Diffusive Hot-Spot Model Overview

• Diffusive Hot-Spot Model is a framework where a spatially localized source couples with transport and loss mechanisms such as diffusion, conduction, and radiative cooling.
• It is applied across fields—from astrophysical plasmas near black holes and cosmological settings to engineered hot spots in fusion, photovoltaics, and nanoscale electronics.
• The model emphasizes the balance between localized injection and environmental smoothing, providing insights into stability, cooling breaks, and dynamic transport phenomena.

“Diffusive Hot-Spot Model” is not a single canonical formalism shared verbatim across disciplines. In this entry, the phrase is used as an Editor's term for a family of models in which a localized hot spot, hot region, or hot-spot-like concentration evolves under diffusion-like transport, radiative loss, conductive smoothing, current redistribution, or related nonlocal coupling. Across the literature, this label spans orbiting synchrotron-emitting plasma near Sgr A*, cooling and continuously injected radio-galaxy hot spots, inertial-confinement-fusion ignition regions, electrothermal runaway in thin-film photovoltaics and nanoscale transistors, and reaction–diffusion hot-spot patterns in urban crime models (Yfantis et al., 2023, Sunada et al., 2022, Fu et al., 2024, Vasko et al., 2013, Kolokolnikov et al., 2012).

1. Core mathematical structure

Taken across these uses, the common structure is a localized source coupled to transport and loss. This suggests a recurring template: a field localized in space is amplified or sustained by a source term, while diffusion, conduction, cooling, expansion, or nonlocal coupling determines its spread, decay, or stability.

Representative formulations differ substantially by domain:

Domain	Transported quantity	Representative equation
PBH hotspot cosmology	Energy density / temperature	$$\partial \rho/\partial t + 4H\rho = (1/a^2)\nabla_c \cdot ( D_E(T)\nabla_c \rho ) + S_\rho$$
Sgr A* diffusive extension	Electron phase-space density	$$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$
Thin-film photovoltaics	Temperature	$$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$
Urban crime	Attractiveness / criminal density	$A_t=\varepsilon^2\Delta A-A+PA+\alpha$, $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$

In the Sgr A* flare-fitting work, diffusion is explicitly absent from the baseline hot-spot fit: “No explicit spatial or momentum diffusion of electrons is modeled in this work.” The proposed extension is to evolve the electron distribution with spatial diffusion $D(\mathbf{x},t)$, momentum diffusion $D_p(p)$, losses $\dot p$, escape $t_{\rm esc}$, and injection $Q$ (Yfantis et al., 2023). By contrast, in PBH hotspot evolution the diffusion equation itself is central, and cosmological expansion modifies both the time evolution and spatial gradients (Kim et al., 8 May 2026). In photovoltaics, the hot spot is an electrothermal instability driven by nonlinear local heating and in-plane thermal diffusion (Vasko et al., 2013). In the crime model, the “hot spot” is a localized spatial concentration generated by a reaction–diffusion–drift system rather than literal temperature (Kolokolnikov et al., 2012).

2. Relativistic plasma hot spots near black holes

For Sgr A*, one line of work models a compact synchrotron source as a Gaussian ball of plasma orbiting in the equatorial plane of a Kerr black hole, embedded in a static background radiatively inefficient accretion flow. The hot spot has standard deviation $$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$0, is centered at radius $$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$1–$$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$2, and emits thermal synchrotron radiation with full covariant polarized radiative transfer solved by ipole (Yfantis et al., 2023). The inferred inclination is $$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$3–$$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$4 deg, the preferred period is approximately $$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$5–$$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$6 min, the orbital radius is $$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$7–$$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$8, and non-Keplerian models prefer $$\partial n_e/\partial t + \nabla\cdot(n_e \mathbf{v}) = \nabla\cdot(D\nabla n_e) - \partial_p[\dot p\,n_e - D_p\,\partial_p n_e/\partial p] - n_e/t_{\rm esc} + Q$$9–$$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$0, indicating sub-Keplerian rotation. The best fit is a model without cooling, with non-Keplerian motion, and with magnetic field pointing toward the observer for $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$1; that model also matches the observed circular polarization variability (Yfantis et al., 2023).

In that framework, cooling is represented only by an exponential decay of the electron temperature, $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$2, while the density and magnetic field remain Gaussian in space. The paper states that this prescription “is not a full synchrotron cooling solution, but a simple knob to explore its impact,” and that particle injection, escape, and spatial spreading are not included (Yfantis et al., 2023). A plausible implication is that the model captures radiative transfer and orbital modulation more directly than transport physics internal to the spot.

A complementary ngEHT imaging study addresses a distinct “shearing hot spot” scenario near Sgr A*. It proposes tracing the dynamical motion of a hot spot with the dynamical image reconstruction algorithm StarWarps. In that picture, the hot spot may form as the exhaust of magnetic reconnection in a current sheet near the black hole horizon, be ejected into an orbit in the accretion disk, and then “shear and diffuse due to instabilities at its boundary during its orbit” (Emami et al., 2022). The motion is divided into two phases: first the appearance of a bright blob, then a shearing phase simulated as a stretched ellipse. Using EHT(2017,2022), ngEHTp1, and ngEHT arrays, the analysis reconstructs the hot spot and infers the first-phase state together with the axes ratio and ellipse area in the second phase. The study concludes that newly added dishes may easily trace the first phase as well as part of the second phase, before the flux is reduced substantially, and therefore identifies the ngEHT as key to directly observing dynamical motions near variable sources such as SgrA* (Emami et al., 2022).

3. Cooling breaks, continuous injection, and cosmological diffusion

In the Cygnus A hot spot D literature, the hot spot is a one-zone, uniform region of radius $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$3 filled with a magnetic field $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$4 and a relativistic downstream flow emerging from a strong relativistic shock at the jet terminus. Electrons are continuously injected at the shock via diffusive shock acceleration and advected downstream. The electron distribution is a broken power law with a cooling break produced by radiative losses during residence in the hot spot downstream region (Sunada et al., 2022). The radio-to-far-infrared spectrum exhibits a break at $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$5, with $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$6, and the paper interprets this as the impact of radiative cooling on an electron distribution sustained by continuous injection from diffusive shock acceleration. Combining the cooling-break relation with the synchrotron-self-Compton X-ray constraint yields $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$7–$$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$8 and $$C\,\partial T/\partial t = \nabla\cdot(k\nabla T) + Q(V,T) - \alpha(T-T_0) - \sigma(T^4-T_0^4)$$9–$A_t=\varepsilon^2\Delta A-A+PA+\alpha$0, while the observed X-ray flux is found to be highly dominated by SSC (Sunada et al., 2022).

The PBH hotspot problem shifts the same broad theme into an expanding-universe setting. There the hotspot is a relativistic plasma heated by Hawking radiation from a light primordial black hole, with transport controlled by a mean free path $A_t=\varepsilon^2\Delta A-A+PA+\alpha$1. In comoving coordinates, the diffusion equation is $A_t=\varepsilon^2\Delta A-A+PA+\alpha$2, and the paper shows that hotspot formation is robust against cosmological expansion (Kim et al., 8 May 2026). The critical scale at which Hubble expansion overtakes diffusion coincides with the decoupling radius, the temperature envelope remains $A_t=\varepsilon^2\Delta A-A+PA+\alpha$3, and the cooling-stage plateau temperature follows $A_t=\varepsilon^2\Delta A-A+PA+\alpha$4, steeper than the flat-spacetime scaling $A_t=\varepsilon^2\Delta A-A+PA+\alpha$5 (Kim et al., 8 May 2026). A central consequence is that all hotspots disappear within a finite time, rather than surviving indefinitely in part of parameter space as in the flat-spacetime treatment.

These two cases share a common logic but not the same microscopic transport. In Cygnus A, the break is set by continuous injection plus radiative cooling in a shock-powered synchrotron source (Sunada et al., 2022). In PBH cosmology, the hotspot is controlled by diffusion in an FRW background and by the suppression of transport through the $A_t=\varepsilon^2\Delta A-A+PA+\alpha$6 factor multiplying spatial gradients (Kim et al., 8 May 2026).

4. Fusion hot spots: non-equilibrium, magnetization, and localized mix

In inertial confinement fusion, hot-spot models are dominated by thermal conduction, alpha-particle deposition, radiation, and expansion. A two-temperature non-equilibrium model for ignition writes separate ion and electron energy balances, with electron conduction as the dominant diffusive loss channel. The paper reports a “spontaneous self-organization evolution,” manifesting as the bifurcation of ion and electron temperatures in both isobaric and isochoric scenarios. The physical reason given is the “preponderant deposition rates of alpha-particles into D-T ions and the decreasing rate of energy exchange between electrons and D-T ions at elevated temperatures,” which produces a higher ion temperature and lower electron temperature during ignition; this directly augments D-T reactions and mitigates electron conduction and bremsstrahlung losses (Fu et al., 2024).

A magnetized extension modifies precisely that conductive channel. A semi-analytic hot-spot model for ICF implosions with an applied axial magnetic field incorporates 2D Braginskii anisotropic heat flow and shows that magnetization alters the radial temperature profile, increasing central peakedness most strongly for moderately magnetized implosions, specifically “with 8-14 T applied field,” relative to both unmagnetized and highly magnetized cases (Spiers et al., 28 Feb 2025). The same work states that indirect-drive NIF experiments with 12 T and 26 T applied magnetic fields demonstrate up to $A_t=\varepsilon^2\Delta A-A+PA+\alpha$7 increase in temperature and $A_t=\varepsilon^2\Delta A-A+PA+\alpha$8 increase in fusion yield. For gas-filled Symcap implosions, the model fits the experimental central temperature amplification accurately, while the yield discrepancy suggests a systematic degradation such as mix, plus additional degradation in the unmagnetized reference shot (Spiers et al., 28 Feb 2025).

Localized mix introduces another transport-controlled degradation. Ion Vlasov–Fokker–Planck simulations with a localized spike of carbon mix totaling $A_t=\varepsilon^2\Delta A-A+PA+\alpha$9 of the hot-spot mass show that the mix region cools and contracts over tens of picoseconds, increasing alpha stopping power and radiative losses (Sadler et al., 2019). The paper concludes that a localized mix region is more severe than an equal amount of uniformly distributed mix, and also identifies a purely kinetic effect: faster ions in the tail of the distribution are absorbed by the mix region, reducing fusion reactivity by several percent on average and by up to approximately $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$0 locally near strong gradients (Sadler et al., 2019). Radiative cooling and contraction of the spike induce fluid motion, which broadens the neutron spectrum and artificially increases inferred ion temperatures, with line-of-sight variation.

Taken together, these fusion studies show three distinct uses of a diffusive hot-spot description: electron thermal diffusion as the dominant sink in ignition theory, anisotropic suppression of that diffusion by magnetization, and localized impurity structures that act as diffusive sinks and alpha-stopping barriers.

5. Engineered electrothermal and non-Fourier hot spots

In forward-biased thin-film photovoltaics based on a-Si:H technology, hot spots spontaneously emerge from coupled electrical and thermal diffusion. The system is modeled as many identical diodes in parallel, connected through the resistive electrode and through thermal connectors, thereby coupling electric and thermal processes (Vasko et al., 2013). The continuum temperature equation is $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$1, with local heat generation $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$2. Experimentally and numerically, the spots evolve by shrinking in diameter and increasing temperature up to approximately $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$3C above that of the surrounding area, while the threshold range for onset is $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$4–$$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$5 A (Vasko et al., 2013). The linearized instability criterion is $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$6, with short wavelengths stabilized by the $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$7 diffusion term. This is a literal diffusive hot-spot instability: positive electrothermal feedback concentrates current into a warmer region, while lateral thermal diffusion and environmental cooling oppose collapse.

At smaller scales, nanoscale hot spots in FinFET and HEMT devices drive transport into a non-Fourier regime. Nonequilibrium MD and FEM benchmarking shows that FEM using bulk thermal conductivity $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$8 significantly underestimates hot-spot temperature, even when the channel thickness is approximately $$\tau P_t=D\nabla\cdot(\nabla P-\tfrac{2P}{A}\nabla A)-PA+\gamma-\alpha$$9 times the phonon mean free path (Abedien et al., 30 Dec 2025). The paper introduces a size-dependent $D(\mathbf{x},t)$0 so that FEM can reproduce MD hot-spot temperatures with high fidelity. It also decomposes the MD-extracted thermal resistance into diffusive spreading, cross-plane ballistic, heat-carrier selective heating, and residual 3D ballistic-spreading contributions (Abedien et al., 30 Dec 2025). The concrete mapping is strongly size-dependent: for $D(\mathbf{x},t)$1, $D(\mathbf{x},t)$2, while for $D(\mathbf{x},t)$3, $D(\mathbf{x},t)$4–$D(\mathbf{x},t)$5.

A related but distinct formulation appears in superconducting nanowire single-photon detectors. There the hot spot is the local region around the photon absorption site in which the quasiparticle distribution is driven far from equilibrium; nonequilibrium quasiparticles diffuse away, suppress the superconducting order parameter, and crowd the transport current (Vodolazov, 2014). The modified hot-spot model solves the Ginzburg–Landau equation self-consistently with a local suppression parameter $D(\mathbf{x},t)$6 inside a finite circular area of radius $D(\mathbf{x},t)$7, rather than prescribing a uniform current “sidewalk” model. The switching threshold is associated with vortex–antivortex pair nucleation inside the hot spot and their subsequent unbinding, and the model fits the current dependence of the red boundary with a single fitting parameter $D(\mathbf{x},t)$8, which denotes the fraction of photon energy that goes into local suppression of superconductivity (Vodolazov, 2014).

6. Reaction–diffusion, mantle forcing, and momentum-space extensions

Some of the most mathematically explicit hot-spot models arise outside literal heat transport. In the urban crime model, hot-spot patterns are steady localized structures of criminal activity generated by a reaction–diffusion–drift system for attractiveness $D(\mathbf{x},t)$9 and criminal density $D_p(p)$0 (Kolokolnikov et al., 2012). Singular perturbation methods construct steady hot spots in one and two dimensions, and nonlocal eigenvalue problems determine their stability on an $D_p(p)$1 time scale. The central result is a critical threshold $D_p(p)$2 such that a pattern consisting of $D_p(p)$3 hot spots is unstable to a competition instability if $D_p(p)$4; the instability is driven by a positive real eigenvalue and triggers the collapse of some of the hot spots (Kolokolnikov et al., 2012). The same paper shows that oscillatory instabilities occur only in a relatively narrow parameter range, and studies a Hopf bifurcation explicitly for a single hot spot in the shadow-system limit.

In planetary core dynamics, a hot spot becomes a localized core–mantle boundary heat-flux anomaly rather than a thermal peak inside the fluid itself. The governing thermal equation contains explicit diffusion, $D_p(p)$5, and the study examines how anomaly amplitude, width, and position modify core convection and dynamo action (Dietrich et al., 2015). For purely hydrodynamic models, the equatorial antisymmetric, axisymmetric symmetry scales almost linearly with CMB amplitude and size, whereas self-consistent dynamo simulations either suppress or drastically enhance that symmetry depending mainly on the horizontal extent of the anomaly. The study concludes that the anomaly should be on the same order as the outer core radius to significantly affect flow and field symmetries, and further argues that such anomalies cannot explain the present-day north–south crustal magnetization asymmetry on Mars within the model range explored (Dietrich et al., 2015).

A further extension occurs in the fermionic hot-spot model of correlated electrons on the square lattice. There the “hot spots” are the eight Fermi-surface points connected by $D_p(p)$6, and the baseline theory is a purely fermionic one-loop RG rather than a diffusion equation (Whitsitt et al., 2014). The paper finds that non-nested hot spots with $D_p(p)$7, $D_p(p)$8 interactions have competing singlet $D_p(p)$9 superconducting and $\dot p$0-form factor incommensurate density-wave instabilities (Whitsitt et al., 2014). The same source explicitly notes that the analyzed model does not include disorder-induced electron diffusion or bosonic collective modes with dissipative or diffusive dynamics; those are presented as plausible extensions through propagators such as

$\dot p$1

or $\dot p$2 (Whitsitt et al., 2014). This is therefore a non-standard usage: the hot spot is momentum-space localized, and the diffusive ingredient enters only in extensions beyond the baseline model.

Across these broader cases, the phrase “diffusive hot spot” denotes not a single ontology but a modeling strategy: localization is represented explicitly, transport is the mechanism that couples local structure to the surrounding medium, and stability or observability is controlled by the competition between source terms and smoothing, leakage, or redistribution processes.