2.6 Attenuation

Causes of attenuation

Attenuation is the reduction in amplitude or intensity of an electromagnetic wave as it propagates through a medium. Every real medium extracts some energy from a wave, and quantifying that loss is essential for designing communication links, waveguides, and any system that moves EM energy from one place to another.

Two main mechanisms drive attenuation: absorption and scattering. How much attenuation you get depends on the wave's frequency and the electromagnetic properties of the medium.

Absorption vs scattering

Absorption converts the wave's energy into heat (or another internal energy form) within the medium. Dielectric heating in a microwave oven is a familiar example: the oscillating electric field forces polar molecules to rotate, and friction-like losses turn that motion into thermal energy.

Scattering redirects the wave away from its original propagation direction when it encounters inhomogeneities (grain boundaries, particles, bubbles, etc.). Two important regimes:

• Rayleigh scattering dominates when the scatterers are much smaller than the wavelength. Scattering intensity scales as $\lambda^{-4}$, so shorter wavelengths scatter far more strongly.
• Mie scattering applies when scatterer size is comparable to the wavelength and produces a more complex angular pattern.

Both mechanisms remove energy from the forward-propagating wave, so total attenuation is the sum of absorptive and scattering losses.

Frequency dependence

Attenuation is almost always frequency-dependent. As a general trend, higher frequencies experience greater attenuation, but the relationship is not always monotonic. Near molecular or atomic resonances, absorption spikes sharply. Relaxation phenomena (discussed below in the dielectrics section) also create frequency bands of elevated loss. This is why choosing an operating frequency always involves a trade-off between bandwidth and acceptable loss.

Material properties

Three bulk parameters control how a medium attenuates an EM wave:

• Conductivity $\sigma$: Higher conductivity means larger ohmic currents and more resistive dissipation.
• Complex permittivity $\epsilon = \epsilon' - j\epsilon''$: The imaginary part $\epsilon''$ captures dielectric loss. A large loss tangent $\tan\delta = \epsilon''/\epsilon'$ means heavy absorption.
• Permeability $\mu$: In magnetic materials, hysteresis and eddy-current losses add another attenuation channel.

Metals (high $\sigma$) and lossy dielectrics (high $\tan\delta$) produce the strongest attenuation. Low-loss dielectrics like PTFE or fused silica are chosen precisely because their $\epsilon''$ is tiny at the frequencies of interest.

Attenuation in conductors

Conductors are everywhere in EM systems (waveguide walls, transmission-line conductors, antenna elements), and their finite conductivity always introduces loss. Two frequency-dependent effects dominate: the skin effect and the proximity effect.

Skin effect

At DC, current distributes uniformly across a conductor's cross-section. As frequency rises, the time-varying magnetic field inside the conductor induces eddy currents that oppose the interior current, pushing the current density toward the surface. The characteristic depth at which the current density falls to $1/e$ of its surface value is the skin depth:

$\delta = \sqrt{\frac{2}{\omega\mu\sigma}}$

For copper at 1 GHz, $\delta \approx 2.1\;\mu\text{m}$. That means virtually all the current flows in a shell only a few micrometers thick. Because the effective cross-sectional area carrying current shrinks, the AC resistance rises well above the DC value, and resistive losses increase accordingly.

Proximity effect

When two or more conductors sit close together, the magnetic field from one conductor distorts the current distribution in its neighbor. Current crowds toward the facing surfaces, further reducing the effective conducting area and raising resistance. The proximity effect compounds the skin effect and becomes significant in tightly spaced transmission lines, transformer windings, and multi-conductor cables.

Conductor losses

The combined result of skin and proximity effects is an increase in effective resistance per unit length. The power dissipated per unit length is $P_\text{loss} = \frac{1}{2}|I|^2 R_\text{eff}$, and this energy comes directly out of the propagating wave.

For a good conductor, the attenuation constant due to conductor loss is:

$\alpha_c = \sqrt{\frac{\omega\mu}{2\sigma}}$

This quantity has units of Np/m (nepers per meter) and grows as $\sqrt{f}$, so doubling the frequency increases conductor attenuation by about 41%.

Attenuation in dielectrics

Dielectrics support electric fields without (ideally) conducting current, but no real dielectric is perfectly lossless. Energy is dissipated through several polarization mechanisms, each with its own characteristic frequency range.

Dielectric losses

When an alternating electric field is applied to a dielectric, bound charges and molecular dipoles must continuously reorient. Each reorientation cycle dissipates a small amount of energy. The loss tangent captures this:

$\tan\delta = \frac{\epsilon''}{\epsilon'}$

where $\epsilon'$ is the real (energy-storage) part and $\epsilon''$ is the imaginary (loss) part of the complex permittivity. A material with $\tan\delta = 0.001$ dissipates 0.1% of the stored energy per cycle. Contributing mechanisms include:

• Dipolar (orientational) relaxation: polar molecules lag behind the field
• Ionic polarization losses: displacement of ions in a crystal lattice
• Electronic polarization losses: distortion of electron clouds (significant only at optical frequencies)

Relaxation phenomena

Each polarization mechanism has a relaxation time $\tau$, the time scale over which the polarization can follow a changing field. The Debye model describes the simplest case (a single relaxation time):

$\epsilon(\omega) = \epsilon_\infty + \frac{\epsilon_s - \epsilon_\infty}{1 + j\omega\tau}$

Here $\epsilon_s$ is the static (low-frequency) permittivity and $\epsilon_\infty$ is the high-frequency limit. Near $\omega = 1/\tau$, the imaginary part $\epsilon''$ peaks, producing a band of maximum absorption. Real materials often have a distribution of relaxation times, broadening the loss peak beyond the simple Debye prediction.

Frequency dependence

At low frequencies, conduction losses and interfacial (Maxwell-Wagner) polarization tend to dominate $\epsilon''$. As frequency increases through the microwave and infrared bands, dipolar and ionic relaxation losses appear in sequence. At optical frequencies, electronic polarization is the remaining mechanism.

Understanding this layered frequency dependence is critical for selecting dielectric materials. For example, water has a strong dipolar relaxation near 20 GHz, which is why microwave ovens operate at 2.45 GHz (on the low-frequency shoulder of that absorption peak, giving good penetration depth while still coupling efficiently).

Attenuation in waveguides

Waveguides confine EM energy within a hollow or dielectric-filled structure. Attenuation in waveguides has a unique feature not present in unbounded media: the cutoff frequency.

Cutoff frequency

Each waveguide mode has a minimum frequency below which it cannot propagate. For a rectangular waveguide with inner dimensions $a \times b$ ($a > b$), the cutoff wavelength of the $\text{TE}_{mn}$ or $\text{TM}_{mn}$ mode is:

$\lambda_c = \frac{2}{\sqrt{\left(\frac{m}{a}\right)^2 + \left(\frac{n}{b}\right)^2}}$

The corresponding cutoff frequency is $f_c = c / \lambda_c$ (in an air-filled guide). Operating below $f_c$ for a given mode means that mode cannot carry power; the fields decay exponentially instead of propagating.

Evanescent modes

Below cutoff, the propagation constant becomes purely imaginary (or, equivalently, the wave vector in the propagation direction becomes imaginary). The fields then decay as $e^{-\alpha z}$ with distance $z$. These evanescent modes store reactive energy near the excitation point but do not transport real power down the guide. They are important in the design of waveguide filters and attenuators, where controlled evanescent sections provide precise attenuation.

Waveguide losses

Above cutoff, three loss mechanisms reduce the power carried by a propagating mode:

• Conductor (wall) losses: Finite conductivity of the metallic walls causes ohmic dissipation. These losses increase as the operating frequency approaches cutoff because the group velocity drops and the fields interact more strongly with the walls.
• Dielectric losses: If the guide is filled (or partially filled) with a lossy dielectric, the bulk $\tan\delta$ of that material contributes directly to attenuation.
• Radiation losses: Slots, joints, bends, or surface roughness can cause energy to leak out of the guide.

For a well-made air-filled metallic waveguide, conductor loss is usually the dominant mechanism.

Attenuation in transmission lines

Transmission lines (coaxial cables, microstrip, stripline, twisted pair) carry EM energy in TEM or quasi-TEM modes. Attenuation here is governed by the line's distributed resistance and dielectric loss, together with any impedance mismatches that cause reflections.

Characteristic impedance

The characteristic impedance $Z_0$ relates the voltage and current waves on the line. For a lossless line:

$Z_0 = \sqrt{\frac{L}{C}}$

where $L$ and $C$ are the inductance and capacitance per unit length. When the source and load impedances both equal $Z_0$, no reflections occur and maximum power is delivered. Any mismatch sends energy back toward the source, effectively increasing the apparent attenuation of the link.

Reflection coefficient

The reflection coefficient at a load $Z_L$ is:

$\Gamma = \frac{Z_L - Z_0}{Z_L + Z_0}$

$|\Gamma| = 0$ means a perfect match (no reflection); $|\Gamma| = 1$ means total reflection (open or short circuit). Even moderate mismatches ($|\Gamma| = 0.2$, corresponding to a VSWR of 1.5) waste about 4% of the incident power in reflection. In a system with multiple mismatches, repeated reflections create standing waves and additional loss.

Transmission line losses

The total attenuation constant $\alpha$ of a transmission line is the sum of conductor and dielectric contributions:

$\alpha = \alpha_c + \alpha_d$

• $\alpha_c$ grows roughly as $\sqrt{f}$ (skin-effect scaling).
• $\alpha_d$ grows roughly linearly with $f$ (since $\epsilon''$ often increases with frequency in the relevant band).

At low frequencies conductor loss dominates; at high frequencies dielectric loss takes over. Manufacturers' data sheets typically give $\alpha$ in dB per unit length at specific frequencies so you can estimate total loss for a given cable run.

Measuring attenuation

Quantifying attenuation accurately is essential for verifying designs and characterizing components.

Attenuation coefficient

The attenuation coefficient $\alpha$ describes how fast the wave amplitude decays with distance:

$A(z) = A_0\, e^{-\alpha z}$

In nepers per meter (Np/m), $\alpha$ gives the exponential decay rate directly. Converting to dB/m:

$\alpha_{\text{dB/m}} = 8.686\;\alpha_{\text{Np/m}}$

For long-haul fiber optics, attenuation is quoted in dB/km (e.g., ~0.2 dB/km for single-mode silica fiber at 1550 nm).

Decibel scale

The decibel is a logarithmic ratio that makes it easy to add up gains and losses in a cascade. For power:

$A_{\text{dB}} = 10\log_{10}\!\left(\frac{P_1}{P_2}\right)$

For voltage or field amplitude (assuming equal impedances):

$A_{\text{dB}} = 20\log_{10}\!\left(\frac{V_1}{V_2}\right)$

A 3 dB loss means half the power is gone. A 10 dB loss means only 10% of the power remains. These quick reference points are worth memorizing.

Measurement techniques

• Vector network analyzer (VNA): Measures S-parameters ($S_{21}$ gives insertion loss, $S_{11}$ gives return loss) over a swept frequency range. This is the standard tool for characterizing cables, connectors, filters, and waveguide components.
• Power meter: Measures absolute power at a point. Comparing input and output power gives total attenuation directly.
• Spectrum analyzer: Useful for frequency-selective measurements and identifying frequency-dependent attenuation.
• Time-domain reflectometry (TDR): Sends a fast pulse down a line and analyzes reflections. TDR locates impedance discontinuities and estimates distributed loss along a transmission line.

Mitigating attenuation

Reducing attenuation is a central goal in most EM system designs. Three main strategies are available.

Impedance matching

Matching eliminates reflection losses and ensures maximum power transfer. Common techniques:

1. Quarter-wave transformer: A transmission line section of length $\lambda/4$ with characteristic impedance $Z_T = \sqrt{Z_0 Z_L}$ transforms $Z_L$ to $Z_0$ at the design frequency.

2. Stub matching: Short- or open-circuited stubs placed at calculated positions along the line cancel the reactive part of the load impedance.

3. Lumped-element matching networks: L-sections, pi-networks, or T-networks built from inductors and capacitors, practical at lower frequencies where lumped components are electrically small.

Each method works perfectly at one frequency; broadband matching requires more complex (multi-section or tapered) designs.

Low-loss materials

Material selection directly controls both conductor and dielectric losses:

• Conductors: Silver has the highest conductivity, but copper is the standard engineering choice (97% of silver's conductivity at a fraction of the cost). Gold plating prevents oxidation on contact surfaces.
• Dielectrics: PTFE (Teflon, $\tan\delta \approx 0.0002$ at 10 GHz), fused silica, and specialized low-loss ceramics are used in high-frequency PCBs, radomes, and waveguide windows.

Surface finish also matters: at microwave frequencies, surface roughness comparable to the skin depth increases conductor loss significantly.

Amplification techniques

When attenuation cannot be reduced below acceptable levels through passive means, active amplification compensates for the lost signal:

• Low-noise amplifiers (LNAs) are placed at the receiver front end to boost weak signals while adding minimal noise.
• Power amplifiers (PAs) at the transmitter end increase the launched signal level.
• Distributed amplification places repeater amplifiers at intervals along a long link (fiber-optic or RF) so the signal never drops below a usable level.

Amplifier placement and gain must be chosen carefully: too much gain introduces distortion and oscillation risk, while too little leaves the signal buried in noise.

Applications of attenuation

Attenuation is not always something to fight against. In many applications, controlled attenuation is the desired function.

Signal processing

• Attenuators are passive components that reduce signal amplitude by a precise amount (e.g., 3 dB, 6 dB, 10 dB pads). They protect sensitive receivers and set signal levels in test setups.
• Equalizers apply frequency-dependent gain (or loss) to flatten the overall channel response, compensating for the fact that cables and channels attenuate higher frequencies more than lower ones.
• Filters exploit the frequency dependence of attenuation to pass desired bands and reject others. A waveguide section operated below cutoff for a particular mode, for instance, acts as a high-pass filter.

Electromagnetic shielding

Shielding relies on attenuation to keep unwanted EM energy out of (or contained within) an enclosure. A conductive shield attenuates an incident wave through two mechanisms:

• Reflection loss: Impedance mismatch at the air-metal boundary reflects most of the wave.
• Absorption loss: The portion that enters the metal decays exponentially over a few skin depths.

Shielding effectiveness (SE) in dB is the total attenuation through the shield. A 1 mm copper sheet provides over 100 dB of SE at 1 GHz, more than enough for most EMI/EMC requirements. Seams, gaskets, and apertures are usually the weak points, not the bulk material.

Microwave heating

Microwave ovens exploit dielectric absorption: water molecules in food absorb energy at 2.45 GHz, converting it to heat. The attenuation (penetration depth) determines how deeply the energy reaches into the material. Foods with high water content absorb strongly and heat quickly near the surface, which is why thick items may need lower power or longer time to heat uniformly. Industrial microwave heating applies the same principle to drying, curing, and sintering processes, with frequency and power chosen to match the loss characteristics of the target material.