Journal Information
Vol. 8. Issue 6.
Pages 6058-6073 (November - December 2019)
Share
Share
Download PDF
More article options
Visits
0
Vol. 8. Issue 6.
Pages 6058-6073 (November - December 2019)
Original Article
DOI: 10.1016/j.jmrt.2019.09.080
Open Access
Investigating a new method to assess the self-healing performance of hardened cement pastes containing supplementary cementitious materials and crystalline admixtures
Visits
0
Park Byoungsuna, Cheol Choi Youngb,
Corresponding author
zerofe@gachon.ac.kr

Corresponding author.
a Construction Technology Research Center, Korea Conformity Laboratories, Seoul 08503, South Korea
b Department of Civil and Environmental Engineering, Gachon University, 1342 Seongnamdaero, Sujeong-gu, Seongnam-si, Gyeonggi-do, 13120 South Korea
This item has received
0
Visits

Under a Creative Commons license
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (28)
Show moreShow less
Tables (3)
Table 1. Chemical compositions and physical properties of raw materials used in this study.
Table 2. Mineralogical phase compositions of the employed raw materials.
Table 3. Mixture proportions of binders.
Show moreShow less
Abstract

The autogenous healing of cementitious materials allows for the self-healing of cracks in concrete structures. It is known to be caused by (a) further hydration occurring when water penetrating the crack reacts with the unreacted binder on the crack face and (b) calcite precipitation resulting from the reaction between Ca2+ diffused from the cement paste and CO32− in penetrated water. Herein, isothermal calorimetry was used to analyze the further hydration of unreacted binder in hardened pastes containing ordinary Portland cement (OPC), supplementary cementitious materials, and crystalline admixtures. The amount of heat produced decreased with increasing sample age because of the reduction in the unreacted binder amount. Long-term further hydration in samples containing ground granulated blast-furnace slag and silica fume caused them to show increased heat production compared to OPC. However, no particular difference in comparison to OPC was detected for samples containing fly ash. When calcium sulfoaluminate was used as an expansion agent, heat production increased for a material age of 7 days and decreased after 28 days (compared to the case of OPC). Finally, self-healing products were analyzed by scanning electron microscopy with energy dispersive spectrometry. The results showed that the products contained calcite and an amorphous material regardless of the binder, while ettringite was observed to form in calcium sulfoaluminate and crystalline admixtures containing samples.

Keywords:
Further hydration
Isothermal calorimetry
Cementitious materials
Healing potential
Scanning electron microscopy
Full Text
1Introduction

Concrete structures generally have a high risk of cracking due to various reasons such as heat of hydration and dry shrinkage, alkali-aggregate reactions and external loads [1–3]. This accelerates the infiltration of harmful outside ions into the concrete interior and reduces durability. Therefore, crack repair is necessary for improving the durability of concrete structures. If the cracks in the concrete structure do not occur or the cracks heal themselves, the cost can be reduced compared to repairing the cracks [4]. Cracks in concrete are inevitable. Therefore, research on self-healing concrete that heals cracks is being conducted.

The ability of cementitious materials to naturally heal fine cracks occurring in concrete is known as autogenous healing. It can largely be modeled by two mechanisms [5,6]. The first mechanism involves the further hydration of unreacted clinker on crack surfaces, i.e., water flowing in from the outside via cracks reacts with the unreacted clinker on crack faces to form new hydration products. In the second mechanism, Ca2+ ions eluted from the cement paste react with the CO32− ions dissolved in the inflowing water to form insoluble calcite.

Autogenous healing performance is determined by the amount of unreacted binder in hardened cementitious materials and the concentration of Ca2+ in the cement matrix. According to Şahmaran et al., cracks with a width of 80 μm or less can be healed in concrete structures produced using ordinary Portland cement (OPC) [7]. Notably, self-healing can improve the durability of concrete structures without any additional repair works, which has inspired numerous studies on improving autogenous healing performance. In particular, many researchers have used supplementary cementitious materials (SCMs) and crystalline admixtures (CAs) to improve crack self-healing performance [8–17].

Tittelboom et al. investigated the autogenous healing properties of cementitious materials containing ground granulated blast-furnace slag (GGBS) and fly ash (FA) [8]. To this end, 50–85/30–50 wt% of OPC was replaced with GGBS/FA, respectively, and the self-healing performance was evaluated using water permeability tests and isothermal calorimetry (IC). Their results showed that the addition of GGBS and FA resulted in an improved self-healing performance as the extent of further hydration was increased. In addition, GGBS was shown to be more effective at improving the self-healing performance than FA.

Şahmaran et al. employed rapid chloride permeability testing to evaluate autogenous healing in the presence of GGBS and FA, demonstrating that better self-healing performance was observed for samples containing GGBS [9]. Abdel-Jawad et al. and Granger et al. studied the effects of the water to binder (W/B) ratio on autogenous healing performance [10,11], revealing that the autogenous healing performance improved as the W/B ratio decreased. This behavior was ascribed to the decrease in unreacted clinker amount with increasing W/B ratio. Jaroenratanapirom et al. used crack closing tests to evaluate the self-healing performance of samples including CAs and SCMs such as FA and silica fume (SF) [12], showing that SF was most effective at the self-healing of cracks with widths of ≥0.25 mm and further demonstrating that the self-healing performance of FA-containing specimens was maximized after 28 day aging. Sisomphon et al. used crack closing and water passing tests to evaluate the self-healing performance of materials prepared using calcium sulfoaluminate (CSA, expansion agent) and CAs [13]. The results of their study revealed that 0.4-mm-wide cracks could be healed and showing that self-healing materials formed on the crack surface mainly comprised calcite. Sisomphon et al. also analyzed the effect of the surrounding environment on crack self-healing [14] by examining the effects of curing water (such as tap water and regularly refreshed tap water) on already cracked specimens and investigating the role of the curing environment (e.g., wet/dry cycling and air exposure). The self-healing performance of specimens was determined by mechanical performance measurements before and after self-healing, and their results showed that the recovery of mechanical performance by self-healing was most affected by wet/dry cycling.

Sherir et al. studied the recovery of mechanical performance owing to self-healing in engineered cementitious materials containing FA and MgO [15]. Darquennes et al. evaluated the self-healing performance of GGBS-containing concrete by evaluating its durability [16] via chloride migration tests, confirming that the addition of GGBS improved durability and increased the efficiency of self-healing by promoting further hydration. Huang et al. studied the autogenous healing of cement composites containing GGBS [17,18] by collecting self-healing products through artificial cracks and analyzing them by a range of instrumental techniques. As a result, the inclusion of GGBS was demonstrated to form a self-healing material different from OPC and improve self-healing performance, as confirmed by measurements of the crack space filling ratios of self-healing products.

An examination of the literature showed that a variety of test methods have been used to evaluate the self-healing performance of cementitious materials. Of these, the water flow test and crack closing test are methods that are often used [19–23]. In particular, the water flow test measures changes in the amount of water flowing out through the crack and quantitatively evaluates the extent of crack healing. In the crack closing test, changes in crack width at the surface are measured by optical microscopy to directly determine the extent of crack healing [5]. These two methods allow one to not only evaluate the mechanism of self-healing caused by the further hydration of inorganic binders but also characterize self-healing caused by calcite precipitation and the effects of crack roughness.

Although the abovementioned techniques allow for an overall evaluation of self-healing performance, they are inadequate for the elucidation of individual mechanisms. Water flow and crack closing tests are not well suited for evaluating the self-healing performance caused by further hydration. Therefore, to describe the autogenous healing properties of cementitious materials, it is necessary to examine their further hydration according to the type of inorganic binder and material age.

Further hydration is a mechanism by which unreacted binders of crack surfaces react with water to produce self-healing products that heal cracks. Therefore, the self-healing performance can be evaluated by measuring the amount of heat generated by further hydration. IC is a technique that measures the amount of heat generated by the reaction of cementitious materials. This technique can be used to evaluate the self-healing performance by further hydration by measuring the amount of heat generated by reacting the powder made by grinding hardened cement paste with water. Tittelboom et al. Investigated the further hydration characteristics of SCMs by calorimetry using isothermal calorimetry [8].

In this study, OPC and SCMs such as GGBS, FA, SF, CSA, and CAs were used as inorganic binders to create paste specimens. Hardened cementitious materials of different age were pulverized, and their hydration heat was determined by IC. SEM-EDS was used to analyze the self-healing material components generated via further hydration.

2Experimental program2.1Materials

In addition to OPC, we used CSA as an expansion agent and SCMs such as GGBS, FA, and SF as binders. CaSO4, Al2(SO4)3, and Na2SO4 (sulfate sources) as well as Na2CO3, NaHCO3, and Li2CO3 (carbonate source) were used as CAs to improve further hydration. Table 1 shows the compositions and physical properties of the raw materials used.

Table 1.

Chemical compositions and physical properties of raw materials used in this study.

Composition (wt%)
OPC  GGBS  FA  CSA  SF 
SiO2  16.91  32.54  53.73  3.87  74.91 
Al2O3  3.95  13.34  20.05  9.63  0.26 
Fe2O3  3.12  0.45  5.57  0.78  0.97 
CaO  57.85  42.28  3.36  48.89  0.19 
MgO  1.66  2.52  0.91  0.65  0.86 
K21.24  0.56  1.45  0.20  0.91 
Na20.23  0.21  0.96  –  0.73 
TiO2  0.25  0.57  0.98  0.44  – 
MnO  0.20  0.24  0.07  –  0.13 
P2O5  0.46  –  0.26  0.42  0.03 
SO3  1.71  2.09  0.45  16.41  0.20 
SrO  0.06  0.06  0.10  0.06  0.01 
Physical properties
Density (kg/m33,190  2,780  2,870  2,100  2,200 
Blaine fineness (m2/kg)  388  463  428  348  17,300* 
*

The specific surface area of silica fume was measured with the BET method.

Table 2 shows the results of quantitative analysis performed using X-ray diffraction (XRD) coupled with Rietveld refinement.

Table 2.

Mineralogical phase compositions of the employed raw materials.

  OPC  GGBS  FA  CSA  SF 
Glass phase  –  95.0  74.4  –  99.3 
C362.0  –  –  3.2   
C216.1  –  –  –   
C32.5  –  –  –   
C4AF  12.1  –  –  –   
Gypsum  2.6  –  –  48.6   
Anhydrite  –  2.9  –  –   
Quartz  0.2  2.1  11.3  –  0.2 
SiC          0.4 
Mullite  –  –  13.9  –   
Magnetite  –  –  0.4  –   
CaO  –  –  –  2.0   
Ca(OH)2  –  –  –  16.6   
Mayenite (C12A7–  –  –  1.7   
CSA  –  –  –  28.0   

Table 3 shows the mixing ratios used to fabricate paste samples. The W/B weight ratio was set to 0.4, and distilled water was used to avoid the effects of interfering ions. To investigate the effect of GGBS specific area, GGBSH (7800 cm2/g) was prepared using a ball mill.

Table 3.

Mixture proportions of binders.

Binder (g)
OPC  GGBS  GGBSH  FA  SF  CSA  CaSO4  Al2(SO4)3  Na2SO4  Na2CO3  NaHCO3  Li2CO3 
SH1-1  100                       
SH1-2  60  40                     
SH1-3  40  60                     
SH1-4  60    40                   
SH1-5  55  40                   
SH1-6  65      35                 
SH1-7  50      50                 
SH1-8  90        10               
SH2-1  90          10             
SH2-2  85          10       
SH2-3  85             
SH2-4  85             
SH2-5  85             

The tests broadly consisted of two series of experiments. In the first series, OPC was replaced with SCMs such as GGBS, FA, and SF. In the second series, CSA and CAs were used. Experiments of the first series were performed to investigate the effects of SCM type, replacement percentage, and particle size, while those of the second series were performed to study the effects of crystallization accelerator type.

2.2Test methods

To evaluate the self-healing potential, the amount of heat produced by the further hydration of unreacted binder in the cement paste was measured by IC [8]. The employed TAM-AIR equipment had eight channels, and the amount of heat was measured using ampoules containing a maximum of 20 g of specimen for each channel. To set the initial temperature before tests, the reference temperature was set at 23 °C, and the baseline was stabilized for 2 h with a reference specimen inserted. After the baseline was stabilized, tests were performed twice for each series.

First, tests were performed to measure the reference heat production of the binder according to the mixture proportions in Table 3. In accordance with these proportions, 100 g of the binder and 40 g of distilled water were placed in a glass beaker and mixed using a glass rod. An approximately 5 g specimen of the obtained paste was placed in an ampoule and weighed with an error tolerance of 0.0001 g, and IC was used to record heat production as a function of time. The reference heat production was measured for 144 h. To measure heat production resulting from the further hydration of unreacted binder in the hardened cement paste, 10 mm × 10 mm × 10 mm cement paste specimens were prepared according to the mixture proportions in Table 3. After mixing, samples were cured in a 20 ± 1 °C, 100% relative humidity chamber for 24 h, unmolded, and cured in a 20 ± 1 °C water container until the tests were performed. To examine the effects of specimen age, tests were performed after 7, 28, and 91 days of aging. Specimens of the target age were dried for 6 h in a 40 °C chamber. After drying, specimens were crushed into powder, and crushed powder was filtered by 200-μm sieve. Sieved powder was mixed with tap water (sieved powder : tap water = 10 g : 4 g). A 5 g of the rehydrated paste was placed in an ampoule, and heat production was measured for 144 h by IC.

Self-healing hydrates precipitated on cracks during autogenous healing were examined by SEM. The cracks in cement and concrete specimens were of irregular type, which complicated the collection of the above hydrates from crack faces. To circumvent this problem, we used the “artificial crack method” employed by Huang et al. for self-healing product collection in cracks [17]. Seven-day old cement pastes (prepared using the mixing proportions of Table 3) were used, and artificially cracked specimen were allowed to self-heal in tap water for seven days.

3Results and discussion3.1Isothermal calorimetry

Fig. 1 shows the specific heat flows of SH1-1, SH1-2, and SH1-3, revealing that the largest initial hydration peak was observed for SH1-1, while peaks of SH1-2 and SH1-3 were similar to each other. This behavior was ascribed to the fact that SH1-1 corresponded to 100% OPC and therefore immediately reacted with water and dissolved, while SH1-2 and SH1-3 contained GGBS and hence had a smaller OPC content. Despite their 40 wt% difference in OPC content, SH1-2 and SH1-3 featured similar peaks, which was attributed to the rapid initial dissolution of SH1-3 caused by its very high water to OPC ratio. The increasing duration of the dormant period with increasing GGBS content and the concomitant decrease of minimal heat flow were rationalized by OPC content differences. That is, the water to OPC ratio increased with increasing GGBS content, i.e., the amount of OPC reacting during initial hydration increased, which was reflected by a decrease of subsequent heat flow. In the case of SH1-3, the peak of the acceleration period was very weak, and notable heat evolution was observed again only after 24 h.

Fig. 1.

Specific heat flows of SH1-1, SH1-2, and SH1-3.

(0.09MB).

The intensity of the peak ascribed to sulfate consumption increased with increasing GGBS content [24] and equaled that of the silicate reaction peak at a GGBS content of 40 wt%. As only a late (appearing after 24 h) sulfate depletion peak was observed at a GGBS content of 80 wt%, increasing GGBS content was concluded to strongly affect the identities of hydration products, promoting the formation of aluminates.

Fig. 2 shows the cumulative heat production curves of SH1-1, SH1-2, and SH1-3, demonstrating that cumulative heat release decreased with increasing GGBS content. The highest initial heat production was observed for pure OPC (SH1-1), while similar values were obtained for SH1-2 and SH1-3. Notably, the heat production values of SH1-1 and SH1-3 almost stabilized at 144 h, contrary to the case of SH1-2. The ongoing heat production reflected the continuing hydration of GGBS and was expected to exhibit a cumulative heat value higher than that of pure OPC if the test would be continued. Therefore, in the case of cement composites cracked after three days, the composite with a GGBS content of 40 wt% was expected to exhibit better self-healing performance than OPC.

Fig. 2.

Cumulative heat production in SH1-1, SH1-2, and SH1-3.

(0.09MB).

Fig. 3 shows the specific heat flow owing to further hydration in SH1-1, SH1-2, and SH1-3 samples of different age, demonstrating the absence of a dormant period in all cases. This finding was ascribed to the very high W/B ratio during further hydration, which allowed for the rapid reaction of these components. At a material age of 7 days, the highest heat flow was observed for OPC. However, the heat flow of the GGBS-rich specimen (SH1-3) markedly increased upon aging, i.e., the corresponding long-term self-healing performance concomitantly improved.

Fig. 3.

Further hydration induced specific heat flows for SH1-1, SH1-2, and SH1-3.

(0.14MB).

Fig. 4 shows curves reflecting cumulative heat production caused by further hydration in SH1-1, SH1-2, and SH1-3 at each material age. At a material age of 7 days, the highest cumulative heat production was observed for SH1-2 (40 wt% GGBS) and continued to increase even after 144 h, with similar behavior observed for SH1-1. Conversely, the value of SH1-3 was almost stabilized after 144 h. For SH1-1 and SH1-2, cumulative heat production decreased with increasing material age, while the reverse was true for SH1-3. These results were similar to those presented in Fig. 3. Therefore, sufficiently aged materials with high GGBS content were expected to exhibit excellent self-healing performance (owing to further hydration).

Fig. 4.

Further hydration–induced cumulative heat production in SH1-1, SH1-2, and SH1-3.

(0.14MB).

Fig. 5 shows the effects of GGBS fineness and Na2SO4 addition on heat flow, revealing that initial heat evolution was the fastest for SH1-2. This finding was explained by the fact that the high Blain surface area of this sample (7800 cm2/g) allowed GGBS to be more finely mixed with OPC and thus inhibited the reaction of the latter with the infiltrating water. The corresponding peak of SH1-5 occurred at a later time because in this case, slow OPC dissolution occurred only after that of Na2SO4. Compared to those of SH1-2 and SH1-4, the dormant period of SH1-5 was very short, which was ascribed to the rapid rate of Na2SO4 reaction [25]. Moreover, SH1-5 also exhibited a higher acceleration period slope than the other two specimens. At this point, it is worth mentioning that the admixture of Na2SO4 into the paste mixture is known to form quick-setting cement. SH1-2 and SH1-4 exhibited very similar dormant period and acceleration period behaviors, and the Blaine surface area of GGBS was therefore concluded not to have any particular effect on the initial characteristics of heat production.

Fig. 5.

Specific heat flows of SH1-2, SH1-4, and SH1-5.

(0.09MB).

Fig. 6 shows the cumulative heat production curves of SH1-2, SH1-4, and SH1-5, revealing that although the largest initial heat release was observed for SH1-5, this sample exhibited the smallest heat release after 144 h. Moreover, the heat release curve of SH1-5 tended to converge more quickly than those of the other two specimens, and it was therefore expected that the amount of produced heat would not significantly increase further if the experiment would be continued beyond 144 h. In contrast, it was expected that the cumulative amount of produced heat would continue increasing even after 144 h in the cases of SH1-2 and SH1-4. Finally, the highest amount of ultimately produced heat was observed for SH1-4, which had a larger Blaine surface area.

Fig. 6.

Cumulative heat production curves of SH1-2, SH1-4, and SH1-5.

(0.08MB).

Fig. 7 shows the specific heat flow owing to further hydration for SH1-2, SH1-4, and SH1-5 samples of different age. The experiment was not performed for 91-day aged SH1-5 because of problems encountered when preparing this specimen. The heat flow of SH1-4 exceeded that of SH1-2 at all material ages, although not by much. Moreover, there was an increase in initial heat flow according to the material age. At 7 days, specific heat flow of SH1-5 behaved similarly to SH1-2 and SH1-4 but exhibited a markedly higher heat flow at an aging time of 28 days, which was ascribed to the much higher self-healing rate of 28-day aged SH1-5 compared to those of other specimens of the same age.

Fig. 7.

Further hydration–induced specific heat flows of SH1-2, SH1-4, and SH1-5.

(0.14MB).

Fig. 8 shows the cumulative heat production caused by further hydration of SH1-2, SH1-4, and SH1-5 at each material age, revealing that the cumulative heat production decreased upon progressive material aging and was highest for SH1-4 at all ages. When the Blaine surface area of GGBS was large, reactivity increased even at a relatively low pH. Therefore, the self-healing performance was expected to improve with increasing Blaine surface area. SH1-5 featured the lowest cumulative heat production at 7 days, whereas a value exceeding that of SH1-2 but still smaller than that of SH1-4 was observed at 28 days. Thus, cumulative heat production was expected to increase further if GGBS with a Blaine surface area of 7800 cm2/g would be used.

Fig. 8.

Further hydration–induced cumulative heat production in SH1-2, SH1-4, and SH1-5.

(0.14MB).

Fig. 9 shows the effect of FA content on initial heat flow, revealing that although the presence of FA decreased the intensity of the initial heat release peak, the exact FA content had no marked effect. This trend was similar to that observed in Fig. 1, which highlighted the effect of GGBS content on heat flow and was ascribed to the increase of the water to OPC ratio (and hence, to the concomitant increase of OPC reactivity) with increasing FA content. In addition, the induction period became longer with increasing FA content (also in agreement with the trend observed in Fig. 1). Finally, the slope and peak values of the acceleration period curves were reduced upon FA incorporation, and almost identical decelerating period lengths were observed in all three cases, while the sulfate depletion peak was absent, in full agreement with the results of Hu et al. [26].

Fig. 9.

Specific heat flow curves of SH1-1, SH1-6, and SH1-7.

(0.09MB).

Fig. 10 shows the effects of FA content on cumulative heat production in SH1-1, SH1-6, and SH1-7, demonstrating that the total amount of heat produced up to 144 h decreased with increasing FA content, while the extent of this decrease was proportional to FA content. For SH1-6 (FA content = 35 wt%), the slope of the heat production curve exceeded those observed for the other two samples during the whole 144-h period. Therefore, an increase of the experimental time period was expected to be reflected in increased cumulative heat production owing to the pozzolanic reaction of FA, similar to the phenomenon observed by Belie et al. [8].

Fig. 10.

Cumulative heat production curves of SH1-1, SH1-6, and SH1-7.

(0.09MB).

Fig. 11 shows the flows of specific heat caused by further hydration in SH1-1, SH1-6, and SH1-7 samples of different ages, demonstrating that the highest initial specific heat flow was observed for SH1-1 (pure OPC). This behavior was ascribed to (a) the long time required to establish a pH allowing for the pozzolanic reaction after rehydration with water in FA-containing samples and (b) the low reactivity of FA. Fig. 3, which shows the effect of GGBS content on specific heat flow, demonstrates that the initial reactivity of GGBS-containing specimens increased with increasing material age. Thus, it was concluded that the incorporation of GGBS enhances self-healing properties to a larger extent than the incorporation of FA.

Fig. 11.

Further hydration–induced specific heat flows of SH1-1, SH1-6, and SH1-7.

(0.14MB).

Fig. 12 shows the effect of FA content on cumulative heat production caused by further hydration in SH1-1, SH1-6, and SH1-7 specimens of different age, revealing that the above heat production decreased with increasing age for SH1-1 and SH1-6. For SH1-7, heat production slightly increased at a material age of 28 days but decreased again after 91 days. This finding was ascribed to the high FA content of SH1-7 (50 wt%), i.e., FA that could not react was hydrated, which resulted in increased cumulative heat production. However, after 91 days of aging, OPC was entirely hydrated, and a pH providing pozzolanic reactivity could not be reached. At a material age of 7 days, SH1-6 and SH1-6 featured lower cumulative heat production than SH1-1. However, the slopes of the corresponding curves at the test ending time of 144 h were larger than that of SH1-1, and the related amounts of cumulative heat were expected to significantly increase further if the test would be continued.

Fig. 12.

Further hydration–induced cumulative heat production in SH1-1, SH1-6, and SH1-7.

(0.13MB).

Fig. 13 shows that SF content significantly affected the magnitude of specific heat flow initial peaks, i.e., the intensity of the SH1-8 peak equaled ˜60% of that of the SH1-1 peak. The SF replacement fraction of SH1-8 equaled 10%, i.e., the OPC content of SH1-8 was 90% of that of SH1-1. Therefore, the initial peaks (ascribed to the initial dissolution of OPC) of SH1-8 and SH1-1 were expected to have a 0.9:1 intensity ratio, whereas a much smaller ratio was observed in practice. This behavior was ascribed to the fact that permeated water was quickly absorbed by SF because of its very small particle size (0.1–1 μm) and high specific surface area (200,000 cm2/g, ˜50 times larger than that of OPC). Apart from the case of the initial period, the behaviors of SH1-1 and SH1-8 were almost identical for all other (induction, acceleration, and deceleration) periods.

Fig. 13.

Specific heat flow curves of SH1-1 and SH1-8.

(0.08MB).

Fig. 14 compares the cumulative heat production curves of SH1-1 and SH1-8, revealing that higher heat production was observed for SH1-1 (by 10%, which corresponded to the SF replacement fraction of SH1-8). Therefore, it was concluded that the heat evolution characterized by IC was owing to the OPC content.

Fig. 14.

Further hydration–induced cumulative heat production in SH1-1 and SH1-8.

(0.08MB).

Fig. 15 shows the effect of SF content on the initial heat flow caused by further hydration in SH1-1 and SH1-8 at each material age, revealing that the above samples exhibited similar patterns at all ages. Therefore, the inclusion of SF was concluded to not significantly affect initial self-healing properties.

Fig. 15.

Further hydration–induced specific heat flows of SH1-1 and SH1-8.

(0.12MB).

Curves in Fig. 16 describe the further hydration–induced production of cumulative heat in SH1-1 and SH1-8 samples of different ages. Notably, the cumulative heat amount decreased with increasing age, with maxima observed for SH1-8 at all ages. Thus, SF incorporation was concluded to benefit self-healing performance. This effect was most pronounced at a material age of 7 days, in which case a 60% heat production difference as observed between the two samples. The slopes of the above curves did not reach a stable value and were therefore expected to further increase with time, which probably reflected the concomitant occurrence of additional reactions. Therefore, SF incorporation was concluded to benefit self-healing properties regardless of material age, being particularly effective for initial crack healing.

Fig. 16.

Further hydration–induced cumulative heat production in SH1-1 and SH1-8.

(0.13MB).

Fig. 17 shows the specific heat flow curves of SH1-1 (pure OPC), SH2-1 (10% OPC replaced by CSA), and SH2-2 (15% OPC replaced by [CSA + sulfate] to promote the formation of ettringite) samples. The highest initial peak was observed for SH1-1, reflecting the dissolution of the C3S component of OPC, while much lower peaks were observed for OPC-poorer SH2-1 and SH2-2. The induction period of SH1-1 was very short compared to those of SH2-1 and SH2-2, which was ascribed to the rapid progress of the acceleration period in the last two cases caused by the high sulfate content of the corresponding samples. Moreover, SH2-2 contained additional sulfate and CSA, and thus featured a shorter induction period than SH2-1. Thus, the reaction rate was concluded to increase with increasing sulfate content. Notably, no sulfate depletion peak was observed for the sulfate- and aluminate-rich SH2-2, whereas such a peak could be well observed in the case of SH1-1. Additionally, SH2-2 experienced a very fast heat flow reduction compared to SH1-1 and SH2-1, which was assumed to explain the fact that after 48 h, the hydration of SH2-2 did not progress as far as that of the other two specimens.

Fig. 17.

Specific heat flow curves of SH1-1, SH2-1, and SH2-2.

(0.09MB).

Fig. 18 shows the effects of sulfate/CSA content on cumulative heat production in SH1-1, SH2-1, and SH2-2, revealing that the corresponding correlations were negative. The initial heat production of SH2-2 was higher than those of the other two specimens but decreased with time and finally became the lowest. As shown in Fig. 16, this behavior was ascribed to the fact that the initial heat production was high and was accompanied by rapid hydration, while OPC could not be properly hydrated because of the rapid volume expansion caused by high CSA and sulfate contents.

Fig. 18.

Cumulative heat production curves of SH1-1, SH2-1, and SH2-2.

(0.09MB).

Fig. 19 shows the initial heat flow caused by further hydration in SH-1, SH2-1, and SH2-2 samples at each material age, revealing that these two parameters were negatively correlated. The highest rehydration-induced heat flow at 7 days was observed for OPC, and the extent of heat flow reduction with age was very small. At 7 days, the SH2-1 specimen (which only contained CSA) exhibited a very large initial heat flow that rapidly decreased with increasing age. For SH2-2, which contained both CSA and sulfate, almost identical heat flows were observed after 7 and 28 days but decreased sharply after 91 days. Thus, SH2-1 was expected to show quick self-healing at an early material age, while the rate of the initial reaction was believed to be reduced to that of OPC with increasing age. The initial reaction rate of SH2-2 was similar to that of OPC at material ages of up to 28 days but was much lower than that of OPC after 91 or more days.

Fig. 19.

Further hydration–induced specific heat flows of SH1-1, SH2-1, and SH2-2.

(0.14MB).

Curves in Fig. 20 describe the further hydration–induced production of cumulative heat in SH1-1, SH2-1, and SH2-2 samples of different age, revealing that the initial self-healing performance decreased in the order of SH2-1 > SH2-2 > SH1-1. This behavior was ascribed to the rapid initial production of ettringite in samples where OPC was replaced by CSA and sulfate as the high sulfate content prevented efficient OPC hydration, leaving a significant fraction of OPC in the non-hydrated state. However, OPC hydration progressed with increasing material age. At a material age of more than 7 days, the hydration of CSA occurred more quickly than that of OPC, and at a material age of 28 days, the cumulative heat production was greatly reduced. In contrast, SH2-2 had a higher sulfate content than SH2-1 and therefore did not exhibit any peaks caused by the aluminate reaction owing to sulfate depletion. At a material age of 28 days, the behavior of SH2-2 was very similar to that of OPC. Thus, specimens containing CSA and sulfate were concluded to feature high initial self-healing performance, which, however, became inferior to that of OPC at aging times equal or exceeding 28 days.

Fig. 20.

Further hydration–induced cumulative heat production in SH1-1, SH2-1, and SH2-2.

(0.14MB).

Fig. 21 shows the specific heat flow caused by further hydration in SH2-3, SH2-4, and SH2-5 samples at each material age. Notably, no particular differences between heat flows at 7 and 28 days were observed for SH2-5, while a noticeable heat flow reduction was observed at 91 days. For SH2-3 and SH2-4, the heat flow at 28 days was greatly reduced compared to that at 7 days, while an increase of aging time to 91 days did not result in marked further reduction.

Fig. 21.

Further hydration–induced specific heat flows of SH2-3, SH2-4, and SH2-5.

(0.14MB).

Fig. 22 shows the cumulative heat production caused by further hydration in SH2-3, SH2-4, and SH2-5 samples of different age. For an aging time of 91 days, the heat production of SH2-3 decreased after 48 h, which, however, was ascribed to an error caused by a problem with the testing equipment. For SH2-4, cumulative heat production decreased with increasing material age, whereas for SH2-5, the lowest and highest cumulative heat production was observed at 7 and 28 days, respectively. The lowest cumulative heat production of SH2-4 was observed for a time of 144 h at all material ages. Therefore, among the carbonate-based crystallization accelerators, Li2CO3 was found to be most useful for improving further hydration.

Fig. 22.

Further hydration–induced cumulative heat production in SH2-3, SH2-4, and SH2-5.

(0.14MB).
3.2SEM-EDS

The morphologies of self-healing materials were observed by SEM, while the corresponding elemental compositions were characterized by EDS. The self-healing products of SH1-1 mostly comprised portlandite (Ca(OH)2) and calcite (CaCO3) and were partly amorphous (Fig. 23). EDS mapping revealed the presence of Ca, O, and C at location 1, while Si was additionally detected at location 3. Thus, materials present al locations 1 and 3 were identified as calcite and portlandite/C-S-H, respectively.

Fig. 23.

SEM imaging and EDS analysis results for self-healing products of SH1-1.

(0.52MB).

Fig. 24 shows SEM imaging and EDS analysis results for self-healing products of SH1-2, revealing that the above products could be identified as calcite and an unknown amorphous material. Unlike in the case of SH1-1, no portlandite was observed as it was consumed during the further hydration of unreacted GGBS particles. EDS analysis revealed the presence of Al ions supplied by GGBS, and the abovementioned amorphous material was therefore tentatively identified as C-A-H.

Fig. 24.

SEM imaging and EDS analysis results for self-healing products of SH1-2.

(0.55MB).

Fig. 25 shows SEM imaging and EDS analysis results for self-healing products of SH1-5. Specifically, SEM imaging showed the presence of an amorphous material, calcite, and needle-shaped hydrates as the major self-healing products. Similar to the case of SH1-2, no portlandite was observed as it was believed to be consumed by the further hydration of GGBS. EDS analysis of needle-shaped hydrates revealed the presence of Ca, O, Al, and S, and the above hydrates were thus identified as ettringite. Moreover, amorphous materials were shown to contain Ca, O, C, Si, and Al, and were thus identified as a mixture of C-S-H, C-A-H, and calcite.

Fig. 25.

SEM imaging and EDS analysis results for self-healing products of SH1-5.

(0.57MB).

Fig. 26 shows SEM imaging and EDS analysis results for self-healing products of SH1-6, revealing that the above products were largely amorphous and contained large amounts of Si and Al (originating from the SiO2 and Al2O3 of FA) in addition to Ca, C, and O. C-S-H and C-A-H were formed via the pozzolan reaction between the cement hydration product Ca(OH)2 and the SiO2 and Al2O3 of FA according to the same mechanism during the self-healing process caused by further hydration.

Fig. 26.

SEM imaging and EDS analysis results for self-healing products of SH1-6.

(0.3MB).

Fig. 27 shows SEM imaging and EDS analysis results for self-healing products of SH1-8, demonstrating that these products mainly comprised calcite and an amorphous material. In this case, SiO2, which accounts for most of the SF, dissolved in water to form H4SiO4, which then reacted with Ca2+ and OH ions supplied by cement to form C-S-H. Therefore, SF inclusion was found to reduce the amount of portlandite and increase that of C-S-H. These findings were supported by the observation of Ca, C, O, and Si as the main elements.

Fig. 27.

SEM imaging and EDS analysis results for self-healing products of SH1-8.

(0.58MB).

Fig. 28 shows SEM imaging and EDS analysis results for self-healing products of SH2-2, revealing the presence of abundant acicular crystals mainly containing Ca, C, O, Al, and S. Based on these results, the above crystals were identified as ettringite, which is an expansive hydrate able to fill cracks when mixed with the normal self-healing product (i.e., calcite).

Fig. 28.

SEM imaging and EDS analysis results for self-healing products of SH2-2.

(0.3MB).
4Conclusions

This study used isothermal calorimetry to examine the further hydration of hardened paste specimens containing a range of unreacted inorganic binders (OPC, CSA expansion agent, CAs, and SCMs such as GGBS, FA, and SF). In particular, the heat produced by 10 mm × 10 mm × 10 mm cubic specimens was measured after 7, 28, and 91 days of material age. In addition, SEM-EDS analysis was performed to identify self-healing materials formed inside cracks.

Heat production caused by further hydration decreased with increasing age (except for SH1-3), which was ascribed to the concomitant decrease of the amount of unreacted binder. Thus, the self-healing performance caused by further hydration deteriorated with increasing material age. Long-term heat production increased with increasing GGBS content, and the heat production profiles of FA-containing specimens were similar to that of SH1-1. SF-containing SH1-8 featured higher cumulative heat production compared to SH1-1 at all material ages, although the cumulative heat production of SH1-8 after 28 days was lower than that of GGBS-containing specimens. SH2-1 contained CSA as an expansion agent, while SH2-2 contained sulfate-type CAs. At a material age of 7 days, the heat production of the above two specimens exceeded that of SH1-1 but decreased after 28 days. On the other hand, SH2-3, SH2-4, and SH2-5 contained carbonate-type CAs. The heat production of these specimens at 7 days was smaller than that of the SH1-1 but became equal to the same at 28 days.

SEM-EDS analysis confirmed that the self-healing products of all specimens corresponded to calcite and an amorphous material. Portlandite was found in SH1-1 but not in samples containing SCMs and CSA. In SH1-1 and SH1-8, the amorphous material was identified as C-S-H, whereas C-A-H was formed in samples comprising GGBS and FA. Moreover, ettringite was produced when a crystallization accelerator and a CSA expansion agent were used.

The obtained results confirmed that GGBS facilitates self-healing caused by further hydration at long-term material ages and illustrated that the identity of the self-healing material produced in the cracks depends on the employed binder.

Declarations of interest

None

Acknowledgements

This research was supported by a grant (19SCIP-B103706-05) from Construction Technology Research Program funded by Ministry of Land, Infrastructure and Transport of Korean government. This research also financially supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (no.20181110200070).

References
[1]
A. Bourchy, L. Barnes, L. Bessette, F. Chalencon, A. Joron, J.M. Torrentia.
Optimization of concrete mix design to account for strength and hydration heat in massive concrete structures.
Cem Concr Compos, 103 (2019), pp. 233-241
[2]
A. Teramoto, M. Watanabe, R. Murakami, T. Ohkubo.
Visualization of internal crack growth due to alkali–silica reaction using digital image correlation.
Constr Build Mater, 190 (2018), pp. 851-860
[3]
X.H. Wang, D.V. Val, L. Zheng, M.R. Jones.
Influence of loading and cracks on carbonation of RC elements made of different concrete types.
Constr Build Mater, 164 (2018), pp. 12-28
[4]
F. Pacheco-Torgal, R.E. Melchers, X. Shi, N.D. Belie, K.V. Tittelboom, et al.
Eco-efficient repair and rehabilitation of concrete infrastructures.
[5]
B. Park, Y.C. Choi.
Quantitative evaluation of crack self-healing in cement-based materials by absorption test.
Constr Build Mater, 184 (2018), pp. 1-10
[6]
H. Huang, G. Ye.
Simulation of self-healing by further hydration in cementitious materials.
Cem Conc Compos, 34 (2012), pp. 460-467
[7]
M. Şahmaran, Li V.C. Li.
Durability properties of micro-cracked ECC containing high volumes fly ash.
Cem Concr Res, 39 (2009), pp. 1033-1043
[8]
K.V. Tittelboom, E. Gruyaer, 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 (2013), pp. 349-359
[9]
M. Sahmaran, G. Yildirim, T.K. Erdem.
Self-healing capability of cementitious composites incorporating different supplementary cementitious materials.
Cem Conc Compos, 35 (2013), pp. 89-101
[10]
Y. Abdel-Jawad, F. Dehn.
Self-healing of self-compacting concrete.
SCC2005 Conference Orlando Florida USA, pp. 1023-1029
[11]
S. Granger, A. Loukili, G. Pijaudier-Cabot, G. Chanvillard.
Experimental characterization of the self-healing of cracks in an ultra-high performance cementitious material. mechanical tests and acoustic emission analysis.
Cem Concr Res, 37 (2007), pp. 519-527
[12]
D. Jaroenratanapirom, R. Sahamitmongkol.
Self-crack closing ability of mortar with different additives.
J Met Mater Min, 21 (2011), pp. 9-17
[13]
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
[14]
K. Sisomphon, O. Copuroglu, E.A.B. Koenders.
Effect of exposure conditions on self-healing behavior of strain hardening cementitious composites incorporating various cementitious materials.
Constr Build Mater, 42 (2013), pp. 217-224
[15]
M.A.A. Sherir, K.M.A. Hossain, M. Lachemi.
The influence of MgO-type expansive agent incorporated in self-healing system of engineered cementitious composites.
Constr Build Mater, 149 (2017), pp. 164-185
[16]
A. Darquennes, K. Olivier, F. Benboudjema, R. Gagné.
Self-healing at early-age, a way to improve the chloride resistance of blast-furnace slag cementitious materials.
Constr Build Mater, 113 (2016), pp. 1017-1028
[17]
H. Huang, G. Ye, D. Damidot.
Characterization and quantification of self-healing behaviors of microcracks due to further hydration in cement paste.
Cem Concr Res, 52 (2013), pp. 71-81
[18]
H. Huang, G. Ye, D. Damidot.
Effect of blast furnace slag on self-healing of microcracks in cementitious materials.
Cem Concr Res, 60 (2014), pp. 68-82
[19]
V.C. Li, Y.M. Lim, Y.W. Chan.
Feasibility study of a passive smart self-healing cementitious composite.
Compos Part B Eng, 29 (1998), pp. 819-827
[20]
S. Sangadji, E. Schlangen.
Self-healing of concrete structures – novel approach using porous network concrete.
J Adv Concr Technol, 10 (2012), pp. 185-194
[21]
A. Jefferson, C. Joseph, R.J. Lark, B. Isaacs, S. Dunn, B. Weager.
A new system for crack closure of cementitious materials using shrinkable polymers.
Cem Concr Res, 40 (2010), pp. 795-801
[22]
Z. Yang, J. Hollar, X. He, X. Shi.
A self-healing cementitious composite using oil core/silica gel shell microcapsules.
Cem Concr Compos, 33 (2011), pp. 506-512
[23]
R.H. Haddad, M.A. Bsoul.
Self-healing of polypropylene fiber reinforced concrete: pozzolan effect.
Civil Engineering Department, Jordan University of Science and Technology, (1999),
[24]
G. Xu, Q. Tian, J. Miao, J. Liu.
Early-age hydration and mechanical properties of high volume slag and fly ash concrete at different curing temperatures.
Constr Build Mater, 149 (2017), pp. 367-377
[25]
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
[26]
X. Hu, C. Shi, Z. Shi, B. Tong, D. Wang.
Early age shrinkage and heat of hydration of cement-fly ash-slag ternary blends.
Constr Build Mater, 153 (2017), pp. 857-865
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.