Metals & alloys — Al alloys (2024, 7075), Ti alloys (Ti-6Al-4V), Ni superalloys (Inconel, Hastelloy), stainless steels: Visual walkthrough

Step 1 — What is "strength", really? A single line that slides

WHAT: We picture the metal not as a solid block but as a grid with one defect line in it.

WHY: Because "strength" is not about breaking atomic bonds all at once — that would need enormous force. Real metals give way far more easily, and the reason is that this line only has to move a little bit at a time. So the honest question "how strong is this metal?" becomes the sharper question: ==how hard is it to make this line move?==

PICTURE: The extra half-plane of atoms (amber) ends at the dislocation line. A small push $τ$ (the sideways shear stress, force per unit area that slides layers past each other) walks it to the right.

Figure — Metals & alloys — Al alloys (2024, 7075), Ti alloys (Ti-6Al-4V), Ni superalloys (Inconel, Hastelloy), stainless steels

Figure s01 — Alt-text: a cyan atomic grid with one amber vertical half-column ending mid-crystal at the dislocation line; a white arrow labelled "push τ" points right; a caption marks the atomic step size b. The picture shows a single defect line about to glide.

Step 2 — Put up a wall: the grain boundary

WHAT: We drop a grain boundary into the path of our gliding line.

WHY: A dislocation lives inside one grain's neat grid. At the boundary the grid direction suddenly changes, so the line cannot simply continue — the slip plane it was riding stops existing on the other side. The boundary is a wall.

PICTURE: Two grains, tilted grids, an amber dislocation gliding right until it slams into the cyan boundary and halts.

Figure s02 — Alt-text: two crystal grains with differently-tilted white atom grids meet at a thick cyan vertical boundary. An amber dot (the dislocation) sits just left of the boundary, halted. A cyan double-headed arrow across the whole left grain is labelled "d = grain diameter", making clear d spans the grain, not the thin seam.

Step 3 — The pile-up: how many cars fit behind the light?

WHAT: We count how many dislocations, $n$ , can queue up between the source that makes them and the boundary that stops them.

WHY: Each dislocation in the queue repels the ones behind it (they carry the same "sign" of strain, like same-pole magnets). The applied stress $τ$ shoves them forward; their mutual repulsion pushes them back. They settle when these balance.

Intuition Where the formula comes from (why

$n \propto τ d$ , and where $π /2$ hides) Build it in three honest pieces — see the labelled sketch below.

The forward shove. The applied stress $τ$ acting on the line's footstep $b$ gives a driving force per unit length of about $τ b$ on each dislocation. Bigger $τ$ → bigger shove.
The back-repulsion. Two like dislocations a distance $x$ apart repel. For the simplest case — screw dislocations in an ideal isotropic medium — the force per unit length is exactly $\frac{G b ^{2}}{2 π x}$ . For edge dislocations the same $1/ x$ shape holds but the prefactor carries an extra $\frac{1}{1 - ν}$ , where $ν$ is Poisson's ratio (how much a material bulges sideways when squeezed, typically $ν \approx 0.3$ ). So the honest edge law is $\frac{G b ^{2}}{2 π ( 1 - ν ) x}$ . Either way the shape is $1/ x$ and the stiffness $G$ sits on top — that $2 π$ (and, for edges, the $1 - ν$ ) is where the final numerical prefactor is born. Stiffer metal (bigger $G$ ) repels harder, so fewer fit.
Fill the runway. Balancing "shove crams them in" against "repulsion spreads them out" along a runway of length $d$ , the equilibrium queue holds a number proportional to (shove) × (runway) ÷ (stiffness × footstep). Doing the elasticity integral exactly turns the "proportional to" into a clean prefactor $π /2$ (screw case). So the $τ$ and the $d$ are pure physics (push × runway); the numerical prefactor is just the leftover of the $1/ x$ repulsion law (with its $2 π$ and, for edges, $1 - ν$ ) after integrating along the pile-up. You never need to memorise the prefactor — only that $n$ grows with $τ d$ and shrinks with $G b$ .

PICTURE: A queue of amber dislocations packed against the boundary; the queue length is the grain size $d$ , and more of them fit when $d$ is larger. Small arrows show each pair repelling (spreading them out) while the applied $τ$ shoves them forward.

Figure s03 — Alt-text: a row of amber dots (dislocations) queued left-to-right against a cyan boundary on the right. A white arrow "push τ" enters from the left source; small opposing cyan arrows between neighbouring dots show mutual repulsion. A cyan double-arrow labelled "d" spans the queue, and an amber note reads "n ∝ τ d". The figure shows how push and runway set the queue length.

Formula Number of piled-up dislocations (screw, isotropic case)

$n = \frac{π τ d}{2 G b}$ Term by term, right where it sits:

$n$ — how many dislocations are in the traffic jam.
$τ$ (top) — the push (piece 1 above); harder push crams in more of them, so $n$ grows with $τ$ . ✔
$d$ (top) — the grain diameter = runway length (piece 3); a wider grain is a longer runway, so $n \propto d$ .
$G$ (bottom) — the shear modulus, the metal's stiffness against sliding; it powers the back-repulsion (piece 2), so it divides.
$b$ (bottom) — the footstep size from Step 1.
$π /2$ — the geometry leftover from integrating the $1/ x$ repulsion along the runway (piece 2); it fixes the number, not the shape of the law. For edge dislocations it would carry an extra $(1 - ν)$ factor.

The one thing to carry forward: $n \propto τ d$ — the jam is longer when you push harder and when grains are bigger.

Step 4 — The jam amplifies stress at the wall

WHAT: We ask what the tip of the queue feels — the stress right at the boundary, where the front dislocation is crushed against the wall.

WHY: This is the crux. A single dislocation pushes with stress $τ$ . But a queue of $n$ of them all leaning on the front one multiplies the force there. The wall doesn't feel $τ$ ; it feels roughly $n$ times $τ$ . This is why a pile-up is dangerous — it concentrates a modest applied stress into a huge local one.

PICTURE: A gentle push $τ$ on the left; at the front of the queue the arrows fan into a huge amber "hot spot" of stress pressing on the boundary.

Figure s04 — Alt-text: a short queue of amber dots meets a cyan boundary; a small white arrow "gentle τ" enters from the left, but at the front dot many amber arrows fan inward onto a bright amber star labelled "hot spot: τ_tip = n τ". The figure shows a modest applied push amplified into a large local stress at the wall.

Now substitute the count from Step 3 into this amplification:

Step 5 — Breaking into the next grain: the critical condition

WHAT: We say the metal yields (the next grain starts to deform) exactly when the tip stress reaches a fixed critical value $τ_{c}$ — the stress needed to nucleate slip on the far side of the wall.

WHY: The boundary can only hold so much. Whatever the grain size, there is one fixed "last straw" stress $τ_{c}$ that starts a new dislocation in the neighbouring grain. So we set the tip stress equal to that constant and solve for the applied push $τ$ that gets us there.

PICTURE: Left grain's pile-up finally lights up a new dislocation (amber flash) in the right grain the instant the tip hits $τ_{c}$ .

Figure s05 — Alt-text: a queue of amber dots on the left presses a cyan boundary; to the right of the boundary a bright amber star marks a newly-nucleated dislocation in grain B, labelled "when τ_tip = τ_c". Grains A and B are named. The figure shows the moment yield begins in the next grain.

Step 6 — From shear to pull: assembling the full Hall–Petch law

WHAT: We translate the microscopic shear result into the everyday tensile yield strength $σ_{y}$ you measure by pulling a bar in a testing machine — and we do the $τ \to σ$ conversion out loud instead of hiding it, then add in the baseline term to reach the complete equation from the intro.

Formula The complete Hall–Petch law (floor + boundary bonus)

The total tensile yield strength is the floor plus the boundary bonus: $σ_{y} = σ_{0} + σ_{boundary} = σ_{0} + k d^{- 1/2}, k = \frac{1}{m} \frac{2 G b τ _{c}}{π}$ $σ_{y} = σ_{0} + k d^{- 1/2}$ Term by term:

$σ_{y}$ — the measured tensile yield strength (MPa): the stress at which the pulled bar permanently bends. Same $σ_{y}$ named in the intro — we have now built it.
$σ_{0}$ — the friction stress: the strength floor with no grain-boundary help; the value at $d \to \infty$ .
$k$ — the Hall–Petch coefficient: now fully unpacked as the Schmid-factor conversion $1/ m$ times the pile-up bundle $2 G b τ_{c} / π$ . A material constant; big $k$ = grains matter a lot.
$d^{- 1/2}$ — the star: strength climbs as grains shrink, but with diminishing returns because it's a square root, not a straight line.

PICTURE: Plot $σ_{y}$ against $d^{- 1/2}$ — a straight line of slope $k$ crossing the axis at $σ_{0}$ . Metallurgists use exactly this straight line to measure $k$ and $σ_{0}$ .

Figure s06 — Alt-text: a cyan straight line rising left-to-right on axes of σ_y (MPa) versus d^(−1/2). It crosses the vertical axis at an amber dashed line marked "σ₀ = 100 MPa (floor)"; its slope is labelled "k". Two amber points mark a coarse grain (150 MPa) and a fine grain (600 MPa). The figure shows Hall–Petch as a straight line.

Step 7 — The edge cases: where the law lives and where it dies

Intuition Case B — Grains extremely tiny (nanograins,

$d \to 0$ ): the inverse Hall–Petch effect The formula screams " $σ_{y} \to \infty$ " — but reality doesn't, and this is the most important limit on the law's domain of validity. Below roughly $d \approx 10$ – $20$ nm the pile-up picture breaks: a grain becomes too small to hold even two dislocations, so there is no queue to amplify anything (Step 3's $n$ drops below 1). With no pile-up, the entire derivation from Step 3 onward is void. Deformation then switches to grain-boundary sliding, and measured strength actually falls as grains shrink further. This regime — where the slope of the $σ_{y}$ vs $d^{- 1/2}$ line goes negative — is called the inverse Hall–Petch effect. So the classic $d^{- 1/2}$ law is valid only for grain sizes from roughly tens of nanometres up to the coarse-grain limit; outside that band it does not apply.

PICTURE: The Hall–Petch straight line climbs, then bends over and drops at the far right (very small $d$ , i.e. very large $d^{- 1/2}$ ) — the amber "breakdown" zone.

Figure s07 — Alt-text: on axes of σ_y versus d^(−1/2), a cyan curve rises as a straight line (dotted continuation shown) then bends over and falls inside a shaded amber "breakdown (no pile-up)" band at the far right. An amber arrow labels the drop "inverse Hall–Petch: strength falls". The figure shows where the law stops being valid.

The one-picture summary

Figure s08 — Alt-text: a flow diagram in blueprint style. Top row of boxes: "line slides" → "hits wall" → "pile-up n" → "tip stress n×τ". A curved arrow drops to a bottom row: "set = τ_c" → "d×τ² algebra" → "sqrt appears" → "σ_y = σ₀ + k d^(−1/2)". An amber caption at the base reads "valid only while a pile-up can form (fails for nanograins)". The figure compresses the whole derivation into one chain.

The whole chain in one frame: a single sliding line → a wall → a traffic-jam pile-up → stress amplified $n$ -fold at the tip → set that equal to the fixed break-through stress → the $τ^{2}$ -vs- $d$ algebra forces a square root → convert shear to pull with the Schmid factor → add the friction floor $σ_{0}$ → $σ_{y} = σ_{0} + k d^{- 1/2}$ → valid only while grains are big enough to hold a pile-up.

Recall Feynman retelling (plain words, no symbols)

Metal bends because tiny flaw-lines slide through its crystal grid. Grain walls are barriers those lines can't cross, so the lines stack up behind a wall like cars behind a red light. The longer the queue, the harder the front car crushes the wall — a big grain gives a long queue and a big crush. The wall breaks (and the metal finally yields) when that crush hits a fixed limit. Because the crushing force grows with the square of how hard you push, undoing that square to find the push you need brings in a square root — and the queue length is set by the grain size — so the strength bonus goes like one-over-the-square-root-of-grain-size. One extra bookkeeping step: pulling a bar and sliding a plane aren't aligned, so only a fraction of your pull (the Schmid factor) does the sliding — dividing by that fraction just rescales the constant. And even with no walls at all there's a baseline stickiness — the friction floor — that you add underneath. Smaller grains, shorter queues, less amplification per push, so you must push harder to break through: more strength on top of that floor. But squeeze the grains too small and you can't even fit two cars in a queue — no jam, no amplification, and the whole trick quits, so strength actually starts to drop. That is the inverse Hall–Petch effect, and it is why nanograins stop getting stronger, and why red-hot turbine blades throw the rule out entirely and become one giant grain.

Recall Quick self-test

Why $d^{- 1/2}$ and not $d^{- 1}$ ? ::: Pile-up count $n \propto d$ makes tip stress $\propto d τ^{2}$ ; setting that to a constant and solving for $τ$ divides by $d$ and takes a square root, giving $τ \propto d^{- 1/2}$ . What does $d$ measure — the grain or the boundary? ::: The grain diameter (the crystal between two boundaries); the boundary itself is only a few atoms thick. What is $σ_{0}$ physically? ::: The friction stress — the strength floor with no grain-boundary help (lattice/Peierls friction + solute + forest dislocations); the value as $d \to \infty$ . What is the full assembled law and what is $k$ ? ::: $σ_{y} = σ_{0} + k d^{- 1/2}$ with $k = \frac{1}{m} \frac{2 G b τ _{c}}{π}$ (Schmid conversion $1/ m$ times the pile-up bundle). What does the Schmid factor $m$ do in $k$ ? ::: It converts shear $τ$ to tensile pull $σ = τ / m$ ; only the fraction $m = cos ϕ cos λ$ of your pull resolves onto the slip plane, so $k$ carries a $1/ m$ . Where does Poisson's ratio $ν$ enter? ::: In the repulsion law between edge dislocations, $\frac{G b ^{2}}{2 π ( 1 - ν ) x}$ ; the screw case has no $ν$ . It nudges the numerical prefactor, not the $d^{- 1/2}$ shape. When does the law fail? ::: For nanograins ( $d ≲ 10$ – $20$ nm) grains are too small to hold a pile-up; strength falls — the inverse Hall–Petch effect. With $σ_{0} = 100$ , $k = 0.5 MPa\cdotm^{1/2}$ , $d = 1 μ$ m, what is $σ_{y}$ ? ::: $100 + 0.5 \times 1000 = 600$ MPa.