=Monteiro |first11=Michael |date=June 2018 |title=Electron Microscopy Imaging of Zinc Soaps Nucleation in Oil Paint |url=https://www.cambridge.org/core/journals/microscopy-and-microanalysis/article/abs/electron-microscopy-imaging-of-zinc-soaps-nucleation-in-oil-paint/B442AD4847D1ABF091A4A43CA4C0E2A2 |journal=Microscopy and Microanalysis |language=en |volume=24 |issue=3 |pages=318–322 |doi=10.1017/S1431927618000387 |pmid=29860951 |bibcode=2018MiMic..24..318H |s2cid=44166918 |issn=1431-9276}}</ref> Forensic scientists use SEM to detect [[Art forgery|art forgeries]].