Metals strengthen with increasing temperature at extreme strain rates

The study investigates how temperature affects the strength of metals under extreme deformation rates where conventional thermally activated mechanisms may give way to dislocation drag. Under normal and moderately high strain rates, metals typically soften with increasing temperature because dislocation motion is thermally activated. However, theory suggests that at very high strain rates (>10⁴ s⁻¹), drag-controlled, ballistic-like dislocation transport can dominate, potentially reversing the temperature dependence. Traditional macroscale high-rate methods often conflate strength with shock effects and rarely provide quantitative strength beyond ~10⁴ s⁻¹. The authors pose the question: Do pure metals exhibit thermal hardening (hotter-is-stronger) at extreme strain rates in the absence of strong shock effects, and can this be quantified via microballistic testing?

Prior work shows ubiquitous thermal softening across more than ten orders of magnitude in strain rate under conditions dominated by thermally activated dislocation motion. At higher rates (beyond ~10⁴ s⁻¹), additional mechanisms (e.g., dislocation drag) are anticipated to contribute substantially to strength, with some models predicting diminished or reversed temperature dependence at extreme rates. Conventional high-rate methods (Kolsky bars, flyer plates, shock impacts, high-power lasers, gas guns, isentropic compression) require high velocities and large samples, often introducing shock and limiting clean measurement of intrinsic strength beyond ~10⁴ s⁻¹. Literature data at ~2×10⁴ s⁻¹ for copper still show thermal softening, suggesting a mechanistic transition occurs at rates higher than those previously explored. Advances in optically driven microballistic systems enable micron-scale impactors at velocities up to ~1,000 m s⁻¹, achieving 10⁶–10⁸ s⁻¹ strain rates without strong shock, and provide two quantitative routes to strength: (1) dynamic yield strength from coefficient of restitution (CoR) scaling for plastic impacts; (2) dynamic hardness from crater volume and deposited energy.

• Experimental approach: Laser-induced microballistic impact tests on metal substrates, enabling strain rates of ~10⁶–10⁸ s⁻¹ without entering strong shock regimes.
• Impactors and targets: Spherical alumina microparticles (≈12.5 ± 1 µm diameter) impacted annealed copper substrates; alumina chosen for much higher hardness (≈15.7 GPa vs ≈0.56 GPa for Cu) to confine plasticity to the substrate. Tests also extended to pure Au and Ti (trends shown in extended data).
• Kinematics measurement: High-speed imaging tracked inbound and rebound particle trajectories to extract impact velocity (v_i) and rebound velocity (v_r). Coefficient of restitution CoR = v_r/v_i used to quantify rebound behavior.
• Temperatures and rates: Experiments conducted at 20 °C, 100 °C, and 177 °C; strain rate for a given test approximated as v_i/d (with d the particle diameter).
• Strength from rebound scaling: For ideally plastic impacts with undeforming impactor on deforming substrate, CoR follows CoR = α (Y_d/E)^{1/2} (with α ≈ 0.78, E effective elastic modulus). Fitting CoR vs v_i across datasets at each temperature yields the dynamic yield strength Y_d (averaged over impact duration and deformation field). Velocity measurement uncertainty ≈ ±2%.
• Hardness from crater measurements: Post-impact craters characterized by 3D laser scanning confocal microscopy to quantify diameter, depth, and volume V. Dynamic hardness H_d computed per test from H_d = 0.5 m_p (v_i^2 − v_r^2)/V, with m_p particle mass. This provides point-wise hardness versus assigned strain rate v_i/d.
• Mechanistic modeling: Total strength decomposed into three additive components at fixed strain rate: (1) thermally activated strength (short-range barriers; softens with T), (2) athermal strength (dislocation–obstacle interactions scaling with elastic modulus; softens with T), and (3) dislocation drag (phonon drag; strengthens with T). Standard models (Supplementary Sections 2.1–2.3) evaluated for copper; apparent activation energy for plasticity Q_app computed via Q_app = R^{-1} (∂ ln ε̇ / ∂ (1/T))_ε̇ to identify sign change indicating mechanistic transition.

• Direct observation of thermal hardening at extreme strain rates: For copper, both rebound trajectories and crater metrics show higher strength at higher temperatures for strain rates ≳10⁵ s⁻¹.
• Quantitative dynamic yield strength (from CoR fits at ~10⁷ s⁻¹): • 20 °C: Y_d ≈ 280 MPa • 100 °C: Y_d ≈ 310 MPa • 177 °C: Y_d ≈ 364 MPa → ≈30% increase over ~157 °C temperature rise.
• Crater metrics: At similar impact velocities, crater depth and diameter decrease with temperature; crater volume decreases by about a factor of two from room temperature to 177 °C, consistent with increased strength (since crater volume ∝ 1/strength).
• Dynamic hardness (per-test analysis): H_d increases with temperature across the tested strain-rate range; for ε̇ > ~10⁵ s⁻¹, copper hardens with increasing T.
• Generality: Similar anomalous temperature dependence observed for pure Au and Ti (extended data), indicating a broader phenomenon across pure metals at extreme rates.
• Mechanistic decomposition: Summed contributions of thermal, athermal, and dislocation drag components predict a net increase in total strength with temperature at these rates; the magnitude and T-dependence align with experimental data.
• Crossover behavior: Comparison with literature at ~2×10⁴ s⁻¹ shows conventional thermal softening at lower rates; modeling of apparent activation energy Q_app indicates a sign change near ~10⁴ s⁻¹, marking transition from thermally activated to phonon-drag-controlled (ballistic-like) dislocation transport.
• Example implication: At ε̇ ~10⁷ s⁻¹ and 177 °C, copper strength >300 MPa, comparable to conventional strength of 304 steel at similar temperature.

The findings confirm that, at extreme strain rates, the temperature dependence of metal strength inverts relative to conventional behavior. The increased CoR and reduced crater volume with rising temperature directly show higher resistance to plastic deformation, which is captured quantitatively by increased dynamic yield strength and hardness. Mechanistic modeling attributes this to dislocation–phonon drag dominating over thermally activated and athermal mechanisms, producing thermal hardening. The work delineates a clear crossover in controlling mechanisms: below ~10⁴ s⁻¹, positive apparent activation energies are consistent with thermally assisted dislocation motion and softening; above ~10⁵ s⁻¹, negative apparent activation energy reflects ballistic-like dislocation transport with increasing phonon drag at higher temperatures. By eliminating strong shock effects via microballistic testing at micron scales, the study isolates intrinsic high-rate strength, resolving ambiguities in prior macroscale impact data. These results refine constitutive understanding and provide benchmarks for high-strain-rate modeling relevant to high-speed manufacturing, erosion, additive processes, and hypervelocity impacts.

Microballistic impact testing enables quantitative, shock-free measurement of strength and hardness at extreme strain rates (10⁶–10⁸ s⁻¹). Using this approach, the study demonstrates that pure metals (Cu, and similarly Au and Ti) exhibit anomalous thermal hardening with increasing temperature at high rates. For copper, dynamic yield strength rises from ~280 MPa at 20 °C to ~364 MPa at 177 °C, corroborated by reduced crater volumes and elevated dynamic hardness. A mechanistic framework combining thermally activated, athermal, and dislocation drag components accurately predicts the observed trend and identifies a mechanistic transition around ~10⁴–10⁵ s⁻¹ where apparent activation energy changes sign. These insights caution against extrapolating low-rate behavior to extreme conditions and suggest new avenues for materials selection and design for high-rate, high-temperature environments. Future research should extend microballistic quantification to broader materials classes (alloys, microstructures), map the transition regime with finer strain-rate resolution, and integrate data to calibrate and validate high-rate constitutive models.

• The primary quantitative demonstrations are on pure copper, with Au and Ti trends shown in extended data; broader generalization to alloys and complex microstructures remains to be established.
• The strain-rate crossover region (~10⁴–10⁵ s⁻¹) is inferred through modeling and stitched datasets; direct measurements densely spanning this interval are limited.
• Measurements are at microscale with spherical alumina impactors and specific particle sizes; scale effects and geometry dependence (e.g., different projectile sizes/shapes) were not exhaustively explored.
• Dynamic yield strength from CoR represents an average over complex deformation during impact; local heterogeneities are averaged and may mask spatial variability.
• Temperature range is modest (20–177 °C); behavior at higher temperatures, near melting, or over wider ranges was not reported here.
• Uncertainties exist in effective modulus selection and assumptions (e.g., ideally plastic scaling) used to fit CoR and compute hardness, though velocity uncertainties are small (~±2%).