When Strings Learn to Dance: Quantum Simulation Captures the Life of Confining Flux Tubes
26 Jul 2025, Yanjiang
What happens when you try to pull two quarks apart? In the everyday world, pulling something apart is straightforward — you apply force, and the object separates. But inside the quantum vacuum, the answer is stranger, almost paradoxical: the harder you pull, the more energy you must invest, until eventually that energy spawns new particles. Quarks, unlike magnets or rubber bands, can never be truly isolated. They are permanently imprisoned by a force that grows stronger with distance — a phenomenon physicists call confinement.
This is not a metaphor. It is a precise mathematical statement about the structure of the strong nuclear force, the same force that binds protons and neutrons into atomic nuclei. But confinement has always been a theoretical prediction, inferred from indirect evidence rather than directly observed in real time. The equations that govern it are too complex for classical computers to simulate dynamically, and the time scales involved are too short for any experiment to watch unfold.
Now, a team led by Jesús Cobos at the University of the Basque Country, in collaboration with IBM Quantum and researchers across Europe, has done something remarkable. Using a superconducting quantum processor with up to 144 qubits, they have directly observed the real-time dynamics of confining strings — the flux tubes that connect quark-like charges — as they oscillate, fragment, and recombine. Their work appears in a preprint (arXiv:2507.08088) that reads less like a physics paper and more like a nature documentary about the quantum vacuum.
The Prison of the Strong Force
To understand what the team achieved, we first need to grasp why confinement is so hard to study. In quantum chromodynamics (QCD) — the theory of the strong force — quarks carry a property called color charge. Unlike electric charge, which comes in two flavors (positive and negative), color charge comes in three, and the force between colored objects behaves nothing like electromagnetism.
Imagine two quarks separated by a short distance. At close range, they barely interact — a phenomenon called asymptotic freedom. But as you pull them apart, something strange happens. The gluon field between them doesn’t weaken with distance like an electric field. Instead, it forms a narrow tube, a flux tube, that behaves like a stretched spring. The energy stored in this tube grows linearly with separation. Pull hard enough, and the tube snaps — but not into nothing. The energy released creates a new quark-antiquark pair, and the tube reforms, leaving you with two separate bound states instead of two isolated quarks.
This is confinement. It is why no one has ever seen a free quark. It is also why understanding the dynamics of flux tubes is essential for understanding everything from the mass of protons to the behavior of matter in neutron stars.
But here’s the problem: the equations of QCD are non-perturbative. They can’t be solved with the usual approximation methods that work for electromagnetism. Lattice QCD, which discretizes spacetime, can compute static properties of flux tubes, but real-time dynamics — the actual motion and evolution of these strings — has remained largely inaccessible. Classical computers simply cannot handle the exponential complexity of quantum time evolution for systems with more than a handful of degrees of freedom.
A Quantum Stage for a Gauge Theory
The team circumvented this limitation by building a quantum simulator — a purpose-built quantum system that mimics the behavior of a gauge theory. Specifically, they implemented the Z₂-Higgs model, a simplified version of a gauge theory that retains the essential features of confinement while being tractable on current quantum hardware.
The key insight was mapping the problem onto IBM’s heavy-hex superconducting qubit architecture. In this mapping, matter fields live on the vertices of the lattice (where the qubits sit), while gauge fields live on the links connecting them. This is not merely an abstract analogy — it is a direct encoding: each qubit’s quantum state represents either the presence or absence of matter, or the orientation of the gauge field. The local gauge symmetry of the theory becomes a constraint on the allowed quantum states, a constraint that the team used to suppress errors and maintain consistency.
The scale of the experiment is worth pausing over. The team used up to 144 qubits — not a record in absolute terms, but the circuit depths they achieved are staggering: up to 192 two-qubit layers. For context, most quantum simulations of comparable complexity operate at depths of 10-30 layers. This is like running a marathon when everyone else is sprinting a hundred meters. The team achieved this through an optimized embedding that minimized cross-talk and a comprehensive suite of error suppression, mitigation, and correction strategies — including dynamical decoupling, readout error mitigation, and zero-noise extrapolation.
The result is an experimental platform capable of 300,000 measurement shots per circuit, totaling 600,000 shots per time step. These are not the statistics of a proof-of-principle demonstration. These are the statistics of a precision measurement.
The String That Breathes and Bends
What did they see? The most striking result is the direct observation of the dynamical hierarchy of string motion. When a flux tube connects two static charges, it can oscillate in two distinct ways: longitudinal oscillations, where the string stretches and compresses like a spring, and transverse bending modes, where the string sways from side to side like a rope in the wind.
The team resolved both modes in real time. They watched as the string’s endpoints — the positions of the dynamical charges — underwent longitudinal oscillations at a characteristic frequency, while the string itself exhibited transverse undulations. These are not merely aesthetic features. They are the precursors to hadronization — the process by which quarks and gluons combine into observable particles — and the rotational spectra of mesons, the bound states of quarks and antiquarks.
More dramatically, the team observed multi-string processes: fragmentation and recombination. When a single string was stretched beyond a critical length, it broke into two shorter strings, each connecting a new charge-anticharge pair. Then, under the right conditions, those fragments could recombine into a single string again. This is the quantum analogue of a rubber band that can snap and re-form, and it is happening on time scales of nanoseconds, inside a cryostat cooled to millikelvin temperatures.
The team validated their experimental results against extensive tensor network simulations using the basis update and Galerkin method — a sophisticated numerical technique that can predict large-scale real-time dynamics. The agreement between experiment and theory was excellent, confirming that the quantum simulator was faithfully reproducing the physics of confinement.
What This Means
This experiment is not a replacement for lattice QCD, nor does it directly simulate the full Standard Model. The Z₂-Higgs model is a toy — a simplified version of reality that captures the essential mechanism of confinement without the full complexity of three color charges and dynamical fermions.
But that is precisely its power. By stripping away unnecessary complexity, the team has isolated the core phenomenon and watched it unfold in real time. This is the scientific method at its most elegant: reduce, simulate, observe, understand.
The implications extend beyond particle physics. Gauge theories appear throughout modern physics — in condensed matter systems, in the fractional quantum Hall effect, in quantum spin liquids, and in proposals for topological quantum computing. The techniques developed here — for mapping gauge theories onto quantum hardware, for suppressing errors at scale, for extracting real-time dynamics from noisy measurements — are transferable to any of these domains.
Perhaps most importantly, this work demonstrates that current quantum processors, despite their limitations, can already probe physics that is inaccessible to classical computation. The regime of non-perturbative real-time dynamics — the realm of hadronization, thermalization, and quantum chaos — is no longer purely theoretical. It is becoming experimental.
The Unanswered Questions
As with any landmark experiment, this work raises as many questions as it answers. Can these techniques be extended to non-Abelian gauge theories, where the gauge fields interact with each other? Can they handle dynamical fermions, which introduce additional complexity? Can the system size be scaled to observe string breaking in the continuum limit, where lattice artifacts become negligible?
The team is already working on these questions. The path from the Z₂-Higgs model to full QCD is long, but the direction is clear. Each step requires better qubits, deeper circuits, more sophisticated error mitigation. But the first step — the demonstration that real-time gauge dynamics can be simulated on a quantum processor — has been taken.
What we are witnessing is the birth of a new kind of experimental physics: one where the laboratory is a quantum computer, the particles are virtual, and the phenomena are simulated rather than observed directly. It is a strange reversal of the usual order. For centuries, physicists built experiments to test theories. Now, they are building theories to test experiments — experiments that exist only as quantum circuits, waiting to be executed.
The prison of the strong force still holds. But we have built a window, and for the first time, we can watch the prisoners dance.
Yanjiang is an online editor of Loom Science
References
- Jesús Cobos et al., Real-Time Dynamics in a (2+1)-D Gauge Theory: The Stringy Nature on a Superconducting Quantum Simulator, arXiv:2507.08088