# Dielectric Spectroscopy and the Debye Model *Complex permittivity ε(ω) = ε′(ω) + iε″(ω) is the canonical test-bed for Kramers–Kronig: the Debye relaxation is an exactly K–K-consistent rational function, and broadband dielectric spectrometers routinely use K–K as an internal sanity check on measured data.* ## Complex permittivity Place a linear isotropic dielectric in an oscillating electric field E(t) = Re{E₀ e^{−iωt}}. The induced polarisation P(t) = Re{P₀ e^{−iωt}} defines a complex susceptibility χ_e(ω) = P₀/(ε₀ E₀), and the relative permittivity ``` ε(ω) = 1 + χ_e(ω) = ε′(ω) + iε″(ω) ``` where, with the physics convention `e^{−iωt}`, the imaginary part ε″ ≥ 0 for a passive medium and represents the dielectric loss. The loss tangent is tan δ = ε″/ε′. Causality of the polarisation response — P(t) at time t cannot depend on E(t′ > t) — is the entire input to K–K in this setting. Subtracting the instantaneous response ε(∞) = ε_∞ (which does not go to zero at large ω) and applying Titchmarsh's theorem to χ(ω) = ε(ω) − ε_∞ gives ``` ε′(ω) − ε_∞ = (2/π) P ∫₀^∞ ω′ ε″(ω′) / (ω′² − ω²) dω′ ε″(ω) = −(2ω/π) P ∫₀^∞ [ε′(ω′) − ε_∞] / (ω′² − ω²) dω′ ``` ## Debye relaxation The simplest physical model is a relaxor: dipoles reorient toward the field with a single exponential time constant τ. The macroscopic response is ``` ε(ω) = ε_∞ + (ε_s − ε_∞) / (1 − iωτ) ``` with ε_s = ε(0) the static permittivity and ε_∞ the high-frequency limit. Split into real and imaginary parts: ``` ε′(ω) = ε_∞ + (ε_s − ε_∞) / (1 + ω²τ²) ε″(ω) = (ε_s − ε_∞) · ωτ / (1 + ω²τ²) ``` ε′(ω) is a sigmoid descending from ε_s at low ω to ε_∞ at high ω, with its steepest slope at ω = 1/τ. ε″(ω) is a Lorentzian-like peak centred at the same ω = 1/τ, with full width at half maximum spanning roughly a decade in ω. The peak height is (ε_s − ε_∞)/2. Debye's function is a rational function of ω with a single pole at ω = −i/τ, which in the physics convention sits in the **lower** half of the complex ω-plane. It is therefore analytic in the UHP: it satisfies K–K automatically. Substituting the Debye ε″ into the K–K integral reproduces the Debye ε′ exactly — a textbook worked example sometimes used to reassure students that the integral really does what it claims. ## When Debye is not enough — Cole–Cole, Davidson–Cole, Havriliak–Negami Real dielectrics — glass formers, polymers, ionic liquids, biological tissue — almost never show a single Debye peak. The observed ε″(ω) is broader and often asymmetric, reflecting a *distribution* of relaxation times rather than a single τ. Three empirical extensions dominate the literature. - **Cole–Cole** (1941): ``` ε(ω) = ε_∞ + (ε_s − ε_∞) / (1 + (−iωτ)^{1−α}) ``` with stretching parameter 0 ≤ α < 1. Symmetric broadening. On a Cole–Cole plot (ε″ vs ε′), the semicircle of the Debye response becomes a depressed arc of radius chord. - **Davidson–Cole**: `ε − ε_∞ = (ε_s − ε_∞) / (1 − iωτ)^β`, skewed to high frequency. - **Havriliak–Negami** (1967): the superset with both exponents, ``` ε(ω) = ε_∞ + (ε_s − ε_∞) / (1 + (−iωτ)^{1−α})^β ``` used to fit polymer relaxation spectra. All three are **analytic in the UHP** as long as the fractional powers are defined with the branch cut on the negative real axis of (−iωτ), so they automatically satisfy K–K. This is by design: an empirical spectral shape that violated causality would be unphysical, and the fractional-power ansatz is chosen so that the analyticity property is preserved. Jonscher's **universal dielectric response** (1977) is a different kind of phenomenology — power-law behaviour ε″(ω) ∝ ω^{n−1} with 0 < n < 1 over many decades, observed in astonishingly disparate systems from glasses to biological membranes. K–K predicts the accompanying ε′(ω) power law; fits across experimental decades are a form of Jonscher's argument for universality. ## K–K as a consistency check Broadband dielectric spectroscopy covers 10^{−5} Hz to 10^{11} Hz by concatenating several instruments (low-frequency bridges, impedance analysers, network analysers, coaxial reflectometers, THz time-domain spectroscopy). Splicing across instrument boundaries, contact impedance, electrode polarisation, and drift all introduce artefacts. The K–K transform of the measured ε″(ω) should reproduce the measured ε′(ω) (modulo ε_∞ and truncation of the integration range); any systematic deviation is a diagnostic of instrumental error rather than a property of the sample. Commercial software packages for dielectric fitting (Novocontrol's WinDETA, for example) include K–K-residual plots as a standard output. Two cautions are worth stating. - **Finite-range data.** Real measurements span a bounded frequency window, so the K–K integrals are evaluated with extrapolated or truncated tails. The most common trick is to fit the high- and low-frequency tails to a Cole–Cole or Havriliak–Negami form and analytically continue the integral to ∞ using the model. - **DC conductivity.** Mobile charge carriers add a −iσ_dc/(ε₀ ω) term to ε(ω), an imaginary part diverging at low ω. It must be subtracted before applying K–K, otherwise the dispersion integral diverges. The dc conductivity itself is extracted from the low-ω tail of ε″ω. ## Bulk-characterisation workflow A typical broadband-dielectric study of a new material proceeds: 1. Measure ε(ω) over the accessible band. 2. Subtract σ_dc/(ε₀ ω) from ε″. 3. Fit the remaining data to Havriliak–Negami (or a sum of such terms for multi-peak samples) — guarantees K–K consistency. 4. Run the K–K transform numerically and compare to the measured ε′; residuals flag bad splice points. 5. Extract relaxation times τ_k, their temperature dependence (Arrhenius or Vogel–Fulcher–Tammann), and glass-transition parameters. In every step, K–K is the background assumption that gives the data a physical skeleton. ## See also - [01_introduction.md](01_introduction.md) - [22_optical_constants.md](22_optical_constants.md) - [24_fdtd_causal_materials.md](24_fdtd_causal_materials.md) ## References - F. Kremer and A. Schönhals (eds.), *Broadband Dielectric Spectroscopy*, Springer, 2003. - A. K. Jonscher, *Dielectric Relaxation in Solids*, Chelsea Dielectric Press, 1983. - K. S. Cole and R. H. Cole, "Dispersion and absorption in dielectrics, I. Alternating current characteristics", *J. Chem. Phys.* 9, 341 (1941). - S. Havriliak and S. Negami, "A complex plane representation of dielectric and mechanical relaxation processes in some polymers", *Polymer* 8, 161 (1967). - [Wikipedia: Dielectric spectroscopy](https://en.wikipedia.org/wiki/Dielectric_spectroscopy)