A flare has to stay lit and safe at the top of a tall stack while the pipe below it is full of flammable gas, and the great danger is that air finds its way down the stack, mixes with that gas, and ignites inside the pipe. A molecular seal is a mechanical device fitted below the flare tip whose whole purpose is to stop that from happening by trapping a pocket of light gas that air cannot easily displace. This guide explains how a molecular seal works, why it relies on gas density rather than a moving part, and how a control system watches purge flow and header oxygen to back it up.
Molecular Seal in one line: A molecular seal is a chamber fitted below a flare tip that traps a pocket of light purge gas whose low density resists being pushed aside by heavier air, creating a barrier against air ingress and flashback into the flare stack. It has no moving parts, working purely on the density difference between the trapped light gas and the surrounding air, and it is monitored indirectly through purge-gas flow and header-oxygen alarms.
The molecular seal, sometimes called a density or labyrinth seal, is a specially shaped internal geometry built into the top of a flare stack, just below the burner tip. Its form creates a change in direction and an inverted region where a light purge gas, typically natural gas or another low-molecular-weight fuel, collects and sits. Because that trapped gas is lighter than air, it resists being displaced downward by the heavier atmosphere above. Air trying to work its way into the stack has to push against a pocket of gas that wants to float upward, and the geometry means it cannot simply flow straight through.
This is a purely passive mechanism. There is no valve, no flap, and nothing to actuate; the seal works because of the physics of two gases of different density meeting in a shaped space, which is why it keeps functioning without power or maintenance for long periods. As long as a small continuous purge keeps replenishing the light-gas pocket, the seal holds and the interface between the trapped gas and the outside air stays parked in the seal region rather than migrating down the stack.
The reason all of this matters is flashback. If air travels down into a header full of flammable gas, it can form an explosive mixture inside the pipe, and an ignition source, the flare flame itself, sits right at the top. An internal explosion or a flame front travelling back down the stack can rupture equipment and is one of the most serious hazards in flare operation. The molecular seal is one of the primary defenses against that scenario, sharply reducing how much sweep gas is needed to keep air out compared with relying on purge velocity alone.
A molecular seal does not work in isolation; it needs a continuous supply of purge gas to keep its light-gas pocket topped up. Over time, diffusion and thermal effects slowly let the pocket degrade, and a small steady purge flow refreshes it so the density barrier is maintained. The seal's great advantage is that it lets the required purge rate be much lower than it would be without a seal, because the operator no longer has to rely on brute velocity to keep air from creeping in. That lower purge saves gas and, because purge gas is itself burned, reduces the associated emissions.
The seal has limits worth understanding. It is designed to resist the slow ingress of air by diffusion and buoyancy, not to hold back a strong reverse flow, so anything that pulls a vacuum or draws gas back down the stack, such as sudden condensation of steam or rapid cooling of hot gas in the header, can defeat it. That is one reason a liquid seal drum or other protection is used in combination on many systems. The molecular seal is a strong first line of defense against ordinary air ingress, not a guarantee against every upset that could pull air in.
Because the seal is passive and its condition cannot be read directly, operators monitor the conditions that keep it working rather than the seal itself. The two most important are that purge gas keeps flowing at or above its minimum rate, and that oxygen does not appear in the flare header where it should not be. If either of those goes wrong, the assumption that the seal is protecting the stack no longer holds, and that is exactly what the control system is set up to detect.
A control system cannot see inside a molecular seal, so it watches the measurable signals that tell it whether the seal's job is being done. The first is purge-gas flow. A low-flow alarm on the purge line warns that the sweep keeping the light-gas pocket refreshed has fallen below its safe minimum, which means the seal may no longer be maintained and air could begin to creep in. Because purge is a continuous, life-safety function, that alarm is typically given high priority and clear operator instruction, since the consequence of ignoring it is a flammable mixture forming where a flame is waiting.
The second and more direct signal is oxygen in the flare header. An oxygen analyzer on the header detects the early presence of air before it can accumulate into an explosive mixture, and a rising oxygen reading is an unambiguous sign that the defenses, including the molecular seal, are being overcome. A header-oxygen high alarm is one of the most safety-critical points in a flare control scheme, because it is the closest thing to a direct warning that flashback conditions are developing inside the pipe. Operators respond by increasing purge, finding the source of the ingress, and treating the situation as an emergency.
In a cloud SCADA platform such as Merobix, these purge-flow and header-oxygen tags from a remote flare are streamed continuously to a hosted system where they are trended, alarmed, and stored. Because flares frequently stand at unmanned sites, having the oxygen and purge signals visible off-site means the warning of a failing seal reaches an operator wherever they are rather than waiting for a site visit, and the alarm can be escalated by notification if it is not acknowledged. The trend history also lets engineers see a slow decline in purge flow or a creeping oxygen reading developing over days, so a seal problem can be addressed before it becomes an ignition event rather than after.
A molecular seal is a dry, passive device below the flare tip that traps a pocket of light gas whose low density resists air ingress. A liquid seal drum sits at the base of the flare and passes gas through a body of liquid, which blocks flashback from travelling back into the header. They protect against ingress and flashback in different ways and are often used together for layered protection.
The molecular seal traps a pocket of light gas, but that pocket slowly degrades through diffusion and thermal effects, so a small continuous purge is needed to keep it refreshed. The seal's benefit is that it lets the required purge rate be much lower than relying on velocity alone. If purge falls below its minimum, the seal can no longer be assumed to be protecting the stack.
The seal cannot be read directly, so the control system watches the conditions that keep it working: purge-gas flow and oxygen in the flare header. A low purge-flow alarm warns that the light-gas pocket may not be maintained, and a rising header-oxygen reading is a direct sign that air is entering where a flame is present. Both are treated as safety-critical alarms.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.