**Photocatalytic Recycled Mortars: Circular Economy as a Solution for Decontamination**

**Auxi Barbudo 1,\*,**†**, Angélica Lozano-Lunar 1,**†**, Antonio López-Uceda 2, Adela P. Galvín <sup>1</sup> and Jesús Ayuso <sup>1</sup>**


Received: 2 September 2020; Accepted: 14 October 2020; Published: 19 October 2020

**Abstract:** The circular economy is an economic model of production and consumption that involves reusing, repairing, refurbishing, and recycling materials after their service life. The use of waste as secondary raw materials is one of the actions to establish this model. Construction and demolition waste (CDW) constitute one of the most important waste streams in Europe due to its high production rate per capita. Aggregates from these recycling operations are usually used in products with low mechanical requirements in the construction sector. In addition, the incorporation of photocatalytic materials in construction has emerged as a promising technology to develop products with special properties such as air decontamination. This research aims to study the decontaminating behavior of mortars manufactured with the maximum amount of mixed recycled sand without affecting their mechanical properties or durability. For this, two families of mortars were produced, one consisting of traditional Portland cement and the other of photocatalytic cement, each with four replacement rates of natural sand by mixed recycled sand from CDW. Mechanical and durability properties, as well as decontaminating capacity, were evaluated for these mortars. The results show adequate mechanical behavior, despite the incorporation of mixed recycled sand, and improved decontaminating capacity by means of NOx reduction capacity.

**Keywords:** recycled aggregates; recycled mortar; construction and demolition waste; decontaminating; photocatalysis

#### **1. Introduction**

A current environmental challenge is to optimize natural resources. For this, minimizing and recovering waste materials are essential from a productive point of view of the main economic sectors. The circular economy model is a sustainable production strategy where the reuse of waste as secondary raw materials is underlined, achieving comprehensive management of waste materials [1]. So, the circular economy concept describes a cyclical system in which economic and environmental aspects are integrated [2]. Reducing the resources used and the waste generated can save resources and help reduce environmental pollution.

At the European level, this model is implemented through the Action Plan for the Implementation of the Circular Economy of the European Parliamentary Commission [3]. This proposed transition toward a more circular economy brings great opportunities. It is important to make efforts to modernize and transform the economy, shifting it toward a more sustainable direction, which will enable companies to make substantial economic gains and become more competitive in the market. This not only offers important energy savings and environmental benefits, but also creates local jobs

and opportunities for social integration. The wider benefits of the circular economy also include reduced energy consumption and carbon dioxide emission levels [3].

The model highlights the important role played by the construction sector, which is still associated with strong negative environmental effects due to the high natural resource consumption for manufacturing and the large production of waste [4]. Waste from construction activities, including excavation or land formation, earthworks, civil construction and building, road works, and building renovation, are considered as construction and demolition waste (CDW) [5,6]. These represent a relevant part of the total generation of solid waste worldwide [7]. Due to this, European Directive 2008/98/EC [8] established that EU member states must reach a recycling rate of 70% for CDW in 2020 to avoid sending it to landfills. However, the potential for reuse and recycling of this waste stream is not being fully exploited. One obstacle is a lack of confidence in the quality of recycled construction and demolition materials. For this reason, on 9 November 2016, the European Commission proposed an industry-wide voluntary protocol on the management of construction and demolition waste [3]. The aim of the protocol was to improve the identification, source separation, and collection of waste, as well as logistics, processing, and quality management. The protocol could thus increase trust in the quality of recycled materials and encourage their use in the construction sector.

After CDW is properly treated, its use as recycled aggregates (RAs) in the construction sector has been widely addressed to mitigate environmental problems such as the consumption of raw materials and waste landfilling. This leads to an increase the recycling rate [9–15], becoming a practical reality that has allowed the development of specific regulations [16].

RAs have different physical, mechanical, and chemical properties from natural aggregates (NAs). They have lower density, higher water absorption, lower resistance to fragmentation, and a higher content of sulfur compounds and soluble salts [17,18]. Depending on the original waste material, recycled aggregates could be concrete, ceramic, or a mixture (recycled mixed aggregate, RMA). RMAs come from building demolitions and contain a wide range of materials, such as concrete waste, pavement material, ceramic products, and, in smaller quantities, other materials such as gypsum, glass, wood, etc. [19]. On some occasions, they are used in works with lower requirements, such as roads with low traffic intensity [20,21], bike lanes [22], urban and pedestrian roads, and unpaved rural roads [23–25], where RAs provide similar functional and structural characteristics as NAs.

Another widely studied application is in the manufacture of concrete. There are studies that support its use, even for structural concrete [26,27]. The use of recycled aggregates in the production of concretes and mortars has the following competitive advantages: (i) decreased extraction of aggregates from rivers, coasts, and quarries; (ii) exclusion of aggregates from landfills, reducing the volume of waste to be treated; and (iii) implementation of the circular economy model and approach to set recycling targets [14].

In most of these studies, only the coarse fraction was used [7,11,12,28]. It was concluded that concrete strength decreased when recycled concrete was used and the reduction could be as low as 40% [29]. No decrease in strength was reported for concrete containing up to 20% fine or 30% coarse recycled aggregates, but beyond these levels, there was a systematic decrease in strength as the content of recycled aggregates increased [29].

Of the different types of RAs, recycled crushed concrete (RCA) is the most widely used in the manufacture of new concrete [9,26,27,30–32]. Even the Spanish concrete standard (EHE-08) [16] allows the use of up to 20% RCA, but only referring to fractions greater than 4 mm. Because of the lower density of RMA, concrete made with it has lower density and higher water absorption than reference concrete [19].

Regarding the influence of recycled coarse and fine fractions, studies on recycled concrete incorporating fine recycled aggregate (FRA) from CDW did not obtain satisfactory results [19,29,30,32,33]. Kathib [29] studied the incorporation of FRA, and showed that properties such as density, dynamic modulus of elasticity, and compressive strength were reduced, the latter resulting in a decrease of 10% and more than 15% with a 50% and 100% incorporation ratio, respectively, at

90 days, in agreement with the results by Kou et al. [33]. Because of these differences, the use of fine fractions in concrete should not be dismissed, but more research on it is needed.

For this reason, the main application of FRA is in recycled mortar, for which the requirements are more tolerant. Within this group, there are different uses depending on where it will be placed: masonry mortar, interior or exterior rendering mortar, or mortar for the manufacture of paving blocks. Most of the investigations carried out to date were aimed at incorporating FRA in mortar for masonry [31,34–36]. However, this research in particular is intended for the use of recycled mortar for any application in contact with the atmosphere, such as wall cladding or paving blocks. With the results obtained, it is hoped that its use can be further expanded.

As in the case of recycled concrete, the use of mixed FRA (MFRA) in mortar has been found to be harmful to the properties of mortar [14,36]. However, this research aims to prove its use as suitable if it meets a series of quality requirements, normally linked to correct CDW treatment.

In addition to the circular economy, the mortar manufactured in this research is intended to provide a second benefit for environmental sustainability: atmospheric decontamination and reduced carbon footprint. The atmospheric pollution produced by accelerated population growth and industrialization can cause serious damage to the health of both people and ecosystems, and even to infrastructures and historical heritage. Among the main harmful emissions are nitrogen oxide gases (NO and NO2, commonly known as NOx) generated by transportation and various industries. These gases have high toxicity and can cause serious problems in human health, as well as environmental problems (acid rain, photochemical smog, destruction of the ozone layer, etc.) [37].

It is essential to reach the highest air quality that does not create a risk to people's health and does not cause deterioration or permanent damage to ecosystems. The current measures for reducing air pollution in cities associated with CO2 emissions and other pollutant gases fall within two lines of action: citizen awareness policies, which advocate avoiding the use of private vehicles in favor of alternative means of transport, such as bicycles or public transport; and policies restricting the circulation of these vehicles, either with speed reductions or by prohibiting movement in downtown areas, as is carried out in cities such as Paris, London, and Madrid [38].

However, these measures are not fully effective in eliminating or reducing air pollution in urban environments, which is why complementary measures are needed in addition to conventional pollution control methods. Within these new measures, it is interesting to consider the advantages of photocatalysis.

The photocatalysis process starts from the natural principle of decontamination of nature itself. It is a similar technology to that of photovoltaic solar panels [39]. Like photosynthesis, which, thanks to sunlight, can remove CO2 to generate organic matter, photocatalysis removes other usual pollutants in the atmosphere such as NOx and SOx (inorganic compounds) and volatile organic compounds (VOCs) through an oxidation process activated by solar energy. Through photocatalysis, most of the pollutants present in urban areas can be reduced, such as NOx, SOx, VOCs, CO, methyl mercaptan, formaldehyde, chlorinated organic compounds, poly-aromatic compounds, etc., which are aggressive in terms of both the properties of the material and the environment.

Construction materials treated with photocatalysts reduce, above all, NOx particles that are produced by vehicles, industry, and energy production. During photocatalysis, the photocatalyst agent absorbs light energy, transfers it to a reactive compound, and triggers a chemical reaction through the formation of radicals. Titanium dioxide (TiO2) and the products derived from it are the most widely used photocatalysts, and it is these that trigger the transformation of NOx (nitrogen oxides) into nitrate through the action of sunlight. This can then be used to increase the shelf life of cement-based materials, while it can also be used to substantially decrease the concentration of some air pollutants, especially in semi-enclosed places such as important urban avenues, tunnels, or heavily polluted places like gas stations and some specific industries [39].

In the process of decontamination by photocatalysis, the contaminant is absorbed on the surface of the material to be later oxidized, in two stages, to an inert nitrate compound (NO3). Finally, the inert compound is removed from the surface of the material by rain.

The incorporation of photocatalytic substances in construction materials has emerged as a promising technology to develop products with special characteristics/properties [40]. Among the multiple advantages of photocatalysis is that it is a clean technology that does not need any maintenance, and once applied, its effect is permanent, and it "cleans" the contaminated air. It also saves on costs, since areas where this mortar has been used remain clean for many years, and it destroys the dirt that is deposited on it, which favors the growth of microorganisms [41–45]. Even the possibility to eliminate pollen [46] or deposited soot [47] from the air has been studied.

A building's façade is one of its most important parts since it gives it a distinct personality. On the other hand, pavements in urban areas, steels, parking lots, etc., are among the most important parts of civil construction. In all these cases, it is important to assess the useful life of the chosen material and hence there is a need to investigate new solutions that extend any maintenance operation over time and directly contribute to improving environmental sustainability [38].

With an awareness of the needs of today's society in terms of waste reduction, specifically CDW, through the use of RA and actions on pollution in urban environments, this research aims to contribute to the development of solutions to both problems by developing recycled mortars with photocatalytic capacity that contribute to the preservation of the environment with sustainable initiatives, and are also technically viable. This study investigates the effect of photocatalytic mortars on reducing air pollution produced by traffic emissions of CO2 and NOx, with the added value of being made with FRA. The intention is to serve to advance the research work and thus achieve the necessary objectives for sustainable development in accordance with European and Spanish regulations. It contributes to the circular economy by using recycled materials to reduce the waste generated and the need to obtain new raw materials, and improves the quality of life of citizens who live in population centers that have severe pollution problems, which is increasingly present in urban environments.

For this, two families of mortar were manufactured with different types of cement using different replacement rates. The aim was to obtain a recycled mortar that provides the highest decontamination capacity using the highest possible percentage of recycled sand without significantly affecting its mechanical and durability properties.

The rest of the article is structured as follows: the "Materials" section details the materials for the production of the mortars; the "Experimental Program and Methods" section specifies how the research has been carried out; in the "Results and Discussion" section, the data obtained is analyzed; and finally, in the "Conclusions" section, the main advances obtained are highlighted, as well as future lines of research.

#### **2. Materials**

The specific materials selected for the research are described below.

#### *2.1. Cements*

Two types of cement were used in this research. Cement for the conventional mortar was CEM I 52 5 N (Cement without additions, high strength (52.5 MPa at 28 days), and normal initial strength). The photocatalytic cement used was i.tech ULTRA, hereinafter called Ph. CEM I. This is a Portland cement similar to CEM I, but with an addition of titanium oxide (TiO2). Both cements come from the same manufacturer.

#### *2.2. Aggregates*

This research used two aggregates with a granule size of 0–4 mm:


The composition test carried out on the coarse fraction of said aggregate (>4 mm), according to EN 933-11, showed 18% ceramic particles (Rb), 34% concrete and mortar (Rc), 47% natural aggregate (Ru), and 1% other particles (X), among which plaster stands out. The water absorption and dry specific density of NS and RS, according to EN 1097-6, are shown in Table 1.

**Table 1.** Specific gravity and water absorption of aggregates. NS, natural sand; RS, recycled sand.


The results indicate that RS had a lower specific gravity and higher percentage of water absorption compared with the corresponding properties of natural aggregates, agreeing with what was indicated by other authors [31,35]. This may be due to a higher percentage of mortar and ceramic particles. However, the percentage of water absorption after 24 h of immersion was lower than the 6–9% obtained by other authors [36,48–50]. This may be due to better treatment of the recycled aggregate, or because the original waste contained fewer porous particles.

#### **3. Experimental Program and Methods**

#### *3.1. Experimental Program*

A total of 8 mortars were manufactured and divided into 2 families, one produced with CEM I (conventional mortar family) that would be used as a reference, and one with Ph. CEM I (photocatalytic mortar family) was used to determine the decontaminating power of mortars made with TiO2.

In order to increase to the sustainability of the sector, as much RS as possible is expected to be used. For it, each family was produced according to 4 replacement rates of NS by RS (0%, 20%, 40%, and 100%) by weight. Table 2 shows the nomenclature of mortar families produced in this research.


**Table 2.** Nomenclature of mortar families.

The mortar dosage was calculated based on EN 196-1 and is shown in Table 3. The weights of aggregates shown in this table refer to dry weight.

The amount of mixing water shown in Table 3 was constant, resulting in a water/cement (w/c) ratio equal to 0.58. Due to the low water absorption of RS compared to other recycled aggregates, an increase in the water content as a percentage of increased incorporated RS was not considered, since a high w/c ratio could produce a weaker and more porous mortar [51].

The mixing procedure was in accordance with EN 196-1. A total of 12 prismatic specimens were produced with dimensions of 40 × 40 × 160 mm. These specimens were cured in a climatic chamber at 20 ± 1 ◦C and 65% relative humidity until the age test.


**Table 3.** Mortar dosage (per m3).

#### *3.2. Mortar Characterization*

The consistency of the mortar families in the fresh state was measured in accordance with EN 1015-3. The hardened state properties analyzed in the mortar families were compressive and flexural strength (EN 1015-11), water absorption by capillarity (EN 1015-18), water absorption capacity, bulk and skeletal density and open porosity for water (Spanish Standard UNE 8398), carbonation depth (EN 13295), and photocatalytic activity (Spanish Standard UNE 83321 EX). These properties were evaluated after 28 days of curing time.

In addition, X-ray diffraction (XRD) analysis was carried out to identify the main crystalline mineral components. For that purpose, a piece of the central part of each mortar specimen was crushed and sieved through a 0.125 mm sieve. The machine used for this technique was a Bruker D8 Discover A25 with Cu-Kα radiation, and the goniometric exploration used was swept from 5◦ to 80◦ (2θ◦) at a speed of 0.0142◦ min<sup>−</sup>1. The Joint Committee on Powder Diffraction Standards database was used to identify the phases formed in the mortars [52]. A JEOL JMS-7800 scanning electron microscope (SEM) was used to determine the mortar's chemical composition.

For the carbonation depth test, the mortar specimens were introduced into a carbonation chamber under conditions of relative humidity of 55–65%, temperature of 23 ± 3 ◦C, and CO2 concentration of 5% ± 0.1%. After 56 days of CO2 exposure, a phenolphthalein pH indicator spray was used on the mortar fracture surface. The noncarbonated mortar surface showed a purple color due to its high alkaline pH. The carbonation depth was measured on the mortar fracture surface from the edge of the specimen to the purple area.

Despite the successful application of TiO2 photocatalysis to cement-based materials, an ideal method to determine the photocatalytic activity is still not available. The experimental conditions and data treatment differ in many aspects (light source, UV intensity, temperature, humidity, flow rate, characteristics of test samples, contaminant analyzed), even leading to noncomparable results [42]. The experimental method proposed by Spanish standard UNE 83321 EX was used in this research. This was aimed at evaluating the degradation of nitrogen oxide, in the gas phase, of inorganic photocatalytic materials contained in cement concretes by a continuous flow test method. For the measurements and calculations required in this test, the concentration of nitrogen oxides (NOx) was defined as the stoichiometric sum of nitrogen oxide (NO) and nitrogen dioxide (NO2).

Likewise, a sample was extracted from the center of each prismatic specimen mortar to analyze its photocatalytic power according to the standardized methodology through the reduction capacity of NOx. The test was carried out on the photocatalytic mortar family in addition to the reference mortar (M0).

#### **4. Results and Discussion**

#### *4.1. Consistency*

The average consistency results obtained after two perpendicular measurements are shown in Table 4. It can be seen that as the percentage of RS increases, the consistency of the mortar decreases, with a minimum value corresponding to 100% RS, mainly due to its higher water absorption. Silva et al. [38] stated that this loss in consistency can also be attributed to the greater angularity of the recycled particles, avoiding effective slippage between them. To compensate for this, an amount of water corresponding to the absorption could be added [31,36,49,53,54], although the mechanical properties

could be affected [35]. For this reason, the most suitable solution would be to add plasticizers to the mortar mix, thus increasing its consistency/workability, as advised by Ledesma et al. [14].


**Table 4.** Consistency values of mortar families (mm).

It was also observed that the photocatalytic mortar family showed less consistency in all cases compared to the conventional family. The lowest consistency, registered by PM100, may be responsible for its porous appearance (Figure 1) and difficult compaction.

**Figure 1.** Porous appearance of PM100 versus smooth appearance of other mixes.

#### *4.2. Mechanical Strengths*

The values registered for mechanical strength at 28 days are shown in Figure 2, which illustrates the comparison of results between the two families.

Figure 2a shows that for the family made with conventional cement and 40% of NS replaced by RS (M40), the compressive strength is equal to that corresponding to 20% replacement, approximately 46 MPa, which is only 10% less than the reference mortar (M0) at 52 MPa. For full replacement (M100), the drop in compressive strength increases to 16%, with an average value of 44 MPa. This agrees with Silva et al. [51], suggesting that as the RS content increases, the compressive strength remains similar to or larger than that of the control mortar. This may be due to the reduction of effective water as the percentage of RS is increased, as explained by López Gayarre et al. [35], or because of the greater number of fine particles that can fill the gaps [49]. In this investigation, the content of fine RS was higher than NS (12.4% vs. 2.26%), favoring the filling of voids in the mortar matrix and diluting the loss of mechanical strength. However, in most mortar mixes, there is a loss of compressive strength as the replacement percentage increases [34]. In all mortar mixes, the compressive strength is greater than the value recommended by the GB 28635-2012 standard [55] (average strength ≥ 30 MPa; any individual strength ≥ 25 MPa) and values obtained in other studies [19,56].

**Figure 2.** Mechanical strength comparison: (**a**) compressive strength; (**b**) flexural strength.

In all cases, the compressive strength obtained in the photocatalytic family is greater than that of their counterparts made with conventional cement. In this way, they start from an average resistance of 59 MPa for PM0, reaching 48 MPa for PM100.

Flexural strength can be correlated with other characteristics such as susceptibility to cracking and adhesive strength of mortar [51]. In this test (Figure 2b), a similar value of flexural strength was obtained for the 0%, 20%, and 40% replacement, 15 MPa, except for the M100 and PM100 families, for which a value of 13 MPa was found. This agrees with values obtained by Silva et al. [51], whose research used the same type and dosage of cement. Therefore, no improved behavior of the photocatalytic family was observed in this test. All values registered in the flexural strength test show very good performance of these mortars for their possible use in pavement, with values greater than 12 MPa in all mortars, exceeding the values obtained by other authors [56,57].

The differences between the strength of mixtures with 0% and 40% RS are practically insignificant, which confirms that increasing the percentage to 40% does not just mean decreased mechanical properties. This agrees with Ledesma et al. [14], who established a maximum replacement ratio of up to 50% of natural sand by mixed recycled sand without significantly affecting the hardened mortar properties.

Generally, a decrease in strength is observed as the percentage of recycled aggregate is increased, as reported in previous works. This decrease is more gradual and more noticeable in the photocatalytic family. However, the increase in RS up to 40% showed only a 7% decrease in strength with respect to 20% replacement, and up to 9% with respect to the reference mortar.

#### *4.3. Mineralogical Analysis*

The mineral phases formed in conventional and photocatalytic mortar families are shown in Figures 3 and 4, respectively. In all specimens, the main detected phase corresponds to quartz (SiO2; 33-1161) [52]. The intensity decreased in all specimen patterns because the NS was replaced with RS, which contains less silica [58]. The main phase in M100 was sanidine ((Na, K)(Si3Al)O8; 10-0357) [52].

The Portlandite phase (Ca(OH)2; 04-0733) [52] and ettringite (Ca6Al2(SO4)3(OH)12·26H2O; 41-1451) [52] were also observed in both families. The presence of these phases is an indicator of the Portland cement reaction [57], as shown by the mechanical performance of both mortar families (Figure 2). For this reason, the incorporation of RS into the mortar is compatible with the common Portland cement reaction.

The other detected phases in the mortars corresponded to silicates such as illite (KAl2Si3AlO10(OH)2; 02-0056) [52] and albite (Na(Si3Al)O8; 10-0393) [52], which is in agreement with authors such as Jiménez et al. [36] and Ledesma et al. [14]. Regarding carbonates, all mortars showed calcite (CaCO3; 05-0586) [52]; additionally, dolomite (CaMg(CO3)2; 36-0426) [52] was detected in M100 and all specimens of the photocatalytic mortar family. In both mortar families, it was observed that calcite (CaCO3; 05-0586) [52] had a greater presence as the replacement of NS by RS increased. This behavior has also been recorded by authors such as Gonçalves et al. [58], who studied the replacement of natural siliceous sand with recycled aggregate. Gypsum (CaSO4·2H2O; 33-0311) [52] was also detected in all specimens in both families, which could correspond to the RA or cement composition [36].

**Figure 3.** X-ray diffraction patterns of conventional mortar family.

**Figure 4.** X-ray diffraction patterns of photocatalytic mortar family.

Additionally, an SEM study was carried out, which supported the results obtained in the mineralogical study, as shown in Table 5. In the photocatalytic mortar family, no mineralogical phase attributed to Ti was detected; however, SEM confirmed the presence of TiO2 in all mortars with Ph. CEM I.


**Table 5.** Average percentage, by weight, of elements by scanning electron microscopy (SEM).

#### *4.4. Water Absorption by Capillarity*

The capacity to absorb water indicates the ability of an unsaturated porous material to absorb and drain water by capillary action, thus making it a suitable property to indirectly assess the durability of cementitious materials. Normally, higher water absorption by capillarity contributes to worse performance since it impairs the protection against external agents [51].

The results shown in Figure 5 indicate an upward trend in water absorption by capillarity as the percentage of RS increased in both families. However, for 20% of RS (M20), the values achieved in the conventional mortar family were even lower than those of the reference mortar (M0). The higher water absorption by capillarity of recycled mortars can be due to the high absorption of RS. Similar tendencies have been observed in studies by other authors [14,31,34,36].

**Figure 5.** Water absorption by capillarity comparison.

According to López Gayarre et al. [35], the reduction in the amount of effective water as the percentage of substitution is increased (greater water absorption) reduces the porosity of fresh mortar, and for this reason, recycled mortar could present this slight increase of water absorption by capillarity.

#### *4.5. Water Absorption Capacity, Bulk and Skeletal Density, and Open Porosity*

The water absorption capacity (Figure 6a) was almost the same for all mortars produced with photocatalytic CEM I. However, for the conventional mortar family, there was a slight increase with 40% of RS, and a more appreciable increase for 100% of RS of around 25%.

**Figure 6.** Comparison of results for (**a**) water absorption capacity and (**b**) open porosity.

Open porosity (Figure 6b) was higher as the RS replacement ratio increased. Poon and Cheung [59] explained this, affirming that materials with lower density lead to higher porosity of mortar blocks.

However, the density values (Figure 7a,b) were approximately the same in all cases, regardless of the type of cement or the amount of recycled aggregate. This differs from the values obtained in some studies [31,34], in which the density decreased as the replacement ratio increased. Other studies [31,49,60,61] found no significant differences in replacement ratios below 20–25%, while for higher replacement ratios, the lower dry density of FRA decreased the dry density. This result was attributed to the fact that a higher fine content (<0.063 mm) in RS allows for filling of voids at replacement ratios up to 10% of the hardened mortar.

**Figure 7.** Comparison of results of (**a**) bulk density and (**b**) skeletal density.

#### *4.6. Carbonation Depth*

The results of the carbonation test are shown in Figure 8. The carbonation depth was greater when the RS increased, although this increase was slight, up to 40%. For 100% replacement, the increase in carbonation depth was much greater than in the reference mortar, possibly due to the greater porosity of this mixture, as observed in the previous section.

**Figure 8.** Carbonation depth comparison at 56 days.

This behavior was similar in both families, although it was more appreciable in the photocatalytic mortar family. Specimens of the photocatalytic mortar family showed an increase between 7% and 25% (17% average) higher than their counterparts in the conventional mortar family. This property showed the same trend as reported by Moro et al. [50].

Of all the mortars, PM100 is the one with the highest CO2 absorption, with a 78% increase in carbonation depth as compared to the reference conventional mortar (M0), and a 66% increase as compared to the reference photocatalytic mortar (PM0). Thus, there is an added beneficial effect for the environment, since carbonation involves the absorption of CO2 from the air. This reacts with the Ca(OH)2 from cement hydrolysis, producing CO3Ca and immobilizing CO2. This process is detrimental to reinforced concrete, since it greatly affects durability due to the risk of corrosion of reinforcements by reducing the pH of the concrete. However, since mortars are not armored, that would not be a problem.

#### *4.7. Photocatalytic Activity Test*

According to the results of Figure 9, incorporating RS instead of NS slightly improved the decontaminating capacity. This can be explained by the greater porosity of the mortar as the percentage of recycled sand was increased, as observed in the previous sections, and agrees with the results reported by Poon and Cheung [59].

According to Spanish Standard UNE127197-1, most of the mixes were classified as category 1; that is, with decontaminating power, measured as the reduction of NOx varying between 4\$ and 6% (4.2%, 4.3%, and 4.8% for PM0, PM20, and PM40, respectively). However, the mix made with photocatalytic cement (PM100) and 100% RS showed a porous appearance and greater decontaminating capacity within category 2 (7.2%), which represents an increase of 71% of decontaminating power compared to mortar made with NS (PM0). This is consistent with Poon and Cheung [59], who indicated that

the porosity of the surface layer is important, as it effectively increases the area available for reacting with pollutants.

#### **5. Conclusions**

This research produced mortars with decontaminating capacity by introducing waste into the productive cycle and contributed to the implementation of the circular economy model. The results show suitable mechanical behavior despite the incorporation of recycled aggregates, since total replacement of NS by RS meant a decrease in compressive and flexural strength of only approximately 18%, with average values of 46 MPa and 13 MPa, respectively. The strength obtained with photocatalytic cement was slightly higher compared to its counterpart made with traditional cement. The properties of water absorption by capillarity, water absorption capacity, and open porosity showed a slight increase for 40% and 100% RS, while in the results obtained for 20% RS, the values obtained were very similar to those of the reference mortar. The bulk and skeletal density showed very similar values in the two families and for all replacement ratios.

The penetration of CO2 obtained in the carbonation test carried out showed a clear benefit with the incorporation of recycled sand. This could be a positive aspect, as it reduces the carbon footprint in the environment.

From a photocatalytic point of view, the incorporation of up to 40% RS slightly favored the elimination of NOx, but the mortar with 100% RS had a significant increase relative to the conventional mortar, moving to a better classification (category 2). These mortars were more porous than conventional mortars, facilitating the entry of light into the interior and, consequently, the elimination of polluting gases. The results obtained add value to the use of recycled aggregates, clearing up the uncertainties that still exist in the use of this material.

Therefore, a photocatalytic mortar with 100% mixed recycled sand is proposed in order to produce the greatest environmental benefits, due to the greater absorption of CO2 and NOx and the greater use of recycled aggregates. Also, the strength is slightly lower but compatible with its use in low-requirement applications in contact with the atmosphere, such as pavement blocks or cladding mortar.

The findings of the present study prove that reducing natural sand mining, minimizing energy consumption and CO2 emissions, reducing global warming, preventing illegal deposition and landfilling of the fine fraction of CDW, and complying with the limits of the European Waste Framework Directive are possible in order to achieve and promote cleaner production in the construction sector (eco-efficiency).

Future lines of research are intended to improve the photocatalytic capacity of mortars based on the good mechanical behavior of the mortars studied in this research by studying the influence of the content and origin of fine aggregates, different aggregates and their nature, reducing the amount of cement due to the good mechanical behavior obtained, or increasing the amount of water according to the percentage of increased RS, or pre-saturating it.

**Author Contributions:** Conceptualization, A.B., A.L.-L., and J.A.; Methodology, A.B. and A.L.-L.; Software and validation, A.B. and A.L.-L.; Formal analysis, A.P.G.; Investigation, A.L.-U. and A.P.G.; Data curation, A.B. and A.L.-L.; Writing—original draft preparation, A.B. and A.L.-L.; Writing—review and editing, A.B., A.L.-L., and J.A.; Visualization, A.L.-U. and A.P.G.; Supervision, A.B., A.L.-L., and J.A.; Funding acquisition, J.A.. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Industrial Low-Clinker Precast Elements Using Recycled Aggregates**

**Carlos Thomas 1,\*, Ana I. Cimentada 1, Blas Cantero 2, Isabel F. Sáez del Bosque <sup>2</sup> and Juan A. Polanco <sup>1</sup>**


Received: 31 August 2020; Accepted: 21 September 2020; Published: 23 September 2020

**Abstract:** Increasing amounts of sustainable concretes are being used as society becomes more aware of the environment. This paper attempts to evaluate the properties of precast concrete elements formed with recycled coarse aggregate and low clinker content cement using recycled additions. To this end, six different mix proportions were characterized: a reference concrete; 2 concretes with 25%wt. and 50%wt. substitution of coarse aggregate made using mixed construction and demolition wastes; and others with recycled cement with low clinker content. The compressive strength, the elastic modulus, and the durability indicator decrease with the proportions of recycled aggregate replacing aggregate, and it is accentuated with the incorporation of recycled cement. However, all the precast elements tested show good performance with slight reduction in the mechanical properties. To confirm the appropriate behaviour of New Jersey precast barriers, a test that simulated the impact of a vehicle was carried out.

**Keywords:** recycled concrete; low clinker cement; precast; mechanical properties; physical properties; New Jersey barriers

#### **1. Introduction**

Construction and demolition waste (CDW) is non-hazardous, inert waste generated in any construction, rehabilitation or demolition work. The industrial and construction sectors generate practically the same amount of non-hazardous waste (industry 37,417 kt<sup>ᑽ</sup> and construction 35,869 kt<sup>ᑽ</sup> ) in Spain [1]. The European Commission estimates that the volume of CDW comprises one third of all waste generated in the European Union, which constitutes the largest waste stream [2]. Recycling this CDW would lead to more sustainable growth, replacing a linear economy based on use of materials with a more circular economy. This is important, as aggregates are the second-most-used raw material by humans, behind only water [3]. There is European legislation to encourage recycling CDW [4] and many countries have specific norms for the use of recycled aggregates (RA) for concrete [5–8]. In addition, the use of RA could lead to cheaper concrete [9].

Several studies have corroborated that the inclusion of RA produces concrete with a lower density and increased heterogeneity [10–12]. RA normally has a higher porosity than natural aggregate (NA) [13]. In a fresh state, Silva et al. [11] concluded that recycled aggregate concrete (RAC) is less workable and, to achieve a workability equivalent to that of NA, RA could be pre-saturated, or water added during mixing to compensate [14]. However, the incorporation of completely saturated aggregates might cause an excessive water supply [15,16]. Once the RAC hardens, these aggregates make

the concrete more susceptible to detrimental environmental effects, resulting in a lower durability [17,18], which should be taken into consideration. Consequently, Annex 15 of the Spanish Instruction for Structural Concrete EHE-08 [19] and other studies [14,20] propose solutions, such as increasing the cement content, reducing the water/cement ratio, or increasing the coating thickness in the case of reinforced concrete.

Generally, it is known that the incorporation of RA into concrete reduces its mechanical properties [21,22], due to the presence of contaminants such as plastics, glass, adhered mortar, etc., ref. [23] and the type of source material (crushed concrete, ceramic or mixed) of the RA [24–26]. The elastic modulus of RAC is lower than that of conventional concrete [15], reaching 45% less for 100% replacement [25]. The results obtained in the characterization of RAC with intermediate replacements present greater variation of results [20]. Other authors have demonstrated the viability of other types of recycled aggregates from waste, such as steel slag [27]. Moreover, the RA affects the fatigue behavior of the concrete [28–32], showing a greater loss of properties than with the static properties. Further research has evaluated the recycling of concrete which incorporates RA [33,34].

With regard to precast concrete elements, it should be noted that, according to the ANDECE (National Association of the Prefabricated Concrete Industry, based in Spain), although the initial cost of elements is higher, the final cost is lower [35]. Other studies such as López-Mesa et al. [36] indicate an almost 18% higher cost of precast slabs versus in situ slabs; although the former have a lower environmental impact and the quality may be higher. Normally, precast elements have a quality seal guaranteeing their properties. Due to a manufacturing process with complete exhaustive control, precast slabs can be: tailored with special properties more easily as they are not manufactured on site; designed with flexibility difficult to achieve in-situ; and incorporate RA in their fabrication. In the case of precast elements using RA, a lower density and strength is observed [37]. Poon et al. [37] investigated the factors that affect the properties of precast concrete blocks with RA, concluding that the compressive strength increases with the reduction in the aggregate/cement ratio (A/C), and that the water absorption of concrete blocks is significantly related to the absorption capacity of the aggregate. Katz [21] investigated the use of precast elements at different ages to produce RA for new precast elements, concluding that the mechanical properties (strength, modulus of elasticity, etc.) when using this type of aggregate in concrete, resemble those when using lightweight aggregates, such as those manufactured using fly ash.

This paper presents the effect on physical and mechanical properties of six types of mixes with different degrees of substitution. The physical properties and durability of these concretes will be analyzed first, then the mechanical properties will be assessed. Finally, the behavior of precast elements will be addressed.

#### **2. Materials and Methodology**

The natural siliceous aggregate used in this study is present in three different sizes: 6/0 mm (NS), 12/6 mm (NG-M), and 22/12 mm (NG-C). Mixed recycled aggregates (MRA) were used by substituting NG-M for MRA-M and NG-C for MRA-C. These MRA were obtained from CDW and were principally made up of concrete and mortars (≈ 45%), unbound aggregate, and natural stone (≈ 45%). Figure 1 shows the different size grading for each aggregate.

**Figure 1.** Grading of the aggregates.

Table 1 displays physical and mechanical properties: where *SSS* is the saturated dry surface density according to EN 1097-6 [38]; *A* is the water absorption by weight according to EN 1097-6 [38]; *LA* is the Los Angeles index according to EN 1097-2 [39]; and *FI* is the flakiness index according to EN 933-3 [40].


**Table 1.** Physical and mechanical properties of the aggregates.

The conventional cement (OPC) was CEM I 42.5 R, and the low clinker content cement (RC) was constituted of 75% CEM I 42.5 R and 25% ceramic waste from CDW. The tests performed with the cement revealed a compressive strength 20% higher in the case of OPC.

Mixing the aggregates in different proportions with the two existing types of cement produced six concrete mixtures, as shown in Table 2. HP signifies a combination of natural aggregates and conventional cement. HPR is a mixture of natural aggregates and low clinker content cement. HR25 and HR50 were fabricated with conventional cement and substitutions of NA by 25%wt. and 50%wt. proportions of RA, respectively. Finally, HRR25 and HRR50 were obtained by amalgamating low clinker content cement with natural aggregates, substituted by 25%wt. and 50%wt. of recycled aggregates accordingly.


**Table 2.** Concrete mix proportions (by m3).

#### *2.1. Physical and Mechanical Properties*

Densities were obtained according to EN-12390-7 [41]. Sub-specimens (10Ø × 10 cm) obtained by cutting 10Ø × 20 cm cylindrical specimens were used. The porosity coefficient is the result of comparing the absorbed water and specimen volume, while the absorption coefficient is the result of comparing the absorbed water and specimen weight. Compressive strength was determined using 10Ø × 20 cm cylindrical specimens according to EN-12390-3 [42], with an application strength rate of 0.5 MPa/s. Elastic modulus was determined with 10Ø × 20 cm cylindrical specimens according to EN-12390-13 [43], at a strength rate of 0.5 MPa/s.

#### *2.2. Durability*

A water penetration test was performed according to EN-12390-8 [44]. Sub-specimens (10Ø × 10 cm) obtained by cutting 10Ø × 20 cm cylindrical specimens were used. The samples were subjected to a pressure of 5 bar for 72 h. After 72 h water penetration under pressure, it was necessary to analyze how deep the water reached. To be able to observe the interior of the sample, it had to be opened. During this research, the Brazilian method (or indirect tensile strength method) was used to open the sample and analyze its interior. In general, when a cylindrical specimen is subjected to tension along its generatrix, it breaks into two halves, which allows the interior to be analyzed. Once the specimen had been opened, it was possible to measure the penetration depth of the water into the porous concrete. This technique also provided another interesting result: the indirect tensile strength of the concrete. For the determination of oxygen permeability, UNE-83981 [45] was taken as a reference. The 10Ø × 20 cm cylindrical specimens were cut to discard the upper and lower face obtaining a new sample of 10Ø × 10 cm. Silicone was impregnated perimetrically in the samples so that the oxygen could only pass longitudinally. A regulated oxygen pressure was applied on the upper face. Digital flow meters registered the oxygen escaping from the lower face.

#### *2.3. Precast Element Preparation*

Two different types of precast elements were manufactured: unreinforced concrete ditches and steel-reinforced New Jersey barriers. Both were manufactured with an industrial concrete mixer, poured in metallic molds and vibrated by hand (Figure 2). In the case of reinforced concrete, reinforcements were set into the mold before the pouring of concrete. In both cases, precast elements were unmolded and cured at ambient temperature.

**Figure 2.** Precast element manufacturing sequence.

#### *2.4. Precast Element Mechanical Characterization*

Concrete ditches have approximate measurements of 50 × 50 × 15 cm. In order to characterize concrete ditches, the tests were carried out by bending. The horizontality of the set was verified, and force was applied by a roller (10Ø × 22 cm) in the central section with a displacement rate of 0.1 mm/s (Figure 3).

**Figure 3.** Precast element characterization (concrete ditches left, New Jersey barriers right).

New Jersey barriers have a section with approximate measurements of 47 × 80 cm and a length of 100 cm. In order to characterize New Jersey barriers, a small crane was used to support the precast element on steel beams. These steel beams were placed at one end to correct the inclination of the face on which the test was to be performed, achieving horizontality on that face (Figure 3). The test consisted in applying a stress with a roller (3Ø × 40 cm). The time of the test was very short (0.1–0.2 s) to simulate an impact. The strength and displacement data of the actuator were recorded during the test.

#### **3. Results and Discussion**

#### *3.1. Physical Properties*

Figure 4 shows the relative and saturated densities of the concretes. As demonstrated, the density decreases as the percentage of NA replaced by RA increases. This is due to the lower density of RA. It also becomes clear that the use of this RC does not affect density significantly.

**Figure 4.** Density vs. RA content.

Figure 5a shows porosity, and Figure 5b shows the absorption coefficient vs. substitution of NA by RA. A decrease in both properties is found in the concretes containing OPC as the percentage of replacement of aggregate increases. However, in the case of concrete made with RC, both properties increase as the percentage of RA increases. This may be because this type of cement interacts more with RAs of different nature, making it difficult to fill all the gaps amongst aggregates. Alternatively, it may be because the RA is able to absorb more water during kneading, causing a small deficit in this type of cement, which is very susceptible to variations in the water dosage. It is possible that there may be another reason that has not been identified.

**Figure 5.** Porosity (**a**) and absorption coefficient (**b**) vs. RA content.

#### *3.2. Compressive Strength and Modulus of Elasticity*

Figure 6a shows the compressive strength-strain curves for each concrete at 160 days. Several studies [25,46,47] show that the concrete's compressive strength decreases with the degree of substitution of RA for NA, but in strain terms, concretes show similar values around 2500 μm/m for the failure. The exception is the HRR50 mix, which exceeds the values of the rest by almost 1000 μm/m. Figure 6b shows the same mixtures but at an age of 365 days. The decrease in strength may also be due to

the randomness of the type of RA and its distribution into the mortar matrix, which causes greater uncertainty than conventional mixtures.

**Figure 6.** Compressive strength-strain at 160 (**a**) and 365 (**b**) days.

Table 3 shows the different values of compressive strength obtained at different ages.


**Table 3.** Compressive strength at different ages.

Table 4 displays the modulus of elasticity, and shows that when using RC, the decrease in the elastic modulus is around 4%. The substitution of 25% by RA implies a decrease in elastic modulus of 5.6%, while the substitution of OPC in this case does not seem to have an influence. In the case of replacing 50% of aggregate by RA, the influence of the substitution of OPC by RC is meaningful, decreasing the elastic modulus by 15%. As for the loss of elastic modulus over time, a greater influence of the cement is observed than the type of aggregate, with a limit that tends to an asymptotic value of around 27 GPa.


**Table 4.** Modulus of elasticity.

Some organizations such as EHE-08, ACI, and Eurocode present their expressions to predict elastic modulus at 28 days from the compressive strength. In Expressions (1)–(3): *E* is elastic modulus at 28 days [GPa] and *f* <sup>28</sup> is the compressive strength at 28 days [MPa].

EHE-08 [48]

ACI [49]

$$E = 8.5 \sqrt[3]{f\_{28}} \tag{1}$$

$$E = 4.7\sqrt{f\_{28}}\tag{2}$$

Eurocode 2 [50]

$$E = 22 \left( f\_{28} / 10 \right)^{0.3} \tag{3}$$

These expressions can be used to obtain the predictions and comparisons, with the experimental results shown in Table 5. The ACI method fits quite well in most cases but predicts higher values when the percentage of substitution is 50%. The EHE-08 method is safer, although when the substitution is 50% and the OPC is replaced by RC, higher values are produced due to the heterogeneity of the RA affecting the compressive strength. These types of expressions only satisfactorily fit ordinary concrete models.

**Table 5.** Elastic modulus obtained with different expressions.


Figure 7 shows that from approximately 48 MPa, concrete with RA achieved the same compressive strength as concrete with OPC. RA concrete increases its elastic modulus significantly. This might be due to the addition of a new variable, such as RA compared with OPC, which is much more standardized throughout its production process.

**Figure 7.** Compressive strength vs. modulus of elasticity.

#### *3.3. Oxygen and Water Permeability*

Figure 8a shows the oxygen permeability and Figure 8b shows the maximum penetration of water vs. percentage of substitution, respectively.

**Figure 8.** Oxygen permeability coefficient (**a**) and water penetration depth (**b**).

The oxygen permeability coefficient increases with the substitution of the NA by RA. This behavior has been reported in some studies, such as Ismail et al. [51], Medina et al. [52], and Thomas et al. [14]. This increase is higher in concrete with RC than OPC; the type of cement being used is an important factor.

The penetration of water increases with the increase in RA substitution. With these results, only HP and HPR comply with the standard EHE-08 [48] for structural concrete in the case of IIIa, IIIb, IV, etc. environment exposition, which requires an average penetration depth of 30mm, and maximum penetration depth of 50 mm. Penetration of water is related to typology and distribution of the RA, and its impurities with high absorption coefficients.

Figure 9 shows cross-sections of concrete where different colors can be seen. These are caused by the RC in HPR and HRR50 mixtures, and some kind of RA and impurities (such as wood or fired clay) in HRR50 mix.

**Figure 9.** Concrete specimen sections.

#### *3.4. Testing Precast Elements*

Figure 10a shows the results of flexural tests on concrete ditches. It can be observed that the concrete composed of RC and RA (HRR50) behaves similarly to HP concrete, which is consistent with the results of splitting tensile strength shown in Table 6. Figure 10b shows the results of the impact test on reinforced precast New Jersey barriers, in which the force applied by the test machine and the position of the actuator are recorded. As expected, the concrete with OPC and NA displayed superior mechanical behavior than concrete with RC and RA. HRR50 could resist only 60% of the force, and 66% of the displacement that HP resisted.

**Figure 10.** Mechanical characterization of precast elements: Bending test on ditches (**a**), impact test on barriers (**b**).



Figure 11 shows the results of the test performed with both types of precast elements. Different sections of cracks in OPC and RC concrete ditches, and the fissure produced in a New Jersey barrier are visual results of the tests.

**Figure 11.** Precast test and cracking.

*Appl. Sci.* **2020**, *10*, 6655

Equation (4) indicates whether a New Jersey barrier could withstand the perpendicular impact of a vehicle. Velocity and mass are variables, and it would be necessary to incorporate a restitution coefficient in order to avoid the elastic impact.

This coefficient relates the velocity before impact with the velocity after collision, considering the barrier is without velocity before and after impact.

$$\mathbb{C}\_{R} = -\frac{V\_{1f} - V\_{2f}}{V\_{1i} - V\_{2i}}; \text{when } V\_{2f}, V\_{2i} = 0 \to \mathbb{C}\_{R} = -\frac{V\_f}{V\_i} \tag{4}$$

García and Cabreiro [53] proposed a method for obtaining the coefficient of restitution based on experimental processes in "*Use of dynamic models in the investigation of road accidents*" (text in Spanish), for which they suggested two equations:

$$\mathcal{C}\_R = 0.45 \cdot e^{(-0.040278 \cdot v)}, \text{ For } v < 54 \text{ km/h} \tag{5}$$

$$C\_R = 0.45 \cdot e^{(-0.015278 \cdot v)}, \text{ For } v \ge 54 \text{ km/h} \tag{6}$$

With Equations (4)–(6), considering the maximum force that a barrier resists, and the duration of the impact as 0.1 s, Equations (7) and (8) are obtained, shown in Figure 12.

$$m = \frac{0.1 \cdot F}{-\left(\frac{v\_i}{3.6}\right) \cdot \left(0.45 \cdot e^{(-0.040278 \cdot v)} + 1\right)}, \text{ For } v < 54 \text{ km/h} \tag{7}$$

$$m = \frac{0.1 \text{-} F}{-\left(\frac{v\_1}{3.6}\right) \cdot \left(0.12 \cdot e^{(-0.015278 \cdot v)} + 1\right)}, \text{ For } v \ge 54 \text{ km/h} \tag{8}$$

**Figure 12.** Simulated behavior of reinforced barriers.

These curves are conservative, as the barrier can withstand strains that absorb energy before cracking, and the parapet would not always be immobile (they are only anchored to the ground on viaducts).

#### **4. Conclusions**

Characterization tests on concrete specimens and precast elements have been carried out using low-clinker cements and recycled aggregates, obtaining the following conclusions. Firstly, the physical-mechanical properties of mixed recycled aggregates are suitable for the manufacture of

concrete and precast elements when the medium and coarse fraction is used. Secondly, the use of mixed recycled aggregates causes a loss of density and compressive strength slightly higher than that which occurs when using recycled concrete aggregates. Recycled concretes made from low-clinker cement are slightly more porous than concretes made with ordinary Portland cement. Finally, regarding the mechanical properties of recycled concrete, a loss of around 10% of the compressive strength is observed when using low-clinker cement. In addition, recycled concrete made with ordinary Portland cement evolves slightly more when over 1 year of curing has elapsed.

**Author Contributions:** Conceptualization, C.T., A.I.C., J.A.P.; methodology, C.T., A.I.C., J.A.P., I.F.S.d.B., B.C.; validation, C.T., I.F.S.d.B., B.C.; formal analysis, C.T.; investigation, C.T., A.I.C., J.A.P.; resources, C.T., J.A.P.; writing—original draft preparation, C.T., I.F.S.d.B., B.C.; writing—review and editing C.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by SODERCAN, S.A. (SODERCAN/FEDER) and BIA2013-48876-C3-2-R awarded by the Ministry of Science and Innovation.

**Acknowledgments:** The authors would like to express our gratitude to Jaime de la Fuente and César Medina for their support and participation in part of the project.

**Conflicts of Interest:** Authors declare no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Mechanical Properties and Flexural Behavior of Sustainable Bamboo Fiber-Reinforced Mortar**

#### **Marcus Maier 1,\*, Alireza Javadian 2,\*, Nazanin Saeidi 2, Cise Unluer 3, Hayden K. Taylor <sup>4</sup> and Claudia P. Ostertag <sup>5</sup>**


Received: 31 August 2020; Accepted: 17 September 2020; Published: 21 September 2020

**Abstract:** In this study, a sustainable mortar mixture is developed using renewable by-products for the enhancement of mechanical properties and fracture behavior. A high-volume of fly ash—a by-product of coal combustion—is used to replace Portland cement while waste by-products from the production of engineered bamboo composite materials are used to obtain bamboo fibers and to improve the fracture toughness of the mixture. The bamboo process waste was ground and size-fractioned by sieving. Several mixes containing different amounts of fibers were prepared for mechanical and fracture toughness assessment, evaluated via bending tests. The addition of bamboo fibers showed insignificant losses of strength, resulting in mixtures with compressive strengths of 55 MPa and above. The bamboo fibers were able to control crack propagation and showed improved crack-bridging effects with higher fiber volumes, resulting in a strain-softening behavior and mixture with higher toughness. The results of this study show that the developed bamboo fiber-reinforced mortar mixture is a promising sustainable and affordable construction material with enhanced mechanical properties and fracture toughness with the potential to be used in different structural applications, especially in developing countries.

**Keywords:** fiber-reinforced; natural fibers; bamboo; sustainable mortar; mechanical characterization; by-products; toughness

#### **1. Introduction**

The construction industry is a major consumer of energy and raw materials and contributes immensely to environmental pollution, especially to greenhouse gas (GHG) emissions [1]. Since the 1970s, annual GHG emissions have steadily increased and reached 53.5 GtCO2e in 2017 [2]. Within the construction industry, concrete is the dominant building material with a global production of <sup>20</sup> <sup>×</sup> <sup>10</sup><sup>12</sup> kg per annum which exceeds the amount of all other construction materials combined. With an increase in the demand for new infrastructure demonstrated by developing countries, the use of Portland cement (PC), the main component in concrete, has been rising rapidly [3]. Accordingly, the global use of PC has increased from 2.22 to 4.10 Gt/year within the past decade. The production of PC accounts for ~5% of the global anthropogenic CO2 emissions [4]. Moreover, concrete is commonly reinforced with steel, whose production involves high energy emissions and consumption of fossil fuels that additionally contributes to CO2 emissions. Furthermore, the fast pace of development in many developing countries has led to an increased demand for reinforced concrete for housing

and infrastructure projects. Unfortunately, the majority of developing countries lack the resources to produce their own cement and steel for the production of reinforced concrete elements which forces them to import the majority of their needs from highly industrialized countries, and as a result of the import surge, trade deficits, economic slow-downs, and loss of jobs are prominent in those countries. Besides the economic challenges from the cement and steel import, environmental issues also need to be addressed. The construction industry is facing an urgent need for the use of sustainable materials incorporating locally available renewable resources as well as industrial by-products with lower environmental impacts.

One potential material is fly ash, a by-product of the combustion of coal, oil and biomass. Fly ash contains Silicon dioxide (SiO2) and Aluminum oxide (Al2O3) as major components that can contribute to the hydration of cement. Furthermore, low-cost and renewable materials such as bamboo and wood can be found in abundant supply in many developing countries, where the bamboo and wood industry produce a large number of waste products. Replacement of cement with by-products such as fly ash or bamboo and wood waste can enable the reduction of the carbon footprint associated with the cement industry and improve the mechanical and thermal properties of the developed formulations. Other performance aspects such as the ductility of these mixes could be further enhanced via the use of other renewable materials such as natural fibers to avoid the brittle failure that is characteristic of plain concrete.

Previous studies [5–11] revealed improvements in the mechanical properties and durability of concrete mixtures, in which PC was partially replaced with wood ash or fly ash. Substituting aggregates with wood process waste such as wood chips, flax or hemp was also shown to enhance the mechanical or thermal properties of concrete mixtures [11–18]. Further research on fiber-reinforced concrete reported that the addition of synthetic fibers—such as polypropylene (PP), polyethylene (PE), polyvinyl alcohol (PVA)—or steel fibers could increase the fire resistance, ductility, tensile strength, impact resistance and toughness of concrete mixtures [19–22]. However, synthetic fibers, which are mainly derived from petroleum-based sources, and steel fibers require energy-intensive and expensive production processes. In contrast, natural fibers, such as those obtained from wood and bamboo industry by-products, can provide a low-cost and sustainable alternative for the construction industry. Challenges of resource scarcity and the negative environmental impacts of synthetic fiber production have led many researchers to search for alternative, green, sources of fibers for the production of fiber-reinforced concrete. Natural fibers represent a sustainable source of raw materials from renewable resources and can help to alleviate the need for synthetic fibers. While there is growing interest in the use of wood fibers to enhance the mechanical behavior and fracture toughness of concrete [23–25], there has so far been relatively little investigation of the use of bamboo fibers for this purpose. Only a few studies [26–30] investigated the performance of bamboo fiber-reinforced concrete and mortar mixtures through a series of mechanical tests. The bamboo fibers in those studies were obtained from bamboo forests and were subsequently processed as fibers for concrete mixtures. Furthermore, the studies showed that only concrete's tensile property had obvious improvement when bamboo fibers were added, while the enhancement to the compression property and flexural property was not obvious. The studies on the application of bamboo fiber-reinforced concrete and mortar mixtures are rather limited. Both bamboo fibers and fly ash present a great opportunity as sustainable and affordable replacements for cement and steel for developing countries. Bamboo belongs to the botanical family of grasses and shows high resistance to tensile stresses. The tensile strength of natural bamboo is superior to that of wood. This attribute marks bamboo as an attractive option to incorporate into fiber-reinforced concrete, especially in developing countries where demand for reinforced concrete is growing rapidly [31–33]. Bamboo is a gigantic grass, which belongs to the angiosperms (seed-bearing vascular plants) group and monocotyledon (flowering plants) subgroup. Bamboo attains maturity in 3 to 5 years, in favorable contrast to wood, which takes at least 20 years, depending on the species [34]. The growth behavior of bamboo culm and the extreme wind loads it has to sustain during its life cycle require a precise mechanical adaptation to the environment. Therefore, material optimization

has to be achieved effectively from the bamboo fibers and their cell structures. This results in an optimized microstructure with superior material performance when compared to various wood species. Furthermore, bamboo can directly address global warming as it rapidly grows and sequesters carbon in biomass and soil faster than almost any wood species. The main components of bamboo culms are cellulose, hemicellulose and lignin. The minor components are resins, tannins, waxes and mineral salts. However, the percentage of each component differs from species to species and depends on the conditions of bamboo growth and the age of the bamboo, as well as the location of the section on the culm [31,34]. In general, cellulose in bamboo culms accounts for more than 50% of the bamboo chemical components. After cellulose, lignin is the next largest component, and normally accounts for more than 20% of the bamboo's mass. Bamboo displays a round-shaped cell cross-section, in contrast to the nearly rectangular and relatively large cells of wood species. Furthermore, bamboo culms have a particular multi-layered cell wall structure with alternating thick and thin layers of fibers, unlike the typical three-layered cell wall of wood species which have a structure with a dominating middle layer [34–36].

In recent years various methods have been developed to employ bamboo through new processing technologies for the fabrication of high-performance bamboo-composite materials in such a way that the inherent mechanical capacities of the fibers are retained, while the durability issues, specifically water absorption, swelling, shrinking and chemical resistance, of the composite could be enhanced for application as structural elements in buildings [31–33,37]. The bamboo-composite materials display high mechanical properties and have been used as either reinforcement in concrete, replacing steel or as structural elements in the form of a beam or column. However, the process through which natural bamboo culms transform into bamboo-composite materials employs only certain sections of the culms and therefore the remaining parts usually become part of the waste of the production process which can be safely utilized for applications as sustainable and affordable fibers in fiber-reinforced mortar.

Therefore, the objective of this study is to develop a sustainable and affordable mortar mixture incorporating by-products and renewable materials (i.e., fly ash and bamboo fibers) that show improved mechanical properties and fracture behavior and could be employed for the construction of low-cost and low-rise housing solutions in developing countries. The developed mixture was characterized via compression, splitting and bending tests, whereby the fracture properties including toughness and absorption energy were also assessed. The findings generated through this work set the foundation for further research on bamboo fibers and bamboo-reinforced concrete and mortar mixtures for structural applications.

#### **2. Materials and Methodology**

#### *2.1. Bamboo Plant*

There are about 1200 species of bamboo under some 90 genera. The physical and mechanical properties of bamboo culms are correlated with the specific gravity and the fiber content. Therefore, the physical and mechanical properties of bamboo differ from species to species and even within the same species or same culm, due to changes in chemical composition as well as specific gravity [38]. For the purpose of this study *Dendrocalamus asper*, known as Petung Putih bamboo, was selected from a bamboo forest on the Java island of Indonesia. *Dendrocalamus asper* is widely used for low-rise and low-cost housing across Indonesia.

#### 2.1.1. Bamboo Fibers

A new processing technology was developed to process bamboo culms into fibers that were suitable for use in a novel engineered bamboo composite materials [33]. Accordingly, the fibers were obtained by processing entire bamboo culms. They were then added to epoxy resin and fabricated into high-tensile-strength bamboo composite materials by using a hot-press fabrication method. The process yielded an engineered bamboo composite material, which was then cut into different sizes to be used in concrete as reinforcement in place of steel bars in previous studies [39,40].

Throughout this process, some portions of the fibers are not used and remain as the waste by-products of the fabrication process. This study adopts the use of these waste portions of these fibers to reinforce the mortar matrix. In an earlier study [37], the mechanical properties of *Dendrocalamus asper* bamboo from Indonesia were investigated with respect to culm physical properties including culm diameter, wall thickness, height, moisture content and specific density. Correlations were drawn between the culm's physical properties and the resulting mechanical properties, including tensile strength, modulus of rupture and modulus of elasticity in flexure and tension.

Bamboo fibers were ground and sieved with a sieve tower using sieve sizes of 1 mm and 500, 300 and 125 μm. The final bamboo fibers used in this study were those remained in the 500 μm and 300 μm sieves (Figure 1) and are referred to as "500 μm" and "300 μm" fibers throughout this study, respectively. Bigger and smaller sieves were used for a better separation of the fiber sizes. This ensured more constant fiber sizes and avoided clogging of the 500-micron sieves with bigger fibers. Trail mixes with 212 μm fibers showed no promising results and were therefore not considered in this article.

**Figure 1.** Bamboo fibers. (**a**) Raw waste material; (**b**) 300 μm-diameter fibers after grinding and sieving; and (**c**) 500 μm-diameter fibers after grinding and sieving.

#### 2.1.2. Bamboo Fiber Treatment

A variety of treatments to improve the durability and the bonding of natural fibers embedded in concrete and mortar matrix are available, with different levels of complexity including but not limited to cement surface coating, lime surface coating, cement–lime coating and oil impregnation [41–44]. For the present study, two treatments were selected for the bamboo fibers. To enhance the durability, lignin was partially removed by simply heating the bamboo fibers in water at 85 ◦C for 72 h and drying them at 80 ◦C for 24 h [45]. In addition, the bond between the fibers and the mortar matrix was improved by an alkaline treatment, which involved the stirring of the bamboo fibers in a lime (Ca(OH)2) solution for 2 h. The solution contained 40 g of lime per liter of water. This suspension was stirred repeatedly during the entire treatment duration to avoid the sedimentation of the fibers and any undissolved lime particles. After the lime treatment, the bamboo fibers were dried at 80 ◦C for 24 h. This treatment is known to modify the surface of the fibers and improve their mechanical strength [45].

#### 2.1.3. Characterization of the Bamboo Fiber Geometry

The fiber geometry was assessed by microscopy, during which a total of 200 fibers within each size category were measured with an optical microscope and analyzed with the Zeiss software "ZEN 2 (Blue edition). Table 1 presents the mean value of the length and diameter as well as the standard deviation obtained from these microscopy measurements. The mean value *x* is calculates by *x* = <sup>1</sup> *n <sup>n</sup> <sup>i</sup>*=<sup>1</sup> *xi* whereby the standard deviation is calculated by σ = <sup>1</sup> *n <sup>n</sup> <sup>i</sup>*=1(*xi* − *x*) 2 . The intentionally chosen simplified manufacturing and sieving process leads to the relatively high standard deviation of the fibers, but allows replication with minimal equipment and labor costs, especially in developing countries with limited access to such facilities. The average density of the bamboo fibers used in this study was 840 kg/m3 and was obtained from [31].


**Table 1.** Geometry of the bamboo fibers.

#### *2.2. Mortar Mixtures*

A mortar mixture containing 547 kg/m3 of Ordinary Portland cement, CEM I, according to Singapore Standard SS EN 197-1 [46] and 656 kg/m3 of fly ash Class C, according to ASTM C 618 [47], was used as a control mixture. Different contents of bamboo fibers (4, 6, and 8% by volume of the concrete) were added to investigate the resulting mechanical and fracture behavior. Fine sand with a maximum grain size of 2 mm was used as aggregates. The sieve curve of the sand is shown in Figure 2. To adjust the workability of the mixtures, a superplasticizer (ACE 8538, BASF), was added to the mixture. To ensure a constant mix procedure and fiber distribution trail mixes were prepared to evaluate mixing time and the amount of superplasticizer needed to achieve good workability. At first Cement, Fly Ash and aggregates were mixed for 90 s before water containing superplasticizer was added and mixed for 5 min. Finally, the bamboo fibers were gradually added and the constituents were mixed for another 3 min before filled into the molds. To evaluate the workability of the mixtures a flow table test was carried out according to ASTM C1437 [48]. All mixtures showed a spread of more than 255 mm diameter (maximum diameter of the flow table test apparatus according to [49]) and therefore could be filled well in the molds and compacted.The compositions of all mixtures used in this study are provided in Table 2.

**Figure 2.** Particle size distribution of the fine aggregates used in this study.


**Table 2.** Compositions of mortar mixtures.

<sup>1</sup> Density of bamboo fibers: 840 kg/m<sup>3</sup> obtained from [31].

#### *2.3. Test Procedure*

The experimental program included compressive and splitting tensile tests for the mechanical characterization of the prepared mixtures. A four-point bending test was also used to evaluate the post-cracking behavior of the bamboo fiber mixtures. These were assessed on 50 mm cubes and beams with dimensions of 50 × 50 × 300 mm. Compression tests were performed at a loading rate of 55 kN/min, following the ASTM C109 standard [50]. Splitting tests were performed in accordance to ASTM C496 [51], at a loading rate of 1.0 mm/min. Four-point bending tests were performed according to ASTM C1609 [52] on prisms with a span, *L*, of 150 mm. The load was applied in the 1/3 points at a rate of 0.2 mm/min. The deflection of the beam was measured with two Linear Variable Differential Transducers (LVDTs) at the mid-span (Figure 3).

**Figure 3.** Four-point bending test performed according to ASTM C1609 [52].

With the recorded load–deflection data at hand, the load at the limit of proportionality, Fmax, and the corresponding modulus of rupture (MOR) could be assessed according to the RILEM TC 162-TDF [53] recommendation. According to the RILEM recommendation, the limit of proportionality is equal to the maximum load recorded up to 0.05 mm. The modulus of rupture, corresponding to the maximum force in a four-point bending test, can be calculated by the following expression:

$$\text{MOR} = \text{F}\_{\text{max}} \ l/(b \, h^2) \,\text{(MPa)} \tag{1}$$

where *b*, *h* and *l* are the width, height and span of the tested specimens and equal to 50 mm, 50 mm and 150 mm respectively.

The density as well as the compressive and tensile splitting tests were evaluated on three specimens after 28 days of curing. The notation throughout this study was chosen as XXX-YP, where XXX referred to the bamboo fiber diameter in microns and Y referred to the volumetric fiber content. Therefore, notation 300-6P referred to the mixture containing 6 V% (volume %) of bamboo fibers with a nominal diameter of 300 μm.

#### **3. Results and Discussion**

#### *3.1. Mechanical Properties*

#### 3.1.1. Density and Compressive Strength

The control mix exhibited a density of 2132 kg/m3, which was slightly higher than the mixes containing fibers. The addition of bamboo fibers had little effect on the density, revealing a reduction ranging between 0.7% and 2.5%, as shown in Table 3. In terms of performance, the control mixture achieved the highest 28-day compressive strength of 75.1 MPa, as shown in Figure 4. The inclusion of fibers led to a reduction in strength, which was directly correlated with the number of bamboo fibers added to the mix design as shown in Table 3. Accordingly, higher fiber contents resulted in lower compressive strengths. This behavior can be attributed to the difference in the compressive strength and E-modulus of the bamboo fibers and the cement matrix. A previous study of the authors on different grades of the bamboo culms revealed a compressive strength between 43.2 and 68.4 MPa and an E-modulus between 18.1 and 28.2 GPa [31]. The incorporated bamboo fibers reduce the overall strength of the matrix resulting in a higher loss of strength with increasing fiber content which is in compliance with the findings of other researchers [20,54,55]. In addition, a study of Li et al. on natural fibers found that air pockets were formed at some of the fibers resulting in a reduced compressive strength compared to the control mixture [56]. Furthermore, a lower aspect ratio, thicker and longer fibers (i.e., "500 μm" batch), generally indicated a greater reduction in the compressive strength than thinner and shorter fibers (i.e., "300 μm" batch), which was also found in [54]. In this respect, the reduction of the compressive strength for the 4 V% to 8 V% of fibers was within a range of 7.8–19.9% for the 300 μm fibers and 9.1–27% for the 500 μm fibers (see Figure 4 and Table 3). An average compressive strength of 60.2 MPa was recorded for the 8 V% mixture with 300 μm fibers, while the corresponding figure was 54.8 MPa for the 8 V% mixture with 500 μm fibers. Although a slight reduction in strength was observed, the overall findings reveal the comparable performance of the bamboo fiber reinforced mortar mixtures with those of commonly used mix designs for structural members in similar studies [20,31,54].


**Table 3.** Physical and mechanical properties of the Bamboo Fiber-reinforced mortar mixtures.

<sup>1</sup> SD: Standard deviation calculated from mean value in kg/m3 or MPa; <sup>2</sup> SD: Standard deviation calculated from mean value in percentage.

**Figure 4.** Compressive strength (**a**) and splitting tensile strength (**b**) of the bamboo fiber-reinforced mortar mixtures.

#### 3.1.2. Splitting Tensile Strength

The control mixture showed a splitting tensile strength of 7.2 MPa. Similar to the trends observed in the compressive strength results, the splitting tensile strength of the samples with 8/6/4 V% of fibers was reduced by 6.9/25.0/26.4% for the 300 μm fibers and 23.6/30.6/31.9% for the 500 μm fibers (see Figure 4 and Table 3). It is worth noting that the splitting test setup compression is applied to a small area along the specimen resulting in tensile stress (lateral force) within the matrix which causes splitting. The splitting tensile strength is the maximum load before cracks appear and therefore the crack bridging behavior of the fibers is not considered in this test. This is investigated with bending tests and discussed later on. According to [57], the tensile strength of a mortar matrix can be assumed to be in the range of 10% of the compressive strength which matches the results obtained in this study.

#### 3.1.3. Flexural Tensile Behavior

The control mix showed linear elastic behavior up to the peak load, which was followed by an abrupt failure, resulting in complete separation of the specimens into two parts. In contrast, the bamboo fiber mixtures exhibited improved post-crack behavior, resulting in a strain-softening behavior, where a single crack occurred within the central third of the prismatic samples. This effect can be seen in Figures 5 and 6 where the pictures of the tested specimens taken after the bending tests are shown for both sets of samples containing 300 μm and 500 μm fibers. Within these figures, it is possible to visualize the fibers bridging the developed crack, which was much more evident as the fiber content increased from 4 V% to 8 V%.

**Figure 5.** Photographs of test samples after bending tests involving 300 μm fibers. Fiber content: (**a**) 4 V%, (**b**) 6 V% and (**c**) 8 V%.

**Figure 6.** Photographs of test samples after bending tests involving 500 μm fibers. Fiber content: (**a**) 4 V%, (**b**) 6 V% and (**c**) 8 V%.

The load–deflection behaviors of all test specimens are shown in Figure 7. Test results with different fiber contents are plotted with an offset of 0.1 mm from each other for better visual comparison. Both fiber groups involving the use of 300 μm and 500 μm fibers, revealed similar trends, during which 8 V% mixtures exhibited the highest peak load and the mixtures with 4 V% the lowest. Furthermore, the bamboo fibers were able to bridge the developing crack, resulting in a strain-softening behavior in all bamboo mixtures. The post-crack behavior was more pronounced for mixtures with higher fiber contents, which was in line with the findings with Aydin Serdar and Hameed et.al. where it was shown that a higher fiber content results in a higher load-carrying capacity due to the crack bridging behavior of the fibers [58,59].

**Figure 7.** Bending test results of mixtures with 4 V%, 6 V% and 8 V% of (**a**) 300 μm fibers and (**b**) 500 μm fibers.

#### 3.1.4. Fracture Properties and Toughness

The limit of proportionality and the modulus of rupture are shown in Figure 8. It should be noted that one out of the 3 samples of the 300 μm with 8 V% and 6 V% exhibited a significantly lower bending performance as can be seen in Figure 7. This resulted in a high standard variation shown in Figure 8. Samples with the 500 μm fiber batch revealed a lower modulus of rupture compared to those containing 300 μm fibers. These values were around 0.0%, 7.7% and 13.0% lower for the 8/6/4 V% fibers specimens, respectively. These findings were in agreement with ACI 544.1R-96, where a lower MOR was reported with decreasing aspect ratio of the fibers [60]. Longer and thinner fibers exhibit a higher bond within a concrete or mortar matrix resulting in a higher pull-out resistance compared to shorter and thicker fibers as shown which was also found by other researchers in [55,61].

**Figure 8.** Bending test assessment: (**a**) limit of proportionality Fmax; (**b**) modulus of rupture.

The toughness was calculated as the area under the load–deflection curve and expressed as energy in Joules. Accordingly, the resulting cumulative energy up to 2 mm deflection is shown in Figures 9 and 10. As expected, a higher fiber content resulted in higher toughness. Thicker and longer fibers, as found in the 500 μm batch, showed an overall reduction in toughness when compared to the smaller and shorter fibers in the 300 μm batch. This is explained due to the higher bridge bearing capacity of the mixtures with a higher amount of fibers and a higher pull-out resistance for fibers with a higher aspect ratio as discussed earlier and shown in [59,61].

**Figure 9.** Cumulative toughness of the bamboo fiber reinforced specimens up to 2 mm deflection: (**a**) 300 μm batch; (**b**) 500 μm batch evaluated from the four-point bending test according to ASTM C1609 [52].

To characterize the post-cracking behavior of the different mixes, two toughness parameters were evaluated in accordance with RILEM RC 162-TDF [53]. The toughness values at a mid-span deflection of span/150, which corresponded to 1 mm deflection, and the toughness at a mid-span deflection of span/75 (2 mm deflection) were evaluated. The corresponding results, presented in Figure 10 and Table 4, revealed that a higher fiber content enhanced the ductility and toughness of the prepared mixes. The toughness values of the 300 μm fiber specimens at 1 mm deflection were 40.0% and 73.3% higher for the 6 V% and 8 V% mixes when compared to 4 V% mixes, respectively. Assessing the 500 μm fiber batch, an increase of 16% and 47% increase in toughness was recorded for the 8 V% and 6 V% mixes when compared to 4 V% mixes, respectively (see Figure 10). Furthermore, the results of the mixes containing 500 μm fibers showed lower energy up to both 1.0 mm and 2.0 mm deflection when

compared to the 300 μm fiber results and showed the influence of the aspect ratio of the fibers on the bond strength of the specimen which is in compliance with findings in [55,61]. The high standard deviation of the 300 μm with 8 V% and 6 V% is the result of the lower bending performance of one out of three tested specimens shown in Figure 7.

**Figure 10.** Toughness of the mixtures at 1 mm deflection measured according to RILEM RC 162-TDF [53].


\* The high standard deviation is a result of a significantly lower bending performance of one specimen.

#### **4. Conclusions**

This study focused on the development of a sustainable and affordable mortar mixture consisting of a high amount of fly ash (656 kg/m3) and varying contents of bamboo fibers (4/6/8 V%) from waste by-products of engineered bamboo-composite fabrication. The bamboo fibers, obtained from bamboo composite production waste, were categorized into two groups of 300 μm and 500 μm and were incorporated at three different volumes of 4/6/8 V% within each mix. The mechanical performance of the developed formulations was assessed via the measurement of the mechanical and fracture properties. From these results, the following conclusions can be drawn:

The addition of 4/6/8 V% of bamboo fibers has a reasonably modest effect on the compressive strength of the mixtures. Compared to the control mixture without fibers, the reduction of the compressive strength of the mixtures with 300 μm fibers is between 7.8% and 19.9%, with larger fiber volume fractions corresponding to a greater reduction of strength.


The findings emerging from this study demonstrate the suitability of using natural bamboo fibers, obtained from process waste, to improve the ductility of high-volume fly ash mortar. The resulting formulations can enable the development of a sustainable and low-cost mixture for structural members. These results can be utilized for the construction of low-cost and low-rise housing units in developing countries, especially in Southeast Asia, Latin and Central America, where there is access to bamboo and low-cost cementitious materials with low demand for ductility. Further studies on the durability of bamboo fibers and the replacement of steel reinforcement with engineered bamboo composites and natural bamboo fiber members are being performed to broaden the application range of these materials on a larger scale.

**Author Contributions:** Conceptualization, M.M.; Methodology, M.M.; Validation, M.M.; Formal analysis, M.M.; Investigation, M.M., A.J., and N.S.; Resources: M.M., A.J., and N.S.; Data curation, M.M., A.J., and N.S.; Writing—original draft preparation, M.M.; Writing—review and editing, A.J.; N.S.; H.K.T.; C.U. Visualization, M.M.; Supervision, H.K.T. and C.P.O.; Project administration, M.M. and A.J.; Funding acquisition, M.M., A.J., and N.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research is funded by the Republic of Singapore's National Research Foundation through a grant to the Berkeley Education Alliance for Research in Singapore (BEARS) for the Singapore-Berkeley Building Efficiency and Sustainability in the Tropics (SinBerBEST) Program. BEARS has been established by the University of California, Berkeley, as a center for intellectual excellence in research and education in Singapore. The research was conducted at the Future Cities Laboratory at the Singapore-ETH Centre, which was established collaboratively between ETH Zurich and Singapore's National Research Foundation (FI 370074016) 766 under its Campus for Research Excellence and Technological Enterprise program.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Durability Assessment of Recycled Aggregate HVFA Concrete**

#### **Valeria Corinaldesi 1,\*, Jacopo Donnini 2, Chiara Giosué 2, Alessandra Mobili <sup>2</sup> and Francesca Tittarelli <sup>2</sup>**


Received: 22 August 2020; Accepted: 14 September 2020; Published: 16 September 2020

**Abstract:** The possibility of producing high-volume fly ash (HVFA) recycled aggregate concrete represents an important step towards the development of sustainable building materials. In fact, there is a growing need to reduce the use of non-renewable natural resources and, at the same time, to valorize industrial by-products, such as fly ash, that would otherwise be sent to the landfill. The present experimental work investigates the physical and mechanical properties of concrete by replacing natural aggregates and cement with recycled aggregates and fly ash, respectively. First, the mechanical properties of four different mixtures have been analyzed and compared. Then, the effectiveness of recycled aggregate and fly ash on reducing carbonation and chloride penetration depth has been also evaluated. Finally, the corrosion behavior of the different concrete mixtures, reinforced with either bare or galvanized steel plates, has been evaluated. The results obtained show that high-volume fly ash (HVFA) recycled aggregate concrete can be produced without significative reduction in mechanical properties. Furthermore, the addition of high-volume fly ash and the total replacement of natural aggregates with recycled ones did not modify the corrosion behavior of embedded bare and galvanized steel reinforcement.

**Keywords:** fly ash; HVFA; recycled aggregate; RAC; sustainable building; reinforced concrete; corrosion of concrete

#### **1. Introduction**

In order to contribute to sustainable construction processes, some building materials, no longer able to fulfill their original task, can be reused as aggregate for concrete after being adequately processed [1]. Replacing natural aggregate (Nat) with recycled aggregate (Rec) in concrete allows the protection of the environment, since it reduces both the impact of quarries from which virgin aggregates are extracted and the volume of rubble disposed to landfills.

Similarly, the employment of fly ash (FA) in concrete enables the recycling of an industrial waste product. In particular, due to its pozzolanic activity, FA can partially replace cement, thus reducing the energy consumption and carbon dioxide emissions related to cement production [2].

Unlikely, replacing Nat with Rec can significantly reduce the performances of concrete in terms of workability. Moreover, Rec, due to its higher porosity with respect to Nat [3,4], also penalizes the concrete's compressive and tensile strength, the stiffness, the permeability, and the adherence between steel reinforcing bars and cement paste.

However, the literature has reported that the addition of mineral admixtures as fly ash (FA), metakaolin, silica fume, and ground granulated blast furnace slag in the mix is able to mitigate these

worsening effects both in traditional [5] and in self compacting concretes (SCC) [6]. In fact, generally, these additions seem able to improve more the properties of Rec concrete than those of Nat concrete [7].

Previous experiments [1,8] have already shown the feasibility of manufacturing structural concretes with Rec and high-volume fly ash (HVFA) since FA, by refining the pore structure, reduces the macro-pores volume. In this way, performance similar to Nat concrete can be achieved except for somewhat lower stiffness of the Rec mixture.

Water absorption [9], chloride ion penetration [10], sulphate attack [11], and shrinkage [12] increase with the increasing incorporation level of Rec. However, the addition of HVFA counteracts this effect [5] thanks to the chemical reaction between some particles of FA, that act as a pozzolanic addition instead of a filler, with Rec [12].

Wei et al. [13] have indicated that an adequate amount of Rec can even increase the frost resistance of concrete, especially when a low amount of FA is added, thanks to optimization of the concrete pore distribution [14].

Moreover, since the thermal expansion coefficient of the new cement paste is similar to that of the cement paste adhered to Rec, Rec concrete deteriorates less in terms of mechanical and durability properties than Nat after high temperature exposures [15], especially when FA is used as mineral admixture [16,17]. FA as bacteria immobilizer also improves the crack healing capacity of Rec concrete [18].

Concerning carbonation resistance, Rec and HVFA concrete suffer a deeper carbonation depth with respect to Nat concrete [19], also in SCC [20]. However, again, the incorporation of FA in Rec concrete allows the counteracting of this problem thanks to a synergistic effect between Rec and FA [9,21,22].

According to Limbachiya et al. [11], the best amount of coarse Rec in concrete is 30%, whereas up to 30% Rec does not significantly affect the concrete's properties. Regarding FA, European standards EN 197-1 and EN 206 limit the incorporation level of FA to 35% by cement mass, since at higher amounts, FA behave as a filler rather than as a binder. However, these two limits for Rec and FA can be exceeded in concrete mixes incorporating both FA and Rec [5]. After 90 days of curing, concretes manufactured with about 50% Rec and 50% FA can be generally classified at the same strength class of the control mix.

Rawaz Kurda et al. [22], thanks to a multicriteria decision method for concrete optimization (CONCRETOP), have shown that the best concrete mixes in terms of both concrete properties and cost and environmental impact, are those manufactured with both FA and Rec additions, rather than with only FA or Rec. In particular, the Global Warming Potential (GWP) of concrete mixes depends on the FA and Rec dosage ratio rather than the dosage of the single materials [23]. Moreover, the GWP of Rec strongly depends on the transportation scenario, but this effect significantly decreases with FA addition [24].

Therefore, it has been already widely proved that replacement of Nat with Rec and the replacement of cement with HVFA, given a little bit of compromise towards strength and durability aspect, can give great benefits to both economic and ecological aspects.

As reported above, many researchers have already studied the different properties of Rec concrete with HVFA. However, durability, which is a key property to ensure sustainable application of these materials in the construction sector, still needs more research to be fully investigated.

In this field, in particular, the literature still reports very few works on the protection offered by HVFA and Rec concrete to the corrosion of reinforcing bars. Stambaugh et al. [25], thanks to the theoretical development, validation, and implementation of a 1D numerical service-life prediction model for RCA, have affirmed that the use of either FA or slag allows the achievement of a 50-year service life for Rec concrete in chloride-laden environments. By the salt ponding test, Rehvati et al. [26] have stated that impermeable and high-quality Rec concrete, able to give high corrosion resistance to reinforcements, can be produced by replacing 20–30% cement with FA. In Gurdián et al. [27], no significant differences in the corrosion resistance of reinforced Rec concretes, manufactured with

15% of spent fluid catalytic cracking catalyst and 35% of FA, and Nat concretes under a natural chloride attack have been observed.

Moreover, in our knowledge, the corrosion behavior of galvanized steel reinforcements in reinforced RCA in HVFA has never been investigated.

Therefore, the purpose of this work is to determine whether the sustainability issue introduced in concrete design by Rec and HVFA would have any adverse effect on the durability of reinforced concrete in terms of penetration speed of chloride and carbon dioxide, and in terms of corrosion of bare or galvanized steel reinforcement embedded in concrete, if cracked.

To investigate the single and combined effect of Rec and HVFA addition on concrete properties, four different concrete mixes were prepared and compared:


The different mixtures were compared in terms of mechanical performances, carbonation and chlorides penetration, and corrosion behavior of embedded bare and galvanized steel reinforcements.

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

#### *2.1. Materials*

As cementitious binder, Portland limestone blended cement type CEM II/A-L 42.5 R was used. The cement's Blaine fineness was 0.418 m2/g and its density was 3.04 kg/m3. The cement's chemical composition is reported in Table 1.


**Table 1.** Chemical composition of cement and fly ash.

Two virgin aggregate fractions were used: limestone aggregate from a quarry (up to 15 mm particle size) and quartz sand (up to 6 mm particle size). Their grain size distribution is shown in Figure 1 and their physical properties are reported in Table 2.

**Figure 1.** Grain size distribution curves of the aggregate fractions.


Coarse Recycled Fraction 2320 8 0.3

**Table 2.** Physical properties of the aggregate fractions.

Two recycled aggregate fractions were used: coarse aggregate (up to 15 mm) and fine aggregate (up to 6 mm). The origin of these recycled aggregates was a recycling plant in Villa Musone (Italy), where debris coming from building demolition has been selected, crushed, cleaned, and finally, sieved. The grain size distribution of recycled aggregates is shown in Figure 1 and their physical properties are reported in Table 2.

As water-reducing admixture, an acrylic-based superplasticizer (in the form of 30% aqueous solution) was added, when required, to reach the optimal workability degree.

Low calcium fly ash (according to the definition of ASTM C 618 Class F) coming from a thermal power plant located in La Spezia (Italy), with Blaine fineness of 0.458 m2/g and density of 2.23 kg/m3, was used. Fly ash chemical composition is detailed in Table 1.

#### *2.2. Mixture Proportions*

Four different concrete mixtures were designed as reported in Table 3. The two different kinds of aggregate particles, either recycled or natural, were used by maintaining the same grain size distribution (up to 15 mm). The optimization of grain size particle distribution was achieved by suitable combining of fine and coarse aggregate fractions, according to Bolomey [12]. All mixtures were designed so to have the same workability, with a slump value in the range 150–180 mm. According to that, when recycled aggregates and fly ash have been used, an acrylic-based superplasticizer was added at dosages up to 2.0% by weight of cement.


**Table 3.** Concrete mixture proportions (kg/m3).

The first reference mixture (Nat-0.6) was designed with only the addition of virgin aggregates and a water-to-cement ratio (*w*/*c*) of 0.60.

In the second mixture (Nat + FA-0.6), fly ash (FA) at the same dosage of cement was added, in replacement of natural sand. However, to reach the same workability of the reference mixture, acrylic-based superplasticizer as water-reducing admixture was added at a dosage of 2.0% by weight of cement.

In the third concrete mixture (Rec-0.3), virgin aggregates were fully substituted with recycled ones, while a lower *w*/*c* equal to 0.3 was adopted, to recover the strength loss due to the addition of weaker aggregate.

Finally, in the fourth mixture (Rec + FA-0.6), fine and coarse recycled aggregates were used in complete substitution of natural aggregates, while cement was partially replaced with fly ash. The use of superplasticizer at dosage of 1.8% by weight of cement allowed us to keep unchanged the water-to-cement ratio (equal to 0.6).

#### **3. Preparation and Testing of Specimens**

#### *3.1. Compression Tests*

Compression tests were performed on cubic specimens (100 mm edge) after 3, 7, 28, and 56 days of wet curing at 20 ◦C (according to the procedure of UNI EN 12390-1 [28]). Compression tests were carried out according to the procedure described in the Italian Standards UNI EN 12390-3 [29]. Three specimens for each curing time and each type of mixture were tested.

#### *3.2. Carbonation Depth*

Carbonation depth was evaluated through a phenolphthalein test (following the indication reported in RILEM CPC-18 [30]) on three cubic concrete specimens (100 mm edge) for each mixture, exposed to the open air at an average temperature of 20 ◦C (only for the first day of curing, wet curing was adopted).

#### *3.3. Chloride Penetration*

Chloride penetration speed into concrete was evaluated by means of silver nitrate and fluorescein test [13]. Both solutions were sprayed on the two cracked surfaces obtained by splitting the cubic (100 mm edge) concrete specimens. These specimens were previously wet-cured for 7 days, air-cured for 21 days at a temperature of 20 ◦C, and finally, exposed to 10% sodium chloride aqueous solution.

#### *3.4. Corrosion Tests*

Concrete specimens with dimensions of 280 <sup>×</sup> <sup>70</sup> <sup>×</sup> 70 mm<sup>3</sup> were manufactured for electrochemical tests. Each specimen was reinforced with a bare or galvanized steel plate (210 <sup>×</sup> 40 <sup>×</sup> 1 mm3), embedded within a 3 cm concrete cover. Steel plates were used instead of common bars because they can allow specimen cracking without splitting and they offer a higher anodic area at the crack apex. The galvanized steel plates, obtained by molten zinc immersion, were covered by a 100 μm thick zinc layer, with an outer pure zinc layer (η phase) about 20 μm thick. The galvanized reinforcements, just before being embedded in the fresh concrete, were submerged for 5 s in a 15% sodium hydroxide solution to dissolve the ZnCO3 layer formed during air exposure. The electric contacts between the metallic plates and the measurement equipment were arranged as reported in [14].

After 1 month of air curing at T = 20 ± 3 ◦C and RH = 50 ± 5%, a crack width of 1 mm was produced in a pre-formed notch area of the specimens by applying a flexural stress with the apex crack reaching the metallic plates. Then, the specimens were exposed to weekly wet–dry cycles (2 days dry and 5 days wet) in a 10% NaCl solution.

The corrosion risk of the reinforcement in the concrete specimens exposed to the chloride environment was evaluated by corrosion potential measurements by using a saturated calomel electrode (SCE) as a reference. The kinetics of the corrosion process was followed by polarization resistance measurements through the galvanodynamic method, where an external graphite bar was used as a counter-electrode. The polarization resistance was calculated as the average value between the anode and the cathode branch.

In the following graphs, the reported electrochemical values are averages of the measurements carried out on 3 specimens of each type during the full immersion period.

In order to validate the electrochemical tests, after 7 wet–dry cycles in the chloride solution, the concrete specimens were saw-cut and all the metallic plates were removed after splitting the concrete specimens, to evaluate the reinforcement corrosion by visual observation. The surface of the corroded area on bare steel plates was evaluated after pickling, whereas metallographic analysis was carried out on the cross-section of the galvanized steel plates to evaluate the coating thickness decrease due to the corrosive attack.

#### **4. Results and Discussion**

#### *4.1. Compression Tests*

The experimental results of compressive tests on the four mixtures at different curing times are reported in Figure 2. All mixtures showed a compressive strength greater than 27 MPa at 28 days. At early ages (3 and 7 days), the different *w*/*c* influenced the compressive strength more than the kind of aggregate. The total substitution of natural aggregates with recycled ones, simultaneously with the reduction in *w*/*c* and the use of superplasticizer (Rec-0.3), did not substantially modify the compressive strength with respect to the concrete mixture with natural aggregates (Nat-0.6). When cement was partially replaced with FA (Rec + FA-0.6), the mixture showed a lower compressive strength at early ages (3 and 7 days) and a slightly higher compressive strength after 56 days of curing, thanks to the fly ash pozzolanic activity which develops at long ages. The use of lighter aggregate (i.e., recycled aggregate) increasingly influenced the compressive strength value as the cement matrix became stronger, since it represents the "weak link" in the chain [15]. The best results have been obtained for the mixture realized with natural aggregates and fly ash (Nat + FA-0.6), which showed a compressive strength of about 44 MPa after 56 days of curing.

**Figure 2.** Compressive strength vs. curing time.

#### *4.2. Carbonation Depth*

In Figure 3, the measured values of carbonation depth (*x*, in mm) were reported vs. days of air exposure (*t*). Experimental data can be quite accurately described by a linear relationship between the carbonation depth and square root of time (following the law *x* = *k*· √ *t*). The higher carbonation depth, equal to about 7 mm after 1 year of exposure, was found for the mixture with natural aggregates (Nat-0.6), while the lower carbonation depth, equal to 2.8 mm, was found in the concrete mixture realized with natural aggregates and fly ash. These results confirm that the addition of high volumes of fly ash is able to significantly reduce the carbonation process, even when a porous aggregate (such as the recycled aggregate) is used, thanks to the refinement of the pores and a consequent improved microstructure.

**Figure 3.** Carbonation depth vs. time of air exposure.

#### *4.3. Chloride Penetration*

In Figure 4, chloride penetration depth values are reported vs. time of exposure to 10% NaCl aqueous solution after saturation with water of concrete specimens. Figure 4 shows only data obtained after about 28-day water immersion due to initial instability phenomena (in practice, the chloride binding with cement matrix can interfere with the chloride migration [16]).

**Figure 4.** Chloride penetration depth vs. immersion time.

Collepardi et al. [13] demonstrated that chloride penetration depth (*x*) varies with immersion time (*t*) according to Equation (1), which can be obtained from the solution of Fick's second law under non-steady-state conditions for diffusion in a semi-finite solid:

$$\mathbf{x} = 4 \cdot \sqrt{D \cdot t} \tag{1}$$

where *<sup>D</sup>* is the diffusion coefficient of Cl ions into wetted concrete pores, expressed as cm2 · <sup>s</sup>−<sup>1</sup> · <sup>10</sup><sup>−</sup>8. The *D* values, coming from (1) by interpolation of the results reported in Figure 4, have been calculated (see Table 4). It can be noticed a significant beneficial effect due to fly ash addition on the chloride penetration depth, which has been measured for Rec + FA-0.6 and Nat + FA-0.6 concretes, respectively. As a matter of fact, Cl diffusion coefficients into these mixtures are 10 times less than those measured for the other concrete mixtures.

**Table 4.** Chloride diffusion coefficients at 20 ◦C.


Collepardi et al. [13] found experimental results, which support the hypothesis that the different structure of pore surfaces created with the addition of pozzolanic materials played an important role in influencing concrete porosity while chloride ions penetrate it. In fact, if concrete is prepared by adding fly ash, chloride binding operated by the cement matrix significantly increases.

A lower water/cement resulted beneficial in terms of hindering of chloride penetration. In fact, the Cl diffusion coefficient into the Nat-0.6 mixture is double than Rec-0.3, even if recycled instead of virgin aggregate was employed. Nevertheless, aggregate pore structure could have a significant effect on concrete permeability to chloride penetration (as you can see by comparing Nat + FA-0.6 and Rec + FA-0.6 mixtures in Figure 4). Zhang et al. [17] experimented several lightweight aggregate concretes and the conclusion was that concrete permeability depends more on cement matrix porosity than lightweight aggregate porosity.

#### *4.4. Corrosion Tests*

#### 4.4.1. Bare Steel Plates

Figures 5 and 6 show, respectively, the free corrosion potential values and the corrosion rates of bare steel plates embedded in cracked concrete as a function of wet–dry cycles. Just after the exposure to the chloride environment, all the steel plates assumed activation values lower than −500 mV/SCE, reflecting a general high corrosion risk, regardless of the type of concrete mixture. At the same time, the related polarization resistance values did not change significantly with the addition of FA or when natural aggregates were replaced with recycled ones, thus indicating similar corrosion rates for the different concretes. Therefore, the total replacement of natural aggregates, with or without high FA volume, does not seem to negatively affect the corrosion behavior of embedded steel reinforcements when an adequate strength class is guaranteed. Moreover, despite the reduction in concrete pore solution alkalinity due to the pozzolanic reaction of fly ash, the corrosion behavior of steel reinforcements in high-volume fly ash concrete seems not to be negatively affected, at least for cracked concrete.

**Figure 5.** Corrosion potential of bare steel plates in cracked concrete as a function of wet–dry cycles.

**Figure 6.** Polarization resistance of bare steel plates in cracked concrete as a function of wet–dry cycles.

To better analyze the electrochemical behavior, all specimens were autopsied after 7 wet–dry cycles in order to visually evaluate the corroded area and to assess the weight loss of the bare steel plates after pickling. A visual observation of the corrosive attack on the steel plates embedded in the different concrete mixtures is reported in Figure 7. All the bare steel plates showed significant

corrosive attack in the crack area, due to the high Cl concentration detected on the steel reinforcement surface (2–5% by weight of cement), thus overcoming the concentration threshold (0.4% by cement weight), which is considered the critical value able to induce the corrosion of bare steel. However, from a morphological point of view, in the presence of recycled aggregate (Rec-0.3) or when a high amount of fly ash is used (Nat + FA-0.6), the corrosive attack appears more diffuse and less penetrating (see Figure 7).

**Figure 7.** Visual observation of the corrosive attack on the bare steel plates in reference natural aggregate concrete (Nat-0.6, on the left), in recycled aggregate concrete (Rec-0.3, in the middle), and high-volume fly ash concrete (Nat + FA-0.6, on the right).

#### 4.4.2. Galvanized Steel Plates

The free corrosion potential and the polarization resistance of galvanized steel plates embedded in cracked concrete, as a function of wet–dry cycles, are reported in Figures 8 and 9, respectively. The electrochemical tests showed no significant increase in the corrosion rate of galvanized steel embedded in concrete mixtures containing recycled aggregate and/or FA. Indeed, the corrosion risk seems to be even lower in recycled aggregate concrete (see Rec-0.3 in Figure 8) than in the other mixtures.

**Figure 8.** Corrosion potential of galvanized steel plates in cracked concrete vs. wet–dry cycles.

**Figure 9.** Polarization resistance of galvanized steel plates in cracked concrete vs. wet–dry cycles.

The visual observation of the interface between galvanized steel plates and concrete added unexpected information to that obtained by the electrochemical tests. Indeed, in the concrete without additions (Nat-0.6, Figure 10a), far from the crack apex, a Fe-Zn alloy appeared on the surface of the reinforcement, meaning total consumption of the η zinc layer due to the corrosive attack, as later confirmed by metallographic observation (Figure 10b). On the other hand, the galvanized steel plates extracted from recycled aggregate concrete (Rec-0.3) showed a less deep corrosive attack. Zinc grains were still visible on the galvanized surface after exposure to wet–dry cycles (Figure 11a). Moreover, the metallographic observation revealed that a continuous thick η zinc layer was still present on the metallic plate (Figure 11b). Similar results were observed when high-volume fly ash was added in the mix.

**Figure 10.** Visual obs. (**a**) and metallographic cross section (**b**) of galvanized steel plate in Nat-0.6

(**a**) (**b**)

**Figure 11.** Visual obs. (**a**) and metallographic cross section (**b**) of a galvanized steel plate in Rec-0.3

#### **5. Conclusions**

Based on the experimental investigation carried out on recycled aggregate HVFA concrete to study its mechanical properties and durability characteristics, the following conclusions can be drawn:


**Author Contributions:** Investigation, V.C., J.D., C.G., A.M. and F.T.; Methodology, J.D. and F.T.; Supervision, V.C.; Writing—review and editing, J.D. and F.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **E**ff**ect of Fly Ash as Cement Replacement on Chloride Di**ff**usion, Chloride Binding Capacity, and Micro-Properties of Concrete in a Water Soaking Environment**

#### **Jun Liu, Jiaying Liu, Zhenyu Huang \*, Jihua Zhu, Wei Liu and Wei Zhang**

Guangdong Provincial Key Laboratory of Durability for Marine Civil Engineering, Shenzhen University, Shenzhen 518060, China; liujun@szu.edu.cn (J.L.); 1910472060@email.szu.edu.cn (J.L.); zhujh@szu.edu.cn (J.Z.); liuwei@szu.edu.cn (W.L.); zhangwdnv@gmail.com (W.Z.)

**\*** Correspondence: huangzhenyu@szu.edu.cn; Tel.: +86-755-8697-5402; Fax: +86-755-2673-2850

Received: 16 August 2020; Accepted: 7 September 2020; Published: 9 September 2020

#### **Featured Application: The study on the di**ff**usion, bonding and micro-properties of chloride penetration in concrete in a water soaking environment can provide fundamental information to evaluate the durability of the marine concrete.**

**Abstract:** This paper experimentally studies the effects of fly ash on the diffusion, bonding, and microproperties of chloride penetration in concrete in a water soaking environment based on the natural diffusion law. Different fly ash replacement ratio of cement in normal concrete was investigated. The effect of fly ash on chloride transportation, diffusion, coefficient, free chloride content, and binding chloride content were quantified, and the concrete porosity and microstructure were also reported through mercury intrusion perimetry and scanning electron microscopy, respectively. It was concluded from the test results that fly ash particles and hydration products (filling and pozzolanic effects) led to the densification of microstructures in concrete. The addition of fly ash greatly reduced the deposition of chloride ions. The chloride ion diffusion coefficient considerably decreased with increasing fly ash replacement, and fly ash benefits the binding of chloride in concrete. Additionally, a new equation is proposed to predict chloride binding capacity based on the test results.

**Keywords:** fly ash; carbon dioxide emission; chloride diffusion; binding capacity of chlorine

#### **1. Introduction**

Since Portland cement was invented in 1824, cement has become the most important and irreplaceable building material in infrastructure construction. However, problems associated with cement are becoming increasingly prominent. For example, cement production consumes a great amount of energy and released a great deal of toxic pollutants, such as dust, soot, sulfur dioxide, and carbon dioxide, into the environment [1,2]. The production of Portland cement causes about 7% of the world's carbon dioxide emissions, which is essential to be reduced. Using supplement materials in concrete (e.g., fly ash, and slag) is effective in reducing the amount of cement consumption without sacrificing any properties [3]. Fly ash is the main coal combustion by-product from power plants [4–6]. It has become the fifth largest raw material resource in the world and has been used for more than 50 years in concrete [7,8]. Replacing part of the cement admixture in concrete can save cement clinker, thus reducing the environmental pollution from cement production [9,10]. It has been reported that, per ton, replacing Portland cement by fly ash can reduce one ton of carbon dioxide emissions [11]. In addition, fly ash can obviously enhance the workability of fresh concrete, reduce

hydration heat in the early stages, improve corrosion resistance [9,10,12–14], and improve concrete's durability performance [15–19].

It is generally believed that chloride ion penetration in concrete is one of the most critical causes of steel reinforcement corrosion that may further cause structural failure [20–22]. Chloride ions are present in many environments, including marine, road de-icing salt, salinized soil, and industrial wastewater environments. Chloride ions penetrate through pores into concrete and accumulate around steel reinforcements, which destroys the passive film on the steel surface and activates steel reinforcement corrosion [23–25]. For existing reinforced concrete structures, the penetration process of chloride ions is an important parameter in durability life assessment [26]. Liu et al. [24] examined the microstructures of fly ash concrete and normal concrete in a chloride atmosphere environment; the fly ash content was discovered to be sensitive to the chloride binding capacity and diffusion coefficient. Wu et al. [27] found that chloride transportation was affected by drying–wetting cycles and compressive stress ratios—the former significantly increased the chloride ion content, and the latter decreased the chloride ion content. Mehta and Monteriropj [28] pointed out that a pore size of <100 nm has little or no effect on concrete strength and permeability, while a pore size of >100 nm negatively affects concrete strength and permeability. The permeability of fly ash concrete seems physically different from that of normal concrete. Unlike that in salt-spray environments, concrete in seawater environments has different chloride ion concentrations, in which migration speed of chloride ion varies considerably. The chloride ion distributions in concrete are quite different, which is related to the way that chloride ions penetrate concrete. In fact, chloride ion penetration is a complex process involving extensive transportation mechanisms, such as diffusion under a concentration difference [27], penetration under water pressure [29], and capillary action under a humidity gradient [30]. In a soaking environment, when the concrete is saturated, chloride ions are mainly transferred through diffusion to the concrete. Meanwhile, in a salt-spray environment, the migration of chloride ions is achieved mainly by capillary suction and the diffusion effect. Therefore, the chloride profile of concrete can be quite different in different environments, which may cause different corrosion mechanisms of reinforcement in concrete. Thus, the knowledge of such durability behavior of fly ash concrete remains unclear, even though it is essential for promoting durable material design in construction under water. Due to filling and pozzolanic effects [31–33], adopting fly ash to partly replace cement can increase chloride ion resistance and lead to the better durability of concrete [34–36]. Moreover, there are very few references on the distribution of chloride ions in concrete soaked with seawater, thus indicating the demand for a comprehensive study that covers not only the micro-properties and transportation mechanism but also the chloride binding capacity of fly ash concrete.

The current investigation aimed to systematically examine the effect of fly ash on chloride ion diffusion in concrete in a water soaking environment. Specifically, the effect was quantified by the chloride ion diffusion coefficient and chloride ion deposit amounts at different layers. The study also compared the amount of free and binding chloride ions, as well as binding resistances, under different dosages of fly ash replacement. Through a mercury intrusion method and SEM, the study explored the effects of fly ash on the pore structure and cement hydration products.

#### **2. Experimental Investigation**

#### *2.1. Materials*

ASTM (American Society for Testing and Materials) type I Portland cement, Class F fly ash, river sand (fineness: 2.3–3.0), and a coarse aggregate (size: 5–20 mm in diameter) were used for concrete preparation. Industrial chloride salt (purity: 99%) was adopted to simulate the NaCl solution. Table 1 lists the proportions of three mixtures of normal concrete with fly ash replacements of 0%, 15%, and 30% (by weight). Table 2 lists the chemical compositions of the cement and fly ash.


**Table 1.** Mix proportions of concrete. kg/m3.

OPC = ordinary Portland cement; CA = coarse aggregate; FA = fly ash; and W/C = water to cement ratio; PCFA0 represents plain concrete with 0% fly ash.

**Table 2.** Chemical compositions (mass percentages of oxides, %).


#### *2.2. Immersion Test Methodology*

To simulate the concrete deterioration process in a seawater soaking environment, immersion tests of concrete samples were conducted based on NT (NORDTEST) Build 443 [37]. Three standard concrete cylinders were cast for each concrete mix and demolded 48 hours later. After standard curing under the environment with a constant temperature of 20 ± 3 ◦C and a relative humidity of >95% for 90 days, these specimens were first saturated with a Ca(OH)2 solution, ensuring that the weight change of all samples was less than 0.1%. After that, the samples were exposed to a saturated-surface-dry state environment under ambient temperature before the top surface was selected as the exposed surface, while the remaining surfaces were sealed with an epoxy material to guarantee the unidirectional diffusion of the chloride ions during the test. Then, the samples were soaked in an NaCl solution with a 16.5% concentration at a temperature of 23 ± 2 ◦C for 35 days. Figure 1a shows the concrete samples immersed in the NaCl solution.

#### *2.3. Measurement of Chloride Content*

After the immersion process, the crystal salt on the concrete surface was removed. Then, the samples were grinded by a specially-designed pulveriser into powder from the exposed surface, layer by layer. From the height of 0–16 mm, the grinding thickness was 1 mm, while from the height of 16–40 mm, the grinding thickness was 2 mm, as shown in Figure 1b. The powder samples were then placed in an oven at up to 105 ◦C to dry to a constant weight. The determination of the free and total chloride ion contents was based on AASHTO T260 (2009) [38], while the concentration of chloride ions was determined by the automatic potentiometric titrator as shown in Figure 1c.

#### *2.4. Microstructural Characterization*

The study adopted mercury intrusion porosimetry to evaluate the porosity and pore size distribution characteristics of concrete [39,40]. For each mix, three sample tests were conducted, and the average results were used for analysis. The concrete samples were crushed into small sizes and dried by a solvent replacement method, changing the solvent every 6 hours for the first few days and then every day for one week. After that, the samples were vacuum-dried and subjected to mercury intrusion up to 210 MPa. The contact angle between mercury and concrete was set to 130◦. Assuming that the shape of the porous material was cylinder, the applied pressure *p* (in mN/m2) was converted to the pore diameter D (in m) by the Washburn equation [41]:

$$D = (-4\sigma \cos \theta) / p\_\prime \tag{1}$$

where σ is the mercury surface tension force in mN/m and θ is the contact angle between mercury and the capillary surface.

(**a**) Concrete sample immersion in NaCl solution.

**Figure 1.** Measurement of chloride content in concrete.

Microstructural studies of the concrete sample after exposure to seawater immersion were performed using SEM. The concrete samples were taken from the cylinder surface with a diameter of <25 mm and a thinness of <20 mm. The specimens were washed and dried to a constant weight in an oven under 105 ◦C. The specimens were then coated with a gold-plated film due to the poor electric conductive performance of concrete. Ultra-high resolution SEM was adopted to identify the hydration products, observe the fly ash, and evaluate the pore structures.

#### **3. Results and Discussion**

#### *3.1. E*ff*ect of Fly Ash on Chloride Transportation*

The free and total chloride contents were determined in accordance with the AASHTO T260 standard [38]. The former was obtained based on the water-soluble extraction method, while the latter was obtained by acid-soluble extraction in a nitric acid solution. After extraction, the chloride content was measured with an automatic potentiometer titrator. Figure 2 shows that the concentration of free chloride content from each concrete layer reduced significantly as the fly ash replacement ratio increased. Two individual zones of chloride deposition, namely the descending (0–15 mm) and flat zones (15–40 mm), were observed. The curve descended significantly at the 0.0–7.5 mm zone, in twhich the chloride concentration of plain concrete with 0% fly ash (PCFA0) reduced by 0.426% with a reduction percentage of 54.5%. For PCFA15 and PCFA30, the related reductions were 0.4722% and 0.4876% with reduction percentages of 65.9% and 69.4%, respectively. It was found that concrete with fly ash had a lower chloride content, which indicated a larger reduction percentage compared to the normal concrete without fly ash. For the zone of 17–27 mm, the reduction of chloride concentration was negligible for all samples since the chloride ion concentration became lower as concrete depth increased.

**Figure 2.** Free chloride deposition of concrete with different fly ash replacement ratios.

Figure 2 shows that the chloride ion distribution curves in the soaking environment satisfied Fick's second law, while Figure 3 shows that the curves in the salt-spray environment did not follow Fick's second law. In the salt-spray environment, there was a rising branch called the convection zone for the chloride ion on the concrete surface. Beyond this convection zone, the chloride ion concentration decreased. Chloride ion distribution in concrete depends on the way chloride ions penetrate into the concrete. In fact, chloride ion penetration is a complex process including extensive transportation mechanisms, such as diffusion under a concentration difference [27], penetration under water pressure [29], and capillary action under a humidity gradient [30]. In a soaking environment, when concrete is saturated, chloride ions are transferred mainly through diffusion to the concrete. Meanwhile, in a salt-spray environment, the migration of chloride ions is mainly attributed to capillary suction and the diffusion effect. Chloride aerosols deposited on the surface of concrete are absorbed through capillary suction. Then, the free chloride ions are diffused into the concrete due to a concentration difference. When the water evaporates from the concrete pore, the chloride ions left in pores gather to a peak amount through a series of drying and wetting cycles. This results in the convection of the chloride ion migration under the salt-spray environment, which can be seem in Figure 2 [24]. According to the results of chloride penetration in the soaking and salt-spray environments in our experiments, the distribution of chloride ions on the concrete surface varied, but the trend approached consistency as the concrete depth increased.

**Figure 3.** Qualitative expression of the chloride penetration curves [24].

The deposition of chloride ions gradually approached zero at concrete depths over 15 mm. This was mainly due to the maximum exposure day (35 days), and this transition depth could be increased if the number of exposure days increases. In addition, the ingress depth also depended on the specific concrete mix used (water to cement ratio: W/C). The curve in the descending zone in Figure 2 indicates that the free chloride deposition of the samples with 15% and 30% fly ash replacement considerably decreased compared to that of the normal PCFA0 concrete. This is because fly ash has three main effects on concrete, namely the pozzolanic, morphological, and microaggregate filling effects. The pozzolanic effect is triggered by Ca(OH)2 formed from Portland cement hydration. However, the pozzolanic reaction at the beginning of cement hydration is very slow because of the presence of less Ca(OH)2. With the increase of curing age, the secondary hydration reaction between the activated ingredients, such as SiO2 and Al2O3, in fly ash, and Ca(OH)2 produces C–S–H gel, calcium aluminate hydrate products, and so on. These products may absorb more free chloride ions, reducing the deposition of chloride ions in concrete. This also indicates that chloride binding capacity is improved with the addition of fly ash. These hydrates can fill large size pores in cement matrixes and thus reduce porosity, narrow pore diameter, and block the connectivity of the pores, all of which subsequently slow the diffusion and migration of chloride ions. In addition, 70% of fly ash particles are inert, compact hollow microsphere (cenospheres) with smooth surfaces [31,32], which are like activated nanomaterials that improve the microstructures of concrete. Consequently, a concrete with fly ash replacement has a much lower chloride deposition than normal concrete.

#### *3.2. Influence of Di*ff*usion Coe*ffi*cient*

It was found that the chloride deposition curve in Figure 2 generally followed Fick's second law (Equation (2)) [42,43]. Based on the data in Figure 2, Figure 4 plots the regression curves of the chloride

coefficient using the Origin software. The chloride ion diffusion coefficient D and surface concentration of chloride C for each group can be determined by Equation (2), as shown in Table 3.

$$\mathcal{C}(\mathbf{x},t) = \mathbb{C}\_0 + (\mathbb{C}\_s - \mathbb{C}\_0)[1 - erf(\frac{\mathbf{x}}{2\sqrt{Dt}})],\tag{2}$$

where *C*(*x*,*t*) is the chloride percentage in concrete depth *x* at exposure time *t* (%), *Cs* is the chloride percentage at the concrete surface (%), *C*<sup>0</sup> is the initial percentage of chloride in concrete, and *er f*(*z*) is the error function, *er f*(*z*) = <sup>2</sup><sup>√</sup> Π *z* <sup>0</sup> exp(−*z*2)*dz*.

**Figure 4.** Regression curves of chloride coefficient using Fick's second law (by mass%).

**Table 3.** Regression results of chloride diffusion coefficient. D: chloride ion diffusion coefficient; CS: chloride percentage at the concrete surface; and σ2: the mercury surface tension force.


Figure 4 shows the regression curves of the chloride diffusion coefficient for PCFA0, PCFA15, and PCFA30 using Fick's second law. The chloride diffusion coefficient significantly decreased with the 15% and 30% replacements of cement by mass, with reduction ratios of about 26.1% and 37.9%, respectively, compared to normal concrete. The related coefficient reductions were 3.1984 <sup>×</sup> 10−<sup>12</sup> and 4.6441 <sup>×</sup> 10−<sup>12</sup> m2/s. This indicated that the chloride diffusion coefficient reduced more as the fly ash dosage rose. However, the reduction rate was slower as more fly ash was added. This was clear from the fact that the chloride diffusion coefficient of the PCFA30 sample was not proportionally reduced, which was less than half of that (1.4457 <sup>×</sup> 10−<sup>12</sup> m2/s) for the PCFA15 sample. Two main reasons for the reduction of the chloride diffusion coefficient were the (1) pozzolanic effect formed tobermorite and calcium aluminate hydrate product [44], which could have increased the chloride binding resistance, and (2) Al2O3 in the fly ash was the essential ingredient to generate Friedel's salt. The chloride ions in concrete are mainly in the form of Friedel's salt. Al2O3 reacts with Ca(OH)2 and may lower the C/A (calcium/aluminum) product and accelerate AFM (monosulphate) to form Friedel's salt, as explained in Equation (3). This process physically and chemically benefits the chloride binding capacity. Nevertheless, differentiating between the chemical and physical absorption process with

the chosen test setup was difficult in this study since chloride penetration have extensive transport mechanisms in concrete. Consequently, the decrease of free chloride content leads to reduce the chloride diffusion coefficient [45]. However, differentiating between the chemical and physical adsorption process needs further study.

#### *3.3. Influence of Free and Binding Chlorides*

Binding chloride is the solidifying process of chloride ions by cement-based materials. The binding chloride in the cement composite could not transport in water in concrete. Binding chloride can be divided into two types: chemical bound and physical absorption. Chemical bound results from the reaction of chloride and cement, while physical absorption depends on the electrostatic interactions (or van der Waals forces) [46,47] between ions. Equation (3) shows that chemical bound chloride ions follow the formation of Friedel's salt [48]:

$$\rm C\_3A + \rm CaCl\_2 + 10H\_2O \rightarrow C\_3A \cdot CaCl\_2 \cdot 10H\_2O,\tag{3}$$

C3A·CaCl2·10H2O reduces the porosity of concrete, thus increasing the resistance of chloride penetration and reducing the chloride ion binding capacity. According to Table 2, the contents of Al2O3 in the fly ash and cement were 23.95% and 4.62%, respectively. Since Al2O3 is a raw material for salt formation, adding fly ash into the concrete increased Friedel's salt formation, which further improved the chloride binding capacity.

Figure 5 shows the distribution of total chloride, free chloride, and bound chloride of the concrete samples. The bound chloride amount was much lower than the total and free chloride amounts. Figure 6 shows the relationship between bound chloride and concrete depth, in which the bound chloride amount decreased as concrete depth increased, but then the chloride amount went slightly up. PCFA15 and PCFA30 had much higher bound chloride amounts compared to that without any fly ash addition (PCFA0). The bound chloride content decreased significantly within the 0.0–12.5 mm depth, while it slightly increased to a constant value beyond 12.5 mm. In this case, free chloride could be reduced in the pores of concrete, which deceased the potential risk of rebar corrosion. The free chloride content kept a linear relationship with the total chloride content [49], as shown in Figure 7. Equation (4) introduces the index R to represent the chloride binding capacity:

$$R = \frac{\partial \mathcal{C}\_b}{\partial \mathcal{C}\_f} \,' \tag{4}$$

where *Cb* is the concentration of the bound chloride and *Cf* is the concentration of the free chloride.

Equation (5) gives a linear formulation to represent the linear relationship between the free and total chloride contents:

$$\mathbf{C}\_{l} = \mathbb{K}\mathbf{C}\_{f} + b\_{\prime} \tag{5}$$

where *Ct* is the concentration of total chloride and *K* and *b* are constants.

Substituting Equation (1) into Equation (2) results in Equation (6):

$$\mathbb{C}\_b = (\mathbb{K} - 1)\mathbb{C}\_f + b\_\prime \tag{6}$$

Equation (7) introduces chloride binding capacity index *R* shown below:

$$R = K - 1,\tag{7}$$

Based on the calculations of Equation (7), Figure 8 shows the relationship between the chloride bound capacity amount and concrete depth. The chloride binding capacity increased as concrete depth rose. Based on Equation (4), the bound capacity indexes of chloride were 0.1426, 0.1882, and 0.2134, respectively, for PCFA0, PCFA15, and PCFA30. This calculation indicates that the chloride binding capacity improves as fly ash addition increases.

**Figure 5.** Distribution of total chloride, free chloride, and bound chloride in concrete samples (by mass%).

**Figure 6.** Relationship between bound chloride amount and concrete depth (by mass%).

**Figure 7.** Relation between free chloride and total chloride depth (by mass%).

**Figure 8.** Relation between chloride bound capacity amount and concrete depth.

#### *3.4. E*ff*ect of Fly Ash on Porosity and Microstructure*

Physical and chemical reactions may significantly affect the porosity and microstructure of concrete due to fly ash addition. Figure 9 shows the cumulated porosity of the concrete samples. It was found that the cumulative pore volume reduced as the fly ash addition increased. Figure 10 shows the pore size distribution of each sample after 90 days of curing. The cumulative volume of pores reduced as fly ash content increased. The mean pore size of the PCFA0, PCFA15, and PCFA30 samples were about 40, 26, and 13 nm, respectively, which indicated that the fly ash addition led to a denser concrete microstructure.

**Figure 9.** Cumulative porosity of concrete sample with different fly ash additions after 90 days of curing.

**Figure 10.** Pore size distribution after 90 days of curing.

Mehta and Monteriropj [28] classified pore size in concrete into four classes: Class I of <4.5 nm, Class II of 4.5–50 nm, Class III of 50–100 nm, and Class IV of > 100 nm. A pore size of <100 nm has little or no effect, while >100 nm has a negative effect on the concrete strength and permeability. Specifically, the curves in Figure 11 show that the pore volume fraction of Class I and II increased while that of Class III and IV decreased. For example, the pore volume fraction with 4.5–100 nm in PCFA30 increased by about 62.5% and 36.8% compared to PCFA0 and PCFA15, respectively, while that of 100–200 nm and

larger than 200 mm for PCFA15 and PCFA30 considerably reduced to 20.83% and 15.39%, respectively, compared to that of PCFA0. This indicated that pore structures were well-improved as fly ash was added into concrete. This was caused by the second hydration of fly ash with C–H after 90 days of curing generated more C–S–H gel, causing the promotion of concrete compactness and the optimization of pore structures. Figure 12a–c shows the morphologies of concrete samples with different fly ash additions after 90 days of curing. PCFA15 and PCFA30 with fly ash had enhanced concrete density, while more pores could be found for PCFA0 without fly ash replacement. The left-unhydrated fly ash particles may have filled pores in the concrete, as shown in Figure 12d. In this case, fly ash prevented the penetration of free chloride ions and reduced the corrosion of steel rebar, leading to a higher concrete durability.

**Figure 11.** Pore size distribution of concrete with different fly ash additions after 90 days of curing.

**Figure 12.** *Cont*.

(**c**) PCFA30

**Figure 12.** Morphologies of concrete samples with different fly ash additions after 90 days of curing.

#### **4. Conclusions**

This study systematically, experimentally, and analytically investigated the influence of fly ash on chloride diffusion, chloride bonding capacity, and the microstructure of normal concrete in a water soaking environment. The main findings are summarized as follows.


**Author Contributions:** Conceptualization, J.L. (Jun Liu); investigation, J.L. (Jiaying Liu); writing—original draft preparation, Z.H.; validation, J.Z.; writing—review and editing, W.L. and W.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** The financial support from Key-Area Research and Development Program of Guangdong Province (2019B111107002), Natural Science Foundation of China (No. 51708360 and 51978407), Shenzhen City Science and Technology Project (No. JCYJ20180305124008155, JCYJ20170302143133880, JCYJ20180305124106675), and Shenzhen International Cooperation Research Project (No.GJHZ20180928155602083) are gratefully acknowledged.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Properties of Foamed Lightweight High-Performance Phosphogypsum-Based Ternary System Binder**

#### **Girts Bumanis \*, Jelizaveta Zorica and Diana Bajare**

Department of Building Materials and Products, Riga Technical University, LV-1658 Riga, Latvia; jelizaveta.zorica@rtu.lv (J.Z.); diana.bajare@rtu.lv (D.B.)

**\*** Correspondence: girts.bumanis@rtu.lv; Tel.: +371-26062011

Received: 21 August 2020; Accepted: 4 September 2020; Published: 8 September 2020

**Featured Application: This research brings introduction to a novel type of hydraulic binder based on phosphogypsum. Ternary system binder based on phosphogypsum, Portland cement, and pozzolan was elaborated as high-performance material with the strength up to 90 MPa. This binder is characterized by low Portland cement content and low water-binder ratio. Mineralogical and microstructural investigation was performed and data represented. Furthermore, development of lightweight foamed concrete based on developed binder was created using foaming agents.**

**Abstract:** The potential of phosphogypsum (PG) as secondary raw material in construction industry is high if compared to other raw materials from the point of view of availability, total energy consumption, and CO2 emissions created during material processing. This work investigates a green hydraulic ternary system binder based on waste phosphogypsum (PG) for the development of sustainable high-performance construction materials. Moreover, a simple, reproducible, and low-cost manufacture is followed by reaching PG utilization up to 50 wt.% of the binder. Commercial gypsum plaster was used for comparison. High-performance binder was obtained and on a basis of it foamed lightweight material was developed. Low water-binder ratio mixture compositions were prepared. Binder paste, mortar, and foamed binder were used for sample preparation. Chemical, mineralogical composition and performance of the binder were evaluated. Results indicate that the used waste may be successfully employed to produce high-performance binder pastes and even mortars with a compression strength up to 90 MPa. With the use of foaming agent, lightweight (370–700 kg/m3) foam concrete was produced with a thermal conductivity from 0.086 to 0.153 W/mK. Water tightness (softening coefficient) of such foamed material was 0.5–0.64. Proposed approach represents a viable solution to reduce the environmental footprint associated with waste disposal.

**Keywords:** phosphogypsum; ternary binder; high performance; strength; foam; lightweight material; thermal conductivity

#### **1. Introduction**

Gypsum binder is widely used in construction due to its ease of production, availability, and low price [1]. Physically, gypsum is infinitely recyclable; however, the recycling process requires additional energy [2]. Despite these benefits, the disadvantage of gypsum binder is its brittleness, poor resistance to cracking, and unsuitability for damp conditions. Traditional gypsum binder use has been defined in EN 12859, where the main gypsum application is associated with the production of plasters, blocks, tiles, and boards [3]. Besides natural gypsum, synthetic gypsum, produced as chemical by-product, is used widely for the production of gypsum products. There are more than 50 different types of gypsum waste [2]. However, the most common by-product is phosphogypsum (PG), flue gas desulphurization gypsum, and borogypsum [4,5]. Phosphogypsum (PG) is produced as a by-product from phosphate fertilizer production and the annual PG production reaches 280 mlj.t worldwide and

only about 15% of PG is used as secondary raw material, but rest is disposed in open-type stacks [5]. Partially or completely, synthetic gypsum can be a substitute for natural gypsum as cement admixtures, gypsum-based plasters, drywalls. Researches on production of traditional gypsum binders based on PG are widely published, but they have rather limited practical application due to specific nature of PG [6]. Moreover, there are legislation limits and prejudice coming from society regarding PG, so the direct use of PG as substitution of natural gypsum is problematic. More complicated and effective way of the utilization of PG is to create an advanced and new type of binder, which has a much lower carbon footprint comparing to Portland cement, while remaining strength properties similar to Portland cement. It was reported that the addition of blast furnace slag and cement could effectively improve the mechanical strength of PG. However, fly ash played a negative role on the compressive strength of PG [7]. S. Kumar described properties of fly ash–lime–phosphogypsum ternary binder [8] or recently anhydrous gypsum was used to develop lime-pozzolan green binder [9]. Two waste-stream materials were used and only lime was defined as a primary resource, the calcination temperature (900–1100 ◦C) of latter is lower compared to Portland cement. In this case, the amount of PG in the binder was in the range from 10–40%. However, the disadvantage comparing to Portland cement is low compressive strength—which could be in range from 2–4 MPa [8]; nevertheless it is reasonable if it is compared to the lime binder and hydraulic lime binder. Higher strength results were obtained in ternary binder system phosphogypsum–steel slag–granulated blast-furnace slag (GGBS)–limestone cement, where the content of PG was from 25–65%, while the amount of slag was from 22–48%. The obtained strength at the age of 28 days was up to 45 MPa while the obtained binder is characterized with fast setting time (initial setting time 6–9 min, final 10–12 min) [10]. These results are more comparable to traditional binders, while the problem could be a fast setting time. The fast setting may not benefit the engineering application because there is not enough time for casting before the cement sets. In some case it was reported that the citric acid in amount from 0.03–0.15 wt.% of cement could retard setting time significantly. The use of citric acid could increase the open time from 25 to 47 min, while this admixture tended to slightly reduce compressive strength of the binder [11]. In ternary systems where Portland cement is present, superplasticizer can be used and low water-binder ratio can be achieved. Traditionally, supplementary materials that are utilized to replace ordinary Portland cement decrease the workability of the cementitious mixtures and superplasticizers such as polycarboxylate based are usually added to cement to control their fluidity [12]. The use of polycarboxylate acid-based superplasticizer could be used from 0.75–1.75 wt.%; however, reports say that it could slightly reduce early compressive strength of the material while final strength tended to increase [11]. These aspects regarding to the utilization of PG in new types of binders were considered in present research by choosing mixture composition, including use of chemical admixtures.

To continue the development of alternative waste stream binders in a production of new materials and enhance its valorization possibilities, novel lightweight foam material based on developed ternary binder was elaborated. In construction industry, there is growing interest in lightweight concretes. It combines positive properties of constructive and insulation materials and is characterized by moderate strength, low density, and improved thermal properties. Cellular concrete is composed on mortar matrix and specially created system of air cells, which occupies up to 85% of material volume [13]. High porosity limits potential of mechanical strength, but high volume of open pores is the main reason for increased water absorption and drying shrinkage. The density of traditional gypsum ranges from 600 to 1500 kg/m3, as given in Clause 4.8.1. of EN 12859 [3]. This well-known standard covers the gypsum application range, and beyond this range, research is being conducted to make gypsum material more sustainable. Attempts to produce lightweight gypsum with foaming admixtures have yielded a material with density ranging from 300 to 600 kg/m<sup>3</sup> [14]. Such material has low density, superior sound, and thermal insulation and can be considered a sustainable high-performance material. High-efficiency sound-absorbing material was made also with PG, which composite structure was described by Baoguo Ma et al. [15]. Thus, aim of the work was to produce in laboratory conditions lightweight ternary system based material with density less than 600 kg/m3, which is outside the

traditional boundaries to bring the novelty of the research. Bulk density and thermal conductivity were set as target values, which should be determined together with technological properties so that gypsum material could be easily produced and handled (workability, strength). Here, research on development of highly porous ternary system gypsum-based binder material was evaluated and compared.

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

Dihydrate phosphogypsum (CaSO4·2H2O) analyzed in this work is a waste generated by fertilizer production plant AB Lifosa (Kedainiai, Lithuania) in wet-process phosphoric acid production, where ˙ apatite from the Kovdor mine, Kola peninsula, Russia, is decomposed by sulphuric acid. To produce 1t of orthophosphoric acid, about 3.0–4.5 t of PG is obtained [16]. At the enterprise, volcanic origin Cola apatite (Kirov and Kovdor) (containing F2 1–2%, P2O5—about 38% as phosphorus) is used as a raw material in the production of phosphoric acid. Besides, phosphorites of sedimentary origin from other countries (Morocco, Jordan, Kazakhstan (Karatau), Algeria, South Africa Republic) are used. The dihydrate PG was used as secondary raw material in the present research. Chemical composition of raw materials is given in Table 1. The initial pH of PG is in the range from 2.2–2.9 while during the storage it can increase gradually. The PG was dried at 60 ◦C (moisture content 9–12 wt.%) and milled to powder-like particles with collision milling in semi-industrial disintegrator with the rotational speed of 50 Hz. The particle size distribution is given in Figure 1. Then, the calcium sulfate hemihydrate binder was obtained by treatment of milled gypsum powder at temperature 180 ◦C for 4 h. Commercial gypsum plaster (BG) was used to compare the characteristic technological properties of obtained binder.


**Table 1.** Chemical composition of raw materials used to prepare novel building material [weight %].

**Figure 1.** The particle size distribution of the raw materials used to prepare novel building material.

The other components of the binder were metakaolin as supplementary cementitious material and Portland cement CEM I 42.5N (CEM). Chemical composition of CEM is given in Table 1. The initial setting time was 182 min and the final setting was 224 min (according to LVS EN 196-3); the normal consistency was 28.2% (according to LVS EN 196-3). Blaine fineness of cement was 3787 cm2/g (according to LVS EN 196-6). Na equivalent was 1.68. Compressive strength after 1 day was 15.4 MPa, after 2 days—32.9 MPa, after 7 days—48.8 MPa, and after 28 days—60.5 MPa. Waste metakaolin (MKW) was obtained from the porous glass granule production factory "Stikloporas" JSC (Druskininkai, Lithuania). Metakaolin is a by-product from the final stage of expanded glass granule production, where kaolinite clay powder is used as a substance for anti-agglomeration and is heated at 850 ◦C temperature for 40–50 min. MKW is mostly amorphous dehydration product of kaolinite, Al2(OH)4Si2O5, which exhibits strong pozzolanic activity often used in the concrete industry as a supplementary cementitious material. XRD pattern and detailed description of MKW was published before [17]. MKW contains amorphous metakaolin detected in 2θ region from 15 to 30◦, and some minor crystalline phases were also detected in MKW: quartz (SiO2) and Halloysite Al2Si2O5(OH)4. The specific surface area of the MKW is 15.86 m2/g. Loss of ignition for metakaolin was 11.5%. The particle size distribution of materials used in the preparation of ternary binder is given in Figure 1.

Sika ViscoCrete G-2 is a high-performance superplasticizer/water reducer based on polycarboxylate (PCE) polymer technology. PCE is formulated for applications in systems with high calcium sulfate content or pure gypsum-based binders. Gips RETARD (TKK) is a powdered citric acid-based admixture, which was used for regulating the setting time of gypsum. Gips RETARD was added to the dry binder mixture and mixed thoroughly before water was added. Sodium dodecyl sulfate (SDS) ≥85%, pure (Acros Organics), was used as foam stabilizer for preparation of foamed samples. To produce porous, lightweight ternary system material, anion surface-active substance with stabilizing and functional agents PB-Lux was used.

Prepared mixture composition is given in Table 2. First two compositions were ternary system binder pastes (GCP-PG and GCP-BG). The difference between two of them is that in one case, commercial gypsum binder is used (abbreviation with BG), while in the other one PG binder is used (abbreviation with PG). The amount of PCE was 1.5% from the total amount of binder while set retarder R was 0.2%. Set retarder was used for samples with PG as it has a shorter set time comparing to commercial gypsum [6]. The W/B ratio was 0.34 for both mixtures. Other two mixtures were based on two mentioned binder pastes, but additional washed quartz sand (0/4 mm) was incorporated into the mixture composition with binder-sand ratio 1:1.75. The W/B ratio slightly increased with the incorporation of sand.


**Table 2.** Mixture composition of prepared ternary system binder paste and ternary system binder mortar.

Mixing procedure of binder pastes was similar to traditional cement mortars. First, dry components including set retarder were homogenized for 2 min. Then, powder gradually was added into the water and superplasticizer mixture and homogenized for 2 min before casting. Mortar was mixed similar to the pastes, while the sand and extra water were added to the pre-mixed paste. Samples were cast in 20 × 20 × 20 mm or 40 × 40 × 160 mm molds for further testing.

The mixture composition of foamed ternary system binder is given in Table 3. Four mixture compositions were prepared. The amount of foam that was prepared was similar for all mixture compositions. The amount of binder paste homogenized with foams was changed for both compositions. SDS stabilizing admixture was used during the preparation of foams. Mixing procedure of binder pastes was similar to traditional cement mortars. First, dry components including set retarder were homogenized for 2 min. Then, powder gradually was added into the water and superplasticizer

mixture and homogenized for 2 min before introduction in simultaneously prepared foams. Then, binder paste was cast in foams and homogenized for 3 min. Both commercial gypsum BG and phosphogypsum PG was compared. Plate samples with dimensions 35 × 35 × 5 cm were prepared for thermal conductivity measurements. Samples were further cut to cubical specimens 5 × 5 × 5 cm and physical properties as well as mechanical properties were tested.


**Table 3.** Mixture compositions of foamed ternary system binder.

The setting time of the obtained binder was tested using the Vicat apparatus. Consistency of fresh mortar was tested with the Suttard's viscometer. Early age (24 h) and 28 d compressive strength was determined. After 24 h, samples were cured in moist conditions (RH 95%). At the age of 35 d, air dry samples were tested. Samples with dimension of 20 × 20 × 20 mm were tested using Zwick Z100 with testing speed of 0.5 mm/min. The specific gravity and total porosity were determined by using Le Chatelier flask. Hardened samples were ground to powder with planetary ball mill Retch PM400 for 10 min with a speed of 300 rpm, and obtained powder was used to determine specific gravity. Total porosity was calculated from bulk density and specific gravity. Water absorption was measured for prismatic specimens for binder and mortar (40 × 40 × 160 mm) and cubical specimens for porous samples (50 × 50 × 50 mm). Samples were dried at 60 ◦C until a constant mass was obtained, then samples were immersed in water for 72 h and saturated mass was obtained. Then, water absorption was calculated and open porosity determined using the volumetric measurements of individual samples. The mineralogical composition was determined by X-ray diffraction (XRD) (PAN analytical X'Pert PRO). The macrostructure of the material was observed by a digital microscope at a magnification of 40 and 120x. Scanning Electron Microscopy (SEM) with Energy Dispersive Spectroscopy (EDX) was used (JEOL JSM 820 + IXRF systems 500 digital processing, Japan) for microscopic analysis of binder paste with 20 kV voltage. The thermal conductivity of the materials was determined using the LaserComp heat meter FOX600, according to the guidelines of Standard LVS EN 12667. Foamed ternary system binder plate specimens with dimensions of 35 × 35 × 5 cm were prepared for thermal conductivity test. Temperature difference between testing plates was 20 ◦C (the bottom plate was +20 ◦C and the upper plate was 0 ◦C).

#### **3. Results and Discussion**

Results are represented in two main subsections where dense ternary system binder properties are described (Section 3.1) and in Section 3.2 where properties of lightweight ternary binder are represented.

#### *3.1. Properties of High-Performance Ternary System Binder and Mortar*

The ternary system binder was described through its appearance in macro and micro level, chemical and mineralogical composition through hydration processes. Fresh and hardened properties of the binder and mortar were determined and results represented.

#### 3.1.1. Appearance of Ternary System Binder and Mortar

The macrostructure of GCP-PG mortar is given in Figure 2. The difference between binder depending on the gypsum type used (PG or BG) did not influence the appearance of the macrostructure of the material. Mortar has a homogenous structure where individual sand filler grains and the binder can be identified. As 50 wt.% of the binder is gypsum, the material has a white appearance that is characteristic for gypsum binders. The whitish appearance could be an advantage for the creation of exposed architectural or structural elements. Small black inclusions as remains of foam glass granules originated from waste metakaolin were indicated, which could be a negative factor influencing the strength of the binder and mortar.

**Figure 2.** Macrostructure of ternary system mortar (**a**) under magnification of 40x, (**b**) under magnification of 120x.

The microstructure of the ternary binder (GCP-PG and GCP-BG) is given in Figure 3. It has a complex fine-grained structure with a large monolithic structure in smaller magnification (500x). The magnification at 2000x reveals the interaction between binder components and different regions, which can be identified.

**Figure 3.** Microstructure of the ternary binder detected by SEM in magnification of 500x and 1000x. (**a**,**c**) GCP-phosphogypsum (PG), (**b**,**d**) GCP-BG.

#### 3.1.2. Chemical and Mineralogical Characterization

XRD results are given in Figure 4. The mineralogical composition of a ternary binder powder was investigated before hydration (Figure 4c,d). Bassanite (CaSO4·0.5H2O, ref 33-0310), calcium silicate oxide (Ca3(SiO4)O, ref 73-0599), and quartz (SiO2, ref 78-1252) were identified from the source dry materials of GCP-PG. GCP-BG had the same three minerals as for GCP-PG, while additionally a dolomite (CaMg (CO3)2, ref 36-0426) and an anhydrite Ca(SO4), ref 72-0916) were also detected, which is associated with the composition of commercial BG.

**Figure 4.** Mineralogical characterization of raw material mixture and hydrated paste of ternary binder: (**a**) Hydrated GCP-PG; (**b**) hydrated GCP-BG, (**c**) raw mixture of GCP-PG, (**d**) raw mixture of GCP-BG. B—Bassanite (CaSO4·0.5H2O, ref 33-0310), C—calcium silicate oxide (Ca3(SiO4)O, ref 73-0599), Q—quartz (SiO2, ref 78-1252), D—dolomite (CaMg (CO3)2, ref 36-0426), A—anhydrite Ca(SO4), ref 72-0916), G—gypsum (Ca(SO4)(H2O)2 (74-1433), E—ettringite (Ca6(Al(OH)6)2(SO4)3(H2O)26, ref 72-0646).

The phase change after binder hydration was detected (Figure 4a,b). Moreover, ternary binder contained gypsum, cement, and pozzolan, the only phases which were identified in a hydrated GCP-PG mixture where gypsum (Ca(SO4)(H2O)2 (74-1433), quartz (SiO2, ref 78-1252), and ettringite (Ca6(Al(OH)6)2(SO4)3(H2O)26, ref 72-0646) were present. The formation and presence of the ettringite in such a late hydration stage are associated with high gypsum content in the mixture. In such a way, expansions can occur from excessive calcium sulfoaluminate formation after hardening and continue until the gypsum becomes depleted, that is why pozzolan was added and, according to previous studies, addition of pozzolan could even eliminate delayed ettringite formation [18]. In mixture composition, GCP-BG ettringite was not detected. Besides minerals coming from raw materials (dolomite, quartz), gypsum (Ca(SO4)(H2O)2 (74-1433) and also gypsum hemihydrate Ca(SO4)(H2O)0.5 were detected, which could be formed slowly from the hydration of the anhydrite.

The EDX analysis of the GCP-PG binder is given in Figure 5. Large amount of Ca element was identified in most EDX points analyzed, which is associated with the fact that Ca is present in all binder components (phosphogypsum, cement, and metakaolin). The gypsum source in the structure can be identified through element S, which comes from PG binder (CaSO4 0.5H2O). Small quantities of Al and Si were identified as elements coming from metakaolin and cement. This explains the fact that no cement minerals were identified by XRD. The XRD analysis of raw material mixture and hardened binder was investigated while no strong new mineralogical peaks were identified. Mostly transition between gypsum hemihydrate and gypsum dihydrate was observed.

**Figure 5.** Energy Dispersive Spectroscopy (EDX) analysis of GCP-PG binder. (**a**) point positions analyzed by EDX, (**b**) element content in analyzed region.

#### 3.1.3. Physical and Mechanical Properties

The bulk density of hardened binder is given in Table 4. The density of binder paste GCP-BG is slightly lower comparing to GCP-PG (1609 to 1766 kg/m3), which can be explained by the finer particle distribution for PG and denser structure of the paste. Similar tendencies were observed for mortar. The bulk density of a mortar material increased to 1886 and 2027 kg/m3 for GCP-BG-S and GCP-PG-S, respectively. The total porosity for GCP-PG and GCP-BG was 23 and 29 vol.%, while for mortar it was even lower −11 and 17 vol.% for GCP-PG-S and GCP-BG-S. The low porosity is associated with low W/B ratio, which is also similar to traditional concrete nature [19]. This phenomenon had an influence also on the compressive strength of the materials presented.


**Table 4.** Physical properties of high-performance ternary system binder.

The consistency of fresh material was strongly affected by the amount of superplasticizer and low water content. The initial mixing of binder powder with water and plasticizer gives stiff mixture while during the intensive mixing, the effect of superplasticizer takes place and very viscous mixture was obtained with flow diameter −295 mm for GCP-BG and 370 mm for GCP-PG. The low W/B ratio of the paste (GCP-PG and GCP-BG) gives similar consistency as for ultra-high performance concrete [20]. The flow for GCP-PG was higher, which is associated with the finer particle nature of PG. The prepared mortar was less viscous comparing to paste while the effect of superplasticizer was also present and high workability obtained. The flow of mortar was from 210–230 mm.

The setting time was longer for GCP-PG and GCP-PG-S as set retarder was purposely used in mixture composition. The initial setting time was from 130–165 min for binder paste and mortar. The final set time was longer for GCP-PG (280 min) than for mortar GCP-PG-S (170 min) as it was more viscous to initiate the set of the paste. GCP-BG and GCP-BG-S initial setting time was slightly influenced by the mineral additives (tin 70 and 75 min, respectively), while the final set time was longer for GCP-BG (130 min) and was reduced for mortar GCP-BG-S (85 min).

The compressive strength results are given in Figure 6. The compressive strength increased during the water curing similar as for Portland cement binder. The strength after demolding at 24 h was similar among all samples (13–15 MPa). The further curing resulted in slower strength gain for mortars (GCP-PG-S and GCP-BG-S). At the age of 28 d, the strength of moist samples reached 50 MPa for GCP-PG and 30 MPa for GCP-BG. The strength increase was not observed between 7 and 28 d for GCP-BG. The mortar strength reached 26 MPa for GCP-BG-S and 43 MPa for GCP-PG-S. After sample drying, material strength increased significantly at the age of 35 d. The GCP-PG reached 88 MPa while GCP-PG-S mortar reached 50 MPa. Binder GCP-BG had 49 MPa and mortar GCP-BG-S had 39 MPa, respectively. The failure of the mortar was mostly in the transition zone of aggregate and binder. Such result is highly competitive and promising compared to traditional Portland cement mortars. The higher strength results obtained for materials based on PG could be explained by the fine nature of PG giving more reactive nature for ternary binder. Moreover, commercial gypsum BG could contain other additives that could lead to lower strength results.

**Figure 6.** The compressive strength of ternary binder and mortar.

#### *3.2. Properties of Foamed Ternary System Binder*

The porous ternary system binder was described through its appearance in macro and micro levels. Physical properties such as density, thermal conductivity as well as mechanical properties such as compressive strength and compressive strength softening coefficient regarding water saturation were determined.

#### 3.2.1. Appearance of Foamed Ternary System Binder

The macrostructure and microstructure of foamed ternary system binder are given in Figure 7. The material appearance is light grey, similar to Portland cement-based materials. The structure of all samples is monolithic and materials are easy to handle. The macrostructure is highly porous for all samples. The pores are distributed uniformly throughout the sample. The pore macrostructure for samples made with BG has larger pores comparing to samples made with PG. This was also confirmed by the pore measurements with digital microscope. Microstructure is generally homogenous while still different particles can be observed in the matrix. The samples with PG have smaller pores that could be evaluated more precisely in micro-level. Depending on the mixture composition, the increase of binder paste in the same foam amount leads to smaller pores remaining the characteristics coming from gypsum source described before. For GCP-BG-1, macro pores with size in range from 1–4 mm were formed while smaller pores were detected within the large pore walls. The smaller pores were in

range from 0.2–1.0 mm. GCP-BG-2 had more uniform large pores in a range of 2 mm, while smaller pores were observed in the range from 0.2–0.6 mm. GCP-PG-1 had pores up to 0.7 mm. While most of the pore size was in range from 0.1–0.5 mm. For GCP-PG-2, the structure was similar to GCP-PG-1. The pore size was in range from 0.1–0.6 mm.

(**a**)

(**b**)

(**c**)

(**d**)

**Figure 7.** Macrostructure and microstructure of ternary system binder foams. (**a**) GCP-BG-1, (**b**) GCP-BG-2, (**c**) GCP-PG-1, (**d**) GCP-PG-2.

The microstructure of samples observed with SEM is given in Figure 8. The pore hierarchical structure remained the same as previously observed with digital microscope. For mixture GCP-BG, large pores are also visible in SEM. In large magnification, the structure similar to cement-based composites is visible. It has a complex fine-grained structure with the large monolithic structure in smaller magnification. The magnification at 500x reveals the interaction between binder components and different regions, which can be identified.

(**a**)

(**b**)

**Figure 8.** The micrographs of ternary system binder foam observed with SEM. (**a**) GCP-BG-1, (**b**) GCP-PG-1.

#### 3.2.2. Physical and Mechanical Properties

The physical properties of ternary system binder foams are given in Table 5. Due to larger pores, the bulk density of GCP-BG foams is lower comparing to GCP-PG foams. The lightweight material with density 368 and 486 kg/m<sup>3</sup> was obtained depending on the amount of paste mixed with foams. GCP-PG foams density increased to 634 and 697 kg/m3, respectively, which is associated with the smaller pores formed in the production process. The total porosity was higher for GCP-BG, which explains the lower bulk density. The total porosity was 84 and 79 vol.% for GCP-BG and 72 and 70 wt.% for GCP-PG. Important finding was that open porosity was formed for GCP-BG foams, while for GCP-PG more pores were closed. This is also evidenced by the fact that during the water absorption test the GCP-PG foams remained flowing after total immersion in water (total test time was 72 h). The water absorption was 105 and 91 wt.% for GCP-PG foams, while for GCP-PG it was 46 and 39 wt.%, respectively. Density for water-saturated samples increased and was in range from 758–901 kg/m3.



The lowest thermal conductivity of produced specimens was for GCP-BG-1. With the lowest bulk density, the thermal conductivity was as low as 0.086 W/m·K. The increase of bulk density for GCP-BG increased the thermal conductivity to 0.123 W/m·K. Furthermore, GCP-PG-1 had higher bulk density, and due to finer pore size distribution, the thermal conductivity was lower comparing to GCP-BG-2 and it was 0.111. The GCP-PG-2 had thermal conductivity of 0.153 W/m·K.

The compressive strength results of prepared ternary system binder foams are given in Figure 9. The compressive strength of GCP-BG-1 was 0.33 MPa and increased to 1.0 for GCP-BG-2. The increase of compressive strength was observed for GCP-PG. For GCP-PG-1, it was 2.4 MPa and increased to 5.4 MPa for GCP-PG-2. The strength increase is associated with the increase of foam bulk density and with the size of macropores formed during production. For samples made with PG, pores were smaller. Foam samples after water absorption tests were tested for compression to determine softening coefficient of the foams. Since gypsum is not watertight material and in proposed binder it is a major part of the binders' component (50 wt.%), concerns regarding water tightness were analyzed (Table 6). It was detected that water softening coefficient was from 0.50–0.64. Similar trends were observed with concrete foams before—the compressive strength decreases with the increase of volumetric moisture content for every density ranging from 300–800 kg/m3, but the softening rate was lower—the compressive strength in the immersed saturation state is about 0.86–0.89 times of that in the standard curing state [21].

**Figure 9.** The relation between bulk density and thermal conductivity of foamed ternary system binder.

**Table 6.** Physical properties of foamed ternary system binder.


#### **4. Conclusions**

High-strength ternary system binder was developed containing major part of gypsum, metakaolin, and Portland cement. Valorization option for waste phosphogypsum and metakaolin was offered by incorporating these materials in high-performance cementitious composites and by producing highly porous lightweight foam materials. Proposed binder has technological properties similar to traditional cementitious binders based on Portland cement. The mixture composition with low W/B ratio of 0.34 was developed with the addition of superplasticizer. The set time of binder was adjusted by the help of the set retarder and it is comparable to Portland cement. Impressive compressive strength of up to 88 MPa was reached. The binder proved to be suitable to produce mortar with a strength of up

to 50 MPa. Samples were prepared by a simple methodology that, together with the use of wastes, contributes to improve sustainability of the process. Foamed ternary system binder material with density in range from 368–697 kg/m<sup>3</sup> was obtained with compressive strength from 0.33 to 3.5 MPa. Besides technological properties, long-term properties such as durability, shrinkage/swelling should be evaluated, as gypsum with cement can form hazardous compounds such as ettringite, which could lead to loss of integrity of the material.

**Author Contributions:** Conceptualization, G.B. and D.B.; methodology, G.B. and D.B.; validation, G.B., D.B., and J.Z.; formal analysis, G.B.; investigation, G.B and J.Z.; resources, G.B. and D.B.; data curation, G.B.; writing—original draft preparation, G.B.; writing—review and editing, G.B., J.Z., and D.B.; visualization, G.B. and J.Z.; supervision, G.B. and D.B.; project administration, G.B. and D.B.; funding acquisition, G.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work has been supported by the European Regional Development Fund within the Activity 1.1.1.2 "Post-doctoral Research Aid" of the Specific Aid Objective 1.1.1 "To increase the research and innovative capacity of scientific institutions of Latvia and the ability to attract external financing, investing in human resources and infrastructure" of the Operational Program "Development of sustainable and effective lightweight building materials based on secondary resources" (No. 1.1.1.2/VIAA/1/16/050).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **A Low-Autogenous-Shrinkage Alkali-Activated Slag and Fly Ash Concrete**

#### **Zhenming Li 1,\*, Xingliang Yao 1,2, Yun Chen 1,3, Tianshi Lu <sup>1</sup> and Guang Ye 1,4**


Received: 17 August 2020; Accepted: 31 August 2020; Published: 2 September 2020

**Abstract:** Alkali-activated slag and fly ash (AASF) materials are emerging as promising alternatives to conventional Portland cement. Despite the superior mechanical properties of AASF materials, they are known to show large autogenous shrinkage, which hinders the wide application of these eco-friendly materials in infrastructure. To mitigate the autogenous shrinkage of AASF, two innovative autogenous-shrinkage-mitigating admixtures, superabsorbent polymers (SAPs) and metakaolin (MK), are applied in this study. The results show that the incorporation of SAPs and MK significantly mitigates autogenous shrinkage and cracking potential of AASF paste and concrete. Moreover, the AASF concrete with SAPs and MK shows enhanced workability and tensile strength-to-compressive strength ratios. These results indicate that SAPs and MK are promising admixtures to make AASF concrete a high-performance alternative to Portland cement concrete in structural engineering.

**Keywords:** alkali-activated concrete; shrinkage; cracking; internal curing; metakaolin

#### **1. Introduction**

An important way to reduce the CO2 emissions from the construction sector is to use "greener" alternative binders to ordinary Portland cement (OPC). Alkali-activated materials (AAMs), or so-called geopolymers, which can be made of industrial by-products, have been reported to entail much lower CO2 emission and embodied energy than OPC systems [1].

Blast-furnace slag (indicated as "slag" hereafter) and fly ash, as by-products from steelmaking and coal-fired electricity plants, respectively, are the two most commonly utilized precursors to produce AAMs. The literature has illustrated that alkali-activated slag and fly ash (AASF) shows superior strength, excellent durability and good fire resistance compared to OPC systems [2–4]. However, a wider application of this material has not been reached yet, partly due to the large autogenous shrinkage and the potential risk of cracking of this material, especially when NaOH and Na2SiO3 are used as activators [5,6].

A number of studies have been conducted to reduce the shrinkage of AASF. However, it has been found that the shrinkage-reducing agents and expansive additives that are widely adopted in OPC may be ineffective or cause side effects (e.g., strength loss) in AAMs due to the differences in microstructures and chemical environments between AAMs and OPC [7,8]. Elevated temperature curing can mitigate the shrinkage of AASF [9], but this strategy has high requirements on the curing equipment and can accelerate the setting of AASF, which is already more rapid than usually needed. Feasible strategies to mitigate the autogenous shrinkage of AASF are desired to widen the commercial acceptance of this material.

The results of previous studies [10–13] indicate that internal curing and incorporation of a small amount of metakaolin (MK) are helpful to reduce the autogenous shrinkage of alkali-activated slag (AAS) and AASF. However, both of these two strategies have their limitations. Internal curing can significantly mitigate the autogenous shrinkage caused by self-desiccation, but on the first day when autogenous shrinkage rapidly develops, the effect of internal curing is quite limited due to the possible involvement of other shrinkage mechanisms [13]. By contrast, the incorporation of MK can effectively mitigate the early-age autogenous shrinkage of AAS and AASF by reducing the high reaction rate in the acceleration period and coarsening the gel pores [10]. However, the effect of MK on later-age autogenous shrinkage is not evident. These results indicate that these two admixtures might be a good complement to each other to further lower the autogenous shrinkage of AAMs. However, the combined effect of them on the autogenous shrinkage and cracking properties of AASF systems have not been studied yet.

In this study, superabsorbent polymers (SAPs) and MK are applied to reduce the autogenous shrinkage of AASF. Experiments are conducted at both paste and concrete scales. The cracking potential of paste and concrete is evaluated by the ring test and Temperature Stress Testing Machine (TSTM), respectively. The workability and the mechanical properties of the concrete are also investigated. Eventually, with the addition of SAPs and MK, a high-performance eco-efficient alkali-activated concrete is produced.

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

#### *2.1. Raw Materials*

The main precursors used in this study were slag and Class F fly ash. The material parameters of these two precursors are shown in Tables 1 and 2.


**Table 1.** Chemical compositions of slag, fly ash and MK.


**Table 2.** Particle size and density of slag, fly ash and MK.

Bulk-polymerized SAPs with particle sizes up to about 200 μm in the dry state were used. The SAPs were a cross-linked copolymer based on acrylamide and acrylate. The composition and physical properties of MK are also shown in Tables 1 and 2, respectively. It should be noted that part of the high percentage of silica content in MK was due to the presence of quartz (43% in weight), which remains inert during the reaction process [14]. No admixture besides SAPs and MK was used.

Pellets of NaOH, deionized water and commercial water glass solution were used to synthesize the alkaline activator. For 1000 g of precursor, an activator containing 384 g of water, 1.146 mol of SiO2 and 0.76 mol of Na2O was applied.

Extra activator for

#### *2.2. Mixtures*

The mixture design of the plain AASF paste and AASF paste with SAPs and/or MK is shown in Table 3.


internal curing - 0.064 - 0.064 MK - - 0.1 0.1

**Table 3.** Mixture proportions of AASF paste with SAPs and/or MK.

The dosage of SAPs was designed based on the absorption capacity of the SAPs (20 g/g activator [11]) and the chemical shrinkage of AASF paste. The ultimate chemical shrinkage of AASF paste was determined by taking the chemical shrinkage measured by dilatometry untill the age of 28 days [15,16], which was 0.026 mL/g. Taking also into account the density of the activator as well, 1.23 g/cm3, the extra liquid provided by the SAPs should be 0.032 g per gram of binder. Therefore, the adding amount of 0.16 wt. % SAP was applied for the internal curing of AASF.

According to the results from previous studies [10,17], 5 wt. % of MK in the binder can already show a significant mitigating effect on the autogenous shrinkage of AASF paste. Higher amounts of MK may lead to considerable strength loss in the matrix. Therefore, the dosage of MK in this study was chosen as 5 wt. % of the binder as a substitution for 10 wt. % of slag.

Based on the results at the paste scale (details will be given in Sections 3.1 and 3.2), two mixtures, AASF and AASFICMK, were chosen for experiments at the concrete scale. The mixture proportions of the concrete mixtures are shown in Table 4.


**Table 4.** Mixture proportions of AASF and AASFICMK concrete.

#### *2.3. Experimental Methods*

#### 2.3.1. Autogenous Shrinkage of Paste

The autogenous shrinkage of the paste was measured by the corrugated tubes method according to ASTM C1968 [18]. The detailed procedure can be found in [19].

#### 2.3.2. Cracking Potential of Paste

The cracking potential of the paste induced by restrained autogenous shrinkage was indicated by the dual ring test [20,21]. The geometry of the rings is shown in Figure 1. The strain gauges attached to the inner surface of the inner steel ring started to record the strains of the inner ring when the paste was cast in between the two steel rings. The paste ring was sealed by aluminium foil fixed with asphalt

tape during the test to avoid moisture loss. The apparatus was put in a temperature-controlled room with the temperature fixed at 20 ◦C.

**Figure 1.** Dimensions of the steel rings, after [22].

The maximal stress in the paste was calculated according to Equation (1) based on the strain and dimensions of the rings [23].

$$
\sigma\_{\text{max}} = -\boldsymbol{\upsilon} \cdot \boldsymbol{E}\_{\text{steel}} \left( \frac{R\_{IP}^{-2} - R\_{II}^{-2}}{2R\_{IP}^{-2}} \right) \left( \frac{R\_{OP}^{-2} + R\_{IP}^{-2}}{R\_{OP}^{-2} - R\_{IP}^{-2}} \right) \tag{1}
$$

where σ*max* is the maximum stress in the paste ring; ε is the measured strain of the inner steel ring; *Esteel* is the elastic modulus of the ring; *RII*, *RIP* and *ROP* are the inter radius of the inner steel ring (75 mm), the inner radius of the paste (87.5 mm) and its outer radius (125 mm), respectively.

#### 2.3.3. Workability of Concrete

The workability of AASFICMK concrete was measured. The slump was measured according to NEN 12350 [24]. The largest diameter of the flow spread of the concrete and the diameter of the spread at right angles to it were measured immediately after the cone was lifted up. The setting time of AASFICMK concrete was determined by the Vicat method [25] on the corresponding paste.

#### 2.3.4. Autogenous Shrinkage of Concrete

The autogenous shrinkage of the concrete was measured with an Autogenous Deformation Testing Machine (ADTM) [26]. The prismatic mold for the concrete was made of thin steel plates and external insulating materials. The size of the mold is 1000 <sup>×</sup> <sup>150</sup> <sup>×</sup> 100 mm3 (see the top left corner of Figure 2). The mold was connected with cryostats by a series of circulation canals located between the plates and the insulating material.

The length change of the concrete was measured with two external quartz rods located next to the side mold. Linear Variable Differential Transformers (LVDTs) were installed at both ends of the rods. The LVDTs measured the movement of the steel bars, which were cast in the concrete. The distance between the two embedded steel bars was 750 mm. The measurement of the deformation of AASF and AASFICMK concrete started at 11h and 12 h after casting, respectively, when the concrete was stiff enough to hold the measuring bars and the LVDTs that were connected with them. Attention was paid to the sealing of the molds in order to avoid moisture loss to the environment.

#### 2.3.5. Cracking Potential of Concrete

The cracking initiation in the concrete was monitored by a TSTM. The TSTM was equipped with a horizontal steel frame in which compressive and tensile force could be applied on the concrete specimen. A temperature-controlled mold was used for the concrete casting in order to obtain any prescribed thermal condition. The mold was similar to the ADTM mold described in Section 2.3.4. The whole specimen was of a dog-bone shape and the testing area of interest was of a prismatic shape

(1000 <sup>×</sup> <sup>150</sup> <sup>×</sup> 100 mm3), see the bottom left corner of Figure 2. The deformation of the concrete was kept at zero (nominally, in reality within ±1 μm range) so that a full restraint condition could be reached. When the total deformation of the concrete went beyond the threshold, a load was applied to pull or push the concrete back to the original position. The load was recorded with the load cell with a loading capacity of 100 kN and a resolution of 0.049 kN. A sudden drop in the load to around zero indicated the occurrence of cracking in the concrete.

#### 2.3.6. Strength of Concrete

Concrete cubes (150 <sup>×</sup> 150 <sup>×</sup> 150 mm3) for the compressive and splitting strength tests were cast and cured in sealed and temperature-controlled steel molds (see the top middle of Figure 2). The moulds were connected with cryostats by parallel circulation tubes and the upper surface was sealed by plastic film.

The compressive strength and splitting strength of the concrete were measured according to NEN-EN 12390 [27]. The measurements were conducted at the age of 1, 3, 7 and 28 days and the day when the concrete beam in the TSTM cracked. One cube was tested for compressive strength and two for splitting strength.

**Figure 2.** Overview of the set-up for the concrete properties, after [28]. Top left: ADTM. Top middle: Cubes. Bottom left: TSTM. Bottom right: controlling systems.

To keep consistency of the materials, the concrete samples for strength, autogenous shrinkage and cracking potential measurements were from the same batch of casting. All the samples were cured in a sealed condition. The whole set-up, including the TSTM, ADTM, cubes and controlling systems is schematically shown in Figure 2. Various thermocouples were used to monitor the temperatures of the samples. To minimize the influence of thermal deformation on the autogenous shrinkage, the temperatures of the middle parts of the specimen, i.e., T3 for the TSTM, T7 for the ADTM, and T9 for the cubes, were controlled at 20 ◦C.

#### **3. Results and Discussion**

#### *3.1. Autogenous Shrinkage of Paste*

The autogenous shrinkage curves of the paste mixtures are shown in Figure 3. The autogenous shrinkage of AASF paste reached more than 2000 μm/m at 1 day and around 4000 μm/m at the age of 7 days. This magnitude is higher than the autogenous shrinkage of common OPC-based systems irrespective of the presence of supplementary materials [29,30]. The shrinkage mechanism was discussed in a previous study [19]. It can be seen from Figure 3 that both the additions of SAPs and MK resulted in lower autogenous shrinkage of AASF paste. In particular, the addition of SAPs greatly mitigate the autogenous shrinkage of AASF paste after the first day. By contrast, MK was more effective on the first day; afterward, the effect of MK became less evident.

**Figure 3.** Autogenous shrinkage of AASF paste with SAPs and/or MK.

When SAPs and MK were added together into AASF, the autogenous shrinkage of the paste was the lowest among all the four mixtures in the whole week. The mitigating effect of the combination of SAPs and MK was more evident than when they were applied individually. Both the early-age and later-age autogenous shrinkage were significantly mitigated compared to those of the plain AASF paste. For example, the 1-day and 7-day autogenous shrinkage of AASFICMK paste was only 30% and 40% of that of AASF paste, respectively. This result indicates that SAPs and MK complement each other in mitigating the autogenous shrinkage of AASF.

#### *3.2. Cracking Potential of Paste*

It should be noted that low autogenous shrinkage does not necessarily mean low cracking potential. If the mitigating effect on the autogenous shrinkage is at the cost of dramatic loss in strength loss, the material may be subject to higher cracking risk [31]. To investigate the effect of SAPs and MK on the cracking potential of the paste, the ring test was used to measure the stress in the paste mixtures under a restrained condition. The results are shown in Figure 4. The sudden drop in the stress to around zero indicated the occurrence of cracking.

**Figure 4.** Autogenous-shrinkage-induced stress in AASF paste with SAPs and/or MK. A logarithmic scale is used on the x-axis in order to distinguish individual curves. The small fluctuation of the stress in AASFICMK paste on around 20 days and 30 days was due to the temperature fluctuation in the curing room.

Figure 4 shows that the plain AASF paste cracked on the third day after casting when the internal stress reached around 2.7 MPa. Substituting 10 wt. % slag by MK prolonged the cracking time by about 1 day and the paste broke at a stress of 3.7 MPa. The cracking potentials of AASF and AASFMK pastes were both "high" according to ASTM C1581 [22]. With internal curing by SAPs, the paste did not crack until 29 days of curing when the stress reached 6 MPa. Since the cracking time of AASFIC was close to 28 days, and the stress rate at the cracking time was 0.14 MPa/day, the cracking potential of AASFIC could be classified as "medium-low" according to ASTM C1581 [22].

The results in Figure 4 indicate that both SAPs and MK were helpful in reducing the cracking potential of the paste. Meanwhile, the addition of SAPs or MK did not lead to low strength of the matrix, as indicated by the high failure stress of the pastes. When SAPs and MK were applied together into AASF, the paste showed no cracking within 3 months of curing, which could not be realized by using only SAPs or MK. According to the low stress rate (<0.1 MPa/day), the cracking risk of AASFICMK paste was rather low [22].

Since the combined incorporation of SAPs and MK led to the lowest autogenous shrinkage and the lowest cracking potential, the mixture AASFICMK was further studied at the concrete level to develop low-shrinkage and low-cracking-potential AAMs concrete. The plain AASF concrete was studied as a reference mixture.

#### *3.3. Workability and Consistence of Fresh Concrete*

During the casting of AASFICMK concrete, a good flowability was observed. The slump of AASFICMK concrete was measured to be 280 mm (Figure 5a). The concrete quickly spread over the whole flow table (700 × 700 mm) after the cone was lifted up (Figure 5b). This slump flow value corresponded to the class SF2 for self-compacting concrete [32]. The initial and final setting times of AASFICMK measured by the Vicat method were 58 min and 117 min, respectively. The long setting time and the large slump flow indicated very good workability of AASFICMK concrete.

**Figure 5.** Slump (**a**) and flowability (**b**) of AASFICMK concrete.

#### *3.4. Strength of Concrete*

The strength development of AASFICMK concrete is shown in Figure 6 with the plain AASF concrete for comparison. It can be seen that with the incorporation of SAPs and MK, the compressive and splitting tensile strength of AASFICMK concrete was generally lower than that of AASF concrete. The reduced strength was contributed by both SAPs and MK. To provide internal curing to the concrete, extra liquid was added during mixing to be absorbed by the SAPs (see Table 4). The SAPs after absorption would act as liquid reservoirs during reaction and also as defects due to the large voids left when the liquid was released. The increased porosity of the concrete led to reduced strength [11]. Besides, the incorporation of MK was found to hinder the reaction rate in the acceleration period and could therefore reduce the strength in the very early age [17], although its impact on the 28-day strength was minor. When SAPs and MK were added together, their reducing effects were combined. Nonetheless, the 1-day compressive strength of AASFICMK concrete reached 2.1 MPa, which enabled a successful demolding at that age. The 28-day compressive strength of AASFICMK concrete reached 51 MPa, which was already sufficient for most structural uses as specified, for example, in the standard ACI 318 [33].

Besides strength values, the splitting tensile strength-to-compressive strength (ft/fc) ratio is also an important parameter that allows for the estimation of ft by knowing fc or vice versa [34]. The ratio also provides insight into the stress type (compression or tension) to which the concrete is more prone. The ft/fc ratio of AASFICMK concrete is compared with that of AASF concrete in Figure 7. On the first day, the ft/fc ratio of AASFICMK concrete was lower than that of AASF concrete which was probably because that the bonding between the aggregate and the paste in AASFICMK was still weak due to the retarding effect of MK and SAPs on the early-age reaction rates of AASF [10,11]. After the first day, however, the ft/fc ratios of AASFICMK concrete were always higher than those of AASF concrete. The higher ft/fc ratio of AASFICMK indicates that the incorporation of MK and SAPs could improve the tensile resistance of AASF concrete.

According to [31,35], a low ft/fc ratio is related to the development of microcracking in the concrete, for example in the paste surrounding aggregates, which harms the tensile strength more than the compressive strength of concrete. As shown in Figures 3 and 4, the incorporation of SAPs and MK reduced the autogenous shrinkage and the cracking potential of AASF paste. Therefore, the development of microcracking in AASFICMK concrete was supposed to be less severe than in AASF concrete. This may be the main reason why AASFICMK concrete showed a higher ft/fc ratio than the

plain AASF concrete. Whether the combination of SAPs and MK can reduce the autogenous shrinkage and the potential for cracking of concrete at the macro level will be verified in the next sections.

**Figure 6.** Compressive (**a**) and splitting strength (**b**) of AASFICMK concrete in comparison with AASF concrete. For splitting strength, the error bar is shown in the diagram, but it is too small to be clearly distinguished from the marker.

**Figure 7.** Splitting tensile strength-to-compressive strength (ft/fc) ratios of AASFICMK concrete, in comparison with AASF concrete.

#### *3.5. Autogenous Shrinkage of Concrete*

Figure 8 shows the autogenous shrinkage of the concrete. The plain AASF concrete showed large autogenous shrinkage, reaching more than 340 μm/m at the age of 28 days. In comparison, the autogenous shrinkage of AASFICMK concrete was less than 120 μm/m after a month of curing. This indicates that the utilization of SAPs and MK could effectively mitigate the autogenous shrinkage of AASF concrete. The autogenous shrinkage of AASFICMK was even lower than that of OPC concrete (see the results in [29,36]). The slight expansion of the concrete at an early age as shown in Figure 8 might be due to artifacts rather than a material behavior, since AASFICMK paste did not show expansion (see Figure 3). When stiffness of the concrete was low, the small pushing force from the LVDTs could move the embedded measuring bars a little bit, which enlarged the distance between the two measuring bars, even if the concrete itself did not expand [31]. After the first 3 days, the "expansion" was compensated by the shrinkage of the concrete.

**Figure 8.** Autogenous shrinkage of AASF and AASFICMK concrete.

#### *3.6. Cracking Potential of Concrete*

The stress evolutions in the plain AASF concrete and AASFICMK concrete are shown in Figure 9. A sudden drop in the stress to around zero indicated the failure of the concrete due to tensile stress. It can be seen that the stress generated in AASFICMK was much lower than that in AASF. In the first 4 days, a small compressive stress was detected in AASFICMK due to the slight "expansion" of the concrete (see Figure 8). Afterwards, a tensile stress started to develop. The stress in AASFICMK was only 50% and 30% of the stress in the plain AASF concrete at the age of 7 days and 14 days, respectively.

**Figure 9.** Self-induced stress in AASF and AASFICMK concrete.

According to the cracking time and stress rate, the cracking potential of AASF concrete was classified as "moderate" [22]. With the incorporation of SAPs and MK, AASFICMK concrete did not crack within 56 days. The stress rate after the first week reached below 0.01 MPa/day, indicating a very "low" cracking potential of the concrete [22].

The superior workability, the high 28-day strength, and the low cracking potential indicate that AASFICMK concrete could be considered as a highly commercially competitive construction material. Furthermore, due to the very low cracking potential of AASFICMK concrete, there is a lot of room for further tailoring the current mixture design in order to reach optimal overall performance of the concrete for different applications. For example, for the cases where the autogenous shrinkage is not very critical, lower liquid/binder ratios, lower dosages of SAPs/MK or higher amounts of slag could be used, by which higher strength of the concrete can be easily achieved.

#### **4. Conclusions**

In this paper, internal curing by SAPs and incorporation of MK were used to mitigate the autogenous shrinkage of slag-and-fly-ash-based AAMs activated by NaOH/Na2SiO3. The ring test and TSTM were used to track the shrinkage-induced stress and cracking potential of the paste and concrete, respectively.

It was found that both SAPs and MK were effective in mitigating the autogenous shrinkage and the self-induced stress of AASF paste and concrete. The dosages of 0.16 wt. % of SAPs and 5 wt. % of MK are recommended, which yielded an alkali-activated concrete (AASFICMK) with very low autogenous shrinkage and cracking potential and high enough strength. AASFICMK concrete also showed satisfactory workability. These results indicate that SAPs and MK are promising admixtures to produce high-performance AASF concrete with low shrinkage.

**Author Contributions:** Z.L.: conceptualization, methodology, investigation, writing—original draft. X.Y.: investigation. Y.C.: investigation. T.L.: methodology. G.Y.: project administration, writing—review & editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research is funded by the Netherlands Organisation for Scientific Research (NWO) under grant number of 15803 and China Scholarship Council (CSC) under grant number of 201506120072 and 201906050022.

**Acknowledgments:** Maiko van Leeuwen is also gratefully acknowledged for his help with the experiments.

**Conflicts of Interest:** The authors declare that they have no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Mechanical Properties and Freeze–Thaw Durability of Basalt Fiber Reactive Powder Concrete**

#### **Wenjun Li 1, Hanbing Liu 1, Bing Zhu 1, Xiang Lyu 1, Xin Gao <sup>2</sup> and Chunyu Liang 1,\***


Received: 20 July 2020; Accepted: 14 August 2020; Published: 16 August 2020

**Abstract:** Basalt fiber has a great advantage on the mechanical properties and durability of reactive powder concrete (RPC) because of its superior mechanical properties and chemical corrosion resistance. In this paper, basalt fiber was adopted to modified RPC, and plain reactive powder concrete (PRPC), basalt fiber reactive powder concrete (BFRPC) and steel fiber reactive powder concrete (SFRPC) were prepared. The mechanical properties and freeze–thaw durability of BFRPC with different basalt fiber contents were tested and compared with PRPC and SFRPC to investigate the effects of basalt fiber contents and fiber type on the mechanical properties and freeze–thaw durability of RPC. Besides, the mass loss rate and compressive strength loss rate of RPC under two freeze–thaw conditions (fresh-water freeze–thaw and chloride-salt freeze–thaw) were tested to evaluate the effects of freeze–thaw conditions on the freeze–thaw durability of RPC. The experiment results showed that the mechanical properties and freeze–thaw resistance of RPC increased as the basalt fiber content increase. Compared with the fresh-water freeze–thaw cycle, the damage of the chloride-salt freeze–thaw cycle on RPC was great. Based on the freeze–thaw experiment results, it was found that SFRPC was sensitive to the corrosion of chloride salts and compared with the steel fiber, the improvement of basalt fiber on the freeze–thaw resistance of RPC was great.

**Keywords:** reactive powder concrete; basalt fiber; chloride-salt corrosion; freeze–thaw durability; mechanical properties

#### **1. Introduction**

In the practical application of concrete structural engineering, the deterioration of concrete and the shortening of its service life are related to freeze–thaw damage, chloride ion penetration, sulfate attack and carbonization [1,2]. Accordingly, it is important to improve the durability of concrete structures in practical applications, which is a crucial factor to reduce the cost of infrastructure construction and maintenance. In cold areas, most concrete structures appear in the natural environment, and the freeze–thaw cycle reduces the service life and becomes one of the main sources of damage to the concrete structure [3]. Especially in cold areas, where de-icing salt has been adopted, the concrete damage caused by the coupling action of chloride corrosion and freeze–thaw cycle is serious [4]. The destruction of concrete structures caused by freeze–thaw cycle has aroused widespread concern. The durability of concrete is closely related to the internal pore structure. Concrete structure with low-porosity could effectively reduce the pore solution inside the concrete, thus reducing the internal water pressure caused by the freezing of the pore solution and improving the freeze–thaw resisrance of concrete [5]. Besides, the compact concrete structure could hinder the transmission of chloride and greatly reduce the damage caused by chloride ion penetration [6,7].

Reactive powder concrete (RPC), proposed in the 1990s, is an ultra-high-performance concrete characterized by dense microstructure and uniform concrete matrix [8]. The traditional RPC is developed through microstructure enhancement technology and the main improvement is related to matrix uniformity, porosity and microstructure, including eliminating coarse aggregates and reducing the water–binder ratio to improve matrix uniformity [9], optimizing particle gradation to decrease matrix porosity [10] and increasing the silica component by adding silica fume to improve the microstructure of the matrix structure [11]. RPC is seen as a potential material for the field of special prestressed and precast concrete components due to its low permeability, superior mechanical properties and durability [12,13]. Although the manufacturing cost of RPC is generally high, the thickness of concrete component is possible to be reduced due to its ultra-high mechanical properties, resulting in materials and costs saving. Therefore, RPC has certain economic advantages in practical applications [14]. However, RPC still has the characteristics of high brittleness of ordinary concrete and the brittleness of RPC increases as the strength increase. Therefore, the addition of fiber is usually adopted to increase the toughness of RPC [15].

Adding fiber into brittle concrete is a widely used technology. The bridging and pulling effects of fiber could improve the brittleness and impact resistance of the mixture [16,17], and the three-dimensional randomly scattered fiber could effectively inhibit the generation and propagation of cracks [18]. Previous reports have found that polypropylene fiber [19], carbon fiber [20] and steel fiber [21] can be used as reinforcing materials for RPC, and the fiber commonly used to reinforce concrete is steel fiber [22]. To improve the flexural strength, compressive strength, elastic modulus and ductility of RPC, the recommended content of steel fiber is 2% volume fraction [23]. Moreover, adding steel fiber into concrete is judged as an effective method to prevent spalling. However, steel fiber with high content will converge into a spherical shape and the steel fiber is sensitive to chemical corrosion, which results in a decrease in properties of concrete. Besides, adding steel fiber into the concrete will reduce the workability and increase the cost [24].

Basalt fiber is an environmentally friendly and high-performence fiber made of natural volcanic basalt rock, which can be used to reinforce concrete. Due to its excellent acid and alkali resistance, great mechanical properties and high temperature stability, the application prospect of basalt fiber as fiber-reinforced material is broad [25]. Basalt fiber can not only effectively increase the strength and durability of concrete but also has a positive effect on the toughness and crack resistance [26,27]. The excellent reinforcement effect of basalt fiber promotes its application in RPC. Wang et al. [28] researched the effect of basalt fiber on the mechanical properties of high-performance concrete. The results showed that the strength of high-performance concrete increased with the increase in basalt fiber content, and the compressive improved slightly, while flexural strength increased significantly. Grzeszczyk et al. [29] evaluated the effect of basalt fiber content on the mechanical properties of RPC; their results showed that the flexural strength of RPC increased with the increase in basalt fiber content. Liu et al. [30] reported that the basalt fiber could increase the toughness and bending strength of basalt fiber RPC beams, and improve the resistance to steel–concrete interface damage. Most research is about the mechanical properties of basalt fiber-reinforced reactive powder concrete. There is not enough research about the freeze–thaw durability of fiber-reinforced RPC, and the most common fiber in the freeze–thaw durability research of fiber-reinforced RPC is steel fiber [2,31]. Therefore, there are few studies were reported on the effect of basalt fiber on freeze–thaw durability of RPC. Moreover, in cold regions, where de-icing salt has been adopted, the coupling effect of chloride corrosion and freeze–thaw poses a challenge to the application of RPC and, because of the excellent corrosion resistance of basalt fiber, the improvement of basalt fiber on the freeze–thaw durability of RPC is worth exploring. Therefore, it is necessary to investigate the enhancement effect of basalt fiber on the performance of RPC.

In this paper, the mechanical properties and freeze–thaw durability of basalt fiber reactive powder concrete (BFRPC) with different basalt fiber contents (4, 8 and 12 kg/m3) were tested and compared with plain reactive powder concrete (PRPC) and steel fiber reactive powder concrete (SFRPC) to investigate the impacts of basalt fiber contents and fiber type on the mechanical properties and freeze–thaw durability of RPC. Moreover, the freeze–thaw resistance of RPC under the fresh-water freeze–thaw cycle and the chloride-salt freeze–thaw cycle was tested to investigate the effects of freeze–thaw conditions on the freeze–thaw durability of RPC. The purpose of this paper is to provide a reference for the practical application of BFRPC in cold regions.

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

#### *2.1. Materials*

Type P.O Portland cement with a strength of 42.5 MPa, manufactured by Yatai Cement Co., Ltd., Jilin, China, was used in this paper. The properties of this cement are given in Table 1. The SF93 silica fume with a specific surface area of 18,100 m2/kg obtained from Si'ao Technology Co., Ltd., Changchun, China, was used in this study. Table 2 shows the chemical composition of silica fume and cement. Three types of quartz sand, with sizes of 20–40, 40–80 and 80–120 mesh, manufactured by Zhenxing quartz sand factory, Luoyang, were used. The proportion of three types of quartz sand was 2:2:1. The quartz powder (400 mesh) was used to fill micro pores. The chemical composition of quartz sand is given in Table 3. HPWR-Q8011 polycarboxylic superplasticizer (water reduction rate 25%) obtained from Qinfen Building Materials Co., Ltd., Shanxi, China, was used to prepare all tested concrete. Chopped basalt fiber, produced by Anjie Composite Material Co., Ltd., Haining, China, was used to reinforce RPC in this study and the properties of basalt fiber are shown in Table 4. Steel fiber with a length of 13 mm and a diameter of 200 μm was adopted in this paper, obtained from Daxing Matel Fiber Co., Ltd., Ganzhou, China, which had a tensile strength of 2850 MPa. The appearance of basalt fiber and steel fiber is given in Figure 1. The tap water was used as mixed water.



**Table 2.** The chemical composition of cement and silica fume.




**Figure 1.** Fiber: (**a**) Basalt fiber (**b**) Steel fiber.

#### *2.2. Mixture Proportion and Specimen Preparation*

The best mixture proportion of RPC was obtained from the response surface method in our group's previous study [32], and based on the optimal mixture proportion; basalt fiber and steel fiber were added into RPC with an external mixing method. The content of basalt fiber was 4, 8 and 12 kg/m3, respectively. The content of steel fiber was 2% volume fraction [23]. Moreover, PRPC specimens without fiber were prepared as the control group to investigate the effects of basalt fiber and steel fiber on the mechanical properties and freeze–thaw durability of RPC. The mixture proportion of BFRPC and SFRPC is shown in Table 5.


**Table 5.** The mixture proportion of basalt fiber reactive powder concrete (BFRPC) and steel fiber reactive powder concrete (SFRPC).

All RPC mixtures were prepared by using an ISO 679 mixer with a capacity of 5 L. The degree of dispersion of basalt fiber and steel fiber in the mixture has a significant impact on the properties of RPC. During the process of mixing, the fibers were put into the mixer step by step to ensure the dispersion of fibers. The specific of mixing were as follows. (1) Put the quartz sand and fiber into the mixer and mix for 2 min. The friction between the quartz sand particles makes the fiber evenly dispersed. (2) Put the cement, quartz powder and silica fume into the mixer and mix for 3 min. (3) Dissolve the polycarboxylic superplasticizer into the water and add the solution into the mixer in two lots, mixing for 3 min each time. After stirring, the mixture was filled into a metal mold with a size of 40 × 40 × 160 mm, immediately, and compressed by vibrating on the ZT-96 vibrator table. The specimens were placed in an environment of 95% humidity and 20 ◦C for 24 h and then de-moulded. The specimens were cured for 48 h under the condition of 90 ◦C steam curing, the heating speed was 12 ◦C/h and the cooling speed was 15 ◦C/h.

#### *2.3. Testing Methods*

#### 2.3.1. Flexural Strength

According to the three-point flexural test described in Chinese national standard GB/T 17671-1999 [33] to evaluate the flexural strength of RPC. The distance between the two fulcrum points was 100 mm. The specimens were placed with the side face up in the testing machine and align the center line of the specimens with the upper fixture. The loading speed was set to 50 N/s. Three specimens with volumes of 40 × 40 × 160 mm were maded in each group for testing and the mean of the three measured results was used as the flexural strength. The specific calculation process is shown in Formula (1). The specific experimental device of flexural strength is shown in Figure 2.

$$f\_f = \frac{1.5 F\_f \text{L}}{\text{b}^3} \tag{1}$$

where *ff* refers to the flexural strength of specimens (MPa), *Ff* refers to the failure load (N), L refers to the distance between the two fulcrum and L= 100 mm, b refers to the cross-sectional width of the specimens and b = 40 mm.

**Figure 2.** The specific experimental device.

#### 2.3.2. Compressive Strength

According to the Chinese national standard GB/T 17671-1999 [33] to evaluate the compressive strength of RPC. The six fracture blocks obtained after the flexural strength test were adopted for the compressive strength test. Removed the debris on the surface of the fracture blocks and placed it in the fixture. The two sides of the fracture blocks were used as the compression surface. Adjusted the position of the specimen so that the compression surface and the fixture were in full contact. The loading rate was set to 2.4 kN/s. The compressive strength of RPC was the mean of the six measured results. The specific calculation process is shown in Formula (2). The specific experimental device of compressive strength is shown in Figure 2.

$$f\_{\mathbf{C}} = \frac{F}{\mathbf{A}}\tag{2}$$

where *fC* refers to the compressive strength of RPC (MPa), *F* refers to the failure load (N), A refers to the compression surface and A = 1600 mm2.

#### 2.3.3. Freeze–Thaw Cycle

In this paper, the freeze–thaw durability of PRPC, SFRPC and BFRPC under two freeze–thaw conditions was investigated. The specific freeze–thaw cycle test grouping is shown in Table 6. According

to the rapid freeze–thaw method described in Chinese national standard GB/T50082-2009 [34] to investigate the freeze–thaw durability of PRPC, SFRPC and BFRPC. The freezing temperature is −18 ± 2 ◦C and the freezing time is 3 h. The melting temperature is 5 ± 2 ◦C and the melting time is greater than one quarter of the entire freezing and melting cycle time. The specific arrangement of the freeze–thaw cycle test is shown in Figure 3.


**Table 6.** Freeze–thaw cycle test grouping.

Note: W refers to fresh water, N refers to 5 wt% NaCl solution.

**Figure 3.** Freeze–thaw cycle test.

ದ The cured specimens were immersed in fresh water and 5 wt% NaCl solution for 48 h. The groups of WR, WB12 and WS were immersed in fresh water and the groups of NR, NB4, NB8, NB12 and NS were immersed in 5 wt% NaCl solution. After immersion for 48 h, the the moisture on the surface of the specimens wiped off; the initial mass of the specimens was weighed and the initial compressive strength was measured. Then, the remaining specimens were put into the freeze–thaw test machine and injected fresh water and 5 wt% NaCl solution until the water level was 5 cm above the top surface of the specimens. The number of freeze–thaw cycle was set to 800, and the mass and compressive strength of the specimens were measured every 100 cycles. The mass loss rate and compressive strength loss rate were adopted to investigate the freeze–thaw durability of BFRPC and SFRPC. The specific calculation formulas are as follows:

$$MR = \frac{M\_i - M\_0}{M\_0} \times 100\% \tag{3}$$

$$CR = \frac{f\_{\text{Ci}} - f\_{\text{Ci}}}{f\_{\text{Ci}}} \times 100\% \tag{4}$$

where *MR* refers to the mass loss rate, *Mi* refers to the mass at the Nth cycles, *M*<sup>0</sup> refers to the initial mass, *CR* refers to the compressive strength loss rate, *fCi* refers to the compressive strength at the Nth cycles, *f <sup>C</sup>*<sup>0</sup> refers to the initial compressive strength.

#### **3. Results and Discussion**

#### *3.1. Mechanical Properties*

In this paper, the mechanical properties of BFRPC with different basalt fiber contents were tested and compared with PRPC and SFRPC to evaluate the effects of basalt fiber contents and fiber type on the mechanical properties of RPC.

#### 3.1.1. Compressive Strength

Figure 4 shows that the effects of basalt fiber contents and fiber type on the compressive strength of RPC. It can be known from Figure 4 that the addition of basalt fiber could improve the compressive strength of RPC and the compressive strength of RPC increases with the increase in basalt fiber content, which is consistent with the conclusion of Wang et al. [28]. When the basalt fiber content is 12 kg/m3, the compressive strength of BFRPC reaches the maximum value of 149.40 MPa, which is 6% higher than that of PRPC. Liu et al. [35] pointed out that evenly distributed basalt fiber inside the RPC could inhibit the initiation of cracks and bear part of the load, thus improving the compressive strength of RPC. Besides, basalt fiber is composed of oxides such as silica, alumina and magnesia, and its chemical composition is similar to cement. The bond strength between basalt fiber and cement paste is great [36], thus improving the compressive strength of RPC. Compared with SFRPC, the compressive strength of BFRPC with a basalt fiber content of 12 kg/m3 is 4.3% lower, which indicates that basalt fiber is not as effective as steel fiber in improving the compressive strength of RPC.

**Figure 4.** Effect of basalt fiber content and fiber type on the compressive strength of RPC.

#### 3.1.2. Flexural Strength

The experiment result of the flexural strength test is shown in Figure 5. Figure 5 shows that the flexural strength of RPC increases as the basalt fiber content increases, which is the same as the improvement of basalt fiber on the compressive strength. The increased trend of flexural strength is a slightly different from that of compressive strength. Compared with compressive strength, when the basalt fiber is from 8 to 12 kg/m3, a significant increase in flexural strength of BFRPC occurs and the flexural strength of BFRPC reaches 16.23 MPa, which is 18.5% higher than that of PRPC. Compared with compressive strength, the improvement of basalt fiber on the flexural strength of RPC is greater, aggreging with the conclusion conducted by Wang et al. [28]. The improvement of the flexural strength of BFRPC is related to the uniform distribution of basalt fiber inside the RPC. The uniform distributed basalt fiber could inhibit the generation and propagation of cracks and bear part of the stress, reducing the stress concentration near the crack and redistributing the stress [16], thus improving the flexural strength of RPC. Besides, the basalt fiber inside the RPC submits a three-dimensional random distribution; this distribution system can limit the deformation of RPC under loading conditions, which leads to an increase in the flexural strength of RPC. The three-dimensional distribution system of basalt fiber is gradually improved as the basalt fiber content increase, which is reflected in the significant increase in flexural strength of BFRPC with a basalt fiber content of 12 kg/m3. Moreover, comparing the flexural strength of SFRPC with that of BFRPC (12 kg/m<sup>3</sup> content), it can be known that the flexural strength of BFRPC and SFRPC is similar, which indicates that when the basalt fiber content is 12 kg/m3, the reinforcement effect of basalt fiber on the flexural strength of RPC is about same as steel fiber; Branston et al. [25] reported a similar conclusion.

**Figure 5.** Effect of basalt fiber content and fiber type on the flexural strength of RPC.

#### *3.2. E*ff*ect of Fiber on Freeze–Thaw Durability of RPC*

#### 3.2.1. Mass Loss

The mass loss of BFRPC with different basalt fiber contents was tested under the chloride-salt freeze–thaw cycle and compared with the PRPC and SFRPC to evaluate the effects of basalt fiber contents and fiber type on the freeze–thaw durability of RPC. Figure 6 shows that the effects of basalt fiber contents and fiber type on the mass loss of RPC. It can be known from Figure 6 that the mass loss rate of RPC increases with the increase in the number of freeze–thaw cycles. The mass loss of RPC is mainly caused by surface scaling and the edges of the specimen surface will fall off as the freeze–thaw cycle test continue, resulting in fiber exposure [2]. It is worth noting that the mass loss rate of BFRPC (8 and 12 kg/m3 content) is negative in the initial part of the freeze–thaw test, which is related to the water absorption of basalt fiber. In the initial part of the freeze–thaw test, the mass of water absorbed by the basalt fiber is greater than that of the surface scaling of RPC caused by freeze–thaw damage. When freeze–thaw cycle levels are the same, the mass loss rate of RPC decreases as the basalt fiber content increase, which indicates that the basalt fiber could improve the freeze–thaw resistance. When the number of freeze–thaw cycles is less than 600, the mass loss rate of SFRPC is in a stable increase stage and the mass loss is mainly caused by surface scaling. When it exceeds 600, a significant increase in the mass loss rate occurs, and the reason is that the steel fiber exposed to the chloride ion environment corrodes rapidly, thus leading to more peeling off the edge of the specimen surface. The mass loss rate of SFRPC reaches 3.03% after 800 freeze–thaw cycles, which is higher than that of BFRPC. This indicates that steel fiber is not as effective as basalt fiber in improving the freeze–thaw durability of RPC.

**Figure 6.** Mass loss rate versus freeze–thaw cycles.

#### 3.2.2. Compressive Strength Loss

The compressive strength of BFRPC with different basalt fiber contents was tested under the chloride-salt freeze–thaw cycle and the compressive strength loss rate was calculated by Formula (4), and compared with the PRPC and SFRPC to evaluate the impacts of basalt fiber contents and fiber type on the freeze–thaw durability of RPC.

The compressive strength and compressive strength loss rate vs. freeze–thaw cycles are shown in Figure 7. It can be seen from Figure 7a that, for all RPC specimens, the compressive strength of RPC decreases as the freeze–thaw test continues, which indicates that the freeze–thaw durability of RPC gradually deteriorates under the chloride-salt freeze–thaw cycle. During the freeze–thaw test, the frost expansion of the pore solution inside the RPC leads to the initiation of micro cracks, thus reducing the compressive strength of RPC, which is consistent with the conclusion reported by An et al. [31]. Comparing the compressive strength of SFRPC with that of BFRPC (12 kg/m<sup>3</sup> content), it can be known that the decline in the compressive strength of SFRPC is higher and the compressive strength of SFRPC and BFRPC is almost the same after 800 freeze–thaw cycles, which indicates that the corrosion of steel fiber not only increases the mass loss of RPC, but also has a negative effect on the compressive strength of RPC. It can be seen from Figure 7b that, for PRPC, BFRPC and SFRPC specimens, the compressive strength loss rate increases as the freeze–thaw test continues. At all freeze–thaw cycle levels, the compressive strength loss rate of BFRPC is lower than that of PRPC, and the higher the basalt fiber content, the lower the compressive strength loss rate of BFRPC, which indicates that adding basalt fiber could improve the freeze–thaw durability of RPC. This is because the addition of basalt fiber increases the number of harmless pores which could effectively reduce the freezing pressure of the pore solution, thus improving the freeze–thaw durability of RPC [37]. Moreover, comparing the compressive strength loss rate of SFRPC with that of BFRPC, the compressive strength loss rate of BFRPC (8 and 12 kg/m<sup>3</sup> content) is lower than that of SFRPC throughout the freeze–thaw cycle test. It indicates that steel fiber is not as effective at enhancing the freeze–thaw resistance of RPC as basalt fiber.

**Figure 7.** Compressive strength and compressive strength loss rate versus freeze–thaw cycles: (**a**) Compressive strength; (**b**) Compressive strength loss rate.

#### *3.3. E*ff*ect of Freeze–Thaw Condition on the Freeze–Thaw Durability of RPC*

#### 3.3.1. Mass Loss

The mass loss of RPC was tested under two freeze–thaw conditions and the mass loss rate was calculated by Formula (3) to evaluate the effects of freeze–thaw conditions on the freeze–thaw durability of RPC. The mass loss rate vs. freeze–thaw cycles is shown in Figure 8. This figure shows that the mass loss rate of RPC increases as the freeze–thaw cycle test continues, and the presence of basalt

fiber water absorption leads to a negative value of the mass loss rate of BFRPC in the early stage of the freeze–thaw test; as the density of the 5-wt% NaCl solution is greater than that of fresh water, when the number of freeze–thaw cycles is less than 300, the mass loss rate of BFRPC (chloride-salt freeze–thaw) is lower than that of BFRPC (fresh-water freeze–thaw). Compared with the fresh-water freeze–thaw cycle, the mass loss rate of RPC is higher under the chloride-salt freeze–thaw cycle, it can be known that chloride-salt freeze–thaw cycles causes more surface scaling, which indicates that the chloride-salt freeze–thaw cycle accelerates the deterioration of the freeze–thaw durability of RPC. After 800 freeze–thaw cycles, the mass loss rate of BFRPC under two freeze–thaw conditions is less than 1%. Besides, at all freeze–thaw cycle levels, the difference in mass loss rate of BFRPC under two freeze–thaw conditions is less than 0.25%. This indicates that the basalt fiber could significantly improve the freeze–thaw durability and chloride ion resistance of RPC.

**Figure 8.** Mass loss versus freeze–thaw cycles.

#### 3.3.2. Compressive Strength Loss

The compressive strength of RPC was tested under two freeze–thaw conditions and the compressive strength loss rate was calculated by Formula (4) to evaluate the effects of freeze–thaw conditions on the freeze–thaw durability of RPC.

The compressive strength and its loss rate vs. freeze–thaw cycles are shown in Figure 9. Figure 9a shows that, for all RPC mixtures, the compressive strength of RPC decreases with the freeze–thaw test continues. Before the freeze–thaw test, the compressive strength of RPC soaked in 5 wt% NaCl solution for 48 h is lower than that of RPC soaked in water, which shows that the immersion of 5 wt% NaCl solution has a certain negative effect on the compressive strength of RPC. When the number of freeze–thaw cycles is the same, the compressive strength of RPC under chloride-salt freeze–thaw cycle is lower than that of RPC under fresh-water freeze–thaw cycle, which indicates that compared with the fresh-water freeze–thaw cycle, chloride-salt freeze–thaw cycle causes greater damage to RPC, similar conclusion is reported by Vaitkeviˇcius et al. [38]. It can be seen from Figure 9b that the compressive strength loss rate of RPC increases as the freeze–thaw cycles increase. At all freeze–thaw cycle levels, compared with the RPC under fresh-water freeze–thaw cycle, the compressive strength loss rate of RPC under chloride-salt freeze–thaw cycle is higher. Due to freeze–thaw cycle, the pore cracked and the microcracks generated provide channels for further penetration of chloride ions. Therefore, the coupling effect of freeze–thaw cycle and chloride ion erosion has more aggressive impact on the durability of RPC [2,39]. After 800 freeze–thaw cycles, the compressive strength loss rate of BFRPC under two freeze–thaw conditions is less than 20% and the difference in compressive strength loss rate of BFRPC under two freeze–thaw conditions is less than 0.9%. It indicates that the basalt fiber could significantly improve the freeze–thaw durability and chloride ion resistance of RPC and the conclusion is consistent with the mass loss rate test.

**Figure 9.** Compressive strength and Compressive strength loss rate versus freeze–thaw cycles: (**a**) Compressive strength; (**b**) Compressive strength loss rate.

#### **4. Conclusions**

In this paper, the mechanical properties and freeze–thaw durability of BFRPC with different basalt fiber contents were tested and compared with PRPC and SFRPC to investigate the effects of basalt fiber contents and fiber type on the mechanical properties and freeze–thaw durability of RPC. Besides, the freeze–thaw durability of RPC under two freeze–thaw conditions was tested to investigate the effects of freeze–thaw conditions on the freeze–thaw durability of RPC. Based on the results of the experiment, the conclusions are as follows:


**Author Contributions:** Conceptualization, H.L. and X.G.; methodology, C.L. and X.G.; formal analysis, W.L. and X.L.; investigation, W.L., B.Z. and X.L.; writing—original draft preparation, W.L.; writing—review and editing, C.L.; funding acquisition, H.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Industrial Technology Research & Development Special Project of Jilin Province (2018C042-1); the Transportation Science and Technology Program of Jilin Province (2018-1-9); and the Science Technology Development Program of Jilin Province (20200403157SF).

**Acknowledgments:** The authors would like to express their appreciation to the anonymous reviewers for their constructive suggestions and comments on improving the quality of the paper.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Alkali Activated Paste and Concrete Based on of Biomass Bottom Ash with Phosphogypsum**

#### **Danute Vaiˇ ˙ ciukyniene˙ 1,\*, Dalia Nizeviˇciene˙ 2, Aras Kantautas 3, Vytautas Bocullo <sup>1</sup> and Andrius Kiele˙ <sup>1</sup>**


Received: 9 July 2020; Accepted: 24 July 2020; Published: 28 July 2020

**Abstract:** There is a growing interest in the development of new cementitious binders for building construction activities. In this study, biomass bottom ash (BBA) was used as aluminosilicate precursor and phosphogypsum (PG) was used as a calcium source. The mixtures of BBA and PG were activated with the sodium hydroxide solution or the mixture of sodium hydroxide solution and sodium silicate hydrate solution. Alkali activated binders were investigated using X-ray powder diffraction (XRD), X-ray fluorescence (XRF) and scanning electron microscopy (SEM) test methods. The compressive strength of hardened paste and fine-grained concrete was also evaluated. After 28 days, the highest compressive strength reached 30.0 MPa—when the BBA was substituted with 15% PG and activated with NaOH solution—which is 14 MPa more than control sample. In addition, BBA fine-grained concrete samples based on BBA with 15% PG substitute activated with NaOH/Na2SiO3 solution showed higher compressive strength compered to when NaOH activator was used −15.4 MPa and 12.9 MPa respectfully. The NaOH/Na2SiO3 activator solution resulted reduced open porosity, so potentially the fine-grained concrete resistance to freeze and thaw increased.

**Keywords:** biomass bottom ash; phosphogypsum; alkali activated fine-grained concrete

#### **1. Introduction**

In recent years, the amount of biomass bottom ash (BBA) originating from Lithuanian combustion plants is constantly increasing. This type of ash is classified as nonhazardous wastes, so BBA is deposited in local landfills. Consequently, it is very important to reuse ash and to reduce discarding at landfill site. According to Carrasco-Hurtado et al. [1] environmental study showed that the amount of heavy metals in BBA is usually lower than that in fly ash so for that reason it is possible to recycle it in construction materials.

Giergiczny et al. [2] investigated composite cement and concrete containing low-calcium and high-calcium fly ash and granulated blast furnace slag. When large quantities of ash or/and slag were incorporated in the cement system, the properties (e.g., long setting time, low early strength, etc.) of samples were improved. One utilization method for BBA could be the incorporation into construction materials. The chemical and mineral composition of BBA is appropriate for reusing in the production of new, low-carbon building materials. In this way, the replacement of traditional initial material [3,4] such as fly ash or slag in alkali activated materials (AAM) by BBA leads to important environmental benefits [5–7]. As demand for ecological alternatives to Portland cement like alkali activated materials (AAM) is growing, there is interest to utilize phosphogypsum (PG) in AAM. AAM

binders are aluminosilicate materials like fly ash, slag, red clay that can be activated with an activator solution—NaOH, Na2SiO3, KOH, etc. Concrete produced with these raw materials has shown potential results: the compressive strength of alkali activated fly ash paste reaching over 25 MPa [8,9].

PG can expand the base of the AAM raw materials. Approximately 4.5–5.5 tons of PG is generated per ton of phosphoric acid production using wet process [10]. It is estimated annually 100–280 million metric tons of PG are generated globally. The PG waste is usually stockpiled in landfills. Landfilling stocks results leaching, and hazardous constituents get into groundwater and underlying soils [11]. Pérez-López et al. investigated PG deposited over Tinto river saltmarshes for 40 years until 2010. Study have shown the high potential of contamination of the whole PG stack, including those stack zones that were restored and supposedly should have stop leaching of toxic solutions [12].

Previous studies [13–17] show interest in using PG or gypsum in AMMs. Multiple studies have investigated the optimum amount of gypsum compounds and it was determined that the optimal amount of CaSO4 in alkali activated systems is close to 10% wt [13,14]. Gypsum takes significant part in the activation processes–it completely dissolves and participates in solid product formation. In the alkali's activation reactions PG is a supplier of SO4 <sup>2</sup><sup>−</sup> and Ca2<sup>+</sup> ions enhancing the formation of secondary reaction products. When PG is present, portlandite and ettringite initially forms after dissolution form in AAM system. The hardened AAM consists mainly of amorphous hydration products, intermixed with thenardite and minor amounts of secondary gypsum. The incorporation of PG results in shorter initial setting time, but longer final setting time. There is significant increase of compressive strength when activator is NaOH [13]. The compressive strength development can be attributed to lower porosity [13,14]. PG decreases of Ca/Si ratios in the C–A–S–H gels and it could be the reason of a higher polymerized network [13]. The AAM samples with PG inclusion exhibited an average of 1.2 times greater residual strength than samples without PG, after being treated at 400, 600, 800 and 1000 ◦C temperatures [15]. Boonserm et al. [16] had found that the additive of flue gas desulfurization gypsum significantly improved the geopolymerization of the mixtures of bottom ash and fly ash. The compressive strength of samples increased too in that samples were up to 10% of gypsum was used. This increase is explained with the formation of additional amount of CSH. Similar results gave Khater et al. [17]. A 10% PG additive improved samples mechanical properties and microstructure. Samples were formed from the fly ash, PG and cement kiln dust mixtures. Rashad et al. [14] investigated the alkali activated fly ash and PG. When 5% or 10% of semi hydrate PG was incorporated in the system, the mechanical properties, improvement was detected. Chang et al. [18] investigated the influence of phosphoric acid and gypsum on the sodium silicate-based alkali-activated slag pastes. It was determined that the addition of phosphoric acid acted as a retarder.

In previous already published papers, alkali activated systems based on fly ash or fly ash and bottom ash with phosphogypsum were investigated. In this work, only biomass bottom ash was used as an aluminosilicate source. Some amount of BBA was substituted with PG. Both (BBA and PG) are local availability byproducts. The aim of this work is to investigate alkali activated paste and fine-grained concrete with BBA and PG, and to describe the effect of PG on the properties of newly formed AAM systems. Two types of alkali activators were used: NaOH solutions and the mixtures made from NaOH solution and sodium silicate hydrate (WG).

#### **2. Methodology**

#### *2.1. The Characterization of Raw Materials and the Mixing Composition of Alkali Activated Biomass Bottom Ash Pastes and Concretes*

In this study, the precursor was made from biomass bottom ash (BBA) which was obtained from the combustion plant located in Lithuania. First, this BBA was dried at 100 ◦C temperature for 24 h and then it was milled in ball mill.

*Appl. Sci.* **2020**, *10*, 5190

The PG used in this work had α hemihydrate type. It is the waste product of orthophosphoric acid production. PG is formed by the reaction of sulfuric acid from natural apatite according this Equation (1) [19]:

$$\text{Ca}\_5\text{(PO}\_4\text{)}\_3\text{F} + 5\text{H}\_2\text{SO}\_4 + 2.5\text{H}\_2\text{O} = 3\text{H}\_3\text{PO}\_4 + 5\text{CaSO}\_4 \cdot 0.5\text{H}\_2\text{O} + \text{HF};\tag{1}$$

The powder of PG was taken from the conveyor belt of waste removal and dried at 100 ± 5 ◦C temperature.

According to SEM analysis the BBA particles have irregular shape with angular morphologies (Figure 1a). Microscopic analysis showed that semihydrate phosphogypsum crystals are of dense structure and irregularly shaped parallelepipeds (Figure 1b).

**Figure 1.** Microstructure of (**a**) biomass bottom ash and (**b**) phosphogypsum.

The particle size distribution for BBA is presented in Figure 2a. The particles size is in wide range from 0.9 μm to 460 μm. The PG particles are finer (Figure 2b) and they are in the range from 0.8 μm to 38 μm. The specific surfaces areas according to Blaine for PG is 201 m2/kg and for BBA is 396 m2/kg.

**Figure 2.** Granulometric composition of biomass bottom ash (**a**) and phosphogypsum (**b**). The mean diameter of PG particles is 74.7 μm and for the particles of BBA is 58.6 μm.

The XRD analysis showed that in BBA dominated quartz, anorthoclase, gehlenite and calcium hydroxide with small amounts of calcium carbonate calcium oxide and magnesium oxide (Figure 3a). BBA has semi-amorphous semi-crystalline structure with a broad peak close to silica [20].

**Figure 3.** The mineral composition (XRD analysis) of (**a**) biomass bottom ash and (**b**) phosphogypsum. Notes: Q—quartz, SiO2 (77–1070); CH—calcium hydroxide, Ca(OH)2 (84–1271); CC—calcium carbonate, Ca(CO)3 (72–1652); A—anorthoclase, (Na,K)(Si3Al)O8 (75–1631); G—gehlenite, Ca2 Al(AlSiO7) (79–2421); CO—calcium oxide, CaO (4–777); MO—magnesium oxide, MgO (78–430); C—bassanite, CaSO4·0.5H2O (33–310); B—brushite, CaPO3(OH)·2H2O (11–293).

In the mineral composition of PG, according to XRD analysis (Figure 3b), dominated bassanite (CaSO4·0.5H2O) and a small amount of brushite (CaPO3(OH)·2H2O). PG exhibit well definite crystalline structure [21].

This byproduct of biomass combustion in power plants (BBA) has the relatively high calcium, silicon and alkali contents (Table 1).


**Table 1.** Chemical composition of initial materials, wt%.

According to the chemical composition of the PG, CaO and SO3 are the major components of this material (Table 1). There is some amount of acidic impurities such as P2O5–0.81%; including water-soluble—0.10% and F—0.14%, which make PG difficult to reuse. Loss on ignition was 6.4%. The pH of the water suspension–4.7.

Two types of alkali activators were used (Table 2). The first type was sodium hydroxide solution made with commercial NaOH pellets (analytical grades). The second one was made from the mixture of NaOH solution and the sodium silicate hydrate (WG) solution (silicate modulus 3.0, concentration 36%).


**Table 2.** The mixing composition of alkali activated biomass bottom ash pastes.

Paste samples size was 20 × 20 × 20 mm; their composition is given in Table 2. The BBA was substituted for PG at various amounts: 5%, 10%, 15%, 20% and 25%. The ratio of water and solid materials (BBA + PG) in the mixtures was regulated and ranged from 0.23 to 0.30. First, dry components were thoroughly mixed. Then, the mixtures were filled with the solutions of alkali activator (Table 2). These solutions were prepared by dissolving NaOH in water. When the complex alkali activator was used WG solution was filled to sodium hydroxide solution.

The total hydration duration was 28 days. The first day samples hydrated in room temperature, the second day at 60 ◦C temperature—and for the remaining 26 days, in room temperature again. All this time samples were covered with polyethylene covering materials which protect the samples from dehydration.

The sand from Kvesai quarry (Lithuania) was used as fine aggregate to produce alkali activated concrete samples the 0/4 fraction sand. The particle density of sand was 2.65 Mg/m3. The amount of initial materials for concrete samples is shown in Table 3.


**Table 3.** Mixing composition of alkali activated biomass bottom ash concrete, g.

The compressive strength of hardened AAM paste was evaluated after 7 and after 28 days. To perform the test a hydraulic press ToniTechnik 2020 was used. The compressive strength of samples

was determined in accordance with EN 196-1:2005. At least of three samples were tested of each type and the Sample Standard Deviation (SSD) was calculated according to Equation (2):

$$SSD = \sqrt{\frac{\sum \left(x - \overline{x}\right)^2}{\left(n - 1\right)}};\tag{2}$$

where, *x* takes on each value in the set; *x* is the average (statistical mean) of the set of values; *n* is the number of values.

#### *2.2. The Experimental Techniques*

The mineral composition of initial materials and AAM hardened pastes was carried out by using X-ray powder diffraction analysis. Data were collected by DRON-6 X-ray diffractometer with Bragg–Brentano geometry using Ni-filtered Cu Kα radiation and graphite monochromator, operating with the voltage of 30 kV and emission current of 20 mA. The step-scan covered angular range of 2–70◦ in steps of 2 = 0.02◦. The powder X-Ray diffraction patterns were identified with references available in PDF-2 data base [22].

The chemical composition of BBA and PG was evaluated by using XRF analysis. For this purpose, a Bruker X-ray S8 Tiger WD using a rhodium (Rh) tube, an anode voltage Ua up to 60 kV and an electric current I up to 130 mA were used. The compressed samples were measured in a helium atmosphere [23]. The hydration water (loss on ignition, %) in phosphogypsum was calculated after heating the material at the temperature 400 ◦C. The pH phosphogypsum was measured with the AD8000 professional multi-parameter pH-ORP-Conductivity-TDS-TEMP bench meter, with a measuring range of −2.00 to 16.00 pH, a resolution of 0.01 pH and an accuracy of ±0.01 pH. The pH measurements of water suspensions were conducted when the ratio of water (W) and solid material (S) W/S was 10.

Microstructure investigation of BBA, PG and hardened pates was performed using a high-resolution scanning electron microscope ZEISS EVO MA10 [24]. The resolution of the images (of secondary electrons in a high vacuum) of this microscope is at least 3 nm with 30 kV and at least 10 nm with 3 kV. In the performed analysis, the acceleration voltage was 5 kV.

A laser particle size analyzer (CILAS 1090 LD) was used for the evaluation of the particle size of the BBA and PG. The distribution of solid particles in the air stream was 12 wt%–15 wt%. Compressed air (2500 mbar) was used as a dispersing phase [25].

The compressive and flexural strength of hardened AAM paste and fine-grained concrete samples were determined by using hydraulic press Toni Technik 2020 according to the EN 196–1. The size of hardened AAM paste samples was 20 × 20 × 20 mm. Fine-grained concrete samples were 40 × 40 × 160 mm prisms. For each data point at least tree samples were tested. The compressive strength of hardened AAM paste was evaluated after 2 and 28 days of hydration. The mechanical properties of fine-grained concrete samples were tested after 7 and 28 days.

The total and open porosity of alkali activated fine-grained concrete samples was evaluated by water absorption according to Skripkiunas et al. [26].

#### **3. Results and Discussion**

The compressive strength of alkali-activated BBA samples is shown in Figure 4. There are two types of samples: one type of hardened AAM pastes was alkali activated by using NaOH solutions and the mixture of NaOH and sodium silicate hydrate solutions was used in the second type of samples. When SiO2/Na2O molar ratio was 2 the samples containing 20% PG substitute had the highest compressive strength. In this case compressive strength reached 24.3 MPa and precursors were alkali activated with NaOH solutions (Figure 4a). Similar values of compressive strength (23.0 MPa) were obtained for that samples which were activated with the mixtures of NaOH and sodium silicate hydrate solutions. In this case the optimal content of PG substituting was 15% (by mass of BBA). In all investigated cases (Figure 4a) the substitution of BBA to PG had gains in compressive strength. This substitution is recommended not to exceed 25%. Similar compressive strength (25.83 MPa) had geopolymer samples formed with circulating fluidized bed combustion coal bottom ash according to Topçu et al. [27].

**Figure 4.** Compressive strength of alkali-activated biomass bottom ash pastes when SiO2/Na2O molar ratio is (**a**) 2 and (**b**) 3 (Table 2).

Figure 4b shows the compressive strength of the alkali activated BAA paste with SiO2/Na2O molar ratio 3. The positive effect was detected in this case. In the PG 15-3 samples with the alkali activator of NaOH solutions the highest compressive strength reached 30.0 MPa after 28 days. By using the same molar ratio but as alkali activator the mixture of NaOH and sodium silicate hydrate solutions was the compressive strength was reached 23.0 MPa after 7 days of hardening. After longer duration (28 days) of hardening, the reduction was observed of more than 3 times of compressive strength (8.0 MPa). This reduction of compressive strength may be explained by the fast alkali reactions resulted in quick strength gain after 2 days. This gain should be due the increase of gel like matrix. After 28 days the structure samples showed cracks on the surface which could be caused by the drying shrinkage (Figure 4b) [28]. In all investigated cases the use of calcium promoter such as PG which substituted BBA had positive effect to compressive strength gain. After 28 days the compressive strength was higher than compressive strength of reference samples. Similar results related with positive effect of calcium promoters in bottom ash geopolymer fine-grained concrete report Hanjitsuwanet al [29].

The mineral composition of alkali activated biomass bottom ash is shown in Figure 5. The X-ray diffraction study is carried out only on the 8 pastes because they are the ones that have shown the highest compressive strength values. The reference compositions were investigated as well. In all X-ray diffraction patterns it is possible to detect quartz and calcium hydroxide which left unreacted from BBA. During alkali reactions calcium silicate hydrate, calcium aluminum oxide hydroxide hydrate, sodium aluminum silicate hydrate formed. When PG was incorporated in the system, additional mineral calcium aluminum hydroxide hydrate formed (PG 20-2, PGWG 15-2, PG 15-3 and PGWG 15-3). The crystal phases remained the same in all samples and it did not depend on the molar SiO2/Na2O ratios which were used in this work. By using lower SiO2/Na2O molar ratio (SiO2/Na2O = 2) the higher amount of alkali had an impact on the formation of Na2CO3(H2O) (without PG) and Na2SO4 (with PG). The formation of Na2CO3(H2O) had negative affect on the development of compressive strength [30]. When PG was inserted in the alkali activated BBA, PG reacted with NaOH and this reaction products Na2SO4 with Ca(OH)2 were (Figure 5a). During hydration process, Na2SO4 is an effective activator for alkali activated binders [31]. Sodium sulfate could motivate the formation of calcium aluminum silicate hydrate and calcium silicate hydrate. As seen in Figure 5a,b, the main peak of calcium silicate hydrate is more intensive in the samples where PG was incorporated.

**Figure 5.** X-ray diffraction patterns of alkali activated biomass bottom ash when SiO2/Na2O molar ratio is (**a**) 2 and (**b**) 3. Notes: Q—quartz, SiO2 (83–2465); Ch—calcium hydroxide, Ca(OH)2 (84–1268); K—calcium silicate hydrate, Ca1.5SiO3.5x H2O (33–306); T—thenardite Na2SO4 (74–2036); Z—sodium aluminum silicate hydrate Na96Al96Si96O384216H2O (39–222); N—sodium carbonate hydrate Na2CO3(H2O) (70–845); Hs—hydroxy–sodalite Na6(AlSiO4)6 8H2O (72–2329); H—alcium aluminum hydroxide hydrate Ca2Al(OH)7·3H2O (33–255).

In the samples with higher amount of SiO2 the molar SiO2/Na2O ratio was 3. The peaks of new formed hydrates appear more intensive (Figure 5b). This could be related with formation of higher amount of polymerization products in alkali activated system. The hydroxy–sodalite was detected in the sample PGWG 15-3 [32].

Figure 6 shows the morphology of alkali activated biomass bottom ash after 28 days of hardening. PG 15-3 and PGWG 15-3 samples exhibited different microstructures. In the microstructure PG 15-3 sample varied honeycomb-like C–S–H and honeycomb type amorphous gel structures (Figure 6a) [33].

**Figure 6.** Microstructure of alkali activated biomass bottom ash samples (**a**) PG 15-3 and (**b**) PGWG 15-3.

It can be observed that in the PGWG 15-3 sample showed a higher degree of microcracking and unreacted the particle of BBA were detected as well (Figure 6b) [34]. This PGWG 15-3 sample had a more compact microstructure by comparing with the microstructure of PG 15-3 sample. This compact microstructure is closely related to the increased amount of hydration products which increased the amount of microcracks [35].

The compressive and flexural strength of alkali activated fine-grained concretes are shown in Table 4. As the aluminosilicate precursor the mixture of BBA with PG was used. The two types of alkali activator solutions were chosen: sodium hydroxide solution and the mixture of sodium hydroxide solution and sodium silicate hydrate. The proportions of PG and BBA were chosen according to the values of paste samples compressive strength (Figure 4). According to Ding et al. [36] the compressive strength values of the alkali-activated pastes, fine-grained concretes and concretes with the same pastes were unequal. Fine-grained concretes had significantly lower compressive strength compared with the paste samples.

**Table 4.** Compressive strength, flexural strength and density of alkali activated fine-grained concretes samples.


Chindaprasirt et al. [37] investigated and compared fly ash and bottom ash fine-grained concretes. The values of compressive strength are different for alkali activated fly ash and for bottom ash. Fly ash fine-grained concrete reached 35 MPa while bottom ash fine-grained concrete had compressive strength in the range of 10–18 MPa. Such a difference is explained by the degree of polymerization. The polymerization of bottom ash is lower than the fly ash during alkali activation. All these samples were cured at 65 ◦C for 48 h. In this work, samples were cured at lower 60 ◦C temperature and duration was shorter-24 h. The fine-grained concretes samples had similar compressive strength 12.9 MPa and 15.4 MPa when activated with NaOH solution and the mixture of NaOH/Na2SiO3 solution, respectively (Table 4). The higher compressive strength could be related with the higher amount of active silicon (sodium silicate hydrate solution) [38]. The flexural strength was similar for both types of fine-grained concretes. A little bit higher value of flexural strength (2.4 MPa) were obtained for the sample with mixture of NaOH solution and Na2SiO3 solution (CPWG 15-3) compared with CPG 15-3 sample.

The porosity study is carried out on the two-alkali activated fine-grained concrete samples shown in Figure 7. It is considered to be because they are the ones that have shown the highest compressive strength values. The X-ray diffraction study is carried out only on the 8 pastes because they are the ones that have shown the highest compressive strength values. The reference compositions were investigated as well. It is possible to predict the durability (freeze–thaw resistance) of alkali activated fine-grained concrete according to these parameters of porosity. The total porosity (P) is almost the same for both types of fine-grained concretes. The open porosity (Pa) which determined by water absorption of alkali activated fine-grained concrete was less (10.9%) for CPG 15-3 samples compared with CPWG 15-3 samples 13.4%. Different situation is with close porosity (Pu). In this case CPG 15-3 samples had higher 16.8% close porosity compared with CPWG 15-3 samples which had 14.1%. Therefore, the alkali activator of NaOH and Na2SiO3 solutions had influence on the formation higher amount of close porosity and lower amount of open porosity while the total porosity remained the almost the same in activated fine-grained concrete samples. According to Nagrockiene et al. [ ˙ 39] concrete with higher closed porosity have better freeze–thaw resistance. Hence, the fine-grained concrete activated with the NaOH and Na2SiO3 solution should have higher freeze–thaw resistance compared with fine-grained concrete activated with just NaOH solution.

**Figure 7.** The porosity of alkali activated fine-grained concrete samples. P—total porosity; Pa—open porosity; Pu—closed porosity.

#### **4. Conclusions**

In this study, the compressive strength of hardened alkali activated past and fine-grained concretes was determined. To improve the reaction degree of BBA calcium promoter such as PG was used. The following observation were made:


**Author Contributions:** We all authors: D.V., D.N., A.K. (Aras Kantautas), V.B. and A.K. (Andrius Kiele) declare ˙ that we contributed to this article in equal parts about 20% everybody: D.V. and V.B. prepared the introduction part. Characterization of initial materials and the part of experimental procedures was prepared by D.N. and A.K. (Andrius Kiele). The parts of "Results and discussion" and "Conclusion" were written and evaluated by all ˙ authors. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** This research work was supported by the Lithuanian Science Council project "The utilization of industrial waste in alkali-activated concrete", project code P-MIP-17-363.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2020 by the authors. 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 (http://creativecommons.org/licenses/by/4.0/).
