EPA Method 21 Requirements for VOC Leak Detection

The EPA’s Method 21 requirements set out how a facility must screen individual pieces of process equipment for volatile organic compound (VOC) leaks: which instrument to use, how to calibrate it, how to move the probe around each valve, pump, flange, or connector, what reading counts as a leak, and how quickly a leak must be fixed and documented. Method 21 is a screening procedure, not a mass emissions measurement. It tells you which components are leaking, not how much VOC your facility is releasing overall.

Where Method 21 Applies

Method 21 is the monitoring procedure written into federal Leak Detection and Repair (LDAR) programs under two frameworks: New Source Performance Standards in 40 CFR Part 60, and National Emission Standards for Hazardous Air Pollutants in 40 CFR Parts 61, 63, and 65.¹ The components typically covered are valves, flanges, threaded connections, pumps, compressors, pressure relief devices, agitators, and process drains.

Every applicable rule sets its own leak definition, meaning the concentration above which a reading counts as a leak. Many NSPS rules use 10,000 parts per million. NESHAP rules are often stricter: under 40 CFR Part 63 Subpart H, Phase III valve leak definitions drop to 500 ppm, with pump thresholds ranging from 1,000 ppm in general service down to 500 ppm for pumps handling ethylene oxide.² Monitoring frequency also depends on the rule and the component: some valves require monthly checks, while certain connectors may only need monitoring once every eight years under performance-based schedules. Check the specific subpart that applies to your unit before setting a schedule.

Instrument Specifications

Method 21 requires a portable VOC analyzer, usually a Flame Ionization Detector (FID) or Photoionization Detector (PID). The meter scale must be readable to within plus or minus 2.5 percent of the applicable leak definition. The instrument needs an electrically driven pump delivering a constant sample flow between 0.10 and 3.0 liters per minute to the detector.³

Response speed is critical. The instrument must reach 90 percent of its final stable reading in 30 seconds or less. The probe inlet cannot exceed 6.4 millimeters (one-quarter inch) in outside diameter, and it must have a single opening for sample admission.³ Because you are sampling flammable vapors, the instrument must also be intrinsically safe for use in potentially explosive atmospheres.

When concentrations exceed the instrument’s upper range, a dilution probe assembly can bring readings back into the linear response range. The 30-second response time and probe diameter limits still apply with the dilution probe attached.

Reference Compounds and Response Factors

Every Method 21 instrument is calibrated against a reference compound specified by the applicable regulation. Methane is common for natural gas and petroleum facilities; isobutylene is used in other settings. A response factor is the ratio of the known concentration of a VOC to the meter reading observed when the instrument is calibrated with the reference compound.³

The response factor for every VOC species the instrument will encounter must be less than 10, unless the applicable rule says otherwise. If no available instrument meets that limit using the rule’s specified reference compound, you can calibrate with one of the target VOCs or another compound, as long as every response factor stays under 10. Published response factors for the compounds and detector type in use can be relied on in place of new tests.

Calibration Before Use

Before field monitoring, the instrument must pass a calibration precision test using two gases: a zero gas containing less than 10 ppm of VOC, and a calibration gas at a known concentration of the reference compound approximately equal to the applicable leak definition.³

The check alternates between zero gas and calibration gas three times, recording each reading. The average difference between the readings and the known calibration value, expressed as a percentage of that value, must be 10 percent or less.³ This must be done before the analyzer goes into service, and again at three-month intervals or at the next use, whichever comes later. If the instrument will not adjust to the calibration gas value, it cannot be used until corrective action brings it back into specification. Response factor testing must also be completed before the instrument is placed in service.

How to Conduct the Survey

Place the probe inlet at the component interface surface where leakage is most likely, then move it slowly along the periphery while watching the meter. When the reading rises, slow down and home in on the maximum. Hold the probe at that maximum-reading location for approximately twice the instrument’s response time.³

For valves, work the probe around the stem circumference at the packing gland where the stem exits the body. For rotating shafts on pumps or compressors, position the probe inlet within one centimeter of the shaft-seal interface and traverse the circumference.³ That close proximity is non-negotiable, because VOC concentrations fall off sharply with distance from the emission point.

If the maximum observed reading exceeds the leak definition in the applicable rule, the component is a leaker and must be repaired.

The “No Detectable Emissions” Case

Some rules require a “no detectable emissions” determination instead of a straight concentration threshold. Method 21 handles this by measuring local ambient VOC concentration: move the probe randomly upwind and downwind at one to two meters from the source. If a nearby source interferes, the background reading may be taken closer to the component, but never less than 25 centimeters away. A component has no detectable emissions when the reading at the surface, minus the background reading, is less than 2.5 percent of the applicable leak definition.³

Repair Deadlines Once a Leak Is Found

Once a component exceeds the leak definition, deadlines start running. Under both NSPS and NESHAP programs, the standard timeline is a first attempt at repair within 5 calendar days and final repair no later than 15 calendar days after detection.⁴² First attempts include practical steps like tightening packing gland nuts or verifying that seal flush systems are operating at design pressure and temperature. The leaking component must be tagged with its equipment identification number and re-monitored after repair to confirm the reading now falls below the leak definition.

Delay of Repair

Repair within 15 days can be delayed when it would require a process unit shutdown to complete. In that case, repair must be finished before the end of the next scheduled shutdown, and verification monitoring must happen within 15 days after the unit restarts.⁵ Other delay provisions cover specific situations:

If the component is isolated from the process and no longer in VOC service, repair may be delayed.
For valves and connectors, delay is allowed when the facility demonstrates that emissions from purging during immediate repair would exceed the fugitive emissions from the continued leak, provided purged material is routed to a control device during eventual repair.
For pumps requiring a dual mechanical seal system with a barrier fluid system, repair may be delayed up to 6 months.

Unsafe-to-monitor equipment is exempt from routine frequency requirements when monitoring would put personnel in immediate danger, and difficult-to-monitor equipment (generally components requiring personnel to be elevated more than two meters above a support surface) typically must be monitored at least once per calendar year. Both classifications require a written plan.

Records and Reports

Documentation is not optional. For every monitoring event, records must include the instrument identification number, the operator’s name or initials, the unique equipment identification number of each component monitored, the date, and the measured concentration reading.

Calibration records must include:

Date of calibration
Gas cylinder identification
Certified concentration of the calibration gas
Any corrective actions taken when the instrument failed to adjust to the calibration gas value

The results of any end-of-day calibration drift assessment must also be recorded. For every leak, the file must show the date of discovery, the date of first repair attempt, the date of final repair, and the post-repair reading confirming the leak was resolved. Records must generally be retained for at least five years.⁶

Most LDAR programs also require semi-annual reporting, due within 60 days after the end of each 180-day period. If nothing went wrong, the report can be a simple statement of compliance. When exceptions did occur, reports must detail each leak event: the component involved, the concentration reading, the repair dates, and for pressure relief device releases, an estimate of the total hazardous air pollutant quantity released along with the corrective measures taken to prevent recurrence.

Optical Gas Imaging as an Alternative

The EPA’s Alternative Work Practice allows optical gas imaging (OGI) cameras in place of a Method 21 monitor for identifying leaking equipment, provided the applicable subpart requires Method 21 monitoring.⁷ OGI cameras detect gas plumes visually, letting an operator scan many components quickly. Any emissions visible through the camera are treated as leaks.

The OGI instrument must undergo a daily check before each monitoring session, and the facility must pick one of three frequency-sensitivity pairs:

Bi-monthly monitoring at 60 grams per hour detection sensitivity
Semi-quarterly monitoring at 85 grams per hour
Monthly monitoring at 100 grams per hour

These schedules replace whatever monitoring frequency the applicable subpart would otherwise require.⁸ OGI does not fully replace Method 21, though. Facilities using the alternative practice must still perform an annual Method 21 survey at the applicable leak definition. Reduced monitoring for good performance, skip-period schedules, and allowable leak percentage provisions do not apply when using OGI. The alternative is also unavailable for closed vent systems, equipment designated as leakless, and equipment classified as having no detectable emissions below 500 ppm.⁷

Penalties for Falsification or Missed Deadlines

Under the Clean Air Act’s criminal enforcement provisions, knowingly making a false statement in any required record or report, or knowingly tampering with a monitoring device or method, carries a penalty of up to 2 years of imprisonment and a fine per violation. On a second conviction, the maximum fine and prison term double.⁹ These penalties reach not only outright fabrication but also omitting material information or failing to maintain required documentation.

Civil penalties apply to missed monitoring deadlines and unrepaired leaks. Enforcement often targets patterns of neglect, but a single missed repair deadline still constitutes a violation from the moment the deadline passes. Facilities that act to repair within the specified window but do not succeed in fixing the leak are not in violation, so long as they continue pursuing the repair under the applicable delay-of-repair provisions.

1
eCFR. 40 CFR Part 60 – Standards of Performance for New Stationary Sources
2
eCFR. 40 CFR Part 63 Subpart H – National Emission Standards for Hazardous Air Pollutants for Equipment Leaks and Fenceline Monitoring
3
Environmental Protection Agency. Method 21 – Determination of Volatile Organic Compound Leaks
4
eCFR. 40 CFR 60.482-7a – Standards: Valves in Gas/Vapor Service and in Light Liquid Service
5
eCFR. 40 CFR 60.482-9a – Standards: Delay of Repair
6
eCFR. 40 CFR 63.1259 – Recordkeeping Requirements
7
eCFR. 40 CFR 60.18 – General Control Device and Work Practice Requirements
8
eCFR. Table 1 to Subpart A of Part 60 – Detection Sensitivity Levels
9
Office of the Law Revision Counsel. 42 USC 7413 – Federal Enforcement