Environ. Sci. Technol. 2010, 44, 6269–6274
Use of Reactive Tracers To Determine Ambient OH Radical Concentrations: Application within the Indoor Environment IAIN R. WHITE,† DAMIEN MARTIN,† ˜ OZ,† M. PAZ MUN FREDRIK K. PETERSSON,† STEPHEN J. HENSHAW,† GRAHAM NICKLESS,† GUY C. LLOYD-JONES,† KEVIN. C. CLEMITSHAW,‡ AND D U D L E Y E . S H A L L C R O S S * ,† School of Chemistry, Bristol University, Bristol BS8 1TS, U.K., Department of Earth Sciences, Royal Holloway, University of London, Egham TW20 0EX, U.K.
Received June 9, 2009. Revised manuscript received December 29, 2009. Accepted February 23, 2010.
The hydroxyl radical (OH) plays a key role in determining indoor air quality. However, its highly reactive nature and low concentration indoors impede direct analysis. This paper describes the techniques used to indirectly quantify indoor OH, including the development of a new method based on the instantaneous release of chemical tracers into the air. This method was used to detect ambient OH in two indoor seminar rooms following tracer detection by gas chromatographymass spectrometry (GCMS). The results from these tests add to the small number of experiments that have measured indoor OH which are discussed with regard to future directions within air quality research.
Introduction Due to the importance of OH in the photochemical processing of atmospheric pollutants, outdoor OH chemistry has received a great deal of attention (1). In contrast, there has been little research conducted to characterize the range of OH concentrations indoors. It was previously assumed that indoor OH originated solely from the infiltration of outdoor air and that levels were therefore negligible given its extremely short lifetime. Indeed once indoors OH rapidly degrades through reaction with the many gaseous species that accumulate inside buildings. However, there is evidence to suggest that a significant level of OH is also produced there. Nazaroff and Cass (2) first suggested OH could be generated indoors based on calculations from a chemical model; a hypothesis later supported when it was shown that reactions between ozone and reactive alkenes cause indirect production of OH (3). As many alkenes react with ozone and produce OH fast enough to significantly alter indoor chemistry (4), detailed modeling studies were eventually carried out to investigate the oxidative capacity of the indoor environment (e.g. refs 5 and 6). OH concentrations from selected theoretical and experimental studies are provided in Table 1. * Corresponding author e-mail:
[email protected]. † Bristol University. ‡ University of London. 10.1021/es901699a
2010 American Chemical Society
Published on Web 07/19/2010
Ventilation or air exchange is the primary removal mechanism for many important indoor pollutants and so plays a major role in influencing the chemistry of indoor air. It is commonly expressed in terms of the air exchange rate, λ, usually in units of h-1. λ is defined as the exfiltration rate, Ev, of air from a room or building of volume, V, i.e. λ)
Ev V
(1)
The ventilation rate can be directly measured using a tracer gas technique to determine the first order exponential decay rate. Poor indoor air quality has been linked to inadequate building ventilation due to insufficient dilution of pollutants such as volatile organic compounds (VOCs) which originate indoors, e.g. from building materials and furnishings (7), cleaning products and air fresheners (8), and office equipment (9). Gas phase chemistry is also promoted, as the residence time of airborne pollutants is increased. This is thought to be a possible cause of Sick Building Syndrome (SBS) - a general term describing a variety of symptoms suffered by dwellers indoors which has ultimately lowered productivity in society (10). However, a correlation in VOC levels and SBS symptoms has not yet been observed (11), but while many compounds are measured below their threshold odor or irritation limits, potentially it may be their oxidation products that lead to discomfort indoors (12). Therefore, to fully understand workplace air chemistry, knowledge of the oxidative capacity of the indoor environment is vital. To date, the only published measurements of indoor OH have been obtained using tracer techniques. In this study, these techniques are described and include a new method developed specifically to detect ambient indoor OH which, to the best of the authors’ knowledge, has been used to make the first measurements of OH in a totally unperturbed office environment. Tracer Methods of OH Detection. Constant Emission Method. The first measurements of indoor OH were made by Weschler and Shields (13), included in Table 1 as the “constant emission method”, a novel indirect method of OH concentration determination in many ways similar to the method applied here. In this method, the compound 1,3,5trimethylbenzene (TMB), which is known to react almost exclusively with OH, is used as an OH indicator. This and an inert reference compound, perchloroethylene (PCE), are simultaneously released at a known rate. PCE reacts with OH at a negligible rate and is therefore used to take into account variation in the rate of ventilation. TMB and PCE are first released under negligible OH conditions yielding the initial ratio between the tracers, R0. Weschler and Shields (13), in their initial experiments, found this ratio was constant over time and equal to the ratio of emission rates, rS, of TMB and PCE: RO )
rSTMB
(2)
rSPCE
The concept of the constant emission method is that at sufficient OH levels, the ratio R0 changes (the magnitude of which provides a measure of the OH concentration). This new ratio, ROH, is dependent on the bimolecular rate constant, kOH, used to calculate OH thus VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6269
[OH] )
[
λ R0 -1 kOH ROH
]
(3)
Tracer Decay Method. The tracer technique used in the present application is yet another method of OH detection, the concept of which was first proposed by Prinn (14). This method, used in this case to detect indoor OH, has been developed such that with minor modification it could also be used in the outdoor environment. Henceforth termed the tracer decay method, it is based upon the deliberate, simultaneous release of an inert tracer (or the dispersant), D, and a reactive tracer, (or titrant), T. If both tracers are chosen such that the only loss mechanism is titrant oxidation from OH and assuming the two tracer species are well mixed upon release then the ratio, R ()[T]/[D]) is spatially invariant and OH can be calculated through [OH] ) -
1 ∂ln R kOH ∂t
(4)
And so plotting ln(R)/kOH against time yields the OH concentration as the slope. In using the tracer decay method, only precise measurement of the ratio change between the tracers in time is needed for accurate OH determination. As experimental conditions vary less in the indoor environment, they can be determined more easily and may be manipulated through the ventilation rate and generation of reactive species, and this environment provides an interesting case study to test the reactive tracer technique. In the indoor environment, OH is likely to naturally vary over time and at different positions within a room. This method therefore gives rise to an OH concentration that is both temporally averaged (over t) and spatially averaged. In order that the latter represents the OH concentration across a room, titrant sampling must commence when the tracers are well mixed within the room at a concentration low enough so as not to perturb local chemistry. This method was used to ascertain the oxidative capacity inside two very different unoccupied University of Bristol (UoB, Bristol, U.K.) seminar rooms over two separate experiments. In both experiments, a small amount of tracer was released and then periodically monitored throughout the day. Ozone, temperature, and selected light hydrocarbons were also monitored over the tracer measurement periods.
Experimental Section Tracer Selection. While the basic notion of using chemical tracers to detect atmospheric OH is conceptually simple, selecting a suitable titrant for such a task represents a significant challenge (15). Many compounds exist that exhibit favorable properties in terms of their use as inert tracers: they are nontoxic and chemically stable and do not readily deposit on surfaces. However, in selecting a reactive tracer, one or more of these properties generally become compromised as reactivity toward OH increases. For the present
application, the chosen compound must be highly reactive to detect low levels of OH yet must remain largely insensitive toward reaction with other oxidants. Vapor pressure must be high, photolysis and deposition processes must not compete with oxidation, and the tracer must be nontoxic. Background levels of the tracer must be low with few sources and as the tracer is by its very nature highly reactive, low levels must be detectable to avoid chemical perturbation during an experiment. Following an extensive study, two compounds were chosen as tracers for method development: perfluoromethylcyclopentane (C6F12, PMCP) as the inert tracer and d5isoprene (C5D5H3) as the reactive tracer. The tracer decay method introduces several new challenges: the analytical system must be stable over several orders of magnitude and both sensitive and selective. The inert tracer, PMCP, can be analyzed with an extremely high degree of sensitivity using negative ion chemical ionizationmass spectrometry (NICI-MS) especially if preconcentrated on carbon based adsorbents (16). The hydrocarbons, however, do not show the same response to chemical ionization. Instead, a preconcentration device was developed specifically to couple with a mass spectrometer in electron impact mode. d-Isoprene Synthesis. The synthesis of d5-isoprene involved three main steps: reaction of d6-acetone with lithium acetylide ethylenediamine forming d6-3-methylbutyn-1-ol, reduction to d6-3-methylbuten-1-ol by hydrogenation using a Lindlar catalyst, and formation of the deuterated isoprene by treatment with HBr.
Direct distillation from the reaction mixture isolated d5isoprene at 93.6% purity. Site Description. The first experiment was conducted in a room situated on the fourth floor of the Synthetic Chemistry building within the School of Chemistry at the UoB on ninth August 2006. This experiment is referred to as the SoC experiment. The room is a mechanically ventilated seminar room with one door and several well-sealed operable windows, primarily used for small group meetings. The volume of the room is 48 m3. The second experiment was conducted in a room on the second floor of the Graduate School of Education building in the UoB on seventh April 2008. This experiment is referred to as the GSE experiment. This room is a naturally ventilated seminar room much larger
TABLE 1. Previous Indoor OH Studies citation
[OH]/cm-3
[O3]/ppb(v)
τOH/ha
notes
1 4 13 5 40 29 6
5.4 × 10 1.7 × 105 7.0 × 105 1.2 × 105 4.1 × 105 1.0-4.6 × 105 1.2-4.0 × 105
15 20 62-192 100 40 60-120 3-13
50.9 16.2 3.9 22.9 6.7 9.8 10.6
mathematical model of indoor air chemistry steady state model constant emission method used SAPRC-99 model to simulate indoor chemistry steady state model constant emission method indoor air box model based on MCM chemistry
a
6270
4
The lifetime of isoprene due to reaction with OH.
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
than the SoC room with a volume of 316 m3. Two well-sealed doors connect the room to a corridor; the room has several windows. Preliminary Ventilation Studies. Ventilation studies were conducted to ascertain the variability in λ prior to performing the SoC experiment. Short bursts of pure CO2 were released into the room, and the decay back to ambient levels was monitored using infrared detectors (17). Five ventilation studies were performed over three days, with an average rate of 0.5 h-1. λ varied a great deal over the three days by (0.4 h-1. However, λ, controlled by an external system, appeared to change smoothly over time Through examining the concentration profiles following CO2 releases, these studies also suggest that the mixing time of the tracer (i.e., the time taken for the release mixture to fully disperse to a homogeneously distributed concentration) is around 15 min. This time was approximately determined as the point where all of the detectors, situated at different positions within the room, measured the same CO2 concentration, decaying at the same rate. This is supported by several ozone tests that were conducted in the room, whereby ozone was monitored via UV photometry while windows in the room were opened: maximum ozone levels were recorded around 15 min after opening. Ozone was monitored over several days and was found to remain fairly constant at around 3 ppb(v). The indoor/outdoor ratio of ozone over this period was between 0.2 and 0.1. The ventilation rate of the GSE room was determined prior to the experiment by observing the decay of the inert tracer; λ was found to be on average 0.3 h-1. Through monitoring the concentration of other fluorinated compounds, present as impurities within the release mixture, the assumption that different species diffuse at the same rate could be tested as the ratio between these compounds remained the same. Extremely variable results were obtained based upon preliminary ozone measurements and ventilation studies. General Experimental Procedure. Both experiments were conducted during hot weather to increase the likelihood of moderately high outdoor ozone levels. Ozone was monitored inside the rooms over the duration of the experiments. Temperature and humidity were also monitored. In the second tracer experiment, CO was also monitored. The experiments commenced upon release of a gaseous mixture containing the tracers. This mixture was prepared by gravimetric dilution of d5-isoprene (prepared as described previously) and PMCP (obtained as a pure liquid, F2 chemicals Ltd., Lancashire, U.K.) into air (Linde Gases Ltd., U.K.) creating gaseous concentrations of 20.6 ppm(v) and 20.1 ppm(v), respectively, with a certification accuracy of (5%. In both cases this tracer mixture was released directly from a filled Tedlar bag (SKC Inc., PA, U.S.A.). Air samples, once transported back to the laboratory, were analyzed for the inert and reactive tracers (an example is shown in Figure 1) as well as C2 to C9 hydrocarbons. Release and Sampling. During the SoC experiment, approximately 2 L of the tracer mixture was introduced into the room at 09:00. Samples were taken at approximately hourly periods throughout the day through a 2 m Nathgal sampling line. Samples were taken using a variable flow rate diaphragm pump (SKC Inc., PA, U.S.A.) into 10 L Tedlar bags (SKC Inc., PA, U.S.A.). The sample bags were immediately stored in black polythene once filled. The successful use of Tedlar bags in the storage of isoprene has been documented by several authors (18–20). However, VOC degradation inside Tedlar has been documented (21), and dramatic losses of isoprene have been observed following the sampling of human breath (22, 23). Our own observations, based upon tests using outdoor air, suggest the reactive tracer is stable inside Tedlar for several days before showing any significant degradation and is stable for longer inside indoor air samples.
FIGURE 1. Tracer decay during the SoC experiment. Furthermore, during tests where high concentrations of ozone (approximately 500 ppb(v)) were artificially generated and sampled into Tedlar bags, a small fraction of the initial ozone concentration was measured immediately after sampling (lowest concentrations were sampled when using a flow restrictor required for low flow applications (a stainless steel needle valve) at around 1% of the initial concentration). The restrictor and sampling equipment remain unretentive toward the reactive tracer. Although these tests imply minimal losses during storage, the samples were analyzed promptly in case ambient ozone concentrations were not depleted by the sampling and storage equipment. During the GSE experiment, 5 L of the tracer mixture was released into the room at 11:45. Samples were taken at roughly 15 min intervals using the same sampling method as the SoC experiment. After tracer release, sampling commenced at least 45 min later to allow the tracer to disperse and equilibrate. Analytical Procedure. Inert tracer (PMCP) concentrations were determined by first cryogenically trapping and focusing the analytes on a trap containing Carboxen 569, 40-50 mesh before analysis by GCMS on an Agilent 5973 mass selective detector (Agilent Technologies, Ltd.) in NICI mode using methane as a reagent gas (Grade 4.5, Air Products Ltd., held at a pressure of 2 × 10-4 Torr). Selected ion monitoring (SIM) mode was employed so only relevant molecular anions were detected. The analytical method is described in full by Cooke et al. (24) with one important alteration: the use of a 30 m × 0.32 mm graphitized carbon (Carbograph) PLOT capillary column (30 m × 0.32 mm i.d, Lara s.r.l, Rome, Italy) capable of separating all six isomers of perfluorodimethylcyclohexane (PDMCH). Samples were identified and quantified through use of a calibrated external standard. Any concern regarding the linear concentration range of the system toward high levels of tracer could be addressed by examining PFC impurities contained within the tracer mixture. For example, the ratio between PMCP and the impurity PDMCH (contained within the mixture at
