Diplomová práce pana Jakuba Krčmy se zabývá koherentním buzením a současnou optickou detekcí nanoskopických spinových vln v tenkých vrstvách yttrito-železitého granátu (YIG) pomocí metody fokusovaného Brillouinova rozptylu světla zesíleného Mieho rezonancemi (Mie-BLS). Pan Krčma zvládl celou problematiku v plné šíři – od teorie přes mikromagnetické a FDTD simulace a výrobu vzorků elektronovou litografií až po vyhodnocení intenzitně a časově rozlišených BLS měření – a jako první demonstroval detekci koherentně buzených spinových vln o vlnové délce až 180 nm v technologicky významném materiálu YIG. Oceňuji jeho vědeckou poctivost, s níž uvádí i negativní výsledky, i jeho dlouhodobou vědeckou aktivitu: je prvním autorem publikace v časopise Science Advances, na niž práce přímo navazuje, a spoluautorem pěti dalších odborných článků. S potěšením konstatuji, že všechny cíle diplomové práce byly splněny, diplomovou práci doporučuji k obhajobě a hodnotím ji stupněm A.

The thesis by Bc. Jakub Krcma addresses a timely and technically demanding problem at the frontier of experimental magnonics: the simultaneous coherent excitation and nanoscale-resolved optical detection of spin waves in yttrium iron garnet (YIG) thin films. The work is of high quality overall, though the presentation balance deserves comment.

The thesis is already extensive and impressively documented, but to be easily followed in its experimental depth it would arguably need roughly twice as much space devoted to the experimental sections. As it stands, the theoretical introduction, covering micromagnetic theory, LLG dynamics, the Kalinikos-Slavin model, and a broad state-of-the-art review, is longer than the experimental part of the thesis. While this foundation is competently written, much of it is standard textbook material that could be significantly condensed without loss of clarity. Redirecting this space toward the experimental results would substantially improve the thesis. For example, Fig. 5.4 presents a beautifully reconstructed spin-wave dispersion relation across the dipolar-exchange crossover, but no representative raw BLS spectra are shown from which the data points were extracted. Including a selection of these would greatly help the reader appreciate the experimental quality and the fitting procedure.

The most scientifically interesting observation of the thesis is the discrepancy in excitation efficiency between the full Py grating coupler and the single Py resonator strip. The grating successfully excites and launches propagating spin waves in excellent agreement with theory, and a decay length of approximately 16 um, which is a compelling demonstration of the technique on YIG. However, excitation from a single Py strip yields only a marginal signal, barely distinguishable from background. This contrast deserves deeper investigation. The thesis correctly identifies resonance localization within the Py strip and wavevector mismatch with the Si detection window as contributing factors, but does not systematically explore the crossover between single- and multi-strip behavior. A set of micromagnetic simulations varying the number of strips from 1 to roughly 20, tracking how emitted spin-wave amplitude scales with N, would have been highly valuable and would directly reveal whether the enhancement is coherent (proportional to N squared) or incoherent (proportional to N), shedding light on the fundamental physics of the emission process. This is identified as the main missed opportunity in the thesis.

A further notable finding, which the thesis itself presents somewhat in passing, is the hybridization between the Py resonance and the YIG Bloch mode satisfying the Bragg condition. The field-sweep measurements in Figs. 5.17 and 5.18 clearly show an anti-crossing behavior, and the strongest propagating spin-wave emission occurs when the Py resonance is tuned slightly below the crossing. This is a non-trivial result with practical implications for the optimal design of future grating couplers and merits more prominent discussion.

The logic of the presentation is mature and the documentation is careful and complete. Some of the presented data would benefit from more careful visualization, as the spectra in Fig. 5.13 contain features that are barely visible at the chosen color scale. A small inset with a rescaled color range would help considerably. The thesis contains a minimal number of errors, e.g. a mislabeled panel in Fig. 4.2 and a small number of additional editorial oversights that do not affect the scientific content.

Overall, this is an impressive thesis. The ultimate experimental results may not be as striking as initially envisioned: the single-resonator geometry shows limited emission efficiency, and the practical applicability of the full excitation-detection scheme on YIG remains constrained. Nevertheless, the thesis checks all necessary boxes. It delivers a rigorous theoretical framework, validates Mie-BLS on YIG for the first time, demonstrates coherent spin-wave propagation at nanoscale wavelengths in a tabletop setup, and provides genuine physical insight into the challenges of excitation and detection of nanoscale spin waves. The work is recommended for defense with the grade of excellent. Topics for thesis defence:

1. In the single-strip geometry, dynamic dipolar coupling between the Py resonator and YIG creates a hybridized system with its own modified dispersion, distinct from bare YIG. When forming a periodic array, this gives rise to an effective medium with a collective dispersion clearly visible in Fig. 5.8(b), which could in principle be described using a Grunberg-type bilayer model to quantify the refractive index mismatch at the grating-to-bare-YIG boundary that may impede spin-wave out-coupling. At the same time, Fig. 5.14 shows an accumulation of dynamic magnetization amplitude along the propagation direction inside the grating. Based on your intuition and the results presented, which effect do you consider dominant for the weak single-strip signal: the dispersion mismatch between the hybridized bilayer system and bare YIG, or simply insufficient amplitude from a single emitter that the array overcomes through coherent addition?
2. Can you estimate, even order-of-magnitude, how many strips are needed before the grating signal becomes detectable above background in your setup? What does this tell you about the coherence of the emission process?
3. The thesis motivates the need to access the exchange-dominated regime in YIG for high-velocity spin-wave devices. Given the detection and excitation limitations you encountered, what is your honest assessment of when Mie-BLS will become a practical characterization tool for real magnonic devices?
4. The reciprocity model and the Green's function model give very similar results in Fig. 4.4, yet they rely on different mathematical frameworks. In what regime would you expect them to diverge, and which would you trust more for Mie resonator geometries?