Jeffrey J. Thomas

Shrinkage can be critical for the strength and durability of drying cement pastes. Shrinkage becomes particularly severe at very low relative humidity, < 20%, which can be met in some activities involving extreme temperatures. Experiments and simulations suggest that small pores in the cement paste, with approximate thickness ≤ 1 nm, stay saturated unless the humidity drops below 20%. Here we suggest that this pore size can define two different categories of pores in the paste: pores thicker than 1 nm, where the Kelvin’s equation and the corresponding capillary (Laplace) pressure apply, and pores thinner than 1 nm, which can be considered as part of the solid skeleton if the humidity stays above 20%. We show that a continuum model, incorporating a pore-blocking mechanism for desorption and equilibrium thermodynamics for adsorption, explains well the sorption hysteresis for a paste that remains above ∼ 20%. At lower humidities, we assume that (1) during adsorpion water re-enters the smallest pores throughout the entire RH range (supported by experiments and simulations) and (2) there exists a simple linear relationship between water and strain in the smallest pores. These minimal assumptions are sufficient to explain the low-humidity hysteresis of water content and strain, but the underlying mechanistic explanation is still an open question. Combining the low-humidity and high-humidity models allows capturing the entire drying and rewetting hysteresis, and provides parameters to predict the corresponding dimensional changes.

The kinetics of hydration of magnesium oxide (MgO) powder to form magnesium hydroxide (Mg(OH) ₂ ) were measured using isothermal calorimetry at different temperatures, and the morphology of the powders before and after hydration were examined. The hydration kinetics of light‐burned MgO exhibit a hydration rate peak similar to that of portland cement hydration, whereas the hydration kinetics of hard‐burned MgO are comparatively slower at the same temperature, and exhibit a continuously declining hydration rate after the first several minutes of reaction. The hydration kinetics of both light‐burned and hard‐burned MgO can be fit using a boundary nucleation and growth model that has previously been applied to the hydration of portland cement and tricalcium silicate. Activation energy values for MgO hydration were determined from the fitted rate constants and were also measured directly using small temperature excursions according to a recently proposed method. For light‐burned MgO the resulting values are in good agreement and indicate a value of 77 kJ/mol. For the hard‐burned MgO the activation energy values vary considerably depending on temperature and how the activation energy is measured, but are always lower than the value obtained for the light‐burned MgO.

An apparatus has been designed to study the volumetric changes under controlled curing conditions. The apparatus is capable of measuring the chemical shrinkage as well as the macroscopic volumetric changes of sealed and saturated specimens under controlled pressure and temperature. Specifically, when measuring separately the macroscopic deformation (bulk shrinkage) and the water absorption of a saturated specimen, we find that the sum of the bulk deformations and water absorption adds up to the chemical shrinkage. Furthermore, while we observe pore pressure drops in the hydrating cement paste, even in the presence of an infinite supply of pressurized water (i.e. chemical shrinkage test conditions), bulk shrinkage occurs even in the absence of negative liquid pressure and at (almost) zero effective stress. That is, at (very) early ages, driving forces other than capillary forces are at work to cause bulk shrinkage.

A new calorimetry‐based method for calculating the activation energy of cement hydration is described. The method requires a scanning calorimeter that can change the sample temperature relatively quickly and has a stable baseline, but the technique is straightforward to perform. Activation energy values can be calculated at multiple hydration times from a single specimen, allowing any changes with time to be assessed. With this method, the measured activation energy at a given time does not depend on the previous microstructural development, as is the case with traditional methods of calculating the activation energy from parallel specimens hydrated isothermally at different temperatures. For tricalcium silicate ( C ₃ S ) hydrated in pure water or in a calcium chloride solution, the activation energy remains statistically constant from the earliest times until the degree of hydration exceeds 0.65, with an overall average value of (51.1 ± 1.8) kJ/mol. For cement hydration, the activation energy is also constant until well past the main hydration peak, with slightly lower average values. These results suggest that the rate‐controlling step in the hydration process does not change during the early hydration period (the first few days after mixing), as different reaction steps tend to have different activation energies.

With the search for hydrocarbons moving to more extreme environments, includingdeepwater, one of the challenges associated with cementing is ensuring thelong-term integrity and mechanical properties of the cement at hightemperatures (HT). To avoid strength retrogression at temperatures above about 110°C, silicais added to the cement. This makes the hydration process (setting and curing ofthe cement) more complex, as initially formed hydration products are replacedby more stable phases over time. To study the nanostructure and mechanicalproperties of HT-cured cement, we applied a variety of techniques includingsmall-angle neutron scattering (SANS), transmission electron microscopy (TEM), nanoindentation, and micro-scratch testing on small cement samples andinvestigated their structure and properties under a variety of conditions. Weobserved that, at HT, there is a general coarsening of the nanometer-scalestructure of the set cement paste over time, with associated degradation of the properties. We showthat the rate of coarsening depends strongly on the initial curing conditions, providing possible strategies for improving the properties and performance ofHT-cured cement. The findings may have particular application to geothermal wells andsteam injection wells. 1. Introduction: Oilwell cementing, in which the annular space between the drilled formationrock and the steel casing pipe is filled with a cement slurry, presentssignificant challenges due to the wide range of environmental conditionsencountered. In particular, the high temperatures and pressures experienced in deep wells generatesignificant changes in the chemical nature and morphology of the hydrationproducts (Taylor, 1997). At near-ambient conditions, the main hydration productof Portland cement is calcium-silicate-hydrate (C-S-H) gel with a Ca/Si molar ratio (C/S)of about 1.7 (Richardson and Groves, 1992, Taylor, 1997). The C-S-H phase hasan amorphous or poorly crystalline nature and a high specific surface area, andis responsible for the generally excellent engineering properties of cement andconcrete cured under normal conditions. As the curing temperature increases toabout 110!C, the C-S-H phase becomes increasingly crystalline but retains itsgeneral structure and properties. Above 110!C, a different phase, alpha-dicalcium silicate hydrate (Ca2SiO2(OH)2) becomes stable and replacesC-S-H. As the C-S-H phase converts to Ca2SiO2(OH)2, the permeability increasesand the strength decreases. To avoid this, oilwell cements for use above 110!C contain a significant amountof silica (approximately 35% by weight of cement) to lower the C/S of thehydration products to about 1. Under these conditions, a range of crystallinecalcium silicate hydrate minerals with generally acceptable engineering properties can form(Taylor, 1964, Kalousek, 1968, Eilers and Root, 1976). While the equilibrium phases over a wide range of conditions are wellestablished (e.g. Shaw et al., 2000, Meller et al., 2009), metastable phasesoften form and linger for extended periods of time. In addition, the morphologyof the products are also highly variable. Thus, it is quite difficult to predict the nature of thehydration products that will develop under particular conditions of time, temperature and pressure. Here we apply novel techniques to investigate theproperties of cement/silica blends cured at high temperature and pressure. Small-angle neutron scattering (SANS) and transmission electron microscopy(TEM) are used to obtain information about the nanometer-scale structure, andnanoindentation and microscratch testing are used to determine the mechanicalproperties.