Construction and Building Materials 63 (2014) 281–293

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Review

Iron-based shape memory alloys for civil engineering structures:An overview

http://dx.doi.org/10.1016/j.conbuildmat.2014.04.0320950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Address: Mateu Orfila Building, Ctra. Valldemossa km7.5, 07122 Palma, Spain. Tel.: +34 971 17 1378; fax: +34 971 17 3426.

E-mail address: antoni.cladera@uib.es (A. Cladera).

A. Cladera a,⇑, B. Weber b, C. Leinenbach b, C. Czaderski b, M. Shahverdi b, M. Motavalli b

a University of Balearic Islands, Department of Physics, Ctra. Valldemossa km 7.5, 07122 Palma, Spainb Empa, Swiss Federal Laboratories for Materials Science and Technology, Überlandstrasse 129, CH-8600 Dübendorf, Switzerland

h i g h l i g h t s

� The development of iron-based SMAs is presented, focusing on features for civil engineering.� Differences between the martensitic transformation in Ni–Ti and Fe–Mn–Si SMAs are highlighted.� High recovery stresses, which are necessary for prestressing, can be obtained for FeMnSi alloys.� Pilot experiences on the application of FeMnSi alloys are presented.� This paper collects unsolved aspects for future research.

a r t i c l e i n f o

Article history:Received 20 December 2013Received in revised form 12 March 2014Accepted 2 April 2014

Keywords:Iron-based shape memory alloyShape memory effectCivil engineeringMartensiteAusteniteStructural concreteRecovery stressPrestressingStrengtheningDamping

a b s t r a c t

Iron-based shape memory alloys (SMAs), especially Fe–Mn–Si alloys, are materials that have greatpotential in civil engineering structures, but their application is still in a pioneer stage. Recent develop-ments in alloy composition and manufacturing envisage new perspectives, especially in the field ofrepairing structures as well for new structures, when using these SMAs as prestressing tendons. Thispaper presents the fundamentals of the martensitic transformation from an engineering perspective aswell as some key properties, such as recovery stress, corrosion resistance, weldability and workability.Finally, some unsolved aspects are collected, and new perspectives for the use of these SMAs arepresented.

� 2014 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2822. The martensitic transformation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2833. Material properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

3.1. Recovery stresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2883.2. Corrosion resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2883.3. Weldability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2883.4. Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2893.5. Workability at room temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

4. Pilot experiences on the application of Fe–Mn–Si alloys. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

4.1. Shape memory effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2894.2. Damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290

Fte

282 A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293

5. Research needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2906. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

1. Introduction

Shape memory alloys (SMAs) are unique materials that have theability to achieve great deformations and to return to a predefinedshape after unloading or upon heating [1]. This shape memoryeffect is the result of the reversible phase transformation thatSMAs undergo, the so-called martensitic transformation. The mar-tensitic transformation can be produced by changes in tempera-ture or by the action of stresses. In the former case, themartensitic transformation takes place at defined temperatures(Fig. 1). The martensitic transformation, or forward transformation,is induced upon cooling the austenite phase (stable at high temper-atures), and consists of the appearance of the martensite phase(stable at low temperatures). In the absence of applied stresses,the temperature at which the process begins is known as Ms (mar-tensite start), whereas Mf (martensite finish) is the temperature atwhich the transformation finishes (Fig. 1). If the material is in mar-tensite (T < Mf), then the reverse transformation can be induced byheating the material. The formation of austenite will start at tem-perature As (austenite start) and will finish at temperature Af (aus-tenite finish). The transformation shows thermal hysteresis, inother words, the forward and reverse transformations do not takeplace at the same temperature [2].

The first discovery of a material with shape memory was docu-mented by Chang and Read, who observed a reversible phasetransformation in an Au–Cd alloy [1]. Buehler et al. discovered in1962 the shape memory effect in a Ni–Ti alloy [3], a fact that ledto the boom in international research in this field and the appear-ance of the first real applications of these alloys. Since then, differ-ent types of alloys with the shape memory effect have beendiscovered. Ni–Ti alloys, in some cases having a third component,are the ones that, to date, hold the first position in the industrialmarket. The main drawback of these alloys for their applicationin civil engineering structures is their cost. In 1982, the shapememory effect (SME) in an Fe–Mn–Si alloy was discovered [4],and since then, new iron-based SMAs with improved SME proper-ties have been developed. It is assumed that this progress will con-tribute to lowering the price of these materials and to making themmuch more competitive for civil engineering applications.

Janke et al. [5] presented possible applications of SMAs in civilengineering structures: passive vibration damping and energy dis-sipation, active vibration control, actuator applications and the uti-lization of the SME for tensioning applications or sensors. However,these authors stated that due to the size of the civil engineering

ig. 1. Schematic definition of the forward and reverse martensitic transformationmperatures.

structures and the actions of relatively high forces, low cost SMAswere needed.

The repairing of structures using the SME for prestressing ten-dons is a promising field for the use of iron-based SMAs, as wellas their use in new structures. If iron-based SMAs can be usedfor prestressing applications, they have several advantages com-pared to the traditional prestressing/posttensioning technologies,for example, there are no friction losses, no anchor heads or ductsare required, and no space is necessary for applying the force witha hydraulic device. The reason is that the prestressing of the SMAtendons is not performed mechanically, as in conventional pre-stressing steel, but with heating, as will be explained later in thispaper.

There are two different groups of iron-based SMAs [6]. The firstgroup contains alloys such as Fe–Pt, Fe–Pd and Fe–Ni–Co, whichexhibit the typical characteristics of thermoelastic martensitictransformations similar to Ni–Ti, with a narrow thermal hysteresis.However, in spite of extensive studies, no pseudoelasticity at roomtemperature has been reported with Fe–Pt or Fe–Pd alloys. In 2010,Tanaka et al. [7] presented an Fe–29Ni–18Co–5Al–8Ta–0.01B(mass %) SMA that shows a recovery strain of over 13% at roomtemperature and a very high tensile strength of 1200 MPa, placingit at the cutting edge of knowledge as far as new materials are con-cerned [8,9]. This iron-based SMA could be very useful for applica-tions that are related to pseudoelasticity and damping capacity.Additionally, good superelastic properties at room temperaturehave been found in an Fe–36Mn–8Al–8.6Ni (mass %) alloy [10],with a recovery strain of over 5% and a fracture tensile strain ofover 8%. However, these two new alloys still need further develop-ment to be able to produce them in large amounts for real-scaleelements in the construction industry, and the cost of the materialwould most likely be too high for construction standards becausethey should be cast in special conditions due to their composition.

The second group is a group of alloys such as Fe–Ni–C and Fe–Mn–Si, which have a larger thermal hysteresis in transformationbut still exhibit the SME. The Fe–Mn–Si SMAs have received con-siderable attention over the past two decades due to their low cost,good workability, good machinability and good weldability [11],although the real applications are still limited except for someremarkable exceptions, i.e., large size joining pipes for tunnel con-struction and crane rail joint bars [6].

The SME of the Fe–Mn–Si alloys is attributed to the stress-induced martensite transformation from a parent c-austenite (fcc– face-centered cubic) phase to an e-martensite phase (hcp – hex-agonal closed-packed) (Fig. 2) at low and intermediate tempera-ture and the reverse transformation (e- to c-phase) at hightemperature. In fact, it had long been known that Fe–Mn alloyscould undergo this transformation, but the desired SME had notbeen obtained [12]. The problem was that for high amounts ofMn, c-austenite was stabilized, making it difficult for the stress-induced martensitic transformation to occur. On the other hand,for lower amounts of Mn, when the alloy was subjected to stress,not only e-martensite but also a0-martensite (bct – body-centeredtetragonal) was generated (Fig. 2c). This phase is irreversible. Theoccurrence of a0-martensite induces dislocations markedly, pre-venting the SME from developing [12]. Sato and his co-workers dis-covered that the addition of Si allowed having the SME [4]. It wasalso observed that Cr, among other elements, was effective to aminor extent. This finding suggested that Cr was suitable as a

Fig. 2. Crystal lattices: (a) c-Austenite (fcc – face-centered cubic); (b) e-Martensite (hcp – hexagonal close-packed structure); (c) a0-Martensite (bct – body-centeredtetragonal).

Table 2Representative Fe–Mn–Si based SMAs developed in the second stage, all containingfine precipitates.

Composition in mass% Year References

Fe–28Mn–6Si–5Cr–0.5(Nb, C) 2001 [18,23]Fe–28Mn–6Si–5Cr–1(V, N) 2004 [24]Fe–14Mn–5Si–8Cr–4Ni–0.16C 2007 [25,26]Fe–17Mn–5Si–10Cr–4Ni–1(V, C) 2009 [27,28]Fe–16Mn–5Si–10Cr–4Ni–1(V, N) 2013 [29]

A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293 283

fourth element to be added to an Fe–Mn–Si alloy to improve itscorrosion resistance [12]. If the amount of Cr exceeds 7%, the brittler phase intrudes in the alloy, impeding the SME. It was found thatNi was the most effective component in restraining the formationof r phase and, therefore, Ni had to be added for increasingamounts of Cr [12]. The research at the materials level has beenprincipally focused on anti-corrosion, training effects, cyclic defor-mation and strengthening [13].

The development of these Fe–Mn–Si alloys can be summarizedby having two distinct stages. The most significant alloys of thefirst stage are summarized in Table 1. The main components thatdifferentiate them are the amount of Mn and Cr. With these alloys,an excellent recovery strain has been achieved by means of cyclicthermo-mechanical treatments, the so-called ‘‘training’’ [14]. Ingeneral, the recovery strain due to the SME in these alloys withouttraining is limited to only 1–2.5%, but the training improves therecovery strain, enabling very good shape recovery and enhancingthe recovery strain by up to 4% [14]. Other more simple heat treat-ments combined with different rolling conditions have beenrecently reported [15–17]. The second stage of the developmentof Fe–Mn–Si alloys started when Kajiwara and his co-workersobserved, in 2001, that the recovery strain and stress of iron-basedalloys could be improved, without any training, by introducing fineNbC precipitates into the microstructure [18]. Several alloys werelater developed that contained other precipitates, such as VN,Cr23C6 or VC, as seen in Table 2. In general terms, the key pointfor the improvement of the recovery strain is related to the forma-tion of a large elastic strain field in the vicinity of the precipitates,which provides a preferential nucleation site for the stress-inducedtransformation from the c-austenite to the e-martensite phase[18]. However, different opinions are still found in the technical lit-erature. Stanford and Dunne [19] propose that the improved recov-ery strain that is attributed to precipitation is likely to be the resultof the thermo-mechanical processing that is used in those experi-ments because they obtained the same shape recovery propertiesin precipitate-free alloys that underwent a similar thermo-mechanical treatment.

Table 1Representative Fe–Mn–Si based SMAs developed in the first stage. Adapted from [13].

Composition in mass% Year References

Fe–30Mn–1Si (single crystal) 1982 [4]Fe–30Mn–6Si (single crystal) 1984 [20]Fe–32Mn–6Si 1986 [21]Fe–28Mn–6Si–5Cr 1990 [22]Fe–20Mn–5Si–8Cr–5Ni 1990 [22]Fe–16Mn–5Si–12Cr–5Ni 1990 [22]

In this paper, the current state-of-the-art related to iron-basedshape memory alloys for their application in civil engineeringstructures will be presented, focusing on Fe–Mn–Si alloys becausethey are the most promising candidates for fast development in theconstruction industry. The paper will highlight the differencesbetween these alloys and other SMAs, i.e., Ni–Ti alloys, and it willcollect unsolved aspects that can be a starting point for futureresearch in this field. Moreover, based on the unique propertiesof iron-based SMAs, new perspectives for the use of these SMAswill be presented.

2. The martensitic transformation

The forward phase transformation in SMAs is a martensitictransformation. The martensitic transformation is, in general, a dif-fusionless solid state transformation in which the atoms move inan organized manner relative to their neighbors. This homoge-neous shearing of the parent phase creates a new crystal structurethat does not have any compositional change (no diffusion).Although the variation in the relative position of the atoms is verysmall, the coordinated movement of all of the atoms leads tochanges in the volume and can bring about significant macroscopicdeformations. The martensitic transformation can be induced by atemperature change, e.g., during quenching and/or by mechanicaldeformation.

Fig. 3a shows schematically the atomic arrangement in a hypo-thetic lattice after mechanically induced martensitic transforma-tion. The lattice is distorted without reorganization of the atoms(neighbors stay neighbors). This martensitic transformation issometimes called thermoelastic and is typical for Ni–Ti-alloys,and it usually shows a small hysteresis. In a plastic deformation,on the other hand, the atoms are rearranged by slip, as shown inFig. 3b. In this case, the neighbors are changed, but the latticestructure remains intact. This transformation can occur in a generalalloy or in an SMA and cannot be reversed by a temperature changebecause the crystal structure is identical to the original state.

The martensitic transformation in Fe–Mn–Si-based alloys is alsoachieved by slip, but in contrast to plastic transformation, the

Fig. 3. Different atomic behaviors of: (a) stress-induced martensitic transformationby distortion in Ni–Ti; (b) plastic deformation by slip. Adapted from [6].

Fig. 5. Shockley partial dislocation in transformation of fcc to hcp structure in Fe–Mn–Si-based alloys. Burgers vector shows the magnitude and direction of thelattice distortion of dislocation.

284 A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293

atoms are rearranged into a different lattice structure, namely fromface-centered cubic (fcc) to hexagonal close packed (hcp) (seeFig. 2). This transformation cannot be visualized in a simple 2D pic-ture and instead must be explained in a 3D setup. To this end, thefcc and hcp lattice structures are represented by close-packedspheres, whose centers correspond to the position of the atoms.The correspondence can easily be verified by looking at the cellsshown in Fig. 4 on the right. Alternatively, the two lattice struc-tures can be represented by three layers of close-packed spheresarranged in two different ways as shown in Fig. 4 on the left. Inthe first case, the third layer is arranged exactly above the firstlayer. This arrangement can be denoted as ABA and correspondsto the hexagonal close-packed lattice structure. In the second case,all of the three layers are in a different position, which is denotedby ABC. This arrangement corresponds to the face-centered cubiclattice when rotated in such a way that the center layer (B) coin-cides with a plane that intersects three diagonally opposite verticesin the fcc cell. The fcc structure can thus be transformed into thehcp structure by moving the top layer from the C-position to theA-position (Fig. 5).

The martensitic transformation from the c-austenite (fcc struc-ture) to the e-martensite (hpc structure) by translating one atomicplane as described above and shown in Fig. 5 is called partialShockley dislocation. Partial dislocations are movements of anatomic layer with less than one atomic distance, which thereforechange the lattice structure. For an Fe–Mn–Si SMA, this martensitictransformation can be reversed by a temperature change. The slip

Fig. 4. The hcp and fcc lattices built from three close-pack

that is involved in the transformation is responsible for the largehysteresis.

In an fcc structure, there are four possible slip planes. These caneasily be visualized by rotating the fcc cube in Fig. 4 around thevertical axis in steps of 90�. Each of them contains three sheardirections, which leads to twelve possible martensite variants inthe Fe–Mn–Si alloys. If only one variant is generated in a singlecrystal, then the corresponding mode is called a ‘monopartialstacking’ [11]. An SMA alloy is normally a polycrystalline materialconsisting of several crystals or grains. In such a material, the dif-ference in grain orientation as well as in the internal stress distri-bution will cause the conditions for monopartial stacking to not befulfilled, and several variants will be activated. At these variantintersections, other phases can be formed [30].

Experimentally, the generation and propagation of the partialdislocations in the fcc matrix has been enhanced by lowering thestacking fault energy (SFE) [31,32]. It has been demonstrated thatadding alloying elements such as Ti and Cu increase the SFE [33],whereas elements such as N, Ta, Ce or Sm lead to a decrease inthe SFE [33,34].

The mechanically induced martensitic transformation occursnot only in shape memory alloys; it is also a well-known transfor-mation in carbon steel, stainless steel and many other alloys. Forexample, high Mn steels (15–30% mass) are used in the automotiveindustry. Due to their high ductility with strains at failure up to80–90% and quite high ultimate strengths, they offer very goodconditions for the crash resistance of structural car body parts.The amount of Mn plays a significant role in the behavior of thesesteels [35]. Ultra high manganese steels exhibit the formation ofmechanical twins, specifically, stress-induced e- and a0-martensitein the c-austenite matrix (Fig. 2) under internal or externalstresses. These steels are commonly called TWIP (twin induced

ed layers of equal spheres in Fe–Mn–Si-based alloys.

A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293 285

plasticity) steels. The TWIP effect is dominant when the manga-nese content is equal to or higher than 25%. TWIP steels (i.e., Fe–25Mn–3Si–3Al) exhibit a total elongation of approximately 92%and an ultimate tensile strength of 650 MPa at room temperature[35]. For steel alloys that have lower amounts of manganese con-tent, approximately 15–20%, the alloy has a so-called TRIP (trans-formation induced plasticity) effect. TRIP Fe–Mn steels offer aslightly lower elongation than TWIP steels but a higher ultimatestrength. For example, an Fe–20Mn–3Si–3Al steel exhibited a totalelongation of approximately 82% and an ultimate tensile strengthof approximately 830 MPa at room temperature [35]. The TRIPeffect is due to the formation of a0-martensite (Fig. 2c), whichenhances the tensile elongation due to a retardation of the slipping.Ding et al. [36] showed that alloys that have an intermediateamount of Mn (23.8%) show simultaneously both effects. There-fore, the very high ultimate strains in the TRIP/TWIP steels aredue to the martensite transformation, but without SME, since thea0-martensite formation prevents the SME in the TRIP/TWIP steels.

Figs. 6 and 7 show some schematic phase diagrams (stress –temperature) and stress–strain curves for different thermome-chanical paths, for a Ni–Ti alloy in Fig. 6a–c and for an Fe–Mn–Sialloy in Fig. 6d–f. As discussed above, SMAs have two differentcrystalline phases: austenite and martensite. Austenite is com-posed of a highly symmetric cubic phase, with the existence of onlyone single possible structure in this case, whereas martensite has astructure that has a lower symmetry, which allows, in the case ofNi–Ti for example, up to 24 different configurations or variants.The assembly of martensitic variants can exist in two forms inalloys that have a thermoelastic martensite transformation, i.e.,Ni–Ti: twinned martensite, which is formed by a combination ofself-accommodated martensitic variants, and detwinned martens-ite, in which a specific variant is dominant. When an SMA with athermoelastic martensitic transformation is cooled below Mf inthe absence of an applied load, the crystal structure changes fromaustenite to martensite (forward transformation, see Fig. 1). Thearrangement of variants occurs such that the average macroscopic

(a) (b)

(d) (e)

Fig. 6. Schematic phase diagrams for Ni–Ti alloy (a–c) and Fe–Mn–Si alloy (d–f), showsubsequent heating to austenite under no deformation constraint for Ni–Ti; (c) pseudoela(e) stress-induced martensite and subsequent heating to austenite under no deformationthe Fe–Mn–Si alloy.

shape change is negligible, which results in twinned martensite(path 1 in Figs. 6a and 7). If the twinned martensite is deformeddue to external forces, its crystal structure changes to the variant,or variants, which enable(s) it to accommodate the maximumelongation and, as such, allow permanent deformations (detwin-ned martensite), as can be schematically seen in path 2 ofFigs. 6a and 7. In this step, the applied stress must be sufficientlylarge (rs) to start the detwinning process. The total detwinningof martensite is reached with the detwinning finish stress (rf).After unloading (path 3 in Figs. 6b and 7), the martensite remainsdetwinned. The change in the phase from detwinned martensite toaustenite can be brought about by raising the temperature (path 4in Figs. 6b and 7), and the SMA regains its cubic crystal structure,returning to its original shape if the deformations are uncon-strained or generating recovery stresses otherwise, as will be seenin detail later in this paper.

For an Fe–Mn–Si alloy, it is also possible to create thermal mar-tensite without an external apparent shape change (path 1 inFig. 6d). During path 1, the full austenite–martensite transforma-tion is not reached in an Fe–Mn–Si alloy, in contrast to alloys thatshow the typical thermoelastic martensitic transformation (Fig. 1).Moreover, the temperature-induced martensite variants cannot beeasily changed by applying a stress (monopartial stacking), and it isnot generally interesting for engineering applications to conductpath 2 from a temperature that is lower than Ms or Mf, as shownin Fig. 6d. The reorientation of different variants can hardly be acti-vated in Fe–Mn–Si alloys, due to the high level of the barrierenergy between the martensite variants, and plasticity, or the TRIPeffect, is activated before this type of energy level can be reached.

If the initial temperature of an Fe–Mn–Si alloy is higher than Ms

but not too close to As, then stress-induced martensite (path 2 inFig. 6e) is recoverable. However, the full austenite–martensitetransformation is still not reached for this stress-induced path.For an Fe–Mn–Si–Cr alloy, the percentage of stress-induced e-mar-tensite has been reported to be 30% or less [12] and approximately48% for an Fe–Mn–Si–Cr–Ni [38]. Plasticity (or irreversible slip) as

(c)

(f)

ing: (a) the detwinnning of Ni–Ti with an applied stress; (b) the unloading andstic loading path for Ni–Ti; (d) creation of thermal martensite in an Fe–Mn–Si alloy;constraint for Fe–Mn–Si alloy; and (f) plastic deformation with an irreversible slip in

Fig. 7. Typical stress–strain–temperature diagrams for Ni–Ti. For Fe–Mn–Si alloys, the pseudoelastic curve disappears, passing from the shape memory effect to an ordinaryplastic deformation. Adapted from [37].

286 A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293

well as the formation of a0-martensite can give rise to an incom-plete shape memory. At high temperatures, irreversible slip is eas-ier to activate and does not require a high mechanical loading level.Consequently, only plasticity occurs at high temperatures, while atlow temperatures, phase transformations occur first, followed byplasticity [39]. At moderate temperatures, both mechanisms arecoupled. It must be highlighted that the limits between the plastic-ity and the phase transformations related to the TRIP/TWIP effectsare still not clearly reported in the scientific literature for these Fe–Mn–Si alloys. In a recently published paper [40], the authorsreported a study on scanning electron microscopy and opticalmicroscopy images of an Fe–Mn–Si alloy in undeformed samplesand after a 4% tensile deformation at �45 �C and 100 �C. Whenthe strain was applied at �45 �C, the microstructure mostly con-sisted of the c- and e-phase. However, in the case of the samplethat was deformed at 100 �C, a significantly lower amount of thee-phase appeared, and on the other hand, a larger amount of thea0-phase was found. Because the formation of the a0-phase doesnot contribute to the SME in Fe–Mn–Si alloys, it is clear that thestrain-induced transformation must be kept at temperatures thatare not too high.

During unloading (path 3 in Fig. 6e), no reverse transformation(e-martensite to c-austenite) takes place in the region in whichboth phases are stable. This region extends over a broad tempera-ture range due to the very large hysteresis of these Fe–Mn–Sialloys, although some pseudoelastic effect has been reporteddepending on the temperature, which indicates that the reversetransformation takes place partially during unloading [40]. Heatingabove As (path 4 in Fig. 6e) will definitely activate the reversetransformation, which is the key point for the recovery strainand the generation of recovery stresses. When the temperaturereaches the austenite finish temperature Af, a part of the strain isrecovered (not all due to trapped martensite plates retained byvarious defects).

In Fig. 6, the slopes of the lines, or boundaries, that define thetransformation temperatures are known as ‘‘stress influence coef-ficients’’. It is typically assumed that each pair of lines for thetwo transformations shares a characteristic slope [41], althoughthis assumption is not necessarily the case for Fe–Mn–Si alloys.Moreover, in these alloys, the Ms boundary for stress-induced mar-tensite crosses the x-axis (zero-stress level) of the phase diagramfor a temperature below the measured Ms for thermally inducedmartensite. This property is an indication of the non-thermoelasticcharacter of its stress-induced martensitic transformation [42]. It

must be highlighted that the phase diagrams shown in Fig. 6 foran Fe–Mn–Si alloy will change considerably depending on thecomposition of the alloy and the thermomechanical treatmentsthat are conducted. Specific phase diagrams for the Fe–Mn–Sialloys can be found elsewhere [39,40,42,43].

Fig. 8 schematically shows the generation of recovery stressesfor three different alloys: Fig. 8a for a narrow hysteresis alloy(i.e. Ni–Ti), Fig. 8b a wide hysteresis alloy (i.e. Ni–Ti–Nb) andFig. 8c for an Fe–Mn–Si alloy. In the two first cases, the prestrainingfollows path 2 from Fig. 6a, starting with the material beingtwinned martensite and evolving with an almost horizontal pla-teau, in which the detwinning process takes place (Fig. 8a and b).

For Ni–Ti alloys, the recovery stress increases during heatingbut decreases during subsequent cooling (Fig. 8a). The recoverystress during heating may exceed the detwinning stress value(rs) because the energy (or stress) needed to detwin the martensiteis significantly lower than the energy (temperature) needed tocarry out the reverse transformation. However, when cooling, therecovery stress may drop down almost to zero because of forwardtransformation. To avoid any loss of recovery stress, the Af temper-ature is often chosen to be below the ambient temperature (situa-tion not represented in Figs. 8 or 9). In this case, the SMA istypically cooled down below the Mf temperature for prestrainingand stored at a temperature below As, by liquid nitrogen if needed.Throughout the service life, the SMA then remains in the high-tem-perature austenitic phase where it keeps the high recovery stress[44]. For Ni–Ti–Nb alloys (Fig. 8b), which have a larger hysteresis,the prestraining is still done at cooled conditions but storage canbe at ambient temperature, if it is below As [45].

For an Fe–Mn–Si alloy, prestraining (Fig. 8c) is typically done atambient temperature following path 2 in Fig. 6e. The transforma-tion takes place directly from austenite to martensite withouttwinning and detwinning. The unloading process for the Fe–Mn–Si alloys, a straight line in Fig. 8c, is sometimes curved, which dem-onstrates some pseudoelasticity [40,42].

Fig. 9 shows the evolution of the recovery stress in the phasediagrams during heating and cooling. During the heating step,the elastic stress in the SMA first decreases due to the suppressedthermal expansion. Once the stress-temperature path crosses theAs boundary, a reverse transformation to the austenite phaseoccurs. This transformation activates the SME, which generatestensile stresses in the SMA, because the contraction is inhibited.However, some of the stress is lost due to thermal expansion.The thermal expansion effect is recovered by thermal contraction

Fig. 8. Schematic stress–strain diagrams of prestraining and generation of recovery stresses during the activation of the reverse transformation in a constrained SMA: (a)Narrow hysteresis alloy, i.e., Ni–Ti; (b) Wide hysteresis alloy, i.e., Ni–Ti–Nb; (c) Fe–Mn–Si alloy.

Fig. 9. Schematic thermomechanical paths during the activation and cooling of a constrained SMA: (a) Narrow hysteresis alloy, i.e., Ni–Ti; (b) wide hysteresis alloy, i.e., Ni–Ti–Nb; (c) Fe–Mn–Si alloy.

A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293 287

during subsequent cooling [46]. The slope of the stress variationdue to pure thermal expansion/contraction varies depending onthe alloy. For example, for an Fe–Mn–Si alloy, considering an elas-tic modulus that is equal to 200 GPa and a coefficient of thermalexpansion (CTE) of 13 � 10�6/�C, this slope would be 2.6 MPa/�C.For a Ni–Ti alloy in the austenite phase, considering an elasticmodulus that is equal to 60 GPa and a CTE of 11 � 10�6/�C, theslope would be 0.66 MPa/�C, which is much lower than the slopefor the Fe–Mn–Si SMA, as seen in Fig. 9.

The behavior during cooling back to ambient temperature (AT)depends mainly on the width of the hysteresis. For a Ni–Ti SMAwith a narrow hysteresis, the recovery stress drops down due toforward transformation (Figs. 8a and 9a). For a wide hysteresisNi–Ti–Nb alloy, the austenite remains stable as long as the AT isabove the Ms boundary, maintaining a high recovery stress(Figs. 8b and 9b). For an Fe–Mn–Si alloy, the recovery stress dropsdown slightly (Figs. 8c and 9c) or even increases due to thermalcontraction. An alloy with AT between Ms and As and a narrow tem-perature hysteresis (Fig. 9a, i.e., Ni–Ti) is not appropriate for actingas permanent prestressing because the recovery stresses decreasesconsiderably after heating and cooling back to the AT. The recoverystresses would even become negligible if the martensite finishtemperature, Mf, was higher than the AT, a situation that is notshown in Fig. 9.

In the absence of heating or cooling, the SMA is at AT. Therefore,this temperature defines the phase in which the alloy should bestable in civil engineering. Janke et al. [5] proposed that, for exter-nal applications in structures, the ambient temperature can beassumed to be situated between �20 �C in winter and 60 �C underintense solar radiation in summer. For this reason, high hysteresisSMAs are required for structural applications.

Going back to SMAs with thermoelastic martensitic transforma-tions, i.e., NiTi, it is also possible to induce martensitic transforma-tion through the application of an external mechanical stress on anSMA that is in the austenite phase (T > Af), as seen in path 5 of

Figs. 6c and 7. In this case, only variants with an intrinsic changeof shape in the direction of the applied stress will appear (detwin-ned martensite). The stress–strain diagram would show pseudo-elasticity (or superelasticity when referring to engineeringapplications). This is made up of an initial elastic branch (path 5in Fig. 7), with the initial elastic modulus of the austenite, a hori-zontal branch, in which the phase transformation from austeniteto martensite is produced by mechanical induction, and anotherelastic branch, which has the initial modulus of the martensite.After unloading (path 6 in Figs. 6c and 7), the material will returnto the origin of the diagram without permanent deformations andby performing a hysteresis loop that dissipates energy, producing adamping effect. For higher initial temperatures (T > Md, where Md isthe threshold temperature above which martensite induced byexternal forces is not produced), the SMA would undergo an initialelastic deformation followed by a plastic slip deformation at highstresses, similar to the deformation that occurs in ordinary steel(path 7 in Fig. 7, effect not shown in Fig. 6 for Ni–Ti).

For Fe–Mn–Si alloys, it is not possible to have the superelastic-ity effect because when applying stresses at a temperature that ishigher than Af, the alloy will suffer plasticity with an irreversibleslip (path 7 in Fig. 6f). However, new findings reveal that someiron-based alloys, i.e., Fe–29Ni–18Co–5Al–8Ta–0.01B (mass %) orFe–36Mn–8Al–8.6Ni (mass %) alloys, show superelasticity at roomtemperature, although as was mentioned in the introduction sec-tion, these alloys still require further development to be producedat the scale that is needed for civil engineering applications, andtheir cost would probably be very high.

3. Material properties

The mechanical properties of each Fe–Mn–Si alloy will strongly depend on itscomposition, heat treatments and the hot and cold work that is performed. For thisreason, only some trends can be commented on in this section. Table 3 shows, as anexample, the fundamental properties of a solution treated Fe–28Mn–6Si–5Cr SMAafter hot working [6].

Table 3Fundamental properties of an Fe–28Mn–6Si–5Cr SMA. Adapted from [6].

Property Unit Values

Stress at 0.2% deformation MPa 200– 300Ultimate tensile strength MPa 680–1000Maximum strain % 16–30Hardness (HV) - 190–220Density (25 �C) kg/m3 7200–7500Melting point �C 1320–1350Thermal expansion (0–500 �C) �C�1 16.5 � 10�6

Thermal conductivity W/m �C 8.4Specific heat J/kg �C 540Specific resistance X cm 100 � 130 � 10�6

Young’s modulus GPa 170Shear modulus GPa 65Poisson ratio – 0.359Ms �C �20–+25Af �C 130–185Recovery strain – no training % 2.5Recovery strain after training % 4.5Recovery stress – no training MPa 130Recovery stress after training MPa 180Magnetic property – Paramagnetism

288 A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293

3.1. Recovery stresses

The key factor for the use of iron-based SMAs as prestressing tendons in civilengineering structures is the recovery stress. As seen in Table 3, shape recoveryproperties for this Fe–28Mn–6Si–5Cr SMA increase significantly after thermome-chanical training. However, as different authors note, this treatment requires addi-tional processing steps, which increases the production cost and is applicable onlyto components that have simple geometries [28,29].

The modulus of elasticity of the Fe–28Mn–6Si–5Cr alloy considered in Table 3 is170 GPa, which is much higher than that reported for Ni–Ti alloys, which variesfrom 30 to 98 GPa in austenite to 21–52 GPa in martensite [47]. Other authors havereported a modulus of elasticity in the range between 160 and 200 GPa for an Fe–17Mn–5Si–10Cr–4Ni–1(V, C) alloy [28,40].

Most of the research that has been conducted to date with Fe–Mn–Si alloys hasbeen focused on studying the SME and on having the maximum recovery strain,paying secondary attention to the recovery stresses. However, for prestressing incivil engineering applications, the recovery stress is a fundamental property. Thealloy with the maximum recovery strain does not enjoy the maximum recoverystress, as was observed for an Fe–Mn–Si–Cr–Ni alloy with varying amounts of C[48]. The carbon in this alloy increases the plastic yield strength and thus favorsthe production of stress-induced martensite during prestraining. A moderate addi-tion of carbon (0.12%) can thus improve the recovery strain. However, an excessiveaddition of carbon (0.18%) will shift the Ms temperature much below room temper-ature and thus diminish the amount of martensite that is produced during pre-straining, which leads to a lower recovery strain. For the recovery stress, on theother hand, a low Ms temperature will prevent the austenite that was produced dur-ing heating from transforming back to martensite during cooling. The recoverystress in the Fe–Mn–Si–Cr–Ni alloys therefore increases with increasing carboncontent [48].

The recovery stresses are not only dependent on the alloy composition but alsoon the microstructural features such as the grain size or the presence, size anddistribution of fine second-phase particles. These features have an influence onthe mechanical properties of the alloys on the one hand, but also on the stackingfault energy or the transformation temperatures on the other hand. They are

Table 4Recovery stresses and transformation temperatures reported for different Fe–Mn–Si alloy

Composition in mass% Comments

Fe–28Mn–6Si–5Cr Without training – 5–8% pre-strainWith training – 5–8% pre-strain

Fe–28Mn–6Si–5Cr–0.5(Nb, C) Non pre-rolled6% pre-rolled14% pre-rolled70% pre-rolled

Fe–19Mn–5Si–8Cr–5Ni Equal channel angular pressing and 4.5% pre-straFe–16Mn–5Si–10Cr–4Ni–1(V, N) 4% pre-strain at �45 �C

4% pre-strain at RTFe–15Mn–4Si–8Cr–4Ni–0.012C Cold-drawn alloy – 6% pre-strain at RT – AnnealinFe–15Mn–4Si–8Cr–4Ni–0.12C Cold-drawn alloy – 8% pre-strain at RT – AnnealinFe–15Mn–4Si–8Cr–4Ni–0.18C Cold-drawn alloy – 4% pre-strain at RT – AnnealinFe–17Mn–5Si–10Cr–4Ni–1(V, C) 4% pre-strain at RT

influenced by the thermo-mechanical pre-treatment of the alloy. Table 4 presentsthe recovery stresses that are reported for different Fe–Mn–Si alloy compositionsand different thermomechanical procedures. For an Fe–28Mn–6Si–5Cr SMA, therecovery stresses vary from approximately 130 MPa to 180 MPa, without and witha cycle of thermomechanical training. However, the recovery stresses are substan-tially higher for alloys that contain small VN or VC precipitates, up to values of 500and 580 MPa, respectively. Values of 460 MPa and 565 MPa were also reported foralloys that are deformed by equal channel angular pressing (a thermomechanicalprocedure that has been successfully applied to produce various ultra-fine grainedmaterials) and subsequent annealing as well as cold rolling and annealing [49,50].

No data on the long-time relaxation of the recovery stresses is reported in theliterature, although good creep and stress relaxation characteristics for high stresslevels have been reported for Fe–Mn–Si SMAs [51], the relaxation being lower forthe alloys that contain VC precipitates compared to that of alloys without such pre-cipitates [52].

For structural concrete applications, it is also very important to consider thetemperature at which the transformation occurs because high temperatures insidethe concrete could cause micro-cracking, deteriorating the mechanical properties ofthe surrounding concrete. In this sense, the reported Fe–17Mn–5Si–10Cr–4Ni–1(V,C) [28] reached the recovery stresses of 580 MPa after heating to only 130 �C(Table 4). The heating temperature is limited in internal prestressing applicationsby the bond strength reduction due to the increased temperature in the surround-ing concrete. Pull-out tests with ribbed reinforcing bars reported in the literaturesuggest a reduction in the bond strength of 10–20% for a concrete temperature ofup to 200 �C [53–55], although this reduction varies depending on the concretecompressive strength and the ratio between the concrete cover depth and the rebardiameter [56].

3.2. Corrosion resistance

The corrosion resistance of Fe–Mn–Si SMAs has been studied for different alloycompositions in aggressive environments such as H2SO4 and NaCl solutions [57–66]. To the authors’ best knowledge, the corrosion characteristics under the alkalineenvironment of concrete (pH of 12–13) that surrounds embedded bars have not yetbeen reported.

In highly oxidizing environments, i.e., H2SO4, Fe–Mn–Si–Cr–Ni–(Co) alloys thatcontain from 8.8 wt% to 12.80 wt% of Cr exhibit a similar or better corrosion resis-tance than stainless steel 304 in spite of the lower amount of Cr. This behavior isattributed to the high amount of Si in the iron-based SMAs [63,65,66].

The corrosion behavior of the Fe–Mn–Si alloys in NaCl solution is a key pointthat is still unclear, with some contradictory results in the available literature[66]. On the one hand, Fe–15Mn–7Si–9Cr–5Ni did not show passivation or localizedpitting behavior in a 3.5% NaCl solution and showed the best behavior when testedas single-phase austenite [60]. However, the training process or the cold rollingreduced the corrosion resistance of the alloys [57]. On the other hand, a recentstudy reports that Fe–Mn–Si–Cr–Ni–(Co) alloys that contain from 8.8 wt% to12.80 wt% of Cr show poor corrosion resistance in a 3.5%NaCl solution comparedto that of stainless steel 304 and that these SMAs are prone to pitting corrosionin chloride environments due to their high manganese content [66]. In any case,the higher the Cr content is, the higher the corrosion resistance [58]. The additionof Cu or Rare Earths, mainly La and a small quantity of Ce, could improve the cor-rosion resistance in NaCl solutions [59,61].

3.3. Weldability

The welding characteristics of Fe–Mn–Si alloys have been studied while consid-ering different welding technologies, such as tugsten-inert gas (TIG), laser beam(LB) and electron beam welding (EB) [67–71]. Although some researchers reportonly the welding optimization parameters, only a few of them focus on shaperecovery strain or stress after welding. The methodology of the tests was, in generalterms, as follows: welding two small pieces of a specified Fe–Mn–Si alloy in the

s.

Recovery stress (MPa) Temperature (�C) Reference

130 350 [6]180 350 [6]145 400 [23]255 400 [23]295 400 [23]200 400 [23]

in at RT 460 500 [49]500 225 [29]440 160 [29]

g temperature 650 �C 520 – [50]g temperature 650 �C 535 – [50]g temperature 750 �C 565 – [50]

580 130 [28]

A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293 289

c-austenite phase, and after that, predeforming the sample and obtaining the shaperecovery strain or stress. After the welding, the phase structure of the weld zone isstill austenite, which is the same as that of the base alloy but includes dendrite crys-tals. There is no obvious change in the microstructure of the heat-affected zoneexcept for some degree of growth of austenite grains. In general terms, the variationin the shape recovery strain due to welding is within ±10% for the small testedpieces [67,68]; as a result, insignificant changes would be produced for a completeelement in which the heat-affected zone is significantly small compared to theoverall dimension of the element. However, the corrosion resistance of the weldedzones is worse than that of the base materials, due to the micro-segregation andwelding stress that occurs within the welded zones. The recovery stresses thatare generated in small 10 mm long samples, including the weld zone of an Fe–28Mn–6Si–5Cr–0.53Nb–0.06C alloy, were 250 MPa, independent of the weldingtechnique used (TIG, LB or EB), compared to the 300–330 MPa reported for thenon-welded material [68]. The dendrite structures that formed during the weldingprocedure disappeared after an annealing treatment of the welded pieces, improv-ing the corrosion resistance. Finally, a manufacturing process for shaft and pipecoupling by forming and welding of an Fe–15Mn–5Si–9Cr–5Ni shape memory alloywas recently proposed [72]. The results showed that welding affects mechanicaland shape memory properties: the material fractured in the welded zone and thedegree of shape recovery decreased 15%. However, the couplings manufacturedwith this new procedure could develop sufficient coupling force and this manufac-turing procedure should be more studied [72].

3.4. Production

In general, Fe–Mn–Si based SMAs are produced by melting and subsequentcasting under high vacuum (or in air) followed by thermo-mechanical processing[73]. The composition of Fe–Mn–Si alloys is very similar to a high-Mn stainlesssteel. For that reason, some authors think that these alloys could probably be con-veniently melted and processed by utilizing the production facilities of steels orstainless steels [6]. However, a mass production electric furnace cannot be adoptedfor the melting of any SMA that contains a large amount of Mn elements and thathas a high heat capacitance [6]. As Maruyama and Kubo remark in the case of anSMA that contains over 20% Mn, the impurities that are contained in the raw Mnmaterials give rise to fatal faults in the mechanical properties of the SMA. For thisreason, decreasing the amount of Mn would be better for a mass production furnaceand would decrease the total cost reasonably [6]. In fact, the alloy Fe–17Mn–5Si–10Cr–4Ni–1(V, C), which has a relatively low Mn content, has already been pro-duced under normal atmospheric conditions [28,40]. A recent research has high-lighted that Fe–Mn–Si alloys with the Mn content above 13 mass% exhibit a goodshape recovery strain of more than 2% with Ni and Cr substitutes, while the alloyswith the Mn content below 13 mass% indicate a sudden decrease in the shaperecovery strain with all the substitutes [74].

An alternative production methodology for the Fe–Mn–Si SMAs by mechanicalalloying and subsequent sintering has been explored [73]. This mechanical alloyingis a synthetic technique involving solid state reaction among powders due to highenergy collisions. Powder metallurgy may offer several advantages for manufactur-ing industrial products, as it enables the production of alloys in near net shape, min-imizing the additional machining required to form the final product. The resultsshowed that the mechanical properties and the shape memory effect could be com-parable to that alloys manufactured by conventional casting [73].

3.5. Workability at room temperature

Although many authors refer to the good workability of Fe–Mn–Si SMAs, notmuch specific research has been conducted on this topic. As Sato et al. [13] note,it is fortunate that these Fe–Mn–Si SMAs are quite similar to the TRIP/TWIP steels,which have good workability at room temperature. However, it is sometimes nec-essary to prevent the SMA from breaking when heavy deformation is needed.Superposition of either a hydrostatic pressure or a proper constraint is effective[13]. Finally, it must be taken into account that the addition of C or N could producecarbide and nitride, respectively, which reduce workability [22].

4. Pilot experiences on the application of Fe–Mn–Si alloys

The application of iron-based SMAs in civil engineering struc-tures is still in a pioneering stage, and only a few pilot applicationscan be found in the literature. Nevertheless, two applications havesucceeded in other related industries: the production of crane railfishplates [75] to connect finite lengths of rail for heavy dutycranes (i.e., steel factories) and pipe joints for steel pipes for a con-struction method for tunneling work [6], with both products man-ufactured by Awaji Materia Co. Application of other SMAs in civilengineering structures, i.e., Ni–Ti or Cu-based alloys, can be foundelsewhere [5,47,76–86].

4.1. Shape memory effect

A bridge in Michigan was strengthened through external post-tensioning that was performed with Fe–Mn–Si–Cr SMA rebars in2001 [87], after a laboratory research study on the possibilities ofexternal reinforcement of shear cracks. In this case, an Fe–28Mn–6Si–5Cr alloy was used that could develop recovery stresses of255 MPa after heating it to 300 �C. The alloy was manufacturedby Nippon Steel Corporation. The effect of extreme environmentaltemperatures on the restrained recovery stresses of the selectediron-based alloy was studied, and the conclusion was that the vari-ations in the stress were relatively small. The study showed thatpost-tensioning with SMAs enabled the shear crack to be closedto a large extent, recovering the load capacity of the rehabilitatedbeam.

Watanabe et al. [88] worked on the reinforcement of an 80 mmlong plaster prism specimen with a 1 mm diameter prestressedwire of Fe–27Mn–6Si–5Cr–0.05C alloy. The wires were subjectedto pretensile strain at room temperature (1%, 2% and 3%), and theywere embedded into a plaster matrix. The specimens were thenheated up to 250 �C to generate a compressive stress in the matrix.Three-point bending tests were performed for mechanical propertycharacterization as well as pull-out tests. These tests showed thatthe fracture toughness of plaster could be improved by the wiresand that the bending strength of the composite specimen increasedwith increasing levels of prestrain in the SMA wires.

Small mortar prisms were reinforced with square Fe–28Mn–6Si–5Cr–1(NbC) bars by Sawaguchi et al. [89] (Fig. 10). As was pre-viously mentioned in this paper, Kajiwara et al. had reported thatthe shape memory properties of Fe–Mn–Si based SMAs could beimproved by the fine dispersion of NbC carbides [18]. The mortarspecimens, with embedded SMA square bars (2 � 2 mm squaresection), or with steel bars with the same dimensions, or withoutany reinforcement, were first cured for two days. The hardenedmortar prisms were extracted from the mold, and a high-temper-ature curing was performed in an autoclave. The first stage of theautoclave curing was performed at 87 �C (360 K) for 24 h, toincrease the compressive strength of the mortar matrix. The sec-ond stage of the curing was intended to generate the prestress inthe SMA, for which two conditions, i.e., 177 �C for 6 h and 247 �Cfor 30 min, were applied. The mechanical properties of the mortarincreased with the 177 �C curing, but they worsened for the 247 �Ctreatment, except for the surprising results they got from prismsstrengthened with steel subjected to the high temperature curing(Fig. 10). The authors concluded that the iron-based SMAs wereusable for producing prestress in the mortar because the bendingstrength increased significantly in the specimens with the SMArespect to the specimens reinforced with steel or the non-rein-forced specimens. Sawaguchi et al. also highlighted that furtherstrengthening could be achieved by lowering the reverse transfor-mation temperatures of the SMAs, thus avoiding significant ther-mal damage to the mortar matrix.

Watanabe et al. [90] investigated the fabrication method andmechanical properties of smart composites that were reinforcedwith Fe–28Mn–6Si–5Cr SMA machining chips. These chips are aby-product of the fabrication of one of the main industrial applica-tions of iron-based shape memory alloys: the pipe joining of steelpipes for tunneling constructions. The study was also conducted ata feasibility level, using 80 mm long plaster prims and obtaininggood preliminary results.

Lee et al. [46] studied the recovery stress behavior of anFe–17Mn–5Si–10Cr–4Ni–1(V, C) shape memory alloy forprestressing concrete structures. The prestressing effect due tothe shape memory effect was simulated by a series of tests thatinvolved pre-straining the material followed by heating andcooling at a constant strain. The tested SMA specimens came from

Fig. 10. Specimens tested by Sawaguchi et al. [89] and bending strength of the mortar specimens with SMA, steel and without reinforcement.

290 A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293

a 15 kg alloy ingot that had been induction melted under normalatmospheric conditions. The iron shape memory alloy used in thisresearch is very promising because it shows a very wide hysteresisfrom �60 �C to 103 �C that is determined from a differential scan-ning calorimetry (DSC) analyses. The authors conducted tests fordifferent heating temperatures and initial pre-strains, obtainingthe highest recovery stress at 400 MPa, for a pre-strain of 4% anda heating temperature of 160 �C. In this study, the behavior underadditional loading after prestressing was also studied, both undercyclic mechanical loading and under cyclic thermal loading. Thesetests showed that the prestressed SMA undergoes an inelasticdeformation during the first cycle, which leads to a reduction inthe recovery stresses. However, after the first cycle, the SMAbehaved elastically, and the level of stress remained stable on sub-sequent cycles. Finally, the authors experimentally showed thatthis loss could be fully recovered by reheating the SMA element.

The same alloy but from a much larger cast has recently beenused to manufacture 20 mm wide and 750–950 mm long ribbedFe-SMA strips for near-surface mounted reinforcements [91]. Therecovery stress after heating to 160 �C and cooling down to roomtemperature measured in a tensile machine with climate chamberwas in the range of 250–300 MPa. The bond behavior tested in lap-shear experiments was in the usual range of reinforcing steel. Thestrips were embedded in 700 mm long mortar prisms of35 � 50 mm2 cross-section and heated by a current supply withapproximately 14 A/mm2 and a voltage of approximately 15 V/mduring less than 7 s. Strain measurements on one of the prism indi-cated a compression stress in the concrete of more than 3 MPa.

4.2. Damping

As far as the authors know, no real application has beenreported to date that is related to damping in civil engineeringstructures when using iron-based shape memory alloys. However,the large damping effect in the TRIP/TWIP Fe–Mn alloys due to themartensitic transformation has been studied in depth [92–95].More recently, based on very similar principles, studies on thevibration mitigation by the martensitic transformation in Fe-basedalloys for their application as a seismic damping material has alsobeen reported [96,97]. The Fe–28Mn–6Si–5Cr–05NbC SMAs thatwas studied by Sawaguchi et al. [96,97] showed a significantdamping capacity in the large-strain amplitude region above0.1%. The specific damping capacity increased with increasingstrain amplitude and reached a saturation value of approximately80% above a strain amplitude of 0.4% (Fig. 11). Their analysisrevealed that the tension-induced martensite reversely transformsinto austenite by compression. The authors thus concluded that

the vibration mitigation during the cyclic tension–compression isaccomplished by the reversible motion of the c/e interfaces.

The hypothetical use of the superelastic Fe–Ni–Co–Al–Ta–Balloy developed by Tanaka et al. [7] in a wire-based smart naturalrubber bearing was recently investigated [98]. The authors con-ducted a study that was based on finite element modeling of differ-ent configurations of bearings. They concluded that the ferroussuperelastic SMA is a good candidate due to its high superelasticstrain range and very low austenite finish temperature (Af).

5. Research needs

A large number of researches that study the behavior of Fe–Mn–Si alloys have been performed, but some aspects are still not clearlyunderstood, and more research is needed on specific topics. Forexample, most of the research on thermomechanical treatmentsduring the production process has focused on improving the recov-ery strain, whereas the recovery stresses are the true key point forthe application of the SMAs as prestressing reinforcements. There-fore, a systematic study on the optimization of the recovery stres-ses for different alloy compositions is needed. Additionally, withrespect to the material properties, relaxation and fatigue proper-ties have not been studied in-depth. More information can befound on the corrosion behavior, although for prestressing applica-tions, it would be necessary to know the corrosion characteristicsunder the alkaline environment of concrete. Large-scale produc-tion also needs research to allow having the large amounts ofmaterials that are needed in civil engineering applications. Forthe development of some products or devices, it is also necessaryto improve the knowledge on weldability. Different papers addressthe weldability properties of the iron-based SMAs in the austenitephase, but it would be very useful to know the temperature-affected zone in a welded alloy in martensite for different weldingtechnologies.

At the application level, most of the research that was con-ducted during recent years for Ni–Ti alloys exploiting the SMEand the damping capacity can be adapted for iron-based shapememory alloys. In this paper, some pilot experiences on the appli-cation of Fe–Mn–Si alloys have been highlighted, but many otherscould be conducted in the near future. For example, it is possible toconsider concrete that is prestressed by Fe–Mn–Si alloy shortfibers, which has been already proposed for Ni–Ti alloys [99,100],or on the confinement of concrete columns, which has beenbroadly tested with Ni–Ti or Ni–Ti–Nb alloys [101–103]. For manyconcrete applications, it would be necessary to have more informa-tion about the bond strength reduction due to the temperatureincrease of the embedded SMA. In fact, the tests that were reported

(a) (b)

Fig. 11. (a) Stress–strain hysteresis loops that were tested at varying strains: ±0.1%, ±0.2%, ±0.4%, ±0.6%, ±0.8%, ±1.0%. (b) Specific damping capacity (SDC) as a function of thestrain amplitudes. Adapted from [97].

A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293 291

in the literature and commented on in this paper address the heat-ing of the concrete and not the direct heating of the rebars.

6. Conclusions

Research on new iron-based SMAs has experienced a consider-able boost during the past decade, developing Fe–Mn–Si alloys thatcan generate high recovery stresses at lower temperatures com-pared to the first alloys developed in the 1980s. The current stateof the art related to these SMAs for their application in civil engi-neering structures has been presented in this paper. The followingconclusions can be drawn:

� The use of iron-based SMA tendons for the repairing ofexisting structures or the reinforcing of new structures isa promising field for the near future. Such iron-basedSMA tendons have several advantages, such as no frictionlosses, no anchor heads and ducts required, and no spacenecessary for applying the force with a hydraulic device.

� The lower cost of these new iron-based SMAs results notonly from the use of cheap iron (Fe) but also from the pos-sibility of melting and producing these SMAs under normalatmospheric conditions.

� These new Fe–Mn–Si alloys have a wider temperaturetransformation hysteresis and a higher elastic stiffness thanother SMAs, i.e., Ni–Ti alloys. They also present a goodworkability, corrosion resistance and weldability.

� New Fe–Mn–Si alloys with fine precipitates have beendeveloped during recent years and allow high recoverystresses without the need for thermomechanical training.

� More research is still needed to allow a better understand-ing of the material behavior. The key topics have been sum-marized in this paper, and they constitute a starting pointfor future research in this field.

Acknowledgments

This paper was written during the first author’s stay at Empa,Swiss Federal Laboratories for Material Science and Technology,and was funded by the Spanish Ministry of Education, Cultureand Sports through the National Human Resources Mobility Pro-gram of the R&D National Plan (project PRX12/00187). Moreover,it has been developed in the framework of project BIA2012-31432, which was co-funded by the Spanish Ministry of Economyand Competitiveness (MINECO) and the European Regional Devel-opment Fund (ERDF).

Furthermore, support from the Swiss Commission for Technol-ogy and Innovation (CTI Project 14496.1 PFIW-IW) and from thecompany re-Fer is highly appreciated.

References

[1] Otsuka K, Wayman CM. Shape memory materials. UnitedKingdom: Cambridge University Press; 1998.

[2] Kumar PK, Lagoudas DC. In: Lagoudas DC, editor. Shape memory alloys.modeling and engineering applications. Springer; 2008. p. 1–51 [Chapter 1].

[3] Buehler WJ, Gilfrich JV, Wiley RC. Effect of low-temperature phase changes onthe mechanical properties of alloys near composition TiNi. J Appl Phys1963;34:1475–7.

[4] Sato A, Chishima E, Soma K, Mori T. Shape memory effect in c¢etransformation in Fe–30Mn–1Si alloy single crystals. Acta Metall1982;30:1177–83.

[5] Janke L, Czaderski C, Motavalli M, Ruth J. Applications of shape memory alloysin civil engineering structures – overview, limits and new ideas. Mater Struct2005;38:578–92.

[6] Maruyama T, Kubo H. In: Yamauchi K, Ohkata I, Tsuchiya K, Miyazaki S,editors. Shape memory and superelastic alloys: technologies andapplications. Cambridge (UK): Woodhead Publishing Limited; 2011. p. 141–59 [Chapter 12].

[7] Tanaka Y, Himuro Y, Kainuma R, Sutou Y, Omori T, Ishida K. Ferrouspolycrystalline shape-memory alloy showing huge superelasticity. Science2010;327:1488–90.

[8] Ma J, Karaman I. Expanding the repertoire of shape memory alloys. Science2010;327:1468–9.

[9] Ma J, Hornbuckle BC, Karaman I, Thompson GB, Luo ZP, Chumlyakov YI. Theeffect of nanoprecipitates on the superelastic properties of FeNiCoAlTa shapememory alloy single crystals. Acta Mater 2013;61:3445–55.

[10] Omori T, Ando K, Okano M, Xu X, Tanaka Y, Ohnuma I, et al. Superelastic effectin polycrystalline ferrous alloys. Science 2011;333:68–71.

[11] Kajiwara S. Characteristic features of shape memory effect and relatedtransformation behavior in Fe-based alloys. Mater Sci Eng A 1999;273–275:67–88.

[12] Otsuka H. Fe–Mn–Si based shape memory alloys. Mat Res Soc Symp Proc1992;246:309–20.

[13] Sato A, Kubo H, Maruyama T. Mechanical properties of Fe–Mn–Si based SMAand the application. Mater Trans 2006;47:571–9.

[14] Wang D, Chen Y, Gong F, Liu D, Liu W. The effect of thermomechanicaltraining on the microstructures of Fe–Mn–Si shape memory alloy. Le, J PhysIV 1995;5:C8,527-C8-530.

[15] Stanford N, Dunne DP. Thermo-mechanical processing and the shape memoryeffect in an Fe–Mn–Si-based shape memory alloy. Mater Sci Eng A2006;422:352–9.

[16] Baruj A, Troiani HE. The effect of pre-rolling Fe–Mn–Si-based shape memoryalloys: mechanical properties and transmission electron microcopyexamination. Mater Sci Eng A 2008;481–482:574–7.

[17] Druker A, Perotti A, Baruj A, Malarría J. Heat treatments of Fe–Mn–Si basedalloys: mechanical properties and related shape memory phenomena. ASTMSpec Tech Publ 2012;1532 STP:287–300.

[18] Kajiwara S, Liu D, Kikuchi T, Shinya N. Remarkable improvement of shapememory effect in Fe–Mn–Si based shape memory alloys by producing NbCprecipitates. Scripta Mater 2001;44:2809–14.

[19] Stanford N, Dunne DP, Li H. Re-examination of the effect of NbC precipitationon shape memory in Fe–Mn–Si-based alloys. Scripta Mater 2008;58:583–6.

[20] Sato A, Chishima E, Yamaji Y, Mori T. Orientation and compositiondependencies of shape memory effect IN Fe–Mn–Si alloys. Acta Metall1984;32:539–47.

292 A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293

[21] Murakami M, Otsuka H, Suzuki H, Matsuda S. Complete shape memory effectin polycrystalline Fe–Mn–Si alloys. Proc Int Conf on Mart Transf 1986:985–90.

[22] Otsuka H, Yamada H, Maruyama T, Tanahashi H, Matsuda S, Murakami M.Effects of alloying additions on Fe–Mn–Si shape memory alloys. ISIJ Int1990;30:674–9.

[23] Baruj A, Kikuchi T, Kajiwara S, Shinya N. Effect of pre-deformation ofaustenite on shape memory properties in Fe–Mn–Si-based alloys containingNb and C. Mater Trans 2002;43:585–8.

[24] Farjami S, Hiraga K, Kubo H. Shape memory effect and cristallographicinvestigation in VN containing Fe–Mn–Si–Cr alloys. Mater Trans2004;45:930–5.

[25] Wei Z, Yuhua W, Ning L, Wenling X, Shanhua W. Directional precipitation ofcarbides induced by c/e interfaces in an FeMnSiCrNiC alloy aged afterdeformation at different temperature. Mater Sci Eng A 2007;459:324–9.

[26] Wen YH, Xiong LR, Li N, Zhang W. Remarkable improvement of shapememory effect in an Fe–Mn–Si–Cr–Ni–C alloy through controllingprecipitation direction of Cr23C6. Mater Sci Eng A 2008;474:60–3.

[27] Dong Z, Klotz UE, Leinenbach C, Bergamini A, Czaderski C, Motavalli M. Anovel Fe–Mn–Si shape memory alloy with improved shape recoveryproperties by VC precipitation. Adv Eng Mater 2009;11:40–4.

[28] Leinenbach C, Kramer H, Bernhard C, Eifler D. Thermo-mechanical propertiesof an Fe–Mn–Si–Cr–Ni–VC shape memory alloy with low transformationtemperature. Adv Eng Mater 2012;14:62–7.

[29] Li K, Dong Z, Liu Y, Zhang L. A newly developed Fe-based shape memory alloysuitable for smart civil engineering. Smart Mater Struct 2013:22.

[30] Dong ZZ, Kajiwara S, Kikuchi T, Sawaguchi T. Effect of pre-deformation atroom temperature on shape memory properties of stainless type Fe–15Mn–5Si–9Cr–5Ni–(0.5–1.5)NbC alloys. Acta Mater 2005;53:4009–18.

[31] Xuejun J, Jihua Z, Hsu TY. Thermodynamic calculation of stacking fault energyin Fe–Mn–Si shape memory alloys. Mater Des 2000;21:537–9.

[32] Guo Z, Rong Y, Chen S, Hsu TY. Crystallography of fcc(c) ? hcp(e) martensitictransformation in Fe–Mn–Si based alloys. Scripta Mater 1999;41:153–8.

[33] Li CL, Cheng DJ, Jin ZH. Influence of deformation temperature on shapememory effect of Fe–Mn–Si–Ni–Cr alloy. Mater Sci Eng A 2002;325:375–9.

[34] Yang CH, Lin HC, Lin KM. Improvement of shape memory effect in Fe–Mn–Sialloy by slight tantalum addition. Mater Sci Eng A 2009;518:139–43.

[35] Grässel O, Krüger L, Frommeyer G, Meyer LW. High strength Fe–Mn–(Al, Si)TRIP/TWIP steels development – properties – application. Int J Plast2000;16:1391–409.

[36] Ding H, Tang Z-, Li W, Wang M, Song D. Microstructures and mechanicalproperties of Fe–Mn–(Al, Si) TRIP/TWIP Steels. J Iron Steel Res Int2006;13:66–70.

[37] Desroches R, McCormick J, Delemont M. Cyclic properties of superelasticshape memory alloy wires and bars. J Struct Eng 2004;130:38–46.

[38] Bergeon N, Guenin G, Esnouf C. Characterization of the stress-induced emartensite in a Fe–Mn–Si–Cr–Ni shape memory alloy: microstructuralobservation at different scales, mechanism of formation and growth. MaterSci Eng A 1997;238:309–16.

[39] Khalil W, Saint-Sulpice L, Arbab Chirani S, Bouby C, Mikolajczak A, Ben ZinebT. Experimental analysis of Fe-based shape memory alloy behavior underthermomechanical cyclic loading. Mech Mater 2013;63:1–11.

[40] Lee WJ, Weber B, Feltrin G, Czaderski C, Motavalli M, Leinenbach C. Phasetransformation behavior under uniaxial deformation of an Fe–Mn–Si–Cr–Ni–VC shape memory alloy. Mater Sci Eng A 2013;581:1–7.

[41] Hartl DJ, Lagoudas DC. In: Lagoudas DC, editor. Shape memory alloys.modelling and engineering applications. Springer; 2008. p. 53–119 [Chapter 2].

[42] Baruj A, Bertolino G, Troiani HE. Temperature dependence of critical stressand pseudoelasticity in a Fe–Mn–Si–Cr pre-rolled alloy. J Alloys Compd2010;502:54–8.

[43] Bouraoui T, Jemal F, Zineb TB. Tensile properties of a Fe–32Mn–6Si shapememory alloy. Strength Mater 2008;40:203–11.

[44] Melton K, Simpson J, Duerig T. A new wide hysteresis NiTi based shapememory alloy and its applications. Proc Int Conf on Mart Transf 1986:1053–8.

[45] Uchida K, Shigenaka N, Sakuma T, Sutou Y, Yamauchi K. Effects of pre-strainand heat treatment temperature on phase transformation temperature andshape recovery stress of Ti–Ni–Nb shape memory alloys for pipe jointapplications. Mater Trans 2008;49:1650–5.

[46] Lee WJ, Weber B, Feltrin G, Czaderski C, Motavalli M, Leinenbach C. Stressrecovery behaviour of an Fe–Mn–Si–Cr–Ni–VC shape memory alloy used forprestressing. Smart Mater Struct 2013;22:1–9.

[47] Alam MS, Youssef MA, Nehdi M. Utilizing shape memory alloys to enhancethe performance and safety of civil infrastructure: a review. Can J Civil Eng2007;34:1075–86.

[48] Wen YH, Xie WL, Li N, Li D. Remarkable difference between effects of carboncontents on recovery strain and recovery stress in Fe–Mn–Si–Cr–Ni–C alloys.Mater Sci Eng A 2007;457:334–7.

[49] Zhang W, Wen YH, Li N, Huang SK. Remarkable improvement of recoverystress of Fe–Mn–Si shape memory alloy fabricated by equal channel angularpressing. Mater Sci Eng A 2007;454–455:19–23.

[50] Wang CP, Wen YH, Peng HB, Xu DQ, Li N. Factors affecting recovery stress inFe–Mn–Si–Cr–Ni–C shape memory alloys. Mater Sci Eng A2011;528:1125–30.

[51] Li JC, Lü XX, Jiang Q. Creep behavior and calorimetric measurement of iron-based shape memory alloy. J Mater Sci Lett 2000;19:389–91.

[52] Liu L, Lin C, Zhou C. The stress relaxation characteristic and martensitictransformation of Fe–Mn–Si shape memory alloys under differentdeformation condition. Adv Mater Res 2012;430–432:106–9.

[53] Diederichs U, Schneider U. Bond strength at high temperatures. Mag ConcrRes 1981;33:75–84.

[54] Hertz K. Anchorage capacity of reinforcing bars at normal and hightemperatures. Mag Concr Res 1982;34:213–20.

[55] Morley PD, Royles R. Response of the bond in reinforced concrete to hightemperatures. Mag Concr Res 1983;35:67–74.

[56] Pothisiri T, Panedpojaman P. Modeling of bonding between steel rebarand concrete at elevated temperatures. Constr Build Mater 2012;27:130–40.

[57] Söderberg O, Liu XW, Yakovenko PG, Ullakko K, Lindroos VK. Corrosionbehaviour of Fe–Mn–Si based shape memory steels trained by cold rolling.Mater Sci Eng A 1999;273–275:543–8.

[58] Lin HC, Lin KM, Lin CS, Ouyang TM. The corrosion behavior of Fe-based shapememory alloys. Corros Sci 2002;44:2013–26.

[59] Huang X, Chen S, Hsu TY, Zuyao XU. Corrosion behavior of Fe25Mn6Si5Crshape memory alloys modified with rare earth in a NaCl solution. J Mater Sci2004;39:6857–9.

[60] Maji BC, Das CM, Krishnan M, Ray RK. The corrosion behaviour of Fe–15Mn–7Si–9Cr–5Ni shape memory alloy. Corros Sci 2006;48:937–49.

[61] Hu BQ, Bai PK, Dong ZZ, Cheng J. Effect of Cu addition on corrosion resistanceand shape memory effect of Fe–14Mn–5Si–9Cr–5Ni alloy. Trans NonferrousMet Soc China 2009;19:149–53.

[62] Charfi A, Bouraoui T, Feki M, Bradai C, Normand B. Surface treatment andcorrosion behaviour of Fe–32Mn–6Si shape memory alloy. C R Chim2009;12:270–5.

[63] Rovere CAD, Alano JH, Otubo J, Kuri SE. Corrosion behavior of shape memorystainless steel in acid media. J Alloys Compd 2011;509:5376–80.

[64] Charfi A, Gamaoun F, Bouraoui T, Bradai C, Normand B. Shape memory effectimprovement and study of the corrosion resistance of the Fe–8Mn–6Si–13Cr–6Ni–12Co alloy. Adv Mater Res 2012;476–478:2162–70.

[65] Della Rovere CA, Alano JH, Silva R, Nascente PAP, Otubo J, Kuri SE.Characterization of passive films on shape memory stainless steels. CorrosSci 2012;57:154–61.

[66] Rovere CAD, Alano JH, Silva R, Nascente PAP, Otubo J, Kuri SE. Influence ofalloying elements on the corrosion properties of shape memory stainlesssteels. Mater Chem Phys 2012;133:668–73.

[67] Lin HC, Lin KM, Chuang YC, Chou TS. Welding characteristics of Fe–30Mn–6Siand Fe–30Mn–6Si–5Cr shape memory alloys. J Alloys Compd2000;306:186–92.

[68] Dong ZZ, Sawaguchi T, Kajiwara S, Kikuchi T, Kim SH, Lee GC. Microstructurechange and shape memory characteristics in welded Fe–28Mn–6Si–5Cr–0.53Nb–0.06C alloy. Mater Sci Eng A 2006;438–440:800–3.

[69] Qiao Z, Li L, Wang D, Li Z. Microstructure and shape recovery characteristicsin a TIG-welded Fe–Mn–Si–Cr–Ni shape memory alloy. Proc SPIE Int Soc OptEng 2007:6423.

[70] Zhou C, Lin C, Liu L. Study on YAG laser welding process of Fe–Mn–Si shapememory alloy. Appl Mech Mater 2010;37–38:1364–7.

[71] Zhou C, Lin C, Liu L. Study on CO2 laser weldability of Fe–Mn–Si shapememory alloy. Proc SPIE Int Soc Opt Eng 2012:8409.

[72] Druker AV, Perotti A, Esquivel I, Malarría J. A manufacturing process for shaftand pipe couplings of Fe–Mn–Si–Ni–Cr shape memory alloys. Mater Des2014;56:878–88.

[73] Saito T, Kapusta C, Takasaki A. Synthesis and characterization of Fe–Mn–Sishape memory alloy by mechanical alloying and subsequent sintering. MaterSci Eng A 2013;592:88–94.

[74] Min XH, Sawaguchi T, Ogawa K, Maruyama T, Yin FX, Tsuzaki K. An attempt tolower Mn content of Fe–17Mn–6Si–0.3C shape memory alloy. J Alloys Compd2013;577:S478–82.

[75] Maruyama T, Kurita T, Kozaki S, Andou K, Farjami S, Kubo H. Innovation inproducing crane rail fishplate using Fe–Mn–Si–Cr based shape memory alloy.Mater Sci Technol 2008;24:908–12.

[76] Song G, Ma N, Li HN. Applications of shape memory alloys in civil structures.Eng Struct 2006;28:1266–74.

[77] Czaderski C, Hahnebach B, Motavalli M. RC beam with variable stiffness andstrength. Constr Build Mater 2006;20:824–33.

[78] Wu M, Johannesson B, Geiker M. A review: self-healing in cementitiousmaterials and engineered cementitious composite as a self-healing material.Constr Build Mater 2012;28:571–83.

[79] Ozbulut OE, Hurlebaus S, Desroches R. Seismic response control using shapememory alloys: a review. J Intell Mater Syst Struct 2011;22:1531–49.

[80] Sun W. Seismic response control of high arch dams including contractionjoint using nonlinear super-elastic SMA damper. Constr Build Mater2011;25:3762–7.

[81] Branco M, Guerreiro L, Mahesh KK, Braz Fernandes FM. Effect of load cyclingon the phase transformations in Ni–Ti wires for civil engineeringapplications. Constr Build Mater 2012;36:508–19.

[82] Isalgue A, Auguet C, Carreras G, Torra V. SMA (Cu-Based, NiTi) for use indamping: the implications of reliability for long time applications and agingbehavior. Funct Mater Lett 2012:5.

[83] Dommer K, Andrawes B. Thermomechanical characterization of NiTiNb shapememory alloy for concrete active confinement applications. J Mater Civ Eng2012;24:1274–82.

A. Cladera et al. / Construction and Building Materials 63 (2014) 281–293 293

[84] Muntasir-Billah AHM, Alam SM. Seismic performance of concrete columnsreinforced with hybrid shape memory alloy (SMA) and fiber reinforcedpolymer (FRP) bars. Constr Build Mater 2012;28:730–42.

[85] Cladera A, Oller, E., Ribas, C. Pilot experiences in application of shape memoryalloys in structural concrete. J Mater Civ Eng 2013, in press; <http://dx.doi.org/10.1061/(ASCE)MT.1943-5533.0000974>.

[86] Torra V, Auguet C, Isalgue A, Carreras G, Terriault P, Lovey FC. Built indampers for stayed cables in bridges via SMA. The SMARTeR-ESF project: amesoscopic and macroscopic experimental analysis with numericalsimulations. Eng Struct 2013;49:43–57.

[87] Soroushian P, Ostowari K, Nossoni A, Chowdhury H. Repair and strengtheningof concrete structures through application of corrective posttensioning forceswith shape memory alloys. Transp Res Rec 2001:20–6.

[88] Watanabe Y, Miyazaki E, Okada H. Enhanced mechanical properties of Fe–Mn–Si–Cr shape memory fiber/plaster smart composite. Mater Trans2002;43:974–83.

[89] Sawaguchi T, Kikuchi T, Ogawa K, Kajiwara S, Ikeo Y, Kojima M, et al.Development of prestressed concrete using Fe–Mn–Si-based shape memoryalloys containing NbC. Mater Trans 2006;47:580–3.

[90] Watanabe Y, Wakatsuki T, Sato H, Maruyama T. Bending strength of Fe–Mn–Si–Cr shape memory alloy machining chips reinforced smart composite. ICCMInt Conf Compos Mater; 2007.

[91] Czaderski C, Shahverdi M, Brönnimann R, Leinenbach C, Motavalli M.Feasibility of iron-based shape memory alloy strips for prestressedstrengthening of concrete structures. Constr Build Mater 2014;56:94–105.

[92] Lee YK, Jun JH, Choi CS. Effect of e martensite content on the dampingcapacity of Fe–17%Mn alloy. Scripta Mater 1996;35:825–30.

[93] Jee KK, Jang WY, Baik SH, Shin MC, Choi CS. Damping capacity in Fe–Mn basedalloys. Scripta Mater 1997;37:943–8.

All in-text references underlined in blue are linked to publications on Rese

[94] Frommeyer G, Brüx U, Neumann P. Supra-ductile and high-strengthmanganese-TRIP/TWIP steels for high energy absorption purposes. ISIJ Int2003;43:438–46.

[95] Watanabe Y, Sato H, Nishino Y, Kim T. Training effects on dampingcapacity in Fe–Mn and Fe–Mn–Cr alloys. Mater Sci Forum 2010;638–642:2201–6.

[96] Sawaguchi T, Kikuchi T, Ogawa K, Yin F, Kajiwara S, Kushibe A, et al.Internal friction of Fe–Mn–Si-based shape memory alloys containing Nband C and their application as a seismic damping material. Key Eng Mat2006;319:53–8.

[97] Sawaguchi T, Sahu P, Kikuchi T, Ogawa K, Kajiwara S, Kushibe A, et al.Vibration mitigation by the reversible fcc/hcp martensitic transformationduring cyclic tension–compression loading of an Fe–Mn–Si-based shapememory alloy. Scripta Mater 2006;54:1885–90.

[98] Hedayati Dezfuli F, Shahria Alam M. Shape memory alloy wire-based smartnatural rubber bearing. Smart Mater Struct 2013:22.

[99] Moser K, Bergamini A, Christen R, Czaderski C. Feasibility of concreteprestressed by shape memory alloy short fibers. Mater Struct/Mater Constr2005;38:593–600.

[100] Shajil N, Srinivasan SM, Santhanam M. Self-centering of shape memory alloyfiber reinforced cement mortar members subjected to strong cyclic loading.Mater Struct 2013;46:651–61.

[101] Krstulovic-Opara N, Thiedeman PD. Active confinement of concrete memberswith self-stressing composites. ACI Struct J 2000;97:297–308.

[102] Shin M, Andrawes B. Experimental investigation of actively confined concreteusing shape memory alloys. Eng Struct 2010;32:656–64.

[103] Park J, Choi E, Park K, Kim HT. Comparing the cyclic behavior of concretecylinders confined by shape memory alloy wire or steel jackets. Smart MaterStruct 2011;20:1–11.

archGate, letting you access and read them immediately.