According to the results of XRD analysis (Figure 10), a phase composition of the alkali-activated slag cements without whitening additives is presumably represented by C-S-H gel (d - 0.333; 0.307; 0.296; 0.281; 0.215; 0.187;0.281;0.182 nm), xonotlite (d - 0.3871; 0.3061; 0.2021; 0.193 nm) and calcite (d - 0.383; 0.302; 0.281; 0.227; 0.209; 0.191; 0.187 nm).

The prevailing phases of the white cement with the 5% additive of ТіО₂ by mass are: calcite (d - 0.383; 0.303; 0.281; 0.21; 0.229; 0.209; 191; 0.187; 0.161; 0.152 nm), CaMg(CO₃)₂ (d -0.536; 0.4; 0.362; 0.287; 0.267; 0.218; 0.177 nm) and quartz (d - 0.425; 0.333; 0.182; 0.152 nm), TiO₂ (d - 0.324; 0.248; 0.23; 0.169; 0.162 nm).

A phase composition of hydration products of the white cement with 15% kaolin by mass is presumably represented by C-S-H gel (d - 0.305; 0.28; 0.18 nm), C-(A)-S-H(B) (d - 0.304; 0.28; 0.180 nm), tobermorite (d - 0.333; 0.307; 0.297; 0.286; 0.214; 0.201; 0.182 nm), calcite (d - 0.385; 0.303; 0.249; 0.228; 0.209; 0.192 nm), weak diffractions of analcime (d - 3.43; 2.92; 0.286; 0.272; 2.52; 1.74 nm) and hydronepheline (d - 0.467; 0.435; 0.407; 0.385; 0.369; 0.343; 0.297; 0.283; 0.259 nm). The quantities of analcime and hydronepheline, according to (6) tend to increase constantly in the long-term. Non reacted phases of the slag are represented by ghelenite (d - 0.286; 0.242; 0.228 nm) and quartz (d - 0.424; 0.333; 0.183 nm). Diffractions lines (d - 0.435; 0.417; 0.385; 0.357; 0.309; 0.234 nm) are characteristic of kaolinite.

The cement composition with 24% СаСО₃ by mass is represented by CSH(B) (d - 0.304; 0.28; 0.180nm), presumably tobermorite gel (d - 0.305; 0.279; 0.18 nm), tobermorite (d - 0.33; 0.307; 0.298; 0.285; 0.24; 0.214; 0.201; 0.183 nm), calcite (d - 0.383; 0.303; 0.248; 0.228; 0.209; 0.192; 0.191 nm),and quartz (d - 0.426; 0.334; 0.183; 0.152 nm).

The study of the morphology of the hydration products (see Figure 11) showed that the following types of particles were visualized: plate-like crystals with layered structure, thread (string)-like crystals, spherulites and hexagonal plates, grains and aggregations of irregular shape. Microprobe analysis of elemental composition and the results of XRD analysis showed that the plate-like and thread (string)-like crystals are characteristic of low-basic calcium silicate hydrates; spherulites and hexagonal plates - aluminosilicate hydrates of alkaline and alkaline-alkali-earth composition; aggregation of irregular shape - various gels. The obtained microstructure of the resulted cement stone is homogeneous and without defects.

Acceleration of the binding of the Na⁺-ion in the alkali-activated slag cement pastes helps to prevent risk of efflorescences on its surface during service life. Homogeneity of distribution of the Na⁺-ion from the surface of the cement pastes to its center was evaluated with the help of a scanning electron microscope (SEM) equipped with microprobe analyzer.

Figure 12 shows the alkali-activated slag cement pastes prepared without additives at 28 of curing. The elemental (Na⁺-ion) distribution in the center (19.8% by mass) and on the surface (18.7% by mass) are very similar. This is evidence of the absence of critical mass transfer of Na⁺-ion and its binding in insoluble compounds.

The elemental composition of the alkali-activated slag cement pastes with TiO₂ is shown in Figure 13. With account of the Na₂O-content in the center (20.5% by mass) and on the surface (34.7% by mass). An assumption can be put forward that the whitening this additive (TiO₂) does not prevent migration of free (non-bound) alkali to the surface.

The elemental composition of the alkali-activated slag cement pastes with kaolin is shown in Figure 14. Analysis of the elemental distribution of this paste shows that in the center and on the surface the kaolin affected positively the degree of binding of free alkali with the formation of insoluble aluminosilicate hydrates. Kaolin is known to possess weak acidic properties and for this reason, considerably increases its cation-exchange capacity in an alkaline medium (55). A quantity of Na₂O in the center (22.02% by mass) and on the surface (14.2% by mass) is evidence of the absence of mass transfer of free alkali from the center to the surface. That was confirmed by XRD (see Figure 10) with the formation of aluminosilicate hydrates of analcime and hydronepheline: and also from the studies by SEM analysis (see Figure 11).

The elemental composition of the alkali-activated slag cement pastes with CaCO₃ is shown in Figure 15. Quantities of Na₂О in the center (17.56% by mass) and on the surface (16.64% by mass) in the cement paste composition with СаСО₃ is an evidence of the absence of mass transfer of alkali and its binding with the formation of aluminosilicate hydrate. This correlates well with the data reported in (40) on the deeper hydration processes taking place in the slag-containing cements in the presence of СаСО₃.

Curves of heat release of the white alkali-activated slag cement, depending upon its constituent composition, are represented in Figure 16.

When the cement composition is mixed with water, the heat starts to release almost immediately and its nature is connected with hemosorption processes, dissolution of the slag glass and formation of colloidal and crystalline products. A conclusion can be made that the higher quantities of the whitening additives in the cement composition the lower is a cumulative heat release. Peak of heat released is shifted from 7 or 8 h (the reference cement composition without additive) to 9 or 11 h (the compositions with the additives). This can be attributed, first of all, to the lower contents of the slag in the cement composition. This is especially evident in the compositions with the additives of kaolin and CaCO₃. Accumulative heat release of the composition with the additive of CaCO₃ (55.1 J/g) is higher than that of the composition with the additive of kaolin (44.0 J/g), though a quantity of the additive of CaCO₃ - 24% is much higher compared to that of the additive of kaolin - 15%. This can, evidently, be attributed to the binding or absorption of some portions of alkali by kaolin and, as a result, the lower reaction ability of the liquid medium relatively to the slag.

The influence of optimal quantities of the whitening additives strengthen standard-specified ages and in the long-term are given in Figure 17. As it see in this Figure 17, the white alkali-activated slag cements have the values of compressive strength in the range 49.0-56.8 MPa at 28 days and 66.8-67.5 MPa at 180 days. All these white cements have good strength gain, and, judging by their strength at 2 days (35.0-37.0 MPa), can be considered as quick-hardening cements (class 42.5R or 52.5R, according to EN 197).

Kaolin possess weak acidic properties and for this reason in the conditions of a highly alkaline medium increases considerably its cation-exchange capacity (55). For this reason, it starts actively participate in structure formation processes with the formation of alkaline and alkaline-alkali-earth zeolite-like aluminosilicate hydrate (6), which guarantee the high performance properties of the final hardened cement pastes.

On the contrary, to ТіО₂ and СаСО₃, kaolin acts not only as a whitening additive, but also as active mineral additive capable to affect considerably technology-related, physico-mechanical, and performance properties.

Since a part of alkali is bound by kaolin, strength values of the alkali-activated slag cement pastes at standard-specified ages are somewhat lower (by 10-11%) compared to those of the composition without additive. However, after 3 or 6 months, this difference disappears, and the high strength values are achieved due to the deepening of hydration processes, synthesis of low-basic calcium silicate hydrates, synthesis of analogs to zeolites and feldspars and absence of the processes of destruction.

On contrary to the additive of kaolin, TiO₂ is almost inert with regard to components of the alkali-activated slag cement. А small optimal quantity of this additive (approx. 5% wt.) almost does not affect the compressive strengths (Figure 17). Of interest is the influence of the additive of СаСО₃ on strength values. As it follows from Figure 17, despite large quantities of this additive (approx. 24% wt.) the strength of the alkali-activated slag cement pastes at 28 days was only by 1.5% lower compared to that of the cement composition without additive, and by the age of 180 days this difference practically not exist. This can be attributed to the following: known-in-the-art is that one of the methods of how to increase strength of concrete is to fill a cement matrix with mineral additives - finely dispersed mineral particles of various nature and fractional composition (54, 56, 57).

Finely dispersed carbonate rocks (58, 59) have certain chemical similarity with the alkali-activated slag cement and in size of particles, close to those of particles of the alkali-activated slag cement. Despite some chemical inertness, this circumstance, nevertheless, causes their interaction with hydration products of the alkali-activated slag cement as centers of crystallization or nucleation and enables to form crystal contacts. As a result, a more perfect microstructure is formed (60), thus ensuring the higher performance properties of the concretes and mortars.

Shrinkage deformations were measured on the prismatic mortar specimens 4×4×16 cm (cement : sand = 1:3); at RH of 60% and a temperature of 20±2ºC. Side by side with taking measurements of shrinkage deformations, the values of moisture loss were measured.

The results obtained from measurements of shrinkage of the white alkali-activated slag cement mortars and moisture loss with optimal quantities of the whitening additives are given in Figure 18. A rather good correlation is observed between shrinkage and moisture loss, except for the mortar with the addition of СаСО₃.The lowest values of shrinkage until stabilization were observed in the specimens with the additive of CaCO₃ - 0.54 mm/m. The compositions without additive and with TiO₂have somewhat higher value of shrinkage - 0.55-0.60 mm/m. The highest value of shrinkage (0.77 mm/m) were observed in the compositions with the additive of kaolin (see Figure 18 a). Due to its small quantity (5% by mass) and chemical inertness, the additive of TiO₂ almost does not affect shrinkage. For this reason, the values of shrinkage deformation in the compositions without additive and with TiO₂ are very close.

The effect of kaolin on shrinkage is greater due to its higher content (15% by mass). This can be attributed to the fact that the alkaline cations of a liquid medium enter into interaction followed by peptization and following swelling of kaolin ( 61 ). Later, as soon as physical water leaves the specimen, a reverse process takes place, resulting in the higher value of shrinkage. This correlates well the values of moisture loss of the composition with the additive of kaolin obtained in the study (see Figure 18 b).

As was mentioned earlier, the lowest value of shrinkage deformation are measured in the composition with the additive of CaCO₃, despite its high content in the cement composition - 24% by mass (Figure 18 a) and despite the fact that the values of moisture loss of this composition are among the highest values (Figure 18 b). This can be due to the following: high values of moisture loss are joined with higher content of calcium carbonate (24% by mass) and its finer particles. However, on contrary to kaolin, calcium carbonate is practically inert in an alkaline environment and is not subjected to peptization and following swelling. Besides, as was mentioned earlier, the particles of finely dispersed calcium carbonate can act as centers of crystallization and to form quickly a rigid crystal framework, which restrict shrinkage, and to fill a pore space of the cement stone and to form a denser and more rigid structure. For this reason, in this case, high values of moisture loss do not mean the higher shrinkage deformations.

The results of test of the white alkali-activated slag cement mortars for freeze/thaw resistance are given in Table 6. Mortar specimens 4×4×16 cm (cement : sand = 1:3) after hardening for 28 days curing were tested. The specimens contained 5% mineral pigment by mass. With the consideration of appropriateness, the testing was restricted by150 cycles of freezing/thawing.

After 150 cycles of freezing/thawing (being equivalent to F200) the mass losses of the specimens were absent, the maximum compressive strength decline was 3.87%, scaling of the surface of the specimens was not observed. Thus, according to the results of the test represented in Table 6, all the mortars correspond in freeze/thaw resistance to brand F200.

The results of test of the pigmented alkali-activated slag cement mortar for resistance to wet/dry cycles are given in Table 7. Beam specimens 4×4×16 cm (cement : sand = 1:3) cured in normal conditions for 28 days were tested. The hardened specimens were exposed to alternate drying for 6 h at t = 105 ± 5^оС and wetting (immersion into water) at 20±2^оС during 6 h. The strength was determined after a certain number of cycles. The cements are considered weather resistant, if after 100 cycles of alternate wet/dry, the strength decline of the specimens is less than 25%. As it follows from Table 7, a conclusion can be mortar composition have passed successfully test for weather resistance.

Colour stability was assessed in accordance with the testing procedure prescribed in the National standard of Ukraine DSTU B V.2.7-268:2011 Portland cement, colored. Technical specification (48) (see Figure 19).

Colour stability of the pigmented alkali-activated slag cement was determined on disc-shaped specimens prepared from a cement paste of normal consistency in accordance with the Ukrainian national standard DSTU B V.2.7-181:2009 (43). Two disc-shaped specimens prepared from each composition were stored in the air as reference samples, two disc-shaped specimens were subjected to steam curing, and two disc-shaped specimens - to ultraviolet radiation. Mineral pigments in a quantity of 5% by mass were used.

Irradiation of the disc-shaped specimens by ultraviolet rays is done with the help of a mercury-quartz lamp with a power of 240±20 W for 48 hrs. The disc-shaped specimens were place data distance of 0.5 m from a source of ultraviolet radiation and a luminous flux was directed on the munder the angle of 45±2°.

Colour stability of the pigmented alkali-activated slag cement pastes was evaluated visually by comparing the colors of the disc-shaped specimens subjected to ultraviolet radiation or steam curing with colors of the reference disc-shaped specimens. As it follows from Figure 19, after exposure to ultraviolet radiation or steam curing the color almost did not change.

Risk of efflorescence in the pigmented alkali-activated slag cement mortars was assessed under procedure of DSTU B V.2.7-69-98 (48) on specimens 4×4×16 cm (cement: sand = 1:3). After 28 days of hardening in normal conditions the mortar specimens were dipped a in container filled with distilled water to a depth of 4-5 cm and blowed with air with a temperature of about 25^оС for a least of 3 h per day during 7 days.

A phenomenon of efflorescence on the open surface of the specimen was evaluated visually by the presence of salt deposits. As it follows from Figure 20, no salt deposits were observed.

In addition to standard test for the evaluation of the risk of efflorescence, other tests were also performed in the conditions of indoor and outdoor application. In order to model the exposure of various service conditions of indoor and outdoor application. Mortars (cement : quartz sand = 1:3; W/C = 0.46-0.53) were applied on a walls of the building from the outside to evaluate the exposure of appropriate weather factors in winter-spring-summer periods: (t = -15…+30^оС, snow, rain, sun radiation, R.H. = 70-100%); and from the inside (t = 18-24^оС, R.H. = 55-80%. The test period was 9 months. In order to prevent efflorescence, a water retaining additive (sodium carboxymethyl cellulose), in a quantity of 0.25% by mass, was added in the decorative alkali-activated slag cement mortars. Test results are shown in Table 8 and Table 9.

From Table 8 and Table 9 is deduced that no salt deposits were formed on the decorative mortars tested in indoor and outdoor applications.

The use of red mud as pigment resulted in nice uniform terracotta color. Above all, according to (28, 62), the use of red mud allows to solve ecological problems, to increase strength of the alkali-activated slag cement due to the alkali contained in it, as well as to improve decorative characteristics of the resulted cement stone.

A final consideration from this study: the surface of the decorative mortars based on the white clinker cement usually have an inhomogeneous glossy surface, which is not acceptable from the point of use for decoration purposes, however, the hardened alkali-activated slag cement mortars have an uniform matte surface.

Adherence tests were carried out in accordance with the methodology set out in EN 1542-1999 (50). A cement-sand mortar with a composition of 1:3 was applied to a concrete slab 30×30×10 cm. The concrete class of the slab was B20, and the water absorption of the slab was 6.2%.

After the mortar hardened, the samples were made on a concrete base in the form of cylindrical cores with a diameter of 5 cm in the amount of 5 pieces and were stored under laboratory conditions at a temperature of 20±2°C and relative humidity of the air about 65% for 27 days. The test of samples for separation from the base was carried out on a tensile testing machine FM-250 (Germany). The destruction occurred in the zone of contact of the mortar with the slab or with a partial detachment of the slab of concrete. The test results are presented in Table 10.