Modeling Surfactant-Tuned Separation Processes for Carbon Nanotubes (CNT)

Single wall carbon nanotubes (CNTs) were first discovered in the early 1990s. This new form of carbon material possesses remarkable thermal, mechanical, and electrical properties, positioning it as a promising candidate for a wide range of applications, including field-effect transistors, display technologies, hydrogen storage, and so on. The market value of CNTs was valued at $2.6 billion in 2019 and is expected to grow to $5.8 billion by 2027. These highly-anticipated qualities have made CNTs the focus of significant research efforts since their initial discovery.

The application of CNTs is far from realizing its full potential and is largely not commercially available in most cases. Methods for synthesizing nanotubes typically generate a carbonaceous soot-like powder containing a minority population of numerous different CNT species. These CNT species are uniquely identified by their diameter and chiral vector (𝑎𝑎1, 𝑎𝑎2) and each species has different functionalities. To date, neither synthesis methods nor separation processes can produce a CNT sample containing only a single species. Using mixtures of distinct CNTs can significantly compromise their material properties and reduce their advantage compared with other newly-developed nanomaterials.

Isolating specific carbon nanotube species, however, is not trivial. Current separation methods are expensive to implement, require specialized equipment, offer low throughput, or present challenges to cost-effective scale-up. The lack of scalable technology for species-based CNT separation has prevented the widespread application of carbon nanotubes. A simpler and cheaper separation technique is thus called for to facilitate large-scale processing.Aqueous two-phase systems (ATPSs), formed by mixing two polymers or a polymer and a salt, have been used as a separation and purification tool for more than 50 years. With nearly 80% water content, ATPS can achieve separations with minimal damage to particles with delicate structures such as CNTs. Because of its potential for continuous operation and process integration, ATPS is particularly promising to meet the downstream processing needs created by the fast-growing production rate of carbon nanotubes. The partitioning behavior involved, however, is complex and difficult to predict due to the large number of factors that can alter the process. The application of ATPS is therefore limited without the technology to easily optimize the desired partitioning behavior across a very large parameter space. Here, we develop and analyze a simplified model to systematically investigate the impact of various factors on the partitioning behavior of carbon nanotubes in an ATPS process. The model is based on high-resolution data previously obtained from droplet-based millifluidic experiments characterizing the partitioning behavior of CNT in PEG-DEX two-phase systems with added surfactant as the primary tuning factor. Using the model we developed, we have identified several dimensionless parameters that capture the salient features of the partitioning behavior of different CNT species, making this model an efficient and flexible tool for rapidly exploring separations over a large parameter space. In this thesis, we use the model to simulate one-stage and multi-stage aqueous two-phase separations between two CNT species. Yield and purity are reported as the separation performance metrics for both species after every stage. A threshold of 0.8 is proposed as the minimum requirement for both yield and purity to mimic the industrial operation process. For some types of CNT samples, only one-stage separation is needed to reach the threshold. Other types of samples require multi-stage processes. Under specific conditions, one CNT species can be concentrated 5000 times with a two-stage process, validating the purifying power of this separation technique.Three key parameters in the model are varied to simulate a broader representation of different CNT species. One-stage separation results are reported for two CNT species under different parameter variations. Parameter variations that can make the partition coefficient curve of one CNT species more distinguishable than other species allow more experimental conditions to be used to reach the same separation performance.