Entanglement in Quantum Computing

What Entanglement Actually Is

Two qubits are entangled when their combined quantum state cannot be written as a product of individual qubit states. Consider two qubits in the state (|00> + |11>)/sqrt(2). This state means there is a 50% probability of measuring both qubits as 0 and a 50% probability of measuring both as 1, with zero probability of measuring 01 or 10. If you measure the first qubit and get 0, the second qubit is guaranteed to be 0. If you get 1, the second is guaranteed to be 1. The outcomes are perfectly correlated.

Classical correlations can also be perfect: if you put one red ball in one box and one blue ball in another box, then shuffle the boxes, opening one box to find the red ball tells you the other box contains the blue ball. The difference with entanglement is that the qubits do not have definite values before measurement. Each qubit individually is in a 50/50 superposition, with no predetermined outcome. The correlation exists without the values being determined in advance. This was demonstrated experimentally through Bell inequality violations, which prove that the correlations cannot be explained by any theory where the qubits carry hidden predetermined values. The 2022 Nobel Prize in Physics was awarded for these experimental confirmations.

The state (|00> + |11>)/sqrt(2) is called a Bell state or EPR pair (after Einstein, Podolsky, and Rosen, who first discussed entanglement in 1935). There are four Bell states, each maximally entangled: (|00> + |11>)/sqrt(2), (|00> - |11>)/sqrt(2), (|01> + |10>)/sqrt(2), and (|01> - |10>)/sqrt(2). They differ in whether the qubits are correlated (same measurement outcomes) or anti-correlated (opposite outcomes) and in the relative phase between the terms. Bell states are the fundamental building blocks of quantum communication, teleportation, and many quantum computing protocols.

Creating Entanglement in a Quantum Computer

Entanglement is created by two-qubit gates, with the CNOT (controlled-NOT) gate being the most common. The CNOT gate has a control qubit and a target qubit. If the control qubit is |0>, nothing happens to the target. If the control qubit is |1>, the target qubit is flipped. When the control qubit is in superposition before the CNOT is applied, the result is entanglement.

The standard recipe for creating a Bell state starts with two qubits both initialized to |0>. First, apply a Hadamard gate to the first qubit, putting it in the superposition (|0> + |1>)/sqrt(2). The two-qubit state is now (|0> + |1>)/sqrt(2) tensor |0> = (|00> + |10>)/sqrt(2). Then apply a CNOT with the first qubit as control and the second as target. The |00> component stays |00> (control is 0, target is unchanged). The |10> component becomes |11> (control is 1, target is flipped). The result is (|00> + |11>)/sqrt(2), a maximally entangled Bell state. This two-gate circuit (Hadamard plus CNOT) is the most fundamental operation in quantum computing, appearing as a subroutine in virtually every quantum algorithm.

Other two-qubit gates also create entanglement. The CZ (controlled-Z) gate applies a phase flip to the target when the control is |1>. The SWAP gate exchanges the states of two qubits. The iSWAP gate, native to some superconducting architectures, swaps states while adding a phase. The choice of entangling gate depends on what is physically natural for the qubit technology: CNOT is standard for trapped ions, CZ is common for superconducting qubits, and Rydberg-mediated gates in neutral atom systems naturally implement CZ-like operations.

Why Entanglement Is Computationally Essential

A theoretical result called the Gottesman-Knill theorem establishes that quantum circuits using only superposition (Hadamard gates), CNOT gates, and measurements in certain bases can be efficiently simulated by a classical computer. Adding non-Clifford gates (like the T gate) to the mix creates states that are both entangled and impossible to simulate classically. This means entanglement is necessary but not sufficient for quantum computational advantage. The combination of entanglement with the specific interference patterns created by universal gate sets is what produces exponential quantum speedups.

In practical quantum algorithms, entanglement serves several roles. It correlates different parts of the computation, allowing information computed in one part of the register to influence operations in another part without direct communication. In Shor's algorithm, entanglement between the input register and the output register during modular exponentiation creates a joint state whose structure encodes the periodicity of the function. The quantum Fourier transform then uses this entanglement structure to extract the period through interference.

In Grover's search algorithm, entanglement between the oracle qubit and the search register creates a joint state that allows the Grover operator to systematically increase the amplitude of the target state. Without entanglement between the oracle and the register, the amplitude amplification process cannot function, and the algorithm degenerates to random guessing.

Variational quantum algorithms use entangling gates to explore the space of quantum states that cannot be efficiently represented classically. A circuit with only single-qubit gates produces states that are tensor products of individual qubit states, which can be described using only 2N parameters for N qubits. Adding entangling gates creates states requiring up to 2^N parameters to describe. The entangling gates give the variational ansatz access to the exponentially large Hilbert space, which is necessary for capturing the quantum correlations present in molecular wave functions, optimization landscapes, and other problems of interest.

Entanglement in Quantum Communication

Quantum teleportation, despite its name, does not transmit matter or energy faster than light. It transfers the quantum state of one qubit to another qubit at a distant location, consuming one shared entangled pair and two bits of classical communication in the process. The protocol works as follows: Alice and Bob share a Bell pair. Alice performs a joint measurement on her qubit (whose state she wants to teleport) and her half of the Bell pair. This measurement yields one of four possible results, which she communicates to Bob over a classical channel. Based on Alice's result, Bob applies one of four correction operations to his half of the Bell pair, which transforms it into an exact copy of the original state. The original state is destroyed at Alice's location (consistent with the no-cloning theorem), and the transfer requires classical communication, so it cannot exceed the speed of light.

Quantum key distribution (QKD) uses entanglement to generate encryption keys that are provably secure against any computational attack. In the E91 protocol (named after Artur Ekert, 1991), Alice and Bob each receive one qubit of many entangled pairs from a shared source. They independently choose random measurement bases and record their results. After measuring all pairs, they publicly compare which bases they chose (without revealing results) and keep only the results where they happened to choose the same basis. These matching results are perfectly correlated due to entanglement and form the shared secret key. Any eavesdropper attempting to intercept and measure the qubits would disturb the entanglement, introducing detectable errors in the correlation statistics.

Quantum networks that distribute entangled pairs across long distances are an active area of development. The main challenge is that entanglement degrades as qubits interact with the environment during transmission. Quantum repeaters, which use entanglement swapping and purification to extend entanglement across multiple shorter links, are being developed to enable long-distance quantum communication. China's quantum communication infrastructure has demonstrated satellite-based entanglement distribution over 1,200 kilometers, and ground-based fiber networks carry entangled photons across metropolitan distances in several countries.

Entanglement in Error Correction

Quantum error correction uses entanglement in a fundamentally different way: to protect quantum information from noise. A logical qubit is encoded as an entangled state of many physical qubits. In the simplest error correction code, Shor's 9-qubit code, a single logical qubit is encoded across 9 physical qubits in an entangled state that is designed so that any single-qubit error (bit flip, phase flip, or both) can be detected and corrected without measuring the encoded information.

The detection works through syndrome measurements on auxiliary qubits. These measurements reveal whether an error occurred and on which qubit, without revealing the logical qubit's value. This is possible because the syndrome qubits become entangled with the data qubits in a specific way that is sensitive to errors but insensitive to the encoded information. The pattern of syndrome measurements identifies the error type and location, allowing a correction operation to restore the original state.

The surface code, the leading error correction scheme for superconducting and neutral atom qubits, arranges physical qubits on a 2D lattice with data qubits at vertices and syndrome qubits at faces and edges. The entangled state that encodes a logical qubit extends across the entire lattice, and the code distance (the size of the smallest uncorrectable error) grows with the lattice dimension. A surface code with distance d can correct up to (d-1)/2 errors, meaning larger lattices provide stronger protection at the cost of more physical qubits per logical qubit. Current experiments have demonstrated surface codes operating below the break-even threshold, where adding more physical qubits actually reduces the logical error rate.