Quantum Entanglement, Superposition, and Bell’s Theorem
Quantum Entanglement, Superposition, and Bell's Theorem
Quantum entanglement is one of the most striking and philosophically profound phenomena in all of physics. When two particles interact and form a composite quantum system, the resulting state cannot in general be written as a product of individual particle states — the particles become entangled. Measuring one particle instantaneously influences what can be said about the other, no matter how far apart they are. Einstein famously called this "spooky action at a distance" and rejected it as evidence that quantum mechanics must be incomplete, proposing that hidden variables should restore local realism. Bell's theorem, proved in 1964, showed that no local hidden variable theory can reproduce all the predictions of quantum mechanics, and Aspect's experiments in the 1980s confirmed quantum mechanics to high precision.
The mathematical arena for two-particle systems is the tensor product Hilbert space $\mathcal{H} = \mathcal{H}_A \otimes \mathcal{H}_B$. For two qubits, $\mathcal{H} = \mathbb{C}^2 \otimes \mathbb{C}^2 \cong \mathbb{C}^4$ with basis $\{|\uparrow\uparrow\rangle, |\uparrow\downarrow\rangle, |\downarrow\uparrow\rangle, |\downarrow\downarrow\rangle\}$. A general state $|\Psi\rangle = \sum_{ij} c_{ij}|i\rangle|j\rangle$ is separable (unentangled) if and only if $c_{ij} = a_i b_j$ for some vectors $a_i, b_j$. The key diagnostic is the Schmidt decomposition: write $c_{ij}$ as a matrix $C$ and perform singular value decomposition $C = U\Lambda V^\dagger$; the state is entangled if and only if more than one Schmidt coefficient $\lambda_k > 0$.
The four maximally entangled two-qubit states — the Bell states — are the EPR pairs: $|\Phi^\pm\rangle = \frac{1}{\sqrt{2}}(|\uparrow\uparrow\rangle \pm |\downarrow\downarrow\rangle)$ and $|\Psi^\pm\rangle = \frac{1}{\sqrt{2}}(|\uparrow\downarrow\rangle \pm |\downarrow\uparrow\rangle)$. These form an orthonormal basis for $\mathcal{H}$ and are eigenstates of the parity operators. In any Bell state, neither particle has a definite state on its own — the reduced density matrix of either subsystem is the maximally mixed state $\rho_A = \text{Tr}_B|\Psi\rangle\langle\Psi| = \frac{1}{2}I$.
The entanglement entropy $S = -\text{Tr}(\rho_A \ln \rho_A)$ quantifies how entangled the state is. For a pure bipartite state with Schmidt coefficients $\{\lambda_k\}$, $S = -\sum_k \lambda_k^2 \ln \lambda_k^2$ (where we treat $\lambda_k^2$ as probabilities). Separable states have $S = 0$; maximally entangled states (with equal Schmidt coefficients) have $S = \ln d$ where $d$ is the Schmidt rank. For two qubits, maximum entanglement entropy is $\ln 2 \approx 0.693$ nats or $1$ ebit.
Bell's theorem derives a mathematical inequality that must be satisfied by any local hidden variable (LHV) theory. Consider measuring spin along directions $\mathbf{a}$ and $\mathbf{b}$ on the two particles. The correlation function is $E(\mathbf{a},\mathbf{b}) = \langle A(\mathbf{a})\cdot B(\mathbf{b})\rangle$ where $A, B = \pm 1$. Bell showed that for LHV theories, the CHSH inequality holds: $|S| = |E(A,B) - E(A,B') + E(A',B) + E(A',B')| \leq 2$. Quantum mechanics allows $|S| \leq 2\sqrt{2} \approx 2.828$ (Tsirelson bound), achieved with the singlet state and optimal detector angles $45°$ apart.
Quantum teleportation is perhaps the most dramatic application of entanglement. Alice and Bob share a Bell pair. Alice has an unknown qubit $|\phi\rangle = \alpha|\uparrow\rangle + \beta|\downarrow\rangle$ to teleport. She performs a Bell-basis measurement on her qubit and her half of the shared pair, obtaining one of four outcomes. Depending on the outcome (communicated classically), Bob applies a Pauli correction to reconstruct $|\phi\rangle$ perfectly. No information travels faster than light — the classical communication is essential. Yet the quantum state, with its continuous parameters $\alpha$ and $\beta$, is transferred using only 2 classical bits.
The no-cloning theorem states that there is no unitary operation that copies an arbitrary unknown quantum state: $U(|\phi\rangle|0\rangle) \neq |\phi\rangle|\phi\rangle$ for all $|\phi\rangle$. The proof is remarkably simple: if cloning existed, the inner product $\langle\phi|\chi\rangle$ would equal $\langle\phi|\chi\rangle^2$, which is only true for $|\langle\phi|\chi\rangle| \in \{0,1\}$, meaning we could only clone orthogonal states. No-cloning is the foundation of quantum key distribution (QKD) — any eavesdropper who intercepts a qubit must measure it, disturbing the state and revealing their presence.
The BB84 protocol for QKD uses four qubit states in two conjugate bases: $\{|0\rangle, |1\rangle\}$ and $\{|+\rangle, |-\rangle\}$. Alice sends randomly chosen states; Bob measures in randomly chosen bases. When bases agree (50% of the time on average), their bits are correlated and form the raw key. Security is guaranteed by the no-cloning theorem and the disturbance caused by measurement in the wrong basis. Modern loophole-free Bell tests have confirmed quantum mechanics at the $>11\sigma$ level and provided security proofs for device-independent QKD.
For two Hilbert spaces $\mathcal{H}_A$ and $\mathcal{H}_B$, the composite system lives in $\mathcal{H}_{AB} = \mathcal{H}_A \otimes \mathcal{H}_B$. A state $|\Psi\rangle \in \mathcal{H}_{AB}$ is separable if $|\Psi\rangle = |\psi_A\rangle \otimes |\psi_B\rangle$; otherwise it is entangled. The Schmidt decomposition guarantees any pure bipartite state can be written:
$$|\Psi\rangle = \sum_{k=1}^r \lambda_k |\alpha_k\rangle \otimes |\beta_k\rangle, \quad \lambda_k > 0, \quad \sum_k \lambda_k^2 = 1$$
where $r$ (the Schmidt rank) equals the number of nonzero singular values of the coefficient matrix $C_{ij}$. Entanglement occurs iff $r \geq 2$.
The four Bell states form a maximally entangled orthonormal basis for $\mathbb{C}^2 \otimes \mathbb{C}^2$:
$$|\Phi^+\rangle = \frac{|\uparrow\uparrow\rangle + |\downarrow\downarrow\rangle}{\sqrt{2}}, \quad |\Phi^-\rangle = \frac{|\uparrow\uparrow\rangle - |\downarrow\downarrow\rangle}{\sqrt{2}}$$
$$|\Psi^+\rangle = \frac{|\uparrow\downarrow\rangle + |\downarrow\uparrow\rangle}{\sqrt{2}}, \quad |\Psi^-\rangle = \frac{|\uparrow\downarrow\rangle - |\downarrow\uparrow\rangle}{\sqrt{2}}$$
The reduced density matrix for either subsystem of any Bell state is $\rho_A = \frac{1}{2}I$, meaning neither particle has a definite spin in any direction. Entanglement entropy $S = \ln 2$ (1 ebit) — the maximum for two qubits.
Let $A, A' \in \{-1,+1\}$ be measurement outcomes for detector settings $\mathbf{a},\mathbf{a}'$ on particle 1, and $B, B'$ for settings $\mathbf{b},\mathbf{b}'$ on particle 2. Define the CHSH parameter:
$$S = E(A,B) - E(A,B') + E(A',B) + E(A',B')$$
Classical (LHV) bound: $|S| \leq 2$ (Bell inequality).
Quantum bound (Tsirelson): $|S| \leq 2\sqrt{2}$. For the singlet state $|\Psi^-\rangle$, the quantum correlation is $E(\mathbf{a},\mathbf{b}) = -\cos(\theta_{ab})$ where $\theta_{ab}$ is the angle between detectors. Choosing $\mathbf{a}=0°$, $\mathbf{b}=45°$, $\mathbf{a}'=90°$, $\mathbf{b}'=135°$ gives $S = -\cos45° + \cos135° - \cos45° - \cos45° \cdot(-1) = 2\sqrt{2}$.
There is no unitary operator $U$ on $\mathcal{H} \otimes \mathcal{H}$ satisfying $U(|\phi\rangle|0\rangle) = |\phi\rangle|\phi\rangle$ for all $|\phi\rangle \in \mathcal{H}$.
Proof: Suppose $U$ clones two states $|\phi\rangle$ and $|\chi\rangle$. Then $\langle\phi|\chi\rangle = \langle\phi|\langle0|U^\dagger U|\chi\rangle|0\rangle = \langle\phi\phi|\chi\chi\rangle = \langle\phi|\chi\rangle^2$. This requires $\langle\phi|\chi\rangle \in \{0,1\}$, meaning only orthogonal or identical states could be cloned — contradicting the assumption that all states are cloned. $\square$
Show that the state $|\Psi\rangle = \frac{1}{2}(|\uparrow\uparrow\rangle + |\uparrow\downarrow\rangle + |\downarrow\uparrow\rangle + |\downarrow\downarrow\rangle)$ is separable, and write it as a product state.
Form the coefficient matrix $C = \frac{1}{2}\begin{pmatrix}1&1\\1&1\end{pmatrix}$ where rows are particle-A basis states and columns are particle-B basis states.
$\det(C) = \frac{1}{4}(1\cdot1 - 1\cdot1) = 0$, so $\text{rank}(C) = 1$ — Schmidt rank 1, hence separable.
The singular value decomposition gives one nonzero singular value: the state factors as:
$$|\Psi\rangle = \left(\frac{|\uparrow\rangle + |\downarrow\rangle}{\sqrt{2}}\right) \otimes \left(\frac{|\uparrow\rangle + |\downarrow\rangle}{\sqrt{2}}\right) = |+\rangle \otimes |+\rangle$$
This is the product of two $|+x\rangle$ spin states — maximally uncertain in the $z$-basis but perfectly definite in the $x$-basis. No entanglement: Schmidt rank = 1.
Compute the entanglement entropy of the state $|\Psi\rangle = \frac{\sqrt{3}}{2}|\uparrow\uparrow\rangle + \frac{1}{2}|\downarrow\downarrow\rangle$. Is this more or less entangled than a Bell state?
The Schmidt coefficients are $\lambda_1 = \sqrt{3}/2$, $\lambda_2 = 1/2$ (the state is already in Schmidt form).
The reduced density matrix for particle A: $\rho_A = \lambda_1^2|\uparrow\rangle\langle\uparrow| + \lambda_2^2|\downarrow\rangle\langle\downarrow| = \frac{3}{4}|\uparrow\rangle\langle\uparrow| + \frac{1}{4}|\downarrow\rangle\langle\downarrow|$.
Entanglement entropy: $S = -\frac{3}{4}\ln\frac{3}{4} - \frac{1}{4}\ln\frac{1}{4} = \frac{3}{4}\ln\frac{4}{3} + \frac{1}{4}\ln 4 = \frac{3}{4}(0.2877) + \frac{1}{4}(1.386) = 0.2158 + 0.3466 = 0.5624$ nats.
A Bell state has $S = \ln 2 = 0.693$ nats. This state has less entanglement (0.562 nats $< 0.693$ nats) because its Schmidt coefficients are unequal. Maximum entanglement requires equal Schmidt coefficients $\lambda_1 = \lambda_2 = 1/\sqrt{2}$.
Verify that quantum mechanics violates the CHSH inequality for the singlet state $|\Psi^-\rangle$ with detector angles $A=0°$, $B=45°$, $A'=90°$, $B'=135°$.
For the singlet state, the correlation function is $E(\alpha,\beta) = -\cos(\alpha-\beta)$.
Compute each term of $S = E(A,B) - E(A,B') + E(A',B) + E(A',B')$:
$E(A,B) = E(0°,45°) = -\cos(45°) = -\frac{\sqrt{2}}{2} \approx -0.707$
$E(A,B') = E(0°,135°) = -\cos(135°) = +\frac{\sqrt{2}}{2} \approx +0.707$
$E(A',B) = E(90°,45°) = -\cos(45°) = -\frac{\sqrt{2}}{2} \approx -0.707$
$E(A',B') = E(90°,135°) = -\cos(45°) = -\frac{\sqrt{2}}{2} \approx -0.707$
$$S = -0.707 - 0.707 + (-0.707) + (-0.707) = -2\sqrt{2} \approx -2.828$$
$|S| = 2\sqrt{2} > 2$. The CHSH inequality $|S| \leq 2$ is violated by $\sqrt{2}-1 \approx 41\%$ above the classical limit. ✓
Describe the quantum teleportation protocol step-by-step for teleporting state $|\phi\rangle = \alpha|0\rangle + \beta|1\rangle$ using the shared Bell pair $|\Phi^+\rangle_{BC} = \frac{1}{\sqrt{2}}(|00\rangle + |11\rangle)_{BC}$. Show Alice's four measurement outcomes and Bob's required corrections.
The total initial state is $|\phi\rangle_A \otimes |\Phi^+\rangle_{BC} = \frac{1}{\sqrt{2}}(\alpha|0\rangle_A + \beta|1\rangle_A)(|00\rangle + |11\rangle)_{BC}$.
Expand: $= \frac{1}{\sqrt{2}}[\alpha|000\rangle + \alpha|011\rangle + \beta|100\rangle + \beta|111\rangle]_{ABC}$.
Rewrite in Bell basis for qubits $A,B$: $|\Phi^\pm\rangle = (|00\rangle \pm |11\rangle)/\sqrt{2}$, $|\Psi^\pm\rangle = (|01\rangle \pm |10\rangle)/\sqrt{2}$.
$$= \frac{1}{2}[|\Phi^+\rangle_{AB}(\alpha|0\rangle + \beta|1\rangle)_C + |\Phi^-\rangle_{AB}(\alpha|0\rangle - \beta|1\rangle)_C$$
$$\quad + |\Psi^+\rangle_{AB}(\alpha|1\rangle + \beta|0\rangle)_C + |\Psi^-\rangle_{AB}(\alpha|1\rangle - \beta|0\rangle)_C]$$
Alice measures in Bell basis (2 cbits), Bob applies: $|\Phi^+\rangle\to I$; $|\Phi^-\rangle\to Z$; $|\Psi^+\rangle\to X$; $|\Psi^-\rangle\to ZX$. In all cases Bob reconstructs $|\phi\rangle_C = \alpha|0\rangle + \beta|1\rangle$. ✓ No faster-than-light signaling: without Alice's 2-bit message, Bob's state is $\rho = \frac{1}{2}I$ (maximally mixed).
Practice Problems
Show Answer Key
1. $\hat{S}_{\text{tot}}^2|\Psi^-\rangle = (\hat{S}_1^2 + \hat{S}_2^2 + 2\hat{S}_1\cdot\hat{S}_2)|\Psi^-\rangle$. Use $\hat{S}_1\cdot\hat{S}_2|\Psi^-\rangle = -\frac{3}{4}\hbar^2|\Psi^-\rangle$ (the singlet). So $\hat{S}_{\text{tot}}^2|\Psi^-\rangle = (\frac{3}{4}+\frac{3}{4}-\frac{3}{2})\hbar^2|\Psi^-\rangle = 0$. ✓
2. The coefficient matrix is $C = \begin{pmatrix}1/2 & \sqrt{3}/2 \\ 0 & 0\end{pmatrix}$. Rank 1, so separable: $|\Psi\rangle = |\uparrow\rangle_A \otimes (\frac{1}{2}|\uparrow\rangle + \frac{\sqrt{3}}{2}|\downarrow\rangle)_B$. Schmidt rank = 1, Schmidt coefficient = 1.
3. $\rho_A = \text{Tr}_B|\Phi^+\rangle\langle\Phi^+| = \frac{1}{2}(|\uparrow\rangle\langle\uparrow| + |\downarrow\rangle\langle\downarrow|) = \frac{1}{2}I$. Then $\rho_A^2 = \frac{1}{4}I$, so $\text{Tr}(\rho_A^2) = \frac{1}{2} < 1$. This confirms the state is mixed (not pure), consistent with maximal entanglement.
4. Schmidt rank $\leq \min(d_A, d_B) = 2$. Maximum entanglement entropy is $\ln 2$ (limited by the smaller Hilbert space dimension $d_A = 2$). Always use the smaller subsystem's entropy.
5. $E = -\cos(90°) = 0$. The spins are uncorrelated at $90°$ separation — measuring one gives no information about the other. This happens because perpendicular spin components are independent.
6. $\langle\Phi^+|\Phi^-\rangle = \frac{1}{2}(\langle\uparrow\uparrow| + \langle\downarrow\downarrow|)(|\uparrow\uparrow\rangle - |\downarrow\downarrow\rangle) = \frac{1}{2}(1-1) = 0$. Similarly all other pairs vanish. All four Bell states have norm 1 and are mutually orthogonal. ✓
7. $|0\rangle = \frac{1}{\sqrt{2}}(|+\rangle + |-\rangle)$. Bob measures $|+\rangle$ with probability $\frac{1}{2}$ and $|-\rangle$ with probability $\frac{1}{2}$. The result is random and gives no information about Alice's bit — bases are mismatched and these runs are discarded in BB84.
8. If the first qubit of $|\Psi^-\rangle$ is measured as $|\uparrow\rangle$, the state collapses to $|\uparrow\rangle_1|\downarrow\rangle_2$ (from the $|\uparrow\downarrow\rangle$ term). The second qubit collapses to $|\downarrow\rangle$.
9. The LHV bound is $|S|\leq 2$. Since $2.7-0.1=2.6 > 2$, this is inconsistent with LHV at $>6\sigma$. The QM bound is $2\sqrt{2}\approx 2.828$; $2.7\pm0.1$ is consistent with QM. This constitutes a Bell inequality violation, ruling out local hidden variable theories.
10. Before Alice sends her 2-bit classical message, Bob's qubit is in the maximally mixed state $\rho_B = \frac{1}{2}I$ regardless of Alice's measurement outcome. No measurement Bob can perform allows him to distinguish the four teleportation branches. Only after receiving Alice's classical bits (limited to speed $c$) can Bob apply the correct Pauli correction and reconstruct $|\phi\rangle$.
11. Quantum discord $D(A|B) = I(A:B) - C(A:B)$ where $I$ is mutual information and $C$ is classical correlations (maximum information extractable by local measurements on $B$). Discord captures quantum correlations present even in separable mixed states — states can have nonzero discord without being entangled. Entanglement entropy applies only to pure states; discord extends to mixed states and captures a broader notion of quantumness.
12. A PR box with $|S|=4$ would allow: Alice chooses input $x$, Bob input $y$, outputs $a,b$ with $a\oplus b = x\cdot y$. If $|S|>2\sqrt{2}$ were achievable with quantum states, one could construct a protocol where Bob's marginal distribution $p(b|y)$ depends on Alice's input $x$, violating the no-signaling principle. The Tsirelson bound $2\sqrt{2}$ is the exact threshold at which quantum correlations remain consistent with no-signaling.