< Back to previous page

Project

Understanding Interface Interactions in Graphene-Ruthenium Hybrids for Next Generation Interconnects

With every technology generation, the modern IC continues to shrink to accommodate a higher packing density of devices. For 5 nm node, the critical load interconnects would have metal pitch as low as 28 nm. This is much smaller than the mean free path of electrons in Cu resulting in a dramatic and nonlinear rise in interconnect resistance. Moreover, the drop in the metal pitch distances and scaling of oxides increase the capacitance of the BEOL. The overall consequence is a rise in the RC delay and severe performance degradation with every iteration in technology advancement. To mitigate this dominant interconnect delay, alternative materials as well as novel architectures are being explored.

Among the most promising alternative conductors is graphene. Due to its atomic thickness, high current carrying capacity and very high carrier mobility, graphene can outperform Cu interconnects at scaled dimensions, but this is possible for a large number of graphene layers with sufficiently high doping levels. To achieve such high carrier density in graphene, several doping methods (surface doping, intercalation doping, substitutional doping etc.) have been demonstrated in literature but they are difficult to control, often introduce reversible/irreversible defects in graphene lattice and are not compatible with BEOL integration.

In this thesis, we investigate a practical material based solution for advanced interconnects - ‘graphene-metal hybrids’, where ruthenium is used as the metal film. Ru is an excellent choice due to its low bulk resistivity, small mean free path, high tolerance to electromigration and absence of the need for a barrier material. To realize alternating stacks of Ru and graphene, it is essential to study both configurations of the hybrid stack: ‘graphene capped Ru’ (Ru/graphene) and ‘Ru deposited on graphene’ (graphene/Ru).

For Ru/graphene stack, we develop a lab-based process flow by transferring CVD grown graphene on PVD sputtered Ru film whereas graphene/Ru devices are achieved by depositing Ru thin film by e-beam evaporation on the surface of graphene transferred on SiO2 substrate. We focus on studying the interface interactions between graphene and Ru in both configurations.

First, we present a systematic experimental study on measuring the metal induced doping in graphene capped Ru devices qualitatively as well as quantitatively. By resolving electron photoemission spectra from two different conductors, we measure the relative band offset at graphene-Ru interface. We measure the increase in graphene carrier density alongwith the corresponding shift in the position of its Fermi level. We observe that graphene is p-doped on Ru substrate resulting in a ~19% improvement in hybrid conductivity as compared to uncapped Ru devices for smaller film thickness. Then, we study the effect of number of graphene layers on the electrical performance of the hybrid devices and find the lowest resistivity for Ru(5nm)/FLG stack which is ~ 26% higher than that of Ru/SLG. Also, by achieving a 30% smaller temperature coefficient of resistivity upon graphene capping, we demonstrate higher thermal reliability of the hybrid structures as compared to their metal counterpart.

Secondly, to obtain a clean and stable interface in graphene/Ru devices, we optimize a downstream plasma surface treatment process for graphene. By measuring the plasma induced doping and defectivity, we find that both SLG and BLG become n-doped upon plasma exposure however the electrical response is different in both cases. While SLG is found to incur significant damage due to its high chemical reactivity, the electrical conductivity of BLG/Ru is found to improve by ~18% upon plasma treatment.Unlike in ‘graphene capped Ru’, we find that the hybrid resistivity in ‘Ru deposited on graphene’ configuration is higher than the reference Ru. The nature of interface interactions can be very different when Ru is deposited on graphene as compared to when graphene is transferred on a Ru substrate. To understand this phenomenon, we assess the thin film properties of the evaporated metal film on graphene. We find that even though the interface adhesion force implies a stable interface, the degradation of device resistivity is mainly attributed to a discontinuous, porous metal film with several current percolation paths. The evaporated film has much smaller grain size composition and higher residual stress when deposited on graphene resulting in larger scattering events and lower conductivity.

Finally, to elucidate the role of graphene in modulating the current conduction in graphene-Ru hybrid systems, we perform semi-classical theoretical resistivity modelling. For this, we perform a direct search method on full numerical formulation of Mayadas-Shatzkes model. Upon calculating the surface specularity parameter and the grain boundary reflection coefficient, we find that since Ru surface is already almost specular, the impact of graphene related suppression of surface scattering would not contribute significantly to the hybrid resistivity. We then compute the parallel conductor model where the resistivities of graphene and Ru are bound to have influences from each other as well from the interface between them. We find a good correlation between the theoretical and experimental values of the carrier density and shift in graphene Fermi level associated with the metal induced doping in graphene due to contact with the underlying Ru film, thereby validating the parallel conduction model.

Hence, by performing a systematic study on understanding the interface interactions in graphene-Ru hybrid systems, this thesis provides experiential insights that form a viable and practical basis for integrating graphene-metal hybrids as conductors in next generation interconnects.

Date:6 Nov 2017 →  12 Jan 2022
Keywords:Graphene, Interconnects, Nanoelectronics
Disciplines:Nanotechnology, Design theories and methods, Ceramic and glass materials, Materials science and engineering, Semiconductor materials, Other materials engineering
Project type:PhD project