6.1 — Aircraft Materials — Ferrous

Introduction to Ferrous Materials

Ferrous materials are metals and alloys whose primary constituent is iron (Fe). The word "ferrous" comes from the Latin ferrum, meaning iron. Ferrous materials are characterised by their high strength, hardness, and — with the notable exception of stainless steel — their susceptibility to corrosion (rust). In aircraft construction, ferrous materials are used selectively where their superior strength and wear resistance are essential, despite their relatively high density compared to aluminium alloys.

The most important ferrous material in aviation is steel — an alloy of iron and carbon (typically 0.05% to 2.0% carbon by weight). By adding other alloying elements and applying specific heat treatments, the properties of steel can be tailored to a wide range of requirements.

Why an Aircraft Uses Steel at All

The first thing most students are told about steel is that it is heavy, and the density figure listed among its properties later in this note reinforces that impression. Taken on its own it is misleading, and it leads to the wrong mental model of when a designer will reach for a ferrous material. Weight per unit of volume is not the quantity that matters in a structure; weight per unit of load carried is. That is called specific strength, and it is simply the ultimate tensile strength divided by the density.

Material	Typical UTS	Density	Specific strength (UTS / density)	Specific stiffness (E / density)
Low-alloy steel, heat treated (e.g. 4340)	about 1,800 MPa	about 7,850 kg/m³	about 229 kN·m/kg	about 26 MN·m/kg
High-strength aluminium alloy	about 570 MPa	about 2,810 kg/m³	about 203 kN·m/kg	about 26 MN·m/kg
Titanium alloy	about 950 MPa	about 4,430 kg/m³	about 214 kN·m/kg	about 26 MN·m/kg

Two results in that table are worth pausing over, because both are counter-intuitive. First, a properly heat-treated low-alloy steel has a higher specific strength than a high-strength aluminium alloy, not a lower one. Per kilogram of material, the steel carries more tensile load. It is not a heavy material in the sense that matters to a designer; it is a dense material, which is a different thing. Second, the specific stiffness of steel, aluminium and titanium is essentially identical at around 26 MN·m/kg. Young's modulus and density rise together as you move between the common structural metals, so you cannot make a stiffness-critical part lighter simply by changing which metal it is made from. You change the section instead — and that is exactly why a wing skin is aluminium: for the same mass you get a thicker sheet, and a thicker sheet resists buckling far better.

So the real reasons an aircraft carries ferrous parts are these:

• Load in a small envelope. A landing gear leg must absorb an enormous load inside a bay that has to close. Steel puts the required strength into the smallest volume of any common airframe material, and volume, not mass, is what constrains a gear bay, a bearing housing or a bolt hole.
• Bearing and wear surfaces. Nothing else available at reasonable cost survives repeated rolling or sliding contact the way a hardened steel surface does. Ball races, gear teeth, cam faces, bushes and pins are ferrous almost without exception.
• Temperature. Aluminium alloys lose useful strength above roughly 150 °C. Steels, and stainless steels in particular, keep working far beyond that, which is why firewalls, exhaust systems, engine bay structure and hot-air ducting are ferrous.
• Stiffness in a slender member. Because Young's modulus is roughly three times that of aluminium, a steel rod of a given diameter is three times as stiff. For a long, thin, tension-loaded member such as a control cable or a tie rod, that matters more than density.
• Weldability and repairability. A welded steel tube fuselage or engine mount can be repaired with equipment available in a small workshop, which is why the light aircraft fleet still uses it.
• Cost. Steel is the cheapest structural metal by a wide margin, so it is the default anywhere the weight penalty is acceptable.

Where the Engineer Actually Meets Ferrous Materials

Aircraft area	Typical ferrous items	Why ferrous
Landing gear	Shock strut cylinders and pistons, side stays, drag braces, axles, torque links, pins and bushes	Very high load in a confined volume; wear surfaces; the highest-strength steels in the aircraft live here
Powerplant installation	Engine mounts, mount bolts, firewall, exhaust stacks and shrouds, turbocharger and bleed ducting, clamps	Strength plus resistance to heat, vibration and exhaust gas
Engine internals	Crankshafts, connecting rods, camshafts, gears, valve springs, bearings, turbine shafts and discs	Fatigue strength, wear resistance and hot strength
Flight controls	Control cables, turnbuckles, bellcrank bushes, push-pull rod ends, hinge pins	Tensile strength, stiffness and wear resistance in small sections
Structure and mechanisms	Flap and slat tracks and carriages, wing attachment fittings, door latches and hinges, seat tracks and studs	Concentrated point loads that would need an impractical volume of aluminium
Fasteners	Bolts, studs, screws, nuts, self-locking inserts, pins, some rivets	High clamp-up and shear capacity in a small diameter
Systems hardware	Hydraulic tubing in high-pressure zones, actuator rods, springs, valve internals, bearing races	Pressure containment, elasticity and wear resistance

The Ferrous Family

"Ferrous" is a family, not a single material, and the members differ enormously. Sorting them out by carbon content is the quickest way to hold the family in your head, because carbon content is what separates them.

Material	Approximate carbon	Character	Aircraft relevance
Pure iron	essentially none	Soft, very ductile, weak, rusts freely, strongly magnetic	None structurally; used for magnetic cores and shielding
Wrought iron	below about 0.08%	Almost pure iron containing threads of slag; tough, easily forged and welded, notably corrosion resistant	Obsolete. Found only on historic and vintage airframes
Plain carbon steel	0.05% to about 1.5%	Properties set almost entirely by carbon content and heat treatment	Tubing, brackets, general hardware, springs, tools
Alloy steel	usually 0.1% to 0.5%, though the tool grades and the through-hardening bearing grades run higher	Deliberate additions of chromium, nickel, molybdenum and others to control strength, toughness, hardenability and hot strength	The workhorse of highly stressed aircraft parts
Stainless steel	as little as 0.03% in the low-carbon welding grades, up to about 1.2% in the hardenable ones	At least 10.5% chromium, giving a self-repairing passive film	Firewalls, exhaust, fasteners, cable, hot and wet zones
Tool steel	0.6% to about 1.5%	Very hard and wear resistant after heat treatment; brittle	Workshop tooling, drills, reamers, dies; not airframe parts
Cast iron	2.0% to about 4.5%	Cheap, easily cast, excellent in compression, brittle in tension, good vibration damping	Effectively absent from airframes; some piston engine components and ground equipment

Note where the boundary between steel and cast iron falls. Above roughly 2% carbon the iron can no longer dissolve all of the carbon at any temperature, so the excess appears as a separate constituent that cannot be removed by heat treatment. That is the physical reason steel is defined as an iron-carbon alloy of up to about 2% carbon: it is the practical limit of the solid solution, not an arbitrary line.

Why Pure Iron Is Not Used

Iron on its own is a poor structural material. It is soft, extremely ductile, and its tensile strength is only a small fraction of that of even a mild steel. It cannot be hardened by heat treatment either, because there is no carbon present to form the hard constituents that hardening depends upon. Note carefully that the objection to pure iron is lack of strength, not brittleness — pure iron is the opposite of brittle. It bends and stays bent. Adding a fraction of one per cent of carbon transforms it, and adding the ability to heat treat it transforms it again.

How Steel Is Made

Understanding the manufacturing route explains several things an engineer meets later: why aerospace steel costs many times what commercial bar costs, why a mill certificate quotes a cast or heat number, and why some specifications insist on a particular melting process.

• Ironmaking. Iron ore, coke and limestone are charged into a blast furnace. The coke reduces the ore to metallic iron and, in doing so, saturates it with carbon. The product, pig iron, carries roughly 4% carbon along with silicon, sulphur, phosphorus and manganese. It is hard, weak and brittle, and useless as it stands.
• Steelmaking — removing carbon. The molten pig iron is refined by blowing air or, in the modern basic oxygen process, pure oxygen through or on to it. The oxygen burns out the excess carbon as carbon monoxide, and also oxidises silicon, manganese and phosphorus into a slag. The essential operation of steelmaking is the removal of carbon, taking it down from about 4% to whatever the specification calls for. An electric arc furnace does the same job starting mainly from scrap.
• Deoxidation and alloying. Oxygen dissolved in the bath would form blowholes on solidification, so it is removed with aluminium or silicon. A fully deoxidised steel is called killed; the aerospace grades are always killed. Alloying elements are added at this stage.
• Secondary refining. For highly stressed aerospace parts, ordinary air-melted steel is not good enough. Non-metallic inclusions left over from melting act as internal notches and become fatigue crack starters. Premium routes such as vacuum induction melting, vacuum arc remelting and electroslag remelting re-melt the steel under vacuum or under a refining slag to strip out those inclusions and dissolved gases. The improvement shows up mainly in transverse ductility and fatigue life rather than in the tensile strength figure, which is precisely why a specification, and not the tensile test alone, has to control it.
• Forming. The steel is cast, then hot rolled, forged, drawn or extruded into bar, sheet, tube or forgings.

Four Words You Need Before Going Further

• Alloy — a metallic material made from two or more elements, of which at least one is a metal. Steel is an alloy of iron and carbon even though carbon is not a metal.
• Solid solution — one element dissolved into the crystal lattice of another while both remain solid, in the same sense that salt dissolves in water. The dissolved atoms are distributed uniformly and do not form a separate substance.
• Phase — a region of material that is uniform in both composition and crystal structure. A steel at room temperature normally contains two phases at once, and knowing which they are is the whole basis of heat treatment.
• Grain — a single crystal within the metal, typically a few hundredths of a millimetre across. A solid metal is a mosaic of grains, each with its lattice pointing in a different direction, meeting at grain boundaries. Grain size turns out to be one of the few things that improves strength and toughness at the same time.

The Iron-Carbon System

Everything that follows in this note — why carbon content decides a steel's character, why heat treatment works at all, why one steel can be hardened and another cannot — comes out of a single piece of physics: iron changes its crystal structure when it is heated, and the two structures dissolve wildly different amounts of carbon. Learn this section properly and the rest of the material stops being a list to memorise.

Iron Changes Its Crystal Structure With Temperature

Most metals keep one crystal structure from room temperature up to melting. Iron does not. It is allotropic, meaning it exists in more than one crystalline form, and it swaps between them at fixed temperatures on the way up:

Temperature range (pure iron)	Name	Crystal structure	Atoms per unit cell	Carbon it can dissolve
Room temperature to 910 °C	Alpha iron (ferrite)	Body-centred cubic (BCC)	Corner atoms plus one at the centre of the cube	Almost none — about 0.02% at best
910 °C to 1,394 °C	Gamma iron (austenite)	Face-centred cubic (FCC)	Corner atoms plus one at the centre of each face	Up to about 2.0%
1,394 °C to 1,538 °C	Delta iron	Body-centred cubic (BCC)	As alpha iron	Very little; of interest only in casting and welding
Above 1,538 °C	Liquid	None	—	Unlimited within the alloy range

The face-centred cubic structure is the more tightly packed of the two, which sounds as though it ought to have less room for carbon, not more. The resolution is that what matters is not how much empty volume there is in total, but how big the individual holes between the iron atoms are. FCC iron has fewer holes but each is larger, and a carbon atom fits into one of them. BCC iron has more holes but every one of them is too small, so a carbon atom forced into the BCC lattice has to push the surrounding iron atoms apart. That single geometric fact drives the whole of steel heat treatment.

The Constituents of Steel

The names below recur throughout the rest of this note, so learn them here. Each is a distinct constituent with its own structure and its own mechanical character.

Constituent	What it is	Crystal structure	Carbon content	Hardness and character	Magnetic?
Ferrite (alpha iron)	Almost pure iron; a solid solution holding a trace of carbon	Body-centred cubic	up to about 0.02%	Soft, weak, very ductile	Yes, strongly
Austenite (gamma iron)	Carbon fully dissolved in iron as a single uniform solid solution	Face-centred cubic	up to about 2.0%	Soft and very ductile at temperature; the starting point of every hardening treatment	No
Cementite (iron carbide, Fe3C)	A chemical compound, not a solution: three iron atoms combined with one carbon atom	Orthorhombic compound	6.67% by weight, fixed	Extremely hard, extremely brittle; the hard constituent of ordinary steel	Weakly, below about 210 °C
Pearlite	Not a phase but a mixture: alternating thin plates of ferrite and cementite	Lamellar (layered) mixture	0.83% overall	Strong without being brittle — the ferrite carries the ductility, the cementite the strength	Yes
Martensite	Carbon trapped in a distorted iron lattice by cooling too fast for it to escape	Body-centred tetragonal — a BCC cell stretched along one axis	Whatever the steel contained	The hardest constituent; also the most brittle and the most highly stressed	Yes
Bainite	Fine needles of ferrite with carbide between them, formed at cooling rates between those giving pearlite and martensite	Acicular (needle-like)	Whatever the steel contained	Hard, and notably tough for its hardness — a useful compromise between pearlite and martensite	Yes

Two of those entries are the ones examiners return to again and again, so state them to yourself precisely. Austenite is carbon fully dissolved in iron and uniformly distributed in solid solution. Cementite is one carbon atom chemically combined with three iron atoms, written Fe3C. And pearlite is neither: it is cementite laminated with ferrite in alternate layers, so it combines the properties of ferrite and cementite rather than being a substance in its own right.

Critical Points and Thermal Arrests

Heat a piece of plain carbon steel steadily and plot its temperature against time. The line does not rise smoothly. At certain temperatures it pauses, or at least slows markedly, even though the furnace is still supplying heat at the same rate. These pauses are called thermal arrests or critical points, and each marks a change of crystal structure. The heat going in is being consumed by the transformation itself — the latent heat of the change — instead of raising the temperature.

• Lower critical point (LCP), also called A1. The first arrest, at about 723 °C. Here pearlite begins to transform into austenite. It sits at 723 °C for every plain carbon steel regardless of carbon content, which makes it a useful fixed landmark.
• Upper critical point (UCP), also called A3. The second arrest, reached after a further temperature rise. Here the last of the remaining ferrite finishes dissolving and the steel becomes fully austenitic. Unlike the LCP it moves with carbon content: 910 °C in pure iron, falling steadily as carbon is added until it meets the lower critical line at 723 °C.
• The magnetic change point (A2), about 770 °C. Iron loses its ferromagnetism here without any change of crystal structure. It matters practically because a steel part heated above this temperature cannot be checked or inspected magnetically until it has cooled.

On a low-carbon steel, the interval between the two critical points is substantial — the temperature climbs a further 60 to 160 °C or so above the first arrest before the second one is reached, the gap being widest at the leanest carbon contents and closing steadily as carbon is added. Up to the pearlitic composition it can never be larger than the 187 °C between 910 °C and 723 °C, and it approaches that figure only in the limit of no carbon at all, where there is no pearlite to produce a first arrest. On a steel at the pearlitic composition the two arrests coincide and there is only one.

Reading the Iron-Carbon Diagram

Plot temperature on the vertical axis and carbon content on the horizontal, and the critical points trace out lines. This is the iron-carbon diagram, and three points on it carry names you must be able to separate.

Point	Composition	Temperature	What happens there	Where it matters
Eutectoid point	0.83% carbon (commonly rounded to 0.8%)	about 723 °C	Solid austenite transforms directly into two solids at once — ferrite and cementite — producing 100% pearlite. This is where the upper and lower critical lines meet	Steel. The reference composition for the whole of steel heat treatment
Eutectic point	4.3% carbon	about 1,147 °C	Liquid transforms directly into two solids at once, at the lowest melting temperature of the whole system	Cast iron. Far outside the steel range
Maximum solubility in austenite	about 2.0% carbon	about 1,147 °C	The most carbon austenite will ever dissolve. Beyond it, free carbide is present at every temperature	The dividing line between steel and cast iron

Eutectoid and eutectic are different words for different things, and the difference is worth fixing firmly. The tell is what the material is before the transformation. A eutectic change starts from a liquid: liquid becomes two solids. A eutectoid change starts from a solid: one solid becomes two different solids. The suffix -oid means "like" — a eutectoid reaction is eutectic-like but happens entirely in the solid state. In the iron-carbon system the eutectoid sits at 0.83% carbon and 723 °C and belongs to steel; the eutectic sits at 4.3% carbon and 1,147 °C and belongs to cast iron. A steel never reaches its eutectic composition, so an engineer working on airframes deals with the eutectoid and only the eutectoid.

What the Carbon Content Does to the Room-Temperature Structure

Cool a steel slowly from the austenitic range and the structure it ends up with depends entirely on how much carbon it started with, measured against that eutectoid composition of 0.83%.

Class	Carbon	Slow-cooled structure	Consequence
Hypo-eutectoid (below the eutectoid)	under 0.83%	Grains of free ferrite, with pearlite filling the space between them. The lower the carbon, the more free ferrite	Soft, ductile, tough, weldable. Nearly every structural and fastener steel lives here
Eutectoid	0.83%	100% pearlite and nothing else	The maximum strength obtainable in the slow-cooled condition, with usable toughness
Hyper-eutectoid (above the eutectoid)	0.83% to about 2.0%	Pearlite with a network of free cementite along the grain boundaries	Very hard and wear resistant, but the brittle carbide network makes it weak in shock. Tool and bearing steels

The reason 0.83% carbon is the composition that produces a wholly pearlitic structure is now visible. Below it there is more iron than the carbon can pair with, and the surplus iron is left over as free ferrite. Above it there is more carbon than the iron can absorb into the pearlite, and the surplus appears as free cementite. Only at the eutectoid composition does everything pair up.

Why Fast Cooling Produces Something Different

Every transformation described so far assumes cooling slowly enough for atoms to move. Carbon atoms escaping from austenite have to diffuse through the lattice to reach the places where cementite is forming, and diffusion takes time. Take that time away and the transformation cannot happen as written.

• Cool slowly (furnace). The carbon has all the time it needs. It forms coarse pearlite with thick, widely spaced plates. The structure is soft and very ductile.
• Cool moderately (still air). The carbon has less time, so it travels shorter distances before it is trapped. The pearlite plates are finer and more closely spaced. Finer pearlite is harder and stronger than coarse pearlite of the same composition, because the closely spaced hard plates obstruct deformation more effectively.
• Cool fast enough (quench). Once the cooling is faster than a certain rate — the critical cooling rate for that steel — the carbon cannot diffuse at all. The face-centred cubic austenite still has to become body-centred cubic because that is what iron does at low temperature, but it must do so with all its carbon still inside it. It shears over into a distorted, stretched body-centred cell called body-centred tetragonal, and that structure is martensite.

Martensite is hard for a specific and satisfying reason. The trapped carbon atoms wedge the lattice open, and a lattice under that much internal strain resists any further slipping. Hardness is resistance to permanent deformation, and permanent deformation in a metal happens by planes of atoms sliding over one another, so a lattice that cannot slip is a hard one. The same strain is why martensite is brittle and why it is left full of internal stress: nothing about the structure has relaxed.

Two consequences follow directly, and both are examinable:

• The hardness of martensite depends on the carbon content, not on the alloying. More carbon means more lattice distortion means more hardness. Hardness climbs steeply with carbon up to roughly 0.6% and then flattens off, so past that point extra carbon buys wear resistance and brittleness rather than more hardness.
• A steel with too little carbon cannot be usefully quench hardened. A mild steel of 0.15% carbon forms a martensite so lightly strained that it is barely harder than the pearlite it replaced. This is the reason a low-carbon component that needs a hard surface has to have carbon put into that surface first — the process covered later under case hardening.

Transformation Is a Race Against Time

It helps to picture the choice between pearlite, bainite and martensite as a race. On cooling, the austenite is unstable and will transform into something; which product it gets depends on how much time it is given at each temperature on the way down.

• Held just below the lower critical point, where diffusion is easy but the driving force is small, austenite forms coarse pearlite slowly.
• Held a couple of hundred degrees lower, where the driving force is large and diffusion still just possible, it forms fine pearlite quickly. This is the fastest region of all, and it is why a quench has to get the part past this temperature band without lingering.
• Lower again, where diffusion is sluggish, it forms bainite — a fine mixture that is hard and unusually tough.
• Below a temperature called the martensite start temperature, no diffusion is possible and the remaining austenite shears over into martensite instantly, without any time dependence at all. Transformation continues only as the temperature keeps falling and stops at the martensite finish temperature.

Two practical results come out of this picture. First, alloying elements slow diffusion down, which delays the formation of pearlite and lets a slower, gentler quench still reach martensite — the property called hardenability, dealt with fully in the heat treatment section. Second, the martensite finish temperature is not necessarily above room temperature, so a quench does not necessarily transform everything — the untransformed remainder is called retained austenite, and what it does to a component is dealt with in the heat treatment section along with the treatment used to remove it.

Properties of Ferrous Materials

Property	Description
Tensile strength	Resistance to being pulled apart — steel has very high tensile strength, typically 400–2000 MPa depending on alloy and treatment
Hardness	Resistance to surface indentation or scratching — can be greatly increased by heat treatment and alloying
Ductility	Ability to be drawn into wire or deformed without fracturing — low-carbon steels are very ductile, high-carbon steels less so
Malleability	Ability to be hammered or rolled into shape without cracking
Toughness	Ability to absorb energy and deform plastically before fracturing — a combination of strength and ductility
Elasticity	Ability to return to original shape after load is removed — steel has a well-defined elastic limit
Fatigue strength	Resistance to failure under repeated cyclic loading — critical in aircraft structures
Corrosion resistance	Plain carbon steels rust readily; stainless steels and surface treatments improve corrosion resistance
Density	Approximately 7,850 kg/m³ — about 2.8 times heavier than aluminium
Stiffness (Young's modulus)	Resistance to elastic deflection under load, quite separate from strength — close to 200 GPa for practically every steel, and essentially unaffected by alloying or heat treatment
Yield strength	The stress at which permanent deformation begins; the practical design limit, since a part that has yielded is out of tolerance even though it has not broken
Brittleness	The opposite of toughness — fracture with little or no plastic deformation and little energy absorbed. Rises with carbon content, with hardness, and as temperature falls
Resilience	The energy a material can absorb and give back within the elastic range — the property a spring is chosen for, and not the same thing as toughness
Hardenability	The depth to which a steel will harden on quenching, as distinct from the hardness value reached. Set mainly by the alloying elements, not by carbon
Wear resistance	Resistance to material loss by rubbing, rolling or abrasive contact; broadly tracks surface hardness, which is why bearing and gear surfaces are case hardened
Creep resistance	Alone among the properties here this one is time-dependent: a hot, steadily loaded part can be within every limit on the day it is fitted and outside them years later. It governs exhaust, turbine and hot-zone fastener life
Machinability	The ease with which the material can be cut, and the surface finish and tool life obtained. Best in the medium-hardness range; both very soft and very hard steels machine poorly
Weldability	The ability to be joined by fusion without cracking or forming a brittle heat-affected zone. Falls sharply as carbon and alloy content rise

Stress and Strain

Most of the entries in the table above are mechanical properties, and a mechanical property is defined in terms of stress and strain, so the two definitions have to be exact.

• Stress is force divided by the cross-sectional area carrying it. In SI units it is newtons per square millimetre, which is the same as the megapascal. Stress is what the material feels; two rods carrying the same load are at different stresses if their diameters differ.
• Strain is the change in length divided by the original length. It is a ratio, so it has no units, and it is often quoted as a percentage.

Three kinds of direct stress appear on an aircraft — tensile (pulling apart), compressive (pushing together) and shear (sliding one plane over the next) — together with the combinations engineers call bending and torsion. Steel behaves almost identically in tension and compression up to the yield point, which is not true of every material: cast iron, for example, is several times stronger in compression than in tension, and that asymmetry is why it is used for housings but never for a tie rod.

Elasticity and Stiffness Are Not the Same Property

This distinction causes more confusion than any other in the property list, and getting it clear pays for itself repeatedly.

Elasticity is the ability to return to the original shape when the load is removed. Stiffness, measured by Young's modulus, is how far the material moves while it is loaded. A rubber band is highly elastic and extremely un-stiff. A steel bar is elastic over a much smaller range but is enormously stiffer within it.

Within the elastic range, stress and strain are proportional. That proportionality is Hooke's law, and the constant of proportionality is Young's modulus:

Now the point that matters, and it is one of the most useful single facts in this module: Young's modulus for steel is essentially the same whatever the alloy and whatever the heat treatment. Mild steel, 4130, 4340 hardened to its maximum, and annealed stainless all sit within a few per cent of 200 GPa. Austenitic stainless grades sit slightly lower, but nowhere near enough to change the conclusion. Heat treatment changes how much load the steel can take before it deforms permanently. It does not change how much it deflects while still elastic.

The maintenance consequence is direct. If a control run feels spongy, or a bracket flexes more than the drawing allows, fitting a part in a stronger steel will not help. Only a larger section, a shorter unsupported length, or an extra support will. Conversely, if a part is yielding — taking a permanent set — then strength is exactly what is short, and heat treatment or a stronger grade is the right answer.

Hardness, and What It Really Tells You

Hardness is resistance to being permanently indented. What makes it so useful is that it is measured on the finished part, in minutes, without destroying anything, and for steels it correlates closely with tensile strength. Both properties depend on the same underlying thing: how strongly the crystal structure resists planes of atoms sliding over one another.

This correlation is why a hardness check is the standard way of confirming that a heat treatment has been carried out correctly. It is also why an unexpectedly low hardness reading on a highly stressed steel part is treated so seriously: it means the part is not at the strength the design assumed.

Ductility and Malleability

Both describe the ability to deform permanently without cracking, but they describe it under opposite kinds of load, and the exam distinguishes them.

• Ductility is deformation under tension — the property that lets a metal be drawn out into wire. It is measured by percentage elongation and by reduction of area in a tensile test.
• Malleability is deformation under compression — the property that lets a metal be hammered, rolled or pressed into shape without cracking.

The two usually go together in steels but they are not identical: lead is highly malleable and only moderately ductile. Both fall as carbon content rises and as hardness rises, which is why annealing — the treatment that produces the softest condition — is the one that makes the metal easiest to work.

Toughness, Brittleness and the Trade-Off That Governs Steel Selection

Toughness is the ability to absorb energy before fracturing. That is a different question from "how much load can it take", and the difference is easiest to see by noticing that a fully hardened, untempered steel is very strong and very hard yet has almost no toughness at all: it takes a large load, but a sharp knock shatters it. Energy absorbed is load multiplied by the distance moved, so a material that fractures without deforming absorbs almost none, however high the load was.

Brittleness is simply the absence of toughness. A brittle fracture happens suddenly, at right angles to the load, with a bright crystalline appearance and no measurable stretching. A ductile fracture is preceded by visible necking and leaves a dull, fibrous, cup-and-cone surface.

The governing rule for steel selection follows: strength and toughness generally trade against one another. Every treatment that raises hardness and tensile strength lowers the energy the steel will absorb before it breaks, and vice versa. The engineer's job is not to maximise either but to select the point on that curve the application needs. There is one important exception to the trade-off, covered in the heat treatment section: refining the grain size raises strength and toughness together.

Resilience — the Property a Spring Needs

Resilience is the energy a material stores and returns while remaining elastic. On a stress-strain diagram it is the area under the elastic part of the curve only, whereas toughness is the area under the whole curve to fracture. A spring steel is chosen for high resilience: a high yield strength combined with the ordinary steel value of Young's modulus means it can be deflected a long way, store a lot of energy, and give all of it back. Spring behaviour itself, including rate and the effect of coil geometry, belongs with the springs material later in this module.

Fatigue and Creep in Service

Two of the properties in the table above describe failure under conditions where the load never approaches the material's static strength, and between them they account for the majority of metallic failures on aircraft.

Fatigue is progressive cracking under repeated or alternating load. The defining and slightly alarming feature is that it occurs at stresses well below the ultimate tensile strength — often below the yield strength — so a part can fail after enough cycles without ever having been overloaded in the ordinary sense. It is not a single-overload failure and it is not an impact failure; it is a reduction in load-carrying capacity caused by the cycling itself. This is why aircraft components carry life limits expressed in flight cycles and hours rather than in load.

Creep is slow, continuous, permanent stretching under a steady load at elevated temperature. Time and temperature do the work that stress alone could not. It becomes a design consideration in steels above roughly four tenths of the absolute melting temperature, which for steel means the hot end of an engine installation: turbine discs and blades, shafts, exhaust components and the fasteners holding them. Creep proceeds in three distinct stages:

Stage	What is happening	Rate of strain	Engineering response
Primary	Initial rapid extension that then slows as the material work hardens	Decreasing	Accepted; allowed for in the design clearances
Secondary	Steady state — work hardening and thermal recovery are in balance. The longest stage by far	Constant and minimum	This is the design regime. The permitted life is calculated from this rate
Tertiary	Internal voids link up and the section necks down, so the true stress climbs and accelerates the process further	Rapidly increasing	Irreversible and self-accelerating. Rupture follows quickly

How Treatments Move the Properties

The single most useful thing to hold in your head is the direction each process moves each property. Each row below states its starting condition, because a claim such as "annealing softens the steel" only means something relative to what it started as.

Change, and from what starting state	Hardness and tensile strength	Ductility	Toughness	Internal (residual) stress	Grain
More carbon (comparing two steels given the same treatment)	Higher	Lower	Lower	No direct effect	No direct effect
Quench hardening, from the normalised condition	Much higher	Much lower	Much lower	Much higher	Unchanged in size; structure becomes martensitic
Tempering, from the as-quenched condition	Slightly lower	Higher	Much higher	Much lower	Unchanged
Full annealing, from any condition	Lowest obtainable	Highest obtainable	Moderate — limited by the coarse structure produced	Lowest obtainable	Coarse
Normalising, from a coarse as-forged or as-welded condition	Moderately higher than annealed	Good	Higher than annealed	Much lower than as-forged	Refined and uniform
Cold working, from the annealed condition	Higher	Lower	Lower	Higher	Distorted and elongated in the working direction
Grain refinement at constant composition	Higher	Little change	Higher	No direct effect	Finer — the one change that improves strength and toughness together
Case hardening, from a low-carbon base	Much higher at the surface only	Lower at the surface; core unchanged	Core unchanged — that is the point of the process	Surface left in useful compression	Unchanged in the core

Physical Properties and What They Cost You in Practice

Physical property	Typical value for steel	Practical consequence
Melting point	about 1,400 to 1,540 °C depending on carbon content	Comfortably above any airframe temperature, but relevant to welding, brazing and fire resistance requirements
Thermal conductivity	about 50 W/m·K for carbon steel; about 15 for austenitic stainless	Heat generated at a cutting edge stays at the cutting edge in stainless. This is a direct cause of the drilling difficulties covered later
Coefficient of linear expansion	about 12 microns/m/°C for carbon steel; about 17 for austenitic stainless	A steel part fitted into an aluminium structure sees its clearance change with temperature, because the aluminium expands about twice as much
Specific heat capacity	about 480 J/kg·K — roughly half that of aluminium	Together with the conductivity above, it governs how long a part takes to heat through in a furnace and how steeply the temperature varies through a thick section during a quench
Electrical resistivity	roughly ten times that of copper for carbon steel, and several times higher again for austenitic stainless	Steel is a poor conductor, so a steel structure is a poor bonding path and steel is never used as a primary electrical conductor
Magnetic permeability	High for ferrite and martensite; effectively that of air for austenite	Enables magnetic particle inspection and the magnet identification check; forces ferrous items to be kept out of compass-safe zones and demands demagnetisation after magnetic inspection

The expansion figures deserve one extra sentence, because they explain a family of problems that look unrelated. Steel expands roughly half as much as aluminium alloy for the same temperature rise. Where the two are bolted together over a long run — a steel fitting on an aluminium spar, or a steel cable strung along an aluminium fuselage — the aluminium moves further, and the joint or the run has to accommodate the difference. Corrosion resistance is the remaining property whose consequences reach furthest, and it is dealt with in full with the corrosion material later in this module; here it is enough to say that plain carbon and low-alloy steels have essentially none of it and depend entirely on a protective scheme, while the stainless grades carry their own.

Types of Steel

Plain Carbon Steel

Plain carbon steel contains iron and carbon with only small amounts of other elements (manganese, silicon, sulphur, phosphorus). It is classified by carbon content:

Type	Carbon %	Properties	Aircraft Applications
Low carbon (mild steel)	0.05–0.30%	Soft, ductile, easily welded; low strength	General fittings, non-structural brackets, wire, welded tube structures (some light aircraft fuselages)
Medium carbon	0.30–0.50%	Harder, stronger; responds to heat treatment	Bolts, studs, axles, forgings
High carbon	0.50–1.50%	Very hard after heat treatment; brittle; holds sharp edge	Springs, cutting tools, wire rope; not common in primary structures
Dead mild (ultra-low carbon)	below 0.15%	Extremely ductile; folds and deep-draws without cracking; cannot be usefully quench hardened because there is too little carbon to form a hard martensite	Pressed and formed sheet parts, shims, clips, and the base stock for components that will be carburised
Free-cutting (resulphurised)	0.08–0.45%, with sulphur deliberately raised	Machines cleanly and breaks its own chips, but the sulphide stringers act as internal notches and cut ductility and fatigue strength across the grain	Shop tooling and non-structural items only — specifically excluded from highly stressed aircraft parts

What Each Carbon Range Can and Cannot Do

The classification is only useful if you know what it buys you, so take one grade from each band and follow it through:

• A 0.20% carbon steel can be cold bent around a tight radius, welded with no preheat and no post-weld treatment, and formed on a folder. Quench it and it will barely change hardness — there is not enough carbon to distort the lattice. If a part in this material needs a hard surface, carbon has to be diffused into the surface first.
• A 0.45% carbon steel responds properly to quenching and tempering, reaching a useful hardness through the section on small parts. It can still be welded, but a preheat and a controlled cooling rate are needed or the heat-affected zone hardens and cracks. This is the band that bolts, studs, axles and shafts come from.
• A 0.95% carbon steel hardens to a file-hard condition and holds a cutting edge, but it snaps rather than bends, it cannot practically be welded, and it is used only where hardness is the whole point — springs, cutting tools, and wire that must resist wear.

Notice that the direction of every property is the same as carbon rises: attainable hardness, tensile strength and wear resistance all go up; ductility, malleability, toughness, weldability and formability all go down. There is no carbon content that is good at everything, which is exactly why alloying elements exist — they let a designer move one property without paying the full price in the others.

Alloy Steels

Alloy steels have deliberate additions of other elements to improve specific properties. Common alloying elements in aircraft steels:

Element	Effect
Chromium (Cr)	Increases hardness, wear resistance, and corrosion resistance; key element in stainless steel (≥10.5% Cr)
Nickel (Ni)	Increases toughness and impact resistance; improves low-temperature properties; corrosion resistance
Molybdenum (Mo)	Increases strength at elevated temperatures; reduces temper brittleness; improves hardenability
Vanadium (V)	Refines grain structure; increases strength, toughness, and wear resistance
Tungsten (W)	Increases hardness at high temperatures; used in high-speed tool steels
Manganese (Mn)	Increases strength and hardness; improves hardenability; counteracts sulphur brittleness
Silicon (Si)	Improves strength and elasticity; used in spring steels
Cobalt (Co)	Raises hot hardness and is the element that lets a tool steel keep its edge at the highest cutting temperatures; also a constituent of maraging steel. Unusually among the alloying elements it does not improve hardenability
Boron (B)	In quantities of a few thousandths of one per cent it produces a large increase in hardenability, allowing a leaner and cheaper steel to harden through a given section
Aluminium (Al)	A deoxidiser and grain refiner in ordinary steels; in nitriding grades it is the essential element, because it forms the extremely hard aluminium nitrides that give a nitrided case its hardness
Titanium (Ti) and Niobium (Nb)	Strong carbide formers used to "stabilise" stainless steel: they lock up the carbon so that chromium carbides cannot form at the grain boundaries during welding, which is what preserves corrosion resistance in a weld zone
Sulphur (S)	Normally an unwanted impurity that causes cracking during hot working. Added deliberately only in free-cutting steels to improve machinability, at the cost of ductility and fatigue strength across the grain
Phosphorus (P)	An impurity that raises strength slightly but embrittles the steel, particularly at low temperature. Aerospace specifications set a tight maximum on it
Copper (Cu)	Gives a modest improvement in atmospheric corrosion resistance and, in the precipitation-hardening stainless grades, forms the fine particles that produce the strengthening
Lead (Pb)	Added only to improve machinability in general engineering steels; not used in aircraft structural grades

How an Alloying Element Actually Works

The table above lists what each element does. Understanding how it does it turns a list into something you can reason with, and it explains several results that otherwise have to be memorised.

An added element has to go somewhere in the iron crystal, and there are only two places available:

• Substitutional. The atom is roughly the same size as an iron atom, so it simply takes an iron atom's place in the lattice. Chromium, nickel, manganese, molybdenum, silicon, tungsten, vanadium and cobalt all behave this way. Because the fit is reasonably good, the lattice is only mildly distorted, and useful effects need percentage-level additions.
• Interstitial. The atom is much smaller and squeezes into the gaps between the iron atoms. Carbon, nitrogen and boron do this. The fit is poor, so the lattice is heavily distorted, and a very small amount produces a large effect. This is the reason a tenth of one per cent of carbon changes a steel more than one per cent of nickel does, and the reason carbon and nitrogen are the elements used for surface hardening.

From there, alloying elements do three distinct jobs:

• They strengthen the solid solution. Any foreign atom distorts the lattice around it and makes slip harder. This is a modest, general effect.
• They form carbides. Chromium, molybdenum, vanadium, tungsten, titanium and niobium bond with carbon more readily than iron does, producing carbides that are harder and far more stable at temperature than plain iron carbide. Hard carbides raise hardness and wear resistance directly; stable carbides are what let a tool steel keep its edge when it is running hot, and what give the alloy steels their resistance to softening and to creep.
• They change when and how the steel transforms. This is the most important effect and the least obvious. Alloying elements slow down the diffusion that pearlite formation depends on. Delay the pearlite and a slower quench still reaches martensite — which is hardenability, the depth of hardening. Molybdenum, chromium, manganese, nickel and boron all raise it.

Austenite Stabilisers and Ferrite Stabilisers

One further effect explains most of what you need to know about stainless steel, so it is worth stating on its own. Some elements widen the temperature range over which austenite is stable, and some narrow it.

• Austenite stabilisers — nickel, manganese, carbon, nitrogen. Adding these lowers the temperature at which austenite gives way to ferrite. Add enough and austenite survives all the way down to room temperature.
• Ferrite stabilisers — chromium, molybdenum, silicon, titanium, niobium, aluminium. Adding these raises that temperature, and enough of them removes the austenite field altogether so the steel is ferritic at every temperature up to melting.

Now work three real cases through, and the whole stainless family falls out of one mechanism:

• 17% chromium, very low carbon, no nickel. Chromium is a ferrite stabiliser and there is nothing to oppose it, so the steel stays body-centred cubic ferrite from room temperature upwards. It never becomes austenite on heating, so it can never be quenched to martensite. Result: magnetic, and not hardenable by heat treatment — only by cold work. This is the ferritic family.
• 18% chromium plus 8% nickel. The nickel is a strong enough austenite stabiliser to overcome the chromium and hold the face-centred cubic structure down to room temperature. Result: non-magnetic (austenite has no ferromagnetism), not hardenable by heat treatment (there is no transformation left to exploit — it is already austenite, so quenching it changes nothing), very ductile, very formable, and strengthened only by cold work. This is the austenitic family, and it is the 18-8 composition.
• 12 to 17% chromium with the carbon raised to 0.15% or more. Carbon is an austenite stabiliser, and enough of it re-opens the austenite field at high temperature. The steel can therefore be heated to austenite and quenched to martensite exactly like an ordinary carbon steel. Result: magnetic and hardenable by quench and temper, with corrosion resistance lower than the austenitic grades because much of the chromium is tied up as carbide. This is the martensitic family.

Two more consequences of the austenitic structure are worth carrying with you. Austenitic stainless expands about half as much again as carbon steel for the same temperature rise, and conducts heat only about a third as well, so a welded austenitic assembly distorts far more than an equivalent carbon-steel one. And austenite work hardens very rapidly, which is why machining it demands a particular technique — covered later under working and handling.

Important Aircraft Steel Alloys

Designation	Composition	Properties	Typical Aircraft Use
SAE 4130 (chromoly)	Cr-Mo steel (0.30% C, 1% Cr, 0.2% Mo)	Excellent strength-to-weight ratio; good weldability; responds well to heat treatment	Engine mounts, fuselage tubing (light aircraft), landing gear, structural fittings
SAE 4340	Ni-Cr-Mo steel (0.40% C, 0.8% Cr, 1.8% Ni, 0.25% Mo)	Very high strength (up to 1800 MPa heat-treated); excellent fatigue strength and toughness	Landing gear components, crankshafts, connecting rods, high-stress structural fittings
Stainless steel (18-8)	18% Cr, 8% Ni (austenitic type 304/321)	Excellent corrosion resistance; non-magnetic; good high-temperature properties	Exhaust systems, firewalls, high-temperature zones, fasteners
Maraging steel	18% Ni, plus Co, Mo, Ti	Ultra-high strength (up to 2400 MPa); tough; good fatigue resistance	Landing gear, high-performance structural parts
SAE 4140	Cr-Mo steel (0.40% C, 1% Cr, 0.2% Mo)	The higher-carbon partner to 4130: a greater attainable strength through a heavier section, but noticeably harder to weld	Machined fittings, shafts, axles, pins, torque links, and heavy-duty tooling
300M (modified 4340)	Ni-Cr-Mo with about 1.6% silicon and vanadium added (0.42% C)	Ultra-high strength, around 1,900–2,000 MPa, with better toughness at that strength than 4340; very notch-sensitive and demands an intact protective finish	Main landing gear cylinders, pistons and axles on large transport aircraft
SAE 8620	Ni-Cr-Mo case-hardening steel (0.20% C)	Low-carbon core stays tough while the carburised surface hardens; good hardenability for its cost	Gears, splines, cam followers, bushes and other wear surfaces
SAE 9310	Ni-Cr-Mo carburising steel with higher nickel (0.10% C)	Premium aerospace gear steel; exceptional core toughness under a hard case, and normally supplied vacuum melted	Engine and gearbox gears, accessory drive gears, splined shafts
SAE 52100	High-carbon chromium bearing steel (about 1% C, 1.45% Cr)	Through-hardens to a very high, uniform hardness; made to extreme cleanliness standards because a single inclusion becomes a spall	Ball and roller bearing races and rolling elements
Nitriding steel (e.g. En41B, the 1.5% Cr-Al-Mo type)	About 0.4% C with roughly 1.5% Cr, 1% Al and 0.25% Mo	The aluminium and chromium form extremely hard nitrides, giving a surface hardness no carburised case can match, with almost no distortion	Crankshaft and camshaft journals, cylinder liners, gears and spindles needing dimensional stability
17-4 PH stainless	About 17% Cr, 4% Ni, 4% Cu with niobium	Precipitation-hardening stainless: solution treated soft, then aged to high strength with very little dimensional change and good corrosion resistance	Highly loaded fittings, actuator components, shafts and valve parts in wet or hot zones
Stainless steel (321)	18-8 austenitic stabilised with titanium	Titanium locks up the carbon so chromium carbides cannot form during welding, preserving corrosion resistance in and beside the weld	Welded exhaust systems, hot-air and bleed ducting, firewall assemblies
Martensitic stainless (410 / 431)	12–17% Cr with the carbon raised enough to allow transformation	Magnetic and hardenable by quench and temper; corrosion resistance lower than the austenitic grades because chromium is tied up as carbide	Compressor blades, fasteners, valve parts, shafts, knife-edge seals and cutting edges
A286 (iron-base superalloy)	About 25% Ni, 15% Cr and 2% Ti, balance iron	Age hardened like the PH grades but retains useful strength to around 700 °C; non-magnetic and corrosion resistant	High-temperature bolts and studs, turbine and hot-section hardware, exhaust clamps
Spring steel (silicon-manganese or chrome-vanadium)	About 0.6% C with 2% Si, or about 0.5% C with Cr and V	Hardened and tempered to a very high yield strength, giving a large elastic range and the ability to return fully to shape after deflection	Valve springs, control-system return springs, retaining clips, circlips, tab washers
Low-carbon aircraft tube (e.g. SAE 1025)	Plain carbon steel, about 0.25% C	Soft, ductile and very easily welded; low strength, so sections are larger than the equivalent in 4130	Non-structural tube, secondary structure and brackets on light aircraft

The Stainless Steel Families

"Stainless" is not one material. The defining requirement is a minimum of about 10.5% chromium, which reacts with oxygen to form an invisible, tightly adherent chromium-oxide film a few atoms thick. That film is passive — it blocks further attack — and it is self-repairing, re-forming within seconds if it is scratched, provided oxygen is available. Nickel is frequently added as well, but nickel is not what makes a steel stainless; chromium is.

Beyond that minimum the family splits according to the crystal structure the composition produces, and structure decides everything else:

Family	Structure	Magnetic?	Strengthened by	Corrosion resistance	Typical aircraft use
Austenitic (300 series — 304, 316, 321, 347)	Face-centred cubic austenite, held down to room temperature by nickel	No	Cold work only — heat treatment cannot harden it	Best of the families; 316 adds molybdenum for pitting resistance	Firewalls, exhaust systems, hot-air ducting, clamps, safety wire, some fasteners
Ferritic (400 series — 430)	Body-centred cubic ferrite at all temperatures; 11–30% Cr with low carbon	Yes	Cold work only — it never becomes austenite, so it cannot be quench hardened	Moderate	Trim, non-structural panelling, low-stress hot items
Martensitic (410, 416, 420, 431, 440C)	Quenched to martensite from austenite, made possible by the higher carbon	Yes	Quench and temper, exactly like a carbon steel	Lowest of the families — chromium is tied up as carbide	Compressor blades, fasteners, valve parts, bearings, cutting edges
Precipitation hardening (17-4 PH, 15-5 PH, 17-7 PH)	Martensitic or semi-austenitic, strengthened by fine particles rather than by carbon	Usually yes	Solution treatment followed by ageing	Good — close to the austenitic grades	Highly loaded fittings, actuator rods, shafts, structural parts in wet zones
Duplex	Roughly half austenite and half ferrite by design	Yes	Cold work; the mixed structure gives high strength as supplied	Very good, especially against stress corrosion in chlorides	Rare in airframes; met mainly in marine and process plant

Tool Steels and High-Speed Steel

Tool steels are not airframe materials, but an engineer uses them every day in drills, reamers, taps, punches and dies, and the exam treats them as part of the ferrous map. They are high-carbon steels, usually with heavy alloy additions, heat treated to very high hardness and correspondingly high wear resistance. The price is brittleness: a tool steel is chosen on the understanding that it will chip rather than bend.

High-speed steel deserves separate treatment because of one property. An ordinary carbon-steel cutting tool loses its hardness as soon as friction heats the edge much above the temperature at which it was tempered — the edge simply tempers itself further and goes soft. High-speed steel does not. It keeps its hardness when the edge is running at a dull red heat, a property called red hardness or hot hardness, and that is what allows it to cut at high speed. The element principally responsible is tungsten, which forms carbides so stable that they do not dissolve or coarsen at cutting temperatures. The classic composition is quoted as 18-4-1 — about 18% tungsten, 4% chromium and 1% vanadium — and molybdenum can substitute for much of the tungsten in the cheaper grades.

Where the very hardest and hottest cutting is required, cobalt is added on top. Cobalt raises hot hardness further, and cobalt-bearing tool and high-speed steels are among the hardest ferrous materials in normal use — a Brinell hardness in the region of 600 to 700 is typical of them, against roughly 100 to 175 for a soft annealed low-carbon steel. Note that this is one of the few practical uses of the Brinell test at the top of its range; a hardened tool steel is normally checked on the Rockwell C scale instead.

Cast Iron

Cast iron carries so much carbon — 2.0% to about 4.5% — that the iron cannot dissolve it all under any conditions. The surplus separates out, and the shape it takes decides the properties completely.

Type	Form the carbon takes	Character
Grey cast iron	Graphite flakes throughout the structure	The commonest type. The flakes behave as thousands of internal cracks, so tensile strength is low and ductility is essentially zero, but compressive strength is three to four times the tensile figure. Excellent vibration damping and outstanding machinability
White cast iron	Carbon locked up as cementite, produced by rapid cooling	Extremely hard, extremely brittle, and effectively unmachinable. Used as a wear surface or as the starting stock for malleable iron
Malleable iron	Graphite in irregular rosettes, produced by a long anneal of white iron	Some genuine ductility and shock resistance, at the cost of a lengthy heat treatment
Spheroidal graphite (ductile or nodular) iron	Graphite as discrete spheres, produced by treating the melt with magnesium or cerium	A sphere concentrates stress far less than a flake, so this iron is genuinely tough and ductile while keeping castability. The best of the family mechanically

The reason cast iron is described simply as heavy and brittle in an examination is that the question is about grey iron, and grey iron is neither tough nor malleable. It fractures without warning under tension or shock. It has essentially no place in airframe structure; where it appears at all it is in piston-engine ancillaries, valve guides, some accessory housings, and ground support equipment, always in compression and never carrying a shock load.

Wrought Iron

Wrought iron is almost pure iron containing fine threads of slag left over from its manufacture. Those slag stringers give it a fibrous grain that makes it tough, easy to forge-weld and notably resistant to corrosion. It was the structural metal of early aviation and it is obsolete: any airframe carrying it is a historic aircraft, and it has been entirely displaced by mild steel, which is stronger, more consistent and far cheaper to make.

Maraging Steel — Strength Without Carbon

Maraging steel is worth understanding in its own right because it strengthens by a route that does not depend on carbon, and because that route is the same one the aluminium alloys use.

The name is a contraction of martensitic and ageing. The carbon content is deliberately kept extremely low — below about 0.03% — so the martensite that forms when the steel is cooled from about 820 °C is not the hard, brittle, carbon-distorted martensite described earlier. It is a soft, tough iron-nickel martensite, and the steel in that condition can be machined, drilled and formed conventionally. The strength comes afterwards, from two heat-treatment steps that are defined the same way for every alloy that uses them:

• Solution treatment — heating to a temperature at which the strengthening constituents dissolve fully into solid solution, then cooling fast enough to hold them there in supersaturated solution. The alloy is comparatively soft and workable immediately afterwards.
• Precipitation hardening (also called age hardening) — holding the supersaturated alloy at room temperature (natural ageing) or at a moderately elevated temperature (artificial ageing) so that the dissolved constituents come back out as extremely fine, closely spaced particles distributed through the grains. Those particles obstruct the movement of dislocations, so hardness and strength rise while ductility falls somewhat. Held too long or too hot, the particles coarsen and spread apart, and the strength falls again — that is over-ageing.

In maraging steel the ageing is carried out at around 480 °C for a few hours, and the precipitates are intermetallic compounds of nickel with molybdenum and titanium. The practical advantages are considerable: an ultra-high tensile strength is reached with better toughness than a carbon-hardened steel of comparable strength, and because the ageing involves no quench there is almost no distortion, so parts can be finish machined before the strengthening treatment. The disadvantages are cost and the fact that maraging steels have poor corrosion resistance and depend entirely on a protective scheme. These two treatments are the principal strengthening route for the light alloys and are covered fully with the non-ferrous materials.

Bearing Steels

Rolling-contact bearings make demands nothing else does: the contact patch between a ball and a race is tiny, so the local stress is enormous, and it is applied and released millions of times. The material has to be very hard where the rolling contact takes place, dimensionally stable, and above all clean, because a single non-metallic inclusion just beneath the surface becomes the origin of a spall, which is why aircraft bearing steels are made to premium melting standards. Two families of steel meet those demands, and they meet them in opposite ways. The commoner of the two is the through-hardening family: high-carbon steels — around 1% — with a chromium addition, hardened uniformly right through the rolling element rather than only at the surface. The standard grade in that family is SAE 52100 and its equivalents — En31 in the older British system, 100Cr6 in the German — all of them roughly 1% carbon with about 1.5% chromium. The second family goes the other way: a low-carbon steel is carburised, so that a hard case is produced over a core that stays tough. That is also where nickel belongs in a bearing steel. It is not an addition to the through-hardening grades, in which it is present only as a residual; it is a deliberate addition in the carburising bearing steels such as SAE 4320 and 9310, whose carbon is around 0.1 to 0.2%, and its job there is to keep the core tough under the hardened case. Bearing types, loads and construction are covered with the bearings material later in this module.

Identification of Ferrous Materials

Aircraft steels are identified by several methods:

• Designation systems: SAE/AISI (USA), BS (British), DIN (German), or specification numbers (e.g. AMS, MIL-S)
• Colour coding: Bars and sheets may be colour-coded on the end or edge per specification standards
• Spark test: Grinding the steel on a wheel produces characteristic spark patterns — carbon content affects spark length, branching, and colour. High carbon produces more branching and bursting sparks
• Magnetic test: Most ferrous materials are magnetic (attracted to a magnet). Exception: austenitic stainless steels (18-8 type) are non-magnetic
• Markings: Part numbers stamped or etched on components; material certificates and traceability documentation

The Order in Which Evidence Counts

The five methods listed above are not equal in weight, and treating them as though they were is the mistake that gets an engineer into trouble. They divide sharply into two classes:

• Documentary evidence — the release paperwork, the material certificate, the heat or cast number, the part number on the item. This is what actually establishes what a piece of metal is, because it links the physical item to a record of its composition, its melting route, its heat treatment and its test results.
• Physical tests — spark, magnet, file, hardness, chemical spot, portable analysis. These are confirmatory. They can tell you that a piece of metal is not what it was supposed to be, which is genuinely useful, but no combination of them can establish that it is the specified material.

The reason is worth stating plainly. Two bars can have identical chemical analysis, identical hardness and identical spark patterns and still be different materials for aviation purposes, because one was vacuum remelted and heat treated to a certificate and the other was not. Composition is only part of a specification. This is why an unidentified piece of steel is scrap: it cannot be made airworthy by testing it, because the tests available cannot recover the history that was lost.

Designation Systems in Detail

The SAE/AISI four-digit system is the one an engineer meets most often on aircraft drawings, and it is genuinely readable once the scheme is known. It encodes composition, not the manufacturer and not any physical property:

• The first digit gives the major alloying class.
• The second digit gives, approximately, the percentage of the principal alloying element in that class.
• The last two digits give the carbon content in hundredths of one per cent.

First digit	Class	Common series
1	Plain carbon steels	10xx plain; 11xx and 12xx free cutting; 15xx higher manganese
2	Nickel steels	23xx, 25xx
3	Nickel-chromium steels	31xx, 33xx
4	Molybdenum-bearing steels	40xx molybdenum; 41xx chromium-molybdenum; 43xx nickel-chromium-molybdenum; 46xx and 48xx nickel-molybdenum
5	Chromium steels	51xx; 52xxx high-carbon high-chromium bearing steels
6	Chromium-vanadium steels	61xx
7	Tungsten-chromium steels	72xx — largely obsolete
8	Nickel-chromium-molybdenum, in lower proportions than the 43xx series	86xx, 87xx, 88xx
9	Silicon-manganese steels and further Ni-Cr-Mo grades	92xx silicon-manganese; 93xx, 94xx, 98xx

The stainless steels use a different, three-digit AISI scheme in which the number is a series identifier rather than a code to be decoded:

• 200 series — austenitic chromium-nickel-manganese, with manganese substituting for some of the nickel.
• 300 series — austenitic chromium-nickel. 304 is the basic 18-8; 316 adds molybdenum for pitting resistance; 321 and 347 are the titanium- and niobium-stabilised welding grades.
• 400 series — straight chromium. Some are ferritic (430) and some martensitic (410, 420, 440C); the carbon content decides which.
• A suffix L means a reduced carbon maximum for welding; H means a raised carbon content for high-temperature strength.

Alongside those you will meet specification numbers rather than grade names, and the distinction matters:

System	Issued by	What it controls
SAE / AISI	SAE International	Chemical composition only
AMS (Aerospace Material Specification)	SAE International	Composition plus the product form, the melting practice, the heat-treated condition, the required mechanical properties, the testing and the acceptance criteria. This is what an aerospace drawing calls out
MIL-S / MIL-DTL	US Department of Defense	Comparable coverage to AMS, on military programmes
BS, EN, DIN, W.Nr	British, European and German standards bodies	European equivalents; EN 10027 uses a descriptive form in which the elements and their percentages appear in the name itself
UNS	SAE and ASTM jointly	A single unified number for every metal and alloy, used to cross-reference between the systems above

Colour Coding — Read the Scheme, Never Assume It

Coloured paint or lacquer on the end or the edge of a bar, sheet or tube is a stock-control convenience. The important thing to understand about it is that there is no single universal colour scheme. The meaning of a colour is defined by the specification, the standard or the supplier that applied it, and it varies between them. A colour is therefore a pointer to the paperwork, not a substitute for it, and a bar whose colour code cannot be matched to a document is an unidentified bar. Colour marks are also the first thing lost when a bar is cut, cleaned or handled, which is exactly why the heat or cast number is stamped or etched rather than painted.

The Spark Test in Practice

Grinding a steel throws off particles that are heated to incandescence and then burn in air. The carbon in them oxidises explosively, and each explosion splits the streak — so the amount of branching and bursting is a direct visual read-out of carbon content, while the colour and length of the stream are influenced by the alloying elements. More carbon means more, and more repeated, bursts.

Material	Stream	Colour	Bursts
Wrought iron	Long, straight streaks that widen at the far end into leaf or spear shapes	Straw to white	None
Low-carbon (mild) steel	Long and straight, with a few forks appearing near the end of the stream	Bright straw-yellow	Very few
Medium-carbon steel	Shorter, with forking along much of its length and small bursts	Bright yellow	Moderate in number
High-carbon steel	Short, dense and bushy, with many repeated star-shaped bursts starting close to the wheel	Bright white-yellow	Numerous and repeating
Grey cast iron	Short stream that stays close to the wheel	Red near the wheel, straw further out	Many small, fine bursts
Austenitic stainless steel	Short and only moderately forked	Straw with a reddish tint	Few
High-speed steel	Short and interrupted, with the streak often ending in a spearhead or bulb	Dark red	None or very few
Nickel and high-nickel alloys	Short, ending in a characteristic wedge or tongue	Orange	None

Three rules make the test worth doing and its results defensible:

• Always compare against a known sample. The test is comparative. Grind a piece of certified material of the suspected grade on the same wheel, at the same pressure, immediately before or after the unknown, and compare the two streams directly. An absolute judgement from memory is unreliable.
• Control the conditions. Use a clean, sharp wheel and light, steady pressure — heavy pressure lengthens and brightens the stream and masks the differences. View the stream against a dark background in subdued light; in bright sunlight the bursts are simply invisible.
• Observe the safety precautions. Eye protection, the correct guard and rest settings, and awareness that grinding sparks are an ignition source. Never spark test a material of unknown composition where magnesium may be present, and never grind a component that is to remain in service — take the sample from scrap stock or from a part already condemned.

The Magnet Test and Its Traps

Most ferrous metals are attracted to a magnet, which makes a pocket magnet the fastest first check available. The response comes from ferrite and martensite, both of which are ferromagnetic; austenite is not. Used carefully the test is genuinely informative, but it has three specific traps.

• A magnetic response does not prove the metal is ferrous. Nickel and cobalt are ferromagnetic in their own right, and so are some of their alloys. A magnet cannot by itself distinguish a steel from a nickel alloy.
• A weak magnetic response does not prove an austenitic part is the wrong material. Cold working transforms some austenite into martensite, so a heavily formed, drawn or machined 18-8 component — a bent bracket, a swaged fitting, a cut thread — is often weakly magnetic at the worked areas while the parent material beside them is not. The same applies to welds in austenitic stainless, whose filler frequently contains a deliberate small proportion of ferrite to prevent hot cracking.
• The response disappears with heat. Iron loses its ferromagnetism at about 770 °C, well below its melting point, so a hot part gives no response at all until it has cooled.

The clean, reliable conclusion the test does support is the one worth memorising: austenitic stainless steels of the 18-8 type are non-magnetic in the annealed condition, whereas ferritic and martensitic stainless steels are magnetic. A magnet is the quickest way to separate a 300-series grade from a 400-series one.

File Test, Hardness Comparison and Chemical Spot Tests

A sharp file drawn across an edge tells you a surprising amount. It bites readily into an annealed or normalised steel, cuts with more effort on a hardened and tempered one, and simply skates over a fully hardened surface without removing metal. Because it is quick and needs no equipment it is the classic way of confirming that a case-hardening or hardening operation has actually taken effect. Like the spark test it is comparative and it removes metal, so it is done on a non-critical area or on scrap.

Chemical spot tests apply a drop of reagent and read the colour produced. They can distinguish, for instance, a nickel-bearing stainless from a plain chromium one. They are useful, but they are chemically specific, they mark the surface, and their use on an aircraft part requires approved data.

Instrumental Identification

Method	How it works	What it can do	What it cannot do
Portable X-ray fluorescence	The surface is irradiated and each element re-emits X-rays at its own characteristic energy	Reads chromium, nickel, molybdenum, manganese and the other alloying metals in seconds, without marking the part	It cannot read carbon — the element is too light to fluoresce usefully. So it separates 4130 from 4340, but it cannot separate 1020 from 1095, or 410 from 420, because those differ only in carbon
Optical emission spectrometry (arc or spark)	A controlled electrical discharge vaporises a trace of the surface and the light emitted is split into its spectrum	Reads carbon as well as the alloying elements, giving a full analysis	Leaves a small burn mark, so it needs approved data before it is used on a part that is to remain in service
Laboratory combustion analysis	A machined sample is burned in oxygen and the carbon dioxide produced is measured	The reference method for carbon and sulphur	Destructive — it consumes a sample, so it applies to stock rather than to an installed part

The Magnetic Property Is Also an Inspection Tool

The same ferromagnetism that makes the magnet test work is what makes magnetic particle inspection possible. The part is magnetised and fine iron particles, dry or suspended in a liquid and often fluorescent, are applied. A crack or inclusion interrupts the magnetic field and forces some of it out of the surface; that leakage field holds the particles and draws a visible line along the defect. The technique finds surface and near-surface discontinuities, and it is fast, sensitive and cheap.

Its limitation follows directly from its principle: it works only on ferromagnetic materials. Ferritic and martensitic steels can be inspected this way; austenitic stainless, aluminium, titanium, magnesium and composites cannot. The non-magnetic metals go instead to dye penetrant, eddy current, ultrasonic or radiographic inspection. Composites do not follow them onto the first two of those. Eddy current needs a good conductor: a glass or aramid laminate is not one at all, and carbon fibre conducts far too poorly and too directionally for a probe to read. Dye penetrant soaks into a porous laminate and contaminates it for any bonded repair afterwards. A damaged laminate is examined ultrasonically or radiographically, by tap testing, or — where the laminate is translucent — by stripping the paint and shining a strong light through it. A part must also be demagnetised (degaussed) after inspection, or residual magnetism will attract ferrous debris in service and can disturb nearby instruments. The methods themselves belong with the maintenance practices material.

Traceability — What the Paperwork Actually Records

Every batch of aerospace steel is made from a single melt, and that melt carries a heat number or cast number that follows the material from the furnace to the finished part. The certificate issued against that number records the chemical analysis, the melting practice, the product form and size, the heat-treated condition, and the results of the mechanical tests carried out on samples taken from that same melt.

Three practical rules follow, and they are the reason a stores procedure looks the way it does:

• Segregate. Different grades, and different heats of the same grade, are stored separately and never allowed to become mixed. Once two bars of similar appearance are in the same rack without markings, both are unidentified.
• Re-identify on cutting. When a bar is cut, the offcut loses whatever identification was on the removed end. The heat number and specification must be transferred to the remaining piece before it leaves the saw.
• Keep the release documentation with the material. The certificate and the release document are what allow the finished part to be certified; without them the material cannot be used, however good it may be.

Heat Treatment of Steel

Heat treatment is the controlled heating and cooling of metals to alter their mechanical properties without changing the shape. It is one of the most important processes in aircraft steel manufacture and maintenance.

Hardening (Quenching)

The steel is heated to above its upper critical temperature (typically 750–900°C depending on alloy) until the crystal structure transforms to austenite. It is then rapidly cooled (quenched) in oil, water, or air. The rapid cooling traps the carbon atoms in the crystal lattice, forming martensite — an extremely hard but brittle structure.

Result: Maximum hardness and strength, but very low ductility and toughness. The steel is too brittle for most applications in this state — it must be tempered.

Tempering

After hardening, the steel is reheated to a moderate temperature (150–650°C) and held for a period, then cooled. Tempering reduces brittleness while retaining most of the hardness gained from quenching. The higher the tempering temperature, the softer and tougher (but less hard) the steel becomes.

Result: The desired balance between hardness and toughness. Different applications require different tempering temperatures — springs are tempered at around 300–400°C; cutting tools at 150–250°C.

Annealing

The steel is heated to above its critical temperature and then cooled very slowly (typically in the furnace itself). This produces the softest possible condition with maximum ductility.

Purpose: To soften the steel for machining, forming, or cold working; to relieve internal stresses; to refine the grain structure.

Normalising

Similar to annealing, but the steel is cooled in still air (faster than furnace cooling but slower than quenching). This produces a finer, more uniform grain structure than annealing.

Purpose: To refine grain structure after hot working (forging, welding); to produce a uniform structure; to improve machinability.

Case Hardening

A process that hardens only the outer surface of the steel while leaving the core soft and tough. This is ideal for components that need a wear-resistant surface but must withstand shock and fatigue loads (like gears and bearing races). Methods include:

• Carburising: The steel is heated in a carbon-rich environment (solid, liquid, or gas). Carbon diffuses into the surface, increasing the surface carbon content. The part is then quenched to harden the carbon-rich surface layer.
• Nitriding: The steel is heated in an ammonia atmosphere at around 500°C. Nitrogen diffuses into the surface, forming extremely hard nitride compounds. No quenching is needed — the surface is hard as-treated. Excellent for fatigue resistance.
• Induction hardening: The surface is rapidly heated by electromagnetic induction and then quenched. Only the heated surface layer hardens.

The Anatomy of a Heat Treatment Cycle

Every treatment described above — and every one that follows — is built from the same four steps, and a specification controls all four. Understanding what each step is for makes the individual processes much easier to keep apart.

• Heating rate. Steel expands as it heats, and it does so unevenly if the outside heats much faster than the inside. A thick or complex part is therefore heated slowly, or in stages with a holding period part-way up, to keep the temperature difference through the section small. Rushing this step distorts or cracks the part before the treatment has even begun.
• Soak temperature. This is the step that decides which structure you are aiming at, and it is set relative to the steel's critical points, not to a universal number. A hardening or normalising temperature is above the upper critical point; a tempering or stress-relieving temperature is below the lower critical point.
• Soak time. Time at temperature is needed for two separate reasons: the centre of the part has to reach the same temperature as the surface, and the transformation itself has to complete. A common workshop guide is about one hour for every 25 mm of the ruling section — the thickest part of the component — but the approved process specification governs. Too short a soak leaves the core untransformed; too long a soak coarsens the grain, and coarse grain lowers toughness.
• Cooling rate. This is the step that decides what the austenite turns into, and therefore what the properties are. Everything from a furnace cool to a brine quench is a choice of cooling rate.

The Furnace Atmosphere Matters As Much As the Temperature

Steel heated in air above roughly 700 °C does two unwanted things at once, and both degrade the surface that will eventually carry the highest stress.

• Scaling. Iron oxidises to a loose, flaky black scale. The part loses dimension, the surface finish is destroyed, and the scale has to be removed afterwards.
• Decarburisation. Oxygen and water vapour react with the carbon in the surface layer and remove it as carbon monoxide. The surface is left with less carbon than the specification calls for, so after quenching it is softer than the core rather than harder, and its fatigue strength is reduced. This is a serious defect on a spring, a gear tooth or a fastener thread, because those are exactly the features where the stress peaks at the surface.

The remedies are to heat in a controlled protective atmosphere, in a vacuum furnace, or in a molten salt bath that excludes air from the surface; or, where none of those is available, to leave a machining allowance and remove the affected layer afterwards. A decarburised layer cannot be put right by re-heat-treating — the carbon has gone.

Hardening in Detail

Two decisions define a hardening operation: what temperature to austenitise at, and how fast to cool.

The austenitising temperature is normally set a modest amount — typically some tens of degrees — above the upper critical point for the grade. High enough to be certain the transformation is complete through the section; not so high that the grain coarsens or the surface scales and decarburises. There is one important exception. For a hyper-eutectoid steel — one carrying more than 0.83% carbon, such as a bearing or tool steel — the part is deliberately heated to above the lower critical point but below the upper, so that the existing hard cementite particles are not dissolved. Keeping them undissolved gives a harder, more wear-resistant result and avoids the excess retained austenite that dissolving them would produce.

The quench medium is chosen to cool the part faster than its critical cooling rate, and no faster than necessary. Overshooting costs distortion and risks cracking.

Quench medium	Relative severity	Behaviour	Where it is used
Brine (salt solution)	Most severe — roughly twice water on the standard comparative scale	The dissolved salt disrupts the insulating vapour film, giving very rapid and unusually even cooling	Plain carbon steels of low hardenability in simple shapes. Highest distortion and cracking risk of all
Water	Severe — the reference value of 1.0 on that scale	Fast, but a clinging vapour blanket makes the cooling uneven, which is what produces distortion	Plain carbon steels; simple, symmetrical sections
Polymer solution	Adjustable between water and oil	The polymer concentration sets the severity, so one bath can be tuned to the steel	A controllable substitute for water or oil, and non-flammable
Oil	Moderate — roughly a quarter to a third of water	Slower and much more even; there are fast, medium and slow quenching oils, and warming the oil speeds it up by thinning it	The great majority of alloy steels, including 4130 and 4340. The standard aerospace quench
Molten salt	Moderate, but held at a chosen temperature rather than cold	Cools rapidly to the bath temperature and then stops, which is what makes martempering and austempering possible	Distortion-critical and thin-section parts
Gas (nitrogen or helium under pressure)	Mild, and adjustable with pressure and flow	Cools cleanly with no residue and no fire risk, inside a vacuum furnace	High-hardenability tool, die and aerospace steels
Still air	Mildest — roughly one fiftieth of water	Slow and gentle	Normalising, and the genuinely air-hardening steels whose hardenability is high enough to reach martensite in air

Two operating details change the result more than students expect. Agitation — moving the part through the quenchant or circulating the quenchant past it — breaks up the vapour blanket and can raise the effective severity of a given medium substantially, as well as making the cooling far more uniform. And the direction of entry matters: a long slender part is quenched vertically, end-on, so that it cools symmetrically, because entering it flat causes one side to cool first and bows it permanently.

Two further points about the quench close the topic:

• The transformation to martensite does not finish at room temperature in every steel. It begins at a temperature called the martensite start and continues only while the temperature keeps falling. In high-carbon and heavily alloyed steels the finish temperature lies below room temperature, so a proportion of retained austenite survives the quench. Retained austenite is soft, and it may transform later in service or in storage, changing the part's dimensions. Where that matters — bearing races, gauges, precision spindles — a sub-zero treatment immediately after quenching drives the transformation to completion before tempering.
• Hardenability decides how much of the section actually hardens. The hardness reached at the surface is set by the carbon content; the depth to which that hardness penetrates is set by the alloy content. A large plain-carbon part quenched in water can be file-hard on the outside and completely untransformed a few millimetres in. The standard measurement of hardenability is the Jominy end-quench test, described with the other material tests later in this note.

Interrupted Quenching — Martempering and Austempering

Both of these use a molten salt bath held at a chosen temperature, and both exist to reduce the internal stresses an ordinary quench generates.

• Martempering. The part is quenched into a bath held just above the martensite start temperature and held there until the whole section has reached the bath temperature. Only then is it removed and allowed to cool in air, so the martensite forms slowly and, crucially, at the same time throughout the section rather than surface-first. The result is the same martensitic structure with far less distortion and a much lower risk of cracking. It still requires tempering afterwards.
• Austempering. The part is quenched into a bath held above the martensite start temperature, in the band where austenite transforms to bainite — commonly in the region of 260 to 400 °C — and held there long enough for that transformation to go to completion. The difference from martempering is what the hold is for: martempering holds only until the section has equalised in temperature and then forms its martensite on the way out, while austempering holds until there is no austenite left to become martensite. Bainite is hard and unusually tough for its hardness, and because no martensite is formed the process produces very little distortion and needs no subsequent tempering. Its limitation is section thickness: the part must be thin enough for the whole section to reach the bath temperature before transformation begins.

Tempering in Detail

Tempering allows carbon to come out of the strained martensite lattice as fine carbide particles. As it does so the lattice relaxes, so the internal stress falls, ductility and toughness rise, and hardness falls. The amount of each depends almost entirely on the temperature chosen; time has a much weaker effect. As a rule, in plain carbon and ordinary low-alloy steels a higher tempering temperature gives a softer, tougher result. Two exceptions to that rule are set out below and both matter in practice: secondary hardening, where in steels carrying molybdenum, vanadium, tungsten or chromium the hardness stops falling and can rise again; and temper embrittlement, where certain steels lose impact toughness in a particular band with no change in hardness to warn you.

Approximate tempering range	What is happening in the steel	Result and typical application
150–250 °C	Carbon segregates and a very fine transition carbide precipitates. The lattice strain falls sharply while the carbide is still extremely fine	A large reduction in internal stress for only a small loss of hardness. Cutting tools, dies and other parts where hardness is everything
250–400 °C	Retained austenite transforms, and the transition carbide converts into cementite	Hardness drops noticeably while resilience reaches its best. Springs are tempered in this band
400–650 °C	The cementite particles coarsen and become rounded, and the ferrite matrix recovers	Hardness falls further and toughness climbs strongly. Highly stressed shafts, gears, landing-gear parts and structural fittings
500–600 °C, in steels containing molybdenum, vanadium, tungsten or chromium	Fine, stable alloy carbides precipitate for the first time	Secondary hardening — hardness stops falling and can rise again. This is why high-speed steel is tempered in this range rather than at a low temperature

Two refinements are worth knowing because they explain requirements that otherwise look arbitrary:

• Double and triple tempering. In high-alloy and tool steels the first temper transforms the retained austenite into fresh, untempered martensite, which is brittle. A second temper is therefore run to temper that new martensite, and on some grades a third. The number of tempers is specified, and skipping one leaves brittle material in the part.
• Temper embrittlement. Certain steels lose impact toughness if they are tempered in, or cooled slowly through, particular temperature bands, without any change in hardness to warn you. In nickel-chromium steels the effect is associated with slow cooling through roughly the 375 to 575 °C region and with trace impurities gathering at the grain boundaries. It is countered by adding molybdenum — which is precisely the reason molybdenum appears in the alloying list as reducing temper brittleness — and, where the steel is susceptible, by cooling briskly from the tempering temperature rather than letting the part cool slowly in the furnace. The general rule that tempering ends in a gentle cool still holds for ordinary carbon and low-alloy steels; this is a specific alloy exception, and the process specification says which applies.

Temper Colours — a Workshop Guide and Nothing More

A clean, bright steel surface grows a thin oxide film as it is heated, and the film's thickness — and therefore its interference colour — changes progressively with temperature. This gives a traditional and genuinely useful way of judging tempering temperature by eye at the bench.

Approximate temperature	Oxide colour on bright steel	Condition produced
about 220 °C	Pale straw	Almost all the quenched hardness retained; cutting edges
about 240 °C	Dark straw	Slightly tougher; drills, taps, reamers
about 260 °C	Brown	Punches, shear blades
about 280 °C	Purple	Impact tools such as cold chisels and screwdrivers
about 300 °C	Dark blue	Springs and other resilient parts
about 320 °C and above	Light blue, then grey	Distinctly softer and much tougher

The Annealing Family

Annealing is not one process. Four distinct treatments carry the name, and they differ in the temperature reached and therefore in what they can achieve.

Treatment	Temperature	Cooling	What it achieves
Full annealing	Above the upper critical point	Very slowly, in the furnace	The softest and most ductile condition obtainable, with all internal stress relieved and the previous structure completely replaced. The grain ends up coarse, which is the price
Process (sub-critical) annealing	Below the lower critical point, typically in the 550–650 °C region	Air or furnace	Recrystallises a cold-worked structure and restores ductility so that further cold working can continue, without any phase change. Used between drawing or pressing operations
Spheroidise annealing	A long soak just below the lower critical point, or cycling gently around it	Very slowly	Rounds the cementite plates into discrete spheres. This is the softest and most machinable condition a high-carbon steel can be put in, and it is the standard condition for supplying tool and bearing steel for machining
Stress relieving	Below the lower critical point, typically 550–650 °C for carbon and low-alloy steels	Slowly, to avoid introducing fresh thermal stress	Reduces residual stress without changing the structure, the hardness or the strength. Applied after welding, heavy machining or forming

Residual stress is the stress locked into a component after manufacture and all heat treatment, with no external load applied at all — as distinct from working or applied stress, which exists only while the part is actually carrying a load. It comes from uneven cooling, from welding, from machining and grinding, and from cold forming. It matters because it adds to the service stress: a part carrying 300 MPa of locked-in tension reaches its yield point at a lower applied load than the drawing assumes, it distorts unpredictably when metal is machined off and releases the balance, and tensile residual stress at a surface makes stress-corrosion cracking and fatigue cracking far more likely. Stress relieving exists to remove it.

Annealing Against Normalising — Choosing Between Them

Both start from above the upper critical temperature and both relieve stress; the only difference in the process is the cooling rate, and every difference in the result follows from that one thing.

Aspect	Full annealing	Normalising
Where the part cools	Inside the furnace, with the furnace switched off	Out of the furnace, standing in still air at room temperature
Resulting pearlite	Coarse — thick, widely spaced plates	Fine — thin, closely spaced plates
Grain size	Coarse	Finer and more uniform
Hardness and strength	The lowest obtainable	Moderately higher than annealed
Ductility	The highest obtainable	Good, but below the annealed value
Impact toughness	Moderate — limited by the coarse grain	Better than annealed, because of the finer grain
Furnace time and cost	Long — the furnace is occupied throughout the cool	Short — the furnace is released as soon as the part comes out
Typically specified when	The part must be as soft and workable as possible for heavy machining or cold forming, or must be fully softened before re-hardening	The part must be returned to a uniform, fine-grained, stress-free condition after forging, welding, casting or rolling — the standard preparation before hardening

Case Hardening — Comparing the Processes

The object of every one of these processes is the same — a hard case over a tough core — but they reach it by two quite different principles, and confusing the two is a common error.

• Carburising, carbonitriding and nitriding change the composition of the surface. They diffuse an element into it that was not there before. They are the answer when the base steel does not contain enough carbon to harden.
• Induction and flame hardening change only the temperature of the surface. The composition is untouched; they simply austenitise and quench a thin layer of a steel that already contains enough carbon to respond. They are the answer when the base steel does contain enough carbon but you want only part of it hard.

Process	Temperature	Medium	Diffuses	Quench needed?	Typical case depth	Base steel required	Distortion
Pack carburising	about 900–950 °C	Solid charcoal and an energiser in a sealed box	Carbon	Yes	Deep — up to a few millimetres over a long cycle	Low carbon, roughly 0.1–0.25%	High — the whole part is soaked at temperature for many hours
Gas carburising	about 900–950 °C	A hydrocarbon gas atmosphere in a sealed furnace	Carbon	Yes	Controllable, from a few tenths of a millimetre upwards	Low carbon	Moderate, and far more controllable than pack
Liquid (salt bath) carburising	about 800–900 °C	Molten cyanide-bearing salts	Carbon, with some nitrogen	Yes	Shallow, and quickly produced	Low carbon	Moderate. The salts are extremely toxic and their use is tightly controlled
Carbonitriding	about 800–870 °C	Carburising gas with ammonia added	Carbon and nitrogen together	Yes, but a milder one	Shallow	Low carbon	Lower than carburising, because both the temperature and the quench are gentler
Nitriding	about 500 °C — well below the lower critical point	Ammonia gas, a salt bath, or a plasma	Nitrogen only	No — the case is hard as treated	Thin, a few tenths of a millimetre	Alloy steels containing aluminium, chromium, molybdenum or vanadium	Minimal — no phase change and no quench
Induction hardening	Locally above the upper critical point	None — an induction coil heats the surface electrically	Nothing; it uses the carbon already present	Yes, usually an integral water spray	Set by the depth heated, and controllable	Must already contain roughly 0.35% carbon or more	Low and localised; can be applied to a finished part
Flame hardening	Locally above the upper critical point	None — an oxy-fuel torch heats the surface	Nothing; it uses the carbon already present	Yes, a following water quench	Generally deeper than induction and less precisely controlled	Must already contain roughly 0.35% carbon or more	Low and localised

Four practical points govern how case hardening is used on real components:

• Case depth grows with the square root of time, not in proportion to it. Diffusion is slow and gets slower as the concentration gradient flattens, so doubling the depth of a carburised case takes roughly four times as long at the same temperature. Deep cases are therefore expensive, and this is one reason deep-case components are specified only where the contact stress genuinely demands it.
• Areas that must stay soft are masked. A thin copper plating, or a proprietary stop-off paste, prevents carbon reaching the surface underneath. It is used where a feature will be machined after treatment, where a thread must not be made brittle, or where a hole is to be drilled.
• Carburising coarsens the grain, so a refining treatment usually follows. Hours at 900 °C or more coarsens the grain of both the case and the core. The classic remedy is a two-stage reheat and quench: first from above the core's critical point to refine the core, then from above the case's critical point — which is lower, because the case is now high-carbon — to refine the case, followed by a low-temperature temper. Modern fine-grained steels often permit a direct quench from the carburising temperature instead.
• Nitrided parts are hardened and tempered before nitriding. The tempering temperature is chosen to be above the nitriding temperature, so that the core does not soften during the long nitriding cycle. This ordering is not optional; reversing it destroys the core properties. It also follows that a nitrided surface cannot be machined afterwards except by grinding, and any brittle white compound layer that forms is removed if the specification calls for it.

Nitriding earns a further note because of what it does to fatigue strength. The nitrides formed in the surface occupy more volume than the metal they replaced, so the case is left in a state of compression. An applied tensile stress at that surface must first cancel the built-in compression before the metal sees any net tension at all, which delays the initiation of a fatigue crack and raises the fatigue strength markedly. It is the same principle that makes shot peening effective.

Finally, the depth of a case is specified in one of two ways and they are not interchangeable. The total case depth is the full depth over which the composition has been altered. The effective case depth is the depth down to a stated hardness — commonly taken as 550 HV for a carburised case — and it is the figure that matters structurally, because it is the depth of genuinely hard material. The specification states which is required and what the value is.

Hydrogen Embrittlement — a Heat Treatment That Undoes Damage

High-strength steels — broadly, those heat treated above about 1,000 MPa tensile strength, though process specifications set the threshold in their own terms and some state it as a hardness — will absorb atomic hydrogen from acid pickling, acid cleaning, electroplating and some corrosion reactions. The hydrogen diffuses through the steel and collects where the internal stress is highest, and there it causes a delayed, brittle fracture at a load well below the design strength. The failures characteristically appear after the part has been in service for some time, which makes them particularly dangerous, and landing-gear steels are the classic case.

The remedy is a low-temperature bake carried out within a short specified period after the process that introduced the hydrogen, at a temperature commonly in the region of 190 °C for a number of hours set by the process specification. The bake drives the hydrogen back out before it has done its damage. Two things are worth being clear about: the bake must be done promptly, because once cracking has initiated it cannot be reversed; and no mechanical treatment — peening, polishing, stress relieving at the wrong temperature — substitutes for it.

Dimensional Change, and What You Must Never Do to a Heat-Treated Part

Martensite occupies slightly more volume than the austenite it came from, so a part grows very slightly on hardening. Combined with the distortion that uneven cooling produces, this is why highly accurate components are rough machined, heat treated, and then finish ground to size — and why a machining allowance is left on the drawing for exactly that purpose.

Working and Handling Ferrous Materials

Heat treatment is not the only thing that changes a steel's properties. Every mechanical operation an engineer performs on it — bending, drawing, machining, grinding, welding — alters the metal as well, and several of those changes are permanent and invisible. This section covers what happens to steel when you work it, and the handling rules that follow.

Cold Working and Strain Hardening

Deform a metal permanently at room temperature and it becomes harder and stronger. This is cold working, and the strengthening it produces is called strain hardening or work hardening.

The mechanism is worth knowing because it explains every consequence. Permanent deformation happens by the movement of line defects called dislocations through the crystal lattice. As they move they multiply, and they run into one another and become entangled. A tangled dislocation is a dislocation that cannot move, so the more the metal is deformed the harder it becomes to deform it further.

Every property that moves at all does so in a definite direction, and those directions are opposite to the ones annealing produces. The third column below is the part worth learning, because it is what lets you predict the answer for a property this table does not list:

Property	Effect of cold working	Why it follows from the dislocation mechanism
Yield strength	Increases — markedly, and this is the largest effect of all	Yield is by definition the stress at which dislocations start moving in bulk, so obstructing them raises it directly and immediately
Tensile strength	Increases, but by less than the yield strength does	The material is being pushed up its own strain-hardening curve, so it starts closer to the top of it. The yield and tensile figures converge
Hardness	Increases	Hardness measures resistance to permanent indentation, which is the same dislocation movement measured a different way
Ductility and malleability	Decrease	Part of the total deformation the metal can ever undergo has already been spent, so less remains before fracture
Toughness	Decreases	Toughness is the energy absorbed before fracture, and with less remaining deformation available there is less distance over which to absorb it
Internal (residual) stress	Increases	Different grains deform by different amounts and end up locked against one another. This is why a process anneal, not just a rest, is needed to relieve it
Young's modulus (stiffness)	Unchanged	Stiffness comes from the strength of the atomic bonds themselves, and tangling dislocations does not alter those. Nothing an engineer does to a steel changes its stiffness
Grain shape	Distorted and elongated in the direction of working	The grains are physically dragged out with the metal, which is why a cold-worked material has different properties along and across the working direction

Check the direction against a concrete case. A mild-steel rod is drawn through a series of progressively smaller dies to make wire. Each pass deforms it permanently. The finished wire has a tensile strength several times that of the rod it started as, and where the rod could be bent double, the wire snaps if it is bent too sharply. Strength up, ductility down — and it was the drawing, not any heat treatment, that did it.

The same mechanism explains a feature of the tensile test that puzzles students. Once mild steel has passed its yield point it continues to accept more load, not less, all the way up to the ultimate tensile strength. It does so because the plastic deformation is strain hardening it as it goes: the material is getting stronger faster than the specimen is getting thinner. Only when that balance tips does the specimen begin to neck, and necking — the local thinning near the end of the test — is the consequence, not the cause.

Cold working is used deliberately wherever it helps:

• Wire drawing, cold rolling and cold heading produce a finished item that is stronger than the stock it was made from.
• Thread rolling forms a thread by displacing metal rather than cutting it. The result is stronger in fatigue than a cut thread, both because the grain flow follows the thread form instead of being severed by it and because the root is left in residual compression.
• Shot peening bombards a surface with hard shot, deforming a thin layer and leaving it in compression. Since a fatigue crack must start under tension, a surface that begins in compression has to have that compression cancelled before it sees any net tension at all, which is why peening raises fatigue life.
• Austenitic stainless steel is strengthened by cold work and by nothing else, since heat treatment cannot harden it.

It also has to be controlled. A heavily cold-worked part has spent much of its ductility and is closer to cracking, which is why sheet-metal work has minimum bend radii, why a formed part must not be straightened and re-bent, and why a process anneal is inserted between severe forming operations to recrystallise the structure and restore the ductility before work continues.

Hot Working and Why Forgings Beat Billets

Work the same metal above its recrystallisation temperature and the picture changes completely. New, strain-free grains form as fast as the deformation distorts the old ones, so no strain hardening accumulates. Enormous shape changes become possible with modest force, and the metal never runs out of ductility.

The important by-product is grain flow. Hot working stretches the grains, and the inclusions and segregation between them, into continuous fibres that follow the shape being formed. A forging die shapes those fibres so that they run along the part, around its corners and through its fillets — that is, along the load path.

Now compare two identical-looking parts. One is closed-die forged to near its final shape and then finish machined. The other is machined out of a solid rolled bar. The forged part has continuous grain flow following its contours. The machined part has the bar's straight flow lines cut clean through, so their ends are exposed at every curved surface, and each exposed end is a potential crack origin. The two have the same chemistry, the same hardness and the same tensile test result — and markedly different fatigue lives. This is why landing gear components, engine mount fittings and highly stressed brackets are forgings, and it is one more reason a part may not be locally reshaped or built up in a way the designer did not intend.

Hot working leaves a coarse grain and a scaled, decarburised surface, so it is normally followed by normalising and by removal of the affected surface layer.

Machining Steel

Cutting speed and tool life are governed principally by the hardness of the workpiece, and the relationship is simple: harder material means a lower cutting speed. A fully hardened steel cannot be turned or milled at all with conventional tooling and must be ground.

Machinability itself peaks in the middle of the hardness range and falls away at both ends, which is not obvious. Very soft, very ductile low-carbon steel is difficult to machine well because the chip tears rather than shearing cleanly and builds up on the tool edge, leaving a poor finish. Very hard steel wears the tool out. A normalised medium-carbon steel — or a spheroidise-annealed high-carbon one — machines best.

Grinding Damage — the Defect You Cannot See

Grinding is the normal way of finishing a hardened steel part to size, and it is also the operation most capable of ruining one. A grinding wheel puts a great deal of energy into a very thin surface layer. If the wheel is blunt or loaded, the cut is too heavy, or the coolant is inadequate, that layer gets hot enough to change its own heat treatment:

• Over-tempering (grinding burn). The surface is reheated above its tempering temperature and becomes locally soft. It usually shows as a straw or blue discolouration, and it means the surface no longer has the hardness or the wear resistance the design assumed.
• Re-hardening and grinding cracks. If the layer gets hot enough to re-austenitise, the cold mass of metal underneath quenches it the instant the wheel passes, forming a thin skin of untempered martensite. That skin is brittle and in tension, and it cracks — typically in a fine network running at right angles to the grinding direction. Grinding cracks are a well-known cause of fatigue failure in hardened components.

The damage may be invisible to the naked eye, which is why hardened, ground aerospace parts are commonly subject to a nital etch inspection: a controlled acid etch that reveals over-tempered and re-hardened areas as distinct dark and light patches. Prevention is a sharp, correctly dressed, open-structure wheel, light cuts, and a generous flow of coolant delivered where the wheel meets the work.

Welding Ferrous Materials

Welding is a local heat treatment that nobody controls. The weld metal itself is a small casting; beside it is a heat-affected zone in which the parent metal has been heated to every temperature between melting and ambient, and then cooled at whatever rate the surrounding cold metal dictated — which in a thick or restrained joint is fast enough to count as a quench.

The consequence is that the heat-affected zone of a hardenable steel can end up as hard, brittle, untempered martensite, full of residual stress and vulnerable to cracking. How likely that is depends on how hardenable the steel is, and hardenability is estimated for welding purposes by the carbon equivalent, which converts the alloy content into an equivalent amount of carbon:

A higher carbon equivalent means a more hardenable heat-affected zone and therefore a greater need for preheat, for controlled interpass temperature, for low-hydrogen consumables and for post-weld heat treatment. The two figures above show why a mild-steel bracket is welded without ceremony while a 4130 joint needs a procedure. But the carbon equivalent is only one input: joint thickness, restraint, heat input, the hydrogen content of the consumables and the cooling rate all bear on the outcome, which is why thin-wall 4130 tube is routinely welded under one set of conditions and a heavy 4130 section is not. The approved welding procedure is what governs, and it is written for the material, the thickness and the joint concerned.

Three further points cover most of what an engineer meets:

• Welding a heat-treated part destroys its heat treatment locally. There is no way around this. A part supplied hardened and tempered cannot be welded and left in that condition; either the whole assembly is re-heat-treated afterwards, or the repair is not permitted. This is why welded 4130 tube structures are welded first and normalised or stress relieved afterwards, and why the repair scheme — not the welder — decides what may be welded.
• Austenitic stainless welds differently from carbon steel. It expands about half as much again and conducts heat about a third as well, so the same weld puts far more distortion into it; fixturing and a controlled welding sequence matter more. And the heat-affected zone is at risk of sensitisation, which is why stabilised grades are used for welded assemblies.
• Dissimilar joints need thought. Joining a stainless to a carbon steel, or a high-carbon to a low-carbon steel, produces a weld metal of intermediate composition whose properties are not those of either parent. The filler is specified for that reason and is not a free choice.

Forming, Bending and Springback

Three material properties govern what can be done to sheet and tube:

• Ductility sets the minimum bend radius. The outside of a bend is in tension, and the tighter the radius the greater the strain there. Exceed the material's ductility and the outer fibre cracks. Bend radii are therefore quoted in terms of material thickness, and the governing figure comes from the structural repair manual or the drawing.
• Grain direction matters because rolled sheet is not the same in both directions. A bend made across the rolling direction — so that the bend line runs at right angles to the grain — tolerates a tighter radius than one made along it. This is why sheet is marked with its rolling direction and why nesting a part on a sheet is not arbitrary.
• Yield strength sets the springback. Only the plastic part of a bend is permanent; the elastic part recovers when the tool is released. Since Young's modulus is the same for every steel, the amount of stored elastic strain depends on the yield strength — so a stronger steel springs back more. Check the direction against a case: annealed mild steel bent to 90 degrees comes off the tool at very nearly 90 degrees, while a spring steel bent over the same tool opens out noticeably, and the tool has to over-bend it to compensate.

Marking Out

Accurate marking out on bright steel needs a background that a scribed line will show against clearly. The traditional and still-standard preparation is a swabbed coat of copper sulphate solution, which reacts with the clean steel surface to deposit an extremely thin, adherent film of copper. A scribed line cuts through that dull copper film to bright metal beneath and is easy to see, and it does not rub off the way a chalk or crayon mark does.

Two related products are often confused with it. Engineer's blue is a non-drying paste used to check the contact between two mating surfaces, not to mark out. A wax crayon gives a quick temporary mark on hot or oily surfaces but is far too coarse for accurate work.

Protecting a Steel Surface

Plain carbon and low-alloy steels have essentially no natural corrosion resistance and depend entirely on an applied protective scheme; unprotected steel in a damp environment rusts, and the rust does not stop the attack the way an aluminium oxide film does. One treatment belongs here rather than with the corrosion material because it is specific to ferrous metals: phosphating. The steel is immersed in a solution of phosphoric acid and metal phosphates, which reacts with the surface to grow an integral crystalline phosphate layer. It is neither a paint nor a plating — it is a chemical conversion of the steel itself — and it does two jobs: it inhibits corrosion in its own right, and its crystalline texture provides an excellent mechanical key for the paint scheme applied over it. Note that other acid baths are used for other purposes; a nitric or sulphuric acid solution is a pickling or cleaning bath, not a phosphating one. The full range of protective treatments, the plating systems and the corrosion mechanisms themselves are covered with the corrosion material.

Handling and Storage

• Keep grades and heats segregated, and transfer the identification to the remainder whenever stock is cut.
• Protect bare machined surfaces. A finish-machined high-strength steel part with no protective finish will begin to rust in hours in a humid store. Oil, vapour-inhibiting paper or an approved temporary coating is applied immediately.
• Avoid contact with dissimilar metals in storage and in assembly, and keep steel swarf and steel wool away from aluminium structure.
• Treat mechanical damage seriously. A nick, a scratch or a tool mark on a high-strength steel part is a stress raiser and a fatigue origin, and the higher the strength of the steel the more sensitive it is to one. Landing-gear components are protected in transit for this reason and not merely to keep them looking clean.

Testing of Ferrous Materials

Material testing answers two quite different questions, and it is worth being clear which one is being asked before choosing a test.

• What are this material's properties? This is the development and qualification question, answered once for a grade and condition and then written into a specification.
• Is this particular batch, or this particular part, what it is supposed to be? This is the acceptance and verification question, and it is the one an engineer meets in service. It is usually answered by a small number of quick, targeted checks rather than by a full characterisation.

Tests also divide by what they cost you:

• Destructive tests — tensile, impact, fatigue, bend, metallographic examination — consume the specimen. They are therefore never carried out on the part itself. Instead a test coupon, cut from the same bar or forging, from the same heat of steel, and passed through the same heat-treatment furnace load, travels with the batch and is tested in its place. The coupon is the evidence, and this is the practical reason heat and cast numbers have to be tracked so carefully.
• Non-destructive tests — the whole range of crack-detection methods, and hardness testing, which leaves only a very small indentation — can be applied to the part itself. Hardness is the workhorse for exactly that reason.

One point applies to every test in this section: a result only means anything if the test is standardised. The specimen dimensions, the notch geometry, the indenter, the applied load, the rate of loading, the dwell time and the temperature are all laid down, because changing any of them changes the number. A hardness figure quoted without its scale, or an elongation figure quoted without its gauge length, is not a measurement — it is a number without units.

The four families that follow each answer a different question. Hardness asks how well the surface resists being pushed in, and by proxy how strong the material is and whether it was heat treated correctly. The tensile test asks how much load the material carries and how much it stretches first. Fatigue testing asks how long it survives loads that are repeated. Impact testing asks how much energy it absorbs when the load arrives suddenly. A material can pass any one of them and fail another, which is why all four appear in a specification.

Hardness Testing

• Brinell test: A hardened steel or tungsten carbide ball is pressed into the surface under a known load. The diameter of the indentation is measured and converted to a Brinell Hardness Number (BHN). Suitable for large, rough surfaces.
• Rockwell test: A diamond cone (Rockwell C scale) or steel ball (Rockwell B scale) is pressed into the surface. Hardness is read directly from the machine dial based on depth of penetration. Fast and widely used.
• Vickers test: A diamond pyramid indenter is pressed into the surface. The diagonal of the resulting square indentation is measured. Very accurate; suitable for thin materials and surface-hardened layers.

All three of the methods above work on the same principle: press a hard indenter into the surface under a controlled load, and measure how big a mark it leaves. A large mark means a soft material and a small one means a hard material. What separates the three is the shape of the indenter, the size of the load, and — the most important difference in practice — whether the indentation is measured optically after the load is removed or electrically as a depth while the load is still applied.

The Brinell Test in Detail

A ball indenter is pressed into the surface under a specified load, held there for a specified time, and removed. The diameter of the round indentation left behind is measured with a calibrated measuring microscope, and the hardness number is the applied load divided by the curved surface area of the impression.

The load is held for a specified dwell time so that the material finishes deforming and a stable, repeatable impression is produced before the load is released. The figure conventionally quoted for both the Brinell and the Vickers tests is about 15 seconds. Too short a dwell gives an undersized impression and an inconsistent, falsely high reading.

Strengths. The indentation is large, so it averages over a substantial volume of material. That makes Brinell the right test for materials that are not uniform on a small scale — castings, forgings, coarse-grained stock — where a small indenter might land entirely inside one constituent and read the wrong thing. It also tolerates a comparatively rough surface.

Limitations. The large indentation is a defect on a finished part and can only be placed where it does no harm. The test is unsuitable for thin material or for a hardened case, because the indenter reaches through the layer being tested into the softer material beneath. And there is an upper limit: above roughly 450 on the scale a hardened steel ball begins to flatten and reads low, which is why a tungsten carbide ball is used for harder materials — and even then the test is not used at the very top of the range.

The Rockwell Test in Detail

The Rockwell test measures a depth, and it does so in two stages, which is what makes it fast and free of optical measurement.

• A small minor load — conventionally 10 kgf — is applied first. This seats the indenter, beds down any surface roughness and any dirt or scale, and establishes the zero datum. Because the datum is set after this seating load, surface imperfections are largely removed from the result.
• The major load is then applied and removed. The depth the indenter has moved beyond the datum, measured with the minor load still applied, is the permanent penetration, and the machine converts it directly into a hardness number on a dial or display.

Scale	Indenter	Total (major) load	Used for
HRC	Diamond cone of 120 degrees included angle, with a small spherical tip (the "brale")	150 kgf	Hardened and tempered steels — the standard aerospace scale for heat-treated parts
HRB	1.588 mm (1/16 inch) hardened steel ball	100 kgf	Softer steels, annealed and normalised material, copper and aluminium alloys
HRA	Diamond cone	60 kgf	Very hard and thin material, cemented carbides, thin case-hardened layers
Superficial scales (15N, 30N, 45N and the corresponding T scales)	Diamond cone or ball	15, 30 or 45 kgf, with a 3 kgf minor load	Thin sheet, shallow cases, nitrided surfaces and coatings, where a normal load would punch through the layer being measured

Strengths. It is fast — a reading takes seconds; it needs no optical measurement and no operator judgement; the indentation is small; and it works over the whole range from soft to fully hardened simply by changing scale. This is why it is the test found in nearly every workshop and why heat-treatment acceptance is normally written in Rockwell.

Limitations. The part must be properly supported on the anvil, because any flexing of the specimen is measured as penetration and reads falsely soft. The surface must be reasonably clean and flat. Readings on different scales are not directly comparable. And with a thin specimen the indenter can deform the anvil face beneath — the classic "anvil effect" — which again reads falsely soft.

The Vickers Test in Detail

A diamond indenter ground as a square-based pyramid with 136 degrees between opposite faces is pressed into the surface. It leaves a square impression, and the two diagonals are measured optically and averaged.

The property that makes Vickers special follows from that geometry. Because a pyramid is geometrically similar at every size, a bigger load simply makes a bigger impression of exactly the same shape, and the calculated hardness comes out the same. So the Vickers number is independent of the load, and a single continuous scale covers everything from soft aluminium to the hardest tool steel. Brinell and Rockwell both need the scale to be changed as the material changes; Vickers does not.

Strengths. One scale for all materials; very small, precisely defined indentations that can be placed on thin sheet, on tiny features and, most usefully, in a line running inward from a surface to survey a case-hardening depth; and the highest accuracy of the three.

Limitations. The surface must be prepared to a fine, often polished, finish, because the diagonals of a small impression cannot be measured on a rough surface. The measurement is optical and therefore slower, and it depends on the operator setting the measuring lines correctly. This is a laboratory test rather than a shop-floor one.

Comparing the Three

	Brinell	Rockwell	Vickers
Indenter	Hardened steel or tungsten carbide ball	Diamond cone (C scale) or steel ball (B scale)	Diamond pyramid, 136 degrees
Quantity measured	Diameter of the impression	Depth of penetration	Diagonal of the impression
How it is read	Optically, then calculated or read from a table	Directly from the machine — no calculation	Optically, then calculated
Indentation size	Large	Small	Very small
Number of scales	One, with the load matched to the material	Many, and they are not interchangeable	One, for all materials
Best suited to	Castings, forgings, coarse or non-uniform material, rougher surfaces	Routine shop and acceptance testing across the whole hardness range	Thin sections, small parts, case-depth surveys, laboratory work
Main drawback	Large mark; upper limit on hardness; thin sections unsuitable	Needs solid support; scales not comparable with each other	Needs a prepared surface and careful optical measurement

Portable and In-Situ Hardness Testing

Bench machines require the part to be brought to them and supported rigidly, which rules out anything installed. Two families of portable instrument exist for in-situ work, and both carry conditions:

• Rebound (Leeb) testers fire a small impact body at the surface and compare its rebound velocity with its impact velocity. A hard surface returns more of the energy. The distractor in a Rockwell question that mentions "the rebound height of a ball" describes this principle, not the Rockwell one, and the classic instrument of this kind is the Shore scleroscope.
• Ultrasonic contact impedance testers press a Vickers diamond into the surface under a fixed load and detect the change in the resonant frequency of the probe rod, which depends on the contact area.

Both require the part to be massive enough and thick enough not to move or ring under the probe; on a thin panel or a light unsupported component they read low, and a correction or a fixture is required. Both need a good surface finish. They are comparative field tools, and where an acceptance figure is in dispute the bench test is the reference.

Typical Hardness Values

Some idea of the range is useful, both for sanity-checking a reading and for recognising when a part has not been heat treated as intended.

Material and condition	Approximate Brinell hardness
Annealed low-carbon steel	about 100 to 175
Normalised medium-carbon steel	roughly 170 to 220
Alloy steel, hardened and tempered for structural use	roughly 300 to 450
Cobalt-bearing tool and high-speed steel	about 600 to 700
Carburised or nitrided case	Beyond the useful Brinell range — measured in Vickers or on a superficial Rockwell scale

Tensile Testing

A standard test specimen is pulled in a tensile testing machine until it breaks. The test measures:

• Ultimate tensile strength (UTS): Maximum stress the material can withstand
• Yield strength: Stress at which permanent deformation begins
• Elongation: Percentage increase in length at fracture (measure of ductility)
• Reduction of area: Percentage decrease in cross-section at the fracture point

The Tensile Specimen

A tensile test measures a material property, so the answer must not depend on the size of the piece tested. That requires the specimen to be standardised, and in one respect the standardisation is not obvious. The gauge length — the marked length over which extension is measured — is made proportional to the cross-section, because a short specimen concentrates all of its necking inside a short gauge length and so reports a higher percentage elongation than a long one cut from identical material. The usual proportional gauge length works out at about five diameters for a round specimen.

The consequence for the engineer is a rule with no exceptions: an elongation figure is meaningless without the gauge length it was measured over. Two certificates quoting 18% and 22% elongation may describe the same steel measured to different standards.

The specimen has enlarged, threaded or shouldered ends so that it grips securely and always fails in the reduced parallel portion, and generous blend radii where the section changes so that the failure is not triggered by a stress concentration at the shoulder. It is pulled at a controlled rate — pull too fast and the measured strength comes out high — while an extensometer records the extension across the gauge length, until it fractures.

Reading the Stress-Strain Curve

Plot stress against strain and the whole mechanical character of the steel appears in one line. Take it region by region.

Region	What is happening in the metal	What is read off
Straight-line portion	Atomic bonds stretch elastically. Everything is recoverable	The slope is Young's modulus. The top of the straight line is the limit of proportionality
Elastic limit	Slightly beyond the straight line; strain is still fully recoverable, but no longer proportional	The last point from which the specimen returns to its original length
Yield	Dislocations break free and move in large numbers; deformation becomes permanent	Yield strength — the stress at which permanent deformation begins
Strain-hardening region	Dislocations multiply and tangle, so the material strengthens faster than the specimen thins	The rising portion; the material accepts increasing load
Maximum	Strengthening can no longer keep pace with the reduction in area	Ultimate tensile strength — the maximum stress the material withstands
Necking	Deformation localises into one narrow band, which thins rapidly	The curve falls, because engineering stress is calculated on the original area while the real area is shrinking fast
Fracture	The neck separates	Elongation and reduction of area, both measures of ductility

Two features of the yield region are worth separating, because they distinguish the steels an engineer meets.

• Mild steel has a genuine, visible yield point — in fact two. The load rises to an upper yield point, then drops abruptly to a lower yield point at which the specimen extends appreciably with no increase in load at all. The cause is that interstitial carbon and nitrogen atoms have gathered around the dislocations and pinned them; once the stress is high enough to tear the dislocations free, they move much more easily, so less stress is needed to keep them going. This is a low-carbon-steel phenomenon and it produces the visible surface bands seen on mild steel pressings.
• Heat-treated alloy steels have no distinct yield point. The curve simply bends over smoothly, and there is no single point that can be identified as "yield". For these the specification quotes a proof stress instead: the stress that leaves a defined permanent strain — commonly 0.1% or 0.2% — after unloading. It is found by drawing a line parallel to the elastic portion, offset along the strain axis by that amount, and reading where it cuts the curve. The 0.2% proof stress is what "yield strength" means on a certificate for an alloy steel.

What the Numbers Are Used For

• Proof or yield stress is the practical design limit for a structural part. A component that has yielded is out of tolerance and is unserviceable, even though it has not broken — which is exactly why a bolt stretched beyond its elastic limit is scrapped rather than reused. It deforms permanently; it does not spring back, and being stretched does not make it more ductile.
• Ultimate tensile strength sets the ultimate load capability and is the figure the hardness correlation predicts.
• Elongation and reduction of area are the ductility measures, and they are the numbers that tell you whether the material will give warning before it fails. A steel with high strength and 2% elongation is a very different proposition from one with the same strength and 15%. Reduction of area is the more sensitive of the two, because it measures what happened at the neck rather than averaging over the whole gauge length.

Heat treatment moves the whole curve in a predictable way. Annealed steel gives a low, long curve: low strength, large elongation. Quenched and lightly tempered steel gives a high, short one: high strength, very little elongation, and a fracture with no necking at all. Increasing the tempering temperature moves the curve progressively from the second shape towards the first. The tensile test therefore verifies a heat treatment directly, and does it more completely than a hardness check because it reports the ductility as well.

Finally, the broken specimen is itself evidence. A ductile steel necks visibly and breaks with a fibrous, grey cup-and-cone fracture, the cone of one half fitting the cup of the other. A brittle one shows no necking, breaks square across, and leaves a bright, faceted, crystalline surface. The appearance of a fracture face is the first thing an investigator looks at, and the same distinction applies to a service failure as to a test piece.

Fatigue Testing

A specimen is subjected to repeated cyclic loading (bending, torsion, or axial) at a specified stress level until it fails. The number of cycles to failure is recorded. Testing at various stress levels produces an S-N curve (stress vs number of cycles). Steel typically has a definite fatigue limit — a stress level below which it can endure an infinite number of cycles without failing.

How a Fatigue Failure Actually Develops

Fatigue is not a gradual weakening of the whole part. It is a single crack, growing a little on each load cycle, in three distinct phases:

• Initiation. A crack starts at the point of highest local stress, which is almost always at a surface and almost always at a stress raiser — a fillet, a hole, a thread root, a keyway, a tool mark, a corrosion pit, an inclusion just below the surface. This phase can occupy the great majority of the component's life and produces nothing detectable.
• Propagation. The crack extends a minute distance on each cycle, opening under the tensile part of the cycle and closing again. The two faces rub as they open and close, so the fracture surface produced in this phase is smooth and often burnished.
• Final fracture. Eventually the remaining uncracked section is too small to carry the peak load, and it fails in one cycle as a straightforward overload. This last stage is instantaneous.

The fracture face records all of this and is the most useful single piece of evidence in a failure investigation:

Feature	What it looks like	What it tells you
Origin	A single point, often at a corner, a hole edge or a change of section, from which the markings radiate	Where the crack started, and therefore what caused it
Beach marks (progression marks)	Smooth, curved bands like the tide lines on a beach, curving away from the origin	Interruptions or changes in the loading — each band is a period of different service. Visible to the naked eye
Striations	Extremely fine parallel lines, visible only under an electron microscope	One line per load cycle — they allow the crack growth rate to be measured directly
Ratchet marks	Small steps running back towards the surface at the origin end	Several cracks started at slightly different points and merged — a sign of a high stress or a severe stress raiser
Final overload zone	A rough, dull, fibrous or crystalline area quite unlike the smooth fatigue region	The part that broke in one go. Its size relative to the smooth region indicates the stress level: a small final zone means the crack grew a long way before the part could no longer carry the load, so the stress was low; a large final zone means the load was high and only a short crack was needed

The S-N Curve, the Endurance Limit and What Modifies It

Testing many specimens at different stress levels and plotting stress against cycles to failure produces the S-N curve. For steel the curve falls steeply at first and then flattens into a horizontal line — the "knee" — typically somewhere between one million and ten million cycles. The stress at that flat portion is the fatigue limit, also called the endurance limit. Below it the steel survives indefinitely; above it the life is finite and falls sharply as the stress rises. The stress level, not the proof stress, is what governs which of those two regimes a part is in.

The useful rule of thumb is that for most steels the fatigue limit in fully reversed loading lies at roughly 40% to 60% of the static ultimate tensile strength, and it is commonly taken as one half. Two qualifications belong with it. Above a tensile strength of roughly 1,400 MPa the fatigue limit stops rising in proportion — making a steel stronger past that point buys very little extra fatigue performance while making it much more sensitive to notches. And the flat knee is characteristic of ferrous materials tested in air: in a corrosive environment it disappears entirely and the curve keeps falling, so a steel exposed to corrosion has no true endurance limit at all. The behaviour of the light alloys, which do not show the knee in the first place, is covered with the non-ferrous materials.

The S-N data themselves come from small, carefully polished laboratory specimens, and a real component never matches one. The published figure has to be reduced to allow for the differences:

Influence	Direction of the effect on fatigue strength	Why
Rougher surface finish	Reduces it	Every machining mark is a small notch at the surface, and the surface is where cracks start
Notches, holes, fillets, thread roots, keyways	Reduces it, often severely	They multiply the local stress well above the nominal value
Larger component size	Reduces it	A bigger part contains more material at high stress, so it is more likely to contain a critical defect
Corrosion, or a corrosive environment	Reduces it, and removes the endurance limit	Pits act as notches and the environment keeps attacking the freshly exposed crack tip
Fretting at a clamped joint	Reduces it	Micro-movement damages the surface and generates debris, creating an origin in an area under high clamp-up
Decarburisation of the surface	Reduces it	The surface layer is left softer and weaker than the material beneath it, exactly where the stress peaks
Tensile residual stress at the surface	Reduces it	It adds directly to the applied tensile stress
Shot peening, nitriding, cold rolling of fillets and threads	Increases it	All three leave the surface in compression, which the applied tension must first cancel before the metal sees any net tension
A polished surface, and clean vacuum-remelted steel	Increases it	Fewer surface notches and fewer internal inclusions means fewer places for a crack to start

How the Test Is Run

• Rotating bending. The classic machine, and the one the original S-N work was done on. A round specimen is loaded in bending and rotated, so every point on its surface passes from maximum tension to maximum compression once per revolution. This gives fully reversed loading — equal tension and compression, zero mean stress — which is the reference condition most published data are quoted for.
• Axial (push-pull). The specimen is pulled and pushed along its axis, which allows any combination of mean and alternating stress to be applied. This matters because a tensile mean stress lowers the fatigue strength while a compressive mean stress raises it, and most real components carry a mean load rather than a purely alternating one.
• Torsional and resonance machines apply the loading in other modes, and resonance machines run at high frequency to accumulate very large cycle counts in reasonable time.

Two features of fatigue testing catch people out. First, the scatter is large: two identical specimens at the same stress can differ in life by a factor of several, because the outcome depends on where a crack happens to start. Fatigue data are therefore statistical and require many specimens, and design values are set at a low probability of failure rather than at the average. Second, a specimen that survives a pre-agreed number of cycles is stopped and recorded as a run-out rather than being run to destruction, since no test can literally demonstrate infinite life.

In service, all of this appears as the life limits and inspection intervals on the maintenance schedule. A component whose failure would be critical is either given a safe life — retired after a set number of cycles, well short of the demonstrated fatigue life, whether or not anything has been found wrong with it — or is designed to be damage tolerant, so that a crack can be found by inspection and the structure will carry its load until it is. Heat treatment cannot restore fatigue life that has already been consumed, and no maintenance action resets the count; this is why life-limited parts are tracked individually by serial number.

Impact Testing

• Charpy test: A notched specimen is placed as a beam and struck by a heavy pendulum. The energy absorbed in breaking the specimen (measured in joules) indicates toughness. Low values indicate brittle material.
• Izod test: Similar to Charpy but the specimen is held as a vertical cantilever. Less commonly used than Charpy in aviation.

Charpy and Izod Compared

Both tests swing a calibrated pendulum hammer into a notched bar and measure the energy the bar absorbs in breaking. The energy is found from the difference in the pendulum's height before and after the blow: whatever it fails to carry through to the far side of its swing went into fracturing the specimen. The machine reads it directly in joules. The differences between the two tests are in how the specimen is held and where it is struck.

	Charpy	Izod
Specimen	Square bar 10 mm × 10 mm × 55 mm long	Square bar 10 mm × 10 mm × 75 mm long
How it is held	Laid horizontally and simply supported at both ends, like a beam	Clamped vertically in a vice, as a cantilever
Notch position	At mid-span, on the face away from the striker	Level with the top of the vice jaws, facing towards the striker
Where the hammer lands	At mid-span, on the face opposite the notch	Above the notch, on the same side as the notch
Striking energy of the standard machine	Commonly around 300 J	Approximately 163 J
Practical position	The international standard for structural steels, and the easier of the two to run at controlled sub-zero temperatures because the specimen is simply laid on the anvil and struck within seconds of leaving its cooling bath	Long used in Britain; less common in aviation work, and awkward for low-temperature testing because clamping the specimen in a vice takes time during which it warms

The notch is not incidental — it is the whole point of the test. An unnotched ductile bar would simply bend around the hammer. The notch forces the fracture to start at a known place and, more importantly, creates the sharp, three-dimensionally constrained stress state in which even a normally ductile steel can behave in a brittle manner. What the test measures is therefore not raw toughness but notch toughness: how the material behaves when there is already a defect in it, which is the realistic case. The notch geometry is standardised precisely because a sharper notch gives a lower result on the same material.

A result may be reported in three complementary ways, and a specification often calls for more than one:

• Absorbed energy, in joules. The primary figure. A low value means a brittle material.
• Lateral expansion — how much the specimen bulged sideways at the compression face before breaking. It measures how much plastic deformation actually took place.
• Fracture appearance — the percentage of the broken face that is dull and fibrous (a ductile, energy-absorbing failure) as opposed to bright and crystalline (a brittle one). It is judged against reference charts and is the most immediate of the three.

The Ductile-to-Brittle Transition

Run the same impact test on the same steel at a series of temperatures and the result is not a constant. It falls as the temperature falls, and over a fairly narrow band it falls steeply, from a high "upper shelf" value to a low "lower shelf" one. The temperature at the middle of that drop is the ductile-to-brittle transition temperature.

Direction check with a concrete case: an ordinary ferritic structural steel that absorbs a large amount of energy and breaks with a fibrous, heavily deformed face at room temperature may, at forty degrees below zero, absorb only a small fraction of that energy and break flat and crystalline with no deformation at all. Nothing about the steel has changed except its temperature. Colder means more brittle, without exception, for this class of material.

Whether a metal shows the transition at all depends on its crystal structure, and this is one of the most useful practical consequences of the phases covered earlier in this note:

• Body-centred cubic metals show a transition. That covers ferrite and martensite, and therefore plain carbon steels, low-alloy steels, and the ferritic and martensitic stainless grades.
• Face-centred cubic metals do not. Austenitic stainless steels keep their toughness all the way down to cryogenic temperatures, which is one reason they are chosen for very cold service as well as very hot.

What moves the transition temperature is worth knowing, because it is what a specification is controlling when it calls for a Charpy value at a stated temperature. Raising the carbon content, coarsening the grain, allowing phosphorus to remain, sharpening the notch, increasing the section thickness and increasing the rate of loading all push the transition to a higher temperature — that is, they make the steel brittle in conditions it would otherwise have survived. Adding nickel and refining the grain push it lower. This is why nickel steels are chosen for low-temperature duty, and it is a further reason grain refinement is valued.

The aviation relevance is direct. Airframe structure is cold-soaked at cruise altitude, ground equipment and aircraft operate in arctic conditions, and components exposed to ram air or in contact with cold fuel run well below ambient. A steel that is perfectly tough in the hangar can be at the bottom of its transition curve in service, and a designer specifies impact energy at the lowest temperature the part will see, not at room temperature.

Other Tests an Engineer Meets

The Jominy End-Quench Test

This is the standard measurement of hardenability — the depth to which a steel hardens, as opposed to the hardness it reaches. A standard bar, 25 mm in diameter and 100 mm long, is heated to its austenitising temperature and then mounted vertically in a fixture where a jet of water of specified size, flow rate and temperature plays on the lower end only. The bottom of the bar is therefore quenched violently while the top cools only in air, and every intermediate cooling rate exists somewhere along its length.

Once cold, two flats are ground along the bar and Rockwell C hardness is measured at intervals from the quenched end. Plotting hardness against distance gives the steel's hardenability curve, and it separates the two variables cleanly:

• The hardness at the quenched end is set by the carbon content, because that end has certainly cooled fast enough to form martensite.
• The rate at which the hardness falls off with distance is set by the alloy content. Two steels of the same carbon content start at the same hardness at the quenched end; the more hardenable of the two falls away more slowly, and that shallower slope is exactly what "high hardenability" means.

Bend Testing

A strip or a welded coupon is bent around a former of specified radius through a specified angle, and it passes if no crack opens on the outside of the bend. It is a simple, direct measure of ductility in the form that matters for sheet work, and it is a standard part of welding procedure qualification, where it demonstrates that the weld and its heat-affected zone are no less ductile than the parent metal.

Metallographic Examination

A sample is cut, mounted, ground, polished and then etched with a suitable reagent, and examined under a microscope. It is the only method that shows directly what the structure is, and it answers questions no mechanical test can:

• Is the structure the one the heat treatment was supposed to produce — pearlite, tempered martensite, bainite?
• How coarse is the grain? Grain size is reported as a standard number, and the convention catches people out: the higher the number, the finer the grain, because each step up doubles the number of grains counted in a given area.
• How deep is the case, and how deep is any decarburised layer?
• How clean is the steel — how many non-metallic inclusions, and of what type?
• A coarse macro-etch of a whole section, examined by eye rather than under a microscope, reveals grain flow in a forging, and with it forging laps, bursts and flow lines that have been machined through.
• An etch applied to the surface of a finished ground part reveals grinding damage as patches of differing shade, as described earlier under grinding: an over-tempered area is attacked more readily and comes up dark, while a re-hardened skin of untempered martensite resists the etch and comes up white.

Creep and Stress-Rupture Testing

A specimen is held at a constant load and a constant elevated temperature, and its extension is recorded against time. A creep test runs long enough to establish the steady, secondary-stage strain rate, which is the number used to calculate how long a part will last before it stretches beyond its allowable limit. A stress-rupture test simply runs to failure and records the time taken. Both are needed to set the temperature and stress limits on hot-section components, and both take a very long time, which is why the data are extrapolated using established time-temperature relationships rather than measured directly at service durations.

Testing the Part Rather Than the Material

Two categories of test apply to a finished component instead of a specimen. A proof load test applies a specified load, higher than the working load but below the design yield, and the part passes if it shows no permanent set and no damage. It is used on shackles, lifting equipment and certain fittings, and because it consumes a little fatigue life it is done only where the design data permit it. And the whole family of non-destructive testing methods — penetrant, magnetic particle, eddy current, ultrasonic and radiographic inspection — look for cracks and internal defects in the actual part without harming it. Which of them can be used on a given item depends on the material: magnetic particle inspection needs a ferromagnetic material, eddy current needs an electrical conductor, and penetrant needs only a non-porous surface but finds only surface-breaking defects. The methods themselves are covered with the maintenance practices material.

Choosing the Right Test

The question being asked	The test that answers it	What you get
Was this part heat treated to the correct condition?	Hardness, normally Rockwell C	A number to compare against the range on the drawing, obtained in seconds without damaging the part
How deep and how hard is the case?	A Vickers hardness traverse across a sectioned coupon	The hardness profile, and from it the effective and total case depth
How strong is the material, and how much warning does it give?	Tensile test	Proof stress, ultimate tensile strength, elongation and reduction of area
Will it survive repeated loading, and for how long?	Fatigue test	The S-N curve and the fatigue limit
Will it take a sudden blow, and at what temperature does it stop doing so?	Charpy or Izod impact test at a series of temperatures	Absorbed energy and the ductile-to-brittle transition temperature
How deeply will this grade harden in a given quench?	Jominy end-quench test	The hardenability curve
What is this material?	Optical emission spectrometry, or portable X-ray fluorescence for the alloying elements alone	A chemical analysis
Is the internal structure right, and is there decarburisation, grinding damage or coarse grain?	Metallographic examination and surface etching	The microstructure, the grain size, and the depth of any affected layer
Will it stretch under a steady load at temperature?	Creep and stress-rupture testing	The secondary creep rate and the time to rupture
Is there a crack in this particular part?	The appropriate non-destructive method for the material	The presence, position and extent of the defect