JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 27 October 2006
Published in J Environ Qual 35:2383-2394 (2006)
DOI: 10.2134/jeq2006.0065
© 2006 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Miller, D. N.
Right arrow Articles by Woodbury, B. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Miller, D. N.
Right arrow Articles by Woodbury, B. L.
Agricola
Right arrow Articles by Miller, D. N.
Right arrow Articles by Woodbury, B. L.
Related Collections
Right arrow Volatile Organic Compounds
Right arrow Air Pollution
Right arrow Animal Waste

TECHNICAL REPORTS

A Solid-Phase Microextraction Chamber Method for Analysis of Manure Volatiles

Daniel N. Miller* and Bryan L. Woodbury

USDA-ARS, U.S. Meat Animal Research Center, P.O. Box 166, Clay Center, NE 68933. D.N. Miller, current address: USDA, ARS, Soil and Water Conservation Research Unit, 121 Keim Hall, East Campus, Lincoln, NE 68583

* Corresponding author (dmiller15{at}unl.edu)

Received for publication February 15, 2006.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Odors from livestock operations are a complex mixture of volatile carbon, sulfur, and nitrogen compounds. Currently, detailed volatiles analysis is both time consuming and requires specialized equipment and methods. This work describes a new method that utilizes a dynamic flux chamber, solid-phase microextraction (SPME), and gas chromatography–mass spectroscopy (GC–MS) to describe and compare the odorous compounds emitted from cattle and swine feces. Evaluation of method parameters produced a protocol for comparing relative emissions based on fixed sample temperature (20°C) and exposed surface area (approximately 523 cm2), air flow rates (1 L min–1 or 16 cm s–1), SPME exposure time (5 min), and chamber cleaning procedures (70% ethanol rinse and drying for 30 min at 105°C) to minimize cross-contamination between samples. A variety of volatile organic compounds (VOCs) including alcohols, volatile fatty acids, aromatic ring compounds, ketones, esters, and sulfides were routinely detected and the relative emissions from fresh and incubated (37°C overnight) swine and cattle feces were compared as a measure of potential to produce odorants during manure storage. Differences in the types and relative quantities of volatiles emitted were detected when animal species (cattle or swine), diet, fecal incubation, or sample storage conditions (20, 4, or –20°C) were varied.

Abbreviations: DM, dry matter • GC, gas chromatography • IC, ion current • MS, mass spectroscopy • OM, organic matter • SPME, solid-phase microextraction • VFA, volatile fatty acids • VOC, volatile organic compounds


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
OVER 150 DIFFERENT CHEMICAL COMPOUNDS have been detected in the air emitted from livestock facilities (O'Neill and Phillips, 1992). Ascribing the human-perceived odor measures to specific volatile chemical constituents has been a challenge, but a growing body of work has emerged implicating volatile organic compounds (VOCs), in particular volatile fatty acids (VFAs) and aromatic ring-containing compounds, as the dominant constituents associated with odor (Zahn et al., 1997, 2001a, 2001b; Wright et al., 2005). Diet is thought to play a significant role in manure composition and in the emissions of VOCs and other gases. Although the links between diet, manure composition, and numerous manure emissions are becoming clearer (Sutton et al., 1999), the specific relationships with odor compound production and volatilization are poorly established. In vitro incubations of cattle feedlot and swine manures imply that unutilized starch in the manure is the predominant source of VFA during manure slurry storage (Miller and Varel, 2002, 2003; Miller and Berry, 2005), whereas aromatic compounds arise primarily from protein fermentation (Mackie et al., 1998). Establishing the links between diet, manure composition, VOC production, and VOC volatilization from the manure would speed efforts to reduce agricultural odors through low-cost diet modification as opposed to potentially costly manure storage and treatment alternatives.

Analysis of VOCs is a challenging multi-step process that involves initial concentration on an adsorbent matrix, elution of the compounds from the matrix by thermal or chemical methods, separation using gas chromatography (GC) techniques, and finally detection and quantification of individual compounds using a mass spectroscopy (MS) or flame ionization. Currently, adsorbent tubes are utilized for the capture and concentration of VOC (Hobbs et al., 1997, 1999; Zahn et al., 1997, 2001b; Keener et al., 2002; Willig et al., 2004). However, the potential for incomplete capture or the breakthrough of select compounds in the VOC mixture during long sampling times is an issue. Furthermore, there are processing issues associated with adsorbent tubes. Thermal desorption tubes require a high initial investment for desorption equipment, tubes, and tube conditioners, plus a GC must be dedicated to the thermal desorption equipment. For chemical extraction of adsorbed compounds, the sensitivity of detection is much lower (only a fraction of the extract is analyzed), and often the solvent peak can mask a number of highly volatile compounds during analysis.

Solid-phase microextraction (SPME) is similar to other adsorbent methods but is not reliant on expensive equipment for thermal desorption nor is chemical extraction necessary for sample analysis. As an alternative to other methods, SPME is fast, sensitive, and cost-effective (Koziel and Pawliszyn, 2001). In SPME, volatiles are first concentrated on an exposed fiber, and then the fiber is injected into a standard GC inlet, which volatilizes the entire sample. Odor compounds are separated and quantified using standard GC methods. Solid-phase microextraction technology has seen some application analyzing odorous compounds associated with agriculture (Lin et al., 2002; Begnaud et al., 2003; Spinhirne et al., 2003, 2004; Razote et al., 2004; Wright et al., 2005) and biosolids (Kim et al., 2005). Conversion of the SPME results into atmospheric concentrations is a challenge, because the amount of VOC captured on the fiber is not only dependent on the airborne concentration of VOC, but also on numerous environmental and experimental factors. Temperature, humidity, wind speed, length of time exposed, and type of fiber coating affect the mass of VOC bound to the fiber under nonequilibrium conditions (Pawliszyn, 1999; Scheppers Wercinski, 1999; Koziel et al., 2000; Alexander, 2004).

The objective of this study was to develop a quick, reliable, and inexpensive laboratory method utilizing SPME and dynamic flux chambers to compare the types and relative amounts of volatiles emitted from manures. We hypothesized that a number of chamber and manure factors would impact our measure of volatiles, and thus, these factors would need to be controlled to achieve consistent measurements. After developing the general protocol, this method was used to determine whether manures of differing origin (cattle or swine), diet, and storage conditions affected the types and relative amounts of emitted volatile compounds.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Fecal Sample Collection and Analysis
Fresh feces (<24 h old), which varied principally by animal and by diet (Table 1), were collected from cattle and swine production pens and individual stalls at the USDA-ARS, U.S. Meat Animal Research Center during the study. The feces were placed in plastic zip-lock bags and transported to the laboratory. Fecal samples were analyzed immediately or after storage (details in Table 2). Immediately before analyzing VOC emissions, two fecal subsamples (5 to 10 g) were taken for analysis of fecal constituents. Dry matter (DM) and organic matter (OM) contents were determined in one subsample by mass loss after drying overnight at 105°C and by mass loss-on-ignition at 425°C overnight, respectively (Nelson and Sommers, 1996). In the second fecal subsample, an equal mass of H2O was mixed into the sample, and the pH was determined in the slurry using a PHM 83 pH meter (Radiometer, Copenhagen, Denmark). After pH determination, the solids were separated from the liquid fraction by centrifugation at 2000 x g for 15 min, and the concentrations of H2O-extractable constituents (alcohols, VFA, aromatic ring-containing compounds, and L-Lactate) in the liquid phase were determined. Total alcohol and L-lactate were determined using the membrane-immobilized alcohol and L-lactate enzyme system, respectively, of the YSI Model 2700 autoanalyzer (Yellow Springs Instrument Company, Yellow Springs, OH). Other fermentation products extracted in the H2O (propanol, isobutanol, butanol, pentanol, hexanol, acetate, propionate, isobutyrate, butyrate, isovalerate, valerate, isocaproate, caproate, heptanoate, caprylate, phenol, {rho}-cresol, 4-ethyl phenol, indole, skatole, benzoate, phenylacetate, and phenylpropionate) were quantified using a Hewlett-Packard 6890 gas chromatograph (Agilent Technologies, Palo Alto, CA) equipped with flame ionization and mass selective detectors. Liquid extract (0.5 mL) was added to a 2-mL vial with ethyl butyrate internal standard (1 mM final conc.), 100 µL of 3 M HCl, and 800 µL of ether. The vials were crimp-capped, shaken for a minute, and a 2-µL volume from the upper ether phase was injected by autoinjector into a split/splitless inlet operated at 275°C and at a 30:1 split. Compounds were then separated and detected using previously published gas chromatographic operation parameters (Miller and Berry, 2005). Fecal solids remaining after cold H2O extraction were dried overnight at 105°C and ground using a mortar and pestle. The starch and protein contents of the dried solids were determined as previously reported (Miller and Berry, 2005).


View this table:
[in this window]
[in a new window]
  		Table 1. Diet formulations.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Evaluation of operational parameters for reliable chamber performance.{dagger}

 
Construction and Operation of the Dynamic Flux Chamber
The volatiles emitted from fecal samples were collected and sampled using a dynamic flux chamber (Fig. 1), which was slightly modified from the design described by Woodbury et al. (2006) that functioned as a continuous flow, stirred-headspace reactor. The chamber was constructed from two stainless steel bowls, one inverted over the other, with a butyl rubber gasket between the edges of the bowls and three binder clips equally spaced around the edges to ensure a tight seal. The lower bowl was filled with the fecal sample to a depth that corresponded to an exposed surface area of approximately 523 cm2 (diameter of 25.8 cm). Alternatively, a low volume pan insert was used to evaluate smaller sample volumes at equivalent surface area, but varying depth (<1 to 3 cm), depending on the amount of sample available. A small fan, switched and powered externally, was mounted within the upper bowl to stir the headspace above the sample. The upper bowl was modified with four air inlets, equally distributed around the sides, and one air exit port centered at the top of the inverted, upper bowl. Compressed air was purified by passage through a bed of activated carbon, humidified to >99% relative humidity using a 250-mL gas washing bottle containing 100 mL of distilled H2O, and supplied to the chamber via copper tubing through the four air inlets in the upper bowl. Humidity was monitored in selected experiments using a temperature/relative humidity probe (Pace Scientific, Mooresville, NC). A SPME sampling port was constructed using a three-way female tee NPT fitting, two Swagelok to male NPT connector fittings (one containing a septa and the other mounted to the top of the upper chamber), and a male NPT to tubing barb connector. As the air left the chamber, it passed through the three-way tee, where a SPME fiber (75 µm PDMS/Carboxen) from a portable field sampler (Supleco, Bellefonte, PA) inserted through the septum was exposed to the air stream. PDMS/Carboxen fibers were selected because recent studies found them to be superior to a variety of other SPME fibers when tested on malodorous gases (Begnaud et al., 2003; Kleeberg et al., 2005). Airflow through the system was balanced by regulating both the inflow of compressed air by fine metering valve (Supelco, Bellefonte, PA) and the outflow of air by vacuum pump using an in-line flow controller (SKC, Eighty Four, PA) down stream from the SPME port. Air inflow and outflow rates were verified using a digital flow meter (Humonics, Folsom, CA). All SPME fibers were initially conditioned for 1 h in a helium air stream at 300°C and subsequently conditioned in a helium air stream for 15 min at 300°C between samples.


Figure 1
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1. General schematic of the dynamic flow chamber slightly modified from the chamber described by Woodbury et al. (2006) showing the (1) low volume manure pan, (2) head space mixing fan, (3) four ports for clean air flow into the chamber, and (4) the exit port holding a portable solid-phase microextraction sampler.

 
Analysis of Volatile Organic Compounds by Gas Chromatography–Mass Spectroscopy and Relative Emission of Fecal Volatile Organic Compounds
The VOCs captured by the SPME fiber were analyzed using an Agilent 6890 gas chromatograph (Agilent Technologies, Palo Alto, CA) equipped with a 5973 MS. Exposed fibers were introduced through a split/splitless inlet (splitless mode) operated at 300°C and were separated on a 30 m x 0.25 mm diameter (0.25 µm film thickness) Innowax PEG column using the following program parameters: flow rate = 2 mL·min–1, initial temp = 50°C, initial time = 3 min, first temp ramp = 30°C min–1 to 170°C, second temp ramp = 20°C min–1 to 240°C, and a final hold at 240°C for 1.5 min. As compounds were eluted from the column, they were carried to the MS by heated transfer line (temp = 250°C) and analyzed using the following MS acquisition parameters: EM voltage = 1388 V, scan mode (30 to 550 atomic mass units), MS quad temp = 150°C, and MS source temp = 230°C. Tuning was performed before starting an experiment and at roughly weekly intervals using the autotune function of the Chemstation software package (Agilent). Peaks areas were calculated, and the identities were determined (quality score >50%) from the spectra (apex minus start of peak) using the National Institute of Standards and Technology 1998 reference library (Washington, DC). Quality scores and retention times were tracked throughout each experiment and used to aid in compound identification when compound identification scores were less than the 50% benchmark for compound identification. Major peaks (see Table 3) routinely achieved quality scores >75%. The identities of select peaks were further confirmed using a standard mixture (Miller and Varel, 2003) containing alcohols, VFAs, and aromatic ring-containing compounds.


View this table:
[in this window]
[in a new window]
  		Table 3. Specific volatile organic compounds routinely identified in swine and cattle fecal emissions by mass spectroscopy grouped according to compound class.

 

Figure 2
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 2. Total ion current (IC) profiles for fresh (A) cattle C1 and (B) swine S2 feces. Major peaks marked with letters are identified in Table 3. Fiber exposure time = 5 min; flow rate = 1 L min–1 (16 cm s–1); chamber temperature = 20°C.

 
Relative emissions from fecal samples were compared in two ways—total ion current (IC) signal or the combined IC signal of compounds belonging to the same compound class. We initially assumed that the peak area (in IC) was proportional to the amount of VOC adsorbed to the fiber during its exposure to the air leaving the flux chamber. For total IC, background trace contaminants, predominantly the CO2 injection peak and siloxanes arising from the fiber and/or septa, were subtracted from the total peak area.

Evaluation of Chamber and Manure Parameters for Comparison Test Conditions
Several factors were evaluated to determine their effect on the amount of VOC captured by the SPME fiber to determine operational parameters for reliable chamber performance (Table 2). Fresh feces from the C1 and S2 diets (Table 1) was stored between analyses at –20°C and used to determine chamber performance. In the basic procedure, a sample was placed in the chamber, the fan was turned ‘ON’ to mix the headspace contents, and humidified air was pulled through the chamber for at least 5 min to stabilize the concentrations of odorant compounds in the chamber headspace. The SPME fibers were then exposed for 5 min to the air stream leaving the chamber. Whenever different samples or treatments were evaluated, the chamber was cleaned between samples using a 70% ethanol spray, followed by a thorough water rinse, and finally drying at least 30 min at 105°C to remove any volatiles adhering to the inside of the chamber.

The basic evaluation procedure was modified (see details in Table 2) to examine the effect of several chamber-related parameters on volatiles captured by the SPME fiber from a single swine and a single beef cattle manure source (S2 and C1, respectively) and included: (1) SPME exposure time, (2) airflow rate through the chamber, (3) SPME fiber orientation (parallel vs. perpendicular), and (4) cleaning procedures to reduce the carryover of volatiles from earlier samples. Additionally, several feces-related parameters were also evaluated: (5) feces surface area and mass, (6) the amount of time needed for stable VOC emission and the length of time VOC emission were stable, and finally (7) the effect of sample temperature. Statistical analyses of the various tests were made using the regression procedure of the SAS statistical software package (SAS Institute, 2001) with total IC signal as the dependent variable and exposure time (1), airflow rate (2), feces surface area or mass of feces (5), feces time in chamber for stable VOC emission (6), or sample temperature (7) as the independent variables. Two fecal samples were used to evaluate chamber/feces-related parameters 1, 5, 6, and 7, whereas triplicate samples from the two fecal sources (C1 and S2) were used to evaluate chamber/feces-related parameters 2, 3, and 4. Differences (P < 0.05) between means of SPME fiber orientation in the inlet (3) and effect of cleaning (4) were detected using t tests.

After establishing the best operational conditions for comparing fluxes from fecal (or other) samples, the effects of feces storage temperature and time on fluxes were evaluated. Three sample storage temperatures (–20, 4, and 20°C) were evaluated using fresh cattle (C1) and swine (S2) feces. Several kg of fresh feces were collected, mixed, and placed in plastic bags, which were subsequently stored at one of the three temperatures for varying periods of time. Triplicate samples were analyzed at each time point, and the resulting total IC signals were analyzed by regression (SAS Institute, 2001) and nonlinear regression techniques of the SigmaPlot graphics software program (SPSS Incorporated, 2002).

As an example of the utility of the flux chamber/SPME method, the fluxes from feces of cattle and swine consuming a variety of diets were surveyed. Samples of six different fresh (noncrusted) feces (C2, C3, C4, C5, S1, and S2) were collected from three different pens. After measuring the VOC signal from each sample, the samples were placed in plastic bags and incubated overnight at 37°C. This incubation was used to gauge the relative potential of the fresh feces to produce additional volatiles as manure accumulated in the pen. Total IC measurements of compound groups were analyzed using a general linear model (SAS Institute, 2001). Data was first sorted by manure treatment (fresh or incubated), and the effect of diet was tested by the model. Differences (P < 0.05) between diet means were determined using t tests. Comparisons between fresh and incubated feces were made using paired t test. Pearson correlation was used to identify significant relationships between the increase in total IC due to incubation and changes in fecal parameters.

Investigating Relationships between Total Ion Current and Volatile Organic Compound Emission
Two series of tests were conducted using pure butanol in the first series and an aqueous mixture of alcohols, VFA, and aromatic compounds in the second series as emission standards. In the butanol tests, pure butanol was drawn up into a capillary tube, which was weighed and placed in the emission chamber. Over a period of several hours, SPME samples were collected, the IC signals of the butanol peaks were determined, and the capillary tube was weighed to determine the rate of mass loss. Several different diameters of capillary tubes were used in this series of tests to produce a variety of butanol emission rates. Regression analysis was used to determine the relationship between total IC signal and the rate of butanol mass loss.

In the second series of tests, an artificial mixture of volatile compounds was utilized to mimic the emission of volatile fecal compounds. The mixture contained alcohols (ethanol, propanol, and butanol), VFA (acetic, propionic, isobutyric, butyric, isovaleric, valeric, and hexanoic acids), and aromatic (phenol, {rho}-cresol, 4-ethyl phenol, indole, and skatole) compounds in water at concentrations ranging from 0.02 to 5 mM. Similar to the butanol tests, the mixture was added to an aluminum pan, weighed, and placed in a chamber. Over several hours, SPME samples were collected, the IC signal was measured, and the mixture was weighed and sampled. Concentrations of volatile compounds were determined by GC as described above. The rate of volatile mass loss was determined from the change in fluid mass and concentration of volatiles. Regression analysis was used to calculate the VOC emission rate and standard error associated with the rate estimate. Total IC (seven or eight depending on the test) responses were averaged. Varying rates of VOC emission were accomplished using different pans with greater surface area or by diluting the VOC mixture 1:10 or 1:2 in distilled water.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Detection and Characterization of Fecal Volatile Organic Compounds
Two typical VOC profiles obtained using the SPME-DFC method with feces collected from animals fed the C1 or S2 diets are depicted in Fig. 2 and indicate that a wide variety of organic compounds were detected at various levels belonging to a variety of compound classes (Table 3). A similar diversity of compounds has been previously reported in animal production environments utilizing adsorbent tubes (Hobbs et al., 1997, 1999; Zahn et al., 1997; McGinn et al., 2003; Rabaud et al., 2003; Willig et al., 2004) and recently applied SPME methods (Begnaud et al., 2003; Spinhirne et al., 2003; Razote et al., 2004). In this study, the volatiles emitted from the two feces were dominated by VFA and aromatic compounds, and for the cattle feces, alcohols also formed a major component of the compounds captured by the SPME fiber. Contrasting the two emissions illustrates the differences in both the quantity of compounds and in the composition. For instance, VOC emitted from the S2 feces were at least three-fold higher in VFA, enriched in branched-chain VFA and aromatics, but depleted in alcohols relative to the C1 feces. Although not easily distinguishable in Fig. 2, differences in the presence and content of major and minor volatile compounds, such as sulfides, were identified, which could influence odor offensiveness and intensity.

Microbial fermentation of protein and starch, before defecation and during manure storage, has been identified as the primary source of malodorous compounds (Mackie et al., 1998; Miller and Varel, 2003). It is likely that accumulation of fermentation products would affect emission primarily through changes in concentration and secondarily by modifying volatility. Acid accumulation and enhanced VFA volatility has been well established (Derikx et al., 1994) and is purported to be a major factor associated with cattle feedlot odors (Watts et al., 1994; Miller and Berry, 2005). Shifts in the solubility of certain odor compounds due to increasing alcohol content may also affect emissions and needs to be evaluated.

Evaluation of Chamber-Related Parameters
The adsorption of VOC onto the exposed SPME fiber with time was nonlinear and fit a hyperbolic equation (Fig. 3A). For these samples, the regression equation predicted that the maximum amount (equilibrium concentration) of volatiles to bind to the fiber would produce a total IC of 2.08 and 1.00 x 109 units for the S2 and C1 feces, respectively. To achieve 99% of the predicted equilibrium value, fibers would need to be exposed for more than 14 and 39 h to the air stream from these swine and cattle feces, respectively. Clearly, waiting for equilibrium conditions would vastly constrain the number of observations and, furthermore, the stability of emissions over the lengthy time needed for equilibration is doubtful (see below). Similar nonlinear adsorption behavior by SPME fibers with time has been reported in complex waste gas exhausts (Kleeberg et al., 2005) and in controlled laboratory experiments (Koziel et al., 2000). In both of those instances, short exposure times were used to reduce the effects of competitive adsorption. Therefore, a 5 min SPME exposure was utilized in the general protocol.


Figure 3
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 3. Relationship between total ion current (IC) bound to the solid phase microextraction (SPME) fiber for cattle C1 and swine S2 feces and (A) the time of fiber exposure to the air stream leaving the dynamic flux chamber [flow rate = 1 L min–1 (16 cm s–1); chamber temperature = 20°C] or (B) air flow rate leaving the chamber [fiber exposure time = 5 min; chamber temperature = 20°C].

 
Although the total mass of volatile compounds increased with time, examination of classes of volatile compounds showed that the percent composition of some of the compound classes also changed (i.e., they were enriched or depleted as a proportion of the total IC signal) with increasing exposure time (Table 4). For both feces, there was an enrichment of aromatic and sulfur-containing compounds at the expense of straight-chain VFA. The percentage composition of branched-chain VFA also increased with longer SPME exposure times in the S2 feces. These differences are likely related to differences in equilibration times between compounds. Although these changes in relative contribution complicate the interpretation of the total IC profile, the magnitudes of the changes (a fraction of a percent per minute) are quite small in relation to the 5-min period that the SPME fiber is exposed.


View this table:
[in this window]
[in a new window]
  		Table 4. Relative enrichment or depletion of SPME-bound volatile compounds emitted from swine and cattle feces according to compound class with increasing SPME exposure time.{dagger}

 
The effect of air flow rate into the chamber and past the SPME fiber was evaluated and found to have little effect on volatiles bound to the SPME fiber at the flow rates tested (Fig. 3B). The rate of total IC increase was insignificant (P > 0.192) for both fecal samples tested, indicating that the range of flow rates did not effect volatilization from the feces or enhance the binding of volatiles to the SPME fiber. For S2, the relationship between total IC and flow rate appeared to be more ambiguous than for the C1 samples, but the slope of the regression line was heavily influenced by measurements of total IC at the 0.5 L min–1 flow rate (8 cm s–1). Normally one would expect higher air flows to strip more VOC from the exposed feces surface, but the continual mixing of headspace by the fan likely overshadows any effect on increased air flow into the chamber. Thus, the rate of total IC increase due to enhanced air flow was very small (1.3% per 1 L min–1 increase in flow rate), and other factors (i.e., SPME exposure time, fecal surface area, fecal temperature, and carryover of volatiles) would ultimately have a larger effect on the amount of volatiles bound to the SPME fiber.

The third chamber-related component tested, the orientation of the SPME fiber within the inlet, did not affect the amount of VOC bound to the fiber when tested on the S2 feces, which was incubated overnight at 37°C. In the perpendicular orientation (fiber perpendicular to a linear air stream), the total IC obtained from the fiber was 8.38 ± 1.13 x 108 units. The total IC obtained from the fiber exposed in the parallel orientation (air stream diverted 90° within the sampling inlet) was 7.08 ± 0.36 x 108 units, which did not differ (P = 0.314) from the perpendicular orientation. Thus, SPME orientation within the air stream did not affect the amount of volatiles bound to the fiber. One factor, however, influencing the orientation of the SPME inlet was the increased incidence that the SPME fiber would break when it was jostled in the perpendicular orientation. Thus, the orientation adopted for the standard protocol utilized the parallel orientation (90° bend of the air stream in the inlet).

Earlier versions of the flux chamber utilized a flow-through plastic container to funnel volatiles onto solid-phase extraction tubes and/or SPME fibers. Although VOC could be characterized and quantified using the plastic pan systems, there was considerable carryover (>10% of the original total IC), even after vigorously cleaning the chamber using a variety of detergents, solvents, and heating treatments. We attributed the carryover in the plastic pan flux chamber to the slow release of adsorbed fecal volatiles from the surfaces which were in direct contact with the fecal sample. Because of the large carryover in the plastic chamber, the stainless steel chamber reported in this study was developed. Carryover from the emptied stainless steel chamber was considerable (14.4% of total IC obtained from the fecal sample) likely due to a larger chamber surface area. However, a simple cleaning technique (spray with 70% ethanol, rinse with H2O, and drying 30 min at 105°C) reduced carryover to <1.3% of the total IC obtained from the fecal sample. Overnight drying completely eliminated carryover.

Evaluation of Feces-Related Parameters
The stability of fecal volatile emissions (i.e., how long it takes to reach a stable emission, and how long emissions are stable) was also a concern, but experiments demonstrated that emission of fecal volatiles stabilized very quickly and remained stable over several hours (Fig. 4A). Emissions from the cattle C1 feces stabilized almost immediately and remained remarkably stable; the initial measurement deviated from the mean of all other cattle measurements by <0.5%. The emission of volatiles from the swine S2 fecal samples were also quite stable, but the initial emission measured from 0 to 5 min was lower (8.9%) compared with the average of emissions measured after 10 min of equilibration time. There was a tendency (P ≤ 0.067) toward increased total IC signal with increasing time in the chamber for both feces tested, but that trend was minor (<0.05% increase per hour of chamber time) relative to the overall total IC signal.


Figure 4
View larger version (18K):
[in this window]
[in a new window]
  		Fig. 4. Amount of (A) total ion current (IC) and (B) chamber air relative humidity and temperature with increasing time in the dynamic flux chamber for the cattle C1 and swine S2 feces. Fiber exposure time = 5 min; flow rate = 1 L min–1 (16 cm s–1); chamber temperature = 20°C.

 
Two environmental factors that could influence the sorption of VOC onto the SPME fiber are the relative humidity and temperature of the air stream (Fig. 4B). Linear regression with backward elimination of nonsignificant variables was used to determine the significance of relative humidity, air temperature, and time in the chamber on total IC measurements. For swine S2 feces, relative humidity proved to be the only significant (P = 0.020) factor; time in chamber and air temperature were not significant (P > 0.15). However, the contribution of relative humidity toward total IC signal was small—for every 10% rise in relative humidity, there was a corresponding 1.6% increase in total IC signal. Moreover, the initial time point, which was substantially lower in both total IC and relative humidity than the rest of the data, strongly influenced whether the slope of the regression line was significant. A follow-up linear regression analysis, which omitted the initial data point, revealed that humidity had only a tendency to affect (P = 0.095) the total IC signal from the swine S2 feces. For cattle C1 feces, regression analysis using relative humidity, air temperature, and time in chamber indicated that air temperature was the only significant factor (P = 0.047); humidity and time in the chamber did not affect (P > 0.37) total IC. However, only a narrow range of air temperatures were evaluated (20.1 to 22.4°C) in the cattle feces experiment; the increase in temperature (2.3°C) was associated with an 11.8% rise in total IC signal from the cattle C1 feces. Based on these results, we determined that: (1) long periods of time (up to 3 h) with sample in the chamber did not affect total IC measurements, (2) a short equilibration period (5 to 10 min) within the chamber was prudent to establish consistent relative humidity (>90%), and (3) further evaluation of fecal temperature (see below) was warranted.

The mass and area of feces within the chamber could also affect VOC emissions. Although there was a moderate relationship between total IC signal and mass of feces (Fig. 5A), the relationship between total IC signal and the exposed surface area was much stronger (Fig. 5B). To ensure a consistent comparison between fecal samples, an arbitrary surface area of approximately 523 cm2 of feces was adopted (the area of the insert pan used). Smaller fecal samples could be accommodated within the chamber with smaller insert pans, but the total IC signal would need to be scaled to the 523 cm2 benchmark for meaningful comparisons. Hand-in-hand with the need for a standard surface area (or total IC signal per unit area) was the need to have a flat, uniform fecal surface. Great care was taken to smooth the fecal surface using a kitchen spatula, because a rough or clumpy surface increased the total IC signal by more than 10% relative to the total IC signal from the flat surface (Miller, unpublished data, 2004).


Figure 5
View larger version (13K):
[in this window]
[in a new window]
  		Fig. 5. Dependence of total ion current (IC) signals on (A) mass or (B) area of cattle C1 feces in the chamber. To increase the total IC signal, cattle C1 feces was stored overnight at 37°C and subsequently at 20°C for 2 wk. Fiber exposure time = 5 min; flow rate = 1 L min–1 (16 cm s–1); chamber temperature = 20°C.

 
The temperature of the feces proved to have a strong affect on the VOC bound to the SPME fiber (Fig. 6). For both feces, the mass of volatiles bound to the SPME fiber increased as the temperature increased with a quadratic function fitting the data. Linear regression with backward elimination of nonsignificant factors eliminated the first-order temperature term and yielded a model relating total IC to the square of the feces temperature. Both the y intercept term and the temperature coefficient significantly differed from zero (P ≤ 0.02) and differed between animal feces. One interpretation of these differences is that the quantity of compounds volatilizing from the surface of the feces differed (i.e., differing y intercept terms) and the composition of compounds subject to volatilization differed (i.e., differing temperature coefficients). The application of the temperature model was limited, because the model predicts increasing emissions below 0°C. In retrospect, one would expect emissions to approach zero with ever lower temperatures. However, the data clearly demonstrate the need for measurements made at a defined temperature, if reliable comparisons of volatile emissions from samples are to be made. All future measurements were conducted using samples equilibrated to 20°C.


Figure 6
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 6. Relationship between total ion current (IC) signal obtained from cattle C1 and swine S2 feces and the temperature of the feces. Fiber exposure time = 5 min; flow rate = 1 L min–1 (16 cm s–1).

 
Application of the General Solid-Phase Microextraction-Flux Chamber Method
Based on the results of the experiments establishing chamber performance and significant sample parameters, a general chamber procedure was adopted. This general method ensures that enough volatiles are collected by the fiber for GC–MS analysis while enabling an analyst to process enough samples for replicate analysis of multiple fecal samples within a day. In the general method, fecal equilibration time (5 min) before SPME exposure to attain consistent chamber humidity >90%, SPME exposure time (5 min), air flow rate (1 L min–1 [16 cm s–1 past the fiber]), inlet orientation (SPME fiber from above with air drawn off to the side), air supply (humidified, compressed air), sample temperature (20°C), and exposed sample surface (approximately 523 cm2) were controlled. Chambers were also cleaned using a 70% ethanol spray followed by 30 min at 105°C and cooling to room temperature between different fecal samples. Multiple chambers were used in rotation to continuously analyze samples.

Effect of Feces Storage Temperature on Volatile Organic Compound Emissions
A variety of parameters associated with manure and fecal samples can change during sample storage depending on storage conditions (Peters, 2003). Requiring sample analysis on the same day as sample collection would likely limit the application and usefulness of the SPME-flux chamber method. Thus, the first application of the chamber technique was to evaluate the best sample storage conditions that minimize changes in total IC measurements. Storage at ambient (20°C), refrigeration (4°C), or freezer (–20°C) temperatures was evaluated, and freezing proved to have the lowest impact on the emission of volatiles from the samples (Fig. 7). Regression analysis revealed no change in total IC with time for both frozen cattle and frozen swine feces (P > 0.66). Refrigeration yielded a mixed result with a significant (P < 0.01) increase in the total IC associated with swine feces, but a nonsignificant (P = 0.12) increase in the total IC associated with cattle feces. However, the increases in total IC associated with refrigeration were quite small (<2% per day of sample storage) relative to the total IC signal of the fresh fecal samples and indicate that short-term refrigeration (up to 3 d) would not overly influence the final total IC measurement. Ambient storage of fecal samples had a dramatic effect on the total IC measured from both the swine and cattle fecal samples during the initial week of storage. Nonlinear regression using a three-parameter, exponential rise to maximum equation fit the observations very well (r2 > 0.914) for both swine and cattle feces stored at ambient temperatures. Linear regression using data collected during the first week of storage indicated that the total IC signal increased (P < 0.01) at a rate of 87 and 9% per day of storage at ambient temperature for the cattle C1 and swine S2 fecal samples, respectively.


Figure 7
View larger version (18K):
[in this window]
[in a new window]
  		Fig. 7. The effect of storing (A) cattle C1 and (B) swine S2 fecal samples up to 5 wk in a freezer (–20°C), a refrigerator (4°C), and at ambient (20°C) temperatures on total ion current (IC) signal. Fiber exposure time = 5 min; flow rate = 1 L min–1 (16 cm s–1); chamber temperature = 20°C.

 
Effect of Diet on Volatile Organic Compound Emissions
The real utility of the SPME-flux chamber technique depends on its ability to discriminate differences in the emissions of volatile compounds from different samples. The measurement of liquid-phase malodorous compounds in manure storage systems (lagoons, pits, and slurries) is often used to evaluate the effectiveness of odor intervention strategies (Varel et al., 2004). Measuring volatilized odor compounds (especially under controlled conditions) is achievable (Hobbs et al., 1997, 1999), but technically challenging and labor intensive. To illustrate the utility of the SPME-flux chamber method, we applied the general protocol to six different fecal samples to determine whether there were any differences in the volatiles emitted from the various sources. Additionally, differences in potential VOC emission were determined as the increase in total IC signal after overnight incubation of the feces.

Striking differences were detected between the different fresh fecal samples before incubation (Table 5). For fresh cattle and swine feces, the total IC signal was greatest when ground corn was the primary dietary ingredient. Notable differences were detected between cattle and swine samples, particularly in the total IC and IC associated with particular compound classes (i.e., straight-chain VFA, branched-chain VFA, and aromatic compound content). Only minor differences were detected between the feces collected from swine on the two diets; the content of straight- and branched-chain VFA was greater on S1 compared with S2. Within the cattle diets, alcohols, straight-chain VFA, and aromatic compounds were the primary classes of compounds detected. One interesting pattern involved increasing sulfur compound signal with increasing corn silage content. Feces from animals fed C3 and C5 (70% and 66% corn silage, respectively) emitted more sulfur compounds than the other two diets, which contained <20% corn silage. Conversely, ester signal from cattle feces increased with decreasing corn silage content. Future comparisons of the volatiles of fresh feces from cattle and swine fed other diets should clarify these initial relationships.


View this table:
[in this window]
[in a new window]
  		Table 5. Mass spectrometer signal (ion current) produced from SPME-captured volatiles emitted from a variety of fresh and incubated fecal samples.{dagger}

 
Like the fresh fecal samples, differences were observed in total IC associated with incubated fecal samples when diet and animal source were varied (Table 5). The total IC produced differed with diet and was largest for the cattle fed the C2 diet (ground corn), followed by feces from the two swine diets (ground corn), cattle feces from the C4 diet (predominantly high moisture corn), and finally cattle feces from the C5 and C3 diets (predominantly corn silage). For differences between animal source, swine total IC signal from incubated samples was enriched in branched-chain VFA and aromatic ring compounds compared with cattle samples. Differences based on diets by animal source were evident. For instance, the total IC signals from swine feces on the S1 diet were enriched in straight-chain VFA and sulfur compounds relative to the feces of swine fed the S2 diet. For cattle, the total IC signals from the various diets in general were consistent with differences observed in the fresh fecal samples. Feces from cattle fed corn silage as a dietary component produced total IC containing a higher proportion of sulfur compounds, whereas ketone signal diminished with corn silage content. Conversely, the signal from branched-chain VFA increased with decreasing corn silage content in cattle diets.

Finally, contrasting fresh and incubated feces, larger IC signals of the incubated feces were observed in all compound classes except aldehydes (Table 5). Total IC signal increased in every manure sample, but to differing degrees. The potential increase in total IC from incubated feces ranged from 49.1 to 754.6 x 106 units for the cattle fed C3 and C2 diets, respectively. Expressed as a percentage, the increase in total IC signal due to incubation ranged from 100 to 750% for swine fed the S1 diet and cattle fed the C4 diet, respectively. Alcohol was the only compound class that increased in all manure incubations. Branched-chain and straight-chain VFA increased in incubated feces from diets where either high moisture corn or ground corn was the principal ingredient. For swine feces samples, the IC from nearly all compound classes increased, except for aldehydes in the S1 or S2 feces and aromatic compounds in the S1 feces. For all cattle feces tested, aromatic compounds did not increase during incubation. Clearly, differences in potential VOC emission based on animal species and diet were detected, and the incubation protocol provides useful insights about the potential for odor compound emissions as manure accumulates in production systems.

What substrate fuels odor compound production during incubation? Fecal starch and protein, both separately or in combination, are the precursors of odor compounds and may contribute to manure acidification, which enhances the volatility of particular odorants (Mackie et al., 1998). Consistent with earlier studies in our laboratory (Miller and Varel, 2002, 2003; Miller and Berry, 2005), starch loss in the feces was more closely correlated to VFA accumulation in the incubated feces on a DM basis (r = –0.65, P < 0.01) than crude protein losses (r = –0.30, P = 0.22) regardless of animal type or diet. Furthermore, the relationship between starch loss and the increase in total IC signal resulting from overnight incubation was also strong (r = –0.48, P < 0.05), further emphasizing the links between manure starch, odor compound accumulation, and emissions. Other correlations (P < 0.05) with the increase in total IC due to overnight incubation included pH (r = –0.48), alcohol content (r = 0.73), and lactic acid content (r = 0.75).

Relating Total Ion Current to Volatile Organic Compound and Odor
A primary goal of manure emissions research is to measure or predict VOC and/or odor emissions. How appropriate is total IC as a proxy for VOC? Recent studies in the literature have utilized total IC to compare the atmospheres of livestock buildings (Begnaud et al., 2003). There is some question regarding the appropriateness of total IC as a single measure of odor or VOC. In these experiments, two assumptions were made regarding the use of total IC as proxy for VOC emission. The first assumption is that the IC produced for all the compounds is equivalent (i.e., ionization is independent of chemistry). Injection of a standard mixture containing known concentrations of alcohols, VFA, and aromatic compounds shows that this assumption may be questionable—a sixfold difference in IC response was observed between compounds when standardized to mass. Within VFA and aromatics compound classes, variation was reduced to less than threefold and indicates that general comparisons of compound class IC between different fecal sources may be made if the abundance profile of individual compounds was similar. However, the most conservative assessment of the relative emission could still be made by comparing the IC of individual compounds.

The second assumption assumes that total IC signal is proportional to VOC emission. Permeation tubes have been used as emission standards for calibration of airborne VFA concentrations using SPME samplers (Spinhirne and Koziel, 2003). In a series of tests, we found that the mass loss of pure butanol from capillary tubes placed in the flux chamber was directly proportional (r2 = 0.984) to the IC measured for SPME fibers exposed to a range of butanol emission rates. Further testing with a mixture containing alcohols, VFA, and aromatics also demonstrated a strong positive relationship between total IC and VOC emission rate (Fig. 8). It is important to note in these tests that the relative uncertainty associated with both total IC signals and VOC emission rates (error in x and y measurements) was similar, and measuring VOC emissions by mass loss in the most controlled environment is difficult. Thus, we deemed that comparisons of total IC or the IC of compound classes between samples was appropriate because flux conditions (temperature, air flow rate, air entering the chamber, and humidity) were consistent between measurements. Given the limitations of our assumptions about total IC and VOC in these experiments, care was taken to discuss total IC in comparisons as broad indicators of trends in VOC emissions because total IC was not an exact predictor of VOC concentrations.


Figure 8
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 8. Relationship between total ion current (IC) signal and volatile organic compound (VOC) flux from standard mixtures of VOC. The standard error (SE) for VOC flux was calculated by regression analysis from rate of VOC mass loss. Mean total IC signal and SE calculated from seven to eight total IC measurements. Fiber exposure time = 5 min; flow rate = 1 L min–1 (16 cm s–1); chamber temperature = 20°C.

 
Going the final step, how well does VOC predict odor? Available literature indicates that there is a relationship between the two. In fact, models describing human odor perception have been developed utilizing specific VOC compounds (Zahn et al., 2001a). Other researchers have also advocated for specific VOC mixtures as artificial odor sources. Future work utilizing olfactometry needs to be made to further define the relationships between odor, VOC, and total IC signals obtained using the SPME flux chamber technique.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
A general protocol was developed to compare the relative emissions of VOC from feces. The SPME-flux chamber technique was robust and provided unique insights into the type and general magnitude of VOC emissions that emanate from feces from different animal species fed differing diets. Initial application of the technique illustrated the need for careful preservation of fecal samples (freezing) to ensure representative emissions measurements in the laboratory. A broad survey of emission from fecal samples obtained from cattle and swine fed a variety of diets identified significant differences in the total emission and the emission of specific odor compound classes from fresh and incubated samples.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
B. L. Woodbury, S. M. Lesch, R. A. Eigenberg, D. N. Miller, and M. J. Spiehs
Electromagnetic Induction Sensor Data to Identify Areas of Manure Accumulation on a Feedlot Surface
Soil Sci. Soc. Am. J., October 21, 2009; 73(6): 2068 - 2077.
[Abstract] [Full Text] [PDF]

Home page
 J ANIM SCIHome page
A. Shabtay, U. Ravid, A. Brosh, R. Baybikov, H. Eitam, and Y. Laor
Dynamics of offensive gas-phase odorants in fresh and aged feces throughout the development of beef cattle
J Anim Sci, May 1, 2009; 87(5): 1835 - 1848.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
Y.-C. M. Lo, J. A. Koziel, L. Cai, S. J. Hoff, W. S. Jenks, and H. Xin
Simultaneous chemical and sensory characterization of volatile organic compounds and semi-volatile organic compounds emitted from swine manure using solid phase microextraction and multidimensional gas chromatography-mass spectrometry-olfactometry.
J. Environ. Qual., March 1, 2008; 37(2): 521 - 534.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Miller, D. N.
Right arrow Articles by Woodbury, B. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Miller, D. N.
Right arrow Articles by Woodbury, B. L.
Agricola
Right arrow Articles by Miller, D. N.
Right arrow Articles by Woodbury, B. L.
Related Collections
Right arrow Volatile Organic Compounds
Right arrow Air Pollution
Right arrow Animal Waste

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome