---
title: Time-Domain Maxwell Equations
url: https://www.emergentmind.com/topics/time-domain-maxwell-equations
type: topic
---

# Time-Domain Maxwell Equations

The time-domain Maxwell equations are the fundamental hyperbolic system governing the dynamics of electric and magnetic fields in media, expressed as a coupled first-order system of partial differential equations. Their precise mathematical formulation and rigorous analysis are central to theoretical and computational electromagnetics. The subject encompasses continuous well-posedness, structure-preserving discretizations, and convergence analysis in the presence of rough data, complex geometries, and heterogeneous material parameters.

## 1. Weak Formulation and Function Space Setting

Let $\Omega \subset \mathbb{R}^3$ be a bounded Lipschitz domain and $[0,T]$ a finite time interval. Material tensors $\varepsilon, \mu, \sigma \in L^\infty(\Omega;\mathbb{R}^{3\times 3})$ are symmetric, with $\varepsilon, \mu$ uniformly elliptic (i.e., $\xi^T\varepsilon(x)\xi \ge \varepsilon_0|\xi|^2$, $\xi^T\mu(x)\xi \ge \mu_0|\xi|^2$ a.e.), and $\sigma$ positive semi-definite. For given data $f \in L^2(0,T;L^2(\Omega)^3)$ and initial fields $E_0, B_0 \in L^2(\Omega)^3$ with $\operatorname{div} B_0 = 0$, the first-order time-domain Maxwell system reads:

\[
\begin{aligned}
&\langle \varepsilon \partial_t E, \psi \rangle + (\sigma E, \psi) - (\mu^{-1} B, \operatorname{curl} \psi) = (f, \psi), \\
&\langle \partial_t B, \phi \rangle + (E, \operatorname{curl} \phi) = 0,
\end{aligned}
\]
for all $\psi \in H_0(\operatorname{curl};\Omega)$, $\phi \in H(\operatorname{curl};\Omega)$ a.e. in $(0,T)$, with $E(0)=E_0$, $B(0)=B_0$. The field regularities are $E \in H^1(0,T; H_0(\operatorname{curl};\Omega)^*) \cap L^\infty(0,T;L^2(\Omega)^3)$ and $B \in H^1(0,T; H(\operatorname{curl};\Omega)^*) \cap L^\infty(0,T; L^2(\Omega)^3)$ [2510.20752].

The function spaces are:
- $H(\operatorname{curl};\Omega)= \{v \in L^2(\Omega)^3: \operatorname{curl} v \in L^2(\Omega)^3\}$,
- $H_0(\operatorname{curl};\Omega)= \{v \in H(\operatorname{curl};\Omega): v \times n=0 \text{ on } \partial\Omega\}$,
- $H(\operatorname{div}^0;\Omega)=\{v \in H(\operatorname{div};\Omega): \operatorname{div} v=0\}$.

## 2. Well-Posedness: Existence, Uniqueness, and Energy Identity

The weak formulation admits a unique solution under the given data and boundary conditions. The argument for uniqueness utilizes interior-in-time mollification, circumventing reflection or extension techniques. For two weak solutions $(E_1,B_1)$, $(E_2,B_2)$ with identical data, setting the differences $\delta E, \delta B$ and constructing regularized mollifications, energy testing yields:

\[
\frac{d}{dt} \mathcal{E}(\delta E, \delta B) + (\sigma \delta E, \delta E) = 0,
\]
with energy functional
\[
\mathcal{E}(E, B)(t) = \frac{1}{2}\left( \|\sqrt{\varepsilon} E(t) \|^2 + \|\sqrt{\mu^{-1}} B(t)\|^2 \right).
\]
Since $\mathcal{E}(\delta E, \delta B)(0)=0$, Grönwall's lemma enforces $\delta E \equiv \delta B \equiv 0$ [2510.20752].

Existence follows via Galerkin approximation: choosing bases in $H_0(\operatorname{curl};\Omega)$ and $H(\operatorname{curl};\Omega)$, seeking finite-dimensional representations for $E_N$, $H_N$ and $B_N$, and passing to the limit using uniform energy estimates and weak-* compactness.

The (continuous) energy identity for any weak solution is:
\[
\mathcal{E}(t) + \int_0^t (\sigma E(s), E(s)) ds = \mathcal{E}(0) + \int_0^t (f(s), E(s)) ds, \qquad t \in [0,T].
\]

Furthermore, strong continuity holds: $E,B \in C([0,T];L^2(\Omega)^3)$.

The solution obeys the a priori estimate:
\[
\|\partial_t E\|_{L^2(0,T;H_0(\operatorname{curl})^*)} + \|\partial_t B\|_{L^2(0,T;H(\operatorname{curl})^*)} + \|E\|_{C([0,T];L^2)} + \|B\|_{C([0,T];L^2)} \le C(\|f\|_{L^2(0,T;L^2)} + \|E_0\|_{L^2} + \|B_0\|_{L^2}).
\]

## 3. Structure-Preserving Semi-Discrete Finite Element Scheme

On a shape-regular tetrahedral mesh $\mathcal{T}_h$, employ:
- $\mathcal{N}_h^k =$ Nédélec edge elements (conforming to $H(\operatorname{curl})$) with zero tangential trace;
- $RT_h^k =$ Raviart–Thomas face elements (conforming to $H(\operatorname{div})$);
- $\mathcal{Q}_h^k =$ piecewise polynomials in $L^2$.

The de Rham sequence is:
\[
P_h^1 \stackrel{\nabla}{\longrightarrow} \mathcal{N}_h^0 \stackrel{\operatorname{curl}}{\longrightarrow} RT_h \stackrel{\operatorname{div}}{\longrightarrow} \mathcal{Q}_h \rightarrow 0.
\]

The semi-discrete problem seeks $E_h(t) \in \mathcal{N}_h^0$, $B_h(t) \in RT_h$ for a.e. $t \in (0,T)$ such that:
\[
\begin{aligned}
&(\varepsilon \partial_t E_h, \psi_h) + (\sigma E_h, \psi_h) - (\mu^{-1} B_h, \operatorname{curl} \psi_h) = (f, \psi_h), &&\forall\psi_h\in\mathcal{N}_h^0, \\
&(\partial_t B_h, \phi_h) + (\operatorname{curl} E_h, \phi_h) = 0, &&\forall\phi_h\in RT_h.
\end{aligned}
\]

Divergence-free initial magnetic field is enforced either by reconstructing a potential $A_0\in H_0(\operatorname{curl})$ with $B_0=\operatorname{curl}A_0$ or by a constrained $L^2$ projection onto solenoidal subspaces.

The scheme preserves a discrete Gauss law:
\[
(\operatorname{div} B_h(t), q_h) = (\operatorname{div} B_h(0), q_h) \qquad \forall q_h\in \mathcal{Q}_h,\,\,\text{a.e. } t.
\]
Hence, if $B_h(0)$ is discretely divergence-free, so is $B_h(t)$ for all time [2510.20752].

The discrete energy functional
\[
\mathcal{E}_h(t) = \frac{1}{2} \left( \|\sqrt{\varepsilon} E_h(t)\|^2 + \|\sqrt{\mu^{-1}} B_h(t)\|^2 \right)
\]
satisfies
\[
\frac{d}{dt} \mathcal{E}_h(t) + \|\sqrt{\sigma} E_h(t)\|^2 = (f(t), E_h(t)).
\]
An a priori bound follows via Grönwall's inequality.

## 4. Convergence Analysis and Limit Identification

Uniform energy bounds provide weak-* compactness:
\[
\begin{aligned}
E_h &\rightharpoonup^* E \quad \text{ in } L^\infty(0,T;L^2(\Omega)^3), \\
B_h &\rightharpoonup^* B \quad \text{ in } L^\infty(0,T;L^2(\Omega)^3).
\end{aligned}
\]

Projector consistency guarantees $L^2$-strong convergence of projections:
\[
\|E_0-Q_h^N E_0\|_{L^2}\to0, \quad \|\Phi-P_h \Phi\|\to0, \quad \|u-R_h u\|_{H(\operatorname{curl})} \to 0.
\]

To pass to the limit in semidiscrete equations, one integrates in time against smooth cutoffs, uses projection consistency, and identifies time derivatives in dual spaces. The preservation of the discrete Gauss law ensures that the limiting magnetic field is solenoidal, i.e., $\operatorname{div} B=0$.

Main convergence result: as $h\to0$, $(E_h,B_h)$ converges weak-* to the unique weak solution $(E,B)$ of the continuous problem, preserving initial data and Gauss law. No convergence rates are given for purely $L^2$-data; only convergence without rate is established [2510.20752].

## 5. Auxiliary Results: Stability, Regularity, and Functional Analytic Tools

Key technical propositions include:
- H(curl) characterization: $(v, \operatorname{curl} \phi) = (g, \phi)$ for all $\phi \in C_c^\infty\implies v\in H(\operatorname{curl})$.
- Differentiability of pairings: $u \in H^1(0,T; V^*)\implies t\mapsto\langle u(t), v\rangle\in H^1(0,T)$.
- Divergence-free evolution: initial solenoidal $B(0)$ implies $\operatorname{div} B(t) = 0$ for all $t$.
- Projections: $L^2$-projection onto $RT_h$ is idempotent, contractive, and convergent; $H(\operatorname{curl})$-Riesz projection $R_h$ is orthogonal and contractive.

These results underpin identification of functional limits, passage to the time derivative in the weak sense, and preservation of constraints in the limit.

## 6. Significance and Scope

The time-domain Maxwell equations, rigorously cast in weak form and discretized using structure-preserving finite element spaces linked by the de Rham complex, provide a robust mathematical foundation for the simulation of electromagnetic phenomena in complex domains. The preservation of divergence constraints, exact discrete energy identities, and convergence to the continuous weak solution reflect the physical and geometric structure of Maxwell dynamics even in the presence of rough coefficients and minimal regularity assumptions.

The methodology and analysis in [2510.20752] make no regularity assumptions beyond $L^\infty$ coefficients and $L^2$ initial data, allowing for application in heterogeneous or nondifferentiable media common in applications. The results serve as a rigorous baseline for subsequent analysis of fully discrete schemes, error estimates under refined data regularity, and the treatment of additional physical phenomena such as lossy or nonlinear materials.

Source: https://www.emergentmind.com/topics/time-domain-maxwell-equations