SEM samples are different in height, ranging from several millimetres to tens of millimetres, and the vertical position of the SEM stub can be changed if we wish to. How do we determine the actual length of things that we are interested in since they are enlarged in the picture? Are the scale at the bottom of each picture and the measurement tool provided by the SEM software reliable?
I doubt its accuracy because the surface of the sample is unknown to the machine.
Modern SEMs have "brains", they calculate working distance (9.1 and 9.2 on your pictures) and in accordance with it change magnification and length of a scale bar. So, no problems with height of a specimen. For as long as SEM was not badly mishandled you can expect accuracy of a scale bar to be not worse than 10% (and usually much better). If you want better accuracy, you need to calibrate your SEM. For this you can buy special calibrating specimen or to ask your service engineer during his next visit to make a calibration for you (service engineer should have a calibrating specimen). Now you can perform measurements on a flat specimen with high precision. For rough, 3D surfaces, things are complicated, with some exceptions like diameter of a fiber. Whatever is orientation of a fiber, it's thickness (means diameter) is always the same and easy to measure. Much worse with fiber's length. If, for example, fiber belongs to a plane which makes 45 degrees with an electron beam, then (with help from Pythagoras) we can calculate that length of a fiber we measure from a picture is just 1/\sqrt{2} of its real length. There are software packages that can perform 3D measurements for two (or more) SEM pictures, but they are rather expensive.
Fortunately (or unfortunately?) most researchers report just measurements they made from pictures, not paying any attention to Pythagoras. And why not? Most reviewers do not know better too.
Modern SEMs have "brains", they calculate working distance (9.1 and 9.2 on your pictures) and in accordance with it change magnification and length of a scale bar. So, no problems with height of a specimen. For as long as SEM was not badly mishandled you can expect accuracy of a scale bar to be not worse than 10% (and usually much better). If you want better accuracy, you need to calibrate your SEM. For this you can buy special calibrating specimen or to ask your service engineer during his next visit to make a calibration for you (service engineer should have a calibrating specimen). Now you can perform measurements on a flat specimen with high precision. For rough, 3D surfaces, things are complicated, with some exceptions like diameter of a fiber. Whatever is orientation of a fiber, it's thickness (means diameter) is always the same and easy to measure. Much worse with fiber's length. If, for example, fiber belongs to a plane which makes 45 degrees with an electron beam, then (with help from Pythagoras) we can calculate that length of a fiber we measure from a picture is just 1/\sqrt{2} of its real length. There are software packages that can perform 3D measurements for two (or more) SEM pictures, but they are rather expensive.
Fortunately (or unfortunately?) most researchers report just measurements they made from pictures, not paying any attention to Pythagoras. And why not? Most reviewers do not know better too.
- Badges
- Science method