Metamaterials &amp; Negative Refraction — Engineering ε and μ

3D Simulations

Article

Optics · Electromagnetism · ⏱ ~13 min read · Last updated: 9 July 2026

Metamaterials & Negative Refraction — Engineering the Optical Constants of Nature

Ordinary materials have a positive refractive index because both permittivity ε (how strongly a material responds to an electric field) and permeability μ (its response to a magnetic field) are positive. In 1968, Victor Veselago showed on paper that a hypothetical material with simultaneously negative ε and μ would bend light the "wrong" way — a negative refractive index. Nature offers almost no such materials at optical frequencies, so physicists built them artificially: sub-wavelength metallic resonators arranged in periodic lattices, engineered to respond to fields the way atoms respond to light. These metamaterials underlie superlenses, cloaking devices, and perfect absorbers.

TL;DR: Metamaterials are artificial structures — periodic arrays of split-ring resonators and metal wires — engineered so their effective permittivity and permeability both go negative in the same frequency band, producing a negative refractive index that bends light backward. This enables flat superlenses beating the diffraction limit, transformation-optics invisibility cloaks, and near-perfect microwave absorbers, though loss and narrow bandwidth still limit optical-frequency devices.

Written by MySimulator Team · Reviewed by MySimulator Editorial Review

1. Negative Refractive Index

Refractive index: n = ±√(εᵣ·μᵣ) Standard rule: take the + root when ε>0 and μ>0. Veselago (1968): if BOTH ε<0 AND μ<0 simultaneously, causality requires the MINUS root: n = −√(εᵣ·μᵣ) < 0 Consequences of n < 0: Snell's law: n₁sinθ₁ = n₂sinθ₂ still holds, but refracted ray bends to SAME side of normal as incident ray (not opposite) Phase velocity v_p = c/n is negative — antiparallel to Poynting vector S (energy still flows forward, phase runs backward) → "left-handed medium" (E, H, k form a left-handed triad instead of right-handed) Doppler effect reversed: source approaching → redshift Cherenkov radiation emitted backward relative to particle motion

2. Split-Ring Resonators (Negative μ)

A split-ring resonator (SRR): conducting ring with a gap, behaves as an LC circuit driven by the magnetic field component of light. Effective permeability (Pendry, 1999): μ_eff(ω) = 1 − (F·ω²) / (ω² − ω₀² + iΓω) F = fractional area filled by rings ω₀ = 1/√(LC) = resonant frequency (L from ring inductance, C from gap capacitance) Γ = damping (ohmic + radiative losses) Near ω₀, μ_eff swings from large positive → negative → back to positive as ω crosses the resonance (anomalous dispersion). Negative μ band: ω₀ < ω < ω_mp where ω_mp is the "magnetic plasma frequency" F·ω₀²/(1−F). Example: copper SRR, ring radius r=3mm, gap 0.2mm → resonance near 4-10 GHz (microwave); scaling ring size down to ~100 nm pushes resonance into the near-infrared/visible.

3. Wire Arrays (Negative ε)

A periodic lattice of thin metal wires behaves as a diluted plasma for the electric field, with a reduced effective plasma frequency: ε_eff(ω) = 1 - ω_p,eff² / (ω² + iΓω) ω_p,eff² = (2πc²) / [a² ln(a/r)] a = wire spacing (lattice period), r = wire radius Below ω_p,eff, ε_eff < 0 (metallic/plasmonic response) — same physics as a bulk metal below its plasma frequency, but ω_p,eff is pushed down from the UV (bulk metal) into the GHz-THz range by geometry alone. Combine an SRR array (negative μ band) with a wire array (negative ε band) so the two negative bands overlap in frequency \to double negative (DNG) metamaterial with real n < 0 in that overlap window.

4. The Veselago Superlens

Flat slab of n = −1 material, thickness d, embedded in vacuum (n=1). Ordinary lenses need curvature to focus; a flat n=−1 slab focuses rays from a point source purely through the sign flip of Snell's law — two internal focal points. Pendry's key 1999 insight: amplitude, not just phase Ordinary lens: only propagating waves (transverse wavevector k_t < ω/c) reach the image → resolution limited to ~λ/2 (diffraction limit) n=−1 slab: evanescent waves (k_t > ω/c, normally decaying as e^(−κz)) are AMPLIFIED inside the slab by surface-plasmon resonance, restoring their amplitude at the image plane → perfect (sub-diffraction) image, in the lossless ideal case Losses in real SRR/wire metamaterials cap resolution well above the ideal, but hyperlens and superlens experiments (Zhang group, 2005-08) achieved resolution below λ/6 in the near field.

5. Transformation Optics & Cloaking

Maxwell's equations are form-invariant under coordinate transformations — a spatial warp of coordinates can be re-expressed as a new material with position-dependent ε(r), μ(r). Cloaking transformation: map a point (r=0) to a small sphere of radius a (cylindrical/spherical "hole in space"), squeezing all coordinate space r\in(0,b) into r'\in(a,b). Required material (spherical cloak, Pendry 2006): ε_r' = μ_r' = (r'-a)/r' ε_θ' = μ_θ' = ε_φ' = μ_φ' = r'/(r'-a) Rays are steered smoothly around the inner region (r'<a) and recombine on the far side with undisturbed phase \to the hidden region and anything inside it become invisible from outside, for the design frequency and (ideally) all angles. First experimental demo: Schurig et al. 2006, microwave cloak using concentric SRR rings; optical-frequency cloaking remains bandwidth- and loss-limited.

6. Dispersion and Loss

Kramers-Kronig relations force any negative-ε or negative-μ resonance band to be narrow-band and strongly DISPERSIVE (n changes rapidly with ω) — a fundamental consequence of causality, not an engineering flaw. Figure of merit (FOM): FOM = |Re(n)| / Im(n) Typical microwave SRR metamaterials: FOM ~ 1-3 (lossy) Best optical metamaterials (fishnet structures): FOM ~ 2-5 near telecom wavelengths Loss mechanisms: ohmic damping in metal (\propto surface resistance, worsens at optical frequencies as metals become more plasma-like/lossy) and radiative leakage from the resonators back into free space. Loss compensation via embedded gain media (quantum dots, dyes) has demonstrated partial FOM improvement but not a lossless optical DNG medium to date.

7. JavaScript Negative-Index Ray Tracer

// Refract a 2D ray through a flat interface, supporting n < 0
// Returns refracted direction (unit vector) or null on TIR

function refract(incidentDir, normal, n1, n2) {
  // incidentDir, normal: {x, y} unit vectors, normal points into medium 1
  const cosI = -(incidentDir.x * normal.x + incidentDir.y * normal.y);
  const eta = n1 / n2;
  const sin2T = eta * eta * (1 - cosI * cosI);
  if (sin2T > 1) return null; // total internal reflection
  const cosT = Math.sqrt(1 - sin2T);
  // sign of cosT flips automatically when n2 < 0 via eta -> bends to same side
  const sign = n2 < 0 ? -1 : 1;
  return {
    x: eta * incidentDir.x + (eta * cosI - sign * cosT) * normal.x,
    y: eta * incidentDir.y + (eta * cosI - sign * cosT) * normal.y,
  };
}

// Trace a point source through a flat n=-1 slab of thickness d
// Demonstrates the two internal focal points of a Veselago lens
function veselagoLensFocus(sourceX, slabStartX, d, n = -1) {
  // distance from source to slab entry
  const s1 = slabStartX - sourceX;
  // inside slab: focus at depth proportional to |n| * s1 (paraxial approx)
  const internalFocusDepth = Math.abs(n) * s1;
  const firstFocusX = slabStartX + Math.min(internalFocusDepth, d);
  // remaining distance re-focuses just outside the exit face
  const overshoot = internalFocusDepth - d;
  const secondFocusX = slabStartX + d + Math.max(0, -overshoot);
  return { firstFocusX, secondFocusX };
}

// Lorentzian resonance model for SRR effective permeability
function muEff(omega, F, omega0, gamma) {
  const denom = omega * omega - omega0 * omega0;
  const real = 1 - (F * omega * omega * denom) / (denom * denom + (gamma * omega) ** 2);
  const imag = (F * omega * omega * gamma * omega) / (denom * denom + (gamma * omega) ** 2);
  return { real, imag };
}

// Example: F=0.4, omega0=2*pi*5e9 rad/s (5 GHz SRR), sweep for negative band
const omega0 = 2 * Math.PI * 5e9;
for (let f = 4e9; f < 8e9; f += 0.5e9) {
  const omega = 2 * Math.PI * f;
  const mu = muEff(omega, 0.4, omega0, 2e8);
  console.log(`f=${(f/1e9).toFixed(1)}GHz μ'=${mu.real.toFixed(2)}`, mu.real < 0 ? 'NEGATIVE' : '');
}

8. Applications

Superlenses & Hyperlenses

Fishnet metal-dielectric stacks and curved hyperlens geometries beat the diffraction limit for near-field imaging, useful in nanolithography metrology and biological super-resolution microscopy.

Invisibility Cloaking

Transformation-optics cloaks route microwaves or specific optical bands around a hidden object; demonstrated at microwave frequencies, actively researched for broadband optical cloaks.

Perfect Absorbers

Metamaterial absorbers tune ε and μ so the impedance matches free space (Z = √(μ/ε) = 1) while heavily damping the wave — near-100% absorption at a target frequency for stealth coatings and thermal photovoltaics.

Antenna Miniaturisation

Metamaterial substrates with engineered ε/μ shrink antenna size below the usual λ/4 rule, used in compact GPS, RFID and 5G front-end designs.

💡 Open Maxwell's Equations Simulation →