Depositional history of sedimentary linear alkylbenzenes (LABs) in a large South American industrial coastal area (Santos Estuary, Southeastern Brazil)

Environmental Pollution 158 (2010) 3355e3364

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Depositional history of sedimentary linear alkylbenzenes (LABs) in a large South American industrial coastal area (Santos Estuary, Southeastern Brazil) César C. Martins a, b, *, Márcia C. Bícego b, Michel M. Mahiques b, Rubens C.L. Figueira b, Moyses G. Tessler b, Rosalinda C. Montone b a b

Centro de Estudos do Mar da Universidade Federal do Paraná, Caixa Postal 50.002, 83255-000, Pontal do Sul, Pontal do Paraná, PR, Brazil Instituto Oceanográfico da Universidade de São Paulo, Praça do Oceanográfico, 191, 05508-900, São Paulo e SP, Brazil

The contamination history of a large South American industrial coastal area indicated by molecular indicator of sewage input.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 March 2010 Received in revised form 10 July 2010 Accepted 26 July 2010

This paper reports the reconstruction of the contamination history of a large South American industrial coastal area (Santos Estuary, Brazil) using linear alkylbenzenes (LABs). Three sediment cores were dated by 137Cs. Concentrations in surficial layers were comparable to the midrange concentrations reported for coastal sediments worldwide. LAB concentrations increased towards the surface, indicating increased waste discharges into the estuary in recent decades. The highest concentration values occurred in the early 1970s, a time of intense industrial activity and marked population growth. The decreased LAB concentration, in the late 1970s was assumed to be the result of the world oil crisis. Treatment of industrial effluents, which began in 1984, was represented by decreased LAB levels. Microbial degradation of LABs may be more intense in the industrial area sediments. The results show that industrial and domestic waste discharges are a historical problem in the area. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Linear alkylbenzenes Sediments Sewage Santos estuary

1. Introduction The Santos Estuary, on the Southeast coast of Brazil, lies close to one of the most important economic areas of the country, the Baixada Santista metropolitan region (Fig. 1). The permanent population is over 1,200,000 and during the summer there is an estimated floating population of 780,000 (Luiz-Silva et al., 2002; Hortellani et al., 2005) due to intense tourism. The largest commercial harbour in South America and the main petrochemical and metallurgical industrial centre in the country, with approximately 1100 industries, are also in this estuary (CETESB, 2001). There have been anthropogenic activities in this area for the last 120 years, since the construction of Santos Harbour. However, urbanization and industrialization around the estuary increased 60 years ago when the Cubatão Industrial Complex was built and the Santos Harbour was enlarged (Luiz-Silva et al., 2006; Hortellani et al., 2008). During the early 1980s, the region became known as one of the worst polluted in the world, as a result of the high levels of organic contaminants in the air and water (Blacksmith Institute, 2006). Since 1970, a large quantity of solid waste from the domestic activities of the Baixada Santista has been disposed in an area * Corresponding author. E-mail address: [email protected] (C.C. Martins). 0269-7491/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2010.07.040

located on the estuary banks. Drainage from this area after successive rainfalls results in a major discharge of pollutants into Santos Estuary (Martins et al., 2007). Unregulated discharges of sewage occur through drains and channels and the ships and docks of Santos Harbour are a significant source of domestic waste in the estuary. Approximately 1.5 million tons of raw sewage is drained to the main river that discharges into the estuary (Medeiros and Bícego, 2004a). A wide range of organic contaminants, (e.g. petroleum hydrocarbons, benzothiazoles, polychlorinated biphenyls and polybrominated diphenyl ethers) have entered the environment as a result of the discharge of untreated industrial effluents and domestics wastes (Eganhouse and Sherblom, 2001; Ni et al., 2009; Zhang et al., 2010). Therefore, the depositional history of molecular markers related to domestic and industrial wastes may be used to reconstruct the input of organic contaminants into the environment (Eganhouse and Kaplan, 1988; Zeng and Venkatesan, 1999; Hartmann et al., 2005). Linear alkylbenzenes (LABs) were first discharged to the environment in the early 1960s as a by-product of linear alkylbenzene sulfonate (LAS) detergents which are the most widely used anionic surfactants in the world. There has been widespread interest in their use as markers of sewage effluent in the marine environment (Eganhouse, 1986; Takada et al., 1994; Zeng et al., 1997). Generally,

3356

C.C. Martins et al. / Environmental Pollution 158 (2010) 3355e3364

N

Rio de Janeiro

B

2 3 .0 º S

São Paulo

Guanabara Bay

Brazil

2 4 .0 º S

São Sebastião Island

Cananéia

(C)

2 5 .0 º S

Atlantic Ocean

(B)

Atlantic Ocean

b 48.0º W

46.0º W

42.0º W

a D

Cubatão

44.0º W

canal de Bertioga

23.9º S

LSR

CQ1

CQ2 C B S ão Vicente

V i c ente de Carvalho

San t os

A

Potential sources of organic pollution: A – Santos Harbour B – Alemoa Oil Terminal C - Casqueiro Garbage Disposal (Alemoa hill) D – Cubatão Industrial Complex

G uar u j á Praia Grande

24.0º S

Baía de Santos Santos

Bay

2 km

c 046.4º W

046.3º W

Fig. 1. Map of the study area showing the sediment sampling sites (CQ1, CQ2 and LSR) in the Santos Estuary and (a) location in Brazil, (b) the Southeastern Brazilian coast, and (c) Baixada Santista.

a set of 26 secondary phenylalkanes with chain lengths ranging from 10 to 14 carbon atoms occurs as trace components in detergent formulations and is introduced into waste streams from detergent use (Takada and Eganhouse, 1998). Under anaerobic conditions they may persist for over 20 years as evidenced by downcore distributions (e.g. Heim et al., 2004; Hartmann et al., 2005) and laboratory experiments in which >99% remained after 50 days incubation (Steber et al., 1995). Recent work indicated that LABs may not be as persistent under reducing conditions. The half life of post-sedimentary LABs was estimated to be 7.8e10.8 years (Eganhouse and Pontolillo, 2008). However, further research could be needed to determine whether the LABs susceptibility to degradation is the same in different environments (Boonyatumanond et al., 2007). The main LAB consumer worldwide is the detergent industry. Demand totalled 2.3
106 tonnes in 1999 and is expected to outstrip the current manufacturing capacity, reaching 3.4
106 tonnes y1 by 2010 (Johnson, 2003). LABs production in Brazil began in 1981 with 35,000 tonnes and it has increased, reaching 220,000 tonnes in 2008, representing around 10% of global production (Penteado et al., 2006). Despite the possible sewage contamination, few data are available describing the input of organic markers of municipal waste such as LABs (Medeiros and Bícego, 2004a), hydrocarbons and PCBs (Bícego et al., 2007) and coprostanol (Martins et al., 2008a) into the study area. The existing data involve upper surficial sediments.

Here we present vertical distributions of long-chain linear alkylbenzenes in sediment cores from the Santos Estuary, showing the historical increase of human activities and contamination in the area. The profiles were used to demonstrate the temporal distribution in recent decades and to study the fate of LABs in estuarine sediments in this region. Dated sediment cores have the potential to provide detailed chronologies of organic contaminant input as long as diagenetic processes of bioturbation, molecular diffusion, and biotransformation are (or are considered to be) negligible (Elsenreich et al., 1989). The historical record of pollution level is also needed to evaluate countermeasures (e.g. regulation of chemicals). Furthermore, understanding the historical trends of pollution is helpful to predict further tendencies of pollution (Boonyatumanond et al., 2007). 2. Experimental methods 2.1. Sampling Three vibro-cores (3 m long, 0.75 mm wide aluminium barrels, 10 cm i.d.) named CQ1, CQ2 and LSR were collected in the Santos Estuary, São Paulo, Brazil (Fig. 1) in September, 2002. Cores CQ1 and CQ2 were collected on the estuary banks, near a solid waste disposal area used by the Baixada Santista cities and close to small towns. The LSR core was collected close to an intensive shipping area where waste is discharged from metallurgical and petrochemical industries. It is located close to petroleum, kerosene

C.C. Martins et al. / Environmental Pollution 158 (2010) 3355e3364 and alcohol storage facilities, near a channel where loading and unloading operations are carried out by several factories in the Cubatão complex (Martins et al., 2007). The cores were sub-sampled at 2-cm intervals, wrapped in aluminium foil (baked at 450  C) and stored at 20  C until analysis in the laboratory. Sediments were analyzed every 2-cm section starting at the surface (0e40 cm for CQ1 and 0e60 cm for LSR) and then in one sub-2 cm-sample every 20 cm section until the bottom of the core. Samples from CQ2 were analyzed every 1 cm until 15 cm depth. The sediment samples were freeze-dried, carefully homogenized in a mortar, sieved through a stainless steel mesh (250 mm) and stored at room temperature in glass bottles (for LABs) and plastic containers (for gamma spectrometry counting). 2.2. Particle size and total organic carbon Particle size distributions in the samples were measured by laser granulometry, using a Malvern Mastersizer 2000 analyzer. Total organic carbon was determined with a LECO CNS200 Analyzer after the total elimination of the calcium carbonate of the samples with 1 mol L1 HCl solution. Sulphametazin (C12H14N4O2S) standards were analyzed for each set of 30 samples to verify accuracy. 2.3. Dating of the sediment cores Sediment samples (20 g) were counted for 90,000e120,000 s using a hyper-pure Ge detector (model GEM60190, by EGG&ORTEC) with a 1.9 keV resolution for the 1332.40 keV 60Co peak. Cesium-137 activity was assayed by means of its peak at 661 keV (Figueira et al., 1998). The detailed method (calibration, detector counting efficiency and errors) was fully described in Martins et al. (2010). International Atomic Energy Agency (IAEA) reference materials were employed to determine the detector counting efficiency in the radionuclide photopeak region. 2.4. LABs: extraction and fractionation The analytical procedure for LAB analyses is based on UNEP (1992) and Martins et al. (2002). Briefly, about 25 g of dry sediment samples were Soxhlet-extracted with hexanes (95% n-hexane) and dichloromethane (J. Baker both) (1:1) for a period of 8 h. The LAB recovery surrogate consists of 1-C19 LAB (Supelco, 99% purity). LAB structures are expressed as “n-Cm”, where “n” is the phenyl substitution position on the alkyl chain and “m” is the number of alkyl carbons. The solvent extract was concentrated in a rotary evaporator to approximately 2 mL. The extract was fractionated by adsorption liquid chromatography into aliphatic and aromatic hydrocarbons using a column of alumina (1.8 g) and silica-gel (3.2 g), and hexanes and 3:1 dichloromethane/hexanes for aliphatic (F1) and aromatic (F2) fractions as eluent, respectively. The fractions were concentrated once more in a rotary evaporator, transferred to a vial, and then the volume was adjusted to precisely 1 mL using a stream of N2 gas. 2.5. LABs: instrumental analyses and quality assurance procedures The LAB analyses were performed with an Agilent 6890 gas chromatograph coupled to an Agilent 5973N mass spectrometer and an Agilent Ultra-2 capillary fused silica column coated with 5% diphenyl/dimethylsiloxane (50 m, 0.32 mm ID and 0.17 mm film thickness). Helium was used as the carrier gas. The oven temperature was programmed from 40  C, holding for 2 min, from 40 to 60  C at 20  C min1, then to 250  C at 5  C min1, and finally to 320  C at 6  C min1, held for 20 min.

3357

The data acquisition was undertaken in the SIM mode and the Agilent Enhanced Chemstation G1701 CA was used to perform the measurements. Calibration was based on an external standard solution containing 1-Cm-LABs (m ¼ 10e14) (Supelco, 99% purity) at different concentrations (0.25, 0.50, 0.75, 1.00, 1.50, 2.50 and 5.00 ng ml1). Compounds were identified by ion mass fragments (m/z 91, 92 and 105) and matching retention times with an effluent sewage sample containing all the n-Cm-LABs (m ¼ 10e13) analyzed (Fig. 2). The n-C14-LABs were identified by ion mass fragments and estimated retention index based on general literature. Quantitation ions used for the LABs were m/z 91 and 105. Potential TAB interference was absent or present at low abundance, confirmed by the data acquisition of some samples (n ¼ 10) undertaken in the mode SCAN and by the monitoring of characteristics TABs ions (e.g. m/z 119) in the mass spectra (Zeng and Yu, 1996; Zeng et al., 1998). Procedural blanks did not contain contaminant peaks which interfere with the analyses of the target compounds. Surrogate recovery, based on the relationship with an internal standard (tetradecene) added at the end of laboratory analyses, ranged from 60.0 to 119.5%. Sediment and blank samples were spiked with a mixture of 1-Cm-LABs (m ¼ 10e14) (Supelco, 99% purity) and the standard recoveries ranged from 78.1 to 88.2%. Detection limits (DL), defined as the concentration value that corresponds to three times the standard deviation of replicate instrumental measurements for the analyte in blanks (Kimbrough and Wakakywa, 1993), ranged from 0.03 to 1.37 ng g1 for all the compounds analyzed. Precision ranged from 5.5 to 10.2% based on the analysis of five replicate samples from the LSR core (40e42 cm).

3. Results and discussion 3.1.

137

Cs sediment profiles

The vertical distribution of 137Cs activity in each core is presented in Fig. 3. The age of each core section was estimated based on the fact that the 137Cs can be used as a time-marker and that the maximum activity of this radionuclide corresponds, theoretically, to the interval between the years 1963 and 1965, the period of maximum fallout in the Southern Hemisphere as a result of nuclear weapon tests (Mackenzie, 2000; Abril, 2003). Based on the 137Cs activity, the average sedimentation rates obtained for sediments deposited above the 137Cs maximum were: 1.47  0.13 (CQ1), 0.47  0.04 (CQ2) and 1.24  0.13 (LSR), in cm y1. The uncertainties in the sedimentation rates were based on counting errors of 137Cs activity (9%). Previous study on sedimentation found rates between 1.9 cm y1 (1987e2004) and 7.6 cm y1 (1959e1976) in the Northeastern area of Santos Estuary, around 7 Km from CQ1 (Luiz-Silva et al., 2008). The authors related the anthropogenic Fe source to a steel plant, and alternatively, the annual steel industrial production (1966e2004) was compared to the Fe distribution along the core. The apparent discrepancy in relation to our results may be explained by the differences in the sampling sites, the technique used to obtain the sedimentation rates, hydrologic and depositional conditions. Higher sedimentation rates are expected in the north of Santos Estuary, as verified by

C12-LAB

C13-LAB

C11-LAB

Abundance 80000 60000

C10-LAB

40000

m/z 91 20000

30000 20000

m/z 105

10000

Time-->

22.00

23.00

24.00

25.00

26.00

27.00

28.00

29.00

30.00

31.00

32.00

Fig. 2. Gas chromatogram of effluent sewage sample containing all the n-Cm-LABs (m ¼ 10e13) analyzed. For m/z ¼ 91, dots indicate peaks of LABs as: n-C10-LAB (n ¼ 3e5); n-C11LAB (n ¼ 3e6); n-C12-LAB; (n ¼ 3e6)-C13-LAB. For m/z ¼ 105, dots indicate peaks of 2-Cm-LAB (m ¼ 10e13).

3358

C.C. Martins et al. / Environmental Pollution 158 (2010) 3355e3364

Fig. 3.

Cs activity concentration (in Bq kg1 dry weight) and sedimentation rate (SR) for different sediment cores from the Santos Estuary, Brazil.

137

Luiz-Silva et al. (2008), due to the proximity to the sources of land material (Fulfaro and Ponçano, 1976). The 210Pb dating techniques for determining sediment accumulation rates are not reliable for the Santos estuary due to high levels of radionuclide contamination from industrial wastes (Machado et al., 2008). 3.2. Grain size and total organic carbon The total organic carbon (TOC) concentration and % silt þ clay are shown in Table 1 and Fig. 4. An irregular vertical distribution in the % silt þ clay was observed in the CQ1 core. The lowest silt þ clay content (70%) and minimum (5.00%) at the middle of the core, between 33 and 37 cm and 19e25 cm, and in the surface layers (0e7 cm). The human settlement and solid waste disposal at Alemoa hill, where the CQ1 core was taken, may explain the increased fine sediment and organic material inputs in this study area. Since the 1970s, this area has received large quantities of solid wastes produced by the inhabitants of São Vicente and Santos, with a population estimated at about 330,000 in the early 1970s and 720,000 in 2000 (Couto, 2003). The CQ2 and LSR cores had similar TOC concentrations, varying between 2.39 and 3.19 (mean value: 2.74  0.18; RSD ¼ 6.8%) and 2.20 and 3.86 (mean value: 3.28  0.32; RSD ¼ 9.7%), respectively. The vertical distribution showed no significant variations towards the deepest layers, confirmed by low relative standard deviation (RSD) values calculated for each core. On the other hand, the sediment profile of % silt þ clay showed a small variation in the distribution of the fine fractions. For the CQ2 core, the highest percentage of fine sediments occurred at depths >17 cm but remained constant between this section and 5 cm, where silt þ clay content presented a small increase followed by a minimum at 3 cm. The recent layers (65%), except at 35 cm, where the silt þ clay content exhibited a minimum (33.1%). Towards the bottom, the fine fraction showed little variation, ranging from 35.5 to 47.9% (mean value: 41.1  4.0; RSD ¼ 9.9%). The site where the CQ2 core was collected may be considered to present a more homogeneous sedimentation pattern when compared to CQ1 and LSR; it is possible that the lower sedimentation rates at this site obscure the high frequency temporal changes in sedimentation. A possible increase in the amount of fine sediments in recent layers due to anthropogenic activities that were performed close to where the CQ1 was collected was not verified for core CQ2, despite the proximity of the CQ1 and CQ2 cores. However, the grain size variation in the LSR core may be associated with periodic dredging activities, mainly between 1980 and 1996, along the main axis (EasteWest) of the estuary. This could indirectly affect the sediment composition near the LSR coring site. The relatively low variation in organic carbon concentration in the CQ2 and LSR cores suggests that natural inputs of organic matter may be more important than anthropogenic sources, because background concentrations are found in the most recent sections even though human activities have increased around both areas in the last few decades. Organic carbon content is commonly associated with the amount of silt and clay present in sediments. Fine sediment particles have relatively larger surface areas and can adsorb colloidal and dissolved organic matter forming sedimentary complexes (Kowalska et al., 1994). According to Pearson correlation coefficients between TOC and silt þ clay, this tendency was not verified in the Santos Estuary sediment cores. The correlation coefficients were: 0.05 (CQ1), 0.79 (CQ2) and 0.02 (LSR). These results suggest that sediment grain size was not a determining factor for organic carbon deposition in CQ1 and LSR, as verified by previous studies (e.g. Medeiros and Bícego, 2004a). The negative relationship between TOC and silt þ clay in the CQ2 core suggests that coarse plant particles contribute to organic matter composition, as may be expected for an intertidal area close to the mangrove forest (Luiz-Silva et al., 2008). These particles may represent aggregates and flocs formed by clay minerals and organic matter and/or decomposing particulate organic matter, including fragments of mangrove leaves, trunks and roots (Glasby et al., 2004). Despite several known anthropogenic sources of organic matter, the terrigenous input derived from higher plant waxes from mangrove vegetation is extremely high, and the accumulation of

Table 1 Total organic carbon (TOC), grain size (silt þ clay), and concentrations of the isomers and total LABs in sediments collected in the Santos Estuary. Depth (cm)

Agea

TOC (%)

silt þ clay (%)

total LABsb (ng g1)

S-C10-LAB

S-C11-LAB (ng g1)

S-C12-LAB

S-C13-LAB

S-C14-LAB

I/E ratioc

Degradation estimate (%)d

CQ1 0e2 2e4 6e8 10e12 16e18 18e20 20e22 22e24 24e26 26e28 28e30 30e32 32e34 34e36 36e38 38e40 60e62 80e82

2002 2001 1998 1995 1991 1990 1988 1987 1986 1984 1983 1982 1980 1979 1978 1976