5.4 Emissivity Tables: Metals and Non-Metals
Key Takeaways
- Published emissivity tables are starting estimates, not certified values for your exact surface, angle, wavelength band, and oxidation state
- Polished metals commonly fall near ε ≈ 0.05–0.15 in thermal IR and require reference or measurement methods for quantitative work
- Oxidized or rough metals often rise to roughly ε ≈ 0.30–0.80 depending on alloy and scale — still verify
- Paints, many plastics, organic materials, and water typically fall near ε ≈ 0.85–0.98 in LWIR; water is about 0.98 in the 8–14 µm band
- Never trust a single table entry blindly: surface condition, spectral band, temperature, and angle can move ε enough to change reported temperature and priority
Every Level II course includes emissivity tables. Used wisely, they speed setup on cooperative surfaces. Used blindly, they produce confident wrong numbers — especially on metals. This section gives exam-ready ranges, explains why tables disagree, and builds the habit of verification from Sections 5.2–4.3.
How to Read an Emissivity Table
A table entry is typically:
- A material name (often ambiguous: “aluminum” could mean polished, sandblasted, or anodized)
- A single number or range
- Sometimes a temperature and wavelength of measurement
- Rarely your exact camera band, viewing angle, and plant surface history
Treat each entry as:
ε_table ≈ guess under stated conditions
not as:
ε_truth for my bus bar today.
| Table limitation | Field implication |
|---|---|
| Surface finish not matching | Polished vs oxidized can differ by 0.5 or more |
| Wavelength mismatch | MWIR vs LWIR metal values may differ |
| Angle assumed near normal | Grazing views lower effective ε |
| Temperature of determination | High-T oxidation changes ε |
| Sample purity / alloy | “Steel” is not one optical material |
| Specular vs diffuse measurement methods | Different lab techniques → different published numbers |
Metals: Polished vs Oxidized
Polished / bright metals (typical IR behavior)
Clean, polished, or freshly machined metals are poor emitters and strong reflectors in the thermal infrared.
Exam working range: ε ≈ 0.05–0.15 for many polished non-oxidized metals in LWIR (some even lower; some slightly higher — use the range as a warning flag, not a precise passport).
| Material condition (examples) | Approximate ε range (order of magnitude) | Level II stance |
|---|---|---|
| Polished aluminum | ~0.04–0.10 | Tape/paint/RAT; do not invent precise T from default ε |
| Polished copper | ~0.02–0.15 (condition-sensitive) | Same |
| Polished stainless / chrome | ~0.05–0.20 | Same |
| Bright galvanized (shiny) | Often low to moderate; highly variable | Verify |
| Mirror finishes | Very low | Reflection methods critical |
If the surface looks like a mirror in visible light, assume thermal IR trouble until proven otherwise. Visible shininess is not a quantitative IR meter, but it is a useful hazard flag.
Oxidized, rough, or weathered metals
Oxidation, scale, roughness, and dust generally raise emissivity because the optical surface becomes less metal-like and more dielectric/rough.
Exam working range: ε ≈ 0.30–0.80 for many oxidized metals — a wide band that should scare you away from one-number certainty.
| Condition | Qualitative ε change | Practice |
|---|---|---|
| Light straw oxidation | Moderate increase | Still prefer measurement |
| Heavy mill scale / rust | Often much higher than polished | Better for IR, still table-uncertain |
| Cast rough surfaces | Higher effective ε | Cavities raise effective ε |
| After wire brushing to bare metal | ε drops again | Re-characterize after maintenance |
| Painted over metal | Use paint ε, not metal ε | Coating dominates if opaque |
Why metal tables disagree so much
Two handbooks can list “oxidized steel” as 0.65 and 0.80 and both be “right” for different samples. Level II competence is saying:
“Tables suggest mid-to-high ε once oxidized, but I will measure or use a reference emitter for quantitative priority.”
Non-Metals: Paints, Plastics, Organics, Water
Most nonmetallic materials of interest in plant and building thermography are high-ε in LWIR.
Exam working range: ε ≈ 0.85–0.98 for many paints, plastics, rubber, wood, brick, human skin, and water-related surfaces (with exceptions for some thin films and shiny polymers).
| Material class | Typical LWIR ε (approx.) | Notes |
|---|---|---|
| Flat black paint | 0.90–0.97 | Workhorse reference |
| Other matte paints / powder coats | 0.85–0.95 | Color in visible ≠ IR ε, but matte organics usually high |
| Electrical tape (matte vinyl) | ~0.90–0.97 | Section 5.2 |
| Many plastics (thick, matte) | 0.85–0.95 | Shiny plastic may be lower; thin film may transmit |
| Rubber, wood, paper, cloth | ~0.90–0.95 | Usually cooperative |
| Concrete, brick, stone | ~0.90–0.95 | Good for building IR |
| Human skin | ~0.97–0.98 | Medical/occupational contexts |
| Water | ~0.95–0.98 | See below |
| Ice / snow | High but condition-dependent | Angle and structure matter |
| Glass (window) | High emission from glass surface in LWIR; does not “see through” like visible | You measure glass temperature, not indoor scene through glass in LWIR |
Water ≈ 0.98 in 8–14 µm
Liquid water is an excellent LWIR emitter/absorber. For exam and field memory:
- Water emissivity in the long-wave infrared (about 8–14 µm) is approximately 0.98 (often cited near 0.95–0.98 depending on angle and state).
- That is why wet surfaces and water films can appear radiometrically cooperative.
- Angle still matters: near-grazing views reduce effective emissivity for water as for other dielectrics (Fresnel behavior).
- A thin layer of water on metal may optically behave like water if the layer is optically thick in-band; extremely thin molecular films may not fully replace metal optics — field judgment and verification still apply.
| Context | Why ε_water ~ 0.98 matters |
|---|---|
| Roof moisture surveys | Wet insulation patterns are thermal and moisture-content problems; surface ε of wet membranes still often high |
| Leak detection | Water films change heat transfer and can change local IR appearance |
| Steam systems | Condensate and wet insulation behave differently from dry cladding |
| Exam recall | Non-metal group: water near blackbody in LWIR |
Practical Table for Level II Memory (Composite)
Use this as a study scaffold, not a substitute for measurement:
| Category | Approx. ε (LWIR) | Quantitative approach |
|---|---|---|
| Polished metals | 0.05–0.15 | Tape, paint, contact ε, careful RAT |
| Oxidized / rough metals | 0.30–0.80 | Prefer measure; tables only rough |
| Paints, plastics, organics | 0.85–0.98 | Table OK if matte/opaque; still set RAT |
| Water | ~0.98 (8–14 µm) | High-ε non-metal; watch angle |
| Unknown shiny anything | Treat as suspect | Reference emitter |
Never Trust a Single Table Blindly — Decision Rules
Rule 1 — Surface condition beats material name
“Aluminum” in a table is useless without finish. Ask: polished, brushed, anodized, painted, powder-coated, greasy, wet?
Rule 2 — Band and camera matter
A value measured at 2 µm may not equal effective ε in 8–14 µm. Prefer tables that state spectral conditions, or measure in situ with your imager band.
Rule 3 — Geometry matters
Threads, slots, cracks, and cavities raise effective emissivity via multiple reflections. A polished bolt head may measure more reliably at a cavity than on the flat chrome top.
Rule 4 — Temperature and aging matter
Heat-tinted stainless after a process upset is not the same ε as new polished stainless. Re-baseline after surface-changing maintenance.
Rule 5 — When in doubt, measure
Hierarchy of trust for Level II quantitative work:
- In-situ characterization (contact + correct RAT → ε)
- Reference emitter on the component (tape/paint)
- Site-specific library you built for recurring assets
- Table ranges with documented assumptions and higher uncertainty
- Guess from memory without documentation — not professional practice
Rule 6 — Consistency for ΔT
If you only need ΔT between two identical surfaces (same paint, same angle, same ε error), relative comparison can remain useful even if absolute T is slightly biased — provided the bias is common-mode. Mixing bare copper vs painted steel in one ΔT without correction is not common-mode.
Exam Scenarios Using Tables
Scenario 1: Question shows polished aluminum bus and offers ε = 0.95 from a “metal table.”
Answer path: Reject. Polished aluminum is ~0.05–0.15 class; use reference methods.
Scenario 2: Heavily oxidized steel tank, matte brown scale. Table says 0.70.
Answer path: Plausible order of magnitude (0.30–0.80 band), but best practice still verifies; better than polished, not gospel.
Scenario 3: Water on a floor for a training question: pick ε nearest 0.98 among options 0.05 / 0.3 / 0.5 / 0.98.
Scenario 4: Plastic safety cover, thick and matte.
Answer path: High-ε non-metal range; default 0.95 often acceptable with RAT set.
Scenario 5: Same asset last year ε = 0.65 oxidized; this year freshly polished.
Answer path: Old table/site value is obsolete; re-measure or tape.
Linking Tables to Camera Defaults
Many cameras default to ε = 0.95. That default is a non-metal / painted-world assumption. It is a reasonable starting point for building envelopes and painted gear, and a terrible silent assumption for bare bus and chrome. Level II inspectors change the default deliberately and document the value used.
Summary for Recall
Emissivity tables organize experience into numbers, but surface condition, band, angle, and temperature control the truth. Remember three clusters: polished metals ~0.05–0.15, oxidized metals ~0.30–0.80, paints/plastics/organics/water ~0.85–0.98 with water ~0.98 in 8–14 µm. Never apply a single table entry blindly — prefer reference emitters and contact back-calculation when quantitative decisions depend on the reading. Tables support judgment; they do not replace Level II measurement discipline.
Which approximate emissivity range is most appropriate for polished bare metals in thermal infrared Level II work?
Approximately what emissivity is commonly cited for water in the long-wave infrared band (~8–14 µm)?
Why should a Level II thermographer refuse to treat a single handbook emissivity for “oxidized steel” as exact for every plant asset?
A camera default of ε = 0.95 is most nearly appropriate as a starting assumption for which target class?