In Silico Fluorescence of Molecules and Nanoparticles.


A significant part of the research activities developed at SNS@CECAM  is aimed at studying photophysical and excited state properties of fluorophores embedded in different environments, from aqueous solution to silica nanoparticles and biological systems such as cellular membranes and proteins.

Organic fluorescent dyes come in a multitude of absorption and emission wavelengths; the high bio-stability, biocompatibility and high brightness reached when incorporated in silica nanocrystals, make them ideal candidates for applications in medical biosensing, in particular as fluorescent tumoral markers.

Nanoscale systems present a challenge for computation, since their properties cannot be modelled in the most effective way by methods developed for bulk simulations, and, at the same time, their dimensions are too large for the standard methods developed for medium size molecular systems.

In this framework, researchers at the SNS@CECAM node are continually developing new integrated computational multilevel approaches combining classical molecular dynamics (MD) simulations and mixed quantum-mechanical/molecular mechanical (QM/MM) methods including last generation mean field models (e.g. the so called polarizable continuum model, PCM) which are fundamental ingredients to accurately reproduce the physical-chemical properties  of the fluorophore molecules embedded in silica nanoparticles as well as their absorption and emission spectra.

Descrizione: fig4.png


Among the organic dyes currently employed in nano-medicine, the ground and excited properties of perylenes, coumarins, bodipy derivatives, rodhamines, fluorescein, cyanines and conjugated polymers have been investigated at SNS@CECAM. A schematic representation of a rhodamine fluorophore adsorbed onto a silica surface is reported in the Figure.


DNA-based Biosensors

Another important area of research at SNS@CECAM is related to the development of models for realistic simulation of interactions between organic molecules (e.g. protein or DNA building blocks) and semiconductor surfaces which are receiving increasing attention due to their application in several disciplines of fundamental science and engineering.  Biochips or microarrays are based on the effects of the hybridization between single-strand DNA fragments (probes), immobilized on the surface of an electrode through suitable functionalization procedures, and complementary or partially complementary oligonucleotide sequences present in solution (targets). The behaviour of the probes is influenced by both the type of immobilization on the substrate and the surface characteristics leading to highly efficient and specific hybridization reactions. Understanding in detail the various steps of the hybridization mechanisms is extremely important when designing or improving DNA biosensors.


The characteristics of the supports and their functionalization, coverage, and packing of the bonded molecules, sequence, length, and concentration of the attached DNA strands, cross-linker properties, and other aspects related to DNA-detection techniques are investigated theoretically. Computational approaches applied at SCS@CECAM cover both high- and low-resolution models (classical all-atom simulations, coarse-grained descriptions and quantum calculations) which are used in a synergic way to describe the behavior of DNA strands when linked to the substrate, free in solution and complexed with dye molecules. Such an integrated approach allows to benefit from the accurate QM computations applied to develop tailored force-field models, which are subsequently adapted for molecular dynamics simulations. The results of such computational investigations provided already insight into the concerted processes of melting, hybridization and DNA/dye complex formation mechanisms and identified various metastable conformations adopted by the oligonucleotides during the various phases processes, in agreement with experimental data.




New developments in SCC-DFTB method

The Tight Binding method stands as a very effective and well tested approach allowing  for the treatment of electronic properties of large molecular systems.  The Tight-Binding method under development at SNS@CECAM is based on the DFT energy expression in the nearby of the equilibrium configuration (so the shortcut DFTB). The DFTB approaches allow to combine (i) feasibility of computations due to crucial computational costs reduction, maintaining at the same time, in contrast to MM approaches, the treatment of the systems at a (ii) quantum-mechanical level.

The first point (i) is made possible by proper sets of parameters employed for the computation of the various contributions of the total energy, permitting to avoid the computation of mono- and bi- electronic integrals, which are the most expensive part of QM calculations. Regarding the latter point (ii), a key issue concerns the possibility to compute self consistently the atomic charges at a QM level, which can be used for calculating all the most important spectroscopic properties of the molecular systems. Our recent implementation of time dependent DFTB (TD-DFTB) is a very powerful tool for calculating electronic vertical absorption energies with the target to calculate excited state geometries of large systems under development. Further developments concern the interface between (TD-)DFTB and PCM protocols, in order to evaluate solvation effects and to perform simulations of molecules in solutions.

Our recent implementation of the analytic fitting of DFTB parameters allows to efficiently compute analytic gradients and hessians, with significant savings of computational time.

Due to the spreading importance of silicon in technological applications, a particular attention has been devoted in to the DFTB treatment of silicon-based nano-systems. In particular, one of the goals aim to establish a set of parameters which are suited for calculation of spectroscopic properties of silicon nano-particles. An improved parameterization of the DFTB energy expression at equilibrium geometry and the fitting of the off-equlibrium energy expression are under development at the moment, in order to compute accurate vertical excitation energies and vibrational properties, respectively. The image below shows 1-amino-3-cyclopentene adsorbed onto Si(100) surface modelled by cluster models of increasing dimensions (Si33H32, Si82H60, Si158H96), as an example of the systems which have been studied with DFTB methodology.



In silico studies of pigments: stuctural and electronic characterization of alizarin metal chelates


Stability and reactivity of ancient dyes and pigments are of central interest for the analysis and preservation of art objects. An integrated experimental-computational approach stems as a very promising way to better understanding of modern analytical techniques applied in the field of cultural heritage. Concerning specific projects at SNS@CECAM first study on alizarin metal complexes has been set in order to gain additional insights on optical properties of this well-known madder-lake pigment. The alizarin-based pigments were widely used until last century in paint and textile tincture. Today these pigments are completely substituted by other more stable synthetic molecules.

However, the study of alizarin based systems, due to relatively large experimental data available stands as a suitable testing to establish computational approaches for the analysis of ancient pigments. In this respect it has been possible to describe electronic and structural characteristics of free alizarin for which experimental data are available (UV-VIS, vibrational calculus) and then extend such studies toward its various metal chelates in different solvation, coordination and environmental conditions. In fact such effects had shown significant influence of the adsorption band of the complex changing its electronic characteristics and finally the madder lake color. The solid state powder pigment require also suitable sites allowing the interaction of the complex with a polymer structure, as a cellulose fiber, probably the most employed support for both paints and textile artworks, which also need to be considered in the model. In general, such preliminary studies confirmed the ability of ‘in silico’ approach to provide an in depth comprehension of the nature and composition of original art-work materials allowing to understand the physical and chemical changes that occurred over the years affecting both material composition and their chromatic properties, with the description of the factors responsible for the colour modifications. In this view the application of computational approaches permit to limit invasive interventions on the unique-sample objects of cultural heritage interest, allowing a detailed knowledge of the structural and electronic properties with the description of degradation mechanisms. Such studies can be also seen as the main door for a more rational employ meant of new dyes as well as suitable tools for powerful, stable and durable restore works on the historical art objects.

Description of systems in solution: explicit and continuum solvent models.

Concerning the description of molecular systems in solution, development effort at SNS@CECAM is related to combine various computational schemes to create user-defined and/or problem-tailored approaches. This is particularly straightforward exploiting theoretical models for solute solvent systems using hybrid methods with non-periodic boundary conditions and localized basis sets, which are at the same time more reliable from a physical point of view and computationally very effective. Additional extension of such schemes toward discrete/continuum models allows conveniently reduce the number of degrees of freedom, while keeping all the important interactions with the bulk, modeled as a continuum. In this respect, the general liquid optimized boundary (GLOB) model developed by the SNS@CECAM team can be successfully applied to perform molecular dynamics simulations of complex molecular systems in solution. Within integrated explicit/implicit solvent scheme specific intermolecular interactions between the solute and the solvent (e.g., hydrogen bonds) can be also retained, especially if they play a crucial role in determining the solute structural, dynamic, or spectroscopic properties, by including solvent molecules in the explicit treatment. Further, such conditions avoid the appearance of possible correlation effects and other problems with charged systems  that may affect molecular calculations and simulations using periodic boundary conditions (PBCs). Such combined explicit/implicit solvent models allow to accurately simulate properties of molecular systems in solution, including large-amplitude motions and solvent librations.


Computational spectroscopy: time-dependent and time-independent routes


The accuracy of a simulated spectrum depends a proper choice of a computational model: a reliable description of equilibrium structures, vibrational properties, and electronic structure is necessary. In the case of macromolecular systems this task is not trivial and in this respect, integrated computational approaches set within both time-independent and time-dependent frameworks are well suited for systems where the most important spectroscopic features have a local character.  For the former approach several computational tools, covering a large panel of spectroscopies, in particular those of vibrational and electronic origin, have been developed and coded. Particular attention has been put in order to establish computational tools allowing also use by a non-expert, in view of a broader adoption of advanced theoretical models. In this context general-purpose modules for the simulation of the line-shapes for vibrational spectroscopy (e.g. infrared) at the anharmonic level and vibrationally-resolved electronic spectroscopy (one-photon, electronic circular dichroism) allowing direct vis-à-vis comparison with experimental outcomes have been developed and validated. Additionally, eigenstate-free time-dependent methods which are the main (when not the only) route to deal with systems affected by significant anharmonic or non-diabatic interactions are developed at SNS@CECAM. Such models relay on the description of dynamic effects through classical MD approaches where dynamic simulations allow sampling the general features of the configurational space with one or more trajectories. Then, spectroscopic observables may be computed on the fly or in a second step by averaging over the corresponding estimators and suitable number of snapshots.  The a posteriori calculation of spectroscopic properties, compared to other on-the-fly approaches, allows exploiting different computational schemes for the MD simulations and the calculation of physical chemical properties. In this way, a more accurate treatment for the more demanding molecular parameters, of both first [e.g., hyperfine coupling constants (hcc’s)] and second (e.g., electronic g-tensor shifts) order, could be achieved independently of structural sampling methods provided the accuracy in reproducing reliable structures and statistics is proven for the latter.

Those approaches represent a great improvement with respect to the methods still commonly used to evaluate spectroscopic properties of larger molecular systems which the neglect influence of dynamical effects on the spectra line-shapes (harmonic approximation, vertical electronic transitions) and contribute to a better understanding of experimental spectra routinely studied nowadays.