Skip to content

Worked example

A coated sample, one question, and where the evidence points.

A team has a coated sample. They want to know the coating’s thickness and layering and how it behaves optically, from the visible into the infrared — and they would rather not split the work across the country. This is what the answer looks like, fixed from the current dataset.

1 · the question, as typed

Plain language is enough.

Example input

We need XRR in Enschede plus UV-Vis spectroscopy and IR emissivity for a coated sample. Can one facility cover all three?

2 · the direction

Two methods, and what each cannot tell you alone.

The Field Agent separates what the team wants to know from the methods that measure it, and lists the questions a laboratory would ask back.

XRRPrimary

X-ray reflectivity — for thickness and layering

Models film thickness, density and interface roughness. The answer depends on the assumed layer stack and says nothing about chemistry by itself.

OPTCHARComplementary

Optical characterisation — for the UV–Vis response and infrared emittance

Spectral range, temperature, angle, and whether emissivity is measured directly or inferred still have to be fixed before a laboratory can answer.

Questions that change the plan

  • What are the coating, substrate and expected thickness range?
  • Which UV–Vis and infrared wavelength ranges matter to the decision?
  • At what temperature and observation angle must emissivity be established?
  • What sample dimensions, preparation limits and damage budget apply?

3 · the laboratories

What the current evidence supports: a plan across three laboratories.

Nothing in the dataset we hold shows one facility covering all three needs, so the result is a multi-facility plan: the Enschede preference is kept for XRR and the two optical needs are covered elsewhere rather than returning nothing. That describes the evidence we hold, not what exists. Each lead shows its source and its evidence level.

Leads, not bookings. No laboratory was contacted. Sources were retrieved on 29 August 2026 and not independently reviewed for current operation; configuration, sample acceptance, availability, turnaround and price remain to be confirmed with each laboratory. A live answer differs as the problem and the evidence change, and stays bounded to three leads per search.