Airborne Nanoparticle Release Associated with the Compounding of Nanocomposites using Nanoalumina as Fillers

Twin screw extrusion is the preferred process to commercially produce nanocomposites by compounding the nanoparticles and polymer melts. Polymer nanocomposites, which contain nanoparticles dispersed in a polymer matrix, provide improved properties at low filler loadings. Nanoalumina particles recently have been used as fillers to polymer matrix that contributed enhanced physical properties of nanocomposites. Recently, concerns had been expressed that airborne nanoparticles particularly of nanoalumina released during compounding might present serious contamination of the air in the workplace. Researchers with experience in environmental health and polymer manufacturing monitored the compounding process for a model nanoalumina-containing nanocomposite using a TSI Fast Mobility Particle Spectrometer (FMPS). FMPS measurements were taken at background locations, source locations, and operators’ breathing zones; in parallel to the FMPS real time measurement, airborne nanoparticles were collected using polycarbonate filters fitted with filmed grids driven by a personal air sampling pump. Filter samples were analyzed for particle morphology and elemental composition. It was found that the nanoparticle number concentration was elevated during processing. The released nanoparticles are a complex mixture of the individual nanoalumina particles, agglomerates of those particles, polymer fume particles, and perhaps others.


INTRODUCTION
second inorganic or organic phase; traditionally, micrometer-sized particles have been used as the filler.In polymer nanocomposites these fillers have at least one dimension less than 100 nm (Kojima et al., 1993).These fillers include alumina, carbon black, silica, talc, calcium carbonate, layered Polymers are often reinforced using a Corresponding author.Tel.: +1-978-934-4366; Fax: +1-978-934-3050 E-mail address: candace.umass@gmail.comsilicates (nanoclays), and recently, silver and engineered nanoparticles such as carbon nanotubes.Although the nanometer-sized particles allow low filler loadings (< 10% w ) in nanocomposites with a retention of flexibility and impact properties, the resulting nanocompound's properties are highly dependent on dispersion of the primary filler particles through the polymer matrix.With good dispersion, each particle is wetted completely by the melted polymer, creating a very high interfacial surface area that can improve properties of the polymer (McCarrie and Winter, 2003).
Since commercial compounding (mixing) of nanocomposites is typically achieved by feeding the nanoparticles and polymer into a twin screw extruder, the airborne particles associated with nanoparticle reinforcing agents are of particular concern, as they can readily enter the body through inhalation.Recent research has suggested that nanometer-sized particles of many materials, including nanoalumina, display greater toxicity than for larger particles, and aggregated nanoparticles can be deaggregated in the lung after inhalation (Ferin et al., 1991;Wolff et al., 1988;Zhang et al., 2000;Renwick et al., 2004;Warheit, 2004).
In addition, Maynard et al. (2005)   During these experiments, the hood was positioned 30 cm (12 in.) above the extruder to collect polymer fumes given off from the melted mixture of polymer and nanoalumina; it was not placed near the feeding port.

Twinscrew feeder Feeding throat
A typical operation of compounding process includes five time periods as illustrated in Fig. 2  operations of experiments.This compounding process was operated using twin screw feeder feeding in the primary feeding port (the first port).
(3 in.) from the feeding port.Source concentration also was only measured during 5% nanoalumina feeding.
Particle number concentrations normalized for channel width were calculated in each of the 32 channels for each phase of the measurement time period.Background concentration measured before feeding nanoalumina into the twin screw extruder was used as the baseline for subtraction from the various concentrations measured during processing.
The magnitude of nanoparticle exposure associated with nanoalumina compounding was measured before, during and after feeding nanoalumina particles to the compounding process.Particle size for each measurement period was evaluated by determining the particle size distribution, geometric mean, and total concentration, which were calculated by the FMPS software.

Particle sampling and characterization
A new nanoparticle aerosol filter sampler was developed and used in these experiments.
A schematic layout of the sampling setup is shown in Fig. 3. TEM-copper grids (400 mesh with a titanium dioxide film) were taped on 47 mm diameter polycarbonate membrane filters (0.2 μm pore size).

Changes in total particle concentration and median size
The results of one typical processing Phase III represents the first step of the compounding process, when the virgin PMMA polymer was loaded into the extruder.
Concentrations were measured at room background, breathing zone and source location.
Particle median sizes dropped to close to 30 nm.
This was likely due to the absence of any agglomerated nanoalumina being fed and the formation of very small polymer fume particles when the polymer pellets were dropped into the heated extruder through the feeding throat.
During Phases IV and V, nanoalumina particles were loaded at 2% and 5% weight of polymer (PMMA) respectively in parallel with loading PMMA pellets.The peak total concentrations at the source were on the order of 10 6 particles/cm 3 , and the particle median size increased again due to the feeding of nanoalumina particles.The particle median diameter rose above 70 nm when more nanoalumina agglomerates were detected in periods IV and V.     from compounding process is a fugitive process and varies considerably from second to second.

Mobility particle size distribution
The unique feature of measurement by FMPS, with its one-second sampling time, is to present the real time change of particle concentration and its size distribution.Using this characteristic, the instantaneous release of particular nanoparticles and the magnitude of release could be identified by the reading in one second.This is an advantage in using the FMPS to investigate the fugitive release of nanoparticles.For the measurement at background before warming up the TSE, the particle geometric mean size was 38 nm, the mode was 52 nm and the particle total concentration was 1.8 x 10 4 particles/cm 3 (Table 1).
As shown in Fig. 4(a), curve (I), the particle concentration was 2.2 x 10 4 particles/cm 3 at the peak particle size of 52 nm at the time before warming up the extruder.This particle size distribution and concentration represents the background airborne nanoparticles that existed in the laboratory.

2) Background particle concentration after warm up and before feeding nanoalumina
When measuring the background location after warming up the TSE, the particle geometric mean became 29 nm and the mode was 34 nm, as shown in Table 1.These diameters are smaller than those found before warming up the extruder.The count median diameter (CMD) (or geometric mean) (Hinds, 1999) of line (II) shown in Fig. 4(b) calculated based on the exposure data of curve (II) in Fig. 4(a) is consistent.The particle total concentration after warming up was 3.7 x 10 5 particles/cm 3 which is 20 fold higher than the total concentration of 1.8 x 10 4 particles/cm 3 before warming up the extruder.As shown in Fig. 4(a), curve (II), the particle concentration was 6.4 x 10 5 particles/cm 3 at the peak particle size of 34 nm after warming up the extruder which is 30 times the peak concentration before warming up the extruder.
The dramatic increase in the number of smaller nanoparticles likely was caused by the release of polymer fume and nanoparticle residues from the TSE by heating the extruder.This particle size distribution and concentration curve represents the background airborne nanoparticles that existed in the laboratory after heating the extruder but prior to feeding the nanoalumina particles and polymer.

3) Background particle concentration during the feeding of nanoalumina
For the measurement at background during feeding nanoalumina particles into the TSE, the particle geometric mean became 54 nm and the mode was 60 nm (Table 1), which are almost double the particle sizes after warming up the TSE.The concentration at the mode as seen in Fig. 4(a), curve (III) was 4.8 x 10 5 particles/cm 3 ; the mode was shifted to a larger value during the feeding of nanoalumina particles.In Fig. particles/cm 3 as shown in Table 1 which is slightly higher than the total concentration of 3.7 x 10 5 particles/cm 3 after warming up the

4) Background particle concentration after feeding nanoalumina
For the measurement at the background location after feeding nanoalumina particles into the TSE, the particle geometric mean was 54 nm and the mode was 52 nm (Table 1), which are very similar to the values during feeding nanoalumina to the TSE.The concentration at the mode as seen in Fig. 4(a), curve (IV) was 5.6 x 10 5 particles/cm 3 which was reduced below the concentration after warming up the TSE.However, the count median diameters as seen in Fig. 4 (III, IV) do not have noticeable difference between lines of during and after feeding Al 2 O 3 .The particle total concentration after feeding nanoalumina became 4.6 x 10 5 particles/cm 3 (Table 1), which is the highest concentration measured at the background location.Released

Particle concentration and size distribution at source
The measurement at the source location during feeding nanoalumina particles into the TSE are summarized in Table 2 and Figs.5(a), curve (I).Here, the particle geometric mean was 44 nm and the mode was 190 nm (Table 2).
The total concentration at the source location was 1.3 x 10 6 particles/cm 3 (Table 2), the highest concentration measured during the experiment at any time or location, which is an order higher than the highest measurement at the background location.The concentration at the mode as seen in Figs.5(a), curve (I) is 1.3 x 10 6 particles/cm 3 which is a large peak found only in the measurement at the source location.X-axis: Log scale particle diameter in nanometer, it's based on 32 channels, each channel is a range size of particles being collected.For example: channel 3.5 represents the average of particle size range of 0-7 nm Y-axis: Normalized particle number concentration.
Note: Normalized concentration is defined as the concentration of particles in a size bin divided by the width of this bin.
If the ith bin has N i particle concentration, thus normalized concentration in the ith bin is n Ni = N i / D i where  D i is the width of the ith bin.For example: D 1 is 7nm Fig. 6.Average particle number concentration and size distribution change at the room background.The large increase in the concentration of nanoparticles at the source was on the order of 10 5 particles/cm 3 or greater, and the concentration of nanoparticles less than 30 nm increased by up to 3 x 10 5 particles/cm 3 at the feeding port.That more agglomerated nanoalumina particles were measured at the source location is indicated by the curve peaking at 200 nm shown in Figs.5(a) and 7.
The larger agglomerated nanoalumina particles escaping from the feeding throat contributed to the large mode of 191 nm measured at the source location.

Particle concentration and size distribution at the breathing zone
The instantaneous data of measurements at the breathing zone location during feeding nanoalumina particles into the TSE are shown in Table 2 and Fig. 5(a), curve (II); here, the particle geometric mean became 42 nm and the mode was 45 nm (Table 2) which are less than the geometric mean (54 nm) and the mode (60 nm) at the background location.The total concentration at the breathing zone location was 2.8 x 10 5 particles/cm 3 (Table 2), which is lower than the concentration of 4.0 x 10 5 particles/cm 3 at the background location during

CONCLUSIONS AND RECOMMENDATIONS
1 μm This study demonstrates conclusively that the xxx Tsai et al., Aerosol and Air Quality Research, Vol. x, No. x, pp. xxx-xxx, 2008 xxx Fig. 1.(a) Layout of twin screw extruder and measurement locations.

Fig. 2 .Fig. 2 .
Fig. 2. (a) Illustration of the time sequence during compounding and the timing of particle measurement.Roman numerals indicate the phases of a typical experiment; numbers indicate sampling times.
experiment performed on March 22, 2007 are presented and discussed here; in addition, two more experiments performed in November 2006 and February 2007 are presented for comparison of measurements at the source.The temporal patterns of total number concentration and median diameter are presented in Fig. 2(b).This long-term monitoring shows total number concentration measured from the beginning of warm up to the end of the operation; background concentrations were measured in the sequence shown in Fig. 2(a) while source and breathing zone concentrations were only measured during the feeding of 5% nanoalumina.During phase I, the TSE was warmed up from room temperature to above 200°C, and concentrations were measured at a central location of room background about 55 cm (22 in) from the TSE both before heating .The laboratory door and windows were closed and the local ventilation system was on, and there were no nanoalumina particles introduced into this room during the warm up period.As shown in Phase I of Fig. 2(b), particle median size decreased from 50 nm to values less than 30 nm after 20 min of heating, due to a large quantity of complex mixtures of nano sized fumes released from the heated extruder.During Phase II, the calibration period, the twin screw feeder of nanoalumina particles placed nearby the primary feeding port of extruder was calibrated for the feeding rate by loading nanoalumina particles into a cup.Free nanoalumina particles were introduced and released into the room and particle concentrations were measured at room background and breathing zone locations.Peaks shown in the black curve of particle total concentration seen in Fig. 2(b) were agglomerated nanoalumina particles released during the compounding process.Meanwhile, the particle median size shown by the light gray curve of Fig. 2(b) gradually increased to above 40 nm at the end of Phase II due to the larger size of the agglomerated nanoalumina particles that were released.
Mobility particle size distributions measured at various locations and at various times as listed in Tables 1, 2, and the cumulative count distribution in Fig. 4(b) are discussed in this section.For each measurement location and time, data concerning particle size are shown in the tables, and the particle concentration as a function of size is shown in Figs.4(a) and 5(a).The differences in particle concentration and size distribution at various locations and times can be identified and compared from these tables and figures.Particle concentrations and the mobility size distributions were measured at the laboratory background location at four different times, i.e., before warming up the machine, after warming up the machine, during feeding nanoalumina particles, and after feeding nanoalumina particles.1) Background particle concentration and size distribution before warm-up Particle concentration and the mobility size distribution measured before warming up the machine represent the original particles present in the laboratory air which was used as the baseline for comparison to any subsequent particle release activities.The statistical data at each second of measurement were automatically calculated by the FMPS software, and the data for one second during a stable measurement period were selected to represent the background measurement for discussion in this section.The same one second of data was selected for use in all of the tables and figures in this section.Release of airborne nanoparticle

Fig. 4 .
Fig. 4. (a) Instantaneous background concentration and size distribution at four different time periods.

Fig. 5 .
Fig. 5. Concentration and size distribution during feeding 5% nanoalumina (a) Instantaneous measurement at source and breathing zone; (b) 3D number concentration at source; (c) Instantaneous measurement at source of 3 experiments.
extruder and before feeding nanoalumina.The background concentration was dramatically increased after warming up the TSE, and the concentration remained at a high level throughout the feeding of nanoalumina particles into the TSE.In other words, the background concentration was not affected as much by the feeding of nanoalumina particles, as it was by heating the TSE.The likely reason is that the measurement location for the background concentration is 55 cm behind the extruder which is farther from the feed port than the other measurement locations.The nanoparticles released from the TSE were diffused in threedimensional space and thus many fewer nanoparticles were carried out to the background measurement location; also, more agglomeration could have occurred during the transport of nanoparticles to the background location, which could have formed agglomerated particles having sizes beyond the measurement limit of the FMPS (560 nm).
Fig. 6 shows the average data measured at the background location before and after feeding nanoalumina particles into the TSE.These curves represent the average data shown in Figs.4(a), curves (II) and (III).The concentration measured before feeding nanoalumina is the same time as the measurement after warm up.The increased nanoparticle concentration at background through feeding nanoalumina is the difference of Fig. 4(a), curves (II) and (III) and the average data were shown in the dotted line in Fig. 6.Obviously, the concentration at background was affected by the process of The dynamic nature of the particle release is shown in Fig.5(b), which plots the second-bysecond change in the particle size distribution at the source during one location.On the other hand, the peak concentrations and distribution measured during different experiments were quite similar, as shown in Fig. 5(c), which plots the source distributions from three experiments performed under identical conditions but in different months.The particle size distribution at the source location shows multiple peaks that have a different pattern compared to concentrations measured at other locations.The average data measured at the source location before and during feeding nanoalumina particles into the TSE is shown in Fig. 7.These curves of "BG after warm up" and "source during feeding" represent the average data shown in Figs.4(a), curve (II) and 5(a), curve (I), respectively.The increased nanoparticle concentration at the source while feeding nanoalumina is the difference between Figs. 4(a), curve (II) and Fig. 5(a), curve (I) and the average increase in nanoparticle concentration is shown by the dotted line in Fig. 7. Released nanoparticles during feeding nanoalumina particles into the TSE at the source location caused a high concentration increase at both particle size ranges of less than 30 nm and above 50 nm.

Fig. 7 .
Fig.7.Average particle number concentration and size distribution change by feeding 5% nanoalumina at the source (primary feeding port).

Fig. 8 .
Fig. 8. Average particle number concentration and size distribution change at the breathing zone by feeding 5% nanoalumina.

Fig. 9 .
Fig. 9. SEM image of nanoparticles collected at the source.
compounding of polymers and nanoalumina in a TSE can release large quantities of nanoparticles into the air.Large quantities of nanoparticles were released during the extruder heating phase, while feeding polymer pellets only, and feeding the nanoalumina/polymer mixture.Nanoparticles released in the heating and polymer feed phases are likely to be polymer fume.When the nanoalumina particles were used instead of micrometer-size particles as fillers, the nanoparticles released during nanocompounding are a complex mixture of the individual nanoalumina particles, agglomerates of those particles, polymer fume particles, and perhaps others.Elevated nanoparticle levels were measured at the source, the room background, and the operators' breathing zone.necessary to use both a particle size-measuring instrument (in this case, the FMPS) and a particle collection and analysis system.Further work is necessary in two areas.(b) Nanoalumina particles Fig. 10.SEM images of nanoparticles collected at the breathing zone.

First
feed method, feed rate, temperature, air flow pattern, etc., must be systematically investigated.Second, the optimum sampling method to completely characterize complex nanoparticle mixtures must undergo further study and optimization.

Table 1 .
Particle size statistical data at background.

Table 2 .
Particle size statistical data during feeding 5% nanoalumina at source and breathing zone