Order you can count on
Copper can be drawn into wire thinner than a hair at room temperature; magnesium, almost as soft, cracks if you try the same thing cold. A steel blade can be hardened by heating it red hot and quenching it, but a copper one cannot. A silicon wafer is cut so that its surface lies along one particular plane of atoms, and a chip designer cares which one. Tin left for months in severe cold can slowly crumble into grey powder; old stories blame this "tin pest" for ruined organ pipes in unheated churches.
Each of these facts comes down to how atoms are arranged in a crystal. In Bonding in solids you saw how two atoms hold each other; this lesson is about the patterns that billions of them form. Copper and magnesium are both metals with non-directional bonds, but they stack their atoms differently, and the difference decides how many ways their crystals can slip. Steel can be hardened because iron changes its crystal structure when heated. Tin crumbles because it has two crystal structures with different densities.
The language of this lesson (lattices, unit cells, Miller indices) is the everyday language of materials science. Every later lesson uses it, starting with X-ray diffraction, which is how crystal structures are measured.
A lattice plus a basis
Look at a tiled floor or a patterned wallpaper. You can describe it with two pieces of information: a motif (one tile, one flower) and the rule for repeating it (shift by this much to the right, that much upwards). A crystal is the same in three dimensions. The rule is a lattice, an infinite array of points, each with identical surroundings. The motif is the basis, the group of atoms placed at every lattice point.
In copper the basis is a single atom. In sodium chloride it is a pair, one Na⁺ and one Cl⁻. In silicon it is two silicon atoms a quarter of a cube diagonal apart. In a protein crystal the basis may contain tens of thousands of atoms. The lattice says only how the basis repeats, never what it is made of.
Analogy, not a formal construction
A lattice is like the grid of seat positions in a theatre; the basis is who sits in each seat. A family of four can occupy each "seat" as easily as one person can. The analogy stops at the seats: a lattice point is a mathematical point, not a place where an atom must be. The origin can be put anywhere, as long as every lattice point then has the same surroundings.
Unit cells and the fourteen Bravais lattices
Definition— lattice, lattice vectors
A three-dimensional lattice is the set of points for all integers , where are three vectors not in one plane, the lattice vectors.
Definition— unit cell, lattice parameters
A unit cell is a region of space which, repeated by the lattice translations, fills space without gaps or overlaps. It is described by six lattice parameters: the edge lengths and the angles between them. A primitive cell contains exactly one lattice point. A conventional cell may contain more, but is chosen to show the symmetry; the cubic cell of a face-centred cubic lattice holds four lattice points.
Definition— coordination number, packing fraction
The coordination number (CN) is the number of nearest neighbours of an atom. The atomic packing fraction (APF) of a structure of identical atoms is the fraction of space filled when each atom is modelled as a hard sphere of radius just touching its nearest neighbours: , with atoms in a cell of volume .
Requiring that a lattice keeps certain rotation symmetries forces relations among its parameters, and there are exactly seven crystal systems. Within some systems, extra lattice points can be added at the centre of the cell (I, from the German innenzentriert), at the centres of all faces (F) or of one pair of faces (C), without breaking the symmetry and without the result being a lattice already counted. Auguste Bravais showed in 1848 that this gives fourteen distinct lattices.
| Crystal system | Constraints on the cell | Bravais lattices | Examples |
|---|---|---|---|
| Cubic | , all angles 90° | P, I, F | polonium (P), iron (I), copper and NaCl (F) |
| Tetragonal | , all angles 90° | P, I | rutile TiO₂ (P), white tin (I) |
| Orthorhombic | , all angles 90° | P, C, I, F | olivine (P), α-uranium (C) |
| Hexagonal | , | P | magnesium, zinc, graphite |
| Trigonal (rhombohedral) | , | R | corundum α-Al₂O₃, calcite, bismuth |
| Monoclinic | , one angle not 90° | P, C | zirconia at room temperature, gypsum |
| Triclinic | no constraints | P | kyanite, microcline feldspar |
"P" means primitive. The count is 3 + 2 + 4 + 1 + 1 + 2 + 1 = 14. Combining the fourteen lattices with all the ways of placing a symmetric basis gives 230 space groups, enumerated independently by Evgraf Fedorov and Arthur Schoenflies in 1891. Every periodic crystal belongs to one of them. (Quasicrystals, discovered in 1982, are perfectly ordered but not periodic, and need a different description; see X-ray diffraction.)
Simple cubic, BCC, FCC and HCP
Most metallic elements crystallise in one of three structures, because non-directional metallic bonding rewards packing atoms tightly.
- Simple cubic (SC): one atom at each corner of a cube. Each corner atom is shared by eight cells, so the cell holds atom. Its packing is poor, and only one element, the α form of polonium, adopts it.
- Body-centred cubic (BCC): an extra atom at the centre of the cube, 2 atoms per cell. Atoms touch along the body diagonal, so . Iron at room temperature, chromium, tungsten, molybdenum, vanadium and the alkali metals are BCC.
- Face-centred cubic (FCC): an extra atom at the centre of each face, each shared by two cells: atoms per cell. Atoms touch along a face diagonal, . Copper, aluminium, nickel, silver, gold, lead and austenitic stainless steel are FCC.
- Hexagonal close-packed (HCP): a hexagonal lattice with a two-atom basis. Magnesium, titanium at room temperature, zinc, cobalt and zirconium are HCP.
FCC and HCP are the two ways of stacking layers of touching spheres as tightly as possible. In one close-packed layer (A), each sphere has six neighbours, and the hollows between them come in two sets, B and C. The next layer sits in one set of hollows, say B. The third layer can sit either directly above the first (giving the sequence ABAB…, which is HCP) or above the C hollows (ABCABC…, which is FCC).
The program builds each structure as a block of atoms, counts neighbours by brute force and computes packing fractions, then compares them with the exact values.
import numpy as np
def atoms_near(cell, basis, n=3):
"""Cartesian positions of every atom in a (2n+1)^3 block of unit cells."""
rng = np.arange(-n, n + 1)
shifts = np.array(np.meshgrid(rng, rng, rng, indexing="ij")).reshape(3, -1).T
frac = (shifts[:, None, :] + np.array(basis, float)[None, :, :]).reshape(-1, 3)
return frac @ np.array(cell, float)
def describe(cell, basis):
pos = atoms_near(cell, basis)
d = np.linalg.norm(pos, axis=1) # distances from the atom at the origin
d = np.sort(d[d > 1e-9])
shells = [] # (distance, number of atoms) for the first two shells
for r in d:
if shells and abs(r - shells[-1][0]) < 1e-6:
shells[-1][1] += 1
elif len(shells) < 2:
shells.append([r, 1])
d_nn = shells[0][0]
volume = abs(np.linalg.det(np.array(cell, float)))
apf = len(basis) * 4 / 3 * np.pi * (d_nn / 2) ** 3 / volume # spheres of radius d_nn/2
return len(basis), d_nn, shells, apf
cubic = np.eye(3) # lattice parameter a = 1
fcc = [[0, 0, 0], [0.5, 0.5, 0], [0.5, 0, 0.5], [0, 0.5, 0.5]]
c = np.sqrt(8 / 3) # ideal c/a for hexagonal close packing
hexagonal = [[1, 0, 0], [-0.5, np.sqrt(3) / 2, 0], [0, 0, c]]
structures = {
"simple cubic": (cubic, [[0, 0, 0]]),
"BCC": (cubic, [[0, 0, 0], [0.5, 0.5, 0.5]]),
"FCC": (cubic, fcc),
"HCP (ideal c/a)": (hexagonal, [[0, 0, 0], [1 / 3, 2 / 3, 0.5]]),
"diamond cubic": (cubic, fcc + [[x + 0.25, y + 0.25, z + 0.25] for x, y, z in fcc]),
}
print("structure atoms/cell d_nn/a CN second shell packing fraction")
for name, (cell, basis) in structures.items():
n, d_nn, shells, apf = describe(cell, basis)
(r1, n1), (r2, n2) = shells
print(f"{name:16} {n:6} {d_nn:6.4f} {n1:3} {n2:3} at {r2 / r1:5.3f} d_nn {apf:.4f}")
print("\nexact: SC pi/6 =", round(np.pi / 6, 4), " BCC pi*sqrt(3)/8 =", round(np.pi * np.sqrt(3) / 8, 4),
" FCC and HCP pi/sqrt(18) =", round(np.pi / np.sqrt(18), 4), " diamond pi*sqrt(3)/16 =", round(np.pi * np.sqrt(3) / 16, 4))FCC and HCP both reach with twelve neighbours; BCC fills 68 per cent with eight. But BCC's second shell, six atoms only 15 per cent further away, makes it nearly a 14-coordinated structure, which is one reason it competes closely with the close-packed ones. The diamond structure, an FCC lattice with a two-atom basis, fills only 34 per cent of space: covalent bonds insist on four neighbours at tetrahedral angles, whatever the waste of space.
Example— The packing fraction of BCC by hand
A BCC cell of edge holds 2 atoms. Atoms touch along the body diagonal, of length , which spans one radius, one diameter and one radius: , so . Then
Density from the lattice parameter
If you know the structure and the lattice parameter, you know the density: count the mass in one cell and divide by its volume,
with atoms of molar mass in the cell and Avogadro's constant. Lattice parameters are measured by X-ray diffraction to five or six significant figures, so this is one of the most precise ways to know a density.
import numpy as np
NA = 6.02214076e23 # per mole
atoms_per_cell = {"BCC": 2, "FCC": 4, "diamond": 8}
def density_cubic(structure, a_nm, molar_mass):
"""Density in g/cm^3 of a cubic crystal with lattice parameter a (nm)."""
a_cm = a_nm * 1e-7
return atoms_per_cell[structure] * molar_mass / (NA * a_cm**3)
# Room-temperature lattice parameters (nm), molar masses (g/mol), measured densities (g/cm^3), rounded
crystals = [("iron (alpha)", "BCC", 0.28665, 55.845, 7.87), ("copper", "FCC", 0.36149, 63.546, 8.96),
("aluminium", "FCC", 0.40495, 26.982, 2.70), ("silicon", "diamond", 0.54310, 28.086, 2.33),
("carbon (diamond)", "diamond", 0.35670, 12.011, 3.51)]
print("crystal structure a/nm rho computed rho measured")
for name, s, a, M, rho in crystals:
print(f"{name:18} {s:8} {a:.5f} {density_cubic(s, a, M):7.3f} {rho:5.2f}")
# HCP magnesium: 2 atoms in a hexagonal cell of volume (sqrt(3)/2) a^2 c
a, c, M = 0.32094e-7, 0.52107e-7, 24.305
V = np.sqrt(3) / 2 * a**2 * c
print(f"magnesium HCP a = 0.32094, c/a = {c / a:.4f}: rho = {2 * M / (NA * V):.3f} (measured 1.74)")
# Iron changes from BCC to FCC at 912 C. Lattice parameters measured near that temperature (nm):
a_bcc, a_fcc = 0.2904, 0.3647
v_bcc, v_fcc = a_bcc**3 / 2, a_fcc**3 / 4 # volume per atom
print(f"\niron at 912 C: volume per atom {v_bcc:.5f} nm^3 (BCC), {v_fcc:.5f} nm^3 (FCC)")
print(f"change on heating BCC -> FCC: {100 * (v_fcc - v_bcc) / v_bcc:+.2f} % in volume, "
f"{100 * ((v_fcc / v_bcc) ** (1 / 3) - 1):+.2f} % in length")
print(f"nearest-neighbour distance: BCC {a_bcc * np.sqrt(3) / 2:.4f} nm, FCC {a_fcc / np.sqrt(2):.4f} nm")
# Tin: white (beta, tetragonal, 4 atoms per cell, a = 0.5832 nm, c = 0.3182 nm)
# and grey (alpha, diamond cubic, a = 0.6489 nm)
M_Sn = 118.71
rho_white = 4 * M_Sn / (NA * (0.5832e-7) ** 2 * 0.3182e-7)
rho_grey = density_cubic("diamond", 0.6489, M_Sn)
print(f"\ntin: white {rho_white:.3f} g/cm^3, grey {rho_grey:.3f} g/cm^3; "
f"white -> grey increases the volume by {100 * (rho_white / rho_grey - 1):.0f} %")The computed densities of iron, copper, aluminium, silicon and diamond agree with the measured ones to within about 0.3 per cent, and magnesium's ratio, 1.624, is within 1 per cent of the ideal for touching spheres. (Not all HCP metals are so close: zinc and cadmium have near 1.86 and 1.89, titanium and zirconium about 1.59.)
The last lines contain a surprise. When iron is heated through 912 °C and changes from BCC to FCC, the distance between nearest neighbours increases, from 0.2515 to 0.2579 nm, yet the volume per atom decreases by about 1 per cent, because FCC packs its atoms more efficiently. A bar of pure iron heated through 912 °C shrinks while the furnace keeps heating it, and steels show the same contraction over their own transformation range. Measuring this length change (dilatometry) is a standard way to find transformation temperatures. The expansion when austenite transforms back on cooling, larger still when it forms martensite, is part of why quenched steel parts can distort or crack (Phase transformations).
Miller indices: naming planes and directions
Crystals behave differently along different directions: they cleave along some planes, slip along others, and diffract X-rays from all of them at different angles. We need names. The system used everywhere is due to the mineralogist William Hallowes Miller (1839).
Definition— Miller indices of a direction
A direction is written : the components of a lattice vector along that direction, in units of the lattice parameters, reduced to the smallest whole numbers. A negative component is written with a bar, . A family of directions equivalent by symmetry is written ; in a cubic crystal means the six cube-edge directions.
Definition— Miller indices of a plane
To name a plane : find where it cuts the three axes, in units of , , (if the plane passes through the origin, use a parallel plane that does not); take the reciprocals of the intercepts (a plane parallel to an axis has intercept and reciprocal 0); clear fractions to the smallest whole numbers. A family of symmetry-equivalent planes is written .
Example— Naming a plane
A plane cuts the x-axis at , the y-axis at , and never meets the z-axis. The intercepts are ; their reciprocals are ; there are no fractions to clear, so this is the plane. A plane with intercepts has reciprocals ; multiplying by 6 gives . A plane cutting at is .
In a cubic crystal three useful facts follow from the geometry: the direction is perpendicular to the plane ; the spacing between neighbouring planes is ; and the angle between two directions is found from their dot product. The program uses these, and counts how densely atoms are packed on the most important planes and along the most important directions.
import numpy as np
from itertools import product
fcc = np.array([[0, 0, 0], [0.5, 0.5, 0], [0.5, 0, 0.5], [0, 0.5, 0.5]])
bcc = np.array([[0, 0, 0], [0.5, 0.5, 0.5]])
def lattice(basis, n=4):
"""Atom positions of a cubic lattice (a = 1) in a block of (2n+1)^3 cells."""
cells = np.array(list(product(range(-n, n + 1), repeat=3)))
return (cells[:, None, :] + basis[None, :, :]).reshape(-1, 3)
def planar_density(basis, hkl):
"""Atoms per a^2 on the plane (hkl) through the origin: 1 / area of its 2D unit cell."""
pts = lattice(basis)
normal = np.array(hkl, float)
in_plane = pts[np.abs(pts @ normal) < 1e-9]
vecs = in_plane[np.linalg.norm(in_plane, axis=1) > 1e-9]
vecs = vecs[np.argsort(np.linalg.norm(vecs, axis=1))]
v1 = vecs[0] # shortest in-plane vector
for v2 in vecs[1:]: # shortest one not parallel to v1
area = np.linalg.norm(np.cross(v1, v2))
if area > 1e-9:
return 1 / area
def linear_density(basis, uvw):
"""Atoms per a along the direction [uvw]: 1 / shortest lattice vector in that direction."""
pts = lattice(basis)
u = np.array(uvw, float) / np.linalg.norm(uvw)
on_line = pts[(np.linalg.norm(np.cross(pts, u), axis=1) < 1e-9) & (pts @ u > 1e-9)]
return 1 / np.min(on_line @ u)
print("planar density (atoms per a^2) (100) (110) (111)")
for name, b in [("FCC", fcc), ("BCC", bcc)]:
print(f" {name} " + " ".join(f"{planar_density(b, p):.3f}" for p in [(1, 0, 0), (1, 1, 0), (1, 1, 1)]))
print("linear density (atoms per a) [100] [110] [111]")
for name, b in [("FCC", fcc), ("BCC", bcc)]:
print(f" {name} " + " ".join(f"{linear_density(b, d):.3f}" for d in [(1, 0, 0), (1, 1, 0), (1, 1, 1)]))
# Interplanar spacings and angles in a cubic crystal: copper, a = 0.36149 nm
a = 0.36149
print("\ncopper: spacing d_hkl = a / sqrt(h^2 + k^2 + l^2)")
for hkl in [(1, 1, 1), (2, 0, 0), (2, 2, 0), (3, 1, 1)]:
print(f" ({''.join(map(str, hkl))}) d = {a / np.sqrt(sum(i * i for i in hkl)):.4f} nm")
def angle(u, v):
return np.degrees(np.arccos(np.dot(u, v) / np.linalg.norm(u) / np.linalg.norm(v)))
print(f"\nangle between [100] and [111]: {angle((1, 0, 0), (1, 1, 1)):.2f} degrees")
print(f"angle between [110] and [111]: {angle((1, 1, 0), (1, 1, 1)):.2f} degrees")
print(f"angle between the planes (1 1 1) and (1 1 -1): {angle((1, 1, 1), (1, 1, -1)):.2f} degrees")The densest planes and directions are where crystals slip. In FCC the densest plane is , with 2.309 atoms per , and the densest direction ; there are four planes, each with three directions in it, giving the twelve slip systems that make copper, aluminium and gold so ductile. In BCC the densest plane is and the densest direction , the body diagonal along which BCC atoms touch. HCP metals have fewer easy slip systems. Magnesium at room temperature slips easily only on its basal plane, which provides just two independent ways to slip, far fewer than the five that Richard von Mises showed a grain in a polycrystal needs to take up an arbitrary change of shape; this is the root of magnesium's poor cold formability. Why slip happens on dense planes, and how, is the subject of Defects and dislocations.
Note also the spacings in copper: the (200) planes are apart. In FCC, atoms sit on planes parallel to the cube faces at every half lattice parameter, not just at the cube faces themselves. Such "unreduced" indices like (200) and (220) are exactly the ones that appear in diffraction, as X-ray diffraction will explain.
Note
The rule that is perpendicular to is a cubic luxury. In a tetragonal or hexagonal crystal it fails in general, and the spacing formula changes too. Hexagonal crystals are usually described with four Miller–Bravais indices , with , so that symmetry-equivalent planes get similar-looking names.
Polymorphism: same atoms, different crystals
Many elements and compounds can crystallise in more than one structure, each stable in a range of temperature and pressure. These forms are called polymorphs, or allotropes for an element.
Iron is the most important example in engineering. At atmospheric pressure it is BCC (α-iron, ferrite) up to 912 °C, FCC (γ-iron, austenite) from 912 to 1394 °C, and BCC again (δ-iron) from 1394 °C to its melting point at 1538 °C. Under high pressure, above about 13 GPa, it becomes HCP (ε-iron), which is believed to be the main form in the Earth's inner core. The α–γ change is what makes steel so versatile: FCC austenite can dissolve far more carbon than BCC ferrite, because its largest interstitial holes are bigger, so heating steel into the austenite range dissolves carbon that is then trapped or precipitated on cooling. That is the basis of heat treatment, in Phase diagrams and Processing and heat treatment.
Carbon shows how much the arrangement matters. In diamond each atom has four covalent neighbours in a tetrahedral network: the hardest natural material, an electrical insulator, transparent. In graphite atoms form flat hexagonal sheets with three neighbours each, stacked with weak van der Waals bonds between them: soft, slippery, black, and a good conductor along the sheets. Graphite is the stable form at ordinary pressure; diamond is metastable, prevented from turning into graphite only by an enormous energy barrier. Since the 1980s further forms have been found: fullerenes such as the football-shaped C₆₀ (Harold Kroto, Robert Curl and Richard Smalley, 1985), carbon nanotubes and single-layer graphene, all met again in Glasses and nanomaterials.
Tin is the cautionary tale. Metallic white tin (β-Sn, tetragonal) is stable above 13.2 °C; below it, grey tin (α-Sn, with the diamond structure) is stable, and, as the density program shows, takes up about a quarter more volume. The transformation is very slow and needs a nucleus to start, but once it starts, the swelling crumbles the metal into powder, hence the name tin pest. It is a real concern for the tin-rich lead-free solders used in electronics, and small additions of elements such as bismuth or antimony are known to suppress it.
Other polymorphs worth knowing: silica (quartz, cristobalite, tridymite, and the high-pressure coesite and stishovite), titanium dioxide (rutile, anatase, brookite), zirconia (monoclinic, tetragonal, cubic, which is why zirconia ceramics are "stabilised" with yttria; see Ceramics and glasses), and ice, which has well over a dozen known crystalline phases at high pressures. In pharmaceuticals, different polymorphs of the same drug can dissolve at different rates, so manufacturers must control which one they make.
From snowflakes to space groups
In 1611 Johannes Kepler wrote a small essay on why snowflakes have six corners, and in it guessed that stacking cannonballs in the FCC arrangement fills space as densely as possible. In 1669 Nicolas Steno noticed that the angles between corresponding faces of quartz crystals are always the same, however different the crystals look. René-Just Haüy, in the 1780s, explained this by imagining crystals built from tiny identical blocks. Bravais's fourteen lattices (1848) and the 230 space groups of Fedorov and Schoenflies (1891) completed the geometry, two decades before anyone could see inside a crystal; X-ray diffraction, from 1912, confirmed it.
Kepler's guess about sphere packing was proved only in 1998 by Thomas Hales, with a proof so heavily dependent on computer calculations that referees could not fully check it. Hales and collaborators then produced a complete computer-verified formal proof, the Flyspeck project, published in 2017. The packing fraction computed above is the densest possible for equal spheres.
The professional view
In research and industry a crystal structure is a file, not a drawing. The Crystallographic Information File (CIF), maintained by the International Union of Crystallography, records the space group, the lattice parameters and the positions of the atoms in the asymmetric unit (the smallest part from which symmetry generates the rest). Structures are archived in databases: the open Crystallography Open Database, the Inorganic Crystal Structure Database (ICSD, from FIZ Karlsruhe), the Cambridge Structural Database for organic and metal–organic crystals, and computed structures in the Materials Project and similar projects. The symmetry operations of every space group are tabulated in the International Tables for Crystallography.
Structures are handled in code with libraries such as pymatgen or the Atomic Simulation Environment (ASE), and drawn with programs such as VESTA. For example, building copper from its space group and asking for its density (not run here; it needs pymatgen installed):
# Not run here: needs pymatgen (pip install pymatgen)
from pymatgen.core import Lattice, Structure
cu = Structure.from_spacegroup("Fm-3m", Lattice.cubic(3.6149), ["Cu"], [[0, 0, 0]])
print(len(cu), "atoms in the conventional cell")
print(cu.density, "g/cm^3")
print(cu.get_space_group_info()) # the symbol and number of the space groupOne site, (0, 0, 0), plus the symmetry operations of space group (number 225) generates the four atoms of the FCC cell. Professionals rarely type coordinates by hand: they download a CIF, check its source and temperature, and let the software apply the symmetry.
Common misconceptions
Misconception
"The unit cell is the crystal's building block, and there is only one." Any cell that tiles space under the lattice translations will do; there are infinitely many, primitive and not. The conventional cell is a choice made to show symmetry. FCC copper's primitive cell is a rhombohedron with one atom, a quarter of the volume of the cubic cell.
Misconception
"Hexagonal close-packed is one of the Bravais lattices." HCP is a hexagonal lattice with a two-atom basis; the atom in the middle layer does not have the same surroundings as the corner atoms (its neighbours are rotated), so it is not a lattice point. Likewise diamond is an FCC lattice with a two-atom basis, and sodium chloride is an FCC lattice with a Na–Cl basis.
Misconception
"BCC is close-packed." It is not: it fills 68 per cent of space, against 74 per cent for FCC and HCP, and it has no close-packed planes. Its densest planes, , are flat, but their atoms form a slightly distorted triangular pattern rather than the perfect triangular layers of FCC and HCP.
Misconception
"Atoms in a metal are hard spheres touching their neighbours." The hard-sphere model is a geometric convenience. Atoms have no sharp edges; the "radius" is half the nearest-neighbour distance and changes with structure, temperature and pressure, as the iron numbers above show.
Misconception
"Diamond is the most stable form of carbon." At ordinary temperature and pressure graphite is (slightly) more stable. Diamond survives because converting it requires breaking strong covalent bonds, a barrier so high that the change is unmeasurably slow at room temperature.
Exercises
1. Crystal structures
Warm-upHumanWhich of these arrangements of points are Bravais lattices (every point having identical surroundings), rather than a lattice plus a basis of more than one atom? Select all that apply.
2. Crystal structures
CoreHumanIn a BCC crystal with lattice parameter , how many atoms lie within times the nearest-neighbour distance of a given atom (not counting the atom itself)?
3. Crystal structures
Warm-upHumanSilicon has the diamond structure: an FCC lattice with a basis of two atoms, at and . How many atoms are in its conventional cubic unit cell?
4. Crystal structures
CoreHumanA plane in a cubic crystal cuts the x-axis at and the y-axis at , and is parallel to the z-axis. What are its Miller indices? Write them in brackets, e.g. (111).
5. Crystal structures
CoreHumanWhen iron is heated through 912 °C it changes from BCC to FCC. The nearest-neighbour distance increases (from about 0.2515 to 0.2579 nm), but the volume per atom decreases by about 1 per cent. Why?
6. Crystal structures
CoreComputationalWrite density(structure, a_nm, molar_mass) that returns the density in g/cm³ of a cubic crystal of one element, where structure is "sc", "bcc", "fcc" or "diamond", a_nm is the lattice parameter in nanometres and molar_mass is in g/mol. Use mol⁻¹. For BCC iron ( = 0.28665 nm, = 55.845 g/mol) you should get about 7.87 g/cm³.
7. Crystal structures
StretchComputationalWrite miller_indices(intercepts) that takes the intercepts of a plane on the x, y and z axes, in units of the lattice parameters, and returns its Miller indices as a tuple of the smallest whole numbers. An intercept of None means the plane is parallel to that axis. Negative intercepts give negative indices. Intercepts may be ints, floats such as 0.5, or fractions.Fractions.
Examples: (1, None, None) → (1, 0, 0); (0.5, 1, None) → (2, 1, 0); (2, 3, None) → (3, 2, 0); (2, 2, 2) → (1, 1, 1).
Further reading and resources
- International Tables for Crystallography · International Union of CrystallographyThe reference work for lattices, space groups and their symmetry operations.
- Crystallography Open DatabaseAn open collection of several hundred thousand crystal structures, downloadable as CIF files.
- The Materials ProjectComputed structures and properties; compare its lattice parameters with the measured ones used here.
- A proof of the Kepler conjecture · Thomas C. Hales · 2005The proof that no packing of equal spheres is denser than face-centred cubic or hexagonal close packing.
- A formal proof of the Kepler conjecture · Thomas Hales and 21 co-authors · 2017The Flyspeck project's computer-checked version of the proof.
- pymatgenAn open-source Python library for crystal structures, used by the Materials Project.
- VESTA · Koichi Momma and Fujio IzumiFree software for drawing crystal structures in three dimensions.
- DoITPoMS teaching and learning packages · University of CambridgeFree interactive packages, including ones on crystallography and lattice planes.