Release from Composites by Mechanical and Thermal Treatment - Safety of Nanomaterials along Their Lifecycle - page 252

of the lathe. The test materials with 0-4 mass% CNTs (1% gradation) were fed by a
carriage, which was driven towards the sanding disk with a speed of 0.0025 cm/s by
a turning screw beneath the lathe. Three levels of sanding paper roughness were used
(P80 with 201 µm grit size, P150 with 100 µm grit size, and P320 with 46.1 µm grit
size) with three disk sander rotation rates (586; 1,425; and 2,167 rpm).
The results showed that the highest particle number concentrations (4,670 #/cm³,
mainly <100 nm diameter) were produced with coarse sanding paper, 2% CNT,
and medium disk sander speed. The lowest concentration (92 #/cm³) was produced
with medium sanding paper, 2% CNT, and slow disk sander speed. No free CNTs
were observed by electron microscopy in airborne samples, except for tests with 4%
CNTs. Also, heating and thermal decomposition were observed due to the friction
between the samples and the sanding paper leading to the formation of particles
<50 nm diameter.
Hirth et al. (2013) used the same experimental setup as Huang et al. (2012). The
test samples were prepared by mixing 2 mass% multiwalled CNTs with epoxy resin
or neat epoxy resin. The automated sanding simulation system was equipped with
sanding paper (68 µm grit size). Other sanding parameters were not given. Hirth
et al. (2013) found that mechanically released fragments in some cases show tubular
protrusions on their surface. With the help of chemically resolved microscopy and
a suitable preparation protocol, they identified these protrusions unambiguously as
naked CNTs. Size-selective quantification of the fragments revealed that as a lower
limit at least 95% of the CNTs remain embedded.
Koponen et al. (2011) investigated the particle size distribution and the total
number of dust particles generated during sanding of paints, lacquers, and fillers
doped with nano-objects as compared to their conventional counterparts. The sand-
ing experiments were conducted inside a human exposure chamber of 20.6 m³ with
a HEPA filtered air supply and an air exchange rate of 9.2 per hour. A commercial
handheld orbital sander (Metabo Model FSR 200 Intec) with sanding paper P240
(58.5 µm grit size) was used as the sander. The sander has an internal fan for dust
removal and normally a filter bag attached to the exhaust air. Instead, for this study
the outlet was connected with a flexible tube to the aerosol sampling chamber. The
measurement devices used were an aerodynamic particle sizer (APS), FMPS, and
a modified commercial electrostatic precipitator (ESP) for particle sampling. All
sanding experiments were performed by the same person using the same protocol.
The test objects were 13 paints, lacquers, and fillers with nano-objects (TiO 2 , SiO 2 ,
kaolinite, carbon black, and perlite) applied on wooden plates produced by several
manufacturers and, additionally, reference plates without nano-objects for each
product type.
The investigation results show that the particle number size distribution of sand-
ing dust (corrected for the emissions from the sanding machine) was dominated
by 100-300 nm sized particles, whereas the mass spectra were dominated by
micrometer particles. The added nano-objects had no or only very little influence
on the particle modes in the sanding dust when compared to the reference products.
Nevertheless, Koponen et al. (2011) observed considerable differences in the release
rates as expressed in number concentrations of the different size modes. Still no
significant link to the investigated nano-objects could be made.