How Online Infrared Moisture Analyzers Work

An online infrared moisture analyzer does not “see” water directly. It sees what water does to light, and then infers the rest. The entire measurement rests on a single molecular fact: the O–H bond in water vibrates at frequencies that match specific near-infrared wavelengths, most notably around 1.45 and 1.94 micrometers. 

When near-infrared light strikes a moving material, water molecules absorb part of that energy; the more water present, the stronger the absorption and the weaker the reflected signal that returns to the instrument.

If that were the whole story, a simple intensity reading would suffice. It is not, because the reflected light is also shaped by material color, surface texture, distance from the sensor, and lamp aging. The decisive design response is the two-wavelength or multi-wavelength ratio method. One wavelength is chosen where water absorbs strongly—the measurement wavelength—and another nearby wavelength where water absorption is weak or negligible serves as the reference. 

The instrument alternately measures the reflected energy at both wavelengths and computes their ratio. Because the interfering factors affect both channels similarly, the ratio cancels much of their influence, leaving a signal whose variation tracks moisture rather than surface appearance or optical drift.

The optical path that makes this possible is straightforward in principle. A quartz halogen lamp or an LED emits broadband near-infrared light; optical filters or a rotating filter wheel isolate the measurement and reference bands; the selected light illuminates the material; a collection mirror gathers the diffusely reflected light and focuses it onto a photodetector, commonly lead sulfide or indium gallium arsenide. The detector converts light intensity into an electrical signal, which is then digitized and processed.

Signal processing completes what the optics begin. The raw ratio is converted into absorbance and then into a moisture value through a calibration model established against a reference method. Temperature compensation is built in, because water’s absorption behavior shifts with material temperature, and dynamic averaging or outlier rejection handles the noise that comes from a moving, sometimes uneven surface. The result is a continuous, non-contact reading that tracks the moisture of the material as it passes—not a laboratory number, but a process variable available in real time.