Higher-Order Fractional Laplacians

• Higher-order fractional Laplacians are nonlocal operators that extend the classical Laplacian to arbitrary real orders, playing a key role in PDE analysis and physical phenomena.
• They are defined via spectral, hypersingular, and finite-difference formulations, enabling explicit kernel representations and detailed insights into boundary regularity.
• These operators underpin practical applications in conformal geometry, potential theory, anomalous diffusion, elasticity, and guide advanced numerical discretization methods.

Higher-order fractional Laplacians generalize both the classical Laplacian and its integer iterates to non-integer powers, providing a unified framework for nonlocal operators of arbitrary real order. They are fundamental in the analysis of partial differential equations (PDEs) involving nonlocality and play crucial roles in conformal geometry, potential theory, elliptic regularity, and various applications in physics, such as anomalous diffusion and elasticity. Their study encompasses operator theory, variational methods, explicit kernel representations, maximum principle failures, boundary regularity, and computational discretization schemes.

1. Definitions and Operator Formulations

Several equivalent definitions of the higher-order fractional Laplacian $(-\Delta)^s$ for $s>0$ are standard:

• Spectral (Fourier) Definition: For $u\in\mathcal{S}(\mathbb{R}^n)$,

$$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$

where $\mathcal{F}$ denotes the Fourier transform. This definition immediately extends to tempered distributions and to Sobolev/Bessel potential spaces by duality (Fontana et al., 2017).

• Integral (Hypersingular) Formulation (Riesz, pointwise): For non-integer $s>0$, write $s=m+\sigma$ with $m\in\mathbb{N}_0$, $\sigma\in(0,1)$. For smooth $u$

$s>0$0

with normalization so that the Fourier symbol is $s>0$1 (Abatangelo et al., 2016, Saldaña, 2018, Fontana et al., 2017).

• Finite-Difference ("Grünwald-Letnikov"-type) Formula: Alternative representation using a finite number of function evaluations, typically as

$s>0$2

which recovers classical polyharmonic operators for integer $s>0$3 (Saldaña, 2018, Abatangelo et al., 2022).

• Spectral Fractional Laplacian in Domains: For a bounded domain $s>0$4 with Dirichlet Laplacian eigenpairs $s>0$5,

$s>0$6

mapping the Sobolev domain $s>0$7 to its dual (Otárola et al., 6 Mar 2026).

2. Fundamental Solutions, Integral Representations, and Extension Problems

• Riesz Potential and Green Kernels: For $s>0$8, the fundamental solution is $s>0$9, satisfying $u\in\mathcal{S}(\mathbb{R}^n)$0 distributionally (Abatangelo et al., 2016). For the Dirichlet problem in the unit ball $u\in\mathcal{S}(\mathbb{R}^n)$1, explicit representations via Boggio's formula are available, giving $u\in\mathcal{S}(\mathbb{R}^n)$2 with precise boundary decay and regularity (Saldaña, 2018, Abatangelo et al., 2017).
• Integral Equation Equivalence: For PDEs of the form $u\in\mathcal{S}(\mathbb{R}^n)$3 on $u\in\mathcal{S}(\mathbb{R}^n)$4, solutions often coincide with solutions of the nonlinear Riesz potential equation

$u\in\mathcal{S}(\mathbb{R}^n)$5

under suitable decay or integrability (Zhuo et al., 2016, Yang, 2021, Cao et al., 2019).

• Extension Problem (Polyharmonic and Weighted): The Caffarelli-Silvestre extension for $u\in\mathcal{S}(\mathbb{R}^n)$6 and its generalizations for $u\in\mathcal{S}(\mathbb{R}^n)$7 realize $u\in\mathcal{S}(\mathbb{R}^n)$8 as a Dirichlet-to-Neumann map for a $u\in\mathcal{S}(\mathbb{R}^n)$9-th order degenerate elliptic equation in $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$0, with $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$1-weighted Laplacian: $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$2 and boundary data prescribed at $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$3, where $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$4, $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$5 (Cora et al., 2021, Yang, 2013). This framework provides variational characterizations and explicit extension kernels.

3. Boundary Value Problems and Regularity

• Dirichlet and Nonlocal Boundary Conditions: For fractional orders, the correct notion of Dirichlet problem is to prescribe $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$6 on $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$7. For $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$8, corresponding traces ("Edenhofer kernels") correspond to consecutive boundary operators recovering standard normal derivatives as $$(-\Delta)^s u(x) = \mathcal{F}^{-1}\left(|\xi|^{2s} \widehat{u}(\xi)\right)(x),$$9 (Abatangelo et al., 2017, Saldaña, 2018).
• Boundary Regularity: Solutions $\mathcal{F}$0 to

$\mathcal{F}$1

satisfy $\mathcal{F}$2 for all $\mathcal{F}$3 ($\mathcal{F}$4 distance to $\mathcal{F}$5), with $\mathcal{F}$6 if $\mathcal{F}$7 (Ros-Oton et al., 2014, Abatangelo et al., 2016).

• Positivity and Maximum Principle: The maximum principle holds for $\mathcal{F}$8, but fails for $\mathcal{F}$9 in general, as sign-changing Green functions and eigenfunctions emerge in nontrivial domains (especially disconnected or nonconvex), with explicit characterizations via kernel sign patterns (Abatangelo et al., 2016, Abatangelo et al., 2022, Abatangelo et al., 2020).

4. Liouville Theorems, Radial Symmetry, and Classification

• Liouville-type Theorems: For subcritical powers $s>0$0 in semilinear equations with $s>0$1 or Hardy-Hénon weights, all nonnegative entire solutions must vanish, with the critical exponent given by $s>0$2 for $s>0$3 (Yang, 2021, Zhuo et al., 2016, Cao et al., 2019).
• Symmetry and Moving Spheres/Planes: In subcritical or critical cases, all nonnegative solutions are radially symmetric and monotone (up to conformal transformation). These are established both via the method of moving spheres (integral forms) and via splitting systems for the operator factorization (Zhuo et al., 2016, Yang, 2021, Cao et al., 2019).
• Super Polyharmonic Properties: Nonnegative solutions $s>0$4 of higher-order equations satisfy $s>0$5 for all $s>0$6 for $s>0$7 (Cao et al., 2019).

5. Oscillatory Phenomena, Maximum Principle Failure, and Rearrangement Inequalities

• Oscillatory Behavior and Eigenfunction Sign Changes: For $s>0$8 the kernel and first Dirichlet eigenfunction may change sign, especially for disconnected or large aspect-ratio domains. Explicit domains (unions of balls, ellipsoids) realize this anti-maximum behavior (Abatangelo et al., 2022, Abatangelo et al., 2020, Abatangelo et al., 2016).
• Breakdown of Rearrangement Inequalities: For $s>0$9, polarization and Pólya–Szegő type inequalities (reduction of Dirichlet energy by symmetric decreasing rearrangement) can fail, and may even reverse direction in certain ranges of $s=m+\sigma$0 (Abatangelo et al., 2022).
• Exceptions in One Dimension: Despite these oscillations, the Faber–Krahn inequality holds for $s=m+\sigma$1 in dimension one: the interval always minimizes the fundamental gap among open sets of fixed length (Abatangelo et al., 2022).

6. Variational Properties, Integration by Parts, and Pohozaev Identities

• Fractional Energy Forms: The natural energy is $s=m+\sigma$2. Minimizing this over $s=m+\sigma$3 vanishing outside $s=m+\sigma$4 yields the unique weak solution to $s=m+\sigma$5 (Abatangelo et al., 2016, Saldaña, 2018).
• Integration by Parts and Boundary Terms: For $s=m+\sigma$6 on $s=m+\sigma$7, and smooth $s=m+\sigma$8,

$s=m+\sigma$9

with corresponding Pohozaev-type identities for fractional orders (Ros-Oton et al., 2014).

• Unique Continuation and Runge Approximation: Solutions exhibit strong unique continuation from open sets and support Runge approximation results, providing the basis for inverse problem identifiability (Covi et al., 2020).

7. Numerical Methods, Discretization, and Implementation

• Finite Difference & High-Order Schemes: Accurate discretizations for $m\in\mathbb{N}_0$0 employ either high-order central difference stencils (with composite quadrature or lattice-Boltzmann–Hermite expansions for isotropy), or Fourier-based weights, yielding $m\in\mathbb{N}_0$1 convergence for $m\in\mathbb{N}_0$2-th order stencils (Wang et al., 29 Jan 2026, Lam et al., 2022). The discretized operators are Toeplitz and amenable to FFT acceleration.
• PDE-Based Polyharmonic Extensions: Extension-based approaches directly discretize the degenerate polyharmonic problem for the extension variable, coupling high-regularity finite elements in $m\in\mathbb{N}_0$3 and spline approximations in $m\in\mathbb{N}_0$4 to achieve algebraic convergence in both variables (Otárola et al., 6 Mar 2026, Cora et al., 2021).
• Kernel Evaluation on General Domains: Explicit formulas for $m\in\mathbb{N}_0$5 acting on polynomials or radial profiles are available for balls and ellipsoids, but general domains require weighted $m\in\mathbb{N}_0$6 spaces for integrability, and explicit kernel assembly becomes intractable outside balls (Abatangelo et al., 2017, Abatangelo et al., 2020).

8. Sharp Inequalities and Extremal Sequences

• Adams and Moser–Trudinger Inequalities: For the critical Sobolev exponent $m\in\mathbb{N}_0$7, the sharp exponential integrability bound holds for $m\in\mathbb{N}_0$8 with $m\in\mathbb{N}_0$9,

$\sigma\in(0,1)$0

sharpness demonstrated by explicit extremal ("bubble") sequences (Fontana et al., 2017).

• Extremal Solutions and Criticality: While true maximizers do not exist, carefully constructed squeezing sequences achieve blow-up at the critical constant in Adams-type inequalities (Fontana et al., 2017).

---

References

• (Fontana et al., 2017) Fontana & Morpurgo, "Sharp exponential integrability for critical Riesz potentials and fractional Laplacians on ℝ^n"
• (Yang, 2021) Chen, Li & Wang, "Liouville-type theorems, radial symmetry and integral representation of solutions to Hardy-Hénon equations involving higher order fractional Laplacians"
• (Parisis et al., 2018) Parisis & Aifantis, "Fractional Generalization of Higher-Order Diffusion"
• (Abatangelo et al., 2017) Abatangelo & Jarohs, "Integral representation of solutions to higher-order fractional Dirichlet problems on balls"
• (Abatangelo et al., 2020) Gal & Jarohs, "Fractional Laplacians on ellipsoids"
• (Covi et al., 2020) Ghosh, Salo, Uhlmann, "Unique continuation property and Poincaré inequality for higher order fractional Laplacians"
• (Zhuo et al., 2016) Zhuo & Li, "A Liouville Theorem for the Higher Order Fractional Laplacian"
• (Abatangelo et al., 2016) Abatangelo, Jarohs & Saldaña, "On the maximum principle for higher-order fractional Laplacians"
• (Cao et al., 2019) Cao, Dai, Qin, "Super poly-harmonic properties, Liouville theorems and classification of nonnegative solutions to equations involving higher-order fractional Laplacians"
• (Saldaña, 2018) Saldaña, Abatangelo & Jarohs, "On fractional higher-order Dirichlet boundary value problems: between the Laplacian and the bilaplacian"
• (Yang, 2013) Stinga & Torrea, "On higher order extensions for the fractional Laplacian"
• (Ros-Oton et al., 2014) Ros-Oton & Serra, "Local integration by parts and Pohozaev identities for higher order fractional Laplacians"
• (Abatangelo et al., 2022) Abatangelo & Jarohs, "Oscillatory Phenomena for Higher-Order Fractional Laplacians"
• (Cora et al., 2021) Cora & Musina, "The $\sigma\in(0,1)$1-polyharmonic extension problem and higher-order fractional Laplacians"
• (Wang et al., 29 Jan 2026) Wang et al., "Higher-Order Finite Difference Methods for the Tempered Fractional Laplacian"
• (Lam et al., 2022) Ghaffari, "On Generalisation of Isotropic Central Difference for Higher Order Approximation of Fractional Laplacian"
• (Otárola et al., 6 Mar 2026) Otárola & Salgado, "Approximation of higher-order powers of the spectral fractional Laplacian via polyharmonic extension"