If you cannot view the document above, please access the PDF file.

15th August 2022 update: further explanatory remarks and corrections:

• The protocol describes an analog to a classical OTP protocol but applied to quantum information only.
• Quantum information is here regarded as a set of unknown states of qubits (unknown complex coefficients is computational basis representation of a qubit; \alpha\ket{0}+\beta\ket{1}, \left|\alpha\right|+\left|\beta\right|=1, \alpha,\beta\in\mathbb{C}). A known state of a qubit represents a classical information – there is no need to quantumly encrypt a classical information as it can be encrypted in a non-destructive manner classically (eg. with OTP) – in contrary to the quantum information represented by the unknown quantum state which cannot be encrypted classically in a non-destructive manner.
• Proposed protocol allows to encrypt and decrypt quantum information (not a classical one) without destroying it – allowing to reuse it in another quantum information processing system (eg. communication between two quantum computers).
• If the message to be encrypted is known to be either |+>|+>...|+> or |->|->...|->, then one can determine from the encrypted quantum register which of the 2 was encrypted message, by measuring the encrypted quantum register in the Hadamard basis. The case of OQP encrypting information in form of the example above is irrelevant to the protol. The states \ket{-}\ket{-}\ket{-}\ket{-}\ket{-}…. or \ket{+}\ket{+}\ket{+}\ket{+}\ket{+}…. are in first place cases of encrypting classical information – which is not the case of use of presented protocol. Secondly this scenario is based on fact that a CNOT gate regardless of the control qubit state (eg. arbitrary state \alpha\ket{0}+\beta\ket{1}, \left|\alpha\right|+\left|\beta\right|=1, \alpha,\beta\in\mathbb{C}) applied to states from Hadamard basis will has following results:
• CNOT\left(\left(\alpha\ket{0}+\beta\ket{1}\right)\ket{+}\right) = (\left(\alpha\ket{0}+\beta\ket{1}\right)\ket{+}
• CNOT\left(\left(\alpha\ket{0}+\beta\ket{1}\right)\ket{-}\right) = (\left(\alpha\ket{0}-\beta\ket{1}\right)\ket{-}
• As above states of target qubit \ket{+} and \ket{-} do not change upon CNOT this is easily avoided by wrapping the encrypting CNOT gate (wrapping with 4 gates – 2 for each qubit; the first two before CNOT gate; the second two after CNOT gate) with arbitrary rotation single qubit gates. Rotation angle (form the set \mathbb{R}) can be randomly chosen along with the control qubit state and kept secret as a composite key. After the wrapping an adversary applying technique indicated above (although encrypting of classical information is not considered in the protocol) will receive ambiguous results – to succeed one would have to guess the exact rotation angle of the wrapping gates form an infinite and dense set of possible values to find a proper basis for a wrapped CNOT, similar as the Hadamard basis is found for a simple CNOT.

31st July 2021 update: explanatory remarks:

• The protocol does not consider securing classical information encoded on qubits in terms of maximum entanglement level. The protocol addresses encryption of quantum information which does not require maximum entanglement level.
• The protocol is not equivalent to coin tossing and applying quantum negation, as such operation obviously does not introduce quantum entanglement (this would be the case only if the single qubit-key was a classical state, or a bit-key, and by its definition its quantum superposition qubit-key).
• The protocol description contains discussion of generalized quantum security (in short defined as inability to use the encrypted quantum information in its further controlled processing).
• The reduced density matrix of the encrypted quantum message obviously contains original information of the quantum message, but in order to find the form of the reduced density matrix one has to have in disposal the whole entangled state including the single-qubit key (without the single qubit-key there is no access to the whole entangled state and hence one cannot find the form of the reduced density matrix).

If you cannot view the document above, please access the PDF file.