Graphite: Structure, Properties & Applications

• Graphite is a layered carbon allotrope characterized by stacked graphene sheets with strong sp² covalent bonds and weak interlayer interactions.
• Its anisotropic bonding yields exceptional electronic structure, easy cleavage, and effective intercalation chemistry for energy storage applications.
• Advanced analyses such as XRD, Raman spectroscopy, and electron microscopy reveal its hierarchical structure, defect taxonomy, and pressure-induced transformation pathways.

Graphite is the thermodynamically stable form of carbon under ordinary conditions. Its defining structural motif is a stack of graphene sheets with strong in-plane $sp^2$ covalent bonding and much weaker interlayer interactions, usually in staggered $AB$ stacking; this extreme bonding anisotropy governs its semimetallic electronic structure, anisotropic transport, easy cleavage, intercalation chemistry, and sensitivity to pressure, defects, and mesostructure (Amsler et al., 2011, Zhmurikov, 2015).

1. Crystal chemistry and structural hierarchy

At the crystallographic level, hexagonal graphite is described with space group $C6/mmc-D_{6h}^4$, 4 atoms per unit cell, $a = 2.46\ \text{\AA}$, $c = 6.71\ \text{\AA}$, in-plane C–C distance $1.42\ \text{\AA}$, interlayer spacing $3.35\ \text{\AA}$, and theoretical density $2.267\ \text{g/cm}^3$ (Zhmurikov, 2015). The monograph also distinguishes rhombohedral graphite as a metastable polymorph that rarely occurs pure, although up to 30% rhombohedral phase may occur in natural graphite (Zhmurikov et al., 2013). The interlayer-bonding contrast is large: the cited bond energies are $418$–$461\ \text{kJ/mol}$ for in-plane $AB$0-bonds and $AB$1–$AB$2 for interlayer bonding, which explains why graphite behaves as a layered solid rather than a three-dimensionally rigid network (Zhmurikov, 2015).

The ideal crystal is only part of the material description. Engineering graphite is repeatedly treated as a hierarchical composite consisting of graphitic microcrystallites, filler and binder phases, pores, and microcracks; the most important structural parameters include density, meso- and microporosity, texture, anisotropy, and crystallite sizes $AB$3 and $AB$4 (Zhmurikov, 2015). X-ray diffraction is used to extract interlayer spacing, preferred orientation, and coherent scattering region sizes through Bragg’s law,

$AB$5

and Scherrer-type line-broadening analysis (Zhmurikov, 2015).

The defect taxonomy is correspondingly broad. The review literature describes interlayer defects such as stacking faults and turbostratic disorder, and in-plane defects including vacancies, implanted impurity atoms, $AB$6-type bond rearrangements, edge defects, screw and edge dislocations, and twin defects (Zhmurikov et al., 2013). For non-ideal graphitic carbons, “turbostratic carbon” denotes material made of fairly good hexagonal sheets but with random relative displacement and loss of regular stacking order (Zhmurikov, 2015). In fine-grained artificial graphites, HRTEM and SEM resolve inclined intercrystallite boundaries, edge dislocations, mesopores down to 10 nm, fullerene-like formations of size 1–3 nm, and layered aggregates ranging from hundreds of nanometers to several microns (Zhmurikov, 2015).

2. Electronic, optical, and thermal response

Graphite is a layered 3D semimetal. The review literature states that weak interlayer interaction produces a slight overlap of valence and conduction bands of about $AB$7–$AB$8, with equal electron and hole concentrations of approximately $AB$9 at $C6/mmc-D_{6h}^4$0 (Zhmurikov, 2015). Near the Dirac points, the low-energy dispersion is linearized as $C6/mmc-D_{6h}^4$1, with $C6/mmc-D_{6h}^4$2 (Zhmurikov et al., 2013). Infrared optics refines this picture: in the range $C6/mmc-D_{6h}^4$3–$C6/mmc-D_{6h}^4$4, the conductivity per graphite layer is close to the universal graphene value $C6/mmc-D_{6h}^4$5, while interlayer hopping $C6/mmc-D_{6h}^4$6 produces characteristic kinks and thresholds, including a kink at $C6/mmc-D_{6h}^4$7; using an observed kink at $C6/mmc-D_{6h}^4$8, the interlayer parameter was inferred as $C6/mmc-D_{6h}^4$9 (Falkovsky, 2010).

The electromagnetic response is tensorial and strongly anisotropic even at X-ray energies. The dielectric tensor is written in the principal-axis frame as

$a = 2.46\ \text{\AA}$0

with $a = 2.46\ \text{\AA}$1 for $a = 2.46\ \text{\AA}$2 and $a = 2.46\ \text{\AA}$3 for $a = 2.46\ \text{\AA}$4 (Draine, 2016). Two optically active lattice resonances are especially noted: an in-plane mode at $a = 2.46\ \text{\AA}$5 and an out-of-plane mode at $a = 2.46\ \text{\AA}$6–$a = 2.46\ \text{\AA}$7, the latter being the more promising observational signature (Draine, 2016). For randomly oriented single-crystal grains, the commonly used $a = 2.46\ \text{\AA}$8–$a = 2.46\ \text{\AA}$9 approximation is exact as $c = 6.71\ \text{\AA}$0, stays within about $c = 6.71\ \text{\AA}$1 in the optical and UV, but can err by as much as $c = 6.71\ \text{\AA}$2 at $c = 6.71\ \text{\AA}$3 (Draine, 2016).

Thermal transport is equally exceptional. In thin highly oriented pyrolytic graphite, the room-temperature in-plane thermal conductivity of an $c = 6.71\ \text{\AA}$4-thick specimen was measured as about $c = 6.71\ \text{\AA}$5, exceeding natural diamond and slightly exceeding isotopically purified graphene in the comparison used there (Machida et al., 2019). The same study found that thinning graphite along the $c = 6.71\ \text{\AA}$6-axis increases, rather than decreases, in-plane thermal conductivity, and that the thermal diffusivity exhibits a Knudsen minimum near $c = 6.71\ \text{\AA}$7 and a Poiseuille-type maximum that shifts upward with decreasing thickness, reaching about $c = 6.71\ \text{\AA}$8 in the thinnest sample (Machida et al., 2019). The proposed interpretation is partially hydrodynamic phonon flow enabled by extreme phonon-dispersion anisotropy.

3. Interlayer energetics, cleavage, and intercalation

The weak interlayer coupling that distinguishes graphite from diamond is quantifiable. A direct measurement based on self-retraction of superlubric twisted bicrystal graphite gave the incommensurate basal-plane cleavage energy

$c = 6.71\ \text{\AA}$9

with near-invariance from $1.42\ \text{\AA}$0 to $1.42\ \text{\AA}$1 and over twist angles $1.42\ \text{\AA}$2 (Wang et al., 2015). Using an anisotropic Peierls–Nabarro grain-boundary model with $1.42\ \text{\AA}$3, the ideal $1.42\ \text{\AA}$4 graphite cleavage energy was inferred as

$1.42\ \text{\AA}$5

so that the basal-plane surface energy is about $1.42\ \text{\AA}$6, with approximate derived values $1.42\ \text{\AA}$7 and $1.42\ \text{\AA}$8 (Wang et al., 2015). The same work formalized the measurement through

$1.42\ \text{\AA}$9

because in the superlubric limit the applied shear force balances the retraction force associated with creating new basal-plane area (Wang et al., 2015).

Graphite’s weakly bonded galleries also make it the canonical lithium-ion anode host. The intercalation reaction is written as

$3.35\ \text{\AA}$0

with staging defined by the number of graphene layers between adjacent Li layers (Pande et al., 2016). In the first-principles phase-diagram study, pristine graphite favored $3.35\ \text{\AA}$1 over $3.35\ \text{\AA}$2 stacking by $3.35\ \text{\AA}$3, whereas fully lithiated $3.35\ \text{\AA}$4 favored $3.35\ \text{\AA}$5-stacked intercalated layers (Pande et al., 2016). The predicted sequence included low-$3.35\ \text{\AA}$6 stage-4 compounds, stage-2 compounds at intermediate $3.35\ \text{\AA}$7, and stage-1 compounds approaching $3.35\ \text{\AA}$8, but the confidence analysis showed that many intermediate phases are much less robust than the low-$3.35\ \text{\AA}$9 stage-4 regime and the end member $2.267\ \text{g/cm}^3$0 (Pande et al., 2016). The same analysis found essentially no change in the phase diagram from $2.267\ \text{g/cm}^3$1 to $2.267\ \text{g/cm}^3$2, with only about $2.267\ \text{g/cm}^3$3 decrease in the final intercalation plateau, suggesting that low-temperature lithium plating is probably dominated more by kinetics than thermodynamics (Pande et al., 2016).

4. Graphitization and pressure-driven phase transformations

Graphite is thermodynamically stable yet difficult to synthesize from less ordered carbon because graphitization is kinetically hindered. In graphitizing carbon derived from PVC, carbonization below about $2.267\ \text{g/cm}^3$4 yields aromatic or graphenic fragments that remain small, misoriented, and separated by an interlayer distance larger than that of graphite; full graphitization occurs only in the $2.267\ \text{g/cm}^3$5–$2.267\ \text{g/cm}^3$6 range, with a critical temperature near $2.267\ \text{g/cm}^3$7 in earlier work and about $2.267\ \text{g/cm}^3$8 under short pulsed heating (Martin et al., 2022). The mechanistic claim of that study is that screw dislocations are the dominant topological defects blocking graphitization. XRD showed that the decisive structural changes are growth of $2.267\ \text{g/cm}^3$9 and reduction of $418$0, while $418$1 remains essentially unchanged; the inferred activation barrier was $418$2 (Martin et al., 2022). HRTEM and molecular dynamics then linked graphitization to the glide and annihilation of screw dislocations, rather than to simple thickening of graphitic stacks (Martin et al., 2022).

Under compression, graphite can also transform away from its layered $418$3 structure. For cold-compressed graphite, one study predicted a new orthorhombic $418$4 allotrope, Z-carbon, that becomes more stable than graphite at $418$5 and matches unexplained X-ray diffraction and Raman features observed between about 10 and 25 GPa (Amsler et al., 2011). The proposed pathway starts from naturally staggered $418$6-stacked graphite, with sheet sliding along $418$7 to $418$8 stacking, decreasing interlayer distance, and alternating armchair-zigzag buckling that converts the network to fully $418$9 bonding (Amsler et al., 2011). Z-carbon is predicted to be dynamically stable, to have bulk modulus $461\ \text{kJ/mol}$0, Vickers hardness $461\ \text{kJ/mol}$1, and an indirect $461\ \text{kJ/mol}$2 gap of about $461\ \text{kJ/mol}$3 (Amsler et al., 2011).

At higher pressures and temperatures, graphite transforms to diamond by a different defect-mediated route. Large-scale molecular dynamics found that diamond nuclei originate at graphite grain boundaries and then propagate in two preferred directions, $461\ \text{kJ/mol}$4 and the faster $461\ \text{kJ/mol}$5 direction (Zhu et al., 2020). In that scenario, cubic diamond is the kinetically favored product, whereas hexagonal diamond appears only as minor twinning structures in two main directions (Zhu et al., 2020). High-resolution TEM confirmed the coherent interface

$461\ \text{kJ/mol}$6

supporting the argument that graphite microstructure, corrugation, and grain boundaries control the transformation pathway (Zhu et al., 2020).

5. Engineered, chemically modified, and architected graphites

Graphite is also a platform for deliberate mesoscale and chemical engineering. A direct-ink-writing study developed a colloidal graphite ink from commercial graphite powders and a small amount of nanoclay, enabling room-temperature printing of complex architectures without further heat treatment (Sajadi et al., 2020). Relative to an unmodified graphite slurry, the modified ink had viscosity $461\ \text{kJ/mol}$7 at $461\ \text{kJ/mol}$8 instead of $461\ \text{kJ/mol}$9, and storage modulus $AB$00 at $AB$01 strain instead of $AB$02 (Sajadi et al., 2020). The printed structures retained about $AB$03 thermal conductivity versus $AB$04 for bulk graphite, had resistivity of approximately $AB$05, and showed compressive specific energy absorption and fracture strain increased by more than 6 times and 3 times, respectively, relative to bulk graphite (Sajadi et al., 2020).

Chemical modification under high pressure can push graphite far from its undoped state. Using adamantane and ortho-carborane precursors at $AB$06, boron-doped graphite was synthesized with $AB$07 at $AB$08 and $AB$09 at $AB$10, with coherent scattering region sizes increasing from about 480 nm to 720 nm as temperature increased (Bagramov et al., 2021). XPS assigned the main B1s peak at $AB$11 to substitutional B–C bonding in graphite layers, and the authors argued that at boron concentrations above about $AB$12, electrostatic repulsion between negatively charged substitutional boron centers could favor periodic, equidistant distribution within the layers (Bagramov et al., 2021).

A more radical proposal replaces graphene sheets altogether with porous graphtriyne sheets to form a “nanoporous graphite.” In that first-principles design, the equilibrium interlayer spacing is $AB$13, only $AB$14 larger than pristine graphite, yet the pore topology yields vertically connected nanochannels and no significant barrier for gas diffusion (Bartolomei et al., 2016). For the inner pore of the trilayer channel model, the reported binding energies were $AB$15 for $AB$16, $AB$17 for $AB$18, $AB$19 for $AB$20, and $AB$21 for $AB$22, with an estimated $AB$23 storage capacity of $AB$24, equivalent to $AB$25 (Bartolomei et al., 2016). This suggests a porous graphite analogue that is nearly as compact as ordinary graphite but functionally closer to a molecular sieve.

6. Nuclear graphite, irradiation, and unresolved claims

Graphite’s technological centrality is most explicit in nuclear systems. Review articles describe it as a primary construction material and moderator because of low neutron absorption, high-temperature stability, thermal conductivity, passively safe fuel reactivity control, and low impurity requirements; reactor graphite should have total impurity content not exceeding about $AB$26, while semiconductor-grade graphite may require impurity levels below $AB$27 (Zhmurikov et al., 2013). Those same reviews emphasize that high-strength artificial graphite is a filler–binder composite whose durability depends on density, pore morphology, crystallite size, anisotropy, and intergranular phase, and they state explicitly that the finer the graphite composite structure, the higher its strength and the longer its service life (Zhmurikov, 2015).

The pore network of nuclear graphite is itself multiscale. A neutron-scattering and neutron-imaging study of highly pure nuclear graphites RID and PGA found a fractal morphology extending from $AB$28 to $AB$29, with both surface and mass fractal dimensions very close to $AB$30 (Zhou et al., 2015). The fitted upper cutoff length was about $AB$31–$AB$32, and the crossover scale $AB$33 was about $AB$34–$AB$35 (Zhou et al., 2015). This means that graphite porosity is not concentrated at one dominant scale but distributed over rough interfaces and connected pore networks across roughly six orders of magnitude (Zhou et al., 2015).

Ion irradiation studies show how quickly this structure can disorder. Confocal Raman imaging of IG-110, NBG-17, and an in-house ASR-0RB graphite after $AB$36 Ar$AB$37 and He$AB$38 irradiation at $AB$39 found that heavier Ar$AB$40 ions produce substantially stronger disordering than He$AB$41 (Gawęda et al., 2023). At $AB$42, SRIM gave $AB$43 for Ar and $AB$44 for He; at $AB$45, the values were $AB$46 and $AB$47, respectively (Gawęda et al., 2023). Raman signatures showed little change from pristine material to $AB$48, then a threshold-like increase in disorder at $AB$49 and $AB$50, with stronger G- and D-band broadening, loss of second-order bands, and stronger crystallite-size reduction for Ar$AB$51 (Gawęda et al., 2023).

Graphite is also the subject of unresolved electronic claims. One recent study on HOPG and flexible graphite gaskets reported “superconducting-like” magnetization loops persisting to $AB$52 and extrapolated a transition temperature $AB$53, while resistance anomalies near 520–560 K were interpreted as evidence for non-percolative granular superconductivity (Rousset-Zenou et al., 2022). The same paper, however, depends on subtraction of the intrinsic graphite diamagnetism, fitting and removing a sigmoidal ferromagnetic contribution, and extrapolating loop widths beyond the measured temperature range; it does not report a zero-resistance state or bulk Meissner proof, and it explicitly interprets the effect as granular and generally non-percolative (Rousset-Zenou et al., 2022). The most defensible reading is therefore not uniform bulk superconductivity in graphite, but a claim of localized superconducting-like regions whose existence remains contingent on model-based signal decomposition (Rousset-Zenou et al., 2022).