autogeneous self healing of high performance fibre...

High Performance Fiber Reinforced Cement Composites (HPFRCC7), Stuttgart, Germany – June 1-3, 2015

71

AUTOGENEOUS SELF HEALING OF HIGH PERFORMANCE FIBRE REINFORCED CEMENTITIOUS COMPOSITES

L. Ferrara (1), M. Geminiani (1), R. Gorlezza (1), V. Krelani (1), M. Roig-Flores (2), G. Sanchez-Arevalo and P. Serna Ros (2)

(1) Department of Civil and Environmental Engineering, Politecnico di Milano, Italy

(2) Institute of Concrete Science and Technology (ICITECH), Universitat Politècnica de València, Spain

Abstract In this paper the results are shown of a thorough characterization of the self-healing

capacity of High Performance Fibre Reinforced Cementitious Composites (HPFRCCs). In detail, the capacity of the material to completely or partially re-seal cracks will be investigated, as a function of its composition, maximum crack opening and exposure conditions. The topic has been investigated including the effect of different flow-induced alignment of fibres, which can result into an either a strain hardening or softening behaviour, whether the material is stressed parallel or perpendicularly to the fibres. Specimens were initially pre-cracked in 4-point bending and up to different values of crack openings, and submitted to different exposure conditions, including water immersion, exposure to humid or dry air, and wet-and-dry cycles. After scheduled exposure times, from one to six months, specimens were tested up to failure according to the same test set-up and outcomes of the self-healing, if any, were quantified in terms of recovery of stiffness, strength and ductility. In a durability-based design framework, self-healing indices to quantify the recovery of mechanical proper-ties were also defined and their significance checked.

1. INTRODUCTION

The fundamental idea underlying the concept of HPFRCCs is that, once a crack if formed in the matrix and fibres crossing it start working, the energy required to pull out the fibres at the cracked section must be higher than the energy required to form a crack at a new position. The iteration of this concept/mechanism up to complete saturation of the crack spacing results, before the localization into a single unstable crack, in a stable multi-cracking process in which the opening of each single crack is very effectively controlled and restrained thanks to the bridging effect provided by the fibres. The aforementioned stable multi-cracking process is associated to a unique “plastic” or even strain hardening behaviour in direct tension, as well as to a deflection hardening behaviour in bending, which is likely to bring substantial innovation into concept and design of engineering structures.


72

The composition of HPFRCCs, which can yield the mechanical behaviour explained above, is characterized by low maximum aggregate size, high cement and binder content for likewise high compactness of the matrix, low water/binder ratio and a fibre volume fraction usually higher than 1%. Because of the high binder content and low water/binder ratio, it is likely that, even after quite longer aging, a significant amount of the same binder phase remains un-hydrated. Once the cracks form the un-hydrated binder materials, which generally remain as such in the inner part of a structural element or building component, may come in contact with water, even simply in the form of air moisture, and undergo delayed hydration reactions whose products, precipitating onto the crack surfaces, can be able to seal it and even heal it, i.e. provide some recovery of the pristine level of performance in terms of engineering and mechanical properties (i.e. permeability, strength and stiffness, etc.).

A few investigations were performed in the past on the self-healing capacity of ordinary Fibre Reinforced Concrete (FRC), with different types of fibres and under different exposure conditions, demonstrating the ability of the tested specimens to recover their strength and stiffness, at different levels depending on the aforementioned variables. Interestingly, it was also observed that in the case of higher volume percentages of fibres, the same fibres were able to better promote self-healing in the sense that they constitute a network supporting the formation and increase of crystals between the faces of the cracks.

Yang et al. [1], investigating the self-healing capacity of Engineered Cementitious Composites (ECCs) subjected to different wet and dry conditioning regimes, found that if crack opening is kept below 150 m, and even better if below 50 m, an even complete recovery of the mechanical performance, in terms of strength and ductility can be obtained. Moreover, higher temperatures in the drying stages of the cycles, worsened the healing capacity. Tziviloglou [2] and Antonopoulos [3] studied combined synergistic effects of fibres and Super Absorbent Polymers (SAPs) in ECCs, whereas both Sahmaran et al. [4] and Li et al. [5] found that even in aggressive conditions, such as high chloride atmosphere or water, the autogeneous healing capacity of ECCs is maintained, at levels depending on the crack width and aggressiveness of the environment.

In most recent years, a few studies have also been published on the effect that natural fibres, either alone or in combination with other types of fibre reinforcement, may have on the self-healing capacity of HPFRCCs, mainly in the case of exposure to wet and dry cycles [6]. It was found that, due to their porous structure, natural fibres can absorb water during the wetting stages of the cycles and then release it and diffuse throughout the matrix during the drying stages, thus better and more effectively promoting the healing processes.

In this study, a thorough investigation of the autogeneous healing capacity of a typical HPFRCC mix, containing 100 kg/m3 (1.28% by volume) of short straight steel fibres has been performed, considering different exposure conditions. As a distinctive feature of this study, the influence has been investigated of the flow induced alignment of the fibres on the material behaviour, either deflection hardening or softening, and on the related self-healing capacity of the cementitious composite. By means of four-point bending tests performed on specimens in the pre-cracked and post-conditioning stages, the recovery of load bearing capacity, ductility and stiffness has been evaluated, as affected by crack opening, in the deflection softening or hardening behaviour. Suitable healing indicators for the recovery of the aforementioned properties have been defined [7], quantified and related, in a design oriented durability based framework, to an index of crack healing, also herein defined through a tailored methodology.


73

Table 1: Mix design of the employed HPFRCC

Constituent Dosage (kg/m3)Cement 600

Slag 500 Sand (0-2 mm) 982

Water 200 Superplasticizer 33 (lt/m3)

Straight steel fibers (lf = 13 mm; df = 0.16 mm) 100

2. EXPERIMENTAL PROGRAMME

The composition of the HPFRCC employed in the present investigation is shown in Table 1. Slabs 30 mm thick, 1m long and 0.5 m wide were casted. Fibre-reinforced material was poured directly from the mixer onto a chute pouring the material at one short edge of the moulds, and allowing it to flow parallel to the long sides (Figure 1). From the slabs, once hardened, beam specimens 100 mm wide and 500 mm long were cut, according to the schematic also shown in Figure 1, to be tested in 4-point bending. The beam specimens were cut from the slabs so that their axis, and hence the direction of the principal tensile stresses due to the bending action to be applied during the tests, was either parallel or perpendicular to the flow direction of the fresh concrete, along which the fibres are aligned [6-8].

After one to two or eleven months aging in lab environment, beam specimens were tested in 4-point bending, according to the set-up shown in Figure 2. Results of typical tests on specimens bent parallel or perpendicular to the preferential fibre alignment are shown in Figure 3, in terms of nominal bending stress vs. COD curves: the material evidently features a deflection hardening, with multiple cracking, or softening, with single crack, behaviour whether stressed parallel or orthogonal to the flow induced alignment of the fibres.

Specimens were pre-cracked, up to different openings as related to their softening or hardening behaviour, and then submitted to different exposure conditions for different exposure durations, up to six months. A synopsis of the experimental programme, with details of pre-crack opening and exposure conditions and durations and test specimen repeatability is given in Table 2. After the aforementioned scheduled exposure times, specimens were removed from the conditioning environment, wiped, in case, and tested again up to failure according to the same set-up shown in Figure 2. “Superposition” between pre-cracking and post-conditioning N-COD curves allowed self-healing capacity and its effects on mechanical performance of the material to be evaluated, as a function of the testing variables, as it will be described and analysed in detail in forthcoming sections.

slab castingflow induced

fiber alignment slab cutting

4pb testing

Figure 1: Slab casting scheme and beam specimen cutting procedure


74

500 mm150 mm

450 mm

200 mm

A

A sect. A-A

150 mm

30 mm7 mm

Figure 2: 4-point bending tests set-up for beam specimens obtained as in Figure 1

Table 2: Synopsis of experimental programme

Softening Hardening Pre-crack opening 0.5 mm 1 mm 2 mm CODpeak+ 0.5

mm Exposure duration (months)

Exposure conditions Pre-crack age 1 6 12 1 6 12 1 6 12 1 6 12

Water immersion 2 m 1 1 1 1 1 1 2 1 1 2 2 2 11 m 3 2 3 1 1 = 1 1 1 1 1 1

Air exposure 2 m 1 2 1 1 1 1 2 1 1 1 2 1 20°C – RH = 95% 2 m 1 2 1 1 1 1 2 1 1 2 2 2 20°C – RH = 50% 2 m 1 2 2 1 1 1 2 2 2 2 2 2

Wet and dry 2 m 1 2 2 1 1 1 2 2 2 2 2 2

0 2 4 6COD (mm)

0

2

4

6

8

10

N (N

/mm

2 )

0.5 mm

(a)

(b)

0 2 4 6COD (mm)

0

10

20

30

N (N

/mm

2 )

CODpeak + 0.5 mm

CODpeak1 mm

(c)

(d)

Figure 3: Nominal stress N - COD curves (a, c) and crack patterns (b, d) for specimens featuring a deflection softening (a, b) and deflection hardening (c, d) behavior

100 mm


75

3. EXPERIMENTAL RESULTS: ANALYSIS AND DISCUSSION

In Figure 4 a-d some example results, in terms of nominal bending stress N vs. COD curves are shown, for both deflection hardening and deflection softening specimens and selected representative exposure conditions (water immersion from one up to two months). A recovery of the load bearing capacity, with respect to the reference case of a specimen monotonically tested up to failure, can be immediately appreciated. A similar behaviour, even if the amount of recovery obviously depended on the exposure condition, was obtained in all investigated cases.

In order to provide a quantitative evaluation of the effects of the occurred self healing, if any, on the recovery of the mechanical properties, mainly with reference to the load bearing capacity, the Self Healing Indices defined as hereafter defined have been calculated: - for deflection softening specimens and deflection hardening specimens pre-cracked up to

0.5 mm beyond the peak:

Index of stress recovery ISR = unloading,

unloading , -

Ncrackingpreft

Nngconditioniafterpeak

f

f

(1)

- for deflection hardening specimens pre-cracked in the pre-peak regime:

Index of stress recovery ISR =

1 -

crack-pre unloading N,

virginunloading,N,

virginunloading, N,,

crack-pre unloainhg N,

virginfpeak

ngconditionipostpeak

f

f

(2)

with the notation meaning explained in Figure 5. The values of the Index of Stress Recovery calculated as above have been plotted, vs. the

exposure duration and for all the different investigated cases, in Figure 6. In all cases, with due differences for the absolute values of the indices which also take into

account the specificity of each single behavior, it can be evidently observed that: - specimens immersed in water together with those exposed to natural environment (in a

quite humid climate, like the Northern Italy one), featured, in average, the highest, and a quite similar actually, recovery trend; remarkably, even deflection softening specimens which were pre-cracked up to the quite large 0.5mm crack opening, were able to gain, in the post-conditioning stage a strength even higher that the cracking strength of the virgin specimen;

- significantly, specimens pre-cracked at 11 months and immersed in water, even continued to show some moderate healing capacity, at a moderately increasing trend with prolonged immersion;

- specimens exposed to a humid environment featuring an appreciable healing rate since from earlier immersion times, but which did not show any significant improvement with prolonged exposure time;

- as expectable specimens exposed to a dry environment featured an almost negligible healing, even if somewhat increasing with prolonged exposure time;

- quite poor, and again as coherently expectable, intermediate between the one of specimens in dry environment, was the performance exhibited by the specimens subjected to the wet and dry cycles.


76

0 2 4 6 8COD (mm)

0

4

8

12

(

MP

a)

water

pre-crack 0,5 mm

after 1 month

pre-crack 0,5 mm

after 6 months

reference

0 2 4 6 8 10COD (mm)

0

8

16

24

32

(

MP

a)

water

pre-crack 2 mm

after 1 month

pre-crack 2 mm

after 6 months

pre-crack 2 mm

reference

0 2 4 6 8 10COD (mm)

0

8

16

24

32

(

MP

a)

water

pre-crack 0.5mm A.P.

after 1 month

pre-crack 0.5mm A.P.

after 6 months

reference

(a) (b) (c)

Figure 4: Nominal stress N - COD curves for deflection softening (a) and hardening speci-mens pre-cracked at 2 mm (b), 0.5 mm after the peak (d) immersed in water up to 6 months

fpeak, after conditioning

σN,unloading

fpeak,pre‐crack

Pre‐crack

Post‐Conditioning

COD [mm]

N[N/m

m2 ]

(a) (b)

Figure 5: Significance and notation of Index of Stress Recovery

0 4 8 12conditioning time (months)

0

0.4

0.8

1.2

1.6

2

ISR

water (2 months pre-crack)water (11 months pre-crack)airhumid chamber (RH 90%)dry chamber (RH 50%)wet/dry cycles (water/50%)


0

4

8

12

16

ISR



-10

0

10

20

ISR


(a) (b) (c)

Figure 6: ISR for deflection softening (a) and deflection hardening specimens pre-cracked up to 2 mm (b) and up to 0.5 mm beyond the peak (c)

It is worth remarking that in the case of deflection hardening specimens the values of the computed Index of Stress Recovery are actually quite high; this is “inborn” in the definition of the Index itself and also considering that for the measured deflection hardening performance the two pre-crack openings (2 mm and 0.5 mm beyond the peak) are both very close to the peak fall in a range where a moderate increase of the stress corresponds to even significant crack opening increase.


77

From the experimentally measured evolution of stiffness all along the load path values of the scalar internal damage variable were calculated and damage evolution curves were built, the effects of healing resulting (Figure 7a) in a general “slowering” of the damage growth. By comparing the damage curves in the pre-cracking and in the post-conditioning stages the crack healing could be estimated (Figure 7a) and an Index of Crack Healing was defined:

Index of Crack Healing ICH = 1 - cracking-pre

ngconditionipost

COD

COD (6)

0

0.2

0.4

0.6

0.8

1D

amag

e pre-crackpost-treatment

(Dj)(Di)

COD (mm)CODpost-conditioning

Crack Closure

CODpre-cracking

0 0.2 0.4 0.6 0.8 1ICH

0

2

4

6

8

ISR

water (2 months pre-crack)water (11 months pre-crack)airhumid chamber (RH 90%)dry chamber (RH 50%)wet/dry cycles

black solid markers - defl. softeninggrey solid markers - defl. hard. 1 mm pre-crackblack hollow markers - defl. hard. 2 mm pre-crackgrey hollow markers - defl. hard. 0.5 mm after the peak

Figure 7: Example of damage evolution curves with ICH estimation (a) and ISR vs. ICH (b)

The ISR vs. ICH plots, shown in Figure 7b highlight, despite the high sparsity of the results a common trend confirming the increasing recovery of the load bearing capacity for closer and closer cracks. The influence of exposure conditions and of the pre-crack opening can also be appreciated. It is worth reminding that for deflection hardening specimens for an integral crack opening equal to e.g. 2 mm, the opening of each single crack, because of the obtained multi-cracking pattern and considering the employed pre-cracking set-up, is roughly equal to 150 m. This hence results in the highest observed crack closure and recovery performance resulting from a synergistic compromise between the larger cluster of unhydrated particles exposed and the sealability/healability of the crack. Pictures of healed cracks in Figures 8, 9 with values of calculated ISR and ICH provide visual confirmation to these statements but also highlight the need to further investigate the effects of exposure conditions on the quality of the self-healing products and hence on the amount of crack sealing and mechanical healing.

1 month in water

(ISR 0.93 – ICH 0.9)1 month air

(ISR 0.48 – ICH 0.4)1 month 90% RH

(ISR 0.64 – ICH 0.64)1 month 50% RH

(ISR 0.11 – ICH 0.2)1 month wet/dry

(ISR 0.42 – ICH 0.52)

Figure 8: Healed cracks for deflection softening specimens after one month exposure to different conditioning environments (with related values of ISR and ICH)


78

1 month in water (ISR 3.3 – ICH 0.88)

1 month air (ISR 1.78 – ICH 0.7)

1 month 90% RH (ISR 1.8 –ICH 0.68)

1 month 50% RH (ISR 1.3 – ICH 0.51)

1 month wet/dry (ISR 2.7 – ICH 0.75)

Figure 9: Healed cracks for deflection hardening specimens precracked at 2 mm after one month exposure to different conditioning environments (with related values of ISR and ICH)

4. CONCLUSIONS

HPFRCCs possess an autogeneous self healing capacity due to the positive synergy between the crack closure effect provided by the fibres and the peculiar material composition, featuring high cement/binder content and low water binder ratio. Cracking expose clusters of binder particles remained unhydrated because of the mix composition as above to outdoor moisture or water and this makes possible delayed hydration reactions whose products seal and heal the cracks. A three-stage dedicated experimental methodology has been employed in this paper to investigate the effects of crack healing in HPFRCCs on the recovery of mechanical performance. Recovery of the load bearing was observed and consistently correlated to a crack closure estimate, based on a tailored procedure based on damage evolution curves. This adds interesting value to the whole life-cycle performance of structures made of or retrofitted with this category of advanced cement based composites.

REFERENCES

[1] Yang, Y., Lepech, M.D., Yang, E.H. and Li, V.C., ‘Autogenous healing of ECCs under wet-dry cycles’, Cem. Concr. Res., 39 (2009) 382-390.

[2] Tziviloglou, E., ‘Self healing in ECC materials with low content of different micro-fibers and micro-particles’, PhD thesis, Delft University of technology, 2009.

[3] Antonopoulou, S., ‘Self-healing in ECC materials with high content of different micro-fibres and micro-particles’, MSc thesis, Delft University of Technology, 2009.

[4] Sahmaran, M., Li, M. and Li, V. C. ‘Transport Properties of Engineered Cementitious Composites under Chloride Exposure,’ ACI Mater. J., 104 (2008) 303–310.

[5] Li, V.C., Lim, Y.M. and Chan, Y.W., ‘Feasibility study of a passive smart self-healing cementitious composite’, Compos. Part B Eng., 29 (1998) 819–827.

[6] Ferrara, L., Ferreira, S.R., Krelani, V., Silva, F. and Toledo Filho, R.D., ‘Effect of natural fibres on the self-healing capacity of HPFRCCs", Proceedings SHCC3, Dordrecht, The Netherlands, November 3-5, 2014 9-16.

[7] Ferrara, L., Krelani, V. and Carsana, M, ‘A fracture testing based approach to assess the effects of crack healing on the recovery of mechanical properties of concrete with and without crystalline admixtures’, Constr. & Build. Mats., 68 (2014) 515-531.

autogeneous self healing of high performance fibre...

Documents