An infrared thermometer infers the temperature of a surface from the thermal radiation it emits, with no physical contact. A lens focuses infrared energy onto a detector, commonly a thermopile or pyroelectric element, and the instrument converts the radiant power into a temperature reading after compensating for emissivity and ambient conditions. Closely related fixed-mount industrial instruments aimed at higher temperatures are called radiation pyrometers, and both rest on the same blackbody radiation physics.
Non-contact measurement is the reason these instruments exist: they read moving, hot, electrically live, hazardous, or hard-to-reach targets in milliseconds, where a contact probe would be impossible, unsafe, or too slow. The trade-off is that accuracy now depends on the operator understanding emissivity, spectral band, and spot size, which is what this guide covers.
Photo: Tony Webster, CC BY 2.0, via Wikimedia Commons
This guide is written for industrial purchasing engineers and design engineers. It spans 6 chapters from radiation physics and history, through sensor types and spectral bands, emissivity and target materials, spec-sheet decoding, to a selection decision sequence, with 7 selection FAQs and manufacturer references. Parameters reference Planck's law and the Stefan-Boltzmann law of blackbody radiation, the ASTM E20.02 radiation thermometry framework, and the clinical standards ASTM E1965 and ISO 80601-2-56.
Chapter 1 / 06
What is an Infrared Thermometer
An infrared thermometer is a non-contact temperature instrument that measures the infrared radiation emitted by a surface and infers its temperature. Every object above absolute zero (-273.15 degrees Celsius) radiates electromagnetic energy, and the amount and spectral distribution of that energy depend on the object's temperature. The instrument collects a portion of this radiation through optics, focuses it onto a detector, and converts the resulting electrical signal into a temperature value. Because nothing touches the target, the device can read objects that are moving, rotating, energized, corrosive, sterile, or simply too hot for a contact probe.
Structurally, an infrared thermometer has four functional parts: (1) the optical system, a lens or mirror that gathers radiation and defines the field of view and the measured spot; (2) a spectral filter that limits the detector to a chosen waveband; (3) the detector, typically a thermopile or pyroelectric element in handheld units, which turns radiant power into a small electrical signal; and (4) signal-processing electronics that apply the emissivity setting, ambient-temperature compensation, and linearization, then display or transmit the result. Aiming aids such as a single or dual laser, or a through-lens video sighting system on fixed pyrometers, help the user place the spot, but the laser dot is only a pointer and does not itself measure anything.
The physics dates to 1800, when William Herschel, measuring the temperature of sunlight split by a prism, found a heating effect beyond the red end of the visible spectrum and so discovered infrared radiation. The quantitative laws followed: the Stefan-Boltzmann law (radiated power rises with the fourth power of absolute temperature), Wien's displacement law of 1893 (the peak wavelength shifts shorter as temperature rises), and Planck's law of 1900, which gives the full spectral distribution and introduced the quantum constant. Macedonio Melloni built the first practical thermopile detector in the 1830s, and Henri Le Chatelier developed the thermoelectric pyrometer in 1886. The disappearing-filament optical pyrometer, built by Holborn and Kurlbaum in 1901, let operators match a furnace glow to a calibrated filament and dominated high-temperature work for decades before electronic detectors replaced it.
In application scale, infrared thermometry spans an enormous temperature range. Cryogenic and ambient work reaches down to roughly -50 degrees Celsius for HVAC, cold-chain, and building diagnostics; general industrial and clinical use covers room temperature to a few hundred degrees; and short-wave industrial pyrometers read molten steel, glass furnaces, and forging operations up to 3000 degrees Celsius or beyond. No single instrument covers this whole span, because the optimal detector, spectral band, and optics differ at each level. The discipline of selection, the subject of this guide, is matching a specific target and temperature to the right physical configuration.
Four engineering factors decide whether a field reading is trustworthy: the emissivity of the target surface, the spectral band of the instrument, the ratio of spot size to target size, and the level of reflected and absorbed background radiation. Unlike contact thermometry, where the main concern is sensor accuracy and response time, infrared measurement makes the operator responsible for these four conditions. A laboratory-grade instrument pointed carelessly at shiny metal can read hundreds of degrees low, while a modest handheld used correctly on a matte surface can be accurate to about one degree. Information density in the next five chapters is aimed squarely at closing that gap.
Chapter 2 / 06
Types and Classification
Infrared temperature instruments classify along three axes: form factor (handheld versus fixed-mount), measurement principle (single-band versus two-color ratio versus thermal imaging), and target class (general-purpose, metals, glass, plastic film, or clinical). These overlap in practice, but the five categories in the table below cover the great majority of industrial and commercial selections. Choosing the wrong category, for example using a general handheld on molten metal, is the most common and most expensive beginner mistake.
General handheld infrared thermometers are point-and-shoot guns with a long-wave 8 to 14 micron detector, an adjustable emissivity setting, and a laser pointer. They suit non-metallic and matte surfaces from cold-chain to a few hundred degrees, and are the right tool for HVAC duct and register checks, motor and bearing temperatures, electrical-panel hot spots, and food safety. The Fluke 62 MAX, for example, covers -30 to +500 degrees Celsius with a 10:1 optics. They are not designed for shiny metal or for spots smaller than the optics resolve.
Fixed single-band pyrometers are mounted permanently to view a process point and feed a controller, often through a 4-20 mA or digital output. They are configured at order time for a specific waveband and range, so a glass-line pyrometer at 5.0 microns, a metals pyrometer at 1.6 microns, and a plastic-film pyrometer at 3.43 microns are different models even from one maker. Their value is repeatable, automated control of a continuous process such as extrusion, annealing, or float glass.
Two-color, or ratio, pyrometers measure two close wavelengths and use the ratio of intensities, which cancels emissivity error and tolerates dust, steam, partial spot fill, and contaminated sight glasses. They are the standard for molten and forged metal, induction heating, and any scene where emissivity changes or the target is smaller than the spot. Optris CTratio and Fluke Endurance two-color models cover roughly 250 to 3500 degrees Celsius. Their limitation is the grey-body assumption, discussed in Chapter 3.
Thermal imaging cameras use a focal-plane detector array to produce a temperature image rather than a single spot, trading per-point accuracy for spatial coverage. They excel at surveys and predictive maintenance, finding the hottest connection in a switchgear cabinet or the failing bearing in a row of motors. SpecForge treats them as a separate category; see the related thermal imaging camera guide. Clinical and screening units are narrow-range forehead and ear thermometers governed by ASTM E1965 and ISO 80601-2-56, with tight laboratory accuracy over the body-temperature span and built-in skin-to-core compensation. They must never be used for industrial work, and industrial guns must never be used on people.
Chapter 3 / 06
Detector Technologies and Spectral Bands
Two design choices define an infrared thermometer's behavior more than any other: the detector type and the spectral band the optics and filter admit. The detector converts radiation to signal; the band decides which materials the instrument can read correctly. The table below summarizes the spectral bands used in practice and the target class each one serves.
Spectral band
Best-fit target
Typical range
Reason for the band
8 to 14 µm
Non-metals, general surfaces
-50 to +1000 °C
Broad long-wave window, stable high emissivity
5.0 to 5.2 µm
Glass surfaces
100 to 1900 °C
Glass is opaque, so surface is read
3.43 µm
Thin PE / PP film
50 to 300 °C
Absorption band of polyolefins
7.9 µm
Thin polyester film
50 to 300 °C
Film opaque where PE is transparent
2.3 µm
Metals, lower range
250 to 1800 °C
Higher, steadier metal emissivity
1.0 to 1.6 µm
Metals, high temperature
450 to 3000 °C
Short wave minimizes emissivity error
Thermopile detectors are the workhorse of handheld and many fixed instruments. A thermopile is a series of thermocouples whose hot junctions sit on a tiny absorbing membrane and whose cold junctions sit on the chip frame; incoming radiation warms the membrane and the array produces a DC voltage proportional to the temperature difference. Thermopiles respond to steady radiation, need no chopper, and are inexpensive and robust, which is why they dominate point-and-shoot guns. Their response time is modest, on the order of tens to hundreds of milliseconds, and signal level is small, so good electronics matter.
Pyroelectric detectors respond to changes in radiation rather than steady level, so they require an optical chopper or a moving target, but they are fast and sensitive, suiting motion-line and modulated applications. Photonic detectors such as InGaAs and silicon photodiodes are used in short-wave 1.0 to 2.3 micron pyrometers for metals; they are far faster, with response times down to about 1 millisecond, and give the high speed needed for forging, rolling, and induction. Microbolometer arrays are the imaging equivalent, used in thermal cameras rather than spot instruments.
The spectral band must match the target. A general 8 to 14 micron unit reads non-metals well because most non-metals have high, stable emissivity across the long-wave band. The same unit fails on bright metal, where short-wave 1.0, 1.6, or 2.3 micron optics give higher and steadier emissivity and far less error. Glass is transparent in the long wave but opaque near 5.0 microns, so glass pyrometers use a 5.0 to 5.2 micron filter to read the true surface rather than seeing through it. Thin plastic films are the reverse: polyethylene and polypropylene absorb at 3.43 microns, while polyester and many other films are best read at 7.9 microns, where they are opaque. Picking the wrong band makes the instrument measure a background or an interior layer instead of the surface you care about.
Two-color (ratio) measurement deserves a closer note because it changes the rules. Instead of one band, the instrument reads two closely spaced wavelengths and computes temperature from the ratio of intensities. If emissivity, a dirty window, steam, or a partly filled spot lowers both signals in the same proportion, the ratio is unchanged and the reading holds. That robustness is why ratio pyrometers own the metals and induction market. The catch is the grey-body assumption: the method is exact only if emissivity is equal at both wavelengths. For non-grey materials whose emissivity differs between the two bands, the ratio introduces its own error, so two-color is a powerful but not universal tool.
Chapter 4 / 06
Emissivity, Materials and Target Conditions
Emissivity is the single largest source of error in non-contact thermometry. It is the ratio of the radiation a real surface emits to that of an ideal blackbody at the same temperature, on a scale from 0 to 1. The instrument measures radiant power and divides by the emissivity setting to recover temperature, so an incorrect setting translates directly into a wrong reading. Most handheld units default to 0.95 because that value fits a wide range of non-metallic, organic, oxidized, and painted surfaces, but the same default is badly wrong for bright metals.
Non-metals cluster near 0.90 to 0.95. Concrete is about 0.92, common brick about 0.85, water and ice near 0.95, and human skin about 0.98, which is why clinical and screening units assume a high fixed emissivity. Paints, rubber, paper, wood, food, and most plastics also fall in this high band, so a default of 0.95 gives readings within a degree or two without adjustment. These are exactly the surfaces a long-wave 8 to 14 micron handheld is designed for.
Bright metals are the trap. Polished aluminum sits near 0.05 and polished copper near 0.04, meaning they emit only a few percent of blackbody radiation and reflect the rest. A thermometer set to 0.95 on polished aluminum can read hundreds of degrees too low, and it also picks up reflected heat from nearby ovens, lamps, or even the operator. Oxidation and roughness raise emissivity sharply: rough or oxidized cast iron reaches about 0.80 to 0.85, and heavily oxidized steel approaches 0.90, so a weathered surface is far easier to read than a machined one. The table below collects representative values for first-pass selection.
Material / surface
Typical emissivity
Reading strategy
Aluminum, polished
~0.05
Short-wave or tape / paint the spot
Copper, polished
~0.04
Short-wave or tape the spot
Cast iron, rough / oxidized
~0.80 to 0.85
Long-wave, default near 0.85
Steel, heavily oxidized
~0.85 to 0.90
Long-wave, default near 0.90
Concrete
~0.92
Long-wave, default 0.95
Brick, common
~0.85
Long-wave, default 0.95
Water / ice
~0.95
Long-wave, default 0.95
Human skin
~0.98
Clinical unit, fixed high value
Three target conditions matter as much as the material itself. Reflected ambient radiation from hot equipment or sunlight adds to the signal and biases readings high on low-emissivity surfaces; shade the target or change the viewing angle to avoid the reflection. Atmospheric path matters over distance or through flames: water vapor and carbon dioxide absorb specific bands, which is why the 8 to 14 micron window is preferred for long paths and why flame pyrometry uses dedicated bands. Surface preparation is the simplest fix of all: a patch of matte tape (with a known emissivity, often near 0.95) or flat paint converts an unreadable shiny metal into a reliable target, and is standard practice for spot checks where short-wave optics are not available.
Chapter 5 / 06
Key Specification Parameters
Spec sheets for infrared thermometers list a dozen or more parameters, but a handful drive every selection: temperature range, spectral band, optical resolution (distance-to-spot ratio), accuracy, response time, emissivity adjustability, and output. Reading them correctly is what separates a reliable purchase from a returned instrument. Each is decoded below.
Distance-to-spot ratio, written for example as 10:1, 30:1, or 60:1, defines how the measured spot grows with distance. A 12:1 unit reads a roughly 25 mm (1 inch) spot from 300 mm (12 inches); at 600 mm the spot is 50 mm. The target must completely cover the spot, and a margin of at least 2:1 (target twice the spot) is good practice, otherwise the reading averages the target with cooler surroundings and falls low. Higher ratios let you measure small or distant targets, while precision-optics pyrometers resolve spots down to about 1 mm. Critically, the spot is smallest only at the optical focus distance; nearer or farther, it grows, so a quoted ratio is meaningful only with its focus distance.
Accuracy is usually stated as plus-or-minus a percent of reading or a fixed temperature, whichever is greater. A representative industrial handheld specifies about plus-or-minus 1 percent of reading or plus-or-minus 1.0 to 1.5 degrees Celsius near room temperature, widening at the cold and hot extremes of its range. Fixed process and laboratory pyrometers reach plus-or-minus 0.3 to 0.5 percent of reading. Remember that this figure is the instrument's contribution under ideal conditions; in the field, emissivity error and reflected radiation usually dominate the total uncertainty.
Response time is the time to reach a specified percentage, commonly 95 percent, of a step change. Thermopile handhelds are typically under about 500 milliseconds, fast enough for manual point-and-shoot use. Photonic short-wave pyrometers reach about 1 millisecond, which is necessary to catch a fast-moving billet or a single induction cycle. For control loops on moving product, response time directly limits how fast the process can run while staying in spec.
Other parameters that decide selection:
Temperature range: Must bracket the process with margin. General handhelds span roughly -50 to +1000 degrees Celsius; short-wave pyrometers reach 3000 degrees Celsius or more. No single band covers the whole span.
Spectral band: 8 to 14 microns for non-metals, 1.0 to 2.3 microns for metals, 5.0 microns for glass, 3.43 and 7.9 microns for plastic films, as set out in Chapter 3.
Emissivity adjustment: Quality handhelds adjust from 0.10 to 1.00. Fixed-emissivity units are limited to one material class. Two-color models need no emissivity entry.
Display resolution: Commonly 0.1 degrees Celsius. Resolution is not accuracy; a fine display does not improve a coarse measurement.
Output and interface: Handhelds give a local display, MIN/MAX/AVG, and sometimes Bluetooth. Fixed pyrometers output 4-20 mA, thermocouple-equivalent mV, or digital (RS-485/USB/Ethernet) for controller integration.
Environmental rating: Housing IP54 to IP65 for handhelds; fixed heads add water cooling and air purge, and fiber-optic heads tolerate ambient up to about 315 degrees Celsius without cooling.
Two figures are routinely confused. Optical resolution (distance-to-spot) is a geometric property of the lens, while display resolution is just the smallest digit shown. A 60:1 instrument with a coarse 1-degree display can be far more useful on a small distant target than a 10:1 instrument with a 0.1-degree display, because the first actually places its energy on the target. Always read optical resolution, accuracy, and the focus distance together.
Chapter 6 / 06
Selection Decision Factors
To turn the preceding five chapters into a specific model, follow the decision sequence below. Most selection errors come not from one wrong answer but from deciding a later step before an earlier one is settled, for example fixing on a brand before confirming the spectral band the target requires. These eight steps double as an RFQ template.
Target material and surface: First identify what you are measuring and its finish. Non-metal or matte? A long-wave 8 to 14 micron handheld fits. Bright metal? You need short-wave optics, a two-color pyrometer, or a tape/paint workaround. Glass or thin plastic film? You need a dedicated narrow band. This single decision drives the spectral band and rules out most of the catalog.
Temperature range: Bracket the process minimum and maximum with margin, then confirm a single instrument can cover it. Cold-chain and HVAC need the sub-zero end; metals and kilns need short-wave high-temperature optics. Do not assume one model spans both.
Spot size and distance: Measure the smallest target dimension and the working distance, then require optics whose spot at that distance is at most half the target. Convert the distance-to-spot ratio at the actual focus distance, not the catalog headline number.
Accuracy and repeatability needed: Distinguish a rough maintenance check (plus-or-minus a few degrees is fine), a process control point (plus-or-minus 0.5 to 1 percent), and a reference or custody application (laboratory-grade fixed pyrometer). Higher accuracy raises price and tightens the installation discipline required.
Form factor and integration: Handheld for spot checks and surveys; fixed sensor head with 4-20 mA or digital output for continuous control; fiber-optic head where ambient heat or access forbids mounting the electronics at the target. Decide where the electronics can physically live.
Environment and protection: Ambient temperature at the mounting point (water cooling or air purge above limits), dust and steam (favor two-color or a purged window), vibration, washdown (IP65 or better), and any hazardous-area certification the site requires.
Standards and conformity: Industrial radiation thermometry follows the ASTM E20.02 framework and supplier calibration traceable to a national lab. Clinical and fever-screening units must conform to ASTM E1965 or ISO 80601-2-56; an industrial gun is not a substitute and must never be used on people.
Total cost of ownership: Purchase price plus mounting, cooling, calibration, and the cost of a wrong reading. A cheap handheld misused on shiny metal that under-reports by 100 degrees can scrap a heat-treat batch worth far more than the instrument; the right tool plus correct emissivity is the real economy.
One frequently overlooked dimension is serviceability and calibration support: local calibration capability, spare optics and windows, firmware updates for fixed pyrometers, and documented traceability. For industrial fixed and handheld instruments, Fluke and Fluke Process Instruments (62 MAX and 64 MAX handhelds, Endurance and Thermalert pyrometers), Optris (CT, CTlaser, CSmicro, CTratio), LumaSense/Advanced Energy Impac, Williamson IR for two-color pyrometry, and Land/AMETEK for combustion and metals are established suppliers. For handheld maintenance, food, and HVAC use, Fluke, Flir/Teledyne, Testo, and ThermoWorks are common choices. Confirm the exact band, range, and optical resolution against your own target before committing to any series, because one brand sells many incompatible variants.
FAQ
What is the difference between an infrared thermometer and a pyrometer?
The terms overlap heavily. Historically a pyrometer meant any non-contact instrument for high-temperature measurement, originally the optical disappearing-filament type used to judge furnace glow. In modern usage, handheld point-and-shoot units sold for HVAC, food, and maintenance work are called infrared thermometers, while fixed-mount industrial sensors aimed at metals, glass, and kilns above roughly 600 degrees Celsius are called pyrometers. Both sense the same thermal radiation and obey the same physics. The practical distinction is form factor and temperature range: infrared thermometers typically cover -50 to +1000 degrees Celsius with an 8 to 14 micron detector, while pyrometers reach 3000 degrees Celsius or higher using short-wave 1 to 2.3 micron optics.
What is emissivity and why does it matter so much?
Emissivity is the ratio of the radiation a real surface emits to that of a perfect blackbody at the same temperature, on a scale of 0 to 1. An infrared thermometer measures radiant power and back-calculates temperature, so it must assume an emissivity value. Most units default to 0.95, which fits non-metals such as concrete, paint, rubber, food, water, and human skin. The problem is bright metals: polished aluminum sits near 0.05 and polished copper near 0.04, so a thermometer set to 0.95 will read far too low, sometimes by hundreds of degrees. Fixes are to enter the correct emissivity (handheld units adjust from 0.10 to 1.00), apply matte tape or flat paint to the target, or use a two-color pyrometer that cancels emissivity error.
What does the distance-to-spot ratio mean for measurement?
The distance-to-spot ratio, written for example as 12:1 or 50:1, states how the measured spot diameter grows with distance. A 12:1 unit measures a roughly 25 mm (1 inch) spot from 300 mm (12 inches) away; double the distance and the spot doubles to 50 mm. The target must completely fill the spot, ideally being at least twice the spot diameter, or the reading averages the target with cooler background and reads low. To measure small or distant objects, choose a higher ratio such as 50:1 or 60:1, or a precision-optics pyrometer that resolves a 1 mm spot. The spot at the optical focus distance is smallest; never assume the visible laser dot equals the measured spot.
Why can't a standard infrared thermometer measure shiny metal correctly?
Shiny metals have very low emissivity, often 0.02 to 0.10, which means they emit little thermal radiation and instead reflect surrounding infrared. A long-wave 8 to 14 micron thermometer set to 0.95 then under-reports badly and also picks up reflected heat from nearby ovens or lamps. Three engineering solutions exist: enter the metal's true emissivity from a reference chart, although polished-metal values are uncertain and finish-dependent; coat the spot with high-emissivity matte tape or paint and read that; or use a short-wave pyrometer at 1.0, 1.6, or 2.3 microns, where metal emissivity is higher and more stable, or a two-color ratio pyrometer that divides out emissivity entirely. Short-wave optics are the standard approach for molten metal, forging, and heat-treat lines.
Which spectral band should I choose for my material?
The detector waveband must match where the target emits and transmits. General non-metallic surfaces below 1000 degrees Celsius use the broad long-wave 8 to 14 micron band, the default for handheld units. Metals and ceramics use short-wave bands, 1.0 micron and 1.6 micron for the highest temperatures and 2.3 micron for lower metal ranges, because short wavelengths reduce emissivity error. Thin plastic films use specific narrow bands, 3.43 microns for polyethylene and polypropylene and 7.9 microns for polyester and other thin films, since plastics are transparent at other wavelengths. Glass surface temperature uses 5.0 to 5.2 microns, where glass is opaque, and flame gases use a 4.5 micron CO2 band. Choosing the wrong band makes the instrument see through the target or read a background instead of the surface.
How accurate are infrared thermometers and what limits accuracy?
A quality industrial handheld unit specifies accuracy of about plus-or-minus 1 percent of reading or plus-or-minus 1.0 to 1.5 degrees Celsius, whichever is greater, near room temperature, widening at the extremes of range. Laboratory and fixed process pyrometers reach plus-or-minus 0.3 to 0.5 percent of reading. In the field, however, instrument accuracy is rarely the limiting factor: emissivity error, reflected ambient radiation, an under-filled spot, a dirty or fogged lens, and the sensor not being at thermal equilibrium with ambient typically dominate. Clinical forehead thermometers are governed by ASTM E1965 and ISO 80601-2-56, which require laboratory accuracy within plus-or-minus 0.2 to 0.3 degrees Celsius over the body-temperature span, a far tighter target than industrial units need.
What is a two-color or ratio pyrometer and when do I need one?
A two-color, or ratio, pyrometer measures radiation at two closely spaced wavelengths and computes temperature from the ratio of the two intensities rather than the absolute level. If emissivity, dust, steam, a dirty window, or a partially filled spot reduces both signals by the same proportion, the ratio is unchanged, so the reading stays correct. This makes ratio pyrometers the right choice for molten and forged metal with changing emissivity, targets smaller than the spot, induction heating with smoke and scale, and viewing through contaminated sight glasses. The limitation is the grey-body assumption: if emissivity differs between the two wavelengths, as it can for some non-grey materials, the ratio method introduces its own error. Optris CTratio and Fluke Endurance two-color models cover roughly 250 to 3500 degrees Celsius.
