**1. Introduction**

As one of the most sustainable approaches to reducing carbon dioxide emissions from the production of cement-based construction materials, various types of industrial by-products are currently used worldwide as Portland cement replacements, called supplementary cementitious materials (SCMs), e.g., ground granulated blast-furnace slag (GGBFS), fly ash, and silica fume. It is well known that the replacement level of cement with SCMs depends on the reactivity, local availability, and legislation [1]. For the strength evolution of SCMs with cement, the pozzolanic reaction of SCMs is an essential chemical process that uses high alkalinity compounds, such as Ca(OH)2, from cement hydration products [2]. Although blending cement with SCMs has many advantages—such as reducing carbon dioxide emissions by saving cement, and improving long-term strength, durability, and chemical stability [3–7]—the setting time and early-age strength development can be drastically delayed if the amount SCMs in the blend is excessive, without adequate alkali activators [8,9].

Various alkali activators have been suggested for adequately promoting strength of blended concrete using SCMs to either partially or fully replace ordinary Portland cement (OPC). For example, it is well known that sodium hydroxide (NaOH), sodium silicate (Na2O·*r*SiO2), sodium carbonate (Na2CO3), and sodium sulphate (Na2SO4) are effective and common ones used to activate GGBFS [10,11]. These alkali activators are effective in promoting the initial reactivity of GGBFS and lead to rapid

strength development [3]. However, the radical chemical reaction can also dramatically accelerate the reaction of GGBFS, which can significantly affect the workability of concrete—such as its setting time and flowability—depending on the alkali content and the slag/activator ratio [10,12]. In addition, the manufacturing process of alkali activators is usually both energy intensive and costly [13], which could limit wider application of alkali activated slag concrete. Alternative studies have been attempted to evaluate sustainable alkali activators for SCMs using industrial by-products [14]. For example, Maraghechi et al. [1] tested recycled glass powder as an alkali activator for binary mixtures with OPC, slag, and fly ash. The findings suggested that elevated temperature curing (60 °C) is preferable to effectively consuming glass powder for alkali activated slag mortar. However, limited work has been conducted on the use of industrial by-products as alkali activators for slag cement and concrete.

The current mainstream liquid crystal display (LCD) design primarily consists of color filter (CF) substrate glass on which RGB pixels are deposited and of thin-film-transistor (TFT) substrate glass that is painted with thin film circuitry that delivers signals to liquid crystal. In order to maintain a high quality and a high resolution, TFT uses an alkali-free glass that is obtained through a high refining process that involves the generation of by-product. Because the by-product of this refining process contains a large amount of alkali ingredients—such as SO3, Na2O, and K2O—it is expected to be used as an alternative alkali activator for GGBFS blended cement concrete. Although the LCD by-product that results from the refining process does not require an additional treatment because it is in the form of powder, it is necessary to investigate its fundamental properties in order to verify its applicability as an alternative alkali activator for GGBFS blended cement concrete. However, to the best of the authors' knowledge, alkaline activation of LCD by-product has not been investigated as a means of developing sustainable construction material that can be utilized for the production of high strength concrete. In order to adapt the alkali industrial by-product to slag concrete, it is necessary to find suitable combinations of base binders by investigating the reactivity of the by-product with existing binding materials. Therefore, the effects of the alkali by-product on the fundamental properties of concrete—such as workability, setting time, and strength enhancement under various mixing conditions—need to be identified. This research gap motivated the study presented in this paper. It should be pointed out, however, that multiple attempts have been conducted to use thin-film transistor liquid-crystal display (TFT-LCD) waste glass to partially replace OPC [15–18]. For example, Lin et al. [15] tested up to 40% of TFT-LCD waste glass to replace OPC, and concluded that as the amount of TFT-LCD waste glass increases the strength of the paste distinctly decreases. Jang et al. [16] used additional activator to promote reactivity of TFT-LCD waste glass with OPC, and concluded that the pozzolanic reaction between the waste glass and the activator leads to enhance compressive strength of high strength concrete products. Kim et al. [18] found that smaller particle size of ground TFT-LCD waste glass would lead to the decreased porosity of TFT-LCD waste glass concrete, which is expected to enhanced durability and permeability.

In this study, the applicability of the LCD by-product of the refining process as an alkali activator was evaluated by characterizing the reactivity of GGBFS with OPC under a normal curing condition with the aim of developing practical applications. The binary paste was prepared by mixing GGBFS and OPC with the LCD by-product-based activator (LCDBA) and the variation of the hydration products in the paste, according to the curing age, was characterized through X-ray diffraction (XRD) as well as through thermogravimetry and differential thermal analysis (TG-DTA). In addition, fresh and hardened properties of OPC-slag concrete, activated by the alkali by-product, were assessed by examining the slump, bleeding, setting time, and compressive strength of concrete.

#### **2. Experimental Program**

The applicability of the LCDBA was investigated with the aim of developing practical applications as a sustainable and alternative alkali activator for GGBFS blended cement concrete. The activation effects of the LCDBA were characterized using GGBFS blended cement paste. In addition, the fresh and hardened properties of GGBFS blended cement concrete incorporating LCDBA were investigated.

#### *2.1. Raw Materials*

OPC (ASTM C 150 Type I) and GGBFS were used as binders to prepare the paste and concrete and their chemical compositions are summarized in Table 1. The OPC and GGBFS that were used had a Blaine specific surface of 339 m2/kg and 449 m2/kg and a density of 3150 kg/m3 and 2910 kg/m3, respectively. The major chemical components of GGBFS were CaO, SiO2, and Al2O3, as shown in Table 1. The basicity coefficient (Kb = (CaO + MgO)/(SiO2 + Al2O3)) was 0.977, which is similar to the neutral value of 1.0 for ideal alkali activation [19]. The hydration modulus, according to a formula proposed in the literature (HM = (CaO + MgO + Al2O3)/SiO2) of GGBFS, was 1.92. This was higher than the value of 1.4, which is required for good hydration properties of GGBFS [19]. The LCD by-product of the refining process (see Figure 1), LCDBA—the chemical composition of which is given in Table 1—had a Blaine specific surface of 770 m2/kg and a density of 2540 kg/m3. XRD analysis was performed to determine the nature of LCDBA, as shown in Figure 2, which indicated that this was crystallized mainly with K2SO4 and Na2SO4. The particle size distributions of GGBFS and LCDBA are shown in Figure 3, which were measured using a laser particle size analyzer (PSA, Beckman Coulter LS 13 320, Brea, CA, USA). The overall particle size distribution of LCDBA was similar to that of GGBFS, showing a narrow distribution with the peak around 10 μm. Although LCDBA had about 1.7 times higher specific surface in comparison to GGBFS, the particle size of LCDBA was slightly larger than that of GGBFS, which indicated that LCDBA has porous microstructures resulting from the refining process.

**Table 1.** Chemical composition of OPC, GGBFS, and LCDBA (wt.%).


**Figure 1.** Image of LCDBA used in this study.

**Figure 3.** Particle size distribution of GGBFS and LCDBA.

Non-reactive river sand, with a specific gravity of 2.62, a fineness modulus of 2.77, and an absorption capacity of 1.1%, was used in preparation of all mortars and concrete. The crushed basalt aggregates were adopted as the coarse aggregate for the concrete mix in which the maximum size, specific gravity, absorption capacity, and fineness modulus were 25 mm, 2.63, 0.8%, and 6.4, respectively. In addition, a polycarboxylate-based superplasticizer with 17% solid content by weight was used to enhance particle dispersions within the mixture.

#### *2.2. Mix Proportions*

The characterization of the reactivity of the binders was carried out based on XRD and TG-DTA tests using GGBFS blended cement paste with the OPC/GGBFS ratio of 55:45 and the water/binder ratio of 30%. LCDBA replaced the binder by 0%, 3%, and 5%. Table 2 shows the concrete mix proportions according to the usage of LCDBA. As in the paste mix design, the ratio of OPC and GGBFS was set to 55:45. In addition, the binder substitution ratio of LCDBA was set to 0%, 3%, and 5% in order to evaluate the effects of the LCDBA substitution level on the fresh and hardened properties of GGBFS blended cement concrete. The GGBFS blended cement concrete was designed to have the total volume

of mortar in the 615 ± 5 L range and the slump of concrete in the 180 ± 25 mm range by controlling the amount of superplasticizer and the sand to total aggregate volume ratio (s/a).


**Table 2.** Mix proportions of concrete.
