3.1 Parallel plate waveguides

Parallel plate waveguide basics

Parallel plate waveguides are one of the simplest structures for guiding electromagnetic waves: two parallel conducting plates separated by a dielectric medium. Despite their simplicity, they illustrate all the core concepts you'll need for more complex waveguide geometries.

The dimensions and material properties of the waveguide determine which modes can propagate and at what frequencies. This section covers the structure, boundary conditions, and the three families of propagation modes (TEM, TE, and TM).

Waveguide structure and dimensions

A parallel plate waveguide consists of two conducting plates separated by a distance $d$, extending infinitely in the transverse directions (perpendicular to wave propagation). In practice, "infinitely" means the plate width $w$ is much larger than $d$, so edge effects are negligible and fields are uniform across the width.

• Plate separation $d$: sets the cutoff frequencies for TE and TM modes
• Plate width $w \gg d$: ensures field uniformity along the width direction
• Waveguide length $L$: the propagation distance of interest

Boundary conditions in waveguides

The conducting plates impose constraints on the fields inside the guide. For perfect electric conductors (PEC):

• The tangential electric field must vanish at each plate surface. A perfect conductor cannot sustain a tangential E-field.
• The normal magnetic field must vanish at each plate surface.

These two conditions are what select the allowed mode shapes and their discrete cutoff frequencies. Every field solution inside the waveguide must satisfy them simultaneously.

Modes of propagation

Parallel plate waveguides support three families of modes, classified by which field components lie purely transverse (perpendicular) to the propagation direction $\hat{z}$:

• TEM (Transverse Electromagnetic): Both $\vec{E}$ and $\vec{H}$ are entirely transverse. No longitudinal field components.
• TE (Transverse Electric): $\vec{E}$ is purely transverse, but $\vec{H}$ has a longitudinal ($z$) component.
• TM (Transverse Magnetic): $\vec{H}$ is purely transverse, but $\vec{E}$ has a longitudinal ($z$) component.

Each TE and TM mode has a cutoff frequency; below it, the mode is evanescent and cannot carry power. The TEM mode has no cutoff. As frequency increases, more modes "turn on," so the number of propagating modes grows with frequency and plate separation.

Transverse electromagnetic (TEM) mode

The TEM mode is the fundamental mode of the parallel plate waveguide. Because it has no cutoff frequency, it propagates at all frequencies, making it the dominant mode for broadband applications and the easiest to analyze.

Electric and magnetic field distributions

In the TEM mode, the fields are uniform between the plates and have no variation in the transverse direction:

• $E_y = E_0 \, e^{-j\beta z}$ (perpendicular to the plates)
• $H_x = \frac{E_0}{\eta} \, e^{-j\beta z}$ (parallel to the plates)

Here $\eta = \sqrt{\mu/\epsilon}$ is the intrinsic impedance of the dielectric filling. The field structure looks exactly like a uniform plane wave, just confined between two plates.

Propagation constant and phase velocity

For the TEM mode, the propagation constant is purely imaginary (no attenuation in the lossless case):

$\gamma = j\beta, \quad \beta = \omega\sqrt{\mu\epsilon}$

The phase velocity is:

$v_p = \frac{\omega}{\beta} = \frac{1}{\sqrt{\mu\epsilon}}$

This equals the speed of light in the dielectric medium. For free space, $v_p = c = 1/\sqrt{\mu_0 \epsilon_0}$. Notice that the TEM phase velocity is independent of frequency, so TEM propagation is non-dispersive.

Characteristic impedance

The characteristic impedance relates voltage to current for a propagating TEM wave:

$Z_0 = \eta \frac{d}{w}$

where $\eta$ is the intrinsic impedance of the dielectric, $d$ is the plate separation, and $w$ is the plate width. Matching $Z_0$ to source and load impedances minimizes reflections and maximizes power transfer.

Power flow and Poynting vector

The time-averaged Poynting vector for the TEM mode points along $\hat{z}$:

$\vec{S}_{avg} = \frac{1}{2} \text{Re}(\vec{E} \times \vec{H}^*) = \frac{1}{2} \frac{E_0^2}{\eta} \hat{z}$

Integrating over the cross-sectional area $w \times d$ gives the total average power:

$P_{avg} = \frac{1}{2} \frac{E_0^2}{\eta} \, w \, d$

The maximum power the guide can handle is limited by the dielectric breakdown field strength and thermal effects from conductor/dielectric losses.

Transverse electric (TE) modes

TE modes have $E_z = 0$ everywhere. The electric field is purely transverse, while the magnetic field has both transverse and longitudinal components. These modes only propagate above their respective cutoff frequencies.

TE mode field equations

For the $\text{TE}_{m0}$ mode (with plates at $x = 0$ and $x = d$, propagation along $z$):

$E_y = E_0 \sin\!\left(\frac{m\pi x}{d}\right) e^{-j\beta z}$

The transverse wavenumber is $k_x = m\pi/d$, where $m = 1, 2, 3, \ldots$ is the mode index. The magnetic field components are derived from $E_y$:

$H_x = -\frac{\beta}{\omega\mu} E_0 \sin\!\left(\frac{m\pi x}{d}\right) e^{-j\beta z}$

$H_z = -\frac{jk_x}{\omega\mu} E_0 \cos\!\left(\frac{m\pi x}{d}\right) e^{-j\beta z}$

The $\sin$ dependence in $E_y$ ensures the tangential electric field vanishes at both plates, satisfying the boundary conditions.

Cutoff frequencies for TE modes

Each $\text{TE}_{m0}$ mode has a cutoff frequency:

$f_{c,m} = \frac{m \, c}{2d\sqrt{\epsilon_r}}$

where $c$ is the speed of light in vacuum and $\epsilon_r$ is the relative permittivity of the dielectric. The $\text{TE}_{10}$ mode has the lowest cutoff. Below $f_{c,m}$, the mode is evanescent ($\beta$ becomes imaginary and the fields decay exponentially).

Dispersion relation and phase velocity

The dispersion relation for TE modes is:

$\beta = \sqrt{k^2 - k_x^2} = k\sqrt{1 - \left(\frac{f_c}{f}\right)^2}$

where $k = \omega\sqrt{\mu\epsilon}$. The phase velocity is:

$v_p = \frac{\omega}{\beta} = \frac{v}{\sqrt{1 - (f_c/f)^2}}$

Here $v = 1/\sqrt{\mu\epsilon}$ is the speed of light in the medium. Notice that $v_p > v$ always for $f > f_c$. This doesn't violate relativity because energy travels at the group velocity $v_g = v^2/v_p$, which is always less than $v$. The relation $v_p \cdot v_g = v^2$ is a useful check.

Attenuation constant vs frequency

In real waveguides with finite conductivity, the attenuation constant for TE modes due to conductor losses is:

$\alpha_{TE} = \frac{2R_s}{\eta \, d} \cdot \frac{1}{\sqrt{1 - (f_c/f)^2}}$

where $R_s = \sqrt{\omega\mu_c / (2\sigma)}$ is the surface resistance of the conductor (with conductivity $\sigma$ and permeability $\mu_c$). Two trends to note:

• As $f \to f_c^+$, $\alpha \to \infty$. Losses spike near cutoff because the group velocity approaches zero and energy spends more time interacting with the lossy walls.
• At frequencies well above cutoff, $\alpha$ decreases and the mode propagates more efficiently.

Transverse magnetic (TM) modes

TM modes have $H_z = 0$ everywhere. The magnetic field is purely transverse, while the electric field has both transverse and longitudinal components. Like TE modes, TM modes have cutoff frequencies.

TM mode field equations

For the $\text{TM}_{m0}$ mode:

$H_y = H_0 \cos\!\left(\frac{m\pi x}{d}\right) e^{-j\beta z}$

The electric field components are:

$E_x = \frac{\beta}{\omega\epsilon} H_0 \cos\!\left(\frac{m\pi x}{d}\right) e^{-j\beta z}$

$E_z = \frac{jk_x}{\omega\epsilon} H_0 \sin\!\left(\frac{m\pi x}{d}\right) e^{-j\beta z}$

The $\sin$ dependence in $E_z$ ensures it vanishes at the plates, while $E_x$ (normal to the plates) need not vanish there. The $\cos$ dependence in $H_y$ satisfies the condition that the normal component of $\vec{H}$ is zero at the conductors.

Cutoff frequencies for TM modes

The cutoff frequencies for TM modes are identical to those of TE modes:

$f_{c,m} = \frac{m \, c}{2d\sqrt{\epsilon_r}}$

This degeneracy between $\text{TE}_{m0}$ and $\text{TM}_{m0}$ cutoff frequencies is specific to the parallel plate geometry. In rectangular waveguides, TE and TM modes of the same index generally have different cutoffs.

Dispersion relation and phase velocity

The dispersion relation is the same as for TE modes:

$\beta = \sqrt{k^2 - k_x^2} = k\sqrt{1 - \left(\frac{f_c}{f}\right)^2}$

The phase and group velocities follow the same formulas as the TE case. The difference between TE and TM modes lies in the field distributions and impedance, not in the dispersion behavior.

Attenuation constant vs frequency

The conductor-loss attenuation for TM modes behaves differently from TE modes:

$\alpha_{TM} = \frac{2R_s}{\eta \, d} \cdot \sqrt{1 - (f_c/f)^2}$

Compare this with the TE result. For TM modes, $\alpha$ actually decreases as $f \to f_c^+$ (going to zero at cutoff) and increases at higher frequencies. This is the opposite trend from TE modes. The physical reason is that the wall currents for TM modes vanish at cutoff, so there's no conductor loss there. This distinction between TE and TM attenuation behavior is a common exam topic.

Higher-order modes

Beyond the fundamental TEM mode and the lowest TE/TM modes, parallel plate waveguides support an infinite set of higher-order modes with increasingly complex field patterns. Each higher mode has a higher cutoff frequency.

Mode excitation and coupling

Higher-order modes can be excited by:

• Asymmetric excitation at the waveguide input (e.g., a feed that doesn't match the TEM field pattern)
• Discontinuities in the waveguide, such as steps, bends, or obstacles
• Coupling from external sources or adjacent waveguides

Mode coupling occurs when energy transfers between modes due to structural irregularities or field overlaps at junctions. Careful design of transitions and feeds is needed to suppress unwanted mode excitation.

Mode interference and beating

When two or more modes propagate simultaneously, they interfere. Because different modes have different propagation constants, their relative phase shifts as they travel, producing a periodic spatial variation in field intensity called mode beating.

The beat length between modes $m$ and $n$ is:

$L_b = \frac{2\pi}{|\beta_m - \beta_n|}$

Over one beat length, the interference pattern completes a full cycle. In systems requiring clean signal transmission, mode beating causes unwanted amplitude and phase fluctuations.

Multimode vs single-mode operation

The choice between single-mode and multimode operation depends on the application:

• Single-mode operation (only TEM propagating): achieved by keeping the operating frequency below the $\text{TE}_1$ / $\text{TM}_1$ cutoff, i.e., $f < c/(2d\sqrt{\epsilon_r})$. This gives the lowest dispersion and cleanest signal.
• Multimode operation (several modes propagating): occurs at higher frequencies or with larger plate separations. Offers higher power capacity but suffers from intermodal dispersion and beating.

For most communication and signal-processing applications, single-mode operation is preferred. Multimode operation is sometimes used in high-power systems where the extra modes help distribute energy.

Losses in parallel plate waveguides

Real waveguides are lossy. The two main loss mechanisms are conductor losses (from finite plate conductivity) and dielectric losses (from a non-ideal filling medium). Both cause the fields to attenuate as they propagate.

Conductor losses and skin effect

At high frequencies, current crowds near the surface of the conducting plates due to the skin effect. The skin depth is:

$\delta_s = \sqrt{\frac{2}{\omega\mu_c\sigma}}$

where $\sigma$ is the conductivity and $\mu_c$ is the permeability of the conductor material. At microwave frequencies, $\delta_s$ is typically on the order of microns (e.g., ~0.66 μm for copper at 10 GHz).

The surface resistance is:

$R_s = \frac{1}{\sigma \delta_s} = \sqrt{\frac{\omega\mu_c}{2\sigma}}$

Conductor loss increases with frequency (since $R_s \propto \sqrt{f}$) and is the dominant loss mechanism in most metallic waveguides. Using higher-conductivity materials (copper, silver) and polishing the plate surfaces reduces these losses.

Dielectric losses

If the dielectric filling has a nonzero loss tangent $\tan\delta$, it absorbs energy from the propagating wave. The dielectric attenuation constant is:

$\alpha_d = \frac{k \, \tan\delta}{2}$

where $k = \omega\sqrt{\mu\epsilon}$ is the wavenumber in the medium. Dielectric loss increases linearly with frequency (through $k$). For low-loss dielectrics like PTFE ($\tan\delta \approx 0.0002$), this contribution is small compared to conductor losses, but it becomes significant at millimeter-wave frequencies or with lossier materials.

The total attenuation is the sum of both contributions: $\alpha_{total} = \alpha_c + \alpha_d$.