Life Cycle Inventory Analysis
The assessment of indoor exposure needs to be facilitated by including emission factors, intake fractions and human-toxicological effect factors for indoor air sources into existing LCA software tools and databases. The present study helps by providing the methodological framework to estimate intake fractions in indoor settings in a structured, transparent and consistent manner. However, indoor emission data also need to be provided in inventory databases, similar to those available for outdoor emissions. This could be achieved by including an indoor air compartment, in addition to the existing air, water, and soil compartments in life cycle inventory databases, e.g., Ecoinvent (51
). Further work is planned to establish a ready-to-use list of relevant emission factors that can be incorporated in the life cycle inventory analysis.
Model Choice and Parameter Values
In LCA, information about specific exposures will not always be available. This will often restrict the choice of the model to the one-box model. The one-box model seems to be a good default choice, as the level of detail matches that of environmental models used in LCIA. More sophisticated models with indoor spatial differentiation may be used as well, if specific information, for instance on the spatial distribution of sources and people in the room, is available (Table ). However, in this case, the level of detail would deviate from the environmental fate and exposure models commonly used, as the latter do not consider inhomogeneous mixing within the environmental media.
Setting up a list of recommended values for exposure parameters may be, in general, difficult. In this paper, we tried to provide ranges of parameters for the models that were considered suitable for use in LCA (Tables and ). However, these ranges are very broad. Especially with regard to the more abstract, but very sensitive parameters, such as the air exchange rates between the conceptual inner and outer boxes in a two-zone model or the eddy-diffusion constant, it is often difficult to find representative values. Contaminant dispersion phenomena within the rooms can be influenced by complex interactions between variables such as the room geometry, the direction of the principal air flows, and the presence and movements of occupants (52
). The models identified suitable for use in LCA (Table ) do not take into account such detailed information. Further, the intake fraction in indoor environments was assumed to not depend on the chemical. This is a valid assumption if removal by ventilation is large in comparison to adsorption to surfaces and degradation (which is often the case in household and occupational settings). However, intake fractions may be reduced significantly by sorption to indoor surfaces (58
), especially for strongly sorbing substances in furnished residential homes.
For generic LCAs, a good approach is to calculate intake fractions for several generic workplace and household environments, which are characterized by air exchange rates, volumes, and numbers of people exposed. The parameter values for the indoor models (i.e., room volumes, air exchange rates, etc.) may vary geographically, e.g., because of climate conditions, cultural aspects, or different ventilation practices. For instance, the number of people per cubic meter working in the chemical industry in countries with cheap labor costs, such as China or India, is probably much higher than in industrial countries with a high degree of automation. Therefore, an important requirement for the final implementation of the model within the USEtox model is that the user can adapt the parameter values to his or her specific circumstances and that default parameter lists for various workplace settings and geographical regions are provided to facilitate the application.
The results assessed with the generic characterization factors based on the one-box model calculations will give only an indication of whether indoor exposure may be important. In such cases, it is advisable to refine the model parameter values or even change the model according to Figure a. The implementation of various indoor model options into the USEtox model, among which the user can choose, will make this recommendation feasible as well for LCA practitioners with limited time availability.
The use of a model can be circumvented in the case that monitored concentration values and production volumes are available. Instead of multiplying the emissions to the intake fractions, as usually done in LCA, the amounts taken in by the people exposed would be directly calculated from the monitored concentrations and the number and inhalation rates of people. This approach requires that pollutant concentrations can be directly linked to the functional unit (source appointment), which is possible in some indoor settings.
In a later stage, indoor exposure to radioactive gases such as radon can also be incorporated within the impact category “radiation” in LCIA methods such as Eco-Indicator 99, similarly to the framework shown in this paper. This is especially important for household settings, where radon can be an important factor for the total health damage as a result of indoor exposure.
Routine Assessment of Indoor Exposure within Life Cycle Assessment
The framework suggested in this paper is the first in putting forward a general procedure for indoor exposure assessment within LCA. From a practical point of view this is relevant, as the model results suggest that intake fractions from indoor emissions are often larger than intake fractions from outdoor emissions. This finding is confirmed by previous studies (6
) showing that indoor chemical concentrations often surpass outdoor concentrations by many orders of magnitude. This stresses the need to consider indoor exposure in LCA. It could even lead to human toxicity becoming a dominant impact category for certain products such as paints, furniture, or carpets. A routine assessment of indoor exposure in LCA will be facilitated by including the indoor model in the USEtox model (20
). Similar developments can be anticipated for the field of risk assessment, as European REACH legislation also calls for exposure scenarios, including worker or consumer exposure, for example. Such integrated assessment will point to the most important exposure pathways and improvement potentials, considering the whole life cycle of chemicals (60
). Moreover, the past has witnessed several cases in which chemicals were banned for one reason, such as ecological impacts, but got substituted by chemicals with other problems, i.e., occupational health effects (e.g., the market introduction of n-hexane/acetone based brake cleaning products due to air quality rules in California in 1990 (61
)). Such trade-offs between the various possible effects of chemicals can be revealed when applying integrated models for indoor and outdoor exposure and, ultimately, such problem shifting may be avoided.