Subsurface Eddy Current Phase Lag Characterization in Aluminum Plates
Subsurface phase lag scales linearly with depth, demanding low-frequency excitation and sub-0.05-degree front-end phase resolution for accurate flaw depth inversion.

Diffusion

Standard Skin Depth and Spatial Delay
As planar electromagnetic waves penetrate a conductive medium, they undergo progressive attenuation and phase shifts. In non-magnetic metals like aluminum, induced oscillating magnetic fields decay exponentially into the material. The standard skin depth (δ) marks the depth where current density falls to 1/e ~ roughly 36.8 percent ~ of its surface value, determined by drive frequency, magnetic permeability, and electrical conductivity.
δ = 1 / sqrt(π · f · μ0 · μr · σ)
Here, f is the excitation frequency in Hertz, μ0 is the permeability of free space (4π × 10^-7 H/m), μr is relative magnetic permeability (1.00 for non-magnetic aluminum alloys), and σ is electrical conductivity in Siemens per meter (S/m). Higher frequencies constrain signals near the surface, whereas lower frequencies expand the skin depth to reach deeper into thick aluminum plate.

Phase Angle Dependency on Plate Thickness and Permeability
The time required for an electromagnetic wave to propagate through a conductor creates a measurable phase lag relative to the surface driver. Inside a homogeneous aluminum plate, the phase lag θ of the magnetic field and eddy currents increases linearly with depth d when scaled to the skin depth.
θ = d / δ radians = d · sqrt(π · f · μ0 · μr · σ) · (180 / π) degrees
Each penetration distance of one skin depth introduces 1.0 radian (57.3 degrees) of phase delay. When the field hits a back-wall or subsurface flaw at depth d, the secondary magnetic field picked up by the receiver coil carries this accumulated shift. Evaluating this phase shift gives a direct measurement of flaw depth or remaining wall thickness without relying solely on signal amplitude, which fluctuates with surface roughness and probe coupling.
One skin depth of electromagnetic diffusion introduces fifty-seven point three degrees of phase lag.
Ignoring depth-dependent attenuation when interpreting phase lag leads to overestimating defect depth, which risks scrapping sound structural airframe components.

Demodulation

Analog Front End Phase Noise Floor
Extracting small phase shifts from low-level sensor voltages requires a phase-sensitive detection architecture behind the front-end amplifier. Driving the probe coil with a clean sine wave demands a high-compliance current driver, as synthesizer phase jitter translates directly into receiver uncertainty. The front end amplifies the microvolt-level coil signal through ultra-low-noise operational amplifiers before passing it to the demodulator.

Digital Quadrature Extraction and Signal Bandwidth
Isolating orthogonal signal vectors converts time-domain coil voltages into complex impedance components. Quadrature demodulation multiplies the amplified receiver signal by reference sine and cosine waves synchronized to the coil driver. Low-pass filters then strip out the double-frequency carrier terms, leaving the In-Phase (I) and Quadrature (Q) voltage components.
θ = arctan(Q / I)
Signal-to-noise ratio degrades exponentially with penetration depth as amplitude decays by exp(-d / δ). At two skin depths (2δ), the secondary field amplitude drops to 13.5 percent of its surface magnitude while round-trip phase lag reaches 2.0 radians (114.6 degrees). Front-end phase noise ultimately defines the minimum detectable change in depth.

Worked Calculation of Signal to Noise Ratio at Depth
A calculation on a six point three five millimeter structural airframe panel demonstrates these subsurface detection limits. Consider a target plate of 2024-T3 aluminum alloy with an electrical conductivity σ of 17.4 MS/m (30.0 percent IACS), inspected at 500 Hz to resolve a 1.00 mm far-surface wall-loss notch on the back wall.
First, compute the skin depth at 500 Hz:
δ = 1 / sqrt(π · 500 · 4π × 10^-7 · 1.0 · 17.4 × 10^6) = 0.00539 m = 5.39 mm
Next, determine the nominal back-wall phase lag at full thickness (d = 6.35 mm):
θ_nominal = (6.35 / 5.39) radians = 1.178 radians = 67.50 degrees
When the probe scans over the 1.00 mm wall-loss defect, remaining thickness decreases to 5.35 mm. Compute the phase lag over the defect:
θ_flaw = (5.35 / 5.39) radians = 0.993 radians = 56.88 degrees
The net phase shift Δθ produced by the 1.00 mm wall loss equals:
Δθ = 67.50 degrees – 56.88 degrees = 10.62 degrees
Resolving this 10.62-degree shift at an amplitude signal-to-noise ratio of 20 dB (a 10:1 voltage ratio) requires keeping total front-end phase noise below 1.06 degrees during bench scanning.
| Frequency (Hz) | Skin Depth δ (mm) | Normalized Depth (d/δ) | Theoretical Phase Lag (°) | Amplitude Factor exp(-d/δ) |
|---|---|---|---|---|
| 100 | 12.06 | 0.526 | 30.14 | 0.591 |
| 250 | 7.63 | 0.832 | 47.67 | 0.435 |
| 500 | 5.39 | 1.178 | 67.50 | 0.308 |
| 1000 | 3.81 | 1.667 | 95.50 | 0.189 |
| 2500 | 2.41 | 2.635 | 150.98 | 0.072 |
A phase noise floor below zero point zero five degrees across a ten hertz measurement bandwidth allows detection of two percent wall-loss defects at two skin depths.
Digital low-pass filters following the demodulator set the phase response time, balancing noise suppression against scanning speed during automated plate inspections.

Bench

Which Drive Frequency Separates Lift off from Depth?
Air gap variations between probe and target surfaces yield impedance vectors along distinct angular trajectories. At low drive frequencies, lift-off produces horizontal impedance shifts on complex plane displays, whereas subsurface phase lag rotates the vector into the fourth quadrant. Rotating the instrument’s phase frame aligns lift-off strictly along the horizontal axis, isolating depth-related phase shifts on the vertical channel.

Probe Topologies for Subsurface Signal Isolation
Reflection probes decouple excitation drivers from pickup coils to maximize dynamic range. Absolute configurations merge driver and pickup functions into one coil, making them sensitive to thermal drift, while differential setups cancel common-mode noise. Driver coil diameter dictates the field geometry; it must exceed twice the target plate thickness to establish uniform planar diffusion under the sensor face.
- Air gap variance causes immediate phase shifts that mimic thickness changes if phase rotation is off by more than one degree.
- Edge proximity interference distorts diffusion trajectories when the probe comes within two coil diameters of plate boundaries.
- Tilt misalignment creates asymmetric flux coupling, introducing phase errors that degrade depth inversion accuracy.
- Sensor heating changes driver coil resistance, causing phase baseline drift during long scan runs.
Whether multi-frequency excitation arrays can completely separate local conductivity variations from far-surface thinning in multi-layer aircraft joints remains under active laboratory evaluation.

Conductivity

Alloy Composition and Heat Treatment Variance
Alloying elements like magnesium, zinc, or copper shorten the electron mean free path in structural aluminum, lowering electrical conductivity, while precipitation hardening further modifies electron mobility. Aircraft-grade 7075-T6 plate exhibits a conductivity near 19.0 MS/m (32.8 percent IACS), structural 6061-T6 measures roughly 24.5 MS/m (42.2 percent IACS), and commercially pure 1100 aluminum reaches 34.5 MS/m (59.5 percent IACS). Temper variations can shift conductivity by up to ten percent, altering baseline phase calibrations.

Thermal Coefficients of Electrical Conductivity
Ambient test cell temperatures directly affect electron scattering within the metal lattice. Aluminum carries a thermal conductivity coefficient α of roughly 0.0039 per degree Celsius. A 10-degree Celsius rise in plate temperature reduces electrical conductivity by 3.9 percent, expanding skin depth by 1.95 percent and producing systematic phase shifts that simulate material thinning.
| Alloy and Temper | Conductivity (MS/m) | Percent IACS | Skin Depth at 500 Hz (mm) | Phase Lag Gradient (°/mm) |
|---|---|---|---|---|
| 7075-T6 | 19.0 | 32.8 | 5.16 | 11.10 |
| 2024-T3 | 17.4 | 30.0 | 5.39 | 10.63 |
| 6061-T6 | 24.5 | 42.2 | 4.55 | 12.59 |
| 5083-H116 | 16.0 | 27.6 | 5.62 | 10.19 |
| 1100-O | 34.5 | 59.5 | 3.83 | 14.96 |
- Place the reference calibration block on a thermal stabilization plate until its surface temperature is within zero point five degrees Celsius of ambient test cell temperature.
- Null the instrument in free air to clear residual probe offsets.
- Position the probe on an unblemished reference area of known conductivity to establish the zero-phase baseline.
- Adjust phase rotation settings so physical lift-off traces purely along the horizontal complex axis.
- Verify phase response across precision milled step blocks matching target plate thickness specifications.
Standard reference blocks certified under ASTM E1004 stipulate matching target plate alloy composition within one percent IACS to prevent systematic phase errors.
Batch phase discrepancies often fall within permissible nominal conductivity bands established by standard mill heat-treatment specifications.

Margin
Phase Angle Acceptance Thresholds in Production Inspection
Defect detection limits depend on maintaining robust signal-to-noise ratios against background electronic noise. Because operating temperatures shift conductivity and thicker plates require lower frequencies, instruments must match driver dynamic range to the attenuation expected at target depths. Procurement specifications typically require a 10 dB phase signal margin above the system noise floor across the entire operating temperature range.

Verification Dossiers for Instrument Specification Compliance
Procurement documents for non-destructive evaluation equipment specify strict thermal and frequency calibration limits, requiring phase stability evaluations over extended time windows to prevent drift-induced false positives.
- Phase noise specification requires certified limits below zero point zero two degrees at a ten hertz sampling bandwidth.
- Frequency sweep range covers ten hertz to one hundred kilohertz to support depth profiling across different plate gauges.
- Temperature compensation software integrates real-time thermal probe sensors to offset conductivity drift during long mill runs.
- Probe driver linearity maintains constant current output across changing coil load impedances to prevent phase distortion.
Under ISO 15549 Clause 6.4, any phase calibration drift exceeding zero point two degrees during a shift requires re-examining all plate lots inspected since the last valid verification check.




