**3. Results and Discussion**

The setting times of the compositions based on gypsum binder are presented in Figure 2. As can be seen, the addition of both HEMC and sodium bicarbonate strongly delay the setting times. The delay caused by the HEMC addition could be due to the formation of a polymer film on the surface of gypsum plaster grains which prevents their hydration.

**Figure 2.** Setting times of gypsum binders with 0.3% HEMC (IC) and various amounts of sodium bicarbonate (0.5%, 1% and 2%) and FGD gypsum (1%, 5% and 10%).

The presence of sodium bicarbonate addition also inhibits the hydration of gypsum binder and delays the setting time; these data are in good correlation with those reported by Umponpanarat and Wansom [14].

To mitigate this effect FGD gypsum was added to the binding system. Previous results reported by our research group showed an important decrease of setting time when FGD gypsum is used as addition to gypsum binder [16]. The calcium sulphate dihydrate (CsD) crystals present in FGD gypsum can act as nucleation sites for the new CsD crystals, formed by the hydration of calcium sulphate hemihydrate (CsH) from gypsum plaster, therefore shortening the setting time (Figure 2).

As can be seen in Figure 2, the decrease of sodium bicarbonate (B) dosage from 2% up to 0.5% in the compositions with FGD gypsum (1%, 5% and 10%), decreases the setting times. The increase of FGD gypsum dosage from 5% up to 10% decreased the setting times.

The relative geometrical density and relative compressive strength (calculated with reference to the values obtained for I specimen) for the gypsum binder with 0.3% HMEC and various amounts of sodium bicarbonate and FGD gypsum are presented in Figure 3.

The addition of HEMC to the gypsum binder determines an increase of the porosity (as will be further presented), and so therefore the values of geometrical density and compressive strengths are smaller for IC as compared with I.

The addition of sodium bicarbonate also decreases the geometrical density and compressive strengths in correlation with its dosage (Figure 3); for the highest dosage of sodium bicarbonate (2 wt.%) the compositions IB2 and IG5B2 have lower values of geometrical density (as compared with I), but no recordable compressive strength. The increase of FGD gypsum dosage up to 10 wt.% (IG10B2) determines a small increase of compressive strength values for a similar value of geometrical density.

As expected, the decrease of sodium bicarbonate dosage determines the increase of compressive strengths correlated with higher values of geometrical density. This evolution is in good correlation with the porosity of these materials (as will be presented further).

The microstructure of gypsum pastes hardened for 7 days in air was assessed by scanning electron microscopy (Figures 4–9).

In the SEM images of I specimen one can notice the presence of large round pores with sizes comprised between 0.1–0.25 microns (see arrow in Figure 4a) formed by the air entraining during the mixing operation. In Figure 4b one can notice the presence of smaller pores with various shapes and sizes (see arrows) specifically for the binding matrices formed by the hydration of CsH; the long CsD crystals are interlocked and forms the binding matrix [13,15,28].

For the IC specimen (Figure 5) the presence of HEMC addition determines an important increase of pores numbers and sizes, up to 0.75–1 mm. A close-up on the binding matrix reveals the presence of interlocked needle-like CsD crystals specific for hardened gypsum binder (see arrow in Figure 5c,d). The increase of the number of round closed pores assessed on SEM images (see arrows in Figure 5a) when HEMC is added in the binding system, can be explained by its specific ability to stabilize the entrained air (high affinity for the water-air interface [21,22]).

**Figure 4.** SEM images of gypsum binder (I) after 7 days of hardening in air, with various magnifications: (**a**) ×100; (**b**) ×500.

**Figure 5.** SEM images of gypsum binder with HEMC (IC) after 7 days of hardening in air with various magnifications: (**a**) ×100; (**b**) ×500; (**c**) ×5000; (**d**) ×5000.

For the specimens with sodium bicarbonate an important increase of volume was noticed shortly after the pouring of the paste in the mold. This phenomenon is due to the CO2 generation in the reaction of sodium bicarbonate with water and calcium sulphate hemihydrate (CsH) [13]. This process is more intense for a higher dosage of sodium bicarbonate and the resulted porosity is higher, as can be noticed for the SEM images presented in Figures 6–9.

For the gypsum plaster with 2% sodium bicarbonate (IB2), one may assess in the SEM images the presence of big pores (over 1 mm) formed in the binding matrix by the released CO2 gas (Figure 6a); the irregular shapes of these pores suggest the coalescence of smaller pores.

It is interesting to note the shape of crystals in IB2 i.e., short thick crystals and plate like crystals (Figure 6c,d). The modification of CsD crystal shape and size can be due to the modification of reaction conditions in the presence of sodium bicarbonate addition [28] or/and to an oriented growth of crystals due to selective adsorption of retarding addition [24].

In the SEM images of plaster with 5% FGD gypsum and 2% sodium bicarbonate (IG5B2) one may notice the presence of numerous round pores with sizes comprised between 0.25–1 mm (Figure 7a) as well as smaller pores (10–20 microns) formed between the CsD crystals (Figure 7b,d). Short and plate-like CsH crystals also present in this composition (Figure 7c,d).

As expected, the reduction of sodium bicarbonate content reduces the average size of pores assessed on SEM images (Figures 8a and 9a). The CsD crystals continues to be shorter with an average size of 10–20 microns. The interlocking of gypsum crystals increases (Figure 9b,c) and this contributes to the increase of mechanical strengths.

Rubio Avalos et al. [13] reported the presence of Na2SO4 as secondary phase in the CaSO4·0.5H2O–NaHCO3–H2O system; according to these authors, the sodium sulphate (small round crystals) precipitates inside the gypsum crystals bulk. This phase was not detected in this study by XRD in the specimens with 0.5% and 1% sodium bicarbonate (Figure 10); this could be due to the low dosage of sodium bicarbonate in these compositions.

Nevertheless, the EDX analysis presented in Figure 11 shows the presence of sodium in high quantity in some specific areas, which could be associated to a phase with sodium content.

The European norm EN 13279-1 sets the requirements for gypsum binders and plasters; these requirements refer to the flexural strength (higher than 1 N/mm2) and compressive strength (higher than 2 N/mm2) [29].

**Figure 6.** *Cont.*

**Figure 6.** SEM images of gypsum plaster with 2% sodium bicarbonate addition (IB2) with various magnifications: (**a**) ×100; (**b**) ×500; (**c**) ×5000; (**d**) ×5000.

**Figure 7.** SEM images of gypsum plaster with 5% FGD and 2% sodium bicarbonate addition (IG5B2) with various magnifications: (**a**) ×100; (**b**) ×500; (**c**) ×1000; (**d**) ×5000.

**Figure 8.** SEM images of gypsum binder with 1% FGD and 1% sodium bicarbonate addition (IG1B1) after 7 days of hardening in air with various magnifications: (**a**) ×200; (**b**) ×2000; (**c**) ×5000; (**d**) ×5000.

The compositions IC and IG1B1 fulfill the above-mentioned requirements (Table 2), and therefore thermal conductivity was assessed on these specimens. The thermal conductivity at 10 ◦C (set I), assessed in accordance with the norm EN ISO 10456 [30], is the thermal conductivity value usually declared by the European producers of this type of construction material based on the fact that 10 ◦C is considered the average yearly temperature at which the thermal insulation of buildings must operate.

As expected, the decrease of the geometrical density and increase of open porosity of IC and IG1B1, due to the presence of hydroxyethyl methyl cellulose (HEMC) and sodium bicarbonate, improves the thermal properties of these materials (i.e., reduces the values of thermal conductivity at 10 ◦C).

**Table 2.** Geometrical density, open porosity, flexural and compressive strengths and thermal conductivity of gypsum-based materials.


**Figure 9.** SEM images of gypsum binder with 1% FGD and 0.5% sodium bicarbonate addition (IG1B0.5) after 7 days of hardening in air, with various magnifications: (**a**) ×100; (**b**) ×500; (**c**) ×10000; (**d**) ×2000.

**Figure 10.** XRD patterns of gypsum binder I—anhydrous and I and IG1B1 hardened pastes.

**Figure 11.** SEM and EDX analyses of IG1B1 paste.

In conformity with European norm EN 13279-1 [29], gypsum plasters or gypsum binders are classified as reaction to fire Class A1 when they contain less than 1% by weight or volume of organic materials, without supplementary testing. The only material presented in this paper which contains organic material is the one with hydroxyethyl methyl cellulose but its amount (0.3%) is below the previously mentioned limit; therefore, the gypsum based materials obtained in this study can be classified as Class A1 (reaction to fire).

Based on the properties assessed for the studied gypsum-based materials a potential practical application could be for the manufacture of light gypsum blocks/boards for non-load-bearing walls with improved thermal insulation properties and good fire behavior.
