Every thermal measurement method in this knowledge hub is excellent — inside its jurisdiction, and unreliable outside it. The flash method owns the bulk disc1; the hot disk owns the clampable solid2; 3ω owns the fabricated film3; thermoreflectance owns the nanoscale layer and its interfaces4; the micro-bridge is the most established route for the single nano-object5; and the suspended electrothermal family owns everything filamentary, free-standing and micro-scale that the rest cannot hold6. Method choice is therefore not a ranking exercise — it is a matching exercise, and the matching key is geometric before it is anything else. This hub page compresses the entire decision into a six-question interactive selector, backs it with the reasoning, and links every branch to its deep-dive article.
- 1Questions 1–3: form, support state, direction
- 2Interactive: the method selector
- 3Question 4: electrical and optical behavior
- 4Question 5: characteristic scale
- 5Question 6: what temperature range?
- 6The jurisdiction map in one table
- 7The honest gray zones
- 8Frequently asked questions
- 9Keep exploring the knowledge hub
- 10References
Questions 1–3: form, support state, and measurement direction
Geometry eliminates more candidates than any other question, because each architecture physically requires a sample form. A self-supporting disc or block enters the flash method’s home ground — standardized, furnace-ready, six decades validated17 — with the transient plane source as the low-preparation alternative for anything that can be clamped in two halves28. A supported film belongs to the fabricated-heater and pump-probe families: 3ω for dielectric films down to tens of nanometers39, thermoreflectance where interfaces and sub-100-nm layers are themselves the question410. A fiber, wire, filament or free-standing strip is the suspended electrothermal family’s native geometry — the sample bridges two electrodes and serves as its own heater and thermometer611. A 2D crystal splits by state: suspended membranes to optothermal Raman12, supported sheets — the state devices actually use — to the supported-differential electrothermal protocol13. A single nano-object too small to be its own instrument goes to the microfabricated suspended bridge514.
Support state splits several of those branches in two. A film freed into a strip and a film on its substrate are different measurement problems — the strip suspends like a fiber6, the supported film belongs to fabricated-heater and pump-probe methods34 — and a 2D crystal’s suspended and supported states differ in both the applicable methods and the number itself1215. Measure the state the application uses.
Measurement direction can move the same sample between families outright. Cross-plane transport of a supported film is 3ω and TDTR territory34; the in-plane direction of the same film needs deliberately anisotropy-sensitive configurations or a freed strip; an oriented bulk solid’s two axes may require two different setups16; and an interface conductance question points at TDTR’s layered fitting regardless of what the sample otherwise is10. A direction left unstated is the single most common source of mismatched expectations in quoted numbers.
Interactive: the method selector
Answer the six questions below and the selector returns the geometry-native primary method, the credible alternates, and the reasoning — with links straight into each method’s deep-dive article. The map above the controls places the principal method families along the sample-scale axis; seven method labels are shown. It places them by the sample scale they were built for.
Question 4: electrical and optical behavior
Electrical character mostly selects the variant inside a family rather than the family itself — with the caveat that whether a coating is permissible, and whether a usable optical response exists, can genuinely redirect the route. In the suspended electrothermal family, a conductive specimen carries its own sensing current directly6; an insulating fiber takes the metallic-coating route, with the composite correction stated explicitly and validated on known materials17; a current-sensitive or delicate specimen can take laser heating with resistive readout in the photo-electro-thermal variant18. On the fabricated-heater side, the sample itself need not conduct at all — 3ω fabricates its own conductive line, requiring only electrical isolation from the sample: free on dielectrics, one insulating spacer layer on conductors3. Optothermal Raman ignores electrical character entirely and asks instead for a temperature-sensitive Raman peak12. The variant choice is where measurement-quality disciplines attach: zero-power extrapolation removing the modeled first-order self-heating bias within the validated local-linear regime19, differential designs subtracting what models would otherwise guess20.
Question 5: characteristic scale
Scale is not decoration on the other five answers — it re-ranks and sometimes vetoes them. At nm–μm, a supported film’s cross-plane question becomes an interface-and-thin-layer problem where pump-probe sensitivity leads and a fabricated heater line is the differential alternative; at μm–mm the heater-line route leads; at mm–cm a “film” has quietly become a bulk sheet and the disc methods own it. And when the declared form and the declared scale cannot coexist — a bulk disc at nanometers, an atomic layer at centimeters — the selector refuses to guess: it flags the combination as internally inconsistent and asks what the specimen physically is before recommending anything.
Question 6: what temperature range?
Temperature decides which parasitic channel leads and which defense is mandatory. High temperature: radiation grows as T³, so flash work leans on its Cape–Lehman and standardized loss corrections217, suspended work graduates from modeled corrections to the experimental length-series subtraction6, and under validated heater geometry, frequency and temperature conditions, the 3ω architecture can strongly suppress the relative contribution of radiation3. Cryogenic: sensing coefficients and signal levels shrink, power discipline tightens, and the evacuated-cryostat architectures — micro-bridge foremost — own the deep range where phonon physics unfolds519. Near room temperature: the widest field of candidates, where preparation economics and the uncertainty budget, not survival, decide2223. And at every temperature, suspended architectures share one prerequisite: vacuum below the geometry’s Knudsen knee, verified by a pressure-independence check rather than assumed6.
The jurisdiction map in one table
| Sample form | Primary method | Credible alternates | Deciding factor |
|---|---|---|---|
| Bulk disc / block | Laser flash1 | TPS2; 3ω on dielectrics3 | Standardization vs preparation economics |
| Clampable solid, minimal prep | TPS (Hot Disk)28 | LFA if disc machinable1 | Window discipline vs disc machining |
| Supported film · cross-plane | 3ω (dielectrics)39 | TDTR, esp. sub-100 nm + interfaces410 | Interface sensitivity and parameter identifiability vs instrument and modeling complexity |
| Supported film · in-plane | Anisotropy-sensitive TDTR (beam-offset / multi-frequency) or specialized 3ω | Release the film and measure the strip electrothermally | Hard direction: anisotropy-sensitive configurations or re-preparation |
| Fiber / wire / free strip · axial (cross-plane → purpose-built) | Suspended TET family611 | Steady-state / pulsed-laser variants2425 | Transient vs steady signal quality |
| 2D crystal, suspended | Optothermal Raman12 | Micro-bridge for flakes5 | Absorbance budget vs placement art |
| 2D crystal, supported · in-plane (cross-plane/interface → TDTR) | Supported-differential TET13 | Raman on supported state15 | Application-faithful state |
| Single nano-object | Suspended micro-bridge514 | TDTR where a transducer lands4 | Background + contact budgets |
The honest gray zones
Real specimens straddle jurisdictions, and pretending otherwise produces confident wrong numbers. Anisotropic bulk: a through-thickness flash reads one tensor component; the in-plane number needs dedicated configurations or a different family, and reporting one axis as “the” conductivity is the classic datasheet failure16. Coated and layered stacks: every method returns a composite unless the decomposition is explicit — TDTR fits it with sensitivity analysis10, the coating protocol subtracts it with a validated correction17, and anything less is an assumption wearing a number. Thin fast plates: too thin for the TPS window, too thick for film methods — run the window arithmetic before booking instrument time2. Ensembles: mats, yarns and bundles measure their junction networks, not their constituents; decide first whether the application needs the object’s number or the network’s1426. In every gray zone the resolution is the same: state what was measured, in what state, with what budget22 — and when the choice is genuinely borderline, a specimen-specific feasibility review with the thermal testing team settles in one exchange what a comparison table cannot.
Frequently asked questions
Which thermal measurement method is the most accurate?
Inside their jurisdictions, the mature methods all reach the few-percent regime — standardized flash on conforming discs7, validated TPS across four decades of k8, protocol-audited suspended TET on filaments19. Outside them, accuracy is undefined, not merely degraded. The question that has an answer is jurisdictional: which method’s assumptions does your sample actually satisfy?
My sample fits two methods — should I run both?
When stakes justify it, yes: cross-family agreement provides a strong independent check and reduces the plausibility of dominant uncorrelated biases — it does not eliminate systematic effects shared by both methods, the strongest validation available — optical and electrothermal families corroborating on 2D materials is the canonical example1227. For routine work, one method inside its jurisdiction with a stated budget is sufficient and honest22.
What single mistake most often ruins method choice?
Forcing the sample to fit a method on hand instead of matching the method to the sample — pressing fibers into a disc, averaging an ensemble and calling it the object, quoting a suspended-state number for a supported application1314. Geometry first; instruments second.
Keep Exploring the ACS Thermal Metrology Knowledge Hub
This decision hub connects every chapter of the ACS thermal metrology knowledge hub:
- Thermal conductivity & diffusivity testing: the pillar guide — the full landscape and buyer’s framework.
- Laser flash analysis explained — the bulk-disc standard and its six failure classes.
- Transient plane source (Hot Disk) explained — the pragmatist’s method and its time window.
- The 3ω method explained — frequency-domain rigor for films and solids.
- TDTR explained — picosecond access to films and interfaces.
- Raman thermometry for 2D materials — the optical route and its error budget.
- Suspended micro-bridge methods — single nano-objects, two budgets.
- The TET technique: a complete guide — the suspended family’s foundation.
- Thermal testing services — send the sample; get defensible numbers back.
References
This decision guide summarizes method jurisdictions for educational purposes; recommendations are starting points for a specimen-specific feasibility review, not a substitute for one. The interactive selector encodes the same tree with links to the deep-dive articles. For formal assessment and quotes, contact our thermal testing team.