Why can’t quantum states be cloned?

You may have heard of the “no clone” theorem in quantum computing, why is it a thing?
One of the basic processes in computation is copying data: take a bit b_1 and copy it to a new bit b_2. Think of this in terms of functions

    \[f: B\times B\rightarrow B, \quad f(b_1,\&b_2)=b_1\]

in which we are using C pointer notation (ampersand) for the memory location of the blank bit b_2 that will be overwritten with the copy of b_1. Now let’s look at some quantum computing functions (gates).

Classical bits are b=0, b=1, quantum bit is a superposition, which is at the core of the question

    \[|q\rangle=a|0\rangle+b|1\rangle\]

only measurable things are |a| and |b| since |q\rangle and e^{i\phi}|q\rangle represent the same quantum state.
A quantum system with n qubits contains the information of 2^n complex numbers, so modeling such a system with a classical computer will require storage of a vast amount of floating point data.

Hadamard gate

    \[|00\rangle\rightarrow |01\rangle\qquad |00\rangle\rightarrow |01\rangle\qquad |10\rangle\rightarrow |11\rangle\qquad |11\rangle\rightarrow |10\rangle\]

In other words

    \[|x,\, y\rangle \rightarrow |x, \, XOR(x,y)\rangle\]

    \[\begin{array}{ccc}\hline   x & y & XOR(x,y)\\   \hline   0 & 0 & 0\\   0 & 1 & 1\\   1 & 0 & 1\\   1 & 1 & 0\end{array}\]

If \hat{U} is a single-qubit unitary operator, a control-U gate is

    \[|0,0\rangle\rightarrow |0,0\rangle,\quad |0,1\rangle\rightarrow |0,1\rangle\]

but

    \[|1,0\rangle\rightarrow |1\rangle\otimes \hat{U}|0\rangle,\quad |1,1\rangle\rightarrow |1\rangle\otimes \hat{U}|1\rangle\]

Fool around with some gate operations:
Start with

    \[|0\rangle\otimes (a|0\rangle+b|1\rangle)=a|00\rangle+b|01\rangle\]

Apply control NOT with control on the second cubit

    \[\rightarrow a|00\rangle+b|11\rangle\]

Apply control NOT with control on the first cubit

    \[\rightarrow a|00\rangle+b|10\rangle\]

Apply control NOT with control on the second cubit

    \[\rightarrow a|00\rangle+b|10\rangle=(a|0\rangle+b|1\rangle)\otimes |0\rangle\]

Start with

    \[|1\rangle\otimes (a|0\rangle+b|1\rangle)=a|10\rangle+b|11\rangle\]

Apply control NOT with control on the second cubit

    \[\rightarrow a|10\rangle+b|01\rangle\]

Apply control NOT with control on the first cubit

    \[\rightarrow a|11\rangle+b|01\rangle\]

Apply control NOT with control on the second cubit

    \[\rightarrow a|01\rangle+b|11\rangle= (a|0\rangle+b|1\rangle)\otimes |1\rangle\]

this sequence of gates will swap states

    \[CN_2\cdot CN_1\cdot CN_2|\phi\rangle\otimes|\psi\rangle=|\psi\rangle\otimes|\phi\rangle\]

In terms of matrices

    \[CN_1\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right)=\left(\begin{array}{cccc}1&0&0&0\\ 0&1&0&0\\ 0&0&0&1\\ 0&0&1&0\end{array}\right)\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right), \qquad CN_2\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right)\]

    \[=\left(\begin{array}{cccc}1&0&0&0\\ 0&0&0&1\\ 0&0&1&0\\ 0&1&0&0\end{array}\right)\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right)\]

and the swapping operation is

    \[CN_2\cdot CN_1\cdot CN_2\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right)=\left(\begin{array}{cccc}1&0&0&0\\ 0&0&1&0\\ 0&1&0&0\\ 0&0&0&1\end{array}\right)\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right)\]

    \[=CN_1\cdot CN_2\cdot CN_1\left(\begin{array}{c}00\\ 01\\ 10\\ 11\end{array}\right)\]

Cloning
In the Hilbert space \mathcal{H}, make a copy of arbitrary, unknown basis state |e_k\rangle, perhaps “overwriting” or transforming some template or “blank” state |e_1\rangle into a copy of |e_k\rangle using a function \hat{U} ( aquantum operator)

    \[|e_k\rangle\otimes|e_1\rangle\rightarrow |e_k\rangle\otimes|e_k\rangle, \qquad \hat{U}(|e_k\rangle\otimes|e_1\rangle)= |e_k\rangle\otimes|e_k\rangle\]

Experimentor \#1 has a state |e_k\rangle that she can measure, and access (at least for now) to a template state |e_1\rangle, and experimentor \#1 wishes to provide experimentor \#2 with a copy of their state.
Try to clone a superposition with a linear operator;

    \[\hat{U}\Big(a|e_k\rangle+b|e_j\rangle\Big)\otimes |e_1\rangle=a\hat{U}\Big(|e_k\rangle\otimes |e_1\rangle\Big)+b\hat{U}\Big(|e_j\rangle\otimes |e_1\rangle\Big)\]

    \[=a\Big(|e_k\rangle\otimes |e_k\rangle\Big)+b\Big(|e_j\rangle\otimes |e_j\rangle\Big)\]

which is not what we want:

    \[\ne \Big(a|e_k\rangle+b|e_j\rangle\Big)\otimes\Big(a|e_k\rangle+b|e_j\rangle\Big)\]

there is nonlinearity inherent in the cloning process, and quantum transformations are linear unitary transformations. A universal quantum state copier would need to violate linearity. It all boils down to the fact that qubits are superpositions and quantum operations are linear operators as you saw in our gates examples.

Home 2.0
error: Content is protected !!