Students and faculty of UC Santa Barbara’s physics department gathered in Corwin Pavilion for the department’s annual Nobel Prize-focused colloquium on Oct. 14. This year’s event was particularly special as the department celebrated two of its own faculty members — professors Michel Devoret and John Martinis — the latest laureates of the Nobel Prize in Physics, along with their colleague at the time, professor John Clarke at UC Berkeley, for their pioneering mid-1980s experiment that observed quantum tunneling and quantized energy levels on a macroscopic scale.
Quantum mechanics describes the universe at its most fundamental level by revealing the probabilistic behavior of particles. The theory successfully explains phenomena that classical physics could not, such as observations of blackbody radiation, the photoelectric effect and the discrete emission of light by atoms, which produces spectral lines used to identify elements in stars. In the quantum mechanical framework, particles exist as a superposition of multiple states, such as different positions, spin orientations or energies, simultaneously. Only once measured does the particle collapse into one definitive state that we can observe. This means that in two identical experiments, the same measurement can lead to different outcomes — a fundamental difference between quantum mechanics and the definite predictability of classical mechanics.
As a fundamental theory, quantum mechanics should, in principle, encompass all physical phenomena. However, a conceptual issue with this is depicted in the well-known Schrödinger’s cat paradox. This thought experiment places a cat inside a box along with a radioactive atom and vial of poison. When the atom decays, the vial is broken and the cat is exposed to the poison. Before a measurement is made, the atom can be described as a superposition of decayed or undecayed, and by extension, the cat is both dead and alive.
“[The] Schrödinger’s cat paradox is a paradox because quantum mechanics is supposed to be universal. It is a universal law of any variable, whether the position of an electron or collective motion of avogadro number of electrons,” Devoret said, explaining the motivation behind their experiment at the colloquium. “It should work on all scales, but we know from measurements that these superpositions do not occur on a macroscopic scale.”
The experiment conducted by Clarke, Devoret and Martinis sought to test this boundary, as conveyed in the simple yet striking question they posed in the first line of their paper published in 1988: “Are macroscopic degrees of freedom governed by quantum mechanics?”
There are various hypotheses that attempt to explain the Schrödinger cat paradox. For instance, the many-worlds interpretation suggests that rather than the wave function collapsing, the other state does occur and is observed in a parallel world. In other words, the cat is dead in one world but alive in another world. Though proposed by Schrödinger himself, this interpretation remains untestable and unsatisfying to many realists.
An alternative possibility, and the one explored by Clarke, Devoret and Martinis, is that quantum mechanics is not an accurate way to describe the world on larger scales. “If we would find a way in which quantum mechanics would fail, that would solve this measurement problem and solve the unease of this paradox,” Devoret explained.
The scientists used a Josephson junction to test if quantum mechanical phenomena could be observed. The junction setup involves two superconducting leads separated by an insulating barrier. The group sent current from one electrode to the other and measured the potential difference. This current was made up of billions of electrons, ensuring that the experiment measured macroscopic degrees of freedom. “[Current] dipole moment is many orders of magnitude larger than an electron around a hydrogen atom, in this way it was a macroscopic experiment,” Devoret said.
Their experiment was unprecedented in its ability to reduce noise to isolate genuine quantum effects. The superconducting chip containing the junction made up only a small fraction of their entire setup. The rest included filters to isolate the chip from external noise, as well as coaxial connectors used to probe and control the dynamics of the experiment at the microwave range. This innovative technique allowed for precise characterization of parameters such as resonant frequency of the junction and damping resistance, so that each parameter needed to apply the theory could be determined independently. These experimental achievements testify to the group’s innovation and dedication which made the success of the experiment possible.
Martinis revealed the time and effort that went into making every part of the setup: “We had to make our own coaxials, we had to make our own filters; various things were custom designed, and a lot of microwave engineering went into this.” The group confirmed that the setup matched classical expectations before cooling the experiment down to be sensitive to quantum tunneling effects.
It “looks like a superconducting short when current is less than the critical current of the junction, which is set by the properties of the material,” Martinis explained while describing the experiment’s behavior at cold temperatures. “A little before it reaches this critical current, due to thermal fluctuations or quantum tunneling, you can see a resistance across it.”
One can consider the generalized flux associated with the junction as a particle confined by the potential function of the Josephson junction, known as the tilted washboard potential. If this particle were classical-like, it would be trapped in a local minimum of the potential and no voltage would be measured across the capacitor. By contrast, if the particle were quantum-like, it can tunnel above the potential, allowing the observation of a non-zero voltage.
“This phenomenon is similar to alpha decay, where the alpha particle tunnels out of the nucleus,” Devoret said. The same occurs in this experiment, “except what tunnels is the state of the circuit.” Having observed quantum tunneling, the researchers quantified their observations by measuring the time duration between the initial preparation of the experiment and the moment it escaped and obtained the tunneling rate. Knowing the measured parameters of the experiment meant that they could compare their experimental result with theory, and they found the two to be consistent. “We are seeing macroscopic quantum tunneling,” Martinis confirmed.
Another key way in which they confirmed quantum behaviour in the experiment was by observing discrete energy levels of the system.
“We put in a single frequency and then vary the frequencies of the qubit,” Martinis explained. By tuning the current into the junction to allow the system to reach resonance, they observed three transitions between the first, second, third and fourth excited states, revealed by distinct spikes in the voltage readings. “I could see three dots,” Martinis said as he described how the groundbreaking result first looked on the oscilloscope. “Before even analyzing the data, I knew we had a really interesting sample.”
These observations agreed with theory for a single particle, and served as a confirmation that the macroscopic degrees of freedom in the group’s experiment behave like a single quantum mechanical particle.
The work of Clarke, Devoret and Martinis has had a lasting impact on the field of quantum computation. The core idea of a quantum computer is to store information as a superposition of states. By showing that macroscopic variables in a superconducting circuit can behave like a single quantum particle, the scientists have confirmed that quantum mechanics manifests in systems “big enough to be held in the hand.” The group’s experimental results have also paved the way for the use of superconducting quantum circuits in novel studies of quantum optics as well as investigations into how quantum mechanics applies to other macroscopic systems.
Martinis later extended the work during his time at Google, where he worked on the Sycamore processor, a superconducting quantum processor that contained 53 qubits. His group demonstrated the processor’s exponentially faster performance, completing a task in 200 seconds that would have taken the best classical supercomputer 10,000 years to complete. “We did a computation on a Hilbert space with 53 qubits in a macroscopic space and showed that quantum mechanics still worked,” he explained.These results have sparked an increased drive for physicists to explore these quantum systems further, with superconductors now recognized as a viable platform for a quantum computer. “The biggest reward I see is creating all this opportunity for physicists,” Martinis remarked, reflecting on the broader impact of the work. The pioneering work of Clarke, Devoret and Martinis has not only deepened our understanding of quantum mechanics but has also laid the foundation for future quantum technologies.
A version of this article appeared on p.11 of the October 23, 2025 edition of the Daily Nexus.
