Thermostats without conjugate points

This paper generalizes Hopf's theorem to thermostats by establishing that non-positive total curvature characterizes systems without conjugate points, proving the equivalence between transverse Green bundles and projective Anosov flows, and providing a counterexample to Hopf's rigidity on the 2-torus via the first known instance of a projectively Anosov thermostat that is not Anosov.

The Gist

Imagine you are walking across a vast, hilly landscape. In the world of standard physics (Riemannian geometry), if you walk in a straight line, you follow the path of least resistance, known as a geodesic. If the hills are shaped just right (like a saddle or a bowl), two people starting from the same point but walking in slightly different directions will eventually drift apart and never meet again. This is a "good" landscape for navigation.

However, if the landscape is shaped like a sphere (like a ball), those two walkers might start apart, drift, and then suddenly cross paths again at the other side of the world. In math, these crossing points are called conjugate points. They are like traffic jams where your path becomes unpredictable.

Now, imagine adding a twist to this story. What if, while walking, a mysterious wind starts blowing on you? This wind doesn't just push you forward; it pushes you sideways, and the strength of the push depends on how fast you are walking. This is a Thermostat.

In the real world, thermostats are used to model how particles move in non-equilibrium systems (like a gas that isn't perfectly still). In this paper, the authors are studying these "windy" landscapes to see how the wind changes the rules of the road.

Here is the breakdown of their discovery, translated into everyday language:

1. The "No-Intersection" Rule (No Conjugate Points)

The authors are obsessed with landscapes where walkers never cross paths again. They want to know: What kind of wind (force) makes sure that if you and a friend start walking slightly differently, you will never meet again?

In the old days (standard geodesics), the rule was simple: "If the ground is curved like a saddle everywhere, you never meet."
The authors found a new, more flexible rule for these windy thermostats. They discovered a "curvature formula" (a complex math equation) that acts like a weather report. If this report says the "thermostat curvature" is negative enough, you are guaranteed to never meet your friend again.

2. The "Green Bundles" (The Invisible Compass)

To understand these paths, the mathematicians invented a tool called Green Bundles. Think of these as invisible, magical compasses attached to every walker.

• One compass points toward the "stable" future (where you are likely to end up).
• The other points toward the "unstable" past (where you came from).

In a perfect, chaotic system (called Anosov), these two compasses always point in completely different directions. They are "transverse." This means the system is very predictable in its chaos; you know exactly how paths diverge.

The authors proved a beautiful connection:

• If the compasses are always pointing in different directions: The system is chaotic but stable (Projectively Anosov).
• If the compasses collapse and point in the exact same direction: The system is too rigid or flat.

3. The Big Surprise: The "Flat Torus" Trick

Here is the most exciting part. In standard geometry, there is a famous rule (Hopf's Theorem) that says: "If you live on a donut-shaped world (a 2-torus) and you never cross paths, the world must be perfectly flat."

The authors asked: Does this rule still hold if we add our mysterious "wind" (the thermostat)?

The Answer: NO.

They constructed a specific example of a windy thermostat on a donut-shaped world.

• The world is curved (not flat).
• The wind is blowing in a very specific, velocity-dependent way.
• Result: Walkers never cross paths (no conjugate points), yet the world is not flat.

This is like saying, "I can build a roller coaster that never loops back on itself, even though the track is twisted and curved, simply by adding the right amount of wind."

4. Why This Matters

This paper is a bridge between pure math and real-world physics.

• For Mathematicians: It shows that the old rules of "flatness" don't apply when you introduce complex forces like thermostats. It generalizes a 70-year-old theorem to a much wider, more complex world.
• For Physicists: Thermostats model things like friction, heat, and energy dissipation. Understanding how these systems behave without "traffic jams" (conjugate points) helps us predict how energy moves in complex systems, from superconductors to climate models.

The Takeaway Metaphor

Imagine a dance floor.

• Standard Geodesics: Dancers move in straight lines. If the floor is curved, they might bump into each other.
• Thermostats: Dancers are pushed by a DJ's wind machine.
• The Discovery: The authors found a specific way to tune the wind machine so that even on a curved, bumpy floor, the dancers can spin and weave without ever colliding. Furthermore, they found that you can have this perfect, collision-free dance on a "donut" floor without the floor being flat, which was previously thought to be impossible.

They also found that if the dancers' "compasses" (Green bundles) stop pointing in different directions and start pointing the same way, the dance loses its chaotic energy and becomes boringly predictable.

In short: They rewrote the rules of navigation for windy, curved worlds, proving that chaos can exist on a donut without the floor being flat.

Technical Summary

1. Problem Statement and Context

The paper investigates the dynamical properties of thermostats on closed oriented Riemannian surfaces $(M, g)$. A thermostat is a dynamical system defined by the equation:
$$\nabla_{\dot{\gamma}} \dot{\gamma} = \lambda(\gamma, \dot{\gamma}) J \dot{\gamma}$$
where $\nabla$ is the Levi-Civita connection, $J$ is the complex structure (rotation by $\pi/2$), and $\lambda \in C^\infty(SM, \mathbb{R})$ is a smooth function on the unit tangent bundle. Unlike geodesic flows ($\lambda=0$) or magnetic flows ($\lambda$ depends only on position), thermostats allow $\lambda$ to depend on velocity, creating systems that can be dissipative while preserving kinetic energy.

The central problem is to generalize classical results regarding conjugate points and curvature from geodesic flows to the broader class of thermostats. Specifically, the authors aim to:

1. Establish a generalized Hopf rigidity theorem for thermostats.
2. Characterize thermostats without conjugate points using curvature conditions.
3. Analyze the relationship between the absence of conjugate points, the Green bundles (stable/unstable subbundles), and the projectively Anosov property.
4. Determine if classical rigidity results (e.g., on the 2-torus) extend to thermostats.

2. Methodology

The authors employ a combination of differential geometry, dynamical systems theory, and symplectic geometry. Key methodological innovations include:

• Cotangent Bundle Approach: Instead of working on the tangent bundle $T(SM)$, the authors lift the dynamics to the cotangent bundle $T^*(SM)$. They utilize the symplectic lift of the flow $\tilde{\phi}_t$ and focus on the characteristic set $\Sigma \subset T^*(SM)$, defined as the kernel of the Liouville form restricted to the flow direction. This allows them to define a smooth invariant subbundle that replaces the contact distribution found in geodesic flows.
• Adapted Frames and Gauges: They introduce an adapted coframe $(\alpha, \beta, \psi_\lambda)$ on $\Sigma$. A crucial technique is the introduction of a gauge function $p: SM \to \mathbb{R}$ to define a family of curvature quantities $\kappa_p$. This gauge freedom allows them to characterize the absence of conjugate points more flexibly than in the geodesic case.
• Damped Jacobi Equations: By analyzing the evolution of covectors in $\Sigma$, they derive a system of differential equations. They introduce a "damped" variable $z(t)$ related to the vertical component of the Jacobi field, satisfying a modified Jacobi equation $\ddot{z} + \tilde{\kappa}z = 0$, where $\tilde{\kappa}$ is a damped thermostat curvature.
• Riccati Analysis: The authors translate the convergence of Green bundles into the existence of global solutions to Riccati equations. They use comparison theorems for Riccati equations to relate the transversality of Green bundles to the negativity of curvature.
• Lyapunov Exponents and Ergodic Theory: They utilize Oseledets' theorem and Lyapunov exponents to analyze the asymptotic behavior of the flow, linking the collapse of Green bundles to the vanishing of topological entropy.

3. Key Contributions and Definitions

A. Thermostat Curvature and Gauges

The paper defines a generalized curvature $\kappa_p$ dependent on a gauge function $p$:
$$\kappa_p := \pi^* K_g - H\lambda + \lambda^2 + Fp + p(p - V\lambda)$$
where $K_g$ is the Gaussian curvature, $H = [V, X]$, and $V, X$ are vertical and geodesic vector fields.

• The standard thermostat curvature $K$ corresponds to the choice $p = V\lambda$.
• The damped curvature $\tilde{\kappa}$ corresponds to $p = V\lambda/2$.

B. Green Bundles in the Cotangent Bundle

The authors construct the Green bundles $G^*_s$ and $G^*_u$ as limits of the cohorizontal subbundle $H^*$ under the inverse transpose of the flow. These are subbundles of the characteristic set $\Sigma$.

• Transversality: $G^*_s(v) \cap G^*_u(v) = \{0\}$.
• Collapse: $G^*_s(v) = G^*_u(v)$.

C. Projectively Anosov Flows

Since thermostats are generally not volume-preserving, the standard Anosov definition (which requires volume preservation for certain equivalences) is insufficient. The authors work with projectively Anosov flows, defined by a splitting of the quotient tangent bundle $T(SM)/\mathbb{R}F = E_s \oplus E_u$ with exponential contraction/expansion rates, without requiring the subbundles to be transverse in the full tangent bundle (only in the quotient).

4. Main Results

Theorem 1.1 & 1.2: Characterization of No Conjugate Points

• Criterion: If there exists a gauge $p$ such that $\kappa_p \leq 0$, the thermostat has no conjugate points.
• Characterization: A thermostat has no conjugate points if and only if there exists a Borel measurable function $p$ (smooth along the flow) such that $\kappa_p = 0$. This is a new characterization even for geodesic flows.

Theorem 1.3: Generalized Hopf Rigidity

For a thermostat without conjugate points and any gauge $p$:
$$\int_{SM} \kappa_p \, d\mu \leq \int_{SM} (p - V\lambda)^2 \, d\mu$$
If equality holds, the thermostat curvature $K$ is identically zero.

• Corollary: On the 2-torus $T^2$, if the flow is volume-preserving (magnetic case), the metric must be flat. However, for general thermostats, this rigidity breaks down.

Theorem 1.4: Existence on the 2-Torus

Contrary to the geodesic/magnetic case, for any Riemannian metric $g$ on $T^2$, there exists a function $\lambda$ such that the thermostat $(T^2, g, \lambda)$ has no conjugate points. Furthermore, $\lambda$ can be chosen such that its zeroth Fourier mode (the magnetic component) is non-zero. This provides a counterexample to the extension of Hopf's rigidity theorem to general thermostats.

Theorem 1.6: Projectively Anosov vs. Transverse Green Bundles

A thermostat without conjugate points is projectively Anosov if and only if its Green bundles are transverse everywhere ($G^*_s \cap G^*_u = \{0\}$).

• This generalizes Eberlein's theorem for geodesic flows.
• Significance: The authors construct examples (Theorem 1.4) of thermostats that are projectively Anosov but not Anosov, answering a question by Mettler and Paternain.

Theorem 1.8: Curvature and Transversality

If a thermostat without conjugate points has transverse Green bundles, there exists a continuous gauge $p$ such that $\kappa_p < 0$ everywhere. This provides a partial converse to the fact that Anosov flows do not necessarily have negative Gaussian curvature.

Theorem 1.9: Collapse of Green Bundles

• If $K=0$, the flow has no conjugate points, and the Green bundles collapse to a line almost everywhere.
• Optimality: The authors provide an example (Proposition 4.1) where $K=0$ but the Green bundles do not collapse everywhere (they are transverse at some points). This shows that the Freire-Mañé conjecture (which links flat metrics to collapsing Green bundles) does not directly extend to thermostats using the standard thermostat curvature $K$.
• However, the damped curvature $\tilde{\kappa}$ is better suited: $\tilde{\kappa} = 0$ almost everywhere if and only if the Green bundles collapse almost everywhere.

5. Significance and Implications

1. Breaking Rigidity on the Torus: The paper demonstrates that the rigidity of the 2-torus (where no conjugate points implies flatness) is a feature of volume-preserving systems (geodesic/magnetic). By introducing velocity-dependent forces ($\lambda$), one can construct non-flat thermostats on $T^2$ without conjugate points.
2. New Class of Anosov-like Systems: The discovery of projectively Anosov thermostats that are not Anosov expands the classification of hyperbolic flows. It highlights that in dissipative systems, the distinction between "Anosov" and "Projectively Anosov" is non-trivial.
3. Refined Curvature Concepts: The introduction of the gauge-dependent curvature $\kappa_p$ and the damped curvature $\tilde{\kappa}$ provides a more nuanced toolkit for analyzing non-reversible, dissipative flows. It shows that the "correct" curvature to look at depends on the specific dynamical property being investigated (e.g., existence of conjugate points vs. collapse of Green bundles).
4. Counterexamples to Conjectures: The paper refines the understanding of the Freire-Mañé conjecture, showing it fails for general thermostats but holds for magnetic systems under specific conditions.

6. Conclusion

The paper successfully generalizes the theory of conjugate points and Green bundles to the setting of thermostats. By leveraging the cotangent bundle formalism and gauge freedom, the authors establish that while many structural results from geodesic flows persist (e.g., the link between transversality and projective Anosov behavior), the rigidity results (like Hopf's theorem) do not. The work provides a comprehensive framework for understanding the interplay between curvature, conjugate points, and hyperbolicity in non-equilibrium statistical mechanical models.