As in situ liquid cell systems for the TEM have become more common, a more comprehensive understanding of how the analytical tools on the microscope can be applied with a liquid cell is required. Sample analysis extending beyond traditional TEM and STEM imaging can provide crucial information about sample behavior, especially when samples are in their native liquid environment.
Electron energy loss spectroscopy (EELS), including energy-filtered TEM (EFTEM), is one such analysis tool that is frequently used by microscopists. This method provides sample information such as element identification and details about the electronic structure, and also offers quantifiable liquid thickness measurements for liquid cell systems.
EELS systems are incorporated into a TEM via an Omega filter or a post-column filter (GIF by Gatan), which is located in the column right below the objective lens. When an EEL spectrum is produced, electrons from the primary beam interact with electrons in or near the valence shell (low-loss spectrum) of the sample as well as the electrons close to the atomic nucleus (core-loss spectrum).
The electrons are inelastically scattered, and the amount of scatter is identified and quantified by a spectrometer. The EEL signal relies on the thickness of the sample.
When samples become increasingly thick, scattering also increases and the signal degrades. The limitations of EELS analysis within an in situ liquid cell had not been strictly quantified until recently.
Researchers at the National Institute of Standards and Technology (NIST) in Bethesda, MD and in David Muller’s group at Cornell University examined how in situ EELS can benefit materials analysis in both spectroscopy and imaging (EFTEM) modes using the Protochips Poseidon liquid cell system.
This article examines the capabilities of EELS in liquid, and assesses the core and low-loss areas of the EELS spectrum as a function of liquid thickness. It also discusses EFTEM imaging in the zero-loss and low region.
Figure A. A series of EEL spectra as a function of increasing liquid thickness. The oxygen K-edge is only visible in the thinnest liquid layer, highlighted by the box. The inset shows the low-loss spectrum of the same sample. The optical gap of water is visible through thicknesses up to ~650 nm.
Figure B. The core-loss spectra of water and LiFePO4 in water. The liquid layer is ~180 nm in this case, and the oxygen K-edge and iron L-edge is evident.
Figure C. Low-loss EFTEM image of LiFePO4 in water. The red areas indicate Li rich material, and the green area indicates Li poor material. The scale bar is 500 nm.
Figure D. An EFTEM image of 30 nm Au nanoparticles using a 10 eV slit on the zero-loss peak.
Each experiment employed two Poseidon system E- chips, each with a 50 nm silicon nitride window, to contain liquid in the TEM column. David Muller’s lab used an FEI Tecnai F20 to take EEL spectra and low-loss EFTEM images. The TEM was operated at 200 kV in both traditional TEM and STEM modes, and a Gatan 865 HR-GIF was used for EELS analysis.
Pure water was used to evaluate the effects of liquid thickness on the EEL spectrum. Zero-loss EFTEM experiments were performed using a Philips/FEI CM300FEG operating at 300 kV with a post column Gatan GIF, by Ian Anderson and Kate Kline at NIST. Agglomerates of 30 nm gold nanoparticles in pure water were imaged in these experiments.
The researchers in Muller’s group discovered that if the ratio of liquid/sample thickness and inelastic mean free path (t/λ) of the electron in the sample is lower than a threshold value, meaningful low and core-loss signals can be obtained.
The core-loss spectrum is more sensitive to sample thickness, because the low-loss electrons can undergo multiple scattering and overwhelm the core-loss signal. This means that meaningful data can be obtained in the low-loss signal for thicker layers, and that meaningful coreloss data can only be extracted from thin liquid layers.
An example of this is illustrated in Figure A, where the low, coreloss spectrum of pure water as a function of increasing thickness was measured. The oxygen K-edge appears in just the thinnest liquid layers (<300 nm, t/λ = ~2.7), and disappears with increasing thickness.
However, the optical gap of water, 6.9 eV, was noticed in the low-loss spectrum up to liquid thicknesses of ~650 nm (t/λ = ~6.5) (inset). The equation I=I0 exp(t/λ) (Beer’s law) was used to determine the liquid thicknesses illustrated in this figure. Where I represents the number of unscattered electrons, established via the zero-loss peak in the EEL spectrum, and I0 is the number of incident electrons.
Figure B depicts a second instance of EELS analysis, showing the iron L-edge from a nanoparticle of LiFePO4, and the oxygen K-edge from liquid and the sample. Spectra of LiFePO4 in water were taken, reavealing that with thin enough liquid layers - 180 nm in this case - core-loss signatures are easily detected. Figure C shows a low-loss EFTEM image of the same sample.
In this experiment, a 5 eV slit was centered at 5 eV in the EELS signal. The FePO4 signal at this point is strong and highlighted green. Next, the slit was centered at 10 eV where the LiFePO4 signal is strong and highlighted red. An overlay of the images is illustrated in the figure, revealing the location of the Li-rich regions at the nanometer scale.
Additional benefits can be provided by energy filtering when imaging via thick samples. The depth of field is increased when a small energy selecting slit is centered on the zero-loss peak, which results in a crisper image. This effect is similar to using a small aperture in photography, where a smaller aperture extends the depth of field.
Using a 10 eV slit centered at the zero-loss peak, an agglomeration of gold nanoparticles in water is imaged in EFTEM mode, as displayed in Figure D. The same area is imaged without a slit, showing the effect of energy filtering on the depth of field wE. In the EFTEM image, more of the image is in focus, though the signal intensity reduces because of the small slit used.
A much clearer image of the structural detail and arrangement of the gold nanoparticle agglomerates can be acquired by EFTEM imaging compared to that obtained directly from the unfiltered image.
In materials science, EELS analysis serves as a useful tool when applied to in situ liquid samples. It helps to detect elements and maps them spatially, and also provides electronic structural information. In particular, EELS analysis is a powerful tool when detecting sample behavior in dynamic liquid environments. It also enables a better detection of material changes and better quantification of the product of reactions.
This information has been sourced, reviewed and adapted from materials provided by Protochips.
For more information on this source, please visit Protochips.