Ofloxacin sorption in soils after long-term tillage: The contribution of organic and mineral compositions

Science of the Total Environment 497–498 (2014) 665–670

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Ofloxacin sorption in soils after long-term tillage: The contribution of organic and mineral compositions Dandan Zhou, Bingfa Chen, Min Wu ⁎, Ni Liang, Di Zhang, Hao Li, Bo Pan Faculty of Environmental Science & Engineering, Kunming University of Science & Technology, Kunming 650500, China

H I G H L I G H T S • • • • •

Mineral compositions tend to be similar after tillage. Increased SOM decreases OFL sorption for soils from the same geological location. Tillage activities or dense vegetations greatly decrease OFL sorption. The summed sorption of individual soil components is higher than the intact soil. Soil should be treated as a dynamic environmental matrix for antibiotic sorption.

a r t i c l e

i n f o

Article history: Received 4 April 2014 Received in revised form 26 July 2014 Accepted 31 July 2014 Available online xxxx Editor: Eddy Y. Zeng Keywords: Antibiotics Drug resistance Mineral compositions Montmorillonite Nonlinear sorption

a b s t r a c t Intensive human activities in agricultural areas resulted in significant alteration of soil properties, which consequently change their interactions with various contaminants. This process needs to be incorporated in contaminant behavior prediction and their risk assessment. However, the relevant study is missing. This work was designed to examine the change of soil properties and ofloxacin (OFL) sorption after tillage. Soil samples were collected in Yuanyang, Mengzi, and Dianchi areas with different agricultural activities. Although the mineral compositions of soils from Yuanyang and Dianchi differed greatly, these compositions are similar after tillage, especially for paddy soils. Soil pH decreased generally after OFL sorption, suggesting that ion exchange of OFL with protons in soil organic matter (SOM) was important for OFL sorption. However, a positive relationship between SOM and OFL sorption was not observed. On the contrary, increased SOM decreased OFL sorption when soils from the same geological location were compared. Generally speaking, tillage activities or dense vegetations greatly decreased OFL sorption. The higher OFL sorption in B horizon than A horizon suggested limited leaching of OFL through soil columns. The summed sorption calculated based on the sorption of individual soil components and their percentages in soils was higher than the intact soil. This phenomenon may be understood from the interactions between soil components, such as the coating of SOM on mineral particles. This study emphasizes that soil should be treat as a dynamic environmental matrix when assessing antibiotic behaviors and risks, especially in the area with intense human activities. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The wide application of antibiotics has resulted in their ubiquitous presence in the environment, such as surface water, seawater, groundwater, soils, and sediments (Nikolaou et al., 2007; Kemper, 2008; Thiele-Bruhn, 2003). For example, as high as 1560 ng/g ofloxacin (OFL) was detected in the sediments of the Pearl Rivers (Yang et al., 2010). Lots of research attention have been attracted regarding the negative effects of antibiotics to ecological systems and the frequently reported drug resistant genes (Cirz et al., 2005; Dhar and McKinney, 2007). Proper assessment of antibiotic risks requires a comprehensive ⁎ Corresponding author. Tel./fax: +86 871 65102829. E-mail address: [email protected] (M. Wu).

http://dx.doi.org/10.1016/j.scitotenv.2014.07.130 0048-9697/© 2014 Elsevier B.V. All rights reserved.

understanding of their environmental behaviors. The interactions between antibiotics and environmental media, primarily soil particles, are essential for the above purpose. In addition, soil is the environmental matrix with the most abundance of microorganisms, and a very important zone for the development of drug resistant genes. Thus, the occurrence of antibiotics in soil system and antibiotic–soil interaction has been the research hotspot recently. Previous studies have suggested that both organic and inorganic compositions of soil particles are important for antibiotic sorption (Wu et al., 2014). It was also reported that different soil compositions have distinct sorption characteristics to antibiotics (Peterson et al., 2009). During the chemical weathering of soils as well as the biological activities in soil systems, soil compositions will be changed, which consequently alter soil–antibiotic interactions. Thus, soils could not be

666

D. Zhou et al. / Science of the Total Environment 497–498 (2014) 665–670

Table 1 Selected soil and mineral properties. pHzpc

Y-N Y-P Y-D M-N M-G M-A M-B B-N B-P B-D B-T

2.24 2.87 2.78 3.69 3.63 5.34 3.35 7.36 2.89 6.66 3.54

CEC

BET

Mineral analysis

cmol(+)/kg

m2/g

QZ

MS

HB

KL

ML

CT

OH

HM

CC

Elemental analysis N

C

H

S

O

C/H

C/(N + O)

2.12 6.68 4.83 5.33 8.18 8.16 9.69 13.94 13.74 12.78 7.81

33.9 18.4 18.3 82.0 35.0 22.6 65.8 65.6 30.0 22.7 2.62

43.4 45.1 46.0 31.8 24.8 32.1 28.4 17.8 49.1 48.8 41.1

35.2 27.0 30.7 21.2 15.0 18.3 14.1 19.5 25.5 21.2 21.6

12.6 10.8 8.4 nd nd nd nd nd nd nd nd

5.3 8.3 6.5 23.4 30.0 23.4 26.1 27.4 8.8 11.3 9.5

nd 4.7 4.4 7.4 9.5 12.8 22.5 10.9 6.1 6.8 7.4

nd nd nd nd nd nd nd 13.0 nd nd nd

1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 nd

nd nd nd 11.4 16.3 5.1 5.0 8.3 6.5 6.5 nd

2.5 3.1 3.0 3.8 3.4 7.3 2.9 2.1 3.0 4.4 20.4

0.06 0.18 0.19 0.23 0.18 0.93 0.32 0.05 0.22 0.32 1.27

0.45 1.93 1.94 2.66 1.95 11.87 2.85 1.53 2.16 3.53 22.04

0.97 1.01 1.02 1.80 1.71 2.39 1.85 1.54 1.00 1.06 3.01

0.08 0.04 0.04 0.07 0.05 0.09 0.09 0.05 0.06 0.19 0.38

7.90 8.38 8.59 13.96 15.51 24.33 14.43 15.30 6.79 9.22 17.03

0.04 0.16 0.16 0.12 0.10 0.41 0.13 0.08 0.18 0.28 0.61

0.08 0.30 0.29 0.25 0.17 0.62 0.26 0.13 0.41 0.49 1.59

The abbreviations for mineral compositions: QZ (quartz); MS (muscovite); HB (hydrobiotite); KL (kaolinite); ML (montmorillonite); CT (calcite); OH (octahedrite); HM (hematite); and CC (crystalline carbon).

treated as static environmental media for antibiotics sorption. The change of soil properties (such as soil compositions) should be incorporated when predicting the environmental behavior of antibiotics. This is especially true when human behavior is involved. For example, long-term intensified agricultural activities altered greatly soil properties, such as degree of clay dispersion, aggregate sizes, SOM stability, as well as mineral elemental contents (Abdollahi et al., 2014; Blazewicz-Wozniak et al., 2008). Soil organic matter (SOM) is the soil component with the primary interest when the sorption mechanisms of organic contaminants are investigated (Mechlinska et al., 2009; Pan et al., 2007). The interactions between SOM and antibiotics are reported to be significant (Bao et al., 2009; Pan et al., 2012a, 2012b). But different from hydrophobic organic contaminants, the sorption of antibiotics on mineral particle was also important. The mineral fractions usually showed comparable or even higher sorption than organic matter. SOM, such as low-molecularweight organic acids, could coat on mineral particles and compete with antibiotics for sorption sites (Zhang and Dong, 2008). We have presented that the organic matter-removed soil particles showed higher sorption than the original soil (Hou et al., 2010). The sorption of antibiotics on different mineral fractions will be important for predicting antibiotic behavior. Therefore, both organic and mineral composition change during tillage activities will alter antibiotic sorption characteristics on soil particles. However, previous studies mostly focus on the property change of organic fractions, with limited interest on soil mineral compositions. This study is specifically designed to investigate antibiotic sorption in soils after tillage. The same soil without tillage will also be collected for reference. Both organic and inorganic compositions will be

characterized and be incorporated in understanding antibiotic sorption mechanisms. OFL will be used as a model antibiotic because of its wide application (Pico and Andreu, 2007; Van Wierene et al., 2012), highconcentration occurrence in the environment (Yang et al., 2010), and strong sorption in soils/sediments in comparison to other antibiotics (Cirz et al., 2005). This study will provide useful information for antibiotic environmental behavior prediction and risk assessment. 2. Experimental section 2.1. Preparation of the adsorbents Soil samples with different tillage activities were collected in different locations of Yunnan province, southwest of China. The common tillage behavior in these sampling areas generally involves the input of soil amendments and ploughing. Yuanyang terrace area is a very famous farming area in Yunnan. The terrace area has been cultivated for more than one thousand years. Both soils from paddy field (Y-P) and dry farmland (Y-D) were collected in the area of latitude 102.74178–102.74451 and longitude 23.07307–23.07341. Natural soil (Y-N) from the same region without cultivation was also collected in the mountain area at higher elevation for comparison. A large area of panax pseudo-ginseng planting was located in Mengzi mountain area. Panax pseudo-ginseng planting was known as an intensified farming process with massive application of fertilizers and pesticides. Three-year Panax pseudo-ginseng planting generally involves 8–10 years fallow period before the next cultivation. Soil samples before (M-N) and after (M-P) three-year panax pseudo-ginseng planting were collected. In addition, soil A layer (M-A) and B layer (M-B) were collected in the mountain area without

Table 2 Fitting results of OFL sorption isotherms using Freundlich equation. logKF

Y-N Y-P Y-D M-N M-G M-A M-B B-N B-P B-D B-T B-N-M QZ MS KL ML

3.25 3.14 3.23 3.60 3.56 3.31 3.57 2.41 3.21 2.67 3.97 4.27 1.42 1.85 3.35 4.98

SE

0.01 0.04 0.02 0.04 0.04 0.03 0.02 0.05 0.03 0.03 0.02 0.02 0.02 0.05 0.02 0.04

n

0.461 0.421 0.446 0.524 0.463 0.627 0.531 1.080 0.689 0.813 0.509 0.336 0.478 0.404 0.380 0.289

SE

0.014 0.048 0.022 0.040 0.039 0.040 0.025 0.053 0.031 0.028 0.024 0.020 0.022 0.049 0.019 0.036

radj 2

0.992 0.883 0.969 0.928 0.926 0.955 0.971 0.970 0.974 0.985 0.981 0.961 0.970 0.836 0.965 0.911

Kd(L/kg) Ce = 0.1Cs

Ce = 0.01Cs

75.8 46.6 67.8 250 158 234 239 409 264 156 530 384 1.3 2.2 60.5 1510

263 177 243 748 544 552 705 340 540 240 1640 1770 4.2 8.7 252 7780

pHa

SE

pHo

6.51 6.73 6.78 6.38 6.68 6.97 6.60 8.04 7.08 7.50 6.23 6.61 6.80 9.40 6.57 6.11

0.10 0.09 0.08 0.16 0.10 0.05 0.12 0.20 0.09 0.13 0.10 0.25 0.11 0.22 0.14 0.21

5.55 6.19 5.41 5.45 6.31 6.93 6.32 8.21 6.13 7.54 5.37 nd

D. Zhou et al. / Science of the Total Environment 497–498 (2014) 665–670

60

were washed using deionized water until a negative test of chloride using AgNO3. The particles were freeze-dried, ground and contained in brown vials for future use. Selected characteristics of all the mineral particles were presented in Table 1.

: Yuanyang : Dianchi

50

Fraction (%)

Black: Nature soil Grey: Dry farmland

40

White: Paddy field

2.2. Batch sorption experiments

30

Batch sorption experiments of OFL were conducted for all the solid particles. OFL was dissolved in background solution of 0.1 M CaCl2 and 200 mg/L NaN3 as a stock solution (100 mg/L). This stock solution was diluted by the background solution to eight different concentrations (1–64 mg/L) in 4–40 mL vials with teflon-lined screw caps. The adsorbents were mixed with OFL solutions with different aqueous/solid ratios in the range of 40:1 to 40,000:1 depending on different types of solid particles based on preliminary experiments. All the vials were kept in dark and shaken in an air-bath shaker at 25 °C for 7 d (Pan et al., 2012a, 2012b). According to our previous experiments, the period of 7 d is sufficient to reach apparent OFL sorption equilibrium. After 7 d equilibration, all the vials were centrifuged at 2000 g for 15 min, and the supernatants were subjected to quantification of OFL on high performance liquid chromatography (HPLC, Agilent Technologies 1200).

B-N

20 B-D B-P

10 0

667

Y-N Y-D

QZ

MS

KL

Y-P

ML

CC

Mineral compositions Fig. 1. Soil mineral composition change with tillage in Yuangyang and Dianchi soils. QZ, MS, KL, ML and CC denote quartz, muscovite, kaolinite, montmorillonite and crystalline carbon, respectively. Although the parent materials were different for soils from two locations, the mineral compositions tended to be similar after human agricultural activities.

agricultural activities. The soil samples collected in Mengzi were located within the area of latitude 103.78113–103.78977 and longitude 23.40204–23.41552. Samples of paddy soil (B-P), dry farmland soil (BD), and soil without tillage (B-N) were collected near Dianchi Lake in the area of latitude 102.69186–102.69202 and longitude 24.73001– 24.73101. Peat soil (B-T) around Dianchi Lake was frequently applied as soil amendment and thus was also collected. All the surface soils were collected within the depth of 0–10 cm. All the collected samples were air-dried, ground and sieved through a 0.8-μm sieve. Visible plant residues were picked out manually. The soil sample of B-N was treated to remove carbonates. Briefly, the soil particles were washed successively using dilute hydrochloric acid and distilled water until pH 7. The soil particles were then freeze-dried. The modified NB was noted as MNB. The model mineral particles used in the study were quartz (QZ), muscovite (MS), kaolinite (KL) and montmorillonite (ML). MS was purchased from Shanghai Yuejiang Taibai Chemical Product Co., Ltd. and the other mineral particles were from Aladdin Co. All the mineral particles were saturated using Ca2+ before use to avoid the influence from impurity ions. Briefly, sufficient CaCl2 was dissolved in deionized water reaching its saturation as a stock solution. Mineral particles were mixed with CaCl2 solution and the mixtures were shook every 5 h for two days. The mixture was centrifuged at 2000 g for 15 min, and the supernatant was replaced by CaCl2 stock solution. This process was repeated another two times, and then the solid particles

2.3. Characterization of adsorbents The bulk elemental compositions of the solid particles were analyzed using X ray fluorescence (8420, ARL, Switzerland). The organic elemental compositions were obtained using an elemental analyzer (Vario EL III, Elementar, Germany). Inductively coupled plasma optical emission spectrometer was combined with X ray diffraction (D/max2200, Rigaku, Japan) analysis to determine the mineral compositions of soil particles. The specific surface area was also measured by BET method (Autosorb-1C, Quantachrome) using N2 as the carrier gas. The pH, pHzpc and CEC of soils were measured using traditional chemical methods (Qiao, 2012). All the results of soil characteristics are listed in Table 2. 2.4. Quantification of OFL The concentration of OFL was measured by HPLC equipped with a reversed-phase C8 column (5 μm, 4.6 × 150 mm). The mobile phase was acetonitrile:deionized water (10:90) with 0.8% acetic acid. The injected sample volume was 20 μm with the flow rate of 1 mL/min. The wavelength of the detector was 286 nm. No apparent OFL degradation was observed in the 7-d equilibration period, as suggested by no significant decrease of peak area and clear HPLC spectrum (Van Wierene et al., 2012).

5

5

logSe (mg / kg)

M-P M-N 4

4

3

2

B-T B-P B-D B-N -1

0

1

Y-P Y-D Y-N M-A M-B

3

2

-1

0

1

2

logCe (mg / L) Fig. 2. OFL sorption in different soils. The solid lines are fitting results using Freundlich equation. Soil samples from Yuanyang, Mengzi and Dianchi areas are noted as the first letters of Y, M, and B, respectively. Natural soils without tillage, paddy soils, and dry forming land soils are noted as the second letters of N, P, and D, respectively. M-A and M-B are A and B horizons in the mountain area without human activities, respectively. B-T is the peat collected in the Dianchi area.

Fraction (%)

668

D. Zhou et al. / Science of the Total Environment 497–498 (2014) 665–670

100

50

Y-P Y-D Y-N M-P M-N M-A M-B B-P B-D B-N B-T

0

6.10

8.28

pH Fig. 3. The change of soil pH values after sorption. The pH values generally decreased after sorption except B-N. Ion exchange may be involved in OFL sorption in most of the soils and van der Waals interactions were important in B-N.

2.5. Data processes Model regression was performed using SigmaPlot 10.0 and statistical analysis was conducted with the aid of SPSS 15.0. Freundlich model was applied for adsorption isotherm analysis: Freundlich model :

log Q e ¼ logK F þ n logC e

ð1Þ

where Qe (mg/kg) and Ce (mg/L) are the equilibrium solid phase and aqueous phase concentrations, respectively. KF [(mg/kg)/(mg/L)n] is the Freundlich affinity coefficient and n is the Freundlich nonlinearity factor. Because the numbers of data points were different for different sorption isotherms, common coefficient of determination (r2) could not be compared directly (Pan and Xing, 2010). The adjusted r2 (r2adj) was calculated and compared: h
 i 2 2 r adj ¼ 1– 1–r ðm–1Þ=ðm–b–1Þ

ð2Þ

where m is the number of data points used for fitting, and b is the number of coefficients in the fitting equation. 3. Results and discussion 3.1. Soil characterization Soil mineral compositions were measured and presented in Table 1. For all the collected soil samples, QZ, MS, KL, and ML are the main mineral components, which generally accounted for more than

Kd (L/kg)

1000

10 0.1

10000

A

100

1000

r = 0.418 P = 0.201

1

10

80% of soil mass. QZ was the most abundant mineral component in most soils. Soil mineral compositions showed very clear geological groups. For example, soils from Yuanyang area (Y-P, Y-D, and Y-N) contain significant contents of HB, which was not observed for other soils. Mengzi soil samples showed much higher KL content than other soils. The cluster analysis (SPSS 15.0) based on soil mineral compositions could clearly group soils from different geographic locations. Thus, soil properties were compared among the soils from the same geographic area. Human agricultural activities significantly changed soil mineral compositions. The main soil mineral fractions were compared between natural, dry farmland, and paddy soils from Yuanyuang and Dianchi areas in Fig. 1. It is interesting to notice that although soil mineral compositions from these two regions differed greatly, these compositions were similar after tillage, especially for paddy soils. For example, the natural soil from Yuanyang (Y-N) showed MS content was almost two times higher than that of the natural soil from Dianchi (B-N) (35.2% vs. 19.5%). After long-term tillage in paddy, the MS contents from the two locations were very similar (27.0% vs. 25.5%). On the other hand, B-N has KL content 5 times higher than that of Y-N (27.4% vs. 5.3%). The paddy tillage behavior brought soils from these two locations similar KL contents (8.8% vs. 8.3%). Among all the soils, B-N is the only soil that has a significant fraction of CT (13.0%). CT was not detectable in soils after tillage (B-D and B-P). The main content of CT is CaCO3, which could be easily weathered or reacted in the acidified environment during the tillage (as indicated by decreased pHzpc after tillage). It should be noted that the increase of a certain mineral composition may not suggest the generation of this composition, but the decreased mass contents of other fractions. For example, QZ in Dianchi soils increased to over 40% after tillage. The reason may be that QZ is rather stable during tillage, while the other mineral contents, such as CT, were chemically altered or degraded. Soil surface areas (SSA) of the investigated soils were mostly in the range of 18.4–82.0 m2/g. A general observation is that SSA decreased greatly after tillage, while organic carbon content increased. This phenomenon may be understood from the common knowledge that SOM blocks the pores in mineral particles and promotes the aggregation of soil particles. Although SOM is abundant of micropores, these pores were not available in SSA measurement using N2 (Pignatello, 1998). The decreased SSA with increased fOC was also reported in previous studies (Drillia et al., 2005). The only exception was soils from panax pseudo-ginseng planting area. The harvest of panax pseudo-ginseng root prevented organic carbon from returning to soil. For acidic soils, the tillage behavior did not change soil pHzpc significantly. However, the soil collected near Dianchi Lake contains CT of 13.0% and its pHzpc was 7.36. The agricultural activities in dry farmland decreased soil pHzpc to 6.66. This acidification process was even more distinct in paddy soil as evidenced by pHzpc decreased to 2.89 for B-P, with comparable values for Y-P.

100

100 0.1

B r = 0.519 P = 0.056

1

10

100

fOC (%) Fig. 4. The relationship between OFL sorption coefficients (Kd) and soil organic carbon content. Panels A and B present Kd values calculated at Ce = 0.1 Cs and Ce = 0.01 Cs, respectively.

D. Zhou et al. / Science of the Total Environment 497–498 (2014) 665–670

6.0

pH after sorption probably suggested that ion exchange was involved in OFL sorption, i.e., the exchange of OFL with protons in soil. The rather stable pH and linear n value may suggest that van der Waals interactions were involved in OFL sorption on B-N.

logSe (mg / kg)

B-N-M

4.0

QZ

3.3. OFL sorption as affected by soil organic carbon content

MS KL ML

2.0

0.0 -1.5

669

0.0

1.5

3.0

logCe (mg / L) Fig. 5. OFL sorption in different minerals. The solid lines are fitting results using Freundlich equation. QZ, MS, KL, and ML denote quartz, muscovite, kaolinite, and montmorillonite, respectively.

3.2. OFL sorption in different soils The sorption isotherms of OFL in soils were presented in Fig. 2 and the fitted results using Freundlich equation were illustrated as solid lines in this figure. The obtained parameters are listed in Table 2. The r2adj values were mostly higher than 0.95, indicating good fitting performance of Freundlich equation to the sorption isotherms. Y-P gave very low r2adj value of 0.88. OFL-Y-P sorption isotherm showed deviation to Freundlich-type sorption with an obviously curved feature, probably indicating the saturation of OFL sorption sites because of the low sorption (with the lowest Kd values in comparison to other soil samples). A general nonlinear sorption was observed for OFL sorption in the investigated soils, with most of the n values lower than 0.7. The only exception was B-N (n = 1.080 ± 0.053). The treated B-N (B-N-M) showed distinct nonlinear sorption to OFL with n values of 0.336 ± 0.020. Although acid treatment could not dissolve SOM significantly, the C content decreased from 1.53% (B-N) to 0.69% (B-NM). It should be noted that the CT content was effectively removed through acid treatment using diluted HCl according to XRD measurement. Thus, the decreased C content may be related with CT removal. With the evidence of both C content change and n value change after acid treatment, we propose that CT forms complexes with SOM and this organo-mineral complex showed rather linear sorption to OFL. The removal of CT also resulted in the release of SOM on the CT surface, which consequently resulted in stronger nonlinear sorption. It is interesting to notice that soil pH values decreased after OFL sorption for all the soils with the only exception of B-N (Fig. 3). The decreased

The relationship between sorption coefficients and fOC was widely discussed in literature when investigating the sorption behavior of organic contaminants. Hydrophobic organic contaminants (HOCs) usually showed significant correlation with soil organic carbon content (Schwarzenbach and Westall, 1985; Pan et al., 2006). Mader et al. indicated that the contribution of soil mineral components to HOCs sorption could be neglected with soil organic carbon content higher than 0.1% (Mader et al., 1997). The relationship between OFL sorption and fOC was also analyzed in this study. OFL sorption was characterized using single-point sorption coefficients calculated at Ce = 0.1 Cs (K1d) and Ce = 0.01 Cs (K2d) as indicated in Table 2. The correlation analysis between K1d and fOC provided a correlation coefficient (r) of 0.418 and K2d–fOC relationship showed r = 0.518 in Fig. 4. Both correlations were not significant at P b 0.05. This result suggested that organic carbon had limited contribution to OFL sorption in soils. On the contrary, SOM may have decreased OFL sorption. For example, M-A and M-B have comparable mineral compositions, and organic carbon content of M-A was much higher than M-B. But the sorption on M-B was higher than M-A. The above statement could be further evidenced when comparing B-N and B-N-M sorption. Although acid treatment removed more than half of the organic carbon (0.69% vs. 1.53%) in B-N, OFL sorption increased 5 times at Ce = 0.01 Cs (1770 L/kg vs. 340 L/kg). Thus, organic carbon removal may have exposed more sorption sites for OFL. The above results suggested a very important contribution of inorganic minerals to OFL sorption, which will be further addressed in the next section. As discussed earlier in this paper, the comparison between soils from different locations may not be proper because of the very complicated soil compositions. For soils from three locations, the natural soils without tillage and with limited vegetation always showed high sorption to OFL. Tillage activities or densely overgrown vegetations (such as M-A) greatly decreased OFL sorption. For example, OFL sorption in Y-D and Y-P was lower than Y-N. Both naturally growing vegetations (M-A) and human intensified farming activities (M-G) decreased OFL sorption in Mengzi soils. But for samples with much less biological activities (M-B, as indicated by much lower organic carbon content), OFL sorption was similar to natural soils (M-N). This result may suggest that the leaching of OFL through soil column could be limited because of the high sorption of OFL in the B layer in comparison to A layer. However, the general conclusion of negative relationship between fOC and OFL sorption could not be made. In reference to previous researches that the adsorption of antibiotics in soils is not determined by organic

160

Kdm/Kd 100 (%)

A 120

ML KL MS QZ

200

ML KL MS QZ

B

80 100 40 0

Y-P Y-D Y-N M-P M-N M-A M-B B-P B-D B-N B-T

0

Y-P Y-D Y-N M-P M-N M-A M-B B-P B-D B-N B-T

Soils Fig. 6. The contributions of different mineral fractions to the overall sorption coefficients. Panels A and B present the ratios of (Km d × P / Kd) calculated at Ce = 0.1 Cs and Ce = 0.01 Cs, respectively.

670

D. Zhou et al. / Science of the Total Environment 497–498 (2014) 665–670

fractions along (Wu et al., 2014), the following analysis focusing on the role of mineral compositions is conducted. 3.4. OFL sorption as affected by soil mineral composition OFL sorption isotherms on four dominant mineral compositions are presented in Fig. 5. These isotherms were well fitted using Freundlich equation and the fitting results are listed in Table 2. Single-point sorption coefficients were also calculated to facilitate the comparison between different adsorbents. OFL sorption coefficients on mineral fractions varied more than four orders of magnitudes. ML showed much higher sorption than the other minerals, probably because of its very high surface area and layer structures (Wu et al., 2010). The contributions of different mineral fractions to the overall m sorption coefficients were calculated as Km d × P / Kd (Kd is the sorption coefficients of mineral particles and P is the percentage of mineral fraction in the soil) and the resulted contributions are presented in Fig. 6. The contribution of ML to the overall sorption exceeded 100% for some of the samples. It should be noted that this calculation provided the highest estimation of sorption contribution without any consideration of interactions between soil components (such as the coating of organic matter on mineral particles). It was reported that SOM could adsorb OFL strongly (Pan et al., 2012a, 2012b). When the sorption contributed by SOM was accounted for in the above calculation, the summarized sorption contributed by soil components will show even higher sorption than the intact soil. We have presented in our previous study that antibiotics may compete with SOM for sorption sites on mineral particle surface (Wu et al., 2014). Thus, organo-mineral complex formation may decrease the apparent OFL sorption. However, SOM may form organo-mineral complexes of different properties and how these complexes interact with OFL could not be provided in this study. It was also noted in Fig. 6 that the ratios of (Km d × P / Kd) were higher at Ce = 0.01 Cs in comparison to those at Ce = 0.1 Cs. This result could be understood from the lower n values of OFL sorption on mineral particles than on soils. Both MS and QZ were dominantly comprised by SiO2 and they showed very low sorption to OFL, with Kd values below 10 mg/L. For the soils investigated in this study, MS and QZ accounted for 37–79% of soil mass. Clearly, the contributions of these two fractions to OFL sorption in soil were very low. Among the four mineral investigated, the ML contents were the lowest. However, the calculated sorption contributed by ML was the highest, with Km d × P / Kd values generally higher than 50%. This result suggested a very important contribution of ML to OFL sorption. However, the relationship between ML content and sorption coefficient was not observed, most likely because of the interactions between mineral particles and SOM. Again, this result suggested that the interactions between soil components provide very important information to understand antibiotic sorption behavior in soils/sediments. Acknowledgments This research was supported by the National Scientific Foundation of China (41173124, 41222025, and 41303092), the Recruitment Program of High-Qualified Scholars in Yunnan (2010CI109), and the General Program of Science and Technology Department of Yunnan Province (2013FZ004).

Appendix A. Supplementary data Supplementary data related to OFL physicochemical properties and sampling locations can be found at: http://dx.doi.org/10.1016/j. scitotenv.2014.07.130 References Abdollahi L, Schjonning P, Elmholt S, Munkholm LJ. The effects of organic matter application and intensive tillage and traffic on soil structure formation and stability. Soil Tillage Res 2014;136:28–37. Bao YY, Zhou QX, Wang YY. Adsorption characteristics of tetracycline by two soils: assessing role of soil organic matter. Aust J Soil Res 2009;47:286–95. Blazewicz-Wozniak M, Kesik T, Wach D, Konopinski M. The influence of conservation tillage on the mineral elements content in soil and chemical composition of onion. Acta Sci Polonorum-Hortorum Cultus 2008;7:61–72. Cirz RT, Chin JK, Andes DR, de Crecy-Lagard V, Craig WA, Romesberg FE. Inhibition of mutation and combating the evolution of antibiotic resistance. Plos Biol 2005;3: 1024–33. Dhar N, McKinney JD. Microbial phenotypic heterogeneity and antibiotic tolerance. Curr Opin Microbiol 2007;10:30–8. Drillia P, Stamatelatou K, Lyberatos G. Fate and mobility of pharmaceuticals in solid matrices. Chemosphere 2005;60:1034–44. Hou J, Pan B, Niu X, Chen J, Xing B. Sulfamethoxazole sorption by sediment fractions in comparison to pyrene and bisphenol A. Environ Pollut 2010;158:2826–32. Kemper N. Veterinary antibiotics in the aquatic and terrestrial environment. Ecol Indic 2008;8:1–13. Mader BT, Goss KU, Eisenreich SJ. Sorption of nonionic, hydrophobic organic chemicals to mineral surfaces. Environ Sci Technol 1997;31:1079–86. Mechlinska A, Gdaniec-Pietryka M, Wolska L, Namiesnik J. Evolution of models for sorption of PAHs and PCBs on geosorbents. Trac-Trends Anal Chem 2009;28:466–82. Nikolaou A, Meric S, Fatta D. Occurrence patterns of pharmaceuticals in water and wastewater environments. Anal Bioanal Chem 2007;387:1225–34. Pan B, Xing B. Adsorption kinetics of 17 alpha-ethinyl estradiol and bisphenol A on carbon nanomaterials. I. Several concerns regarding pseudo-first order and pseudo-second order models. J Soils Sediment 2010;10:838–44. Pan B, Xing B, Liu W, Tao S, Lin X, Zhang Y, et al. Two-compartment sorption of phenanthrene on eight soils with various organic carbon contents. J. Environ. Sci. Health B 2006;41:1333–47. Pan B, Xing B, Liu W, Xing G, Tao S. Investigating interactions of phenanthrene with dissolved organic matter: Limitations of Stern–Volmer plot. Chemosphere 2007;69: 1555–62. Pan B, Qiu M, Wu M, Zhang D, Peng H, Wu D, et al. The opposite impacts of Cu and Mg cations on dissolved organic matter–ofloxacin interaction. Environ Pollut 2012a; 161:76–82. Pan B, Wang P, Wu M, Li J, Zhang D, Xiao D. Sorption kinetics of ofloxacin in soils and mineral particles. Environ Pollut 2012b;171:185–90. Peterson JW, O'Meara TA, Seymour MD, Wang W, Gu BH. Sorption mechanisms of cephapirin, a veterinary antibiotic, onto quartz and feldspar minerals as detected by Raman spectroscopy. Environ Pollut 2009;157:1849–56. Pico Y, Andreu V. Fluoroquinolones in soil—risks and challenges. Anal Bioanal Chem 2007; 387:1287–99. Pignatello JJ. Soil organic matter as a nanoporous sorbent of organic pollutants. Adv Colloid Interface Sci 1998;77:445–67. Qiao S. Experimental instructions to soil physical and chemical property analysis. Beijing: China University of Geosciences Press; 2012. Schwarzenbach RP, Westall J. Sorption of hydrophobic trace organic-compounds in groundwater systems. Water Sci Technol 1985;17:39–55. Thiele-Bruhn S. Pharmaceutical antibiotic compounds in soils—a review. J Plant Nutr Soil Sci 2003;166:145–67. Van Wierene EM, Seymour MD, Peterson JW. Interaction of the fluoroquinolone antibiotic, ofloxacin, with titanium oxide nanoparticles in water: adsorption and breakdown. Sci Total Environ 2012;441:1–9. Wu QF, Li ZH, Hong HL, Yin K, Tie LY. Adsorption and intercalation of ciprofloxacin on montmorillonite. Appl Clay Sci 2010;50:204–11. Wu D, Li H, Liao SH, Sun XL, Peng HB, Zhang D, et al. Co-sorption of ofloxacin and Cu(II) in soils before and after organic matter removal. Sci Total Environ 2014;481:209–16. Yang JF, Ying GG, Zhao JL, Tao R, Su HC, Chen F. Simultaneous determination of four classes of antibiotics in sediments of the Pearl Rivers using RRLC-MS/MS. Sci Total Environ 2010;408:3424–32. Zhang JQ, Dong YH. Effect of low-molecular-weight organic acids on the adsorption of norfloxacin in typical variable charge soils of China. J Hazard Mater 2008;151:833–9.