Q4Electromagnetics Waves

(a) Propagation of waves and their types: electromagnetic waves are classified by the orientation of their field components relative to the direction of propagation. A TEM (transverse electromagnetic) wave has both E and H entirely transverse to the propagation direction (Ez=Hz=0), the type supported by two-conductor lines (coax, parallel-wire) and by unbounded free space; it has no cutoff frequency and propagates non-dispersively at 1/√(με). A TE (transverse electric) wave has Ez=0 but Hz≠0, possible in single-conductor waveguides, with a nonzero cutoff frequency below which it cannot propagate. A TM (transverse magnetic) wave has Hz=0 but Ez≠0, likewise supported in waveguides with a mode-dependent cutoff (as derived explicitly for the rectangular-waveguide case in Part B, Q.7 of this paper). Beyond these three main types found in guided electromagnetic systems, waves can also be classified as plane waves (infinite planar wavefronts, an idealization) versus spherical/cylindrical waves (the actual wavefront shape radiated by a point or line source before the far-field plane-wave approximation applies), and, more broadly in physics, as transverse waves (oscillation perpendicular to propagation, the case for essentially all free-space and guided EM waves) versus longitudinal waves (oscillation parallel to propagation, common in acoustics but occurring in electromagnetics only in special contexts such as plasma (Langmuir) oscillations, not in ordinary dielectric or vacuum propagation).

(b) Wave propagation in a conducting medium — phase and group velocity: starting from Maxwell's curl equations in a conducting medium (conductivity σ) and combining them into the wave equation ∇²E = jωμ(σ+jωε)E = γ²E, the complex propagation constant γ=α+jβ is found by equating real and imaginary parts of γ²=jωμσ−ω²με, giving the standard results:

From β, the phase velocity is directly vp=ω/β, which — because β is a nonlinear function of ω through the loss-tangent term σ/ωε — is frequency-dependent, making a lossy conducting medium inherently dispersive; the group velocity vg=dω/dβ, obtained by differentiating β(ω) and inverting, generally differs numerically from vp at the same frequency, this vp≠vg condition being the defining signature of dispersion. The complete step-by-step derivation of these α, β, vp and vg expressions, including their simplified forms in the good-dielectric (σ/ωε≪1) and good-conductor (σ/ωε≫1) limiting cases, is given in full in Part C, Q.2 of the companion 2024 (Main/Back) paper on this same subject.

Elaborating part (a) — guided/surface wave types: beyond the TEM/TE/TM classification given above, it is worth noting the broader category of surface waves — electromagnetic waves that propagate bound to (and decaying exponentially away from) a single interface or a thin guiding structure without being enclosed by conducting walls on all sides, as in a genuine waveguide. Examples include the Zenneck surface wave that can propagate along a lossy ground/air interface (relevant to long-range low-frequency ground-wave radio propagation), and the dielectric-rod or optical-fibre guided wave, which is confined near the core-cladding interface by total internal reflection rather than by a conducting boundary. These surface/guided waves share the general mathematical feature of an evanescent (exponentially decaying) field transverse to the direction of propagation outside the guiding structure, in exact analogy with the below-cutoff evanescent decay discussed for ordinary waveguides elsewhere in this paper, even though the physical confinement mechanism (total internal reflection at a dielectric interface, rather than reflection from a conducting wall) is different.

Standing waves as a further propagation type: a standing wave, formed by the superposition of two oppositely travelling waves of equal amplitude (as occurs on a mismatched transmission line, discussed at length in Part C, Q.3 of the companion 2024 Jan/Feb paper, or inside a resonant cavity), is a further important wave 'type' in the sense that it does not itself transport net energy in either direction — its field pattern oscillates in place (nodes and antinodes fixed in space) rather than translating, distinguishing it fundamentally from a travelling wave even though it is mathematically constructed from two oppositely-directed travelling-wave solutions of the same wave equation. Recognizing when a physical situation produces travelling waves (unmatched systems, one-way propagation) versus standing waves (fully or partially reflected systems, resonant structures) is essential to correctly interpreting the field and power-flow behaviour of any specific electromagnetic system.

Elaborating part (b) — the two limiting regimes in physical detail: the general α, β formulas simplify usefully at the two extremes of the loss tangent σ/ωε. In the good-dielectric (low-loss) limit (σ/ωε ≪ 1, typical of most practical insulators well below any relaxation resonance), a binomial expansion of the square root gives α ≈ (σ/2)√(μ/ε) (independent of frequency to leading order) and β ≈ ω√(με)[1+(1/8)(σ/ωε)²] ≈ ω√(με), so vp ≈ vg ≈ 1/√(με), essentially indistinguishable from the lossless-dielectric result — physically, a good dielectric supports wave propagation that is only weakly attenuated and very nearly non-dispersive, which is why ordinary dielectric materials used in cables and substrates are treated as effectively lossless for many practical engineering calculations. In the good-conductor (high-loss) limit (σ/ωε ≫ 1, typical of metals at RF and microwave frequencies), the square root approximation instead gives α ≈ β ≈ √(ωμσ/2), so vp = ω/β = √(2ω/μσ), a result that depends on √ω and is therefore strongly dispersive — different frequency components travel at markedly different phase speeds inside a good conductor. This high-loss limit is exactly the regime that produces the skin effect discussed in the electrical-materials portion of this examination series, with skin depth δ = 1/α = √(2/ωμσ) directly derived from this same α expression, confirming that skin effect and general lossy-medium wave propagation are simply two descriptions of the same underlying physics viewed in the good-conductor limit.

Why this matters for waveguide and transmission-line engineering: the distinction between these two limiting regimes explains why waveguide interiors are deliberately kept air-filled (or filled with a good, low-loss dielectric) rather than conducting — operating in the good-dielectric regime keeps α small and propagation nearly dispersion-free — while the waveguide walls themselves are made of the best available conductor (often silver- or gold-plated) specifically to minimize the small but nonzero conductor-loss contribution to overall guide attenuation, which arises from exactly the good-conductor skin-effect physics described above applied to the finite (not infinite) conductivity of the real wall material.