Spectroscopic Reflectometer: Working Principle, Capabilities and How to Choose One
A spectroscopic reflectometer is a non-contact optical instrument that measures a thin film by recording how much light it reflects across a band of wavelengths. It is the thickness tool of choice where throughput matters: measurement is fast, it tolerates process vibration, and it can run inline without sample preparation. This guide covers what the instrument actually does, what it can and cannot report, and which published TDM specifications apply. If you are deciding between a reflectometer and an ellipsometer, that comparison is covered separately in the spectroscopic reflectometry vs ellipsometry guide. This page assumes you have already chosen the reflectometric route and explains what to look for in one.
What a Spectroscopic Reflectometer Measures
The instrument's primary output is a reflectance spectrum: the ratio of reflected to incident intensity at each wavelength. Everything else is derived from that curve. Depending on configuration and on what is already known about the film, a spectro reflectometer can report:
- Physical film thickness — the routine output, from the interference modulation in the spectrum.
- Reflectance and transmittance — directly measured, and the basis for colour, sheet-resistance and optical-loss checks.
- Refractive index (n) — recoverable when thickness is independently known or when the layer count is low enough that parameters do not correlate.
- Uniformity and mapping — thickness across a wafer, sheet or coated area, point by point.
- Process values — in in-situ configurations, wet-film thickness and its drift during coating.
Absorption-related outputs such as band gap depend on knowing the extinction coefficient k, which is where an ellipsometer or a combined platform is usually required.
Working Principle: Interference, Then Spectral Fitting
The measurement comes from optical interference, in four steps.
1. Broadband illumination. White light is directed onto the film through an objective, at a defined angle and numerical aperture.
2. Two reflections interfere. Part of the beam reflects at the air/film surface, part at the film/substrate interface. The two returned waves travel different path lengths; the difference is the optical path 2·n·d·cosθ, where n is the film's refractive index, d its thickness and θ the internal propagation angle.
3. The spectrum shows fringes. Where the two waves arrive in phase, reflectance rises; where they arrive out of phase, it falls. Across a wavelength sweep this becomes a repeating modulation — a fringe pattern. Thicker films pack fringes closer together, which is why a single measurement can resolve a wide thickness span without moving the sample.
4. Software fits a model, not a peak count. The instrument compares the measured curve with a calculated spectrum for a defined layer stack, adjusts thickness (and any permitted optical constant) until the curves match, and reports the best fit. This is why the result is a model-derived value: the reported thickness is only as valid as the layer order, materials and n/k assumptions supplied to the fit. A reflectometer also measures absolute intensity, so it requires a reference — a documented calibration practice, not an optional step.
Published Measurement Range and Resolution (TDM Models)
There is no universal range, resolution or accuracy for this instrument class. Those figures belong to a specific model, spot size, sample and test condition. The values below reproduce TDM's published, model-specific claims; where a model does not publish a numerical range, no range has been invented for it.
| Model | Stage | Published thickness range | Published resolution / accuracy | Other published conditions |
|---|---|---|---|---|
| SE200 | Offline | 4 nm – 450 µm | Repeatability ≤ 0.02 nm; absolute accuracy 0.2% or 1 nm | Spot 1–100 µm; 4–12 inch wafers; n/k, reflectance spectrum and mapping atlas outputs |
| SE100 | Offline | 5 – 1200 nm | Repeatability ≤ ±1% | Samples 10 × 10 to 220 × 220 mm; calculation under 10 s; combines ellipsometry, reflectometry and transmission |
| SE8000 | Offline | Not published | Not published | Imaging 3D thickness, multi-layer metrology, uniformity visualisation; full-range continuous spectrum at preset positions |
| SE600X | Inline | 5 – 1200 nm | 10 Hz sampling | 31 measurement points; no sample preparation; thickness with absorption, reflectance, transmittance and PL |
| SE310 | In-situ | 300 nm – 20 µm | Repeatability < 1%; 15 Hz | Spot 1.5–2 mm; working distance under 5 mm; continuous single-spot wet-film monitoring |
| SE31-FM | In-situ | Spectral range 350 – 2400 nm | Wavelength resolution < 0.2 nm; thickness accuracy ±0.1% on standard optical-film systems | Transmission/reflection dual mode; PLC integration |
| SE31 | In-situ | Not published | Not published | Wet-film thickness and reflectance at a slot-die coating head; SDK, Ethernet and machine start/stop triggers |
For the complete cross-model matrix, see film thickness measurement equipment by model and range. For definitions, units and the difference between repeatability, accuracy, resolution and detection limit, see thin film thickness.
Limits: Where Reflectometry Stops Being the Right Tool
Stating these limits is part of choosing the instrument, because each one is a spec you must confirm before purchase.
- Very thin films with unknown optical constants. Fringe amplitude shrinks as the film thins, so below roughly 10 nm the signal is weak unless n is well known and assumed. Above about 100 nm the modulation is comfortable, which is where reflectometry is strongest.
- Unknown multilayer stacks. Each additional free layer adds a parameter. Beyond two or three unknown layers the fit stops being unique — thickness and index trade against each other.
- No optical contrast. If the film index is close to the substrate or surrounding medium, there is almost nothing to fit. Index-matched systems are the classic failure case.
- Strongly absorbing films. k has to come from somewhere; if it is unknown and the film absorbs across the band, the fit needs constraints it may not have.
- Rough or scattering surfaces. Scattered light leaves the collection cone and depresses measured intensity, biasing the fitted thickness.
- Patterned or inhomogeneous areas. The spot averages whatever is inside it; a spot larger than the feature reports a mixture, not the film.
Typical Configuration and the Specifications That Decide Selection
Reflectometers are sold as a small number of subsystems. Specify each one independently rather than accepting a single "best" figure.
- Spectral range and wavelength resolution — set which films can be resolved. A thin film needs enough spectral points across its fringe period; a 350–2400 nm band with sub-0.2 nm resolution is a different capability from a visible-only instrument.
- Spot size and optics — determines whether you can measure a micro-area pad or must average over a whole die. Selectable 1–100 µm on SE200 is the difference between mapping a wafer edge and reading it.
- Measurement speed and sampling count — the throughput spec on a line. SE600X publishes 10 Hz at 31 points; SE310 publishes 15 Hz at a single spot. Cycle time, not resolution, usually decides an inline installation.
- Measurement stage — offline laboratory, inline after coating, or in-situ at the coating head. These are separate hardware and interface routes; wet-film monitoring during coating cannot be added to an offline instrument afterwards.
- Outputs required — thickness only, or thickness plus T/R, mapping, PL, band gap or n/k. If k-derived values are mandatory, select a platform that measures them.
- Model and reference handling — check which layer models, material libraries and calibration routines ship with the instrument, and whether you can save an approved model per product.
- Integration — PLC, Ethernet or SDK, machine start/stop triggers, working distance and environmental limits.
Where Spectroscopic Reflectometers Are Used
- Semiconductor and wafer films — dielectrics, antireflective and mask layers, thick films up to the hundreds-of-micrometres end, 4–12 inch wafer formats.
- Photovoltaic and perovskite production — full-stack thickness inline at line speed, alongside reflectance, transmittance and PL.
- Optical coatings — filter, AR and mirror layer verification, where thickness plus T/R is the acceptance criterion.
- Slot-die and roll coating process control — wet-film thickness and drift at the head, before drying fixes the defect in.
- Display, sheet and flexible films — large-area uniformity across sheet samples rather than single-point readings.
What to Send Before Requesting a Configuration
A reflectometer quotation is only meaningful with the stack in hand. Provide: layer order and materials, substrate, expected minimum and maximum thickness, whether n is known, sample dimensions and format, required spot size or number of mapping points, required outputs, cycle time, measurement stage, and the interface available on the line. Missing any of these, two different instruments can look correct — and one of them will not measure your film.
Spectroscopic Reflectometer: Quick Answers
What is a spectroscopic reflectometer?
A spectroscopic reflectometer is a non-contact optical instrument that records how much broadband light a film reflects as a function of wavelength, then fits that reflectance spectrum to an optical model to report film thickness and, where conditions allow, reflectance, transmittance and refractive index. It measures intensity rather than polarization change.
How does a spectroscopic reflectometer measure thickness?
Light reflected from the film surface and from the film/substrate interface interferes. The optical path difference 2·n·d·cosθ sets where constructive and destructive interference occur, so a film of thickness d produces a repeating modulation across the spectrum. The instrument fits the measured curve against a layer model to solve for d, and reports it together with the model used and the assumptions about n and k.
Can spectroscopic reflectometry measure multilayer films?
It can measure a known stack with a defined layer order, where each layer's optical constants are available or constrained. It cannot freely solve an unknown multilayer stack: as layer count grows, parameters become correlated and the fit is no longer unique. TDM's SE8000 publishes imaging 3D thickness and multi-layer metrology functions, but does not publish a numerical range, so multilayer capability should be confirmed per stack.
What thickness range do TDM reflectometric instruments publish?
Published ranges are model-specific. SE200 lists 4 nm–450 µm micro-area measurement with a 1–100 µm spot, repeatability ≤0.02 nm and absolute accuracy 0.2% or 1 nm. SE100 lists 5–1200 nm point measurement with repeatability ≤±1%. SE600X lists inline 5–1200 nm at 10 Hz across 31 measurement points. SE310 lists 300 nm–20 µm wet-film monitoring at 15 Hz. SE31-FM lists 350–2400 nm with wavelength resolution below 0.2 nm. SE31 and SE8000 do not publish a numerical thickness range.
Is a spectroscopic reflectometer the same instrument as an ellipsometer?
No. A reflectometer records reflected intensity; an ellipsometer records the polarization change of that light. The practical consequence is a speed-versus-precision trade-off, and some platforms combine both — TDM's SE100 publishes spectroscopic ellipsometry, reflectometry and transmission on one benchtop instrument. See the full comparison of when to use each technique in the spectroscopic reflectometry vs ellipsometry guide.
Related Guides and Products
- Spectroscopic Reflectometry vs Ellipsometry
- How to Measure Thin Film Thickness
- Thin Film Metrology for Semiconductor Manufacturing
- Thin Film Measurement Equipment Buyer's Guide
- Film Thickness Measurement Equipment by Model and Range
