Odorant Emissions from Intensive Pig Production Measured by Online

Jun 29, 2010 - Dezhao Liu , Michael Jørgen Hansen , Lise Bonne Guldberg , and Anders .... Monica Martinez , Anna Novelli , Uwe Parchatka , Gavin J. P...
4 downloads 0 Views 433KB Size
Environ. Sci. Technol. 2010, 44, 5894–5900

Odorant Emissions from Intensive Pig Production Measured by Online Proton-Transfer-Reaction Mass Spectrometry A N D E R S F E I L B E R G , * ,† D E Z H A O L I U , † ANDERS P. S. ADAMSEN,† MICHAEL J. HANSEN,† AND KRISTOFFER E. N. JONASSEN‡ Department of Biosystems Engineering, Aarhus University, Blichers Alle´ 20, DK-8830 Tjele, Denmark, and Danish Agriculture & Food Council, Pig Research Centre, Axeltorv 3, DK-1609 Copenhagen, Denmark

Received February 26, 2010. Revised manuscript received May 20, 2010. Accepted June 15, 2010.

Emission of odorous compounds from intensive livestock production is a cause of nuisance in populated rural areas. Knowledge on the chemical composition of odor and temporal variations in emissions are needed in order to identify factors of importance for emission rates and select proper abatement technologies. In this work, a method based on protontransfer-reaction mass spectrometry (PTR-MS) has been developed and tested for continuous measurements of odorant emissions from intensive pig production facilities. The method is assessed to cover all presently known important odorants from this type of animal production with adequate sensitivity and a time resolution of less than one minute. The sensitivity toward hydrogen sulfide is demonstrated to exhibit a pronounced humidity dependency, which can be included in the calibration procedure in order to achieve quantitative results for this compound. Application of the method at an experimental pig facility demonstrated strong temporal variations in emissions, including diurnal variation. Based on these first results, air exchange and animal activity are suggested to be of importance for emission rates of odorants. Highest emissions are seen for hydrogen sulfide and acetic acid, whereas key odorants are evaluated from tabulated odor threshold values to be hydrogen sulfide, methanethiol, 4-methylphenol, and butanoic acid.

Introduction Livestock production is associated with emissions of odorous volatile organic compounds and hydrogen sulfide (H2S), which adversely influence air quality in the vicinity of the sources (1, 2). Particularly, intensive pig production is a source of offending malodorous compounds including carboxylic acids, phenols, indoles, aldehydes, ketones, amines, and volatile reduced sulfur compounds (3-8). Many of these are degradation products of proteins and are therefore ubiquitous in emissions from intensive pig production facilities (9). In addition to odor, these compounds may also contribute to the burden of volatile organic compounds (VOC) in regions * Corresponding author phone: (+45) 8999 1937; fax: (+45) 8999 1619; e-mail: [email protected]. † Aarhus University. ‡ Danish Agriculture & Food Council. 5894

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

with intensive production, but this contribution is presently not well-known. Accurate chemical characterization of this complex mixture is necessary in order to estimate the agricultural contribution to ambient VOC levels but also for developing efficient emission abatement solutions such as waste management, ventilation optimization, or air treatment. In the past, a number of chemical methods have been applied, typically based on sample collection and laboratory analysis by gas chromatography with mass spectrometric detection (GC/MS) (3-7). However, the analytes present in ventilation air from pig production include compounds that are polar, volatile, and/or reactive, and problems related to sampling the full range of odorants therefore remain (7, 8, 10). Quantitative methods for measuring volatile organic sulfur compounds, e.g. dimethyl sulfide and methanethiol, and estimates of the contribution of these to odor nuisance are particularly challenging partly due to the high humidity of livestock ventilation air (8). Especially for methanethiol, improved methods are needed due to the low odor threshold value (11) as well as low sampling stability (8, 10) of this odorant. So far, a method for quantitative and selective measurements of relevant volatile sulfur compounds, including H2S, as well as other odorous VOC that are less volatile and more polar, has not been demonstrated. Furthermore, most published work include a limited number of discrete samples and provide very little insight into the dynamics of emission, which is important in order to identify potentials for emission reduction. Recently, Feilberg et al. (12) presented continuous membrane inlet mass spectrometry (MIMS) data on odorant reduction in biological air filters treating air from pig houses. These measurements suggested that low removal of reduced organic sulfur compounds is a major limitation in odor removal efficiency of such filters. However, lack of adequate field calibration, reduced selectivity, and sensitivity were identified as major shortcomings of the method with respect to quantitative and specific odorant emission estimates. PTR-MS (proton-transfer-reaction mass spectrometry) has been demonstrated to be a promising tool for time-resolved measurement of emissions of selected VOC from dairy farms (13, 14). The method is based on chemical ionization by protonated water and is characterized by very high selectivity and sensitivity together with short response times (15-17). Since PTR-MS is a direct method with no sample collection, it holds a potential for overcoming sampling issues observed in previous methods for measuring odorants. PTR-MS has previously been used for characterizing an olfactometer (18) as well as for measuring odorous compounds from e.g. composting plants (19, 20) and household waste (21). In this paper, a method for measuring odorous VOC based on PTR-MS is characterized and applied at an experimental pig production unit. The purpose of the work presented is to test PTR-MS as a quantitative method under field conditions and to evaluate the importance of obtaining timeresolved odorant data.

Methods and Materials Location. Field experiments were carried out at the experimental pig production facility, Grønhøj, located in MidJutland, Denmark, and managed by Danish Agriculture & Food Council, Pig Research Centre. Continuous measurements were performed at the outlet of a pig section consisting of two finisher pens. The section is equipped with mechanical ventilation via a ventilation duct through the roof and with diffuse air intake through the ceiling. It was not possible to collect inlet air after passing through the ceiling without bias 10.1021/es100483s

 2010 American Chemical Society

Published on Web 06/29/2010

from the room air. However, the inlet air is filtered through the highly porous glass wool ceiling insulation at relatively high residence time, and, furthermore, the air intake for the ventilation system was extended to 5 m above the roof ridge in order to avoid taking in air from the exhaust duct. It is therefore assumed that background concentrations are negligible. Each pen contained 16 pigs and was equipped with a simple dry feeder and a nipple drinker. Feed was supplied ad libitum. The floor consisted of 1/3 drained floor and 2/3 slatted floor and a 60 cm deep slurry pit. The ventilation rate was controlled by the room temperature, which was attempted to be maintained at 19 °C during the period of measurements. The pigs entered the pens on May 25 2009 with an average weight of 31.7 kg. The average weight on July 6 (after the measurement period) was 64.2 kg. The slurry pit was emptied before the pigs entered and again on July 13. Instrumentation and Calibration. A High-Sensitivity PTR-MS (Ionicon Analytik, Innsbruck, Austria) was applied for the continuous measurements. The principle of PTR-MS has been described in detail in recent review papers (16, 22). The PTR-MS was operated under standard ion drift tube conditions applying a total voltage of 600 V and maintaining the pressure in the range of 2.1-2.2 mbar. The temperature of the drift tube was controlled at 60 °C. The inlet flow during measurement as well as calibration was ∼100 mL/min. The E/N-value was ∼135 Td for all measurements. A Dynacal permeation calibrator (VICI, Switzerland) was used to calibrate the PTR-MS for a range of compounds known to be present in exhaust air from pig facilities. Permeation tubes containing the following compounds were available: acetic acid (AA), propanoic acid (PrA), butanoic acid (BA), hydrogen sulfide (H2S), methanethiol (MT), dimethyl sulfide (DMS), and 3-methylphenol (used as surrogate for 4-methylphenol; 4MP). The emission rates of the permeation tubes were determined gravimetrically, and the uncertainty ranged from 2-15%. Permeation rates were in the range of 50 to 4000 ng/min. Dry zero-air for the laboratory tests and calibrations were produced from pressurized air by using a cold trap (obtained dewpoint: ∼-30 °C) and a charcoal filter. Further hydrocarbon removal was ensured by purification via a Supelpure HC filter (Supelco, USA). Zeroair controlled by a mass flow controller (Sierra Instruments, USA) was passed through the permeation oven. The output was further diluted by using a mass flow controller. In addition, calibration was performed by zero-air dilution of certified gas standards (Air Liquide, Germany) containing close to 5 ppm in N2 of H2S, MT, DMS, dimethyl disulfide (DMDS), and trimethylamine (TMA), respectively. The standards were diluted by a factor of 20 or higher. The accuracy of the gas standard concentrations was (10%. The humidity dependency of the sensitivity of the PTRMS was investigated by diluting the output from the permeation oven with zero-air bubbled through a water trap via a frit diffuser. Calibrations were performed in the laboratory before and after the measurement campaign. For compounds for which calibration standards were not available, the sensitivity was calculated based on the rate constant for proton transfer and the estimated drift tube residence time, as described by e.g. de Gouw and Warneke (16) and applied for quantification of livestock emissions previously (13, 14). Proton transfer rate constants for sensitivity calculation were calculated for pentanoic acid, 4-ethylphenol, indole, and 3-methylindole based on the method of Su and Chesnavich (23). We used either tabulated experimental values of polarizability and dipole moment (24) or values from NIST (25) estimated by a method comparable to the one used by Zhao and Zhang (26). Estimated sensitivities are presented in the Supporting Information, Table S1.

For the field measurements presented in the current paper, we used factory set transmission factors as the measurements were initiated less than four months after the PTR-MS was received. These factors were subsequently checked with a mixture of aromatic compounds in concentrations close to 100 ppb ((10%; Restek, P/N 34423-PI), see the Supporting Information (Table S2). The uncertainties in concentration measurements were estimated by error propagation to be in the range of 10-26% depending on the calibration method (see the Supporting Information). Detection limits were determined based on blank measurements of zero air. Values in the range of 20 to 200 ppt were generally observed (Table 1) using a mass dwell time of 2 s. The detection limits are well below reported average odor threshold values of known odorants from livestock production (Table 1). Fragmentation patterns of selected compounds were investigated by placing a glass vial sealed with a silicone septum in a 0.25 L glass bottle. Diffusion through the septum prevented overloading the instrument. The surrounding air containing compounds diffusing through the septum was flushed and analyzed by PTR-MS in scan mode. Field Measurements. Field measurement data were collected in a period of four weeks from May 28, 2009 to June 25, 2009. The PTR-MS was located next to the climate chamber, and sampling air was drawn from ∼1 m above the exhaust duct entrance with the intake positioned approximately in the middle of the duct. A 6 m Teflon (FEP) sampling line (outer diameter 6 mm, inner diameter 4 mm) was used to draw air to the PTR-MS. This was connected to the instrument inlet system and drift tube via a 1 m PEEK sampling line (inner diameter: 1 mm). The main sampling line was insulated and temperature controlled at 40 °C by a heater strip, whereas the PEEK tube was controlled at 60 °C. A 5 µm PTFE filter (Millipore Millex-LS) was used to remove larger particles. Measurements were carried out continuously by means of selected ion monitoring (SIM) mode with each ion being detected for 2 s. A complete measurement cycle was 45 s. Full mass spectral scans were performed using a dwell time of 0.5 s per mass at the start-up and occasionally during the measurement period. Masses for SIM mode were selected based on chemical compositions reported for similar systems, known PTR-MS fragmentation patterns, and ion abundance in full scan mode. The ions monitored together with compound assignments are presented in Table 1. Supplementary screenings of compounds present in ventilated air was carried out by gas chromatography with mass spectrometric detection (GC/MS) on 9 air samples collected on three occasions by adsorption tubes packed with a Tenax TA/Carbograph 5TD combination (12). Samples were collected at 100 mL min-1 for 15 min. Concentrations of selected compounds were estimated based on 1-point calibrations by adding aliquots of compound solutions (in pentane and/or methanol) to the sorbent tubes and flushing the solvents by a flow of He. Ventilation rate was measured at 5 min intervals with a measurement wing (Fancom, Holland). Room temperature was measured with a VE10 temperature sensor (Veng System A/S, Denmark) placed immediately below the ventilation duct. NH3 was monitored every 140 min by an electrochemical sensor (Polytron 1, Dra¨ger, Germany) as described by Lyngbye et al (27). The NH3 sensor was calibrated on weekdays by comparison with detection tubes (Kitagawa, Japan). Slurry height was measured manually on a daily basis.

Results and Discussion Humidity Effects. Prior to field measurements, the influence of humidity on PTR-MS response was tested. As a part of this, the response toward H2S was tested since this odorant VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5895

TABLE 1. Compound Assignment Together with Ions Monitored by PTR-MS, Mean Concentrations with Ranges, Mean Emission Rates with Ranges, Detection Limits, Odor Threshold Values (OTV) for Relevant Compounds (11) and Mean Odor Activity Values (OAV) for Selected Compounds compound assignmenta

ion(s) (m/z)

Cair,mean (ppbv)

rangeb (ppbv)

emissionmean (mg h-1)

range (mg h-1)

DLc (ppbv)

OTV (ppbv)

OAVmean

hydrogen sulfide methanethiol acetone trimethylamine acetic acid dimethyl sulfide C4-carbonyls (e.g., 2-butanone) propanoic acid 2,3-butanedione butanoic acide phenol (+ dimethyl disulfide) C5 carboxylic acidsf 4-methylphenol indoleg 4-ethylphenol dimethyl trisulfide 3-methyl-1H-indole

35 49 59 60 61 + 43 63 73 75 + 57 87 89 + 71 95 103 + 85 109 118 123 127 132

265 4.0 5.1 9.4 209.5 4.9d 2.2 66.0 1.3 55.6 0.7 9.5 3.9 0.12 0.34 NDh 0.38

14-1723 0.9-10 2.3-12 2.5-21 69-393 1.5-15 0.5-10 20-137 0.5-2.5 16-115 0.1-1.8 2.8-19 1.1-6.8 0.04-0.25 0.14-0.53 0.08-0.9

643 14 24 41 956 21 13 365 8.7 369 6.3 75 36 1.4 3.4 3.4

30-3653 1.8-37 3.3-93 6.1-78 112-2545 3.2-48 1.4-99 47-921 1.3-21 46-1014 0.6-16 9.2-216 4.2-98 0.2-4.1 0.3-7.3 0.5-9.0

1-5 0.02 0.14 0.15 0.21 0.07 0.15 0.12 0.06 0.16 0.09 0.08 0.12 0.04 0.03 0.06 0.02

2.3 0.07

115 57

2.1 236 5.9

4.8 0.9 0.9

25 0.6 1.9 54.6 1.4 0.3 0.4 1.3 0.4 0.09

2.6 2.5 29 6.9 15.1 0.3 0.3 4.2

a

See text for discussion of potential interferences not mentioned in this table. b The range is presented by the 0.1%-percentile and the 99.9% percentile (n ) 6371). c DL: detection limits determined as three times the noise level of blank measurements. d Data have been corrected for O18 isotopic contribution from acetic acid. e Minor contribution from 2-methylpropanoic acid (10-20%). f Pentanoic acid and 3-methylbutanoic acid (ratio ∼2:1). Weighted average used for OTV and OAV. g Corrected for isotopic contributions from compounds giving m/z 117, assumed to be hexanoic acid. h ND: not detected.

is known to be abundant in livestock emissions and is expected to contribute significantly to odor nuisance. H2S has a relatively low proton affinity (705 kJ mol-1) only slightly higher than that of water (691 kJ mol-1) and is considered difficult to measure by PTR-MS due to the back-reaction of protonated H2S being relevant H3S+ + H2O f H2S + H3O+

(1)

Although the rate constant of (1) is significantly lower than the rate constant of H2S protonation by H3O+, the presence of water from sampling air and the ion source enhances the influence of (1). This results in a significant loss of sensitivity for H2S and a dependency on sampling humidity. However, it has previously been demonstrated for other compounds with relatively low proton affinities, hydrogen cyanide and formaldehyde, that PTR-MS calibration could be obtained by taking into account reaction 1 (28, 29). The humidity dependence of the PTR-MS response toward H2S as measured by m/z 35 was therefore investigated. Results of measurements on the output of an H2S permeation source diluted with air at varying humidity are presented in Figure 1. As can be seen, an empirical equation can be fitted to the observed data enabling a correction of the responses based on simultaneous measurements of the monohydrate cluster of H3O+ (H3O+(H2O); m/z 37) relative to H3O+ (measured as the O18 isotope at m/z 21) representing the air concentration of water. It should be emphasized that the amount of water clusters formed in the drift tube is dependent on the instrumental settings and therefore humidity correction was performed with settings identical to those used during measurements. A significant effect of humidity on the fragmentation of carboxylic acids has been identified, as shown in the Supporting Information (Figure S1) for AA. However, the sum of the protonated parent molecule and the fragments is constant within (5%. A similar trend was observed for the other carboxylic acids, but the fragmentation decreases significantly as a function of carbon number (e.g., for BA, the ratio of the fragment relative to the protonated molecular 5896

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

FIGURE 1. Relative response of PTR-MS to H2S as a function of humidity at two different concentrations (b: 830 ppb; O: 280 ppb) expressed as the concentration of the diluted output from the permeation oven (Cstd) relative to the observed concentration using a protonation rate constant of 2 × 10-9 cm-3 s-1 (Cobs). The full line represents the empirical logarithmic function (y ) a × ln(x) + b) fitted to the data. Relative humidity (2) is included as a function of the water cluster signal intensity of m/z 37 (H3O+(H2O)) relative to the primary ion (m/z 21). ion is 1 ppbv. A signal at m/z 35 ascribed to Table S3, Supporting Information), although there are H2S was clearly detected throughout the measurement period. exceptions as discussed in the following. Major compounds By applying the humidity correction described earlier, this observed in PTR-MS scans include oxygenated VOC such as important odor compound can be monitored by PTR-MS methanol (m/z 33) and acetaldehyde (m/z 45). In addition, despite the reduced sensitivity discussed above. All m/z 35 m/z 41 is ascribed to alcohol fragments based on the results data were corrected according to the relationship presented presented in the Supporting Information (Table S6). This in Figure 1. For most masses monitored, unambiguous fragment corresponds to C3H5+, which could hypothetically be formed by dehydration of e.g. protonated 1-propanol and compound assignment is possible, but a few special cases subsequent elimination of H2. m/z 46, if assumed to be a are discussed in the following. The signal at m/z 43 is mainly protonated parent molecule should contain an uneven ascribed to AA based on the ratios of m/z 43 to m/z 61 (see number of nitrogen atoms and may tentatively be ascribed Figures S1 and S2; Supporting Information). Alcohols present to dimethyl amine or ethylamine. The detection of this in the sample air will also contribute to m/z 43, due to the compound, however, can presently not be confirmed by other considerable fragmentation of these compounds (32); see methods. m/z 47 may be ascribed to ethanol (30) and/or also the Supporting Information (Table S6). However, alformic acid (35). Ethanol undergoes significant fragmentation cohols were observed by TD-GC/MS to be present at levels by loss of -H2O with the primary fragment being protonated less than 5% of AA. Therefore it is reasonable to use the sum ethylene (C2H5+) (13). Since ethylene has a lower proton of m/z 43 and m/z 61 to quantify AA, and the humidity effect affinity than water, this fragment ion will not be detected, on AA fragmentation can be disregarded according to the and ethanol can therefore not be quantified in a straightresults presented in Figure S1 (Supporting Information). This forward way. Screenings by TD-GC/MS revealed only a small was confirmed by plotting the concentration based on the ethanol peak (less than ∼1 ppb), whereas formic acid has sum of m/z 43 and m/z 61 versus the concentrations based previously been reported in air samples from pig houses on m/z 61 corrected for humidity and fragmentation (Sup(35). It is proposed that m/z 47 is mainly due to formic acid, porting Information, Figure S2). The signal at m/z 73 can but a contribution by ethanol cannot be excluded. The theoretically be ascribed to several C4 carbonyl compounds. Based on TD-GC/MS, it is proposed that 2-butanone is the magnitude of m/z 48 signals was too low to evaluate the compound dominantly contributing to this mass. Likewise, presence of ethanol from isotopic patterns. m/z 87 is mainly ascribed to 2,3-butanedione. m/z 95 may Other signals observed in full scan mode are mainly be due to the presence of both phenol and DMDS. Of these, ascribed to low level VOC fragments, aromatic compounds, the latter gives a fragment at m/z 79 (33), with a relative and C13 isotopic contributions (see Table S3, Supporting Information). abundance of ∼30% of m/z 95 in the specific instrument Odor Contributions. The measured concentration levels applied in current work (see the Supporting Information, and relative composition of odor compounds agree well with Table S6). Background benzene also gives m/z 79, and thus previous quantitative measurements of e.g. sulfur compounds the presence of DMDS cannot be determined from this (8, 31, 34) and other odor compounds recovered from fragment. Both phenol and DMDS have been reported to be VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5897

FIGURE 3. Diurnal variation of hourly averaged emissions of 4MP (∆), BA (2), MT (O), and DMS (b) together with ventilation rate divided by 5 for visual clarity (VR/5; s) and room temperature in °C (gray line) during a selected period of three days. sorption tubes e.g (7, 12). In order to assess the contributions to odor, an approach based on odor threshold values (OTVs) for individual compounds has been used. OTVs were estimated from ref 11 using geometric means of detection values. OTVs are included in Table 1 together with odor activity values (OAVs) calculated as mean concentration divided by OTV. The compounds estimated to mainly contribute to odor from the pig house are H2S, MT, BA, and 4MP. It should be emphasized that the application of OAV is valid only under the assumption that antagonistic and synergistic effects can be disregarded. For many compounds, OTVs are highly variable (up to several orders of magnitude), and more accurate estimates would strongly improve this method for estimating odor contribution. The presence and concentration levels of MT are supported by previous investigations (31, 34). The current study emphasizes the significant odor contribution of MT, which may have been overlooked in some previous investigations (3, 6, 30) since in these studies, conversion of MT to DMDS during sampling and/or analysis was not evaluated or corrected for. Observed Emission Patterns of Odorants. The dominant compounds emitted were observed to be H2S and volatile carboxylic acids. Compared to studies on emissions from dairy and cattle buildings (13, 14), emissions of alcohols are estimated to be relatively low. This is most likely because alcohols from dairy and cattle systems are emitted from fermented feed (silage) (36, 37), which does not occur in a pig facility using dry feed. For all compounds, the highest concentrations in the outlet air were observed at nighttime coinciding with low ventilation rates (for examples, see the Supporting Information). To observe any temporal variation in emission of odorants, emission rates (mg/h) for the whole section (32 pigs) for each compound were estimated by multiplying 5 min averaged concentrations measured by PTR-MS with the corresponding ventilation rates measured every 5 min. For most compounds, the emission rate increased gradually during the measurement period as shown in Figures S7 and 8 (Supporting Information) for selected compounds. Apparently, there was a steady increase in emissions of e.g. BA and 4MP with a 2-3-fold increase in emissions during this period of one month. Similar patterns were observed for other phenols, carboxylic acids, and indoles (see the Supporting Information). H2S emissions are relatively low in the early phase but increase rapidly after June 7 (13 days after the pigs entered the chamber) followed by a decline to intermediate levels. No stirring or other perturbations of the manure slurry, which could lead to increased H2S emission, 5898

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

took place during the period from June 7 to June 12. The observed trend in H2S emissions could be due to e.g. changes in the microbial production of H2S, changes in microbial surface oxidation of H2S, and changes in pH affecting the H2S/HS- equilibrium. A better understanding of the factors influencing emission of H2S may lead to development of manure management strategies aimed at reducing the emission of this important odorant. The emissions generally exhibit a diurnal variation with daytime maxima. An example of this pattern for a period of 3 days is presented in Figure 3 with hourly averaged data for the sake of visual clarity. For several compounds, emission peaks were observed during morning and afternoon with an intermediate minimum at noon. This pattern very much follows the animal activity as expected from the results of e.g. Blanes-Vidal et al. (38), who observed increasing activity from early morning, followed by decreased activity at noon and another activity peak in the afternoon. High animal activity is associated with excretion of urine and feces, which influences the slurry surface and may lead to increased emissions. In addition to activity, ventilation rate increases significantly during daytime, whereas room temperature increases moderately, especially on warm days where maximum ventilation is not sufficient to remove the heat. Room temperature does not directly reflect the temperature of the slurry surface, but on days with both high ventilation rates and high temperatures, the slurry surface temperature is expected to increase. The diurnal variation in emissions is most likely due to a combination of variations in activity, ventilation rate, and temperature. Separating the effects of temperature and ventilation rate is difficult since these are to some extent correlated. However, the highest temperatures are not associated with increased ventilation rates, since there is a maximum set point of 3200 m3 h-1 (see the Supporting Information, Figure S11A). As can be seen from Table S7 and Figure S11A-B, water-soluble compounds are to some extent correlated with temperature, whereas less soluble compounds (i.e., sulfur compounds and trimethylamine) are only weakly correlated with temperature. However, comparing these results with correlations with ventilation rate (Figure 4 and Figures S9 and S10), it is seen that relatively higher correlations with ventilation rate are observed. Previous results indicated that odor compounds primarily originate from manure (27, 31). Higher ventilation rates increase the air velocity near the manure surface (39), which has previously been observed to lead to increased emissions of ammonia (39) but may also increase the temperature of the surface as mentioned above. The manure surface area remained

Supporting Information Available Additional information on the analytical technique; assignments of detected masses other than those included in Table 1 together with the results of acetic acid fragmentation as a function of humidity, a test of the acetic acid correction method, and fragmentation patterns of selected compounds; results of supplementary GC/MS analyses and comparison with PTR-MS; and compound concentration data for the whole period and diurnally averaged emission data together with additional results of emissions versus ventilation rate, temperature, and ammonia emission. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited FIGURE 4. Diurnally averaged emission data versus diurnally averaged ventilation rates for the period from June 6 to June 25 shown for selected compounds: 4MP (4), BA (2), MT (O), and H2S (b). Full lines represent linear regressions of the data. constant, whereas manure volume increased linearly with time (R2 ) 0.99, data not shown). DMS represents a special case, since this was the only compound for which a negative correlation with ventilation rate was observed (Supporting Information, Figure S10). The reason for this is presently not clear, but a different source than other compounds, e.g. animal breath (40), could explain the results. A definitive conclusion regarding the causes of the variation in diurnally averaged data cannot presently be drawn. However, a plausible interpretation of the differences between emission patterns of sulfur compounds and other compounds (Figure 4 and Supporting Information) is as follows: 1) For compounds with low Henry’s law constants (e.g., carboxylic acids, phenols, and indoles; KH ) P/Caq), emission rate is increased with higher air velocity above the surface and/or higher temperature. This reduces the air side mass transfer resistance, which dominate over liquid side resistance for these relatively soluble compounds (41). 2) For compounds with high KH (sulfur compounds), the liquid side mass transfer resistance is expected to dominate , and the influence of air velocity and air temperature is therefore less important. An effect of manure volume/height cannot be excluded in this case since both manure volume and ventilation rate increase over time and therefore are correlated. In addition to the above, correlations with NH3 emissions have also been investigated. The results are presented in the Supporting Information, Table S7. This can be used to estimate rough emission inventories on a larger scale using NH3 as a reference, for which relatively detailed emission data exist. However, compared to ref 14, weaker correlations are observed and more data are needed in order to provide more valid regional emission data. Understanding the emission patterns of odorants is important for development of future pig houses and manure management systems with reduced odor impact. PTR-MS measurements has for the first time enabled insights into the temporal variation in emissions of odorants from intensive pig production. Future experiments under more controlled conditions with selected parameters kept constant, e.g. using slurry in wind tunnels, will further aid in elucidating factors of importance for emissions of odorants.

Acknowledgments The Danish Food Industry Agency under the Ministry of Food, Agriculture and Fisheries (No. 3412-07-02074) and the Danish Strategic Research Council (No. 2104-08-0017) are acknowledged for financial support.

(1) Schiffman, S. S.; Studwell, C. E.; Landerman, L. R.; Berman, K.; Sundy, J. S. Symptomatic effects of exposure to diluted air sampled from a swine confinement atmosphere on healthy human subjects. Environ. Health Perspect. 2005, 113 (5), 567– 576. (2) Aneja, V. P.; Schlesinger, W. H.; Erisman, J. W. Effects of Agriculture upon the Air Quality and Climate: Research, Policy, and Regulations. Environ. Sci. Technol. 2009, 43 (12), 4234– 4240. (3) Zahn, J. A.; Hatfield, J. L.; Do, Y. S.; DiSpirito, A. A.; Laird, D. A.; Pfeiffer, R. L. Characterization of volatile organic emissions and wastes from a swine production facility. J. Environ. Qual. 1997, 26 (6), 1687–1696. (4) Schiffman, S. S.; Bennett, J. L.; Raymer, J. H. Quantification of odors and odorants from swine operations in North Carolina. Agric. Forest Meteorol. 2001, 108 (3), 213–240. (5) Wright, D. W.; Eaton, D. K.; Nielsen, L. T.; Kuhrt, F. W.; Koziel, J. A.; Spinhirne, J. P.; Parker, D. B. Multidimensional gas chromatography-olfactometry for the identification and prioritization of malodors from confined animal feeding operations. J. Agric. Food. Chem. 2005, 53 (22), 8663–8672. (6) Bulliner, E. A.; Koziel, J. A.; Cai, L. S.; Wright, D. Characterization of livestock odors using steel plates, solid-phase microextraction, and multidimensional gas chromatography-mass spectrometryolfactometry. J. Air Waste Man. Assoc. 2006, 56 (10), 1391–1403. (7) Trabue, S. L.; Scoggin, K. D.; Li, H.; Burns, R.; Xin, H. W. Field sampling method for quantifying odorants in humid environments. Environ. Sci. Technol. 2008, 42 (10), 3745–3750. (8) Trabue, S.; Scoggin, K.; Mitloehner, F.; Li, H.; Burns, R.; Xin, H. W. Field sampling method for quantifying volatile sulfur compounds from animal feeding operations. Atmos. Environ. 2008, 42 (14), 3332–3341. (9) Hobbs, P. J.; Webb, J.; Mottram, T. T.; Grant, B.; Misselbrook, T. M. Emissions of volatile organic compounds originating from UK livestock agriculture. J. Sci. Food Agric. 2004, 84 (11), 1414– 1420. (10) Lestremau, F.; Andersson, F. A. T.; Desauziers, V. Investigation of artefact formation during analysis of volatile sulphur compounds using solid phase microextraction (SPME). Chromatographia 2004, 59 (9-10), 607–613. (11) Van Gemert, L. J. Compilations of Odour Threshold Values in Air, Water and Other Media; Boelens Aroma Chemical Information Service: Huizen, The Netherlands, 2003. (12) Feilberg, A.; Adamsen, A. P. S.; Lindholst, S.; Lyngbye, M.; Scha¨fer, A. Evaluation of Biological Air Filters for Livestock Ventilation Air by Membrane Inlet Mass Spectrometry. J. Environ. Qual 2010, 39, 1085–1096. (13) Shaw, S. L.; Mitloehner, F. M.; Jackson, W.; Depeters, E. J.; Fadel, J. G.; Robinson, P. H.; Holzinger, R.; Goldstein, A. H. Volatile organic compound emissions from dairy cows and their waste as measured by proton-transfer-reaction mass spectrometry. Environ. Sci. Technol. 2007, 41 (4), 1310–1316. (14) Ngwabie, N. M.; Schade, G. W.; Custer, T. G.; Linke, S.; Hinz, T. Abundances and flux estimates of volatile organic compounds from a dairy cowshed in Germany. J. Environ. Qual. 2008, 37 (2), 565–573. (15) Lindinger, W.; Hansel, A.; Jordan, A. Proton-transfer-reaction mass spectrometry (PTR-MS): on-line monitoring of volatile organic compounds at pptv levels. Chem. Soc. Rev. 1998, 27 (5), 347–354. (16) de Gouw, J.; Warneke, C. Measurements of volatile organic compounds in the earths atmosphere using proton-transferreaction mass spectrometry. Mass Spectrom. Rev. 2007, 26 (2), 223–257. VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5899

(17) Hewitt, C. N.; Hayward, S.; Tani, A. The application of proton transfer reaction-mass spectrometry (PTR-MS) to the monitoring and analysis of volatile organic compounds in the atmosphere. J. Environ. Monit. 2003, 5 (1), 1–7. (18) Beauchamp, J.; Frasnelli, J.; Buettner, A.; Scheibe, M.; Hansel, A.; Hummel, T. Characterization of an olfactometer by protontransfer-reaction mass spectrometry. Meas. Sci. Technol 2010, 21 (2), 025801. (19) Biasioli, F.; Gasperi, F.; Odorizzi, G.; Aprea, E.; Mott, D.; Marini, F.; Autiero, G.; Rotondo, G.; Mark, T. D. PTR-MS monitoring of odour emissions from composting plants. Int. J. Mass Spectrom. 2004, 239 (2-3), 103–109. (20) Biasioli, F.; Aprea, E.; Gasperi, F.; Mark, T. D. Measuring odour emission and biofilter efficiency in composting plants by proton transfer reaction-mass spectrometry. Water Sci. Technol. 2009, 59 (7), 1263–1269. (21) Mayrhofer, S.; Mikoviny, T.; Waldhuber, S.; Wagner, A. O.; Innerebner, G.; Franke-Whittle, I. H.; Mark, T. D.; Hansel, A.; Insam, H. Microbial community related to volatile organic compound (VOC) mission in household biowaste. Environ. Microbiol. 2006, 8 (11), 1960–1974. (22) Blake, R. S.; Monks, P. S.; Ellis, A. M. Proton-Transfer Reaction Mass Spectrometry. Chem. Rev. 2009, 109 (3), 861–896. (23) Su, T.; Chesnavich, W. J. Parametrization of the Ion-Polar Molecule Collision Rate-Constant by Trajectory Calculations. J. Chem. Phys. 1982, 76 (10), 5183–5185. (24) Handbook of Chemistry and Physics, 79th ed.; Lide, D. R., Ed.; CRC Press: Boca Raton, FL, 1999; p 79. (25) National Institute of Standards and Technology. Computational Chemistry Comparison and Benchmark DataBase. 2010. http:// cccbdb.nist.gov/ (accessed May 18, 2009). (26) Zhao, J.; Zhang, R. Y. Proton transfer reaction rate constants between hydronium ion (H3O(+)) and volatile organic compounds. Atmos. Environ. 2004, 38 (14), 2177–2185. (27) Lyngbye, M.; Hansen, M. J.; Riis, A. L.; Jensen, T. L.; Sørensen, G. In 1000 Olfactometry Analyses and 100 TD-GC/MS Analyses to Evaluate Methods for Reducing Odour from Finishing Units in Denmark; Aneja, V. P., Schlesinger, W. H., Knighton, R., Jennings, G., Niyogi, D., Gilliam, W., Duke, C. S., Eds.; North Carolina State University: Raleigh, NC, 2006; pp 138-152. (28) Inomata, S.; Tanimoto, H.; Kameyama, S.; Tsunogai, U.; Irie, H.; Kanaya, Y.; Wang, Z. Technical Note: Determination of formaldehyde mixing ratios in air with PTR-MS: laboratory experiments and field measurements. Atmos. Chem. Phys 2008, 8 (2), 273–284. (29) Knighton, W. B.; Fortner, E. C.; Midey, A. J.; Viggiano, A. A.; Herndon, S. C.; Wood, E. C.; Kolb, C. E. HCN detection with a proton transfer reaction mass spectrometer. Int. J. Mass Spectrom. 2009, 283 (1-3), 112–121.

5900

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

(30) Blunden, J.; Aneja, V. P.; Lonneman, W. A. Characterization of non-methane volatile organic compounds at swine facilities in eastern North Carolina. Atmos. Environ. 2005, 39 (36), 6707– 6718. (31) Willig, S.; Lacorn, M.; Claus, R. Development of a rapid and accurate method for the determination of key compounds of pig odor. J. Chromatogr., A 2004, 1038 (1-2), 11–18. (32) Warneke, C.; Kuczynski, J.; Hansel, A.; Jordan, A.; Vogel, W.; Lindinger, W. Proton transfer reaction mass spectrometry (PTRMS): Propanol in human breath. Int. J. Mass Spectrom. Ion Processes 1996, 154 (1-2), 61–70. (33) Aprea, E.; Biasioli, F.; Carlin, S.; Versini, G.; Mark, T. D.; Gasperi, F. Rapid white truffle headspace analysis by proton transfer reaction mass spectrometry and comparison with solid-phase microextraction coupled with gas chromatography/mass spectrometry. Rapid Commun. Mass Spectrom. 2007, 21 (16), 2564– 2572. (34) Kim, K. Y.; Ko, H. J.; Kim, H. T.; Kim, Y. S.; Roh, Y. M.; Lee, C. M.; Kim, H. S.; Kim, C. N. Sulfuric odorous compounds emitted from pig-feeding operations. Atmos. Environ. 2007, 41 (23), 4811– 4818. (35) Martensson, L.; Magnusson, M.; Shen, Y.; Jonsson, J. A. Air concentrations of volatile organic acids in confined animal buildings-determination with ion chromatography. Agric., Ecosyst. Environ. 1999, 75 (1-2), 101–108. (36) Montes, F.; Hafner, S.; Rotz, C. A. In Measuring emissions of volatile organic compounds from silage; Paper N0. 096184; ASABE: St. Joseph, MI, 2009. (37) Chung, M. Y.; Beene, M.; Ashkan, S.; Krauter, C.; Hasson, A. S. Evaluation of non-enteric sources of non-methane volatile organic compound (NMVOC) emissions from dairies. Atmos. Environ. 2010, 44 (6), 786–794. (38) Blanes-Vidal, V.; Hansen, M. N.; Pedersen, S.; Rom, H. B. Emissions of ammonia, methane and nitrous oxide from pig houses and slurry: Effects of rooting material, animal activity and ventilation flow. Agric., Ecosyst. Environ. 2008, 124 (3-4), 237–244. (39) Ye, Z.; Zhang, G.; Seo, I. H.; Kai, P.; Saha, C. K.; Wang, C.; Li, B. Airflow characteristics at the surface of manure in a storage pit affected by ventilation rate, floor slat opening, and headspace height. Biosystems Eng. 2009, 104 (1), 97–105. (40) Hobbs, P.; Mottram, T. New Directions: Significant contributions of dimethyl sulphide from livestock to the atmosphere. Atmos. Environ. 2000, 34 (21), 3649–3650. (41) Schwarzenbach, R.; Gschwend, P. M.; Imboden, D. Environmental Organic Chemistry, 2nd ed.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003.

ES100483S