Figure 4 (a) to (d) shows the self-healing process of MBA-BDA polymer at room temperature. The polymer plate was made with MBA-BDA powder heated to 100 °C and then pressed against a round plate and cooled down to room temperature. At room temperature, the broken polymer plate can heal multiple times. To determine if adding MBA-BDA co-polymer to the cement formulation brings about self-healing capability to the cementitious material, MBA-BDA copolymer in the form of a powder was added to the cement slurry in a concentration of 10wt% based on dry cement. The copolymer-modified cement slurry was transferred to a cylindrical mold and cured for 24 hours. Then, the cylinder was mechanically split into two pieces followed by holding together the two pieces with the assistance of a rubber band to improve physical contact. The cement sample demonstrated to heal itself at room temperature since the two pieces “glued” together after 24 hours (Fig. 4(e)).
To quantify the self-healing capability of polymer-modified cement and how it compares to unmodified cement, compressive strength analysis of original (un-damaged) cement samples and the same samples after damage followed by an aging (healing) event was performed with (minimum) three samples per formulation. After the compressive strength test, the samples were aged for a week to allow for regeneration of the microstructure before performing a new compressive strength test. As described in the Methods section, the compressive strength tests were performed in two stages, applying fast load increments (vertical load displacement of 0.001” /sec up to 300 lbs) followed by a slow (vertical load displacement of 0.0001”/sec) to generate micro-fractures while preventing irreversible damage of the cement’s structural integrity. These damage (compressive strength test) / healing events (sample aging for a week) were repeated six times. The results shown in Fig. 5(a) for unmodified (control) cement indicate a continuous compressive strength decline, particularly after the fourth test. In addition, the uncertainty or variability of the compressive strength data is significantly high. This is even more evident in the recovery ratio values where the uncertainty of the recovery % is higher than the average for nearly every test (Fig. 5(c)). Therefore, the recovery ratios for all six tests in unmodified samples ranging from 49–101% are misleading.
In contrast, the MBA-BDA-modified cement samples show more consistent compressive strength values across all six tests and the resulting recovery ratios seem to indicate that the cement increases its healing capability over time (tests 1 to 3 and tests 4 to 6) with recovery increasing from 48–193% and from 59–177%, respectively. The reason for a drop in compressive strength recovery ratio between the third and fourth test is unclear but could be due to a significantly higher damage post third test where instead of the controlled formation of microfractures, larger fractures were produced. However, the increase in recovery of mechanical strength after the fourth test seems to indicate that the self-healing capability combined with longer hydration times are responsible for the increase in recovery ratio. A recovery ratio higher than 100% suggests that the recovery of mechanical strength is a combination of both, the continued hydration of the cement sample (autogenous healing) and the self-healing process (autonomous healing). This is not the case for control cement due to the absence of autonomous healing. Plots in Fig. 5(b) and Fig. 5(d) show that MBA-BDA-modified cement is the only formulation with self-healing ability.
The rheological properties of cement slurries are important to ensure that the cement is workable during the application process. The rheological characteristics of a standard cement consisting of 70 wt % cement, 30 wt % silica flour, and a water-to-cement weight ratio (w:c) of 0.54 (control sample) were determined followed by the evaluation of MBA-BDA-modified cement. As shown in Fig. 6, 10 wt% MBA-BDA-modified cement shows a unique plastic viscosity development history in the first 3 h of hydration. The plastic viscosity decreases as a function of time with a higher value than plain cement but a lower value after three hours of hydration. We hypothesized that the MBA-BDA polymer dissolution in the aqueous slurry is kinetically controlled in this process. As time evolves MBA-BDA polymer slowly dissolves in the cement slurry with the resulting adsorption of polymer moieties on the surface of cement particles (via hydrogen bonding between -NH functionalities in the polymer and oxygen atoms in the unhydrated cement grains as well as at the Calcium Silicate Hydrate (C-S-H) hydration products). In this fashion, the cement hydration process is gradually delayed with a resulting reduction in viscosity of the three-dimensional C-S-H hydration gel system. Unexpectedly, the yield stress of the MBA-BDA cement is about one order of magnitude lower than that of control unmodified cement samples as shown in Fig. 6 (b). This indicates that MBA-BDA cement (in the first three hours of hydration) behaves closer to a Newtonian fluid which makes it easier to pump especially under low shear rate conditions.
To better understand the physical and chemical properties of these polymer-modified cements, including proposing potential mechanisms for self-healing behavior, we performed X-ray microtomography (XMT) with density contrast analysis, X-ray diffraction (XRD) spectroscopy, and 13C NMR of control cement samples and polymer-modified cement samples. Figure 7 shows XMT micrographs (left) and segmentation images (right) for conventional (unmodified) cement and MBA-BDA copolymer-modified cement. The copolymer-modified cement shows the homogeneous distribution of the polymer (in blue) in the cement matrix available for autonomous healing of cracks.
Figure 8 shows XRD spectra of unmodified and polymer-modified cement (MBA-BDA-cement). The virtually complete overlapping of the XRD data of both samples highlights their similar mineralogy. X-ray diffraction analysis indicated that the samples were composed of quartz (SiO2, 033-1161), portlandite [Ca(OH)2, 044-1481], and gypsum (CaSO4•2H2O, 033–0311). The most prominent difference between the two patterns was in the relative intensities of the quartz reflections, attributed to some preferential orientation of durable quartz grains. The analysis also demonstrates that the cement signature is not altered by the addition of polymer.
13C solid-state NMR results are shown in Fig. 9 for unmodified cement, neat polymer, and the corresponding polymer-modified cement samples. The most important carbon shifts are those associated to the carbonyl groups found between 150–200 ppm. While the carbonyl originating from cement (see cement control A) remains constant in the polymer-cement sample with a carbon peak at 168.5 ppm assumed to be the calcium carbonate feature of the cement, the carbonyl of the polymer shifts. This is evidenced in trace C where the carbonyl of the MBA-BDA polymer is shifted downfield.
The peak moves downfield from 174 to 181 ppm which is consistent with the hypothesis that the carbonyl moiety is hydrogen bonding with the hydrogen of C-S-H in cement. The fact that the peak broadens in spectrum C is further evidence of H exchange between polymer and cement. This result suggests that at least one of the mechanism responsible for self-healing is the formation of multiple hydrogen bonds between the polymer and cement.11.