The Scherer-Prévost Group’new experimental techniques, including time-lapse photography, have provided improved measurements of the kinetics of corrosion of cement. The results are being used to verify numerical models and to test the effects of reservoir conditions (e.g., pH, CO2 concentration) on the rate of corrosion. An important observation, predicted by their corrosion model, is that the reaction is faster at a given pH when CO2 is present (that is, carbonic acid is more aggressive to cement than hydrochloric acid). Moreover, time-lapse images show that the corrosion layer contracts, which could contribute to opening of an annular gap by flowing acid.


Measurement of corrosion kinetics

The results of the group’s initial flow-through and batch experiments carried out by Andrew Duguid (now with Schlumberger) are now being published. Graduate student Ed Matteo has been extending this work by using time-lapse photography to measure the rate of corrosion of cement by acidic brine with greatly enhanced precision (Figure 7). His results show two new features: (1) at a given pH, the rate of attack is much faster when COis dissolved in the water (as predicted by Bruno Huet’s model); (2) the corroded layers contract (that is, the residual silica gel occupies less volume than the cement from which it formed), particularly at lower pH. The kinetics of attack in these experiments are clearly diffusion-controlled, as the depth of the corroded layer increases strictly in proportion to the square root of time.

The pore structure and composition of the corroded samples is being examined by nitrogen sorption and electron microscopy. The sample shown in Figure 8 was soaked in ethanol and then dried at 100°C. The point of the alcohol exchange is to reduce the surface energy, so that less capillary pressure develops during drying; nevertheless, the corroded layer (primarily silica gel) cracks during drying. In the future, the researchers will put the water-saturated sample into the ESEM and examine it wet, so that this artificial cracking does not occur. In the spring, the diffusivity within the pores will be measured by NMR, in collaboration with Dr. Leo Pel at the Technical University in Eindhoven.

Figure 7. USB video camera focused on the surface of a sample of cement that is in contact with an acidic solution. The insert at right shows the corroded layers (white and orange layers over the gray unaltered cement) as seen by the camera.


Figure 8. A cross‐section of cement paste exposed to hydrochloric acid with a pH of zero (1 Molar HCl), which is 1000 times stronger than what might be found in a well. The corroded surface layer (~0.5 mm thick) is clearly seen at the top of this image, made with the Environmental Scanning Electron Microscope (ESEM).

A new series of experiments has been designed in collaboration with Bruno Huet to test his predictions regarding the effect of carbonate concentration in the water on corrosion kinetics. At certain concentrations, precipitation of calcium carbonate is expected to cause pore blocking, which will significantly reduce the rate of attack. If this effect is verified, it could explain some discrepancies among the corrosion studies performed at Princeton, NETL, and Schlumberger, where different solution compositions were employed.


Reactive transport modeling

In collaboration with Dr. Bruno Huet (now at Schlumberger), the Scherer-Prévost Group has published an analysis of the kinetics of corrosion of cement paste by a carbonated brine, using a reactive transport module within Dynaflow. The methodology for coupling transport and geochemical modules is derived and its assumptions are discussed. The modules are coupled in a sequential iterative approach to model: (1) mineral dissolution/ precipitation, (2) aqueous phase speciation, and (3) porosity-dependent transport properties. Simulation results reproduce qualitatively the dissolution of cement hydrates (calcium hydroxide, calcium-silicate-hydrate, and sulfate phases) and intermediate products (CaCO3) that have been observed experimentally. However, when using a standard power law to relate effective transport properties to porosity, modeling and experimental results do not coincide; here, agreement between simulations and observations is obtained by assuming the functional dependence of effective diffusivity on mineralogy. This issue is being addressed experimentally be Ed Matteo. The group is also working with Bruno Huet to evaluate how the boundary conditions (pH, and concentrations of CO2 and Ca2+ ions) control cement reactivity. Complementary experiments are being done by Ed Matteo.


Evaluation of field samples

In collaboration with colleagues at NETL, the Scherer-Prévost team completed a microstructural and compositional analysis of cement samples retrieved from Teapot Dome. Attempts to retrieve cores from a 19-year old well at Teapot Dome yielded two samples of cement. One sample, from the Wall Creek (Frontier) Sandstone (3060 ft), was clearly contaminated by the lead cement, as it contained fly ash that was not included in the tail cement. That sample experienced sulfate attack, apparently from reaction with the formation water, resulting in formation of ettringite. The sample that was in contact with a dolomitic (Tensleep) formation showed calcium loss and magnesium enrichment, apparently owing to reaction of dolomite with calcium hydroxide in the cement. These results indicate that significant modification of the well cement can be expected in older wells, and this should be taken into account in modeling of the chemical durability of cements exposed to geosequestered CO2.