Sustainable Cements Containing Sugarcane Bagasse Ash and Limestone: Effects on Compressive Strength and Acid Attack of Mortar

Abstract

Sustainable cements are an important alternative to reduce the environmental impact of the cement industry by lowering the clinker-to-cement ratio with supplementary cementitious materials. In this respect, the present study aimed to evaluate the influence of partial clinker replacement by sugarcane bagasse ash (SCBA) and limestone filler (LF) on the mechanical and durability performance of mortars. Four blended Portland cements were produced with binary and ternary mixes of clinker, SCBA, and LF. An ordinary cement was also produced for comparison purposes. All five cements were characterized and applied in mortars in order to assess compressive strength and water absorption. Next, 28-day specimens were immersed in a sulfuric acid solution for 56 days to investigate deterioration using mass loss, length variation, water absorption, and compressive strength tests. In general, the combination of SCBA and LF produced more sustainable cements with suitable properties, as SCBA improved the mechanical behavior, while LF improved the durability performance of mortars. In this context, ternary mixes with 14% SCBA and 14% LF are indicated for mechanical uses, while 7% SCBA and 14% LF are recommended for durability purposes, as both maintained the respective properties of the reference cement.

1. Introduction

Cement is the main building material used worldwide. The scale of cement production is related to the development and industrialization levels of countries [1]. However, cement manufacturing is responsible for large energy and raw material consumption, in addition to the high carbon dioxide emissions caused by clinker production [1,2,3,4]. It is estimated that the cement industry is responsible for around 8% of all anthropogenic carbon dioxide emissions to the atmosphere [5]. Furthermore, cement production is expected to increase until at least 2050 [3]. As such, sustainable development of the cement industry has been a topic of growing interest among researchers. In fact, various wastes have been used in different cement–based building materials [6,7,8,9,10,11,12,13,14,15]. Another important alternative used worldwide consists of incorporating supplementary cementitious materials (SCM) as cement (or clinker) replacements [16,17,18]. The use of SCM can reduce CO₂ emissions by up to 40% without significantly changing the mechanical strength, durability, and cost of Portland cement [17].

Traditional SCMs such as fly ash, blast furnace slag, metakaolin, and natural pozzolans have been used to reduce the clinker-to-cement ratio [3,19]. Recently, studies have looked to SCMs from agro-industrial wastes with significant pozzolanic potential. In this case, the effective use of these materials in the cement industry could result in a sustainable method to discard large amounts of waste and minimize the environmental damage caused by cement production [16]. In this scenario, sugarcane bagasse ash is a promising and abundant SCM, especially in tropical countries. The main constituent of bagasse ash is silica, which makes it extensively studied [20,21]. Numerous studies have shown the positive effects of bagasse ash on different properties of cementitious materials [22,23,24,25,26,27,28,29]. The use of processing methods such as burning, grinding, and densimetric fractionation increases the specific surface area, amorphous content, and pozzolanic activity of ash due to the reduction of contaminants [30,31,32,33,34,35].

Several studies have recently investigated the incorporation of limestone and a pozzolanic material into cementitious compounds in order to combine the physical and chemical effects of these different materials [36,37,38,39]. Limestone provides an extra surface for hydrate precipitation and nucleation [40,41,42,43], but high levels of this material can result in loss of mechanical strength due to the dilution effect, since it behaves like aquasi-inert material [38,39,44]. Thus, the combination of limestone and pozzolan is an interesting alternative for the cement industry, given that the latter can enhance the mechanical performance at later ages [27].

In this respect, the present study aimed at evaluating the influence of partial clinker replacement by sugarcane bagasse ash and limestone on the mechanical and durability performance of mortars. To that end, five different cements were produced, characterized, and applied in mortars with low water-to-cement ratios. The isolated and combined effects of SCBA and limestone filler were evaluated and compared in terms of compressive strength, water absorption, and sulfuric acid attack durability tests.

2. Materials and Methods

For cement production, the clinker and gypsum were provided by a cement factory in Espírito Santo State, Brazil. The oxide composition of these materials was obtained by X-ray fluorescence spectrometry using an EDX-720 spectrometer (Shimadzu, Kyoto, Japan), as shown in

Table 1. Chemical composition (% mass) and physical properties of raw materials.

	Clinker	Gypsum	SCBA	LF
CaO	71.5	52.3	3.4	52.8
SiO2	14.2	4.5	53.8	3.7
Al2O3	5.2	-	18.1	-
Fe2O3	5.7	0.1	10.8	1.8
K2O	0.6	-	4.3	0.5
SO3	1.7	40.6	2.7	0.6
TiO2	0.3	-	1.0	-
P2O5	-	-	1.6	-
MnO	0.1	-	0.2	-
LOI *	0.8	2.6	3.9	40.5
Physical properties
D50 (μm)	-	-	4.18	4.17
Density (g/cm³)	-	-	2.61	2.77
BET specific surface area (m²/kg)	-	-	17,000	5000
Pozzolanic activity index with cement (%)	-	-	127	-

* LOI: Loss on Ignition.

.

Sugarcane bagasse ash (SCBA) was collected from the ash collector in a Brazilian sugarcane plant in Rio de Janeiro State, Brazil. The ash was obtained from the uncontrolled burning of sugarcane bagasse (Figure 1a) in boilers at temperatures around 800 °C. The as-received SCBA was treated with densimetric fractionation [30] (Figure 1b) and autogenous recalcination [45] processes that reduced the contaminant content, resulting in ash with good pozzolanic activity, a high specific surface area, and adequate chemical and mineralogical compositions, as shown in Table 1 and Figure 2, respectively. Figure 1c shows the SCBA after grinding.

Limestone filler (LF), composed of around 94% CaCO₃, was obtained from a mining company in Rio de Janeiro State, Brazil. Its chemical composition and physical properties are shown in Table 1. Both SCBA and LF were ground in a ball mill to obtain similar granulometry (see D₅₀ values) in order to minimize different physical effects. Although the particle distribution of these materials was similar, the specific surface area of SCBA was larger than that of LF. This was justified by the porous structure of SCBA, which is responsible for most of the porous volume of the ash, as described in previous studies [30,35,46,47].

The Portland cement samples were produced at a laboratory scale with the joint grinding of clinker and gypsum and mixes of SCBA and LF in specific proportions established for each cement, as presented in

Table 2. Cement mix proportions (% mass).

Cement	Clinker	SCBA	Limestone	Gypsum
PC-1	95	-	-	5
PC-2	81	14	-	5
PC-3	81	-	14	5
PC-4	74	7	14	5
PC-5	67	14	14	5

. Five types of cement were produced: a reference cement (PC-1, ordinary cement), two blended cements with binary mixes of clinker-SCBA and clinker-LF (14% of clinker replacement), and two other blended cements with ternary mixes of clinker, SCBA, and LF (21 and 28% of total clinker replacement). The grinding parameters and cement production procedures are described elsewhere [45]. The gypsum content was fixed at 5% to maintain the same cement formulation and avoid the undersulfation phenomenon, which could result in sulfate depletion before C₃S hydration [36,40].

After production, all five types of cement were characterized and applied in mortars. Figure 3 summarizes the methodological procedures. First, the particle size distribution was determined by laser diffraction (Mastersizer 2000, Malvern Instruments, Malvern, UK) with dispersion in absolute ethanol for 15 min. Next, the chemical composition, in terms of oxides, was obtained by semi-quantitative analysis using an X-ray fluorescence spectrometer (EDX-720, Shimadzu, Kyoto, Japan) and loss on ignition according to Brazilian standard NBR NM 18 [48]. The initial and final setting times were determined using the Vicat method [49]. Density and Blaine fineness were determined according to ASTM C188-17 [50] and ASTM C204-18E1 [51], respectively. Finally, the pozzolanic behavior of the three cements containing SCBA (PC-2, PC-4, and PC-5) was evaluated by the Frattini test [52] to determine whether the SCBA content was effective in producing pozzolanic cements.

Next, mortars were produced to assess the influence of cement type on compressive strength and durability to acid attack. In addition to the cements produced, deionized water and normalized sand [53] were used to produce the mortar specimens, with constant water-to-cement and sand-to-cement ratios of 0.48 and 3.0, respectively. Polycarboxylate-based superplasticizer (Basf Glenium 51, with 28.9% solid content by mass) content was fixed at 0.012% of cement mass. The mix was prepared in a standard mortar mixer. First, water and superplasticizer were homogenized, and the cement was then added and manually mixed with a spatula for 30 s. This was followed by mechanical mixing at low velocity for 30 s. The sand was added and mixed for 30 s at low velocity and then at high velocity for an additional 60 s. The mixture was allowed to rest for 90 s, and the last mix was carried out for 60 s at high velocity. The mortars were designated, according to cement type, as M-PC-1, M-PC-2, M-PC-3, M-PC-4, and M-PC-5.

The compressive strength tests were performed at 7, 28, and 84 days of curing, in line with ASTM C109M-16 [54]. For each age, three cubic specimens (50 mm edge) were molded for each mortar, demolded after 24 h, and kept in a limewater solution until the testing ages. The ruptures occurred in a UH-F500kNI hydraulic universal testing machine (Shimadzu, Kyoto, Japan) operating at 0.5 mm/min. Water absorption tests [55] were performed with four cylindrical specimens (25 mm diameter and 50 mm high) of each mortar at 28 and 84 days of curing. The molding and curing processes were the same as those described for cubic specimens.

Finally, a durability test was performed with three cubic (50 mm edge), four cylindrical (25 mm diameter and 50 mm high), and three prismatic specimens (25 × 25 × 285 mm) of each mortar. After 28 days of limewater curing, the specimens were immersed in a sulfuric acid solution for 56 days. The procedures developed by Khan et al. [56] were adapted, and the mortars were immersed in a solution with 1.5% H₂SO₄. The specimen-to-solution volume ratio was kept at 0.25. The pH was measured weekly, and the solution was completely changed when it reached the established limit of 1.5, which occurred about every 2 weeks. During the exposure period, mass loss (cubic specimens) and length variation (prismatic specimens) were measured twice a week and calculated using Equations (1) and (2), according to ASTM C267-01 [57] and ASTM C1012-18b [58], respectively:

$$\Delta M_t=\frac{M_t-M_i}{M_i}\times 100$$

(1)

$$\Delta L_t=\frac{L_t-L_i}{L_g}\times 100$$

(2)

where ΔM_t is mass loss over time (%); M_t is mass at time t (g); M_i is initial mass (g); ΔL_t is length variation over time (%); L_t is specimen length at time t (mm); L_i is the initial specimen (mm); and L_g is the nominal gauge length (250 mm).

After the exposure period, the cubic and cylindrical specimens (aged 84 days) were submitted to compressive strength and water absorption tests, respectively, to evaluate the mortar performance after the attack. The results were compared to those of the same age specimens not submitted to the attack. A degraded layer formed and adhered to the specimen surface. The average thickness of this degraded layer was measured using ImageJ software to analyze the images that were taken after the specimen ruptured. One-way analysis of variance (ANOVA) and Duncan’s multiple-range tests (p ≤ 0.05) were used to compare the test data.

3. Results and Discussion

3.1. Characterization of Blended Cements

The addition of SCBA and LF slightly changed the physical properties and chemical compositions of blended cements compared to the reference, as presented in

Table 3. Chemical composition (% mass) and the main physical properties of cements.

	PC-1	PC-2	PC-3	PC-4	PC-5
CaO	69.96	55.89	68.24	61.18	54.53
SiO2	15.72	23.74	12.96	17.46	21.34
Al2O3	4.10	7.59	3.26	5.18	5.94
SO3	3.62	3.77	3.48	3.55	3.63
Fe2O3	4.71	5.62	4.40	4.63	5.12
K2O	0.60	1.32	0.62	0.84	1.21
TiO2	0.28	0.42	0.44	0.30	0.42
MnO	0.08	0.11	0.08	0.09	0.11
Na2Oeq *	0.39	0.87	0.41	0.56	0.80
LOI	0.91	1.55	6.52	6.77	7.68
D50(μm)	12.8	10.9	11.4	10.5	9.5
Density (g/cm³)	3.07	2.90	3.01	2.90	2.87
Blaine fineness (m²/kg)	325	418	405	527	633
Initial setting time (min)	120	155	120	130	130
Final setting time (min)	180	200	160	180	200

* Na₂O_eq = Na₂O + 0.658 K₂O.

. As expected, incorporating SCBA increased the silica content of blended cements but also raised the cement alkaline equivalent (Na₂O_eq), increasing the expansion potential via the alkali–silica reaction in concretes [59]. Both SCMs increased the loss on ignition of blended cements, especially LF, due to its carbonate composition. The particle size distribution (Figure 4) of all the cements was very similar, allowing them to be compared. However, the incorporation of ultrafine SCBA and LF resulted in lower density and higher Blaine fineness of blended cements compared to PC-1, especially with SCBA because of its high BET specific surface area.

The setting times of blended cements were also influenced by SCM incorporation (Table 3). While SCBA increased the initial and final setting times of PC-2 compared to PC-1, the opposite was observed for PC-3, since LF shortened the final setting time. These effects were observed in other studies with SCBA and LF, separately. The hydration delay promoted by SCBA is related to the presence of contaminants [30,34,60], such as carbonaceous compounds (3.9% LOI) and SO₃ (Table 1) in the ash. On the other hand, the acceleration provided by the ultrafine LF can be attributed to the extra surface area for hydrate nucleation [40,61]. A combination of these effects was observed in the ternary mixes, with a slight increase in the initial setting time of both PC-4 and PC-5 compared to the reference, and the final setting time of PC-5 because of its higher SCBA content compared to PC-4. Nevertheless, the setting times of all blended cements were in accordance with the Brazilian standard for blended Portland cements [62]. Moreover, Figure 5 shows the results of the Frattini test, which indicated that all cements with SCBA (PC-2, PC-4, and PC-5) were classified as pozzolanic Portland cements. This result may be associated with good pozzolanic activity and the large specific surface area of SCBA [20,22,23].

3.2. Mechanical Performance of Mortars

Figure 6 presents the average compressive strength for three specimens of each mortar after 7, 28, and 84 days of curing. According to the results, M-PC-2 showed significantly higher strength than that of the other mortars at all ages. In the present case, the heterogeneous nucleation and pozzolanic effects of SCBA were more effective and significant than the dilution effect of clinker replacement in PC-2, a finding also reported in previous studies [27,34]. On the other hand, the isolated effect of LF in PC-3 was negative for mortar mechanical behavior, since M-PC-3 exhibited significantly lower compressive strength than that of M-PC-1 at all ages and the worst result between all the mortars at 28 and 84 days of curing. The lower compressive strength of M-PC-3 was related to the dilution effect [40,43,44]. As expected, the ternary mixes (M-PC-4 and M-PC-5) showed intermediate results compared to their binary counterparts. The compressive strength of both ternary mortars at 7 days was affected by the high clinker replacement content in PC-4 and PC-5 cements. However, increases in strength were observed in these mortars with long-term curing, especially due to the pozzolanic effect of SCBA. At 28 days, no significant difference was observed in the compressive strength of M-PC-5 and M-PC-1. The LF effect was more pronounced in M-PC-4, which displayed lower strength than M-PC-1 and M-PC-5.

Figure 7 presents the results of water absorption (average of four specimens) after 28 and 84 days of curing. An increase in water absorption was observed for all blended cements compared to the reference at 28 days. However, at 84 days of curing, there was a significant reduction in the water absorption of all the cements related to the previous age, especially for the blended varieties. These results indicate a better effect of SCM at later ages, particularly SCBA because of its pozzolanic activity, which causes gradual pore closing [63,64]. In our case, M-PC-2 exhibited higher water absorption than that of M-PC-1 at 28 days, but after 84 days the binary mix showed significantly lower absorption. On the other hand, LF increased the water absorption of M-PC-3 compared to the reference at both ages because of the ultrafine particles used in the study. The behavior of ternary mixes was similar to that of the binary mortars. While M-PC-4 obtained an increase in absorption owing to the higher LF content, M-PC-5 exhibited a decrease, caused by the higher SCBA content. Furthermore, no significant differences were found between M-PC-5, M-PC-1, and M-PC-2 at 84 days.

3.3. Durability to Sulfuric Acid Attack

As expected, sulfuric acid attack degraded all the mortars, as shown in Figure 8. It is important to emphasize that sulfuric acid consumes the CH from cement pastes, resulting in the formation of calcium sulfate or even ettringite, which are expansive products. The expansion of these products in a hardened mortar (or concrete) causes the degradation of cementitious composites, and C–S–H decalcification caused by acid exposure results in a loss of strength [44,65]. Visually, M-PC-3 was the least affected by the attack, while M-PC-2 showed the greatest apparent degradation.

In this context, the degradation of mortars caused by sulfuric acid attack resulted in mass loss in all the specimens, as shown in Figure 9. M-PC-3 showed better behavior due to the presence of LF. The calcite from LF reacted with the sulfuric acid, protecting the mortar against C–S–H decalcification and enhancing its capacity to neutralize the acid [66,67]. On the other hand, M-PC-2 exhibited the highest mass loss among the mortars studied, including the reference, as previously observed in visual inspection. After 56 days of exposure, the mass loss for M-PC-2, M-PC-1, and M-PC-3 was 43.9, 33.8, and 20.9%, respectively. These results suggest that CH consumption by the pozzolanic reaction of SBCA was not enough to improve the durability of this mortar to sulfuric acid attack, despite the good mechanical behavior of M-PC-2. Similar findings were reported by Senhadji et al. [44] in mortars containing silica fume. A more in-depth investigation on the mechanisms of sulfuric acid attack is needed to better understand the influence of SCBA on the durability properties of cementitious material. Ternary mixes displayed intermediate behavior compared to their binary counterparts and results very similar to those of the reference. At the end of the attack, M-PC-4 and M-PC-5 showed 31.8 and 34.3% mass loss, respectively, which confirmed the positive effect of adding LF to these mortars.

As expected, expansion trends were observed in all mortars during 56 days of exposure, as presented in Figure 10. The four mortars with blended cements had significantly lower expansion than that of the reference, indicating better mortar performance with SCBA and LF during the sulfuric acid attack. At the end of the exposure period, M-PC-1 showed 0.039% expansion, while M-PC-2, M-PC-3, M-PC-4, and M-PC-5 obtained 0.022, 0.008, 0.008, and 0.006%, respectively. Unlike the mass variation results, the mortars containing SCBA (M-PC-2, M-PC-4, and M-PC-5) showed the best length variation results, probably due to pozzolanic reactions, and their expansion values were considered non-significant according to ASTM C1012-18b [58].

Figure 11 shows the formation of a degraded layer on the surface of mortar specimens after acid exposure. The photographs were taken after the rupture of cubic specimens. This layer was composed mainly of calcium sulfate, the main product of cementitious degradation caused by sulfuric acid attack [44]. The average thickness of this degraded layer was 0.619, 0.817, 1.666, 0.869, and 1.724 mm for M-PC-1, M-PC-2, M-PC-3, M-PC-4, and M-PC-5, respectively. Mortars containing blended cements were thicker than the reference, especially mortars containing LF, indicating that blended cements were more degraded, since more gypsum was formed in the presence of SCM.

Figure 12 and Figure 13 present the compressive strength and water absorption results of mortars, respectively, after 56 days of acid exposure compared to specimens that were not attacked. As expected, there was a clear decline in the strength of all the mortars after the sulfuric acid attack due to C–S–H decalcification [44,67] and a significant increase in water absorption after the exposure period caused by mortar deterioration in the acid environment. Figure 12 shows that despite the good performance of M-PC-3 and M-PC-5 in terms of mass loss, these mortars presented significantly lower strength than that of the reference after the attack and the greatest decline in strength compared to the specimens cured in limewater. The loss of strength was 49 and 47% for M-PC-3 and M-PC-5, respectively, while M-PC-1 lost 31% of its compressive strength compared to the nonattacked mortar. On the other hand, M-PC-2 and M-PC-4 lost 41 and 31% of their strength, respectively, albeit with no significant difference compared to the reference mortar after the attack.

Figure 13 shows that the presence of SCBA did not favor water absorption of M-PC-2 and M-PC-5, which obtained increases of 29 and 32%, respectively, when compared to mortars of the same age cured in limewater. The rise in water absorption may have caused mortar deterioration and mass loss during acid exposure. M-PC-1, M-PC-3, and M-PC-4 showed increased water absorption of 12, 7, and 12% respectively, indicating good LF performance. Calcium sulfate formation due to acid exposure may have caused densification of the degraded layer, as observed in Figure 11, enhancing the absorption of LF mortars [68]. After acid exposure, M-PC-1 exhibited the lowest water absorption, but no significant difference was observed between this mix, M-PC-2, and M-PC-3. The higher SCM content influenced ternary mortar absorption, which displayed the highest values after the attack.

The results showed that adding LF decreased the water absorption and mass variation of mortars, while SCBA showed good mechanical behavior and reduced expansion potential during sulfuric acid exposure. The combined effects of these materials resulted in cements with good durability to sulfuric acid attack. Limestone acted as a sacrificial medium, causing less deterioration because of its higher calcium content, and SCBA decreased the CH content in mortars due to the pozzolanic reaction [44,66,67]. Different contents of these SCMs as clinker replacements should be studied to optimize their effects on cementitious systems and enable the production of sustainable cements.

4. Conclusions

The present study aimed at evaluating the effects of partial clinker replacement by SCBA and LF on the compressive strength and durability against a sulfuric acid attack of mortars. Based on the test results, the following conclusions can be drawn:

• The incorporation of ultrafine SCBA and LF into Portland blended cements significantly increased the Blaine fineness of cements, especially those containing ash due to its higher specific surface area. In addition, the binary clinker–SCBA mix and both ternary mixes were classified as pozzolanic Portland cements, indicating that the reactivity of SCBA was more effective than the dilution effect of 28% clinker replacement. However, SCBA increased cement setting times, delaying hydration due to the presence of contaminants in the ash. On the other hand, LF accelerated hydration because of the nucleation effect, reducing the final setting time compared to the reference.
• SCBA had a positive effect on the compressive strength and water absorption of mortars due to its pozzolanic activity, especially at later ages. However, the dilution effect was more pronounced with the addition of LF, and the binary clinker–LF mortar displayed the worst performance among all the mortars after 28 days. In the ternary mixes, strength was compromised in the early days due to the dilution effect, but the action of SCBA contributed to the mechanical performance of these mortars after 28 days of curing.
• With respect to durability tests, blended cements showed good performance compared to the reference. Cements containing LF exhibited excellent mass loss and water absorption, while SCBA provided lower length variation and good mechanical behavior after the sulfuric acid attack. The combined effect of these materials was evident in ternary mixes, which displayed good durability, with little mass and length variation, and good mechanical performance after the attack.

The results obtained in this work revealed the significant potential of SBCA and LF combinations for sustainable cement production with up to 28% clinker replacement. Future research should investigate higher clinker replacements with different contents of SCBA and LF, optimizing the effects of both SCMs. In addition, a life cycle assessment study could be performed to evaluate the environmental influence of blended cements on the reduction of CO₂ emissions. In addition, the durability of cements in other aggressive environments should be evaluated.