Angle Resolved Scatterometry

Angle-resolved scatterometry is a variant of optical scatterometry that measures a periodic structure's diffraction signature as a function of incidence angle rather than wavelength, scanning the angle of an incoming light beam across a range and recording the reflected or diffracted intensity at each angle to build an angular "fingerprint" that structural-model fitting can then invert back into actual dimensions. It's a genuinely different information channel than the more commonly discussed spectroscopic scatterometry approach, which instead holds angle fixed and varies wavelength — and the two approaches, rather than competing, provide meaningfully complementary sensitivity when combined.

Fixed-wavelength angle-resolved measurement, commonly using a single-wavelength source like a 633nm HeNe laser swept across multiple incidence angles, isolates angle as the sole varying measurement parameter, which produces a genuinely different information channel about the structure than wavelength-varying spectroscopic scatterometry provides at any single fixed angle. This isolation of angle as the sole variable is exactly what makes angle-resolved and spectroscopic scatterometry meaningfully complementary rather than redundant — each approach's model-fitting sensitivity to different structural parameters comes from probing a genuinely different physical dependency of the diffraction response, not simply from collecting more data points along the same dependency.

The θ-2θ scan geometry, varying both incidence and detection angle together to trace the angular distribution of scattered light, captures the full angular reflectance curve that functions as the specific structure's unique diffraction fingerprint — a curve shape that model-based fitting algorithms can then invert to recover the actual critical dimension, sidewall angle, and height values that produced it. This fingerprint framing is literally how the measurement works in practice: rather than reading dimensions directly off the instrument, scatterometry recovery requires fitting a parameterized structural model against the measured angular curve until the model's predicted curve matches the measured one closely enough to trust the resulting extracted parameters.

Hybrid metrology approaches that combine both angle and wavelength variation in a single measurement meaningfully improve parameter-extraction accuracy and robustness relative to either variable alone, because the combined dataset constrains the structural model-fitting problem more tightly than either angle-only or wavelength-only data could by itself. This combined-constraint advantage is exactly why modern scatterometry tools increasingly integrate both angular and spectroscopic measurement capability into a single platform rather than offering them as separate, mutually exclusive instrument configurations — the two data types genuinely reinforce each other's parameter-extraction confidence rather than simply duplicating information.

Diffraction-based overlay (DBO) measurement depends specifically on angle-resolved data — measuring first-diffraction-order intensity as a function of angle — because overlay error between two patterned layers produces a specific, angle-dependent asymmetry in that diffraction order's intensity that a purely spectroscopic, fixed-angle measurement simply wouldn't capture. This dependency is exactly why angle-resolved scatterometry occupies a critical, not merely supplementary, role in advanced overlay metrology — DBO as a measurement technique is fundamentally built on the angular dependency that only angle-resolved measurement provides access to in the first place.

Specular reflectance versus angle, and specific diffraction-order intensity versus angle, represent two genuinely distinct measurement types available within the angle-resolved approach, each sensitive to somewhat different aspects of the underlying periodic structure's geometry and symmetry. Choosing which measurement type to emphasize for a given metrology application isn't arbitrary — a measurement specifically targeting overlay asymmetry naturally favors diffraction-order intensity tracking, while a measurement targeting general critical-dimension and profile extraction might rely more heavily on specular reflectance behavior, illustrating how angle-resolved scatterometry isn't one single measurement technique but a family of related approaches sharing the same underlying angular-scan principle.