Self-healing materials: A review of advances in materials, evaluation, characterization and monitoring techniques

Composites Part B 87 (2016) 92e119

Contents lists available at ScienceDirect

Composites Part B journal homepage: www.elsevier.com/locate/compositesb

Review

Self-healing materials: A review of advances in materials, evaluation, characterization and monitoring techniques D.G. Bekas, K. Tsirka, D. Baltzis, A.S. Paipetis* Department of Materials Science and Engineering, University of Ioannina, Ioannina 45110, Greece

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 July 2015 Received in revised form 3 September 2015 Accepted 27 September 2015 Available online 31 October 2015

Self-healing materials are attracting increasing interest of the research community, over the last decades, due to their efficiency in detecting and “autonomically” healing damage. Numerous attempts are being presented every year focusing on the development of different self-healing systems as well as their integration to large scale production with the best possible propertyecost relationship. The current work aims to present the most recent breakthroughs in these attempts from many different research groups published during the last five years. The current review focuses in polymeric systems and their composites. The reviewed literature is presented in three distinct categories, based on three different scopes of interest. These categories are (i) the materials and systems employed, (ii) the experimental techniques for the evaluation of materials properties and self-healing efficiency of the materials/structures and (iii) the characterization techniques utilized in order to evaluate (off-line) and monitor (on-line) the healing efficiency of the proposed systems. Published works are presented separately in all the different categories, thus the interested reader is advised to follow the structure of the review and refer to the chapter of interest. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Self-healing materials A. Polymerematrix composites (PMCs) B. Mechanical properties C. Mechanical testing

1. Introduction Self-healing materials are a relatively new class of smart materials that possess the ability to fully or partially recover a functionality that is mediated by operational use. Local functionality loss can be defined as the situation when a section of a material or structure exhibits degraded performance when compared with the rest of the material/structure. Global functionality loss can be defined as the situation when the material or structure exhibits degraded performance when compared to its properties prior to any exposure to operational loads. This work focuses on self-healing polymers and their composites. The incorporation of healing agents in polymeric materials inadvertently leads in a new material with altered properties when compared to the material that does not possess the healing functionality. The performance and life-time of the new composite in conjunction with the efficiency of the selfhealing functionality are of primary importance as they are often competing with each other.

* Corresponding author. E-mail address: [email protected] (A.S. Paipetis). http://dx.doi.org/10.1016/j.compositesb.2015.09.057 1359-8368/© 2015 Elsevier Ltd. All rights reserved.

The scope of the first section of this work is to describe the three primary self-healing approaches (intrinsic, capsule based and vascular) as well as the critical issues and challenges associated with each approach. A review of the literature on the materials that have been used as healing agents over the last five years is presented. The aim of the second section of this review is to present the most frequently used testing procedures and specimen geometries found in research publications during the last five years. Special consideration is given to the ones that provide, either qualitatively or quantitatively, insight on the self-healing performance of the composites. Associated ASTM standards are also presented. Finally, in order to gain an overall insight into the behavior of self-healing materials, their structure, performance and selfhealing effectiveness is evaluated via the use of various characterization and monitoring techniques. The combination of mechanical testing and materials characterization techniques can exploit the actual capabilities as well as restraints associated with these materials. The ongoing development of the microscopic, spectroscopic and other characterization methods during the last decades, renders them invaluable tools, which can provide knowledge about the structure of materials, their chemical composition, as well as the way the react. In the last section of the current review, an extensive

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

overview of the most common monitoring methods is presented. Typical microscopic methods such as optical and scanning electron microscopy as well as analytical and spectroscopic methods like NMR, AFM, FTIR and Raman spectroscopy employed for the evaluation and monitoring of self-healing are presented. The present review covers the literature published during the last five years, with more extensive information and paradigms in each different case. 2. Materials The ability of self-healing materials to regain autonomously or externally assisted, their initial properties is primarily affected by the selection of the healing agents. Inspired by the biological systems, intrinsic, capsule-based and vascular methods, are the main approaches used in order to impart self-healing functionalities to materials or structures. A variety of self-healing agents have been extensively studied to meet the requirements of the new highly demanding applications of smart materials. This section is dedicated to the review of the research on the aforementioned three approaches as well as the materials that have been proposed as healing agents over the last five years. 2.1. Intrinsic self-healing materials In the recent past, polymer science has reached at a point where it is possible to synthesize “smart” polymers that possess the remarkable, bio-inspired ability of regaining their initial properties completely, ideally without external input. These polymers constitute one of the most important categories of self-healing materials, that of the intrinsic or remendable healing polymers. In this case, repair is achieved through the inherent reversibility of bonding in the matrix phase, which acts as a healing agent. Despite the good healing performance that was achieved in the first generation of intrinsic self-healing epoxy systems [1], the incorporation of dicyclopentadiene (DCPD)/Grubbs' catalyst within the matrix e an expensive and unstable in the hostile environment Grubbs' catalyst-limits its applications [2]. Within the aim of the research community is to maximize the healing efficiency and minimize the cost. Therefore, several other material and techniques have been developed in order to satisfy these criteria. Thermally reversible reactions, especially the DielseAlder (DA) reaction, for cross-linking linear polymers have been extensively studied by many researchers. Their main advantage is the theoretically infinite number of repetitions of the healing process without any further addition of chemical or healing agents [3e6]. Hermosilla et al. [7] presented a novel reversible thermoset polymer based on chemical modification of aliphatic polyketones into the corresponding derivatives containing furan and/or amine groups along the backbone. The furan moieties allow for the thermal setting of the polymer by the DielseAlder (DA) and retro-DA sequence (bis-maleimide), while amine moieties allow for the tuning of the hydrogen bonding density. This new class of polymer material showed improved Tg values with respect to the respective counterparts containing only furan groups. Via this modification, these materials recover their mechanical properties after three thermal cycles. In another study, Joost Brancart et al. [8] investigated the ability of furan-maleimide building blocks to create reversible covenant networks in an epoxy based coating. Furanfictionalized precursors were synthesized via reaction of amines with furfuryl glycidyl ether (FGE). The reversible cross-linking of the furan-precursors with a bis-maleimide was achieved in a twostep procedure. Thermal analysis of these composites showed that modification of the polymer network structure allows for the tailoring of the temperature for the self-healing process. Jenifer Ax

93

and Gerhard Wenz [9] created a processible, remendable and highly oriented polymeric material with pending furane substituents, esterifying hydroxyethylcellulose with both furoyl chloride and acetic anhydride. In order to achieve crosslinking (DA reaction), 1,6-bis(N-maleimido)hexane was used. They have shown that both constituents can be mixed without premature formation of gels due to the low rate DA reaction under 70
C. Yoshifumi Amamoto et al. [10] have successfully produced a cross-linked polymer based on reshuffling of thiuram disulfide (TDS) units. Stimulation of the self-healing process occurred under ambient conditions (visible light, air, room temperature) in the absence of a solvent. To carry out the self-healing reaction in a bulk material at room temperature, the reactive TDS units, capable of re-shuffling, were incorporated in the main chain of a low Tg polyurethane. In a more recent work, Claudio Toncelli and co-workers [11] presented the successful synthesis and crosslinking of functionalized (varying amounts of furan groups) polyketones with (methylene-di-p-phenylene)bis-maleimide. In addition, they managed to modify thermal and mechanical properties of the material by controlling the furan reactions. This self-healable polymer exhibited an almost full recovery of thermal and mechanical properties for seven consecutive self-healing cycles, independently of the furan intake. Guadalupe Rivero et al. [12] managed to produce polyurethane networks with healing capability, based on PCL and furanmaleimide chemistry, at mild temperature conditions via one-pot synthesis. A combination of a quick shape memory effect (contact of the free furan and maleimide moieties) followed by a progressive DielseAlder reaction (reformation of the covenant bonds) allows the remendable process to take place at 50
C, resulting in a complete recovery of the structural integrity without complete melting of the polymer. A schematic representation of the DielseAlder based shape memory assisted self-healing process is depicted in Fig. 1. A new approach for the development of self-healing nanocomposites was proposed by Sandra Schafer and Guido Kickelbick [13]. In their study surface-functionalized silica nanoparticles were used as cross-linking agents in thermally triggered self-healing

Fig. 1. Schematic depiction of the DielseAlder based shape memory assisted selfhealing process in a polyurethane material based on PCL and furan-maleimide chemistry [12]. Reprinted with permission from Rivero G, Nguyen L-TT, Hillewaere XKD, Du Prez FE. One-Pot Thermo-Remendable Shape Memory Polyurethanes. Macromolecules. 47(6):2010e8. Copyright (2014) American Chemical Society.

94

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

nanocomposites based on Diels-Alder (DA) chemistry of poly (butyl methacrylates) and structurally varied polysiloxanes. It has been observed that the healing properties of the nanocomposite are highly affected by the molecular structure of the crosslinker (spacer length), the length of the polymeric chain but also by the type of the polymer. DA reaction seemed to be favored by the use of modified polymers (high mobility) and by the presence of particles that have long spacer groups along with lower molar amount of coupling agent. Apart from the DielseAlder reaction, different healing chemistries have been explored to meet the application requirements of self-healing polymers in different occasions. Healing functionality was successfully incorporated in a polyurethane (PU) elastomer by crosslinking the tri-functional homopolymer of hexamethylene diisocyanate (tri-HDI) and polyethylene glycol (PEG) with alkoxyamine-based diol. It has been shown that the design of the polyurethane molecules can be used to optimize not only the mechanical properties but also the healing performance. Moreover, the healing process is completed only by a single step dynamic equilibrium of CeON bonding, unlike to self-healing based on reversible DA bonds, which has to be heated up to a certain temperature for disconnecting the intermonomer linkages and then cooled down for reconnection [14]. Additional efforts have been made towards the production of autonomously self-healing materials using non-covalently bonded systems where the polymerization and/or the crosslinking occur by intermolecular interactions of the monomer units and/or the side chains. Compared to chemically cross-linked hydrogels, supramolecular hydrogels demonstrate better reproducibility of the healing procedure. Takahiro Kakuta et al. [15] used the hosteguest interaction to produce selfhealing materials that can recover their initial strength even after being sectioned in the middle. The aforementioned self-healable supramolecular materials consisted of cyclodextrins (CD) - guest gel crosslinked between poly(acrylamide) chains with inclusion complexes. The obtained CDeguest gels exhibit a self-standing property without chemical crosslinking reagents, indicating that the newly formed hosteguest interactions between the CD and the guest units stabilize the conformation of the CDeguest gels. In another study, [16], a promising non-covalent thermal-switchable self-healing hydrogel was developed by mixing hydrophobically modified chitosan (hm-chitosan) with thermal-responsive vesicle composed of 5-methyl salicylic acid (5 mS) and dodecyltrimethylammonium bromide (DTAB). By altering the temperature, the hydrogel can be switched from sol to gel state (Fig. 2). These transitions can be reversibly performed for several cycles in a similar way to a supramolecular gel. The gelation temperature in particular, can be easily controlled by varying the ratio of DTAB to 5 mS. In a more recent work, Lafont and his team [17], created a multifunctional self-healing composite capable of multiple healing by mixing an uncured thermoset rubber with reversible disulphide bonds, loading it with inert, thermally conductive graphite and hexagonal boron nitride (hBN) as fillers. They proved that higher the healing temperature the better was the cohesion recovery even for highly loaded composites. A very promising concept that combines reversible covalent linkages through imine bond formation with non-covalent interactions through hydrogen bonds between urea-type groups inside the same polymer structure was presented by Nabarun Roy and his team [18]. Through polycondensation reaction between siloxane-based dialdehyde and carbohydrazide they managed to address reversibility to carbinol (hydroxyl) e terminated polydimethylsiloxane (PDMS) via the formation of bisiminourea type subunits. Acylhydrazone units and lateral hydrogen bonding interactions impart to the polymer structure reversible covalent and non-covalent linkages respectively, resulting to a soft

dynamic polymer film capable of autonomous healing. In addition, alterations to mechanical properties of the polymer can be achieved by modifying the length of the siloxane spacer units. In another study, So Young An et al. [19] reported a novel dual sulfideedisulfide crosslinked networks which exhibited a rapid (30 se30 min) and effective self-healing ability at room temperature without external stimuli. The method that has been used for the synthesis of dual-sulfideedisulfide crosslinked network, produced a sufficient density of disulfide crosslinkages which was necessary for the completion of the self-healing process at room temperature. 2.2. Capsule-based self-healing materials An alternative approach to achieve self-repair polymeric materials is the incorporation of capsules within the polymer. Inside these microcapsules lies the healing agent which will be delivered to the damaged area upon rupture of the capsule. The first capsulebased self-healing concept was proposed by White et al. [2]. They embedded microcapsules containing healing agent and catalyst particles into a matrix material achieving a very promising selfhealing efficiency. Since then, microcapsules were extensively studied by many researchers due to their ease of applicability and their potential for mass production. Several epoxy monomers have been easily encapsulated using various methods [20e25]. However, the encapsulation of suitable hardeners remains an issue. Recently, Li Yuan et al. [26] demonstrated the self-healing ability of a cyanate ester (CE) resin by the addition of microcapsules within the volume of the material. The capsules consisted of a poly(ureaformaldehyde) shell filled with an bisphenol A epoxy (EP) as curing agent. Diaminodiphenylsulfone (DDS) catalyst was also used in the CE formulation to decrease the polymerization reaction temperature. Specimens exhibited an 85% self-healing efficiency proving the effectiveness of the microcapsule approach for the development of self-healing polymer materials, as well as for fiberreinforced CE composites. Henghua Jin et al. demonstrated a selfhealing epoxy adhesive suitable for bonding steel substrates using DCPD filled microcapsules and Grubbs' first generation catalyst [27]. It was noteworthy that the addition of both components to the neat resin epoxy (EPON 828) increased the virgin fracture toughness by 26% and a recovery of 56% of fracture toughness was reported. Capsule-based self-healing coatings have been studied by many researchers over the last three years to due to the increased importance of maintaining the potential of protection of the underlying substrate [28e33]. In detail, Xiuxiu Liu et al. [33] prepared a smart self-healing coating consisted of an epoxy resin (diglycidyl ether of bisphenol A) as matrix and microcapsules filled with the same polymer as curing agent. Capsules were synthesized by interfacial polymerization of epoxy droplets with ethylenediamine (EDA). These microcapsules exhibited high shell strength while they could rupture under external force, releasing the healing agent to the damaged area. It should be noted that the complete absence of catalyst along with the high level of healing efficiency, make epoxy-capsule loaded polymers excellent candidates for the development of self-healing films. In another study, Erica Manfredi and co-workers [34] produced glass fiber reinforced polymer (CFRP) containing a solvent (ethyl phenylacetate e EPA) capsulebased healing system using vacuum assisted resin infusion molding technique. Capsules were manually dispersed into the composite and the maximum pressure threshold, in order to avoid premature capsule rupture was 0.3 bar. It must be pointed out that the healing process is based on the swelling mechanism of the polymeric matrix (Epon 828/DETA) in the presence of EPA solvent, and filling the defects that have been created due to static loading

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

95

Fig. 2. Thermal-switching of the vesicle-based gel. Photographs of a sample in aqueous solution: (a) before and (b) after heating. (c) Schematic illustration of the solegel transition [16]. Reprinted from Colloid and Polymer Science, Vol. 291(7), 2013, pp. 1749e58, Thermal-responsive self-healing hydrogel based on hydrophobically modified chitosan and vesicle. Hao X, Liu H, Xie Y, Fang C, Yang H., Figure 2, Original caption: “Thermal-switching of the vesicle-based gel. Photographs of a sample of hm-chitosan (0.4%) and DTAB/5 mS (16 mM/ 20 mM) in aqueous solution: a before and b after heating. Before heating, the sample is strongly viscoelastic and holds its weight in the inverted vial. After heating at 55
C for 10 min, the sample is transformed into a low viscosity fluid that flows easily. c Schematic illustration of the solegel transition.”, Copyright Springer-Verlag Berlin Heidelberg 2013, with kind permission from Springer Science and Business Media.

in Mode I and II. Using a single capsule, resin-solvent self-healing chemistry, Jones and co-workers [35] managed to obtain a full recovery of interfacial shear strength (IFSS) for a glass/epoxy composite. Microcapsules contained EPON 862 (diglycidyl ether of bisphenol-F) dissolved in ethyl phenylacetate (EPA) while the shell material consisted of poly(urea-formaldehyde) e pUF. Moreover, several parameters that can affect the healing efficiency of the system, like the resin-solvent ratio, the capsule coverage and the

Fig. 3. SEM micrographs of glass fibers with varying capsule [35]. Reprinted from Composites Science and Technology, Vol. 79, Jones AR, Blaiszik BJ, White SR, Sottos NR., Full recovery of fiber/matrix interfacial bond strength using a microencapsulated solvent-based healing system. pp. 1e7, Copyright (2013), with permission from Elsevier.

capsules size were also examined. Results indicated that the critical resin-solvent ratio in order to obtain submicron capsules (0.6 lm diameter) was 30:70 in which a total 83% recovery of IFSS was reported. Fig. 3 depicts SEM images of fibers with varying capsule coverage that have been used for the IFSS experiment. Dual-component microcapsules also drew the attention of the research community. The approach lies in fabricating a self-healing epoxy composite by embedding a healing agent consisting of epoxy and its hardener inside separate capsules. He Zhang and co-worker [36] created two types of healing agent carriers, i.e. microcapsules containing epoxy solution (Epolam 5015 and hardener 5015) and etched hollow glass bubbles (HGBs) loaded with amine solution (diethylenetriamine and ethyl phenylacetate) which they incorporated in self-healing epoxy system (Epolam 5015 and hardener 5015). Using TGA, SEM and optical microscopy they managed to characterize both capsules and bubbles. The results indicate that the amine in the etched HGBs shows high thermal stability during the curing stage. A mathematical model has been also formulated in order to calculate the available healants and the diffusion distance on the crack plane of a two-part epoxy-amine. Based on the simple cubic array model, the diffusion distance of the released healing agent was calculated to be inversely proportional to the cubic root of the concentration of the healing agent carrier. In a more recent study [37], Jin and his team focused on the encapsulation of epoxy and amine reactants in separate polymeric microcapsules. In the case of the epoxy resin, a polyurethane (PU)poly(urea-formaldehyde) (PUF) double shell wall was used. The core consisted of disphenol-A epoxy resin diluted with a low viscosity reactive diluent (o-cresyl glycidyl ether). As for the amine capsules, they were produced following a method of vacuum infiltration of polyoxypropylenetriamine (POPTA) into polymeric hollow (PUF walled) microcapsules, demonstrating thus a simple approach for the encapsulation of a highly reactive core material. Both epoxy and amine microcapsules can be seen in Fig. 4.

96

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 4. (a) Epoxy capsules consist of a polyurethane e poly(UF) double shell wall and a DGEBA/o-CGE core. (b) Amine capsules contain a poly(UF) shell wall and a POPTA core [37]. Reprinted from Jin H, Mangun CL, Griffin AS, Moore JS, Sottos NR, White SR. Thermally stable autonomic healing in epoxy using a dual-microcapsule system. Advanced Materials. 26(2):282e7, Copyright (2013) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Afterward, the capsules were embedded into an epoxy matrix system (Araldite/Aradur 8615) while taking into account the required stoichiometry. Maintaining the total capsule concentration at 10 wt% while varying the ratio of epoxy to amine capsules, they managed to obtain the highest average healing efficiency which was at an equal mass ratio of amine:epoxy capsules (5:5). It was demonstrated, higher exposure temperature caused more loss of core contents for both types of capsules leading to a poor mixing of the reactants in the damaged area. Apart from the well-known poly(urea-formaldehyde)-shell microcapsules, a generalized silica coating scheme was developed by Jackson et al. in order to functionalize and protect sub-micron and micron size dicyclopentadiene monomer-filled capsules and Grubbs' catalyst particles [38]. Fluoride-catalyzed silica condensation chemistry was used for the construction of the protective and functional silica coatings resulting to an improvement of the dispersion of the capsules and catalyst particles inside the epoxy matrix. Unlike many other studies, a successful incorporation of both capsules into the epoxy was achieved without significant loss of healing agent. In Fig. 5, a TEM image of a silica coated DCPD-filled capsule is presented. In an effort to improve the self-healing efficiency of epoxy resin, Qi Li and his co-workers [39] prepared a dual-component microcapsule of diglycidyl ether of bisphenol A epoxy (DGEBA) (resin) and polyether amine (hardener) using a water-in-oil-inwater emulsion solvent evaporation technique with polymethyl methacrylate (PMMA) as shell material. They have shown that the healing efficiency of epoxy was affected by the content and ratio of the dual-component microcapsules. Self-healing was carried out successfully at room temperature, but as was indicated, increase in temperature led to higher levels of the selfhealing efficiency.

A very interesting concept developed by Dong Yu Zhu and his team [40] constitutes the construction and development of a multilayered microcapsules used for self-healing thermoplastics (Fig. 6). By optimizing the synthesis conditions, robust poly(melamine-formaldehyde) (PMF)-walled microcapsules containing fluidic glycidyl methacrylate (GMA) monomer with proper size and core content were produced. Second and third (outer/ protective) layers consisted of living poly(methyl methacrylate) (PMMA-Br) and wax respectively. Results concerning the performance and stability (thermal and chemical) indicate that the multilayered microcapsules might be applicable for manufacturing not only self-healing thermoplastics but also self-healing thermosets. 2.3. Vascular self-healing materials Similar to blood vessels in biology, vascular self-healing systems incorporate healing agents into a polymer matrix through microchannels. The original idea as proposed by Toohey et al. [41], concerned the incorporation of a microchannel network containing dicyclopentadiene (DCPD) in the material. Microchannels delivered DCPD to an epoxy surface coating containing Grubbs' catalyst. Over the years vascular self-healing materials were extensively studied due to the variety of healing that can be used and the large scale of damage that can be healed [42e47]. Patrick et al. [48] demonstrated that in situ self-healing can be achieved in structural fibercomposites via microvascular delivery of isolated, reactive healing reactants. Diglycidyl ether of bisphenol A (DGEBA) based epoxy resin (EPON 8132) and aliphatic triethylenetetramine (TETA) based hardener (EPIKURE 3046) were used as healants due to their reaction kinetics and their post-polymerized mechanical properties. In order to create the microvascular network, pre-vascularized

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

97

Fig. 5. A representative TEM image of a microtomed cross-section of a silica coated DCPD-filled capsule. The DCPD core is removed during the microtoming process [38]. Reprinted from Aaron C. Jackson, Jonathan A. Bartelt, Kamil Marczewski, Nancy R. Sottos, Paul V. Braun. Silica Protected Micron and Sub-Micron Capsules and Particles for SelfHealing at the Microscale. Macromolecular Rapid Communications. Copyright (2010), Wiley Periodicals. Inc.

composite textile reinforcement was produced by stitching catalyst-infusion, in a precise pattern of aerospace-grade woven fabric. The fiber composite preform was then consolidated into a structural laminate via vacuum assisted resin transfer molding (VARTM) of a thermoset epoxy matrix. It is noteworthy that after the final thermal PLA evacuation step (three-dimensional (3D) microvasculature), no significant change to fracture properties was observed. They have also shown that vascular architectures not only provide efficient and repetitive delivery of healing agents, but they also contribute to increased resistance to delamination initiation and propagation. The self-healing cycle of the aforementioned microvascular system is depicted in Fig. 7. The effect of microvascular channels on the in-plane tensile properties and damage propagation in a 3D orthogonally woven/ glass epoxy has been successfully described by Coppola et al. [49]. Using Vaporization of Sacrificial Components (VoSC) process they managed to produce composites consisted of two part epoxy matrix (EPON 862 epoxy/EPIKURE W curing agent) with straight and wave shaped channels (Fig. 8). Sacrificial fibers (SF) were prepared using poly(lactic acid) (PLA) monofilament fibers treated with

tin(II) oxalate (SnOx) catalyst so as to decrease their thermal degradation temperature. SF removed during the post-curing process leading to an insignificant alteration on the tensile properties, strength and modulus of the composite material. In another work, A. R. Hamilton et al. [95] reported the use of active pumps that can deliver a two-part healing system (Epon 8132/Epikure 3046) inside a material through microvascular networks. This technique allows a small vascular system to deliver large volumes of healing agent to the damaged area. Moreover, dynamic pumping leads to an enhancement of component's mixing in the target region, improving with that way the self-healing efficiency. The construction of self-healing materials with embedded ternary interpenetrating microvascular networks by direct-write assembly of fugitive inks has been reported by Hansen and his team [50]. The matrix of the material consisted of a two part epoxy system diglycidyl ether of bisphenol-A resin (EPON 8132) and an aliphatic amidoamine (Epikure 3046) as hardener. It was noteworthy that they managed to accelerate the recovery of mechanical properties of the resin by exploring the effect of temperature on the healing reaction kinetics of the healing agent. They report a

Fig. 6. Profile of multilayered microcapsule [40]. Reprinted from Polymer, Vol. 54 (16), Zhu DY, Rong MZ, Zhang MQ. Preparation and characterization of multilayered microcapsule-like microreactor for self-healing polymers. pp. 4227e36, Copyright (2013), with permission from Elsevier.

98

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 7. Life-cycle of a self-healing microvascular fiber-composite. Pristine woven composite laminate showing stacked textile reinforcement with dual-channel (red/blue), liquid filled vascular network [48]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Reprinted from Patrick JF, Hart KR, Krull BP, Diesendruck CE, Moore JS, White SR, et al. Continuous self-healing life cycle in vascularized structural composites. Advanced materials. 26(25):4302e8, Copyright (2014) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Fig. 8. (aec) Schematics of the unit cell of the preforms. Optical micrographs show surfaces (def) normal to the warp direction and surfaces (gei) normal to the weft direction. Scale bars represent 1 mm [49]. Reprinted from Composites Part A: Applied Science and Manufacturing, Vol. 59, Coppola AM, Thakre PR, Sottos NR, White SR. Tensile properties and damage evolution in vascular 3D woven glass/epoxy composites. pp. 9e17, Copyright (2014), with permission from Elsevier.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

reduction in healing times by over an order of magnitude. In attempting to improve the performance of self-healing materials, Richard S. Trask and co-workers [51] presented the construction of a combined sensing and healing vascular network within an advanced fiber-reinforced composite. Poly(tetrafluoroethylene) (PTFE)-coated steel wires used in order to create the microvascular network inside a fiber reinforced polymer composite. A lowpressure sensor was directly connected within the perceived damage zone, while the output signal of the sensor was monitored via open-source microprocessors. The laminates were subjected to a 10-J energy impact and the healing agent was delivered through a pump from an external reservoir. Two different healing chemistries were tested, a commercial system ResinTech RT151 and the wellknown epoxy based system of diglycidyl ether of bisphenol-A (DGEBA), ethyl phenylacetate (EPA) and diethylenetriamine (DETA) resulting to a recovery of 91% and 94% in post-impact compression strength respectively. The use of Hollow Glass Fibers (HGFs) filled with a single component epoxy resin (Envirez 70301) in e-glass/epoxy composites has been reported by S. Zainuddin et al. [52]. The matrix consisted of a two part epoxy system. Part-A was a blend of diglycidyl ether of bisphenol-A (DGEBA), aliphatic deglycidylether and epoxy terminated polyether polyol. The curing agent (Part-B) was a mixture of 70e90% cycloaliphatic amine and 10e30% polyoxylalkylamine. A commercially available woven fabric oriented in two directions (warp at 0
and fill at 90
) was used as reinforcement. Fig. 9 depicts the fabrication of composite embedded with HGFs. Using this methodology the managed to achieve significant regaining of the mechanical properties after multiple Low Velocity Impact (LVI). Koralagundi Matt et al. [53] produced a vascular network within a conventional glass fiber reinforced polymer matrix composites (PMC) in order to address self-healing capability to a composite

99

structure that can be used in wind turbine blades. Via dynamic mechanical analysis (DMA), they proved that the vascular network does not affect the dynamic mechanical properties of the final composite. Moreover, it has been shown that the most effective way to produce vascular self-healing structures is to arrange the tube network parallel to the resin flow direction during the vacuum infusion process. 3. Material properties as a means to self-healing evaluation The challenge in the design of self-healing materials is to create a new composite material with an autonomous or externally stimulated damage healing capability in order to extend the performance life time of the newly developed material or product. The presence of local regions in the material with lower or degraded performance than that of the surrounding areas can be defined as damage [38]. Thermal or electrochemical degradation can also be included under this definition. The incorporation of self-healing agents (SHA) in a material such as a typical polymer matrix would certainly alter its properties. Hence it is crucial to monitor those changes in order to assess the performance of the new composite material. These changes can also be used as a mean to characterize qualitatively or quantitatively the healing performance. In the most favorable scenario, the new material properties will be equal or better to that of the unmodified one. The researcher or engineer should study the modified self-healing material as compared to the unmodified, virgin material in order to assess its performance. The following sections are dedicated to the presentation of several methods, techniques, and specimen geometries to describe, either qualitatively or quantitatively, the unmodified and modified self-healing material properties and damage focusing on those that can be used to characterize the healing performance.

Fig. 9. Filling of HGFs and fabrication of e-glass/epoxy composite [52]. Reprinted from Composite Structures, Vol. 108, Zainuddin S, Arefin T, Fahim A, Hosur MV, Tyson JD, Kumar A, et al. Recovery and improvement in low-velocity impact properties of e-glass/epoxy composites through novel self-healing technique. pp. 277e86, Copyright (2014), with permission from Elsevier.

100

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

3.1. Self-healing modified and neat material's physical properties Prior to the evaluation of the self-healing properties and performance of the modified material, comparative tests are typically employed in order to assess the modified and unmodified material properties. Most of the self-healing systems that are reported in the literature consist of a polymeric matrix and the self-healing agents. These consist of microcapsules, vascular networks or other polymers in the form of additives. The most frequently used characterization techniques for these materials are Dynamic Mechanical Analysis (DMA), Differential Scanning Calorimetry (DSC) and Thermogravimetric Analysis (TGA). DMA is a widely used technique for materials characterization and it is mostly used to determine the glass transition temperature of the constituent materials [44], the viscoelastic properties in terms of storage and loss moduli or shear storage modulus [54] and measure the coefficient of thermal expansion (CTE) [55]. DSC is also used to measure the glass transition temperature for the matrix and the self-healing agents (SHA) [56], monitor the selfhealing process [39] and the curing process. TGA is commonly used to determine selected characteristics of materials that exhibit either mass gain or loss. In self-healing studies, TGA is employed for the SHA thermal stability evaluation [57] and the evaluation of the amount of encapsulated SHA. It is also useful for decomposition and maximum weight loss temperature measurements [58]. Other techniques that are employed in self-healing materials characterization include Fourier Transform Infrared Spectroscopy (FTIR) [58], Nuclear Magnetic Resonance (NMR) [59,60] and RAMAN spectroscopy [61]. These techniques are widely used for monitoring the self-healing process and will be further discussed in later sections. 3.2. Mechanical properties The following sections are dedicated to the presentation of mechanical performance evaluation and the respective techniques employed. It should be noted that apart from the information that these properties and techniques provide regarding the healing efficiency, they can offer quantitative and qualitative means for comparing the modified and unmodified materials. It is also common that the mentioned techniques are complimented with other qualitative techniques like scanning electron, transmission electron, acoustical, and/or optical microscopy. 3.2.1. Static damage Static damage can occur in structural materials in the form of cracks anywhere in the 3D structure and depending on the application, loading conditions and type of damage, can occur over the span of multiple length scales. For instance cracks in a fiber composite structure can initiate on the fiberematrix interface, propagate to the matrix phase and result in the failure of the structure through fiber eruption, pull out etc. The mechanical properties, damage initiation and propagation and healing performance have been extensively studied under universal testing machines under various loading conditions and scenarios. Tensile testing is one of those loading conditions and has been used extensively for measuring stressestrain relations, ultimate tensile strength and Young modulus [39]. Specimen geometries for such tests include rectangular shaped specimens (Fig. 10) and dog bone specimens [62,63] where ASTM D 3039 [64] and D 638 [65] are the respective standards that describe these two geometries along with all the variables included in the testing procedures, while a representative graph obtained from such tests can be seen in Fig. 11. Coppola et al. [49], refers to

Fig. 10. Schematic of the tensile test specimen [49]. Reprinted from Composites Part A: Applied Science and Manufacturing, Vol. 59, Coppola AM, Thakre PR, Sottos NR, White SR. Tensile properties and damage evolution in vascular 3D woven glass/epoxy composites. pp. 9e17, Copyright (2014), with permission from Elsevier.

both of these standards in order to study the effect of vascular channels on the in-plane tensile properties and damage progression of three-dimensional woven textile composites. Composites specimens were prepared and tested according to ASTM D3039 while epoxy dog bone shaped specimens were prepared and tested according to ASTM D 638. A similar standard is the ASTM D 1078 [66] which describes the microtensile dog bone specimen geometry but has the limitation that it cannot provide data for the determination of modulus of elasticity (Fig. 12) [67].

Fig. 11. Representative stressestrain curves of virgin and repeatedly repaired IDHPEG800-0.5 specimen [63]. Reprinted from Polymer, Vol. 53(13), Ling J, Rong MZ, Zhang MQ. Photo-stimulated self-healing polyurethane containing dihydroxyl coumarin derivatives. pp. 2691e8, Copyright (2012), with permission from Elsevier.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Ultimate tensile strength can be used in order to compare the neat material properties with the modified ones [68] or to provide a metric of the healing performance as described in the works of Yuan et al. [14]. In their work, self-healing is characterized by quasistatic tensile tests performed on rectangle shaped specimens. The specimens are tested till fracture, then put together to heal and retested. The ratio of tensile strength of the healed specimen to that of the virgin one provides the healing efficiency. The improper alignment of the fractured surfaces and surface roughness effects are challenges for the tensile tests because they can lead to porosity of the healed specimens. In addition, tensile tests are inherently designed to characterize bulk continuous deformation leading up to failure and the tensile stress and strain values might be misleading as the material necks prior to failure. Elongation at break, yield point stress [69] and force displacement curves [70] have also been used to quantify and characterize the self-healing performance but due to the aforementioned challenges, these metrics do not fully reflect the healing quality. Tensile loading conditions can be used to study the adhesion behavior of adhesives or composites via lap shear tests. In such a configuration, a thin slice of a self-healing adhesive is sandwiched between two plates and the sample is tested under tension till the lap joint fails and the maximum shear strength can be used as a self-healing efficiency in multiple cycles (Fig. 13) [71e73]. Another approach is according to ASTM D 897 [74] where shear testing is performed under compressive loading conditions. ASTM D 3846 [75] refers to reinforced plastics and is concerned with the determination of in plane shear strength of reinforced thermosetting plastics in flat sheet form in thicknesses ranging from 2.54 to 6.60 mm and is adopted by Hondred et al. [76] in order to examine the adhesive properties of thermosetting polymers modified with rare earth triflates. A major advantage of the lap shear tests is that fractured surfaces can be brought into contact in a more controlled manner compared to tensile tests and the alignment and clamping conditions are easily reproducible and less sensitive to topological effects. In addition for self-healing systems with reversible chemistries, the lap shear tests can be designed to study experimental parameters of controlled force, curing temperature, and multiple healing cycles. However the distinction between adhesive and cohesive failure is of paramount importance.

Fig. 12. Microtensile test specimens [67]. Reprinted from Grande AM, Castelnovo L, Landro LD, Giacomuzzo C, Francesconi A, Rahman MA. Rate-Dependent Self-Healing Behavior of an Ethylene-co-Methacrylic Acid Ionomer Under High-Energy Impact Conditions. Journal of Applied Polymer Science. Copyright (2013), Wiley Periodicals. Inc.

101

The cohesion recovery is related to the ability of the material to exhibit temperature-activated mending within its volume. To investigate the cohesive healing ability, Lafont et al. [17] cut their self-healing samples into four pieces using a sharp razor blade. The pieces of material were put back together until visual contact and were placed between two glass slides. The initial cut width and area was recorded under an optical microscope. During the healing procedure, the samples were visually inspected at various intervals in order to measure the evolution of the cut width and area for the healing efficiency quantification. Bending and compressive loading can also be used to provide insight on the healing efficiency and the structural integrity restoration recovery. Wu et al. [77] used for the healing performance evaluation of a self-healing carbon fiber/epoxy composite system (Fig. 14), the stiffness recovery ratio (SRR%) defined as:

SRR ð%Þ ¼

healed flexural stiffness ; Initial flexural stiffness

The SSR was derived from rectangular shaped specimens under 3 point bending testing according to the ASTM D 790 [78] standard. In addition load-deflection curves are used in order to evaluate the core shell nano-fibers effect on the mechanical properties of the laminates. Li et al. [79] adopt the 3 point bending fixture with single edge notched beam (SENB) specimens, that are described in the ASTM D 5045 [80] standard, to evaluate the healing behavior of a modified DGEBA epoxy resin. Healing is presented as a function of load recovery for the modified and unmodified resins. The SENB specimen geometry is also used by Meure et al. [81]. The healing efficiency is determined by comparing fracture toughness KIC values of fractured specimens after healing with those of pre-fractured specimens. The primary advantage of using the fracture toughness values for healing performance evaluation is that it yields a quantitative measure of healing efficiency that is tied to the recovery of an inherent material property. However, the calculations of KIC with the SENB geometry require accurate knowledge of the initial crack length and the crack length after healing. The tapered double cantilever beam (TDCB) specimen geometry can be used in order to overcome the crack length measurement difficulty mentioned for the SENB specimen geometry [24]. The primary feature of this geometry is that it exhibits a linear relationship between critical load P and fracture toughness KIC independent of crack length (Fig. 15). In addition, the short groove of the geometry requires small amounts of self-healing material and the crack initiates and propagates in a more controlled manner. In addition, if it is required, the test can be stopped in a desired crack length. A detailed comparison between the SENB and TDCB geometries is presented in the works of Brown [82]. The TDCB specimens are widely used in self-healing applications for healing performance evaluation either by comparing KIC values [59] or critical peak loads values [83,84]. The TDCB specimens can also be used in fatigue loading scenarios like in the work of Neuser and Michaud [85]. In their work, epoxy TDCB specimens with microcapsules and shape memory alloy wires were subjected to tensionetension fatigue and tension testing. Fatigue testing was conducted in order to compare the behavior of pure epoxy, epoxy with microcapsules, and epoxy with SMA wires while tension testing provided the virgin and healed peak load data for the healing efficiency evaluation. However, a major disadvantage of the TDCB geometry is that the fracture behavior of the specimens, show an important dispersion and unstable fracture behavior that must be taken into account to obtain accurate results [86].

102

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 13. Adhesion recovery as function of the healing temperature and cross-linker type at 65
C using (a) 4-SH or (b) 3-SH [73]. Reprinted from Lafont U, van Zeijl H, van der Zwaag S. Influence of cross-linkers on the cohesive and adhesive self-healing ability of polysulfide-based thermosets. ACS applied materials & interfaces. 4(11):6280e8. Copyright (2012) American Chemical Society.

Fig. 14. (a) Three-point bending test set up, (b) three-point bending specimens [77]. Reprinted from Wu X-F, Rahman A, Zhou Z, Pelot DD, Sinha-Ray S, Chen B, et al. Electrospinning Core-Shell Nanofibers for Interfacial Toughening and Self-Healing of Carbon-Fiber/ Epoxy Composites. Journal of Applied Polymer Science. Copyright (2012), Wiley Periodicals. Inc.

An alternative geometry to the TDCB geometry is the widthtapered double cantilever beam (WTDCB) (Fig. 16). The WTDCB provides a crack length independent measurement of mode I fracture toughness like the TDCB geometry. Jin et al. [27] uses the

WTDCB geometry under quasi static fracture and fatigue testing. Specimens of steel adherents bonded with self-healing epoxy adhesive were prepared and tested on a universal testing machine under quasi static loading and cyclic loading conditions. The

Fig. 15. TDCB specimen geometry and dimensions in mm [86].
mez D, Gilabert FA, Tsangouri E, Van Hemelrijck D, Hillewaere XKD, Du Prez FE, et al. In-depth Reprinted from International Journal of Solids and Structures, Vol. 64e65, Garoz Go numerical analysis of the TDCB specimen for characterization of self-healing polymers. pp. 145e54, Copyright (2015), with permission from Elsevier.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

103

providing information for mode II strain release energy rate GII calculations. The double cleavage drilled compression (DCDC) specimen or open hole specimen (OHS) under compressive load is selected from Hamilton et al. [95]. They study vascular epoxy specimens in order to evaluate the healing performance of pumping protocols as manifested by the recovery of fracture toughness after each healing cycle. The DCDC geometry is more appropriate for studying the fracture toughness of brittle materials like the epoxy matrices and their fiber reinforced composites. Under a uniform axial compression load, the Poison effect produces a tensile stress concentrated around the central hole which induces the initiation of two symmetric mode I cracks at each crown of the hole propagating along the mid-plane of the sample.

Fig. 16. (a) Geometry of WTDCB specimen consisting of adhesively bonded A36 steel adherents. (b) Optical microscopy of cross section of a self-healing adhesive incorporated with Grubbs' catalyst and DCPD microcapsules [27]. Reprinted from Polymer, Vol. 52(7), Jin H, Miller GM, Sottos NR, White SR. Fracture and fatigue response of a self-healing epoxy adhesive. pp. 1628e34, Copyright (2011), with permission from Elsevier.

healing efficiency is assessed by the ratio of the healed fracture toughness to the virgin fracture toughness while the fatigue performance of the self-healing adhesive was investigated under cyclic loading. Another specimen geometry that can be used for fracture toughness calculations in the form of opening mode I interlaminar fracture toughness GIC [87,88], is the double cantilever beam (DCB) [89] as described in ASTM 5528 [90] standard for fiber reinforced composite materials and depicted in Fig. 17. It is widely used in selfhealing vascularized materials research because it has the advantage that the test can be stopped at any prescribed crack length [91]. However, mode I critical strain energy release rate calculations require that applied load, crack opening displacement and crack length values are recorded [92]. The accurate crack length measurement is a disadvantage like in the SENB geometry. DCB specimens can be tested under different testing geometries and loading scenarios like mode II end loaded split (ELS) [43] as seen in Fig. 18 or end notched flexure under three point bending (Fig. 19) [34,93,94]

3.2.2. Impact damage Impact damage is more difficult to describe compared to static and fatigue damage. The impact damage is a dynamic response of the impacted material, the impacting material, as well as the supporting jig. Impact events in fiber reinforced composites can cause significant reduction in mechanical performance whilst leaving little visual evidence of the impact event. Self-healing of impact damage in fiber-reinforced composites is regarded as one of the most difficult areas of ongoing research because of the large damage volume and multiple failure modes. In such cases the testing procedures involve secondary testing for determining the mechanical properties after the impact event like compression after impact (CAI) testing [96] and/or non-destructive techniques, like ultrasonic C-scan, in order to assess the damage and the healing process as well. The impact event can be generated with different testing apparatus like impact drop tower devices on e-glass/epoxy samples (Fig. 21) [97] or glass fiber composites with microcapsules [98], falling weight impact test machine [99] or ballistic pendulum setup on ionomeric polymers [100] covering a wide range of impact energies and projectile velocities [67,98]. These types of tests can provide data for peak load, energy to maximum load and absorbed energy which can be used for the healing performance evaluation (Fig. 20). The impact testing apparatus and procedure are described in standards like ASTM D 7136 [101] while CAI testing can be found in ASTM D 7137 [102]. Norris et al. [96,79] employ the previous mentioned standards (D 7136, D 7137), in order to investigate the impact behavior of a vasculature design in a fiber reinforced composites and the post healing compressive strength of the proposed system is used for the healing performance evaluation. Ultrasonic C-scanning provided information on the delamination location and helped to assess the different proposed vascular networks. Haase et al. [100] investigates the behavior of a self-healing ionomer under dynamic puncture testing. An impactor with a similar shape to the ballistic impact tests projectiles is pushed through a self-healing polymer sheet at a constant speed. The main focus of this research is the temperature rise caused by the impactor which was recorded by three thermocouples embedded in the polymer sheet. 3.3. Corrosion resistance and protection

Fig. 17. Schematic showing the interplay locations of the EMAA fibers and the delamination fracture plane in the carbon fibereepoxy laminates [92]. Reprinted from Composites Part A: Applied Science and Manufacturing, Vol. 43(8), Pingkarawat K, Wang CH, Varley RJ, Mouritz AP. Self-healing of delamination cracks in mendable epoxy matrix laminates using poly[ethylene-co-(methacrylic acid)] thermoplastic. pp. 1301e7, Copyright (2012), with permission from Elsevier.

Most metals in natural environments exist in their oxidized form, which means that metals tend to corrode, leading to loss of mechanical and esthetic properties. The easiest way to protect metals from undesired corrosion is to apply protective coatings that offer active protection, passive protection, or both. The failure of the protective layer leads unavoidably to corrosion of the underlying metal. The self-restoration of this protective coating is a typical self-

104

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 18. Mode II ELS specimen geometry [43]. Reprinted from Composites Science and Technology, Vol. 71(6), Norris CJ, Bond IP, Trask RS. Interactions between propagating cracks and bioinspired self-healing vascules embedded in glass fiber reinforced composites. pp. 847e53, Copyright (2011), with permission from Elsevier.

healing functionality. In order to explore if a system shows selfhealing functionalities the creation of an artificial defect in the coating system, and evaluation of the ability of the system to suppress or decrease corrosion to desired levels and restore the protective functionality has been broadly used. A detailed review regarding the employed self-healing corrosion protection methodologies can be found in Ref. [103]. One of the most used techniques in corrosion science to monitor local corrosion damage is the Electrochemical Impedance Spectroscopy (EIS) via bod plots (Fig. 22) [104e106]. In order to access the charge transfer resistance or polarization resistance that is proportional to the corrosion rate at the monitored interface, EIS results have to be interpreted with the help of a model of the interface. An important advantage of EIS over other laboratory techniques is the possibility of using very small amplitude signals without significantly disturbing the properties being measured. To make an EIS measurement, a small amplitude signal is applied to a specimen over a range of frequencies. The EIS instrument records the real (resistance) and imaginary (capacitance) components of the impedance response of the system. Depending upon the shape of the EIS spectrum, a circuit model or circuit description code and initial circuit parameters are assumed and input by the operator. EIS can provide quantitative information about the electrochemical state of a coating. The EIS set up is used by Garcia et al. [29] in order to assess the healing performance of their proposed self-healing anticorrosive organic coating. Scanning Vibrating Electrode Technique (SVET) is also employed in order to provide further verification on the EIS results. Corrosion activity maps can be obtained by using the SVE Technique [107]. SVET uses a single wire to measure the voltage drop in a solution. This voltage drop is a result of local current at the surface of a sample. Measuring this voltage in the solution,

Fig. 19. End notched flexure specimen geometry as adopted by Ref. [94]. Reprinted from International Journal of Solids and Structures, Vol. 46(13), Ouyang Z, Li G. Nonlinear interface shear fracture of end notched flexure specimens. pp. 2659e68, Copyright (2009), with permission from Elsevier.

the current at the sample surface is mapped. Current can be naturally occurring from a corrosion or biological process, or the current can be externally controlled using a galvanostat. A key application of SVET is to study corrosion process of bare or coated metals. Hollamby et al. [30] employ SVET in order to evaluate the anticorrosive and self-healing behavior of their proposed hybrid polyester coating. Control and coated specimens were scratched and immersed in NaClaq solution and the current density maps from SVET measurements were charted as seen in Fig. 23. Vimalanandan et al. [108] employed the Scanning-Kelvin-Probe (SKP) technique in order to investigate the self-healing performance and the corrosion-driven catholic delamination progress of a conductive polymer (CP) based nano-capsule system. SKP is a scanning probe method where the potential offset between a probe tip and a surface can be measured using the same principle as a macroscopic Kelvin probe (Fig. 24). For the SKP measurements, a scratch was introduced to the CP-coating, this defect was then covered with KCl and introduced to the SKP chamber. The behavior of the corrosion potential and the progress of the delamination were studied. The corrosion potential in the electrolyte defect was monitored by positioning the SKP tip close to the electrolyte drop serving as a reference electrode. The cathodic delamination progress was monitored by scanning the coatings from the defects to reflect the potential

Fig. 20. Energy vs. time curves for microcapsules contained glass fiber reinforced composites [98]. Reprinted from Chowdhury RA, Hosur MV, Nuruddin M, Tcherbi-Narteh A, Kumar A, Boddu V, et al. Self-healing epoxy composites: preparation, characterization and healing performance. Journal of Materials Research and Technology. 2015;4(1):33e43. Copyright (2015), with permission from Elsevier.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

105

state and electro-catalyst materials. The long term anticorrosive efficiency of a damaged epoxy coating containing silyl-ester microcapsules on an aluminum substrate is studied via SECM testing in the works of Gonzalez-Garcia et al. [28]. Combining redox and feedback modes, the long term healing of the coating was demonstrated (Fig. 25).

3.4. Electrical conductivity

Fig. 21. Target mounting in the impact chamber [97]. Reprinted from Advances in Space Research, Vol. 51(5), Francesconi A, Giacomuzzo C, Grande AM, Mudric T, Zaccariotto M, Etemadi E, et al. Comparison of self-healing ionomer to aluminum-alloy bumpers for protecting spacecraft equipment from space debris impacts. pp. 930e40, Copyright (2013), with permission from Elsevier.

distribution as a function of the distance from the defect and time. Scanning Electrochemical Microscopy (SECM), can provide information about the redox activity, redox mode and topography, feedback mode of liquid/gas, liquid/solid and liquid/liquid interfaces. SECM measurements can be used in order to yield topographic information and to probe the surface reactivity of solid-

The self-healing concept has also been implemented and achieved for materials with electrical functionality. Such materials are able to recover conduction paths at different scales and most investigations of conductivity recovery in the literature deal with the healing of such conductive paths. A qualitative way to monitor the recovery of conductivity in self-healing materials can be found in the works of Palleau et al. [109]. Here a simple electronic circuit consisting of a LED, a voltage source and a self-healing stretchable (SHS) wire in series is monitored and captured in video (Fig. 26). The SHS wires are a combination of a self-healing polymer structured with microchannels filled with EGaIn. Scissors are used to cut the wire so that the circuit continuity is lost. When the wires are aligned, the liquid metal components merge together forming a continuous and conductive wire. For studies where a very small change in resistivity is to be monitored, a Wheatstone bridge set-up is preferred. This technique measures an unknown resistance by using an electrical circuit. The

Fig. 22. (a, b) Bode plot and phase angle, (c) of specimens coated with nanocapsules loaded with various types of corrosion inhibitors [106]. Reprinted from Progress in Organic Coatings, Vol. 76(10), Choi H, Kim KY, Park JM. Encapsulation of aliphatic amines into nanoparticles for self-healing corrosion protection of steel sheets. pp. 1316e24, Copyright (2013), with permission from Elsevier.

106

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 23. (A) SVET current density map and (inset) visual appearance of the scratched control sample. (B) SVET current density map and (inset) appearance of the scratched NPs_BTAa sample [30]. Reprinted from Hollamby MJ, Fix D, Donch I, Borisova D, Mohwald H, Shchukin D. Hybrid polyester coating incorporating functionalized mesoporous carriers for the holistic protection of steel surfaces. Advanced materials. 23(11):1361e5, Copyright (2011) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

set up consists of an electrical source with a known voltage, three resistors with known values, and a galvanometer. In self-healing studies, the Wheatstone bridge set up allows monitoring of both the creation and the healing of damage since when the damage occurs the system's resistivity will increase [110]. Blaiszik et al. [111] employ the Wheatstone bridge set up in order to in situ monitor a four point bending test conducted on specimens of microencapsulated metal dispersed in a dielectric material. The specimen acts as one resistor on the Wheatstone bridge circuit. The circuit is monitored throughout the four-point

bend test using a Wheatstone bridge with the specimen as one bridge arm. The performance of the circuit is evaluated by measuring the normalized bridge voltage:

Vnorm ¼ ðVh  V∞ Þ=ðVo  V∞ Þ; where Vo is the bridge voltage before damage, V∞ is the bridge voltage measured for a fully broken circuit, and Vh is the instantaneous bridge voltage of the circuit. The value of Vnorm ranges from zero for a specimen with no electrical conductance to one for

Fig. 24. (A) Scheme depicting the model-coating system and the experimental set-up used to evaluate the self-healing performance of the coating system. (B) Corrosion potential monitored by SKP in the defect. (C) Delamination profiles recorded by SKP [108]. Reprinted from Vimalanandan A, Lv LP, Tran TH, Landfester K, Crespy D, Rohwerder M. Redox-responsive self-healing for corrosion protection. Advanced materials. 25(48):6980e4, Copyright (2013) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

107

Fig. 25. (a) Optical micrograph of AA2024-T3 sample with bare and silyl-treated surface. (b) SECM image of the transition area on (a) using the electroreduction of oxygen. (c) Approaching-curves performed on the bare metal (black line) and on the silyl-covered area (red line). (d) Overlapped approaching-curves corresponding to measurements using electrochemical mediator (red-dashed line) and oxygen reduction (black-solid line) [28]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
lez-García Y, García SJ, Hughes AE, Mol JMC. A combined redox-competition and negative-feedback SECM Reprinted from Electrochemistry Communications, Vol. 13(10), Gonza study of self-healing anticorrosive coatings. pp. 1094e7, Copyright (2011), with permission from Elsevier.

a fully conductive specimen. The efficiency of conductivity restoration, hc, is defined for each specimen as Vnorm after fracture (Fig. 27). Another interesting approach where mechanical and electrical properties are simultaneously recorded and analyzed can be found in the works of Bailey et al. [112]. Here apart from the EIS technique, the authors employ an in-situ electro-tensile technique in order to assess the degree of mechanical and electrical self-healing efficiency of a composite coating. This technique involves the controlled introduction of a crack by pulling the coating in tension while measuring the changes in electrical conductivity on-line. Complementing the EIS results, it was demonstrated that when microcapsules possessing an EPA:ECNT (ethyl phenylacetate: epoxy with carbon nano-tubes) core were incorporated into the coating, electrical conductivity and mechanical properties were restored to 64% (±23) and 81% (±39) respectively (Fig. 28). Furthermore, sequential cracking and healing events were noticed while the coating was pulled in

tension and both EIS and in situ tensile loading and electrical conductivity test revealed a 24 h restoration of this coating analogous to pure ECNT. 4. Characterization of self-healing systems and monitoring of their healing efficiency Next to the technical challenge of realizing a self-healing system, there is an inevitable need both for characterizing the functional components that constitute it and monitor the whole process of self-healing. A variety of characterization techniques can be found in the literature. However, the methods for monitoring the self-healing process are limited. More specifically, in the area of characterization the most common techniques are Fourier Transform Infrared Spectroscopy (FTIR), Nuclear Magnetic Resonance Spectroscopy, Optical and Scanning Electron Microscopy (OM and SEM), Transmission electron microscopy (TEM), Atomic Force Microscopy (AFM), X-ray

108

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 26. a) Schematics illustrating the disconnection and reconnection of a simple electronic circuit using a self-healing wire. b) Variation of the resistance of SHS wires during connection/disconnection/reconnection experiments [109]. Reprinted from Palleau E, Reece S, Desai SC, Smith ME, Dickey MD. Self-healing stretchable wires for reconfigurable circuit wiring and 3D microfluidics. Advanced materials. 25(11):1589e92, Copyright (2013) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

diffraction analysis and rheological studies. In the field of monitoring, the reported techniques include Raman spectroscopy, acoustic emission and ultrasonics. The following section presents an overview of the research conducted by several groups, with specific examples for each technique. 4.1. Characterization techniques 4.1.1. Fourier transform infrared spectroscopy (FTIR) FTIR is a well-established technique based on molecular interactions. In the field of self-healing materials, this technique is by far the mostly employed methodology in order to confirm the healing functionality, compare the virgin and healed materials, as well as to monitor the process of self-healing reactions. It has been employed both for the characterization of microcapsules [26,39,113e115] and intrinsic self-healing systems based on DielseAlder cycloadditions [11,116e118]. FTIR has also been used in other self-healing systems such as self-healing gels [118], intrinsic reversible crosslinked networks healed via photocyclization or on disulfide links [119], mendable epoxy networks and 3D braided composites with vascular channels, or polyurethane/graphene self-healable nanocomposites [120]. An interesting example of the use of FTIR can be found in the work of Araya-Hermosilla et al. [7] who presented a novel reversible thermoset with tunable Tg based on chemical modification of aliphatic polyketones and furan and/or amine groups. In this material system they monitored the cycloaddition through the spectral band of CeO stretching around 1000e1300 cm1. As the molar ratio between furan and maleimide groups increased, the intensity of the band centered around 1180 cm1 (corresponding to CeOeC ether peak) also increased, thus testifying the occurrence of the DielseAlder reaction.

In another case concerning anti-corrosive self-healing organic
et al. [121] investigated the application of linseed oil coatings Szabo e a film former healing material e and octadecylamine (ODA) e a corrosion inhibitor in the core of microcapsules which were added in a self-healing paint using FTIR. Moreover, they tried to specify the influence of Co-octoate, used as a drier in order to reduce solidification time and thus improve the self-healing ability of the paint. They found out that seven days were needed by the linseed oil film in order to dry completely and additionally that this amount of time decreased to several hours with the addition of Co-octoate. Regarding the self-healing functionality they encountered some difficulties with the addition of ODA which weakened the healing process. This difficulties were overcame by increasing the Cooctoate concentration. In another encapsulation attempt, Garcia et al. [29], utilized FTIR in order to confirm the hydrolysis of a water reactive silyl ester which had been encapsulated in a self-healing anticorrosive organic coating. Through the use of FTIR and contact angle measurements they showed that this silyl ester, after its hydrolysis, had the ability to completely coat a metallic surface and form a hydrophobic protective layer, which actually became denser with time. Yuan et al. [122] produced a self-healing system based on cyanate ester resins (CE) with the addition of low molecular weight poly(phenylene oxide) resins (PPO). This CE/PPO system was studied via FTIR in order to quantitatively estimate the extent of conversion, a, of the cyanate ester groups (eOCN) and the amount, x, of unreacted eOCN groups according to the FTIR spectra of uncured and cured CE resin/PPO resin, as is shown in Fig. 29. They used as a reference peak the vibration band of the phenyl ring at 1510 cm1 and chose the vibration bands of eOCN at 2280/ 2238 cm1 to calculate x and a. Moreover, they attributed the improved flexural strength of CE/PPO systems to the higher conversion (a) of eOCN detected by FTIR spectroscopy.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

4.1.2. Near-infrared spectroscopy (NIR) Near infrared spectroscopy has been utilized by Varley et al. [56] as a convenient technique to compare the concentration of different functional groups and determine whether there was any impact on a mendable epoxy network as a result of thermoplastic addition, which may affect the healing process. Using this technique the aforementioned group demonstrated that the modification of their mendable epoxy system, with different healing agents, had a negligible impact upon the network formation, or, the chemistry of polymerization after curing had occurred. This was assumed as NIR did not reveal any changes neither between the spectra of the unmodified and the modified material, nor between the spectra acquired before and after healing of the modified epoxies. In conclusion, they confirmed that healing was more likely via physical processes namely diffusion through free volume and reputation across a crack plane during thermal activation. 4.1.3. Nuclear magnetic resonance spectroscopy (NMR) This experimental technique is typically used in order to exploit the magnetic properties of certain atomic nuclei. Relying on the phenomenon of nuclear magnetic resonance it can provide detailed information about the structure, dynamics, reaction state and chemical environment of the molecules. In the case of self-healing materials numerous studies have used NMR in order to identify interactions between atoms and to confirm the formation of self-healing systems through various chemical reactions. These studies include not only intrinsic chemistries such as the nano-composite self-healing gel produced by

109

Sharma [118] et al., the supramolecular self-healing gels exhibited by Zhang et al. [123]. An example of NMR utilization in microcapsule based selfhealing systems is the work of Zhu et al. [40] who, through the use of NMR, confirmed that capsule rapture led to polymerization achievement, thus confirming that the produced multilayered capsules enclosed the monomer needed for self-healing reaction to occur. Furthermore, Kakuta et al. [15] employed NMR to trace a reason why the formation of inclusion complexes plays such an important role in the formation of their supramolecular preorganized hydrogel system. This system was based on non-covalent hosteguest interactions between polymers and was produced by radical copolymerization of monomers of a complex of a cyclodextrin (CD) host and aliphatic guest in aqueous solution. The group of Kakuta showed that the inclusion complexes undergo a dissolving effect between the CD and guest monomers causing homogeneous radical copolymerization which results to the production of a supramolecular self-healing hydrogel system. Jinhui et al. [124] also, used NMR in order to confirm the successful modification of a commercial epoxy resin with furan groups as well as the occurrence of the self-healing DielseAlder reaction between the modified epoxy and bismaleimide. 4.1.4. Optical microscopy (OM), scanning electron microscopy (SEM) and transmission electron microscopy (TEM) Microscopy is widely used to confirm the self-healing components and structures, after the production step, optical microscopy

Fig. 27. Evolution of the normalized bridge voltage and force during four-point bend tests of a self-healing specimen (a) and a control specimen (b). (c) The percentage of samples where healing was observed [111]. Reprinted from Blaiszik BJ, Kramer SL, Grady ME, McIlroy DA, Moore JS, Sottos NR, et al. Autonomic restoration of electrical conductivity. Advanced materials. 24(3):398e401, Copyright (2012) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

110

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 28. Stress and normalized electrical resistance of an ECNT coating (a) without capsules, with (b) hexyl acetate capsules, (c) EPA:EPON microcapsules, and (d) EPA:ECNT [112]. Reprinted with permission from Bailey BM, Leterrier Y, Garcia SJ, van der Zwaag S, Michaud V. Electrically conductive self-healing polymer composite coatings. Progress in Organic Coatings. 2015;85:189e98, Copyright (2015), with permission from Elsevier.

Fig. 29. Fourier transform infrared spectra of poly(phenylene oxide) (PPO) and the uncured and cured cyanate ester (CE)/poly(phenylene oxide) systems [122]. Reprinted from Yuan L, Huang S, Hu Y, Zhang Y, Gu A, Liang G, et al. Poly(phenylene oxide) modified cyanate resin for self-healing. Polymers for Advanced Technologies. Copyright (2014) John Wiley & Sons Ltd.

(OM), scanning electron microscopy (SEM) or transmittance electron microscopy (TEM) are employed depending on the size of the studied morphology. Numerous researchers employed SEM to study various features including epoxy/hardener containing microcapsules [23,33], fracture surfaces of microcapsules [114], mendable epoxy resins [125], unhealed and healed CFRPs [126], polyurethane/graphene selfhealing nanocomposites [95], solvent-filled microcapsules incorporated into a polyurethane layer which is deposited atop a silver ink line for restoring electrical conductivity of the ink [88], shape memory polymers [104] as well as healing agent containing micro/ nanocapsules embedded in anticorrosive coatings [105]. A very interesting analytical study using SEM in order to view a three-dimensional image of both the inner and outer surface and morphology of various capsules ranging from several tens of microns to below 100 nm in size, has been published by Hodoroaba et al. [127]. In this study SEM was used in the Transmission Mode and the samples were prepared on thin supporting foils (on TEM grids). Fig. 30 shows SEM micrographs with corresponding EDX analyses presented in their study. Li et al. [128] produced a cement based system containing selfhealing microcapsules. Then, they studied those microcapsules

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

composed of diglycidyl ether of bisphenol A epoxy resin as the core material and polystyreneedivinylbenzene as the shell material and the fracture surfaces of them via OM and SEM. SEM was also employed by the group of Jinhui et al. [124] in order to assess the self-healing process of the produced self-healing epoxy system. TEM is mostly used in self-healing systems which incorporate carbon allotrope nanoinclusions as well as in some capsule based self-healing systems with sub-micron sized capsules. Specifically, TEM was used in a study published by Leterrier et al. [112] who synthesized an electrically conductive partially cured epoxy coating incorporating a microcapsule based healing mechanism. The microcapsules contained a mixture of ethyl phenylacetate and a nanoreinforced epoxy resin, the matrix was also reinforced with nanoinclusions. TEM facilitated to the visualization of the carbon nanotube distributions into the core of the microcapsules. 4.1.5. AFM (Atomic Force Microscopy) Atomic force microscopy is generally utilized in self-healing systems in order to assess their healing performance in terms of temperature, time and local mobility of the atoms of the studied materials. Brancart et al. [8] performed an extensive study in order to assess self-healing coatings based on reversible polymer networks using AFM. They studied the self-healing ability of the coatings through the healing of well-defined and reproducible nanosized scratches and other defects applied by nanolithography. The group of Dikic [129] produced a self-replenishing hydrophobic coating based on perfluoroalkyl dangling chains covalently bonded to a cross-linked polymer network through a polymeric spacer. They used AFM to assess the self-replenishing process by comparing the fluorine end groups concentration at the virgin and

111

healed specimen. The decrease in the Force Displacement mean value measured by AFM depicting an increase in fluorine content in the healed materials showed that the prerequisite for replenishing is the mobility of the species within the coating network. This mobility was triggered by annealing. In a another study, Faghihnejad et al. [130] employed AFM to characterize self-healing films with multiple hydrogen bonding groups with respect to changes in the Relative Humidity (RH) of the films. During this contact mechanic testing they observed a transition from elastic to viscous failure. Characteristic profiles are presented in Fig. 31. 4.1.6. X-ray diffraction analysis X-ray diffraction as a technique can be utilized in order to identify certain interactions between molecules and stacking interactions exhibited as a diffraction pattern. Roy et al. [131] produced an amino-acid-based (11-(4-(pyrene-1-yl)butanamido) undecanoic acid) self-repairing hydrogel which contained carbon nanoparticles (graphene and single wall carbon nanotubes) for the incorporation of semiconducting behavior. X-ray diffraction analysis was used in order to evaluate the bonding interactions of the produced hydrogel. The conclusions of their study declared that the gelator molecules being used in their system are self-assembled through hydrogen-bonding interactions between the amide moieties and pep stacking interactions. They, also, recognized the presence of pep stacking interactions of the pyrene p planes of the hydrogelator molecules with the p planes of the graphene, as well as the p walls of the Pr-SWCNTs, in the hybrid gel state. 4.1.7. X-ray photoemission spectroscopy (XPS) XPS is a surface-sensitive quantitative spectroscopic technique that measures the elemental composition at the parts per thousand

Fig. 30. Upper (a) and in-transmission (b) SEM observation of SiO2 submicrocapsules prepared on lacey carbon foil on TEM grids. (c) EDX analysis confirming compositional differences of the two particles [127]. Reprinted from Hodoroaba VD, Akcakayiran D, Grigoriev DO, Shchukin DG. Characterization of micro- and nanocapsules for self-healing anti-corrosion coatings by high-resolution SEM with coupled transmission mode and EDX. The Analyst. 2014;139(8):2004e10. Published by The Royal Society of Chemistry.

112

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 31. Typical topographical AFM images of surface patterns associated with the detachment of two self-healing films (thickness ~100 nm) from adhesive contact in contact mechanics tests: a) more viscous state, bed) more elastic state, T ¼ 40
C [130]. Reprinted from Faghihnejad A, Feldman KE, Yu J, Tirrell MV, Israelachvili JN, Hawker CJ, et al. Adhesion and Surface Interactions of a Self-Healing Polymer with Multiple HydrogenBonding Groups. Advanced Functional Materials. 24(16):2322e33, Copyright (2014) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

range, chemical state and electronic state of the elements that exist within a material. Nowak et al. [104] prepared a self-healing polymer coating which contained the organic corrosion inhibitors encapsulated inside the polyelectrolyte nanocapsules. This coating was painted onto the surface of AA2024 T3 alloy. They utilized XPS in order to confirm the release of the inhibitor from a scratched coating and thus the healing reaction of the produced coating. Fig. 32 shows the acquired spectra from an unscratched surface, a dry scratched surface and a wetted scratched surface of this self-healing organic coating. The peak which appears at a binding energy of 163.5 eV on the spectrum of the dry scratched surface was ascribed to the inhibitor liberation from the destroyed capsules. The shift of this peak to the lower value of 162.4 eV of binding energy in the wetted sample confirms the formation of a sulfurealuminum bond.

4.1.8. Rheometry, rheological studies and thixotropic behavior Rheological measurements are being performed in self-healing systems, mostly hydrogel systems, in order to determine the solegel transition point of these systems and to measure the storage (G0 ) and loss (G00 ) moduli as a function of temperature, in a certain angular frequency, within a linear range of viscoelasticity. Some research groups are also conducting measurements of the thixotropic properties of their systems in order to evaluate the gel recovery time. More specifically, Sharma et al. [118] and Scheltjens et al. [117] examined the linear viscoelastic properties to fully estimate the thermo-mechanical behavior of their systems during healing. They assessed which of their produced materials behaved more like a solid behavior and which like a gel by measuring G0 and G00 in dependence with time. The solegel transition temperature of their reversible covalent bonding self-healing

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

113

ultrasound to monitor in situ the progression of self-healing of cracks in concrete. Ultrasonics were employed to investigate fully or partially fractured tensile tested specimens. Self-healing was assessed via monitoring of distinct diffuse ultrasonic parameters such as Arrival Time of Maximum Energy (ATME) and effective ultrasonic diffusivity. The latter term refers to the ratio of the amount of space per time that the ultrasound waves are able to occupy during their diffusion by the scatterers included in a material like cement. ATME has shown to be correlated with crack width which in this system can represent the extent of healing. An indicative diagram presenting their measurements is shown in Fig. 34. A comparison of the measured crack width and diffusivity over time is depicted in this diagram. The authors conclude by stating an exponentialerecovery model which characterizes their material and relates the measured diffusivity with the amount of self-healing.

Fig. 32. S 2p ionization spectra of a self-healing polymer coating with encapsulated organic corrosion inhibitors. Unscratched coating (down), dry scratch (middle), wetted scratch (top) [104]. Reprinted from Progress in Organic Coatings, Vol. 84, Kope
c M, Szczepanowicz K,
rna K, Socha RP, Nowak P, et al. Self-healing epoxy coatings loaded Mordarski G, Podgo with inhibitor-containing polyelectrolyte nanocapsules. pp. 97e106, Copyright (2015), with permission from Elsevier.

polymer network in correspondence with equilibrium conditions was determined. Roy et al. [131] conducted a rheological study on an epoxy amine based network modified with reversible bonds. This study which confirmed the presence of a stable and stiff gel-phase material. Interestingly, the stiffness of the native gel increased upon the incorporation of Pr-SWCNTs (pristine single walled carbon nanotubes), RGO (reduced graphene oxide), and both Pr-SWCNTs and RGO into the hybrid gel system. The study of the thixotropic properties of the produced hydrogels revealed the gel-recovery time was significantly shortened from 7 min to 2.48 min, 3.25 min and 3 min by the incorporation of both SWCNTs and RGO, and the individual inclusion of RGO and Pr-SWCNTs, respectively, within the native gel. indicating fastest self-healing ability in the case of the hybrid gel with RGO and Pr-SWCNTs, and the slowest self-healing ability in the case of the native gel. Another investigation of the rheological behavior of a selfhealing system is described in the work by Hao et al. [16]. This attempt includes the determination of the sol gel transition point, the G0 and G00 moduli as well as the recovery properties of a thermal-responsive self-healing hydrogel based on hydrophobically modified chitosan and vesicle. They found that depending on temperature and the consequent state of their material (solution, viscoelastic gel etc.) there is a strong frequency dependence of G0 and G00 . Furthermore, the relaxation time, which reflects the life time of the cross-links, was further investigated, coming to the conclusion that the behavior of their hydrogel was governed by the interplay of the hydrophobic attractive interactions and the electrostatic repulsive interactions, with the latter playing a negative part. Fig. 33, depicts the relaxation times as a function of frequency for various concentrations of 5-methyl salicylic (mS) acid at 25
C. 4.2. Monitoring techniques 4.2.1. Ultrasonics Being a very effective non-destructive technique, ultrasonics have been utilized by many researchers for self-healing process monitoring. A study published by In et al. [132] employed diffused

4.2.2. Acoustic emission Acoustic emission is a non-destructive technique used in many cases in order to detect damage evolution and propagation in materials. In the case of self-healing materials, Coppola et al. [49] utilized acoustic emission in order to assess the damage evolution in the vascular 3D woven glass/epoxy composites. In more detail, they correlated the acoustic activity with strain data from mechanical testing. Fig. 35 presents stress and acoustic emission data in relation to the strain that the composites are subjected to. They observed that, for all the specimens, tested acoustic events initiated after a threshold strain was reached, denoted by εAEi. Moreover, acoustic activity initiated earlier for the vascular specimens compared to the control ones (specimens without channels) indicating earlier onset of damage. This early onset indicated that other mechanisms were responsible for the reduction of strength in the wave channel specimens (specimen in which the channels follow a wave shape trajectory) although it did not directly correlate to reduced strength. 4.2.3. Raman spectroscopy Raman spectroscopy is a valuable characterization tool which has been utilized in different ways in order to evaluate the produced structures and assess the healing performance of self-healing systems. A short presentation of the several attempts of different working groups during the most recent years can be found in the next paragraphs. In a review article by Zedler et al. [133], four different aspects related to the use of resonance Raman spectroscopy in the study of self-healing materials are presented. The first aspect is concerned with resonance Raman investigation of self-healing biopolymers living in marine habitats. These biopolymers are presented as model damage tolerant and self-repairing structures, in the work of Holten-Andersen et al. [134]. Other aspects presented in the aforementioned review, concern the use of traditional Raman spectroscopy for the investigation of the role of protein conformational changes in the self-healing behavior of the whelks egg capsules (WEC), along with some examples of monitoring temporally- and spatially-resolved changes in the local chemistry, as well as the stereospecificity of Ring Opening Metathesis Polymerization (ROMP) self-healing reactions. Furthermore, it covers examples of in situ Raman characterization of the epoxy based polymerization products formed in the cracked surface of microcapsule- and micro-vascular based self-healing materials, as well as an investigation and confirmation of DielseAlder (DA), retro-DielseAlder (rDA) self-healing reactions of coatings based on a polymer modified with terpyridineemetal complexations.

114

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 33. Dynamic rheological study. (a) Storage modulus (G0 ) and loss modulus (G00 ) as a function of frequency for 0.4 wt% solution of hm-chitosan with different concentration of 5 mS at 25
C. (b) Relaxation of the concentration of 5 mS at 25
C [16]. Reprinted from Colloid and Polymer Science, Vol. 291(7), 2013, pp. 1749e58, Thermal-responsive self-healing hydrogel based on hydrophobically modified chitosan and vesicle. Hao X, Liu H, Xie Y, Fang C, Yang H., Figure 5, Original Caption: “Dynamic rheological study. (a) Storage modulus G0 (filled symbols) and loss modulus G00 (open symbols) as a function of frequency for 0.4 wt% solution of hm-chitosan with different concentration of 5 mS (DTAB ¼ 16 mM) at 25
C. (b) The relaxation time as a function of the concentration of 5 mS at 25
C, the DTAB concentration is 16 mM.”, Copyright Springer-Verlag Berlin Heidelberg 2013, with kind permission from Springer Science and Business Media.

Fig. 34. ATME and diffusivity plot with tension crack over exposure time for (a) 100% cement material, (b) 60% cement mass:35% slag:5% Metakaolin [132]. Reprinted from NDT & E International, Vol. 57, In C-W, Holland RB, Kim J-Y, Kurtis KE, Kahn LF, Jacobs LJ. Monitoring and evaluation of self-healing in concrete using diffuse ultrasound. pp. 36e44, Copyright (2013), with permission from Elsevier.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 35. A sample plot of stress and cumulative acoustic emissions (AEcum) versus strain. Locations used to record ultimate tensile stress (su), Young's modulus (E), and strain at damage initiation (εAEi) [49]. Reprinted from Composites Part A: Applied Science and Manufacturing, Vol. 59, Coppola AM, Thakre PR, Sottos NR, White SR. Tensile properties and damage evolution in vascular 3D woven glass/epoxy composites. pp. 9e17, Copyright (2014), with permission from Elsevier.

The group of Zhu et al. [40], who synthesized multilayered microcapsules in a production sequence of four steps, used Raman spectroscopy for the characterization of the produced structures. These multi-layered microcapsules are comprised of distinct layers of glycidyl methacrylate (GMA) e loaded poly(melamine-formaldehyde) (PMF) microcapsules as the core, living PMMA as the second layer, cuprous bromide/N,N,N0 ,N0 ,N00 -pentamethyldiethylenetriamine catalyst system (CuBr/PMDETA) as the third layer and a wax coat as the outer shell to protect the air sensitive Cu(I) in the second layer. The chemical structure of the microcapsules was verified via Raman microscopy depth profiling, depicted in Fig. 36. The profile not only described the composition feature of the entire multilayered microcapsule, but also facilitated for a rough measurement of the thickness of each layer. Patrick et al. [135] probed the chemical composition of the healed fracture interfaces of their 3D vascular composites. They

115

produced two different architectures of vascularized composites, namely an isolated-“parallel” and an interpenetrating-“herringbone” one. The epoxy system used was a diglycidyl ether of bisphenol A (DGEBA) based epoxy resin (R) and aliphatic triethylenetetramine (TETA) based hardener (H). Raman spectra at various locations of the parallel configuration sample indicated the resin and hardener rich regions corresponding to distinct stretching and vibrating modes of different molecules, as well as a partially mixed region. In addition, they conducted a series of ex situ, pre-mixed Raman investigations, in order to construct a linear calibration curve which assisted the quantification of the in-situ resin:hardener (R:H) mix proportions of the healed material based on the ratio of phenyl to amide peak intensities. They acquired Raman spectra from three different regions on the herringbone fracture surface, which yielded calculated R:H ratios of roughly 3:1, 2:1 and 3:2, respectively. These values confirmed the ability of the interpenetrating vasculature of approximate intended fluid delivery as well as the ability of the selected healing chemistry to polymerize under non-stoichiometric conditions, as can be seen in Fig. 37. In the category of microcapsule self-healing materials Yuan et al. [136] utilized Raman spectroscopy in order to acquire a live record of the curing reaction of the released healing agent in their system which comprised of glass fabric epoxy composites and epoxy/ mercaptan healing agent. Chipara et al. [137] used Raman spectroscopy in order to characterize a self-healing system containing urea-formaldehyde microcapsules filled with dicyclopentadiene (DCPD) and first generation Grubbs' catalyst dispersed within polyethylene oxide. In another work, Ramachandran et al. [138] produced copolymer films that upon mechanical damage undergo color changes in the damaged area, but upon exposure to sunlight, temperature and/ or acidic vapors, the damaged area is self-repaired, recovering the initial colorless appearance. The aforementioned group used Raman spectroscopy for monitoring the molecular repair processes induced by visible light. In the category of intrinsic photopolymerizable self-healing materials Zhang et al. [61] used Raman spectroscopy in order to estimate the appropriate time for photodimerization until equilibrium would be reached in modified polyurethane network, in which coumarin served as a photosensitive crosslinker. They verified the homogeneity of the photochemical reaction throughout the entire polyurethane film volume.

Fig. 36. (a e left) Raman spectra of the substances used for composing the multilayered microcapsules. (b e right) Typical Raman spectra of a multilayered microcapsule z120 mm in diameter collected at the scanning depths from top to bottom [40]. Reprinted from Polymer, Vol. 54(16), Zhu DY, Rong MZ, Zhang MQ. Preparation and characterization of multilayered microcapsule-like microreactor for self-healing polymers. pp. 4227e36. Copyright (2013), with permission from Elsevier.

116

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

Fig. 37. In situ healing reaction characterization (single cycle, Da ¼ 70 mm) via fluorescent images in combination with Raman spectroscopy [135]. Reprinted from Patrick JF, Hart KR, Krull BP, Diesendruck CE, Moore JS, White SR, et al. Continuous Self-Healing Life Cycle in Vascularized Structural Composites. Advanced materials. Copyright (2014) Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

In a very recent work, White et al. [139] produced a cross-linked epoxy polymeric material which incorporated an imidazole polymerization initiator into the matrix volume. An aliphatic bond formation shown as a peak at 1112 cm1 in the Raman spectrum of the healed material compared to the original spectrum of the virgin material confirmed the success of the healing process as well as the proposed healing mechanism in the modified epoxy system.

5. Concluding remarks A wide variety of novel chemistries regarding both extrinsic and intrinsic self-healing approaches have been developed by the research community over the last years. As has been shown, significant effort has been made towards developing an ideal fully autonomous intrinsic self-healing system capable of regaining its initial properties rapidly at ambient temperature. Reversible reactions and especially DielseAlder mechanism are very promising for the synthesis of processible, remendable and highly oriented polymers. This class of materials is capable of regaining its initial properties at the molecular level, theoretically for an infinite number of repetitions without any further addition of chemicals. However, only small volumes of damage can be healed because material contact is required for healing, and cyclic reactions reduce the healing efficiency after repeated healing cycles.

On the other hand, vascular self-healing materials are capable of multiple healing of large damage volumes since the healing agent can be repeatedly infused to the damaged area through the formed networks. In this case, the challenge lies in the incorporation of a microchannel network to a material. Significant progress has been made towards the integration of a vascular network that will not affect the properties of the existing material through the optimization of the manufacturing process. Apart from the decision for the selection of the self-healing chemistry, the engineer or scientist needs to have an extensive range of testing procedures in order to select the appropriate one depending the application. Depending on the final product, stiffness, impact resistance, adhesion strength, electrical conductivity, corrosion protection are of primary importance. The novel healing material technologies target applications in various sectors such as the aerospace, automotive, communications etc. However, a considerable impediment for their use is the certification, particularly for aerospace affiliated sectors. The various specimen geometries or experimental configurations range from the tapered double cantilever beam which is widely applicable despite its fundamental disadvantages, to typical Wheatstone bridge setups when electrical properties are of interest. Self-healing corrosion resistant coatings can find excellent characterization procedures in electrochemical impedance spectroscopy or cyclic potentiodynamic sweep in aggressive environments. Moreover the available choices in experimental

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

equipment and procedures enable the researcher to even monitor the actual healing process gaining more insight on the self-healing chemistry and engineering. Characterization of self-healing materials through various experimental techniques and evaluation of their self-healing efficiency via monitoring of the occurring self-healing reactions is naturally following the material production step. Microscopic methods like OM, SEM and TEM serve the visualization of the produced structures and are generally used before and after healing, thus confirming its success. These techniques are sometimes combined with AFM when studying the local mobility of atoms and generating topographical images of the virgin and healed materials. Other characterization methods, including NMR, XRD and XPS, that can study interactions in the atomic or molecular level, are also being utilized in order to confirm successful healing. A qualitative estimation of healing and conclusions about its mechanism are also achieved through the use of NIR and FTIR with the latter in some cases also leading to quantitative estimations of healing efficiency. Rheological studies are mostly being used in modified systems, in which the modification accounts for the healing functionality, in a comparative way as to the unmodified ones. Monitoring of the various systems healing process can be achieved via the utilization of ultrasonics, acoustic emission and Raman spectroscopy with the former and the latter one being able to estimate a quantitative value of the healing efficiency. Summarizing, self healing materials and related technologies have substantially progressed in the recent years. However, fundamental to their application is their integration in the component design via the selection of the appropriate technology while always taking in to account apart from the desired healing functionality, production and certification issues. Acknowledgments This research has been co-financed by the European Union (European Social Fund e ESF) and Greek national funds through the Operational Program “Education and Lifelong Learning” of the National Strategic Reference Framework (NSRF) e Research Funding Program: THALES. Investing in knowledge society through the European Social Fund. The authors also acknowledge the “ACP3GA-20l3-6054l2-HIPOCRATES” research programs for financial support. References [1] Chen X, Dam MA, Ono K, Mal A, Shen H, Nutt SR, et al. A thermally remendable cross-linked polymeric material. Science 2002;295(5560): 1698e702. [2] White SR, Sottos NR, Geubelle PH, Moore JS, Kessler MR, Sriram SR. Autonomic healing of polymer composites. Lett Nat 2001;409:794e7. [3] Postiglione G, Turri S, Levi M. Effect of the plasticizer on the self-healing properties of a polymer coating based on the thermoreversible DielseAlder reaction. Prog Org Coat 2015;78:526e31. [4] Park JS, Darlington T, Starr AF, Takahashi K, Riendeau J, Thomas Hahn H. Multiple healing effect of thermally activated self-healing composites based on DielseAlder reaction. Compos Sci Technol 2010;70(15):2154e9. € tteritzsch J, Stumpf S, Hoeppener S, Vitz J, Hager MD, Schubert US. One[5] Ko component intrinsic self-healing coatings based on reversible crosslinking by DielseAlder cycloadditions. Macromol Chem Phys 2013;214(14):1636e49. [6] Zhang W, Duchet J, Gerard JF. Self-healable interfaces based on thermoreversible DielseAlder reactions in carbon fiber reinforced composites. J Colloid Interface Sci 2014;430:61e8. [7] Araya-Hermosilla R, Broekhuis AA, Picchioni F. Reversible polymer networks containing covalent and hydrogen bonding interactions. Eur Polym J 2014;50:127e34. [8] Brancart J, Scheltjens G, Muselle T, Van Mele B, Terryn H, Van Assche G. Atomic force microscopy-based study of self-healing coatings based on reversible polymer network systems. J Intell Mater Syst Struct 2012;25(1):40e6. [9] Ax J, Wenz G. Thermoreversible networks by DielseAlder reaction of cellulose furoates with bismaleimides. Macromol Chem Phys 2012;213(2): 182e6.

117

[10] Amamoto Y, Otsuka H, Takahara A, Matyjaszewski K. Self-healing of covalently cross-linked polymers by reshuffling thiuram disulfide moieties in air under visible light. Adv Mater 2012;24(29):3975e80. [11] Toncelli C, De Reus DC, Picchioni F, Broekhuis AA. Properties of reversible DielseAlder furan/maleimide polymer networks as function of crosslink density. Macromol Chem Phys 2012;213(2):157e65. [12] Rivero G, Nguyen L-TT, Hillewaere XKD, Du Prez FE. One-pot thermoremendable shape memory polyurethanes. Macromolecules 2014;47(6): 2010e8. [13] Sch€ afer S, Kickelbick G. Self-healing polymer nanocomposites based on DielseAlder-reactions with silica nanoparticles: the role of the polymer matrix. Polymer 2015. http://dx.doi.org/10.1016/j.polymer.2015.03.017. [14] Ce Yuan, Rong MZ, Zhang MQ. Self-healing polyurethane elastomer with thermally reversible alkoxyamines as crosslinkages. Polymer 2014;55(7): 1782e91. [15] Kakuta T, Takashima Y, Nakahata M, Otsubo M, Yamaguchi H, Harada A. Preorganized hydrogel: self-healing properties of supramolecular hydrogels formed by polymerization of host-guest-monomers that contain cyclodextrins and hydrophobic guest groups. Adv Mater 2013;25(20):2849e53. [16] Hao X, Liu H, Xie Y, Fang C, Yang H. Thermal-responsive self-healing hydrogel based on hydrophobically modified chitosan and vesicle. Colloid Polym Sci 2013;291(7):1749e58. [17] Lafont U, Moreno-Belle C, van Zeijl H, van der Zwaag S. Self-healing thermally conductive adhesives. J Intell Mater Syst Struct 2013;25(1):67e74. [18] Roy N, Buhler E, Lehn J-M. Double dynamic self-healing polymers: supramolecular and covalent dynamic polymers based on the bisiminocarbohydrazide motif. Polym Int 2014;63(8):1400e5. [19] An SY, Noh SM, Nam JH, Oh JK. Dual sulfide-disulfide crosslinked networks with rapid and room temperature self-healability. Macromol Rapid Commun 2015. http://dx.doi.org/10.1002/marc.201500123. [20] Hondred PR, Salat L, Mangler J, Kessler MR. Tung oil-based thermosetting polymers for self-healing applications. J Appl Polym Sci 2014. http:// dx.doi.org/10.1002/APP.40406. [21] Moll JL, Jin H, Mangun CL, White SR, Sottos NR. Self-sealing of mechanical damage in a fully cured structural composite. Compos Sci Technol 2013;79: 15e20. [22] Haase MF, Grigoriev DO, Mohwald H, Shchukin DG. Development of nanoparticle stabilized polymer nanocontainers with high content of the encapsulated active agent and their application in water-borne anticorrosive coatings. Adv Mater 2012;24(18):2429e35. [23] Haiyan L, Rongguo W, Wenbo L. Preparation and self-healing performance of epoxy composites with microcapsules and tungsten (VI) chloride catalyst. J Reinf Plast Compos 2012;31(13):924e32. [24] Coope TS, Mayer UFJ, Wass DF, Trask RS, Bond IP. Self-healing of an epoxy resin using Scandium(III) triflate as a catalytic curing agent. Adv Funct Mater 2011;21(24):4624e31. [25] Jackson AC, Bartelt JA, Braun PV. Transparent self-healing polymers based on encapsulated plasticizers in a thermoplastic matrix. Adv Funct Mater 2011;21(24):4705e11. [26] Yuan L, Huang S, Gu A, Liang G, Chen F, Hu Y, et al. A cyanate ester/microcapsule system with low cure temperature and self-healing capacity. Compos Sci Technol 2013;87:111e7. [27] Jin H, Miller GM, Sottos NR, White SR. Fracture and fatigue response of a selfhealing epoxy adhesive. Polymer 2011;52(7):1628e34.
lez-García Y, García SJ, Hughes AE, Mol JMC. A combined redox[28] Gonza competition and negative-feedback SECM study of self-healing anticorrosive coatings. Electrochem Commun 2011;13(10):1094e7.
lez-García Y, Mol JMC, et al. [29] García SJ, Fischer HR, White PA, Mardel J, Gonza Self-healing anticorrosive organic coating based on an encapsulated water reactive silyl ester: synthesis and proof of concept. Prog Org Coat 2011;70(2e3):142e9. [30] Hollamby MJ, Fix D, Donch I, Borisova D, Mohwald H, Shchukin D. Hybrid polyester coating incorporating functionalized mesoporous carriers for the holistic protection of steel surfaces. Adv Mater 2011;23(11):1361e5. [31] Huang M, Yang J. Long-term performance of 1H, 1H0 , 2H, 2H0 -perfluorooctyl triethoxysilane (POTS) microcapsule-based self-healing anticorrosive coatings. J Intell Mater Syst Struct 2013;25(1):98e106. [32] Nesterova T, Dam-Johansen K, Kiil S. Synthesis of durable microcapsules for self-healing anticorrosive coatings: a comparison of selected methods. Prog Org Coat 2011;70(4):342e52. [33] Liu X, Zhang H, Wang J, Wang Z, Wang S. Preparation of epoxy microcapsule based self-healing coatings and their behavior. Surf Coat Technol 2012;206(23):4976e80. [34] Manfredi E, Cohades A, Richard I, Michaud V. Assessment of solvent capsulebased healing for woven E-glass fibre-reinforced polymers. Smart Mater Struct 2015;24(1):015019. [35] Jones AR, Blaiszik BJ, White SR, Sottos NR. Full recovery of fiber/matrix interfacial bond strength using a microencapsulated solvent-based healing system. Compos Sci Technol 2013;79:1e7. [36] Zhang H, Yang J. Development of self-healing polymers via amineeepoxy chemistry: I. Properties of healing agent carriers and the modelling of a twopart self-healing system. Smart Mater Struct 2014;23(6):065003. [37] Jin H, Mangun CL, Griffin AS, Moore JS, Sottos NR, White SR. Thermally stable autonomic healing in epoxy using a dual-microcapsule system. Adv Mater 2014;26(2):282e7.

118

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119

[38] Jackson AC, Bartelt JA, Marczewski K, Sottos NR, Braun PV. Silica-protected micron and sub-micron capsules and particles for self-healing at the microscale. Macromol Rapid Commun 2011;32(1):82e7. [39] Li Q, Siddaramaiah, Kim NH, Hui D, Lee JH. Effects of dual component microcapsules of resin and curing agent on the self-healing efficiency of epoxy. Compos Part B Eng 2013;55:79e85. [40] Zhu DY, Rong MZ, Zhang MQ. Preparation and characterization of multilayered microcapsule-like microreactor for self-healing polymers. Polymer 2013;54(16):4227e36. [41] Toohey KS, Sottos NR, Lewis JA, Moore JS, White SR. Self-healing materials with microvascular networks. Nat Mater 2007;6(8):581e5. [42] Hamilton AR, Sottos NR, White SR. Self-healing of internal damage in synthetic vascular materials. Adv Mater 2010;22(45):5159e63. [43] Norris CJ, Bond IP, Trask RS. Interactions between propagating cracks and bioinspired self-healing vascules embedded in glass fibre reinforced composites. Compos Sci Technol 2011;71(6):847e53. [44] Peterson AM, Kotthapalli H, Rahmathullah MAM, Palmese GR. Investigation of interpenetrating polymer networks for self-healing applications. Compos Sci Technol 2012;72(2):330e6. [45] Grunenfelder LK, Suksangpanya N, Salinas C, Milliron G, Yaraghi N, Herrera S, et al. Bio-inspired impact-resistant composites. Acta Biomater 2014;10(9): 3997e4008. [46] Esser-Kahn AP, Thakre PR, Dong H, Patrick JF, Vlasko-Vlasov VK, Sottos NR, et al. Three-dimensional microvascular fiber-reinforced composites. Adv Mater 2011;23(32):3654e8. [47] Yang T, Wang CH, Zhang J, He S, Mouritz AP. Toughening and self-healing of epoxy matrix laminates using mendable polymer stitching. Compos Sci Technol 2012;72(12):1396e401. [48] Patrick JF, Hart KR, Krull BP, Diesendruck CE, Moore JS, White SR, et al. Continuous self-healing life cycle in vascularized structural composites. Adv Mater 2014;26(25):4302e8. [49] Coppola AM, Thakre PR, Sottos NR, White SR. Tensile properties and damage evolution in vascular 3D woven glass/epoxy composites. Compos Part A Appl Sci Manuf 2014;59:9e17. [50] Hansen CJ, White SR, Sottos NR, Lewis JA. Accelerated self-healing via ternary interpenetrating microvascular networks. Adv Funct Mater 2011;21(22): 4320e6. [51] Trask RS, Norris CJ, Bond IP. Stimuli-triggered self-healing functionality in advanced fibre-reinforced composites. J Intell Mater Syst Struct 2013;25(1): 87e97. [52] Zainuddin S, Arefin T, Fahim A, Hosur MV, Tyson JD, Kumar A, et al. Recovery and improvement in low-velocity impact properties of e-glass/epoxy composites through novel self-healing technique. Compos Struct 2014;108:277e86. [53] Koralagundi Matt AK, Strong S, ElGammal T, Amano RS. Development of novel self-healing polymer composites for use in wind turbine blades. J Energy Resour Technol 2015;137(5):051202. [54] Varley RJ, Craze DA, Mouritz AP, Wang CH. Thermoplastic healing in epoxy networks: exploring performance and mechanism of alternative healing agents. Macromol Mater Eng 2013;298(11):1232e42. [55] Wu XL, Huang WM, Seow ZG, Chin WS, Yang WG, Sun KY. Two-step shape recovery in heating-responsive shape memory polytetrafluoroethylene and its thermally assisted self-healing. Smart Mater Struct 2013;22(12): 125023. [56] Varley RJ, Dao B, Pillsbury C, Kalista SJ, Jones FR. Low-molecular-weight thermoplastic modifiers as effective healing agents in mendable epoxy networks. J Intell Mater Syst Struct 2013;25(1):107e17. [57] Zhang H, Yang J. Development of self-healing polymers via amineeepoxy chemistry: II. Systematic evaluation of self-healing performance. Smart Mater Struct 2014;23(6):065004. [58] Yuan L, Huang S, Hu Y, Zhang Y, Gu A, Liang G, et al. Poly(phenylene oxide) modified cyanate resin for self-healing. Polym Adv Technol 2014;25(7): 752e9. [59] Guadagno L, Raimondo M, Naddeo C, Longo P, Mariconda A, Binder WH. Healing efficiency and dynamic mechanical properties of self-healing epoxy systems. Smart Mater Struct 2014;23(4):045001. [60] Guadagno L, Longo P, Raimondo M, Naddeo C, Mariconda A, Vittoria V, et al. Use of HoveydaeGrubbs' second generation catalyst in self-healing epoxy mixtures. Compos Part B Eng 2011;42(2):296e301. [61] Ling J, Rong MZ, Zhang MQ. Coumarin imparts repeated photochemical remendability to polyurethane. J Mater Chem 2011;21(45):18373. [62] Wu AS, Coppola AM, Sinnott MJ, Chou T-W, Thostenson ET, Byun J-H, et al. Sensing of damage and healing in three-dimensional braided composites with vascular channels. Compos Sci Technol 2012;72(13):1618e26. [63] Ling J, Rong MZ, Zhang MQ. Photo-stimulated self-healing polyurethane containing dihydroxyl coumarin derivatives. Polymer 2012;53(13):2691e8. [64] International A. D 3039 e Standard test method for tensile properties of polymer matrix composite materials. 2000. [65] International A. D 638 e Standard test method for tensile properties of plastics. 2002. [66] International A. D 1708 e Standard test method for tensile properties of plastics by use of microtensile specimens. 2002. [67] Grande AM, Castelnovo L, Landro LD, Giacomuzzo C, Francesconi A, Rahman MA. Rate-dependent self-healing behavior of an ethylene-comethacrylic acid ionomer under high-energy impact conditions. J Appl Polym Sci 2013. http://dx.doi.org/10.1002/APP.39384.

[68] Mirabedini SM, Dutil I, Gauquelin L, Yan N, Farnood RR. Preparation of selfhealing acrylic latex coatings using novel oil-filled ethyl cellulose microcapsules. Prog Org Coat 2015;85:168e77. [69] Murphy EB. The return of photoelastic stress measurements: utilizing birefringence to monitor damage and repair in healable materials. J Mater Chem 2011;21:8. [70] Maes F, Montarnal D, Cantournet S, Tournilhac F, Corte L, Leibler L. Activation and deactivation of self-healing in supramolecular rubbers. Soft Matter 2012;8:6. [71] International A. D 1002 e Standard test method for apparent shear strength of single-lap-joint adhesively bonded metal specimens by tension loading (Metal-to-Metal). 2000. [72] International A. D 5868 e Standard test method for lap shear adhesion for fiber reinforced plastic (FRP) bonding. 2001. [73] Lafont U, van Zeijl H, van der Zwaag S. Influence of cross-linkers on the cohesive and adhesive self-healing ability of polysulfide-based thermosets. ACS Appl Mater Interfaces 2012;4(11):6280e8. [74] International A. D 897 e Standard test method for tensile properties of adhesive bonds. 2000. [75] International A. D 3846 e Standard test method for in-plane shear strength of reinforced plastics. 2002. [76] Hondred PR, Autori C, Kessler MR. Rare earth triflate initiators in the cationic polymerization of tung oil-based thermosetting polymers for self-healing applications. Macromol Mater Eng 2014. http://dx.doi.org/10.1002/ mame.201300437. [77] Wu X-F, Rahman A, Zhou Z, Pelot DD, Sinha-Ray S, Chen B, et al. Electrospinning core-shell nanofibers for interfacial toughening and self-healing of carbon-fiber/epoxy composites. J Appl Polym Sci 2012. http://dx.doi.org/ 10.10002/APP.38838. [78] International A. D 790 e Standard test methods for flexural properties of unreinforced and reinforced plastics and electrical insulating materials. 2003. [79] Li G, Zhang P. A self-healing particulate composite reinforced with strain hardened short shape memory polymer fibers. Polymer 2013;54(18): 5075e86. [80] International A. D 5045 e Standard test methods for plane-strain fracture toughness and strain energy release rate of plastic materials. 99. [81] Meure S, Varley RJ, Wu DY, Mayo S, Nairn K, Furman S. Confirmation of the healing mechanism in a mendable EMAAeepoxy resin. Eur Polym J 2012;48(3):524e31. [82] Brown EN. Use of the tapered double-cantilever beam geometry for fracture toughness measurements and its application to the quantification of selfhealing. J Strain Anal Eng Des 2011;46(3):167e86. [83] Raimondo M, Guadagno L. Healing efficiency of epoxy-based materials for structural applications. Polym Compos 2013;34(9):1525e32. [84] Everitt DT, Coope TS, Trask RS, Wass DF, Bond IP. Application of a silvereolefin coordination polymer as a catalytic curing agent for self-healing epoxy polymers. Smart Mater Struct 2015;24(5):055004. [85] Neuser S, Michaud V. Fatigue response of solvent-based self-healing smart materials. Exp Mech 2013;54(2):293e304.
mez D, Gilabert FA, Tsangouri E, Van Hemelrijck D, Hillewaere XKD, [86] Garoz Go Du Prez FE, et al. In-depth numerical analysis of the TDCB specimen for characterization of self-healing polymers. Int J Solids Struct 2015;64e65: 145e54. [87] Pingkarawat K, Wang CH, Varley RJ, Mouritz AP. Effect of mendable polymer stitch density on the toughening and healing of delamination cracks in carboneepoxy laminates. Compos Part A Appl Sci Manuf 2013;50:22e30. [88] Hargou K, Pingkarawat K, Mouritz AP, Wang CH. Ultrasonic activation of mendable polymer for self-healing carboneepoxy laminates. Compos Part B Eng 2013;45(1):1031e9. [89] Varley RJ, Charve F. EMAA as a healing agent for mendable high temperature epoxy amine thermosets. Compos Part A Appl Sci Manuf 2012;43(7): 1073e80. [90] International A. D 5528 e Standard test method for mode I interlaminar fracture toughness of unidirectional fiber-reinforced polymer matrix composites. 2001. [91] Wang CH, Sidhu K, Yang T, Zhang J, Shanks R. Interlayer self-healing and toughening of carbon fibre/epoxy composites using copolymer films. Compos Part A Appl Sci Manuf 2012;43(3):512e8. [92] Pingkarawat K, Wang CH, Varley RJ, Mouritz AP. Self-healing of delamination cracks in mendable epoxy matrix laminates using poly[ethylene-co(methacrylic acid)] thermoplastic. Compos Part A Appl Sci Manuf 2012;43(8):1301e7. [93] Manfredi E, Michaud V. Packing and permeability properties of E-glass fibre reinforcements functionalised with capsules for self-healing applications. Compos Part A Appl Sci Manuf 2014;66:94e102. [94] Ouyang Z, Li G. Nonlinear interface shear fracture of end notched flexure specimens. Int J Solids Struct 2009;46(13):2659e68. [95] Hamilton AR, Sottos NR, White SR. Pressurized vascular systems for selfhealing materials. J R Soc Interface/R Soc 2012;9(70):1020e8. [96] Norris CJ, Meadway GJ, O'Sullivan MJ, Bond IP, Trask RS. Self-healing fibre reinforced composites via a bioinspired vasculature. Adv Funct Mater 2011;21(19):3624e33.

D.G. Bekas et al. / Composites Part B 87 (2016) 92e119 [97] Francesconi A, Giacomuzzo C, Grande AM, Mudric T, Zaccariotto M, Etemadi E, et al. Comparison of self-healing ionomer to aluminium-alloy bumpers for protecting spacecraft equipment from space debris impacts. Adv Space Res 2013;51(5):930e40. [98] Chowdhury RA, Hosur MV, Nuruddin M, Tcherbi-Narteh A, Kumar A, Boddu V, et al. Self-healing epoxy composites: preparation, characterization and healing performance. J Mater Res Technol 2015;4(1):33e43. [99] Kling S, Czig
any T. Damage detection and self-repair in hollow glass fiber fabric-reinforced epoxy composites via fiber filling. Compos Sci Technol 2014;99:82e8. [100] Haase T, Rohr I, Thoma K. Dynamic temperature measurements on a thermally activated self-healing ionomer. J Intell Mater Syst Struct 2012;25(1): 25e30. [101] International A. D 7136 e Standard test method for measuring the damage resistance of a fiber-reinforced polymer matrix composite to a drop-weight impact event. 2015. [102] International A. D 7137 e Standard test method for compressive residual strength properties of damaged polymer matrix composite plates. 2012. [103] Montemor MF. Functional and smart coatings for corrosion protection: a review of recent advances. Surf Coat Technol 2014;258:17e37.
rna K, Socha RP, Nowak P, [104] Kope
c M, Szczepanowicz K, Mordarski G, Podgo et al. Self-healing epoxy coatings loaded with inhibitor-containing polyelectrolyte nanocapsules. Prog Org Coat 2015;84:97e106. [105] Samadzadeh M, Boura SH, Peikari M, Ashrafi A, Kasiriha M. Tung oil: an autonomous repairing agent for self-healing epoxy coatings. Prog Org Coat 2011;70(4):383e7. [106] Choi H, Kim KY, Park JM. Encapsulation of aliphatic amines into nanoparticles for self-healing corrosion protection of steel sheets. Prog Org Coat 2013;76(10):1316e24. [107] Siva T, Kamaraj K, Sathiyanarayanan S. Epoxy curing by polyaniline (PANI) e characterization and self-healing evaluation. Prog Org Coat 2014;77(6): 1095e103. [108] Vimalanandan A, Lv LP, Tran TH, Landfester K, Crespy D, Rohwerder M. Redox-responsive self-healing for corrosion protection. Adv Mater 2013;25(48):6980e4. [109] Palleau E, Reece S, Desai SC, Smith ME, Dickey MD. Self-healing stretchable wires for reconfigurable circuit wiring and 3D microfluidics. Adv Mater 2013;25(11):1589e92. [110] Odom SA, Chayanupatkul S, Blaiszik BJ, Zhao O, Jackson AC, Braun PV, et al. A self-healing conductive ink. Adv Mater 2012;24(19):2578e81. 09. [111] Blaiszik BJ, Kramer SL, Grady ME, McIlroy DA, Moore JS, Sottos NR, et al. Autonomic restoration of electrical conductivity. Adv Mater 2012;24(3): 398e401. [112] Bailey BM, Leterrier Y, Garcia SJ, van der Zwaag S, Michaud V. Electrically conductive self-healing polymer composite coatings. Prog Org Coat 2015;85: 189e98. [113] Tripathi M, Rahamtullah, Kumar D, Rajagopal C, Roy Prasun K. Influence of microcapsule shell material on the mechanical behavior of epoxy composites for self-healing applications. J Appl Polym Sci 2014. http://dx.doi.org/ 10.1002/APP.40572. [114] Jin H, Mangun CL, Stradley DS, Moore JS, Sottos NR, White SR. Self-healing thermoset using encapsulated epoxy-amine healing chemistry. Polymer 2012;53(2):581e7. [115] Wang H-p, Hu S-q, Cai S-j, Yu F. Preparation and properties of bisphenol A epoxy resin microcapsules coated with melamineeformaldehyde resin. Polym Bull 2014;71(9):2407e19. [116] Scheltjens G, Brancart J, De Graeve I, Van Mele B, Terryn H, Van Assche G. Self-healing property characterization of reversible thermoset coatings. J Therm Anal Calorim 2011;105(3):805e9. [117] Scheltjens G, Diaz MM, Brancart J, Van Assche G, Van Mele B. A self-healing polymer network based on reversible covalent bonding. React Funct Polym 2013;73(2):413e20.

119

[118] Sharma M, Mondal D, Mukesh C, Prasad K. Self-healing guar gum and guar gum-multiwalled carbon nanotubes nanocomposite gels prepared in an ionic liquid. Carbohydr Polym 2013;98(1):1025e30. [119] Canadell J, Goossens H, Klumperman B. Self-healing materials based on disulfide links. Macromolecules 2011;44(8):2536e41. [120] Kim JT, Kim BK, Kim EY, Kwon SH, Jeong HM. Synthesis and properties of near IR induced self-healable polyurethane/graphene nanocomposites. Eur Polym J 2013;49(12):3889e96.
T, Telegdi J, Nyikos L. Linseed oil-filled microcapsules containing drier [121] Szabo and corrosion inhibitor e their effects on self-healing capability of paints. Prog Org Coat 2015;84:136e42. [122] Yuan L, Huang S, Hu Y, Zhang Y, Gu A, Liang G, et al. Poly(phenylene oxide) modified cyanate resin for self-healing. Polym Adv Technol 2014. http://dx.doi.org/10.1002/pat.3290. [123] Zhang M, Xu D, Yan X, Chen J, Dong S, Zheng B, et al. self healing supramolecular gels formed by crown ether based host guest interactions. Angew Chem 2012;124:5. [124] Li J, Zhang G, Deng L, Jiang K, Zhao S, Gao Y, et al. Thermally reversible and self healing novolac epoxy resins based on Diels Alder chemistry. 2015. [125] Meure S, Wu DY, Furman S. Polyethylene-co-methacrylic acid healing agents for mendable epoxy resins. Acta Mater 2009;57(14):4312e20. [126] Coope TS, Wass DF, Trask RS, Bond IP. Metal triflates as catalytic curing agents in self-healing fibre reinforced polymer composite materials. Macromol Mater Eng 2014;299(2):208e18. [127] Hodoroaba VD, Akcakayiran D, Grigoriev DO, Shchukin DG. Characterization of micro- and nanocapsules for self-healing anti-corrosion coatings by highresolution SEM with coupled transmission mode and EDX. Analyst 2014;139(8):2004e10. [128] Li W, Jiang Z, Yang Z, Zhao N, Yuan W. Self-healing efficiency of cementitious materials containing microcapsules filled with healing adhesive: mechanical restoration and healing process monitored by water absorption. PloS One 2013;8(11):e81616. [129] Dikic T, Ming W, van Benthem RA, Esteves AC, de With G. Self-replenishing surfaces. Adv Mater 2012;24(27):3701e4. [130] Faghihnejad A, Feldman KE, Yu J, Tirrell MV, Israelachvili JN, Hawker CJ, et al. Adhesion and surface interactions of a self-healing polymer with multiple hydrogen-bonding groups. Adv Funct Mater 2014;24(16):2322e33. [131] Roy S, Baral A, Banerjee A. An amino-acid-based self-healing hydrogel: modulation of the self-healing properties by incorporating carbon-based nanomaterials. Chemistry 2013;19(44):14950e7. [132] In C-W, Holland RB, Kim J-Y, Kurtis KE, Kahn LF, Jacobs LJ. Monitoring and evaluation of self-healing in concrete using diffuse ultrasound. NDT & E Int 2013;57:36e44. [133] Zedler L, Hager MD, Schubert US, Harrington MJ, Schmitt M, Popp J, et al. Monitoring the chemistry of self-healing by vibrational spectroscopy e current state and perspectives. Mater Today 2014;17(2):57e69. [134] Holten-Andersen N, Harrington MJ, Birkedalc H, Leed BP, Messersmith PB, Leea KYC, et al. pH induced metal ligand cross links inspired by mussel yield self healing polymer networks with near covalent elastic moduli. PNAS 2011;108(7):5. [135] Patrick JF, Hart KR, Krull BP, Diesendruck CE, Moore JS, White SR, et al. Continuous self-healing life cycle in vascularized structural composites. Adv Mater 2014. http://dx.doi.org/10.1002/adma.201400248. [136] Yuan YC, Ye Y, Rong MZ, Chen H, Wu J, Zhang MQ, et al. Self-healing of lowvelocity impact damage in glass fabric/epoxy composites using an epoxyemercaptan healing agent. Smart Mater Struct 2011;20(1):015024. [137] Chipara DM, Flores M, Perez A, Puente N, Lozano K, Chipara M. Adding autonomic healing. Adv Polym Technol 2012;00(0):9. [138] Ramachandran D, Liu F, Urban MW. Self-repairable copolymers that change color. RSC Adv 2012;2:9. [139] Hart KR, Sottos NR, White SR. Repeatable self-healing of an epoxy matrix using imidazole initiated polymerization. Polymer 2015;67:174e84.