Characterization, Distribution, and Risk Assessment of Polycyclic Aromatic Hydrocarbons (PAHs) in the Workplaces of an Electric Arc Furnace (EAF) Steelmaking Factory

This study measured workplace polycyclic aromatic hydrocarbons (PAHs) in an electric arc furnace (EAF) factory during regular and maintenance periods and estimated workers’ lung cancer risk from 40 years of exposure using Monte Carlo Simulation. Workers were grouped into three similar exposure groups (SEGs) based on their tasks in melting, ladling and casting areas of the EAF factory. Results showed that the PAH levels (0.0127–0.0310 µ g BaP eq m –3 ) during maintenance period were two to four orders higher than atmospheric PAH concentrations in some industrial sites of Taiwan. PAH levels rose to 0.0533–0.155 µ g BaP eq m –3 during regular work as hotter furnaces released more PAHs. Compared to maintenance period, particle PAHs increased more in melting area, gas PAHs increased more in ladling and casting areas, indicating that melting emissions were mainly particles, but gas PAHs traveled farther and raised gas levels in adjacent areas. The Monte Carlo simulation estimated the 95th percentile of PAH risk for three SEGs (1.26 × 10 –5 –8.00 × 10 –6 ). The health risk assessment showed that PAH exposure put one to three workers per thousand in each area at risk of lung cancer, above the acceptable limit of one per million. The data implied that every five micrograms increase in BaP eq concentrations added ten per million cancer risk with 95% certainty. The study recommends personal inhalation protection for workers and air pollution control devices as long-term solutions.


INTRODUCTION
Electric arc furnace (EAF) steelmaking recycles metal scraps to produce metal products and contributes up to 30% of the total global annual crude steel production.Aside from lower installation costs and shorter construction period compared to traditional blast furnace-(basic oxygen furnace (BOF)), EAFs were also more energy efficient than the BOFs.EAF uses 9.1-12.5GJ of energy per ton steel while the traditional blast furnace-BOF steelmaking method requires up to 31.2 GJ of energy per ton of crude steel, thus making EAFs more energy efficient than BOFs (Yang et al., 2014).Moreover, the use of EAFs results in lower PAH emissions than BOFs since no iron production system is involved in steelmaking such as coking, sintering and blast furnace (Chen et al., 2021;Lee et al., 2022;Li et al., 2018).Even so, metal scraps contain organic residues such as rubber, cable insulators and other plastic compounds (Wang et al., 2021) which when combusted during steel production favors PAH formation in EAF factory.
PAH emissions from stacks of various manufacturing processes in the iron and steel industry have been well investigated.It should be noted that these manufacturing processes that generated PAHs might be mainly exhausted through stacks, however, part of them might also be released to the workplace atmospheres and result in the workers' exposure.Epidemiological studies have shown that long-term exposures to PAHs in workplace atmospheres might result in workers with excessive lung cancer rates and cause respiratory, skin, bladder and many other diseases (Boffetta et al., 1997;Cappelletti et al., 2016;Hoshuyama et al., 2006;Room et al., 2022).Since the toxicity of PAHs varies directly with its number of benzenoid rings (the larger the number of benzenoid rings, the more toxic it is relative to benzo[a]pyrene (BaP) (Petit et al., 2019;Yadav et al., 2020), it is important to know how much PAHs, particularly the HMW-PAHs, workers were exposed to in the EAF factory.
As of now, the monitoring of PAH concentrations in the workplace environment has been reported in sinter factories (Lin et al., 2008;Zhang et al., 2022) and EAF factories (Jackson et al., 2012;Chen et al., 2021).In sinter factory workplace PAHs monitoring, samples were taken from raw materials inlet, sintering grate, rough roll shredder, control room and outdoor environment.It was found out that the air conditioning system promoted the "filter effect" which influenced lower PAHs in the control room, hence, suggesting that PAHs were from emission fugitives in the sinter factory (Lin et al., 2008).In another study about sinter factory workplace emission, it was determined that the coking process emitted highest PAH concentration in the workplace (Zhang et al., 2022).
For EAF factory, PAH workplace monitoring was conducted in the melting shop, casting department, and furnace control cabin.The results suggested that the melting shop and casting department workplaces contained the highest number of PAHs and the data gathered estimated three in one million individuals were at risk at developing lung cancer considering of 40-year occupational exposure period which was within the acceptable range of excess cancer risk by U.S. Environmental Protection Agency (U.S. EPA) (Jackson et al., 2012).In our previous study, personal inhalable PAH exposure of workers in the EAF factory was investigated and the dermal and lung cancer risk were estimated from all the similar exposure groups namely melting, ladling and casting workers.The results showed that workers were exposed to 2.17 × 10 4 -1.52 × 10 5 ng m -3 and 2.73-6.76× 10 3 ng m -3 for gas-and particle-phase PAH concentrations, respectively.The melting and casting workers were exposed to PAHs that were higher than the time-weighted average permissible exposure level (PEL-TWA).Furthermore, results indicated that one in ten thousand individuals were at risk of developing skin cancer over a lifetime exposure (Wang et al., 2021).
To characterize the source apportionment, fluctuations and distribution of PAH concentrations in the workplaces of EAF steelmaking factory, the study was conducted during maintenance and regular work periods.To the best of our knowledge, there have been no reports describing the difference of PAH emissions during maintenance work period (furnaces were shutdown) and regular work period (furnaces were in operation).After characterizing and analyzing the distribution of PAHs during regular and maintenance work period, the lung cancer risk was evaluated for workers using a Monte Carlo Simulation over a 40-year occupational exposure in the EAF factory.

Sampling Methods and Sampling Strategy
Regular work period is described as the involvement of all the four manufacturing processes, including melting, ladling, and casting involved in the production of steel and iron, while maintenance period was when operations are ongoing without turning the furnace on.Through our field observation, workers in the melting area conducted very similar tasks and hence can be considered as a similar exposure group (SEG).Areas of the ladling and casting areas were located independently from the furnace area and workers in each area can also be regarded as one SEG.Therefore, three SEGs namely: melting, ladling and casting areas were selected to characterize PAH concentrations in the EAF factory.
For each selected SEG, static samples were collected from seven sampling sites uniformly distributed in the workplace.At each sampling site, a sampling train was placed at the breathing zone of a mannequin (~1.5 m above the ground level).The sampling train consisted of a filter cassette (IOM personal inhalable aerosol sampler, SKC Inc., Catalog No. 225-70) for collecting particle phase PAHs and followed by a sorbent tube (washed XAD-2, 3.5 g/0.5 g) for collecting gas phase PAHs.The sampling flow rate was specified at 2 L min -1 for the IOM personal inhalable sampler, and subsequently adjusted its flow to 0.2 L min -1 via a needle valve for the sorbent tube.The sampling time was approximately 6 to 8 hours per sample.Time/activity patterns for workers in each of the three selected SEG were recorded according to our field observation.

Sample Analysis
The PAHs contained in each filter and sorbent tube sample were Soxhlet extracted with the solvent (n-hexane and dichloromethane, V:V = 1:1) for 24 h, after which the extracts were concentrated, a series of cleanup and re-concentrated to exactly 0.5 mL or 1.0 mL.Gas Chromatography (GC) and Mass Spectrometry (MS) were used for PAH analyses.The GC instrument (Hewlett-Packard 5890A) was equipped with a capillary column (Hewlett-Packard Ultra 2-50 m × 0.32 mm × 0.17 µm) and an HP 7673A automatic sampler.The MS instrument (Hewlett-Packard 5972) was operated under the selected ion monitoring (SIM) mode using the molecular ions determined from the scan mode for pure PAH standards.The details of the GC-MS, analytical procedures, quality assurance, and quality control are contained in the supplementary materials and our previous study (Chang et al., 2014;Chen et al., 2017;Wang et al., 2009Wang et al., , 2007)).
The concentrations of 22 PAHs species were determined, and they were further sorted into three different molecular weight ranges: low molecular weight (LMW) PAHs (containing two-and three-ringed PAHs), middle molecular weight (MMW) PAHs (four-ringed PAHs), and high molecular weight (HMW) PAHs (five-, six-and seven-ringed PAHs).The carcinogenic potency of each collected sample was also determined by BaP equivalent concentration (BaPeq) which is the product of its toxic equivalent factor (TEF) and its concentration.Detailed information of each individual PAHs and its TEF was described in the supplementary materials.

Probabilistic Exposure and Health-risk Assessment
The 22 PAHs species sampled in this study were further evaluated to assess its health-related risks to EAF workers between regular and maintenance work period.Since BaP is the most carcinogenic PAH species, the toxicity of each sampled PAHs was determined based on BaP equivalent concentration (BaPeq).To calculate the BaPeq, a toxic equivalent factor (TEF, see Table S1) as a multiplier to the concentration of PAHs species was used which was based on the Taiwan Environmental Protection Authority (EPA) advisory note for PAHs classification.In the U.S. EPA probabilistic model approach, estimating cancer risks from PAHs compounds was estimated by calculating the chronic daily intake (CDI, mg kg -1 day -1 ) through inhalation pathway using Eq. ( 1): where BaPeq is the total PAHs concentration in terms of BaP equivalence, IR is the inhalation rate of adults, EF is the exposure frequency of EAF workers (8-12 h day -1 for 6 days week -1 duty duration), ED is the exposure duration throughout EAF factory tenure (25-40 years), BW is the body weight of workers (65.6 ± 13.0 kg) and LE is the average life expectancy (70 years or 25,550 days) (See Table S2).After which, lung cancer risk (CR) is then calculated by multiplying the CDI with the carcinogenic slope factor of PAHs which is 3.14 ± 1.80 mg kg -1 day -1 (Qishlaqi and Beiramali, 2019).
In this study, Monte Carlo Simulation (Gupta and Gupta, 2023) was applied to forecast lung cancer risks caused by PAHs by selecting a random value of each variable parameter in Eq. ( 1) according to their distribution function.Each simulation was performed at 10,000 iterations.If the values for cancer risk estimates that correspond to the 95th percentile of the risk distribution exceed the safe levels of 10 -6 , they are deemed unacceptable (Qishlaqi and Beiramali, 2019).

Concentrations of PAH Homologues during Maintenance Period
The PAH concentrations in the melting, ladling, and casting areas during the maintenance period are listed in Table 1.The maintenance period lacked thermal combustion suggesting that the PAH concentrations found represent background levels for the EAF plant.In this study, the ladling and casting areas had comparable concentrations (0.211 µg m -3 and 0.194 µg m -3 for mass concentrations; 0.0310 µg m -3 and 0.0249 µg m -3 for BaPeq concentrations, respectively), which were slightly higher than the melting area (0.147 µg m -3 ; 0.0127 µg m -3 ).The PAH concentrations during the maintenance period were like the total PAH concentrations in the ambient air found in a multi-industrial sampling site in central Taiwan (mass = 1.65 ± 1.24 µg m -3 ; BaPeq = 0.0585 ± 0.00265 µg m -3 ) (Fang et al., 2004b) and near petrochemical industrial complex in Kaohsiung, Taiwan (mass = 0.158 ± 0.0677 µg m -3 , BaPeq = 0.00492 ± 0.00422 µg m -3 ) (Fang et al., 2004a).Like this study, the LMW and MMW PAHs were the most abundant in gas phase and HMW PAHs were most abundant in particle phase (Lai et al., 2017).Table 2 also shows that in the ladling and casting areas, their gas and particle phase PAH concentrations (0.0920-0.102 µg m -3 in gas vs. 0.0982-0.113µg m -3 in particle) were similar.However, in the melting area, gas phase PAH concentration (0.111 µg m -3 ) was higher than the particle phase (0.0363 µg m -3 ).In terms of total BaPeq concentration, particle phase PAHs in the ladling and casting areas (0.0183-0.0249 µg m -3 ) were three to four times higher than gas phase PAHs (0.00612-0.00663µg m -3 ).However, the casting area had similar BaPeq concentrations for gas and particle phase PAHs.These suggest that the degree of agitation of PAH-bearing particles during maintenance can vary depending on the location.
BaPeq results indicated that PAH concentrations during regular work were below the Occupational Safety and Health Administration (OSHA)-Permissible Emission Levels (PEL) (200 µg m -3 ).Although the melting area was expected to produce the highest emissions, the ventilation system, which created an upward thermal draft force, significantly reduced the emission values.Similar observations were found in our prior investigation into worker PAH exposure using personal sampling procedures.It was noted that workers stationed in the ladling area were exposed to higher levels of PAHs (mass = 158 µg m -3 ; BaPeq = 0.415 µg m -3 ) in comparison to those in the melting and casting areas (mass = 28.4-33.2µg m -3 ; BaPeq = 0.0975-0.161µg m -3 ) (Wang et al., 2021).
Contrary to the maintenance period, in the regular period, the mass concentrations of gas phase PAHs in the three areas were two times higher than the particle phase PAHs.For BaPeq, the gas and particle phase PAHs of ladling and casting areas were similar (ladling: 0.0700 µg m -3 and 0.0848 µg m -3 ; casting: 0.0267 µg m -3 and 0.0266 µg m -3 , for gas and particle phase PAH respectively), however, in the melting area, BaPeq in the gas phase (0.0640 µg m -3 ) remained twice as high as in the particle phase (0.0398 µg m -3 ).
Table 3 lists the ratios of PAH homologue concentrations of regular to maintenance work periods in the melting, ladling, and casting areas, showing that there were significant differences in the PAH concentrations between the two work periods.During the regular work period, PAH mass concentrations were three to six times higher, while BaPeq concentrations were two to eight times higher than those during maintenance period in all areas.This increase can be attributed to the furnaces operating at elevated temperatures, which trigger more fugitive PAHs into the work environments.Therefore, it is essential to calculate the resulting worker exposure risks to safeguard their well-being.
Upon comparing the different SEGs, it was found that in the melting area, the increased levels (8.3 times higher) on particle phase PAH mass concentration were higher than that of gas phase PAH (5.5 times higher), while for the increased levels of BaPeq, they were comparable (8.4 times higher and 7.7 times higher for gas and particle, respectively).However, in the ladling and casting areas, the increase levels in gas phase (mass concentration = 6.9 and 4.0 times higher; BaPeq concentration = 11.4 and 4.0 times higher) were much higher than those of the particle phase (mass concentration = 3.3 and 1.9 times higher; BaPeq concentration = 3.4 and 1.5 times higher).
The overview of the gas-particle partitioning of PAHs in the melting, ladling, and casting areas of the EAF factory was presented in Fig. 1 and it was shown that the mass distribution was dominated by gas-phase PAHs, while the BaPeq distribution was dominated by particle-phase PAHs during both work periods.This suggests that fugitive emissions from the melting area primarily exist as particulate matter, however, the gas phase PAHs were found to be more easily transported than the particle phase, leading to a significant increase in gas-phase PAHs in the adjacent areas.In general, both gas and particle phase PAH pose health risks that depend on exposure duration, concentrations, and individual susceptibility.Gas phase can penetrate the lungs more easily which could result in respiratory irritation, lung inflammation and cancer.On the other hand, particlephase PAHs are more likely to cause respiratory problems due to their ability to penetrate deep into the lungs and cause inflammation, resulting in respiratory symptoms such as coughing, wheezing, and shortness of breath (Låg et al., 2020).
Fig. 3 presents the toxicity distribution of the 22 PAH species in all SEGs.LMW and MMW PAHs in gas phase show no significant difference in BaPeq contribution percentage while HMW PAHs exhibit notable variations.BaP and Dba dominate both gas and particle phases.Compared to the maintenance period, the percentages of most LMW PAHs in the gas phase in the melting area remain the same except for NaP (increased by 2%), while HMW PAHs such as CyP, BaP, Dba, and Cor increased by 1-3% during the regular period.Particle phase PAHs increased by 3-7% for most congeners, except for some congeners which were decreased such as NaP (15%), IdP and Dba (7%), Phe (5%), and AcP, Ace, and Flu (4%).Similarly, in the ladling area, HMW PAHs in both gas and particle phases increased by 3-6% and 3-4%, respectively, causing LMW and MMW PAHs to decrease by 1-5% and 1-2%, respectively.The most notable differences were NaP and BaA (5%) decreased in the gas phase and BaA (3%) increased in the particle phase.In the casting area, gas phase PAHs showed no significant changes, while HMW PAHs in the particle phase increased, especially IdP (11%).Other HMW PAHs such as DbP, Cor, and BgP decreased by 2-4%.
A study examined how PAHs form in an EAF factory and found that the ring growth mechanism was most likely to occur due to addition of C3 or C4 radicals to aromatic rings (Yang et al., 2022).These radicals come from hydrocarbon decomposition or hydrogen abstraction from PAHs.The mechanism occurs at high temperatures and pressures, like in an electric arc furnace, and produces PAHs of various shapes and sizes.Hydrocarbons in steel scrap or natural gas can crack and form C3 or C4 radicals in the furnace.Oxygen can also react with PAHs and create radical sites that combine with other PAHs or hydrocarbons to form larger PAHs (Reizer et al., 2022;Yang et al., 2022).This study confirmed the ring growth mechanism in the EAF factory by observing an increase in MMW and HMW PAHs and a decrease in LMW PAHs during the regular period compared to maintenance period when the EAF was on shutdown.

Probabilistic Exposure and Health Risk Assessment
The results of the Monte Carlo simulation estimates were based on BaPeq obtained during regular periods and are shown in Table 4.The study found that the average PAH risk estimates were similar for all three SEGs, with values of 1.26 × 10 -5 , 1.87 × 10 -5 , and 8.00 × 10 -6 for melting, ladling, and casting areas, respectively.However, the 95 th percentile showed that ten to thirty  per million individuals were at risk of developing lung cancer in each of the SEGs due to continuous PAH exposure, which exceeded the acceptable limit for lung cancer risk (1 × 10 −6 , one case of cancer per million people).Furthermore, the data suggested that for a minimum increment of five micrograms in BaPeq concentrations, an additional ten per million individuals with 95% certainty is at risk of developing cancer.In addition, the higher concentration of gas-phase PAHs is an indicator of increased exposure to PAHs compared to particle-bound PAHs, which can increase the risk of adverse health effects.Thus, it is highly recommended that workers in the EAF factory use personal inhalation protective equipment, and the factory should have air pollution control devices as a long-term solution to the workplace's PAHs air pollution.

CONCLUSIONS
This study examined PAHs in gas and particle phases in three areas and two work periods of an EAF factory in Taiwan.The results showed that PAH levels were higher in regular work (0.585-1.04 µg m -3 in mass and 0.0533-0.155µg m -3 in BaPeq) than in maintenance work (0.147-0.211 µg m -3 in mass and 0.0127-0.0310µg m -3 in BaPeq).The levels of PAHs in particles and gas varied depending on the process and the area.During the regular period, particle PAHs rose more in the melting area, while gas PAHs increased more in the ladling and casting areas.This suggests that melting produced more particle emissions, but gas emissions could spread further and affect nearby areas.The ladling area had the highest PAH exposure in both phases.The gas-particle partitioning of PAHs varied by work period and area, with more gas phase PAHs in regular work (61-69% of mass and 28-50% of BaPeq) than in maintenance work (43-76% of mass and 20-50% of BaPeq).NaP was the most abundant PAH in both phases (40-50% of mass), while BaP and Dba were the most toxic PAHs in both phases (28-50% of BaPeq).The increase of HMW PAHs in regular work (3-7% in mass and 2-8% in BaPeq) suggested ring growth as the main PAH formation pathway in the EAF factory.The probabilistic risk assessment showed that ten to thirty per million workers had lung cancer risk from PAH exposure.The findings provide useful information for health risk assessment and emission control of PAHs in EAF factories.They also improve the understanding of PAH sources and behaviors in EAF processes.Moreover, measures such as personal protection and air pollution control are recommended to reduce PAH exposure and health risks in the workplace.

Fig. 2 .
Fig. 2. PAHs congener profiles mass distribution percent in the melting, ladling, and casting areas during maintenance and regular work period.

Fig. 3 .
Fig. 3. BaPeq congener profiles distribution percent pin the melting, ladling and casting areas during maintenance and regular work period.

Table 1 .
Concentra�ons of PAH homologues during maintenance period.

Table 2 .
Concentrations of PAH homologues during regular period.

Table 3 .
PAH homologue ratios of regular to maintenance work periods in melting, ladling and casting areas.

Table 4 .
Probabilistic estimation of lung cancer risk per million individuals using Monte Carlo Simulation during regular work period (iterations = 10,000).