The 12 nanotube and reference fiber samples assembled for this study behaved quite differently upon long-term exposure to an oxidizing phagolysosomal simulant fluid. Crocidolite asbestos, a positive control for biodurability, showed no apparent changes after 90 days, as expected. The negative control, wollastonite, was completely dissolved at the point of first inspection (30 days), consistent with its low biopersistence reported by Warheit et al. and Macdonald et al.[25
The carbon nanotube samples showed considerable variation depending on type and functionalization. Most nanotube samples showed little change when examined by TEM, as illustrated in . This unfunctionalized sample consisted of a nearly continuous network of entangled SWCNT ropes that was unaffected by 90 day mild agitation in the oxidizing, low-pH biological stimulant fluid (see ). Other SWCNT samples showed subtle or modest changes in aggregation state reflecting gradual disentanglement of the ropes, especially the hydrophilic tube samples functionalized with aryl-sulfonate or ozone, but no evidence of length reduction by tube cutting.
Figure 1 Persistence and morphological stability of unfunctionalized SWCNTs upon 90 day exposure to oxidative phagolysosomal simulant. Typical TEM images before (a) and after (b) 90 day exposure, showing no detectable change. (c) Effective tube/aggregate sizes (more ...)
shows that carboxylated SWCNTs exhibit a very different behavior. They undergo extensive breakup of the entangled rope networks and liberation of short tubes or tube bundles that become progressively shorter by chemical attack on the tubular graphene during incubation. SEM images provide more information on the morphological evolution of carboxylated SWCNTs. For SEM analysis the CNTs were dispersed from ethanol suspensions onto silicon substrates by bar-coating (see example, ). panel b and c show nanotube length distributions determined by manual measurement and counting of about three hundred individual tubes identified on SEM micrographs for the SWCNT-COOH samples. It is evident that shorter nanotubes are significantly more abundant after 90-day exposure to the physiological oxidizing environment (c) relative to the original SWCNTs (b). Histograms derived from TEM images also show tube shortening after 90 day exposure (data not shown).
Figure 2 Degradation and morphological instability of SWCNT-COOH upon 90 day exposure to oxidative phagolysosomal simulant. Typical TEM images before exposure (a), after 1 day (b), 7 day (c), 30 day (d), 60 day (e) and 90 day exposure (f). After 90 days the loss (more ...)
Figure 3 Nanotube length distribution measurements for carboxylated SWCNTs before and after long-term exposure to physiological oxidizing fluids. (a) Example SEM image of fresh tubes distributed on silicon substrates by bar coating; (b, c) Nanotube length distributions (more ...)
Dynamic light scattering is useful as a qualitative tool to track changes in size and aggregation state in situ – without the need for drying or substrate deposition. The annealed, low-functional-group nanotubes in show a bimodal size distribution, which we interpret as the presence of continuous entangled rope structures (the 3–8 um large aggregate peak) coexisting with primary tubes/bundles any equi-axed particulate carbon by-products (the 50 – 800 nm peak), which are free in suspension and not connected to the large entangled rope networks. For most of the nanotube samples studied, the bimodal distribution shows few or modest changes (see ), whereas the carboxylated tubes show highly significant changes in the distribution of both peak heights and locations (). Over time, the SWCNT-COOH large-aggregate peak loses intensity and by 90 days disappears, reflecting the near total loss of the entangled rope networks. In its place the free primary peak grows, and the peak shifts to progressively smaller sizes (). Because DLS estimates hydrodynamic size, it is sensitive to changes in aggregation, length reduction, and the breakup of the primary bundles or ropes, which reduces their apparent diameter. The migration of the peak location thus provides a crude measure of the extent of primary tube debundling in combination with shortening (), while the decrease in large-aggregate peak height provides a measure of disentanglement or large network structures (). DLS size analysis relies on spherical geometry, so cannot be used to provide the primary data on length reduction in this study. That evidence comes from electron microscopy rather ( and ), but DLS is nevertheless useful to track the physical evolution of the samples in situ.
Figure 4 Hydrodynamic size distributions by dynamic light scattering (a) show the progressive loss of the large aggregate peak and the growth of the primary tube peak over the course of 90 days. (b) quantifies the disappearance of the large aggregate peak which (more ...)
Carboxylation is typically carried out by treatment with oxidizing acids, HNO3
, alone or in mixtures, or in combination with peroxide [27
]. In this study, similar behavior was observed with the commercial sample (carboxylated with HNO3
) and samples treated in-house with HNO3
for 1 or 3 hrs (see ).
Control experiments were carried out by incubating carboxylated SWCNTs in DI H2O, only PSF medium, or only 1mM H2O2 with mild agitatition for 90 days. As shown in , there is no evident tube shortening or degradation following incubation in either water () or PSF medium (), whereas long, entangled carboxylated tubes subjected to 90 days treatment degraded into short tubes or debris after exposure to 1mM H2O2 () or 1mM H2O2 plus PSF (), suggesting that continuous exposure to an oxidizing environment is important for further attack and degradation of SWCNTs-COOH.
Figure 5 Morphologies of carboxylated SWCNTs after 90 day exposure to different media: (a) DI H2O; (b) PSF only; (c) 1mM H2O2 only and (d) 1mM H2O2 plus PSF. Tube degradation occurs only under oxidizing conditions, but does not necessarily require the low pH of (more ...)
The large changes in morphology for the SWCNT-COOH samples were accompanied by accumulation of solid particulate debris (). Note that all of the samples contain some equi-axed carbon solids, which in the fresh samples include graphenic carbon shells originally templated on catalyst particles and often found at tube tips. After 90 day exposure of the SWCNT-COOH samples, however, one sees accumulation of additional globular material that appears on tube walls and dispersed throughout the sample in dried SEM specimens as shown in . This debris suggests a chemical degradation process rather than a simple tube cutting and disentanglement by the mild mechanical forces generated by slow sample rotation.
SEM image showing equi-axed debris or “dots” (see arrows) on silicon substrate after 90 day exposure of carboxylated SWCNTs. The dots, which are believed to carbonaceous debris, can also be seen by close inspection of TEM images.
To test the hypothesis of chemical degradation, we examined the UV-visible spectral characteristics of the clear filtrates following removing of the nanotubes using centrifugal ultrafiltration with a 3000 NMWL (ca 2nm) pore cellulose membrane. shows that the ultra-filtrates of the 90-day exposed SWCNT-COOH samples are strongly fluorescent under 365 nm illumination. Neither the simulant fluid itself, nor the ultrafiltrates from fresh nanotube suspension were fluorescent, indicating that the fine (< 2 nm) fluorescent products appear during the 90 day treatment. also shows the excitation and fluorescence spectra, centered at 300 nm and 405 nm, respectively. Other SWCNT types (unfunctionalized, ozone treated, and aryl-sulfonated) exposed to phagolysosomal stimulant for 90 days do not exhibit strong fluorescence in their ultrafiltrates (), indicating again that the SWCNT-COOH sample is unique in its ability to undergo chemical degradation under these conditions. We suspect that the fluorescent degradation products are primarily carbonaceous nanoparticles (< 2nm), which have been reported to exhibit fluorescence [30
]. The excitation (250–300 nm) and fluorescence (400–450 nm) spectra are similar to those reported for graphenic material with polyaromatic substructures [33
Figure 7 Optical properties of nanotube ultra-filtrates showing evidence of < 2nm chemical degradation products in the SWCNT-COOH sample after 90-day exposure to the oxidative phagolysosomal simulant. (a) digital photograph under UV illumination showing (more ...)