Relationship between sources and patterns of VOCs in indoor air

Abstract People spend most of their daytime in indoor environments. Their activities influence the composition of the indoor air by emitting volatile organic compounds (VOCs). The increasing number of different VOCs became the focus of attention in recent years as the question arises from the relationship between exposure to air pollutants and diseases. The present study of flats in Leipzig (Germany) is based on measurements of 60 different VOCs and is unique in the field of indoor air quality due to its enormous size of samples (n = 2 242) and questionnaire data. The main purpose of our analysis was to identify the sources and patterns that characterize airborne VOCs in occupied flats. We combined two methods, principal components analysis (PCA) and non–negative matrix factorization (NMF), to assign compounds to their origin and to understand the coinstantaneous existence of several VOCs. PCA clustering provided a source apportionment and yielded 10 principal components (PCs) with an explained variance of 72%. However, real indoor air quality is often affected by combined sources. NMF reveals characteristic compositions of VOCs in indoor environments and emphasizes that constantly recurring structures are not single sources, but rather fusions of them, so called patterns. Interpreting these sources, we realized that homes were strongly influenced by ventilation, human activities, furnishings, natural processes (such as solar radiation) or their combinations. The very large set of samples and the combination with questionnaires applied on this comprehensive assessment of VOCs allows generalizing the results to homes in middle–scale cities with minor industrial pollution. As a conclusion, single VOC–dose–response relationships are inopportune for situations when indoor sources occur in combination. Further studies are necessary to assess associated health risks.