NANOTECHNOLOGY
Capping Ru with graphene
For this study, the imec researchers realised Ru/graphene hybrid structures by transferring a multilayer graphene film (grown by chemical vapour deposition (CVD)) onto a thin Ru film (typically 5nm) that was grown by physical vapour deposition (PVD). After transfer, graphene was found to adhere well to the large area PVD Ru film.
Figure 2:
Comparing properties of graphene (carbon nanotube (CNT), single layer graphene (SLG) and few layer graphene (FLG) with other interconnect materials of interest (tungsten (W), copper (Cu) and ruthenium (Ru)
Why graphene?
In recent years, graphene has been intensively studied for interconnect applications, where it has potential to fulfill diverse roles. The material has for example been considered as an oxidation barrier and as an ultrathin diffusion barrier for metals. Researchers have also investigated the feasibility of using multilayer graphene wires or nanoribbons as an alternative conductor.
Interest in graphene for interconnect applications comes as no surprise. Graphene exhibits a high intrinsic carrier mobility (up to 200,000cm2V-1s-1) and a large current carrying capacity (up to 108A/cm2). In addition, graphene has a high thermal conductivity and competitive robustness against electromigration. It can also be made atomically thin, which helps alleviating the thickness contribution to the RC delay.
Despite these interesting properties, graphene has one major drawback: intrinsically, it does not hold enough charge carriers to be useful as a local interconnect. The lack of charge carriers severely reduces its electrical conductivity, a key metric for interconnect performance that is proportional to both the mobility and the carrier concentration. For this reason, several layers of graphene will be needed to cross-over Cu for example, for (local) interconnect applications – as confirmed by modeling. The number of layers will be a trade-off between the material’s overall contribution to resistance and capacitance. Fortunately, there are ways to further modulate graphene’s conductivity. This has driven the research of so-called graphene nanoribbons – graphene layers patterned into narrow strips. The specific angular orientation of the graphene layers with respect to their underlying layer provides another knob for improvement. Finally, the conductivity of graphene can be boosted by doping, this way providing graphene with extra electrons or holes to carry the current. Doping can be performed in several ways, for example by metal-induced doping – enabled by bringing graphene in direct contact with metals like Cu or Ru. These hybrid metal/graphene schemes bring together the best of both ‘worlds’: the high carrier concentration of the metal and the high mobility of graphene.
This article looks into the feasibility of using hybrid metal/graphene structures for sub-2nm interconnect applications. Two different structures are being examined: graphene-capped metal and metal-capped graphene devices. The study focuses on Ru as a metal of interest, as it has recently emerged as a potential alternative for Cu metallization. But the concepts presented here are expected to be expandable towards other ‘interconnect’ metals.
Of interest for interconnect applications is the metal-induced doping of graphene that is expected to happen at the interface with Ru. To understand and be able to control the doping, the charge transfer at the Ru/graphene interface was systematically investigated. The two main observations can be summarised as follows. First, the researchers found that the sheet resistance of Ru dropped by an average of 15% after encapsulation with graphene. Second, they observed a downward shift of graphene’s Fermi level into the valence band by ~0.5eV compared to intrinsic graphene, corresponding to a hole concentration of 1.9E13cm-2. This observation is an indication of metal-induced doping that happens at the interface, causing graphene to become p-doped when added as a capping layer to Ru.
From this study, it has become clear that encapsulating Ru with graphene can boost the electrical performance of Ru as an interconnect. Yet, more fundamental insights are needed to determine the exact conduction mechanism taking place within the capped structure. Either Ru remains the main conductor, with graphene helping to reduce its resistivity by suppressing scattering mechanism(s) in the metal. Or the two conductors now act in parallel, with a higher conductivity for graphene (with respect to intrinsic graphene) because of the charge transfer. Modeling work is currently ongoing to get a better understanding. It should also be noted that the Ru lines are observed to be less sensitive to temperature fluctuations when encapsulated with graphene. This could be due to the high thermal conductivity of graphene, providing an alternative/additional path for efficient heat dissipation. This observation is of interest for future interconnect applications, as the self-heating in highly scaled IC wires and an insufficient heat dissipation to the surrounding dielectric can degrade the interconnect’s thermal reliability. Overall, the researchers conclude that these graphene-capped metal hybrid structures provide an answer to the RC delay problem for future interconnects. Imec envisions their introduction in the BEOL technology roadmap for the 1nm node and beyond.
Figure 3: Experimentally measured sheet resistance of bare Ru (black) and graphene-capped Ru (red) devices for different thickness of Ru thin film substrate [as presented at IITC 2019]
JUNE 2021 | ELECTRONICS TODAY 41
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50
