Effects of Copper Slag as a Replacement for Fine Aggregate on the Behavior and Ultimate Strength of Reinforced Concrete Slender Columns

Use of copper slag (CS) as a replacement for fine aggregate (FA) in RC slender columns was experimentally investigated in this study. Twenty columns measuring 150 mm x150 mm x 2500 mm were tested for monotonic axial compression load until failure. The concrete mixture included ordinary Portland cement (OPC) cement, fine aggregate, 10 mm coarse aggregate, and CS. The cpercentage of cement, water and coarse aggregate were kept constant within the mixture, while the percentage of CS as a replacement for fine aggregate varied from 0 to 100%. Four 8 mm diameter high yield steel and 6 mm mild steel bars were used as longitudinal and transverse reinforcement, respectively. Five cubes measuring 100 mm x100 mm x100 mm, eight cylinders measuring 150 mm x 300 mm and five prisms measuring 100 mm x 100 mm x 500 mm were cast and tested for each mixture to determine the compressive and tensile strengths of the concrete. The results showed that the replacement of up to 40% of the fine aggregate with CS caused no major changes in concrete strength, column failure load, or measured flexural stiffness (EI). Further increasing the percentage reduced the concrete strength, column failure load, and flexural stiffness (EI), and increased concrete slump and lateral and vertical deflections of the column. The maximum difference in concrete strength between the mixes of 0% CS and 100% CS was 29%, with the difference between the measured/control failure loads between the columns with 0 and 100% CS was 20% the maximum difference in the measured EI between the columns with 0 and 100% CS was 25%. The measured to calculated failure loads of all specimen varied between 91 and -100.02%. The measured steel strains were proportional to the failure loads. It was noted that columns with high percentages of CS ( 60%) experienced buckling at earlier stages of loading than those with lower percentages of CS.


Introduction
Aggregate is the main constituent of concrete, occupying more than 70% of the concrete matrix.In many countries, there is a scarcity of natural aggregate that is suitable for construction, whereas in other countries the consumption of aggregate has increased in recent years, due to increases in the construction industry.In order to reduce depletion of natural aggregate due to construction, artificially manufactured aggregate and some industrial waste materials can be used as alternatives.Copper slag (CS), the glassy material, produced during matte smelting and copper conversion was previously considered waste and disposed as landfill.It has been estimated that for every ton of copper production about 2.2-3 tons of slag are generated.Slags containing < 0.8% copper are either discarded as waste or sold cheaply (Shi et al. 2008;Gorai et al. 2003).
Processed, air-cooled, and granulated CS has a number of favorable mechanical properties for aggregate use, including excellent soundness characteristics and good abrasion resistance (Queneau et al. 1991).Alter (2005) studied the effect of CS on the environment on the basis of basel convention and characterized it as non-hazardous.Shanmuganathan et al. (2008) carried out toxicity characterization and longterm stability studies on CS and reported that the slag samples are non-toxic and pose no environmental hazard additionally poor leachability of the slag metals assure long-term stability, even in extreme climates.
The tests indicate that the heavy metals present in the slag are stable and are not likely to dissolve significantly even through repetitive leaching under acid rain.
Utilization of CS for applications such as replacement for fine aggregate (FA) in concrete has the dual benefit of eliminating the costs of disposal and lowering the cost of the concrete in some countries.Slight variations in prosperities of the slag may be anticipated depending on source.Taha et al. 2007 explored the use of CS as a replacement in cement in controlled low-strength concrete and reported that mixes which were designed using waste material (i.e., CS), cement, and fine aggregate yielded higher strength values than those mixes where the waste material was used as a full replacement for cement.Shoya et al. (1999) found no major differences in concrete compressive strength due to the use of CS as a replacement for FA.Resende et al. (2008) reported a small reduction in concrete compressive and flexural strengths due to substitution of CS.Khanzadi and Behnood (2009) also explored the possibility of using CS as a replacement for coarse aggregate and reported an improvement in the mechanical properties of the high strength concrete mixture.Al-Jabri et al. (2009a) investigated the use of CS as a replacement for sand in high performance concrete (HPC) with a constant water content.They reported that an addition of up to 50% CS for sand yielded comparable strength to that of a control mixture with no CS.Further addition of CS caused a reduction in strength.Al-Jabri et al. (2009b) studied the use of CS as a replacement for sand in highstrength concrete (HSC) with almost constant workability.The water content was adjusted in each mixture in order to achieve the similar workability as for the control mixture.They noted a remarkable 22% reduction in water demand when they replaced 100% of the sand with CS compared to a control mixture of 0% CS.They reported an increase in the concrete strength due to the increased content of CS as replacement of FA, while super-plasticizer was found to be a very important ingredient in HSC made with CS in order to provide good workability and better consistency.Recently (Wu et al. 2010) recommended that less than 40% CS as sand replacement can achieve a high strength concrete that is comparable or better than a control mix.
No test results were found in the literature on the use of CS as a replacement for FA in structural members.In this research, the CS was used as a replacement for FA in twenty reinforced concrete (RC) slender columns.The objective was to investigate the effect of partial and full replacement of FA with CS on the strength and behavior of columns.The column cross section was 150 mm x150 mm and its length was 2500 mm.The columns were divided into six groups based on the percentage of CS used as a replacement for FA as follows: 0% CS, 20% CS, 40% CS, 60% CS, 80% CS and 100% CS.The water and cement contents as well as the coarse aggregate (10 mm) were kept constant for all specimens.At least three columns were tested from the same group.The results were judged based on the failure load, lateral and vertical deflections, steel strain, concrete compressive and tensile strengths, effects on EI and slump values.

Materials Used
OPC produced by Oman Cement Company, natural fine and 10 mm coarse aggregates from a nearby crusher, and fresh tap water were used in the concrete mixture.The CS was brought from Oman Mining Company, which produces an average of 60,000 tons annually (Taha et al. 2007).Table 1 shows the chemical and physical properties of the CS and cement used, and Figure 1 shows sieve analysis of the FA and CS.It is clear that both fit within the grading limits of zone 1 of the (Omani standard for fine aggregate (OS-2, 1982).
However, many more particles of fine aggregate than CS passed through sieves from 0.1 to 1 mm. Figure 2 shows a picture of the glassy surface CS  C192, 1998) for about 4 minutes before casting.A vibration table was used for the compaction of the samples, while the column specimen was compacted using a special concrete vibrator.For each mix five cubes measuring 100 mm x100 mm x 100 mm (compression), eight cylinders measuring 150 mm x 300 mm (3 for compression and 5 for splitting), and five prisms measuring 100 mm x 100 mm x 500 mm (flexural) were cast.A slump value for the fresh concrete of each mixture was taken as the average of three samples in accordance with British Standards Institute (BS1) 881: part 102 (1983).Curing of the columns and samples was carried out for one week using wet Hessian cloth and then the samples were placed under room temperature.In addition, three cubes of 150 mm x150 mm x150 mm, three cylinders of 150 mm x 300 mm and three prisms of 100 mm x 100 mm x 500 mm Table 1.Chemical composition and physical properties of copper slag and cement produced in Oman

Specimen Construction and Preparation for Testing
Figure 3 shows the steel strain gauges' labelling method.For each longitudinal bar, two strain gauges (A, B) were stuck opposite one another at the mid-span of the column.The letter L identifies the location of the strain gauge, (i.e., L2A means the number of the strain gauge which is stuck on longitudinal bar number 2 at location A).Two stirrups, one 75 mm below and one 75 mm above the mid-span of the column were strain gauged.Four strain gauges were stuck on each stirrup with one in each face of the column.The letters S and B/T were used to differentiate between the strain gauges (i.e., ST1 means the strain gauge in the upper stirrup on face 1 of the column and SB4 means the strain gauge which is in the lower stirrup at face 4 of the column).For measuring lateral deflection of the beam, a linear variable differential transducer (LVDT) was installed on each face at the mid-span of the column as shown in Figure 4.The concrete surface strains were measured using 100 mm horizontal and 200 mm vertical Demec gauges at mid-span.
After installing the steel strain gauges, the steel cage was erected and inserted into a wooden mould.The mould was laid horizontal on a levelled surface and its walls were perfectly vertical to ensure straightness of the specimen and minimize the imperfection.Strain gauges were numbered and casting of the specimen and the samples was carried out.After curing under wet Hessian cloths for one week, the specimen and the samples were left under room temperature for about four weeks before testing.
The column was painted white and the Demec pins were fixed.The column was installed on a 5000 kN Dartec universal testing machine using steel plate caps and a steel ball at the center of the plate at each end to ensure the application of a compressive axial load.Figure 5 shows the cap and ball system used, while Figure 6 shows a typical column installed on a testing machine.The strain gauges and LVDTs were connected to a data logger while the Demec readings were taken manually.

Testing Procedures
The compressive axial load on the column was applied in increments of 50 kN for the first three increments followed by reduced increments of 20 kN till the load reached 210 kN, and then by 10 kN increments until failure.The rate of loading was 1 kN/second.To allow for stable deformation to take place after each load increment, an interval of about one minute was used before recording the readings.The applied load, the steel strain, and the defection readings were directly recorded by a data acquisition system while the DEMEC readings were taken manually.The cube, cylinder, and prism samples were tested in the same day that the column was tested.The cube unconfined compressive strength was tested in accordance with  BS1881 part 116 (1983, with a rate of loading of 2.5 kN/seconds while the cylinder compressive strength was obtained using the ASTM C39 (1986) with a rate of loading of 2 kN/second.For testing the tensile strength, the average strength of the five prisms from each mixture was considered using BS1881 part 118 (1983) with a rate of loading of 0.2 kN/second while the average of splitting strengths of five cylinders was recorded using BS 1881 part 117 (1983) with a rate of loading of 4.4 kN/second.

Test Results and Discussion
The slender column behavior is controlled by material strength, support conditions, type of loading, and  the slenderness ratio.The material effect in this research was evaluated by CS as a replacement for FA.The percentage of replacement varied from 0 to 100%.The reinforcement type, strength, and arrangement were kept the same for all specimens.The cement, coarse aggregate, water types, and quantities were kept constant in all specimen tested.The support condition was maintained to be pinned by introducing steel balls at both ends in all tested columns.The loading was maintained to be concentrically axial as the steel balls guaranteed that no applied moment or shear force were introduced.The slenderness ratio, kl/r, was kept constant as the column length, square cross-section, supporting condition, and load arrangement were maintained to be the same in all specimen.Unlike short columns, the slender columns' failure was usually controlled by buckling.The buckling load decreased as the slenderness ratio increased.Since, in our case, the column is hinged, the k value is equal to one and the effective length is equal to the actual length of the column.

Strength of Materials
Table 3 shows the average slump values, density of concrete, cube and cylinder compressive strengths, the cylinder splitting and prism flexural tensile strengths.
It also gives the average tensile yield strength of the reinforcement used.It is clear that minor reduction in the concrete compressive and tensile strengths were reported due to the increase of CS as replacement for FA.The increase in CS also led to higher slump and the mixture reached collapse in the 80 and 100% lavels of CS.This is due to the fact that the glassy surface of CS absorbs less water (0.17%) than FA.This excess water formed internal voids, which led to concrete bleeding and contributed to the increase of porosity that weakened the concrete strength.A microscopic study made by Wu et al. (2010) showed that with the increase of CS content in the concrete mixture the voids increased.It is also possible that the porosity resulted from the reduction in the content of fine particles, which led to a reduction in the interlocking effect in the concrete matrix.The sieve analysis of CS and FA presented in Figure 1 proves that many more 0.1 to 1 mm particles of FA than CS passed through the sieves.The result is a matrix with a higher percentage of voids than the control mixture (0% CS) and weaker bounds between the coarse aggregate particles.This problem cannot be detected by only studying the values of concrete density recorded in Table 3, due to the fact that CS possess a much higher density (about 3.5 ton/m 3 ) and specific gravity (3.45) than the densi- ty (1.4 -1.5 ton/m 3 ) and specific gravity (2.77) of the FA.In other words, although the concrete density increases due to the increase in CS content, the porosity also increases, which adversely effects the concrete strength.

Failure Load of Specimen
Table 4 shows the measured failure loads.The column was considered collapsed when it could resist no more loads.This was preceded by buckling of the column at mid-height into a half-sine wave shape in all cases.The measured failure load is the maximum load recorded before the column fails.It is clear from Figure 7 that the presence of up to 40% of CS as a replacement for FA resulted in no major change in the load carrying capacity of the column compared to the measured failure load of the control specimen (0% CS).Gradual drops in the load carrying capacity were observed for the columns having 60% CS and more.The maximum difference in the measured/control failure load between the group with 40% CS and that of the larger CS was recorded in the columns with 100% CS as 20%.It was noticed during testing that a large presence of CS resulted in earlier buckling of the column, which led to an earlier failure load than when the percentage of CS was low.The failure of the 0% CS column was more sudden with less time between the start of buckling and failure load than columns with a high percentage of CS.

Steel Strain
Table 6 shows the measured to yield strain ratios / y in the longitudinal steel versus the ratios of measured to control loads.The measured strain reported in this table is the strain recorded for the load increment just before failure.In general, more strains were recorded in the columns with a CS content 40% than those of higher CS contents.This is due to the fact that columns with a CS content 40% have resisted more loads and therefore the steel was subjected to higher stresses than the columns with higher CS.The measured to yield longitudinal steel strain / y versus the measured to control load ratios of the members of each group behaved almost in a similar fashion.Figure 10 shows the effect of CS content on the longitudinal steel strains for a typical member of each group.It is clear that as the percentage of copper slag increases the column stiffness decreases (more strain for same load).
As to be expected, the strain in the stirrups was negligible and no trend could be observed, therefore, it was not reported.

Lateral Deflection
Table 7 shows the maximum measured displacement values in the front and rear faces of the twenty columns tested with different percentages of CS as replacement for FA.The positive values indicate extension.The front face is the face of the column that was on the convex side (negative values) while the rear face is on the concave side (positive values).It should be mentioned that, for the purpose of safety, the LVDT readers were removed immediately after the signs of buckling, which means few final readings were recorded.It was noted that all columns' buckling occurred at the columns' midpoints, as can be seen in Figure 11.The measured lateral displacement versus the measured load to control load ratios of the members of each group behaved almost in a similar fashion.Figure 12 shows typical measured/control load ratios versus lateral displacement in the tested columns.In general, there was a gradual increase in displacement with the increase of CS, which indicates gradual loss in stiffness due to increased CS.The drops were more pronounced in the columns with a large percentage of CS ( 60%).This indicates that a high percentage of CS as a replacement for fine aggregate leads to more

Vertical Deflection
Table 8 shows the maximum measured vertical deflection values one increment before collapse in the twenty columns tested with different CS percentages.
Although the columns with a CS percentage 40% resisted higher loads, they experienced smaller values of vertical deflections than CS = 60%.The comparatively small values of deflections recorded in the CS = 80% and CS = 100% groups are due to the fact that these columns experienced premature buckling and loss of strength due to the large CS content.The measured to vertical displacement versus the measured load to control the load-ratios of the members of each group behaved almost in a similar fashion.Figure 13 shows a typical load to control load ratios versus vertical displacement in the tested columns.These results emphasize the findings recorded in the lateral deflection of gradual loss in stiffness with the increase of CS content.

Conclusion and Recommendation
The use of CS as a replacement for FA is environmentally helpful due to the reduction in the waste produced from the copper manufacturing process.It also contributes to conservation of natural FA.Twenty RC slender columns measuring 150 mm x 150 mm x 2500 mm with different percentages of CS as a replacement for FA were tested in this research.The contents of cement, water, and coarse aggregate were kept constant while the percentages of CS as a replacement for FA varied from 0 to 100%.Results showed that replacement of up to 40% of FA with CS caused no major changes in column failure load, EI or concrete strength.Further increasing the ratio of CS ro FA reduced the concrete strength and column failure load, and increased concrete slump and lateral and vertical deflections.The maximum difference in measured to control failure load between the group with a CS 40% and that of the larger CS was recorded in the columns 100% CS as 20%.The measured steel strains were proportional to the failure loads.It was noticed that columns with high percentages of CS ( 60%) experienced buckling at earlier stages of loading than those of low percentages of CS.
It is possible that the reduction in strength resulting from increasing CS is due to increased voids due to the fact that CS possesses fewer fine particles than FA .It could also be due to the increase of the free water because the CS absorbs less water than the FA.
It is recommended that the effect of CS change on total void volume and amount of free water content be studied separately.

Figure 1 .
Figure 1.Fine aggregate and copper slag sieve analysis

Figure 4 .Figure 5 .
Figure 4. LVDT on each face at mid-span of the column

Figure 6 .
Figure 6.Typical column installed on teh 5000 kN DARTEC testing machine

Figure 12 .
Figure 12.Typical lateral deflection of tested columns Effects of Copper Slag as a Replacement for Fine Aggregate on the Behavior and Ultimate Strength of Reinforced Concrete

Table 3 . Average physical properties of concrete and steel used
Effects of Copper Slag as a Replacement for Fine Aggregate on the Behavior and Ultimate Strength of Reinforced ConcreteSlender Columns

Table 4 .
Measured failure loads of the tested column (kN)

Longitudinal strain ratio vs P measured / P ratio
Effects of Copper Slag as a Replacement for Fine Aggregate on the Behavior and Ultimate Strength of Reinforced ConcreteSlender Columns Typical failure mode and location of buckling of a column

Table 5 .
Effect of copper slag on P measured P calculated

Table
Average ratios of measured to yield strains y in the longitudinal steel AS Alnuaimi Al-Jabri KS, Hisada M, Al-Oraimi SK, Al-SaidyAH  (2009a), Copper slag as sand replacement for high performance concrete.Cement and Concrete Composites 31:483-488.Al-Jabri KS, Hisada M, Al-Saidy AH, Al-Oraimi SK (2009b), Performance of high strength concrete made with copper slag as fine aggregate.Construction and Building Materials 23:2132-2140.Alnuaimi AS (2009), Use of copper slag as replacement to fine aggregate in RC slender columns.The Fourth International Conference on Computational methods and experiments in material characterization, MC'09, New Forest, UK.Alter H (2005), The composition and environmental hazard of copper slags in the context of the basel Convention.Resources, Conservation and Recycling 43:353-360.ASTM C192 (1998), Standard practice for making and curing concrete test specimens in the laboratory.West Conshohocken, PA: international, USA.ASTM (1986), Test for compressive strength of cylindrical concrete specimens.ASTM, USA.BS 1881 (1983), Testing of concrete, Part 102: Method for determination of slump, British Standard Institution UK.BS 1881 (1983), Testing of concrete, Part 116: Method for determination of compressive strength of concrete cubes.British Standard Institution, BSI, UK.

Table 7 .
Average measured lateral displacement in the front and rear faces (mm)Table Average measured vertical displacements (mm)