Evanescent modes feature steep attenuation (e.g., TE₀₁ in rectangular waveguides decays ~0.6dB/μm at 10GHz), trapping >85% energy within 10μm of walls as fields diminish exponentially from surfaces; excited via near-field probes, they never propagate, unlike guided modes.
Table of Contents
Rapid decay with distance
A standard silicon optical waveguide operating at a wavelength (λ) of 1550 nanometers, the evanescent field’s intensity typically falls to about 1/exp(2π) (roughly 0.2%) of its initial value at a distance of just λ/2, or about 775 nm, from the waveguide core. This rapid drop-off is quantified by the penetration depth (δ), which is the distance at which the field amplitude decreases by a factor of 1/e(about 37% of its starting value). In many practical waveguide scenarios, this δ can be as small as 100 nm to 1 μm, effectively constraining the field’s influence to an extremely narrow region.
The spatial decay is governed by the attenuation constant (α), where the electric field amplitude follows E(z) = E₀ * e^(-αz). This means that if the attenuation constant α is 1000 m⁻¹, the field’s amplitude will be halved approximately every 0.69 mm (since ln(2)/α ≈ 0.00069 m). The value of α is not arbitrary; it is directly determined by the discrepancy between the cut-off wavenumber (k_c) and the wave number in the medium. For a rectangular waveguide with a cut-off frequency 10% higher than the signal frequency, α can be on the order of 100s to 1000s of nepers per meter. This exponential relationship is the reason why these modes are effectively “localized.” For instance, increasing the distance from the source by just three times the penetration depth (3δ) reduces the field’s power (which is proportional to the square of the amplitude) to just E₀² * e^(-6), or about 0.25% of its initial power. This is why bringing a second waveguide or a sensor within a distance of a few hundred nanometers is critical for efficient coupling in devices like directional couplers or evanescent field sensors.
| Distance from Interface (z / δ) | Normalized Field Amplitude (E / E₀) | Normalized Power (P / P₀) |
|---|---|---|
| 0 | 1.000 | 1.000 |
| 0.5 | 0.607 | 0.368 |
| 1.0 | 0.368 | 0.135 |
| 2.0 | 0.135 | 0.018 |
| 3.0 | 0.050 | 0.0025 |
A surface plasmon resonance (SPR) biosensor can detect a change in refractive index within a ~200 nm thick layer above a gold film because the evanescent field power drops to near zero beyond that distance. This confinement provides excellent spatial resolution and surface specificity, allowing the sensor to ignore bulk solution effects and focus on molecular binding events occurring immediately at the surface, with a typical sensitivity measured in Refractive Index Units (RIU) on the order of 10⁻⁶ to 10⁻⁷ RIU. In integrated photonics, this property enables the dense packing of waveguides. Engineers can place two waveguides as close as 1-2 μm apart with confidence that cross-talk will be minimal because the evanescent fields decay sufficiently over the gap, ensuring isolation better than -30 dB at the operating wavelength.
No net energy flow
In a propagating mode, these fields are in phase, resulting in a non-zero time-average of the Poynting vector, which points in the direction of propagation. In an evanescent mode, a 90-degree phase shift exists between the transverse electric and magnetic fields. This quadrature phase relationship causes the instantaneous power flow to oscillate back and forth locally, much like a simple harmonic oscillator exchanging energy between kinetic and potential forms, resulting in a net time-averaged power of exactly 0 watts per square meter.
For a wave with a frequency of 200 THz (a common infrared wavelength of 1500 nm), this power oscillation occurs at a staggering 400 THz. The amount of energy sloshing back and forth is directly tied to the field’s strength at a given point. For example, at a distance of 1 micron from the waveguide core where the field amplitude might be 30% of its peak value, the peak instantaneous reactive power density could be on the order of 10-100 watts per square meter, but its time-average remains zero. This is why an isolated evanescent field, by itself, cannot transmit information or energy to a distant point.
The defining characteristic of an evanescent mode is a net energy flow of zero; it acts as a reactive energy storage field, not a radiative power transmitter.
When a second waveguide or a receiver is brought within the decay length (typically < 1 µm), the evanescent field’s reactive energy can interact with it. The presence of this second object perturbs the system, allowing the localized energy to be “tapped” and converted into a propagating mode in the adjacent structure. The efficiency of this transfer is exquisitely sensitive to the gap. A gap increase from 0.5 µm to 1.0 µm can reduce the coupling efficiency by over 50% because the strength of the reactive field available for interaction drops exponentially.
| Characteristic | Propagating Mode (e.g., Fundamental Mode) | Evanescent Mode (Below Cut-off) |
|---|---|---|
| Time-Averaged Net Power Flow | Non-zero (e.g., 1 mW in a single-mode fiber) | 0 W |
| Nature of Power | Real, transmitted power | Reactive, stored power (imaginary Poynting vector) |
| Field Phase Relationship | Electric and Magnetic fields in phase | 90-degree phase shift between transverse E and H fields |
| Typical Application | Long-distance communication (>1 km) | Near-field coupling, sensing over sub-micron distances |
In an evanescent field biosensor, a protein molecule with a diameter of about 5 nm binding to the sensor surface interacts with this reactive field. This interaction changes the local effective refractive index, which subtly alters the propagation constant of the guided mode in the core, shifting the resonance frequency by a measurable amount, perhaps 0.01%. The sensor detects this shift precisely because the evanescent field is not radiating energy away but is storing it locally, making it exquisitely sensitive to minute surface changes.
Existence below cut-off frequency
For a standard rectangular metallic waveguide with a 20 mm x 10 mm cross-section, the cut-off frequency for the dominant TE10 mode is approximately 7.5 GHz. If you attempt to propagate a 5 GHz signal through this guide, which is 33% below the cut-off, it will not travel. Instead, it establishes an evanescent field that decays exponentially with distance, becoming negligible within a short length, often just a few centimeters. The transition from propagation to evanescence is abrupt; a mere 1% decrease in frequency below the cut-off can change the wave’s behavior from traveling kilometers to fading within meters.
- The cut-off condition is determined by the waveguide’s narrowest transverse dimension and the refractive index contrast between the core and cladding.
- Operating below this frequency forces the propagation constant (β) to become a purely imaginary number, which mathematically dictates exponential decay.
- The rate of decay is not constant; it increases sharply as the operating frequency moves further below the cut-off frequency.
The underlying math is straightforward. The propagation constant γ is given by γ² = (π/a)² – ω²με, where ‘a’ is the waveguide width. Above cut-off, ω²με > (π/a)², making γ imaginary (jβ) and representing a propagating wave. Below cut-off, ω²με < (π/a)², forcing γ to be a real number (α), which is the attenuation constant. The value of α in Nepers per meter is α = √((π/a)² – ω²με). This means the attenuation is not a linear function.
For our 20 mm wide waveguide at 5 GHz, α calculates to roughly 0.83 Np/m. Since a field drops by a factor of e(about 37% in amplitude) over a distance of 1/α, the 1/e decay length is about 1.2 meters. If the frequency is lowered further to 3 GHz (60% below cut-off), the attenuation constant α increases to approximately 1.57 Np/m, and the 1/e decay length shrinks to just 0.64 meters. This quantifies why a signal only slightly below cut-off might still have a perceptible field a short distance away, while one far below cut-off vanishes almost instantly. In optical fiber terms, for a single-mode fiber with a 9 µm diameter core and a numerical aperture of 0.12, the cut-off wavelength for the fundamental mode is around 1260 nm. Light at a wavelength of 1310 nm propagates efficiently with an attenuation of about 0.3 dB/km. However, if you inject light with a wavelength of 1550 nm, which is 23% longer than the cut-off wavelength, the fiber can only support the fundamental mode. But if you try to launch a higher-order mode, like the LP11 mode, at 1550 nm, it becomes evanescent because its cut-off wavelength is around 1400 nm; it will be extinguished within a few millimeters, with a loss exceeding 100 dB/km.
Stronger confinement near source
The confinement strength is quantified by the attenuation constant (α) or, more intuitively, the penetration depth (δ), which is the distance at which the field amplitude decreases to about 37% of its value at the interface. For a silicon nitride photonic waveguide operating at 1550 nm, this δ can be as small as 150 nm. This means that within the first 300 nm (twice the penetration depth), the field’s intensity (proportional to the square of the amplitude) will have dropped to roughly (0.37)² ≈ 14% of its surface value. This creates an effective sensing or interaction volume that is exceptionally shallow, often less than 1 µm in total depth, ensuring that any measurement is supremely sensitive to surface conditions rather than bulk properties.
- The field amplitude follows a strict exponential decay formula: E(z) = E₀ * e^(-z/δ), making its presence overwhelmingly dominant within a distance of 1-2 penetration depths from the source.
- The degree of confinement is dynamically tunable; operating further below the cut-off frequency significantly reduces the penetration depth, tightening the confinement.
- This creates a steep energy density gradient, where the power density can change by an order of magnitude over a distance of a few hundred nanometers.
For example, in a microwave waveguide with a cut-off of 10 GHz, a 9 GHz signal might have a penetration depth of 5 cm. However, a 5 GHz signal, which is 50% further below cut-off, will have a much smaller δ, perhaps only 1.5 cm, confining the field more tightly to the discontinuity. This relationship is a critical design parameter. The following table illustrates how the confinement, measured by the normalized power remaining, changes with distance for two different scenarios: one slightly below cut-off (weaker confinement) and one far below cut-off (stronger confinement).
| Distance from Source | Normalized Power (Slightly below cut-off, e.g., δ = 500 nm) | Normalized Power (Far below cut-off, e.g., δ = 150 nm) |
|---|---|---|
| z = δ | 0.37 | 0.37 |
| z = 2δ | 0.14 | 0.14 |
| z = 3δ | 0.05 | 0.05 |
| Absolute Distance: z = 300 nm | P ≈ 0.55 | P ≈ 0.14 |
In scanning near-field optical microscopy (SNOM), a metallic tip with an aperture of just 50 nm is placed deep within the evanescent field (less than 10 nm from the surface). At this distance, the field intensity is still over 90% of its maximum value, allowing the probe to capture details far below the diffraction limit, resolving features as small as 20 nm. In integrated photonic circuits, strong confinement is essential for creating compact devices. A micro-ring resonator with a radius of 10 µm can effectively filter wavelengths because the evanescent tail coupling between the ring and the adjacent bus waveguide is tightly confined to a gap of 200 nm. This tight confinement ensures that the coupling is strong enough to be functional but localized enough to prevent crosstalk with other circuit elements just 5 µm away.
Useful near-field applications
The unique properties of evanescent fields—especially their exponential decay and strong near-field confinement—are not just theoretical curiosities; they are the operational basis for a wide range of high-precision technologies. Because the field intensity is significant only within a fraction of a wavelength from the source (typically < 1 µm for optical frequencies), it provides a perfectly localized probe for sensing, imaging, and signal manipulation at the nanoscale. This allows devices to circumvent the fundamental diffraction limit of light, which restricts conventional optics to resolving features no smaller than about 200-300 nm.
- Evanescent waves enable sensing with extreme surface sensitivity, as the interaction is limited to a depth of ~200 nm, making the signal immune to bulk solution effects.
- They form the basis for key photonic integrated components like directional couplers and ring resonators by allowing controlled energy transfer across nanoscale gaps.
- In imaging, they allow for resolution beyond the diffraction limit by detecting near-field information before it propagates as far-field radiation.
In a surface plasmon resonance (SPR) sensor, a gold film ~50 nm thick is excited to create a plasmon with a very strong evanescent field extending 100-300 nm into the analyte. When a protein with a molecular weight of 50 kDa binds to the sensor surface, it changes the local refractive index within this tiny volume. A high-quality SPR instrument can detect a refractive index shift as small as 10⁻⁶ to 10⁻⁷ RIU, which corresponds to a surface coverage change of less than 1 picogram per square millimeter. This allows researchers to measure binding kinetics in real-time, determining association rates (kₐ) on the order of 10⁵ 1/Ms and dissociation rates (kₑ) of 10⁻³ 1/s with high precision. The evanescent field’s short range is crucial here; it ensures the sensor is >90% insensitive to changes in the bulk solution several microns away, focusing exclusively on the molecular binding events at the interface.
A directional coupler, which splits optical power between two waveguides, functions by placing the cores a precise distance apart, often 0.2 to 0.5 µm. The coupling length (Lc) for a 50/50 power split is inversely proportional to the strength of the evanescent tail overlap. For a silicon photonic chip operating at 1550 nm, this Lc might be 50 µm. The coupling ratio is highly wavelength-dependent; a shift of just 10 nm can change the splitting ratio by ±15%, a property used to build wavelength-division multiplexing (WDM) filters. Similarly, a micro-ring resonator with a radius of 5 µm and a Q-factor of 10,000 relies on evanescent coupling from an adjacent waveguide to filter a specific channel with a bandwidth of just 0.15 nm. The gap between the ring and the waveguide must be controlled to within ±10 nm during fabrication to achieve the designed coupling efficiency, as a deviation of 50 nm can drop the coupled power by over 70%.
