- Email: [email protected]
Chemical characterization of odorous gases at a landfill site by gas chromatography–mass spectrometry
Journal of Chromatography A, 1122 (2006) 222–229
Chemical characterization of odorous gases at a landfill site by gas chromatography–mass spectrometry Faruk Dincer ∗ , Mustafa Odabasi, Aysen Muezzinoglu Department of Environmental Engineering, Dokuz Eylul University, Kaynaklar Campus, 35160 Buca-Izmir, Turkey Received 17 January 2006; received in revised form 12 April 2006; accepted 18 April 2006 Available online 12 May 2006
Abstract The composition of odorous gases emitted from a municipal landfill in the city of Izmir, Turkey was investigated using gas chromatography–mass spectrometry, and these data were examined in relation with the odor concentrations. Several volatile organic compounds (VOCs) were identified and quantified at five sampling sites in May and September 2005. Detected VOCs were monoaromatics (0.09–47.42 g m−3 ), halogenated compounds (0.001–62.91 g m−3 ), aldehydes (0.01–38.55 g m−3 ), esters (0.01–7.54 g m−3 ), ketones (0.03–67.60 g m−3 ), sulfur/nitrogen containing compounds (0.03–5.05 g m−3 ), and volatile fatty acids (VFAs) (0.05–43.71 g m−3 ). High levels of aldehydes (propanal up to 38.55 g m−3 ) and VFAs (formic acid up to 43.71 g m−3 ) were measured in May. However, VOC concentrations were relatively low in September. The monoaromatics and halogenated compounds were the abundant VOCs in landfill air for the both sampling periods. The benzene-to-toluene (B:T) ratio at the landfill site was significantly lower than urban areas indicating the presence of higher amounts of toluene in landfills compared to traffic exhaust rich urban areas. A statistically significant linear relationship was found between odor concentrations determined by olfactometry and total VOC concentrations. The relationships of odor concentrations with the different groups of chemicals were also examined using a step-wise multiple regression analysis. It was found that the concentrations of aldehydes, ketones, and esters are the best estimators, explaining 96% of the variability in odor concentrations (r2 = 0.96, n = 10, P < 0.01). © 2006 Elsevier B.V. All rights reserved. Keywords: Odor concentration; Olfactometry; Volatile organic compounds; Landfill odors; GC–MS; Benzene-to-toluene ratio
1. Introduction Municipal solid waste (MSW) landfills are potential sources of offensive odors creating annoyance in urban areas [1]. Odor pollution has become a growing concern during the last decades for urban communities located near or downwind of MSW landfills. The annoying odors released to the atmosphere from landfills may cause decreased quality of life and possibly more negative consequences on human health and welfare [2]. Landfill gases are generated naturally by anaerobic decomposition of wastes. Summer is the critical season for such processes as the higher temperatures and richer organic matter in the waste composition favor anaerobic processes leading to waste decomposition. Landfill gases are mainly composed of methane and carbon dioxide. They also include some non-methane volatile
∗
Corresponding author. Tel.: +90 232 453 0922; fax: +90 232 453 0922. E-mail address: [email protected] (F. Dincer).
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.04.075
organic compounds (VOCs) [3]. Although the amounts of VOCs are usually below 1% (by volume) of the total emissions, their adverse effects on the environment are not negligible. For example, a range of chlorofluorocarbon compounds contribute to both stratospheric ozone depletion and greenhouse effect [4]. Prolonged exposure to the landfill gases containing benzene, toluene, and xylenes (BTX) and chlorinated hydrocarbons can cause severe health problems especially on landfill operators [5]. Alkylbenzenes, limonene, certain esters, and organosulfur compounds are mainly responsible for odor nuisance [6]. Odor measurement techniques have been based on sensory analysis using human nose as a detector. Compared to the human nose, many of the chemical detectors are not as sensitive for the odor active compounds [7]. Olfactometry based on the human perception capacity is the most common method for measuring odor concentrations. In this method, odor measurement is carried out by presenting a sample of odorous air in a range of dilutions using neutral air to an independent panel of selected and trained persons, and statistically treating the responses from
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
the panelists for odor detection. Dilutions are prepared using an olfactometer, and results are evaluated according to certain protocols such as the CEN methodology [8]. Gas chromatography (GC) in connection with various detectors, especially mass spectrometry (MS) has been applied for the characterization of the chemical composition of odorous gas samples [9]. Davoli et al. [10] reported an analytical approach for characterization of odorous VOCs by solid phase microextraction (SPME) followed by GC–MS in landfills for a wide range of compounds (i.e., from highly polar volatile fatty acids to non-polar hydrocarbons). Zou et al. [5] conducted a study to characterize the ambient VOC levels in different seasons in a landfill in South China. They reported high levels of benzene, toluene, and chlorinated VOCs. However, their study did not cover odor concentrations. The relationship between odor and chemical concentrations was investigated by Defoer et al. [11] for vegetable, fruit, and garden composting plants and for the emissions of an animal rendering plant treated by bio-filtration. These studies compared the profiles of odor concentrations and chemical concentrations of compounds in different chemical classes. The characterization of VOCs [10] and odor assessment based on the dispersion of odors [1] at landfills were previously reported. However, there are no studies in the literature investigating the relationship between landfill odor and VOC concentrations. The objective of the present work was to investigate the relationship between landfill odor and VOC concentrations at the landfill site of Izmir, Turkey. Samples were collected at five sampling points in May and September 2005. Odor levels were measured by olfactometry and the chemical concentrations were determined by GC–MS following thermal desorption. Measured odor and VOC concentrations were analyzed statistically and discussed. 2. Materials and methods 2.1. Site description Harmandalı MSW site is the main landfill in Izmir. It has been in operation since 1990 and receives commercial, industrial, and municipal wastes. The capacity of the landfill is 3000 t day−1 of domestic, medical, and industrial wastes and wastewater treatment sludge. The site does not accept hazardous wastes. The field sampling was conducted on May 6 and September 1, 2005 to represent the beginning and end of the hot and dry summer season in the area having a typical Mediterranean climate. The average temperatures were 19 ◦ C in May and 25 ◦ C in September and the relative humidity was 55% in both May and September sampling days. Same five sampling points were visited at active burial areas in May and September sampling periods. Sites 1 and 2 were selected to represent the areas where thickened sludge from the Izmir municipal wastewater treatment plant and the medical wastes were disposed, respectively. The other three sites were at the domestic solid waste disposal area. Site 3 was the raw waste collection and classification area of
223
the municipal wastes. At the waste classification site there was heavy truck traffic and machinery operations were active. All three of the municipal waste burial areas were selected to represent different stages and ages of solid waste management. 2.2. Sampling and analysis Olfactometry and thermal desorption followed by GC–MS were used to determine the odor and volatile organic compound concentrations of the collected samples. 2.2.1. Sampling method Three samples were collected from underneath a specially designed hood at each sampling point from the air adjacent to the surface of the landfill covers. Sampling was carried out following recommendations described in the European Standard EN 13725 [8]. Air samples were drawn into 5 l Nalophan® bags using a special sampler working with the lung principle. The sampler draws the air directly into the bag by evacuating the tightly closed atmospheric pressure vessel in which it was placed. Odor and taste free bags that are impermeable to water and organics were used in sampling only for once. Polytetrafluoroethylene (PTFE) sampling tubes were used as input and connection lines. Samples were transported to the olfactometry laboratory and analyzed within 24 h. 2.2.2. Olfactometric analysis The odor concentrations of the samples were measured by dynamic olfactometry. The tests were carried out inside an odorfree, clean laboratory with selected and trained panelists. Each sample was diluted in the olfactometer (model TO7, ECOMA, Honigsee, Germany) several times differing from each other by a factor of two and presented to the panelists three times. Dilutions were made using odor-free air supplied by a compressor fitted with carbon filters and an air dryer. The olfactometer is a computer controlled semi-automatic instrument with four panel member places and computes the odor concentration by means of a special computer program based on the perception response data of panelists. This method employs a “yes/no” technique and determines how many times a sample must be diluted with odor-free air to be at the threshold of detection by 50% of the panel. At this instance the number of required dilution defines the odor concentration in odor units per cubic meter (OU m−3 ). 2.2.3. Chemical analysis Grab sampling by Nalophan® bags has been commonly used to collect gas samples. Samples collected in bags are either directly introduced to the analytical systems or concentrated using techniques like solid phase micro extraction (SPME) and adsorbent tubes-thermal desorption (TD). Some recent applications using grab sampling with Nalophan® bags include the analysis of odorous gas emissions from composting plants (gas sample-proton transfer reaction-mass spectrometry) [12], waste gases and landfill gases (SPME-GC–MS) [10,13,14], VOC emissions from apple tree (TD-GC–MS) [15], VOCs emissions from moss roses (solid–liquid extraction-GC–MS) [16], fig tree
224
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
emissions (solid–liquid extraction-GC–MS) [17], volatiles from maize (solid–liquid extraction-GC-FID) [18], mushroom composting emissions (TD-GC–MS) [19]. The use of Nalophan® bags for odor sampling is recommended by the European Standard EN 13725 [8] due to their inertness. A recent study indicated that gas mixtures containing ppbv concentrations of several chlorinated and aromatic hydrocarbons could be safely stored for several days in Nalophan® bags without any significant loss [20]. In the present study, gas samples were collected into Nalophan® bags and were transported to the laboratory. Then, gases were passed through the adsorbent tubes via silicone tubing connected to a vacuum pump (Rena 301, Rena OEM, France). Sampling flow rate and sample volume were 100 ml min−1 and 2–3 l, respectively. The flow rate was measured using a rotameter (Gilmont, Barnant Inc., USA). The rotameter calibration was checked occasionally (at three flow rates in duplicate, n = 6) using a primary standard (soap-bubble meter). The average percent difference between two flow meters was <2.5%. Duplicate tubes obtained from each sampling bag were separately analyzed to determine the concentrations of volatile fatty acids (VFAs) and volatile organic compounds, respectively. The sample tubes were refrigerated and analyzed within 1–3 days as recommended [21]. Glass adsorbent tubes (6 mm O.D., 17.8 mm length) were prepared in the laboratory according to the ambient air sampling methods recommended by USEPA [22]. Each tube was packed at the upstream (sampling) end with 3 mm silanized glass wool followed by a series of sections of 150 mg Tenax TA (60/80 mesh) (Supelco, Bellefonte, PA, USA), 3 mm silanized glass wool, 100 mg Carboxen 1000 (Supelco, Bellefonte, PA, USA) and finally 3 mm silanized glass wool at the downstream end. Filled tubes were conditioned at 260 ◦ C for 1 h with a 50 ml min−1 reverse flow (opposite to the sampling direction) of high purity nitrogen prior to use. Ends of the conditioned tubes were first closed with PTFE caps and each tube was placed into tightly capped special tube containers prior to use and after sampling. Silica gel and activated charcoal were placed to the bottom section of the tube containers for humidity and contamination control, respectively. Samples were analyzed with a gas chromatograph (GC) (Agilent 6890N, Agilent, Wilmington, DE, USA) equipped with a mass selective detector (Agilent 5973 inert MSD, Agilent, Wilmington, DE, USA) and a thermal desorber (Tekmar, Aerotrap 6000, USA). The thermal desorber was modified by replacing the original fused silica transfer line (0.32 mm I.D.) causing flow restriction through the GC inlet with a glass-lined inert steel tubing (0.7 mm I.D., Alltech). The cryogenic internal trap operating with liquid nitrogen was replaced with an adsorbent (100 mg Tenax, Supelco, Bellefonte, PA, USA) filled ambient trap. The cooling for the internal trap was provided by a vacuum pump drawing air at room temperature through the liquid nitrogen line at a rate of 15 l min−1 . VOC samples were desorbed for 5 min at 225 ◦ C using helium flow at the rate of 40 ml min−1 . Internal trap temperature during sample desorption was 35 ◦ C. The trap was desorbed for 1 min at 240 ◦ C. Then, it was baked for 10 min at 250 ◦ C. Valve oven and transfer line temperature of the thermal desorber was 200 ◦ C.
The chromatographic column was HP5-MS (30 m, 0.25 mm, 0.25 m) and the carrier gas was helium at 1 ml min−1 flowrate and 36 cm s−1 linear velocity. The split ratio was 1:40. The inlet temperature was 240 ◦ C. Temperature program for VOCs was: initial oven temperature 40 ◦ C, hold 3 min, 40–120 ◦ C at 5 ◦ C min−1 , hold 1 min. Temperature program for VFAs was: initial oven temperature 50 ◦ C, hold 1 min, 50–225 ◦ C at 35 ◦ C min−1 , hold 2.5 min. Ionization mode of the MS was electron impact (EI). Ion source, quadrupole, and GC/MSD interface temperatures were 230, 150, and 280 ◦ C, respectively. The MSD was run in selected ion monitoring mode. Compounds were identified on the basis of their retention times (within ±0.05 min of the retention time of calibration standard), target, and qualifier ions. Identified compounds were quantified using the external standard calibration procedure. Five levels (0.4, 2, 5, 10 and 20 g ml−1 ) of VOC solutions were prepared in methanol as the calibration standards. Five levels of standard solutions with varying VFAs (i.e., 12, 30, 60, 180, and 300 g ml−1 for acetic acid) were prepared in deionized (DI) water. Thermal desorption tubes used for calibration (150 mg Tenax TA + 100 mg Carboxen 1000) were loaded by spiking with 1 l of the calibration standards. Then, the standard loaded tubes were run at specified conditions to calibrate the analytical system (Thermal desorber-GC–MS) [22]. In all cases linear fit was good with r2 > 0.99. 2.3. Quality control and quality assurance Instrumental detection limits for VOCs (∼5 pg for a split ratio of 1:40) were determined by linear extrapolation from the lowest standard in the calibration curve using the area of a peak having a signal/noise ratio of 3. Six blank thermal desorption tubes (three for VFAs and three for VOCs) for each sampling period were analyzed as process blanks in order to determine the level of contamination during sample handling and preparation. The limit of detection (LOD, pg) of the method was defined as the mean blank mass plus three standard deviations (LOD = mean blank + 3SD) [23–27]. LOD values ranged between 5-15230 pg and 5-3843 pg for May and September sampling periods, respectively. Instrumental detection limits were used for the compounds that were not detected in blanks. In general, VOCs in the samples were substantially higher than the blanks. Sample quantities exceeding the LODs were quantified and corrected by subtracting the mean blank amount from the sample amount. For VFA analyses, no blanks were detected and therefore blank correction was not necessary. For three samples a back-up tube was connected in series with the sample tube during sampling from the Nalophan® bags to check if there was breakthrough. Back-up tubes were found to contain similar quantities of the compounds as the blanks indicating that the breakthrough from the sample tubes was not a problem during sampling. The system performance was confirmed daily by analyzing a midrange calibration standard. The relative standard deviation from the initial calibration was <10%. Analytical precision determined from three pairs of duplicate samples ranged between 2 and 5%.
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
3. Results and discussion
Table 1 VOC concentrations (ranges (g m−3 ) and mean values) at Izmir landfill
3.1. VOC composition and concentrations In both sampling periods the odorous emissions from the landfill have contained several compounds. A total of 53 VOCs were found in May while 48 VOCs in September sampling periods (Table 1). The measured VOCs were classified as monoaromatics (e.g. benzene, toluene), halogenated compounds (e.g. chlorobenzene, trichloroethene), aldehydes (e.g. hexanal, propanal), ketones (e.g. acetone), VFAs (e.g. acetic acid, formic acid), esters (e.g. butyl acetate, butyl formate), and S and N compounds (carbon disulfide and acrylonitrile). In May, the most abundant group was ketones (25% of total VOC concentrations) followed by monoaromatics (21%), aldehydes (20%), volatile fatty acids (17%), halogenated compounds (14%), esters (2%), and S and N compounds (1%). In September, aldehydes had the highest concentration (37% of total VOC concentrations) and they were followed by ketones (36%), monoaromatics (13%), halogenated compounds (6%), VFAs (5%), S and N compounds (2%), and esters (1%). Monoaromatics had significant concentrations in both May and September samples (Fig. 1). Toluene had the highest average concentration in this group and was exceeded only by formic acid and propanal in both sampling campaigns. Benzene-to-toluene (B:T) ratio has been commonly used as an indicator of traffic emissions. A B:T ratio of 0.5 was reported from studies on vehicle exhaust [28]. Recent urban air measurements have also showed B:T ratios ranging between 0.27 and 0.5 [29–31] (Table 2). Therefore, B:T ratios around 0.5 may indicate that the ambient VOC concentrations are mainly affected by traffic emissions. However, landfills have significantly lower B:T ratios (i.e., 0.1 in USA) [32]. Benzene-to-toluene ratios measured in this study were 0.015 and 0.11 in May and September, respectively indicating the presence of higher amounts of toluene than benzene in landfills compared to traffic exhaust rich urban areas (Table 2). Chlorinated organics in the landfill might be related to the solvents that are widely used as cleaners and disinfectants in urban and industrial sources. Sixteen chlorinated compounds were detected in May and September samples. The average concentrations of some chlorinated species such as chloroform, carbon tetrachloride, chlorobenzene, trichloroethene, and tetrachloroethene were 0.08, 0.17, 0.04, 13.06, and 2.37 g m−3 , respectively (Fig. 1). Tetrachloroethylene is present in the textiles, dry cleaning workplaces, and chemical manufacturing units. Trichloroethene is used as a degreaser, extraction, and cleaning solvent for household, commercial, and industrial uses. Thus, their residuals might have reached the landfill along with the domestic, commercial, and industrial wastes. Compared to monoaromatic compounds, the concentrations of chlorinated species were relatively low in May and September samples. In May, all VFAs were detected only at the medical waste burial point of the landfill. Formic and isobutyric acids were the most abundant volatile fatty acids in the landfill burial points with the concentrations of 24.22 and 3.56 g m−3 , respectively.
225
Formic acid Acetic acid Propionic acid Isobutyric acid Butyric acid Isovaleric acid Valeric acid Isocaproic acid Caproic acid Heptanoic acid Acrolein Butanal Crotonaldehyde Decanal Heptanal Hexanal Nonanal Octanal Pentanal Propanal Butyl acetate Butyl formate Butyl propionate Methyl propionate Vinyl acetate Benzene Ethylbenzene o,m-Xylene p-Xylene Styrene Toluene 1,1,1-Trichloroethane 1,1,2,2-Tetrachloroethane 1,1,2-Trichloroethane 1,1-Dichloroethane 1,1-Dichloroethene 1,2-Dichlorobenzene 1,2-Dichloroethane 1,2-Dichloropropane 1,3-Dichlorobenzene 1,4-Dichlorobenzene Bromodichloromethane Bromoform Carbon tetrachloride Chlorobenzene Chloroform cis-1,3-Dichloropropene cis-1,4-Dichloro-2-butene Dibromochloromethane Iodomethane Methylene chloride Tetrachloroethene trans-1,2-Dichloroethene trans-1,3-Dichloropropene trans-1,4-Dichloro-2-butene Trichloroethene 2-Butanone 2-Hexanone 4-Methyl-2-pentanone (MIBK) Acetone Cyclohexanone Carbon disulfide Acrylonitrile Pyridine
May
September
3.44–43.71 (24.22) 0.17–5.34 (1.92) 0.19–3.52 (1.86) 0.50–6.61 (3.56) 0.06–4.39 (2.22) 0.05–6.05 (2.34) 0.07–4.99 (1.86) 2.84a 1.37–2.91 (2.14) 0.25–8.85 (3.39) 0.48–2.02 (1.04) 0.43–1.70 (1.01) 0.05–0.38 (0.14 0.66–1.49 (0.97) 0.16–1.51 (0.73) 0.47–5.94 (2.59) 0.64–2.16 (1.57) 0.27–2.28 (1.54) 0.15–1.92 (0.75) 4.16–38.55 (21.13) 0.12–7.54 (2.70) 0.03–0.10 (0.06) 0.11a N.D. 0.01–2.29 (0.51) 0.09–0.59 (0.29) 0.21–4.94 (2.03) 0.35–10.16 (4.33) 0.34–9.55 (3.59) 0.67–14.44 (3.88) 1.56–47.42 (18.97) 0.02–0.05 (0.04) N.D. 0.05–0.08 (0.06) 0.0014–0.01 (0.006) 0.12–0.37 (0.27) 0.03–0.08 (0.05) 0.01–1.22 (0.30) N.D. 0.002–0.01 (0.005) 0.08–0.40 (0.25) N.D. N.D. 0.12–0.23 (0.17) 0.01–0.12 (0.04) 0.03–0.16 (0.08) 0.03–0.07 (0.05) N.D. N.D. 0.001–0.02 (0.02) 1.62–7.95 (4.42) 0.05–9.16 (2.37) N.D. 0.085–0.09 (0.09) N.D. 0.18–62.91 (13.06) N.D. 0.04–0.80 (0.27) 0.03–0.42 (0.21) 7.95–67.60 (37.17) 0.08–9.13 (3.15) 0.41–5.05 (1.51) 0.09–0.20 (0.14) N.D.
11.10a N.D. N.D. N.D. N.D. 0.20a 0.37a N.D. 0.14–0.34 (0.22) 0.30–1.30 (0.73) 1.25–2.66 (1.83) 0.34–1.33 (0.59) 0.01–0.15 (0.09) 1.25–9.42 (3.95) 0.18–0.64 (0.32) 0.45–3.55 (1.32) 1.28–5.01 (2.64) 0.53–1.42 (0.82) 0.26–0.87 (0.44) 5.15–13.45 (8.49) 0.09–0.42 (0.21) 0.03–0.12 (0.05) 0.10a 0.18a 0.08–1.79 (0.65) 0.34–1.06 (0.53) 0.15–0.76 (0.45) 0.22–1.11 (0.73) 0.15–1.01 (0.62) 0.13–0.40 (0.24) 1.51–11.18 (4.76) 0.02–0.10 (0.04) N.D. N.D. 0.02a 0.04–0.70 (0.22) 0.02–0.09 (0.04) 0.01a N.D. 0.02a 0.07–0.21 (0.14) N.D. N.D. 0.10–0.15 (0.13) 0.009–0.014 (0.01) 0.03–0.13 (0.06) N.D. N.D. N.D. 0.02–0.03 (0.02) 0.68–2.58 (1.43) 0.02–1.02 (0.50) 0.11a N.D. 0.004a 0.11–1.95 (0.81) N.D. 0.06–0.22 (0.10) 0.05–0.18 (0.08) 11.20–28.57 (19.94) 0.004a 0.24–2.36 (1.15) 0.03–0.04 (0.034) N.D.
Data in parenthesis represent the average concentrations; N.D.: not detected. a Identified only in one sampling site.
226
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
Fig. 1. Seasonal variations of concentrations of monoaromatics, halogenated compounds, aldehydes and ketones, and VFAs.
In September, the concentration of formic acid was relatively lower and it could be detected at one site only. High levels of aldehydes and ketones were observed at each sampling location in May and September with the maximum levels at sampling sites 3 and 4. Propanal has a pungent, suffocating and unpleasant odor characterizing the landfill gas emissions. Propanal was found above the odor threshold of 3.6 g m−3 [33] at all sampling points. Some esters and organic S and N compounds were also identified and quantified in this study but the concentrations of these compounds were relatively low compared to other VOC groups. For example, concentrations of butyl acetate were 2.70 and 0.21 g m−3 in May and September, respectively. Carbon disulfide that has a characteristic odor (vegetable sulfide/medicinal) had relatively low average concentrations (1.51 g m−3 in May
Table 2 Comparison of average benzene and toluene levels (g m−3 ) and B:T ratios in ambient air at selected locations Benzene
Toluene
B:T
Site
Reference
0.23 0.53
18.97 4.76
0.015 0.11
Izmir (landfill) Izmir (landfill)
3.31 37.6 52.8 33.8
15.39 102 528 62
0.27 0.37 0.1 0.5
Izmir (urban) Izmir (urban) USA (landfill) Guangzhou (urban)
This study (May) This study (September) [29] [31] [32] [30]
and 1.15 g m−3 in September). Acrylonitrile concentrations were even lower (0.034–0.14 g m−3 ). Recent studies have shown that reduced sulfur compounds are important trace components of landfill gases [34–36]. Thus, sulfur and nitrogen containing organic compounds that are not measured in this study may also be significant in odor characterization. 3.2. Odor and total VOC concentrations The results of olfactometric and chemical analysis for the two sampling periods are given in Table 3. The total VOC concentrations presented in Table 3 are the sum of concentrations of detected compounds. For May samples, olfactometric concentrations varied between 1416 and 116,027 OU m−3 with a mean value of 47,886 OU m−3 and the total VOC concentrations varied between 69 and 258 g m−3 (average, 156 g m−3 ). For September samples the olfactometric and total VOC concentrations ranged between 1070 and 111,980 OU m−3 (average, 29,684 OU m−3 ) and 43–101 g m−3 (average, 56 g m−3 ), respectively. Total VOC concentrations in May and September have shown similar patterns. In September total VOC concentrations were relatively lower at all sampling locations with slightly decreasing odor concentrations (Fig. 2, Table 3). A significant decrease in odor concentration was observed only at Site 5. This decrease was probably due to the fact that during the May sampling there was an ongoing waste burial while the site was covered with soil during the September sampling.
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
227
Table 3 Concurrent data sets of the olfactometric odor concentrations (OU m−3 ) and chemical concentrations (g m−3 ) for May and September samples Sampling point
Odor concentration
Total VOC concentration
May 1 2 3 4 5
3765 1416 116027 35928 82292
± ± ± ± ±
153 219 21939 2050 3215
118 69 258 207 127
± ± ± ± ±
September 1 2 3 4 5
3380 1070 111980 29270 2720
± ± ± ± ±
400 425 20817 2309 800
56 36 101 43 44
± ± ± ± ±
Acids
Aldehydes
Esters
Halogenated compounds
Monoaromatics
Ketones
S and N compounds
10 2 13 11 9
0.11 45 38 45 1
18 9 48 34 48
1 0.17 10 5 0.42
74 2 17 5 7
6 4 76 67 13
19 8 68 50 52
1.2 0.6 0.6 0.6 5.2
5 2 6 2 4
N.D. 1 12 2 0.30
19 14 37 18 14
0.26 0.19 1 1 2
5 1 6 3 2
6 5 15 8 3
25 12 29 11 23
1.0 1.9 2.4 0.27 0.29
Fig. 2. Variations of odor and total VOC concentrations in May and September samples.
An odor control chemical (consisted of nearly 20% calcium hydroxide) applied onto the landfill sites (Izmir Metropolitan Municipal Waste Management Authority, 2005) during the period starting with mid-April and ending in September 2005. This application has focused mainly on sludge and medical
waste burial areas (Site 1 and 2). The burial areas were covered with 1–3 cm of the chemical every third to fourth day. May and September samplings at the landfill sites coincided with 1 month after the beginning and end of this application period. Therefore, it is possible that the composition of the odorous gas samples was affected by this application. However, the effect of the chemical application is not significant on odor concentrations since there was only a slight decrease at all sampling sites except Site 5 that was possibly affected by the soil cover (Table 3). VFA concentrations were very low and some of them could not be detected in September in contrast to the fact that most of them were found at all sampling sites in May (Fig. 1). The VFA is a good indicator of waste decomposition process. Lower VFA concentrations observed in September may be due to decreased decomposition rate as a result of odor control chemical or due to the sorption of the VFAs onto the alkaline chemical. However, the decrease in other non-biogenic VOCs (i.e., monoaromatic and halogenated compounds) concentrations in September cannot be explained by the application of odor control chemical since they are not end products of decomposition or they are not likely to be sorbed by the chemical. The decrease in concentrations of these VOCs was probably due to their lower content in the wastes as a result of loss by evaporation during a long and hot period.
Table 4 Pearson correlation coefficients between the different compound groups and odor (n = 10)
Odor Total VOCs Acids Aldehydes Esters Halogenated compounds Monoaromatics Ketones S and N compounds *
Odor
Total VOCs
Acids
Aldehydes
Esters
Halogenated compounds
Monoaromatics
Ketones
S and N compounds
1
0.64* 1
0.23 0.63 1
0.91* 0.76* 0.21 1
0.52 0.87* 0.61 0.58 1
−0.10 0.23 −0.18 −0.04 0.02 1
0.55 0.93* 0.69* 0.66* 0.94* −0.03 1
0.73* 0.88* 0.37 0.91* 0.79* −0.02 0.83* 1
0.41 0.02 −0.33 0.53 −0.30 −0.03 −0.18 0.29 1
Statistically significant (P < 0.05).
228
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
3.3. Relationship between the odor and total VOC concentration The characterization of VOCs at landfills [10] and odor assessment on the basis of the dispersion of odors [1] were previously reported. However, previous studies did not investigate the relationship between landfill odor and VOC concentrations. The relationship between concentrations of odors and total VOCs measured in the present study was analyzed statistically. A correlation matrix was calculated for different variables and Pearson coefficients were given as a measure of relationship between groups of variables (Table 4). There are statistically significant correlations (P < 0.05, n = 10) between odor concentrations and total VOC, aldehyde and ketone concentrations (Table 4). Linear regression analysis performed on the overall data set (May + September) between odor and total VOC concentrations indicated that 41% of the variance in odor concentrations can be explained by the total VOC concentrations (r2 = 0.41, n = 10, P < 0.05). The relationships of odor concentrations with the concentrations of different groups of chemicals were further examined using a step-wise multiple regression analysis. It was found that the concentrations of aldehydes, ketones, and esters are the best estimators, explaining 96% of the variability in odor concentrations (r2 = 0.96, n = 10, P < 0.01). 4. Conclusions The odor levels and the composition of odorous gases emitted from a municipal landfill site were studied. Up to 53 VOC compounds in different classes (mono aromatics, halogenated compounds, aldehydes, ketones, esters, and volatile fatty acids) were identified and quantified. The benzene-to-toluene ratio at the landfill site in May and September was found to be significantly lower than the values determined in urban areas indicating the presence of higher amounts of toluene than benzene in landfills compared to urban areas. Collected samples were also analyzed for odor concentrations by olfactometry. It is concluded that a relationship exists between odor and total VOC concentrations at the studied landfill site. The relationships of odor concentrations with the different groups of chemicals were also examined and it was found that the concentrations of aldehydes, ketones, and esters are the best estimators, explaining 96% of the variability in odor concentrations. Acknowledgement This work was supported in part by the research fund of Dokuz Eylul University (Project No. 03.KB.FEN.061) and by the Scientific and Technical Research Council of Turkey (TUBITAK) (Project No. ICTAG-A056). Olfactometric analyses were conducted in the laboratory established by the help of the EU project; “Odorous Emissions and Immissions Management Policy in Turkey-LIFE TCY/TR/000009”. Authors are thankful to panel members for their participation in olfactometry.
The help and permission by the Izmir Metropolitan Municipal Waste Management Authority should also be gratefully acknowledged.
References [1] U. Sarkar, S.E. Hobbs, P. Longhurst, J. Environ. Manage. 68 (2003) 153. [2] ATDSR, Agency for Toxic Substances and Disease Registry (http://www. atsdr.cdc.gov). [3] F. Kreith, Handbook of Solid Waste Management, McGraw-Hill, New York, 1995. [4] J.T. Houghton, G.J. Jenkins, J.J. Ephraums (Eds.), Climate Change: The IPCC Scientific Assessment, Cambridge University Press, Cambridge, 1991. [5] S.C. Zou, S.C. Lee, C.Y. Chan, K.F. Ho, X.M. Wang, L.Y. Chan, Z.X. Zhang, Chemosphere 51 (2003) 1015. [6] A. Muezzinoglu, Chemosphere 51 (2003) 245. [7] S.M. van Ruth, Biomol. Eng. 17 (2001) 121. [8] CEN, Air quality—Determination of odor concentration by dynamic olfactometry, European Standard EN 13725, Brussels, Belgium, 2003. [9] E. Smet, H. van Langenhove, I. De Bo, Atmos. Environ. 33 (1999) 1295. [10] E. Davoli, M.L. Gangai, L. Morselli, D. Tonelli, Chemosphere 51 (2003) 357. [11] N. Defoer, I. De Bo, H. van Langenhove, J. Dewulf, T. van Elst, J. Chromotogr. A 970 (2002) 259. [12] F. Biasioli, F. Gasperi, G. Odorizzi, E. Aprea, D. Mott, F. Marini, G. Autiero, G. Rotondo, T.D. M¨ark, Int. J. Mass Spectrom. 239 (2004) 103. [13] K.K. Kleeberg, Y. Liu, M. Jans, M. Schlegelmilch, J. Streese, R. Stegmann, Waste Manage. 25 (2005) 872. [14] P. Pierucci, E. Porazzi, M. Pardo Martinez, F. Adani, C. Carati, F.M. Rubino, A. Colombi, E. Calcaterra, E. Benfenati, Chemosphere 59 (2005) 423. [15] A. Vallat, H. Gu, S. Dorn, Phytochemistry 66 (2005) 1540. [16] J.C. Caisard, V. Bergougnoux, M. Martin, M. Mauriat, S. Baudino, Ann. Bot. 97 (2006) 231. [17] L. Grison-Pige, J.M. Bessiere, T.C.J. Turlings, F. Kjelberg, J. Roy, M.M. Hossaert-McKey, Funct. Ecol. 15 (2001) 551. [18] M.E. Hoballah, T.G. K¨ollner, J. Degenhardt, T.C.J. Turlings, OIKOS 105 (2004) 168. [19] R. Noble, P.J. Hobbs, A. Dobrovin-Pennington, T.H. Misselbrook, A. Mead, J. Environ. Qual. 30 (2001) 760. [20] M. Gorgenyi, J. Dewulf, H. Van Langenhove, Chromatographia 51 (2000) 461. [21] C. Peng, S. Batterman, J. Environ. Monitor. 2 (2000) 313. [22] US-EPA, Compendium Method TO-17, Determination of Volatile Organic Compounds in Ambient Air Using Active Sampling Onto Sorbent Tubes, Center for Environmental Research Information Office of Research and Development, Cincinnati, USA, 1999. [23] T. Harner, M. Shoeib, M. Diamond, M. Ikonomou, G. Stern, Chemosphere, in press. [24] G.O. Thomas, M. Wilkinson, S. Hodson, K.C. Jones, Environ. Pollut. 141 (2006) 30. [25] F. Wong, T. Harner, Q.T. Liu, M.L. Diamond, Environ. Pollut. 129 (2004) 387. [26] A. Sofuoglu, E. Cetin, S.S. Bozacioglu, G.D. Sener, M. Odabasi, Atmos. Environ. 38 (2004) 4483. [27] R. Lohmann, P.A. Brunciak, J. Dachs, C.L. Gigliotti, E. Nelson, D. Van Ry, T. Glenn, S.J. Eisenreich, J.L. Jones, K.C. Jones, Atmos. Environ. 37 (2003) 959–969. [28] P.A. Scheff, R.A. Wadden, Environ. Sci. Technol. 27 (1993) 617. [29] M. Odabasi, Unpublished results, 2005. [30] X.M. Wang, G.Y. Sheng, J.M. Fu, C.Y. Chan, L.Y. Chan, S.C. Lee, Atmos. Environ. 36 (2002) 5141. [31] A. Muezzinoglu, M. Odabasi, L. Onat, Atmos. Environ. 35 (2001) 753.
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229 [32] J. LaRegina, J.W. Bozzelli, R. Harkov, S. Gianti, Environ. Prog. 5 (1986) 18. [33] US-EPA, Reference Guide to Odor Thresholds for Hazardous Air Pollutants Listed in the Clean Air Act Amendments of 1990, Air Risk Information Support Center, USEPA Office of Research and Development, EPA/600/R-92/047, 1992.
229
[34] K.H. Kim, Y.J. Choi, E.C. Jeon, Y. Sunwoo, Atmos. Environ. 39 (2005) 1103. [35] Z.H. Shon, K.H. Kim, E.C. Jeon, M.Y. Kim, Y.K. Kim, S.K. Song, Atmos. Environ. 39 (2005) 4803. [36] S. Lee, Q. Xu, M. Booth, T.G. Townsend, P. Chadik, G. Bitton, Waste Manage. 26 (2006) 526.
Chemical characterization of odorous gases at a landfill site by gas chromatography–mass spectrometry Faruk Dincer ∗ , Mustafa Odabasi, Aysen Muezzinoglu Department of Environmental Engineering, Dokuz Eylul University, Kaynaklar Campus, 35160 Buca-Izmir, Turkey Received 17 January 2006; received in revised form 12 April 2006; accepted 18 April 2006 Available online 12 May 2006
Abstract The composition of odorous gases emitted from a municipal landfill in the city of Izmir, Turkey was investigated using gas chromatography–mass spectrometry, and these data were examined in relation with the odor concentrations. Several volatile organic compounds (VOCs) were identified and quantified at five sampling sites in May and September 2005. Detected VOCs were monoaromatics (0.09–47.42 g m−3 ), halogenated compounds (0.001–62.91 g m−3 ), aldehydes (0.01–38.55 g m−3 ), esters (0.01–7.54 g m−3 ), ketones (0.03–67.60 g m−3 ), sulfur/nitrogen containing compounds (0.03–5.05 g m−3 ), and volatile fatty acids (VFAs) (0.05–43.71 g m−3 ). High levels of aldehydes (propanal up to 38.55 g m−3 ) and VFAs (formic acid up to 43.71 g m−3 ) were measured in May. However, VOC concentrations were relatively low in September. The monoaromatics and halogenated compounds were the abundant VOCs in landfill air for the both sampling periods. The benzene-to-toluene (B:T) ratio at the landfill site was significantly lower than urban areas indicating the presence of higher amounts of toluene in landfills compared to traffic exhaust rich urban areas. A statistically significant linear relationship was found between odor concentrations determined by olfactometry and total VOC concentrations. The relationships of odor concentrations with the different groups of chemicals were also examined using a step-wise multiple regression analysis. It was found that the concentrations of aldehydes, ketones, and esters are the best estimators, explaining 96% of the variability in odor concentrations (r2 = 0.96, n = 10, P < 0.01). © 2006 Elsevier B.V. All rights reserved. Keywords: Odor concentration; Olfactometry; Volatile organic compounds; Landfill odors; GC–MS; Benzene-to-toluene ratio
1. Introduction Municipal solid waste (MSW) landfills are potential sources of offensive odors creating annoyance in urban areas [1]. Odor pollution has become a growing concern during the last decades for urban communities located near or downwind of MSW landfills. The annoying odors released to the atmosphere from landfills may cause decreased quality of life and possibly more negative consequences on human health and welfare [2]. Landfill gases are generated naturally by anaerobic decomposition of wastes. Summer is the critical season for such processes as the higher temperatures and richer organic matter in the waste composition favor anaerobic processes leading to waste decomposition. Landfill gases are mainly composed of methane and carbon dioxide. They also include some non-methane volatile
∗
Corresponding author. Tel.: +90 232 453 0922; fax: +90 232 453 0922. E-mail address: [email protected] (F. Dincer).
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.04.075
organic compounds (VOCs) [3]. Although the amounts of VOCs are usually below 1% (by volume) of the total emissions, their adverse effects on the environment are not negligible. For example, a range of chlorofluorocarbon compounds contribute to both stratospheric ozone depletion and greenhouse effect [4]. Prolonged exposure to the landfill gases containing benzene, toluene, and xylenes (BTX) and chlorinated hydrocarbons can cause severe health problems especially on landfill operators [5]. Alkylbenzenes, limonene, certain esters, and organosulfur compounds are mainly responsible for odor nuisance [6]. Odor measurement techniques have been based on sensory analysis using human nose as a detector. Compared to the human nose, many of the chemical detectors are not as sensitive for the odor active compounds [7]. Olfactometry based on the human perception capacity is the most common method for measuring odor concentrations. In this method, odor measurement is carried out by presenting a sample of odorous air in a range of dilutions using neutral air to an independent panel of selected and trained persons, and statistically treating the responses from
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
the panelists for odor detection. Dilutions are prepared using an olfactometer, and results are evaluated according to certain protocols such as the CEN methodology [8]. Gas chromatography (GC) in connection with various detectors, especially mass spectrometry (MS) has been applied for the characterization of the chemical composition of odorous gas samples [9]. Davoli et al. [10] reported an analytical approach for characterization of odorous VOCs by solid phase microextraction (SPME) followed by GC–MS in landfills for a wide range of compounds (i.e., from highly polar volatile fatty acids to non-polar hydrocarbons). Zou et al. [5] conducted a study to characterize the ambient VOC levels in different seasons in a landfill in South China. They reported high levels of benzene, toluene, and chlorinated VOCs. However, their study did not cover odor concentrations. The relationship between odor and chemical concentrations was investigated by Defoer et al. [11] for vegetable, fruit, and garden composting plants and for the emissions of an animal rendering plant treated by bio-filtration. These studies compared the profiles of odor concentrations and chemical concentrations of compounds in different chemical classes. The characterization of VOCs [10] and odor assessment based on the dispersion of odors [1] at landfills were previously reported. However, there are no studies in the literature investigating the relationship between landfill odor and VOC concentrations. The objective of the present work was to investigate the relationship between landfill odor and VOC concentrations at the landfill site of Izmir, Turkey. Samples were collected at five sampling points in May and September 2005. Odor levels were measured by olfactometry and the chemical concentrations were determined by GC–MS following thermal desorption. Measured odor and VOC concentrations were analyzed statistically and discussed. 2. Materials and methods 2.1. Site description Harmandalı MSW site is the main landfill in Izmir. It has been in operation since 1990 and receives commercial, industrial, and municipal wastes. The capacity of the landfill is 3000 t day−1 of domestic, medical, and industrial wastes and wastewater treatment sludge. The site does not accept hazardous wastes. The field sampling was conducted on May 6 and September 1, 2005 to represent the beginning and end of the hot and dry summer season in the area having a typical Mediterranean climate. The average temperatures were 19 ◦ C in May and 25 ◦ C in September and the relative humidity was 55% in both May and September sampling days. Same five sampling points were visited at active burial areas in May and September sampling periods. Sites 1 and 2 were selected to represent the areas where thickened sludge from the Izmir municipal wastewater treatment plant and the medical wastes were disposed, respectively. The other three sites were at the domestic solid waste disposal area. Site 3 was the raw waste collection and classification area of
223
the municipal wastes. At the waste classification site there was heavy truck traffic and machinery operations were active. All three of the municipal waste burial areas were selected to represent different stages and ages of solid waste management. 2.2. Sampling and analysis Olfactometry and thermal desorption followed by GC–MS were used to determine the odor and volatile organic compound concentrations of the collected samples. 2.2.1. Sampling method Three samples were collected from underneath a specially designed hood at each sampling point from the air adjacent to the surface of the landfill covers. Sampling was carried out following recommendations described in the European Standard EN 13725 [8]. Air samples were drawn into 5 l Nalophan® bags using a special sampler working with the lung principle. The sampler draws the air directly into the bag by evacuating the tightly closed atmospheric pressure vessel in which it was placed. Odor and taste free bags that are impermeable to water and organics were used in sampling only for once. Polytetrafluoroethylene (PTFE) sampling tubes were used as input and connection lines. Samples were transported to the olfactometry laboratory and analyzed within 24 h. 2.2.2. Olfactometric analysis The odor concentrations of the samples were measured by dynamic olfactometry. The tests were carried out inside an odorfree, clean laboratory with selected and trained panelists. Each sample was diluted in the olfactometer (model TO7, ECOMA, Honigsee, Germany) several times differing from each other by a factor of two and presented to the panelists three times. Dilutions were made using odor-free air supplied by a compressor fitted with carbon filters and an air dryer. The olfactometer is a computer controlled semi-automatic instrument with four panel member places and computes the odor concentration by means of a special computer program based on the perception response data of panelists. This method employs a “yes/no” technique and determines how many times a sample must be diluted with odor-free air to be at the threshold of detection by 50% of the panel. At this instance the number of required dilution defines the odor concentration in odor units per cubic meter (OU m−3 ). 2.2.3. Chemical analysis Grab sampling by Nalophan® bags has been commonly used to collect gas samples. Samples collected in bags are either directly introduced to the analytical systems or concentrated using techniques like solid phase micro extraction (SPME) and adsorbent tubes-thermal desorption (TD). Some recent applications using grab sampling with Nalophan® bags include the analysis of odorous gas emissions from composting plants (gas sample-proton transfer reaction-mass spectrometry) [12], waste gases and landfill gases (SPME-GC–MS) [10,13,14], VOC emissions from apple tree (TD-GC–MS) [15], VOCs emissions from moss roses (solid–liquid extraction-GC–MS) [16], fig tree
224
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
emissions (solid–liquid extraction-GC–MS) [17], volatiles from maize (solid–liquid extraction-GC-FID) [18], mushroom composting emissions (TD-GC–MS) [19]. The use of Nalophan® bags for odor sampling is recommended by the European Standard EN 13725 [8] due to their inertness. A recent study indicated that gas mixtures containing ppbv concentrations of several chlorinated and aromatic hydrocarbons could be safely stored for several days in Nalophan® bags without any significant loss [20]. In the present study, gas samples were collected into Nalophan® bags and were transported to the laboratory. Then, gases were passed through the adsorbent tubes via silicone tubing connected to a vacuum pump (Rena 301, Rena OEM, France). Sampling flow rate and sample volume were 100 ml min−1 and 2–3 l, respectively. The flow rate was measured using a rotameter (Gilmont, Barnant Inc., USA). The rotameter calibration was checked occasionally (at three flow rates in duplicate, n = 6) using a primary standard (soap-bubble meter). The average percent difference between two flow meters was <2.5%. Duplicate tubes obtained from each sampling bag were separately analyzed to determine the concentrations of volatile fatty acids (VFAs) and volatile organic compounds, respectively. The sample tubes were refrigerated and analyzed within 1–3 days as recommended [21]. Glass adsorbent tubes (6 mm O.D., 17.8 mm length) were prepared in the laboratory according to the ambient air sampling methods recommended by USEPA [22]. Each tube was packed at the upstream (sampling) end with 3 mm silanized glass wool followed by a series of sections of 150 mg Tenax TA (60/80 mesh) (Supelco, Bellefonte, PA, USA), 3 mm silanized glass wool, 100 mg Carboxen 1000 (Supelco, Bellefonte, PA, USA) and finally 3 mm silanized glass wool at the downstream end. Filled tubes were conditioned at 260 ◦ C for 1 h with a 50 ml min−1 reverse flow (opposite to the sampling direction) of high purity nitrogen prior to use. Ends of the conditioned tubes were first closed with PTFE caps and each tube was placed into tightly capped special tube containers prior to use and after sampling. Silica gel and activated charcoal were placed to the bottom section of the tube containers for humidity and contamination control, respectively. Samples were analyzed with a gas chromatograph (GC) (Agilent 6890N, Agilent, Wilmington, DE, USA) equipped with a mass selective detector (Agilent 5973 inert MSD, Agilent, Wilmington, DE, USA) and a thermal desorber (Tekmar, Aerotrap 6000, USA). The thermal desorber was modified by replacing the original fused silica transfer line (0.32 mm I.D.) causing flow restriction through the GC inlet with a glass-lined inert steel tubing (0.7 mm I.D., Alltech). The cryogenic internal trap operating with liquid nitrogen was replaced with an adsorbent (100 mg Tenax, Supelco, Bellefonte, PA, USA) filled ambient trap. The cooling for the internal trap was provided by a vacuum pump drawing air at room temperature through the liquid nitrogen line at a rate of 15 l min−1 . VOC samples were desorbed for 5 min at 225 ◦ C using helium flow at the rate of 40 ml min−1 . Internal trap temperature during sample desorption was 35 ◦ C. The trap was desorbed for 1 min at 240 ◦ C. Then, it was baked for 10 min at 250 ◦ C. Valve oven and transfer line temperature of the thermal desorber was 200 ◦ C.
The chromatographic column was HP5-MS (30 m, 0.25 mm, 0.25 m) and the carrier gas was helium at 1 ml min−1 flowrate and 36 cm s−1 linear velocity. The split ratio was 1:40. The inlet temperature was 240 ◦ C. Temperature program for VOCs was: initial oven temperature 40 ◦ C, hold 3 min, 40–120 ◦ C at 5 ◦ C min−1 , hold 1 min. Temperature program for VFAs was: initial oven temperature 50 ◦ C, hold 1 min, 50–225 ◦ C at 35 ◦ C min−1 , hold 2.5 min. Ionization mode of the MS was electron impact (EI). Ion source, quadrupole, and GC/MSD interface temperatures were 230, 150, and 280 ◦ C, respectively. The MSD was run in selected ion monitoring mode. Compounds were identified on the basis of their retention times (within ±0.05 min of the retention time of calibration standard), target, and qualifier ions. Identified compounds were quantified using the external standard calibration procedure. Five levels (0.4, 2, 5, 10 and 20 g ml−1 ) of VOC solutions were prepared in methanol as the calibration standards. Five levels of standard solutions with varying VFAs (i.e., 12, 30, 60, 180, and 300 g ml−1 for acetic acid) were prepared in deionized (DI) water. Thermal desorption tubes used for calibration (150 mg Tenax TA + 100 mg Carboxen 1000) were loaded by spiking with 1 l of the calibration standards. Then, the standard loaded tubes were run at specified conditions to calibrate the analytical system (Thermal desorber-GC–MS) [22]. In all cases linear fit was good with r2 > 0.99. 2.3. Quality control and quality assurance Instrumental detection limits for VOCs (∼5 pg for a split ratio of 1:40) were determined by linear extrapolation from the lowest standard in the calibration curve using the area of a peak having a signal/noise ratio of 3. Six blank thermal desorption tubes (three for VFAs and three for VOCs) for each sampling period were analyzed as process blanks in order to determine the level of contamination during sample handling and preparation. The limit of detection (LOD, pg) of the method was defined as the mean blank mass plus three standard deviations (LOD = mean blank + 3SD) [23–27]. LOD values ranged between 5-15230 pg and 5-3843 pg for May and September sampling periods, respectively. Instrumental detection limits were used for the compounds that were not detected in blanks. In general, VOCs in the samples were substantially higher than the blanks. Sample quantities exceeding the LODs were quantified and corrected by subtracting the mean blank amount from the sample amount. For VFA analyses, no blanks were detected and therefore blank correction was not necessary. For three samples a back-up tube was connected in series with the sample tube during sampling from the Nalophan® bags to check if there was breakthrough. Back-up tubes were found to contain similar quantities of the compounds as the blanks indicating that the breakthrough from the sample tubes was not a problem during sampling. The system performance was confirmed daily by analyzing a midrange calibration standard. The relative standard deviation from the initial calibration was <10%. Analytical precision determined from three pairs of duplicate samples ranged between 2 and 5%.
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
3. Results and discussion
Table 1 VOC concentrations (ranges (g m−3 ) and mean values) at Izmir landfill
3.1. VOC composition and concentrations In both sampling periods the odorous emissions from the landfill have contained several compounds. A total of 53 VOCs were found in May while 48 VOCs in September sampling periods (Table 1). The measured VOCs were classified as monoaromatics (e.g. benzene, toluene), halogenated compounds (e.g. chlorobenzene, trichloroethene), aldehydes (e.g. hexanal, propanal), ketones (e.g. acetone), VFAs (e.g. acetic acid, formic acid), esters (e.g. butyl acetate, butyl formate), and S and N compounds (carbon disulfide and acrylonitrile). In May, the most abundant group was ketones (25% of total VOC concentrations) followed by monoaromatics (21%), aldehydes (20%), volatile fatty acids (17%), halogenated compounds (14%), esters (2%), and S and N compounds (1%). In September, aldehydes had the highest concentration (37% of total VOC concentrations) and they were followed by ketones (36%), monoaromatics (13%), halogenated compounds (6%), VFAs (5%), S and N compounds (2%), and esters (1%). Monoaromatics had significant concentrations in both May and September samples (Fig. 1). Toluene had the highest average concentration in this group and was exceeded only by formic acid and propanal in both sampling campaigns. Benzene-to-toluene (B:T) ratio has been commonly used as an indicator of traffic emissions. A B:T ratio of 0.5 was reported from studies on vehicle exhaust [28]. Recent urban air measurements have also showed B:T ratios ranging between 0.27 and 0.5 [29–31] (Table 2). Therefore, B:T ratios around 0.5 may indicate that the ambient VOC concentrations are mainly affected by traffic emissions. However, landfills have significantly lower B:T ratios (i.e., 0.1 in USA) [32]. Benzene-to-toluene ratios measured in this study were 0.015 and 0.11 in May and September, respectively indicating the presence of higher amounts of toluene than benzene in landfills compared to traffic exhaust rich urban areas (Table 2). Chlorinated organics in the landfill might be related to the solvents that are widely used as cleaners and disinfectants in urban and industrial sources. Sixteen chlorinated compounds were detected in May and September samples. The average concentrations of some chlorinated species such as chloroform, carbon tetrachloride, chlorobenzene, trichloroethene, and tetrachloroethene were 0.08, 0.17, 0.04, 13.06, and 2.37 g m−3 , respectively (Fig. 1). Tetrachloroethylene is present in the textiles, dry cleaning workplaces, and chemical manufacturing units. Trichloroethene is used as a degreaser, extraction, and cleaning solvent for household, commercial, and industrial uses. Thus, their residuals might have reached the landfill along with the domestic, commercial, and industrial wastes. Compared to monoaromatic compounds, the concentrations of chlorinated species were relatively low in May and September samples. In May, all VFAs were detected only at the medical waste burial point of the landfill. Formic and isobutyric acids were the most abundant volatile fatty acids in the landfill burial points with the concentrations of 24.22 and 3.56 g m−3 , respectively.
225
Formic acid Acetic acid Propionic acid Isobutyric acid Butyric acid Isovaleric acid Valeric acid Isocaproic acid Caproic acid Heptanoic acid Acrolein Butanal Crotonaldehyde Decanal Heptanal Hexanal Nonanal Octanal Pentanal Propanal Butyl acetate Butyl formate Butyl propionate Methyl propionate Vinyl acetate Benzene Ethylbenzene o,m-Xylene p-Xylene Styrene Toluene 1,1,1-Trichloroethane 1,1,2,2-Tetrachloroethane 1,1,2-Trichloroethane 1,1-Dichloroethane 1,1-Dichloroethene 1,2-Dichlorobenzene 1,2-Dichloroethane 1,2-Dichloropropane 1,3-Dichlorobenzene 1,4-Dichlorobenzene Bromodichloromethane Bromoform Carbon tetrachloride Chlorobenzene Chloroform cis-1,3-Dichloropropene cis-1,4-Dichloro-2-butene Dibromochloromethane Iodomethane Methylene chloride Tetrachloroethene trans-1,2-Dichloroethene trans-1,3-Dichloropropene trans-1,4-Dichloro-2-butene Trichloroethene 2-Butanone 2-Hexanone 4-Methyl-2-pentanone (MIBK) Acetone Cyclohexanone Carbon disulfide Acrylonitrile Pyridine
May
September
3.44–43.71 (24.22) 0.17–5.34 (1.92) 0.19–3.52 (1.86) 0.50–6.61 (3.56) 0.06–4.39 (2.22) 0.05–6.05 (2.34) 0.07–4.99 (1.86) 2.84a 1.37–2.91 (2.14) 0.25–8.85 (3.39) 0.48–2.02 (1.04) 0.43–1.70 (1.01) 0.05–0.38 (0.14 0.66–1.49 (0.97) 0.16–1.51 (0.73) 0.47–5.94 (2.59) 0.64–2.16 (1.57) 0.27–2.28 (1.54) 0.15–1.92 (0.75) 4.16–38.55 (21.13) 0.12–7.54 (2.70) 0.03–0.10 (0.06) 0.11a N.D. 0.01–2.29 (0.51) 0.09–0.59 (0.29) 0.21–4.94 (2.03) 0.35–10.16 (4.33) 0.34–9.55 (3.59) 0.67–14.44 (3.88) 1.56–47.42 (18.97) 0.02–0.05 (0.04) N.D. 0.05–0.08 (0.06) 0.0014–0.01 (0.006) 0.12–0.37 (0.27) 0.03–0.08 (0.05) 0.01–1.22 (0.30) N.D. 0.002–0.01 (0.005) 0.08–0.40 (0.25) N.D. N.D. 0.12–0.23 (0.17) 0.01–0.12 (0.04) 0.03–0.16 (0.08) 0.03–0.07 (0.05) N.D. N.D. 0.001–0.02 (0.02) 1.62–7.95 (4.42) 0.05–9.16 (2.37) N.D. 0.085–0.09 (0.09) N.D. 0.18–62.91 (13.06) N.D. 0.04–0.80 (0.27) 0.03–0.42 (0.21) 7.95–67.60 (37.17) 0.08–9.13 (3.15) 0.41–5.05 (1.51) 0.09–0.20 (0.14) N.D.
11.10a N.D. N.D. N.D. N.D. 0.20a 0.37a N.D. 0.14–0.34 (0.22) 0.30–1.30 (0.73) 1.25–2.66 (1.83) 0.34–1.33 (0.59) 0.01–0.15 (0.09) 1.25–9.42 (3.95) 0.18–0.64 (0.32) 0.45–3.55 (1.32) 1.28–5.01 (2.64) 0.53–1.42 (0.82) 0.26–0.87 (0.44) 5.15–13.45 (8.49) 0.09–0.42 (0.21) 0.03–0.12 (0.05) 0.10a 0.18a 0.08–1.79 (0.65) 0.34–1.06 (0.53) 0.15–0.76 (0.45) 0.22–1.11 (0.73) 0.15–1.01 (0.62) 0.13–0.40 (0.24) 1.51–11.18 (4.76) 0.02–0.10 (0.04) N.D. N.D. 0.02a 0.04–0.70 (0.22) 0.02–0.09 (0.04) 0.01a N.D. 0.02a 0.07–0.21 (0.14) N.D. N.D. 0.10–0.15 (0.13) 0.009–0.014 (0.01) 0.03–0.13 (0.06) N.D. N.D. N.D. 0.02–0.03 (0.02) 0.68–2.58 (1.43) 0.02–1.02 (0.50) 0.11a N.D. 0.004a 0.11–1.95 (0.81) N.D. 0.06–0.22 (0.10) 0.05–0.18 (0.08) 11.20–28.57 (19.94) 0.004a 0.24–2.36 (1.15) 0.03–0.04 (0.034) N.D.
Data in parenthesis represent the average concentrations; N.D.: not detected. a Identified only in one sampling site.
226
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
Fig. 1. Seasonal variations of concentrations of monoaromatics, halogenated compounds, aldehydes and ketones, and VFAs.
In September, the concentration of formic acid was relatively lower and it could be detected at one site only. High levels of aldehydes and ketones were observed at each sampling location in May and September with the maximum levels at sampling sites 3 and 4. Propanal has a pungent, suffocating and unpleasant odor characterizing the landfill gas emissions. Propanal was found above the odor threshold of 3.6 g m−3 [33] at all sampling points. Some esters and organic S and N compounds were also identified and quantified in this study but the concentrations of these compounds were relatively low compared to other VOC groups. For example, concentrations of butyl acetate were 2.70 and 0.21 g m−3 in May and September, respectively. Carbon disulfide that has a characteristic odor (vegetable sulfide/medicinal) had relatively low average concentrations (1.51 g m−3 in May
Table 2 Comparison of average benzene and toluene levels (g m−3 ) and B:T ratios in ambient air at selected locations Benzene
Toluene
B:T
Site
Reference
0.23 0.53
18.97 4.76
0.015 0.11
Izmir (landfill) Izmir (landfill)
3.31 37.6 52.8 33.8
15.39 102 528 62
0.27 0.37 0.1 0.5
Izmir (urban) Izmir (urban) USA (landfill) Guangzhou (urban)
This study (May) This study (September) [29] [31] [32] [30]
and 1.15 g m−3 in September). Acrylonitrile concentrations were even lower (0.034–0.14 g m−3 ). Recent studies have shown that reduced sulfur compounds are important trace components of landfill gases [34–36]. Thus, sulfur and nitrogen containing organic compounds that are not measured in this study may also be significant in odor characterization. 3.2. Odor and total VOC concentrations The results of olfactometric and chemical analysis for the two sampling periods are given in Table 3. The total VOC concentrations presented in Table 3 are the sum of concentrations of detected compounds. For May samples, olfactometric concentrations varied between 1416 and 116,027 OU m−3 with a mean value of 47,886 OU m−3 and the total VOC concentrations varied between 69 and 258 g m−3 (average, 156 g m−3 ). For September samples the olfactometric and total VOC concentrations ranged between 1070 and 111,980 OU m−3 (average, 29,684 OU m−3 ) and 43–101 g m−3 (average, 56 g m−3 ), respectively. Total VOC concentrations in May and September have shown similar patterns. In September total VOC concentrations were relatively lower at all sampling locations with slightly decreasing odor concentrations (Fig. 2, Table 3). A significant decrease in odor concentration was observed only at Site 5. This decrease was probably due to the fact that during the May sampling there was an ongoing waste burial while the site was covered with soil during the September sampling.
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
227
Table 3 Concurrent data sets of the olfactometric odor concentrations (OU m−3 ) and chemical concentrations (g m−3 ) for May and September samples Sampling point
Odor concentration
Total VOC concentration
May 1 2 3 4 5
3765 1416 116027 35928 82292
± ± ± ± ±
153 219 21939 2050 3215
118 69 258 207 127
± ± ± ± ±
September 1 2 3 4 5
3380 1070 111980 29270 2720
± ± ± ± ±
400 425 20817 2309 800
56 36 101 43 44
± ± ± ± ±
Acids
Aldehydes
Esters
Halogenated compounds
Monoaromatics
Ketones
S and N compounds
10 2 13 11 9
0.11 45 38 45 1
18 9 48 34 48
1 0.17 10 5 0.42
74 2 17 5 7
6 4 76 67 13
19 8 68 50 52
1.2 0.6 0.6 0.6 5.2
5 2 6 2 4
N.D. 1 12 2 0.30
19 14 37 18 14
0.26 0.19 1 1 2
5 1 6 3 2
6 5 15 8 3
25 12 29 11 23
1.0 1.9 2.4 0.27 0.29
Fig. 2. Variations of odor and total VOC concentrations in May and September samples.
An odor control chemical (consisted of nearly 20% calcium hydroxide) applied onto the landfill sites (Izmir Metropolitan Municipal Waste Management Authority, 2005) during the period starting with mid-April and ending in September 2005. This application has focused mainly on sludge and medical
waste burial areas (Site 1 and 2). The burial areas were covered with 1–3 cm of the chemical every third to fourth day. May and September samplings at the landfill sites coincided with 1 month after the beginning and end of this application period. Therefore, it is possible that the composition of the odorous gas samples was affected by this application. However, the effect of the chemical application is not significant on odor concentrations since there was only a slight decrease at all sampling sites except Site 5 that was possibly affected by the soil cover (Table 3). VFA concentrations were very low and some of them could not be detected in September in contrast to the fact that most of them were found at all sampling sites in May (Fig. 1). The VFA is a good indicator of waste decomposition process. Lower VFA concentrations observed in September may be due to decreased decomposition rate as a result of odor control chemical or due to the sorption of the VFAs onto the alkaline chemical. However, the decrease in other non-biogenic VOCs (i.e., monoaromatic and halogenated compounds) concentrations in September cannot be explained by the application of odor control chemical since they are not end products of decomposition or they are not likely to be sorbed by the chemical. The decrease in concentrations of these VOCs was probably due to their lower content in the wastes as a result of loss by evaporation during a long and hot period.
Table 4 Pearson correlation coefficients between the different compound groups and odor (n = 10)
Odor Total VOCs Acids Aldehydes Esters Halogenated compounds Monoaromatics Ketones S and N compounds *
Odor
Total VOCs
Acids
Aldehydes
Esters
Halogenated compounds
Monoaromatics
Ketones
S and N compounds
1
0.64* 1
0.23 0.63 1
0.91* 0.76* 0.21 1
0.52 0.87* 0.61 0.58 1
−0.10 0.23 −0.18 −0.04 0.02 1
0.55 0.93* 0.69* 0.66* 0.94* −0.03 1
0.73* 0.88* 0.37 0.91* 0.79* −0.02 0.83* 1
0.41 0.02 −0.33 0.53 −0.30 −0.03 −0.18 0.29 1
Statistically significant (P < 0.05).
228
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229
3.3. Relationship between the odor and total VOC concentration The characterization of VOCs at landfills [10] and odor assessment on the basis of the dispersion of odors [1] were previously reported. However, previous studies did not investigate the relationship between landfill odor and VOC concentrations. The relationship between concentrations of odors and total VOCs measured in the present study was analyzed statistically. A correlation matrix was calculated for different variables and Pearson coefficients were given as a measure of relationship between groups of variables (Table 4). There are statistically significant correlations (P < 0.05, n = 10) between odor concentrations and total VOC, aldehyde and ketone concentrations (Table 4). Linear regression analysis performed on the overall data set (May + September) between odor and total VOC concentrations indicated that 41% of the variance in odor concentrations can be explained by the total VOC concentrations (r2 = 0.41, n = 10, P < 0.05). The relationships of odor concentrations with the concentrations of different groups of chemicals were further examined using a step-wise multiple regression analysis. It was found that the concentrations of aldehydes, ketones, and esters are the best estimators, explaining 96% of the variability in odor concentrations (r2 = 0.96, n = 10, P < 0.01). 4. Conclusions The odor levels and the composition of odorous gases emitted from a municipal landfill site were studied. Up to 53 VOC compounds in different classes (mono aromatics, halogenated compounds, aldehydes, ketones, esters, and volatile fatty acids) were identified and quantified. The benzene-to-toluene ratio at the landfill site in May and September was found to be significantly lower than the values determined in urban areas indicating the presence of higher amounts of toluene than benzene in landfills compared to urban areas. Collected samples were also analyzed for odor concentrations by olfactometry. It is concluded that a relationship exists between odor and total VOC concentrations at the studied landfill site. The relationships of odor concentrations with the different groups of chemicals were also examined and it was found that the concentrations of aldehydes, ketones, and esters are the best estimators, explaining 96% of the variability in odor concentrations. Acknowledgement This work was supported in part by the research fund of Dokuz Eylul University (Project No. 03.KB.FEN.061) and by the Scientific and Technical Research Council of Turkey (TUBITAK) (Project No. ICTAG-A056). Olfactometric analyses were conducted in the laboratory established by the help of the EU project; “Odorous Emissions and Immissions Management Policy in Turkey-LIFE TCY/TR/000009”. Authors are thankful to panel members for their participation in olfactometry.
The help and permission by the Izmir Metropolitan Municipal Waste Management Authority should also be gratefully acknowledged.
References [1] U. Sarkar, S.E. Hobbs, P. Longhurst, J. Environ. Manage. 68 (2003) 153. [2] ATDSR, Agency for Toxic Substances and Disease Registry (http://www. atsdr.cdc.gov). [3] F. Kreith, Handbook of Solid Waste Management, McGraw-Hill, New York, 1995. [4] J.T. Houghton, G.J. Jenkins, J.J. Ephraums (Eds.), Climate Change: The IPCC Scientific Assessment, Cambridge University Press, Cambridge, 1991. [5] S.C. Zou, S.C. Lee, C.Y. Chan, K.F. Ho, X.M. Wang, L.Y. Chan, Z.X. Zhang, Chemosphere 51 (2003) 1015. [6] A. Muezzinoglu, Chemosphere 51 (2003) 245. [7] S.M. van Ruth, Biomol. Eng. 17 (2001) 121. [8] CEN, Air quality—Determination of odor concentration by dynamic olfactometry, European Standard EN 13725, Brussels, Belgium, 2003. [9] E. Smet, H. van Langenhove, I. De Bo, Atmos. Environ. 33 (1999) 1295. [10] E. Davoli, M.L. Gangai, L. Morselli, D. Tonelli, Chemosphere 51 (2003) 357. [11] N. Defoer, I. De Bo, H. van Langenhove, J. Dewulf, T. van Elst, J. Chromotogr. A 970 (2002) 259. [12] F. Biasioli, F. Gasperi, G. Odorizzi, E. Aprea, D. Mott, F. Marini, G. Autiero, G. Rotondo, T.D. M¨ark, Int. J. Mass Spectrom. 239 (2004) 103. [13] K.K. Kleeberg, Y. Liu, M. Jans, M. Schlegelmilch, J. Streese, R. Stegmann, Waste Manage. 25 (2005) 872. [14] P. Pierucci, E. Porazzi, M. Pardo Martinez, F. Adani, C. Carati, F.M. Rubino, A. Colombi, E. Calcaterra, E. Benfenati, Chemosphere 59 (2005) 423. [15] A. Vallat, H. Gu, S. Dorn, Phytochemistry 66 (2005) 1540. [16] J.C. Caisard, V. Bergougnoux, M. Martin, M. Mauriat, S. Baudino, Ann. Bot. 97 (2006) 231. [17] L. Grison-Pige, J.M. Bessiere, T.C.J. Turlings, F. Kjelberg, J. Roy, M.M. Hossaert-McKey, Funct. Ecol. 15 (2001) 551. [18] M.E. Hoballah, T.G. K¨ollner, J. Degenhardt, T.C.J. Turlings, OIKOS 105 (2004) 168. [19] R. Noble, P.J. Hobbs, A. Dobrovin-Pennington, T.H. Misselbrook, A. Mead, J. Environ. Qual. 30 (2001) 760. [20] M. Gorgenyi, J. Dewulf, H. Van Langenhove, Chromatographia 51 (2000) 461. [21] C. Peng, S. Batterman, J. Environ. Monitor. 2 (2000) 313. [22] US-EPA, Compendium Method TO-17, Determination of Volatile Organic Compounds in Ambient Air Using Active Sampling Onto Sorbent Tubes, Center for Environmental Research Information Office of Research and Development, Cincinnati, USA, 1999. [23] T. Harner, M. Shoeib, M. Diamond, M. Ikonomou, G. Stern, Chemosphere, in press. [24] G.O. Thomas, M. Wilkinson, S. Hodson, K.C. Jones, Environ. Pollut. 141 (2006) 30. [25] F. Wong, T. Harner, Q.T. Liu, M.L. Diamond, Environ. Pollut. 129 (2004) 387. [26] A. Sofuoglu, E. Cetin, S.S. Bozacioglu, G.D. Sener, M. Odabasi, Atmos. Environ. 38 (2004) 4483. [27] R. Lohmann, P.A. Brunciak, J. Dachs, C.L. Gigliotti, E. Nelson, D. Van Ry, T. Glenn, S.J. Eisenreich, J.L. Jones, K.C. Jones, Atmos. Environ. 37 (2003) 959–969. [28] P.A. Scheff, R.A. Wadden, Environ. Sci. Technol. 27 (1993) 617. [29] M. Odabasi, Unpublished results, 2005. [30] X.M. Wang, G.Y. Sheng, J.M. Fu, C.Y. Chan, L.Y. Chan, S.C. Lee, Atmos. Environ. 36 (2002) 5141. [31] A. Muezzinoglu, M. Odabasi, L. Onat, Atmos. Environ. 35 (2001) 753.
F. Dincer et al. / J. Chromatogr. A 1122 (2006) 222–229 [32] J. LaRegina, J.W. Bozzelli, R. Harkov, S. Gianti, Environ. Prog. 5 (1986) 18. [33] US-EPA, Reference Guide to Odor Thresholds for Hazardous Air Pollutants Listed in the Clean Air Act Amendments of 1990, Air Risk Information Support Center, USEPA Office of Research and Development, EPA/600/R-92/047, 1992.
229
[34] K.H. Kim, Y.J. Choi, E.C. Jeon, Y. Sunwoo, Atmos. Environ. 39 (2005) 1103. [35] Z.H. Shon, K.H. Kim, E.C. Jeon, M.Y. Kim, Y.K. Kim, S.K. Song, Atmos. Environ. 39 (2005) 4803. [36] S. Lee, Q. Xu, M. Booth, T.G. Townsend, P. Chadik, G. Bitton, Waste Manage. 26 (2006) 526.