What Is Hidden in the Volume Plasmon?
EFTEM Plasmon to Carbon Map for Soft Matter Characterization
- Fig.1: Volume plasmon image of negatively stained functionalized polystyrene spheres.
- Fig. 2: (A, D) elastic filtered TEM images of organic polymer systems; (B, E) corresponding PCR images; (C, F) AFM phase images of the block face of the same samples.
- Fig. 3: Low-loss EELS spectra obtained from the organic materials: (A) original polymers, (B) stained/non stained polyamide section.
- Fig. 4: PCR images (A, B, C) which were calculated using different energy window position and width for the bulk plasmon images. (D) Intensity line profiles of corresponding PCR images.
- Fig. 5: TEM and EFTEM images of the edge of negatively stained polyamide section. (A, B) PCR images of intact and beam damaged sample, (C) U O concentration map in terms of atoms per unit volume, (D, E) corresponding to (A, B) elastic filtered TEM images, (F) corresponding to (B, E) relative thickness map (the t/λ values marked in the top of the image).
The volume plasmon excitation of valence electrons within soft organic matter contains a variety of characteristics responsible for their intrinsic physical and chemical properties. Up to now, the analysis of those properties (hardness, surface related signal) within Transmission Electron Microscopy (TEM) remains problematic due to the lack of methodological approaches which enable the extraction of required information from the ATEM data. Here we propose an empiric EFTEM based method that allows one to identify the latter.
The interaction of a charged particle (e.g. electron) or an electromagnetic field (e.g. photons) with a many-particle system may lead either to the excitation of a single particle state of the system or to the excitation of collective states involving many particles. The latter case is described by the formation of the so-called giant resonances which are characterized by the collective motion of charged particles against that of the particles of opposite charge. Being a general physical phenomenon, this effect has been considered in nuclei , many-electron atoms , atomic clusters  and condensed media . Soft materials (e.g. vast majority of biological and polymer samples) which in general are amorphous or polycrystalline insulators have delocalized (valence) electrons which oscillate against the positively charged ions forming collective plasmon excitations as well . However due to the high complexity of the phenomenon, very little is currently known regarding plasmon-energy loss peaks formation in soft matter. The theoretical calculations regarding plasmon energy (Ep) determination which have been reported so far are based on a fully crystalline state of the described materials . Usage of electron energy loss spectroscopy (EELS) gives an important insight into the variety of excitation of a solid. The information is obtained by measuring the energy losses which the electron suffers by transmitting a thin film of the solid material . Probing of low-loss electron excitations (less than 50 eV energy loss) that involves inter- and intraband transitions and collective excitations of bonding electrons by analytical TEM (ATEM) allows one to study a variety of characteristics and processes responsible for intrinsic physical and chemical properties of condensed matter.
In accordance with the universal binding energy relationship (UBER), Ep - material property scaling laws are universal in their applicability to crystalline materials with metallic, covalent and partially ionic bonding . Here we demonstrate strong correlations between volume plasmon energy of amorphous and polycrystalline polymers and their mechanical properties.
Plasmon to Carbon Ratio (PCR) Imaging for Determination of the Mechanical Properties of Organic Systems
Figure 2 represents examples of usage of the EFTEM based PCR analytical approach for evaluation of the ultrastructural-mechanical characteristics of organic polymer systems within ATEM. The PCR technique based on the fact that the bulk (π + σ) plasmon peak features (energy and intensity) are a sensitive function of both: the valence electron density and the thickness of the sample . Since both Ep - electrical resistivity (ρ) and Young´s modulus (E) - ρ correlations were shown to exist, a relation between Ep and E was expected and then verified for carbonized samples . For amorphous/polycrystalline organic materials any theoretical work concerning Ep - mechanical property relation has not been done yet. However low-loss EELS spectra obtained from the amorphous/polycrystalline polymers (fig. 3 (A)) indicate that Ep are slightly shifted from each other, and considerably shifted from the graphite. On the other hand, the intensity of a volume plasmon peak is mainly a function of the thickness of the sample, although it also can be influenced by sample anisotropy and diffraction effects [7, 10]. The thickness-related effects, which substantially mask a material contrast (due to high input of intensity in the EFTEM bulk plasmon image), for the organic samples can be adequately removed by normalizing the bulk plasmon image on the carbon elemental map . Normalization by the specimen thickness (t/λ map) which is used in absolute quantification algorithm, cannot eliminate thickness (t) variation effects properly because the relative thickness map, in addition to sample thickness variation, also reflects variation in Z (atomic number), and depends on the accuracy of inelastic mean free path (λ) determination . The carbon map reflects mainly the intensity distribution of the signal obtained from the inner shells of carbon atoms within the sample volume, and for thin organic samples can be considered as a more accurate representative of the sample thickness . Therefore the resulting PCR image contains enhanced signal from intrinsic physical characteristics of the sample. Important that PCR data according to statistics are very similar to atomic force microscopy (AFM) phase images, where phase shift related to the hardness and elastic modulus of the materials. The major parameters which influence PCR image contrast are the energy window position and width (fig. 4) for the bulk plasmon image. These parameters have to be optimized for each particular sample composition. The Ep of polycaprolactone (PCL) and polylactid (PLA) are localized near 21 eV in contrast to carbon supportive film, which has Ep of nearly 25 eV (fig. 3 (A)). By centering the energy window near the Ep of particular phases within the sample, one can expect to get a maximum contrast from them (fig. 4). Since the material contrast almost solely depends on Ep, the energy slit width has to be chosen as small as possible in order to get highest possible contrast .
PCR Imaging for Extracting Surface Related Signal of Negatively Stained/Metal Shadowed Organic Materials
As it was mentioned above volume plasmons result from the collective excitation of the valence electrons and also reflect the solid-state character of the specimen . For the sample area which contains a metal layer, the physical density also increases resulting in a higher valence electron density. Therefore the signal from the surface replicated metal atoms also included in the EFTEM volume plasmon image. However in contrast to conventional EFTEM core-loss elemental mapping which shows only the presence of the chemical elements within the sample in 2D image, the volume plasmon image seems to contain additional information about collective interaction of those elements with each other. Therefore in PCR image the intact metal layer and the broken one (due to electron beam damage) (fig. 5 (A, B)) exhibit a completely different contrast appearance (continuous 3D-like and grainy 2D like correspondently). This phenomenon allows one to extract the surface related signal from stained organic samples by PCR. In summary the EFTEM volume plasmon imaging contains a tremendous potential for the soft matter characterization as well as for the theoretical evaluation of the phenomenon.
We are very grateful to Franz P. Schmidt, Michaela Albu and Ferdinand Hofer for fruitful discussions. This work was supported by Austrian Cooperative Research (ACR).
 Eisenberg J. M. et al.: Nuclear Models: Collective and Single-Particle Phenomena, North-Holland (1970)
 Connerade J. P. et al.: Giant Resonances in Atoms, Molecules, and Solids, Plenum Press (1987)
 De Heer W.: A. Rev. Mod. Phys. 65, 611- 676 (1993)
 Pines D. et al.: Phys. Rev. 85, 338-353 (1952)
 Egri I. J.: Phys. C: Solid State Phys. 18, 1191-1196 (1985)
 Oleshko V. P. J.: Nanosc. Nanotechnol. 12(11), 8580-8588 (2012)
 Egerton R.: Electron Energy-loss Spectroscopy in the Electron Microscope, Springer (2011)
 Matsko N. B. et al.: Microscopy and Microanalysis 19, 642-651 (2013)
 Laffont L. et al.: Carbon 40, 767-780 (2002)
 Matsko N. B. et al.: Microscopy and Microanalysis (in press) (2014)
Dr. Nadejda B. Matsko (corresponding author via e-mail request button below)
Dr. Ilse Letofsky-Papst
Institute for Electron Microscopy and Nanoanalysis (FELMI-ZFE)
Graz University of Technology
Dr. Vikas Mittal
Chemical Engineering Department
The Petroleum Institute
Abu Dhabi, UAE