Measurement of Pozzolanic Activity Index of Scoria , Pumice , and Rice Husk Ash as Potential Supplementary Cementitious Materials for Portland Cement

1Department of Materials, Energy Science and Engineering, The Nelson Mandela Institution of Science and Technology, P.O. Box 447, Arusha, Tanzania 2Department of Built Environment Engineering, Mbeya University of Science and Technology, P.O. Box 131, Mbeya, Tanzania 3Department of Chemistry, South Eastern Kenya University, P.O. Box 170-90200, Kitui, Kenya 4Department of Structural and Construction Engineering, University of Dar es Salaam, P.O. Box 35131, Dar es Salaam, Tanzania


Introduction
Mortar is a workable paste prepared by adding water to a mixture of sand and binding materials like cement or lime.Cement-sand mortar is used for binding building materials such as clay bricks, concrete blocks, stones, and coarse aggregates in concrete and for plastering.The strength and durability of a mortar or a concrete is a function of the volume and quality of the binder and the water/binder ratio used.Hydration of cement gives the strength of mortar and concrete; the hydration products fill the pores between the aggregates.The addition of supplementary cementitious materials (SCMs) reduces the amount of cement and improves the properties of mortar and concrete such as mechanical strengths, resistance to chemical attack, reduced permeability, and shrinkage [1,2].SCMs are pozzolans with a high amount of silica (SiO 2 ) and alumina (Al 2 O 3 ).As such, they react, in the presence of water, with calcium hydroxide (CH) at ordinary temperature to form calcium silicate hydrates (C-S-H) and calcium aluminosilicate hydrates (C-A-H).Both the C-S-H and C-A-H bind the aggregates and give additional strength to mortar and concrete.Fly ash, silica fume, ground granulated blast furnace slag, diatomaceous earth, and metakaolin are frequently used SCMs.Sugar cane bagasse ash and rice husk ash (RHA) obtained from agricultural byproducts are also used as SCMs.The use of SCMs is driven by reasons of their abundance and economic and environmental benefits and today's advances in cement technology.The influence of SCMs on the properties of blended cement in mortar and concrete such as setting times and early strength depends on the quantity and quality of the reactive SiO 2 and Al 2 O 3 present, particle size distribution (PSD), and the specific surface area.These three dictate the amounts of CH consumed by SCMs and the degree of C-S-H and C-A-H formation.In contrast, PSD
A number of methods, categorized as direct and indirect, have been developed for the qualitative and quantitative determination of these indices.Direct and indirect methods measure the amount of CH consumed and the relative compressive strength, respectively [4].The direct methods involve chemical titration, electrical conductivity measurement, and thermogravimetric analysis and so forth [3].The indirect method compares the relative compressive strength of pozzolan blended cement mortar to that of Portland cement mortar.The reliability of each method is an important factor.For instance, Luxán et al. 's method (direct method) is accurate for early age determination of PAI but McCarter and Tran's method (also a direct method) is more accurate for the later ages [5].Both of these methods are based on conductivity [6].Luxán et al. [4] studied PAI of natural rock pozzolan by monitoring the change in conductivity during the first 2 min of a reaction between 5 g of pozzolan and 200 mL of saturated Ca(OH) 2 solution at 40 ± 1 ∘ C. 2 min were found sufficient for the stabilization of the reaction.Also, McCarter and Tran [5] measured conductivity of various pozzolans mixed with Ca(OH) 2 powder and water for 48 h.Stiffening of the paste was observed between 24 and 48 h.On the other hand, Wansom et al. [6] studied and compared various electrical methods of measuring PAI of RHA for 1 h in a saturated Ca(OH) 2 -pozzolan-water system.Paya et al. [7] modified Luxán et al. 's method by using 800 mg of Ca(OH) 2 powder in 1000 mL of deionized water instead of the saturated solution of Ca(OH) 2 .In their study, the proposed observation time was 120 min but the actual time used was ∼17 min.The above studies limited the time for PAI determination to 48 h.In this study, the time was prolonged to 28 days to reflect the required curing period of cement to achieve ultimate compressive strength.Furthermore, time for taking the readings was tailored to 2, 7, and 28 days to match the curing period for measuring compressive strength of cement-sand mortar.This work aimed at exploring new SCMs (scoria and pumice), comparing their pozzolanic activity index and performance in the mortar with those made with RHA.Scoria and pumice are pyroclastic ejecta which is found in many areas around the world [8].Scoria is composed of red colored fragments indicating the presence of iron.It is highly vesicular and composed of crystalline structure.Pumice is an extrusive volcanic rock material, produced when lava, which is full of water and gases, is ejected on the earth surface by volcanic action [9].It is light-colored and extremely vesicular and composed of high silica and alumina and low in iron and magnesium.The chemical compositions of scoria and pumice from different sources vary depending on the mineralogical composition, formation temperature, and rate of cooling.The chemical composition of scoria and pumice is shown in Table 1.Although scoria and pumice have pozzolanic properties, they have not yet been used as SCMs to the best of our knowledge.
Rice husk ash is produced when rice husk is incinerated at low temperature (<800 ∘ C) with a controlled supply of air to achieve the reactive silica phase [6].When incinerated above 800 ∘ C, amorphous silica converts into crystalline silica phase.The properties of the ash depend on chemical composition, purity, fineness, and the amount of amorphous silica which decreases with calcination [10,11].According to Della et al. [12], RHA burned at 700 ∘ C has 94.95% of reactive silica (Table 1).Wansom et al. [6] and Givi et al. [13] state that burning of RHA at different temperatures and supply of oxygen results in different amounts of reactive silica.

Materials and Methods
2.1.Materials.This study was carried out using scoria obtained from Uchira, Kilimanjaro, Tanzania, and pumice and rice husk from Mbeya, Tanzania.Analytical grade calcium hydroxide (Ca(OH) 2 ) and Portland Limestone Cement (PLC) CEM II/A-L class 42.5 N conforming to BS 197-1 were used.Standard sand prepared according to SS-EN 196-1 was also used.

Experimental
Procedures.Samples of scoria, pumice, and rice husk were calcined at different temperatures ranging from 600 to 900 ∘ C in a muffle furnace at a rate of 7 ∘ C/min.They were kept at the respective temperature for 1 h and then cooled to room temperature.Natural samples of scoria and pumice and a sample of rice husk ash incinerated in uncontrolled condition were included in this study for comparison purposes.The incineration in uncontrolled condition was set in a small room of 2 m 2 and the rice husks were sprayed to 0.5 m depth.The ash obtained from this condition was designated as RHA-uc.All samples were milled using disc mill model 4A100L6T1 SN 535277 at a rate of 2 kg/h and sieved through 75  BS standard sieve mounted on automatic sieve shaker [14].95% of the weight of each sample passed this sieve.The samples designations and notation are shown in Table 2.

Determination of Chemical and Phase Composition.
Chemical and phase compositions of the samples in Table 2 were analyzed by X-ray fluorescence (XRF) and X-ray diffraction (XRD) techniques [18,19].XRF analysis was performed using Bruker XRF-S8 Tiger spectrometer, while XRD analysis was done using Bruker S8 diffractometer with Cu K  radiation.Scanning was performed within 2 ∘ ≤ 2 theta ≤ 60 ∘ for the XRF samples and 5 ∘ ≤ 2 theta ≤ 70 ∘ for the XRD samples.The scanning rate for both XRF and XRD was 0.02 degrees/sec.The analyses were done according to ASTM C 618 [20] which requires the sum of the basic oxides, silica, alumina, and ferrous, to be higher than 70% for pozzolana.

Determination of Setting Time and Soundness.
Setting times and soundness were determined according to SS-EN 196-3 [19] from standard consistency paste prepared by replacing cement with 10, 20, 30, and 40% of scoria, pumice, or RHA treated at different temperatures (Table 2).
In principle, standard consistency paste is the one having water/cement ratio between 26 and 33% [26].Initial setting time of cement should not be less than 30 min.SS-EN 197-1 [21], however, specifies an initial setting time of 60 min and a final setting time of no more than 10 h for cement strength class 42.5.There is no limit which is given for the final setting time.Neville and Brooks, [26] suggested 90 min + 1.2 times the initial setting time as the final setting time.

Materials Chemical Composition and Physical Properties.
Table 3 shows results for the chemical composition of scoria, pumice, and RHA materials treated at different temperatures.
The mean value for silica (SiO 2 ) content was found to be 42.9, 57.5, and 92.7%; basic oxide 71.1, 78.4, and 93.3%; calcium oxide (CaO) 10.1, 1.1, and 0.66%; alkali 3.2, 9.1, and 1.6%; and loss on ignition 5.2, 7.4, and 1.9% for scoria, pumice, and RHA samples, respectively.Scoria, pumice, and RHA samples indicated significant difference with the degree of marginal difference  < 0.05 for the basic oxides and loss on ignition.Observation on the chemical composition revealed that significant influence was caused more by the type of materials than the calcination temperature.Scoria samples showed high CaO content indicating high latent hydraulic properties.Pumice, on the other hand, had low calcium content with low hydraulic properties.The chemical compositions of scoria and pumice were comparable with those reported in the literature (Table 1).The sum of the basic oxides observed was over 70% as required by ASTM C 618 [20] for a material to qualify as a pozzolan.Table 3 shows the specific gravity (SG) and specific surface area of scoria, pumice, and RHA.The specific gravity increased with calcination temperature for both the pumice and RHA samples.The natural pumice sample (P-N) showed an SG of 2.39.Upon calcination at 900 ∘ C, the specific gravity increased to 2.50.Similarly, SG increased from 2.17 for RHA-uc sample to 2.28 for RHA sample calcined at 900 ∘ C (RHA 900) (Table 3).There was no obvious trend between specific gravity and calcination temperature observed for the scoria samples.The increase in specific gravity with temperature for pumice and RHA samples was probably due to the reduction of porosity and increased crystallinity as the materials sintered at higher temperatures.Similarly, specific surface area of scoria and RHA decreased with calcination temperatures from 575 (for S-N) to 325 m 2 /kg (for S 900) and from 611 (for RHA-uc) to 282 m 2 /kg (for RHA 900), respectively.This was probably due to the strengthening of the microstructure via neck growth and agglomeration, densification, and coarsening of the microstructure, which occurred as the materials sintered at elevated temperatures.The low specific gravities of scoria, pumice, and RHA as compared to Portland cement (3.15) are attributed to the porous microstructure.No obvious trend between surface area and calcination temperature was observed for pumice samples.
Figure 1 shows the XRD pattern of materials at different calcination temperatures.Figure 1(a) shows multiple sharp peaks indicative of polycrystalline nature of scoria.There was no observable change in crystallinity of scoria with calcination temperature as peak intensities and widths remained almost the same except for a peak around 6 ∘ 2 theta which disappeared at higher temperatures.The main phases observed were pyroxene (augite/diopside) and hematite.Similar to scoria materials, no changes in crystallinity with calcination temperature were observed for pumice samples (Figure 1(b)).The peak at 33 ∘ 2 theta for P 600-900 could be due to preferred orientation.The main phases observed were feldspar (anorthoclase/anorthite) and mica.On the contrary, RHA samples showed dramatic changes in peak intensities and widths with calcination temperature (Figure 1(c)).The XRD pattern for RHA-uc and RHA 600 showed one broad peak at around 22 ∘ 2 theta attributed to amorphous silica [11,30].This peak becomes narrow and intense at temperatures beyond 600 ∘ C. At 800-900 ∘ C, more sharp peaks evolved correspondingly to crystalline phase (cristobalite).

Particle Size Distribution (PSD).
Figure 2 shows the PSD profiles of scoria, pumice, and RHA.The PSD profiles indicated well-graded particles where all particles, from coarse to fine, were present.Fine particles are more reactive and have good parking effect.The specific surface area and PSD not only are critical to pozzolanic reactions but also translate to reduced setting times and enhanced final strength and packing density of mortar and concrete.This   also implies reduced cracking, shrinkage, permeability, and chemical attack of a mortar or concrete.of Ca 2+ ions reduced with time, the rates of PAI decreased and stabilized when all the Ca 2+ ions were consumed.High PAI of RHA samples correlates well with the XRF and the XRD results in Table 3 and Figure 1(c) which show high amorphous silica content in RHA samples compared to scoria and pumice.Calcination temperature and hence the degree of crystallinity were found to have a strong influence on PAI.This was attributed to the fact that, at low temperatures, less crystalline silica is obtained, which is more reactive than the highly crystalline silica obtained at high temperatures.The increase in the degree of crystallinity with temperature is clearly exhibited by XRD patterns in Figure 1(c), which show an evolution of narrow and intense peaks with the increase in temperature.The RHA-uc and RHA 600 and 700 samples were rather amorphous compared to their scoria and pumice counterparts (Figures 1(a) and 1(b)).This explains why the PAI of RHA-uc, RHA 600, and RHA 700 samples were higher at 88, 86, and 87% compared to 82, 75, and 75 and 70, 52, and 49% for S-N, S 600, and S 700 and P-N, P 600, and P 700, respectively, at 7 days.In contrast, pumice attained 86, 75, and 75% for P-N, P 600, and P 700 at 28 days.The failures of pumice samples to achieve the minimum PAI of 75% stipulated by ASTM C 618 within 7 days indicated their inability to attain sufficient strength within 7 days.The PAI results show, however, that composite cement made of these materials would achieve the ultimate compressive strength within 28 days.

Setting Times.
Prior to setting times and soundness tests, standard consistency paste was prepared from cement blended with scoria, pumice, or RHA at 10, 20, 30, and 40% replacement levels.The standard consistency paste indicated  water-cement ratio requirement of between 0.26 and 0.33 except for RHA-uc, RHA 600, and RHA 700, which showed water-cement ratio of 0.38 at 40% replacement level.Table 4 shows the influence of materials and fineness on setting times of blended cement.The analysis of experimental data indicated that materials and fineness affected the setting times significantly.The level of marginal significance observed for the initial setting time and final setting times was  < 0.05 for fineness and materials, correspondingly (Table 4, Figures 4 and 5).The difference was due to variation in mineralogy, chemical composition, and fineness, which dictate the rate and type of chemical reactions to take place.Natural scoria (S-N) was found to be very reactive; as such, it gave the least mean initial setting time of 80 min (Figure 5(a)).This was attributed to the low crystallinity and high fineness of 575 m 2 /kg.Nevertheless, natural pumice (P-N) reacted slowly despite its high fineness of 506 m 2 /kg because of delayed pozzolanic reaction.Although RHA 600 was amorphous and had a fineness of 462 m 2 /kg, it showed significantly delayed setting times (154 min initial setting time and 190 min final setting time) owing to the high water content (Figure 5(c)).The highest mean initial setting times observed were 103, 146, and 154 min for S 800, P 900, and RHA 600, respectively.Likewise, the highest final setting times (FST) were 156, 182, and 190 min for S 900, P 900, and RHA 600, respectively (Figures 4 and 5).Initial and final mean setting times observed were shorter compared to those of Portland pozzolana cement, which were 166 and 285 min, respectively.6 and 7 show shrinkage and soundness test results.The statistical analysis indicated that the type of material and fineness had a major influence on soundness and shrinkage.Marginal significance for the type of material and fineness  < 0.05 (Table 4) was observed between pumice-RHA and pumice-scoria (Figure 6) correspondingly.The maximum values observed for the materials were 2 and 1.5 mm for shrinkage and soundness, respectively.The minimum and maximum mean values for shrinkage and soundness observed were 0.5 and 1.0 mm, respectively.These results imply that MgO and alkali had no effect on the blended cement and hence the MgO and alkali contents in our SCMs are within the acceptable levels.The values obtained for soundness and shrinkage in our study are acceptable according to SS-EN 196-3 [19].

Compressive Strength of Scoria and Pumice Blended
Cement.Table 5 shows statistical analysis of compressive strength test results of scoria, pumice, and RHA blended  Advances in Civil Engineering  cement.The level of cement replacement, fineness, and type of material was found to have a huge influence on the compressive strength development.The level of marginal significance observed for the level of cement replacement, fineness, and type of materials was  < 0.05.Compressive strengths of all samples increased with curing age but decreased with the percent of cement replaced (Figure 8).Compressive strength increased fast up to 28 days beyond Advances in Civil Engineering which the growth rate decreased gradually with curing period as the pozzolanic reaction depleted both the CH and silica content.
Mean compressive strength of all samples after 2, 7, and 28 days indicates that pumice and RHA samples performed better compared to scoria samples (Figures 9(a)-9(c)).Scoria samples treated at high temperature reacted poorly due to low amorphous silica content hence contributing to low mean compressive strength.This was probably due to the decrease in the amount of reactive amorphous silica content and evolution of more nonreactive crystalline silica content with temperature.Pumice reacted slowly with CH from hydrating cement and gained the mean ultimate compressive strength of over 42.5 MPa at 28 days.The mean ultimate compressive strengths at 28-day curing period for scoria, pumice, and RHA blended cement were rather similar since over 95    percent of the pozzolanic reaction was over.The ultimate mean compressive strengths achieved at 28-day curing period were 44.8, 43.0, and 42.5 MPa for pumice, RHA, and scoria, respectively (Figure 9(c)).With regard to fineness, the ultimate mean strength at 28 days for scoria samples with the fineness of 575 m 2 /kg was 52.2 MPa compared to 43.1 and 44.6 MPa for pumice and RHA with the fineness of 506 and 611 m 2 /kg, respectively (Figure 10).Natural scoria achieved high compressive strength probably due to latent hydraulicity, high reactive silica content, and good parking effects.

Conclusion
Scoria, pumice, and rice husk ash materials were successfully characterized according to ASTM C 618 criteria and tested to establish their potential as SCMs.Finely milled uncalcined samples of scoria and pumices and rice husk ash calcined in uncontrolled condition with specific surface area of 575, 506, and 611 m 2 /kg, respectively, showed high PAI of 82, 70.7, and 87.8% at 7 days and 89.0, 85.5, and 88.9% at 28 days in that order.These values meet the criteria for SCMs.Pozzolanic activity indices achieved by all scoria samples at 7 days were above the specified minimum of 75%.In comparison, rice husk ash achieved over 80% PAI after 7 days.Contrary, all pumice samples achieved minimum pozzolanic activity index after 28 days.Scoria and rice husk ash blended cement reached the ultimate compressive strength in 7 days while pumice blended cement achieved that in 28 days.The ultimate mean compressive strengths of 44.8, 43.0, and 42.5 MPa for pumice, RHA, and scoria, respectively, were achieved after 28-day curing period.In addition, setting time observed for all samples ranged from a minimum mean initial setting time of 80 min for S-N to the maximum mean initial setting time of 154 min for RHA 600.Likewise, the final setting time ranged from a minimum mean final setting time of 156 min for S 900 to maximum final setting time of 190 min for RHA 600.Initial and final setting times for scoria, pumice, and RHA were lower compared to those of Portland pozzolana cement (PPC), which were 166 and 285 min as initial setting time and final setting times in that order.On the other hand, soundness and shrinkage determined for all samples had a maximum mean value of 1.5 mm.The combination of all the above parameters demonstrated the suitability of natural scoria and pumice as SCMs.

Figure 4 :
Figure 4: Influence of materials on setting times: (a) initial setting time and (b) final setting time.

Figure 6 :
Figure 6: Effect of materials on (a) shrinkage and (b) soundness.

Table 1 :
Chemical compositions of scoria, pumice, and RHA from different literatures.

Table 2 :
Sample designation and notation.Natural scoria and pumice; # combustion in an uncontrolled condition.
[27]28,29]he compressive strength of cement mortar made of PLC blended with 10, 20, 30, and 40% of scoria, pumice, or RHA treated at different temperatures was measured according to SS-EN 196-1[27].Normally, this test uses cement mortar prism of 40 × 40 × 160 mm or 70.7 mm cubes.Compression test of cement mortar is also measured on cubes of 50 mm with cement/sand ratio 1 : 2.75 and water/cement ratio 0.485[26,28,29].In this research, SS-EN 196-1[27]was adopted using 40 mm cubes.Three test and reference cubes were cast and crushed at curing periods of 2, 7, 28, 56, and 90 days using compression machines conforming to SS-EN 196-1[27].The loading rate was adjusted to a range between 12 and 24 kN/min.The results are as shown in Figures8 and 10.

Table 3 :
Temperature effects on chemical composition and physical properties of the materials used.

Table 4 :
The influence of materials, temperature, and chemical composition on initial setting times, final setting times, shrinkage, and soundness.