A dihedral angle is the angle formed by two half-planes that share a common boundary line, called the edge of the angle. In three-dimensional Euclidean space, it measures how far one half-plane opens away from the other, as with the two covers of a book. The term also denotes the inclination between two intersecting planes, the interior angle between adjacent faces of a polyhedron, or an oriented angle describing molecular conformation. These usages require different conventions about which angle is measured.
Geometric definition
Let two half-planes have the same boundary line . Choose a point on , and in each half-plane draw a ray from perpendicular to . The angle between these rays is the plane angle of the dihedral angle and gives its measure.
Equivalently, intersect the half-planes with a plane perpendicular to . Their intersections are two rays, and the angle between them measures the dihedral angle. The result is independent of the point chosen along the edge.
A non-reflex dihedral angle has a measure from to radians, or from to . If a particular region around the edge is specified, a reflex angle greater than may also be relevant.
The edge is essential to this construction: an arbitrary oblique cross-section generally does not show the true dihedral angle.
Half-planes versus complete planes
Two distinct intersecting planes divide the space around their intersection line into four regions. Their cross-sectional angles alternate between and . Thus, a pair of complete planes does not by itself select one of these supplementary angles.
The angle between two unoriented planes is conventionally the smaller angle,
A polyhedron’s interior dihedral angle, by contrast, is selected by the region occupied by the solid and may be obtuse or reflex. Confusing these conventions can produce supplementary answers to the same geometric calculation.
Calculation using normal vectors
A normal vector to a plane is perpendicular to every direction lying in that plane. Normal vectors provide a convenient way to calculate angles between planes.
Suppose two planes have nonzero normal vectors and . Their smaller unoriented angle is
Here, is the dot product, and is the Euclidean norm. The absolute value makes the result independent of the arbitrary choice of sign for either normal.
For planes written in Cartesian coordinates as
normal vectors are
Consequently,
The constants affect the positions of the planes, not their inclination. Parallel planes are assigned an inclination of zero, although distinct parallel planes do not form a dihedral angle with an actual intersection edge.
Interior angles from outward normals
For adjacent faces of a convex polyhedron, let be outward unit normals and let be the interior dihedral angle. Then
or equivalently,
The interior angle is therefore supplementary to the angle between outward normals. This formula depends on the normals being consistently outward-pointing and on the local convexity of the edge. A reflex interior angle requires additional orientation or inside–outside information.
Angles in polyhedra
A dihedral angle belongs to an edge, whereas a face angle belongs to a vertex within a polygonal face. Neither should be confused with a solid angle, which measures a three-dimensional angular region around a point.
For regular polyhedra, symmetry makes all interior dihedral angles equal. Examples include:
| Polyhedron | Interior dihedral angle |
|---|---|
| Regular tetrahedron | |
| Cube | |
| Regular octahedron | |
| Regular dodecahedron | |
| Regular icosahedron |
In a regular tetrahedron, for example, every face is an equilateral triangle with face angles of , but adjacent faces meet at approximately .
For a nonconvex polyhedron, an interior dihedral angle may exceed . The two supporting planes alone cannot distinguish this reflex angle from its non-reflex counterpart; the solid’s interior must also be specified.
Signed dihedral angles
An oriented or signed dihedral angle distinguishes the two senses of rotation about an edge. It requires a directed axis, an ordering of the planes, and a sign convention. A common range is .
Let be a unit vector along the chosen edge direction, and let be consistently chosen unit normals to the planes. Since both normals are perpendicular to the edge, a signed angle between them can be written as
where denotes the cross product and is the two-argument inverse tangent.
The dot product determines the cosine, while the scalar triple product determines the signed sine. Reversing the axis direction or exchanging the two planes reverses the angle’s sign, modulo a full turn. Reversing just one normal changes the represented angle by , so normal orientation is part of the definition rather than an incidental choice.
Four-point and molecular torsion angles
In chemistry, four ordered atomic positions define a dihedral or torsion angle between planes and , measured about the central bond . It describes the relative orientation of the outer bonds, rather than merely the smaller angle between the complete planes.
Write the position vectors as , and define
One explicit signed convention is
In this convention, viewing along the central bond gives a torsion of zero when the projected outer bonds point in the same direction, and a torsion of when they point in opposite directions. Other conventions can reverse the sign, so comparisons between software packages require attention to atom ordering and orientation.
Torsion angles are central to describing molecular conformations. In proteins, backbone torsion angles conventionally denoted , , and describe rotations associated with successive backbone bonds. They are distinct from bond angles, which involve three atoms rather than four.
If either consecutive triple is collinear, its plane normal vanishes and the torsion angle is undefined.
Applications and computational considerations
In geometry and geometric modelling, dihedral angles describe the shape of polyhedral surfaces and the folding between adjacent mesh faces. In architecture and fabrication, they help specify joints and folds. In finite element meshing, very small or very large tetrahedral dihedral angles are among the indicators used to assess element quality.
Dihedral angles also provide a discrete measure of bending. Adjacent faces that form a locally flat surface have an interior angle of ; their departure from flatness can be represented by . Such constructions connect polyhedral geometry with differential geometry.
For reliable computation, several distinctions matter:
- Convention: the smaller plane angle, interior angle, exterior turning angle, and signed torsion are different quantities.
- Degeneracy: zero-length edges and collinear point triples can make a required axis or normal undefined.
- Numerical precision: floating-point arithmetic may produce a normalized dot product slightly outside ; implementations commonly clamp it before applying inverse cosine.
- Near-degenerate geometry: nearly collinear point triples yield small cross products and poorly determined normals.
- Angle wrapping: signed angles jump at the chosen branch boundary, even when the underlying geometry changes continuously.
For suitable vectors, an formulation using both sine and cosine information generally offers better numerical stability near zero and than inverse cosine alone.