Carbon Allotropes

Take two pencils' worth of graphite, compress the carbon hard enough, and heat it to the temperature of a blowtorch. Nature does exactly this a hundred miles under your feet. After a billion years, you get a diamond. Same element, same fourteen electrons per atom's worth of nuclear charge. Yet one form is the softest thing you'd want to write with. The other is the hardest natural material known. This split personality is not a chemical trick. It is pure structural bookkeeping. It is decided entirely by how carbon's valence orbitals hybridize. It also depends on how the resulting bonds tile space. Carbon's ability to build wildly different solids from a single element is unmatched in the periodic table. This matters far beyond geology. Diamond anvils generate the pressures used to study planetary interiors. Graphite lubricates machinery and moderates nuclear reactors. Graphene and carbon nanotubes are reshaping electronics, composites, and sensors. Fullerenes carry electrons in organic solar cells and cages for drug delivery. All of this traces back to one structural fact. Carbon can adopt $sp^3$ , $sp^2$ , or curved- $sp^2$ bonding. Each choice locks in a distinct set of mechanical, electrical, and optical properties. The big picture is that allotropy in carbon is a case study in structure-dictates-property reasoning. The same logic is used across p-block periodic trends. It is revisited more deeply in Catenation and Carbon-Carbon Bonding. Diamond's isotropic $sp^3$ network gives three-dimensional rigidity and an insulating band gap. Graphite's layered $sp^2$ sheets give in-plane strength with interlayer slip. They also give metallic-like conduction. Fullerenes curve $sp^2$ sheets into closed cages by inserting exactly twelve pentagons. They obey a topological law (Euler's formula) as strictly as any physical law. Nanotubes are graphene rolled into cylinders. Their electronic character is fixed by a single pair of integers. Graphene itself is the two-dimensional parent sheet. Its Dirac-cone band structure has become a textbook example. It shows relativistic-like physics appearing in a lattice. This article builds all five structures from first principles. These principles include hybridization, bond angles, packing, and band formation. It then asks, quantitatively, why density, hardness, conductivity, and color differ so sharply. These differences occur between forms of the same element. Expect orbital diagrams and lattice geometry. Expect a topological derivation of the fullerene pentagon rule. Expect a band-structure argument for why graphite conducts and diamond does not.

Key takeaways

Diamond: every carbon is $sp^3$ , tetrahedral, bonded to four others in a 3-D network — hardest natural substance, wide band gap (~5.5 eV), electrical insulator, best natural thermal conductor.
Graphite: $sp^2$ carbons form flat hexagonal sheets (bond length 141.5 pm) held together by weak van der Waals/π-stacking (interlayer spacing 335 pm) — soft, slippery, conducts electricity in-plane.
Fullerenes (C₆₀, C₇₀…) are closed cages built from hexagons plus exactly 12 pentagons, a consequence of Euler's polyhedron formula — molecular, soluble, discrete HOMO–LUMO gap.
Carbon nanotubes are rolled-up graphene sheets; the roll-up (chiral) vector $(n,m)$ deterministically fixes whether the tube is metallic or semiconducting.
Graphene is the 2-D single sheet: massless Dirac-fermion-like carriers, record electron mobility, single-atom thickness, zero band gap at the Fermi point (the "Dirac point").
All differences reduce to one variable: the hybridization and dimensionality of the carbon network — 3-D $sp^3$ vs. 2-D/0-D $sp^2$ vs. curved $sp^2$ .
Key idea: carbon's allotropes are a controlled experiment in orbital hybridization — changing $sp^3$ to $sp^2$ (and flat to curved) rewires bonding topology and completely rewrites mechanical and electronic behavior.

Where this lives today

Diamond's mechanical and thermal extremes make it indispensable in cutting and grinding tools, and synthetic diamond (grown by high-pressure-high-temperature or chemical vapor deposition) is now the dominant industrial source, sparing natural stones for gemology. Diamond's wide band gap and radiation hardness make it attractive for high-power/high-temperature electronics and for diamond-based quantum sensors, where nitrogen-vacancy (NV) color-center defects act as atomic-scale magnetometers sensitive enough for single-molecule NMR and even neuroscience imaging. Diamond anvil cells — two gem-quality diamond tips compressing a sample between them — remain the standard tool for reproducing planetary-core pressures (hundreds of gigapascals) in the laboratory, letting chemists and geophysicists study how matter (including, ironically, other allotropes of carbon) behaves under extreme conditions. Graphite's softness and lubricity make it a dry lubricant and a component of pencil "leads" (mixed with clay to tune hardness), while its electrical conductivity and chemical inertness make it the electrode material of choice in lithium-ion battery anodes (lithium ions intercalate between the sheets) and in electrolysis and arc-furnace electrodes. Its neutron-moderating properties (slowing fast neutrons without much absorption) made it central to early nuclear reactor design, including the graphite-moderated reactors of the mid-20th century. Fullerenes are used as electron-accepting components in organic photovoltaic blends (PCBM, a soluble C₆₀ derivative, is a benchmark acceptor material), and endohedral fullerenes — cages with a metal ion trapped inside — are explored as MRI contrast agents and radiopharmaceutical carriers, since the carbon cage shields the body from a reactive or radioactive guest ion while still allowing the whole assembly to be functionalized for targeting. Carbon nanotubes, exploiting the metallic/semiconducting split derived above, are used as conductive additives in battery electrodes and composites, as field-effect transistor channels in nanoelectronics research, and — bundled into macroscopic fibers — as extremely strong, lightweight reinforcing elements, since individual SWCNTs have tensile strengths far exceeding steel on a per-mass basis. Graphene's combination of mechanical strength, electrical conductivity, and optical transparency underlies its use in flexible transparent electrodes for touchscreens and photovoltaics, as a barrier coating impermeable to gases, and as a sensing surface where adsorption of even single molecules measurably perturbs its conductivity — a direct consequence of its Dirac-point band structure making it exceptionally sensitive to any perturbation near the Fermi level. Understanding all these devices at a deeper level connects back to the same p-block periodic trends that make carbon uniquely able to catenate and hybridize flexibly, a theme picked up directly in Catenation and Carbon-Carbon Bonding and thrown into sharp relief by comparison with its heavier congener in Silicon vs Carbon: Group 14 Contrast.

Ask the tutor: How does introducing a single pentagon-heptagon defect pair (a Stone–Wales defect) into an otherwise perfect graphene or nanotube lattice locally change the electronic structure without changing the total pentagon/hexagon topology count?

The physical picture

Carbon has the ground-state configuration $1s^2 2s^2 2p^2$ , but in condensed phases it almost never uses these atomic orbitals directly — it hybridizes them. Mixing one $2s$ and three $2p$ orbitals gives four equivalent $sp^3$ hybrids pointing to the corners of a tetrahedron (109.5°); mixing one $2s$ and two $2p$ orbitals gives three coplanar $sp^2$ hybrids at 120° plus one unhybridized $p$ orbital perpendicular to the plane. Every property discussed below is a direct consequence of which of these two hybridization schemes carbon adopts, and how the resulting units are stitched into an extended solid or a closed molecule. In diamond, all four valence electrons per carbon go into $sp^3$ hybrids, each forming a single, strong $\sigma$ bond (C–C bond length 154 pm, bond enthalpy ≈ 346 kJ/mol) to a neighboring carbon. The result is a fully three-dimensional covalent network — technically a single "supermolecule" — in the cubic diamond structure (space group $Fd\bar3m$ ), which can be viewed as two interpenetrating face-centered cubic lattices offset by $(1/4,1/4,1/4)$ . Because every bond is a strong $\sigma$ bond and every atom is rigidly locked to four neighbors in every direction, diamond resists shear and compression isotropically: it is the hardest known natural material (Mohs 10, Vickers hardness ≈ 100 GPa) and, because all four valence electrons are tied up in localized $\sigma$ bonds with a large HOMO–LUMO (valence-conduction) gap of about 5.5 eV, it is an excellent electrical insulator. Paradoxically it is also the best thermal conductor among natural bulk materials (up to ~2200 W/(m·K)) because the same stiff, low-mass, defect-free lattice that resists deformation also carries lattice vibrations (phonons) extremely efficiently. In graphite, each carbon uses $sp^2$ hybrids to form three in-plane $\sigma$ bonds to neighboring carbons, generating a flat hexagonal (honeycomb) sheet with C–C = 141.5 pm — shorter and stronger than diamond's bond, because of partial double-bond character. The fourth electron sits in an unhybridized $2p_z$ orbital perpendicular to the sheet; these $p_z$ orbitals overlap sideways across the whole sheet to form a delocalized π system, exactly analogous to benzene's aromatic ring but extended over the entire two-dimensional lattice. Stacks of these sheets (the "AB" or Bernal stacking, alternating registry) are held together only by weak van der Waals forces and π–π interactions, with an interlayer spacing of 335 pm — more than double the in-plane bond length. This is why graphite cleaves easily along the layers (pencil "lead," lubricants) while being strong within a layer, and why it conducts electricity well in-plane (delocalized π electrons act like a 2-D electron gas) but poorly perpendicular to the sheets.

Fullerenes take an $sp^2$ sheet and close it into a hollow cage by introducing pentagons among the hexagons — a topological necessity explored rigorously below. C₆₀, "buckminsterfullerene," has 12 pentagonal and 20 hexagonal faces arranged with full icosahedral ( $I_h$ ) symmetry, exactly the pattern of a soccer ball, with two distinct bond lengths: 139 pm at hexagon–hexagon fusions ([6,6] bonds, more double-bond-like) and 145 pm at hexagon–pentagon fusions ([5,6] bonds, more single-bond-like). Because the cage is a discrete, closed molecule rather than an infinite network, fullerenes are molecular solids: they dissolve in nonpolar solvents (a red-brown solution of C₆₀ in toluene is a signature undergraduate-lab observation), sublime, and have discrete HOMO–LUMO gaps (~1.7 eV for C₆₀) rather than continuous bands.

Carbon nanotubes are, geometrically, a single graphene sheet rolled into a seamless cylinder (single-walled, SWCNT) or several concentric cylinders (multi-walled, MWCNT). The roll-up direction is specified by a chiral vector built from the graphene lattice vectors $\mathbf{a}_1, \mathbf{a}_2$ :

\mathbf{C}_h = n\mathbf{a}_1 + m\mathbf{a}_2 \equiv (n,m)

The integers $(n,m)$ completely determine the tube's diameter, chirality, and — remarkably — whether it behaves as a metal or a semiconductor, a purely topological/quantum-mechanical selection rule derived below.

Graphene is simply an isolated single sheet of the graphite lattice: one atom thick, the thinnest possible stable 2-D crystal, first isolated by mechanical exfoliation ("Scotch-tape method") in 2004. Its band structure has conical points (Dirac points) at the corners of the hexagonal Brillouin zone where the conduction and valence bands touch with linear dispersion, making its charge carriers behave as massless Dirac fermions — giving graphene extraordinarily high carrier mobility (>10⁴–10⁵ cm²/(V·s)), remarkable mechanical strength (Young's modulus ≈ 1 TPa), and near-total optical transparency (2.3% absorption per layer, set by the fine-structure constant).

Figure

Hybridization contrast: diamond sp³ tetrahedron vs. graphite sp² sheet with π system

Figure

Fullerene pentagon rule and nanotube chiral vector

The core ideas

The bond-angle/hybridization relationship, the interlayer cohesion energetics, the topological pentagon rule, and the nanotube metallicity rule are the four load-bearing quantitative relationships in this territory.

\theta_{sp^n} = \cos^{-1}\!\left(\frac{-1}{n}\right)

where: $\theta$ is the ideal bond angle for hybridization $sp^n$ (so $n=3$ for $sp^3$ gives 109.47°, and the $sp^2$ case is handled by trigonal planar geometry giving exactly 120°). This single formula explains why diamond's tetrahedra (109.5°) and graphite's flat sheets (120°) are geometrically incompatible — you cannot smoothly deform one lattice into the other without breaking bonds.

V - E + F = 2

where: $V$ , $E$ , $F$ are the numbers of vertices, edges, and faces of any convex polyhedron (Euler's formula). Applied to a fullerene cage — where every vertex is a 3-coordinate carbon and every face is a pentagon or hexagon — this equation forces the number of pentagons to be exactly 12, independent of the number of hexagons, as derived rigorously below.

n - m \equiv 0 \ (\mathrm{mod}\ 3) \implies \text{metallic tube}; \quad n-m \not\equiv 0 \implies \text{semiconducting tube}

where: $(n,m)$ are the chiral indices of the nanotube's roll-up vector. This rule comes from where the allowed circumferential wavevectors intersect graphene's Dirac points in the Brillouin zone.

E_g \approx \frac{2\hbar v_F}{d}

where: $E_g$ is the band gap opened in a semiconducting nanotube, $v_F \approx 8\times10^5\ \text{m/s}$ is graphene's Fermi velocity, and $d$ is the tube diameter. Reading: smaller-diameter semiconducting tubes have larger gaps — curvature-induced confinement is inversely proportional to tube radius, just as a particle-in-a-box gap grows as the box shrinks.

Derivation AWhy fullerenes need exactly 12 pentagons — Euler's theorem applied to a trivalent carbon cage

Goal. Show that any closed, convex carbon cage built only from pentagonal and hexagonal faces, with every carbon three-coordinate, must contain exactly 12 pentagons, regardless of how many hexagons are present.

Step 1 — Set up the counting variables. Let $P$ be the number of pentagonal faces and $H$ the number of hexagonal faces. Total faces:

F = P + H

Every carbon atom (vertex) is 3-coordinate ( $sp^2$ , three σ bonds), so counting bond-ends at vertices,

3V = 2E

since each edge is shared by exactly two vertices. This gives $V = 2E/3$ .

Step 2 — Count edges from faces. Each pentagon contributes 5 edges and each hexagon 6 edges, but every edge is shared by exactly two faces, so

2E = 5P + 6H

Step 3 — Substitute into Euler's formula. Euler's theorem for any convex polyhedron states

V - E + F = 2

Substitute $V = 2E/3$ and $F = P+H$ :

\frac{2E}{3} - E + (P+H) = 2

-\frac{E}{3} + P + H = 2

Now use $2E = 5P+6H$ , i.e. $E = \dfrac{5P+6H}{2}$ , and substitute:

-\frac{1}{3}\left(\frac{5P+6H}{2}\right) + P + H = 2

-\frac{5P+6H}{6} + P + H = 2

Multiply through by 6:

-(5P+6H) + 6P + 6H = 12

P = 12

Step 4 — Interpret. The number of hexagons $H$ cancels completely — it can be any non-negative integer (0, as in the small, unstable, highly-strained dodecahedrane-like C₂₀ cage; or 20, as in C₆₀; or hundreds, as in larger fullerenes and the caps of nanotubes) — but the pentagon count is rigidly fixed at 12. This is why every closed fullerene, from C₂₀ to C₇₀ to giant onion-like cages, and every capped nanotube end, must carry exactly twelve pentagonal defects distributed over its surface. Isolated-pentagon-rule (IPR) structures such as C₆₀ additionally require that no two pentagons share an edge, which is an energetic refinement (adjacent pentagons strain bond angles further and destabilize the cage) layered on top of this topological floor.

Go deeper · Derivation B · 2 minWhy n − m mod 3 sets metal vs. semiconductor in a carbon nanotube

Derivation BWhy n − m mod 3 sets metal vs. semiconductor in a carbon nanotube

Goal. Show that rolling a graphene sheet into a cylinder along chiral vector $(n,m)$ produces a metal when $n-m$ is a multiple of 3, and a semiconductor otherwise — using graphene's Dirac-point band structure and the periodic boundary condition imposed by rolling.

Step 1 — Graphene's band structure has Dirac points at K and K′. Graphene's honeycomb lattice has two atoms per unit cell (sublattices A and B). A tight-binding treatment with nearest-neighbor hopping integral $t$ gives the π-band dispersion

E(\mathbf{k}) = \pm t\left|1 + e^{i\mathbf{k}\cdot\mathbf{a}_1} + e^{i\mathbf{k}\cdot\mathbf{a}_2}\right|

This vanishes ( $E=0$ , valence and conduction bands touch) at two inequivalent corners of the hexagonal Brillouin zone, called K and K′ — the Dirac points. Everywhere else in the zone the two bands are split by a nonzero gap. Undoped graphene is thus a "zero-gap semiconductor" / semimetal precisely because the Fermi level sits exactly at these touching points.

Step 2 — Rolling imposes a periodic boundary condition. When the sheet is rolled along $\mathbf{C}_h = n\mathbf{a}_1+m\mathbf{a}_2$ , the electron wavefunction must be single-valued around the circumference:

\mathbf{k}\cdot\mathbf{C}_h = 2\pi q, \qquad q \in \mathbb{Z}

This quantizes the component of $\mathbf{k}$ around the tube circumference into a discrete ladder of allowed lines ("cutting lines") in the 2-D Brillouin zone, spaced by $2\pi/|\mathbf{C}_h|$ , while $\mathbf{k}$ along the tube axis remains continuous (1-D subbands).

Step 3 — Ask whether a cutting line passes exactly through K. The tube is metallic if and only if one of these discrete allowed lines passes exactly through a Dirac point K (or K′), because only then does a 1-D subband touch zero energy with no gap; otherwise the nearest cutting line misses K by a finite offset, and confinement opens a gap of order $E_g \sim 1/d$ . Writing the reciprocal-lattice vectors dual to $\mathbf{a}_1,\mathbf{a}_2$ and evaluating the condition $\mathbf{k}\cdot\mathbf{C}_h = 2\pi q$ at $\mathbf{k}=\mathbf{K}$ , the algebra (standard in tight-binding treatments of graphene) reduces to the simple integer condition

\frac{n-m}{3} \in \mathbb{Z}

i.e., a cutting line hits K exactly when $n - m$ is divisible by 3.

Step 4 — Read off the rule.

n-m \equiv 0 \ (\mathrm{mod}\ 3) \implies \text{metallic}

n-m \not\equiv 0 \ (\mathrm{mod}\ 3) \implies \text{semiconducting}, \quad E_g \approx \frac{2\hbar v_F}{d}

Special named cases: armchair tubes $(n,n)$ always have $n-m=0$ , hence are always metallic. Zigzag tubes $(n,0)$ are metallic only when $n$ is a multiple of 3. Chiral tubes follow the same $n-m$ rule regardless of the specific chirality angle. This is a purely topological/quantum-confinement selection rule — no doping or defect is required to switch a tube between conducting regimes, only the geometric choice of how the sheet was rolled.

Worked examples

Example 1 (foundational) — Bond-angle and coordination check for diamond vs. graphite. Confirm the hybridization assignment using the bond-angle formula and simple valence-electron bookkeeping. For diamond, every carbon has 4 σ bonds and 0 lone pairs, VSEPR/hybridization electron count = 4, so hybridization is $sp^3$ and

\theta = \cos^{-1}(-1/3) = 109.47^\circ

matching the experimentally measured C–C–C angle in diamond (109.5°) essentially exactly — a clean confirmation that no lone pairs or π-bonding perturb the geometry. For graphite, each carbon has 3 σ bonds (to 3 neighbors) and no lone pair, but retains 1 electron in a $p_z$ orbital for π bonding. The σ-framework hybridization is $sp^2$ , and trigonal planar geometry gives exactly

\theta = 120.0^\circ

also matching experiment. As a limiting-case check: if graphite's carbons were instead $sp^3$ (tetrahedral, 109.5°), the sheets could not be flat — this is exactly the geometric reason diamond and graphite cannot interconvert without breaking and reforming bonds (i.e., high-pressure/high-temperature synthesis is required industrially, not a simple flex of the lattice).

Example 2 (applied) — Estimating the semiconducting gap of a real, non-round nanotube. A single-walled carbon nanotube has chiral indices $(11,4)$ . Determine whether it is metallic or semiconducting, estimate its diameter, and if semiconducting, estimate its band gap. Step 1: metallicity test.

n - m = 11 - 4 = 7

7 \bmod 3 = 1 \neq 0 \implies \text{semiconducting}

Step 2: diameter. The nanotube diameter is

d = \frac{a_{cc}\sqrt{3}\sqrt{n^2+nm+m^2}}{\pi}

where $a_{cc} = 0.142$ nm is the graphene C–C bond length. Compute the radical term first:

n^2+nm+m^2 = 121 + 44 + 16 = 181

\sqrt{181} = 13.454

d = \frac{0.142 \times 1.732 \times 13.454}{3.1416}

d = \frac{3.309}{3.1416} = 1.053\ \text{nm}

Step 3: band gap. Using $E_g \approx 2\hbar v_F/d$ with $\hbar v_F \approx 0.53\ \text{eV}\,\cdot\,\text{nm}$ (a standard graphene parameter, since $v_F\approx 8\times 10^5$ m/s):

E_g \approx \frac{2 \times 0.53\ \text{eV}\,\cdot\,\text{nm}}{1.053\ \text{nm}} = \frac{1.06}{1.053}\ \text{eV} = 1.007\ \text{eV}

This places the (11,4) tube's optical/electronic gap near 1.0 eV — in the same range as silicon's indirect gap (1.12 eV), which is why semiconducting nanotubes are of interest for nanoscale transistors and infrared photodetectors. As a limiting-case sanity check: an armchair tube such as $(11,11)$ has $n-m=0$ , always metallic regardless of diameter — no gap estimate applies there at all, consistent with the derivation.

Example 3 (exam/competition level) — Density comparison and packing-efficiency argument between diamond and graphite. Diamond has a cubic unit cell with lattice parameter $a = 356.7$ pm and 8 carbon atoms per conventional cell. Graphite has a hexagonal unit cell with $a = 246.4$ pm (in-plane) and $c = 671.1$ pm (twice the interlayer spacing, since the unit cell spans two layers), containing 4 carbon atoms. Compute both densities and rationalize the difference structurally. Diamond:

V_{\text{diamond}} = a^3 = (356.7\times10^{-12}\ \text{m})^3 = 4.538\times10^{-29}\ \text{m}^3

Mass of 8 carbon atoms:

m = 8 \times \frac{12.011\ \text{g/mol}}{6.022\times10^{23}\ \text{mol}^{-1}} = 8 \times 1.995\times10^{-23}\ \text{g} = 1.596\times10^{-22}\ \text{g}

\rho_{\text{diamond}} = \frac{1.596\times10^{-22}\ \text{g}}{4.538\times10^{-29}\ \text{m}^3} = \frac{1.596\times10^{-22}}{4.538\times10^{-29}}\ \text{g/m}^3 = 3.518\times10^{6}\ \text{g/m}^3

Converting to g/cm³ (1 m³ = 10⁶ cm³):

\rho_{\text{diamond}} = 3.518\ \text{g/cm}^3

(literature value 3.51 g/cm³ — excellent agreement). Graphite: hexagonal cell volume is

V_{\text{graphite}} = \frac{\sqrt{3}}{2} a^2 c = 0.8660 \times (246.4\times10^{-12})^2 \times (671.1\times10^{-12})\ \text{m}^3

a^2 = (246.4\times10^{-12})^2 = 6.071\times10^{-20}\ \text{m}^2

V_{\text{graphite}} = 0.8660 \times 6.071\times10^{-20} \times 6.711\times10^{-10}\ \text{m}^3

V_{\text{graphite}} = 0.8660 \times 4.075\times10^{-29} = 3.529\times10^{-29}\ \text{m}^3

Mass of 4 carbon atoms:

m = 4 \times 1.995\times10^{-23}\ \text{g} = 7.981\times10^{-23}\ \text{g}

\rho_{\text{graphite}} = \frac{7.981\times10^{-23}\ \text{g}}{3.529\times10^{-29}\ \text{m}^3} = 2.262\times10^{6}\ \text{g/m}^3 = 2.262\ \text{g/cm}^3

(literature value 2.26 g/cm³ — excellent agreement). Rationalization.

\frac{\rho_{\text{diamond}}}{\rho_{\text{graphite}}} = \frac{3.518}{2.262} = 1.555

Diamond is roughly 55% denser than graphite despite having a longer individual C–C bond (154 pm vs. 141.5 pm) — because diamond packs bonds isotropically in three dimensions with no wasted interlayer gap, whereas graphite's structure is dominated by the large 335 pm van der Waals gap between sheets, which contributes volume without contributing mass-bearing bonds. This is the structural basis of the historical graphite-to-diamond high-pressure synthesis route: squeezing out that interlayer gap and reorganizing $sp^2$ sheets into an $sp^3$ network releases roughly 35–40% of the volume, which is exactly why enormous pressure (typically 5–6 GPa with metal-catalyst assistance, or >10 GPa uncatalyzed) is needed to force the transformation.

Limiting cases, scope & common traps

Regime / limit	What happens	Why
Perfect $sp^3$ , 3-D network	Diamond: insulator, hardest solid, isotropic	All 4 electrons in localized σ bonds; large HOMO–LUMO/band gap
Perfect $sp^2$ , 2-D infinite sheet	Graphite/graphene: anisotropic conductor	Delocalized π electrons free to move in-plane only
$sp^2$ sheet + exactly 12 pentagons, closed	Fullerene: molecular, soluble, discrete gap	Euler's theorem forces $P=12$ regardless of $H$
$sp^2$ sheet rolled, $(n,m)$ with $n-m\equiv0$ (mod 3)	Nanotube: metallic	Allowed k-line passes through Dirac point K
$sp^2$ sheet rolled, $(n,m)$ with $n-m\not\equiv0$	Nanotube: semiconducting, $E_g\propto 1/d$	Confinement misses K, opens curvature gap
Very small fullerenes (e.g. C₂₀, all-pentagon)	Highly strained, reactive, rare	Adjacent pentagons violate isolated-pentagon rule, raise strain energy
Bulk amorphous carbon / soot	Neither perfect diamond nor graphite lattice	Mixture of $sp^2$ / $sp^3$ domains, disordered network

The chemist behind it

Origins · 1 minHarold W. Kroto and how this was discovered

Harold W. Kroto

1939–2016 · Fullerene discovery, Sussex/Rice collaboration, 1985

Kroto, working with Robert Curl and Richard Smalley at Rice University, was using laser-vaporization mass spectrometry to simulate the carbon chemistry of red-giant stars and interstellar dust clouds — an astrochemistry motivation, not a materials-science one. In September 1985, mass spectra of laser-ablated graphite showed an anomalously intense peak at mass 720, corresponding to a C₆₀ cluster, with a companion peak at C₇₀. Within days the team proposed a closed, cage-like, soccer-ball structure — later named buckminsterfullerene after the geodesic domes of Buckminster Fuller — to explain the exceptional stability of exactly this mass. Kroto, Curl, and Smalley shared the 1996 Nobel Prize in Chemistry for the discovery. The structure was confirmed structurally once Krätschmer and Huffman developed a bulk arc-discharge synthesis in 1990 that produced enough C₆₀ for X-ray and NMR characterization, opening the entire field of fullerene and (soon after) nanotube chemistry.

History

- 1772 — Antoine Lavoisier shows diamond, on combustion, yields only carbon dioxide, proving diamond is a form of pure carbon.

- 1779 — Carl Wilhelm Scheele demonstrates graphite also burns to carbon dioxide, establishing graphite as another form of the same element.

- 1985 — Kroto, Curl, Smalley discover C₆₀ buckminsterfullerene via laser-vaporization mass spectrometry.

- 1991 — Sumio Iijima reports carbon nanotubes in the soot of arc-discharge fullerene synthesis, using transmission electron microscopy.

- 2004 — Andre Geim and Konstantin Novoselov isolate single-layer graphene by mechanical exfoliation, earning the 2010 Nobel Prize in Physics.

Why can't diamond conduct electricity if it's made of carbon, which is often a conductor?

Conductivity in solids depends on whether electrons can move freely near the Fermi level, not on the element itself. In diamond, all four valence electrons per carbon are locked into strong, localized $sp^3$ σ bonds, and the gap between the filled (valence) and empty (conduction) bands is large, about 5.5 eV. There are essentially no mobile electrons at room temperature, so diamond is an excellent insulator — in fact it's used as an electrical insulator in high-power electronics precisely because it also conducts heat so well, letting it remove waste heat without leaking current. Graphite conducts because its structure is different: one electron per carbon sits in a delocalized π system that can move freely within a sheet.

Is graphene just a very thin sheet of graphite, or is it fundamentally different?

Structurally, graphene is exactly one atomic layer of the graphite lattice — same hexagonal $sp^2$ network, same 141.5 pm bond length. But removing the interlayer coupling changes the physics substantially: an isolated sheet has different phonon behavior, extreme mechanical flexibility, and, most importantly, direct experimental access to its 2-D Dirac-cone band structure without interference from neighboring layers. Bulk graphite's electronic structure is also influenced by weak interlayer coupling that graphene, by definition, does not have. So graphene is graphite's parent sheet, but isolating it created a genuinely new research field rather than a trivial "thin graphite."

Why does C₆₀ specifically form, rather than some other fullerene, in the original mass spectrometry experiments?

C₆₀ is the smallest fullerene in which all twelve pentagons can be arranged so that no two pentagons share an edge — the isolated-pentagon rule (IPR). Fused pentagon-pentagon edges introduce extra angle strain and localized reactive double-bond character, destabilizing the cage. C₆₀'s truncated-icosahedron structure spaces the twelve pentagons maximally apart among the twenty hexagons, giving it unusually high symmetry ( $I_h$ ) and low strain, which is why it appeared as a dominant, unusually stable peak against a background of many possible cluster sizes.

Are diamond and graphite really the same total energy, or is one intrinsically more stable?

At standard conditions (298 K, 1 atm), graphite is the thermodynamically stable allotrope; diamond is metastable, lying about 2.9 kJ/mol higher in Gibbs energy per mole of carbon. The reason diamonds do not spontaneously turn into graphite on a human timescale is kinetic: converting the rigid $sp^3$ network into flat $sp^2$ sheets requires breaking strong covalent bonds and reorganizing the whole lattice, a process with a very high activation barrier at room temperature. Diamond is thus often cited as the textbook example of a kinetically persistent metastable phase.

How is graphene actually made, both in research labs and at industrial scale?

The original 2004 isolation used mechanical exfoliation — repeatedly peeling graphite with adhesive tape until single layers remained, which gives extremely high-quality but tiny, irregular flakes suited to fundamental research. Scalable methods include chemical vapor deposition (CVD), where a hydrocarbon gas decomposes on a catalytic metal foil (commonly copper) to grow large continuous graphene sheets that are then transferred onto a target substrate, and liquid-phase exfoliation, where graphite is sonicated in a solvent to shear layers apart in bulk, at some cost to sheet size and defect density. Each method trades off crystal quality against scalability and cost.

Do all carbon nanotubes with the same diameter have the same electronic properties?

No — diameter alone is insufficient because it depends only on $\sqrt{n^2+nm+m^2}$ , while metallicity depends on the separate condition $n-m \bmod 3$ . Two tubes can have nearly identical diameters but different $(n,m)$ pairs, placing one on the "hits the Dirac point" side and the other on the "misses it" side, so one is metallic and the other semiconducting with a finite gap. This is why bulk nanotube synthesis, which typically produces a statistical mixture of chiralities, yields a mixture of metallic and semiconducting tubes that must be separated (e.g., by density-gradient ultracentrifugation or selective chemistry) for electronic applications requiring a single type.

Why is diamond less dense than close-packed metals despite being the "hardest" material?

Hardness measures resistance to localized deformation (indentation, scratching), which in diamond comes from the strength and directional rigidity of covalent $sp^3$ bonds — any attempt to distort a tetrahedral angle meets strong resistance from orbital overlap energetics. Density measures how much mass is packed per unit volume, which depends on packing efficiency, not bond strength. Diamond's tetrahedral network is actually a fairly open structure (packing efficiency ~34%, the same topology as ice or the diamond cubic lattice), far less efficient than close-packed metals (~74%), so diamond can be both "hard" and "not especially dense" simultaneously — these are independent structural consequences of directional covalent bonding versus packing geometry.