EM Theoretical performance vs Sample dependent performance

Welcome to the Resolve blog. My hope for this series of blogs is to dwell on the most important considerations as a buyer when choosing an electron microscope. Over the next few months I’ll be focusing on what I believe to be the most important criteria when trying to choose a microscope.

First, the most important step is to establish the range of samples the machine will most likely be imaging and what kind of imaging techniques will be used for those samples. However, as consultants, we often find that the people placed in charge of the purchase always have a bias towards their own work and are therefore blinkered by their own requirements. As a result they purchase a machine specifically for their imaging purposes, which limits the usability of the instrument. Only a totally independent person will be able to correctly assess the machine requirements and then create an unbiased strategy for testing the most suitable microscopes available.

The first subject to be discussed in this series of blogs is one very important specification on a machine i.e. the performance specifications. This is a very important aspect that the buyer should be aware of, as there is a huge difference between theoretical performance and sample dependent performance.

Electron microscopes, in similar price ranges, perform equally to each other on paper. In these circumstances the performance is measured based on a standardized sample. Standardized samples provide ideal qualities for imaging. They are electrically conductive, have high contrast against a low contrast background and are very stable under vacuum. One such example is the gold particle on carbon sample. The gold particles, with superior feature details, generate high secondary and/or backscatter electron signals with excellent contrast against the carbon background.

Generally, performance is also measured using ideal imaging conditions minimizing any possible distortions that could affect the image quality.

  1. Machine will be placed in a vibration free environment
  2. Filament saturation and optimum signal to noise (on a tungsten machine)
  3. High keV (minimizing lens aberrations in SEM & TEM)
  4. High vacuum (a clean sample and high vacuum can influence imaging in both SEM & TEM)
  5. Shortest working distance
  6. Smallest spot size (using samples with a high signal to noise)
  7. Smallest aperture (using samples with a high signal to noise)

These types of imaging conditions do not really reflect those used to image a variety of samples on a day-to-day basis on a microscope therefore it is important to test the machine using realistic samples that are going to be imaged. Additionally the imaging conditions used should be considered when testing the resolution capabilities of the machine.  For example, with SEM, if sample surface imaging is the goal then assessing resolution at low keV would be an absolute requirement when testing the microscopes’ resolution capabilities. In TEM, if low dose imaging was the goal, then testing which machine that generates the greatest image contrast with a known sample would be one way of evaluating its practical capability. These are just a few examples that should be on the list for consideration.

One other consideration is what types of samples should be chosen for the test. Time allocation on a demonstration day will be limited therefore; it is crucial to find a few samples (no more than 6) that would cover the range of imaging conditions and sample behavior to get a reasonable idea of the practical resolution capabilities of the machine. Generally the samples should reflect day-to-day examples and each one should be chosen to test imaging performance under different conditions. As a guide it is essential to choose samples that will push the capabilities of the vacuum system, detectors, and sample holder/stage robustness and accuracy. This should weed out systems that may not be optimal for your needs. Other samples for particular specialized applications can also be included at a later stage before the purchasing negotiations begin.

In conclusion, the message that I’m trying to put across here is that under ideal imaging conditions using ideal samples many machines perform equally when compared to each other. But, their strengths and weaknesses can be tested using real day-to-day samples and imaging with more realistic imaging conditions. These particulars can be exploited by developing a logical and defined testing plan before the demonstration and will be invaluable to finally determine which machine to eventually purchase.