Lawrence Berkeley National Laboratory

Chemical reactions involving ozone of outdoor origin and indoor materials are known to be significant sources of formaldehyde and other irritant gas-phase oxidation products in the indoor environment. HVAC filters are exposed to particularly high ozone concentrations--close to outdoor levels. In this study, we investigated chemical processes taking place on the surface of filters that included fiberglass, polyester, cotton/polyester blend and synthetic (e.g., polyolefin) filter media. Ozone reactions were studied on unused filter media, and on filters that were deployed for 3 months in two different locations: at the Lawrence Berkeley National Laboratory and at the Port of Oakland. Specimens from each filter were exposed to ozone under controlled conditions in a laboratory flow tube at a constant flow of dry or humidified air (50percent relative humidity). Ozone was generated with a UV source upstream of the flow tube, and monitored using a photometric detector. Ozone breakthrough curves were recorded for each sample exposed to ~;;150 ppbv O3 for periods of ~;;1000 min, from which we estimated their uptake rate. Most experiments were performed at 1.3 L/min (corresponding to a face velocity of 0.013 m/s), except for a few tests performed at a higher airflow rate, to obtain a face velocity of 0.093 m/s, slightly closer to HVAC operation conditions. Formaldehyde and acetaldehyde, two oxidation byproducts, were collected downstream of the filter and quantified. Emissions of these volatile aldehydes were consistently higher under humidified air than under dry conditions, at which levels were near the limit of detection. Our results confirm that there are significant reactions of ozone as air containing ozone flows through HVAC filters, particularly when the filters are loaded with particles and the air is humidified. The amount of ozone reacted was not clearly related to the types of filter media, e.g., fiberglass versus synthetic. Specific fiberglass filters that were coated with an impaction oil showed significantly higher formaldehyde emissions than most other samples. Those emissions were magnified in the presence of particles (i.e., in used filters), and were observed even in the absence of ozone, which suggests that hydrolysis of filter binder or tackifier additives may be the reason for those high emissions. Mass balance calculations indicate that the emission rates of formaldehyde and acetaldehyde from the filters are generally not large enough to substantially increase indoor formaldehyde or acetaldehyde concentrations.