Voids and Porosity in HDPE Welds: Gas, Not Strength

Nov 09, 2026

Leave a message

Every void in a fusion joint is a place where material did not meet material. That sounds like a truism, and it is the most useful thing to know about porosity, because it splits the problem in two at once: either something was in the way, or there was never enough material to close the gap. The first is contamination. The second is fit-up. They look similar on a section and completely different in the cause.

The distinction matters because the standard response to porosity is to clean harder. Cleaning harder fixes one of the causes, does nothing for the other two, and several of the cleaning habits used on site actively make the first one worse.

Four sources, four signatures

Porosity is not one defect with one cause. The four entries below cover almost everything found in a PE joint, and they can be separated by eye once the joint is opened.

What was in the wayWhat the voids look likeWhere they sitWhat it tells you
Water, as a surface filmMany small round cavities, roughly even in sizeA band along the fusion plane, across the whole faceSteam expanded at melt temperature before the faces closed
Air, trapped by the fit-upA few large smooth cavities, often one dominant oneAt the top of the joint, or at the widest part of the gapThe gas had no escape path and the melt closed over it
Volatiles from degraded materialFine, dense foaming, like a spongeA band just below the heated surface, with a dull or brown edgeThe material was damaged before it was joined
Oil, dust or sandFlat unbonded patches rather than cavitiesAt the interface, following the contaminationWetting was prevented; no chain diffusion took place

The last row is not porosity in the strict sense, but it is reported as porosity and it is treated the same way on site. It has a different mechanism and a different fix, so it is worth keeping separate. The wider catalogue of fusion defects, including the ones that have nothing to do with trapped gas, is in HDPE pipe fitting defects and their causes.

Four fusion joint sections showing different void patterns: scattered small round voids across the face, one large cavity at the top, a fine foam band at the surface, and a grey unbonded strip with no cavity
Four things look like porosity and are not. The pattern identifies the source before anything is cut.

Water is a surface problem, not an absorption problem

Polyethylene absorbs almost nothing. Its water uptake is well under a tenth of a percent, which is why PE pipe is used for potable water and why a pipe that has been in a wet trench for a year is not "full of water" in any meaningful sense. The water that causes porosity is a film on the surface, and it does not have to be visible.

At the melt temperature the film does not stay a film. Water expands by more than a thousand times its liquid volume when it turns to steam, and at 200 °C it is expanding into a polymer melt that is viscous, not watery. The steam cannot bubble up through the melt the way it would through a liquid. It pushes the melt front apart, the front closes behind it, and the result is a cavity whose walls are smooth, round and fully fused. That is what a water void looks like: not a tear, a bubble.

Two joint sections: a thin water film between two faced ends before heating, and the same joint after heating with round steam cavities along the fusion plane
A film a few microns thick becomes a bubble at melt temperature. The cavity walls are fully fused.

This is also why the amount of water involved is counter-intuitive. A film a few molecules thick over a 315 mm face is enough to produce a visible band of cavities, because each patch of film produces a disproportionately large volume of gas. Inspectors who expect to find a wet pipe are often looking at a pipe that was dry to the touch.

Where the water comes from on a dry day

Most sites report porosity on days when it did not rain. The water arrives by routes that are easy to overlook.

Pipe end seen face on with five numbered markers pointing to condensation, water in the bore, a contaminated cloth, bare hands and rain, with an explanatory list
Five routes for water on a day it did not rain, on a pipe that is dry to the touch in all five.

Condensation is the first. A pipe is a large thermal mass, and a length that spent the night in a shaded stack or in a trench stays below the dew point well into the morning. Warm humid air moving over it deposits a film that is invisible against a matt black surface. The same thing happens when a pipe is brought out of a cool store into a warm humid day: it sweats, exactly like a cold glass.

The second is the pipe bore. Cutting, wet coring and washing all leave water inside, and water runs down the bore to the end that is about to be faced. The face is machined wet, then heated. The evidence is a concentration of voids near the inner surface rather than across the whole face.

The third is the cleaning process itself, which is covered in its own section below, because most of the water that arrives at the fusion face on a dry day is delivered by the cloth that was meant to remove contamination.

Air: the gap that has nowhere to go

Air is the second source and it is a fit-up problem rather than a cleanliness problem. In butt fusion, a gap between the faced ends can come from a blunt or chipped blade, from insufficient facing feed, from pipe ends that are not square to the axis, or from an oval pipe clamped round. The bead closes over the top of the gap before the melt has filled it, and the air inside has no route out.

In electrofusion the same mechanism is more pronounced, because the whole joint is enclosed. The clearance between pipe and fitting is a designed value, and the scraper depth is what sets it. Scrape too deep and the clearance grows, which means more air to displace and less melt pressure at the interface. Scrape too little and the oxide layer stays. Neither error is visible after the fitting is on. The difference between the two processes, and why the same contamination matters more in one of them, is set out in electrofusion versus butt fusion. Water that reaches the terminals rather than the bore fails in a different way, as an electrical fault, and that case is covered under electrofusion short circuit and smoke.

The signature is worth remembering because it is the one that can be diagnosed without cutting: a small number of large, smooth, round cavities, sitting at the top of the joint. Air rises in a melt. Water does not, because it is generated in place across the whole face.

When the void is a symptom of overheating

The third source is the material itself. Hold the plate too long, or run it too hot, and the surface layer degrades before it is ever joined. Degradation produces low molecular weight volatiles, and those volatiles expand exactly as water does, at the same interface, with the same result.

The difference is in the pattern. Thermal volatiles come off the surface that was in contact with the plate, so the foaming sits in a band immediately below that surface, is finer and denser than water voids, and is usually accompanied by a dull, brown or chalky edge. The smell is a useful confirmation if anyone is still at the machine.

This case cannot be cleaned out. Lowering the plate temperature, shortening the heating time and checking the plate surface temperature at the face rather than at the controller are the corrections, and they are the subject of over-fusion in PE pipe welding.

Contamination that never becomes a gas

Dust, sand and soil are inert. They do not expand and they do not burn off at plate temperature. What they do is sit between two surfaces that need to be in molecular contact, which means that wherever a particle is, there is no bond. A dust film produces a scatter of small unbonded patches and, around each particle, a void nucleus where the melt could not close.

Oil and grease behave differently again. At plate temperature an oil film pyrolyses into a carbonaceous residue, and that residue stays on the face through the joining phase. The result is not round cavities but a dark line at the fusion plane and a bond that is weak across the whole contaminated area. Release agents, cutting fluid and hydraulic oil all produce this, and all of them are transferred by hands and gloves rather than by the pipe.

Skin oil is in the same family and is the most commonly transferred. A prepared surface that has been handled by bare hands looks clean, feels clean and is not clean. That is the entire justification for the glove rule, and it is the rule most often broken when a fitter is checking a dimension.

What voids actually do to the joint

A single round cavity in a ductile material is not a crack. It blunts rather than sharpens, and a joint with scattered small voids will usually pass a bend test, because the material around each cavity deforms and the cavity closes. This is why "we have some porosity" is not on its own a reason to cut a joint out, and why acceptance criteria are written around extent rather than around the presence of voids.

What changes the verdict is connectivity. Voids that touch form a channel. A channel that reaches either surface becomes a leak path, and under pressure it becomes a leak path with a driving force behind it. The same total void volume can be harmless as isolated bubbles and fatal as a connected chain, which is why the count of voids says very little and their distribution says almost everything.

The second effect is on the bonded section. A void at the interface removes bonded area, and the stress that area would have carried moves to its edge. Small voids spread evenly are a minor reduction. A cluster at one point is a stress concentration at the edge of the cluster, and that is where a service failure will start.

A joint that fails with a flat, smooth surface and no cavities at all is not a porosity case, and the distinction is worth making early because the corrective action is completely different. No cavities means no gas, which means no bond was formed in the first place. The two routes to that are an interface that cooled before it was pressed, covered under cold weld in HDPE butt fusion, and two materials that could not bond to each other at all, covered under material mismatch in PE pipe welding.

Finding them

Visual inspection catches the gross cases. Water contamination often announces itself during the weld: a spitting or hissing sound at the joint, small pits on the bead surface where gas escaped through the melt, and occasionally a bead that is porous to the touch. A bead with a blistered or pitted surface on a joint that was otherwise made correctly is a strong indication of a surface film.

The one visual finding that gets mistaken for porosity is a cracked bead. It looks like a defect in the same family and it is not: a crack is a cooling problem, with no gas involved, and it is covered under bead cracking after butt fusion.

Air pockets and internal voids do not show on the outside. Ultrasonic examination can find them, but PE is a difficult material for it: the attenuation is high, the fusion plane is itself a reflector, and the examination needs a procedure written for PE rather than one transferred from steel. The practical methods and their limits are set out under NDT inspection of HDPE welds.

Sectioning settles it. A slice through the joint, polished and examined at low magnification, shows the void band, its position relative to the fusion plane and whether the voids are connected. That is also the only way to tell a water band from a thermal foam band with certainty, which matters because the two have completely different corrections. Where a contract names a joining standard, the acceptance provisions differ between them, and those differences are compared in DVS 2207 and ISO 21307.

Cleaning that works, and cleaning that makes it worse

The procedure is short and almost every step in it can be done wrong in a way that adds contamination rather than removing it.

What works: scrape the full circumference with a dedicated PE scraper, to the depth the procedure specifies, with a sharp blade, and remove the shavings rather than brushing them along the surface. Wipe once, with a clean lint-free cloth and a solvent that evaporates completely and leaves nothing behind; 90 percent isopropyl alcohol is the usual choice. Let it flash off rather than wiping it dry. Do not touch the surface afterwards. Cap or wrap the end if the weld is delayed beyond the interval the procedure allows, and re-scrape if that interval is exceeded.

StepDone properlyDone wrong, and what it adds
ScrapeSharp blade, full circumference, to the specified depthBlunt blade or skipped area: oxide layer remains, and shavings are pushed along the face
WipeClean lint-free cloth, fresh solvent, one passCloth from a pocket, or wiping dry with a rag: dust, oil and fibres go onto the face
EvaporateLet the solvent flash off completelyWelding while the surface is still wet: the solvent itself becomes the contamination
HandleGloves on, no contact with the prepared faceBare hands checking a dimension: skin oil over the whole area touched
TimingWeld inside the interval the procedure allowsRe-wiping instead of re-scraping after a delay: the oxide layer is not removed by solvent
WeatherJoint sheltered, pipe above the dew pointWelding between showers, or waiting for the sun to dry the pipe
FittingsKept in the bag until the moment of assemblyBore left open on the ground: the fitting collects what the pipe collects

Two of these are worth spelling out because they are not obvious. A container of solvent left open in humid air absorbs water until the concentration is no longer what the label says, so a cloth soaked from it delivers water to the face while smelling correctly of alcohol. And a rag used to dry a surface leaves fibres in it; those fibres are not removed by the next wipe, and each one is a point where the melt cannot close.

For electrofusion there is one more, and it is the one most often lost to housekeeping: the fitting bore is a prepared surface too. The handling rules that apply to the pipe apply equally to the fitting, and they are set out in daily care of electrofusion fittings.

Four step process for preparing a fusion face, scrape, wipe, evaporate and weld, each with the failure mode that adds contamination instead of removing it
Every step has one way to put back more than it takes off.

Checklist

  • Check the pipe surface temperature against the dew point before facing. If the pipe is colder than the dew point, it is wet whether or not it looks wet.
  • Drain and dry the bore before facing. Water runs to the end that is about to be machined.
  • Scrape to the specified depth with a sharp blade, and discard the shavings away from the joint.
  • Wipe once with a clean lint-free cloth and fresh solvent, then let it evaporate. Do not wipe dry.
  • Gloves on from the moment the surface is prepared. No bare-hand contact with the fusion face.
  • Weld within the interval the procedure allows. If it is exceeded, re-scrape rather than re-wipe.
  • Shelter the joint from rain, fog and heavy dew. Wet weather is covered in weather conditions for PE pipe welding.
  • Listen during the weld. A hiss or a spit is information, not background noise.

None of this is difficult, and none of it is expensive. The reason porosity persists is that every step depends on a person doing it in the same order on the tenth joint of the day as on the first, which is the subject of welder qualification and operator training. For the sequence around the joint itself, including how the pipe is supported while it is prepared, see the construction sequence.

Frequently asked questions

Does polyethylene absorb water?

Almost none, well under a tenth of a percent. The water that causes porosity is a film on the surface or inside the gap, not water held in the material. This is why drying a pipe in storage achieves nothing and why a dry-looking pipe can still produce a wet joint.

Why do water voids form a band across the whole face?

Because the film is spread across the whole face. Each patch of film turns to steam where it sits, so the cavities appear along the entire interface rather than in one place. Air behaves the opposite way: it collects and rises, which is why trapped air produces one or two large cavities at the top.

Can I weld PE pipe in light rain?

Not without shelter over the joint. The rain itself is less of a problem than the film it leaves and the humidity it maintains. A tent that covers the machine but not the joint does not solve it, and a joint welded in a break between showers is being welded onto a wet surface.

Will a few small bubbles cause a leak?

Not on their own. Isolated round voids blunt rather than propagate, and they usually survive a bend test. The risk is connectivity: voids that touch form a channel, and a channel that reaches a surface leaks under pressure. Distribution matters more than count.

Can higher pressure squeeze the voids out?

No. Pressure moves melt, not gas, and the gas has to go somewhere. Raising the pressure to close a void usually displaces material along the path of least resistance and can make the joint worse by thinning the bonded section at the edge. The fix is to stop the gas getting in.

Why is porosity more common in electrofusion than in butt fusion?

Because the joint is enclosed. In butt fusion the interface is open at the bead, so some gas escapes. In electrofusion the melt is contained by the fitting, the clearance between pipe and fitting holds a fixed volume of air, and the heating wire sits in the middle of it. Anything on either surface ends up inside the joint rather than at its edge.

Send Inquiry