Skip to content

Ultrahigh flow strength in shocked nanopolycrystalline diamond

FieldValue
JournalScience Advances
Published2026-09-18
AuthorsAnirudh Hari, Kento Katagiri, Wanghui Li, Dorian P. Luccioni, Rayen Lin
DOI10.1126/sciadv.aed5546
Citations2

Shock-compressed nanopolycrystalline diamond reached a peak flow strength of 107 ± 5 GPa at a shock stress of 227 ± 8 GPa. The experiment recorded this value at a total strain of 0.262 ± 0.006. Femtosecond in situ X-ray diffraction and molecular dynamics associated the peak with stacking fault-mediated strengthening.

Flow strength rose with shock stress below 227 ± 8 GPa and fell above that stress. The measured strength became negligible at 565 ± 14 GPa, where it was 2 ± 6 GPa. Molecular dynamics showed the same trend, with 4 ± 1 GPa at 623 ± 2 GPa.

ConditionShock stressFlow strengthTotal strain
Experimental peak227 ± 8 GPa107 ± 5 GPa0.262 ± 0.006
Experimental softening565 ± 14 GPa2 ± 6 GPa0.452 ± 0.006
MD transition region260 ± 4 GPa87 ± 9 GPa0.290 ± 0.011
MD high stress623 ± 2 GPa4 ± 1 GPa0.564 ± 0.003

The peak was five times the 23 GPa ultimate flow strength measured under quasi-static compression. It also exceeded prior shock measurements on micro-polycrystalline and single-crystalline diamond. The authors report values more than ten times the flow strengths for alumina and silicon carbide under similar conditions.

Earlier single-crystal work at 184 ± 16 GPa found diagonal bands along {111} planes. The authors attributed these bands to stacking faults bound by partial dislocations. That work did not connect the pressure-induced defects to flow strength.

Nanopolycrystalline diamond was synthesized from highly oriented pyrolytic graphite at 15 GPa and between 2,300 and 2,500 °C. The synthesis used a 6,000-tonne multi-anvil press. The starting material had a density of 3.514 ± 0.003 g/cm³ and average grains of 30 to 50 nm.

A 15 μm polypropylene film served as the ablator. The nanosecond drive laser was focused onto a 140 or 260 μm diameter spot. Laser intensity was about 10¹² W/cm², producing multi-megabar shock loading. The arrangement sought a single planar shock front while limiting hard X-ray emission from the ablator.

The X-ray probe used about 10 fs pulses at 10, 11, or 12 keV. Its 50 × 20 μm² spot probed the sample about 1 ns after shock entry. Target tilts were 108°, 110°, or 115° relative to the beam. Pixel-array detectors recorded the diffraction patterns.

Each reflection contained peaks for uncompressed, elastically compressed, and plastically compressed material. Peak positions encoded lattice d-spacings, while azimuthal intensity changes recorded grain orientation. The analysis divided each pattern into 10° azimuthal sectors and fitted each line profile with a PearsonVII peak shape. Longitudinal and transverse lattice strains were optimized from the fitted 2θ values.

NPD had a longitudinal sound speed of about 18 km/s, compared with a bulk sound speed of about 11 km/s. This separation supported an elastic wave followed by a plastic wave at multi-megabar stress.

Flow strength, defined as 2τ, was the differential lattice strain multiplied by a density-functional-theory shear modulus. The authors used a Voigt iso-strain approximation for the modulus. Peak longitudinal stress came from shocked density through a measured US-up relation and Rankine-Hugoniot conservation. This method measures the differential stress sustained during plastic flow.

Large-scale LAMMPS simulations used a quantum-accurate machine-learning potential for diamond. The modeled sample was about 50.5 × 50.5 × 150.8 nm³ and contained up to 66 million atoms. It included about 900 grains with an average diameter of 8 nm.

After equilibration at 300 K for 30 ps, a virtual piston generated planar uniaxial shock loading. The simulations used a 1 fs timestep under the NVE ensemble. Local stresses, temperature, density, and particle velocity were averaged along the shock direction. Dislocation analysis tracked partial and full dislocations, while inverse pole figures quantified texture.

The model used 8 nm grains rather than the experimental 30 to 50 nm grains. It therefore contained a higher grain-boundary fraction. The simulations also followed picoseconds rather than the experiment’s roughly 1 ns observation time.

A weak <110> fiber texture appeared at 425 ± 14 GPa and above. The <110> directions aligned with the shock axis, while transverse orientations remained random. Forward diffraction models and MD-derived diffraction agreed with this texture. Inverse pole figures from MD also showed the same reorientation at 623 GPa.

The authors interpret the texture as evidence of increasing dislocation slip within grains. They compare this response with compressed FCC metals, which form <110> fiber textures through {111}<110> slip. The texture is indirect evidence for dislocation activity, while the simulations resolved the dislocation structures.

Diamond dislocations can glide as full dislocations with Burgers vector a/2<110> on {111} planes. They can split into two a/6<112>{111} partials separated by a stacking fault. Below 260 GPa, both dislocation densities increased, with partial dislocations more abundant.

At 260 ± 4 GPa, MD gave 3.26 × 10¹⁶ m⁻² partials and 3.02 × 10¹⁶ m⁻² full dislocations. At 623 ± 2 GPa, partial density fell to 1.68 × 10¹⁶ m⁻², while full density rose to 1.39 × 10¹⁷ m⁻².

The authors propose that stacking faults shorten the mean free path for dislocations on other planes. This raises resistance to plastic flow and explains the hardening regime. Compression also raises the shear modulus through denser atomic packing, adding strengthening.

Above 260 GPa, rising full-dislocation density enables greater plastic strain and contributes to softening. The shift also produces the observed <110> texture. The authors identify the increasing full-dislocation population as the primary cause of softening above the transition.

Shock temperature was calculated from a carbon equation of state for experiments and directly from transverse atomic velocities in MD. The low-stress regime remained below 2,000 K, limiting thermal softening. At 565 ± 14 GPa, the fitted experimental temperature reached 6,000 ± 150 K.

MD also found increasing amorphization, mainly at grain boundaries. Disordered atoms accounted for 5% at 260 GPa, 9% at 366 GPa, and 30% at 623 GPa. The authors suggest that bond weakening, amorphization, and grain-boundary thickening can all contribute to softening.

The experiments and simulations differ in grain size, time scale, texture onset, and high-stress strength. Texture began at 425 ± 14 GPa in experiments and 623 GPa in MD. MD retained 20 ± 5 GPa at 484 GPa, while experiments had nearly no strength at 565 ± 14 GPa.

The MD also began with a lower density of 3.39 g/cm³ because its smaller grains increased the grain-boundary fraction. Control experiments could not separate grain-boundary and strain-rate contributions. Quantitative Hall-Petch behavior also requires more precise measurements across other grain sizes.

The simulations found no clear nano-twin formation. The authors suggest that nano-twins may form over longer times or after pressure release.

The measured peak defines a strength window near 227 ± 8 GPa rather than a general increase with pressure. Engineers should not transfer the 107 ± 5 GPa result directly to a different shock stress, temperature, strain rate, or grain size. At higher stress, MD showed more full dislocations and disordered grain boundaries as strength fell.

For nano-polycrystalline diamond under shock, the authors identify stacking faults as a pressure-enabled strengthening mechanism absent from their cited quasi-static comparison. They suggest using this mechanism to design ultrahard materials for aerospace and nuclear-energy conditions.

Material selection for those applications should consider the full stress path and the transition from stacking-fault strengthening to dislocation-driven softening. The concrete design constraint is that 107 ± 5 GPa applies to 30 to 50 nm NPD near 227 ± 8 GPa, with rapid softening as shock stress rises.

View Original Abstract

Extreme pressures and temperatures create conditions that allow even hard and brittle materials to flow plastically. Despite extensive research, the limits of flow strength under such conditions remain uncertain, and the mechanisms driving deformation at the relevant stresses are a subject of debate. Using femtosecond in situ x-ray diffraction experiments and large-scale molecular dynamics simulations, we demonstrate that stacking fault-mediated strengthening enables shock-compressed nanopolycrystalline diamond to achieve a peak flow strength of 92 ± 3 GPa at a stress of 212 ± 6 GPa. Our findings show that extreme conditions can unlock ultrahigh strength via a complex array of competing deformation mechanisms and thermodynamic effects.