ARTICLE IN PRESS

1352-2310/$ - se

doi:10.1016/j.at

�CorrespondHealth Sciences

and Rutgers Un

08854, USA. T

E-mail addr

Atmospheric Environment 40 (2006) 6122–6133

www.elsevier.com/locate/atmosenv

Factors affecting ozone removal rates in a simulatedaircraft cabin environment

Gyongyi Tamasa, Charles J. Weschlera,b,�, Zsolt Bako-Biroa,David P. Wyona, Peter Strøm-Tejsena

aInternational Centre for Indoor Environment and Energy (ICIEE), Department of Mechanical Engineering,

Technical University of Denmark, 2800 Kgs. Lyngby, DenmarkbEnvironmental and Occupational Health Sciences Institute (UMDNJ-RW Johnson Medical School and Rutgers University),

Piscataway, NJ 08854, USA

Received 15 December 2005; received in revised form 8 May 2006; accepted 10 May 2006

Abstract

Ozone concentrations were measured concurrently inside a simulated aircraft cabin and in the airstream providing

ventilation air to the cabin. Ozone decay rates were also measured after cessation of ozone injection into the supply

airstream. By systematically varying the presence or absence of people, soiled T-shirts, aircraft seats and a used HEPA

filter, we have been able in the course of 24 experiments to isolate the contributions of these and other factors to the

removal of ozone from the cabin air. In the case of this simulated aircraft, people were responsible for almost 60% of the

ozone removal occurring within the cabin and recirculation system; respiration can only have been responsible for about

4% of this removal. The aircraft seats removed about 25% of the ozone; the loaded HEPA filter, 7%; and the other

surfaces, 10%. A T-shirt that had been slept in overnight removed roughly 70% as much ozone as a person, indicating the

importance of skin oils in ozone removal. The presence of the used HEPA filter in the recirculated airstream reduced the

perceived air quality. Over a 5-h period, the overall ozone removal rate by cabin surfaces decreased at �3%h�1. With

people present, the measured ratio of ozone’s concentration in the cabin versus that outside the cabin was 0.15–0.21,

smaller than levels reported in the literature. The results reinforce the conclusion that the optimal way to reduce people’s

exposure to both ozone and ozone oxidation products is to efficiently remove ozone from the air supply system of an

aircraft.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Aircraft cabin air quality; Indoor air chemistry; Heterogeneous chemistry; Human bioeffluents; Skin oil; Deposition velocities

e front matter r 2006 Elsevier Ltd. All rights reserved

mosenv.2006.05.034

ing author. Environmental and Occupational

Institute (UMDNJ-RW Johnson Medical School

iversity), 170 Frelinghuysen Rd., Piscataway, NJ

el.: +1 732 235 4114; fax: +1 732 235 4569.

ess: weschlch@umdnj.edu (C.J. Weschler).

1. Introduction

The presence of ozone in the indoor environmenthas received increased attention over the last decadedue to the recognition of its direct effect on humanhealth and its important contribution to indoorchemistry. When inhaled, ozone can damage thelung cells and aggravate chronic diseases such as

.

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–6133 6123

emphysema, bronchitis and asthma (US EPA,2006). Based on data from 95 US urban commu-nities during a 14-year period, Bell et al. (2004)estimated a 0.52% increase in daily mortality for a10 ppb increase in the previous week’s local ambientozone level. Ozone also initiates reactions thatgenerate oxidation products, some of which arerecognized irritants and potential health hazards(e.g., Weschler et al., 1992; Morrison and Nazaroff,2002a; Nazaroff and Weschler, 2004; Wisthaleret al., 2005). The negative effects of productsoriginating from some of these reactions have beendirectly evaluated with human subjects (Klenø andWolkoff, 2004; Nøjgaard et al., 2005; Tamas et al.,2006). The US EPA (2006) health-based standard forozone is 120ppb for 1 h and 80ppb for 8 h exposures.

At typical cruising altitudes, the concentration ofozone in the cabin of a commercial aircraft can besignificantly higher than it is in normal indoorenvironments. This is due to the high concentration(500–800 ppb) of ozone in the outside air at higheraltitudes (410 000m; SAE International, 2000) thatis transported through the aircraft ventilationsystem to the cockpit and passenger area. A recentstudy using passive samplers on 106 trans-U.S. andtrans-Pacific flight segments reported average con-centrations of 80 (730) ppb in airplane cabinsduring winter and early spring; 20% of themeasurements exceeded 100 ppb and 11% exceeded120 ppb (Spengler et al., 2004). Spicer et al. (2004)continuously measured the ozone levels on fourflights within the continental United States using areal time ozone monitor. On all four flights, theauthors found that ozone levels increased fromtakeoff until cruise altitude had been reached. Atcruise altitude average ozone levels were between 31and 106 ppb.

Catalytic converters are sometimes installed in theair supply system of aircraft to reduce ozoneexposures of passengers and crew. Without suchconverters or when a converter malfunctions, alarge amount of ozone may pass into the cabin. The‘‘retention ratio’’ is defined as the ratio between theozone concentration in the cabin and in the airoutside the cabin (ambient air) when no devices arepresent to deliberately remove ozone. Averageretention ratios of 0.47 and 0.83 have beenmeasured for a Boeing 747-100 and 747SP, respec-tively (Nastrom et al., 1980). The default retentionratio for demonstrating compliance with the USFederal Aviation Administration (FAA) regulationsaddressing cabin ozone levels is 0.7 (SAE AIR910

cited in NRC, 2002). Ozone is partly removed whenpassing through an aircraft air delivery systemincluding air conditioning units and surfaces in therecirculation system, and is further removed bysurfaces in the cabin, including the clothing,exposed skin and exposed hair of passengers. Theremoval is due to surface reactions and simpledecomposition (Weschler, 2000). The removal/reac-tion of ozone with various materials has beenstudied both in real environments (see Weschler,2000; Grøntoft and Raychaudhuri, 2004) and underspecial laboratory conditions (Klenø et al., 2001).A few studies have examined the interaction ofozone with ventilation filters (Bek +o et al., 2005,2006; Hyttinen et al., 2003, 2006). The effect ofpeople on ozone removal has not been system-atically investigated, but some data are availablefrom air quality studies made in indoor environ-ments with and without occupants (Bako-Biroet al., 2005).

Since the late 1980s, little new data has beenpublished on ozone removal rates in aircraft cabinenvironments. Wisthaler et al. (2005) examined theproducts formed when ozone, at concentrationstypical of those encountered at cruising altitudes, ispresent in a simulated commercial aircraft cabin(the same simulated cabin as was used in the presentstudy). Although the study was conducted withouthuman occupants, T-shirts that had been slept in theprevious night were used as surrogates for some ofthe less volatile bioeffluents associated with passen-gers and crew. Ozone-initiated reactions signifi-cantly altered the composition of the cabin air.Concentrations of aldehydes (saturated and unsa-turated) and squalene oxidation products weremuch larger when ozone was present.

The present study examines the effects of variousfactors (people and their clothing, aircraft seats,other surfaces within the cabin, ventilation filters)on the ozone removal rate within a simulatedaircraft cabin. Ozone removal by such processesinfluences the exposure of passengers and crew toboth ozone and the products of ozone-initiatedchemistry. This information can be used to betterdetermine the need and required efficiency of ozoneremoval devices in today’s aircraft.

2. Methods

Over a 2-year period (2004 and 2005) experimentswere conducted to study ozone dynamics in asimulated aircraft environment at the Technical

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–61336124

University of Denmark. In each of these experiments,the simulated aircraft cabin and the integratedventilation system were exposed to elevated ozoneconcentrations and possible interactions of ozonewith various sinks were evaluated.

2.1. Description of simulated aircraft cabin

The simulated aircraft cabin contained 21 seats(three rows of seven). As illustrated in Fig. 1, it wasinstalled inside an existing climate chamber that iscapable of providing two separately controlled airsupply streams. One of these was used to cool theclimate chamber, to ensure that the rate of heat lossthrough the simulated cabin walls was realistic, theother to ventilate the simulated cabin.

The total airflow to the cabin, including recircu-lated air, was always 200L s�1 (equivalent to 23 h�1)of which 25–75L s�1 (3–8.8 h�1) was outdoor air,depending on the experimental condition. Therecommended value for the outdoor airflow rate,as specified by Federal Aviation Regulation (FAR25), would be 75L s�1 at altitude if this cabin had 16occupants. The cabin air temperature was con-trolled by the ventilation system at 23.370.3 1C.Relative humidity varied with the experimentalconditions and spanned the range from 2% to24% (in all but three conditions it was below 15%).Experiments were conducted at ground level baro-metric pressure. The pressure in the simulated cabinwas 4–8 Pa higher than that in the outer chamber inorder to avoid contamination of the cabin air withair that might otherwise have infiltrated in anuncontrolled manner.

Chiller

Outdoor Air

Filter +

chamber

Pollution Chamber

Dehumidifier

System for external

C

Fig. 1. Schematic of the simulated air

Emissions to the cabin air, other than thosedescribed above, occurred from the materials andfittings specified in this paragraph. Used aircraftseats were obtained from a supplier of aftermarketparts. The carpet had been in use in a passengeraircraft for the duration of its normal service life.The total surface area of the carpet was 15.6m2

while that of the aircraft seats was 24.6m2. Thewalls consisted of six panels with windows from aused aircraft cabin interior. For additional detailsregarding the simulated cabin, see Wisthaler et al.(2005).

2.2. Generation of ozone

All of the outdoor supply air first passed througha 10-m3 chamber (designated ‘‘pollution chamber’’in Fig. 1). Pure oxygen (99.9999%) from acompressed gas cylinder flowed through six UVozone generators, providing ozone to this chamberand thus to the cabin. Given the high rate of totalair supply (200L s�1) relative to the cabin volume,the mixing time was short (o5min). The rate atwhich ozone was generated was regulated by turn-ing one or more ozone generators on or off.

2.3. Physical measurements

Cabin temperature, relative humidity and pres-sure; outdoor air supply rate; and total air supplyrate were monitored continuously during theexperiments. Between replicate conditions within aset of experiments, the outdoor and total supply

Exhaust

Exhaust

HEPAfilter

hamber

Cabin

craft cabin’s ventilation system.

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–6133 6125

rates varied by less than 3%. The ozone concentra-tion in the cabin air and in the pollution chamberwas continuously measured with two UV photo-metric analyzers operating at 254 nm (Dasibi 1003-AH and Dasibi 1003-RS). These instruments have arange of 0–500 ppb, a sensitivity of 1 ppb, and aprecision of 1 ppb or 71%, whichever is greater.The sampling points for the ozone instruments wereclose to the exhaust of the pollution chamber andnear the center of the cabin at a height of 1.2m.The ozone concentrations in the pollution chamberand the cabin environment varied between 96–615and 41–341 ppb, respectively. The levels were setaccording to the purpose of each experiment,the outdoor air supply rate and the absence orpresence of human subjects in the cabin. Whenpeople were present the ozone levels ranged from 60to 80 ppb.

2.4. Experimental procedures

Ozone removal by HEPA filter: A used HEPAfilter that had been in a passenger aircraft for 18months (its maximum recommended service life)was included in the air recirculation system (seeFig. 1) in some of the conditions. Two-thirds of itscross-sectional area was blocked off so that the areathrough which the recirculated air passed was incorrect proportion to the length of the cabin section.A total of five experiments, at air exchange rates ofeither 4.4 or 6.5 h�1, were conducted to evaluate theeffect of HEPA filters on ozone removal rates in thecabin. Two of the experiments focused on theinfluence of a loaded HEPA filter; in one experimenta new HEPA filter was installed instead; and twoexperiments were conducted with no HEPA filterinstalled. The length of the ozonation period rangedfrom 1 to 2 h. In four additional experiments,conducted at low and high air change rates witheither a used or new HEPA filter in the recirculationsystem, ozone removal efficiencies were directlydetermined by measuring ozone concentrationsupstream and downstream of the filter.

Sensory evaluation of the cabin air: Sensoryevaluation of the cabin air was carried out for twoconditions: used HEPA filter or no HEPA filterinstalled in the recirculation duct of the ventilationsystem. The outdoor air supply rate to the cabinduring the sensory evaluations was 4.4 h�1. Toavoid having subjects in the sensory panel enter thecabin, the air quality assessment took place, aftersteady-state conditions had been achieved, in an

adjacent chamber that was continuously suppliedwith air (�20L s�1) from the aircraft cabin. Thesystem was exposed to elevated ozone concentra-tions for at least 2 h before the sensory assessmentsbegan. A panel of untrained subjects consisting of26 staff members and students (age 22–60 years)from the ICIEE assessed the air quality shortly afterentering the chamber, recording their assessmentson a pseudo-continuous acceptability scale and onan odor intensity scale (Clausen, 2000). The panelwas blind to conditions.

Ozone removal by aircraft seats: Aircraft seats,due to their large surface area, present a significantsink for ozone. The contribution of the aircraft seatsto ozone removal was evaluated by measuring theozone removal rates in the presence and absence ofseats, keeping all other parameters the same. Suchcomparisons were made at low and high outdoor airsupply rates (4.4 and 8.8 h�1), always with a usedHEPA filter in place.

Ozone removal by soiled T-shirts: To simulate thepresence of humans in the cabin, 17 soiled T-shirtswere placed over the back of the aircraft seats. Malesubjects had slept in these T-shirts throughout theprevious night. Two sets of experiments wereconducted, one at an outdoor air supply rate of3 h�1 with a loaded HEPA filter in the system, and asecond at an outdoor air supply rate of 6.5 h�1 withno HEPA filter in the system. In both casesmeasurements were made with and without T-shirtsin the cabin.

Ozone removal by people: After obtaining ethicsapprovals from boards in Denmark and the UnitedStates, four experiments were conducted in thepresence or absence of 16 young female subjects atoutdoor air supply rates of either 4.4 or 8.8 h�1.During these experiments, a used HEPA filter wasin place in the recirculation system. On each of thefour experimental days there were two periods whenozone was generated — first, for a 1-h period beforeanybody entered the cabin; then, after a 1-h break,for a 4-h period with subjects present.

Surface aging: Changes in the ozone removal rate(i.e., surface aging) were evaluated during a 5-hexposure at an outdoor air supply rate of 8.8 h�1 inthe presence of a used HEPA filter, but with nosubjects in the cabin.

3. Data analysis

Ozone removal rate: The ozone removal ratewas determined from measurements of ozone’s

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–61336126

first-order decay within the cabin when the ozonegenerator was turned off. This method was aug-mented with mass-balance calculations (Weschler,2000) for each pair of cabin/pollution chamberozone concentrations when steady-state conditionshad been achieved (e.g. the aging effect experiment).When comparisons were possible, results obtainedwith these different approaches were similar. Mostof the results reported in this paper are based on thefirst-order decay measurements.

Sensory assessments: All data obtained fromquestionnaires were tested for normality using theShapiro–Wilks’ W test with a rejection region ofpo0.01. As all these data were normally distributed,one-way analysis of variance (ANOVA), supple-mented with Duncan’s post hoc comparison, orpaired t tests were used to evaluate differencesbetween the conditions. All reported p values are1-tailed.

Table 1

Summary of experiments

Experiment HEPA Seats T-shirts Peop

HEPA filter None + — —

Used + — —

New + — —

None + — —

Used + — —

HEPA filter,

upstream/

downstream ozone

measurements

Used + — —

New + — —

Used + — —

New + — —

HEPA filter,

sensory

None + — —

Used + — —

Aircraft seats Used + — —

Used — — —

Used + — —

Used — — —

Soiled T-shirts Used + — —

Used + 17 —

None + — —

None + 17 —

People Used + — —

Used + — 16

Used + — —

Used + — 16

Aging Used + — —

aAir exchange with outdoors (l).

4. Results and discussion

4.1. Retention ratio

The experiments are summarized in Table 1. Theozone retention ratio was calculated based onconcurrent ozone levels in the cabin and in thepollution chamber, which in this case is equivalentto ambient air for an actual aircraft. In the presenceof a used HEPA filter, at a low outdoor air changerate (4.4 h�1), the retention ratio was 0.15 withpeople present and 0.33 without people. For similarconditions, but at a high air change rate (8.8 h�1),the retention ratio was 0.21 with people present and0.52 without people. Over the small range of cabinhumidities in these experiments (10–22% RH), therewas no discernable influence of humidity on theozone retention ratio. With people present, themeasured ozone retention ratios in the simulated

le la (h�1) n Aim

4.4 1 Determine ozone removal

rates for used and new HEPA

filters at two outdoor air

change rates

4.4 3

4.4 1

6.5 2

6.5 1

4.4 2 Direct measurements of ozone

removal efficiencies of used

and new HEPA filters at two

outdoor air change rates

4.4 2

8.8 2

8.8 2

4.4 1 Sensory evaluations of the

cabin air with and without a

used filter

4.4 1

4.4 3 Determine ozone removal

rates for the aircraft seats at

two outdoor air change rates

4.4 2

8.8 2

8.8 2

3.0 1 Simulate contribution of

human skin oils to ozone

removal rates at two outdoor

air change rates

3.0 1

6.5 1

6.5 1

4.4 2 Determine ozone removal

rates due to people at two

outdoor air change rates

4.4 2

8.8 2

8.8 2

8.8 1 Examine ‘‘surface aging’’

effect on ozone removal

during an extended exposure

period

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–6133 6127

cabin were much lower than the average retentionratios measured by Nastrom et al. (1980) oncommercial aircraft (0.47 on a Boeing 747-100 and0.83 on a Boeing 747SP). Although passengers werepresent on these flights, there is no information ontheir numbers, and the authors state that theinfluence of load factors on the retention ratio wasnot addressed in their study. It should be noted thatthe outdoor air supply rate may have been higher onthese 1977 flights and that ozone in the cabin air wasmeasured close to the point where it entered thecabin; both differences would result in higherretention ratios if all else were equal.

4.2. The effect of a used or new HEPA filter

Table 2 summarizes the ozone removal rates inthe cabin with no HEPA filter, a new HEPA filter,or a used HEPA filter installed in the recirculationsystem. Two different airflow rates were used. Thenumber of independent measurements for a givencondition is also shown. When more than onemeasurement was carried out under the sameconditions, the ozone removal rate is given as themean. Due to a tight schedule for the use ofthe simulated aircraft cabin, we were limited in thenumber of filter experiments that we could conduct.

Compared to the condition without a filter, thepresence of the used HEPA filter in the systemincreased the ozone removal rate by 0.7 h�1 at thelower outdoor air supply rate (4.4 h�1), and by0.9 h�1 at the higher outdoor air supply rate(6.5 h�1). A new HEPA filter increased the ozone

Table 2

The effect of a used or new HEPA filter on the ozone removal rate

Condition Airflow (h�1) Ozone rem

Totala

Used HEPA Outdoor 4.4; recirc 18.8 9.170.6

No HEPA Outdoor 4.4; recirc 18.8 8.4

Difference

New HEPA Outdoor 4.4; recirc 18.8 9.8

No HEPA Outdoor 4.4; recirc 18.8 8.4

Difference

Used HEPA Outdoor 6.5; recirc 16.7 12.2

No HEPA Outdoor 6.5; recirc 16.7 11.370.0

Difference

For the ‘‘total’’ values, in cases with multiple measurements, the meanaIncludes removal due to air exchange with outdoors (l).

removal rate by 1.4 h�1 at the lower outdoor airsupply rate. This isolated finding is contrary to theexpectation that a filter will remove more ozonewhen loaded, an expectation that is confirmed byexperiments described in the following paragraph.

The results presented in Table 2 were less thanideal for determining the fractional contribution ofthe loaded HEPA filter or a new HEPA filter tooverall ozone removal. As reported in Table 3, wewere able to make direct measurements of the ozoneremoval efficiencies by measuring ozone concentra-tions immediately upstream and downstream ofboth the loaded and new HEPA filters at low andhigh outdoor air supply rates. The removal efficien-cies in Table 3 translate to ozone removal rates of0.5 and 1.1 h�1 for the loaded HEPA filter and 0.4and 0.9 h�1 for the new HEPA filter. These directmeasurements of ozone removal efficiencies for theloaded HEPA filter and the new HEPA filter areconsistent with expectation.

The effects of used and new HEPA filters on theoverall ozone removal rate were similar in magni-tude. In the case of the new HEPA filter, ozone maybe reacting with tackifiers and binders that areassociated with the filtration media (Bek +o et al.,2006). As the filter ages and particles cover thesurface of the media, ozone reacts with organicconstituents of the captured particles. The measuredozone removal efficiency (3.3–5.7%) of the usedHEPA filter is slightly less than values reported byHyttinen et al. (2003) for nine used supply air filters(8–26%) and by Bek +o et al. (2006) for samples froma loaded EU7 filter (12%). However, both of these

oval rate (h�1) Number of measurements

Total—l

4.7 3

4.0 1

0.7

5.4 1

4.0 1

1.4

5.7 1

4.8 2

0.9

s and standard deviations are shown.

ARTICLE IN PRESS

Table 3

Ozone removal efficiencies, based on upstream and downstream ozone measurements, for used and new HEPA filters in the recirculation

duct

Airflow (h�1) Ozone removal efficiency (%) Ozone removal rate (h�1)

Used HEPA New HEPA Used HEPA New HEPA

Outdoor 4.4; recirc 18.8 (3.1m s�1 face velocity) 5.770.8 4.771.0 1.1 0.9

Outdoor 8.8; recirc 14.4 (2.3m s�1 face velocity) 3.371.8 2.571.7 0.5 0.4

Ozone removal rates calculated from the measured ozone removal efficiencies.

Table 4

The effect of HEPA filter on IAQ

Condition Airflow (h�1) Cabin ozone (ppb) Acceptabilitya Odor intensityb

Average7SD Mean 95% c.i. Mean 95% c.i.

No HEPA Outdoor 4.4; recirc 18.8 5272.6 0.06 �0.11y0.23 14.9 11.1y18.7

Used HEPA Outdoor 4.4; recirc 18.8 5571.4 �0.18 �0.37y0.01 20.6 15.6y25.6

aScale: �1 (clearly unacceptable) to +1 (clearly acceptable).bScale: 0 (no odor) to 50 (overpowering odor).

G. Tamas et al. / Atmospheric Environment 40 (2006) 6122–61336128

studies were conducted using face velocities in therange of 0.1–0.2m s�1, almost an order of magni-tude smaller than the face velocities in the currentinvestigation (Table 3).

The subjective results shown in Table 4 clearlyindicate the negative effect of the used HEPA filteron sensory evaluations. The perceived air quality(PAQ) was significantly poorer with the used HEPAfilter in the system (�0.18) compared to no HEPAfilter in the system (+0.06; po0.03). Furthermore,the odor intensity was stronger compared to thecondition without a HEPA filter (po0.006). This isconsistent with results from numerous other studiesindicating that a loaded ventilation filter adverselyimpacts the quality of the air that passes through it(Clausen, 2004 and references therein). This effectmay be amplified by oxidation processes occurringon the surface of the loaded HEPA filter (Bek +oet al., 2006).

4.3. The effect of aircraft seats and other surfaces

The effect of aircraft seats on the measured ozoneremoval rate can be seen from the results presentedin Table 5. In the presence of the used HEPA filterat the lower air change rate, the ozone removal ratewas 2.7 h�1 larger with seats present than with seatsabsent. The same comparison at the higher airchange rate indicates a contribution of 3.1 h�1 fromthe aircraft seats. The difference between the

removal rates measured at low and high outdoorair change rates is not large and is within the errorof measurement. With the seats removed, theoverall ozone removal rate was between 1.7 and2.0 h�1. Given that the loaded HEPA filter con-tributed approximately 0.7–0.9 h�1, the other sur-faces in the cabin and ventilation systemcontributed between 0.8 and 1.3 h�1 to the ozoneremoval rate. We assume, based on previous studiesexamining interactions between ozone and carpet(Weschler et al., 1992; Morrison and Nazaroff,2000; Morrison and Nazaroff, 2002b), that thecarpet within the cabin was primarily responsiblefor scavenging by ‘‘other surfaces’’. Although thetotal surface area of the aircraft seats was onlyabout 50% higher than that of the carpet, the seatsremoved roughly twice as much ozone. We suggestthat this reflects soiling of the seats by human skinoils. It should be noted that the exposed surfacearea of the seats will be on average substantially less(20–25% less by our estimates) when the airplane isoccupied.

4.4. The effect of soiled T-shirts

The ozone removal rate increased in the presenceof soiled T-shirts compared to the condition withoutT-shirts (Table 6). These data indicate that the 17 T-shirts contribute 3.6–5.1 h�1 to the ozone removalrate. Our previous study (Wisthaler et al., 2005) of

ARTICLE IN PRESS

Table 5

The effect of aircraft seats on ozone removal rates

Condition Airflow (h�1) Ozone removal rate (h�1) Deposition velocity (cm s�1) Number of measurements

Totala Total—l

Seats Outdoor 4.4; recirc 18.8 9.170.6 4.7 0.09 3

No seats Outdoor 4.4; recirc 18.8 6.470.2 2.0 2

Difference 2.7

Seats Outdoor 8.8; recirc 14.4 13.671.8 4.8 0.11 2

No seats Outdoor 8.8; recirc 14.4 10.570.2 1.7 2

Difference 3.1

For the ‘‘Total’’ values, means and standard deviations are shown. Deposition velocities calculated as outlined in the text.aIncludes removal due to air exchange with outdoors (l).

Table 6

The effect of soiled T-shirts on ozone removal rates

Condition Airflow (h�1) Ozone removal rate (h�1) Deposition velocity (cm s�1) Number of measurements

Totala Total—l

T-shirtsb Outdoor 3; recirc 20.2 14.4 11.4 0.27 1

No T-shirtsb Outdoor 3; recirc 20.2 9.3 6.3 1

Difference 5.1

T-shirtsc Outdoor 6.5; recirc 16.7 14.9 8.4 0.19 1

No T-shirtsc Outdoor 6.5; recirc 16.7 11.3 4.8 1

Difference 3.6

Deposition velocities calculated as outlined in the text.aIncludes removal due to air exchange with outdoors (l).bUsed HEPA in system.cNo HEPA in system.

G. Tamas et al. / Atmospheric Environment 40 (2006) 6122–6133 6129

the effects of ozone-initiated chemistry on thechemical composition of cabin air examined freshT-shirts as well as soiled T-shirts. We found thatfresh T-shirts, in the presence of ozone, had littleeffect on the chemicals present in the cabin,indicating that it was primarily skin oils transferredto the soiled T-shirts that were responsible for theozone removal.

These T-shirt experiments should not be mis-construed. They tell us little regarding the contribu-tion of people’s clothing to ozone removal. Theextent to which clothing is soiled with skin oil onactual flights is unknown and may be greater or lessthan in these experiments. Furthermore, somefabrics are expected to scavenge ozone at fasterrates than other fabrics. Further experiments arenecessary to elucidate the role of clothing in ozoneremoval.

4.5. The effect of people

Four experiments, two at low and two at highoutdoor air supply rates, were conducted withhuman subjects in the simulated aircraft cabin.The results presented in Table 7 show that thepresence of people increased the ozone removal ratein the cabin by 6.2 and 7.4 h�1 at low and highoutdoor air supply rates, respectively. Again, thedifference between these removal rates is within theerror of measurement.

Ozone reacts on both exposed body surfaces andclothing; it is also removed by respiration. Thefractional ozone removal attributable to exposedskin/hair versus clothing could not be determinedfrom the present experiments. However, it ispossible to estimate removal attributable to respira-tion. The breathing rate for sedentary adult females

ARTICLE IN PRESS

Table 7

The effect of people on ozone removal rates

Condition Airflow (h�1) Ozone removal rate (h�1) Deposition velocity (cm s�1) Number of measurements

Totala Total—l

People Outdoor 4.4; recirc 18.8 15.370.1 10.9 0.20 2

No people Outdoor 4.4; recirc 18.8 9.170.6 4.7 2

Difference 6.2

People Outdoor 8.8; recirc 14.4 21.071.3 12.2 0.23 2

No people Outdoor 8.8; recirc 14.4 13.671.8 4.8 2

Difference 7.4

For the ‘‘Total’’ values, means and standard deviations are shown. Deposition velocities calculated as outlined in the text.aIncludes removal due to air exchange with outdoors (l).

G. Tamas et al. / Atmospheric Environment 40 (2006) 6122–61336130

averages 0.48m3 h�1 (US EPA, 1997). It is reason-able to assume that all of the ozone inhaled isremoved (Thorp, 1950). Hence, 16 human subjectswould remove 7.7m3 h�1. Given that the totalvolume of the simulated cabin plus recirculationsystem is 31m3, breathing by the 16 human subjectswould be equivalent to an ozone removal rate of0.25 h�1. Hence, respiration can only make a smallcontribution (�4%) to the additional removal ofozone that was observed when female subjects werepresent in the simulated cabin.

4.6. Deposition velocities

To put ozone removal by passengers and crew inperspective, it is useful to convert the values inTable 7 to ozone deposition velocities (Nazaroffet al., 1993), which can be compared with otherdeposition velocities reported in the literature.Deposition velocities were calculated by multiplyingthe rate constants for ozone removal (h�1) by thevolume of the system (31m3) and dividing by thetotal surface area of the passengers (16� 1.7m2).The latter value for the female subjects wasestimated using the DuBois formula (ISO, 1990)and takes no account of the effective increase insurface area that is contributed by the porosity androughness of clothing textiles. The calculateddeposition velocities associated with subjects in thesimulated cabin were 0.20 and 0.23 cm s�1. Based onthe total surface area of the cabin seats (24.6m2)and the estimated total surface area of the T-shirts(16.2m2), ozone deposition velocities were alsocalculated for the seats and T-shirts. For the aircraftseats, the deposition velocities are 0.09 and0.11 cm s�1 and for the T-shirts 0.27 and

0.19 cm s�1. The fact that the deposition velocitiesfor soiled T-shirts were comparable to those forpeople themselves illustrates the relatively largecontribution that readily transferred human skinoils make to the ozone removal process. Indeed, thecalculated deposition velocities for the T-shirts andpeople are close to the anticipated upper limit forsuch values based on mass transport considerations(see Fig. 1 of Morrison and Nazaroff, 2002b). Thecalculated ‘‘deposition velocities’’ for T-shirts andpeople may include contributions from gas phasereactions with off-gassing unsaturated oxidationproducts (e.g., 6-methyl-5-heptene-2-one) derivedfrom ozone/skin oil reactions. Further studies willbe required to clarify this issue.

4.7. The effect of surface aging

The influence of prolonged ozone exposure on thesurface removal rate (aging) was examined in anexperiment with constant ozone generation over a5-h period (Fig. 2). Given the outdoor air exchangerate of 8.8 h�1, a pseudo steady-state conditionshould have been attained within the first hour, andduring the final 4 h of the experiment, the ozoneconcentration should have remained constant if theozone removal rate within the cabin was constant.However, it is apparent in Fig. 2 that theconcentration of ozone in both the pollutionchamber and the aircraft cabin continued to slowlyincrease during this period. Using a mass balancemodel and the ozone concentrations in the pollutionchamber and cabin, the ozone removal rates werecalculated for the final 3.5 h of this experiment(open triangles in Fig. 2). Ozone was removed withan average rate constant of 3.8 h�1 during the first

ARTICLE IN PRESS

0

20

40

60

80

100

120

140

160

180

Ozo

ne c

once

ntra

tion

[ppb

]

0

1

2

3

4

5

6

7

8

9

Ozo

ne r

emov

al r

ate

[h-1

]

14:30 15:30 16:30

Time

17:30 18:30 19:30 20:30

Old HEPA

3.72 ± 0.50 h-1

3.62 ± 0.45 h-13.81 ± 0.53 h-1

8.79 h-1

Poll chamber Cabin Removal

Fig. 2. The calculated ozone removal rates (triangles and right y-axis) derived from the measured ozone concentrations in the pollution

chamber (diamonds and left y-axis) and aircraft cabin (squares and left y-axis). See text for details.

58%

25%

10%

7%

0%

20%

40%

60%

80%

100%

HEPA filter

Other surfaces

Seats

People

Fig. 3. Relative contribution (%) of various sinks to ozone

removal in the simulated aircraft cabin. Median values have been

used in preparing figure; the range of values were 0.7–0.9 h�1 for

HEPA filter, 2.7–3.1 h�1 for seats, 0.8–1.3 h�1 for other surfaces,

and 6.2–7.4 h�1 for people.

G. Tamas et al. / Atmospheric Environment 40 (2006) 6122–6133 6131

half of this period compared with an average of3.6 h�1 during the second half (these values aresignificantly different; po0.006). The trendlineshown on the figure indicates that the ozoneremoval rate decreased by roughly 3%h�1 duringthis experiment.

Similar observations have been reported in theliterature for both carpets (Morrison and Nazaroff,2002a) and building filters (Bek +o et al., 2006).Presumably, pollutants present on the varioussurfaces are gradually consumed by reaction withozone. ‘‘Aged’’ surfaces can regain some of theirozone scavenging potential if they are not exposedto ozone for several hours or days. During suchrecovery periods, we suggest that reactive materialswithin the bulk of various aircraft cabin materialsdiffuse to the surface, becoming available for futurereaction with ozone. Note that passengers, themajor sink for ozone in the aircraft cabin, arecontinually ‘‘renewed’’ in successive flights.

4.8. Comparisons of ozone sinks

Fig. 3 presents the relative contribution of HEPAfilters, other surfaces, seats and people to ozoneremoval in the simulated aircraft cabin. Of thesesinks, people make the largest relative contribution,almost 60%. The next largest contribution comesfrom the aircraft seats, approximately 25%. Theseseats had been in actual service for several years. Itseems probable that they remove ozone largelybecause they are soiled with the skin oils of thepassengers who occupied them throughout theirservice life. In our previous study (Wisthaler et al.,

2005) we saw chemical evidence that these residualoils are likely contributing to ozone removal. Thecontribution from the HEPA filter is about 7%. Theother surfaces in the cabin, excluding the seats andHEPA filter, contribute 10%, a relatively smallfraction of the total.

5. Conclusions

Taken together, these measurements provide amore complete picture of the sinks that remove

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–61336132

ozone as it is transported from outside the aircraftto the breathing zone of passengers and crew.However, the results should not be over-interpreted.They have been obtained for one type of simulatedaircraft cabin environment with carpet, seats,HEPA filter, wall panels, and ductwork of a certaincomposition. Additional simulations involving al-ternative materials and other aircraft types, as wellas measurements on actual flights, will be necessaryto more fully understand the factors influencingozone removal in aircraft cabins. A general conclu-sion that does come out of the present study is thatpassengers and crew remove large amounts of ozoneand are dominant factors in determining ozonelevels in the air that surrounds them.

Ozone removal is desirable in terms of reducingthe exposure of passengers to ozone, but ozoneremoval by surfaces is not without consequencesbecause, to a large extent, this removal is due toozone reactions with organic compounds on thesurfaces. The more volatile oxidation products cansubsequently desorb from the surfaces and becomepart of the mix of chemicals within the cabin towhich passengers and crew are exposed. Detailedchemical analyses conducted in this same simulatedcabin (Wisthaler et al., 2005) indicate that theoxidation products include saturated and unsatu-rated aldehydes and squalene oxidation products. Inthe cited study, many of the resulting aldehydeswere present at concentrations above their odorthresholds. Formaldehyde and acrolein were presentat concentrations that exceeded California’s refer-ence exposure levels designed to protect individualsfrom chronic health effects (OEHHA, 2005).

Retention ratios measured in this study are lowerthan the default value of 0.7 used by the FAA indeciding whether or not ozone-removing devicesshould be used on aircraft. However, this should notbe construed as an extra ‘‘margin of protection’’.For the reasons outlined above, a low retentionratio indicates significant surface chemistry and,potentially, significant exposures to the consequentoxidation products that desorb from surfaces. Theozone removal associated with the presence ofpassengers produces products to which otherpassengers are exposed. From the T-shirt experi-ments, we know that these products include thesqualene oxidation products acetone, 6-methyl-5-heptene-2one and 4-oxopentanal (Wisthaler et al.,2005). The presence of lower ozone concentrationsin aircraft cabin air does not, by itself, indicate aclean or healthful environment. Rather, reduction

of ozone concentrations may relate directly toheterogeneous reactions that form other air pollu-tants of concern. This fact underscores the need forefficient removal of ambient ozone from the airsupply system of an aircraft.

Acknowledgements

We thank William W. Nazaroff for suggesting theuse of deposition velocities and for observationsrelated to ozone transport from bulk air to thesurface of human subjects. We thank Julita Zar-zycka, Danuta Myskow, Pawe" Wargocki, KirilNaydenov, and Jørn Toftum of the TechnicalUniversity of Denmark, each of whom contributedto the completion of this project. Finally, we thankthe subjects who patiently participated in theseexperiments. This work has been supported by theDanish Technical Research Council (STVF) as partof the research program of the International Centrefor Indoor Environment and Energy at the Techni-cal University of Denmark and by the US FederalAviation Administration (FAA) Office of Aero-space Medicine through the Air TransportationCenter of Excellence for Airliner Cabin Environ-ment Research (ACER), Cooperative Agreement04-C-ACE-UMDNJ. Although the FAA has spon-sored this project, it neither endorses nor rejects thefindings of this research.

References

Bako-Biro, Zs., Weschler, C.J., Wargocki, P., Fanger, P.O., 2005.

Effects of indoor pollution sources and ventilation rate on

ozone’s surface removal rate and the occurrence of oxyge-

nated VOCs in an office space. In: Proceedings of Indoor Air

2005, Beijing, The 10th International Conference on Indoor

Air Quality and Climate, vol. II(2), pp. 2320–2324.

Bek +o, G., Tamas, Gy., Halas, O., Clausen, G., Weschler, C.J.,

2005. Ultra-fine particles as indicators of the generation of

oxidized products on the surface of used air filters. In:

Proceedings of Indoor Air 2005, Beijing, The 10th Interna-

tional Conference on Indoor Air Quality and Climate, vol.

II(1), pp. 1521–1525.

Bek +o, G., Halas, O., Clausen, G., Weschler, C.J., 2006. Initial

studies of oxidative processes on filter surfaces and their

impact on perceived air quality. Indoor Air 16 (1), 56–64.

Bell, M.L., McDermott, A., Zeger, S.L., Samet, J.M., Dominici,

F., 2004. Ozone and short-term mortality in 95 US urban

communities, 1987–2000. Journal of the American Medical

Association 292, 2372–2378.

Clausen, G., 2000. Sensory evaluation of emissions and indoor

air quality. In: Proceedings of Healthy Buildings 2000, Espoo,

vol. 1, pp. 53–62.

ARTICLE IN PRESSG. Tamas et al. / Atmospheric Environment 40 (2006) 6122–6133 6133

Clausen, G., 2004. Ventilation filters and indoor air quality: a

review of research from the International Centre for Indoor

Environment and Energy. Indoor Air 14 (s7), 202–207.

Grøntoft, T., Raychaudhuri, M.R., 2004. Compilation of tables

of surface deposition velocities for O3, NO2 and SO2 to a

range of indoor surfaces. Atmospheric Environment 38 (4),

533–544.

Hyttinen, M., Pasanen, P., Salo, J., Bjorkroth, M., Vartiainen,

M., Kalliokoski, P., 2003. Reactions of ozone on ventilation

filters. Indoor and Built Environment 12 (3), 151–158.

Hyttinen, M., Pasanen, P., Kalliokoski, P., 2006. Removal of

ozone on clean, dusty and sooty supply air filters. Atmo-

spheric Environment 40 (2), 315–325.

ISO, 1990. Ergonomics—Determination of Metabolic Heat

Production, ISO 8996.

Klenø, J.G., Wolkoff, P., 2004. Changes in eye blink frequency as

a measure of trigeminal stimulation by exposure to limonene

oxidation products, isoprene oxidation products and nitrate

radicals. International Archives of Occupational and Envir-

onmental Health 77 (4), 235–243.

Klenø, J.G., Clausen, P.A., Weschler, C.J., Wolkoff, P., 2001.

Determination of ozone removal rates by selected building

products using the FLEC emission cell. Environmental

Science and Technology 35 (12), 2548–2553.

Morrison, G.C., Nazaroff, W.W., 2000. The rate of ozone uptake

on carpets: experimental studies. Environmental Science and

Technology 34 (23), 4963–4968.

Morrison, G.C., Nazaroff, W.W., 2002a. Ozone interactions with

carpet: secondary emissions of aldehydes. Environmental

Science and Technology 36 (10), 2185–2192.

Morrison, G.C., Nazaroff, W.W., 2002b. The rate of ozone

uptake on carpets: mathematical modelling. Atmospheric

Environment 36 (11), 1749–1756.

Nastrom, G.D., Holderman, J.D., Perkins, P.J., 1980. Measure-

ments of cabin and ambient ozone on B747 airplanes. Journal

of Aircraft 17, 246–249.

Nazaroff, W.W., Weschler, C.J., 2004. Cleaning products and air

fresheners: exposure to primary and secondary air pollutants.

Atmospheric Environment 38 (18), 2841–2865.

Nazaroff, W.W., Gadgil, A.J., Weschler, C.J., 1993. Critique of

the use of deposition velocity in modeling indoor air quality.

In: Nagda, N.L. (Ed.), Modeling of Indoor Air Quality and

Exposure, ASTM STP 1205. American Society for Testing

and Materials, Philadelphia, PA, pp. 81–104.

National Research Council (NRC) Committee on air quality in

passenger cabins of commercial aircraft, 2002. The Airliner

Cabin Environment and the Health of Passengers and Crew.

National Academy Press, Washington, DC.

Nøjgaard, J.K., Christensen, K.B., Wolkoff, P., 2005. The effect

on human eye blink frequency of exposure to limonene

oxidation products and methacrolein. Toxicology Letters 156

(2), 241–251.

OEHHA, California’s Office of Environmental Health Hazard

Assessment, Air, http://www.oehha.ca.gov/air.html (accessed

on November 30, 2005).

SAE International (Society of Automotive Engineers Interna-

tional), 2000. Ozone in High Altitude Aircraft. Aerospace

Information Report, SAE AIR910 Rev. B. Society of

Automotive Engineers.

Spengler, J., Ludwig, S., Weker, R., 2004. Ozone exposures

during trans-continental and trans-pacific flights. Indoor Air

14 (s7), 67–73.

Spicer, C.W., Murphy, M.J., Holdren, M.W., Myers, J.D.,

MacGregor, I.C., Holloman, C., James, R.R., Tucker, K.,

Zaborski, R., 2004. Relate air quality and other factors to

comfort and health symptoms reported by passengers and

crew on commercial transport aircraft (part I). ASHRAE

Report 1262-TRP. American Society for Heating, Refrigerat-

ing and Air Conditioning Engineers, Atlanta, GA.

Tamas, Gy., Weschler, C.J., Toftum, J., Fanger, P.O., 2006.

Influence of ozone–limonene reactions on perceived air

quality. Indoor Air 16 (3), 168–178.

Thorp, C., 1950. Toxicity of O3. Industrial Medicine and Surgery

19 (2), 45–57.

US Environmental Protection Agency (US EPA), 1997. Exposure

Factors Handbook. General Factors, vol. 1. National Center

for Environmental Assessment, Washington, DC EPA/600/P-

95/001Fa.

US Environmental Protection Agency (US EPA), 2006. Air

Quality Criteria for Ozone and Related Photochemical

Oxidants (Final). US Environmental Protection Agency,

Washington, DC EPA/600/R-05/004aF-cF.

Weschler, C.J., 2000. Ozone in indoor environments: concentra-

tion and chemistry. Indoor Air 10 (4), 269–288.

Weschler, C.J., Hodgson, A.T., Wooley, J.D., 1992. Indoor

chemistry—ozone, volatile organic compounds, and carpets.

Environmental Science and Technology 26 (12), 2371–2377.

Wisthaler, A., Tamas, Gy., Wyon, D., Strøm-Tejsen, P., Space,

D., Beauchamp, J., Hansel, A., Mark, T.D., Weschler, C.J.,

2005. Products of ozone initiated chemistry in a simulated

aircraft environment. Environmental Science and Technology

39 (13), 4823–4832.