Stylus Profilometry vs Ellipsometry vs SEM: Comparing

Stylus profilometry, spectroscopic ellipsometry, and Scanning Electron Microscopy (SEM) are the three main techniques for measuring thin-film thickness and surface topography. Each uses a different physical mechanism: mechanical contact, optical polarization change, or electron–matter interaction. This article compares their principles, instrumentation, sample preparation, workflows, accuracy, repeatability, and best-fit application scenarios.

Principle & mechanism

Stylus profilometry is a contact method. A diamond-tipped stylus of defined tip radius is drawn across the surface at constant low force. A transducer converts the vertical displacement of the stylus arm into a height profile. The measured signal is a direct geometrical trace. Step height and roughness values therefore follow from surface coordinates without a material model.

Ellipsometry is a non-contact optical method based on polarization. Linearly polarized light reflected from the sample becomes elliptically polarized. The measured quantities psi and delta depend on film thickness and refractive indices. Thickness is obtained by fitting an optical model to the data. The result is therefore model-dependent and requires reasonable input of optical constants.

SEM builds images from secondary and backscattered electrons generated by a focused electron beam. Contrast arises from differences in material composition and surface topography. Cross-sectional imaging yields direct visual thickness information with nanometer-level detail.

Measurement Principle & Instrumentation

A stylus profilometer comprises a precision scanning stage, a stylus assembly with interchangeable tips, a force-feedback or lever transducer, and software for leveling, filtering, and step-height computation. Vertical resolution reaches the sub-nanometer range on vibration-isolated instruments. Lateral resolution is limited by tip radius and flank angle.

A spectroscopic ellipsometer uses a light source, a rotating-compensator or nulling polarization analyzer, and a detector spanning ultraviolet to near-infrared wavelengths. Measurements are typically taken at multiple incident angles. Regression against a layered optical stack model yields thickness and refractive index simultaneously.

The SEM consists of an electron gun, electromagnetic lenses, scan coils, and detectors for secondary and backscattered electrons. Focused ion beam (FIB) cross-sectioning is frequently combined with SEM to prepare damage-free cross sections for direct film-thickness measurement. Each platform demands its own calibration artifacts. Examples include certified step-height standards, reference silicon oxide wafers, or magnification standards.

Sample Preparation Requirements

Profilometry samples must present a measurable step or distinct topographic feature. Patterned films, masked etch steps, or masked deposition areas are typical configurations. The surface should be free of loose particles that could deflect the stylus. The sample must be rigidly mounted because vibration corrupts the trace. Maximum sample height and weight are constrained by the stage.

Ellipsometry requires a specular, optically smooth surface with a film that is at least partially transparent at the probe wavelengths. Opaque or strongly roughened layers restrict the technique. Multilayer stacks need a known nominal layer structure so that the optical model remains solvable.

SEM specimens must be vacuum-compatible and electrically conductive, or coated with a thin conductive layer to suppress charging. Cross-section preparation may involve mechanical polishing and ion milling, or FIB lift-out for site-specific locations. All three methods benefit from cleanroom handling to avoid particulate contamination.

Step-by-Step Test Procedure

For stylus profilometry, the operator mounts the sample, selects an appropriate tip radius and contact force, and levels the stage. A pre-scan at reduced force locates the step. A measurement scan then crosses the feature at controlled speed. Data processing applies datum leveling, noise filtering, and a step-height algorithm that averages the two plateau regions.

For ellipsometry, the workflow starts with definition of the layer stack and selection of optical constants from a material library or from prior characterization. Data are acquired at multiple wavelengths and angles of incidence. Model regression then iterates film thickness until calculated and measured psi/delta spectra agree within acceptable error. Results are reported together with goodness-of-fit indicators such as mean squared error.

For SEM cross-sectioning, the region of interest is marked and the cross section is prepared by FIB or polishing. The sample is loaded and pumped down. Imaging is performed at defined accelerating voltage and magnification, with the film layers measured against the instrument scale.

Performance

Metrics: Accuracy & Repeatability
Accuracy and repeatability are assessed against traceable reference standards and repeated measurements at fixed locations. Stylus profilometry offers direct traceability to length standards. It excels at step heights from nanometers to hundreds of micrometers. Repeatability depends on stage flatness, vibration isolation, and operator consistency in defining measurement regions. Its lateral resolution is the main limitation.

Ellipsometry achieves sub-nanometer precision on smooth transparent films and can resolve very thin layers. Accuracy is bounded by model quality. Incorrect optical constants introduce systematic error that repeatability statistics cannot reveal. Goodness-of-fit and parameter correlation checks should therefore accompany every reported value.

SEM delivers high lateral resolution and direct visualization of layer interfaces. This makes it valuable for multilayers and abrupt interfaces. Measurement uncertainty is influenced by cut angle, edge rounding, and magnification calibration. Cross-lab agreement is generally strongest when methods are combined. Examples include verifying ellipsometric results against SEM cross sections or profilometric step heights.

Application Scenarios & Method Selection

Method choice follows from film transparency, thickness range, and required throughput. Stylus profilometry suits patterned wafers, etch-depth control, coating step heights, and roughness qualification where a physical step exists. It is robust, fast, and directly traceable.

Ellipsometry is preferred for uniform transparent or semi-transparent films such as oxide and nitride layers on wafers. It provides non-contact, non-destructive mapping of thickness and refractive index across a surface. It is the standard approach for high-volume process monitoring of thin dielectric films.

SEM-based cross-sectioning applies when layers are opaque, extremely thin, or part of complex multilayer stacks. It is also used when interfacial structure, defects, or grain morphology must be visualized alongside thickness. In practice, laboratories often use ellipsometry or profilometry for routine measurement. SEM cross-section analysis is reserved for qualification, failure analysis, and arbitration of disputed results, accepting its destructive nature and longer cycle time.