The Thread Sealants and Where Each One Belongs

Why this matters

Almost every tech carries two or three thread sealants and picks between them by habit. The habit works until it meets a service the product was never rated for, and then it fails in one of three ways: it does not seal, it seals and then dissolves, or it becomes a hazard. The last one is not theoretical. Hydrocarbon-bearing compounds in an oxygen system are an ignition source, and a compound that is not listed for potable water has no business in a line somebody drinks from. This card is organized around where each family must not go, because that is the part nobody is taught and the part that generates the callbacks.

On oxygen or oxygen-enriched service, use only a sealant specifically labeled for oxygen service. Ordinary pipe dopes, tapes, and thread lubricants carry hydrocarbons, and hydrocarbons in an oxygen system can ignite from the heat of rapid pressurization with no spark and no flame present. There is no "small amount is fine" version of this rule.

On fuel gas, leak test with an approved leak-detection solution or a combustible gas detector, never a flame, and use a compound listed for the gas you are working with.

What a sealant actually does, and what makes the joint tight

A tapered thread joint does not seal because of the sealant. It seals because the tapered male thread wedges into the tapered female thread and the two thread flanks bear against each other under load. What the sealant does is fill the helical gap that remains at the crest and root of the thread form, and lubricate the flanks so you can reach full engagement without galling.

That is the whole function, and three consequences fall out of it.

• A sealant cannot fix a joint that is not making up properly. Cross-threaded, under-engaged, damaged, or wrong-form threads leak with any product on them. More sealant is the response that turns a leak into a split fitting.
• A sealant changes the torque-to-engagement relationship, because it changes friction. A makeup spec expressed as turns past finger tight assumes a sealant type, and switching products changes where snug arrives.
• A sealant that is wrong for the fluid stops being a sealant. It softens, washes out, or is chemically attacked, and the joint that passed its test leaks months later with nothing having moved.

The families

Tape densities differ, and a heavier tape needs fewer wraps than a standard one to reach the same fill. The common practice for standard-density tape is two to three wraps in the direction the fitting turns; check the tape's own instruction rather than carrying one wrap count across every product.

Where each one must not go

This is the section worth memorizing.

PTFE tape must not go on:

• Flare fittings, compression fittings, or any face or cone seal. The seal there is metal to metal on the cone or the ferrule, and tape holds those faces apart. Where a flare needs lubrication, a trace of the specified lubricant goes on the back of the flare nut, never on the sealing cone.
• Straight-thread ports that seal on an O-ring or a gasket. Nothing on those threads seals anything.
• Anywhere shreds can enter the flow path. Tape wrapped past the first thread, or wrapped against the direction of makeup so it unwinds, sheds material into valves, orifices, and cartridges. Start the wrap a thread or two back from the end.

Pipe joint compound must not go on:

• Oxygen service, unless specifically labeled for it.
• Potable water, unless certified for contact with drinking water. NSF/ANSI 61 is the standard that certification is written against, and plumbing codes generally require it for materials in potable contact. Check the label, not the shelf.
• Plastic threads, unless the compound states plastic compatibility. Some petroleum-based compounds attack certain plastics and cause slow cracking that shows up as a split fitting weeks later rather than a leak on the day.
• Sealing faces of any kind, for the same reason as tape.

Anaerobic thread sealant must not go on:

• Plastic threads. It needs metal to cure, and some formulations craze certain plastics.
• Inert or heavily plated metals without an activator. Stainless and some platings cure slowly or not at all; the product instruction names which need a primer.
• Joints that will be broken in the field with hand tools, if you reached for a high-strength grade. Removing the strong grades generally requires heat, and heat on an assembled system is a separate hazard decision the next person may not be able to make. Say what makes it safe rather than leaving it as a judgment call: the system is depressurized, drained, purged and verified free of flammable contents before any heat is applied. Note also that PTFE tape or residue above roughly 500 degrees F releases decomposition products that cause polymer fume fever, so clean the threads before heating and do not carry taped hands to a break.
• Oxygen service, unless specifically labeled.

Threadlocker is not a pipe sealant, however similar the bottle looks. Locking and sealing are different formulations with different gap-filling and different removal behaviour, and grabbing the wrong one of a matched pair is among the most common mix-ups on a van.

The threads that do not seal on the threads

A large share of sealant misuse comes from not identifying the thread form, so learn to tell them apart by feel before you reach for anything.

• A tapered thread visibly narrows along its length, and a mating fitting hand-starts only a few threads before it begins to snug. Sealing happens on the threads, so the thread is where the sealant goes.
• A straight thread runs at constant diameter and hand-threads freely all the way down to a shoulder, a captive washer, or an O-ring in a boss. The seal is that element, at the face or in the groove. Sealant on those threads adds friction, masks the feel of the seal bottoming out, and gets you either an under-compressed seal that weeps or a crushed one that fails later.
• A dryseal form is a tapered thread designed for the crests and roots to interfere metal to metal. Sealant may still be specified, but the mechanism differs, so follow the fitting manufacturer rather than a general habit.

If you cannot tell which you have, thread the fitting on dry by hand. Freely to a shoulder means straight; snug within a few turns means tapered.

Service conditions that decide for you

Some services take the decision out of your hands, and the label is the only acceptable evidence.

• Oxygen: only a product labeled for oxygen service, as covered above.
• Fuel gas: a compound listed for the gas in question. Gas-rated tape is a distinct product from general-purpose tape, and a rating for one gas is not a rating for another.
• Potable water: certified for potable contact.
• Refrigerant and refrigerant oils: most joints in a refrigerant circuit are flare, brazed, or gasketed rather than thread-sealed, and where a threaded joint does exist, most tapes and compounds are not qualified for the refrigerant and its oil.
• High temperature: every product has a ceiling on its data sheet, and a compound that thins and runs at temperature leaves the joint empty.

The consumable nobody dates

The reactive products lose performance before they look bad. An anaerobic that has thickened, gone stringy, or set in the neck is finished, and a partly aged bottle can cure slowly enough that the joint goes into service before it has developed strength. The date is on the container.

Paste compounds separate and skin over; a tin with a hard crust and oil floating on top is no longer the formulation on the label. Tape does not chemically expire, but a roll that has been loose in a truck bed picks up grit, and grit in a thread is an abrasive during makeup.

The practical control is small containers, an open date written on them, and one place they live rather than five partial bottles spread across a van.

A worked case: the same habit on two thread materials

A shop tracked weeps after an install with a mix of metal and plastic threaded connections, all made by the same two-person crew, all sealed with the same standard-density tape and the same habit of turning past snug until it felt solid.

Ninety days after the install, four of the eleven plastic female threaded joints were weeping. Zero of the nineteen metal-to-metal threaded joints were. That is roughly 36 percent of the plastic joints against none of the metal ones, on the same job, same crew, same product, same day.

Reading it. The variable is not the crew, the tape, or the workmanship, because all three were held constant across both populations. The variable is what the female half is made of and how it responds to the wedging force a tapered thread generates.

A tapered thread makes up by wedging, so the deeper it goes, the higher the hoop stress in the female fitting. Metal absorbs that. A plastic female fitting has far less capacity for it, and it does not fail at the moment of assembly, which is the trap. It takes a stress it can survive briefly, and then relaxes under it over weeks. The joint that felt right on the day opens up quietly.

Piling on more tape makes it worse rather than better. Extra wraps take up radial clearance, so the joint reaches its snug point at less thread engagement while the wedging force at that point is higher. The tech feels a firmer joint and reads it as a better one.

The fix, which was three changes and not one. Fewer wraps on plastic, per the tape's own instruction rather than by feel. A sealant the fitting manufacturer approves for plastic. And makeup by the manufacturer's stated method, which for plastic threaded fittings is usually hand tight plus a specified small amount of wrench, rather than by feel - because feel is calibrated on metal and metal is the material that forgives it.

What would change the reading. If the metal joints had also weeped at a similar rate, the shared variable would be the crew's makeup habit or the product, and the fix would be different. If the plastic joints had failed on the day rather than at ninety days, that points to splitting during assembly rather than slow relaxation, and the fix moves to assembly torque alone. The delay is what identifies the mechanism, so record when a weep appeared, not just that it did.

Checking you got this right

• Identify the thread form before you open any container. Tapered, straight with a separate seal, or dryseal - each takes a different answer, and one of them takes nothing.
• Read the label for the service, not the shelf. Oxygen, potable, and fuel gas are label-verified, not remembered.
• Check the material of both halves of the joint, not just the male half. Plastic on either side changes the product and the makeup method.
• Confirm nothing went on a sealing face. Run a finger over the cone or the ferrule before assembly.
• Confirm the reactive products on the van are in date and that opened containers carry the date they were opened.
• After the joint is in service, note the date any weep appeared. The delay between assembly and leak is the single most diagnostic piece of information about which of these rules was broken.

References

• NSF/ANSI 61, drinking water system components, for the certification required on materials in contact with potable water, as referenced by plumbing codes
• Sealant manufacturer technical data sheets for temperature limits, chemical compatibility, cure requirements, activator needs, and shelf life
• Fitting and equipment manufacturer instructions for makeup method on plastic threaded fittings and for any specified sealant
• Compressed gas industry guidance on cleanliness and hydrocarbon exclusion for oxygen service components
• See related: Why Threaded Connections Leak; Sealant vs Tape vs Gasket: Which Seal to Use; Cross-Threading Prevention