Chemosphere 84 (2011) 1556–1562

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Degradation of 4-nonylphenol, 4-t-octylphenol, bisphenol A and triclosan followingbiosolids addition to soil under laboratory conditions

K.A. Langdon a,b, M.St.J. Warne b,⇑, R.J. Smernik a, A. Shareef b, R.S. Kookana b

a School of Agriculture, Food and Wine, University of Adelaide, South Australia, 5005 Adelaide, Australiab Water for a Healthy Country Research Flagship, Commonwealth Scientific and Industrial Research Organisation (CSIRO), PMB 2, Glen Osmond, South Australia,5064 Adelaide, Australia

a r t i c l e i n f o

Article history:Received 24 February 2011Received in revised form 24 May 2011Accepted 25 May 2011Available online 23 June 2011

Keywords:BiosolidsSoil degradation4-nonylphenol4-t-octylphenolBisphenol ATriclosan

0045-6535/$ - see front matter Crown Copyright � 2doi:10.1016/j.chemosphere.2011.05.053

⇑ Corresponding author. Present address: Water QuHealth Division, Department of Environment and Reland, 4102 Brisbane, Australia. Tel.: +61 7 31705571;

E-mail address: Michael.Warne@derm.qld.gov.au (

a b s t r a c t

Land application of biosolids is common practice in many countries, however, there are some potentialrisks associated with the presence of contaminants within the biosolids. This laboratory study examinedthe degradation of four commonly found organic compounds, 4-nonylphenol, 4-t-octylphenol, bisphenolA, and triclosan, in soil following the addition of two biosolids over 32 weeks. The pattern of degradationwas assessed to determine if it followed a standard first-order decay model or if a biphasic model with adegrading and a recalcitrant fraction better described the data. The time taken for the initial concentra-tions to decrease by 50% (DT50), based on a first-order model, was 12–25 d for 4-nonylphenol, 10–14 dfor 4-t-octylphenol, 18–102 d for bisphenol A, and 73–301 d for triclosan. For 4-nonylphenol, bisphenol Aand triclosan, the biphasic model fitted the degradation data better than the first-order model, indicatingthe presence of a degrading fraction and a non-degrading recalcitrant fraction. The recalcitrant fractionfor these three compounds at the completion of the 32 week experiment was 17–21%, 24–42%, and30–51% of the initial concentrations, respectively. For 4-t-octylphenol, the first-order model was suffi-cient in explaining the degradation data, indicating that no recalcitrant fraction was present. This studyshowed that biphasic degradation occurred for some organic compounds in biosolids amended soil andthat the use of standard first-order degradation models may underestimate the persistence of someorganic compounds following land application of biosolids.

Crown Copyright � 2011 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Biosolids may contain a broad range of organic contaminants(e.g. Kinney et al., 2006; USEPA, 2009; Langdon et al., 2010; Leivaet al., 2010) that can enter the environment when this product isapplied to agricultural land as a replacement or supplement forinorganic fertilisers. Four specific organic compounds that have re-ceived increasing interest recently due to their potential adverseenvironmental effects, as a result of their toxicity and/or their abil-ity to mimic natural hormones, are the surfactant metabolites, 4-nonylphenol and 4-t-octylphenol, the plasticiser bisphenol A andthe antimicrobial agent triclosan. These four compounds have beendetected in biosolids at a range of concentrations, up to 440, 2.4,4.6, and 130 mg kg�1 (Kinney et al., 2006; USEPA, 2009), respec-tively. When assessing the potential risk that these compoundsmay pose to the environment following the application of biosolidsto land, the time required for the compounds to degrade and their

011 Published by Elsevier Ltd. All r

ality and Aquatic Ecosystemsource Management Queens-fax: +61 7 31705799.M.St.J. Warne).

potential to accumulate are important factors that need to beconsidered.

Soil degradation experiments conducted on the four target com-pounds (4-nonylphenol, 4-t-octylphenol, bisphenol A, and triclo-san) have reported half-lives or DT50 values (time taken for theinitial concentration of the compound to decrease by 50%) rangingfrom 1 to 17 d (Topp and Starratt, 2000; Roberts et al., 2006),approximately 5 d (Ying and Kookana, 2005), 1 to 7 d (Ying andKookana, 2005; Xu et al., 2009), and 13 to 58 d (Ying et al., 2007;Wu et al., 2009a; Xu et al., 2009), respectively. In addition to theabove studies, the degradation of 4-nonylphenol in soil followingthe addition of biosolids has been examined in more detail in sev-eral studies. For example, Marcomini et al. (1988) conducted a deg-radation experiment on several compounds, including 4-nonylphenol, following sewage sludge application to soil and re-ported multiple phases of degradation for this compound. This con-sisted of an ‘‘initial’’ fast degradation phase, followed by a slower‘‘transition’’ phase and then a final ‘‘persistent’’ phase. The averagenon-degrading residual concentration of 4-nonylphenol at the com-pletion of the experiment was 0.5 mg kg�1 (Marcomini et al., 1988).In a more recent study, Brown et al. (2009), reported half-life valuesfor 4-nonylphenol in a biosolids amended soil of 16–23 d, however,

ights reserved.

K.A. Langdon et al. / Chemosphere 84 (2011) 1556–1562 1557

they found that 15–30% of the initial 4-nonylphenol remained inthe soil at the completion of the 45 d study.

It has been highlighted that the calculation of a single half-lifeor DT50 value for a compound in soil may be an oversimplificationof a complex system, and that the degradation is often more accu-rately described by models with multiple degradation phases (e.g.Hill and Schaaljem, 1985; Marcomini et al., 1988; Ma et al., 2004;Krogh et al., 2009; Sarmah and Close, 2009), for example, biphasicdegradation. Some research has shown that although a compoundmight show a small DT50 value in a soil (calculated from a stan-dard first-order decay model), therefore predicting a fast degrada-tion rate, accumulation over time can still be observed (Ciglaschet al., 2006). This indicates that the single DT50 value may not beadequate to explain the degradation behaviour of some com-pounds. Biphasic degradation of organic compounds is often de-scribed using a two-compartment model, with one compartmentdegrading rapidly and the other compartment degrading slowly(e.g. Hill and Schaaljem, 1985; Ma et al., 2004; Krogh et al., 2009;Sarmah and Close, 2009). In several studies however, the slowlydegrading fraction is characterised by a rate constant of zero, indi-cating this fraction to be non-degrading or recalcitrant (e.g.Sjöström et al., 2008; Wu et al., 2009b).

The aim of this study was to examine the degradation of 4-nonylphenol, 4-t-octylphenol, triclosan, and bisphenol A, followingbiosolids addition to a soil, under controlled laboratory conditionsover a period of 32 weeks (i.e. 224 d). By conducting the study un-der controlled laboratory conditions, with constant temperatureand moisture, any external influences caused by variations in cli-matic conditions were removed. The rate and pattern of degrada-tion of the compounds was assessed to determine if it wasconsistent with a standard first-order degradation model orwhether a biphasic model including a non-degrading or recalci-trant fraction was more appropriate.

2. Materials and methods

2.1. Soil and biosolids

A bulk soil sample was collected from a field site at Mount Com-pass in South Australia (SA) (35� 21044.95 S and 138� 32044.95 E),which is located approximately 70 km south of Adelaide, for usein this study. The site had no history of previous biosolids or sew-age sludge applications. This soil had a pH of 4.4, which was deter-mined from a soil solution ratio of 1:5 in 0.01 M CaCl2, an organiccarbon content of 2.5%, and consisted of 96% sand, 2.5% silt, and1.5% clay. The bulk sample was dried at 40 �C prior to beinghomogenised by grinding with a mortar and pestle and sieved to2 mm. Three subsamples were removed from the dried homoge-nised soil for chemical analysis using the method outlined belowto ensure that there were no background concentrations of the tar-get compounds (i.e. 4-nonylphenol, 4-t-octylphenol, triclosan, andbisphenol A) prior to the commencement of experimental work.

Two South Australian biosolids were collected and used in thisstudy. Both biosolids had been treated by anaerobic digestion, butthereafter one of the biosolids had been centrifuge dried (CDB) andthe other had been solar dried in a lagoon system (LDB). The CDBwas collected immediately following centrifugation, whereas theLDB was collected from a stockpile that had completed treatmentless than 1 month prior to collection. At the time of collectionthe biosolids were mixed by hand. The moisture contents of thebiosolids were 63% for the CDB and 52% for the LDB and for theexperimental work undertaken in this study, the biosolids wereused as collected (i.e. wet). Triplicate sub-samples were removedfrom each of the biosolids samples and freeze dried for analysisof the target compounds using the method outlined in Langdonet al. (2011).

2.2. Experimental design and set up

Individual 50 g samples were weighed from the dried bulk soilinto glass jars and hydrated to 50% of their maximum water hold-ing capacity (MWHC) with Milli Q (MQ) water (the method used todetermine the MWHC is outlined in Jenkinson and Powlson(1976)). All samples were then placed in closed containers in thedark and pre-incubated at 22 �C for 14 d to rejuvenate and stabilisesoil microbial communities. After the pre-incubation either theCDB or LDB sample was added to the hydrated soil, in a random-ised manner, at a rate equivalent to 50 dry t ha�1(where 1 ha isequal to 10 000 m2) (assuming a soil bulk density of 1.3 g cm�3

and an incorporation depth of 10 cm) and mixed throughout thesample. Five replicate sample jars from each biosolids treatmentwere then immediately freeze dried and homogenised by grindingand sieving to 2 mm before being stored in the dark until analysedas the initial sample (t0). All the remaining sample jars wereweighed, then placed on wet paper towel in containers with lidsand kept in the dark at a constant temperature of 22 �C. The sam-ples were opened to the air on a daily basis and the moisture con-tent in the soil was maintained throughout the experiment byweight at 50% MWHC. At eight additional sampling intervals (3,7, 14, 28, 56, 112, 168, and 224 d post biosolids addition), triplicatesample jars were removed from each of the biosolids treatmentsand freeze dried, ground and sieved for immediate analysis ofthe target compounds.

2.3. Sample extraction and gas chromatography–mass spectrometryanalysis

The method used for sample extraction and analysis in thisstudy was based on that outlined in Langdon et al. (2011), withthe only variation being that the current study used a 10 g samplefor extraction and analysis. In brief, each freeze dried sample wasextracted three times with a 1:1 mixture of methanol and acetone(15 mL) in an ultrasonic bath. For each sample the extracts werecombined then diluted with MQ water and loaded onto OasisHLB� solid phase extraction (SPE) cartridges. Elution of the sam-ples was conducted using 3 � 2.5 mL methanol, followed by3 � 2.5 mL acetone and 3 � 2.5 mL ethyl acetate and reconstitutedin 4 mL of methanol. Each sample was then derivatized in 400 lL ofpyridine and 100 lL of the silylation agent N,O-bis-(trimethyl-silyl)-trifluorocetamide (BSTFA) + 1% trimethyl-chlorosilane(TMCS) (based on the method of Shareef et al. (2006)), and anthra-cene-d10 was added to each sample as an instrument internal stan-dard (IS). Along with each batch of samples, a method blank wasrun (i.e. a tube containing no biosolids) to detect any backgroundcontamination from any of the solvents or sample preparationsteps. Samples were analysed using an Agilent 6890 Series GC sys-tem that was interfaced with an Agilent 5973 Network Mass Spec-trometer (MS). The specific details of the GC–MS parameters, thetypical retention times of each of the compounds and target andqualifier ions are reported in Langdon et al. (2011). The concentra-tions of each of the compounds were determined from relative re-sponse factors based on the IS and then adjusted for extractionrecoveries based on labelled surrogates (i.e. triclosan-13C12, bisphe-nol A-d16, and 4-n-nonylphenol-d8) which were spiked into thesamples 1 d prior to extraction. The recoveries of the labelled sur-rogates were used to determine the reproducibility of the methodwithin and between analytical runs. The limit of detection (LOD)and limit of quantification (LOQ) for each of the compounds weredetermined as 3- and 10-times the signal to noise ratio and were,30 and 100 lg kg�1 respectively for 4-nonylphenol, 0.6 and2.0 lg kg�1 respectively for 4-t-octylphenol, 0.3 and 1.0 lg kg�1

respectively for bisphenol A, and 0.8 and 2.7 lg kg�1 respectivelyfor triclosan.

1558 K.A. Langdon et al. / Chemosphere 84 (2011) 1556–1562

2.4. Statistical analysis and interpretation

Prior to all statistical analyses, the concentration data at eachsampling interval were converted to a ratio of the initial concentra-tion (Ct/C0). This normalised the data to an initial mean value of 1and removed any variation at t0 between the biosolids treatmentsand the compounds.

The statistical analyses conducted on the degradation data in-cluded a univariate analysis of variance (ANOVA) to determine ifthe compounds significantly decreased over the 224 d of the exper-iment, using SPSS� Version 17. Nonlinear regressions were also con-ducted to determine the degradation patterns of each compound.There were two nonlinear regression models fitted to the degrada-tion data of each compound based on first-order kinetics, using Sig-maPlot�. The first model was a standard first-order exponentialdecay model (i.e. with two fitting parameters) and the second modelwas a biphasic model (i.e. with three fitting parameters) that ac-counted for a degrading fraction and a recalcitrant fraction of thecompounds. The rate constant from the first-order model was usedto determine the DT50. The rate constant from the biphasic modelproduced a DT50biphasic, which indicated the degradation rate ofthe degrading fraction, and a y-intercept (x0), which indicated themagnitude of the recalcitrant fraction. The residual sums of squareswere then used to statistically compare the two models and producea probability value (p-value) to determine which model providedthe best fit to the data. If the p-value from this comparison was lessthan 0.05 then the biphasic model was considered to provide a sig-nificantly better fit to the data. A detailed outline of the nonlinearregression models is provided in Supplementary material.

3. Results

3.1. Data quality assurance and extraction recoveries

The method blanks run with each batch of samples were belowdetection for all of the compounds except for 4-nonylphenol. Theconcentrations of 4-nonylphenol in the method blanks varied fromapproximately 60–130 lg L�1 in the final extract solution. The con-centration of 4-nonylphenol in the method blank was subtractedfrom each of the samples within the same run prior to the concen-trations being converted to lg kg�1.

The relative standard deviations (RSDs) for the labelled surro-gates within each run of samples were in the majority of cases lessthan 25%, however, in few cases did vary up to 30%. This indicatedthat within each run the variation between the samples was rela-tively small. Comparison of the average recoveries of the labelledsurrogates between each run of samples was used to provide anindication of the variation in the method from one run of samplesto the next. The overall average recovery for 4-n-nonylphenol-d8

was 66% in the CDB treatment and 65% in the LDB treatment, forbisphenol A-d16 it was 104% and 96%, respectively, and for triclo-san-13C12 it was 84% and 91%, respectively. The RSDs for theseaverage recoveries were below 20% for 4-n-nonylphenol-d8 andtriclosan-13C12 in each of the treatments, whereas for bisphenolA-d16 they were below 25%. This indicated that the method used

Table 1The average and range of concentrations of 4-nonylphenol, 4-t-octylphenoldried biosolids (CDB) and lagoon dried biosolids (LDB) treatments.

Biosolids treatment Initial compound concentration ( lg k

4-Nonylphenol 4-t-O

CDB 11800 (7780–16600) 73 (LDB 1690 (607–2480) 129

a The actual upper limit of this range was 462 lg kg�1, however this va

for extraction and analysis of the samples in this study was repro-ducible both within and between runs of samples.

3.2. Preliminary field soil and biosolids analysis

Analysis of the field soil prior to the commencement of theexperiment showed that there were no background concentrationsof any of the target compounds, 4-nonylphenol, 4-t-octylphenol,bisphenol A and triclosan (i.e. all compound concentrations < LOD).The analysis of the two biosolids samples prior to their addition tothe soil showed detectable levels of the four target compounds. Inthe CDB sample the average concentrations of 4-nonylphenol, 4-t-octylphenol, bisphenol A and triclosan were 280, 2.3, 0.19 and3.1 mg kg�1, respectively. There was minimal variation betweenthe replicate sub-samples of the CDB, indicated by RSD values ofless than 10% for each compound. In comparison, analysis of theLDB sample showed concentrations of 43, 2.4, 0.17 and5.9 mg kg�1, respectively. The LDB sample showed more variationbetween the replicates when compared to the CDB, with RSD val-ues of less than 20% for 4-nonylphenol, 4-t-octylphenol and triclo-san and approximately 30% for bisphenol A. This indicated thatthere was higher heterogeneity in the LDB sample compared tothe CDB sample.

3.3. Degradation of 4-nonylphenol from biosolids amended soil

In the initial t0 soil samples, the average concentration of 4-nonylphenol across the five replicates in the CDB treatment was11 800 lg kg�1, whereas in the LDB treatment it was 1690 lg kg�1

(Table 1). This large difference between the treatments was ex-pected due to the large variation in 4-nonylphenol concentrationin the original biosolids samples. There was significant degradationof 4-nonylphenol following the addition of both biosolids treat-ments to the soil over the 224 d of the study (p < 0.0005) (Fig. 1).For both biosolids treatments, from 28 d post biosolids additionto the completion of the experiment (i.e. 224 d), there was no sig-nificant change (p > 0.05) in concentration of 4-nonylphenol.

The fit of the first-order model for the 4-nonylphenol degrada-tion data in both biosolids treatments was significant (both p-val-ues <0.001) and had R2 values of 0.62 and 0.68 for the CDB and LDBtreatments respectively (Fig. 1 and Table 2). The DT50 values for 4-nonylphenol obtained from this model were 12 and 25 d for theCDB and LDB treatments, respectively. The statistical comparisonof the two models (i.e. first-order and biphasic) to the 4-nonylphe-nol degradation data showed that the biphasic model explainedthe data significantly better than the first-order model (both p-val-ues 60.04, Table 2). The effect of adding the third parameter in thebiphasic model was more marked for the CDB treatment(p < 0.001) than for the LDB treatment (p = 0.04). The DT50biphasic

values for 4-nonylphenol were 5.8 d in the CDB treatment and14 d in the LDB treatment (Table 2). The biphasic model fitted tothe normalised degradation data produced x0 values for the CDBand LDB treatments of 0.21 and 0.17 respectively, indicating 21%of the initial concentration of 4-nonylphenol in the CDB treatmentand 17% of the initial concentration of 4-nonylphenol in the LDB

, bisphenol A and triclosan in the initial (t0) sample for the centrifuge

g�1)

ctylphenol Bisphenol A Triclosan

40–105) 5.9 (4.1–8.1) 184 (146–236)a

(53–193) 9.8 (5.0–15) 361 (238–503)

lue was removed as an outlier.

(a) CDB

Time since biosolids addition (days)0 50 100 150 200 250

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

(b) LDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Fig. 1. Degradation of 4-nonylphenol following the addition of (a) centrifuge driedbiosolids (CDB) and (b) lagoon dried biosolids (LDB) to a soil. All concentration datais normalised as a ratio of the concentration at each sampling interval to the initialconcentration (Ct/C0). The nonlinear regression fits for the first-order and biphasicmodels are represented by the dashed line and the solid line, respectively.

(a) CDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

(b) LDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Fig. 2. Degradation of 4-t-octylphenol following the addition of (a) centrifuge driedbiosolids (CDB) and (b) lagoon dried biosolids (LDB) to a soil. All concentration datais normalised as a ratio of the concentration at each sampling interval to the initialconcentration (Ct/C0). The nonlinear regression fits for the first-order and biphasicmodels are represented by the dashed line and the solid line, respectively.

K.A. Langdon et al. / Chemosphere 84 (2011) 1556–1562 1559

treatment was persistent through to the completion of the exper-iment (i.e. after 224 d). These recalcitrant fractions correspondedto 4-nonylphenol concentrations of 2500 lg kg�1 in the CDB treat-ment and 290 lg kg�1 in the LDB treatment at the completion ofthis study.

3.4. Degradation of 4-t-octylphenol from biosolids amended soil

The initial t0 soil samples showed concentrations of 4-t-octyl-phenol in the CDB and LDB treatments of 73 and 129 lg kg�1,

Table 2Summary of the degradation information from the first-order and biphasic models for the codried biosolids (CDB) and lagoon dried biosolids (LDB) treatments. The degradation halfrespectively) are shown in days and the y-intercept (x0) values correspond to the Ct/C0 val

Model Measure 4-Nonylphenol 4-t-Octylphen

CDB LDB CDB

First-order R2 0.62 0.68 0.81DT50 12 25 14p-valuea <0.001 <0.001 <0.001

Biphasic R2 0.78 0.73 0.83DT50biphasic 5.8 14 9.9x0 0.21 0.17 0.10p-valueb <0.001 0.04 0.07

Best fit Biphasic Biphasic First order

a Significance of the first-order model.b Significance of the biphasic model compared to the first-order model.

respectively (Table 1). There was significant degradation of thecompound 4-t-octylphenol over the 224 d of this study(p < 0.0005) (Fig. 2). The samples analysed from 28 d post biosolidsaddition to the completion of the experiment (i.e. 224 d), in bothbiosolids treatments, showed no significant changes in the concen-tration of 4-t-octylphenol (p > 0.05).

The fit of the first-order model to the 4-t-octylphenol norma-lised degradation data was significant for both the biosolids treat-ments (both p-values <0.001), and also produced high R2 values(0.81 for the CDB treatment and 0.79 for the LDB treatment)

mpounds 4-nonylphenol, 4-t-octylphenol, bisphenol A and triclosan for the centrifugelives determined using the first-order and biphasic models (DT50 and DT50biphasic,

ues. The significance values were calculated using Eq. (5) in Supplementary material.

ol Bisphenol A Triclosan

LDB CDB LDB CDB LDB

0.79 0.29 0.55 0.17 0.5710 102 18 301 73<0.001 0.003 <0.001 0.03 <0.001

0.80 0.53 0.68 0.58 0.768.7 8.7 7.7 1.2 6.30.06 0.42 0.24 0.51 0.300.34 0.001 0.003 <0.001 <0.001

First order Biphasic Biphasic Biphasic Biphasic

1560 K.A. Langdon et al. / Chemosphere 84 (2011) 1556–1562

(Fig. 2 and Table 2). The DT50 values obtained for 4-t-octylphenolfrom this model were 14 d for the CDB treated soil and 10 d for theLDB treated soil. The fit of the biphasic model to the 4-t-octylphe-nol degradation data produced marginally higher R2 values thanthe first-order model, however, it did not significantly improvethe fit of the degradation data for 4-t-octylphenol (p-values = 0.07and 0.34 for the CDB and LDB treated soils respectively, Table 2).

3.5. Degradation of bisphenol A from biosolids amended soil

At the commencement of the experiment, bisphenol A was de-tected in the soil treated with the CDB at a concentration of5.9 lg kg�1 and the LDB at a concentration of 9.8 lg kg�1 (Table 1).The concentrations of bisphenol A in each of the biosolids treat-ments were found to significantly decrease over the 224 d study(p < 0.0005) (Fig. 3). From 28 d post biosolids addition through tothe completion of the experiment, there was no significant changesin bisphenol A concentration (p > 0.05).

The fitting of the first-order model to the bisphenol A degrada-tion data was significant (both p-values 60.003), and produced R2

values of 0.29 for the CDB treated and 0.55 for the LDB treated soils(Fig. 3 and Table 2). The DT50 values that were obtained from thefirst-order model differed considerably between the two biosolidstreatments, being 102 d for the CDB treated soils and 18 d for theLDB treated soils. The additional parameter in the biphasic modelsignificantly improved the fit to the bisphenol A degradation datafor both the CDB and LDB treated soils (both p-values60.003).

(a) CDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.5

1.0

1.5

2.0

(b) LDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.5

1.0

1.5

2.0

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Fig. 3. Degradation of bisphenol A following the addition of (a) centrifuge driedbiosolids (CDB) and (b) lagoon dried biosolids (LDB) to a soil. All concentration datais normalised as a ratio of the concentration at each sampling interval to the initialconcentration (Ct/C0). The nonlinear regression fits for the first-order and biphasicmodels are represented by the dashed line and the solid line, respectively.

The biphasic model accounted for 53% of the variation in the deg-radation data from the CDB treated soils and 68% of the variationfrom the LDB treated soils (Table 2). The DT50biphasic values calcu-lated from this model were 8.7 d in the CDB treated soils and 7.7 din the LDB treated soils (Table 2). The proportion of the initialbisphenol A concentration that was predicted by the biphasic mod-el to be recalcitrant at the completion of the experiment was 42%in the CDB treated soils and 24% in the LDB treated soils (Table 2).These recalcitrant fractions corresponded to virtually the sameconcentration in the two biosolids treatments at the end of theexperiment, with values of 2.5 and 2.4 lg kg�1, respectively.

3.6. Degradation of triclosan from biosolids amended soil

The initial t0 soil samples showed concentrations of triclosan inthe CDB and LDB treatments of 184 and 361 lg kg�1, respectively(Table 1). The concentrations of triclosan significantly decreasedthroughout the duration of the experiment (p < 0.0005) (Fig. 4).There was a rapid decrease in the concentration of triclosan atthe commencement of the experiment and this resulted in a signif-icant decrease observed 3 d post biosolids addition for both bioso-lids treatments. The concentrations of triclosan did not changesignificantly, however, from 14 d post biosolids addition until thecompletion of the experiment (i.e. 224 d) in both of the biosolidstreatments.

The fit of the first-order model to the triclosan degradationdata was significant for both the CDB and LDB treated soils (both

(a) CDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

(b) LDB

Time since biosolids addition (days)0 50 100 150 200 250

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Nor

mal

ised

con

cent

ratio

n (C

t/C0)

Fig. 4. Degradation of triclosan following the addition of (a) centrifuge driedbiosolids (CDB) and (b) lagoon dried biosolids (LDB) to a soil. All concentration datais normalised as a ratio of the concentration at each sampling interval to the initialconcentration (Ct/C0). The nonlinear regression fits for the first-order and biphasicmodels are represented by the dashed line and the solid line, respectively.

K.A. Langdon et al. / Chemosphere 84 (2011) 1556–1562 1561

p-values60.03), however, this model only accounted for 17% ofthe variation in the data for the CDB treatment, whereas, for theLDB treatment it accounted for 57% of the variation (Fig. 4 and Ta-ble 2). The DT50 values calculated from the first-order model var-ied considerably between the two biosolids treatments and were301 d and 73 d for the CDB and LDB treatments, respectively (Ta-ble 2). The fit of the biphasic model to the degradation data fortriclosan from both biosolids treatments showed higher R2 valuesof 0.58 and 0.76 for the CDB and LDB treatments, respectively (Ta-ble 2), when compared to the first-order model. When the fits ofthe two models were compared statistically, the biphasic modelsignificantly improved the explanation of variation in the datain both treatments (both p-values <0.001, Table 2). The DT50bipha-

sic values obtained for triclosan were 1.2 d in the CDB treatmentand 6.3 d in the LDB treatment. The x0 values obtained from thebiphasic model for the triclosan degradation data in the CDBand LDB treatments were 0.51 and 0.30, respectively, indicatingthat 51% and 30% of the initial triclosan remained in the soil.These recalcitrant fractions corresponded to 94 and 108 lg kg�1

of triclosan in the CDB and LDB treatments, respectively.

4. Discussion

When the DT50 values that were obtained in this study arecompared to those that have been reported from the literature,they are generally similar or only slightly higher when the majorityof the variation in the data is explained by the model. For example,Brown et al. (2009) reported half life values for 4-nonylphenolfrom a biosolids amended soil of 16–23 d, which is in the samerange as those reported in this study of 12–25 d. For 4-t-octylphe-nol, in a study, where the compound was spiked into a soil, theaverage half life was reported to be 5 d (Ying and Kookana,2005), which is approximately 2- to 3-times smaller than those re-ported in this study of 10–14 d. These small differences may be dueto variations in experimental conditions and also from the additionof the compound through spiking rather than indigenous to thebiosolids. For the two compounds bisphenol A and triclosan, inthe cases where the fit to the first-order model was reasonablygood (i.e. P55%), the DT50 values were only marginally higherthan those reported in the literature. In this study the DT50 valuefor bisphenol A in the LDB treatment was 18 d, whereas othershave reported values ranging from 1 to 7 d (Ying and Kookana,2005; Xu et al., 2009). For triclosan also in the LDB treatment,the DT50 value in this study was 73 d, which is only slightly largerthan the value reported by Wu et al. (2009a) (i.e. 58 d), however itis approximately 4-times larger than that reported by Ying et al.(2007). For bisphenol A and triclosan in the CDB treatment, wherea low proportion of the variation in the data was explained by thefirst-order model (i.e. 629%), the DT50 values are considerably lar-ger than those calculated for the LDB treatment and those in otherstudies. The DT50 value for bisphenol A in the CDB treatment wasapproximately 15-times larger than the highest value reported inother studies (Ying and Kookana, 2005) and for triclosan in thesame biosolids treatment, the value was approximately 5-timeslonger than the highest reported elsewhere (Wu et al., 2009a).Due to the poor fit of the first-order model to the degradation datafor these two compounds in the CDB treatment, it is likely that theuse of a single DT50 value is not sufficient in explaining the degra-dation of these compounds and provides an unreliable predictionof their persistence.

When an additional parameter was used in the biphasic model,the fit to the data was significantly improved for 4-nonylphenol,bisphenol A and triclosan, indicating that there was a fraction ofthese compounds that was resistant to degradation. This was notthe case for 4-t-octylphenol, where the additional fitting parame-

ter in the biphasic model provided no statistically significantimprovement in explaining the variation in the data. This is likelyto be due, in part, to the fit of the first-order model being quitegood for this compound (R2 = 0.79 and 0.81) and the fact that theconcentrations of this compound decreased more than the othersover the duration of the study (Table 2 and Fig. 2). The results ob-served in this study for the compounds 4-nonylphenol, bisphenol Aand triclosan are consistent with other research. For exampleMarcomini et al. (1988) reported a persistent fraction of approxi-mately 10% for 4-nonylphenol following sewage sludge additionto soil and Sjöström et al. (2008), reported recalcitrant fractionsof 26–35% for nonylphenol following the addition of sewage sludgeto soil. There are several possible suggestions for the presence of arecalcitrant fraction of these compounds in a biosolids amendedsoil. It has been suggested that this may be due to the distributionof the compound throughout heterogeneous aggregates of bioso-lids (Hesselsoe et al., 2001; Sjöström et al., 2008). The formationof biosolids aggregates tends to produce aerobic zones in the outerareas and anaerobic zones in the centre of aggregates, which canresult in persistent or recalcitrant concentrations of the com-pounds contained within the biosolids (Hesselsoe et al., 2001). Asthe compounds assessed in this study degrade predominately un-der aerobic conditions (e.g. McAvoy et al., 2002; Ying and Kookana,2005; Ying et al., 2007; Press-Kristensen et al., 2008), the presenceof anaerobic zones in the biosolids aggregates is likely to have re-sulted in the degradation slowing considerably or halting. A furtherhypothesis is that the recalcitrant fraction is due to sorption that isnon-reversible which means that there is a sorbed fraction that isnot available to microorganisms and hence non-degradable (Wuet al., 2009b; Katayama et al., 2010). This is likely to be an impor-tant factor for the four compounds examined in this study due totheir hydrophobic nature (octanol–water partition co-efficient val-ues (log Kow) range from 3.3 to 6.8) (e.g. Langdon et al., 2010). Inaddition to these above suggestions, it should be noted that gener-ally a biosolids matrix is complex and may involve many compo-nents. Various organic compounds may sorb more strongly to thematrix or to different components of the matrix, resulting in thepresence of recalcitrant fractions. This may explain the differingproportions of each of the compounds in this study that were recal-citrant using the same soil amended with different biosolids andthe lack of a recalcitrant fraction (statistically) for 4-t-octylphenol.It should also be noted however, that if the presence of the recalci-trant fraction of these compounds is due to a decrease in their bio-availability, they are therefore less likely to exert toxic effects tosoil organisms. Further studies examining the bioavailability ofthese compounds at different stages of degradation following bios-olids addition would therefore be warranted.

The results obtained in this study also showed some variation be-tween the two biosolids treatments within the same soil. For exam-ple, the non-degrading fraction of the compounds was always lowerin the LDB treated soil compared to the CDB treated soil. These dif-ferences further indicate the potential effect of differing biosolidsmatrices on the degradation patterns of organic compounds follow-ing land application of biosolids, therefore highlighting the impor-tance of conducting degradation studies on compounds that areindigenous to the biosolids. The variations observed between thetreatments in this controlled laboratory study are likely to be furtheremphasised in a ‘real-world’ field scenario therefore follow-up fieldwork assessing the degradation of these compounds is required.

Overall, the results from this study raise concerns relating to thepotential accumulation of organic compounds in biosolidsamended soils, particularly if repeat applications are made. Thisis due to the presence of recalcitrant fractions of compounds thatshow resistance to degradation. This is particularly the case forthe compounds bisphenol A and triclosan which had the highestrecalcitrant fractions (42% and 51% respectively). In addition, this

1562 K.A. Langdon et al. / Chemosphere 84 (2011) 1556–1562

study also showed that the use of a single value, for example DT50,is in some cases insufficient in explaining the degradation of organ-ic compounds following biosolids addition to soil. Although inmost cases a large proportion of the data was explained by thefirst-order model, this was significantly improved by the biphasicmodel. The use of the most appropriate model is crucial whendetermining the risks associated with these compounds followingthe addition of biosolids to land, as use of an incorrect model couldlead to significant underestimation of the persistence of organiccompounds in biosolids amended soils.

5. Conclusion

The four compounds 4-nonylphenol, 4-t-octylphenol, bisphenolA and triclosan were found to degrade over time when added to asoil via two biosolids. The time taken for the initial concentrationto decrease by 50% (DT50), based on a standard first-order decaymodel, was 12–25 d for 4-nonylphenol, 10–14 d for 4-t-octylphe-nol, 18–102 d for bisphenol A and 73–301 d for triclosan. The useof the first-order model produced DT50 values that were consistentwith other research only when a considerable portion of the vari-ation was explained by the model. When the first-order modeldid not explain a considerable portion of the variation, the calcu-lated DT50 values were markedly longer than those reported inthe literature. For 4-nonylphenol, bisphenol A and triclosan, a bi-phasic model, which accounts for a degrading fraction and a recal-citrant fraction, fitted the degradation data significantly betterthan the first-order model. The recalcitrant fractions for thesethree compounds remaining at the completion of the experiment,as predicted by the biphasic model, were 297–2480 lg kg�1 for4-nonylphenol, 2.4–2.5 lg kg�1 for bisphenol A and 94–108 lg kg�1 for triclosan, which corresponded to 17–21%, 24–42%, and 30–51% of the initial concentrations, respectively. Thisindicates that there is the potential for accumulation of these com-pounds with repeat applications of biosolids. In contrast, for 4-t-octylphenol, the first-order model was sufficient for predicting itsdegradation thus indicating that there was no statistical evidencefor a recalcitrant fraction of this compound. It appears that differ-ent biosolids matrices may influence the degradation of these com-pounds. The biphasic degradation pattern of some organiccompounds found in biosolids is possibly related to anaerobic con-ditions within biosolids aggregates and differential non-reversiblesorption of compounds to the biosolids matrix. This study showsthat an understanding of the degradation pattern is crucial whenassessing the persistence of compounds in soils following the addi-tion of biosolids.

Acknowledgements

The principal author (KL) wishes to acknowledge the financialsupport of Water Quality Research Australia (WQRA) (formerlythe Co-operative Research Centre for Water Quality and Treatment– CRCWQT), Western Australia Water Corporation, VictorianDepartment of Human Services and the Commonwealth Scientificand Industrial Research Organisation (CSIRO).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.chemosphere.2011.05.053.

References

Brown, S., Devin-Clarke, D., Doubrava, M., O’Connor, G., 2009. Fate of 4-nonylphenolin a biosolids amended soil. Chemosphere 75, 549–554.

Ciglasch, H., Busche, J., Amelung, W., Totrakool, S., Kaupenjohann, M., 2006.Insecticide dissipation after repeated field application to a northern Thailandultisol. J. Agric. Food Chem. 54, 8551–8559.

Hesselsoe, M., Jensen, D., Skals, K., Olesen, T., Moldrup, P., Roslev, P., Mortensen,G.K., Henriksen, K., 2001. Degradation of 4-nonylphenol in homogeneous andnonhomogeneous mixtures of soil and sewage sludge. Environ. Sci. Technol. 35,3695–3700.

Hill, B.D., Schaaljem, G.B., 1985. A 2-compartment model for the dissipation ofdeltamethrin on soil. J. Agric. Food Chem. 33, 1001–1006.

Jenkinson, D.S., Powlson, D.S., 1976. The effects of biocidal treatments onmetabolism in soil – V. A method for measuring soil biomass. Soil Biol.Biochem. 8, 209–213.

Katayama, A., Bhula, R., Burns, G.R., Carazo, E., Felsot, A., Hamilton, D., Harris, C.,Kim, Y.-H., Kleter, G., Koerdel, W., Linders, J., Peijnenburg, J.G.M.W., Sabljic, A.,Stephenson, R.G., Racke, D.K., Rubin, B., Tanaka, K., Unsworth, J., Wauchope, R.D.,2010. Bioavailability of xenobiotics in the soil environment. Rev. Environ.Contam. T 203, 1–86.

Kinney, C.A., Furlong, E.T., Zaugg, S.D., Burkhardt, M.R., Werner, S.L., Cahill, J.D.,Jorgensen, G.R., 2006. Survey of organic wastewater contaminants in biosolidsdestined for land application. Environ. Sci. Technol. 40, 7207–7215.

Krogh, K.A., Jensen, G.G., Schneider, M.K., Fenner, K., Halling-Sørensen, B., 2009.Analysis of the dissipation kinetics of ivermectin at different temperatures andin four different soils. Chemosphere 75, 1097–1104.

Langdon, K.A., Warne, M.St.J., Kookana, R.S., 2010. Aquatic hazard assessment forpharmaceuticals, personal care products and endocrine disruptingcompounds from biosolids-amended land. Integr. Environ. Assess. Manage.6, 663–676.

Langdon, K.A., Warne, M.St.J., Smernik, R.J., Shareef, A., Kookana, R.S., 2011. Selectedpersonal care products and endocrine disruptors in biosolids: An Australia-widesurvey. Sci. Total Environ. 409, 1075–1081.

Leiva, C., Ahumada, I., Sepúlveda, B., Richter, P., 2010. Polychlorinated biphenylbehavior in soils amended with biosolids. Chemosphere 79, 273–277.

Ma, Q.L., Rahman, A., Holland, P.T., James, T.K., McNaughton, D.E., 2004. Fielddissipation of acetochlor in two New Zealand soils at two application rates. J.Environ. Qual. 33, 930–938.

Marcomini, A., Capel, P.D., Giger, W., Hani, H., 1988. Residues of detergent-derivedorganic pollutants and polychlorinated-biphenyls in sludge amended soil.Naturwissenschaften 75, 460–462.

McAvoy, D., Schatowitz, B., Jacob, M., Hauk, A., Eckhoff, W., 2002. Measurement oftriclosan in wastewater treatment systems. Environ. Toxicol. Chem. 21, 1323–1329.

Press-Kristensen, K., Lindblom, E., Schmidt, J.E., Henze, M., 2008. Examining thebiodegradation of endocrine disrupting bisphenol A and nonylphenol inWWTPs. Water Sci. Technol. 57, 1253–1256.

Roberts, P., Roberts, J.P., Jones, D.L., 2006. Behaviour of the endocrine disruptingchemical nonylphenol in soil: assessing the risk associated with spreadingcontaminated waste to land. Soil Biol. Biochem. 38, 1812–1822.

Sarmah, A.K., Close, M.E., 2009. Modelling the dissipation kinetics of six commonlyused pesticides in two contrasting soils of New Zealand. J. Environ. Sci. Heal. B44, 507–517.

Shareef, A., Angove, M.J., Wells, J.D., 2006. Optimization of silylation using N-methyl-N-(trimethylsilyl)-trifluoroacetamide, N, O-bis-(trimethylsilyl)-trifluoroacetamide and N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide for the determination of the estrogens estrone and17-alpha-ethinylestradiol by gas chromatography-mass spectrometry. J.Chromatogr. A 1108, 121–128.

Sjöström, A.E., Collins, C.D., Smith, S.R., Shaw, G., 2008. Degradation and plantuptake of nonylphenol (NP) and nonylphenol-12-ethoxylate (NP12EO) in fourcontrasting agricultural soils. Environ. Pollut. 156, 1284–1289.

Topp, E., Starratt, A., 2000. Rapid mineralization of the endocrine-disruptingchemical 4-nonylphenol in soil. Environ. Toxicol. Chem. 19, 313–318.

USEPA (United States Environmental Protection Agency), 2009. Targeted nationalsewage sludge survey and analysis technical report, EPA-822-R-08-016. USEnvironmental Protection Agency, Washington, USA.

Wu, C.X., Spongberg, A.L., Witter, J.D., 2009a. Adsorption and degradation oftriclosan and triclocarban in solis and biosolids-amended soils. J. Agric. FoodChem. 57, 4900–4905.

Wu, C.X., Spongberg, A.L., Witter, J.D., 2009b. Sorption and biodegradation ofselected antibiotics in biosolids. J. Environ. Sci. Health A 44, 454–461.

Xu, J., Wu, L.S., Chang, A.C., 2009. Degradation and adsorption of selectedpharmaceuticals and personal care products (PPCPs) in agricultural soils.Chemosphere 77, 1299–1305.

Ying, G.G., Kookana, R.S., 2005. Sorption and degradation of estrogen-like-endocrinedisrupting chemicals in soil. Environ. Toxicol. Chem. 24, 2640–2645.

Ying, G.G., Yu, X.Y., Kookana, R.S., 2007. Biological degradation of triclocarban andtriclosan in a soil under aerobic and anaerobic conditions and comparison withenvironmental fate modeling. Environ. Pollut. 150, 300–305.