Noise cancellation by superposition of channels and superactivation of quantum capacity: Experimental realization by NMR

This paper experimentally demonstrates, using NMR registers, that superposing noisy quantum channels can cancel dephasing effects to recover coherence and activate quantum capacity in zero-capacity depolarizing channels.

The Gist

Imagine you have a very delicate message written on a piece of paper. Now, imagine you have to send this message through a hallway filled with two different types of "noise machines."

• Machine A is a fan that blows dust on the paper, smudging the ink (this is like a dephasing channel, which scrambles the timing or phase of quantum information).
• Machine B is a shredder that tears the paper into confetti (this is like a depolarizing channel, which completely randomizes the information).

Usually, if you send your message through either machine, the message is ruined. If you send it through both, it's definitely destroyed. But what if you could send your message through both machines at the exact same time, in a "superposition"?

This is the core idea of the paper by Deepika Bhargava and colleagues. They didn't just send the message through the machines; they put the path the message takes into a quantum superposition. They used a clever trick to make the noise from Machine A and the noise from Machine B cancel each other out, like two waves in a pond meeting and flattening the water surface.

Here is a breakdown of their findings using everyday analogies:

1. The Setup: The "Quantum Switch" vs. The "Superposition of Paths"

In the past, scientists used a "Quantum Switch" to put the order of events in superposition (e.g., "Machine A then Machine B" AND "Machine B then Machine A" happening at once).

This team did something different. They created a superposition of the machines themselves. Imagine a quantum coin flip that decides whether your message goes through Machine A or Machine B. But in the quantum world, the message goes through both simultaneously. By carefully tuning this "coin flip," they could make the destructive effects of the two machines interfere with each other.

2. Experiment One: Cleaning the Smudge (Dephasing Channels)

The Problem: Imagine your message gets smudged by dust (dephasing).
The Experiment: They used a 3-qubit NMR system (think of it as a tiny, ultra-precise radio receiver using atoms in a liquid) to simulate two different levels of dustiness.
The Result:

• When they tuned the superposition just right, the "smudges" from the first machine perfectly cancelled out the "smudges" from the second.
• The Analogy: It's like noise-canceling headphones. The headphones create a sound wave that is the exact opposite of the outside noise, silencing it. Here, the two noisy channels created "anti-noise" that silenced the damage to the quantum message.
• The Catch: To get this perfect silence, they had to throw away most of the messages. Only a tiny fraction (about 6%) made it through the "clean" path. But the ones that did were perfectly preserved.

3. Experiment Two: The Magic Shredder (Depolarizing Channels)

The Problem: Imagine a machine that shreds your message into random confetti. In quantum terms, this is a channel with zero capacity—it destroys all information. You can't send anything useful through it.
The Experiment: They used a more complex 5-qubit system to simulate two different "shredders." Both were so bad that, individually, they had zero ability to transmit information.
The Result:

• When they superposed these two "zero-capacity" channels, something magical happened: The resulting channel suddenly had positive capacity.
• The Analogy: Imagine two broken radios that can't pick up a single station. If you combine their signals in a specific quantum way, suddenly they start broadcasting a clear, perfect signal.
• The "Superactivation": The paper calls this "superactivation." They showed that two channels that are individually useless can, when superposed, become a perfect highway for information.
• The Catch: Again, the success rate was very low (less than 1% of the time). They had to discard almost all the data to find the few instances where the noise cancelled out perfectly.

4. How They Did It (The NMR Lab)

They didn't use sci-fi lasers or space stations. They used Nuclear Magnetic Resonance (NMR), which is the same technology used in MRI machines, but applied to tiny molecules in a liquid.

• They used molecules like a specific type of fluoromethane and a benzene derivative.
• They treated the atomic nuclei (the protons and fluorine atoms) as tiny magnets (qubits).
• By applying precise radio waves (pulses), they could control these atoms to act as the "noise machines" and the "control switches."
• They essentially programmed the atoms to simulate the math of the superposition and then measured the result.

Summary of Claims

The paper claims to have:

1. Proven the math: They figured out the exact conditions under which superposing two noisy channels results in a valid, working channel.
2. Demonstrated "Noise Cancellation": They showed that two dephasing channels can destructively interfere to restore a quantum state.
3. Demonstrated "Superactivation": They showed that two channels with zero capacity (entanglement-breaking) can be superposed to create a channel with positive capacity.
4. Experimental Proof: They verified all of this in a real lab using NMR, not just on a computer simulation.

Important Note: The authors emphasize that while the quality of the information is restored, the quantity (the number of successful messages) is very low. It's a trade-off: you get perfect data, but very little of it. The paper does not claim this is ready for commercial use yet, but rather that it proves a fundamental principle of quantum physics: that noise can be cancelled out by superposition, even turning "useless" channels into useful ones.

Technical Summary

Technical Summary: Noise Cancellation by Superposition of Channels and Superactivation of Quantum Capacity: Experimental Realization by NMR

Problem Statement
Noisy quantum channels degrade essential quantum resources such as coherence and entanglement, posing significant challenges for the realization of quantum technologies. While coherent control of noisy channels can minimize their effects, specific phenomena like the "superposition of channels" offer a novel mechanism for noise mitigation. This paper addresses the theoretical and experimental realization of canceling two noisy quantum channels by superposing their corresponding Stinespring dilation unitaries. Specifically, the authors investigate whether superposing two zero-capacity Entanglement Breaking (EB) channels can result in a channel with positive quantum capacity (superactivation), and whether two dephasing channels can destructively interfere to recover quantum coherence.

Methodology
The authors develop a framework for engineering the interference of two quantum channels using a common ancillary register.

1. Theoretical Framework:

   • The quantum channel $\Phi$ is modeled via Stinespring dilation as a joint unitary evolution $U$ of the system and an ancillary environment, followed by tracing out the environment.
   • To superpose two channels $\Phi$ and $\tilde{\Phi}$ (with Kraus operators $\{K_i\}$ and $\{\tilde{K}_i\}$), a control qubit $C$ is initialized in a superposition state $|\alpha\rangle = \cos\alpha|0\rangle + \sin\alpha|1\rangle$.
   • A controlled unitary operation is applied on the joint space of the ancilla ($A$), system ($S$), and control ($C$): $U_{total} = U_{AS} \otimes |0\rangle\langle 0|_C + \tilde{U}_{AS} \otimes |1\rangle\langle 1|_C$.
   • The control qubit is measured in the Pauli-X basis, and the $|+\rangle$ outcome is postselected. The ancilla is then traced out to obtain the final system state.
   • Validity Conditions: The authors derive necessary and sufficient conditions for the resulting superposed map to be a valid, linear, trace-preserving quantum channel. They prove that if $\tilde{U}^\dagger U$ is proportional to the identity in the subspace of the initial ancilla state, the superposed channel is linear. For general Pauli channels, this condition is satisfied, resulting in a valid Pauli channel with effective error probabilities dependent on the superposition parameter $\alpha$.

2. Experimental Realization (NMR):

   • Dephasing Channels: Implemented using a three-qubit NMR register (1H, 19F, 13C nuclei in 13C-dibromofluoromethane). The system qubit was initialized in the $|+\rangle$ state, and two dephasing channels with strengths $p=0.1$ and $\tilde{p}=0.45$ were superposed.
   • Depolarizing Channels: Implemented using a five-qubit NMR register (three 19F and two 1H nuclei in 1-bromo-2,4,5-trifluorobenzene partially oriented in MBBA). The system qubit (F1) was subjected to two depolarizing channels with strengths $p=0.55$ and $\tilde{p}=0.7$, both of which are Entanglement Breaking (EB) channels (zero quantum capacity).
   • Measurement: The normalization factor $\gamma$ (probability of postselection) and the expectation values of Pauli operators were measured to determine the effective channel strength and verify coherence recovery or capacity activation.

Key Results

1. Cancellation of Dephasing Channels:

   • The superposition of two dephasing channels resulted in a valid dephasing channel with an effective strength dependent on $\alpha$.
   • Destructive Interference: At specific values of $\alpha$ (e.g., $\alpha \approx 2.70$), the two noisy channels destructively interfered, effectively recovering the original un-dephased state $|+\rangle$.
   • Constructive Interference: At other values (e.g., $\alpha \approx 1.68$), the channels constructively interfered, outputting a maximally mixed state.
   • Trade-off: The recovery of coherence came at the cost of a reduced postselection probability ($\gamma \approx 0.06$), indicating a trade-off between ensemble size and transmission quality.

2. Superactivation of Quantum Capacity:

   • The authors superposed two zero-capacity EB depolarizing channels ($p=0.55, \tilde{p}=0.7$).
   • Non-EB Result: For a range of $\alpha$ (specifically $\alpha \in [2.35, 2.90]$), the resulting superposed channel was found to be non-EB, possessing a positive quantum capacity.
   • Perfect Activation: At $\alpha \approx 2.416$, the effective channel strength vanished ($p=0$), corresponding to an ideal, noiseless channel. This demonstrates the "perfect activation" of quantum capacity where two individually useless channels, when superposed, enable perfect quantum communication.
   • Experimental Verification: While experimental data showed deviations from theory (likely due to RF inhomogeneity and noise), specific data points fell within the theoretical bounds for non-EB and perfect activation regimes, confirming the phenomenon within experimental error bars.

Significance and Claims
The paper claims to provide an alternative structure for superposing quantum channels by utilizing a single control qubit and a common ancillary register, distinct from the "quantum switch" which superposes the temporal order of operations.

• Resource Efficiency: The authors note that previous models involving the quantum switch for superactivating EB channels required significant resources (multiple ancillary qubits and complex control). This work demonstrates that superactivation and perfect activation can be achieved with fewer resources by using a common ancillary register.
• Feasibility: The work challenges the notion that superposing two independent noise channels with separate registers is infeasible for perfect activation, showing instead that a common register enables this outcome.
• Novel Control: The study demonstrates a novel approach to the coherent control of quantum channels, offering potential applications in quantum information processing and communication, specifically in mitigating noise and enhancing channel capacity.

The authors conclude that their experimental realization using NMR validates the theoretical framework for channel superposition, interference, and the superactivation of quantum capacity, while acknowledging the trade-off between the probability of successful postselection and the quality of the resulting channel.