Ultrafast Reactive Sintering

• Ultrafast reactive sintering is a one-step process that combines rapid phase synthesis with densification on timescales of seconds to minutes under ultrahigh heating rates.
• It leverages coupled thermal, electrical, and chemical effects, such as flux assistance and thermal runaway, to produce tailored microstructures.
• Applications include solid electrolytes, thermoelectrics, and high-entropy ceramics, offering improved performance and significantly reduced processing times.

Ultrafast reactive sintering is a class of processing routes in which synthesis, phase formation, and densification occur simultaneously on ultrashort timescales, typically from seconds to a few minutes, under ultrahigh heating rates and, in some variants, strong electric fields and currents (Luo, 24 Mar 2026). It is distinguished from both conventional sintering, which relies on furnace heating for hours, and non-reactive ultrafast sintering, in which densification is rapid but no new phase is chemically generated during the thermal excursion. Across the current literature, ultrafast reactive sintering encompasses field-coupled modes such as reactive flash sintering and reactive spark plasma sintering, as well as field-decoupled modes such as reactive ultrafast high-temperature sintering and selective laser reaction sintering. The defining feature is not a particular heat source, but the coupling of reaction chemistry to densification and microstructural evolution during an ultrafast thermal cycle (Curcio et al., 2022).

1. Definition, scope, and relation to adjacent processes

Ultrafast reactive sintering refers to one-step ultrafast processing that combines reactive synthesis with densification (Luo, 24 Mar 2026). In this sense, it is a subset of the broader ultrafast sintering landscape, which also includes flash sintering, rapid thermal annealing, ultrafast high-temperature sintering using graphite felt heaters, black-light laser or UV sintering, atmospheric-pressure plasma sintering, induction ultrafast sintering, and ultrafast pressure-assisted sintering. The broader category is unified by ultrahigh heating rates, typically on the order of $10^2\ \mathrm{K/s}$, whereas the reactive subset adds concurrent chemical reaction and phase formation.

A central distinction in the literature is between reactive and non-reactive ultrafast sintering. Reactive sintering proceeds while a chemical reaction creates a new phase or compound at particle contacts or interfaces. By contrast, an ultrafast process may be rapid yet non-reactive if no interfacial reaction layer or compound formation is reported. This distinction is explicit in work on zirconia–copper cermets processed by ultrafast high-temperature sintering: the method is ultrafast, but it is not reactive sintering because no interfacial reaction between Cu and $\mathrm{ZrO_2}$ is reported and the retained phase arises from kinetic control of zirconia’s transformation rather than chemical reaction (Zheng et al., 11 Jun 2026).

A related misconception is that ultrafast reactive sintering is synonymous with flash sintering. The literature does not support that equivalence. Flash sintering is one route, characterized by coupled thermal–electrical runaway under an applied field, but reactive ultrafast synthesis and sintering have also been demonstrated by field-free or field-decoupled routes such as ultrafast high-temperature sintering and selective laser reaction sintering (Luo, 24 Mar 2026). This suggests that ultrafast reactive sintering is better understood as a kinetic regime than as a single apparatus-specific method.

2. Governing mechanisms: heating-rate effects, runaway, and reaction–densification coupling

The mechanistic basis of ultrafast reactive sintering is the coupling between rapid heat delivery, thermally activated transport, and reaction chemistry. In field-assisted systems, Joule heating is often written as

$q = J \cdot E = \sigma(T) E^2,$

with a corresponding lumped energy balance

$$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$

so that runaway occurs when the temperature sensitivity of heating exceeds that of heat dissipation (Luo, 24 Mar 2026). A mathematical treatment of flash sintering formalizes this as incipient thermal runaway and identifies it as a necessary condition for the flash event (Lacey et al., 2015). Under fixed-voltage operation, the Joule source increases with temperature and can produce blow-up; under fixed-current operation, the source decreases with temperature and the system relaxes to a steady state.

Reactive ultrafast sintering adds a second thermal contribution from chemistry. In the Perspective on ultrafast sintering, the total heat generation in a reactive field-assisted route is written as

$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$

where $q_{\mathrm{rxn}}$ is the reaction heat release rate and $\xi$ is reaction extent (Luo, 24 Mar 2026). Exothermic reactions and transient liquid phases can therefore intensify ultrafast densification or even trigger “forced” runaway when phase changes or eutectic events abruptly increase conductivity.

In field-decoupled routes, the decisive variable remains the temperature profile, particularly $dT/dt \sim 10^2\ \mathrm{K/s}$ (Luo, 24 Mar 2026). In glass-derived electrolytes such as $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$ (LAGP), rapid heating shifts crystallization to higher temperature and shorter time, widening the effective interval between $T_g$ and $\mathrm{ZrO_2}$0. This permits viscous-flow densification in the amorphous state before crystallization becomes significant (Curcio et al., 2022). For crystallization itself, the transformed fraction can be described isothermally by the Johnson–Mehl–Avrami–Kolmogorov form

$\mathrm{ZrO_2}$1

Under ultrafast heating, the time spent where $\mathrm{ZrO_2}$2 is large is reduced, delaying nucleation during the densification phase (Curcio et al., 2022).

The general kinetic interpretation advanced across the literature is that ultrafast heating suppresses particle and pore coarsening, may generate non-equilibrium grain boundaries with enhanced diffusivity, and can access premelting-like disordered grain-boundary states at high temperature (Luo, 24 Mar 2026). The standard Arrhenius form

$\mathrm{ZrO_2}$3

remains the basic representation of transport and reaction rates, but the apparent activation behavior in ultrafast regimes may differ from conventional long-dwell sintering because the relevant microstructural state is not the same.

3. Process architectures and characteristic operating windows

Several distinct process architectures now realize ultrafast reactive sintering. Reactive spark plasma sintering forms dense products directly from reactive powder mixtures under pressure and current-assisted heating. In bulk BiCuSeO, a nominal “BiCuSeO” mixture of Bi, Cu, Se, and $\mathrm{ZrO_2}$4 was processed in an SPS system under Ar, at 50 MPa, with a heating rate of $\mathrm{ZrO_2}$5, dwell of 5 minutes, and reactive sintering temperatures of 903 K and 973 K; the one-step route was completed in less than $\mathrm{ZrO_2}$6 hours including loading and cooling (Novitskii et al., 2020). The shrinkage profile showed densification starting at $\mathrm{ZrO_2}$7, about 50% densification by 473 K, and densification effectively finished by $\mathrm{ZrO_2}$8, far below the $\mathrm{ZrO_2}$9 required to densify pre-formed BiCuSeO powder.

Ultrafast high-temperature sintering employs external resistive heating of graphite felt or carbon fabric. In LAGP, two graphite felt strips were Joule heated by a programmable DC supply, with the pellet located in the center of the hot zone under ambient pressure and ambient air. A current of $q = J \cdot E = \sigma(T) E^2,$0 was applied for 180 s total, with a peak temperature of $q = J \cdot E = \sigma(T) E^2,$1 and an effective heating rate $q = J \cdot E = \sigma(T) E^2,$2 (Curcio et al., 2022). In this configuration, no plasma is involved; the pellet is heated primarily by radiation and conduction from the hot carbon.

Flash microwave pressing represents another route, combining resonant-field amplification, the negative temperature coefficient of zirconia resistivity, and mechanical pressure. In the reported zirconia system, a 2.45 GHz WR340 resonant cavity, 2 kW generator, and alumina die compaction geometry produced flash within 10–20 s, with activation from near room temperature, at pressures of 5 MPa and 10 MPa (Manière et al., 2020). The electric field maxima in the applicator were on the order of $q = J \cdot E = \sigma(T) E^2,$3–$q = J \cdot E = \sigma(T) E^2,$4, and thermal runaway began near $q = J \cdot E = \sigma(T) E^2,$5.

Selective laser reaction sintering operates by laser rastering in a reactive gas. A 445 nm CW diode laser scanned powder beds in 100 vol% $q = J \cdot E = \sigma(T) E^2,$6 or 100 vol% $q = J \cdot E = \sigma(T) E^2,$7, converting transition metals and their oxides into carbides and nitrides during sub-second local interaction times (Peters et al., 2022). The process is compatible with powder bed fusion logic: deposit a precursor layer, scan in reactive gas to convert and reaction-bond, then deposit the next layer.

A further route uses direct insertion of green compacts into a preheated furnace. For $q = J \cdot E = \sigma(T) E^2,$8, disks were inserted directly into a muffle furnace at $q = J \cdot E = \sigma(T) E^2,$9, held for minutes, and then air-quenched (Orr et al., 2021). Although this is conventional in equipment, it is ultrafast in the kinetic sense because phase formation peaks within a narrow $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$0-minute window.

4. Reaction pathways and material-specific exemplars

The reactive chemistry in ultrafast sintering varies widely by material class. In oxide solid electrolytes, LAGP exemplifies a glass-to-crystal route in which densification of an amorphous precursor precedes crystallization into a NASICON electrolyte. At $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$1, the compact remained a pressed powder with disjoint particles; at $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$2, necks formed as viscous flow initiated; at $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$3, a fully dense amorphous monolith was obtained; and at $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$4, the specimen became fully crystalline NASICON LAGP, with porosity appearing due to the density change upon crystallization (Curcio et al., 2022). This sequence is significant because it shows that the reaction step is not always simultaneous with densification at every instant; rather, ultrafast heating can temporally decouple them within a single cycle.

In chalcogenide thermoelectrics, reactive SPS of BiCuSeO proceeds directly from elemental and oxide precursors. XRD and EDS of the pre-milled mixture showed only precursor phases, confirming no reaction during the 5 minute planetary milling step, whereas reactive densification during SPS yielded BiCuSeO as the main phase with minor $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$5 and $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$6 in most specimens (Novitskii et al., 2020). The authors noted mechanistic similarity to self-propagating high-temperature synthesis, though direct calorimetry or exotherm measurements were not reported.

In $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$7, the target reaction is

$$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$8

The reported work showed that at $$\frac{dT}{dt} = \frac{\sigma(T)E^2 - Q_{\mathrm{loss}}(T)}{\rho C_p},$$9, the $$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$0 phase fraction increases rapidly and reaches a maximum after about 3.5 minutes, after which longer dwell leads to decomposition toward $$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$1 and $$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$2 (Orr et al., 2021). With 6.5 wt% NaCl added, molten salt appears above 801°C and promotes a dissolution–precipitation pathway that suppresses competing phases during the critical 3–4 minute interval.

In ultra-high-temperature ceramics, selective laser reaction sintering uses gas–solid chemistry such as

$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$3

$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$4

$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$5

and

$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$6

for $$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$7 (Peters et al., 2022). Conversion is reaction-bonding rather than full densification, but it remains an ultrafast reactive route because synthesis and consolidation occur within the laser pass.

In high-entropy borides, reactive SPS from elemental boron and metals leverages a large reactive thermodynamic driving force to form single-phase AlB$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$8-structured solid solutions containing 10–20 mol.% WB$$q_{\mathrm{total}}(T)=\sigma(T)E^2 + q_{\mathrm{rxn}}(T,\xi),$$9 and/or MoB$q_{\mathrm{rxn}}$0 equivalents that could not be stabilized by slower routes using pre-formed binary borides (Qin et al., 2019). This is an important counterexample to equilibrium intuition: the kinetic path itself determines whether W- and Mo-rich monoborides segregate or whether a single high-entropy diboride forms.

5. Microstructure, transport, and property consequences

The microstructural consequence most consistently associated with ultrafast reactive sintering is suppression of coarsening during the short high-temperature excursion. In LAGP, the average grain size was $q_{\mathrm{rxn}}$1 for ultrafast high-temperature sintering versus $q_{\mathrm{rxn}}$2 for conventional annealing, with relative densities of $q_{\mathrm{rxn}}$3 and $q_{\mathrm{rxn}}$4, respectively (Curcio et al., 2022). Conventional slow annealing also produced local compositional inhomogeneity, specifically Al-rich, Ge-poor secondary-phase precipitates attributed to Li loss, whereas the ultrafast cycle showed no detectable secondary phases by STEM-EDX.

These microstructural differences directly affected ion transport. The total ionic conductivity at $q_{\mathrm{rxn}}$5 was $q_{\mathrm{rxn}}$6 for conventionally processed LAGP and $q_{\mathrm{rxn}}$7 for ultrafast-processed LAGP, with similar activation energies, $q_{\mathrm{rxn}}$8 and $q_{\mathrm{rxn}}$9, respectively (Curcio et al., 2022). The similarity in $\xi$0 indicates comparable bulk pathways, while the lower total conductivity in the ultrafast sample derives from larger grain-boundary resistance. Using the Brick-Layer Model,

$\xi$1

the grain-boundary specific conductivity at $\xi$2 was estimated as $\xi$3 for ultrafast LAGP versus $\xi$4 for conventional LAGP.

Additive engineering can reverse this limitation. Under ultrafast high-temperature sintering, 1 wt% $\xi$5 yielded $\xi$6 and $\xi$7, while 1 wt% $\xi$8 yielded $\xi$9, $dT/dt \sim 10^2\ \mathrm{K/s}$0, and $dT/dt \sim 10^2\ \mathrm{K/s}$1 (Curcio et al., 2022). The interpretation given is transient-liquid-assisted wetting and intergranular filling; $dT/dt \sim 10^2\ \mathrm{K/s}$2 also supplies excess lithium to the boundary region.

In thermoelectrics, one-step reactive SPS of BiCuSeO yielded relative densities $dT/dt \sim 10^2\ \mathrm{K/s}$3 of theoretical, whereas a second milling and short SPS step produced $dT/dt \sim 10^2\ \mathrm{K/s}$4 density and the highest power factor above $dT/dt \sim 10^2\ \mathrm{K/s}$5, with $dT/dt \sim 10^2\ \mathrm{K/s}$6 for pristine BiCuSeO (Novitskii et al., 2020). The lattice thermal conductivity remained almost unaffected by synthesis route, with $dT/dt \sim 10^2\ \mathrm{K/s}$7 at room temperature $dT/dt \sim 10^2\ \mathrm{K/s}$8 for all bulks and $dT/dt \sim 10^2\ \mathrm{K/s}$9 at 773 K.

In high-entropy borides, reactive SPS delivered dense single-phase products with relative densities from 95.5% to 99.2% and Vickers hardness values from $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$0 to $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$1 (Qin et al., 2019). Notably, the W/Mo-containing compositions were harder than the W/Mo-free baseline, despite rule-of-mixture expectations based on binary diborides predicting the opposite. The paper attributes this to solid-solution strengthening and electronic effects within the high-entropy AlB$\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$2 lattice.

6. Processing strategies, limitations, and open questions

Several processing strategies recur across the literature. One is deliberate separation of fast and slow phenomena. In LAGP, the recommended approach is to ramp rapidly $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$3 through the $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$4–$\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$5 interval, use only a brief hold near $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$6, and quench rapidly to preserve fine grains and suppress Li loss (Curcio et al., 2022). In $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$7, the essential strategy is strict timing: the optimal window is $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$8 minutes at $\mathrm{Li_{1.5}Al_{0.5}Ge_{1.5}(PO_4)_3}$9, followed by immediate extraction and quenching, because only a few minutes of over-dwell drive decomposition (Orr et al., 2021).

Another common strategy is transient-liquid or flux assistance. Molten NaCl dramatically increased the peak $T_g$0 fraction from about 50% to about 98.5% at the same temperature and similar dwell (Orr et al., 2021). In LAGP, low-melting $T_g$1 and $T_g$2 improved intergranular connectivity, though higher amounts of 5 wt% were detrimental to conductivity and lower amounts $T_g$3 wt% were ineffective (Curcio et al., 2022). This suggests that boundary engineering is often the central materials problem once reaction completeness is achieved.

A third strategy is compensation of reaction-induced volume change. In selective laser reaction sintering, metal precursors expand during conversion whereas oxide precursors contract. Composite metal/oxide mixtures were therefore designed for near-zero net volume change, for example $T_g$4 mol for carbides and $T_g$5 mol (Peters et al., 2022). The rationale was to reduce the constrained eigenstrain associated with $T_g$6, for which a film-stress estimate was written as

$T_g$7

The literature is equally clear about limitations. One-step reactive SPS of BiCuSeO was reproducibly limited to densities below 90% of theoretical, probably because of inhomogeneous phase formation and/or Se evaporation (Novitskii et al., 2020). Flash microwave pressing of zirconia showed persistent central hotspots and weakly sintered edges, even after adding boundary insulation, indicating unresolved thermal-uniformity problems (Manière et al., 2020). Single-component laser reaction sintering frequently produced cracking because gas–solid reactivity induced large volumetric changes (Peters et al., 2022). In $T_g$8, post-sinter water rinsing to remove NaCl left pitting and cavities, raising practical concerns for target integrity (Orr et al., 2021).

Several open scientific questions remain unresolved. The 2026 Perspective identifies the relative roles of suppressed coarsening, non-equilibrium grain boundaries, and premelting-like grain-boundary disorder in ultrafast densification as a major unresolved issue (Luo, 24 Mar 2026). It also emphasizes the need to separate intrinsic heating-rate effects from electric-field-induced effects such as defect generation, electromigration, and reaction acceleration. For reactive systems specifically, the interaction between reaction exotherms, conductivity changes, and thermal runaway remains a central modeling problem. This suggests that ultrafast reactive sintering is now experimentally broad but still mechanistically incomplete.

7. Industrial and research significance

The technological significance of ultrafast reactive sintering lies in time compression, localized heat input, and compatibility with scalable hardware. Ultrafast high-temperature sintering uses inexpensive carbon felts and DC supplies, operates in ambient air in the LAGP example, and is conducive to continuous processing such as conveyor-fed felt or roll-to-roll layouts (Curcio et al., 2022). Reactive SPS of BiCuSeO reduced total processing time from 48–72 h of powder synthesis plus consolidation in the conventional route to $T_g$9 h for one-step RSPS and $\mathrm{ZrO_2}$00 h for the two-step RSPS+SPS route (Novitskii et al., 2020). The Perspective generalizes this advantage to high-throughput discovery of high-entropy and compositionally complex ceramics, where vast compositional spaces must be explored rapidly (Luo, 24 Mar 2026).

The field also has direct relevance to energy materials and extreme-environment ceramics. In solid electrolytes, ultrafast reactive sintering offers a route to phase-pure NASICON-type conductors while limiting alkali volatility (Curcio et al., 2022). In thermoelectrics, reactive SPS demonstrates that a direct bulk synthesis route can deliver $\mathrm{ZrO_2}$01 at 773 K equal or slightly superior to conventional pristine BiCuSeO while eliminating lengthy precursor synthesis (Novitskii et al., 2020). In ultra-high-temperature ceramics, selective laser reaction sintering offers a single-step route to carbides and nitrides that are otherwise difficult to realize by direct laser sintering or melting (Peters et al., 2022). In high-entropy borides, reactive SPS shows that kinetically selected pathways can stabilize compositions that slower routes cannot (Qin et al., 2019).

A final point of significance is conceptual. The literature increasingly treats ultrafast reactive sintering not as an isolated curiosity, but as a general method for exploiting kinetic route dependence. High heating rates, short peak-temperature dwells, transient liquids, reactive exotherms, and rapid quenching can together access microstructures and phase assemblages that conventional equilibrium-guided processing does not reach (Luo, 24 Mar 2026). A plausible implication is that future work will rely less on a single “best” sintering method and more on designing temperature–time–reaction trajectories that deliberately favor desired phase-formation and densification pathways on ultrashort timescales.