Wet Film Thickness Is a Control-Loop Problem, Not a Resolution Problem

Q: How do I calculate wet film thickness?

On a slot-die line the set point is arithmetic, not measurement: wet thickness follows from volumetric flow rate divided by coating width and web speed (solids content then converts it to dried thickness — dried ≈ wet × solids fraction). What a sensor adds is not the formula, it is knowing whether the line is actually delivering that number: the published SE310 samples at 15 Hz, which is one reading every 0.67 mm at 10 mm/s and every 3.3 mm at 50 mm/s.

Q: How to read a wet film thickness gauge?

A contact wet film gauge is a comb pressed into the fresh coating; you read the highest wetted scale value, which makes it destructive, single-point and operator-dependent. Non-contact reflectometry reads before the film leaves the head, returns a value per sample period rather than per operator walk-by, and leaves the substrate untouched. TDM's published wet-film instrument (SE310, spot 1.5–2 mm, working distance under 5 mm) is the second type.

Q: How to measure film thickness?

Wet and dry are two different purchases. Wet film during coating needs an in-situ channel at the head (see the sampling arithmetic above). Dry film needs either an offline bench or an inline station: published ranges are SE100 at 5–1200 nm and SE200 at 4 nm–450 µm offline, SE600X at 5–1200 nm inline at 10 Hz across 31 points. The full model-by-model matrix is on the reflectometer page.

Q: What is the 80/20 rule for DFT measurement?

It is a dry-film acceptance convention from coating inspection: most readings must reach the specified DFT, and the remainder may fall below it but only within a stated margin — the exact threshold comes from the standard or contract in force, so confirm it there rather than from a blog. What matters for metrology is that the rule is a statistical acceptance test over many points, which is exactly what a continuous inline channel produces and what a handful of manual spot checks cannot support.

Wet-film thickness measurement during slot-die coating is not a precision problem. It is a control-loop problem. The specs that decide whether it works are sampling rate, dead time and working distance. Resolution is what makes the correction trustworthy. Getting them in the wrong order is why in-situ channels get installed and then get switched off. This guide covers the arithmetic and the geometry that decide whether a wet-film channel can actually close a loop.

Measurability is a product of three things

A closed correction loop has one requirement the lab never has to meet: the sample has to exist before the film dries. Write it as a condition:

samples per unit length = sampling rate ÷ line speed

That ratio, not the accuracy figure, is what determines whether you have a control signal or a very expensive historical record.

The SE310 slot-die wet film metrology module publishes 15 Hz continuous single-spot sampling, a 1.5–2 mm spot, <1% measurement repeatability and a 300 nm–20 µm wet-film thickness range. Put the published 15 Hz against some line speeds and the shape of the problem appears:

Distance per reading at 15 Hz at various line speeds
Line speed	Substrate per reading	vs a 1.5–2 mm spot
10 mm/s	0.67 mm	overlapping — redundant edge-to-edge coverage
30 mm/s	2.0 mm	abutting — one independent sample per spot width
50 mm/s	3.3 mm	gapped — the film between readings is unsampled
100 mm/s	6.7 mm	sparse — short-lived excursions can fall between samples

Line speeds here are variables for you to substitute, not claims about any line. The arithmetic is `distance = v ÷ 15 Hz`.

Two readings follow from it. Below roughly 30 mm/s the optic is oversampled — extra readings buy averaging, not information. Above it, the sampling grid coarser than the spot means a transient pump pulsation can pass through the coated area unseen. The same 15 Hz is either redundant or marginal depending only on a number that is yours.

Dead time is not the same as processing time

The published data processing time is <1 s. That is the instrument's contribution to latency, not the loop's. Downstream of it sit the Ethernet hop to the computing server, the SDK, whatever your controller does with the value, and the actuator lag on pump speed or gap height. Add them before you judge the loop against your process.

At 50 mm/s, a one-second response is 50 mm of already-coated substrate. Nothing the measurement does can recover that length. So the honest question is not "how fast is the sensor" but "how much film am I willing to lose per correction" — and the answer sets the sampling rate you need, not the datasheet.

Working distance is a control decision

The published working distance is <5 mm, and the module is specified to mount on the coating head with the slot-die machine providing start/stop trigger signals. That is a geometry choice with a consequence: the measurement happens at the point of deposition, so the correction applies to the film that has not been made yet.

A sensor standing 15–25 mm off the surface is measuring a downstream location, and every millimetre of standoff is process drift you cannot attribute. If the optics can sit on the head, they should — because that placement is the whole reason the channel is called in-situ.

The trade-off is real and physical: <5 mm inside a coating enclosure means solvent vapour, temperature and mechanical vibration live with the optic. The published <500 g module at 190 × 176 × 70 mm, 24 V, with an Ethernet link to the computing server, is small enough to go on the head — which is why the head is where it belongs.

Limits: Where Wet-Film Metrology Stops Being the Right Answer

State these before sizing anything, because each one is a disqualifier, not a caveat.

Opaque or highly absorbing inks. The method reads interference fringes. A carbon-loaded or metal-paste coating returns no usable fringe structure.
Index-matched systems. If the wet ink's refractive index sits close to the substrate or an underlying layer, there is almost nothing to fit.
Very thin wet films. Fringe amplitude shrinks as the film thins; a sub-micron wet layer is inside the published 300 nm floor of the SE310's 300 nm–20 µm range.
Cross-width uniformity. Published sampling mode is continuous single spot — a high-rate time series at one location, not an edge-to-edge map.
Post-dry verification. Once solvent leaves, the wet number is history. Dried-film QC stays a separate measurement.
Rough or scattering substrates. Scattered light leaves the collection cone and biases the fitted thickness.

Which TDM Instrument Handles Which Wet-Film Job

TDM in-situ wet-film instruments — published figures only
Model	Stage	Published wet-film range	Published rate / repeatability	Published geometry
SE310	In-situ, slot-die head	300 nm – 20 µm	15 Hz; repeatability < 1%	Spot 1.5–2 mm; working distance under 5 mm; continuous single spot
SE31	In-situ, slot-die head	Not published	Not published	Thickness and reflectance at the coating head; SDK, Ethernet, machine start/stop triggers
SE31-FM	In-situ, optical films	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

Where a model does not publish a numerical range, none has been invented for it. The full cross-model matrix, including offline and inline instruments, is maintained on the reflectometer page and on film thickness measurement equipment by model and range — those two pages are the source of record; if this table and they ever disagree, they govern.

If your question is about a *dried* film on a completed part, this is not the right page — see how to measure thin film thickness.

What a single spot does not tell you

The SE310's published sampling mode is continuous single spot. Be precise about what that buys you: a high-rate time series at one location across the web.

It is not a cross-width uniformity map. Coat-to-coat drift at the centreline is exactly what single-spot sampling is good at; edge-to-edge thickness variation is a different question and needs either multiple heads or a scanning measurement — decide which problem you are solving before you size the system.

Worth flagging: SE310 is also the model number of our offline 3D optical profiler (light-section triangulation). Same number, different instrument. Everything above refers to the slot-die wet film metrology module.

Before you buy a wet-film channel

Fix your line speed first. Rate ÷ speed gives the sampling grid. Nothing else matters until that ratio is acceptable.
Budget the whole loop, not the instrument's <1 s.
Confirm your film is measurable. Interference-fringe reflectometry needs a transparent or semi-transparent wet film on a reflective or partially reflective substrate. Carbon-loaded and metal-paste coatings do not return usable fringes.
Map wet to dried through your solids fraction (dried ≈ wet × solids fraction) before setting a target; the coater is controlled in wet units, the spec is written in dried units.
Run a feasibility test on your own ink. Refractive index that moves with temperature or evaporation rate degrades any reflectometric readout.
Decide single-spot versus cross-width — the two are not the same purchase.

Frequently Asked Questions

How do I calculate wet film thickness?