The Effect of Fly Ash and Recycled Aggregate on the Strength and Carbon Emission Impact of FRCCs

Concrete is the most widely used construction material in the world. In particular, cementitious composites that contain reinforced fibers have a relatively large amount of cement. Therefore, in this study, the target is to maintain at least 90% of the performance along with a reduction of CO2 emissions from the material stage during fiber‐reinforced cementitious composites (FRCCs) production by applying fly ash (FA) and recycled sand (RS). To calculate the amount of CO2 emission at the stage of manufacturing the FRCCs, life cycle inventory database (LCI DB) were referenced from domestic and Japan. As the performance efficiency indicators, simple equations to evaluate the amount of CO2 emis‐ sion of FRCCs were then formulated as a function of the replacement ratio of FA and RS. And binder and CO2 intensi‐ ties were analyzed by FRCCs strength. Based on the results for the test, a mix of FA 25% and RS 25% of W/B 0.45, is considered most suitable in terms of performance of FRCCs and reduction of CO2 emissions.


Introduction
Fiber-reinforced cementitious composites (FRCCs) can improve tensile strength and deformation capacity by adding about 2% of reinforcing fibers to compensate for brittle fracture characteristics of ordinary concrete. Furthermore, by forming an infinite number of microcracks through the bridging action of reinforcing fibers, excellent durability can be achieved, which means that the durability of the structure itself can be increased . However, FRCCs are high-energy consumption materials affiliated with mortar that require a relatively large amount of cement (944 kg of CO 2 is emitted in the production of 1 ton of cement based on the KOREA LCI DB) compared to ordinary concrete (Yoon 2007). Therefore, cementitious composites that contain reinforced fibers have a relatively large amount of cement. In particular, it is reported that CO 2 emissions of Strain Hardening Cement-Based Composite (SHCC) are 2.61 times that of ordinary concrete (Rister and Graves 2002).
The reduction of carbon dioxide (CO 2 ) is becoming a priority in the global cement industry's energy consumption reduction and greenhouse gas reduction (Park and Park 2008;Kim et al. 2011). As such, various efforts have been made to reduce CO 2 emissions and conserve natural resources around the world recently (Gartner 2004).
Green concrete research to reduce CO 2 emissions is actively utilizing blast furnace slag powder and fly ash (FA), which are industrial byproducts produced in a large volume of industries. Additionally, research on desulfurization gypsum and red mud are actively conducted domestically and internationally (Prasad et al. 2009;Ginés et al. 2009).
In particular, while considerable developments have been made for geopolymer concretes, efforts have also been made to replace the cement-based binder in the current FRCCs with geopolymeric cement resulting in fiber-reinforced geopolymer composites, which are greener than the former. Although a great deal of important research has been performed on fiber-reinforced geopolymer composites, the field is still relatively young Shaikh (2014). It has been estimated that the manufacture of geopolymeric cement emits about 80% less CO 2 than the manufacture of ordinary portland cement (OPC) (Duxson et al. 2007;Xu and Van Deventer 2002) primarily because limestone does not need to be calcined to produce geopolymeric cement. FA based geopolymers have extremely low embodied energies. For a geopolymer based concrete (produced from FA and a soluble silicalike activator, and cured under mild heating), Tempest et al. (2009) estimated that 70% less energy is consumed when compared with OPC-based concrete of similar strength. Yang et al. (2007) showed the feasibility of creating green engineered cementitious composite (ECC) using a high volume of fly ash while maintaining tensile ductility. The fly ash was replaced up to 85% by weight of cement. The research results indicate that ECC with fly ash tends to reduce the crack width and improve the robustness of the tensile ductility, and also reduces the free drying shrinkage while slightly reducing the compressive strength after 28 days. Also, research on ECC using limestone powder as an inorganic binder has been actively conducted internationally (Hyun and Kim 2016;Tsivilis et al. 2003;Voglis et al. 2005).
Meanwhile, as part of the research for conservation of natural resources, research is underway to recycle construction waste into an alternative resource for natural aggregates (Park and Bae 2007;Lee et al. 2009). The use of recycled aggregates should be expanded in order to actively solve environmental destruction caused by the depletion of natural aggregates and the disposal of waste concrete, but they have not been widely used due to the lack of awareness on recycled aggregates and the lack of previous research on a recycled aggregate and recycled aggregate concrete. Thus, the use of recycled materials can reduce CO 2 emissions as well as provide a positive environmental impact of recycling construction waste and preserving natural resources (Cho and Chae 2015;Cho and Chae 2016). However, a previous investigation indicated that compared to the natural aggregate concrete (NAC) with same water to binder ratio (W/B), the recycled aggregate concrete (RAC) obviously degrade in compressive strength, splitting and flexural tensile, modulus of elasticity and frost resistance, while its drying shrinkage, creep, water absorption increased (Marinkovic et al. 2010).
Data that can easily evaluate CO 2 reduction due to recycled material substitution are very limited 21 . In particular, there are only a few studies regarding CO 2 reduction and using recycled aggregates (Yang et al. 2007) due to that the current research on FRCCs have mostly focused on improvement and identification of tensile performance of materials based on micromechanics (Prasad et al. 2009;Ginés et al. 2009). Marinković et al. (2010) reported that concrete recycling leads to CO 2 emission reductions only if the recycling plant is located near the building site. Authors argued that the recycling process is similar to crushed stone aggregate production and it is reasonable to assume that its energy demand is similar for both processes. However, due to the additional cement for compensating the low quality of coarse recycled aggregate, energy consumption and CO 2 emissions of RAC are highly dependent on the condition of the specific project, such as transportation, landfills and the LCI, which may be quite different from site to site. Akbarnezhad and Nadoushani (2015) The effects of various concrete recycling operations on CO 2 emissions were studied and showed that the level of CO 2 emissions can vary considerably depending on the recycling process adopted and the expected coarse recycled aggregate quality. Chowdhury et al. (2010) identified the limits for the transport requirements within which concrete recycling remains an attractive option in terms of energy and associated carbon emissions. A life cycle assessment (LCA) on both RAC and NAC was conducted by Turk et al. (2015) and found that the carbon emission of RAC was reduced to 96%; this number was, however, reported to be dependent on the transportation distance. Thus, the studies currently focus on CO 2 emissions of RAC rather than FRCCs.
Therefore, the objective of this study is to manufacture and develop FRCCs using recycled materials that maintain more than 90% performance of conventional FRCCs using Ordinary Portland Cement (OPC) for applying FRCCs using recycled materials such as FA and RS to structural members. In addition, an analysis was carried out on the CO 2 emissions and the recyclable material produced during the manufacture of FRCCs, by analyzing the compressive, flexural and tensile strength of FRCCs according to CO 2 emissions. Furthermore, this will present basic data for developing optimal FRCCs that can provide not only sustainability, economic feasibility but also environmental load reduction performance.

Material Used
The material foreground (Fig. 1) and physical, chemical and mechanical used in this study are shown in Tables 1, 2 and 3, respectively. The cement used was ordinary Portland cement with a density of 3.14 g/cm 3 produced by Korea company A as defined in ASTM C150 (2003), and the FA was F grade, a by-product of Korea B thermal power plant with a density of 2.11 g/cm 3 . The fine aggregate used was Jumunjin sea sand, manufactured in Jumunjin-eup, Gangwon-do, and specified compressive strength (f ck ) of original concrete used for manufacturing RS in this study was 18 MPa. The pieces of concrete were crushed and sieved into required aggregate sizes using a jaw-crusher by following Korea Standard sieves. Micro-steel fiber (MSF) from Japanese company K was used as reinforcing fiber, and polycarboxylic acid-based water reducing agent which improves workability and unit water quantity reduction by dispersing cement and taking in fine air was used as an admixture.

Experiment Method
In this study, to identify the mechanical characteristics of cement composites reinforced with MSF by replacement of recycled materials, the water-binder ratio (W/B = 0.35, 0.45) was determined by preliminary experiment, and the MSF was determined as 1% binder volume ratio. The replacement rates of the recycled materials were set at 25% and 50%, and the mixing conditions are shown in Table 4. A total of 240 specimens were prepared, 20 specimens for each mix, including 10 specimens to test flexural and compressive strength and 10 specimens to test direct tensile strength. Compressive and flexural strength specimens were manufactured and tested in accordance with ISO 679 (2009) by casting them in a 40 × 40 × 160 (mm) mold. For direct tensile strength tests, a dumbbell type was fabricated as shown in Fig. 2a, and a direct tensile strength tester was used to quantitatively evaluate the tensile behavior characteristics, as shown in Fig. 2b. The specimens were removed from the form 1 day after casting Page 5 of 13 Lee et al. Int J Concr Struct Mater (2020) 14:17 and water cured at 20 ± 2 °C for 28 days in a constanttemperature water bath.
3 Assessment Procedure for CO 2

System Boundary
LCA system of a typical structure consists of production and transportation of component materials, construction of the structure, life cycle and disposal, etc. (Yang and Moon 2012). CO 2 emissions from LCA should consider all of the above factors, but the contribution to LCA on recycled materials which are addressed as variables in this study is concentrated in the production part of the component materials. In addition, system boundaries were set to the raw material production stage in this study, in other words, the material stage, of each component material in order to analyze the efficiency of strength versus environmental load according to each mix. At this time, because cementitious composites are ordered and produced in units of 1 m 3 , the functional unit for CO 2 evaluation in the FRCC mixing stage was set to 1 m 3 .

LCI Data Base
The life cycle inventory data base (LCI DB) for performing CO 2 evaluation of FRCCs must include the input and output factors required for each activity involved in the production of each material, the FRCCs production stage. FRCCs List data collected from component materials, transportation, and production activities are different in each country depending on the differences in temperature, energy sources, and natural resources, so it is advisable to use data available in your own country (Jung and Yang 2015). Therefore, in this study, the national LCI DB of Korea and LCI DB (KEITI 2017) of the Ministry of Land, Infrastructure, and Transport (2015) were investigated, and the CO 2 emission factor of the raw material production phase of FRCCs mix materials is summarized in Table 5. However, as there is no LCI DB in Korea and other countries for Jumunjin sea sand and MSF, reference has been made from prior research papers (Kim et al. 2015). And for SP, data lists presented by the Japan Society of Civil Engineers used (Sakai and Kawai 2006).

CO 2 Evaluation Procedure
The total amount of CO 2 emitted ( C d ) from the system, from the outlet of each material of FRCCs functional units to the stage before FRCCs construction can be calculated by the following Eqs. (1) and (2).
In this regard, CO 2−M is the CO 2 emissions from the material stage, including cement, each admixture, aggregate, water and admixture (water reducing agent), CO 2−T is the CO 2 emissions from each material and the transportation stage of the produced FRCCs, and CO 2−P is the CO 2 emissions at the production stage, calculated from the power consumed to mix the FRCCs.
where, i is the individual materials used for the production of FRCCs, n is the number of these materials, W i and CO 2(i)−LCI are the unit mass (kg/m 3 ) and CO 2 emission factor (CO 2 -kg/kg) of the individual materials i (Cho and Chae 2015).
The application of FRCCs' CO 2 emission reduction technology can be evaluated by material, transportation, and manufacturing stages. At the material stage, reduction performance of CO 2 emissions can be evaluated by substituting recycled materials instead of cement and the transportation stage is divided into short distance procurement and energy reduction. The manufacturing stage can be divided into new and renewable energy application and manufacturing facilities (Kim and Tae 2010). In this study, because the environmental load reduction performance of FRCCs by using recycled materials is evaluated, CO 2 emissions were only calculated up to the material stage, and the transportation and manufacturing stages were excluded.
(1)  Lee et al. Int J Concr Struct Mater (2020) 14:17 Table 6 shows an example of FRCC-2 (W/B = 35%) calculated by Eq. (2), which is the equation for estimating CO 2 emissions at the material level of FRCCs. Column A is the unit mass of each material, with information related to the FRCCs mix, and column B is the amount of CO 2 provided by the LCI database of each material, so the CO 2 emissions from the production of component materials from FRCCs production were calculated by the times of columns A and B, respectively. At this time, the amount of water used for each mix was additionally calculated by considering the water absorption rate of the recycled aggregate (4.32%) as compared with that of the natural aggregate.

Binder Intensity and CO 2 Intensity
As shown in Table 6, the OPC accounts for the largest proportion of CO 2 emissions associated with the materials used to make FRCCs. Therefore, in order to reduce the CO 2 emissions of FRCCs, it is necessary to reduce the usage of OPC which is very important in terms of the environmental impact of binders. For this reason, Yang et al. (2013), Damineli et al. (2010) and Aïtcin (2000) conducted research on the efficient use of binders to reduce CO 2 emissions. In particular, Damineli et al. suggests that the environmental efficiency of concrete needs to be defined in terms of total binder consumption regarding the required strength and life cycle. On the other hand, the increase in the strength of the concrete can reduce the cross-sectional size of the structural member, so that the CO 2 reduction effect can be expected by reducing the amount of concrete. Therefore, the amount of binder and the corresponding amount of CO 2 required for expressing unit strength (1 MPa) needs to be evaluated. In short, the binder index ( B i ) measures the amount of binder (kg/m 3 ) necessary to deliver 1 MPa of mechanical strength, and consequently express the efficiency of using binder materials. The CO 2 index ( C i ) allows estimating the global warming potential of concrete formulations. It is considered that the combination of low B i and C i is an important index to improve the environmental load reduction performance of FRCCs. In this study, the environmental performance of FRCCs binders according to the replacement of FA and RS was evaluated by using the B i and C i presented by Damineli et al. (2010).
where B is the amount of binder (kg/m 3 ) for FRCCs functional unit (1 m 3 ); C is the total CO 2 emission (kg/m 3 ) for the functional unit (1 m 3 ); f is the strength (MPa) for the respective mechanical characteristics such as compressive, flexural and direct tensile strength. Table 7 shows the mean values and standard deviations of the compressive, flexural and direct tensile strengths of 28 days according to the replacement rate of recycled materials when mixed with W/B 0.35 and W/B 0.45. As the W/B decreased, the compressive, flexural and direct tensile strengths of each mix increased, and the flexural load-displacement curves did not show unusual features according to the recycled materials. The strength of FA25RS25 and FA25RS50, and FA50RS25 and FA50RS50 of W/B 0.35 and W/B 0.45 showed a slight decrease or increase as the replacement ratio of RS increased. Therefore, it is considered that the replacement ratio of RS relatively slightly affects to strength. In the mix designs of FA25RS25 and FA50RS25 and FA25RS50 and FA50RS50 at the same W/B, the flexural, compressive, and direct tensile strength were lower in the 50% of FA replacement ratio compared to 25% of FA replacement ratio. As the FA replacement ratio increased, it was determined that the strength decreased slightly due to FA which delayed early-age strength development ).

CO 2 Emission Assessment of FRCCs
In this study, the target is to maintain at least 90% of the performance of OPC100 along with a reduction in CO 2 emission from the material stage during FRCCs production by applying FA, RS, etc. CO 2 emissions are affected by the type of binder and the higher the W/B, the lower the CO 2 emission (Fig. 3). In general, the compressive strength of concrete tends to increase as the W/B decreases, and a large amount of binder is required to exhibit high compressive strength. As a result, as the W/B is decreased from 0.45 to 0.35, the amount of the binder   Page 8 of 13 Lee et al. Int J Concr Struct Mater (2020) 14:17 CO 2 emissions caused by RS increased by 0.8% in FA25RS50. For FA50RS25 and FA50RS50 replaced with 50% FA, the difference in CO 2 emissions due to RS showed that FA50RS50 increased by 1.3%. Meanwhile, for FA25RS25 and FA50RS25 with a 25% RS replacement ratio, the CO 2 emissions decreased by 31.8% in FA50RS25 as the FA replacement ratio increased, and for FA25RS50 and FA50RS50 with a 50% RS replacement ratio, the CO 2 emissions decreased by 31.4% in FA50RS50.
As a result of analyzing the CO 2 emissions of FRCCs using recycling materials, CO 2 emissions decreased as the FA replacement rate increased. However, as the replacement ratio of RS increased, CO 2 emissions for each mix of W/B 0.35 increased about 0.4% and 1.0%, respectively, and for W/B 0.45 they increased about 0.8% and 1.3%, respectively. The higher the W/B, a lower amount of binder is required, but the amount of fine aggregate increases relatively. In addition, the analysis showed that the CO 2 emissions increased as the RS replacement ratio increased. The CO 2 emission factor of the recycled aggregate in the Korea LCI DB is calculated by considering the construction waste transportation stage and the manufacturing stages which are grinding and crushing processes. In particular, because there is a screening process of below 5 mm added to the RS used in FRCCs, it is considered that the CO 2 emission factor is higher than that of Jumunjin sea sand. However, it is advisable in terms of reducing construction waste and recycling as a natural aggregate alternative resource, and it should be actively considered in the future development of resource recycling type materials. Meanwhile, regardless of W/B, Plain CO 2 emissions were reduced by 1.6% compared to OPC100. It is considered that the CO 2 emission decreased because Plain does not contain MSF.

Binder Intensity
In this study, the experimental results of the mechanical characteristics according to each mix are summarized and the relationship between compression, bending and direct tensile strength according to each experiment. At this time, the Plain without MSF did not show tensile characteristics of FRCCs and therefore was excluded from the analysis. Additionally, the relationship between B i according to compression, flexural and direct tensile strength is shown in Table 8.
For compressive strength, B i tends to decrease in the order of FA 50% > FA 25% > OPC 100% (Fig. 4). This indicates that the amount of binder to reach unit compression strength (1 MPa) is higher than that of OPC. In the case of FA25RS25 and FA25RS50 replaced with 25% FA at W/B 0.35, compared with OPC100, the compressive strength decreased to 60.13 MPa and 51.16 MPa, respectively, and the increase in bi was noticeable due to the decrease in strength according to the RS. Meanwhile, the mix of FA50RS25 and 5 replaced by 25% and 50% of RSs was not much difference from FA 50%, so the B i was similar. For W/B 0.45, it was confirmed that B i was varied according to the replacement ratio of RS at 50% replacement ratio of FA. It was also found that B i according to the replacement ratio of FA increased more than W/B 0.35.
The flexural strengths of W/B 0.35 and 0.45 were similar to those of compressive strength (Fig. 5). It is considered that this is due to the compressive strength higher than W/B 0.45 due to the relatively low W/B, rather than the influence of fiber bridging action of Page 9 of 13 Lee et al. Int J Concr Struct Mater (2020) 14:17 the mixed MSF. FA25RS25 and FA25RS50 showed a tendency to decrease in flexural strength as the RS replacement ratio increased, but the difference was insignificant.
In the case of direct tensile strength, the replacement ratio of 25% of FA was measured to have high strength, showing a lower B i compared to strength (Fig. 6). However, as the replacement ratio of RS increased, the direct tensile strength decreased, thus B i increased. Meanwhile, the mix of 50% replacement of FA, the direct tensile strength tended to increase as the replacement ratio of RS increased, but the difference was not significant and B i also showed no difference. Therefore, when considering the direct tensile strength alone, the best mix was FA50RS25 mixed with 25% FA and 25% RS.
As a result, a relatively high correlation coefficient (R 2 ) between each strength data and the B i is obtained. Hence, the relationship between the binder and strengths is helpful in designing a sustainable FRCCs.

CO 2 Intensity
The experimental results of the mechanical characteristics of each mix and the relationship between the non-dimensional value of the strength and C i for each experiment are summarized in this study. At this time, Page 10 of 13 Lee et al. Int J Concr Struct Mater (2020) 14:17 the Plain without MSF did not show tensile characteristics of FRCCs and therefore was excluded from the analysis. Additionally, the relationship between C i according to compression, flexural and direct tensile strength is shown in Table 8. The mix of FA50RS25 and FA50RS50 with 50% replacement of FA regardless of W/B showed higher CO 2 reduction rate than OPC100 by 20%, and in terms of compressive strength against CO 2 , W/B 0.45 showed better results (Fig. 7). However, the compressive strength of FA50RS25 and FA50RS50 decreased to less than 20% compared to OPC100. CO 2 emissions were slightly higher for FA25RS25 and FA25RS50 with 25% replacement of FA than FA50RS25 and FA50RS50 with 50% replacement of FA. The mix of FA25RS25 and FA25RS50 of W/B 0.45 is satisfactory within the performance maintenance range suggested in this study, and it is considered to be the best mix in terms of compressive strength compared to CO 2 emissions.
The relationship between C i and normalized flexural strength (Fig. 8). Similar to compressive strength, the combination of FA50RS25 and FA50RS50 showed a high CO 2 reduction rate, but except for the FA50RS25 mix of W/B 0.45, the target performance maintenance range in this study was not met. FA25RS25 and FA25RS50 with 25% replacement of FA were found to be a better mix to reduce CO 2 emissions and secure flexural strength.
The relationship between C i and normalized direct tensile strength (Fig. 9). Unlike compression strength and flexural strength, The FA25RS25 and FA25RS50 mixes of W/B 0.45 and the FA25RS25 containing 25% of FA and 25% RSs of W/B 0.35 were satisfactory in the performance maintenance range proposed in this study. In particular, the FA25RS25 mix of W/B 0.45 reduced CO 2 emissions by about 30% compared to OPC100, while the direct tensile strength improved by more than 5%. Therefore, satisfied with the performance maintenance range proposed in this study, and it is considered the best mix in terms of direct tensile strength compared to CO 2 emission.
As a result, a relatively high correlation coefficient (R 2 ) between the relationship between the non-dimensional value of the strength and C i for each experiment is obtained. Hence, the relationship between the CO 2 emission and strengths is helpful in designing a sustainable FRCCs.
Page 12 of 13 Lee et al. Int J Concr Struct Mater (2020) 14:17 between B i and C i and mechanical characteristics of FRCCs was compared and analyzed. From the experimental and analytical results, conclusions are summarized as follows.
(1) As a result of evaluating the mechanical characteristics of MSF FRCCs using recycled materials, comparing to the objective of this study which was OPC 100, the mixtures that resulted more than the 90% of the performance of mechanical characteristics, were FA25RS25 and FA25RS50 of W/B 0.45. Additionally, in comparison with OPC 100, the compressive strength, the flexural strength, and the direct tensile strength of FA25RS25 and FA25RS50 were 95.1% and 94.8%, 105.2% and 110.0%, and 105.3% and 90.2% respectively. (2) The lower the W/B, the higher the CO 2 emissions and the difference in CO 2 emissions between W/B 0.35 and W/B 0.45 was 15.3% on average. As the replacement rate of FA increased, the average CO 2 emission decreased by 24.1%, and as the replacement rate of RS increased, the average CO 2 emission increased by 0.7%. Therefore, as a result of the CO 2 emissions at the material stage according to the FRCC mix tested in this study, a greater reduction in CO 2 emission resulted from the replacement rate of FA than W/B, and the CO 2 emission by RS increased, but the difference was not significant. (3) B i was smaller on average as W/B was lower, and the mix of FA replacement was higher than OPC.
As the replacement ratio of FA increased, B i also increased, indicating that the amount of binder to produce unit strength (1 MPa) was higher for FA than that of OPC. As a result of analyzing the relationship between compression, flexural, tensile strength, and C i , it was found that C i decreased as the replacement rate of the recycled material increased. The replacement ratio of FA and RS for maintaining the performance of mechanical characteristics of 90% or more compared with the case of no replacement were, the mix of FA 25% and the mix of RS 25%, of W/B 0.45 for compressive strength, the mix of FA 25% and 50% and the mix of RS 25%, of W/B 0.45 for flexural strength. In the case of direct tensile strength, the mix of FA 25% and 50% and the mix of RS 25%, of W/B 0.45 and the mix of FA 25% and the mix of circulating fine aggregate 25%, of W/B 0.35. (4) A relatively high correlation coefficient (R 2 ) between the strengths data and the B i and C i is obtained. Hence, the relationship between the binder and CO 2 intensities is helpful in designing a sustainable FRCCs. Furthermore, the CO 2 emission of a FRCCs can be predicted simply based on information about the type and amount of binder. (5) In particular, the mix of FA25RS25 replaced with FA 25% and RS 25% of W/B 0.45, is considered most suitable in terms of performance of FRCCs and reduction of CO 2 emissions.