|Title||Evaluation of a Combined Ultraviolet Photocatalytic Oxidation (UVPCO) / Chemisorbent Air Cleaner for Indoor Air Applications|
|Year of Publication||2007|
|Authors||Alfred T Hodgson, Hugo Destaillats, Toshifumi Hotchi, William J Fisk|
We previously reported that gas-phase byproducts of incomplete oxidation were generated when a prototype ultraviolet photocatalytic oxidation (UVPCO) air cleaner was operated in the laboratory with indoor-relevant mixtures of VOCs at realistic concentrations. Under these conditions, there was net production of formaldehyde and acetaldehyde, two important indoor air toxicants. Here, we further explore the issue of byproduct generation. Using the same UVPCO air cleaner, we conducted experiments to identify common VOCs that lead to the production of formaldehyde and acetaldehyde and to quantify their production rates. We sought to reduce the production of formaldehyde and acetaldehyde to acceptable levels by employing different chemisorbent scrubbers downstream of the UVPCO device. Additionally, we made preliminary measurements to estimate the capacity and expected lifetime of the chemisorbent media. For most experiments, the system was operated at 680 - 780 m3/h (400 - 460 cfm). A set of experiments was conducted with common VOCs introduced into the UVPCO device individually and in mixture. Compound conversion efficiencies and the production of formaldehyde and acetaldehyde were determined by comparison of compound concentrations upstream and downstream of the reactor. There was general agreement between compound conversions efficiencies determined individually and in the mixture. This suggests that competition among compounds for active sites on the photocatalyst surface will not limit the performance of the UVPCO device when the total VOC concentration is low. A possible exception was the very volatile alcohols, for which there were some indications of competitive adsorption. The results also showed that formaldehyde was produced from many commonly encountered VOCs, while acetaldehyde was generated by specific VOCs, particularly ethanol. The implication is that formaldehyde concentrations are likely to increase when an effective UVPCO air cleaner is used in buildings containing typical VOC sources. The magnitude of the expected increase will depend upon a number of interrelated factors. Series of experiments were conducted to determine if the oxidizer, sodium permanganate (NaMnO4·H2O), has sufficient reaction rates and capacity to counteract formaldehyde and acetaldehyde production and enable a 50 % reduction in building ventilation rate without net increases in indoor aldehyde concentrations. A commercially produced filter element and two laboratory-fabricated media beds containing NaMnO4·H2O chemisorbent media were evaluated. The effectiveness of a device for removal of formaldehyde, acetaldehyde and other VOCs was determined by measurement of concentrations immediately upstream and downstream of the device. In some experiments, conversion efficiencies and byproduct generation by the UVPCO device also were determined. Six experiments were conducted with the commercial filter element installed downstream of the UVPCO reactor. Eleven experiments were conducted with a single panel media bed (30 cm by 61 cm by 2.5 cm deep) installed downstream of the UVPCO reactor; in these, the effects of temperature and air residence time on conversion efficiency were examined. Two experiments were conducted with a four-panel, folded, media bed (approximately four times the size of the single panel bed) installed downstream of the reactor. Because the commercial unit contained activated carbon as an additional component, it was effective at removing lower volatility compounds that typically have low oxidation rates in the UVPCO reactor. The filter element also met the minimum efficiency objective for formaldehyde. However, the removal of acetaldehyde was less than required. The air residence time in the single panel bed was not optimized as the removal efficiencies for both formaldehyde and acetaldehyde were strongly inversely related to the air flow rate through the device. In addition, the acetaldehyde removal efficiency decreased to less than 10 % with extended use of the device. The folded bed was considerably more effective; formaldehyde was removed with greater than 90 % efficiency, and acetaldehyde was removed at about 70 % efficiency. With the combined UVPCO/chemisorbent system, the net removal efficiencies for formaldehyde and acetaldehyde were 90 % and 40 %, respectively. Two pairs of replicated experiments were conducted with the UVPCO system operating within a 50-m3 environmental chamber in a simulated HVAC mode with recirculation of chamber air. For one pair, the UVPCO air cleaner was operated alone, and for the other, the combined system of UVPCO air cleaner plus a downstream chemisorbent was used. The results showed that the chemisorbent media contributed substantially to the removal of VOCs in this mode. Concentrations were pulled down within the first hour. Net reductions for formaldehyde and acetaldehyde at near steady-state conditions were in the range of 50 to 70 %. From an analysis of NaMnO4·H2O in new and used media and the conditions of the experiments with the single panel media bed, we estimated that, on average, about nine moles of NaMnO4·H2O were needed to mineralize one mole of VOCs, and about three moles of the reactant were needed to mineralize one mole of carbon. These values were used to make estimates of the media consumption rate for the experimental conditions and for a hypothetical building application. In summary, the use of a multi-panel, folded scrubber filled with NaMnO4·H2O chemisorbent media downstream of the prototype UVPCO air cleaner effectively counteracted the generation of formaldehyde and acetaldehyde due to incomplete oxidation of VOCs in the UVPCO reactor. Thus, this combined UVPCO air cleaner and chemisorbent system appears to have sufficient VOC removal efficiency to enable a 50 % reduction in ventilation rate without increasing indoor aldehyde concentrations.
|LBNL Report Number|| |