Durability of concrete with Belite-Gehlenite clinker as fine aggregate

. The Japanese cement industry uses large quantities of industrial waste and by-products as raw materials in the production of cement clinker. Although the amount of industrial waste generated annually has remained almost constant, domestic demand for cement has been falling. In order to maintain the amount of waste re-used by the cement industry, there is a need to explore new ways of utilizing clinker besides in cement production. The proportion of waste used in the production of Belite-Gehlenite clinker featured in this study is about twice as much as in normal clinker. Previous studies have shown that when clinker is used as aggregate in mortar and concrete, clinker hydration products fill cracks as they occur for added self-healing performance. In this study, in addition to the basic characteristics of concrete containing Belite-Gehlenite clinker as fine aggregate, the resistance to cracking of specimens made with the concrete is investigated. Compared with concrete using natural sand, it is confirmed that compressive strength is improved, drying shrinkage is reduced, carbonation is suppressed, and freeze-thaw resistance is maintained. It is also demonstrated that resistance to cracking is improved.


INTRODUCTION
Large quantities of industrial waste and by-products are currently used by the cement industry in Japan as raw materials for the production of cement clinker.While the amount of such waste has neither increased nor decreased in recent years, domestic demand for cement has been falling, as shown in Figure 1.In Japan, there is a shortage of final disposal sites for waste and the amount of waste must be reduced, and the cement industry needs to explore new applications of cement clinker in order to fully and effectively utilize these wastes.Large amounts of the waste, including waste soil from construction, sewage sludge, waste clay and coal ash, are used to produce Belite-Gehlenite clinker (GCL).GCL is sintered from raw materials consisting of virgin material (limestone, quartzite, etc.) that are blended and pulverized with the industrial waste.Japanese cement companies feed these raw materials into their rotary kilns to produce clinker at a rate of 2000 tons [4,409,245 lb] per day using a firing temperature of around 1200˚C [2,192℉].Each tonne [2,204.62lb] of GCL uses 0.6 tonnes [1,322 lb] of waste, which is a waste usage ratio 1.5 to 2 times better than for the ordinary Portland cement clinker (NCL) commercially produced by Japanese cement plants.In this study, the focus is on the application of this GCL as a concrete aggregate.The durability of concrete structures and how to improve it is an important issue.Any cracks that form must be repaired before they become harmful.However, concrete is used in a vast range of applications and locations, and there are many concrete bridges and other large structures where inspection and repair absorbs a great deal of time, money and human resources.On the other hand, it has been confirmed that when clinker is used as aggregate in concrete, a self-healing property develops because any unhydrated clinker reacts with water to fill cracks that develop [1].Therefore, in this study, the crack resistance performance of concrete using GCL as fine aggregate is evaluated in addition to the basic properties of the concrete.

Experiments on fundamental properties of concrete using GCL as fine aggregate
Some results regarding the fundamental properties of mortar using GCL as fine aggregate can be found in the literature.As the replacement ratio of GCL aggregate increases, compressive strength increases, but drying shrinkage and the carbonation rate decrease [1] [2].The focus of this paper is concrete using GCL as fine aggregate: the fundamental properties of concrete samples with varying proportions of GCL are evaluated by experiment.

Mix Proportions
Table 2 shows the mix proportions of the concrete samples.By controlling the amount of SP and AE, slump was adjusted to be within 10±2cm [3.94±0.79in] and air volume to within 4.5±1.5%.For symbols in Table 2, see 2.1 Materials .

Test Methods
Tests were carried out on the fresh and cured concrete as follows.

Fresh property tests
Slump tests were carried out in accordance with ISO 4109 and air volume measurements were carried out in accordance with ISO 4848.

Compressive strength test
Compressive strength tests were carried out in accordance with ISO 1920-4, after 7 and 28 days of curing in water at 20˚C [68℉].The test specimens were φ100 mm × h200 mm [φ3.937 in × h7.874 in] in size and three specimens were tested for each GCL/S ratio.

Drying shrinkage test
Drying shrinkage tests were carried out in accordance with ISO 1920-8.Three prismatic test specimens measuring 100 mm × 100 mm × 400 mm [3.937 in × 3.937 in × 15.748 in] were tested for each GCL/S ratio.Measurements were started after curing for 7 days in water at 20˚C [68℉].Drying conditions were 20˚C [68℉] and 60% relative humidity.

Freezing and thawing test
Freeze-thaw tests were carried out in accordance with ASTM C666.Two prismatic test specimens measuring 100 mm × 100 mm × 400 mm [3.937 in × 3.937 in × 15.748 in] were tested for each GCL/S ratio.Testing began after curing in water at 20°C [68℉] until the age of 28 days.Freezing and thawing shall take not less than 3 hours and not more than 4 hours per cycle.The core temperature of the test piece was reduced to between 5 °C and -18 °C and increased to between − 18 °C and 5 °C.The maximum and minimum temperatures at the centre of the specimen during each cycle are in the range of 5 ± 2 °C and − 18 ± 2 °C.

Accelerated carbonation test
Accelerated carbonation tests were carried out in accordance with ISO 1920-12.Two prismatic test specimens measuring 100 mm × 100 mm × 400 mm [3.937 in × 3.937 in × 15.748 in] were tested for each GCL/S ratio.The tests were conducted after curing for 28 days in water at 20˚C [68℉], followed by curing for 28 days in air at 20˚C [68℉] and 60% relative humidity.The accelerated carbonation conditions were 20˚C [68℉], 60% relative humidity and a carbon dioxide concentration of 5%.

Fresh property tests
Table 3 shows the slump and air content results.SP and AE in the Table 3 indicate the addition rate of the admixture described in the 2.1 Materials.Values for all samples were within the target range as a result of using SP and AE.

Compressive strength
Figure 3 shows results of compressive strength testing.Samples containing GCL exhibit higher compressive strength than the one with sand only.Hosoda et al. reported that when clinker is used for aggregate, the hydration reaction proceeds and the transition zone at the aggregate interface becomes denser [3].In this study as well, it is considered that hydration of GCL resulted in increased compressive strength because the transition zone at the interface between cement paste and fine aggregate became more dense and number of fragile voids was reduced.

Drying shrinkage
Figure 4 shows results of the drying shrinkage tests.Overall, the drying shrinkage of specimens containing GCL is less than that of the one without GCL.Minimum drying shrinkage was exhibited with a GCL content of 25%.This may be because the amount of bleeding increases as the volume ratio of GCL increases.Figure 5 shows the amount of bleeding from GCL mortar as reported in a previous study [4].It is clear that bleeding increases as the volume ratio of GCL increases.When the volume ratio exceeds 25%, the amount of bleeding increases significantly and this is thought to be the reason for the 25% volume ratio showing the best drying shrinkage results.However, further study of this matter is needed.

Freezing and thawing test
Figure 6 shows the freeze-thaw test results.Although the increased bleeding associated with GCL addition was a cause for concern that freeze-thaw resistance might be compromised, it was found that resistance to freezing and thawing did not decrease when GCL was added to the mix.

Accelerated carbonation test
Figure 7 shows the results of accelerated carbonation testing.As the volume ratio of GCL increases, the carbonation rate decreases.The calcium hydroxide derived from GCL may delay the progress of carbonation.Densification in the cement-aggregate transition zone also may have some effect.

Crack Resistance of Concrete with Bealite-Gehlenite Clinker as Fine Aggregate
Past research has proven that, when GCL is in the mix, hydration products fill the voids which seems to be the crack, thereby providing a self-healing property.In this section, the resistance of concrete containing GCL fine aggregate to crack generation is examined.

Materials
The materials used in this examination are ordinary Portland cement (Symbol: C; density: 3.15 g/cm 3

Formulation
Table 4 shows the planned concrete mix proportions.The amounts of SP and AE were adjusted to target a slump value of 10±2 cm [3.94±0.79in] and an air volume of 4.5±1.5%.For symbols in Table 4, see 3.1 Materials .

Fresh property tests
The fresh properties were examined in the same manner as in Section 2.3.River sand (Symbol: RS, surface dry density: 2.58 g/cm 3  [0.0932lb/in3]) was used instead of crushed sand as fine aggregate.

Crack resistance test
The evaluation of crack resistance was based on the autogenous shrinkage stress measurement method [5] described in the report of the autogenous shrinkage research committee of the Japan Concrete Institute (JCI).A D32 deformed steel bar with a strain gauge attached was cast in a 100 mm × 100 mm × 1500 mm [3.937 in × 3.937 in × 59.055 in] specimen as shown in Figure 8 and the strain generated in the reinforcement was measured.The test procedure is described below.(c) The sample concrete is placed in the mould and compacted.At this time, care is taken to ensure that the strain gauges do not move.(h) The crack resistance is evaluated using the shrinkage stress calculated from the strain measurement and the time to crack generation.The shrinkage stress is calculated using the following equation (1).

Fresh property tests
The results are shown in Table 5. SP and AE in the Table 5 indicate the addition rate of the admixture described in the 3.1 Materials.By adjusting the admixture, the target values were achieved for all samples.

Crack resistance test result
The test results are shown in Figure 9 and Table 6.In this graph, the vertical axis represents the shrinkage stress calculated using Equation ( 1), and the horizontal axis represents the age of the material, indicating that the crack occurred in the test piece when the value of the shrinkage stress suddenly changed at the beginning.In the period up to 28 days of age, when drying was prevented, the stress due only to autogenous shrinkage was greater in the GCL 50% sample than in GCL 0%.In this condition, when drying of the sample is prevented, the measured shrinkage derives from hydration shrinkage caused by the reaction of water with cement [6].The water-cement (W/C) ratio greatly affects the amount of autogenous shrinkage, with shrinkage increasing as the W/C ratio is reduced [7].When clinker aggregate is used, particles of 0.15 mm [0.006 in] or less can be regarded as binder.Figure 2 shows that 9% or less of the GCL mass meets this criterion, so in the GCL 50% sample the GCL would contribute about 40 kg [88.185 lb] of binder per cubic metre.This would mean that the actual W/C ratio was 39.1%, or 4.1% lower than in the GCL 0% sample.This reduced W/C resulted in the increased autogenous shrinkage seen in GCL 50% in comparison with the concrete containing no GCL.
In the period after drying began, shrinkage stress was also greater in GCL 50% than in GCL 0%, but crack initiation was slower.Table 6 shows that the first cracks occurred in the GCL 0% sample within 10 days, while no cracks were observed in the GCL 50% sample until 31 days.It is surmised that this is because the hydration reaction proceeds at the clinker surface, hardening and densifying the structure and resulting in improved adhesion with the reinforcement.This results in a highly elastic structure deriving from the rigidity of GCL itself and this constrains deformation.These results confirm that crack resistance is improved when GCL clinker is added to the concrete mix.

Conclusion
In this study, researchers investigated the properties of concrete containing Belite-Gehlenite clinker (GCL) as fine aggregate.Based on the results of the study, the following conclusions can be drawn.
(1) Compressive strength increases and drying shrinkage decreases compared with concrete containing no GCL.
Carbonation is suppressed and does not adversely affect freezing resistance.However, as the proportion of GCL is increased, bleeding increases.This may affect the cured properties and further studies of this effect are needed in the future.
(2) Crack resistance tests confirm that the addition of GCL leads to more drying days before crack initiation, in spite of greater shrinkage stress.That is, crack resistance is improved by the addition of GCL.

Fig. 6 .Fig. 7 .
Fig. 6.Freeze thaw testing (a)  In order to prevent the free deformation of the sample from being restricted, a Teflon sheet (thickness: 1 mm [0.039 in]) is placed inside the assembled form on the bottom surface, and a polystyrene panel (thickness: 3 mm [0.118 in]) is placed inside each end.Next, a polyester film (thickness: 0.1 mm [0.004 in]) is laid within the mould on the side, end and bottom surfaces to prevent the sample from contacting the mould.(b) A D32 deformed steel bar with ribs and joints removed from the central 300 mm [11.811 in] section is used as the restraint reinforcement and is installed in the position shown in Figure8.The diameter of the centre 300 mm section of the reinforcing bar is 30 mm[1.181 in]and strain gauges are attached above and below the centre.

Fig. 8 .
Fig. 8. Form used in "Method of Testing Self-Shrinkage Stress of Concrete

Table 1 .
Mineral composition of clinker

Table 3 .
Fresh property results

Table 4 .
Concrete mix proportions 3]) as coarse aggregate, tap water (symbol: W) as kneading water, polycarboxylic acid-based high-performance AE water reducing agent (symbol: SP), and AE agent (symbol: AE) as composite high alkylcarboxylic acidbased ionic surfactant and nonionic surfactant.The particle size distribution of the fine aggregates was adjusted to be the same as in Figure2.
After removal from the form, the strain generated in the reinforcing bar is measured, and the stress generated in the concrete is calculated from the axial strain of the reinforcing bar.The strain is measured at 1 h intervals up to 48 h of age and at 4 h intervals thereafter.(g) On the 28th day of age, the aluminum foil tape is removed from the test piece and drying begins.During this period, strain measurements are continued.They continue until cracks are generated in the test piece.

Table 5 .
Fresh property results

Table 6 .
crack resistance test result