Graphite: Structure, Properties & Innovations

• Graphite is a layered carbon allotrope defined by ABAB stacking, strong in-plane sp² bonding, and a semimetallic electronic structure with Dirac-like dispersion.
• Optical and thermal analyses reveal marked anisotropy, with a dielectric tensor showing hyperbolic behavior and in-plane thermal conductivity reaching record levels in thin specimens.
• Defect engineering and processing routes—from irradiation and boron doping to 3D printing—critically influence graphitization and give rise to interfacial phenomena and emergent states.

to be structured by crystallographic, transport, defect, processing, and interfacial phenomena. Must avoid concrete claims not in block, such as "ambient conditions" unless from paper. It is in several. Good.

Need maybe mention space group P63/mmc from 2025 optical paper and C6/mmc from review. Hmm discrepancy? One says hexagonal layered structure space group P6_3/mmc; review says C6/mmc likely typo? Better avoid space group conflict maybe only mention layered hexagonal structure and ABAB stacking. But if mention space group, use one paper only and maybe "reported as space group P6_3/mmc" (Toksumakov et al., 17 Nov 2025). Could avoid conflict.

Need ensure each concrete claim traceable. For example "graphite is thermodynamically stable form of carbon" appears in (Martin et al., 2022) abstract and others. Good.

Let's use citations once per section maybe at relevant paragraphs. Need not cite every sentence.

Need likely not mention paper titles unless useful. Since citations okay. But article style may not need titles.

Need include equations only those from data. Could include dielectric tensor, conductivity low-energy maybe not too much. Technical audience expects. But keep concise.

Need distinguish inferences: e.g., "This suggests greater relative dissipation/absorption at higher frequency" from 1806 paper. Need mark as such.

Need handle "GRAPHITE" as topic, not Graphite ML paper unless maybe as disambiguation? But the topic is material graphite, not the graph ML algorithm named Graphite. Since data includes irrelevant (grover et al., 2018) Graphite graphs method. But user asked encyclopedia article on GRAPHITE ambiguous maybe maybe should focus on material because majority data about carbon material and title. Need ignore algorithm? Since topic "GRAPHITE" likely carbon allotrope. Should not include disambiguation unless from data. But maybe mention "The term 'Graphite' also names an algorithmic framework..."? Since provided data includes that. However task likely on topic below GRAPHITE and dataset includes one non-material paper because arxiv id includes Graphite acronym. But task says encyclopedia article on topic GRAPHITE, not necessarily carbon allotrope? Yet majority of data clearly carbon material. Should not include ML algorithm; it would confuse and violate fidelity? It is in data but topic likely carbon allotrope. Could mention that "In unrelated machine-learning literature, 'Graphite' denotes a graph generative model" but that would be weird and likely outside scope. Better ignore irrelevant paper.

Need maybe mention range of applications: RF attenuation, nanophotonics, nuclear reactors, 3D printing, HPHT synthesis. Could include. Need be careful with numbers.

Need maybe mention cold-compressed graphite transformation to Z-carbon and graphite-to-diamond transition; yes under extreme conditions.

Need maybe discuss nanoscale gap emergence and superconductivity claims as controversies. Good.

Need maybe section 1 crystal structure and allotropy: layered sp2, hexagonal graphite ABAB, rhombohedral up to 30%, interlayer spacing 3.35 Å, a=2.46 Å c=6.71 Å, in-plane C-C 1.42 Å, bond energies, anisotropy; semimetal, band overlap 0.03-0.04 eV, Dirac-like E = ħv_F k, c-axis anisotropy in effective mass. All from review (Zhmurikov, 2015). Also optical anisotropy from (Toksumakov et al., 17 Nov 2025), infrared conductivity per layer close to e^2/4ħ (Falkovsky, 2010).

Need maybe mention dielectric tensor and in-plane vs out-of-plane optical. Thermal section: 8.5 μm graphite κ~4300 W/mK at room temp, fivefold increase from 580 to 8.5 μm due to partially hydrodynamic phonon flow and extreme phonon-dispersion anisotropy (Machida et al., 2019). Should not say "highest among solids" unless paper says one of best and above diamond, okay. Could state "room-temperature in-plane thermal conductivity of 8.5 μm-thick graphite was 4300 W/m-K, above natural-abundance diamond and slightly above isotopically pure graphene" (Machida et al., 2019).

Need section on defects and graphitization from (Martin et al., 2022, Zhmurikov, 2015, Gawęda et al., 2023). Could explain screw dislocations control graphitization, XRD metrics d{002}, La, Lc, critical temp 2250°C literature and 2500°C nonequilibrium, barrier 6 ± 1 eV, HRTEM interdigitated c/2 fringes. Also irradiation in nuclear graphite with Ar+/He+ causing disorder, L_a reduction to 10-15 nm under Ar at 1e16 ion/cm^2, etc. Good.

Need section on processing and engineered forms from (Bagramov et al., 2021) boron-doped graphite, (Sajadi et al., 2020) 3D printing, (Bartolomei et al., 2016) nanoporous graphtriyne-based “graphite”, (Morarka et al., 2018) pencil-drawn films maybe in applications. Need not overdo. Could discuss forms and composites. Need mention low-cost pencil coating ~1.8 μm thickness, conductivity 457.67 S/m, attenuation increasing with frequency and layers; radio noise suppression with six A4 sheets rolled to 12 layers. Good. And 3D printed graphite room temperature with clay additive, thermal conductivity ~9 W/mK, resistivity ~0.18 Ω cm, etc. And highly boron-doped graphite 1.13/1.4 at.% at 5.5 GPa 1400/1700°C.

Need section on interfaces, topology, and emergent states: graphite-on-graphite proposal near-flat bands, spin ordering or superconducting pairing (Volovik et al., 2017); nanoscale patterned HOPG bandgap 112±15 meV due mechanical distortions (Chaiyachad et al., 2024); granular superconductivity above 500 K claims (Rousset-Zenou et al., 2022) controversial. Need phrase "claimed" and note evidence indirect and analysis-dependent exactly from data. Maybe mention "the paper argues" rather than asserting as established. Need objective. Could include cold compressed Z-carbon maybe maybe in extreme conditions section instead.

Need maybe 6 sections: 1 Crystallographic and electronic framework 2 Anisotropic optical and thermal transport 3 Defects, disorder, and graphitization 4 Processing, composites, and engineered architectures 5 High-pressure and irradiation behavior 6 Interfacial, nanoscale, and contested emergent phenomena

Need maybe include numbered headings as "## 1. ..." etc.

Need citations at most once per section for each paper. We can cite maybe two-three papers per section. Must avoid repeated same id multiple times within same section. Can cite same paper next section again okay. Should be fine.

Need ensure all claims from data block. For example if we say "graphite is semimetallic" appears in optical/2014/2411. yes. If say "thermodynamically stable under ordinary conditions" from 1508 and 2206. Could use.

Need maybe include formulas: dielectric tensor; thermal diffusivity; attenuation formula; lifetime law? Maybe some essential ones. Could include the infrared conductivity scale $\sigma_0 = e^2/(4\hbar c_0)$, or maybe not. Maybe include E = ħv_F k and density of states? Hmm maybe too many. Need concise but comprehensive. Since target is technical audience, some equations okay. Let's include a few: dielectric tensor, Dirac dispersion, thermal diffusivity maybe, Zhurkov lifetime law? But article topic graphite broad. Maybe include only a couple significant equations. Since no multiple headings? Could still include in sections.

Need no references to writing process. Start with intro paragraph no heading. Let's draft around ~1800-2200 words maybe. Need only article itself.

Let's carefully craft with claims.

Graphite is the thermodynamically stable allotrope of carbon under ordinary conditions in the materials literature represented here, defined by $sp^2$-bonded carbon sheets arranged in a layered crystal with strong in-plane bonding and much weaker interlayer coupling. Across the cited work, graphite appears simultaneously as a semimetal, a strongly anisotropic optical and thermal material, a technologically important high-temperature ceramic, a defect-sensitive precursor to other carbon phases, and a host for interface-driven or nanoscale emergent phenomena. Its scientific description is therefore inseparable from anisotropy, stacking, disorder, and mesoscale structure (Zhmurikov, 2015).

1. Crystal structure, bonding, and anisotropic electronic framework

Graphite is described as a layered carbon crystal in which each carbon atom forms three in-plane bonds using $sp^2$ hybridization, leaving a $p_z$ orbital perpendicular to the layers. The hexagonal phase is reported with stacking sequence $ABAB\ldots$, lattice parameters $a = 2.46\ \text{\AA}$ and $c = 6.71\ \text{\AA}$, in-plane C–C distance $1.42\ \text{\AA}$, and interlayer spacing $3.35\ \text{\AA}$; rhombohedral graphite is treated as a metastable stacking variant that can occur in natural graphite at up to $30\%$ (Zhmurikov, 2015). The strong contrast between in-plane bond energy, reported as $sp^2$0–$sp^2$1, and interlayer bonding, reported as $sp^2$2–$sp^2$3, underlies the marked anisotropy of the material (Zhmurikov, 2015).

The low-energy electronic structure is described in two complementary ways. In the graphene-like approximation, the dispersion near $sp^2$4 and $sp^2$5 is linear,

$sp^2$6

with $sp^2$7, while the density of states vanishes linearly at the Dirac point in the ideal two-dimensional limit (Zhmurikov, 2015). When interlayer coupling is included, graphite becomes a semimetal with valence and conduction bands overlapping by about $sp^2$8–$sp^2$9, and with equal electron and hole concentrations of about $sp^2$0 at $sp^2$1 (Zhmurikov, 2015). In the infrared theory based on the low-energy Slonczewski-Weiss-McClure description, the in-plane conductivity per graphite layer is very close to graphene’s universal optical conductance $sp^2$2, while graphite-specific spectral features arise from $sp^2$3-dependent interlayer hopping and produce a kink and threshold structure at $sp^2$4 with $sp^2$5 (Falkovsky, 2010).

This combination of nearly two-dimensional in-plane electronic structure and finite interlayer coherence explains why graphite occupies an intermediate position between graphene and fully three-dimensional carbon crystals. A plausible implication is that many of graphite’s unusual transport and interface phenomena originate from that same coexistence of layered Dirac-like physics and weak but consequential interlayer coupling.

2. Optical and thermal response

Graphite is treated as a strongly anisotropic uniaxial optical material with dielectric tensor

$sp^2$6

where the in-plane and out-of-plane responses differ qualitatively rather than merely quantitatively (Toksumakov et al., 17 Nov 2025). Over $sp^2$7–$sp^2$8, the in-plane dielectric function is reported as metallic-like or plasmonic-like in the ultraviolet, with $sp^2$9 there, a zero crossing near $p_z$0, a strong $p_z$1 absorption peak near $p_z$2, and increasing near-infrared loss attributed to Drude-like free-carrier absorption (Toksumakov et al., 17 Nov 2025). By contrast, the out-of-plane response remains dielectric-like, with $p_z$3 and $p_z$4 across the measured range (Toksumakov et al., 17 Nov 2025). Because $p_z$5 while $p_z$6 in the ultraviolet, graphite is identified there as a natural type-II hyperbolic material (Toksumakov et al., 17 Nov 2025).

At longer wavelengths, the in-plane optical response can be continued from the measured $p_z$7–$p_z$8 range to the millimetric and DC limits by combining a Drude term, modified Drude terms, and a $p_z$9-resonance Lorentzian (Papoular et al., 2014). In this description, both the plasma frequency and the scattering rate decrease with temperature and level off near $ABAB\ldots$0, while spectral weight is transferred from an intermediate interband plateau into a low-frequency intraband response (Papoular et al., 2014). This picture sharpens the distinction between graphite’s bulk optical anisotropy and earlier extrapolations that produced artificial far-infrared structure (Papoular et al., 2014).

Thermally, graphite is presented as one of the best heat conductors among solids, but with behavior that depends strongly on thickness and direction. In thin highly oriented pyrolytic graphite, the in-plane thermal conductivity of an $ABAB\ldots$1-thick specimen reaches about $ABAB\ldots$2 near room temperature, about five times the value of a $ABAB\ldots$3 sample from the same mother crystal (Machida et al., 2019). The thermal diffusivity

$ABAB\ldots$4

is nonmonotonic, with a Knudsen minimum near $ABAB\ldots$5 and a Poiseuille maximum near $ABAB\ldots$6 in thick samples; these hydrodynamic signatures strengthen and shift upward in temperature as thickness decreases (Machida et al., 2019). The proposed mechanism is that graphite’s extreme phonon-dispersion anisotropy suppresses some momentum-relaxing Umklapp processes when thickness along the $ABAB\ldots$7-axis is reduced, so thinning can enhance rather than degrade in-plane heat transport (Machida et al., 2019).

3. Defects, disorder, and graphitization

The cited literature treats graphite not simply as an ideal lattice but as a defect-rich, mesostructured material whose properties depend on stacking faults, vacancies, impurities, twisted layers, dislocations, pores, and grain-boundary phases (Zhmurikov, 2015). In graphitizing carbons, the key kinetic barrier to formation of ordered Bernal graphite is proposed to be the screw dislocation: a ramp-like interplanar defect that pins lateral graphenic crystallite growth $ABAB\ldots$8, maintains turbostratic disorder, and prevents ideal ABA stacking until the defects glide and annihilate at high temperature (Martin et al., 2022). X-ray diffraction tracks this process through the interlayer spacing $ABAB\ldots$9, the stacking-direction coherence length $a = 2.46\ \text{\AA}$0, and the lateral graphenic size $a = 2.46\ \text{\AA}$1; in pulsed graphitization experiments, $a = 2.46\ \text{\AA}$2 increases substantially only above about $a = 2.46\ \text{\AA}$3, while $a = 2.46\ \text{\AA}$4 remains approximately constant, and the inferred graphitization barrier is $a = 2.46\ \text{\AA}$5 (Martin et al., 2022). High-resolution transmission electron microscopy identifies these screws as interdigitated fringes offset by $a = 2.46\ \text{\AA}$6, with image contrast disappearing upon defocus changes of only $a = 2.46\ \text{\AA}$7 (Martin et al., 2022).

Irradiated nuclear graphite provides a complementary picture of disorder. In IG-110, NBG-17, and an ASR-0RB-type in-house graphite, $a = 2.46\ \text{\AA}$8 Ar$a = 2.46\ \text{\AA}$9 and He$c = 6.71\ \text{\AA}$0 irradiation at $c = 6.71\ \text{\AA}$1 drives increasing Raman disorder as fluence rises from $c = 6.71\ \text{\AA}$2 to $c = 6.71\ \text{\AA}$3 (Gawęda et al., 2023). The principal disorder metric is the D-to-G intensity ratio $c = 6.71\ \text{\AA}$4, while crystallite size is inferred from G-band broadening using a relation of the form

$c = 6.71\ \text{\AA}$5

with $c = 6.71\ \text{\AA}$6 and $c = 6.71\ \text{\AA}$7 (Gawęda et al., 2023). Ar$c = 6.71\ \text{\AA}$8 causes much stronger degradation than He$c = 6.71\ \text{\AA}$9: at $1.42\ \text{\AA}$0, Ar$1.42\ \text{\AA}$1 already drives the estimated crystallite size down to about $1.42\ \text{\AA}$2–$1.42\ \text{\AA}$3, accompanied by band broadening and disappearance of second-order Raman features above $1.42\ \text{\AA}$4, which the authors interpret as progression toward amorphization (Gawęda et al., 2023).

These studies converge on a common point: graphite’s functional state is controlled less by nominal composition than by the evolution of mesoscale disorder. This suggests that graphitization, irradiation tolerance, and lifetime prediction are all fundamentally defect-engineering problems.

4. Processing routes, composites, and engineered graphite forms

Artificial graphites and graphite composites are described as hierarchical materials composed of graphitic crystallites, filler grains, binder-derived carbon, pores, and intergranular phases, with strength and durability governed by crystallite size, porosity, anisotropy, and boundary structure rather than density alone (Zhmurikov, 2015). Fine-grained graphites such as SGL, LeCL, and MPG-6 exemplify this point: they share broadly similar graphitic lattice parameters but differ sharply in texture, packet thickness, porosity, and intercrystalline transport barriers (Zhmurikov, 2015). At the lifetime level, fracture is treated as a thermofluctuational process with characteristic durability

$1.42\ \text{\AA}$5

and with a strong bias toward microcrack formation at grain boundaries rather than within crystallites (Zhmurikov, 2015).

Processing can also deliberately modify graphite chemistry. Under $1.42\ \text{\AA}$6, adamantane plus ortho-carborane produces highly boron-doped graphite with estimated substitutional boron contents of $1.42\ \text{\AA}$7 at $1.42\ \text{\AA}$8 and $1.42\ \text{\AA}$9 at $3.35\ \text{\AA}$0, inferred from X-ray peak shifts through

$3.35\ \text{\AA}$1

The same study reports coherent scattering region sizes increasing from about $3.35\ \text{\AA}$2 to $3.35\ \text{\AA}$3 as temperature rises, together with XPS evidence for B–C bonding at $3.35\ \text{\AA}$4, consistent with substitutional boron in graphite layers (Bagramov et al., 2021). The authors argue that above about $3.35\ \text{\AA}$5 boron, Coulomb repulsion between negatively charged substitutional boron centers may favor approximately equidistant distribution, though this remains an interpretive inference rather than a direct structural observation (Bagramov et al., 2021).

At a much larger length scale, room-temperature direct ink writing can convert commercial graphite powders into self-supporting three-dimensional graphite architectures by adding a small amount of nano-clay as rheology modifier and binder (Sajadi et al., 2020). The modified ink reaches a viscosity of $3.35\ \text{\AA}$6 at $3.35\ \text{\AA}$7 and $3.35\ \text{\AA}$8 at $3.35\ \text{\AA}$9, with storage modulus $30\%$0 at $30\%$1 strain and $30\%$2, enabling extrusion and shape retention (Sajadi et al., 2020). The resulting printed graphite exhibits thermal conductivity $30\%$3, resistivity $30\%$4, and a compressive plateau beyond $30\%$5 strain, in contrast to bulk graphite collapse near $30\%$6 compressive strain (Sajadi et al., 2020).

Other engineered forms operate at still lower cost. Pencil-drawn graphite composite films on ordinary paper, deposited using a 9B pencil, produce coatings of average thickness $30\%$7 and conductivity $30\%$8, with Raman D, G, and 2D bands at $30\%$9, $sp^2$00, and $sp^2$01, respectively (Morarka et al., 2018). In the $sp^2$02–$sp^2$03 range they attenuate transmitted RF power increasingly with both frequency and number of layers; ten coated sheets yield roughly $sp^2$04 near $sp^2$05 and about $sp^2$06 to $sp^2$07 near $sp^2$08, while a 12-layer roll suppresses nearby relay-induced radio interference in a qualitative AM-radio test (Morarka et al., 2018).

5. High-pressure, transformed, and nanoporous graphite-related phases

Under compression and high temperature, graphite functions as a precursor and a reference state for several distinct carbon transformations. In cold compression, one proposed product is Z-carbon, an orthorhombic $sp^2$09 allotrope with pure $sp^2$10 bonding, lattice parameters $sp^2$11, $sp^2$12, $sp^2$13, and an enthalpy crossing below graphite at about $sp^2$14 (1109.01313). Its proposed formation pathway involves sliding of graphite from $sp^2$15 to $sp^2$16 stacking along $sp^2$17, followed by buckling and three-dimensional bond formation, and the paper identifies it as the best match to unexplained Raman and X-ray features of cold-compressed graphite in the $sp^2$18–$sp^2$19 range (1109.01313).

Under hot compression, graphite-to-diamond conversion is treated as microstructure-controlled rather than homogeneous. Large-scale molecular dynamics and high-resolution transmission electron microscopy indicate that diamond nuclei originate at graphite grain boundaries, then propagate preferentially along both $sp^2$20 and $sp^2$21, with the latter faster; cubic diamond is the kinetically favored product, while hexagonal-diamond-like motifs appear mainly as twins or stacking-fault structures (Zhu et al., 2020). The experimentally confirmed coherent interface

$sp^2$22

encapsulates the central role of graphite’s anisotropy in determining the transformation path (Zhu et al., 2020).

A different line of work asks whether a graphite-like solid can be intrinsically porous. First-principles calculations on stacked graphtriyne sheets predict a layered “nanoporous graphite” with interlayer spacing $sp^2$23, preferred lateral offset around $sp^2$24–$sp^2$25, and continuous subnanometer channels that selectively stabilize $sp^2$26 over $sp^2$27, $sp^2$28, and $sp^2$29 (Bartolomei et al., 2016). In the three-pore channel model, the inner-pore binding energy for $sp^2$30 is $sp^2$31 with adsorption enthalpy $sp^2$32, compared with $sp^2$33 for $sp^2$34, $sp^2$35 for $sp^2$36, and $sp^2$37 for $sp^2$38 (Bartolomei et al., 2016). This remains a theoretical prediction rather than an experimentally realized graphite polymorph.

6. Interfaces, nanoscale distortions, and contested emergent states

Several papers recast graphite as a platform for interfacial or mesoscale electronic phases not expected from ideal Bernal graphite alone. One proposal introduces a graphite-on-graphite geometry in which a vertical Bernal graphite wall on a flat graphitic substrate creates an atomically smooth lateral interface expected to host approximate flat-band states derived from vertical Dirac nodal lines; with electron-electron interaction, these states are argued to favor either spin ordering or superconducting pairing at unusually high temperatures (Volovik et al., 2017). Because the argument relies on idealized nodal-line topology and exceptionally smooth edges, the paper is explicitly programmatic rather than demonstrative (Volovik et al., 2017).

At the nanoscale, patterned highly oriented pyrolytic graphite can develop a measurable bandgap. In square HOPG patterns with minimum size $sp^2$39, ARPES yields a gap of $sp^2$40, while Raman shows a $sp^2$41-band redshift of about $sp^2$42, interpreted as tensile strain (Chaiyachad et al., 2024). One-step photoemission calculations reproduce a comparable gap of $sp^2$43 for a mechanically distorted surface geometry, and supplementary strained trilayer calculations give $sp^2$44 at $sp^2$45 biaxial strain (Chaiyachad et al., 2024). The paper argues against pure quantum confinement, twist as the dominant mechanism, or disorder alone, and instead assigns the gap to mechanically induced lattice distortion during patterning (Chaiyachad et al., 2024).

The most controversial claims concern superconductivity. In highly oriented pyrolytic graphite and commercial flexible graphite gaskets, superconducting-like hysteresis loops are extracted only after subtracting both graphite diamagnetism and an empirically fitted ferromagnetic sigmoidal contribution; the residual loop width is then fit with a Bean-like temperature dependence,

$sp^2$46

yielding an extrapolated $sp^2$47 in HOPG and comparable high-temperature scales in Grafoil and Papyex, together with lower-temperature transitions near $sp^2$48 and $sp^2$49–$sp^2$50 in the gasket materials (Rousset-Zenou et al., 2022). High-temperature resistance measurements on selected HOPG contact configurations show transition-like anomalies near $sp^2$51–$sp^2$52 but no zero-resistance percolation (Rousset-Zenou et al., 2022). The paper interprets these results as hidden granular superconductivity associated with graphite interfaces, stacking faults, or twisted regions, but it also acknowledges that the microscopic origin remains speculative and that the evidence is indirect and analysis-dependent (Rousset-Zenou et al., 2022).

Taken together, these studies portray graphite not as a single, exhausted material class, but as a layered carbon system whose observable behavior depends acutely on stacking, strain, defect topology, thickness, and mesoscale architecture. This suggests that the most active frontiers in graphite research lie less in ideal bulk Bernal graphite than in controlled deviations from it: interfaces, confinement, disorder, graphitization pathways, and multiscale engineered forms.