Introduction

In modern occupational health and medicine, the issue of the impact of nanoparticles (NPs, < 100 nm) on workers in metallurgical and metalworking industries is becoming particularly pressing. Technological processes characterized by high temperatures and significant energy consumption (smelting, welding, cutting, and electrical discharge machining) are accompanied by the incidental generation of large volumes of nanoparticles. According to the results of air quality monitoring, concentrations of ultradisperse aerosols in industrial settings can reach peak values of 7.8 million particles per cm3 [1–11]. Experimental and epidemiological data indicate that inhalation exposure to fine particulate matter triggers the development of systemic respiratory, cardiovascular, immune, and neurological disorders [1, 12–19].

Despite the proven risks specific to nanoparticles, current regulations are largely based on mass-based indicators of total aerosol, which leads to an underestimation of the actual threat posed by the nanofraction. This regulatory and methodological gap represents a key barrier to ensuring safety of workers at metallurgical plants, where complex mixtures of metal and metal oxide NPs are regularly generated [2, 6, 20, 21]. Therefore, the development of a reliable methodology for the identification and quantification of nanoparticles in the workplace air is a priority. Scanning electron microscopy (SEM), combined with energy-dispersive X-ray spectroscopy (EDX), enables the analysis of individual nanoscale aerosol particles, including the determination of the multi-element composition of samples [22–26]. The development of a comprehensive approach to identifying the nanofraction of aerosols based on these methods is crucial for risk assessment and the implementation of effective preventive measures in industrial settings.

The aim of this study is to identify nanoparticles within polydisperse aerosols of complex composition in the workplace air of a crude lead production facility using electron microscopy and X-ray fluorescence analysis.

Materials and Methods

Description of the production process. The study was conducted at a plant producing crude lead from secondary raw materials (scrap batteries and waste from copper smelting). The site houses two short drum furnaces equipped with extraction systems. The furnaces are loaded with prepared charge materials using a charging machine. Smelting is a reduction process at temperatures of 1,100–1,200 °C; crude lead and slag-dust product are periodically poured into moulds and transported further along the production line. A diagram of the site and sampling points is shown in Fig. 1.

Particle density analysis. Samples of settled dust were collected around the perimeter of the site near the sampling point at furnace 1 from horizontal surfaces free of overhanging elements. Collection was carried out using a natural-bristle brush into polypropylene zip-lock bags. A total of 10 dust samples were collected. Prior to analysis, the dust was filtered through a 100 µm mesh sieve and dried in an oven. The particle density of the settled dust was determined using the pycnometric method with a neutral liquid in accordance with GOST 5180–2015*. It should be noted that this standard was developed for soils; however, within the scope of this study, it was used to obtain estimated values for the density of settled dust as an accessible integrated indicator. PMS-0.65 silicone oil (SiliconSnab LLC, Russia) was used as the neutral liquid.

Aerosol analysis. For the study of the aerosol by scanning electron microscopy, 47-mm nylon (NY) disc membrane filters with a nominal pore size of 0.2 µm by Sartorius (Sartorius AG, Germany) were used; these had previously been evaluated for their ability to capture the nanofraction of the aerosol [27]. A custom-designed filter holder with a cone (intake orifice of 1.5 cm³) was used [28]. Samples were collected using an AirChek TOUCH Pump personal aspirator (SKC Inc., USA). During the smelting of lead-containing recycled materials, 16 aerosol samples were collected; sampling times ranged from 10 to 60 minutes depending on the stage of the process, at an air flow rate of 1.5 L/min (Table 1).

The filters were affixed to workbenches using conductive carbon tape for analysis. The filter surfaces were cleaned in a ZONE II ozone cleaning system (Hitachi, Japan). The samples were then coated with a 10-nm-thick conductive carbon layer using a Leica ACE600 vacuum sputtering system (Leica, Austria). The filter surfaces were imaged using a Hitachi REGULUS SU8220 ultra-high-resolution scanning electron microscope equipped with a five-section backscattered electron detector (Hitachi, Japan). The filter surfaces were examined at magnifications of ×600, ×1,500, ×6,000 and ×50,000. The number of fields of view ranged from 40 to 150 depending on the particle load on the filter. The lower threshold for the number of particles analyzed was 1,000 at each magnification.

The elemental composition of the nanoparticles was identified using an Ultim® Extreme windowless 100 mm² EDAX detector at an accelerating voltage of 20 kV. Prior to the analysis of each sample, the detector was optimized using a standard cobalt sample for X-ray fluorescence analysis. Both individual nanoparticles and clusters of up to several dozens of nanoparticles were analyzed, provided there were no larger particles in the field of view. The least number of spectra obtained for a single sample was 100.

Data analysis. Electron micrographs were analyzed using the open-source ImageJ software offering automated object detection, measurement of their projected area (A), the minimum diameter of the particle’s projected area (dm), and the aspect ratio (AR). For each particle, the equivalent projected diameter (dp) and the particle volume (V) were calculated.

The diameter dp represents the diameter of a circle having the same area as the dust particle, as projected onto a two-dimensional image, and is calculated as follows:

 dp = √(4 ∙ А) / π                                     (1)

The volume of the particles was determined following the approach of Anderson et al. [29], according to which the height of a particle is equal to its smallest dimension (the minimum diameter of the particle’s projected area):

V = (π ∙ А ∙ dm) / (6 ∙ √АR)                          (2)

To calculate the number concentrations, the particles falling within the range were counted, and for volume concentrations, the sum of the volumes of these particles was determined. Measurement results obtained at different magnifications were combined by means of a proportional conversion from large particles to small ones.

When calculating the mass concentrations of nanoparticles, their elemental composition and the identification of unique nanoparticle compounds (NPCs) formed during the production of crude lead, the following assumptions were made:

Thus, the total mass concentration of nanoparticles (Cm, µg/m³) was calculated using the formula:

 Cm нч = Cv нч ∙ ρ                                     (3)

where Cv is the total volumetric concentration of NPs, mm³/m³ and ρ is the density of particles, g/cm³.

The mass concentration of identified chemical elements (Xm, µg/m³) was calculated using the formula::

where xi is the mean relative weight percentage of the element in the sample, %, and xi is the relative weight percentage of the element in the i-specter, %.

Results

Electron microscopy findings revealed that the samples of suspended particles collected at the smelter’s workplace in a lead-containing scrap processing plant were characterized by size heterogeneity: the particle diameter varied over a wide range of 0.01–25 µm. The particles were formed as a result of a condensation process and therefore were predominantly regular spherical or cubic in shape (Fig. 2).

Furthermore, heterogeneity in the morphology of the aerosol particles was observed (Fig. 3).

The examination of the overall size distribution of the aerosols showed that the total count of aerosol particles ranged from 0.13 to 138.58 billion/m³, depending on the stage of the crude lead smelting (Table 2).

To calculate the mass concentrations of fine particles, the mean density of aerosol particles was determined by estimating the densities of the particulate matter. According to the study results, the density of dust particles in the furnace area was 4.76 ± 0.61 g/cm³ (mean ± standard deviation), with a minimum/maximum range of 3.84/6.05 g/cm³. When interpreting the results, it was taken into account that settled dust may differ in morphology, degree of agglomeration and phase composition from aerosol particles suspended in the workplace air. Consequently, the density values obtained were regarded as approximate and were used for a rough estimate of particle mass concentrations.

Analysis of the particle size distribution of the aerosols revealed the predominance of fine particles. However, the proportion of fine particles in the peak volume concentration did not exceed 2.1% (Fig. 3). The number, volume, and mass concentrations of fine particles ranged from 0.06 to 100.08 billion/m³, 0.006–9.958 million µm³/m³, and 0.03–47.47 µg/m³, respectively.

When investigating the elemental composition of fine particles, both individual particles and their aggregates were analyzed (Fig. 4).

Based on the results of the elemental analysis, 22 elements from the periodic table were identified in the composition of nanoparticles (Table 3).

Elemental analysis of the nano-sized aerosol generated during crude lead smelting revealed significant heterogeneity and variability in the elemental composition of the nanoparticles depending on the smelting stage and the operations carried out simultaneously (see Table 1). The most common elements in the nanoparticles were Pb, Na, Zn, As, Sn, and S. Their frequency of occurrence varied depending on the smelting stage; however, they were detected in the nanoparticles of all the samples tested. Furthermore, 90–100% of the NPs contained Pb, whilst the occurrence of other priority elements varied significantly (from 7–34 to 49–100%).

Taking into account the results of the preliminary analysis of the elemental composition of the NPs and the morphological heterogeneity of the aerosol particles, it was decided to identify NP compositional units (Fig. 3). The identification of the nanoparticle composition was based on data obtained during the elemental composition identification stage using the EDX method. For each aerosol sample, a list of nanoparticle compositions was determined. The number of nanoparticles belonging to each identified nanoparticle composition was then calculated. The number concentration of the nanoparticle composition (Cn con.np, units/m³) was determined using the formula:

 Cn con.np = Cn np ∙ (ncon.np / nnp)                        (5)

where Cn np is the total number concentration of nanoparticles, units/m³; ncon.np is the number of identified nanoparticle compositions in the sample; nnp is the total number of identified nanoparticles in the sample.

Thus, it was established that during the smelting process, up to 153 NP compounds (ranging from 8 to 45 types depending on the smelting stage) were generated in the workplace air, comprising between one and eight elements. NP compounds containing three to five elements prevailed. Table 4 shows the most common types of identified NP compounds and their concentrations.

Despite the diversity of the elemental composition of the nanoparticles, only six types of single-element nanoparticles were detected over the entire observation period – Na, Al, Zn, As, Sb, and Pb. Furthermore, the proportion of Pb, Sb, and Zn particles reached 12.6%, 1.4%, and 0.6% of the total number of particles at certain stages of smelting, respectively, whilst the proportions of other types of single-element particles did not exceed 0.22%.

Discussion

Microscopic findings showed that samples of suspended particles collected at the smelter’s workplace in a lead-containing scrap processing plant were characterized by size heterogeneity: particle diameters varied over a wide range of 0.01–25 µm. The study by Rovin et al. also indicates an upper size limit of 25 µm. However, in the present study, a large number of particles smaller than 100 nm were detected [30].

Furthermore, heterogeneity in the morphology of the aerosol particles was observed. The presence of multiple phases within the aerosol particles points to a mechanism of heterogeneous nucleation, in which vapours of more volatile elements form a surface layer on pre-existing particles of a different elemental composition.

The study of changes in the concentrations and particle size distribution of aerosols over time has established the following:

the aerosol formed is capable of persisting in the air for a long period even after the relevant process operations have ceased;

despite the presence of local ventilation systems, operations carried out simultaneously in other areas of the workshop can significantly affect the overall (background) concentration and composition of aerosols throughout the entire workshop;

the highest concentrations of large particles (≥ 1 µm in size) were observed during furnace charging (re-charging) and mechanical operations (e.g. cutting of tap holes) as disintegration aerosols; and

peak concentrations of submicron particles (< 1 µm), including fine particles, occurred during the discharge and pouring of molten materials (slag, dross, and crude lead), when the particles formed mainly by condensation.

Furnace operators, who monitored the smelting process and tapped the metal, were most exposed as they worked at the emission source. Thus, the spikes in fine particle concentrations were caused precisely by operations involving the molten metal – its tapping and transfer through the runners.

The study showed that the entire smelting period can be divided into four phases based on the qualitative elemental composition of the molten metal: 1) during the stages from ZP to P-4, Pb predominated in the molten metal, with a gradual decrease and increase in the proportions of Zn, Na, and Sn; 2) during the period from P-5 to P-7, Pb also predominated in the composition of the molten metal, but its proportion decreased significantly. At the same time, the proportion of Zn increased whilst that of Sn gradually decreased; 3) during stages P-8 and P-9, Pb continued to predominate in the composition of the crude lead, but the proportion of Zn decreased whilst that of As increased, and 4) during the period from VPSh-1 to VS-2, Zn has already predominated in the composition of the crude lead.

This diversity of qualitative characteristics of the elemental composition of crude lead is explained by differences in raw materials used during the smelting process. Under continuous production conditions, an average of four smelting runs were carried out per shift (two furnaces, each with two runs), during each of which either identical or unique raw materials could be used depending on production requirements (production volume targets, order volumes for the supply of specific products by the enterprise). These uncertainties introduce significant distortions into the identification of contributing factors and determination of their levels, thus preventing an objective assessment of the potential impact on workers’ health and elaboration of effective preventive measures.

No correlations were found between the size of ultrafine particle agglomerates and the number of all identified elements in the sample, nor with the particle concentrations. This may indicate the specific nature of the technical process (temperature regime, additives), leading to the formation of ultrafine particle agglomerates with a complex composition. Furthermore, as described earlier, there are possible scenarios in which the particles formed had a non-homogeneous composition rather than a multilayered structure. Primarily, nanoparticles might have formed from elements or substances with a high condensation temperature, whilst substances with a lower condensation temperature condensed directly onto their surfaces.

Conclusion

Application of a suite of electron microscopy techniques in combination with EDS enabled the detailed identification of nanoscale components within complex polydisperse aerosols. Owing to their high spatial resolution and elemental microanalysis, these methods provided reliable information not only on the morphology and dispersibility of the particles, but also on their unique multicomponent composition, enabling establishment of the patterns governing the formation of multicomponent particles and determination of their distribution across the phases of the technological process.

The approaches to identifying fine particles in the workplace air presented in this paper have provided valuable additional information on the risk factors for metallurgical workers; however, the methodology is labour-intensive and is suitable only for expert studies. Further development of the methodology should focus on its optimization. The most promising approach appears to be the integration of the proposed method with techniques for determining real-time particle counts in aerosols (for particles larger than 5 nm). Such an approach would significantly reduce labour and financial costs of investigating nanoparticles within polydisperse aerosols of complex composition.

* GOST 5180–2015 Soils. Methods for laboratory determination of physical characteristics. Moscow: Standartinform; 2016.