A metal bellows seal removes the dynamic secondary sealing element because the welded corrugated bellows performs both spring-loading and axial compensation as a single static-mounted metal component, making the sliding elastomer O-ring used in pusher seals redundant [S3][S6].
In a conventional pusher-type mechanical seal, a secondary elastomer (typically an O-ring or wedge) slides along the shaft or sleeve to transmit spring force and accommodate axial face movement; the bellows design folds that function into the metal convolutions, so no elastomer travels with the shaft [S2][S3].
What the Dynamic Secondary Seal Does in a Pusher Seal
In a pusher-type mechanical seal, a dynamic secondary seal (most commonly an elastomer O-ring, chevron, or U-cup) is the only component that can move axially with the seal face while remaining sealed against the shaft or sleeve [S4][S5].
It carries two jobs simultaneously: transmit spring force to the rotating face, and accommodate the small axial excursions caused by thermal growth, pressure fluctuation, and face wear [S4]. Because it slides on a metal counterface every time the seal face opens or closes, the O-ring is the part most exposed to fretting, hang-up, chemical attack, and solid build-up [S2][S3].
Industry data referenced for non-pusher bellows designs attributes roughly three-quarters of early-life seal failures in moving-parts service to friction at these dynamic secondary interfaces, a number that is widely cited but rarely re-validated in the open literature (2024-11) [S2].
How a Metal Bellows Replaces That Function
A metal bellows is a stack of edge-welded or formed convolutions, usually 8 to 12 plies in modern designs, that compresses and extends axially like a spring while sealing statically at both ends [S1][S3].
Because the bellows itself supplies the spring force, the elastomer is demoted to a static gasket (often a PTFE wedge or graphite ring) that never moves against the shaft [S2][S3]. The shaft therefore sees only a single rotating metal-to-face interface; there is no sliding polymer interface to wear, fret, or hang up [S3][S6].
The pressure-balanced variant described in US5662335A goes further: an internal smaller bellows bottoms out under reverse pressure, leaving only the correctly sized active bellows to carry hydraulic load, a design pattern that keeps the secondary seal zone truly static across pressure reversals [S8].
Failure Modes Eliminated by Removing the Sliding Elastomer
Removing the dynamic secondary seal eliminates five recurring failure mechanisms: O-ring hang-up on solid-laden shaft surfaces, shaft/sleeve fretting corrosion under the O-ring, chemical attack on the elastomer, cold-flow extrusion into the shaft gap, and stick-slip chatter driven by O-ring friction (2025-08) [S3][S5][S6].
The hang-up failure is particularly common in paper-stock, refinery, and crystallizing-chemical services, where a build-up on the shaft can clamp the O-ring and prevent the face from closing, causing catastrophic leakage [S3][S6]. With a bellows design, the shaft is clean metal all the way to the face, so there is no elastomer land for solids to grip [S6].
For a deeper look at how face materials and surface finish drive leakage rates on these seals, the mechanical seal reference covers face combinations down to sub-1 µm Ra finishes and 0.25 µm fluid film thickness, both of which matter more once the O-ring damping is removed [S2].
Material and Design Trade-offs That Come With the Trade
Replacing a sliding elastomer with a metal bellows trades O-ring damping for a much stiffer, fatigue-limited spring, so bellows seals are more sensitive to misalignment, runout, and pressure spikes than pusher designs (2025-08) [S2][S5].
Common bellows alloys and their published operating envelopes:
316L stainless steel: suited to chloride service below roughly 5,000 ppm Cl- at temperatures near 200 °F; widely used for general chemical duty [S2].
Alloy 276 (UNS N10276): AESSEAL's standard bellows material, with corrosion resistance above 90% across a wide pH range, and superior mechanical properties to 300-series stainless or Alloy 20 [S2][S3].
Inconel 718: used in high-temperature hydrocarbon service, retaining structural integrity up to about 800 °F per NACE-cited work (2024-11) [S2].
Convolutions must be sized for the full axial travel plus pressure end-load; common industrial cartridges use 12 convolutions versus an 8-ply historical baseline, dropping per-ply stress and lengthening fatigue life [S3]. Designs that cannot accept the bellows stiffness, for example very high runout or heavy thermal cycling, may still need a pusher seal with its dynamic secondary [S2].
Where a Bellows Seal Is the Wrong Choice
Bellows seals are not a universal upgrade: they fail predictably in services with high solids, severe dry running, or extreme misalignment, where the lack of O-ring damping allows stick-slip and face chatter [S5].
They are also a poor fit for very large shaft diameters above the fatigue envelope of standard edge-welded convolutions, where multi-ply formed bellows or a hybrid pusher design is more economical (2024-09) [S4].
For rotating equipment whose secondary sealing requirements are dominated by an elastomer's chemical compatibility rather than its sliding behaviour, such as strong oxidizers that attack 300-series stainless, the oil seal and bellows seal reference pages outline material substitutions that keep the seal static without forcing a full metal-bellows conversion. Where shaft speed, not containment, is the limiting factor, the same engineering logic pushes designers toward a dynamic balancing machine verification step rather than a seal re-design.
Standards, Sourcing, and Specification Cues
Bellows seal specification draws on the same API 682 / ISO 21049 mechanical seal architecture used for pusher seals, with the secondary seal category reclassified as static rather than dynamic [S4].
For hazardous-area service, the same IEC 60079-series and ATEX 2014/34/EU equipment-protection levels apply to the seal cartridge regardless of internal geometry, so Ex d, Ex e, or Ex i ratings transfer without re-test when only the secondary seal is changed from dynamic to static [S4].
Trackable signals to watch over the next procurement cycle: rising specification of Alloy 276 over 300-series stainless in chemical tenders, more 12-ply bellows stacks appearing as the cost premium over 8-ply narrows, and more API 682 Category 3 seal plans built around static secondary sealing elements rather than floating O-rings (2024-09) [S3][S4].
This topic is covered further in Metal Seated Butterfly Valve Leakage Class: Class IV to V, Rarely Class VI.