Towards the Automation of Scalable Carbon Nanotube Sorting via Aqueous Two-Phase Extraction
MetadataShow full item record
Carbon nanotubes are an exciting quantum material with numerous interesting and unique properties. Notable electric properties are their variable band-gap and metallic and semi- conducting variants. Carbon nanotubes are a prime candidate for beyond-silicon devices and have already been used in experimental computer processors. Carbon nanotubes have valuable optical, chemical, and mechanical properties as well. Many of these novel properties such as the band-gap and absorption spectrum depend on the carbon nanotube’s chirality, which is related to its diameter. Unfortunately, no method for carbon nanotube synthesis has yet been able to produce pure, single-chirality nanotubes. Applications requiring a specific chirality must thus sort the synthesised nan- otubes after the fact. The most scalable and affordable sorting method is aqueous two- phase extraction (ATPE). However, this procedure requires significant time and skill as it involves many stages of micro-pipetting. The subject of this thesis is the automation of ATPE. Since this is a water-based procedure, the core around which the automated system is built is an off-the-shelf liquid handling robot. The majority of this text describes the custom extensions to this robot that make it capable of ATPE. The Photo Stage provides an environment to capture consistent images of the two-phase system, and the accompanying computer vision software analyses the image and locates the phase boundary. This information is used to programmatically extract the top phase. Transportation of vials is accomplished by the addition of the Gripper. The caps are removed from vials and tubes using the Vial Opener. The Powder Handling System has the goal of automating the creation of chemical stocks. The Inventory System is designed to track the contents and history of ATPE process vials. Finally, a robotic arm will integrate external instruments including a centrifuge, sonicator, and spectrophotometer. However, this is not complete and is planned for the second generation of the project. Without the integration of the centrifuge, which is required at every step of ATPE to partition the solution into two phases, it is impossible to carry out ATPE from start to finish. Thus, the system is being used in a mode referred to as hybrid ATPE. Under these conditions, when the centrifuge is needed, the system is paused so that a human operator can transfer the vials to the centrifuge. When the centrifuge has finished, the vials are transferred back to the liquid handling robot and the system is resumed. This hybrid system takes advantage of the Gripper, Vial Opener, and Photo Stage. The Inventory System is incomplete in that the camera has not been mounted, however the associated code can already be used to track the contents of vials if the identifier is entered manually rather than automatically being read from a barcode or other marker. The Powder Handling System, which is for the pre-processing step of stock-making and not for the main ATPE process, is not complete. A precision balance has been purchased and a conveyor has been developed, but there have been multiple failed attempts to fabricate a powder dispenser. Thus, the Powder Handling System is not used and stocks are still created manually. Finally, the external robotic system is not complete and is not used for hybrid ATPE. Under hybrid ATPE, it has been possible to automatically extract the top phase from two-phase systems. The reliability of the computer vision boundary recognition system has been increased to > 95 %. However, the results of automated separation leave room for improvement. The robotic system leaves more discarded waste around the boundary than when ATPE is performed manually, and there is a heightened risk of contaminating the boundary. The reason for this is that the pipettes of the liquid handling robot are always vertical, parallel to the vial. This prevents the liquid that wicks up the vial walls by capillary action from being properly extracted. This will be addressed by future additions to the robotic system.
Cite this version of the work
Marcel Robitaille (2022). Towards the Automation of Scalable Carbon Nanotube Sorting via Aqueous Two-Phase Extraction. UWSpace. http://hdl.handle.net/10012/18839