Femtosecond Noise Correlation Spectroscopy of Exchange-Coupled Spin Systems

Abstract

Fluctuations are often seen as a nuisance, yet they are fundamental to many physical phenomena. In exchange‑coupled spin systems such as ferromagnets and antiferromagnets, these fluctuations originate from the thermal population of quantised spin waves — magnons — which, for example, play a key role in driving phase transitions. To date, spin noise spectroscopy has served as the primary method for probing magnetic fluctuations. However, electronic detection constraints fundamentally limit its bandwidth to the gigahertz range and to spin noise with nanosecond coherence times. This restricts its application largely to paramagnetic systems. In contrast, antiferromagnets exhibit significantly faster magnon dynamics reaching into the terahertz regime, making them promising candidates for next‑generation high‑speed spintronic technologies, but — combined with their significantly shorter spin coherence times — render antiferromagnetic spin fluctuations inaccessible to conventional spin noise spectroscopy. By contrast, much research has focused on the deterministic control of their spin excitations. Yet accessing these stochastic processes is crucial for understanding fundamental decoherence mechanisms and guiding the development of noise‑resilient spintronic devices. Capturing the ultrafast dynamics of antiferromagnetic spin noise therefore calls for a new measurement approach. To detect terahertz fluctuations of the vacuum electric field via its amplitude autocorrelaton, femtosecond noise correlation spectroscopy (FemNoC) has recently emerged as a powerful tool. However, its application to magnetic systems — such as antiferromagnets — has yet to be realised. This thesis presents a novel two‑colour variant of FemNoC, combining a dual‑wavelength probe scheme with subharmonic lock‑in detection. This approach bridges the gap between conventional spin noise spectroscopy and ultrafast subcycle fluctuation measurements, thereby enabling access to antiferromagnetic spin fluctuations. Building on prior work, several key experimental advancements were implemented, including the design and integration of a custom dual‑axis electromagnet, enhancements to the detection electronics, and the development of fully automated control software. In parallel, a comprehensive mathematical model was developed, validated through experiment, and supported by simulations to optimise signal acquisition and interpretation. To enable quantitative analysis, calibration protocols were established that express FemNoC signals in absolute physical units. This allows direct comparison with other physical observables and supports the design of experiments tailored to a broad range of material systems. Leveraging these innovations, spin noise dynamics in the canted antiferromagnet Sm0.7Er0.3FeO3 orthoferrite were investigated, revealing spontaneous picosecond spin switching near the spin‑reorientation transition. Additional studies on the ferrimagnetic insulator Bi:YIG demonstrated a clear scaling of FemNoC signals with magnon population, introducing a unique thermodynamic degree of freedom not accessible through conventional deterministic probing techniques such as pump–probe. Together, these results establish two‑colour FemNoC as a unique, non‑invasive probe of femtosecond incoherent magnetisation dynamics, opening new opportunities for the investigation of ultrafast magnetic phenomena through spin fluctuations.