(113d) Catalyst Nanoparticle Generation Routes for Deep Injection Floating Catalyst Chemical Vapor Deposition

Carbon nanotubes (CNTs) are among the most promising candidates to meet the demands of next-generation materials applications, from high-strength light-weight construction to the development of field effect transistors. The future development and adoption of many of these technologies depends on a supply of CNTs that are high-quality (i.e., long and highly crystalline) and affordable enough to enable competition with legacy technologies.

Floating catalyst chemical vapor deposition (FCCVD) is the most scalable method for the production of high quality few walled CNTs due to its continuous nature and lack of a catalyst substrate. FCCVD operates by using an aerosol of catalytic nanoparticles composed of iron or a similar metal suspended in the process gas. These floating particles serve as active sites for CNT growth. This reduces the number of post-processing steps needed to purify CNTs. The FCCVD technique also allows for the direct production not only of CNT powders, but also of fibers, films, and foams directly from the reactor. The recent development of deep injection floating catalyst chemical vapor deposition (DI-FCCVD), in which reactants are injected deep into the deep into the high temperature zone of the reactor has dramatically improved the production rates of high-crystallinity CNTs [1]. Despite these advantages, FCCVD and DI-FCCVD are still limited by poor reactant utilization and low process intensity, leading to high production costs for high-quality CNTs.

Generating a catalyst nanoparticle aerosol flow of the appropriate diameter and flow, with narrow polydispersity, is a critical step in the FCCVD process [2]. Most catalyst nanoparticle production methods rely on the thermal decomposition of organometallic precursors in situ without process control or validation, resulting in poor control over catalyst particle size and composition [3]. Prior observations suggest that poorly controlled particle diameter may contribute to low catalyst utilization and low efficiency in the FCCVD process [3] Catalyst utilization, which we defined as the proportion of catalyst material that successfully grows a CNT, remains below 5 % in current reactors. DI-FCCVD is also limited by the coagulation of catalyst particles. Thus, we hypothesize that the apparent trade-off between high throughput particle synthesis and particle size is due to poor catalyst particle size control.

To date, DI-FCCVD has only been studied with catalyst formed by precursor decomposition. The use of pre-formed catalyst particles offers a potential route to further improve catalyst efficiency in these processes. To this end, we have developed a reliable catalyst generation system using an atmospheric pressure radio frequency plasma in a mixture of argon and hydrogen. Our system decouples the catalyst formation process from the CNT growth process, allowing for process parameters to be individually optimized for both. The high energy density of plasma makes it ideal for the formation of catalyst particles, as it can utilize metallic catalyst sources, such as an iron wire, rather than a precursor molecule, removing extraneous carbon from the reaction. In addition, it is well known that nanoparticles created in plasma are primarily negatively charged which slows particle agglomeration, enabling narrow size distributions [4] We developed a plasma based particle generation system and characterized the effects of process geometry, the effects of plasma power, post-plasma dilution of the particles, and the inclusion of process additives, such as sulfur compounds by scanning mobility particle sizer, transmission electron microscope, and scanning electron microscope. We supplement our empirical characterizations by modeling the plasma in COMSOL. We then coupled this system to a DI-FCCVD process in order to synthesize high quality CNTs. We present our results in optimizing this process to improve productivity, catalyst utilization, and reactant conversion to few walled, high aspect ratio CNTs. We expect this work to serve as a starting point for studies demonstrating the enhancement of control and scalability of catalyst particle delivery for DI-FCCVD reactors.