Nanographene was produced in one of two proprietary processes shown in FIGURE 1. Although the equipment was similar, Process A was used for larger batches than Process B. For both processes, the final products were deposited in two stainless steel containers with larger particles being collected in container 1 and smaller particles in container 2. The collection container was then unscrewed from its bayonet mount and carried to a weigh-out booth. To prepare for further processing, workers used a scoop to transfer the materials from the collection container to a different container on a scale for weighing. Following production, the inside walls of process vessels were coated with residual materials and periodically required cleaning. To clean the process tanks, the operator opened the access hatch and used a hand tool to scrape accumulated powder from the walls, which created obvious dust exposure.

Process B had two design features that allowed for better containment of product and reduced emissions to the work environment, specifically: 1)the blower was located downstream of both product recovery vessels, and; 2) butterfly valves were incorporated on the upstream of the collection vessels. Since the fan was downstream of the process, all components of the process were under negative pressure during operation, minimizing the potential for system leakage to the work environment. In addition, the inclusion of butterfly valves on the bottom of the product recovery vessels allowed for system isolation during product harvest. Split butterfly valves have been widely used in the pharmaceutical industry to minimize particle emissions during transferring products or materials from one process vessel to another. These valves may be closed so that process leakage does not occur when collection containers are removed during product harvesting.

At Site A, the following task-based exposures and process evaluation were conducted to assess the impact of engineer controls and process changes on exposure mitigation:

This study site was a small business whose main product was a thin, paper-like membrane coated with CNTs. These materials were combined to produce a nanocomposite paper. By optimizing the formulation, properties of interest such as mechanical strength, conductivity, handle-ability, thickness, weight, and other properties can also be customized to meet end-user requirements.

Dry powders were weighed out into a 2.5 liter beaker in a ventilated enclosure located inside a downflow room (FIGURE 2). During weighing, the enclosure ventilation was momentarily turned off, because air movement affected the accuracy of the scale. Solid ingredients in plastic bags or bottles were set in the enclosure and opened. A scoop was used to transfer the specified mass of various ingredients to the beaker. The total volume of solid nanomaterials transferred to the beaker was about 1 liter. Following weigh-out of materials, a solvent was added to the dry materials for initial mixing. The beaker was then taken out of the enclosure and more solvent was added to further mix the material, forming a slurry. The incorporation of nanomaterials into a slurry should suppress dust emissions during subsequent handling activities. Finally, the beaker was transferred to the production room where the customized sheets of membrane are produced.

A ventilated enclosure (Xpert Filtered Balance System, Labconco Co., Kansas City, Missouri) was used to weigh-out material on a perforated bench top. The face of the enclosure had an opening of 8 inches by 34 inches (20.3 cm by 86.4 cm) with a moveable front sash that allows for the moving of equipment into or out of the enclosure. This unit included a fan that exhausted air through a baffle plate in the back of the enclosure and discharged air through an ultra-low penetration air (ULPA) filter towards the ceiling of the room returning air to the lab. The ceiling panel directly over the position where the worker stands to perform the task is a light fixture; there is no ceiling air flow over the worker. Enclosure face velocities were measured using a hot wire anemometer, and air flow patterns were visualized with smoke tracers. Airborne concentrations were measured near the source during weigh-out of the dry nanomaterials and the mixing of the slurry using the DustTrak aerosol photometer and the FMPS. In addition, background concentrations both inside and outside of the down flow room were measured to assess non-process derived particulates.

The preparation of the slurry for the process was conducted inside a down flow room. Unidirectional flow booths, or down flow booths are commonly used in pharmaceutical applications for large-scale powder packing, process loading, and tray dryer loading ⁽²⁷⁾. In general, these booths supply air from overhead over the full depth of the booth. Particles generated by processes carried out in the booths are captured and carried to the exhaust registers, which are located along the back wall of the booth. The down flow room in this facility was 14 feet (4.3 m) wide, 28 feet long (8.5 m), and 9 feet (2.7 m) high. The majority of air in the room was recirculated by 36 2-ft (0.6 m) x 4-ft (1.2 m) fan-powered HEPA filters (Model No. SAM 24 MS GS, Clean Rooms International) located in the ceiling. Except for the air exhausted by the fume hood and the ventilated enclosure, the remaining air returned to these fan-powered HEPA filters through a bank of seven filters that covered the inlets to the return air plenum. There were no provisions for make-up air; leakage into the return air plenum and around the room’s door was likely. During this study, air flow from the fan-powered HEPA filters was measured using an air capture hood (model 8371, TSI Inc.) where possible. Because access to some of the fan-powered HEPA filters was obstructed (e.g. sprinkler heads for fire suppression), a hot wire anemometer was used to measure some filter face velocities and estimate system flow.

The small company produced vertically aligned CNTs for use in electronics. These CNTs can be processed into a fiber or thread that is highly electrically conductive, flexible, bendable, fatigue-resistant, and load-bearing for multifunctional applications.

Aligned CNTs were manufactured in an EasyTube^™ 3000 reactor (First Nano, Ronkonkoma, NY). This system is a customizable chemical vapor deposition/annealing process tool for nanomaterials synthesis, thin film depositions and anneals. The unit included digital control of production and maintenance parameters. The only human interaction during the operation of this reactor occurred during the loading and unloading of the reactor. To load the reactor, the access door was opened to allow the worker to mount the collection substrates on a holder. The reactor was sealed after the holder moved into the reactor. Then, gases such as argon, helium, methane, ethane, ethylene, acetylene, and hydrogen flowed into the reactor and, in the presence of a proprietary catalyst, formed aligned CNTs on the substrate. At the end of the process, the reactor was opened, and the collection substrate was moved into the loading/unloading port. The substrate was then placed in a container for shipping.

The reactor was comprised of four cabinets, including: the control cabinet, load compartment, reactor and burn box cabinet (see Figure 3). Each cabinet was maintained under negative pressure with respect to the room to prevent leakage.

The function of each cabinet is listed below:

This system uses a hierarchical process control scheme to prevent process leakage:

Cabinet inlet air velocities and static pressures were measured using a hot wire anemometer in front of the slots of loading/unloading modules and the inlet port of the burn box. During routine production operations, the DustTrak and FMPS monitored the area concentrations in front of the reactor. Worker exposures were monitored when CNTs were unloaded from the reactor.

The difference between similar Processes A and B provide us a good case for exploring the effects of engineering controls on reducing worker aerosol exposure during product harvest. FIGURE 4 shows dust exposures measured at the source with the DustTrak and FMPS when collection containers were removed from Processes A and B following completion. The dust exposures measured by the DustTrak (FIGURE 4a) did not increase when the collection containers were removed from Process B. However, mass concentrations measured by the DustTrak exceeded 1 mg/m³ (Container 1) and 10 mg/m³ (Container 2) when harvesting product from Process A. Product harvest from Process B showed an increase of 0.002 mg/m³ above background, but the same task performed on Process A resulted in a much higher WBZ concentration of 2.25 mg/m³. Worker exposure during the harvest task for Process B was 99.9% lower than that from Process A.

The data from the FMPS (FIGURE 4b) showed similar trends as the DustTrak, however, the magnitude of the fluctuation in fine particle concentrations (5.6–560 nm) was much smaller than that shown on the DustTrak during product harvest. Unlike the DustTrak, the FMPS also did not show any significant increases in fine particle concentrations during the transfer of nanomaterials inside a ventilated enclosure. These results suggest that the contaminants released from these tasks were primarily larger agglomerates.

The implementation of a waiting time prior to product harvest was evaluated on Process A because no dust exposure was found during this task on Process B. The monitoring results have shown that the particulate mass concentrations measured on Process A were largely unaffected by container removal and brief concentration spikes were below 0.5 mg/m³. The average concentrations at the WBZ decreased from 2.36 to 0.01 mg/m³ while removing Container 1 and from 0.32 to 0 mg/m³ while removing Container 2 after implementing a 30 minute wait. Overall, the implementation of the wait time resulted in a reduction of 99.6% and 100% dust concentrations measured at the WBZ during the removal of Containers 1 and 2, respectively.

Operating the process blower (see FIGURE 1a) while cleaning the process tank significantly reduced worker dust exposure. When the downstream fan was tuned on, an average inward air velocity of 130 feet per minute (fpm) (40 m/min) was measured at the face of the open access door. The release of smoke tracers near the access door also showed that the smoke was being effectively captured. FIGURE 5 shows that the operation of the process ventilation reduced dust exposures during tank cleaning. The mass concentration in the WBZ averaged 0.71 mg/m³ (FIGURE 5a) and 0.18 mg/m³ (FIGURE 5b), when the process ventilation was turned off and on, respectively. The implementation of this simple process change resulted in a nearly 75% reduction in exposure during harvesting.

In down flow rooms, the air flows from the ceiling moves toward the floor and away from the WBZ before being captured at floor level air return registers. The ventilation measurements for the downdraft room showed that the total estimated air flow was 19,000 cubic feet per minute (cfm) (538 m³/min). The chemical fume hood located inside the room exhausted 1,100 cfm (31 m³/min) from the room. The ventilated enclosure had an average face velocity of 98 fpm (30 m/min) and recirculated this flow into the room following integral HEPA filtration.

As shown in FIGURE 6, the dust concentrations measured within the room at 0.002 mg/m³ were lower than that of the general facility due to the HEPA filtration (0.035 mg/m³ measured outside of the room). WBZ dust concentration measured with the DustTrak during weigh-out activities in the ventilated did not differ from the room background measurement of 0.002 mg/m³. Likewise, the nanoparticle number concentrations measured with the FMPS in the downflow room were below 500 particles/cm³, a factor of 10 to 40 less than the typical urban number concentrations of reportedly 5,000 to 20,000 particles/cm^{3.(28, 29)} The fine particle number concentration did not show significant increases above background during weigh-out and handling of material inside the ventilated enclosure. However, when materials were handled outside of the ventilated enclosure, the DustTrak detected some transient increases in mass concentration likely due to contaminant emissions from the mixing operation suggesting that this task should be conducted inside the ventilated enclosure.

The observed air flow into the load compartment (84 cfm or 2.38 m³/min) was slightly lower than the design air flow (100 cfm or 2.83 m³/min). This minor difference between the measured and design air flows may have been due to measurement difficulties. However, static pressures measured were approximately −0.03 inches of water (or −7.5 Pa) inside the cabinets indicating that they were under slight negative pressure with respect to the ambient environment.

Particle number concentrations at Site C were below 1,000 particles/cm³, much less than the 5,000 to 20,000 particles/cm³ reported for urban environments.^{(28, 29)} The average background mass concentration obtained from the DustTrak was as low as 0.004 mg/m³ at Site C. Workplace ambient mass concentrations during task monitoring were generally under 0.04 mg/m³ (FIGURE 7) consistent with ambient PM10 concentrations which ranged between 0.006 and 0.06 mg/m³ for this area.⁽³⁰⁾ The average concentration measured during the task of loading and unloading of the reactor was 0.01 mg/m³. Although most of the particulate mass is likely due to ambient air pollution background, there were concentration spikes approaching 1 mg/m³ (FIGURE 7) over 2–4 seconds. These spikes were due to the handling of the CNTs outside of the enclosure. Neither real-time nor filter-based concentration measurements detected leakage from the Easy Tube 3000 reactor. These findings indicate that the proper enclosure and ventilation of the reactor and associated process cabinets effectively contained process-generated nano-aerosols.

The results from these field surveys are summarized in Table I. Overall, the engineering controls and process changes had resulted in substantial reductions in airborne contaminants from tasks or processes involving nanomaterials. They showed promise for eliminating particle emissions at sources (e.g., the use of process ventilation and harvest wait time at Site A, and the application of a down flow booth and ventilated enclosure at Site B) and minimize particle release into the workplace (e.g., the use of a closed production system at Site C). The use of process ventilation at Site A largely mitigated particle emissions from the task of tank cleaning resulting in worker exposure to agglomerated nanomaterials. And the implementation of a 30 minute waiting time for product harvesting at Site A also reduced potential for exposure to nanomaterials as well as contact burns due to high equipment surface temperature.

The harvesting of material from reactors has been identified as a potentially high exposure activity in several manufacturing plants ^(34–38). In addition, the cleanout of the reactors has contributed to increasing facility concentrations and exposures to operation and maintenance workers. Leakage from pressurized reactors can also contribute to background concentrations and result in exposure to employees throughout the facility. At site A, the airborne concentration at the source during the collection of products from the discharge vessel was measured at 2.27 mg/m³ and 0.017 mg/m³ for Processes A and B, respectively. This difference was due to differences in the process design, specifically, the use of isolation valves at the collection vessel for Process B.

To investigate the potential for reducing worker exposure through the implementation of a simple process changes at Process A, a 30 minute wait time was added to allow the aerosol within the process to settle. The addition of this wait time nearly eliminated the worker exposure during product harvesting (Table I) showing a reduction in worker breathing zone concentration from 2.4 mg/m³ to 0.06 mg/m³. This simple process change could be used until more permanent control measures can be put into place including the isolation of the collection vessel from the process using butterfly (or other) valves and/or the implementation of an enclosure around the collection point. The use of isolation valves in the Process B effectively eliminated emissions during harvesting at this site. Another approach could be to use a ventilated enclosure around the discharge point to prevent the loss of nanomaterials into the work environment during product harvesting (FIGURE 8). Two studies have shown that when a reactor is housed in a well-designed and operated enclosure, particle loss to the work environment is low. ^{(38, 39)}

As part of normal operations, a worker cleaned the process tank with a hand tool creating a personal exposure of 0.71 mg/m³. To reduce worker exposure during this task, the system exhaust fan was operated keeping the process equipment under negative pressure. Worker breathing zone concentrations were reduced to 0.18 mg/m³ when the blower was kept on during reactor cleaning. The use of ventilation has been evaluated during reactor cleaning in other settings. Methner assessed the use of a portable LEV unit for controlling exposure during cleanout of a vapor deposition reactor used for producing nanoscale metal catalytic materials comprised of manganese, cobalt or nickel ⁽³⁷⁾. Analysis of real-time data and airborne metals showed an average reduction in concentrations of 88–96% during cleanout procedures when using a portable fume extractor. In this case, the operation of an exhaust fan during process cleaning operations reduced worker exposures while not requiring the purchase of any additional hardware.

Small-scale weighing and handling of nanoscale powders are common tasks; examples include working with a QA/QC sample, processing smaller quantities, and packaging/opening nanomaterials in production and downstream facilities. In these processes, workers may weigh out a specific amount of nanomaterials to be added to a process such as mixing or compounding. The tasks of weighing out nanomaterials can lead to worker exposure primarily through the scooping, pouring, and dumping of these materials. At site B, the use of a down flow booth significantly reduced particulate levels from ambient background pollution. In general, dust concentrations in the room were about an order of magnitude lower than the ambient facility background (0.002 vs. 0.059 mg/m³). However, when the nanofibers were manually mixed with solvent, some release of particulates was measured. Concentrations measured at the source showed transient peak concentrations of up to 0.08 mg/m³ when the nanofibers were manually stirred on the benchtop.

Methner et al. evaluated a university-based research lab that used CNFs to produce high performance polymer materials. ⁽⁴⁰⁾ Several processes were evaluated during the survey: chopping extruded materials containing CNFs, transferring and mixing CNFs with acetone, cutting composite materials, and manually sifting oven-dried CNFs on an open bench top. Real-time monitoring did not identify any process as a substantial source of airborne CNF emissions; however, weighing/mixing of CNFs in an unventilated area resulted in elevated particle concentrations compared to background. Other studies have shown that bench top activities such as probe sonication of nanomaterials in solution can also result in emission of airborne particles. ^{(36, 41)} Conducting the mixing tasks inside the ventilated enclosure used for weigh-out is an effective means for mitigating the noise and aerosol exposure.

Many different types of commercially available laboratory fume hoods can be employed to reduce exposure during the handling of nanopowders. Other controls have also been used in the pharmaceutical and nanotechnology industries for containment of powders during small quantity handling and manipulation. They include glove boxes, glove bags, biological safety cabinets or cytotoxic safety cabinets, and homemade ventilation enclosures. Newer nano hoods based on pharmaceutical weigh-out enclosures may be a reasonable alternative to larger fume hoods when only small-scale, bench-top manipulation of powders is needed. Overall, the published studies suggest that the selection of a fume hood with improved operating characteristics such as a variable air volume hood provides better operator protection than conventional fume hoods when handling dry nanomaterials.⁽⁴²⁾ When using any hood, the worker should strive to maintain the face velocity in the recommended range of 80–120 ft/min. ⁽⁴³⁾ Additionally, proper use of the engineering control by the operator and validation of the performance of the control equipment is essential for risk mitigation.

Site C used automated and closed processing systems designed and built to significantly control process emissions. Hazard control for this carbon nanotube reactor is largely integral to the equipment’s design and is intended to contain air contaminants. There are interlocks on the doors or access panels to safely shut down the operation if these doors are inadvertently opened. Ventilation is used as a secondary, redundant control measure that removes any process leakage, thermal decomposition products from the reactor exterior, and heat from the enclosed spaces around the reactor. Air samples suggested that the operation of the equipment did not contribute to air contamination in the workplace. Individual fibers were not detected in the workplace air, and elemental carbon concentrations were less than 1 μg/m³. However, short transient peak concentrations were seen when the worker unloaded the CNTs from the process suggesting that a higher exhaust flow for the cabinet may be required to fully contain the contaminants.