EPR Paradox Explained: Entanglement to Nobel

The EPR paradox is the 1935 argument by Einstein, Podolsky and Rosen that quantum mechanics, because of the strong correlations it predicts between separated particles, cannot be a complete description of physical reality. Decades later, Bell tests turned that philosophical claim into a measurable one. In 2015 a loophole-free experiment showed that nature violates the limit any local-realist completion of quantum mechanics would have to obey.

This article gives the EPR paradox explained step by step: what the original paper claimed, how Bohr and Schrödinger responded, how Bell’s theorem made the question testable, what the 2015 Delft experiment measured, why the 2022 Nobel Prize in Physics went to entanglement experiments, and what entanglement can and cannot do in practice.

What is quantum entanglement, and what is it not?

Two particles are entangled when their joint quantum state cannot be written as a separate state for each particle. Measure both, and the results are correlated in ways that no assignment of pre-existing, independent properties to each particle can fully reproduce. That last part is what Bell’s theorem makes precise.

Entanglement is often described as one particle “instantly affecting” the other. That wording is misleading. In standard quantum mechanics, each side’s measurement results look random on their own, no matter what the other side chooses to measure. The correlation only becomes visible when the two sets of records are brought together and compared, which requires an ordinary communication channel limited by the speed of light. This property is known as no-signalling: the correlations are real, but they cannot be used to send a message.

So entanglement is not a hidden wire between particles. It is a feature of how quantum systems share a single description, and it shows up as statistics, not as signals.

The EPR paradox explained: what the 1935 paper argued

The paper is A. Einstein, B. Podolsky and N. Rosen, “Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?”, published in Physical Review volume 47, pages 777–780, on 15 May 1935.

Its logic runs in three moves:

A criterion of reality. If you can predict the value of a physical quantity with certainty without disturbing the system, that quantity corresponds to an element of physical reality.
A completeness requirement. A complete theory must contain a counterpart for every element of reality.
A two-particle setup. For two particles that have interacted and then separated, measuring the position of one lets you predict the position of the other, and measuring the momentum of one lets you predict the momentum of the other. Since the distant particle was not disturbed, EPR argued, both quantities should be real, even though quantum mechanics does not let them have definite values at the same time.

The authors framed their conclusion as a choice: either the description of reality given by the wave function is not complete, or the two quantities cannot be real at the same time. They chose the first option. (This summary follows the abstract as reproduced on SciSpace; the APS original is the authoritative text.)

The “paradox” is therefore not a contradiction inside quantum mechanics. It is a tension between quantum mechanics and two intuitive assumptions: that physical properties exist independently of measurement, and that what happens here cannot instantly influence what is real over there.

Bohr’s and Schrödinger’s replies

Niels Bohr answered the same year, under the same title, in Physical Review 48, 696–702 (1935). Bohr disputed the EPR criterion of reality, arguing that what counts as a well-defined property depends on the whole experimental arrangement, including which measurement is chosen.

Erwin Schrödinger also responded in 1935 with “Discussion of probability relations between separated systems” in the Mathematical Proceedings of the Cambridge Philosophical Society, volume 31, pages 555–563, as listed in the same reference list compiled by Stacey (2018). His paper took the correlations between separated systems seriously as a central feature of the theory.

Bohm’s spin version

EPR’s original example used position and momentum, which is awkward to test. David Bohm reformulated the argument using spin in his 1951 textbook Quantum Theory (Prentice-Hall). Spin has only a few possible outcomes along any chosen axis, which turns EPR into a setup where two observers each pick a measurement direction and record discrete results. Nearly every modern discussion and experiment builds on this version.

How Bell’s theorem made the EPR question testable

For decades the debate looked like a matter of interpretation. Bell’s theorem changed that. It shows that any theory in which outcomes are fixed by local hidden properties, carried by each particle and unaffected by the distant measurement choice, must produce correlations that obey certain limits, called Bell inequalities. Quantum mechanics predicts that suitably chosen measurements on entangled particles violate those limits.

That turns the EPR question into an experiment. If the measured correlations stay within the limit, a local-realist completion of quantum mechanics remains possible. If they exceed it, and the experiment is free of loopholes, no such completion can exist.

The CHSH inequality in plain language

The most widely used form is the CHSH inequality. Two observers, often called Alice and Bob, each choose randomly between two measurement settings. After many rounds, they compute the correlation between their results for each of the four setting combinations and combine those four numbers into a single value, S.

Any local-realist model gives S ≤ 2.
Quantum mechanics predicts that entangled particles measured at well-chosen settings give values of S above 2.
A measured S significantly above 2, in an experiment without loopholes, rules out local realism.

From Clauser’s 1972 test to the 2015 loophole-free Bell test

John Clauser ran the first Bell test in 1972 at UC Berkeley. According to Physics World’s coverage of the 2022 Nobel Prize, he measured polarization correlations between photon pairs produced in an atomic transition and found a violation of Bell’s inequality. The experiment still left loopholes open, so it could not be considered conclusive.

The three main loopholes

Detection (fair-sampling) loophole: if detectors miss many particles, the detected subset might be unrepresentative, and a local model could exploit that to fake a violation.
Locality loophole: if one side’s setting choice could, in principle, reach the other side at light speed before its measurement finishes, a local mechanism could explain the correlations.
Freedom-of-choice loophole: if the measurement settings were somehow correlated with the hidden properties of the particles, the test’s assumptions would fail.

The first two are experimental problems that can be closed with better hardware and geometry. The third can be narrowed with independent random number sources, but it can never be eliminated entirely by experiment, because it questions the independence of the settings themselves. The extreme version of that objection is usually called superdeterminism.

The Delft experiment (2015)

A team at Delft reported “Loophole-free Bell inequality violation using electron spins separated by 1.3 kilometres”, published as Hensen et al., Nature 526, 682–686 (2015). Key details:

Element	What Delft did
Entangled system	Electron spins in two labs 1.3 km apart
Entanglement method	Event-ready scheme, estimated state fidelity 0.92 ± 0.03
Detection loophole	Closed by efficient spin readout
Locality loophole	Closed by fast random basis choice and the 1.3 km separation
Trials	245 CHSH trials
Result	S = 2.42 ± 0.20 against a local-realist limit of 2
Significance	p = 0.039 for a local-realist model producing a violation this large

The headline is that both main experimental loopholes were closed in the same experiment, as also recorded by TU Delft’s research portal. The statistical margin was modest, with a p-value of 0.039, which is why the result is best read as a decisive step in a longer series of experiments rather than a single final proof.

Why the 2022 Nobel Prize in Physics went to entanglement

On 4 October 2022 the Royal Swedish Academy of Sciences awarded the 2022 Nobel Prize in Physics jointly to Alain Aspect (Institut d’Optique / Université Paris-Saclay and École Polytechnique), John F. Clauser (J.F. Clauser & Assoc., Walnut Creek, California) and Anton Zeilinger (University of Vienna). The official press release gives the citation: “for experiments with entangled photons, establishing the violation of Bell inequalities and pioneering quantum information science.”

Each laureate received one third of the prize, according to Optica’s announcement. The citation covers two things: settling, through experiment, the question EPR raised, and turning entanglement into a working resource for quantum information science.

Was Einstein simply wrong? The more careful framing is that EPR’s thought experiment holds up, but its preferred conclusion does not. A 2026 arXiv preprint (not yet peer reviewed) describes the EPR setup as correct and as demonstrated in real experiments. What Bell tests rule out is the local-realist completion EPR hoped for.

How is quantum entanglement created and used today?

The experiments above already show two routes. Clauser’s 1972 setup used photon pairs emitted in an atomic transition, and the 2022 laureates’ work centred on entangled photons. Delft entangled electron spins in separate labs using an event-ready scheme, where a successful entangling event is heralded before the measurement round counts. Research groups also use other platforms, such as trapped atoms and ions and superconducting circuits, each with different trade-offs in distance, speed and fidelity.

The main applications follow directly from the physics:

Quantum key distribution. In entanglement-based QKD, two parties measure shared entangled pairs and check a sample of results over a public channel. An eavesdropper does not trigger an instant alarm; interference shows up as degraded correlations in that statistical check. Prepare-and-measure schemes work differently, as covered in our guide to how the BB84 QKD protocol works and where it breaks. Security agencies remain cautious about QKD in practice; see why the NSA still won’t back quantum key distribution.
Teleportation and quantum networks. Entanglement plus a classical message lets a quantum state be transferred between nodes, which is the basis of proposed quantum internets. Our explainer on quantum repeaters, teleportation and real network links covers the engineering. Note that teleportation still needs the classical message, so no-signalling is respected.
Quantum computing. Entanglement between qubits is a core resource that quantum algorithms rely on.

Entanglement-based security relies on special hardware links between two endpoints. Securing public systems such as blockchains against future quantum computers takes a different, purely mathematical route. Hadamard’s BB84 rollup on Ethereum, for example, replaces ECDSA keys with hash-based one-time addresses rather than relying on quantum hardware; its developer docs describe the design.

FAQ

Does quantum entanglement allow faster-than-light communication?

No. Each party’s results look random on their own regardless of what the other party does. The correlations only appear when both records are compared, which requires an ordinary channel limited by light speed. This no-signalling property holds even in teleportation.

What did the 2015 loophole-free Bell test prove, and what remains open?

The Delft experiment closed the detection and locality loopholes at once and measured S = 2.42 ± 0.20 against a local-realist limit of 2, with p = 0.039. It rules out local hidden-variable models under the assumption that the measurement settings are freely chosen. That freedom-of-choice assumption, challenged by superdeterminism, cannot be tested away completely.

Was Einstein wrong about the EPR paradox?

EPR’s reasoning about the two-particle setup was sound, and the correlations it described are real. What experiments contradict is the conclusion that a local, realist theory could complete quantum mechanics.

How do scientists create entangled particles in the lab?

Common approaches include photon pairs from atomic transitions or other optical sources, matter qubits such as electron spins linked by heralded event-ready schemes, and platforms like trapped ions and superconducting circuits. The choice depends on whether the goal is long distance, high fidelity or computation.

EPR Paradox Explained: From Einstein’s 1935 Doubt to the 2022 Nobel