All lessons

Module 9: Condensed matter physics · Lesson 9.2 · about 45 min

Defects and dislocations

Point defects in thermal equilibrium (vacancies, interstitials, Schottky and Frenkel pairs) from the free energy, their effect on density and diffusion; edge and screw dislocations, the Burgers vector and circuit, glide and climb, why real crystals are a thousand times weaker than the ideal shear strength, stress fields and line energy, Frank-Read multiplication and work hardening.

Key points, meanings and formulas on one page: choose "Save as PDF" in the print box.

Module 9 · about 45 minutes. A perfect crystal is a useful idealisation, but no real crystal is perfect. A few missing atoms change how a solid diffuses, conducts and colours, and a few line defects decide whether a metal bends or snaps. NET Part C asks both: an Arrhenius count of point defects, and the geometry of a dislocation.

Point defects are present in thermal equilibrium: making one costs energy but gains entropy, so at any T > 0 some always exist, in a fraction about exp(−E/kT). Dislocations are not equilibrium defects (their energy per length is far too large), but they are always there, and they are why crystals yield at small stresses: slip spreads one atomic row at a time.

1. Point defects in equilibrium

Take a crystal of N sites and remove n atoms to the surface. Each vacancy costs Ev; the n vacancies can sit in N!/[n!(N−n)!] ways.

Free energy. F = nEv − kT ln[N!/(n!(N − n)!)].

Stirling. ln N! ≈ N ln N − N, so ∂F/∂n = Ev − kT ln[(N − n)/n].

Minimise. Setting ∂F/∂n = 0 gives n/(N − n) = exp(−Ev/kT), and with n ≪ N, n ≈ N exp(−Ev/kT).

Add vibrational entropy. The atoms around a vacancy vibrate more softly, adding Sv per defect: n/N = exp(Sv/k) exp(−Ev/kT), with exp(Sv/k) typically 2 to 10.

nN=e−Ev/kBT,nSN=e−ES/2kBT,nF=NN′e−EF/2kBT
Vacancy (Schottky defect in a metal)
an atom moved from inside to the surface; `$E_v \approx 1$` eV in Cu, Al, Au
Schottky pair `$n_S$`
in an ionic crystal, one cation and one anion vacancy together, so the crystal stays neutral; the factor 2 in the exponent comes from `$n_+ n_- = N^2 e^{-E_S/kT}$` with `$n_+ = n_-$`
Frenkel pair `$n_F$`
an ion moved from its site to one of `$N'$` interstitial sites; common in AgBr and AgCl, where the small Ag⁺ fits between the anions
Density test
Schottky defects lower the density (lattice sites stay, atoms go to the surface); Frenkel defects do not change it

Worked example. Copper has Ev=1.28 eV. At 1300 K, kBT=0.112 eV and Ev/kBT=11.4, so n/N=1.1×10−5: about 11 vacancies per million sites. At 300 K the same formula gives 3×10−22, effectively none. A crystal quenched from near its melting point keeps its high-temperature vacancies for a while, which is why quenched metals are harder and anneal on heating.

The figure plots log10⁡(n/N) against 1000/T. A straight line means Arrhenius behaviour; its slope is −E/(2.303kB) for vacancies and half that for pairs. This is exactly how formation energies are read from experiments.

Arrhenius plot of the equilibrium defect fraction. Change the formation energy and entropy, and switch between single vacancies and Schottky pairs (half the slope). The marker follows the chosen temperature.

Diffusion and ionic conduction

Atoms move by jumping into a neighbouring vacancy. The jump rate needs a vacancy next door (fraction e−Ev/kT) and a hop over a barrier Em, so

D=D0e−(Ev+Em)/kBT,σion=nq2DkBT

The second form is the Nernst-Einstein relation: in an ionic crystal, vacancies carry charge, so the conductivity also rises in Arrhenius fashion. Doping NaCl with CaCl₂ adds one cation vacancy per Ca²⁺ (to keep neutrality), so at low T the vacancy count is fixed by the dopant and only Em shows in the slope: the two-slope plot NET likes to show. Colour centres (F centres) are anion vacancies that trap an electron; their absorption band makes alkali halides coloured after irradiation.

2. Why crystals are weak: the dislocation

Slide the top half of a crystal over the bottom half by one spacing b, all at once. The shear stress rises like τ=(Gb/2πa)sin⁡(2πx/b), so the ideal shear strength is

τideal≈G2π∼a few GPa

Measured yield stresses of pure single crystals are near 1 MPa, about 10−3 to 10−4 of this. Taylor, Orowan and Polanyi (1934) explained the gap: slip spreads by moving a line defect, a dislocation, across the slip plane, breaking only the bonds along the line at each step. It is how you move a heavy carpet by pushing a ruck along it.

Edge dislocation. An extra half-plane of atoms ends inside the crystal. The line is the bottom edge of the half-plane. Above it the lattice is squeezed, below it stretched.

Burgers circuit. Walk a closed loop of lattice steps around the line (say 5 up, 5 right, 5 down, 5 left). In a perfect crystal it closes. Around the dislocation it fails to close by one lattice vector: the Burgers vector b.

Edge versus screw. Edge: b ⊥ line. Screw: b ∥ line, and the planes around the line form a helical ramp. A general (mixed) dislocation curves, and b stays the same along it while the line direction changes.

Conservation. b is the same all along a dislocation; a dislocation cannot end inside a crystal (only at a surface, a grain boundary, or a node where the Burgers vectors add to zero).

An edge dislocation in a simple square lattice (displacements from the elastic solution u = (b/2π) arctan(y/x)). Press play to drive it by the shear stress: the line glides one row at a time and leaves a step of size b at the surface. Switch on the Burgers circuit to see the closure failure.

Glide and climb

The slip plane contains both b→ and the line. Motion in that plane, glide, needs only small shifts of atoms and happens at low temperature under shear. An edge dislocation moving out of its slip plane, climb, needs atoms to be added to or removed from the half-plane, so it requires vacancy diffusion and is important only at high temperature (creep). A screw dislocation has b→ along the line, so any plane containing the line is a slip plane: it can cross-slip.

When a dislocation with Burgers vector b sweeps fully across a crystal, the top half shifts by b relative to the bottom. Many dislocations of density ρ (line length per volume, m⁻²) moving at speed v give the Orowan strain rate

γ˙=ρbv

Stress, energy and force

Around a screw dislocation the elastic shear strain is b/2πr, so the stress is τ=Gb/2πr. Integrating the strain energy density τ2/2G from a core radius r0 to an outer cut-off R gives the energy per unit length:

EL=Gb24πln⁡Rr0≈12Gb2(edge: divide by 1−ν)
`$E/L \propto b^2$`
Frank's rule: the dislocation with the smallest `$b$` is favoured; `$\vec b_1 \to \vec b_2 + \vec b_3$` goes ahead if `$b_1^2 > b_2^2 + b_3^2$`. In fcc, `$\frac a2[110]$` splits into two Shockley partials `$\frac a6[211] + \frac a6[12\bar1]$` with a stacking fault between them
Size
with `$G = 40$` GPa, `$b = 0.25$` nm: `$Gb^2/2 \approx 1.3$` nJ/m, about 2 eV for every atomic spacing of line (a few eV once the logarithm is kept). That is why dislocations are never in thermal equilibrium: the entropy cannot pay for it
Peach-Koehler force
a resolved shear stress `$\tau$` pushes on the line with force per length `$F/L = \tau b$`
Line tension
`$T \approx Gb^2/2$`; a segment pinned at both ends bows out to radius `$R = T/\tau b \approx Gb/2\tau$`

Worked example: Frank-Read source. A segment of length L=1 μm is pinned at both ends in a crystal with G=45 GPa, b=0.256 nm. It bows into a semicircle (the hardest shape to pass) when R=L/2, which needs τ=Gb/L=45×109×0.256×10−9/10−6=11.5 MPa. Beyond that the loop grows, wraps around, pinches off a closed loop and restores the segment, ready to emit another. One source can produce hundreds of loops: this is how the dislocation density rises from about 1010 m⁻² in an annealed metal to 1015 to 1016 m⁻² after heavy cold work.

Work hardening

Dislocations on crossing planes get in each other’s way. The mean spacing between them is ρ−1/2, and freeing a segment of that length needs τ∼Gb/ℓ, which is Taylor’s law:

τ=αGbρ,α≈0.2 to 0.5

Grain boundaries also block slip: in the Hall-Petch law, yield stress = σ0+kyd−1/2 for grain size d. A low-angle tilt boundary is itself a wall of edge dislocations spaced D=b/θ apart. With b=0.25 nm and a 1° tilt, D=14 nm.

“Dislocations are frozen-in thermal defects like vacancies, and the Burgers vector changes along a curved line.” Vacancies have a Boltzmann population because their entropy of placement is enormous compared with Ev/kT; a dislocation line costs a few eV per atomic length and gains almost no entropy, so its equilibrium density is zero. It exists because of growth and deformation history. And b→ is fixed along the whole line: what changes along a curved loop is the line direction, so the character changes from edge to screw and back.

  • Vacancies: n/N=eSv/ke−Ev/kT; Schottky pairs in ionic crystals e−ES/2kT; Frenkel pairs NN′e−EF/2kT. Schottky lowers density, Frenkel does not.
  • Diffusion D=D0e−(Ev+Em)/kT; ionic conductivity σ=nq2D/kT; aliovalent dopants fix the vacancy count at low T.
  • Ideal shear strength ≈G/2π; dislocations explain yield at 10−3 to 10−4 of it.
  • Burgers circuit gives b→: edge b→⟂ line, screw b→∥ line; constant along the line; lines end only at surfaces, boundaries or nodes.
  • Glide in the plane of b→ and line; climb needs diffusion; screws cross-slip. Orowan: γ˙=ρbv.
  • E/L≈Gb2/2 (Frank’s b2 rule), force τb per length, Frank-Read stress Gb/L, Taylor τ=αGbρ, tilt boundary D=b/θ.

Practise: exam questions on this lesson

Next: lesson 9.3 turns from mechanical to electrical response: polarisation, the local field and the optical constants of insulators and metals. For the elastic moduli used here, see the JAM lesson on elasticity.

Check yourself

1In an ionic crystal the Schottky pair formation energy is 2.0 eV. At 1000 K the fraction of vacant cation sites is closest to
2For a pure screw dislocation, the Burgers vector is
3Why does a metal deform at shear stresses about 10³ times smaller than the ideal estimate G/2π?

Sources

  • C. Kittel, Introduction to Solid State Physics, 8th ed. (2005), chapters 18 and 20
  • D. Hull and D. J. Bacon, Introduction to Dislocations, 5th ed. (2011), chapters 1 to 4
  • N. W. Ashcroft and N. D. Mermin, Solid State Physics (1976), chapter 30
  • CSIR-UGC NET Physical Sciences syllabus, Part C condensed matter physics, published by the National Testing Agency

Your ticks stay in this browser. Save them as a file on the course home.

Updated 2026-11-22. Not affiliated with CSIR, UGC or the National Testing Agency. Check the official CSIR NET website for the current pattern. All questions are original.