A 2025 line-spot confocal prototype demonstrated 30 nm resolution and a 0.013 μm standard deviation across an 800 μm measurement range, beating its point-sensor benchmark of 0.064 μm SD on the same setup [S4].
Point-spectral confocal displacement sensors remain the workhorse for single-point sub-micron gauging on polished, transparent, or liquid targets, while line-spectral variants extend the same physics into parallel 1D profiling of lithium-battery wafers, optical component topography, and semiconductor stacks [S4][S1].
Operating Principle: Chromatic Aberration as a Position Encoder
Both architectures share the confocal chromatic basis: a broadband white-light source is dispersed by a multi-lens objective with strong controlled axial chromatic aberration, so each wavelength focuses at a different axial position and only the wavelength conjugate to the target surface returns peak intensity through a confocal pinhole to a spectrometer [S5][S8].
The peak-wavelength-to-distance mapping is calibrated at the factory, and the spectrometer reports a single peak for a discrete surface or multiple peaks when the beam crosses transparent layers, enabling one-sided thickness measurement with refractive-index correction [S5]. Micro-Epsilon controllers using this principle hit measurement rates up to 30 kHz with dynamically regulated CCD exposure, and the same controllers expose Ethernet, EtherCAT, Profinet, EtherNet/IP, RS422, and analog outputs to plant networks [S5]. For background on how the displacement map relates to other sensing families, see the displacement sensor reference page and the confocal displacement sensor entry.
Point-Spectral Sensor: Single-Axis Precision at Low Throughput
Point-spectral sensors concentrate the dispersed white light into one micro-spot, typically 3-25 μm diameter across the full range, and resolve that single point to nanometer-class repeatability on polished mirrored, transparent, matt, or liquid targets [S5]. Tilt tolerance up to 48° is achievable on commercial units, which is the reason they are specified over laser triangulation when a curved or angled surface would otherwise decouple a triangulation sensor's geometry [S5].
For an in-depth treatment of how that resolution-range tradeoff plays out in selection, see Confocal Displacement Sensor: Resolution-Range Tradeoff and Selection Rules. The mechanical architecture is coaxial, which is why the technology is also called a coaxial spectral confocal instrument: the same lens path is used for illumination and collection, suppressing specular noise that hurts triangulation on shiny parts [S2].
Line-Spectral Sensor: Parallel Acquisition, Larger Optics, Higher Cost
Line-spectral confocal sensors replace the single micro-spot with an elongated line focus, typically produced by a cylindrical lens plus a binary diffractive element, so a 2D camera or line-CCD can read out N axial channels simultaneously [S4]. The 2025 anti-interference line-spot design reported a 3.8 mm line length, 0.2 mm line width, 800 μm measurement range, and 30 nm resolution after calibration on a nanometer displacement stage [S4].
For longer profiles, a 2024 line-swept design combined a prism-grating spectrometer with a custom dispersive objective and demonstrated 24 mm scan length, 0.8 μm axial resolution, and 3.9 mm dispersion range across the full optical path [S1]. GRIN-lens variants shrink the head into ultra-small probe form factors for endoscopic or tight-clearance applications, at the cost of reduced measurement range compared with multi-element refractive objectives [S2]. Pricing on these probes scales with optics, spectrometer complexity, and channel count; per-channel economics are broken out in Confocal Displacement Sensor Pricing per Channel: What Drives the Bill.
Side-by-Side Spec Matrix
Decision criteria line up as follows. Spot geometry: point sensors use a 3-25 μm micro-spot, line sensors produce a 3.8 mm (binary-diffractive) to 24 mm (prism-grating) elongated focus [S5][S4][S1]. Measurement range: point heads span hundreds of μm to a few mm per optical stack, while the demonstrated line-spectral prototypes reach 800 μm (3.8 mm line) and 3.9 mm dispersion (24 mm line) [S4][S1]. Resolution: point-spectral units hit nanometer-class repeatability on static targets, whereas the line-spot prototype reports 30 nm resolution with 0.013 μm SD versus 0.064 μm SD for the point sensor in the same test rig [S4][S5]. Throughput: a point sensor reads one Z value per exposure, so full-field profiling needs XY raster scanning; a line sensor reads N values per exposure, cutting the number of mechanical axes moves for 1D profiles [S4].
Tilt tolerance, optics complexity, and cost are the trade axes that close the matrix: point heads tolerate up to 48° tilt and use simpler multi-lens optics, while line heads add a cylindrical plus diffractive element, a 2D imaging spectrometer, and more demanding alignment, which raises unit cost but removes the raster-stage bottleneck for inline profilometry [S5][S4]. For alternative non-contact gauging on similar targets, the laser displacement sensor and capacitive sensor pages contrast triangulation and capacitive tradeoffs against the chromatic-confocal family.
Selection Rules by Use Case
Specify point-spectral confocal when the target is a single known feature (polished mirror, lens vertex, liquid level, glass thickness, wafer flatness spot) and the budget favors one axis of motion, nanometer repeatability, and 48° tilt tolerance [S5]. Point heads also fit clean rooms, R&D benches, and lab-grade thickness gauges where the smallest possible probe envelope matters more than areal throughput [S3].
Specify line-spectral confocal when the application requires a 1D profile in a single pass, such as lithium-battery electrode thickness, wafer warpage strips, optical-component surface topography, or any line of transparent layers that must be read simultaneously for inline SPC [S4][S1]. The line-spot architecture in particular improved dynamic stability by roughly 5x (0.013 μm vs 0.064 μm SD) over the point sensor in vibration-prone measurements of lithium-battery wafer thickness, which is the headline reliability gain for production-floor deployment [S4].
Known Failure Modes and Constraints
Point sensors inherit low scan efficiency, since they only acquire one Z value per exposure, so any full-field surface recovery has to be reconstructed from XY motion, which slows 3D topographic measurement and amplifies the impact of stage drift [S1]. Line sensors address throughput but at the price of larger, more complex optics and stricter alignment, and the longer the line focus the more vulnerable the system is to angular deflection and surface-quality variation across the line, which is the main motivation behind the 2025 anti-interference line-spot design [S4].
Both architectures assume the wavelength-to-distance mapping is stable; thermal drift, mechanical flexure, and refractive-index mismatch in multi-layer thickness measurement all shift that mapping, so factory calibration plus real-time exposure control are the standard mitigations, with confocal sensors overall characterized as more thermally stable than laser triangulation or eddy-current alternatives in published comparisons [S5]. Signal-to-noise ratio drops when the peak broadens, which happens on rough or low-reflectivity surfaces, and adaptive correction, peak-fitting with neural networks, and angular-incidence color confocal models are the active research directions for tightening that error budget [S4].
Where the Technology Is Heading Next
Three verifiable signals to track over the next 6-12 months: line-spectral designs extending scan length past the 24 mm mark while keeping axial resolution in the sub-micrometer band [S1]; ultra-small GRIN-lens probes commercialized for medical, endoscopic, and tight-clearance inspection where conventional multi-element objectives will not fit [S2]; and binary-diffractive plus cylindrical-lens line-spot modules moving from ZEMAX simulation into production cells for lithium-battery thickness and warpage control [S4]. Watch for revised commercial datasheets on the SCI-series class of heads, which already cover sub-micron displacement and thickness measurement on transparent, polished, and liquid surfaces, and which sit at the same architectural fork between point and line spectral designs [S3].