Indoor Air Quality Implications of Additive Manufacturing in Hybrid Workplaces

Abstract

Additive manufacturing (AM) technologies are increasingly being integrated into work environments, making the coexistence of production and office activities within hybrid workplaces more common and raising concerns regarding indoor air quality (IAQ) in shared or adjacent workplace environments. Although emissions associated with AM processes have been examined, limited attention has been given to IAQ conditions in hybrid workplaces where production and office functions share the same indoor environment. This study investigates the IAQ conditions associated with selected powder bed fusion and vat photopolymerization operations in a hybrid workplace through real-time environmental monitoring and employee perception analysis. Field experiments were conducted during pre-printing, printing, post-printing, and post-processing phases. Two custom-designed instruments, the Indoor Network-Based Air-Sensing Ecosystem (INBASE) and the Indoor Air Quality Polling Instrument (AIRPOLL), were developed to monitor particulate matter, volatile organic compounds, formaldehyde, benzene, ethanol, nitrogen dioxide, carbon dioxide, and oxygen and collect employee feedback. Findings showed that pollutant concentrations varied by process type, operational phase, and monitored space. Higher concentrations, particularly of volatile organic compounds, were observed in production spaces, while most other parameters remained within reference values. Employee feedback indicated positive perceptions in office spaces and greater concerns in production spaces. The findings support the importance of ventilation, spatial separation, and continuous monitoring in AM-integrated hybrid workplaces.

1. Introduction

People spend a significant portion of their time indoors [1], with workplaces playing a central role in daily life. The transformation of workplaces with Industry 4.0 [2] has introduced novel manufacturing techniques such as additive manufacturing (AM), an emerging technology that fabricates three-dimensional objects through an additive process [3]. Unlike traditional manufacturing processes, AM enables flexible, small-scale production without molds or large facilities, making it feasible to integrate production activities within office environments [4]. This shift enhances operational efficiency but also raises concerns about indoor air quality (IAQ) due to the emission of pollutants [5]. The integration of additive manufacturing into office environments introduces a hybrid workplace concept in which production and non-production functions coexist in the indoor ecosystem. Unlike conventional industrial facilities, where manufacturing activities are spatially separated from administrative areas, hybrid AM workplaces may involve adjacent zones, shared circulation paths, and overlapping occupancy patterns. These conditions may create IAQ concerns as pollutants generated during AM operations may extend beyond production zones and influence the environmental conditions experienced by non-production occupants. Consequently, IAQ assessment in such environments requires consideration of process-related pollutant variations within production spaces and potential changes in office spaces.

The nature and intensity of AM emissions depend on the process type, operational phase, and working environment [6], and the associated hazards, including inhalation of fine and ultrafine particles, exposure to volatile organic compounds, and dermal exposure to powders, binders, resins, solvents, and radiation, are well-documented [7,8]. Despite the expanding use of AM applications and the growing body of literature in this field, existing studies have predominantly focused on isolated laboratory or production settings [9], with emphasis on particle and VOC emissions [10,11], material-related exposure [10,12], occupational hygiene [6,9,13], and ventilation or control strategies [14,15,16]. Additionally, some studies have investigated the potential transfer of pollutants [9,13,17], phase-dependent changes in IAQ [18,19], and employee perceptions [17,20]. However, despite this initial focus, limited attention has been paid to hybrid workplaces where office and production activities coexist within indoor environments [21]. As AM moves from industrial facilities into smaller, mixed-use workplaces, further research is needed to understand how IAQ conditions vary across production and office spaces and whether conventional spatial separation and ventilation strategies are sufficient for these settings. To this end, this research investigates the IAQ conditions associated with selected AM operations in an operational hybrid workplace by combining real-time environmental monitoring with employee perception feedback. Two custom-designed instruments were developed and deployed to monitor IAQ variations associated with powder bed fusion (PBF) and vat photopolymerization (VP) operations across the monitored operational phases in both production and office spaces. The contribution of the present study is threefold. First, it develops and implements two tailored, low-cost IAQ assessment instruments for AM-integrated hybrid workplaces. Second, it evaluates real-time IAQ conditions together with employee perceptions in an operational AM-integrated hybrid workplace. Third, it provides a case- and phase-based comparison of monitored pollutant concentrations across separately monitored production and office spaces during selected AM operations.

The rest of the paper is structured as follows. Section 2 reviews the relevant literature. Section 3 details the methodology and the field experiments, whereas Section 4 presents the results and discusses the findings. The paper is concluded in Section 5 by summarizing the findings, limitations, and future research directions.

2. Research Background

2.1. Integrated Production–Office Spaces

Although work spaces such as office spaces and production spaces [22] were historically distinct entities, these divisions have begun to blur with the advent of Industry 4.0, which accelerates methods like advanced manufacturing. AM enables production in flexible, multipurpose, or smaller spaces, increasing the overlap between office and production environments and driving their convergence.

2.2. Additive Manufacturing

Additive manufacturing (AM) is an emerging manufacturing field in which three-dimensional objects are fabricated through additive processes [3], enabling complex geometries at lower cost, weight, and time [4]. According to the American Society for Testing and Materials (ASTM), AM processes have seven categories [23,24]: binder jetting, directed energy deposition, material extrusion, material jetting, powder bed fusion, sheet lamination, and vat photopolymerization [6,24,25], each involving distinct operations for pre-printing, printing, post-printing, and post-processing stages. This study focuses on powder bed fusion and vat photopolymerization, which were selected based on their availability for monitoring at the workplace under real operating conditions, their relevance to fabrication through different materials, and their distinct emission profiles.

2.2.1. Powder Bed Fusion

In powder bed fusion (PBF), powder-based materials are solidified using a heat source, offering high dimensional accuracy and desirable mechanical properties. PBF technologies include electron beam melting (EBM), direct metal laser sintering (DMLS), selective laser sintering (SLS), and selective laser melting (SLM) [24].

In PBF, pre-printing involves powder preparation and loading into the printer. Printing consists of repeated powder spreading and selective melting or sintering until the build cycle is completed. During post-printing, the printed component is retrieved, the surplus powder is removed, and used powder is sieved and replenished. During post-processing, the printed part is detached from the build platform, and necessary finishing treatments are applied [6].

2.2.2. Vat Photopolymerization

In vat photopolymerization (VP), a liquid photopolymer is selectively cured into a solid state in a vat through light-induced polymerization [25]. VP technologies include digital light processing (DLP), liquid crystal display (LCD), and stereolithography (SLA) [25,26].

In VP, pre-printing operations involve resin mixing, dispensing, and vat loading [6]. Printing consists of the layer-by-layer photopolymerization process. During printing in stereolithography (SLA), a scanning laser beam selectively cures resin in the prescribed geometry [25,26]. During post-printing, the printed component is retrieved. Post-processing activities include washing and curing to enhance final part properties. Additional steps include removing supports, grinding, gluing, sealing, painting, polishing, coating, varnishing, sterilizing, inspecting, and measuring [27].

2.3. Indoor Air Quality

Essential for maintaining a healthy indoor environment [28], indoor air quality (IAQ) refers to air free of harmful contaminant levels that satisfies most occupants [29]. The World Health Organization (WHO) reported that 90% of the global population is exposed to inadequate air quality [30]. Given its effects on comfort, productivity, and well-being [31], IAQ represents a major public health concern.

As AM continues to gain widespread adoption, concerns regarding health and safety have emerged [5]. Although all AM processes involve certain risks, the nature and severity of these hazards depend on the utilized technology, process stage, and working environment [7,8,32]. Key health concerns include inhalation of IAQ pollutants, dermal exposure to AM materials, and radiation [7,8]. The relevant IAQ pollutants are summarized in Table 1 with their main characteristics, health relevance, and guideline values.

Table 1. Characteristics, health relevance, and reference values of monitored IAQ parameters.

Parameter	Characteristics	IAQ and Health Relevance	Reference Values Used for Contextual Interpretation
Particulate Matter (PM1, PM2.5, PM10)	PM consists of airborne particles with different sizes and compositions [33]. PM10, PM2.5, and PM1 [34] denote particles with diameters below 10, 2.5, and 1 μm, respectively [35]. It may also contain polymer-based micro- and nanoplastics [36].	Exposure to PM has been associated with respiratory and cardiovascular diseases [37], with different-sized particles accumulating in different regions of the respiratory system.	WHO [38] provides guideline values for PM2.5 and PM10, presented in Table A1. In this study, these values were used as contextual references rather than compliance thresholds because the monitoring durations did not correspond to annual or 24 h averaging periods.
Volatile Organic Compounds (VOC/TVOC)	VOCs are common indoor pollutants with boiling points below 400 °C [39]. Total VOC (TVOC) represents their cumulative concentration [40].	VOC exposure is associated with odor discomfort, sensory irritation, neurological symptoms, and compound-specific severe health effects [37].	The five-level TVOC classification of the Federal Environmental Agency in Germany [41] was used for contextual interpretation of hygiene evaluation.
Formaldehyde (HCHO)	Formaldehyde is a colorless and flammable VOC [42] whose emissions are influenced by thermal conditions, ventilation, building materials, and indoor furnishings [43].	Documented effects of its exposure include neurophysiological symptoms, eye and airway irritation, pulmonary inflammation, pneumonia, edema, coma, and death [44,45].	WHO [46] recommends the guideline value of 0.1 mg/m3 for a 30 min average. In this study, these values were used as contextual references rather than compliance thresholds because the monitoring durations did not correspond to 30 min averaging periods.
Benzene (C6H6)	Benzene is a human genotoxic carcinogen VOC.	Benzene exposure is linked to carcinogenicity, genotoxicity, and hematotoxicity [46], leading to fatigue, headaches, blood and bone marrow disorders, and fatality.	WHO [37] states that no safe exposure level can be defined for benzene.
Ethyl alcohol (C2H5OH)	Ethyl alcohol is a volatile and flammable compound [47].	Inhalation of ethyl alcohol is associated with various irritations in sensory organs and neurological effects [47].	The odor threshold is 84 ppm [48].
Nitrogen dioxide (NO2)	Nitrogen dioxide is a combustion byproduct [37].	The exposure to nitrogen dioxide increases respiratory and allergic risks and blood coagulation [37].	WHO [38] provides guideline values for NO2 presented in Table A2.
Carbon dioxide (CO2)	CO2 is a common IAQ and ventilation indicator. It is mainly released through metabolic activities [49].	CO2 exposure is associated with impaired cognitive abilities, headaches, fatigue, drowsiness, increased arousal, blood acidosis, and respiratory failure [50].	The typical outdoor level is 400 ppm, while poor ventilation is represented by 1100 ppm [50].

2.4. Indoor Air Quality in Additive Manufacturing Environments

While all AM processes present hazards, exposure levels and types vary by process type, phase, and environment [7,8,32]. As this research focuses on PBF and VP in hybrid workplaces, related studies are presented to contextualize the findings.

Studies on IAQ in PBF environments examine emissions from metal and polymer feedstocks across process stages. Accordingly, operational tasks such as powder handling, post-processing, and equipment cleaning increase exposure risks due to the release of ultrafine particles, airborne nanoparticles, and potentially hazardous substances [51,52,53,54,55]. Research on laser powder bed fusion demonstrated that transient emission peaks occur during cleaning and maintenance stages, emphasizing the phase-dependent exposure behavior [56]. For instance, pre-printing and post-processing stages significantly exceed IAQ threshold limits [57]. Similarly, in laser metal deposition (LMD) processes, the release of metal nanoparticles is associated with applied laser power, with powder-based feedstocks releasing considerably higher nanoparticle concentrations than wire-based alternatives [56]. Studies on nylon-based materials have reported elevated dust and respirable particle levels during post-printing [58]. Recent research identifies the polymer powders used in SLS, such as polyamide and polyurethane, as intentionally engineered microplastics [59]. Although these fuse during printing, residual and recycled material can still emit airborne microplastic particles during handling and post-processing [36,59]. Furthermore, the reuse of powders may alter emission characteristics [60,61]. In metal and polymer environments, formaldehyde and particulate concentrations increase, particularly with reinforced powders [62]. Beyond feedstock characteristics, emission behavior is also influenced by process parameters and environmental conditions, such as extrusion temperature, thermal decomposition behavior, filament feed rate, and air exchange conditions [63]. Studies conducted in non-industrial AM environments have demonstrated that indoor particle and VOC concentrations were affected by simultaneous machine operation, inadequate ventilation, and filtration performance [64,65]. While several studies report pollutant levels below occupational exposure limits, improved ventilation, personal protective equipment, and other control measures are recommended [66,67,68]. Continuous monitoring and preventive strategies are also emphasized due to nanoscale particles and metal residues generated during AM operations [69,70,71,72].

Studies investigating IAQ during VP operations highlight emissions from photopolymer resins during printing and post-processing stages. Researchers have shown that exposure risks persist beyond printing due to chemical residues and cleaning agents [73] associated with post-processing tasks involving isopropyl alcohol rinsing, soaking, and drying [74]. Emission modeling studies have further demonstrated that printing and post-curing contribute to VOC release, varying by process parameters and resin type [75]. Investigations comparing different machines, feedstocks, and ventilation setups have reported elevated VOC and PM levels, especially during post-processing [76]. While some studies have reported minimal particle emissions under well-ventilated conditions, common VOC types such as acetone and isopropanol were still detected at measurable levels [58]. To proactively control these emissions, studies indicate that retrofitted activated carbon adsorbents or photocatalytic oxidation systems can reduce TVOC levels [72].

Owing to the impact of AM operations on IAQ, previous studies have proposed various approaches for monitoring and mitigating indoor air pollution in AM environments, including structured environmental management frameworks [64,77], real-time and low-cost monitoring systems [78], and simulation-based airflow and ventilation optimization methods [79].

3. Methodology

This study employs field experiments to examine IAQ conditions associated with operations of PBF-SLM, PBF-SLS, and VP-SLA processes using two custom-designed instruments. The experiments were conducted in a workplace specializing in AM. The 375 m² and 3.7 m high AM facility represents a hybrid workplace in which office and production activities operate within the same indoor environment, as illustrated in Figure 1.

Figure 1. The plan of the hybrid workplace showing the office zone in the blue frame; production zones, AM equipment, and post-processing area in the pink frame; and INBASE/AIRPOLL monitoring locations in the red circles.

The plan documents the spatial configuration of the monitored hybrid AM workplace, including production and office spaces, room areas, circulation zones, doors/spatial connections, AM equipment, post-processing equipment, and INBASE/AIRPOLL monitoring locations. Detailed ventilation and airflow characteristics, including supply and exhaust airflow rates, local exhaust capture velocity, pressure differences, airflow direction, and return-air configuration, were not quantitatively measured. Therefore, these features are treated as contextual limitations.

Field measurements were performed in production and office spaces for each AM process. This configuration enabled the comparison of IAQ parameters in production and office spaces occupied by employees with different work patterns and occupancy durations.

Although the study was conducted in a single operational facility, the research is composed of a structured phase-based monitoring approach, including real-time environmental monitoring and gathering employee perception during selected AM technologies, operational phases, and workplace zones in a hybrid workplace. The experimental setup is illustrated in Figure 2.

Figure 2. The diagram of the field experiments.

Ventilation and spatial characteristics were documented descriptively based on the existing workplace configuration. However, air exchange rates, supply and exhaust airflow rates, local exhaust capture velocity, pressure differences between rooms, airflow direction, door and window opening conditions, shared or separate return-air systems, outdoor pollutant concentrations, and detailed occupancy/activity levels were not measured during the field experiments. Therefore, ventilation and spatial features are treated as contextual workplace characteristics.

3.1. Development of INBASE and AIRPOLL

Two instruments, the Indoor Network-Based Air-Sensing Ecosystem (INBASE) and the Indoor Air Quality Polling Instrument (AIRPOLL), were custom-designed to be used in field experiments.

INBASE, illustrated in Figure 3, is an IoT-based multi-sensor-arrayed real-time IAQ monitor developed to record pollutant-related sensor outputs for PM_1.0, PM_2.5, PM₁₀, TVOC, HCHO, C₆H₆, C₂H₅OH, CO₂, O₂, and NO₂. However, the system does not measure ultrafine particle number concentration, particle-size distribution below 1 μm, particle chemical composition, metal nanoparticles, polymeric micro- or nanoparticles, resin-specific VOCs, or individual solvent compounds using analytical methods. The hardware includes quantitative sensors for PM (HM3301) (Seeed Studio, Shenzhen, China), TVOC (SGP30) (Seeed Studio, Shenzhen, China), HCHO (SFA30) (Seeed Studio, Shenzhen, China), CO₂ (SCD30) (Seeed Studio, Shenzhen, China), and O₂, (ME2-O₂-Φ20) (Seeed Studio, Shenzhen, China); a qualitative sensor (Multichannel V2) (Seeed Studio, Shenzhen, China) for C₆H₆, C₂H₅OH, and NO₂; and a Wi-Fi module (ESP8266) (Seeed Studio, Shenzhen, China), a display unit (SSD1306) (Solomon Systech, Hong Kong), a base shield (Grove Base Shield V2.0) (Seeed Studio, Shenzhen, China), a breadboard (Robolink, Ankara, Turkey), and a development board (Arduino Mega 2560 Rev3) (Arduino S.r.l., Monza, Italy).

Figure 3. Indoor Network-Based Air-Sensing Ecosystem (INBASE).

The software of INBASE was developed using Arduino IDE (Version 2.3.2). The configuration included installing open-source sensor libraries, defining connection methods for each sensor, and entering Wi-Fi credentials, server IP addresses, and API keys. Sensor readings were formatted and transmitted to the ThingSpeak (MathWorks, Natick, MA, USA) server over TCP through the ESP8266 module for real-time storage, visualization, and remote data access. Additionally, when the system was connected via USB, measurements were displayed on the Arduino IDE for local data access.

Sensor calibration procedures were performed in accordance with the manufacturer’s guidelines to maintain accuracy and consistency. The factory-calibrated sensors, including HM3301 and SFA30, were operated by integrating manufacturer demo codes into the configuration code. Sensors requiring user-calibration, including SGP30, Multichannel V2, SCD30, and ME2-O₂-Φ20, underwent different procedures in line with the recommended calibration protocols. For SGP30, a baseline calibration procedure, consisting of initialization and continuous operation for 12 h to store baseline values, was applied. The Multichannel V2 module was recalibrated by using manufacturer calibration curves. By using inverse equations derived from these curves, voltage outputs were converted into concentration values, which were embedded in the configuration code. Additionally, the sensor was preheated for 168 h to reach equilibrium. For SCD30, the Automatic Self-Calibration algorithm was utilized while the sensor was operated continuously for 168 h with daily fresh-air exposure, after which the calculated parameters were stored in memory. ME2-O₂-Φ20 was preheated for 30 min and exposed to fresh air; voltage readings were recorded as reference and converted to percentage concentrations using the manufacturer’s sample code.

The performance characteristics of the sensors were taken from manufacturer specifications. HM3301 provided PM_1.0, PM_2.5, and PM₁₀ concentrations with a 1 µg/m³ resolution. SGP30 reported TVOC values from 0–60,000 ppb with stepwise resolutions ranging from 1–32 ppb. The SGP30 TVOC output was converted from ppb to mg/m³ using ethanol as the reference compound. The conversion followed the ideal gas relationship at 25 °C and 1 atm: TVOC (mg/m³) = TVOC (ppb) × 46.07/24,450, where 46.07 g/mol is the molecular weight of ethanol, and 24.45 L/mol is the molar volume at 25 °C and 1 atm. Dynamic temperature and pressure corrections were not applied. Therefore, the reported TVOC values should be interpreted as ethanol-equivalent TVOC concentrations used for contextual comparison. SFA30 detected HCHO with 1 ppb resolution and ±20 ppb or ±20% accuracy at low concentrations. SCD30 measured CO₂ within 400–10,000 ppm at ±30 ppm + 3% accuracy. ME2-O₂-Φ20 covered 0–25% O₂ volume with ≤15 s response time. Multichannel V2 was qualitative at the sensor level; therefore, ppm values for C₆H₆, C₂H₅OH, and NO₂ were obtained through interpolation equations derived from the manufacturer’s calibration curves. These converted values were interpreted as semi-quantitative sensor outputs to identify possible phase-related changes. As the sensor may respond to multiple gases and VOC mixtures rather than to a single target compound, cross-sensitivity may affect the C₆H₆, C₂H₅OH, and NO₂ channels, particularly in solvent-rich environments. These reported sensor specifications and conversion procedures form the basis of interpreting measurements in this study.

Detection and quantification limits, repeatability, drift, cross-sensitivity, and uncertainty ranges were not independently determined for the assembled INBASE system; therefore, the related measurements were interpreted as indicative environmental monitoring data.

The operational checks and calibration procedures adopted in this study primarily followed manufacturer-recommended protocols for each sensor. These procedures comprised (i) determining background baselines before each session; (ii) performing bench checks: HM3301 zero-check in clean air, SGP30 baseline storage, Multichannel V2 preheating to data-sheet durations, SCD30 ASC with daily outdoor referencing, and ME2-O₂-Φ20 fresh-air reference; and (iii) ensuring procedural controls: ESD-safe operation during calibration and fixed instrument placement on a 1.2 m stand at 1 m from the active printer, consistent across sessions. In addition, the study design included background, pre-printing, printing, post-printing, and post-processing phases, enabling phase-wise validation of expected emission patterns against baselines. However, the INBASE system was not validated against reference-grade analytical instruments. Environmental conditions, including temperature and humidity, which may affect low-cost sensor responses, were not independently corrected in this study. Therefore, the measurements obtained should be interpreted as indicative and comparative environmental monitoring data intended to characterize relative variations, operational trends, and phase-based IAQ patterns within the monitored hybrid workplace environment rather than as regulatory-grade exposure measurements.

AIRPOLL, illustrated in Figure 4, is a simplified real-time voting device developed by the authors for perception feedback, enabling employees to rate perceived IAQ conditions during their routine work. It consisted of five tact switches (Robolink, Ankara, Türkiye), five resistors (Robolink, Ankara, Türkiye) to keep the input signal at logic 0 when the corresponding switch was not pressed, a card shield (SD Card Shield V4) (Seeed Studio, Shenzhen, China), a real-time clock (DS1302) (Analog Devices, Wilmington, MA, USA), a display unit (SSD1306) (Solomon Systech, Hong Kong, China), a breadboard (Robolink, Ankara, Türkiye), and a development board (Arduino Mega 2560 Rev3) (Arduino S.r.l., Monza, Italy).

Figure 4. Indoor Air Quality Polling Instrument (AIRPOLL).

The software of AIRPOLL was developed in Arduino IDE (Version 2.3.2). The configuration code included the connections of switches, resistors, the clock, and the display to the development board and formatting of the vote records for storage. Each button press was recorded on an SD card with a timestamp in a structured format and displayed in OLED screen.

Field experiments in production and office spaces were conducted separately due to spatial and infrastructural conditions of the facility. Therefore, the production and office measurements were not simultaneous, and the results were not used to calculate direct pollutant transfer rates between spaces. Comparisons between production and office spaces were interpreted as phase-based and case-specific IAQ observations under the monitored operational conditions. Data collection followed the facility’s workflow based on received orders, leading to variations in printed components and experiment durations, even for the same AM method. Therefore, the monitored operations should be interpreted as case-based field observations rather than controlled repeated trials. Each monitored case reflected the specific combination of equipment, material, room volume, ventilation condition, filtration system, phase duration, occupancy, and workplace activity present during the measurement period. As the field experiments were conducted under active workplace operating conditions, occupant presence and routine workplace activities were not experimentally controlled. Instead, the monitoring sessions were designed to capture realistic IAQ conditions associated with ongoing AM operations within the hybrid workplace environment. During the experiments, production-related activities, operational phases, and employee presence within the monitored spaces were observed and documented to support contextual interpretation of the measurements. For confidentiality reasons, details of the printed components cannot be disclosed. The timeline of the field experiments is illustrated in Figure 5.

Figure 5. Timeline of the field experiments.

3.2. Field Experiments in Production Spaces

IAQ was monitored in real-time in the production spaces using INBASE during various operational phases of PBF-SLM, PBF-SLS, and VP-SLA, while the employee perceptions of IAQ were captured in real-time using AIRPOLL. Both instruments were positioned on a 1.2-m-tall movable stand, approximately corresponding to the seated breathing-zone height of occupants and 1 m away from the relevant printers, marked by the burgundy circles in Figure 1. These positions were selected to characterize IAQ conditions at fixed workplace locations. However, the measurements represent area monitoring rather than personal breathing-zone sampling. The following sections detail the monitoring sessions for each AM technology.

3.2.1. Powder Bed Fusion—Selective Laser Melting

PBF-SLM operations took place in a 127.7 m³ (34.5 m² × 3.7 m) PBF-SLM production space, highlighted in green in Figure 1. This space accommodates two SLM printers, with different building volumes and brands, and a pre-printing workspace. The monitored SLM printer features a 160 mm × 160 mm × 100 mm build volume, 10–40 μm layer thickness, and 500 W laser power. Titanium alloy (Ti-6Al-4V) and cobalt–chromium alloy powders were used as feedstock materials. The monitoring process covered four phases:

• Background Phase: Baseline IAQ measurements were taken during the period of inactivity;
• Pre-printing Phase: Titanium alloy powder (Ti-6Al-4V) was measured, mixed, and introduced into the printer. The build platform was cleaned to eliminate debris or residue and fixed into the build chamber, and the design file was uploaded to the control software;
• Printing Phase: The SLM printer operated continuously to fabricate the component;
• Post-printing Phase: The printer cover was opened. As the printed component was embedded within the uncured powder, the excess powder was vacuumed, and the build platform was removed. Any remaining powder on the build platform was transferred into a tray, sieved, and replenished for reuse.

Following the post-printing phase, as the printed component with the build platform was sent to another facility for post-processing activities, the post-processing phase was not monitored.

3.2.2. Powder Bed Fusion—Selective Laser Sintering

PBF-SLS operations were conducted in a 49.2 m³ (13.3 m² × 3.7 m) PBF-SLS production space, indicated in purple in Figure 1. This space is equipped with an SLS printer with 165 mm × 165 mm × 300 mm building volume, 110 microns layer thickness, intake and exhaust filters to control air contaminants, an SLS powder management station equipped with a HEPA filter and negative air pressure system for powder recovery, debris filtration, and material storage, and an FDM printer. The monitoring process covered five phases:

• Background Phase: Baseline IAQ measurements were taken during the period of inactivity;
• Pre-printing Phase: The powder cartridge was filled with recovered Nylon 12 powder, loaded into the hopper, and prepared for printing. The cleaned optical cassette and the build chamber were inserted, and the design file was uploaded;
• Printing Phase: The printing phase comprised three subsequent steps occurring inside the SLS printer. These steps were pre-heating (51 min), where the building chamber was heated up to 179 °C; printing (105 min), where the plastic powder was fused layer by layer to print the intended object; and cooling (313 min), where the printed object remained inside the printer to acquire the desired mechanical properties until the bed temperature decreased to 40 °C;
• Post-printing Phase: The build chamber was extracted and transferred to the powder management station;
• Post-processing Phase: The print cake, consisting of printed components and unfused and semi-fused powder, was removed from the build chamber and transferred to the powder management station. The excess powder was scrubbed and brushed off from the components, and the components were collected in a wire basket. The remaining powder on the station was collected into a sifter, filtered, and reclaimed for reuse. Any leftover powder was removed using a vacuum system.

3.2.3. Vat Photopolymerization—Stereolithography

VP-SLA operations took place in a 48.1 m³ (13.0 m² × 3.7 m) VP-SLA production space for pre-printing, printing, and post-printing phases and in a 51.4 m³ (13.9 m² × 3.7 m) post-processing atelier for the post-processing phase, shown in turquoise and pink in Figure 1, respectively.

The VP-SLA production space housed seven SLA printers with varying build volumes, a DLP printer, and an open-shelf system for sealed resin cartridges. The monitored SLA printer has a 145 mm × 145 mm × 193 mm build volume, 25–300 μm layer thickness, and 250 mW laser power. It does not include a filter or other form of emissions capture. The post-processing atelier includes UV cure equipment, large and small wash equipment, two workstations, an air compressor, and post-processing tools. The utilized wash equipment features a wash bucket, wash basket, electric solvent pump, flexible scraper, scraper, flush cutters, rinse bottle, build platform adaptor, and gloves, while the cure system uses 500 W dual heaters to reach 80 °C for curing. The monitoring process covered five phases:

• Background Phase: Baseline IAQ measurements were taken during the period of inactivity;
• Pre-printing Phase: The design file was uploaded to the printer, and the standard grey resin cartridge was shaken and inserted into the printer;
• Printing Phase: The SLA printer continuously operated to fabricate the component;
• Post-printing Phase: The printer cover was opened, and the resin tank and build platform were removed;
• Post-processing Phase: The printed component was transferred to the post-processing atelier, where it was washed with isopropyl alcohol to remove residual resin, dried with an air compressor, cured for 60 min under heat and UV light, detached from the build platform, and had its supports removed. Sanding was not required and was therefore not performed.

3.3. Field Experiments in Office Space

The field experiments in office space, marked in blue in Figure 1, involved real-time IAQ monitoring during various operational phases of PBF-SLM, PBF-SLS, and VP-SLA using INBASE and collection of real-time employee feedback on IAQ perception via AIRPOLL. Both instruments were positioned on an office desk accessible to employees and located within the routinely occupied working area to support representative monitoring of indoor conditions and facilitate real-time AIRPOLL feedback collection during daily workplace activities, highlighted by the burgundy circle in Figure 1. AM operations took place in the respective production spaces using the same equipment and procedures described in production-spaces experiments. For PBF-SLM operations, four operational phases, including background, pre-printing, printing, and post-printing, were monitored. For PBF-SLS and VP-SLA operations, five operational phases, including background, pre-printing, printing, post-printing, and post-processing, were monitored.

3.4. Data Processing and Quality Control

INBASE generated real-time sensor records at a recording interval of 15 s and transmitted the data to the ThingSpeak server. The raw time-series data were exported as CSV files and assigned to operational phases using the documented start and end times of each monitored phase.

Operational phase boundaries were defined according to AM activities recorded during the field experiments. However, other routine workplace activities occurring during the same monitoring periods were not separated from AM-related activities. Therefore, phase averages represent combined IAQ conditions during the monitored phase. Accordingly, the background phase referred to the period before AM-related activity began in the monitored space. The pre-printing phase included preparation activities before printer operation, such as material preparation, machine setup, powder handling, resin preparation, or loading procedures, depending on the AM process. The printing phase began when the printer operation started and ended when the print job was completed. The post-printing phase included activities immediately after print completion, such as printer opening, cooling, part removal, powder removal, or initial handling of the printed component. The post-processing phase referred to additional finishing or cleaning procedures performed after part removal, such as washing, cleaning, or other process-specific treatment steps. Since not all AM operations included all phases, only the phases observed during each monitored operation were reported.

Sensor response time and possible carryover between operational phases were not corrected quantitatively. Therefore, short-term changes recorded around phase transitions were interpreted descriptively as transient observations and were considered together with the documented operational activity at that time.

For each pollutant and phase, the number of valid observations and phase average value were calculated. Missing timestamps caused by communication interruptions were treated as missing data and were not interpolated. Duplicate timestamps and physically implausible values caused by communication or recording errors were removed before analysis. No smoothing was applied to the raw time-series data. Because consecutive sensor readings were temporally autocorrelated, individual real-time observations were not treated as independent replicates. Therefore, no inferential statistical testing, analysis of variance, or confidence interval calculation was conducted. Phase averages were used only as descriptive summaries of the monitored operational phases because independent repeated trials under standardized equipment, material, ventilation, occupancy, and print-job conditions were not available. Reported concentration values represent calculated phase averages derived from INBASE sensor outputs. The displayed decimal places were retained to preserve calculation transparency; however, they should not be interpreted as certified analytical precision. Since measurement uncertainty was not independently quantified through laboratory calibration or co-location with reference-grade instruments, interpretation focused on relative changes, phase-based trends, transient peaks, and contextual reference comparisons. Possible effects of environmental conditions, including temperature and humidity, were considered as part of the overall measurement uncertainty and were not corrected quantitatively.

Sensor warm-up and stabilization procedures followed the manufacturer-based protocols described in Section 3.1. Before each monitoring session, background measurements were collected to document baseline conditions. No additional drift correction was applied because the assembled INBASE system was not co-located with certified reference-grade instruments. Sensor lag at phase transitions was not corrected quantitatively; therefore, short-term peaks occurring around operational transitions, such as printer opening, powder handling, washing, cleaning, or component removal, were interpreted descriptively as transient observations.

AIRPOLL responses were collected voluntarily and anonymously using five response categories: “very good”, “good”, “neutral”, “bad”, and “very bad”. Before the field experiments, employees were informed about the purpose of AIRPOLL and the meaning of the five response categories. The device was used as a simplified practical feedback tool and was not psychometrically validated. These responses were scored from 5 to 1, respectively, for descriptive summary purposes. Because the AIRPOLL scale is ordinal, response distributions, total vote counts, median, and range were used as descriptive indicators. Weighted averages and standard deviations were retained only as supplementary descriptive summaries and were not interpreted as inferential statistics. Because AIRPOLL did not collect personal identifiers, the same employee could potentially vote more than once during a monitoring period. Individual participant identity, participant-level vote frequency, exact task, and exact location at the time of voting were not recorded. A fixed minimum interval between votes was not imposed beyond the device recording procedure.

4. Results and Discussion

Real-time IAQ measurements from INBASE were evaluated for production and office spaces during PBF-SLM, PBF-SLS, and VP-SLA operations. The reference values given in Table A3 were used to contextualize the measured parameters. Where the averaging period of a reference value differed from the monitored phase duration, the comparison was interpreted only as a contextual benchmark. The interpretation was limited to the parameters measured by INBASE and did not represent a comprehensive assessment of all possible AM-related emissions. Additionally, production and office results were compared descriptively because the two spaces were monitored during different time periods.

AIRPOLL responses were scored on a five-point scale ranging from 1 for “very bad” to 5 for “very good” to calculate the weighted averages of IAQ perception values. These responses were used as supportive perception feedback. Because the dataset was limited, anonymous, and ordinal and potentially included repeated votes from the same employees, the observed correspondences between AIRPOLL scores and monitored IAQ parameters were interpreted as exploratory co-occurrences.

4.1. Field Experiments for Powder Bed Fusion—Selective Laser Melting

Indoor air quality pollutant concentrations, operational phase durations, temperature, and relative humidity in the PBF-SLM production space and the office space during PBF-SLM operations are summarized in Table 2.

Table 2. IAQ pollutant concentrations in production and office spaces during Powder Bed Fusion–Selective Laser Melting (PBF-SLM) operations.

Pollutant	Monitored Spaces	Monitoring Phases/Unit	PBF-SLM Operational Phases
			Background	Pre-Printing	Printing	Post-Printing	Post-Processing
Time	Production_Space	min	46	26	66	14	-
	Office_Space	min	30	12	260	14	-
Temp	Production_Space	Avg (°C)	23	24	25	26	-
	Office_Space	Avg (°C)	24	24	23	22	-
RH	Production_Space	Avg (%)	31	32	31	31	-
	Office_Space	Avg (%)	31	32	32	32	-
PM1.0	Production_Space	Avg (μg/m3)	3.00	7.75	3.60	8.15	-
	Office_Space	Avg (μg/m3)	4.13	3.72	3.75	2.07	-
PM2.5	Production_Space	Avg (μg/m3)	4.00	11.03	4.69	11.50	-
	Office_Space	Avg (μg/m3)	4.52	4.86	4.97	3.07	-
PM10	Production_Space	Avg (μg/m3)	4.00	11.25	4.69	11.53	-
	Office_Space	Avg (μg/m3)	5.52	4.86	4.97	3.07	-
TVOC	Production_Space	Avg (mg/m3)	0.05	0.04	0.67	0.82	-
	Office_Space	Avg (mg/m3)	0.13	0.14	0.07	0.19	-
HCHO	Production_Space	Avg (mg/m3)	0.004	0.003	0.003	0.003	-
	Office_Space	Avg (mg/m3)	0.002	0.003	0.003	0.004	-
C6H6	Production_Space	Avg (mg/m3)	0.00	0.00	0.00	0.00	-
	Office_Space	Avg (mg/m3)	0.00	0.00	0.00	0.00	-
C2H5OH	Production_Space	Avg (ppm)	0.00	0.00	0.00	0.00	-
	Office_Space	Avg (ppm)	0.00	0.00	0.00	0.00	-
NO2	Production_Space	Avg (μg/m3)	0.00	0.00	0.00	0.00	-
	Office_Space	Avg (μg/m3)	0.00	0.00	0.00	0.00	-
CO2	Production_Space	Avg (ppm)	439.75	718.97	765.39	804.64	-
	Office_Space	Avg (ppm)	758.48	728.83	563.30	388.63	-
O2	Production_Space	Avg (%)	20.89	21.02	21.10	21.07	-
	Office_Space	Avg (%)	20.98	20.93	20.92	20.93	-

In the PBF-SLM production space, every monitored parameter except for TVOC remained within the reference thresholds across all operational phases. In the office space, all parameters remained within the reference values used for interpretation, indicating that the influence of PBF-SLM operations on the office environment was limited. Although PM_1.0 levels showed variations across operational phases in both production and office spaces, they remained close to background levels. Similarly, PM_2.5 and PM₁₀ concentrations exhibited short-term increases during pre-printing and post-printing in the production space due to powder handling and cleaning operations. The values remained below the selected contextual reference concentrations during the monitored periods.

The principal exception was TVOC, which rose to concentrations of Level 2 based on phase-average concentrations in the production space during the printing and post-printing phases. Real-time measurements showed higher transient peaks at Level 3 during these phases, which are reported separately from the phase-average classification. These increases may reflect thermal degradation and emissions associated with powder fusion and part removal processes. This indicates potential hygienic concerns and a need for increased local ventilation in the PBF-SLM production space. In the office space, TVOC remained within reference ranges. The IAQ pollutant concentrations during PBF-SLM operations are presented in Figure 6 for the production space and in Figure 7 for the office space.

Figure 6. IAQ pollutant concentrations measured in the production space during Powder Bed Fusion–Selective Laser Melting (PBF-SLM) operations (PM: Particulate Matter; TVOC: Total Volatile Organic Compounds; HCHO: Formaldehyde; C₆H₆: Benzene; C₂H₅OH: Ethyl Alcohol; NO₂: Nitrogen Dioxide; CO₂: Carbon Dioxide; O₂: Oxygen).

Figure 7. IAQ pollutant concentrations measured in the office space during Powder Bed Fusion–Selective Laser Melting (PBF-SLM) operations (PM: Particulate Matter; TVOC: Total Volatile Organic Compounds; HCHO: Formaldehyde; C₆H₆: Benzene; C₂H₅OH: Ethyl Alcohol; NO₂: Nitrogen Dioxide; CO₂: Carbon Dioxide; O₂: Oxygen).

Employee ratings from AIRPOLL for the PBF-SLM production and office spaces are summarized in Table 3, including vote counts, response distributions, weighted averages, and standard deviations across operational phases. AIRPOLL responses were interpreted as supportive perception feedback. Although the limited AIRPOLL dataset did not permit robust statistical interpretation of relationships between employee perception and IAQ parameters, the declining trend from good to very bad was observed in IAQ perception in the PBF-SLM production space across operational phases. This pattern co-occurred with increases in TVOC levels in the PBF-SLM production space and may suggest a possible association between TVOC variation and perceived discomfort. In contrast, IAQ perception in the office space was generally positive across operational phases, which can correspond to the environmental conditions observed in the office space. The IAQ perception of employees for the PBF-SLM production space and the office space is presented in Figure 8.

Table 3. IAQ ratings in production and office spaces during Powder Bed Fusion–Selective Laser Melting (PBF-SLM) operations.

Monitored Spaces	Operational Phase	Very Bad (n)	Bad (n)	Neutral (n)	Good (n)	Very Good (n)	Total Votes (n)	Weighted Average	Standard Deviation
Production Space	Background	0	1	1	2	3	7	4.00	1.15
	Pre-printing	0	0	1	0	0	1	3.00	0
	Printing	0	1	4	0	0	5	2.80	0.45
	Post-printing	1	0	0	0	0	1	1.00	0
Office Space	Background	0	0	1	2	0	3	3.67	0.58
	Pre-printing	0	0	1	0	0	1	3.00	0
	Printing	2	0	1	3	3	9	3.56	1.59
	Post-printing	0	1	1	0	1	3	3.33	1.53

Figure 8. Employee perception of IAQ in production and office spaces during Powder Bed Fusion–Selective Laser Melting (PBF-SLM) operations.

4.2. Field Experiments for Powder Bed Fusion—Selective Laser Sintering

IAQ pollutant concentrations from INBASE for the PBF-SLS production space and the office space were presented in Table 4, with corresponding phase durations, temperature, and relative humidity.

Table 4. IAQ pollutant concentrations in production and office spaces during Powder Bed Fusion–Selective Laser Sintering (PBF-SLS) operations.

Pollutant	Monitored Spaces	Monitoring Phases/Unit	PBF-SLS Operational Phases
			Background	Pre-Printing	Printing	Post-Printing	Post-Processing
Time	Production_Space	min	54	33	469	6	154
	Office_Space	min	39	25	321	6	28
Temp	Production_Space	Avg (°C)	26	27	24	25	26
	Office_Space	Avg (°C)	25	26	26	26	26
RH	Production_Space	Avg (%)	39	36	42	45	44
	Office_Space	Avg (%)	50	48	46	45	45
PM1.0	Production_Space	Avg (μg/m3)	5.17	5.82	5.36	7.00	5.19
	Office_Space	Avg (μg/m3)	7.21	7.06	8.74	10.42	10.00
PM2.5	Production_Space	Avg (μg/m3)	7.48	8.03	7.52	10.00	7.27
	Office_Space	Avg (μg/m3)	10.17	10.06	12.43	14.83	14.00
PM10	Production_Space	Avg (μg/m3)	7.48	8.03	7.52	10.00	7.27
	Office_Space	Avg (μg/m3)	10.17	10.06	12.62	16.17	14.09
TVOC	Production_Space	Avg (mg/m3)	0.05	0.13	0.26	0.18	0.22
	Office_Space	Avg (mg/m3)	0.28	0.01	0.33	0.81	0.72
HCHO	Production_Space	Avg (mg/m3)	0.002	0.001	0.002	0.002	0.002
	Office_Space	Avg (mg/m3)	0.002	0.002	0.002	0.002	0.002
C6H6	Production_Space	Avg (mg/m3)	0.00	0.00	0.00	0.00	0.00
	Office_Space	Avg (mg/m3)	0.00	0.00	0.00	0.00	0.00
C2H5OH	Production_Space	Avg (ppm)	0.00	0.00	0.00	0.00	0.00
	Office_Space	Avg (ppm)	0.00	0.00	0.00	0.00	0.00
NO2	Production_Space	Avg (μg/m3)	0.00	0.00	2.91	0.00	0.00
	Office_Space	Avg (μg/m3)	0.00	0.00	0.00	0.00	0.00
CO2	Production_Space	Avg (ppm)	408.12	459.21	405.20	411.73	514.67
	Office_Space	Avg (ppm)	481.28	518.26	585.70	684.33	602.53
O2	Production_Space	Avg (%)	21.08	21.15	21.03	21.05	21.09
	Office_Space	Avg (%)	21.03	21.08	21.08	21.04	21.05

During PBF-SLS operations, most pollutant concentrations in both the production space and the office space remained below the selected contextual reference concentrations during the monitored periods. Particulate matter concentrations remained relatively low in the production space. In the office space, PM_2.5 and PM₁₀ showed short-term increases during post-printing. HCHO, C₆H₆, C₂H₅OH, and O₂ stayed within the reference values used for interpretation throughout; NO₂ showed only a brief rise during printing in the production space; and CO₂ varied with powder-handling activity while remaining within the reference values used for interpretation. Although the room configuration and existing filtration system may have contributed to this pattern, ventilation parameters were not measured quantitatively, and causal attribution is therefore not possible.

TVOC again emerged as the most active parameter. Based on phase-average concentrations, TVOC reached Level 2 during printing in the production space and Level 2 during printing, post-printing, and post-processing in the office space. However, transient real-time peaks reached Level 3 during post-printing and post-processing, revealing short-term increases not reflected by the phase-average values and suggesting a need for enhanced ventilation and source investigation. The IAQ pollutant concentrations during PBF-SLS operations are presented in Figure 9 for the production space and in Figure 10 for the office space.

Figure 9. IAQ pollutant concentrations measured in production space during Powder Bed Fusion–Selective Laser Sintering (PBF-SLS) operations (PM: Particulate Matter; TVOC: Total Volatile Organic Compounds; HCHO: Formaldehyde; C₆H₆: Benzene; C₂H₅OH: Ethyl Alcohol; NO₂: Nitrogen Dioxide; CO₂: Carbon Dioxide; O₂: Oxygen).

Figure 10. IAQ pollutant concentrations measured in the office space during Powder Bed Fusion–Selective Laser Sintering (PBF-SLS) operations (PM: Particulate Matter; TVOC: Total Volatile Organic Compounds; HCHO: Formaldehyde; C₆H₆: Benzene; C₂H₅OH: Ethyl Alcohol; NO₂: Nitrogen Dioxide; CO₂: Carbon Dioxide; O₂: Oxygen).

AIRPOLL ratings for the PBF-SLS production space and the office space are summarized in Table 5, including vote counts, response distributions, weighted averages, and standard deviations across operational phases. The responses were interpreted as supportive perception feedback. The rating indicated that IAQ was perceived as above neutral in the PBF-SLS production space and positive in the office space across all operational phases. Therefore, the feedback may suggest that employees did not perceive significant IAQ deterioration during PBF-SLS operations. The IAQ perception of employees for the PBF-SLS production space and the office space is presented in Figure 11.

Table 5. IAQ ratings in production and office spaces during Powder Bed Fusion–Selective Laser Sintering (PBF-SLS) operations.

Monitored Spaces	Operational Phase	Very Bad (n)	Bad (n)	Neutral (n)	Good (n)	Very Good (n)	Total Votes (n)	Weighted Average	Standard Deviation
Production Space	Background	0	0	2	1	0	3	3.33	0.58
	Pre-printing	0	0	1	0	0	1	3.00	0
	Printing	0	0	1	3	0	4	3.75	0.50
	Post-printing	0	0	0	1	0	1	4.00	0
	Post-processing	0	1	1	2	0	4	3.25	0.96
Office Space	Background	0	0	0	1	0	1	4.00	0
	Pre-printing	0	0	0	2	0	2	4.00	0
	Printing	0	1	0	4	2	7	4.00	1
	Post-printing	0	0	0	1	0	1	4.00	0
	Post-processing	0	0	1	1	0	2	3.50	0.71

Figure 11. Employee perception of IAQ in production and office spaces during Powder Bed Fusion–Selective Laser Sintering (PBF-SLS) operations.

4.3. Field Experiments for Vat Photopolymerization—Stereolithography

IAQ pollutant concentrations gathered from INBASE for the VP-SLA production spaces and the office space, with duration, temperature, and relative humidity, are given in Table 6.

Table 6. IAQ pollutant concentrations in production and office spaces during Vat Photopolymerization–Stereolithography (VP-SLA) operations.

Pollutant	Monitored Spaces	Monitoring Phases/Unit	VP-SLA Operational Phases
			Background	Pre-Printing	Printing	Post-Printing	Post-Processing
Time	Production_Space	min	60	15	573	2	232
	Office_Space	min	42	16	94	2	101
Temp	Production_Space	Avg (°C)	24	24	24	24	24
	Office_Space	Avg (°C)	23	26	23	27	26
RH	Production_Space	Avg (%)	32	30	31	32	40
	Office_Space	Avg (%)	27	34	31	29	35
PM1.0	Production_Space	Avg (μg/m3)	3.87	5.53	5.35	4.50	6.15
	Office_Space	Avg (μg/m3)	4.05	4.63	4.78	9.00	4.70
PM2.5	Production_Space	Avg (μg/m3)	5.03	7.67	7.40	6.00	8.61
	Office_Space	Avg (μg/m3)	5.40	6.50	6.65	13.00	6.62
PM10	Production_Space	Avg (μg/m3)	5.03	7.67	7.40	6.00	8.73
	Office_Space	Avg (μg/m3)	5.40	6.50	6.65	13.00	6.62
TVOC	Production_Space	Avg (mg/m3)	0.09	0.10	0.45	0.44	3.26
	Office_Space	Avg (mg/m3)	0.06	0.08	0.05	0.17	0.12
HCHO	Production_Space	Avg (mg/m3)	0.003	0.003	0.005	0.004	0.004
	Office_Space	Avg (mg/m3)	0.002	0.003	0.003	0.002	0.003
C6H6	Production_Space	Avg (mg/m3)	0.00	0.00	0.00	0.00	2.67
	Office_Space	Avg (mg/m3)	0.00	0.00	0.00	0.00	0.00
C2H5OH	Production_Space	Avg (ppm)	0.00	0.00	0.00	0.00	1.35
	Office_Space	Avg (ppm)	0.00	0.00	0.00	0.00	0.00
NO2	Production_Space	Avg (μg/m3)	0.00	38.89	5.16	0.00	0.49
	Office_Space	Avg (μg/m3)	0.00	0.00	0.00	0.00	0.00
CO2	Production_Space	Avg (ppm)	372.96	688.46	512.86	394.32	630.66
	Office_Space	Avg (ppm)	608.86	652.75	632.56	559.51	689.07
O2	Production_Space	Avg (%)	21.00	21.11	21.01	21.16	20.92
	Office_Space	Avg (%)	20.94	20.99	20.97	20.98	20.22

During VP-SLA operations, pollutant concentrations in the production and office spaces generally remained within the reference concentrations. However, certain indicators approached or were higher than the contextual reference values.

Particulate matter showed slight increases during post-processing, and HCHO, CO₂, and O₂ remained within the reference values used for interpretation. Multichannel V2 sensor responses assigned to the C₆H₆ and C₂H₅OH channels were observed only during VP-SLA post-processing in the production space and were not observed in the office space. Because isopropyl alcohol was used during this phase, the C₂H₅OH channel response was interpreted cautiously as an alcohol-related sensor response. NO₂ showed intermittent short-term sensor responses in the production space during pre-printing, printing and post-processing and was not detected in the office space.

TVOC dominated the VP-SLA emission profile. During post-processing, the phase-average TVOC concentration reached Level 4, while transient peak concentrations exceeded Level 5 in the real-time measurements, underscoring the need for effective emission control and ventilation during this phase. In the office space, phase-average TVOC concentrations remained at Level 1 across all monitored phases, with transient real-time peaks reaching Level 2, indicating no hygienic concern. The IAQ pollutant concentrations during VP-SLA operations are presented in Figure 12 for the production space and in Figure 13 for the office space.

Figure 12. IAQ pollutant concentrations measured in the production spaces during Vat Photopolymerization–Stereolithography (VP-SLA) operations (PM: Particulate Matter; TVOC: Total Volatile Organic Compounds; HCHO: Formaldehyde; C₆H₆: Benzene; C₂H₅OH: Ethyl Alcohol; NO₂: Nitrogen Dioxide; CO₂: Carbon Dioxide; O₂: Oxygen).

Figure 13. IAQ pollutant concentrations measured in the office space during Vat Photopolymerization–Stereolithography (VP-SLA) operations (PM: Particulate Matter; TVOC: Total Volatile Organic Compounds; HCHO: Formaldehyde; C₆H₆: Benzene; C₂H₅OH: Ethyl Alcohol; NO₂: Nitrogen Dioxide; CO₂: Carbon Dioxide; O₂: Oxygen).

Employee ratings gathered from AIRPOLL for the VP-SLA production and office spaces, along with vote counts, response distributions, weighted averages, and standard deviations across operational phases, are given in Table 7. AIRPOLL responses were interpreted as supportive perception feedback. Employee perception of IAQ in the VP-SLA production spaces, given in Figure 14, demonstrated below-neutral ratings across phases except for the background. Due to the short duration of the post-printing, no feedback was collected. Despite the insufficiency of the AIRPOLL dataset for statistical assessment of relationships between perceived IAQ and pollutant concentrations, the below-neutral ratings recorded during operational phases suggest that emissions associated with VP-SLA activities may have led to an uncomfortable environment.

Table 7. IAQ ratings in production and office spaces during Vat Photopolymerization–Stereolithography (VP-SLA) operations.

Monitored Spaces	Operational Phase	Very Bad (n)	Bad (n)	Neutral (n)	Good (n)	Very Good (n)	Total Votes (n)	Weighted Average	Standard Deviation
Production Space	Background	0	0	1	1	0	2	3.50	0.71
	Pre-printing	1	1	1	0	0	3	2.00	1.00
	Printing	1	0	4	0	0	5	2.60	0.89
	Post-printing	0	0	0	0	0	0	-	-
	Post-processing	0	2	3	0	0	5	2.60	0.55
Office Space	Background	0	0	2	5	0	7	3.71	0.49
	Pre-printing	0	0	0	3	0	3	4.00	0
	Printing	0	0	1	2	0	3	3.67	0.58
	Post-printing	0	0	0	0	0	0	-	-
	Post-processing	0	0	0	1	0	1	4.00	0

Figure 14. Employee perception of IAQ in production and office spaces during Vat Photopolymerization–Stereolithography (VP-SLA) operations.

To illustrate, elevated TVOC levels during operational phases, together with sensor responses in the C₆H₆, C₂H₅OH, and NO₂ channels, may have contributed to perceived IAQ discomfort. However, the Multichannel V2 outputs should be interpreted as indicative sensor responses. In contrast, the office space, given in Figure 14, was rated positively, with no feedback collected during post-printing due to its two-minute duration. Although the available observations were insufficient for detailed statistical interpretation of relationships between measured IAQ parameters and employee perception, the generally stable pollutant concentrations within acceptable thresholds may have contributed to the positive IAQ perception.

4.4. Case-Based Comparison of Monitored AM Operations and Discussion

The following discussion distinguishes between measured observations and plausible interpretations. The measured observations refer to the monitored IAQ parameters and AIRPOLL responses recorded during the selected operational phases. Explanations related to emission sources, filtration, ventilation, negative pressure, spatial separation, or pollutant movement are interpreted as plausible mechanisms consistent with the observed patterns. Additionally, because the study was conducted in an active workplace, the observed IAQ patterns may reflect combined effects of AM-related activities, occupancy variation, cleaning, equipment use, door openings, and other routine tasks.

Across the monitored AM operations, TVOC was the parameter that most consistently exceeded the reference levels. It breached its reference thresholds in every production space, and its magnitude differed sharply by technology. Among the monitored cases, the VP-SLA operation showed the highest concentrations. During VP-SLA post-processing, the phase-average concentration reached Level 4, while transient real-time peaks exceeded Level 5. By comparison, PBF-SLM production-space TVOC reached Level 2 based on phase-average values, and PBF-SLS production- and office-space TVOC increases also remained at Level 2 based on phase-average values. Perception ratings declined during the phases in which these elevated concentrations were recorded. Employee ratings declined most, falling below neutral, in the VP-SLA and PBF-SLM production spaces, whereas the PBF-SLS production space, where TVOC peaked at lower levels and the unit incorporated filtration and negative-pressure powder handling, retained above-neutral ratings. On the other hand, the office environment stayed within the reference values used for interpretation and was rated positively during all operations, with only moderate TVOC rises during PBF-SLS and VP-SLA. Because AIRPOLL responses were anonymous, ordinal, and limited in number and because participant identity, repeated voting, task, and exact location were not recorded, the perception data were used only to support interpretation of phase-based IAQ observations.

These findings are well-positioned against earlier research on IAQ in AM environments. In PBF processes, previous studies reported PM and VOC emissions, particularly during powder handling, machine cleaning, and post-processing phases [51,52,54,55]. Our monitoring also showed increases in PM and TVOC during pre- and post-printing activities in SLM and SLS operations, which indicates a similar trend. Other studies generally reported concentrations below occupational exposure limits [66,69,71]; however, concerns related to cobalt exposure, nanoparticle peaks, and recycled powder behavior were also highlighted [60,61,70]. These exposure types were outside the scope of this study due to differences in monitored parameters.

For polymer-based SLS, Damanhuri et al. [67,68] and Zisook et al. [58] reported increases in PM_2.5, CO₂, and TVOC. The concentration levels were higher with recycled powders than virgin powders. In this study, PBF-SLS operations resulted in TVOC increases occasionally, which is consistent with their observation that post-printing and powder recovery stages contribute to emissions. CO₂ and PM_2.5 levels occasionally exceeded guideline values in their study, whereas relatively lower levels of PM_2.5 and CO₂ were observed in this study. Moreover, Väisänen et al. [62] reported higher PM and HCHO concentrations with reinforced nylon powders. By comparison, the results of this research showed moderate increases in PM. These variances can be attributed to the differences in feedstock material, powder condition, printer configuration, room volume, ventilation context, filtration, differences in monitoring design, and workplace conditions.

In the case of vat photopolymerization, prior studies highlighted TVOC increases during cleaning and curing steps [73,75,76]. This study also recorded significant TVOC rises during SLA post-processing, along with sensor responses in the C₆H₆ and C₂H₅OH channels during the same phase, which is consistent with their findings on solvent- and resin-related emissions. However, they should not be interpreted as confirmed compound-specific concentrations because TVOC composition was not analytically identified, and Multichannel V2 cross-sensitivity could not be excluded. By contrast, Zisook et al. [58] reported relatively stable conditions in a well-ventilated laboratory, whereas our results in a production atelier indicated higher levels, suggesting that ventilation differences may account for this divergence. They may also reflect occupancy, time of day, ventilation state, door conditions, cleaning, or other routine workplace activities.

Overall, the monitored cases were consistent with earlier research that emission activity may increase during material handling, cleaning, and post-processing tasks. Where differences occur, they appear to reflect variations in feedstock material, process configuration, and ventilation practices across facilities. However, because the present study did not include controlled repetitions, the results should be interpreted as case-based observations.

4.5. Implications for Hybrid Workplace Design and Monitoring

These findings carry practical implications for the design and operation of hybrid AM workplaces, but the recommendations should be interpreted as precautionary design and operational strategies. These implications are drawn from the monitored cases and the existing body of literature and should be adapted to the specific equipment, materials, ventilation systems, and spatial layouts of other AM workplaces. As higher pollutant concentrations were mainly observed in production spaces, while office-space increases were more limited, enclosing fabrication activities in dedicated rooms with buffer zones and controlled circulation paths can be an effective first line of defense. Ventilation should be zoned to the source, prioritizing the production and post-processing areas where TVOC peaks rather than relying on general office ventilation. Phase-specific peaks, particularly during VP-SLA post-processing and powder-handling stages, warrant localized exhaust and source capture, which can substantially reduce TVOC at the breathing zone [10]. Finally, the phase- and process-dependent variation in IAQ conditions demonstrates the need for continuous IAQ monitoring rather than spot checks, ideally paired with occupant feedback to capture VOC-driven discomfort that instruments alone may underprioritize.

5. Conclusions

This study investigated IAQ conditions and employee perception during selected PBF-SLM, PBF-SLS, and VP-SLA operations, within an AM-integrated hybrid workplace. Field experiments captured real-time pollutant data on PM_1.0, PM_2.5, PM₁₀, TVOC, HCHO, C₆H₆, C₂H₅OH, NO₂, CO₂, and O₂, together with employee feedback across various operational phases using custom-designed instruments, INBASE and AIRPOLL. The results are summarized in Table 8. However, these results refer only to the monitored parameters and do not establish comprehensive safety conditions for all AM-related emissions. The labels summarize measurements according to the reference values used for contextual interpretation in this study. As there is no reference value for PM_1.0, it is excluded from the table.

Table 8. Summary of the Results of the Research.

	PBF-SLM		PBF-SLS		VP-SLA
	Production Space	Office Space	Production Space	Office Space	Production Space	Office Space
PM2.5	within reference range	within reference range	within reference range	within reference range	within reference range	within reference range
PM10	within reference range	within reference range	within reference range	within reference range	within reference range	within reference range
TVOC	above reference value	within reference range	above reference value	above reference value	substantial concern	above reference value
HCHO	within reference range	within reference range	within reference range	within reference range	within reference range	within reference range
C6H6	no sensor response	no sensor response	no sensor response	no sensor response	channel response—above reference value	no sensor response
C2H5OH	no sensor response	no sensor response	no sensor response	no sensor response	within reference range	no sensor response
NO2	no sensor response	no sensor response	channel response—within reference range	no sensor response	channel response—within reference range	no sensor response
CO2	within reference range	within reference range	within reference range	within reference range	within reference range	within reference range
O2	within reference range	within reference range	within reference range	within reference range	within reference range	within reference range
IAQ Perception	below neutral	above neutral	above neutral	above neutral	below neutral	above neutral

Across all technologies and spaces, PM_2.5, PM₁₀, HCHO, CO₂, and O₂ levels remained within reference values. C₆H₆, C₂H₅OH, and NO₂ were generally not observed except for the VP-SLA production spaces. TVOC showed the clearest process- and phase-dependent variation. In production spaces, phase-average TVOC concentrations reached Level 2 during PBF-SLM and PBF-SLS operations and Level 4 during VP-SLA post-processing, with transient real-time peaks exceeding Level 5. In office spaces, phase-average TVOC concentrations remained at Level 1 during PBF-SLM and VP-SLA monitoring and reached Level 2 during PBF-SLS monitoring. Employee perception of IAQ was positive in the office space during all operations. While the employee feedback was positive in PBF-SLS production space, it declined below neutral in the production spaces of PBF-SLM and VP-SLA, suggesting a perceived impact of emissions in those environments.

To summarize, IAQ conditions varied across the monitored cases, operational phases, and workplace spaces, with higher pollutant concentrations observed in production spaces. Office spaces maintained a more stable IAQ and received positive perceptions, while production spaces raised concerns during specific phases. Although the employee feedback dataset was too limited to conduct statistical analysis to interpret relationships between IAQ parameters and IAQ perception, the observations were used to acquire insights to address IAQ challenges in hybrid workplaces. The impact of many IAQ pollutants was limited in the office space, while certain exceptions were captured with TVOC concentrations. These observations indicate that IAQ conditions in office spaces should also be monitored in hybrid AM workplaces; however, direct pollutant transfer could not be confirmed because production and office spaces were not monitored simultaneously. The observations support the consideration of facility adjustments; spatial organization, including enclosed fabrication rooms, buffer zones, and controlled circulation paths; improvements for ventilation conditions, such as localized exhaust ventilation and negative pressure containment; maintenance of the fabrication and filtration equipment; and encouraging the use of PPE during AM operations as precautionary strategies for reducing potential pollutant movement and occupant exposure.

This study has several limitations. The field experiments were conducted in a single hybrid workplace specializing in AM and covered selected PBF-SLM, PBF-SLS, and VP-SLA operations. Independent repetitions under standardized equipment, material, ventilation, occupancy, and print-job conditions were not conducted. Therefore, analysis of variance, confidence intervals, and statistical generalization across AM processes were not performed. The reported phase averages were interpreted as descriptive summaries of monitored case-specific IAQ conditions. The monitoring process was conducted following the facility’s production schedule; therefore, monitoring durations, printed components, material conditions, and occupant activities could not be fully standardized, limiting the temporal and spatial representativeness of the results. Although AM-related activities defining each operational phase were documented, a detailed time–activity log covering all occupant movements, door openings, cleaning activities, non-AM equipment use, and unrelated workplace tasks was not recorded. Therefore, these variables were treated as contextual workplace conditions that may have contributed to the observed IAQ variation. Accordingly, the findings should be interpreted as a case-based assessment of IAQ behavior in the monitored hybrid workplace under operational conditions.

Several reference values used for contextual interpretation have annual, 24 h, or 30 min averaging periods that did not always correspond to the monitored phase durations. Therefore, these comparisons were not used to determine compliance or safety. Due to the availability of only one air quality monitor, production and office spaces were not monitored simultaneously. Therefore, production–office comparisons should be interpreted as case-specific observations.

The sensors utilized for INBASE were commercially available low-cost sensors, which are calibrated according to manufacturer protocols and may introduce accuracy and precision limitations since the system was not validated against certified reference-grade analytical instruments. In addition, detection and quantification limits, repeatability, drift, cross-sensitivity, temperature- and humidity-related response effects, and uncertainty ranges were not independently determined for the assembled INBASE system. This limitation is particularly relevant for the Multichannel V2 sensor. Therefore, the results should be considered indicative and comparative environmental measurements. In addition, TVOC values were converted from SGP30 ppb output to ethanol-equivalent mg/m³ values using a fixed conversion basis. The TVOC classification used in this study was originally intended as a hygiene-evaluation guide and was not treated as a health-based limit. Because TVOC composition was not analytically identified, the TVOC results should be interpreted as relative VOC-related variation. Drift correction was not applied beyond the manufacturer-based calibration and stabilization procedures. Overall, the real-time sensor data were used as descriptive time-series records to identify relative changes, phase-based trends, and transient peaks.

INBASE monitored selected IAQ parameters but did not provide a complete characterization of all AM-related emissions. The measurements were based on fixed-location area monitoring. Individual time–activity patterns and the duration employees spent in different spaces were not recorded. Therefore, the results should not be interpreted as personal exposure measurements, exposure-dose estimates, or occupational health-risk assessments. Instead, they represent monitored indoor concentrations for the monitored parameters only at selected workplace locations.

AIRPOLL employee feedback was limited by the small number of participants and operational constraints. The device was developed as a simplified practical feedback tool and was not psychometrically validated. Responses were anonymous and ordinal, and the same employee could potentially vote more than once. Participant identity, participant-level vote frequency, exact task, and exact location at the time of voting were not recorded. Therefore, the independence of individual votes could not be verified. The data were also subjective and may have been affected by individual sensitivity, task type, and location within the workplace. Therefore, inferential statistical analysis was not conducted, and AIRPOLL results were interpreted as supportive, phase-based perception observations.

Ventilation and spatial performance were not quantitatively evaluated in this study. Air exchange rates, supply and exhaust airflow, local exhaust capture velocity, pressure differences between rooms, airflow direction, door and window conditions, return-air configuration, outdoor pollutant concentrations, and detailed occupancy/activity levels were not measured. Therefore, differences between production and office spaces cannot be directly attributed to specific ventilation, filtration, negative-pressure, or spatial-separation strategies. The design recommendations presented in this study should be interpreted as precautionary strategies informed by the monitored case and previous literature.

Future research should address the methodological limitations of this case-based study through longer monitoring periods, repeated measurements, simultaneous multi-location monitoring, and validation against reference-grade instruments. Such studies should document occupancy, time–activity patterns, door conditions, ventilation state, airflow direction, pressure differences, and other relevant spatial and operational variables to better evaluate temporal relationships and possible pollutant movement between spaces. Future work should also include additional AM technologies, materials, and workplace configurations to improve the generalizability of the findings.

Further validation of INBASE is also needed. Laboratory calibration, uncertainty analysis, and cross-sensitivity testing would be particularly important for improving the interpretation of Multichannel V2 gas sensor outputs. Indicative sensor responses should be verified using reference-grade analytical methods such as GC-MS, calibrated electrochemical comparison, or laboratory co-location. Future studies should include personal breathing-zone sampling, ultrafine particle number concentration, particle-size distribution below 1 μm, particle composition analysis, metal and polymeric particle characterization, and reference-grade analytical VOC measurements. In addition, the effectiveness of mitigation strategies such as enclosure, buffer zones, localized exhaust ventilation, filtration, and controlled circulation using airflow measurements, pressure-difference monitoring, capture-efficiency assessment, simultaneous multi-location IAQ monitoring, and personal exposure measurements should be evaluated. Larger AIRPOLL datasets would also support stronger analysis of relationships between IAQ parameters and perceived air quality.