Views: 0 Author: J-VALVES Publish Time: 2026-10-15 Origin: Site
The useful divisions are not the words "low temperature" and "cryogenic." They are these five lines:
−29 °C. The minimum design temperature for carbon steel (ASME B16.34 material Group 1.1). Above it, standard carbon steel needs no impact testing. Below it, Charpy V-notch impact testing at the minimum design metal temperature (MDMT) is required, with absorbed energy per the applicable standard.
−46 °C. The floor for low-temperature carbon steel castings. In ASTM A352 the usual minimum impact test temperatures are −32 °C for LCA and −46 °C for LCB and LCC. This is also the customary industry threshold for "low-temperature service"; MSS SP-134 uses the −46 °C / −50 °F order of magnitude as its entry point.
−101 °C. The usual minimum impact test temperature for 3.5 % nickel steel (ASTM A352 LC3) — the ethylene band (ethylene is around −101 °C).
−162 °C. Atmospheric LNG. This is past every ferritic and martensitic low-temperature casting grade and into austenitic stainless or 9 % nickel steel territory.
−196 °C. Liquid nitrogen, with liquid oxygen at −183 °C and liquid argon at −186 °C nearby. It is also the common lower bound of cryogenic valve standards and test media: BS 6364 and ISO 28921-1/-2 all cover −50 °C to −196 °C. Within ASTM A352, only LC9 (9 % Ni) carries a −196 °C usual test temperature.
Temperature band | Acceptable pressure-boundary materials (examples) | Explicitly not acceptable | Typical service |
Above −29 °C | ASTM A216 WCB; A105 forgings | Carbon steel used below −29 °C without impact testing | Standard process, ambient tank farms |
−46 °C to −29 °C | A352 LCB / LCC; A350 LF2 forgings (impact tested at MDMT) | Untested WCB | Propylene refrigeration (≈ −46 °C), low-temperature separation |
−101 °C to −46 °C | A352 LC3 (3.5 % Ni); austenitic stainless as a more conservative choice | LCB / LCC | Ethylene (≈ −101 °C), cryogenic separation |
−162 °C | A351 CF8M / CF3M; A182 F316 / F316L forgings; 9 % Ni steel (mainly tanks and some equipment) | Every ferritic / martensitic low-temperature casting grade, including LC3 and LC4 | LNG liquefaction, storage, loading, terminals, BOG |
−196 °C | Austenitic stainless (CF8M / CF3M / F316L); A352 LC9 | Carbon and low-nickel steels; free-machining sulfur grades such as 303 | LIN, LOX, LAR, air separation |
Boundary 1 — the misconception worth correcting: LC3 is not "the LNG standard."
A lot of selection literature describes ASTM A352 LC3 as the preferred material for LNG main lines, and it is common to hear that LC3 is the default valve body for large LNG projects. By the standard itself, LC3's usual minimum impact test temperature is −101 °C; it is a 3.5 % nickel steel positioned at the ethylene level, not the methane level. At −162 °C, the pressure boundary belongs to austenitic stainless (CF8M / CF3M / F316L) or 9 % nickel steel.
Specifying LC3 at −162 °C means using a low-temperature steel beyond its own impact test floor. That sentence belongs in the technical agreement, not in a brochure.
Boundary 2 — austenitic stainless is not exempt from testing.
ASTM A352 Note 1 is explicit: certain A351 austenitic casting grades have been found suitable down to −184 °C, and others down to −254 °C, when impact tested in accordance with A352 with energy levels and test temperatures mutually agreed between purchaser and manufacturer. In other words, austenitic suitability still needs impact data behind it — "austenitic has no ductile-to-brittle transition, so it must be fine" is not a qualification. The same heat, a different heat treatment or a heavier wall section, can behave differently.
On a cryogenic ball valve, "material" is never just the body. The parts that fail first are usually these, and they are rarely listed in the materials column of a purchase spec:
Bolting. Pressure-boundary bolting should be a low-temperature grade — ASTM A320 L7 with A194 Gr.4 nuts, for example. Carrying ambient-service A193 B7 / B7M over to −162 °C is a common error. Note also the differential contraction: austenitic stainless has a coefficient of thermal expansion around 16–17 × 10⁻⁶/K against roughly 11–12 × 10⁻⁶/K for carbon steel. Between assembly temperature and cryogenic service that difference shows up as bolt preload loss. Long-neck, large-flange assemblies should use disc-spring compensation or tensioned bolting, and be re-torqued after warm-up.
Stem. Use austenitic stainless forgings (A182 F316 / F304 class) with an anti-blowout design. Precipitation-hardening grades such as 17-4PH, even in the H1150 double-age condition, have substantially lower low-temperature impact toughness than austenitic stainless and are not a first choice for cryogenic stems or balls — unless the supplier can show impact test data at the design minimum temperature.
Sulfur-bearing free-machining stainless is excluded. Grades such as 303, with sulfide inclusions creating notch sensitivity, are not acceptable for cryogenic bodies or trim.
Packing and gaskets. Flexible graphite or PTFE-based packing is typical. The packing box has to hold through cool-down, steady cold operation and warm-up — which is the direct reason the extended bonnet exists (Section 4).
Documentation goes in the contract too: body, bonnet (including the extension) and pressure-boundary parts should come with EN 10204 3.1 or 3.2 certification, and the impact test temperature on that certificate must not be higher than the design minimum temperature.
At cryogenic temperature the sealing interface has to survive two things at once: unequal contraction of the components, and the change in mechanical behaviour of a polymer seat. A floating-ball design that relies on line pressure to push the ball into the seat cannot build enough contact stress in a low-pressure cryogenic line (tank-top BOG, truck-loading return lines, for example). That is why large-bore, high-differential cryogenic ball valves are normally trunnion-mounted with spring (disc-spring) energized seats: the spring holds sealing at low differential, and line pressure supplements it at high differential.
The classic error here is preload calculated at ambient. From +30 °C to −162 °C, the contraction difference between ball, seat ring and body changes contact stress along the sealing band directly. If nothing in the assembly absorbs that displacement, contact stress drains away in one direction only.
Seat material | Usable range (typical) | Ambient leakage capability | Main failure mode and limits |
PCTFE (Kel-F / CTFE) | −196 °C up to about +80 to +120 °C | Can achieve ISO 5208 Rate A (no visible leakage) | The default polymer seat for cryogenic service; the ceiling is low, so it fails first on the warm side — warm-up sections, BOG discharge, flare lines. Use the specific grade datasheet |
Virgin PTFE | Lower limit around −73 °C (−100 °F) | Rate A | Below the usual limit requires justification; cryogenic composite / reinforced PTFE seats are offered to −196 °C, but creep and rebound data must come from the supplier's testing |
PEEK and modified PEEK | About −150 °C to +250 °C (literature varies) | Rate A | Better wear and high-temperature performance than PCTFE, but low-temperature rebound and resulting torque must be verified separately; strongly filler-dependent |
Spring-energized PTFE seat | −196 °C to +200 °C | Rate A | Common LNG solution: preload compensates contraction. The failure mode is preload decay over cycles, not the material |
Metal-to-metal hard seat | −196 °C into the high-temperature range (limited by body material) | Typically ISO 5208 Rate D (0.1 × DN, mm³/s) | Wear-resistant and fire-safe, but leakage class is below soft-seated designs; demands tighter sealing-surface geometry and thermal contraction compensation, and higher operating torque |
Boundary 3 — with a soft seat on flammable service, anti-static bonding and fire performance are gates, not options.
On LNG, ethylene and similar flammable media, a soft-seated ball valve needs a conductive path (the industry norm is ≤10 Ω across body–ball–stem) plus demonstrated fire performance (API 607 / ISO 10497 class testing). A polymer seat loses its elasticity early in a fire; what actually provides fire-safe sealing is the metal back-up seat and the body/seat-ring geometry, not the polymer itself.
The extended bonnet is not cosmetic insulation. Its engineering purpose is to create a vapour column between the body and the stuffing box, moving the packing and the operator out of the cold zone and keeping the stuffing box above 0 °C. Field acceptance commonly uses something like ≥5 °C at the packing-box wall. Once that temperature goes below 0 °C, condensed moisture freezes between packing and stem: torque climbs, then the stem seizes.
The length cannot be guessed. MSS SP-134 (current edition MSS SP-134-2025, the third revision, replacing 2012; first published 2005) covers gate, globe, ball, butterfly and check valves from −73 °C to −254 °C, and sets minimum extension lengths separately by valve type (rising-stem versus quarter-turn) and by service type (cold-box liquid, cold-box gas, non-cold-box liquid and so on). Citing the commonly reproduced Table 1 values: NPS 4 (DN100) in cold-box liquid service is about 26 in / 650 mm, appreciably longer than non-cold-box service. Extensions may be cast, forged or made from seamless tube, but wall thickness must satisfy the ASME B16.34 requirement for the applicable class; if the extension material differs from the body, thickness must be compensated proportionally.
Criterion / parameter | Requirement or typical value | Engineering consequence |
Packing-box temperature | Above 0 °C; ≥5 °C commonly used as a witnessed criterion | Below freezing, the stem seizes and the packing fails |
Minimum extension length | Per MSS SP-134 Table 1 by valve type and service type (e.g. NPS 4 cold-box liquid ≈ 650 mm) | Do not shorten to suit space; a shortened bonnet falls outside the qualified design |
Installation orientation | Stem essentially upward (manufacturers typically allow about ±45° from vertical) | The vapour column depends on orientation; a horizontal stem defeats the whole purpose |
Insulation / cold box | The upper extension and packing box must not be wrapped in insulation | Insulating the extension conducts cold straight to the packing |
Stem stability | Check buckling by slenderness ratio (MSS SP-134 Annex X1 gives the method; safety factor 2); quarter-turn stem torsional deflection usually limited to 2° | A long extension changes the stem load case; large valves must be re-checked |
Stem-to-extension bore clearance | Keep as small as practical | Excess clearance opens a convection path and shortens the effective vapour column |
Read together, the three boundaries above produce six chains worth knowing, each with an observable symptom:
Frost line in the wrong place → extension length or insulation practice is wrong. In normal operation there should be a stable cold-to-warm transition. If frost reaches the packing box or the operator, the vapour column is not holding the packing above 0 °C. The fix is to review extension length and orientation — not to add insulation.
Cavity overpressure → no relief path, or it is installed backwards. Liquid trapped in a ball cavity vaporises as the valve warms, expanding by a factor on the order of several hundred (LNG is commonly taken as about 600:1 liquid-to-gas). Without a relief path, cavity pressure rapidly exceeds the shell design capability. Cryogenic ball valves therefore use a drilled ball or self-relieving seats (SPE) — which makes the valve directional. It must be installed following the flow arrow, with the relief side facing the specified pressure side. This is the most common and most damaging field error in cryogenic plants.
Frozen, seized stem → water ingress at the packing, or an insulated extension. The signature is abnormally high torque when cold that recovers on warm-up, unlike the sustained high torque from seat wear.
Internal leak after warm-up → seat permanent deformation or preload decay. Tight when cold, leaking after warm-up points at permanent deformation (cold flow) of the polymer seat and loss of disc-spring preload — not at scoring on the sealing surface. It gets worse after repeated thermal cycles, which is why high-cycle duty such as loading arms should have post-cycling sealing stability written into the specification.
Leak at a flange or body joint → lost bolt preload. Differential contraction between dissimilar materials (austenitic stainless against carbon steel) reduces preload, and long-neck geometry amplifies it. Countermeasures are disc-spring compensation, re-torquing after warm-up, and welded joints where feasible.
Static charge accumulation → no conductive path. A soft seat electrically isolates ball from body; friction-generated charge is an ignition source on flammable media. The anti-static device (a conductive spring, on the order of ≤10 Ω) is a requirement, not an option.
Cryogenic ball valve testing is two-stage. The ambient stage runs to API 598 (11th edition, 2023) or ISO 5208: shell at 1.5× the 38 °C rating, seat at 1.1× — the same multipliers as API 6D. Plenty of valves ship after this stage alone. The cryogenic stage is a different exercise.
Two verification routes dominate. BS 6364 is the document most often made mandatory in LNG project valve specifications. It covers −50 °C to −196 °C, applies from DN15 upward, and requires an extended bonnet / packing box design, impact testing of materials at the minimum temperature, and verification of sealing and operation at liquid nitrogen temperature; its typical pressure values are again 1.5× shell and 1.1× seat, with helium used as the leak-detection medium. It is an older document; the international counterpart is the ISO 28921 family. ISO 28921-1 (design, manufacturing and production testing) covers −50 °C to −196 °C, DN10–DN900 (NPS 3/8–36), PN16 to PN250 and Class 150–1500; its 2013 edition was superseded by the 2022 edition. ISO 28921-2 (type testing), current as the 2015 edition, also spans −50 °C to −196 °C and explicitly does not apply to cryogenic vessel valves designed to ISO 21011. The routes differ in kind: BS 6364 is per-valve cryogenic testing before shipment; ISO 28921-2 is a type-approval regime (first-article testing plus periodic follow-up sampling). Choose the route the owner's specification names — do not "upgrade" the standard number on your own initiative.
Stage | Content | Notes and frequent omissions |
Ambient pressure test | Shell 1.5 × 38 °C rating; seat 1.1 × rating (API 598 / ISO 5208) | A baseline, not evidence of cryogenic suitability. Test medium should be dry and oil-free; residual moisture freezes during later cryogenic testing and damages internals |
Materials and documentation | EN 10204 3.1 / 3.2 certificates; impact test temperature ≤ design minimum | Impact acceptance should be set jointly by the material standard and the project specification (a common practice is ≥27 J average of three specimens and ≥20 J minimum individual). Do not treat a single catalogue number as a universal acceptance line |
Cryogenic treatment (process) | GB/T 24925-2019 explicitly requires cryogenic treatment for austenitic stainless valves serving below −100 °C | Austenitic stainless can partially transform to martensite at low temperature, with dimensional change. Stabilisation is normally scheduled between semi-finish and finish machining to keep sealing-surface distortion acceptable |
Cool-down and soak | Cool to target temperature with liquid nitrogen and stabilise | Cooling rate must be controlled (industry practice commonly limits it to tens of °C/min). Liquid level must cover the body and the body-to-bonnet joint, with thermocouples confirming uniform temperature before testing begins |
Cold operation and leak test | Full open/close cycles while cold; leak test at seat and stem (helium commonly used) | Acceptance limits come from the project specification (e.g. volumetric leakage per DN for metal seats; no visible leakage for soft seats). Helium is more sensitive than air or water and finds micro-paths an ambient test never sees |
Warm-up re-test | Return to ambient and re-test sealing and operating torque against pre-cool-down data | The step most often skipped. Tight when cold and leaking after warm-up is the characteristic cryogenic failure; skipping the warm-up re-test abandons the only verification of seat rebound and bolt preload |
Boundary 4 — standards have scope limits; do not cite them out of range.
EN 1626 covers cryogenic vessel valves only, up to DN150 and oriented to vacuum-jacketed construction; large-bore full-port cryogenic ball valves for LNG loading and terminal lines fall outside it. ISO 28921-2 explicitly excludes cryogenic vessel valves designed to ISO 21011. Citing an inapplicable standard number may pass unnoticed at bid stage and become a non-conformance at third-party witness or owner review.
If | Then |
Minimum design temperature above −29 °C, non-corrosive service | WCB / A105 carbon steel body is acceptable, with the pressure class checked at temperature per ASME B16.34 |
Minimum design temperature −46 °C to −29 °C (e.g. propylene refrigeration) | ASTM A352 LCB / LCC or A350 LF2, with impact test records at MDMT |
Minimum design temperature −101 °C to −46 °C (e.g. ethylene) | ASTM A352 LC3 (3.5 % Ni); if the stream is also corrosive, go directly to austenitic stainless |
Minimum design temperature −162 °C (LNG) | A351 CF8M / CF3M or A182 F316 / F316L pressure boundary. Do not use LC3 / LC4 as an "LNG default" |
Minimum design temperature to −196 °C (LIN, LOX, LAR) | Austenitic stainless or A352 LC9; oxygen service additionally requires oxygen cleaning (ASTM G93 class) |
Bore ≥ DN200, or significant differential pressure | Switch to a trunnion-mounted design with spring (disc-spring) energized seats, and re-calculate preload for the cold contracted geometry |
Small bore, low differential, very low torque required (tank-top sampling, return lines) | Floating ball with a PCTFE seat is viable, provided sealing contact stress can be established at the low differential |
Flammable medium with a soft seat | Anti-static conductive path (≤10 Ω order) is mandatory, with API 607 / ISO 10497 fire test basis |
Valve has an extended bonnet | Take the minimum length from MSS SP-134 Table 1 (cold box versus non-cold box, and service type); check stem buckling by slenderness ratio (safety factor 2); respect the manufacturer's permitted tilt angle |
Space is tight and someone suggests shortening the extension | Do not accept the shortening. Re-route the piping or change the valve type — the shortened valve is outside the qualified design envelope |
Q1. Can I use LCB or LC3 for the body of an LNG ball valve at −162 °C?
No — LC3 is unsuitable and LCB is worse. In ASTM A352 the usual minimum impact test temperatures are −46 °C for LCB and LCC and −101 °C for LC3, both above the −162 °C LNG condition. At −162 °C the pressure boundary should be austenitic stainless (A351 CF8M / CF3M, A182 F316 / F316L) or 9 % nickel steel. Treating LC3 as the "LNG standard" is a widespread misconception; the technical agreement should simply state that the impact test temperature of the material is not higher than the design minimum temperature.
Q2. Does a cryogenic ball valve always need an extended bonnet, and how long?
Whenever the service temperature would take the stuffing box below 0 °C, yes — the purpose is to form a vapour column between body and stuffing box so that packing and operator stay above 0 °C (a witnessed criterion of ≥5 °C is common). The length should not be estimated: use the minimum values in MSS SP-134 Table 1, taken by valve type and service type — for example NPS 4 (DN100) in cold-box liquid service at about 26 in / 650 mm. The extension must not be wrapped in insulation, the stem should be essentially upward (typically within about ±45° of vertical), and the stem must be checked for buckling by slenderness ratio.
Q3. Why is the cavity relief hole directional?
Because the cavity between the seats is a closed volume. Trapped LNG that absorbs heat vaporises and expands by a factor of several hundred (about 600:1 liquid-to-gas is the usual planning figure), and cavity pressure can exceed shell design capability in a short time. Cryogenic ball valves therefore provide a drilled ball or self-relieving seats to connect the cavity to one designated side. That path is directional: installation must follow the body flow arrow, with the relief side facing the side the design specifies. Installed backwards, the valve looks normal at ambient and becomes a pressurisation hazard the moment it is cooled down.
Q4. Can a cryogenic ball valve use PTFE seats?
It depends. Virgin PTFE's usual lower limit is around −73 °C (−100 °F), so service at −162 °C needs more evidence. PTFE-based composite and spring-energized PTFE seats do exist for cryogenic service, but their suitability depends on the specific formulation, the preload arrangement and the number of thermal cycles — it must rest on the supplier's rebound and leakage test data at the design minimum temperature. PCTFE (Kel-F) is the polymer more commonly specified in cryogenic service, usable to −196 °C, at the cost of a distinctly lower continuous service ceiling (roughly +80 to +120 °C, varying by grade) — which is why it is unsuitable for warm-up sections or hot-side duty.
Q5. If the valve passes all ambient tests, is it proven at −162 °C?
No. Ambient testing (1.5× shell, 1.1× seat) proves the valve is a sound industrial valve at room temperature. It says nothing about what happens after cool-down: components contract by different amounts, clearances change, and a sealing interface that fits well at ambient can open a path when cold. Cryogenic valves therefore need separate low-temperature testing — either the per-valve BS 6364 route or the first-article-plus-sampling ISO 28921-2 route — and that testing must include a warm-up re-test, because tight-when-cold, leaking-after-warm-up is the characteristic cryogenic failure mode.
Q6. Can I insulate the extended bonnet section to stop it frosting?
Do not wrap the upper extension and the packing box in insulation. The extension works by letting a vapour column form between the cold end and the stuffing box and keeping the packing side above 0 °C by natural convection. Insulating that section removes the convective path and conducts cold straight to the packing. Insulation should stop at the specified interface; the extension is left bare or treated for condensation per the manufacturer's instruction. Frost itself is not the problem — frost reaching the stuffing box is.
The materials line is only the first row of a cryogenic ball valve specification. A complete technical requirement covers five groups of parameters: design minimum temperature and the corresponding temperature band; pressure-boundary material plus its impact test evidence at MDMT; seat material with its temperature ceiling, along with the preload arrangement and its ability to hold contact stress through thermal cycling; extension length, installation orientation and stem buckling check; and the test route (per-valve cryogenic test or type test), leak-detection medium, and warm-up re-test requirement.
Leave any one of those five out, and the valve is perfectly sound in the warehouse. It only becomes a problem after the first cool-down — and it becomes a problem not because of the alloy, but because of the multipliers, lengths and directions that nobody checked and that decide the mechanics of the seal and the temperature distribution in the neck.
