"Nano" is different !
Nanogenomics: a new way to think nanocatalysis Mind the isomer!
Recently we show how it is possible to classify univocally metallic nanoparticles on the basis of the occurrence of their atop and bridge sites, introducing a geometrical genome, defined upon the generalised coordination number. We like the idea that a nano-genomics --in parallel to bio-genomics-- is possible (PCCP Adv. paper). We apply it to study how the catalytic activity of morphologically diverse samples can be predicted and varied. Combining classical MD and linear scaling relationship we predict on-the-fly the mass activity of supported Pt-nanoparticles for ORR, as recently discussed in Rossi et al. ChemPhysChem 2019) Mind the sample! Thanks to the NanoCHE approach, a modify version of the CHE model to study electrochemical reactions, we show how we can move to design nanosamples and not only individual nanoparticles. Again the classical motion of atoms affect the catalytic activity, don't forget if small they move fast! The full story is on ACS Cat. 2020 Born to be different! We show that the formation process can lead to significant geometrical differences that affect the catalytic. There is hope that the design of the best MNPs can be done choosing how to produce them! Thanks Elena Gazzarrini (MSci 2020) for her exciting results! See the whole story on a recent Nanoscale paper. We show that the activity of Cu nanoparticles for the CO2 conversion in methane can be 'predicted' by the formation process (in the gas phase). |
|
Machine Learned force field
|
As classical MD is becoming a very useful if not mandatory tool to understand physics-chemical properties of metallic nanoparticles, we need to have fast but accurate potentials. A way is to learn from available DFT-data and using ML techniques to train better force field. The MFF package is a way to do it. See our mini review on the field, Zeni et al. Adv. Phys. X 2019.
Employing Gaussian regression, we develop a novel machine-learned interparticle potential (MLIP) for Au nanoparticles. We map it onto a classical potential to be used on LAMMPS. The MLIP is cross-size (from few hundreds to thousands of atoms and to bulk), able to reproduced the complex solid melting transition. We also propose an automatic and robust classification to discern the core and surface of Au-NPs using a unsupervised learning scheme (k-means). An excellent agreement with experimental data is shown, Zeni et al. Nat. Comms (2021) |
|
Chemo-physical propertiesAt the nanoscale, the break of translations symmetry might cause that standard methods in condensed matter cannot straightforwardly applied. Even small morphological changes, distortions of the geometrical shape and relative positions of the chemical species alter the nanoparticle/nanoalloy properties.
Catalytic properties By means of density functional theory and the development of classical tools, we study the importance of the chemical ordering at the interface oxide/metallic nanoparticle for green chemical reactions, such as oxygen reduction (Asara, et al. ACS Catalysis 2016) coming from a careful classification of supported sub-nano sized nanoparticles (Paz-Borbon, Inorganics 2017). A very exciting result -combining classical MD and nanoCHE model - suggests that the formation process might affect the nanoparticle morphology and hence its catalytic properties, see the case of Cu nanoparticles for the CO2 to methane conversion, Gazzarrini Nanoscale 2021. Magnetic properties We study how geometrical distortions in the second shell of neighbours, which are strongly shape dependent, induce a peculiar charge transfer and then the appearance of a magnetic state (DiPaola, et al. Nanoletters 2016). We demonstrate how Pt13 could loose, reduce or maintain their magnetic properties after embedding into a zeolite environment (DiPaola et al. Nanoscale 2017). We investigate the magnetic behaviour of Al NP, doped with Ni and Pt. By BOMD, we discover three new architectures at 19 atoms, characterised by original piling of pentagons. Each isomer has a peculiar PDOS. The full story on JMCC 2020. Optical properties The HOMO-LUMO gap of Au-alloyed clusters depends on the presence of a cavity (Baletto et al. PCCP 2015).. We show that breathing modes of the fullerene cages of Ag, Cu, and Au can affect strongly their optical spectra. Ag is the chemical specie with the strongest plasmonic character (Zhao et al. JCP 2022). We study the effect of doping of Pt and Rh. Just a dopant of Pt affects strongly the optical spectra. Most significant, the relative position of Pt changes the absorption spectrum, Jones, EPJAP 2022. A new evidence that Au and noble gases (Ng) can make chemical bonds with a covalent nature. The very small (4 to 6 atoms) Au clusters binds up to 6 atoms of Ng, for the full story see Ferrari et al. JCP, 2022. |
|
Shape characterisation, isomer fluctuation(s) and dynamicsNano-objects move fast along their free energy space, and change their configuration. As their physical properties depends on their shape, elucidating how they transform from one minimum to another and the typical time scale is of paramount importance.
Working in close collaboration with Lievens' and Janssens' group in Leuven, we see the effect of Pd doping on Au-clusters, Ferrari et al. (Nanoscale. Adv 2021). We further discover a super stable isomer for the cationic AgAu with 14 atoms, as it appears in experiments! See the full work on JCP 2020. The reader interested in benchmarking DFT and experiments could find our work on cationic metal-Ar complexes of interest, Delgado-Callico et al. (Theo. Chem. Acc., 2021). A challenge task is of course a good characterisation of the each minimum and of the pathway connecting two minima. We show a Metadynamics at low temperatures to explore the potential energy surface of small Pt clusters and automatically get whether minima are connected (Pavan et al. EPJD 2013, selected by Springer for an outreach paper on "New taxonomy of platinum nanoparticles"). We apply Metadynamics to characterise the shape dynamics in mono- and bi-metallic nanosystem (Pavan et al. JCP 2015). A comparison with iterative molecular dynamics and DEPTS confirm metadynamics result (Gould et al., JPCL 2016). The influence of mismatch has been investigated in our latest publication appeared on PCCP 2017 and EPJB 2017. The video is also available from Francesca's latest interview for the Science and Engineering South Consortium (SES, ses.ac.uk). If you want to know more about on how nanogeometry affects catalysis, please see my recent lecture at the CATSENSE school on Nanocatalysis. If you are interested in the characterisation and classification of MNPs, have a look at this topical review on small nanoparticles, JPCM 2019. We also identify very peculiar transition pathways -without any melting or disorder phase- that brings decahedra into FCC shapes -of course with stacking faults. Here two videos for Au nanoparticles of about 900 atoms. Simulations obtained using metadynamics at 300K, collective variables as stacking fault number and coordination umber. SMATB potentials for Au-Au interaction. Details of the transformations are available in JPCM 2019. A new numerical tool, named Sapphire, is freely available to help the community getting their analysis done. The code will be presented at the incoming Faraday Discussion on Nanoalloys. Thanks Robert for his massive coding! |
|
Formation processes
From Francesca'a early works, modelling the growth process is a quite attracting features. A one-by-one model (Baletto et al., PRL 2000) has found to be extremely peowerful in predicting new shape and chemical odering. Some recent accomplishments regards the quasi-Janus (Parsina, et al. JPCC 2010) and core-shell ordering change depending on the Co-concentration and the CoPt (Parsina et al. Nanoscale 2012).
Thanks to the work by Matteo Tiberi (MSci 2020, now PhD at Cambridge) our LoDiS gains a new tool to mimic the coalescence in the gas phase and constrained between plates. |
|
Nanothermodynamics
Nanothermodynamics is fascinating!
Although many descriptors have been proposed, a measurable method to define the temperature at which a metallic nanoparticle melts was not available yet. We find an universal signature (at least for monometallic nanoparticles) to easily detect when the transition occurs. See the full story on Nanoscale 2021B. Such signature is based on the position of the second peak of the pair-distance distribution function. It can be measured and easy to be calculated. It is more robust than other methods and also an easy tool to discern objects with/without a geometrical order. Nanoalloys at high temperatures The nucleation and the thermal behaviour itself is one of the open question at the nanoscale and not fully understood yet. We show what is the size effect on the melting temperature for CuPt and we extrapolate a critical size after which a classical approach is faster (Pavan et al. PCCP 2015). In collaboration with Csanjy's group (Cambridge) and Partay (Reading), we are demonstrating the closure of the hysteresis loop in CuPt nano alloys. |
|
Support, embedding and 'real' environment
Pt nanoparticles in zeolite |
Pt onto MgO(100)
|
Anionic water system
|
Naive structural properties of embedded Pt clusters, just accepted to appear on Nanoscale (2017)
|
Evolution of supported icosahedral Pt nanoparticles during a melting transition. See our recent publication on JPCM 2017.
|
Charged water systems play a fundamental role in our life. We show that the extra electron charge may influence the shape of water clusters. We recently published a chapter on the book in the memory of Roy Johnston.
|
