Soft and conductive nanomaterials like carbon nanotubes, graphene, and nanowire scaffolds have expanded the family of ultraflexible microelectrodes that can bend and flex with the natural movement of the brain, reduce the inflammatory response, and improve the stability of long-term neural recordings. However, current methods to implant these highly flexible electrodes rely on temporary stiffening agents that temporarily increase the electrode size and stiffness thus aggravating neural damage during implantation, which can lead to cell loss and glial activation that persists even after the stiffening agents are removed or dissolve. A method to deliver thin, ultraflexible electrodes deep into neural tissue without increasing the stiffness or size of the electrodes will enable minimally invasive electrical recordings from within the brain. Here we show that specially designed microfluidic devices can apply a tension force to ultraflexible electrodes that prevents buckling without increasing the thickness or stiffness of the electrode during implantation. Additionally, these "fluidic microdrives" allow us to precisely actuate the electrode position with micron-scale accuracy. To demonstrate the efficacy of our fluidic microdrives, we used them to actuate highly flexible carbon nanotube fiber (CNTf) microelectrodes for electrophysiology. We used this approach in three proof-of-concept experiments. First, we recorded compound action potentials in a soft model organism, the small cnidarian Hydra. Second, we targeted electrodes precisely to the thalamic reticular nucleus in brain slices and recorded spontaneous and optogenetically evoked extracellular action potentials. Finally, we inserted electrodes more than 4 mm deep into the brain of rats and detected spontaneous individual unit activity in both cortical and subcortical regions. Compared to syringe injection, fluidic microdrives do not penetrate the brain and prevent changes in intracranial pressure by diverting fluid away from the implantation site during insertion and actuation. Overall, the fluidic microdrive technology provides a robust new method to implant and actuate ultraflexible neural electrodes.

Boron nitride nanotubes (BNNT) are poised to fill an electrically insulating, high-temperature, highstrength niche. Despite significant progress over the past two decades, BNNTs are not yet synthesized in high enough quantity and quality to permit their use in engineering applications. The next necessary step to make BNNTs accessible for research and applications is to improve the availability of high-quality BNNTs. Here, we present a scalable bulk purification technique that yields high-purity BNNTs. Bulk synthesized material is introduced to a wet oxygen environment at elevated temperatures to remove elemental boron and hexagonal boron nitride impurities with a final yield of purified BNNTs near 10 wt %. This process shows full removal of impurities, as observed by scanning electron microscopy (SEM), cryogenic transmission electron microscopy (TEM), and high-resolution TEM. X-ray photoelectron spectroscopy (XPS) and infrared spectroscopy show minimal BNNT functionalization, while high-resolution TEM shows damage to large-diameter BNNTs.

Boron nitride nanotubes (BNNTs) are of interest for their unique combination of high tensile strength, high electrical resistivity, high neutron cross section, and low reactivity. The fastest route to employing these properties in composites and macroscopic articles is through solution processing. However, dispersing BNNTs without functionalization or use of a surfactant is challenging. We show here by cryogenic transmission electron microscopy that BNNTs spontaneously dissolve in chlorosulfonic acid as disentangled individual molecules. Electron energy loss spectroscopy of BNNTs dried from the solution confirms preservation of the sp hybridization for boron and nitrogen, eliminating the possibility of BNNT functionalization or damage. The length and diameter of the BNNTs was statistically calculated to be ∼4.5 μm and ∼4 nm, respectively. Interestingly, bent or otherwise damaged BNNTs are filled by chlorosulfonic acid. Additionally, nanometer-sized synthesis byproducts, including boron nitride clusters, isolated single and multilayer hexagonal boron nitride, and boron particles, were identified. Dissolution in superacid provides a route for solution processing BNNTs without altering their chemical structure.

Due to recent advances in high-throughput synthesis, research on boron nitride nanotubes (BNNTs) is moving toward applications. One future goal is the assembly of macroscopic articles of high-aspect-ratio, pristine BNNTs. However, these articles are presently unattainable because of insufficient purification and fabrication methods. We introduce a solution process for extracting BNNTs from synthesis impurities without sonication or the use of surfactants and proceed to convert the extracted BNNTs into thin films. The solution process can also be used to convert as-synthesized material-which contains significant amounts of hexagonal boron nitride ( h-BN)-into mats and aerogels with controllable structure and dimension. The solution extraction method, combined with further advances in synthesis and purification, contributes to the development of all-BNNT macroscopic articles, such as fibers and 3-D structures.

Carbon nanotube (CNT) fibers are a promising material for wearable electronics and biomedical applications due to their combined flexibility and electrical conductivity. To engineer the bending properties for such applications requires understanding how the bending stiffness of CNT fibers scales with CNT length and fiber diameter. We measure bending stiffness with a cantilever setup interpreted within Euler Elastica theory. We find that the bending stiffness scales with a power law of 1.9 for the fiber diameter and 1.6 for the CNT length. The diameter scaling exponent for fiber diameter agrees with results from earlier experiments and theory for microscopic CNT bundles. We develop a simple model which predicts the experimentally observed scaling exponents within statistical significance.