On face angles of tetrahedra with a given base

This paper investigates the geometric properties of tetrahedra with a fixed base by characterizing the closure and boundary of the set formed by the cosines of the face angles at the apex.

The Gist

Imagine you are holding a triangular piece of cardboard on a table. This is your base (let's call it triangle $ABC$). Now, imagine you have a magical, invisible point $D$ floating somewhere in the air above or below this table. If you connect $D$ to the three corners of your triangle, you create a 3D pyramid shape called a tetrahedron.

The paper by Nikitenko and Nikonorov is essentially a map-making expedition. They want to answer a very specific question: "If I move point $D$ around in the entire 3D universe (except right on the table), what are all the possible shapes of the three corners at the very top of my pyramid?"

Specifically, they aren't looking at the lengths of the edges. They are looking at the angles where the three top corners meet at point $D$. Let's call these angles $\alpha$, $\beta$, and $\gamma$.

The "Cosine" Translation

Angles are tricky to work with mathematically. So, the authors decide to translate these angles into cosines (a number between -1 and 1).

• Think of the three angles as a triplet of numbers: $(\cos \alpha, \cos \beta, \cos \gamma)$.
• The goal is to find the shape formed by all possible triplets you can get by moving $D$ around.

The "Pillow" and the "Pillowcase"

The authors discover that all these possible triplets fit inside a specific 3D shape they call the "Pillow" ($P$).

• The Pillow ($P$): This is a soft, rounded, 3D blob. It represents the absolute physical limits of geometry. No matter how you arrange your pyramid, the angles can never be so extreme that they fall outside this pillow.
• The Pillowcase ($BP$): This is the surface or the "skin" of the pillow. If you flatten your pyramid so that point $D$ touches the table (making the pyramid flat), your angles land exactly on this skin.

The Mystery of the "Danger Cylinder"

Here is where it gets interesting. The authors find that the "map" of all possible angles isn't just a solid blob; it has a complex boundary. To understand this boundary, they introduce a Cylinder.

Imagine a vertical tube standing up, passing through the center of your triangle $ABC$.

• If point $D$ is inside this tube, the angles behave one way.
• If point $D$ is outside this tube, they behave another way.
• If point $D$ is exactly on the surface of this tube, something magical happens: the geometry becomes "degenerate" or "singular." It's like a fold in the fabric of space. The map of angles gets crinkled here.

The authors prove that the "edges" of their final shape (the boundary of the Pillow) are formed by the images of this special cylinder.

The Three Types of Triangles

The shape of the final map depends entirely on the shape of your base triangle $ABC$:

1. The Sharp Triangle (Acute): All angles are less than 90°.

   • The Result: The map is a complex, multi-faceted shape. The "center" of the map is deep inside the solid volume. The boundary is made of three distinct "patches" coming from the cylinder, stitched together with the flat "skin" of the pillow. It looks like a 3D puzzle with three main pieces.

2. The Right Triangle: One angle is exactly 90°.

   • The Result: The map simplifies. The "center" point of the map moves right to the edge of the shape. The boundary is formed by just one special patch from the cylinder.

3. The Dull Triangle (Obtuse): One angle is greater than 90°.

   • The Result: Similar to the right triangle, but the geometry shifts slightly. The "center" point of the map actually moves outside the main shape, meaning the valid angles are restricted to a specific corner of the Pillow. Again, only one special patch from the cylinder defines the boundary.

The "Limit" Points

What happens if you move point $D$ infinitely far away?

• The angles get smaller and smaller, approaching zero.
• In the "cosine" world, this means the triplet approaches $(1, 1, 1)$. This is the "North Pole" of their Pillow.

What happens if $D$ gets very close to one of the corners of the triangle (say, $A$)?

• The angles behave wildly.
• The authors calculate that as $D$ approaches $A$, the possible angle triplets fill up a specific solid ellipse (a flattened, 3D oval).
• There are three such ellipses, one for each corner ($A$, $B$, and $C$). These act like "caps" or "end-caps" on the main shape.

The Big Picture: Why Does This Matter?

You might ask, "Who cares about the angles of a floating pyramid?"

The authors suggest this is crucial for real-world problems:

• Camera Tracking: If you have a camera taking a picture of three known points, this math helps figure out exactly where the camera is in 3D space.
• Molecular Geometry: Atoms form tetrahedral shapes. Understanding the limits of their angles helps chemists predict how molecules can or cannot twist and turn.
• Computer Graphics & Meshes: When building 3D models for movies or video games, you need to know the valid ranges of angles to ensure the models don't glitch or break.

Summary Analogy

Imagine you are a sculptor trying to carve a statue out of a block of clay (the Pillow).

• You can't just carve anywhere; the clay has a hard outer shell (The Pillowcase) that you can't break.
• Inside the clay, there are three "hard cores" (The Limit Ellipses) near the corners that you have to carve around.
• There is a special "fold" in the clay (The Cylinder) where the texture changes.
• Depending on whether your base triangle is sharp, right, or dull, the final shape of your statue changes:
  • Sharp Base: You carve three distinct wings.
  • Right/Dull Base: You only need to carve one wing.

The paper provides the exact blueprint for this sculpture, telling you exactly where the boundaries are and how the shape changes based on the triangle you start with.

Technical Summary

1. Problem Statement

The paper addresses a geometric problem concerning the set of face angles of tetrahedra sharing a fixed, non-degenerate base triangle $ABC$ in 3-dimensional Euclidean space ($E^3$).

• Given: A fixed triangle $ABC$ with side lengths $x=d(B,C)$, $y=d(A,C)$, and $z=d(A,B)$.
• Variable: A point $D$ lying outside the plane of $ABC$ (denoted as $\Pi_0$).
• Target Set: The set $\Sigma(\triangle ABC)$ consisting of all triples of cosines of the face angles at vertex $D$:
  $$\Sigma(\triangle ABC) = \{ (\cos \alpha, \cos \beta, \cos \gamma) \in \mathbb{R}^3 \mid \alpha = \angle BDC, \beta = \angle ADC, \gamma = \angle ADB \}$$
• Objective: To determine the closure of $\Sigma(\triangle ABC)$ in $\mathbb{R}^3$ and explicitly describe its boundary. The authors analyze how this set depends on the geometry of the base triangle (acute, right, or obtuse).

2. Methodology

The authors employ a combination of algebraic geometry, differential topology, and classical Euclidean geometry.

A. Parametrization and Mapping

The problem is formalized using a mapping $F$ from the space of possible positions of $D$ to the space of cosine triples.

• Domain: The set $G_T = \Pi_+ \cup (\Pi_0 \setminus \{A, B, C\})$, where $\Pi_+$ is the half-space above the base plane.
• Mapping: $F: G_T \to \mathbb{R}^3$ defined by $F(D) = (\cos \angle BDC, \cos \angle ADC, \cos \angle ADB)$.
• Coordinates: The study utilizes both distance coordinates (squared distances from $D$ to vertices) and Cartesian coordinates. Cartesian coordinates are preferred for analyzing the smoothness and degeneracy of the map.

B. The "Pillow" and "Pillowcase"

The authors identify a universal bounding set $P$ (the "pillow") defined by the inequality:
$$1 + 2xyz - x^2 - y^2 - z^2 \geq 0$$
where $x, y, z$ are the cosines of the face angles. This inequality is derived from the volume formula of a tetrahedron (specifically, the condition that the volume squared must be non-negative).
The boundary of this set, $BP$ (the "pillowcase"), is defined by the equality $1 + 2xyz - x^2 - y^2 - z^2 = 0$.

C. Analysis of Degeneracy

A critical step involves identifying points where the map $F$ is degenerate (i.e., the Jacobian determinant vanishes).

• Result: The map $F$ is degenerate if and only if $D$ lies in the plane of the base ($\Pi_0$) or on the right circular cylinder ($Cyl$) passing through the circumcircle of $\triangle ABC$.
• Implication: The boundary of the image set $\Sigma(\triangle ABC)$ is largely determined by the images of these degenerate sets.

D. Limit Analysis

The authors analyze the behavior of $F(D)$ as $D$ approaches:

1. The vertices $A, B, C$: This leads to "limit solid ellipses" ($Lim(A), Lim(B), Lim(C)$).
2. The circumcircle (at infinity or on the cylinder): This generates specific curves and surfaces on the boundary.
3. The plane $\Pi_0$: The image of the plane (excluding vertices) lies on the "pillowcase" $BP$.

3. Key Contributions and Results

A. Topological Structure

• Theorem 3: The closure of the set $\Sigma(\triangle ABC)$, denoted $\overline{\Sigma}$ (or $C_{FT}$), is homeomorphic to a closed 3-dimensional ball.
• The boundary of this set, $\partial \overline{\Sigma}$, is homeomorphic to a 2-dimensional sphere ($S^2$).

B. Decomposition of the Boundary

The boundary $\partial \overline{\Sigma}$ is constructed from several distinct geometric components:

1. The "Pillowcase" ($BP$): The image of the base plane $\Pi_0$ (excluding vertices) lies on $BP$.
2. Limit Solid Ellipses: Convex hulls of ellipses formed by the limits as $D$ approaches the vertices $A, B, C$.
3. Images of Special Regions on the Cylinder: The map $F$ restricted to the cylinder $Cyl$ (over the circumcircle) produces a smooth surface of positive Gaussian curvature (except at three "edges"). The boundary of the final set includes specific "special regions" of this cylinder image.
4. The Point at Infinity: The point $(1, 1, 1)$ corresponds to $D$ moving infinitely far away.

C. Dependence on Base Triangle Type

The structure of the boundary changes significantly based on whether $\triangle ABC$ is acute, right, or obtuse:

• Acute-angled Base:
  • The point $I = (\cos A, \cos B, \cos C)$ lies in the interior of the closure.
  • The boundary contains images of three distinct special regions on the cylinder.
  • The set $P \setminus \overline{\Sigma}$ consists of three disjoint convex sets.
• Right-angled Base:
  • The point $I$ lies on the boundary.
  • The boundary contains the image of one special region on the cylinder.
• Obtuse-angled Base:
  • The point $I$ lies outside the closure.
  • The boundary contains the image of one special region on the cylinder.

D. Explicit Formula for the Set

The paper provides a precise characterization of the set $\Sigma(\triangle ABC)$:
$$\Sigma(\triangle ABC) = \text{int}(\overline{\Sigma}) \cup F(SRT)$$
where $SRT$ is the set of points on the cylinder belonging to the "special regions" or their boundaries. This formula separates the interior points (generated by non-degenerate points off the cylinder) from the boundary points (generated by points on the cylinder and the base plane).

4. Significance and Applications

• Geometric Characterization: The paper provides a complete solution to the inverse problem of determining which triples of face angles are realizable for a tetrahedron with a given base.
• Connection to Perspective 3-Point (P3P) Problem: The mapping $F$ is directly related to the P3P problem in computer vision (determining camera position from three known points). The degeneracy analysis (the "danger cylinder") is a known phenomenon in P3P, and this paper rigorously characterizes the resulting solution space in terms of face angles.
• Molecular Geometry and Meshing: The results are applicable to molecular geometry (bond angles in tetrahedral structures) and the generation of tetrahedral finite-element meshes, where constraints on face angles are crucial.
• Topological Insight: The proof that the solution space is a 3-ball with a spherical boundary provides a robust topological framework for understanding the space of tetrahedra.

5. Conclusion

Nikitenko and Nikonorov successfully characterize the set of face angles for tetrahedra with a fixed base. By analyzing the mapping from spatial coordinates to cosine triples, they identify that the solution set is a topological 3-ball bounded by a complex surface composed of parts of a "pillowcase" (algebraic surface), limit ellipses, and images of specific regions on a cylinder. The nature of this boundary is shown to be fundamentally dependent on the acute, right, or obtuse nature of the base triangle.