Journal Information
Vol. 8. Issue 4.
Pages 3443-3452 (July - August 2019)
Share
Share
Download PDF
More article options
Visits
270
Vol. 8. Issue 4.
Pages 3443-3452 (July - August 2019)
Original Article
DOI: 10.1016/j.jmrt.2019.06.010
Open Access
Enhanced autogenous healing of ground granulated blast furnace slag blended cements and mortars
Visits
270
Young Cheol Choia, Byoungsun Parkb,
Corresponding author
pbs0927@kcl.re.kr

Corresponding author.
a Department of Civil and Environmental Engineering, Gachon University, 1342 Seongnamdaero, Sujeong-gu, Seongnam-si, Gyeonggi-do, 13120, South Korea
b Construction Technology Research Center, Korea Conformity Laboratories, Seoul, 08503, South Korea
This item has received
270
Visits

Under a Creative Commons license
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (13)
Show moreShow less
Tables (3)
Table 1. Chemical compositions and physical properties of the raw materials.
Table 2. Mixture proportions.
Table 3. Slump flow test results for mortar.
Show moreShow less
Abstract

In this study, the autogenous healing characteristics of cementitious materials with ground granulated blast furnace slag (GGBFS) and crystalline admixtures were investigated. The main variables were the replacement of cement by GGBFS (15–50%) and crystalline admixture types (anhydrite, Na2SO4, and Na2CO3). Isothermal calorimetry analysis was performed to evaluate the potential for crack self-healing. The water flow test was used to evaluate the autogenous healing performance after inducing cracks in 7-day aged specimens. The autogenous healing products were analyzed through scanning electron microscopy (SEM). Cumulative heat was the highest when the GGBFS replacement percentage was 15% and the crystalline admixtures were used simultaneously, and it decreased when the replacement percentage of GGBFS increased. The results of the water flow test showed that crack self-healing performance was enhanced as the replacement ratio of GGBFS was increased up to 30%. The crack self-healing performance was the highest when anhydrite and Na2SO4 were used at the same time. SEM analysis showed that the types of self-healing products affected the crack self-healing performance.

Keywords:
Autogenous healing
Water flow test
Isothermal calorimetry
Ground granulated blast furnace slag
Crystalline admixture
Full Text
1Introduction

Concrete has the ability to autogenously heal small cracks occurring due to shrinkage and hydration heat induced by the properties of its component materials. This is typically called autogenous healing [1–3]. The autogenous healing mechanism of concrete cracks is roughly divided into the precipitate formation mechanism induced by further hydration of unreacted materials existing on the crack surfaces and the calcite formation mechanism induced by the reaction of CO32− existing on the crack surfaces and Ca2+ flowing in from the pores inside the cement matrix [4]. The autogenous healing properties of cementitious materials improve the durability of concrete structures through the self-healing of cracks. However, according to result of previous studies, the crack width healable by autogenous self-healing is, at maximum, only about 0.1 mm [5]. Cracks accelerate penetration of external harmful ions inside the concrete, and this decreases the durability performance of concrete structures since damages, such as rebar corrosion, are accelerated [6–8]. The problem of durability does not occur when the crack width is 0.1 mm or less, but when it is larger, the penetration rate of harmful ions increases sharply [9,10]. Therefore, to improve the durability of concrete through autogenous healing, a technology that can heal cracks with widths larger than 0.1 mm is needed.

Numerous studies have been carried out to improve the autogenous healing performance of concrete by using an inorganic binder such as SCMs, expansion agents, swelling materials, and crystalline admixtures. Termkhajornkit et al. conducted an autogenous healing performance evaluation study according to the content of fly ash (FA) using a method of measuring the changes of the chloride ion diffusion coefficient according to crack recovery [11]. FA was used to replace up to 50% of ordinary Portland cement (OPC), and through the experiment, it was confirmed that, as the mixing ratio increased, autogenous healing improved. Sahmaran et al. performed an autogenous healing performance evaluation of concrete according to the blending of FA and ground granulated blast furnace slag (GGBFS) [12]. The durability performance recovery by autogenous healing after occurrence of cracks was evaluated through a rapid chloride permeability test (RCPT), and it was confirmed through the test that the durability recovery performance by autogenous healing was significant when GGBFS was blended. Van Tittelboom et al. analyzed the autogenous healing properties according to mixes of FA and GGBFS [13]. The autogenous healing performance was evaluated for replacements of up to 50% of OPC by FA and up to 85% of OPC with GGBFS. Isothermal calorimetry, crack closing, and water flow tests were conducted for evaluating the autogenous healing performance. According to this study, FA and GGBFS both enhanced autogenous healing compared with OPC. GGBFS enhanced autogenous healing more than FA, and was led to better mechanical performance of concrete. Roig-Flores et al. investigated the self-healing capacity of cementitious materials incorporating crystalline admixtures (CAs) [14]. OPC was replaced with 4% by weight of CA and cracking was induced at 2 days. The specimen was cured for self-healing 42 days after cracking, and self-healing performance was evaluated through a water permeability test and crack closing test. The experimental results showed that the self-healing performance of the specimens immersed in water increased but decreased in wet/dry conditions. Cuenca et al. investigated the self-healing performance of cement composites incorporating CAs according to curing conditions [15]. The test results indicated that the self-healing performance of specimens containing CAs was improved when immersed in water. However, in wet/dry and outdoor conditions, the self-healing performance of specimens with CAs was similar to that of specimens without CAs. They concluded that the self-healing performance was higher for specimens with CAs immersed in water as compared with those for control. Sisomphon et al. investigated the self-healing performance of cement-based materials with calcium sulfo-aluminate (CSA) as an expansive agent and CAs [16]. It was concluded from the test results that it is possible to heal cracks up to 150 μm autogenously by water curing for 28 days. ACI committee 212 defined CAs as hydrophilic materials that react with water and cement particles in concrete to form hydrates and fill microcracks and capillary pores [17]. Therefore, the permeability of concrete can be improved by incorporating CAs into concrete.

According to results of previous studies, the mixing of GGBFS is advantageous for improving autogenous healing performance in comparison with FA, and it was confirmed that the autogenous healing performance can be enhanced through blending of calcium sulfoaluminate (CSA) expansion agents and crystalline admixture (CA) [18]. In the present study, analysis of physical and autogenous healing properties of concrete was conducted using varying mixes of GGBFS and crystalline admixtures. The characteristics of fresh concrete were investigated using a setting test and a slump flow test. The characteristics of hardened concrete were investigated using autogenous shrinkage measurements and compressive strength tests. To analyze autogenous healing properties, an isothermal calorimetry and water flow test were performed. SEM analysis was carried out to analyze the components of autogenous healing products in the cracks.

2Material and methods2.1Materials

In the present study, OPC, GGBFS (a byproduct of the steel industry), and a high-strength admixture (HSAD) were used as inorganic binders. As the HSAD, a product typically used in South Korea for the purpose of strength increase of secondary concrete products was chosen. Additionally, anhydrite, Na2SO4, and Na2CO3 were used as crystalline admixtures for the stimulation of autogenous healing. Table 1 shows the chemical compositions and basic physical properties of raw materials obtained using X-ray fluorescence (XRF).

Table 1.

Chemical compositions and physical properties of the raw materials.

Chemical compositions (%)
OPC  GGBFS  HSAD  Anhydrite 
SiO2  18.55  29.13  21.9  0.78 
Al2O3  4.41  11.82  10.3  0.29 
Fe2O3  3.23  0.44  1.58  0.12 
CaO  62.13  42.51  49.3  45.3 
MgO  2.04  2.43  4.33  – 
K21.22  0.52  0.55  0.06 
Na20.26  0.2  0.45  – 
TiO2  0.27  0.59  0.52  – 
MnO  0.2  0.23  0.17  – 
P2O5  0.13  –  0.14  0.02 
SO3  3.03  3.34  10.7  53.3 
SrO  –  0.05  0.06  0.08 
  Physical properties
Blaine fineness [m2/kg]  383  428  392  405 
Density [kg/m33,120  2,950  2,550  2,975 

Fig. 1 shows the particle size distributions of OPC, GGBFS, HSAD, and anhydrite measured via laser diffraction analysis (LA-960, HORIBA). The average particle sizes were 17.84, 14.02, 11.90, and 17.40 μm, respectively.

Fig. 1.

Particle size distributions of used raw materials.

(0.06MB).

The main mineral components of OPC are C3S, C2S, C3A, and C4AF, with contents of 62.0%, 16.1%, 2.5%, and 12.1%, respectively. As shown in Fig. 2(a), OPC contains 2.6% gypsum and as a mineral admixture, 4.5% limestone powder. Based on X-ray diffraction (XRD) analysis results, GGBFS is mostly composed of a 95% amorphous phase, and a crystalline phase composed of anhydrite (2.9%) and quartz (2.1%).

Fig. 2.

XRD patterns of OPC and GGBFS.

(0.08MB).

Table 2 presents the composition ratios of binders for the autogenous healing performance evaluation of cracks. The HSAD replaced 30% of the total binder weight. The binders in Table 2 represent the pastes used for the measurement of isothermal calorimetry and the mortars used for the water flow test; water-binder (W/B) ratios were all fixed at 0.3. The mortars were fabricated with equal amounts of sand and binder.

Table 2.

Mixture proportions.

LabelsBinder (g)
OPC  HSAD  GGBFS  Anhydrite  Na2SO4  Na2CO3 
Plain  1820  780  –  –  –  – 
S15  1430  780  390  –  –  – 
S30  1040  780  780  –  –  – 
S15A5NS3  650  780  390  130  78  – 
S30A5NS3  1040  780  780  130  78  – 
S30NS3NC3  650  780  780  –  78  78 
S50NS3NC3  962  780  1300  –  78  78 
2.2Test methods2.2.1Setting time and slump flow

In the present study, the setting time of paste was measured by using a PA8 automatic setting tester (ACMEL) in accordance with ISO 9597 [19]. After putting the fresh paste—mixed in accordance with proportions given in Table 2—in a container of 40 mm height, the penetration depth in the paste of a 1.13 mm diameter needle under gravity was measured. The initial setting time was determined to be the moment when the penetration depth of the needle became 25 mm, and the final setting time was determined to be the moment when the penetration depth became 0.5 mm or less. The penetration depth was measured at 15 min intervals, and when the measurement depths corresponding to the initial and final setting times were in between the measurement times, the initial and final setting times were calculated using linear interpolation between measurement times.

2.2.2Isothermal calorimetry

The isothermal calorimeter used for the measurement of the heat flow of hydration was a TAM Air isothermal calorimeter with eight-channels operating in the milliwatt range. To evaluate the autogenous healing potentials under further hydration of the binder, an isothermal calorimetry test was carried out at the age of 0, 7, and 91 days. According to the mixture proportions in Table 2, 100 g of binder and 30 g of tap water were mixed in a glass beaker, and about 5 g of this was put in an ample and after measuring the mass, the occurring heat flow of hydration was measured. To evaluate the autogenous healing potentials after 7 and 91 days, 10 × 10 × 10 mm specimens were fabricated with a W/B ratio of 0.30 using the mixture proportion of Table 2. The fabricated specimens were cured in a chamber at 20 ± 3 °C and 100% of relative humidity (RH) for 24 h. Afterwards, before conducting the autogenous healing potential evaluation, the specimens were water-cured in a water container of 20 ± 3 °C. The specimen that was water-cured up to the goal age (7 and 91 days) was dried in a 40 °C chamber for 24 h before grinding it to obtain a powdered specimen. After obtaining enough powder, the powder was sieved through a 200 μm sieve, thereby preparing the specimen for mixing. After mixing 10 g of the prepared powder specimen and 4 g of tap water, about 4.5 g of rehydrated paste was put into an ample and the test was carried out. The temperature changes were measured every minute and the isothermal calorimetry was performed for three days (72 h).

2.2.3Autogenous shrinkage

Using the fresh paste fabricated in accordance with Table 2, autogenous shrinkage was measured in a chamber with a temperature maintained at 23 °C according to the standard ASTM C 1698 “Standard Test Method for Autogenous Strain of Cement Paste and Mortar” [20]. The corrugated plastic tube used in the test had a length of 420 ± 5 mm, a diameter of 29 ± 0.5 mm, and a wall thickness of 0.5 ± 0.2 mm. The fresh paste mixed according to Table 2 was filled up inside the corrugated plastic tube, and a vibration table was used to make sure that there was no empty space inside the tube. After closing the entrance with a plug, it was put on a support and the initial length was measured. Afterwards, the length of the specimen was measured using a data logger and a linear variable differential transformer (LVDT).

2.2.4Water flow test

In the present study, the autogenous healing performance was evaluated through the water flow test. In the test, ϕ100 mm × 50 mm specimens, fabricated with the mixture proportion in Table 2, were used. After fabrication, the specimens were cured in a 20 ± 3 °C and 100% RH chamber for 24 h, and afterwards, cracks were induced by water-curing in the water container with a temperature of 20 ± 3 °C until the age of 7 days. The cracks were induced through a cleavage tensile strength test method, and after completely separating a specimen into two pieces, a copper wire, about 0.2 mm in diameter, was inserted so that the crack width would be constant. Afterwards, both sides of the specimen were fixed with an epoxy. To make sure that the crack width would not change during the test process, each specimen was inserted into a rubber ring and the specimen was fixed with a steel band; under this state, the crack width was measured for each specimen. The water flow test was carried out in a static head state, in which the height of the head was 160 mm (including the length of specimen), and tap water with a temperature of 20 ± 3 °C was used.

3Results and discussion3.1Setting time and flow test

Fig. 3 shows the setting time test results. It was confirmed that as the replacement of GGBFS increased, the initial and final setting times increased; this was caused by the fact that the hydration rate of GGBFS is lower compared to that of OPC, and accordingly, the strength development of the specimens slowed down [21]. In the case of specimens S15A5NS3, S30A5NS3, S30NS3NC3, and S50NS3NC3, using the crystalline admixtures, the initial and final setting times decreased compared with the plain specimen. In general, it is known that the SO42− ions supplied by anhydrite and Na2SO4 stimulate the ettringite production inside the concrete, thereby increasing the initial strength development [22]. Mota et al. investigated the effect of OH and SO42− ions on the compressive strength and initial hydration properties of cementitious materials by mixing water containing NaOH and Na2SO4. An isothermal calorimetry test showed an increase in initial heat flow when Na2SO4 was added. It was confirmed that the initial compressive strength was higher than that of distilled water [23]. In the case of Na2CO3 used for S30NS3NC3 and S50NS3NC3, although a small amount of about 3% binder content was used as a material to produce a rapid set, it was confirmed that the setting time decreased sharply. Especially, in the case of a 50% mixing ratio of GGBFS, the initial setting time of specimen S50NS3NC3 was shorter than that of specimen S30NS3NC3, and through this, it was confirmed that the setting time changed significantly depending on the Na2CO3 content.

Fig. 3.

Initial and final setting time.

(0.07MB).

Table 3 shows the slump flow test results for mortar. It was confirmed that as the mixed amount of GGBFS increased, the flow increased as well. In the case of GGBFS, because the hydration reaction did not occur during the initial period of mixing, the water content increased as much as the replacement amount of GGBFS compared with OPC; hence, the water to cement ratio increased. Owing to this effect, when the replacement amount of GGBFS increased, the slump flow increased as well. In the case of specimens S15A5NS3 and S30A5NS3, containing crystalline admixtures, flow increased compared with specimens using only GGBFS, unlike the results of setting test; and in the case of the mixture containing Na2CO3, no slump flow was measured due to rapid setting. In the case of specimens S15A5NS3 and S30A5NS3, it was determined that the initial flow increased because the hydration reaction of anhydrite progressed slowly, as for mixtures S15 and S30.

Table 3.

Slump flow test results for mortar.

  Plain  S15  S30  S15A5NS3  S30A5NS3  S30NS3NC3  S50NS3NC3 
Flow [mm]  169.23  176.79  181.23  198.07  194.26  a  a 
a

No slump flow by rapid set occurred.

3.2Isothermal calorimetry

Fig. 4 shows the test results of specific heat flows of specimens Plain and S30. In the case of the Plain specimen, the specific heat flow occurring at an age of 7 days was higher than at the age of 91 days; the heat flows became identical after about 60 h. The reason for this was determined to be the presence of more unreacted clinkers inside the paste after 7 than after 91 days. After 60 h, the specific heat flow of the 7-day aged specimen became equal to that of the 91-day aged specimen because the unreacted clinkers reacted rapidly due to the high W/B ratio. On the other hand, in the case of specimen S30, the specific heat flow of the 91-day aged specimen was higher than that of 7-day aged specimen; the heat flows became similar after 48 h.

Fig. 4.

Specific heat flow of specimens Plain and S30.

(0.1MB).

Fig. 5(a and b) is the result of measuring the cumulative heat by further hydration of unreacted substances inside the 7-day aged specimens. In the graphs, a heat of about 20–30 J/g-binder occurred within 1 h, regardless of mixture type, which then increased slowly with time; it was confirmed that the initial heat was not significantly different between the mixtures. Regarding cumulative heat after 72 h, the cumulative heat of specimen S15A5NS3 was the highest and that of specimen S30 was the lowest. For the other five mixtures, the cumulative heat after 72 h was similar. Because the age of 7 days is a relatively early age when considering the hydration rate of common OPC, the effect of unreacted cement clinkers on the further hydration of unreacted substances was large. Therefore, in the case of specimen S30 that replaced 30% of cement with GGBFS, cumulative heat decreased; and in the case of specimens S30A5NS3, S30NS5NC3, and S50NS5NC3, using the crystalline admixtures, the cumulative heat increased compared with specimen S30. On the other hand, specimens Plain and S15, in which 15% of cement was replaced by GGBFS, showed very similar cumulative heats, and in the case of specimen S15A5NS3 with crystalline admixtures added to mix S15, cumulative heat increased.

Fig. 5.

Cumulative heats of specimens.

(0.28MB).

Fig. 5(c and d) shows the cumulative heat measurement results with respect to the further hydration of unreacted substances of the 91-day aged specimens. In the graph, it can be confirmed that the cumulative heats decreased compared with the 7-day aged specimens. This was because the amount of unreacted binders decreased as the age increased. In the cumulative heat measurement results of after 91 days, the cumulative heat of the Plain specimen was the lowest, unlike the cumulative heat measurement results after 7 days. The reason for this is the decrease of unreacted cement clinkers due to increasing hydration of OPC. Another reason for this was the increasing effect of further hydration of supplementary cementitious materials (SCMs) and crystalline admixtures. In the case of 91 days, specimen S15 that used 15% of GGBFS showed higher cumulative heat than specimen S30 that used 30% of GGBFS. These results stand in opposition to the results of Tittelboom that the cumulative heat increases as the mixed amount of GGBFS increases [13]. It was determined that the results by Tittelboom were different from the results of the present study because the former measured the cumulative heat for 336 h for the cases of 50% and 80% of GGBFS. The amount of unreacted binder is larger in S30 than in S15 at 28 days because the degree of hydration of GGBFS is generally lower than that of OPC [24]. Therefore, the hydration heat of S30 generated by further hydration was expected to be higher than that of S15. However, the experimental results differed from expectations. This is because a certain amount of time is required for the additional hydration reaction of unreacted GGBFS. In this study, the cumulative heat measurement period is 72 h, which is considered to be sufficiently short to react with unreacted GGBFS and water. As shown in Fig. 5, cumulative heat increased continuously at 72 h because the unreacted material was continuously hydrated at the crack surface. Therefore, a long-term experiment, such as that conducted by Tittelboom et al., may result in a higher cumulative heat of S30. When the blending ratio of GGBFS was 30% or higher, unreacted substances were produced due to further hydration on the crack surface because of the slow hydration rate of GGBFS, because a relatively long time was required; it was determined that when the blending ratio of GGBFS increased, the measurement period of cumulative heat should also increase.

Also, in the case of 91 days, the cumulative heat of specimen S15A5NS3 was the highest. The cumulative heat of S15A5NS3 and S30A5NS3 containing Na2SO4 and anhydrite was increased compared to that of S15 and S30. Na2SO4 as crystalline admixture accelerated the early hydration of OPC, and the setting time of specimens containing Na2SO4 was reduced. Na2SO4 is also used as an alkali activator to accelerate the reaction of GGBFS. In the case of specimens containing Na2SO4, Na2SO4 accelerates the hydration of unreacted OPC and GGBFS so that the cumulative heat value increased compared with that of S15 and S30. On the other hand, further hydrations of S30NS5NC3 and S50NS5NC3 decreased compared to those of S30. In the case of Na2CO3, slump flow could not be measured because of rapid setting. For this reason, it was supposed that OPC and GGBFS have not been further reacted. Similar phenomena was observed in the measurement of cumulative heat by the further hydration of the unreacted binder. As shown in Fig. 5(c) and (d), the slope of S30NS5NC3 and S50NS5NC3 at 72 h is smaller than that of S30A5NS3. Therefore, the use of Na2CO3 as a crystalline admixture may inhibit further hydration.

3.3Autogenous shrinkage

Figs. 6 and 7 present autogenous shrinkage test results. In case of the Plain specimen, the autogenous shrinkage strain was approximately 290 × 10−6 mm/mm after 14 days. As the replacement ratio of GGBFS increased, so did the autogenous shrinkage. After 14 days, the shrinkage strains of specimens S15 and S30 were 5.2% and 19.0% larger, respectively, compared with the Plain specimen. In the case of specimen S15A5NS3, a different pattern was observed compared with specimens Plain, S15, and S30. After setting, strain increased continuously as time passed because swelling occurred instead of shrinkage. This was due to the anhydrite and Na2SO4 used besides GGBFS. Owing to the supply of SO4 contained in the anhydrite and Na2SO4, the expandable hydrates such as ettringite were produced, thereby expanding the volume of the specimen. Similar results were obtained for specimen S30A5NS3, and it was confirmed that the strains of specimens S15A5NS3 and S30A5NS3 were similar. Specimens S30NS3NC3 and S50NS3NC3, in which anhydrite and Na2CO3 were used, showed almost no volume change by autogenous healing; probably because hardening had already started when the paste was put into the mold due to rapid setting caused by mixing of Na2CO3. When hydration continued, overall volume change did not occur; it was determined that the overall volume change did not appear because of the parts not filled inside the mold.

Fig. 6.

Autogenous shrinkage of specimens Plain, S15, S30.

(0.06MB).
Fig. 7.

Autogenous shrinkage of specimens S15A5NS3, S30A5NS3, S30NS3NC3.

(0.07MB).
3.4Water flow test

Figs. 8–10 show the water flow test results. The water flow reduction rates caused by autogenous healing were measured for 28 days, and they were represented as water flow rates in comparison with the initial water flow. Fig. 8 shows the water flow test results of specimens Plain, S15, and S30. As the content of GGBFS increased, the reduction rate of water flow increased. This was different from the isothermal calorimetry result, in which the cumulative heat of specimen S15 was higher than that of specimen S30; and through this, it was confirmed that it was difficult to evaluate the autogenous healing performance using only the heat flow of hydration of unreacted substances. The types and volume changes of autogenous healing products precipitated on the crack surface must be also considered besides the heat flow of hydration, and for this, it can be viewed that the component analysis of autogenous healing products should be conducted in parallel. The reason for the decrease of water flow in specimen S30 compared with specimen S15 was the increased amount of unreacted binders as the content of GGBFS increased, and in addition, C-A-S-H was produced as aluminate ions were mixed.

Fig. 8.

Water flow test results of specimens Plain, S15, S30.

(0.06MB).
Fig. 9.

Water flow test results of specimens S15A5NS3 and S30A5NS3.

(0.05MB).
Fig. 10.

Water flow test results of specimens S30NS3NC3 and S50NS3NC3.

(0.05MB).

Fig. 9 shows the results for specimens S15A5NS3 and S30A5NS3. In the cases of specimens S15A5NS3 and S30A5NS3, the water permeability reduction appeared rapidly up to initial 4 days compared with specimens Plain, S15, and S30, due to the crystalline admixtures. It was determined that due to SO42− supplied from anhydrite and Na2SO4, the expandable hydrates such as AFm, and ettringite were produced on the crack surfaces, thereby healing the cracks more efficiently with small hydration reactions. Na2SO4 acts as a crystalline accelerator and produces hydrate quickly. These results were similar in setting time. In the case of specimens containing Na2SO4 and anhydrite, ettringite formation was accelerated and setting time decreased. In the water flow test, hydration products were generated on the cracked surface in a short time through the same mechanism and self-healing performance was improved. Furthermore, the water permeability reduction increased in specimen S30A5NS3 compared with specimen S15A5NS3, due to the increases in C-A-S-H, AFm, and ettringite production induced by aluminates supplied from GGBFS.

Fig. 10 shows the water flow test results of specimens S30NS3NC3 and S50NS3NC3. In the cases of specimens S30NS3NC3 and S50NS3NC3, the cumulative heat measured by isothermal calorimetry was small compared with specimens S15 and S30, but the water flow reduction rate was high. This result was caused by the difference in autogenous healing products produced on the crack surfaces, and was determined to be an effect of Na2SO4 used as a crystallization accelerator. NaCO3 was used to induce the production of calcite through the supply of CO32−, but in reality, it did not have much effect on autogenous healing.

3.5SEM analysis

From the results of isothermal calorimetry and water flow test, it can be concluded that autogenous healing performance was affected by mix proportions of the specimens. The SEM analysis was performed to analyze the autogenous healing products according to mixture and evaluate which autogenous healing product is advantageous for crack healing. Figs. 11–13 show the SEM analysis results of autogenous healing products produced on the cracks of specimens Plain, S30, and S30A5NS3. In the SEM analysis result of the Plain specimen, Portlandite and calcite, which were the plate type hydrates produced by hydration of OPC, were mainly investigated. In the case of specimen S30, besides calcite, it was confirmed that monosulfates and C-A-H compounds were mainly produced by aluminum supplied from GGBFS. Because GGBFS consumes Ca(OH)2 during the hydration process, Portlandite was not found. Monosulfates are, as shown in Fig. 11, needle-type hydrates having a long shape, unlike Portlandite and calcite, and they can have a bridge role since they are produced across the crack surfaces; further, since C-A-H is produced around the monosulfates, the cracks can be effectively filled up. Therefore, in the case of specimen S30, a crack can be effectively healed even when higher hydration reaction heat occurs compared with that the Plain specimen. As a result of analyzing the autogenous healing substances of specimen S30A5NS3, it was confirmed that a needle type hydrate, ettringite was produced, as shown in Fig. 12. The ettringite was produced on the cracks by sulfur ions flown out from the anhydrite and Na2SO4, which were used as crystalline admixtures. The ettringite can have not only a bridge role at the cracks as a needle-type hydrate like monosulfates, but can also be effective in filling the cracks since volume is expanded during the process of hydrate production. Therefore, in the case of specimen S30A5NS3, it is determined that despite of less cumulative heat, the autogenous healing performance is better.

Fig. 11.

SEM observations of self-healing products of specimen Plain.

(0.17MB).
Fig. 12.

SEM observations of self-healing products of specimen S30.

(0.18MB).
Fig. 13.

SEM observations of self-healing products of specimens S30A5NS3.

(0.22MB).
4Conclusions

In the present study, an experimental investigation was conducted to determine possibilities for improving the autogenous healing performance of cementitious materials with GGBFS and crystalline admixtures. GGBFS replaced 15%, 30%, and 50% of binders, while anhydrite, Na2SO4, and Na2CO3 were used as crystalline admixtures. For the basic properties evaluation of the mixtures used in the tests, the setting and slump flow tests were performed. Setting time increased as the replacement ratio of GGBFS increased, and decreased when the crystalline admixtures were used. Slump flow increased as the GGBFS replacement increased and increased even more when crystalline admixtures were used. However, when Na2CO3 was mixed, no slump flow was measured due to rapid setting.

As the methods of autogenous healing performance evaluation, the isothermal calorimetry and the water flow test were used, and the autogenous healing substances produced on the cracks were analyzed using SEM. As a test result, it was confirmed through isothermal calorimetry that the cumulative heat decreased as the specimen age increased. This possibly happened because unreacted clinkers, which facilitated further hydration, decreased as the degree of hydration increased. In the comparison of results according to the replacement ratios of GGBFS, the blend that replaced 15% of binder with GGBFS showed the highest cumulative heat; in contrast, when the replacement ratio of GGBFS increased, cumulative heat decreased. One explanation for this result is that when the mixing amount of GGBFS increased, the OPC decreased, reducing OH- ions needed for the reaction with GGBFS. However, it was confirmed that the cumulative heat increased continuously after 72 h, which was the termination time of test. Therefore, it was determined that the test results would have been different if the cumulative heat had been measured over a longer period.

In the water flow test results, the autogenous healing performance was enhanced as the GGBFS replacement increased, and it was confirmed that the autogenous healing performance was maximum when 30% of binder was replaced with GGBFS and anhydrite and Na2SO4 were used as crystalline admixtures. This was a different result compared with the result of cumulative heat measured using isothermal calorimetry, and SEM analysis confirmed that the autogenous healing products were different depending on the type of autogenous healing material. Particularly, when mixing GGBFS, anhydrite, and Na2SO4, it was confirmed that the volume of autogenous healing products increased even when the cumulative heat was small because a large amount of needle-type hydrates, ettringite, was produced.

In this study, it was confirmed that the incorporation of GGBFS and CAs increased the cumulative heat through further hydration compared with that of Plain, and the effect increased with the passage of time. However, it is not proportional to the self-healing performance by the water flow test and cumulative heat by further hydration. The self-healing performance was influenced not only by the amount of unreacted binder but also by the type of self-healing products generated by further hydration. Therefore, it was found that it is necessary to use a material capable of inducing an expandable material to improve self-healing performance.

Acknowledgements

This work was supported by Korea Environmental Industry and Technology Institute (KEITI) through Public Technology Program based on Environmental Policy Project, funded by Korea Ministry of Environment (MOE) (2016000700003) and grants 19SCIP-B103706-05 from Construction Technology Research Program funded by Ministry of Land, Infrastructure and Transport of Korean government.

References
[1]
S. Jacobsen, E.J. Sellevold.
Self-healing of high strength concrete after deterioration by freeze/thaw.
Cem Concr Res, 26 (1995), pp. 55-62
[2]
S. Jacobsen, J. Marchand, L. Boisvert.
Effect of cracking and healing on chloride transport in OPC concrete.
Cem Concr Res, 26 (1996), pp. 869-881
[3]
N. Hearn.
Self-sealing, autogenous healing and continued hydration: what is the difference?.
Mater Struct, 31 (1998), pp. 563-567
[4]
T.H. Ahn, T. Kishi.
Crack self-healing behavior of cementitious composites incorporating various mineral admixtures.
J Adv Concr Technol, 8 (2010), pp. 171-186
[5]
M. Maes, D. Snoeck, N.D. Belie.
Chloride penetration in cracked mortar and the influence of autogenous crack healing.
Constr Build Mater, 115 (2016), pp. 114-124
[6]
L. Marsavina, K. Audenaert, G.D. Schutter, N. Faur, D. Marsavina.
Experimental and numerical determination of the chloride penetration in cracked concrete.
Constr Build Mater, 23 (2009), pp. 264-274
[7]
D.P. Bentz, E.J. Garboczi, Y. Lu, N. Martys, A.R. Sakulich, W.J. Weiss.
Modeling of the influence of transverse cracking on chloride penetration into concrete.
Cem Concr Compos, 38 (2013), pp. 65-74
[8]
H.-L. Wang, J.-G. Dai, X.-Y. Sun, X.-L. Zhang.
Characteristics of concrete cracks and their influence on chloride penetration.
Constr Build Mater, 107 (2016), pp. 216-225
[9]
A. Djerb, S. Bonnet, A. Khelidj, V. Baroghel-bouny.
Influence of traversing crack on chloride diffusion into concrete.
Cem Concr Res, 38 (2008), pp. 877-883
[10]
M. Sahmaran.
Effect of flexure induced transverse crack and self-healing on chloride diffusivity of reinforced mortar.
J Mater Sci, 42 (2007), pp. 9131-9136
[11]
P. Termkhajornkit, T. Nawa, Y. Yamashiro, T. Saito.
Self-healing ability of fly ash–cement systems.
Cem Concr Compos, 31 (2009), pp. 195-203
[12]
M. Sahmaran, G. Yildirim, T.K. Erdem.
Self-healing capability of cementitious composites incorporating different supplementary cementitious materials.
Cem Concr Compos, 35 (2013), pp. 89-101
[13]
K.V. Tittelboom, E. Gruyaert, H. Rahier, N.D. Belie.
Influence of mix composition on the extent of autogenous crack healing by continued hydration or calcium carbonate formation.
Constr Build Mater, 37 (2012), pp. 349-359
[14]
M. Roig-Flores, F. Pirritano, P. Serna, L. Ferrara.
Effect of crystalline admixtures on the self-healing capability of early-age concrete studied by means of permeability and crack closing tests.
Constr Build Mater, 114 (2016), pp. 447-457
[15]
E. Cuenca, A. Tejedor, L. Ferrara.
A methodology to assess crack-sealing effectiveness of crystalline admixtures under repeated cracking-healing cycles.
Constr Build Mater, 179 (2018), pp. 619-632
[16]
K. Sisomphon, O. Copuroglu, E.A.B. Koenders.
Self-healing of surface cracks in mortars with expansive additive and crystalline additive.
Cem Concr Compos, 34 (2012), pp. 566-574
[17]
ACI Committee 212.
Report ACI 212-3R-10, 2010. Report on chemical admixtures for concrete.
American Concrete Institute (ACI), (2010), pp. 46-50
[18]
M. Roig-Flores, S. Moscato, P. Serna, L. Ferrara.
Self-healing capability of concrete with crystalline admixtures in different environments.
Constr Build Mater, 86 (2015), pp. 1-11
[19]
ISO 9597:2008, Cement — test methods — determination of setting time and soundness.
[20]
ASTM C 1698-09, Standard test method for autogenous strain of cement paste and mortar.
[21]
P.J. Wainwright, H. Ait-Aider.
The influence of cement source and slag additions on the bleeding of concrete.
Cem Concr Res, 25 (1995), pp. 1445-1456
[22]
M. Zajac, A. Rossberg, G.L. Saout, B. Lothenbach.
Influence of limestone and anhydrite on the hydration of Portland cements.
Cem Concr Res, 46 (2014), pp. 99-108
[23]
B. Mota, T. Matschei, K. Scrivener.
Impact of NaOH and Na2SO4 on the kinetics and microstructural development of white cement hydration.
Cem Concr Res, 108 (2018), pp. 172-185
[24]
A. Arora, G. Sant, N. Neithalath.
Ternary blends containing slag and interground/blended limestone: hydration, strength, and pore structure.
Constr Build Mater, 102 (2016), pp. 113-124
Copyright © 2019. The Authors
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Tools
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.