Reports: ND1054135-ND10: Understanding Nanoscale Calcium Silicate Hydrate (C-S-H) Interfaces in Cement and Interaction Mechanisms with Polymer Additives

Hendrik Heinz, PhD, University of Colorado-Boulder

We employed molecular simulations to study the adsorption of various polymers on low energy surfaces of tobermorite 14 Å, including (001), (004), and (100) facets, to understand surface interactions and strength of adhesion. Tobermorite 14 Å is a key model phase for hydrated Portlandite cement, and the polymers chosen in this study include widely used polycarboxylate ethers. The main chain of the polymers consists of polyacrylate and the side chains of polyethylene oxide (PEO) whereby the degree of esterification and the length of the PEO side chains is varied. Results for (001) facets show that the attraction per individual acrylate group is on the order 0 to -2 kcal/mol, consistent with calorimetry measurements by experimental collaborators that indicate 2-3 RT= -1.2 to -1.8 kcal/mol acrylate adsorption energy for the best binding polymers. The adsorption strength increases for longer PEO chains as well as for larger degree of esterification with PEO. The differences are significant and data for other facets are in progress. Binding mechanisms involve pairing of calcium ions on the tobermorite surface with carboxylate side groups in the polymer chain (all groups are deprotonated at pH ~12), as well as hydrogen bonds between silanol groups on the (100) surface and oxygen atoms in acrylate. Depletion forces could also play a role to enhance adsorption of some of the polymer chain to the surface in order not to disrupt the hydrogen bonded network in liquid water. The understanding of mechanisms of attraction as a function of polymer composition and morphology is helpful to establish guidance in designing better formulations and to control materials properties. A manuscript explaining the mechanisms and adsorption data, including facet-specific differences, is in preparation.

Steered molecular dynamics calculations are also in process to determine free energy contributions, energy contributions, and entropy contributions to adsorption, as well as resistance to lateral movement across the surface for the same systems. This work in progress may form the basis of another manuscript.

Upon completion of the understanding of adhesion properties of the polymers we plan to quantify agglomeration energies between the organically modified surfaces. It has been shown that reduction in agglomeration properties determines the extent of mobility and, therefore likely fluidity, of hydrated cement particles. This hypothesis is based on recent work that demonstrated organic molecules (grinding aids) to function as spacers between cement particles that reduce the attraction between two nearby surfaces, represented by a lower agglomeration energy (J. Phys. Chem. C 2013, 117, 10417). The efficiency in reducing the agglomeration energy for a comparable amount of organic additive employed was shown to correlate with energy savings in a ball mill. A lower agglomeration energy might, in an analogous way, have an impact on strength development in cement/concrete and potentially lower the necessary water content to process cement.

A first publication acknowledging support from ACS-PRF developes similar arguments into a multi-scale modeling scheme for grinding cement clinker , entitled “En Route to a Multimodel Schemefor Clinker Comminution with Chemical Grinding Aids” in Adv. App. Ceram. 2015, DOI: http://dx.doi.org/10.1179/1743676115Y.0000000023(published online).

The involvement of aqueous interfaces in cement also requires reliable parameters for inorganic ions, including alkali, earth alkali, and halide ions. Efforts towards force field refinements for these ions are under way to expand the INTERFACE force field, as well as any force fields of similar energy expression (e.g. CHARMM, PCFF, OPLS-AA, DREIDING). The new force field parameters allow the reproduction of hydration energies in better than 5% agreement with experiment. At the same time, the Lennard-Jones diameters correlate well with van-der-Waals radii observed in experiment, and the computed coordination numbers with water are close to experimentally observed values. We also successfully tested that crystal structures of several metal halides could be reproduced using exactly the same parameters without adjustments (typically <3% deviation in computed density). Trends in epsilon values in the Lennard-Jones potential also correlate with atomic/ionic polarizabilities. The new set of parameters for common ions is thermodynamically better supported than previous parameters sets and could have long term impact on computations of cement and mineral interfaces, as well as for aqueous solutions of various ionic strength with biomolecules. A manuscript is also in preparation.

The PhD student supported in this project, Mr. Tariq Jamil, received an NSF fellowship to attend the "23rd World Forum on Advanced Materials (PolyChar 23)" in Lincoln, NE, May 11-15, 2015, and received a diploma of distinction (best poster award) in the poster session of the same conference. Work related to this project was also presented at several National Meetings (ACS, MRS), and seminar talks.