---
title: Hemispherical TE Mode Cavity
url: https://www.emergentmind.com/topics/hemispherical-transverse-electric-mode-cavity
type: topic
---

# Hemispherical TE Mode Cavity

Searching arXiv for the specified papers and closely related work on hemispherical TE-mode cavities.
A hemispherical transverse-electric mode cavity is an electromagnetic resonator whose confining boundary includes a hemispherical surface and whose operative mode is transverse-electric in the relevant formulation. In closed spherical-coordinate treatments, this means TE-to-\(r\) modes with \(E_r=0\); in monolithic semiconductor hemispherical microcavities, it refers to a TE-like fundamental transverse mode whose electric field is predominantly tangential to the planar mirror and transverse to the optical axis. The concept spans several distinct but related implementations: perfectly conducting hemispheres with an equatorial conducting plane, layered spherical or hemispherical conducting cavities with dielectrics, and open plano-concave semiconductor Fabry–Perot microcavities used for single-photon extraction. Across these realizations, the hemispherical boundary is used to control angular spectra, suppress higher-order modes, shape Gaussian far fields, and tune the balance among confinement, radiation loss, and usable output coupling [2512.18498] [1103.1240] [1905.10181] [1810.10785].

## 1. Geometric archetypes and TE definitions

The subject encompasses both closed and open cavities. In a perfectly conducting hemispherical cavity, the boundary consists of a spherical shell \(r=R\) and a perfectly conducting equatorial plane \(z=0\), with a homogeneous medium \((\epsilon,\mu)\) inside. In that setting the natural modal definition is TE-to-\(r\): the electric field has no radial component, while the magnetic field generally has both radial and tangential components [2512.18498].

A second archetype is the layered conducting cavity derived from a full spherical resonator filled concentrically with dielectric or vacuum layers and then adapted to the hemisphere by adding a perfectly conducting equatorial plane. In the spherical parent problem, TE modes have no radial electric field and are represented with vector spherical harmonics; the hemisphere is obtained by imposing the equatorial boundary condition on the angular part of the spherical solution [1103.1240].

A third archetype is the monolithic semiconductor hemispherical microcavity. There the geometry is plano–concave rather than fully closed: a planar back mirror lies below the emitter, and a curved top mirror is etched into the semiconductor surface. The device is realized as a GaAs slab of thickness \(L\) hosting an emitter, with a hemisphere of radius of curvature \(R\) milled into the top surface. The paper emphasizes TE-like coupling rather than a strict spherical-coordinate TE-to-\(r\) classification, because the relevant mode is the fundamental transverse mode of a hemispherical Fabry–Perot cavity and the electric field is predominantly tangential to the planar interface [1905.10181].

These realizations are linked by a common design objective: to isolate a small set of transverse-electric or TE-like modes whose field geometry is compatible with the boundary conditions and useful output channel. A plausible implication is that “hemispherical TE cavity” is best understood as a family resemblance among hemispherical resonators, rather than a single canonical device class.

## 2. Angular spectra, parity selection, and radial quantization

In the closed homogeneous hemispherical cavity, the TE problem is formulated from a scalar Debye potential \(\Phi(r,\theta,\phi)=R(r)\Theta(\theta)e^{im\phi}\) satisfying the scalar Helmholtz equation. The separated radial and angular equations are
$$
r^2 R'' + 2 r R' + [k^2 r^2 - \nu(\nu+1)] R = 0,
$$
and
$$
\frac{1}{\sin\theta}\frac{d}{d\theta}\!\left(\sin\theta \frac{d\Theta}{d\theta}\right) + \left[\nu(\nu+1)-\frac{m^2}{\sin^2\theta}\right]\Theta = 0,
$$
with \(k=\omega\sqrt{\epsilon\mu}\). The TE fields are generated by \(E=i\omega\mu\nabla\times(r\Phi)\) and \(H=(1/(i\omega\mu))\nabla\times E\), which makes \(E_r=0\) identically [2512.18498].

The north-pole endpoint condition selects the regular associated-Legendre branch \(\Theta(\theta)=P_\nu^m(\cos\theta)\). At the equatorial perfectly conducting plane \(\theta=\pi/2\), TE implies \(E_r=0\) already, so the tangential-field condition reduces to \(E_\phi(\pi/2)=0\), equivalently
$$
\left.\frac{d\Theta}{d\theta}\right|_{\theta=\pi/2}=0,
$$
or
$$
P_\nu^{m\,\prime}(0)=0.
$$
This defines a self-adjoint Sturm–Liouville problem on \([0,\pi/2]\) for each integer \(m\), and the resulting \(\nu\) spectrum is discrete but need not be integer [2512.18498].

For the integer subset \(\nu=\ell\), parity gives a simple selection rule:
$$
P_\ell^{m\,\prime}(0)=0 \quad \text{if and only if} \quad \ell+m \text{ is even}.
$$
Hence zonal modes with \(m=0\) survive only for even \(\ell\), sectoral modes with \(\ell=m\) are always admissible, and tesseral modes are admissible when \(\ell+m\) is even. The sectoral case \(\nu=m\) is especially simple, because \(\Theta(\theta)=\sin^m\theta\) solves the angular equation exactly for any real \(m>0\) and automatically satisfies the equatorial Neumann condition [2512.18498].

The radial solution regular at the origin is \(R(r)=A j_\nu(kr)\), and the perfectly conducting spherical wall requires the tangential electric field to vanish at \(r=R\). For TE-to-\(r\) modes this gives
$$
j_\nu(kR)=0,
$$
so that
$$
k_{\nu,n}R=x_{\nu,n}, \qquad f_{\nu,n}=\frac{c}{2\pi R}x_{\nu,n},
$$
with \(x_{\nu,n}\) the zeros of \(j_\nu\). The lowest nontrivial TE resonance reported for the hemispherical cavity is \(TE_{1,1,1}\), with \(kR\approx 4.493409\) [2512.18498].

An important clarification follows from comparing the closed-form hemisphere analysis with the earlier layered-sphere adaptation. The earlier construction identified \(m=0\) and even \(l\) as the simplest rigorous subset obtained directly from spherical \(TE^{l0}\) modes that satisfy the equatorial boundary [1103.1240]. The later complete angular analysis showed that this is not the full integer spectrum: the exact integer condition is \(\ell+m\) even, and additional discrete non-integer \(\nu_q(m)\) generally exist on the hemisphere [2512.18498]. This resolves the misconception that angular indices must remain integer merely because the underlying equations are Maxwellian; the integer restriction on the full sphere is tied to full-domain endpoint constraints rather than to Maxwell’s equations alone.

## 3. Open hemispherical semiconductor microcavities

The semiconductor realization is a monolithic hemispherical Fabry–Perot microcavity designed for enhanced light–matter interaction and photon extraction from an embedded quantum emitter such as an InAs/GaAs quantum dot or a diamond color center. A hemisphere of radius of curvature \(R\) is milled into the top surface of a GaAs slab of thickness \(L\), with refractive index \(n\approx 3.48\) at \(\sim 940\) nm. The cavity is plano–concave: a flat back mirror lies below the emitter and a curved top mirror lies above it. All three device classes reported use \(R=2L\), which is within the stability range \(0<L<R\) and is identified as the choice that minimizes far-field divergence while maintaining cavity stability [1905.10181].

The mirror design sets the finesse regime. The low-\(Q\) structure uses a gold planar back mirror and the bare GaAs–air interface as the top curved surface. The medium-\(Q\) structure keeps the gold back mirror but adds a conformal 5-layer top DBR. The high-\(Q\) structure uses DBRs on both sides, with a 9-layer top DBR and a 20-layer bottom DBR, all centered at 914 nm. The shallow hemispherical etch is characterized by \(L-h\), chosen to be small, \(\le 190\) nm in the designs, so that higher-order Laguerre–Gaussian modes are clipped more strongly than the fundamental while the latter is not apodized [1905.10181].

For the fundamental mode, the waist is located at the planar mirror. The Gaussian approximation used in the paper gives
$$
w_0^2 \approx \frac{\lambda L}{\pi n}\sqrt{\frac{R-L}{L}},
$$
which reduces to \(w_0^2\approx (\lambda L)/(\pi n)\) when \(R=2L\). In the medium-\(Q\) structure, FDTD finds a mode radius of approximately \(238\) nm at the waist. The beam radius at the curved mirror is
$$
w(R)=w_0\sqrt{1+\left(\frac{L}{z_R}\right)^2},
$$
with \(z_R=\pi n w_0^2/\lambda\), and for \(R=2L\) the paper states that \(z_R\approx L\), so \(w(R)\approx w_0\sqrt{2}\) [1905.10181].

| Configuration | Mirrors and geometry | Reported values |
|---|---|---|
| Low-\(Q\) | Gold back mirror; GaAs–air top surface; \(L=1572\) nm; \(h=1375\) nm | \(\Delta\lambda \approx 14\) nm; \(\eta_{\mathrm{ext}}\approx 0.3\); \(F_P\approx 2.2\); FOM \(\approx 0.57\) |
| Medium-\(Q\) | Gold back mirror; 5-layer top DBR; \(L=1502\) nm; \(h=1312\) nm | \(\Delta\lambda \approx 3.3\) nm; \(\eta_{\mathrm{ext}}\approx 0.5\); \(F_P\approx 2.4\); FOM \(\approx 1.3\) |
| High-\(Q\) | 20-layer bottom DBR; 9-layer top DBR; \(L=1600\) nm; \(h=1200\) nm | \(\Delta\lambda \approx 0.95\) nm; \(\eta_{\mathrm{ext}}\approx 0.6\); \(F_P\approx 6\); FOM \(\approx 1.85\) |

The free spectral range is
$$
\mathrm{FSR}=\frac{c}{2nL},
$$
and for \(L\approx 1.5\,\mu\mathrm{m}\) and \(n\approx 3.48\) the paper reports \(\mathrm{FSR}\approx 2.9\times 10^{13}\) Hz, corresponding near 940 nm to \(\Delta\lambda\approx 80\)–85 nm when DBR penetration is neglected [1905.10181].

## 4. Extraction physics, polarization, and mode purity

The operative mode in the semiconductor device is the fundamental transverse mode \(TEM_{q,0,0}\), which FDTD shows to have a highly directional Gaussian \(k\)-space profile. In the medium-\(Q\) case, the far field is Gaussian with angular spread \(\pm 25^\circ\), corresponding to \(\mathrm{NA}\approx 0.42\), and the simulations show strong collection within numerical apertures up to \(\mathrm{NA}\approx 0.68\). The paper attributes up to a 20-fold improvement in figure of merit relative to an unetched planar structure to this reshaping of the emission into a fundamental Gaussian mode [1905.10181].

The spectral trade-off is set by finesse and reflectivity. For a two-mirror cavity with power reflectivities \(R_1\) and \(R_2\),
$$
F=\frac{\pi\sqrt{R_1R_2}}{1-R_1R_2},
$$
and
$$
Q=\frac{\nu}{\Delta\nu}=F\,\frac{\nu}{\mathrm{FSR}}=\left(\frac{2nL}{\lambda}\right)F.
$$
Increasing reflectivity raises \(F\), narrows the linewidth, and increases \(Q\), but it also reduces bandwidth and can increase large-angle loss at the top DBR in the high-\(Q\) configuration. This is reflected directly in the reported trend \(\Delta\lambda\approx 14\) nm \(\rightarrow 0.95\) nm and \(F_P\approx 2.2 \rightarrow 6\) as the mirror stack is strengthened [1905.10181].

The Purcell factor is expressed as
$$
F_P=\frac{3}{4\pi^2}\left(\frac{\lambda}{n}\right)^3 \frac{Q}{V},
$$
with
$$
V=\frac{\int \epsilon(\mathbf r)|\mathbf E(\mathbf r)|^2\,d^3r}{\max(\epsilon(\mathbf r)|\mathbf E(\mathbf r)|^2)}.
$$
For resonant coupling to a single cavity mode, the paper gives the estimate
$$
\beta \approx \frac{F_P}{F_P+1},
$$
which places \(\beta\) roughly between \(0.69\) for the low-\(Q\) design and \(0.86\) for the high-\(Q\) design. Extraction into the useful far-field depends not only on \(\beta\) but also on mirror transmissions, so \(\eta_{\mathrm{ext}}\) rises from approximately \(0.3\) to \(0.5\) between low- and medium-\(Q\) as in-plane guided modes are suppressed, then saturates near \(0.6\) in the high-\(Q\) design because of DBR loss at large angles [1905.10181].

Polarization behavior is central to the TE-like designation. In a high-index semiconductor with DBR mirrors, polarization anisotropy can split the fundamental into two orthogonal linear polarizations. The TE-like mode is typically favored by the boundary conditions at the planar interface and by the DBR polarization response, and the simulations use an emitter dipole in the \(XZ\) plane to match a TE-like coupling scenario. Strain, DBR birefringence, and slight mirror figure errors can split the fundamental mode; the paper states that this splitting is benign in low- and medium-\(Q\) cavities because the cavity bandwidth exceeds the splitting, but that in narrowband high-\(Q\) cavities the splitting can exceed the linewidth, which precludes robust circular polarization and complicates spin-selective protocols [1905.10181].

## 5. Perturbation theory for open hemispherical TE cavities

For open dielectric cavities, the transverse-electric polarization can be formulated rigorously through a scalar Helmholtz equation with outgoing radiation:
$$
\Delta u + \omega^2 \epsilon(x)\mu\,u = 0 \quad \text{in } \mathbb R^d,
$$
together with the outgoing radiation condition. With \(\Omega\subset\mathbb R^d\) the cavity domain, the volume integral operator
$$
K_\Omega[u](x)=-\int_\Omega u(y)\,\Gamma_m(x-y;\omega)\,dy
$$
gives the Lippmann–Schwinger form
$$
(1-\omega^2\tau\epsilon_c\mu_m K_\Omega)[u]=0.
$$
A scattering resonance \(\omega_0\) is defined by the existence of \(j_0\) such that \(1-\omega^2\tau\epsilon_c\mu_m\lambda_{j_0}(\omega_0)=0\), where \(\lambda_j(\omega)\) is an eigenvalue of \(K_\Omega\) [1810.10785].

Near a non-exceptional resonance, the outgoing Green’s function admits a pole-pencil decomposition
$$
G(x,y;\omega)=\Gamma_m(x-y;\omega)+C_{j_0}(\omega)e_{j_0}(x;\omega)e_{j_0}(y;\omega)\frac{1}{\omega-\omega_0}+R(x,y,\omega),
$$
which separates the resonant pole from the analytic background. Because the cavity is open, \(\omega_0\), \(C_{j_0}\), and the mode \(e_{j_0}\) are complex, so the perturbative frequency shift is also complex and encodes both resonance shift and linewidth change [1810.10785].

For a small internal particle \(D=z+\delta B\subset\Omega\) with permeability contrast, the leading-order small-volume shift is
$$
\omega_s-\omega_0=C_{j_0}(\omega_0)M(\mu_m/\mu_c,B)\,\nabla e_{j_0}(z;\omega_0)\cdot \nabla e_{j_0}(z;\omega_0),
$$
where \(M(\mu_m/\mu_c,B)\) is the magnetic polarization tensor. For a particle outside the cavity, the analogous formula uses the field
$$
g_{j_0}(x;\omega)=\omega^2\tau\epsilon_c\mu_m\int_\Omega e_{j_0}(z;\omega)\Gamma_m(z,x;\omega)\,dz
$$
and becomes
$$
\omega_s-\omega_0=\delta^d C_{j_0}(\omega_0)M(\mu_m/\mu_c,B)\,\nabla g_{j_0}(z;\omega_0)\cdot \nabla g_{j_0}(z;\omega_0).
$$
The theory is geometry-agnostic, so it applies to hemispherical open cavities once the unperturbed TE mode and the cavity Green’s function are known [1810.10785].

The paper also analyzes two cases that are especially relevant to high-sensitivity hemispherical resonators. First, plasmonic or dispersive particles can strongly enhance shifts when the particle resonance nearly coincides with the cavity resonance. Second, exceptional scattering resonances require a higher-order-pole analysis rather than the simple-resonance formula. A plausible implication is that hemispherical TE cavities, especially those with strong field gradients near curved boundaries or apertures, are natural platforms for non-Hermitian perturbation engineering, but only within the validity regime \(D\cap\partial\Omega=\varnothing\), \(\delta\to 0\), and isolated \(\omega_0\) [1810.10785].

## 6. Layered conducting and dielectric hemispheres, \(Q\), and accelerator use

The layered conducting-cavity analysis begins from a full spherical conducting shell containing dielectric or vacuum layers. In Gaussian units with \(\mu=1\), harmonic fields proportional to \(e^{-i\omega t}\) satisfy
$$
\nabla\times \mathcal E = ik\mathcal B, \qquad \nabla\times \mathcal B=-ik\epsilon \mathcal E, \qquad k=\omega/c.
$$
For TE modes, \(\mathcal E\) has only tangential components and \(\mathcal B\) has both tangential and radial components. In a uniform dielectric layer,
$$
\mathcal E^{lm}(r,\omega)=\mathcal E_t(kr)\,\mathcal X_{lm}(\theta,\phi),
$$
with
$$
\mathcal E_t(r)=A_l^{(1)}j_l(\sqrt\epsilon kr)+A_l^{(2)}y_l(\sqrt\epsilon kr),
$$
and the tangential and radial magnetic components are expressed through \(j_l^D\) and \(y_l^D\), the derivatives of the Riccati–Bessel and Riccati–Neumann functions [1103.1240].

The perfectly conducting wall at \(r=b\) imposes, for TE modes,
$$
A_{lN}^{(1)} j_l(\sqrt{\epsilon_N}kb)+A_{lN}^{(2)} y_l(\sqrt{\epsilon_N}kb)=0.
$$
At dielectric interfaces, continuity of tangential \(\mathcal E\) and tangential \(\mathcal B\) yields a transfer-matrix recurrence. For a completely filled sphere the TE frequencies are determined by \(j_l(\sqrt{\epsilon_1}kb)=0\). For the special case of a dielectric sphere of radius \(a\) inside vacuum, the paper gives an explicit two-region characteristic equation for the TE modes, which is solved numerically [1103.1240].

Losses are summarized by
$$
Q=\frac{\omega_0 U}{P}\equiv \frac{2\pi U}{T_0P},
$$
with
$$
U=\frac{1}{16\pi}\int_V(\epsilon'|\mathcal E|^2+|\mathcal H|^2)\,dv.
$$
The metallic-wall contribution is
$$
Q_{\mathrm{met}}=\frac{2}{\delta}\frac{\int_V |\mathcal B|^2\,dv}{\int_S |\mathcal B|^2\,ds},
$$
while dielectric loss gives
$$
Q_{\mathrm{diel}}=-\frac{\omega_0}{2\,\mathrm{Im}\,\omega},
$$
and the total quality factor is
$$
\frac{1}{Q_s}=\frac{1}{Q_{\mathrm{met}}}+\frac{1}{Q_{\mathrm{diel}}}.
$$
For \(\epsilon'=10\) and \(\epsilon''=10^{-6}\), the paper reports \(Q_s\approx Q_{\mathrm{diel}}\) and values of order \(10^7\), three orders of magnitude greater than the cylindrical pill-box TM cavity value \(Q_c\approx 1.633\times 10^4\) for copper with \(R_c=h=d=4\) cm [1103.1240].

The hemispherical version is obtained by adding a perfectly conducting plane at the equator. The curved hemispherical wall obeys the same radial boundary condition as in the sphere; the equatorial plane imposes vanishing tangential electric field there. In the adaptation developed from the spherical solutions, the simplest exact subset is obtained from spherical \(TE^{l0}\) modes with even \(l\), which automatically satisfy the plane boundary. The paper remarks that the flat equatorial plane offers straightforward mechanical surfaces for coupling apertures, feeds, and beam ports, while the tangential electric field vanishes on the plane, reducing risk of breakdown there [1103.1240].

The same study uses spherical TE oscillations with a central dielectric sphere to accelerate strongly relativistic electrons on a nearly tangential trajectory. For a specific geometry \(a=0.708\) cm, \(b=2.124\) cm, and \((l,m)=(6,-2)\), with normalization \(U=1/(8\pi)\) erg, the reference effective accelerating field is reported as \(E_{\mathrm{eff}}^{\mathrm{ref}}=1.2366\) kV/m. The transverse deflections are small and obey the scaling
$$
\alpha = 5.9\times 10^{-4}\,\frac{E_{\mathrm{eff}}/E_{\mathrm{eff}}^{\mathrm{ref}}}{\gamma}\ \mathrm{cm},
$$
and the minimum electron energy for exit displacement constrained to \(p\%\) of the trajectory clearance is
$$
m_e c^2[\mathrm{GeV}] = \frac{20.69}{p}\left(\frac{E_{\mathrm{eff}}[\mathrm{MV/m}]}{100}\right).
$$
This accelerator context is distinct from single-photon extraction, but it shows that hemispherical TE cavities are not confined to quantum-optical use [1103.1240].

## 7. Fabrication, imperfections, and experimental characterization

The semiconductor hemispherical microcavities were fabricated monolithically in GaAs by water-assisted FIB milling at \(50\) pA, with target etch depths of \(150\)–\(190\) nm and milling time of approximately \(20\) minutes per device. AFM profiles were fitted to a hemisphere-on-plane model
$$
h(x,y)=z_s+\sqrt{R^2-r^2}\quad \text{if } (z_s+\sqrt{R^2-r^2})>z_p; \qquad h(x,y)=z_p \quad \text{otherwise},
$$
where \(z_s\) is the sphere-center position, \(z_p\) the plane \(z\)-position, and \(r\) the radial coordinate. A representative device yielded \(R\approx 3.2\,\mu\mathrm m\), etch depth \(151\) nm, and RMS roughness approximately \(4\) nm [1905.10181].

The same work analyzed fabrication tolerances numerically. In the medium-\(Q\) design, source offsets up to \(100\) nm from the cavity axis caused only slight changes in far-field pointing angle and minimal degradation of \(F_P\). A Monte Carlo analysis with RMS roughness of \(5\) nm, implemented through a 2D FDTD proxy, showed that performance remains close to optimum with high probability; the measured roughness of approximately \(4\) nm is well below the optical wavelength and is therefore consistent with low scattering loss and preserved mode purity. Mirror figure errors and height variations can induce polarization splitting and minor TE/TM mixing, particularly in high-\(Q\) structures where the linewidth is narrow, while low- and medium-\(Q\) cavities mitigate those effects through broader spectral acceptance [1905.10181].

Optical characterization used differential reflectance spectroscopy in a confocal microscope with a fiber-coupled IR LED centered at approximately \(940\) nm. Low-\(Q\) devices exhibited broad resonances near \(920\) nm and \(960\) nm with FWHM approximately \(24\) nm and \(31\) nm. These resonance dips were associated with the fundamental axial resonances predicted by FDTD, with an approximately \(20\) nm redshift relative to simulation attributed to fabrication tolerances. Simulated \(k\)-space emission maps showed a Gaussian far field consistent with the fundamental transverse mode. Polarization-resolved measurements were not reported, so the experimental confirmation of TE-mode behavior is indirect, resting on the measured cavity geometry, the observed resonances, and the simulated resilience of low- and medium-\(Q\) designs to polarization splitting [1905.10181].

Taken together, these results place hemispherical TE cavities at the intersection of exact angular-spectrum theory, open-cavity non-Hermitian resonance analysis, and fabrication-driven mode engineering. Closed hemispherical PEC cavities support a parity-selected and, in general, non-integer angular spectrum with TE radial quantization \(j_\nu(kR)=0\) [2512.18498]. Layered conducting hemispheres inherit the spherical transfer-matrix structure and can achieve total quality factors dominated by dielectric loss at the \(10^7\) level [1103.1240]. Open semiconductor hemispherical microcavities exploit a shallow plano–concave geometry to suppress higher-order modes, concentrate emission into a fundamental TE-like Gaussian beam, and balance bandwidth, Purcell enhancement, and extraction efficiency for embedded quantum emitters [1905.10181]. Open-cavity perturbation theory then supplies explicit formulas for how small particles shift those TE resonances, including the roles of radiation leakage, plasmonic enhancement, and exceptional points [1810.10785].

Source: https://www.emergentmind.com/topics/hemispherical-transverse-electric-mode-cavity