Human exposure to metals in consumer-focused fused filament fabrication (FFF)/3D printing processes

Abstract

Fused filament fabrication (FFF) or 3D printing is a growing technology used in industry, cottage industry and for consumer applications. Low-cost 3D printing devices have become increasingly popular among children and teens. Consequently, 3D printers are increasingly common in households, schools, and libraries. Because the operation of 3D printers is associated with the release of inhalable particles and volatile organic compounds (VOCs), there are concerns of possible health implications, particularly for use in schools and residential environments that may not have adequate ventilation such as classrooms bedrooms and garages, etc. Along with the growing consumer market for low-cost printers and printer pens, there is also an expanding market for a range of specialty filaments with additives such as inorganic colorants, metal particles and nanomaterials as well as metal-containing flame retardants, antioxidants, heat stabilizers and catalysts. Inhalation of particulate-associated metals may represent a health risk depending on both the metal and internal dose to the respiratory tract. Little has been reported, however, about the presence, speciation, and source of metals in the emissions; or likewise the effect of metals on emission processes and toxicological implications of these 3D printer generated emissions. This report evaluates various issues including the following: metals in feedstock with a focus on filament characteristics and function of metals; the effect of metals on the emissions and metals detected in emissions; printer emissions, particle formation, transport, and transformation; exposure and translation to internal dose; and potential toxicity on inhaled dose. Finally, data gaps and potential areas of future research are discussed within these contexts.

Graphical Abstract

1. Introduction

Fused filament fabrication (FFF), also referred to as 3D printing, is a rapidly growing technology used in consumer venues such as homes, schools, libraries, and universities as well as numerous manufacturing sectors (de Leon et al., 2016; Ford and Minshall, 2019; Wasti and Adhikari, 2020). The operation of 3D printers results in the release of particles and volatile organic compounds (VOCs) which can be inhaled, raising concerns of possible health impacts for those exposed to 3D printer emissions. Although poly-lactic acid (PLA) and acrylonitrile butadiene styrene (ABS) polymers compose the primary filaments used for consumer applications, other materials such as poly(e-caprolactone) (PCL), poly carbonate (PC), and nylon are also used for specialized commercial and industrial applications (Liu et al., 2019; Reyes-Rodríguez et al., 2017). For consumer applications, additives such as dyes, organometallic compounds, metal particles, metal oxides or carbon nanomaterials may be used to modify thermal, mechanical, electrical, flame resistance and aesthetic properties of the print objects. These filaments may also contain relatively high percentages of micrometer-scale particles of copper, bronze, steel, tungsten, gold, aluminum nitride, boron nitride, etc. (Alberts et al., 2021; Quill et al., 2018). Particles emitted in the printing process are typically in the ultrafine1 or nano-size range implying that micrometer size metallic particles present in the filament mostly remain in the print object. The inhalation of airborne particles and VOCs are associated with a variety of health concerns (USEPA, 2019, USEPA, 2021), which has raised concerns for exposures to 3D printer emissions. While the concern for exposures has often focused on the organic components of the emissions in both vapor and particulate phase, until recently (Chen et al., 2020), relatively less attention was devoted to the potential emissions and health concerns of metals.

Consumer and occupational exposures to 3D printer emissions (i.e., particles and VOCs) have been investigated by number of laboratories (Azimi et al., 2016; Byrley et al., 2019; Stefaniak et al., 2019). Depending on the filament composition and printing conditions, millions to billions of ultrafine particles and numerous VOCs can be released over a typical print cycle (Azimi et al., 2016; Stefaniak et al., 2017a). For context, particle emission rates and average geometric mean for 3D printers were comparable to HP2055dn laser printers (Stefaniak et al., 2017a); however, because of the time required to print a typical object is on the order of hours compared to min for laser printers potential exposures may be expected to be greater for 3D printer operation provided that the density of emitted particles are comparable. In addition, due to recent reductions in price for low-cost printers and increased popularity among children and teens, these devices are finding their way into homes, schools and libraries that may not have adequate ventilation comparable to occupational workplaces. Although the use of 3D printers is increasing for consumer applications, uncertainties remain concerning user exposures.

The characteristics of particle emissions such as chemical composition, concentration (mass, surface area, or number), size and distribution, density, and shape depend on a wide range of variables including filament compositions, printers, and operational conditions (e.g., nozzle temperature and size of chamber containing the printer). Transport and transformation processes including aerosol dispersion and particle aggregation are impacted by spatial and temporal factors (e.g., room configuration, ventilation, time of operation and distance from the printer) and will determine the effective particle size distributions and concentration of the resultant exposure (Azimi et al., 2017; Poikkimaki et al., 2019).

Polymer compositions are typically reported for filaments, print objects and emissions; but less has been reported concerning the composition, amounts or the contributing effects of polymer additives. This issue may be relevant to exposure given the wide range of additives (e.g., metals, dyes, organometallics, and nanomaterials) and the possible inclusion of these additives in the resultant particle emissions. Inhalation of toxic metals as a component of particulate matter has been shown to be a significant exposure route to systemic circulation (Fortoul et al., 2015). Many of the metals associated with negative health effects of exposure to polluted outdoor air, such as Fe, Ni, V, Cr, Zn, Al, Mn, and Pb are also found in both printer filaments and emissions. To date, most assessments of exposure to 3D printer emissions have involved the use of either PLA or ABS filaments without considering potential contributions of metallic or other constituents of the printer filaments. The primary purpose for this report is to review the presence and effects of metal-containing additives on particle emissions from consumer FFF use with respect to human exposure, consider translation to internal dose, evaluate the toxicological implications of such inhaled dose, and to identify potential data gaps related to these particulate emissions (Fig. 1).

Figure 1. Consumer exposure to emissions in (FFF).

. A. Pushing the 3D printer filament through the heated nozzle releases emissions, B. Emissions consist of particles, free and particle-bound vapors of volatile organic compounds (VOCs), and trace metals, C. These emissions can deposit along the respiratory tract, including in the nasal or extra-thoracic, tracheobronchial and pulmonary regions. Ultrafine material may translocate directly to the brain from the extrathorasic region or from the lower respiratory to systemic circulation. Material physically cleared from the tracheobronchial region and subsequently swallowed to the gut may also be distributed systemically. Secondary cytokines and inflammation may also result in systemic effects from epithelial perturbation in the respiratory tract, including impacts on the neuroendocrine system. Although transport of FFF emitted particles through nasal and gastrointestinal pathways is plausible, this process has yet to be demonstrated.

2. Literature Search and Selection Process

Three separate pools of peer-reviewed literature were collected to be included in this review. The first literature search was performed to capture study characteristics of metal-fill filaments that have been synthesized experimentally. These studies included (1) synthesized composite polymers that contained a polymer commonly used in FFF 3D printers and one or more metal additives and (2) used the composite polymer in a FFF 3D printer. Most studies not meeting these criteria were excluded. Search terms used included “3D Printing”, “fused deposition modeling”, FDM, “fused filament fabrication”, “FFF”, “filament synthesis”, “composite”, “Fabrication”, “extrusion” and “nanocomposite”. Information that was extracted included composite type, metal component size and form, percent of metal additives, extruder type, printer type and analytical characterization method (Table 1).

Table 1.

Summary of metal thermoplastic composites: synthesis, extrusion, printing, and characterization.

Composite type	Metal component size & form	Percent of metal additives	Extruder type	Printer type	Analytical characterization	References
ABS-stainless steel	Stainless steel Cr, 14% wt; Si, 0.63% wt; Ni, 0.07% wt; Fe, 85% wt; 30 μm, Pa	10, 15, 23% wt	Filabot Extruder Model FB00071	Hyrel System 30 M	SEM,b Calorimetry, tensile strength	(Ryder et al., 2018)
ABS-TiO2	TiO2, 0.05 μm, P	TiO2 5% wt	Twin screw extruder/compounder (model ZK 25T)	MakerBot replicator	SEM, tensile properties, Stress-strain	(Torrado Perez et al., 2014)
PVA-Au	Au, 0.01, 0.02 μm, P	Au 0.07% wt	Thermo Fisher PolyLab OS single screw	Ultimaker 2+ FDM printer using a 0.4 mm nozzle and 0.6 mm nozzle	UV–Vis,c TEMd	(Kool et al., 2019)
ABS-TiO2,	TiO2, 0.021 μm, P	TiO2 1, 5, 10% wt	DSM Xplore Micro 15 cc Twin Screw Compounder, Filabot Wee Extruder	Flash forge creator	XRD,e Perkin Elmer LS-55 luminescence spectrometer, FTIR,f tensile properties, UV–Vis	(Skorski et al., 2016)
ABS- Cu & Fe	Cu < 24 μm, P Fe < 43 μm, P	Cu & Fe 10–50% wt	Filastruder	NP-Mendel FDM- (open creators)	Tensile and thermal properties, optical microscope	(Hwang et al., 2014)
ABS 5% TiO2 ABS 2% ZnO ABS 5% SrTiO3 ABS 5% Al2O3	TiO2, NRg ZnO, nanorods SrTiO3, 5 μm, P Al2O3, 0.5 μm, P	TiO2 5% wt ZnO 2% wt SrTiO3 5% wt Al2O3 5% wt	Twin screw extruder/compounder (Model ZK 25 T),	MakerBot Replicator	SEM, mechanical properties including tensile strength and elongation to break	(Torrado et al., 2015)
ABS-Fe ABS-Cu	Fe, 45 μm, P Cu,10 & 45 μm, P	Fe & Cu, 5–40% by volume	Single screw extruder	Stratasys FDM300	SEM, Thermal conductivity, heat capacity, mechanical properties, stress strain	(Nikzad et al., 2011)
Al2O3 in Nylon	100 μm, P 100, 120, 150 μm, mix, P	Al2O3 13–50% wt	Single screw extruder	Stratasys FDM	SEM, tensile strength, elongation to break, young's modulus, yield strength	(Singh et al., 2016)
Nylon-Fe	Fe- < 30 μm, P Fe-50–80 μm, P	Fe 30–40% by volume, 70–76% wt	Brabender Plasti-Corder single screw extruder	Stratasys FDM	Tensile properties	(Masood and Song, 2004)

^a.P-powder.

^b.SEM, scanning electron microscope.

^c.UV–Vis, ultraviolet-visible spectrometry.

^d.TEM-transmission electron microscopy.

^e.XRD, X-ray diffraction analyses.

^f.FTIR-Fourier-transform infrared spectroscopy.

^g.NR, not reported.

The second literature search was performed to collect study characteristics of trace metals detected in FFF 3D printer emissions. These studies included trace metals detected in the emissions from a FFF 3D printing process. Most studies not meeting these criteria were excluded. Search terms included “metal exposure”, “emission”, “particulate matter”, “PM”, “aerosol”, “metal” “metal-toxicity”, “additives”, “extruder”, “FDM”, “fused deposition modeling”, “fused filament fabrication”. Information that was extracted included material, metal detected, sampling materials, sampling technique, polymer manufacturer, and color (Table 3, Table 4).

Table 3.

Summary metals detected in FDM 3D emissions.

Material	Metals detected	Sampling	Techniques (characterization)	References
ABS	Cr, Ni, Si, Ca, Mg, Na, Al	Filters (47 mm, 2 μm)	SEM-EDSa (gold/palladium sputter coated)	(Stefaniak et al., 2017b)
PLA	Fe	Filters (47 mm, 2 μm)	SEM-EDS (gold/palladium sputter coated	(Stefaniak et al., 2017b)
ABS, 95–100%,	Fe, Zn	Filters and TEM grids	ICP-OES,b TEM-EDS,c ASM,d miniDISC,	(Steinle, 2016)
PLA, 90–90%	Fe, Zn	Filters and TEM grids	ICP-OES, TEM-EDS, ASM, miniDISC,	(Steinle, 2016)
ABS/PLA	Al, Fe	Filters, 0.8 μm polycarbonate membrane	ICP-MS,e FE_SEM-EDS,f	(Stefaniak et al., 2019)
ABS	Na, Si, Cu, Ca, Al	TEM grids	TEM-EDS, CPC,g SMPS-OPS,h	(Zontek et al., 2017)
PLA	Na, Si, Cu, Mg, Ti, Al	TEM grids	TEM-EDS, CPC, SMPS-OPS	(Zontek et al., 2017)
PLA-Cu, Zn, Si, Fe infused		SEM stub covered with carbone tape, direct exposure	SEM-EDS, ICP-MS	(Vance et al., 2017)
ABS	Ca, Na, Si, Ni, Cr, Fe, Al	Biosampler, Cell culture Medium filtered through 0.2 μm, mounted on SEM stub	FE-SEM-EDS, NTA,i ELSj	(Farcas et al., 2019)
PC	Ca, Na, Si, Ni, Cr, Fe, Al	Biosampler, cell culture Medium filtered through 0.2 μm, mounted on SEM stub	FE-SEM-EDS, NTA, ELS,	(Farcas et al., 2019)
PLA	Si, K	TEM grids	SEM-EDS, TEM-EDS, SMPS	(Youn et al., 2019)
PLA Clear Yellow Clear Orange Translucent-Blue	Si Mo, Na, Ni, Si Ca, Fe, Mg, Mo, Mn, Si, Sr, Zn	3 μm polycarbonate filter cassettes, filters for ICP	ICP-MS, ELPI,k SMPS, FE-SEM-EDS	(Yi et al., 2019)
ABS Black	Na, Mg, Al, Si, Ca, K, Fe	Surface swabs	SEM-EDS	Brinsko-Bechert and Palenik (2020)

^a.SEM-EDS, Scanning Electron Microscopy, Energy Dispersive X-Ray Analysis.

^b.ICP-OES, inductively coupled plasma atomic emission spectroscopy.

^c.TEM-EDS, Transmission Electron Microscopy-Energy Dispersive X-Ray Analysis.

^d.ASM-Aerosol Spectrometer.

^e.ICP-MS, inductively coupled plasma-mass spectrometry.

^f.FE_SEM-EDS, Field emission Scanning Electron microscope -Energy dispersive X-Ray Analysis.

^g.CPC-Condensation particle counters.

^h.SMPS-OPS, Scanning mobility particle sizer-optical particle sizer.

^i.NTA, Nanoparticle tracking analysis.

^j.ELS, Electrophoretic light scattering.

^k.ELPI, electrical low-pressure impactor.

Table 4.

Filament color and metals detected in FDM 3D printing emissions.

Material	Manufacturer	Color (terms taken from literature as it is)	Detected metal	Reference
ABS	MakerBota	Natural	Cr, Ni	(Stefaniak et al., 2017b)
		Blue	Cr, Na, Mg, Si, Ca, Ni
		Red	Cr, Al, Ni
		Black	Mg, Ca
PLA	MakerBotb	True red	Fe	(Stefaniak et al., 2017b)
		Army green	Fe
		Ocean blue	Fe
		Transparent blue	Fe
ABS	Rock Hill	Yellow	Fe, Zn	(Steinle, 2016)
PLA	Rock Hill	Yellow	Fe, Zn
PLA		White	Si, K	(Youn et al., 2019)
PLA	MakerBot	Clear yellow Clear orange Translucent blue	Si Mo, Na, Ni, Si Ca, Fe, Mg, Mo, Mn, Si, Sr, Zn	(Yi et al., 2019)
ABS	3DXTech	Black	Ni, Cr, Fe, Ca, Na, Al	(Farcas et al., 2019)
PC	Gizmo Dorks	Black	Ni, Cr, Fe, Co	(Farcas et al., 2019)

^a.This filament was cross-referenced for use in the Stefaniak et al. (2017b) study by Yi et al. (2019).

^b.This filament was cross-referenced for use in the Stefaniak et al. (2017b) study by Yi et al. (2019).

The third literature search was performed to collect study characteristics of toxicity studies involving 3D printer generated emissions or 3D printed objects. Studies that were included contained toxicological information from FFF 3D printer emissions or 3D printed products associated with metal additives (Table 5). Most studies not meeting these criteria were excluded. Search terms included “metal exposure”, “emission”, “particulate matter”, “PM”, “aerosol”, “metal”, “metal-toxicity”, “additives”, “extruder”, “FDM”, “fused deposition modeling”. Information that was extracted includes material, study type, experimental regimen, and toxicological implications (Table 5). Databases used for all three literature searches included Web of Science, ProQuest Agricultural & Environmental Science, Science Direct, PubMed and Google Scholar. Different combinations of terms and wildcards were used for all three literature searches.

Table 5.

Toxicity studies related to 3D printer generated emissions and 3D printed objects

Material	Study type	Experimental regimen	Toxicological implications (examples)	Reference
ABS, PLA	Human clinical exposure	1-hour exposure to 26 healthy adults (13 female and 13 male) with average age 25 years in a single-blinded, randomized, cross-over design with median ([25th; 75th percentiles]; min; max) average LDSAa (μm2/cm3) values of 81.0 ([47.1; 113]; 25.7; 358) for ABS and 7.2 ([4.8; 10]; 2.9; 17) for PLA; spirometry, questionnaire, FeNOb and urine sample immediately post exposure and nasal secretion, FeNO and urine sample at 2–3 h post exposure. Specific inflammatory cytokines were measured in nasal and 8-isoprostaglandin F2α (8-isoPGF2α) in urine samples.	Slight increase in FeNO suggesting possible eosinophilic inflammation in ABS exposed	(Gumperlein et al., 2018)
ABS and PLA	Human case report	28-year-old self-employed adult with a history of asthma in childhood reported respiratory symptoms 10 days after using printers	Respiratory problems, asthma demonstrated with MeCh challenge; symptom worse with ABS than PLA	(House et al., 2017)
Not specific	Self-reported health and occupational exposure survey	Questionnaire administered to 46 workers across 17 companies.	Working more than 40 h per week with 3D printers was significantly associated (P < 0.05) with having been given a respiratory-related diagnosis (asthma or allergic rhinitis).	(Chan et al., 2018)
Black ABS	In Vivo, Male, Sprague-Dawley rats	3-hour nose-only exposure 0.9 ± 0.1 mg/m3 mean aerodynamic diameter (ELPIc): 70 ± 2 nm	Microvascular dysfunction, elevated mean arterial pressure	(Stefaniak et al., 2017a)
ABS (2 different filaments), PLA, and Nylon	In Vivo, Male C57BL/6 mice	Single 50 μL intra-tracheal instillation Particle diameters estimated roughly from SEM images were 71 ± 20 nm (mean ± standard deviation) for “ABS d”, 106 ± 20 nm for “ABS c”, and 14 ± 25 nm for PLA. Concentrations varied by assay performed.	Strong inflammatory response, increase neutrophils number	(Zhang et al., 2019)
ABS	In Vivo, Male, Sprague-Dawley rats	Whole-body exposure for 4 h/day, 4 days/week and five exposure durations (1, 4, 8, 15, and 30 days) at 240 ± 90 mg/m3 average geometric mean particle mobility diameter of 85 nm and geometric standard deviation 1.6 nm.	Minimal and transient pulmonary toxicity: Increased cytokines (IFN-γ and IL-10)d in BALFe at days 1 and 4 post-exposure with peak of IL-10 on day 15; increased macrophages at day 15. Increased serum biomarkers of renal and hepatic function on day 1. No significant histopathology.	(Farcas et al., 2020)
ABS (2 different filaments), PLA, and Nylon	In Vitro epithelial cells (A549), rat alveolar macrophages (NR8383)	Submerged culture Particle diameters estimated roughly from SEM images were 71 ± 20 nm (mean ± standard deviation) for “ABS d”, 106 ± 20 nm for “ABS c”, and 14 ± 25 nm for PLA. Concentrations varied by assay performed; duration of incubation not provided.	Cell death, oxidative stress, inflammatory responses	(Zhang et al., 2019)
Black ABS, Black PC	In Vitro Human SAECf	Cells exposed as undiluted (0%), 25% dilution, and 50% dilution in serum-free SABM™, resulting in six doses for each filament type Mean particle sizes in cell culture medium were 201 ± 18 nm and 202 ± 8 nm for PC and ABS	24-hour post exposure increase in cytotoxicity, oxidative stress, apoptosis, necrosis, and production of pro-inflammatory cytokines and chemokines	(Farcas et al., 2019)
Multiple 3D printed or molded materials	In Vitro, bovine embryo and ER activation in BG1Luc4E2 cell line	Submerged co-cultured with leachate from printed or molded parts for 20–22 h	Inhibited embryo cleavage	(de Almeida Monteiro Melo Ferraz et al., 2018)
SLA-3D printed material	In Vivo Zebrafish	Fish embryo toxicity assays performed with wild type, double transgenic and single transgenic lines incubated with leachate for 48 h at various dilutions	Developmental toxicity correlated with in situ generation of reactive oxygen species (ROS),g an increase in lipid peroxidation and protein carbonylation markers and enhanced activity of superoxide dismutase (SOD)h and glutathione-S-transferase (GST)i in embryos exposed to concentrations as low as 20% v/v for plastic extracts; ROS- induced cellular damage led to induction of caspase-dependent apoptosis; significantly decreased acetylcholinesterase (AChE)j activity with lack of any CNS-specific apoptotic phenotypes as well as lack of changes in motor neuron density, axonal growth, muscle segment integrity or presence of myoseptal defects	(Walpitagama et al., 2019)

^a.LDSA - Lung deposited surface area.

^b.FeNO - fractional exhaled nitric oxide; PGF2α Prostaglandin 2α.

^c.ELPI- Electrical Low-Pressure Impactor.

^d.(IFN-γ and IL-10)- Interferon gamma and interleukin 10.

^e.BALF-Broncho-alveolar lavage fluid.

^f.SAEC-Small airway epithelial cells.

^g.ROS- reactive oxygen species.

^h.SOD- superoxide dismutase.

^i.GST- glutathione-S-transferase.

^j.AChE- acetylcholinesterase.

2.1. Commercially available metal-containing filaments

Commercially available filaments can be broadly classified as filaments containing trace amounts of metals, and composite filaments containing relatively high percentages of micrometer size metal particles (Table 2). Polymer materials may include ABS, PA, PC, PLA, nylon, PET or PETG (Ngo et al., 2018; Singh et al., 2020).

Table 2.

Examples of commercial filaments that contain larger proportions of metallic powders.

Manufacturer/provider	Metal filament types	Metal percentage	Reference
ColorFabb	CopperFill BronzeFill SteelFill BrassFill	Cu, ~80% Bronze, ~80% Stainless steel ~80% Brass, ~80%	(Colorfabb, 2020, Laureto et al., 2017)
Proto pasta	Composite Steel fiber PLA Bronze Composite HTPLA Brass Composite HTPLA Copper Composite Fiber- HTPLA Composite Iron PLA Magnetic Iron PLA	Fe < 35.7, Cr < 10.2, Ni < 12, Mo < 1.5, Si < 0.6% Cu < 42.3, Tin < 17.7% Cu < 44.1%, Zn <15.1% Cu < 60% Fe < 45% Fe, 43–45%	(Protopasta, 2020)
BASF	Ultrafuse 316L Metal (stainless steel)	Stainless steel 90%	(MatterHackers, 2020)

Trace metals detected in commercial filaments may vary by filament color or product type (Table 3, Table 4). Metals may be added to plastics in FFF print filaments for a range of purposes, which are in some cases stated by the manufacturer and in other cases are not specified. Metal-based colorants have some advantages over organic-based colorants due to their better heat and light stability. The wide range of industrial metal-based colorants and variety of trace metals present in colored and non-colored filaments, make it difficult to associate specific trace metals with filament colors. It is also likely that the higher relative percentages of Al, Fe, and Zn have been added for their flame-retardant activities. For example, Al, Fe, and Zn appear to be most commonly detected in commercial filaments at concentrations ranging from 0.008–0.047 mg/kg (Yi et al., 2019). Additional trace metals reported in filaments include V, Ti, Cu, Cr, Cd, Co, Mn, Ni, and metalloids Sb and As (Yi et al., 2019).

Filaments possessing high percentages of metals primarily added for aesthetic purposes or to enhance the structural characteristics of printed objects are also on the market including PLA-alloy composites such as CopperFill, BrassFill, BronzeFill, Magnetic iron, steel and stainless steel (Table 2) (MatterHackers, 2020a). There is no defined size range of metallic particles that are added into commercial filaments. Laureto et al. (2017) reported particles size analyses for metal-containing filaments by combining back-scattered electron images and Image J. The particles size distribution peaks were ~ 15 μm for CopperFill and BronzeFill and ~ 10 μm for Magnetic-iron PLA and Stainless-steel PLA (Laureto et al., 2017). The minimum sizes of filler-particles seemed to be lower for Magnetic-iron PLA and Stainless-fill PLA (below 1 μm) compared to CopperFill and BronzeFill. However, it has to be clear that the method is known to underestimate sizes (Laureto et al., 2017).

In addition to the broad range of commercially available filament types, colors, and special effects (e.g., clear, or opaque colors, fluorescence, phosphorescence etc.), the ability to fabricate custom filaments is now possible using lower cost consumer-grade filament extrusion devices. These devices have become more popular due to their lower prices and greater availability.

3. Metals in feedstock (printer filaments)

3.1. Filament characteristics & function of metals

Metals in the form of elemental particles, salts, metal oxides and organometallics have been blended into polymers for a variety of functional and aesthetic purposes. These metallic compounds may be added to plastics as inorganic colorants (Zn, Pb, Cr, Co, Cd and Ti) (Campanale et al., 2020), flame retardants (Sb₂O₃, Al(OH)₃, zinc borate), antioxidants (cadmium and lead compounds), UV or heat stabilizers (Pb, Sn, Ba, Cd and Zn) (Hahladakis et al., 2018) and metal catalysts (zinc compounds) (Bahramian et al., 2016). Metals can also serve as slip agents (zinc stearates) to reduce surface friction, or in the form of oxides (Fe₂O₃, CuO and ZnO) as pro-oxidants and photo-oxidation catalysts (Hahladakis et al., 2018). The addition of metals has been reported for a range of polymers, commonly used in 3D printing, including ABS, PLA, and nylon (Alassali et al., 2020) (Table 1).

Various metals (0.02–100 μm diameter particles) are added to 3D printing polymer filaments at relatively high concentrations (≥5%) for the modification of physical and chemical properties of print objects (Ryder et al., 2018; Torrado et al., 2015) (Table 1). In addition to commercially available specialty filaments, extruding devices are available that allow blending of additives, such as metals, into base polymers to create unique formulations. By modifying existing filament materials with metals, it is possible to achieve desired physical, mechanical, thermal, electrical, or chemical properties of printed objects as well as to simplify the extrusion processes. For example, metals have been added to reduce tensile strength, improve thermal conductivity, and enhance the printing process as well as increase the stability of large-scale 3D structures (Hwang et al., 2014). Thermal and mechanical properties of ABS were improved by adding iron and copper fillers (Nikzad et al., 2011). Further, the presence of TiO₂ increased tensile strength (Torrado Perez et al., 2014) and decreased degradation of ABS (Skorski et al., 2016).

The inclusion of metal and metal oxide particles, ranging in size from nanometers to micrometers, alters the characteristics of filaments and print objects constructed from several polymer types (Table 1). ABS was the most used polymer as well as nylon and polyvinyl alcohol (PVA). Cu, Fe, Au or stainless-steel particles in the micrometer size range have been blended with ABS at 5–50% by weight (Hwang et al., 2014; Nikzad et al., 2011). Other particles such as Al₂O₃ (Singh et al., 2016) or Fe have been added to nylon (Masood and Song, 2004) and to ABS (Torrado Perez et al., 2014; Waheed et al., 2019) at various weight percentages. Mechanical and fracture characteristics typically remain similar with the base polymer at the lower percentage inclusion; while the heat capacity, magnetic and conductivity characteristics are modified as anticipated by the character of the metallic inclusions (Hwang et al., 2014; Masood and Song, 2004; Nikzad et al., 2011; Ryder et al., 2018; Singh et al., 2016). Nanoscale TiO₂ added to ABS has also been shown to result in photocatalytic activity (Skorski et al., 2016).

Addition of metals to polymers can either accelerate or retard the melting process (Lalia-Kantouri, 2005; Lee et al., 2006). While the accelerating effect could be advantageous in reducing the melt temperature of filaments to ease the extrusion process in feedstocks, the retarding effect is mainly intended to minimize possible flaming. The modification of polymer melting temperature may also influence the emission rates of particulates and VOCs. Although metals have been added to filaments to modify print object characteristics (electrical, physicochemical, etc.) for industrial applications, consumer applications have mostly focused on the aesthetic characteristics of print objects with metal additives. Non-industrial filament extruders have recently come onto the market allowing consumers to make their own custom formulations, potentially leading to emissions to the home environment for a wide range of filament compositions (Byrley et al., 2020).

4. 3D Printer Emissions

4.1. Emissions during printing

Metal-associated particle emissions show size fractions that range between 0.01 and 0.25 μm (Stefaniak et al., 2017b; Yi et al., 2019; Youn et al., 2019; Zontek et al., 2017). In several of these cases, metals were mapped to emitted particles using SEM-EDS or TEM-EDS. For consumer applications, metals are typically added at lower concentrations as plasticizers, pigments, or flame retardants (Farcas et al., 2019) and at higher concentrations to simulate the appearance of cast metal for the print objects (Table 2). For example, metals likely included for their flame retardant rather than aesthetic function have been reported in PLA emissions at metal concentrations of up to 3168 ng/g of iron, 92.3 ng/g of Mn, 42.8 ng/g of Zn, per mass printed (Yi et al., 2019). Although some of these studies report metal concentrations in the emissions, others report only detection of metals by methods such as SEM-EDS (Table 3). We are not aware of any studies that report the speciation of the metal components. There are several studies that compare high percentage metals found in filaments to metals in the emissions. For example, Vance et al. (2017) reported that for Copper-Fill filament with a high percentage of copper (21%), no copper was discernable in the emitted printer particles using SEM-EDS. However, they also suggested further investigations on the composition of emissions to verify the absence of copper. Similarly, metals were not detected in emissions collected from tungsten and copper containing filaments (Alberts et al., 2021) using SEM-EDS. In another study, Copper-Fill PLA showed a comparatively high emission of particles compared to non‑copper PLA filaments; however, the copper content of emission was not measured (Stabile et al., 2017). Similarly, Poikkimaki et al. (2019) reported that PLA with copper additive showed an increased particulate emission rate with high nanocluster (below 3 nm particles) content compared to printing PLA without copper; however, the presence or absence of copper on the emissions was not reported.

Common metal emissions detected in ABS and PLA include Cu, Al, Fe, Zn, Ca, Mg and Na (Table 3). Cr and Ni were specifically found in ABS emissions and Ti in PLA emissions. Emissions from PC were reported to contain Cr, Al, Fe, Ni, Ca, and Na (Table 3). In most cases only qualitative information is available from TEM/SEM-EDS evaluations of particulate emissions; however, (Yi et al., 2019) reported particulate emission values for trace levels of metals that may be relevant to health effects including Co, Mn, Mo, Pb, Sn, and Sr at levels ranging from 0.03–92 ng/g filament printed for 3D pens or printer operations. Another study of 3D pens, which are often marketed for children, reported the possible risk of exposure to metals during printing with filaments some of which contained copper and steel (Sigloch et al., 2020). Emissions on the order of 10⁵–10⁶ particles/cm³ were observed and reported to contain thermoplastic materials, metal components as well as carbon nanotubes (Sigloch et al., 2020; Singh et al., 2021).

Although specific colors are not typically reported to be associated with specific metals, the relative occurrence of Cr was reported to be higher than other elements in natural, blue, and red colored ABS (Stefaniak et al., 2017b). Also, emissions from black PC and ABS printing both contained Cr, Ni and Fe (Farcas et al., 2019). The concentrations, speciation, and bioavailability of metals in particulate emissions compared to their presence in filaments has not been widely explored. Although in cases where metal concentrations in particulate emissions have been reported, they are typically quite low, and a better understanding of metal speciation and leaching characteristics may contribute to more meaningful determination of exposure, dose, and effects.

4.2. Particulate formation

Sampling for particulate emissions for FFF 3D printers has typically been conducted in controlled room settings using sealed chambers with filtered and controlled air flows, and at various distances from the printer nozzle. For use in a FFF 3D printer, the end of the filament is first placed inside of the printer head. The filament is then mechanically pushed through a 3D printer nozzle as it is heated to partially liquefy it. The recommended nozzle temperature is determined by the filament material being used. Nozzle temperatures used in 3D printer studies have ranged from 180 °C to 280 °C (Byrley et al., 2019). The filament then cools as it exits the nozzle so that the melted polymer can be added as a thin layer and solidified in the form of a 3D printed object (Ding et al., 2019).

During the liquefaction stage, filaments release emissions primarily in the form of semi-volatile organic compounds (SVOCs) and VOCs (Ding et al., 2019; Vance et al., 2017; Zhang et al., 2018). The SVOCs, compounds with lower vapor pressures, likely nucleate to form particles while other more volatile chemicals stay as free vapors (Zhang et al., 2018). The species of the individual VOCs and SVOCs being released depends on the chemical composition of polymer (Byrley et al., 2020; Davis et al., 2019) and also by the polymer additives (Potter et al., 2021).

Based on chemical analysis of particles released from the use of ABS, particles are likely to be formed, in part, from SVOCs originating from additives or breakdown products of the ABS polymer (J. Gu et al., 2019; Potter et al., 2019; Vance et al., 2017; Zhang et al., 2018). There are several lines of evidence that support this concept. First, for ABS, Raman spectra for the aerosol fraction is different from that of the filament and print object (Vance et al., 2017). Next, the boiling points for styrene (145 °C) and acetonitrile (77 °C) are significantly lower than the typical extrusion temperatures (210–270) (Vance et al., 2017). Although these monomers may condense on aerosol particles, they are not likely to form nucleation sites for aerosol emissions. These SVOC emissions may be comprised of oligomers, solvents, catalysts, and other additives such as organic and inorganic colorants. These SVOCs nucleate to form nanosized particles which then grow into larger particles through the adsorption of other volatile compounds (Zhang et al., 2017). These larger particles further agglomerate in the environment as they cool and move away from the printer head (Ding et al., 2019). Experiments that record particle measurements over the print time have shown a shift in median particle size to larger particles (100–200 nm, even up to 400 nm depending on material type) over time (Vance et al., 2017; Zhang et al., 2017). PLA, another popular filament, is theorized to follow a similar particle generation pathway as ABS; however, the dynamics of PLA emission particle formation has not been investigated as thoroughly (Ding et al., 2019; Vance et al., 2017; Zhang et al., 2018).

A study by Ding et al. (2019) found that most of the weight loss from the filament occurred during liquefaction, before solidification of polymer in the print object. Interestingly, a study by Simon et al. (2018) found that using different operating procedures for a 3D printer resulted in different levels of emissions. Emissions were reduced when the nozzle was cleaned and the filament was retracted from the printer head before heating to reduce excess heating of the filament (Simon et al., 2018). Generally, higher temperatures are associated with greater emission rates, especially those above recommended printing temperatures which can lead to degradation/depolymerization of a filament (Byrley et al., 2019; Potter et al., 2019; Wojtyla et al., 2017). Oxygen was also found to play a role in the emission of styrene and other volatile compounds from ABS and CNT enabled ABS filaments (Potter et al., 2019).

4.3. Effects of metals on emissions

The thermal decomposition behavior of polymers depends on polymer type, structure of the complexes, heating rate, sample mass, metal type or composition and the atmosphere around the sample (Lalia-Kantouri, 2005; Ou et al., 2016). In general, the accelerating or retarding effect of transition metals could be related to reduction of the activation energy in the thermal degradation process or hindering of the mobility of polymer chains in the heating process (Lee et al., 2006).

The trace metal content of the particulate emissions varies with filament polymer and color, however, no specific relationship to function (e.g., dye, flame retardant, etc.) or metal speciation has been reported. In addition, a wide range of trace metals have been detected in the particulate emissions from filaments composed of various polymers (see Table 3, Table 4). Aerosol formation has been associated with less volatile filament additives SVOCs rather than the more volatile VOCs (Vance et al., 2017). Consequently, trace metals are likely to be associated with SVOCs that are involved in the nucleation process for aerosol particle growth.

The addition of trace substances to filaments can change the printer emission characteristics (Pelley, 2018; Zhang et al., 2017). A higher rate of particle emissions and smaller particle size was noted for the use of Tungsten-fill ABS filament (Alberts et al., 2021). Addition of Cu to PLA also resulted in higher particle emission rates and smaller particle sizes (Alberts et al., 2021; Poikkimaki et al., 2019; Stabile et al., 2017). This finding differs from Vance et al. (2017) who reported fewer particles for the Copper-fill filaments compared to neat filament.

Because the inclusion of relatively high Cu content in the filament has been reported to increase thermal conductivity of the print product as compared to product printed with neat filament (Nikzad et al., 2011), and higher temperatures tend to increase the particulate emissions (Byrley et al., 2019), we suggest the possibility that the higher particle emissions for the copper-containing filaments could be a result of more efficient heat transfer from the nozzle surface throughout the filament. However, this concept has not been investigated with respect to the 3D printing process.

5. Human Exposure and Translation to Internal Dose

5.1. Exposure Modeling

Exposure modeling provides powerful tools for predicting inhalation exposure to aerosol particles and VOCs emitted from consumer products and processes. Because exposure models do not include detailed anatomical and physiological parameters, exposure models typically only predict either airborne concentration or a total inhaled dose in units of mg/kg body weight. And as with any model, the accuracy of the prediction is dependent on the quality and nature of the input data. The models differ in their characterization of exposure parameters such as product information and chemical characteristics, room size and ventilation, spray nozzle features, etc. It should also be noted that the average ventilation rate and body weight might differ from available dosimetry models as exposure models only use an average ventilation rate and typical body weight as the normalizing factor to express an “inhaled dose”. Only the ConsExpo model addresses different size metrics (e.g., surface area and particle number) for average exposure estimates in a similar fashion.

Inhalation of aerosols from consumer spray products has been studied using a range of exposure models such as the consumer exposure model (CEM), the European Center for Ecotoxicology of Chemicals targeted risk assessment model (ECETOC TRA), SprayExpo, ConsExpo and ConsExpo Nano (Park et al., 2018; Williams et al., 2010). Park et al. (2018) compared these different exposure models with respect to translation to internal deposition with predictions from dosimetry models such as that of the International Commission on Radiological Protection (ICRP, 1994) and the Multiple-path particle dosimetry (MPPD) model (ARA, 2016). A USEPA version of the MPPD model was recently peer reviewed for deployment in Agency assessments and will be available in early 2022 (FRN #: 10020-77-ORD.). These dosimetry models (see next section) can predict regional or local internal deposited and retained doses over time and account for various physiological factors and physicochemical properties. For example, the MPPD accounts for differences in breathing mode (nasal, oronasal or mouth), breathing rate and airway architecture with life stage as well as physicochemical properties of hygroscopicity, polydisperse and multi-modal distributions, and clearance mechanisms including solubility.

Park et al. (2018) conducted short term exposure scenarios (10, 30 and 120 min.) in a classroom setting. Total average inhaled dose predicted by the exposure model was compared to the experimental data to determine the difference between the experimental data and modeled estimates. Under these exposure assumptions, the exposure model was able to closely estimate short-term inhalation doses (i.e., up to 10 min) from consumer spray products. However, the model deviated significantly from the data for longer-term experimental estimates (Park et al., 2018). In a similar manner, an evaluation of two consumer exposure model simulations with experimental data, resulted in significant deviations due to incomplete air mixing (Delmaar and Meesters, 2020). Modeling discrepancies were also partly attributed to a lack of data on spatial and temporal changes in mass and size distributions (Delmaar and Meesters, 2020). Uncertainties in particle size distributions are also reported for 3D printer generated emissions. Additional differences between exposure factors for 3D printing processes and aerosol sprays are also expected. These differences involve fixed vs. mobile location of the emission sources (i.e., spray can vs. printer) as well as the expected release rates and periods. Further, input parameters such as chemical properties of particles are not so clear for 3D emitted particles compared to consumer spray products. Yet, these aerosol spray models may provide reasonable first-approximation exposure estimates for a short-term scenario using input parameters that are relevant to the 3D printer emissions.

Some of the limitations in exposure modeling discussed above were recently addressed by MacCuspie et al. (2021) who used a workflow incorporating computational fluid dynamics (CFD) modeling to characterize temporal and spatial heterogeneity of nanoparticle emissions from 3D printers. The model predictions were in good agreement with measurements (MacCuspie et al., 2021). This workflow highlighted the need to characterize the complexity and heterogeneity of emissions to quantify the distribution of particle concentrations within a room and would provide needed input data to similar CFD models of inhaled particles and deposition in the respiratory tract (MacCuspie et al., 2021).

Understanding the exposure concentration, particle density and size distribution over time is required to properly model exposures associated with 3D printing. These parameters are also necessary inputs to dosimetry models that predict potential internal doses within the respiratory system. An understanding of the distribution of the VOC within the aerosol (e.g., as a separate component or adhered to the particles) is also necessary. It is the internal dose that ultimately best predicts risk for aerosol particles and VOCs. An exposure model which accurately estimates temporal and spatial variations of all 3D printer generated emissions, has yet to be described. Nevertheless, several reports have focused on printing exposures in occupational settings (Byrley et al., 2019; Chýlek et al., 2019; Floyd et al., 2017; Mendes et al., 2017; Poikkimaki et al., 2019; Stabile et al., 2017; Stefaniak et al., 2019; Stephens et al., 2013; Vance et al., 2017). In addition to general exposure estimates, Azimi et al. (2017) reported exposure predictions to particles and VOCs for a range of desktop printer-filament combinations using published data. The use of the modeling platform, CONTAM allowed significant versatility applied to modeling a multi-room work facility. In addition, these authors reported factors for exposure to several polymer types and printer models as well as control strategies for reducing exposures to both particles and VOCs (Azimi et al., 2017). Compared to occupational settings, however, consumer settings such as residential rooms (bedrooms, garages, etc.), schools and libraries present greater uncertainties for exposure assessment due to differences in key input parameters including low and variable ventilation, room configuration and operator and bystander position relative to the printer over time.

For studies in occupational settings, the sampling inlet is typically placed at a defined location with respect to the printer, and room conditions are typically defined for the purpose of the exposure estimation. The particulate emission rates, however, also depend on variable factors such as polymer type and manufacturer, additives, print temperature, etc. (Alberts et al., 2021; Azimi et al., 2016; Davis et al., 2019; Gu et al., 2019; Kim et al., 2015; Poikkimaki et al., 2019; Stefaniak et al., 2017b; Yi et al., 2016; Zhang et al., 2017). In addition, the particle populations show significant spatial and temporal variation due to transport and transformation processes of particle formation, aggregation and possible settling or adherence to surfaces. Understanding intrinsic factors related to the printer process (polymer type, polymer additives, printer, print temperature, print time, etc.) and extrinsic factors such as room configuration, ventilation, operator proximity, etc., emissions exposure assessments are in early stages of development. Consequently, due to uncertainties in multiple factors associated with 3D printer particulate emissions, exposure estimates for these processes are limited.

In addition to occupational studies that simply measure particulates, several initial studies have reported metal contents in the particles. In a study of 23 printers and four occupational settings with a variety of filament feed stocks, Stefaniak et al., 2019) reported exposures to Al and Fe, which fell below the National Institute for Occupational Safety and Health (NIOSH) recommended exposure limits. Exposure levels for these metals yielded no adverse effects for an occupational setting which used engineering controls that were commensurate with industrial operations rather than for residential settings. Nevertheless, for 3D printers and printing pens sold as children’s toys, Yi et al. (2019) suggested that emissions, particularly in poorly ventilated residential rooms contain trace levels of metals which may be of health concern.

As discussed above, a wide range of studies are available that could serve as foundations for the development of a more comprehensive exposure model. Nevertheless, few reports have focused on the characterization of filament additives that may carry over into SVOCs that are likely to form the particulate emissions. The effect of chemical composition and additive percentages on emissions concentrations and compositions are other crucial information that needs to be quantified for the development of a universal model specific to 3D emissions.

5.2. Dosimetry Modeling

Although comprehensive reviews and dosimetry models for inhaled particle deposition and retention of exposures to particles are available (Asgharian et al., 2001; Jarabek et al., 2005; Koullapis et al., 2020; Kuempel et al., 2015; Millage et al., 2010), reports concerning application of dosimetry models for 3D printing processes are limited (Byrley et al., 2021; Stabile et al., 2017; Stefaniak et al., 2018; Yi et al., 2019; Youn et al., 2019). In general, particle deposition along the respiratory system is dependent on physicochemical properties of particles such as size, distribution, density, hygroscopicity, and shape; and these interact intimately with physiological parameters including the airway (geometry) and airflow (ventilation) within the respiratory system. Ventilation rate, composed of tidal volume and frequency, can vary dramatically with activity pattern (e.g., increased ventilation rate with exertion). Breathing mode, for example by the nose or mouth, is also an important determinant. The anatomical and ventilatory characteristics can vary considerably across age, sex, and race. Retained dose is influenced by clearance mechanisms including physical clearance by the mucociliary escalator, as well as translocation to the interstitium or dissolution.

For translation of exposure to internal dose estimates, these parameters can be used as input to dosimetry models such as the MPPD to predict internal deposition in various regions of the respiratory tract (Manojkumar et al., 2019). Retained dose metrics can also be computed. Standard values for anatomical and physiological parameters can be obtained from sources such as the International Commission on Radiological Protection (ICRP, 1994) or the USEPA exposure factors handbook (USEPA, 2011). More localized predictions can be obtained from computational fluid dynamics (CFD) dosimetry models (Garcia et al., 2015; Kabilan et al., 2016; Schroeter et al., 2013; Schroeter et al., 2006) sometimes referred to as computational fluid-particle dynamic (CFPD) models when applied to particle exposures (Talaat et al., 2019). Due to technology constraints on imaging that underlies rendering of computational meshes, most CFPD models are reliable only for the nasal region and major tracheobronchial branches but multi-scale models are emerging that will allow accurate full respiratory tract modeling for particles (Kuprat et al., 2021). However, reliable characterization of the physicochemical properties required as input is necessary to use any of these dosimetry models.

Application of dosimetry models requires exposure input data such as particle size characterization (mass, surface area or number), geometric standard deviation, density, and concentration. Because exposure parameters such as particle size, shape and concentration change with time and proximity to the printer, these uncertainties will be inherent in any dosimetry modeling efforts. A spherical shape is typically assumed, but other forms such as fibrous or nanotube can be considered by changing the aspect ratio parameter (typically setting it to greater than three) and specific shape parameters to help refine predictions are possible (Asgharian et al., 2018).

Several dose-modeling studies have been reported for 3D printing processes (Byrley et al., 2021; Stabile et al., 2017; Stefaniak et al., 2018; Yi et al., 2019; Youn et al., 2019). Studies have reported relative doses for various regions of the respiratory track including the nasal or extra-thoracic, tracheobronchial, and pulmonary regions (Byrley et al., 2021). Additional output parameters such as respiratory generation numbers (Youn et al., 2019) as well as simulations for sex and age categories (Byrley et al., 2021; Yi et al., 2019) have also been reported. Although informative, interpretation of model predictions has inherent uncertainties due to lack of particle-specific information such as spatial and temporal-dependent particle size distribution, particle composition, solubility, surface reactivity, or leaching of additives. It is expected that dosimetry modeling will advance with improvements in more situationally relevant exposure assessments for input data as well as information concerning particle interaction with target tissues. In addition to better clarification of exposure parameters, characterization of the physicochemical properties is a clear research need to advance these predictions and inform potential toxicity.

6. Potential Toxicity

6.1. Potential Involvement of Metals in 3D Printer Particle Emission Toxicity

Although particles emitted from 3D printing processes vary in average size depending on the polymer, additives, and print conditions, they are primarily in the nano-size range (Poikkimaki et al., 2019; Stefaniak et al., 2017b) which also allows them to be potentially respirable. In addition, many of these emitted particles are also associated with toxic metals. Evaluation of potential toxicity from inhaled particles from 3D printers must consider impacts from emerging evidence in both the nanomaterials literature and from the air pollution literature regarding ultrafine ambient aerosols (Costa et al., 2019; Stone et al., 2017; USEPA, 2019). As shown in Panel C of Fig. 1, these impacts can include direct toxicity of the particles to the respiratory tract as well as translocation of either the particles or circulating cytokines and inflammation from epithelial perturbation to the brain or other systemic tissues including their mitochondria (Li et al., 2003). Consideration of the interplay between toxicity of the particle per se and any metal moiety reactions is requisite (Efremenko et al., 2017). In addition, while much of the research to date concerns potential health risks in occupational settings (Chen et al., 2020; Mohammadian and Nasirzadeh, 2021), it is also important to consider non-occupational settings such as schools, libraries, and home settings where exposed individuals may be children or other individuals of heightened susceptibility (Byrley et al., 2021).

Depending on their composition and size, non-soluble nanoparticles can cause significant damage to epithelial or other cells by perturbation of the cell membranes followed by inflammation, oxidative stress and organelle injury (Buzea et al., 2007). The toxicity of the metal moiety per se and contribution to the toxicity of inhaled particles depends on their release from a particulate carrier which is often determined by either an acid dissociation process at low pH (such as in macrophages) or by bond formation with an organic compound (Fang et al., 2017). Toxic metals associated with particulate matter (e.g., PM₁₀, PM_2.5, PM < 1) have been shown to enter systemic circulation primarily through the respiratory system and to a smaller extent through the gastrointestinal system (Fortoul et al., 2015; Stone et al., 2017; USEPA, 2019). Particle bound heavy metals typically associated with fossil fuel combustion products (motor vehicle and diesel), industrial processes, mining activities, wood stove burning, and cigarette smoke have been extensively studied. Inhalation exposure to metals such as Fe, Ni, V, Cr, Cu, Zn, Al, Mn, Ca, Pb and Cd associated with particles have been linked to a wide range of adverse health outcomes such as damage to cardiovascular, respiratory, immunological, metabolic, and neurological systems (Fortoul et al., 2015; USEPA, 2019). Exposure to Cr, Cd and Mn have also been associated with adverse neurological effects in children (Caparros-Gonzalez et al., 2019; Hessabi et al., 2019; Yi et al., 2019). Although many of these metals have been detected in 3D printer emissions, (Table 3), their potential transport into systemic circulation or potential involvement in toxicological effects has not been explored.

The emissions from laser printers, like those from 3D printers, contain a complex mixture of VOCs, particle-phase organic compounds and transition metals such as zinc, chromium, nickel, iron, titanium, and aluminum (Carll et al., 2020). Carll et al. (2020) exposed rats by inhalation to the emission products of laser printers, which were present at approximately 500,000 particles / m³ (particle number count) or 71.5 μg/m³ (mass concentration) repeatedly for 21 days. The particles were approximately 44–49 nm in mean diameter. The emissions also contained volatile organic gases with a total VOC concentration ranging from 245 to 363 parts per billion (ppb). After exposure the rats showed alterations in multiple parameters of cardiac function including increased ventricular pressure, decreased heart rate variability, arrhythmia, and hypertension. These and other changes suggested augmented sympathetic tone, impaired ventricular function, and increased risks of cardiovascular disease, but the relative contribution of components was not characterized.

There is an extremely limited database for studies of toxicity from 3D printer emissions (Table 5). While some toxicity studies of 3D printer emissions as well as studies with leachates from printed objects have been reported, many have been from in vitro assays or acute in vivo exposures and most with only one exposure concentration precluding dose-response analysis. There are also some human data including a case report, clinical study, and an occupational health survey.

Farcas et al. (2019), using an in vitro submerged culture of human small airway epithelial cells (SAEC), demonstrated that emissions from PC and ABS resulted in dose-dependent cytotoxicity, oxidative stress, apoptosis, necrosis, and production of pro-inflammatory cytokines and chemokines. Similarly, epithelial cells and macrophages exposed to ABS and PLA emissions exhibited toxicity such as cell death, oxidative stress, and inflammatory responses (Zhang et al., 2019). Zhang et al. (2019) also observed an oxidative potential of ABS/PLA emissions using a cell-free dithiothreitol assay.

Farcas et al. (2020) subsequently conducted a repeat whole-body inhalation exposure study to ABS in male Sprague Dawley rats, 4 h/day for 4 days / week for up to 30 days. The exposure concentration was 240 ± 90 mg/m³ with an average geometric mean particle mobility diameter of 85 nm and GSD of 1.6. Exposure to ABS emissions resulted in increased levels of IFN-y and IL-10 in bronchoalveolar released lavage cells as well as histopathological changes in lungs and nasal passages. Exposure to emissions from ABS, PLA and nylon also resulted in strong inflammatory response and increase in neutrophils number in mice. Similarly, Stefaniak et al. (2017a) reported rodent inhalation exposures to ABS emissions resulted in compromised macrovascular function compared to control animals.

In addition to evaluating toxicity of airborne emissions, there has also been interest in potential toxicity of the printed objects, particularly when synthesizing objects for medical or dental applications (Inoue and Ikuta, 2013; Macdonald et al., 2016; Popov et al., 2004). Zebra fish embryos exposed to 3D print object leachate from a photoreactive resin resulted in the generation of reactive oxygen species and lipid peroxidation as well as enhanced activity of superoxide dismutase and glutathione-S-transferase (Walpitagama et al., 2019). The toxicity of objects created from polymerizing acrylate or methacrylate monomers can be attributed by the authors to the amount of unpolymerized monomers that leach out in aquatic media (Oskui et al., 2015; Xu et al., 2021).

A human occupational health and exposure study reported upper and lower respiratory tract symptoms in workers who operate 3D printers in an occupational settings (Chan et al., 2018). A questionnaire based survey on exposure to emissions from printing PLA, ABS and nylon reported respiratory symptoms, nasal congestion, rhinorrhea, cough, itchiness of nose throat or eyes of workers (Chan et al., 2018). In a well-conducted clinical study, 26 adults sitting 40 cm from a 3D printer to test the effect of emissions from a 1-hour operating exposure to ABS and PLA printing indicated 81 μm²/cm³ and 7 μm²/cm³ lung deposited surface area (LDSA) values respectively (Gumperlein et al., 2018). This short exposure increased cytokine levels (IL-1β, IL-6, TNF-α) in nasal secretions and an increase in exhaled nitric oxide during exposure to emissions from ABS printing (Gumperlein et al., 2018). In addition, the level of pro-inflammatory cytokine IL-6 was shown to change from 38 to 68 pg/mL and from 27 to 89 pg/mL for use of ABS and PLA, respectively, however, no clinically significant acute health effects were observed (Gumperlein et al., 2018).

The presence of redox active metals, even at low concentrations may exert toxic effects on biological systems resulting from the production of reactive oxygen species (ROS) (Egorova and Ananikov, 2017; Jomova and Valko, 2011; Tchounwou et al., 2012). Because ROS and associated biological indicators such as lipid peroxidation and increased activity of superoxide dismutase and glutathione-S-transferase, have been reported following cellular exposure to 3D printer emissions, the relationship between metals in particulate emissions and toxicity warrants further investigation.

Exposure studies have reported trace concentrations of metals associated with bulk filaments and emitted particles (Farcas et al., 2019; Stefaniak et al., 2019; Stefaniak et al., 2017b; Steinle, 2016; Vance et al., 2017; Yi et al., 2019; Youn et al., 2019; Zontek et al., 2017) (Table 3). A few of these studies have reported polymer type, brand, color and observed metals associated with the particulate emissions from specific filaments (Farcas et al., 2019; Stefaniak et al., 2017b; Steinle, 2016; Yi et al., 2019; Youn et al., 2019) (Table 4). However, those studies that report metals in particulate emissions did not focus on toxicity assays (Farcas et al., 2019; Stefaniak et al., 2017a; Steinle, 2016; Yi et al., 2019; Youn et al., 2019) and those reporting toxicity results did not report relationships between metals in the emissions and toxicity (Chan et al., 2018; de Almeida Monteiro Melo Ferraz et al., 2018; Farcas et al., 2020; Gumperlein et al., 2018; House et al., 2017; Oskui et al., 2015; Stefaniak et al., 2017a; Walpitagama et al., 2019; Zhang et al., 2019).

As discussed above, the toxicity reports in FFF emissions are not specifically attributed to the presence of metals. This is not necessarily surprising given the significant number and possible combinations of polymer types, colors, and associated metals. It is also interesting to note that common colors such as blue, red, orange, and black tend to contain the most variety of metals. They are not, however, consistent among brands even for similar colors (Table 4). Another informational gap in our understanding of metal containing polymer filaments involves the metal speciation and potential for leaching or metabolism of particulate emissions and print objects. One of the challenges for toxicological evaluation of 3D printer emissions is that there are two basic components (particles and vapors) that may interact in the exposure atmosphere as well as contribute in unique ways to observed toxicity at the cellular, tissue and organism levels. Careful consideration of the relative contribution of these components will be important to characterize and aid understanding of exposure, dose and risk (Price et al., 2020). Because the effects of these components may be additive or synergetic, it is difficult to determine the relative contribution to exposure, dose, and risk.

Dosimetry modeling results of Byrley et al. (2021) showed potential persistence of particles in the pulmonary region of the lower respiratory tract where slow mechanical clearance opens the possibility for accumulation of particles if there is repeated exposure. Solubility may mitigate these doses, however, persistence of particles will induce inflammation (Nassan et al., 2021), macrophage responses, and release of acidic macrophage fluids. Although metals concentration in the emissions is low, extended exposure accompanied with a release of acidic macrophage fluids may lead to an increased dissolution of metals from the particles. Increased dissolution could lead to more metals being translocated across the air/blood interface and higher systemic concentrations of circulating toxic metals.

6.2. Possible Nasal-olfactory uptake

Toxicology studies related to FDM emissions mostly targeted bronchial, pulmonary, or cardiovascular functions (Table 5). However, another potential pathway of exposure that could raise health concerns involves exposure to the brain via size-dependent deposition of inhaled particles in the respiratory and nasal olfactory epithelium, and subsequent uptake along the trigeminal or olfactory nerves and transport into the brain (Fig. 1). Unlike particles deposited into the bronchial or pulmonary regions, which must be taken up into the blood, transported to the cerebral arteries, and cross the diffusion limiting blood-brain-barrier to enter the brain tissue, particles traveling along the trigeminal or nasal olfactory pathways can be taken up directly into the brain tissue.

Of these two cranial nerves, the olfactory pathway has received greater attention. A variety of materials has been observed to be transported into the brain via nasal olfactory uptake including viruses, prion proteins, chemicals, including toxic soluble metals, quantum dots and other nanoparticles (Brenneman et al., 2000; Elder et al., 2006; Hopkins et al., 2018; Hopkins et al., 2014; Kincaid et al., 2012). The parameters determining nasal olfactory uptake have become a focus for development of drugs for CNS treatment. Particle size, coating and surface charge all influence nasal olfactory uptake. Smaller particles are taken up more readily than larger particles, coatings with organic materials such as polyethylene glycol (PEG), peptides or curcumin enhance uptake, and positively charged particles are taken up more than negatively charged or neutral particles (Bonaccorso et al., 2020; Fechter et al., 2002; Lin et al., 2016). Both intracellular and extracellular pathways have been proposed for particle uptake and transport along the olfactory nerve including: (1) endocytosis of deposited particles into olfactory sensory neurons and axoplasmic transport within the neurons to the olfactory bulb, from there trans-synaptic movement to other CNS neurons and further distribution across the brain has been observed for some materials; and (2) absorption into extracellular fluid in the lymphatic/perivascular/perineural spaces and diffusion into the brain extracellular fluid compartment (Lochhead and Thorne, 2012; Mathison et al., 1998). Beyond particle uptake, relatively little information is available regarding particle clearance from the brain. More information is needed on rates of both particle uptake and clearance from the brain to better understand the potential for brain bioaccumulation of particles or the metal components of particles.

Exposure of the brain to nanoparticles traveling via nasal olfactory uptake is of interest not only for the development of potential CNS therapeutics, but is also a concern for adverse effects from exposure to air pollution particles (Block and Calderón-Garcidueñas, 2009; Costa et al., 2017; Costa et al., 2019; Oberdörster et al., 2009). Inflammatory and/or neurodegenerative changes have been observed in the brains of dogs and humans living in polluted parts of Mexico City relative to those from regions with lower air pollution (Calderón-Garcidueñas et al., 2002; Calderón-Garcidueñas et al., 2003; Calderón-Garcidueñas et al., 2004). Particles having a shape consistent with combustion generation, such as in automobile engines or in power plants, were identified in the autopsied frontal cortex of people from both Mexico City and Manchester England (Maher et al., 2016). Epidemiological studies have shown correlations between air pollution particle concentrations and increased risks of neurodevelopmental or neurodegenerative diseases (Jayaraj et al., 2017; Rivas et al., 2019; Sunyer et al., 2015). The USEPA’s, 2019 Integrated Science Assessment of the health risks of exposure to ambient air particulate matter concluded that there was likely to be a causal link between PM_2.5 exposure and nervous system effects (USEPA, 2019). More recently, a nationwide survey of US Medicare records combined with the spatial distribution of PM_2.5 concentrations showed a low-dose linear function between PM_2.5 exposure and increased risk of chronic debilitating brain disorders (dementia, Parkinson’s disease), which showed no mathematical threshold, suggesting that any level of exposure increased risks (Yitshak-Sade et al., 2021).

Particles generated by 3D printers are within the size range of particles deposited into the respiratory and olfactory epithelium (Poikkimaki et al., 2019). Many of the metal constituents of 3D printer filaments and 3D printer particles identified in Tables 3 and 4 can be neurotoxic, and/or can induce inflammatory responses. In addition, particles themselves can generate inflammatory responses regardless of the metallic content. For these reasons, nanomaterial-sized 3D printer generated particles should be considered to have potential for deposition into the nasal olfactory epithelium, uptake into the brain and, potentially, the generation of inflammatory or toxic reactions there. It is important to note, however, that to date we are not aware of any experimental investigations of the uptake or transport of 3D generated particles by the nasal olfactory route or other pathways leading to the brain.

6.3. Regulatory Guidance or Controls

We are aware of no current occupational exposure limits or other regulatory guidelines or standards specifically related to the emissions from 3D printers. There are, of course, standards for many components of 3D printer emissions such as single VOCs, particulate matter, and individual metals (ATSDR, 2021; NIOSH, 2020b; OSHA, 2021; USEPA, 2021b). In comparing any given exposure situation to health standards, it is important to consider the appropriate metric since occupational standards are designed for adult healthy workers assumed to work a standard 40-hour work week. School or home users, for example, may not fit this description in age, health status, or exposure patterns, and therefore occupational exposure limits may not be appropriate. Environmental exposure guidelines, such as those set by the EPA typically do consider children and other susceptible members of the population but may also assume exposure scenarios reflecting outdoor ambient air exposures including that exposures occur 24 h/day, 7 days/week for a lifetime, which also may not fit the profile of non-professional 3D printer users. Acute Exposure Guideline Levels (AEGLs) consider short term exposure durations, but are intended to prevent human health effects of varying severity in the case of rare or once-in-a lifetime exposures to airborne chemicals and are intended for use by emergency responders dealing with chemical spills or other catastrophic exposures (USEPA, 2021a) and therefore also may not be suited for non-occupational use of 3D printers.

In the absence of legal or official exposure standards, informational materials providing guidance for safely operating 3D printers have been issued by reputable organizations including the National Institute of Occupational Safety and Health, (NIOSH, 2020a), and the Consumer Product Safety Commission (CPSC, 2020) In essence, these guidance documents suggest enclosing the printers in a cabinet or other enclosure that is exhausted to the outdoors, providing adequate room ventilation, reducing proximity of individuals to the printers, using filaments with lower rates of emissions, and other measures to reduce exposure concentrations and durations. Please note that devices printed with 3D printers, such as medical or dental materials, toys, or weapons, may be regulated separately under the appropriate authorities. By following the recommendations for safely operating 3D printers, it should be possible to greatly reduce exposures during printer operations and the associated potential health risks.

7. Summary & recommendations to address data gaps

Non-polymer additives such as colorants, flame retardants and metals which are included in 3D printer filaments are not often well characterized with respect to emissions studies in this emerging field of research. Here we address potential exposures in consumer focused FDM 3D printing, with emphasis on metals present in filaments, emissions, and print objects. Given the increasing variety of materials such as metals (e.g., copper bronze and steel), organic and inorganic colorants, wood, and nanomaterials (carbon fibers, etc.) added for aesthetic purposes, particularly in consumer products, there is a growing need to characterize the exposure and effects of these materials. Below we summarize and suggest potential data gaps:Although trace levels of metals have been detected and, in some cases, measured in emissions from FDM 3D printing, metal speciation, leaching and the potential for redox cycling have not been reported.

• Although trace levels of metals have been detected and, in some cases, measured in emissions from FDM 3D printing, metal speciation, leaching and the potential for redox cycling have not been reported.

• Correlations among filament color, reported metals in filaments and particulate emissions are also limited. Further, the relationships among filament colors and toxicity have yet to be characterized.

• Metal containing filaments, added for aesthetic purposes, are reported to emit higher concentrations of particles compared to similar base material filaments. These commercially available “metal fill” filaments typically contain high amounts (10–80%) of micrometer-sized metal particles. The composition of the particles emitted from these filaments including the amount and forms of metals, however, has not been clearly elucidated.

• Due to the increased availability of filament extruder equipment, fabrication of custom filaments is becoming more prevalent among consumers. Reports for custom extruded filaments which incorporate metals or metal oxides have mainly focused on enhancing the mechanical properties of filaments and print objects. Reports concerning the effects of additives such as metals, wood, glitter, etc. on potential exposures to both extruder emissions and printer emissions for custom filaments are also currently limited.

• Particles emitted from 3D printing processes show a high degree of spatial and temporal heterogeneity with respect to size and shape. Emissions are affected by printing conditions such as print temperature, extrusion rate, filament polymer, color, etc. As a result, exposure modeling has been, for the most part, limited to a single individual exposure (monitoring position) and specific printing factors over the printing time course. Since one or more users may be moving with respect to the emission source over the printer operation period, mapping the dynamic room variables may provide a more relevant exposure scenario. These limitations also impact dosimetry modeling which requires exposure input parameters such as relevant particle density and size distribution over a time course.

• Exposure and survey studies for 3D printing operations have been primarily focused on occupational settings. Comparatively little attention has been paid to how children and teenagers might interact with low-cost 3D printers and printer pens in residential and school settings. Given that critical determinants of inhaled dose such as airway anatomy and ventilation vary considerably with age. Characterizing exposures to children and teenagers represents a key data gap. Consequently, we suggest that more consumer-based studies should be conducted, particularly in home and school settings. Prior to results of these studies, the precautionary principle would indicate the use of particles masks or printer enclosures. We also suggest evaluating the use and characterization of specialized filaments that are likely to contain additives such as metals and nanomaterial.

• Available toxicity studies on FFF emissions, even though are very limited compared to the variability of filaments used by consumers, focus on bronchial and pulmonary regions or on the assessment of the effect of emissions on cardiovascular functions. However, the presence of millions of nanoparticles in those emissions and the fact that metals have been reported to be associated to those nanoparticle emissions suggests the importance of further investigation on translocation from the nasal region and contributions of inhaled nanoparticles to systemic toxicity.

Acronyms

ABS

Acrylonitrile Butadiene Styrene

AChE

acetylcholinesterase

ASM

Aerosol Spectrometer

BALF

Broncho-alveolar lavage fluid

CNS

Central Nervous System

CNT

Carbon Nanotube

CPC

Condensation Particle Counters

CPFD

Computational Fluid Particle Dynamic model

EDS

Energy Dispersive x-ray Spectroscopy

ELPI

Electrical Low-Pressure Impactor

ELS

Electrophoretic Light Scattering

FDM

Fused Deposition Modeling

FeNO

fractional exhaled nitric oxide

GST

glutathione-S-transferase

HTPLA

High Temperature Polylactic Acid

ICPOES

Inductively Coupled Plasma optical emission spectroscopy

ICRP

International Commission on Radiological Protection

IFWγ

Interferon gamma and interleukin 10

LDSA

Lung deposited surface area

LDSA

Lung deposited surface area

MPPD

Multiple path particle dosimetry model

NIOSH

National Institute for Occupational Safety and Health

NTA

Nanoparticle tracking analysis

PC

polycarbonate

PCL

poly (e-caprolactone)

PET

polyethylene terephthalate

PETG

polyethylene terephthalate glycol

PLA

polylactic acid

PVA

polyvinyl alchol

ROS

Reactive oxygen species

SAEC

Small airway epithelial cells

SEM

Scanning electron microscope

SMPS-CPC

Scanning mobility particle sizer-optical particle sizer

SOD

superoxide dismutase

SVOC

semi-volatile organic compound

TEM

Transmission electron microscope

TNF-α

Tumor necrosis factor alpha

VOC

Volatile organic compound