The SCBA was collected from a sugar mill located in the community of Tezonapa, Veracruz, México. This ash is generated as a combustion by-product of sugarcane bagasse at temperatures between 550 and 700^oC and recovered by sprinkling water during the sugar production. The collected ash was homogenized and dried for 24 hours in an electric oven at 105^oC.

In order to select the SCBA to be used in subsequent phases of this research, its pozzolanic potential was estimated at 7, 14 and 28 days after the ash was subjected to two low-energy post-treatments. The first post-treatment consisted of sieving the SCBA for five minutes through No. 8 (2.36 mm), No. 100 (150 μm) and No. 200 (75 μm) ASTM sieves (SCBA8, SCBA100 and SCBA200). The fractions removed in the sieving process were 0, 20 y 30% for the No. 8 (2.36 mm), No. 100 (150 μm) and No. 200 (75 μm) sieves, respectively.

The second post-treatment consisted of sieving through the same sieves for the same period of time followed by two hours of grinding (SCBA8g, SCBA100g and SCBA200g). Sieving was carried out to reduce the particle size and to remove organic material, and posterior grinding was done solely to increase the specific surface area of the SCBA (7, 9,12,19–21). Ordinary Portland Cement CPO 40 Lafarge^®, SCBA from each post-treatment (at 20% cement replacement), standard silica sand and distilled water were used to prepare mortar cubes in order to obtain the strength activity indexes (SAI) (ASTM C 311) of the six post-treated SCBAs. A mortar mixture (FA20) containing 20% Admix Tech^® Class F Fly Ash (FA) was prepared as a reference; the FA meets the requirements described in the ASTM C 618 standard.

Originally, water/cement or cementitious materials and sand/cement ratios of 0.484 and 2.75 were used to prepare the mortars in accordance with the ASTM C 311. However, a water/cement ratio of 0.57 for the control mortar and a water/cementitious material ratio of 0.68 for mortars containing 20% SCBA were necessary to keep the sand/cementitious material ratio of 2.75 (ASTM C 311) and to fulfill the mortars’ flow of 110±5% (ASTM C 1437). A water/cementitious material ratio of 0.56 was used for the mortar containing 20% FA to keep the same sand/cement ratio and as well fulfill the mortars’ flow requirements. No superplasticizer was used in these experiments.

SAIs for sieved and ground SCBAs and for the FA estimated at 7, 14 and 28 days are presented in Figure 1. In accordance with the literature, the SAI value increases when the particle size of the SCBA decreases as a result of sieving or grinding; the SAI also increases with age as a consequence of the pozzolanic reaction (7–9,12,21).

Most mixtures had 28-day compressive strength values higher than 75% of the control mixture specified by the ASTM C 618 (this percentage was considered as a reference point for the analysis since the ASTM C 618 accounts for the various materials from natural resources requiring calcination to induce satisfactory properties). The only value smaller than 75% was for the mortar containing the material that passed the No. 8 sieve (SCBA8).

Poor performance of grinding SCBA after passing through the No. 8 and 100 sieves is supported by the fact that the wider sieves allow the passage of larger carbon particles, while a posterior grinding is not effective in increasing the SAI, as carbon particles are ground together along with particles rich in silica and alumina.

The mixture with 20% FA (FA20) yielded the highest SAIs at the studied ages (98, 95 and 104% at 7, 14 and 28 days respectively) followed by the mixture with the combination of sieved plus ground SCBA (SCBA200g) (90, 87 and 95 at 7, 14 and 28 days respectively), and then only sieved SCBA (SCBA200) (85, 91 and 92% at 7, 14 and 28 days respectively). As may be observed, the difference between SCBA200g and SCBA200 was not significant.

Based on the above information, it was decided to use only SCBA sieved through sieve No. 200 (75 μm) (SCBA200) for the remaining stages of the present research. The SAI at the three ages of the selected SCBA was more than 75%; therefore, this material can be considered a natural pozzolan, in accordance with the ASTM C 618.

From analyzing the results of the previous section, SCBA200, the material in which the post-treatment consisted of sieving it only through the No. 200 (75 μm) sieve for five minutes, was selected as the study subject for the following phase of this research. For brevity SCBA200 will be further designated only as UtSCBA (Untreated Sugarcane Bagasses Ash).

Blended Portland Cement 30R Holcim Apasco® (CPC) (approximately 5% ground granulated blast furnace slag), which is readily available in southwestern México, was chosen as the cementing material.

Chemical composition, particle size distribution (PSD), density (ASTM C 188) and specific surface area (SSA) of the CPC and UtSCBA were obtained. The PSDs were obtained by wet method (via isopropyl alcohol) using a laser analyzer Microtrac S3500®; shape particles and refractive indices of 1.70 and 1.54 for CPC and UtSCBA were considered respectively. The SSAs were obtained by physical adsorption of nitrogen using a Quantachrome Nova 2000e® analyzer; the BET method was used for the analysis. The morphology and mineralogical phases of the UtSCBA were analyzed by SEM and XRD, respectively. The morphology of the materials was observed using a scanning electron microscope JEOL JSM-6490LV® equipped with an X-ray scattering analyzer (EDS) from Oxford Instruments 7573®. Mineralogical phases were identified using a diffractometer Bruker D8 Advance®, which uses radiation of CuK and a wavelength of 1.5418Ǻ to a passage of 0.05^o and a time of incidence of 0.5 seconds per step, maintaining an interval 2θ from 10^o to 70^o.

UtSCBA, river sand (with a density of 2.7g/cm³ and a fineness modulus of 2.45 (ASTM C 33), distilled water and a high-range water reducer (HRWR), Plastol 4000^® (ml per kg of cementitious materials), were used to prepare the mortars. All mortars had a 0.63 water/cementitious material, a 1:3 cementitious material to sand ratio, and were prepared replacing 0, 10 and 20% of the weight of cement by UtSCBA (Control, UtSCBA10 and UtSCBA20 mixtures, respectively) (Table 1).

Flow table (ASTM C1437), temperature (ASTM C1064/C1064M), volumetric weight (ASTM C138/C138M) and air content (ASTM C231) tests of the mortars were carried out. Cylindrical specimens with a diameter of 75 mm and a height of 150 mm were prepared to study the microstructure and obtain the compressive strength. All specimens were cast in two layers and compacted using a vibrating table. Only the mixtures with an addition of UtSCBA required HRWR to maintain the workability. Specimens were cured in a 3% calcium hydroxide solution until the time of the test.

The morphology of gold-palladium-covered mortar samples at 1, 7, 28, 90 and 600 days were analyzed using a scanning electron microscope JEOL JSM-6490LV®. The mineralogical phases of mortar samples ground and sieved through a 150 μm mesh (ASTM) at 1, 7, 28, 90 and 600 days were identified by XRD using a diffractometer Bruker D8 Advance® and were analyzed by the software EVA version 11.0.0.3^®. The compressive strengths (ASTM C 39/C39M) at 1, 7, 28, 56, 90, and 450 days were obtained by using a hydraulic press ELVEC E 659®. The compressive strength at 600 days was obtained from small mortar cubes (50 x 50 x 50 mm). These cubes were obtained from cylinders used during the microstructural evaluation. The results from the cubes were corrected by size and shape. For this purpose, the conversion factors for size and shape for normal-strength concrete proposed by Seong-Tae et al. (22) were implemented.

Chemical analysis of the selected UtSCBA (Table 2) shows that the sum of major oxides (SiO₂ + Al₂O₃ + Fe₂O₃ = 76%) is slightly larger than 70% of the overall material composition (ASTM C 618); this result along with the SAI estimate corroborates the pozzolanic activity of the UtSCBA despite the high loss on ignition (LOI) content.

The analysis of the physical characteristics shows that the UtSCBA has lower density, greater mean PSD and larger SSA than the CPC (Table 3).

Figure 2a shows the PSDs for CPC and UtSCBA. It can be observed that the UtSCBA has a PSD between 3.8 and 250 μm and CPC between 0.7 and 400 μm. The results indicate that UtSCBA has larger size particles when compared to CPC.

A detailed analysis of the PSD of these materials can be done analyzing their density distributions (Figure 2b). UtSCBA shows a monodisperse distribution, which can be associated with a sieved material (23). On the other hand, it can be observed that for the CPC the PSD shows two particle populations (bimodal) at 7 and 30 μm. This type of distribution can be caused by two factors: the presence of particles from the different mineralogical components because a compound cement was used (13), and the result of grinding because fine particles tend to cluster (7).

The UtSCBA show grains of varied shapes and sizes in accordance with the calcination conditions (Figure 3) (12,24). There are agglomerated particles with high porosity, typical for organic materials, and prismatic particles with well-defined edges (9,25–26).The EDS microanalysis revealed that the agglomerated particles (A) are rich in Si, Al and C oxides and not so rich in Ca oxides (8). The prismatic particles (B) are constituted mostly by Si. Furthermore, spherical (C) and fibrous carbon particles (unburned carbon) (D) were observed. The spherical particles are rich in Ca and Si oxides as well as Al and Fe, and the fibrous particles contain mainly Si and Al. In this respect, recent studies (27) found that some fibrous carbon particles from SCBA may be covered by Si and O depending on the pyrolysis of the bagasse, affirming that the unburned carbon can present amorphicity.

In summary, the high LOI, lower density, greater PSD, larger SSA (in comparison with CPC), simultaneous appearance of amorphous and crystalline phases, and variability of the the UtSCBA’s shapes and sizes can be attributed to the variations of temperature and air flow during the bagasse’s calcination process in the sugar mill, in accordance with results observed by other authors (24).

The results of the XRD mineralogical analyses show that the CPC (Figure 4a) contains mainly di-calcium and tri-calcium silicate, tri-calcium aluminate and tetra-calcium aluminoferrite, and to a lesser extent calcium oxide, calcium carbonate and gypsum. The results of the XRD mineralogical analyses (Figure 4b) show that the UtSCBA exhibited amorphous and crystalline phases. The amorphous phase in the UtSCBA was observed as a diffuse halo in the interval 2θ from 10^° to 35^°. This halo, which is characteristic of pozzolanic materials, suggests the presence of amorphous substances such as silica (6,8,15,24,28–29). The crystalline phases of quartz and cristobalite in the UtSCBA were detected; both phases have been reported by others (6, 25, 29). Also gibbsite (Al₂O₃), hematite (Fe₂O₃) and calcium (Ca) phases were identified, and the residual carbon (C) phase in the UtSCBA, commonly attributed to the bagasse’s LOIs (8,13), was detected at 2θ of 22.8^° and 26.6^°. However, the identification of carbon in the crystalline phase was little reliable due to the small quantity as well as the overlapping observed in the signal.

Mortars prepared with UtSCBA presented workability problems which were overcome with a superplasticizer, as previously demonstrated in the section on the proportioning of the mortar mixture. The high LOI content, an indication of the high level of unburned carbon particles, can be blamed for this. Chandara et al. (30) mentioned that high levels of unburned carbon particles increase the requirement for water, and Jiménez-Quero et al. (31) reported that the variety of shapes of sieved UtSCBA particles can reduce the flow of the mortar mixtures since these particles can increase friction during the mixing process.

Mortar with 20% UtSCBA was less dense than the other mortars (Table 4). This was expected, as UtSCBA is a porous material with a lower density, greater mean PSD, and larger SSA than CPC.

The micrograph of the control mortar at day 1 shows a heterogeneous matrix. The phases of C-S-H and CH created from the Portland cement hydration were found to be in the majority of the cementitious products (32–33). At 1 and 7 days, unhydrated particles of cement and particles with layers of hydration products were observed (Figures 5a and 5b) (32–34). Moreover, discontinuities in the interphase aggregate-cementitious paste were observed (Figure 5a) at day 1. At 28, 90 and 600 days the matrix of the mortars was denser (Figures 5c, 5d and 5e); nevertheless, porous zones with cementitious materials and cement particles were still observed.

Similarly to the control mortar, micrographs of mortars containing 10 and 20% UtSCBA taken at day 1 (Figures 6a and 7a) show some cement particles, unreacted particles of UtSCBA and some cementitious products on the surface of the UtSCBA particles. The EDS microanalyses (Figure 6a) show the different elements on the surface of a prismatic particle of UtSCBA. In UtSCBA mortars it was observed that as time passes, the cementitious matrixes were denser than the control mortar. The same effect was found by Govindarajen et al. (35) in cement pastes with 10% SCBA evaluated at 7 and 28 days, and in concrete when up to 30% SCBA was added and evaluated until 180 days of age (36).

The information mentioned above supports the occurrence of pozzolanic reactions as a consequence of the CH consumption and the formation of C-S-H and calcium-aluminum silicate hydrate (C-A-S-H) (Figures 6b, 6c, 6d, 6e, 7b, 7c, 7d and 7e) improving in this way compressive strength at later ages. However, at 90 days unreacted prismatic particles of UtSCBA were observed (Figure 6d) in the UtSCBA10 mortar; these particles were the largest in size and could improve the cementitious matrix at later stages. Finally, at 600 days it can be seen as a compacted matrix in both UtSCBA mortars when compared to earlier ages.

The XRD patterns of the control mortar (Figure 8) show that the C-S-H (4CaO.5SiO₂.5H₂O) and CH (Ca(OH)₂) were the main mineralogical phases formed; both phases were produced during the Portland cement hydration. At day 1 the phase of calcium silicate (CaSiO₃), attributed to unhydrated particles in the cement, was detected; at later ages the intensity of the peaks of this phase decreased. At 28, 90 and 600 days a phase of C-S-H with different stoichiometry (CaSiO₃.H₂O) and the phase of C-A-S-H (CaAl₂Si₃O_10.6H₂O) were detected. The appearance of two or more phases of C-S-H is known and reported by other authors (32, 37–38). The C-A-S-H is a product formed after activation of the aluminate and silicate phases in the presence of CH and could improve the microstructure and increase the durability of mortars (38). Other phases like quartz (SiO₂) and anorthite attributed to the sand, and Calcite (CaCO₃) attributed to the CPC, were also detected.

The XRD patterns of the UtSCBA10 and UtSCBA20 mortars (Figures 9 and 10) show the increase of the peak intensities of the C-S-H and C-A-S-H phases on the mortars as age progressed. Both products were created as the result of pozzolanic reactions between the UtSCBA, CH and water; this observation is supported by the decreased intensity of the CH phase and the increase of the C-S-H phase in comparison to the control mortar, as observed at 28,90 and 600 days. The consumption of CH and the increase of the C-S-H in lime and cement pastes with the addition of SCBA at ages between 3 and 28 days has been reported (25,35).

In the present research, when 10% of cement is replaced by UtSCBA a pozzolanic reaction occurs and CH still remains in the matrix until the active material in the pozzolan is depleted. When 20% cement is replaced by UtSCBA, the CH is once again consumed but at a slower rate than for 10% UtSCBA. In this mixture there is a larger amount of silica present but also a larger amount of carbon, and carbon can be blamed for significantly inhibiting the pozzolanic reaction (25). When 20% UtSCBA is incorporated there is still a considerable amount of CH because only a very large replacement of cement by a pozzolan depletes the CH content. For example, when more than 50% FA is used to replace cement the CH is totally consumed (39).

The above mentioned information and the physical characteristics of the UtSCBA explain the increase in the compressive strength of mortars with the addition of UtSCBA at later ages.

Moreover, the phase of cristobalite showed variations at different ages. These variations can be attributed to the disordered structures or its metastable state, which could improve pozzolanicity (40–41).

Compressive strength of the mortar mixtures containing UtSCBA increased over time. The addition of 10 and 20% UtSCBA (UtSCBA10 and UtSCBA20) increased the strength at 28, 56, 90, 450 and 600 days when compared to the control mortar (Figure 11). However, the beneficial effect on the strength when adding UtSCBA was noted only after 28 days. The same effect was observed by other authors when ground SCBA was used in concrete mixtures (36).

Moreover, it was observed that when the content of UtSCBA increased from 10 to 20%, the compressive strength at 1, 7, 28, 56 and 90 days decreased between 7 and 12%. Chusilp et al. (11) reported this effect in mortars with ground bagasse ash and 10% LOI (similar to the LOI obtained in this research). In contrast, compressive strength results at 450 and 600 days show that when the content of UtSCBA increased from 10 to 20% in mortars mixtures, a slight increase in strength occurred. Results show that the addition of 20% UtSCBA increased the compressive strength by 10.6% and 11.2% at 450 and 600 days respectively compared with the compressive strength at 90 days, while the mortar with addition of 10% UtSCBA remains practically constant at 90, 450 and 600 days. In this respect, Chusilp et al. (16) reported a similar behavior using 10 and 20% ground SCBA (95% of the ash passing the No. 325 sieve and LOI of 8.16%) in concrete mixtures at 28 and 90 days, in which the compressive strength of the concrete with the addition of 20% SCBA was greater than that of the concrete with the addition of 10% SCBA.

The above mentioned could be attributed to the fact that after 90 days the UtSCBA20 mortar has more unreacted UtSCBA, which improves the pozzolanic activity in comparison with the UtSCBA10 mortar. Likewise, the large fibrous-porous particles of UtSCBA (which are covered with a layer of Si and O), as well as the small particles, increase the specific surface area for the pozzolanic activity; consequently there is a gain in strength when 20% UtSCBA is added to the mortar.

The results indicate that, despite using a basically untreated SCBA as a cement replacement for preparing mortars, ash like this helps improve the long-term compressive strength and the microstructure of such mortars.