Views: 0 Author: J-VALVES Publish Time: 2026-11-19 Origin: Site
Chlorine is the service where a valve catalogue number tells you almost nothing. Two valves can look identical, carry the same DN and Class, and be ordered on the same day: one is a carbon steel body with a Monel ball and filled PTFE seats, and it is the correct choice; the other is the same casting with a 316 trim, and it is a toxic release waiting for a humid afternoon. The difference between them is not a preference or a price tier. It is a single number - the water content of the chlorine, expressed in parts per million - and almost every other decision in the specification follows from it.
The second thing that makes chlorine different is that the fluid sits at its own vapour pressure at ambient temperature. A line that nobody would describe as "high pressure" is an 11 bar line on a warm day, and a liquid-full section between two closed valves can be driven to several hundred bar by nothing more than afternoon sun. Chlorine valves therefore fail in two completely different ways: slowly, through a packing or gasket leak that creates a toxic cloud, and violently, through a cavity that was never given a relief path. This article deals with both, and converts the physics into clauses that can be written into an enquiry.
The Chlorine Institute defines dry chlorine as liquid or gaseous chlorine containing no more than 150 ppm of water by weight, and wet chlorine as anything above that. The definition is about moisture, not about phase: chlorine can be dry gas, dry liquid, wet gas or wet liquid, and the four combinations behave very differently towards metal. The reason a 150 ppm threshold carries so much weight is a passivation film. On carbon steel in dry chlorine, a ferric chloride layer forms and holds the surface in a passive state, and the corrosion rate becomes negligible. Water breaks that film: chlorine hydrolyses to hydrochloric acid and hypochlorous acid, the film is undermined, and the wall begins to lose metal at a rate measured in millimetres per year rather than micrometres. Published rates for carbon steel in wet chlorine run to several millimetres per year, which puts a 6 mm wall at roughly twelve months to perforation - the reason the wet side of a chlor-alkali plant is a different materials world from the dry side.
The acid arithmetic explains why such a small moisture figure matters. One mole of water produces one mole of HCl and one mole of HOCl, so each gram of water carries about 4.94 g of acid potential. At the 150 ppm convention that is 0.74 g of acid per kilogram of chlorine - a small mass, but it is delivered as a liquid film on the wetted surface rather than diluted in the bulk, so the attack is local and concentrated at exactly the places a valve cannot afford to lose: seat bands, stem guides and gasket faces.
Characteristic | Dry chlorine (150 ppm water or less) | Wet chlorine (above 150 ppm water) |
|---|---|---|
Mechanism on steel | Ferric chloride film passivates the surface; attack effectively stops | Hydrolysis to HCl and HOCl destroys the film; rapid general and pitting attack |
Default pressure-boundary material | Carbon steel (A106 Gr B pipe, A216 WCB or A105 valve bodies) | Hastelloy C-276 or C22, titanium (wet service only), tantalum, PTFE or PFA lining, FRP and PVC at low pressure |
Stainless steel 304/316 | Resistant below about 121 deg C, but one moisture excursion invites chloride stress corrosion cracking | Not acceptable; pitting and chloride SCC |
Copper alloys | Attacked; not permitted | Attacked; not permitted |
Titanium | Prohibited - reacts exothermically and can sustain a metal fire | Excellent, because the oxide film needs moisture to stay stable |
Where it appears in a plant | Chlorine storage and vaporization, dried gas headers, tank car and cylinder unloading, liquid chlorine transfer | Cell room gas cooling and drying train inlet, hypochlorite and bleach, scrubber liquor, chlorination injection lines |
Two operational consequences are easy to miss in a purchase order. First, dryness is a property of containment, not of the product. Commercial chlorine is shipped dry, and the bulk assay will still read dry after a valve changeout performed in humid air, because the moisture that matters is the film that formed on the seat while the valve was open to atmosphere. Second, the purge specification and the material limit have to be read together: a dew point of -40 deg C, the usual requirement for purge air or nitrogen before breaking a dry-chlorine joint, corresponds to roughly 127 ppmv of water over ice or about 189 ppmv over supercooled water - one value on each side of the 150 ppm convention. That is why a plant can be inside its moisture specification by bulk analysis and still create a local wet cell on a gasket face.
Liquid chlorine boils at about -34 deg C at atmospheric pressure and is stored and transferred under its own vapour pressure, which rises steeply with temperature. This is the first thing a valve specification has to absorb: the design pressure of a chlorine line is set by the hottest day the line will ever see, not by the pump or the compressor.
Temperature | Vapour pressure (bar absolute) | Liquid density (kg/L) | What it means for the valve |
|---|---|---|---|
-34 deg C | 1.0 | 1.56 | Atmospheric boiling point; any uninsulated liquid line can reach this on depressurization |
0 deg C | 3.7 | 1.47 | Cold-climate ambient; still above the set pressure of many low-pressure relief devices |
10 deg C | 5.0 | 1.44 | Cooling water temperature; the "normal" operating point in temperate plants |
20 deg C | 6.7 | 1.41 | Indoor ambient; the value most datasheets quote, and the one most often treated as the design case |
30 deg C | 8.7 | 1.38 | Warm indoor plant; Class 150 still comfortable |
40 deg C | 11.3 | 1.35 | Outdoor line in summer; a "low pressure" system is now an 11 bar system |
50 deg C | 13.9 | 1.31 | Solar gain on an uninsulated or white-painted line; margin to Class 150 is down to about 1.4 |
60 deg C | 17.8 | 1.28 | Beyond the practical limit for most chlorine equipment; also approaching material temperature ceilings |
The rating consequence is narrow but real. A Class 150 carbon steel valve at 50 deg C is rated about 19 bar, so 13.9 bar of chlorine vapour pressure still fits - but the margin is 1.4, not the 3 that a buyer assumes when he sees "150 lb" and "7 bar" in the same sentence. Add a relief valve set at 1.1 times the maximum operating pressure and the two numbers converge. This is one of the reasons chlorine practice tends to specify Class 300 for liquid service even where the pressure calculation would allow Class 150, and why the temperature ceiling, not the pressure class, is usually the governing limit.
The phase change matters as much as the pressure. One volume of liquid chlorine vaporizes to roughly 460 volumes of gas at ambient conditions, so a one litre cavity - a valve body, a short spool, a gauge root - contains enough liquid to fill 460 litres of space with gas. Released into a 100 cubic metre room, that single litre gives a concentration around 4,600 ppm, which is more than four hundred times the NIOSH immediately dangerous to life or health value of 10 ppm. Two design rules come straight out of that arithmetic: never bleed a liquid chlorine valve cavity into an occupied space, and size any vent or scrubber connection by the gas volume that a liquid release generates, not by the mass in the cavity.
Trapped liquid is the violent failure mode, and it is arithmetic rather than opinion. Liquid chlorine has a high volumetric expansion coefficient and a low compressibility; when a fixed volume is completely full of liquid and the temperature rises, the liquid cannot expand, so the pressure rises instead. The classic demonstration is the liquid-full cylinder, and the published numbers are large enough to be worth reproducing because they set the scale for what a valve cavity can do.
Temperature (liquid-full from 0 deg C) | Saturation pressure (MPa) | Hydraulic pressure increment (MPa) | Total pressure (MPa) |
|---|---|---|---|
0 deg C | 0.27 | 0 | 0.27 |
5 deg C | 0.33 | 7.09 | 7.42 |
10 deg C | 0.40 | 13.88 | 14.28 |
15 deg C | 0.48 | 20.35 | 20.83 |
20 deg C | 0.56 | 26.53 | 27.09 |
30 deg C | 0.77 | 37.40 | 38.17 |
40 deg C | 1.03 | 46.58 | 47.61 |
Read the increments rather than the totals. Between 0 and 20 deg C the hydraulic term grows at about 1.33 MPa per kelvin, and between 20 and 30 deg C at about 1.25 MPa per kelvin. In round terms, a liquid-full, rigid chlorine volume gains 12 to 14 bar for every degree of warming, which is an order of magnitude more than water (about 4.6 bar/K) and two orders more than any gas in a fixed volume. The pressure rise is independent of volume - a 50 ml gauge pocket behaves like a 50 litre spool - but the stored energy is not, and it is the energy that decides whether the failure is a weep or a fragmentation event.
Now apply that to a valve. A closed ball valve with liquid on both sides has a body cavity that is liquid-full by definition. Its seats are designed to seal, so the cavity is isolated from both the upstream and downstream line. Thirty kelvin of solar gain takes that cavity to roughly 390 bar - about 1.5 times the ambient rating of a Class 1500 valve and within 10 percent of Class 2500. No practical wall thickness or bolting arrangement survives that with margin, and the part that fails is the body joint, the body cap or the flange, not the pipe. This is precisely why the Chlorine Institute's dry chlorine piping pamphlet requires every liquid chlorine ball valve to have a means of relieving excess pressure from the body cavity, and it is the single most frequently omitted clause in a chlorine valve enquiry.
Two methods are approved, and they are not interchangeable in terms of behaviour:
Relief method | How it works | Consequence for the valve | Where it goes wrong |
|---|---|---|---|
Upstream vent hole (vented ball or body vent) | A hole connects the body cavity to the upstream throat in the closed position, or a body port permanently bypasses the upstream seat | Positive relief independent of seat flexure; the valve becomes unidirectional and must be installed with the arrow upstream | Installed backwards, the valve neither seals nor relieves; a vented ball fitted in the wrong orientation turns a safety feature into a leak path |
Self-relieving seats | The downstream seat always seals; the upstream seat flexes and relieves into the upstream line once cavity pressure exceeds a threshold | Bidirectional shut-off is retained; relief depends on controlled, repeatable seat flexure and on the seat material | Seat swell, ferric chloride deposits or a cold set can lock the upstream seat shut, leaving the cavity with no relief and no visible sign until the body splits |
Both methods fail in the same direction when they are applied to the wrong phase. A vented ball is required for liquid service, and for gas service wherever the gas can condense - which includes any gas line whose temperature can fall below the saturation temperature at its pressure, a real condition on outdoor lines in winter and on lines that are depressurized. A valve prepared for liquid duty and installed on a gas line is not dangerous in the same way, but it is unidirectional and will leak through the vent in the reverse direction, which is discovered only at the seat leak test or, worse, at the first attempt to isolate. The cast arrow on the body is therefore a safety device, not a flow decoration, and the installation drawing should show it.
The same physics explains why the cavity relief has to be treated as a pressure-relief device in its own right during commissioning. The relief path is small - a 3 mm hole is typical - and it can be blocked by ferric chloride scale, by frozen moisture or by a seat that has taken a compression set. A functional check that proves the path is open, performed before the line is put into chlorine service and repeated at each maintenance interval, is worth more than any certificate.
Chlorine materials selection is often presented as a ladder from carbon steel up to Hastelloy, with the implication that the top of the ladder is the safe choice. It is not a ladder. The correct material depends on moisture, on phase, on temperature and on whether the surface is a pressure boundary or a high-velocity wetted part, and the same alloy can be the best and the worst choice in the same plant.
Material | Dry chlorine | Wet chlorine | Notes for valve specification |
|---|---|---|---|
Carbon steel (A216 WCB, A105, A106 Gr B) | Standard; resistant below about 121 deg C, service classes to 149 deg C | Rapid attack; not acceptable | The default body material for dry service. Corrosion rate rises sharply above 121 deg C; the 149 deg C class limit is a code limit, not a comfortable operating target |
Ductile iron (A395) | Acceptable for Class I gas service valves and strainers only | Not acceptable | Cast iron and malleable iron fittings and general-purpose valves are not recommended for chlorine at all |
304 / 316 stainless steel | Resistant in the absence of moisture, but one wet excursion can initiate chloride SCC | Not acceptable | The most common wrong answer in chlorine procurement, because "stainless" reads as "corrosion resistant" to a buyer who has not been told about SCC |
Alloy 20 | Used for bodies and trim where higher strength than nickel alloys is needed | Limited; check chloride and acid conditions | Common in chlorine valve catalogues as an intermediate option |
Monel 400 | Good; the traditional ball, stem and disc material | Not recommended | Nickel-copper alloy; the copper content that helps in dry chlorine is what limits it in wet chlorine |
Hastelloy C-276 / C22 | Excellent | Excellent | The only common alloy that is a genuine answer on both sides of the 150 ppm line; C22 is often the wet-side default |
Titanium | Prohibited - exothermic reaction, metal fire | Excellent, provided the chlorine is genuinely wet | The inversion is absolute and has caused real releases; titanium must never be used where the chlorine can dry out |
Tantalum | Inert to about 149 deg C | Inert to about 149 deg C | Used for diaphragms and instrument wetted parts; expensive and easily mis-identified by appearance |
Copper alloys (brass, bronze, gunmetal) | Attacked | Attacked | Excluded from trim, seats, gauge internals and packing followers |
Aluminium | Prohibited | Not acceptable | Also excluded from nameplates, tags and actuator housings that can contact chlorine |
PTFE, modified PTFE, TFM | Excellent | Excellent | Standard seat and body seal material; filled grades resist the abrasive ferric chloride debris that forms in dry service |
PCTFE | Good | Good | Lower permeability and higher stiffness than PTFE; used where seat extrusion is the limit, at the cost of temperature range |
PVDF, PFA, FEP | Good | Good | Lining and diaphragm materials; PVDF is common in wet chlorine at moderate temperature |
Elastomers (FKM, EPDM, NBR) | Marginal to poor | Poor | Chlorine attacks most elastomers and diffuses through them; rubber-seated valves are a recurring field failure |
Carbon steel also carries two temperature numbers that are often confused. The first is a corrosion threshold: above about 121 deg C the attack rate in dry chlorine rises significantly, which is why 121 deg C is treated as a practical ceiling even though the piping classes extend to 149 deg C. The second is an ignition threshold: iron and steel can ignite in chlorine at around 226 deg C, releasing heat and metal chloride fumes. The gap between 149 and 226 deg C is not a safety margin to be consumed by a badly insulated line or a heat-traced valve; it is the reason chlorine equipment is normally kept well below its class limit, and the reason a chlorine valve should never be the hottest point on a trace-heated run.
The titanium inversion deserves its own paragraph because it is the one materials rule that is genuinely counter-intuitive. Titanium owes its corrosion resistance to a titanium dioxide film that needs moisture to remain stable, so it is excellent in wet chlorine, in chloride brine and in hypochlorite. In dry chlorine the same metal reacts exothermically and can sustain a metal fire that consumes the component and breaches whatever contains it. A documented 2014 release at a chlorine packaging facility is instructive: a gauge with a titanium housing and a tantalum diaphragm was taken from a storeroom and fitted to dry chlorine service because it looked identical to the stainless gauges normally used there, and the resulting metal fire is believed to have initiated the release. The lesson is not "train the fitters harder". It is that a specification which identifies materials by appearance, or by a generic "chlorine service" label, leaves that failure mode available, and that the physical segregation of chlorine-prepared parts from visually identical non-compatible parts is a control measure.
Between the two extremes, the practical material ladder follows moisture. Published valve-maker tables differ in the exact bands - some switch at 20 ppm, some at 50 ppm - but they agree on the direction: carbon steel body with Monel or C-276 trim up to about 50 ppm, a nickel-alloy body from 50 ppm towards the 150 ppm limit, and C-276 or C22 for everything once the chlorine is wet. A representative ladder, condensed from two major chlorine valve programmes, looks like this:
Water content | Body | Ball, disc or stem | Seats and seals |
|---|---|---|---|
0 to 20 ppm | Carbon steel (WCB / A105) | Alloy C22 ball with Monel 400 stem | PTFE or modified PTFE; filled PTFE where ferric chloride abrasion is expected |
20 to 50 ppm | Carbon steel | Alloy C22 ball and stem, or Monel | As above, with live-loaded PTFE stem seals |
50 to 150 ppm | Monel 400 or Alloy C22 | Monel or C22; Hastelloy C where loading and unloading admits atmospheric moisture | PTFE, modified PTFE or PCTFE; graphite only where fire-safe qualification is required |
Above 150 ppm (wet) | Hastelloy C-276 or C22, titanium, PTFE or PFA lined steel, FRP and PVC at low pressure | C-276 or C22; titanium only where the chlorine cannot dry out | PTFE and PFA; no elastomer seats |
Two clauses follow. The first is that the moisture basis must be stated on the enquiry, because a valve maker cannot choose between carbon steel and Monel without it, and will default to the more expensive option or to whatever is in stock. The second is that the trim and the body can legitimately be different materials: the pressure boundary sees stagnant chlorine, while the ball, stem and seat see the highest velocity and the abrasion from ferric chloride debris. Specifying the whole valve in one alloy is either wasteful or unsafe, depending on which alloy is chosen.
Chlorine duty is usually described as isolation or control, and the valve type follows. But each type has a characteristic weakness in this service, and the weakness is what the specification has to compensate for.
Type | Best chlorine duty | Characteristic weakness | Specification response |
|---|---|---|---|
On-off isolation on dry gas and liquid headers; the traditional choice for cylinder and tank car connections | Body cavity traps liquid; wedge pocket collects ferric chloride; rising stem packing is a chronic leak path; a partially open gate erodes | Vent or drain the body cavity; specify cavity relief where liquid can be trapped; live-loaded packing; full open or full closed only | |
Throttling at vaporizers, chlorinators and injection points; small bore isolation where a positive shut-off is needed | Seat and plug erode under throttling; the body has a low pocket that holds liquid and debris; high pressure drop generates cold spots | Hard-faced or C-276 trim; drain the pocket; avoid throttling at pressure ratios that produce flashing; extended bonnet where the body runs cold | |
Quarter-turn isolation; the default for liquid chlorine transfer and tank connections | Cavity traps liquid; seats and ball score on ferric chloride; not a throttling valve; the vent makes it unidirectional | Mandatory cavity relief; filled PTFE seats; C-276 or Monel ball; mark and install in the arrow direction; gear operator on larger sizes | |
Large diameter, low pressure gas and wet chlorine lines; cooling water and scrubber circuits | Disc edge accelerates flow at the seat; shaft seal is a leak path; elastomer seats are attacked; the disc obstructs the bore and traps debris | PTFE or modified PTFE seat; double-offset or triple-offset geometry; no elastomer seat; not for lethal-service tight isolation without a proven seat test | |
Preventing backflow into storage, vaporizers and tank car connections | Chatter and impact damage the disc and seat; a check valve is not a positive isolation; debris holds it open | Spring-assisted or cushioned closure; C-276 or Monel trim; do not credit a check valve as a block valve in a line-break procedure | |
Sleeved plug designs on wet chlorine and slurry-adjacent duties | Lubricated plug valves inject a hydrocarbon lubricant into the flow path, which is both a fuel and a contaminant | Lubricated plugs are excluded; if a plug valve is used, specify a sleeved, un lubricated design with PTFE sleeve | |
Diaphragm valves | Low pressure wet chlorine, hypochlorite and dosing lines | Diaphragm is the pressure boundary and the seat at the same time; elastomer diaphragms age and permeate | PTFE-faced diaphragm; body lining matched to wet service; limited to modest pressure and temperature |
The type decision interacts with the material decision in a way that catches out cost-driven projects. A butterfly valve is the cheapest way to isolate a large line, and in chlorine service the saving is real - but the leak path moves from the seat, which can be tested, to the shaft seal, which usually cannot be tested in place and which sits directly under the insulation or the actuator. A project that saves 40 percent on the valve and then adds a permanent gas detection point above the shaft has not saved anything. The honest comparison is between total installed leak paths, not between valve prices.
Non-metallics are where chlorine valves actually leak, and they are also where the cheapest specification mistakes are made, because "PTFE" is treated as a single material when in fact the grade, the filler and the manufacturing route all change the behaviour.
Material | Dry chlorine | Wet chlorine | Practical limit | Comment |
|---|---|---|---|---|
Virgin PTFE | Excellent | Excellent | Creep and cold flow above about 100 deg C under load | Chemically inert, mechanically soft; the default but not always the best |
Modified PTFE / TFM | Excellent | Excellent | Better creep resistance than virgin PTFE | Preferred for seats that must hold class A leakage over thermal cycles |
Filled PTFE (glass, carbon, bronze-free) | Excellent | Good | Filler must be chlorine-compatible | Recommended where ferric chloride debris abrades the seat; bronze-filled grades are excluded |
PCTFE | Good | Good | Lower temperature ceiling than PTFE | Higher stiffness and lower permeability; used for barrier seals and small seats |
PFA / FEP | Good | Good | Lower strength; usually a lining or diaphragm | Lining for wet chlorine bodies; see the lined valve permeation guide for the failure modes |
Graphite | Acceptable in packing and gaskets | Acceptable | Oxidation above about 450 deg C in air | The fire-safe packing choice; must be qualified by the valve maker, not added by the buyer |
FKM / EPDM / NBR | Marginal | Poor | Swelling, hardening and permeation | Chlorine attacks most elastomers; elastomer seats and O-rings are a recurring cause of seeps |
Packing deserves separate treatment because it is the chronic exposure path. A chlorine valve that passes its seat test can still emit through the stem packing for years, and the emission is invisible until it is smelled or detected. The practical package is a PTFE chevron set, live-loaded so that the load does not decay as the rings settle, tested to ISO 15848-1 Class A or the equivalent tightness class in the purchaser's specification. Graphite is used where fire-safe qualification is mandatory, but the choice belongs to the valve maker's chlorine-qualified bill of materials - a buyer who adds "graphite packing" to a chlorine order without checking the qualification has changed the valve, not specified it.
Gaskets follow the same logic. The Chlorine Institute's gasket pamphlet is the reference, and the materials that perform best are restructured or filled PTFE sheet rather than skived PTFE, because the creep relaxation of skived sheet is what produces the slow weep at a body joint after a few thermal cycles. Spiral-wound gaskets with PTFE filler are used, but the filler and the winding material both have to be chlorine-compatible, and the gasket must be listed for the moisture condition. A generic "PTFE gasket" clause is not sufficient.
Lubricants are a separate prohibition. Chlorine supports the combustion of hydrocarbons, and a hydrocarbon grease in a stem thread or a seat pocket is both a fuel and a contaminant that will chlorinate over time. Chlorine-prepared valves are assembled with no hydrocarbon lubricant, or with a named PTFE-based or chlorinated lubricant approved by the valve maker. Any clause that says "lubricate as required" is a defect in a chlorine specification.
The preparation sequence for a chlorine valve is not the same as oxygen cleaning, and confusing the two produces either an over-specified valve or an under-dried one. Oxygen cleaning targets hydrocarbon residue because hydrocarbon is the fuel. Chlorine preparation targets both hydrocarbon and moisture, because moisture is what turns a passive surface into a corroding one, and because a hydrocarbon film is a combustion risk in a fluid that supports combustion. The acceptance criteria overlap, but the drying step is the one that chlorine cannot do without.
Step | What is done | Acceptance criterion | Common failure |
|---|---|---|---|
Degrease and clean | Remove machining oils, preservatives and handling contamination from all wetted parts | No visible residue; solvent or aqueous process compatible with the soft goods | Cleaning after assembly, which leaves solvent trapped behind seats |
Dry | Hot dry air or nitrogen through the assembly until the outlet dew point is stable | Dew point of -40 deg C or below, measured at the outlet | Judging dryness by time rather than by measurement |
Assemble | Assembly in a controlled area, with chlorine-compatible lubricant only where the design requires it | Named lubricant, recorded batch; no hydrocarbon grease | Assembly on an open shop floor in humid weather |
Pressure test | Shell and seat test with dry nitrogen or dry air | Test medium dew point -40 deg C; pressure and duration per the purchase specification; seat leakage class stated | Hydrostatic test with water, followed by incomplete drying |
Seat leak test | Quantified leakage measurement, not a bubble count | Leakage rate and test medium recorded; class consistent with the API 598 test requirements | Bubble test on a valve that will carry liquid chlorine, where the acceptance is a rate in ml/min or bubbles/min |
Cavity relief check | Prove the vent or self-relieving seat is open and relieves in the correct direction | Flow through the relief path demonstrated; direction verified against the body arrow | Relief path blocked by sealant, paint or a misplaced seat |
Cap, plug, bag | All ports sealed, desiccant where specified, tagged "prepared for chlorine service" | Seals intact on receipt; tag identifies the preparation standard | Caps removed on site and refitted after the valve has sat open in the yard |
The hydrostatic question needs a clear answer because it comes up on every project. A water test is the normal way to prove shell integrity, and refusing it outright is impractical. What matters is what happens afterwards: the valve must be drained, dried to the specified dew point, and the dew point recorded, and the test water must not be left in a cavity that will later be closed off. A chlorine valve that is hydrotested, drained and shipped wet is a valve that has been prepared for corrosion, not for chlorine. Where the specification allows it, testing with dry nitrogen avoids the problem entirely, and many chlorine valve programmes do exactly that. The storage and preservation guide covers the packaging side of the same problem.
Chlorine is about 2.5 times as heavy as air, so a leak does not disperse upward - it finds the lowest point in the room, the pit, the trench or the sump, and stays there. That single property invalidates several installation habits that are normal in other services.
No valves in pits or low points. A valve installed at the bottom of a pit puts the packing gland and the body joint in the place where a leak accumulates. If the layout forces a low point, the valve belongs on the vertical run above it, with the drain piped to a closed system.
Stem orientation matters. A vertical stem with the gland accessible is the default; an inverted valve puts the packing in a pocket where liquid chlorine can sit against it, and makes gland adjustment impossible without scaffolding.
Drain and vent to a closed system. Every cavity drain, vent and gauge root is a potential release point. They should be piped to a scrubber or a safe outdoor location, sloped so that moisture cannot collect, and never manifolded together - a common header lets a single leak travel to every connection on it. Vent headers must also be kept separate from ammonia-bearing vents, because chlorine and ammonia can form nitrogen trichloride, a shock-sensitive compound.
Do not rely on the valve for isolation during maintenance. Positive isolation for a line break needs a double block and bleed, a spectacle blind or a physical spade, plus a documented procedure. A single seat, especially a metal-seated or a vented one, is not a maintenance isolation.
Gas detection at low level, plus a second point at the gland. Detection at the ceiling is the wrong installation for a gas that is heavier than air. The second point belongs where the packing is, because that is where the slow leaks originate.
Access for packing replacement. The most common maintenance task on a chlorine valve is the one that requires the gland to be broken. If the valve is installed with 100 mm of clearance above the yoke, the task becomes a line break instead of a gland adjustment.
Chlorine is a regulated toxic substance in most jurisdictions, and the regulatory thresholds change what the valve specification has to contain. In the United States, a process holding 1,500 lb (about 680 kg) or more of chlorine falls under OSHA process safety management, and 2,500 lb (about 1,130 kg) triggers the EPA risk management programme. Those thresholds are low: a single one-ton container exceeds the process safety threshold on its own, and two of them cross the risk management threshold. Once a plant is inside either programme, the valves stop being a procurement item and become part of a documented mechanical-integrity system, with written procedures, change control and proof testing.
Function | Typical arrangement | Specification point | Failure to avoid |
|---|---|---|---|
Normal isolation | Manual or gear-operated ball or gate valve, chlorine-prepared | Cavity relief; ISO 15848-1 Class A packing; prepared and tagged | Standard valve ordered with a "chlorine" note in the description only |
Emergency isolation | Pneumatic spring-return fail-close actuator on the main liquid or gas connection | Fail-close on loss of air; remote manual trip; position feedback that proves the closed position | Fail-last or double-acting actuator with no stored energy to close |
Control / throttling | Globe valve with C-276 trim and a positioned actuator | Characterised trim; anti-cavitation or noise trim only if the process requires it | Ball valve used as a throttling device, eroding the seat and the ball |
Relief | Rupture disc plus relief valve, or a relief valve with a chlorine-compatible seat | Set pressure below the rating of the weakest component at the maximum ambient temperature | Relief sized for gas only, on a line that can be liquid-full |
Proof testing | Partial stroke test on actuated valves; full stroke and seat test at turnaround | Test interval and acceptance recorded; a partial stroke proves movement, not tightness | Treating a successful partial stroke as evidence of seat integrity |
Actuator selection follows the toxic-service logic rather than the process-control logic: the safe position is closed, the actuator must be able to close against the full differential with the available air pressure (including the minimum instrument air pressure, not the compressor nameplate), and the closing time should be fast enough to limit the release but not so fast that the liquid column produces a surge. The general sizing arithmetic is in the pneumatic actuator sizing guide; what changes for chlorine is the consequence of a wrong spring, which is a release rather than a nuisance trip.
The following clauses are the ones that decide what arrives on site. They are written so that they can be copied into an enquiry, and each one closes a failure mode described above.
# | Clause | Why it is there |
|---|---|---|
1 | State the water content of the chlorine in ppm by weight, and whether the service is dry or wet, on the first page of the datasheet | Every material decision depends on it; a valve maker cannot choose between carbon steel and C-276 without it |
2 | State the design temperature as the maximum ambient plus solar gain, not the normal operating temperature | Vapour pressure at 50 deg C is twice the value at 20 deg C |
3 | For liquid service, require cavity relief to the upstream side and require the body to be marked with a direction arrow | Liquid-full expansion reaches hundreds of bar; the arrow is part of the safety function |
4 | State whether the relief method is a vented ball or self-relieving seats, and accept that a vented valve is unidirectional | The two methods behave differently in service and cannot be substituted at the factory's discretion |
5 | Specify body, trim and seat materials separately, with the moisture band that applies to each | The pressure boundary and the high-velocity parts see different conditions |
6 | Prohibit copper alloys, aluminium, titanium in dry service, and lubricated plug valves | Each is a known, repeated failure mode rather than a theoretical risk |
7 | Require PTFE or modified PTFE packing, live-loaded, tested to a stated tightness class | Packing is the chronic emission path and the most common leak on a chlorine valve |
8 | Require cleaning and drying to a measured dew point of -40 deg C, with the measurement recorded | Dryness judged by time rather than measurement is not dryness |
9 | State the test medium, the test pressure and the seat leakage class, and require the test to be dry | A water test that is not followed by drying prepares the valve for corrosion |
10 | Require the cavity relief path to be functionally proven and recorded before shipment | A blocked relief path is invisible until the body fails |
11 | Require "prepared for chlorine service" tagging, caps on all ports, and packaging that survives outdoor storage | Preparation is lost between the factory and the flange |
12 | Require material certificates to EN 10204 3.1 with heat numbers and a PMI check on the trim alloys | Monel, C-276 and 316 are visually similar; identification by appearance is how the 2014 titanium fire happened |
Chlorine valve failures are rarely sudden, and the evidence is usually present at the previous inspection. The signatures below are the ones worth training a maintenance team to recognise, because each one points to a specific cause rather than to "the valve is bad".
Signature | Most likely cause | Immediate action | Design fix |
|---|---|---|---|
White or green deposit at the gland | Moisture ingress plus corrosion of the stem; packing not live-loaded | Do not wipe it off and continue; plan a gland rebuild with the line isolated | Live-loaded PTFE packing; stem material upgraded to Monel or C-276 |
Body joint weep that appears after a thermal cycle | Creep relaxation of a skived PTFE gasket | Retorque to the maker's figure, then replace if it returns | Restructured PTFE gasket; controlled bolt tightening sequence |
Seat leak that appears only at low differential pressure | Seat compression set, or ferric chloride debris on the seat band | Cycle the valve; if it persists, plan replacement | Filled PTFE or PCTFE seat; filtration upstream; cavity drain |
Valve will not open after standing closed | Corrosion at the seat or guide, or thermal contraction of the body onto the wedge or ball | Do not force it with a bar; apply the maker's breakaway torque and investigate | Correct trim material; cavity drain; exercise procedure at fixed intervals |
Body distortion or a cracked body cap | Liquid-full cavity with no relief, or relief path blocked | Treat as a release risk; evacuate and depressurize from a remote point | Cavity relief; vented ball; relief path check at every maintenance |
Brown or black staining on the ball, with a scored seat band | Ferric chloride abrasion in dry service | Replace seats and ball; inspect upstream for debris | Filled PTFE seats; C-276 ball; upstream strainer |
Pressure creeping up in a closed cavity | Seat passing, or relief path blocked while the cavity warns up | Investigate before the next thermal cycle | Pressure indicator on the cavity; relief path proven at commissioning |
Rapid corrosion of a stainless trim | Moisture excursion with 304/316 in service | Replace with C-276 or Monel; find the moisture source | Remove 300-series stainless from the chlorine bill of materials |
What is the difference between dry and wet chlorine for valve selection?
Dry chlorine contains no more than 150 ppm of water by weight, and wet chlorine contains more. In dry chlorine, carbon steel passivates under a ferric chloride film and is the standard pressure-boundary material; in wet chlorine the same steel corrodes at millimetres per year. The distinction is about moisture, not phase, so dry gas, dry liquid, wet gas and wet liquid are four different material problems. The trap is that dryness belongs to the containment, not to the product: a dry line becomes locally wet when a valve is changed in humid air, and the bulk assay will never show it.
Do chlorine ball valves need a vent hole in the ball?
Not all of them, but any valve that can trap liquid chlorine does. A closed ball valve isolates its body cavity between two seats, and liquid-full chlorine gains roughly 12 to 14 bar for every kelvin of warming, so 30 K of solar gain reaches about 390 bar. The Chlorine Institute's dry chlorine piping pamphlet requires a relief path, and there are two accepted methods: a vent hole connecting the cavity to the upstream throat, or self-relieving seats. A vented valve seats in one direction only and must be installed with the body arrow upstream; a self-relieving valve keeps bidirectional shut-off but depends on repeatable seat flexure.
Can 316 stainless steel be used in chlorine service?
It resists dry chlorine below about 121 deg C, and it is still the wrong default. The 300 series is susceptible to chloride stress corrosion cracking, and chlorine service is full of moisture excursions - a changeout in humid air, a wet purge, a leaking gland. Once cracking starts it is not visible from the outside and it does not need a high stress to propagate. Most chlorine valve programmes therefore exclude 300-series stainless from the bill of materials and use carbon steel for the body with Monel or Hastelloy C-276 for the ball, disc and stem, or C-276 for everything once the chlorine is wet.
Why is titanium prohibited in dry chlorine but used in wet chlorine?
Titanium's corrosion resistance comes from a titanium dioxide film that needs moisture to stay stable. In wet chlorine, brine and hypochlorite that film works well. In dry chlorine, titanium reacts exothermically and can sustain a metal fire that consumes the component and breaches its containment. The failure is not theoretical: a documented 2014 release at a chlorine packaging plant involved a gauge with a titanium housing and a tantalum diaphragm that was taken from a storeroom because it looked like the stainless gauges normally used there. Aluminium, tin and gold are excluded from dry chlorine for the same reason.
What test medium should be used for a chlorine valve?
Dry nitrogen or dry air with a dew point of -40 deg C or below, wherever the purchase specification allows it. A hydrostatic test with water is the normal way to prove shell integrity and is acceptable if the valve is then drained, dried to the specified dew point and the measurement is recorded. What is not acceptable is a water test followed by a drain and a shipment, because the valve then arrives prepared for corrosion rather than for chlorine. The seat test should be quantified as a leakage rate rather than a bubble count, and the cavity relief path should be functionally proven before the valve is packed.