Jump to main content.

Leaves as an Indicator of Exposure to Airborne Volatile Organic Compounds

Michael H. Hiatt
U.S. Environmental Protection Agency, National Exposure Research Laboratory
Environmental Sciences Division. P.O. Box 93478, Las Vegas, Nevada 89193-3478

  Phone: 702 798 2381. Fax: 702 798 2142. 
E-mail: hiatt.mike@epa.gov.
[Note:  minor content and formatting differences exist between this web 
version and the published version]


The concentration of volatile organic compounds (VOCs) in leaves is primarily a product of airborne exposures and dependent upon bioconcentration factors and release rates. The bioconcentration factors for VOCs in grass are found to be related to their partitioning between octanol and air equivalent to a relationship previously determined for PCBs. The rate that leaves release VOCs is dependent upon meteorological conditions and the enthalpy of phase change between air and plant. The enthalpy of phase change (DHPA) for a compound in leaves is closely related to its enthalpy of vaporization. The BCF and DHPA for a compound vary among plants but are highly correlated to each other. The change in BCF by plant (and correlated change in DHPA) is likely due to differences in the amount of octanol-equivalent matter contained in their leaves.

The concentration of airborne VOCs is predicted to maximize near dawn simultaneous with natural inversion patterns. A model incorporating this phenomenon with other meteorological data, DHPA and BCF is a useful tool predicting concentrations of VOCs in leaves. Vegetation can be especially useful in capturing VOCs at the critical time that air exposures are greatest. How long a leaf might retain a compound after uptake is dependent on the compound, the leaf type, and the magnitude of the wind and temperature. During calm weather leaves can be used as a record of these early morning exposures. However, windy conditions quickly clear leaves of their VOC content.


The presence of volatile organic compounds (VOCs) throughout the environment is well documented. The ubiquitous nature of VOCs is substantiated by the detection of VOCs in leaves from the Mt. Everest tree line (1), seawater throughout the globe (2), and both sediments (3) and air (4) from Antarctica. It has been suggested that semivolatile compounds, such as polychlorinated biphenyls and pesticides, are transported to colder regions by global distillation (5), although additional studies are necessary to prove this hypothesis (6). Studies of the migration of more volatile compounds would certainly shed light on the ultimate destination of airborne organic compounds.

Concurrent with the airborne movement of organic compounds is their foliar uptake. The amount of chemical that can be present in leaves is related to a bioconcentration factor (BCF) which for lipophilic organic compounds has been shown to correspond to their partition coefficients between air and octanol (7, 8, 9), and to vary by plant type (7) and temperature (7, 10). While a portion of the lipophilic compounds may reside interior to the plant cuticle (11) the leaf may be considered as having only octanol-equivalent, aqueous, and gaseous compartments (12). This simplified leaf model accumulates the lipophilic compounds in the cuticle independent of routes (12).

It cannot be assumed that the concentration of a compound in leaves would always reflect an equilibrium with air although the more volatile a compound is the more likely it is to be in equilibrium (13). The uptake of lipophilic semivolatile compounds by vegetation has been established as primarily leaf surface adsorption at a rate limited by atmospheric resistance (13). The lack of influence by route (stomata or cuticular surfaces) on uptake of lipophilic compounds has also been shown for semivolatile compounds (11). As passive diffusion has been identified as the primary force in moving these compounds, it is reasonable to assume the trends found describing semivolatile uptake by plants would describe the behavior of lipophilic VOCs. While VOCs would certainly diffuse more rapidly than semivolatiles, the relative importance of diffusion through the different media would remain.

The purpose for this study was to observe the content of VOCs in leaves and air and identify how the concentration in leaves might be explained in relation to their concentration in air and if these findings were consistent with investigations involving less volatile compounds. After establishing how the content of VOCs in leaves reflects exposure to airborne VOCs, a secondary goal of this work was to assess the limitations of using leaves as a means to measure an exposure of vegetation to VOCs. Expanding on this work, the analyses of leaves from remote sites will gauge wether the content of VOCs in leaves can indicate a pristine condition or, potentially, a sink.

The impact of meteorological conditions on BCF and rates of equilibration is critical to the interpretation of results from the analysis of vegetation (14). The monitoring of vegetation in a natural environment provides an opportunity to assess the impact of meteorological conditions with theory. The drawback to this uncontrolled approach is that multiple variables are impacting leaf-air dynamics simultaneously, and the isolation of a single variable for study (i.e., air concentrations, temperature, laminar air flow) was not possible. Therefore, all variables had to be addressed simultaneously.


Vacuum Distiller.  Samples were analyzed using a vacuum distiller coupled with GC/MS. The vacuum distiller serves to remove analytes from a sample by volatilizing the analytes in a reduced pressure environment. A condenser column serves to condense most water that is also being volatilized at the reduced pressure. The material that passes through the condenser is cryogenically trapped in the vacuum distiller cryotrap. After the vacuum distillation is complete, the cryotrap is ballistically heated to volatilize the distillate while a helium carrier gas sweeps the trap transferring the analytes to the GC/MS via a transfer line. The direction of flow at the cryotrap is controlled by a 6-port valve.

The vacuum distiller used in this study has been previously described in detail (7, 15). The principal components of the apparatus are a cryogenically cooled condenser (Associated Design & Manufacturing Company, Alexandria, VA), solenoid valves (Peter Paul Series 70 Model 72 solenoid-action valves from Peter Paul Electronics Co., Inc., New Britain, CT), a 6-port Valco electronic actuated valve, and a cryofocusing trap (Specialty Fitting and Assembly, Cincinnati, OH). An Edwards vacuum pump (Model 1) was used to supply vacuum to the distiller. Operation of the apparatus was controlled by a microprocessor (Bitstream Technologies, Inc., Las Vegas, NV).

The temperature of the cryotrap six-port valve was maintained at 150 C (Valcon E rotor). All transfer lines were heated to 90 C. The transfer line between the vacuum distiller and the GC was heated to 170 C. A Pirani vacuum gauge (Edwards Model 1000) was placed at the vacuum pump to monitor the integrity of the apparatus under vacuum. A nitrogen flushing of the condenser between distillations was by a line (5 psi nitrogen) connected at the top of the condenser column with a nitrogen vent at the bottom of the condenser.

The sample chamber containing the spiked leaf sample (room temperature) was distilled under vacuum for 5 min. Water vapor was collected on the condenser column (5.0 2.5 C) while the fraction containing the analytes was collected in the cryoloop cooled with liquid nitrogen (-196 C). The sample chamber valve closed at the completion of the vacuum distillation and the cryoloop valve switched to allow the GC carrier gas to sweep the cryoloop and pass to the capillary column. The cryoloop temperature was ballistically heated to 120 C to volatilize the distillate. The transfer of the distillate to the GC was complete after 3 minutes. The cryoloop valve was returned to its original position, and the cryoloop heated to 200 C for 7 min. After the sample was vacuum distilled, the condenser column was heated to 90 C and flushed with nitrogen gas, while the nitrogen gas/condenser valve and vent valve were opened to remove most of the condensed material. After 3 min the condenser-nitrogen gas line and vent valves were closed. The last step was an evacuation of the condenser for an additional 10-min period to remove any condensed water and contaminants that might remain after the nitrogen flushing.

GC/MS Apparatus.  A Hewlett-Packard mass spectrometer (Model 5972) and gas chromatograph (HP5890 Series II with Model MJSC metal jet separator) with a 60-m x 0.53-mm i.d., 3.0-m film thickness, VOCOL capillary column (Supelco, Bellefonte, PA) was used for the determination of analytes from the vacuum distillation apparatus. Gas chromatograph operating conditions were: 3 min at 10 C; 50 C/min ramp to 40 C; 5 C/min ramp to 120 C; 20 C/min ramp to 220 C; and isothermal at 220 C for 3.4 min, resulting in a total run time of 28 min. The jet separator was held at 210 C and the transfer line at 280 C. The injector was interfaced to the vacuum distillation apparatus by connecting the carrier inlet gas line to the cryoloop valve and then back to the injector. The injection or inlet temperature was 240 C and the inlet pressure was 10 psi.

The mass spectrometer was operated in the selected ion mode (SIM) with 100 ms dwell on each ion being monitored (Table 1). The additional sensitivity gained by using SIM became necessary to consistently detect and measure the analytes.

Leaf Analysis. Leaf samples of grass (mixed), mock orange (Pittosporum tobira), pine (Pinus eldavica), rosemary (Rosmarinus officinalis prostratus), and juniper (Juniperus sabina tamariscifolia) were analyzed. Before this study was completed, the juniper plants that were being monitored were removed during landscaping renovation, resulting in fewer plant species being investigated than initially planned.

As in the previous study, leaf samples were collected at the University of Nevada-Las Vegas campus. Leaves that were exposed to the wind and shaded during the sampling period were selected. Grass samples consisted of blades between 2 and 5 in. long and cut approximately 1 in. above the ground. The outer leaves approximately 2 ft above ground were collected from mock orange. Pine needles were taken at heights between 5 and 7 ft and removed from outer whorls. Samples of rosemary were taken from raised containers at a height of 3 ft. The rosemary samples consisted of leaves and stems. The individual leaves of rosemary were not separated from the stem. Rather, the outer portions of the plant where the stems had new growth and were still green (not woody) were selected. Measurements of the leaves physical properties that were used in calculations are presented in Table 2.

Fresh leaves (10 g wet weight) and 1 mL water were placed in the sample vessel. The presence of water helped suppress the recovery of alcohols that would decrease chromatographic resolution. There were no attempts to mince or mix the sample. The samples were then spiked with 10 mL methanol, containing surrogate compounds, directly into sample vessels (a 100 L round bottom flask fitted with a 15 mm O-ring connector), which Were used to contain the samples during both vacuum spike and vacuum distillation. Equilibration of the surrogate compounds and the tissue was accomplished using overnight vacuum spiking as previously described (16). Samples were analyzed by vacuum distillation/gas chromatography/mass spectrometry (VD/GC/MS) the following day.

The concentrations of the analytes in the leaves were determined using the surrogate-based matrix correction technique (16). For this study, the recoveries of the analytes from the matrix during a vacuum distillation were compared with the surrogates as a function of relative volatility between octanol and air (aKOA). With the analyte recovery prediction, their responses were corrected for matrix effects and then the concentration was determined.

Three groups of surrogates were used. The first group of surrogates was used to determine the matrix effects for analytes with aKOA values below 5000. The surrogates in the first group were: fluorobenzene; 1,4-fluorobenzene; toluene-d8; and 1,2-dibromoethane-d4. The second group of surrogates was used to predict recoveries for those analytes with aKOA values between 5000 and 22,000. The second group's surrogates were: 1,2-dibromoethane-d4; bromobenzene-d5; and 1,2-dichlorobenzene-d4. The third group of surrogates was used to predict the recovery of analytes with aKOA greater than 22,000. This group of surrogates was: 1,2-dichlorobenzene-d4; 1,2,4-trichlorobenzene-d3; and naphthalene-d8. The list of aKOA values is presented in Table 1.

The concentrations of volatile organic compounds were determined relative to the dry-weight of the leaves. The dry weight for leaves was determined as the weight after heating at 115 C overnight. The samples were dried in the sample vessel after their vacuum distillation. The sample dry-weight was used in the determination of the BCF.

Air Analysis.  Air samples were collected using a 1 L Erlenmeyer flask modified with an O-ring connector (15 mm). The Erlenmeyer flask was sealed using a cap made from a stainless steel O-ring connector (15 mm) attached to a stainless steel toggle valve with Swagelock fittings (Nupro 4BKT). Before sample collection, the flask and valve/cap were attached (1/4 in. Swagelock fittings) to a port on the vacuum distillation apparatus and evacuated. An air grab sample was then collected in the flask (disconnecting the O-ring connection), 5 mL surrogate solution was added, and then the valve/O-ring cap was connected. The flask assembly was then reattached to the vacuum distiller for analyses. The air sample was evacuated from the flask (open cap-toggle valve), passed through the vacuum distiller condenser, and focused in the system cryotrap (-196 C). The condenser was held at 90 C during the sample transfer. After 5 min of collecting the analytes in the cryotrap, the valving was switched to desorb mode, and the cryotrap heated to 100 C, transferring the focused material to the GC/MS. The determination of analytes in the air samples was performed using external standard quantitation. Results were not corrected to STP.

The concentrations of VOCs varied among the different days. Benzene, tetrachloroethene, ethylbenzene, 1,4-dichlorobenzene, and naphthalene concentrations ranged from less than 0.1 to 10 ng/L. Toluene varied between 1 and 113 ng/L, and the sum of m- and p-xylenes varied between 0.3 and 19 ng/L.

Sample Collection.  Samples were collected on 11 different days during the period of January 1998 to August 1998. Leaf and air samples were collected at roughly 1 h intervals from 7:00 a.m. to 11:00 a.m. At each hour, a sample of air and leaves were collected. In this study the collection of samples on a given day is referred to as a sampling event and the time between successive samples is referred to as interval. During a sampling event, only two plant species were sampled as they were within several feet of each other and the sampling of leaves and air could be completed within 10 minutes. Mock orange was directly adjacent to grass at one sampling location and pine adjacent to rosemary at the other sampling location. There were six sampling events undertaken to investigate pine and rosemary leaves and five sampling events investigating grass and mock orange leaves.

Meteorologic data used in this study were collected from the Los Alamos National Laboratory NEWNET station located at University of Nevada at Las Vegas. The data are accessed on the internet (htpp://newnet.jdola.lanl.gov/data_frame.asp?number=0214). Every 15 min, meteorological readings of temperature, average wind speed, and wind direction were recorded by the station less than 100 yd from both sampling locations. The 7:00 a.m. to 11:00 a.m. sampling period was selected as it was the most likely to provide the greatest changes in air concentrations. It was typical for concentrations of VOCs in air to maximize and then drop quickly during this period. This pattern is standard for the Las Vegas Valley as early morning air inversions are quickly dispersed as the morning temperature increases. The greater the change in VOC concentrations in air, the greater the potential to observe an impact on leaf concentrations.

Initial sampling events (one for the grass and mock orange sampling and one for pine and rosemary sampling) were designed to determine BCF. For those sampling events one sample of air and leaves were taken at 5:30 a.m. prior to the time before the concentration of VOCs would be at a maximum (approximately 7:30 a.m.). This sampling pattern was used to detect if the concentration of VOCs in leaves reflected current air concentrations or a prior exposure. If the concentration of VOCs in air were maximum and the concentration of VOCs in leaves increased from the 5:30 leaf sample then the BCF could be determined. If the leaf concentration did not increase, it would be established that the concentrations of VOCs in leaves were greater than equilibrium (the leaves would be releasing VOCs) and a determination of BCF would be biased by a previous exposure. For both of the initial sampling events the samples did increase and BCF values were determined.

The remaining sampling events were designed to collect samples during the periods when the leaves would be releasing VOCs. During the sample events there were typically five sample sets (leaves and air samples) collected. These sample collections resulted with 23 intervals that could be used to evaluate concentrations of VOCs in air and pine-rosemary leaves. There were 19 intervals for the evaluation of VOCs in air and grass-mock leaves.

The determination of release rates over an interval was not always useful and criteria had to be established. The first criterion that was used was the concentration of the VOC in air had to drop 25% or more over the interval. This requirement was intended to minimize the impact of bias from analytical variability (less than 10%) or an exposure spike from a minor random source. This criterion was the most restrictive and resulted in the rejection of 54% of all the intervals (for all plants and all compounds). Another criterion was that the wind direction did not vary more than 90 during the interval. This limitation was a means to eliminate using intervals when the change in air concentration of VOCs could not be considered as uniform during the interval. This criterion eliminated the use of 11% of the intervals.

It was also necessary that the VOC concentration in leaves had to show a decline over the interval. If the concentration in leaves increased while the concentration in air decreased a determination of release rates would not be reliable. Two possible conditions would result in these observations and would result in erroneous calculations of release rates. One was that while the concentration in air would drop, it would still be so elevated (while the concentration in leaves less than its equilibrium concentration) that the leaves continued to uptake the VOC. The second possibility was that the concentration of a VOC in air increased as a spike (intermittent source) after the start of an interval and had subsided by the time the end of the interval. This criterion was not met for 11% of the intervals.

The final precondition that was used was an instrumental sensitivity criterion. That requirement was that the analyte had to have a response (GC/MS) greater than 3 times background. This criterion eliminated few data (6%) with most rejections (80%) due to a low concentration of benzene in leaves.


Bioconcentration Factors.  In a previous study, the BCFs for the VOCs used in this study were determined. That study took both air and leaf samples early in the morning at the time when the concentration of VOCs in air was predicted to be the greatest. Noted in that publication was a potential high bias in the determination of BCF values as the study did not ensure that the concentration of a compound in the leaves was not biased from a previous exposure.

For this study, BCF values (concentration of VOC on a dry-weight basis compared to concentration in air) were determined from sampling events where leaves were shown to not reflect a concentration of VOC from an exposure prior to the sampling event. The BCF values where then determined using the samples collected when the air concentration was maximized. The BCF values were standardized to 20 C by factoring the impact of compound vapor pressure changes (8), and their values are presented in Table 3. While it is useful to report BCF in terms of dry weight of leaves it is noted that in a later discussion the BCF is converted to a volume of leaves basis for the determination of enthalpy of phase change between plant and air.

The BCF values determined in this study compare closely to previous work (8) for the leaves with the exception of rosemary. It appears that the earlier rosemary BCFs were biased high, potentially by a residual content of VOCs from previous days' exposures.

The BCF values for grass determined by this study as well as those that had been determined for PCBs in grass (10) are presented in Table 4. The VOC data is seen to follow the same trend established using just the PCB data (Figure 1). The correlation of BCF to KOA for PCBs in grass had been shown to be log BCF = 1.0928 log KOA - 2.5258 with r2 = 0.9906. The additional VOC data indicate that the range of compounds that can be described by a linear relationship of BCF to KOA can be expanded. The resultant relationship of BCF to KOA is changed little with log BCF = 0.9728 log KOA - 1.517 ( r2 = .9909). This revised relationship should be a useful guide to predict the BCF for a wide range of organic compounds that would be expected to be transported via air.

The higher BCF value for some VOCs in pine and rosemary is likely due to the presence of organic matter (i.e., oils, lipids, and waxes) in their leaves that can dissolve greater amounts of VOCs as compared to grass. Grass has a reported 1% lipid content (cutin being 0.7%) by leaf volume (9) whereas conifer needles have been reported to have just essential oils up to 1% dry-weight or 0.44% by volume (17). In contrast, for the least volatile VOCs the BCF appears to vary less among the plants studied. These data suggests (with some risk) that for a lipophilic compound that is semivolatile, such as the PCBs and pesticides, its BCF will tend to be less a variable among different plant types.

Release ratesThe change in concentration of a compound in leaves in response to its concentration can be expressed as

dcl/dt = k1ca - k2cl (1)

where cl is the concentration of a VOC in leaves as ng/kg dry weight, k1 is the uptake rate in h-1, k2 is the release rate in h-1, and ca is the concentration in air as ng/L (18). The release of VOCs from leaves is at a much slower rate than their uptake, with the ratio of uptake-to-release rates expressed as

k1/k2 = BCF (2)

where BCF is expressed as ng/kg leaf dry weight to ng/L air.

As it was not feasible to collect air samples more frequent than an hourly basis it was necessary to assume that the change in a VOC concentration in air over an interval t (approximately 1 h) was gradual. A simplification that made integration of eq 1 more palatable was the BCF over an interval was considered to be a constant and equal to the BCF for the temperature at the beginning of the interval. Integrating eq. 1 over time t and substituting Bk2 for k1 yields

cl t = cl o e-k2t + BCF ( rk 2 -1 e-k2t - cao e-k2t + rt - rk2-1 + cao ) (3)

where cl t is equal to the concentration (ng/kg dry weight) of a compound in the leaf at the end of the interval t, cl o is equal to the concentration at the start, cao is the concentration in air at the start of the interval in ng/L, and r is the rate the concentration in air changes over the time interval (cat = rt + cao).

The analyses of air and leaf samples at the start and end of interval, t, provide the empirical data necessary to determine the uptake and release rates. It was found that when the VOC concentrations in air and leaves were near equilibrium, the determined rates were unreliable or the equation was not solvable. Therefore rates were only determined when the concentration of VOCs in leaves was not near equilibration with air (greater than 5% difference).

As these sample collections were done on different days over an eight-month period, the determined rates reflected significantly different meteorological conditions. The release rates required an elaboration to incorporate these meteorological variables in order to compare results. In this work the release of VOCs by leaves is related to the enthalpy (DHPA) of phase change between the plant and air (10). The Arrhenius equation

k2 = Ae-DHPA/RT (4)

was used to describe this dependence of the rate of release on this energy and the influence of temperature. This equation also provides a simple approach for addressing the other variables affecting release rates (with DHPA in KJ mol-1, the gas constant R in units of KJ mol-1 K-1, temperature T in Kelvin, and the frequency term A in units of h-1). Because A is a geometric term describing diffusion of compounds through a boundary of air surrounding a volume of leaf, eq 4 is in terms of leaf volume. Therefore the determination of k2 by eq 3 must be in terms of volume (BCFv is ng/L leaves per ng/L air and concentrations of VOCs in leaves are in ng/L leaves).

The frequency term was equated to be the ability of a chemical to pass through the laminar flow boundary to the surface of the leaves, corresponding to the mass transfer coefficient times leaf surface area (MTC S) in ref 12. This term was determined by the equation

A = [Da/(Lc)]S (5)

where Da is the diffusion coefficient for a chemical in air in cm2/h at temperature T, Lc is the thickness (in cm) of the laminar flow boundary at the leaf's surface, and S is the leaf surface per unit volume (cm-1). The diffusion coefficients for the VOCs at the temperature of sampling were calculated using the Fuller, Schettler, and Giddings estimation method (19).

Wind speed and the geometric shape of a leaf govern the thickness of the laminar flow boundary. The laminar boundary thickness (in mm) surrounding flat leaves (mock orange and grass) is described by the equation

L = 4.0 (l/v) (6)

where l is the mean length in the downwind direction in meters, and v is the wind speed in m/s, and the factor 4.0 has units of mm s1/2 (20). The laminar boundary thickness (mm) surrounding cylindrical leaves (pine and rosemary) is described using the equation

L = 5.8 (d/v) (7)

where d is the leaf diameter in meters and v is the wind speed in m/s and the 5.8 factor has units of mm s1/2 (20). Table 2 lists the values for physical properties of leaves necessary to solve eqs 6 and 7. With the measurement of wind speed and temperature from the NewNet station, A can be calculated.

With values for k2 (calculated from eq 3) and A, the DHPA can be determined by eq 3. Their results are presented in Table 3. One generalization would be that the DHPA for all the VOCs in this study is approximately equal to the heat of vaporization at the boiling point (DHPA = * Hvb where = 0.92 .10, 0.93 .07, 1.05 .06, and 1.04 .09 respectively for grass, mock orange, pine, and rosemary). It was also found that the DHPA for a compound varies among the plants as does the BCF. The correlation of DHPA to BCF was 0.002, .79, .53, .63, .88, .57, and .90 for benzene, toluene, tetrachloroethene, ethylbenzene, m,p-xylenes, 1,4-dichlorobenzene, and naphthalene respectively. It is reasonable to expect that the DHPA correlates to the content of an octanol-equivalent matter as does the BCF. The fact that benzene does not fit this pattern is most likely due to the difficulty of measuring the compound in leaves and distinguishing its response from background.

The averages of the DHPA results for VOCs in grass are presented in Table 4. Included in Table 4 are DHPA values determined for PCBs in grass (10). A linear relationship of DHPA to Hvap that was determined for PCBs is no longer linear when the VOC data are included (Figure 3). Such an inconsistency had been considered likely with a suggestion that the more lipophilic a compound is, the interaction with grass is greater than its interaction with the pure liquid (10). While a trend is not readily observable with just the VOC DHPA to Hvap data Figure 3 indicates a nonlinear relationship. The equation DHPA = 30.7 + 138.02/(1+(Hvap/85.74)-7.065) is accurate over the range of analytes with r2 = .99.

Leaves as passive samplersThe concentration of VOCs in leaves reflects their concentration in air, BCF and DHPA values, leaf geometry, and meteorological conditions. All these factors must be considered when evaluating the concentration of a VOC in leaves in terms of an airborne exposure. The changes in meteorological conditions over a day influence the concentrations of VOCs in leaves throughout the day. Therefore how the meteorological conditions change is a critical factor that must also be considered. The following discussion focuses on how well leaves would be expected to retain VOCs uptaken during an early morning exposure.

In order that the impact of the these factors can be discussed a pattern for how VOC concentrations in air and meteorological parameters change is defined. That is at 7:00 a.m. the concentration of VOCs in air are maximum and the temperature is a minimum, and at noon the concentration of VOCs in air would be at a minimum and the temperature at its maximum. The changes between maxima and minima are at an even rate. The wind speed is taken to be a constant throughout the day. An additional pattern is considered to represent an extreme where there is a significant change in wind direction that introduces air flow with negligible VOC content. This condition is simulated by making the concentration of VOCs in air negligible one hour after their maximum at 7:00 a.m. With these patterns, eq 3 can be used to incrementally determine the concentration of VOCs in leaves.

The influence of temperature on how well leaves retain VOCs is evaluated using two temperature ranges. One temperature range represents a cold day (0 to 10 C) and one a warm day (20 to 30 C). Three wind speeds (0.5, 1.0 and 5.0 m s-1) are considered. The minimum concentration of VOC in air is taken to be 1/10 its maximum. Starting with negligible concentrations of VOCs in leaves on Day 1 eq 3 is used to predict the concentrations over 7 days. Seven days were characterized to avoid an influence by an arbitrary starting concentration of VOCs in leaves for Day 1. The concentrations of VOCs in leaves for Day 7 are those being evaluated.

The effects of wind speed and the two patterns of VOC concentrations in air are presented in Figure 3. This figure illustrates how the concentration of toluene in pine needles varies on a cold day for the different wind speeds and the two air concentration patterns. This figure shows that while wind speed can have a significant effect, a "clean" wind (possibly from a different direction) can clear the leaves of toluene quite rapidly.

One characteristic observed in Figure 3 is that the relative concentration of toluene in pine needles still increases while the concentration of toluene air is dropping. This increase is because with the gradual changing air concentration pattern, the content of toluene in air gradually increases from the previous days low (noon) until 7:00 a.m. and the content of toluene in the pine needles does not reach equilibration with the air at 7:00 a.m. In fact the lag is such that for the 0.5 m/s wind speed (and therefore slower uptake rates) the content of toluene in pine needles continues to increase until 9:00 a.m.

It is apparent from Figure 3 that if leaves are expected to retain VOCs over a morning, a calm day (low wind speed and steady direction) is preferred. Likewise, the concentration of VOCs in leaves on a windy day would be very time dependent and less reliable. An interpretation of this figure would be that the concentration of toluene in pine would not be expected to vary from its 7:00 a.m. concentration by more that 10% on a cold day when the winds are less than 1ms-1.

The 10:00 a.m. concentrations of the VOCs in grass and pine are presented in Table 5 as relative to their concentrations at 7:00 a.m. From this table it can be seen that the VOC content in leaves can be representative of airborne content during a predictable morning inversion. If the concentration of VOCs in air change gradually during a morning (a reasonable assumption if wind is not gusting and changing directions) the loss of VOC over three hours would be expected to be less than 50% on even a warm windy day. Under calm conditions the concentration of VOCs in leaves hours after a peak exposure can be useful in measuring that peak exposure. The data in Table 5 also indicate that if there is a rapid change in air concentration the VOCs could be quickly cleared, especially for grass.

The use of leaves to measure airborne exposure requires the knowledge of meteorological conditions. While leaves can provide a useful record of an early morning exposure, they can be quickly cleared of VOCs by gusting conditions. The use of leaves to measure exposure to a morning high requires morning sampling. How well the samples represent a peak exposure is dependent on plant type, compound, temperature and wind conditions. It would be unlikely that meteorological conditions would justify using leaves to measure exposures after early morning.


The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD), funded and performed the research described. This manuscript has been subjected to the EPA's e review and has been approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.


  • Ciccioli, P.; Brancaleoni, E. Cecinato, A.; Sparapani, R. J. Chromatog. 1993, 643, 55-69.
  • Wiedman, T.; Guethner, B.; Class, T.; Ballschmiter, K. Environ. Sci. Technol. 1994, 28, 2231-2329.
  • Lenihan, H. Mar. Pollut. Bull. 1992, 25, 318-323.
  • Cicciola, P.; Cecinato, A.; Brancaleoni, E.; Frattoni, M. Internat. J. Environ. Anal. Chem. 1996, 62, 245-253.
  • Mackay, D.; Wania, F. Sci. Total Environ. 1995, 160/161, 25-38.
  • Bignert, A.; Olsson, M.; Persson, W.; Jensen, S.; Zakrisson, S.; Litzen, K.; Eriksson, U.; Haggberg, L.; Alsberg, T. Environ. Pollut. 1998, 99, 177-198.
  • Hiatt, M. Anal. Chem. 1998, 70, 851-856.
  • Welke, B.; Ettlinger, K.; Riederer, M. Environ. Sci. Technol. 1998, 32, 1099-1104.
  • Tolls, J.; McLachlan, M. Environ. Sci. Technol. 1994, 28, 159-166.
  • Komp, P.; McLachlan, M. Environ. Sci. Technol. 1997, 31, 886-890.
  • Riederer, M., Environ. Sci. Technol. 1990, 24, 829-837.
  • Paterson, S.; Mackay, D.; Bacci, E.; Calamari, D. Environ. Sci. Technol. 1991, 25, 866-871.
  • McLachlan, M.; Welsh-Paush, K.; Tolls, J. Environ. Sci. Technol. 1995, 29, 1998-2004
  • Ockenden, W.; Steinnes, E.; Parker, C.; Jones, K. Environ. Sci. Technol. 1998, 32, 2721-2726.
  • Hiatt, M.; Farr, C. Anal. Chem. 1995, 67, 426-433.
  • Hiatt, M. Anal. Chem. 1997, 69, 1127-1134.
  • Orav, A.; Kailas, T.; Liiv, M. Chromatographia 1996, 43, 215-219.
  • Bacci, E.; Cerejeira, M.; Gaggi, C.; Chemello, G.; Calamari, D.; Vighi, M. Chemoshpere 1990, 21, 525-535.
  • Lyman, W.; Reehl, W.; Rosenblatt, D. Handbook of Chemical Property Estimation Methods: Environmental Behavior of Organic Compounds; American Chemical Society: Washington, DC, 1990.
  • Nobel, P. Physicochemical and Environmental Plant Physiology; Academic Press: San Diego, 1991.
  • Lide, D. Handbook of Chemistry and Physics, 74th ed., CRC Press: Boca Raton, 1993.




Analytical Environmental Chemistry Home Page
Vacuum Distillation Home Page
Environmental Sciences | Research & Development
National Exposure Research Laboratory
Author:  Mike Hiatt / Email:  Hiatt.Mike@epa.gov
Send questions or comments to the Information Desk
ESD Info Desk (contractor operated)

Local Navigation

Jump to main content.