In recent years many studies were published on toxicological effects of nanoparticles (NPs) on human tissue. Many dose-response studies rely on in vitro assays, in which cultured cells are exposed to a suspension of cell-culture medium and NPs. Limbach et al. [1] stated, that the sedimentation process of NPs in this setup is more complex than in the case of microparticles (MPs). While sedimentation of NPs is primarily driven by diffusion, it is mainly gravitational force, which influences MPs. To compare and quantify the sedimentation of particles of both scales, sedimentation studies and simulations with SiO₂ particles were performed in this work. Scanning electron microscopy (SEM) was applied to measure the direct cellular dose, i.e., the sedimented areal densities of particles (AD) on the cells, as a more appropriate definition of dose according to Teeguarden et al. [2]. Simulations are based on ISDD, a computational model by Hinderliter et al. [3].

The samples were prepared as follows: A549 lung cancer cells were seeded onto indium-tin-oxide coated glass substrates in culture plates containing Dulbecco's Modified Eagle's Medium supplemented with fetal calf serum (FCS). After adhering overnight, the cells were incubated with SiO₂ particles (from 70 nm up to 500 nm diameter) in cell culture medium for 1 and 4 hours. Using ISDD and assuming, that sedimented particles are homogeneously distributed, targeted ADs were calculated according to predefined incubation concentrations. After fixation with paraformaldehyde the cells were dehydrated with graded ethanol series, dried by critical point drying and investigated in a FEI Quanta 650 SEM. Particles were imaged with secondary electrons (SE) in intercellular regions between cells. Backscattered electron (BSE) images were taken to detect particles on cells. To increase the BSE contrast, a retarding bias was applied to the sample stage.

Fig. 1a shows a representative SE image of an intercellular region with 200 nm SiO₂ particles after 1 h incubation. The particles appear homogeneously distributed. Fig. 1b depicts a cell surface of the same specimen with a substantially smaller AD (0.15 NP/µm² compared to 0.84 NP/µm² in Fig. 1a) and an inhomogeneous NP distribution. A smaller cellular AD is also observed for all other particles sizes and incubation times. This is shown in Fig. 2a, where measured cellular and intercellular ADs for each specimen are compared. In Fig. 2b, the simulated AD is plotted against the measured intercellular AD for all samples. Calculated and measured ADs agree well within the error bars, which can be considered as a verification of the ISDD model. However, the cellular ADs are significantly lower indicating that another effect must be taken into account. Cellular uptake can be ruled out as an explanation, because focused-ion-beam sectioning of whole cells did not show a high particle density in cells.

Fluorescence microscopy (FM) investigations with rhodamine labelled SiO₂ NPs (Ø = 70 nm) and A549 cells qualitatively confirm the differences between cellular and intercellular ADs observed by SEM (Fig. 3a). The red fluorescence is much stronger in intercellular regions than on cells, however the signals stem from small agglomerates. Dynamic light scattering shows, that this agglomeration is caused by FCS coating of NPs applied before incubation. To study the impact of FCS in more detail, further in vitro experiments with and without FCS precoatings of NPs and/or substrates were performed. Preliminary results indicate, that protein coatings induce an attractive interaction between NPs with their protein corona and surface proteins. Since FM is unable to resolve single NPs, SEM is best suited for confirmation.

SEM is convenient to quantitatively determine ADs of NPs and reveals distinct differences between cellular and intercellular ADs. Our results highlight a major problem of bulk investigation methods relying on lysates, because intercellular and cellular regions cannot be distinguished.

References

[1] L.K. Limbach, et al., Environ. Sci. Technol., 39 (2005), pp. 9370–9376.

[2] J.G. Teeguarden, et al., Toxicol. Sci., 95 (2007), pp. 300–312.

[3] P.M. Hinderliter, et al., Part. Fibre. Toxicol., 7 (2010), p. 36.

We acknowledge the support of the BIF graduate school funded by the Helmholtz Association.

One current and much-debated topic in the characterization of nanomaterials (NM) is the implementation of the recently introduced recommendation on a definition of a nanomaterial by the European Commission [1]. According to this definition [1], a material is a NM when for 50% or more of the particles in the number size distribution, one or more external dimensions is in the size range 1 nm-100 nm. The European NanoDefine project [2] was set up to develop and validate a robust, readily implementable and cost-effective measurement approach to obtain a quantitative particle size distribution and to distinguish between NMs and non-NMs according to the definition [1].

All currently available sizing techniques able to address nanoparticles were systematically evaluated. It was demonstrated that particle sizing techniques like: analytical centrifugation, particle tracking analysis, single-particle inductively coupled plasma mass-spectrometry, differential electrical mobility analysis, dynamic light scattering, small angle X-ray scattering, ultrasonic attenuation spectrometry, but also gas adsorption analysis based on the BET-method can be applied for a screening classification. However, the quality of the results depends on the individual material to be classified. For well-dispersed, nearly spherical (nano)particles most of the sizing techniques can be applied in a quick and reliable way. In contrast, the classification of most real-world materials is a challenging task, mainly due to non-spherical particle shape, large polydispersity or strong agglomeration/aggregation of the particles. In the present study it was shown that these issues can be resolved in most cases by electron microscopy as a confirmatory classification technique [3-6].

Electron microscopy techniques such as TEM, STEM, SEM or TSEM (transmission in SEM) are capable of assessing the size of individual nanoparticles accurately (see Figures 1 and 2). Nevertheless the challenging aspect is sample preparation from powder or liquid form on the substrate, so that a homogeneous distribution of well-separated (deagglomerated) particles is attained. The systematic study in this work shows examples where the extraction of the critical, smallest particle dimension - as the decisive particle parameter for the classification as a NM - is possible by analysing the sample after its simple, dry preparation. The consequences of additional typical issues like loss of information due to screening of smaller particles by larger ones or the (in)ability to access the constituent particles in aggregates [5] are discussed.

By means of practical examples the inherent statistical evaluation of the particle size is highlighted together with all its pitfalls such as setting of a suitable threshold for delimitation of the particle boundaries in the electron micrograph or consideration of systematic (bias) deviations from the true particle size because of evaluation via surface sensitive secondary electron detectors, e.g. In-Lens.

[1] European Commission, Commission Recommendation of 18 October 2011 on the definition of nanomaterial (2011/696/EU). Official J. Europ. Union 54 (2011) p. 38.

[2] www.NanoDefine.eu

[3] F Babick et al, submitted.

[4] K Yamamoto, Microsc. Microanal. 21 (Suppl 3) (2015) p. 2399.

[5] P-J de Temmerman et al, Powder Technol. 261 (2014) p. 191.

[6] P Müller et al, Microsc. Microanal. 21 (Suppl 3) (2015), p. 2403.

[7] The research leading to these results has received funding from the European Union’s Seventh

Framework Programme (FP7/2007-2013) under grant agreement n° 604347-2. (www.nanodefine.eu).

Nowadays, nanoparticles with sizes between 2 to 50 nm become more and more popular, because they can be applied in various fields such as materials science, chemistry, catalysis, medicine or biology. In particular nanoparticles with fancy shapes gain a lot of attention, also because of the low-cost synthesis methods. This study is focused on several types of catalytic nanoparticles (NPs), silver nanoparticles synthesized using green chemistry methods and in particular on their morphological and chemical analysis using HR(TEM). Three-dimensional (3D) catalysts [1] being promising for application in fuel cells were studied. These nanoparticles have a dodecahedron shape with Pt skin at the edges and a Ni core, Figure 1(a). When etching away the core, the remaining empty Pt frame offers a much larger active surface compared to spherical nanoparticles. The morphology of these 3D PtNi particles strongly depends on the synthesis parameters allowing fabricating dodecahedrons, Figure 1(a), core-shell or even dendritic structures, Figure 1(b).

SnO₂ nanoparticles are excellent supports for noble metal NPs, as their combination exhibits good catalytic activity towards ethanol oxidation reaction [2]. Various synthesis routes including polyol and microwave assisted methods allowed producing different SnO₂ NPs, Figure 2(a). The break of the C=C bond in the ethanol molecule occurs at the interface between PtRh and SnO₂ particles, Figure 2(b). Therefore their physical contact is imperative for the effectiveness of the catalyst. Structural aspects and chemical analysis by TEM characterization techniques of the PtRh/SnO₂/C catalysts were analyzed.

Green synthesis method using camomile extract was applied to synthesize silver nanoparticles in order to tune their antibacterial properties merging the synergistic effect of camomile and Ag [3]. Scanning transmission electron microscopy (STEM) revealed that camomile extract (CE) consisted of porous globular nanometer sized structures, which were a perfect support for Ag nanoparticles, Figure 3(a). The Ag nanoparticles synthesized with the camomile extract (AgNPs/CE) of 7 nm average size, were uniformly distributed on the CE support, Figure 3(b). The EDX chemical analysis showed that camomile terpenoids, Figure 3(c) act as a capping and reducing agent being adsorbed on the surface of AgNPs/CE, Figure 3(d), enabling their reduction from Ag⁺ and preventing them from agglomeration. Antibacterial tests using four bacteria strains, showed that the AgNPs/CE performed five times better compared to CE and AgNPs/G samples, reducing totally all the bacteria in 2 hours, Figure 4.

References

1. C. Chen, Y. Kang, Z.Huo, Z. Zhu, W. Huang, H.L. Xin, J.D. Snyder, D. Li, J.A. Herron, M. Mavrikakis, M. Chi, K.L. More, Y. Li, N.M. Markovic, G.A. Somorjai, P. Yang, V.R. Stamenkovic, Science 343 (2014) 1339-1343.

1. A. Kowal, M. Li, M. Shao, K. Sasaki, M.B. Vukmirovic, J. Zhang, N.S. Marinkovic, P. Liu, A.I. Frenkel and R.R. Adzic, Nature Materials 9 (2009) 325-330.
2. M. Parlinska-Wojtan, M. Kus-Liskiewicz, J. Depciuch, and O. Sadik submitted to Bioprocess and Biosystems Engineering.

Acknowledgements

We thank the Center for Innovation and Transfer of Natural Sciences and Engineering Knowledge of the University of Rzeszow and the Division of Materials Processing Technology, Management and Computer Techniques in Materials Science, Institute of Engineering Materials and Biomaterials, of the Silesian University of Technology in Poland for using the TEM instruments. Financial support from the Polish National Science Centre (NCN), grant UMO-2014/13/B/ST5/04497 is acknowledged.

Applications of nanoparticles for chemistry, biochemistry and biomedicine require a detailed understanding of the nature of the nanoparticles and of their surface interactions with both functionalizing organic groups and biological environments. Here we present ongoing studies of the nature and surface interactions of iron oxide nanoparticles (IONPs) intended for magnetic resonance imaging detection and hyperthermia treatment.

The IONPs are synthesized by a novel low temperature aqueous route devoid of surfactant chemistry or capping agents. In addition to routine characterization of the IONPs by TEM, XRD, FTIR, XPS, DLS, and magnetic characterization, we use aberration-corrected high resolution TEM to monitor their structural quality. Because of an observed sensitivity to the electron beam at a 200 kV high tension (HT), we use a HT of 80 kV, with effects of chromatic aberration reduced by implementing a monochromatic “rainbow” illumination in the incident beam [1]. This imaging allows observations of minute changes in surface structural quality. Initially formed particles are rounded, showing signs of slight structural disorder in the first atomic layers at the surface (Fig. 1). In contrast, aged particles have atomically sharp structural ordering at surfaces which are markedly more faceted (Fig. 2). No beam damage effects are observed other than the hopping of atoms on surface ledge sites.

For bioengineering, it is critical to understand how molecules and proteins interact with these IONPs. Beginning with the former, the particles are functionalized with folic acid, the molecule most often used for "nonspecific" targeting. Using the same monochromated, Cs-corrected imaging conditions as above, in Figure 3 we demonstrate the ability to image ligand attachment and surface coverage, similarly to Lee et al. [2]. The folic acid molecules show phase contrast with definition down to about 2 Å; this phase contrast will be compared to simulations for better interpretability. Ligands are observed to form two or three strand thick shells around the surfaces of the IONPs. Using fast frame acquisition, very specific effects of electron irradiation are recorded: the molecules wriggle under the beam, often with one end appearing to detach and reattach to the surface of the IONP. This behavior is explained in terms of relative bonding potentials for either the amino acid or the carboxylic ends of the molecules absorbed to the IONP surfaces.

Once the IONPs are injected into a biological environment it is known that any such functionalization is rapidly replaced by a surface coverage of proteins – the protein corona – which then dictates how the IONPs interact with this environment [3]. To understand this biomedically-important process, for the first time, we obtain in vivo conditions which create a protein corona whose nature mimics that observed for in vitro studies. As well as again using phase contrast imaging of dry samples, we also study the morphology of this corona by simple negative-staining to create contrast between the protein and the carbon support film. Contrary to the commonly held view that the proteins make a uniform encapsulating layer around nanoparticles, their coverage is distinctly patchy (Fig. 4); consistent with the attaching proteins having widely varying sizes. As a next step HAADF STEM tomography will be used to map the 3D morphology of this non-uniform coronal coverage.

With research ongoing on these IONPs for "magnetotheranostics", the other main perspective of this work is to pursue imaging of such interactions in representative liquid environments, particularly using high resolution cryo-TEM imaging to monitor better the nature of organic functionalization in an aqueous environment.

References and Acknowledgements

[1] P.C. Tiemeijer et al., Ultramicrosc. 114 (2012) 72–81.

[2] Z. Lee et al., Nano Lett. 9 (2009) 3365–3369.

[3] M.P. Monopoli et al., Nature Nano. 7 (2012) 779–786.

The authors acknowledge funding from Nano-Tera.ch, project Magnetotheranostics number 530 627.

CoCrMo Metal on Metal (MoM) hip implants were designed to be durable, targeting a better quality of life for young, active patients. Current evidence suggests that such implants can release wear particles and metal ions due to a bio-tribocorrosion process [1,2]. This involves both corrosion of the implant surface itself, which is stimulated by the wear process removing a chromium-rich passivating film, and also mechanical wear of the surface to produce nanoparticulate wear debris that can be spread by lymphatic circulation and subsequently corrode. Both processes can therefore give rise to the release of metal ions which may have an inflammatory or toxic effect on cells and tissues, notably because certain metal ions can complex with proteins and disable the primary function.

This study focuses on the synthesis and characterization of CoCrMo nanoparticles, which mimic metal-on-metal (MoM) wear debris from hip implants. We have used a hitherto unexplored approach of mechanochemical milling to produce a large amount of CoCrMo mimetic wear debris over short time scales [3]. Using TEM, the nanoparticles produced were found to be similar in size, shape and composition to wear debris from CoCrMo hip implants produced both in-vivo and using hip simulators. An efficient separation of nanoparticles from solution prior to free ion concentration analysis by inductively coupled plasma mass spectrometry (ICP-MS) was developed using centrifugation combined with ultrafiltration (2 kDa ultrafilters; Figure 1). This revised preparation method has allowed the identification of the dissolution of specific alloy elements both in hip simulator lubricant systems and in different biological media and pH environments during both dynamic tribocorrosion studies (using milling) and also during static corrosion (Figure 2). The results indicate a much lower dissolution of cobalt than previously reported.  We also identified a significant fraction of Mo ions in solution that could be linked to the preferential binding of Mo by bovine serum albumin (BSA) proteins identified by FTIR and TEM [4; Figure 3]. Electrochemical corrosion tests in the presence of BSA confirm that the proteins play an important role in Mo dissolution from CoCrMo. We suggest that the interaction of Mo-rich surfaces with amide groups in serum proteins and the possible formation of metal carbonyl complexes are important because both can modify biological molecules, potentially altering the function.

In summary, the role of Mo as well as Co ions should be accounted for in the tribocorrosion of CoCrMo implant alloys, particularly in terms of inflammatory and toxicological responses.

[1] D. Cohen (2012) How safe are metal-on-metal hip implants?, Br. Med. J., 344, e1410–e1410.

[2] Y. Yan, A. Neville, D. Dowson, S. Williams, J. Fisher (2009) Effect of metallic nanoparticles on the biotribocorrosion behaviour of Metal-on-Metal hip prostheses, Wear, 267, 683–688.

[3] T.A. Simoes, A.E. Goode, A.E. Porter, M.P. Ryan, S.J. Milne, A.P. Brown, et al. (2014) Microstructural characterization of low and high carbon CoCrMo alloy nanoparticles produced by mechanical milling. J. Phys. Conf. Ser., 522, 012059.

[4]T.A. Simoes, A.P. Brown, S.J. Milne, R.M.D. Brydson (2015) Bovine Serum Albumin binding to CoCrMo nanoparticles and the influence on dissolution. J. Phys. Conf. Ser., 644, 012039.

Nanoparticles associating a noble metal and a ferromagnetic metal are appealing from a magneto-plasmonics point of view, in addition to the problematics of magnetic anisotropy tailoring (interface anisotropy, phase transformation) and of nanoalloy original geometries. Because Co and Au are immiscible in the bulk phase, and since fcc cobalt and gold have highly different cell parameters, chemically separated structures (core-shell type) are expected for nanoparticles.

We have studied CoAu cluster assemblies, with a diameter between 3 and 10 nm, prepared by low energy cluster beam deposition (LECBD) where nanoparticles are formed in out-of-equilibrium conditions by laser vaporization, then deposited on a substrate under ultrahigh vacuum conditions and protected by a capping layer (amorphous carbon, to avoid oxidation). The nanoparticles’ structure and chemical arrangement (see figures) have been investigated by HRTEM, STEM-HAADF and STEM-EELS before and after annealing (2h around 500°C). A mapping of the low energy electronic excitations has also been performed by STEM-EELS, which is a challenge on such small nano-objects.

As prepared particles are found to be inhomogeneous (as deduced from EELS measurements), with interatomic distances always corresponding to pure gold, and they appear to be surrounded by a shell of lighter HAADF intensity which rapidly transforms upon electron beam exposure in STEM. After annealing, a phase separation is observed and CoAu nanoparticles adopt a core-shell structure where, as observed by HRTEM (with a clearly visible difference of inter-plane distances between Co and Au regions), STEM-HAADF and STEM-EELS, an off-centered cobalt core is surrounded by a gold shell (see figure). For both as-prepared and annealed nanoparticles, the Co is not oxidized thanks to the efficient protection of the thin carbon layer. Using low-loss STEM-EELS mapping, we are able to observe some surface contributions, which may reflect collective electronic resonances (surface plasmons), with significant differences between as-prepared and annealed CoAu nanoparticles (in particular, an intense peak is detected on the gold-rich regions, around 6 eV).

These results shed light on the atomic-scale behavior of the Co-Au nanoalloy, here in an amorphous carbon matrix, which will be compared to other dielectric matrices, and will help us to understand the original magnetic properties of these hybrid nanoparticles. 

The unique properties of nanoparticle structures are directly determined by particle composition, size, shape and surface faceting; each offering a tuneable parameter with which one can tailor properties to a given application. Advances in wet-chemistry techniques now enable the synthesis of nanoparticles with shape anisotropy and novel surface faceting that exhibit exciting optical and catalytic properties.  Here, we develop and apply quantitative scanning transmission electron microscopy methods to observe key structural changes that enable anisotropic growth; characterise the nanorod faceting in three dimensions; and probe the relative surface “energies” of both high and low index facets.

Gold nanorods are an archetypal system with which to study anisotropic nanoparticle growth.  Yet despite intense research it remains unclear how or why a single crystal seed particle, with a cubic lattice, grows preferentially in two of six nominally symmetry-equivalent directions. Observations at various stages of gold nanorod growth reveal the onset of asymmetry occurs only in single crystal seed particles that have reached diameters between 4 and 6 nm.  In this size range only, small, asymmetric truncating surfaces with an open atomic structure become apparent, and in the presence of Ag⁺ ions are stabilised, becoming side facets in the embryonic nanorod structure [1,2].   These results provide the first direct observation of the structural changes that break the symmetry of the seed particle and provide key insights into the mechanism of anisotropic growth.

The various facets exhibited by the nanoparticle [3] and their relative surface energies are a crucial driver of shape control whilst also directly determining how the resulting particle will interact with its environment.  We apply a quantitative STEM technique [4] to count the number of atoms in each atomic column as identified in STEM images of a single crystal gold nanorod orientated in two different zone axes.  Using this method we are able to determine the morphology and facet crystallography of the nanorod, finding it is comprised of both high {0 1 1+√2} and low {110}, {100} index side-facets, each of comparable size and shape.  Furthermore, by applying this method at successive time intervals and comparing the images it is possible to quantify atomic movement on the surface and therefore determine the relative stability of different crystallographic facets and the overall stability of the nanoparticle shape [5].  These results provide important information on the effect of surfactants on the relative surface energies of high and low index facets, and shed new light on the growth kinetics of Au nanorods.

Acknowledgements: This work was supported by the Australian Research Council (ARC) grants DP120101573, DP160104679 and LE0454166

1. M. J. Walsh, S. J. Barrow, W. Tong, A. M. Funston and J. Etheridge. ACS Nano, 2015. 9(1). 715-724.
2. W.Tong, M. J. Walsh, P. Mulvaney, J. Etheridge and A.M. Funston.  In preparation 2016
3. H. Katz-Boon, C. J. Rossouw, M. Weyland, A.M. Funston, P. Mulvaney and J. Etheridge. Nano Letters, 2010. 11(1): p. 273-278
4. C. Dwyer, C. Maunders, C. L. Zheng, M. Weyland. P. C. Tiejmeijer and J. Etheridge.  Appl. Phys. Lett. 2012, 100, 191915
5. H. Katz-Boon, M. J. Walsh, C. Dwyer, P. Mulvaney, A.M. Funston and J. Etheridge. Nano Letters. 2015, 15 (3), 1635

Anisotropic metal nanoparticles (NPs), and especially nanorods (NRs) exhibit interesting optical properties, which arise from their localized surface plasmon. Unlike nanospheres, gold NRs have a longitudinal surface plasmon resonance (LSPR) in the visible or near-infrared range of the spectrum. By altering e.g. the shape or the dimensions (aspect ratio) of the NRs, the LSPR can be tuned, which makes them interesting materials for a broad range of light based applications, such as photocatalysis [1], data storage [2] and photothermal applications. The optical properties of gold nanoparticles can be extended even further by introducing a second metal. However, synthesizing bimetallic NRs with a good control over the metal composition and distribution while retaining the rod shape is challenging.

In this study we present bimetallic systems composed of gold-based NRs coated with a protective mesoporous silica layer. We show that it is possible to synthesize bimetallic core-shell nanorods within a mesoporous silica shell, by etching away part of the gold and overgrowing the remaining Au-core with a second metal while precisely controlling the core-size, metal-shell thickness and thus the metal-to-metal ratio [3]. Depending on the choice of metal, different growth behavior was observed. Overgrowth with Ag resulted in a smooth shell whereas the Pt and Pd metal shells had a rough morphology (Figure 1). The different types of bimetallic NRs were characterized in detail with advanced electron microscopy techniques such as Energy-dispersive X-ray spectroscopy (EDX), high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and electron tomography. Subsequently, we used these silica coated bimetallic core-shell structured rods as a starting material to make fully alloyed NRs, whereby the two metals were mixed via thermal treatment without loss of anisotropy [4]. The alloying process was followed in detail with in-situ HAADF-STEM heating and EDX measurements by making use of a special heating holder (Figure 2).

[1]           Suljo Linic et al., Nat. Mater. 14, (2015), 567-576. 
[2]           P. Zijlstra et al., Nature, 459, (2009), 410–413
[3]           T.S. Deng, J.E.S. van der Hoeven, A. Yalcin, H.W. Zandbergen, M.A. van Huis, A. van Blaaderen, Chem. Mater. 27, (2015), 7196-7203. 
[4]           W.A. Albrecht and J.E.S. van der Hoeven, T. Deng, P.E. de Jongh, A. van Blaaderen., submitted

Since the first nanoparticle syntheses, research has focused on producing various nano-catalysts of different compositions and uses. Over the years, the importance of their shape, hence their exposed facets, grew in interest. The differences between facets, may sometimes remain unclear and controversial, but more and more effects in terms of reactivity and/or selectivity are shown in recent works. In this framework, rod and cubic-shaped cerium oxide nanoparticles were synthesized using a hydrothermal microwave-assisted process. Systematic TEM imaging and analysis reveal that rods are progressively replaced by cubes with increasing synthesis time (with fixed temperature and pH). Besides, a more careful look at these cubic-shaped particles highlights a slight anisotropy - depending on the zone axis they are imaged along - leading us to reconsider their nucleation and growth mechanisms.

These processes were first studied using radiation damages induced in ceria rods’ complex microstructures under focussed electron beam. Their periodical packing composed of single crystals aligned along a common low-index axis were firstly reduced and new nanometer-sized domains were formed with different (micro)structures. These zones tend to split from the initial rod in case of severe irradiation conditions and may, in-fine, lead to cuboidal particles via a solid-state transformation process.

As-synthesized cubes were then characterized by means of high resolution TEM coupled with molecular dynamics simulations. As expected, they exhibit single crystalline structure and {001} lateral facets are enclosed by {111} sharp corners and {011} average flat edges made of alternating {111} steps. More surprisingly, whatever the particle size might be, projected edges and corners width remains constant and corresponds to the diagonal of a four by four square formed by {002} planes. By keeping constant this truncation, surface plane ratios can be tuned only by adjusting ceria particles synthesis conditions.

Further on, when submitted to reducing conditions, both simulations and experiments tend to show a partially reversible flattening of the {011} edges combined with the observation of superlattice reflections and Moiré fringes. This indicates that oxygen anions have been removed from the fluorine CeO₂ structure inducing an ordered oxygen vacancies lattice. Finally, oxygen vacancies movement into the ceria network will be discussed.

Most of our current understanding of the deformation mechanisms active in nanocrystalline (nc) metals stems from in-situ deformation experiments on bulk materials using x-ray diffraction (XRD). However, XRD cannot directly resolve the local deformation processes, e.g. grain growth or twinning. For a local analysis, these processes are traditionally investigated using BF/DF-TEM. However, varying contrast due to local orientation changes, bending and defects during in-situ BF-TEM straining experiments make an accurate interpretation for nanometer sized grains difficult. On the other hand, Automated Crystal Orientation Mapping (ACOM-TEM) allows for the identification of the crystallographic orientation of all crystallites with sizes down to around 10 nm, well below the limit of electron back scatter diffraction (EBSD)¹. Using template matching to reveal the crystal orientation, the ASTAR (Nanomegas) ACOM-TEM analysis offers an angular resolution that is nearly as good as EBSD^1,2.

Recently, ACOM-TEM imaging in STEM modus was combined with in-situ straining inside a TEM^3–5. This combination was the key to new data evaluation based on orientation maps. By tracking individual crystallites through a straining series the change of their orientation can be evaluated in order to distinguish between an overall crystallite rotation and sample bending. In addition, twinning/detwinning and grain growth can be directly followed and the automatic data evaluation leads to user independent quantitative statistical information such as grain size distribution³.

Recent investigations revealed some challenges using ACOM-TEM for in-situ experiments if there are overlapping crystallites. Overlapping crystallites lead to superimposed diffraction patterns that confuse the matching procedure. Tilting nc Pd in-situ and tracking the changes using ACOM-TEM, it became apparent that some orientations are more dominant than others during the matching procedure. Further, we present an ambiguity filter that reduces the number of pixels with a '180° ambiguity problem' (Fig. 1). The challenges discussed here for orientation mapping of nc materials do not only appear with ACOM-TEM, but are mostly an inherent problem of any TEM projection technique. However, using ACOM-TEM these limitations become apparent and can be properly analyzed, e.g. by mapping the rotation of many crystallites for a given area of interest. This enables to detect sample bending or tilting in a (in-situ) series of orientation maps, which cannot be measured by BF/DF-TEM.

The ACOM-TEM measurement and evaluation routine was applied to magnetron sputtered Pd_xAu_1-x thin film samples of about 50 nm. Grain growth and grain fragmentation as well as twinning and detwinning have been observed to take place simultaneously at different locations. In addition, large angle grain rotations with ~39° and ~60° occur that can be related to twin formation, twin migration and twin-twin interaction as a result of partial dislocation activity (Fig. 2). Furthermore, plastic deformation in nanocrystalline thin films was found to be partially reversible upon rupture of the film. In conclusion, conventional deformation mechanisms are still active in nanocrystalline metals, however, with different weighting than in conventional materials with coarser grains.

We would like to acknowledge Christian Brandl, Edgar Rauch, Florian Bachmann, Ralf Hielscher, Ankush Kashiwar. This work was supported by the DFG under grant FOR714 and Karlsruhe Nano Micro Facility (KNMF).

References

(1)            Rauch, E. F.; Portillo, J.; Nicolopoulos, S.; Bultreys, D.; Rouvimov, S.; Moeck, P. Zeitschrift für Krist. 2010, 225 (2-3), 103–109.
(2)            Zaefferer, S. Cryst. Res. Technol. 2011, 46 (6), 607–628.
(3)            Kobler, A.; Kashiwar, A.; Hahn, H.; Kübel, C. Ultramicroscopy 2013, 128, 68–81.
(4)            Kobler, A.; Kübel, C. Imaging Microsc. 2014, No. 1.
(5)            Mompiou, F.; Legros, M. Scr. Mater. 2015, 99, 5–8.

Charge distribution resulting in the formation of a space charge zone (SCZ) in ionic materials has a critical role on functional properties [1].  Even though significant advances in theoretical models have been accomplished, experimental evidence in nanoscale granular materials is indirect.

Here, we investigated the distribution of cations and defects on the formation of a SCZ in a nanoscale granular model system of non-stoichiometric MgO∙nAl₂O₃ (MAS, n= 0.95 and 1.07). The SCZ was investigated experimentally by electron energy-loss spectroscopy (EELS) and off-axis electron holography (OAEH).

EEL spectra were collected along directions perpendicular to grain boundaries (GB’s), from which the magnesium-to-aluminum relative cation concentrations were calculated, as presented in Fig.1. We found that regardless of annealing processes, the vicinity of GB’s of the Mg rich spinel has excess Mg⁺² cations while the vicinity of GB’s of the Al rich spinel has excess of Al⁺³ cations. Additionally, the cation distribution shows strong dependency on the grain size. For non-stoichiometric MAS, cation concentration is proportional to the defect concentration, because deviation from stoichiometry results in adjacent defects that compensate for the electric charge [2, 3, 4]. In both materials, the cation distribution is inhomogeneous for grains smaller than 40 nm. For larger grains, the defect concentration approaches the bulk value at the center of the grain. Furthermore, excess of Mg (Al) cations at the vicinity of the GB decreased with increase of grain size. Maier et al. [1] calculated that for grain size at the scale of the Debye length (estimated at 9nm for non-stoichiometric MAS studied here [7]), the GC is no longer electrically neutral, instead influenced by accumulation or depletion of charge at the boundaries.

Due to the lack of accurate values for defect formation energy [5, 6], we applied OAEH to measure directly the electrostatic charge distribution in nano-sized MAS. We show that charge distribution and the buildup of electrostatic potential between GB and core are linked to the spatial distribution of defects rather than the overall composition of MAS (Fig. 2). At the vicinity of GB’s, excess Mg⁺² or Al⁺³ cations accumulate depending on the composition, the magnitude of which increases with decreasing grain size. Indeed, the potential distributions show the relation between the excess cation species, grain size and the Debye length, in agreement with theoretical models [1].

References:

[1] J. Maier, Prog. Solid State Chem. 23, 171-263 (1995).

[2] M. Rubat du Merac et al., J Am Ceram Soc. 96 (2013) 3341-3365.

[3] S.T. Murphy et al, Philosophical Magazine. 90 (2010) 1297-1305.

[4] Y. Chiang, W.D. Kingery, J Am Ceram Soc. 73 (1990) 1153-1158.

[5] K. Lehovec, J. Chem. Phys. 21, 1123-1128 (1953).

[6] K. Kliewer & J. Koehler,. Phys. Rev. 140, A1226 (1965).

[7] M. Halabi, V. Ezertzky, A. Kohn and S. Hayun, submitted.

Selective laser melting (SLM) is an additive manufacturing technique where successive laser beam passes are used to melt metal powder which forms a solid layer on solidification with high densification, little material waste, and large design freedom [1]. The application of SLM to repair high temperature components that often need reconditioning requires an understanding of the microstructural and compositional developments of the chosen material throughout the SLM process. Alloy 718 is a Ni-Cr-Fe-Nb-Ti-Al alloy used in applications where high strength is needed while maintaining corrosion and creep resistance, making this alloy a prime candidate for SLM structural and compositional characterization. Precipitation hardening is one of the primary strengthening mechanisms of Alloy 718, where intermetallic phases of L1₂-ordered Ni₃₍Al, Ti) (γ’) or D2₀₀-ordered Ni₃(Nb, Ti) (γ’’) may form coherent precipitate particles in the face-centered cubic matrix (γ) [2]. Additional phases that may be present in the microstructure of Alloy 718 include D0_a-ordered Ni₃Nb (δ), MC, M₆C, M₂₃C₆, and (Ni, Cr, Fe)₂(Nb, Mo, Ti) Laves [3, 4]. The complex microstructures in this alloy system are further complicated by the multiple heating and cooling cycles present in the SLM process, thus requiring characterization on the nanoscale in order to understand the microstructural development during processing.                           

Analytical electron microscopy allows the identification of the particular microstructural components on the micro and nano scales.  Alloy 718 is of particular interest in that the γ’ precipitates can nucleate on the (001) surface of γ’’ precipitates in the as-SLM condition. The structure and chemical composition of these precipitates was investigated through X-ray energy dispersive spectrometry (XEDS) using an aberration-corrected FEI Titan G2 ChemiSTEM equipped with the Super X EDX X-ray detector configuration. Figure 1 shows a γ’’ precipitate with γ’ precipitates on the two elongated sides of the γ’’ in both scanning transmission electron microscopy (STEM) bright-field (BF) and high-angle annular dark-field (HAADF) imaging modes. The understanding the formation of γ’/γ’’ requires both structural information about the interface and chemical analysis across the interface of the two precipitates. Figure 2 displays 4 XEDS spectrum images of Ni, Nb, Ti, and Al showing the location of these elements throughout the precipitates present in the γ matrix. Quantitative XEDS analysis was performed on an as printed Alloy 718 specimen where the γ matrix composition was found to be 50.5 wt % Ni, 1.4 wt % Nb, 0.3 wt % Al, and 0.09 wt % Ti with γ’ and γ’’ having compositions of 66.6 wt % Ni, 7.13 wt % Nb,  3.18 wt % Ti, and 2.4 wt % Al and 65.0 wt % Ni, 25.4 wt % Nb, 0.37 wt % Al and 3.6 wt % Ti, respectively.             

References:

[1] Gu, D.D., Meiners, W., Wissenbach, K., Poprawe, R., Int. Mater. Rev., 2012, 57, 133-164.

[2] Azadian, S., Wei, L.Y., and Warren, R., Mater. Charact., 2004, 53, 7.

[3] Burke, M.G. and Miller, M.K., J. de Physique, 1989, 50, C8 395-400.

[4] Rama, J.G.D., Reddya, A.V., Raob, K.P., Reddyc, G.M., Sundar, J.K.S., J. Mater. Process. Tech., 2005, 167, 73.

The kinetic behavior of a Fe–24 Ni–0.4 C (weight percent) martensitic steel during aging at room temperature has been investigated by transmission electron microscopy. Electron diffraction, coupled with imaging techniques in a transmission electron microscope, including high resolution studies, have been used at various stages during the aging process, in order to characterize the microstructural features of the analyzed samples.

Rapid cooling (quenching) in liquid nitrogen of the austenitic state of the above-mentioned material leads to the formation of a martensitic phase which, at room temperature (RT), begins to transform, through a process called spinodal decomposition. As a result of this process, a modulated structure is formed, in which carbon-rich regions (precipitates) occur in a periodic manner throughout the matrix, leading to the presence of diffuse streaks or satellite spots around each fundamental (matrix) reflection on the electron diffraction patterns, see Figure1. This process is thus accompanied by a reduction of the tetragonality of the martensitic phase (bct), which evolves towards the formation of a cubic structure (bcc), see Figure2, corresponding to alpha-iron (ferrite). After long ageing times, Fe3C is observed. The presence and structure of intermediate carbides is also studied.  

The microstructure of the carbon-rich phase is also analyzed by electron diffraction. The evolution of the tetragonality with time, translated as the c/a ratio, which is also function of the carbon composition, is studied. Moreover, the variation of the distance between the carbon-rich precipitates (in fact, the periodicity of the modulated structure), along with the evolution of their width and length, are also observed.

Thanks are due to the Clym ( Centre Lyonnais de Microscopie) for the access to the TEM 2010F.

Most materials used in nuclear reactors are prone to radiation damage and their mechanical properties degrade with time, limiting their service life. Defects produced by high energy particle radiation can be highly mobile at high temperatures with their mobility being influenced by the local stress fields associated pre-existing defects and grain boundaries.  The relaxation of radiation defects can produce clusters and induce defect diffusion to interfaces and other pre-existing defects, where they can be absorbed and alter the microstructure that can detrimental to the material’s mechanical properties.  Radiation mechanisms that promote biased point defect diffusion can induce significant microstructural change and degrade properties. For instance, interstitials, being more mobile than the vacancies, are quickly absorbed by nearby dislocations, inducing creep by dislocation climb and dislocation multiplication that results in work hardening and embrittlement. Likewise, a small excess of remnant vacancies can agglomerate, leading to the formation of voids that cause swelling, increased residual stresses, the formation of microcracks, and the eventual failure of the material. The long-term stability of a microstructure under irradiation depends on its neutrality towards biased point defect dynamics. Materials with a high density of sites that can act as sinks for the point defects produced by high-energy particles would thus be an enabling technology for reliable and clean nuclear energy. 

We have explored radiation damage mechanisms in one such material, nanotwinned copper, which has a microstructure comprising a high density of coherent twin boundaries distributed in a “latter-like” morphology within columnar grains having high misorientation angle intercolumnar boundaries. By performing in-situ particle irradiation in a MeV TEM and characterizing the evolution in the microstructure using the Nanomegas ASTAR system, we have studied the irradiation induced damage and correlated the radiation induced grain boundary migration and void preferential void formation to the local grain boundary character and network.  In-situ observations were made at two different temperatures, room temperature and temperature above 573 K that stimulate void nucleation and growth. Figure 1 shows an example montage of the irradiation induced void formation and grain boundary migration with increased time and dose.  The radiation induced grain boundary migration (RIGM) of several high-angle random and low–angle grain boundaries within the irradiated zone was observed at both temperatures.  However, low-energy CSL type boundaries were not observed to migrate even at elevated temperatures as well as grain boundaries connected to junctions that contain one or more Σ3 boundaries. Thus, coherent twins are thought to stabilize the microstructure from radiation induced coarsening.

At temperature above 573 K during high-energy electron irradiations, copious amounts of voids were observed to nucleate and growth. They were predominantly observed to nucleate and grow in the regions with a high density of the coherent twin boundaries (CTBs).  Very few voids were observed around HARGBs and LAGBs as shown in Figure 2, and only a limited of number of small voids were observed in the vicinity of incoherent twin boundaries. Non-coordinated boundaries having excess free volume are efficient sinks for both vacancies and interstitials, and thus do not stimulate biased point defect diffusion that promote void formation, explaining the observed lack of voids in the vicinity of HARGBs. On the other hand, the observed high density of large voids in regions with finely spaced CTBs may arise from the biased diffusion of point defects. Interstitials being more mobile that are able to diffuse through CTBs can rapidly migrate and annihilate at free surfaces or other sinks such HARGBs. This leaves an excess of vacancies in the vicinity of the CTBs which can coalesce and cause the nucleation of voids.  Given these observations, a nanocrystalline material having a high fraction of CTBs, though being resistant to radiation induced coarsening, may not possess the ideal topology for radiation resilience and may be prone to increased void formation and swelling.

Over the last years, progress in nanomaterials design and manufacturing has revolutionized technology and opened up prospects for many scientific researches. The investigations of material properties (optical, electronic, mechanical...) at small scales have revealed amazing behaviors, different from those currently observed in bulk samples. For instance, silicon, which is known to behave at room temperature as a brittle material, shows an unexpected ductile behavior when the sample size decreases below a few hundreds of nanometers [1]. The mechanisms leading to this phenomenon remain, however, poorly understood. In this context, this research project aims at investigating in more details the deformation behavior of silicon nanopillars by combining experimental techniques (SEM, FIB, HRTEM) and molecular dynamics simulations. In this work, various nanopillars with different orientations and diameters (from 100 nm to 1 µm), were patterned by Reactive-Ion Etching and FIB micromachining. These pillars were then compressed with a slow-strain-rate (10^-4 s^-1) at room temperature using a nanoindenter equipped with a flat punch and operated in displacement-control mode. Post mortem observations of deformed nanopillars performed by SEM and TEM reveal the activation of different slip systems. The comparison between experimental and simulated HRTEM images notably evidences the simultaneous propagation of partial and perfect dislocations in {111} planes. In addition, unexpected plastic events have also been observed in {113} planes. On the basis of the microscopic observations, various possible deformation mechanisms involved during the nano-compression of the pillars are proposed.

[1] F. Östlund et al., Brittle-to-Ductile Transition in Uniaxial Compression of Silicon Pillars at Room Temperature, Adv. Func. Mat., 19, p1(2009).

This work is performed within the framework of the ANR-funded research project « BrIttle-to-Ductile Transition in Silicon at Low dimensions » (ANR-12-BS04-0003-01, SIMI4 program).

Due to the differences in lattice parameters and thermal expansion coefficients, Ge films grown on Si substrates suffer from a high density of threading dislocations, cracks and wafer bowing. One method to avoid these problems is to grow Ge crystals on (001)-Si pillars by low energy plasma enhanced chemical vapor deposition (LEPECVD). With this strategy, threading dislocations of 60° type with Burgers vector (b) b=1/2 can be avoided in the bulk as they end at the sidewalls of the Ge crystals. However, the misfit strain is not totally relaxed; it is accommodated by misfit dislocations (MDs) of 60° and 90° types, while planar defects have not been reported in these Ge crystals.

This work presents a detailed analysis of defects formed to relax the misfit strain between the Ge crystals and Si pillars using high angle annular dark-field scanning transmission electron microscopy (HAADF-STEM).

In Ge/Si interface, the misfit strain is accommodated by 60° and 90° MDs. The 90° MDs are Lomer MDs lying on (001) planes which are formed by interaction of two 60° MDs. These pairs of MDs form steps at the interface leading to an atomic roughness.

Most interestingly, besides the MDs, there is a high density of planar defects which has not been reported before. These 2D defects are formed at the Ge/Si interface and extend between 3 to 40 nm along the {111} planes. We observed coherent twin boundaries (CTB), incoherent twin boundaries (ITBs) and stacking faults (SF) bounded by partial dislocations (PDs). Figure 1a shows CTBs of the ∑3{111}-type (red arrows), the distance between them is 2.95 nm. The inset of Fig. 1a is its Fourier transform (FFT), the twin plane correspond to GA(1-1-1)=GB(1-11) (GA and GB are grains A and B) and the corresponding planes for GA and GB are (1-11) and (1-1-1) respectively. The CTBs ends inside the crystal with an ∑3{112}-ITB (Fig 2b) having 6, 7, 5 rings along the boundary. Figure 1c shows an image of the CTB at the interface, the yellow arrows are the PDs accommodating the mismatch in the CTB. They are 30° Shockley PDs since the Ge grows in compressive films. The CTB are formed nearby the steps formed by the pairs of MDs.

Figure 1d shows a misfit partial dislocation (MPD) at the end of a SF.  The Burgers circuit around the dislocation gives b=1/6[1-12] which corresponds to a Shockley PD. The stacking sequence of the SF changes from AaBbCcAaBbCc… to AaBbCcBbAaBbCc (Fig. 2e). The SF is of extrinsic type since there is an extra plane in the stacking sequence. For the compressive Ge, the Shockley MPD with an extrinsic SF corresponds to 90° at the interface which in addition has  a perfect 60°  dislocation close to the PD. Figure 2f shows two SFs in (1-11) and (1-1-1) planes, they annihilate each other forming a stair rod dislocation (SRD). The stacking sequence of the SFs changes from AaBbCcAaBbCc… to AaBbCcBbCcAaBb… The SFs are of intrinsic type since there is a missing plane in the stacking sequence. The configuration of the MPDs at the interface in Ge with intrinsic SFs is 30° MPD at the interface and 90° MPD in the Ge. The two 90° PDs in the Ge interact with each other and form the SRD with b=1/3[-110].

SFs are also found in the Ge crystal (Fig. 1g). The Burgers circuit around the PDs gives b=1/6[1-12] which corresponds to Shockley PDs. These intrinsic SFs (Fig. 1h) are formed by the dissociation of perfect 60° dislocations.

We conclude that the misfit strain between Si and Ge is accommodated by 60° and 90° MDs, the 60° MDs splitting to form MPDs and CTB. The MPDs form extrinsic or intrinsic SFs. Intrinsic SFs are interacting with each other forming SRD. Overall, it can be summarized that the defect chemistry in Ge pillars is more complex than reported in earlier studies. This might have significant impact on the electronic properties of these binary semiconductor heterostructures.

Organic-inorganic metal-halide perovskite solar cells are emerging as a promising photovoltaic technology to harvest solar energy, with latest efficiencies now surpassing 22%¹ - an impressive increase from the first reported value of 3% in 2009.² In addition to low manufacturing costs, the optical properties of such cells can be tailored to form efficient tandems when combined with high-efficiency silicon solar cells.³ A typical perovskite cell structure as investigated here is based on a methylammonium lead trihalide absorber (MAPbI₃) that is placed between hole- (Spiro-OMeTAD) and electron-selective contacts (a fullerene-based material).

While new record efficiencies are frequently reported, the commercial application of this solar cell technology remains hindered by issues related to thermal and operational stability. Different mechanisms that are still debated modify cell properties with time, temperature, illumination and general operating conditions.⁴ In order to correlate applied voltage (V) and resulting current (I) to changes in active layer chemistry and structure on the nanometre scale, we performed both ex situ and in situ transmission electron microscopy (TEM) experiments, involving (scanning) TEM (STEM) imaging, selected-area electron diffraction, energy-dispersive X-ray spectroscopy and electron energy-loss spectroscopy. Samples were prepared by focused ion beam (FIB) milling, with exposure to air during transfer to the TEM minimised to <5 minutes to reduce any degradation of MAPbI₃.

First, the effects of exposure to air and electron beam irradiation were assessed in relation to FIB final thinning parameters. Once adequate sample preparation and observation conditions were identified, changes in morphology during cell characterisation were assessed ex situ by comparing lamellae extracted from as-manufactured and tested cells and then in situ by contacting FIB-prepared samples to a microelectromechanical systems (MEMS) chip mounted in a TEM specimen holder⁵ (Fig. 1a). Cell manufacturing parameters led to iodine diffusion into the hole collector, with the width of this diffused layer remaining constant during I-V characterisation. Similarly to ex situ experiments, the MAPbI₃/Spiro interface was observed to delaminate during in situ electrical measurements, resulting in the presence of a ~5 nm Pb-rich layer on the hole-transparent-layer side (Figs. 1b-c). In addition, PbI₂ nanoparticles were observed to nucleate within the MAPbI₃ layer at the hole-collector interface and at the positions of structural defects (Figs. 1b-d).

Overall, the active MAPbI₃ layer was observed to be sensitive to sample preparation, exposure to air, observation conditions and I-V stimulus, resulting in the need for great care to deconvolute each effect. Different mechanisms that may all contribute to the decrease in efficiency of the cell were identified both ex situ and in situ, including ionic migration, PbI₂ formation and local delamination of interfaces.

Acknowledgments

Financial support is gratefully acknowledged from the Swiss National Science Foundation (SNSF) Sinergia project DisCO and the European Union Seventh Framework Program under Grant Agreement 312483 – ESTEEM2 (Integrated Infrastructure Initiative – I3). The authors also wish to thank Max Kruth, Doris Merteens and Vadim Migunov.

1         NREL research cell efficiency records, 2016, http://www.nrel.gov/ncpv/.

2         A. Kojima et al., J. Am. Chem. Soc., 2009, 131, 6050–6051.

3         J. Werner et al., J. Phys. Chem. Lett., 2016, 7, 161–166.

4         S. D. Stranks et al., Nat. Nanotechnol., 2015, 10, 391–402.

5         M. Duchamp et al., Microsc. Microanal., 2014, 20, 1638–1645.

Among complex oxides Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-d (BSCF) exhibits excellent oxygen permeability due to its high oxygen non-stoichiometry. However, the favoured cubic perovskite phase tends to partly decompose in the desired operation temperature between 700 and 900 °C. Various secondary phases with high Co-content and low O-non-stoichiometry are formed close to grain boundaries [1]. The decomposition can be correlated with thermally activated O-vacancies. By lowering the temperature (from e.g. 1000 °C), the O-vacancy concentration decreases forcing the multivalent transition metal atoms (mainly Co) to increase their valence state to maintain charge neutrality. This leads to the collapse of the cubic phase because the tolerance factor shifts into the hexagonal regime [2]. The formation of secondary phases leads to a substantial reduction in the O-permeation [2].

To overcome this issue, doping was suggested which proved to be beneficial for several different cations (e.g. Zr, Y) [3,4]. Among these candidates, Y was chosen due to its monovalent character and large ionic radius which are both beneficial for stabilizing the cubic phase. Transmission/scanning electron microscopy (TEM/SEM) investigations were carried out on 1% (BSCF1Y) to 10% Y (BSCF10Y) B-site doped samples to analyse the phase constitution. O-permeation measurements were performed to assess the long-time permeability. Due to the intermediate ionic radius of Y the lattice position was checked using atom location by channeling enhanced microanalysis (ALCHEMI) [5].

BSCF tends to form a variety of secondary phases. Especially at high temperatures Co tends to diffuse out of the cubic lattice. This happens during the sintering process at high temperatures of 1050…1150 °C when CoO grains form close to grain boundaries. At lower temperatures partial phase decomposition to Ba_n+1Co_nO_3n+3(Co₈O₈) (n ≥ 2, BCO), Co₃O₄ and Ba_0.5+xSr_0.5-xCoO₃ (hexagonal) phases was demonstrated. Valence-state analysis of Co employing the Co-L_2,3 white-line distance method [7] revealed that secondary phases with higher Co-valence state tend to be more stable at lower temperatures. Permeation measurements of BSCF10Y show that the degradation of O-permeation is reduced compared to undoped BSCF. This can be understood by the suppression of secondary phase formation in BSCF10Y at temperatures ≥800 °C (cf. Fig. 1 and Fig. 2). Secondary phases like BCO (cf. Fig. 3) and Co_xO_y are completely supressed by ≥ 3 at% Y-doping. The surfaces of permeation pellets show only minor amounts of BaSO₄ which can be attributed to sulphur impurities in the feed gas. At lower temperatures (~700 °C) small volume fractions of the hexagonal phase are formed even in BSCF10Y (cf. Fig. 4). However, the volume fraction was negligible compared to the amount of secondary phases in undoped BSCF revealing a stabilizing effect on the cubic BSCF phase even at 700 °C.

ALCHEMI experiments showed unintended partial Y-occupation of up to 55 % on the A-site [8]. Therefore, Y doping is expected to generate B-site vacancies. These can contribute to the increased stability of the cubic phase because BSCF seems to tend towards B-site deficiency. To verify this hypothesis BSCF with 5% B-site deficiency was intentionally prepared which indeed showed less Co-outdiffusion and less pronounced secondary phase formation.

References

[1] P. Müller et al., Chem. Mater. 25, 564–573 (2013).

[2] S. Švarcová et al., SSI 178, 1787–1791 (2008).

[3] S. Yakovlev et al., Appl. Phys. Lett. 96, 254101 (2010).

[4] P. Haworth et al., Sep. Purif. Technol. 81, 88–93 (2011).

[5] J.C.H. Spence & J. Taftø, J. Microsc. 130, 147–154 (1983).

[6] M. Arnold et al., J. Membrane. Sci. 293, 44–52 (2007).

[7] P. Müller et al., Microsc. Microanal. 19, 1595–1605 (2013).

[8] M. Meffert et al., Microsc. Microanal. 22, 113-121 (2016).

[9] Financial support from the German Science Foundation (DFG) is gratefully acknowledged.

Over the last few years, the interest in perovskite based solar cells has boomed, due to a surprisingly fast increase in terms of their efficiency that has now reached values comparable with established photovoltaic technologies. Nevertheless, the understanding of the optoelectronic properties of such nanostructured materials is still an open problem and issues related to their stability and degradation pathways represent the current hot topic in this research area.

Organic-inorganic solar cells present a complex composition as well as a composite structure that are strongly related to device fabrication. In this work four processing methods of the organic-inorganic halide perovskite have been investigated, varying the deposition method (single step or double step) and the atmospheres in which the synthesis has been carried out¹. We compared interface quality, morphology, chemical composition and efficiency of the resulting cells. A fluorine doped tin oxide (FTO) glass layer was coated first by a compact (hole blocking) TiO₂ layer and then by a nanoporous TiO₂ layer. The TiO₂ scaffold was infiltrated and capped by a methyl-ammonium lead iodide. Spiro-MeOTAD, acting as hole transport layer was spin coated on the perovskite layer; Au contacts were deposited on top. The devices were analysed using several complementary characterisation techniques: Scanning Transmission Electron Miscroscopy (STEM) used in conjunction with EDX analysis, time of flight secondary ion mass spectrometry (ToF-SIMS) and X-rays photoelectron spectroscopy (XPS). In particular, the use of FIB specimen preparation, combined with analytical transmission electron microscopy, represents a powerful and versatile tool for the characterization of devices based on hybrid composites with nano- and micro-scale structural and chemical features. EDX maps were then treated using multivariate analysis in order to optimise signal-to-noise ratio and obtain high quality EDX maps. The application of this method plays a key role in the analysis of data acquired with low electron doses to minimise specimen damage, and is unique in allowing the identification of different materials present in the sample as compounds rather than individual elements.

Having fully characterised the devices, we investigated the different degradation processes that affect the perovskite based solar cell. Air exposure and temperature were proven to be  responsible for the drastic reduction in device performance. STEM-EDX analysis was thus used in conjunction with the in situ heating, bringing the cells to 250 °C². The main result was the direct observation of elemental migration, particularly evident in iodine maps, and the simultaneous formation of metallic Pb precipitates, resulting in the depletion of the initial perovskite region.

Lastly, we studied the changes in chemical composition and morphology after 2 months of air exposure in dark³. In this case the principal variations in local elemental composition and in morphology concerned respectively the migration of lead and iodine into the HTL layer towards the Au electrode, resulting in a severe degradation of the photoactive layer and the physical formation of bubbles in the Spiro-OMeTAD.

[1] F Matteocci, Y Busby, JJ Pireaux, G Divitini, S Cacovich, C Ducati, A. Di Carlo, Interface and composition analysis on perovskite solar cells, ACS applied materials & interfaces 7 (47), 26176-26183 (2015)

[2] G. Divitini, S. Cacovich, F. Matteocci, L. Cinà, A. Di Carlo, C. Ducati, In situ observation of heat-induced degradation of perovskite solar cells, Nature Energy, 15012 (2016)

[3] S. Cacovich, G. Divitini, C. Ireland, F. Matteocci, A. Di Carlo, C. Ducati, Study of ageing processes in perovskite solar cells using Scanning Transmission Electron Microscopy and multivariate analysis, Under review

 CdTe is one of the most promising materials for thin film solar cells due to its near-optimum bandgap, high efficiency and low cost of fabrication. In order to reduce the contact resistance, it is common to deposit a thin Cu layer at the back of the cell. However, Cu is found to diffuse easily into the p-type CdTe layer and even to the n-type CdS layer. This diffusion causes a problem because too much Cu diffusion has been shown to cause performance degradation [1-2]. Therefore it is important to figure out the pathways of Cu diffusion, and the configurations that Cu stabilizes inside the CdTe solar cells.

 In this research, we employed aberration-corrected Scanning Transmission Electron Microscopy (STEM) combined with Electron Energy Loss Spectroscopy (EELS) or Energy Dispersive X-Ray Spectroscopy (EDX) to probe Cu diffusion in CdTe solar cells. We first investigated the cells with a typical Cu-annealing process at 150° C for 45 min in dry air under ambient pressure, but did not find very much Cu diffusion. Therefore we extended the annealing at 150° C for 8 hours in vacuum using our sample-baking system. The vacuum annealed samples show clear Cu diffusion traces along some of the grain boundaries (GBs). However not all the GBs became Cu-enriched, indicating that Cu diffusion may have a preference for different GBs (Fig.1). Moreover, STEM-EDX reveals Cu-rich precipitates embedded inside high-density twin boundaries in CdTe (Fig.2). The Cd concentration decreases while Te does not change on the precipitates, and no other elements were found, indicating the precipitates are most likely to be Cu_xTe_y. The Cu_xTe_y precipitates exist all the way along the high density of twin boundaries from the contact layer to the CdS layer, demonstrating another pathway for Cu diffusion.

 Similar Cu_xTe_y precipitates were found at the CdTe/CdS interface. A precipitate located at the edge of the STEM specimen without overlapping CdTe or CdS was used to quantify the ratio of Cu to Te via both EELS and Z-contrast image intensity. Both results show that the ratio of Cu to Te is slightly lower than 2:1, therefore the precipitate can be described as Cu_2-xTe. Moreover, the atomic structure of the precipitates show variations from the normal CuTe structure, for example formation of 2×2 supercells (Fig.3). More detailed results will be shown and the photovoltaic behavior of the Cu_2-xTe precipitates will also be discussed.

References:

[1] K.D.Dobson, et al, Stability of CdTe/CdS thin-film solar cells, Solar Energy Materials & Solar Cells, 62, 295 (2000)

[2] S.H.Demtsu, et al, Cu-related recombinations in CdS/CdTe solar cells, thin solid films, 516, 2251 (2008)

[3] This research was supported by the US DOE, Office of Energy Efficiency and Renewable Energy (F-PACE, DE-FOA-0000492), (CL, NP, YFY, SJP), the Office of DOE-BES, Materials Science and Engineering Division (ARL). The EDX Research was sponsored in part by the UK EPSRC through the UK National Facility for Aberration-Corrected STEM (SuperSTEM) (TJP, SJH).  

One challenge in the development of efficient organic photovoltaic devices (OPVs) is the optimization of the donor/acceptor(D/A) interface morphology, as this has tremendous impact on the electrical transport properties. In an optimized absorber layer morphology, excitons should be generated always in close proximity to the D/A interface to reach it within their lifetime, where theydissociate into free charge carriers [1-3]. In OPVs however the free charge carriers exhibit a small mobility which requires short pathways towards the electrodes to prevent recombination. Substrate temperature during deposition as well as surface chemistry are important process parameters [4-7]. In this respect alternating D/A layers are promising candidates for an optimized active layer morphology [8-10].

In this contribution we present results obtained from a stack assembly of three Zn-Phtalocyanine/C₆₀ layer (3 nm/3 nm nominal thickness) pairs (for further details of the preparation see [11]). The complete OPV device is shown in Fig. 1. The morphology of the individual D/A layers was investigated after each deposition step of ZnPc or C₆₀, respectively. For this, energy filtered electron transmission microscopy in a Zeiss LIBRA 200 FE operated at 80 kV was used. In order to be able to investigate the same area after each deposition step, a special finder TEM grid, coated with a graphene monolayer was used. Zero-loss filtered images were acquired with an energy filter slit width of 10 eV. In Fig. 2 we show images which where obtained after the first, second and third deposition, respectively (deposition at room temperature (RT)). The first image (Fig. 2a) shows that the ZnPc film is not closed, but islands have formed. The same can be observed for the successive C₆₀ layer (Fig. 2b). However, the latter agglomeration is stronger, thus larger clusters are formed, indicating a higher surface mobility of C₆₀. In the further course of the deposition steps (not shown) the contrast between image features of the first two layers decreases. This indicates that the successive deposition steps lead to closed films. Such a morphology is unfavorable as there are limited percolation paths from the different interfaces towards the respective electrodes. The morphology in the layer stack changes substantially when the substrate temperature is increased to 80 °C. Figure 3 shows the corresponding morphologies after the 1^st, 2^nd and 3^rd deposition step, respectively. In contrast to the experiment at RT agglomeration is observed in all following layers. In order to investigate the inter-layer links the morphology in layer n was visualized by calculating the difference image d_n = Image_n – Image_n-1. The result for the first six layers is also depicted in Fig. 3. We observe a high porosity and a well established crosslinking between the ZnPc on the one hand and the C₆₀ layers on the other. These morphology results in an improved efficiency of η=2.1 % compared to η=0.5 % for the device deposited at RT.

References:

[1] B. Johnev, M. Vogel, K. Fostiropoulos, B. Mertesacker, M. Rusu, M.-C. Lux-Steiner, A. Weidinger, Thin Solid Films, vol. 488, no. 1, pp. 270 – 273 (2005).

[2] S. Senthilarasu, S. Velumani, R. Sathyamoorthy, A. Subbarayan, J.A. Ascencio, G. Canizal, P.J. Sebastian, J.A. Chavez, and R. Perez, Appl. Phys. A, 77:383 – 389 (2003).

[3] B. P. Rand, D. Cheyns, K. Vasseur, N. C. Giebink, S. Mothy, Y. Yi, V. Coropceanu, D. Beljonne, J. Cornil, J. L. Bredas, J. Genoe, Adv. Func. Mater., 22:2987–2995 (2012).

[4] K. Fostiropoulos, W. Schindler, Phys. Status Solidi B246, 2840 – 2843 (2009).

[5] A. F. Bartelt, C. Strothkämper, W. Schindler, K. Fostiropoulos, R. Eichberger, Appl. Phys. Lett. 99, 143304 (2011).

[6] S. Pfuetzner, J. Meiss, A. Petrich, M. Riebe, K. Leo, Appl. Phys. Lett. 94, 253303 (2009).

[7] P. Peumans, A. Yakimov, and S.R. Forrest, J. Appl. Phys., 93(7), 3693–3723 (2003).

[8] J. Xue, B. P. Rand, S. Uchida, and S.R. Forrest, J. App. Phys., 98(12):124903 (2005).

[9] B. P. Rand, J. Xue, S. Uchida, and S. R. Forrest, J. Appl. Phys., 98, 124902 (2005).

[10] Z. R. Hong, B. Maenning, R. Lessmann, M. Pfeiffer, and K. Leo, Appl. Phys. Lett. 90, 203505 (2007).

[11] G. Chouliaras, W. Schindler, M. Wollgarten, M. Ch. Lux-Steiner, K. Fostiropoulos, submitted.

Mesoporous single crystals have been a matter of intense discussion in the last years in the context of solar energy harvesting, since it is expected that this material class can contribute significantly to the improved design of highly efficient solar energy conversion devices [1]. Especially for photocatalytic or photoelectrochemical hydrogen generation high surface area and good charge-transport properties are key features to enhanced device performance [2, 3]. Good conductivity is usually obtained in large defect free structures such as single crystalline materials, where consequently the surface area is small. High surface area, however, is obtained best by porous agglomerates of nanoparticles, where the conductivity is low because of multiple grain boundaries. The possibility to achieve performance improvement by combining both concepts has been demonstrated via the fabrication of large single crystals on the micrometer scale with a mesoporous structure using a template based approach [4]. In comparison to nanocrystalline materials, the improved charge carrier conductivity has been shown in this material class as well as its competitive surface area. Moreover, other template-free synthesis routes are known, where porous structures in solids are formed spontaneously. One example is the solid-gas phase reaction carried out for the synthesis of oxynitrides, i.e. thermal ammonolysis [5]. However, the control of pore size and density requires a detailed understanding of the reaction mechanisms during synthesis.  

Some of these oxynitride materials i.e. LaTiO₂N (LTON) or LaTaON₂ are photocatalytically active [6-8]. The main characterization techniques used to evaluate the pore quantity and quality have been powder techniques (for example physisorption), giving information about the open porosity, and qualitative scanning electron microscopy (SEM) and transmission electron microscopy (TEM), which suggested that open and closed pores are formed [6-8]. However, little is known about the size and shape distribution especially of the closed porosity or about the pore formation process.

In this contribution we will focus on microscopic pore characterization of LTON as a function of the synthesis method by combining several TEM techniques (Figure 1, 2) [9,10]. The pores themselves were explored mainly by electron tomography and by scanning TEM (STEM) with an high angle annular dark field (HAADF) detector, while the crystallinity was investigated using a combination of high resolution transmission electron microscopy (HREM), selected area diffraction (SAD) and nanobeam diffraction. The local chemical composition was studied by electron energy loss spectroscopy. With the improved understanding of the pore formation mechanism in LTON we enabled porosity tuning in large oxynitride single crystals leading to enhanced performance in photocatalytic and photoelectrochemical water-splitting.

[1]          C. Ducati, Nature 495 (2013) 180-181.

[2]          T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 43 (2014) 7520-7535.

[3]          A. Kay, I. Cesar and M. Grätzel, J. Am. Chem. Soc., 128 (2006) 15714-15721.

[4]          E.J.W. Crossland, N. Noel, V. Sivaram, T. Leijtens, J.A. Alexander-Webber, and H.J. Snaith, Nature,

              495 (2013) 215-219.

[5]          S.G. Ebbinghaus, H.-P. Abicht, R. Dronskowski, T. Müller, A. Reller and A. Weidenkaff,

              Prog. Solid State Chem., 37 (2009) 173-205.

[6]          A.E. Maegli, S. Pokrant, T. Hisatomi, M. Trottmann, K. Domen, and A. Weidenkaff, J. Phys. Chem. C,

              118 (2014) 16344-16351.

[7]          T. Takata, C. Pan and K. Domen, ChemElectroChem, 3 (2016) 31-37.

[8]          N.-Y. Park and Y.-I. Kim, J. Mater. Sci., 47 (2012) 5333-5340.

[9]          S. Pokrant, M.C. Cheynet, S. Irsen, A.E. Maegli and R. Erni, J. Phys. Chem. C, 118 (2014)

              20940-20947.

[10]        S. Pokrant, S. Dilger and S. Landmann, J. Mater. Sci., (2016) DOI: http://dx.doi.org/10.1557/jmr.2016.9

Acknowledgements: The authors thank the SNSF for the PrecoR grant 20PC21_155667.