Areal Energy Density (Laser Powder Bed Fusion)

Areal energy density is the same accounting with one fewer division: $E_a = P/(v h)$, the joules delivered per unit of scanned area. It sits between two other numbers worth naming. Above it is the volumetric energy density, which divides once more by the layer thickness. Below it is the line energy, $P/v$, which is the joules delivered to each millimetre of a single track before the hatch spacing spreads them sideways.

Of the three, line energy is the most physical, because a single track is a real object you can build, section and photograph, while a hatched area is an accounting construct and a layer of powder is an accounting construct with a thickness. Areal energy density is the middle case and it earns its place for one specific reason: it leaves out the layer thickness, and the layer thickness is a property of the recoater rather than of the beam. When you are comparing two laser settings at one layer thickness — which is what most parameter development actually is — the layer thickness in $E_v$ is a constant that adds no information and slightly obscures the comparison. Dropping it is honest.

The unit deserves a word. Energy per unit area in SI is J/m², and a laser powder-bed figure of 2.5 J/mm² is 2,500,000 J/m². This site pins the picker to MJ/m² because 1 MJ/m² is 1 J/mm², exactly, so the number on the screen reads the way the trade writes it while the engine keeps its SI honesty underneath. The same type carries laser fluence, wave energy density and arc-flash incident energy, which are all the same dimension wearing different trade names.

The warnings that apply to volumetric energy density apply here too, only slightly less severely. It still averages a Gaussian beam over a rectangle the beam never illuminated uniformly — the centre of a track receives several times the mean while the edges receive almost nothing, and melting is a threshold process that does not care about means. It still contains no absorptivity, so it is not comparable between materials or between powder conditions. And it still cannot distinguish a fast beam at high power from a slow one at low power, which is the whole of Bertoli's finding. Use it for what it is good at, which is ranking settings on one machine at one layer thickness, and go to the normalised enthalpy when the question is what the melt pool is actually doing.