Assessing Oxygen Vacancies in Bismuth Oxide ... - ACS Publications

Oct 24, 2017 - This type of information has not been accessed previously from EELS ... spectra or how vacancies reflect in Bi2O3 spectra were unreport...
0 downloads 0 Views 4MB Size
Subscriber access provided by University of Missouri-Columbia

Article

Assessing Oxygen Vacancies in Bismuth Oxide Through EELS Measurements and DFT Simulations Pau Torruella, Catalina Coll, Gemma Martin, Lluis Lopez-Conesa, María Vila, Carlos Díaz-Guerra, María Varela, Maria Luisa Ruiz-González, Javier Piqueras, Francesca Peiró, and Sonia Estrade J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.7b06310 • Publication Date (Web): 24 Oct 2017 Downloaded from http://pubs.acs.org on October 26, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Assessing Oxygen Vacancies in Bismuth Oxide through EELS Measurements and DFT Simulations Pau Torruella*1,2, Catalina Coll1,2, Gemma Martín1,2, Lluís López-Conesa1,2,3, María Vila4, Carlos Díaz-Guerra4, María Varela4,5, María Luisa Ruiz-González6, Javier Piqueras4, Francesca Peiró1,2, Sònia Estradé1,2. 1

LENS-MIND, Departament d'Enginyeries: Electrònica, Universitat de Barcelona, 08028 Barcelona, Spain. 2 Institute of Nanoscience and Nanotechnology (IN2UB), Universitat de Barcelona, 08028 Barcelona, Spain. 3 TEM-MAT, CCiT, Universitat de Barcelona, 08028 Barcelona, Spain. 4 Departamento de Física de Materiales, Facultad de Ciencias Físicas, Universidad Complutense de Madrid, 28040 Madrid, Spain. 5 Instituto de Magnetismo Aplicado. Facultad de Ciencias Físicas, Universidad Complutense de Madrid, 28040 Madrid, Spain. 6 Departamento de Química Inorgánica, Facultad de Químicas, Universidad Complutense de Madrid, 28040 Madrid, Spain. *Corresponding autor: [email protected] Abstract Pioneering Electron Energy Loss Spectroscopy (EELS) measurements of α-Bi2O3 are performed on three samples obtained through different synthesis methods. Experimental low-loss and core-loss EELS spectra are acquired. By combining them with detailed structural characterization and Density Functional Theory (DFT) simulations, we are able to detect and evaluate the presence of oxygen vacancies in the samples. This type of information has not been accessed previously from EELS data in bismuth oxide, since high-resolution EELS spectra or how vacancies reflect in Bi2O3 spectra were unreported. This novel measurement is further validated through comparison with photoluminescence data. Therefore, the technique has the ability to probe oxygen vacancies in Bi2O3 at an unprecedented resolution which might allow solving material science and technological issues related to this material. Introduction Bismuth oxide is a polymorph with four main crystallographic phases, namely the monoclinic α-phase which is stable at low temperature; the tetragonal β and face-centered cubic γ phases which are metastable; and the δ-phase which is stable at high temperatures1,2. It is a widely used material in the state-of-the-art of many technological fields such as gas sensors3,4, optical devices5,6 and solid oxide fuel cells7,8. However, the functional properties of Bi2O3 are strongly dependent on its crystalline phase. As an example, the conduction mechanism of α-Bi2O3 is mainly electronic whereas for δ-Bi2O3 is mainly ionic9,10. Moreover, with the advent of nanotechnology a huge miscellany of Bi2O3 nanostructures is being reported, from 1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

nanocoatings to nanowires, nanotubes, or hollow nanoparticles, each with its own applications and virtues11–15. Therefore, when tailoring these complex nanostructures, a way to assess Bi2O3 phase, defects and other structural properties with high resolution is mandatory. In this context, the use of a high resolution transmission electron microscope (TEM) provides a great tool to evaluate crystalline properties at the nanoscale16. Additionally, when coupled with electron energy loss spectroscopy (EELS) it can also probe chemical and electronic properties of a given material at high resolution. In particular, the band structure of a material is related to low-loss EELS spectra, while information on the oxidation state, coordination or vacancy presence of a given element in the sample can be obtained from core-loss spectra by analyzing its energy loss near edge structure (ELNES) features17. The latter properties are of particular importance when dealing with bismuth oxide, since its high ionic conductivity is discussed in terms of oxygen vacancy ordering, generation and transport9,10,18,19. However, obtaining this information through the ELNES fine structure is not straightforward and requires previous knowledge on how microscopic properties of the given material alter the spectra. This can only be obtained from reference samples or simulations20,21. Surprisingly, reported EELS measurements of Bi2O3 only use low-loss spectra to determine band gap energy19 and none, to the best of our knowledge, presents any spectrum from the core-loss region. The present work aims to bridge this gap of knowledge and provide Bi2O3 core-loss EELS data, while linking experimental and simulation EELS results to microscopic properties of three different Bi2O3 samples, namely bismuth oxide nanowires synthesized by thermal evaporation22, commercial Bi2O3 powder, and a Bi2O3 ceramic23 obtained in the form of a sintered pellet from this powder. The crystalline structure of the samples has been fully characterized through scanning-TEM (STEM) and Raman spectroscopy. Careful EELS measurements have been performed in low-loss and core-loss regimes. In order to understand the EELS features from the obtained spectra, the measurements are supported with density functional theory (DFT) simulations that can calculate the band structure and the ELNES features of a given crystalline structure24. Finally, photoluminescence (PL) measurements are performed and the results are compared with the conclusions obtained from the novel EELS measurements. Methods Commercial Bi2O3 powders (99.999%) were purchased from Sigma – Aldrich. In order to obtain the Bi2O3 ceramics, these powders were compacted under compressive load, forming disk – shaped samples of about 1 mm thick and 7 mm diameter. These samples were then annealed in air at 750 0C for 10 hours23. Bi2O3 nanowires were grown mixing 80 wt. % Bi powder (Goodfellow, 99.997%) and 20 wt. % Er2O3 powder (STREM, 99.9%). This mixture was compressed to form disk – shaped pellets, which were then annealed at 800 0C for 4 h in a horizontal tube furnace under Ar flow. This treatment led to the growth of a high density of αBi2O3 nanowires22. It must be pointed out that Er was not incorporated into the nanowires25. Its role is to alter the oxidation kinetics, favouring the formation of a high density of nanostructures. Micro-Raman and micro-PL measurements were carried out at room temperature in a Horiba Jovin-Ybon LabRAM HR800 system. In order to collect Raman spectra, the samples were 2 ACS Paragon Plus Environment

Page 2 of 21

Page 3 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

excited by a 633 nm He–Ne laser on an Olympus BX 41 confocal microscope with a 100× objective. A 325 nm He-Cd laser and a 40x objective were used for PL measurements. Scanning TEM (STEM) high angle annular dark field (HAADF) and EELS measurements from the samples were acquired in a JEOL ARM200cF TEM operated at 200 kV and equipped with a cold field emission fun (c-FEG) and a GATAN Quantum GIF EEL spectrometer. A Gatan double tilt cryo-holder (model 636) was used for the ceramic sample in order to minimize electron beam damage, with a working temperature of 97 K. The collection and convergence semi-angles were 27.78 mrads and 35 mrads respectively throughout all the EELS measurements. These measurements were performed using the dual-EELS acquisition mode, which allows acquiring both low-loss and core-loss EELS spectra simultaneously. The latter were focused in the 520560 eV energy loss range, were the O K edge can be observed. The dual acquisition enables aligning core-loss spectra with the zero-loss peak, therefore making the absolute energy-loss position of different features of the peaks reliable. Single spectra for each sample were obtained by summing larger spectrum images after zero-loss peak alignment. These spectrum images were acquired with atomic resolution and the corresponding co-acquired HAADF images (Figure S1) allow identification of the zone axis of the crystal during the acquisition as well as confirming the lack of beam damage. They were also acquired so that there were no carbon contributions either from the TEM grid or from contamination. Theoretical calculations of the electronic structure of α-Bi2O3 were carried out using the linearized augmented plane wave (LAPW) method based on DFT, implemented on WIEN2k 26,27 ab initio simulation package. The DFT calculation was performed considering the local density approximation28,29 (LDA). The atomistic input model was built according to the crystallographic information of the unit cell (UC) from α- Bi2O3 structure as reported in 30, which has a P21c space group (5.8486, 8.1661 and 7.5097 Å lattice parameters and beta angle of 113 0, giving 5 non-equivalent atomic positions). The cutoff energy of the simulation was set at -11.9 Ry, defining the following valence and semicore electrons: 2s² and 2p⁴ for oxygen and 4f 14, 5d10, 6p³ and 6s² for bismuth. DFT Self-consistency cycles were performed were performed until the total energy was converged to 10-5 eV and the residual forces on atoms were below 0.01 eV/Å with 1000 k-points. The plane-wave basis set cutoff was set as RKmax=7. Results Normalized Raman spectra representative of the Bi2O3 powders, ceramics and nanowires, are shown in Figure 1A. No significant differences, in terms of peak positions and bandwidths, were found in Raman spectra of the different samples. Raman bands appear peaked at 84, 119, 139, 152, 184, 210, 280, 313, 411, 448, and 526 cm -1, which are all characteristic of the αBi2O3 phase31–33. PL measurements, Figure 1B, show three narrow, well-resolved emission bands, peaked at 1.86, 1.95 and 2.05 eV. Besides, two weaker and broader emissions, respectively centered near 2.10 eV and 2.27 eV, can be also appreciated. The PL intensity varies strongly amongst the specimens. The Bi2O3 ceramic shows the maximum PL intensity, which halves for the nanowires, and is three times weaker in the powder. In addition, the relative intensity of the 2.10 and 2.27 eV bands varies for each sample, with the ceramic showing the highest 2.27 eV relative emission and the commercial powder the lowest. Moreover, the peak positions of the 3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

abovementioned three narrow emissions observed centered between 1.86 and 2.05 eV in PL spectra of powders and nanowires appear shifted 0.3 eV towards higher energies in the ceramic. Additionally, PL spectra from the later sample shows a weak and broad new emission at about 3.0 eV as well as an intense band centered near 1.72 eV. The commercial powder sample was prepared for TEM observation by the standard dilution in hexane and drop pouring onto a TEM holey carbon grid. The TEM structural analysis showed that it was comprised of crystals with lateral sizes varying from 50 nm to 500 nm, depicted in Figure 2A, which consisted of Bi2O3 in the α-phase, seen in the [001] zone axis in Figure 2B. As for the pellet, the sample was prepared by the standard FIB lift-out technique34. TEM characterization evidenced it consisted in large (several microns) α-phase Bi2O3 crystals (Figure 2C-D). The Bi2O3 nanowires were prepared by solution and sonication in hexane, followed by pouring a drop of the solution onto a TEM holey carbon grid. As can be seen in the lowmagnification TEM image (Figure 2E), they were found to form hierarchical structures and also consisted mainly of bismuth oxide in the α-phase (Figure 2F). Sparse, smaller nanowires of the metastable β-Bi2O3 nanowires were also observed but as Raman measurements show, they are not statistically relevant. The EELS measurements at 200 kV for each sample are shown in Figure 3. Clear differences between samples appear in the spectra. In order to discuss these differences, the main features of the spectra have been labelled. The core-loss spectra show a main peak at 534.6 eV (b1) which is at the same position for all samples. The onset of this peak shows a shoulder at 532.5 eV (a1) that varies in height when comparing the commercial powder sample with the other two, which were thermally treated. The a1 intensity goes from being almost 25% lower than the b1 peak in the commercial powder to being equal in the nanowire and the sintered pellet. Moreover, the powder shows some additional peaks in the continuum region of the spectrum at 540.2 eV (c1) and 545.7 eV (d1). To understand information from the EEL spectra, an initial DFT simulation for the standard unit cell was carried out. Importantly, the LDA approximation used in the calculations is known to not reliably calculate absolute onset energies for EELS edges35. Because of this, a shift of 1.2 eV will be applied to all simulations. This value has been chosen so that the b1 peak energy of the experimental spectrum of the commercial powder matches the maximum of the simulated edge from the unit cell, de facto taking the commercial powder as a reference sample. The simulated O K edge from the unit cell is shown in Figure 4A, where the energies of the experimental features (a1,b1,c1,d1) are also labelled. An excellent match with the experimental spectrum from the reference Bi2O3 powder can be appreciated: the simulation shows all the features of the experimental spectra, with only small differences in the position of the c1 peak (at a lower energy in the simulation) and the d1 peak (at a higher energy). Remarkably, this is the first reported experimental spectrum of α-Bi2O3, with additional confirmation from the simulation, and will be made available at the EELS DataBase service36. In contrast, the a1 shoulder is much more intense in the pellet and nanowire spectra, as well as both showing a steeper onset than the reference sample. The good match between the powder sample and the UC simulation, whilst small discrepancies show up for nanowire and pellet, points to significant changes between these samples.

4 ACS Paragon Plus Environment

Page 4 of 21

Page 5 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

An important fact to consider is that in the UC of α-Bi2O3 there are 6 oxygen atoms located at three different non-equivalent crystallographic positions (two in each one), each one seeing a different electronic environment in the structure. These positions will be referred as OI for x=0.2337a, y=0.4533b, z=0.1266c and equivalent positions; OII for x=0.265a, y=0.0294b, z=0.0115c; and OIII for x=0.7783a, y=0.3037b, z=0.2080c where a, b, c are the unit cell parameters. The contributions of oxygen atoms in OI, OII, OIII can also be extracted from the simulated UC, and are shown in Figure 4A. The total spectra can be evaluated as the sum of these contributions weighted by their multiplicity (the number of atoms in each position and equivalents, two for all in this case). In Figure 4A it is clear that these contributions are different. In fact, the most notable difference is also the height of the a1 shoulder, OII having the lowest, and the b1 peak distance respect to the edge onset, OII having the highest. This fact hints that the difference in the experimental spectra might be due to an imbalance on the occupation of the three oxygen sites. In order to further investigate this, simulations of unit cells with oxygen vacancies were undertaken. A few considerations must be taken into account when removing an atom from a unit cell in a DFT simulation. Even though a unit cell with one missing oxygen atom might seem a rather small defect in a big crystal, the simulations are performed with periodic boundary conditions, which means actually an infinite crystal with one less oxygen in each unit cell is being simulated. Because of this, if the vacancy proportion wants to be kept low, a large “supercell” consisting of multiple unit cells in which only one oxygen vacancy is being introduced must be simulated. Three additional DFT simulations were performed to evaluate oxygen vacancy concentration impact on EELS spectra. A 2x1x1 “supercell”, which will be referred to as V simulation, with one oxygen atom missing, a 3x1x1 “supercell” with one vacancy and a 3x1x1 “supercell” with two vacancies. Cosidering that the unit cell of Bi2O3 has 12 oxygen atoms in total, vacancy concentrations are 1/24=4.17%, 1/36=2.78% and 2/36=5.56% respectively. The crystallographic position in which the vacancy is formed must be carefully considered. The vacancy formation energy for each position was calculated yielding similar values for each one within the reliability of the simulation. Because of this, vacancies in the OII site were chosen, as their contribution can account for the observed discrepancies between the experimental spectra of the powder and the nanowire/pellet sample. The same offset of 1.2eV applied to the UC simulation is applied to the vacancy simulations calculated spectra. UC and V simulation were compared both in the core-loss (Figure 4B) and the low-loss regime (Figure 4C). Remarkably, no major differences appear in the low-loss spectra, apart from a small peak at 1.9 eV. Similarly, the experimental low-loss spectra barely changes between each sample, and unfortunately at 1.9 eV the zero-loss peak overlaps with the loss function of the samples too much to see any peak of this kind in experimental data. But the most interesting feature in this comparison is the appearance of a 1.3 eV shift to lower energies of the O K edge in the V simulation. This seems very reasonable if one bears in mind that this shift has also been seen in EELS spectra of other transition metals, such as Fe, Co or Mn, when they are reduced37–40. By comparing the O K EELS edge position of all the simulations (Figure 5) it is clear that the energy shift is proportional to the vacancy concentration. Discussion

5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

As far as the cationic structure is concerned, the samples show no significant differences, all of them being mainly α-Bi2O3. However, the experimental EEL spectra show clear differences and their origin must be elucidated. At first, one could think that the origin of these small differences could be found in the orientation dependence of EEL spectra41 since the samples were not in the same zone axis during the acquisition. In order to explore this possibility, UC simulations for the relevant zone axis were carried out. As can be seen in Figure S2, they show no significant differences in our experimental conditions, so this factor can be ruled out. Additionally, because of the known dependence on oxygen vacancies of bismuth oxide properties, it is reasonable to suspect that they might have a role in these measurements. In order to perform a rigorous analysis of the core-loss spectra from the nanowire and pellet sample in terms of oxygen vacancies, each configuration of missing oxygen atoms for increasingly large supercells should be simulated. This approach is not feasible in terms of computational time. In this regard, it must be kept in mind that bismuth is one of the heaviest stable elements on the periodic table, with 83 electrons, 29 of which have been treated rigorously within the DFT formalism (orbitals 4f to 6s as stated in the methods section), and therefore costly to simulate. However, based on the knowledge that the previous simulations have yielded, that vacancies shift the O K edge, and that b1, a1 features vary for different ( ) ( ) symmetry sites, the following simple model can be inferred: ( ) ( ) . Properly normalized, the ki parameters represent occupancies of the different oxygen sites and the parameter s accounts for the energy shift that the V simulation demonstrates. This model can now be used to fit the experimental data and give some grounds for discussion on the presence and nature of oxygen vacancies on the material. Importantly, it must be kept in mind that the positions of the simulated oxygen spectra have been calibrated by using the commercial powder sample, and therefore any property calculated by this method will be relative to the commercial powder. However, this seems reasonable since it is the purest form in which we can obtain this material. The model was implemented in Python using a Scipy curve fitting routine42 and the results of the fit with the experimental are summarized in Figure 6. The obtained ki and s parameters (detailed in the supplementary information) show that the occupancy of OI and OIII barely changes in the three samples. Nonetheless, a small shift of approximately 0.45 eV for the nanowire sample and 0.58eV for the pellet is introduced and the occupation OII drops to zero for these samples. Furthermore, from Figure 5 the linear relation between O K edge energy shift and vacancy concentration can be calculated by a linear regression, yielding an r-value of 0.97 (details in the supplementary information). From the regression values, the O K shift value of the nanowire and pellet sample can be translated into approximately 1.2% of vacant oxygen sites in the nanowires and 1.6% in the pellet sample, relative to the powder. These results would indicate that, although the commercial powder has a certain density of vacancies, they are uniformly distributed in all crystallographic positions while the ceramics and nanowires have a higher concentration of vacancies preferably located at the OII site. The findings are in good agreement with PL results. Previous luminescence studies of α- Bi2O3 ceramic samples annealed in different atmospheres had shown that a 2.10 eV band is related 6 ACS Paragon Plus Environment

Page 6 of 21

Page 7 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

to oxygen vacancies23. Other works reported that the emission in the low-energy spectral range (≈below 2.1 eV) can also be attributed to oxygen vacancies that form defect donor states43, and that the higher the emission intensity in this spectral range, the greater the vacancy density44. Hence, the higher PL intensity observed in ceramics and nanowires - as compared to that measured in powder samples - as well as the higher relative intensity of the PL emissions in the low-energy spectral range - support the fact that our EELS results reveal a higher concentration of oxygen vacancies in the studied α- Bi2O3 nanowires and ceramics. Conclusions Structural characterization and PL measurements have been carried out. Core-loss EELS spectra of Bi2O3 have been reported for the first time for three different α-Bi2O3 samples. These measurements have been interpreted with the help of DFT simulations of the α-Bi2O3 unit cell and of super cells containing oxygen vacancies. The DFT simulations establish the relation between oxygen vacancy concentration and O K edge onset. The EELS spectra analysis confirms an increase in oxygen vacancies in the nanowires and the pellet sample, occupying preferably the OII crystallographic site. Remarkably, the PL results are in good agreement with these findings. Therefore, the current work allows ELNES analysis to be an efficient tool to assess oxygen vacancies in Bi2O3, extracting a type information which was not possible to access previously from EEL spectra. Supporting information. HAADF images coacquired the spectra in figure 3. Bi2O3 O K Edge orientation dependence simulations. Vacancy formation energies obtained from the DFT calculations. Fitting description and obtained parameters from figures 5, 6. Acknowledgments This work has been carried out in the frame of the Spanish research projects MAT2013-41506P, MAT2016-79455-P, MAT2013-48628-R, MAT2015-65274-R, MAT2015-066888-C3-3-R, CSD2009-0013 and FIS2013-46159-C3-3-P. It was also supported by the Spanish Agencia Estatal de Investigación (AEI) and the European Regional Development Fund (ERDF). Catalan Government support from the SGR2014-672 and 2014-SGR-1015 projects are also acknowledged. Financial support from the ERC PoC MAGTOOLS is also acknowledged. Measurements were performed in the Centro Nacional de Microscopía Electrónica (CNME) at the Universidad Complutense de Madrid (UCM) and also at the Centres Científico-Tècnics of the Universitat de Barcelona (CCIT-UB). The DFT simulations were carried out in the LUSITANIA II supercomputer located in Cáceres, a node of the Spanish supercomputing net (Red Española de Supercomputación – RES).

7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. Optical characterization. A) Normalized Raman spectra representative of the three kinds of Bi2O3 samples investigated. B) Normalized PL spectra from the three Bi2O3 investigated samples.

8 ACS Paragon Plus Environment

Page 8 of 21

Page 9 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 2. STEM-HAADF imaging. A) Low magnification STEM-HAADF image of a group of Bi2O3 nanocrystals from the reference powder. B) Atomic resolution STEM-HAADF image of a Bi2O3 nanocrystal seen along the [001]α-Bi2O3 zone axis. C) Low magnification STEM-HAADF image of the FIB-prepared Bi2O3 pellet. The top bright layer seen in the image is due to the preparation of the sample. D) Atomic resolution STEM-HAADF image from the pellet sample seen along the [001] zone axis of the α-phase. E) Low magnification TEM image of the nanowire sample. F) STEM-HAADF image of an α-phase nanowire along the [010] zone axis of the α-phase.

9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Experimental EELS. Core-loss and low-loss from the three investigated samples. Note that each spectrum has been divided by the intensity at its maximum. The main peaks of the spectra have been marked and labeled: a1=532.5 eV, b1=534.6 eV, c1=540.2 eV, d1=545.7 eV, a2=11.0eV, b2=18.7eV, c2=29.5eV.

Figure 4. DFT simulations of EEL spectra. A) Simulated O K Edge spectrum for α-Bi2O3 (plotted in black). The contributions from each non-equivalent oxygen in the unit cell are also plotted. An inset with the α-Bi2O3 unit cell seen along the [010] zone axis shows the position of each of these oxygen in blue, red and green (the Bi atoms are shown in purple). The energy positions of the spectral features identified in the experimental spectra are also shown as vertical lines. 10 ACS Paragon Plus Environment

Page 10 of 21

Page 11 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

B) Simulated core-loss and C) low-loss spectra for α-Bi2O3 and for α-Bi2O3 with one oxygen less for every two unit cells.

Figure 5. DFT simulations of oxygen vacancies. O K energy shift for each of the DFT simulations.

Figure 6. EELS fitting. Fitting of the core-loss experimental spectra from each α-Bi2O3 sample with the weighted contributions of each non-equivalent oxygen atoms and an energy shift. 11 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

REFERENCES (1)

Harwig, H. A.; Gerards, A. G. The Polymorphism of Bismuth Sesquioxide. Thermochim. Acta 1979, 28, 121–131.

(2)

Harwig, H. A. On Structure of Bismuth Sesquioxide: The Α, Β, Γ and Δ-Phase. Zeitschrift für Anorg. und Allg. Chemie 1978, 444, 151–166.

(3)

Gou, X.; Li, R.; Wang, G.; Chen, Z.; Wexler, D. Room-Temperature Solution Synthesis of Bi2O3 Nanowires for Gas Sensing Application. Nanotechnology 2009, 20 (49), 495501.

(4)

Cabot, A.; Marsal, A.; Arbiol, J.; Morante, J. R. Bi2O3 as a Selective Sensing Material for NO Detection. Sensors Actuators B Chem. 2004, 99 (1), 74–89.

(5)

Vila, M.; Díaz-guerra, C.; Lorenz, K.; Piqueras, J.; Alves, E.; Nappini, S.; Magnano, E. Structural and Luminescence Properties of Eu and Er Implanted Bi2O3 Nanowires for Optoelectronic Applications. J. Mater. Chem. C 2013, 1, 7920–7929.

(6)

Eberl, J.; Kisch, H. Visible Light Photo-Oxidations in the Presence of Α-Bi2O3. Photochem. Photobiol. Sci. 2008, 7.

(7)

Wachsman, E. D.; Lee, K. T. Lowering the Temperature of Solid Oxide Fuel Cells. Science (80-. ). 2011, 334 (June 2005).

(8)

Zhong, G. H.; Wang, J. L.; Zeng, Z. Ionic Transport Properties in Doped Δ-Bi2O3. J. Phys. Conf. Ser 2006, 29, 1–5.

(9)

Takahashi, T.; Iwahara, H.; Arao, T. High Oxide Ion Conduction in Sintered Oxides of the System Bi2O3-Y2O3. J. Appl. Electrochem. 1975, 5, 187–195.

(10)

Shuk, P.; Wiemhöfer, H. D.; Guth, U.; Göpel, W.; Greenblatt, M. Oxide Ion Conducting Solid Electrolytes Based on Bi2O3. Solid State Ionics 1996, 89 (1996), 179–196.

(11)

Niu, K. Y.; Park, J.; Zheng, H.; Alivisatos, a. P. Revealing Bismuth Oxide Hollow Nanoparticle Formation by the Kirkendall Effect. Nano Lett. 2013, 13 (11), 5715–5719.

(12)

Kumari, L.; Lin, J.; Ma, Y. One-Dimensional Bi2O3 Nanohooks: Synthesis, Characterization and Optical Properties. J. Phys. Condens. Matter 2007, 19.

(13)

Qiu, Y.; Liu, D.; Yang, J.; Yang, S. Controlled Synthesis of Bismuth Oxide Nanowires by an Oxidative Metal Vapor Transport Deposition Technique. Adv. Mater. 2006, 18 (19), 2604–2608.

(14)

Sarma, B.; Jurovitzki, A. L.; Smith, Y. R.; Mohanty, S. K.; Misra, M. Redox-Induced Enhancement in Interfacial Capacitance of the Titania Nanotube/bismuth Oxide Composite Electrode. ACS Appl. Mater. Interfaces 2013, 5 (5), 1688–1697.

12 ACS Paragon Plus Environment

Page 12 of 21

Page 13 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(15)

Mahmoud, W. E.; Al-Ghamdi, A. A. Synthesis and Properties of Bismuth Oxide Nanoshell Coated Polyaniline Nanoparticles for Promising Photovoltaic Properties. Polym. Adv. Technol. 2011, 22 (6), 877–881.

(16)

Williams, D. B.; Carter, C. B. Transmission Electron Microscopy: A Textbook for Materials Science; 2009; Vol. V1-V4.

(17)

Egerton, R. F. Electron Energy-Loss Spectroscopy in the Electron Microscope (3rd Edition); 2011.

(18)

Boyapati, S.; Wachsman, E. D.; Jiang, N. Effect of Oxygen Sublattice Ordering on Interstitial Transport Mechanism and Conductivity Activation Energies in PhaseStabilized Cubic Bismuth Oxides. Solid State Ionics 2001, 140, 149–160.

(19)

Shuk, P.; Wiemhöfer, H. D.; Göpel, W. Electronic Properties of Bi2O3 Based Solid Electrolytes. Zeitschrift für Anorg. und Allg. Chemie 1997, 623, 892–896.

(20)

Yedra, L.; Xuriguera, E.; Estrader, M.; López-ortega, A.; Baró, M. D.; Nogués, J.; Roldan, M.; Varela, M.; Estradé, S.; Peiró, F. Oxide Wizard : An EELS Application to Characterize the White Lines of Transition Metal Edges. Microsc. Microanal. 2014, 20, 698–705.

(21)

Varela, M.; Oxley, M. P.; Luo, W.; Tao, J.; Watanabe, M.; Lupini, a. R.; Pantelides, S. T.; Pennycook, S. J. Atomic-Resolution Imaging of Oxidation States in Manganites. Phys. Rev. B - Condens. Matter Mater. Phys. 2009, 79 (8), 1–14.

(22)

Vila, M.; Diaz-Guerra, C.; Piqueras, J.; Lopez-Conesa, L.; Estrade, S.; Peiro, F. Growth, Structure, Luminescence and Mechanical Resonance of Bi2O3 Nano- and Microwires. CrystEngComm 2015, 17 (1), 132–139.

(23)

Vila, M.; Díaz-Guerra, C.; Piqueras, J. Luminescence and Raman Study of Α-Bi2O3 Ceramics. Mater. Chem. Phys. 2012, 133 (1), 559–564.

(24)

Hébert, C. Practical Aspects of Running the WIEN2k Code for Electron Spectroscopy. Micron 2007, 38, 12–28.

(25)

Vila, M.; Díaz-Guerra, C.; Piqueras, J. Laser Irradiation-Induced Α to Δ Phase Transformation in Bi2O3 Ceramics and Nanowires. Appl. Phys. Lett. 2012, 101 (7), 071905.

(26)

Blaha, P.; Schwarz, K.; Madsen, G.; Kvasicka, D.; Luitz, J. WIEN2K, an Augmented Plane Wave + Local Orbital Program for Calculating Crystal Properties. Technische universität Wien 2001.

(27)

Blaha, P.; Schwarz, K.; Sorantin, P. Full-Potential, Linearized Augmented Plane Wave Programs for Crystalline Systems. Comput. Phys. Commun. 1990, 59 (2), 399–415.

(28)

Hohenberg, P.; Kohn, W. Inhomogenous Electron Gas. Phys. Rev. B 1964, 136 (3), 864– 871.

(29)

Kohn, W.; Sham, L. J. Self-Consistent Equations Including Exchange and Correlation Effects. Phys. Rev. B 1965, 140 (4), 1133–1139. 13 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(30)

Kalinchenko, F. V.; Borzenkova, M. P.; Novoselova, A. V. The BiF3 - Bi2O3 System. Russ. J. Inorg. Chem. 1981, 26 (1), 118–120.

(31)

Betsch, R. J.; White, W. B. Vibrational Spectra of Bismuth Oxide and the SilleniteStructure Bismuth Oxide Derivatives. Spectrochim. Acta Part A Mol. Spectrosc. 1978, 34 (5), 505–514.

(32)

Denisov, V. N.; Ivlev, A. N.; Lipin, A. S.; Mavrin, B. N.; Orlov, V. G. Raman Spectra and Lattice Dynamics of Single-Crystal Α-Bi2O3. J. Phys. Condens. Matter 1997, 9 (23), 4967– 4978.

(33)

Narang, S. N.; Patel, N. D.; Kartha, V. B. Infrared and Raman Spectral Studies and Normal Modes of Α-Bi2O3. J. Mol. Struct. 1994, 327 (2-3), 221–235.

(34)

Giannuzzi, L. A.; Kempshall, B. W.; Schwarz, S. M.; Lomness, J. K.; Prenitzer, B. I.; Stevie, F. A. FIB Lift-Out Specimen Preparation Techniques. In Introduction to Focused Ion Beams; Kluwer Academic Publishers: Boston, 2005; pp 201–228.

(35)

Hamann, D. R.; Muller, D. A. Absolute and Approximate Calculations of Electron-EnergyLoss Spectroscopy Edge Thresholds. Phys. Rev. Lett. 2002, 89 (12), 126404–4.

(36)

Ewels, P.; Sikora, T.; Serin, V.; Ewels, C. P.; Lajaunie, L. A Complete Overhaul of the Electron Energy-Loss Spectroscopy and X-Ray Absorption Spectroscopy Database: Eelsdb.eu. Microsc. Microanal. 2016, 22, 717–724.

(37)

Schmid, H. K.; Mader, W. Oxidation States of Mn and Fe in Various Compound Oxide Systems. Micron 2006, 37 (5), 426–432.

(38)

López-Ortega, A.; Roca, A. G.; Torruella, P.; Petrecca, M.; Estradé, S.; Peiró, F.; Puntes, V.; Nogués, J. Galvanic Replacement onto Complex Metal-Oxide Nanoparticles: Impact of Water or Other Oxidizers in the Formation of Either Fully Dense Onion-like or Multicomponent Hollow MnOx/FeOx Structures. Chem. Mater. 2016, 28 (21), 8025– 8031.

(39)

Wang, Z.; Bentley, J.; Evans, N. Valence State Mapping of Cobalt and Manganese Using near-Edge Fine Structures. Micron 2000, 31 (4), 355–362.

(40)

Torruella, P.; Arenal, R.; de la Peña, F.; Saghi, Z.; Yedra, L.; Eljarrat, A.; López-Conesa, L.; Estrader, M.; López-Ortega, A.; Salazar-Alvarez, G.; et al. 3D Visualization of the Iron Oxidation State in FeO/Fe3O4 Core−Shell Nanocubes from Electron Energy Loss Tomography. Nano Lett. 2016, 16, 5068–5073.

(41)

Schattschneider, P.; Jouffrey, B.; Hebert, C. Orientation Dependence of Ionization Edges in EELS. Ultramicroscopy 2001, 86, 343–353.

(42)

Jones, E.; Oliphant, T.; Peterson, P.; others. {SciPy}: Open Source Scientific Tools for {Python}. 2001.

(43)

Schmidt, S.; Kubaski, E. T.; Volanti, D. P.; Sequinel, T.; Bezzon, V. D. N.; Beltrán, A.; Tebcherani, S. M.; Varela, J. A. Effect of Pressure-Assisted Heat Treatment on

14 ACS Paragon Plus Environment

Page 14 of 21

Page 15 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Photoluminescence Emission of Α-Bi2O3 Needles. Inorg. Chem. 2015, 54 (21), 10184– 10191. (44)

Wu, Y.; Lu, G. The Roles of Density-Tunable Surface Oxygen Vacancy over Bouquet-like Bi2O3 in Enhancing Photocatalytic Activity. Phys. Chem. Chem. Phys. 2014, 16 (9), 4165– 4175.

TOC Graphic

15 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. Optical characterization. A) Normalized Raman spectra representative of the three kinds of Bi2O3 samples investigated. B) Normalized PL spectra from the three Bi2O3 investigated samples. 83x36mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 16 of 21

Page 17 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 2. STEM-HAADF imaging. A) Low magnification STEM-HAADF image of a group of Bi2O3 nanocrystals from the reference powder. B) Atomic resolution STEM-HAADF image of a Bi2O3 nanocrystal seen along the [001]α-Bi2O3 zone axis. C) Low magnification STEM-HAADF image of the FIB-prepared Bi2O3 pellet. The top bright layer seen in the image is due to the preparation of the sample. D) Atomic resolution STEM-HAADF image from the pellet sample seen along the [001] zone axis of the α-phase. E) Low magnification TEM image of the nanowire sample. F) STEM-HAADF image of an α-phase nanowire along the [010] zone axis of the α-phase. 150x225mm (300 x 300 DPI)

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Experimental EELS. Core-loss and low-loss from the three investigated samples. Note that each spectrum has been divided by the intensity at its maximum. The main peaks of the spectra have been marked and labeled: a1=532.5 eV, b1=534.6 eV, c1=540.2 eV, d1=545.7 eV, a2=11.0eV, b2=18.7eV, c2=29.5eV. 160x197mm (150 x 150 DPI)

ACS Paragon Plus Environment

Page 18 of 21

Page 19 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 4. DFT simulations of EEL spectra. A) Simulated O K Edge spectrum for α-Bi2O3 (plotted in black). The contributions from each non-equivalent oxygen in the unit cell are also plotted. An inset with the αBi2O3 unit cell seen along the [010] zone axis shows the position of each of these oxygen in blue, red and green (the Bi atoms are shown in purple). The energy positions of the spectral features identified in the experimental spectra are also shown as vertical lines. 81x42mm (300 x 300 DPI)

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 5. EELS fitting. Fitting of the core-loss experimental spectra from each α-Bi2O3 sample with the weighted contributions of each non-equivalent oxygen atoms and an energy shift. 191x201mm (150 x 150 DPI)

ACS Paragon Plus Environment

Page 20 of 21

Page 21 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 5. DFT simulations of oxygen vacancies. O K energy shift for each of the DFT simulations.

ACS Paragon Plus Environment