Shock-induced Melting Research Articles

Speed of sound for understanding metals in extreme environments

Knowing material behavior is crucial for successful design, especially given the growing number of next-generation energy, defense, and manufacturing systems operating in extreme environments. Specific applications for materials in extreme environments include fusion energy, semiconductor manufacturing, metal additive manufacturing, and aerospace. With increased applications, awareness of foundational science for materials in extreme environments is imperative. The speed of sound provides insights into phase boundaries, like shock-induced melting. Thermodynamic integration of the speed of sound enables the deduction of other desirable properties that are difficult to measure accurately, like density, heat capacity, and expansivity. Metrology advancements enable the speed of sound to be measured at extreme conditions up to 15 000 K and 600 GPa. This comprehensive review presents state-of-the-art sound speed metrology while contextualizing it through a historical lens. Detailed discussions on new standards and metrology best practices, including uncertainty reporting, are included. Data availability for condensed matter speed of sound is presented, highlighting significant gaps in the literature. A theoretical section covers empirically based theoretical models like equations of state and CALPHAD models, the growing practice of using molecular dynamics and density functional theory simulations to fill gaps in measured data, and the use of artificial intelligence and machine learning prediction tools. Concluding, we review how a lack of measurement methods leads to gaps in data availability, which leads to data-driven theoretical models having higher uncertainty, thus limiting confidence in optimizing designs via numerical simulation for critical emerging technologies in extreme environments.

Porosity and specific surface area dependence of shock-induced plasticity and melting in open-cell nanoporous Cu

We systematically study the plasticity and melting behavior in shock loading, as well as their dependence on porosity (ϕ) and specific surface area (γ) for nanoporous copper (NPC), by conducting large-scale non-equilibrium molecular dynamics simulations. During shock compression, the plasticity (i.e., dislocation slips) is dominant at lower impact velocities, while melting is governing at higher impact velocities. With increasing ϕ, both the plasticity and melting undergo the transitions from “heterogeneity” to “homogeneity” along the transverse directions. The increase in γ prompts an apparent heat release and gives rise to the transition from local plasticity to uniform solid disordering at lower impact velocities, while accelerates the melting at higher impact velocities, by converting more surface energy into internal energy. Upon impact, shock-induced pores collapse accelerates the consolidation of NPCs and is controlled by two mechanisms, i.e., the shearing ligament, prompted by plasticity, under low-velocity impact, and the internal micro-jetting facilitated by melting under high-velocity impact.

Structure and titanium distribution of feiite characterized using synchrotron single-crystal X-ray diffraction techniques

Abstract A solid solution of the mineral feiite (Fe3TiO5) was recently discovered in a shock-induced melt pocket of the Shergotty martian shergottite. It is particularly interesting for its potential as an indicator of pressure-temperature (P-T) and oxygen fugacity in martian crustal and mantle material. To date, complete crystallographic analysis of feiite has not been conducted, as the mineral was previously analyzed by electron backscatter diffraction on micrometer-size grains (Ma et al. 2021). Here we report a convergent crystal-structure model for feiite based on synchrotron single-crystal X-ray diffraction data collected on three grains of feiite synthesized at 12 GPa and 1200 °C. Feiite adopts the CaFe3O5 structure type (Cmcm, Z = 4), which is composed of two octahedral M1 and M2 sites and one trigonal prismatic M3 site (M = metal) in a ratio of 1:2:1. The three feiite grains with composition Ti0.46–0.60Fe3.54–3.40O5 were best modeled by substituting Ti4+ into only the octahedral M2 site, accounting for 30% of this site. Comparisons of the measured average bond lengths in the coordination polyhedra with the optimized Ti4+–O, Fe2+–O, and Fe3+–O bond lengths suggest that ferrous iron occupies the trigonal M3 site, while iron is mixed valence in the octahedral M1 and M2 sites. The Ti4+ and Fe3+ content constrained by our crystal-chemical analyses suggests that at least ~30% of the available iron must be ferric (i.e., Fe3+/Fetotal = 0.3) for the sample synthesized at 12 GPa and 1200 °C and higher P-T conditions may be needed to form the end-member feiite (Fe32+TiO5).

Shock-induced melting and crystallization in titanium irradiated by ultrashort laser pulse

Modification of titanium microstructure after propagation of a melting shock wave (SW) generated by a femtosecond laser pulse is investigated experimentally and analyzed using hydrodynamic and atomistic simulations. Scanning and transmission electron microscopy with analysis of microdiffraction is used to determine the microstructure of modified subsurface layers of titanium. We found that two layers are modified beneath the surface. A top surface polycrystalline layer of nanoscale grains is formed from shock-molten material via rapid crystallization. In a deeper subsurface layer, where the shock-induced melting changes into plastic deformation due to attenuation of SW, the grain structure of solid is considerably affected, which results in a grain size distribution differing from that in the intact titanium. Molecular dynamics simulation of single-crystal titanium reveals that the SW front continues to melt even after its temperature drops below the melting curve Tm(P). The enormous shear stress of ∼12 GPa generated in a narrow SW front leads to free slip of atomic planes, collapse of the crystal lattice, and formation of a supercooled metastable melt. Such melt crystallizes in an unloading tail of SW. The mechanical melting ceases after drop in the shear stress giving rise to the shock-induced plastic deformation. The last process triggers a long-term rearrangement of atomic structures in solid. The overall depth of modified layers is limited by SW attenuation to the Hugoniot elastic limit and can reach several micrometers. The obtained results reveal the basic physical mechanisms of surface hardening of metals by ultrashort laser pulses.

Real-time (nanoseconds) determination of liquid phase growth during shock-induced melting

Melting of solids is a fundamental natural phenomenon whose pressure dependence has been of interest for nearly a century. However, the temporal evolution of the molten phase under pressure has eluded measurements because of experimental challenges. By using the shock front as a fiducial, we investigated the time-dependent growth of the molten phase in shock-compressed germanium. In situ x-ray diffraction measurements at different times (1 to 6nanoseconds) behind the shock front quantified the real-time growth of the liquid phase at several peak stresses. These results show that the characteristic time for melting in shock-compressed germanium decreases from ~7.2nanoseconds at 35gigapascals to less than 1nanosecond at 42gigapascals. Our melting kinetics results suggest the need to consider heterogeneous nucleation as a mechanism for shock-induced melting and provide an approach to measuring melting kinetics in shock-compressed solids.

Shock-induced melting of [100] lithium fluoride: Sound speed and Hugoniot measurements to 230 GPa

Although [100] lithium flouride (LiF) is the most widely used optical window material in dynamic compression experiments, its high stress ($>100\phantom{\rule{0.16em}{0ex}}\mathrm{GPa}$) shock compression response, including melting, is not well understood. To address this need, we measured wave profiles in plate impact experiments to determine the Hugoniot states and longitudinal sound speeds in [100] LiF crystals shock compressed to 231 GPa. The measured peak states are fitted well by a linear shock velocity--particle velocity relation, providing an accurate determination of the LiF Hugoniot curve to significantly higher stresses than previous experiments. The longitudinal sound speeds show a near linear increase with density compression to 182 GPa. Between 182 GPa and 195 GPa, the sound speed and the longitudinal modulus decrease abruptly, due to shock-induced melting. The increasing sound speeds and moduli at higher stresses suggest that shock compressed LiF is fully liquid at 195 GPa and above, allowing determination of the Gr\"uneisen parameter for liquid LiF. The melt stress determined here differs from that predicted by current multiphase equations of state for LiF. Our results provide important insight into the high stress solid and liquid states of shock compressed LiF and point to the need for an improved multiphase equation of state at high pressures and high temperatures.

Newly detected shock-induced high-pressure phases formed in amphibolite clasts of the suevite breccia (Ries impact crater, Germany): Liebermannite, kokchetavite, and other ultrahigh-pressure phases

Amphibolite clasts in the suevite of the Ries impact crater contain shock-induced melt veins (SMVs) with high-pressure phases such as majoritic garnet, jadeitic clinopyroxene and others. In addition, heat conduction from hot SMVs into adjacent rock portions locally produced further high P–T melt pools. These melts were preferentially generated in rock domains, where the SMVs cross older (‘pre-Ries’) veinlets with analcime or prehnite and larger grains of sericitized plagioclase. Melting of such chemically different local bulk systems (Na-, Ca-, Ca-Na- and K-Na-rich) was facilitated by low solidus temperatures of the original secondary OH-bearing phases. From the resulting shock-induced melts, liebermannite, kokchetavite, jadeite, nonstoichiometric and albitic jadeite, grossular, vuagnatite, lawsonite + coesite, and clinozoisite crystallized during pressure release. Vuagnatite is now proven to be a genuine high-pressure phase. Its ubiquitous distance of 20–35 μm from the hot shock veins suggests a temperature sensitivity typical for an OH-bearing phase. In local Na-rich melts albitic jadeite appears instead of the assemblage jadeite + SiO2. Liebermannite, a dense polymorph of K-feldspar was identified by Raman spectroscopy. After stishovite, liebermannite constitutes the second known high-pressure phase in the Ries that contains silicon exclusively in six-fold coordination. The KAlSi3O8-polymorph kokchetavite was formed in alkali-rich melt glasses. Pressure and temperature values in the range of about 8–11 GPa and ~ 800–1100 °C were estimated from the chemical compositions of locally occurring majoritic garnets (Si = 3.21–3.32 and 3.06–3.10 apfu), respectively, and the presence of fine-grained aggregates of lawsonite and coesite. Generally, the neighboring areas of the veins are characterized by a sequence of variable high-pressure phases documenting strongly falling P–T conditions with increasing distance from the vein. These novel features enlighten the dynamic event during passage of a shock wave.

The Los Angeles martian diabase: Phosphate U-Th-Pb geochronology and mantle source constraints

In situ laser ablation inductively coupled mass spectrometry (LA ICP-MS) is used to determine the U-Th-Pb age of the phosphates ferromerrillite and apatite in the Los Angeles shergottitic meteorite. The initial 207Pb/206Pb was refined by analyzing K-rich diaplectic glass. LA ICP-MS mapping was used to document zones of elevated U and Th content and to establish textural controls on isotope ages. By critically assessing dispersion in the U-Th-Pb dataset due to Pb-diffusion in phosphates during high-temperature shock metamorphism, and as a result of subsequent terrestrial contamination, we obtain a best-estimate U-Pb age of 169 ± 5 Ma anchored at an initial 207Pb/206Pb of 0.98390 ± 0.00018. This is statistically indistinguishable from a joint-isochron age of 179 ± 6 with initial 208Pb/206Pb of 2.5151 ± 0.0028. These results complement previously determined Rb-Sr and Sm-Nd isotope ages and provide independent evidence for LA having crystallized as a medium-grained basic rock from a thick lava flow or high-level intrusion in the late Amazonian at ∼170 Ma. In the context of martian mantle evolution, the initial common-Pb values suggest that Los Angeles originated from a source (µ2 ∼ 3.2) that is similar to enriched members of the shergottite meteorite clan. The U-Th-Pb systematics of both ferromerrillite and apatite were locally affected by diffusive Pb-loss in thin U-enriched marginal domains and more profoundly in shock-induced melt pockets where temperatures briefly exceeded 2000 °C. The results reveal: (1) how precise U-Pb ages can be attained from phosphates; (2) the importance of microtextural contextualization of isotope data; (3) that the timescales of cooling from shock conditions were sufficient to promote local diffusive re-equilibration of Pb over 10 s of microns; and (4) that LA ICP-MS mapping can be used to locate domains with the highest U/Pb and Th/Pb, which increases precision on lower intercept ages and isochron regression lines.

A mesoscale study of micro-spallation of Cu through coarse-grained molecular dynamics modeling

Micro-spallation in metals is a complex dynamic fragmentation process accompanied by shock-induced overheating and melting. However, the damage evolution involved in this process, as well as the underlying mechanism, remain poorly understood. Here, a computationally efficient coarse-grained molecular dynamics (CGMD) method is used to study the micro-spallation of Cu. We demonstrate the capability of this method to reproduce results obtained by the classical molecular dynamics (MD) method in predicting spall damage of solid Cu under shock loading. CGMD simulations, however, give a higher spall strength and a later nucleation time compared with MD simulations, owing to the higher stress required to create collective voids in the former compared with the smaller individual voids in the latter. By contrast, the calculated values (including those of the compressive pressure, strain rate, and spall strength) and the predictions of microstructural evolution during micro-spallation of Cu obtained from CGMD simulations are in good agreement with those from MD simulations. This is attributed to the temperature immediately before spallation being sufficiently high for a strong shock to exist with a dominant effect on spallation, such that the collective motion of voids in CGMD simulations has a negligible effect on spall strength. A dependence of the spall strength on the strain rate of liquid Cu is proposed. This CGMD method allows the investigation of micro-spallation of Cu at the mesoscale.

Shock induced melting of sapphire

The method for calculation the melting curves of crystalline bodies based on the Debye model of heat capacity and the Lindemann melting rule is proposed. Hugoniot shock adiabate, determined in dynamic experiments and thermophysical characteristics of the substance under normal conditions are used as input data. Mathematically, the calculation of the melting curveis reduced to the Cauchy problem for a system of ordinary differential equations. This system is solved numerically by the Runge-Kutta method. Using the proposed method, the melting curves of copper, silver, gold, and sapphire at high pressures are calculated. The results obtained for copper, silver and gold were compared with available calculated and experimental data to validate the method. Experiments on shock compression of transparent sapphire sampleswere performed, using a Mach-type cumulative explosive generators. Investigated pressure range (280-1350 GPa) covered a region of shock-induced melting. The temperature of shock front was registered by fast optical pyrometer together with shock velocity. Particle velocity andpressure were obtained by impedance matching technique. Satisfactory agreement of calculatedand experimental data on temperature of melting behind the shock wave front in sapphire was obtained.

Natural Fe-bearing aluminous bridgmanite in the Katol L6 chondrite

Bridgmanite, the most abundant mineral of the Earth's lower mantle, has been reported in only a few shocked chondritic meteorites; however, the compositions of these instances differ from that expected in the terrestrial bridgmanite. Here, we report the first natural occurrence of Fe-bearing aluminous bridgmanite in shock-induced melt veins within the Katol L6 chondrite with a composition that closely matches those synthesized in high-pressure and temperature experiments over the last three decades. The Katol bridgmanite coexists with majorite and metal-sulfide intergrowths. We found that the natural Fe-bearing aluminous bridgmanite in the Katol L6 chondrite has a significantly higher Fe3+/ΣFe ratio (0.69 ± 0.08) than coexisting majorite (0.37 ± 0.10), which agrees with experimental studies. The Katol bridgmanite is arguably the closest natural analog for the bridgmanite composition expected to be present in the Earth's lower mantle. Textural observations and comparison with laboratory experiments suggest that the Katol bridgmanite formed at pressures of ∼23 to 25 gigapascals directly from the chondritic melt generated by the shock event. Thus, the Katol L6 sample may also serve as a unique analog for crystallization of bridgmanite during the final stages of magma ocean crystallization during Earth's formation.

Apollo 15 regolith breccia provides first natural evidence for olivine incongruent melting

Abstract The Apollo 15 mission returned various samples of regolith breccias, typical lunar rocks lithified by impact events on the Moon’s surface. Here we report our observations on shock features recorded in a section of the Apollo Sample 15299. We observe the presence of ferropericlase crystals confined in a shock-melt pocket and conclude that their formation is related to a shock-induced incongruent melting of olivine. While predicted by experiments, this phenomenon has never been observed in a natural sample. The incongruent melting of olivine provides an important signature of melting under high-pressure conditions and allows for estimating the pressure-temperature (P-T) experienced by the studied sample during the impact event. We infer that the fracture porosity that likely characterized the studied sample prior to the shock event critically affected the P-T path during the shock compression and allowed the studied sample to be subjected to elevated temperature during relatively low shock pressures.

Formation mechanisms of ringwoodite: clues from the Martian meteorite Northwest Africa 8705

Ringwoodite and wadsleyite are the high-pressure polymorphs of olivine, which are common in shocked meteorites. They are the major constituent minerals in the terrestrial mantle. NWA 8705, an olivine-phyric shergottite, was heavily shocked, producing shock-induced melt veins and pockets associated with four occurrences of ringwoodite: (1) the lamellae intergrown with the host olivine adjacent to a shock-induced melt pocket; (2) polycrystalline assemblages preserving the shapes and compositions of the pre-existing olivine within a shock-induced melt vein (60 μm in width); (3) the rod-like grains coexisting with wadsleyite and clinopyroxene within a shock-induced melt vein; (4) the microlite clusters embedded in silicate glass within a very thin shock-induced melt vein (20 μm in width). The first two occurrences of ringwoodite likely formed via solid-state transformation from olivine, supported by their morphological features and homogeneous compositions (Mg# 64–62) similar to the host olivine (Mg# 66–64). The third occurrence of ringwoodite might fractionally crystallize from the shock-induced melt, based on its heterogeneous and more FeO-enriched compositions (Mg# 76–51) than those of the coexisting wadsleyite (Mg# 77–67) and the host olivine (Mg# 66–64) of this meteorite. The coexistence of ringwoodite, wadsleyite, and clinopyroxene suggests a post-shock pressure of 14–16 GPa and a temperature of 1650–1750 °C. The fourth occurrence of ringwoodite with compositional variation (Mg# 72–58) likely crystallized from melt at 16–18 GPa and 1750–1850 °C. The presence of the four occurrences of ringwoodite was probably due to their very fast cooling rates in and/or adjacent to the thin shock-induced melt veins and small pockets. In addition, the higher Fa-contents of the host olivine (Fa35–39) in NWA 8705 than those in ordinary chondrites (Fa16–32) makes the olivine–ringwoodite transformation prolong to a lower pressure.

Dynamic mechanical contact behaviors and sintering mechanism of Al nanoparticles subjected to high-speed impact

In this work, dynamic mechanical contact behaviors and sintering mechanism of Al nanoparticles subjected to high-speed impact have been investigated using molecular dynamic simulations by analyzing radial distribution function (RDF), atomic average displacement, mean square displacement (MSD), the sintering neck radius, shrinkage, nanoparticle temperature and the proportion of amorphous atoms. The results shows that the long-ranged ordered structure of Al nanoparticles gradually transforms into short-range ordered structure and finally amorphous structure during the head-on impact. The atomic average displacement, MSD, radius ratio (x/R) and shrinkage at different initial impact speeds first experience a rapid increase to a maximum, followed by a slow decrease and subsequent a recovery to an almost constant. The displacement of surface atoms is always larger than the internal atoms during the head-on impact. There exists one critical impact velocity (e.g., 2800 m/s for given 4 nm in diameter) beyond which the ordered aluminum atoms completely transform into amorphous structure. The contact force, contact radius and contact stress obtained under a given low impact velocity and small strain are quite consistent with the predictions of the continuous Hertz model, but as the impact velocity and normal displacement further increase, the predictions of Hertz model underestimate these contact parameters. The influence of nanoparticle size on the relationship of contact radius and number of contacting atoms is inappreciable. The nanoparticles undergo elastic deformation, and then plastic deformation, followed by mechanical failure and shock-induced melting process into a whole. This work could strengthen our understanding of the industrial processes with applications in energetic nanomaterials, rocket propellant, metal sintering, etc.

Occurrence of tuite and ahrensite in Zagami and their significance for shock-histories recorded in martian meteorites

AbstractWe report on the discovery of two high-pressure minerals, tuite and ahrensite, located in two small shock-induced melt pockets (SIMP 1 and 2) in the Zagami martian meteorite, coexisting with granular and acicular stishovite and seifertite. Tuite identified in this study has two formation pathways: decomposition of apatite and transformation of merrillite under high-P-T conditions. Chlorine-bearing products, presumably derived from the decomposition of apatite, are concentrated along the grain boundaries of tuite grains. Nanocrystalline ahrensite in the pyroxene clast in SIMP 2 is likely to be a decomposition product of pigeonite under high-P-T conditions by a solid-state transformation mechanism. The pressure and temperature conditions estimated from the high-pressure minerals in the shock-induced melt pockets are ~12–22 GPa and ~1100–1500 °C, respectively, although previous estimates of peak shock pressure are higher. This discrepancy probably represents the shift of kinetic relative to thermodynamic phase boundaries, in particular the comparatively small region that we examine here, rather than a principal disagreement between the peak shock conditions.

Discovery of Elgoresyite, (Mg,Fe)5Si2O9: Implications for Novel Iron-Magnesium Silicates in Rocky Planetary Interiors

As the most abundant material of rocky planets, high-pressure polymorphs of iron- and aluminum-bearing magnesium silicates have long been sought by both observations and experiments. Meanwhile, it was recently revealed that iron oxides form (FeO)m(Fe2O3)n homologous series above ∼10 GPa according to laboratory high-pressure experiments. Here, we report a new high-pressure iron-magnesium silicate, recently approved by the International Mineralogical Association as a new mineral (No. 2020-086) and named elgoresyite, in a shock-induced melt vein of the Suizhou L6 chondrite with a chemistry of (Mg,Fe)5Si2O9. The crystal structure of this new silicate is the same as the iron oxide Fe7O9, strongly suggesting that silicates also form ((Mg,Fe)O)m + n(SiO2)n series that are isostructural to iron oxides via (Mg2+,Fe2+) + Si4+ = 2Fe3+ substitution. To test this hypothesis, the phase relationships of the silicates and iron oxides should be further investigated at higher temperature conditions. Newly found iron-magnesium silicate is a potential constituent mineral in rocky planets with relatively high MgO + FeO content.

Shock-induced melting of two-dimensional Yukawa systems from TH−PH Hugoniot curves

The TH−PH Hugoniot curves of compressional shocks in 2D Yukawa systems are derived from the combination of the Rankine–Hugoniot relation around the shock front and the universal relationship for the temperature in the postshock region. From the equation of state of 2D Yukawa liquids, the equilibrium melting curve for 2D Yukawa systems is derived using the two variables of the temperature T and the pressure P. It is found that the obtained TH−PH Hugoniot curves are intercepted by the equilibrium melting curve, indicating the existence of shock-induced phase transition at these crossing points. To confirm this prediction, molecular dynamical simulations of 2D Yukawa systems of κ=0.75 for the conditions around the crossing point are performed. In the postshock region, the calculated various diagnostics of static structural measures, like the Voronoi diagram, the defect ratio, the probability distribution of the shape factors ξ, the pair correlation function g(r), and the static structure factor S(q), suggest that, for our studied system, the shock-induced melting happens when the compressional speed of the boundary is 0.212a0ωpd<vleft<0.283a0ωpd, the same as the prediction from the crossing point.

Hypervelocity impact tests on a Whipple shield using a flyer plate in the velocity range from 4 km/s to 12 km/s

A Mylar flyer plate, being launched by an electric gun, has been applied to explore the ballistic performance of a 2024 alloy Whipple shield consisting of a 1.0-mm-thick bumper and a 3.2-mm-thick rear wall with a standoff of 70 mm between them. In comparison with a spherical projectile, experimental results show that the polymer plate, having about one-fourth mass of an aluminum sphere, can penetrate the same Whipple at impact velocities of 3 km/s and 5 km/s. The ballistic limit curves (BLCs) of the Whipple are obtained in the velocity range from 3 km/s to 12 km/s using 0.25-mm-thick and 0.5-mm-thick Mylar flyers at normal and/or oblique impacts. Different from spherical projectile, the BLCs obtained by a flyer plate oscillate abruptly with impact velocity. The development of the BLC depends on both the spreading feature of the fragmented bumper sheet and the kinetic energy it carries. In normal impact condition, the BLC turns to go upward at the flyer/bumper shock pressures of 41 GPa, 71 GPa and 121 GPa, exactly corresponding to the occurrence of melting in bumper materials under the interactions among rebounded stress pulses, under the first releasing wave propagating through shock-compressed state, and under direct shock compression, respectively. Among the three melting mechanisms, the shock-induced melting produces the most widely spread debris cloud with the finest fragments.

Effect of the initial temperature on the shock response of Cu50Zr50 bulk metallic glass by molecular dynamics simulation

The effect of initial temperature on the shock response of Cu50Zr50 bulk metallic glass (BMG) is investigated by molecular dynamics simulations using the multiscale shock technique. The shock Hugoniot relationship of Cu50Zr50 BMG shows an obvious temperature dependence. At the Hugoniot elastic limit (HEL), the critical shear stress decreases with increasing initial temperature; this is modeled based on the activation of shear transformation zones. Shock at high temperatures shows a pressure-dependent HEL. The deformation map reveals that there are three deformation regimes under different shock intensities and temperatures: elastic, plastic, and shock-induced melting. The flow stress decreases with increasing initial temperature confirming that high temperatures reduce the shear resistance of the Cu50Zr50 BMG. At high temperatures and pressures, a shock-induced melting regime is identified using a diffusivity-based analysis.

Atomistic study on the anomalous temperature-dependent dynamic tensile strength of ice under shock loading

ABSTRACT Although the compressive strength of ice under both quasi-static [M. Arakawa and N. Maeno, Mechanical strength of polycrystalline ice under uniaxial compression. Cold Reg. Sci. Tech 26 (1997), pp. 215–229.] and dynamic [X. Wu and V. Prakash, Dynamic compressive behavior of ice at cryogenic temperatures. Cold Reg. Sci. Tech 118 (2015), pp. 1–13.] loadings shows an anomalous temperature effect that the compression strength is insensitive to temperature in a specific temperature range below −100oC, it is still unclear whether the anomalous temperature exists for the tensile strength of ice at cryogenic temperatures. In this paper, the temperature-dependent dynamic tensile strength of ice 1 h under shock loading is investigated by molecular dynamics simulations. It is intriguing to see that the dynamic tensile strength of the ice exhibits a similar anomalous temperature effect, i.e. it is almost insensitive to temperature in the range 117 ∼ 163 K, which could be interpreted by the competitive mechanism between shock-induced pulverisation and melting. The evolution of the pentagonal-heptagonal defects and the ductile-to-brittle transformation are also observed with decreasing temperature, leading to the unique dynamic tensile behaviour of ice under shock.