Parallel analysis of volatile fatty acids, indole, skatole, phenol, and trimethylamine from waste-related source environments

Journal of Chromatography A, 1314 (2013) 241–248

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Parallel analysis of volatile fatty acids, indole, skatole, phenol, and trimethylamine from waste-related source environments Md Mahmudur Rahman 1 , Ki-Hyun Kim ∗ Department of Environment and Energy, Sejong University, 98 Gun-Ja Dong, Gwang-Jin Gu, Seoul 143-747, Republic of Korea

a r t i c l e

i n f o

Article history: Received 25 July 2013 Received in revised form 6 September 2013 Accepted 10 September 2013 Available online 14 September 2013 Keywords: Gas chromatography Thermal desorber Mass spectrometry Response factor Odorants Volatile fatty acids

a b s t r a c t An experimental technique based on sorbent tube–thermal desorption–gas chromatography (ST–TD–GC) was investigated for the simultaneous determination of a cluster of eight volatile odorants (propionic acid, n-butyric acid, i-valeric acid, n-valeric acid, trimethylamine, phenol, indole, and skatole) and a reference compound (benzene). Calibration was made by direct injection of a liquid working standard (L-WS) into a quartz tube packed with three bed sorbent (Tenax TA, Carbopack B, and Carbopack X). To assess the relative performance between different detector systems, a comparative analysis was made using both mass spectrometry (MS) and a flame ionization detector (FID) with the aid of a TD system. In the TD–GC–MS analysis, calibration results were evaluated in two different modes, namely total ion chromatogram (TIC) and extracted ion chromatogram (EIC). In both FID and MS, the elution order of investigated odorants complied with the retention time index (RTI) values for the polar column with a coefficient of determination (R2 ) at or above 0.99. As a means to validate our detection approach, environmental samples from a bathroom and manhole (vacuum samples) as well as cat stool and wastewater (headspace samples) were also collected. The ST–TD method tested for the concurrent analysis of diverse odorants allowed us to measure a list of offensive odorants from those samples. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Odor, which refers to objectionable smells, is now considered an important environmental pollution phenomenon [1]. In our communal life, citizens are concerned about the presence of unusual odors because of their probable health effects, the long-term degradation of their surrounding environment, and the financial impact on their property [2–5]. A list of offensive volatile odorants (e.g., sulfur containing compounds, ammonia, amines, aldehydes, volatile fatty acids (VFA)) have been identified as produced from such environmental sources as storm water catch basins [6], lake sediments [7], food decay [8], charcoal combustion [9], livestock manure and feces [10–12], and municipal waste treatment plants [13]. In light of the environmental significance of air pollution, the emission of odorants, if made in excess quantity, is subject to administrative regulation in many countries (e.g., Japan and Korea [14]). As part of such regulation efforts, a list of offensive odorants such as several volatile fatty acids (VFA: propanoic, butyric, iso-valeric, and valeric acid) are designated as target offensive

∗ Corresponding author. Tel.: +82 02 499 9151; fax: +82 02 3408 4320. E-mail addresses: [email protected], [email protected] (K.-H. Kim). 1 Now at: International Laboratory for Air Quality and Health, Queensland University of Technology, Brisbane, QLD 4001, Australia. 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.09.035

odorants for tight administrative regulation [14]. Along with VFA, many other odorants (such as phenol, indole, skatole, and amines) are also found from many strong source environments such as waste water treatment facilities and animal feed operations [15–18]. In fact, due to the presence of indole and skatole, boar taint in pork often poses a problem when meat products are heated by the consumer prior to consumption [19–21]. In light of the high abundance of these odorants in many polluted mediums, an accurate technique to simultaneously measure a wide range of odorants is in great demand (Table 1). To investigate the pollution status of volatile odorants in ambient air, the gas chromatography–mass spectrometry (GC–MS) method has been commonly used as a reliable tool for analysis. Although many pretreatment options are available for the analysis of various odorants, their specifications for GC–MS analysis need to be established further to attain the maximum reliability. In this respect, the sorbent tube (ST) method combined with the thermal desorption (TD) technique is highly advantageous to induce the optimization of the GC-based quantitation for VOC [22]. However, as its application is often found to be case-specific, its feasibility needs to be validated further to apply to different target compounds [23]. To help resolve such limitations, many attempts have been made to simultaneously collect a suite of odorant homologues [24]. However, only a limited number of studies have been successful in producing and validating the data with sufficient confidence. In addition, sources of experimental biases in previous studies have

242

M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248

Table 1 Basic information of eight odorants and one reference compound selected as target compounds in this study. Order

Group

Full name

Short name

CAS number

Reagent purity (%)

MW (g mol−1 )

Density (g cm−3 )

Formula

Conc. (ng ␮L−1 )c

1

Amine

Tri methyl amine

TMA

75-50-3

25

59.1

0.93

(CH3 )3 N

233,000

2 3 4 5

Fatty acid

Propionic acid n-Butyric acid i-Valeric acid n-Valeric acid

PPA BTA IVA VLA

79-09-4 107-92-6 503-74-2 109-52-4

99 99 99 99

74.1 88.1 102 102

0.99 0.96 0.93 0.93

C3 H6 O2 C4 H8 O2 C5 H10 O2 C5 H10 O2

980,100 949,905 915,750 920,700

6 7 8 9

Aromatic

Phenola Indolea Skatolea Benzeneb

PHN IND SKT B

108-95-2 120-72-9 83-34-1 71-43-2

100 99 98 100

94.1 117 131 78.1

1.07 1.17 1.11 0.88

C6 H6 O C8 H7 N C9 H9 N C6 H6

37,125 37,125 37,125 872,118

a b c

As these compounds are in solid form, primary standard was prepared individually by dissolving 0.15 g into 4 mL of methanol. Benzene was analyzed as a reference compound. Concentration = purity (%) × density.

also been demonstrated experimentally, e.g., loss of VFA due to the use of stainless steel tube [25]. In this study, the applicability of the ST–TD–GC–MS was tested for the simultaneous quantification of various volatile odorants including VFA, indole, skatole, phenol, and TMA by following the optimum sampling procedure recommended for their collection [25]. Our study is thus meaningful in that simultaneous quantitation of different odorant groups such as VFA and amine compound is made with a highly reliable sampling approach. In addition, the compatibility of TD–GC–MS for such application was also investigated against TD–GC–FID. To assess the reliability of these approaches, analysis of all target odorants was also made from five types of environmental air samples (e.g., bathroom, manhole, cat stool, and wastewater). For the analysis of these odorant samples, the sorbent tube method was employed using a quartz tube packed with three bed sorbent (Tenax TA + Carbopack B + Carbopack X). The results of this analysis were evaluated to account for the feasibility of the ST-TD method for the analysis of key odorant species released from some strong source environments.

2. Materials and methods 2.1. Preparation of working standards A total of nine compounds including eight odorants (and one reference compound (benzene)) were selected in this study as the target compounds (Table 1): (1) tri methyl amine (TMA), (2) propionic acid (PPA), (3) n-butyric acid (BTA), (4) i-valeric acid (IVA), (5) n-valeric acid (VLA), (6) Phenol (PHN), (7) indole (IND), (8) skatole (SKT), and (9) benzene (B). To validate the feasibility of the ST–TD–GC–MS method for all the target odorants, the raw chemicals were purchased at a purity of 25% (for TMA), 98% (for SKT), 99.5% (PHN and B), and 99% for the rest (Sigma–Aldrich, USA). Concentrations of all primary grade compounds are shown in Table 1. In order to prepare the liquid phase primary standard (PS) of the three raw compounds (PHN, IND, and SKT) obtained in solid phase, 0.15 g of each were dissolved into three separate vials (4 mL) containing methanol (MeOH: J.T. Baker). As shown in Table 2, PS for each of three chemical groups

Table 2 Preparation of liquid-phase VOC standard for the sorbent tube analysis by GC–FID and GC–MS. A. Step-1: primary standard (PS) Compounds

Group-1 (All in liquid phase) Reagent volume (␮L) Solvent (␮L)a Concentration (ng ␮L−1 )

Amine

Fatty acid

TMA

PPA

250 1250 38,833

61

Compounds

Reference BTA

IVA

63 1246 39,896

39,857

VLA

B

65

65

39,683

39,897

70 1430 40,699

Odorants PHN

(2) Group-2 (purchased as solid) Reagent mass (g) Solvent (␮L) Concentration (ng ␮L−1 )

IND

0.15 4000 37,125

SKT

0.15 4000 37,125

0.15 4000 37,125

B. Step-2 Compounds

BTA

IVA

VLA

PHN

IND

SKT

B

(1) First stage working standard (L-WS) Concentration (ng ␮L−1 ) 485 498 499 Methanol as solvent (␮L) = 3700 PS volume (␮L) = 50 × 6 = 300

496

499

464

464

464

509

Order

TMA

L-WS (␮L)

PPA

Solvent (␮L)

(2) Second stage (final) working standard (final L-WS) 30 1470 1 60 1440 2 150 1350 3 300 1200 4 a

Methanol as solvent.

Concentration (ng ␮L−1 ) 9.7 19.4 48.5 97.1

10 19.9 49.8 99.6

10 19.8 49.9 99.7

10 19.8 49.6 99.2

10 19.9 49.9 99.7

9.3 18.6 46.4 92.8

9.3 18.6 46.4 92.8

9.3 18.6 46.4 92.8

10.2 20.3 50.9 101.7

M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248

was prepared independently in 1.5 mL MeOH. Then their liquid phase working standards (WS) were prepared by mixing four types of PS through a two-step gravimetric dilution. To prepare the first L-WS in the second step, a total of 300 ␮L (6 individual vials × 50 ␮L of each PS) was mixed with 300 ␮L of MeOH solvent (Table 2). In the next step, by further diluting this first L-WS, the final L-WS for each target compound was prepared for four point calibration (the mass of each compound at approximately 10, 20, 50, and 100 ng ␮L−1 ). At last, 1 ␮L volume of the final L-WS was withdrawn and used for the TD–GC calibration analysis across all four concentration levels. 2.2. Collection of environmental air samples In order to simultaneously collect samples of all target odorants from diverse environmental media, a quartz tube prepared as a three-bed type (50 mg of Tenax TA, Carbopack B, and Carbopack X: namely, TBX) to induce optimal adsorption of odorants (and VOCs) with a wide range of volatility; breakthrough was not observed under the experimental setups tested in this study. It was found that stainless steel material has been used most commonly as sorbent packing tubing materials for the collection of VOC. However, in this study, quartz-based sorbent tube was used exclusively to avoid the sorptive loss problem associated with the collection of gaseous VFA on stainless steel. In our recent study, we demonstrated that the use of stainless steel should be avoided and replaced with such material as quartz to minimize or eliminate such problem [25]. The collection of odorants investigated in this study was hence made using a quartz tube sampler. The collection of sorbent tube samples was made by employing the two different procedures: (i) collection of two headspace samples from a waste water treatment tank and a cat stool (with impinger) and (ii) vacuum sampling of ambient air onto ST from a bathroom and a manhole site. To collect the dynamic headspace sample, 10 g of a cat stool was put into an impinger (after softening the cat stool with 20 mL distilled water). Then, ultrapure air from a cylinder was brought into the impinger as a sweep gas through one end at a flow rate of 100 mL min−1 . The exiting gas was then collected into a Polyester Aluminum (PEA) bag (5 L capacity). The headspace sample collected in PEA was then absorbed onto sorbent tubes for the TD–GC analysis. (For this second-hand sampling, we did not consider the actual loss of target compounds due to the sorptive loss in the bag sampler.). In the case of wastewater samples, 200 mL of headspace above the waste water reactor chamber were collected directly onto the sorbent tube. Likewise, 1 L of ambient air was collected onto the tube at the bathroom and manhole site. All sorbent tubes used for the collection of each sample were subsequently analyzed by TD–GC–MS in the laboratory. 2.3. Instrumental analysis In this study, the GC systems equipped with GC/MS (SHIMADZU GCMS-QP2010, Japan) and GC/FID (IGC-7200, Donam Inc., Korea) were employed for comparative calibration of all target odorants. The quartz tubes used for the collection of the standard (or samples) were desorbed using a TD system equipped with an electrically cooled focusing trap (UNITY, Markes International Ltd., UK). It should be noted that each detection system (MS and FID) was equipped with the same TD system mentioned above for comparison. The TD focusing trap was packed with an equivolume ratio of Carbopack B and Tenax TA. The sorbent tubes for sampling were prepared in the laboratory by packing with 50 mg of Tenax TA (60/80 mesh, Restek, USA), Carbopack B (60/80 mesh, Superclo, USA), and Carbopack X (40/60 mesh, Superclo, USA). These sorbents were packed in the order to allow for quantitative recovery of the target odor compounds in consideration of their reactivities. For the calibration analysis, 1 ␮L

243

of the final L-WS was injected into the sorbent tube with a supply of ultrapure N2 gas, as described in our previous study [22]. The tubes loaded with samples/standard were subsequently desorbed at 310 ◦ C in the TD system and detected by each detector system. Detailed information regarding the instrumental setup is summarized in Table 3. Target odorants separated by both GC were measured by MS and FID for comparative purposes. During the software analysis of the GC–MS method, peak areas were calculated in two different modes, namely total ion chromatogram (TIC) and extracted ion chromatogram (EIC) (Fig. 1). In the EIC mode, the response of individually selected ions can be identified. In contrast, the TIC mode facilitates the analysis of total ion count representing a particular mass range. An actual chromatogram of L-WS used for the calibration (3rd point) analysis (48.5 (TMA)–50.9 (B) ng) is shown in Fig. 1. The response factor (RF) values of all target compounds were obtained in both the TIC and EIC modes (Fig. 2). The concentration of TMA was calculated only in the EIC mode due to significant interference in the TIC mode (Fig. 2). In addition, although the relative performance of the liquid standard was tested by both FID and MS, the quantification of environmental samples was made only by TD–GC–MS to simplify the identification of selected odorants.

3. Results and discussion 3.1. Relative performance of MS and FID for the quantification of odorants In this study, the key performance characteristics of the GC–MS and GC–FID systems for important odorants were assessed with the aid of the TD technique. The calibration data for individual target odorants (and a reference compound) are compiled in terms of response factors (RF) and the coefficient of determination (R2 ) for both TD–GC–MS (in TIC and EIC mode) and TD–GC–FID (Fig. 2). The results of this comparison confirm that the RF values derived by the EIC mode are systematically lower than those of the TIC mode. It is also interesting to note that the magnitude of RF values between EIC mode (GC–MS) and GC–FID are in most cases highly comparable with each other despite the fact that their signals are integrated by the two independent systems (Table 4). According to our analysis, the elution order of investigated compounds were seen on the order of TMA, B, PPA, BTA, IVA, VLA, PHN, IND, and SKT, which complied well with the relative ordering of retention time index (RTI) values in the polar column: 679 [26], 938 [27], 1525 [28], 1628 [28], 1691 [29], 1698 [28], 2004 [30], 2450 [31], and 2494 [32], respectively. As mentioned earlier, the RF value of TMA was calculated only in the EIC mode of GC–MS because of co-eluting interferences from a sorbent tube in the TIC mode (in MS). Due to this interference problem, the quantification of TMA was made only by the TD–GC–MS. To calculate the RF values of each target compound in the EIC mode (by TD–GC–MS), the peak areas of individual compounds were calculated using the base peak of the mass spectrum (mass to charge ratio or (m/z)). A chromatogram of target odorants (∼50 ng) is shown in Fig. 1. The R2 value of the calibration data was found at around 0.99 for most compounds by the GC/MS analysis (both TIC and EIC mode). However, RF values obtained in the EIC mode were significantly smaller by about four times than those of TIC because only base peak of mass spectrum was considered in the former. The RF values of target compounds in the TIC mode ranged between 16,229 (PPA) to 66,088 (B). Likewise, GC–FID coupled with TD also yielded fairly reliable R2 values (about 0.99) (Fig. 2) with much a reduced RF value (on average 127%) relative to the TIC mode (GC–MS). The RF values of the FID system ranged between 9989 (PPA) to 32388 (B).

244

M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248

Table 3 Instrumental setup for the analysis of selected target odorants. A. GC/MS system for odorant/analysis (i) GC/MS (SHIMADZU GCMS-QP2010, Japan) (a) Oven setting 1st oven temp 1st oven rate 2nd oven temp. 2nd oven rate Final oven temp. Total time

60 ◦ C (0 min) 10 ◦ C min−1 100 ◦ C 15 ◦ C min−1 220 ◦ C (18 min) 30 min

(c) Column (Vocol, PA, USA) Column (Vocol, PA, USA) Length: Film thickness

0.32 mm 60 m 1.8 ␮m

(b) Detector (MS) Ionization mode Ion source temp. TIC scan range Threshold

EI (70 eV) 200 ◦ C 35–600 m/z 100

Trap low Trap high Flow path temperature

5 ◦C 320 ◦ C 120 ◦ C

Temp.

310 ◦ C

(b) Detector (FID) Temp. Air H2 N2

250 ◦ C 300 mL min−1 30 mL min−1 30 mL min−1

(ii) Thermal desorber (UNITY, Markers International Ltd., UK) Carbopack B + Tenax TA Cold trap 1:5 Split ratio 5 mL min−1 Split flow Trap hold time 10 min

Trap low Trap high Flow path temperature

5 ◦C 310 ◦ C 150 ◦ C

(iii) Sampling tube Absorbent Desorb. time

Temp.

310 ◦ C

(ii) Thermal desorber (UNITY, Markers International Ltd., UK) Carbopack B + Tenax TA Cold trap 1:5 Split ratio 5 mL min−1 Split flow 5 min Hold time (iii) Sampling tube Absorbent Desorb time

Tenax TA + Carbopack B + Carbopack X 10 min

B. GC/FID system for odorant analysis (i) GC/FID system (IGC-7200, Donam Inc., Korea) (a) Oven setting 60 ◦ C (0 min) 1st oven temp. 1st oven rate 10 ◦ C min−1 100 ◦ C 2nd oven temp. 2nd oven rate 15 ◦ C min−1 Final oven temp. 220 ◦ C (6 min) Total time 20 min (c) Column (CP Wax) Column (Vocol, PA, USA) Length Film thickness

0.25 mm 30 m 1.8 ␮m

Tenax TA + Carbopack B + Carbopack X 10 min

Table 4 Basic quality assurance of target odorant analyzed by both TD–GC–MS and TD–GC–FID. Order

Compounds

MDLa

RSEb (%)

MS

FID

TIC Mass (ng) 1 2 3 4 5 6 7 8 9 a b c d

PPA BTA IVA VLA PHN IND SKT B TMA

0.79 1.24 0.98 0.75 0.78 0.51 0.53 0.27 –

EIC Conc. (ppb)c 0.26 0.34 0.23 0.18 0.2 0.11 0.1 0.08 –d

Mass (ng)

Conc. (ppb)

0.12 1.50 0.29 0.30 0.70 0.05 0.17 0.85 0.33

0.04 0.42 0.07 0.07 0.18 0.01 0.03 0.27 0.1

MS

FID

Mass (ng)

Conc. (ppb)

TIC

EIC

6.95 2.86 5.12 4.32 5.36 5.56 6.56 2.69 –

3.91 1.29 2.13 2.46 1.82 3.4 2.53 1.23 –

3.78 3.11 1.95 3.87 1.79 2.98 2.47 1.53 –

3.41 1.12 2.28 0.99 1.53 2.25 2.4 1.28 1.5

Method detection limit (MDL) was calculated by 7 replicate analysis of 10 ng standards of each sample. Relative standard error (RSE) was calculated from three replicate injections (about 10 ng) of individual sample. Concentration values in molar ratio terms were computed assuming 1 L of total sampling volume for each target compound. Concentration values of TMA were calculated in EIC mode only due to interference problem.

2.95 1.06 1.49 2.61 1.2 3.03 2.4 3.24 –

M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248

245

Fig. 1. Chromatograms of investigated odorants: (A) standard injection of about 50 ng in TIC mode (a) TMA, (b) B, (c) PPA, (d) BTA, (e) IVA, (f) VLA, (g) PHN, (h) IND, and (i) SKT, (B) 50 ng standard injection in EIC mode (refer to Table 3 for absolute mass of the third calibration point), and (C) headspace of cat stool sample.

3.2. Quality assurance (QA) of TD–GC coupled with MS and FID In order to assess the relative performance of TD–GC setup between the two detectors (MS and FID), the basic quality assurance (QA) parameters were evaluated in terms of the method detection limit (MDL) and precision (e.g., reproducibility expressed in terms of a relative standard error (% RSE) of measurements). These QA parameters were determined by seven and three replicate analyses of the lowest calibration point of the final L-WS (about 10 ng), respectively (Fig. 2). In the case of GC–MS, MDL and RSE were calculated in both TIC and EIC. In both detector systems (MS and FID), the QA parameters were quantified in an identical manner. As with the case of the RF calculations, QA parameters for TMA were compiled in the EIC mode of GC–MS. In all cases, RSE tended to fall below 5%. The MDL value for each odorant was derived by multiplying the SD of seven replicate analyses with 3.14 (Student’s t-value at the 99.9% confidence interval) and then divided by the relevant RF value [24]. In the TD–GC–MS analysis, the MDL (in TIC mode) values were found in between 0.27 ng (or 0.08 ppb) for B to 1.24 ng (or 0.34 ppb) for BTA. (Here for the conversion of MDL from mass (ng) into molar ratio (ppb), a total sampling volume of 1 L was assumed for each target compound.) In contrast, MDL in the EIC mode was found on average three times lower than the TIC counterpart. In TD–GC–FID,

the MDL values were found in between 2.69 ng (or 1.23 ppb) for B and 6.95 ng (or 3.91 ppb) for PPA, which were on average about 15 times higher than GC–MS counterparts (in TIC mode) (Fig. 2). 3.3. Environmental sample analysis for target odorants and comparison with previous studies Considering the abundance of odorants between different environmental reservoirs, many scientists have attempted to develop diverse measurement techniques. Such methods have been employed in various studies to derive quantitative data sets of various odorants. In order to validate the applicability of the ST-TD method, a total of five air samples were collected from four different environmental setups for TD–GC–MS analysis (Table 5). To measure the concentrations of all target odorants, samples from the bathroom and manhole were collected on sorbent tubes via mini-vacuum pump and subsequently analyzed by TD–GC–MS. In contrast, in the case of the cat stool and wastewater, their headspace samples collected on the sorbent tube were analyzed by the GC–MS system. It should be noted that the concentrations of target compounds were calculated in both TIC and EIC modes (Table 2). Among investigated odorants, PPA was found more frequently in the concentration range of 0.58 ppb (manhole) to around 20 ppb

246 M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248

Fig. 2. Comparison of calibration results of investigated target odorants between TD–GC–MS (TIC vs. EIC) and TD–GC–FID.

M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248

247

Table 5 Concentration of odorants investigated in the five different environmental samples collected in this studya . Order

1 2 3 4 5 a b c d e f

Sample name

Bathroom Cat stoold Manhole Waste water-1 Waste water-2

Sample code

BT CS MH WA1f WA2f

Investigated compounds TMAb

PPA

BTA

IVA

VLA

PHN

IND

SKT

0.61 3.47 0.10e 0.10 0.10

0.94/3.51c 13.9/14.5 0.58/0.44 17.9/25.6 22.5/14.6

0.34/0.42 24.92 0.34/0.42 19.6/13.9 14.7/11.6

0.23/0.07 9.88/9.55 0.23/0.07 11.2/6.95 8.52/5.6

0.18/0.07 2.51/3.74 0.18/0.07 6.33/4.04 3.98/3.39

0.50 1.57/1.56 0.20/0.18 0.20/0.18 0.20/0.18

0.11/0.01 7.62/7.81 0.11/0.01 0.11/0.01 0.11/0.01

0.1/0.03 0.1/0.03 0.1/0.03 0.1/0.03 0.1/0.03

Total sampling volume loaded on sorbent tube (Tenax TA + Carbopack B + Carbopack X) was 1 L. Concentration calculated by EIC mode only. Comparison between TIC mode/EIC mode concentration. 5 L headspace sample was collected in Polyester Aluminum (PEA) bags by passing ultrapure air as sweep gas through an impinge system at flow rate of 100 mL min−1 . All underlined sample concentration values imply below MDL value. 200 mL headspace air of waste water treatment tank collected directly on a 3-bed sorbent tube.

(wastewater). Interestingly, almost all odorants were detected successfully from the cat stool samples, except for SKT, which was found below the detection limit in almost all cases. In our waste water sample, four VFAs were identified as the dominant component. The reliability of our experimental approach to quantify a suite of odorants can be assessed further through comparison of previous studies. It is found that a number of studies have been made previously to measure most or part of our target odorants. However, the performance of experimental approaches used by other researchers cannot be assessed directly in most cases due to limited availability of detailed QA procedure, especially the bias associated with loss of VFA due to the selection of sorbent tube material (e.g., stainless steel) [25]. Although the data collected from a number of studies can be compared with our data, it is suspected that many of those data may be subject to underestimation of VFA due to such problem. In an investigation of odorants from the air surrounding cattle feedlots, Trabue et al. [17] found most of the volatile odorants (investigated in our study) in the sub ppbv level (e.g., BTA, PHN, IND, and SKT: consistently below one ppbv). Considering the abundance of those odorants in livestock operation, Zhang et al. [33] developed a method to analyze 15 odorants in TD–GC–MS/olfactometry (O). Although they completed their calibrations successfully, they did not validate their methods from real sample analysis. Like Zang et al. [33], Sato et al. [34] investigated odor compounds from the dynamic headspace air of human waste (feces and urine) collected from a septic tank using Tenax TA sorbent material. According to their study, the concentrations of many target odorants also included in this study were seen at much higher levels (e.g., by about three order of magnitude) than ours: TMA (0.80–1.20 ppm), PPA (5.30–27 ppm), BTA (1.50–9.20 ppm), IVA (0.53–2.62 ppm), VLA (0.41–1.60 ppm), IND (0.02–0.35 ppm), and SKT (0.1–0.48 ppm). Schiffman et al. [10] also conducted the measurements of odorants emitted from swine operation in North Carolina, US. They identified a total of 331 VOCs from swine facilities and reported that concentrations of several organic acids (BTA (4 ppb) and IVA (15 ppb)) and indole (0.1 ppb) were above the standardized odor threshold level. Furthermore, in a similar swine facility area, indole and skatole were also designated as the odorants with the lowest known odor threshold [10]. 4. Conclusion In this study, a method to simultaneously measure a total of eight odorants was investigated using the GC method coupled with a TD system. The relative performance of the two detector systems ((i) mass spectrometry (MS) and (ii) flame ionized detector (FID)) was also tested for their applicability to a suite of key offensive odorants. Prior to analysis, samples of both liquid standard and

environmental matrix were absorbed onto three bed sorbent tubes made of quartz (Tenax TA + Carbopack B + Carbopack X). Although TMA was co-eluted with certain interfering compounds (possibly from sorbent materials), its quantification was completed by TD–GC–MS (EIC mode). Overall, both TD–GC–MS and FID were found as potent tools to quantify most of the selected target compounds reliably (R2 = ≥0.98). In light of the abundance of VFA, indole, skatole, and TMA, wellknown offensive odorants in certain polluted environmental media (e.g., swine facilities, cattle feedlots, municipal waste, and livestock), the applicability of ST–TD–GC experimental approach was investigated in this study. The reliability of our analytical method, when tested against four types of environmental media, suggests that our method can be used reliably for the quantification of most target compounds at sub-ppb levels. As a result, our analytical method can be used to acquire quantitative data for diverse offensive odorants in air samples under a variety of environmental settings. References [1] A. Yuwono, P.S. Lammers, Int. Comm. Agric. Eng. 6 (2004) 1. [2] J.R. Miner, Agric. Eng. 61 (1980) 22. [3] D. Bundy, National Livestock Poultry and Aquaculture Waste Management, ASAE Pub., St. Joseph, 1992, pp. 3. [4] D. Shusterman, Arch. Environ. Health 47 (1992) 76. [5] S.S. Schiffman, J.M. Walker, P. Dalton, T.S. Lorig, J.H. Raymer, D. Shusterman, C.M. Williams, J. Agromed. 7 (2000) 7. [6] E. Kabir, K.-H. Kim, J.-W. Ahn, O.-F. Hong, Y.S. Chang, Chemosphere 81 (2010) 327. [7] J. Susaya, K.H. Kim, Y.S. Chang, Atmos. Environ. 45 (2011) 1236. [8] K.-H. Kim, R. Pal, J.-W. Ahn, Y.H. Kim, Waste Manage. (Oxford) 29 (2009) 1265. [9] M.M. Rahman, K.-H. Kim, J. Hazard. Mater. (2012). [10] S.S. Schiffman, J.L. Bennett, J.H. Raymer, Agric. For. Meteorol. 108 (2001) 213. [11] L.S. Cai, J.A. Koziel, Y.C. Lo, S.J. Hoff, J. Chromatogr. A 1102 (2006) 60. [12] A. Shabtay, U. Ravid, A. Brosh, R. Baybikov, H. Eitam, Y. Laor, J. Anim. Sci. 87 (2009) 1835. [13] K.H. Kim, S.H. Jo, H.C. Song, S.K. Pandey, H.N. Song, J.M. Oh, Y. Sunwoo, K.C. Choi, Int. J. Environ. Sci. Technol. 10 (2013) 261. [14] K.-H. Kim, S.-Y. Park, Atmos. Environ. 42 (2008) 5061. [15] A. Banel, A. Jakimska, M. Wasielewska, L. Wolska, B. Zygmunt, Anal. Chim. Acta 716 (2012) 24. [16] S. Trabue, B. Kerr, B. Bearson, C. Ziemer, J. Environ. Qual. 40 (2011) 1510. [17] S. Trabue, K. Scoggin, L. McConnell, R. Maghirang, E. Razote, J. Hatfield, Atmos. Environ. 45 (2011) 4243. [18] D.W. Wright, D.K. Eaton, L.T. Nielsen, F.W. Kuhrt, J.A. Koziel, J.P. Spinhirne, D.B. Parker, J. Agric. Food. Chem. 53 (2005) 8663. [19] M. Bonneau, A.J. Kempster, R. Claus, C. Claudi-Magnussen, A. Diestre, E. Tornberg, P. Walstra, P. Chevillon, U. Weiler, G.L. Cook, Meat Sci. 54 (2000) 251. [20] F.M. Whittington, G.R. Nute, S.I. Hughes, J.D. McGivan, I.J. Lean, J.D. Wood, E. Doran, Meat Sci. 67 (2004) 569. [21] K. Verheyden, H. Noppe, M. Aluwé, S. Millet, J. Vanden Bussche, H.F. De Brabander, J. Chromatogr. A 1174 (2007) 132. [22] Y.-H. Kim, K.-H. Kim, Anal. Chim. Acta 780 (2013) 46. [23] Y.-H. Kim, K.-H. Kim, Anal. Chem. 84 (2012) 4126. [24] Y.-H. Kim, K.-H. Kim, Anal. Chem. 84 (2012) 8284. [25] Y.-H. Kim, K.-H. Kim, Anal. Chem. 85 (2013) 7818.

248 [26] [27] [28] [29] [30] [31]

M.M. Rahman, K.-H. Kim / J. Chromatogr. A 1314 (2013) 241–248 A. Giri, K. Osako, A. Okamoto, T. Ohshima, Food Res. Int. 43 (2010) 1027. R.G. Binder, R.A. Flath, T.R. Mon, J. Agric. Food. Chem. 37 (1989) 418. K.L. Goodner, LWT – Food Sci. Technol. 41 (2008) 951. L. Cullere, A. Escudero, J. Cacho, V. Ferreira, J. Agric. Food. Chem. 52 (2004) 1653. Y. Sekiwa, K. Kubota, A. Kobayashi, J. Agric. Food. Chem. 45 (1997) 826. K. Kumazawa, H. Masuda, J. Agric. Food. Chem. 47 (1999) 5169.

[32] P. Schnermann, P. Schieberle, J. Agric. Food. Chem. 45 (1997) 867. [33] S. Zhang, L. Cai, J.A. Koziel, S.J. Hoff, D.R. Schmidt, C.J. Clanton, L.D. Jacobson, D.B. Parker, A.J. Heber, Sens. Actuators B: Chem. 146 (2010) 427. [34] H. Sato, H. Morimatsu, T. Kimura, Y. Moriyama, T. Yamashita, Y. Nakashima, J. Health Sci. 48 (2002) 179.