The study explores spin-orbit coupling (SOC)-driven phenomena in ferromagnetic metal (FM)/heavy metal (HM) interfaces, focusing on perpendicular magnetic anisotropy (PMA) and the interfacial Dzyaloshinskii-Moriya interaction (IDMI). These phenomena are crucial for developing spin-orbitronic devices like field-free magnetization switching, skyrmion racetrack memory, and SOT nanooscillators. While IDMI is known to depend on SOC and broken inversion symmetry, it's also influenced by 3d-5d(4d) orbital hybridization and interface roughness. Existing theoretical models often assume atomically smooth interfaces, neglecting the impact of real-world interface roughness and intermixing on the atomic scale. This roughness causes fluctuations in the interaction between localized spins and heavy metal atoms with high SOC, affecting the IDMI, PMA, and dynamics of chiral spin textures. To stabilize spin textures like skyrmions, IDMI values exceeding critical values for Néel domain wall formation are needed. Enhancing IDMI requires breaking structural inversion symmetry and improving interface quality; however, even with perfect epitaxial growth, local IDMI variations due to lattice strain occur. Using different heavy metals at the top and bottom interfaces can induce IDMIs of opposite signs, potentially increasing the interaction additively. Correlated roughness at both interfaces is a promising approach to cancel FM layer thickness variation and maximize IDMI. The IDMI is highly sensitive to ferromagnetic layer thickness (tFM), with interface degradation at small tFM values leading to nonlinear IDMI behavior. Previous work has investigated the influence of interface quality, considering surface roughness and atomic intermixing, in symmetric Pt/Co/Pt systems. Asymmetry is essential to induce IDMI, achievable through methods like high-temperature deposition or the introduction of an ultrathin intermediate heavy metal layer. The research hypothesizes that maximizing IDMI can be achieved through correlated roughness, which minimizes the quality parameter (difference in top and bottom interface roughness), theoretically enhancing PMA.

The introduction thoroughly reviews the existing literature on spin-orbit coupling (SOC) driven phenomena in ferromagnetic/heavy metal interfaces, particularly focusing on PMA and IDMI. It cites numerous previous studies on the theoretical and experimental aspects of these phenomena, their applications in spin-orbitronic devices, and the challenges in controlling and enhancing them. The literature review highlights the limitations of existing theoretical models which assume atomically smooth interfaces, contrasting them with the realities of real-world interfaces with roughness and intermixing. It discusses the dependence of IDMI on factors like SOC, broken inversion symmetry, 3d-5d(4d) orbital hybridization and the impact of interface quality, including roughness and intermixing, on the IDMI and the need for larger IDMI values to stabilize spin textures like skyrmions. The review also mentions strategies like using different heavy metals at the top and bottom interfaces and the concept of correlated interface roughness to enhance IDMI and stabilize PMA.

The study employed a hybrid growth process combining molecular beam epitaxy (MBE) and magnetron sputtering. Si(111) substrates were prepared with a Cu buffer layer, followed by the deposition of a Pd seed layer with varying thicknesses (0-12.6 nm) using MBE to control interface roughness. The roughness was characterized in situ using reflection high-energy electron diffraction (RHEED) and scanning tunneling microscopy (STM). After MBE, three series of samples were deposited via magnetron sputtering: MgO-series (Pt/CoFeSiB/MgO/Ta), Ta-series (Pt/CoFeSiB/Ta), and Ru-series (Pt/CoFeSiB/Ru/Ta). The Co70.5Fe4.5Si11B10 ferromagnetic layer was consistently 1.5 nm thick. The surface morphology was further analyzed ex situ using atomic force microscopy (AFM). To analyze crystal structure and interface quality (roughness, intermixing, thickness variation), X-ray diffraction (XRD) and X-ray reflectivity (XRR) were employed, with XRR data fitting using GlobalFit software. Scanning transmission electron microscopy (STEM), electron energy loss spectroscopy (EELS), and high-resolution transmission electron microscopy (HRTEM) were used to investigate the atomic structure and elemental depth profiles. Secondary ion mass spectrometry (SIMS) and energy-dispersive X-ray spectroscopy (EDX) helped to determine the elemental distribution and intermixing. Magnetic properties were assessed through magneto-optical Kerr effect (MOKE), vibrating sample magnetometry (VSM), magnetic force microscopy (MFM), and Kerr microscopy to analyze magnetization reversal processes. The study's methodology involved precise control over the growth process and a combination of surface and structural characterization techniques to correlate interface roughness with magnetic properties.

The research found that atomic-scale surface modulation significantly impacted magnetic properties and IDMI. The root-mean-square (rms) roughness (Rq) of the Pd seed layer varied with thickness (tpd), affecting subsequent layer growth. Three growth modes were observed for Pd: 2D growth (up to 0.6 nm), layer-by-layer growth (0.6-2.9 nm), and 3D island growth (above 2.9 nm). RHEED analysis confirmed these modes. The correlated roughness of the top and bottom interfaces, achieved at a specific Pd thickness (tpd = 10.35 nm), maximized IDMI for all three sample series. HAADF-STEM and HRTEM images revealed epitaxial growth of fcc(111)-Pd and the repetition of surface modulation in subsequent Pt and even CoFeSiB layers (quasi-epitaxial growth), despite CoFeSiB's usually amorphous nature. EELS, SIMS, and EDX confirmed intermixing and boron diffusion at interfaces. The correlated interface roughness, evidenced by a convergence of Rq values for Pd, Pt, CoFeSiB, and the capping layer at tpd = 10.35 nm, led to the highest IDMI values. The samples showed varying magnetic anisotropy depending on the capping layer (Ta, Ru, MgO). The Ta-series displayed fourfold out-of-plane anisotropy, the Ru-series in-plane anisotropy, and the MgO-series primarily out-of-plane anisotropy. The morphological correlation at tpd = 10.35 nm resulted in the peak IDMI values. The study directly demonstrated the dependence of IDMI magnitude on interface quality, showing that controlled atomic-scale modulation via a correlated interface roughness can significantly enhance both IDMI and PMA.

The findings address the research question by demonstrating that precise control of interface roughness at the atomic scale significantly enhances IDMI and PMA in ultrathin ferromagnetic films. The correlated roughness of the top and bottom interfaces, as opposed to simply smooth interfaces, is identified as the key factor for maximizing IDMI. This contrasts with theoretical models which often neglect interface roughness. The results highlight the importance of interface engineering in spin-orbitronic device development, showing that controlled interface roughness provides a powerful tool for manipulating magnetic properties. The observed quasi-epitaxial growth of the CoFeSiB layer despite its typically amorphous nature is surprising and warrants further investigation. The different magnetic anisotropies observed in the different sample series underscore the importance of the capping layer material in determining the overall magnetic behavior. The study's success in achieving correlated roughness provides a significant advance towards reliable and precise control of spin-orbitronic device functionality.

This study successfully demonstrated a method for enhancing perpendicular magnetic anisotropy (PMA) and the Dzyaloshinskii-Moriya interaction (IDMI) in ultrathin ferromagnetic films through atomic-scale modulation of interfaces. The correlated roughness of the top and bottom interfaces proved crucial in maximizing IDMI. The findings highlight the significant impact of interface engineering on magnetic properties and offer a promising pathway for developing advanced spin-orbitronic devices with improved controllability. Future research could explore optimizing the thickness of each layer, exploring other material combinations, and investigating the effect of different types of interface roughness on the magnetic properties.

The study focused on specific material combinations and layer thicknesses. The generalizability of the findings to other materials and structures requires further investigation. The precise mechanism leading to the quasi-epitaxial growth of the CoFeSiB layer needs further exploration. The analysis relied on several sophisticated characterization techniques, which might have limitations in precisely measuring interfacial properties at the atomic scale. The study did not explore the long-term stability of the enhanced IDMI and PMA under various operational conditions.