Cryogenic Ball Valve -196°C to -50°C for LNG and Air Separation

A cryogenic ball valve is a quarter-turn isolation valve built for service between -50°C and -196°C, in which the body, trim, seat, bolting and bonnet length are selected for thermal contraction and verified by test to BS 6364 and ISO 28921-1 instead of by ambient-temperature rules. The ball and stem are austenitic stainless with recorded impact energy, the seat is normally PCTFE, the body joint carries a graphite-based gasket, and the bonnet is long enough to keep the packing above 0°C.

J-VALVES manufactures cryogenic ball valves from 1/2 inch to 56 inch in ASME Class 150 to 2500 at its works in Wenzhou, Zhejiang. The buyers are EPC contractors, air separation packagers and LNG terminal operators who need a valve that still seals after a cold soak and still turns by hand. This page sets out the temperature bands, the failure modes that decide the design, the material and seat matrices, the BS 6364 test sequence and the route from a datasheet line to a construction type.

Why a Ball Valve and Not Something Else

A ball valve gives a quarter turn, a straight bore and shutoff in both directions, and those three properties matter more at -162°C than at ambient. A gate valve has a lower pressure drop, but its wedge slides, and a sliding fit that is correct at 20°C can bind when the body shrinks around it. The ball rotates against a seat that moves a few tenths of a millimetre, and the seal is surface contact rather than a wedging action. The trade is that the seat and the packing, not the metal, set the temperature limit of the whole valve.

Cryogenic ball valves appear in four kinds of plant: LNG terminals, where the fluid is about -162°C; air separation units, where liquid nitrogen at -196°C, liquid oxygen at -183°C and liquid argon at -186°C are distributed through a cold box; liquid storage and transport, including tank nozzles and pump suction; and petrochemical plants with refrigerated storage.

Temperature Bands and What Changes at Each

For a cryogenic ball valve, service is generally taken as -50°C and below. Between -29°C and -50°C many specifications still call the duty low temperature rather than cryogenic, and the design can stay close to a standard valve with better materials. Below -50°C the seat, the bonnet and the test programme all change.

The first band is a materials decision. A PTFE seat will seal once at -40°C, but it has taken a compression set and will not recover through a thermal cycle, so the second cycle leaks. The second band is where the bonnet becomes a thermal device rather than a pipe connection. The third is where the test programme becomes the product, because at -162°C nobody accepts a calculation in place of a cold test. The fourth covers helium and research duties, quoted case by case, where a metal seat or a PCTFE insert in a metal carrier takes over.

Why Standard Ball Valves Fail in Cryogenic Service

A standard ball valve fails at cryogenic temperature for five separate reasons, and a valve that fixes only four of them still leaks. Each one can be predicted from the materials and the temperature difference, which is why a cryogenic design can be reviewed on paper before anything is cast.

1. Packing and gland freezing. Moisture condenses on the cold stem and freezes into a collar of ice around the gland. When the valve turns, that ice is dragged into the packing, scoring the stem and tearing the rings. Water in the process does the same from the inside.

1. Elastomer and PTFE embrittlement. PTFE and RPTFE are flexible because their polymer chains slide at room temperature. Below about -29°C that movement is gone, and a seat that relied on elastic recovery keeps whatever shape it was pressed into. The valve may pass a leak test cold and fail on the second cycle, because the first cycle is what set the shape.

1. Differential contraction between ball and seat. Austenitic stainless contracts by roughly 0.3 percent between 20°C and -196°C. PTFE and PCTFE contract several times as much, while the seat pocket in the body contracts with the steel. A seat that was in correct interference at ambient can lose preload, or be crushed, depending on which side of the pocket traps it. This is why cryogenic seats are spring loaded.

1. Body and bonnet joint gasket relaxation. The joint is a bolted assembly of stainless flange, gasket and stainless bolt. At low temperature the bolt and the flange stack contract by different amounts, the gasket changes thickness, bolt preload drops and the joint weeps. It cannot be retightened in service because the joint sits under insulation or inside a cold box.

1. Moisture trapped in the cavity. Test water left in the body cavity freezes, expands by about nine percent, and either cracks the body or jams the ball. Even a little ice blocks the pressure relief path, so trapped liquid expands against a closed cavity on warm-up.

A sixth effect decides the actuator. Seat squeeze, contraction of the ball against the seat and stiffened grease raise break-to-open torque, often by a factor of two or more over the ambient figure. An actuator sized on the ambient torque will stall on the first cold cycle.

Body and Trim Material Matrix

Austenitic stainless is the default body material for a cryogenic ball valve because it has no ductile-to-brittle transition in this range. It is still impact tested, not because the test is expected to fail, but because ASME B16.34 and every LNG specification require a recorded result at the design temperature. Carbon and low alloy steels do have a transition, which is why their minimum temperature is a hard number.

CF8M is the workhorse. Minimum yield is 205 MPa and minimum tensile 485 MPa to A351, which covers Class 150 to 600 bodies at almost any size in the range. CF3M is the low carbon version, chosen where the body or adjoining pipe will be welded. Forged F316 and F316L are used for Class 900 and above, and above about 24 inch where casting soundness is harder to guarantee.

A350 LF2 is the material buyers ask about most often and misapply most often. It is a sound low temperature carbon steel with 250 MPa minimum yield, 485 to 655 MPa tensile and 0.30 percent maximum carbon, and it is genuinely impact tested. It is limited to about -46°C. That makes it correct for cold carbon steel service and wrong for LNG at -162°C, where the Charpy shelf has already collapsed. LCC behaves the same way.

9% nickel steel appears in specifications because LNG tanks are built from it. It is not normally used for valve bodies in this size range: it is expensive, hard to cast, and offers nothing CF8M does not already give at -196°C. When a datasheet asks for a 9% nickel body, the requirement usually belongs to the tank nozzle.

Seat, Seal and Packing Selection

The seat decides whether a low temperature ball valve works, and there are only a few honest options.

PCTFE is a chlorotrifluoroethylene polymer and it is the reason a cryogenic ball valve can be soft seated at all. It keeps measurable toughness at -196°C, has low permeability to nitrogen and methane, and takes the seat load without flowing. Its thermal expansion is still several times that of stainless, so it cannot simply be pressed into a pocket and left to hold itself. A spring or a wave washer behind the seat keeps sealing stress positive through the contraction range. Above Class 900 the seat is usually a PCTFE insert in a metal carrier.

Packing follows the same logic, in reverse. Flexible graphite is standard because it does not melt, does not take a set when cold, and tolerates a stem that moves as it cools. PTFE packing is acceptable only where the extended bonnet keeps the gland above 0°C.

Bonnet Length, Heat Leak and Insulation

An extended bonnet on a cryogenic ball valve is not cosmetic. It puts distance between the cold body and the packing, and that distance is thermal resistance in series with a very good conductor. Stainless conducts about 16 W/m·K at room temperature and less when cold, but a 100 mm stem still carries enough heat to pull the gland below freezing when the body is at -162°C.

BS 6364 requires the bonnet to be long enough for the insulation and the service temperature. In practice the length is set by the cold box wall and its insulation, the distance needed to keep the gland above 0°C at the specified ambient, and the room needed for the handle outside the insulation. A valve in a cold box with 300 mm of insulation needs a bonnet that clears all of it, and the dimension is quoted from body centre to gland face.

Two details decide if the extension works. Orientation comes first: a bonnet pointing down collects liquid and ice in the extension, and a bonnet pointing sideways puts a bending load on the joint. Vertical up is the default. The vapour barrier comes second. It has to be continuous around the stem, or moist air reaches the cold metal, condenses, and the ice collar forms inside the insulation where nobody can see it.

Heat leak has a process cost as well: every watt down the stem boils liquid, and on a small liquid nitrogen line that boil-off shows as two-phase flow.

Test Sequence to BS 6364 and ISO 28921-1

A cryogenic ball valve is proved by test, and the test is a sequence rather than a single event. The ambient tests find the ordinary defects cheaply. The cold tests find the ones that appear only after contraction. The warm test exists because some failures surface only when the seat has to recover.

The shell test checks the pressure boundary and it is done warm because water is the practical medium and a shell failure at -196°C would be destructive rather than an inspection. The ambient seat test sets the baseline leak rate. ISO 5208 gives the rate letters, and the specification usually calls for rate A or B. A valve that cannot hold the rate warm will not hold it cold.

Chill down is where the test becomes cryogenic. The valve is immersed, or cold gas passed through it, until the body, bonnet and seat area reach the target temperature and stop changing. Thermocouples sit on the body near the seat, not in the bath, because the bath can be at -196°C while the seat is still at -80°C. Soak time follows the heaviest section.

The cold seat test uses helium. Its molecule is small and stays mobile at cryogenic temperature, so it finds leaks nitrogen would miss, and the leak is measured on a mass spectrometer instead of counted as bubbles. The valve is cycled at temperature, closed, and the leak measured on both seats where the design allows it. Torque is measured on the same cold valve, because that number sizes the actuator and rising torque is the first sign the seat is being crushed.

The final warm test is the one buyers forget. After the valve returns to ambient the seat has to recover its shape and hold the same leak rate. If it does not, the seat has taken a permanent set and the valve will fail on the second cool-down in the plant. The closing examination checks the stem for scoring, the body for ice damage and the dimensions for distortion.

J-VALVES records the shell and seat results, the cold leak rate, the torque and the temperature trace by serial number, and issues an EN 10204 3.1 material certificate with the impact results for the heats used. NACE MR0175 is available on request for sour and cold duty.

Sizes, Pressure Classes and Dimensional Practice

The range of a cryogenic ball valve is 1/2 inch to 56 inch and Class 150 to 2500, but the construction changes across that range, and the table below is the practical split.

Pressure rating at temperature needs care. For CF8M in ASME B16.34 group 2.2 the room temperature ratings are approximately 19.6 bar at Class 150, 51.1 bar at Class 300, 102.1 bar at Class 600, 153.2 bar at Class 900 and 255.3 bar at Class 1500. Austenitic grades keep those figures down to -196°C subject to impact testing, so the shell is rarely the limiting item. The limit sits in the seat, the packing and the bolting, which is why the datasheet has to state the design temperature.

Class to PN equivalence is quoted in most international specifications: 150 to PN20, 300 to PN50, 600 to PN100, 900 to PN150, 1500 to PN250, 2500 to PN420. The equivalence is one of rating, not of flange dimensions, and a PN50 flange will not bolt to a Class 300 flange without an adapter. Face-to-face dimensions follow ASME B16.10, flanges follow B16.5 up to 24 inch and B16.47 above it, and butt weld ends follow B16.25. An insulated line often calls for a longer face-to-face so the insulation runs continuously across the body.

How to Choose Within the Family

The first question is size and class. Above about 6 inch the seat load from line pressure becomes large enough that a floating ball needs an impractical stem, so the valve becomes trunnion mounted with spring loaded seats. A cryogenic trunnion mounted ball valve is the normal choice for LNG send-out lines and cold box headers.

The second is where the valve sits. A valve inside a cold box, behind insulation, or on a line that runs through a wall needs an extended stem, and the length is agreed as a dimension rather than taken from a catalogue. A branch that is fully insulated with the pipe needs only a standard bonnet. A line that cannot carry flange joints, or one that will be buried, takes a fully welded body.

The third is body material and fluid. Carbon steel at -46°C is legitimate for cold service, and an LF2 body valve is cheaper than CF8M where the temperature allows it. LNG at -162°C is not that duty. Sour service adds NACE MR0175 and hardness control on the bolting and the trim.

The fourth is function. Instrument isolation usually needs double block and bleed so the impulse line can be vented without draining the main line. Small bore instrument valves are three piece, because the body has to come out of a manifold without cutting pipe. Abrasive particles, thermal cycling that would destroy a plastic seat, or a fire survival requirement point to a metal seated valve, with the leak rate agreed against ISO 5208.

A datasheet line resolves into one of these combinations, and the sibling pages take each one further: trunnion, extended stem, three piece, LF2 body, double block and bleed, metal seated, LNG terminal duty and fully welded.

Typical delivery is 4 to 8 weeks depending on configuration and material availability. Dimensional reports, test records and material certificates are issued with the valve, and drawings are released before machining on any non-standard face-to-face or stem length.

Cryogenic Ball Valve FAQ

What temperature counts as cryogenic service for a ball valve?

Cryogenic service is generally taken as -50°C and below. Between -29°C and -50°C most specifications call the duty low temperature, and a valve with RPTFE or PCTFE seats and impact tested steel can be used. Below -50°C the seat, bonnet and test programme change, and BS 6364 covers valves down to -196°C. Below -196°C, helium duty, each case is engineered and tested to the project specification.

Why do standard ball valves fail in cryogenic service?

Five mechanisms act together. Moisture freezes on the cold stem and tears the packing. PTFE and RPTFE seats lose elasticity below -29°C and keep a permanent set. The seat contracts several times more than the stainless pocket, so preload disappears or the seat is crushed. Bolt preload at the body joint relaxes as the flange stack cools. Water left in the cavity freezes and expands. A valve that corrects only some of these still leaks.

Which seat material is used in a cryogenic ball valve?

PCTFE is the standard choice and works down to about -196°C. It keeps useful toughness at liquid nitrogen temperature, has low permeability and enough compressive strength to hold seat load. PTFE and RPTFE are limited to about -29°C and must not be offered for LNG at -162°C. PEEK reaches about -60°C. Above Class 900 the PCTFE seat is usually an insert in a metal carrier, and metal seated valves are used below -196°C.

What is the difference between a cryogenic ball valve and a low temperature ball valve?

Low temperature is the wider category, covering roughly -29°C to -50°C, where the design is a standard valve with impact tested steel and a suitable seat. Cryogenic is the subset at -50°C and below, where the bonnet is extended to protect the packing, the seat is spring loaded to survive contraction, the body joint is engineered for bolt relaxation, and the valve is proved by a cold test to BS 6364 rather than by an ambient test alone.

How is a cryogenic ball valve tested to BS 6364?

The valve gets a shell test at 1.5 times the 38°C rating and a seat test at 1.1 times, both at ambient. It is then chilled to the design temperature with liquid nitrogen or cold gas until body thermocouples stabilise, and the seat is tested with helium, with torque measured cold. Gland and body joint are probed. After warm-up the seat test is repeated to prove recovery, and the valve is examined for distortion.

What information do you need to quote a cryogenic ball valve?

Size, ASME class, design minimum temperature, fluid, end connection and face-to-face dimension, body material, seat material, bonnet or stem extension length, insulation thickness, whether the valve sits inside a cold box, actuation, test standard and certification, and quantity. Sour service, fire testing and NACE MR0175 should be stated at enquiry. A line class sheet or datasheet shortens the review considerably.