In this interview, AZoMaterials speaks with Dr. Liam Spillane, Application Scientist at Gatan Inc., about recent advances in electron energy loss spectroscopy (EELS) using dose-fractionated spectrum imaging and direct detection cameras.
Dr. Spillane explains how this technique enhances atomic resolution spectroscopy, supports cryo-EM materials analysis, and enables ultra-low and ultra-high energy loss mapping.
Can you please introduce yourself and your role at Gatan?
My name is Dr. Liam Spillane. I’m an Application Scientist at Gatan. I’ve been working in the field of electron microscopy for just over 18 years. I started my career in EM with a Ph.D. at Imperial College London, where I focused on oxide thin films using STEM and monochromated EELS.
Following that, I completed postdoctoral research in advanced EM, exploring a range of topics including nanoparticle toxicology, ferroelectrics, catalysts, and fuel cell and joined Gatan in 2012. I relocated to the United States in 2019, where I continue to support and collaborate with researchers to advance their work using our STEM EELS and GIF solutions.
What are direct detection counting cameras, and why do they matter?
Direct detection counting cameras are used in electron microscopy to register individual electron events with minimal readout noise. There are two main types: monolithic sensors (e.g., K2®/K3®) and hybrid-pixel detectors (e.g., Stela®).
Single electron counting or thresholding hybrid pixel counting, enables the incoming analog signal we measure to be completely digitized. This means that readout noise in the detection pipeline can be effectively remoted, so these cameras are ideal for rapid frame acquisition and dose fractionation, where thousands of low-dose or sparse detector frames are captured and summed. This approach only works because of the near-zero readout noise architecture. Summing huge numbers of sparse frames from conventional indirect cameras like this just doesn't work.
Dose fractionation helps maintain signal quality without damaging beam-sensitive samples.
What is dose fractionation, and why is it important in EELS?
Dose fractionation involves acquiring many fast, low-dose frames instead of a single long exposure. It was first used in cryo-EM for motion correction.
The same methodology can be followed in STEM-EELS, with electron counting cameras. 2D array spectrum imaging traditionally involves scanning an area of interest once – acquiring all the data in one pass. Instead of this “one pass” approach, we can collect time series of spectrum images (or dose fractions) and then sum a number of SI passes after acquisition. This gives us the option to discard those affected by drift or sample damage which improves data quality significantly. High speed acquisition give access to lower dose rates which helps reduce radiation damage in many samples. High speed acquisition also gives access to higher frequency spatial drift correction - which improves spatial resolution in the final summed data.
How does dose fractionation apply in the context of materials science?
Counting cameras now allow EELS data to be streamed to disc at thousands of spectra per second. Multi-pass STEM EELS was first performed with older CCD camera technology (e.g., Lewys Jones, Peter Nellist). With a CCD there are benefits as long care is taken to make sure each spectrum is above the readout noise floor of the CCD camera, but counting cameras really take this to approach to the next level and beyond. Because of the near-zero readout noise characteristics of counting cameras, they can always run at the maximum speed, even if the signal is effectively zero. Eventually when enough frames are summed in a dose fractionation series, we end up with high quality data which simply is not the case with CCD. Using Digital Micrograph’s in situ SI support, continuous multi-pass datasets can be recorded directly to high-capacity RAID storage, with drift correction applied during acquisition and continuous multi-pass scanning performed with Digiscan™ 3.
Can you give a real-world example in atomic-resolution oxide mapping?
The K3 combines fast acquisition speeds, down to 339 microseconds per frame, with near-zero noise readout, making it possible to spread the total electron dose across many short exposures. i.e., large numbers of relatively sparse or low intensity frames can be summed post-acquisition for a high SNR final dataset.
This approach, known as dose fractionation, is well established for imaging in cryo-EM. The method enables high-speed motion correction, which improves spatial resolution in the final images, but also enables data acquisition from beam-sensitive materials. The dose fractionated data can be collected at a low dose rate, which is advantageous. That data is stored in a "dose fractionated" time series, so compromised frames can be discarded post-acquisition. The same methodology followed in STEM-SI allows acquisition of high-quality, atomic-resolution data that would otherwise be difficult or impossible to obtain.
Can you share an example of a challenging experiment made possible by these technologies?
Absolutely. A good example is a recent study of strontium ruthenate (SRO) films grown on dysprosium scandate (DSO) substrates. This was in collaboration with Berit Goodge from the Max Planck Institute for Chemical Physics of Solids in Dresden, Germany. SRO is a low temperature quantum oxide material. So - all the data acquisition was carried out at liquid nitrogen temperature using a double tilt side entry cooling-holder. SRO is very beam sensitive for an oxide, so it was also necessary to use a low probe current of 20 pA for all the experiments. With the Stela and K3 cameras, we collected individual SI passes in 3–6 seconds in dose fractionation series up to a total of 420 passes. This data rate is only meaningfully achievable with counting cameras, because we can acquire a single camera readout for each pixel in the EELS spectrum image, and be confident that temporal summation post-acquisition will improve our SNR. In one example K3 dose fractionation series, 234 passes were summed after spatial drift refinement, and the result gave us atomic-resolution ADF and EELS maps. Scandium (Sc-L2,3), dysprosium (Dy-M4,5), strontium (Sr-L2,3), and ruthenium (Ru-L2,3) mapping was possible at liquid nitrogen temperature with a 20 pA probe using dose fractionation. No denoising or filtering was required, even for Ru-L2,3 which starts at 2,838 eV!
How does this compare with traditional CCD acquisition?
The accumulation time needed for good SNR core loss EELS, is around 10 ms for typical probe currents and energy-loss. For CCD, this 10 ms acquisition would typically be done in one long readout. Spectrum imaging performed with pixel dwell times of 10 ms is slow - it takes multiple minutes to acquire a single pass for a normal sized 2D array. SI data acquired this way is highly susceptible to sample drift and distortion. In contrast, modern counting cameras can acquire the same 10 ms of total accumulation time in dose fractions of 110 µs - 340 µs per pixel. This means the same 2D array SI can be acquired in seconds, with on-the-fly drift correction improved spatial accuracy. This really is game changing for EELS spectrum imaging. Routine atomic EELS experiments become much easier and more predictable. Experiments that were previously impossible or extremely changing - like atomic EELS at liquid nitrogen temperature, high energy loss, and low probe current, become possible.
Can dose-fractionated EELS support cryo-biology applications?
Yes. We used a very similar workflow in collaboration with Wah Chiu and his group at Stanford, to perform EELS mapping in frozen cell sections. A combination of ultra-low probe current (3.7 pA), a defocused STEM probe (0.5 - 1nm), sub-pixel scanning, and K3 direct detection, allowed us to acquire STEM EELS dose fractionation series at a dose rate of just 18 e-/Å2 per SI pass. Elemental mapping was performed on 15 pass dose fractionation series which allowed us to map carbon, nitrogen, oxygen and calcium. Calcium was clearly localized in mitochondrial granules.
Are there high-dose applications as well?
Yes. Andrew Thron (Gatan, USA) has used the Stela to map K-edges of cobalt, iron, and manganese in a mixed iron–cobalt oxide. A long total acquisition time was needed - well over an hour as these high energy loss transition metal K-edges have incredibly weak scattering. Dose fractionation is ideal for this scenario. The ultra-low signal can be accumulatively acquired over a long dose fractionation experiment, and eventually there is enough signal for mapping and ELNES analysis. Andrew used a pixel time of 220 µs per pixel, and an SI pass time of a few seconds. This meant it was possible to perform drift correction every few seconds, to preserve spatial resolution. It also meant the total dose was more evenly spread out. Dose fractionation enabled the detection of localized manganese that would have been lost in the averaged data.
How do you choose between K3 and Stela?
K3 is my choice when energy resolution and energy range are critical - for example, in the SRO experiment I was able to perform ELNES analysis on the full set of edges with an effective energy resolution of 0.9 eV. 200 kV Stela data at the same energy range has a resolution of 6-9 eV. Stela is my go-to choice when mapping speed is the priority; the 9,000 spectra per second acquisition rate is very well-suited for high-throughput or routine EELS mapping, or in very challenging sample drift conditions.
What are your recommendations for labs adopting this technology?
Use dose-fractionated multi-pass acquisition routinely. Set the SI pixel time to be a single readout of the counting camera and keep each pass under 10 seconds, or under 5 seconds for samples that have a high drift rate. Acquire the data quickly, discard poor-quality frames, and adjust the number of summed passes based on your priorities: spatial resolution, energy resolution, or dose limits.
This approach makes atomic-resolution EELS extremely reliable and repeatable.
About Dr. Liam Spillane
Dr. Liam Spillane is an Application Scientist at Gatan Inc. He holds a Ph.D. in electron microscopy from Imperial College London, where his research focused on aberration-corrected STEM and monochromated EELS of oxide thin films.
Following postdoctoral work in materials characterization, including nanoparticle catalysts and solid oxide fuel cells, he joined Gatan in 2012. Based in Pleasanton, California, Dr. Spillane specializes in STEM EELS and GIF technologies, particularly advanced spectrum imaging techniques with direct-detection cameras.
He has collaborated with leading groups in cryo-EM (e.g., Wah Chiu’s lab) to pioneer ultra-low-dose and cryogenic EELS mapping. His contributions are shaping next-generation atomic-resolution spectroscopy workflows across materials science and structural biology.
This information has been sourced, reviewed and adapted from materials provided by Gatan, Inc.
For more information on this source, please visit Gatan, Inc.
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.