Refrigerant Leak Detection Methods & Tools

By Jennifer Kirstens · August 30, 2026 · 8 min read

PES Supply is expanding into refrigerant service tools — including AHRI 740-certified electronic leak detectors and UV leak-detection kits. Contact us to join the waitlist. This guide covers the four leak-detection methods professionals use, AHRI 740 sensitivity standards, A2L-compatible detector requirements, and step-by-step inspection procedures.

The EPA's Refrigerant Management program (40 CFR Part 82 Subpart F, updated most recently under the AIM Act) requires operators of appliances containing more than 50 lb of refrigerant to repair leaks that exceed 10% annualized leak rate on commercial refrigeration or 20% on comfort cooling. Above those thresholds you either fix within 30 days or retrofit / retire the system. That regulation, plus new A2L flammability considerations, plus real-world compressor failure costs pushing $4,000+ per residential unit, is why every professional-grade service call now starts with leak isolation before evacuation and charging.

For A2L refrigerants (R-32, R-454B, R-1234yf), leak detection also becomes a life-safety measure. A2L classification means the refrigerant is mildly flammable in air at concentrations between roughly 3.5% and 14% by volume. Slow leaks into unventilated equipment closets can build up over hours. Every A2L equipment installation under UL 60335-2-40 requires a fixed leak-detection sensor in the enclosure or a ventilation solution — and every service call requires a portable detector on the truck.

Method 1 — Electronic (heated diode, IR, or MEMS sensor)

Electronic detectors sniff air near suspected leak points and alarm when refrigerant vapor is detected. Three sensor technologies dominate the pro market:

• Heated diode: A platinum-tungsten sensor detects halogen atoms. Consumable — sensor life 400–1500 hours. Very sensitive (0.05 oz/yr on best models). Popular with Inficon Tek-Mate, Fieldpiece SRL8, Bacharach H-10 Pro.
• Infrared (non-dispersive IR): IR beam passes through the sample cell; a photodetector measures absorption at refrigerant-specific wavelengths. Long life (10+ years), false-alarm resistant, expensive ($400–$650). Popular with Testo 316-3, Inficon D-TEK Select.
• MEMS chemical sensor: Newer solid-state technology, low cost, moderate sensitivity. Common in shop-tier detectors.

When to use: Primary leak-detection method for every service call. Fast, quantifiable, and covers all common refrigerants.

Method 2 — UV fluorescent dye

A fluorescent dye compatible with the system's oil (SAE J2297) is injected into the refrigerant charge. Under a 365 or 400 nm UV lamp, the dye glows at the leak site. Best for confirming slow or intermittent leaks that electronic detection struggles with.

When to use: After electronic detection narrows the search area, or on customer callback for "recharged 6 months ago and now low again" symptoms.

Method 3 — Bubble solution

Soapy water or purpose-made bubble solution (Big Blu, Nu-Calgon Snoop) applied to fittings, brazed joints, and Schrader cores under pressure. The oldest method, and still the most decisive for locating a leak at a specific fitting once electronic detection points to a general area.

When to use: Final confirmation after electronic detection identifies a general area. Also the go-to for verifying a repair after brazing.

Method 4 — Ultrasonic

Detects the ultrasonic hiss of gas escaping through a leak point. Works even when the system pressure is very low or in noisy environments where electronic detectors struggle. Common in commercial refrigeration and industrial applications.

When to use: On systems with residual pressure, in noisy mechanical rooms, or when electronic detection is defeated by wind or ambient contamination.

AHRI Standard 740 (current version 740-2020) is the industry benchmark for refrigerant leak-detector performance. It defines three test conditions and a rated sensitivity per refrigerant type. The 2020 revision was significant: it explicitly added A2L refrigerants (R-32, R-454B, R-1234yf, R-1234ze) as test refrigerants and updated the false-alarm criteria to reflect field conditions.

AHRI 740 tests a detector at three sensitivity levels: low, medium, and high. Certification means the detector meets or beats the manufacturer-rated sensitivity at each level, with a false-alarm rate below the standard's threshold.

Table 2 — AHRI 740-2020 sensitivity classes

Not every detector is rated at every level. A detector might carry AHRI 740 certification for Class II R-410A but not Class I R-32. Read the certificate carefully — the phrase "AHRI 740 Certified" without a class or refrigerant listing is marketing.

Table 3 — Refrigerant type detectability by sensor technology

Legacy detectors tuned for R-410A and R-22 (both A1 refrigerants) commonly fail on A2L blends in one of two ways:

• False-alarm on R-32's higher vapor pressure — the sensor sees more molecules and interprets it as a large leak.
• Under-detect R-454B because the primary detection wavelength is off from the blend's IR absorption peak.

A2L-rated detectors solve both by recalibrating the sensor response, adding refrigerant-specific detection modes, and (for heated-diode units) using an A2L-safe sensor cell that won't act as an ignition source in a flammable atmosphere.

Recommended A2L-certified detectors (2026)

• Inficon D-TEK Select A2L — dual-IR sensor with dedicated A2L mode. AHRI 740 Class I for R-32, R-454B, R-410A. $580–$650.
• Fieldpiece SRL8 — heated-diode with A2L calibration. AHRI 740 Class II for R-32 and R-454B. $350–$390.
• Testo 316-3 A2L — IR-based, ATEX-rated for use in A2L atmospheres. AHRI 740 Class I. $520–$580.
• Bacharach H-10 Pro A2L — heated-diode, long-standing pro brand. AHRI 740 Class II. $290–$340.
• Yellow Jacket 69365 — MEMS-based, budget option. AHRI 740 Class III. $230–$270.

A proper AHRI 740-style field leak inspection is more structured than "wave the sniffer around." The industry procedure most OEMs will accept for warranty documentation:

After running electronic detection on tens of thousands of systems, the industry has a consistent picture of where leaks originate. In rough order of frequency:

Structured inspection means hitting these locations in order rather than randomly waving the probe.

A system can be tested for leaks two ways: pressurized (with dry nitrogen or trace refrigerant), or under vacuum (via decay test). Both have their place, but they detect different things.

Pressure test reveals leaks flowing outward at operating-like conditions. Fittings that seal fine at 30 psi may leak at 400 psi — pressure testing to 550 psi high side on R-410A/R-454B is how you catch these. Downside: you can't isolate the exact leak location without a detector or bubble solution.

Vacuum decay test reveals leaks flowing inward at low pressure. Some slow leaks (formic-acid pinholes in aluminum evaporators) show up on vacuum decay but seal themselves under positive pressure. Downside: vacuum decay does not distinguish between a leak and outgassing moisture — you may chase a phantom leak that is actually residual water.

Professional practice: pressure-test with nitrogen first (locates any large leaks), then evacuate and run a decay test after triple-evacuation (confirms tight system before charging).

Above we mentioned UL 60335-2-40 fixed leak sensors. Here's the practical breakdown for installers.

• Charge limits by room area. The standard defines a maximum refrigerant charge (M1) that can be installed in a room of a given area without any leak mitigation. Below that, no sensor needed. R-32 residential systems 3 tons and under typically fit under M1 for average-sized rooms.
• M2 range. Above M1 but below the M3 hard limit, a fixed sensor with mitigation (typically forced ventilation) is required.
• M3 hard limit. Above this charge, the equipment cannot be installed in the space at all without a fundamental redesign (larger room, secondary loop, chilled water).
• Sensor location. Mount 12–18 inches above the floor (R-32 and R-454B are heavier than air). Not in dead-air corners. Wired to the OEM control board or a standalone shutoff/ventilation controller.
• Bump-test schedule. Fixed sensors need annual bump-testing with a calibration gas. Owners often skip this; it invalidates the warranty.

Electronic detectors alarm on many things that are not refrigerant. Field techs recognize:

• Cleaning solvent residue. Fresh brake cleaner, contact cleaner, or WD-40 near a service port will trigger a heated-diode detector for 5–15 minutes after application.
• Silicone caulk vapors. Newly-cured silicone off-gasses methanol; false-alarms IR detectors for hours.
• Windshield washer fluid, alcohol-based cleaners. Common in service vans, false-alarm heated diodes.
• High humidity. Some cheap MEMS sensors alarm on water vapor. Pro-tier units auto-compensate.
• Cross-contamination on the probe tip. If you touched a leaky fitting, the probe now reads leak everywhere. Bump-test in clean air to zero it.

Ultrasonic leak detectors listen for the 20–40 kHz turbulence signature of gas escaping through a small orifice under pressure. Two applications where ultrasonic beats electronic:

• Noisy commercial mechanical rooms. Fans, pumps, and compressor noise contaminate electronic-detector sniff samples. Ultrasonic uses a directional parabolic microphone that isolates the leak signal.
• Low system pressure. When system pressure is below refrigerant saturation temperature (residual charge, cold ambient), electronic sensors don't get enough refrigerant vapor to alarm. Ultrasonic still detects the gas-flow turbulence.

Modern ultrasonic detectors (SDT270, UE Systems Ultraprobe 550, Fluke ii910) run $500–$2,500 and include contact probes, air-borne parabolic dishes, and headphones. They are complementary to electronic detection, not a replacement — electronic remains faster and quantifiable for typical residential and light-commercial refrigerant leaks.

When a system is charged, an electronic detector can locate the leak with the refrigerant already in the system. When a system is empty (post-repair, new install), you have two options:

• Nitrogen with trace refrigerant. Add 5–10 psi of the target refrigerant, then pressurize to 550 psi with dry nitrogen. Leak-test with electronic detector. Recovery is required after test. Uses about 4 oz of refrigerant per test.
• Pure nitrogen with bubble solution. Pressurize with straight nitrogen and locate leaks with bubble solution. Slower, but avoids the refrigerant recovery step.

Trace-refrigerant testing is banned for A2L installations — the flammable-vapor risk during test doesn't warrant the small time savings. Use pure nitrogen with bubble solution or ultrasonic for A2L work.

UV dye is a confirmation tool, not a primary detector. Common mistakes:

• Overdosing. Manufacturer spec is typically 1/8 oz per 1.5 lb refrigerant. Doubling that causes the whole system exterior to glow, not just the leak.
• Wrong dye chemistry. R-32 and R-454B use POE oil. Only POE-compatible UV dyes (Spectroline OIL-GLO 44, Tracerline TP-3820) work. Mineral-oil dyes cause acid formation.
• Wrong lamp wavelength. Different dyes fluoresce at different wavelengths. 365 nm lamps pair with green-fluorescent dyes; 400 nm lamps pair with yellow-fluorescent dyes. Buy them as a kit.
• Not enough run time. Dye migrates slowly. A system running only 30 minutes after injection often shows no glow at the leak. Give it a week of normal operation and re-inspect.

UL 60335-2-40 mandates that A2L equipment either limit refrigerant charge to below a room-dependent threshold or include a fixed leak-detection sensor that triggers ventilation. Common fixed sensors from Honeywell, Bacharach, Testo, and MSA are wall-mounted at 12–18 inches from the floor (R-32 and R-454B are heavier than air), wired to the equipment control board, and calibrated to trip at 25% of LFL (lower flammable limit).

These are installer-provided in most commercial applications; some residential A2L equipment ships with the sensor integrated in the air handler and does not require a separate wall unit. Read the OEM install manual — this is where most first-time A2L installers get tripped up on inspection.

Different contractor profiles justify different kits. Rough guidance:

• Solo residential contractor, mostly R-410A + A2L retrofits: Mid-tier heated-diode detector (Fieldpiece SRL8), UV dye kit with universal POE-compatible dye, and a small bottle of bubble solution. Total kit around $450.
• Mini-split / ductless specialist: Class I dual-IR detector (Inficon D-TEK Select A2L), UV kit, and torque-wrench-verified flare-fitting tools. Total around $780. The Class I sensitivity matters here because mini-split leaks are often at fittings and small in absolute mass.
• Commercial rooftop and light industrial: Ultrasonic detector plus mid-tier electronic (dual-technology approach). Class II sensitivity is enough. Add a 4-oz mini-cylinder of R-134a for trace-leak testing on A1 systems. Total around $1,200.
• Supermarket / refrigeration rack: Multi-gas fixed sensor package for ammonia and CO₂ systems, plus ultrasonic and IR portable units. Total around $3,000+.

PES Supply's forthcoming refrigerant-tools bundles will target the first two categories — solo residential and mini-split — as our primary customer base.

PES Supply currently stocks:

• R-410A Refrigerant — 25 lb Cylinder
• R-134a Refrigerant — 30 lb Cylinder
• R-22 Refrigerant — 30 lb Cylinder (Reclaimed, AHRI 700)

For related buying guides see HVAC Refrigerant Service Tools Buying Guide and HVAC Vacuum Pump & Evacuation Guide.

PES Supply is expanding into refrigerant leak detection tools. AHRI 740-2020 certified electronic detectors, UV dye kits, ultrasonic detectors, and fixed A2L monitoring sensors. Contact us to join the waitlist.

What is AHRI 740 certification?

AHRI Standard 740 is the industry benchmark for refrigerant leak-detector performance. AHRI 740-2020 defines three sensitivity classes (0.05, 0.10, 0.25 oz/yr) and explicitly covers A2L refrigerants including R-32, R-454B, and R-1234yf.

Do I need an A2L-rated leak detector to service R-32 systems?

Yes. Legacy A1-tuned detectors either false-alarm on R-32's higher vapor pressure or fail to detect R-454B at rated sensitivity. Modern A2L-listed detectors use dual-IR or A2L-safe heated-diode sensors.

How sensitive should my leak detector be?

For residential and light commercial: AHRI 740 Class II (0.10 oz/yr) is the pro standard. For MVAC (mobile A/C) and mini-split: Class I (0.05 oz/yr) catches slow leaks that Class II misses.

Can I use UV dye instead of an electronic detector?

UV dye is a confirmation tool for slow or intermittent leaks; it is not a substitute for electronic detection during initial inspection. Use them together — electronic first to isolate the area, then UV dye or bubble solution to pinpoint the fitting.

Why does my leak detector false-alarm at every fitting?

Most often the probe tip is contaminated by residual refrigerant from a prior alarm. Move to clean air, wait 30 seconds for the sensor to zero, and re-approach the fitting slowly.

Do heated diode sensors expire?

Yes. Sensor life is 400–1500 hours depending on model and refrigerant load. When sensitivity drops below AHRI-rated levels, replace the sensor cartridge. Infrared sensors last 10+ years and do not require replacement.

Is bubble solution accurate enough for professional service?

Bubble solution reliably locates leaks above roughly 1 oz/yr — fine for confirming a fitting after electronic detection points to it, but not sensitive enough for the initial isolation of a small leak.

What is the EPA leak-rate limit for HVAC equipment?

Under EPA 40 CFR Part 82 Subpart F: 20% annualized leak rate for comfort cooling (residential/commercial A/C) or 10% for commercial refrigeration. Above those thresholds, the leak must be repaired within 30 days or the appliance retrofitted or retired.