CHARACTERIZATION OF VOLATILE ORGANIC COMPOUNDS
ON AIRBORNE DUST IN A SWINE FINISHING BARN
E. B. Razote, R. G. Maghirang, L. M. Seitz, I. J. Jeon
ABSTRACT. Three methods of extracting volatile organic compounds (VOCs) adsorbed on the airborne dust in a swine finishing
building were investigated: solvent extraction using dichloromethane, solid−phase microextraction (SPME) using carboxen/
polydimethylsiloxane (CAR/PDMS) and PDMS fibers, and purge and trap. Airborne dust was first collected in pre−baked
glass−fiber filters and analyzed using each of the three methods. Solvent extraction with dichloromethane extracted only some
high−boiling point carboxylic acids. The SPME CAR/PDMS fiber extracted the low− to mid−boiling point VOCs such as the
carboxylic acids, phenols, and indoles; while the PDMS fiber extracted more of the mid−boiling point compounds, specifically
the aliphatic hydrocarbons, indoles, and some aldehydes. The purge and trap method extracted compounds with low− to mid−
boiling points including volatile carboxylic acids, aldehydes, alcohols, ketones, indoles, and esters. Quantitative analysis
of five selected VOCs (i.e., acetic acid, propionic acid, butyric acid, hexanal, and nonanal) using the purge and trap method
showed acetic acid as generally the most abundant and nonanal as the least abundant.
Keywords. Airborne dust, Purge and trap, Solvent extraction, SPME, Swine.
evelopment of appropriate systems to control the Previous studies have also indicated that dust can
odor emanating from livestock facilities requires transport and amplify the odor from livestock operations
knowledge of the odor’s major components. Pre- (Hammond et al., 1979; Takai et al., 1998). Dust−borne odors
vious studies have focused mainly on characteriz- can be transported over long distances where they can be
ing odorants coming from the manure, as well as the air in and perceived as a nuisance. However, limited research has
around these facilities. O’Neill and Phillips (1992) listed characterized the odorous compounds that are adsorbed on
168 compounds associated with the odor from livestock op- the dust in livestock buildings (Hammond et al., 1979;
erations. Schiffman et al. (2001) identified a total of 331 dif- Hammond et al., 1981; Hartung, 1985; Oehrl et al., 2001).
ferent compounds from swine facilities in North Carolina; Characterization of the dust−borne compounds involves
the compounds included many acids, alcohols, aldehydes, dust sampling and collection, extraction of the compounds
amides, amines, aromatics, esters, ethers, fixed gases, halo- from the dust, and identification and/or quantification of the
genated hydrocarbons, hydrocarbons, ketones, nitriles, other compounds using gas chromatograph−mass spectrometry
nitrogen−containing compounds, phenols, sulfur−containing (GC−MS). Several methods can be used to extract the
compounds, and volatile steroids. According to Yu et al. adsorbed compounds from the dust in livestock buildings:
(1991), indole, p−cresol, phenol, skatole, volatile carboxylic solvent extraction, solid−phase microextraction (SPME),
acids (e.g., acetic, propionic, isobutyric, butyric, isovaleric, and purge and trap (P&T). Solvent extraction has served as
valeric, caproic, and heptanoic), and ammonia appeared to be the primary method in previous studies (Hammond et al.,
the most important constituents of odor from swine waste. 1979; Hammond et al., 1981; Hartung, 1985; Oehrl et al.,
Zahn et al. (1997) indicated that, based on their research and 2001). However, solvent extraction is time−consuming and
available odor threshold data, the C2 through C9 carboxylic may result in loss of volatile compounds during the extraction
acids represented the greatest threat to air quality because of and concentration process (Hartung, 1985; Zhang et al.,
their high experimental transport efficiencies, high airborne 1994).
concentrations, and low odor thresholds. Recently, SPME has been used to characterize airborne
compounds in livestock buildings (Zahn et al., 1997; Yo,
1999; Powers et al., 2000; Gralapp et al., 2001; Kim−Yang et
Article was submitted for review in April 2003; approved for
al., 2001; Razote et al., 2002). Various researchers have
publication by the Structures & Environment Division of ASAE in May discussed the theory and chemistry of SPME (Zhang et al.,
2004. Presented at the 2002 ASAE Annual Meeting as Paper No. 024162. 1994; Pawliszyn, 1997; Lord and Pawliszyn, 2000). The
The authors are Edna B. Razote, ASAE Member Engineer, Research SPME method is simple, rapid, and sensitive; it combines
Assistant, and Ronaldo G. Maghirang, ASAE Member Engineer, sampling, pre−concentration, and direct transfer of com-
Professor, Department of Biological and Agricultural Engineering, Kansas
State University, Manhattan Kansas; Larry M. Seitz, Research Chemist, pounds into the GC (Pawliszyn, 1997). A small amount of
USDA−ARS Grain Marketing and Production Research Center, sample can be used for analysis. Furthermore, the cost of
Manhattan, Kansas; and Ike J. Jeon, Professor, Institute of Food Science, using SPME is relatively low because only slight modifica-
Department of Animal Science and Industry, Kansas State University, tion to the GC injector is needed. About 50 to 100 analyses
Manhattan, Kansas. Corresponding author: Ronaldo G. Maghirang,
Department of Biological and Agricultural Engineering, Kansas State
can be done per fiber. However, aside from PDMS fiber,
University, 147 Seaton Hall, Manhattan, KS 66506; phone: 785−532− quantification using other types of fibers needs to be further
2908; fax: 785−532−5825; e−mail: firstname.lastname@example.org. studied. Quantification relies mostly on the use of models and
Transactions of the ASAE
Vol. 47(4): 1231−1238 2004 American Society of Agricultural Engineers ISSN 0001−2351 1231
careful calibration (e.g., constant temperature, sampling Eighty Four, Pa.) placed in open−face filter cassettes. A
time, pH, etc.) is crucial. sampling pump was used to pull in air at a flow rate of
Another method that has been used to extract volatile 2.3 m3/h for 4 h. After each sampling, the filters were placed
compounds is P&T. Compared to solvent extraction, the P&T in pre−weighed 10 mL vials and sealed using an aluminum
method is simpler and faster. This method involves place- cap and Teflon−covered septa, and then immediately trans-
ment of the sample in a purge vessel where the volatile ported to the laboratory for analysis. Average temperature
organic compounds (VOCs) are purged off by an inert gas and inside the building during sampling was 21.4°C, ranging
subsequently trapped on a solid sorbent. Heating of the from 20.2°C to 22.6°C. The samplers were run in triplicates,
sorbent desorbs the compounds, which are then carried to the one for each extraction method. For solvent extraction and
GC. This method has been applied in determining and SPME, all three replicates were analyzed; for P&T, only two
quantifying the VOCs associated with grains (Ram et al., replicates were analyzed because of instrument malfunctions
1999; Seitz et al., 1999), fish (Refsgaard et al., 1992; Santos for one of the replicates.
et al., 2001), and soil (Askari et al., 1996). Detection and The second experiment was conducted for quantitative
identification of most compounds can be done using a small analysis of selected VOCs using the P&T method. Airborne
amount of sample. However, the initial setup for this method dust was collected at an exhaust fan in the swine building
is more complex and costly since it requires a P&T unit, GC using pre−baked and pre−weighed 20 × 25 cm glass−fiber
modification to accept the P&T unit, high−purity gases for filters (type A/E, Pall Life Sciences, Ann Arbor, Mich.)
sample purging, and liquid nitrogen for cryofocusing the placed in a high−volume sampler (model 500, Bendix Corp.,
compounds at the top of the GC column. Lewisburg, W.V.). The sampler was operated at an airflow
Further research is needed to characterize the compounds rate of 1.13 m3/min for 2.5 h. Dust was collected on three
that are adsorbed on airborne dust in livestock buildings. This different days with 2 to 3 replicates for each day. Average
study was conducted to: (1) identify compounds adsorbed on temperature at the exhaust during the sampling was 26.8°C,
the airborne dust from a swine finishing building using three ranging from 26.0°C to 27.5°C. After sampling, each
extraction methods, namely solvent extraction, SPME, and collection filter was cut in three 3.8 cm diameter circles.
P&T; and (2) quantify the major compounds identified using These cut filters were placed in pre−weighed 10 mL vials and
the P&T method. sealed using an aluminum cap and Teflon−covered septa, and
then immediately transported to the laboratory for analysis.
MATERIALS AND METHODS FILTER CONDITIONING
The filters were baked at 400°C for 4 h in a muffle furnace
AIRBORNE DUST SAMPLING
Two sets of experiments were conducted: (1) qualitative prior to sampling to remove most of the VOCs from the
filters. After baking, the filters were quickly weighed and
analysis of the VOCs on the dust using three extraction
wrapped in baked aluminum foil prior to transport to the barn.
methods, and (2) quantitative analysis of preselected VOCs
on the dust using P&T. All airborne dust samples were To check the effectiveness of this conditioning procedure,
unbaked and freshly baked filters were analyzed for presence
collected from a swine finishing barn at the Kansas State
of compounds using SPME 100 mm polydimethysiloxane
University Swine Teaching and Research Unit (Manhattan,
Kansas). The barn was 34 m long, 12 m wide, and 2.5 m high, (PDMS) and 75 mm carboxen (CAR)/PDMS fibers (Supelco,
Bellfonte, Pa.). The procedure for SPME extraction, de-
with 80 pens arranged in four rows over fully slatted floors.
scribed later in a separate section, was followed. Comparison
Each pen (1.62 × 1.62 m) had an automatic self−feeder and
waterer and held two pigs during the study. Ground feed with of the ion chromatograms for the two filters indicated that
conditioning reduced the number of compounds from at least
5% added fat was distributed through overhead augers to the
34 (fig. 1a) to approximately six (fig. 1b) using the
feeders four times a day. The alley floor was cleaned at least
once a week. Manure was collected in two deep pits under the CAR/PDMS fiber. Furthermore, the amount of these six
compounds in the baked filter was negligible compared to the
pen rows and removed through a pull−plug system after each
unbaked filter. No compounds were detected after the
finished batch. Ventilation air entered through 21 sidewall
inlets (0.53 m wide each) distributed along the two sidewalls, conditioning procedure using the PDMS fiber.
passed through the two underfloor pits running longitudinally
under the pens, and exhausted by three 0.61 m exhaust fans ANALYTICAL METHODS
at one end of the building. Two 51.3 kW gas heaters located Solvent Extraction
at the middle of the room provided supplemental heat. The Dichloromethane has been successfully used as a solvent
first−stage (minimum ventilation) fan was always in opera- to extract VOCs concentrated on fabric swatches and in
tion. The initial weight of the pigs when they were brought lagoon samples (Schiffman et al., 2001), as well as indole and
into the barn was about 25 to 35 kg each. They remained in skatole on dust samples (Travis and Elliott, 1977). For this
the barn for about 15 to 17 weeks until they reached a market study, dichloromethane (analytical grade; Fischer Scientific,
weight of about 110 to 125 kg. During the course of sampling St. Louis, Mo.) was used to extract VOCs from the airborne
there was a change in the batch of pigs, which might have dust. Based on a preliminary test, 5 mL of dichloromethane
caused some of the variability in the results, as noted later. was added to the 10 mL vial containing the sample. The vial
The first experiment used the solvent extraction, SPME, was sealed with an aluminum cap and Teflon−covered
and P&T methods to determine the VOCs that were adsorbed septum, and was shaken for 10 min in a shaker (Eberbach
on the dust samples. Airborne dust was collected from the Corporation, Ann Arbor, Mich.) with a speed setting of 100.
center of the building, 1.5 m from the floor using pre−baked Each filter was extracted three times. Each time, the extract
and pre−weighed 37 mm glass−fiber filters (SKC, Inc., was allowed to settle and was filtered using Whatman No. 1
1232 TRANSACTIONS OF THE ASAE
5 10 15 20 25
Figure 1. Total ion chromatograms of filters (a) before and (b) after the filter conditioning procedure. Filters were analyzed using SPME (CAR/PDMS)
filter paper. To concentrate down to 0.5 mL, the total extract purge was performed to remove excess moisture from the
was placed in a water bath at 40°C under a constant stream trap. The trap was preheated at 220°C, and then the volatiles
of nitrogen (ultra−high purity grade) at 50 mL/min flow rate. were desorbed from the trap at 225°C for 6 min. With the
Five mL of the concentrate was directly injected to the capillary interface module, the desorbed volatiles were
GC−MS for analysis. cryofocused at −140°C (liquid N2). The cryofocused zone
was heated at 200°C for 0.85 min before the start of the
analytical run. The temperature of the injector zone under the
Two types of fibers were used for the SPME extraction: capillary interface was maintained at 200°C.
100 mm PDMS and 75 mm CAR/PDMS. Adsorbent−type
As mentioned earlier, the P&T method was also used in
fibers like CAR/PDMS are best used for extracting low
the second experiment to quantify selected VOCs. Five
molecular weight compounds (<200 amu), while absorbent−
compounds that were observed to be present in all of the
type fibers like PDMS are more efficient in extracting higher
samples and/or with large peak areas from the first experi-
molecular weight compounds (>200 amu) (Shirey, 2000a,
ment were quantified. These compounds included three
2000b). The SPME fibers were conditioned as recommended
carboxylic acids (acetic, propionic, and butyric) and two
by the manufacturer prior to first use. The PDMS fiber was
aldehydes (hexanal and nonanal). The three acids have been
conditioned at 250°C for 60 to 90 min, while the CAR/PDMS
reported to be major contributors to the swine odor (Ham-
fiber was conditioned at 280°C for 30 to 60 min. Based on a
mond et al., 1979; Yu et al., 1991), while the two aldehydes,
preliminary test, the following protocol was adopted to
although less odorous by themselves, may contribute to the
extract the VOCs: (1) place the 10 mL vial containing the
characteristic swine odor when mixed with the other
sample in a water bath at 80°C, (2) pierce the septum using
compounds. Quantifying target ions were identified for each
the SPME needle and expose the SPME fiber to the
compound. One mL of ethylbenzene−d10 in methanol (50 ng/
headspace for 30 min, and (3) immediately inject the SPME
mL) was chosen as an internal standard to the samples. The
fiber into the GC for analysis.
target mass of m/e 116 selected for quantitative detection of
Purge and Trap ethylbenzene−d10 had no interference from other com-
The filter was placed in a U−shaped sparge tube attached pounds, and the compound itself does not occur naturally. A
to a P&T instrument (model G1901−60500, Hewlett−Pack- calibration plot consisting of three to five data points was first
ard, Palo Alto, Cal.) equipped with a sample pocket heater obtained for each compound (Razote, 2003). To correct for
(model 14−5737−020, Hewlett−Packard, Palo Alto, Cal.) and differences in purging and recovery among different runs, the
a capillary interface module (model G1908−60500, Hewlett− ratio of response of the target compound to that of the internal
Packard, Palo Alto, Cal.). The protocol for extracting and standard (i.e., ethylbenzene−d10 ) was plotted against the ratio
analyzing the compounds was based on the procedure of the concentration of the target compound to that of the
developed by Seitz et al. (1999) for analyzing volatiles from internal standard. A linear fit with zero intercept was used in
grain samples. Each filter was preheated to 80°C for 3 min, quantifying the compounds. The SAS General Linear Models
and then the volatiles from the heated filters were purged with (GLM) procedure and Least Squares Means (LSMeans)
helium at 40 mL/min onto a glass−lined Tenax trap (Type 1G, (SAS ver. 6.12, SAS Institute, Inc., Cary, N.C.) were
Tekmar) for 10 min. After the sample purge, an 8 min dry performed on concentration means to determine the variabil-
Vol. 47(4): 1231−1238 1233
ity in the amount of the compounds between sampling dates Tests on the recovery of 10 preselected standard compounds
and between compounds. with this method showed mean recoveries ranging from 67%
to 90% and losses ranging from 10% to 33%.
GAS CHROMATOGRAPHY AND MASS SPECTROMETRY A total of 38 different compounds were identified for the
Compounds extracted by solvent extraction and SPME two SPME fibers (table 1). Compounds extracted depended
methods were analyzed using an HP 5890A GC coupled with on the type of SPME fiber used. The PDMS fiber has a
an HP 5970 mass selective detector (MSD) (Hewlett−Pack- liquid−phase coating that works as an absorbent; rapid
ard, Palo Alto, Cal.). The column was a fused silica HP−5 diffusion of volatiles occurs in a liquid coating but the small
capillary column (30 m × 0.25 mm i.d. × 0.25 mm; Agilent analytes are not well retained (Shirey et al., 1998). On the
Technologies, Wilmington, Del.). The oven temperature was other hand, the CAR/PDMS fiber works both as adsorbent
programmed as follows: initial temperature of 40°C for (CAR) and absorbent (PDMS). The pores in carboxen create
2 min, then ramped to 200°C at 10°C/min and held for 5 min, a surface where the volatile analytes are physically trapped,
followed by a 10°C/min temperature increase to 250°C and resulting in better retention of smaller analytes (Shirey et al.,
held for 4 min. Injection port and MSD transfer line 1998).
temperatures were 250°C and 280°C, respectively. The In this study, the CAR/PDMS fiber was able to extract
carrier gas was helium at 1 mL/min. Desorption time was more of the low− to mid−boiling point compounds, particu-
2 min. The compounds were identified by their mass spectra larly the carboxylic acids. These results were similar to those
utilizing the probability−based matching software program obtained by Razote et al. (2002) in a related study of VOCs
with NIST Library Data Base (ver. 4.5, Agilent Technolo- in the air of the same swine barn. The PDMS fiber, on the
gies, Palo Alto, Cal.). Some of the compounds were also other hand, was able to extract more of the mid−boiling point
identified by their GC retention times. The above tempera- compounds, specifically the hydrocarbons and the alde-
ture program was based on a study by Kim (2002) on VOCs hydes. Other studies also observed this selective property of
in soymilk. the SPME fibers. Shirey (2000a) and Popp and Paschke
For the P&T method, the GC−MS was an HP 5890 series (1997) observed that the CAR/PDMS fiber had greater
II GC coupled with an HP 5971 MSD (Hewlett−Packard, Palo response in extracting VOCs compared to the other fibers,
Alto, Cal.). The column was BPX5 (50 m × 0.32 mm i.d. × including PDMS. Similarly, Abalos and Bayona (2000) were
0.25 mm; Scientific Glass Engineering, Austin, Texas). The able to extract the C2−C7 carboxylic acids in aqueous
oven temperature program was the same as that used in the samples using the CAR/PDMS fiber. In this study, 17 and
solvent extraction and SPME analysis. The transfer line 27 compounds were identified out of 30 and 49 compounds
temperature of the MSD was 280°C. The carrier gas was extracted by CAR/PDMS and PDMS fibers, respectively.
helium at a constant flow rate of approximately 1.0 mL/min. Hexanal, diethyl phthalate, p−cresol, indole, methyl−1H−
Compounds were identified by comparing their mass spectra indole, and skatole were extracted by both fibers. Most of the
with standard spectra in the HP59943B Wiley PBM MS compounds extracted by the SPME fibers had been reported
database. To verify the presence of indole and skatole in the to be present in swine facilities either on the dust, in the air,
sample, standards of these two compounds were injected in and/or in the manure/lagoon (table 1). Compounds that have
the U−shaped sparge tube and their retention times were not been reported include octadecanoic acid; 2−butyl−2−oc-
noted. tenal; 2−heptadecanol; methyl−1H−indole; 2,6,10,14−tetra-
methylhexadecane; hexamethylcyclotrisiloxane; and octa−
methylcyclotetrasiloxane. The two siloxanes might be
coming from the GC column (Manura, 1995). Four acids
RESULTS AND DISCUSSION (acetic, propionic, butyric, and valeric acids) were picked up
QUALITATIVE ANALYSIS by the CAR/PDMS in all three replicates analyzed. These
A total of 84 different compounds were tentatively acids have been reported to be major contributors to swine
identified using the three extraction methods, most of which odor (Hammond et al., 1979; Yu et al., 1991; Zahn et al.,
have been reported to be present in dust, air, and/or manure 1997).
in previous studies (table 1). It was observed that not all The phthalates, which were only detected and identified
compounds identified were present in all replicates for each using the solvent extraction and SPME methods, might be
method. The samples/replicates for each method were coming from the septa since only these two methods used the
collected at different times over two batches of pigs. This septa and 10 mL vial during the process of extraction.
might account for the difference in the compounds detected Phthalate esters are used to soften septa, and these might have
between replicates for each method. been released during the process of shaking for solvent
With solvent extraction, only high−boiling point com- extraction and septum piercing for SPME fiber exposure to
pounds were extracted. Seven out of an average of 18 com- the headspace. Restek Corporation (2003), in their guide to
pounds detected by the GC−MS were tentatively identified minimizing septa problems, mentioned release of these
(table 1). From the seven compounds identified, only volatiles from septa in what is known as septum bleed.
hexadecanoic acid and dibutyl phthalate have been reported For the P&T method, 57 compounds were extracted and
to be present in a swine environment (Schiffman et al., 2001). tentatively identified (table 1). Forty of these compounds
The presence of small amounts of approximately 11 other were not seen using SPME, most of which were ketones,
VOCs, as indicated by the gas chromatograms, suggests that alcohols, aldehydes, ethers, and other compounds. In addi-
solvent extraction requires a large amount of dust sample tion, 21 of these compounds have not been previously
(>5 mg) for better detection and identification of com- reported, including 4−heptanone and 5−methyl−3−hepta-
pounds. Furthermore, some of the more volatile compounds none, which were reported to be irritants, and 2−methylpro-
were lost during the process of extraction and concentration. panal, which has a pungent odor. Indole and skatole, which
1234 TRANSACTIONS OF THE ASAE
Table 1. Frequency of occurrence of the compounds extracted on the airborne dust from inside a swine finishing
building using solvent extraction (SE), SPME, and P&T methods and identified by GC−MS from this work.
Frequency of Occurrence
Compounds SE[a] PDMS CAR/PDMS P&T[b] Dust Air Manure/Lagoon
Acetic acid 3 1 ii, iii, iv v, vi, vii, viii, xi v, vi, vii, x, xii, xiii, xiv
Propionic acid 3 1 ii, iii, iv v, vi, vii, viii, xi v, vi, vii, xii, xiii, xiv
Isobutyric acid 2 iii, iv v, vi, viii, xi v, vi, vii, xii, xiii, xiv
Butyric acid 3 1 ii, iii, iv v, vi, vii, viii, xi v, vi, vii, xii, xiii, xiv
Isovaleric acid 1 1 iii, iv v, vi, viii, xi vi, vii, x, xii, xiii, xiv
Valeric acid 3 1 ii, iii, iv v, vi, vii, viii, xi v, vi, vii, xii, xiii, xiv
Hexadecanoic acid 2 v
9−Hexadecenoic acid 1
Octadecanoic acid 2 1
9−Octadecenoic acid 1
9,12−Octadecadienoic acid 2
2−Methylbutanoic acid 1 v vi, x
Benzoic acid 1 v, vi, vii v, vi, vii
Acetone 1 ii v, vi v
2−Heptanone 1 v v
2−Octanone 2 v, vi v, vi
3−Octanone 1 v vii
2−Nonanone 1 v
2−Decanone 1 v
2−Nonadecanone 1 v
Pentanal 2 ii v, vi v, vi, vii, xii
Hexanal 1 1 2 i, ii v, vi v, vi
Heptanal 1 2 i, ii v, vi v, vi
Octanal 2 v, vi v, vi
Nonanal 2 2 i v v
2−Nonenal 2 2 i v
2−Heptenal 2 ii
Decanal 2 1 ii v, vi v, vi
2,4 Nonadienal 2 i, ii
2−Butyl−2−octenal 2 1
Benzaldehyde 2 ii v, vi, ix v, vi
1−Pentanol 2 v xiii
1−Octanol 1 v
1−Hexadecanol 2 v
Vol. 47(4): 1231−1238 1235
Table 1 (cont’d). Frequency of occurrence of the compounds extracted on the airborne dust from inside a swine finishing
building using solvent extraction (SE), SPME, and P&T methods and identified by GC−MS from this work.
Frequency of Occurrence
Compounds SE[a] PDMS CAR/PDMS P&T[b] Dust Air Manure/Lagoon
Methyl butyrate 1
Diethyl phthalate 3 3 v v
Dibutyl phthalate 2 v v
2−Ethylhexyl acetate 1
2−Ethylhexyl butyrate 1 v
Diisobutyl phthalate 2
Phenol 1 i, iii, iv v, vi, vii, viii, xi v, vi, vii, x, xii, xiii, xiv
p−Cresol 1 1 1 i, ii, iii, iv v, vi, vii, viii, xi vi, vii, x, xii, xiii, xiv, xv
4−Ethylphenol 1 v, vi, vii, viii, xi v, vi, vii, x, xiii
Indole 1 1 1 iii, iv v, vi, viii, xi v, vi, vii, x, xii, xiii, xiv, xv
Methyl −1H−indole 1 1
Skatole 1 1 1 i, iii, iv vi, viii, xi v, vi, vii, x, xii, xiii, xiv, xv
Benzothiazole 2 1 xv
Heptane 1 v v
Dodecane 1 v
Tridecane 1 v
Tetradecane 2 1 v v
Pentadecane 1 1 v
Hexadecane 1 v v
Heptadecane 2 v v
Octadecane 2 v v
Nonadecane 1 v
Eicosane 1 v v
Methylene chloride 2 v v
Dimethyldisulfide 1 v, vi, viii, ix v, vi, vii, xiii, xv
2−Methylfuran 1 v, vi v
2−Pentylfuran 2 i v, ix v
Diethyl ether 1
Limonene 1 ix
[a] Out of three replicates.
[b] Out of two replicates.
[c] i = Hammond et al., 1979; ii = Hammond et al., 1981; iii = Hartung, 1985; iv = Oehrl et al., 2001; v = Schiffman et al., 2001; vi = Spoelstra, 1980; vii
= Zahn et al., 1997; viii = Zahn et al., 2001; ix = Kim−Yang et al., 2001; x = Hobbs et al., 1995; xi = Gralapp et al., 2001; xii = Schaefer, 1977; xiii =
Yasuhara et al., 1984; xiv = Yu et al., 1991; and xv = Hammond et al., 1989.
were identified by their retention times and by molecular ion lists only the compounds that were tentatively identified.
extraction, had smaller peaks compared to the other com- Numerous other compounds were extracted, although the
pounds, suggesting that their concentration in the sample amounts of most of those compounds were too low for proper
might be low. identification by GC−MS.
Comparison of the three methods suggests that the P&T
method extracted more of the odorous compounds compared QUANTITATIVE ANALYSIS
to solvent extraction and SPME. However, it appears that a The mean concentration of acetic acid on the airborne dust
combination of methods should be used to extract most of the was significantly higher than that of the other compounds
compounds from airborne dust. It should be noted that table 1 tested (p < 0.05), and nonanal had the lowest concentration
1236 TRANSACTIONS OF THE ASAE
(table 2). Acetic acid also had the highest variability on all S Solvent extraction with dichloromethane lacked the
sampling dates except for June, when the concentration of the sensitivity needed for detection and identification of
propionic acid measured was more variable. Significant (p < the more volatile compounds, as it extracted the least
0.05) differences in concentration were also observed number of compounds. In addition, 10% to 33% of the
between sampling times for all compounds except for butyric compounds were lost in this method, perhaps due to ex-
acid and nonanal. Such variability in concentration between traction and concentration.
sampling times is expected and was also observed by S Most of the compounds reported to be present in the air
Hammond et al. (1981) and Hartung (1985) in their and manure/lagoon of swine barns including volatile
quantitative analyses of odorous compounds from swine carboxylic acids, aldehydes, alcohols, ketones, hydro-
house airborne and settled dust, respectively. The age and carbons, phenols, indoles, other nitrogen−containing
diet of swine, waste handling, and the temperature and compounds, sulfur−containing compound, ethers, es-
humidity in the building, among other parameters, could ters, and other compounds were also found in the air-
account for such variability. Measured concentration of the borne dust.
acids from this study are much lower than those reported by S Quantitative analysis of five compounds using the
Hartung (1985) and Oehrl et al. (2001) in their analyses of purge and trap method showed that acetic acid was the
settled dust. Differences of 148 and 732 mg/g for acetic acid, most abundant (72 mg/g), and nonanal was the least
99 and 143 mg/g for propionic acid, and 52 and 120 mg/g for abundant (5 mg/g).
butyric acid were calculated from those reported by Hartung
(1985) and Oehrl et al. (2001), respectively. On the other ACKNOWLEDGEMENTS
hand, Hammond et al. (1981) reported much lower con- This study was supported in part by the National Science
centrations of these acids in their analysis of airborne dust. Foundation (Grant No. EPS−0082800), the Kansas Center for
Differences of 86, 30, 11, and 35mg/g for acetic, propionic, Agricultural Resources and the Environment, and the Kansas
and butyric acids and hexanal, respectively, were calculated Agricultural Experiment Station (Contribution No. 03−34−
based on the highest concentration measured for each J). We would also like to acknowledge Dr. Ram and
compound for both studies. The differences in the amount Dr. Howard of the USDA−ARS−GMPRC and Dr. Erickson
reported could be attributed to differences in sampling of the Chemical Engineering Department, Kansas State
methods and analysis and differences in test conditions, University, for their valuable input.
including those mentioned earlier. For example, the above
studies collected settled dust by free settling (Hartung, 1985)
and by brushing from exhaust fans (Oehrl et al., 2001), while
Hammond et al. (1981) analyzed airborne dust collected by REFERENCES
Abalos, M., and J. M. Bayona. 2000. Application of gas
an electrostatic precipitator. Furthermore, the sampling
chromatography coupled to chemical ionization mass
system used in this study did not account for possible errors spectrometry following headspace solid−phase microextraction
associated with sorption/desorption of compounds on the for the determination of free volatile fatty acids in aqueous
filter and collected dust. As such, measured concentration of samples. J. Chromatogr. A 891: 287−294.
the five compounds might have been overestimated. Further Askari, M. D. F., M. P. Maskarinec, S. M. Smith, P. M. Bean, and C.
improvement on the sampling and analysis system is needed. C. Travis. 1996. Effectiveness of the purge−and−trap for
measurement of volatile organic compounds in aged soils. Anal.
Chem. 68(19): 3431−3433.
Gralapp, A, W. Powers, and D. Bundy. 2001. Comparison of
CONCLUSIONS olfactometry, gas chromatography, and electronic nose
The following conclusions were drawn from this study: technology for measurement of indoor air from swine facilities.
S Eighty−four compounds were identified by the three Trans. ASAE 44(5): 1283−1290.
methods as being present in the airborne dust from the Hammond, E. G., C. Fedler, and G. Junk. 1979. Identification of
swine facility. The SPME and purge and trap methods dust−borne odors in swine confinement facilities. Trans. ASAE
extracted 38 and 57 compounds, respectively, having 22(5): 1186−1189.
Hammond, E. G., C. Fedler, and R. J. Smith. 1981. Analysis of
low− to mid−boiling points.
particle−borne swine house odor. Agric. Environ. 6(4): 395−401.
Hammond, E. G., C. Heppner, and R. Smith. 1989. Odors of swine
Table 2. Amounts of five major compounds identified in the airborne waste lagoons. Agric. Ecosyst. Environ. 25(2−3): 103−110.
dust from the exhaust fan of a swine finishing barn using purge and
trap method. Values in parentheses are the standard deviations.
Hartung, J. 1985. Gas chromatographic analysis of volatile fatty
acids and phenolic/indolic compounds in pig house dust after
Concentration [a] (µg/g dust)
ethanolic extraction. Environ. Tech. Letters 6(1): 21−30.
Compound June 20[b] July 11[c] August 21[b] Overall Mean[d] Hobbs, P., T. Misselbrook, and B. Pain. 1995. Assessment of odours
Acetic acid 119 (10) x 22 (8) y 98 (12) x 72 (48) * from livestock wastes by a photoionization detector, an
Propionic acid 41 (19) x 8 (2) y 24 (4) x,y 22 (17) + electronic nose, olfactometry, and gas chromatography−mass
Butyric acid 21 (18) x 4 (3) y 19 (4) x 13 (12) + spectrometry. J. Agric. Eng. Res. 60(2): 137−144.
Hexanal 35 (1) x 35 (2) x 12 (1) y 29 (11) + Kim, H. 2002. Binding characteristics of cyclodextrins in model
Nonanal 3 (2) x 7 (3) x 5 (1) x 5 (3) + systems and their effectiveness on entrapping beany flavor
compounds in soymilk. PhD diss. Manhattan, Kansas: Kansas
Means with the same letter in the same row were not significantly differ-
ent at the 5% level.
[b] Average of two replicates. Kim−Yang, H., S. Davies, R. D. von Bernuth, and E. A. Kline.
[c] Average of three replicates. 2001. A comparison of sampling methods for the
[d] Means with the same symbol were not significantly different at the 5% characterization of odorous compounds in livestock facilities
Vol. 47(4): 1231−1238 1237
using gas chromatography−mass spectrometry. ASAE Paper No. Schaefer, J. 1977. Sampling, characterization, and analysis of
014037. St. Joseph, Mich.: ASAE. malodours. Agric. Environ. 3: 121−127.
Lord, H., and J. Pawliszyn. 2000. Evolution of solid−phase Schiffman, S. S., J. L. Bennett, and J. H. Raymer. 2001.
microextraction technology. J. Chromatogr. A. 885: 153−193. Quantification of odors and odorants from swine operations in
Manura, J. J. 1995. Elimination of “memory” peaks in thermal North Carolina. Agric. Forest Meteor. 108(3): 213−240.
desorption. SISWEB Application Note, Technical Bulletin E. Seitz, L. M., M. S. Ram, and R. Rengarajan. 1999. Volatiles
Available at: www.sisweb.com/referenc/articles/gcback.htm. obtained from whole and ground grain samples by supercritical
Accessed 20 March 2003. carbon dioxide and direct helium purge methods: Observations
Oehrl, L. L., K. M. Keener, R. W. Bottcher, R. D. Munilla, and K. on 2,3−butanediols and halogenated anisoles. J. Agric. Food
M. Connelly. 2001. Characterization of odor components from Chem. 47(3): 1051−1061.
swine housing dust using gas chromatography. Applied Eng. in Shirey, R. E. 2000a. Optimization of extraction conditions for
Agric. 17(5): 659−661. low−molecular−weight analytes using solid−phase
O’Neill, D. H., and V. R. Phillips. 1992. A review of the control of microextraction. J. Chromatogr. Sci. 38(3): 109−116.
odour nuisance from livestock buildings: Part 3. properties of Shirey, R. E. 2000b. Optimization of extraction conditions and fiber
the odorous substances which have been identified in livestock selection for semi−volatile analytes using solid−phase
wastes or in the air around them. J. Agric. Eng. Res. 53(1): microextraction. J. Chromatogr. Sci. 38(7):270−288.
23−50. Shirey, R. E., V. Mani, and R. Mindrup. 1998. On−site sampling for
Pawliszyn, J. 1997. Applications. In Solid Phase Microextraction: volatiles and pesticides using solid−phase microextraction.
Theory and Practice, 161−180. New York, N.Y.: Wiley−VCH. American Environ. Lab. 10: 21−22.
Popp, P., and A. Paschke. 1997. Solid phase microextraction of Spoelstra, S. 1980. Origin of objectionable odorous components in
volatile organic compounds using piggery wastes and the possibility of applying indicator
carboxen−polydimethylsiloxane fibers. Chromatographia components for studying odour development. Agric. Environ.
46(7−8): 419−424. 5(3): 241−260.
Powers, W., T. van Kempen, D. Bundy, A. Sutton, and S. Hoff. Takai, H., S. Pedersen, J. O. Johnsen, J. H. M. Metz, P. W. G. Groot
2000. Objective measurement of odors using gas Koerkamp, G. H. Uenk, V. R. Phillips, M. R. Holden, R. W.
chromatography/mass spectrometry and instrumental Sneath, J. L. Short, R. P. White, J. Hartung, J. Seedorf, M.
technologies. In Air Pollution from Agricultural Operations: Schroder, K. H. Linkert, and C. M. Wathes. 1998.
Proc. 2nd International Conference, 163−169. St. Joseph, Concentrations and emissions of airborne dust in livestock
Mich.: ASAE. buildings in Northern Europe. J. Agric. Eng. Res. 70(1): 59−77.
Ram, M. S., L. M. Seitz, and R. Rengarajan. 1999. Use of an Travis, T., and L. Elliott. 1977. Quantitation of indole and skatole in
autosampler for dynamic headspace extraction of volatile a housed swine unit. J. Environ. Qual. 6(4): 407−410.
compounds from grains and effect of added water on the Yasuhara, A., K. Fuwa, and M. Jimbu. 1984. Identification of
extraction. J. Agric. Food Chem. 47(10): 4202−4208. odorous compounds in fresh and rotten swine manure. Agric.
Razote, E. 2003. Gas chromatographic analysis of gaseous and Biol. Chem. 48(12): 3001−3010.
dust−borne volatile organic compounds in a swine finishing Yo, S. 1999. Analysis of volatile fatty acids in wastewater collected
building. MS thesis. Manhattan, Kans.: Kansas State University. from a pig farm by a solid phase microextraction method.
Razote, E., I. Jeon, and R. Maghirang. 2002. Dynamic air sampling Chemosphere 38(4): 823−834.
of volatile organic compounds using solid phase Yu, J. C., C. E. Isaac, R. N. Coleman, J. J. R. Feddes, and B. S.
microextraction. J. Environ. Sci. Health B 37(4): 365−378. West. 1991. Odorous compounds form treated pig manure.
Refsgaard, H., A. Haahr, and B. Jensen. 1992. Isolation and Canadian Agric. Eng. 33(1): 131−136.
quantification of volatiles in fish by dynamic headspace Zahn, J. A., J. L. Hatfield, Y. S. Do, A. A. Dispirito, D. A. Laird,
sampling and mass spectrometry. J. Agric. Food Chem. 47(3): and R. L. Pfeiffer. 1997. Characterization of volatile organic
114−118. emissions and wastes from swine production facility. J. Environ.
Restek Corporation. 2003. Guide to minimizing septa problems. Qual. 26(6): 1687−1696.
Bellefonte, Pa.: Restek Corporation. Available at: Zahn, J. A., A. A. Dispirito, Y. S. Do, B. E. Brooks, E. E. Cooper,
www.chromtech.net.au/pdf/rtxsepta. PDF. Accessed 20 March and J. L. Hatfield. 2001. Correlation of human olfactory
2003. responses to airborne concentrations of malodorous volatile
Santos, G. V., M. C. da Cunha Veloso, P. A. de Paula Pereira, and J. organic compounds emitted from swine effluent. J. Environ.
B. de Andrade. 2001. Fish off−flavor analysis by headspace and Qual. 30(2): 624−634.
off−line purge−and−trap followed by HRGC−MS. American Zhang, Z., M. J. Yang, and J. Pawliszyn. 1994. Solid−phase
Lab. 33(24): 28−30. microextraction. Anal. Chem. 66(17): 844A−853A.
1238 TRANSACTIONS OF THE ASAE