Gyroid Infill: Strongest Lattice for AM

If you ask a slicer to fill the interior of a 3D print, it will default to rectilinear or triangular infill patterns. These patterns are fast to compute and easy to tool-path. They are also mechanically mediocre. When lightweight structures need to carry real loads, the gyroid is the lattice of choice.

The gyroid is a triply periodic minimal surface (TPMS) that partitions space into two interlocking labyrinths. It occurs in nature in butterfly wing scales and certain lipid bilayer configurations. In engineering, it provides an extraordinary combination of stiffness, strength, energy absorption, and manufacturability that no strut-based lattice can match.

What Is a TPMS?

A triply periodic minimal surface is a surface that repeats in three independent spatial directions and has zero mean curvature at every point. “Minimal surface” means it locally minimizes area for its boundary constraints, like a soap film stretched across a wire frame.

The mathematical family includes several named surfaces:

• Gyroid: sin(x)cos(y) + sin(y)cos(z) + sin(z)cos(x) = 0
• Schwarz P: cos(x) + cos(y) + cos(z) = 0
• Schwarz D (Diamond): cos(x)cos(y)cos(z) - sin(x)sin(y)sin(z) = 0
• Neovius: 3(cos(x) + cos(y) + cos(z)) + 4cos(x)cos(y)cos(z) = 0

Each surface is defined by a simple implicit equation. The zero-level set of the equation is the surface. Thickening the surface by offsetting in both normal directions creates a solid sheet lattice. Alternatively, solidifying one side of the surface creates a network lattice.

The gyroid is unique among TPMS in that it contains no straight lines and no planar symmetry elements. This makes it isotropic under loading: its mechanical response is nearly identical regardless of load direction. The Schwarz P surface, by contrast, has cubic symmetry planes that create directional weakness.

Mechanical Properties

Stiffness-to-Weight Ratio

For sheet-based gyroid lattices at 20% relative density (80% porosity), the Young’s modulus in any direction reaches approximately 30% of the Hashin-Shtrikman upper bound. This is the theoretical maximum stiffness achievable by any isotropic microstructure at that density. No strut-based lattice comes close.

Strut lattices like BCC (body-centered cubic) or octet-truss achieve 10-20% of the upper bound at the same density. The difference is because struts carry load through bending (inefficient), while TPMS sheets carry load through membrane stretching (efficient).

Strength

Gyroid lattices exhibit a stretching-dominated deformation mode, meaning they fail by material yielding rather than structural buckling. This gives them a more predictable and higher strength than bending-dominated lattices.

For AlSi10Mg printed by laser powder bed fusion, a 30% density gyroid lattice achieves compressive yield strengths of 40-60 MPa, compared to 20-35 MPa for a BCC strut lattice at the same density and material.

Energy Absorption

Under compression beyond yield, gyroid lattices exhibit a long, flat stress plateau before densification. This makes them excellent energy absorbers for crash and impact applications. The energy absorption per unit volume at 50% strain is typically 1.5-2x that of equivalent-density strut lattices.

The deformation mechanism during the plateau phase involves progressive layer-by-layer crushing, which is stable and predictable. Strut lattices often exhibit catastrophic shear band failure, which makes their energy absorption less reliable.

Fatigue

TPMS sheet lattices have inherently better fatigue resistance than strut lattices because sheet structures have smooth, continuous geometry without the stress concentrations at strut nodes. Node stress concentrations in strut lattices are typically 3-5x the nominal stress, which nucleates fatigue cracks.

Design Parameters

Wall Thickness

The primary design variable for a gyroid lattice is the wall thickness, which directly controls relative density. For the equation sin(x)cos(y) + sin(y)cos(z) + sin(z)cos(x) = t, the parameter t controls the offset from the minimal surface. Increasing t thickens the walls and increases density.

Practical wall thickness is constrained by the manufacturing process. Laser powder bed fusion in metals requires minimum wall thickness of 0.2-0.4 mm. Polymer SLS requires 0.5-1.0 mm. FDM requires 0.8-1.2 mm.

Cell Size

The unit cell size determines the number of repetitions across the part. Smaller cells create more surface area per volume (better for heat exchange) but require finer print resolution. Larger cells are easier to manufacture and inspect but may not adequately fill thin-walled regions.

Typical cell sizes for structural applications range from 2 mm to 10 mm. For heat exchangers, 1-3 mm cells maximize surface area. For energy absorption, 5-15 mm cells provide the desired progressive crushing behavior.

Relative Density

The ratio of solid volume to total volume. For load-bearing applications, 15-35% relative density is typical. Below 15%, the walls become too thin for reliable manufacturing. Above 35%, the weight savings diminish and solid material becomes more efficient.

Graded Gyroids

Uniform gyroids are useful but limited. Real engineering structures have non-uniform load distributions. A bracket is heavily loaded near its mounting points and lightly loaded at its center. A uniform lattice over-designs the center and under-designs the mount points.

Graded gyroids solve this by varying the lattice parameters spatially:

Density grading. The wall thickness parameter t varies as a function of position, creating denser lattice where loads are high and lighter lattice where loads are low. This is directly informed by topology optimization results.

Cell size grading. The unit cell dimensions vary across the part. Smaller cells near surfaces for better surface finish and larger cells in the interior for easier powder removal.

Morphology grading. Transitioning between different TPMS types (e.g., gyroid near the surface for isotropy, Schwarz P in the interior for directional stiffness) in a continuous, smooth manner.

Mathematically, grading is straightforward in implicit representation. The gyroid equation becomes:

sin(x/s(p))cos(y/s(p)) + sin(y/s(p))cos(z/s(p)) + sin(z/s(p))cos(x/s(p)) = t(p)

where s(p) is the spatially varying cell size and t(p) is the spatially varying thickness. This is a single field evaluation at every point. No geometry stitching, no transition zones, no Boolean operations.

NeuroCAD’s field-based architecture treats graded gyroids as a native operation. The density field from a topology optimization result directly drives the lattice parameter field. The designer connects the optimization output to the lattice generator, and the result is a continuously graded structure evaluated in real time.

Applications

Aerospace Brackets

Topology-optimized brackets for satellite and aircraft structures frequently use gyroid infill to replace the organic shapes that topology optimization produces. The optimized density distribution maps directly to a graded gyroid, producing a printable part with near-optimal stiffness-to-weight ratio.

Orthopedic Implants

Bone tissue ingrowth requires porous structures with 60-80% porosity and pore sizes of 300-800 microns. Gyroid lattices match trabecular bone morphology and mechanical properties better than strut lattices. The continuous, smooth surface minimizes stress shielding and promotes cell attachment.

Titanium gyroid implants printed by electron beam melting are now in clinical use for spinal fusion cages and acetabular cups.

Heat Exchangers

The gyroid surface partitions space into two non-intersecting channel networks. Each network can carry a different fluid, creating a compact heat exchanger with extremely high surface-area-to-volume ratio. At 30% relative density, the specific surface area exceeds 1,000 m^2/m^3.

Energy Absorption

Automotive crash structures, helmet liners, and packaging all benefit from the gyroid’s stable, predictable crushing plateau. Graded gyroids can be tuned to produce specific force-displacement curves, matching the protection requirements of different body regions in a helmet or different crash scenarios in a vehicle.

CAD Workflow

Designing gyroid structures in B-rep CAD is effectively impossible at scale. A single 5 mm gyroid unit cell contains a complex, non-planar surface that cannot be represented exactly by NURBS. Tiling this into a 100x100x100 mm volume creates millions of surface patches with intersection artifacts.

The practical workflow is:

1. Define the design domain as a boundary or field.
2. Run topology optimization to determine the density distribution.
3. Map the density field to gyroid parameters (wall thickness, cell size).
4. Evaluate the graded gyroid field over the design domain.
5. Boolean-intersect with the outer boundary (trivially, a max operation in SDF).
6. Export to STL/3MF for printing or convert to STEP via the B-spline fitting pipeline.

Steps 3-5 are single-expression field evaluations in an implicit geometry system. In NeuroCAD, this entire workflow is a connected graph of field operations that updates interactively when any parameter changes.

The Strongest Lattice, Naturally

The gyroid is not the strongest lattice because someone optimized it. It is the strongest isotropic lattice because minimal surfaces inherently maximize membrane load transfer. Nature discovered this billions of years ago. Engineers are finally catching up, enabled by additive manufacturing that can build what subtractive processes cannot, and by CAD systems that can represent what B-rep kernels cannot.

The gyroid equation is five terms long. Its engineering impact is unbounded.