Thermal Spray Coating: Painting Surfaces With Molten Metal

How It Works: Melt, Accelerate, Splat, Freeze

Every thermal spray process follows the same four-beat cycle: a heat source melts or softens a feedstock (powder or wire), a gas stream accelerates the particles toward the substrate, each particle impacts and flattens into a thin lamella, and the splat quenches almost instantly against the cold part. The coating is not a casting — it is a stack of overlapping frozen droplets, mechanically interlocked and metallurgically welded at their edges.

The governing physics is a race between two clocks. The particle must reside in the hot zone long enough to melt (or reach its plastic state) but impact before it re-solidifies in flight. Particle heating is diffusion-limited, so the melting time scales as t_melt ≈ ρ c_p d² / k, where d is particle diameter and k the thermal conductivity — meaning a 45-µm particle needs roughly 4× the dwell time of a 22-µm particle. On impact, the flattening ratio (splat diameter D over initial diameter d) obeys the classic scaling D/d ≈ 1.29 · Re^0.2, where Re = ρvd/µ is the droplet Reynolds number. A well-melted particle hitting at 300 m/s spreads to 3–5× its diameter, producing lamellae 1–5 µm thick.

• Cooling rate on impact is extreme: 10⁶–10⁸ K/s, comparable to melt-spinning. This splat quenching produces fine grains, metastable phases, and amorphous regions you could never get from bulk casting.
• Bonding is dominantly mechanical (interlocking on a grit-blasted, Rₐ ≈ 5–10 µm surface) with local metallurgical or diffusion bonds where the splat is hot enough to fuse the asperity.

The Physics of the Splat: Why Speed Beats Heat

For decades the intuition was "hotter is better." HVOF and cold spray proved the opposite: kinetic energy often matters more than thermal energy. A dense, well-bonded coating needs each particle to arrive with enough momentum to flatten fully and deform the underlying layer, expelling trapped gas and closing porosity.

The particle's kinetic energy density is ½ρv², so a jump from 100 m/s (arc) to 800 m/s (HVOF) is a 64× increase in impact energy per unit volume — even though the HVOF flame (~2800 °C) is far cooler than a plasma jet (~12000 °C). That is why HVOF tungsten carbide comes out at <2% porosity while plasma-sprayed WC is spongier and partly decarburized. In cold spray, particles never melt at all: they bond purely by solid-state adiabatic shear instability above a material-specific critical velocity (~550 m/s for copper, ~750 m/s for titanium), where the impact interface locally softens and the surfaces jet together like explosion welding.

• Oxidation control: hotter, slower, longer-dwell particles pick up more oxide films. Every oxide skin is a weak interface, so processes are traded on the melt-vs-oxidation balance — inert-gas plasma and cold spray minimize it.
• Residual stress: each quenching splat contracts and pulls tension into the layer (tensile quenching stress), while the peening of high-velocity impacts adds compression. Net through-thickness stress governs whether the coating stays put or delaminates.

Controlling Variables and Design Trade-offs

A thermal spray recipe has a dozen coupled knobs, and every one is a trade-off. The engineer's job is to hit a target microstructure — porosity, oxide content, phase composition, and bond strength — within a stable deposition window.

• Spray distance (standoff): typically 100–300 mm. Too short overheats the part and cracks the coating; too long lets particles cool and oxidize, dropping density and deposition efficiency (mass deposited / mass sprayed, typically 40–75%).
• Particle size distribution: fine powder (5–25 µm) melts easily but chokes feeders and oxidizes; coarse powder (45–90 µm) resists melting and bounces off. HVOF favors 15–45 µm.
• Gun traverse and part rotation: to build ~10 µm per pass evenly, a shaft might spin at 100–300 rpm while the gun traverses at 200–1000 mm/s, so surface speed × pass overlap sets thickness uniformity.
• Fuel/oxygen ratio (HVOF): stoichiometry sets flame temperature and, critically, carbon activity — a fuel-rich flame suppresses WC → W₂C decarburization.
• Cooling: CO₂ or air jets keep the substrate below its tempering/aging temperature (often 150–200 °C for hardened steel) to protect the base metal and manage residual stress.

The master trade-off: density and bond strength climb with particle velocity, but so do equipment cost and stress. Flame spray is cheap and forgiving for zinc; HVOF buys you aerospace-grade carbide at 5–10× the capital and gas cost.

Sizing, Scale, and the Numbers That Matter

A production HVOF gun runs on 20–40 kg/h of oxygen, 10–20 L/h of kerosene, delivers 80–200 kW of thermal power, and deposits WC-Co at 3–6 kg/h with a deposition efficiency near 50–65%. A plasma torch draws 40–80 kW electrically (500–1000 A at 40–80 V DC) through argon/hydrogen at 40–120 slpm.

Coating design starts from the required thickness and its tolerance. For a hydraulic rod, a designer might specify a 250 µm WC-Co-Cr layer, ground back to 200 µm final at a 0.4 µm Rₐ finish. Bond-strength qualification uses ASTM C633 (glue a stub to the coating, pull it, measure MPa) — HVOF carbides routinely exceed 70 MPa, comfortably above the 20–40 MPa of flame or plasma coatings. Hardness maps to wear life: HVOF WC-Co lands at 1000–1300 HV, versus ~800 HV for hard chrome.

• Thickness limit per pass is set by residual stress: build too thick too fast and the tensile quenching stress (often 50–200 MPa) exceeds the interface strength, and the coating spalls. Thick builds (>1 mm) are laid in many thin passes with interpass cooling.
• Thermal-barrier coatings exploit low conductivity: 7–8 wt% yttria-stabilized zirconia (YSZ) has k ≈ 1 W/m·K, so a 300 µm layer drops blade-surface temperature by ~150 °C, using Fourier's law q = k ΔT / t to trade thickness for temperature.

Real Applications and Hardware

Thermal spray is a quiet enabler across heavy industry, and the hardware is refreshingly tangible — a robot arm, a spray gun, a powder feeder, a grit-blast booth, and a dust-collection system rated for metal fines.

• Aerospace turbines: APS or EB-PVD YSZ thermal-barrier coatings over an MCrAlY bond coat let combustor and turbine sections run hundreds of degrees above the superalloy's creep limit. This single technology has been worth roughly a full stage of turbine efficiency.
• Chrome replacement: landing gear, hydraulic actuators, and print rolls now carry HVOF WC-Co-Cr or Cr₃C₂-NiCr, driven by the phase-out of hexavalent chromium electroplating under REACH and OSHA rules.
• Corrosion protection: twin-wire arc-sprayed zinc or 85/15 Zn-Al on bridges, ship ballast tanks, and wind-turbine towers gives 25+ year cathodic protection — the same galvanic sacrifice as galvanizing, but field-applicable on any size structure.
• Repair and rebuild: worn shafts, bearing journals, and pump impellers are restored to size with sprayed steel or bronze, salvaging castings that would cost 10× to replace.
• Biomedical: porous plasma-sprayed hydroxyapatite and titanium on hip and knee implants promote bone in-growth.

Governing specs include ISO 14917 (terminology/classification), AWS C2.23/SSPC-CS 23 (thermal-sprayed metallic coatings for corrosion), and MIL-STD-1687 for defense hardware.

Failure Modes, Limits, and Best Practice

Coatings fail at their weakest link — almost always an interface. Understanding the failure taxonomy is how you avoid a $1M engine unscheduled removal.

• Delamination / spallation: driven by residual stress and thermal-expansion mismatch. If the coating and substrate CTEs differ by Δα, each ΔT thermal cycle injects strain ε ≈ Δα·ΔT at the interface — the primary driver of TBC life. Graded bond coats and CTE-matched materials mitigate it.
• Poor adhesion: nearly always a surface-prep failure. The substrate must be grit-blasted to a clean, angular Rₐ ≈ 5–10 µm within a few hours (before re-oxidation), or the mechanical interlock never forms.
• Interconnected porosity: open pores let corrosives reach the substrate, so under-dense coatings can accelerate galvanic corrosion rather than prevent it. Dense HVOF or a sealer is required for barrier (non-sacrificial) coatings.
• Decarburization / phase change: overheated WC loses carbon to brittle W₂C and metallic W, gutting wear resistance — the reason HVOF (cooler, faster) beats plasma for carbides.
• Fatigue debit: hard, brittle coatings and tensile residual stress can lower the substrate's fatigue limit. Best practice compensates with a compressive shot-peen or a compressive-stress process (HVOF, cold spray).

Best practice checklist: grit-blast and spray within the same shift; validate bond strength per ASTM C633; keep substrate temperature under control; cross-section and measure porosity/oxide by image analysis (ASTM E2109); and never use a porous coating as a standalone corrosion barrier without sealing.

Comparison of the major thermal spray processes by energy source, particle velocity, and typical use.

Process	Heat source	Particle velocity	Porosity	Typical use
Flame (combustion) spray	Oxy-fuel flame, ~3000 °C	40–100 m/s	10–15%	Low-cost zinc/Al, rebuild
Twin-wire arc	Electric arc, ~6000 °C	100–150 m/s	5–10%	Corrosion-resistant Zn/Al, cheap
Atmospheric plasma (APS)	Plasma jet, 8000–15000 °C	200–400 m/s	2–8%	Ceramic TBCs (YSZ), Al₂O₃
HVOF	Supersonic oxy-fuel, ~2800 °C	600–1000 m/s	0.5–2%	WC-Co wear, dense carbides
Cold spray	Compressed gas, solid state	500–1200 m/s	<1%	Cu/Al repair, oxide-free