Physikalisches Kolloquium

Magnetic resonance imaging is an inherently non-invasive technique with biological applications from the cellular to human size-scales. A major technological push has been towards stronger magnetic fields, which can be >20 Tesla for preclinical studies and >10 Tesla for humans, since these increase the signal strength and ultimate imaging resolution. Such systems, however, require advances in hardware design, acquisition sequences and image processing algorithms to achieve optimal performance. The first part of this talk will concentrate on technical challenges and practical approaches for human scanning at 7 Tesla and above. The challenges include B_1 and B_0 inhomogeneities, increased specific absorption rate, and high sensitivity to movement. Neurological and neuroscience applications discussed include ocular and neurological tumours, epilepsy, neuromuscular diseases, glymphatic clearance and mechanistic studies of lithium for bipolar disorders. The second part will discuss the opposite end of the MRI spectrum, ultra-low field systems at ~50 mT which have been designed to address the challenges of global healthcare accessibility. The challenges here are diametrically opposite to those at high field, and topics of system design, characterization and in vivo applications will be highlighted.

Life is physics and chemistry in action. While molecular simulations of systems as complex as whole cells are now within reach, predicting chemical reactivity on relevant time and length scales remains a challenge. I will present our recent work towards bringing action – here: chemistry – to classical simulations and molecular design through machine learning. Among others, we substitute costly quantum mechanical calculations with a graph neural network-based emulator. Our framework can deal with the plethora of life’s chemistry amidst the ‘jiggling and wiggling’ of biomolecules. Importantly, we also uncover unexpected biomolecular processes that we in turn put to test in experiments. Finally, I will demonstrate how we harness a flow-matching model to predict biomolecular dynamics. Our method paves the way for generating novel flexible and functional proteins.

Symmetries play a key role in many areas of modern physics. For example, the symmetry-breaking paradigm describes how various phases of matter emerge. In magnetism, spontaneous symmetry breaking leads to well-known phases of ferromagnets and antiferromagnets. Ferromagnets have a net magnetization, while antiferromagnets have atomic magnetic moments that cancel out. Surprisingly, recent research shows this magnetic dichotomy, developed in the 1930s, is incomplete [1-5]. In this talk, we introduce our recently developed classification of magnetic phases based on spin-lattice symmetries. These are pairs of operations in spin and lattice space. This unorthodox perspective has led us to identify two unconventional magnetic phases: altermagnets [4] (see figure) and antialtermagnets [5]. Like antiferromagnets, both have compensated magnetic order and thus no net magnetization. But unlike antiferromagnets—and similar to ferromagnets—they induce spin polarization in the electronic structure. The key distinction between altermagnets and antialtermagnets lies in their behaviour under time-reversal symmetry. Altermagnets break time-reversal symmetry in their electronic structure, resulting in features like d-wave spin order [2]. In contrast, antialtermagnets preserve time-reversal symmetry and exhibit properties such as p-wave spin order [5,6]. We’ll also explore how the concept of altermagnetism was inspired by our earlier theoretical prediction[1-2] and experimental observation of an unconventional spontaneous Hall effect[7]. Additionally, we’ll highlight recent photoemission experiments that have confirmed altermagnetic order in materials like MnTe and CrSb [8]. Finally, we’ll discuss the broader implications of altermagnetism and spin symmetries. These findings have potential applications in areas such as spintronics, magnonics, topological materials, 2D materials, and multiferroics [9]—all of which could lead to faster, smaller, and more energy-efficient AI-era information technologies [1,9].

In the ΛCDM cosmological model the Universe is assumed to be isotropic and homogeneous when averaged on large scales. That the Cosmic Microwave Background has a dipole anisotropy is interpreted as due to our peculiar (non-Hubble) motion because of local inhomogeneity. There must then be a corresponding dipole in the sky distribution of sources at high redshift. Using catalogues of radio sources and quasars we find that this expectation is rejected at >5σ, i.e. the distribution of distant matter is not isotropic in the 'CMB frame’. This calls into question the standard practice of boosting to this frame to analyse cosmological data, in particular to infer acceleration of the Hubble expansion rate using Type Ia supernovae, which is then interpreted as due to a Cosmological Constant Λ. We find that the inferred acceleration is anisotropic (in the direction of the CMB hotspot) and likely illusory because of our being embedded in a coherent bulk flow, rather than due to dark energy.