Mauritius: A study of the potential use of Mauritian bagasse ash in concrete*

* By B. K. Baguant and G.T.G. Mohamedbhai, Department of Civil Engineering, University of Mauritius

This paper was presented to the seminar on Local Building Materials for Housing, Third International seminar of the African Network of Scientific and Technological Institutions (ANSTI), civil engineering subnetwork, held at the university of Mauritius, Reduit, March 1990. ANSTI is a UNESCO-sponsored Network.

Abstract

Bagasse is the fibrous residue of sugarcane after crushing and extraction of juice. Bagasse ash is the waste product of the combustion of bagasse for energy in sugar factories. An estimated 20,000 tons of bagasse ash are produced every year in Mauritius, of which about 11,000 tons are potentially usable.

This paper investigates the technical feasibility of using bagasse ash as a partial replacement of cement in concrete (i.e., as a pozzolana), and as a fine aggregate in concrete. The low chemical reactivity of the ash is demonstrated. Its use as sand in concrete shows no adverse effects on strength or strength development up to one year. However, the longer-term properties of the concrete, such as shrinkage and durability need to be further investigated.

Introduction

The economy of Mauritius, a small, 1800 km², Indian Ocean island, has been and is still very much dependent on the cultivation of sugarcane. The production of sugarcane varies from year to year, the main factor for such variations being climatic conditions, the average annual yield being about 5.6 million tons (1). Bagasse, the residue of sugarcane after crushing and extraction of juice, consists of water (about 50 per cent), fibres (above 48 per cent) and relatively small amounts of soluble solids. Nearly all bagasse produced in Mauritius is burnt for energy needed for sugar processing. The surplus energy is convened into electricity. Only a very small amount (about 3750 tons per year) is used for making resin-bonded boards.

Bagasse ash is at present considered to be a waste product with little or no use. It has negative value, in that, the sugar factories have to spend money to dispose of it. Moreover, it is a potential environmental pollutant. It is estimated that about 20,000 tons of bagasse ash are produced every year. This represents about 0.3 per cent of cane crushed or about 2.8 per cent of the dry weight of bagasse.

Since all cement used in Mauritius is imported, the prime use of bagasse ash in concrete would be as pozzolana, that is, as a partial replacement for cement. Another possibility is its use as a fine aggregate, even though the quantities available are small compared with the national annual consumption of about 1.2 million tons of fine aggregate. The latter requirement is currently being satisfied by crushing basalt rock and quarrying in dwindling reserves of natural coral sand.

Although Mauritius is sparing little effort in developing its industrial sector, its economy will continue to depend heavily on sugar production for many years to come. The sugar factories on their side, will continue to depend entirely on bagasse burning for energy. The supply of bagasse ash is, therefore, ensured for years ahead.

Characteristics of bagasse ash

Types and availability

The sugar factories in Mauritius use two types of bagasse-fired water-tube boilers: a hearth-type and a suspension-burning type (2). In the former, bagasse fed in at the top of the furnace falls in a pile on the floor of the hearth, while cold or hot air is blown on to the burning bagasse. This type of burning produces three types of ash:

(a) Furnace bottom ash, which accumulates on the floor of the hearth and is removed manually at every weekend shutdown.

(b) Hopper ash, which is blown by horizontal air blasts into hoppers situated at the back of the furnace.

(c) Fly ash, which is carried by the exhaust gases up the chimney and is partially removed by sprays of water.

In the suspension-burning type, bagasse fed in from the sides is distributed by a spreader-stoker on to the plan area of the combustion chamber, while primary air is blown upwards through a grate enabling the bagasse to burn in suspension in air without heaps forming on the grate. This type of furnace also produces three types of ash:

(a) Grate ash, which accumulates on the grate and is removed by tilting the grate every six to eight hours.

(b) Hopper ash, which is blown into hoppers by secondary air fed in from the sides of the furnace.

(c) Fly ash, which is carried up the chimney.

Of the estimated 20,000 tons of bagasse ash produced every year, about 4000 tons are fly ash (2). The remainder consists mainly of grate ash (about 11,000 tons).

Table 1. Chemical composition of bagasse ash (excluding fly ash) from different sources

Constituents	Percentage composition
	Ref. 9	Ref. 10	Ref. 11	Ref. 2	Dundee, Hopper ash	Dundee grate ash
SiO2	74.2	75.0	71.4	73.07	50.19	57.61
Al2O3	12.12	3.6	1.6	12.00	10.12	10.11
Fe2O3	12.12	2.7	1.9	12.00	15.75	10.79
CaO	3.93	3.9	4.3	4.3	4.94	4.58
MgO	0.32	4.1	3.8	2.66	8.14	5.31
K2O	1.67	7.1	11.0	2.65	5.56	6.22
Na2O	0.36	0.2	0.8	0.16	0.41	0.29
P2O5	5.58	2.3	3.4	2.57	2.09	2.38
SO3	0.20	0.3	-	0.23	-	-
Cl2	0.07	-	-	0.01	-	-
MnO	-	-	-	-	0.26	0.29
TiO2	-	-	-	-	1.57	1.46
L.O.I*	1.63	1.65	-	1.30	1.23	1.24

* Loss-on-ignition

Chemical composition

Results of chemical tests on grate and hopper ashes carried out in Mauritius and at the University of Dundee are shown in table 1. The local data were obtained using classical laboratory techniques whereas the Dundee results were obtained by an X-ray fluorescence (XRF) method of oxide analysis.

The results indicate a fairly high degree of variability in chemical composition of the ashes, with the Dundee data showing marked departures from typical results. But, in general, the grate and hopper ashes have a high silica content, which, if present in an amorphous chemically reactive form, could enable the ashes to exhibit pozzolanic properties. Also, being alkaline, the ashes are compatible for mixing with Portland cement.

The grate and hopper ashes have reasonably low loss-on-ignition values (see table 2), with the exception of ashes from fuel where the factory burns pulverized coal in addition to bagasse for energy production. Fly ash has a high carbon content. It is also soft, compressible and highly absorptive. It is, therefore, not suitable for inclusion in concrete.

Table 2. Loss-on-ignition (L.O.I) values of bagasse ashes from various sources

	Type of ash
Factory	Grate	Hopper	Fly
Mon Desert Alma	3±0.5	1±0.3	71±3
Belle Vue	4±0.3	1±0.2	-
F.U.E.L.	14±1	4±0.3	80±2
Rose-Belle	3±0.5	1±0.2	7±0.5

Table 3. Specifications of BS 3892 for pulverized fuel ash (pfa) for use in concrete

Property	Limit (percentage)
Fineness (expressed as the mass proportion retained on a 45 micron sieve)	Not greater than 12.5
Loss-on-ignition	Not greater than 7
Water requirement of a mixture of pfa and ordinary	Not greater than 95 of that of OPC
Portland cement (OPC)
Pozzolanic activity index	Not less than 85

Physical properties

The grate and hopper ashes are granular, rough, vascular particles whose maximum sizes can vary extensively as the material forms lumps in the furnace. The relative density of the ashes on a saturated surface dry basis range between 1.90 and 2.12. The ashes also have very high absorption values of 10 ± 2 per cent.

The grading of the ashes can be extremely variable, the percentage passing 600 um sieve ranging from 14 per cent to 80 per cent. But, in general, the ashes tend to be rather coarse, with 15-25 per cent passing the 600 µm sieve (2,3).

Pozzolanic activity

There is no national or international standard for testing bagasse ash. The specifications of BS 3892 (4) for pulverized fuel ash (pfa) for use in concrete (see table 3) were therefore, used to establish whether bagasse ash could be an effective pozzolanic material for use in concrete.

Samples of grate ash and hopper ash obtained from one factory were used in all the tests. Fly ash was excluded because of its high carbon content The ash samples were oven-dried, ground in a ball-mill and sieved through a 45 µm test sieve. The samples thus automatically satisfied the fineness requirement. The loss-on-ignition values for grate ash varied between 1.4 and 1.7 per cent and those for hopper ash between 0 and 0.2 per cent. These values are well below the 7 per cent specified in BS 3892.

The water requirement of grate ash (110 per cent) and hopper ash (104 per cent) were greater than the 95 per cent specified in BS 3892, probably because of greater friction between the particles of ash compared with pfa (5). The greater angularity of grate ash particles compared with hopper ash is reflected in its higher water requirement to give the same flow characteristics as that of hopper-ash mortar.

The pozzolanic activity index values for grate-ash and hopper-ash mortars were found to be 68 per cent and 65 per cent, respectively, which are much lower than the specified minimum of 85 per cent in BS 3 892, showing the relatively poor reactivities of the ashes. This is not too surprising in view of the relative absence of cenospheres revealed by the scanning electron microscope (SEM) photographs. However, many pozzolanas are known to exhibit low reactivities at early ages and yet develop significantly higher strengths at later ages. It was therefore, decided to investigate the strength development with time of concrete specimens incorporating ground grate-and hopper-ashes. The mix proportions of the normal, grate-and hopper-ash concretes (see table 4) were such that the cement content was kept constant and the natural sand was reduced by an amount equal to the volume of ash added. The strength results of the specimens tested over a period of one year are shown in figure 1 and in table 5.

Table 4. Mix proportion of normal, hopper- and grate-ash concrete

Concrete mix	Normal	Hopper ash	Grate ash
Cement (kg/m3)	320	320	320
20 mm gravel (kg/m3)	744	744	744
10 mm gravel (kg/m3)	372	372	372
Natural sand (kg/m3)	744	558	558
Ash (kg/m3)	-	95	60
Water (kg/m3)	200	200	200
Total (kg/m3)	2380	2289	2254
Slump (mm)	85	50	70

Table 5. Strength development with time

	Compressive strength
Age (days)	Normal concrete		Hopper-ash concrete		Grate-ash concrete
	N/mm2	Percentage	N/mm2	Percentage	N/mm2	Percentage
3	19.9	48.3	21.5	48.5	20.4	46.9
7	30.3	73.5	30.9	69.7	29.9	68.7
28	41.2	100.0	44.3	100.0	43.5	100.0
90	45.4	110.2	50.1	113.1	47.8	109.9
180	45.2	109.7	50.7	114.4	48.4	111.3
365	45.9	111.4	51.5	116.2	51.5	118.4

Figure 1. Strength development as a percentage of 28-day strength

The strengths of both grate- and hopper-ash mixes were comparable with those of the normal mix at all corresponding ages. There was no evidence of any significant additional strength development due to pozzolanic activity. These results confirm the poor reactivity of the ashes and make it most unlikely that bagasse ash in Mauritius will ever be used as pozzolana in concrete.

Bagasse ash as sand replacement in concrete

Characteristics of ash selected

Grate ash is granular requiring no grinding for use as sand, whereas hopper ash usually contains clinkery lumps which must be ground. Also, because grate ash is available in larger quantities than hopper ash, it was decided to use grate ash in this series of investigations.

Table 6. Series 1: Concrete mixes with basalt sand

Mix	Free water/cement ratio	Cement content (kg/m3)	Basalt sand content (kg/m3)	Coarse aggregate content (kg/m3)
1.1	0.41	561	747	912
1.2	0.48	480	800	940
1.3	0.59	390	878	952
1.4	0.67	343	939	939
1.5	0.84	274	1031	915

Table 7. Series 2: Concrete mixes with grate ash

Mix	Free water/cement ratio	Cement content (kg/m3)	Grate ash content (kg/m3)	Coarse aggregate content (kg/m3)
2.1	0.41	561	498	912
2.2	0.48	480	553	940
2.3	0.59	390	585	952
2.4	0.67	343	626	939
2.5	0.84	274	688	915

The ash was oven-dried for 24 hours and then passed through a BS 4.75 nun standard test sieve to remove the coarser particles. Tests were carried out on grate ash and crushed basalt sand to determine their grading, relative density, and water absorption in accordance with BS 812 (6).

The grading of grate ash was coarser (20 per cent passing a 600 um sieve) than that of basalt sand (41 per cent passing a 600 um sieve). The relative density of the ash (2.00) was about two thirds that of basalt sand and the replacement of the sand by the ash was carried out by volume rather than by weight in order not to affect the volume stability of the concrete. The absorption value of the ash (10 ± 2 per cent) was very high compared with that of basalt sand (0.40 ± 0.01 per cent).

Compressive strength tests

Three series of tests were carried out.

Series 1: These tests consisted of five concrete mixes prepared with ordinary Portland cement, basalt sand and 12 nun maximum size basalt aggregate. These mixes were designed in accordance with the United Kingdom Department of the Environment (DOE) (7) design procedure for normal concrete mixes so as to obtain a range of 28-day strengths varying between 20 and 60 N/mm² and a slump of 30-60 mm. The aggregates used were oven-dried and the actual amount of water added to each of the five mixes was the sum of the free water and the absorption of the aggregates. The cement content of the mixes ranged from 274 kg/m³ to 561 kg/m³ and the corresponding free water/cement ratio varied from 0.84 to 0.41 (see table 6).

Table 8. Series 3: Concrete mixes with grate ash and superplasticizer

Mix	Free water/Cement ratio	cement content (kg/m3)	Grate ash content (kg/m3)	Coarse aggregate content (kg/m3)	Superplasticizer content (kg/m3)
3.1	0.29	561	498	912	17
3.2	0.34	480	533	940	14
3.3	0.41	390	585	952	12
3.4	0.47	343	626	939	10
3.5	0.58	274	688	915	8

Table 9. Compressive strength results of mixes in the three series, 28-day cube strength (N/mm²)

Mix	Series 1 (basalt sand)		Series 2 (grates ash)		Series 3 (grate ash + plasticizer)
1	(0.41)	62	(0.41)	33	(0.29)	70
2	(0.48)	53	(0.48)	28	(0.34)	65
3	(0.59)	42	(0.59)	21	(0.41)	55
4	(0.67)	34	(0.67)	14	(0.47)	45
5	(0.84)	24	(0.84)	8	(0.58)	35

(Figures in brackets indicate the free water/cement ratios)

Table 10. Average initial absorption

	Average initial surface absorption (ml/m2/s)
Time (Minutes)	Series 1 (basalt sand)	Series 2 (grate ash)	Series 3 (grate ash + plasticizer)
10	0.150	0.143	0.045
30	0.117	0.073	0.022
60	0.070	0.058	0.015
120	0.067	0.033	0.010

Series 2: In these test specimens, basalt sand in each of the five mixes was totally replaced, by volume, by grate ash (see table 7). The free water/cement ratios were kept the same but the total amount of water for each mix was adjusted to take into account the higher absorption value of grate ash. Nevertheless, a reduced slump of 0-10 mm was obtained.

Series 3: A water-reducing superplasticizer (Melment L10 - a modified condensate product of melamine and formaldehyde) was added to the mixes of series 2. The cement, ash and coarse aggregate contents of series 2 and series 3 mixes were, thus, the same (see table 8). Preliminary tests were carried out to determine the optimum dosage of superplasticizer to be used, that is, the amount of superplasticizer for which no further significant water reduction was observed. This was found to be about 3 per cent by weight of cement, and this was the amount used in each of the five mixes of series 3, causing a 22 per cent reduction in the total amount of water used, but again giving a slump of 0-10 mm. Three 100-mm cubes were prepared from each of the five mixes in each of the three series. The cubes were stored in water and tested in compression at the age of 28 days. The results are shown in table 9 and figure 2.

Shrinkage

Figure 2. Change of strength with change in cement content

Figure 3. Shrinkage - time graph

Shrinkage tests were carried out using standard 300 × 75 × 75 mm prisms, in accordance with BS 1881 (8), on mix 3 of the three series of concrete described earlier. After storing the specimens in water for seven days, they were removed and shrinkage readings were started. Shrinkage was recorded over a period of 49 days during which the specimens were continuously stored in air.

The results (see figure 3) showed that shrinkage increased with age in all the three series but the increase was more marked in grate-ash concrete (with or without superplasticizer) (3).

Although both concrete mixes had the same free water contents, the total water contents in the mixes were significantly different because of the higher absorption of the grate ash. This could have been responsible for the greater drying shrinkage in the ash concrete. The superplasticizer caused only a slight reduction (about 12.5 per cent) in the shrinkage of the ash concrete.

Permeability

Permeability of concrete, that is the measure of the ability of liquids or gases to penetrate the material, is an indirect measure of its durability.

There are various methods of measuring permeability and the one used in this study was the initial surface absorption test (ISAT) described in BS 1881. The test measures the rate of flow of water, under a small constant head, into oven-dried concrete per unit surface area after a stated interval of time from the beginning of the test. The initial surface absorption of concrete decreases with time until no more water is absorbed and the surface has reached saturation.

Tests were carried out on concrete cubes of mix 3 in each of the three series of concrete. Three cubes, at the age of 28 days, were tested in each series. The average initial surface absorption values of the three series are shown in table 10. (3)

Discussion

The variability in the chemical and physical characteristics of bagasse ash is an important consideration in the evaluation of the material for potential use in concrete. Hopper ash and grate ash are fairly similar in chemical composition, whereas fly ash has a very high carbon content. Grate ash is fairly granular, which makes it quite suitable for use as a fine aggregate in concrete. Hopper ash tends to form clinkery lumps which would need grinding or crushing before incorporation in a concrete mix. Fly ash is soft, compressible and highly absorptive, which makes it quite unsuitable for mixing into concrete. Quantities also favour the use of grate ash in concrete in preference to the other ashes.

Although both grate ash and hopper ash satisfy the requirements of BS 3892 for loss-on-ignition, silica, magnesia and sulphuric anhydride contents, they fail to meet the specification for water requirement and pozzolanic activity index. Moreover, the ashes require grinding before they can comply with the fineness requirements.

There appears to be no adverse effect on the strength or strength development of concrete incorporating bagasse ash over a period of up to one year. Further tests are required to assess the longer-term effects, in particular the shrinkage and durability properties.

Grate-ash concrete has significantly lower strength than the corresponding basalt sand concrete. The 28-day strength of the ash concrete varies from about half to one third of the equivalent control mix. However, the addition of a superplasticizer to the grate-ash concrete causes a considerable increase in strength. With the latter concrete, strengths about three times those of unplasticized ash concrete and about 35 per cent higher than those of basalt-sand concrete are obtained. Also, the plasticized grate-ash concrete appears to reach a strength limit at about 70 N/mm².

In the fresh state, the ash-concrete mixes show lower workability. The slump decreases from 30-60 mm in the basalt-sand concrete to 0-10 mm in the grate-ash concrete (with or without superplasticizer). This is due to the coarse grading of the ash and to the rough angular nature of the grate-ash particles which increase aggregate friction and interlock. This results in poor cohesion of the fresh concrete, the workability of which hardly improves even with further addition of water.

Conclusions

Bagasse ash in Mauritius shows too poor reactivity with Portland cement to make it an effective pozzolana in concrete.

Bagasse ash can be effectively used as a fine aggregate in concrete to produce a range of compressive strengths up to about 70 N/mm². However, generally reduced workability is observed in the fresh concrete, due to the coarse grading and angular texture of the ash particles.

The shrinkage of bagasse-ash concrete is not excessively higher than that of normal concrete. However more data are required to assess this property of the concrete.

The initial surface absorption characteristics of bagasse-ash concrete do not give any indication of low durability in comparison with basalt-sand concrete.

The addition of a superplasticizer significantly reduces the initial absorption capacity of bagasse-ash concrete.

Acknowledgement

The invaluable contribution of Dr. R.K. Dhir, Reader in Civil Engineering, Department of Civil Engineering, University of Dundee, in connection with this project is gratefully acknowledged.

References

1. Mohamedbhai, G.T.G., and Baguant, B.K., “Possibility of using bagasse ash and other furnace residues as partial substitute for cement in Mauritius”, Revue agricole et sucri de l'Ile Maurice, vol. 64, no. 3, September-December 1985.

2. Li Pin Yuen, L.Y.N.G., “Availability of bagasse ash and its characteristics in Mauritius”, B. Tech. final year project (Reduit, University of Mauritius, 1986).

3. Kim Currun, G.S., “The effects of plasticizer on bagasse ash concrete”, B. Tech. final year project (Reduit, University of Mauritius, 1988).

4. British Standard, BS 3892: 1982, Pulverised fuel ash.

5. Choollun, V.K., “Pozzolanic activity of bagasse ash and a laterite”, B. Tech. final year project (University of Mauritius, 1988).

6. British Standard, BS 812: 1975, Methods for sampling and testing of mineral aggregates, sands and fillers.

7. Department of the Environment, Design of Normal Concrete Mixes (Garston, Building Research Establishment, Transport and Road Research Laboratory, 1975).

8. British Standard, BS 1881: 1990. Methods of testing concrete.

9. Uteene, A.F., “Investigation into the properties of bagasse ash cement”, B. Tech. final year project (Reduit, University of Mauritius, 1980).

10. Cementia Engineering and Consulting Ltd., A Report on Tests Carried out on Bagasse Ash from Mauritius, 1978.

11. Paturau, J.M., By-products of the Cane Sugar Industry. An Introduction to Their Industrial Utilisation, Sugar Series, 3 (Elsevier Scientific Publishing Company).