*3.3. 29Si*

The 29Si NMR spectrum of the anhydrated cement (Figure 3 (top)) shows a single signal, which is superposition of a narrow line at about −70.6 ppm from silicon atoms in the structure of belite and a broad asymmetric line in the range from −65 to −75 ppm, which corresponds to silicon atoms in the structure of alite [20] that are located in various local environments.

**Figure 3.** 29Si NMR spectra of the (**top**) anhydrated Portland-limestone cement, and (**bottom**) cement paste sample without additives at the 7th day of hydration.

During the initial stage of hydration of all three cement mixtures, broad unresolved lines appear in the spectrum initially in the range from −75 to −95 ppm, and then the specified range expands to chemical shift values of about −120 ppm (Figure 3 (bottom)) shows the 29Si NMR spectrum from the sample **C07**; the rest of the spectra are shown in Supplementary Figures S7–S9). This behavior corresponds to the appearance of an inhomogeneous phase, containing silicon atoms in tetrahedral environment of oxygen atoms, which are characterized by the presence of one to two linked silicon tetrahedra (Q<sup>1</sup> and Q<sup>2</sup> structural elements, from −75 to −90 ppm) or of three to four linked silicon tetrahedra (Q<sup>3</sup> and Q<sup>4</sup> structural elements, from −90 to −120 ppm) [12]. The presence of the former two structural elements in the cement paste characterizes the formation of the mostly amorphous C–(A−)S−H gel, while the latter two point to the formation of crosslinked silicate chains (C−(A−S−H of low Ca/Si ratio)) and amorphous hydrous silica [21].

At the later stages of hydration, a narrow line at about −86 ppm appeared in the spectra. This line corresponds to silicon atoms in Q2 structural elements of the cement paste, in paired (Q2 P) and/or bridged (Q2 b)silicon tetrahedra in silicate chains, which form the bulk structure of the resulting cement paste [22].

#### **4. Discussion**

Figure 4 shows the time dependences of the relative integrated intensities of the 13C NMR signals for all three studied samples. It can be seen that the intensities of the lines corresponding to amorphous calcium carbonate sharply decrease at the initial stage of the hydration process for **C** and **CAA** samples, and then a slight increase is observed. For the **CSA** sample, a gradual decrease in the relative proportion of amorphous calcium carbonate is observed. The proportion of calcite increases for all the samples, while that of aragonite decreases, as is especially noticeable for **C** sample. However, at the end of the studied hydration period (15–34 days), the relative content of calcite and aragonite essentially stabilizes and for **CAA** sample even slightly reverses. It should be noted that a significant amount of aragonite is observed on the first day of hydration only for **C** sample.

**Figure 4.** Time-evolution (in logarithmic scale) of the relative integrated areas of the isotropic components recognized in the 13C NMR spectra of the samples: **C** (**a**), **CAA** (**b**) and **CSA** (**c**).

The increase in the content of the poorly soluble calcium carbonate polymorphs (calcite and aragonite), as noted for all samples, leads to their precipitation in the pores of the hardening cement paste, and this might cause an increase in its strength and a decrease in porosity [1,2]. The smallest amount of these calcium carbonate polymorphs is observed for the sample **CAA**.

Evidently, the formation of aragonite in the samples containing organic additives begins only on the second day of hydration, while the fraction of the initial amorphous CaCO3 in the **CSA** sample, on the first day, is much lower than for the other samples. This observation can be attributed to the fact that the acetate ion (CH3COO−) can be adsorbed on the surface of the anhydrated cement microparticles and prevent their hydration [10,23], and also the crystallization of new phases [24]. However, in the case of acetic acid addition, the acidity of the pore solution increases, and this increase, at the initial stage of hydration, contributes to the dissolution of the fine particles of the anhydrated cement. At the same time, the presence of sodium cations hinders this process, forming a weakly alkaline medium in **CSA** mixture.

It is worth noting that it is not possible to quantify the amount of calcium monocarboaluminate hydrate (AFm) in **CAA** sample from 13C NMR spectra, although this compound is resolved in the corresponding 27Al NMR spectra. This is because of the negligible 13C NMR chemical shift difference between AFm phase and other calcium carbonates (calcite and vaterite) that prevents a reliable deconvolution of the overlapped signals [14].

Figure 5 illustrates the time dependences of the relative integrated intensities of the 27Al NMR signals for all three studied samples. Considering the observed changes in the intensities of the aforementioned spectral components, it can be concluded that, after the initial dissolution of the aluminate phases of the cement used, a large amount of aluminate hydrate forms, whose quantity gradually decreases. Then the amount of various compounds in the form of C–S–H, A–S–H or C–(A–)S–H gels grows in volume, further gradually decreasing. Finally, a gradual increase in the amount of various calcium hydroaluminates and ettringite is observed.

It should be noted that, for the **CAA** and **CSA** samples, on the first day, there is an insignificant amount of residual aluminum impurity in C2S and C3S. This can be associated with incomplete hydration of the anhydrous cement particles, due to the adsorption of the acetate ions on their surface.

The diagrams of Figure 5 show that, for all samples, the amount of ettringite initially increased and then decreased, and, at later stages, it again increased [2]. According to the generally accepted theory of hydration of aluminum-containing cements, ettringite crystallizes in two stages. In the initial stage, long narrow crystals form, which contribute to the initial binding of the hydrated cement grains. Later, during the deceleration of the hydration process, the initially formed ettringite recrystallizes in the form of large crystals in the voids of the matrix. Moreover, for the **C** sample, recrystallization of ettringite practically did not occur, while, in the **CAA** sample, the amount of the primary and secondary ettringite is noticeably larger than for all the other samples. The presence of acetate groups in the **CAA** and **CSA** samples can partially replace the sulfate groups [24]. Hence, the excess of the latter facilitates the formation of primary ettringite, whose content is larger than that in the **C** sample, resulting in the kinetics observed.

Since the surface of the aluminum-containing clinker phases is more electronegative than that of C3S and C2S [11], their dissolution occurs faster and, in parallel, a deficiency of calcium arises. Thus, in the **CAA** and **CSA** samples, at the early stages of hydration, an increased amount of aluminate hydrate is observed, which subsequently, with an increase in the calcium content, gradually transforms into the more stable C3AH6 phase.

Moreover, aluminum actively passes into the crystalline phases of ettringite and calcium hydroaluminate; hence, its content in the amorphous C–(A–)S–H phase decreases. It should be noted that the increased ettringite content observed in the **CAA** sample may act as a risk factor for sulfate corrosion.

**Figure 5.** Time-evolution (in logarithmic scale) of the relative integrated areas of the isotropic components recognized in the 27Al NMR spectra of the samples: **C** (**a**), **CAA** (**b**) and **CSA** (**c**).

Figure 6 illustrates the time dependences of the relative integrated intensities of the 29Si NMR signals for all the three studied samples. When analyzing the change in the relative integrated intensities of the observed spectral components during hydration, we observed that the mass fraction of the silicate-containing clinker phases gradually decreases for all samples, while the mass fraction of the C–(A–)S–H phase increases proportionally, as well as the fraction of paired Q2 terahedra. It should be noted that the Q2 tetrahedra

resolved in the spectrum appear on the first day for the **CSA** sample, on the second day for the **C** sample, and only on the third day for the **CAA** sample. Moreover, for the **CAA** sample, the spectral component, which is visually distinguishable from the baseline and is characteristic for Q3 and Q<sup>4</sup> structural elements, also appears only on the second day of hydration.

**Figure 6.** Time-evolution (in logarithmic scale) of the relative integrated areas of the isotropic components recognized in the 29Si NMR spectra of the samples: **C** (**a**), **CAA** (**b**) and **CSA** (**c**).

These observations, along with the fact that no resolved peaks arise from other characteristic Q1, Q2 and Q<sup>3</sup> structural elements [22], may indicate that such a characteristic layered structure of hydrated cements remains mainly amorphous; however, the number of paired Q<sup>2</sup> tetrahedra increases, and this increase can correspond to an increase in the length of silicate chains, consisting of paired silicate tetrahedra. The presence of such phase corresponds to an increase in strength of the cement matrix. The formation of this phase for the **CAA** sample is observed at later stages of hydration.

It should be noted that, during cement hydration, a relative redistribution of the amounts of alite and belite occurs (Figure 7). For all the pastes, an increase in the relative content of belite is observed that is much larger for the **CAA** sample as compared to the other two.

**Figure 7.** Time-evolution (in logarithmic scale) of the relative integrated areas of the isotropic component of calcium silicates (belite (C2S, blue markers) and alite (C3S, red markers)) recognized in the 29Si NMR spectra of the studied samples (**C** (filled points), **CAA** (empty points) and **CSA** (crossed points)).

Over the entire investigated time interval, the silicon-containing anhydrous phases did not completely hydrate. It should be noted that, in the **CAA** sample, the remaining amount of such phases is slightly less than for the other two samples; the addition of acetic acid leads to the involvement of a larger amount of alite in the formation of the C–(A–)S–H gel. That is, at the later stages of the hydration process of the **CAA** paste a smaller amount of alite remains anhydrated, and, thus, it is more effectively transformed to amorphous hydrate phase, affecting the strength properties of the hardened material.

#### **5. Conclusions**

In contrast to X-ray studies, one of the advantages of the NMR method is the ability to directly observe the signals of the nuclei both in the crystalline and amorphous local environments. As a result, in this work, it was possible to trace the time dependences of a set of chemical phases in the studied cement pastes. Despite the natural difficulty in obtaining unambiguous deconvolution of strongly overlapped signals of some cases, it was possible to identify the main components the presence of which is assumed in the chemistry of cementitious materials, including the amorphous phases, especially the crucial ones containing 29Si nuclei.

Considering all the above, it can be deduced that the addition of acetic acid and sodium acetate changes the kinetics of the cement paste phase composition during the hydration process. Adsorption of the acetate ion on the surface of the anhydrated and hydrated phases has a significant effect on the hydration process when the studied organic substances are added in the mixtures. Moreover, the presence of sodium ions slightly increases the alkalinity of the pore solution, partially reducing the efficiency of such adsorption.

It can be concluded that the addition of 3% acetic acid or sodium acetate, by cement mass, to the cement paste hindered the initial stages of the hydration process. The addition of sodium acetate led to the formation of a large amount of poorly soluble forms of calcium carbonate and a significant increase in the amount of polymerized silicon-containing phases.

Concerning the sulfate degradation of the cement paste, we see that the addition of acetic acid led to the development of favorable conditions for the formation of ettringite; in contrast, the addition of sodium acetate slightly slowed down this process. Thus, in the future studies, it would be interesting to investigate whether sodium acetate is a useful additive for improving the durability of hardened cementitious materials against sulfate attack. As only cement pastes were investigated in this work, further applied studies in this field are needed.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/ma15062004/s1. Figure S1: 13C NMR spectra of the anhydrated Portland-limestone cement and cement paste samples without additives (**C**) (a) and the chemical shift of spectra components (b) at the different hydration times. Deconvolution of C07 spectra per three (c) and two (d) components. Figure S2: 13C NMR spectra of the anhydrated Portland-limestone cement and cement paste samples with acetic acid (**CAA**) (a) and the chemical shift of spectra components (b) at the different hydration times. Figure S3: 13C NMR spectra of the anhydrated Portland-limestone cement and cement paste samples with sodium acetate (**CSA**) (a) and the chemical shift of spectra components (b) at the different hydration time. Figure S4: 27Al NMR spectra of the anhydrated Portland-limestone cement and cement paste sample without additives (**C**) (a) and the chemical shift of spectra components (b) at the different hydration time. Figure S5: 27Al NMR spectra of the anhydrated Portland-limestone cement and cement paste samples with acetic acid (**CAA**) (a) and the chemical shift of spectra components (b) at the different hydration time. Figure S6: 27Al NMR spectra of the anhydrated Portland-limestone cement and cement paste samples with sodium acetate (**CSA**) (a) and the chemical shift of spectra components (b) at the different hydration time. Figure S7: 29Si NMR spectra of the anhydrated Portland-limestone cement and cement paste sample without additives (**C**). Figure S8: 29Si NMR spectra of the anhydrated Portland-limestone cement and cement paste samples with acetic acid (**CAA**). Figure S9: 29Si NMR spectra of the anhydrated Portland-limestone cement and cement paste samples with sodium acetate (**CSA**).

**Author Contributions:** Conceptualization, P.T. and A.M.; formal analysis, P.T.; investigation, A.M.; funding acquisition, P.T. and K.S.; writing—original draft preparation, A.M.; writing—review and editing, P.T. and K.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Russian Foundation for Basic Research (grant number: RFBR 20-52-26021) and the Czech Science Foundation (grant number: GACR 21-35772J). ˇ

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data available upon request to the authors.

**Acknowledgments:** The work was performed by using the equipment of the Saint Petersburg State University Research Park at the Resource Centers for Magnetic Resonance Research and X-ray Diffraction Studies.

**Conflicts of Interest:** The authors declare no conflict of interest.
