Prior research suggests that chemical processes taking place on the surface of particle filters employed in buildings may lead to the formation of harmful secondary byproducts. We investigated ozone reactions with fiberglass, polyester, cotton/polyester and polyolefin filter media, as well as hydrolysis of filter media additives. Studies were carried out on unused media, and on filters that were installed for 3 months in buildings at two different locations in the San Francisco Bay Area. Specimens from each filter media were exposed to {approx}150 ppbv ozone in a flow tube under a constant flow of dry or humidified air (50percent RH). Ozone breakthrough was recorded for each sample over periods of {approx}1000 min; the ozone uptake rate was calculated for an initial transient period and for steady-state conditions. While ozone uptake was observed in all cases, we did not observe significant differences in the uptake rate and capacity for the various types of filter media tested. Most experiments were performed at an airflow rate of 1.3 L/min (face velocity = 0.013 m/s), and a few tests were also run at higher rates (8 to 10 L/min). Formaldehyde and acetaldehyde, two oxidation byproducts, were quantified downstream of each sample. Those aldehydes (m/zmore » 31 and 45) and other volatile byproducts (m/z 57, 59, 61 and 101) were also detected in real-time using Proton-Transfer Reaction - Mass Spectrometry (PTR-MS). Low-ppbv byproduct emissions were consistently higher under humidified air than under dry conditions, and were higher when the filters were loaded with particles, as compared with unused filters. No significant differences were observed when ozone reacted over various types of filter media. Fiberglass filters heavily coated with impaction oil (tackifier) showed higher formaldehyde emissions than other samples. Those emissions were particularly high in the case of used filters, and were observed even in the absence of ozone, suggesting that hydrolysis of additives, rather than ozonolysis, is the main formaldehyde source in those filters. Emission rates of formaldehyde and acetaldehyde were not found to be large enough to substantially increase indoor concentrations in typical building scenarios. Nevertheless, ozone reactions on HVAC filters cannot be ignored as a source of low levels of indoor irritants.« less