UC Berkeley

Vertically-aligned carbon nanotube (VACNT) arrays are both an important technological system, and a fascinating system for studying basic principles of nanomaterial synthesis. However, despite continuing efforts for the past decade, important questions about this process remain largely unexplained. Recently, nanotube research investigations have been conducted, aiming at revealing the underlying growth mechanisms, rather than merely studying the feasibility on new growth methods. Nonetheless, growth deactivation and the accompanying termination mechanisms still remain a topic of nanotube synthesis science. Due to the extremely small size, however, direct characterization of various transport and conversion events occurring at the catalyst surface is not an easy task. Thus investigations on growth kinetics are the first step to resolve questions about growth mechanism.

Before exploring kinetic aspects of the growth process, one must achieve reliable growth conditions since growth non-reproducibility retards obtaining reliable growth data and undermines the scientific value of the data. In order to improve growth reliability, several factors that may contribute to growth non-reproducibility were identified and thereafter mitigated. Firstly, a simulation study was conducted to achieve insight into temperature and velocity profile of gases inside the reactor since gas flow dynamics can render growth environment near the substrate non-uniform. Interestingly, when argon gas was used as the main carrier gas, natural convective flow emerged, generating flow circulation before the gas reached the substrate placed at the center of the tube reactor. This flow circulation was not favorable for controlled gas introduction. This problem could be resolved by using a more heat- and momentum- conductive gas such as helium.

Secondly, atomic force microscopy of annealed catalyst revealed that the aluminum sub-layer was not thermally stable at the growth temperature although this material has been widely used as a barrier layer to avoid silicide formation of catalyst on silicon substrates. In this respect, aluminum oxide should be a better choice, but under-stoichiometry of the aluminum oxide layer, which originated from sputter target degradation, affected thermal stability of the layer. Reactive sputtering by oxygen addition greatly enhanced thermal stability, and finally defect-free catalyst nanoparticles were formed by thermal annealing.

Thirdly, the effect of the small part-per-million levels of oxygen-containing species on VACNT growth revealed that oxygen-containing gas impurities in nominally pure gas sources have a great influence on growth kinetics in a positive way; their presence increases catalyst lifetime and growth yield. However, the kinetic behavior that is highly sensitive to gas purity is prone to showing an interfering kinetic trend where the real mechanism is masked by the significant gas impurity effect. The stark difference in catalytic lifetime after the introduction of high-performance gas purifiers shows that extremely tight control of the reaction gas composition purity is necessary to obtain controlled growth of CNTs under atmospheric chemical vapor deposition (CVD) conditions. Finally, more reliable growth of VACNTs was achieved, and thereafter the next step for fundamental growth kinetics measurement was followed.

Finally, the CVD system was equipped with an optical micrometer that enables in-situ measurement of the height of growing VACNTs, which have advantageous structure facilitating measurement of growth kinetics since the array height has a robust correlation with growth yield and thereby growth rate. Various ethylene and hydrogen combinations were examined to capture growth kinetics related to different gas environment. The measured initial growth rates were linearly proportional to ethylene concentration, whereas a reciprocal relation was observed with respect to hydrogen concentration. The apparent activation energy was higher than reported in references. Flow rate variation experiments revealed that gas phase reaction is involved as the crucial growth step, which supports the observed high activation energy. Consequently, a growth model was proposed so that it could reasonably fit the initial growth rate data.

Kinetic aspects related to growth deactivation were explored by measuring the final growth height and catalyst lifetime. Unlike growth with unpurified gases, growth became much less sensitive to gas composition after purification. Importantly, it was observed that growth deactivates by deficit of carbon source when relatively low ethylene was introduced. This result is surprising since ethylene pressure should be high enough at the catalyst, considering the calculated sticking coefficient of ethylene is very low, approximately 10^-5. Thus it substantiates the idea that catalyst-mediated gas pretreatment process is critical to sustain nanotube growth. Importantly, this idea challenges the widely accepted growth termination concept whereby nanotube stops growing due to catalyst encapsulation by excessive carbon. Indeed, reduced flow rate of gas mixture increased growth yield remarkably by promoting the gas pretreatment over the catalyst. Catalyst ripening, or steric hindrance by interaction of nanotubes can be an alternative reason for growth termination, but analysis of morphologies of the annealed catalyst and as-grown nanotubes revealed that their effects were not significant for the corresponding growth conditions.