Combined Functionalization of Carbon Nanotubes (CNT) Fibers with H 2 SO 4 /HNO 3 and Ca(OH) 2 for Addition in Cementitious Matrix

: Acid treatment is commonly used to improve the dispersion of carbon nanotubes (CNT) in a cementitious matrix, but it causes undesired delay on cement hydration kinetics. This work reports a combined CNT functionalization method with H 2 SO 4 /HNO 3 and Ca(OH) 2 for addition in a cementitious matrix. Results showed that the Ca(OH) 2 exposure neutralized the active sites generated by acid exposure, compensating the delay in hydration. As a result, CNT exposed to H 2 SO 4 /HNO 3 for 9 h and further Ca(OH) 2 treatment led to equivalent hydration kinetics than un-treated CNT did with improved stability.


Introduction
Carbon nanotubes (CNT) incorporation in cement-based materials can improve their mechanical performance by acting as bridges across cracks and voids, therefore increasing the strength and toughness of the composite [1][2][3][4]. In addition, the very high specific surface area (SSA) of CNT tends to accelerate the early hydration kinetics of cement, resulting in mechanical strength gains even at early ages [5][6][7][8]. However, properly dispersing CNT is a challenge due to its high SSA, clustering formation, and Van der Waals attractive forces [9][10][11][12].
When CNTs are added to fresh cement mix, the polar groups formed by the functionalization process can bind with ions from the pore solution, delaying the cement hydration kinetics. This delay can slow down the formation of the microstructure of the cementitious matrix [4,32,33], impairing the mechanical strength of the composite at early ages [34][35][36][37]. In this context, Mendoza Reales et al. [38] evaluated the effect of a commercial dispersion of CNT on the hydration of a class G cement paste, finding that CNT addition delayed the induction period by up to 15 h, and this delay was proportional to the CNT content added. Torabian Isfahani et al. [39] produced cement pastes with 0 and 0.1 wt. % of functionalized CNT, finding that the incorporation of functionalized CNT reduced the 24-h cumulative heat by 3% compared with plain cement paste. Indirectly, Ahmed et al. [40] observed lower strength increases when functionalized CNT were added compared with un-treated CNT.
In order to prevent such undesired delay on cement hydration, this work investigated a combined functionalization of CNT with acid and calcium hydroxide for addition in cementitious matrix.

Materials
CNT manufactured by Cheap Tubes Inc. was used, with density of 1.7 g/cm 3 , external diameter of 8-15 nm, length of 10-50 nm, and 95% purity. Scanning electron microscopy (SEM) images of CNT are shown later in Section 3.1. A Portland cement type CP I-S [41] was used for paste preparation, the characteristics of which were provided by the manufacturer, as presented in Table 1. A polycarboxylate-based superplasticizer was used to achieve proper flowability, with pH of about 6, density of 1.09 g/cm 3 , and solid content of 30.5%. The pastes were prepared with distilled water, using a water/cement ratio of 0.30 by weight, with the addition of both superplasticizer and CNT in 0.1 wt. % over the cement weight. Firstly, CNTs, water, and superplasticizer were mixed and sonicated for 20 min at 55 kHz in a low energy water bath sonicator (to prevent damaging the CNTs). Then, this suspension was mechanically mixed with the cement for 3 min.

CNT Treatment
Functionalization was conducted by immersing the CNT in a solution containing one part of HNO 3 (65%) and three parts of H 2 SO 4 (92%) in volume, for three different exposure times: three, six, and nine hours. The mixtures contained 100 mL of the acid solution and 10 mg of nanotubes. During the exposure, samples were kept in agitation using an ultrasonic bath. After the exposure to acid, samples were diluted in distilled water in the ratio of 100 mL of water for every 1 mL of acid + CNT solution. The solution was kept resting for 1 h, then filtered. The material retained on the filtration membrane was washed with distilled water and filtered until the residue reached a pH of 6 ± 1. In addition, a sample of CNT exposed to 9 h of acid treatment was further immersed in an aqueous solution of calcium hydroxide (Ca(OH) 2 ) with a concentration of 1 g/L for 24 h, subsequently being washed and air-dried.
The samples were named according to the CNT treatment: CNT-0, -3, -6, and -9 respectively for zero, three, six, and nine hours of acid exposure, in addition to CNT-9-CH for the sample with coupled acid and Ca(OH) 2 treatment.

Testing Methods
Zeta potential was conducted using a Zetasizer Nano ZS (Malvern) equipment in pH of 2-12. Raman spectrometry was conducted in a Ínvia (Renishaw) equipment with argon laser and wavelength of 514.5 nm within the green range of the visible light spectrum. SEM was conducted using a JSM 6701F (JEOL) microscope operating at 15 kV. Isothermal calorimetry was conducted on a TAM Air (TA Instruments) calorimeter at 21 • C for up to 50 h in paste samples of about 10 g. Figure 1 shows the zeta potential and sedimentation of the CNT with different treatments. Un-treated CNT had an isoelectric point at a pH of 9.5 while increasing the H 2 SO 4 /HNO 3 exposure time progressively decreased the pH of the isoelectric point. Consequently, CNT-9 had the highest stability among the samples (as visually confirmed in Figure 1b), with no isoelectric point and zeta potential from about −20 to −36 mV for pH of 8.5-12.0. The increase in exposure time to acid treatment progressively increases the formation of defects, i.e., active sites, as previously reported by [26,42]. In turn, subsequent Ca(OH) 2 neutralization (CNT-9-CH) reduced the absolute zeta potential values compared with the non-neutralized suspension (CNT-9), indicating that the rapidly ionizable active sites created by the functionalization-which generate negative charge in solution-have been reduced.    Furthermore, the pH of cement pore solution is around 12, at 25 • C [43]. For this pH, all the CNT submitted to acid treatment had zeta potential from −33 to −36 mV, while un-treated CNT had −22.8 mV, confirming that H 2 SO 4 /HNO 3 functionalization increased CNT stability in the pH of cement paste. The additional Ca(OH) 2 neutralization led to a zeta potential of −26.3 mV, partially reducing the dispersibility of CNT when compared to acid-traded CNTs, but still 15% higher than un-treated CNT. Figure 2 shows the Raman spectra of the CNT with different exposure times to H 2 SO 4 /HNO 3 solution, where two bands are identified: D-bands (ID) at around 1340 cm −1 , which corresponds to the disorder-induced band associated with structural defects, and Gbands (IG) around 1570 cm −1 , associated with carbon-carbon stretching from graphite [44]. An increase in the ID/IG ratio was observed when CNTs were treated regardless of the exposure time, confirming the increase in the number of structural defects on the surface of the nanotubes [45]. Despite the increase in the number of defects with the functionalization, SEM images in Figure 3 showed that no significant changes in the overall morphology of CNT occurred, indicating that the acid treatment was not harmful to the integrity of the CNT.   Figure 4 shows the isothermal calorimetry curves of the cement pastes, and Table 2 summarizes the testing results. The "Control" mix corresponds to a plain cement paste, i.e., without CNT addition. The incorporation of un-treated CNT (CNT-0) enhanced the early-age hydration of the cement compared with the Control mix, anticipating the occurrence of the main peak of heat release from 21.5 to 20.2 h and increasing the heat flow peak value from 2.03 to 2.18 mW/g of cement while reducing the induction period. The latter corresponds to the period of low reaction rate before the main cement reaction peak and is pointed out in Figure 4a. This behavior can be attributed to the high SSA of CNT: the incorporation of small particles provides extra surface to the nucleation and growth of the hydrated products resulting from the cement reaction, thus enhancing the early-age hydration kinetics [46][47][48][49], as reported by [5] for CNT incorporation in cement composite. In contrast, CNT functionalization delayed the cement hydration kinetics, and such delay increased as the acid exposure time increased: the main heat flow peak reduced down to 1.98 mW/g of cement and occurred up to 25.6 h for CNT-9. The polar groups created by the acid functionalization (e.g., -COH and -COOH) can bind with the Ca(OH) 2 formed Fibers 2021, 9, 14 5 of 8 by cement hydration reactions [29,30]. Since the saturation of Ca 2+ and consequent precipitation of Ca(OH) 2 is recognized as the trigger for the end of the induction period [35], the capture of Ca 2+ ions by those groups can delay that saturation and consequently the overall cement reactions.   Figure 4 shows the isothermal calorimetry curves of the cement pastes, and Table 2 summarizes the testing results. The "Control" mix corresponds to a plain cement paste, i.e., without CNT addition. The incorporation of un-treated CNT (CNT-0) enhanced the early-age hydration of the cement compared with the Control mix, anticipating the occurrence of the main peak of heat release from 21.5 to 20.2 h and increasing the heat flow peak value from 2.03 to 2.18 mW/g of cement while reducing the induction period. The latter corresponds to the period of low reaction rate before the main cement reaction peak   Figure 4b shows the results for the pastes containing CNT with (CNT-9-CH) and without (CNT-0 and -9) subsequent Ca(OH) 2 neutralization. The neutralization compensated the hydration delay caused by the acid treatment even for the highest exposure time investigated, resulting in a heat flow peak value of 2.11 mW/g of cement and occurrence time of 20.5 h for CNT-9-CH, similar to those of the CNT-0 mix (2.18 mW/g of cement and 20.2 h, respectively). This indicates that the polar groups formed in CNT bound with Ca 2+ ions from the Ca(OH) 2 neutralization solution before its addition in the paste, preventing the undesired delay in the cement hydration kinetics. without (CNT-0 and -9) subsequent Ca(OH)2 neutralization. The neutralization compensated the hydration delay caused by the acid treatment even for the highest exposure time investigated, resulting in a heat flow peak value of 2.11 mW/g of cement and occurrence time of 20.5 h for CNT-9-CH, similar to those of the CNT-0 mix (2.18 mW/g of cement and 20.2 h, respectively). This indicates that the polar groups formed in CNT bound with Ca 2+ ions from the Ca(OH)2 neutralization solution before its addition in the paste, preventing the undesired delay in the cement hydration kinetics.

Conclusions
This work investigated the combined functionalization of CNT with H 2 SO 4 /HNO 3 and Ca(OH) 2 for addition in a cementitious matrix. Zeta potential and visual sedimentation indicated that the acid functionalization improved the dispersion and stability of CNT in water. Despite the increase in ID/IG ratio in Raman, SEM showed that H 2 SO 4 /HNO 3 exposure did not lead to evident damage in the CNT's morphology. However, the increase in the acid exposure time progressively delayed the early-age hydration kinetics of cement paste with the CNT. Further Ca(OH) 2 neutralization limited the dispersion improvement promoted by acid treatment but compensated the undesired hydration delay, leading to equivalent cement hydration kinetics to un-treated CNT with improved stability.