A ball valve leaves the works with a coating on it, and the coating is usually the line in the specification that nobody argues about until the valve has spent two winters on a coastal laydown. Then the blisters open along the body joint, the flange faces are rusted under the gasket, and the paint on the stem has been scraped off by the stem's own travel. None of that is a bad batch of paint. It is a wrong corrosivity category, a missing stripe coat and a masking decision taken in a hurry, and all three are settled before the first litre is mixed.
This article is about the external surface of a valve specifically, not about pipe coating in general. A valve is a poor shape to paint: it has crevices where two castings meet, small radius edges, threads, a moving stem, a nameplate that must stay legible, flange faces that must stay flat, and weld ends that must stay bare. Roughly speaking, every one of those features is a place where coating work is either done deliberately or fails quietly two years later.
1. The corrosivity category is the decision, not the paint brand
ISO 12944-2 classifies the environment, not the product. C1 and C2 are heated interiors and dry stores. C3 is urban and inland industrial with moderate sulphur dioxide and low chloride. C4 is coastal or industrial with meaningful salt or chemical load, and it is where most outdoor pipe racks sit within a few kilometres of a coast. C5 is the salt-laden coastal belt and heavy industrial atmosphere, and CX is the offshore and tropical marine extreme. Then there is the immersion family, Im1 to Im4, for structures in fresh water, buried in soil, or permanently in sea water.
Two mistakes recur. The first is specifying C4 for a site that is 800 m from the sea, because the chloride deposition rate is not a function of distance in the way people assume; onshore winds carry salt several kilometres inland and the rate depends on terrain, prevailing wind and rainfall washing. The second is specifying C5 and then accepting a 160 micron system because the painter had it in stock. Category and dry film thickness are a pair. One without the other means nothing.
| Category | Environment | Where a valve actually meets it | Typical total DFT | Durability band |
|---|---|---|---|---|
| C2 | Heated, dry, low pollution | Indoor plant room, warehouse stock | 80 to 120 micron | Low to medium |
| C3 | Urban inland, moderate SO2 | Sheltered outdoor rack, water treatment hall | 160 to 200 micron | Medium to high |
| C4 | Industrial or coastal, moderate chloride | Refinery pipe rack, 2 to 10 km inland | 240 to 280 micron | High, 15 to 25 years |
| C5 | Coastal belt, high chloride, aggressive | Jetty, dock, desalination, 0 to 1 km from sea | 280 to 350 micron | Very high, over 25 years |
| CX | Offshore, tropical marine industrial | Topsides, FPSO module, splash-adjacent | 320 to 450 micron, or TSA | Very high, over 25 years |
| Im2 | Buried in soil or immersed in fresh water | Direct buried valve, sump, tank farm pit | 400 to 600 micron glass flake, or tape | Very high |
| Im3 | Immersed in sea water | Outfall, seawater intake, subsea splash zone | Not a paint job: TSA, duplex, or CP | Design life dependent |
2. What a C5 system looks like in micrometres
A working C5 system on a carbon steel valve body is three coats and about 300 micron: a zinc-rich epoxy primer at 60 to 80, a micaceous iron oxide epoxy intermediate at 100 to 120, and an aliphatic polyurethane topcoat at 60 to 80. The primer provides galvanic protection at scratches, the MIO intermediate slows water transmission and gives the film its barrier thickness, and the polyurethane topcoat provides colour and UV stability. A zinc silicate primer is an alternative with better heat resistance, but it is unforgiving: it must be topcoated within its recoat window, and a zinc silicate left exposed to moisture before topcoating will form zinc salts that destroy adhesion.
| System build | Primer | Intermediate | Topcoat | Total DFT | Use it for |
|---|---|---|---|---|---|
| Alkyd, one or two coats | Zinc phosphate alkyd 40 | none | Alkyd 40 | 80 to 120 micron | C2 indoor only, never for a wet or coastal site |
| Epoxy two coat | Zinc phosphate epoxy 60 | MIO epoxy 120 | none | 180 to 220 micron | C3, will chalk if exposed to sun |
| Epoxy plus polyurethane | Zinc rich epoxy 70 | MIO epoxy 110 | Aliphatic PU 70 | 240 to 280 micron | C4, the default for outdoor process valves |
| High build C5 | Zinc rich epoxy 80 | MIO epoxy 120 | PU or polysiloxane 80 | 280 to 350 micron | C5 coastal, jetty, desalination |
| Offshore qualified | Zinc silicate 75 | Epoxy 125 | Polysiloxane 75 | 275 to 350 micron | CX, qualified to NORSOK M-501 system 1 |
| Glass flake epoxy | Glass flake epoxy 2 x 200 | none | none | 400 to 500 micron | Buried and immersed, Im2 |
| Inert multipolymeric matrix | Matrix coat 2 x 150 | none | none | 300 to 350 micron | CUI service to 150 degrees C, ISO 12944-9 |
3. Surface preparation decides more than the coating does
On our own rework records, the majority of coating failures traced back to preparation rather than to the product: blast profile too low, salt not washed off, blasting done in the morning and coating in the afternoon after rain, or edges left as sawn. The enforceable numbers are these. Blast to Sa 2.5 of ISO 8501-1, with a profile of 50 to 85 micron for a zinc-rich system, measured on the blast comparison plates. Salt contamination on the steel before coating below 20 mg per square metre for immersed or C5 duty, and below 50 for general atmospheric, measured by the Bresle method of ISO 8502-9. Dust class 2 or better. Relative humidity below 85 percent and steel temperature at least 3 K above dew point, checked with a hygrometer rather than a hand.
Edges matter more on a valve than on a plate because a valve has many of them. A sharp arris cannot hold 300 micron of paint: the film pulls back by surface tension and the exposed steel at the corner begins to rust, and the rust lifts the film back from the edge. The specification should require all edges, weld toes and flame-cut surfaces to be ground to a radius of about 2 mm before blasting, which is ISO 12944-3 territory and is routinely written out of contracts to save an hour of grinding per valve.
Blasting a valve also means protecting it. Grit enters the bore, the cavity, the stem threads and the flange stud holes, and a single particle left in the seat area will score the ball on first operation. Blind the bore and the ports, plug the grease nipples, tape the flange faces, and blow out the cavity afterwards. A valve that was blasted without masking is a valve that will be returned.
4. CUI: the failure mode that begins at 50 degrees C
Corrosion under insulation is not a coating problem in the ordinary sense, because the coating is only one of the defences and often the weakest. API 583 sets out the mechanism: water enters the insulation through a damaged jacket, a failed sealant, a vent that lets rain in, a steam trace leak, or simply condensation on a cold surface, and it sits against the steel in a warm, aerated, chloride-concentrating environment. The temperature band for carbon steel is roughly 50 to 175 degrees C, with the fastest attack between about 80 and 120 degrees C, where the water is hot enough to drive the corrosion reaction but not hot enough to boil away. Below 50 degrees C the water is cool and the rate is lower. Above 175 degrees C most of the water flashes off at the surface.
Austenitic stainless has the opposite problem in the same band. External chloride stress corrosion cracking of 304 and 316 occurs between about 60 and 175 degrees C where chloride from insulation leachate, sea salt or washdown concentrates on the surface. A painted stainless valve is not automatically protected: standard epoxy is permeable to water vapour and the chloride still arrives. This is why hot stainless lines in coastal plants are either insulated with low-chloride insulation and coated with a CUI-rated system, or left uninsulated, and why "stainless so no coating" is a decision that needs a temperature and a chloride level behind it.
| Body material | CUI risk band | Mechanism | What fails first | Mitigation that works |
|---|---|---|---|---|
| Carbon steel, uncoated | 50 to 175 degrees C, worst 80 to 120 | General and pitting corrosion under wet insulation | Wall loss at the body joint and under support saddles | TSA 300 to 500 micron, or inert matrix coating, plus closed cell insulation |
| Carbon steel, standard epoxy | above 120 degrees C | Film hydrolysis and blistering, then corrosion | Blister at the body joint, coating lifts off in sheets | Use a coating rated for the temperature, not a 90 degrees C epoxy on a 150 degrees C line |
| 304 or 316 stainless | 60 to 175 degrees C | External chloride SCC | Branching cracks at the body joint and under the insulation | Low chloride insulation, CUI-rated coating, or remove the insulation |
| Duplex 2205 | 60 to 175 degrees C | SCC resistance is better but not unlimited | Pitting at chloride deposits | Coat anyway in C5, keep the insulation dry |
| Any material, wet insulation | all temperatures | Water trapped at the low point of the jacket | The bottom of the valve, invisible until the insulation is cut | Drainable jacket, sealed penetrations, inspection ports at known low points |
5. The details a valve needs and a pipe does not
Flange faces. Mask them before coating, and keep the masking on through the topcoat. A raised face with 300 micron of paint on it will not seat a spiral wound gasket properly: the gasket bites into the film, the film creeps, and the joint leaks at pressure. The accepted practice is to coat the flange outside diameter and the back face, mask the raised face and the gasket seating area, and apply a thin film rust preventive to the bare face for shipment. Where a coating is required on the seating face by project specification, the DFT must be stated and the gasket selected for it.
Stripe coats. Every edge, weld, bolt head, threaded connection and awkward corner gets a brush-applied stripe coat in the same or a contrasting colour before the spray coat. On a valve this is not cosmetic: the body joint, the bonnet flange and the packing gland are exactly the places where spray misses and where crevice corrosion starts.
Stem and gland. The stem must move. Painting the exposed stem is pointless because the coating will be sheared off within a hundred cycles, and painting the part of the stem that enters the packing is actively harmful because paint flakes migrate into the packing and damage it. Protect the stem with a grease film, a plastic or rubber boot, or a stainless stem in the first place, and leave the travel zone bare. The gland nuts and studs get the same treatment as body bolting.
Bolting. Hot dip galvanised to ISO 1461, or a PTFE-based or zinc flake coating, or stainless where the temperature and chloride allow. Bare carbon steel studs on a C5 valve will show rust before the valve is installed, and rust staining on a new coating is what a client's inspector photographs. Note that 316 bolting in a coastal chloride environment above about 60 degrees C is a SCC risk, not a safe default.
Weld ends. Leave 50 to 100 mm bare at each end for the site weld, and protect that bare band with a weldable shop primer or a removable tape. Coating over the weld prep means the welder grinds it off anyway, and the grinding takes the adjacent coating with it.
Nameplate and fittings. Do not paint over the nameplate, the heat number, the tag plate or the grease nipple threads. Use a stainless tag wired on, and mask threads with plugs rather than tape, because tape adhesive remains and gauls the fitting.
6. Weld heat, field joints and repair
A weld destroys the coating 100 to 200 mm either side of the arc, and post weld heat treatment destroys it much further: a zinc-rich primer exposed above roughly 400 degrees C is gone, and a heat resistant aluminium-filled primer is the correct choice for PWHT zones because it survives the cycle. Where a valve is welded into a line after coating, the sequence should be: weld, NDT, PWHT if required, then field joint coating. Doing it in the other order is how a new valve acquires a 150 mm ring of rust at each end before commissioning.
For field joints, a heat shrink sleeve is the default on buried and splash-zone work, and a two-component liquid epoxy with a compatible primer is used where the geometry is awkward. The overlap onto the factory coating must be at least 50 mm and the surface under the overlap must be abraded and solvent wiped, because a sleeve that lifts at the overlap is a water trap. On buried valves, a petrolatum tape wrap over the whole body and both joints is a cheap and very effective addition, and it is compatible with cathodic protection in a way that a thick dielectric coating on the whole valve sometimes is not.
7. Inspection that actually rejects bad work
Dry film thickness is measured to SSPC-PA 2, and the rule that matters is the 80/20 one: the average of all readings must be at or above the specified DFT, no single reading may fall below 80 percent of it, and no reading should exceed 120 percent without the coating supplier's agreement, because an over-thick film cracks, traps solvent and loses adhesion. Measure on a frequency based on area, and measure after each coat rather than only on the finished system, otherwise the non-conforming layer is buried.
Holiday detection finds pinholes that a thickness gauge cannot. A wet sponge at low voltage is used for thin films, and a high voltage spark tester per NACE SP0188 for thick or immersed systems, with the voltage set from the film thickness and the coating type. Set too high and the test burns holes through a sound coating and creates the defect it was looking for, which is a common way to fail a good valve.
Adhesion is checked by pull-off to ISO 4624, typically requiring 5 MPa or cohesive failure in the substrate for an epoxy on Sa 2.5 steel, and by a knife test on site. Cure is checked before the valve is packed: a solvent rub on a sample, and a check that the recoat window has not been exceeded. The whole thing should be documented against a coating procedure specification with batch numbers, environmental readings at each coat, and a signature per coat, because the argument about a failed coating two years later is settled by those records and by nothing else.
8. When coating is the wrong answer
There are four cases where the right move is not a better paint. Thermal spray aluminium at 300 to 500 micron, sealed, is the standard answer for CUI service and for immersed carbon steel, because it is metallic, conductive, and does not blister; it also survives mechanical damage far better than a film. A duplex or super duplex body solves the seawater case outright and is often cheaper than a coated carbon steel valve over a 25 year life once rework is counted. A GRP or polyethylene shroud over a buried valve replaces the coating with a barrier that cannot chip. And cathodic protection, correctly coordinated with a coating that is not so thick and dielectric that it starves the CP current, is the standard for buried and subsea work.
What does not work is using coating thickness to compensate for a category that was specified too low, or adding a fourth coat to a system that is failing because the salt was never washed off the steel.
9. The cost of a coating that fails
Re-blasting and recoating a valve in a yard costs several times the original coating price, because the work is now manual, the masking is more difficult, and the valve has to be disassembled to protect the seat. Doing the same work offshore or on a live plant multiplies it again by the cost of access, scaffolding, permits and production deferral, and it is not unusual for the remedial cost of one valve to exceed the purchase price of ten. The commercial argument for specifying the correct system is not the coating cost. It is the probability of ever having to touch it again.
10. A coating clause you can paste into a specification
External surfaces of valves shall be prepared and coated in accordance with ISO 12944 for the corrosivity category stated in the data sheet, with the category, the system, the nominal DFT and the durability band all stated in the purchase order. Carbon steel surfaces shall be blast cleaned to Sa 2.5 of ISO 8501-1 with a profile of 50 to 85 micron, salt contamination below 20 mg per square metre, and coating applied at a steel temperature at least 3 K above dew point and a relative humidity below 85 percent. Systems shall comprise a zinc rich or zinc silicate primer, an epoxy intermediate and an aliphatic polyurethane or polysiloxane topcoat, with a contrasting stripe coat applied by brush to all edges, welds, bolt heads and crevices before each full coat. Flange faces shall be masked and left with a rust preventive only. Stems shall be protected but not coated over the travel zone. Weld ends shall be left bare for 50 to 100 mm and protected with weldable primer. Where the operating temperature falls between 50 and 175 degrees C and the valve is insulated, a CUI-rated system shall be used and shall be stated separately from the atmospheric system. Dry film thickness shall be verified to SSPC-PA 2 with the 80/20 rule, holiday tested, and adhesion tested to ISO 4624, with records submitted per valve. J-VALVES supplies ball valves from 1/2 to 56 inch (DN15 to DN1400) in ASME Class 150 to 2500, in carbon steel, stainless, duplex and high alloy bodies, for service from minus 196 to 650 degrees C, and can supply with ISO 12944 C3 to CX systems, NORSOK M-501 qualified systems, thermal spray aluminium for CUI and immersed duty, and hot dip galvanised or PTFE coated bolting, tested to API 598 and API 6D with 1.5 times shell and 1.1 times seat pressure, with coating records and 3.1 certificates in the documentation package, and quotations returned within three working hours.
11. Frequently Asked Questions
What corrosivity category should I specify for a coastal site?
C5 for anything within about a kilometre of the sea in an exposed position, and C4 only where the site is sheltered and the chloride deposition rate has actually been measured. Distance alone is a poor guide because onshore wind carries salt several kilometres inland. If the plant is offshore or in a tropical marine industrial setting, CX applies and NORSOK M-501 is the system most buyers reference rather than ISO 12944 alone.
Can a stainless steel valve be left uncoated?
For atmospheric C3 and below, usually yes, provided the grade is right and the surface is clean. For insulated service between 60 and 175 degrees C in a chloride environment, no: external chloride stress corrosion cracking of 304 and 316 is a real failure mode and it starts under the insulation where nobody looks. Duplex is more resistant but is still coated on C5 projects as a matter of practice.
Why does paint fail at the flange face and the body joint?
Because those are edges and crevices. A film thins by surface tension over a sharp arris, and a crevice holds water and salt against the steel for longer than any flat surface does. That is why the specification needs edges ground to about 2 mm radius, a brush applied stripe coat on every edge and weld, and flange faces masked rather than coated, since a thick film under a gasket will creep and leak.
How thick should the coating be on a C5 valve?
Around 280 to 350 micron total, built as a zinc rich primer at 60 to 80, an epoxy intermediate at 100 to 120 and a polyurethane topcoat at 60 to 80. More is not better: above roughly 120 percent of the specified thickness the film is prone to solvent entrapment, cracking and loss of adhesion, and the 80/20 rule of SSPC-PA 2 sets both a floor and a practical ceiling.
Is a standard epoxy suitable for an insulated line at 140 degrees C?
No. A conventional epoxy is typically rated to about 90 to 120 degrees C for continuous service, and above that it hydrolyses and blisters from the steel side, which is indistinguishable from CUI until the insulation is removed. For that band the options are a novolac epoxy rated to 150 degrees C or above, an inert multipolymeric matrix coating qualified to ISO 12944-9, or thermal spray aluminium.
