Shape-memory alloy
A shape-memory alloy (SMA) is a class of metallic materials that exhibit the unique ability to return to a predefined shape after undergoing deformation, either through heating (shape memory effect) or under specific stress conditions (superelasticity), due to a reversible, diffusionless martensitic phase transformation between a high-temperature austenite phase and a low-temperature martensite phase.[1][2] This transformation enables SMAs to recover strains up to 8-10% without permanent damage, distinguishing them from conventional metals.[2][3] The shape memory effect was first observed in 1932 by Arne Ölander in gold-cadmium alloys, but practical development accelerated with the discovery of nickel-titanium (NiTi, or Nitinol) in 1962 by William J. Buehler and Frederick Wang at the U.S. Naval Ordnance Laboratory.[1][2] Common SMAs include NiTi, which offers biocompatibility and high recovery strains of about 6%; copper-based alloys like Cu-Al-Mn and Cu-Zn-Al, providing cost-effective alternatives with recovery strains up to 10%; and iron-based alloys such as Fe-Mn-Si, noted for their higher elastic moduli around 184 GPa but lower recovery strains of about 2%.[1][2] These alloys are characterized by key transformation temperatures—martensite start (Ms), martensite finish (Mf), austenite start (As), and austenite finish (Af)—which determine their operational range, typically tunable from -196°C to over 100°C depending on composition.[1] Mechanically, SMAs display austenite moduli of 30-68 GPa and exhibit phenomena like twinning and detwinning during deformation, allowing energy dissipation and damping capabilities.[2][3] Superelasticity occurs above the Af temperature, where stress induces martensite formation that reverts upon unloading, enabling applications requiring large, reversible deformations.[1][2] Constitutive models, such as those by Auricchio and Taylor (1997), describe these behaviors for engineering simulations, accounting for hysteresis and phase fractions.[2] SMAs find diverse applications across industries due to their actuation, sensing, and energy absorption properties.[1] In medicine, NiTi is used in stents, orthodontic wires, and implants for its biocompatibility and superelasticity.[1][3] Aerospace employs them in actuators, such as F-14 aircraft pipe connectors and Boeing's vortex generators for drag reduction, as well as space missions like the Nimbus satellite's thermal louvers.[1][4] In civil engineering, they enhance seismic resilience in bridges and concrete structures by reducing residual strains by over 80%.[2] Consumer products include self-adjusting eyeglass frames and toys, while emerging uses involve high-temperature variants like Ni-Ti-Hf for engine components.[1][3]Introduction
Definition and Overview
Shape-memory alloys (SMAs) are a class of smart materials capable of returning to a predefined shape after undergoing significant deformation, either through heating above a critical temperature (shape memory effect) or upon removal of applied stress (superelasticity or pseudoelasticity).[5] These properties arise from a reversible, solid-state phase transformation in the alloy's crystal structure, distinguishing SMAs from conventional metals that exhibit plastic deformation.[1] At the core of this behavior is the diffusionless martensitic phase transformation between two distinct phases: the high-temperature austenite phase, which is rigid and typically cubic in symmetry, and the low-temperature martensite phase, which is softer, deformable, and often monoclinic or orthorhombic.[5] In the martensite phase, the material can be easily deformed via twinning, but heating induces reversion to the austenite phase, recovering the original shape with strains up to 8-10%.[1] This thermomechanical coupling enables SMAs to function as actuators, dampers, and sensors in engineering applications, offering high power-to-weight ratios and silent operation.[1][6] The engineering significance of SMAs is underscored by their growing adoption in diverse fields, from aerospace morphing structures to biomedical devices, with the global market valued at approximately USD 15.5 billion in 2024 and projected to reach USD 45.4 billion by 2034 at a compound annual growth rate of 11.3%.[7] In everyday applications, nickel-titanium (NiTi) SMAs are used in self-expanding cardiovascular stents that deploy upon body temperature activation and in flexible eyeglass frames that recover from bending without fracture.[8][9]Historical Development
The shape-memory effect was first observed in the early 1930s by Swedish metallurgist Arne Ölander, who noted unusual recovery behavior in a gold-cadmium alloy during deformation experiments, marking the initial recognition of this phenomenon in metallic materials.[10] This discovery laid the groundwork for subsequent research, though practical applications remained limited due to the alloy's properties and cost. A major breakthrough occurred in 1962 when William J. Buehler, a metallurgist at the U.S. Naval Ordnance Laboratory in White Oak, Maryland, and his colleague Frederick E. Wang identified the shape-memory effect in a near-equiatomic nickel-titanium alloy during studies for high-strength materials in missile technology.[11] They named the alloy Nitinol, an acronym derived from its primary elements (nickel and titanium) and the laboratory's initials (Naval Ordnance Laboratory). This development shifted focus toward more viable, corrosion-resistant alloys with tunable transformation temperatures. Commercialization accelerated in the 1970s, driven by the medical sector's adoption of Nitinol for orthodontic arch wires around 1975, which exploited superelasticity to apply gentle, continuous forces and minimize patient discomfort.[12] Simultaneously, U.S. government patents facilitated broader uses, including actuators for couplings and thermal devices, transitioning shape-memory alloys from laboratory curiosities to engineered components.[13] The 1980s and 1990s saw expanded applications, including the introduction of pseudoelastic Nitinol eyeglass frames that resisted permanent deformation, enhancing durability in consumer products.[12] In aerospace, shape-memory alloys gained traction for adaptive structures, such as variable-geometry chevrons and deployment mechanisms in satellites, leveraging their high work density and reliability in harsh conditions.[14] From the 2000s, integration into consumer electronics and robotics proliferated, with Nitinol wires enabling miniaturized actuators in cell phones, cameras, and biomimetic robots for precise motion control.[15] By 2024-2025, emphasis shifted to high-temperature shape-memory alloys, such as NiTiHf variants, designed for extreme environments like jet engines and hypersonic vehicles, achieving transformation temperatures above 200°C through advanced processing techniques.[16][17]Fundamental Mechanisms
Shape Memory Effect
The shape memory effect (SME) is a unique thermomechanical behavior exhibited by certain alloys, enabling them to recover their original shape after significant deformation upon heating, driven by a reversible martensitic phase transformation. In shape-memory alloys like Nitinol (NiTi), this effect arises from the diffusionless transformation between the high-temperature austenite phase and the low-temperature martensite phase, allowing for large recoverable strains without permanent damage. The material demonstrates a "memory" of its austenitic configuration set during processing.[1] In the one-way SME, the alloy is first cooled below the martensite start temperature (Ms) and martensite finish temperature (Mf), transitioning fully into the twinned martensite phase, which is soft and easily deformable by detwinning under applied stress, accommodating strains up to 8%. The deformed martensitic structure is then heated above the austenite start temperature (As) and austenite finish temperature (Af), prompting the reverse transformation to the rigid austenite phase and full recovery of the original shape. This process exhibits thermal hysteresis, characterized by the temperature gap between the forward (martensite formation on cooling) and reverse (austenite formation on heating) transformations, typically spanning 20–50°C in NiTi depending on composition and processing. The recovery strain is given by \epsilon_{recovery} = \epsilon_{deformation}, where the maximum recoverable deformation reaches 6–8% in well-processed Nitinol, establishing its utility for applications requiring one-time actuation.[1][18] The two-way SME extends this capability, allowing the alloy to spontaneously deform into a low-temperature martensitic shape upon cooling and recover to the high-temperature austenitic shape upon heating, without applied external stress in either direction. This bidirectional actuation is not inherent but is induced through a shape-memory training process, involving repeated thermomechanical cycles where the alloy is deformed in the martensite phase, heated to recover the austenite shape, and cooled to fix the martensite configuration, often requiring 10–50 cycles for stabilization via internal stress fields or oriented precipitates. In NiTi, training via martensitic deformation around a fixture at low temperature followed by heating above Af and quenching achieves attack angles or strains of several percent, enabling cyclic applications like actuators. Unlike the one-way effect, which requires heating solely for recovery and external force for resetting, the two-way variant supports reversible operation over temperature cycles, though with potentially reduced maximum strain (typically 2–4%) due to training-induced fatigue limits.[19][1]Pseudoelasticity
Pseudoelasticity, also known as superelasticity, refers to the ability of certain shape-memory alloys to undergo large deformations and fully recover their original shape upon removal of the applied stress, without any permanent strain, when tested at temperatures above the austenite finish temperature (Af). This behavior is primarily observed in alloys like NiTi (Nitinol), where the mechanism involves a stress-induced martensitic phase transformation from the high-temperature austenite phase to the low-temperature martensite phase during loading. The transformation produces favorably oriented martensite variants that accommodate the deformation through a detwinning process, allowing strains far exceeding conventional elastic limits. Upon unloading, the reverse martensitic transformation occurs, restoring the austenite structure and the original shape.[2] The characteristic stress-strain response under pseudoelastic conditions features an initial linear elastic regime in the austenite phase, governed by the austenite modulus (typically around 50-80 GPa for NiTi), followed by a nearly flat upper plateau associated with the forward stress-induced transformation, where significant strain (up to 6-8%) develops at relatively constant stress levels around 400-600 MPa. This plateau reflects the progressive formation of martensite, after which a brief elastic loading in martensite may occur before unloading begins with a lower plateau corresponding to the reverse transformation, exhibiting a pronounced hysteresis loop due to the energy barrier between phases. The full recovery of the transformation strain distinguishes pseudoelasticity from plastic deformation, with representative recoverable strains of 6-8% in well-processed NiTi alloys enabling applications beyond typical metallic elasticity (0.1-0.2%).[20][2] One practical application of this hysteresis is in structural damping, where the energy dissipation during the loading-unloading cycle—quantified by the area enclosed in the stress-strain loop—provides vibration control and seismic isolation in bridges and buildings, as demonstrated in NiTi wire-reinforced systems that enhance passive damping by up to several times conventional levels.[21] Pseudoelasticity is temperature-dependent and confined to the austenite stability range, typically 10-100°C above Af for NiTi, where the material remains austenitic under no load but the stress overcomes the transformation barrier. As temperature increases within this range, the critical stresses for forward and reverse transformations rise (following the Clausius-Clapeyron relation), while the hysteresis width narrows, reducing energy dissipation but improving recovery efficiency; below Af, the behavior shifts toward the shape memory effect, with incomplete stress recovery.[2]Phase Transformations and Crystal Structures
Martensitic Transformation
The martensitic transformation in shape-memory alloys is a diffusionless, shear-dominated solid-state phase change that occurs without atomic diffusion, involving the cooperative and nearly simultaneous movement of atoms over small distances to form a new crystal structure. This transformation typically proceeds from the high-temperature austenite phase, which possesses a body-centered cubic (BCC) or face-centered cubic (FCC) lattice, to the low-temperature martensite phase, which can have a monoclinic, orthorhombic, body-centered tetragonal (BCT), or other structures depending on the alloy composition.[22] The shear nature of the process results in a significant shape change, often accommodated by invariant plane strain, enabling the material's functional properties. The driving force for the martensitic transformation arises from both chemical and non-chemical contributions, with the chemical component primarily stemming from undercooling below the equilibrium temperature, and the non-chemical from applied stress that favors the martensite variant. This driving force is thermodynamically described by the change in Gibbs free energy, given by \Delta G = \Delta H - T \Delta S, where \Delta H is the enthalpy change, T is the temperature, and \Delta S is the entropy change associated with the phase transition. In shape-memory alloys, the balance of these terms ensures that the transformation is reversible under appropriate conditions, with stress lowering the effective transformation temperature by contributing to \Delta G.[23][24] The kinetics of the martensitic transformation in shape-memory alloys can be either athermal, where the extent of transformation depends directly on temperature without time dependence, or isothermal, involving time-activated nucleation and growth at constant temperature. Athermal transformations are predominant in most shape-memory alloys like NiTi, occurring rapidly upon cooling through critical temperatures, while isothermal variants, observed in some compositions, exhibit slower progression due to thermal activation barriers. To accommodate the shear strain without introducing significant dislocations, the martensite phase forms with internal twinning, where thin layers of differently oriented martensite variants stack to maintain compatibility with the surrounding austenite and minimize elastic strain energy.[25][26] Critical temperatures define the boundaries of the martensitic transformation and are denoted as M_s (martensite start), M_f (martensite finish), A_s (austenite start), and A_f (austenite finish), marking the onset and completion of forward and reverse transformations, respectively. These temperatures are precisely measured using differential scanning calorimetry (DSC), which detects the exothermic and endothermic heat flows associated with the phase change, providing enthalpies and hysteresis widths that characterize the material's thermal stability. In shape-memory alloys, the hysteresis between forward and reverse peaks in DSC traces reflects the energy barriers for nucleation and interface motion.[27] The martensitic transformation underpins the shape memory effect (SME) and pseudoelasticity by enabling thermoelastic variants, where the reverse transformation is driven by the release of stored elastic energy, ensuring high reversibility over multiple cycles. In thermoelastic martensite, the fine twinned structure and low fault energy allow the phase boundaries to move with minimal irreversibility, facilitating shape recovery upon heating for SME or stress removal for pseudoelasticity. This reversible character distinguishes shape-memory alloys from conventional martensitic steels, where transformations are often non-thermoelastic and lead to permanent deformation.[28]Crystal Structures
Shape-memory alloys exhibit distinct crystal structures in their austenite and martensite phases, which underpin their functional properties. The high-temperature austenite phase, prevalent above the transformation temperature, typically adopts a B2 structure in binary alloys like NiTi. This is an ordered body-centered cubic (BCC) lattice of the CsCl type, characterized by high cubic symmetry where Ni atoms occupy the cube corners and Ti atoms the body center (or vice versa), resulting in a rigid, equiaxed arrangement. The lattice parameter for this phase in NiTi is approximately a = 0.301 nm at room temperature.[29][30] In contrast, the low-temperature martensite phase in NiTi features a monoclinic B19' structure, which arises from the diffusionless shear transformation of austenite and possesses lower symmetry to accommodate lattice distortion. This structure enables the formation of multiple twinned variants—self-accommodating domains of martensite plates oriented to minimize overall strain during formation. Deformation in the martensitic state occurs via twinning, where applied stress reorients these variants through detwinning, aligning them to produce macroscopic shape change recoverable upon heating. In certain ternary alloys, such as Ti-Ni-Cu systems, the martensite adopts an orthorhombic structure instead, altering the variant interfaces and transformation behavior.[29][31][32][33] Ni-rich variants of NiTi alloys often involve an intermediate R-phase during transformation, exhibiting a rhombohedral crystal structure that distorts the cubic austenite lattice slightly, with lattice parameter a \approx 0.301 nm and angle \alpha < 90^\circ. This phase facilitates a two-step transformation sequence (B2 to R-phase, then to B19'), influencing the overall hysteresis and recoverable strain. The precise lattice parameters of these phases critically determine the maximum transformation strains, as deviations in atomic spacing directly modulate the shear components and compatibility between variants.[29][34][35]Materials and Composition
Common Binary Alloys
The most prominent binary shape-memory alloy is Nitinol, composed of approximately 55 wt% nickel and 45 wt% titanium, which exhibits a tunable austenite finish temperature (Af) ranging from -200°C to 100°C through adjustments in composition and heat treatment.[36] This alloy demonstrates a transformation strain of 4-8% during the martensitic phase change, enabling reliable shape recovery in applications such as actuators and medical devices.[37] Its near-equiatomic structure supports both the shape-memory effect and pseudoelasticity, depending on the operating temperature relative to the transformation range.[38] Copper-based binary alloys, such as Cu-Zn-Al, are near-eutectoid compositions typically containing 22-28 wt% Zn and 3-5 wt% Al, offering lower production costs compared to Nitinol while achieving Af temperatures up to 200°C.[39] These alloys provide a broader transformation temperature range of -200°C to 200°C, making them suitable for high-temperature actuators, though their recovery strain is generally around 5%.[36] Cu-Al-Ni variants, with about 4 wt% Ni added to similar Cu-Al bases, exhibit improved corrosion resistance over Cu-Zn-Al due to the stabilizing effect of nickel on the protective oxide layer.[37] Iron-based binary alloys, exemplified by Fe-Mn-Si with roughly 30 wt% Mn and 5 wt% Si, represent the most economical option among common shape-memory alloys, with costs significantly lower than those of NiTi or Cu-based systems.[37] They deliver a recovery strain of up to approximately 4% (typically 2% without training), constrained by interactions between martensite variants and grain boundaries, and are primarily applied in civil engineering for prestressing concrete structures where high recovery stress is prioritized over large strains.[40]| Alloy Type | Transformation Temperature Range (°C) | Relative Cost | Corrosion Resistance |
|---|---|---|---|
| NiTi (Nitinol) | -200 to 100 | High | Excellent (TiO₂ passivation layer) |
| Cu-based (e.g., CuZnAl, CuAlNi) | -200 to 200 | Low | Poor to moderate (improved in Ni variants) |
| Fe-based (e.g., FeMnSi) | 50 to 300 (tunable via annealing) | Very low | Moderate (suitable for non-corrosive environments) |