3D Printing for Archaeology: Studying Dinosaur Fossils in Albumen Fossils

Some dinosaur fossils come out of the ground encased in a pale, translucent mineral layer that preparators describe as albumen-like. It is not egg white. It is a fine crystalline overgrowth, often calcite or silica, that formed after burial. Under a hand lens it looks milky; under X-rays it behaves almost like bone.

That similarity is the core problem. A standard medical CT scan cannot separate the two materials because their density and atomic number sit too close together. Mechanical preparation with an air scribe works on large bones but destroys the surface detail you wanted to measure. Acid prep is slower and can undercut thin features.

Micro-CT pushes the resolution down to a few micrometers per voxel. At that scale the matrix and the bone start to separate in the histogram, and a threshold can be drawn between them. The result is a voxel volume you can slice, measure and mesh without ever touching the original specimen.

The specimen itself never leaves the collection room. That matters for holotype material, for loaned specimens, and for any fossil where the curator will not approve destructive sampling.

Raw micro-CT output is a stack of grayscale slices. A typical 2,000-slice dataset at 5 μm voxels runs to tens of gigabytes. Before any printer sees it, the volume needs segmentation, smoothing and decimation. Skip that step and you get a file that chokes the slicer or prints with stair-stepped surfaces.

Segmentation is where the judgment lives. A single global threshold rarely works across a whole bone. We usually see labs use region growing plus manual paint passes on the ambiguous voxels near the matrix boundary. For a small bipedal herbivore bone, that can take a full day per specimen.

Decimation is the trade you make next. A 40-million-triangle mesh carries detail no printer can resolve and no viewer will open. Dropping to 2 to 5 million triangles keeps surface features down to the voxel size while making the file workable on a normal workstation.

One more pass matters for articulation: watertightness. Meshes from CT often have holes where the threshold failed. Slicers will repair small ones silently, which can shift geometry. Check the mesh in a dedicated tool before you commit a print run.

• 1
  Voxel size5–20 μm is typical for small bones; large limb bones can run coarser.
• 2
  Triangle budget2–5 million triangles keeps detail without freezing the slicer.
• 3
  Scale checkVerify a known landmark length against the mesh before printing.

Additive wins when the geometry is organic, hollow, or impossible to reach with a cutter. A braincase interior or a sinus cavity is a natural fit. So is any part where you only need one copy and the surface is the point.

Machining wins when the printed part has to hold a thread, carry a load, or stay flat over time. Resin creeps under sustained load. A mount bracket printed in resin will sag within months under a heavy cast. The same bracket cut from 6061-T6 aluminum holds its shape for years.

The two processes are often used in sequence on the same project. A resin print confirms the fit and the pose. Then the structural brackets, pins and base plate get machined from aluminum or stainless. The scan data feeds both paths from the same CAD reference.

For fossil work we keep tolerances realistic. A printed replica of a 200 mm bone might be printed at ±0.2 mm, which is fine for study. A machined aluminum mount for that bone can hold ±0.005 mm, which is what keeps the specimen from rocking on the base.

Standard SLA resin gives the sharpest surface for the least money. It is the default for study copies and for masters that will be molded. Keep the printed part out of direct sun and away from solvents, and it will last for years on a shelf.

When the replica has to survive handling by students or visitors, a cast urethane part from a printed master takes impact better than raw resin. Shore hardness can be tuned to match the feel of real bone.

For mounts and fixtures, aluminum 6061-T6 and 304 stainless cover most museum needs. Aluminum is light and machines fast. Stainless resists corrosion in humid storage rooms and takes a thread without inserts.

If the part has to be non-magnetic near a scanner, or needs to be lighter than aluminum, we look at titanium and engineering plastics like POM or PEEK. These are not cheap options, so they only make sense for a specific constraint.