The most consequential limitation a measurement technique can have is a class of materials it simply cannot see. For electrothermal methods that blind spot ought to be enormous: polymers, glasses, ceramics, biological fibers — the insulating majority of interesting materials — carry no current, so they can neither heat themselves nor report their temperature through resistance. The metal-coating protocol substantially narrows that blind spot — wherever deposition compatibility and the composite model can be validated — with a few nanometers of sputtered metal and one piece of transport physics. The coating makes the sample measurable; the Wiedemann–Franz relation helps estimate the film’s thermal conductance — subject to geometry, continuity and effective-Lorenz-number assumptions — so its share can be removed within a composite model. What remains is the insulator’s own thermal story, told through a conductor’s voice.
- 1The blind spot
- 2What the coating does — and adds
- 3Wiedemann–Franz: computing the film’s share
- 4Interactive: subtract the coating
- 5Choosing the measurement path
- 6What the protocol has unlocked
- 7The Lorenz number honesty clause
- 8Protocol discipline
- 9Frequently asked questions
- 10Keep exploring the knowledge hub
- 11References
The blind spot
TET’s elegance rests on the sample serving as its own heater and thermometer1 — which quietly presumes the sample conducts electricity. A polymer fiber, a glass filament, a ceramic whisker, a strand of spider silk: none can pass the heating current or register a temperature-dependent resistance. Yet these are precisely the materials whose thermal transport is least documented and most structure-sensitive, where processing decisions swing conductivity by large factors and datasheet values barely exist. A characterization family that stopped at conductors would forfeit the samples that need it most.
The need is sharpened by how these materials behave. Insulating fibers and films are overwhelmingly processing-defined: drawing ratio, crystallinity, moisture history and cross-linking move their transport properties over ranges that dwarf typical measurement uncertainty, so a literature value — where one even exists — describes a cousin of the specimen at hand, not the specimen itself. The practical consequence is that whole industries size thermal margins for polymer packaging, insulation layers and natural-fiber composites from estimates, because the specimens that matter were never individually measurable. A protocol that extends electrothermal measurement to suspended insulators — wherever deposition compatibility and the composite-model assumptions can be validated — converts much of that data vacuum into routine metrology.
What the coating does — and adds
The fix is disarmingly direct: sputter a nanometer-scale metallic film onto the sample’s surface — gold and iridium are the demonstrated choices, valued for chemical stability and controllable deposition, though continuity and substrate compatibility remain process-dependent — and the coated composite becomes electrically alive2. Current flows through the film; Joule heat is generated at the surface and conducts into the substrate; the film’s temperature-dependent resistance reports the composite’s thermal response. The measurement then uses the same suspended transient framework as a conductive specimen and reduces through the same routes3 — while the recovered property must be interpreted through the coating–substrate composite conductance and heat-capacity model.
But the measurement now sees a composite, and honesty requires saying so precisely. The film adds two things: a parallel thermal channel along the sample’s length, and a heat capacity contribution. The capacity term may be neglected only when ρfcp,fAf ≪ ρscp,sAs — often small for a nanometer film on a micrometer fiber, but a criterion to check quantitatively rather than assume, especially for very thin substrates, repeated coatings, low-density fibers or atomic-scale supports. The parallel conduction channel is never negligible by default — metals conduct heat well, and even a thin film’s contribution can rival a poorly conducting substrate’s own transport. Making the film thin helps, but thinness alone never establishes negligibility; the contribution must be computed and removed4.
The deposition itself lives inside a window. Below a material- and substrate-dependent threshold, a sputtered film is not yet a film but an archipelago — islands whose resistance is set by percolation hops rather than metallic transport, electrically noisy and thermally ill-defined. Above the window, the film’s parallel channel grows toward parity with a poorly conducting substrate and the eventual subtraction balloons. The craft is to sit just above continuity: thick enough for ohmic, stable, low-noise resistance; thin enough that the substrate remains the majority carrier of heat. Gold’s chemical inertness and iridium’s track record toward fine-grained continuity are what earned the two metals their standing in this role — though continuity remains substrate- and process-dependent, with gold in particular prone to island nucleation on many surfaces — and whatever the choice, verifying a stable, ohmic resistance before the thermal run is the checkpoint that keeps archipelagos out of the dataset.
Wiedemann–Franz: computing the film’s share
The removal leans on one of transport physics’ oldest regularities — but it must be applied to the right quantity. What the transient measurement returns is the composite’s effective response, in which substrate and film are parallel axial heat channels that also share the heat capacity. In quantities, the additive object is axial thermal conductance, not conductivity: the composite’s effective diffusivity behaves as
In a metal the same electrons carry charge and heat — ke = LσT — and for a highly insulating substrate the coated sample’s measured resistance is, to good approximation, the film’s. Convert it through the film’s Lorenz number and temperature into the film’s axial thermal conductance, remove that parallel channel from the composite’s conductance, account for the heat-capacity denominator, and the substrate’s transport remains. The modern review formalizes exactly this reduction, computing and eliminating the metallic coating’s contribution from the measured composite response4.
The procedure’s quiet strength is that its central input is measured, not assumed: the film’s actual resistance on this sample enters the correction directly. Measured resistance is not a cure-all — converting it into film thermal conductance still requires the film’s geometry, verified continuity, a defensible effective Lorenz number and a clear attribution of the current path — but it anchors the largest single input in data. The Lorenz number deserves its own section.
Interactive: subtract the coating
Choosing the measurement path by sample class
| Sample class | Path | Heater / thermometer | Subtraction burden | Representative cases |
|---|---|---|---|---|
| Metallic & conductive fibers | Direct TET | Sample itself / its own resistance | None (radiation & self-heating handled by the companion protocols) | Metal wires, carbon fibers5 |
| Semiconductive samples | Direct TET, characterization first | Sample / sample | None, but the resistance–temperature behavior needs mapping before it can serve as thermometer | Semiconducting nanowires4 |
| Insulating fibers & films | Coated TET (this article) | Nanometer Au/Ir film / film resistance | Film’s parallel channel via Wiedemann–Franz, with measured effective Lorenz number | Protein and cellulosic fibers67, glasses, polymers |
| Insulating and coating-intolerant | Optical heating with non-contact thermometry, or a separately fabricated sensor platform | Raman/thermoreflectance where optically applicable; otherwise an external microfabricated thermometer | No coating subtraction — but requires suitable optical response or a separate sensor structure; no universal route | Deposition-sensitive polymers, delicate biological specimens |
| Conductive but low-TCR or current-sensitive | Optical heating + Johnson-noise or other electrical thermometry | Laser / thermal-noise spectrum | None from coating; noise thermometry trades the TCR requirement for exacting electronics8 | Specimens where minimal current matters most |
What the protocol has unlocked
The coating protocol’s justification is the measurement record it opened. The family’s reach into biological and natural insulators is the showcase: spider silk, coated and suspended, revealed exceptionally high thermal conductivity for a protein fiber and a remarkable, reversible response to stretching6; lignocellulose has been characterized down to the single-cell scale across a wide temperature range, mapping thermophysical behavior in a material class that underpins everything from structural timber to paper7. Neither specimen conducts a microampere on its own; both became measurable through nanometers of metal and the subtraction this article describes.
The lineage is older than the modern refinements. The earliest optical-heating member of the family was explicitly built for “conductive and nonconductive” wires alike9 — nonconductive coverage was a founding ambition, not an afterthought — and the platform-level integration that industrialized these methods carries the coating workflow as a first-class citizen rather than a workaround10. What has changed over the years is not the idea but the honesty of the subtraction: from assuming a bulk Lorenz number toward measuring the film’s own effective value, the refinement the next section takes up.
The Lorenz number honesty clause
The Sommerfeld value of the Lorenz number is a bulk-metal result, and nanometer films are not bulk metals. At film thicknesses comparable to the electron mean free path, boundary and grain-boundary scattering reshape charge and heat transport differently, and the direct measurement campaign on iridium films from 7 nm down to 0.6 nm found both conductivities far below bulk with an effective Lorenz number substantially above the textbook constant11; companion work on sub-7-nm Ir films on silk substrates mapped where near-bulk Wiedemann–Franz behavior does and does not survive12. Nor is thin-film geometry the only offender: in graphene near charge neutrality, Johnson-noise thermometry has caught the Wiedemann–Franz ratio blowing through the textbook value by an order of magnitude as the electron system enters a hydrodynamic regime13 — a Science-grade reminder that the “law” is a well-behaved metal’s habit, not a constant of nature. Import the bulk value into the subtraction and the error lands, amplified, on the substrate: the thinner the film and the less conductive the substrate, the more the final answer hinges on this single constant.
The defensible practice is therefore to measure or bracket the film’s effective Lorenz number rather than assume it — and the family’s own differential machinery provides the tool. The coating-differential analysis introduced in the previous article, run across systematically varied coating resistance, returns the effective Lorenz number of the deposited film as a fitted parameter — closing the loop so the subtraction rests on the film actually present, not on a constant from a different century’s metallurgy4.
Protocol discipline
Five disciplines keep coated measurements defensible. Continuity before physics: below a material-dependent thickness a sputtered film breaks into islands whose resistance is dominated by percolation, not transport — verify ohmic, stable resistance before trusting any thermal reading. Know whose resistance you measured: for a highly insulating substrate the measured resistance is the film’s; for semiconductive or partially conductive substrates the electrical channels must be separated before any Wiedemann–Franz conversion is attempted. Characterize the film as deposited: resistance at measurement temperature, thickness by the deposition record or direct metrology, and an effective Lorenz number by the differential route where the accuracy target warrants it11. Respect the coupling assumption: the analysis treats film and substrate as thermally intimate; delamination or poor conformity violates it in ways that surface as transient anomalies — inspect residuals3. Budget the subtraction: report the film’s computed share alongside the final value; a substrate number extracted as a small difference between two comparable quantities carries amplified uncertainty, and saying so is what separates characterization from arithmetic2. The uncertainty framework makes the arithmetic explicit: the variances of the composite measurement, the film-resistance measurement and the Lorenz-number input combine into the substrate value’s budget, and when the film carries a large share of the total, the combined relative uncertainty on the small remainder inflates accordingly14 — which is the quantitative reason the discipline insists on thin-but-continuous films and measured, not imported, film properties.
With the protocol in place, the family’s reach broadens across dielectric, semiconductive and metallic samples in the 1D and 2D formats4 — wherever coating feasibility and the composite-model assumptions are validated for the specimen at hand. Our fiber and film testing service applies the coating workflow to insulating specimens where sample feasibility review supports it, and the 2D material service extends it to the supported atomic-thickness measurements the next article takes up. The transient model and reduction routes behind everything here live in the TET complete guide.
Frequently asked questions
Keep Exploring the ACS Thermal Metrology Knowledge Hub
This article is one chapter of the ACS thermal metrology knowledge hub. To keep going:
- Thermal conductivity & diffusivity testing: the pillar guide — methods, samples and a buyer’s framework in one place.
- Differential TET — correcting radiation, residual-gas transfer and coatings.
- Differential TET for 2D materials — supported-sheet transport down to atomic thickness.
- Fiber & film testing service — wires, fibers and films measured on this platform.
- 2D material testing service — supported-composite characterization for atomic-thickness samples.
References
This article describes the metal-coating measurement protocol for electrically non-conductive samples in general, idealized terms for education. Real workflows involve deposition-specific film characterization, sample-dependent validation and full uncertainty budgets; consult the published protocols and instrument documentation for specifics. The interactive tool above is a schematic teaching aid, not an instrument.