--- title: Helmholtz-Schrödinger Equation Overview url: https://www.emergentmind.com/topics/helmholtz-schrodinger-equation type: topic --- # Helmholtz-Schrödinger Equation Overview The Helmholtz–Schrödinger equation denotes a family of stationary wave equations that interpolate between the Helmholtz operator and the time-independent Schrödinger operator, often at fixed frequency or fixed energy. In the literature covered here, the term is not tied to a single canonical PDE: it includes the semilinear fixed-frequency boundary value problem \[ -\Delta u-k^2u+q(x)u^2=0, \] the one-dimensional stationary Schrödinger equation \[ \psi''(x)+[k^2-V(x)]\psi(x)=0, \] and variable-coefficient formulations of \[ -\Delta \psi-k^2(x)\psi=f \] with \(k^2(x)=E-V(x)\) or \(k^2(x)=\frac{2m}{\hbar^2}(E-V(x))\) [2206.15006] [1710.07199] [1705.08139] [2505.00648]. A related paraxial usage arises when the Helmholtz equation is reduced, by a multiple-scaling ansatz, to a Schrödinger equation for a slowly varying envelope [2104.00133]. ## 1. Terminological scope and canonical forms A central feature of the terminology is that “Helmholtz” and “Schrödinger” describe the same stationary operator from different viewpoints. In one dimension, the time-independent Schrödinger equation \[ \frac{d^2}{dx^2}\psi(x)+[k^2-V(x)]\psi(x)=0 \] is explicitly of Helmholtz form, with local effective wave number \(\kappa^2(x)=k^2-V(x)\) [1710.07199]. In higher dimensions, the general Helmholtz equation is \[ \nabla^2\psi+\kappa^2(\mathbf r)\psi=0, \] so the stationary Schrödinger equation is recovered by identifying the potential with an energy shift in the coefficient \(\kappa^2\) [1710.07199]. In PDE and inverse problems, the term is used for fixed-frequency elliptic equations whose linear part is the Helmholtz operator \(-\Delta-k^2\) and whose interpretation is Schrödinger-like after rewriting \(-\Delta-k^2\) as \(-\Delta+V\) with \(V=-k^2\). The nonlinear model studied in the partial-data Calderón setting is \[ \begin{cases} -\Delta u(x)-k^2u(x)+q(x)u(x)^2=0 & \text{in }\Omega,\\ u(x)=f(x) & \text{on }\partial\Omega, \end{cases} \] with \(\Omega\subset\mathbb R^n\), \(n\ge 2\), bounded with smooth boundary, \(k\in\mathbb R\) fixed outside a discrete resonance set, and \(q\in L^\infty(\Omega)\) or more generally \(L^p\) [2206.15006]. In computational work, the same identification is used to transfer Helmholtz solvers to stationary quantum problems. The stationary Schrödinger equation \[ -\Delta \psi+V(x)\psi=E\psi \] is written as \[ -\Delta \psi-k(x)^2\psi=f,\qquad k(x)^2=E-V(x), \] or, in nonrelativistic normalization, \[ -\Delta \psi-k^2(x)\psi=0,\qquad k^2(x)=\frac{2m}{\hbar^2}(E-V(x)), \] making it a variable-coefficient Helmholtz problem [1705.08139] [2505.00648]. A common source of ambiguity is therefore terminological rather than mathematical. Inverse problems, spectral theory, numerical linear algebra, and scattering theory all use closely related stationary operators, but emphasize different structures: fixed frequency, fixed energy, semilinearity, variable refractive index, or potential recovery. This suggests that “Helmholtz–Schrödinger equation” is best understood as a class label for stationary elliptic wave equations rather than a unique normal form. ## 2. Transformations, invariants, and operator reductions A major analytical theme is the reduction of Helmholtz–Schrödinger equations to alternative normal forms. For the one-dimensional equation \[ \psi''(x)+[k^2-V(x)]\psi(x)=0, \] a similarity transformation \(\psi(z)=\phi(z)u(z)\) with \(z=z(x)\) yields \[ u_{zz}+f(z)u_z+g(z)u=0. \] Its invariant is \[ I(z)=g(z)-\frac12 f_z(z)-\frac14 f(z)^2, \] and the Schwarzian derivative of the coordinate change, \[ \{z,x\}=\frac{z'''(x)}{z'(x)}-\frac32\left(\frac{z''(x)}{z'(x)}\right)^2, \] enters the identity \[ \rho^2 I(z)+\{z,x\}=k^2-V(x)\equiv I_S(x),\qquad \rho=z'(x), \] which the paper calls the Schrödinger invariant [1710.07199]. This furnishes a constructive route to exactly solvable potentials: specify a target ODE such as the Heun equation, choose \(z(x)\), and recover \(V(x)\) from \(I_S(x)\). The same paper uses the general Heun equation as the target canonical form, \[ u_{zz}+\left(\frac{\gamma}{z}+\frac{\delta}{z-1}+\frac{\epsilon}{z-a}\right)u_z+\frac{\alpha\beta z-q}{z(z-1)(z-a)}u=0, \] and derives families of solvable potentials from the quadratic relation \[ (z')^2=\alpha_1 z^2+\beta_1 z+\gamma_1. \] Special choices produce exponential, hyperbolic, trigonometric, quadratic, and linear changes of variables, with resulting potentials described as Heun–Eckart type, Heun–Pöschl–Teller I type, Heun–Pöschl–Teller II type, rational, and Coulomb-plus-centrifugal type [1710.07199]. A different reduction appears in paraxial wave theory. Starting from the Helmholtz equation \[ \Delta v+\omega^2 v=0 \] and treating \(z\) as the evolution variable, a multiple-scaling ansatz \[ v(x,y,z)\approx e^{ik_z z}w(X,Y,Z),\qquad X=\varepsilon x,\ Y=\varepsilon y,\ Z=\varepsilon^2 z, \] with \(\omega^2=k_z^2\), leads at leading order to the Schrödinger equation \[ 2ik_z\partial_Z w=-\partial_X^2 w-\partial_Y^2 w. \] The approximation is justified in Sobolev norms with an \(O(\varepsilon)\) error on propagation distances \(z=O(\varepsilon^{-2})\), despite the fact that the Helmholtz equation is ill-posed as an initial-value problem in \(z\) without a low-transverse-frequency restriction [2104.00133]. In axially symmetric settings, the stationary equation \[ \left(\partial_r^2+\frac1r\partial_r+\partial_z^2-u(r,z)\right)Y(r,z)=0 \] admits a nonlocal Darboux transformation after the substitution \(Y=Pe^h\), with \[ u=-h_{rr}+h_r^2+\frac1r h_r+h_{zz}+h_z^2. \] The resulting first-order system involves an auxiliary nonlocal variable \(Q\), and in the special case \(V=s+2h+\ln r=0\) the transformation reduces to a generalized Moutard transformation adapted to cylindrical coordinates [2101.10654]. ## 3. Inverse boundary value theory The most developed inverse-problem formulation in the supplied corpus is the partial-data Calderón problem for the nonlinear equation \[ -\Delta u-k^2u+q(x)u^2=0 \] in a bounded smooth domain \(\Omega\subset\mathbb R^n\) [2206.15006]. For \(p>n\), a discrete exceptional set \(E\subset\mathbb R\), and sufficiently small boundary data in \(W^{2-1/p,p}(\partial\Omega)\), the forward problem \[ -\Delta u-k^2u+q(x,u)=0,\qquad u|_{\partial\Omega}=f \] has a unique solution \(u\in W^{2,p}(\Omega)\) with estimate \[ \|u\|_{W^{2,p}(\Omega)}\le C\|f\|_{W^{2-1/p,p}(\partial\Omega)}. \] The nonlinear Dirichlet-to-Neumann map is \[ \Lambda_q:f\mapsto \partial_\nu u_f|_{\partial\Omega}, \] and the partial map restricts Dirichlet data to \(\Gamma_1\subset\partial\Omega\) and Neumann observations to \(\Gamma_2\subset\partial\Omega\). The principal uniqueness theorem states that if \[ \Lambda_{q_1,\Gamma_1,\Gamma_2}(f)=\Lambda_{q_2,\Gamma_1,\Gamma_2}(f) \] on \(\Gamma_2\) for all sufficiently small \(f\in C^s(\Gamma_1)\), \(s>1\), \(s\notin\mathbb Z\), then \(q_1=q_2\) in \(\Omega\). The method is higher-order linearization. With boundary data \(f=\varepsilon_1 f_1+\varepsilon_2 f_2\), the first derivatives \[ v_j^{(l)}=\partial_{\varepsilon_l}u_j|_{(0,0)} \] solve the linear Helmholtz equation \[ (-\Delta-k^2)v_j^{(l)}=0. \] The second derivatives \[ w_j=\partial_{\varepsilon_1}\partial_{\varepsilon_2}u_j|_{(0,0)} \] satisfy \[ (-\Delta-k^2)w_j+2q_j(x)v^{(1)}v^{(2)}=0,\qquad w_j|_{\partial\Omega}=0. \] After integration by parts with a third linear solution \(v^{(0)}\), one obtains the key identity \[ \int_\Omega (q_1-q_2)\,v^{(1)}v^{(2)}v^{(0)}\,dx=0. \] The density argument needed to force \(q_1-q_2=0\) has two levels. First, a local density theorem, proved by a Dos Santos Ferreira–Kenig–Sjöstrand–Uhlmann type microlocal construction, shows that if \[ \int_\Omega q(x)\,v_1(x)v_2(x)\,dx=0 \] for all solutions of \[ (-\Delta-k^2)v=0,\qquad v|_{\partial\Omega\setminus\Gamma}=0, \] then \(q\equiv 0\) near a boundary point of \(\Gamma\). Second, Runge approximation and unique continuation propagate this vanishing along curves to the whole domain. The same strategy also yields simultaneous identification of unknown cavities and coefficients, and recovery of unknown outer boundaries together with the nonlinear coefficient [2206.15006]. ## 4. Exact solutions, separability, and solvable geometries Closed-form solutions are available in several highly structured settings. For the inhomogeneous Helmholtz equation in \(\mathbb R^3\), \[ (\Delta+k^2)u_k=\chi_{B_{x_0,r}}, \] with the Sommerfeld radiation condition, the outgoing solution can be written explicitly in terms of \(d=|x-x_0|\): \[ u_k(x)= \begin{cases} \displaystyle \frac{(i-kr)e^{ik(d-r)}-(i+kr)e^{ik(d+r)}}{2dk^3}, & x\notin B_{x_0,r},\\[3mm] \displaystyle \frac{(i+kr)\big(e^{ik(r-d)}-e^{ik(r+d)}\big)-2dk}{2dk^3}, & x\in B_{x_0,r}. \end{cases} \] The same paper derives a closed formula for the free Schrödinger evolution with initial data \(\chi_{B_{x_0,r}}\), expressed by complex error functions, and extends these exact ball-data solutions to approximation formulas for general radial \(H^1\)-data by shell decomposition [2108.08108]. Separation of variables can also be generalized beyond classical orthogonal systems. Polyelliptic coordinates are built by gluing several local elliptic coordinate systems associated with the sides of a convex polygon, producing a global orthogonal coordinate system outside the polygonal “protected area.” The metric assumes Stäckel form, \[ H_{\theta_c}^2=g_1(\theta_c)\bigl[h_1(\mu_c)+h_2(\theta_c)\bigr],\qquad H_{\mu_c}^2=g_2(\mu_c)\bigl[h_1(\mu_c)+h_2(\theta_c)\bigr], \] which yields separated ODEs for the 2D Helmholtz equation and, by extension, for the time-independent Schrödinger equation with piecewise constant potential [1409.6761]. This framework is proposed for polygonal scattering and finite-wall polygonal quantum wells, including the square well. Exactly solvable potential construction via Heun functions fits naturally into the same landscape. Choosing the Heun equation as canonical target and using the invariant relation \[ I_S(x)=\rho^2 I_H(z)+\{z,x\} \] produces explicit potential families with Heun-function wavefunctions \[ \psi(x)=\rho(x)^{-1/2}[z(x)]^{\gamma/2}[z(x)-1]^{\delta/2}[z(x)-a]^{\epsilon/2} \,Hl(a,q;\alpha,\beta,\gamma,\delta,\epsilon;z(x)) \] [1710.07199]. ## 5. Numerical solution methods and computational formulations High-frequency Helmholtz–Schrödinger problems are dominated by indefiniteness, oscillation, and dispersion error. Domain decomposition methods address this through absorbed local problems and coarse corrections. For the interior Helmholtz problem \[ -\Delta u-(k^2+i\varepsilon)u=f, \qquad \partial_n u-i\eta u=0, \] two-level ORAS/ImpRAS preconditioners combine local solves with a coarse space \(Z\), either from an absorbed coarse mesh or from local Dirichlet-to-Neumann eigenproblems [1705.08139]. In the DtN construction, interface modes satisfy \[ \big(A_{\Gamma_i\Gamma_i}^{(i)}-A_{\Gamma_i I}A_{II}^{-1}A_{I\Gamma_i}\big)\mathbf u =\lambda M_{\Gamma_i}\mathbf u, \] and modes with \(\Re(\lambda_j)