“The reciprocal lattice only dictates where diffraction peaks may appear; whether each peak shines brightly or vanishes altogether is decided by complex phasor interference among the atoms inside a single unit cell.”From the Braggs to Patterson, Hauptman & Rietveld
Body-centered cubic \(\alpha\text{-Fe}\) has real \((100)\) planes spaced by \(a\), yet its \((100)\) XRD peak is identically zero; diamond-cubic \(\mathrm{Si}\) loses not only \((100)\) but also the fcc-allowed \((200)\) reflection. What cancels the scattered waves?
Detectors measure only intensity \(I\propto |F_{hkl}|^2\), losing all phases \(\varphi_{hkl}\) required for inverse Fourier synthesis—how are crystal structures solved? And how do a few \(2\theta\) peaks from an unknown cubic powder reveal its Bravais lattice, \(a\), and grain size \(D\)?
Part 1's Ewald sphere gave the necessary condition for diffraction directions; Part 2 enters the unit cell with the atomic scattering factor \(f(\theta)\) and structure factor \(F_{hkl}\) to establish the sufficient condition (intensities & systematic absences), then turns theory into lab tools via powder indexing and the Scherrer equation.
| Crystal Structure (Atoms / Cell) | Structure Factor \(F_{hkl}\) | Allowed Condition | Extinction / Weak-Reflection Rule | First Allowed \(N=h^2+k^2+l^2\) |
|---|---|---|---|---|
| Simple cubic sc (1) | \(F=f\) | All integers \((hkl)\) | No systematic extinction | \(1_{(100)},\,2_{(110)},\,3_{(111)},\,4_{(200)}\) |
| Body-centered bcc (2) | \(f\bigl[1+(-1)^{h+k+l}\bigr]\) | \(h+k+l=2n\) (even, \(F=2f\)) | \(h+k+l=2n+1\) (I-centering lattice ext.) | \(2_{(110)},\,4_{(200)},\,6_{(211)},\,8_{(220)}\) |
| Face-centered fcc (4) | \(f\bigl[1+(-1)^{h+k}+(-1)^{h+l}+(-1)^{k+l}\bigr]\) | \(h,k,l\) unmixed parity \((F=4f)\) | Mixed parity (F-centering lattice ext.) | \(3_{(111)},\,4_{(200)},\,8_{(220)},\,11_{(311)}\) |
| Diamond Si / C (8, \(Fd\bar{3}m\)) fcc × basis \((0,0,0),(\tfrac{1}{4},\tfrac{1}{4},\tfrac{1}{4})\) | \(F^{\mathrm{fcc}}\bigl[1+\mathrm{e}^{\mathrm{i}\frac{\pi}{2}(h+k+l)}\bigr]\) | ① All odd \((|F|^2=32f^2)\) ② All even & \(h+k+l=4n\) \((|F|^2=64f^2)\) | ① Mixed parity \((210)\): lattice ext. ② All even, \(h+k+l=4n+2\) \((200),(222)\): structure ext. | \(3_{(111)},\,8_{(220)},\,11_{(311)},\,16_{(400)}\) |
| Ordered \(\mathrm{CsCl}\) / \(\mathrm{NaCl}\) | \(\mathrm{CsCl}: f_{\mathrm{A}}+f_{\mathrm{B}}(-1)^{h+k+l}\) \(\mathrm{NaCl}: 4\bigl[f_{\mathrm{A}}+f_{\mathrm{B}}(-1)^{h+k+l}\bigr]\) | Fundamental \((f_{\mathrm{A}}+f_{\mathrm{B}})\): \(h+k+l\) even (\(\mathrm{CsCl}\)); all even (\(\mathrm{NaCl}\)) | Antiphase \(\lvert f_{\mathrm{A}}-f_{\mathrm{B}}\rvert\neq 0\): weak superlattice (\(\mathrm{CsCl}\,100\); \(\mathrm{NaCl}\,111\)); vanishes in isoelectronic \(\mathrm{KCl}\) | \(\mathrm{CsCl}: \mathbf{1^*},2,3^*,4\) \(\mathrm{NaCl}: \mathbf{3^*},4,8,11^*\) |
1 : 2 : 3 : 4 : 5 : 6 : 82 : 4 : 6 : 8 : 10 : 12 : 141 : 2 : 3 : 4 : 5 : 6 : 7 (7th peak (321) breaks tie!)3 : 4 = 1 : 1.3333 : 4 : 8 : 11 : 12 : 163 : 8 = 1 : 2.6673 : 8 : 11 : 16 : 19 : 24\(\alpha\text{-Fe}(100)\) vanishes because the body-center plane at \(a/2\) introduces a \(\lambda/2\) path shift (\(h+k+l=1\) odd, lattice extinction); diamond \(\mathrm{Si}\) additionally loses \((200)\) and \((222)\) because its \((\tfrac{1}{4},\tfrac{1}{4},\tfrac{1}{4})\) sublattice shifts by \(\pi\) against the origin sublattice when \(h+k+l=4n+2\) (structure extinction).
Lost phases are bypassed via Patterson autocorrelation \(P(\boldsymbol{u})\) for heavy-atom vectors or recovered probabilistically via \(\rho\ge 0\) direct methods; unknown cubic powders are indexed from \(\sin^2\theta\) integer ratios (e.g. \(3:4:8:11\Rightarrow\text{fcc}\)), and FWHM \(\beta\) (in radians) gives crystallite size \(D\) via the Scherrer equation.
Next lecture: X-rays struggle with light \(\mathrm{H/Li}\) atoms and cannot image individual dislocations—Part 3 introduces electron & neutron diffraction, operando battery XRD, Abbe imaging, and TEM/SEM characterization.
1. For an elemental body-centered cubic (bcc) metal, which pair among \((210)\), \((211)\), \((300)\), and \((310)\) produces allowed diffraction peaks?
bcc requires \(h+k+l\) even: \((211)\to 4\) and \((310)\to 4\) are allowed, whereas \((210)\to 3\) and \((300)\to 3\) suffer body-centered lattice extinction.
2. In diamond-cubic \(\mathrm{Si}\) (\(Fd\bar{3}m\)), both \((210)\) and \((200)\) reflections are absent. What are their respective extinction mechanisms?
\((210)\) has mixed parity (\(F^{\mathrm{fcc}}=0\), F-centering lattice extinction); \((200)\) is all-even with \(h+k+l=2=4n+2\), canceling the two-atom basis (\(d\)-glide structure extinction).
3. A nanopowder peak at \(2\theta=38^\circ\) has instrument-corrected FWHM \(\beta=0.5^\circ\) \((\lambda=0.154\text{ nm},\,K=0.9)\). Its mean crystallite size \(D\) is closest to:
Convert \(\beta=0.5^\circ=0.00873\text{ rad}\) and use Bragg half-angle \(\theta=19^\circ\) \((\cos 19^\circ\approx 0.9455)\): \(D=\frac{0.9\times 0.154}{0.00873\times 0.9455}\approx 16.8\text{ nm}\).