Materials Science · Metals, alloys and microstructural design

Grains, dislocations and strengthening

Explain how grain boundaries, dissolved atoms, second-phase particles and other dislocations resist slip. Derive a bounded pinning model, compare grain-size strengthening with total yield stress, and use processing and structural evidence to test a mechanism.

Useful background: Atoms, bonding, structure and defects; Processing, structure and useful properties. Each worked case explains the quantities it uses.

Back to the chapter

After reading, you can

Explain strengthening through obstacles to plastic deformation.
Distinguish grain-boundary, solid-solution, precipitation and work-hardening mechanisms.
Derive and apply a supplied line-tension model with consistent shear-stress and spacing units.
Use a finite-range grain-size relationship without confusing its strengthening term with total yield stress.
Judge a proposed mechanism using structural evidence, matched conditions and the actual deformation route.

Sliding happens through defects

Plastic flow in many crystalline metals proceeds through dislocation movement. Increasing the difficulty of that movement can raise yield stress. Grain boundaries, solutes, precipitates and other dislocations can act as obstacles through different mechanisms.

A wire can become harder to bend again after it has been plastically bent, even though no new element was added. The change need not be stronger bonds between every pair of atoms. It can be a changed population and arrangement of obstacles to the motion that permits permanent slip.

A dislocation is a line bounding slipped and unslipped regions on a crystal plane. As it glides, neighbours rearrange locally and the slipped area grows. A complete passage can leave a relative displacement b, the magnitude of the Burgers vector for that slip. This avoids moving an entire plane of atoms over its barrier simultaneously. Defects, boundaries and interfaces develops the structural picture; Applied force and stress over an area introduces force-per-area units.

Shear stress τ acts tangentially to the chosen slip plane in the slip direction. It is the local driving component used below, not automatically the tensile stress σ applied to a polycrystal. Crystal orientation and the neighbouring grains affect how applied loading drives its slip systems. More resistance to one line segment is therefore a mechanism contribution, not by itself a complete tensile yield criterion.

Four obstacles resist motion in different ways

A grain boundary joins differently oriented crystals. The easy slip planes and directions on one side generally do not continue unchanged on the other. Slip transmission or activation of a new dislocation in the neighbouring grain needs a compatible rearrangement and sufficient local driving stress. Dislocations can accumulate near a boundary and interact through their stress fields. Shortening a grain can change that accumulation and the distance travelled before another boundary is encountered. The boundary is not an impenetrable wall, and not every boundary has the same resistance.

A solute is an atom of a dissolved species within the matrix phase. A substitutional solute occupies a normal lattice site; an interstitial lies between normally occupied sites. Its local size, bonding and elastic response can alter the energy of a nearby dislocation. If a line settles into a favourable association, moving it away requires additional work. These interactions can resist motion without a separate particle phase. Solute strengthening is not a row of literal hard balls: dislocation character, species, concentration and temperature affect the interaction.

A precipitate is a second-phase particle formed within the matrix. A dislocation may cut a shearable particle, changing its interfaces, internal order or elastic distortion. For example, slip through an ordered particle can misalign the ordered sequence across the cut; making that disturbed boundary costs energy. Alternatively, a sufficiently resistant particle can pin the line while the segment between particles bows. On completing bypass, a dislocation loop can remain around the particle. Cutting and bypass are different routes, with different size and spacing dependence.

The easier available route matters. Making a shearable particle larger can increase the cost of cutting in a particular regime, yet a coarser, more widely spaced population can make bypass easier. Coherence describes how well atomic lattices match across a particle–matrix interface; its strain and its loss can also change resistance. There is no universal particle radius at which all alloys switch route. Volume fraction alone does not specify the actual gaps on a slip plane.

Other dislocations also exert stress fields and intersect moving lines. A forest means dislocations crossing the active slip plane. Intersections can form junctions or steps that obstruct further glide, depending on their geometry. Deformation can create, store and rearrange dislocations, increasing such interactions: work hardening. The defects that permit slip can thus also obstruct later slip. Not every intersection is an identical permanent lock, and not every stored line remains an equally effective obstacle.

These mechanisms are distinguished by what the obstacle is and how the line passes it. An optical grain image does not resolve every dissolved atom, small precipitate or dislocation. Use an observation method capable of resolving the claimed feature, and retain phase, particle-distribution and deformation-history evidence when those features can also change.

A structural explanation needs more than the word obstacle. Each route is conditional on the actual crystal, state and loading.
Mechanism	Feature resisting glide	Useful structural distinction
Grain-boundary strengthening	An orientation-changing junction; continuation needs compatible slip	Grain size, orientations and boundary character
Solid-solution strengthening	Energy interaction with dissolved species in the matrix	Species and dissolved concentration, not just total alloy content
Precipitation strengthening	Cutting a particle or bowing between resistant particles	Phase, size, coherence and actual slip-plane spacing
Work hardening	Fields, junctions and steps involving other dislocations	Stored dislocation arrangement and prior deformation

Derive one spacing effect from a pinned line

Consider a deliberately idealised segment on one glide plane, fixed at two strong point pins separated by λ along their straight connecting chord. Here λ is the distance between the pinning points, not a grain diameter or a particle centre spacing. Between the pins the line can glide but cannot cut through, escape the plane or release a pin. We follow its quasi-static minor circular-arc equilibria up to a semicircle.

Supply a positive, constant line tension T. In this model an additional line length Δℓ costs energy TΔℓ, so the line tends to straighten. T has units J/m = N; it is a force, not a stress. Real dislocation energy and effective line tension depend on elastic properties, line character and surroundings. Treating T as constant makes this a model, rather than an exact law for every bowed mixed-character line.

The driving force per unit line length is τb. One way to see its meaning is work: advancing a straight length ℓ by dx sweeps area ℓdx and leaves slip b, so the shear work is τbℓdx. Dividing by the advance gives force τbℓ. Thus τb has units Pa × m = N/m. For a small circular arc, the transverse components of the two line tensions balance this distributed drive: T/R = τb, where R is the radius. Equivalently τ = T/(bR).

For a circle joining the fixed pins, let θ be half the angle subtended by the arc. The chord gives λ = 2R sin θ. The bowed minor-arc branch has 0 < θ ≤ 90°; the straight zero-drive line is the limit as θ tends to zero and R increases without bound. Its largest supported drive occurs at the semicircle: R = λ/2. The model threshold is therefore τ_c = 2T/(bλ). Halving spacing doubles this threshold when T, b and pin strength remain unchanged. Beyond it there is no minor-arc equilibrium of this type; that is not proof that the entire metal has yielded or fractured.

Choose original teaching inputs T = 1.20 × 10⁻⁹ N and b = 0.250 nm = 2.50 × 10⁻¹⁰ m. At λ = 100 nm = 10⁻⁷ m, τ_c = 2.40 × 10⁻⁹/(2.50 × 10⁻¹⁷) = 9.60 × 10⁷ Pa = 96.0 MPa. At 50 and 200 nm the model gives 192 and 48.0 MPa. These are invented pinning inputs, not measured properties of a named alloy.

The figure holds the same two pins 100 nm apart. At θ = 30°, R = 100 nm and τ = 48.0 MPa. At θ = 60°, R ≈ 57.735 nm and τ ≈ 83.138 MPa. At 90°, R = 50.0 nm and τ = 96.0 MPa. A larger applied shear produces a more strongly bowed line with a smaller radius, not a longer pin spacing.

This construction isolates a spacing effect relevant to strong pinning. It does not decide whether a real particle will instead be cut, whether thermal activation permits escape, or how many slip systems carry a polycrystal’s deformation. It omits changes in line tension, interactions between many segments and the evolving loops left after bypass. Do not add this local shear threshold to the tensile grain-size model below as if the two were measured independent pieces of one alloy strength.

Independent values for the constant-line-tension pinning model. T = 1.20 nN and b = 0.250 nm remain fixed. No tensile yield stress is inferred.
Pin spacing λ (nm)	Semicircle radius (nm)	τ_c (MPa)
50	25	192
100	50	96.0
200	100	48.0

Two fixed point pins A and B are 100 nm apart on one slip plane. At half-angle θ of 30, 60 and 90 degrees, minor circular arcs have radii 100, about 57.735 and 50.0 nm, and require resolved shear 48.0, about 83.138 and 96.0 MPa. The 30 degree line is solid blue, the 60 degree line long-dashed violet and the 90 degree line short-dashed orange. A radius joins the chord centre to the semicircle apex; it is half the 100 nm chord. With supplied constant line tension T = 1.20 nN and slip displacement b = 0.250 nm, force balance gives τ = T/(bR) and the semicircle threshold τ_c = 2T/(bλ). The lower plot covers only pin spacings 50–200 nm: thresholds are 192, 96.0 and 48.0 MPa at 50, 100 and 200 nm. These are model shear thresholds, not bulk tensile yield stresses.

A grain-size term is not the whole yield stress

For many conventional polycrystalline metals under suitable conditions, a useful empirical form is the Hall–Petch relationship σ_y = σ₀ + k/√d. Here d is a stated representative grain size, σ_y a consistently defined tensile yield stress, σ₀ a fitted baseline for the controlled state, and k the grain-size coefficient. The formula is not obtained by replacing λ with d in the pinning equation. Grain-boundary transmission and plastic accommodation involve another physical problem.

The boundary picture explains why size can matter: a smaller grain interrupts slip over a shorter distance, and can reduce the dislocation accumulation available to drive slip in the next grain. It does not prove the square-root exponent for every structure. Use the relation only where its form and parameters are supported for the material, processing state, temperature, rate and grain-size convention. Cold-worked dislocation boundaries are not automatically interchangeable with the grain boundaries of an otherwise matched, undeformed state.

Use a separate invented example in which σ_y = 60.0 MPa + (0.400 MPa·m¹ᐟ²)/√d is stipulated over d = 25–100 μm only. Composition, phases, other defect state, orientation distribution, test temperature, rate and yield definition are held comparable. Here the 60.0 MPa baseline includes the fixed contributions of this model; it is not an experimentally separated universal lattice resistance. The coefficient’s length unit requires d in metres.

At d = 100 μm = 10⁻⁴ m, √d = 0.0100 m¹ᐟ², so the size term is 40.0 MPa and total yield stress 100 MPa. At 25 μm, √d = 0.00500 m¹ᐟ²: the term is 80.0 MPa and total 140 MPa. Fourfold refinement doubles this term but raises the total by 40%, not 100%. The fixed baseline matters.

The second figure shows only the stipulated 25–100 μm interval. It compares total stress with the baseline and displays the two contributions at three chosen sizes. It supplies no evidence that the formula remains valid at 5 nm, at another temperature or after a change in work hardening. A small-grain observation and one high strength value cannot identify the mechanism by themselves.

Invented, finite-range tensile yield model. The baseline is 60.0 MPa; all other stated conditions remain matched. These are supplied model values, not an experimental fit or an alloy catalogue.
d (μm)	d in metres	k/√d (MPa)	Total σ_y (MPa)
25	2.50 × 10⁻⁵	80.0	140
64	6.40 × 10⁻⁵	50.0	110
100	1.00 × 10⁻⁴	40.0	100

An independent invented Hall–Petch tensile model is stipulated only from 25 to 100 micrometres. Its constant baseline is 60.0 MPa, and its size term is 0.400 MPa square-root metres divided by the square root of grain size in metres. A solid curve gives total tensile yield stress; a long-dashed line gives the fixed baseline. At grain sizes 25, 64 and 100 micrometres, the size terms are 80.0, 50.0 and 40.0 MPa and totals 140, 110 and 100 MPa. Three bars use one common stress scale and separate the diagonally hatched left fixed 60.0 MPa segment from the right size-dependent segment by position. Refining 100 to 25 micrometres doubles the size term but raises total yield stress by 40 percent. These are model inputs and calculations, not measured material data or predictions outside the stated interval.

A stronger state has trade-offs

Finer grains often strengthen conventional polycrystalline metals over an appropriate range, but the trend is not universal at every size or temperature. Grain-boundary processes can matter in creep, and excessive barriers can reduce ductility. Specify which deformation mechanism and structural range the explanation concerns.

A boundary can impede glide while also providing a route for atom transport or participating in sliding and accommodation. At sufficiently different sizes or temperatures, deformation can rely more on such boundary processes or on dislocation emission and absorption at boundaries. There is no universal nanometre cutoff at which every metal reverses its size trend. Sustained loads and creep mechanisms develops the time-dependent alternatives; no creep-rate law is supplied by the short-test grain-size equation.

Structure can change during use. Recovery can remove or rearrange stored dislocations; recrystallisation can replace a deformed structure with new lower-dislocation-density grains. Particles can coarsen, transform or dissolve, changing the route and spacing resisting glide. A strong initial state is not necessarily a stable service state. Why processing history matters follows the order of operations; Thermal treatment and evolving microstructure considers the alloy-specific transformations.

Greater yield stress does not determine ductility, fracture toughness, elastic modulus or conductivity. Some obstacle changes reduce remaining tensile ductility, but neither model here calculates that reduction or makes it inevitable for every grain-refinement route. Metallic bonding and electron behaviour explains why transport and dislocation motion are separate responses to structure.

To test a proposed explanation, compare states with the same relevant chemistry, orientation, temperature, rate and yield definition, then examine the structural variable that was changed. If grain refinement also introduced cold work or altered precipitates, the stress difference cannot be assigned wholly to d. A model that matches one number is not a unique mechanism diagnosis. Retain the process record and enough spatial evidence to establish the structure represented by the mechanical test.

Worked examples

Work through the case

Cold working raises a metal’s yield stress without adding alloying elements. What mechanism can explain it?

Deformation increases and rearranges dislocations.
Their interactions impede further motion.
More stress may be required for continued plastic flow.

Work hardening can strengthen a metal through its changed defect structure.

Use spacing and force units, then test the interpretation

In the supplied constant-T model, keep T = 1.20 nN and b = 0.250 nm. A changed obstacle distribution reduces λ from 100 to 80.0 nm. Calculate the new τ_c and its percentage change. Does this prove a 25% rise in the metal’s tensile yield stress?

Convert spacing: 80.0 nm = 8.00 × 10⁻⁸ m. The denominator bλ is (2.50 × 10⁻¹⁰)(8.00 × 10⁻⁸) = 2.00 × 10⁻¹⁷ m².
The model gives τ_c = 2T/(bλ) = (2.40 × 10⁻⁹ N)/(2.00 × 10⁻¹⁷ m²) = 1.20 × 10⁸ Pa = 120 MPa.
Relative to 96.0 MPa, the factor is 120/96.0 = 1.25: a 25% increase. The spacing fell by 20%; an inverse relation does not give an equal percentage rise.
The result applies to the fixed-pin segment with the prescribed T and b. If pins release, particles are cut or another route operates first, that threshold is not the governing event. Relating it to bulk tensile yield needs the active slip systems, orientation, other obstacles and a suitable aggregate model.

The segment threshold rises from 96.0 to 120 MPa, or 25%. A corresponding percentage change in bulk tensile yield has not been established.

Separate the grain-size contribution from the total

Use only the independent grain-size model over 25–100 μm. Compare d = 100 and 25 μm, then calculate the result at an unfamiliar d = 36.0 μm. Why is doubling total stress an error?

At 100 μm, √d = 0.0100 m¹ᐟ² and k/√d = 0.400/0.0100 = 40.0 MPa. Total σ_y = 60.0 + 40.0 = 100 MPa.
At 25 μm, √d = 0.00500 m¹ᐟ² and the term is 80.0 MPa. Total is 140 MPa. Thus the term doubles while total changes by (140 − 100)/100 = 0.40, or 40%.
At 36.0 μm = 3.60 × 10⁻⁵ m, √d = 0.00600 m¹ᐟ². The term is 66.666… MPa, giving total σ_y ≈ 126.67 MPa. This size lies within the stipulated interval.
The ratio rule applies to the size-dependent term, not the fixed 60.0 MPa baseline. None of these computations supplies ductility, a processing schedule or an allowable component load.

The totals are 100, 140 and about 126.67 MPa at 100, 25 and 36 μm. Fourfold refinement doubles the grain-size term, not total stress.

Solve for size without turning a model into a recipe

For the same stipulated grain-size model and matched conditions, what d meets the model criterion σ_y ≥ 125 MPa? Does the answer specify how to process a real alloy?

Subtract the fixed baseline: the size term must satisfy 0.400/√d ≥ 125 − 60.0 = 65.0 MPa, using d in metres.
With positive d and k, this is √d ≤ 0.400/65.0 m¹ᐟ² and d ≤ (0.400/65.0)² m ≈ 3.78698 × 10⁻⁵ m = 37.8698 μm.
Intersect that inequality with the stated domain. The model qualifies 25 μm ≤ d ≤ about 37.87 μm; it says nothing about sizes below 25 μm. Substitution at the boundary recovers 125 MPa.
A real process must achieve and retain the required structure and other controlled state variables, with measured property evidence. The reverse calculation neither selects a treatment nor accounts for uncertainty, ductility, crack resistance or component safety.

Within this model’s domain, 25–about 37.87 μm meets the stated yield criterion. It is a conditional model interval, not a production or safety specification.

Try the reasoning

Is “smaller grains always stronger” an unrestricted law?
Show the answer and reasoning
No.
Mechanisms can change with size, temperature and loading. Boundary-assisted transport or sliding, boundary emission of dislocations and a changed stored-defect structure can change the response. Apply the supplied grain-size model only in its stated 25–100 μm domain, not at every smaller size.
Can precipitates strengthen without being the majority phase?
Show the answer and reasoning
Yes.
Their distribution and interaction with dislocations can matter. A small fraction can intersect many slip paths when distributed as suitable particles. The amount alone does not specify their spacing, shearability or phase-interface structure.
For the constant-T segment, λ = 125 nm and the other inputs are unchanged. Find τ_c. At applied resolved shear 70.0 MPa, is this model segment already beyond its semicircle threshold? Would that answer rule out every microscopic motion in the specimen?
Show the answer and reasoning
τ_c = 76.8 MPa. The 70.0 MPa drive is below this threshold; it does not rule out bowing of this segment or motion elsewhere.
Use 2.40 × 10⁻⁹/[(2.50 × 10⁻¹⁰)(1.25 × 10⁻⁷)] = 7.68 × 10⁷ Pa. Subcritical drive can change the equilibrium curvature. Other segments can have different spacing, pin strength and orientation; this is not a universal no-motion criterion.
A calculation uses λ = 100 and b = 0.250 in τ_c = 2T/(bλ), but keeps T = 1.20 × 10⁻⁹ N and reports the result in pascals. Repair the units and explain why λ is not simply a precipitate diameter.
Show the answer and reasoning
Use λ = 10⁻⁷ m and b = 2.50 × 10⁻¹⁰ m to obtain 96.0 MPa. λ is the separation between the stipulated pinning points on a slip-plane chord, not a particle diameter.
The product must have units m² for N/m². Unconverted numerical lengths expressed in nanometres do not supply that product. Particle size, centre spacing and clear distance between actual pins are different measurements; the last is independently supplied in this model.
Two specimens contain the same precipitate volume fraction. One has closely spaced small particles and the other fewer coarse particles. A report says their strengthening must match. What is missing, and can “smaller always better” replace that claim?
Show the answer and reasoning
Their actual slip-plane spacing, particle phase/interface state and cutting-versus-bypass resistance are missing. Neither equal fraction nor “smaller always better” fixes the strength.
Cutting can be easier for some small shearable particles; wider gaps can make bypass easier for a coarser population. Identify the operative route and compare its resistance under matched conditions. The fixed-pin equation applies only after its strong-pin assumptions are established, not to every particle population.
In the stipulated grain-size model, calculate σ_y at d = 81.0 μm. A second calculation inserts 81 directly and then doubles the total when d is reduced by four. Identify both errors.
Show the answer and reasoning
At 81.0 μm the term is about 44.444 MPa and total about 104.444 MPa. Use metres in the coefficient and scale only the size term. A fourfold reduction to 20.25 μm also leaves the stated domain.
√(81 × 10⁻⁶ m) = 0.00900 m¹ᐟ², so 0.400/0.00900 = 44.444… MPa. The 60.0 MPa baseline is not multiplied by the size factor. Although a formula can be evaluated algebraically at 20.25 μm, this lesson has supplied no supported prediction there.
A controlled specimen is a single matrix phase with more dissolved substitutional species, and its resistance to slip rises. Must unseen second-phase particles explain the change? What evidence would distinguish the proposed routes?
Show the answer and reasoning
No. A solute–dislocation energy interaction is a distinct possible explanation. Establish dissolved composition and use phase/structural observations capable of detecting the relevant particles before assigning precipitation strengthening.
A substitutional species need not form a separate phase to affect a dislocation. Conversely, failing to see particles in an insufficiently resolving image does not prove their absence. The single-phase condition in this question is supplied; in an experiment it needs appropriate evidence, together with matched state and test conditions.
A finer-grained sample is stronger after a process that also adds substantial cold work. The analyst subtracts the coarse sample’s yield stress and calls the entire difference grain-boundary strengthening. Is that attribution established? Propose a better comparison.
Show the answer and reasoning
No. The changed dislocation structure is a competing contribution. Compare states with other relevant defects, phases, orientation and test conditions controlled, and measure the structural features that can have changed.
The same observed increase can be consistent with more than one mechanism. Grain size alone does not count forest junctions or determine their resistance. A treatment used to remove cold work can itself change grains and particles, so verify the final matched states rather than assuming the operation isolates one variable.

What to carry forward

Name the obstacle and the motion it impedes.
A strengthening mechanism has a range of applicability.
Dissolved solutes, particle phases, orientation-changing boundaries and forest dislocations are different obstacles.
A shorter fixed-pin distance raises the supplied segment threshold; its shear stress is not automatically a bulk tensile yield stress.
A grain-size term and total yield stress scale differently when a nonzero baseline is present.
Use matched conditions and resolved structural evidence; a property increase alone does not identify its mechanism or service stability.

References

Teaching resources and research supporting this lesson’s explanations and examples.