We got used to have faster, smaller and more power-efficient gadgets every year. This fantastic progress has been enabled by miniaturization of silicon transistors (so-called Moore’s law). Nowadays, the classical scaling is reaching its physical limits with critical...
We got used to have faster, smaller and more power-efficient gadgets every year. This fantastic progress has been enabled by miniaturization of silicon transistors (so-called Moore’s law). Nowadays, the classical scaling is reaching its physical limits with critical dimensions as small as just a few nanometers. This explains the growing interest towards novel materials that can potentially replace silicon in future high performance electronic devices. Two-dimensional materials such as graphene and transition metal dichalcogenides (TMDs) have attracted significant interest due to their unique electronic, optical and mechanical properties. Field effect transistors with channels made of atomically thin 2D semiconductors are less prone to short channel effects which allows device scaling to a few nanometer scale. Due to the absence of surface dangling bonds 2D materials can be assembled in heterostructures with sharp interfaces and specific band alignments paving the way to fabrication of novel low power devices.
While exceptional properties of TMDs have been demonstrated on mechanically exfoliated flakes (using the Nobel-winning scotch tape technique) large scale integration of these materials into sophisticated devices remains very challenging. Due to the extreme sensitivity and fragility of 2D crystals in comparison to classical bulk materials, all processing technologies used in semiconductor fabs (e.g. deposition, etching, cleaning) have to be reexamined. The goal of the PULSE2D project is to develop plasma-based technology for cleaning, functionalization and etching of TMD materials with atomic-scale precision (Figure 1).
Application of developed processes will allow integration of 2D TMDs on full wafers in semiconductor fabs leading to fabrication of new generations of high-performance devices.
WP1 focused on the elucidation of the fundamental mechanisms of plasma-TMD interactions. It has been shown that TMD monolayers are easily damaged even by low energy (about 10 eV) ions arriving on the surface in the absence of any bias. This means that for low-damage processes such as cleaning or functionalization, ion bombardment should be fully excluded. That is why the application of a remote plasma source has been thoroughly investigate in this work package. In particular we used a remote H2 plasma to selectively remove polymer residues from WS2 (MoS2) surface. The residues originate from material transfer and pattering and they can not be fully removed by any existing technique which is a major roadblock for TMD integration. We have demonstrated that polymer residues are fully eliminated by the remote H2 plasma with only a minor modification of material properties. Moreover, the mechanisms of plasma-surface interactions have been investigated using ab-initio simulations performed in collaboration with the Moscow State University and with the University of Antwerp. In addition, remote hydrogen plasma in combination with molecular doping has been successfully applied to reduce contact resistance in TMD-based field effect transistors.
WP2 focused on the application of atomic layer etching (ALE) technique for wafer-scale integration of TMDs in collaboration with Oxford Instruments Plasma Technology (OIPT). Selective etching of gate dielectrics is a critical step for contacting the TMD channel in field-effect transistors. ALE of high-k dielectric (ZrO2, HfO2, Al2O3) layers in BCl3/Cl2 chemistry has been investigated. We have shown that by using ALE gate dielectric materials can be removed selectively to SiO2 deposited at the TMD – high-k interface. This allowed for the first time formation of top contacts to 2D layers in a fab-compatible manner.
WP3 deals with the improvement of film closure during atomic layer deposition of gate dielectrics on MoS2 (WS¬2). We have shown that remote hydrogen plasma pre-treatment combined with molecular doping with Cl2 and OCS can be used to improve ALD nucleation on TMD surface. However, the improved nucleation comes at the expense of material damage (nucleation sites are associated with surface defects). Therefore an alternative ALD process with extended exposure of the TMD surface to precursor molecules has been proposed. This new approach is being investigated at the moment when this report is being written.
Plasma processing of TMD materials on full wafers in fab environment is a completely new field of research and development. Imec is a global leader in wafer scale integration of 2D monolayers, which means that the PULSE2D project has a unique position with an access to state of the art materials and technologies. Low damage remote hydrogen plasma cleaning of polymer residues from transferred WS2 and MoS2 developed in WP1 opens up routes for integration of high quality 2D materials grown on special substrates and then transferred on the target wafer. None of the existing cleaning technologies are able to fully eliminate polymer residues without damaging 2D layers. We have also shown that plasma cleaning can reduce contact resistance which is of foremost importance, because contacts limit the performance of ultra-scaled TMD devices. The fundamental understanding of interaction between remote H2 plasma and TMDs achieved through a combination of experiments and ab-initio simulations is important for defect engineering in 2D semiconductors which can be used in future electronic and optoelectronic applications of TMDs.
Overall, plasma processes developed in the framework of the project together with the fundamental understanding of plasma-surface interactions with 2D materials might be applied in the coming years for fabrication of new devices based on 2D materials for logic, memory and optoelectronic application.