Solid-liquid interactions in ionanofluids : experiments and molecular simulation

Tese de doutoramento, Química (Química Física), Universidade de Lisboa, Faculdade de Ciências, 2017One of the main areas of research in chemistry and chemical engineering involves the use of ionic liquids and nanomaterials as alternatives to many chemical products and chemical processes, as the latter are currently considered to be environmentally non-friendly. Their possible use as new heat transfer fluids and heat storage materials, which can obey to most principles of green chemistry or green processing, requires the experimental and theoretical study of the heat transfer mechanisms in complex fluids, like the ionanofluids. It was the purpose of this dissertation to study ionanofluids, which consist on the dispersion of nanomaterials in an ionic liquid. The first objective of this work was to measure thermophysical properties of ionic liquids and ionanofluids, namely thermal conductivity, viscosity, density and heat capacity in a temperature range between -10 e 150 ºC and at atmospherical pressure. In this sense, the thermophysical properties of a considerable set of ionic liquids and ionanofluids were measured, with particular emphasis on the thermal conductivity of the fluids. The ionic liquids studied were [C2mim][EtSO4], [C4mim][(CF3SO2)2N], [C2mim][N(CN)2], [C4mim][N(CN)2], [C4mpyr][N(CN)2], [C2mim][SCN], [C4mim][SCN], [C2mim][C(CN)3], [C4mim][C(CN)3], [P66614][N(CN)2], [P66614][Br] and their suspensions with 0.5% and 1% w/w of multi-walled carbon nanotubes (MWCNTs). The results obtained show that there is a substantial enhancement of the thermal conductivity of the base fluid due to the suspension of the nanomaterial, considering both mass fractions. However, the enhancement varies significantly when considering different base ionic liquids, with a range between 2 to 30%, with increasing temperature. This fact makes it more difficult to unify the obtained information in order to obtain a model that allows predicting the enhancement of the thermal conductivity. Current models used to calculate the thermal conductivity of nanofluids present values that are considerably underestimated when compared to the experimental ones, somewhat due to the considerations on the role of the solid-liquid interface on heat transport. Considering density, the impact from the addition of MWCNTs on the base fluid’s density is very low, ranging between 0.25% and 0.5% for 0.5% w/w and 1% w/w MWCNTs, respectively. This was fairly expected and is due to the considerable difference in density between both types of materials. However, viscosity was the property for which the highest values of enhancement were verified, ranging between 28 and 245% in both mass fractions of MWCNTs. The heat capacity was the only of the four properties mentioned above not to be studied in this work due to technical issues with the calorimeter to be used. Nevertheless, the amount of data collected on the remainder thermophysical properties was extensive. It is believed that the latter contributes meaningfully to a growing database of ionic liquids and ionanofluids’ properties, while providing insight on the variation of said properties obtained from the suspension of MWCNTs in ionic liquids. The second objective of this work consisted on the development of molecular interaction models between ionic liquids and highly conductive nanomaterials, such as carbon nanotubes and graphene sheets. These models were constructed based on quantum calculations of the interaction energy between the ions and a cluster, providing interaction potentials. Once these models were obtained, a second stage on this computational approach entailed to simulate, by Molecular Dynamics methods, the interface nanomaterial/ionic liquid, in order to understand the specific interparticle/molecular interactions and their contribution to the heat transfer. This would allow to study both structural properties, such as the ordering of the ionic fluid at the interface, and dynamic ones, such as residence times and diffusion. The first stage consisted on adjusting a site-site interaction potential model between each atom of the ionic liquid and graphene, in order to reproduce the calculated energy values using quantum chemistry methods. This step had an exploratory feature, since no similar work was available at the time in the literature. The choice of potential function that better described the energy values lead to a n-m potential, with n and m exponents different from the typically used for simple organic molecules (12-6). The transferability of the adjusted parameters to interaction energy values between ionic liquids and graphene was also evaluated. This was performed through the calculation of interaction energy values between an ionic liquid and a nanotube, without further adjustment. A more modest set of calculations were also performed to evaluate the charge distribution in graphene sheets and nanotubes, considering both zig-zag and armchair structures. As such, a model that correctly represents the interactions of several ionic liquids with carbon nanomaterials with different structures was obtained. These models were subsequently used in molecular dynamics simulations, considering two types of carbon nanomaterial (a singlewalled carbon nanotube, SWCNT, and a stack of graphene sheets) solvated by an ionic liquid ([C4mim][N(CN)2], [C4mim [SCN], [C4mim][C(CN)3] e [C4mim][tf2N]). The structural information obtained, i.e., the organization of the ions at the solid-liquid interface, demonstrated a greater proximity of the anion to the carbon surface. By studying two different SWCNTs (with different chirality indexes, namely (7,7) and (10,10), entailing distinct diameter values), it was possible to observe that the cations of ionic liquid are organized differently inside and outside of nanotube relatively to their position and orientation. In the widest nanotube, the alkyl chains are directed towards the center of the tube, creating a non-polar domain. Using this structural information, NEMD (non-equilibrium molecular dynamics) simulations were performed with the aim of calculating the thermal conductivity of each system at different temperatures. The results obtained agree reasonably with the experimental results, with a good prediction of the relative order between ionic liquids. The thermal conductivity was calculated in the composite systems considering different directions of the heat flow. In the case of the system composed by ionic liquids and a SWCNT, the thermal conductivity was calculated along the direction of the axis of the nanotube. The enhancement verified in this case is due to the conduction of heat trough the nanomaterial and is system-dependent, as seen in the experimental results. Regarding the systems composed by a stack of graphene sheets and ionic liquid, the thermal conductivity was calculated considering a heat flux perpendicular to the nanomaterial. As such, the global values of thermal conductivity of these composite systems were lower than the ones containing solely ionic liquid. However, this configuration allowed the study of the thermal conductivity of the interfacial region, where a significant enhancement was observed when compared to the bulk liquid in the same simulation box. In addition, no significant discontinuity was verified in the temperature profile at the solid-liquid interface, which represents a useful contribution for future predictive models for heat transport and properties. Lastly, the final segment of this dissertation consisted on using the information obtained both from experiment and from simulation. In this sense, the feasibility of the studied substances as heat transfer fluids was evaluated. This was attained through the calculation of heat transfer areas (the main design parameter in heat exchangers) considering a specific process and using the values of thermophysical properties obtained. The results were compared with heat transfer areas calculated for currently used heat transfer fluids. It was verified that some of the studied ionanofluids could compete with the commercial heat transfer fluids in terms of area necessary to transfer the same amount of heat. This fact supports further research on considering ionanofluids as heat transfer fluids