Mechanisms of Aging and Release from Weathered Nanocomposites - Safety of Nanomaterials along Their Lifecycle - page 325

(Figure 14.2c), the finding that strong shear is required for the release of MWCNTs
matches the experience from the formulation of as-produced MWCNTs into aqueous
suspensions, where ultrasound is critical for dispersing MWCNTs.
A closely related approach to facilitate the collection of spontaneous release
from run-off waters by instead performing a UV + immersion induced release was
investigated at CEREGE. They subjected acrylic coatings with 7% nano-CeO 2 to an
adaptation of standard UNE EN 927-6 (realized by Suntest XLS+, UV 300-400 nm
at 65 W/m²), performing rain only four times per week and collecting the run-off
waters for ICP-MS analysis of Ce content (neglecting structural information in a first
step). Alternatively, they irradiated by UV 105 W/cm² (300-400 nm) and immersed
specimen for 1.5 h four times a week (Scifo et al. 2012). The UV apparatus was
identical and the irradiated area of the repeatedly immersed specimens (12 cm²) at
CEREGE was in the same range as used for single immersion after irradiation as
used at BASF (Hirth et al. 2013). After an initial lag time for the immersion protocol,
the Ce amounts released grew faster than for the modified spontaneous protocol, but
were in the exact same range of 350-800 µg/m² (at 400 kJ/m²). Only the Ce content
is known, but if one assumes that the fragments are of the same 7% CeO 2 + poly-
mer composition as the original specimens, this range corresponds to 5-11 mg/m²
total release, which is more than an order of magnitude below the values obtained
for SiO 2 -PA (Figure 14.6) or MWCNT-PP (Figure 14.4), but in the same range as
the values obtained for MWCNT-TPU (Hirth et al. 2013). Related UV + immer-
sion protocols determined by ICP-OES the release of TiO 2 from photocatalytic glass,
finding 25 mg/m² release after 1.1 MJ/m² UV energy (Olabarrieta et al. 2012). The
same setup was later upgraded with mechanical shear by brushing and rolling wheels
(Zorita 2013).
Similar ranges of release are obtained by another variant of induced release,
namely gently spraying water onto vertically held large (71.2 cm²) samples for
10 min between or after irradiations. This protocol was proposed by Nguyen on
the example of nanoSiO 2 -epoxy, and the collected run-off was measured by ICP-
OES (Nguyen 2012). These authors also observed a lag time and then measured a
total 14-56 mg/m² Si release after 800 kJ/m² UV radiation, which one may again
extrapolate to 280-1,123 mg/m² release of assumed SiO 2 /epoxy fragments. For the
same samples, the order of magnitude of release (50 mg/m² Si after 800 kJ/m²) was
confirmed by another complementary approach, which unfortunately is applicable
to SiO 2 composites only. Degraded nanocomposites (shown in Figure 14.3) were
immersed in a 5% HF solution to dissolve the SiO 2 nanofillers that are accessible for
the acid from the surface (Nguyen et al. 2012). The extracted solution was then ana-
lyzed by ICP-OES. Sampling in this protocol was limited to extraction and quanti-
fication, and the characterization of morphology was necessarily lost. The extracted
Si amounts scaled with the 5% or 10% SiO 2 content as expected and confirmed that
more Si (roughly doubled amounts) can be dissolved after UV irradiation as com-
pared to the nonirradiated nanocomposites, from which significant Si was extracted,
too, possibly by diffusion of the acid through percolating nanofiller networks. The
HF protocol for induced release is not a realistic measure of release probability,
but a quick method to estimate for a given nanocomposite the total SiO 2 mass with
vanishing coverage by the polymer matrix. It may serve as worst-case benchmark to