In a string of accurate quantum experiments, physicists have violated Heisenberg’s Uncertainty Principle – one of the icons of quantum weirdness, with its counter-intuitive statement that all measurements are essentially fuzzy and cannot be improved beyond a certain limit.

When Heisenberg first proposed the Uncertainty Principle nearly a century ago, it seemed to be merely a curiosity, as technology at that time was nowhere near sensitive enough to challenge quantum limits. 

However modern technology has enabled physicists to explore those limits, and now eclipse them.

An international team led by physicists from the Department of Quantum Science (DQS) in the ANU Research School of Physics developed theory and then ran experiments confirming their calculation of actual uncertainty bounds which improve on those proposed by Heisenberg and the 2003 successor, Ozawa’s 'Universally Valid Uncertainty Relation'.

Their experiments used a number of different quantum computing systems around the globe, each with multiple correlated qubits – it is this correlation that enables better accuracy than was previously possible, said Lorcan Conlon, a PhD student in the DQS. 

“Ozawa’s uncertainty relation assumes that measurements are made on individual systems, but by taking a pair of systems, entangling them together and measuring them, we showed you can do better,” Mr Conlon said.

“We actually gain access to a much broader class of measurements, and this lets us show that supposedly universally valid uncertainty principles are not in fact universally valid.” 

“But the question of a truly universally valid uncertainty principle remains open.”

Heisenberg’s Uncertainty Principle arises from the fact that certain pairs of quantum properties are connected to each other. If you measure one accurately, then the other is disturbed in a way that makes an accurate measurement impossible. These pairs are known as conjugate properties – for example the position and momentum of an object are a conjugate pair.

Separate quantum objects can also be linked, a phenomenon known as entanglement. When entangled, some properties of the objects become shared and dependent on one another – in contrast to the either/or nature of conjugate properties.

The DQS team leveraged the properties of entanglement to get higher accuracy of conjugate pair measurements, an approach first devised in 2005 by Hiroshi Nagaoka, who found the accuracy improves the more systems are entangled.

First, the team explored Nagaoka’s theory to find a specific example of quantum properties for which collective measurements improved the accuracy enough to be observed experimentally. 

Next, they needed to devise an experiment to map the abstract quantum parameters to a physically measurable quantum system, which is not straightforward. Initially they tried to use a photonics set up that they could implement in their own lab, but could not do it successfully.

With a number of quantum computing systems around the world now available for public use via the internet, the team decided to try to test Nagaoka’s theory on the various machines, which included the Rigetti superconducting silicon architecture in the United States, a trapped ion computer at Innsbruck, Austria, IBM Quantum System One, (a superconducting device based in Germany) and an entangled photonics set-up in Germany.

Numerical optimisation enabled the DQS team to find the optimum experimental parameters: the rotation of the individual qubits, the coupling between the qubits to generate the necessary entanglement, done using CNOT gates, and the measurement angle. 

The team put the scheme into practice on all the quantum computers available, finding significantly varying performance among the different devices. The best system was the IBM Quantum System One, based in Germany.

On that system a single qubit performed well, so the team tried for a two-qubit system, but found that noise became significant. Modelling the noise showed the source was a systematic rotation discrepancy, which could be corrected. With this, the team were able to take good measurements and matched the Nagaoka calculation for a two-qubit system.

“You can imagine my excitement when I realised we had actually violated uncertainty relations which are assumed to be universally valid,” Mr Conlon said.

The team then set their sights on linking three qubits together to achieve an even better accuracy. However, in practice, the noise in the three-qubit system was too high for uncertainty limits to be explored.

Nonetheless Professor Ping Koy Lam, from DQS and A*STAR chief quantum scientist at Institute of Materials Research and Engineering (IMRE) in Singapore, said observing quantum-enhancement with two qubits in a noisy scenario was a significant success.

"For practical applications, such as in biomedical measurements, it is important that we can see an advantage even when the signal is inevitably embedded in a noisy real-world environment," he said.

The research has been published in Nature Physics.