For Presentation at the Air & Waste Management
Association's 91st Annual Meeting & Exhibition, June
14-18, 1998, San Diego, California
William A. Jacoby§ and Nicholas Chornet
National Renewable Energy Laboratory, 1617 Cole Boulevard, Golden, Colorado 80401-3393
§ current address: University of Missouri at Columbia, Department of Chemical Engineering, Columbia, Missouri 65211
Although the use of chemical filtration devices has shown to improve indoor air quality,3 commercial gas phase filtration has both chemical and economic disadvantages. In a previous study by the authors, the use of activated carbon/potassium permanganate filtration was shown to increase the indoor concentration of acetaldehyde over outside feed air, presumably from the partial oxidation of higher molecular weight volatile organic compounds (VOCs) in outdoor air.4 In addition, the use of ozone generators to reduce indoor VOCs may lead to IAQ complaints as low concentrations of ozone act as an irritant to the upper respiratory tract.
In more recent research, the use of ultraviolet radiation combined with a titanium dioxide catalytic surface has shown promise in the overall reduction of VOCs in air.5-8 Photocatalytic oxidation utilizes ultraviolet or near-ultraviolet ( < 385 nm) radiation to promote electrons from the valence band into the conduction band of a titanium dioxide semiconductor. Destruction of organic compounds takes place through reactions with molecular oxygen or through reactions with hydroxyl radicals and super-oxide ions formed after the initial production of highly reactive electron and hole pairs. The oxidation reaction with VOCs is heterogeneous catalysis, commonly referred to as photocatalytic oxidation.
Photocatalytic oxidation (PCO) provides significant advantages for the mitigation of pollutants associated with poor indoor air quality. Due to low pressure drop across the reactor and ambient temperature operation, PCO reactors can be incorporated into existing HVAC systems. Rather than transferring pollutants from the gas phase to the solid phase, PCO provides a reduction of absolute toxicity as the gaseous products from the complete photocatalytic oxidation of volatile organic compounds are carbon dioxide and water. Several authors have reported very high or complete conversion efficiencies of VOCs by photocatalytic oxidation reactors.
While PCO can produce complete oxidation of VOCs, a 'complete' conversion is difficult to define as a carbon mass balance is limited by the analytical instrumentation used to quantify possible intermediate products. Intermediate products from the oxidation of organic compounds are likely to contain aldehydes, ketones, esters, and acids which may be difficult to detect over wide concentration ranges using a single analytical technique.
Previous chemical and kinetic research on photocatalytic oxidation of organic compounds has dealt primarily with organic concentrations in the low to high parts-per-million by volume (ppmv) range which is more typical of chemical process stream concentrations than that associated with indoor air quality. In these studies, parts-per-billion by volume (ppbv) concentrations of intermediate VOC oxidation products such as formaldehyde and acetaldehyde are possible even under 'complete' oxidation parameters for a feed stream of ppmv concentrations. Aldehyde and ketone concentrations in the parts-per-billion range can present significant indoor air quality health problems. Only recently has a PCO technique been investigated at levels of organic compounds more closely associated with indoor air quality.9 An evaluation of PCO as a technique for the improvement of indoor air quality requires evaluation of the organic oxidation of gaseous contaminants in the sub parts-per-million range.
Cost effectiveness of PCO as an air cleaner technology has recently been reported.10 This analysis estimates current PCO technology to cost significantly more than granular activated carbon technology. A major factor in determining costs associated with PCO reactor operation is the ultraviolet light intensity required to destroy a range of organic compounds at the low inlet concentrations involved with indoor air quality issues. As with chemical kinetics, little research has been carried out on UV energy levels necessary to carry out oxidation of low level VOCs.
The assessment of a photocatalytic oxidation reactor used in the mitigation of indoor air quality problems requires an evaluation of conversion efficiency of the reactors for a range of organic compounds at various flow rates, inlet concentrations and ultraviolet light intensities. The objective of this research is to assess the photocatalytic oxidation of organic compounds at concentration levels more closely associated with indoor air quality issues. Specifically, the research objectives are to:
The design of the reactors and sampling unit allows for an easy change in flow rate, inlet concentration of carbonyl compounds and ultraviolet light intensity. A schematic of the experimental apparatus is shown in Figure 1.
The four PCO reactor designs used in this investigation are shown in Figure 2. Physical and operating characteristics of the four PCO reactors are listed in Table 1. The design objectives for a photocatalytic reactor are to mix photons, catalyst and reactants as intimately as possible. Control of UV light intensity is also desirable. The packed bed annular reactor, the coated wall annular reactor and the box reactor were used to evaluate the photocatalytic oxidation destruction of the low molecular weight carbonyl compounds as a function of flow rate. The flat bed PCO reactor was specifically designed for the light intensity studies.
Comparison of the packed bed annular reactor, the coated wall annular reactor, the box reactor and the flat bed reactor can provide insight into photon transport, mass transport and surface reaction processes. The coated-wall annular reactor is coated on the inside surface of its outer glass tube with the titanium dioxide photocatalyst. This design provides a well-characterized reactive catalyst surface along the length of the reactor body and allows uniform light distribution. The packed-bed annular reactor is similar in design to the coated-wall annular reactor, but has a greater surface area of titanium dioxide provided by the coated 3 mm glass beads packed within the annular volume. The box reactor has the largest reaction surface area provided by two sections of fiberglass 'furnace' filter (American Air Filter) coated with titanium dioxide and placed on either side of the ultraviolet lamp. The flat bed reactor utilizes the fiberglass filter material positioned under a flat quartz window that permits uniform light exposure and external variations in the ultraviolet radiation intensity.
Procedures for applying a thin, uniform coating of Degussa P-25 titanium dioxide are detailed elsewhere.7 The titanium dioxide surface of the fiberglass filter in the box and flat bed PCO reactors was applied with a Vega Inc. air brush and dried with a hot air dryer. The near ultraviolet and ultraviolet radiation was supplied by three different light sources. The two annular reactors used an 8-W fluorescent black light, Sylvania F8T5/BLB, with a maximum spectral intensity at 356 nm. The box reactor used a 4-W germicidal lamp, GE G4511, with a maximum spectral intensity at 254 nm. The flat bed reactor utilized two 40-W black lights, GE F40/BLB, with a maximum spectral intensity of 356 nm. Ultraviolet light intensity on the flat bed catalytic surface was varied by adjusting the distance between the lamps and the reactor quartz window. For the flat bed reactor, surface ultraviolet radiation intensity was measured at the quartz window using a Black-Ray Model J-221 radiometer. Average light intensity across the packed-bed, box and flat bed PCO reactors could not be measured due to shadowing created by the fiberglass or glass bead supporting materials.
Simultaneous carbonyl sampling, 1) downstream of the PCO reactor and 2) bypassing the PCO reactor, was carried out using a dual diaphragm pump. Volume flow measurements were carried out using identical 0-3 L/min rotameters calibrated against a Singer Model DTM-115 dry test meter. Sampling times varied from 15 minutes to 12 hours depending on the inlet concentration of the carbonyl compounds. All samples were taken at ambient room temperature (23-25 C). Relative humidity was not considered a variable in this investigation, although it is expected to influence reactor efficiency.9 Estimates of the relative humidity were 30-40% for the duration of the sampling procedures. Following sample collection, the cartridges were capped, sealed and stored at -10 C until laboratory analysis could be performed, typically within a week of sampling.
The sampling and analysis procedures of formaldehyde, acetaldehyde and acetone are detailed elsewhere 11,12. Gaseous carbonyls are trapped on a silica cartridge coated with acidified 2,4-dinitrophenyl hydrazine. The trapped carbonyls are extracted with acetonitrile and analyzed by high performance liquid chromatography. The chromatographic analyses used a Varian Model 5020 HPLC with Varian Model 9090 autosampler and Varian Model 2550 variable wavelength detector. All carbonyl concentrations are field-blank corrected. The statistically determined detection limits for the carbonyl analyses were typically in the range of 0.2 to 0.5 ppbv. The range of carbonyl compound concentrations in the inlet stream varied from 2 to approximately 1500 ppbv.
Low concentrations of formaldehyde, acetaldehyde and acetone were achieved by using ambient laboratory air. Higher concentrations of formaldehyde were achieved through the use of a Teflon-walled permeation tube containing approximately 50 mg of paraformaldehyde. With the exception of the experiment involving ppmv levels of trichloroethylene, there was no attempt to hold the inlet concentrations of organic compounds constant. The dual, simultaneous sampling from a single carbonyl source allows for an accurate measurement of carbonyl compound destruction independent of variations in the inlet source concentration.
RESULTS AND DISCUSSION
An experiment utilizing a 50 ppmv inlet concentration of ethanol flowing through a flat bed PCO reactor was carried out to validate the carbonyl sampling procedures. The experiment yielded approximate concentrations of 0.68 ppmv formaldehyde, 6.5 ppmv acetaldehyde and negligible acetone. The three carbonyl compounds represent intermediate products from the photocatalytic oxidation of ethanol, and are consistent with the results of Nimlos, et al.13 This initial experiment verifies that the 2,4-dinitrophenyl hydrazine analysis technique is capable of measuring the sub part-per-million concentrations of carbonyl compounds expected in the photocatalytic oxidation of indoor levels of organic compounds. Furthermore, the experiment points out the need for analytical procedures capable of measuring a wide range of carbonyl concentrations associated with a photocatalytic oxidation experiment.
Figures 3 and 4 represent the percentage destruction of formaldehyde, acetaldehyde and acetone for the two annular PCO reactors, coated-wall and packed-bed, using an identical ultraviolet light source. A comparison between the two annular reactors is aided by the similar design and residence times (Table 1). The destruction of formaldehyde and acetone is nearly 100 percent for both the coated-wall and packed-bed annular reactor designs over the range of 0.5 to 2.0 L/min volume flow rates through the reactor. The data from Obee9 suggest that the PCO oxidation rate for formaldehyde decreases below the maximum at 10 ppmv, yet the results of these experiments indicate the nearly 100 percent destruction of ppbv levels of formaldehyde by photocatalytic oxidation.
The photocatalytic destruction of acetaldehyde shown in Figures 3 and 4 is noticeably lower than that for formaldehyde and acetone. This is consistent with a comparison of oxidation rates of formaldehyde and acetaldehyde by Nimlos, et al.13 at the ppmv concentration level. Although not definitive, the results in Figure 4 suggest a drop in acetaldehyde destruction for the packed-bed annular reactor at the higher flow rates above 1 L/min. This could be caused by a combination of lower oxidation kinetics for acetaldehyde, insufficient reactor residence time, or inadequate radiation intensity for molecules tunneling in the dark side of the reactor caused by shadowing from the coated glass beads.
Quantitation confidence is reduced for carbonyl concentrations at, or near, the detection limit. For inlet carbonyl concentrations of 2 ppbv, a 90 percent destruction from the PCO reactor results in an outlet carbonyl concentrations at or below the detection limit of the chromatographic analysis. Therefore, all carbonyl data below an inlet concentration of 2 ppbv were discarded.
Low molecular weight carbonyls below 15 ppbv do not pose a health threat for indoor air quality. The inlet concentration of formaldehyde was augmented above ambient laboratory levels to produce a range of formaldehyde concentrations more typical of IAQ situations. A summary of formaldehyde destruction by photocatalytic oxidation over a wide range of concentrations is shown in Figure 5. Over the 2-1500 ppbv range of formaldehyde concentrations there is nearly 100 percent destruction by all three photocatalytic reactor designs. The coated-wall and packed-bed PCO reactors have considerably less titanium dioxide loading and shorter residence times but more light intensity compared to the box PCO reactor (Table 1).
An important factor in the economic evaluation of PCO reactors is the ultraviolet light intensity required for efficient destruction of organic compounds. Figure 6 represents the photocatalytic oxidation destruction of formaldehyde for light intensities ranging between 0.1 and 3.8 mW/cm2. Even at the lowest ultraviolet intensities used in this project, 0.1 mW/cm2, destruction of formaldehyde is virtually 100 percent.
The light intensity data in Figure 6 also show the destruction of formaldehyde in moles formaldehyde per Joule of ultraviolet radiation. As expected, the photooxidation efficiency is highest at the lower light intensities. The photooxidation efficiency of 0.06 moles/Joule at the lowest light intensity compares favorably with photooxidation destruction efficiency of the well characterized trichloroethylene (TCE) in the ppmv range (see Figure 6).
Under dark conditions, with no ultraviolet light source and the reactor shielded from room light, the PCO reactor removed 35 percent of the formaldehyde as shown in Figure 6. This suggests that formaldehyde, to a small extent, is reactive or adsorbed at the titanium dioxide surface in the absence of incident ultraviolet radiation.
Evaluation of photooxidation efficiency for carbonyl compounds between the three PCO reactor designs is precluded due to the nearly 100 percent destruction by all three reactor designs. To assist the evaluation of the reactor designs, the organic compound was changed to trichloroethylene and the inlet concentration increased to the ppmv range. TCE concentrations were monitored by FT-IR instrumentation, and results are shown in Figure 7. The photocatalytic oxidation efficiency for the three PCO reactors is similar, but the box reactor shows the highest efficiency, followed by the coated-wall annular reactor, then the packed-bed annular reactor. Given that the box reactor has a lower light intensity than the two annular reactors (Table 1), this suggests the ability to further reduce ultraviolet light intensity and maintain efficient VOC destruction.
This paper addresses photocatalytic oxidation as a means of mitigating volatile organic compounds as associated with poor indoor air quality. Four PCO reactor designs were evaluated at sub parts-per-million concentrations ranges of carbonyl compounds.
The operating conditions selected for this project did not reach mass transport limits and the nearly complete conversion of carbonyls prohibited the calculation of intrinsic reaction rates. However, the results clearly demonstrate that (1) the destruction of low molecular weight carbonyl compounds at concentrations associated with IAQ issues is nearly 100 percent using photocatalytic oxidation techniques, (2) within the parameters used in this project, the destruction of carbonyl compounds was independent of reactor design, residence time (flow rate) and light intensity and (3) the box reactor shows the greatest destruction efficiency for trichloroethylene.
Additional experimentation must be carried out to evaluate the scale-up viability of photocatalytic oxidation as a technique to mitigate VOCs associated with IAQ issues. The data indicate efficient destruction of VOCs with light intensities below 0.1 mW/cm2 and reactor residence times under 1 second, suggesting that the operating conditions were not optimized for this study. Two major questions will be the focus of our continuing research on efficient VOC destruction under IAQ conditions: How far can the light intensity be lowered, and how much can the residence time be shortened?
The authors would like to thank the National Renewal Energy Laboratory (NREL) for providing laboratory space and equipment used in this project. This work was supported by the University of Colorado at Denver Department of Chemistry and the Department of Energy Solar Industrial Program at NREL.
This paper, in part, was presented at Engineering Solutions to Indoor Air Quality Problems sponsored by the US EPA and AWMA, at Research Triangle Park, NC, July, 1997.
1. Lanning, John A. and Anderson, Larry G.; in Proc. of 1993 Annual Meeting of Air & Waste Management Assoc.; Denver, Colorado; June, 1993; No. 93-WA71.03.
2. Godish, T., Indoor Air Pollution Control, Lewis Publishers, Chelsea MI, 1990, Chapter 1.
3. VanOsdell, Douglas W. and Sparks, Leslie E.; ASHRAE Journal, 1995, February, 34-40.
4. Wenger, Kelly; Lanning, John A. and Anderson, Larry G; in Proc. of 1996 Annual Meeting of Air & Waste Management Assoc.; Nashville, Tennessee; June, 1996; No. 96-WP85.04.
5. Dibble, L.A. and Raupp, G.B.; Catal. Lett., 1990, 4, 345-351.
6. Peral, J. and Ollis, D.F.; Journal of Catalysis, 1992, 136, 554-560.
7. Jacoby, William A.; Blake, Daniel M.; et al.; J. Air & Waste Management Assoc., 1996, 46, 891-898.
8. Anderson, Marc A.; Zeltner, Walter A.; Fu, Xianzhi; et al.; Center for Indoor Air Research Currents, 1996, December; 1,4.
9. Obee, Timothy N.; Env. Science & Tech., 1996, 30, 3578-3584.
10. Henschel, Bruce; Inside IAQ, U.S. Environmental Protection Agency: Research Triangle Park, North Carolina; Fall/Winter, 1996, EPA-600-N-97-001, 1-3.
11. Anderson, L.G.; Wolfe, P.; Barrell, R.A. and Lanning, J.A.; in Alternative Fuels and the Environment, F.S. Sterrett, Ed.; Lewis Publishers; Boca Raton, Florida: 1994, Chap. 5.
12. Anderson, Larry G.; Lanning, John A.; et al.; Atmospheric Environment, 1996, 30, 2113-2123.
13. Nimlos, Mark R; Wolfrum, Edward J.; Brewer, Matthew L.; et al.; Env. Science & Tech.,
1996, 30, 3102-3110.
Table 1. Physical and Operating Characteristics of Various Photocatalytic Oxidation Reactors
|PCO Reactor Type||Symbol||Reactor Dimensions||Loadingg TiO2||, Residence @ 1.0 L/min||Light Sources|
|Coated-wall Annular||CW||1.75 cm OD 1.43 cm ID 18.10 cm length||0.034 g||0.87 sec||8W Black 356 nm 4.0 mW/cm2|
|Packed-bed Annular (glass beads)||PB||3.18 cm OD 1.43 cm ID 9.21 cm length||0.075 g||1.40 sec||8W Black 356 nm|
|Box (2 fiberglass filters)||BX||13.66 cm 6.67 cm 2.54x2 cm length||14.2 g||40 sec||4W Germicidal 254 nm|
|Flat Bed (fiberglass filter)||FB||5.56 cm 1.75 cm 9.37 cm length||1.35 g||5.48 sec||2-40W Black 356 nm|