Squashed Entanglement

How do we measure the true, "spooky" connection of quantum entanglement when it's mixed with the noise of the real world? While perfect correlations are easy to quantify, most quantum systems exist in messy, mixed states where quantum entanglement is tangled up with ordinary classical information. The challenge lies in isolating the purely quantum part—the private, secure resource that powers quantum technologies. This problem of distinguishing quantum signal from classical noise is a central question in quantum information theory.

This article introduces ​​squashed entanglement​​, a powerful theoretical tool designed to solve this very problem. We will embark on a journey to understand this profound concept across two main sections. First, in "Principles and Mechanisms," we will explore the core idea behind squashed entanglement, defining it through the intuitive lens of an ultimate eavesdropper and the mathematics of conditional quantum information. Next, in "Applications and Interdisciplinary Connections," we will see how this abstract measure provides concrete limits for quantum cryptography and offers insights into the fundamental structure of reality itself, linking entanglement to the very fabric of spacetime.

Imagine you're handed a strange, glowing crystal that connects two points in space, let's call them A and B. When you poke the crystal at A, something instantly happens at B. They are correlated. But what is the nature of this connection? Is it a simple, classical link, like a hidden wire sending signals? Or is it the ethereal, spooky connection of quantum entanglement? How would you tell the difference? And more importantly, how would you measure how much entanglement is there? This is one of the deepest questions in quantum information, and its answer leads us to a beautiful and powerful idea: ​​squashed entanglement​​.

Let's start with the simplest case. Alice and Bob share a pair of particles in a perfectly ​​pure entangled state​​, like the famous Bell state $|\Phi^+\rangle = \frac{1}{\sqrt{2}}(|00\rangle + |11\rangle)$. In this idealized world, their systems are completely isolated from everything else. The correlation between them is perfect and entirely quantum. The amount of entanglement is easy to define: it's the ​​entropy of entanglement​​, given by the von Neumann entropy of one of the halves, say $S(\rho_A)$. For a Bell state, this gives exactly one bit of entanglement. This single number tells us that they share one "ebit" – a fundamental unit of entanglement that can be used for tasks like quantum teleportation.

But the real world is messy. Quantum states are rarely pure; they are almost always ​​mixed states​​. A mixed state is a statistical cocktail of different pure states. This means the correlations between Alice and Bob might be a mix of true quantum entanglement and mundane classical noise. Think of it this way: their particles might have interacted with the environment, or the machine that produced them might be imperfect. Now, if we just measure the total correlation between A and B, we're mixing up the good stuff (entanglement) with the junk (classical noise).

How can we isolate the pure, private entanglement? Let's use a thought experiment. Imagine a spy, Eve, who is exceptionally powerful. Her goal is to understand the correlation between Alice and Bob. The classical part of the correlation is like a conversation shouted in a crowded room—anyone can listen in and learn what's being said. The entangled part, however, is like a secret code, a private key that only Alice and Bob possess.

Eve's power comes from her ability to access an auxiliary system, let's call it $E$, that might hold information about how Alice and Bob's state was created or how it interacted with the world. Eve's strategy is to learn everything she can from her system $E$ to try and "explain away" the correlations she sees between A and B. The true, unshakeable entanglement is whatever correlation is left over—the part that remains a secret even from this ultimate spy.

In the language of quantum mechanics, Eve's knowledge is captured by a quantity called the ​​quantum conditional mutual information​​, $I(A:B|E)$. It's defined as:

This formula looks a bit dense, but its meaning is wonderfully intuitive. Think of the entropy $S(\rho)$ as a measure of our uncertainty about a system. The expression $I(A:B|E)$ quantifies how much our uncertainty about Bob's system ($B$) decreases when we learn the state of Alice's system ($A$), given that we already have complete knowledge of Eve's system ($E$).

If $I(A:B|E) > 0$, it means that even after Eve has learned everything she possibly can from her system $E$, there is still some residual correlation between Alice and Bob that she cannot explain. This leftover correlation is her blind spot; it is the private information shared between A and B.

What if $I(A:B|E) = 0$? This signifies a special situation called a ​​quantum Markov chain​​, where the systems form a chain $A-E-B$. It means that from Eve's perspective, Alice and Bob are conditionally independent. All the correlation between A and B can be completely explained by their individual correlations with E. In our analogy, it means Eve has cracked the code entirely. For instance, if we have a state where a classical bit held by Alice is correlated with a quantum state held by Bob, we can always construct an environment $E$ that is just a copy of that classical bit. For this choice of $E$, the conditional mutual information vanishes, $I(A:B|E) = 0$. This brilliantly demonstrates that purely classical correlations contribute nothing to this quantity. Similarly, if the correlation between A and C can be entirely mediated by a third party B, we can choose B as our environment and find that the private correlation is zero.

Here comes the crucial step. There isn't just one possible spy or one possible environment $E$. The laws of physics allow for countless ways the environment could be correlated with our system. To find the true, irreducible entanglement, we must be pessimistic. We must assume Eve is the most powerful spy imaginable. This means she has access to whatever extension $E$ gives her the most leverage in explaining away the $A-B$ correlation.

This corresponds to finding the extension $E$ that minimizes the conditional mutual information $I(A:B|E)$. We have to search over all possible physical extensions and find the absolute minimum value—a process called taking the ​​infimum​​. This rock-bottom value represents the correlation that is impossible to "explain away," the part that is truly private to Alice and Bob.

This leads us to the formal definition of ​​squashed entanglement​​:

The name is perfect. We are metaphorically "squashing" the total correlations, squeezing out all the classical noise and any information that could be publicly known, until only the pure, incompressible kernel of quantum entanglement remains. The factor of $\frac{1}{2}$ is a historical convention that ensures for a pure state, the squashed entanglement equals the entropy of entanglement, our gold standard from the simple case.

The power of this definition becomes clear when we apply it.

• ​​For Pure States:​​ If Alice and Bob share a pure state, there is nothing for Eve to grab onto. Any extension $E$ is necessarily uncorrelated with their system. The calculation confirms our intuition: the "squashing" does nothing, and the squashed entanglement is simply the entropy of entanglement, $E_{\mathrm{sq}}(A:B) = S(\rho_A)$. For a Bell state, this is 1 ebit. The measure passes its first and most important sanity check.

• ​​For Classical States:​​ If the state shared by Alice and Bob is purely classical (e.g., they both have a copy of the same random number), squashed entanglement gives zero. It correctly identifies that despite the perfect correlation, there is no quantum entanglement. It is not fooled by classical mimicry.

• ​​For Mixed States:​​ This is where squashed entanglement truly shines. Consider a two-qubit state that is a probabilistic mixture of two different Bell states. The system is certainly entangled, but it's also noisy. Calculating the squashed entanglement reveals exactly how the entanglement is diluted by the classical uncertainty of which Bell state the system is in. The result beautifully interpolates between the pure case (1 ebit of entanglement) and a completely random mixture.

For a quantity to be a trustworthy measure of entanglement, it must obey a certain set of physical rules. Squashed entanglement passes with flying colors, solidifying its role as one of the most important theoretical tools we have.

• ​​It cannot be created by local operations.​​ Alice and Bob, working separately on their own systems and talking on the phone, cannot increase their entanglement. Squashed entanglement respects this fundamental rule.

• ​​It is monogamous.​​ Entanglement is not a resource that can be freely shared. If Alice is maximally entangled with Bob, she cannot be entangled at all with a third person, Charlie. Squashed entanglement mathematically captures this "monogamy of entanglement." The relationships between conditional mutual informations for multipartite systems, as seen in, are a deep reflection of this principle, which itself stems from a fundamental property of quantum entropy known as ​​strong subadditivity​​.

• ​​It is continuous.​​ A good physical measure shouldn't jump dramatically if the system changes only slightly. Squashed entanglement is robust. If you perform a very "gentle" measurement on one of the particles—a measurement that barely disturbs the state—the squashed entanglement will only change by a tiny amount. This stability is crucial for it to be a meaningful quantity in a world of unavoidable small perturbations.

• ​​It provides bounds for other measures.​​ While its definition involving an infimum makes it notoriously difficult to calculate in general, it serves as a powerful benchmark. For example, it provides a fundamental upper bound on the amount of secret key that Alice and Bob can distill from their shared state. Furthermore, it can be bounded from below by more easily calculated quantities, like the mutual information, giving us a practical way to get a handle on the minimum amount of entanglement present in a system.

In the end, squashed entanglement is more than just a formula. It is a concept that gets to the very heart of what makes quantum correlations special. It provides a rigorous answer to the question we started with: by imagining the ultimate spy and finding the correlations that remain stubbornly private, we can distill the essence of entanglement from the noise of the classical world, revealing a measure that is as beautiful as it is profound.

In our journey so far, we have explored the elegant mathematical architecture of squashed entanglement. We defined it through a rather abstract minimization process, involving a mysterious third party and the subtle dance of conditional information. It is a beautiful construction, to be sure, a theorist's delight. But one might fairly ask: What is it for? Does this carefully defined quantity connect to anything tangible, anything we can build or measure? Does it help us understand the world any better?

The answer, perhaps surprisingly, is a resounding yes. Like a master key that unlocks doors in seemingly unrelated corridors, squashed entanglement reveals its power in some of the most critical and fascinating areas of modern science. It is not merely a mathematical curiosity; it is a fundamental quantity that governs the flow of quantum information, secures our deepest secrets, and even hints at the very fabric of reality. Let's step out of the abstract realm of definitions and see this remarkable concept at work.

Imagine two individuals, Alice and Bob, who need to communicate a secret message. In our classical world, they might use a key to encrypt their message, but they always face a nagging worry: how did they share the key in the first place? If they met in person, could they be sure no one was watching? If they sent it with a courier, could the courier be trusted? This is the age-old problem of key distribution.

Quantum mechanics offers a revolutionary solution: Quantum Key Distribution (QKD). The idea is to have Alice and Bob share a pair of particles in a state of pure, unadulterated entanglement. The spooky correlations inherent in this entanglement allow them to generate a secret key, with a guarantee from the laws of physics themselves that any eavesdropper—let's call her Eve—who tries to intercept the key will inevitably disturb the system and reveal her presence.

This sounds perfect, but reality is often messy. What if the source that generates these entangled pairs is imperfect? What if, with some probability $p$, it produces the perfect, maximally entangled state, but with probability $1-p$, it sputters and sends out a pair of completely uncorrelated, useless particles? This is a realistic model of a noisy quantum device. The state Alice and Bob share is no longer pure but a probabilistic mixture. The crucial question becomes: how much of a secret key can they salvage from this noisy resource?

This is precisely where squashed entanglement enters the stage. It provides the ultimate, unbreakable upper bound on the rate at which a secret key can be distilled. It quantifies the amount of entanglement that is truly "private"—the part that remains secure no matter what clever strategy Eve might employ. And for the scenario we just described, the answer provided by squashed entanglement is astonishingly simple and intuitive. The maximum secret key rate they can hope for is simply... $p$.

Think about what this means. If the source has a 90% success rate ($p=0.9$), they can generate at most $0.9$ bits of secret key for every pair of particles they use. The squashed entanglement perfectly isolates the useful part of the quantum state, discarding the fraction that was corrupted by noise. It tells us that the fundamental limit of privacy is set directly by the quality of our entanglement source.

This beautiful idea is not limited to the simple two-level systems of qubits. Suppose Alice and Bob are more ambitious and use qutrits—quantum systems with three levels. A maximally entangled pair of qutrits can carry more entanglement, a total of $\log_2(3)$ ebits. If their qutrit source is also faulty, succeeding only with probability $p$, squashed entanglement once again gives a wonderfully clear answer. The maximum secret key rate is now $p \log_2(3)$. The principle holds: the usable private correlation is the probability of success multiplied by the entanglement of the successful outcome. This is a direct consequence of a powerful property of squashed entanglement known as additivity. For mixtures of states that are physically distinct (having orthogonal supports), the total squashed entanglement is simply the weighted average of the entanglement of the parts. It behaves exactly as our intuition would demand.

Let's look closer at the definition of squashed entanglement. It involves an "infimum," a minimization over all possible ways an eavesdropper, Eve, could hold a system E that is correlated with Alice and Bob's system AB. This is the origin of the "squashing." We are finding the worst-case scenario, where Eve's knowledge is most damaging to the correlations between Alice and Bob, and we are asking what entanglement survives. The entanglement that remains is the truly private part, the part Eve can never access.

We can illustrate this with a thought experiment. Imagine a situation where the state shared by Alice and Bob, $\rho_{AB}$, originated from a larger pure state involving Eve, $|\Psi\rangle_{ABE}$. This means Eve's particle is intrinsically correlated with the pair Alice and Bob hold. She has quantum "side-information." Now, what happens if something happens to Eve's particle?

Consider a dramatic scenario: Eve's quantum memory is hit by a cosmic ray, or she passes it through a channel so noisy that it completely scrambles her qubit, replacing it with random noise. This is what physicists call a "fully depolarizing channel." In this case, her system E becomes completely uncorrelated with Alice and Bob's system AB. She has lost all her side-information.

If we calculate the conditional mutual information $I(A:B|E)$ in this scenario, it collapses to the standard quantum mutual information, $I(A:B)$. The "conditioning" on E becomes irrelevant because E contains no information. This value provides an upper bound on the squashed entanglement, but it is no longer the infimum. We have not found the most damaging scenario for Eve; we have engineered the most benign one! The resulting value represents the total correlation between Alice and Bob, both public and private. Squashed entanglement, by seeking the minimum over all of Eve's possible states, distills the private part from this total. This example beautifully demonstrates the meaning of the "squashing": it is the process of subtracting out any part of the correlation that an eavesdropper could possibly know about.

The applications of squashed entanglement are not confined to quantum communication. Its unique mathematical properties make it a powerful theoretical tool for understanding the very structure of quantum mechanics. One of the most profound features of entanglement is its "monogamy." Unlike classical correlations, which can be shared freely, quantum entanglement is exclusive. If Alice's particle is maximally entangled with Bob's, it cannot be entangled with anyone else's particle, not even a little bit.

Squashed entanglement respects this principle of monogamy in a way that many other entanglement measures do not. Its additivity for orthogonal states is a reflection of this deep property. This makes it an invaluable tool for theorists studying the complex web of correlations in multipartite quantum systems.

And here, we venture to the frontiers of modern physics. A revolutionary idea, sometimes summarized by the slogan "It from Qubit," suggests that the geometry of spacetime itself might not be fundamental. Instead, it may be an emergent property of the intricate network of entanglement between quantum systems. In the context of the ER=EPR conjecture, physicists are exploring the possibility that a wormhole connecting two distant points in spacetime (an Einstein-Rosen bridge, or ER) is the geometric manifestation of two black holes being quantum mechanically entangled (an Einstein-Podolsky-Rosen pair, or EPR).

If spacetime is indeed woven from the threads of entanglement, then understanding the fundamental quantities that govern entanglement is paramount. A measure like squashed entanglement, which so cleanly isolates the intrinsic, private correlations and respects the principle of monogamy, could be more than just a tool for communication theory. It might be one of the fundamental quantities needed to write the dictionary that translates the language of quantum information into the language of gravity and geometry.

From the practical task of securing our secrets to the grand quest of understanding the origin of spacetime, squashed entanglement proves to be a concept of surprising depth and utility. It is a testament to the beautiful unity of physics, where an idea conceived in pure information theory can provide the ultimate limit on a practical technology and, at the same time, offer a new lens through which to view the deepest mysteries of the cosmos.