Environ. Sci. Technol. 2006, 40, 5962-5970
Sorption of Heterocyclic Organic Compounds to Reference Soils: Column Studies for Process Identification ERPING BI, T O R S T E N C . S C H M I D T , * ,† A N D STEFAN B. HADERLEIN Chair of Environmental Mineralogy, Centre for Applied Geoscience, University of Tuebingen, Sigwartstr. 10, D-72076 Tuebingen, Germany
In this study, the sorption behavior of a wide variety of N-, S-, and O-heterocyclic compounds (NSOs) to reference soils (Eurosoils 1-5) was characterized by a soil column chromatography (SCC) approach. The major goal was to identify the compound specific and environmental factors influencing sorption processes. The sorption of S- and O-heterocyclic compounds (thiophene, benzothiophene, 5-methylbenzo[b]thiophene, benzofuran, 2-methylbenzofuran, and 2,3-dimethylbenzofuran) was generally controlled by nonspecific interactions with soil organic carbon (OC). With regard to non-ionizable N-heterocyclic compounds, pyrrole, 1-methylpyrrole, and pyrimidine were hardly retarded in any soil. The sorption of indole, 2-hydroxyquinoline, and benzotriazole was dominated by specific interaction (e.g., complexation of surface-bound cations) rather than partition to soil OC. The sorption of ionizable N-heterocyclic compounds (quinoline, isoquinoline, quinaldine, 2-methylpyridine, and pyridine) can be described by a conceptual model including partitioning to soil OC, cation exchange, and an additional sorption process (probably surface complexation of the neutral species). Cation exchange was usually the dominant mechanism in the sorption of ionizable compounds if the protonated fraction of the compound exceeded 5%. Otherwise, surface complexation became dominant. Soil pH was the most important factor influencing the sorption of ionizable NSOs. Our study suggests that a fairly precise assessment of sorption in most soils can be expected for N-, S-, and O-heterocyclic compounds if the three sorption mechanisms are taken into account where appropriate. Deviations from this behavior indicated special cases where additional soil specific properties (e.g., accessible surface, CEC, charge density) need to be considered such as for 2-methylpyridine and pyridine sorption to Eurosoil 1.
Introduction Heterocyclic organic compounds, containing nitrogen, oxygen, or sulfur in the ring-structure, are frequently found in groundwater at sites contaminated with coal tar (containing * Corresponding author phone: +49-203-379-3311/-3308; fax +49203-379-2108; e-mail:
[email protected]. † Current address: University Duisburg-Essen, Chair of Instrumental Analysis, Lotharstr. 1, MF147, D-47048 Duisburg. 5962
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006
about 5% N-, S-, and O-heterocyclic compounds (NSOs) (1), and they have been shown to be persistent in some cases (2). Sorption to soil is one of the key processes that determine the fate of these compounds in the subsurface. Sorption of nonpolar organic compounds strongly depends on the organic carbon (OC) content in soil or sediment (3, 4). For non-ionizable NSOs, such as benzothiophene, nonspecific interaction (namely van der Waals forces) appears to be dominant in the sorption to clay till (1). Concerning the sorption of ionizable NSOs, however, which show a certain polarity, additional interactions might play an important role. Studies on sorption of N-heterocyclic compounds to soils and minerals have shown that the overall sorption is dominated by cation exchange of protonated quinoline (510), pyridine (11-13), and 2-methylpyridine (14). Even at low fractions of the cationic form in aqueous solution, the overall sorption of such compounds may be dominated by this process (5). In many cases, the highest sorption of such organic bases occurs in the pH range corresponding to the pKa of the corresponding cation (8). The conventional approach to investigate the sorption of organic contaminants in water-soil systems is to carry out batch experiments. For weakly sorbing solutes (e.g., NSOs of high hydrophilicity) and/or for soil of low organic matter content (e.g., BT > THIO and DMBF > 2MBF > BF. This can also be seen in Figure 2. The Koc values for the five investigated soils differ by a factor of ca. 3 or 0.5 log units (Figure 2). As shown in Figure 2, for all compounds (Tol, Benz, DMBF, 2MBF, BF, 5MBT, BT, and THIO), for which nonspecific interaction is a dominant sorption mechanism, Koc decreases in the following order: ES1 = ES5 > ES3 > ES2 > ES4. This variation probably reflects the structural variability of the organic material in the five soils. Both Eurosoil 1 and 5 are of high humification degree (30), and there is a rather high amount of micropores in these soils. Therefore, micropore filling might contribute to the overall sorption in these two soils as suggested by Celis et al. (31). Due to similar properties of humic acids in Eurosoil 1 and 5 (30), the sorption of compounds dominated by nonspecific interaction shows the same pattern. Non-ionizable N-heterocyclic compounds (PR, 1MPR, IND, and PM) in the natural pH range exhibit less consistent 5966
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006
compounds (pKa)
E5Q-4b pH 5.2-5.7
E3Q-1 pH 6.0
E1Q-2 pH 6.0-6.3
E4Q-1 pH 6.9
E2Q-3 pH 8.6
IND (-2.4) 2QUI (-0.31/11.76) BTA (1.6/8.6) QUI (4.9) PY (5.23) IQUI (5.42) 2MQUI (5.71) 2MPY (6.00)
60
75
61
67
63
98
98
97
98
97
86
91
87
91
74
97
93
98
93
79
97
99
100
98
81
99
99
99
97
79
95
87
94
29
99
99
99
71
100
a K data are from Table 2, K d d, part (sorption due to partitioning) values are calculated by inserting the compound Kow (Table 2) into the equations fitted in Figure 1; fion (fraction of ion species) is shown in Table 5. b The line with E5Q-4 is for the column names; below it are the measured pH values in each column.
sorption patterns. PR, 1MPR, and PM show very low sorption and thus are hardly retarded in all soils (Table 2). Weak nonspecific interaction appears to dominate the sorption of PR and 1MPR. There is a large deviation of PM from the NSI-line in Figure 1. However, due to low overall sorption of PM, a quantitative interpretation of the data seems inappropriate. With regard to IND, 2QUI, and BTA, the deviation from the NSI line (Figure 1) shows a significant contribution of specific interactions to their overall sorption. This contribution was quantified by calculating Kd expected for the partitioning into organic matter based on the equations given in the inset and the compound’s log Kow values. The contribution of nonspecific interaction to overall sorption was then subtracted from the observed Kd and normalized to the observed Kd as shown in Table 4. 2QUI is expected to exist mainly as the keto form (oxo tautomer, see eq 3) in aqueous solution, and no sorption to aluminum and iron oxide was found in batch experiments (29).
The contribution of specific interaction to the overall sorption of BTA has also been mentioned in a previous study (34). Except for Eurosoil 2, BTA (pKa ) 8.6) is present mainly in its neutral form in the solution. In Eurosoil 2, 50% of BTA is present as anion (Table 5). Repulsion between the anion and the negatively charged sites of the soil reduces the sorption of BTA significantly. For IND, BTA, and 2QUI the dominant interaction between the neutral compounds and the solid matrix is possibly via surface complexation to surface-bound cations, as described for the sorption of veterinary pharmaceuticals in soils (35) and discussed in detail for BTA in a forthcoming paper. For IND it has indeed been shown that there is enhanced interaction between cations and the π-electron system of N-heterocyclic compounds compared with aromatic hydrocarbons (e.g., benzene) (32, 33). This might also be the case in the sorption of 2QUI and BTA. Sorption of Ionizable N-Heterocyclic Organic Compounds. The group of ionizable basic N-heterocyclic com-
TABLE 5. Speciation of Compounds Under Different Conditionsa E5Q-4 compounds
pH
ion (%)
PY 2MPY QUI IQUI 2MQUI BTA
5.7 5.6 5.5 5.7 5.7 5.2
28 71 20 34 51 0
E3Q-1 pH
ion (%)
6.0 6.0 6.0 6.0 6.0 6.0
15 52 8 22 34 0
E1Q2 pH
ion (%)
6.0 6.3 6.1 6.2 6.0 6.3
15 35 6 14 34 0.4
E4Q-1
E2Q-3
pH
ion (%)
pH
ion (%)
6.8 6.9 6.7 6.7 6.7 6.7
3 11 2 5 9 1
8.6 8.6 8.6 8.6 8.6 8.6
0 0 0 0 0 50
a Percentage of BH+ ) 1/(1 + 10pH-pKa) × 100%; percentage of BTA) (1 - 1/(1 + 10pH-pKa)) × 100%; pH values were monitored in the column experiments.
pounds (PY, 2MPY, QUI, IQUI, and 2MQUI) significantly deviates from the NSI lines (data points outside 95% confidence interval of NSI lines in Figure 1). Their overall
FIGURE 3. Relationship between cation fraction and sorption resulting from specific interaction. Fitting equation is y ) KB × (1 - x/100) + KBH+ × x/100. Empty triangle points (from Eurosoil 1 Column) in plots (b) and (d) were excluded. In plot (e) the data in the dashed circle means a higher value is expected but could not be quantified. VOL. 40, NO. 19, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5967
FIGURE 4. Predicted sorption based on eq 5 (resulting from contribution of various mechanisms) normalized by division by measured values.
sorption is dominated by specific interactions, except for 2MQUI in Eurosoil 2 (Table 4). These compounds are only slightly retarded in Eurosoil 2 (Table 2), which has the highest OC content and a pH of 8.6, making these compounds present only in the neutral form. The methyl group of 2MQUI increases its hydrophobicity, and thus the sorption of 2MQUI is dominated by nonspecific interaction to soil foc when there is no cation form present. For the other compounds within this group, the specific interaction contributes still more than 70% to the overall sorption (e.g., 71% for 2MPY in Table 4). In the other soils studied, a significant, albeit small, fraction of the compound is present as cation. Previous studies on PY (11-13, 36), 2MPY (14), and quinoline (8, 9, 11) showed that cation exchange is a dominant sorption mechanism for the protonated compounds. N-heterocyclic compounds (e.g., PY, 2MPY, QUI) have been found to be sorbed preferentially as the cation even when pH of the solution was greater than the pKa values (5, 9, 12, 14). In this study, a great impact of the cation fraction on sorption was also observed. Conceptual Model. We propose the following model (eq 4) to assess the contribution of each sorption mechanism to the overall sorption. 5968
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006
Kd ) Kd,part + Kd,BH+ + Kd,B
(4)
Kd ) Kpart(1 - fion) + KBH+ fion + KB(1 - fion)
(5)
Where Kd,part is the sorption due to partition to soil OC; fion is the cation fraction of protonated N-heterocyclic compounds; Kd,BH+ is the contribution of cation exchange of the protonated form of ionizable NSOs; Kd,B is the contribution of specific interaction (likely via surface complexation) of the neutral form of ionizable N-heterocyclic compounds and of non-ionizable N-heterocyclic compounds [L/kg]; and Kpart, KBH+, and KB are the constants for each mechanism, respectively. In eq 5, Kd is the overall sorption (data from Table 2). Kd,part values were calculated using the compound Kow (Table 2), and the equations shown for each soil are inset in Figure 1. The fraction present in cationic form (fion) was taken from Table 5. With these parameters known, KB and KBH+ were determined by combining eqs 4 and 5 and plotting Kd,BH+ + Kd,B vs fion:
Kd - Kd,part ) Kd,BH+ + Kd,B ) fion(KBH+ - KB) + KB
(6)
Results are shown in Figure 3. QUI and IQUI sorption resulting from specific interactions is well-correlated to the cation fraction in all soils. Compared with the other soils, 2MPY and PY show extraordinarily high sorption resulting from specific interaction in Eurosoil 1. Therefore, sorption data points from Eurosoil 1 for these two compounds were excluded in fitting. Sorption of QUI and IQUI in Eurosoil 1 shows the same phenomenon, but not so pronounced. The fitting results of KBH+ indicate cation exchange decreasing in the order IQUI (64.8) > QUI (36.6) > 2MQUI (16.6) > PY (1.45) > 2MPY (1.05). This order is consistent with an expected impact of hydration (or surface charge density) on cation exchange. The molecular size of two-ring compounds is larger than those of one-ring compounds. Therefore, they have a lower surface charge density and a smaller hydration shell. This reduces the distance between compounds and solid surface and causes a higher electrostatic attraction, and consequently leads to a higher sorption affinity. The effect of KB follows the order 2MQUI > IQUI > QUI . 2MPY > PY. The contribution of Kd,B to the sorption of 2MPY and PY is relatively small, and consequently results in small KB values. According to the order of KB and KBH+ for these compounds, it seems that the presence of a methyl group on the ring has influence on cation exchange and complexation interaction of a compound. However, further work is needed to clarify this issue. With the fitted KB and KBH+, the sorption of each compound and contribution of each sorption mechanism can be calculated by using eq 5 in different soils. The predicted values were normalized by dividing them with the measured values and the results are shown in Figure 4. The dominant contribution of cation exchange to overall sorption even at small cation fractions is clearly visible in the plots, which corroborates previous findings (5). In all soils, sorption of QUI and IQUI was well predicted. For PY and 2MQUI, the estimated values are within a factor of 2 of the experimental data. The relatively high deviation of 2MPY in Eurosoil 2 is caused by higher uncertainty of the data due to very low retardation (see Figure 3). In the plots of PY, 2MPY, and 2MQUI, their sorption to Eurosoil 1 was not calculated. The reason is that these compounds showed higher sorption in the Eurosoil 1 column than in the other soil columns, which indicates that there must be additional interactions besides that included in eq 5, especially for PY and 2MPY sorption. Based on previous studies, in addition to pH, other factors such as the total accessible surface area (1), and other soil properties (37) might also be important for the sorption of these compounds in a certain system. Due to high clay content, Eurosoil 1 has a high BET specific surface area even after dilution with quartz (12.5 m2/g in column E1Q-2, data shown in SI). The accessible surface might be a reason for higher sorption observed in Eurosoil 1. However, this additional interaction cannot be characterized based on the available data in this study. Overall, the results show that SCC is a suitable approach to investigate sorption of polar organic compounds to natural sorbents. The environmental pH is a critical index to assess the sorption of ionizable N-heterocyclic compounds in the suggested conceptual model. However, the effect of environmental factors such as aqueous phase composition and ionic strength should be further studied.
Acknowledgments We thank Dr. Stefan Lamotte from Bischoff for his suggestions and valuable discussions on soil column packing and soil column chromatography development. We acknowledge Peter Grathwohl, Christian Niederer, Satoshi Endo, and three anonymous reviewers for fruitful comments on the manuscript.
Supporting Information Available Tables of brief information on chemicals investigated, properties of the column packing materials, specific surface area of the packing materials, and sorption resulting from specific interaction. Figure showing representative soil column chromatograms. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Broholm, M. M.; Broholm, K.; Arvin, E. Sorption of heterocyclic compounds on natural clayey till. J. Contam. Hydrol. 1999, 39, 183-200. (2) Zamfirescu, D.; Grathwohl, P. Occurrence and attenuation of specific organic compounds in the groundwater plume at a former gasworks site. J. Contam. Hydrol. 2001, 53, 407-427. (3) Chiou, C. T. Partition and Adsorption of Organic Contaminants in Environmental Systems; John Wiley & Sons: New York, 2002. (4) Grathwohl, P. Influence of organic matter from soils and sediments from various origins on the sorption of some chlorinated aliphatic hydrocarbons: Implications on Koc correlations. Environ. Sci. Technol. 1990, 24, 1687-1693. (5) Ainsworth, C. C.; Zachara, J. M.; Schmidt, R. L. Quinoline sorption on Na-montmorillonite: Contributions of the protonated and neutral species. Clays Clay Miner. 1987, 35, 121-128. (6) Burgos, W. D.; Pisutpaisal, N.; Mazzarese, M. C.; Chorover, J. Adsorption of quinoline to kaolinite and montmorillonite. Environ. Eng. Sci. 2002, 19, 59-68. (7) Helmy, A. K.; De Bussetti, S. G.; Ferreiro, E. A. Adsorption of quinoline from aqueous solutions by some clays and oxides. Clays Clay Miner. 1983, 31, 29-36. (8) Thomsen, A. B.; Henriksen, K.; Møldrup, P.; Grøn, C. Sorption, transport, and degradation of quinoline in unsaturated soil. Environ. Sci. Technol. 1999, 33, 2891-2898. (9) Zachara, J. M.; Ainsworth, C. C.; Felice, L. J.; Resch, C. T. Quinoline sorption to subsurface materials: Role of pH and retention of the organic cation. Environ. Sci. Technol. 1986, 20, 620-627. (10) Zachara, J. M.; Ainsworth, C. C.; Schmidt, R. L.; Resch, C. T. Influence of cosolvents on quinoline sorption by subsurface materials and clays. J. Contam. Hydrol. 1988, 2. (11) Zachara, J. M.; Ainsworth, C. C.; Cowan, C.; Thomas, B. L. Sorption of binary mixtures of aromatic nitrogen heterocyclic compounds on subsurface materials. Environ. Sci. Technol. 1987, 21, 397-402. (12) Zachara, J. M.; Ainsworth, C. C.; Smith, S. C. The sorption of N-heterocyclic compounds on reference and subsurface smectite clay isolates. J. Contam. Hydrol. 1990, 6, 281-305. (13) Zhu, D. Q.; Herbert, B. E.; Schlautman, M. A. Sorption of pyridine to suspended soil particles studied by deuterium nuclear magnetic resonance. Soil Sci. Soc. Am. J. 2003, 67, 1370-1377. (14) O’Loughlin, E. J.; Traina, S. J.; Sims, G. K. Effects of sorption on the biodegradation of 2-methylpyridine in aqueous suspensions of reference clay minerals. Environ. Toxicol. Chem. 2000, 19, 2168-2174. (15) Das, B. S.; Lee, L. S.; Rao, P. S. C.; Hultgren, R. P. Sorption and degradation of steroid hormones in soils during transport: Column studies and model evaluation. Environ. Sci. Technol. 2004, 38, 1460-1470. (16) Fesch, C.; Simon, W.; Haderlein, S. B.; Reichert, P.; Schwarzenbach, R. P. Nonlinear sorption and nonequilibrium solute transport in aggregated porous media: Experiments, process identification and modeling. J. Contam. Hydrol. 1998, 31, 373407. (17) Lee, L. S.; Rao, P. S. C.; Brusseau, M. L. Nonequilibrium sorption and transport of neutral and ionized chlorophenols. Environ. Sci. Technol. 1991, 25, 722-729. (18) Mader, B. T.; Goss, K. U.; Eisenreich, S. J. Sorption of nonionic, hydrophobic organic chemicals to mineral surfaces. Environ. Sci. Technol. 1997, 31, 1079-1086. (19) Brusseau, M. L. Transport of reactive contaminants in heterogeneous porous media. Rev. Geophys. 1994, 32, 285-314. (20) Gawlik, B. M.; Bo, F.; Kettrup, A.; Muntau, H. Characterisation of a second generation of European reference soils for sorption studies in the framework of chemical testingsPart I: chemical composition and pedological properties. Sci. Total. Environ. 1999, 229, 99-107. (21) Bi, E. 2006. Sorption and transport of heterocyclic aromatic compounds in soils. Ph.D. dissertation, Tuebingen university, Tuebingen, Germany. VOL. 40, NO. 19, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5969
(22) Valocchi, A. J. Validity of the local equilibrium assumption for modeling sorbing solute transport through homogeneous soils. Water Resour. Res. 1985, 21, 808-820. (23) Young, D. F.; Ball, W. P. Column experimental design requirements for estimating model parameters from temporal moments under nonequilibrium conditions. Adv. Water Resour. 2000, 23, 449-460. (24) Dyson, N. Chromatographic Integration Methods; Second ed.; Royal Society of Chemistry: Cambridge, 1998. (25) TorresLapasio, J. R. .; BaezaBaeza, J. J.; GarciaAlvarezCoque, M. C. A model for the description, simulation, and deconvolution of skewed chromatographic peaks. Anal. Chem. 1997, 69, 38223831. (26) Altfelder, S.; Streck, T.; Maraqa, M. A.; Voice, T. C. Nonequilibrium sorption of dimethylphthalatesCompatibility of batch and column techniques. Soil Sci. Soc. Am. J. 2001, 65, 102-111. (27) Turin, H. J.; Bowman, R. S. Sorption behavior and competition of bromacil, napropamide, and prometryn. J. Environ. Qual. 1997, 26, 1282-1287. (28) Albert, A.; Goldacre, R. J.; Phillips, J. Strength of heterocyclic bases. J. Chem. Soc. 1948, 2240-2249. (29) Vasudevan, D.; Dorley, P. J.; Zhuang, X. Adsorption of hydroxy pyridines and quinolines at the metal oxide-water interface: Role of tautomeric equilibrium. Environ. Sci. Technol. 2001, 35, 2006-2013. (30) Senesi, N.; D’Orazio, V.; Ricca, G. Humic acids in the first generation of EUROSOILS. Geoderma 2003, 116, 325-344. (31) Celis, R.; Real, M.; Hermosin, M. C.; Cornejo, J. Sorption and leaching behaviour of polar aromatic acids in agricultural soils
5970
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006
(32)
(33)
(34)
(35) (36) (37)
by batch and column leaching tests. Eur. J. Soil Sci. 2005, 56, 287-297. Zhu, W.-L.; Jiang, H.-L.; Tan, X.-J.; Chen, K.-X.; Ji, R.-Y.; Puah, C. M.; Cao, Y. Density functional theory (DFT) study on the interaction of ammonium (NH4+) and aromatic nitrogen heterocyclics. J. Chem. Soc., Perkin Trans. 2 1999, 2615-2622. Chatterjee, A.; Ebina, T.; Iwasaki, T. Best dioctahedral smectite for nitrogen heterocyclics adsorption - A reactivity index study. J. Phys. Chem. A. 2001, 105, 10694-10701. Hart, D. S.; Davis, L. C.; Erickson, L. E.; Callender, T. M. Sorption and partitioning parameters of benzotriazole compounds. Microchem. J. 2004, 77, 9-17. Tolls, J. Sorption of veterinary pharmaceuticals in soils: A review. Environ. Sci. Technol. 2001, 35, 3397-3406. Baker, R. A.; Luh, M. D. Pyridine sorption from aqueous solution by montmorillonite and kaolinite. Water Res. 1971, 5, 839-848. Gawlik, B. M.; Feicht, E. A.; Karcher, W.; Kettrup, A.; Muntau, H. Application of the European reference soil set (EUROSOILS) to a HPLC-screening method for the estimation of soil adsorption coefficients of organic compounds. Chemosphere 1998, 36, 2903-2919.
Received for review February 28, 2006. Revised manuscript received July 13, 2006. Accepted August 14, 2006. ES060470E
