**1. Introduction**

Supplementary cementitious materials (SCMs) are used increasingly in cement-based products, either for improving their properties or for reducing the carbon footprint of cement. Given that the hydration of ordinary Portland cement (OPC) is not yet understood in full, these materials bring even more complex reactions into the hydration process [1,2]. The benefit from utilization of SCMs lies either in their reactivity [3,4] or in the enhancement of cement hydration, as explained by the filler effect [5,6]. The environmental benefit from cement replacement with SCMs increases with the rate of replacement [7,8], but cement substitution must be limited to the extent that the performance of the final product is not undermined. The effect of SCM use could be beneficial to the performance of the produced concrete, depending on its fineness and on its cement substitution rate [9,10]. Cement substitution by a SCM is expected to affect the rate of strength development and the final strength, but also the water requirement and consistency of the cement paste [11,12].

SCM particles have a different size and specific surface area compared to Portland cement and, therefore, alter the microstructure and packing density of the cement paste. The particle size distribution of the cementitious materials has been found to influence both workability and hydration of cement pastes by improving their packing density [13,14]. Mixture design optimization by considering particle packing has been utilized in the design of ultra-high performance concrete [15,16]. There have been several mathematical packing models proposed to predict the packing density of multi-component mixes [17,18]. In order to predict the effect of cement substitution with SCMs, Yu et al. [19] proposed

**Citation:** Anastasiou, E.K. Effect of High Calcium Fly Ash, Ladle Furnace Slag, and Limestone Filler on Packing Density, Consistency, and Strength of Cement Pastes. *Materials* **2021**, *14*, 301. https://doi.org/10.3390/ ma14020301

Received: 4 December 2020 Accepted: 4 January 2021 Published: 8 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional clai-ms in published maps and institutio-nal affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

a linear packing model, considering the surface characteristics (sphericity) of particles. De Larrard [20] proposed the compressive packing model, considering the degree of compaction rather than particle surface characteristics, while Fennis et al. [21] proposed the compaction-interaction packing model, considering particle interaction and aggregate inclusion. According to Zhang et al. [22], the fresh cement paste can be seen as a suspension, with water either filling the voids between particles (filling water) or coating particles and providing fluidity (excess water). SCMs can act as fillers, improving the packing density of the suspension, which means that they can reduce the required amount of filling water. They can also provide space for the hydration of cement, accelerating strength development (filler effect). On the other hand, large SCM particles could block smaller cement particles from hydrating (wall effect) and a large proportion of small SCM particles could increase the distance between hydrated cement particles (loosening effect), affecting the consistency of cement pastes [23]. Mehdipour and Khayat [24] suggested that the presence of more fine particles than the amount required to fill the voids in the cement matrix contributes to the flowability of concrete. SCM particles themselves, on the other hand, may be contributing directly to strength development, if they exhibit hydraulic or pozzolanic properties.

There are several well-known SCMs, such as siliceous fly ash, silica fume, ground granulated blastfurnace slag, and limestone filler (LF). Their availability, however, varies locally, and several other materials are being researched, based on local availability, such as high calcium fly ash (HCFA), metakaolin, ladle furnace slag (LFS), and rice husk ash [25]. The present research investigates the use of HCFA, a by-product of lignite-fired power plants; LFS, a by-product of the steelmaking process; and LF, ground natural limestone, in cement pastes. HCFA is known to exhibit both pozzolanic and self-cementing properties and has been used for the past decades in blended cement manufacturing [26,27]. LFS is a weak pozzolan with some latent hydraulic properties and is mostly considered as filler [28,29]. LF is receiving increasing attention in the literature as a SCM since it seems to promote cement hydration [30–32]. A combination of SCMs in ternary systems is often proposed since there seems to be some synergy between alternative materials of different chemical composition and of different fineness [33,34].

The aim of the present research was to investigate the effect of cement replacement with the above SCMs, considering packing density, in binary and ternary mixtures. Since the fineness and reactivity of the SCMs have an impact on fresh paste consistency and strength development, it is important to understand how increasing cement replacement and altering packing density affects these properties. Furthermore, the ternary binders consisting of OPC, HCFA and LFS or OPC, HCFA, and LF were studied to identify possible synergistic effects. The goal was to identify ways of designing cement pastes with SCMs in the most beneficial way possible, since successful implementation can result in maximizing the positive effects of cement substitution and increasing SCM use.

#### **2. Materials and Methods**

CEM type I 42.5 N according to EN 197-1 [35] was used as OPC for all the tests. HCFA was used unprocessed, as received from the power plant. LFS was water-quenched and air-cooled and then sieved through the 100 μm sieve. LF was used as received from the cleaning of the limestone aggregate silos in a ready-mixed concrete plant. Table 1 shows the chemical composition of all the materials used, measured by atomic absorption spectroscopy (AAnalyst 400, Perkin Elmer, Waltham, MA, USA). The loss on ignition (LOI) and the chloride, sodium, and sulfate ion contents were determined by ionic chromatography (Thermo Scientific, Waltham, MA, USA, Dionex ICS-1100) for all the materials used.

Figure 1 shows the particle size distribution of the materials used, and Table 2 shows their median particle size diameter d50, specific surface area, and apparent specific density values. The particle size distribution, d50, and specific surface area of the fine materials were measured using a laser particle size analyzer (Malvern Mastersizer 2000, Worcestershire, United Kingdom). The apparent specific density of the fine materials was determined using a Le Chatelier flask, according to ASTM C188-14 [36].


**Table 1.** Chemical properties of the fine materials used in the present research (%wt.).

**Figure 1.** Particle size distribution of the fine materials used.



The first step in the experimental program was to identify water demand when substituting OPC with each of the three alternative binders. The required water to cementitious material (w/cm) ratios were determined both for maximum packing density and for equal consistency. The wet packing density approach, as proposed by Wong and Kwan [36], was followed in order to determine packing density for pastes with 100% OPC and for pastes with 10%, 20%, and 30% cement replacement with HCFA, LFS, and LF. The w/cm for maximum packing was recorded, referred to as optimum water demand. The reduction of the w/cm ratio increases packing density up to the point that the water fills the voids amongst solid particles, but further water reduction decreases packing. Thus, the optimum water demand is determined at the point where packing density is maximized. However, at maximum packing, the workability of the fresh paste is typically very low and serves as a

measure of the effect of SCM use on packing and maximum strength development. In order to determine water demand for workable pastes, the w/cm ratio for normal consistency, according to the Vicat method as described in European Standard EN 196-3 [37], was also determined for the same replacement rates.

The cement pastes were prepared in a laboratory mixer by adding water first and then adding the dry-mixed binders and mixing for 120 s. Additional mixing time of 30 s was allowed if required. The fresh pastes were placed in the 40 mm deep truncated conical Vicat mold and compacted on a vibration table for elimination of entrapped air and weighed for the determination of wet packing. Wong and Kwan [38] have identified packing density as the solid concentration *ϕ*, which is calculated from Equations (1) and (2) as follows:

$$
\varphi = \frac{V\_c}{V} \tag{1}
$$

where *V*c is the solid volume of the cementitious materials and *V* is the volume of the mold. The solid volume *V*c can be calculated from the following formula:

$$V\_{\mathbb{C}} = \frac{M}{\rho\_w u\_w + \rho\_a R\_a + \rho\_\beta R\_\beta} \tag{2}$$

where *M* is the mass of the paste in the mold; *ρw*, *ρα*, *ρβ* are the densities of water and cementitious materials *α* and *β*, respectively; *uw* is the water to cementitious material ratio by volume (w/cmV); and *R<sup>α</sup>* and *R<sup>β</sup>* are the volumetric ratios of cementitious materials *α* and *β*, respectively.

After weighing, the truncated conical specimens, still in the mold, were subjected to measurement of Vicat plunger penetration, according to EN 196-3, and the depth of penetration was recorded. The depth of plunger penetration, with a minimum of 0 mm and a maximum of 40 mm, characterizes the consistency of the paste and was used as a measure of workability. Normal consistency is described in the standard as the consistency that allows the plunger to penetrate the specimen 34 mm. Lecomte et al. [39] followed a similar approach to characterize the packing ability of various cement pastes. A series of pastes, 8 to 12 for each paste formulation, was prepared for various w/cm ratios in order to identify the optimum water demand for maximum packing and to determine the w/cm ratio for a paste of normal consistency, which was selected as a suitable level of workability. The above procedure was carried out for 100% OPC as reference and for 10%, 20%, and 30% wt. OPC replacement with HCFA, LFS, and LF, resulting in ten different formulations.

Based on the w/cm ratios for optimum water demand and for normal consistency, binary pastes with 20% OPC replacement with HCFA, LFS, or LF were prepared and tested for compressive strength development at 3, 7, 28, and 90 days. At least six 40 mm cubic specimens were tested at each age and paste, after curing at 20 ◦C and 95% relative humidity. These were compared to reference 100% OPC cement pastes and were used to assess the contribution of each SCM to strength development, either at maximum packing (optimum water demand), or at equal fresh state workability (normal consistency). Since the tested SCMs had varying effects on workability and strength development, it was decided to test ternary binders to identify possible benefits from the interaction of binders. Based on the strength development test results, two ternary cement pastes, one with 70% OPC, 20% HCFA, and 10% LFS and one with 70% OPC, 20% HCFA, and 10% LF were prepared and tested for strength development at 3, 7, 28, and 90 days at optimum water demand. Scanning Electron Microscopy (SEM-JEOL 840A JSM, Tokyo, Japan) was also used to assess the crystals formation and the microstructure of the reference and ternary cement pastes.

#### **3. Results**
