Explosive Welding: Bonding Metals With a Controlled Blast

The Setup: A Flyer, a Standoff, and a Sweeping Detonation

An explosion weld is assembled cold, like a sandwich. The heavy base (parent) plate — usually cheap carbon or low-alloy steel — rests on a rigid anvil of steel, concrete, or packed sand. Above it sits the flyer (cladder) plate, the expensive corrosion- or heat-resistant metal, held either parallel with a small standoff gap or tilted at a slight preset angle. The gap is only a fraction to a few times the flyer thickness (typically a few millimetres) and is set by spacers or foam pegs. A layer of explosive is spread across the top of the flyer, often over a thin rubber or PVC buffer that shields the surface from the hot detonation gases, and a single detonator sits at one edge.

When it fires, the detonation front races across the sheet at its detonation velocity (VOD), and the expanding gases behind that front hammer the flyer downward. Low-VOD explosives are chosen deliberately — ammonium-nitrate mixtures (ANFO), amatol, or PETN cut with salt or sand — so the VOD lands in the 1,500–3,000 m/s band, safely below the sound speed of the metals. Because the detonation sweeps rather than firing everywhere at once, the flyer peels down progressively from the initiation edge. Any given point on the interface is welded in microseconds, and the front crosses a metre-wide plate in a few hundred microseconds, so it is done outdoors in remote firing bays, in sand pits, or underwater to absorb the blast.

The Moving Collision Point and the Metal Jet

The heart of the process is the collision point — the line where the descending flyer slaps into the base. As the flyer peels down, that line sweeps across the plate at the collision-point velocity V_c, which for a parallel setup is essentially the detonation velocity. The flyer itself arrives at only a modest impact velocity V_p of roughly 150–500 m/s, but it meets the base at a sharp dynamic bend angle β of about 5–25°.

At that closing wedge the transient contact pressure reaches on the order of 10–30 GPa — one to two orders of magnitude above the yield strength of steel. At such stresses the metals' own strength is negligible: they momentarily behave like fluids, which is why the whole event can be analysed with hydrodynamics rather than solid mechanics. A thin reentrant jet of surface material is squirted forward out of the wedge, exactly like the spray thrown up where two streams of water collide, or the jet of a shaped charge. That jet strips away the oxide films, adsorbed gas, and grime on both faces — the single greatest obstacle to any solid bond — leaving atomically clean, virgin metal that is instantly pressed into contact under gigapascal pressure. With nothing between them, the two lattices bond metallurgically. Crucially, the jet only forms while the collision point stays subsonic relative to the metals' bulk sound speed (~5 km/s in steel). Push V_c supersonic and an attached shock replaces the jet, no surface is cleaned, and no bond forms — the same physics that governs oblique-shock reflection.

Why the Interface Goes Wavy

Cut and polish an explosion weld and you see the signature: not a flat seam but a train of interlocking sine waves. The wavelength is typically 0.1–3 mm and the amplitude tens to hundreds of microns, both growing with flyer thickness and impact energy. The wave is generated behind the sweeping collision point, where the freshly bonded metal is under enormous shear. The favoured explanations are the jet-indentation mechanism of Bahrani, Black and Crossland (1967), in which the reentrant jet periodically indents a hump just ahead of the collision point, plus a Kelvin–Helmholtz-like instability of the sheared interface and a von Kármán vortex-street analogue — the collision point acting like an obstacle in a fast stream that periodically sheds vortices. Cowan, Bergmann and Holtzman (1971) showed the onset of waviness tracks a Reynolds-number-like parameter, R = (ρ₁+ρ₂)V_c² / 2(H₁+H₂), built from the two metals' densities and hardnesses (hardness standing in for an effective dynamic viscosity) and the collision-point velocity — with the flat-to-wavy transition falling near a critical value of about 10.6.

The waves are not just cosmetic. They mechanically interlock the two metals, adding to the metallurgical bond and giving the joint its high shear strength. But their swirling vortices can also trap small pockets of material that briefly melted and re-froze — and that is exactly where any brittle intermetallic collects. So bigger waves are a double-edged sign: more interlock, but larger melt pockets. Push the energy too high and those pockets merge into a continuous brittle layer that cracks. Good practice keeps the waves modest and the melt pockets small and discontinuous.

The Welding Window: Governing Numbers and Trade-offs

The three master quantities are tied together by a simple kinematic relation, V_p = 2 V_c sin(β/2): the impact velocity, the collision-point velocity, and the bend angle are not independent. The engineer's job is to land the operating point inside the weldability window — a region in the (collision velocity, impact velocity) plane first mapped by Deribas in Novosibirsk and by Wittman, bounded on four sides:

• Lower limit: a minimum impact energy is needed to eject a jet and exceed the dynamic yield stress. Below it, the surfaces are never cleaned and no bond forms.
• Upper limit: too much energy melts the whole interface, forming a continuous brittle intermetallic layer that cracks or spalls.
• Sonic limit (right side): a collision point faster than the metals' sound speed suppresses the jet entirely.
• Load ratio R: the mass of explosive per unit area divided by the flyer's areal mass (typically 0.5–2) fixes V_p through the Gurney equation, V_p = √(2E)·f(R); the Gurney velocity √(2E) is roughly 1.5–3 km/s for the diluted, low-VOD explosives used. The standoff sets how far the flyer accelerates, and thus β.

The failure modes map directly onto the window edges: too little charge gives unbonded patches; too much gives interfacial melt, intermetallics, and spallation; the wrong VOD kills jetting. The safety factor is bought by centring the operating point in the window rather than hugging an edge, and confirming it on sacrificial test coupons before the production shot.

No Melting, So No Brittle Intermetallics

This is the whole reason the process exists. Try to fusion-weld aluminium to steel and the shared molten pool grows thick, brittle Fe–Al intermetallics — FeAl₃, Fe₂Al₅ — that fracture under the slightest load; the same happens with titanium–steel (Fe₂Ti, FeTi). Explosive welding sidesteps this because the intense heating is confined to a layer only microns thick and lasts only about a microsecond, then is quenched by conduction into the massive cold plates on either side at cooling rates of order 10⁵–10⁷ K/s. Diffusion simply has no time to build a thick intermetallic; at most you get a sub-micron, discontinuous diffusion zone or isolated melt pockets in the wave crests.

That is why explosion welding can join pairs that fusion welding cannot touch: aluminium–steel, titanium–steel, copper–steel, tantalum–steel, zirconium–steel, and aluminium–copper. The bulk of both plates never sees more than a brief thermal pulse, so their tempers and corrosion resistance survive. The trade-off is heavy cold work and residual stress near the bond line, and the parent plate arrives dished from the blast; producers flatten it and often follow with a low-temperature stress-relief anneal, deliberately kept below the temperature at which intermetallics would resume growing.

How It Is Built, Tested, and Standardised

Explosion cladding is an industrial product, not a laboratory curiosity, and it is bought and sold against real standards. After the shot the clad plate is flattened and its bond is proven by three kinds of test:

• Ultrasonic testing for unbonded areas — 100% coverage per ASTM A578 / A435; any disbond shows as a reflection.
• Bond shear strength — ASTM A263 (chromium-stainless clad steel) requires at least 140 MPa (20 ksi), generally far exceeded — with ram-tensile and side-bend tests to confirm ductility.
• Metallography of the wave interface, checking wavelength, melt-pocket size, and the absence of a continuous intermetallic band.

The product standards cover each clad family: A263 (chromium stainless clad), A264 (chromium-nickel, i.e. austenitic stainless clad), A265 (nickel and nickel-alloy clad), B432 (copper and copper-alloy clad), and B898 for the reactive and refractory metals — titanium, zirconium, tantalum. The American Welding Society abbreviates the process EXW. Today the field descends from DuPont's 1960s Detaclad business — built on the jetting and wave-formation work of G. R. Cowan and A. H. Holtzman — bought in 1996 by Dynamic Materials Corporation (today DMC Global) and now sold under the NobelClad brand, which can bond plates of tens of square metres in a single detonation.

Where It Is Used: Cladding, Transition Joints, and Bimetals

Clad plate is the biggest market. Bonding a few millimetres of titanium, zirconium, or a nickel alloy onto thick carbon steel gives a vessel wall that is corrosion-proof on the wetted face and cheap and strong underneath — used for pressure vessels, heat-exchanger tubesheets, chlor-alkali cells, desalination gear, and chemical reactors, where a solid plate of the exotic alloy would be ruinously expensive.

Bimetallic transition joints solve the dissimilar-metal problem structurally. DMC's Detacouple aluminium-to-steel strips let a shipyard fusion-weld an aluminium superstructure to a steel hull by welding each side to its own metal, sidestepping a direct Al–Fe fusion joint that would crack — standard practice on naval and cruise ships, sometimes with a pure-aluminium or titanium interlayer to further starve intermetallics. Aluminium-to-stainless transitions serve cryogenic and LNG plant and superconducting-magnet cryostats. Aluminium–copper transition pieces carry current in busbars and the cell-to-cell connectors of aluminium smelters, where a bolted Al–Cu joint would corrode and overheat.

Explosion welding also expands tubes into tubesheets, plugs failed heat-exchanger tubes, and repairs pipe. Fittingly, the effect was first noticed on the battlefield — the wavy bonds found where wartime fragments and shaped-charge liner metal had stuck fast to armour plate are what put researchers like L. R. Carl (1944), and later Cowan, Holtzman, and Deribas, onto turning a weapon's physics into a joining process.

Explosive welding versus other ways to join metals — especially dissimilar pairs

Property	Explosive welding (EXW)	Arc / fusion welding	Roll / diffusion bonding
Bonding mechanism	Jetting + gigapascal impact, solid state	Melt a shared pool, then solidify	Sustained pressure + heat, atomic diffusion
Bulk melting?	None — micron-thin, µs interface heating	Yes — weld pool and heat-affected zone	None, but held hot for minutes to hours
Al–steel / Ti–steel	Excellent — routine industrial product	Poor — thick brittle Fe–Al / Fe–Ti cracks	Thin roll-bonded strip only (e.g. Al–stainless cookware); thick plate impractical
Interface	Wavy, mechanically interlocked	Fusion zone, dendrites, HAZ	Flat diffusion bond
Geometry & size	Large flat or cylindrical clad plates (m²)	Any joint, but thermal distortion	Thin flat sheet or strip
Energy source	Chemical explosive (one shot, µs)	Electric arc or beam (seconds–minutes)	Rolling mill / hot press (minutes–hours)