In this interview, Dr. Jacob P. Hoogenboom, from the Faculty of Applied Sciences at TU Delft, talks to AZoM about the development of electron microscopy with FAST-EM, new automated ultra-fast multibeam electron microscope.
Could you give a brief introduction to your group and what you are working on?
I am an associate professor at Delft University of Technology, and I am leading my own research group. I am also heading a section on microscopy instrumentation and techniques, consisting of five principal investigators like myself.
My group is developing novel techniques and novel instruments for microscopy, specifically at the boundary between light and electron microscopy. We try to bridge the limitations of both individual modalities by combining the two in a single system. That offers us new capabilities for imaging that go beyond what is currently possible.
Could you tell us a bit about the consortium that developed the FAST-EM system and the role that the Delft University of Technology played in this?
The consortium aims to make a commercial product out of the prototype multi-beam scanning electron microscope (SEM) developed here at TU Delft. This development has been ongoing for a long time, from before I started at TU Delft.
The group of Pieter Kruit, who is also part of the same section, started developing a multi-beam electron source and microscope around 2000, at that time mostly for electron lithography applications, so not yet for imaging.
Over the past years, our group, especially in the work I was doing, looked more at biological applications. We became interested in multi-beam imaging for biological research questions. When I started at TU Delft about 12 years ago, we developed an integrated microscope which combined a light and electron microscope, and from that work, Delmic spun out.
Later, we realized that the integrated microscope could also be used as a detector for a multi-beam SEM specifically for biological samples. This was then incorporated in the prototype that had shown initial results and for which by then also other forms of detection and imaging had been explored. The detection based on the integrated microscope turned out to be very successful. The consortium was founded to build a first functional model and later a commercial product.
Besides TU Delft and Delmic, this consortium also included Technolution and Thermo Fisher, formerly FEI, with whom we already had a very long-standing relationship.
What are some of the limitations of current scanning electron microscopes?
Electron microscopy is the go-to technique for imaging materials at the highest possible resolution. For a scanning electron microscope, this is about one nanometer. That resolution is reached by focusing a beam on a small area and then collecting the signal from that single position.
Now, you have to acquire a signal for some time to get a detectable signal that you can translate into a pixel and move on to the next pixel. A major limitation comes from the fact that if you look at things with those length scales, with such a pixel size, this takes too long to go to macroscopic length scales.
In a reasonable time, the area you can scan is very small, typically on the order of hundreds of micrometers-squared. This is a disadvantage that a multi-beam microscope tackles.
One other limitation with electron microscopy is that it can only image fixed samples - dead materials and no life or dynamic processes.
How did you consider these limitations when developing FAST-EM?
FAST-EM is really targeted to the throughput limitation - the limitation in the acquisition time of images. You could go about this in two ways. One could reduce the time you need to scan a single pixel, and we are looking into approaches to do that. The detector we developed for FAST-EM tackles that because it allows a little faster imaging per pixel than regular electron detectors would do.
However, throughput would go up the most if you could multiply the number of beams, as then you multiply the area you scan by the number of beams. For example, if you can scan with 100 beams, you can go 100 times faster than with a single beam system.
One of the main realizations that underlie this, which was already done in the work of Pieter Kruit, was that if you have a scanning electron microscope, only a fraction of the current you draw from the source, goes into the electron beam to do the imaging.
For this, you use a very small aperture, which leads to a very small opening angle of the beam in the electron microscope. This is different from light microscopy, where typically you use a wide opening angle. The aperture needs to be small because electron microscopy is limited by the aberrations you induce if you have a large opening angle beam.
You throw away a lot of the current that is extracted from the source when you put this aperture immediately behind the source. The main idea is that if you put an array of apertures instead, each aperture creates a beamlet with a small opening angle. So, you can use current you would otherwise throw away and you have a multitude of beams that are focused by the electron microscope onto the sample.
FAST-EM is also an automated electron microscope. How is this beneficial when imaging and analyzing large and multiple samples?
Some scanning electron microscopes already operate in a semi-automated fashion. But typically, if you scan an area, you have to move the sample or the beam to scan the next area.
Image Credit: Delmic
If you want to make a composite image, all these areas need to have a small overlap region, and then they have to be stitched together. The FAST-EM system already has all the scripts to do that. Navigating from one visual section to the next is already included in its user software.
If an operator has to select the next region or the next section, you gain some throughput with the multibeam, but you still lose time in the manual operations. This then becomes the next bottleneck. I think it is a logical step that you should go to an automated acquisition if you speed up the acquisition. It becomes very boring work if the operator constantly has to change things manually.
Discover the FAST-EM Multibeam Electron Microscope here
FAST-EM could have many applications in life sciences. Could you describe some of its main uses?
In biomedical sciences, most electron microscopy is carried out in imaging facilities. Here you have an experienced facility with operating staff and people that help you prepare your sample, do the imaging, or help you with the imaging.
Most of these facilities are limited by the number of samples they can handle. I think that FAST-EM can lift this limitation. It can allow a facility to process more samples from a faculty or out of a hospital.
The other thing is that at present, as it takes so long to scan a region, the operator will select an area to scan and give that data to the customer. With FAST-EM, you can acquire the full sample and send that data to the user or share it with the user. The user can browse through the data as if he is behind a microscope.
The user can also look outside the region of interest, and others can look to that region of interest. It can be shared between people. The data can become open, available to people who do not have the opportunity to use an electron microscope or to prepare those samples. It is a prerequisite for discovery to look beyond what you are actually looking for.
More samples can be processed. This is important in cases such as where you have a disease type and looking for deviations. Here, it is important to compare data between human or animal models or different cell or tissue types to reach a conclusion and look at inter-individual variability. This is something that could be done with this system.
In the time you would normally acquire one sample, you can now do 100. You can get statistics on this. That is one application area which I think will become really important.
The other area that people are exploring is volume imaging. Typically, you section a sample for electron microscopy, and if you collect sequential sections and you keep track of the order of the sections, then by scanning them, you can reconstitute a 3D image.
This has been explored in the past years, and the number of people using such techniques has increased drastically, but throughput is a bottleneck. The volume is typically limited.
With FAST-EM, we can go to larger volumes. For instance, it could be used for a full model animal brain to map out all the neurons' connections. There have been initial results with existing machines on this, which typically have over a year's acquisition times.
This can now be reduced to a few days, which means that you can track multiple specimens, multiple model systems, multiple animal brains, and get to comparative analysis.
Do you believe that this microscope could help overcome some of the challenges faced by other electron microscopes and provide answers to scientific problems left unanswered till now?
There have also been other attempts to multi-beam electron microscopy, and there are also people that try to use, for instance, a few microscopes parallel to each other.
There is a big advantage to doing everything in one microscope because it is not realistic to have 100 parallel electron microscopes, especially when the price of FAST-EM is only around twice the price of a regular SEM. That is a significant difference, given it competes with 100 microscopes in parallel.
Image Credit: Islet of Langerhans imaged with FAST-EM. Sample courtesy of the Giepmans lab, University Medical Center Groningen.
Another advantage is that the multi-beam uses transmission detection, which is very well suited for tissue imaging. The images you get are very similar to what biologists normally get from a regular SEM or a regular transmission electron microscope (TEM). I think that is also a big advantage, as it means they do not have to do anything different in their sample preparation.
What does the future of electron microscopy look like? Do you believe that as technology continues to advance, the throughput will continue to increase?
I think this is just a start. Now we can go to more samples, to larger volumes, and get to inter-sample variability, which will open many new questions on the technology side and the biology side.
It needs automation in the microscope process, and it also needs automation in the data analysis. I think in the end, we might move away from looking and analyzing the images ourselves to actually extracting data from these images to get the answers to our research questions.
Obviously, people want to go to a larger scale. Once you map the brain, you can think of how that relates to the entire nervous system, which is an interconnected biological system. This will carry out work on a much larger length scale.
Your research group is an early adopter of the system. How will the system change your research, and what are the next steps in your research?
We are early adopters, and we are the initiators, and I think it is a bit of a double role. The work we do here in TU Delft offers us unique opportunities for our research on the application side.
It offers new possibilities for my field of research - correlative light electron microscopy, especially in terms of neurons or system development. The electron microscope is now going to give volumes and areas that used to be the domain of light microscopy only.
We can correlate these two modalities on larger length scales. This brings in many research questions and requests for new technology, specifically at combining that data.
We will have a commercial multi-beam system here, and the fact that this attracts users with biological life science-related questions will also bring us into contact with their feedback and further requests.
It will be a driver for us to think of the next technology and what developments we should start thinking about based on the experience and the research questions that people bring to the microscope and think of when they are working with the microscope.
Another is that it allows us to look at further innovation. Based on how the system performs, how we can adjust this performance, and how we can optimize its performance, it will give us clues about how to further expand to more beams, higher throughput, higher resolution, and how to move from one magnification to another in a more or less seamless fashion.
About Jacob Hoogenboom
Jacob Hoogenboom is an associate professor at Delft University of Technology. His research group focuses on development of instrumentation and methodology for multi-modal and multi-scale microscopy. Research topics include correlative light and electron microscopy, superresolution and cathodoluminescence, ultrafast light-electron pump-probe microscopy, large-scale imaging, and the fundamentals of light-electron-matter interactions. In addition, Hoogenboom actively collaborates, already in the conceptual phase of new technology, with experts in biomedical microscopy to discuss and explore efforts needed in, e.g., probe development, sample preparation, and biological applications. He is co-founder and supervisory board member of Delmic BV and board member of the Netherlands Electron Microscopy Infrastructure (NEMI).
This information has been sourced, reviewed, and adapted from materials provided by Delmic B.V.
For more information on this source, please visit Delmic B.V.
Disclaimer: The views expressed here are those of the interviewee and do not necessarily represent the views of AZoM.com Limited (T/A) AZoNetwork, the owner and operator of this website. This disclaimer forms part of the Terms and Conditions of use of this website.