Cold Welding in a Vacuum: How Metals Bond Without Heat (and Why Space Engineers Fear It)

In April 1991, NASA’s Galileo spacecraft unfurled its high-gain antenna like an umbrella and then simply stopped — three ribs had welded themselves together in the silence of vacuum, with no heat, no arc, and no filler metal. That same invisible bond is now being deliberately engineered to patch spacecraft hulls in orbit, and it is the reason a bolt in a university vacuum chamber can fuse shut overnight.

A Stuck Antenna 400 Million Miles from Home: Why This Topic Matters

On April 11, 1991, engineers at NASA’s Jet Propulsion Laboratory sent a deployment command to the Galileo spacecraft’s high-gain antenna, a 4.8-metre umbrella of ribs and mesh meant to swing open like a folding fan and carry the mission’s data home. The ribs started to move, then stopped. The team retried, retracted and commanded the mechanism again and again. Nothing. Three ribs had locked together in the folded configuration, and one of the most expensive communication antennas ever built stayed shut for the rest of the mission.

There was no arc, no molten puddle, no filler rod and no glue. The investigation summarised in NASA’s formal Galileo failure investigation (LLIS Lesson 492) traced the anomaly to the rib retention mechanism’s design, preload-induced damage to the ceramic coating on the retention pins, and the breakdown of the molybdenum disulfide dry lubricant meant to keep the parts sliding. Once those films failed, bare metal met bare metal under pressure in hard vacuum, and in that environment metal does what it naturally wants to do: it bonds. Three ribs cold welded into place, quietly, at ambient temperature, 400 million miles from anyone who could pry them apart.

The Galileo story is not a curiosity, it is a design lesson. JPL had already run a cold-welding qualification program in 1969 to work out the correct ground test environment for spacecraft mechanisms in vacuum, work preserved in NASA technical literature such as the early NASA reports on cold welding in vacuum. The physics was known. The failure happened anyway, because a mechanism that behaves predictably in air can act like a completely different machine once the atmosphere is pumped away.

Here is the flip side: the same phenomenon that jammed Galileo’s antenna is a process engineers use on purpose. Cold welding is discussed as a hazard by NASA, ESA and the American Welding Society’s solid-state welding family, and it is simultaneously a commercial joining technique for wires, tubes and seals that fusion welding cannot make cleanly. Whether you are trying to stop it or start it, you are dealing with the same physics.

What This Guide Will Teach You

This article is for people who arrived with a real question: a mechanism that stuck, a vacuum chamber that surprised them, a materials decision they cannot afford to get wrong, or simple curiosity about why metal sticks to metal in space. The topic runs in order, from first principles to shop-floor practice:

1. The definition and the physics. What cold welding actually is, how the American Welding Society and NASA define it, the 0.1–0.5 melting-temperature criterion, and what happens at the atomic level.
2. Why vacuum is the trigger. The oxide, nitride and adsorbed gas layers that protect metals on Earth, the pressure thresholds where bonding starts, and why gold and platinum behave differently.
3. How a vacuum cold weld is made. Ion-beam cleaning, fracturing, controlled deformation and the contact parameters, briefed step by step.
4. Which metals cooperate and which refuse. A practical selection guide covering ductile nonferrous metals, steels, brittle materials and dissimilar pairs, with real adhesion data.
5. Where it is used. Wire and tube joints, hermetic pinch-off seals, ion-beam joining, aluminium-to-copper joints, and the ASTROBEAT experiment now testing in-orbit hull repair on the International Space Station.
6. Where it goes wrong. Spacecraft mechanism failures, fretting and stiction, bolts fusing shut inside chambers, and the testing mistakes that cause them.
7. Standards and test methods. Qualification chamber requirements, Hertzian contact pressure limits, outgassing screening and acceptance criteria.
8. Design rules you can apply. Coatings, dry films, material pairing, contact-area control, inspection advice and what to do at the workbench.

If you are a design engineer, you will leave with a checklist of high-risk contact types to audit. Fabricators get process parameters and material guidance; students get the physics without hand-waving; and welders who have only known arcs will see how much the rules change without an atmosphere. We also compare cold welding against the processes you already know. For the same physics-first treatment of a pressure-based process you can practise today, start with our guide to how a spot welder works: physics, process and practical setup.

Keep one image in mind as you read: two perfectly flat, perfectly clean metal surfaces touching in the dark. On Earth they sit apart forever under a film a few nanometres thick. In vacuum, that film is the only thing standing between a working hinge and a permanent weld — and as Galileo proved, it is not always enough.

What Cold Welding in a Vacuum Actually Is

Take away the arc, the torch, the filler rod and the shielding gas, and you are left with something that sounds almost too simple to be a real weld: push two pieces of metal together hard enough, and they become one. That is cold welding in a vacuum. No melt pool, no sparks, no orange glow lighting up the shop. Just pressure, clean metal, and a bond that forms in the time it takes to close a clamp.

It sounds like a party trick, but the definition is formal and well established. The American Welding Society defines cold welding as a solid-state welding process in which pressure is used at room temperature to produce coalescence of metals — the exact wording NASA reproduces in its own technical literature on the subject (NASA Lessons Learned: Cold Welding). Two words in that sentence do all the heavy lifting: pressure and room temperature. There is no melting, so there is no liquid metal to solidify, and no filler metal to contaminate the joint.

The 0.1–0.5 Tm Rule: Where “Cold” Really Comes From

If nothing gets hot, why call it a welding process at all? NASA patent literature draws the line numerically: cold welding occurs when the interface temperature stays low relative to the metal’s melting point — typically between 0.1 and 0.5 Tm, where Tm is the melting point expressed in degrees Kelvin (NASA Technical Reports Server). That ratio is the technical licence for the word “cold.” Check it against everyday metals and it holds up:

Every one of those sits comfortably inside the 0.1–0.5 Tm window, which is precisely why ambient-temperature pressure welding counts as “cold” even though the metal is technically above absolute zero by a few hundred degrees. Nothing approaches the melting range, so there is no heat-affected zone, no grain growth, no distortion, and none of the residual stress problems that keep fabricators awake after a heavy arc pass.

At the Atomic Level: Why Metal Cannot Tell Where It Ends

Here is the part that trips people up. A polished piece of copper and a polished piece of copper do not have a natural boundary between them — only the atoms sitting on their surfaces do. Once those surfaces are forced into intimate contact, the atoms at the interface cannot distinguish which piece they originally belonged to. As Penn State professor Sven Bilén explains, once atoms are in intimate contact the electrons cannot tell which piece they came from, so they begin sharing and effectively weld the pieces together (Live Science).

That electron sharing is the same metallic bonding that holds the rest of the part together. The interface does not become a joint in the usual sense; it stops existing as an interface at all. The atoms reorganise into a continuous lattice, and the seam is, chemically speaking, indistinguishable from the surrounding material.

A good analogy is two brand-new panes of window glass. Press them together flat and clean, and they cling as if glued — because the surfaces are close enough at the molecular scale for intermolecular forces to act across the gap. Metal goes one step further: the atoms actually intermingle and share electrons, forming a real metallic bond rather than a weak surface attraction. The catch is that glass only needs to be clean; metal in air is never clean, because oxygen instantly rebuilds the oxide film that keeps the two pieces apart. That is the whole reason vacuum matters, and it is the subject of the next section.

The practical takeaway is simple: cold welding is not adhesion, brazing, or a mechanical interlock. It is genuine coalescence of two solid metals created by pressure alone, defined by temperature relative to melting rather than by any flame, arc, or laser. Once you accept that, the rest of the topic — why vacuum enables it, how it is controlled, and when it sabotages a spacecraft mechanism — follows logically from a single fact: fresh metal, held close enough together, simply refuses to stay separate.

Why Vacuum Changes Everything: Oxides, Clean Surfaces, and the 1 Pa Threshold

If you pick up a polished steel bar in your workshop, you are not holding bare steel. You are holding a sandwich: adsorbed water vapour and gases from the room on top, nitrides and carbon contamination beneath, and underneath it all a thin, tenacious oxide skin that formed the instant the metal first met air. That sandwich is why your steel bar does not weld itself to the workbench. On Earth, almost every metal you can see is separated from every other metal by a few nanometres of chemistry that atoms simply cannot cross.

Remove the air, and the sandwich collapses.

The Invisible Blanket That Never Lets Metals Touch

Oxide layers are not optional for metals that live in atmosphere. Aluminium, titanium, copper, iron and nickel all grow oxide films continuously, and those films are chemically stable, mechanically stiff and completely unlike the soft metal underneath. When two engineering surfaces meet, the asperities on each side land on oxide, not on metal. Load the joint and the oxide behaves like a thin ceramic interlayer: it carries contact stress, prevents electron sharing across the interface and keeps the bond safely below the threshold where cold welding becomes possible. Adsorbed water and gas molecules add a second barrier on top, and under normal atmospheric conditions the separation is permanent.

Vacuum breaks the arrangement in two steps. First, the gas supply stops: with oxygen and water vapour removed, damaged oxide cannot regrow. Second, mechanical action — sliding, rolling, fracturing or impact — breaks the brittle films apart and exposes fresh, chemically active metal underneath. Once two clean metal surfaces touch, their atoms cannot tell where one piece ends and the other begins. Electrons delocalise across the interface, metallic bonds form, and the joint becomes, in effect, one solid piece of material.

Where the Threshold Actually Sits

The European Space Agency’s long-standing cold-welding compilation puts a hard line under the folklore: at ambient pressure there is no cold welding, because oxygen replenishes surface films faster than contact can damage them. Practical fretting campaigns by ESA’s Materials and Processes group and the Austrian Aerospace Centre mapped the same boundary from a tribology angle, finding the switch from oxidative wear — the harmless regime where debris is oxide powder — to adhesive wear between roughly 0.1 and 10⁻³ mbar. Cross-check the numbers and they agree: somewhere near one pascal, the surface stops repairing itself in time.

Vacuum level still matters below that line. NASA literature on ion-beam cold welding notes that at a vacuum of 10⁻⁶ torr, recontamination of a freshly cleaned surface takes on the order of a few seconds, and the clock stretches as pressure drops further (NASA Technical Reports Server). Even a good laboratory vacuum gives you a narrow working window. Pump a chamber to high vacuum, sputter-clean two faces, and you still have to bring them together promptly, before a monolayer of residual gas sneaks back in.

Why Gold and Platinum Are the Troublemakers

Most metals need their oxide stripped before they can cold weld. A few never had one. Gold and platinum form no oxide layer at all, which is why they stay shiny forever and why they are, in Caltech materials scientist Julia Greer’s words, notoriously prone to cold welding (Live Science). There is no protective blanket to break through in the first place — just clean, noble metal and whatever adsorbed gas happens to be sitting on it. Gold contacts in space mechanisms are therefore among the hardest to protect: a gold-plated connector that slides freely on the ground can tack itself in orbit.

Orbit Is Even Harsher Than a Lab Chamber

Low Earth orbit adds two forces that no ground chamber fully reproduces. Ultraviolet radiation and charged-particle flux drive off adsorbed contaminants, while atomic oxygen — abundant at LEO altitudes — chemically scours and erodes exposed surfaces. Both effects work in the same direction: they keep spacecraft materials cleaner, and in some cases chemically fresher, than anything you can condition on Earth. It is a major reason mechanisms behave differently once the hatch closes.

Keep those two thresholds in mind and the rest of this article becomes much easier to read: every failure story, test standard and design rule that follows traces back to the same root cause — in vacuum, the thin protective skin that keeps metals honest on Earth simply does not come back.

How a Vacuum Cold Weld Is Made: From Ion Beams to Roll Deformation

Knowing that cold welding happens is one thing. Making it happen on purpose is another. Whether you are sealing assemblies in a vacuum chamber or engineering a hermetic joint without a watt of heat, the recipe never changes: strip the surfaces, press them together hard enough that the metal flows, and keep every speck of contamination out until the joint is made.

Here is the process as a fabricator would brief it on the shop floor.

Step 1: Get Into Vacuum — and Understand Why

The chamber goes first. You pump down to remove the atmosphere that would otherwise rebuild an oxide layer the instant cleaning stops. NASA patent US 4,245,768 (NASA-Case-LEW-12982-1) describes cold welding demonstrated successfully at background pressures between 10^-4 and 10^-6 torr — deep enough that recontamination of a freshly cleaned surface slows to a timescale you can work with. At 10^-6 torr, the residual gas is roughly a billion times thinner than room air.

Step 2: Clean the Surface at the Atomic Level

Here vacuum cold welding departs from ordinary shop practice. Solvent degreasing removes oils but leaves the metal’s oxide film intact, and that film is the single barrier stopping two pieces of metal from fusing. So the process has to remove it physically:

• Ion beam sputtering. The NASA patent method. An accelerated beam of xenon or argon ions is aimed at the mating surfaces, blasting away oxide, nitride and adsorbed gas molecules atom by atom. Xenon is preferred for its higher sputter yield, though argon costs far less.
• In-situ fracture. Some designs simply break a preload-loaded member inside the chamber, exposing virgin metal that has never touched air.
• Plasma etching and wire brushing. Lower-tech supplements. Brushing only counts inside vacuum or before an airtight transfer, since aluminium reoxidises in milliseconds in air.

One reassuring detail from the same NASA work: at roughly 1 W/cm² beam power density, ion-beam cleaning of thin aluminium specimens (around 0.127 mm) produces only a modest steady-state temperature rise. This is genuinely a low-temperature process.

Step 3: Apply Pressure — and Force Metal to Flow

Clean surfaces alone are not enough. Two flat plates pressed lightly together touch only at the tips of their microscopic asperities. A bond needs those peaks to collapse so fresh metal from both sides flows together across a boundary that no longer exists. That means deformation, sometimes a lot of it.

Research published by Kaihara and colleagues in Japan gives a clean textbook example: copper bonded to aluminium under ultra-high vacuum of 10^-6 Pa, using 24 rpm rotation for 5 minutes at a bonding pressure of 120 MPa. Rotation provides shear, pressure provides intimacy, vacuum provides freedom from oxide. All three had to be right.

Industrial pinch-off practice shows the aggressive end of the scale. Per NASA’s copper tube pinch-off guidance, OFHC copper tubing can elongate 350% or more during the squeeze, cold welding the collapsing walls into a hermetic seal rated for bake-outs up to 600 °C and pressures to 800 psi. The deformation is the welding mechanism: the tube does not just close, it fuses shut.

The Two Classical Approaches

Process designers generally pick one of two routes:

1. Severe localised shear. Rotate, twist or swage the parts so surface films fracture and disperse into the joint. Higher deformation, but tolerant of less-than-perfect cleaning. Most pinch-off and roll-bonding processes live here.
2. Ultra-clean surfaces with moderate force. Ion-beam or fracture-cleaned parts need only enough pressure for intimate contact, because no oxide is left to break through. The elegant route — and the expensive one.

None of this needs a welding helmet, which is what makes vacuum cold welding counterintuitive for welders raised on arcs. Spot welding follows the same pressure-plus-cleanliness logic but adds heat, and our guide to how spot welders work explains that physics in detail.

The takeaway: cold welding is not magic and not spontaneous — it is controlled deformation where cleanliness replaces heat. Master those three variables and the physics that stranded Galileo’s antenna becomes a joining tool in your hands.

Which Metals Cold Weld — and Which Ones Refuse

The success of cold welding in a vacuum comes down to the material. Get the pairing right and two clean, ductile surfaces will fuse with nothing but pressure. Get it wrong, and no amount of force will create a load-bearing joint. Before committing to a design, it pays to know which metals are likely to bond and which will refuse.

The Metals That Bond With Ease

Soft, ductile metals are the strongest candidates for this process. According to TWI’s technical guidance on solid-state welding, aluminium, copper, nickel alloys and low-carbon steels cold weld most readily. These metals yield under load, allowing their surface asperities to collapse and expose large areas of fresh, oxide-free metal for electron sharing.

This is most apparent in smaller sections. For example, cold wire butt joints are routinely performed on wire from 0.5 mm to 10 mm diameter or larger. Thinner parts can be joined with simple tools, while thicker stock will require a hydraulic or mechanical press. Low-carbon steel is a particularly common surprise, as its ductility allows it to bond despite being ferrous. These same traits make cold welding a natural pressure-only counterpart to processes like TIG welding sheet metal, especially for thin, soft workpieces.

Why Hard and Brittle Materials Refuse

Hardness is the biggest obstacle to cold welding. If a material is too hard or brittle, it will resist plastic deformation. Instead of flattening across the interface, the two faces will only touch at a few high points, leaving far too little real contact area to form a joint. Ceramics, tool steels, heavily work-hardened alloys and glassy materials typically fall into this category.

Real adhesion data reinforces this trend. In fretting tests run by ESA’s Advanced Achievements Centre, tungsten-against-tungsten and molybdenum-against-molybdenum pairs stayed well below 500 mN of adhesion force, while silver, copper and 17-7PH stainless steel showed far higher adhesion under otherwise comparable conditions. The low adhesion of refractory metals makes them ideal for bearings and sliding contacts, but far from ideal for structural bonds.

Beyond hardness, two factors can make or break a joint: surface finish and contact pressure. A finely lapped, flat surface will contact over nearly its entire area, while a rough face will only touch at its peaks. Even the softest metal will fail to bond if parts rock, gap or see uneven pressure. Uniform, high contact pressure is required to collapse those asperities.

The Noble Metal Exception

While most metals rely on ductility, gold and platinum have a different advantage: they carry no protective oxide layer at all. On Earth, oxides, nitrides, adsorbed gases and water vapour block bonding, but gold never forms that skin. In a vacuum, it arrives already clean and primed to stick. Silver shares much of this behaviour, which is why its use in space hardware demands mechanical restraint or deliberate separation. No amount of “cleaning” is needed to trigger it, which is often the problem.

Approved Materials for Vacuum Pinch-Off Seals

For the specific task of cold-welding a tube shut to seal a vacuum, material choice is tightly controlled. NASA’s copper tube pinch-off white paper approves the following materials for reliable cold-weld sealing:

• OFHC (oxygen-free high-conductivity) copper
• High-purity nickel, grades Ni 200 and Ni 270
• Annealed 3003 H14 aluminium
• Pure iron, gold, platinum, silver and columbium (niobium)

Dissimilar Pairs and Practical Takeaways

Dissimilar metals can be tricky. If a soft metal is pressed against a hard one, the soft side will deform, but the hard side contributes little contact, producing a weaker joint. Pairs with large differences in hardness, thermal expansion or galvanic behaviour are best avoided for structural use.

That said, some pressure-based solutions sidestep melt-related issues. Aluminium-to-copper is a prime example: both cold weld readily, but fusion welding them forms brittle intermetallics. Pressure-based approaches, such as spot welding copper and other solid-state methods, avoid that entirely. In practice, the rule is simple: use soft, ductile metals, finish them smooth and flat, and apply uniform, high pressure. Do that, and the joint will hold. Fall outside of those parameters, and it will not.

Where Vacuum Cold Welding Is Used: From Spacecraft Hulls to Pinch-Off Seals

Everything so far has framed cold welding as a hazard. Flip the coin and you find a joining technique that does things no torch can: no melting, no filler, no heat-affected zone, and no distortion. When you want two pieces of metal to become one continuous piece of metal without ever getting them hot, vacuum cold welding is not a bug, it is the entire product.

Wire, Tubing, and the Humble Pinch-Off Seal

On the shop floor, the most common deliberate cold welds are small and unglamorous. Wire manufacturers butt-clean and press strand ends together; tube fabricators squeeze soft metal lines flat until the walls bond into a leak-tight seal. That second one is the hermetic pinch-off, a technique vacuum technicians rely on daily. A copper or aluminium tube is pumped down, then a calibrated tool crushes it shut; the metal-to-metal contact is intimate enough that the joint holds high vacuum indefinitely with no braze, no solder, and no elastomer to outgas. NASA’s own pinch-off guidance for copper tube treats the squeeze as a qualified process, with tool geometry and deformation depth controlling whether the seal is reliable or merely closed (NASA copper tube pinch-off white paper). Get it right and you have sealed a system permanently with nothing but a pair of jaws.

The Aluminium-to-Copper Problem, Solved

Ask an arc welder to fuse aluminium to copper and you will get a roomful of brittle intermetallic compounds and a joint that cracks in service. Cold welding sidesteps the metallurgy entirely because nothing ever reaches the liquid state where those intermetallics form. The result is a strong, durable bond across a pair that fusion processes struggle with — a genuine advantage in busbars, battery tabs, and hybrid electrical assemblies. If you have fought aluminium-to-copper joints before, it is worth reading how spot welding copper handles resistance-welding parameters, since the pressure-and-cleanliness mindset carries straight across.

Ion Beams, Wafer Bonding, and Big Space Structures

Further up the precision ladder sits ion-beam preparation: surfaces are sputtered clean with argon or xenon inside a vacuum chamber and then pressed together before a single molecule of oxygen can return. NASA patent literature covers using this approach to bond thin wafers to hard substrates and to fabricate large space structures, with coating and bonding performed in a single vacuum set-up so the parts are never exposed between steps (NASA NTRS, NASA NTRS). It is also the principle behind semiconductor wafer bonding, where two mirror-flat silicon wafers touch and adhere with no adhesive at all. Anyone building or specifying the chambers this happens in should study how a spot welder works for the pressure-control intuition, then look at vacuum hardware requirements in more detail.

The New Frontier: Patching a Hull in Orbit

The most exciting application is currently 400 kilometres up. The ASTROBEAT experiment, sponsored by the ISS National Laboratory and delivered on SpaceX’s CRS-31 mission, is testing cold welding as a way to patch simulated spacecraft hull breaches — micrometeoroid damage that today demands spare parts nobody can easily resupply. Led by Leonardo Barilaro of the Malta College of Arts, Sciences and Technology, the project explicitly aims to turn a long-standing space hazard into a viable repair mechanism (Live Science). In other words, the same force that jammed Galileo’s antenna is being trained to seal the hole it might punch in the next mission.

Note: The through-line for every application above is the same three inputs — a virgin surface, a ductile material, and enough pressure to force true atomic contact. Change one, and you get either a failed bond or an unwanted one.

That duality is what makes this topic worth understanding in full, and it sets up the failure stories in the next section neatly.

When Cold Welding Is the Villain: Spacecraft Mechanism Failures

Every engineer who wants a cold weld has a mirror image: an engineer who would pay dearly to avoid one. The metal does not care which one you are. The same contact mechanics that let an orbital repair kit patch a hull can also seize a hinge, glue a release pin shut, or turn a routine antenna deployment into a mission-long malfunction. This is where cold welding in a vacuum stops being a laboratory curiosity and starts costing missions.

Galileo: What the Investigation Actually Found

In April 1991, NASA’s Galileo spacecraft deployed its high-gain antenna and stalled with three ribs locked together. The post-flight engineering review, summarised in NASA’s Lessons Learned database, pointed to two compounding failures rather than one freak event. First, the V-groove pin-and-socket interfaces carried very high contact stresses that crushed and displaced the solid lubricant film designed to keep the parts sliding. Once that film was breached, sliding friction between surfaces displaced under load in vacuum climbed by nearly a factor of ten compared with the same contact in air (NASA LLIS Lesson 492).

That tenfold jump is the number mechanism designers quote to each other. Nothing about the hinge geometry changed between the clean room and 400 million miles away; only the environment did. Ground testing that measures friction, release force, and travel in ambient air is measuring a different machine than the one that flies. The broader NASA literature on solid-state and cold welding behaviour reinforces the same warning (NASA NTRS 19720016873, NASA NTRS 19810010928).

The Contacts Every Designer Should Audit

ESA and the Austrian Aerospace Centre (AAC) rank spacecraft contacts by cold-welding risk, and the list reads like a bill of materials for any moving mechanism: hold-down and release mechanisms, stranded wire harnesses, mechanical end stops, gear toothing, bearings, and slip rings (ESA Publication STM-279). The common thread is a combination of clean or freshly displaced metal, sustained contact load, and little or no chance for an oxide layer to regrow.

The wire-harness finding surprises people most. According to work presented in the ESMATS 2024 proceedings, bending a wire harness in vacuum may cause cold welding of single wires within a strand. A cable bundle that flexes freely on the bench can effectively sinter itself into a stiff, unreliable loom in orbit, changing routing loads and connector strain in ways nobody planned for.

Bolts That Fuse Shut on Earth

This is not only an orbital problem. Sven Bilén, a professor of electrical engineering at Penn State, has described how bolts in his own laboratory’s vacuum chamber fused shut after the chamber was moved across campus and pump-down resumed; the fasteners had to be drilled out (Live Science). Two steel faces sat under preload with no oxygen to replenish their oxide films, and the threads quietly became one piece of metal.

Anyone who has to machine a seized fastener out of a chamber knows it is dirty, hands-on work. It is also exactly the moment where eye protection, gloves, and proper shop procedure matter as much as anywhere in fabrication — the same discipline covered in our guide to welding PPE.

Fretting and Stiction: Cold Welding’s Cousins

Two related phenomena live on the same spectrum. Fretting happens when a joint oscillates by tiny amplitudes: the oxide layer is repeatedly broken, fresh metal touches fresh metal, junctions form and shear, and the resulting debris accelerates wear. On a vibrating connector or bearing race, fretting shows up as rising electrical resistance and dimensional change long before anything looks seized. Stiction is the static-friction surprise — a mechanism that broke free easily in air refuses to move in vacuum, or releases with a shock that damages the mechanism it was protecting. Both are what you get when contact surfaces are treated as if air will always be there to keep them apart.

The practical lesson is simple: in vacuum, treat every loaded contact as a candidate weld, and qualify it in the environment where it must actually work.

Common Mistakes, Failure Modes, and How to Test for Them

Most cold welding disasters are not exotic physics. They are ordinary engineering oversights: a contaminated test chamber, a coating chosen by habit, a fastener torqued “nice and tight,” or a mechanism that passed ground test and then jammed in orbit. If your hardware will live in vacuum, read this section twice.

Mistake 1: Trusting a Contaminated Test Chamber

The costliest mistake in vacuum mechanism testing is documented in NASA’s own history. In a 1972 cold-welding program, researchers compared laboratory results with actual flight data and found that only about 50% of the data generated in an oil-diffusion-pumped ultrahigh vacuum chamber agreed with what really happened in space. The culprit was pump oil: hydrocarbon vapour back-streamed from the diffusion pump, deposited a thin organic film on every “clean” metal surface, and acted as an unintended lubricant that suppressed cold welding in the lab but was absent on a real spacecraft. Repeating the same tests in ion-pumped facilities with no hydrocarbons anywhere in the system pushed agreement with flight data to roughly 90%. NASA’s 1972 report remains a sobering reminder that a good test in the wrong chamber is worse than no test at all.

Mistake 2: Relying on MoS₂ Where It Cannot Help

Molybdenum disulfide is the default dry-film lubricant for space mechanisms, for good reason: it works beautifully in vacuum. It is not universal protection. ESA’s classic adhesion study, STM-279, demonstrated the failure mode clearly. An MoS₂ coating applied to 17-7PH precipitation-hardening stainless steel lost its lubricating effect after just 20 cycles — roughly 200 seconds of testing — and could not prevent cold welding. The film was mechanically destroyed at the contact, exposing bare metal that then welded. Hard coatings on the same substrate did protect. MoS₂ needs a compatible, relatively soft substrate and an application process that produces a strongly adherent, well-consolidated film. On hard, smooth, high-strength steels it can fail faster than a deployment sequence runs.

Mistake 3: Over-Tightening Fasteners and Ignoring Contact Pressure

A bolt that feels “secure” in a workshop can be a cold-welding time bomb. Cold welding scales with real contact area and plastic deformation, both driven by Hertzian contact pressure. The governing design limit comes from ECSS: ECSS-E-ST-33-01C Rev.2, section 4.7.4.4.5 requires that peak Hertzian contact pressure remain below 93% of the yield limit of the weakest material in the pair — approximately 58% of the elastic limit. Over-torquing a soft, ductile, oxide-prone austenitic stainless fastener against a hard captive nut is precisely how a hold-down bolt becomes permanently one piece. The same reasoning applies to thread engagement length, washer hardness, and any point contact inside a mechanism.

Mistake 4: Assuming Air-Test Behaviour Predicts Vacuum Behaviour

This is the classic mechanism designer’s trap. A latch that cycles smoothly for ten thousand cycles on the bench can seize within minutes in vacuum, because in air the oxide layer, adsorbed water film, and nitrogen atmosphere all quietly separate the surfaces. Remove them and you remove the safety margin. Two phenomena make it worse: fretting, where micro-sliding repeatedly breaks down surface films and compacts debris into a cold-welded junction, and stiction, the static friction spike that can strand a mechanism at rest. Air testing tells you almost nothing about either. For more on how dramatically atmosphere changes contact behaviour, see our guide to spot welding physics and process setup.

How to Test Properly: Methods and Acceptance Criteria

Modern qualification follows a disciplined recipe, refined by ESA’s adhesion and cold-welding research at Airbus and summarised in recent mission studies:

• Impact testing in chambers with base pressures below 5 × 10⁻⁸ mbar, using ion or getter pumps, to simulate latch closures and release events.
• Fretting testing at a 50 µm stroke and 200 Hz, replicating vibration-induced micro-motion at bolted and pinned interfaces.
• Contact stress screening against the ECSS Hertzian limits above, applied to every touching pair including threads, pins, bearings, and end stops.
• Outgassing screening via the ASTM E595 test, catalogued in NASA Reference Publication 1124, requiring total mass loss below 1.0% and collected volatile condensable material below 0.1% at 125 °C and 10⁻⁶ torr.
• Witness-sample adhesion checks, where sacrificial coupons from the same material lot are cold-welded in a dedicated facility and measured for separation force.

Ground crews are not immune either. Working around vacuum hardware brings high-voltage supplies, trapped-energy releases, and pinch points during chamber loading, so keep current on shop safety practice, including the fundamentals in our MIG welding safety precautions guide, before improvising any bench rig for cold-weld testing.

Test honestly, measure contact pressure honestly, and choose coatings on evidence rather than habit. Do that, and cold welding in vacuum stops being a surprise and starts being a solved problem.

Cold Welding vs. Fusion, Brazing, and Other Solid-State Processes

No joining process wins everywhere. Cold welding solves one set of problems beautifully and fails at another just as decisively, so choosing it well means seeing where it sits relative to the alternatives an engineer might reach for instead.

A Quick Map of the Joining Family

Every metal-joining method falls into one of three camps: it melts the base metal, it melts only a filler, or it melts nothing at all. Cold welding lives firmly in the third. TWI’s technical guide to solid-state welding makes the central point plainly: because these processes do not melt the metals, the joined parts retain their original properties, thermal expansion and conductivity matter far less, and dissimilar metal joints become possible.

ASM Handbook Volume 6A treats cold welding as one member of a broader family that also includes magnetic pulse welding, roll welding, ultrasonic additive manufacturing, and explosion welding. Same physical principle, wildly different ways of delivering the pressure.

Cold Welding vs. Fusion Welding

Arc processes such as TIG and MIG create a molten pool, which means a heat-affected zone, residual stress, distortion, and usually a filler metal chemically different from both parent metals. Cold welding skips all of that: no arc, no melt, no filler, no HAZ. Conductivity and microstructure at the joint stay essentially as they were, which matters for heat-sensitive assemblies and precision parts that would warp under an arc.

The flip side is geometry. Fusion welding runs a bead along a seam of almost any shape, with procedures that are documented, code-qualified, and inspectable by radiography or ultrasonics. Cold welding wants flat, parallel, ductile surfaces and does not like long irregular seams. Before weighing an arc process for sheet or thin-wall tube, read The Complete Guide to TIG Welding Sheet Metal first, because the distortion TIG users fight is precisely what cold welding avoids by design. At the other extreme, Railroad Thermite Welding shows how much heat and metallurgical change a purely fusion-based route introduces.

Cold Welding vs. Brazing and Soldering

Brazing and soldering melt a filler metal rather than the parent metals, so they are gentle, versatile, and excellent at joining dissimilar materials into complex joint geometries. The cost is joint strength and often temperature capability: the joint is only as good as the filler, and the alloy left behind is not the alloy you started with. A cold weld, when it works, is made of the parent material itself. Where a brazed joint would creep or soften, a cold weld does not, because nothing was ever heated.

Cold Welding vs. Diffusion Bonding

This is the closest sibling, and the comparison that most often confuses newcomers. Diffusion welding produces no appreciable deformation and can join refractory metals that cold welding would shatter, but it requires heat plus a vacuum or controlled atmosphere to keep surfaces oxide-free, and the cycle runs hours rather than seconds. Cold welding trades that gentleness for speed and room-temperature processing, and pays with deformation at the weld interface. If your part tolerates flattening, cold welding is dramatically faster; if dimensional fidelity is everything, diffusion bonding wins. NASA’s foundational literature, including its NTRS-hosted research on vacuum cold pressure welding, frames the trade-off in exactly these terms.

Cold Welding vs. Friction, Explosion, Roll, and Ultrasonic Processes

Friction and friction stir welding deliver pressure through mechanical work, generating heat in the process, and can reach butt joints and thicker sections cold welding cannot. Explosion welding achieves enormous instantaneous pressure over large flat areas and is the go-to for cladding dissimilar metals, though it remains a specialty operation with real logistical overhead. Roll welding applies pressure continuously through rolling passes, and ultrasonic methods such as additive manufacturing use high-frequency vibration at relatively low bulk temperatures. The theme is constant: clean surfaces plus pressure, with the energy delivered in a different currency.

The Trade-Offs That Actually Decide the Job

• Strength vs. deformation. Cold welds approach parent-metal strength but visibly reshape the joint area; diffusion bonds barely move the part but cost hours.
• Dissimilar metals. Every solid-state route handles them better than fusion welding, because no brittle intermetallic melt pool forms.
• Equipment and access. Vacuum chambers and ion sources are expensive infrastructure; an arc welder is not.
• Throughput. Cold and ultrasonic welding are fast once set up; diffusion bonding is a batch process.
• Inspectability. Fusion welds carry decades of code-qualified NDT practice; solid-state joints often need metallography or pull testing to qualify.
• Repairability. A fusion weld can be ground out and rerun; an accidental cold weld cannot be reheated to undo it, which is why spacecraft designers treat it as a failure mode rather than a feature.

That last point is the hinge of this entire article and the start of the next section: the same physics that makes vacuum cold welding a manufacturing tool makes it a hazard 400 million miles from the nearest repair shop.

Practical Tips, Design Rules, and the Future of Cold Welding in Vacuum

Cold welding punishes ignorance and rewards preparation. Nothing about it is mysterious once you know the rule: metals bond when clean surfaces touch under pressure, and everything else an engineer does exists to keep those two conditions from coinciding by accident. Whether you are designing a spacecraft release mechanism, building a vacuum chamber in your workshop, or wondering why a bolt seized overnight, the playbook is the same. Keep surfaces apart, keep them covered, keep them clean.

Prevention Strategies That Actually Work

The first line of defence is a coating that stops bare metal ever meeting bare metal. Johns Hopkins University Applied Physics Laboratory spacecraft lubrication guidance is blunt: if a cold-weld-prone material must be used, apply a protective coating such as hard anodised aluminium alloy. Anodising grows an artificial oxide layer directly onto the aluminium and locks it in place, so the interface carries something stable and ceramic-like instead of a pristine metallic face. Conversion coatings, phosphate and chromate treatments, and simple platings do the same job, giving the joint a chemical barrier that vacuum cannot strip away.

Dry lubricants are the second line. Molybdenum disulfide (MoS₂) and tungsten disulfide (WS₂) work by letting weakly bonded layers slide over one another, physically separating the moving surfaces, as materials scientists explained to Live Science. The caveat is critical: NASA’s Galileo lessons warn engineers to evaluate MoS₂ carefully on any hardware that will be tested or operated in air, because its wear rate in oxygen is high and a film that survives a vacuum chamber test may already be destroyed by launch. Pair a soft coating with a hard substrate, or the reverse, and ask how the film will behave during ground handling, not just in orbit.

Geometry matters too. Galileo’s real lesson was preload damage: squeeze a coated pin hard enough to breach the lubricant film and you expose bare metal exactly where slip occurs. Keep contact pressure below the film’s damage threshold, spread load over larger bearing areas, and avoid point contacts where two curved surfaces meet.

Inspection and Qualification Advice

If you are qualifying hardware, the test environment is part of the test. NASA’s Jet Propulsion Laboratory cold-welding program concluded that ion-pumped or magnesium-getter vacuum facilities should be used for qualification work and that oil-pumped environments must not be, because pump-oil vapour contaminates the very surfaces you are trying to keep pristine and quietly invalidates the results. Screen materials for outgassing, run witness samples alongside flight articles, and measure adhesion or breakaway torque before and after thermal cycling rather than trusting a single room-temperature reading. Assume a mechanism proven in air will not behave the same way in vacuum until data proves otherwise.

Workshop Safety and Housekeeping

You do not need a spacecraft to meet cold welding. Anyone pulling a vacuum on a home chamber can find it in threads, bearings, and fittings. Handle vacuum hardware with clean gloves or lint-free wipes, never bare fingers, since skin oils are a contamination source on one side and a slip risk on the other. Torque fasteners to specification instead of guessing, keep solvents and wipes segregated, and treat any mechanism headed for a chamber as clean-room equipment. The usual shop discipline applies long before you reach the chamber: correct protective equipment, ventilation, and hazard awareness are covered in our Welding PPE: The Complete Guide to Protective Gear Every Welder Needs (2026).

The Future: In-Space Manufacturing and Repair

The force that jammed Galileo’s antenna is now being courted deliberately. Researchers are developing ion-beam activation and cold-pressure joining for assembling large orbital structures, bonding semiconductor wafers without heat, and patching micrometeoroid damage on hulls while a spacecraft is still in service. In-space manufacturing suits the process because cold welds need no power-hungry heat source, no filler metal launched from Earth, and leave no heat-affected zone to weaken thin panels. On the ground it already quietly produces wire draws, hermetic pinch-off seals, and aluminium-to-copper joints that fusion welding handles badly. Expect it to spread wherever mass, power, and dissimilar metals are the constraints.

Thirty-five years after three antenna ribs fused themselves shut 400 million miles from home, cold welding has moved from unexplained failure to engineered tool. The physics never changed; only our respect for it did. Design as though your cleanest surface is waiting to become permanent, and the same bond that stranded Galileo’s high-gain antenna can one day build, seal, and repair the structures we send next.

Frequently Asked Questions

What is cold welding in a vacuum?

Cold welding in a vacuum is a solid-state joining process in which two clean metal surfaces are pressed together and bond atom-for-atom without melting, arc, or filler metal. Vacuum removes the protective oxide layer that normally stops metals from bonding, so pressure alone creates a joint as strong as the parent metal.

Why does metal weld in a vacuum but not in normal air?

In air, every metal surface carries a thin oxide and contamination film that physically separates the atoms. In vacuum that film cannot reform once it is broken, so fresh metal atoms from each side come into intimate contact and begin sharing electrons, fusing the pieces together.

At what vacuum level does cold welding start?

ESA research reports no cold welding at ambient pressure, with adhesion appearing at roughly 1 Pascal and below. The shift from oxidative to adhesive wear typically occurs between 0.1 and 10^-3 mbar, and the effect becomes more reliable the deeper the vacuum.

Which metals are most prone to cold welding?

Ductile, soft metals bond most easily: aluminium, copper, nickel, silver, lead and gold are classic examples. Gold and platinum are especially notorious because they form no oxide layer at all, even in normal air, so they have no natural protection against bonding.

Is a cold-welded joint as strong as a fusion weld?

A properly made cold weld can be as strong as the base metal because the interface effectively disappears. The trade-off is that cold welding requires very clean, flat surfaces and produces substantial deformation, and it is limited mainly to ductile materials.

How did cold welding affect the Galileo spacecraft?

In April 1991 Galileo’s high-gain antenna failed to deploy fully; investigation found three ribs had cold welded in the folded position. High contact stress had damaged the pin coating and dry lubricant, and friction in vacuum rose nearly tenfold, locking the mechanism.

How do engineers prevent unwanted cold welding in space?

They separate or protect the surfaces: hard anodising and conversion coatings, MoS2 or WS2 dry-film lubricants, dissimilar material pairings, reduced contact area and Hertzian stress, and in some cases hermetic sealing. Each fix has limits, so hardware is qualified in oil-free vacuum chambers.

Can cold welding happen on Earth?

Yes, though much less often because oxygen constantly rebuilds the oxide layer. It appears when that protection is removed — for example when over-tightening austenitic stainless fasteners, during friction welding, or when bolts sit undisturbed inside a laboratory vacuum chamber for long periods.

What is the difference between cold welding, galling, and stiction?

Cold welding is a genuine metallurgical bond between two clean metal surfaces with no heating. Galling is a form of adhesive wear where surface material transfers and tears under sliding, while stiction is the broader sticking of parts at rest. In vacuum, stiction and galling often escalate into true cold welding.