When Spring Energized Metal Seals Beat a Hollow Metal O-Ring

Spring energized metal seals beat a hollow metal O-ring when the joint must keep residual contact load after flanges move, bolts relax, or temperature cycles. A crushed hollow tube work-hardens and has little springback, so it needs a stiff, highly loaded joint and cannot replace a spring. Specify the jacket-plus-spring when recovery and cycling matter; keep the hollow O-ring when the joint is circular, stiff, and pressure-energized.

What Structural Difference Separates Spring Energized Metal Seals From a Hollow Metal O-Ring?

Spring energized metal seals differ from a hollow metal O-ring in where residual load comes from after bolt-up. The O-ring is a thin-wall tube that work-hardens as it is crushed; the spring-energized part is a metallic C-jacket plus an internal helical or coil spring that keeps contact force when the joint relaxes. System pressure can assist both, but it is not a substitute for a spring when flanges move.

That is a load-path question, not a silhouette question. Whether the jacket is a C-section or the tube is an O-section matters, but geometry C vs O is a different question from whether an internal spring is doing mechanical work after the joint unloads. This article stays on jacket-plus-spring versus hollow-tube recovery. Sonkit Sealing Solutions treats the two as distinct specifications: a hollow metal O-ring on the O-ring family, and a spring-energized Metal C-Ring on the C-ring family.

A hollow metal O-ring is a closed tube, usually circular in plan, seated in a matching groove and compressed axially until the outer surface and its plating form a continuous contact band. If the tube is vented, process pressure can enter the bore and add to the contact force. If the tube is unvented, the trapped gas or vacuum inside the bore still leaves the wall as the only elastic member. Either way, the wall is thin, the crush is plastic, and the remaining spring rate after installation is the leftover elasticity of a work-hardened tube.

A spring-energized metal C-ring is a two-part structure: a formed metallic jacket in a C cross-section, and a helical or coil spring captured inside that jacket. Bolt-up compresses the jacket; the spring stores load at a designed rate. After the jacket takes a set, the spring continues to push the plated contact lands against the flange faces. The open side of the C faces the system pressure so hydrostatic force adds to spring force, but the spring is still present when pressure is low, reversing, or absent during a thermal hold.

The practical difference shows up the first time the joint is not the joint that was torqued in the shop. Bolts relax. Flanges rotate under a bolt circle. Differential expansion opens a gap. Vibration walks residual stress out of a crushed tube. A hollow O-ring that sealed at assembly can lose line load and leak without any change in surface finish. A jacket-plus-spring design is specified for that second state — residual load after the hardware has moved — not for a prettier cross-section drawing.

The internal spring is not a marketing extra. It is the load path that remains after the jacket and the bolts have done what metal joints always do: relax.

Why Does a Hollow Metal O-Ring Need High Seating Load and Lose Recovery?

A hollow metal O-ring needs high seating load because the tube must be crushed far enough for plating to flow into surface valleys, and that crush plastically works the wall. Once the wall work-hardens, elastic recovery is small, so residual load falls if bolts relax, flanges rotate, or thermal growth opens the gap. System pressure can help a vented ring only while pressure is present and the joint stays stiff.

Brookhaven National Laboratory published hollow-metal-O-ring tests for AGS vacuum service that make the seating-load point concrete. Six variations were evaluated, in Inconel-X and Type 304 stainless, with indium or silver plating, at tube sections of 0.218 in OD × 0.016 in wall, 0.225 in × 0.020 in, and 0.250 in × 0.032 in. The most reliable seals in that campaign were Inconel-X, indium-plated rings; the sealing forces for those gaskets ranged from 223 to 402 lbf per linear inch (about 39 to 70 N/mm). That range sat below the AGS bolting capacity of about 600 lbf/in (about 105 N/mm). The note is on the BNL technical note on hollow metal O-ring vacuum seals.

Those figures are not a universal design table. They are a public data point that a hollow tube in a real vacuum flange needed hundreds of pounds per inch of circumference to seal, and that the laboratory treated available bolt load as a hard limit. A designer who copies a hollow O-ring onto a lighter flange, a softer land, or a joint that cannot put ~40–70 N/mm into the seal is asking the tube to seal without the crush the geometry requires.

Work-hardening is the recovery problem that follows that crush. Forming a thin-wall tube already cold-works the metal. Axial compression to seating height works it further. After plastic set, the unloading curve is steep and the elastic rebound is a small fraction of the installed deflection. The ring still occupies the groove, so a first-pass helium test can pass. Residual contact stress is now a function of how little the flanges later move. A stiff, highly loaded joint that never opens the gap can live with that. A joint that sees bake-out, thermal shock, or bolt-load scatter cannot.

Pressure-energization is real and still not a spring. A vent hole lets system pressure into the tube so hydrostatic force adds to wall force at the contact band. That assist scales with pressure and disappears when the line is vented, during a pump-down, or at the low-pressure end of a cycle. NASA’s high-temperature metallic-seal work for aero propulsion treated spring design and seating load as first-order variables for exactly this reason: a high-temperature static joint cannot count on a single crush event to hold contact through thermal transients. The workshop paper is on the NASA Technical Reports Server (high-temperature metallic seal development).

Plating does not restore springback. Silver, gold, nickel, or PTFE on a hollow O-ring improves microscopic conformity so asperities close at a given load. Soft overlay does not change the plastic set of the tube wall. NASA flange-design notes on self-equalizing metal O-rings already required a soft-metal plating and a fine mating finish; that is an interface rule, not a recovery rule. If residual load is the leak risk, plating is necessary and still insufficient.

When Do Spring Energized Metal Seals Win on Thermal Cycling, Vibration, and Flange Rotation?

Spring energized metal seals win on thermal cycling, vibration, and modest flange rotation because the internal spring keeps residual contact load after the jacket and the bolts have relaxed. A hollow tube cannot restore that load once it has work-hardened. The spring is worth the extra part when the joint will see temperature swings, vibration, lower available bolt load, or flange rotation that would unload a crushed O-ring.

The spring alloy is the reason that residual load can be specified rather than hoped for. INCONEL alloy 718 (UNS N07718 / W.Nr. 2.4668) is a precipitation-hardenable nickel-chromium alloy documented for service from −423 °F to 1300 °F (−253 °C to 704 °C). AMS 5662 / 5663 minima for bars, forgings, and rings include 150 ksi (1034 MPa) 0.2% yield at room temperature and 125 ksi (862 MPa) at 1200 °F (649 °C). The same bulletin records relaxation-resistant helical-spring use to 1100 °F (593 °C). Those figures describe the spring metal, not a leak rate. They are published in the Special Metals INCONEL alloy 718 bulletin.

Bar and forging product used for high-temperature springs is covered by ASTM B637-26, the specification for precipitation-hardening and cold-worked nickel-alloy bars, forgings, and forging stock for moderate or high-temperature service, including tension, hardness, and stress-rupture testing. Specifying 718 for the spring is a materials-control decision: keep elastic energy in the coil after the jacket has taken a set, through bake-out or hot-process temperature, without treating the alloy grade as a sealing slogan.

The jacket is a different duty. It has to form, take plating, resist the process media, and present two contact lands. Typical jacket alloys are 304/316 stainless, INCONEL alloy 600, alloy 718, and Alloy 625. INCONEL alloy 625 (UNS N06625 / W.Nr. 2.4856) is a nickel-chromium-molybdenum-niobium alloy used from cryogenic temperature to 1800 °F (982 °C); strength comes from molybdenum and niobium stiffening of the matrix, so precipitation hardening is not required. Special Metals lists springs, seals, and bellows among potential uses. Jacket selection is therefore corrosion, temperature, and formability; spring selection is residual load. Mixing the two into one “nickel alloy seal” line on a drawing hides the split. The 625 data are in the Special Metals INCONEL alloy 625 bulletin.

Four joint conditions make that split pay for itself:

• Thermal cycling. Flange, bolt, and seal metals expand at different rates. The gap at the contact band opens and closes. A work-hardened tube follows the opening only by the small elastic leftover after crush. A helical spring follows the opening along a designed load-deflection curve, so contact stress remains above the plating’s sealing threshold through the cycle rather than only at cold assembly.
• Vibration and mechanical shock. High-frequency unloading walks residual stress out of a plastically set tube. The spring continues to press the jacket lands. This is still a static seal — the spring is not a dynamic wear part — but the joint is not a dead crush either.
• Lower available bolt load. Light flanges, limited bolt circles, and hardware that cannot put hollow-O-ring crush into the land still need metallic contact. The spring rate can be chosen so seating load is high enough to flow plating and low enough that the flange does not rotate or yield. A hollow tube’s seating load is mostly a consequence of wall thickness and crush; it is a poor knob when bolt load is the constraint.
• Modest flange rotation. A bolt circle outside the seal diameter bends the flange. The inner land unloads. Spring-energized metal seals keep a residual line load on that inner land; a crushed O-ring that has already taken a set does not climb back into contact.

Pressure still helps. The open C faces the system pressure, so hydrostatic force adds to spring force as pressure rises. The difference versus a vented hollow O-ring is the low-pressure and transient end of the same duty: pump-down, start-up, cooldown, and reversing pressure. The spring is working when the hydrostatic term is small. That is the condition in which specifying only a hollow tube is a bet on joint stiffness that the hardware may not keep.

When Does a Hollow Metal O-Ring Still Beat the Internal Spring?

A hollow metal O-ring still wins when the joint is stiff, highly loaded, circular, and driven by high system pressure that can re-energize a vented tube. In that hardware the extra spring is unused mass: the flange does not move enough to need recovery, and the high seating load is already available from the bolt circle. Simple circular grooves and high-pressure static duty are the O-ring’s home ground.

High system pressure is the first honest win for the tube. A vented hollow O-ring turns process pressure into contact stress around a closed circumference. As pressure rises, the assist rises. A designer who already has a heavy flange, a full bolt circle, and a pressure that stays high in service is buying a simple part that uses the process as the energizer. Adding a helical spring inside a C-jacket does not create pressure that is not already there; it adds a part, a spring-rate tolerance, and a jacket-forming step. On a stiff, hot, high-pressure static joint that never opens the gap, that extra part is not doing recovery work.

Simple circular grooves are the second win. A hollow O-ring sits in a straightforward circular gland. Groove width and depth follow tube OD and designed crush. Non-circular, racetrack, or interrupted glands are a different manufacturing problem; for a round static face with a conventional groove, the tube is the least fussy metal ring to install. A spring-energized C-ring still needs a controlled free height, a designed compression, and orientation of the open C toward pressure. That is justified when recovery is required. It is overhead when the groove is a simple circle and the joint will not move.

Joint stiffness is the third win, and it is the one engineers skip. ASME PCC-1-2022 is the standard for assembling pressure-boundary flanged joints with ring-type gaskets that sit entirely inside the bolt circle. It exists because leak-tightness after assembly depends on gasket seating, operating bolt load, flange condition, and assembler practice — not on naming a gasket family. A hollow metal O-ring that was crushed correctly in a rigid pair of flanges, with bolt load held through operation, is doing what that assembly discipline is for. If the joint is that joint, the spring is a solution to a movement problem the hardware does not have.

The hollow O-ring also remains the right call when the failure mode is “never enough crush,” not “lost recovery.” Rough lands, thick plating that needs more flow, or a first-closure requirement on a one-shot high-pressure vessel can favor a tube that is meant to take a large plastic set. Spring-energized metal seals are not a way to avoid seating load altogether. They change who supplies load after seating. If the land will not seal until a high line load has flowed metal into every valley, and the flange can take that load, a hollow O-ring is still the direct tool.

What the hollow O-ring should not be asked to do is impersonate a spring. “We vented it, so it is energized” is true only while pressure is up and the gap is stable. “We plated it, so it will recover” confuses interface conformity with structural rebound. “We used 718 tube, so it is a high-temperature spring” confuses alloy capability with geometry: 718 in a crushed tube still work-hardens; 718 in a helical spring is the product the Special Metals bulletin actually documents for relaxation resistance to 1100 °F (593 °C).

How Should Engineers Choose Spring Energized Metal Seals Versus a Hollow Metal O-Ring?

Engineers should choose spring energized metal seals versus a hollow metal O-ring by writing flange stiffness, expected gap change, available bolt load, cycle count, and groove family before naming a profile. If residual load after relaxation is the leak risk, specify the jacket-plus-spring. If the joint stays stiff and highly loaded under pressure, specify the hollow tube. Do not treat catalog family names as a substitute for that load-path check.

The comparison below is a selection frame, not a catalog. Load numbers other than the BNL hollow-O-ring seating-force range are omitted because they are application-specific. Alloy figures are from the Special Metals bulletins and ASTM B637 cited above. A Metal C-Ring seal without an internal spring is a third geometry and is not the subject of this table.

Use the following checklist as the minimum data set before choosing a profile. The Metal Seal Design Guide is the place to put the same inputs in a standard form; the metal ring seal products index is the map of families once the load path is decided.

1. Flange stiffness and rotation. Bolt circle, flange thickness, modulus, and whether the inner land unloads when the bolts are pulled up. If the flange rotates, residual recovery is a requirement, not a preference.
2. Expected gap change in service. Bake-out or process ΔT, CTE stack of flange, bolts, and seal, and any mechanical separation from pressure end-load. If the gap will grow, a crushed tube is the wrong elastic member.
3. Available bolt load versus seating need. Compare the bolt-circle capacity with the crush a hollow O-ring actually needs. BNL’s 223–402 lbf/in band is a reminder that hollow-tube seating is not a light load. If the hardware cannot supply it, do not specify the tube.
4. Pressure history, not just design pressure. Peak, hold, vent, pump-down, and reversing pressure. Hydrostatic assist is only present while pressure is present.
5. Cycle count and vibration. One-shot closure on a stiff vessel is hollow-O-ring territory. Repeated thermal cycles, hot-cold swings, or vibration on the joint favor the internal spring.
6. Groove family. Circular, simple, and already cut for a tube is a reason to stay on the O-ring if the joint is stiff. A new gland can be cut for a spring-energized C-ring; do not drop a jacket-plus-spring into an O-ring groove and call it equivalent.
7. Jacket alloy, spring alloy, and plating as separate lines. Media and temperature set the jacket (304/316, 600, 625, 718). Residual load at temperature sets the spring (718 is the documented high-temperature spring alloy in the Special Metals bulletin). Plating (silver, gold, nickel, PTFE) is the interface layer — specify it, do not use it as a substitute for spring rate. Thickness belongs on a drawing, not as an invented catalog micron value.
8. Assembly discipline. PCC-1 exists because seating load at assembly and operating load later are different states. Write both. A hollow O-ring that is under-crushed will leak immediately; a spring-energized ring that is over-compressed can take a jacket set that the spring was not sized to cover.

When those eight inputs are on the requisition, the choice is usually obvious. Residual load after a moving joint: jacket plus spring. Stiff, circular, high-pressure, highly loaded static face: hollow tube. Ambiguous hardware is an engineering review, not a coin flip between two product photos.

Frequently Asked Questions

These questions cover the load-path decisions that decide a spring-energized metal seal versus a hollow metal O-ring: residual recovery, seating load, cycling, and when the extra spring is unused. Direct answers come first. Geometry C versus O is not re-argued here. Alloy, plating, and groove inputs belong in the requisition, not in a catalog slogan.

When should I specify spring energized metal seals instead of a hollow metal O-ring?

Specify spring energized metal seals when residual contact load must survive thermal cycling, vibration, bolt relaxation, lower available bolt load, or modest flange rotation. A hollow tube that has work-hardened cannot restore that load. Keep the hollow O-ring for a stiff, circular, highly loaded joint that stays under high system pressure.

Does system pressure make a hollow metal O-ring act like a spring?

No. System pressure can add hydrostatic force to a vented tube, but that assist exists only while pressure is present and the gap is stable. A spring is a designed elastic member that still pushes the contact lands during pump-down, cooldown, and reversing pressure. Vented is not the same as spring-energized.

Why does a hollow metal O-ring need a higher seating load?

The tube must be crushed far enough for plating to flow into surface valleys, and that crush is plastic. Brookhaven’s AGS vacuum tests recorded 223–402 lbf/in (39–70 N/mm) for the Inconel-X indium-plated rings that sealed most reliably — a public reminder that hollow-tube seating is a high line load, not a light “metal o ring” push-fit.

Which jacket and spring alloys are typical in spring energized metal seals?

Jackets are commonly 304/316, INCONEL alloy 600, alloy 718, or Alloy 625; the internal spring is often alloy 718 when residual load must be held at temperature. Special Metals documents 718 helical-spring relaxation resistance to 1100 °F (593 °C) and 625 service from cryogenic temperature to 1800 °F (982 °C). Plating is a separate specification: silver, gold, nickel, or PTFE, without treating overlay thickness as a published universal micron value.

Can I drop a spring-energized C-ring into an existing hollow O-ring groove?

No. Groove depth, width, free height, designed compression, and C-opening orientation are not interchangeable with tube OD and crush. A jacket-plus-spring ring that is over-compressed or facing the wrong pressure direction will not deliver the residual-load behavior it was chosen for. Cut the gland for the profile you are specifying.

What should I send Sonkit to choose between the two?

Send the flange and groove drawing, bolt pattern and load limit, pressure and temperature history including vents and cycles, media, required plating constraints, and whether the joint is expected to rotate or open in service. Those inputs let Sonkit Sealing Solutions place the hardware on the hollow-tube path or the jacket-plus-spring path instead of quoting both as equivalent catalog items.