Structural Characterization of Several Cement-Based Materials Containing Chemical Additives with Potential Application in Additive Manufacturing

The rapid increase in additive manufacturing applications in all industries has highlighted the lack of innovative technologies and processes in the construction industry. Several European and international policies are in place to guide the development of the technological processes involved in the construction industry toward a sustainable future. Considering the global concerns regarding this industry, the purpose of this study was to develop new cement-based materials that are capable of accelerated hydration and early strength development for use in additive manufacturing. Ca(NO3)2·4H2O, Al2(SO4)3·18H2O and Na2S2O3·5H2O were used to obtain the accelerating effect in the hydration of Portland cement. Based on results obtained from X-ray diffraction (XRD), scanning electron microscopy and energy-dispersive X-ray spectroscopy (SEM/EDX) techniques, as well as low-field nuclear magnetic resonance relaxometry (LF-NMR) techniques, it was demonstrated that all accelerators used have a quickening effect on cement hydration. The addition of Na2S2O3·5H2O or combined Na2S2O3·5H2O and Ca(NO3)2·4H2O led to obtaining new cement-based materials with early strength development and fast hydration of microorganized internal structures, critical characteristics for 3D printing.


Introduction
The construction industry makes an essential contribution to the economic growth of a country, but it also has a significant impact on the environment due to the generated pollution. Although the COVID-19 pandemic had a negative impact on the development of the construction industry, in 2022, it generated 8.2 trillion USD, with an estimated compound annual growth rate (CAGR) of 7.3% [1] and 7% to 8.5% of global employment [2]. Regarding the environmental issues related to the construction industry, it is considered to have the most serious environmental impact among all industries on air, water, soil, and noise pollution, waste generation, and greenhouse gas emissions [3]. The World Business Council for Sustainable Development (WBCSD) and its partners proposed several guidelines for the sustainable transformation of construction industry at the Stockholm+50 international meeting in June 2022 [4]. Among these guidelines, one focuses on material efficiency and innovation [5]. The recent trends in construction materials follow this guideline to maximize process efficiency and reduce cost by: introducing polymer-derived fibers into concrete to make it bendable; using softwood to reduce emissions and waste in the manufacturing process; continuing to use recycled materials; and using 3D printing to reduce material losses, reduce manpower and execution time [6].
Much work is still needed to find fast, efficient, cost-effective, and environmentally friendly solutions to overcome the shortcomings of AM in the construction industry. This work is being conducted in response to the need to find suitable cement-based material that can meet the needs of the emerging construction industry. Five new cement-based materials using Ca(NO 3 ) 2 ·4H 2 O, Al 2 (SO 4 ) 3 ·18H 2 O and Na 2 S 2 O 3 ·5H 2 O and different combinations thereof were prepared, tested, and compared with white cement 52.5R by XRD diffraction, SEM/EDX, and LF-NMR to provide information about their structure and their role as accelerators in cement-based materials.

Results and Discussion
The samples were analyzed by X-ray diffraction. Figure 1 shows the X-ray diffraction patterns and their assignments using the abbreviations listed in Table 1. The diffraction patterns of all investigated samples show narrow peaks characteristic of several crystalline mineral phases identified by their PDF (powder diffraction file).
hydration of the cement, formation and distribution of hydration products, and the homogeneity of the cement paste [29][30][31][32][33]. LF-NMR is a noninvasive and nondestructive quantitative technique that provides valuable information on hydration kinetics, porosity, and pore size distribution [27], and has been used in cement and concrete research for many years [34,35].
Much work is still needed to find fast, efficient, cost-effective, and environmentally friendly solutions to overcome the shortcomings of AM in the construction industry. This work is being conducted in response to the need to find suitable cement-based material that can meet the needs of the emerging construction industry. Five new cement-based materials using Ca(NO3)2·4H2O, Al2(SO4)3·18H2O and Na2S2O3·5H2O and different combinations thereof were prepared, tested, and compared with white cement 52.5R by XRD diffraction, SEM/EDX, and LF-NMR to provide information about their structure and their role as accelerators in cement-based materials.

Results and Discussion
The samples were analyzed by X-ray diffraction. Figure 1 shows the X-ray diffraction patterns and their assignments using the abbreviations listed in Table 1. The diffraction patterns of all investigated samples show narrow peaks characteristic of several crystalline mineral phases identified by their PDF (powder diffraction file).   Table 2 and Figure 2 show the results of the semiquantitative analysis of the mineral phases present in the samples and their degree of crystallinity (the ratio between the integrated area of crystalline species and the integrated area under the overall XRD profile or sum of crystalline and amorphous species in arbitrary units).   Table 2 and Figure 2 show the results of the semiquantitative analysis of the mineral phases present in the samples and their degree of crystallinity (the ratio between the integrated area of crystalline species and the integrated area under the overall XRD profile or sum of crystalline and amorphous species in arbitrary units).   The results of the X-ray diffraction show different variations in C3S, C2S, CC and CH content when accelerators are introduced to Portland cement. Some of the interpretations are presented below.

•
The accelerators inhibit C3S hydration when added to Portland cement. The addition of Na2S2O3 or Ca(NO3)2 simultaneously with Na2S2O3 and Al2(SO4)3 results in an increase in C3S content over Ca(NO3)2. The highest value for C3S content is obtained for CPNa, while the lowest value is assigned to CPCa. It is well known that Na compounds are very reactive, especially when combined with water [37], accelerating hydration and increasing the amount of hydration products [23]. Rapid hydration of the cement can lead to early freezing of the cement and thus an increase in the C3S content.

•
The addition of accelerators to Portland cement contributes to a reduction in the content of C2S supporting its hydration. The highest value is obtained in the case of CP2CaNa and decreases for the samples CPCa, CCPaNa, and CPNa, which will have similar values, while the lowest value is obtained in the case of CCPaNaAl. The results of the X-ray diffraction show different variations in C 3 S, C 2 S, CC and CH content when accelerators are introduced to Portland cement. Some of the interpretations are presented below.

•
The accelerators inhibit C 3 S hydration when added to Portland cement. The addition of Na 2 S 2 O 3 or Ca(NO 3 ) 2 simultaneously with Na 2 S 2 O 3 and Al 2 (SO 4 ) 3 results in an increase in C 3 S content over Ca(NO 3 ) 2 . The highest value for C 3 S content is obtained for CPNa, while the lowest value is assigned to CPCa. It is well known that Na compounds are very reactive, especially when combined with water [37], accelerating hydration and increasing the amount of hydration products [23]. Rapid hydration of the cement can lead to early freezing of the cement and thus an increase in the C 3 S content.

•
The addition of accelerators to Portland cement contributes to a reduction in the content of C 2 S supporting its hydration. The highest value is obtained in the case of CP2CaNa and decreases for the samples CPCa, CCPaNa, and CPNa, which will have similar values, while the lowest value is obtained in the case of CCPaNaAl.

•
Regarding the reaction products C-H and CC, the following behavior can be observed in the samples containing the accelerator. The content of C-H increases for CPCa, with the lowest value being obtained for CPCaNaAl where the presence of ettringite (AFt) has been detected. The appearance of Aft in Portland cement with the addition of Al 2 (SO 4 ) 3 is due to the increase in sulfate content in the liquid hydrate phase after mixing, which promotes the production of acicular ettringite crystals [38].

•
In the case of CC, all the values obtained are below the value obtained for CP. The highest value is obtained in the case of CPNa and the lowest in the case of CPCa. Based on these results and the crystallinity degree, it can be concluded that the addition of Ca(NO 3 ) 2 in Portland cement increases crystalline C-H, increasing the crystalline degree, while the addition of Na 2 S 2 O 3 has the opposite effect. Calcium silicate hydrate (C-S-H) is the main hydration product of Portland cement, formed by the reaction of C 3 S and C 2 S with water. It is generally amorphous or weakly crystalline [26,39]. A decrease in C 2 S and a sharp decrease in the degree of crystallinity in the case of CPNa may indicate that the addition of Na 2 S 2 O 3 in CP leads to a larger amount of C-S-H hydration as a result of C 2 S hydration than in the rest of the analyzed samples.

•
The C 3 S/C 2 S ratio, which expresses the rate of cement hardening, is shown in Figure 3. It can be seen that the addition of Na 2 S 2 O 3 accelerates the cement hardening process the most and the addition of Ca(NO 3 ) 2 the least. The addition of Al 2 (SO 4 ) 3 accelerates the cement hardening process more than Ca(NO 3 ) 2 when comparing the two samples containing the same content of Na 2 S 2 O 3 , CP2CaNa and, respectively, CPCaNaAl.
• Regarding the reaction products C-H and CC, the following behavior can be observed in the samples containing the accelerator. The content of C-H increases for CPCa, with the lowest value being obtained for CPCaNaAl where the presence of ettringite (AFt) has been detected. The appearance of Aft in Portland cement with the addition of Al2(SO4)3 is due to the increase in sulfate content in the liquid hydrate phase after mixing, which promotes the production of acicular ettringite crystals [38].

•
In the case of CC, all the values obtained are below the value obtained for CP. The highest value is obtained in the case of CPNa and the lowest in the case of CPCa. Based on these results and the crystallinity degree, it can be concluded that the addition of Ca(NO3)2 in Portland cement increases crystalline C-H, increasing the crystalline degree, while the addition of Na2S2O3 has the opposite effect. Calcium silicate hydrate (C-S-H) is the main hydration product of Portland cement, formed by the reaction of C3S and C2S with water. It is generally amorphous or weakly crystalline [26,39]. A decrease in C2S and a sharp decrease in the degree of crystallinity in the case of CPNa may indicate that the addition of Na2S2O3 in CP leads to a larger amount of C-S-H hydration as a result of C2S hydration than in the rest of the analyzed samples.

•
The C3S/C2S ratio, which expresses the rate of cement hardening, is shown in Figure 3. It can be seen that the addition of Na2S2O3 accelerates the cement hardening process the most and the addition of Ca(NO3)2 the least. The addition of Al2(SO4)3 accelerates the cement hardening process more than Ca(NO3)2 when comparing the two samples containing the same content of Na2S2O3, CP2CaNa and, respectively, CPCaNaAl.    The distance between the pores and their dimensions were measured and are shown in Table 3. The minimum value for the pore diameter was for CPCa sample (~2.88 μm) and the maximum for CP (~9.44 μm). These values classified the pores observed in the investigated samples as air holes (>several μm), although they are very close to large capillary pore values (100-1000 nm) [40]. The addition of accelerators in CP leads to a reduction in pore diameter, indicating a more compact structure in these samples. The above idea is also supported by the values obtained for the pore spacing. The addition of accelerators in CP leads to a reduction in pore spacing: a small pore spacing indicates a high pore density [41].  The distance between the pores and their dimensions were measured and are shown in Table 3. The minimum value for the pore diameter was for CPCa sample (~2.88 µm) and the maximum for CP (~9.44 µm). These values classified the pores observed in the investigated samples as air holes (>several µm), although they are very close to large capillary pore values (100-1000 nm) [40]. The addition of accelerators in CP leads to a reduction in pore diameter, indicating a more compact structure in these samples. The above idea is also supported by the values obtained for the pore spacing. The addition of accelerators in CP leads to a reduction in pore spacing: a small pore spacing indicates a high pore density [41]. The distribution of the elements on the sample surface and the change in chemical composition on the sample surface can be observed by analyzing the element map. Figure 5 shows the surface and the distribution of elements on the analyzed surfaces expressed in "false" colors. Table 4 shows the elemental composition obtained from the EDX analysis. the opposite, indicating that increasing the Ca content leads to structural modification in the interior and at the surface of the samples. The ratios of Ca/Si and Ca/(Si+Al) exceed the values reported in the literature for calcium silicate hydrates [42], indicating the existence of Ca-containing phases alongside C-S-H. These results are consistent with the changes in the mineral phases of CH, SS, C2S, and C3S observed in the XRD results, suggesting that the addition of Ca(NO3)2 in Portland cement increases the formation and the content of Ca-dominant phases, like Ca(OH)2 and/or CaCO3 in the examined samples.  The NMR analysis consists of monitoring the hydration responses of the samples over 6 hours by recording the echo series every 15 min throughout the hydration period ( Figure 6). As the initial slope of echo intensity increases with echo time, it becomes obvious that the addition of accelerators leads to rapid hydration of the cement mass, particularly the addition of combined Ca(NO3)2·4H2O and Na2S2O3·5H2O. According to the results, O and Ca have the highest concentration in all samples. The Na content increases with the increase of Na 2 S 2 O 3 content, albeit sometimes unevenly. A direct connection between Ca and Si can be observed. As the Ca content increases, the Si content decreases. Although one might expect that the addition of Ca(NO 3 ) 2 to Portland cement would increase the Ca content at the surface of the samples, the results suggest the opposite, indicating that increasing the Ca content leads to structural modification in the interior and at the surface of the samples. The ratios of Ca/Si and Ca/(Si+Al) exceed the values reported in the literature for calcium silicate hydrates [42], indicating the existence of Ca-containing phases alongside C-S-H. These results are consistent with the changes in the mineral phases of CH, SS, C 2 S, and C 3 S observed in the XRD results, suggesting that the addition of Ca(NO 3 ) 2 in Portland cement increases the formation and the content of Ca-dominant phases, like Ca(OH) 2 and/or CaCO 3 in the examined samples.
The NMR analysis consists of monitoring the hydration responses of the samples over 6 hours by recording the echo series every 15 min throughout the hydration period ( Figure 6). As the initial slope of echo intensity increases with echo time, it becomes obvious that the addition of accelerators leads to rapid hydration of the cement mass, particularly the addition of combined Ca(NO 3   The distribution of transverse relaxation times versus hydration time was plotted (Figure 7). The relaxation time distributions were extracted using a numerical Laplace transform from CPMG echo series.
According to Figure 7, two peaks were observed and were assigned to the presence of two aqueous environments. The first peak, which occurs at the beginning of the hydration phase, is due to the chemically bounded water and occurs when a certain The distribution of transverse relaxation times versus hydration time was plotted (Figure 7). The relaxation time distributions were extracted using a numerical Laplace transform from CPMG echo series. first peak occurs after 1 ms and moves slightly to lower values (below 1 ms) during the hydration process. The second peak appears after 10 ms and moves slightly to lower values (below 10 ms) during the hydration process. This result indicates that the pores in the samples are becoming smaller due to the hydration reactions. A part of the capillary water participates to the formation of microorganized structure or evaporates.  According to Figure 7, two peaks were observed and were assigned to the presence of two aqueous environments. The first peak, which occurs at the beginning of the hydration phase, is due to the chemically bounded water and occurs when a certain amount of water (flocculant water) combines with the cement grains and forms multiple microorganized systems (cement grain flocculation, chemical reactions, ettringite pores). The second peak is due to the water (capillary water) filling the space between these microstructures [43] or from the surface of the sample [44]. For the samples examined, the first peak occurs after 1 ms and moves slightly to lower values (below 1 ms) during the hydration process. The second peak appears after 10 ms and moves slightly to lower values (below 10 ms) during the hydration process. This result indicates that the pores in the samples are becoming smaller due to the hydration reactions. A part of the capillary water participates to the formation of microorganized structure or evaporates.
The plot of peak area versus hydration time shows an initial decrease in area of the first peak (Figure 8) for all samples examined. CP begins a slow rise from minute 195, while samples containing accelerator show an accelerated increase earlier. The second peak continuously decreases for all samples, indicating a decrease in the amount of free water in the capillary pores. Slower decay can be observed for CP than for acceleratorcontaining samples. These results imply that the hydration process is relatively constant at the beginning of the water-cement interaction, followed by an increase in flocculant water and a decrease in capillary water. This behavior may be due to the participation of capillary water in the process of forming new cement-based structures. peak continuously decreases for all samples, indicating a decrease in the amount of free water in the capillary pores. Slower decay can be observed for CP than for acceleratorcontaining samples. These results imply that the hydration process is relatively constant at the beginning of the water-cement interaction, followed by an increase in flocculant water and a decrease in capillary water. This behavior may be due to the participation of capillary water in the process of forming new cement-based structures. In Figure 9, the maximum point of the T2 relaxation time is plotted for each of the analyzed samples, showing a slow decrease for the CP and CPCaNaAl samples and an accelerated decrease for all other samples. The T2 evolution of CP and CPCaNaAl indicates a slow decrease of capillary water, indicative of a slow change in pore volume, whereas for the other samples analyzed, capillary water decreases rapidly, indicative of a rapid and significant change in pore volume.
Based on the data obtained by NMR analysis, it can be stated that the hydration process of the analyzed samples, or rather the change in pore volume and the formation of organized structures, occurs fastest in CPNa samples, followed by CPCa, CPCaNa, CP2CaNa, CPCaNaAl and CP. In Figure 9, the maximum point of the T2 relaxation time is plotted for each of the analyzed samples, showing a slow decrease for the CP and CPCaNaAl samples and an accelerated decrease for all other samples. The T2 evolution of CP and CPCaNaAl indicates a slow decrease of capillary water, indicative of a slow change in pore volume, whereas for the other samples analyzed, capillary water decreases rapidly, indicative of a rapid and significant change in pore volume.
Based on the data obtained by NMR analysis, it can be stated that the hydration process of the analyzed samples, or rather the change in pore volume and the formation of organized structures, occurs fastest in CPNa samples, followed by CPCa, CPCaNa, CP2CaNa, CPCaNaAl and CP.
The level of the accelerator to be used for this study was determined by multiple tests ranging from 1% to 5% accelerators. In order to preserve the cement paste's properties, the accelerator concentration range was set between 1% and 5%. Further study was conducted on concentrations that formed compact materials at the smallest concentrations.
The chemical analyses were conducted on the obtained materials at an age of 90 days, choosing this age to ensure that the hydration process was complete. Table 5. Mix proportion of the paste and abbreviation of the examined samples.

Sample Preparation
White cement 52.5R produced in Romania (composition: 95-100% clinker and 0-5% other components) and accelerators (Ca(NO 3 ) 2 ·4H 2 O, Al 2 (SO 4 ) 3 ·18H 2 O and Na 2 S 2 O 3 ·5H 2 O) from Merck (Darmstadt, Germany) were used for the preparation of the experimental cement-based materials. The samples were prepared with the same water/cement ratio of 0.4 and different proportions of accelerator (1-3%) according to Table 5. The level of the accelerator to be used for this study was determined by multiple tests ranging from 1% to 5% accelerators. In order to preserve the cement paste's properties, the accelerator concentration range was set between 1% and 5%. Further study was conducted on concentrations that formed compact materials at the smallest concentrations.
The chemical analyses were conducted on the obtained materials at an age of 90 days, choosing this age to ensure that the hydration process was complete.

Experimental
XRD Analysis X-ray diffraction was performed using a Bruker D8 Advance powder diffractometer (Bruker, Karlsruhe, Germany) with a LynxEye XE detector and Cukα1 radiation, λ = 1.54060 Å at room temperature. The following measurement conditions were applied: operating tension 40 kV; current 40 mA; scanning speed 0.02 • /s; scanning range 2θ = 10-80 • . The crystalline phases were identified using Bruker DIFFRAC.SUITE 7.5 software, and crystallinity, and semiquantitative percentages of identified phases were estimated using Bruker DIFFRAC.SUITE EVA 3.0 software.

SEM-EDX Analysis
The sample surface (surface topography, pore size and elemental composition) was examined using scanning electron microscopy (VEGA3 SBU, Tescan, Brno-Kohoutovice, Czech Republic) with energy dispersive X-ray spectroscopy (Quantax EDS, Bruker, Karlsruhe, Germany) SEM/EDX at room temperature. The analysis of the EDX data was carried out with Esprit 2.2 software. Small specimens of~2 mm 2 were used for this study. The magnification level was between 350-420 x. No special treatments such as gold coating were applied to these samples.

Low-Field NMR Relaxometry Analysis
Transverse relaxation measurements were performed with a Bruker MINISPEC MQ20 NMR Analyzer, using the well-known Carr-Purcell-Meiboom-Gill (CPMG) technique [35,36]. The evolution of water consumption in the capillary pores was monitored by an experiment carried out under the following conditions: 1000 echoes were recorded in the CPMG pulse sequence, with an echo time of 0.08 ms, and 32 scans/experiment.

Conclusions
The aim of the study was to obtain and characterize several cement-based materials with white cement and accelerators-Ca(NO 3 ) 2 ·4H 2 O, Al 2 (SO 4 ) 3 ·18H 2 O and Na 2 S 2 O 3 ·5H 2 O-in different concentrations, and to test their potential for usage in 3D printing. The obtained materials were studied using the following techniques: X-ray diffraction, scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM/EDX), and low-field nuclear magnetic resonance relaxometry (LF-NMR).
The X-ray diffraction (XRD) patterns of the obtained materials present crystalline mineral phases characteristic of cement materials (CH, CC, C2S, C3S and Aft). Their content modifies with the type of accelerator used. C3S/C2S ratio was used to observe the hardening speed of the samples and significant increases in the materials containing accelerators compared to Portland cement.
Scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDX) revealed information regarding the surface of the analyzed materials. Significant differences were observed when comparing Portland cement with samples containing accelerators in pore size, the distance between pores, and the spatial distribution of the elements on the surface of the material.
Low-field nuclear magnetic resonance (LF-NMR) relaxometry was used to observe the water kinetics in the obtained materials over 6 h. The results show a decrease in the volume of pores for materials containing accelerators compared to Portland cement and faster hydration of these materials.
The obtained results confirm the potential of the obtained materials for use in 3D printing and suggest that cement paste with Na 2 S 2 O 3 ·5H 2 O as accelerator is the most suitable of them for the intended purpose.