An Innovative Method for the Sustainable Utilization of Blast-furnace Slag in the Cleaner Production of One-part Hybrid Cement

Hybrid cement (HC) can be dened as alkali activated-blended-Portland cement (PC). It was prepared by the addition of an alkaline solution to high-volume aluminosilicate-blended-PC. Although this cement exhibits higher mechanical performance compared to conventional blended one (aluminosilicate-PC blend), it represents lower commercial viability because of the corrosive nature of an alkaline solution. Therefore, this study focuses on the preparing one-part-HC using dry activator-based blast-furnace slag (DAS). DAS was prepared by mixing sodium hydroxide (NaOH) with blast-furnace slag (BFS) at low water to BFS ratio, followed by drying and grinding to yield DAS-powder. Different contents of DAS (equivalent to 70 wt.% BFS and 1, 2, and 3 wt.% NaOH) were blended with 30 wt.% PC. A mixture containing 70 wt.% BFS and 30 wt.% PC was used as a reference sample. The mortar was adjusted at sand: powder (BFS-PC and/or DAS-PC) weight ratio of 3: 1. The microstructural analysis proved that DAS-powder is mainly composed of sodium calcium aluminosilicate activated species and unreacted BFS. These species can interact again with water to form calcium aluminum silicate hydrate (C-A-S-H) and NaOH, suggesting that the DAS acts as NaOH-carrier. One-part HC-mortars having 1, 2, and 3 wt.% NaOH recorded 7-days compressive strength values 82, 44, and 27 %, respectively, higher than that of the control sample. At 180-days of curing, a signicant reduction in compressive strength was observed within HC-mortar having 3 wt.% NaOH. This could be attributed to the increase of Ca (within C-S-H) replacement by Na, forming Na-rich-phase with lower binding capacity. The main hydration products within HC are C-S-H, C-A-S-H, and chabazite as one of zeolite family.


Introduction
Portland cement (PC) is the common and dominant binding material in the construction sector [1,2]. Almost 4 billion tonnes is the annual production of PC worldwide, since China produces more than a half of global cement production [3,4]. About 8% of the total anthropogenic global warming potential is resulted from PC-manufacturing [5]. Each metric tonne of PC requires 4.2 GJ energy, releasing approximately in 0.8-1.0 tonne of carbon footprint into surrounding environment [6]. To mitigate the high CO 2 emission and energy demand, several authors replaced a high portion of PC by supplementary cementitious materials such as y ash (FA), silica fume (SF), and blastfurnace slag (BFS) [7][8][9][10][11]. Although the role of supplementary cementitious materials in the mitigation of carbon footprint and the improvement of the durability and the later mechanical properties of PC [12][13][14][15][16][17], the substitution of PC with high volume of these materials caused a noticeable retardation in its early hydration [18][19][20].
BFS is simply de ned as a calcium aluminosilicate rich-waste resulted from the top of smelted iron during the heating of iron ore in a blast furnace. The molten BFS was quenched with water to yield glassy materials with high amorphous content [21]. Statistically, the extraction of 1 tonne of iron from ore generates almost 0.3 to 1.0 tonne BFS [22]. It is well known that BFS was used as a partial substituent to PC to yield what is called as slag cement [23]. Several authors stated that the replacement of PC by high volume BFS has re ected on a signi cant reduction of the heat of hydration, resulting in a retardation in the early compressive strength [24][25][26]. It was found that the performance and hydration characteristics of PC-BFS cement enhanced with increasing the curing temperature and the neness of slag [27][28][29].
The addition of nano silica led to a signi cant improvement in the early compressive strength accompanied by an acceleration in setting time of high volume BFS blended cement [30][31][32][33]. A considerable improvement (16%) in the compressive strength of high volume BFS blended cement was achieved by the addition of 1 wt.% nano alumina [34]. the positive role of silica and alumina nano-particles is mainly originated from the formation of the additional calcium silicate hydrate (C-S-H), ettringite and calcium aluminate hydrate (C-A-H), which have high e ciency in the mechanical properties improvement and the pore size reduction [30][31][32][33][34]. Other study [35] proved that the addition of 1 to 4 wt.% nano calcium carbonate to cement paste individually containing 70, 80, and 90 wt.% BFS led to the enhancement of their strengths by 8 to 24%, 1 to 16% and 2 to 20%, respectively.
As an innovative approach, the addition of alkalis to high volume BFS-blend-PC has resulted in the formation of the alkali-activated BSF-PC, namely hybrid cement (HC) [36,37]. The individual addition of sodium hydroxide (NaOH) and sodium silicate (Na 2 SiO 3 ) to PC containing 80 wt.% produced hardened materials with compressive strength value 4.5 and 10.8 times, respectively, higher than that BFS-PC blend [38]. After dissolving alkali, it should be kept for enough time to cool before its mixing with BFS-PC blend. The corrosive nature of alkaline solutions [39] is the common reason behind the retardation of the commercial viability of this type of cement. It is important to produce hybrid cement, in which alkali incorporated inside its composition. In other word, one-part-HC (just added water) should be prepared to achieve the safe use of this type of cement.
Therefore, this paper focused on the utilization of dry activator-based alkali activated BFS (DAS) in preparing onepart hybrid alkali-activated slag-PC. DAS was prepared by mixing sodium hydroxide solution with BFS (at low water to powder ratio of 0.1), followed by drying and grinding to yield DAS-powder, which acts as a safe and stable NaOH-carrier. The ability of sodium leaching from DAS into surrounding media was evaluated using pH measurement. The impact of sodium oxide content within dry activator on the performance of HC composed of 70 wt.% BFS and 30 wt.% PC was evaluated. This study also suggested the hydration reaction mechanism and the composition of strength-giving-phases resulted from one-part-HC.

Materials resources
One-part hybrid cement (HC) mortar was fabricated from sand, ordinary Portland cement (OPC), blast-furnace slag (BFS), and sodium hydroxide (NaOH). Sand was brought from El-Wasta area (Beni Suef, Egypt). OPC (type I: 42.5 N) was purchased from Beni-Suef Cement Company (Cairo, Egypt). BFS was supplied from Helwan Company for Steel Industry (Helwan, Egypt). NaOH with purity of 99.99% was imparted by LOBA Chemical Company (India). Table 1 shows the chemical oxides compositions of OPC, BFS, and sand. X-ray diffraction (XRD) analysis ( Fig. 1) proved that BFS exhibits a completely amorphous pattern with a hump at 2theta of 20-35º. OPC and BFS demonstrated speci c surface areas of 3410 and 3540 cm 2 /g, respectively.

Preparation of one-part hybrid cement powder
One-part hybrid cement powder was synthesized by mixing dry activator based alkali activated slag (DAS) with OPC. DAS was prepared by activating the BFS with NaOH solution, followed by drying and grinding. As shown in Table 2, BFS was individually activated by 1, 2, and 3 wt.% NaOH at W/BFS ratio of 0.1. The activated slurry immediately dried at 80 ºC for 24 h, followed by grinding to yield DAS-powder. The W/BFS ratio of 0.1 was chosen according to previously published work [40], which reported that W/BFS ratio of 0.1 is appropriate water content for the formation of alkali-activated powder with high capability to re-interact with water, yielding hardened materials. One-part-HC was prepared by mixing DAS with OPC at different weights equivalent to 70 wt. % BFS and 1, 2, and 3 % NaOH (by weight of BFS). Control sample containing 70 wt.% BFS and 30 wt.% OPC was made for comparison. The details of mixing proportions were listed in Table 3.

Preparation of one-part-HC mortar
One-part-HC mortar was designed at sand: powder (BFS-OPC and/or DAS-OPC) weight ratio of 3:1. Sand and powder were dry mixed in a ball mill; after that the dry blend was transferred to mixer, then the water was added (at W/P ratio of 0.47). Slow and rapid rates of wet mixing were applied on the fresh cement mortar to ensure the complete homogeneity. The workable mortar was transferred into stainless steel molds with dimensions 50 x 50 x 50 mm, followed by vibration, smoothing, and curing in relative humidity of 99 ± 1% at 23 ± 2 ºC for 24 h.
Thereafter, the hardened mortar was demolded and cured under tap water for 7, 28, 90, and 180 days. Cement paste with the same BFS-OPC and DAS-OPC weight ratios were prepared for investigating the hydration products.

Experimental methods
Different experimental methods, including owability, setting time, zeta potential, and compressive strength, were carried out on the prepared one-part-HC-paste and/or mortar. The workability of the fresh one-part-HC-mortar and BFS-OPC-mortar was determined by measuring the average spread diameters of the fresh mortar on ow the table [41]. Initial and nal setting times of the fresh pastes were conducted three times on each mixture using Vicat apparatus based on ASTM C191 [42]. Zeta potential of the fresh cement mortars was measured using Malvern Zetasizer (nano-series), in which deionized water was used as a carrier liquid. This test was conducted to determine electrostatic repulsion between hydrated cement particles within cement mortar. Compressive strength of the hardened one-part-HC-mortar was measured according ASTM C109M [43] using German-Bruf-Pressing Machine with a maximum load capacity of 175kN. This test was conducted on three-hardened cubes of each mixture and the average reading was recorded. The broken paste was crushed and washed several times using acetone and methanol solution (at volume ratio of 1:1), then dried at 70 ºC for 3 h, to stop the hydration reaction.
After that, the dried paste was kept in a vessel until analyses. In contrast, the broken cement mortar was immersed in the same solution for 24 h, followed by drying at 70 ºC for 3 h, and then kept until microstructural investigation.

Instrumental techniques
pH of the prepared DAS was conducted on the ltrate of the suspended solution (at DAS to distilled water weight ratio of 0.5) via Delta OHM HD 8705 pH meter and PCFC11 combination electrode with accuracy of 0.01. This test was conducted three times and the pH value was accepted if the variation rate of reading was less than 0.01 / minute. The oxides compositions of the starting materials were determined using X-ray uorescence spectrometer (XRF: Xios, PW1400). The mineralogical compositions of the hardened cement pastes was identi ed using X-ray diffraction (XRD). This analysis was conducted on the powdered-sample of the hydrated cement pastes using Philips PW3050/60 diffractometer. The mineralogical compositions were determined within the 2theta range of 5-50º with 1s/step scanning rate and 0.05°/step resolution. Thermogravemetric analysis and its derivative (TG/DTG) was performed using DT-50 Thermal Analyzer (Schimadzu Co-Kyoto, Japan). This analysis was carried out on the powdered-cement paste to identify the hydration phases within its matrix. Each weight loss appeared at de nite temperature is a liated to the speci c hydration product. This test was conducted by weighing 20 mg of sample in Pt-crucible, then heated in N 2 atmosphere up to 1000 ºC and heating rate of 10 ºC / min. The functional groups within hydration products were identi ed using Fourier transform infrared (FT-IR) spectroscopy (KBrdiscussing Genesis-II FT-IR spectrometer) at the wavelength range of 400-4000 cm − 1 . The microstructural development of the hardened one-part-HC-mortar was investigated using eld emission scanning electron microscopy (FESEM, FEI Company, Holland) provided by an energy dispersive X-ray analyzer (EDS).

Characterization of DAS
The characterization of DAS is very important to determine its reaction mechanism and its role in the hydration of one-part-HC. FESEM-micrographs (Fig. 2) show that the addition of NaOH to BFS during preparing DAS has resulted in a partial dissolution of polygonal-shaped-BFS particle to form sodium calcium aluminosilicate species as con rmed by EDS-analysis. Additionally, the formation of these species enhances with NaOH addition.
As stated by Davidovits [44], gehlenite and akermanite within BFS can be dissolved by NaOH to yield (Na, Ca)ortho-sialate hydrate and calcium silicate hydrate as activated species. After that, these species condense together to form (Na, Ca)-cyclo-ortho-(sialate-disiloxo) and excess of calcium silicate hydrate (Fig. 3). Polymerization process is the following step, in which a long chain of sodium calcium aluminosilicate is formed. The mixing water content plays an important role in the dissolution/condensation process [45]. As prviousely stated [40], the use of low water during activation process plays a circular role in the control of condensation rate. The addition of low water content, followed by drying, causes a retardation in the condensation process, forming DAS with high activated species content. This is the main reason behind the justi cation of mixing water at W/BFS ratio of 0.1 and the application of drying during preparing DAS.
To identify the reactivity of activated species, the leaching test was applied to the prepared DAS. As shown in Fig.   4, a signi cant rising in pH-value in a short time (5 minutes) was recorded after suspension of DAS-powder in water. The possible explanation of this outcome is the moving Na cation at the ortho position within activated species, forming sodium hydroxide in the water medium (Fig. 5). This means the high hydraulic reactivity of DAS, as the activation process can be continued after water addition. It can be said that the DAS acts as a carrier of NaOH, which strongly contributes to resolve the corrosive nature of alkali solutions during preparing one-part-HC.

Flowability and zeta potential of one-part-HC-mortars
The impact of NaOH contents within DAS on the workability and zeta potential of fresh one-part-HC mortar is represented in Fig. 6. Generally, the spreading diameter of the fresh HC-mortars indicates their workability. The fresh HC-mortar with the broadest speeding diameter exhibits the best uidity. There is a direct relationship between the owability of the mortar and the negative zeta potential values. Increasing the NaOH content within DAS leads to a signi cant enhancement in the workability, which coincides with an increment in negative zeta potential value, as in line with the previously published work [46], which stated that the addition of Na 2 O content enhances the workability of one-part alkali activated cement. He et al. [47] proved that increasing negative zeta potential is mainly originated from an increment in the electrostatic repulsive force between cement particles, resulting in an enhancement in the workability of the fresh cementitious material.

Setting time of one-part-HC pastes
The initial and nal setting times (IST and FST, respectively) of the fresh BFS-OPC and DAS-OPC pastes are shown in Fig. 7. The setting time of the fresh cement paste mainly depends on the content of NaOH within DAS.
Depending on the composition of the fresh cement pastes, the IST was identi ed after 183-264 minute; meanwhile, the FST was recorded after 275-367 minute. One-part HC samples record shorter IST and FST compared to BFS- consumption and C-S-H formation rates. Increasing NaOH content up to 3 wt.% within DAS has resulted in a signi cant increment in C-S-H formation at the expense of Ca(OH) 2 phase. This con rms the fact that the activated aluminosilicate species within DAS can easily interact with Ca(OH) 2 to yield C-S-H and/or C-N-A-S-H binding phases [52][53][54][55]. Ettringite has been detected within the pattern a liated to control sample at 7-days of hydration.
With time advanced up to 28 days, ettringite peaks disappeared. This con rmed the dissociation of ettringite with curing time, as in agreement with previous reports [23,56]. On the other hand, the presence of alkali within one-part HC-paste prevents the formation of ettringite. This could be the possible reason behind the appearance of gypsum within the hydrated one-part HC pastes. Halaweh [57] stated that the presence of alkali increases the rate of sulfate release into solution, causing instability of ettringite. Cement hydration and alkali BFS activation are the two synergistic mechanisms of one-part HC-mortar. The activated aluminosilicate species-containing-DAS interacts with water to yield C-A-S-H and NaOH (Fig. 5). The unreacted BFS within DAS was dissolved by the liberated NaOH through alkali-activation process, yielding C-N-A-S-H, as con rmed by XRD-analysis (Fig. 8). On the other hand, the OPC within one-part HC-mortar also interacts with water to produce C-S-H and Ca(OH) 2 . An additional C-N-A-S-H could be formed through the consumption of Ca(OH) 2 by the activated species resulted from alkali BFS-activation. In contrast, the formation of strength-givingphases is resulted from the pozzolanic reaction, as C-S-H and/or C-A-S-H, was formed through the interaction between aluminosilicate within BFS and Ca(OH) 2 resulted from OPC hydration.
It is recognized that the enhancement of hydration products' content strongly re ects on the mechanical properties of the hydrated cement [40,46,58,66,67]. Accordingly, a relationship between compressive strength and hydration products content (determined by TG-analysis) within cement pastes hydrated for 28-days was represented in mortars. Therefore, a relationship between Na/Ca ratio (identi ed by EDS-analysis) and 28-days compressive strength values of one-part HC-mortar was represented in Fig. 13. Increasing the NaOH within DAS induces the incorporation of Na into the hydration products. In other words, the replacement of Ca within C-S-H and/or C-A-S-H enhanced with NaOH addition. Interestingly, the best 28-days compressive value was recorded at Na/Ca mole ratio of 0.11. Whereas a signi cant regression in compressive strength values was achieved when the NaOH content within DAS increases up to 3 wt.%.

Conclusions
This paper reported the synthesis and characterization of one-part hybrid cement mortar, in which dry activatorbased blast-furnace slag and Portland cement were the main ingredients. Dry activator with different contents were blended with Portland cement to achieve 70 wt.% blast-furnace slag and 1, 2, 3 wt.% NaOH. The hydrated one-part hybrid cement mortar containing 1 and 2 wt.% NaOH was found to exhibit shorter setting time, higher workability, and higher compressive strength compared to blended cement containing the same blast-furnace slag content. Although incorporating dry activator content equivalent to 3 wt.% NaOH accelerated the setting time and enhanced the workability and early compressive strength, it demonstrated the lower compressive strength at later ages of curing. Accordingly, sodium hydroxide content within dry activator played an important role in the performance of the prepared mortar, especially at later ages of hydration. As suggested by the reaction mechanism, dry activator is mainly composed of sodium calcium aluminosilicate species with a high ability to reinteract with water, yielding calcium aluminum silicate hydrate and sodium hydroxide. The liberated sodium hydroxide accelerated the dissolution of unreacted slag particles, resulting in an acceleration in the hydration products formation which coincided with compressive strength improvement. As identi ed by X-ray diffraction and thermogravemetric analyses, calcium silicate hydrate, calcium aluminum silicate hydrate, and chabazite zeolite are the dominant hydration products with the prepared hybrid cement. It can be said that preparatig dry activator resolved the drawback of the corrosive nature of alkaline solution, which could strongly re ect on the safe use of hybrid cement in different construction projects.