Berlin-Led Review Shows Tailored Pulses Flip Nuclear Spins 10 Times Faster Than Standard Ones
A Science Advances review marking a century of spin physics says optimal control now reaches from MRI scanners to chiral molecules, and may need machine learning next.
Electromagnetic pulses designed with optimal control theory can invert nuclear spins about 10 times faster than the conventional pulses used until now, according to a review in Science Advances that surveys what physicists have learned about steering quantum systems since spin entered the field a century ago.
Spin is a property of particles with no direct counterpart in everyday experience. Since the early 20th century, researchers have learned to manipulate it with precisely shaped electromagnetic pulses. That work underpins two technologies in wide use today: magnetic resonance imaging, which is routine in hospitals, and nuclear magnetic resonance spectroscopy, which is used to determine the structures of proteins and to develop new drugs.
The review, led by Steffen J. Glaser and colleagues and titled "Optimal control of spin systems," appears in a special issue of the journal devoted to the next 100 years of spin. It was written with researchers from the Technical University of Munich, Aarhus University, Universite de Bourgogne and Dortmund University. Christiane Koch, a professor of theoretical physics at Freie Universitaet Berlin, framed the core question. "Given a quantum system and the fields I can apply to it, how do I reach the state I want as quickly, as reliably and as accurately as possible," she said, "and how well can it be done at all, in principle?"
The speed gain is the headline number. In magnetic resonance imaging, the same methods also let designers build the constraints of a real scanner directly into the design of an experiment, rather than treating hardware limits as an afterthought.
The authors say the approach now reaches well beyond magnetic resonance. Molecules can be controlled with tailored electromagnetic pulses, which matters for chiral molecules that come in left- and right-handed forms and can differ sharply in biological effect. Carefully shaped sequences of microwave pulses can move one of the two forms into a chosen quantum state while leaving its mirror image alone, a prerequisite for separating them.
Related techniques have been used to guide individual atoms through a large range of possible states. Optimal control has improved the preparation of certain atomic states in experiments, and it has produced a nonclassical superposition in which an atom occupies several very different configurations at once. According to the review, no protocol for doing that had existed before. "Improving a protocol that already works is useful," Koch said. "The more interesting case is a state that no one had been able to prepare before."
The review is also candid about limits. Optimal control needs an accurate model of the quantum system and a well-calibrated experiment, so a pulse that works on paper can fail if the model is off. Looking ahead, the authors see the most promising path in combining optimal control with machine learning. Control methods bring precise physical models and fast convergence, while machine learning can infer compact descriptions from experimental data.
There is a catch, which is data. "Quantum measurements are scarce and expensive," Koch said. "Quantum physics is not big data. That is why the two approaches have to be combined, rather than one replacing the other."
The paper is a review, not a new experiment, so the tenfold figure reflects the field's accumulated results rather than a single new measurement. Its DOI is 10.1126/sciadv.aeh1978.
