**2. Materials and Methods**

Cement pastes were prepared with a type CEM II/A-L 42.5N Portland-limestone cement (SLK Cement–Sukhoy Log, Sverdlovsk Region, Russia), distilled water and p.a. organic additives (acetic acid and sodium acetate). A water-to-cement ratio of 0.45 was used. Organic additives amounted to 3% by cement mass each, both added to the mixing water. Cement pastes were cast in cylindrical plastic molds (12 mm in diameter; 30 mm in height), where they remained sealed for 24 h. After demolding, the specimens were immersed in distilled water and kept throughout the investigated hydration periods. To study the kinetics of the changes in the phase composition of the pastes at certain ages (after 1, 2, 3, 5, 7, 14 and 34 days), about 150 mg of hardened material was scraped off the specimens' end and placed in paper bags to prevent any further hydration. Such quantity was sufficient for conducting the NMR experiment. In this work, the following nomenclature was used to label the samples: **C** (cement paste without additives), **CAA** (cement paste with acetic acid) and **CSA** (cement paste with sodium acetate). The stage of hydration was indicated by the age of the cement paste, which was added to the sample marker.

Prior to hydration, the mineralogical composition of the employed cement was determined with X-ray powder diffraction (XRPD) analysis; the results are summarized in Table 1. XRPD analysis was performed at room temperature, using CuKα radiation 2-theta range 5–80◦, step 0.0203◦, voltage 30 kV and current 10 mA. Qualitative X-ray phase analysis was carried out with the software PDXL 2.8.4.0 (Rigaku, Tokyo, Japan), with connection of PDF-2 database (International Diffraction Data Center, ICDD). Quantitative phase analysis (wt.%) was carried out by applying the Rietveld method [13] on the obtained full-profile data with the software TOPAS 4.2 (Bruker, Billerica, MA, USA).

NMR experiments were performed by using an Avance III 400 WB spectrometer (Bruker, Billerica, MA, USA) at constant magnetic field of 9.4 T. 13C, 27Al and 29Si nuclei were studied and characterized by the respective resonance frequencies of 100, 104 and 86 MHz. A probe that is able to rotate the samples at the magic angle to the direction of the constant magnetic field (stabilization accuracy of the rotation frequency ±4 Hz) and stabilize their temperature (temperature stabilization accuracy ±1 ◦C) was used. Powder samples were loaded on a 4 mm zircon oxide rotor and rotated at a frequency of 12.5 KHz at 20 ◦C. Tetramethylsilane, for 13C and 29Si nuclei, and 1 M·D2O AlCl3 solution, for 27Al nuclei, were used as external references.

**Table 1.** Quantitative phase analysis of Portland-limestone cement derived from Rietveld refinements of X-ray powder diffraction data (Bragg R factor −5.2%).


\* The formulas in brackets correspond to cement chemist notation.

All the spectra were recorded by using a single-pulse sequence. The duration of the exciting impulses was 2.5, 4.5 and 2.5 μs; the relaxation delay was 4, 2 and 4 s; the number of scans was 1024, 512 and 1024 for 13C, 27Al and 29Si nuclei, respectively.

Deconvolution of spectra into Gaussian-shape individual components was performed by the least squares method, using the software Origin 9.0. (OriginLab Corporation, Northampton, MA, USA) For all the spectra, the results of approximation were obtained with a coefficient of determination *R<sup>2</sup>* higher than 0.8. Since single-pulse sequence was used to record the NMR spectra, the relative integrated intensities of the signals can be interpreted as mole fractions of the corresponding phase components.

#### **3. Results**
