Despite advances in single-atom doping, the controlled assembly of multi-atom defects in 2D materials remains difficult, limiting the ability to tailor their electronic, magnetic, or catalytic functionality. Here, we demonstrate that ultralow-energy (82 eV) implantation of mass-selected Mn2 cluster ions into monolayer graphene enables the formation of substitutional Mn dimers. Low-temperature scanning tunneling microscopy and spectroscopy, supported by density-functional theory calculations, reveal that these Mn–Mn pairs exhibit a characteristic four-lobe signature and distinct spectroscopic fingerprints. Molecular dynamics simulations show that Mn2 dimers fragment upon impact, displacing several carbon atoms and enabling recombination events that yield substitutional dimers. These results show that cluster-ion implantation can create multi-atom defects in 2D materials, extending ultralow-energy implantation beyond single-atom doping. This provides access to a broader range of defects relevant to spintronics, quantum technologies, and catalysis.

Foundry-compatible transfer of two-dimensional materials is central to angstrom-era complementary FET logic, where atomically thin channels must be moved from growth wafers to target stacks without contamination, strain or interfacial damage. Scalable delamination, bonding and wafer-level metrology are needed to bridge lab-scale device performance with 300-mm CMOS manufacturing.

Strain is a key tuning parameter in solid-state systems, but most studies focus on strain fields extending over tens of nanometers or more. Here we investigate the extreme limit of ultra-localized strain in graphene, introduced through bond defects generated by ultralow-energy implantation of noble gas ions. Using molecular dynamics simulations, Raman spectroscopy, and scanning tunneling microscopy, we identify the formation and thermal stability of bond defects that locally stretch only a few C-C bonds without removing or substituting atoms. Tight-binding calculations reveal that such bond defects induce local charge trapping, leading to substantial Fermi-level shifts. Synchrotron-based angle-resolved photoemission spectroscopy directly confirms these predictions: even at modest defect densities (similar to 1012 cm-2), the graphene Fermi level shifts by up to 0.3 eV. This strong effect is remarkable given that it is achieved without altering graphene's composition, in contrast to conventional impurity doping or vacancy formation. Upon thermal annealing, the electronic structure recovers towards the pristine state, showing that these effects can be tuned and reversed. Our results establish bond defects as a new class of functional disorder in graphene, capable of strongly modifying its electronic properties solely by bond rearrangement.

Graphene has emerged as a promising material for integration into silicon photonics, owing to its ultrafast and broadband photoresponse without the need for an external bias voltage. This photoresponse relies on the photo-thermoelectric effect created by hot carriers. A key factor underlying the performance of graphene photodetectors is the cooling dynamics of these hot carriers. In this work, we engineer these dynamics in a WSe2-graphene-WSe2 waveguide-integrated photodetector. In particular, by introducing proximity screening by a nearby graphite layer to this structure, we prolong the hot-carrier cooling time, leading to an enhanced photoresponse. We characterize the cooling dynamics under continuous-wave laser excitation by employing a photomixing technique, revealing an increase in the cooling time by up to a factor of four. Direct photoresponse measurements show that the internal photoresponsivity improves by approximately 50

We demonstrate the first co-integration of a wafer-scale single-layer graphene electro-absorption modulator (EAM) with a high-speed SiGe bipolar CMOS (BiCMOS) electronic integrated circuit (EIC), achieving robust operation up to 25 GBd. The scalable graphene EAM, fabricated in a 300 mm CMOS pilot line, exhibits an intrinsic electro-optic (EO) bandwidth of 11.2 GHz. To overcome the RC-limitation and simultaneously enable low-voltage operation, the EIC was co-designed with a 30 Omega internal termination impedance and amplifies the voltage swing by a factor of 3. This optimization increased the effective EO bandwidth and allowed the co-integrated system to produce open eye diagrams at 25 GBd from a low 500mVpp input drive voltage. We report measured Q-factors and an estimated bit error rate of 1.37 & times;10-7 at 25 GBd using a 5-tap feed-forward equalizer. This work demonstrates the first functional EO link co-integrating a graphene modulator with a BiCMOS electronic driver. By validating this wire-bonded interface, we establish a feasible and practical route towards energy-efficient, scalable photonic interconnects.

Continued scaling of interconnect dimensions into the sub-10 nm regime challenges conventional Cu dual-damascene metallization due to rising resistivity, reliability degradation, and integration complexity. Ru is a promising alternative; however, polycrystalline Ru remains limited by grain-boundary scattering and patterning constraints at advanced nodes. Here, we present a metal layer-transfer integration scheme that enables single-crystal Ru interconnects by decoupling epitaxial metal growth from BEOL patterning. High-quality epitaxial Ru films are sputter-deposited on AlN (0001)/Si (111) at BEOL-compatible temperatures and transferred to device wafers via Ru–Ru wafer-to-wafer bonding. Structural characterization confirms excellent crystalline quality and coherent epitaxial alignment, while electrical measurements demonstrate near-bulk resistivity with weak thickness dependence. Vacuum bonding with surface activation followed by post-bond annealing yields void-free Ru–Ru interfaces and induces solid-phase epitaxial regrowth across the bonding interface. These results indicate that metal-to-metal bonding-based layer transfer can become a viable pathway for integrating single-crystal interconnects in future technology nodes.

Backside power-delivery-network (BSPDN) schemes require wafer-to-wafer bonding steps that do not leave high order shape changes or localized stresses in the bonded stack to ensure that these don't get transferred to the thinned wafer for further lithographic exposure. However, bonding mechanics involve strong adhesive forces which can inherently create localized distortions that lithography tools must eventually compensate. Some contributors to grid distortion of the target wafer are, method of bond initiation, bond front velocity variations, and lack of symmetry between the wafers. In this work, we evaluate a low-distortion bonding approach in the SUSS XBA tool, where some of these contributors are tackled at the source, namely by initiating the bond without a localized external force, and keeping the wafers compliant symmetric during bonding. Wafers are bonded with a slight pre-stress due to controlled gas pressure applied over the whole backside of the wafers throughout bonding. Wafer-shape measurements of the bonded stacks are used to perform gradient-based in-plane-displacements (IPD) modelling to estimate bonding-induced grid distortion. Dense scanner metrology is used on patterned-bonded wafers to confirm the location and severity of distortion predictions from patterned wafer geometry (PWG) measurements. On the PWG distortion maps and scanner grid readouts, we perform alignment and CPE modelling with different field layouts. Sub-10 nm levels of residual grid distortion are achievable with relatively low-order correction models, and less than or equal to 3 nm by using advanced CPE models. These results demonstrate that pneumatically warped bonding yields low-distortion bonded stacks, and that simple process tuning can decouple the dominant distortion source from edge-related variability.

In this letter, we demonstrate an actively mode-locked laser at telecom wavelength made of graphene electro-absorption modulator on a photonic integrated circuit, inserted in a dispersion-engineered cavity with III-V semiconductor optical amplifiers as gain medium. Stable low-jitter pulse trains at 10 GHz with 2.9 ps deconvoluted pulse widths are achieved, with performances comparable to a commercial lithium niobate high-speed modulator. These results highlight the potential of integrated graphene-based devices for the realization of compact and stable optical clocks operating at high repetition rate.

This work demonstrates the integration of silicon temporary carriers with an infrared (IR) laser release process to enable fine pitch redistribution layer (RDL) fabrication for fanout wafer level packaging (FOWLP). While IR laser debonding has been shown to be highly effective on fully inorganic stacks, its interaction with organic layers used in RDL first flows remains less explored. We investigate the impact of polymer mechanical properties on the release behaviour and introduce two key adaptations to the multilayer release (MLR) stack: replacement of the reflective layer with a TiW optical decoupling layer compatible with standard lithography, and insertion of an organic or inorganic buffer layer to ensure electrical insulation and simplify post release cleaning. Blanket polymer transfer studies establish optimized laser parameter windows that prevent defect formation related to incomplete debonding or delamination between layers. These optimized conditions are then applied to a two level, 1.6 µm pitch damascene RDL stack, which is successfully over-molded, bonded, and transferred with Cu pads exposure after flip and wet etching. Electrical measurements confirm low leakage and line resistance values of 20–30 Ω, validating compatibility with advanced patterning flows. The results confirm that Si carriers combined with IR laser release provide a robust, low warpage and fully production compatible platform for high density RDL first FOWLP manufacturing.

Using substitutional Mn in graphene/Cu(111) as a model point defect, we combine scanning tunneling microscopy (STM) and angle-resolved photoemission spectroscopy (ARPES) to test the predicted fingerprints of resonant scattering. As the Mn concentration increases to 0.44

Wafer-to-wafer hybrid bonding is emerging as a key enabler for dense three-dimensional integration, yet its progress is constrained by misalignment between bonded wafers. Here we present a comprehensive, decade-long analysis of bonding overlay performance across multiple generations of bonding equipment and test vehicles, revealing how systematic and stochastic distortions limit alignment accuracy. Advancement in bonding hardware and integration flow have reduced total overlay errors by nearly an order of magnitude. Additionally, the application of lithography pre-compensation can halve the dominant residual distortion, enabling a practical route toward sub-25-nanometre bonding overlay without relying solely on next-generation bonders. Principal component analysis links remaining variation to incoming wafer shape and recipe-dependent scaling control, defining the fundamental limits and guiding a roadmap for achieving high-yield, nanometre-accurate hybrid bonding in future three-dimensional systems

Temporary bonding and debonding techniques are essential for advanced semiconductor processing, particularly in 3D and 2.5 D integration. Conventional methods using polymer adhesives and UV laser debonding face limitations including low thermal stability, high total thickness variation (TTV), and poor mechanical alignment, associated with the use of glass carrier wafers. This paper further assesses a novel infrared (IR) laser debonding process based on a fully inorganic bonding stack deposited on reusable silicon carriers. The proposed technology enables higher temperature processing, improved wafer thinning and wider applicability due to the use of Si carrier instead of glass. A comparative cost-of-ownership (CoO) analysis between the IR laser release and full wafer grinding approach demonstrates substantial savings when carrier reuse is implemented. Cleaning strategies for both device and carrier wafers are evaluated, and stack optimizations are proposed to further reduce costs while maintaining process performance. These findings support the viability of IR laser debonding as a scalable and cost-effective solution for nextgeneration packaging technologies.

We present our efforts to facilitate 300mm MX2 growth on sapphire and completely automate the layer transfer module. Two concepts are presented; a collective die-to wafer approach (CoD2W), and a 300mm wafer-to-wafer approach (W2W), both with the vision of enabling MX2 integration in foundries for BEOL and/or FEOL applications.

The development of temporary bonding solutions for advanced packaging industry allowing mechanical support during backside processing, is driven by the demand for more robust and reliable solutions in this field. Notwithstanding the growing acceptance of UV laser debonding technology, siliconbased carrier systems persist in their status as the preferred option. This paper details the findings of a study on the application of an infrared (IR) laser debonding technique combined with a temporary bonding adhesive on silicon carrier wafers. The aims of this study were twofold: firstly, to evaluate the potential of IR laser debonding, a technique that permits higher processing temperatures, thereby facilitating advancements in adhesives and other polymer-based processes, such as Redistribution Layer (RDL)-first applications; and secondly, to assess the cleanability of residues remaining after IR laser exposure on both the product and silicon carrier wafer debond interfaces. It has been demonstrated that the well-controlled debond mechanism and interface enable the application of various bonding methods, such as fusion or temporary bonding adhesives, by employing similar inorganic IR laser release layers. It is also shown that after IR laser debonding, the inorganic release layer residuals and the applied polymer layer can be removed using standardized cleaning methods, ensuring the technology's suitability for high- volume utilization and easy integration.

Two dimensional (2D) materials are at the forefront of research in integrated modulators. However, achieving pure phase modulation with low insertion loss remains challenging in 2D material modulators. This work explores phase modulators based on monolayer MoS2 integrated on a Si photonic platform. We investigated the electro-optical response of MoS2 using two device architectures: a monolayer MoS2 integrated on a doped Si waveguide and a MoS2-Al2O3-MoS2 capacitor integrated on an undoped Si waveguide. At a wavelength of 1550 nm, the integration of MoS2 in both architectures induces minimal optical insertion losses. Specifically, the single-layer device on a doped silicon waveguide achieves a phase modulation efficiency of 0.53 Vcm, with propagation losses dominantly originating from the doped silicon (108 dB cm(-1) ). The MoS2-Al2O3-MoS2 modulator has a low propagation loss of 6.4 dB cm(-1 )and modulation efficiency of 1 Vcm. Depending on the resistance-capacitance constant, a trade-off exists between achievable modulation efficiency, electro-optical bandwidth, and insertion loss. The strong and lossless electro-refractive response makes MoS2-based phase modulators suitable for a variety of applications such as light detection and ranging, optical switching, quantum and optical neural networks, and coherent optical communication.

We demonstrate both passive and active mode-locking of fiber lasers using a graphene-based absorber integrated on a silicon photonics platform manufactured using a wafer-scale CMOS process. Passive mode-locking is achieved at a 28 MHz repetition rate, generating 1.7 ps optical pulses, while active mode-locking is demonstrated at repetition rates of 4 GHz and 10 GHz. This work demonstrates the potential of scalable graphene-based saturable absorbers for fiber laser locking and paves the way for future fully integrated mode-locked laser systems.

The continuous expansion of two-dimensional materials research since the first developments of over 15 years ago has enabled tremendous progress in the fundamental understanding of their properties and behavior. The promises held by these materials to facilitate scaling beyond silicon-based device architectures are still valid, but the manufacturability and integration with silicon technology remain challenging. On the metrology side, characterization of the device channel and assessment of the expected performance is lacking, at least in a fully non-destructive and process line-compatible implementation. The current paper demonstrates a clear correlation between metrics associated with the transistor performance on one hand, and parameters from photoluminescence spectra on the other. The concept is demonstrated on state-of-the-art 300 mm process MoS2 devices, without the need for specific measurement conditions or sample preparation. Being truly non-contact and relatively fast, this analysis provides the community with a potential route toward non-invasive material quality assessment, applicable at several stages of the process and with a direct connection to device performance.

As interconnect scaling approaches sub-10 nm critical dimensions, conventional Cu interconnects face increasing challenges, including resistance scaling, degraded electromigration, and complex integration requirements. Recently, Ru has emerged as a promising alternative for barrier-and linerless interconnects due to its more favorable resistance scaling and superior intrinsic electromigration resistance. Epitaxial single-crystal Ru presents additional advantages over polycrystalline Ru, since the absence of grain boundaries reduces resistivity and enhances etching behavior. In this paper, we demonstrate the epitaxial growth of Ru (001) on Si (111) wafers using an AlN (001) buffer layer, an approach that is scalable to 300 mm wafer processing. Furthermore, we explore the feasibility of Ru epitaxy with alternative crystallographic orientations on sapphire substrates. Finally, we discuss potential integration pathways for epitaxial Ru films in advanced interconnects.

Although ultralow energy (ULE) ion implantation is an effective method for substitutional doping of graphene with transition metals, it generally results in substantial nonsubstitutional incorporation, such as atoms intercalated between the graphene layer and the substrate or incorporated in the substrate subsurface. These nonsubstitutional components can have undesired or uncontrolled effects on the electronic properties of the doped graphene layer. Here, we demonstrate that graphene, substitutionally doped with Mn via ULE ion implantation, can be successfully transferred using a standard wet transfer process. This method preserves the substitutional Mn while removing the nonsubstitutional Mn present in the pretransfer surface, as evidenced by X-ray photoelectron spectroscopy, X-ray absorption spectroscopy, and scanning tunneling microscopy. Furthermore, the transferred Mn-doped graphene retains its characteristic Dirac band structure, as shown by angle-resolved photoemission spectroscopy. These results demonstrate the feasibility of transferring substitutionally doped graphene while maintaining its structural and electronic integrity. This work provides a practical route not only for studying graphene doped by ULE ion implantation using surface-sensitive techniques, free from the complications posed by nonsubstitutional components, but also for integrating it into complex structures, such as stacking with other 2D materials or transferring onto virtually any substrate or device structure.