Rafture: Fracture Mechanics & Consensus

• Rafture is a term describing instabilities in cohesive floating aggregates at fluid interfaces, manifesting as rupture, fracture, or cavitation.
• It encapsulates a range of behaviors where capillarity, elasticity, viscous drag, and geometry compete to set crack patterns and scaling laws.
• In distributed systems, Rafture refers to an erasure-coded variant of Raft that prunes post-dissemination to optimize storage while ensuring recovery.

Searching arXiv for the provided topic and related papers. Searching arXiv for "Rafture raft rupture particle raft fracture". Rafture is a polysemous term in the arXiv literature. In soft-matter and interfacial mechanics, it most commonly denotes the rupture, fracture, or cavitation of a floating raft: a particulate, droplet, or colloidal monolayer at an air–liquid or liquid–liquid interface that acquires effective elasticity, jams, or localizes strain and then fails under forcing. A distinct 2026 distributed-systems usage names an erasure-coded variant of the Raft consensus protocol with post-dissemination pruning. In the mechanical literature, the common structure is a two-dimensional aggregate whose failure morphology is selected by the competition between forcing, capillarity, elasticity, viscous drag, disorder, and geometry [(Bandi et al., 2010); (Tô et al., 2022); (Xiao et al., 2020); (Saddier et al., 2023); (Yan et al., 19 Feb 2026); (Illing et al., 2023); (Iqbal et al., 16 Mar 2026); (Kerur et al., 25 Mar 2026)].

1. Scope of the term and recurrent physical motifs

In the interfacial-physics sense, a raft is a floating aggregate of particles or droplets that is cohesive because of capillary attraction, depletion attraction, or interfacial confinement. Depending on the system, the raft can be a tenuous monolayer of hydrophobic particles at an air–water interface, a dense disordered particle pillar at an air–oil interface, a compressed array of adhesive droplets, a graphite powder membrane driven by gravity waves, or a densely packed granular monolayer on a vertically vibrated interface [(Bandi et al., 2010); (Xiao et al., 2020); (Illing et al., 2023); (Saddier et al., 2023); (Yan et al., 19 Feb 2026)].

Across these systems, failure does not reduce to a single crack criterion. In shock-driven rafts, failure follows passive advection into a jammed annulus and then periodic fracture. In stretched rafts, failure can take the form of diffuse rifting whose characteristic spacing is selected by a competition between capillary–viscous healing and imposed differential motion. In quasi-static tensile loading, strain localizes into a shear band, with ductile or brittle outcomes set by the capillary interaction range. In compression of adhesive droplet rafts, fracture appears as abrupt depletion-bond breaking and load drops. Under waves or vibration, the operative modes include progressive cracking, fragmentation, and cavity opening. Under droplet impact, splash, cavity collapse, Worthington jets, and particle ejection become central observables (Tô et al., 2022, Xiao et al., 2020, Illing et al., 2023, Saddier et al., 2023, Yan et al., 19 Feb 2026, Iqbal et al., 16 Mar 2026).

Taken together, these studies suggest that “rafture” is best understood not as a single constitutive law, but as a family of instability-and-failure processes in quasi-two-dimensional interfacial matter. The recurrent control parameters are packing fraction, particle or droplet size, interaction range, wettability, forcing rate, and the degree to which the raft behaves as a compliant sheet, a jammed granular solid, or a rate-dependent dissipative assembly.

2. Shock-driven jamming and periodic fracture

The canonical formulation of rafture in the narrow sense appears in “Shock driven jamming and periodic fracture of particulate rafts” (Bandi et al., 2010). There, a localized droplet of surfactant creates a rapidly spreading surfactant film whose Marangoni stress drives a radially divergent shock front. The shock advects initially non-inertial particles outward, and for a deep subphase the shock radius obeys

$$R_s = K\left(\frac{\Delta\gamma^2}{\mu\,\rho}\right)^{1/4}t^{3/4},$$

with $K \approx 0.665$–$1.52$ and $\approx 0.94$ in the reported experiments. The compaction band’s leading edge $R_L$, trailing edge $R_T$, and width $W$ all grow as $t^{3/4}$, confirming passive advection by the Marangoni-driven flow rather than autonomous particulate dynamics (Bandi et al., 2010).

As particles are swept up, the local packing fraction $\phi(r,t)$ rises until the raft becomes a jammed, disordered elastic annulus with peak packing fraction $\phi^*$ slightly below the two-dimensional random close packing density $K \approx 0.665$0. Jamming and fracture occur almost simultaneously at time $K \approx 0.665$1: once $K \approx 0.665$2 reaches $K \approx 0.665$3, the outward-driven annulus fractures into nearly regular triangular cracks. The crack faces make a narrow distribution of angles $K \approx 0.665$4–$K \approx 0.665$5 with the local radial direction, the pattern grows self-similarly with the band, and after $K \approx 0.665$6 the crack number $K \approx 0.665$7 is fixed, with cracks deepening and widening but not nucleating anew (Bandi et al., 2010).

A simple geometric theory combines annular elasticity with mass conservation. Because the stress-relief scale is comparable to the band width, the circumferential crack spacing is $K \approx 0.665$8, giving

$K \approx 0.665$9

Mass conservation between the initially swept area and the jammed annulus yields

$1.52$0

With $1.52$1 and $1.52$2, this produces the central scaling

$1.52$3

Experimentally, $1.52$4 and $1.52$5 vary monotonically with the initial packing fraction $1.52$6, whereas $1.52$7 is independent of $1.52$8 and scales linearly with particle diameter, with $1.52$9 across $\approx 0.94$0 and multiple particle species (Bandi et al., 2010).

This framework also defines the limits of the periodic-crack regime. If $\approx 0.94$1 is too low, the continuum picture breaks down because the crack scale approaches a few grain diameters. If $\approx 0.94$2 is too high, near $\approx 0.94$3, the raft is pre-jammed and fracture proceeds via local rearrangements, branching, and kinking rather than periodic triangular cracking. The paper therefore isolates a specific mechanism: surfactant-driven advection first creates a jammed elastic solid out of a dilute monolayer and then fractures that solid almost immediately (Bandi et al., 2010).

3. Rate-dependent tensile and compressive rafture

A second major usage concerns rafts driven quasistatically or kinematically in tension or compression. In “Rifts in Rafts,” a floating raft of spherical polyethylene particles is stretched uniaxially by an expanding air–liquid interface. The failure morphology changes continuously with pulling velocity and is governed by a competition between a capillary–viscous re-aggregation speed and the differential speed imposed by the expanding fluid. In the one-dimensional continuum model,

$\approx 0.94$4

with growth rate $\approx 0.94$5 and healing speed

$\approx 0.94$6

Equating this with the per-neighbor differential speed $\approx 0.94$7 selects the cluster scale:

$\approx 0.94$8

Low pulling speed produces large clusters and few rifts; high speed produces clusters with widths approaching a single particle, with near-zero Poisson’s ratio and little change in the transverse direction (Tô et al., 2022).

In “Strain localization and failure of disordered particle rafts with tunable ductility during tensile deformation,” the loading is quasi-static uniaxial tension on dense, disordered pillars of floating Styrofoam spheres. Here the central control parameter is the capillary interaction range in units of particle size, $\approx 0.94$9, with

$R_L$0

Smaller particles have larger $R_L$1 and therefore longer-range attractions and greater ductility. After an initial elastic regime, strain localizes into a system-spanning shear band at $R_L$2 for $R_L$3 mm particles, $R_L$4 for $R_L$5 mm particles, and $R_L$6 for $R_L$7 mm particles. The band angle is approximately $R_L$8–$R_L$9, not the $R_T$0 often associated with metallic glasses or polymers. Local kinematics are quantified by the deviatoric strain-rate magnitude

$R_T$1

and rearrangements by $R_T$2, while structure is characterized by the Voronoi-based scalar $R_T$3 and the local anisotropy measure

$R_T$4

High-$R_T$5 sites rearrange more strongly, and rearrangements increase $R_T$6 on average, producing a positive feedback that focuses deformation into the failure band. Long-range interactions permit large plastic deformation with minimal void growth; short-range interactions confine rearrangements, accelerate void growth, and produce abrupt brittle rafture (Xiao et al., 2020).

A third variant is compressive rafture in adhesive droplet arrays. “Compression and fracture of ordered and disordered droplet rafts” models a two-dimensional raft of adhesive droplets with short-range Asakura–Oosawa depletion attraction and capillary repulsion in an overdamped Durian-type framework. In ordered monodisperse arrays, wall force rises elastically, reaches a maximum, and then drops abruptly when depletion bonds break along a fracture. For a monodisperse $R_T$7 crystal, the equivalent stiffness is

$R_T$8

and the clean row-reduction peak force scales as

$R_T$9

With increasing bidispersity, sharp fracture events are replaced by multiple smaller avalanches, and the excess number of fracture peaks obeys

$W$0

Large or wide arrays can exhibit competing fractures nucleated nearly simultaneously at distinct locations; slight wall tilt strongly suppresses this effect (Illing et al., 2023).

These tensile and compressive studies collectively show that rafture can be coherent or diffuse, ductile or brittle, and single-event or avalanche-like. The selected mode depends on whether structure can heal or rearrange on the forcing timescale, whether interactions are long- or short-ranged, and whether order or disorder concentrates or redistributes stress.

4. Wave-, vibration-, and impact-driven rafture

Wave forcing generates another family of rafture phenomena. In “Breaking of a floating particle raft by water waves,” graphite powder forms a thin, capillary-cohesive raft on water and is driven by gravity waves with wavelength about $W$1 cm, much larger than the raft thickness of order $W$2 $W$3m. The raft develops oblique cracks at $W$4, opening fractures at the upstream edge, and progressively smaller fragments over tens to hundreds of seconds. The area distribution of fragments resembles floe-size statistics used for sea ice, and for rescaled area $W$5 the late-time probability density obeys $W$6 with $W$7 over an intermediate range. The mean fragment area decays as $W$8 in the late regime, while the circularity saturates near $W$9 (Saddier et al., 2023).

A key result of that wave study is mechanistic: bending stresses are far too small to explain fracture, whereas viscous shear beneath the raft is sufficient. The axial in-plane stress scales as

$t^{3/4}$0

for progressive waves and

$t^{3/4}$1

for standing waves. Experimental threshold diagrams align with viscous-stress iso-curves rather than bending-curvature iso-curves, with inferred breaking thresholds of approximately $t^{3/4}$2 Pa in one tank and approximately $t^{3/4}$3 Pa with sliding sidewalls (Saddier et al., 2023).

Vertical vibration produces a different route to failure. In “Densely-packed particle raft at vertically vibrated air-water interface,” a monolayer of buoyant Styrofoam particles exhibits subharmonic Faraday-like standing waves over part of the $t^{3/4}$4 space, but at higher packing fractions the measured dispersion relation requires an emergent elastic-raft correction:

$t^{3/4}$5

Here $t^{3/4}$6 decreases monotonically with increasing $t^{3/4}$7, whereas $t^{3/4}$8 rises rapidly near $t^{3/4}$9. At high frequency and low amplitude, the raft enters a no-wave regime with thermal-like particle motion; the MSD scales as $\phi(r,t)$0, with $\phi(r,t)$1 at $\phi(r,t)$2 and $\phi(r,t)$3 at $\phi(r,t)$4. Increasing amplitude from this supercooled-like state nucleates a cavity inside the raft. The cavity size obeys the empirical scaling

$\phi(r,t)$5

Inside the cavity, free-surface water waves persist; outside, the raft remains glassy. The paper explicitly interprets this coexistence as a rupture- or cavitation-like state (Yan et al., 19 Feb 2026).

Droplet impact introduces yet another operational definition of rafture. In “Droplet Impact on Microparticle Raft,” millimetric impacts onto closely packed microparticle monolayers alter lamella spreading, rim destabilization, cavity collapse, Worthington jet formation, and particle ejection. Splash onset is organized by a roughness-versus-burial criterion: grains destabilize the rim if the protrusion height $\phi(r,t)$6 is comparable to or larger than the lamella thickness $\phi(r,t)$7. Experiments collapse splash threshold data using

$\phi(r,t)$8

and jet-height data using

$\phi(r,t)$9

Two regimes emerge:

$\phi^*$0

$\phi^*$1

Superhydrophobic rafts enable prolific impact-driven and jet-mediated particle ejection, including particle-armoured Worthington jets that fragment into liquid marbles, whereas deeply immersed particles stabilize the lamella and suppress detachment (Iqbal et al., 16 Mar 2026).

These wave-, vibration-, and impact-driven studies show that rafture need not be a static fracture process. It can be progressive fragmentation, cavitation, or jet-mediated disassembly, with the governing balance set by viscous shear, emergent bending elasticity, or inertial–capillary impact dynamics.

5. Relation to raft formation and membrane-raft literature

A persistent misconception is that all “raft” studies concern rupture. Several papers in the supplied corpus instead analyze raft formation, stability, and interaction without fracture. Their relevance to rafture is largely taxonomic and conceptual: they define what a raft is, how finite raft size is selected, and how interfacial elasticity or chirality can stabilize domains before any failure process begins.

“Chiral twist drives raft formation and organization in membranes composed of rod-like particles” studies colloidal monolayer membranes of bidisperse filamentous viruses. There, chiral liquid-crystal twist stabilizes finite-sized rafts and mediates long-range repulsion between them. The effective line tension is renormalized to

$\phi^*$2

and finite raft size follows when chirality-driven line-tension reduction is balanced by higher-order elastic costs. The same work gives a twist-mediated repulsion $\phi^*$3 and frames finite-sized rafts as a consequence of chirality, depletion, and orientational elasticity rather than breakup (Kang et al., 2016).

“Conformational switching of chiral colloidal rafts regulates raft-raft attractions and repulsions” extends this picture by showing that individual colloidal rafts in an achiral background can switch between right-handed and metastable left-handed states. Same-handed pairs repel with an exponential interaction of decay length approximately $\phi^*$4 $\phi^*$5m, whereas opposite-handed pairs have an attractive well with minimum at edge-to-edge separation $\phi^*$6 $\phi^*$7m and depth approximately $\phi^*$8. The paper further describes higher-order architectures explicitly as a rich “Rafture,” including stable tetramers, alternating chains, and “ionic crystallites” on a binary square lattice (Miller et al., 2019).

“The Structure of Cholesterol in Lipid Rafts” is again not a rupture study. By combining coarse-grained simulations and neutron diffraction, it identifies raft-like structures in DPPC–cholesterol membranes, including strongly bound cholesterol pairs at $\phi^*$9 Å in the liquid-disordered phase, a monoclinic lattice in raft-like ordered domains, and triclinic cholesterol bilayer plaques. This literature uses “raft” to denote compositionally distinct membrane domains rather than mechanically failing aggregates (Toppozini et al., 2014).

The distinction matters. In the failure literature, rafture is a dynamical transition from integrity to crack, rift, void, or fragment. In the membrane-domain literature, “raft” denotes a stable or metastable mesoscale structure whose physics is governed by chirality, depletion, line tension, orientational elasticity, and composition. The two literatures share geometric confinement and interfacial energetics, but not the same failure semantics.

6. Rafture as an erasure-coded Raft protocol

A fully separate meaning appears in “Rafture: Erasure-coded Raft with Post-Dissemination Pruning” (Kerur et al., 25 Mar 2026). Here Rafture is not a mechanical phenomenon but a consensus-and-storage algorithm for distributed systems operating under partial synchrony with crash failures. The system has $K \approx 0.665$00 nodes, tolerates up to $K \approx 0.665$01 failures with $K \approx 0.665$02, and preserves Raft’s majority commit rule. The paper’s central claim is that Rafture is the first information-dispersal solution to incorporate post-dissemination pruning, thereby adapting storage cost after dissemination completes while keeping recovery uniform (Kerur et al., 25 Mar 2026).

The coding strategy fixes the threshold once and varies only the number of distinct fragments assigned per node. A leader encodes a log entry payload $K \approx 0.665$03 into a very large fragment space using a

$K \approx 0.665$04

erasure code, creating $K \approx 0.665$05 unique fragments such that any $K \approx 0.665$06 can reconstruct $K \approx 0.665$07. If the leader estimates a responsive quorum of $K \approx 0.665$08, it sends at least

$K \approx 0.665$09

distinct fragments to each node in that set. The same expression governs pruning: if a node later determines that $K \approx 0.665$10 nodes are storing the entry, it only needs to retain

$K \approx 0.665$11

fragments (Kerur et al., 25 Mar 2026).

The implementation introduces two explicit thresholds. After responses from $K \approx 0.665$12 nodes, followers prune to two fragments for all entries up to a specified $K \approx 0.665$13 threshold. After responses from all $K \approx 0.665$14 nodes, followers prune to one fragment. The formal guarantee is that enough fragments persist to reconstruct a log entry despite $K \approx 0.665$15 failures, and liveness is inherited from Raft under partial synchrony. The design differs from prior dissemination-focused systems such as CRaft, HRaft, and FlexRaft by decoupling latency-sensitive dissemination from long-term storage optimization. Empirically, it improves long-term storage consumption under dynamic network conditions and reduces the frequency of retransmission-driven extra network hops by over-sending on the critical path and pruning later (Kerur et al., 25 Mar 2026).

This systems usage is terminologically independent of the physical literature. Its inclusion under the same lexical item is significant because it demonstrates that “Rafture” on arXiv is not a single disciplinary term. In soft matter it typically abbreviates raft rupture; in distributed systems it denotes a specific erasure-coded Raft construction with fixed-threshold recovery and post-dissemination pruning.