Subsurface Driven Size-Dependent Elasticity in Nanodiamond
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Physical Review X |
| Published | 2026-03-16 |
| Authors | Jiaqi Zhang, Chunmeng Liu, Xing Li, Dongyang Wang, Keke Zhao |
| DOI | 10.1103/b3h5-34wt |
| Citations | 1 |
Nanodiamond softened as particle diameter decreased across the studied 4 to 13 nm range. Its effective axial Young’s modulus fell from about 1000 GPa near 12 nm to about 700 GPa near 4 nm. Experiments on about 100 particles and atomistic calculations linked this reduction to a weakened surface-core interface rather than soft outermost layers.
Size-dependent axial stiffness
Section titled “Size-dependent axial stiffness”Particles near the large end of the size range approached the cited bulk diamond modulus of about 1140 GPa. FIB-processed particles around 200 nm had been reported at about 865 GPa. Nanocrystalline films below 150 nm had reached values close to single-crystal diamond under optimized growth.
Those comparisons place the observed softening below the bulk-like regime reported in much of the prior diamond literature. An alternative sphere-flattening analysis produced the same qualitative trend. No systematic modulus offset appeared for indexed [100], [110], and [111] zone axes within experimental scatter.
In situ TEM nanomechanical measurement
Section titled “In situ TEM nanomechanical measurement”Individual particles were clamped between two diamond indenters. One indenter was attached to a quartz length-extension resonator, and the other was fixed to the holder.
Detonation-synthesized particles with diameters below 10 nm came from one batch. They were cleaned in aqua regia for at least 24 hours and dispersed on an indenter. The samples were baked at about 100 °C for at least 24 hours under vacuum before testing.
Measurements occurred at room temperature, about 300 K, in a JEM-2000VF TEM at 200 kV and 10⁻⁷ Pa. The 10 μm and 100 μm tungsten tips carried the diamond indenters. STEM imaging at 300 kV and EELS confirmed crystalline diamond. No detectable sp² signal appeared at 283 to 285 eV, while the sp³-related band spanned 286 to 336 eV. XRD and Raman measurements also supported high crystalline quality.
The quartz resonator oscillated near 1 MHz with a constant amplitude of 15 pm. Frequency modulation tracked sample stiffness through resonance shifts. The complex response separated conservative stiffness from dissipative stiffness. Energy loss tracked irreversible deformation.
Selected measurements at 30 pm gave consistent elastic ranges, dissipation onset, and modulus within scatter. The authors also compared beam-blanked cycles with 80 kV imaging to assess beam-dose and time effects. Images recorded every 0.2 seconds supplied the particle shortening.
Modulus used E = kl/A, with l equal to initial height and A derived from multiple transverse chords. The effective area was calculated as A = π(Deff/2)². Only data before dissipation onset entered the fit.
Direct TEM shortening excluded elastic indentation of the diamond tips from the displacement axis. The result was therefore an effective axial modulus under contact-constrained compression, not a free uniaxial constant.
Early stiffness followed Hertzian sphere-flat scaling and approached zero with vanishing indentation. The data did not require a JKR-type finite-stiffness offset at zero load.
Elastic limit and plastic onset
Section titled “Elastic limit and plastic onset”A representative 7.3 nm particle was compressed along [111] to 6.2 nm, about 15% strain. At that point, dissipation rose to about 0.05 eV per oscillation cycle and then plateaued.
The authors defined plasticity as permanent residual strain after full unloading. Repeat load-unload cycles confirmed the irreversible change. After unloading, the particle retained a lateral size of about 7.2 nm without visible cracking.
The authors attribute the response to atomic rearrangement and local amorphization rather than dislocation-mediated plasticity. Similar dissipation-free elasticity and a reproducible dissipation onset appeared across the dataset. The particle therefore accommodated large elastic strain before an atomistic disordering transition.
Core-shell fit and parameter limits
Section titled “Core-shell fit and parameter limits”The authors represented each particle as a bulk-like core and an outer shell with a different modulus. The modulus was fitted as an area-weighted average of core and shell values. The fit used one size-independent shell thickness for all particles.
The fit gave a core modulus of 1185.8 ± 41.6 GPa and a surface modulus of 341.4 ± 557.7 GPa. The shell thickness was 0.75 ± 0.63 nm. The uncertainties are 1σ values from the fit covariance matrix.
The core estimate agreed with the cited bulk value of about 1140 GPa. The large surface uncertainty reflects strong coupling between the surface modulus and shell thickness. Refitting with shell thickness fixed between 0.6 and 1.0 nm changed the surface modulus along the correlated direction. Fit quality and the core modulus changed little.
The model is a geometry-averaged description, not a strict Voigt isostrain model. Polar contact regions are closer to isostress, while equatorial regions are closer to isostrain. The surface modulus also applies to the batch-specific surface state because baking could only reduce terminal groups.
DFT identifies the weak interface
Section titled “DFT identifies the weak interface”CASTEP calculations modeled (111), (110), and (100) diamond surfaces. They used PBE-GGA, ultrasoft pseudopotentials, a 653 eV cutoff, and 0.04 Å⁻¹ Monkhorst-Pack spacing.
The outermost surface bonds were shortest at about 1.49 Å, compared with about 1.55 Å in the core. Surface-core bonds were longer: 1.64 Å for (111), 1.59 Å for (110), and 1.57 Å for (100).
Charge density was higher within the outer surface and lower at the interface. These results distinguish strong bonding inside the outermost layer from weak bonding across the interface. The authors therefore assign the softening to the interface, not to intrinsically soft surface bonds.
Hydrogen termination retained the bond-length and charge-density contrast. Oxygen termination substantially reduced that contrast. The calculations suggest that termination can alter the mechanical effect of the interface.
Molecular dynamics and bond-level strain
Section titled “Molecular dynamics and bond-level strain”LAMMPS simulations compressed spherical particles derived from pristine bulk single crystals. The modeled diameters ranged from below 1 nm to 12 nm. Tersoff interactions were integrated at 300 K with a 1 fs time step. Each particle was equilibrated for 100 ps in the NVT ensemble.
Compression used a strain rate of 10⁷ s⁻¹. Tests at 10⁶ s⁻¹ produced the same modulus and size dependence within numerical uncertainty. The simulated core modulus was about 1050 GPa, while the surface-region modulus was about 400 GPa. The overall modulus increased with particle size and agreed with the experimental trend.
Below 2 nm, the core modulus rose slightly, while the surface-region modulus fell marginally. Surface atoms often had coordination numbers of 2 or 3, compared with 4 in the core.
Connecting bonds accumulated strain faster than either core bonds or surface bonds. Surface bonds carried the lowest strain throughout deformation. This bond-level pattern localizes plasticlike rearrangement at the interface.
An independent LCBOP potential reproduced the qualitative softening trend. Surface classifications used depth, coordination, and tetrahedral-order descriptors. Changing classification thresholds did not change the main conclusions.
Interfacial dissipation and delayed fracture
Section titled “Interfacial dissipation and delayed fracture”The authors interpret the weak interface through Griffith and Orowan criteria. Classical crack propagation occurs when the strain energy release rate G exceeds 2γs. The modified condition is G ≥ 2(γs + γp), where γp represents energy dissipated by plastic deformation.
At 4 to 13 nm, a sharp continuum crack tip is not strictly applicable. The authors use a crack-tip-like atomistic front for incipient interfacial decohesion and localization.
Observed sliding, bond-length adjustments, and local relaxation support this interpretation. These processes dissipate energy and impede cleavagelike crack advance. The observed plasticity and energy rise beyond the elastic limit provide evidence for an interfacial plastic dissipation zone.
Engineering implications and limits
Section titled “Engineering implications and limits”The results support using particle size and surface state as design levers for nanoscale diamond devices. Reducing diameter lowers effective axial stiffness and can shift resonator frequency or acoustic impedance. Surface termination may alter interfacial bond contrast, although the paper did not test device-scale frequency tuning.
Operating below the dissipation onset should help maximize resonator Q and suppress loss. Interfacial phonon scattering may also make termination relevant to thermal transport devices. These are design implications, not demonstrated device results.
The measured modulus applies to contact-constrained compression of detonation nanodiamond near room temperature. The surface parameter reflects the tested batch after aqua-regia cleaning and vacuum baking. It should not be treated as a universal property of every diamond surface.
The experiment covered about 100 particles, with diameters mainly within 4 to 13 nm. The core-shell model cannot resolve a unique surface modulus because that value and shell thickness are correlated. The simulations began with pristine bulk crystals, whereas the experimental particles were detonation-synthesized.
Temperature dependence was not measured. The authors expect the qualitative trend to persist under moderate heating, with possible changes from surface reconstruction or diffusion. Absolute modulus may shift through anharmonic lattice softening.
View Original Abstract
Diamond, an archetype of hardness and brittleness, can nonetheless display striking elastic compliance at the nanoscale, in sharp contrast to its ultrastiff, fracture prone bulk behavior. Despite recent experimental progress, the atomic scale origin of this anomalous elasticity remains unresolved. Here, we uncover a pronounced size- dependent softening in nanodiamond, with the effective axial Young’s modulus decreasing from about 1000 to about 700 GPa as the particle diameter shrinks from 13 to 4 nm. By integrating transmission electron microscopy nanomechanical testing with first principles calculations and molecular dynamics simulations, we show that this stiffness reduction does not arise from the surface atomic layers, which remain comparatively stiff, but instead from weakened bonding in a distinct interfacial region between the surface and the crystalline core. These under-coordinated regions localize strain and accommodate large elastic deformation. The results uncover a previously overlooked interfacial mechanism that governs elasticity in brittle nanomaterials and offer design principles for nanomechanical resonators, quantum devices, and thermal transport applications with tunable stiffness.
Single Crystal Cvd Diamond Plate
View Specs →
Boron Doped Single Crystal Diamond Plate
View Specs →