Computational characterisation of metal oxide nanoparticles for risk assessment

• Autores: Laura Escorihuela Martí
• Directores de la Tesis: Alberto Fernández Sabater (dir. tes.) , Benjami Martorell Masip (codir. tes.)
• Lectura: En la Universitat Rovira i Virgili ( España ) en 2019
• Idioma: español
• Tribunal Calificador de la Tesis: Alberto Roldan Martínez (presid.) , Vladimir Baulin (secret.) , Alfonso Gallo Bueno (voc.)
• Enlaces
  • Tesis en acceso abierto en: TDX
• Resumen

  • Due to their intrinsic properties, nanomaterials (NMs) are the cornerstone of a wide range of technologically advanced applications, with metal oxide nanoparticles (MeO) NPs being the most used in areas such as electronics, optics, opto-electronics, pharmacy, medicine, cosmetics and textiles. However, there is still an important knowledge gap regarding how size influences their physicochemical properties as well as the risk to human health. Recent studies provided more insight on the size dependence of nanoparticle properties and reactivity, revealing that small sized nanoparticles (NPs) have a more variable behaviour in terms of their properties than larger size NPs, which have a more constant behaviour. Therefore, nanomaterials (NMs) need a specific regulation to assess their toxicity. The generation of in vitro and in vivo toxicity characterization data is essential for risk assessment and establishment of safe use of engineered nanomaterials (ENMs). However, this is a formidable task given the expected growth in number and diversity of ENMs.

    Nonetheless, toxicity assessment of NMs is a daunting task that involves multiple testing conditions and endpoints, and testing of different NP configurations. Computed based methods, in silico methods, based on theoretical and statistical domain, evaluate and determine and predict processes or even substance properties, this methodology is involved in very different disciplines given huge challenges. Apart from the legislation urgency for risk assessment exits a vacuum in literature, given that the data for the environmental risk assessment found in literature is uncertain and present knowledge gaps, though is not useful for the risk assessment for nanoparticles. This nanosafety data needs to be provided by standardised methods. One option to simplified this nanosafety data generation is the in silico methodology. In silico testing methods constitutes a cost-effective approach to fill the existing gaps in nanosafety data for being an effective tool for the safe-by-design of ENM. It exists an urgent necessity to develop nanosafety data for toxicity assessment in NPs, in particular for MeO NPs, and generate valuable QNAR (Quantitative Nano-Structure Activity Relationship) models for their risk assessment legislation using in silico methods to optimize time and resources to this purpose.

    The most popular in silico method based on quantum mechanics for chemistry is Density Functional Theory (DFT), which is postulated on approximations to the exact exchange–correlation functional (e.g. LDA, GGA, hybrid GGA, meta-GGA) that are relatively computationally efficient; it is competitive in accuracy for many interesting chemical phenomena, and it is computationally much less expensive than higher-level alternatives and useful for modelling chemical reaction pathways or for descriptor data in QSAR development. DFT is the base of the huge part of the methodology implemented in this thesis.

    In this thesis we performed a strict and deep study of the best methods to evaluate the band gap and the solubility of MeO NP from a computational point of view. The use of periodical-DFT methods has allowed us to optimise structures computing the electronic ground state energy for nanostructures as the nanotubes or spherical nanoparticles. To get more reliability for band gap determination, the exchange-correlation functional has been improved using the DFT+U methodology. This type of functional gives us a reasonable computational cost and accuracy in band gap calculations or geometry optimisation, without affecting the predictive capability on the influence of experimental environment. After that, to reach large systems up to 3000 atoms in order to simulate more realistic biological systems, it has been used the DFTB development for band gaps determination in large nanoparticles. Furthermore, the coupling of DFTB and molecular dynamics simulations has allowed the description of water-NP interaction, giving and extra value to this work.

    In a work previous to this thesis, determined that the descriptor EC (conduction band energy) in the range between -4.84 and -4.12 eV of standard redox potential, can lead to the generation of cellular oxidative stress between NPs and cells developing then toxicity.3,4,5 Furthermore, the ionic index of metal cations (Z) in MeO is correlated with their hydration energy, which is a measure of the affinity of the metal ion for water molecules. Therefore, how soluble is the MeO in the biology environment is also an important descriptor correlated with the toxicity found in other works. As a consequence, this is the first descriptor chosen for this thesis. Given these two important descriptor postulated a nano-SAR (Structure Activity Relationship) developed from the toxicity data of twenty two MeO NPs from 10 to 70 nm of size, where toxicity it was correlated with these two descriptors..

    The second descriptor chosen was the band gap between the HOMO and LUMO band energy for each size of NP, because it is totally linked in the Ec mentioned above. This is a property easy to measure via DFT or DFTB for small NPs, but it needs a large amount of computing resources for NPs with sizes greater than 2 nm. Therefore, in order to obtain band gaps for bigger NPs, the strategy followed was to create a prediction model of band gap MeO and compare the results obtained with the experimental values found in bibliography.

    The computational results obtained with the methodology developed in this thesis for the ZnO case have been promising and, in order to make more robust the method employed, it has been tested for TiO2 too, showing an excellent efficiency in results.

    Great part of the work has been spent in the solubility evaluation of the MeO NP, because there is a few quantity of bibliography about that (both experimentally and computationally), what indeed is reasonable because the MeO NP has a very low solubility and it is quite difficult its evaluation either experimentally or computationally. In my opinion, the results obtained with the implementation of the use of thermodynamical theory as the Ostwald–Freundlich approach in the Molecular dynamics framework are fixing a new starting point for the solubility evaluation of NP in the near future, and it is a proof for computational engagement of legislation authorities for toxicity risk assessment.

    Finally, the data obtained from the prediction models of band gap as well as the solubility models have been used to create nano-QSAR models, This is the reason why we used a nano-QSAR model published for these two MeO NPs and we added our computed descriptors in the work of Papa et al., in order to improve or equal the accuracy of the method. The data that may be used in these models is not enough to avoid statistical biases, but in this case it has been used Matlab software to implement statistical methods such as cross validation method, leave-on-out technique or the calculation of the root mean square error.

    Finally, I would like to highlight that the work of this thesis has been carried out with the managing of different techniques, software and disciplines. So, this is an additional proof that the perfect combination of different disciplines can produce a great work.