Mutual Information from Modular Flow in General CFTs

This paper establishes a new high-precision analytic approximation for the vacuum mutual information in arbitrary $d$-dimensional conformal field theories by leveraging modular flow and twist operator contributions from primary fields, a method validated against known results and applied to previously unsolved cases like the $d=4$ Maxwell field.

The Gist

The Big Picture: Measuring How Much Two Things "Know" About Each Other

Imagine you have two separate rooms, Room A and Room B. In the quantum world, these rooms aren't just empty spaces; they are filled with invisible fields (like electromagnetic fields or quantum vibrations). Even though the rooms are separated, the quantum fields inside them are deeply connected. They share a secret handshake called Mutual Information (MI).

Think of MI as a measure of how much the two rooms are gossiping about each other. If Room A knows everything about Room B, they have high mutual information. If they are totally strangers, the MI is low.

Physicists want to calculate exactly how much this "gossip" exists for any shape of room and any type of quantum field. The problem is, calculating this is like trying to solve a 10,000-piece puzzle while wearing thick gloves. It's incredibly hard.

The Problem: The "Short Distance" vs. "Long Distance" Trap

For a long time, physicists had two ways to guess the answer, but both had flaws:

1. The Long-Distance Guess: When the rooms are far apart, the connection is weak. Physicists could make a very good guess here, like estimating how loud a whisper is from across a stadium.
2. The Short-Distance Guess: When the rooms are right next to each other, the connection is strong. But the math gets messy and breaks down, often predicting that the "gossip" becomes infinitely loud (a "volume law" explosion), which doesn't make sense in the real world.

The big challenge was: How do we create one single formula that works perfectly whether the rooms are far apart or right next to each other?

The New Tool: "Modular Flow" (The Time Machine)

The authors of this paper found a clever new way to look at the problem using something called Modular Flow.

Imagine the quantum field in Room A is a movie playing on a loop. Modular flow is like a special remote control that can fast-forward or rewind that movie in a very specific, geometric way. It doesn't just move time forward; it stretches and squeezes the space inside the room based on its shape (like a ball).

The authors realized that if you use this "remote control" to watch how the fields in Room A evolve, you can figure out exactly how much information they share with Room B. It's like listening to the echo of a shout in Room A to figure out the shape of the walls in Room B.

The Breakthrough: The "Two-Copy" Trick

To do the math, the authors used a trick called the Replica Method. Imagine you have a photocopy machine.

1. You take the quantum field and make $n$ copies of it (like stacking $n$ sheets of transparent paper).
2. You twist these sheets together at the edges of the rooms (this is the "twist operator").
3. You ask: "If I look at just two of these copies, what does that tell me about the whole stack?"

The authors discovered that by focusing on just two copies of the universe and using their "Modular Flow" remote control, they could derive a precise formula for the Mutual Information.

The Analogy: Imagine trying to understand the weather in a whole city. Instead of checking every single street, you realize that if you watch how the wind moves between just two specific parks (using a special time-lapse camera), you can predict the weather for the entire city.

The Result: A "Universal Approximation"

Using this method, the authors created a new formula that works for:

• Any shape: Spheres, balls, or boosted (moving) spheres.
• Any dimension: Whether the universe has 2, 3, or 4 dimensions.
• Any distance: It works when the rooms are far apart and when they are close.

The "Magic Fix":
The formula they found was perfect for long distances but went crazy (infinite) when the rooms got too close. To fix this, they applied a "patch":

1. They realized the "crazy" part was due to a specific mathematical quirk.
2. They tweaked the formula by pretending the universe had one fewer dimension (a mathematical sleight of hand) and added a small correction term.
3. The Result: This new "patched" formula matched the known exact answers for simple cases (like 2D and 3D simulations) with incredible precision.

Why This Matters

1. It's a Universal Translator: Before this, we needed different math for different types of particles. Now, we have a "universal translator" that can estimate the connection between any two regions in almost any quantum theory.
2. Predicting the Unknown: The authors used their new formula to predict the Mutual Information for a 4D Maxwell field (which describes light and electromagnetism). No one knew the exact answer for this before. Their formula gives us the best possible guess we currently have.
3. Simplicity from Complexity: They showed that you don't need to know every detail of the quantum universe to get a great answer. Just knowing the "lowest energy" particle (the simplest building block) and using this modular flow trick is enough to get a highly accurate result.

Summary in a Nutshell

The authors built a universal calculator for quantum "gossip." They used a geometric time-travel trick (Modular Flow) and a copy-paste strategy (Replica Method) to create a formula that tells us exactly how connected two regions of space are. It works for far-away rooms, close rooms, and even for predicting the behavior of light in 4 dimensions, bridging the gap between what we know and what we want to know.

Technical Summary

1. Problem Statement

The paper addresses the challenge of characterizing the vacuum mutual information (MI) between two disjoint spacetime regions (specifically spherical balls) in a general $d$-dimensional Conformal Field Theory (CFT).

• Context: MI, defined as $I(A, B) = S_A + S_B - S_{A \cup B}$, is a universal quantity that encodes the full data of a QFT. While entanglement entropy suffers from UV divergences, MI is finite and well-defined.
• Current Limitations: Previous methods for calculating MI at long distances relied on expansions involving twist operators in the replica theory ($CFT^{\otimes n}$). However, existing approximations were often limited to specific cases (e.g., scalar operators, unboosted spheres) or failed to provide high-precision analytic formulas valid for arbitrary separations (bridging the gap between short-distance and long-distance regimes).
• Goal: To derive a general, high-precision formula for the MI contribution arising from the two-copy sector of an arbitrary primary operator (of any spin) in a general CFT, and to construct a unified ansatz that accurately interpolates between short and long distances.

2. Methodology

The authors employ a combination of modular flow, replica trick, and conformal geometry.

A. Modular Flow and Twist Operators

• The MI is expressed via the limit of the expectation value of twist operators $\tilde{\Sigma}^{(n)}_A$ and $\tilde{\Sigma}^{(n)}_B$ in the $n$-fold replicated theory:
  $$I(A, B) = \lim_{n \to 1} \frac{1}{n-1} \langle \tilde{\Sigma}^{(n)}_A \tilde{\Sigma}^{(n)}_B \rangle$$
• The authors utilize a key identity connecting the correlation of twist operators with local operators in the replicated theory to modularly evolved operators in the seed CFT. Specifically, for operators $O_l$ and $O_k$ on different copies ($l \neq k$):
  $$\langle \tilde{\Sigma}^{(n)}_A O_l(x_1) O_k(x_2) \rangle \xrightarrow{n \to 1} \langle O(x_1) O_A[x_2, i\tau_{lk}] \rangle$$
  where $O_A[x, s] = \rho_A^{-is} O(x) \rho_A^{is}$ is the operator evolved by the modular flow associated with region $A$.

B. Geometric Nature of Modular Flow

• For spherical regions, the modular flow corresponds to a specific conformal transformation (Eq. 3 in the paper).
• The authors consider two disjoint spheres, $A$ and $B$, where $B$ is boosted relative to $A$ with rapidity $\beta$. This generalizes previous results restricted to concentric, unboosted spheres.
• They construct an ansatz for the twist operator as a bilocal integral over the causal completion of the entangling region, involving a kernel function derived from the inverse of the two-point function.

C. Derivation of the Main Formula

1. Kernel Determination: By inverting the correlation function relation, they determine the kernel $G_{A, \alpha\beta}$ that defines the twist operator in terms of primary operators.
2. Integral Transformation: The discrete sum over replica indices is transformed into an integral over the modular parameter $s$ using analytic continuation ($i\tau_{lk} \to s + i/2$).
3. Evaluation: The integrals are evaluated using delta functions arising from the causal structure, reducing the problem to a single integral over the modular parameter $s$.
4. Character Calculation: The result depends on the character of the Lorentz representation ($\chi_R$) of the primary operator, evaluated for a specific effective boost $\tilde{\beta}$ derived from the geometry and modular parameter.

3. Key Contributions & Results

A. General Formula for Boosted Spheres

The paper derives a closed-form integral expression (Eq. 14) for the MI contribution ($I_\Delta$) from a primary operator of scaling dimension $\Delta$ and spin representation $R$:
$$I_\Delta = \frac{z^\Delta \bar{z}^\Delta}{2^{d+4}} \int_0^1 dw \, \frac{w^{2\Delta} \sqrt{1-w}}{\dots} \chi_R(\Lambda(\tilde{\beta}))$$

• This formula is valid for arbitrary spins and arbitrary relative boosts between the spheres.
• It generalizes and supersedes all previous long-distance expansions.
• The effective boost $\tilde{\beta}$ depends on the modular parameter $s$, the physical boost $\beta$, and the conformal cross-ratios $z, \bar{z}$.

B. Long-Distance Behavior

• In the limit of large separation, the formula correctly reproduces the leading contribution from the lowest scaling dimension primary in the theory.
• The result matches the sum of conformal blocks associated with the two-copy sector (products of two seed operators), confirming consistency with the Operator Product Expansion (OPE) of twist operators.

C. Short-Distance Pathology and the New Ansatz

• Pathology: The pure two-copy sector formula exhibits a volume-law divergence ($I \sim (R/\delta)^{d-1}$) at short distances ($\delta \to 0$), whereas a local CFT should exhibit an area-law ($I \sim (R/\delta)^{d-2}$). This is a known artifact of considering only the two-copy sector (similar to Generalized Free Fields).
• The Solution (High-Precision Ansatz): The authors propose a modified ansatz to interpolate between regimes:
  $$I_{\text{Ansatz}} \equiv I_\Delta(d-1) + \sigma \cdot I_{\Delta^*}(d-1)$$
  • Dimension Shift: They shift the spacetime dimension in the formula from $d$ to $d-1$. This crucially preserves the correct long-distance behavior (which depends only on $\Delta$, not $d$) while altering the short-distance power law.
  • Additional Operator: They add a contribution from an auxiliary operator with dimension $\Delta^*$ and degrees of freedom $\sigma$.
  • Fixing Parameters:
    • $\sigma$ is fixed to ensure the coefficient of the leading short-distance term matches the universal coefficient $k_d$ of the entanglement entropy of a strip.
    • $\Delta^*$ is chosen either to cancel spurious subleading terms or to match the next-to-leading primary dimension of the theory.

4. Validation and Applications

The authors validate their approach against known exact results and numerical data:

1. $d=2$ Free Fermion: The ansatz reproduces the exact MI curve for arbitrary separations.
2. $d=3$ Free Scalar: The results match high-precision lattice calculations.
3. $d=4$ Maxwell Field: The paper provides the first analytic approximation for the MI of a $d=4$ Maxwell field (a case with no prior exact results). They cross-verify this by indirectly calculating it via the $d=4$ free scalar and $d=2$ chiral scalar difference, finding excellent agreement.

5. Significance

• Universal Characterization: The work provides a systematic way to extract CFT data (scaling dimensions, spins, and universal coefficients) from vacuum MI at arbitrary separations.
• Analytic Precision: It offers the most precise analytic approximation for MI to date, overcoming the limitations of previous expansions that were either only valid at large distances or lacked generality regarding spin and boosts.
• Insight into Degrees of Freedom: The observation that the two-copy sector alone yields a volume law (reminiscent of Generalized Free Fields) suggests that the MI formula captures a "coarse-grained" set of degrees of freedom, while the full CFT structure (including multi-copy sectors) is required to restore the area law at short distances.
• Future Directions: The framework sets the stage for incorporating multi-copy contributions to achieve a complete, systematic characterization of any CFT via vacuum mutual information.

In summary, this paper establishes a powerful, geometrically motivated framework for calculating mutual information in general CFTs, bridging the gap between long-distance conformal block expansions and short-distance area laws through a novel high-precision ansatz.