Assessment of the impact of adding fly ash and coal bottom ash to an advanced cement-based composite matrix

: To reduce environmental impact and hydration heat from high cement consumption in advanced cement-based composites, this study proposes the partial cement substitution by fly ash (FA) and coal bottom ash (CBA) in rates of 15% by mass (7.5% FA and 7.5% CBA) and 30% by mass (15% FA and 15% CBA). Matrixes were assessed for their physical and mechanical characteristics though the tests of compressive strength, capillary water absorption and ultrasonic pulse velocity, added to the analysis of images generated by scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS). Results indicated better performance of fly ash due to its particle size, shape and texture, especially for the rate of 15%. In terms of compressive strength, mixture RFA15 yielded the highest strength. The capillary water absorption results were linear and with low variations of about 1% and 5% between 28 and 56 days.

The binary mixture, that is, one with two pozzolans in the composite, causes hybrid and synergistic effect between hydration and pozzolanic reactions, along with physical effects of obstruction and refinement of pores due to the presence of smaller particles.
Coal ash is waste derived from inorganic and organic matter during the incineration and combustion of coal. The main composition of the ashes from power plants consists of crystalline, vitreous and organic matter [44]. Osterreicher-Cunha et al. [45] stress that fly ash, in physical terms, is normally characterized by its silty particle size, with particle specific gravity varying between 2050 kg/m 3 and 2200 kg/m 3 .
Fly ash has been broadly used as a supplementary cementation material in technological concretes as it contributes to the workability of fresh concrete and the formation of the microstructure, as well as mechanical properties and longterm durability. The proportion of fly ash in binding agents is usually adapted to different engineering applications and curing conditions [46].
According to Lenormand et al. [47], the durability and environmental efficiency of pozzolanic materials derived from coal incineration are one of the most decisive points regarding their use as construction materials. Both fly ash and bottom ash can reduce hydration heat and autogenous shrinkage of high-strength concrete, in addition to providing more workability and viscosity to the fresh mixture. The size of fly ash particles is usually within the same range as that of cement [48].
When pozzolanic materials are used, the most important effects on the microstructure of the cement paste are the changes in pore structure, which comprises the reduction of particle size caused by pozzolanic reactions, and the obstruction of pores and voids by the action of finer particles, hence improving the density of the composite. The pozzolanic effect is maximized in binary mixtures [49], i.e., a mixture made with two materials, as is the case in this study, cement and fly ash.
Using a binary mixture may also allow for a benefic effect of pore refinement, generating greater resistance to chlorate mitigation and to concrete corrosion, thus increasing overall durability. Such pore refinement leads directly to a reduction in capillary water absorption, which takes place because of size, distribution, shape, tortuosity and continuity of their pores [50].
It is believed that composites with ternary mixtures can be optimized so that the synergistic effect lets the components compensate for their mutual defects, meaning that individual properties are improved. Results showed that using ternary mixtures improves compressive strength and flexural strength overall. Durability is also improved, along with the resistance to sulfate, acid and chloride ion penetration [51].
Studies have revealed that synergetic action in a ternary mixture greatly improves fresh concrete workability and produces a more uniform mixture, a denser microstructure and grants better performance to hardened concrete [52].
Thus, this research aims to present the analyze of the impact of fly ash and coal bottom ash addition on the properties of ultra-high-performance concrete, considering mechanical and physical properties.

Materials
The materials used to craft the advanced cement-based composite were: Brazilian type CPV-ARI cement, quartz dust, sand and fly ash and coal bottom ash as a partial substitute for cement. There materials presented the particle-size distribution of Figure 1, along with the characteristic diameters of Table 1. In addition to laser particle size to characterize FA and CBA, the technologies of scanning electron microscopy (SEM) and energydispersive X-ray spectroscopy (EDS) were used as well. The sand used has specific weight of 2623,4 kg/m 3 and its particle size distribution is presented in Figure 2.   Regarding the materials particle size distribution is important to notice that Li et al. [53] uses ashes with D50 between 15,94 and 22,05 and in this study, the range is between 5,19 and 49,17, due to fly ash, as expected. Is also important to notice that the difference between coal bottom ash and cement are higher than with the use of fly ash, which has similar values to the cement in all characteristic diameters.

Methods
The mix ratios used in this experimental program are presented in Table 2. After determining the reference ratio, the partial replacement by mass of cement by fly ash (FA) and coal bottom ash (CBA) was carried out. The materials were mixed in a dual speed mortar mixer. The materials were added in the same order as in Christ [3]. After mixing, the samples were cured under standard conditions according to NBR 5738:2015 [54] until the testing age. The relation water/cement was 0,3 for all mixtures. The materials order in mixture was cement and FA or CBA, sand, superplasticizer and quartz dust. The samples were cured in humidity chamber until the age of the tests.

Characterization Parameters
The tests performed during this experimental program are described as follows, along with their respective standards.
a) Characterization of Ashes -Scanning Electron Microscopy (SEM) and Energy-Dispersive X-Ray Spectroscopy (EDS) To perform the SEM analysis, the sample was previously dried in an oven for approximately 24 hours at a temperature of 45°C. Magnifications of 1000 and 5000 times were used for the analysis. The EDS test, which aims to characterize materials in indicative terms of chemical composition, was performed on each ash. The determination of the percentages of each constituting element of the ashes relied on the mean of the results obtained from eight different observation points.
b) Ultrasonic Pulse Velocity The ultrasonic pulse velocity test abided by NBR 8802:2013 [55], searching for a quality indicator for the different concretes (density, void presence). For this test, specimens of 50 mm x 100 mm were analyzed at 7, 28 and 56 days of curing, three specimens per mix ratio developed, which were the same ones used for the compressive strength test. During the test, a thin layer of solid petroleum jelly was added to the surface of the transmitter and receiver device to ensure the perfect coupling between the magnetic wave devices and the specimens. The direct method was used with devices on opposite longitudinal faces.
c) Axial Compressive Strength The compressive strength test was performed as per NBR 5739:2007 [56]. Cylindrical specimens with dimensions of 500 mm x 100 mm were molded and then fractured at 7, 28 and 56 days, considering three specimens per mixing ratio developed, at each age. d

) Capillary Water Absorption
The test for determination of capillary water absorption was performed according to RILEM TC 116-PCD [57]. In this test, specimens with diameter of 50 mm and thickness of 50 mm were used, which were the lower half of cylindrical specimens with dimensions of 50 mm x 100 mm. The analyses took place at 28 and 56 days of curing, considering three specimens per mix ratio developed. The samples had to be dried in an oven at 100°C until constant mass was obtained, which was measured by weighing twice a day. Next, the specimens were wrapped with a balloon on the top face and with tape on the sides, promoting the impermeability of concrete on the external faces, and a 3-mm water layer was maintained on the bottom face. The weighing was done at 10 min, 1h, 4h and 24h.

Scanning Electron Microscopy (SEM)
The SEM images are presented in Figures 3 and 4, considering magnifications of 1000 times (a) and 5000 times (b) respectively. In Figure 3, most of the fly ash grains were uniformly distributed, spherical and with regular shape. Furthermore, in Figure 2(a), the dimension of some spherical grains varied between 0.67 and 44.56 µm, with average diameter of 7.61 µm between measured particles, which was close the characteristic average diameter of 6.56 µm obtained in the laser particle size test. Kutchko and Kim [58] found average particle size varying between 16 and 64 μm, bearing in mind that bigger particles usually correspond to higher unburned carbon concentrations. Rafieizonooz et al. [59] reported the spherical and regular shape of fly ash, which presented smaller particles than the coal bottom ash, also assessed by these authors. Coal bottom ash had an irregular shape with apparent grain porosity. Figure 4 depicts the irregular 59shape with porous and rough texture of coal bottom ash. In Figure 4(a), the dimension of some grains varied from 3.67 to 89.59 µm, with average diameter of 42.16 between measured particles, which is somewhat different from the characteristic average diameter of 56.62 µm obtained in the laser particle size test. Such difference is understandable, because in SEM only a small portion of grains gets identified, while the particlesize test uses broader sampling. Thus, the values obtained comply with the particle-size distribution.
Concerning Rafieizonooz et al. [59], the ashes in this study turned out to be physically compatible with their result, seeing that they reported porosity and irregular shape of coal bottom ash grains. As per Mangulkar and Jamkar [60], it is very important to not only observe the size of the particles but also the predominance of their shape. From the ashes assessment, it was noted that FA had less laminar particles, what can favor particle packing.

Energy-Dispersive X-Ray Spectroscopy (EDS)
Analyzing the EDS test results (Table 3 and Table 4), the elements in greater proportion in the composition of fly ash were, on average, silicon (53.75%), aluminum (25.85%), calcium (18.21%) and iron (4.67%), along with lower amounts of potassium and titanium.  Still on the topic of chemical composition, the fly ash used in this study had a total percentage of SiO 2 , Al 2 O 3 and Fe 2 O 3 equals to 84.27%. As described in ASTM C618-15 [61], the percentage sum of these elements falls into category C or F if above 50% and, above 70%, into class N, which require other variables for classification, such as loss on ignition (LOI), as it represents the mass of moisture and volatile materials in a sample. The chemical analysis performed by Rafieizonooz et al. [59] indicated that the fly ash used was composed mostly of silica, iron and alumina, and the percentage sum of SiO 2 , Al 2 O 3 and Fe 2 O 3 was about 78.82%, therefore making it a class F ash according to ASTM C618-15 [61].
Regarding the EDS test results for coal bottom ash, the elements that make up most of this waste were, on average, silicon (56.38%), aluminum (31.12%) and iron (9.13%), along with small amounts of potassium, titanium and calcium. The percentage sum of SiO 2 , Al 2 O 3 and Fe 2 O 3 of the ash used in this study was equals to 96.63%. Nonetheless, the limits of ASTM C618-15 [61] apply to fly ash and natural pozzolans only.
The chemical analysis of coal bottom ash performed by Rafieizonooz et al. [59] revealed that it was composed mostly of silica, iron and alumina, with small amounts of sulfate, magnesium and calcium. The percentage sum of SiO 2 , Al 2 O 3 and Fe 2 O 3 was equals to 83.24% in their study. It should be noted that the chemical characterization values obtained are based on a semi-quantitative technique, which explains the great variation from the energy-dispersive Xray test and the reduced reliability [62].

Compressive Strength
The compressive strength tests were carried out at 7, 28 and 58 days of specimen curing, and the average test results were summarized in Table 5. Three samples were tested for each trace, and the result presented is the medium. Figure 5 allows a comparison between the average strength of each mixture at each testing age.

Figure 5 -Compressive strength test results
From these tests, it was concluded that the binary mixture with 15% fly ash (RFA15) yielded higher 28-day strength than the binary mixture with 30% fly ash (RFA30) and the other ternary mixtures, and its strength was just 6.8% lower than that of the reference mixture (RR). Mixture RFA15 presented the highest 56-day strength, which was approximately 28% higher than that of RR.
Among the ternary mixtures, the one with lower pozzolanic replacement content yielded the best compressive strength, what may be caused by the particle size distribution and format. Moreover, mixtures RR, RFA7.5BA7.5 and RFA15BA15 did not achieve the linearity that was expected of them, seeing that RR and RFA7.5BA7.5 demonstrated lower values at 28 days and RFA15BA15 at 56 days, what can be noted by curves and correlation coefficients.
Mixture RR achieved the best compressive performance at 7 and 28 days, but due to the pozzolanic activity that occurs at more advanced ages, mixture RFA15 had a very similar performance at 28 days and superior at 56 days of curing. Mixture RFA30 presented an increase of approximately 16% compared to RR at 56 days.
Some results still produced great variability. It was related to the susceptibility of reactive powder concrete to the procedures of mixing, molding and curing, during which any variation can generate significant changes in the results.
Lawrence et al. [63] stated that finer dusts tend to yield better cement hydration, although some dusts reduce the hydration rate, what can lead to a delay in configuration and then affect the development of short-term compressive strength, an effect that is especially linked to fly ashes and silica fume. According to Mehta and Monteiro [27] apud Duarte et al. [64], the reduction of compressive strength owing to the increased rate of substitution of cement by ashes is a result of the generation of low-density secondary hydration products, mostly calcium silicates formed from the reaction of calcium hydroxide with ash particles. Calcium silicates Hydrated calcium silicates are also cement hydration products, although they have high density.
Such delay can be observed in the results since the compressive strength reached values that are lower than the usual for reactive powder concretes, indicating that the ashes used changed the setting time of the composite [6].
Kreuz [65] reported that compressive strength decreases as the rate of substitution of cement by bottom ash increases, though there were cases with 10% of substitution in which the strength remained equals to or even greater than the reference mixture.
The reduction in the value of strength of the reference concrete to 56 days, and the 28 days of the RFA15BA15 was also observed. This variation may result from the molding and sample-making processes, or from the greater heterogeneity that is expected when concretes with a higher strength class are developed [66], [67].
Still related to the results and with the characterization of pozzolans, it is worth pointing out that, according to Lee et al. [68] smaller particles tend to have greater efficiency in the development of resistance resulting from the generation of secondary products.

Ultrasonic Pulse Velocity
According to IS 13311-92 [69], which regards the classification of the quality of concrete with respect to the ultrasonic pulse velocity, all mixing ratios analyzed at 7, 28 and 56 days fit the maximum quality class, graded as excellent ( Figure 6). The reference mixture RR yielded the highest ultrasonic pulse velocity at all testing ages, despite not presenting a linear behavior. The assessment at 7 days of curing revealed no linearity between the values of the ternary mixtures RFA7.BA7.5 and RFA15BA15, but the binary mixtures RFA15 and RFA30 displayed a difference of less than 1% in the results. Mixtures RFA15 and RFA30 remained close and only mixture RFA15BA15 was different from the others, and mixture RFA7.5BA7.5 presented the highest ultrasonic wave velocity after RR. In the next assessment, at 28 days of curing, both binary and ternary mixtures obtained linear results compared to each other. However, the value of mixture RFA15BA15 was 3.5% higher than that of 7 days. At 56 days of curing, the results were like those from the 7 days of curing, considering the proportioning between the mixtures, and in both ages the lowest result was obtained in mixtures with 30% of pozzolanic materials in agreement with other studies [66], [70] Therefore, all mixtures, except for the binary ones, reached higher ultrasonic pulse velocities at 28 days of curing, compared to 7 and 56 days of curing, and the result at 56 days was higher than that at 7 days, except for mixture RR.
The decrease of permeable pore space in ash concrete mixtures results in higher ultrasonic pulse velocities. The values from this test indicate the quality of the mixture in terms of density, homogeneity and uniformity [41] [58]. Yujing et al. [71] pointed that UPHC with pozzolanic materials promotes a pore medium size reduction, and thus, the material is denser than conventional concretes and UPHC with lower pozzolanic materials content.
Finally, the ultrasonic pulse test results were not linear, considering that only RFA15 presented progressive results, and RFA30 remained the same at 28 and 56 days.
Analyzing the results obtained for compressive strength and propagation velocity of ultrasonic waves comparatively, it is worth noting that the expectation brought by Godinho et al. [72] was not verified, since an equivalent relationship between the variables was not perceived. While in the compressive strength the sample with the best performance at 56 days was the RFA15, in the ultrasonic speed there was a superiority of the reference sample.

Capillary Water Absorption
The capillary water absorption test results show that all mixtures presented slight variations between 28 and 56 days, except for mixture RFA30, which at 56 days presented an increase of 32% in absorption at the end of 24 hours, as observed in Figures 7 and 8.
Moreover, the resulting values had great variation, most likely due to the presence of voids in the materials. Capillary pores are reduced by the formation of secondary C-S-H gel from the pozzolanic reaction, therefore reducing concrete capillary water absorption [50].
The results obtained in 28 and 56 days are very similar. Probably, due to the type of cement applied, which has hydronation and strength development occurring in first ages. However, it may be pointed out that the mixtures with the use of pozzolanic materials has reached almost the same values of reference mixture at the age of 56 days, showing its contribution for the mixture [73].

CONCLUSIONS
From the ashes assessment, it was noted that fly ash presented better results in the tests of compressive strength and capillary water absorption, which can be linked to particle size, shape, texture and pozzolanic activity. Chemically, both coal ashes were similar as the fly ash contained silicon, aluminum, calcium and iron, in addition to small amounts of potassium, titanium and sodium, whereas the bottom ash contained silicon, aluminum and iron along with small amounts of potassium, titanium and calcium.
Regarding the test mixtures, the binary mixture with 15% fly ash performed better, considering that its results maintained linearity and progression in the tests of compressive strength and ultrasonic pulse velocity. As for the capillary water absorption test, the results were identical between the samples analyzed. The binary mixture with 30% fly ash was also efficient as it performed better than the reference concrete. There seems to be a need for in-depth analysis of the interaction between ashes, especially due to the nonuniform particle size of coal bottom ash.
Considering the results obtained through this study, the authors recommend that future studies may evaluate different proportions between coal and bottom ash, considering the satisfactory behavior obtained by the tests of ultrasonic velocity and compressive strength.