Device critical dimension (CD) and morphology testing is a core metrology discipline that verifies the dimensional accuracy, geometric profiles, and surface topography of micro- and nano-scale structures in semiconductor wafers, photomasks, MEMS, and advanced packaging components. Accurate CD control directly determines device performance, power efficiency, and manufacturing yield, making this testing essential for lithography, etch, deposition, and process development in the microelectronics industry. This article explains the sample types, measurable parameters, standard testing methods, and typical applications of device critical dimension and morphology inspection.
Technical Overview
Critical dimension (CD) refers to the smallest feature size of a fabricated device structure, such as the linewidth of a gate, the width of an interconnect line, or the diameter of a contact hole. Morphology testing covers the three-dimensional shape and surface characteristics of these features, including sidewall angle, step height, trench depth, line edge roughness, and surface topography. Together, CD and morphology measurements define whether a patterned structure matches its design intent and whether a manufacturing process remains within specification.
As semiconductor technology nodes shrink below 10 nanometers and device architectures become three-dimensional, such as FinFET and gate-all-around structures, dimensional tolerances tighten to the sub-nanometer and angstrom level. Even small deviations in CD can cause significant variations in transistor threshold voltage, leakage current, resistance-capacitance delay, and overall circuit yield. For this reason, CD and morphology metrology is embedded at multiple stages of the manufacturing flow, from lithography overlay verification to final structural qualification.
Modern CD and morphology testing combines multiple complementary techniques. Electron microscopy provides high-resolution top-down and cross-sectional imaging, atomic force microscopy provides traceable three-dimensional profilometry, and optical methods such as white light interferometry and scatterometry provide fast, non-destructive, full-wafer capability. Measurement procedures and calibration practices are guided by international standards from organizations such as ISO, IEC, SEMI, ASTM, JIS, and GB, and accredited third-party laboratories routinely perform these measurements for process qualification, incoming inspection, failure analysis, and litigation or supply-chain disputes.
Testing Samples
Semiconductor Wafers and Patterned Device Structures
Patterned silicon, silicon-on-insulator, GaAs, GaN, and SiC wafers are the most common samples. Measurements are performed on gates, fins, spacers, trenches, STI structures, metal interconnects, contact holes, and damascene patterns at various process steps, including after-lithography, after-etch, and after chemical mechanical planarization.
Photomasks and Reticles
Quartz, ArF, and EUV photomasks require CD uniformity verification on opaque and phase-shift features. Key measurements include main feature CD, linearity, pitch, corner rounding, defect dimensions, and phase and transmission-related structural profiles, which directly transfer into wafer-level pattern fidelity.
MEMS and Micro-Sensor Structures
Micro-electromechanical devices contain deep trenches, cantilevers, membranes, comb drives, and cavities with high aspect ratios. Both lateral dimensions and vertical profiles such as etch depth, sidewall verticality, and surface roughness must be characterized to ensure mechanical resonance, sensitivity, and reliability.
Advanced Packaging Components
Samples include through-silicon vias (TSV), copper pillars, solder bumps, redistribution layers, wafer-level fan-out structures, and bonding interfaces. Critical measurements include bump height and coplanarity, bump diameter, RCD linewidth and spacing, TSV depth and profile, and underfill or molding compound topography.
Thin Films, Coatings, and Display Structures
Transparent conductive oxides, OLED and TFT pixel banks, color filters, photoresist coatings, and functional thin films require step-height, film-thickness, and surface-roughness characterization to evaluate deposition uniformity, etch completeness, and planarization quality.
Testing Items
Typical dimensional and morphological parameters measured in device CD and morphology testing include:
- Critical dimension (CD) and linewidth of lines, trenches, gates, and traces
- Pitch, spacing, and periodicity of repetitive structures
- Contact hole, via, and trench diameter or width
- Sidewall angle (SWA) and profile shape, including re-entrant and tapered geometries
- Trench depth, etch depth, and via depth
- Step height between adjacent material layers or process levels
- Film thickness and film thickness uniformity
- Line edge roughness (LER) and line width roughness (LWR)
- Surface roughness parameters such as Ra, Rq, Rz, Sa, and Sq
- Aspect ratio and depth-to-width ratio of etched features
- CD uniformity (CDU) across die, wafer, and lot
- Buried and recessed feature geometry, such as recess depth in fin or channel structures
- Full-field surface topography, warpage, and planarization maps
- Bump height, coplanarity, and volume for interconnect structures
Testing Methods
Critical Dimension Scanning Electron Microscopy (CD-SEM)
CD-SEM is the workhorse of inline dimensional metrology. A focused electron beam scans the patterned surface, and secondary electron intensity profiles are converted into edge positions with sub-nanometer precision. It provides fast, largely non-destructive top-down measurements of CD, pitch, LER, LWR, and CDU, and supports full-wafer automated pattern recognition for high-throughput process control.
Atomic Force Microscopy (AFM and CD-AFM)
AFM uses a sharp probe tip that interacts with the sample surface to reconstruct a true three-dimensional profile. CD-AFM employs flared or boot-shaped tips to measure sidewall angle, top width, bottom width, and depth of trenches and holes in a single scan. Because AFM measurements can be traceably calibrated, CD-AFM is widely used as a reference method to qualify and cross-check other inline metrology tools.
White Light Interferometry and Optical Profilometry
White light interferometry (also called vertical scanning interferometry) measures surface height by analyzing interference fringes generated by optical path differences. It delivers rapid, non-contact areal topography with sub-nanometer vertical resolution, making it ideal for step height, film thickness of transparent layers, trench depth, warpage, bump coplanarity, and full-field roughness mapping on areas from microns to full wafers.
Stylus Profilometry
Stylus profilometry drags a diamond tip across the surface along a programmed scan path. It provides robust step height, film stress-derived wafer curvature, and depth measurements over long scan lengths, and is frequently used for thicker films, etch depth on large features, and process monitor wafers where direct mechanical contact is acceptable.
Cross-Sectional SEM with FIB Milling and TEM
Focused ion beam (FIB) milling prepares precise cross sections through specific device locations, which are then imaged by SEM to reveal sidewall profiles, layer stacks, CD at multiple heights, and etch or deposition anomalies. For the most advanced nodes, transmission electron microscopy (TEM) provides angstrom-scale imaging of gate profiles, liner and barrier layers, and interface morphology. These are destructive, location-specific reference techniques used for process qualification and failure analysis.
Scatterometry and Optical CD (OCD) Metrology
Scatterometry illuminates periodic arrays with polarized light and fits the measured reflectance spectrum against a rigorous electromagnetic model. It reconstructs multiple parameters simultaneously, such as CD, sidewall angle, height, and film thickness, with excellent precision and speed, and is suitable for in-line monitoring of dense arrays where direct imaging is difficult, including three-dimensional structures.
Confocal Laser Scanning and Confocal Chromatic Microscopy
Confocal optical methods provide high-contrast height discrimination on reflective, transparent, or steep surfaces. They are widely applied to packaging samples, bonding surfaces, and roughness evaluation where large-area coverage and non-contact measurement are required.
Applications
Device critical dimension and morphology testing supports the entire microelectronics product lifecycle:
- Lithography process development and exposure dose or focus optimization, including CD bias and focus-exposure matrix analysis
- Evaluation of etch recipes, including etch bias, sidewall profile control, depth uniformity, and loading effects
- Deposition and chemical mechanical planarization monitoring through film thickness and step-height uniformity mapping
- Photomask incoming inspection and CD uniformity qualification before release to production
- MEMS structural qualification, including deep reactive ion etch profile, stress-induced deformation, and cavity geometry
- Advanced packaging process control for TSV, micro-bump, copper pillar, and redistribution layer geometry
- New product introduction benchmarking and process window establishment in research and development
- Failure analysis, yield excursion root-cause diagnosis, and structural comparison against design targets
- Supply-chain acceptance testing and independent verification by accredited third-party laboratories
FAQ
What is the typical measurement accuracy of CD and morphology testing?
Measurement capability depends on the technique. CD-SEM typically achieves sub-nanometer precision on dense line features, CD-AFM provides traceable three-dimensional measurements with sub-nanometer vertical resolution, white light interferometry reaches sub-nanometer vertical resolution over large fields, and TEM cross-sectional imaging resolves sub-angstrom structural details. The appropriate method is selected based on feature size, material properties, target uncertainty, and whether destructive sampling is acceptable.
Are these tests destructive to the sample?
Top-down CD-SEM, AFM, scatterometry, and optical profilometry are essentially non-destructive and can be performed on product wafers or delivered parts. Cross-sectional SEM with FIB preparation and TEM analysis are destructive because they require cutting through the region of interest, so they are normally applied to dedicated monitor dies or sacrificed samples.
Which method is best for high-aspect-ratio structures such as deep trenches and TSVs?
For deep trenches and vias, CD-AFM with flared tips, cross-sectional FIB-SEM, and optical profilometry with long-working-distance objectives are commonly combined. Scatterometry is also effective for periodic deep-structure arrays because the optical response encodes the full profile geometry rather than only the top opening.
How is measurement traceability ensured?
Reputable laboratories calibrate instruments using certified reference materials and follow documented procedures aligned with internationally recognized standards from organizations such as ISO, IEC, SEMI, ASTM, JIS, and GB. Measurement uncertainty budgets, tool matching protocols, and gauge repeatability and reproducibility studies are established before release of production data.
What information should be provided when submitting samples for testing?
Clients should provide the design layout or measurement location coordinates, target dimensions and tolerances, material stack information, the process step of interest, and any specific reporting requirements. This allows the laboratory to select the correct metrology platform, locate features efficiently, and report results in a format suitable for engineering decision-making.
