Spreading Layer Geometry

• Spreading layer geometry is an extended, state-dependent configuration defined by variables such as thickness, footprint, and asymmetry that mediate observable phenomena.
• In neutron star accretion, wetting dynamics, and powder spreading, its geometric traits control key outputs like spectral properties, contact-line speeds, and material defects.
• Modeling approaches range from height field analyses and DEM simulations to continuum and network theories, highlighting the balance between geometric idealization and real-world inference.

In the cited literature, “spreading layer geometry” does not denote a single formalism. It refers, instead, to the geometry of an extended layer whose latitudinal coverage, footprint, thickness, depth, or coupling pattern controls the observable dynamics of radiation, wetting, granular deposition, or invasion. In neutron-star accretion it is an equatorial belt or pair of bands on a spherical surface; in wetting it is a height field $h(x,y,t)$ over a moving footprint $\Omega(t)$; in powder-bed processes it is the deposited powder-bed state after recoating; and in nonlocal or multilayer propagation it is an anisotropic spreading body, a coupled block structure, or a moving shallow layer with a front (Kajava et al., 2017, Kant et al., 2017, Penny et al., 2022, Guo et al., 11 Jun 2026).

1. Domain-specific meanings and shared descriptors

Across these works, the geometric object is always a layer-like configuration whose evolution cannot be reduced to a single scalar rate. The relevant descriptors are surface coverage, latitudinal or lateral extent, effective thickness or depth, symmetry versus asymmetry, and the relation between local transport laws and global constraints.

Context	Geometric object	Principal descriptors
Weakly magnetized neutron stars	Surface belt between $\theta_{\rm min}$ and $\theta_{\rm max}$ or two equatorial bands	Latitude coverage, occultation by the disc, visibility, emitting area
Wetting, films, puddles	Height field $h(x,y,t)$ over $\Omega(t)$	Contact angle, footprint, terraces, rivulets, depth profile
Powder spreading	Deposited powder bed after recoating	Effective layer thickness, surface coverage, packing density, roughness
Multilayer or nonlocal propagation	Coupled layers, spreading set, or thin current	Interlayer geometry, front position, anisotropic spreading body

This suggests that the central question is not merely whether a layer exists, but how its geometry mediates observables. In some settings geometry determines visibility and spectral hardening; in others it determines contact-line speed, defect morphology, or the large-time location of level sets.

2. Neutron-star spreading layers as latitudinally variable surface structures

In the neutron-star literature summarized here, the spreading layer is the continuation of the disc–star boundary onto the stellar surface. The physical picture adopted from Inogamov–Sunyaev-type models is that accreting gas reaches the equator from a thin disc and then “spreads from the NS equator (disc mid-plane) towards the NS rotational poles,” is “levitating above the NS surface,” and emits in “two bands located above the equator and up to a latitude higher up that is determined by the mass accretion rate.” For 4U 1608–52 in the soft state, the persistent spectrum was fitted with $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$, the diskbb normalization implied $R_{\rm in}\approx 10.7$ km on average, with $R_{\rm in,min}\approx 8.0$ km and $R_{\rm in,max}\approx 13.5$ km, and the comptt component carried about 35–55% of the bolometric persistent flux, all of which was interpreted as consistent with a thin disc reaching the neutron-star surface plus a bright equatorial spreading layer (Kajava et al., 2017).

The burst geometry inferred for the same source is explicitly phase dependent. In soft-state PRE bursts, spectral decomposition requires a burst blackbody plus a second optically thick Comptonized component identified with the spreading layer. During bright phases the spreading-layer component can brighten by a factor up to about 50, and the proposed mechanism is radiation pressure from burst photons passing through the equatorial layer and pushing it to higher latitudes. The strongest geometric claim is therefore interpretive: the layer “could cover almost the entire NS in the brightest phases,” whereas the preferred touchdown scenario is that “a small polar cap is always directly visible, with the rest of the NS being covered underneath the spreading layer.” The authors also argue that the “bottom half” of the NS should be covered by the accretion disc, so soft-state burst analysis violates two standard assumptions used in radius inference: the whole star is not visible, and the burst emission is reprocessed in the spreading layer. They further associate the relevant hardening with spreading-layer values $\Omega(t)$0–1.8 rather than the $\Omega(t)$1–1.4 often assumed in soft-state radius analyses, while hard-state bursts are well fit by a simple blackbody and show no comparable variable SL component (Kajava et al., 2017).

Independent evidence for the same geometry comes from thermonuclear superbursts. In 4U 1636–536, non-negative matrix factorization isolated a quasi-Planckian component with nearly constant shape, characteristic temperature around $\Omega(t)$2–$\Omega(t)$3 keV, and a flux change by a factor of $\Omega(t)$4; at peak it contributed more than 60% of the total bolometric flux. Because its shape remained nearly fixed while its normalization varied strongly, the favored interpretation was a change in emitting area or covering fraction of a belt-like spreading layer, possibly widened by radiation pressure and partially blocking direct burst emission (Koljonen et al., 2016).

A different geometric consequence appears in the multiple-burst model. At sufficiently high accretion rate, $\Omega(t)$5, the standard spreading-layer picture yields two off-equatorial settling rings, one in each hemisphere, and double bursts are interpreted as sequential ignition in these spatially separated fuel reservoirs. Triple bursts motivated a refinement in which some matter also settles in a “central ring zone” near the equator, producing an effective three-zone burning geometry: northern ring, equatorial ring, southern ring (Grebenev et al., 2017).

3. Surface-based relativistic and polarimetric geometry

The polarization model for weakly magnetized neutron stars formalizes the spreading layer as a strictly surface-based emitter. The star is a sphere of radius $\Omega(t)$6 and mass $\Omega(t)$7, and the layer is “a belt located between the minimum and maximum colatitude, $\Omega(t)$8 and $\Omega(t)$9.” The coordinate system is centered on the star, the spin axis is the $\theta_{\rm min}$0-axis, the disc lies in the $\theta_{\rm min}$1-plane, and the observer is at inclination $\theta_{\rm min}$2. Surface geometry enters through

$\theta_{\rm min}$3

A geometrically thin, optically thick disc blocks the lower half of the neutron star, so only the upper half is included as a source of direct spreading-layer photons. Visibility is further modified by Schwarzschild light bending, with surface elements visible when $\theta_{\rm min}$4, and Beloborodov’s approximation is used for the relation between $\theta_{\rm min}$5 and $\theta_{\rm min}$6. The emitting region is axisymmetric, the flow is purely azimuthal, and the local atmosphere is treated as plane-parallel rather than a 3D wedge (Bobrikova et al., 2024).

Within that model, geometry dominates the net polarization. The paper compares the whole sphere, upper hemisphere, narrow $\theta_{\rm min}$7-wide rings, broad belts, and an Inogamov–Sunyaev-like latitudinal weighting. The whole sphere yields nearly complete Stokes cancellation; hemisphere truncation by the disc produces nonzero polarization; narrow rings isolate latitude effects; and broad belts suppress polarization through stronger cancellation of local polarization angles. The fiducial wide SL geometry $\theta_{\rm min}$8, $\theta_{\rm min}$9 already shows this cancellation trend, and broadening the belt toward a hemisphere reduces the polarization degree further. Even with Doppler boosting, aberration, and energy-dependent weighting, the net polarization degree remains below about $\theta_{\rm max}$0, while polarization-angle swings can reach $\theta_{\rm max}$1–$\theta_{\rm max}$2. The low polarization is therefore not a statement about weak local scattering polarization; it is a statement about the near-axisymmetry of the surface belt and the consequent Stokes cancellation. The model is also explicit about what it omits: non-axisymmetric patchiness, finite vertical thickness, magnetic anisotropy, disc reflection, and scattering in winds (Bobrikova et al., 2024).

4. Wetting films, droplets, and puddles

In wetting theory, spreading layer geometry is typically encoded in a height field $\theta_{\rm max}$3 and a moving footprint. For droplets on topographic substrates, the central geometric statement is that substrate slope modifies the local apparent contact angle. With free-surface height $\theta_{\rm max}$4 and substrate topography $\theta_{\rm max}$5, the paper gives

$\theta_{\rm max}$6

The topographic correction $\theta_{\rm max}$7 increases spreading when the contact line moves uphill and suppresses it when it moves downhill. In recessed stadium-shaped pixels, this produces thin pointed rivulets along the floor–wall corner; their appearance is consistent with the Concus–Finn criterion

$\theta_{\rm max}$8

which for $\theta_{\rm max}$9 gives $h(x,y,t)$0. The quasi-static constant-curvature model captures overall morphology, wall-guided propagation, and the threshold-like disappearance of rivulets as $h(x,y,t)$1 increases, but it does not include the viscous resistance that controls late-time rivulet propagation (Kant et al., 2017).

At molecular scales, the same geometric language acquires a different meaning. The effective film height is defined through adsorption,

$h(x,y,t)$2

and the free energy is

$h(x,y,t)$3

When the binding potential $h(x,y,t)$4 is oscillatory, with multiple local minima corresponding to one, two, three, or more molecular layers, the spreading geometry is no longer a smooth wedge above a single precursor film. Instead it becomes terraced: flat or nearly flat segments at preferred heights separated by narrow transitions. The dynamic equation

$h(x,y,t)$5

then has to interpolate between hydrodynamic transport for thick films and diffusive transport for ultrathin adsorption layers. The resulting geometry includes metastable terraces, pinning at favored heights, and “popping” transitions between neighboring minima of the oscillatory free-energy landscape (Yin et al., 2016).

For slowly spreading sessile droplets, the geometry can instead be reduced to a one-parameter family of spherical caps. The exact cap relations

$h(x,y,t)$6

and

$h(x,y,t)$7

close the spreading problem when the capillary number is assumed to be a function of the contact angle. The resulting ODE,

$h(x,y,t)$8

extends de Gennes’ thin-drop construction to arbitrary contact angles. On swellable polymer brushes, the paper shows that the spherical-cap geometry can remain accurate even when the constitutive law $h(x,y,t)$9 ceases to be universal, so geometric closure and dynamical closure separate cleanly (Fricke et al., 2020).

A more general non-axisymmetric generalization appears in the puddle model. There the geometry is split into a moving wetted domain $\Omega(t)$0 and a depth field $\Omega(t)$1 over that domain, with

$\Omega(t)$2

The contact angle is the boundary slope,

$\Omega(t)$3

and the contact-line evolution follows the Euler–Lagrange law

$\Omega(t)$4

This formulation recovers a local Hoffman–Voinov–Tanner law for small non-axisymmetric puddles and a nonlocal Hele–Shaw-like limit for large regular puddles, while also predicting smoothing of protrusions and indentations through the nonlocal dependence of $\Omega(t)$5 and $\Omega(t)$6 on the full footprint (Darrow, 11 Jun 2026).

5. Granular and powder-bed deposition geometries

In metal powder bed fusion, the deposited powder bed is treated explicitly as a process state variable. Geometry is therefore not limited to the nominal layer thickness; it includes effective layer thickness, surface coverage, packing density, height distribution, roughness, and spatial homogeneity. The chapter defines the central metrics as

$\Omega(t)$7

Nominal layer thickness is the machine-imposed displacement, whereas effective layer thickness is the actual deposited-layer descriptor derived from the particle-height distribution. The chapter repeatedly links these geometric quantities to downstream energy absorption, melt-pool stability, lack-of-fusion risk, and mechanical performance, and it emphasizes that particle size distribution, morphology, cohesion, flowability, recoater velocity, recoater type, humidity, and chamber conditions all alter the final bed geometry (Astarita, 22 Jun 2026).

The binder-jet sand study sharpens this geometric description by focusing on empty patches and deposited particle volume. Spreadability is defined there as the ability of powder “to be uniformly spread through a constriction to form a thin and dense layer without any defects such as empty patches and particle agglomerations.” The main quantitative layer metric is

$\Omega(t)$8

where $\Omega(t)$9 is the total particle volume in the spread layer. Empty patches are “the area on the surface of the base which is not covered by any particles,” and their area percentage is obtained by image analysis. Several empty patches with different size and shapes are observed along the spreading direction even when the gap height reaches $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$0. On a rough base the empty-patch area decreases from about 60% at $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$1 to about 16% at $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$2, while on a smooth base there are almost no particles left on the base at small gap because of full-slip flow. The study further identifies a direct correlation between mechanical jamming near the blade gap and local empty spaces in the final layer, and it emphasizes that large rolling resistance is important for the mechanical jamming responsible for these defects (Xu et al., 26 Feb 2025).

Transmission X-ray imaging and DEM then show how blade geometry itself controls the geometry of thin metal powder layers. For a nominal $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$3 Ti-6Al-4V layer, the measured mean effective depth is $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$4 for a rigid $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$5 blade, $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$6 for a rigid $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$7 blade, $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$8 for a V-shaped compliant blade, and $${\tt tbabs}\times({\tt diskbb}+{\tt comptt}+{\tt gauss})$$9 for a cylindrical blade. The central conclusion is that curved or radiused blade profiles increase material deposition. The primary advantage of the V-shaped rubber blade over the $R_{\rm in}\approx 10.7$0 rigid blade is not gross bending, but local deflection of the blade edge that eliminates streaking from large particles while also increasing deposition. The $R_{\rm in}\approx 10.7$1 blade can leave severe full-thickness streaks because large particles become trapped against the substrate, whereas the cylinder yields the densest layers but introduces stick-slip banding. DEM additionally identifies a critical surface energy of about $R_{\rm in}\approx 10.7$2, beyond which layer density is greatly impaired for blade spreading, and shows that lower-density powders are more sensitive to the same cohesive scale (Penny et al., 2022).

6. Multilayer and continuum propagation geometries

In continuum gravity-current theory, spreading layer geometry appears as the thickness profile and front structure of a shallow layer moving against a deep ambient. When one layer is asymptotically deep, the two-layer shallow-water equations reduce consistently to a locally conservative one-layer model for the spreading layer, with reduced gravity $R_{\rm in}\approx 10.7$3. The reduction is not limited to smooth interiors: it also captures fronts, shocks, and contact discontinuities without any extra front condition. In one-dimensional dam-break geometry the reduced model predicts a rarefaction, a finite-thickness shelf, and a sharp leading discontinuity; in the deep-ambient limit the front Froude number approaches $R_{\rm in}\approx 10.7$4, consistent with the von Kármán and Benjamin front conditions (Fyhn et al., 2019).

For nonlocal Fisher–KPP invasion, the geometry is instead encoded in an intrinsic spreading body. The multidimensional theory defines

$R_{\rm in}\approx 10.7$5

and for unbounded initial supports the full spreading set is

$R_{\rm in}\approx 10.7$6

This is a Minkowski-sum representation of the asymptotic spreading geometry. Because the kernel may be asymmetric, the intrinsic spreading set need not contain the origin, so propagation can be biased, one-sided, or conical. The theory proves local Hausdorff convergence of rescaled level sets toward $R_{\rm in}\approx 10.7$7 and, under stronger assumptions, $R_{\rm in}\approx 10.7$8 in physical space (Guo et al., 11 Jun 2026).

In network science, spreading layer geometry is the geometry of coupling between layers. The multilayer review defines nodes as node-layer pairs $R_{\rm in}\approx 10.7$9 and edges as $R_{\rm in,min}\approx 8.0$0, distinguishes intra-layer from inter-layer pathways, and emphasizes that spreading depends on how multiple adjacency structures coexist, overlap, and couple, rather than on a single graph (Salehi et al., 2014). In the single-edge SIS optimization problem, the interlayer geometry of a two-layer system is represented by a block adjacency perturbation, and the optimal edge is not generically “hub-to-hub”; away from threshold, the best bridge tends to connect a highly infected region to a weakly infected, still-receptive region, as encoded in a dynamical Katz-like index built from the steady state on the isolated layers (Pan et al., 2019). In patchy environments with class-specific mobility layers, each class disperses on its own CTMC network over the same patch set, and the effective epidemic geometry is governed by the spectral abscissa $R_{\rm in,min}\approx 8.0$1, where mobility geometry enters through $R_{\rm in,min}\approx 8.0$2 and patch-level inter-class mixing enters through $R_{\rm in,min}\approx 8.0$3 (Abhishek et al., 2022). In two-layer SIR networks, multiple epidemic peaks emerge when the layers have sufficiently different degree distributions and the inter-layer coupling is weak, so that the outbreaks in the two layers occur at different times (Zheng et al., 2017).

7. Recurring themes, limitations, and contested points

A first recurrent theme is that geometry is usually state dependent. The neutron-star spreading layer expands and recedes during a burst; droplet footprints deform because $R_{\rm in,min}\approx 8.0$4 modifies the local contact angle; powder-bed geometry changes with recoater kinematics, cohesion, and jamming; and nonlocal invasion sets depend jointly on intrinsic wave geometry and the geometry of the initial support. This suggests that “spreading layer geometry” is better understood as a coupled dynamical state than as a static boundary condition.

A second theme is the tension between geometric idealization and observable inference. In 4U 1608–52, “almost the entire NS” coverage is presented as an informed interpretation rather than a direct measurement, and a specific tension remains between the theoretical radial SL width of only “a few hundred meters” and the continued visibility of an SL-like component during PRE expansion to “few tens of kilometers” (Kajava et al., 2017). In the polarization model, the surface belt is axisymmetric by construction and omits non-axisymmetric patchiness, finite vertical thickness, and disc reflection, which limits what can be inferred from its $R_{\rm in,min}\approx 8.0$5 polarization prediction (Bobrikova et al., 2024). In topography-controlled droplet spreading, the quasi-static model captures the geometry of rivulets but not their long-time propagation rate because it does not include viscous resistance in the rivulet body (Kant et al., 2017). In spherical-cap spreading, the cap geometry can remain accurate even when the assumed universal law $R_{\rm in,min}\approx 8.0$6 fails, so geometric adequacy does not imply dynamical adequacy (Fricke et al., 2020). In DEM-based powder spreading, high computational cost, simplified particle shapes, restricted domain sizes, and limited multilayer build histories constrain predictive fidelity (Astarita, 22 Jun 2026, Penny et al., 2022).

A third theme is that symmetry usually suppresses distinguishable signatures, whereas asymmetry generates them. Axisymmetric spreading-layer belts lead to Stokes cancellation; broad powder layers with uniform coverage suppress large local defects; and isotropic nonlocal kernels recover centered spreading bodies. By contrast, disc occultation, topographic slopes, terraced binding potentials, weakly coupled heterogeneous network layers, and asymmetric nonlocal kernels all generate qualitatively new observables: hidden stellar area, wall-guided rivulets, stepped contact regions, multiple epidemic peaks, or biased spreading sets. In that sense, the geometry of the spreading layer is often the mechanism by which local microscopic rules become macroscopic, and sometimes directly measurable, structure.