Additive manufacturing (AM) is a broad manufacturing term that encompasses a range of processes that create objects by adding material through a computer-aided design model (Plessis, Preez, and Stefaniak 2022; Stefaniak, Du Preez and Du Plessis, 2021b). Three-dimensional (3D) printing is a form of AM, which builds objects through layer-by-layer deposition of feedstock material using a 3D printer machine and computer software (Plessis, Preez, and Stefaniak 2022) Fused filament fabrication (FFF), also known as Filament Freeform Fabrication), is one 3D printing process in which filaments are melted and extruded from a heated nozzle to deposit material. FFF is an emerging technology and one of the most popular additive manufacturing processes, especially for consumers and small manufacturers such as industrial, academic military and residential sectors (MacCuspie et al. 2021).

Polycarbonate (PC) is a versatile material, and PC filaments are widely used for FFF 3D printing. Polycarbonate filaments are often loaded with additives to achieve different properties of the print objects. These additives range from dyes, photopolymer resins, organometallic compounds, carbon nanomaterials, nanometal oxides to micro-meter-scale particles such as copper, bronze, steel, tungsten, gold, and aluminum nitride (Vance et al. 2017). Several engineered nanomaterials were infused into PC filaments, such as silicon dioxide nanoparticles, titanium nitride nanoparticles (Vidakis et al. 2021), titanium carbide nanopowder (Vaisanen et al. 2022; Vidakis, Petousis, Grammatikos, et al. 2022), aluminum nitride nanoparticles (Vidakis, Petousis, Mangelis, et al. 2022), and carbon nanotubes (Potter et al. 2021).

During heating, PC filaments undergo thermal degradation and release fine particles (0.1 to 2.5 um) and ultrafine particles (d < 100 nm) as well as numerous volatile, and semi-volatile organic compounds (VOCs) that may be derived from PC polymer and additives in the polymer (Alijagic et al. 2022; Azimi et al. 2016; Byrley et al. 2020; Gu et al. 2019; Stefaniak et al. 2017, 2019; Tedla et al. 2022). These emissions may pose a potential hazard to human health. Currently, the potential health hazard attributed to PC filament printing emissions exposure remains to be determined.

A NIOSH research group used a condensation nuclei counter to study PC filament emission rates, and determined that the number-based particle emission rates from an industrial-scale material extrusion AM machine were approximately 2.2 × 10¹¹ number/min and the total VOCs emission rates were approximately 19 mg/min (Stefaniak et al. 2019). In addition, Stefaniak et al. (2019) found low levels of acetone, benzene, toluene, and m,p-xylene during PC filament printing processes. Potter et al. (2019) noted that PC filament emissions contained bisphenol A (BPA), phenol, chlorobenzene, DEHP, and di-tert-butylphenol. Previously, Farcas et al. (2019) examined PC filament printer emission-induced cell toxicity originating from a commercial PC 3D printer which was generated in a chamber using a 3D printer and subsequently collected for incubation in a cell culture medium. The number-based size distribution of the particles within the chamber was between 140 and 170 nm and mean particle sizes in the cell culture medium were 201 ± 18 nm. Analysis of elemental composition of particles collected in the cell culture medium found was as follows: C, O, Ca, Na, Si, Ni, Cr, Fe, S, Al, and Cl. The organic compounds in the emission collection cell culture medium identified were BPA, p-isopropenylphenol, and phenol. At 24 hr post-exposure, PC emissions were internalized in human small airway epithelial cells (SAEC) and induced the following concentration-dependent responses: cytotoxicity, oxidative stress, apoptosis, necrosis, as well as increases in number of pro-inflammatory cytokines and chemokine production in SAEC (Farcas et al. 2019). These results demonstrated that PC filament 3D printing emissions were associated with a cellular toxicity in SAEC.

Although cell-based in vitro toxicity analysis is increasingly applied to (1) screen and rank chemicals for prioritizing toxicity studies, as well as to (2) determine underlying toxic mechanisms, the toxicological significance of in vitro study-generated data in hazard and risk assessment is limited. In comparison with animal-based in vivo studies, in vitro cell investigations (1) lack tissue-specific differentiated functions, (2) physiological context with other cells and tissues, and (3) biological concordance in metabolism of xenobiotic chemicals (Blaauboer 2008). The aim of this investigation was to examine the influence of PC filament printer emission-induced acute and sub-chronic pulmonary and systemic toxicity in rats.

In this study, the particle mass exposure concentration in the chamber calculated from FMPS measurements, assuming bulk density of PC and spherical particle shape, was between 0.3 and 3 mg/m³, and daily mean mass concentration from gravimetric measurements of particles collected on filters, over all exposure days, was 0.529 mg/m³ (Krajnak et al. 2023). Note that the effective density of the aerosol emissions is unknown but was reasonably assumed to be equivalent to bulk PC for the purpose of calculating particle mass. Determination of the effective density of airborne particles is complex; although Katz et al. (2020) determined that during FFF 3D printing with ABS, values were approximately 30% higher compared with bulk, but for PLA, the effective and bulk densities were similar. In addition, it is important to note that the composition of the particles released during 3D printing with PC filament was not analyzed in this investigation and might be made up of plasticizers, other semi-volatile constituents in the filament, and/or polycarbonate polymer. In the aforementioned study by Katz et al. (2020) used an aerosol mass spectrometer and electron energy loss spectrometry to interrogate the composition of aerosols generated from FFF 3D printing. For ABS, both techniques identified aromatic compounds and the energy loss spectra were consistent with polymeric styrene, which supports the concept that bulk composition was at least partially reflective of aerosol composition. It is conceivable that variations of particle emissions over the 4-hr print job may be associated with several factors: 1) build plates are heated and this may lead to extruded plastic cooling faster the farther away it is from the first layer; 2) the 3D printing housing chamber is not temperature controlled, and thus the air temperature inside this chamber might fluctuate during print jobs; and 3) variations of filament composition from the manufacturing process. The precise chemical composition of each roll of the filament was proprietary and not made available. Another factor that might account for the observed variability is that generating the exact same mass concentration in the exposure chamber each day was found to be difficult, partly because three printers were utilized and it was not uncommon for one of them to fail while printing. Prints might fail if (1) the part did not stay attached to the build plate, (2) the print nozzle became clogged, or (3) a software error occurred. During an exposure, if a print failed, that printer was stopped to prevent damage to the printer. If a printer door was opened to try and fix it, this might dilute the concentration down to near zero during an animal exposure, thus the door was never opened to attempt to fix a stopped printer.

The estimated particle alveolar dose over 4 days of exposures (8.1 μg) in the exposed rats can be scaled to a human equivalent dose by using known alveolar surface area values. The alveolar surface area of the human lung is approximately 70 m² (Hasleton 1972; Butler 1976; Gehr, Bachofen, and Weibel 1978), and the surface area of a rat lung is approximately 0.4 m² (Pinkerton et al. 1982; Stone et al. 1992). By using these values, the equivalent human alveolar lung burden was calculated; 8.1 μg × 70 m2/0.4 m2 = 1,418 μg. The human alveolar deposition fraction for the particles generated in this study was 0.2256 based on the MPPD model (Anjilvel and Asgharian 1995; Yeh and Schum 1980). By using this deposition fraction, the total inhaled particle mass was calculated: 1,418 μg/0.2256 = 6.29 mg. A healthy adult human has a minute ventilation of 8 L/min (Pritchard 1976). The total volume of air inhaled by a worker over the duration of 4 days of animal exposures is estimated: 4 days × 4 hr/day × 60 min/hr × 8 L/min = 7,680 L (7.68 m³). Dividing the total inhaled particle mass by the total volume of inhaled air may result in a human equivalent average exposure concentration: 6.29 mg/7.68 m³ = 0.819 mg/m³.

The particle mass concentration of 0.529 mg/m³ is on the same order of magnitude as reported for injection molding and grinding tasks (approximately 0.1 to 0.3 mg/m³) that involved heating PC polymer (Boonruksa et al. 2017). On a count basis, the mean PC particle concentration for all animal exposures during 3D printing was 4.6 × 10⁶ particles/cm³ (Krajnak et al. 2023) although this value was likely an underestimate because the concentration in the inhalation chamber often reached the upper limit of the FMPS instrument (1 × 10⁶ particles/cm³) during the first 45 min of a print. This exposure profile was consistent with test chamber investigations that reported a peak concentration of 1 × 10⁶ particles/cm³ at the start of 3D printing with PC filament and remained above 10⁴ to 10⁵ particles/cm³ for the remaining print time (Ding, Wan, and Ng 2020). Stefaniak et al. (2021a) noted that during large format additive manufacturing with PC polymer in the workplace, an average particle number concentration of 1.2 to 2.5 × 10⁴ particles/cm³ was detected. The count median diameter during exposures was 35 nm (Farcas et al. 2019), consistent with particle sizes observed in 3D printing test chamber studies with PC filament (Ding, Wan, and Ng 2020).

Acetaldehyde, acetone, ethanol, and ethylbenzene were released during 3D printing with PC filament and were a component of the total exposure received by animals (Krajnak et al. 2023). These VOCs are all known thermal breakdown products of PC polymer (Cabanes and Fullana 2021; Guillemot, Oury, and Melin 2017). Acetaldehyde was present in the highest concentration, which was consistent with Guillemot et al. (2017) who reported acetaldehyde as a main breakdown product of PC polymer. Acetone levels may be explained by the method of using ABS Juice for print bed preparation. ABS juice is made by dissolving 40 g of ABS print filament into 100 ml of acetone. A thin layer of the ABS juice is applied to the print bed with a paint brush and then allowed to dry. This results in a thin layer of ABS on top of the print bed which is barely visible. The ABS sticks well to the print surface, and PC melts into the ABS and thus adheres to this layer.

Most PC polymer is made from BPA monomer. When PC polymer is heated, BPA and several other thermal breakdown products are released into the air (Erickson 2007; Huang et al. 2018). Our results showed BPA is one of the many VOCs emitted from 3D PC filaments (Stefaniak, Du Preez, and Du Plessis 2021b). The mean concentration of BPA in the emissions was 5.3 ± 0.18 μg/m³ (Krajnak et al. 2023). MPPD estimates of BPA deposition in the entire respiratory system were 36 ng on day 1, 536 ng on day 15, and 1,072 ng on day 30 of exposure (Krajnak et al. 2023). It was suggested that PC filament emission-induced neuroendocrine toxic effects may be associated with elevated BPA in the PC filament emissions (Krajnak et al. 2023). There is little information regarding BPA-induced pulmonary and systemic toxicity in vivo. One study showed that whole-body inhalation exposure of BPA up to 90 mg/m³ for 8 weeks (6 hr/day, 5 days/week) induced no marked effect on body weight, hematology, serum chemistry, organ weights, or histopathological lesions in rats (Chung et al. 2017). O’Brien et al (2014) found that when exposed to BPA in utero via maternal diet, an individual may experience BPA sensitization without symptoms of pulmonary inflammation when the subject becomes an adult. Although no pulmonary nor systemic toxicity was detected in this study, it is difficult to exclude the potential toxic effects of BPA due to the relative low concentration of BPA in the PC filament emissions.

To our knowledge, there are no published studies of respiratory effects attributed to inhalation of PC polymer particles. The absence of appreciable respiratory toxicity following exposure to PC emissions during 3D printing in the current study might be explained by multiple factors. First, results obtained from this study may reflect that the respiratory tract was not a target organ. The count median diameter of PC particle emissions during 3D printing (35 nm) was sufficiently small to reach the gas-exchange region of the lung. Once deposited, these particles may translocate across lung epithelium cells to the blood, distribute systemically, and subsequently induce adverse effects in other organs. Indeed, Krajnak et al (2023) found that 30 days of PC filament emissions exposure affected the neuroendocrine system by reductions in thyroid stimulating hormone, follicle stimulating hormone and prolactin in rats, which were associated with an elevation in markers of cell injury, decrease in active mitochondria in the olfactory bulb and gonadotropin releasing hormone cells and fibers, fall in number of tyrosine hydroxylase immunolabeled fibers in the arcuate nucleus, and decrease in spermatogonium (Krajnak et al. 2023).

Second, exposure to biologically active components of PC particles might have been attenuated by the presence of polymer matrix. This effect of the polymer matrix is in agreement with findings in prior investigation that demonstrated the toxicity of engineered nanomaterials was attenuated when embedded in a polymer matrix such as paint (Saber et al. 2012; Smulders et al. 2014).

Finally, it is possible that the particle mass concentration used for exposures was simply insufficient to induce adverse respiratory health effects. If the absence of respiratory health effects was a reflection of low mass concentration, some caution is warranted in the interpretation of these results as it might mean that exposure to PC particles is not without risk. In the current study, the mean particle mass exposure (0.529 mg/m³) was generated by three desktop-scale 3D printers. Secondo et al. (2020) noted that some indoor spaces may have up to 30 desktop-scale 3D printers operating simultaneously, which might translate into higher particle mass exposure than what was employed in the current study and potential toxicity may have been masked.

Previously, Farcas et al. (2019) in vitro studies found that 3D PC filament emissions induced a concentration-dependent toxicity in human small airway epithelial cells (SAEC). Further, the PC filament emissions produced concentration-dependent oxidative stress, apoptosis, necrosis, and generation of pro-inflammatory cytokines and chemokines in SAEC. Our in vitro studies also indicated that exposure to 3D PC filament emissions might pose a toxicological risk. However, in this in vivo study, data demonstrated that inhalation exposure to 3D PC filament induced no significant pulmonary or systemic toxicity in rats. It is possible that there are several factors that may be attributed to the divergence of the toxicological observations between our in vivo and in vitro toxicity investigations, including lung burden differences and exposure conditions.

There is a significant difference in the equivalent alveolar lung burdens of PC particles between the in vitro and in vivo studies (1697 μg vs 60.6 μg). In our previous in vitro experiments, the mean delivered PC particle concentration was 3.64 × 10⁷ particles/ml, with a count distribution median size of approximately 120 nm in the solution. By utilizing a lognormal fit of the particle size distribution, and using a MATLAB script, the particle mass per ml of the solution was calculated as 1.4 μg/ml by assuming spherical particles with a density of 1.2 g/cm³. Each plate well had a surface area of 0.33 cm² and was dosed with 0.1 ml of solution such that the particle mass per surface area was calculated: (1.4 μg/ml × 0.1 ml)/(0.33 cm²) = 0.424 μg/cm². By assuming that a typical rat lung has an alveolar surface area of 4000 cm², one can calculate the equivalent alveolar lung burden of 1697 μg in rats (0.424 ug/cm² × 4000 cm² = 1697 μg). In our current studies, the alveolar burden of day 30 exposure without and with clearance were 60.6 μg and 25.5 μg, respectively (determined with MPPD software). Note that MPPD only considers mechanical mechanisms (e.g., mucociliary escalator) when estimating lung clearance, but the actual amount cleared may be higher if removal by chemical dissolution is an important mechanism for particles released during FFF 3D printing with PC filament. Currently, an inhalation project exposing rats to higher concentrations of 3D PC filament emissions might help to elucidate the role of lung burden.

Traditional submerged cell culture technology was applied in our previous in vitro studies. There are several drawbacks of the submerged cell culture that may affect toxicological outcomes: a) our SAEC-based submerged cell culture technology does not mimic the physical features of airway mucosa nor alveolar unit of the lung, b) the physicochemical characteristics of the PC filament emissions may change in the cell culture media (Upadhyay and Palmberg 2018), and c) the submerged cell culture technology is unable to recapture the inhalation conditions nor the particle deposition pattern in lungs (Upadhyay and Palmberg 2018). Indeed, our previous parallel investigations of 3D ABS filament emission-induced toxicity showed a similar pattern to this study, that ABS filament emissions produced a significant toxicity in the SAEC-based submerged cell culture model but not in vivo pulmonary or systemic toxicity in rats (Farcas et al. 2019, 2020). However, when an Air-Liquid Interface (ALI) model of primary normal human-derived bronchial epithelial cells was applied to perform in vitro toxicity assessments, 3D PC filament emissions-mediated toxicity was noted indicating that the results obtained from the ALI model corresponded with those from the in vivo studies (Farcas et al. 2022). It is possible that an ALI model may be a better choice to evaluate 3D PC filament emission-induced in vitro toxicity.

Finally, for live organs in vivo, there are many other cells coexisting that may communicate with the targets cells to compensate emission-initiated toxicity; therefore, a single-type cell-based toxicological observation may not adequately simulate in vivo conditions (Madorran et al. 2020).

Consumer-grade level PC filament printers emitted UFPs and VOCs during real-time printing. Data demonstrated that whole-body inhalation exposure to 3D PC filament emissions at the concentration of 0.529 mg/m³ for 1, 4, 8, 15, or 30 days exposure (4 hr/day, 4 days/week) did not produce significant pulmonary or systemic toxicological responses in rats, which was inconsistent with toxicological results observed in our in vitro studies. Given the significant lower lung burden of the 30-day exposure in rats compared with the in vitro study, it is warranted to perform a toxicity assessment to expose rats to a higher concentration of 3D PC filament emissions to confirm whether the concentration of 0.529 mg/m³ of PC filament emissions is too low to initiate any pulmonary and systemic toxic responses in rats. Currently, there is an ongoing study to determine pulmonary and systemic toxicologic responses to 2.5 mg/m³ of PC filament emissions via whole-body inhalation exposure in rats, and these results might provide further insights for 3D printing exposures.