When an LDAR program says a component must be monitored for leaks, EPA Method 21 is the procedure that defines exactly how to do it. It is the reference method for screening equipment components with a portable instrument, and it specifies everything from what concentration counts as a leak to precisely how the probe is moved around a fitting and how the instrument is calibrated. This guide explains what Method 21 is, walks through its leak-definition threshold, probe technique, and calibration requirements, and shows how digitized readings feed into LDAR recordkeeping and trending.
EPA Method 21 in one line: EPA Method 21 is the reference procedure for detecting leaks from equipment components using a portable vapor-analyzing instrument. It defines the leak as a measured concentration at or above a specified threshold in parts per million, and it prescribes how the instrument is calibrated, how the probe is moved along the component's sealing surfaces, and how the maximum reading is recorded. It is the standard screening method that underpins most LDAR monitoring.
Method 21 is built around a handheld vapor analyzer that the technician carries to each component. The instrument draws in air through a probe and measures the concentration of organic vapor in it, most commonly using a flame ionization detector, which burns the sample in a small hydrogen flame and measures the ions produced, or a photoionization detector, which uses ultraviolet light to ionize the vapor. Either way the instrument reports a concentration in parts per million relative to a calibration gas, giving a numeric reading at each point it samples.
The core concept of Method 21 is the leak definition: a component is considered leaking when the analyzer reads at or above a specified concentration threshold, expressed in parts per million. That threshold is set by the regulation that invokes Method 21, and it can differ by component type, service, and applicable rule, so the same physical reading might constitute a leak under one standard and not another. What Method 21 itself standardizes is how the concentration is measured, so that a reading taken by one technician at one facility means the same thing as a reading taken by another elsewhere.
It is important to understand that Method 21 measures a screening concentration, not a mass emission rate. The parts-per-million reading tells you a component is leaking above the threshold and needs attention, but it does not directly tell you how many kilograms per hour are escaping. That distinction matters because the method's job is consistent detection for the LDAR repair trigger, not quantification. A reading over the threshold starts the repair clock regardless of exactly how large the leak is.
How the probe is moved is a defined part of Method 21, not left to the technician's discretion, because the goal is to find the highest concentration at the leak interface reliably. The probe is placed at the surface where leakage would occur, such as the seal of a valve stem, the seam of a flange, or the seal area of a pump, and moved slowly along the entire accessible sealing surface while watching the reading. The technician records the maximum concentration observed as the component moves through the probe path, and it is that peak reading that is compared against the leak threshold.
Distance and speed matter: the probe is held right at the surface rather than backed away, and it is moved slowly enough that the instrument, which has a finite response time, can register a localized leak rather than sliding past it. Getting the technique right is what separates a valid screening from one that misses real leaks, and it is why Method 21 spells the procedure out. The same disciplined approach applies whether the component is easy to reach or awkward, with provisions for difficult-to-monitor and inaccessible components.
Calibration is the other pillar of a valid Method 21 reading, because an instrument that has drifted gives a meaningless number. The analyzer is calibrated against known reference gases before use, and the method requires checking calibration to control for drift over the monitoring period, since instruments can wander with temperature, battery state, and time. If the instrument fails its drift check, readings taken since the last good calibration may be called into question. This is why disciplined daily calibration records are inseparable from the monitoring data itself in any credible LDAR program.
A Method 21 survey generates a stream of data: for each component, the peak concentration, the date and time, the instrument used, the technician, and the calibration status of that instrument. In an LDAR program, that data is not just a pass or fail; it is the evidence that monitoring happened at the required frequency and the trigger that starts a repair clock when a reading exceeds the threshold. Capturing it accurately and completely for every component is essential, and doing so on paper across thousands of components is slow and error-prone.
Digitizing the readings changes both the reliability and the usefulness of the data. When Method 21 results flow into a records and monitoring system such as Merobix, each reading is stored against the specific component in the inventory, its calibration lineage is attached, and any reading over the leak threshold automatically starts the repair clock the LDAR program tracks. That turns a survey from a stack of forms into structured data that can prove monitoring frequency and repair deadlines were met when an inspector asks.
Storing readings as data over time also unlocks trending, which paper cannot easily provide. Because each component's concentration history is retained, an operator can see which components leak repeatedly, which areas of a plant are getting worse, and whether readings are creeping up toward the threshold before they cross it. Watching those trends turns Method 21 from a pure pass-fail gate into an early indicator of developing problems, letting maintenance be targeted at the components and areas that keep recurring rather than treating every survey as an independent snapshot.
Method 21 uses a portable vapor analyzer, most commonly one with a flame ionization detector that burns the sample and measures the resulting ions, or a photoionization detector that ionizes the vapor with ultraviolet light. The instrument reports the organic vapor concentration in parts per million relative to a calibration gas. The method specifies how that instrument is calibrated and how its probe is used, so readings are consistent across technicians and sites.
A component is considered leaking under Method 21 when the analyzer reads at or above a concentration threshold, expressed in parts per million, at the component's sealing surface. The specific threshold is set by the regulation that invokes the method and can vary by component type and service. Method 21 standardizes how the concentration is measured; the applicable rule sets the number that constitutes a leak.
No. Method 21 measures a screening concentration in parts per million at the leak interface, which tells you a component is leaking above the threshold and needs repair, but it does not directly give a mass emission rate. Its purpose is consistent detection to trigger the LDAR repair process, not quantification of how many kilograms per hour are escaping. Quantifying emissions requires separate methods.
Primary references from the standards bodies and regulators that define this topic:
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.