Francesco Buda

Boron-nitride-rich organic thin materials based on borazines have gained significant attention for their potential in nano(opto)electronic and energy storage devices. We address synthetic challenges in producing effective borazine-based thin films by proposing a dual theoretical and experimental protocol. This combines a multiscale computational approach, using density functional theory and classical molecular dynamics, with synthesis and thin-film formation via the Langmuir-Blodgett technique. The computational modeling focuses on three key properties: π-π stacking interactions, molecular steric hindrance, and dynamic self-assembly orientation. This modeling guided the selection of a borazine molecular building block and enabled the successful experimental formation of a free-standing molecular-thin borazine-based film. Solely π-π stacking interactions were found to drive the formation of a bilayer film with a molecular thickness of 2.1 nm, capable of spanning 0.6 μm diameter holes as a free-standing film. The agreement between theory and experiment confirms that the film retains essential features of the borazine molecular crystal, particularly intermolecular offset face-to-face π-π stacking and hexagonal-based pattern orientations. We thus establish a robust and transferable approach for modeling and synthesizing borazine-based thin materials, deepen the understanding of molecular interactions in borazine self-assembly, and demonstrate the suitability of the Langmuir-Blodgett technique for fabricating borazine-based 2D materials.

Bulk carbon materials can be synthesized through the modulation of precursor structures, stacking mode, and the optimization of pyrolysis conditions. The structure of two-dimensional (2D) carbon materials, however, is mainly driven by the pyrolysis temperature, pressure, and organic gas flow rates. Here, we report the synthesis of a centimeter-sized nanocrystalline graphene monolayer from pyrolyzing a Langmuir monolayer of a hexa(terpyridine)hexaphenylbenzene polyaromatic hydrocarbon (PAH) yielding a 2D carbon material with a thickness of 0.36 ± 0.07 nm. By designing the amphiphilic PAHs and reducing intermolecular π-π stacking interactions, a monolayer of horizontally aligned PAH was obtained at the water-air interface. During the pyrolysis, the monolayer served as a molecularly thin solid carbon source, leading to the controlled preparation of an electrically insulating nanocrystalline graphene monolayer. Our work proposes a 2D carbon material formed from the initial assembly of PAH monolayer at the water-air interface and subsequent pyrolysis and offers a modular strategy for designing nanocrystalline 2D films.

An ideal proton exchange membrane should only allow protons to pass and remain impermeable to fuels, as required in fuel cell applications. In methanol fuel cells, high proton conductivity enables high power density, whereas methanol crossover between the electrodes degrades the catalyst activity and lowers efficiency. In conventional polymer membranes, however, conductivity and selectivity are often antagonistic: long transport pathways are needed to achieve selectivity, but these introduce additional ionic resistance. Graphene, a two-dimensional material consisting of a single atomic layer of carbon, intrinsically addresses both requirements. Its basal plane is impermeable to water and other molecules, while still exhibiting a measurable degree of proton conductivity. Here, we show that chemical functionalization of monolayer graphene with sulfophenyl groups significantly enhances its proton transport properties. The conductance increases from 6.9 ± 1.1 to 30.9 ± 2.3 S·cm^-2, while the energy barrier for proton transport is reduced to 6.9 kJ·mol^-1. These findings propose functionalized graphene as an alternative to polymer membranes for electrochemical energy devices.

Glycosylation, the formation of glycosidic bonds, is a central yet challenging step in the chemical synthesis of complex carbohydrates due to its intricate regio- and stereochemical control. This study explores explicitly solvated, semi-empirical molecular dynamics (MD) simulations combined with multiple walker well-tempered metadynamics to investigate the mechanistic landscape of glycosylation involving a constrained glucose donor and a series of simple alcohol nucleophiles varying in nucleophilicity: ethanol, 2-monofluoroethanol, 2,2-difluoroethanol, and 2,2,2-trifluoroethanol. Our simulations reveal several mechanistic pathways depending on the nucleophile and substitution site. Stronger nucleophiles favor concerted SN2 displacement, while weaker nucleophiles increasingly promote dissociative SN1-like mechanisms and frontside attack pathways. This study demonstrates how semi-empirical MD simulations, combined with explicit solvation, can provide insights to understand the glycosylation reaction pathways.

Photoactivated chemotherapy (PACT) employs light to precisely control the activity of prodrugs, enabling spatial and temporal regulation of therapeutic effects while minimizing systemic toxicity. Transition-metal complexes, particularly Ru(II) polypyridyl compounds, have emerged as promising PACT agents due to their ability to undergo photodissociation via triplet excited states. However, rationalizing and predicting photosubstitution quantum yields remain challenging due to complex excited-state dynamics, solvent effects, and limitations of traditional modeling approaches. In this study, we synthesized nine Ru(II) complexes incorporating a monodentate thioether ligand, a terpyridine ligand, and various bidentate polypyridyl ligands. Upon red-light irradiation in aqueous solution, these complexes showed selective photosubstitution of the thioether by an aqua ligand. Despite similar absorption properties, these compounds exhibited markedly different photosubstitution quantum yields that static DFT calculations failed to explain. These modeling difficulties prompted us to develop a novel triplet-state molecular dynamics protocol using the GFN-xTB method, explicit solvation, and enhanced sampling techniques. Our approach enabled full simulations of the ligand exchange process on the triplet hypersurface and comparison to experimental data. It revealed that both cis and trans substitution of the thioether by a water molecule is possible; it was also able to distinguish between photoactive and photoinactive compounds. Unexpectedly, we found that the deviation from planarity of the bidentate ligand, rather than the energy levels of the different triplet excited states (³MLCT, ³MC) involved in the photosubstitution reaction, was the primary determinant of photosubstitution efficiency, as it promoted access to the trans photosubstitution pathway. Furthermore, our simulations uniquely identified whether dissociative, interchange, or associative mechanisms governed the reactivity of each complex. These results provide for the first time mechanistic insights into Ru(II)-based photosubstitution reactions in a solvent and offer a practical, scalable computational tool for designing next-generation PACT agents with optimized light-responsive properties.

The performance of dye-sensitized photoelectrochemical cells is presently limited by the photocathode component. Here, we investigate the impact of nuclear dynamics on the photoinduced charge separation of the benchmark NiO/P1 system (P1 = 4-(bis-4-(5-(2,2-dicyano-vinyl)-thiophene-2-yl)-phenyl-amino)-benzoic acid). Transient absorption (TA) studies in aqueous environments with different viscosities show that photoinduced hole injection either proceeds ultrafast (<100 fs) or in a sub-ps time window. We assign the fastest component to a surface species strongly coupled to the NiO. Interestingly, the slower injection component and charge recombination are slowed down considerably in more viscous media. Quantum-classical dynamics simulations of a system with the dye standing perpendicular to the surface yield an injection lifetime remarkably close to the slow component from kinetic modeling of the TA results. Simulations including nuclear thermal motion yield a 2-fold increase in hole transfer rate compared to simulations on fixed nuclei, highlighting the role of nuclear motion and providing new design principles for dye-sensitized photocathodes.

Molecular catalysts offer an ideal platform for conducting mechanistic studies of the hydrogen evolution reaction (HER) due to their electronic tunability. This study explores a series of anionic M=Co(III)- and M=Zn(II)-porphyrin complexes with electron-donating ([M(OMeP)] ^n-, [M(MeP)] ^n-) and electron-withdrawing ([M(F8P)] ^n-, [M(F16P)] ^n-) substituents. The activity of these complexes for the HER was analyzed in homogeneous photocatalytic conditions using [Ru(bpy)₃]²⁺ as a photosensitizer under blue light (450 nm) irradiation. The substituent-induced electronic effects were found to tightly control the activity and mechanism of the photocatalytic HER. As expected, the electron-rich [Co(OMeP)]^3- catalyst showed higher activity in acidic media (pH 4.1) with a maximum TOF of 7.2 ± 0.4 h^-1 and TON of 175 ± 5 after 39.5 h. DFT calculations were performed to investigate the HER mechanism. H₂ formation was found to initiate following proton-coupled reduction of a Co^III-H hydride intermediate in such conditions. More surprisingly, however, the electron-poor [Co(F16P)]^3- catalyst was more active at neutral pH (7.0), achieving a maximum TOF of 6.7 ± 0.3 h^-1 and TON of 70 ± 3 after 39.5 h. Instead of forming the Co^III-H hydride, an additional ligand-based reduction led to a ligand-protonated intermediate. This work demonstrates that electron-poor HER catalysts can outperform electron-rich catalysts near neutral pH conditions.

Fluorescence spectra of single terrylene molecules adsorbed on hexagonal boron nitride flakes were recorded at cryogenic temperatures. The pure electronic transitions of terrylene molecules are spread over a broad energy scale from 570 to 610 nm. Surprisingly, peaks in the vibrationally resolved fluorescence spectrum show intensity variations of ≤20-fold between molecules. We find an extreme case in which the Debye-Waller-Franck-Condon factor of the zero-phonon line exceeds 0.8. The vibronic intensity correlates with both the spectral position of the electronic transition and the frequency of the longitudinal stretch mode, which varies between 243 and 257 cm^-1. Using density functional theory calculations, we show that these observations can be explained by terrylene chemisorption on charge-donating defect sites. The electronic states of molecules at such chemisorption sites would be very attractive for the efficient emission of single photons with narrow lines and for the generation of indistinguishable photons.

Fluorescence spectra of single terrylene molecules adsorbed on hexagonal boron nitride (hBN) flakes were recorded at cryogenic temperatures. The pure electronic transitions of terrylene molecules are spread over a broad energy scale, from 570 to 610 nm. Surprisingly, peaks in the vibrationally resolved fluorescence spectrum show intensity variations up to 20-fold between molecules. We find an extreme case where the Debye-Waller-Franck-Condon factor of the zero-phonon line exceeds 0.8. The vibronic intensity correlates both with the spectral position of the electronic transition and with the frequency of the longitudinal stretch mode, which varies between 243 and 257 cm-1. By means of DFT calculations, we show that these observations can be explained by terrylene chemisorption on charge-donating defect sites. The electronic states of molecules at such chemisorption sites would be very attractive for the efficient emission of single photons with narrow lines and for the generation of indistinguishable photons.