Rebound: Cross-Disciplinary Dynamics

• Rebound is a cross-disciplinary concept describing the return, reversal, or secondary evolution of a system following an initial decline or impact.
• It manifests in diverse applications such as rapid climate emission recovery, impact detachment, cavitation bubble re-expansion, epidemic resurgence, and orbital migration.
• The term also appears as proper names in software for N-body simulations and Haskell libraries, emphasizing efficient handling of iterative processes.

Rebound denotes a return, reversal, or renewed propagation following an initial decline, arrest, or collision. Across the scientific literature, the term is used in several distinct but structurally related senses: as rapid recovery of anthropogenic emissions after a transient shock; as bounce, takeoff, or rebound suppression in impact mechanics; as secondary bubble formation after cavitation collapse; as abrupt reversal of carrier motion in relativistic wave dynamics; as outward migration reversal in dispersing protoplanetary disks; and as post-control resurgence in epidemic metapopulations. In some contexts, “REBOUND” and “Rebound” are proper names rather than physical descriptors, referring respectively to an N-body code for collisional dynamics and a Haskell library for well-scoped binding (Jackson et al., 2021, Rein et al., 2011, Santo et al., 16 Sep 2025).

1. Rebound as a cross-disciplinary dynamical motif

A common formal structure links otherwise disparate uses of rebound. An external perturbation or boundary condition first drives a system away from its prior trajectory; the subsequent evolution then exhibits either recovery toward the pre-perturbation state, reversal of motion, or emergence of a secondary state. This broad pattern appears in fossil-carbon emissions after the COVID-19 disruption, where 2021 emissions were projected to return to roughly 99% of the 2019 level (Jackson et al., 2021). It also appears in weak-gravity landing, where a lander can retain enough kinetic energy after first contact to rebound multiple times unless impact energy is dissipated (Wen et al., 5 Nov 2025).

In impact and interfacial hydrodynamics, rebound typically denotes physical detachment after collision. A non-wetting drop may spread, retract, and then leave a rigid or deformable substrate; a cavitation bubble may collapse and re-expand into a rebound bubble; a projectile or impactor may reverse direction after a transient solid-like response of a suspension (Chantelot et al., 2017, Supponen et al., 2018, Egawa et al., 2019). In more abstract dynamical settings, the term can denote reversal of support propagation or migration direction: bounded localized free Dirac wavefunctions exhibit a finite rebound phase between shrinking and expanding carriers, while planets near an expanding photo-evaporative cavity can undergo outward “rebound migration” (Castrigiano, 2020, Liu et al., 8 Sep 2025).

This diversity of usage does not imply a single universal mechanism. Rather, the literature uses the same term for a family of transition phenomena in which stored, redirected, or newly released energy changes the sign or qualitative direction of evolution. This suggests that “rebound” is best treated as a contextual term whose meaning is fixed by the governing transport mechanism: torque in disks, Kelvin impulse in cavitation, capillarity and lubrication in drops, elastic force chains in suspensions, or susceptible-pool rebuilding in epidemics.

2. Recovery and resurgence after temporary suppression

In climate-carbon accounting, rebound refers to rapid recovery after a shock-induced decline. Global fossil CO$_2$ emissions fell by 5.4%, from 36.7 Gt CO2 in 2019 to 34.8 Gt CO2 in 2020, then were projected to rise by 4.9% in 2021, with an uncertainty range of 4.1% to 5.7%, reaching 36.4 ± 0.3 Gt CO2 (Jackson et al., 2021). Because 2021/2019 ≈ 36.4/36.7 ≈ 0.992, the 2021 total was characterized as “near pre-COVID-19 levels.” The rebound was regionally uneven: China was projected at 11.1 Gt CO2, about 7% above 2019; India at 2.7 Gt CO2, about 3% above 2019; while the United States (5.1 Gt CO2), the European Union (2.8 Gt CO2), and the rest of the world (14.8 Gt CO2) remained below 2019 levels. By fuel, coal was projected to rebound above 2019 levels to 14.7 Gt CO2, natural gas also to exceed 2019 levels, while oil remained well below 2019, at about 11.5 Gt CO2, with transport—especially surface transport and aviation—still depressed (Jackson et al., 2021).

In epidemic control, rebound denotes resurgence following apparently successful vaccination. In a metapopulation SIRS model, the “honeymoon period” is the temporary drop in cases after vaccination begins, whereas the “rebound effect” is the later increase in infections after the campaign ends (Castioni et al., 3 Feb 2025). The mechanism is network-structural: vaccinating enough patches can make infection vanish in the reduced network during the campaign, allowing susceptibles to accumulate; once vaccination stops, the network can again support transmission. The paper formulates the criterion through the reduced adjacency matrix $\mathbf{A}_{\mathrm{red}}$: $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$ This condition identifies vaccination strategies likely to precipitate rebound (Castioni et al., 3 Feb 2025).

These two cases illustrate different meanings of “rebound” under temporary suppression. In emissions, rebound is near-restoration of a previous aggregate rate after an exogenous interruption. In epidemics, rebound is delayed resurgence induced by over-effective suppression of transmission in the controlled phase. A plausible implication is that rebound analysis often requires distinguishing short-term success metrics from post-intervention trajectory: the same intervention that reduces a quantity rapidly can also create the state variables that enable later return or overshoot.

3. Rebound in impact mechanics and interfacial transport

In drop impact, rebound is the detachment of a liquid body from a surface after spreading and retraction, but the literature shows that this outcome is highly tunable. On a thin PDMS membrane, rebound differs qualitatively from rigid superhydrophobic impact because the membrane deforms and then “kicks” the liquid upward as it overshoots the horizontal position (Chantelot et al., 2017). For $R=1.8~\text{mm}$, the rigid-substrate contact time was reported as $\tau_0 = 22.7 \pm 0.8~\text{ms}$, while a flexible-membrane example gave $\tau \approx 7~\text{ms}$, about a 70% reduction. The dynamics are modeled as coupled oscillators with effective frequency $f = \sqrt{f_m f_d}$, and the data collapse under $\tau f \approx 0.75$ (Chantelot et al., 2017).

Other studies analyze rebound suppression rather than acceleration. On sublimating dry ice, rebound is progressively penalized by the thickness $\delta$ of a solidified basal layer formed during impact; the coefficient of restitution satisfies

$$\epsilon^2 = 1 - a_1\bar{\delta} - a_2\bar{\delta}^3,$$

and the system exhibits a broader sequence of outcomes than conventional supercooled superhydrophobic impact: complete bouncing $\mathbf{A}_{\mathrm{red}}$0 fragmentation with rebound $\mathbf{A}_{\mathrm{red}}$1 no-rebound (Kulkarni et al., 2022). On a supersolvophobic surface, adding a very small amount of polymer suppresses rebound even though the shear viscosity and liquid-vapor surface tension change only slightly and the reported impact numbers remain roughly $\mathbf{A}_{\mathrm{red}}$2 and $\mathbf{A}_{\mathrm{red}}$3 (Lee et al., 2020). The decisive stage is a late hopping stage, in which anti-rebound arises from resistance against vertical detachment through polymer adsorption and polymer elongation force (Lee et al., 2020).

Rebound can also be enhanced by viscoelasticity. A viscoelastic shear-thinning drop on a superhydrophobic surface can undergo complete rebound in a “Balloon regime”, where a vertical ligament grows, develops a balloon-like head, and detaches without satellites (Díaz et al., 14 Feb 2025). Ligament formation begins around $\mathbf{A}_{\mathrm{red}}$4, with experiments spanning $\mathbf{A}_{\mathrm{red}}$5, and the maximum ligament length scales as $\mathbf{A}_{\mathrm{red}}$6 (Díaz et al., 14 Feb 2025). The paper attributes filament survival and complete rebound to high elasticity, which prevents breakup of the thinning ligament (Díaz et al., 14 Feb 2025).

Weak-gravity landing presents a mechanically distinct but conceptually related problem: rebound as undesired post-impact persistence of kinetic energy. In small-body environments with surface gravity on the order of $\mathbf{A}_{\mathrm{red}}$7 or, more generally, $\mathbf{A}_{\mathrm{red}}$8 to $\mathbf{A}_{\mathrm{red}}$9, even a low-speed touchdown can send a lander bouncing away (Wen et al., 5 Nov 2025). A particle-filled flexible shell suppresses rebound by combining shell deformation, shell hysteresis, particle–particle collisions, particle–shell friction, and granular rearrangement. The reported result is that the flexible shell–granule system dissipates over 90% of impact energy, typically across filling ratios from 10% to 50%, and that impacts on granular beds show a penetration-depth scaling of about $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$0, together with a two-stage velocity decay (Wen et al., 5 Nov 2025).

These examples show that in impact systems rebound is not synonymous with elasticity alone. It can be promoted by cooperative substrate recoil, prevented by freezing or polymer-mediated detachment resistance, or mitigated by passive multi-channel dissipation. This suggests that the operational question is often not whether rebound occurs, but which internal or interfacial degree of freedom controls momentum reversal.

4. Rebound bubbles, cavitation, and viscous collision

Cavitation literature uses “rebound” for the bubble that forms after primary collapse, when compressed gas and vapor re-expand. Experiments on laser-induced cavitation bubbles show that the energy of the first rebound bubble grows logarithmically with initial dipole deformation: $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$1 with $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$2 (Supponen et al., 2018). Here $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$3 is the anisotropy parameter, given in a pressure gradient by $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$4, and for gravity by $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$5. The result was reported to hold for bubbles deformed by gravity and by a near free surface, indicating that rebound energy depends primarily on initial asymmetry rather than deformation source in that regime (Supponen et al., 2018).

A related but distinct phenomenon is reversal of rebound direction in a near-wall cavitation bubble interacting with a free sphere. High-speed imaging shows a transition from away-from-wall to wallward rebound as the initial bubble–sphere separation increases (Ren et al., 24 Jun 2026). The mechanism is expressed through Kelvin impulse on the closed bubble boundary,

$$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$6

where the contact closure on the sphere contributes an additional bubble-side impulse. The paper argues that reversal is not governed primarily by the instantaneous velocity of the sphere, but by contact geometry: the bubble-side contact closure can supply an away-from-wall contribution that competes with the wallward background from the wall-image source and the sphere-induced quadrupolar field (Ren et al., 24 Jun 2026).

Dense-suspension impacts introduce yet another rebound mechanism. Numerical work on dense suspensions finds that rebound is a late-stage phenomenon caused by a percolated elastic contact network, while the observed scaling laws

$$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$7

are early-stage dynamical effects and are not directly caused by rebound (Pradipto et al., 2021). The extended “floating + force chain” model adds an elastic term $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$8, where $$R_0(\mathbf{A}_{\mathrm{red}})=\frac{\beta}{\mu}\left(1+p\,\lambda_{\max}(\mathbf{A}_{\mathrm{red}})\right)<1.$$9 is the number of connected force chains from the impactor to the bottom plate (Pradipto et al., 2021). Experiments with dense potato-starch suspensions likewise treat rebound as evidence that the impact-induced region briefly behaves like a solid. There the restitution coefficient

$R=1.8~\text{mm}$0

is nearly independent of $R=1.8~\text{mm}$1 and decreases as layer thickness $R=1.8~\text{mm}$2 increases; a Voigt model yields effective properties of approximately $R=1.8~\text{mm}$3 and $R=1.8~\text{mm}$4, while the long-timescale viscosity is $R=1.8~\text{mm}$5 (Egawa et al., 2019).

Viscous particle-wall rebound further complicates the concept by separating approach and contact phases. For a quasi-2D cylinder falling in a viscous liquid, the apparent coefficient of restitution is modeled as

$R=1.8~\text{mm}$6

with a reported critical Stokes number $R=1.8~\text{mm}$7 (Aguilar-Corona et al., 2023). The paper interprets rebound through apparent roughness, which sets collision onset, and contact time, which controls viscous loss during contact (Aguilar-Corona et al., 2023).

Taken together, these studies show that “rebound” in fluid–structure and multiphase systems often designates a secondary event that inherits memory of collapse asymmetry, contact geometry, or boundary-enabled elasticity. This suggests that rebound is frequently a diagnostic of hidden state transfer rather than a mere observable reversal.

5. Rebound in wave mechanics and orbital dynamics

In relativistic quantum dynamics, rebound is used in a sharply defined kinematic sense. For a bounded localized free Dirac wavefunction $R=1.8~\text{mm}$8, the carrier boundary in direction $R=1.8~\text{mm}$9 obeys

$\tau_0 = 22.7 \pm 0.8~\text{ms}$0

This means the carrier first shrinks inward at light speed, then after a finite rebound phase expands back outward at light speed, with direction-dependent transition times $\tau_0 = 22.7 \pm 0.8~\text{ms}$1 (Castrigiano, 2020). The motion is isotropic outside the transition region, but the rebound is anisotropic and abrupt because opposite or different directions may switch at different times. The same work also proves asymptotic probability concentration in spherical shells whose outer radius increases at light speed (Castrigiano, 2020).

In protoplanetary-disk dynamics, rebound denotes outward migration reversal near an expanding inner cavity. Two-dimensional hydrodynamical simulations show that X-ray photo-evaporation can create a cavity whose edge moves outward; planets near the steep density gradient can then experience a strong positive corotation torque that overcomes the usual negative Lindblad torque and drives outward migration (Liu et al., 8 Sep 2025). The total torque is written as

$\tau_0 = 22.7 \pm 0.8~\text{ms}$2

and for locally isothermal disks the corotation component is linked to the vortensity gradient through $\tau_0 = 22.7 \pm 0.8~\text{ms}$3, with $\tau_0 = 22.7 \pm 0.8~\text{ms}$4 (Liu et al., 8 Sep 2025). Rebound migration is strongest for super-Earth and Neptune-mass planets, is suppressed for Saturn-mass planets, and for Jupiter-mass planets modest outward drift arises instead from eccentric-disk torque asymmetry (Liu et al., 8 Sep 2025).

This mechanism extends to multi-planet systems during late disk clearing. Using 2D hydrodynamical simulations with the Dusty FARGO-ADSG code, a locally isothermal disk, and stellar X-ray photoevaporation, one study shows that divergent migration induced by the cavity edge can break mean-motion resonances such as $\tau_0 = 22.7 \pm 0.8~\text{ms}$5 and $\tau_0 = 22.7 \pm 0.8~\text{ms}$6, widen period ratios, and trigger instability (Liu et al., 9 Jun 2026). A representative resonant angle for the $\tau_0 = 22.7 \pm 0.8~\text{ms}$7 commensurability is

$\tau_0 = 22.7 \pm 0.8~\text{ms}$8

In one reported low-mass case, a $\tau_0 = 22.7 \pm 0.8~\text{ms}$9–$\tau \approx 7~\text{ms}$0 pair ended with a period ratio of about 2.85 after rebound-driven divergent migration (Liu et al., 9 Jun 2026).

Both the Dirac and disk cases use “rebound” for a reversal driven not by ordinary restitution but by causal propagation constraints or torque reversal. A plausible implication is that the term extends naturally to systems in which the sign of transport changes while the governing equations remain continuous: in one case because support boundaries are light-speed limited, in the other because the local torque budget flips near a moving density edge.

6. Proper names: REBOUND and Rebound as software systems

Not all technical uses of the word denote a physical process. REBOUND is an open-source, multi-purpose N-body code designed primarily for collisional dynamics such as planetary rings, while also supporting general gravitational N-body problems (Rein et al., 2011). It is a modular C99 code distributed under GPLv3, with OpenMP, MPI, and hybrid parallelization. The code includes three second-order symplectic integrators—leap-frog, Wisdom-Holman mapping (WH), and the Symplectic Epicycle Integrator (SEI)—as well as open, periodic, and shearing-sheet boundary conditions, direct and Barnes–Hut gravity, and collision modules including tree-based and plane-sweep algorithms (Rein et al., 2011). In this usage, REBOUND is a framework name, not a description of return motion.

A more recent software usage is Rebound, a Haskell library for well-scoped syntax with binders (Santo et al., 16 Sep 2025). Its central abstraction is the use of first-class environments representing parallel substitutions, together with scope-indexed terms that make many de Bruijn-scope invariants compile-time properties. The interface centers on Env, Bind, SubstVar, and Subst, and the library automates substitution, alpha-equivalence, and related operations while using delayed, composable environments internally for efficiency (Santo et al., 16 Sep 2025). Benchmarks reported in the paper show that Rebound produces faster code than several competing libraries and that a port of pi-forall achieved substantial time and memory reductions on several programs (Santo et al., 16 Sep 2025).

These proper-name usages are historically separate from physical rebound phenomena. Their inclusion in the technical record nonetheless reflects a broader semantic tendency: “rebound” evokes return, redirection, or efficient handling of repeated interaction. In software nomenclature, that resonance is metaphorical rather than mechanistic.