{
  "id": "martensite",
  "title": "Martensite",
  "lang": "en",
  "source_lang": "en",
  "summary": "Martensite is a crystalline transformation product formed by coordinated atomic motion, central to steel hardening and the behavior of shape-memory alloys.",
  "category": "technology",
  "subcategory": "materials",
  "type": "concept",
  "aliases": [
    "Martensitic phase"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/martensite/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/martensite/",
      "json": "https://aiwiki.page/zh/martensite.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Martensite is a crystalline transformation product formed when a parent phase changes structure through coordinated atomic displacements rather than long-range atomic diffusion. In its most familiar form, it develops when steel containing austenite is cooled sufficiently rapidly to suppress competing transformations. The resulting iron-rich structure can retain carbon far above its equilibrium solubility, producing the high strength and hardness associated with quenched steel. The term also applies to transformation products in many nonferrous alloys and some ceramics, where the composition, crystal symmetry, and mechanical properties differ substantially from those of carbon steel.\n\nMartensite is important both as an engineering material and as a model of solid-state phase transformation. Its formation governs the performance of cutting tools, bearings, gears, structural steels, and many welded components. Reversible martensitic transformations also underlie shape memory and superelasticity. Understanding martensite therefore requires distinguishing the transformation mechanism from the properties of any particular alloy: diffusionless formation does not necessarily imply high hardness, brittleness, or rapid cooling."
    },
    {
      "heading": "Definition and scope",
      "text": "In metallurgy, martensite denotes the product of a transformation in which neighboring atoms move cooperatively to establish a new crystal lattice. Atoms generally retain their near neighbors, and the transformation proceeds without the compositional redistribution characteristic of diffusional precipitation. The parent and product consequently have closely related compositions at the moment of transformation. This does not mean that every atomic movement is identical or that defects are absent; local atomic shuffles, interfacial motion, and accommodation by the surrounding material are integral to many martensitic transformations.\n\nThe defining distinction concerns mechanism, not simply cooling rate. Quenching steel is the best-known route to martensite because rapid cooling prevents carbon redistribution and the formation of other products before the martensitic transformation begins. Nevertheless, martensite can also form during slow cooling in alloys whose competing transformations are sufficiently sluggish. Applied stress or plastic deformation can induce it at a constant temperature, while some martensitic transformations proceed measurably with time under isothermal conditions.\n\nTerminology requires particular care in steels. Austenite is the face-centered cubic parent phase, conventionally denoted \\(\\gamma\\). Ordinary carbon-steel martensite is commonly denoted \\(\\alpha'\\), with a body-centered tetragonal structure when sufficient carbon remains in solid solution. At low carbon contents, its tetragonal distortion can be very small, and the lattice is often described as approximately body-centered cubic. Some austenitic steels additionally form hexagonal close-packed \\(\\varepsilon\\)-martensite, either as an intermediate in the formation of \\(\\alpha'\\) or as a product in its own right.\n\nMartensite is not a single chemical compound or a unique crystal structure shared by all materials. Iron–carbon martensite, titanium-alloy martensite, and the martensite of nickel–titanium alloys are distinct phases with different structures and properties. Their common feature is the character of the transformation. Nor is martensite synonymous with every needle-shaped microstructure: diffusional products can also appear acicular, while martensite can occur as laths, plates, or more complex arrangements. Identification therefore depends on crystallography, composition, and transformation behavior rather than optical appearance alone."
    },
    {
      "heading": "Atomic mechanism and crystallography",
      "text": "The martensitic transformation changes the parent crystal structure through a coordinated lattice deformation accompanied, where required, by atomic shuffles. In steels, the Bain correspondence provides a particularly useful geometrical description. Proposed by Edgar C. Bain in 1924, it relates the face-centered cubic austenite lattice to a body-centered lattice through contraction along one principal direction and expansion along the other two. This correspondence describes how lattice sites can be related, but it is not by itself a complete description of the actual transformation pathway or observed plate shape.\n\nA real martensitic region must fit into its surroundings. Its formation therefore produces both a change of shape and elastic or plastic deformation in adjacent material. The overall shape deformation is often approximately an invariant-plane strain: one plane remains essentially undistorted and unrotated, while the material undergoes shear and a smaller normal displacement. This plane is associated with the observed habit plane between martensite and austenite. The resulting surface relief is a characteristic experimental sign of martensitic transformation.\n\nThe lattice deformation alone generally cannot provide the observed compatibility. Additional deformation through slip or twinning supplies what crystallographic theories call lattice-invariant deformation. Slip involves the movement of dislocations, whereas twinning produces regions with a specific mirror-related crystallographic orientation. These mechanisms allow the product to accommodate its shape change while maintaining a relatively favorable interface with the parent. Accommodation is incomplete, leaving residual stresses and a high density of defects.\n\nSpecific orientation relationships connect parent and product crystals. In steels, the Kurdjumov–Sachs relationship approximately aligns \\(\\{111\\}_{\\gamma}\\) planes with \\(\\{110\\}_{\\alpha'}\\) planes and \\(\\langle110\\rangle_{\\gamma}\\) directions with \\(\\langle111\\rangle_{\\alpha'}\\) directions. The Nishiyama–Wassermann relationship shares the same parallel-plane relationship but has a different parallel-direction condition. Actual measurements often show small deviations from these ideal descriptions.\n\nBecause the parent crystal has several symmetry-equivalent directions and planes, a single austenite grain can generate multiple martensitic orientation variants. Their selection is influenced by stress, neighboring grains, interfaces, and the need to accommodate transformation strain. Variant combinations can partially cancel their shape changes, but externally applied loading can favor variants whose transformation strain contributes most favorably to the imposed deformation."
    },
    {
      "heading": "Carbon, lattice distortion, and microstructure",
      "text": "Carbon plays a central role in the structure and strength of martensite in ordinary steels. Austenite can dissolve substantially more carbon than equilibrium body-centered cubic iron at low temperature. When the lattice transforms without long-range diffusion, carbon is initially trapped in interstitial positions within the product. Preferential occupation of particular interstitial sites produces a tetragonal distortion, conventionally expressed by the ratio \\(c/a\\) of the lattice parameters.\n\nFor idealized, relatively homogeneous iron–carbon martensite, a frequently used approximate relationship is\n\n\\[\\frac{c}{a}\\approx 1+0.045C,\\]\n\nwhere \\(C\\) is the carbon content expressed numerically in weight percent. This relation is not universal. Carbon segregation to defects, carbide precipitation, auto-tempering, alloy additions, and local variations in carbon ordering can all alter measured tetragonality. Modern diffraction and microscopy reveal that an apparently simple average lattice parameter may conceal considerable variation within individual martensitic regions.\n\nLow-carbon and many medium-carbon steels predominantly develop lath martensite. Individual laths are elongated crystals containing high dislocation densities. They are organized into a hierarchy of sub-blocks, blocks, and packets within a prior austenite grain. These levels correspond to different crystallographic relationships, and their boundaries do not offer identical resistance to slip or crack propagation. Consequently, the mechanically effective grain size is not necessarily the width of an individual lath or the diameter of the original austenite grain.\n\nHigher-carbon steels more commonly develop plate martensite, often containing extensive internal twinning. Plates may display a midrib, marking an early stage of growth, with later regions accommodated partly through dislocation processes. The transition from predominantly lath to predominantly plate morphology is gradual and depends on composition and transformation temperature rather than on one universal carbon threshold. Mixed lath-and-plate structures occur over intermediate conditions.\n\nThe surrounding material further complicates the microstructure. Untransformed retained austenite can remain as thin films between martensitic units or as larger blocky regions. Carbides may survive incomplete dissolution during austenitizing, or precipitate during and after transformation. Because low-carbon martensite can begin forming at relatively high temperatures, carbon may redistribute while the component is still cooling; this is termed auto-tempering. An as-quenched steel is therefore not necessarily a chemically uniform mixture of pristine martensite and unchanged austenite."
    },
    {
      "heading": "Transformation temperatures and kinetics",
      "text": "Martensite begins to form on cooling when the chemical driving force becomes sufficient to overcome the energetic and mechanical barriers to transformation. The martensite-start temperature, \\(M_s\\), identifies the onset under specified conditions. The martensite-finish temperature, \\(M_f\\), describes the temperature below which transformation is regarded as essentially complete by a particular measurement criterion. Neither is an equilibrium phase boundary, and completion need not mean the literal absence of all residual austenite.\n\nComposition strongly affects these temperatures. Carbon and several common alloying elements stabilize austenite and lower \\(M_s\\). One widely used empirical relationship for steels, the Andrews equation, is\n\n\\[M_s\\;({}^{\\circ}\\mathrm{C}) = 539-423C-30.4Mn-17.7Ni-12.1Cr-7.5Mo,\\]\n\nwhere the element symbols represent concentrations in weight percent. Such equations apply within the compositional ranges from which they were developed. They do not replace measurements for highly alloyed steels, unusual austenitizing conditions, or microstructures with substantial chemical segregation. The relevant composition is that of the austenite actually transforming, which may differ from the bulk analysis if carbides remain undissolved.\n\nFor many steels, martensite formation is approximately athermal: the fraction transformed depends principally on the lowest temperature reached rather than on the duration of a hold. Individual units can form rapidly, while further cooling activates additional transformation events. A commonly used description is the Koistinen–Marburger equation, published in 1959:\n\n\\[f_{\\mathrm{M}}=1-\\exp[-\\alpha(M_s-T)], \\qquad T<M_s,\\]\n\nwhere \\(f_{\\mathrm{M}}\\) is the fraction of initially available austenite transformed and \\(\\alpha\\) is an empirical coefficient. A frequently quoted value for certain plain-carbon steels is about \\(0.011\\;{}^{\\circ}\\mathrm{C}^{-1}\\), but its applicability depends on material and conditions.\n\nThe equation illustrates why retained austenite can persist after substantial undercooling: the predicted transformation approaches completion progressively rather than ending at a sharp universal temperature. Mechanical constraint, austenite grain size, and local carbon enrichment can further stabilize the remaining material. Conversely, applied stresses can promote suitably oriented variants by contributing mechanical work to the transformation.\n\nAthermal behavior is common but not definitional. Isothermal martensitic transformations are documented in some alloys, and time-dependent nucleation can become important under particular temperature and stress conditions. Distinguishing nucleation, rapid growth of individual units, and the overall evolution of transformed fraction prevents the misleading inference that a fast local event must make the entire transformation instantaneous."
    },
    {
      "heading": "Quenching and competing transformations",
      "text": "Producing martensite in steel normally begins with austenitizing: heating into an appropriate temperature range to establish the desired parent phase and dissolve selected constituents. The necessary temperature and holding time depend on composition, component size, initial microstructure, and the intended balance between carbide dissolution and austenite grain growth. Excessive heating can coarsen the prior austenite grains, while insufficient dissolution can leave the austenite lower in carbon or alloying elements than anticipated.\n\nCooling must then avoid excessive formation of competing products. Pearlite consists of an intergrowth of iron-rich ferrite and cementite, with carbon redistribution accompanying its formation. Bainite develops through a different transformation sequence and commonly contains fine ferritic units together with carbides or carbon-enriched austenite. Its detailed formation mechanism has been a major subject of metallurgical research. Unlike martensite formation, producing these microstructures ordinarily involves significant carbon redistribution during the overall reaction.\n\nA time–temperature–transformation diagram describes transformations during isothermal holds, while a continuous-cooling-transformation diagram represents behavior during continuous cooling. Their transformation-start curves show why a given steel requires a particular cooling history to develop a specified martensite fraction. A cooling path that passes through a pearlite or bainite transformation region can consume part of the austenite before the remaining material reaches \\(M_s\\). The final component may therefore contain several products rather than a single phase.\n\nHardenability describes the capacity of a steel to form a hardened structure through a given section under specified cooling conditions. It is distinct from maximum attainable hardness. Carbon largely controls the hardness potential of ordinary as-quenched martensite, whereas elements such as chromium, manganese, molybdenum, and nickel can retard competing transformations and permit martensite formation at lower cooling rates. The Jominy end-quench test assesses hardenability through a hardness profile generated by a controlled cooling-rate gradient.\n\nWater, brine, oil, aqueous polymer solutions, and gases provide different heat-transfer conditions. Cooling rate also varies within a single component: a surface generally cools faster than its interior, and vapor formation or uneven fluid flow can alter local heat extraction. The transformation-related expansion of steel martensite interacts with thermal contraction to generate distortion and residual stress. Quench cracking arises when the resulting tensile stresses exceed local resistance to fracture. Martempering uses an interrupted cooling schedule near the martensitic transformation range to reduce temperature differences before transformation proceeds extensively."
    },
    {
      "heading": "Strength, toughness, and tempering",
      "text": "The strength of carbon-steel martensite results from several interacting mechanisms. Interstitial carbon produces strong local lattice distortions and impedes dislocation motion. Transformation generates a high dislocation density, while lath, block, and packet boundaries constrain subsequent deformation. Carbon segregation and fine precipitates can provide additional obstacles. Tetragonality is therefore an important structural signature, but it is not a sufficient explanation of martensitic hardness by itself.\n\nHardness generally rises with carbon content in the ordinary carbon-steel range, although the relationship becomes complicated by retained austenite, carbides, and changes in morphology. Low-carbon martensite can combine high strength with useful toughness and ductility. High-carbon, untempered martensite is more susceptible to brittle fracture, especially when it contains high residual stresses or unfavorable grain-boundary conditions. Thus “martensitic” does not specify a unique mechanical behavior.\n\nTempering deliberately modifies the as-quenched structure by reheating below the temperature at which extensive austenite formation occurs. Carbon redistributes, carbides develop, dislocations rearrange or annihilate, and residual stresses relax. At relatively low temperatures, carbon clustering and transition-carbide precipitation may occur; with increasing temperature or time, cementite formation and structural recovery become more prominent in many carbon steels. These processes overlap, and the sequence depends on composition and heating conditions.\n\nThe usual consequence is lower hardness accompanied by improved ductility and fracture toughness, but alloy steels can behave differently. Fine alloy-carbide precipitation can produce secondary hardening during tempering, as in many high-speed and hot-work tool steels. Tempering can also destabilize retained austenite, causing it to transform during cooling and creating fresh martensite. Multiple tempering treatments are used in some tool steels to address these successive changes.\n\nNot every tempering response is beneficial. Certain temperature ranges can produce tempered-martensite embrittlement, while other embrittlement phenomena involve impurity segregation during exposure or slow cooling at higher tempering temperatures. These mechanisms should not be treated as one process. Likewise, susceptibility to hydrogen embrittlement depends on strength level, hydrogen availability, trapping sites, stress, and microstructure, not on the martensitic label alone. Engineering specifications therefore define both composition and processing history, together with measurable property requirements.\n\nThe expression “tempered martensite” describes the inherited microstructure after tempering. At sufficiently advanced stages, its phases may be predominantly ferritic iron and carbides rather than supersaturated martensite, even though the original lath or plate arrangement remains recognizable."
    },
    {
      "heading": "Identification and quantitative characterization",
      "text": "Optical metallography provides a first view of martensitic structures after polishing and suitable etching. It can reveal packets, plates, prior austenite boundaries, and heterogeneous regions, but fine laths and internal twins generally require higher resolution. Etching contrast alone does not uniquely identify a phase: heavily deformed ferrite, bainitic structures, and tempered martensite can have overlapping appearances. Reliable interpretation combines morphology with thermal history and complementary measurements.\n\nAn electron microscope allows examination at much smaller length scales. Scanning electron microscopy resolves surface morphology and microstructural boundaries, while transmission electron microscopy can reveal dislocations, twins, carbides, and thin retained-austenite films. Electron diffraction determines local crystal orientation and phase identity. Specimen preparation matters because mechanical polishing, ion milling, or electron-beam exposure can alter delicate regions or induce transformation in metastable austenite.\n\nElectron backscatter diffraction maps crystallographic orientation over polished surfaces. It is especially valuable for identifying martensitic variants and reconstructing prior austenite grain structures from the product orientations. Conventional indexing may not reliably distinguish low-tetragonality martensite from body-centered cubic ferrite because their diffraction patterns are very similar. High defect densities also reduce pattern quality. Phase identification consequently benefits from orientation relationships, local strain indicators, morphology, and independent phase measurements rather than a single automated classification.\n\nX-ray diffraction can measure lattice parameters, estimate tetragonality, and quantify retained austenite from the intensities of selected reflections. Texture, overlapping peaks, residual stress, and the spatial scale of sampling affect accuracy. Neutron diffraction offers greater penetration for bulk measurements in suitable specimens, while synchrotron methods can track transformation during cooling or deformation with high time resolution. Atom-probe tomography provides complementary three-dimensional information about carbon segregation and nanoscale chemical partitioning.\n\nDilatometry follows changes in specimen dimensions during controlled thermal cycles. Because austenite-to-martensite transformation in steel usually causes expansion relative to the extrapolated parent-phase behavior, dimensional curves help determine \\(M_s\\) and transformation progress. Calorimetry detects heat effects, and magnetic measurements can distinguish ferromagnetic \\(\\alpha'\\) from many austenitic matrices. Hardness mapping indicates the resulting mechanical heterogeneity but cannot independently establish phase identity.\n\nQuantification must specify both the measured quantity and its uncertainty. A reported martensite fraction may refer to image area, diffraction-derived volume fraction, or an estimate inferred from dilation. These methods sample different volumes and rely on different assumptions, so agreement is not automatic."
    },
    {
      "heading": "Industrial steels and processing routes",
      "text": "Quenched-and-tempered steels use martensite as an intermediate structure from which a controlled balance of strength and toughness is developed. Applications include shafts, fasteners, pressure-bearing components, and heavily loaded machine parts. The essential processing variables include austenite condition, through-section cooling, tempering temperature, and tempering time. A component described commercially as hardened steel may consequently contain tempered martensite rather than the brittle, highly supersaturated structure associated with an immediate quench.\n\nSurface-hardening routes concentrate martensitic strengthening where contact stresses and wear are greatest. Carburizing enriches the surface with carbon before quenching, producing a hard case over a lower-carbon core. Induction hardening instead heats a selected region rapidly through electromagnetic induction before cooling, usually without deliberately changing its bulk composition. Laser hardening can localize austenitization still further, with heat extraction into the surrounding material sometimes providing the necessary self-quenching. Gears, bearing surfaces, and machine-tool ways illustrate the importance of controlling both case depth and transition to the underlying material.\n\nMartensitic stainless steels combine a martensitic matrix with sufficient chromium to support corrosion resistance under suitable conditions. Their corrosion behavior depends on chromium available in solid solution, carbide distribution, surface condition, and environment. They are used in applications such as cutting instruments, turbine components, and selected mechanical parts. They should be distinguished from austenitic stainless steels, some of which form martensite only after deformation or cooling.\n\nMaraging steels take a different approach. Their low-carbon, commonly nickel-rich martensite provides a matrix that is subsequently strengthened by aging and intermetallic precipitation. High strength therefore does not arise primarily from carbon supersaturation. This illustrates why the same transformation category can support markedly different alloy-design strategies.\n\nAdvanced sheet steels deliberately combine martensite with other constituents. Dual-phase steels contain a relatively soft ferritic matrix and harder martensitic regions, producing useful combinations of strength, work hardening, and formability for automotive structures. Quenching-and-partitioning processing seeks to retain carbon-enriched austenite alongside martensite through controlled redistribution after partial transformation.\n\nMartensite can also arise unintentionally. In welding, rapid cooling of a hardenable heat-affected zone can create a brittle local structure and contribute to hydrogen-assisted cracking. In metal additive manufacturing, repeated localized heating and cooling can generate martensite, temper earlier-formed regions, or alter retained austenite. The resulting properties depend on the complete thermal history, not merely the peak cooling rate."
    },
    {
      "heading": "Martensite beyond carbon steels",
      "text": "The broader significance of martensite becomes clear in shape-memory alloys. These materials undergo transformations between a higher-temperature parent phase and a lower-temperature martensitic phase whose variants can rearrange under stress. If deformation occurs primarily through variant rearrangement rather than irreversible slip, heating can restore the parent structure and recover the original macroscopic shape. This is the shape-memory effect.\n\nIn nickel–titanium, near-equiatomic compositions commonly transform between an ordered cubic B2 parent phase and monoclinic B19′ martensite. Some compositions and processing conditions introduce an intermediate R phase. Transformation temperatures and hysteresis depend sensitively on composition, precipitation, deformation history, and heat treatment. Unlike high-carbon steel martensite, the functional martensite in these alloys is not valuable chiefly because of extreme hardness.\n\nSuperelasticity occurs when a material above its austenite-finish temperature forms stress-induced martensite during loading and transforms back during unloading. Large recoverable strains arise from this reversible structural change. The stress–strain response commonly contains a transformation plateau and a hysteresis loop, reflecting the different loading and unloading conditions required to move interfaces. Applications include flexible mechanical elements, couplings, and compact actuators. Functional performance depends on resistance to irreversible slip and transformation fatigue over repeated cycles.\n\nOther metallic systems broaden the range further. Titanium alloys can form hexagonal \\(\\alpha'\\) or orthorhombic \\(\\alpha''\\) martensite from a body-centered cubic parent phase, depending on composition. Copper-based shape-memory alloys also undergo reversible martensitic transformations. In some austenitic steels, deformation-induced transformation increases work hardening and delays strain localization, producing transformation-induced plasticity rather than a conventional thermally activated shape-memory cycle.\n\nCeramics provide another important example. Partially stabilized zirconia can undergo a stress-assisted tetragonal-to-monoclinic transformation near an advancing crack. The associated strain can oppose crack opening and contribute to transformation toughening. Here, transformation is exploited to improve resistance to fracture rather than to create a quenched metallic cutting edge.\n\nReversibility is therefore conditional, not inherent in all martensite. Thermoelastic transformations have comparatively mobile interfaces and a favorable balance of chemical and elastic energies, permitting substantial reverse motion. Steel martensite commonly produces extensive plastic accommodation and, during heating, may undergo carbon redistribution and carbide precipitation. Its reverse transformation consequently does not generally restore the original component shape."
    },
    {
      "heading": "Historical development and continuing research",
      "text": "The term martensite honors Adolf Martens, a German engineer whose late nineteenth-century work helped establish metallographic examination as an engineering method. Floris Osmond introduced the name during that period as researchers developed terminology for the structures revealed in steels. Early classifications relied heavily on etched appearance and mechanical properties because the atomic structures and transformation mechanisms were not yet accessible directly.\n\nThe development of diffraction methods in the early twentieth century connected hardening with changes in crystal structure. Bain’s 1924 lattice correspondence supplied a geometrical framework for the austenite-to-martensite transformation. Studies of surface relief, habit planes, orientation relationships, and internal twins then showed that the change could not be understood solely as a rearrangement of an isolated unit cell. The mechanical compatibility of a transforming region with the surrounding solid became a central problem.\n\nIn the 1950s, the phenomenological crystallographic theories associated with Morris Cohen and John S. Bowles and with Michael S. Wechsler, David S. Lieberman, and Thomas A. Read linked lattice deformation, lattice-invariant deformation, and macroscopic shape change. These theories successfully explained important features of habit planes and orientation relationships. They provided a bridge between atomic-scale lattice correspondences and observable plate geometry without claiming to resolve every detail of nucleation or interface motion.\n\nSubsequent research increasingly treated martensite as a hierarchical and chemically heterogeneous material. Electron microscopy established the importance of dislocation substructures and fine twinning. Orientation mapping clarified packet and block organization, while high-resolution diffraction and atom-probe methods revealed local tetragonality variations, carbon segregation, and nanoscale precipitation. These findings complicate simple textbook pictures in which carbon remains uniformly trapped until a separate tempering operation begins.\n\nOpen questions concern how martensite nucleates at particular defects, how interface structure controls mobility, and how local stresses select transformation variants. Small austenite regions may remain stable under conditions where larger grains transform, making size and constraint essential variables in modern multiphase steels. Experiments that follow transformation in situ are increasingly combined with atomistic simulation, phase-field modeling, and crystal-plasticity calculations.\n\nCyclic transformation presents another active problem. Repeated interface motion can accumulate defects, shift transformation temperatures, and reduce recoverable strain in shape-memory alloys. In structural steels, service deformation can progressively transform retained austenite and alter local stress distributions. Models must therefore connect the initiation of individual transformation events with evolving microstructure and component-scale behavior rather than treating martensite as a fixed constituent with one invariant set of properties."
    }
  ],
  "links": [
    "diffusion",
    "phase-transition",
    "metallurgy",
    "crystal-structure",
    "hydrogen-embrittlement",
    "electron-microscope",
    "x-ray-crystallography",
    "wear",
    "laser",
    "automobile",
    "3d-printing",
    "actuator"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}