A BRIEF INTRODUCTION TO EMISSIVITY

I. GENERAL OVERVIEW:

All surfaces emit thermal radiation. However, at any given temperature and wavelength, there is a maximum amount of radiation that any surface can emit. If a surface emits this maximum amount of radiation, it is known as a blackbody. There are well known equations, such as Plancks Law that can be used to calculate the amount of radiation emitted as a function of wavelength and temperature.

Most surfaces are not blackbody emitters, and emit some fraction of the amount of thermal radiation that a blackbody would. This fraction is known as emissivity. If a surface emits ½ or 0.5 as much radiation at a given wavelength and temperature as a blackbody, it is said to have an emissivity of 0.5. If it emits 1/10 or 0.1 as much as a blackbody, it has an emissivity of 0.1 and so on. Obviously, a blackbody has an emissivity of 1.0 at all temperatures and wavelengths.

All radiation thermometers (RTs), commonly called pyrometers, contain one or more sensors that convert thermal radiation received through some type of optics into an electrical signal. Often, the RT contains one or more optical filters that allow only a selected wavelength (within a narrow band) to reach the sensor. These RTs are calibrated by aiming them at a blackbody and noting the sensor(s) electrical signal output as a function of blackbody temperature.

The most commonly used type of RT is known as a “brightness” or single wavelength device. There are also two wavelength RTs, know as ratio RTs and special multiwavelength devices. For the purposes of this discussion, I will concentrate on single wavelength RTs, mentioning ratio devices where appropriate. Multiwavelength RTs are highly specialized and are beyond the topic of this discussion.

As I mentioned, RTs are calibrated against a blackbody having an emissivity of 1.0, but are used to measure the temperatures of surfaces having an emissivity of less than 1.0. How does this work? There is a simple trick: If you take any number (other than zero) and divide it by itself, the answer is 1.0. There is an electronic divider circuit built into the RT and there is some sort of dial to enter an emissivity value. The divider circuit divides the signal received by this entered emissivity value. If you entered the right value, the two emissivities cancel to 1.0 and the signal value appears to be that of a blackbody.

However, too low of an entered value will cause the apparent temperature to be too high, and too high of a value entered will cause the apparent temperature to be too low. The error if the wrong emissivity value is entered gets worse at longer wavelengths.

Ratio units, which divide a short wavelength signal by a long wavelength signal, achieve a similar result as the two emissivities (short wavelength emissivity/long wavelength emissivity) divide out to 1.0 if and only if the two are the same, which happens for materials called “greybodies. However, if the two emissivities are not equal, which happens, the result error will be much larger than for many brightness RTs, so caution is advised. Certain types of smoke or water vapors can be tricky.

If it is known, when using a ratio unit, that the ratio of the two emissivities is not exactly one, there is also a divider circuit that has an input called E-slope or ratio correction or certain other names that will (as with brightness units) correct the ratio to 1.0.

Therefore: for an RT to correctly measure temperature, you must (!) know what emissivity value to enter. This is not always easy as many different things (some having nothing to do with emissivity per se) affect the emissivity value you must enter. And for god’s sake, do not just look up a value in a manufacturer’s handbook! The following is a partial list of things that affect emissivity to enter:

What is the actual emitting surface composed of? Often, even very thin coatings on the surface can affect the emissivity value of the surface and (depending on the coating thickness) the correct emissivity value may be that of the coating, not the main substrate. Very thin coatings (such as oxides or oils) can act as “interference filters” that cause the emissivity to vary widely depending on exact film thickness.
Regardless of what the emitting surface is composed of, the microscopic (and macroscopic) roughness of the surface causes differences in emissivity simply because a rougher surface has a larger emitting area. Generally, the emissivity of most opaque emitting surfaces increases as wavelength becomes shorter.
The geometric shape of a surface affects the required emissivity setting. Convex surfaces such as cylinders or balls have emissivity values that drop with increased curvature simply because they spread their radiation over a wider angle. Concave surfaces such as the inside of a bowl or box have higher than expected emissivities.
The angle from normal (straight on or 90 degrees) at which the surface is viewed, expecially for specular (mirror-like) or semispecular (dull reflection as from a smooth metal) surface will cause emissivity to start dropping rapidly at any angle beyond about 45 degrees as such surfaces do not emit equally at all angles. This is less of a problem with matt (nonreflective) surfaces except at long wavelengths. Also be cautioned that viewing a surface at more than about a 30 degree angle from normal may cause polarization effects, and some RTs are sensitive to such effects.
If an RT views a surface through a window, which is sometimes necessary, there will be both absorptive and reflective radiation losses depending on window thickness, window material and viewing angle. These losses will require using a lower emissivity setting than normal.
If the sight path between the RT and surface is partially obscured by smoke, steam, water droplets or particulate matter, all of which may reduce the radiation received by the RT, some folks try to compensate by this by using a lower emissivity setting. Don’t do it. Ratio RTs or Peak picking techniques (covered under signal processing) are better methods.

I didn’t know it was this complicated! How do I know what emissivity setting to use?

Well, that’s why I survived all the massive cutbacks in research at both U.S.Steel and Bethlehem Steel: I know what I’m doing (I hope). But I’ll give you a quick and dirty tip that works in some cases: Take a typical section of the material you wish to view and spray half of it with flat black high temperature paint . Heat the material. Focus the RT on the black section with the emissivity set to 0.95. Note the temperature reading. Then shift the RT to the noncoated section and adjust the emissivity to give the same temperature reading. This is approximately the emissivity to use.