The Complete Guide to Metal 3D Printing: Melt Pool Physics, Process Science, and Material Metallurgy

Complete guide to metal additive manufacturing physics. Covers Rosenthal heat equation, keyhole vs conduction mode, Marangoni flow, 10⁵–10⁷ K/s cooling rates, residual stress, Ti-6Al-4V/Inconel 718/AlSi10Mg/316L metallurgy, process defects, and DFM rules.

2026-07-03 · FabFlow Team

The Complete Guide to Metal 3D Printing: Melt Pool Physics, Process Science, and Material Metallurgy

Metal additive manufacturing doesn't just "3D print metal" — it performs micrometer-scale casting, welding, and heat treatment thousands of times per second, all inside a single build chamber. Understanding the physics behind each of these phenomena is the difference between a part that passes aerospace certification and one that cracks during post-processing.

This guide covers the melt pool dynamics, heat transfer equations, solidification metallurgy, residual stress mechanics, and material-specific behavior that govern every major metal AM process.

1. The Technology Landscape: Four Ways to 3D Print Metal

graph TD
    A["Metal Additive<br/>Manufacturing"] --> B["Powder Bed Fusion<br/>(PBF)"]
    A --> C["Directed Energy<br/>Deposition (DED)"]
    A --> D["Binder Jetting<br/>(BJ)"]
    A --> E["Sheet Lamination<br/>(UAM)"]
    
    B --> B1["Laser PBF<br/>(DMLS / SLM)"]
    B --> B2["Electron Beam PBF<br/>(EBM)"]
    
    C --> C1["Laser DED<br/>(LENS)"]
    C --> C2["Electron Beam DED<br/>(EBAM)"]
    C --> C3["Arc DED<br/>(WAAM)"]
    
    style B1 fill:#ff922b,color:#fff
    style B2 fill:#ff922b,color:#fff

Laser Powder Bed Fusion (LPBF / DMLS / SLM) dominates the market (>80% of metal AM systems). It's what most people mean when they say "metal 3D printing." This guide focuses primarily on LPBF physics, with comparisons to EBM and DED where the differences matter.

2. Melt Pool Physics: The Heart of Metal AM

2.1 The Rosenthal Equation — Steady-State Heat Flow

The Rosenthal solution for a moving point heat source on a semi-infinite plate is the foundational model for understanding melt pool geometry in LPBF:

Where:

T(x,y,z) = temperature at point (x,y,z) relative to the moving heat source
T_0 = preheat / ambient temperature (typically 100–200°C for LPBF build plate)
\eta = laser absorptivity of the powder bed (0.35–0.70 depending on material, wavelength, and powder morphology)
P = laser power (typically 200–500 W for LPBF)
k = thermal conductivity of the solid material (W/m·K)
R = \sqrt{x^2 + y^2 + z^2} = distance from heat source
v = scan speed (typically 500–1500 mm/s)
\kappa = k/(\rho c_p) = thermal diffusivity

This equation reveals the fundamental tradeoff in LPBF: higher scan speeds produce smaller, shallower melt pools; higher laser power makes them larger and deeper.

2.2 Melt Pool Dimensions

The melt pool width W, depth D, and length L are the critical parameters that determine track stability, porosity, and microstructure:

Parameter · Typical Range (LPBF) · What It Controls

Melt pool width W · 80–200 μm · Overlap between adjacent tracks

Melt pool depth D · 30–150 μm · Layer adhesion, lack-of-fusion porosity

Melt pool length L · 200–600 μm · Solidification rate (longer pool = slower cooling)

Aspect ratio D/W · 0.3–0.8 · Above 0.8 → keyhole mode (unstable)

2.3 Conduction Mode vs. Keyhole Mode

The energy density at the laser spot determines whether you're in safe conduction mode or dangerous keyhole mode:

Where d_s is the laser spot diameter (typically 50–100 μm).

Mode · E_d Range · Melt Pool Shape · Result

Conduction · Low–moderate · Wide, shallow (D/W < 0.5) · Stable, low porosity

Transition · Moderate · ~Hemispherical (D/W \approx 0.5–0.8) · Good — target this

Keyhole · High · Deep, narrow (D/W > 0.8) · Vapor depression forms; gas entrapment → spherical porosity

In keyhole mode, the laser intensity is high enough to vaporize metal, creating a deep vapor cavity (keyhole) in the melt pool. The cavity walls are unstable — they collapse, trapping gas that becomes spherical pores in the final part. This is the dominant source of process-induced porosity in LPBF.

3. Marangoni Convection: Why the Melt Pool Stirs Itself

The surface tension of liquid metal decreases with increasing temperature (d\gamma/dT < 0 for most metals). In the melt pool, the center (under the laser) is hottest, creating a surface tension gradient that drives fluid flow outward:

Where Ma is the Marangoni number — the ratio of surface-tension-driven flow to thermal diffusion.

For LPBF conditions, Ma \gg 1, meaning Marangoni convection dominates heat transfer in the melt pool. Flow velocities can reach 1–10 m/s in pools only 100 μm wide — incredibly vigorous stirring at microscale. This is beneficial for mixing alloying elements but can also create surface ripples and spatter.

The Sign of d\gamma/dT Matters

Sign · Behavior · Materials

d\gamma/dT < 0 (normal) · Flow outward from center → wide, shallow pool · Most pure metals, Ti-6Al-4V

d\gamma/dT > 0 (anomalous) · Flow inward → deep, narrow pool · Some steels with active surface elements (S, O, Se)

The presence of oxygen or sulfur at ppm levels can flip the sign of d\gamma/dT for steels, dramatically changing melt pool shape. This is why powder chemistry trace elements matter enormously in LPBF — they aren't just about alloy composition, they affect the physics of the process itself.

4. Solidification: Microstructure from Microseconds

4.1 Cooling Rates

The cooling rate in LPBF is extreme — among the fastest in any manufacturing process:

For comparison:

Conventional casting: 10⁻¹–10¹ K/s
Welding: 10¹–10³ K/s
Melt spinning (rapid solidification): 10⁴–10⁶ K/s
LPBF: 10⁵–10⁷ K/s — rivaling the fastest solidification techniques known

This produces microstructures impossible through conventional processing: supersaturated solid solutions, metastable phases, ultra-fine grains (0.5–5 μm), and suppression of detrimental intermetallic precipitates.

4.2 Thermal Gradient and Growth Rate

The solidification microstructure is determined by two parameters at the solid-liquid interface:

Thermal gradient G = dT/dx (K/m) — how steeply temperature drops ahead of the interface
Solidification growth rate R = dx/dt (m/s) — how fast the interface advances

Their product G \cdot R = \dot{T} gives the cooling rate. Their ratio G/R determines the solidification morphology:

G/R Ratio · Microstructure

High (G/R) · Planar (flat interface) — rare in LPBF

Moderate · Cellular (elongated grains along thermal gradient)

Low (G/R) · Equiaxed (randomly oriented grains) — center of melt pool

LPBF typically operates in the cellular-to-equiaxed range, producing columnar grains that grow epitaxially from the previous layer, creating a characteristic textured microstructure.

4.3 Epitaxial Grain Growth

Because the melt pool partially re-melts the previous layer, solidification starts from the existing grains at the melt pool boundary. Grains grow along the direction of maximum thermal gradient — typically perpendicular to the melt pool boundary, curving to follow the laser path:

graph TD
    A["Layer N: melt pool solidifies<br/>→ columnar grains form"] --> B["Layer N+1: laser scans<br/>→ re-melts top of Layer N"]
    B --> C["Solidification starts from<br/>existing grains (epitaxial)"]
    C --> D["Grains grow across<br/>layer boundary"]
    D --> E["Result: Strong Z-direction<br/>properties, textured microstructure"]
    
    style E fill:#51cf66,color:#fff

This epitaxial growth across layers is why properly processed LPBF parts have excellent Z-direction tensile strength — there's no discrete "layer line" weakness as in polymer AM. The grain structure is continuous across the build.

5. Residual Stress and Distortion: The Hidden Cost

5.1 Why Metal AM Parts Warp

The extreme thermal gradients that make LPBF possible also create its biggest challenge: residual stress. As the laser passes, a thin layer of material expands. The surrounding cold material constrains this expansion, creating compressive plastic strain. Upon cooling, the hot region contracts, but now it's shorter than it "wants" to be — producing tensile residual stress:

Where:

E = Young's modulus (110 GPa for Ti-6Al-4V, 200 GPa for steels)
\alpha = coefficient of thermal expansion (8.6 \times 10^{-6} /K for Ti-6Al-4V)
\Delta T = temperature range over which stress develops (~1000–1500°C)
f_{constraint} = degree of mechanical constraint (0–1)

For Ti-6Al-4V, this gives \sigma_{residual} \approx 110 \cdot 8.6\times10^{-6} \cdot 1200 \approx 1100 MPa — exceeding the yield strength at elevated temperature. This is why supports aren't just for overhangs in metal AM — they're mechanical anchors that resist the part pulling itself off the build plate.

5.2 Stress Relief and HIP

Stress relief (annealing): Heat to 600–800°C for 1–4 hours. Reduces residual stress by 70–90% but doesn't close internal porosity.
Hot Isostatic Pressing (HIP): Heat to 900–1150°C at 100–200 MPa argon pressure for 2–4 hours. Closes internal porosity, homogenizes microstructure, and eliminates residual stress. Required for aerospace-critical parts.
Build plate stress relief: Always stress-relieve before cutting parts off the build plate. Cutting a stressed part off a cold plate can cause immediate distortion.

6. Material-Specific Metallurgy

6.1 Ti-6Al-4V (Titanium Grade 5) — The Aerospace Standard

Ti-6Al-4V accounts for ~40% of all LPBF production. Why:

Phase transformation: Above the \beta-transus (995°C), Ti-6Al-4V is 100% BCC β-phase. Below it, it transforms to HCP α-phase + β. At LPBF cooling rates (>10⁵ K/s), the β→α transformation produces an acicular (needle-like) α′ martensite — extremely fine, with high strength (UTS ~1200–1300 MPa as-built) but low ductility.
Post-process heat treatment: HIP at 920°C + 100 MPa converts martensitic α′ to lamellar α+β, reducing strength to ~950–1050 MPa but restoring ductility to >10% elongation.
Oxygen pickup: Ti-6Al-4V is highly reactive. The build chamber is flooded with argon to maintain O₂ < 1000 ppm. Higher O₂ = higher strength but severe embrittlement.

6.2 Inconel 718 — The Hot Section Champion

Nickel-based superalloy for applications up to 700°C:

Precipitation strengthening: γ″ (Ni₃Nb, body-centered tetragonal) and γ′ (Ni₃(Al,Ti), FCC) precipitates provide high-temperature strength through coherent strain fields that impede dislocation motion.
LPBF advantage: The rapid solidification suppresses Laves phase formation (a brittle Nb-rich intermetallic that forms during slow cooling in casting). LPBF Inconel 718 can achieve better homogenization than cast material.
Post-processing mandatory: Solution treatment (980°C, 1 hr) → aging (720°C, 8 hr + furnace cool to 620°C, 8 hr). This sequence is taken directly from the AMS 5662 aerospace specification.

6.3 AlSi10Mg — The Prototyping Workhorse

Aluminum-silicon casting alloy, the most common aluminum for LPBF:

Near-eutectic composition: Al-10Si-0.4Mg. The Si content depresses the melting point and narrows the freezing range, making it more processable than wrought aluminum alloys (6061, 7075) which crack severely during LPBF.
Microstructure: Fine cellular-dendritic structure with Si particles decorating cell boundaries. Fine (200–500 nm cell size) due to rapid solidification.
Heat treatment: Direct aging at 160–170°C for 4–6 hours precipitates Mg₂Si strengthening particles. Solution treatment at 530°C dissolves the Si network, coarsening grains but improving ductility.

6.4 316L Stainless Steel — Corrosion Resistance

LPBF advantage over wrought: The rapid solidification suppresses Cr-carbide precipitation at grain boundaries (sensitization), a common issue in welded 316L. LPBF 316L can have superior intergranular corrosion resistance compared to wrought 316L.
Hierarchical microstructure: Cellular sub-grains (0.5–1 μm) within larger columnar grains, with dislocation cell walls enriched in Cr and Mo. This unique structure gives LPBF 316L simultaneously higher strength and ductility than wrought.

7. Process Defects and Their Physics

7.1 Lack-of-Fusion Porosity

When energy density is too low, adjacent tracks or layers don't fully fuse. These defects are irregular in shape (not spherical) and act as stress concentrators.

Where A_{abs} is the absorptivity — the minimum energy density to fully melt a layer of thickness equal to the powder layer (~30–50 μm).

7.2 Keyhole Porosity (Spherical Pores)

As discussed in Section 2.3 — too much energy density creates a vapor cavity that collapses and traps gas. These pores are spherical and typically 10–100 μm.

7.3 Balling

When the melt pool breaks up into disconnected spheres rather than a continuous track — occurs when the melt pool length-to-width ratio is too high. The Rayleigh-Plateau instability (the same physics that makes a water stream break into droplets) governs this:

If the melt pool length exceeds \pi \times its diameter, balling is likely.

7.4 Cracking

Solidification cracking (hot cracking): Occurs in the mushy zone during the last stages of solidification when liquid films between dendrites can't accommodate thermal contraction strains. High-alloy content materials (Al 6061, Al 7075, some Ni-based superalloys) are especially susceptible.
Solid-state cracking (cold cracking): Residual stress exceeds the material's fracture toughness after cooling. Most common in crack-sensitive materials (high-carbon steels, some Ti alloys with high interstitial content).

8. DFM for Metal AM

8.1 Support Structures — Not Optional

Metal AM supports serve three functions:

Heat conduction — anchor the part to the build plate and conduct heat away (prevents overheating)
Mechanical restraint — resist the part pulling itself off the plate due to residual stress
Overhang support — prevent collapse of downward-facing surfaces

Feature · Support Required? · Guideline

Overhang >45° from horizontal · Yes · Add block or tree supports

Overhang <45° from horizontal · Usually no · May need heat conduction supports for large surfaces

Horizontal holes/bridges >2mm · Yes · Especially on lower surfaces

Vertical holes · No · Self-supporting

First layer of any part · Yes · Directly on build plate or support structure

8.2 The 45° Rule

The self-supporting angle for LPBF is approximately 45° from horizontal (downskin angle). Below this, the overhanging material sags or curls upward (the "super-elevation" effect) due to residual stress in the newly solidified layer.

8.3 Powder Removal

Any internal channel must have at least two openings for powder removal after the build. Minimum channel diameter: ~1mm for short channels, scaling up with length. Powder trapped in blind holes or single-opening cavities is essentially permanent.

8.4 Wall Thickness

Material · Minimum Wall · Recommended Minimum

Ti-6Al-4V · 0.3 mm · 0.5 mm

Inconel 718 · 0.4 mm · 0.6 mm

AlSi10Mg · 0.3 mm · 0.5 mm

316L · 0.3 mm · 0.5 mm

9. Post-Processing: The Hidden Half of Metal AM

A metal AM part fresh off the build plate is not finished. The full workflow:

graph LR
    A["Build<br/>Complete"] --> B["Stress Relieve<br/>on build plate"]
    B --> C["Wire EDM / Saw<br/>cut from plate"]
    C --> D["Remove supports<br/>manual or CNC"]
    D --> E["HIP<br/>(if required)"]
    E --> F["Heat Treat<br/>to spec"]
    F --> G["Machine critical<br/>surfaces"]
    G --> H["Surface finish<br/>(bead blast, polish)"]
    H --> I["Inspection<br/>(CT, UT, dye pen)"]
    
    style A fill:#ff922b,color:#fff
    style I fill:#51cf66,color:#fff

The post-processing cost is typically 30–60% of the total part cost. Budget for it from the start.

10. Cost Model for Metal LPBF

Where:

C_{machine} = t_{build} \times R_{machine} — hourly machine rate (₹3,000–₹8,000/hr in India, 50–150/hr globally)
C_{material} = m_{part} \times R_{powder} + m_{waste} \times R_{waste}

- Powder cost: ₹5,000–₹15,000/kg for Ti-6Al-4V, ₹8,000–₹25,000/kg for Inconel 718 - Powder reuse: After sieving, 70–90% of unfused powder can be reused (but properties degrade with each cycle due to satellite particle formation and oxygen pickup)

C_{post} = stress relief + support removal + HIP + machining + surface finish + inspection

Build time estimation:

For a 100mm tall part at 50μm layer thickness: n_{layers} = 2000. Recoat time ~7–10 seconds per layer. Total recoating time alone: ~4–5.5 hours. Scan time depends on cross-sectional area but typically adds 2–3× the recoating time for dense parts.

11. When Metal AM Wins (and When It Doesn't)

✅ Use Metal AM For:

Topology-optimized parts — complex organic shapes impossible to machine
Conformal cooling channels — follow the part contour, impossible to drill
Part consolidation — 50 welded components → 1 printed part
Low-volume production (1–500 units) of complex geometries
Lattice structures — bone-like porous scaffolds for medical implants
Rapid tooling — injection mold inserts with conformal cooling in days, not weeks

❌ Skip Metal AM For:

Simple prismatic parts (cheaper to CNC machine)
High-volume production >5,000 units (injection molding, casting, or forging wins)
Parts requiring wrought mechanical properties without post-processing
Very large parts (LPBF build volumes max out at ~400×400×400mm)
Tight-tolerance surfaces (always need post-machining — budget for it)

12. The Indian Metal AM Ecosystem

India's metal AM capacity has grown significantly since 2022. Key developments:

ISRO has qualified LPBF components for satellite propulsion systems (Inconel 718 thrust chambers)
Wipro 3D and Intech Additive operate production-scale LPBF farms in Bangalore
IIT Bombay, IIT Madras, and ARCI Hyderabad run active metal AM research programs
DMRL (DRDO) has developed indigenous metal powder production capabilities for Ti-6Al-4V and Ni-based superalloys

On FabFlow, you can source LPBF metal parts from verified Indian manufacturers — upload your design, specify Ti-6Al-4V, Inconel 718, AlSi10Mg, or 316L, and get quotes from qualified shops.

References

DebRoy, T., Wei, H. L., Zuback, J. S., Mukherjee, T., Elmer, J. W., Milewski, J. O., Beese, A. M., Wilson-Heid, A., De, A., & Zhang, W. (2018). Additive manufacturing of metallic components — Process, structure and properties. Progress in Materials Science, 92, 112–224. — Comprehensive review covering melt pool physics, heat transfer models, solidification, and residual stress across all metal AM processes.

Rosenthal, D. (1946). The theory of moving sources of heat and its application to metal treatments. Transactions of the ASME, 68, 849–866. — The foundational heat flow model used in Section 2.1.

Khairallah, S. A., Anderson, A. T., Rubenchik, A., & King, W. E. (2016). Laser powder-bed fusion additive manufacturing: Physics of complex melt flow and formation mechanisms of pores, spatter, and denudation zones. Acta Materialia, 108, 36–45. — High-fidelity simulation of melt pool dynamics, Marangoni flow, keyhole formation, and defect mechanisms.

Thijs, L., Verhaeghe, F., Craeghs, T., Van Humbeeck, J., & Kruth, J. P. (2010). A study of the microstructural evolution during selective laser melting of Ti-6Al-4V. Acta Materialia, 58(9), 3303–3312. — Detailed microstructural analysis of LPBF Ti-6Al-4V including martensitic transformation, epitaxial growth, and texture development.

Wang, Y. M., Voisin, T., McKeown, J. T., Ye, J., Calta, N. P., Li, Z., Zeng, Z., Zhang, Y., Chen, W., Roehling, T. T., Ott, R. T., Santala, M. K., Depond, P. J., Matthews, M. J., Hamza, A. V., & Zhu, T. (2018). Additively manufactured hierarchical stainless steels with high strength and ductility. Nature Materials, 17(1), 63–71. — Discovery of the unique cellular dislocation structure in LPBF 316L that gives simultaneous strength and ductility improvements.

Mercelis, P. & Kruth, J. P. (2006). Residual stresses in selective laser sintering and selective laser melting. Rapid Prototyping Journal, 12(5), 254–265. — Foundational work on thermal stress development in metal PBF, including the constraint factor model.

FabFlow. (2026). Metal 3D Printing Services — LPBF / DMLS from Verified Indian Manufacturers. [Online]. Available: https://www.fabflow.app