Field guide · Tag engineering
RFID Around Metal and Liquids: Test the Finished Product
Material interaction is an RF design constraint, not a software bug; reliable deployments select and place tags against the real object and workflow.
The Material Is Part of the Antenna System
An RFID tag does not operate independently of the object beneath and around it. Conductive metal changes the electromagnetic boundary around a UHF antenna and can detune a conventional label. Water rich materials absorb UHF energy and can reduce the field available to power and communicate with a passive tag. Software cannot restore energy that never reached the tag.
This is why “RFID compatible” is too broad a description for a product. A design must name the frequency, tag construction, mounting surface, contents, packaging, read geometry, and required workflow.
Metal Requires a Purpose Built Structure
Zebra’s on metal tagging documentation explains that traditional labels can lose substantial performance on metal and that on metal labels use a spacer structure. That added construction changes thickness and can affect printing and encoding equipment.
On metal performance is not one category. A tag designed for a large steel tool may not suit a small curved surgical instrument or a reusable container exposed to cleaning. Mounting method and the available ground plane influence behavior. Buyers should request the exact tag model and test data on a representative object, not a general statement that the product is “metal mount.”
Liquids Change With Fill and Pack
For a bottle or vial, an empty sample is not representative. Fluid composition, fill level, head space, label placement, neighboring containers, tray material, and case packing can change the result. Rotating the package may place the tag behind the liquid relative to the reader.
Near field HF or NFC may be a better fit for some close interactions around water rich materials, but it creates a different workflow and read distance. The choice should come from the required interaction rather than a blanket belief that one frequency always solves liquids.
Test the Operational Envelope
Create a matrix that covers the finished product states: empty and full where relevant, cold and warm if material properties or condensation vary, single and densely packed, expected orientations, movement speeds, and realistic read zone congestion. Measure successful identification at the business event level, not the number of low level reader observations.
Reader and antenna configuration should be fixed and documented during comparisons. Increasing power can enlarge an uncontrolled zone and create stray reads even if it improves one difficult orientation. The goal is dependable discrimination of the intended population.
Treat tag placement as a controlled product attribute. A few millimetres of movement can alter coupling to a metal edge, liquid volume, or neighboring package. Define an allowed placement window, validate the label application process, and inspect production variation rather than relying on a carefully prepared engineering sample. If contents or packaging change, repeat the relevant tests and retain the configuration with the result.
Acceptance testing should use enough samples and production lots to reveal variation. Record misses and unintended reads by configuration rather than averaging them away. The approved setup should name the tag revision, placement, reader firmware, antenna geometry, power, and line conditions.
Deployment Questions
- Is the proposed tag designed for the exact mounting material and available footprint?
- Can printing and encoding equipment handle the tag thickness and construction?
- Was testing performed on filled, sealed, packed product?
- Which orientations and movement speeds represent normal and worst case handling?
- What is the acceptable miss and stray read behavior at the workflow level?
- How do temperature, moisture, cleaning, and aging affect attachment and performance?
- Would a different frequency or interaction model simplify the physical problem?
Limitations
Vendor fact sheets identify the engineering problem and available approaches, but they do not establish universal range. Read distance depends on the complete configuration and regulatory environment. Treat any uncited range claim as a test hypothesis, not a specification for your deployment.