Neutron Radiography Based Visualization and ... - Semantic Scholar

2 downloads 0 Views 10MB Size Report
Apr 26, 2016 - polyurethane-based healing agents to ensure (multiple) crack ... radiography is normally the best technique to visualize the moisture .... self-healing properties, the metal plates were removed after the curing period of 28 days, ...
materials Article

Neutron Radiography Based Visualization and Profiling of Water Uptake in (Un)cracked and Autonomously Healed Cementitious Materials Philip Van den Heede 1,2 , Bjorn Van Belleghem 1,2 , Natalia Alderete 1,3 , Kim Van Tittelboom 1 and Nele De Belie 1, * 1

2 3

*

Magnel Laboratory for Concrete Research, Department of Structural Engineering, Faculty of Engineering and Architecture, Ghent University, Technologiepark Zwijnaarde 904, Ghent B-9052, Belgium; [email protected] (P.V.d.H.); [email protected] (B.V.B.); [email protected] (N.A.); [email protected] (K.V.T.) Strategic Initiative Materials (SIM vzw), project ISHECO within the program ‘SHE’, Technologiepark Zwijnaarde 935, Ghent B-9052, Belgium LEMIT, Laboratory for Multidisciplinary Training in Technological Research, 52 entre 121 y 122 s/n, La Plata 1900, Argentina Correspondence: [email protected]; Tel.: +32-9-264-5522

Academic Editor: Hong Wong Received: 23 February 2016; Accepted: 20 April 2016; Published: 26 April 2016

Abstract: Given their low tensile strength, cement-based materials are very susceptible to cracking. These cracks serve as preferential pathways for corrosion inducing substances. For large concrete infrastructure works, currently available time-consuming manual repair techniques are not always an option. Often, one simply cannot reach the damaged areas and when making those areas accessible anyway (e.g., by redirecting traffic), the economic impacts involved would be enormous. Under those circumstances, it might be useful to have concrete with an embedded autonomous healing mechanism. In this paper, the effectiveness of incorporating encapsulated high and low viscosity polyurethane-based healing agents to ensure (multiple) crack healing has been investigated by means of capillary absorption tests on mortar while monitoring the time-dependent water ingress with neutron radiography. Overall visual interpretation and water front/sample cross-section area ratios as well as water profiles representing the area around the crack and their integrals do not show a preference for the high or low viscosity healing agent. Another observation is that in presence of two cracks, only one is properly healed, especially when using the latter healing agent. Exposure to water immediately after release of the healing agent stimulates the foaming reaction of the polyurethane and ensures a better crack closure. Keywords: neutron radiography; capillary water absorption; mortar; cracks; autonomous self-healing; encapsulated polyurethane; viscosity

1. Introduction Concrete is vulnerable to crack formation due to its low tensile strength. Cracks in concrete may appear due to several causes such as drying shrinkage, external loading, temperature gradients, freeze–thaw action, etc. The presence of cracks leads to an accelerated ingress of water and other aggressive substances deep into the concrete matrix. In this way, harmful substances transported by the water reach the steel reinforcement very fast, which can lead to accelerated corrosion initiation and hence major durability problems. To avoid problems regarding concrete deterioration and reinforcement corrosion, cracks need to be repaired as soon as possible. However, repair works require high direct and indirect costs and Materials 2016, 9, 311; doi:10.3390/ma9050311

www.mdpi.com/journal/materials

Materials 2016, 9, 311

2 of 28

some cracked regions in concrete elements are not even visible or accessible. Therefore, it would be beneficial to give the concrete the ability to heal the cracks by itself without any human intervention. Different methods have already been explored to obtain autonomous crack healing in concrete [1]. One of the promising approaches is the embedment of brittle capsules filled with healing agents inside the concrete matrix [2–4]. Crack formation leads to capsule breakage and release of the healing agent, which fills up the crack and forms a barrier which prevents water ingress through the cracks. Evaluation of the efficiency of the autonomous healing mechanism to prevent ingress of water through cracks can be done by water permeability tests [2,5,6] or gravimetrical water absorption tests [7–9]. Both test methods were originally designed to study uncracked, cracked and repaired cementitious materials [6,7,10–18]. From these tests the water permeability coefficient or sorption coefficient of specimens with healed cracks can be compared to specimens with unhealed cracks as well. This provides useful information regarding the healing efficiency. However, detailed information about the water ingress through healed cracks is not obtained in this way. In order to identify the difference in the development of the water distribution between sound, cracked and healed mortar or concrete, it is necessary to visualize the water flow. In this way, it becomes possible to evaluate whether the autonomous crack healing mechanism provides full sealing of the crack or whether there is still a part of the crack where water can enter. By visualizing the water uptake phenomenon over time, it also becomes possible to investigate whether the crack healing keeps preventing water ingress through the crack for long exposure periods to water. Interesting contributions regarding the visualization of the initial water uptake (up to 7 s) were published by Gardner et al. who studied the capillary rise heights for discrete cracks present in cementitious materials using a high speed digital camera [19]. Nevertheless, this technique cannot provide information regarding the water uptake in the matrix around the crack for longer exposure periods. Radiation attenuation techniques such as X-ray and neutron radiography can be used for visualizing this water uptake in porous materials [20–26]. Both techniques are efficient to visualize water uptake processes, but X-rays interact differently with matter than neutrons. Heavy materials like steel or lead induce strong X-ray attenuation, whereas light materials like water or plastics result in a strong neutron attenuation. Since neutrons interact stronger with hydrogen than X-rays, neutron radiography is normally the best technique to visualize the moisture distribution in mortar samples. Capillary water absorption in sound and cracked mortar and concrete has already been investigated in several studies by X-ray [9,27–29] or neutron radiography [17,30–36] experiments. In the research of Van Tittelboom et al. [30], the self-healing efficiency of mortar with encapsulated healing agents was visualized for the first time by neutron radiography. From the moisture distribution profiles on the neutron radiographs it was found that encapsulation of polyurethane-based healing agents proved to be very efficient for autonomous healing of single cracks. The main aim of the current study was to investigate the efficiency of two polyurethane-based healing agents with different viscosity to prevent water ingress through cracks in mortar. Not only the effect of a single healed crack was studied in this research, also specimens with two cracks were produced to investigate the possibility of multiple crack healing. Moreover, in previous laboratory experiments, some time passed between triggering the autonomous crack healing mechanism and testing of the crack healing efficiency to ensure hardening of the healing agent. However, in real structures cracks can appear during periods where the structure is already exposed to water. Therefore, the effect of instant autonomous crack healing during water exposure was also studied. 2. Materials 2.1. Mortar Mortar with a water-to-cement ratio of 0.5 and a sand-to-cement ratio of 3 was made. Sand with grain sizes ranging from 0 to 2 mm was used in combination with Ordinary Portland cement CEM I

Materials 2016, 9, 311

3 of 28

52.5 N. The proportioning of the different constituents per m³ of mortar was as follows: 1535 kg/m3 sand, 512 kg/m3 cement and 256 kg/m3 water. 2.2. Capsules In order to encapsulate the selected healing agents, borosilicate glass capsules with a length of 30 and 50 mm, an internal diameter of 3.00 mm and an external diameter of 3.35 mm were chosen. As glass is a very brittle material, breakage of the capsules upon crack creation will be assured. 2.3. Healing Agent Two different types of polyurethane-based healing agents were selected for this study. As the viscosity of the healing agent is a very important parameter that determines whether the healing agent will flow out of the capsules and fill the crack, healing agents with different viscosities were chosen. The first agent was developed within the framework of the SHEcon project (Self-healing concrete for structural and architectural applications) where a polyurethane-based healing agent was developed in order to meet the needs for self-healing of thermally induced cracks in concrete sandwich panels [37,38]. It has a viscosity of 6700 mPas at 25 ˝ C. It is a polyurethane (PU) precursor that essentially consists of methylene diphenyl diisocyanate (MDI) and a polyether polyol. This precursor reacts with water to create the foam that heals the cracks. The moisture content of the concrete itself counts as the main water source. The second healing agent is a commercially available healing agent named Flex SLV AF from the company De Neef Conchem with a much lower viscosity of about 200 mPas at 25 ˝ C. It is a MDI and polyether polyol based prepolymer. As such, the backbone of the product is an incomplete PU polymer containing residual isocyanate (-NCO) groups that react with the water present in the mortar upon capsule breakage to form a waterproof PU polymer. The healing agent is free of volatile organic compounds, does not contain catalysts or any water-soluble products. The viscosity and rheological behavior upon release from the capsules is mainly controlled by the presence of inert, hydrophobic compounds. The polyurethane-based healing agent with the highest viscosity is named PU_HV and the one with the lowest viscosity PU_LV. Both healing agents are one-component healing agents that react upon contact with moisture in the matrix. Upon their reaction, foaming occurs which causes expansion of the healing agent. This is desirable to fill the crack space completely and assure that the crack is sealed against the ingress of aggressive substances. 2.4. Mortar Prisms with(out) Encapsulated Healing Agent Table 1 gives an overview of all studied test series of mortar prisms together with their nomenclature and the number of specimens per test series. Prisms with dimensions of 40 ˆ 40 ˆ 160 mm3 were prepared by using wooden molds. In order to create standardized cracks (see Section 3.1), thin metal plates were positioned in the molds up to a depth of 20 mm. These plates had a width of 40 mm, a thickness of 300 µm and were fixed at their position by means of a metal framework that was connected to the mold. In order to evaluate the healing efficiency of the proposed self-healing mechanism, one test series without inserted metal plates, and thus without standard cracks, was provided (UN). The samples of all other test series did contain one or two metal plates and thus one or two standard cracks. For samples with only one crack (CR_1), the metal plate was positioned at the center of the samples’ length. For samples with two cracks (CR_2), both plates were positioned in parallel with an intermediate distance of 20 mm symmetrically, with regard to the center of the sample. Samples with self-healing properties were obtained by embedding encapsulated healing agent in the matrix at the position where the crack would appear. In order to trigger the healing mechanism at the moment of crack appearance, a method described by Van Belleghem et al. [9] was used. Since standardized cracks were created, the capsules were put through the metal plates used for crack creation. Therefore, two holes with a diameter of 3.50 mm were drilled in the metal plates that were used to create the cracks in the samples with self-healing properties. The capsule’s position (and thus also the position of the holes) was chosen in such a way that there was a mortar cover on each capsule

Materials 2016, 9, 311

4 of 28

of 10 mm at the side and the bottom (top in Figure 1) of the sample. The 30 and 50 mm length capsules Materials 2016, 9, 311 4 of 28 were introduced in the samples containing one and two plates, respectively, and positioned through the holes inside thefixed metalby plate(s). capsules werewires, placedwhich through theconnected holes in the their their position was gluing Once them the onto thin plastic were to plates, the walls of position was fixed by gluing them onto thin plastic wires, which were connected to the walls of the the mold (Figure 1). mold (Figure 1). Table 1. Test series under investigation (n = number of specimens). Table 1. Test series under investigation (n = number of specimens).

Code Code UN CR_1 UN CR_1 CR_2

Description Test Series n Description Test Series n 2a Uncracked mortar samples Mortar Uncracked samples with onesamples standard crack mortar 2a3a Mortar samples with one standard crack 3a3a Mortar samples with two standard cracks a CR_2 Mortar samples with two standard cracks Mortar samples with one standard crack autonomously 3 a CR_1_PU_HV 3 +2b Mortar samples with one standard crack autonomously CR_1_PU_HV 3a+2b healed with high viscosity healing agent healed with high viscosity healing agent Mortar onestandard standardcrack crack autonomously a a b b Mortarsamples samples with with one autonomously CR_1_PU_LV CR_1_PU_LV 3 3 + 2+ 2 healed with with low agent healed lowviscosity viscosityhealing healing agent Mortar samples with two standard cracks autonomously CR_2_PU_HV Mortar samples with two standard cracks autonomously 4 a a CR_2_PU_HV 4 healed with high viscosity healing agent healed with high viscosity healing agent Mortar samples with two standard cracks autonomously CR_2_PU_LV Mortar samples with two standard cracks autonomously 4 a healed with low viscosity healing agent 4a CR_2_PU_LV healed with low viscosity healing agent a exposed to water after b exposed PU hardening; to water before PU hardening. a

exposed to water after PU hardening; b exposed to water before PU hardening.

(a) 30 mm long capsule s with he aling age nt

(b) Me tal plate

50 mm long capsule s with he aling age nt

Figure one single single crack crack (a); (a); Figure 1. 1. Mortar Mortar samples samples with with autonomous autonomous crack crack healing healing properties properties containing: containing: one and and two two cracks cracks (b). (b).

Once molds was finished, samples werewere mademade by filling the molds mortar Once preparation preparationofofthe the molds was finished, samples by filling the with molds with in two equal layers. Each layer was compacted through vibration on a vibrating table. After casting, mortar in two equal layers. Each layer was compacted through vibration on a vibrating table. After the samples were stored instored an air-conditioned room at aroom temperature of 20 ˝ C and a relative casting, the samples were in an air-conditioned at a temperature of 20 °C and humidity a relative of more than 95%. Twenty-four hours later, samples were demolded and then stored again in the again same humidity of more than 95%. Twenty-four hours later, samples were demolded and then stored air-conditioned room until the age of 28 days. in the same air-conditioned room until the age of 28 days. 3. Methods 3. Methods 3.1. Crack Creation and Healing 3.1. Crack Creation and Healing For the samples containing cracks without healing agent (CR_1 and CR_2), the metal plates were For the samples containing cracks without healing agent (CR_1 and CR_2), the metal plates were removed from the mortar matrix at the moment of demolding. In this way, standard cracks with removed from the mortar matrix at the moment of demolding. In this way, standard cracks with a a depth of 20 mm, a length of 40 mm and a width of 300 µm were created. For all samples with depth of 20 mm, a length of 40 mm and a width of 300 µm were created. For all samples with self-healing self-healing properties, the metal plates were removed after the curing period of 28 days, except for properties, the metal plates were removed after the curing period of 28 days, except for two samples two samples of both the CR_1_PU_HV and CR_1_PU_LV series (specimens D and E). For these last of both the CR_1_PU_HV and CR_1_PU_LV series (specimens D and E). For these last four samples, four samples, cracks were created and thus healing was triggered just before exposure to the water cracks were created and thus healing was triggered just before exposure to the water absorption test. absorption test. This was done in order to investigate the healing efficiency of the samples if cracking This was done in order to investigate the healing efficiency of the samples if cracking occurs upon occurs upon exposure to water and thus the healing agent had not yet hardened in the crack at the exposure to water and thus the healing agent had not yet hardened in the crack at the moment of moment of contact with water. contact with water. Autonomous crack healing was obtained for the samples with embedded capsules as the capsules were broken at the position of the crack plane at the moment that the metal plates were removed. Breakage of the capsules was followed by release of the healing agent out of the capsules, which

Materials 2016, 9, 311

5 of 28

Autonomous crack healing was obtained for the samples with embedded capsules as the capsules Materials 2016, 9,at 311the position of the crack plane at the moment that the metal plates were removed. 5 of 28 were broken Breakage of the capsules was followed by release of the healing agent out of the capsules, which developed into the cracks. Due to contact of the healing agent with moisture inside the cementitious developed into the cracks. Due to contact of the healing agent with moisture inside the cementitious matrix, the healing agent started to polymerize, resulting in crack repair. matrix, the healing agent started to polymerize, resulting in crack repair.

3.2. Sample 3.2. Sample Preparation Preparation As the the crack crack creation creation technique technique with with metal metal plates plates required the test test surface surface to to be be the the troweled troweled As required the surface, the test surface surface of of the the samples samples was was rough. rough. In test surface, surface, a a layer layer of of surface, the test In order order to to obtain obtain aa flat flat test approximately 1–2 mm was cut off. approximately 1–2 mm was cut off. Neutron radiography radiography images images are produced by by radiation radiation passing passing through through an an object. object. The The spatial spatial Neutron are produced resolution of of the the radiographs radiographs therefore therefore depends depends on on the the thickness thickness of of the the scanned scanned object. object. In In the the case case of of resolution mortar, a sample thickness of 40 mm is quite high in order to obtain a good spatial resolution of the mortar, a sample thickness of 40 mm is quite high in order to obtain a good spatial resolution of the water content content in in the the samples. samples. Therefore, Therefore, all all specimens specimens were were sawn sawn along along the the longitudinal longitudinal axis axis in in two two water equal parts. The two halves had a thickness ranging from 17 to 20 mm. The exact thickness of each equal parts. The two halves had a thickness ranging from 17 to 20 mm. The exact thickness of each half was was measured measured with with aa caliper caliper with with an an accuracy accuracy of of 0.05 0.05 mm. mm. One half One half half of of each each sample sample was was used used for for the water absorption tests. the water absorption tests. All specimens specimens were were dried dried in in an an oven oven at at 40 40 ˝°C less than than 0.1% 0.1% in in All C until until constant constant mass mass (mass (mass change change less 24 h) was achieved. In this way, a uniform moisture distribution was obtained inside the sample in 24 h) was achieved. In this way, a uniform moisture distribution was obtained inside the sample in aa short time time period. period. After specimens were covered over over aa height height of of short After the the drying drying period, period, the the sides sides of of the the specimens were covered 15 mm with a self-adhesive aluminum tape so that the water could only enter the samples through 15 mm with a self-adhesive aluminum tape so that the water could only enter the samples through the test test surface surface (top (top surface surface in in Figure Figure 2). 2). Regular interacts only only the Regular aluminum aluminum tape tape was was used used since since it it interacts weakly with with neutrons neutrons (it (it is is almost almost transparent transparent for for neutrons). neutrons). Since was weakly Since the the main main aim aim of of the the research research was the investigation of the water uptake in the crack region, a large part of the bottom surface was also the investigation of the water uptake in the crack region, a large part of the bottom surface was also covered with with the the aluminum aluminum tape that only only a a 10 10 mm mm wide wide zone zone around around the the cracks cracks was was exposed exposed to to covered tape so so that water during the absorption test (Figure 2). water during the absorption test (Figure 2).

Figure 2. 2. Mortar Mortar sample, sample, with with one one standard standard crack crack (CR_1), Figure (CR_1), covered covered with with aluminum aluminum tape. tape.

3.3. Visualization of 3.3. Visualization of Water Water Ingress Ingress 3.3.1. Neutron Radiography Facility 3.3.1. Neutron Radiography Facility The visualization of the water ingress in cracked mortar was performed by real time neutron The visualization of the water ingress in cracked mortar was performed by real time neutron radiography at the thermal neutron imaging facility NEUTRA. This facility is part of the spallation radiography at the thermal neutron imaging facility NEUTRA. This facility is part of the spallation neutron source SINQ of the Paul Scherrer Institute (PSI) in Switzerland. The neutron beam of the neutron source SINQ of the Paul Scherrer Institute (PSI) in Switzerland. The neutron beam of the spallation source was guided to a fixed size aperture of 2 cm in diameter by means of a convergent inner spallation source was guided to a fixed size aperture of 2 cm in diameter by means of a convergent collimator tube. From there, a divergent outer collimator led the parallelized neutron beam to the object inner collimator tube. From there, a divergent outer collimator led the parallelized neutron beam to space with a useful area of 400 mm diameter. The thermal energy spectrum of the neutron beam was the object space with a useful area of 400 mm diameter. The thermal energy spectrum of the neutron characterized by a Maxwell–Boltzmann distribution with peak energy of 25 meV. The neutron beam beam was characterized by a Maxwell–Boltzmann distribution with peak energy of 25 meV. The then passed through the studied samples to a 100 µm thick LiF/ZnS scintillator screen. This scintillator neutron beam then passed through the studied samples to a 100 µm thick LiF/ZnS scintillator screen. converted the neutrons to visible light, which was deflected by mirrors in a dark room and recorded This scintillator converted the neutrons to visible light, which was deflected by mirrors in a dark by a cooled slow-scan Andor SCMOS camera with a 50 mm AF-S NIKKOR lens. room and recorded by a cooled slow-scan Andor SCMOS camera with a 50 mm AF-S NIKKOR lens. For each series of specimens, open beam, dark current and black body images were acquired. The open beam image represents the spatial distribution of the neutron beam intensity and aimed at removing the inhomogeneity of the beam. The dark current image was taken while all shutters of the beam line were closed. It was used as a correction for the background noise level of the camera. The

Materials 2016, 9, 311

6 of 28

For each series of specimens, open beam, dark current and black body images were acquired. The open beam image represents the spatial distribution of the neutron beam intensity and aimed Materials 2016, 9,the 311 inhomogeneity of the beam. The dark current image was taken while all shutters 6 of 28 at removing of the beam line were closed. It was used as a correction for the background noise level of the camera. black body image was obtained by placing neutron absorbing blocks of boronated polyethylene in The black body image was obtained by placing neutron absorbing blocks of boronated polyethylene in front of the scanned samples. This was used as a background scattering correction. front of the scanned samples. This was used as a background scattering correction.

3.3.2. 3.3.2. Water Water Uptake Uptake during during Scanning Scanning Five Five samples samples could could be be scanned scanned simultaneously simultaneously since since the the detector’s detector’s field field of of view view measured measured 22. Therefore, a test frame with five aluminum shelves was used (Figure 3). Aluminum 310 × 310 mm 310 ˆ 310 mm . Therefore, a test frame with five aluminum shelves was used (Figure 3). Aluminum containers containerswere wereplaced placedon oneach eachof ofthe theshelves. shelves.The Themortar mortarspecimens specimenswere wereplaced placed on on aluminum aluminum line line supports in the containers. A reference image of the samples in the dry state was taken before supports in the containers. A reference image of the samples in the dry state was taken before the the containers containers were were filled filled with with water. water. Water Water was was manually manually added added to to the the containers containers by by means means of of aa syringe syringe so so that that the the immersion immersion of of the the samples samples amounted amounted to to 33 ±˘11 mm. mm. When Whenall all five five containers containerswere were filled, filled, radiographs were taken with an exposure time of 3 s, which resulted in a time step of 4.6 s due radiographs were taken with an exposure time of 3 s, which resulted in a time step of 4.6 s due to to aa process delay. This resulted in a pixel size of 0.273 mm for all images. Radiographic scanning during process delay. This resulted in a pixel size of 0.273 mm for all images. Radiographic scanning during the the water water absorption absorption process process continued continued for for 2–4 2–4 h. h.

Figure 3. Setup of the samples in the aluminum test frame during scanning. Figure 3. Setup of the samples in the aluminum test frame during scanning.

3.4. Image Analysis 3.4. Image Analysis In order to be able to accurately analyze the moisture distribution profiles in the samples, the In order to be able to accurately analyze the moisture distribution profiles in the samples, the corrections mentioned in Section 3.3.1 needed to be taken into account. Dark current image and black corrections mentioned in Section 3.3.1 needed to be taken into account. Dark current image and black body image were subtracted from each of the neutron radiographs in order to take into account the body image were subtracted from each of the neutron radiographs in order to take into account the background noise of the camera and the background scattering. The same was done for the open background noise of the camera and the background scattering. The same was done for the open beam image. Next, a flat field correction was applied by dividing the corrected neutron radiographs beam image. Next, a flat field correction was applied by dividing the corrected neutron radiographs by the corrected open beam image. All image operations were performed using the image processing software ImageJ (1.48v, National Institutes of Health, Bethesda, MD, USA). Visualization of the water front was done by dividing the images in the wet state by the reference image in the dry state. The moisture profile in the mortar samples was visualized in this way for three different exposure times: 5 min, 30 min and 4 h. From these images the crack healing with different

Materials 2016, 9, 311

7 of 28

by the corrected open beam image. All image operations were performed using the image processing software ImageJ (1.48v, National Institutes of Health, Bethesda, MD, USA). Visualization of the water front was done by dividing the images in the wet state by the reference Materials 2016, 9, 311 7 of 28 image in the dry state. The moisture profile in the mortar samples was visualized in this way for three different exposure times: 5 min, 30 at min and 4 h.time From theseSince images crack healing with different polyurethanes could be evaluated different steps. thethe inclusion of the images for all polyurethanes could be evaluated at different time steps. Since the inclusion of the images forper all samples would make this paper too lengthy, only the imaging of one representative specimen samples would make this paper too lengthy, only the imaging of one representative specimen per test test series was included. seriesNext was to included. the visual evaluation, a quantitative evaluation of crack healing was performed by Next to the visual evaluation, quantitative evaluation ofdifferent crack healing was performed byand plotting plotting the water profiles of the aobtained images at three times: 5 min, 30 min 4 h. the water profiles of the obtained images at three different times: 5 min, 30 min and 4 h. Calculation of Calculation of the water content was done in a similar way as described previously [30]. Horizontal the water content was done in a in similar way as previously water water profiles were determined the central 80described mm of each specimen[30]. at aHorizontal height of 11.3 mmprofiles above were determined in the central 80 mm of each specimen at a height of 11.3 mm above the bottom of the the bottom of the samples. Figure 4 shows the as such defined rectangular area for water profiling in samples. Figure 4 shows the as such defined rectangular area for water profiling in a schematic way for a schematic way for a sample neutron image. The height of the rectangular area perpendicular to the a sample neutron The height of the the rectangular perpendicular the cracks (Figure layer 4a) was cracks (Figure 4a)image. was chosen so that profile was area always determinedtoabove the capsule of chosen so that the profile was always determined above the capsule layer of the self-healing specimens. the self-healing specimens. In this way, the presence of the capsules did not interfere with the water In this way, presence of profiles the capsules interfere with the water profile in the mortar. Water profile in thethe mortar. Water were did alsonot taken along the cracks (Figure 4b). Therefore, additional profiles were also taken along the cracks (Figure 4b). Therefore, additional rectangular areas were rectangular areas were defined that coincided with the cracks and extended from the bottom to the defined that coincided with the cracks and extended from the bottom to the top of the sample. top of the sample.

Figure 4. Location a sample sample neutron neutron Figure 4. Location and and dimensions dimensions of of the the rectangular rectangular areas areas for for water water profiling profiling in in a radiography image: (a) and (b) (b) coinciding coinciding with with the the cracks. cracks. radiography image: (a) perpendicular perpendicular to to the the cracks; cracks; and

3.5. Analysis of the Spread Region of the Healing Agents in the Crack After visualization visualizationofofthe thewater water uptake neutron radiography, the specimens with healed uptake by by neutron radiography, the specimens with healed cracks crackssplit were to analyze the region spreadofregion of the polyurethanes on the crack In this were in split orderintoorder analyze the spread the polyurethanes on the crack faces. In faces. this way, the way, theofamount of water uptake through the healed linkedof to the amountthat of amount water uptake through the healed cracks could becracks linkedcould to thebe amount polyurethane polyurethane that filled uphealed the crack. For the containing healed specimens containing two cracks, specimens filled up the crack. For the specimens two cracks, the specimens werethe split twice so were splitcracks twice could so thatbeboth cracks Hence, could be analyzed. Hence, it could be determined the that both analyzed. it could be determined whether the healing whether agents were healing wereinequally distributed in both cracks. equally agents distributed both cracks. Resultsand andDiscussion Discussion 4. Results 4.1. Neutron Neutron Radiography Radiography Imaging Imaging of 4.1. of Crack Crack Healing Healing 4.1.1. (Un)cracked Condition 4.1.1. (Un)cracked Condition Before going into detail on the ability of an encapsulated PU-based healing agent to prevent Before going into detail on the ability of an encapsulated PU-based healing agent to prevent accelerated water ingress in cracks, the water uptake by capillary action of uncracked and cracked accelerated water ingress in cracks, the water uptake by capillary action of uncracked and cracked mortar was studied. When looking at the water uptake of a representative uncracked specimen as a mortar was studied. When looking at the water uptake of a representative uncracked specimen as a function of exposure time (Figure 5), it is clear that the water front only very gradually progressed function of exposure time (Figure 5), it is clear that the water front only very gradually progressed starting from the 10 ˆ 20 mm2 exposed area at the bottom surface of the prism. It took 30 min before starting from the 10 × 20 mm2 exposed area at the bottom surface of the prism. It took 30 min before the height of the water front actually rose above the water level in the tray. After having been in the height of the water front actually rose above the water level in the tray. After having been in contact with water for 4 h, the maximum height of the water front amounted to 14 mm starting from contact with water for 4 h, the maximum height of the water front amounted to 14 mm starting from the bottom surface of the specimen or 11 mm above the water level in the tray. The maximum width the bottom surface of the specimen or 11 mm above the water level in the tray. The maximum width of the lateral water front was around 35 mm. The recorded water ingress through the undamaged cementitious matrix is expected to be present also in the artificially cracked samples, but it will of course be surmounted with additional ingress, which is entirely induced by the presence of the crack.

Materials 2016, 9, 311

8 of 28

of the lateral water front was around 35 mm. The recorded water ingress through the undamaged cementitious matrix is expected to be present also in the artificially cracked samples, but it will of course be surmounted with additional ingress, which is entirely induced by the presence of the crack. Materials 2016, 9, 311

8 of 28

Materials 2016, 9, 311

8 of 28

Figure 5. Visualization of the water uptake as a function of exposure time using neutron radiography

Figure 5. Visualization of the water uptake as a function of exposure time using neutron radiography for one uncracked mortar prism (representative sample: UN_A). Figure 5. Visualization the water uptake as a function exposure time using neutron radiography for one uncracked mortar of prism (representative sample:of UN_A). for one uncracked mortar prism (representative sample: UN_A).

With one or two articial cracks present (Figure 6), logically, there is a faster rise of the water front With twolevel articial cracks present 6), logically, there is a faster of the suction water front aboveone the or water in the tray from the (Figure start of the experiment onwards due torise capillary With one or two articial cracks present (Figure 6), logically, there is a faster rise of the water front the location of the standardized abovephenomena the wateratlevel in the tray from the start crack(s). of the experiment onwards due to capillary suction above the water level in the tray from the start of the experiment onwards due to capillary suction phenomena at the location of of thethe standardized phenomena at the location standardizedcrack(s). crack(s).

Figure 6. Visualization of the water uptake as a function of exposure time using neutron radiography for cracked mortar prisms: (a) one artificial crack (representative sample: CR_1_B); and (b) two Figure 6. Visualization of the water uptake as a function of exposure time using neutron radiography Figure 6. Visualization of the water uptake as a function of exposure time using neutron radiography artificial cracks (representative sample: CR_2_C). for cracked mortar prisms: (a) one artificial crack (representative sample: CR_1_B); and (b) two for cracked mortar prisms: (a) one artificial crack (representative sample: CR_1_B); and (b) two artificial artificial cracks (representative sample: CR_2_C).

After 4 h of exposure, theCR_2_C). maximum height of the water front in the mortar prism containing cracks (representative sample: just one crack (Figure 6a) was 27 mm. When the crack tip is taken as reference point, the water front After 4 h of exposure, the maximum height of the water front in the mortar prism containing is extending around 12 the mmmaximum in upwardheight direction.the In the lateral direction, the water uptake in the just After 4 hcrack of exposure, front the mortar prism containing just one (Figure 6a) was 27 mm. When theof crack water tip is taken asinreference point, the water front vicinity of the crack did not turn out to be completely symmetrical. The final lateral ingress on the is extending 12 27 mm in upward direction. lateral as direction, thepoint, water the uptake in the one crack (Figurearound 6a) was mm. When the crack In tipthe is taken reference water front is left hand side seems to extend further, but also appears a bit more vague at the very end. Most vicinity of the crack did not turn out to be completely symmetrical. The final lateral ingress on the extending around 12 mm in upward direction. In the lateral direction, the water uptake in the vicinity probably, the aluminum tape which was applied on the bottom surface of the specimen next to the left hand side seems extend but also appears a bit more vague the veryonend. Most of the10 crack did turn to out be further, completely final lateralat ingress thesome left hand 2 not × 20 mm exposure area,towas not perfectly symmetrical. attached to theThe mortar surface. This allowed for probably, the aluminum tape which was applied on the bottom surface of the specimen next to the side seems to extend further,inbut also appears bit the more vague at very end.the Most probably, additional water ingress between the tape aand specimen. Bythe multiplying more correct the 10 × 20 mm2 exposure area, was not perfectly attached to the mortar surface. This allowed for some 2 aluminum was on the surface theaspecimen nextindication to the 10 of ˆ 20 lateral tape waterwhich front at theapplied right hand sidebottom of the crack by of two, more reliable themm additional water ingress in between the tape and the specimen. By multiplying the more correct maximum lateral width of the watertofront be obtained. should be equalfor to 47 mm.additional Thus, exposure area,overall was not perfectly attached the can mortar surface.It This allowed some lateral water front at the right hand side of the crack by two, a more reliable indication of the when comparing the dimensions of the final lateral water fronts of the uncracked and cracked samples, watermaximum ingress in between the tapeofand the specimen. multiplying thebemore lateral overall lateral width the water front can beBy obtained. It should equalcorrect to 47 mm. Thus,water 35right mm versus 47 mm, the undesirable effects ofmore the crack are quite significant. Comparison of the frontbeing at the hand side of the crack by two, a reliable indication of the maximum overall when comparing the dimensions of the final lateral water fronts of the uncracked and cracked samples, maximum heights of the water fronts for uncracked and cracked mortar, being 14 mm versus 27 mm lateral width of the water front can be obtained. It should be equal to 47 mm. Thus, when comparing being 35 mm versus 47 mm, the undesirable effects of the crack are quite significant. Comparison of the from the bottom of samples, leads to a similar conclusion. maximum heights the lateral water fronts uncracked cracked mortar, being 14samples, mm versus 27 mm the dimensions of the of final waterforfronts of theand uncracked and cracked being 35 mm from the bottom of samples, leads to a similar conclusion.

Materials 2016, 9, 311

9 of 28

versus 47 mm, the undesirable effects of the crack are quite significant. Comparison of the maximum heights of the water fronts for uncracked and cracked mortar, being 14 mm versus 27 mm from the bottom of samples, leads to a similar conclusion. Materials 9, 311 with two cracks (Figure 6b), the initial water uptake per crack is very 9similar of 28 For the2016, specimen to that of the specimen with a single crack. However, after 30 min of exposure, the water fronts around For the specimen with two cracks (Figure 6b), the initial water uptake per crack is very similar the two individual cracks are about to merge. From then onwards combined effect fronts should be to that of the specimen with a single crack. However, after 30 min their of exposure, the water takenaround into consideration. This time, the overall water front stayed quite symmetrical until the very end the two individual cracks are about to merge. From then onwards their combined effect of theshould experiment. maximumThis risetime, of the front starting from the crack tips amounted be takenThe intofinal consideration. thewater overall water front stayed quite symmetrical until to around mm, overall The lateral frontrise was mmstarting in width near bottom the13 very endwhile of the the experiment. finalwater maximum ofaround the water80front from the the crack tips of amountedThe to around mm, thehand overall lateral water 80 mm inside width the specimen. water 13 front atwhile the left side of the left front crackwas andaround the right hand of near the right bottom for of the specimen. water front at thesides. left hand of thethan left crack and the rightcontaining hand crackthe extended about 30 mmThe outwards on both Thisside is more for the specimen side of the right crack extended for about 30 mm outwards on both sides. This is more than for just one crack (Figure 6a). There, around 23 mm was measured for the more accurate maximumthe ingress specimen containing just one crack (Figure 6a). There, around 23 mm was measured for the more at the right hand side of the crack. accurate maximum ingress at the right hand side of the crack.

4.1.2. Healing of One Crack

4.1.2. Healing of One Crack

Comparison of the water fronts observed for the samples where one crack was healed Comparison of the water fronts observed for the samples where one crack was healed with an with encapsulated an encapsulated PU-based healing agent (Figure with those of the (Figure (un)cracked mortar PU-based healing agent (Figure 7) with those of7)the (un)cracked mortar 5 and 6a), (Figures 5 and 6a), allows a thorough ofability. the crack healing ability. allows a thorough evaluation of the evaluation crack healing

Figure 7. Visualization of the water uptake as a function of exposure time using neutron radiography

Figure 7. Visualization of the water uptake as a function of exposure time using neutron radiography for mortar prisms with one crack healed: (a) crack healing with PU_HV (representative sample: for mortar prisms with crackhealing healed: (a)PU_LV crack (representative healing with sample: PU_HVCR_1_PU_LV_A). (representative sample: CR_1_PU_HV_A); andone (b) crack with CR_1_PU_HV_A); and (b) crack healing with PU_LV (representative sample: CR_1_PU_LV_A).

The water uptake of the specimens just after bringing them in contact with the water (after 5 min)

The watermuch uptake thewhen specimens just after bringing themIninother contact with waterof(after 5 min) is clearly lessof than an unhealed crack was present. words, thethe release the PU upon capsule breakage was able to eliminate the strong capillary suction effects in the artificial cracks is clearly much less than when an unhealed crack was present. In other words, the release of the PU properly. This was statement holds true forthe thestrong two PU-based healing agents, either with a high or uponquite capsule breakage able to eliminate capillary suction effects in the artificial cracks low viscosity andholds PU_LV). that PU-based the white grey coloragents, inside the crackwith represents quite aproperly. This(PU_HV statement trueAlso for note the two healing either a high or a the polyurethane thus indicates a substantial of the crack. low viscosity (PU_HVand and PU_LV). Also note that filling the white grey color inside the crack represents the With increasing exposure time, a water front above the water level gradually becomes visible in polyurethane and thus indicates a substantial filling of the crack. the vicinity of the crack. In terms of grey value it is much vaguer than for the cracked samples. Still, With increasing exposure time, a water front above the water level gradually becomes visible this water front progresses faster than in case of the uncracked sample. After 30 min, it is clearly more in thevisible, vicinity of for thethe crack. In terms of grey value it is much vaguer can than the cracked samples. while uncracked sample it takes longer before something befor seen above the water Still, this water front progresses faster than in case of the uncracked sample. After 30 min, it is clearly level. By the end of the experiment, there is an overall water front that is clearly larger in area than in morecase visible, while for the uncracked it takes longer before something seen above the of the uncracked sample. This issample especially the case for the sample containingcan the be high viscosity (Figure 7a). Based on aismainly visual evaluation on one sample,larger the low waterPU-based level. Byhealing the endagent of the experiment, there an overall water front that is clearly in area healing agent seemed workedthe more effectively PU_LV). the high than viscosity in case ofPU-based the uncracked sample. Thistoishave especially case for the (Figure sample7b, containing Note that for the sample containing encapsulated (Figure 7b) there are traces undesired viscosity PU-based healing agent (Figure 7a). Based onPU_LV a mainly visual evaluation onof one sample, the water sorption at the right hand side of the sample. This is merely due to an improper attachment of low viscosity PU-based healing agent seemed to have worked more effectively (Figure 7b, PU_LV). the aluminum tape onto the bottom surface of the mortar prism. Although the aluminum tape was Note that for the sample containing encapsulated PU_LV (Figure 7b) there are traces of undesired always carefully applied, this problem could not always be avoided unfortunately. water sorption at the right hand side of the sample. This is merely due to an improper attachment of

Materials 2016, 9, 311

10 of 28

the aluminum tape onto the bottom surface of the mortar prism. Although the aluminum tape was always carefully applied, this problem could not always be avoided unfortunately. Materials 2016, 9, 311

4.1.3. Healing of Two Cracks

10 of 28

4.1.3.the Healing of Two with Cracks For specimens self-healing properties containing two artificial cracks, the following conclusions can be drawn (Figure 8). When properties using a high viscosity PUartificial for healing purposes (Figure 8a, For the specimens with self-healing containing two cracks, the following conclusions can be drawn (Figure 8). When a high viscosity PUcracks for healing purposes (Figure 8a, for PU_HV), the initial water uptake after 5 minusing in the vicinity of the is substantially reduced thetime initial water uptake after 5 minwater in the front vicinity of thethat cracks substantially for side both PU_HV), cracks. As passes, the progressing shows theiscrack on the reduced right hand both better cracks. than As time front shows that theperformance crack on the right hand side is healed thepasses, one onthe theprogressing left hand water side. Thus, the healing for the two cracks is healed better than the one on the left hand side. Thus, the healing performance for the two cracks is not equal. The same goes for the specimen that contained a 50 mm long capsule filled with a low is not equal. The same goes for the specimen that contained a 50 mm long capsule filled with a low viscosity PU precursor (Figure 8b, PU_LV). The only difference is that the unequal water uptake for viscosity PU precursor (Figure 8b, PU_LV). The only difference is that the unequal water uptake for the two cracks is already visible just after starting the experiment. The crack on the right hand side the two cracks is already visible just after starting the experiment. The crack on the right hand side barely shows signs of healing, while oneon onthe theleft lefthand hand side seems topreventing be preventing barely shows signs of healing, whilethe thePU PU in in the the one side seems to be waterwater ingress almost completely, even after having been in contact with water for 4 h. ingress almost completely, even after having been in contact with water for 4 h.

Figure 8. Visualization of the water uptake as a function of exposure time using neutron radiography

Figure 8. Visualization of the water uptake as a function of exposure time using neutron radiography for mortar prisms with two cracks healed: (a) crack healing with PU_HV (representative sample: for mortar prisms with two cracks healed: (a) crack healing with PU_HV (representative sample: CR_2_PU_HV_C); and (b) crack healing with PU_LV (representative sample: CR_2_PU_LV_B). CR_2_PU_HV_C); and (b) crack healing with PU_LV (representative sample: CR_2_PU_LV_B).

The difference in behavior between the two PU-based healing agents with respect to the proper

The difference in behavior between the two PU-based healinginagents with respect to the proper healing of two cracks could maybe be attributed to their difference viscosity. When subsequently healing of two maybe be(within attributed to their viscosity. When pulling out cracks the twocould thin metal plates a timespan of difference less than oneinminute) to break thesubsequently capsules andout initiate the release of theplates healing agent, the high viscosity PUthan precursor tends toto flow quite easily pulling the two thin metal (within a timespan of less one minute) break the capsules into thethe tworelease cracks. On the healing other hand, the low PU precursor rather flows into crack thateasily and initiate of the agent, the viscosity high viscosity PU precursor tends tothe flow quite corresponds with On the the thinother metalhand, plate that out PU first.precursor Due to therather lower flows viscosity, into the two cracks. the was low pulled viscosity intohence the crack greater fluidity, PU_LV seeped more quickly into one of the cracks before extracting the second metal that corresponds with the thin metal plate that was pulled out first. Due to the lower viscosity, hence plate and thereby leaving less material to fill the second crack. Whereas in the specimen containing greater fluidity, PU_LV seeped more quickly into one of the cracks before extracting the second metal the high viscosity healing agent, the extraction of the metal plates was relatively fast to allow a more plate even and thereby leaving lessagents material to fill Whereas in the specimen distribution of the along thethe twosecond cracks.crack. However, careful evaluation of containing the other the high specimens viscosity healing agent, the extraction of the metal plates was relatively fast to allow a more even that were tested cannot unambiguously confirm this assumption. For a more in-depth distribution of on thethat agents along the to two cracks. However, careful evaluation of the other specimens evaluation matter, we refer Section 4.4 where the spread region of the self-healing substances that were tested unambiguously confirm this assumption. Forobservation a more in-depth onto the crackcannot walls has been assessed experimentally. Another general that canevaluation be made on is that the of Section PU-based4.4 healing 50 mm capsulessubstances is insufficient to the that matter, weamount refer to whereagent the present spread in region oflong the glass self-healing onto properly heal two adjacent cracks. This is quite logic as the PU amount present in one 30 mm long crack walls has been assessed experimentally. Another general observation that can be made is that capsule of was not enough to fully prevent the water in just onecapsules crack. is insufficient to properly the amount PU-based healing agent present in 50ingress mm long glass heal two adjacent cracks. This is quite logic as the PU amount present in one 30 mm long capsule was 4.1.4. Water Front/Sample Cross-Section Area Ratios not enough to fully prevent the water ingress in just one crack. The water front/sample cross-section area (WFA/SCA) ratios as calculated in ImageJ (1.48v, National Institutes of Health, Bethesda, MD, from the neutron radiography images after 4 h of 4.1.4. Water Front/Sample Cross-Section AreaUSA) Ratios exposure (Figures 5–8), allow a somewhat more quantitative evaluation of the healing ability when The front/sample cross-section areaPU-based (WFA/SCA) ratios calculated in ImageJ (1.48v, usingwater the two considered types of encapsulated healing agentas (Table 2).

National Institutes of Health, Bethesda, MD, USA) from the neutron radiography images after 4 h of

Materials 2016, 9, 311

11 of 28

exposure (Figures 5–8), allow a somewhat more quantitative evaluation of the healing ability when using the two considered types of encapsulated PU-based healing agent (Table 2). Table 2. Water front/sample cross-section area (WFA/SCA) ratios as calculated from the neutron radiography images after 4 h of exposure (Figures 5–8). Sample

WFA/SCA (%)

UN CR_1 CR_1_PU_HV CR_1_PU_LV CR_2 CR_2_PU_HV CR_2_PU_LV

10.8 ˘ 5.0 14.1 ˘ 0.6 16.7 ˘ 3.5 15.7 ˘ 4.5 26.9 ˘ 1.9 25.3 ˘ 3.1 24.0 ˘ 3.8

When comparing the uncracked test series (UN) and the test series containing just one crack (CR_1), the mean WFA/SCA ratio of the latter is higher. Nevertheless, the rather high standard error (5.0%) for the uncracked series indicates that the water uptake could be as high as for cracked mortar. This rather unexpected observation can mainly be attributed to the fact that the aluminum tape at the bottom surface of one of the uncracked samples (UN_B) was not entirely well attached and therefore caused some undesirable extra water uptake (WFA/SCA: 15.8%) through capillary suction phenomena between the tape and the bottom surface. Sample UN_A was characterized by a WFA/SCA ratio of only 5.8%. In comparison with that value, the effect of one crack on the water uptake is evident (WFA/SCA: 14.1%). The self-healing test series with one crack (CR_1_PU_HV and CR_1_PU_LV) have WFA/SCA ratios that are slightly higher than 14.1%, i.e., 16.7% and 15.7%, respectively. This would indicate that the healing mechanism was not effective at all. Nonetheless, their high standard errors (3.5% and 4.5%, respectively) demonstrate that considerably lower values are very well possible. The recorded minimum values amounted to 11.9% and 6.9%, respectively. Comparison of those two values would indicate a slight preference for the low viscosity healing agent. However, given their high susceptibility to variation, it remains difficult to draw definitive conclusions on that matter. Evaluation of this parameter for the specimens in which two cracks should have healed (CR_2_PU_HV and CR_2_PU_LV), shows that for both the high and low viscosity PU-based healing agents the WFA/SCA ratios (25.3% and 24.0%) are only slightly lower compared to the cracked condition (CR_2: 26.9%). Given the substantial standard error on those results (3.1% and 3.8%, respectively) the differences with the cracked reference (CR_2) are almost non-existing. The amount of PU precursor incorporated in the 50 mm long capsules flowing into the cracks is not sufficient to establish a decent healing performance. Based on the WFA/SCA ratios, a higher dosage of healing agent would be preferred. Again, the WFA/SCA ratios do not indicate a preference for a high or low viscosity healing agent. On the other hand, one should be aware of the fact that a low viscosity healing agent mainly flows towards one crack, the one that has been created first. This was clearly visible in the neutron radiography images (Figure 8b). Thus, for multiple crack healing, a low viscosity PU-based healing agent may not necessarily be an advantage, depending on the expected crack distance. A higher dosage of PU_HV may be recommended in that case. 4.2. Water Profiling of Crack Healing Perpendicular to the Crack(s) 4.2.1. (Un)cracked Condition The water profiles corresponding with a rectangular area taken at a height of 11.3 mm above the bottom of the samples (Figure 9), lead to quite similar conclusions as the neutron radiography images shown in Section 4.1.1.

Materials 2016, 9, 311

12 of 28

Materials 2016, 9, 311

12 of 28

Water Water contentcontent (× 10 -1 g/cm³) contentcontent (× 10 -1 g/cm³) (× 10 -1 g/cm³)Water Water (× 10 -1 g/cm³)Water Water contentcontent (× 10 -1 g/cm³) (× 10 -1 g/cm³)

Materials 2016, 9, 311

12 of 28

5 minutes (a) 2.0 5 minutes (a)

UN_A UN_B

2.0

UN_A UN_B

0.0 -40

-30

-20

-10

0 0.0 30 minutes (b)

10

20

30

40

-40

-30

-20

-10

2.0 0 30 minutes (b)

10

20

30

40 UN_A UN_B

2.0

UN_A UN_B

0.0

-40

-30

-20

-10

0 40.0 hours (c)

10

20

30

-40

-30

-20

-10

2.0 0 4 hours (c)

10

20

30

2.0

40

UN_A UN_B

40

UN_A UN_B

0.0

-40

-30

-20

-10 0 10 0.0 Distance from the center (mm)

20

30

40

-40

-30

-20

-10

20

30

40

0

10

Figure 9. Water profiles of the two uncracked mortar Distance standard from the center (mm)prisms: (a) after 5 min; (b) after 30 min; Figure 9. Water profiles of the two uncracked standard mortar prisms: (a) after 5 min; (b) after 30 min; and (c) after 4 h. andFigure (c) after 4 h. profiles of the two uncracked standard mortar prisms: (a) after 5 min; (b) after 30 min; 9. Water

Water Water contentcontent (× 10 -1 g/cm³) (× 10 -1 g/cm³)

and (c) after 4 h. After 30 min of exposure, there is still no increased water content at the selected zone in the After 30 min of exposure, there is still no increased water content at the selected zone in the sample, −1 g/cm3. The increased sample, whereas, by the end of the experiment, it showed an increase of 0.8 ×´10 1 g/cm 3 . The After 30 min of exposure, there isitstill no increased water content at the selected zone in the whereas, by the end of the experiment, showed an increase of 0.8 ˆ 10 increased water content is mainly located around the central 10 × 20 mm2 exposed area. −1The distance from the 3. The increased 2 sample, whereas, by the end of the experiment, it showed an increase of 0.8 × 10 g/cm water content is mainly around the(−10, central 10 ˆto20(−30, mm30exposed expected location of alocated crack ranges from 10 mm) mm). area. The distance from the water content isofmainly located around the central 10 to × 20 mm230 exposed area. The distance from the expected location a crack ranges from (´10, 10 mm) (´30, Note that the water profiles of samples UN_A and UN_B after mm). 4 h of exposure differ significantly. expected location of a crack ranges from (−10, 10 mm) to (−30, 30 mm). Note thatwater the water profiles UN_B after 4due h oftoexposure differ The broad spread in caseof of samples the latterUN_A sampleand is most probably the fact that thesignificantly. aluminum Note thatspread the water profiles of samples UN_Aisand UN_B after 4due h ofto exposure differ significantly. Thetape broad water in case of the latter sample most probably the fact that the aluminum did not attach entirely well to the bottom surface of that mortar prism. The water profile of broad waterentirely spread in case ofthe thebottom latter sample isof most probably due to The the fact thatprofile the aluminum tapeThe did not attach well to surface that mortar prism. water of sample sample UN_B should therefore be interpreted with caution. tape did not attach entirely well to the bottom surface of that mortar prism. The water profile of UN_B should beone interpreted In the therefore presence of artificial with crack,caution. the profiles show an obvious steep peak corresponding sample UN_B should therefore be interpreted with caution. In the of one artificial crack, the profiles showitself an obvious peak corresponding with with thepresence water absorbed by capillary action in the crack (Figure steep 10). The notable water profile In the presence of one artificial crack, the profiles show an obvious steep peak corresponding to this peak represents the water by the(Figure mortar10). matrix the crack Thisadjacent water the adjacent water absorbed by capillary action in theuptake crack itself Thenear notable waterwall. profile with the water absorbed by capillary action in the crack itself (Figure 10). The notable water profile profile progresses laterally as abyfunction of exposure timethe (Figure 10a–c)This while the profile peak to this peakclearly represents the water uptake the mortar matrix near crack wall. water adjacent to this peak represents the water uptake by the mortar matrix near the crack wall. This water corresponding with the crack did not show significant change. clearly progresses laterally as a function of exposure time (Figure 10a–c) while the peak corresponding profile clearly progresses laterally as a function of exposure time (Figure 10a–c) while the peak withcorresponding the crack didwith not show significant change. 5 minutes (a)change. the crack did not show significant 2.0 5 minutes (a)

CR_1_A CR_1_B CR_1_C CR_1_A

2.0

CR_1_B CR_1_C 0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0

0

10

20

30

40

0

10

20

30

40

Figure 10. Cont.

Materials 2016, 9, 311

13 of 28

-1 g/cm³) Water -1 g/cm³) Water content (× 10(×-1 10 g/cm³) content (× 10(×-110 g/cm³) Water content Water content

Materials 2016, 9, 311 Materials 2016, 9, 311

13 of 28 13 of 28 30 minutes (b)

30 2.0 minutes (b)

CR_1_A CR_1_B CR_1_A CR_1_C CR_1_B

2.0

CR_1_C

0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0 0 4 hours (c) 0 42.0 hours (c)

10

10

20

20

30

30

2.0

40

40 CR_1_A CR_1_B CR_1_A CR_1_C CR_1_B CR_1_C

0.0 -40

-30

-20

-40

-30

-20

0.0 0 -10 10 Distance from the center (mm) -10 0 10 Distance from the center (mm)

20

20

30

40

30

40

Figure 10. Water profiles of the three standard mortar prisms containing one artificial crack: (a) after Figure 10. Water profiles of the three standard mortar prisms containing one artificial crack: (a) after 5 min; (b) 30profiles min; and 4 h. Figure 10. after Water of(c) theafter three standard mortar prisms containing one artificial crack: (a) after 5 min; (b) after 30 min; and (c) after 4 h. 5 min; (b) after 30 min; and (c) after 4 h.

WaterWater content (× 10 -1(×g/cm³) content (× 10 -1(×g/cm³) content (× 10 -1 content 10 -1 g/cm³)WaterWater content 10 -1 g/cm³) Water Water content (×g/cm³) 10 -1 g/cm³)

For the specimens containing two cracks (Figure 11), the obtained water profiles can be more or For the specimens containing two cracks (Figure 11), the obtained water profiles can be more less considered as a superposition of two profiles of 11), a singular crack.water Only profiles after about 30 more min of For the specimens containing two cracks (Figure the obtained can be or or less considered as a superposition oftwo two profilesofshould ofa asingular singular crack. Only after about 30Just min exposure the effect the merging ofofthe twoprofiles profiles be taken intoOnly consideration as well. less considered as aofsuperposition crack. after about 30 min of of exposure thespecimens effectof ofthe the merging the two profiles should beextend taken into as Just well. like for the with one crack, the water profiles mainly theconsideration lateral direction as exposure the effect merging ofof the two profiles should be taken into in consideration as well. Justlike likefor forthe theexposure specimens withone onecrack, crack,the thewater waterprofiles profilesmainly mainlyextend extendininthe thelateral lateraldirection directionasas function of time. specimens with function of exposure time. function of exposure time. 5 minutes (a)

5 2.0 minutes (a)

CR_2_A CR_2_B CR_2_A CR_2_C CR_2_B CR_2_C

2.0

0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0 0 30 minutes (b) 2.0 0 30 minutes (b)

10

10

20

20

30

30

2.0

40

40 CR_2_A CR_2_B CR_2_A CR_2_C CR_2_B CR_2_C

0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0 0 4 hours (c) 2.0 0 4 hours (c) 2.0

10

10

20

20

30

30

40 CR_2_A CR_2_B CR_2_A CR_2_C CR_2_B CR_2_C

40

0.0 -40

-30

-20

-10 10 20 30 40 0.0 0 Distance from the center (mm) -40 -30 -20 -10 0 10 20 30 40 Distance from the center (mm) Figure 11. Water profiles of the three standard mortar prisms containing two artificial cracks: (a) after

5 min; (b) after 30 min; and (c) after 4 h. Figure 11. Water profiles of the three standard mortar prisms containing two artificial cracks: (a) after Figure 11. Water profiles of the three standard mortar prisms containing two artificial cracks: (a) after 5 min; (b) after 30 min; and (c) after 4 h. 5 min; (b) after 30 min; and (c) after 4 h.

Materials 2016, 9, 311

14 of 28

Materials 2016, 9, 311 Materials 2016, 9, 311

14 of 28 14 of 28

-1 g/cm³) Water Water content content (× 10 (×-110 g/cm³)

-1 g/cm³) Water Water content content (× 10 (×-110 g/cm³)

-1 g/cm³) Water Water content content (× 10 (×-110 g/cm³)

4.2.2. 4.2.2.Healing Healingof ofOne OneCrack Crack 4.2.2. Healing of One Crack When Whenlooking lookingat atthe thewater waterprofiles profilesfor forthe therectangular rectangulararea areaabove abovethe theposition position of ofthe thecapsule, capsule, When looking at the water profiles for the rectangular area above the position of theviscosity capsule, little difference was observed between the specimens that were healed with the high and low little difference was observed between the specimens that were healed with the high and low viscosity little difference was observed between the specimens that were healed with the high and low viscosity healing healingagent agent(Figures (Figures12 12and and13). 13). healing agent (Figures 12 and 13). 5 minutes (a) 5 minutes (a) 2.0

CR_1_PU_HV_A CR_1_PU_HV_A CR_1_PU_HV_B CR_1_PU_HV_B CR_1_PU_HV_C CR_1_PU_HV_C

2.0

-40 -40

-30 -30

-20 -20

-10 -10

0.0 0.0

0 30 minutes (b) 0 30 minutes (b) 2.0

10 10

20 20

30 30

40 40

CR_1_PU_HV_A CR_1_PU_HV_A CR_1_PU_HV_B CR_1_PU_HV_B CR_1_PU_HV_C CR_1_PU_HV_C

2.0

0.0 0.0 -40 -40

-30 -30

-20 -20

-10 -10

0 4 hours 0 (c) 42.0 hours (c)

10 10

20 20

30 30

40 40

CR_1_PU_HV_A CR_1_PU_HV_A CR_1_PU_HV_B CR_1_PU_HV_B CR_1_PU_HV_C CR_1_PU_HV_C

2.0

0.0 0.0 -40 -40

-30 -30

-20 -20

-10 0 10 Distance from 0the center (mm) -10 10 Distance from the center (mm)

20 20

30 30

40 40

-1 g/cm³) Water Water content content (× 10 (×-110 g/cm³)

-1 g/cm³) Water Water content content (× 10 (×-110 g/cm³)

Figure12. 12.Water Waterprofiles profilesofofthe thethree threestandard standardmortar mortarprisms prismscontaining containingone oneartificial artificialcrack crackhealed healed Figure Figure 12. Water profiles of the three standard mortar prisms containing one artificial crack healed withPU_HV: PU_HV:(a) (a)after after5 5min; min;(b) (b)after after3030min; min;and and(c)(c)after after4 4h.h. with with PU_HV: (a) after 5 min; (b) after 30 min; and (c) after 4 h. 5 minutes (a) 5 minutes (a) 2.0

CR_1_PU_LV_A CR_1_PU_LV_A CR_1_PU_LV_B CR_1_PU_LV_B CR_1_PU_LV_C CR_1_PU_LV_C

2.0

0.0 0.0 -40 -40

-30 -30

-20 -20

-10 -10

0 30 minutes (b) 0 30 2.0 minutes (b)

10 10

20 20

30 30

40 40

CR_1_PU_LV_A CR_1_PU_LV_A CR_1_PU_LV_B CR_1_PU_LV_B CR_1_PU_LV_C CR_1_PU_LV_C

2.0

0.0 0.0 -40 -40

-30 -30

-20 -20

-10 -10

0 0

Figure13. 13.Cont. Cont. Figure Figure 13. Cont.

10 10

20 20

30 30

40 40

Materials 2016, 9, 311

15 of 28

Materials 2016, 9, 311

15 of 28

Water content (× 10 -1 g/cm³)

4 hours (c) 2.0

CR_1_PU_LV_A CR_1_PU_LV_B CR_1_PU_LV_C

0.0 -40

-30

-20

-10 0 10 Distance from the center (mm)

20

30

40

Figure 13. Water profiles of the three standard mortar prisms containing one artificial crack healed Figure 13. Water profiles of the three standard mortar prisms containing one artificial crack healed with PU_LV: (a) after 5 min; (b) after 30 min; and (c) after 4 h. with PU_LV: (a) after 5 min; (b) after 30 min; and (c) after 4 h.

Opposite results were observed when comparing to those of the neutron radiography images of Opposite results were observed when comparing to those of the neutron radiography images samples CR_1_PU_HV_A and CR_1_PU_LV_A. There, the water fronts exhibited by the former of samples CR_1_PU_HV_A and CR_1_PU_LV_A. There, the water fronts exhibited by the former sample looked more pronounced than in the latter. However, this is not completely unexpected as sample looked more pronounced than in the latter. However, this is not completely unexpected as the calculated water profiles only relate to a very thin rectangular area above the capsule while the the calculated water profiles only relate to a very thin rectangular area above the capsule while the visually observed water front in the neutron radiography images relates to the whole cross-sectional visually observed water front in the neutron radiography images relates to the whole cross-sectional area around the crack zone. Still, one should be aware of the fact that the neutron images in Figure 7 area around the crack zone. Still, one should be aware of the fact that the neutron images in Figure 7 do not account for the possible variation within each of the two specimen groups (CR_1_PU_HV and do not account for the possible variation within each of the two specimen groups (CR_1_PU_HV and CR_1_PU_LV). Additional consideration of the neutron radiography image related mean WFA/SCA CR_1_PU_LV). Additional consideration of the neutron radiography image related mean WFA/SCA ratios and their corresponding standard errors (Table 2) indicates no preference for the high or low ratios and their corresponding standard errors (Table 2) indicates no preference for the high or low viscosity healing agent. This is completely in line with the water profile results. viscosity healing agent. This is completely in line with the water profile results. A closer look at the individual profiles shows that the crack itself was not always properly healed. A closer look at the individual profiles shows that the crack itself was not always properly healed. For each type of healing agent (PU_HV and PU_LV), one out of three samples still shows an upward For each type of healing agent (PU_HV and PU_LV), one out of three samples still shows an upward peak indicating presence of water in the crack. Ideally, a healed sample should be characterized by a peak indicating presence of water in the crack. Ideally, a healed sample should be characterized by significant downward peak. The more the downward peak reaches the flat baseline curve for an a significant downward peak. The more the downward peak reaches the flat baseline curve for an unexposed area of the sample, the lower the amount of water. Given this, there is an obvious healing unexposed area of the sample, the lower the amount of water. Given this, there is an obvious healing effect for two out of three samples per type of healing agent. Still, it must be said that the expected effect for two out of three samples per type of healing agent. Still, it must be said that the expected enhanced healing performance in comparison to the observed was actually higher. enhanced healing performance in comparison to the observed was actually higher. 4.2.3.Healing HealingofofTwo TwoCracks Cracks 4.2.3. Thewater waterprofiles profilesfor forthe thespecimens specimenswith withtwo twohealed healedcracks cracks(Figures (Figures14 14and and15) 15)are aretotobe be The interpreted in in aa similar forfor thethe upward and and downward peakspeaks and how interpreted similar way waybybylooking looking upward downward andthey howevolve. they From the comparison of the neutron radiography images of samples CR_2_PU_HV_C and evolve. From the comparison of the neutron radiography images of samples CR_2_PU_HV_C CR_2_PU_LV_B (Figure 8), it seemed that with PU_HV, the two cracks were somewhat more equally and CR_2_PU_LV_B (Figure 8), it seemed that with PU_HV, the two cracks were somewhat more filled than with PU_LV, especially in the initial the experiment. The water profiles equally filled than with PU_LV, especially in thestages initialof stages of the experiment. The water obtained profiles for each test series of samples further confirm this. Except for the water profile of specimen specimen obtained for each test series of samples further confirm this. Except for the water profile of CR_2_PU_HV_B,major major upward peaks are completely in theprofiles water of profiles of replicates the other CR_2_PU_HV_B, upward peaks are completely absentabsent in the water the other replicates containing PU_HV. On the other hand, the downward peaks do not always completely containing PU_HV. On the other hand, the downward peaks do not always completely extend to the extend to of the water profiles. baseline of the the baseline water profiles. WithPU_LV, PU_LV, a substantial upward water content related canseen be seen incrack one crack all With a substantial upward water content related peakpeak can be in one for allfor four four tested samples. The peak water content in all samples is related to the crack which appeared tested samples. The peak water content in all samples is related to the crack which appeared second in second in the sample. Thus, tends the PU_LV tends to flow into the crack that first, was created the sample. Thus, the PU_LV to flow mainly into mainly the crack that was created leaving first, the leaving the other one practically unhealed. Evaluation of the water profiles along the cracks and the other one practically unhealed. Evaluation of the water profiles along the cracks and the spread region spread region of the self-healing substances on the crack faces further confirms this (see Sections of the self-healing substances on the crack faces further confirms this (see Sections 4.3 and 4.4). It 4.3 is and 4.4). It is also worth noticing that the downward peaks after 4 h of exposure in the one healed also worth noticing that the downward peaks after 4 h of exposure in the one healed crack seem a bit crackpronounced seem a bit more than when using PU_HV. more than pronounced when using PU_HV.

Materials 2016, 9, 311

16 of 28

-1 g/cm³) -1 g/cm³) -1 g/cm³) Water content Water content Water content Water content (× 10(×-1 10 g/cm³) Water content (× 10(×-1 10 g/cm³) Water content (× 10(×-1 10 g/cm³)

Materials 2016, 9, 311 Materials 2016, 9, 311

16 of 28 16 of 28 5 minutes (a) 2.0 5 minutes (a)

CR_2_PU_HV_A CR_2_PU_HV_B CR_2_PU_HV_A CR_2_PU_HV_C CR_2_PU_HV_B CR_2_PU_HV_D CR_2_PU_HV_C

2.0

CR_2_PU_HV_D 0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0 0 30 minutes (b) 0 30 2.0 minutes (b)

10

20

30

40

10

20

30

40

CR_2_PU_HV_A CR_2_PU_HV_B CR_2_PU_HV_A CR_2_PU_HV_C CR_2_PU_HV_B CR_2_PU_HV_D CR_2_PU_HV_C

2.0

CR_2_PU_HV_D

0.0

-40

-30

-20

-10

-40

-30

-20

-10

0.0 0 4 hours (c) 0 42.0 hours (c)

10

20

30

40

10

20

30

40

CR_2_PU_HV_A CR_2_PU_HV_B CR_2_PU_HV_A CR_2_PU_HV_C CR_2_PU_HV_B CR_2_PU_HV_D CR_2_PU_HV_C

2.0

CR_2_PU_HV_D 0.0 -40

-30

-20

-40

-30

-20

0.0 -10 0 10 Distance from the center (mm) -10 0 10 Distance from the center (mm)

20

30

40

20

30

40

-1 g/cm³) -1 g/cm³) -1 g/cm³) Water content Water content Water content Water content (× 10(×-1 10 g/cm³) Water content (× 10(×-1 10 g/cm³) Water content (× 10(×-1 10 g/cm³)

Figure 14. Water profiles of the four standard mortar prisms containing two artificial cracks healed Figure 14. Water profiles of the four standard mortar prisms containing two artificial cracks healed Figure 14. Water profiles of the mortar prisms with PU_HV: (a) after 5 min; (b) four after standard 30 min; and (c) after 4 h. containing two artificial cracks healed withwith PU_HV: (a) after 5 min; (b) after 30 min; and (c) after 4 h. PU_HV: (a) after 5 min; (b) after 30 min; and (c) after 4 h. 5 minutes (a) 2.0 5 minutes (a)

CR_2_PU_LV_A CR_2_PU_LV_B CR_2_PU_LV_A CR_2_PU_LV_C CR_2_PU_LV_B CR_2_PU_LV_D CR_2_PU_LV_C

2.0

CR_2_PU_LV_D

0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0 0 30 minutes (b) 0 30 2.0 minutes (b)

10

20

30

40

10

20

30

40

CR_2_PU_LV_A CR_2_PU_LV_B CR_2_PU_LV_A CR_2_PU_LV_C CR_2_PU_LV_B CR_2_PU_LV_D CR_2_PU_LV_C

2.0

CR_2_PU_LV_D

0.0 -40

-30

-20

-10

-40

-30

-20

-10

0.0 0 4 hours (c) 0 42.0 hours (c)

10

20

30

40

10

20

30

40

CR_2_PU_LV_A CR_2_PU_LV_B CR_2_PU_LV_A CR_2_PU_LV_C CR_2_PU_LV_B CR_2_PU_LV_D CR_2_PU_LV_C

2.0

CR_2_PU_LV_D

0.0 -40

-30

-20

-40

-30

-20

0.0 -10 0 10 Distance from the -10 0 center (mm) 10 Distance from the center (mm)

20

30

40

20

30

40

Figure 15. Water profiles of the four standard mortar prisms containing two artificial cracks healed Figure 15. Water profiles of (b) the after four 30 standard mortar prisms with PU_LV: (a) after 5 min; min; and (c) after 4 h. containing two artificial cracks healed Figure 15. Water profiles of the four standard mortar prisms containing two artificial cracks healed with PU_LV: (a) after 5 min; (b) after 30 min; and (c) after 4 h.

with PU_LV: (a) after 5 min; (b) after 30 min; and (c) after 4 h.

Materials 2016, 9, 311

17 of 28

4.2.4. Water Profile Integrals The integrals of all previously shown water profiles (Figures 9–15 area in between the peaks and the baseline of the profiles), give an indication of the healing efficiency of the different samples for a very localized rectangular area at a height of 11.3 mm from the samples’ bottom surface. The mean values of these integrals and the corresponding standard errors have been summarized in Table 3. Table 3. Water profile integrals (ˆ10´1 g/cm2 ). Sample

Integral (ˆ10´1 g/cm2 )

UN CR_1 CR_1_PU_HV CR_1_PU_LV CR_2 CR_2_PU_HV CR_2_PU_LV

1.6 ˘ 0.5 1.8 ˘ 0.1 2.0 ˘ 0.3 1.7 ˘ 0.4 2.4 ˘ 0.3 2.3 ˘ 0.2 2.4 ˘ 0.2

On the local scale, i.e., the thin 80 mm long rectangular area above the capsule near the crack tip, there is a marginal healing effect of incorporating capsules filled with PU_HV or PU_LV to heal one or two cracks. The integrals calculated for the samples with self-healing properties do not differ significantly from the cracked ones. Moreover, based on these results little difference was observed in terms of healing ability depending on whether the polyurethane precursor with the high or low viscosity was used. This is in line with the previously recorded findings of the mean WFA/SCA ratios and their corresponding standard errors (Table 2). Both approaches indicate that the matrix around the crack(s) can still be quite affected by the presence of the crack in terms of water ingress even after (partial) healing. Within the cracks, the healing ability can often be considered more adequate. The water profiles along the cracks that are shown in the next section (Section 4.3) prove this. Nonetheless, the healing performance with encapsulated PU precursors could still be improved substantially, at least in the case where the specimens were exposed to water only after the healing agents had been able to react to a maximum extent with the moisture content of the surrounding mortar and the air humidity. 4.3. Water Profiling of Crack Healing along the Crack(s) 4.3.1. Cracked Condition All water profiles taken along the cracks of the cracked samples without self-healing properties look very similar (Figures 16 and 17). At the crack location (the graph sections in between the full and the dashed horizontal black lines), the water content amounts to 0.6–0.8 ˆ 10´1 g/cm3 . This water content range is maintained over the entire crack depth and does not really change as a function of exposure time. On the other hand, above the crack tips (the graph sections above the dashed horizontal black lines) the water content obviously increases with time. After only 5 min of exposure, an increased water content in the matrix of around 0.4 ˆ 10´1 g/cm3 can only be observed up to a sample height of 20 mm. After 30 min, the affected sample height is only slightly higher, while a 4-h exposure period resulted in a 0.4 ˆ 10´1 g/cm3 water content up to a sample height of 30 mm. These observations are valid for the individual water profiles along the crack for both the specimens containing one and two artificially induced standardized cracks.

Materials 2016, 9, 311

18 of 28

Materials 2016, 9, 311

18 of 28

Materials 2016, 9, 311

Distance from bottom (mm) Distance from bottom (mm)

40 40

30 30

20

18 of 28

40

CR_1_A (a) CR_1_A (a) 5 min 305 min min 4 30 hours min 4 hours

20

40

40

CR_1_B (b)

30

40

CR_1_B (b)

5 min 30 min 5 min 430 hours min

30

20 20

4 hours

30

20

CR_1_C (c) 5 min 30 min 5 min 430hours min

20

4 hours

10

10

10

30

CR_1_C (c)

10

10

10

0

0 00.0 2.0 -1 Water 0.0 content (× 10 g/cm³) 2.0 Water content (× 10 -1 g/cm³)

0 0 0.0 2.0 Water content (× 10 -1 g/cm³) 0.0 2.0 Water content (× 10 -1 g/cm³)

00.0 2.0 -1 Water 0.0 content (× 10 g/cm³) 2.0 Water content (× 10 -1 g/cm³)

Figure 16. Water profiles along the crack as a function of exposure time for the neutron radiography

Figure 16. Water profiles along the crack as a function of exposure time for the neutron radiography Figureof16.the Water along the crack as containing a function of exposure for(a) theCR_1_A; neutron radiography images threeprofiles standard mortar prisms one artificialtime crack: (b) CR_1_B; images of the three standard mortar prisms containing one artificial crack: (a) CR_1_A; (b) CR_1_B; images of the three standard mortar prisms containing one artificial crack: (a) CR_1_A; (b) CR_1_B; and (c) CR_1_C. and (c) CR_1_C. and (c) CR_1_C. Distance from bottom (mm) Distance from bottom (mm)

40 40 30 30 20 20

CR_2_A (a) CR_2_A (a) 5 min 30 5 min 430 hours min 4 hours

10 10 0

40 40 30 30 20 20

00.0 2.0 Water 0.0 content (× 10 -1 g/cm³) 2.0 Water content (× 10 -1 g/cm³)

CR_2_B (b) CR_2_B (b) 5 min 5 min 30 min min 430 hours 4 hours

40 40 30 30 20 20

10 10

10 10

0

0 0

0 0.0 2.0 0.0 content (× 10 -1 g/cm³) 2.0 Water -1 g/cm³) Water content (× 10 40 CR_2_C (c) 40 CR_2_C (c)

30 30 20 20

5 min 5 min 30 min min 430 hours 4 hours

0.0 2.0 0.0 content (× 10 -1 g/cm³) 2.0 Water -1 g/cm³) Water content (× 10 40 CR_2_C (c') 40 CR_2_C (c') 5 min 30 5 min 30 30 min min 430 hours 20 4 hours 20

10 10

10 10

0

0 0

0 0.0 2.0 0.0 content (× 10 -1 g/cm³) 2.0 Water Water content (× 10 -1 g/cm³)

CR_2_B (b') CR_2_B (b') 5 min 5 min 30 min min 430 hours 4 hours

0.0 2.0 0.0 2.0 Water content (× 10 -1-1g/cm³) Water content (× 10 g/cm³) Figure 17. Water profiles along the crack as a function of exposure time for the neutron radiography Figure 17. Water profiles along the crack as a function of exposure time for the neutron radiography images of the three standard prisms two artificialtime cracks: crack on the Figure 17. Water profiles alongmortar the crack as acontaining function of exposure for(a) theCR_2_A neutron radiography images of the three standard mortar prisms containing two artificial cracks: (a) CR_2_A crack on the left; of (a’)the CR_2_A on the right; (b) CR_2_B crack on two the left; (b’) CR_2_B on the right; (c) on images three crack standard mortar prisms containing artificial cracks:crack (a) CR_2_A crack left; (a’) CR_2_A crack on the right; (b) CR_2_B crack on the left; (b’) CR_2_B crack on the right; (c) CR_2_C crack on the left;on and (c’)right; CR_2_C crack on the right. the left; (a’) CR_2_A crack the (b) CR_2_B crack on the left; (b’) CR_2_B crack on the right; CR_2_C crack on the left; and (c’) CR_2_C crack on the right. (c) CR_2_C crack on the left; and (c’) CR_2_C crack on the right.

Materials 2016, 9, 311

19 of 28

Materials 2016, 9, 311

19 of 28

Distance from bottom (mm) Distance from bottom (mm)

Materials 2016, 9, 311 19 of 28 4.3.2. Healing of One Crack 4.3.2. Healing of One Crack 4.3.2. Healing of One Crack In comparison with the water profiles along the crack for the cracked specimens (Figure 16), In comparison withto thethe water profiles along the crack for the cracked specimens (Figure 16), the the ones corresponding samples containing encapsulated PU_HV (Figure(Figure 18) and comparison with water profiles along the crack for the cracked 16),PU_LV the ones In corresponding to thethe samples containing encapsulated PU_HV (Figure specimens 18) and PU_LV (Figure 19) (Figure 19) are characterized by a substantially reduced water content in between the full and the ones corresponding to the samples containing encapsulated PU_HV (Figure 18) and PU_LV (Figure 19) are characterized by a substantially reduced water content in between the full and the dashed dashed horizontal black lines of the graphs. In other words, there is a significant healing effect for the are characterized by aof substantially waterthere content in betweenhealing the full andfor thethe dashed horizontal black lines the graphs. Inreduced other words, is a significant effect crack ´1 g/cm3 still crack itself. Nevertheless, this reduced water content of around 0.2 ˆ 10 tends to increase horizontal black lines of reduced the graphs. In other words, there 0.2 is a×significant effect for the crack itself. Nevertheless, this water content of around 10−1 g/cm3healing still tends to increase a bit 3 after 4 h. −1 g/cm 3 still tends a bit with exposure time to aroundwater 0.4−1ˆcontent 10´3 1 g/cm The evolution in water content with itself. Nevertheless, this reduced of around 0.2 × 10 to increase a bit with exposure time to around 0.4 × 10 g/cm after 4 h. The evolution in water content with exposure −1 3 after exposure time and sample height also still exists for4 the sample matrix above the crack tip.exposure Only now, with exposure time to around 0.4 × 10 g/cm h. The evolution in water content with time and sample height also still exists for the sample matrix above the crack tip. Only now, the thetime affected sampleheight heightisalso isclearly clearly lower than mm (maximum mm). The most promising water andsample sample height stilllower exists for the sample matrix above theThe crack tip.promising Only now, the affected than 3030 mm (maximum 2626 mm). most water affected sample clearly lower than 30 mmfor (maximum 26 mm). The most promising water profiles in terms ofheight healing efficiency were obtained samplesfor CR_1_PU_LV_A and CR_1_PU_LV_C profiles in terms of ishealing efficiency were obtained samples CR_1_PU_LV_A and profiles in terms of healing efficiency were obtained for samples CR_1_PU_LV_A and18). (Figure 19). Such low water contents could not be achieved by using encapsulated PU_HV (Figure CR_1_PU_LV_C (Figure 19). Such low water contents could not be achieved by using encapsulated CR_1_PU_LV_C (Figure 19). Such lowlow water contents could notviscosity beshould achieved bythe using encapsulated This does not necessarily that the viscosity healing agent have preference PU_HV (Figure 18). Thismean does not necessarily mean that the low healing agent shouldbecause have PU_HV (Figure 18). This does not necessarily mean that the low viscosity healing agent should have thethe third sample because of the series (CR_1_PU_LV_B) clearly(CR_1_PU_LV_B) has a higher water content evenwater shows preference the third sample of the series clearly has aand higher the preference because the third sample of the series (CR_1_PU_LV_B) clearly has a higher water signs of water enrichment near the crack tip. content and even shows signs of water enrichment near the crack tip. content and even shows signs of water enrichment near the crack tip.

40 40 30 30 20 20

CR_1_PU_HV_A (a) CR_1_PU_HV_A (a) 5 min 530min min 30 min 4 hours 4 hours

40 40 30 30 20 20

CR_1_PU_HV_B (b) CR_1_PU_HV_B (b) 5 min 530min min 30 min 4 hours 4 hours

10 10

10 10

40 40 30 30 20 20

CR_1_PU_HV_C (c) CR_1_PU_HV_C (c) 5 min 530min min 30 min 4 hours 4 hours

10 10

0 0 0 0 0.0 0 0.0 0 0.0 2.0 2.0 2.0 0.0 2.0 0.0 2.0 0.0 2.0 Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Figure 18. Water profiles along the crack as a function of exposure time for the neutron radiography Figure 18. Water profiles along as of exposure exposuretime time forthe theneutron neutron radiography Figure profiles alongthe thecrack crack as aa function function of radiography images 18. of Water the three standard mortar prisms containing one artificial for crack healed with PU_HV:

Distance from bottom (mm) Distance from bottom (mm)

images ofofthe mortar containing one oneartificial artificialcrack crackhealed healed with PU_HV: images thethree threestandard standard mortar prisms prisms with PU_HV: (a) CR_1_PU_HV_A; (b) CR_1_PU_HV_B; and containing (c) CR_1_PU_HV_C. (a)(a) CR_1_PU_HV_A; CR_1_PU_HV_A;(b) (b)CR_1_PU_HV_B; CR_1_PU_HV_B; and and (c) (c) CR_1_PU_HV_C. CR_1_PU_HV_C. 40 40 30 30 20 20 10 10

CR_1_PU_LV_A (a) CR_1_PU_LV_A (a) 5 min 530min min 30 min 4 hours 4 hours

40 40 30 30 20 20 10 10

CR_1_PU_LV_B (b) CR_1_PU_LV_B (b) 5 min 530min min 30 min 4 hours 4 hours

40 40 30 30 20 20

CR_1_PU_LV_C (c) CR_1_PU_LV_C (c) 5 min 530min min 30 min 4 hours 4 hours

10 10

0 0 0 0 0.0 0 0.0 0 0.0 2.0 2.0 2.0 0.0 2.0 0.0 2.0 0.0 2.0 Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Water content (× 10 -1 g/cm³) Figure 19. Water profiles along the crack as a function of exposure time for the neutron radiography Figure Water profiles alongthe thecrack crack as aa function function of time neutron radiography Figure 19.19. Water profiles along as of exposure exposure timefor forthe thehealed neutron radiography images of the three standard mortar prisms containing one artificial crack with PU_LV: images of the three standard mortar prisms containing one artificial crack healed with PU_LV: images of the three standard mortar prisms containing one artificial crack healed with PU_LV: (a) CR_1_PU_LV_A; (b) CR_1_PU_LV_B; and (c) CR_1_PU_LV_C. CR_1_PU_LV_A; (b)CR_1_PU_LV_B; CR_1_PU_LV_B;and and(c) (c)CR_1_PU_LV_C. CR_1_PU_LV_C. (a)(a) CR_1_PU_LV_A; (b)

4.3.3. Healing of Two Cracks 4.3.3. HealingofofTwo TwoCracks Cracks 4.3.3. Healing In presence of two cracks, the amount of PU-based healing agent is seldom able to prevent presence oftwo twocracks, the amount of PU-based PU-based healing agent isisseldom able to to prevent InIn presence the amount of healing agent seldom able prevent increased waterof ingress incracks, both of them to a sufficient extent (Figures 20 and 21). With incorporation increased water ingress in both of them to a sufficient extent (Figures 20 and 21). With incorporation of encapsulated PU_HV (Figure 20), to only sample extent C showed substantial healing both cracks of increased water ingress in both of them a sufficient (Figures 20 and 21). Withofincorporation of encapsulated PU_HV (Figure 20), only sample C showed substantial of both cracks although thePU_HV crack on(Figure the left sideonly of the sample showed obvious signs ofhealing water enrichment near encapsulated 20), sample Cstill showed substantial healing of both cracks although although the crack on the left side of the sample still showed obvious signs of water enrichment near the crack on the left side of the sample still showed obvious signs of water enrichment near the crack

Materials 2016, 9, 311

20 of 28

Materials 2016, 9, 311

20 of 28

5 min 30 min 4 hours

20 10 0

Distance from bottom (mm)

0.0 40

CR_2_PU_HV_B (b)

30

5 min 30 min 4 hours

20 10 0

Distance from bottom (mm)

0.0

40

5 min 30 min 4 hours

20 10 0

Distance from bottom (mm)

0.0 40 30 20

5 min 30 min 4 hours

10 0 0.0 2.0 Water content (× 10 -1 g/cm³)

5 min 30 min 4 hours

20 10 0 0.0

40

2.0

CR_2_PU_HV_B (b')

30

5 min 30 min 4 hours

20 10 0 0.0 40

2.0 CR_2_PU_HV_C (c')

30

5 min 30 min 4 hours

20 10 0 0.0

2.0 CR_2_PU_HV_D (d)

CR_2_PU_HV_A (a')

30

2.0

CR_2_PU_HV_C (c)

30

40

2.0

Distance from bottom (mm)

30

Distance from bottom (mm)

CR_2_PU_HV_A (a)

Distance from bottom (mm)

40

Distance from bottom (mm)

Distance from bottom (mm)

tip. The other three samples were characterized by one crack that had healed properly, while the other the crackthe tip.mere The cracked other three samples were characterized by one crack that had healed properly, resembled condition. while the other resembled the mere cracked condition.

40 30 20

2.0 CR_2_PU_HV_D (d') 5 min 30 min 4 hours

10 0 0.0 2.0 Water content (× 10 -1 g/cm³)

Figure 20. Water profiles along the crack as a function of exposure time for the neutron radiography Figure 20. Water profiles along the crack as a function of exposure time for the neutron radiography images of the four standard mortar prisms containing two artificial cracks healed with PU_HV: images of the four standard mortar prisms containing two artificial cracks healed with PU_HV: (a) CR_2_PU_HV_A crack on the left; (a’) CR_2_PU_HV_A crack on the right; (b) CR_2_PU_HV_B (a) CR_2_PU_HV_A crack on the left; (a’) CR_2_PU_HV_A crack on the right; (b) CR_2_PU_HV_B crack on the left; (b’) CR_2_PU_HV_B crack on the right; (c) CR_2_PU_HV_C crack on the left; (c’) crack on the left; (b’) CR_2_PU_HV_B crack on the right; (c) CR_2_PU_HV_C crack on the left; (c’) CR_2_PU_HV_C crack on the right; (d) CR_2_PU_HV_D crack on the left; and (d’) CR_2_PU_HV_D CR_2_PU_HV_C crack on the right; (d) CR_2_PU_HV_D crack on the left; and (d’) CR_2_PU_HV_D crack on the right. crack on the right.

Materials 2016, 9, 311

21 of 28

CR_2_PU_LV_A (a) 5 min 30 min 4 hours

20 10 0 40

CR_2_PU_LV_B (b)

30

5 min 30 min 4 hours

20 10 0

Distance from bottom (mm)

0.0

40

5 min 30 min 4 hours

20 10 0

Distance from bottom (mm)

0.0 40 30 20

5 min 30 min 4 hours

10 0 0.0 2.0 Water content (× 10 -1 g/cm³)

5 min 30 min 4 hours

20 10 0 0.0

40

2.0

CR_2_PU_LV_B (b')

30

5 min 30 min 4 hours

20 10 0 0.0

40

2.0

CR_2_PU_LV_C (c')

30

5 min 30 min 4 hours

20 10 0 0.0

2.0 CR_2_PU_LV_D (d)

CR_2_PU_LV_A (a')

30

2.0

CR_2_PU_LV_C (c)

30

40

2.0 Distance from bottom (mm)

Distance from bottom (mm)

0.0

Distance from bottom (mm)

30

Distance from bottom (mm)

40

Distance from bottom (mm)

21 of 28

Distance from bottom (mm)

Materials 2016, 9, 311

40 30 20

2.0

CR_2_PU_LV_D (d') 5 min 30 min 4 hours

10 0 0.0 2.0 Water content (× 10 -1 g/cm³)

Figure 21. Water profiles along the crack as a function of exposure time for the neutron radiography Figure 21. Water profiles along the crack as a function of exposure time for the neutron radiography images of the four standard mortar prisms containing two artificial cracks healed with PU_LV: images of the four standard mortar prisms containing two artificial cracks healed with PU_LV: (a) CR_2_PU_LV_A crack on the left; (a’) CR_2_PU_LV_A crack on the right; (b) CR_2_PU_LV_B (a) CR_2_PU_LV_A crack on the left; (a’) CR_2_PU_LV_A crack on the right; (b) CR_2_PU_LV_B crack on the left; (b’) CR_2_PU_LV_B crack on the right; (c) CR_2_PU_LV_C crack on the left; (c’) crack on the left; (b’) CR_2_PU_LV_B crack on the right; (c) CR_2_PU_LV_C crack on the left; (c’) CR_2_PU_LV_C crack on the right; (d) CR_2_PU_LV_D crack on the left; and (d’) CR_2_PU_LV_D CR_2_PU_LV_C crack on the right; (d) CR_2_PU_LV_D crack on the left; and (d’) CR_2_PU_LV_D crack on the right. crack on the right.

For the healed cracks, an increase in water content with exposure time could still be observed in For theregion healed cracks, increase water the content with exposure time could beprofiles observed the crack (the graphan sections in in between full and dashed lines), while the still water of in thethe crack region (the graph sections in between the full and dashed lines), while the water profiles unhealed ones did not change anymore after 5 min of exposure. It must be mentioned though of thethat unhealed ones did not change min of exposure. It must mentioned that the unhealed ones still clearlyanymore differed after from 5the cracked ones because thebewater contentthough over the the unhealed ones still clearly differed from the cracked ones because the water content over the entire

Materials 2016, 9, 311

Materials 2016, 9, 311

22 of 28

22 of 28

crack depth was much more variable. Usually there was a reduced water content near the bottom of entire the sample enrichment near the Usually crack tip. Thewas water contentwater of thecontent matrixnear above crack and depthwater was much more variable. there a reduced thethe unhealed crack tip strongly resembled thosenear of the The affected sample height bottom of the sample and water enrichment the cracked crack tip.specimens. The water content of the matrix above was again around 30 mm after 4 h of exposure. For the healed cracks, these values ranged between the unhealed crack tip strongly resembled those of the cracked specimens. The affected sample height 22 and 28 again mm. Finally, be noted that the more or less adequately healed cracks couldbetween be linked was around it 30should mm after 4 h of exposure. For the healed cracks, these values ranged systematically to the crackit that wasbecreated (themore plateor that pulled out first).cracks could be 22 and 28 mm. Finally, should noted first that the lesswas adequately healed For the test series to containing the was encapsulated low viscosity water profiles linked systematically the crack that created first (the plate thatpolyurethane, was pulled outthe first). For the test series containing the encapsulated low viscosity polyurethane, the water clearly indicate that the available amount of healing agent could only heal one out of twoprofiles cracks in indicate that the available amount healing agent couldthose only heal of twofirst cracks allclearly studied samples (Figure 21). The healedofcracks were always that one wereout created andinthe all studied (Figure 21). The healed were always those that were first and water contentsamples that could be achieved in themcracks through self-healing is lower thancreated when using thethe high water content that could be achieved in them through self-healing is lower than when using the high viscosity healing agent. The matrix above the crack tip on the left side was barely affected anymore healing agent.ofThe the crack tip onofthe side wascracks barely on affected anymore byviscosity the earlier presence thematrix crack. above The water profiles theleft unhealed the right side of by the earlier presence of the crack. The water profiles of the unhealed cracks on the right side thethe the specimen again strongly resemble those of the cracked test series. The only difference is of that specimen again strongly resemble those of the cracked test series. The only difference is that the water water content near the bottom of the samples is clearly lower than near the crack tip. Most probably, content near the bottom of the samples is clearly lower than near the crack tip. Most probably, a large a large portion of the crack close to the exposure surface was filled with PU, but not completely. As a portion of the crack close to the exposure surface was filled with PU, but not completely. As a consequence, water was still able penetrate and reach the crack tip where almost no healing agent was consequence, water was still able penetrate and reach the crack tip where almost no healing agent present. It gives the typical water enrichment near the crack tip, which is also visible in the neutron was present. It gives the typical water enrichment near the crack tip, which is also visible in the radiography images (e.g., in Figure 8). neutron radiography images (e.g., in Figure 8). 4.4. Spread Region of the Healing Agents in the Crack 4.4. Spread Region of the Healing Agents in the Crack 4.4.1. Healing of One Crack 4.4.1. Healing of One Crack After splitting the specimens, the crack faces of each crack were placed next to each other and After splitting the specimens, the crack faces of each crack were placed next to each other and a a picture of the total crack surface was taken. For each of the crack faces, the total area covered by picture of the total crack surface was taken. For each of the crack faces, the total area covered by polyurethane was determined using photo editing software. The percentage of the crack face covered polyurethane was determined using photo editing software. The percentage of the crack face covered byby polyurethane coverageof ofthe thecrack crackface. face.InIn this way, spread polyurethanewas wasthen thendenoted denotedas as the the surface surface coverage this way, thethe spread region of the polyurethane inside the crack could be determined. region of the polyurethane inside the crack could be determined. For the specimens the pictures picturesofofthe thecrack crackfaces faces are represented For the specimenscontaining containingone one healed healed crack, crack, the are represented in in Figure 22. The mean surface coverage of both crack faces for each crack is also shown in this figure. Figure 22. The mean surface coverage of both crack faces for each crack is also shown in this figure.

Figure 22. Visualization of the spread region of the healing agents in the crack for mortar prisms with Figure 22. Visualization of the spread region of the healing agents in the crack for mortar prisms one healed crack. Mean surface coverage of both crack faces is indicated for each individual specimen. with one healed crack. Mean surface coverage of both crack faces is indicated for each individual (a–c) the surface coverage with the high viscosity PU-based healing agent; (d–f) the surface coverage specimen. (a) CR_1_PU_HV_A; (b) CR_1_PU_HV_B; (c) CR_1_PU_HV_C; (d) CR_1_PU_LV_A; with the low viscosity PU-based healing agent. (e) CR_1_PU_LV_B; and (f) CR_1_PU_LV_C.

Materials 2016, 9, 311

23 of 28

For the samples where the crack was healed with the high viscosity PU-based healing agent (Figure 22a–c), the surface coverage ranged from 79% to 92%. Clearly this healing agent is able to fill up most of the crack volume. This can also be noticed in the neutron radiographs (Figure 7a). The crack has a white color over the entire depth, which means that the crack was filled with healing agent to a large extent and no water could enter anymore through the crack. The samples where the crack was healed with the low viscosity PU-based healing agent (Figure 22d–f) show a much lower surface coverage. When analyzing the crack faces, it can be seen that there is a dark color on both crack faces due to the leakage of polyurethane. However, the crack was not filled by the hardened polyurethane at those regions. A possible explanation for the dark color of the crack faces is that some minor amount of polyurethane might have been absorbed by the mortar matrix. Only at the bottom of the crack hardened polyurethane was noticed at both crack faces. This can be attributed to the very low viscosity of this PU-based healing agent. When the plates were pulled out, most of the PU leaked to the bottom of the crack and some even flowed out of the crack onto the bottom surface of the specimens. This excess PU at the bottom was then removed when the outer layer of the bottom surface was cut off to obtain a flat test face (see Section 3.2). Thus, for the specimens with the low viscosity PU-based healing agent, only 13% to 21% of the crack at the bottom was filled with healing agent. In the neutron radiography images, the incomplete filling of the crack can also be seen. In Figure 7b the inside of the crack for specimen CR_1_PU_LV_A is not completely showing the white color. There is a zone inside the crack that remains dark-colored. It further indicates that the crack is indeed not filled with the healing agent in that region. From the analysis of the crack faces of the specimens it can be concluded that the high viscosity healing agent is able to fill up much more crack volume than the low viscosity healing agent. However, based on a mainly visual evaluation of the water uptake in the neutron radiographs (Section 4.1.2) it was concluded that the low viscosity healing agent worked more efficiently to reduce the water ingress through the crack. This might seem contradictory with the results of the spread region of the healing agents. However, it is possible that even though the low viscosity healing agent does not fill most of the crack volume, there is a good sealing of the crack near the bottom surface of the specimen in contact with water during the capillary absorption test. 4.4.2. Healing of Two Cracks The pictures of the crack faces of the samples containing two healed cracks are shown in Figure 23. For the samples where the cracks were healed with the high viscosity PU-based healing agent, it can be seen that the crack which was created first (Figure 23a,d,f,g) has the highest surface coverage (ranging from 84% to 97%). This is comparable to the values of the surface coverage that were found for the specimens with one crack healed with the same healing agent (Figure 22a–c). The surface coverage of the crack that was created last varies greatly among the different specimens. For sample CR_2_PU_HV_B only a very small amount of healing agent (surface coverage of 17.6%) was noticed in the second crack (Figure 23c). For sample CR_2_PU_HV_C the surface coverage of the second crack was 81.8% (Figure 23e). This means that the healing agent in the capsule was much more equally distributed between both cracks and the amount of healing agent was enough to fill most of the two crack volumes. The fact that both cracks are filled well can also be seen in Figure 8a where both cracks have a white color in the neutron radiographs. However, from the image of the water absorption after 4 h, it can be seen that there is a water enrichment at the crack tip of the left crack. This means that the sealing of the left crack (crack that was created last) was not perfect and due to capillary action there was some water that could penetrate to the zone of the crack that was not covered with polyurethane. For the sample where two cracks were healed with the low viscosity PU-based healing agent, a much lower surface coverage was found for the crack that was created first (Figure 23i,k,m,o) compared to the cracks healed with the high viscosity PU-based healing agent. In accordance with the specimens containing just one healed crack, the hardened polyurethane was only found at the lower part of the crack. Again there was quite some variation in surface coverage for the crack

Materials 2016, 9, 311

Materials 2016, 9, 311

24 of 28

24 of 28

that was created last. For the samples CR_2_PU_LV_A and CR_2_PU_LV_B the surface coverage of second the second 23j,l) more than half the surface coverage of the crack. crack crack (Figure(Figure 23j,l) was stillwas morestill than half the surface coverage of the first crack. For first samples ForCR_2_PU_LV_C samples CR_2_PU_LV_C and CR_2_PU_LV_D, lessagent healing was in thecrack second and CR_2_PU_LV_D, much less much healing wasagent found in found the second crack (Figure 23n,p) compared to the first crack. (Figure 23n,p) compared to the first crack.

Figure 23. Visualization of the spread region of the healing agents in the crack for mortar prisms with Figure 23. Visualization of the spread region of the healing agents in the crack for mortar prisms two healed cracks. Mean surface coverage of both crack faces is indicated for the cracks of each with two healed cracks. Mean surface coverage of both crack faces is indicated for the cracks individual specimen. of each individual specimen. (a) CR_2_PU_HV_A (left crack); (b) CR_2_PU_HV_A (right crack); (c) When CR_2_PU_HV_B CR_2_PU_HV_B (rightabsorption crack); (e) CR_2_PU_HV_C (left crack); comparing(left the crack); visual (d) evaluation of the water in specimen CR_2_PU_LV_B (f) CR_2_PU_HV_C (right crack); (g) CR_2_PU_HV_D (left crack); (h) CR_2_PU_HV_D (right crack); (Figure 8b) to the result of the crack surface coverage (Figure 23k,l) it can be seen that no water enters (i) CR_2_PU_LV_A (left crack); (j) CR_2_PU_LV_A (right crack); (k) CR_2_PU_LV_B (left crack); through the left crack. This means that the surface coverage of 39% at the bottom of the left crack was (l) CR_2_PU_LV_B (right crack); (m) CR_2_PU_LV_C crack);in (n)the CR_2_PU_LV_C (right crack); sufficient to prevent the water from entering the crack. (left However, crack on the right side, water (o) CR_2_PU_LV_D (left crack); and (p) CR_2_PU_LV_D (left crack). enters through the whole crack. Apparently, the sealing of this second crack at the bottom was not

sufficient to prevent ingress of water. When comparing the visual evaluation of the water absorption in specimen CR_2_PU_LV_B (Figure 8b) toUptake the result of the crack surface coverage (Figure 23k,l) it can be seen that no water enters 4.5. Water Immediately after Triggering the Healing Mechanism through the left crack. This means that the surface coverage of 39% at the bottom of the left crack was In the previous experiments, the specimens with self-healing properties were only brought in sufficient to prevent the water from entering the crack. However, in the crack on the right side, water contact with water after the healing agents had enough time to form a foam and harden. Usually this enters whole crack.these Apparently, the sealing of this second at the bottom was not takesthrough around the 24 h. However, are circumstances that are not oftencrack encountered in everyday sufficient prevent ingress of water. concretetoenvironments. Upon cracking there may already be contact with water, and the PU healing must then also work properly. Neutron radiography images of specimens containing one artificial 4.5. Water Uptake Immediately after Triggering the Healing Mechanism crack, where the release of either of the two healing agents (PU_HV and PU_LV) was triggered and exposure to water began not long the afterspecimens the removal of the thin metalproperties plates (D specimens In the previous experiments, with self-healing were onlyPU_HV_1 brought in and PU_LV_1: after 15 min; E specimens PU_LV_2 and PU_LV_2: after 3 h), show that this situation contact with water after the healing agents had enough time to form a foam and harden. Usually is not problematic at all (Figure 24). In contrast with the specimens where the PU in cracks had

Materials 2016, 9, 311

25 of 28

this takes around 24 h. However, these are circumstances that are not often encountered in everyday concrete environments. Upon cracking there may already be contact with water, and the PU healing must then also work properly. Neutron radiography images of specimens containing one artificial crack, where the release of either of the two healing agents (PU_HV and PU_LV) was triggered and exposure to water began not long after the removal of the thin metal plates (D specimens PU_HV_1 and PU_LV_1: after 15 min; E specimens PU_LV_2 and PU_LV_2: after 3 h), show that this situation is 2016, 9, 311 of 28 notMaterials problematic at all (Figure 24). In contrast with the specimens where the PU in cracks had25already hardened, these specimens clearly show a much more pronounced healing efficiency as there is almost these specimens a much pronounced efficiency as there noalready visible hardened, water front. This behaviorclearly can beshow explained asmore follows. Both PUhealing precursors that have been is almost no visible water front. This behavior can be explained as follows. Both PU precursors that in used in this research require water to initiate the foaming reaction. When bringing the specimen have been used in this research require water to initiate the foaming reaction. When bringing the contact with water, upon release of the PU precursor, it apparently creates the optimal conditions for specimen in contact with water, upon release of the PU precursor, it apparently creates the optimal a profound foaming reaction. Normally, the moisture content of the cementitious matrix around the conditions for a profound foaming reaction. Normally, the moisture content of the cementitious crack and the air humidity is sufficient to generate the foam. Still, with even more water around, the matrix around the crack and the air humidity is sufficient to generate the foam. Still, with even more reaction conditions can be even more beneficial. This holds true for both the high and low viscosity water around, the reaction conditions can be even more beneficial. This holds true for both the high healing agents. and low viscosity healing agents.

Figure 24. Visualization of the water uptake as a function of exposure time using neutron radiography Figure 24. Visualization of the water uptake as a function of exposure time using neutron radiography for four mortar prisms immediately after release of the PU-based healing agent: (a) crack healing with for four mortar prisms immediately after release of the PU-based healing agent: (a) crack healing with PU_HV (specimens: PU_HV_1-2); and (b) crack healing with PU_LV (specimens: PU_LV_1-2). PU_HV (specimens: PU_HV_1-2); and (b) crack healing with PU_LV (specimens: PU_LV_1-2).

5. Conclusions 5. Conclusions Neutron radiography imaging was found to be a very effective technique to assess the healing Neutron radiography imaging was found to be a very effective technique to assess the healing performance of two encapsulated polyurethane-based healing agents—one with high and one with performance of two encapsulated polyurethane-based healing4agents—one with high with low viscosity—inside mortar upon cracking and a subsequent h capillary sorption test.and Theone water low viscosity—inside mortar upon be cracking andvery a subsequent uptake as a function of time could visualized clearly. 4 h capillary sorption test. The water uptakeFrom as a function of time could be visualized very clearly. neutron radiography image per test a mere visual interpretation of one representative From a mere visual interpretation of one representative radiography image per series, series, it seemed that one artificially induced crack, 300 µm inneutron width and 20 mm deep, could betest healed it seemed that oneby artificially crack, 300 µm inagent. widthHowever, and 20 mm deep, could be WFA/SCA healed more more efficiently means ofinduced the low viscosity healing comparison of the efficiently by means of the low viscosity healing agent. However, comparison of the WFA/SCA ratios ratios for all samples could not really confirm this finding as the WFA/SCA ratios varied considerably forwithin all samples could not When really confirm finding the WFA/SCA ratios varied within each test series. aiming atthis healing twoasartificially induced cracks, the considerably same conclusions cantest be drawn. Evaluation of the water profiles which provided information onsame a more local scale,can i.e., be each series. When aiming at healing two artificially induced cracks, the conclusions a 80 mm long, thin rectangular area atwhich heightprovided 11.3 mminformation from the bottom of thelocal samples, drawn. Evaluation of the water profiles on a more scale, as i.e.,well a 80as mm their integrals also did not indicate a substantial difference in healing performance depending on the long, thin rectangular area at height 11.3 mm from the bottom of the samples, as well as their integrals viscosity PU precursor. Nevertheless, the in low viscosity healing agent clearly tends to heal only one also did notofindicate a substantial difference healing performance depending on the viscosity of PU of two cracks more effectively, i.e., the one that was created first. Additional evaluation of water precursor. Nevertheless, the low viscosity healing agent clearly tends to heal only one of two cracks profiles taken along theone cracks spread regions of self-healing substances onto the crack faces more effectively, i.e., the thatand wasthe created first. Additional evaluation of water profiles taken along after the capillary sorption tests later confirmed this. Apart from that, it should be noted that the now incorporated amount of each PU precursor in the capsules was insufficient to heal one or two cracks completely. The above-mentioned conclusions refer to the scenario in which the polyurethane foam was given enough time to harden before the mortar specimens were brought in direct contact with water. Alternatively, some specimens with one crack were immediately exposed to water after initiating

Materials 2016, 9, 311

26 of 28

the cracks and the spread regions of self-healing substances onto the crack faces after the capillary sorption tests later confirmed this. Apart from that, it should be noted that the now incorporated amount of each PU precursor in the capsules was insufficient to heal one or two cracks completely. The above-mentioned conclusions refer to the scenario in which the polyurethane foam was given enough time to harden before the mortar specimens were brought in direct contact with water. Alternatively, some specimens with one crack were immediately exposed to water after initiating capsule breakage and release of the PU precursor. Based on visual assessment of the neutron radiography images obtained during a two hour capillary sorption test it was observed that in abundant presence of water the foaming reaction was found to be much more profound and capable of preventing further water ingress almost completely in the crack zone. This was the case for both the high and the low viscosity healing agent. It implies a great potential of using encapsulated PU-based healing agents in very humid environments where the foaming reaction does not have to depend on the moisture content of the mortar and the air humidity alone. Acknowledgments: This research under the program SHE (Engineered Self-Healing materials), project ISHECO (Impact of Self-Healing Engineered materials on steel COrrosion of reinforced concrete) was funded by SIM (Strategic Initiative Materials in Flanders) and IWT (Agency for Innovation by Science and Technology). The financial support from the foundations for this study is gratefully appreciated. Kim Van Tittelboom is a postdoctoral fellow of the Research Foundation—Flanders (FWO) (project number 12A3314N), and acknowledges its support. The authors would like to thank the Paul Scherrer Institute (PSI) for granting neutron beam time at the NEUTRA SINQ facility to execute their research proposal (ID No. 20141533). A special word of thanks also goes to Mr. Jan Hovind of PSI for his technical assistance during the measurements. Author Contributions: Philip Van den Heede, Bjorn Van Bellegehem, Natalia Alderete and Kim Van Tittelboom conceived, designed and performed all experiments. They also all contributed in analyzing the obtained neutron radiography images. Philip Van den Heede, Bjorn Van Belleghem and Kim Van Tittelboom wrote the paper. Nele De Belie was the promoter of all authors regarding their PhD or postdoctoral research and hence provided input at all stages of the research. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations The following abbreviations are used in this manuscript: CR HV LV PSI PU UN WFA/SCA

Cracked High viscosity Low viscosity Paul Scherrer Institute Polyurethane Uncracked Water front/sample cross-section area ratio

References 1. 2. 3.

4.

5.

Van Tittelboom, K.; De Belie, N. Self-healing in cementitious materails—A review. Materials 2013, 6, 2182–2217. [CrossRef] Van Tittelboom, K.; De Belie, N.; Van Loo, D.; Jacobs, J. Self-healing efficiency of cementitious materials containing tubular capsules filled with healing agent. Cem. Concr. Compos. 2011, 33, 497–505. [CrossRef] Maes, M.; Van Tittelboom, K.; De Belie, N. The efficiency of self-healing cementitious materials by means of encapsulated polyurethane in chloride containing environments. Constr. Build. Mater. 2014, 71, 528–537. [CrossRef] Feiteira, J.; Gruyaert, E.; De Belie, N. Self-healing of dynamic concrete cracks using polymer precursors as encapsulated healing agents. In Proceedings of the Concrete Solutions 5th International Conference on Concrete Repair, Belfast, UK, 1–3 September 2014; pp. 65–69. Aldea, C.-M.; Shah, S.P.; Karr, A. Permeability of cracked concrete. Mater. Struct. 1999, 32, 370–376. [CrossRef]

Materials 2016, 9, 311

6. 7.

8.

9.

10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20.

21.

22.

23. 24. 25.

26.

27 of 28

Wang, K.; Jansen, D.C.; Shah, S.P.; Karr, A.F. Permeability study of cracked concrete. Cem. Concr. Res. 1997, 27, 381–393. [CrossRef] Wang, L.C. Experimental study on water absorption by concrete damaged by uniaxial loading. In Proceedings of the 4th International Conference on the Durability of Concrete Structures, West Lafayette, IN, USA, 23–26 July 2014; pp. 198–204. Van Belleghem, B.; Rodrigo, M.; Dewanckele, J.; Van den Steen, N.; De Graeve, I.; Deconinck, J.; Cnudde, V.; Van Tittelboom, K.; De Belie, N. Capillary water absorption in cracked and uncracked mortar—A comparison between experimental study and finite element analysis. Constr. Build. Mater. 2016, 110, 154–162. [CrossRef] Van Belleghem, B.; Dewanckele, J.; De Belie, N.; Cnudde, V. Analysis and visualization of water uptake in cracked and healed mortar by water absorption tests and X-ray radiography. In Proceedings of the 4th International Conference on Concrete Repair, Rehabilitation and Retrofitting, Leipzig, Germany, 5–7 October 2015; pp. 45–53. Audenaert, K. Transport mechanisms of self-compacting concrete related to carbonation and chloride penetration. Ph.D. Thesis, Ghent University, Ghent, Belgium, 2006. (In Dutch) Aldea, C.-M.; Ghandehari, M.; Shah, S.P.; Karr, A.F. Combined effect of cracking and water permeability of concrete. In Proceedings of the 14th Engineering Mechanics Conference, Austin, TX, USA, 21–24 May 2000. Hall, C. Water sorptivity of mortars and concretes: A review. Mag. Concr. Res. 1989, 14, 51–61. [CrossRef] McCarter, W.J.; Ezirim, H.; Emerson, M. Absorption of water and chloride into concrete. Mag. Concr. Res. 1992, 44, 31–37. [CrossRef] Hall, C.; Tse, T.K.-M. Water movement in porous building materials—VII. The sorptivity of mortars. Build. Environ. 1986, 21, 113–118. [CrossRef] Martys, N.S.; Ferraris, C.F. Capillary transport in mortars and concrete. Cem. Concr. Res. 1997, 27, 747–760. [CrossRef] Castro, J.; Bentz, D.; Weiss, J. Effecto of sample conditioning on the water absorption of concrete. Cem. Concr. Compos. 2011, 33, 805–813. [CrossRef] Schröfl, C.; Mechtcherine, V.; Kaestner, A.; Vontobel, P.; Hovind, J.; Lehmann, E. Transport of water through strain-hardening cement-based composite (SHCC) applied on top of cracked reinforced concrete slabs with and without hydrophobization of cracks—Investigation by neutron radiography. Constr. Build. Mater. 2015, 76, 70–86. [CrossRef] Lieboldt, M.; Mechtcherine, V. Capillary transport of water through textile-reinforced concrete applied in repairing and/or strengthening cracked RC structures. Cem. Concr. Res. 2013, 52, 53–62. [CrossRef] Gardner, D.; Jefferson, A.; Hoffman, A. Investigation of capillary flow in discrete cracks in cementitious materials. Cem. Concr. Res. 2012, 42, 972–981. [CrossRef] Dewanckele, J.; De Kock, T.; Fronteau, G.; Derluyn, H.; Vontobel, P.; Dierick, M.; Van Hoorebeke, L.; Jacobs, P.; Cnudde, V. Neutron radiography and X-ray computed tomography for quantifying weathering and water uptake processes inside porous limestone used as building material. Mater. Charact. 2014, 88, 86–99. [CrossRef] Cnudde, V.; Dierick, M.; Vlassenbroeck, J.; Masschaele, B.; Lehmann, E.; Jacobs, P.; Van Hoorebeke, L. High-speed neutron radiography for monitoring the water absorption by capillarity in porous materials. Nucl. Instrum. Meth. B 2008, 266, 155–163. [CrossRef] Roels, S.; Carmeliet, J.; Hens, H.; Adan, O.; Brocken, H.; Cerny, R.; Pavlik, Z.; Ellis, A.T.; Hall, C.; Kumaran, K.; et al. A comparison of different techniques to quantify moisture content profiles in porous building materials. J. Therm. Envel. Build. Sci. 2004, 27, 261–276. [CrossRef] Baker, P.H.; Bailly, D.; Campbell, M.; Galbraith, G.H.; McLean, R.C.; Poffa, N.; Sanders, C.H. The application of X-ray absorption to building moisture transport studies. Measurement 2007, 40, 951–959. [CrossRef] Pleinert, H.; Sadouki, H.; Wittmann, F.H. Determination of moisture distributions in porous building materials by neutron transmission analysis. Mater. Struct. 1998, 31, 218–224. [CrossRef] Brew, D.R.M.; de Beer, F.C.; Radebe, M.J.; Nshimirimana, R.; McGlinn, P.J.; Aldridge, L.P.; Payne, T.E. Water transport through cement-based barriers—A preliminary study using neutron radiography and tomography. Nucl. Instrum. Meth. A 2009, 605, 163–166. [CrossRef] El Abd, A.; Czachor, A.; Milczarek, J. Neutron radiography determination of water diffusivity in fired clay brick. Appl. Radiat. Isot. 2009, 67, 556–559. [CrossRef] [PubMed]

Materials 2016, 9, 311

27. 28. 29. 30.

31.

32. 33.

34. 35. 36.

37.

38.

28 of 28

Weiss, J.; Geiker, M.R.; Hansen, K.K. Using X-ray transmission/attenuation to quantify fluid absorption in cracked concrete. Int. J. Mater. Struct. Int. 2015, 9, 3–20. [CrossRef] Roels, S.; Carmeliet, J. Analysis of moisture flow in porous building materials using microfocus X-ray radiography. Int. J. Heat Mass Transf. 2006, 49, 4762–4772. [CrossRef] Roels, S.; Vandersteen, K.; Carmeliet, J. Measuring and simulating moisture uptake in a fractured porous medium. Avd. Water Resour. 2003, 26, 237–246. [CrossRef] Van Tittelboom, K.; Snoeck, D.; Vontobel, P.; Wittmann, F.H.; De Belie, N. Use of neutron radiography and tomography to visualize the autonomous crack sealing efficiency in cementitious materials. Mater. Struct. 2013, 46, 105–121. [CrossRef] Snoeck, D.; Steuperaert, S.; Van Tittelboom, K.; Dubruel, P.; De Belie, N. Visualization of water penetration in cementitious materials with superabsorbent polymers by means of neutron radiography. Cem. Concr. Res. 2012, 42, 1113–1121. [CrossRef] Zhang, P.; Wittmann, F.H.; Zhao, T.; Lehmann, E.H. Neutron imaging of water penetration into cracked steel reinforced concrete. Phys. B Condens. Matter 2010, 405, 1866–1871. [CrossRef] Zhang, P.; Wittmann, F.H.; Zhao, T.; Lehmann, E.H.; Jin, Z. Visualization and quantification of water movement in porous cement-based materials by real time thermal neutron radiography: Theoretical analysis and experimental study. Sci. China Technol. Sci. 2010, 53, 1198–1207. [CrossRef] Kanematsu, M.; Maruyama, I.; Noguchi, T.; Iikura, H.; Tsuchiya, N. Quantification of water penetration into concrete through cracks by neutron radiography. Nucl. Instrum. Meth. A 2009, 605, 154–158. [CrossRef] Zhang, P.; Wittmann, F.H.; Haist, M.; Müller, H.S.; Vontobel, P.; Zhao, T. Water penetration into micro-cracks in reinforced concrete. Restor. Build. Monum. 2014, 20, 85–94. Zhang, P.; Wittmann, F.H.; Zhao, T.; Lehmann, E.H.; Vontobel, P. Neutron radiography, a powerful method to determine time-dependent moisture distributions in concrete. Nucl. Eng. Des. 2011, 241, 4758–4766. [CrossRef] Van Tittelboom, K.; Wang, J.; Araujo, A.; Snoeck, D.; Gruyaert, E.; Debbaut, B.; Derluyn, H.; Cnudde, V.; Tsangouri, E.; Van Hemelrijck, D.; et al. Comparison of different approaches for self-healing concrete in a large-scale lab test. Constr. Build. Mater. 2016, 107, 125–137. [CrossRef] Van Tittelboom, K.; Gruyaert, E.; De Backer, P.; Moerman, W.; De Belie, N. Self-repair of thermal cracks in concrete sandwich panels. Struct. Concr. 2015, 16, 273–288. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).