An infrared thermometer assumes it knows how efficiently a surface radiates heat, and that assumption is a single number the user sets: the emissivity. Get it right and the reading is accurate; get it wrong - especially on shiny bare metal - and the error can run to hundreds of degrees. Emissivity is quietly the most error-prone setting in all of non-contact thermometry, and it is the parameter an operator troubleshoots first when an infrared reading disagrees with a thermocouple. This guide defines emissivity, explains why shiny surfaces wreck IR readings, and shows how to correct a spot reading against a SCADA-trended contact sensor.
Emissivity in one line: Emissivity is a number between 0 and 1 that describes how efficiently a surface emits thermal radiation compared with a perfect radiator, or blackbody, at the same temperature. Because an infrared thermometer infers temperature from emitted radiation, it must be told the target's emissivity to interpret that radiation correctly. A wrong emissivity setting - particularly too high a value on a low-emissivity shiny metal - throws the reading off, sometimes by hundreds of degrees, which makes it the single most common source of infrared measurement error.
A perfect radiator, called a blackbody, emits the maximum thermal radiation physically possible for its temperature and is assigned an emissivity of 1. Real surfaces always emit somewhat less than that ideal, and emissivity is the ratio that captures how much less - a value from 0 to 1 giving the fraction of blackbody radiation the surface actually emits. A matte black painted surface may emit close to 0.95, behaving almost like a blackbody, while a polished metal may emit only 0.05, radiating a small fraction of what a blackbody would at the same temperature.
This matters because an infrared thermometer measures the radiation arriving at its detector and works backward to a temperature, and that calculation depends entirely on how efficiently the surface was supposed to be radiating. The instrument therefore has an emissivity setting the user must match to the target. If the true emissivity is 0.9 and the setting is 0.9, the thermometer interprets the received radiation correctly. But if the setting does not match the surface, the instrument applies the wrong conversion and reports a temperature that can be far from the truth.
The direction and size of the error follow logically. If the setting is higher than the surface's real emissivity, the instrument assumes the surface should be radiating a lot, sees relatively little, and concludes the surface must be cooler than it is - reading low. The mismatch matters most where emissivity is low, because on a poor emitter a small error in the assumed value corresponds to a large error in inferred temperature. This is why emissivity is described as the most error-prone setting: the physics amplifies any mismatch precisely where surfaces radiate weakly.
Bare, shiny, polished metals are the classic infrared nightmare, and emissivity is why. A polished metal has very low emissivity - it radiates only a small fraction of a blackbody's output - so the instrument receives little thermal radiation to work with and any error in the emissivity setting translates into a huge temperature error. Set a typical default emissivity on shiny stainless or aluminum and the reading can be off by hundreds of degrees, which is the failure that sends operators chasing a phantom problem or, worse, missing a real one.
Low emissivity comes paired with a second curse: high reflectivity. A surface that emits little radiation reflects a lot, so a shiny metal target also mirrors the thermal radiation of its surroundings back into the sensor. The instrument cannot tell the target's own emission from reflected radiation off a nearby hot furnace or the operator's own body heat, so both errors stack - the weak true signal is swamped by reflected surroundings. This is why a shiny surface can read wildly inconsistent depending on what is nearby and where the operator stands.
The practical fixes all address the same root cause. The most reliable is to raise the surface's emissivity locally and predictably - a patch of high-emissivity flat black paint or a piece of matte tape brings the emissivity up near 0.95 and stabilizes it, so a known setting reads true. Where that is not possible, one measures the actual emissivity by comparing the IR reading to a trusted contact measurement and adjusting the setting until they agree, then uses that value thereafter. Some instruments offer wavelength bands less sensitive to emissivity for metals, but for most field work, treating the surface or calibrating the setting is what turns a useless shiny-metal reading into a trustworthy one.
When a handheld infrared reading disagrees with a contact sensor, emissivity is the first thing to suspect, and a SCADA historian is what makes the comparison rigorous. A contact sensor such as a thermocouple or RTD that already feeds a trended tag gives a trustworthy reference temperature for the same or a nearby surface. The procedure is straightforward: aim the infrared thermometer at a spot near the contact sensor, then adjust the emissivity setting until the infrared reading matches the trended contact value. The setting that produces agreement is the surface's effective emissivity, and it can be reused for that material thereafter.
A cloud SCADA such as Merobix makes this reference readily available because the contact sensor's value is historized and stable, so a technician calibrating an IR gun is comparing against a known-good number rather than a momentary guess. Trending both over time also exposes emissivity problems that a single spot check would miss. If a fixed infrared tag consistently reads a fixed amount below a nearby contact tag, that steady offset is the fingerprint of an emissivity mismatch rather than a real temperature difference, and the offset points directly at the correction needed.
This cross-check is especially valuable because emissivity is not always constant. A surface that oxidizes, weathers, gets painted, or accumulates a coating changes its emissivity over time, so a setting that was correct at commissioning can drift out of validity. Watching a fixed infrared tag against a contact reference in the historian catches that drift as a slowly growing disagreement, prompting a re-calibration of the emissivity setting before the infrared reading silently loses its accuracy. In effect, the trended contact sensor becomes the anchor that keeps the non-contact reading honest.
It depends on the surface. Matte, non-metallic, and painted surfaces are typically near 0.9 to 0.95 and are forgiving. Shiny, bare, or polished metals have very low emissivity, often below 0.1, and need either a special low-emissivity setting or, better, a patch of matte black paint or tape to raise the emissivity to a known value. When in doubt, calibrate the setting against a trusted contact measurement.
Shiny metal has very low emissivity, so it radiates little of its own heat and any error in the emissivity setting produces a large temperature error, often hundreds of degrees. Low emissivity also means high reflectivity, so the surface mirrors the thermal radiation of nearby hot objects into the sensor and corrupts the reading further. Covering the spot with matte tape or flat black paint fixes both problems.
Treat the thermocouple as the reference and adjust the infrared thermometer's emissivity setting until its reading matches the contact measurement on the same or a nearby spot. The emissivity value that brings them into agreement is the surface's effective emissivity and can be reused for that material. If a fixed IR tag shows a steady offset from a trended contact sensor, that constant difference usually signals an emissivity mismatch rather than a real temperature gap.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.