Cleaning as high-risk activity for respiratory particulate exposure during additive manufacturing of sand moulds and its preceding silica sand coating process

Abstract

Background

Occupational exposure to respirable crystalline silica (RCS) is a known cause of respiratory diseases, such as silicosis and lung cancer. Binder jetting additive manufacturing (AM) uses silica sand coated with sulphonic acid as feedstock material and operators are potentially exposed to RCS during various activities associated with AM. This includes the cleaning of the AM machine and associated equipment. This study aimed to investigate particulate exposures associated with additive manufacturing of sand moulds and its preceding silica sand coating process.

Methods

The particle size distribution (PSD) and particle shape analysis of different forms of silica sand (virgin, coated, and used) was determined using a Malvern Morphologi G3 automated microscope and the structural characteristics was measured using X-ray diffraction (XRD). Personal exposure and area monitoring for airborne respirable dust and RCS were performed using MDHS 14/4 and NIOSH 7602, while real-time particle number concentrations of 0.3 to 10 µm sized particles was measured using the TSI Aerosol Particle Counter (APC). Monitoring was performed for 2 operators over 8 d and included 3 d of coating, one day of cleaning the AM machine, and 4 d of printing during which 3 identical parts were manufactured.

Results

According to the PSD analysis, virgin and used silica sand particles were mostly in the respirable size range (d(0.9) = 3.98 ± 0.72 µm; and d(0.9) = 6.51 ± 2.71 µm, respectively), while coated sand was mostly in the inhalable size fraction d(0.5) = 29.76 ± 42.91 µm). The wt% results of the XRD analysis for the bulk virgin, coated and used silica sand were 97.3%, 92.6%, and 96.8% quartz, respectively. Personal exposure to RCS exceeded the exposure limit of 0.1 mg/m³ when the operator used compressed air to clean the coating machine’s filter (0.112 mg/m³) and exceeded the action level on the day the AM machine was cleaned (0.70 mg/m³). The results for real-time particle number concentrations of 0.3 to 10 µm sized particles showed peaks while the cleaning activities such as dry sweeping were performed.

Conclusion

The personal exposure to RCS was the highest on days when cleaning activities that used compressed air and dry sweeping took place. The high quartz content of the silica sand feedstock material and the respirable size of the virgin and used silica sand particles means that cleaning activities pose an RCS exposure risk to AM operators. Nine recommendations are made to reduce exposure to RCS during cleaning activities.

Introduction

Additive manufacturing (AM) is defined by the American Society for Testing and Materials (ASTM) as a process of building a printed part layer upon layer to create a 3-dimensional (3D) shape (ISO/ASTM 2021). The introduction of this technology has unlocked new opportunities for various manufacturing industries, such as aerospace, automotive, and medical sectors, among others (Gibson et al. 2015). The metal casting sector has also adopted the use of AM to manufacture sand moulds. Implementation of AM in metal casting means that sand moulds can be manufactured easier, quicker at a lower cost, and with less waste (Anakhu et al. 2018; Sun and Shang 2021).

There are 2 main techniques associated with AM of sand moulds, namely binder jetting and selective laser sintering (Le Néel et al. 2018). Binder jetting is a process, which uses a furan resin binder to selectively bind sand particles together, while selective laser sintering uses heat from a carbon dioxide laser to selectively melt or fuse sand particles together (Hackney and Wooldridge 2017; ISO/ASTM 2021). Silica sand is preferred as feedstock material over other materials, such as zircon, olivine, and chromite due to its ability to consume less furan resin binder, which results in defect-free moulds (Nyembwe et al. 2016).

Additive manufacturing using silica sand involves 3 phases: pre-processing, processing and post-processing. Pre-processing includes cleaning of the AM machine and loading of the AM machine job box with the silica sand. The processing phase involves printing of the actual sand mould, while the post-processing phase includes cleaning of the part as well as other finishing activities. With sand casting, there is an additional step at the beginning of pre-processing which is the coating of virgin (unused) silica sand with sulphonic acid. Virgin sand that has been coated with sulphonic acid is referred to as coated sand, while silica sand that has been supplied to an AM machine during a previous build cycle, and recycled, is referred to as used sand. It is crucial that the silica sand feedstock material used for the AM of sand moulds is coated with sulphonic acid to form a homogeneous mixture as it helps harden the furan resin binder (Dady et al. 2019) which plays an important role in improving the integrity and strength of the mould. The characteristics of the feedstock material used is extremely important since the use of various feedstock materials in AM has been associated with potential adverse health effects for operators (Bours et al. 2017). Evaluating the health hazards posed by feedstock materials involves 2 principal considerations: particle size and chemical composition (McClellan 2002; Du Preez et al. 2018).

The aerodynamic diameter of a particle determines the location within the respiratory tract to which particle will travel and potentially become deposited (Wilson et al. 2002; Brown et al. 2013; Kim et al. 2015). Subsequently, inhaled particles can be categorized into the following fractions: The inhalable fraction (<100 µm), which is the mass fraction of total airborne particulates inhaled through the nose and mouth (Cherrie et al. 2013; DOEL 2021), the thoracic fraction of inhaled particles, which can penetrate beyond the larynx (50% penetration at 10 µm) (Wilson et al. 2002; Brown et al. 2013; Stacey et al. 2018; DOEL 2021) and the respirable fraction which can penetrate to and deposit in the unciliated airways or alveolar region of the lung (50% penetration at 4 μm) (Brown et al. 2013; DOEL 2021). The sand used to manufacture metal casting moulds often contain a high percentage of silica (quartz) and operators are at risk of exposure to airborne respirable crystalline silica (RCS) during the AM processes. Following inhalation, RCS particles can deposit in the alveolar region of the lung and cause pulmonary inflammation and scarring leading to an increased risk for the development of respiratory conditions such as silicosis, lung cancer (Group 1 carcinogen), and chronic obstructive pulmonary disease (IARC 2012; Schroder, 2014; Prajapati et al. 2021). Therefore, the physical and chemical characteristics of the silica sand used to manufacture AM sand moulds should be investigated when determining the risk to health associated with sand mould AM.

In the context of foundries, removal of excess silica sand following the manufacturing of the mould results in particulates becoming airborne, leading to possible RCS exposure (NIOSH 1989). Similarly, the extent of exposure to hazardous feedstock material for AM operators is determined by the activities conducted across the 3 phases of AM. For example, during the pre-processing and post-processing phases, manual handling of the feedstock material can lead to respiratory exposure to the material used which could pose a risk of developing respiratory diseases (Du Preez et al. 2018). The extent of particulate exposure during various AM activities, such as cleaning the AM machine and associated equipment, remains uncertain.

The health risks associated with silica sand coating and the AM of sand moulds are not exclusively limited to particulates as exposure to sulphonic acid, for instance, can result in respiratory, eye and skin irritation, drowsiness, and dizziness (Usiquimica 2018). Furthermore, binder jetting can lead to the emission of volatile organic compounds (VOCs) which can cause eye and skin irritation with a possibility of allergic reactions, nausea, headaches, dizziness, as well as carcinogenic effects (IARC 2012; Tsai 2016; Kellens et al. 2017; Shuai et al. 2018). It is important to note that this study did not focus on exposure to sulphonic acid and VOCs, and therefore, a detailed discussion of these aspects will not be included.

Currently, there is no literature available concerning the exposure of operators to RCS and airborne particulates during the coating of silica sand and the subsequent AM of sand moulds for metal casting. Although it has been found than certain tasks such as the handling of powders increase the risk of exposure (Du Preez et al. 2018), the risk of particulate exposure during the cleaning of silica sand from coating and binder jetting AM machines has not yet been reported. Consequently, the objective of this study was to investigate exposures associated with silica sand coating and the subsequent binder jetting AM process by (i) investigating the physical and chemical characteristics of different types of silica sand particles (virgin, coated, and used), (ii) monitoring personal exposure to airborne RCS and respirable dust, and (iii) measuring particulate number concentrations in the workplace air during the coating and AM processes using a direct reading particle counter.

Methods

Study design

The study was conducted at a South African tertiary education institution’s AM facility over a period of 8 d and included coating and printing activities. This was a cross-sectional study that investigated the exposure associated with the AM of 3 identical sand moulds in order to inform and improve the health and safety procedures of the facility. The coating and printing phases were monitored as follows: Days 1 to 3 consisted of coating processes (including cleaning of the coating machine), on day 4 the Voxeljet VX1000 (Voxeljet, Germany) AM machine was cleaned and on days 5 to 8, 3 identical parts were printed. The coating and printing activities studied were performed by 2 AM operators who both participated in the study.

This study was classified as minimal risk and approved by the North-West University Health Research Committee (NWU-HREC) (NWU-00020-19-A1). Signed informed consent was obtained from the AM operators who participated in the study.

Facility and process description

All activities performed during the study are summarized in Table 1. One AM operator performed all the activities associated with the sulphonic acid coating of the silica sand. The room, in which the coating machine (Fig. 1) was located, was also used as a storage facility and the housekeeping in this room was poor as there was spilled silica sand on the floor from previous days. The only means of reducing airborne contaminants or introducing fresh air was a built-in dust extractor fitted to the coating machine and the occasional opening of the bay door. The coating process involved the AM operator using a built-in vacuum duct hose to draw sand from a bulk bag into the hopper. The sand then gradually flowed from the hopper into the screw conveyor where a tube from a sulphonic acid container supplied the sand with liquid acid. The acid was mixed with the sand inside the screw conveyor and the AM operator then placed another empty bulk bag beneath the screw conveyor wherein coated sand was collected. The built-in dust extractor collected airborne dust particles released when sand was poured into the bulk bag. Following the coating process, the bag was moved outside the building using a forklift.

It was not possible to perform exposure monitoring during all 3 AM phases in a single day as the sand moulds needed ± 12 h to dry between the processing and post-processing phases. Thus, the post-processing for each print was performed on the day following pre-processing and processing. On days 6, 7, and 8, there was one operator working with the AM machine, and on day 5, 2 operators were involved in the printing of the sand moulds. The AM operator manufactured 3 identical sand moulds of the same dimensions so that the exposure scenario could be repeated 3 times. The printed sand moulds had a height, width, and length of 33.5, 565.3, and 1000.0 mm respectively. The AM machine was located in a separate room from the coating machine.

Physical and chemical characterization of silica sand

Bulk samples of virgin, coated, and used silica sand were collected separately into 50 ml vials.

The Malvern Morphologi G3 (Malvern Instruments Ltd, UK) was used to quantify particle size distribution (PSD) in terms of optical diameter to provide a comprehensive understanding of the size ranges present in the silica sand sample as well as the shape (circularity and convexity) of the 3 types of silica sand samples. The 5 mm³ samples were placed in a dispersion unit of the instrument and dispersed onto a glass slide, after which the images of individual sand particles were taken and analysed using the automated microscope. The samples were analysed in triplicates. The Malvern Morphologi G3 counts the particles on the glass slide and reports the results in terms of number of particles in a specific diameter.

X-ray diffraction (XRD) analysis was conducted using the X’Pert Pro X-ray diffractometer (PANalytical Instrument, Netherlands) to determine the structural composition of the raw materials of the 3 sand types. The material for finely pulverized and undiluted sand samples were placed onto the spinner stage inside the XRD using a back-loading technique. The x-rays generated by a PW3376/00 Co LFF tube (PANalytical, The Netherlands) were then used to scan the sand samples. Thereafter, the PANalytical HighScore Plus software suite was used to interpret the data which included the Rietveld analysis as part of the data interpretation process. Peak identification was done using the International Centre for Diffraction Data (ICDD) Powder Diffraction File (PDF)4+ and Inorganic Crystal Structure Database (PAN-ICSD).

Particle number concentration

Real-time particulate number concentrations were measured using a factory calibrated Airborne Particle Counter (APC) (Aerotrak Portable Particle Counter model 9310 TSI Inc., MN, USA). The APC is used to monitor airborne particle concentrations in a work environment and characterize particles within the 0.3–10 µm size range and measures particle concentrations in specific size channels (0.3, 0.5, 1, 3, 5, and 10 µm). By providing real-time data on particle concentrations, immediate assessment of air quality in the workplace can be performed. For each day, the APC recorded data at 1-min intervals and was zeroed prior to each measurement. During sampling, the APC was placed on a table between 1 and 2 m away from the machines. The particle number concentrations (particles/m³) were measured throughout the coating and AM phases, beginning with background measurements in the room prior to the start of the specific process.

Personal and area monitoring

Personal respiratory exposure to respirable dust and RCS were monitored during each day according to the Methods for the Determination of Hazardous Substances (MDHS) 14/4 and National Institute for Occupational Safety and Health (NIOSH) 7602 (HSE 2014; NIOSH 2017). A sampling train consisting of an aluminium cyclone sampler attached to a polystyrene cassette containing a 37-mm polyvinyl chloride (PVC) filter with 5.0 µm pore size, attached to a GilianGilair plus (Sensidyne, Inc., LP, USA) air sampling pump with flexible tubing, was used to measure personal exposure in the operator’s breathing zone. The flow through each sampling train was checked before and after sampling using the Gilian Gilibrator-2 (Sensidyne, Inc., FL, USA) calibrator (2.5 L/min).

On each day of exposure monitoring, the operators performed both administrative and printing duties and the personal exposure measurements were collected while the operators performed printing duties. The time it took to complete the day’s printing duties varied between 45 min on day 8 when only post-processing was performed to 255 min on day 7 when all 3 phases were performed. The sampling times are shown in Table 3.

Area monitoring was also conducted to investigate concentrations of RCS and respirable dust in the workplace air. The sampling procedure was similar to that used for personal monitoring and the sampling trains were placed on tripod stands and positioned between one and 2 m away from the source and 1.5 m above floor level. In the AM machine room one of the sampling trains was placed in front of the AM machine (approximately 2 m from the AM machine door) and another was placed behind the AM machine at the hopper (approximately 1.2 m from the back of the AM machine). A total of 9 personal and 11 area RCS and respirable dust samples were collected and analysed of crystalline silica.

Following the shift, the filters from personal and area samples were transported inside their cassettes to a SANAS accredited analytical laboratory for analysis according to MDHS 14/4 and NIOSH 7602 (HSE 2014; NIOSH 2017). For MDHS 14/4, a Sartorius CPA225D balance (Sartorius, Göttingen, Germany) was used, and the method had a limit of detection of 0.01 mg, a limit of quantification of 0.03 mg, and an uncertainty of 1.09% (k = 2). For NIOSH 7602, a Bruker Tensor 27 FT-IR spectrometer (Bruker, Billerica, MA, USA) was used, and the method had a limit of detection of 0.004 mg, a limit of quantification of 0.012 mg and an uncertainty of 10.24% (k = 2) at >100 µg and 22.34% (k = 2) at >10 µg.

Statistical analysis of results

Descriptive statistics (mean, range, and standard deviation) were used to describe the PSD and shape of virgin, coated, and used silica sand. The mean particle sizes were classified as d(0.1) (10% of the particles were less than the stated diameter); d(0.5) (50% of the particles were smaller than the specified diameter); and d(0.9) (90% of the particles were smaller than the specified diameter). The time-weighted average (TWA) of the respirable dust and RCS exposure results only included printing related activities and were calculated by assuming that the exposure during administrative tasks was negligible. Statistical tests were performed, and graphs obtained using Graphpad Prism 8 (Graphpad Prism, version 8, GraphPad Software, La Jolla, USA).

Results

Physical and chemical properties of silica sand particles

Table 2 shows the PSD and structural composition for virgin, coated, and used sand. Virgin sand contained predominantly respirable sized particles (< 4 µm) [d(0.9) = 3.98 ± 0.72 µm], while coated sand particles were generally larger and concentrated in the inhalable fraction [d(0.5) = 29.76 ± 42.91 µm]. The majority of the used sand particles were also in the respirable fraction range [d(0.5) = 1.88 ± 0.12 µm]. The circularity and convexity analyses showed that the particles of virgin, coated and used sand were smooth, but not spherical. According to the XRD analysis, virgin, coated, and used silica sand all contained > 92% quartz. Mullite, aluminium oxide, and graphite were also present in minute amounts.

Particle number concentrations

Figure 2 shows the results obtained from the APC and illustrates the release of particles into workplace air during the coating (Fig. 2A,B) and printing (Fig. 2C,D) activities. The highest particulate number concentrations emitted during coating and printing were of 0.3 µm sized particles, followed by 0.5, 1, 3, 5, and 10 µm sized particle emissions.

Figure 2A shows the particle number concentrations of particles peaked after approximately 5 min (0.3 µm = 1.90 × 10⁸/m³; 0.5 µm = 1.86 × 10⁸/m³; 1 µm = 1.62 × 10⁸/m³; 3 µm = 4.60 × 10⁷/m³; 5 µm = 8.2 × 10⁶/m³; and 10 µm = 2.8 × 10⁴/m³) when the operator used compressed air to clean the coating machine filter. Figure 2B shows that on day 3, the particle number concentrations increased rapidly (time ≈ 6 min) when the coating machine filter was cleaned (0.3 µm = 1.13 × 10⁸/m³; 0.5 µm = 0.95 × 10⁸/m³; 1 µm = 0.63 × 10⁸/m³; 3 µm = 8.1 × 10⁶/m³; 5 µm = 8.5 × 10⁵/m³; and 10 µm = 4.5 × 10⁴/m³) and again when the bay door was opened (between time ≈ 71 to 95 min). Figure 2C illustrates the real-time airborne concentrations of 0.3 to 10 µm sized particles during cleaning of the AM machine. Particle number concentrations gradually increased when the operator started cleaning and peaked towards the end of the procedure (0.3 µm = 0.87 × 10⁸/m³; 0.5 µm = 0.8 × 10⁸/m³; 1 µm = 0.63 × 10⁸/m³; 3 µm = 0.17 × 10⁸/m³; 5 µm = 2.7 × 10⁶/m³; and 10 µm = 2.3 × 10⁴/m³), when the operator used a broom to sweep the AM machine room. The measured particle number concentrations during the different phases of AM of sand moulds varied depending on the AM phases performed (Fig. 2D). The peak airborne particle number concentrations were measured during the pre-processing phase (0.3 µm = 0.94 × 10⁸/m³; 0.5 µm = 0.43 × 10⁸/m³; 1 µm = 0.15 × 10⁸/m³; 3 µm = 9.3 × 10⁵/m³; 5 µm = 6.7 × 10⁴/m³; and 10 µm = 9.2 × 10²/m³).

Respirable dust and RCS

Although the process is automated, the operator was required to remain in the room during the coating process. Table 3 shows that the daily 8-h TWA personal exposure to RCS during coating ranged from 0.007 to 0.112 mg/m³. During the first day of coating, the personal exposure to RCS experienced by the operator exceeded the South African TWA occupational exposure limit-maximum limit (OEL-ML) of 0.1 mg/m³ (DOEL 2021). Respirable dust concentrations during personal monitoring ranged from 0.159 to 0.424 mg/m³ which were less than 10% of the TWA occupational exposure limit-restricted limit (OEL-RL) of 5 mg/m³ (DOEL 2021). The personal exposure to RCS for the printing process was the highest on day 4 when the machine was cleaned (0.070 mg/m³) and for respirable dust during day 5 (0.242 mg/m³) when pre-processing and processing was performed. The percentage quartz in the samples ranged between 2% and 48%.

Airborne concentrations of RCS measured in the area during coating and printing are also shown in Table 3. For coating, the area concentrations were lower than that of the personal exposure on all 3 d, but for the printing process, the area concentrations measured at position A1 (in front of the AM machine), exceeded the personal exposure concentrations on 2 of the 4 printing days.

Discussion

Information on the risks associated with the coating of silica sand for use in AM as well as personal exposure during the AM of sand moulds is limited. As such, this study was conducted to investigate respiratory exposure to particulates associated with these processes.

Physical and chemical characteristics

Bulk samples of virgin, coated, and used silica sand were analysed using optical microscopy and XRD to determine if the different types of sand pose a risk to the health of operators based on their particle size and chemical composition. According to the PSD results, the virgin and used silica sand were mostly comprised of particles in the respirable size range (virgin sand d(0.9) = 3.98 ± 0.72 µm; used sand d(0.5) = 1.88 ± 0.12 µm), while most of the coated sand was classified as being in the inhalable size fraction (d(0.5) = 29.76 ± 42.91 µm). The larger PSD of the coated silica sand compared to the virgin sand could have been caused by the coating process where the addition of liquid sulphonic acid to the virgin sand bound the silica sand particles together to form agglomerates. The large variation in PSD results for the coated sand shows that the mixture of the silica sand and sulphonic acid formed an uneven mixture, where some particles agglomerated and some did not. The smaller PSD of the used silica sand suggests that the coated sand agglomerates were broken down close to their original size by mechanical action during the printing process.

The results of the XRD analysis showed that the virgin, coated and used silica sand consisted of 97.3%, 92.6%, and 96.8% quartz, respectively. These percentages refer to weight % (wt%) of the crystal structures in the sample. It does, therefore, not necessarily refer to the total content of the sample but to the % of the sample detected by the XRD method. These results are consistent with another study that found that the sand feedstock material of a Voxeljet VX1000 AM machine contained > 90% quartz according to the manufacturer’s safety data sheet (SDS) and > 96% quartz according to Wavelength Dispersive X-ray Fluorescence analysis performed during the study (Meiring 2023). The high silica content of the silica sand feedstock material used during silica sand binder jetting inherently poses a significant risk to operators. Combined with the high percentage of virgin and used silica sand that falls in the respirable size fraction, it can be concluded that the silica sand used as feedstock material poses a risk to the health of the AM operator, as exposure can ultimately lead to an increased risk for the development of respiratory disease (IARC 2012; Upadhyay et al. 2014; Prajapati et al. 2021).

The combined approach of the PSD and XRD analysis allows for a more thorough assessment of the potential health risks associated with silica sand, particularly concerning the presence and quantity of crystalline silica in respirable size fractions. Additionally, it is concerning to note that the manufacturers of the silica sand AM feedstock material did not provide a SDS to the facility when the material was purchased, meaning that the AM operators were not aware of the potential dangers of the feedstock material. Subsequently, it was not possible to compare findings of this study with the information in the SDS as was done by Meiring (2023).

Respirable dust and RCS

Before the AM operator could start the coating process, they cleaned the coating machine filter. This was done using compressed air and caused a visible dust cloud as the particles captured on the filter were liberated into the air. This increase in airborne particles can be observed in Figure 2A and 2B, where real-time particle number concentrations of 0.3 to 10 µm sized particles increased during the cleaning activity. According to Table 3, the personal RCS exposure of the AM operator on day one exceeded the TWA-OEL-ML of 0.1 mg/m³. The combination of the APC, which showed an increase in the particle number concentrations in the room during cleaning of the coating machine, and the operator’s personal exposure measurement, therefore indicate that cleaning was the activity responsible for the increased exposure. Additionally, although the result cannot be compared to the personal exposure limit, the airborne RCS concentration in the work area exceeded the 50% action limit (Table 3). These results confirm that cleaning the coating machine using compressed air led to the emission of particles into the workplace air and subsequent personal exposure of the operator. This result is concerning because both virgin and coated silica sand feedstock material have a substantial quartz content, which means that the AM operators are at risk of exposure to RCS even if they are only exposed to relatively small concentrations of respirable dust. It should, however, be noted that all the personal exposure samples did not contain a high quartz percentage, and that the feedstock material was not the only source of dust at the facility. The operators moved around the facility and were exposed to respirable dust from other sources such as dust from outside the facility. The personal exposure sample of the operator who was over-exposed to RCS on day one contained 29% quartz while the area sample collected on the same day contained 48% quartz, and the personal exposure samples collected on days 2 and 3 contained 4% and 2% quartz, respectively. On day 3 the operator used a forklift to transport bulk bags with coated sand through the open bay door (Table 1). Exposure from this activity caused by the liberation of dust from outside areas (as is shown in Figure 2B) could have contributed to the lower silica content of the collected particles. Additionally, the wt% reported by the bulk sand XRD analysis did not include amorphous compounds that possibly formed part of the feedstock material, which could have contributed to the lower percentage quartz in the respirable dust samples. Nevertheless, although the results from personal RCS exposure sampling vary between day 1, 2, and 3 of coating, because of differences in the activities performed during these days, the cleaning of the coating machine’s filter was identified to contribute significantly to the operator’s RCS exposure and should receive priority when planning exposure reduction strategies and further exposure monitoring.

Prior to the start of the AM printing process, the AM operator also cleaned the AM machine using a vacuum, and the floor of the AM room using a broom. As with the cleaning of the coating machine filter, a dust cloud was visible while the AM operator was sweeping the floor. Figure 2C shows that the real-time particle number concentrations of 0.3 to 10 µm sized particles increased during the cleaning period and peaked while the AM operator was sweeping the AM room. The AM operator’s personal RCS exposure during the cleaning period was 70% of the TWA-OEL-ML which was the only time that the personal exposure of the AM operator exceeded the detection limit for RCS during the AM printing process. This personal exposure sample contained 44% quartz. Even though the current South African exposure limit was not exceeded, this result is still significant since the exposure limit for RCS in many other countries is currently 0.05 mg/m³ (OSHA 2019; Safe Work Australia 2024) and the health-based threshold limit value (TVL) listed by the American Conference of Governmental Industrial Hygienists (ACGIH) is 0.025 mg/m³ (ACGIH 2023). This means that personal RCS exposure of 0.07 mg/m³ still poses a risk to the health of the operator. During days 5 to 8, the AM operator did not experience significant personal exposure to RCS.

Figure 2D shows the real-time particle number concentrations of 0.3 to 10 µm sized particles during day 6 of the study, where the AM operator started with the post-processing of the previous day’s print and continued with the pre-processing and processing of the next print. The particle number concentrations of 0.3 to 10 µm sized particles increased above background levels during post-processing and pre-processing and returned to the background levels during the processing phase. Post-processing involves the manual cleaning of the printed sand mould while pre-processing involves manual loading of the AM machine with feedstock material as well as manual emptying of buckets below the hopper which contain used silica sand. These activities caused silica sand particles to become liberated into the workplace air. The concentrations of respirable dust measured at area A1 (in front of the AM machine) on days 6 to 8, when post-processing was performed were 3 to 5 times higher than day 5 when only pre-processing and processing were performed. This suggests that manual cleaning of the print during post-processing is an important source of airborne respirable dust in the AM room. Since the local extraction ventilation system in the room was not operational, concentration of respirable dust in the AM room air was unnecessarily high. A study conducted by Zontek et al. (2016) reported that poorly ventilated areas led to an increase in particle concentrations and Zhang et al. (2017) also observed that inadequate ventilation resulted in accumulation of particles in workplace air. It is, therefore, necessary for the facility to use local extraction ventilation and general ventilation systems properly in order to subsequently reduce particle build-up of both RCS and respirable dust.

Recommendations

The personal exposure results reported in this study represent 3 d of coating silica sand with liquid sulfonic acid and the subsequent printing of 3 identical parts. Even though the XRD analysis showed that the feedstock material contained a high percentage of quartz (>92%), the quartz content of the personal exposure and area samples collected during the study were substantially lower and varied considerably based on the activities performed by the operators (2 to 48%). It is recommended that the RCS exposure at the facility be measured again with specific focus on the cleaning activities to better understand exposure to RCS during cleaning activities since the scope of this study was to report exposure during one set of coating and printing activities.

Since the virgin, coated, and used types of silica sand feedstock material all contain substantial percentages of quartz, it is essential to keep exposure to airborne dust from the feedstock material as low as possible. Even though the AM operators were exposed to relatively low concentrations of respirable dust (max = 0.424 mg/m³) compared to the TWA-OEL-RL of 5 mg/m³, exposure to RCS still exceeded the RCS TWA-OEL-ML of 0.1 mg/m³ on one of the coating days (0.112 mg/m³). Therefore, relatively small peaks in exposure to respirable dust can lead to substantial RCS exposure. The results from this study show that the risk for exposure to RCS and respirable dust during the AM of sand moulds is the highest during activities that liberate silica sand particles into the workplace air, such as cleaning with compressed air and dry sweeping. Figure 2A, 2B, and 2C show that the real-time particle number concentrations of 0.3, 0.5, 1, and 3 µm sized particles increased as a group, compared to Figure 2D, where the concentrations of 0.3 µm sized particles are higher than 0.5, 1, and 3 µm sized particles. This shows that the cleaning activities which caused the peaks in Figure 2A, 2B, and 2C (cleaning with compressed air and dry sweeping) transferred more energy to the silica sand particles and liberated bigger particles which led to increased personal exposure concentrations as was measured during gravimetric monitoring. The recommendations suggested to the AM facility primarily focusses on reducing the amount of energy transferred to the silica sand particles during cleaning and other pre- and post-processing activities.

The following recommendations were made to help reduce emissions during the cleaning activities at the AM facility: (i) According to the Regulation 13 (a) of the Regulations for Hazardous Chemical Substances, 2021 (DOEL 2021), the use of compressed air to remove hazardous particles from any surfaces or person is strictly prohibited. Therefore, alternative cleaning methods should be used to clean the coating machine filters. The facility can use a glove box to clean filters without releasing particulates into the workplace air. This glove box should be under negative pressure to prevent the release of contaminants; (ii) The coating machine filters should be maintained and kept clean regularly to minimize accumulation of dust on the filter; (iii) A mobile dust extractor with suitable high-efficiency particulate air (HEPA) filter can be placed at the coating machine to avoid the spread of quartz particles into the room from the coating machine. This mobile dust extractor should be able to create a capture velocity of 1 to 2.5 m/s (contaminants released into moving air with moderate energy) and a transfer velocity of 15 m/s (fine dusts) (HSE 2017); (iv) A vacuum cleaner equipped with a HEPA filter should be used, instead of a broom, throughout the cleaning procedure since dry sweeping unnecessarily liberates particles into the air; (v) If a vacuum cleaner is not available or feasible, wet sweeping methods can be utilized to reduce resuspension of silica sand dust in the air; (vi) The general ventilation in the coating room should be improved by providing a system that can introduce fresh air and extract contaminated air, thereby minimizing particle accumulation in the room; (vii) The floor and surfaces of the coating room were visibly dirty and facility should improve the housekeeping of the coating room which will reduce contamination of workplace air caused by resuspension of silica sand dust; (viii) No SDS was provided by the silica sand manufacturer and it is recommended that the AM facility obtain an SDS because the AM operator was unaware of the dangers of working with silica sand. With SDSs, the AM operator will be able to gain information on how to handle and store silica sand as well as the dangers of exposure. The information from the SDSs should be incorporated into a training programme and AM operators should undergo regular training and supervision (DOEL 2021); (ix) Throughout the study, the AM operator was wearing casual clothing and using an FFP2 half-face disposable respirator and latex gloves, to help protect them from exposure to hazardous chemicals. Additional protective clothing such as standard overalls and work boots, should be provided to prevent silica sand dust adhering to the AM operators’ casual clothing. The operator should then change clothes in a separate change room from where dirty clothes can be removed for washing. This will improve the general housekeeping in the facility.

Limitations

The study did not investigate the difference between the emissions generated by the 3 phases of AM since there was not time to allow the emissions return to background concentrations between the phases. This made comparisons between the emissions generated by each printing phase difficult.

The data reported in this study showed that increased particle number concentrations were associated with cleaning activities. However, particle number concentrations were only measured in the area using the APC (Aerotrak Portable Particle Counter model 9310 TSI Inc., MN, USA). Currently, there are personal direct reading particle counters available to better measure and characterize personal exposure to aerosols in the breathing zone of the operators. These instruments should be used in future studies so that personal exposure can be investigated.

During the gravimetric sampling, clear styrene cassettes were used. Although common practice, this was a limitation since they are more prone to wall deposits compared to conductive plastic cassettes and exposure could, therefore, have been underestimated (NIOSH 2014).

Conclusion

Current literature is limited in its coverage of the potential hazards linked to the coating of silica sand used for the AM of sand moulds, as well as the binder jetting AM technique for such moulds. This study aimed to address this gap by investigating respiratory particulate exposure during both the binder jetting AM of sand moulds and the preceding sulfonic acid coating procedure. The silica sand feedstock material was identified as the primary source of exposure in addition to other secondary sources such a dust from outside the facility. Virgin, coated, and used silica sand all pose a risk as they all contained respirable quartz particles. The AM operator experienced RCS exposure that exceeded the exposure limit while cleaning the coating machine filter using compressed air. Additionally, real-time particle number concentrations of 0.3 to 10 µm sized particles peaked during cleaning activities such as dry sweeping and the use of compressed air. Therefore, reduction of exposure during cleaning activities, such as cleaning of the coating filter and dry sweeping, should be prioritized by the facility and the technique used to clean the equipment used for coating and printing should be improved.

Supplementary material

Supplementary material is available at Annals of Work Exposures and Health online.

Acknowledgments

We would like to thank the AM facility and operators for participating in the study.

Author contributions

The manuscript is as a result of a compilation from contributions of all authors.

Funding

This research was funded by the South African Department of Science and Innovation (DSI) and forms part of the Collaborative Programme for Additive Manufacturing.

Conflict of interest

The authors declare no conflict of interest.

Data availability

The data will be made available on reasonable request.

References