The valve came out of the line looking better than most valves that come out of a hydrochloric acid line. Three inch, PFA lined, cast body, Class 150, fourteen months in service on a 32 percent acid transfer duty at 55 to 60 degrees C, about forty open and close cycles a week. The paint was intact. The bore was smooth and black, with the dull sheen a PFA liner keeps for years. It passed a holiday test at twelve kilovolts end to end. And yet, forty millimetres inside the upstream flange, a blister the size of a thumbnail had lifted the liner off the body wall, and underneath it the casting had lost 2.4 millimetres out of 9.5.
Nothing had flowed through that loss. The acid never reached it from the bore, because the liner was continuous and the spark test proved the liner was continuous. The acid arrived from behind, one molecule at a time, through three millimetres of fluoropolymer that is completely inert towards it and still, at 60 degrees C, permeable to it. When it met the steel it did what acid does to steel, and because the reaction produced hydrogen and the hydrogen had nowhere to go, it pushed the liner off the wall from the inside.
That is the whole design problem of a lined valve, and it is the one thing no compatibility chart will tell you. A lined valve is not a valve made of plastic. It is a steel pressure vessel with a plastic membrane inside it, and the membrane has a diffusion rate, a stiffness that falls sharply with temperature, and a coefficient of expansion roughly ten times that of the metal it is bonded to. Get any one of the three wrong at the specification stage and the valve fails in a way that a visual inspection of the wetted surface cannot find.
The sections that follow take the three properties in turn, then the design details that compensate for them, then the duties where a lined valve is simply the wrong answer.
1. Three ways to build a valve that acid cannot eat
Before anything else it is worth being clear which of three quite different products is being bought, because purchasing departments routinely compare quotations that are not for the same object. A solid alloy valve carries both the pressure and the corrosion on the same wrought or cast material. A lined valve splits the duty: ordinary steel carries the pressure, a moulded fluoropolymer carries the corrosion. A solid thermoplastic valve does both in plastic, and is limited by the creep of the plastic rather than by the metal.
Route | What contains the pressure | What resists the acid | Where it fails first | Cost index |
|---|---|---|---|---|
Solid alloy body, C276 or titanium or zirconium | The alloy itself | The same alloy, full section | Crevice corrosion in seat pockets; price and lead time before anything else | 8 to 15 |
Lined steel body, PTFE or PFA | Carbon steel or ductile iron casting | A 2.5 to 4 mm moulded liner | Permeation behind the liner, then cold flow at the flange face | 1.0 |
Solid thermoplastic body, PP or PVDF | The plastic wall, in creep | The same plastic | Temperature and sustained load; thread and flange joints pull out | 0.6 to 0.9 |
The lined route wins on cost by a wide margin and loses on exactly one thing: the interface between the liner and the steel is a place where chemistry can happen unseen. Everything else in this article is a consequence of that interface.
2. How the liner gets into the body, and why the fit matters
There are three production routes and they behave differently in service. The oldest is a loose liner, where a sintered PTFE sleeve is pressed into a machined cavity with a deliberate interference of a few tenths of a millimetre and held by the flange turn-over and the body joint. The second is an adhesive bonded liner, where an etched PTFE or a PFA surface is bonded to a grit blasted and primed body wall. The third, and the one that now dominates higher-quality supply, is direct moulding: PFA or modified PTFE is transfer moulded or rotolined against the heated body so that the liner is formed in place, fills every groove, and shrinks onto the wall as it cools.
The dimensional point that decides field behaviour is not thickness but contact. The liner must stay pressed against the steel across the whole temperature range, because the moment it lifts, the gap becomes a still volume of process fluid and a crevice cell. The problem is that PTFE expands about 12 to 13 parts in 100,000 per kelvin and steel about 1.2. Constrained in a steel cavity, that mismatch turns into compressive stress rather than growth, and the stress is not small: for 120 kelvin of rise the free expansion difference is about 1.3 percent, and with a constrained modulus near 0.5 gigapascal at 100 degrees C the liner sees roughly 6.6 megapascals of compression. The compressive strength of PTFE at that temperature is in the same band, five to eight megapascals. The liner is therefore yielding, slowly, whenever a lined valve sits hot, and it is yielding in the direction that thins it at the flange and thickens it into the bore.
This is the real reason a lined valve is not rated for the full temperature of its resin. Virgin PTFE is sold as a 260 degrees C material, and no responsible liner supplier will rate a bonded or loose liner above about 150 degrees C, not because the polymer degrades but because the assembly creeps. Anchor grooves, a machined dovetail pattern, a moulded-in flange lip and a body joint that traps the liner edge are all attempts to hold a material that is trying to move.
3. Permeation: the column the datasheet does not have
Fluoropolymers are barrier materials against ions and terrible barriers against small molecules. A chloride ion, hydrated and bulky, does not travel through a PTFE lattice at any useful rate, which is why the liner genuinely survives 32 percent hydrochloric acid for years. Hydrogen chloride as an undissociated molecule, water, chlorine, hydrogen fluoride and oxygen are another matter: they dissolve into the polymer, diffuse along the amorphous regions, and leave on the other side. The flux is proportional to the solubility, the diffusion coefficient and the applied activity, and inversely proportional to thickness, so a change from 3 millimetres to 2.5 millimetres is a twenty percent increase in the amount of acid reaching the steel.
Diffusion is thermally activated, and for these polymers the permeability roughly doubles for every 10 to 15 kelvin. A valve on a line at 30 degrees C and the same valve on a line at 60 degrees C are not the same valve with a warmer label; the second one is passing two to eight times as much acid through its liner, and it will show a blister in a fifth of the time.
Liner material | Relative permeation, water and small acid molecules | Practical limit in a lined valve | Vacuum capability | Behaviour to expect |
|---|---|---|---|---|
PTFE, virgin sintered | 100 | 150 degrees C | To about minus 0.5 bar | Highest permeation, highest creep, cheapest |
Modified PTFE | 80 | 160 degrees C | To about minus 0.5 bar | Less creep, better weldability, similar permeation |
PFA | 45 | 200 degrees C | To about minus 0.9 bar | Best high temperature choice; higher cost, fewer moulders |
FEP | 60 | 150 degrees C | To about minus 0.7 bar | Softer, easy to score during assembly |
ETFE | 15 | 120 degrees C | To about minus 0.5 bar | Better abrasion resistance, lower temperature |
PVDF | 3 | 120 degrees C | To about minus 0.6 bar | Rigid and low permeation; swells in ketones and amines |
PP | 2 | 90 degrees C | To about minus 0.3 bar | Cheapest; oxidising acids attack it outright |
The numbers in the second column are order-of-magnitude ratios for the same thickness, not design values, and they are given because relative ranking is what a specification needs. The absolute ranking is stable across the literature even where the constants are not: PTFE is the most permeable of the common liners by a factor of two over PFA and by more than an order of magnitude over PVDF.
4. The gas that lifts the liner
What happens behind the liner depends on the acid. Hydrochloric acid attacks carbon steel to give ferrous chloride and hydrogen: one mole of iron gives one mole of hydrogen, and one mole of gas at room temperature occupies about 24.5 litres. Half a gram of steel consumed behind a liner therefore produces roughly 220 cubic centimetres of hydrogen, and if that gas is generated into a void of 20 cubic centimetres between liner and wall, the void reaches about eleven bar. Eleven bar acting on three millimetres of PFA over even a small area is enough to deform it permanently, and the deformation is what the next inspection finds.
Hydrofluoric acid is worse in a different way, because it permeates readily and the steel behind it goes into rapid general loss rather than into pitting, so the wall thins over a broad patch and the liner blisters across a wide area rather than at a point. Sulphuric acid above about 70 percent behaves the opposite way: it is a strong dehydrating agent, it carbonises and attacks many elastomers, and it permeates more slowly than hydrochloric, but any water dilution at the liner interface produces a locally dilute and far more aggressive acid.
Sodium hydroxide is the case that surprises people, because there is no gas at all. Caustic at 50 percent and 80 degrees C behind a liner does not generate hydrogen; it concentrates. Water evaporates through the liner faster than the salt does, so the film behind the liner becomes stronger as time passes, and carbon steel in concentrated caustic above about 60 degrees C is a classic caustic stress corrosion cracking couple. The failure is a crack through the body wall, not a leak at a seat, and it looks nothing like corrosion.
5. The fix is a hole, and it is usually missing
Because the gas and the concentrate have to go somewhere, a correctly designed lined valve gives them a route to atmosphere. The feature is a set of small drilled vents, typically 3 to 6 millimetres, through the body wall from the annulus behind the liner, and a vented body cavity, so that whatever passes the liner reaches the outside instead of accumulating. In a good design they are at the lowest point of each annulus, they are not tapped and plugged, and the liner has shallow grooves on its back face that lead the permeate to them.
The consequences of leaving them out are exactly the failure at the top of this article. The consequences of having them are a maintenance conversation that has to be had once: those holes weep, and they are supposed to. A weeping vent hole is the design working, and the correct response is to log the rate, because a rise in that rate is the only early warning a lined valve gives. Painting over the vents, plugging them to stop the drip, or blanking them during transport and forgetting to remove the blank are three of the most common ways a lined valve is quietly converted into an unvented pressure vessel.
The second half of the fix is the body cavity. A lined ball valve has an annular space between the two seats, and on a lined design that space is also lined. If the cavity has no relief, permeate from the upstream side accumulates there and eventually finds the body joint, which is why lined valves are normally supplied with a cavity vent or a drain plug at the bottom of the body. It should be open, or it should be a documented manual drain point, and it should not be a plugged hole nobody has looked at since commissioning.
6. Vacuum, blowdown, and the collapse
Permeation sets up a differential in one direction; draining the line sets it up in the other, and that is when liners collapse. When a line is pumped down or drained for maintenance, the bore goes to vacuum while the annulus behind the liner is still at whatever pressure has accumulated there. One bar across a 200 millimetre port is 3.1 kilonewtons, and that load sits on a membrane a few millimetres thick with a modulus of a few hundred megapascals at temperature. The liner does not tear. It folds, and the fold stays.
The field symptom is memorable once seen: a valve that passed its seat test at the last shutdown is found, on restart, to be passing almost nothing with the ball fully open, and the bore has a soft crumpled lining that closes it down. The same damage occurs more slowly on a line with a reciprocating pump and poor damping, where the pressure at the valve oscillates around atmospheric, and on any line where the valve is used as a block valve upstream of a vessel that is evacuated.
The mitigation is in the specification, not in the field: if the duty includes vacuum, the liner has to be PFA rather than plain PTFE, the venting has to be positive, and the body should carry a vacuum rating in writing. A supplier who will not put a vacuum figure on the datasheet is telling you the liner is loose, and a loose liner in a vacuum line is a liner waiting to fold.
7. Temperature and pressure: why a lined Class 150 is not a Class 150
The body is a Class 150 casting and the flange bolts are Class 150 bolts, but the assembly is not a 19.6 bar valve at temperature, and it is worth stating the reason plainly because it is a frequent argument at the witness point. The pressure boundary that fails first is the liner, and the liner fails by creep and cold flow, both of which accelerate with temperature. The seat load that keeps the valve tight is generated by a fluoropolymer seat ring pressing on a lined ball, and that load relaxes as the material softens. The result is a derating curve that is far steeper than the metal one.
Service temperature | Typical maximum differential for a PTFE lined Class 150 valve | Same for PFA | What limits it |
|---|---|---|---|
20 degrees C | 16 bar | 16 bar | The body casting, not the liner |
80 degrees C | 12 bar | 14 bar | Seat relaxation begins |
120 degrees C | 8 bar | 10 bar | Liner creep at the flange, seat set |
150 degrees C | 5 bar | 8 bar | Compressive yield of the liner |
180 degrees C | Not recommended | 5 bar | Only PFA, only with a metal seat backup |
These are working figures for a well-made valve and they will move by a bar or two between suppliers, which is itself the point: the derating has to come from the supplier's own test data and it has to be in the quotation, not inferred from the body class.
8. The ball, the stem, and the bolts that get wet
Three components in a lined valve are usually the first to disappoint, and none of them is the body. The ball is the first, because a lined ball is a steel ball with a moulded cover, and at the port edge the cover is at its thinnest and at its most highly stressed. That edge is what wipes the seat every time the valve cycles, and it is what chips when the valve is closed on a crystal, a gasket fragment or a piece of scale. A chipped port edge is not a leak on the day it happens; it is a leak six months later when the seat has worn into the chip.
The stem is the second. Some designs line the stem, some sleeve it, and some use a solid alloy stem through a lined body, and the three have very different sealing behaviour. A lined stem has a liner seam running along it, and the seam passes through the packing every stroke. A sleeved stem has a joint at the base of the sleeve. A solid alloy stem is the most reliable and the most expensive, and it is the right answer on any duty involving dry chlorine, bromine or hot hypochlorite, where a stem failure is an release rather than a leak.
The third is the body joint. In a through-bolted lined valve the bolts pass through the wetted cavity, which means the bolts are in the process and are usually the least corrosion resistant item in the bill of materials. Carbon steel bolts in a lined hydrochloric acid valve are a two-year item at best; the correct callout is a C276 or a fully encapsulated bolt, and it belongs in the specification rather than in the after-sales conversation.
9. The flange face is part of the liner
On a lined valve the liner is turned over the flange face and becomes the gasket seating surface, which makes the flange joint the most highly loaded part of the liner in the whole valve and the part most often destroyed at installation. The liner there is 2.5 to 3 millimetres thick and it is carrying the full bolt load of a Class 150 joint through a material that cold flows at a few megapascals. Two consequences follow.
The first is that the gasket has to be chosen for the liner, not for the pipe. A full face soft gasket, a PTFE envelope gasket or a virgin PTFE sheet gasket are all reasonable. A spiral wound gasket with a graphite filler, which is the default in most pipe stores, is wrong here: its seating stress is far above what the liner can take, and its inner ring does not protect the liner edge from the process. The second is that torque has to be controlled and re-checked. The correct sequence is to torque to the lining manufacturer's figure rather than to the flange table, to re-torque after the first thermal cycle because the liner will have relaxed, and to accept that a lined joint which has been tightened twice to stop a weep has probably already been damaged.
There is a related installation fault worth mentioning because it is invisible afterwards: a liner face that has been nipped by a misaligned pipe flange, or cut by a raised face that was rough or had a weld spatter on it. The cut is a few tenths of a millimetre deep, it does not show on a spark test because the test is applied from the bore, and it is a permanent path from the process to the bolt circle.
10. Two things the liner cannot do
The first is electrical. PTFE has a surface resistivity above 10 to the eighteenth ohm centimetre, so a lined valve is an insulator in the middle of a conductive pipe, and any flammable solvent flowing through it will accumulate charge on the liner and on the ball. A bonded, non-conductive liner in a solvent transfer line is a static ignition source with a design life, and the fix is a conductive liner, carbon black modified PFA or PTFE, with a resistance to earth specified in the purchase order. Earthing straps across the flanges do not solve it, because the charge is on the wetted surface, not on the pipe.
The second is fire. A lined valve is not a fire safe valve in the sense of API 607, and it will not be certified as one, because the only seal in the valve is a fluoropolymer that melts and decomposes. Above about 400 degrees C PTFE decomposes with the release of hydrogen fluoride, which means a lined valve in a fire is not only a failed seal but a hydrogen fluoride source in the middle of the incident. On a duty where a fire case has to be closed out, the lined valve should not be the primary isolation, or the isolation should be duplicated with a metal seated valve upstream.
11. Four shop tests that find the defects, and one that does not
A lined valve can be checked properly in a workshop in under an hour, and the checks are worth writing into the inspection and test plan because they are the only ones that address the failure modes above. A high voltage holiday test at 10 to 15 kilovolts over the whole wetted surface finds pinholes and thin spots. An ultrasonic thickness survey of the liner at the flange faces, the ball port and the cavity floor finds the thin areas that moulding produces. A visual and borescope inspection of the port edge finds chips before they become leaks. A shell test at 1.5 times the class, performed before the liner is fitted or at a reduced pressure after, confirms the casting.
The test that does not find the defect is the one everybody does: a hydrostatic seat test at ambient on a new valve. It passes on the day the valve is made, and it says nothing about permeation, nothing about creep at 120 degrees C, and nothing about whether the vent holes were drilled. The vent holes are checked with a piece of wire and a light, and that check takes ten seconds, which is roughly the cost of not doing it.
12. Writing the purchase order so it can be met
Most lined valve problems are created in a two-line enquiry that says "lined ball valve, 3 inch, Class 150, PTFE lined, HCl service". That sentence leaves the liner type, the liner thickness, the venting, the ball construction, the bolt material, the flange facing, the vacuum rating and the derating curve all open, and the supplier will fill every one of them at the lowest cost that satisfies the sentence. A specification that closes those gaps is not longer; it is about nine lines.
Name the liner resin and not just the family, because PTFE and PFA are not interchangeable above 150 degrees C. State the thickness at the flange and at the bore separately. Require vent holes and a vented cavity, and require them to be open at despatch. State the maximum differential at the maximum temperature, so the derating is a supplier commitment rather than an inference. Require the holiday test voltage and the report. Name the bolt material and the gasket type. State whether the duty includes vacuum or a pump that can pull the line down. State whether the fluid is flammable, which decides the conductive liner. And state the fire case, so nobody assumes API 607 applies.
Getting those nine lines right costs an hour. The valve at the top of this article cost a shutdown, a replacement and an investigation, and the investigation took three weeks because the liner was perfect and nobody could see where the acid had come from.
The works at J-VALVES covers a half inch to 56 inch, ASME classes 150 to 2500, a cold limit of minus 196 and a hot limit of plus 650 degrees C, with 1.5 times the class on the shell and 1.1 times on the seat as standard routine; lined bodies move through that range on a slower line, because every liner is moulded, spark tested and pressure checked before it is fitted. Hand over the line data, the tag list and the maximum differential at temperature, and the priced reply is out in three working hours.
13. Frequently Asked Questions
Is a PTFE lined ball valve suitable for hydrochloric acid?
Yes, and it is the standard economic answer for 30 to 37 percent acid at moderate temperature, provided the liner is vented and the duty stays inside the derating curve. The liner is inert to the acid; the risk is that hydrogen chloride permeates the liner and attacks the steel body from behind, generating hydrogen that blisters the liner. A vented body and a documented permeation route turn that from a failure into a maintenance item.
Why did the liner blister when the spark test passed?
Because the two tests look at different sides of the liner. A holiday test proves the bore surface has no pinhole. A blister is caused by acid that crossed the liner molecularly and reacted with the steel behind it, so the bore side can be perfectly sound while the back side is full of gas at several bar. Once the liner has lifted, the void behind it is a corrosion cell that grows until the body thins through.
Can a lined ball valve be used on vacuum service?
Only if the supplier states a vacuum rating in writing, and only with PFA or a bonded liner. A loose PTFE liner with no bonding and no venting will collapse into the bore when the line is pumped down, because the annulus behind it is at a higher pressure than the bore. Typical catalogue limits are minus 0.5 bar for PTFE, minus 0.7 for FEP and minus 0.9 for PFA.
What temperature can a PFA lined valve take?
The resin is rated to 260 degrees C, but the valve is not. A PFA lined valve is normally derated above 150 degrees C because the liner expands about ten times as much as the steel body, goes into compression, and creeps thinner at the flange. At 180 degrees C the practical differential is around 5 bar, and above that the seat load relaxes faster than the elastomer can recover.
Is a lined ball valve fire safe?
No. A lined valve has no secondary metal seat and no graphite seal, so it cannot pass an API 607 or ISO 10497 test, and it should not be presented as fire safe in a datasheet. Above roughly 400 degrees C the fluoropolymer decomposes and releases hydrogen fluoride, so on a duty with a credible fire case the lined valve should not be the only isolation.
