Cosmological horizons in regular bouncing backgrounds

This paper challenges the conventional wisdom that decelerated expansion precludes event horizons and accelerated expansion precludes particle horizons by demonstrating, through regular bouncing cosmological models, that the existence of these non-local horizons depends on the spacetime's complete past and future history rather than just its instantaneous expansion phase.

The Gist

Imagine the universe as a giant, expanding balloon. In standard cosmology, we usually think of this balloon starting from a tiny, infinitely dense point (the Big Bang) and inflating forever. In this story, there are two main "rules" about what we can see and what we can never see:

1. The Particle Horizon: Think of this as the "visible edge" of the universe. It's the furthest distance light could have traveled to reach you since the beginning of time. If the universe started with a Big Bang, there is a limit to how far back you can look; you can't see the "other side" of the balloon because light hasn't had enough time to get there yet.
2. The Event Horizon: Think of this as the "future edge." It's the boundary beyond which, if a light signal is sent right now, it will never reach you, no matter how long you wait. This happens if the universe is expanding so fast that the space between you and the signal stretches faster than the signal can travel.

The Old Rule of Thumb
For a long time, scientists thought there was a simple rule:

• If the universe is expanding slowly (decelerating), you have a "visible edge" (Particle Horizon) but no "future edge" (Event Horizon). Eventually, you can see everything.
• If the universe is expanding super fast (accelerating), you have a "future edge" (Event Horizon) but no "visible edge" (Particle Horizon). You can see everything that ever existed, but you can't send a message to the far future.

The New Story: The "Bouncing" Universe
This paper, written by M. Gasperini, asks a "what if" question. What if the universe didn't start with a Big Bang explosion? What if, instead, it was shrinking for a very long time, hit a minimum size (a "bounce"), and then started expanding again? This is called a regular bouncing background.

The author shows that in these bouncing scenarios, the old rules break down. The "horizons" behave in surprising ways because the universe has a long history before the bounce, not just a beginning.

Here are the three main scenarios the paper explores, using simple analogies:

1. The "Infinite Library" Scenario (No Horizons at All)

Imagine a library that has existed forever. You are walking down the aisle looking for books.

• The Setup: The universe was shrinking forever, then bounced, and is now expanding forever.
• The Result: Because the universe has existed for an infinite amount of time in the past, light from everywhere has had infinite time to reach you. There is no "visible edge" (Particle Horizon).
• The Future: Because the universe will expand forever but slow down, light you send out now will eventually reach any point in the universe, no matter how far. There is no "future edge" (Event Horizon).
• The Takeaway: In this specific type of bounce, you can see everything that ever happened, and you can eventually reach everything in the future. Both horizons disappear.

2. The "One-Way Street" Scenario (Only a Future Edge)

Now imagine a different kind of bounce. The universe was shrinking, bounced, and is now expanding, but it's expanding so fast that it's getting "stretched" like taffy.

• The Result: Just like in the standard "fast expansion" model, there is no limit to how far back you can see (no Particle Horizon) because of the infinite past.
• The Catch: However, because the expansion is accelerating, there is a "future edge." If you send a message to a distant galaxy right now, the space between you will stretch so fast that the message will never arrive.
• The Takeaway: You can see the entire history of the universe, but you cannot communicate with the entire future.

3. The "Double-Locked Room" Scenario (Both Horizons Exist)

This is the most surprising part. Imagine a universe that was shrinking slowly, bounced, and is now expanding quickly.

• The Result: In this case, both horizons exist!
  • Particle Horizon: Even though the universe is old, the way it shrank and bounced means there are some regions so far away that light from the very beginning hasn't reached you yet. You have a "visible edge."
  • Event Horizon: Because the universe is now expanding fast, there are regions you can never reach in the future. You have a "future edge."
• The Takeaway: You have a limit on what you can see now, and a limit on what you can reach later. This contradicts the old idea that you can't have both at the same time.

The "Local" vs. "Global" Difference

The paper makes one final, crucial distinction.

• Global Horizons (Particle/Event): These depend on the entire history of the universe (the whole story from start to finish). As shown above, in bouncing universes, these can be weird, missing, or both present.
• Local Horizon (The "Perturbation" Horizon): This is a different kind of boundary. It's like a local speed limit for ripples in the fabric of space (quantum fluctuations). The paper notes that this local boundary behaves very predictably and closely follows the standard "Hubble radius" (the size of the universe at any given moment). It doesn't care about the long history of the bounce; it just reacts to the immediate conditions.

Summary

The paper argues that we cannot assume the rules of the standard Big Bang apply to "bouncing" universes. Depending on how the universe shrank before the bounce and how it expands after, the "edges" of our observable universe can disappear, appear, or exist in pairs. The universe's history is a long story, and the "horizons" are determined by the whole plot, not just the current chapter.

Technical Summary

Technical Summary: Cosmological Horizons in Regular Bouncing Backgrounds

Problem Statement
In standard cosmological scenarios, the existence of cosmological horizons is often linked to the kinematic phase of the universe's expansion. A decelerated expansion phase is typically characterized by the absence of global event horizons, while an accelerated expansion phase is associated with the absence of particle horizons. However, this paper argues that such associations are not absolute. Since particle and event horizons are non-local properties determined by the full past and future history of the spacetime manifold, their existence and behavior in "bouncing" cosmologies—where the universe transitions from a contracting phase to an expanding phase without a singularity—may differ significantly from standard expectations. The central problem addressed is whether regular bouncing backgrounds, connecting an initial regime of growing curvature to a final regime of standard decelerated or accelerated expansion, necessarily exhibit the absence of particle or event horizons, or if they can support scenarios where both, neither, or only one exists.

Methodology
The author employs analytical and numerical methods to investigate the behavior of cosmological horizons within various regular bouncing models. The study focuses on spatially flat Friedmann-Lemaitre-Robertson-Walker (FLRW) metrics with scale factor $a(t)$. The radii of the particle horizon ($R_P$) and the event horizon ($R_E$) are defined by the integrals of the inverse scale factor over the past ($t_{min}$ to $t$) and future ($t$ to $t_{max}$) cosmic time, respectively.

The analysis proceeds by examining specific exact solutions and numerical models across different theoretical frameworks:

1. String Cosmology (T-duality invariant): Exact solutions derived from string cosmology equations, including non-local dilaton potentials and higher-order $\alpha'$ corrections.
2. Modified Gravity (Massive Gravity): Numerical integration of cosmological equations from a modified theory of massive gravity that allows for Null Energy Condition (NEC) violation to facilitate a bounce.
3. Generalized Gravitational Theories: Exact solutions from quadratic gravity ($f(R)$) and scalar-tensor "quintom" models.

For each model, the author evaluates the convergence or divergence of the horizon integrals (Eq. 1) over the infinite time domain ($t \in (-\infty, +\infty)$) and compares the behavior of $R_P$ and $R_E$ against the local Hubble radius $R_H = |H|^{-1}$. The study also contrasts these non-local horizons with the local perturbation horizon $R_\lambda$, which governs the amplification of quantum metric fluctuations.

Key Contributions and Results

The paper demonstrates that the behavior of cosmological horizons in bouncing scenarios is highly dependent on the specific kinematic properties of the pre-bounce and post-bounce regimes, yielding four distinct scenarios:

1. Absence of Both Horizons:
   In T-duality invariant string cosmology models (e.g., solutions with non-local dilaton potentials or $O(d,d)$-symmetric models with higher curvature corrections), the universe evolves from an initial accelerated expansion (growing curvature) to a final decelerated expansion. In these cases, the integrals for both $R_P$ and $R_E$ diverge due to the infinite extension of time in both directions. Consequently, these backgrounds possess neither a particle horizon nor an event horizon.

2. Absence of Particle Horizon, Presence of Event Horizon:
   The paper presents a self-dual string cosmology solution incorporating one-loop corrections. This model describes a bounce connecting an initial phase of accelerated expansion (growing curvature) to a final phase of accelerated expansion (decreasing curvature). Here, the particle horizon integral diverges (no particle horizon), but the event horizon integral converges, resulting in a finite, time-dependent event horizon. This mirrors the behavior of standard inflationary cosmology but within a non-singular, bouncing framework.

3. Presence of Both Horizons (Decelerated Contraction to Accelerated Expansion):
   The author constructs scenarios where the universe transitions from a phase of decelerated contraction to accelerated expansion.

   • Massive Gravity Model: A numerical solution based on modified massive gravity equations, utilizing a time-dependent violation of the NEC, yields a smooth bounce. Both $R_P$ and $R_E$ integrals converge, resulting in finite, well-defined particle and event horizons.
   • Quadratic and Quintom Models: Exact solutions in quadratic gravity and quintom scalar-tensor models also exhibit this behavior. In these symmetric or near-symmetric bounces, the particle horizon is larger than the event horizon for $t > 0$ ($R_P > R_E$) and smaller for $t < 0$ ($R_P < R_E$), reflecting the causal asymmetry relative to the bounce time.

4. Distinction from Perturbation Horizons:
   A crucial distinction is drawn between the non-local causal horizons ($R_P, R_E$) and the local horizon ($R_\lambda$) relevant for the amplification of quantum fluctuations. The paper shows that while $R_P$ and $R_E$ can exhibit "unconventional" behaviors (e.g., both existing simultaneously), the local horizon $R_\lambda$ asymptotically tracks the Hubble radius $R_H$. Therefore, the specific global causal structure (presence/absence of $R_P$ or $R_E$) does not directly dictate the amplification of metric perturbations, which is governed by local dynamics around the bounce.

Significance and Claims
The paper claims to provide a concise illustration of the diversity of causal structures possible in regular bouncing cosmologies. Its primary significance lies in challenging the assumption that bouncing scenarios automatically resolve the horizon problem via the total absence of particle horizons.

• Counter-Example to Standard Resolutions: The existence of bouncing models with finite particle horizons (scenarios 3 and 4) serves as a counter-example to proposed solutions of the horizon problem in bouncing cosmologies that rely on the non-existence of particle horizons.
• Non-Locality of Horizons: The work reinforces that horizons are non-local properties dependent on the full spacetime history, meaning that the local kinematic state (accelerated vs. decelerated) is insufficient to determine their existence without considering the global topology and time extension.
• Model Dependence: The results highlight that the presence or absence of horizons is not a universal feature of all bouncing models but is strictly tied to the specific symmetries and asymptotic behaviors (e.g., self-duality, time-reflection symmetry, or specific matter sources) of the chosen gravitational model.

The paper concludes that while the global causal structure varies widely, the local physics governing primordial perturbations remains robustly linked to the Hubble scale, suggesting that the "horizon problem" in bouncing cosmologies must be addressed with a nuanced understanding of these distinct horizon definitions.