---
title: Inhomogeneous Biharmonic Schrödinger Equation
url: https://www.emergentmind.com/topics/inhomogeneous-biharmonic-schrodinger-equation
type: topic
---

# Inhomogeneous Biharmonic Schrödinger Equation

The inhomogeneous biharmonic Schrödinger equation is a class of fourth-order dispersive equations in which the biharmonic propagator is coupled to a spatially weighted nonlinearity. The canonical whole-space model is
$$
i\partial_t u+\Delta^2u=\lambda |x|^{-b}|u|^\sigma u,\qquad u(0,x)=u_0(x),
$$
with \(u:\mathbb R\times\mathbb R^d\to\mathbb C\), \(u_0\) posed in \(L^2\), \(H^s\), or \(\dot H^s\), and \(b>0\) measuring the strength of the inhomogeneity near the origin. Closely related formulations include mixed-dispersion equations of the form \(i\partial_t u-\Delta^2u+\mu\Delta u+|x|^{-b}|u|^\alpha u=0\), energy-critical focusing models, and equations with a spatially growing factor \(|x|^b\). The weight breaks translation invariance and introduces singular or unbounded coefficients, so the subject couples fourth-order Strichartz theory with weighted Sobolev inequalities, variational thresholds, virial identities, and concentration-compactness. In the broader fourth-order Schrödinger literature, the same inhomogeneous paradigm also includes initial-boundary value problems with non-homogeneous boundary data, internal control, and equations with external potentials on waveguide manifolds [1910.03908], [2211.11824], [2511.06985], [2003.09337].

## 1. Model equations, parameters, and criticality

The standard decaying-weight equation is parameterized by the spatial dimension \(d\), Sobolev index \(s\), inhomogeneity exponent \(b\), power \(\sigma\) or \(\alpha\), and a sign parameter \(\lambda\in\mathbb R\) or \(\lambda\in\{\pm1\}\) distinguishing focusing and defocusing conventions. In whole-space Cauchy problems, the principal linear flow is \(e^{it\Delta^2}\), while mixed-dispersion variants replace it by \(e^{-it(\Delta^2-\mu\Delta)}\). The mixed-dispersion model
$$
i\partial_t u-\Delta^2u+\mu\Delta u+|x|^{-b}|u|^\alpha u=0
$$
is studied for \(d\ge1\), \(\mu\ge0\), and \(0<b<\min\{d,4\}\), with the weight remaining singular at the origin but decaying at infinity [1910.03908].

For the decaying-weight equation, the scaling-critical Sobolev index is
$$
s_c=\frac d2-\frac{4-b}{\sigma}.
$$
This yields the standard trichotomy: mass-critical when \(s_c=0\), equivalently \(\sigma=\frac{8-2b}{d}\); energy-critical when \(s_c=2\), equivalently \(\sigma=\frac{8-2b}{d-4}\) for \(d\ge5\); and intercritical, or mass-supercritical and energy-subcritical, when \(0<s_c<2\). In the mixed-dispersion setting the same regime is encoded by
$$
\gamma_c=\frac d2-\frac{4-b}{\alpha},
$$
and the range \(0<\gamma_c<2\) is exactly the mass-supercritical, energy-subcritical range. When \(\mu=0\), the equation is scale-invariant; when \(\mu>0\), the repulsive harmonic perturbation breaks the scaling symmetry while preserving the same basic criticality heuristics [2211.11824].

Whole-space models typically conserve mass and a fourth-order energy. For the equation \(iu_t+\Delta^2u=\lambda |x|^{-b}|u|^\sigma u\),
$$
M[u(t)]=\int_{\mathbb R^d}|u(t,x)|^2\,dx,
$$
and
$$
E[u(t)]=\frac12\int_{\mathbb R^d}|\Delta u(t,x)|^2\,dx+\frac{\lambda}{\sigma+2}\int_{\mathbb R^d}|x|^{-b}|u(t,x)|^{\sigma+2}\,dx.
$$
In the mixed-dispersion model the energy acquires the additional quadratic term \(\frac{\mu}{2}\int |\nabla u|^2\,dx\). For the growing-weight equation \(i\partial_t v+\Delta^2v=-\epsilon |x|^b|v|^{q-1}v\), the critical index becomes
$$
s_c=\frac N2-\frac{4+b}{q-1},
$$
with mass-critical exponent \(q_m=1+\frac{8+2b}{N}\) and energy-critical exponent \(q_e=1+\frac{8+2b}{N-4}\) for \(N>4\) [2511.06985].

## 2. Cauchy theory in \(L^2\), \(H^s\), and \(\dot H^s\)

The early whole-space theory established local and global well-posedness for the decaying-weight equation in \(L^2\) and \(H^2\), together with a small-data intercritical theory and a long-time perturbation principle. In the mass-subcritical regime, local \(L^2\) well-posedness extends globally by mass conservation. In \(H^2\), global continuation was obtained in the mass-subcritical regime and in the mass-critical regime under smallness of the \(L^2\)-mass, while sufficiently small intercritical data yield global solutions and scattering. A later refinement enlarged the intercritical parameter range, especially in dimensions \(N=5,6,7\), and also developed a \(\dot H^2\)-critical local/global theory and stability result for \(5\le N\le11\) under the restriction \(0<b<\frac{12-N}{N-2}\) [1910.03908], [2105.01509].

Fractional Sobolev local theory was then extended well beyond \(s=0\) and \(s=2\). For
$$
0\le s<\min\left\{2+\frac d2,\frac32 d\right\},\qquad
0<b<\min\left\{4,d,\frac32 d-s,\frac d2+2-s\right\},
$$
and \(0<\sigma<\sigma_c(s)\), the Cauchy problem is locally well-posed in \(H^s(\mathbb R^d)\), with maximal lifespan, uniqueness in the natural Strichartz class, and the standard blow-up alternative. This extended earlier work of Guzmán–Pastor and Liu–Zhang by enlarging the admissible ranges of \(s\) and \(b\). Standard continuous dependence in the full \(H^s\) topology, rather than only in \(H^{s-\varepsilon}\), was subsequently proved on compact subintervals of the lifespan, and the flow map was shown to be locally Lipschitz under additional regularity assumptions on the nonlinearity [2206.06690], [2305.17900].

The critical theory was developed in Sobolev-Lorentz and Lorentz-type settings. For
$$
iu_t+\Delta^2u=\lambda |x|^{-b}|u|^\sigma u,
$$
local well-posedness in \(H^s\) was obtained not only in the subcritical case \(\sigma<\sigma_c(s)\) but also in the critical case \(\sigma=\sigma_c(s)\), together with small-data global well-posedness and scattering when \(\frac{8-2b}{d}\le \sigma\le \sigma_c(s)\). A further refinement treated the \(H^s\)-critical equation
$$
iu_t\pm \Delta^2u=\lambda |x|^{-b}|u|^\sigma u,\qquad \sigma=\frac{8-2b}{d-2s},
$$
for \(d\ge3\), \(1\le s<\frac d2\), and \(0<b<\min\{4,2+\frac d2-s\}\), obtaining local well-posedness and small-data scattering in both \(H^s\) and \(\dot H^s\) under less restrictive regularity assumptions on the nonlinear map \(u\mapsto |u|^\sigma u\) than in earlier critical results [2208.08657], [2409.06278].

## 3. Variational thresholds, scattering, and blow-up

In the intercritical focusing mixed-dispersion problem, a threshold theory is organized around the action
$$
S_{\mu,\omega}(f)=E_\mu(f)+\omega M(f),
$$
the Pohozaev functional
$$
G_\mu(f)=2\|\Delta f\|_2^2+\mu\|\nabla f\|_2^2-\frac{d\alpha+2b}{\alpha+2}\int |x|^{-b}|f|^{\alpha+2}\,dx,
$$
and the constrained minimization level
$$
m_{\mu,\omega}=\inf\{S_{\mu,\omega}(f):f\in H^2(\mathbb R^d)\setminus\{0\},\ G_\mu(f)=0\}.
$$
The invariant set
$$
A_{\mu,\omega}=\{f\in H^2(\mathbb R^d):S_{\mu,\omega}(f)<m_{\mu,\omega},\ G_\mu(f)\ge0\}
$$
controls the global dynamics. If \(u_0\in A_{\mu,\omega}\), then the corresponding solution is global and uniformly bounded in \(H^2\). The same framework yields spacetime estimates for \(\int_I\int |x|^{-b}|u|^{\alpha+2}\), with exponent \((1+\min\{2,b\})/2\) in general and exponent \(1/2\) for radial data. Scattering in \(H^2(\mathbb R^d)\) then follows for all \(d\ge5\), for \(d=4\) with \(1<b<2\), and in the remaining low-dimensional cases under the stated radial hypotheses. When \(\mu=0\), this extends the scattering results of Saanouni and of Campos–Guzmán to dimensions three and four; when \(\mu>0\), the energy scattering result is new. The same paper also identifies \(A_{0,\omega}=B_+\) through a sharp Gagliardo–Nirenberg inequality and the ground state \(Q\) solving \(\Delta^2Q+\omega Q-|x|^{-b}|Q|^\alpha Q=0\) [2211.11824].

At the energy-critical level, the non-radial focusing equation
$$
i\partial_t u+\Delta^2u-|x|^{-b}|u|^pu=0,\qquad p=\frac{8-2b}{N-4},
$$
admits a sharp threshold formulation in terms of the ground state \(W\) solving
$$
\Delta^2W-|x|^{-b}|W|^pW=0.
$$
For \(5\le N\le11\), if
$$
\sup_{t\in I}\|\Delta u(t)\|_{L^2}<\|\Delta W\|_{L^2},
$$
then the solution exists globally and scatters in both time directions in \(\dot H^2\). The proof establishes energy trapping, a Palais–Smale compactness statement, and a concentration-compactness/rigidity argument in the Kenig–Merle style, adapted to the lack of translation invariance and to the fact that the kinetic energy is not itself conserved in the weighted setting [2508.02796].

Complementary blow-up results identify the opposite side of the ground-state barrier. For the focusing energy-critical equation, finite-time blow-up of the \(H^2\)-norm is proved for radial data when either \(E(u_0)<0\) or
$$
0\le E(u_0)<E(W),\qquad \|\Delta u_0\|_{L^2}>\|\Delta W\|_{L^2}.
$$
For non-radial data, the conclusion depends on the size of \(b\): when \(0<b<16/N\), negative energy yields blow-up in the norm-divergence sense; when \(b\ge16/N\), negative energy and the same positive-energy/large-kinetic condition imply finite-time blow-up. The mechanism is a localized virial argument with error terms estimated either by Strauss decay in the radial case or by the decay of the weight \(|x|^{-b}\) in the non-radial case [2507.05518].

## 4. Analytical frameworks and proof technology

The earliest whole-space proofs relied on standard biharmonic Strichartz estimates, weighted Gagliardo–Nirenberg or Caffarelli–Kohn–Nirenberg inequalities, Hardy–Littlewood estimates, and explicit splitting of space into the unit ball and its complement to treat \(|x|^{-b}\). This framework supports local theory, small-data global theory, scattering criteria based on finiteness of critical Strichartz norms, and long-time perturbation results, but it generates dimension-dependent restrictions in low regularity or near-critical regimes [1910.03908], [2105.01509].

A different route uses bilinear Strichartz-type estimates in Besov spaces. For coefficients \(K(x)\) behaving like \(|x|^{-b}\), new Besov-space bilinear estimates for the Duhamel operator were used to prove local well-posedness in the whole \(H^s\)-subcritical case with \(0<s\le2\). The key gain is that the product structure is handled through time-difference Besov norms rather than only pointwise Sobolev control, which is particularly effective when \(f(u)\) has limited differentiability and \(K(x)\) is singular [2103.08154].

The Lorentz-space program was developed to treat the weight more systematically. Sobolev-Lorentz spaces
$$
H^s_{p,q}=\{f:(I-\Delta)^{s/2}f\in L^{p,q}\},\qquad \dot H^s_{p,q}=\{f:(-\Delta)^{s/2}f\in L^{p,q}\},
$$
support Hölder, Hardy, embedding, product, and chain-rule estimates that are well matched to the fact that \(|x|^{-b}\in L^{d/b,\infty}\). A central advance was the extension of the Lorentz-space fractional chain rule from \(0<s\le1\) to all \(s>0\), which made critical \(H^s\)-analysis possible in a unified framework [2208.08657].

In the intercritical scattering theory, Lorentz estimates also reshape the nonlinear decay mechanism. Dispersive and Strichartz estimates for \(U_\mu(t)=e^{-it(\Delta^2-\mu\Delta)}\) were proved in Lorentz spaces, including derivative-gain estimates crucial in dimensions three and four. The weighted nonlinearity could then be estimated without splitting space into near and far regions, and localized virial/Morawetz functionals supplied the spacetime bounds needed for the scattering criterion [2211.11824].

At the critical level, Lorentz-type refinements were pushed further to Besov-Lorentz and Triebel–Lizorkin–Lorentz spaces. Regular Strichartz estimates with gains in the spatial regularity index were established for the biharmonic propagator and then inserted into \(H^s\)-critical fixed-point arguments. In the non-radial energy-critical theory, this functional-analytic machinery is complemented by linear profile decomposition, nonlinear profile analysis, and rigidity via localized virial identities, producing a full concentration-compactness/rigidity scheme for the weighted fourth-order problem [2409.06278], [2508.02796].

## 5. Boundary-driven and controlled fourth-order flows

The inhomogeneous biharmonic Schrödinger equation also appears as an initial-boundary value problem with prescribed boundary traces. On the half-line \(\mathbb R_+\), the problem with inhomogeneous Dirichlet–Neumann data
$$
q(0,t)=g_0(t),\qquad q_x(0,t)=g_1(t)
$$
was analyzed by combining the Fokas unified transform with Fourier analysis. This yielded an explicit representation formula for the linear problem, fractional Sobolev space-time estimates, local well-posedness for \(s\in(\frac12,\frac92)\), low-regularity local well-posedness for \(0\le s<\frac12\) through boundary Strichartz estimates, and global \(H^2\) well-posedness for the defocusing model up to cubic nonlinearities [1802.10499].

On a bounded interval \((0,L)\), local well-posedness has been established for non-homogeneous Navier and Dirichlet boundary conditions. In the Navier case, the boundary traces are \(u\) and \(u_{xx}\); in the Dirichlet case, they are \(u\) and \(u_x\). The admissible boundary data lie in the optimal trace spaces \(H^{(s+3-j)/4}_{\mathrm{loc}}(\mathbb R^+)\), with compatibility conditions at \(t=0\) when the Sobolev regularity is high enough. These results use boundary integral operators and a fixed-point scheme in \(C([0,T^*];H^s(0,L))\), with an additional \(L^4\)-based component at low regularity in the Navier problem [2003.09337].

A low-regularity quarter-plane theory was developed for the cubic equation with inhomogeneous Dirichlet and Neumann data by introducing a Duhamel boundary forcing operator adapted to the fourth-order flow. The solution is constructed on the whole line and then corrected by boundary forcing terms whose coefficients solve an explicit linear system determined by boundary traces. Bourgain-type spaces \(X^{s,b}\) and the associated mixed trace space \(Z^{s,b}\) permit local well-posedness for \(0\le s<\frac12\), analytic dependence on the data, and a formulation that also suggests extensions to star graphs [1812.11079].

Control and stabilization theory on the torus \(\mathbb T\) concerns the fourth-order nonlinear Schrödinger equation
$$
i\partial_tu+\partial_x^2u-\partial_x^4u=\lambda |u|^2u
$$
with internal forcing or damping supported in an arbitrary nonempty open subset. Using Bourgain spaces, observability for the linearized adjoint system, propagation of compactness and regularity, and unique continuation, one obtains local exact controllability near the origin, global exponential stabilization under localized damping, and global exact controllability on bounded \(L^2\)-sets [1807.05264].

## 6. Growing weights, geometric settings, and present limitations

A distinct branch of the theory replaces the decaying weight \(|x|^{-b}\) by a growing factor \(|x|^b\). For
$$
i\partial_t v+\Delta^2 v=-\epsilon |x|^b|v|^{q-1}v,
$$
the main difficulty is that the source term is unbounded and translation invariance is lost in the opposite direction. The available results are radial: local well-posedness in the radial energy space \(H^2_{rd}\) for \(N\ge5\) and in \(H^1_{rd}\) for \(N\ge3\), together with small-data global well-posedness and scattering in \(H^2_{rd}\) in the energy-subcritical range. The decisive estimates are Strauss-type radial inequalities, which replace the Hardy-type tools used in decaying-weight problems; the same results also emphasize that radial symmetry is not merely cosmetic but structurally built into the argument [2511.06985].

Geometric variants arise on product manifolds. On the waveguide manifold \(\Omega_r\times\mathbb T^n\), normalized standing waves have been studied for
$$
\Delta^2 u+V(x,y)u+\lambda u=\mu |u|^{p-2}u+|u|^{q-2}u,
$$
under the mass constraint \(\int u^2=\Theta\), in the mixed regime
$$
2<p<2+\frac{8}{d+n}<q<4^*.
$$
The resulting variational picture depends on the sign of \(\mu\): for \(\mu\le0\), mountain-pass normalized solutions are produced; for \(\mu>0\), both local-minimum and mountain-pass branches occur for small masses. Orbital stability is then proved for the minimizer-type branches by the standard Cazenave–Lions strategy on the mass shell [2410.00032].

Several current limitations are explicit. In the mixed-dispersion scattering theory, the case \(\mu<0\) is not covered because the linear dispersive and Strichartz theory is only local in time. In the growing-weight model, low-dimensional cases and further scattering problems are left open, and the radial assumption remains essential. In the earlier \(\dot H^2\)-critical decaying-weight theory, the restrictions \(5\le N\le11\) and \(0<b<\frac{12-N}{N-2}\) are technical inputs of the contraction argument. A plausible implication is that future progress will continue to depend on replacing symmetry-dependent decay arguments and weight-splitting estimates by non-radial critical techniques of the kind already developed for the energy-critical whole-space equation [2211.11824], [2105.01509].

Source: https://www.emergentmind.com/topics/inhomogeneous-biharmonic-schrodinger-equation