Making the most of imaging and spectroscopy in TEM: computer simulations for materials science problems

Abstract

[eng] Transmission Electron Microscopy (TEM), since its first implementation by Ernst August Friedrich Ruska and Max Knoll in 1931, has been an essential technique in the nanoscience and nanotechnology field. In the beginning, the real resolution was just a small fraction of the potential resolution expected by the fact of using electrons as a “light” source. The wavelength of the electrons accelerated at hundreds of electronvolts would involve a subatomic resolution; however, all the aberrations related to electromagnetic lenses caused a dramatic decrease. In addition, the energy resolution was highly affected by the chromatic aberration of the electron beam. Nowadays, all these initial problems have been solved by the development of the image aberration correctors and the monochromators. Since atomic resolution together with 10 meV energy resolution are a reality for researchers, new and higher horizons have been set for the transmission electron microscopy, such as orbital imaging, phonon imaging, or real time atom monitoring amongst others. TEM could be described at its most fundamental as the analysis of the result of impacting electrons with a specific compound or structure. From, this impact different data can be obtained which can be rapidly classified between imaging and spectroscopy. With the recent increases in energy and spatial resolution, a huge amount of information can be directly extracted from very large experimental datasets; however, for a deeper understanding, most of the times the support from theoretical calculations is also needed. Solid state physics with quantum considerations can contribute to an accurate description of the studied systems. Whereas in the past, materials science, solid state physics, quantum mechanics and chemistry were disciplines with a huge separation between them, nowadays they merge in the field of nanoscience and nanotechnology. When the object size is reduced to the nanoscale the quantum effects cannot be neglected anymore, any change on the synthesis can in turn change the structure which plays an essential role on the compound properties. Thus, modelling has become an essential step in the materials synthesis and characterization. The knowledge of the structure allows to compute the interaction of the electrons with any well described crystalline structure and generate images and spectra comparable with experimental data, but not just as a check, but to gain deeper insight. The interaction of the electrons with matter must be computed by solving the Schrödinger equation of the electrons interacting with the sample. The sample, the system, can be considered as a periodic potential. Imaging, measuring, modelling and manipulating matter are the basis of the promising field of nanoscience, and they can be carried out using a TEM, with the continuous support of theoretical calculations to obtain the most. The present thesis uses three main types of calculations to interpret TEM data: atomic simulations applied to imaging, Boundary Element Method (BEM) based calculations for surface plasmon distributions and Density Functional Theory (DFT) for EELS analysis. Even if they will be presented separately, they are not independent; the essence is always the same but depending on the desired results different considerations are needed. The materials science problems solved through these kinds of simulations presented in the thesis are the analysis of CuPtB ordering effects in GaInP, the influence of oxygen vacancies in the EELS of Bi2O3, the consequences of the Fe3O4 Verwey transition in its electronic structure and how it is observed in EELS and, finally, the surface plasmon distribution in gold-nanodomes as a function of the dome shape. To conclude, the simulations have been presented as an essential tool to complement TEM studies to link the experimental results with the most fundamental aspects which are determined by the structure of the studied materials.

[cat] Aquesta tesi doctoral s'ha centrat en la realització de càlculs teòrics que permetin comprendre i extreure la major quantitat d'informació possible de les dades experimentals de microscòpia de transmissió d’electrons (TEM), i de les tècniques espectroscòpiques relacionades, concretament, l'espectroscòpia de pèrdua d’energia dels electrons (EELS). S’hi utilitzen tres tipus principals de càlculs per interpretar les dades del TEM: simulacions atòmiques aplicades a l'obtenció d'imatges, càlculs basats en el mètode d'elements de contorn (BEM) per a les distribucions de plasmons superficials i la teoria del funcional de la densitat (DFT) per a l'anàlisi d’EELS. Tot i que es presentin per separat, no són independents; l'essència sempre és la mateixa, però depenent dels resultats desitjats es necessiten diferents consideracions. En aquest sentit, primerament s'han presentat les bases físiques de diferents mètodes de simulació: simulació multislice per calcular imatges de contrast de número atòmic i de contrast de fase, càlculs (DFT) per calcular dades EELS de baixa pèrdua i de pèrdues profundes i, simulacions basades en BEM per a plasmons de superfície. Un cop presentades les bases, s’han resolt problemes de la ciència dels materials mitjançant aquest tipus de simulacions: l'anàlisi dels efectes d'ordenació del CuPtB al GaInP, la influència de les vacants d'oxigen a l'EELS del Bi2O3, les conseqüències de la transició Fe3O4 Verwey en la seva estructura electrònica i com s'observa a l'EELS i, finalment, la distribució de plasmons superficials als nanodoms d'or en funció de la forma de la cúpula. En resum, al llarg la tesi doctoral les simulacions han demostrat ser una eina essencial per complementar els estudis de TEM, per vincular els resultats experimentals amb els aspectes més fonamentals determinats per l'estructura dels materials estudiats.