Inductive proximity sensors detect only metallic objects via eddy-current coupling between a sensor coil and the target, so they cannot sense plastics, wood, glass, liquids, or powder; capacitive proximity sensors, by contrast, detect both metallic and non-metallic targets by measuring changes in the dielectric field at the sensing face [S1][S5][S6].
Where the target is non-metallic (plastic pellet, paper stack, resin level, wood grain, glass sheet, granulate in a hopper), the choice is effectively made by physics: only capacitive technology sees the material at all, and the capacitive sensor product family is the default reference design for such duty.
Why inductive fails on non-metallic targets
Inductive sensing couples an AC magnetic field from a coil into the target; a metallic target draws eddy currents that load the coil and reduce its effective inductance, which the oscillator reads as a detect event [S5][S6].
Inductive sensors work by electromagnetic coupling that induces eddy currents in metal targets, so non-metallic materials such as plastics, wood, and ceramics cannot be detected and capacitive sensors are used instead when the target is non-metallic [S2][S5].
The same physics that excludes non-metals also gives inductive sensors high immunity to humidity, condensation, and water films on the face, because water is non-magnetic and does not load the coil, which is the opposite of capacitive behaviour [S5].
Capacitive sensing principle for non-metallic duty
A capacitive proximity sensor forms a sensing field between an active electrode at the face and a reference electrode (often the housing or a guard ring); any material entering the field changes the dielectric coupling and shifts the oscillator frequency or charge-transfer time, which the detection circuit reads as a target [S1][S6].
Because the mechanism is dielectric rather than magnetic, capacitive units respond to nearly any solid, liquid, or powder whose dielectric constant differs from air, and a 1–10 pF shift at the active electrode is enough for a clean switch on typical 12–30 V DC industrial units [S1][S2].
The trade-off is selectivity: a capacitive sensor reacts to anything in its field, including mounting brackets, condensation, and flying chips, so commissioning always includes a sensitivity pot trim and a guard-ring or shielded housing where the environment is wet or dusty [S1].
Decision matrix: target material vs sensor type
The four columns below are the ones that actually break ties in a panel shop, with values grounded in the source material rather than a generic claim; for raw steel, stainless, aluminium, or copper the inductive sensor wins on repeatability and on tolerance to water and coolant films, while for plastic, wood, glass, paper, cardboard, oil, water, and granulate the capacitive sensor is the only technology that detects at all [S1][S2][S3][S5].
Comparison of the two technologies on the criteria that drive a 2026 panel build:
Target material coverage: inductive, only ferrous and non-ferrous metals, no plastic, wood, glass, or liquid; capacitive, metals plus all non-metallics with dielectric constant above roughly 1.5 (covers most engineering plastics, water, alcohols, paper, wood) [S1][S2].
Mounting class: inductive, available in embeddable (shielded) and non-embeddable (unshielded) versions; capacitive, generally treated as unshielded and not embeddable in metal, because the surrounding metal distorts the dielectric field [S1].
Environmental tolerance: inductive, excellent against humidity, water films, coolant, and temperature drift; capacitive, sensitive to humidity, ambient temperature, moisture, and non-metallic residue on the face, which is why a sensitivity adjustment pot is a near-universal feature on capacitive bodies [S1][S5].
Use cases where capacitive is the only option
Liquid level control through a glass or plastic tank wall is a textbook capacitive application, because the wall is non-conductive and the liquid inside has a dielectric constant an order of magnitude above air, so the sensor reads level without ever contacting the process fluid [S1][S2].
Powder and granulate detection in hoppers, silos, and feed bins uses capacitive sensing for the same reason: the dielectric of bulk solids (plastic pellets, flour, wood chips, cement) is high enough relative to air to give a clean switch, and the sensor can be mounted outside a non-metallic sight glass to keep it out of the product zone [S1][S3].
Where inductive still wins on metallic work
For steel and aluminium part detection on a stamping press, an inductive sensor gives better repeatability (typically ±0.1 mm or better at rated Sn) and is immune to the coolant and oil films that swamp a capacitive face [S1][S3].
Underwater and submerged applications, including flow meters and pump chambers, are also an inductive-friendly environment because water does not load the coil the way it loads a capacitive dielectric, so the inductive sense holds its calibration where capacitive units drift [S5].
For a deeper look at how a capacitive sensor's rated distance is set against a grounded metal reference, the capacitive sensor rated distance: grounded metal target as the calibration baseline reference is the natural next read; for the standards angle, the Capacitive Sensor Standard and Functional Principle for Non-Metallic Detection piece traces the IEC 60947-5-2 framework that most 2026 datasheets still quote [S1].
Commissioning and failure modes specific to capacitive
The standard mitigation in 2026 OEM datasheets is a sensitivity adjustment pot, a TEACH-in input on IO-Link capacitive models, and a guard-ring electrode that confines the field and reduces the wall-effect from nearby grounded metal; the Eaton E53 tubular capacitive line is a representative example with sensitivity pot and output LED for commissioning [S1].
For high-temperature or washdown zones (food and beverage, pharmaceutical), capacitive sensors are usually specified in stainless or corrosion-resistant plastic bodies rather than the standard nickel-plated brass used on inductive units, because the housing material is part of the field geometry and metal near the face distorts the dielectric [S1].
Specifying for a 2026 build: what to put on the drawing
For a metallic target application, the callout should read: inductive proximity sensor, shielded or unshielded per mounting, Sn specified against mild steel, same electrical and IP envelope, with the addition that inductive units can be embedded in metal without field distortion, which simplifies the bracket design [S1][S3].
Specifying Sn against steel on a capacitive datasheet and then installing against plastic is a common error that compresses the working range to roughly 20–50% of the headline figure, because the dielectric constant of plastic is far lower than that of grounded steel; the capacitive sensor reference and the Capacitive Sensor Standard and Functional Principle for Non-Metallic Detection article both flag this as the single most common field complaint [S1][S2].
For component-level specifications, see gas detection.