Sensors and Actuators B: Chemical

Transcription

1 Sensors and Actuators B 146 (2010) Contents lists available at ScienceDirect Sensors and Actuators B: Chemical journal homepage: Field air sampling and simultaneous chemical and sensory analysis of livestock odorants with sorbent tubes and GC MS/olfactometry Shicheng Zhang a,b, Lingshuang Cai b, Jacek A. Koziel b,, Steven J. Hoff b, David R. Schmidt c, Charles J. Clanton c, Larry D. Jacobson c, David B. Parker d, Albert J. Heber e a Department of Environmental Science and Engineering, Fudan University, Shanghai, China b Department of Agricultural and Biosystems Engineering, Iowa State University, Ames, IA 50011, USA c Department of Bioproducts and Biosystems Engineering, University of Minnesota, St. Paul, MN, USA d College of Agriculture, West Texas A&M University, Canyon, TX, USA e Department of Agricultural and Biological Engineering, Purdue University, West Lafayette, IN, USA article info abstract Article history: Available online 22 November 2009 Keywords: GC O Livestock Odor Field air sampling VOCs Characterization and quantification of livestock odorants is one of the most challenging analytical tasks because odor-causing chemicals are very reactive, polar, and often present at very low concentrations in a complex matrix of less important or irrelevant gases. The objective of this research was to develop a novel analytical method for characterization of the livestock odorants including their odor character, odor intensity, and hedonic tone and to apply this method for quantitative analysis of the key odorants responsible for livestock odor. Field samples were collected with sorbent tubes packed with Tenax TA. The automated one-step thermal desorption module coupled with multidimensional gas chromatography mass spectrometry/olfactometry system was used for simultaneous chemical and odor analysis. Fifteen odorous VOCs identified from livestock operations were quantified. Method detection limits ranged from 30 pg for indole to 3590 pg for acetic acid per sample. In addition, odor character, odor intensity, and hedonic tone associated with each of the target odorants were also analyzed simultaneously. The mass of each VOC in the sample correlated well with the log stimulus intensity of odor. All of the coefficients of determination (R 2 ) were greater than 0.74, and the top 10 R 2 s were greater than Field air samples from swine and dairy operations confirmed that target compounds quantified represented typical odor-causing compounds emitted from livestock Elsevier B.V. All rights reserved. 1. Introduction Odor emissions from livestock facilities affect air quality in surrounding communities. Many volatile organic compounds (VOCs) have been identified, including acids, alcohols, aldehydes, amines, volatile fatty acids (VFAs), hydrocarbons, ketones, indoles, phenols, nitrogen-containing compounds, and sulfur-containing compounds as those are typically emitted from animal agriculture [1,2]. Compounds contributing to livestock odor have been identified, such as VFAs, p-cresol, 4-ethylphenol, indole, skatole, and sulfur-containing compounds [3 7]. Corresponding author at: Department of Agricultural and Biosystems Engineering, Iowa State University, 3103 NSRIC, Ames, IA 50011, USA. Tel.: ; fax: addresses: zhangsc@fudan.edu.cn (S. Zhang), lscai@iastate.edu (L. Cai), koziel@iastate.edu (J.A. Koziel), hoffer@iastate.edu (S.J. Hoff), schmi071@umn.edu (D.R. Schmidt), cjclanton@umn.edu (C.J. Clanton), jacob007@umn.edu (L.D. Jacobson), dparker@mail.wtamu.edu (D.B. Parker), heber@purdue.edu (A.J. Heber). Livestock odor can be measured with sensory and/or analytical methods. Olfactometry determines odor concentration in terms of dilutions to threshold using dynamic forced-choice olfactometry, which relies on air sample collection in bags for subsequent evaluation with panelists. Analytical methods provide compoundspecific information, such as gas concentration and, if used in conjunction with olfactometry, an identification of individual odorous compounds that might be responsible for causing the odor. Quantification of these odor-causing compounds is useful for air quality measurements and also for the development and assessment of odor mitigation technologies. Gas chromatography (GC) mass spectrometry (MS)/olfactometry (O) offers the advantages of combining sensory assessment with the identification and quantification of compounds. This method was used for identification of odorous compounds from swine facilities [3 7]. Rabaud et al. [8] used thermal desorption (TD) GC O/MS to identify and to quantify odorous compounds from a dairy. Simultaneous GC O approach was used for identification and qualitative assessment of odorous compounds from swine facilities [3 5]. To date, relatively few references exist on (a) /$ see front matter 2009 Elsevier B.V. All rights reserved. doi: /j.snb

2 428 S. Zhang et al. / Sensors and Actuators B 146 (2010) quantification of livestock VOCs and (b) linking typical livestock VOCs and their concentrations with measured odor parameters [9 11]. The focus of this research was to develop an odor characterization method for specific livestock odorants and develop a quantitative analysis method for the key odorous compounds responsible for livestock odor emissions using TDmultidimensional (MD)GC MS/O system. In addition, correlations between odor intensities and VOC concentrations were developed. 2. Experimental and methods 2.1. Thermal desorption multidimensional GC MS/olfactometry (TD MDGC MS/O) system Simultaneous chemical and sensory analyses of livestock odorants were completed using the TD-MDGC MS/O system. The TD system consists of a Model 3200 automated thermal desorption inlet for Agilent 6890 GC developed by Microanalytics (Round Rock, TX, USA) based on a PAL autosampler. The unique design of the Model 3200 system allows for gentle purging of air and water from sorbent tubes prior to a single-step sample desorption and introduction to GC. This system eliminates desorption followed by a separate step of cryotrapping and subsequent rapid desorption. Instead, samples are desorbed directly onto the front of GC column, eliminating problems associated with a typical desorption trapping desorption and problems with the presence of water/air in sorbent tubes. Multidimensional GC MS/O (Microanalytics) was equipped with two columns connected in series. The non-polar pre-column was 12 m, 0.53 mm i.d.; film thickness, 1 m with 5% phenyl methylpolysiloxane stationary phase (SGE BP5) and operated with constant pressure mode at 8.5 psi (0.58 atm). The polar analytical column was a 25 m 0.53 mm fused silica capillary column coated with poly (ethylene glycol) (WAX; SGE BP20) at a film thickness of 1 m. The column pressure was constant at 5.8 psi (0.39 atm). System automation and data acquisition software were MultiTraxTM V and AromaTraxTM V (Microanalytics) and ChemStationTM (Agilent, Santa Clara, CA, USA). The general GC run parameters used were as follows: injector, 260 C; FID, 280 C, column, 40 C initial, 3 min hold, 7 Cmin 1, 220 C final, 10 min hold; carrier gas, GC-grade helium. The GC was operated in a constant pressure mode where the mid-point pressure, i.e., pressure between pre-column and column, was always at 5.8 psi (0.39 atm) and the heart-cut sweep pressure was 5.0 psi. The MS scan range was m z 1. Spectra were collected at 6 scans s 1 using scan and selective ion monitoring (SIM) simultaneously. Electron multiplier voltage was set to 1000 V. MS tuning was performed using the default autotune setting using perfluorotributylamine (PFTBA) weekly. A human panelist assessed odor of each of the separated compounds (at the sniff port) simultaneously with chemical analyses. Odor caused by separated VOCs was evaluated with a 64-descriptor odor character panel, odor intensity scale, and odor hedonic tone scale with Aromatrax (Microanalytics, TX, USA) software summarizing this information in aromagrams Air sampling Sampling sorbent tubes were constructed of 304-grade stainless steel and then double passivated with a proprietary surface-coating process. Then, tubes were packed with 65 mg Tenax TA. Silanized glass wool plugs and stainless steel screens were placed in the two ends of the tubes to hold the sorbent. Before the first use, sorbent tubes were conditioned by thermal desorption (260 C for 5 h) under a 100 ml min 1 flow of N 2. For subsequent uses, pre-conditioning at 260 C for 30 min was tested as sufficient and applied for all tubes. Field air samples were taken using a portable sampling pump with a set flow rate of 70 ml min 1 for 1 h, followed by storage at 4 C, and analysis within seven days. The sampling flow rates were checked with a NIST-traceable digital flow meter (Bios International, Butler, NJ, USA) Standards and calibration Fifteen compounds were selected as the target VOCs for quantification in this work. The selection was based on previous studies relative to typical odorous VOCs emitted from livestock facilities (Table 1) [1,3 7]. Sulfur VOCs were not quantified due to system limitations. Standard solutions for gas standards were prepared by diluting stock standard solutions in methanol and were stored at 4 C in the dark. Stock standard solutions of VFAs and phenolics were prepared by adding known masses of pure chemicals into a 40 ml pre-cleaned vial, and then filled with a known mass of methanol. Detector (MS) response factors for target VOCs were determined by direct injection of 1.0 L of standard solution onto the GC column and measuring recovery of each odorant. Calibration of sorbent tubes was performed as follows. 5 L or 10 L of the standard solution was spiked into a sorbent tube using an ATIS TM adsorbent tube injector system (Supelco, Bellefonte, PA, USA). A N 2 flow of 50 ml min 1 for 5 min with the block heater Table 1 Summary of target odorous compounds quantified in this study, with the method s linear range, and method detection limits. No. Compounds MW Retention time (min) MS ion a Linear range (ng) MDLs (ng) MDLs b (ppbv) 1 Acetic acid , 60, Propanoic acid , 28, Isobutanoic acid , 27, Butanoic acid , 27, Isopentanoic acid , 43, Pentanoic acid , 73, Hexanoic acid , 73, Guaiacol , 124, Heptanoic acid , 73, Phenol , 66, p-cresol , 77, Ethylphenol , 122, Aminoacetophenone , 135, Indole , 90, Skatole , 77, a Quantification ions shown in italics. b Values derived from mass-based MDLs (for 4.2 L air sample, 298 K, 1 atm).

3 S. Zhang et al. / Sensors and Actuators B 146 (2010) Table 2 Sensory analysis of a gas calibration standard (RSD = relative standard deviation). No. Compounds Mass (ng) Odor character Odor intensity (%) Hedonic tone RSD (%) 1 Acetic acid Acidic Propanoic acid Fatty acid, body odor Isobutyric acid Body odor, fatty acid Butyric acid Body odor, fatty acid Isovaleric acid Body odor, fatty acid Valeric acid Body odor, acidic, spicy Hexanoic acid Acidic, spicy Guaiacol 59.7 Burnt, medicinal, phenolic Heptanoic acid Acidic, spicy Phenol Burnt, phenolic p-cresol 68.9 Barnyard, medicinal, phenolic Ethylphenol 56.0 Burnt, phenolic Aminoacetophenone 88.9 Taco shell, medicinal, phenolic, sweet Indole 42.6 Medicinal, taco shell, barnyard, sweet Skatole 34.1 Taco shell, medicinal, sweet, barnyard temperature set to 75 C was optimized and used to transfer target VOCs onto the sorbent tubes. Table 1 summarizes the method detection limits (MDLs) for target VOCs. 3. Results and discussion Quantification of odorous VOC concentrations, odor intensity, character, and hedonic tone was performed simultaneously using the TD MDGC MS/O system. Target compounds were separated in the GC column and isolated compounds were split into MS and the sniff port with the mass split ratio of 1:3. The concentrations of compounds were quantified with the MS detector, and the odor character, intensity, odor event duration time (i.e., time from start to end of detected odor), and hedonic tone was identified and quantified via the sniff port by the panelist (Table 2). Fig. 1 shows the chromatogram and aromagram of a standard calibration gas sample with 15 typical odorous VOCs. The aromagram represents panelist response to separated compounds eluting from the GC column. Peaks represent recorded odor intensity on the scale of 0 100% [1,3 5]. It was observed, that with the increase of compound boiling point (and GC column retention time), the detection of odor associated with specific compounds was increasingly delayed compared Fig. 1. Chromatogram and aromagram of 15 VOCs. (1) Acetic acid; (2) propanoic acid; (3) isobutanoic acid; (4) butanoic acid; (5) isopentanoic acid; (6) pentanoic acid; (7) hexanoic acid; (8) guaiacol; (9) heptanoic acid; (10) phenol; (11) p-cresol; (12) 4-ethylphenol; (13) 2-aminoacetophenone; (14) indole; (15) skatole.

4 430 S. Zhang et al. / Sensors and Actuators B 146 (2010) with the GC column retention time. Thus, in order to quantify the odor event accurately, it was important to train the panelist to correctly separate each odor event. The separation of odors caused by higher boiling point compounds such as 2-aminoacetophenone, 4-ethylphenol, indole and skatole was optimized as follows. The GC MS O analysis of a single compound was performed and compared with the same analysis of the 15 VOCs gas mixture. It was found that the odor events for these four VOCs with the highest boiling points overlapped. This was likely caused by condensation of VOCs in the sniff port. The odor events for the remaining 11 VOCs were completely separated. Thus, the separation of odor events for the four high boiling point VOCs was optimized and set to occur at constant GC column retention times. This was done to analyze and compare odor characteristics between all experiments. Method detection limit (MDL) was determined using standard U.S. EPA methodology [12]. The MDLs were defined as the minimum concentration of a substance that can be measured and reported with 99% confidence when the analyte concentration is greater than zero and is determined from analysis of a sample in a given matrix containing the analyte. The MDLs for our method are listed in Table 1. The MDLs were generally lower than those reported in other previous studies [13,14]. Precision tests were conducted by consecutive analysis of three tubes spiked with the same amount of a standard work solution. Values of repeatability (% relative standard deviation (RSD) values) are reported in Table 2. All of the odorants had reasonable repeatabilities <20% that meets USEPA method performance criteria [15]. To examine VOC breakthrough, two tubes were connected in series into the standards spiking system. Individual analysis of each tube showed that no significant breakthrough (measured as % VOCs in the second-in-series tube) was observed for most of the standard VOCs. Breakthrough was only observed for low molecular weight compounds: acetic acid (maximum 34.7%, mean 27.7%), propanoic acid (maximum 24.0%, mean 21.5%), and isobutanoic acid (maximum 5.6%, mean 4.8%). This was due to the weaker sorption capacity of Tenax TA to low molecular weight VFAs. We also investigated the correlation of measured (at the sniff port) odor intensities associated with each separated target VOC with the VOC masses introduced with a sample. We found that the mass of each VOCs correlate well with the log stimulus intensity (Fig. 2). All of the coefficients of determination (R 2 ) for linear fits were greater than 0.74, and the top 10 R 2 s were greater than This finding was consistent with the Weber Fechner Law, which is referred to as one of fundamental psychophysical laws [16]: odor intensity (VOC) = m log(voc mass in sample) + b Fig. 2. Correlation between the odor intensity and the mass of 15 typical VOCs in one tube. For many odorants used in the food and fragrance industry, there is a linear relationship between log olfactory intensity reported by the individual panelist and the air concentration of the odorant present in air [17]. Zahn et al. [9,10] also reported that the total air concentration of VOCs emitted from swine manure correlate well with the log stimulus intensity. Finally, this method was used to measure the odor emitted from swine and dairy sites. Fig. 3 shows the chromatograms and aromagrams for field samples collected at a swine site (Fig. 3A) and a dairy site (Fig. 3B). Target odorous compounds were labeled and the corresponding VOC concentrations and odor characteristics are shown in Table 3. Typical odorous compounds emitted from a swine site were p-cresol, propanoic acid, butyric acid, and acetic acid. The typical odorous compounds emitted from dairy site were acetic acid, propanoic acid, butyric acid, and p-cresol. Although most of the typical odorous compounds from swine and dairy sites were similar, measured concentration of those from swine sites were much Table 3 Summary of typical odorous compounds quantified for samples collected from swine and dairy barn exhaust. No. Compounds Swine site Dairy site Concentration (ppb) Odor I (%) Hedonic tone Concentration (ppb) Odor I (%) Hedonic tone 1 Acetic acid Propanoic acid Isobutyric acid N/D N/D 4 Butyric acid Isovaleric acid Valeric acid N/D N/D 7 Hexanoic acid N/D N/D N/D N/D 8 Guaiacol N/D N/D N/D N/D N/D N/D 9 Heptanoic acid N/D N/D N/D N/D 10 Phenol N/D N/D N/D N/D 11 p-cresol Ethylphenol N/D N/D 13 2-Aminoacetophenone N/D N/D N/D N/D N/D 14 Indole N/D N/D N/D 15 Skatole(3-methylindole) N/D N/D N/D N/D N/D N/D, not detected; I, odor intensity.

5 S. Zhang et al. / Sensors and Actuators B 146 (2010) Fig. 3. Chromatogram and aromagram of typical air samples from swine barn (A) and dairy barn (B). (1) Acetic acid; (2) propanoic acid; (3) isobutanoic acid; (4) butanoic acid; (5) isopentanoic acid; (6) pentanoic acid; (7) hexanoic acid; (9) heptanoic acid; (10) phenol; (11) p-cresol; (12) 4-ethylphenol; (13) 2-aminoacetophenone; (14) indole. higher. Note that measured odor intensities (aromagram range of 0 100%) from both sites were of similar magnitude, which is consistent with the fundamental psychophysical law. Additional VOCs were also identified in air samples shown in Fig. 3, e.g., toluene, hexanal, heptanal, styrene, 3-octanone, limonene, nonanal, and decanal. These compounds were less odorous and/or had positive hedonic tone (scale used from 4 to +4), and were not apparently contributing to the characteristic overall odor (Table 3). 4. Conclusions The TD MDGC MS/O system can be used to estimate concentrations of VFAs and phenolic compounds associated with odorous gas emissions from CAFOs. Odor character, odor intensity, and odor hedonic tone can be assessed for separated target compounds simultaneously with chemical analyses. Concentrations of odorous compounds correlated well with the measured log stimulus intensity. Field samples from swine and dairy barns confirmed that target compounds quantified included those contributing to the overall characteristic odor. Acknowledgments The authors would like to thank the USDA for supporting this work via following grants: USDA-CSREES grant # Odor emission and chemical analysis of odorous compounds from animal buildings and USDA-CSREES grant # Mass transfer modeling validation for gas and odor emissions from manure storages and lagoons. References [1] Y.C.M. Lo, J.A. Koziel, L.S. Cai, S.J. Hoff, W.S. Jenks, H.W. Xin, Simultaneous chemical and sensory characterization of VOCs/semi-VOCs emitted from swine manure using SPME GC MS O, JEQ 37 (2008) [2] S.S. Schiffman, J.L. Bennett, J.H. Raymer, Quantification of odors and odorants from swine operations in North Carolina, Agric. For. Meteorol. 108 (2001) [3] J.A. Koziel, L.S. Cai, D.W. Wright, S.J. Hoff, Solid-phase microextraction as a novel air sampling technology for improved GC olfactometry-based assessment of livestock odors, J. Chromatogr. Sci. 44 (2006) [4] E.A. Bulliner, J.A. Koziel, L.S. Cai, D. Wright, Characterization of livestock odors using steel plates, solid-phase microextraction, and multidimensional GC MS O, JAWMA 56 (2006)

6 432 S. Zhang et al. / Sensors and Actuators B 146 (2010) [5] L.S. Cai, J.A. Koziel, Y.C. Lo, S.J. Hoff, Characterization of volatile organic compounds and odorants associated with swine barn PM using SPME and GC MS O, J. Chromatogr. A 1102 (2006) [6] K.M. Keener, J. Zhang, R.W. Bottcher, R.D. Munilla, Evaluation of thermal desorption for the measurement of artificial swine odorants in the vapor phase, Trans. ASAE 45 (2002) [7] L.L. Oehrl, K.M. Keener, R.W. Bottcher, R.D. Munilla, K.M. Connelly, Characterization of odor components from swine housing dust using gas chromatography, Appl. Eng. Agric. 17 (2001) [8] N.E. Rabaud, S.E. Ebeler, L.L. Ashbaugh, G.R. Flocchini, The application of thermal desorption GC/MS with simultaneous olfactory evaluation for the characterization and quantification of odor compounds from a dairy, J. Agric. Food Chem. 50 (2002) [9] J.A. Zahn, A.A. DiSpirito, D.A. Laird, Y.S. Do, B.E. Brooks, E.E. Cooper, J.L. Hatfield, Correlation of human olfactory responses to airborne concentrations of malodorous VOCs emitted from swine effluent, J. Environ. Qual. 30 (2001) [10] J.A. Zahn, J.L. Hatfield, D.A. Laird, T.T. Hart, Y.S. Do, A.A. DiSpirito, Functional classification of swine manure management systems based on effluent and gas emission characteristics, JEQ 30 (2001) [11] J.A. Zahn, J.L. Hatfield, D.A. Laird, T.T. Hart, Y.S. Do, A.A. DiSpirito, Assessing the relationship between concentrations of malodor compounds and odor scores from judges, J. Am. Dent. Assoc. 136 (2005) [12] Analytical Detection Limit Guidance & Laboratory Guide for Determining Method Detection Limits, OUTREACH/-Publications/LOD%20Guidance%20Document.pdf (accessed ). [13] S.L. Trabue, K.D. Scoggin, H. Li, R. Burns, H.W. Xin, Field sampling method for quantifying odorants in humid environments., ES&T 42 (2008) [14] A. Ribes, G. Carrera, E. Gallego, X. Roca, M.J. Berenguer, X. Guardino, Development and validation of a method for air-quality and nuisance odors monitoring of VOCs using multi-sorbent adsorption and GC MS TD system, J. Chromatogr. A (2007) [15] U.S. EPA, Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient Air, Method TO-17, Center for Environmental Research Information, Office of Research and Development, U.S. EPA, [16] V. Audouin, F. Bonnet, Z.M. Vickers, G.A. Reineccius, Limitations in the use of odor activity values to determine important odorants in foods, in: J.V. Leland, P. Schieberle, A. Buettner, T.E. Acree (Eds.), Gas Chromatography Olfactometry: The State of the Art, Oxford University Press, New York, 2001, pp [17] A. Turk, A.M. Hyman, Odor measurement and control, in: G.D. Clayton, F.L. Clayton (Eds.), Patty s Industrial Hygiene and Toxicology, John Wiley & Sons, New York, 1991, pp Biographies Shicheng Zhang received his PhD in 2001 from Northeastern University, China. He is currently an associate professor of environmental engineering at Fudan University, Shanghai, China. His research interests are mainly in the areas of environmental catalysis, odor emission measurement and mitigation, indoor air quality and bioenergy. Lingshuang Cai received her PhD in Analytical Chemistry at Wuhan University, China. She manages the Atmospheric Air Quality Laboratory at Iowa State University. She conducts air quality measurements and develops novel analytical methods for quantification of VOCs and semi-vocs with multidimensional GC MS O. Jacek Koziel, PhD (Univesity of Texas at Austin), is an associate professor at Iowa State University. His research focus is on air quality engineering and livestock odor, chemical and sensory analysis of gases, biotechnology, food science and forensics. He received the Wolfgang Göpel Memorial Award, sponsored by the Göpel family and the Institute of Physical Chemistry (University of Tübingen, Germany) for the best presentation at the 2009 International Symposium on Olfaction and Electronic Nose, Brescia, Italy. Steven Hoff, PhD, PE, received his PhD in 1990 from the University of Minnesota. He is currently a professor of Agricultural and Biosystems Engineering at Iowa State University. His research interests are mainly in the areas of animal housing, sensors and instrumentation, air quality, and odor dispersion and modeling from animal production systems. David Schmidt, MS, PE, is a research associate in the Department of Bioproducts and Biosystems Engineering at the University of Minnesota. His primary research efforts have been in the measurement of odor and gas emissions from livestock facilities and the development of techniques to mitigate these emissions. Charles Clanton, PhD, PE, received his PhD in 1985 from the University of Minnesota. He is currently a professor of Bioproducts and Biosystems engineering at the University of Minnesota. His research interests are mainly in the areas of manure management associated with water quality, air quality and odors, pathogens, storage, and land application. Larry Jacobson, PhD, PE, is a professor and extension engineer and livestock housing specialist in the Department of Bioproducts & Biosystems Engineering at the University of Minnesota. Special research and extension emphasis areas are: air quality and emissions for both animal and human health concerns, environmentally safe manure management systems, and economical facilities for livestock and poultry. He is a licenced Professional Engineer (P.E.) in the State of Minnesota. David Parker, PhD, PE, is a professor of Environmental Science and Engineering at West Texas A&M University in Canyon, Texas. He is a member of the Feedlot Research Group, where he specializes in odor and VOC measurement and control at animal feeding operations. Albert Heber, PhD, PE, is a professor in Agricultural and Biological Engineering at Purdue University. His primary research is the assessment and mitigation of air emissions from confined livestock facilities.