RFID frequency guide: LF, HF / NFC and UHF for procurement teams

Yes, but each frequency requires its own reader fleet — the hardware does not overlap. Buildings routinely deploy LF Prox for legacy door access + HF MIFARE for payment + UHF for inventory all in parallel. Multi-band readers exist (some desktop enrollment stations and handheld devices support 125 kHz + 13.56 MHz + 860-960 MHz in one unit) but cost 3-5x single-band reader prices. The only time buyers need to worry about RF interference between frequencies is when antennas are installed within 0.5 m of each other at dock doors — LF inductive fields can desensitize UHF receivers at very close range. Separate mounting by 1-2 m and no practical interference occurs.

Yes — HF 13.56 MHz penetrates water and biological tissue far better than UHF, which is absorbed heavily. In practice, this matters for: (1) pet microchipping and veterinary implants — LF / HF only, never UHF; (2) livestock identification through skin, muscle and rumen fluid — LF primarily, HF secondarily; (3) medical devices implanted in tissue; (4) swimming pool / water park wristbands where the tag sits against wet skin; (5) industrial laundry where tags are tumbling in wash water (though this is engineered around with textile encapsulation rather than relying on HF itself). For non-biological wet environments, HF advantage is minor — waterproof UHF wristbands work fine for cruise ship all-inclusive programs where the tag doesn't need to read while submerged at depth.

Regulatory policy difference. FCC Part 15.247 allows unlicensed 4 W EIRP in the 902-928 MHz ISM band. ETSI EN 302 208 allows 2 W ERP (which converts to about 3.3 W EIRP) in the 865.6-867.6 MHz band. Practical impact on read range: US installations reach ~8-12 m at dock doors, EU installations ~6-9 m — a 20-30% range reduction due to lower transmit power plus narrower frequency band (only 2 MHz vs 26 MHz in US). EU installations compensate with tighter antenna focusing and higher-gain directional antennas. Tag-side: same chip works in both regions, only the reader TX power differs. When specifying a global rollout, design to the EU constraint — US deployments will then have generous headroom.

Rare outside niche applications. Microwave RFID at 2.45 GHz historically served: electronic tolling (now mostly migrated to UHF or DSRC), active container tracking (Savi Technologies, largely BLE-replaced), and a handful of race timing systems. The 2.45 GHz band is heavily shared with WiFi, Bluetooth, and microwave ovens — interference makes it unattractive versus clean UHF. For new B2B deployments in 2024-2026, 2.45 GHz is effectively never the right answer. If a vendor is proposing 2.45 GHz for a new project, ask pointed questions about why — they may be trying to sell legacy stock rather than the best fit. Exception: military ISO 18000-4 deployments for high-mobility asset tracking in certain national defense networks remain at 2.45 GHz by regulation.

Answer four questions in order. (1) Do consumers need to tap with phones? If yes, HF 13.56 MHz NFC — nothing else works with smartphones. (2) Do you need to read many items at once from a few meters away? If yes, UHF 860-960 MHz. (3) Is this an upgrade of an existing LF Prox or HF MIFARE deployment? Match the installed base to avoid reader replacement cost. (4) Is this livestock / pet / implant application? LF 134.2 kHz. If none of these apply, the default for new industrial deployments is UHF 860-960 MHz because of cost-per-tag economics. When still uncertain, request samples of all three frequencies from RFIDAK and test against your actual reader before committing.

One tag can work globally if you specify "worldwide antenna" design. UHF Gen2 chips are inherently broadband and work across all regional sub-bands (FCC 902-928, ETSI 865-868, Japan 916-921, China 920-925) — only the antenna needs optimization. A worldwide-tuned antenna trades 10-15% peak range per region for consistent performance across every region. For supply chain tags moving between regions (export cartons, global RTPs), always specify worldwide. For region-locked inventory (US-only retail, EU-only DPP), region-specific antenna delivers maximum range at minimum cost. HF and LF have no regional variation — 13.56 MHz and 125-134 kHz are globally unified under ISO standards.

At MOQ 100K, UHF sticker inlays are the cheapest at $0.04-0.08 per unit (benefit of massive manufacturing scale from retail mandates). HF NFC stickers run $0.05-0.15. LF stickers are oldest technology and counterintuitively most expensive at $0.18-0.30 because LF manufacturing volume is 1-2% of UHF volume, so scale economies don't apply. For PVC cards, the cost curve inverts — HF MIFARE Classic cards are $0.15-0.40, UHF cards $0.20-0.55 (fewer suppliers make UHF cards vs stickers). The cheapest option for a new deployment is usually UHF stickers unless a use case specifically demands another format. For LF, expect a 2-5x cost premium versus UHF for equivalent format and volume.

No — 5G mmWave operates at 24-71 GHz which is physically unsuitable for passive RFID. The wavelengths are millimeter-scale (matching the name), so tag antennas would need to be micrometer-scale to resonate — currently impractical. 5G does interact with RFID ecosystems via reader backhaul connectivity (fixed readers uploading inventory counts over 5G networks) but not at the tag layer. Emerging research on 5G-powered passive sensors shows promise at 5-10 year horizons but won't disrupt UHF Gen2's 860-960 MHz system in the current procurement cycle. For any project with a 2024-2028 decision window, UHF 860-960 MHz is the safe bet — it has 42 billion tags deployed in 2024 alone (RAIN Alliance) and global chip supply at scale.

Three practical moves. (1) Pick established ISO standards, not proprietary protocols — ISO/IEC 14443 (HF), 15693 (HF vicinity), 18000-63 (UHF Gen2v2), 11784/11785 (LF). All have 20+ year commitment from international standards bodies. (2) Avoid vendor-locked proprietary chips — if your chip supplier goes out of business, a standards-compliant chip has 5+ alternative sources. (3) Document your RFP in generic terms allowing drop-in chip substitution at repeat-order time — e.g., "any EPC Gen2v2 compliant chip with 128-bit EPC" rather than "Impinj Monza R6 specifically". Over 10 years, your current chip will go end-of-life; standards-based spec keeps you supplied. Frequencies themselves have stable 20+ year horizons; the risk is at the chip level, not the frequency level.

Rare but real cases. (1) Retail with consumer engagement: UHF for warehouse + NFC for shopper tap, as discussed in luxury DPP. (2) Access control migration: LF + HF dual-frequency card during transition years when you need both old LF readers and new HF readers to recognize the same card. (3) Global asset tracking with mobile field scanning: UHF for fixed portals + NFC for technician smartphone spot-check. Dual-frequency tags cost 40-80% more than single-frequency and add complexity — avoid unless there is a specific operational driver. Most "maybe we need dual frequency" thinking resolves to "run two parallel single-frequency deployments" once the actual workflow is mapped. Sample-test both before committing to the dual-frequency format because antenna coupling between the two frequencies can reduce range.

All three frequencies have similar operating range when chip and housing are properly engineered — typical range is −25°C to +70°C for standard inlays, extendable to −40°C to +110°C for industrial-grade construction. Differences emerge at extreme conditions: LF and HF tolerate higher humidity better (water is more transparent at lower frequencies), while UHF is more sensitive to water damage if the chip substrate cracks. For applications requiring autoclave sterilization (134°C), PPS-encapsulated chips handle all three frequencies but cost 3-5x baseline PVC. For cold-chain cryogenic applications (−40°C or below), active BLE or cellular trackers are typically more reliable than passive RFID regardless of frequency. Always specify operating temperature range in RFP to get the right substrate.

Yes for most applications. Desktop reader-writer units (uFR Nano, ACR122U for NFC / HF; Impinj Speedway or desktop UHF writers for EPC Gen2) both encode and read within one device. For high-throughput encoding lines (encoding 10K+ cards/day), dedicated auto-encoding stations with conveyor feeding are more efficient and include final read-verification. For field-operations reader fleets (warehouse gates, point-of-sale), reading is usually read-only — encoding happens at the manufacturing facility or during initial provisioning. Clarify encoding needs upfront: if encoding is daily (hotel key programming), buy encoding-capable readers. If encoding is one-time at manufacture, buy read-only readers which are 30-50% cheaper.