Marble Waste Valorization through Alkali Activation

: In the present study, the valorization potential of marble waste in the presence of metakaolin via alkali activation was explored. The activating solution used consisted of NaOH and sodium silicate solutions. The effects of marble waste to metakaolin ratio, particle size of raw materials, curing temperature, and Na 2 O/SiO 2 and H 2 O/Na 2 O molar ratios present in the activating solution on the main properties and the morphology of the produced alkali-activated materials (AAMs) was evaluated. The durability and structural integrity of the AAMs after ﬁring at temperatures between 200 and 600 ◦ C, immersion in deionized water and 1 mol/L NaCl solution for different time periods and subjection to freeze–thaw cycles were also investigated. Characterization techniques including Fourier transform infrared spectroscopy, X-ray diffraction, mercury intrusion porosimetry and scanning electron microscopy were used in order to study the structure of the produced AAMs. The highest compressive strength (~36 MPa) was achieved by the AAMs prepared with marble waste to metakaolin mass ratio of 0.3 after curing at 40 ◦ C. The results indicated that the utilization of marble waste in the presence of metakaolin enables the production of AAMs with good physical (porosity, density and water absorption) and mechanical properties, thus contributing to the valorization of this waste type and the reduction of the environmental footprint of the marble industry.


Introduction
Marbles are natural stones which are formed as a result of recrystallization of limestone and dolomite under high temperature and pressure [1]. They are used since ancient times for the construction of cultural heritage monuments, historical buildings and numerous other applications. Their economic value depends both on their morphological and structural characteristics. Worldwide, there exist various types and significant reserves of marbles which they account for about 50% of the total natural stones production [2,3]. Due to the processing requirements and environmental limitations, only 10% of the quarrying material is used in the market [4]. Cutting and polishing processes in marble plants result in the production of two types of waste, namely marble waste slurry and marble waste powder. Thus far, the ornamental stone industry has not developed an efficient valorization process for these waste types, therefore they are either stored in warehouses for future use in other applications, or disposed of in landfills, causing environmental problems [5,6].
Marble waste can be used as replacement of aggregates or additive in concrete production [7], as well as raw material for up to 30 wt. % substitution of red clay during the production of bricks for the improvement of their properties [8,9]. Thus, the higher use of marble waste in construction applications, via a technically and economically feasible technology, can result in significant economic and environmental benefits for the quarrying industry [10].
Alkali activation is a promising technology for the valorization of various waste types and the production of secondary materials, termed alkali-activated materials (AAMs),  50 17.9 14.5 8.1 8.8 d 10 1.1 1.0 0. 6 2.1 The chemical composition of the raw materials was determined with the use of an X-ray fluorescence energy dispersive spectrometer (XRF-EDS) S2 Range type (Bruker, Karlsruhe, Germany) and is shown in Table 2. As seen from these data, metakaolin may act as source of silicon and aluminum (55.9 wt. % SiO 2 and 38.5 wt. % Al 2 O 3 ), elements which are required for alkali activation, whereas marble waste has, as anticipated, high content of CaO (53.1 wt. %). The activating solution consisted of sodium silicate solution (Na 2 O = 7.5-8.5 wt. %, SiO 2 = 25.5-28.5 wt. %, Merck, Germany) and NaOH solution (6-10 mol/L). The NaOH solution was produced by dissolving anhydrous pellets of NaOH in distilled water so that the desired molarity was obtained. Table 3 shows the mixing proportions of the raw materials and reagents used as well as the SiO 2 /Al 2 O 3 molar ratio in the starting mixture, the Na 2 O/SiO 2 and H 2 O/Na 2 O molar ratios in the activating solution and the overall liquid/solid (L/S) ratio. Specific proportions of marble waste and metakaolin were used in order to obtain suitable SiO 2 /Al 2 O 3 ratios in the precursors, based on the results of earlier studies carried out in our lab as well as by using literature data. [11,23]. The resulting paste was obtained after mechanical mixing of the raw materials with the activating solution for 10 min. The activating solution was prepared by mixing the solutions used and was left overnight prior to use. As seen in Table 3, three different H 2 O/Na 2 O molar ratios were used in the experimental series in order to obtain a slurry with suitable fluidity and workability, while at the same time maintaining in all cases the lowest possible L/S ratio, which in our case varies slightly between 0.49 and 0.54 [24]. It is mentioned that in other earlier studies carried out in our labs with the use of other wastes, e.g., slags and construction and demolition wastes, this ratio was much lower, i.e., 0.20-0.30 [12,17]. The produced paste was poured in steel cubic molds (5 × 5 × 5 cm 3 ) and remained at room temperature for either 2 h (metakaolin-based (MK) specimens) or 4 h (marble wastemetakaolin (MW/MK) specimens) to facilitate sufficient initial setting and hardening. The specimens were then demolded, placed in sealed plastic bags and cured at 40, 60 and 90 • C for 24 h in a laboratory oven ON-02G (Jeio Tech, Billerica, MA, USA). After curing, the specimens were removed from the oven, aged at room temperature for 7 days and their compressive strength was determined using a Matest compression machine with dual range 500/15 kN (Matest, Bergamo, Italy). Measurements were carried out in triplicate and in the paper mean values are given. It is mentioned that deviation of measurements was marginal and did not exceed ±3.5 %. The mineralogy, structure and morphology of selected AAMs was determined by X-ray diffraction (XRD), Fourier Transform Infrared Spectroscopy (FTIR), Mercury Intrusion Porosimetry (MIP) and Scanning Electron Microscopy (SEM). XRD analysis was performed using an X-ray diffractometer (D8-Advance, Bruker AXS, Bruker, Karlsruhe, Germany) with a Cu tube, a scanning range from 4 • to 70 • 2-theta (θ) with a step 0.02 • and a measuring time of 0.2 s/step. Qualitative analysis was carried out using the DiffracPlus Software (Bruker AXS, Bruker, Karlsruhe, Germany) and the PDF database. Fourier Transform Infrared Spectroscopy (FTIR) was performed using KBr pellets with a PerkinElmer 1000 spectrometer (PerkinElmer, Akron, OH, USA). MIP analysis was performed using a Micromeritics AutoPore 9400 porosimeter (Micromeritics, Atlanta, GA, USA). Scanning Electron Microscopy (SEM) was carried out using a JEOL 6380LV (JEOL, Tokyo, Japan) microscope equipped with an Oxford INCA energy dispersive X-ray spectrometer (EDS) microanalysis system (Oxford Instruments, Abingdon, UK). Prior to SEM analysis, AAM specimens were stored in an anaerobic glove box and then embedded in an epoxy resin (EpoFix, Struers, Ballerup, Denmark). The epoxy-embedded samples were ground and polished using diamond abrasives. The apparent density and the water absorption of selected AAMs were also determined according to BS EN 1936 and BS EN 13755:2008 [26] standards, respectively.
In order to study their structural integrity, selected AAMs were (i) subjected to firing at 200, 400 and 600 • C for 2 h, in a laboratory furnace N-8L Selecta (JP Selecta, Abrera, Spain), (ii) immersed in deionized water and 1 mol/L (M) NaCl solution for 7 and 30 days and (iii) subjected to 7 and 15 freeze-thaw cycles, according to ASTM 1262-10 [27]. AAMs were subjected to freezing at minus 17 • C for 4 h and thawing at room temperature for 18 h (one complete cycle). After each test, the weight loss and the compressive strength of the specimens were determined. Figure 1 illustrates the compressive strength of the AAMs produced after curing of the reactive paste at 90 • C for 7 days, using different marble waste (median particle size 17.9 µm) to metakaolin mass ratios (MW/MK 0.3, 0.7 and 1.5 AAMs) as a function of H 2 O/Na 2 O molar ratio present in the activating solution (16.0, 18.5 and 22.0). The compressive strength of MK and MW reference AAMs is also given for comparison. 17.9 μm) to metakaolin mass ratios (MW/MK 0.3, 0.7 and 1.5 AAMs) as a function of H2O/Na2O molar ratio present in the activating solution (16.0, 18.5 and 22.0). The compressive strength of MK and MW reference AAMs is also given for comparison. As seen in Figure 1, the MW reference specimens exhibited very low compressive strength for all the H2O/Na2O molar ratios used, indicating the poor alkali activation potential of marble waste, due to its extremely low content of silica and alumina, as shown in Table 2. On the other hand, the MK reference specimens exhibited much higher compressive strength, which in all cases exceeded 25 MPa and reached in the optimum synthesis conditions 37.3 MPa. This is the reason why marble waste was mixed with metakaolin which due to its high content of silica and alumina possesses a much higher inherent alkali activation potential; the beneficial behavior of metakaolin during alkali activation has been underlined by many researchers who investigated mixtures of metakaolin and various municipal and industrial wastes for the production of AAMs with good mechanical and physico-chemical properties [28][29][30][31][32][33][34].

Effect of Marble Waste to Metakaolin Ratio
The specimens produced after alkali activation of the mixtures containing marble waste and metakaolin acquired quite high compressive strength. The highest value (30.1 MPa) was acquired by the specimens produced with the use of marble waste to metakaolin ratio 0.3 (MW/MK 0.3 AAM) and H2O/Na2O molar ratio 16. Relatively high compressive strength values, namely 22.3 MPa and 15.2 MPa, were also obtained when the same H2O/Na2O molar was used and the marble waste to metakaolin ratio in the starting mixture increased to 0.7 and 1.5, respectively. The decrease of the compressive strength values when higher marble waste to metakaolin mass ratios were used is due to the inherent poor alkali activation potential of marble waste, as mentioned earlier in this paper.
It is also seen in Figure 1 that the increase of H2O/Na2O molar ratio from 16 to 22, resulted in a noticeable decrease of the compressive strength for all produced AAMs. Lower H2O/Na2O ratios denote that stronger activating solution is used which has the ability to provide the required alkalinity in the produced reactive paste that results in sufficient dissolution of Si and Al from the precursors and the formation of stronger Si-O-Al bonds [35,36].

Effect of Curing Temperature
The effect of curing temperature (40, 60 and 90 °C) on the compressive strength of the AAMs produced using marble waste (median particle size 17.9 μm) to metakaolin mass As seen in Figure 1, the MW reference specimens exhibited very low compressive strength for all the H 2 O/Na 2 O molar ratios used, indicating the poor alkali activation potential of marble waste, due to its extremely low content of silica and alumina, as shown in Table 2. On the other hand, the MK reference specimens exhibited much higher compressive strength, which in all cases exceeded 25 MPa and reached in the optimum synthesis conditions 37.3 MPa. This is the reason why marble waste was mixed with metakaolin which due to its high content of silica and alumina possesses a much higher inherent alkali activation potential; the beneficial behavior of metakaolin during alkali activation has been underlined by many researchers who investigated mixtures of metakaolin and various municipal and industrial wastes for the production of AAMs with good mechanical and physico-chemical properties [28][29][30][31][32][33][34].
The specimens produced after alkali activation of the mixtures containing marble waste and metakaolin acquired quite high compressive strength. The highest value (30.1 MPa) was acquired by the specimens produced with the use of marble waste to metakaolin ratio 0.3 (MW/MK 0.3 AAM) and H 2 O/Na 2 O molar ratio 16. Relatively high compressive strength values, namely 22.3 MPa and 15.2 MPa, were also obtained when the same H 2 O/Na 2 O molar was used and the marble waste to metakaolin ratio in the starting mixture increased to 0.7 and 1.5, respectively. The decrease of the compressive strength values when higher marble waste to metakaolin mass ratios were used is due to the inherent poor alkali activation potential of marble waste, as mentioned earlier in this paper.
It is also seen in Figure 1 that the increase of H 2 O/Na 2 O molar ratio from 16 to 22, resulted in a noticeable decrease of the compressive strength for all produced AAMs. Lower H 2 O/Na 2 O ratios denote that stronger activating solution is used which has the ability to provide the required alkalinity in the produced reactive paste that results in sufficient dissolution of Si and Al from the precursors and the formation of stronger Si-O-Al bonds [35,36].

Effect of Curing Temperature
The effect of curing temperature (40, 60 and 90 • C) on the compressive strength of the AAMs produced using marble waste (median particle size 17.9 µm) to metakaolin mass ratio 0.3 and H 2 O/Na 2 O molar ratio in the activating solution 16, is illustrated in Figure 2. The compressive strength of MK and MW reference AAMs is given again for comparison.
Minerals 2021, 11, x FOR PEER REVIEW 6 of 17 ratio 0.3 and H2O/Na2O molar ratio in the activating solution 16, is illustrated in Figure 2.
The compressive strength of MK and MW reference AAMs is given again for comparison.

Figure 2.
Effect of curing temperature on the 7-day compressive strength of alkali-activated materials (AAMs) produced with the use of marble waste to metakaolin mass ratio 0.3 and H2O/Na2O molar ratio in the activating solution 16 (median MW particle size 17.9 μm).
As seen from the data, the effect of the curing temperature over the range investigated was marginal and thus its increase from 40 to 90 °C had only a slightly negative effect on the compressive strength of the produced AAMs. The highest compressive strength (35.7 MPa) was acquired by the AAMs produced after curing at 40 °C. Although in most cases curing at elevated temperatures is known to accelerate the rate of the alkali activation reactions [15,37], in this case the process was carried out effectively even at the low temperature of 40 °C ; this was mainly due to the presence of sufficient metakaolin, which has significant alkali activation potential at low temperature, as also indicated in previous studies [38][39][40][41]. It is thus underlined that the use of low temperature during curing is beneficial since it results in significant economic and environmental benefits, which are associated with lower energy cost and lower emissions due to reduced energy consumption. Figure 3 illustrates the effect of median particle size of marble waste (8.1, 14.5 or 17.9 μm) on the compressive strength of AAMs produced using marble waste to metakaolin mass ratio 0.3 (MW/MK 0.3 AAM), H2O/Na2O molar ratio 16, and curing temperature 40 °C.

Effect of Marble Waste Particle Size on the Compressive Strength
It is seen from these data that the effect of particle size of marble waste on the compressive strength of the produced AAMs, which varied between 35 and 38 MPa, was marginal. It is known from earlier studies that in most cases the decrease in particle size accelerates the rate of alkali activated reactions and results in densification of the matrix of the produced specimens and thus in noticeable increase of the compressive strength. This is due to the increase of the specific surface area of the raw materials and the subsequent increase of their reactivity when attacked by alkali activating solutions [22,38,42,43]. However, in our case the close range of the small particle size of both precursors used, which for the marble waste varied between 8.1 and 17.9 µ m, did not affect alkali activation and resulted in the production of AAMs with high compressive strength. As seen from the data, the effect of the curing temperature over the range investigated was marginal and thus its increase from 40 to 90 • C had only a slightly negative effect on the compressive strength of the produced AAMs. The highest compressive strength (35.7 MPa) was acquired by the AAMs produced after curing at 40 • C. Although in most cases curing at elevated temperatures is known to accelerate the rate of the alkali activation reactions [15,37], in this case the process was carried out effectively even at the low temperature of 40 • C; this was mainly due to the presence of sufficient metakaolin, which has significant alkali activation potential at low temperature, as also indicated in previous studies [38][39][40][41]. It is thus underlined that the use of low temperature during curing is beneficial since it results in significant economic and environmental benefits, which are associated with lower energy cost and lower emissions due to reduced energy consumption.  Table 4 presents the physical properties of selected AAMs prepared from marble waste, metakaolin and their mixture with a marble waste to metakaolin mass ratio 0.3, using activating solution with H2O/Na2O molar ratio 16, after curing at 40 °C for 24 h. It It is seen from these data that the effect of particle size of marble waste on the compressive strength of the produced AAMs, which varied between 35 and 38 MPa, was marginal. It is known from earlier studies that in most cases the decrease in particle size accelerates the rate of alkali activated reactions and results in densification of the matrix of the produced specimens and thus in noticeable increase of the compressive strength. This is due to the increase of the specific surface area of the raw materials and the subsequent increase of their reactivity when attacked by alkali activating solutions [22,38,42,43]. However, in our case the close range of the small particle size of both precursors used, which for the marble waste varied between 8.1 and 17.9 µm, did not affect alkali activation and resulted in the production of AAMs with high compressive strength. Table 4 presents the physical properties of selected AAMs prepared from marble waste, metakaolin and their mixture with a marble waste to metakaolin mass ratio 0.3, using activating solution with H 2 O/Na 2 O molar ratio 16, after curing at 40 • C for 24 h. It can be seen from these results that mixing of marble waste with metakaolin has beneficial effect on the properties of the produced AAMs (MW/MK 0.3) when compared to the MW AAM. The porosity and the water absorption dropped by almost 35% and 50%, compared to the respective values of the reference MW AAMs. On the other hand, their density increased by almost 10% to 1611 kg/m 3 , due to the formation of a dense structure with compressive strength exceeding 35 MPa. Furthermore, the porosity and the water absorption of the MW/MK 0.3 AAM increased by 22% and 26%, while their density decreased by 10% compared to the respective values of the reference MK AAMs due to the presence of marble powder in the starting mixture.  Table 5 compares the results obtained from this study with those derived from other selected studies investigating the alkali activation of raw material mixtures containing ornamental stone industry waste and metakaolin.

Physical Properties of the Produced AAMs
Colangelo et al. [19] produced thermal insulating materials through alkali activation of mixtures containing marble waste, metakaolin and organic additives. Their maximum compressive strength reached was 6 MPa, which is considered sufficient for such materials. Clausi et al. [24], prepared AAMs using mixtures of ornamental stones and metakaolin. The produced specimens, after curing for 28 days at 20 • C and 90% relative humidity, acquired very high compressive strength values (72 MPa). The H 2 O/Na 2 O molar ratio used in the activating solution was 10, lower than the one used in the present study. Simão et al. [21] used mixtures of marble waste, volcanic stones cutting waste and metakaolin as raw materials to produce inorganic polymers, which after curing at ambient conditions acquired compressive strength of 28.6 MPa. The H 2 O/Na 2 O molar ratio in the activating solution was 13, quite similar to the lower one used in the present study.
It is seen from these data that the AAMs produced in the present study acquire compressive strength which is comparable with most values determined in other studies; one noticeable exception is the production of AAMs after alkali activation of ornamental stones which as raw material have more suitable mineralogical composition and thus much higher alkali activation potential compared to marble waste.  Figure 4.  The spectrum of raw metakaolin is shown in Figure 5a. The peaks seen at 452 cm −1 and 790 cm −1 are related to Al-O-Si bonds, while the broad band illustrated between 1180-1060 cm −1 denotes its aluminosilicate matrix [36]. As seen in the spectrum of MK AAM (Figure 5c), the peak illustrated at 790 cm −1 in the raw material almost disappeared. This is attributed to the development of the aluminosilicate network due to alkali activation [39]. The peak seen at 1012 cm −1 , also shown in the MW/MK 0.3 AAM (Figure 5e), corresponds to Si-O-Si and Al-O-Si asymmetrical stretching vibrations due to the SiO4 and AlO4  Figure 4e) contain crystalline phases which exhibit lower intensity compared to the respective ones present in the raw materials. Overall, it can be seen that the peak intensities of the crystalline phases of calcite, dolomite, muscovite and quartz present in the produced AAMs decreased or even disappeared compared to the ones present in the raw materials (waste marble and metakaolin). The decreasing intensities of these phases which identified in all AAMs (MK, MW and MW/MK) indicate their dissolution in the polymeric matrix due to the attack of the alkali activator and the subsequent formation of aluminosilicate amorphous gels, as shown by the following SEM-EDS analysis.  (Figure 6a,b) revealed an heterogeneous in size microstructure that is dominated by large calcite crystals (>30 μm long) along with smaller crystals (<15 μm) of dolomite. In higher magnification (Figure 6b), it is shown that the microstructure of the MW AAM is loose and several undissolved/un-hydrated particles of various sizes and sharp edges as well as large voids (black gaps) are detected; this indicates their limited reaction during alkali activation due to the low reactivity of marble waste precursor, as results from its low inherent content of silica and alumina. The large number of voids and un-reacted particles indicates poor structural integrity that negatively affects the compressive strength of the produced AAM (3.9 MPa). The spectrum of raw metakaolin is shown in Figure 5a. The peaks seen at 452 cm −1 and 790 cm −1 are related to Al-O-Si bonds, while the broad band illustrated between 1180-1060 cm −1 denotes its aluminosilicate matrix [36]. As seen in the spectrum of MK AAM (Figure 5c), the peak illustrated at 790 cm −1 in the raw material almost disappeared. This is attributed to the development of the aluminosilicate network due to alkali activation [39]. The peak seen at 1012 cm −1 , also shown in the MW/MK 0.3 AAM (Figure 5e), corresponds to Si-O-Si and Al-O-Si asymmetrical stretching vibrations due to the SiO 4 and AlO 4 reorganization that took place during alkali activation [37,44].

FTIR Analysis
The spectra of raw marble waste (Figure 5b) and MW AAM (Figure 5d) did not show any considerable changes, indicating that marble waste is poorly alkali activated. The bands at 712, 872 and 1423 cm −1 , illustrated in both spectra represent CO 3 vibrations due to the presence of CaCO 3 [45,46]. In the spectrum of the MW/MK 0.3 AAM (Figure 5e) the peaks at 1414 and 1490 cm −1 are due to atmospheric carbonation [47]. The band seen at 1500 cm −1 corresponds to H-O-H vibrations of absorbed water, while the peak at 1598 cm −1 is due to the presence of chemically bonded water [48]. Figure 6 shows the SEM images of selected AAMs (MW, MW/MK 0.3 and MK) produced under the optimum conditions (H 2 O/Na 2 O molar ratio 16, curing at 40 • C for 24 h). SEM examination of MW AAM surface and microanalysis by EDS (Figure 6a,b) revealed an heterogeneous in size microstructure that is dominated by large calcite crystals (>30 µm long) along with smaller crystals (<15 µm) of dolomite. In higher magnification (Figure 6b), it is shown that the microstructure of the MW AAM is loose and several undissolved/unhydrated particles of various sizes and sharp edges as well as large voids (black gaps) are detected; this indicates their limited reaction during alkali activation due to the low reactivity of marble waste precursor, as results from its low inherent content of silica and alumina. The large number of voids and un-reacted particles indicates poor structural integrity that negatively affects the compressive strength of the produced AAM (3.9 MPa). On the other hand, the microstructure of the MW/MK 0.3 and MK AAMs is mainly characterized by smaller particles and the formation of a more homogenous polymeric On the other hand, the microstructure of the MW/MK 0.3 and MK AAMs is mainly characterized by smaller particles and the formation of a more homogenous polymeric network with less visible cracks compared to MW AAM (Figure 6c,d and e, respectively). However, the MK AAM is characterized by a more homogeneous and compact microstructure due to the formation of a denser matrix in which very few unreactive particles are seen; this is due to the presence of metakaolin and the resulting higher degree of polymerization (Figure 6e). Figure 6d (zoom of rectangular area of Figure 6c) shows an improved pore refinement in the entire matrix of the MW/MK 0.3 AAM compared to MW AAM, due to the addition of metakaolin as precursor in the starting mixture, as well as fewer smaller in size unreactive particles that exhibit better cohesion with the precursors; this dense and compact microstructure justifies the high compressive strength obtained by the MW/MK 0.3 AAM (35.7 MPa) which is slightly lower in comparison to MK AAM (40.3. MPa) ( Table 3).

SEM Analysis
The well-formed microstructure of the MW/MK 0.3 AAM was further evaluated using quantitative EDS analysis in order to elucidate the formation of the hydrated products in the polymeric matrix. SEM-EDS analyses confirmed the excessive formation of two types of gels i.e., a needle-like N-A-S-H (G 1 ) and a continuous flake-like (N,C)-A-S-H (G 2 ) shown in Figure 6d. More specifically, EDS analyses indicate the formation of a N-A-S-H gel rich in Na and Al and poor in Ca with Al/Si, Na/Si and Ca/Si ratios equal to 0.38, 0.82 and 0.14, respectively. As the polycondensation reactions progress, Ca ions dissolve from the Ca-rich marble waste and replace Na ions in the N-A-S-H gel via an ion exchange process that consequently result in the formation of the solid (N,C)-A-S-H gel [49]. Although both gels are characterized by similar Al/Si ratio (0.35) due to the addition of metakaolin in the starting mixture, small amount of Na and big amount of Ca incorporated in the (N,C)-A-S-H gel, as indicated by the corresponding Na/Si (0.21) and Ca/Si (1.6) ratios, respectively. The co-existence of these two types of gels in the structure of the MW/MK 0.3 AAM is the major strength contributor and justifies its good mechanical properties [5,50].
In other related studies, it has been mentioned that during alkali activation some of the unreacted CaCO 3 , which is the major constituent of marble waste, may induce a micro-filler effect in the AAMs produced when marble waste and aluminosilicate minerals are used as precursors [51]. On the other hand, it has been mentioned that CaCO 3 exhibits rather reduced solubility in alkaline solutions [52]. However, some precipitation of C-A-S-H and N-A-S-H phases may take place during alkali activation and result in the increase of the compressive strength of the produced specimens [53,54]. In another recent study it was shown that the incorporation of cotton and viscon fibers improved the microstructure and properties of geopolymer concrete produced from zeolitic tuff and marble waste [55]. Therefore, the final microstructure of the alkali activated materials may be the result of the synergistic action of the filler effect and the precipitation of calcium bearing phases.

Structural Integrity
The durability of the MW/MK 0.3 AAM produced under the optimum conditions (marble waste to metakaolin mass ratio 0.3, H 2 O/Na 2 O molar ratio 16, curing at 40 • C for 24 h) was tested after exposure in different aggressive environments, namely firing at 200, 400 and 600 • C for 2 h, immersion in deionized water and 1 mol/L NaCl solution for 7 and 30 days, and also implementation of freeze-thaw cycles according to ASTM C1262-10. Table 6 shows the evolution of compressive strength and weight loss of the MW/MK 0.3 AAM after firing at 200, 400 and 600 • C for 2 h. It is seen from these data that the AAM lost a substantial percentage of its compressive strength, even after firing at 200 • C which is considered a low temperature, while after firing at temperatures higher than 400 • C it almost disintegrated. The weight loss increased, in accordance with the loss of compressive strength, from 24.6% at 200 • C to 33.5% at 600 • C. The detrimental effect of high-temperature firing on the integrity of the AAM is attributed to the decomposition of aluminosilicate bonds, the loss of water as well as the development and subsequent propagation of cracks. The behavior of the AAM after firing, as determined in this study, was worse compared to the behavior of AAMs produced from other waste types, including slags or construction and demolition wastes [56,57].  Figure 7 shows the evolution of the compressive strength and weight loss of the MW/MK 0.3 AAM after immersion in deionized water and 1 mol/L NaCl solution for 7 and 30 days.  Figure 7 shows the evolution of the compressive strength and weight loss of the MW/MK 0.3 ΑΑΜ after immersion in deionized water and 1 mol/L NaCl solution for 7 and 30 days.
As seen in Figure 7, the MW/MK 0.3 AAM after immersion in 1 mol/L NaCl solution for 30 days exhibited the highest compressive strength loss (58%) due to the disintegration of the Si-O-Al-Si bonds [56,58]. On the other hand, after immersion in deionized water for the same period it retained an acceptable compressive strength level (24.3 MPa). Immersion for a period of 7 days in both deionized water and NaCl solution had a milder impact on the compressive strength and the determined values varied between 25 and 28 MPa. As also anticipated, immersion in 1 mol/L NaCl solution over a period of 30 days resulted in the highest weight loss (12%), while immersion in deionized water for the same period caused lower weight loss (8%).  Figure 8 shows the behavior of the MW/MK 0.3 after subjection to freeze-thaw cycles according to ASTM C1262-10 standard. It can be seen from these data that the AAM exhibited very low resistance, by taking into account that the compressive strength loss after 7 freeze-thaw cycles was close to 40%, while after 15 cycles it increased to values higher than 60%. It is mentioned that the weight loss of the MW/MK 0.3 AAM after 7 and 15 freeze-thaw cycles was also high, namely 17 and 39%, respectively.
The poor performance of the AAM is due to its properties (Table 4) and particularly its high porosity (25.1%) and water absorption (14.2%). The main deterioration mechanism is the freezing of the water absorbed by the capillaries which results in the development and subsequent expansion of cracks [59]. The behavior of the AAM after subjection to more than 15 freeze-thaw cycles could not be evaluated, since its structural integrity was severely affected (Figure 9). As seen in Figure 7, the MW/MK 0.3 AAM after immersion in 1 mol/L NaCl solution for 30 days exhibited the highest compressive strength loss (58%) due to the disintegration of the Si-O-Al-Si bonds [56,58]. On the other hand, after immersion in deionized water for the same period it retained an acceptable compressive strength level (24.3 MPa). Immersion for a period of 7 days in both deionized water and NaCl solution had a milder impact on the compressive strength and the determined values varied between 25 and 28 MPa. As also anticipated, immersion in 1 mol/L NaCl solution over a period of 30 days resulted in the highest weight loss (12%), while immersion in deionized water for the same period caused lower weight loss (8%). Figure 8 shows the behavior of the MW/MK 0.3 after subjection to freeze-thaw cycles according to ASTM C1262-10 standard. It can be seen from these data that the AAM exhibited very low resistance, by taking into account that the compressive strength loss after 7 freeze-thaw cycles was close to 40%, while after 15 cycles it increased to values higher than 60%. It is mentioned that the weight loss of the MW/MK 0.3 AAM after 7 and 15 freeze-thaw cycles was also high, namely 17 and 39%, respectively.
The poor performance of the AAM is due to its properties (Table 4) and particularly its high porosity (25.1%) and water absorption (14.2%). The main deterioration mechanism is the freezing of the water absorbed by the capillaries which results in the development and subsequent expansion of cracks [59]. The behavior of the AAM after subjection to more than 15 freeze-thaw cycles could not be evaluated, since its structural integrity was severely affected (Figure 9).  Based on the evaluation of the results of the present study, it is deduced that the best AAMs obtained, e.g., MW/MK 0.3, may be used as alternative binders which exhibit good mechanical and physico-chemical properties which are similar to those of ordinary Portland cement. Thus, the production of alkali activated materials from marble wastes amended with metakaolin offers significant environmental and economic advantages and results in the valorization of this large waste stream which is usually disposed to the environment and causes several impacts [14,60].

Conclusions
This study showed that marble waste can be valorized, with the addition of metakaolin, via alkali activation for the production of AAMs with beneficial physical and mechanical properties. The highest compressive strength was obtained by the AAMs produced when marble waste was mixed with metakaolin at a ratio of 0.3 and alkali activated under the optimum synthesis conditions, namely H2O/Na2O molar ratio 16 in the activating solution, curing temperature 40 °C, curing time 24 h and ageing time 7 days.
The addition of metakaolin as precursor during alkali activation was considered due to its known high alkali activation potential. The properties and the microstructure of MW/MK AAMs are mainly due to the synergistic action of the filler effect of calcium carbonate present in marble waste and the precipitation of some calcium bearing phases.
The produced AAMs showed a fairly good performance after immersion in 1 mol/L NaCl solution and especially in deionized water. More specifically, the AAMs after immersion in deionized water for 30 days retained an acceptable compressive strength value (24.3 MPa). On the other hand, when the AAMs subjected to firing at 200 °C they lost a  Based on the evaluation of the results of the present study, it is deduced that the best AAMs obtained, e.g., MW/MK 0.3, may be used as alternative binders which exhibit good mechanical and physico-chemical properties which are similar to those of ordinary Portland cement. Thus, the production of alkali activated materials from marble wastes amended with metakaolin offers significant environmental and economic advantages and results in the valorization of this large waste stream which is usually disposed to the environment and causes several impacts [14,60].

Conclusions
This study showed that marble waste can be valorized, with the addition of metakaolin, via alkali activation for the production of AAMs with beneficial physical and mechanical properties. The highest compressive strength was obtained by the AAMs produced when marble waste was mixed with metakaolin at a ratio of 0.3 and alkali activated under the optimum synthesis conditions, namely H2O/Na2O molar ratio 16 in the activating solution, curing temperature 40 °C, curing time 24 h and ageing time 7 days.
The addition of metakaolin as precursor during alkali activation was considered due to its known high alkali activation potential. The properties and the microstructure of MW/MK AAMs are mainly due to the synergistic action of the filler effect of calcium carbonate present in marble waste and the precipitation of some calcium bearing phases.
The produced AAMs showed a fairly good performance after immersion in 1 mol/L NaCl solution and especially in deionized water. More specifically, the AAMs after immersion in deionized water for 30 days retained an acceptable compressive strength value (24.3 MPa). On the other hand, when the AAMs subjected to firing at 200 °C they lost a Based on the evaluation of the results of the present study, it is deduced that the best AAMs obtained, e.g., MW/MK 0.3, may be used as alternative binders which exhibit good mechanical and physico-chemical properties which are similar to those of ordinary Portland cement. Thus, the production of alkali activated materials from marble wastes amended with metakaolin offers significant environmental and economic advantages and results in the valorization of this large waste stream which is usually disposed to the environment and causes several impacts [14,60].

Conclusions
This study showed that marble waste can be valorized, with the addition of metakaolin, via alkali activation for the production of AAMs with beneficial physical and mechanical properties. The highest compressive strength was obtained by the AAMs produced when marble waste was mixed with metakaolin at a ratio of 0.3 and alkali activated under the optimum synthesis conditions, namely H 2 O/Na 2 O molar ratio 16 in the activating solution, curing temperature 40 • C, curing time 24 h and ageing time 7 days.
The addition of metakaolin as precursor during alkali activation was considered due to its known high alkali activation potential. The properties and the microstructure of MW/MK AAMs are mainly due to the synergistic action of the filler effect of calcium carbonate present in marble waste and the precipitation of some calcium bearing phases.
The produced AAMs showed a fairly good performance after immersion in 1 mol/L NaCl solution and especially in deionized water. More specifically, the AAMs after immersion in deionized water for 30 days retained an acceptable compressive strength value (24.3 MPa). On the other hand, when the AAMs subjected to firing at 200 • C they lost a substantial percentage of their compressive strength while after firing at temperatures higher than 400 • C their integrity was deteriorated fast.
In addition, the structural integrity of the produced AAMs was severely affected after subjection to freeze-thaw tests, and the compressive strength loss after 7 cycles reached almost the value of 40%. This is attributed to the high porosity (25.1%) and water absorption (14.2%) values of the initial specimens.
Additional studies are under way to improve alkali activation of marble waste and maximize its valorization potential. These studies explore the use of other precursors including mixtures of marble waste and other waste types with much higher Si and Al content which exhibit higher alkali activation potential, such as metallurgical slags or construction and demolition wastes, to improve the properties of the produced AAMs. Substitution of metakaolin with other precursors may result in a process with reduced cost and also enable the valorization of other waste streams that are produced worldwide in very large quantities. Also, the incorporation of natural fibers in the AAMs produced after alkali activation may result in specimens with beneficial properties. In any case, the selection of an additive that will improve the alkali activation potential of marble wastes depends on the desired properties of the produced specimens that define their final use as for example binder or construction element exhibiting among others suitable compressive strength, porosity or fire resistance.
Author Contributions: K.K. conceived of the idea, designed the experiments, analyzed the results and reviewed the paper. A.S. performed the experiments, analyzed the results and wrote the first draft of the paper, G.B. carried out SEM studies and reviewed the mineralogical part of the paper. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the project "INVALOR: Research Infrastructure for Waste Valorization and Sustainable Management" (MIS 5002495), which is implemented under the Action "Reinforcement of the Research and Innovation Infrastructure," funded by the operational program "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available yet, due to the fact that the final report of the project INVALOR has not been submitted.