Chemical Composition of Thoracic Dust at Workplaces During Cement Production.

Objectives: Cement belongs to the most used building materials. Clinker is the major constituent of cement, and it is believed that the strong increase of pH after hydration of clinker minerals is responsible for the observed decline in lung function of cement production workers. Information on clinker exposure at workplaces in the cement production industry is scarse. The aims of this study are to determine the chemical composition of thoracic dust and to quantify workplace exposure to clinker in cement production.

Methods: The elemental composition of 1250 personal thoracic samples collected at workplaces in 15 plants located in 8 different countries (Estonia, Greece, Italy, Norway, Sweden, Switzerland, Spain, Turkey) was determined by inductively coupled plasma optical emission spectrometry (ICP-OES), separately for water- and acid-soluble fraction. Positive matrix factorization (PMF) was used to determine the contribution of different sources to the dust composition and to quantify the clinker content in 1227 of the thoracic samples. In addition, 107 material samples were analysed to facilitate interpretation of the factors obtained by PMF.

Results: The median thoracic mass concentrations varied for individual plants between 0.28 and 3.5 mg/m3. PMF with 8 water-soluble and 10 insoluble (i.e., acid-soluble) element concentrations yielded a five-factor solution: Ca, K, Na sulfates; silicates; insoluble clinker; soluble clinker-rich; and soluble Ca-rich. The clinker content of the samples was calculated as sum of the insoluble clinker and soluble clinker-rich factors. The median clinker fraction of all samples was 45% (range 0-95%), and varied between 20% and 70% for individual plants.

Discussion: The 5-factor solution of PMF was selected on the basis of several mathematical parameters recommended in the literature as well as the mineralogical interpretability of the factors. In addition, interpretation of the factors was supported by the measured apparent solubility of Al, K, Si, Fe, and to a lesser extent Ca in material samples. The total clinker content obtained in the present study is considerably lower than estimates based on the Ca concentrations in a sample, and somewhat lower than estimates based on Si concentrations after selective leaching with a methanol/maleic acid mixture. The clinker abundance in workplace dust of one plant investigated in the present contribution was also estimated in a recent study by electron microscopy, and the good agreement between both studies gives confidence in the results of PMF.

Conclusions: The clinker fraction in personal thoracic samples could be quantified from the chemical composition by positive matrix factorization. Our results allow for further epidemiological analyses of health effects in the cement production industry. As these estimates are more accurate for clinker exposure than aerosol mass, stronger associations with respiratory effects are expected if clinker is the main cause of these effects.