On the rationality of some real threefolds

This paper investigates the rationality of geometrically rational real threefold bundles with connected real loci and vanishing intermediate Jacobian obstructions, employing unramified cohomology, birational rigidity, and explicit constructions to establish both negative and positive results.

The Gist

Imagine you are an architect trying to figure out if a strange, complex building can be completely rebuilt into a simple, empty warehouse (a 3D grid of empty space). In the mathematical world, this building is called a "variety," and the empty warehouse is called "rational space."

If you can take the complex building, tear it apart, and reassemble it into the empty warehouse without losing any information, mathematicians say the building is rational. If you can't, it's irrational.

For a long time, mathematicians knew how to solve this puzzle for simple 2D buildings (surfaces) and some 3D buildings over "perfect" fields (like the complex numbers). But when you move to real numbers (the numbers we use in everyday life, including negatives and irrationals), the puzzle gets much harder.

This paper by Olivier Benoist and Alena Pirutka tackles a specific group of tricky 3D buildings defined by equations involving squares (like $x^2 + y^2 + z^2 = \dots$). These buildings have a special property: their "real" shape is connected (you can walk from any point to any other point without jumping), which is usually a good sign for being rational. However, the usual tools for proving they are not rational fail to work here.

Here is the breakdown of their findings, using some everyday analogies:

1. The "Ghost" Obstruction (Type 1 Buildings)

The authors first look at a family of buildings defined by the equation:
$$x^2 + y^2 + z^2 = u \cdot p(u)$$
Think of this as a shape where the "width" in the $x, y, z$ directions depends on a variable $u$.

• The Problem: Standard tools (like checking the "intermediate Jacobian," which is like checking the building's internal wiring) say these buildings might be rational. They pass all the standard tests.
• The Discovery: The authors found a way to prove that some of these buildings are actually not rational.
• The Metaphor: Imagine you have a locked box. The standard key (the usual math tools) doesn't fit, so you assume the box is empty (rational). But the authors invented a new, invisible key (called "unramified cohomology"). They showed that for certain specific versions of these buildings, this invisible key fits perfectly, proving there is a hidden lock inside.
• The Result: They proved that for any fixed complexity level, there is no single "magic recipe" that can turn all these buildings into a warehouse. Some are just too weird to be simplified.

2. The "Low vs. High" Complexity (Type 2 Buildings)

Next, they looked at a second family of buildings defined by:
$$x^2 + y^2 = f(v, w)$$
Here, the shape depends on a curve $f(v, w) = 0$.

• The Good News (Low Degree): If the curve $f$ is simple (degree 4 or less, like a circle or a simple squiggle), the building is rational.
  • Analogy: It's like a house with a simple floor plan. You can easily knock down a wall and turn it into an open loft. The authors actually showed you how to do the demolition and reconstruction for these simple cases.
• The Bad News (High Degree): If the curve $f$ is very complex (degree 12 or higher), the building is never rational.
  • Analogy: Imagine a house with a floor plan so twisted and knotted that no matter how hard you try, you can't straighten it out.
• The Tool: To prove this, they used a technique called Birational Rigidity.
  • The Metaphor: Think of the building as a piece of clay. For simple shapes, you can squish and reshape the clay into a cube. But for these high-degree buildings, the clay has become "rigid." It's like a sculpture made of hardened steel; you can't reshape it into a cube without breaking it. The authors proved that these specific steel sculptures are "superrigid"—they cannot be transformed into the simple warehouse.

3. The "Real World" Twist

Why does this matter? Most math is done over "complex numbers," which are like a perfect, infinite universe where everything is smooth. But we live in the real world (real numbers).

• In the real world, things can be "connected" (you can walk everywhere) but still be mathematically "twisted" in a way that prevents them from being simple.
• The authors showed that even if a building looks nice and connected in the real world, it might still be mathematically "broken" (irrational) because of how it behaves when you look at it through a different lens (algebraic closure).

Summary of the Takeaways

1. Not all connected shapes are simple: Just because a 3D shape looks connected and passes standard tests doesn't mean it can be simplified into a basic grid.
2. New Tools are needed: The old tools (intermediate Jacobians) were blind to these specific shapes. The authors had to invent/use a new tool (unramified cohomology) to see the "hidden locks."
3. Complexity matters:
   • Simple curves ($d \le 4$) $\rightarrow$ Rational (Easy to simplify).
   • Very complex curves ($d \ge 12$) $\rightarrow$ Irrational (Too rigid to simplify).
   • Medium complexity ($6 \le d \le 10$)$\rightarrow$ The Mystery: The authors suspect these are also irrational, but they haven't fully cracked that specific code yet.

In a nutshell: This paper is like a detective story where the authors prove that some 3D shapes, which look innocent and connected, are actually mathematically "impossible" to simplify. They did this by finding new ways to look at the shapes and proving that once a shape gets too complex, it becomes permanently "rigid" and can never be turned into a simple, empty space.

Technical Summary

Here is a detailed technical summary of the paper "On the Rationality of Some Real Threefolds" by Olivier Benoist and Alena Pirutka.

1. Problem Statement and Context

The paper investigates the $k$-rationality of specific geometrically rational threefolds defined over a real closed field $k$ (such as $\mathbb{R}$). A variety $X$ is $k$-rational if it is birational to the affine space $\mathbb{A}^n_k$.

The authors focus on varieties that lie at the "edge" of current rationality techniques. Specifically, they study conic bundles and quadric surface bundles where:

1. The real locus is connected (a necessary condition for rationality).
2. The intermediate Jacobian obstructions (a primary tool for proving non-rationality over $\mathbb{C}$ and recently adapted to $\mathbb{R}$) vanish.

The study centers on two families of equations:

• Family (0.1): Quadric surface bundles over $\mathbb{P}^1$ given by $x^2 + y^2 + z^2 = u \cdot p(u)$, where $p(u)$ is a non-negative separable polynomial of even degree $d \geq 2$.
• Family (0.2): Conic bundles over $\mathbb{P}^2$ given by $x^2 + y^2 = f(v, w)$, where the zero locus of $f$ is a nodal rational curve with specific properties regarding its real points.

The Core Question: Are these varieties rational over $k$ when standard obstructions (like intermediate Jacobians) fail to detect non-rationality?

2. Methodology

The authors employ a combination of advanced techniques from birational geometry and arithmetic geometry:

• Unramified Cohomology: To prove non-rationality for Family (0.1), the authors utilize degree 3 unramified cohomology groups ($H^3_{nr}$). This generalizes the Artin–Mumford obstruction (which uses the Brauer group, $H^2$). They construct a specific non-trivial cohomology class $\alpha \in H^3_{nr}(k(X \times Y)/k, \mathbb{Z}/2)$ and demonstrate its non-vanishing.
• Specialization and Ramification: A novel aspect of their proof involves showing that a cohomology class is ramified along divisors in the special fiber of a model over a non-archimedean real closed field (specifically $R((t^{1/n}))$). This allows them to detect non-triviality where standard methods fail.
• Birational Rigidity (Sarkisov Program): To prove non-rationality for Family (0.2) in high degrees, the authors adapt Sarkisov's theorem to non-closed fields. They establish that certain conic bundles are birationally superrigid, meaning any birational map to another Mori fiber space must be an isomorphism of the generic fibers.
• Explicit Constructions: For low-degree cases of Family (0.2), the authors provide direct geometric constructions proving rationality by analyzing the nodal curves and reducing the problem to the rationality of specific quadric surface bundles.

3. Key Contributions and Results

A. Non-Rationality of Quadric Surface Bundles (Family 0.1)

• Theorem 0.1: For any even degree $d \geq 2$, there exist varieties defined by $x^2 + y^2 + z^2 = u \cdot p(u)$ over the real closed field $R = \bigcup_{n \geq 1} \mathbb{R}((t^{1/n}))$ that are not $R$-rational.
  • Method: They construct a non-trivial unramified cohomology class $\alpha$. By analyzing the ramification of this class on a model over a power series ring, they prove $\alpha \neq 0$. Since rational varieties must have trivial unramified cohomology, these varieties are not rational (nor stably rational).
• Corollary 0.2: There is no single rationality construction (or finite set of constructions) using rational functions of bounded degree that can prove the rationality of all such varieties for a fixed $d$. This implies the rationality problem for these varieties is not solvable by a uniform algebraic method.
• Note: The rationality of these varieties over the standard reals $\mathbb{R}$ remains an open question, though the result over $R((t^{1/n}))$ suggests a negative answer is likely.

B. Rationality of Conic Bundles in Low Degree (Family 0.2)

• Theorem 0.3: For conic bundles defined by $x^2 + y^2 = f(v, w)$ with degree $d \leq 4$, the varieties are always $R$-rational over any real closed field.
  • Method: The authors perform a detailed case analysis of the nodal quartic curves ($d=4$). They show that these curves can be transformed via birational maps into quadric surface bundles that possess smooth rational points, thereby proving rationality.

C. Non-Rationality of Conic Bundles in High Degree (Family 0.2)

• Theorem 0.4: For conic bundles defined by $x^2 + y^2 = f(v, w)$ with degree $d \geq 12$, the varieties are never $R$-rational.
  • Method: They apply a variant of Sarkisov's theorem (Theorem 0.5) adapted to non-closed fields. They prove that if the linear system $|4K_S + \Delta|$ is non-empty (where $S$ is the base surface and $\Delta$ the discriminant curve), the conic bundle is birationally superrigid.
  • For $d \geq 12$, the condition $|4K_S + \Delta| \neq \emptyset$ is satisfied, implying the variety cannot be rational.
• Theorem 0.5 (Technical Contribution): The authors rigorously prove that Sarkisov's theorem on birational superrigidity holds over non-closed fields of characteristic 0. This is a significant technical contribution because properties like "relative Picard rank 1" and "Q-factoriality" are not always preserved under field extension, requiring careful adaptation of the Minimal Model Program (MMP) and Sarkisov links.

4. Significance and Impact

1. Overcoming Intermediate Jacobian Obstructions: The paper addresses a gap in the literature where intermediate Jacobian obstructions vanish. It demonstrates that unramified cohomology and birational rigidity are powerful alternative tools for detecting non-rationality in real threefolds.
2. Extension of Birational Rigidity: The authors provide the first application of birational rigidity techniques to $k$-rationality problems for geometrically rational varieties in dimension $\geq 3$ over non-closed fields. They clarify the subtleties of running the Sarkisov program over non-closed fields, specifically regarding the behavior of the relative Picard rank and Q-factoriality.
3. Sharpness of Results: The results establish a clear boundary for rationality in these families:
   • Low degree ($d \leq 4$) $\implies$ Rational.
   • High degree ($d \geq 12$) $\implies$ Non-rational.
   • The authors conjecture that non-rationality holds for all $d \geq 6$, suggesting the condition $|4K_S + \Delta| \neq \emptyset$ might be improved to $|3K_S + \Delta| \neq \emptyset$ or even $|2K_S + \Delta| \neq \emptyset$ in future work.
4. Methodological Innovation: The technique of detecting non-triviality of unramified cohomology classes by analyzing ramification on special fibers of models over non-archimedean real closed fields is presented as a novel approach.

In summary, this paper resolves the rationality status of specific families of real threefolds that were previously inaccessible to standard tools, utilizing a sophisticated blend of cohomological obstructions and birational geometry adapted to the arithmetic setting of real closed fields.