CHSH inequality
In physics, the Clauser–Horne–Shimony–Holt (CHSH) inequality can be used in the proof of Bell’s theorem, which states that certain consequences of entanglement in quantum mechanics cannot be reproduced by local hidden-variable theories.
Classical states
As illustrated in Fig. 1, let us consider two persons, Alice
and Bob
, who have each two independent observables $A,\bar{A}$ and $B,\bar{B}$ with possibles outcomes $\pm 1$.

We define the variable
\[\begin{equation} \label{eq:S} S = \langle AB \rangle + \langle \bar{A}B \rangle + \langle A\bar{B} \rangle + \langle \bar{A}\bar{B} \rangle. \end{equation}\]The usual form of the CHSH inequality is
\[\begin{equation} \label{eq:CHSH_ineq} |S| \leq 2. \end{equation}\]Expliciting all possible combinations of Alice’ and Bob’s observables in Table 1, it is easy to verify \eqref{eq:CHSH_ineq} in the classical case. Indeed,
\[\begin{align} \begin{split} S &= \big| \braket{AB} + \braket{A\bar{B}} + \braket{\bar{A}B} - \braket{\bar{A}\bar{B}} \big| \\ &= \big| \braket{A(B+\bar{B}) + \bar{A}(B-\bar{B})} \big| \\ &= 2 \end{split} \end{align}\]if the only outcomes are $\pm 1$ and cannot be in between.
\(A\) | \(\bar{A}\) | \(B\) | \(\bar{B}\) | \(A(B+\bar{B})\) | \(\bar{A}(B-\bar{B})\) |
---|---|---|---|---|---|
-1 | -1 | -1 | -1 | 2 | 0 |
-1 | -1 | -1 | 1 | 0 | 2 |
-1 | -1 | 1 | -1 | 0 | -2 |
-1 | -1 | 1 | 1 | -2 | 0 |
-1 | 1 | -1 | -1 | 2 | 0 |
-1 | 1 | -1 | 1 | 0 | -2 |
-1 | 1 | 1 | -1 | 0 | 2 |
-1 | 1 | 1 | 1 | -2 | 0 |
1 | -1 | -1 | -1 | -2 | 0 |
1 | -1 | -1 | 1 | 0 | 2 |
1 | -1 | 1 | -1 | 0 | -2 |
1 | -1 | 1 | 1 | 2 | 0 |
1 | 1 | -1 | -1 | -2 | 0 |
1 | 1 | -1 | 1 | 0 | -2 |
1 | 1 | 1 | -1 | 0 | 2 |
1 | 1 | 1 | 1 | 2 | 0 |
Quantum states
In the more general quantum case, illustrated in Fig. 2, we write
\(\begin{equation} \ket{\psi} = a \ket{00} + b \ket{01} + c \ket{10} + d \ket{11} = \begin{bmatrix} a \\ b \\ c \\ d \end{bmatrix} \end{equation}\) any general 2-qbit quantum state, with $|a|^2+|b|^2+|c|^2+|d|^2 =1$. Now, in contrast to the classical case, $A,\bar{A},B,\bar{B}$ are all bases for measuring the same state $\ket{\psi}$ shared by Alice and Bob. They are given by
\[\begin{align} \begin{split} A &= Z = \begin{bmatrix} 1 & 0 \\ 0 & -1 \end{bmatrix} \\ \bar{A} &= X = \begin{bmatrix} 0 & 1 \\ 1 & 0 \end{bmatrix} \\ B &= \sqrtt (Z+X) = \sqrtt \begin{bmatrix} 1 & 1 \\ 1 & -1 \end{bmatrix} \\ \bar{B} &= \sqrtt (Z-X) = \sqrtt \begin{bmatrix} 1 & -1 \\ -1 & -1 \end{bmatrix} \end{split} \end{align}\]
We remind the definition in \eqref{eq:S} of the variable
\[\begin{equation*} S = \langle AB \rangle + \langle \bar{A}B \rangle + \langle A\bar{B} \rangle + \langle \bar{A}\bar{B} \rangle. \end{equation*}\]In quantum algebra, the statistical expectation of the output of the observables $AB$ can be computed as
\[\begin{align} \label{eq:AB} \begin{split} \braket{AB} &= \braket{\psi | A \otimes B | \psi} \\ &= \begin{bmatrix} a & b & c & d \end{bmatrix} \sqrtt \begin{bmatrix} 1 & 1 & & \\ 1 & -1 & & \\ & & -1 & -1 \\ & & -1 & 1 \end{bmatrix} \begin{bmatrix} a \\ b \\ c \\ d \end{bmatrix} \\ &= \sqrtt \begin{bmatrix} a & b & c & d \end{bmatrix} \begin{bmatrix} a+b \\ a-b \\ -c-d \\ -c + d \end{bmatrix} \\ &= \sqrtt (a^2-b^2-c^2+d^2). \end{split} \end{align}\]And, we similarly obtain
\[\begin{align} \braket{\bar{A}B} &= \sqrtt (2ad+2bc) \label{eq:A_B} \\ \braket{A\bar{B}} &= \sqrtt (a^2-b^2-c^2+d^2) \label{eq:AB_} \\ \braket{\bar{A}\bar{B}} &= \sqrtt (-2ad-2bc). \label{eq:A_B_} \end{align}\]Injecting \eqref{eq:AB}-\eqref{eq:A_B_} into \eqref{eq:S} yields
\[\begin{align} \label{eq:S_quantum} \begin{split} S &= \frac{2}{\sqrt 2} (a^2-b^2-c^2+d^2+2ad+2bc) \\ &= \frac{2}{\sqrt 2} [(a+d)^2-(b-c)^2]. \end{split} \end{align}\]From \eqref{eq:S_quantum}, we can distinguish different situations. Firstly, the simplest case $a=1$, which is a classical state
, gives $S=\frac{2}{\sqrt 2}$ and satisfies the CHSH inequality \eqref{eq:CHSH_ineq}. However, choosing $a=d=\sqrtt$ gives the entangled state
and $S=2\sqrt{2} > 2$ and breaks the CHSH inequality
!
Enjoy Reading This Article?
Here are some more articles you might like to read next: