Graphite conducts electricity but is soft and highly anisotropic; very hard carbons are generally brittle. An international team reports an intermediate state—neither an ordinary crystal nor amorphous carbon—that combines, in its tests, high hardness, measurable plasticity and conductivity varying by less than 12 % between 4 and 450 K.
Source: nature.com
In plain terms
Imagine replacing a stack of cards with a lattice of small interwoven bundles: the bundles retain graphite’s local order, but the whole no longer has a single weak direction. According to the mechanism proposed by the authors, some bonds between the sheets break under compression and the bundles rearrange instead of immediately producing a crack. The authors also describe the conduction as nearly isotropic, consistent with the structural texture; however, the study reports no series of direction-resolved electrical measurements that would establish this directly. It is nevertheless neither a ready-to-use component nor evidence of large-scale ductility: the most striking mechanical performance comes from 1 to 2 µm micropillars.
| Parameter | Value |
|---|---|
| Primary source | Nature Communications, article accepted on 10 August 2026 and published online on 22 August 2026; DOI 10.1038/s41467-026-77124-9 |
| Status and integrity | Peer-reviewed and citable Article in Press, before final editing; Chinese and Russian public funding; no competing interests declared |
| Material | PG-2 paracrystalline graphite, made from 99.9 % pure C60 |
| Synthesis | 13 GPa, 900 °C for 1 h; recovered cylinder about 3 mm in diameter |
| Architecture | Distorted graphitic domains about 3 nm across; dominant angle close to 70° with angular dispersion of about 30° |
| Bonding and density | sp2 fraction estimated at 89 % by EELS; density 2.35 g/cm³ |
| Hardness | Nanoindentation: 33.7 ± 1.1 GPa, average over 14 locations; Vickers: 31.6 ± 0.5 GPa under 4.9 N |
| Compression | Representative curve from a 2 µm micropillar: yield strength 17.8 GPa at 12.5 %, peak 22.4 GPa at 22.5 %, plastic strain 12 % before failure; sample size not stated in the main text |
| Electrical transport | 48 160 S/m at room temperature; variation below 12 % between 4 and 450 K |
| Level of evidence | Laboratory synthesis and testing; the article reports no in-service qualification, and no independent replication was identified in the corpus consulted |
Technical explanation
Creating order without recreating a crystal. In the formation scenario proposed by the authors, C60 first forms a three-dimensional polymerized network at around 600 °C under quasi-hydrostatic pressure. At 900 °C, the cages collapse and graphitize locally. The face-centered cubic geometry of the precursor would favor conversion of {111} planes into graphitic (002) planes; two nonparallel {111} planes form an angle of 70.5°, close to the observed texture. In this scenario, hydrostatic pressure prevents the macroscopic alignment that uniaxial compression would impose.
What the structural measurements establish. X-ray diffraction no longer shows sharp peaks from long-range periodicity, but the atomic distribution function retains oscillations out to about 20 Å: order extends beyond the nearest neighbors without becoming long-range crystalline. 4D-STEM probes a diffraction pattern at every position with a 0.5 nm probe; it maps domains about 3 nm across in a texture whose dominant angle is close to 70°, with about 30° dispersion. EELS estimates the sp2 bond fraction and provides a signal consistent with a fraction of covalent interlayer bonds, without directly locating sp3 bridges. Together, these techniques point to a continuous network of distorted graphitic bundles without clear conventional grain boundaries.
A constrained atomic model, not a photograph. The reverse Monte Carlo reconstruction contains 8 919 atoms in a 42.215 Å cubic box. It is constrained by total X-ray scattering, nanodiffraction patterns, microscopy and a force distribution produced by a machine-learning potential. The model reproduces the observed average order: about 5.7 % of the atoms have a configuration very close to graphite, with a root-mean-square deviation (RMSD) below 0.1 Å, while 49.8 % deviate by more than 0.3 Å. The study therefore establishes a configuration compatible with the experimental constraints, without demonstrating that it is unique.
Why hardness does not lead immediately to failure. In the representative curve for the 2 µm micropillar, the elastic regime ends at 12.5 % strain and 17.8 GPa; the material then continues carrying more load up to 22.4 GPa. After compression, microscopy shows weakening of the 70° texture, rearrangement of the bundles and partial amorphization. Interlayer spacing increases from 3.15 Å to 3.50 Å, while EELS indicates more sp2 character. According to the mechanism proposed by the authors, an irreversible conversion of interlayer sp3 bonds into sp2 bonds enables the domains to slide and rearrange. A loading–unloading simulation reproduces a residual strain of about 4 % and an sp2 fraction rising from 90.5 % to 95.0 %; it supports this mechanism without measuring it directly.
Why conductivity changes little with temperature. The graphitic domains form a three-dimensional network rather than an oriented stack. The room-temperature conductivity of 48 160 S/m is well below that of graphite in-plane but above its conduction along the weak axis. Between 4 and 450 K, it varies by less than 12 %. Above 16 K, the authors attribute its slope to two-dimensional weak localization; below that, the behavior becomes semimetallic. The detailed transport mechanism explicitly remains to be resolved.
What the tests do not demonstrate. Nanoindentation measures a local response under a tip, and micropillars introduce size effects; they provide neither standardized fracture toughness, nor the tensile ductility of a macroscopic part, nor fatigue life. The absence of radial cracks even under 49 N indicates unusual local dissipation but does not replace multiscale mechanical qualification. Synthesis also requires 13 GPa, two pressure ramps of about 10 h and a one-hour thermal hold.
Why it worked
The targeted bottleneck was the trade-off between hardness, plastic strain and conduction in a dense carbon. The starting point is not graphite that has simply been compressed. In the scenario proposed by the authors, the geometry of the C60 cages and their crystal supplies local orientations that quasi-hydrostatic pressure transforms without aligning them all. The resulting network thus retains strong covalent bonds and conductive pathways while offering diffuse interfaces that can rearrange.
The quantitative comparison is clear under the tested conditions. The hardness of 33.7 GPa is about two orders of magnitude above the roughly 0.3 GPa cited for crystalline graphite. Unlike the ultrahard sp3 carbons compared in the article, which fail at near-zero plastic strain in micropillar tests, PG-2 reaches 12 % in a 2 µm pillar. This combination clears the bar for overcoming the trade-off under the tested conditions; it does not prove that a bulk part would retain all three properties.
The main gap between the result and an application is therefore scale. The recovered cylinders are about 3 mm in diameter, the strongest mechanical results come from micrometer-scale pillars, and the article reports no standardized test of toughness, fatigue, thermal shock, oxidation or electrical cycling. The authors mention aerospace and robust electronics as possible horizons, not as validated uses.
Proposed causal chain
C60 cages ordered in a face-centered cubic crystal → quasi-hydrostatic compression at 13 GPa → three-dimensional polymerization around 600 °C → cage collapse and local graphitization at 900 °C → statistical inheritance of a dominant angle close to 70° with about 30° dispersion → distorted graphitic domains about 3 nm across, interwoven without clear boundaries → covalent interlayer bonds consistent with an sp3 fraction + three-dimensional sp2 pathways → local hardness of 33.7 ± 1.1 GPa and conductivity of 48 160 S/m → according to the authors’ mechanism, sp3 → sp2 conversion, domain rearrangement and partial amorphization under compression → 12 % plastic strain in the representative curve from a 2 µm micropillar → current need to validate the effect at macroscopic scale and in real environments.
Anecdote
The word “paracrystal” dates back to Rolf Hosemann’s work in 1950. More than seventy years later, the difficulty is still not merely making a solid between a crystal and an amorphous material: its intermediate order must also be demonstrated. Here, the conclusion rests precisely on the convergence of total scattering, 4D-STEM and a constrained atomic model.
Legacy and current data
Structural data, figure source data, a set of simulated configurations and two supplementary videos are announced with the article. The article provides no batch-production data, material yield, reproducibility across presses or cost, however. The current state is therefore a laboratory material characterized along several axes, not an industrial process.
The researcher’s view — open questions
These experiments would extend the study; they are not part of its results.
- Test size. Reproducing the same stress–strain curve on millimeter-scale specimens and then measuring toughness and fatigue with appropriate standards would distinguish intrinsic plasticity from a micropillar size effect.
- Test the mechanism. In situ structural measurements during loading cycles would directly track the sp3 → sp2 conversion, domain angle and possible recovery instead of comparing only before and after states.
- Test process portability. Varying the precursor, hydrostaticity, time and temperature would establish whether the 70° texture is a robust window or a narrow property of C60 at 13 GPa.
- Test extreme use. Cycling from 4 to 450 K, irradiation, oxidation and stable electrical contacts are necessary before linking the measured weak temperature dependence to a robust component.
Sources
Reference verified during the fact-checking audit of 24 August 2026.
- Saisai Wang et al., “Paracrystalline graphite with combined high strength, plasticity and anomalous conductivity,” Nature Communications, accepted on 10 August 2026 and published online on 22 August 2026 — peer-reviewed and citable Article in Press, whose final edited version is still due to replace this version. Funding: National Key R&D Program of China, National Natural Science Foundation of China, Jilin Province programs, China Postdoctoral Science Foundation, Open Project of State Key Laboratory of High Pressure and Superhard Materials and Russian Science Foundation; the authors declare no competing interests. DOI: 10.1038/s41467-026-77124-9
Background references
- Rolf Hosemann, “Röntgeninterferenzen an Stoffen mit flüssigkeitsstatistischen Gitterstörungen,” Zeitschrift für Physik 128, 1–35 (1950) — historical origin of the paracrystalline framework, cited by the study.
Confidence statement
Confidence is high in the synthesis conditions, structural dimensions, hardness, micropillar curves and conductivity reproduced from the peer-reviewed article. Confidence is moderate in the complete atomic mechanism: the reconstruction establishes a compatible configuration without demonstrating its uniqueness, and simulation complements observation. Uncertainty is high regarding translation to a macroscopic or industrial part: the article reports neither standardized qualification nor scale-up data, and no independent replication was identified in the corpus consulted. Uniqueness, DOI, dates, units, formulas, links and limitations were checked; no subject with the same key or central result appears in the Pulse registry.