Conclusions

The ensemble of conformations generated by REMD simulations of the solvated RGD and SPT tripeptides described with the CHARMM force-field with both the TIPS3P and SPC/Fw water potentials has been compared. The effect of the water model on the con- formational equilibrium of these peptides has been analysed using Ramachandran plots, clustering analyses, RDF profiles, hydrogen-bonding analysis and side-chain dihedral angle analysis. The equilibrium distribution of population over the Ramachandran plots compares favourably between the two water models for both the peptides. The popu- lations of all regions of the Ramachandran plots were within error, apart from one in

the RGD case, regiongand even then it followed the same general trend with respect

constant which, for SPC/Fw, was 80 % the size of the value for TIPS3P. For both the backbone and all-atom peptide clustering analyses, sampling of conformational phase space is seen to be very similar. Populations of clusters are within error for the SPT case and follow the same general trend for the RGD case. RDFs between key atoms of the peptide and the oxygen and hydrogen water atoms showed that differences in water structuring around several groups in the tripeptide were also found to be negligi- ble. There was a slight difference in the rate of conformational sampling. Furthermore, there were generally more transitions between the regions of the Ramachandran plot in the TIPS3P case compared to the SPC/Fw case. This was, however, entirely consistent with the 50 % lower viscosity of the TIPS3P model compared to SPC/Fw. Overall, we found that the equilibrium conformations sampled for each force field combination to be sufficiently similar which suggest that CHARMM and SPC/Fw can be combined with a reasonable level of reliability, at least for our two exemplar tripeptides. As with all force fields, regardless of their maturity, further testing will be required to establish a greater measure confidence.

In spite of the suggested compatability of the two force-fields, CHARMM and SPC/Fw, in the validation study of this chapter, it has not been used in the work of sub- sequent chapters. TIPS3P is computationally more efficient and has been used reliably with the CHARMM force-field in many previous studies. Furthmore, the difference in dielectric constant between SPC/Fw and TIPS3P could have a signifcant impact on ion-surface interactions and lead to inaccurate results concerning analysis, such as ion density distributions, which - given the importance of electric double layer (EDL) ef- fects, particularly the change in Debye length with concentration, to many hypotheses of the ’low-salinity’ effect - is particularly important. Furthermore, the silica force-field used in Chapters 4-6 was parameterised to be used with common biomolecule force- fields, including CHARMM and AMBER.

Chapter 4

Water and Ion Structure at the

Aqueous Electrolyte/Amorphous Silica

Interface

Silica is one of the most abundant minerals on Earth and the interface it makes with aqueous electrolyte, for example seawater or physiological solution, is ubiquitous in nature and technology. Atomistic-level insight into the structure of ions and water at the interface is of great importance to a range of commercial applications. For instance, the impact of changes to the bulk concentration of the solution would provide valuable

insight into the mechanism of enhanced oil recovery (EOR) from sandstone reservoirs8;

amorphous silica is often used as a model for the quartz grains in the sandstone pores. In EOR, as discussed in section 1.1.1, the use of low-salinity seawater in the water flooding of oil reservoirs increases oil yields compared to when normally concentrated seawater is used. Interfacial solvent structuring has previously been described as ’gov- erning’ the adsorption of organic molecules at mineral/water interfaces, an example of

which is oil at the silica/water interface.22 Furthermore, many previously developed

hypotheses used to explain the ’low salinity’ effect and EOR, such as fines migration theory, are centered around changes to the electric double layer (EDL), a description of the distribution of ions in the region close to the surface (more details of the hypotheses and a definition of the EDL are given in section 1.1.1) The concentration dependence of interfacial ion and water structure, if there is such a dependence, is therefore of great

interest. In this Chapter, we have taken a reductionist approach, using MD simulation to investigate the interface between amorphous silica and pure saline solutions of some

of the electrolyte types present in seawater: NaCl, KCl, CaCl2and MgCl2, at both high

and low salinity. This is in contrast to previous experimental studies designed to eluci- date the mechansim of enhanced oil recovery, where the full mixture of ions present in

seawater was used. Adsorption of the NH3+ functionality, one of the basic functional

groups in crude oil, at the interfaces of the same eight solutions with amorphous silica is considered in Chapter 5. While the proportion of nitrogen-containing compounds in crude oil, particularly aliphatic amines, is small, a significant change in the adhesion

force of the R-NH+₃-functionalised AFM tip to the aqueous electrolyte/amorphous sil-

ica interface on changing electrolyte concentration is observed experimentally in this work (see Chapter 5) and the presence of polar functionalities is a known requirement

for the ’low-salinity’ effect.12, 13, 8 Furthermore, the interaction of the functionality with

the aqueous electrolyte/amorphous silica interface is thus far underexplored.

The model amorphous silica surface used features both protonated and deproto- nated silanols, as depicted in Figure 2.8. In this Chapter, a number of new terminologies are introduced, the definitions of which can be found in section 2.1.3.

Appendix C contains: details of the number of water molecules used in each simulation, further lateral ion and water density profiles, further angular distributions of

the angle made by the O−-M+ vector and the surface normal, a summary of distances

from the surface of peaks in the ion density profiles, net water orientation as a function

of distance from the surface and full cos(θ) distributions.