Removal Pahs By Fenton

Chemosphere 59 (2005) 1427–1437 www.elsevier.com/locate/chemosphere

Removal of sorbed polycyclic aromatic hydrocarbons from soil, sludge and sediment samples using the FentonÕs reagent process Vanina Flotron, Corine Delteil, Yann Padellec, Vale´rie Camel

*

Institut National Agronomique Paris-Grignon, Laboratoire de Chimie Analytique—UMR Environnement et Grandes Cultures, 16 Rue Claude Bernard, 75231 Paris cedex 05, France Received 19 May 2004; received in revised form 10 December 2004; accepted 24 December 2004

Abstract The use of the FentonÕs reagent process has been investigated for the remediation of environmental matrices contaminated by polycyclic aromatic hydrocarbons (PAHs). Laboratory experiments were first conducted in aqueous solutions, to study the kinetics of oxidation and adsorption of PAHs. Benzo[a]pyrene was more rapidly degraded than adsorbed, while only partial oxidation of fluoranthene occurred. In the case of benzo[b]fluoranthene, its adsorption prevented its oxidation. Besides competition effects between PAHs were found, with slower oxidation of mixtures as compared to single PAH solutions. Apparition of some by-products was observed, and a di-hydroxylated derivative of benzo[a]pyrene could be identified under our conditions. Consequently, application to solid environmental matrices (soil, sludge and sediment samples) was performed using large amounts of reagents. The efficiency of the Fenton treatment was dependent on the matrix characteristics (such as its organic carbon content) and the PAH availability (correlated to the date and level of contamination). However, no pH adjustment was required, as well as no iron addition due to the presence of iron oxides in the solid matrices, suggesting the potential application of Fenton-like treatment for the remediation of PAH-contaminated environmental solids.
2005 Elsevier Ltd. All rights reserved. Keywords: AOP; Fenton; Oxidation; PAH; Remediation; Solid

1. Introduction Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous pollutants that occur in natural media such as soil, sediment, water and air as a result of both natural and anthropogenic processes (Henner et al., 1997). Their presence in such environmental matrices is of great con-

* Corresponding author. Tel.: +33 1 44 08 17 25; fax: +33 1 44 08 16 53. E-mail address: [email protected] (V. Camel).

cern due to the possible bioavailability of these toxic compounds. In some cases, remediation should be considered with a view of reducing the toxicity. Because of the very low biodegradability of these organic pollutants, a chemical oxidative treatment should be preferred. The FentonÕs reagent (Fe(II)–H2O2) seems very attractive for such treatment for several reasons: (i) its moderate cost, (ii) its simplicity of operation, (iii) its advanced oxidative potential due to the formation of hydroxyl radicals OH (Watts, 1992; Venkatadri and Peters, 1993; Schulte et al., 1995; Neyens and Baeyens, 2003). Hydroxyl radicals are the most oxidative species

0045-6535/$ - see front matter
2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2004.12.065

1428

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

that can be formed in aqueous solutions, and they have been shown to degrade many organic compounds including PAHs, using different advanced oxidation processes (AOPs). Nowadays, the FentonÕs reagent is often used to treat wastewaters from various industries (Kuo, 1992; Schulte et al., 1995; Wanpeng et al., 1996; Lin and Lo, 1997; Kang and Hwang, 2000). It can also be applied to the remediation of contaminated solid matrices, such as soils, sediments or sludges (Mustranta and Viikari, 1993; Kawahara et al., 1995; Martens and Frankenberger, 1995; Watts and Dilly, 1996; EPA, 1998; Watts et al., 1999; Nam et al., 2001; Lee et al., 2002; Neyens et al., 2002; Watts et al., 2002; Neyens and Baeyens, 2003). Furthermore, Fenton oxidation is a process that can be used in combination with bioremediation techniques, either as a pre-treatment for the oxidation of PAHs to more biodegradable compounds, or as a posttreatment for residual PAHs (Kelley et al., 1991; Martens and Frankenberger, 1995; Nam et al., 2001). This process may be effective for the treatment of PAHs in waters. The FentonÕs reaction is based on the hydrogen peroxide decomposition in the presence of ferrous iron to produce a hydroxyl radical, which is the main oxidizing species (reaction (1)). Fe2þ þ H2 O2 ! Fe3þ þ OH þ  OH k 1 ¼ 53–76 mol1 l s1

ð1Þ

It is generally conducted in very acidic medium (pH 2–3) to prevent iron salts precipitation (Kuo, 1992). The hydroxyl radicals formed degrade organic compounds either by hydrogen abstraction (reaction (2)) or by hydroxyl addition (reaction (3)). RH þ  OH !  R þ H O k ¼ 107 –1010 mol1 l s1 2

2

ð2Þ R þ  OH !  ROH

ð3Þ

Usually the disappearance of the organic substrate follows a pseudo-first order kinetic law, the OH concentration being considered as a constant (due to their high reactivity). Thus, the pseudo-first order rate constant for PAH oxidation can be estimated experimentally (i.e. kexp) by following the studied PAH disappearance over time. An optimal oxidation of the organic substrate can be obtained if the consumption of the hydroxyl radicals by other reactions, especially with the reagents themselves (reactions (4) and (5)) is limited. Tang and Huang (1996) have reported a 10:1 H2O2/Fe(II) molar ratio to allow for the best degradation of 2,4-dichlorophenol. Fe2þ þ  OH ! Fe3þ þ OH

k 4 ¼ 3–4  108 mol1 l s1 ð4Þ

H2 O2 þ  OH ! H2 O þ HO2 k 5 ¼ 1:2–4:5  107 mol1 l s1

ð5Þ

As the hydroxyl radicals are generated in aqueous solution, a solid/solution suspension should be produced in order to apply the FentonÕs reagent for the remediation of contaminated solid matrices. In such a system, previous studies reported that pollutants should be first desorbed from the solid phase in order to be able to react with hydroxyl radicals generated in the solution (Sedlak and Andren, 1994; Watts et al., 1999). More recently, Watts et al. (2002) showed that the presence of iron minerals in the solid matrix can allow the generation of the reactive species close to benzo[a]pyrene adsorbed or present in non-aqueous phase liquid (NAPL), enabling its degradation. In addition, the presence of iron minerals in the sludges could enable the Fenton oxidation to proceed without any iron addition in presence of large amounts of hydrogen peroxide (Venkatadri and Peters, 1993; Lin and Gurol, 1998; Watts et al., 1999). So, this study was undergone to investigate the feasibility of using the FentonÕs reagent to degrade sorbed PAHs in three different solid matrices: a sewage sludge, an agricultural soil, and a marine sediment. Besides, as Fenton oxidation has not been studied in sufficient detail to provide substantial information on the efficiency of this treatment to degrade PAHs, the behaviour of PAHs in aqueous solutions containing OH has first been studied. Elevated PAH concentrations were chosen in order to study adsorption and oxidation kinetics. Moreover, efforts were done to determine kinetics of appearance of by-products and to identify them.

2. Materials and methods 2.1. Reagents and chemicals Individual standard solutions (10 mg l1 in acetonitrile) of the following PAHs were obtained from CIL Cluzeau (Paris, France): fluoranthene (Fla, purity: 98.3%), benzo[b]fluoranthene (BbF, purity: 99.0%), benzo[a]pyrene (BaP, purity: 98.7%). Their structure and properties are reported in Table 1. Hydrogen peroxide (30% solution as weight) and iron (II) sulfate heptahydrate (purity: 99.5%) were supplied by VWR (France), as well as HPLC-grade methanol (purity: 99.8%), acetonitrile (purity: 99.8%), tetrahydrofuran (purity: 99.7%), n-hexane (purity: 99.8%) and acetone (purity: 99.8%). Deionized water was produced with a Milli-Q system from Millipore (Saint-Quentin-en-Yvelines, France). To avoid cross contamination, before use all vessels were rinsed several times with acetone, and then rinsed with milli-Q water.

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437 Table 1 Structure and properties of the investigated PAHs Water solubility at 25 C (mg l1)a

Log Kowa

Fluoranthene

0.26

5.2

Benzo[b]fluoranthene

0.0015

6.6

Compounds

Structure

1429

X31–147 (AFNOR, 1996); organic carbon content was estimated based on the modified Anne method (which consists in the oxidation of organic matter by potassium dichromate in sulfuric medium). All samples were frozen upon reception in the laboratory for their storage. Before their use, samples (around 100 g) were taken, and carefully dried in an oven (40 C, 72 h) to avoid PAH losses (Berset et al., 1999), before being homogenized with a mortar and the fraction greater than 2 mm removed through sieving. 2.3. Fenton oxidation

Benzo[a]pyrene

a

0.0038

6.0

Manoli and Samara (1999).

2.2. Environmental solid samples Solid sewage sludge samples were obtained from the municipal wastewater station of Ache`res near Paris, and sampled at the end of the sludge treatment (according to a statistical sampling plan, the composite sample obtained being homogenized). The soil is an agricultural soil, which has received wastewaters from the municipal wastewater plant of Ache`res during several decades; several soil samples were taken from the 0–30 cm layer, and mixed together to produce the final soil sample. The sediment comes from the ‘‘Ria de Aveiro’’, a lagoon located in the centre of Portugal that received industrial effluents; several sediment samples were taken from the 15–16 cm layer, and mixed together to produce the final sediment sample. The characteristics of these three matrices are summarized in Table 2. The pH was taken in a solid/water (1:2.5) suspension; total iron was determined according to the French norm NF Table 2 Characteristics of the sludge, sediment and soil samples Characteristics

Sludge

Soil

Sediment

Fluoranthene (mg kg1) Benzo[b]fluoranthene (mg kg1) Benzo[a]pyrene (mg kg1) Organic carbon (%) Organic matter (%) Total iron (mg g1) pH Type

1.20

0.107

0.017

0.43

0.064

0.123

0.31

0.051

0.003

20.8 36.1 94 8.8 nd

1.7 3.0 nd 7.1 Fine sand 70% Silt 17%

nd 10.4 30 4.4 Sand

nd: not determined.

All experiments were done in triplicate. 2.3.1. Oxidation of solubilized and sorbed PAHs in aqueous solutions Experiments were conducted at room temperature in 50 ml Teflon erlenmeyers, covered with aluminium foil to avoid any photolytic degradation. Aqueous solutions of PAHs (25 ml, 80 lg l1 each) were prepared upon dilution with milli-Q water of the stock PAH solutions. Initial pH was adjusted to 3 with H2SO4. The ferrous iron and hydrogen peroxide solutions, freshly prepared every day in stock concentrations, were successively added. Time zero was sampled just before the addition of hydrogen peroxide to the solution. The FentonÕs reaction was then started upon addition of H2O2. At various time intervals (after 45, 120 and 180 min), an aliquote (1 ml) was withdrawn, mixed with 0.25 ml methanol (to quench radicals) before being neutralized with NaOH (final pH around 6) and further injected in HPLC. At the end of the oxidation experiment, the remaining solution was discarded, and the recipient walls as well as magnetic stir bar were rinsed with acetonitrile (2 h contact time with continual mixing). This solution was then injected in the HPLC system, to analyse possible retained PAHs. 2.3.2. Oxidation of sorbed PAHs in solid matrix suspension Solid matrix suspensions were prepared in Teflon recipients by adding 10 ml of milli-Q water to 2 g of oven-dried matrix, followed by the Fe(II) solution (final concentration of 102 mol l1). After stirring the suspension a few minutes, hydrogen peroxide (final concentration of 4.9 mol l1) was slowly added with great care to avoid excessive temperature increase. The suspension was then magnetically stirred during 24 h to allow Fenton oxidation. Experiments have been carried out on non-spiked as well as spiked samples. In the latter case, 80 ll of each PAH standard (at 10 mg l1 in acetonitrile) were carefully added on the matrix surface without mixing to avoid PAH losses on the recipient walls. In addition, blank experiments were carried out under the same conditions, except that no hydrogen peroxide was

1430

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

added, to study possible desorption or degradation of the PAHs. 2.4. Extraction After Fenton oxidation in solid/aqueous suspensions, the solid matrix was separated from the aqueous phase by centrifugation in Teflon tubes (15 min, 1397 g, 15 C). Then the aqueous supernatant was pre-concentrated using solid-phase extraction on styrene-divinylbenzene copolymer, after addition of methanol (4 ml) to the solution to avoid PAH adsorption on recipient walls. Cartridges (Bond Elut PPL, 0.2 g, Varian) were conditioned successively with methanol/tetrahydrofuran (10/90 v/v—2 · 2.5 ml), methanol (2 · 2.5 ml) and milliQ water (2 · 2.5 ml), and PAHs were eluted with methanol/tetrahydrofuran (10/90 v/v—3 + 2 ml). The overall extraction efficiency of the solid-phase extraction has been evaluated: 92.6% (2.4), 67.1% (5.1) and 63.4% (9.5) for fluoranthene, benzo[b]fluoranthene and benzo[a]pyrene respectively (RSDs are given in italic and in brackets). Then the obtained fractions were concentrated under a gentle stream of nitrogen to around 1 ml, filtered (nylon filters, 13 mm, 0.45 lm, CIL Cluzeau) before being injected in the HPLC system. The solid residues were dried in an oven (40 C, 2 · 24 h) before being extracted using pressurized fluid extraction (ASE 100 system, Dionex). The conditions were as follows: 1 g dried matrix mixed with hydromatrix and alumina, hexane/acetone (50/50 v/v), 120 C, 5 min static time, solvent purge volume 6.8 ml, nitrogen purge time 100 s, and two static cycles. The extraction efficiency of this step has also been evaluated: 106.4% (3.3), 103.0% (3.6) and 75.5% (4.7) for fluoranthene, benzo[b]fluoranthene and benzo[a]pyrene respectively (RSDs are given in italic and in brackets). The extracts were further concentrated to dryness, and the solid residue redissolved in 1 ml acetonitrile for HPLC analysis. 2.5. Liquid chromatography analysis The HPLC system consisted of a Varian 9010 low-pressure gradient pump, a Rheodyne Model 7125 injection valve equipped with a 20 ll loop, a Thermo Separation Science fluorimetric detector (FL3000), a PE Nelson 900 Series interface, and a computer (Intel Pentium, speed 90 MHz, hard disk 813 Mo, RAM 16 Mo, operating system Windows 3.11). Data analysis was performed using the TurboChrom TC4 Navigator. A Supelcosil LC-PAH column (250 · 4.6 mm i.d., C18-silica, 5 lm particle size, Supelco) was used, along with a pre-column (Supelguard LC-18, Supelco). Separation was performed using the following gradient: acetonitrile/water (60:40 v/v) for 5 min, followed by a 25-min ramp to attain 100% acetonitrile, this solvent being further maintained for 15 min. The total flow-rate

was 1.5 ml min1. Detection was performed at selected excitation/emission wavelengths, respectively 230/410 nm for fluoranthene, 250/420 nm for benzo[b]fluoranthene and benzo[a]pyrene. Quantitation was performed using external calibrations with standard solutions of PAHs in acetonitrile/water (60/40 v/v) in the range 25– 200 lg l1. 2.6. Formation of oxidation by-products For identification of by-products formed, due to the low sensitivity of the detectors used, pre-concentration of the aqueous solutions was performed by solid-phase extraction onto styrene-divinylbenzene (as retention of the initial PAH was also required). Cartridges (Bond Elut PPL, 0.2 g, Varian) were conditioned successively with methanol/tetrahydrofuran (50/50 v/v, 2 · 2.5 ml), methanol (2 · 2.5 ml) and milli-Q water at pH 3 (2 · 2.5 ml), and PAHs were eluted with methanol/tetrahydrofuran (50/50 v/v, 5 · 1 ml). The obtained fractions were analysed using an HPLC system coupled with a diode-array UV–visible detector (HPLC-DAD). It consisted of a Waters 1525 high-pressure gradient pump, a Rheodyne injection valve equipped with a 20 ll loop, a Waters 2996 DAD detector, and a computer (AMD K7, type Athlon XP, speed 2 MHz, hard disk 9.54 Go, RAM 256 Mo, operating system Windows 2000). Data analysis was performed using the Millenium software. A Supelcosil LC-PAH column (150 · 3.0 mm i.d., C18-silica, 5 lm particle size, Supelco) was used, along with a pre-column (Supelguard LC-18, Supelco). Separation was performed using the following gradient: acetonitrile/water (40:60 v/v) for 4 min, followed by a 11-min ramp to attain 100% acetonitrile, this solvent being further maintained for 10 min. The total flow-rate was 0.8 ml min1. Some samples were also analysed using an HPLC system coupled to a UV–visible detector (detection at 254 nm) and to a mass spectrometer (MS) equipped with an atmospheric pressure chemical ionization (APCI) interface. The HPLC system consisted of a Perkin– Elmer 200LC high-pressure gradient pump, a Rheodyne injection valve equipped with a 20 ll loop, an APCI interface, a PE-Sciex API 100 simple quadrupole, a channel electromultiplier and a computer. Data analysis was performed using the MacDAD software. A Supelcosil LC-PAH column (150 · 3.0 mm i.d., C18-silica, 5 lm particle size, Supelco) was used, along with a pre-column (filled with C18-silica). Separation was performed using a similar gradient as with the HPLC-DAD system, except that the mobile phase was acidified (0.2% acetic acid): acetonitrile/water/acetic acid (39.9:59.9:0.2 v/v/v) for 4 min, followed by a 11-min ramp to attain 100% acetonitrile/acetic acid (99.8:0.2 v/v), this solvent being further maintained for 10 min. The total flow-rate was 0.8 ml min1. Other conditions for the APCI interface were: positive mode, source temperature 400 C, source

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

needle current 5 lA, orifice voltage 30 V, ring voltage 300 V, nebulizer gas 8 l min1, curtain gas (N2) 0.8 l min1.

3. Results and discussion Experiments were first conducted in aqueous solutions, due to the lack of data related to Fenton oxidation of PAHs in water. As we were interested in studying oxidation of both solubilized and sorbed PAHs, we used experimental conditions that enabled partial adsorption of the pollutants on the recipient walls (Flotron et al., 2003). Mild Fenton conditions were used, due to the absence of natural organic matter that may consume OH. Then, oxidation of PAHs sorbed onto solid samples was investigated, using stronger Fenton conditions to take into account the presence of large amounts of organic matter and the subsequent radical scavenging. 3.1. Oxidation of sorbed and solubilized PAHs in aqueous solutions 3.1.1. Kinetics of oxidation Experiments were first conducted with PAH mixture in aqueous solutions. The system was studied in presence of an excess of hydrogen peroxide (i.e. [H2O2]0/

1431

[Fe2+]0 = 2 or 10, this latter value being the theoretically optimal ratio according to Tang and Huang (1996)). In all cases, the reagents were in excess compared to the substrates (see Table 3). Results obtained for the different experiments are shown in Fig. 1, and compared to control experiments (i.e. no reagent addition, enabling the estimation of adsorption kineticsP on recipient walls). For experiment no. 1 (i.e. [Fe2+]0/[ PAHs]0 = 10 and [H2O2]0/[Fe2+]0 = 10), 60% of benzo[a]pyrene were degraded, the remaining benzo[a]pyrene being sorbed on the recipient walls. By comparison with the experiment without FentonÕs reagent, for which around 60% of this PAH was solubilized, we assumed that only the solubilized benzo[a]pyrene could be oxidized by the hydroxyl radicals generated in the solution, suggesting that sorption prevents oxidation under our conditions. In the same time, no degradation of either fluoranthene or benzo[b]fluoranthene could be observed, despite the large soluble content of the former compound. Thus, it is evident that benzo[a]pyrene is more rapidly degraded than these two compounds, probably due to its aromatic structure, the presence of a five-atom ring in the PAH structure leading to a lower oxidation. This efficient oxidation of benzo[a]pyrene using the FentonÕs reagent has already been reported (Kawahara et al., 1995; Nam et al., 2001), and is of great interest due to its recalcitrance to microbial degradation.

Table 3 Experimental pseudo-first order rate kinetic constants determined under different FentonÕs reagent concentrations (and related correlation coefficients R2) P Experiment [Fe2+]0 [H2O2]0 [Fe2+]0/[ PAHs]0 [H2O2]0/[Fe2+]0 PAH k 1exp R21 R22 k 2exp (mol l1) (mol l1) (mol/mol) (mol/mol) (min1) (min1) 0a

0

0

1

1.0 · 105

2

0

0

Fla—mixture BbF—mixture BaP—mixture

0.001 0.002 0.004

0.971 0.961 0.964

– – –

– – –

1.0 · 104

10

10

Fla—mixture BaP—mixture

– 0.017

– 0.991

– –

– –

8.2 · 105

1.7 · 104

80

2

Fla—mixture BaP—mixture

0.020 0.138

0.997 0.976

– 0.047

– 0.984

3

1.6 · 104

3.2 · 104

155

2

Fla—mixture BaP—mixture

0.039 0.422

0.996 1.000

0.001 0.052

0.948 0.992

4

1.6 · 104

1.6 · 103

155

10

Fla—mixture BaP—mixture

0.150 0.648

1.000 1.000

0.002 0.070

0.970 0.994

5

3.1 · 104

3.1 · 103

300

10

Fla—mixture BaP—mixture

0.051 0.088

0.984 0.951

0.003 –

0.947 –

6

1.6 · 104

3.2 · 104 (2-step addition)

155

2

Fla—mixture BaP—mixture

0.033 0.170

0.973 0.944

0.001 0.006

0.861 0.921

7

3.1 · 104

3.1 · 103

300

10

Fla—single BaP—single

0.303 0.539

1.000 1.000

0.006 0.058

0.999 0.975

k 1exp and k 2exp are the kinetic constants of PAH oxidation with the Fe2+/H2O2 or the Fe3+/H2O2 system respectively. a In this experiment, adsorption kinetic constants were estimated, while for the others only oxidation kinetics are presented.

1432

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

Percentage of BaP content (%)

Solubilised

100% 80% 60% 40% 20% 0%

0

1

2

Solubilised

Percentage of Fla content (%)

Degraded

120%

(a)

3

Adsorbed

4

5

6

Degraded

140% 120% 100% 80% 60% 40% 20% 0%

(b)

0

1

2

3

Adsorbed

4

5

6

Degraded

160% 140% 120% 100%

(c)

80% 60% 40% 20% 0%

prone to adsorption). Again, this suggests that adsorption prevents oxidation under these conditions. One explanation lies in the moderate OH concentrations, leading to the consumption of these species in the solution before coming near the recipient walls where the sorbed PAHs are present. Adding hydrogen peroxide in two steps (at time 0 and 30 min) instead of one did not enhance the removals, suggesting that the production of hydroxyl radicals was limited by the ferrous iron concentration in the solution. This was confirmed by the oxidation kinetics observed for fluoranthene and benzo[a]pyrene. A pseudo-first order reaction was noted with a change in the experimental kinetic rate constant after a certain time, leading to a slower degradation as indicated in Table 3. This could be explained by an initial rapid production of OH upon oxidation of Fe(II) into Fe(III) by hydrogen peroxide (leading to rapid substrate degradation and kinetic constant k 1exp ), followed by partial regeneration of Fe(II) when an excess of hydrogen peroxide is present, according to reaction (6). Fe3þ þ H O ! Fe2þ þ Hþ þ HO 2

Solubilised

Percentage of BbF content (%)

Adsorbed

140%

0

1

2

3

4

5

6

Fig. 1. Effect of FentonÕs reagent concentrations on the behaviour of PAHs in acidic aqueous solutions: (a) benzo[a]pyrene, (b) fluoranthene and (c) benzo[b]fluoranthene. Vertical bars represent the 95% confidence interval of the mean on triplicates. The X-axis represent the series of experiment according to Table 3. Experimental conditions: [PAH]0 = 80 lg l1; pH 3; 3 h oxidation.

As the reagent concentrations were increased, almost complete degradation of benzo[a]pyrene could be achieved, indicating that this compound was more rapidly degraded than adsorbed. Significant degradation of fluoranthene also occurred, but as its adsorption remained constant, we assumed that only solubilized fluoranthene could be oxidized under our conditions. On the opposite, benzo[b]fluoranthene remained undegraded, being mainly adsorbed on the recipient walls (this compound being the most hydophobic, it is the most

2

k 6 ¼ 0:01 mol1 l s1

2

ð6Þ

However, this reaction being considerably slower than the reaction (1), the rate constant will be decreased (k 2exp  k 1exp ). Performing the same type of experiments on single PAHs for the two previously degraded compounds revealed competition effects (see experiment no. 7). Hence, the kinetic constants are obviously higher when organic compounds are sole in solution, as more OH are available for the considered PAH oxidation. On the opposite, when both organic compounds are present, competition effects between the two solutes lead to a slower oxidation of both compounds, especially for the more recalcitrant compound, fluoranthene. This suggests that higher reagent concentrations should be added in presence of different organic substrates to achieve significant PAH oxidation, especially with solid matrices that contain natural organic matter. Experiments performed on single benzo[b]fluoranthene confirmed the absence of oxidation under our conditions. 3.1.2. By-products formation Efforts were made to follow the formation of some byproducts, and possibly identify some structures. Results are presented in Fig. 2 for fluoranthene and benzo[a]pyrene. A few compounds appeared on the chromatograms, with much lower retention times than the parent compounds, indicating more polar structures. Some compounds further disappeared, being oxidized, while others remained in the solution over time. However, these compounds could be detected only at trace levels, so that their identification could not be achieved. Yet,

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

1433

Fig. 2. By-product formation during the FentonÕs oxidation in aqueous solution: (a) appearance of peaks on the chromatogram during fluoranthene oxidation, (b) appearance of peaks on the chromatogram during benzo[a]pyrene oxidation and (c) APCI-HPLC-MS mass spectrum of by-product P3. Experimental conditions: [BaP]0 = 80 lg l1; [Fe2+]0 = 3.1 · 104 mol l1; [H2O2]0 = 3.1 · 103 mol l1; pH 3. By-products of fluoranthene are noted P 0 ; by-products of benzo[a]pyrene are noted P.

1434

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

some structural information could be obtained in the case of the more concentrated by-product for benzo[a]pyrene (i.e. P3) as shown in Fig. 2. The molecular mass of this compound should be 284, assuming the m/z 285 being MH+ (which is produced in the APCI interface). The other major fragments can be explained as follows: m/z 286 is the fraction of MH+ corresponding to the natural isotopic 13C content of the molecule; m/z 267 is formed upon the loss of water giving (MH  H2O)+; m/z 257 is the result of the loss of CO leading to (MH  CO)+. All these observed fragments are consistent with a di-hydroxylated derivative of benzo[a]pyrene, but they do not give information about the position of the two hydroxylated groups on the molecule. Due to possible high toxicity of such derivatives, efforts need to be pursued to ascertain the structures of by-products formed. In particular, we attributed the low levels of by-products in our solutions to the fact that aqueous samples were pre-concentrated on a PS-DVB copolymer during solid-phase extraction, which should not efficiently retain polar compounds. So, in order to achieve higher sensitivity, samples should be pre-treated on a more polar sorbent such as porous graphitized carbon to achieve better extraction of by-products. 3.2. Oxidation of sorbed PAHs in solid matrices under strong conditions without pH adjustment Experiments for solid matrices were performed using aqueous solid matrices suspensions, and upon large con-

centrations of FentonÕs reagent, to enable sufficient OH to be available for PAH oxidation, as their consumption by natural organic matter has to be taken into account. No pH adjustment was performed, as previous studies reported efficient oxidation of sorbed compounds under such conditions (Watts et al., 2002). Preliminary experiments conducted with PAHs solubilized in aqueous solutions (at the 40 lg l1 level) under such conditions showed that almost complete oxidation of the three solutes could be achieved, with removals of 99.4%, 88.8% and 99.3% for fluoranthene, benzo[b]fluoranthene and benzo[a]pyrene respectively. 3.2.1. Removal of native PAHs Blank experiments (conducted in the absence of hydrogen peroxide) revealed no desorption of the native PAHs. Therefore, oxidation of the pollutants should occur in the sorbed form. Hence, it is interesting to note that oxidation of the sorbed native PAHs could be observed (see Table 4). Besides, a strong dependence on the nature of the matrix was evident. Large removal of the three tested PAHs could be obtained for the sludge sample, while only benzo[b]fluoranthene was significantly degraded for the sediment sample, and none of these compounds could be efficiently oxidized in the case of the soil sample under these conditions. The influence of the matrix on PAH oxidation by the FentonÕs reagent has already been observed (Nam et al., 2001). Despite no clear correlation between the observed percentages of removal and the natural organic matter

Table 4 Total PAH content (expressed as percentage of the total PAH content) removed from the matrix and recovered in the aqueous phase after FentonÕs reagent application (24 h reaction) (RSDs are given in italic and in brackets) Sample

PAH

Spiked content (%)

Total removed (%) Native (no spike)

Native + spiked

Sludge

Fluoranthene

0.0 25.0 0.0 48.2 0.0 56.3

46.2 (0.0) – 36.2 (0.2) – 48.1 (0.2) –

– 38.0 (5.1) – 42.0 (5.5) – 67.0 (2.0)

0.2 (13.4) nd 0.8 (20.2) nd 0.4 (28.0) nd

0.0 96.0 0.0 76.5 0.0 99.2

9.7 (141.4) – 85.0 (4.4) – 16.9 (123.9) –

– 97.8 (0.2) – 97.6 (0.4) – 99.1 (0.5)

nd nd 5.4 (25.4) nd nd nd

0.0 78.9 0.0 86.1 0.0 88.7

0.0 – 8.0 (40.7) – 0.0 –

– 85.7 (0.2) – 87.4 (0.5) – 88.6 (0.8)

1.5 nd 2.7 0.1 1.5 0.1

Benzo[b]fluoranthene Benzo[a]pyrene Sediment

Fluoranthene Benzo[b]fluoranthene Benzo[a]pyrene

Soil

Fluoranthene Benzo[b]fluoranthene Benzo[a]pyrene

nd: not detected.

Total recovered in solution (%)

(45.2) (47.3) (141.4) (36.8) (20.7)

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

3.2.2. Removal of spiked and native PAHs Spiked PAHs (0.4 mg kg1 each) were added to the investigated environmental matrices (sludge, soil and sediment), as such compounds are expected to be less sorbed than the native pollutants, being more available for OH oxidation. Results presented in Table 4 show the total PAH content removed from the matrix and recovered in solution after Fenton oxidation using strong conditions without pH adjustment. In that case, large removals of the three compounds were obtained, ranging from 38% to nearly 100%. So, we may assume

that spiked PAHs were quantitatively oxidized in each case, while native PAH removals differed depending on the compound and the matrix as already observed. Nam et al. (2001) also reported more efficient degradation of spiked PAHs than native PAHs in soils. 3.2.3. Effect of the solid matrix iron oxides In order to investigate possible effect of iron oxides present in the solid matrix, experiments were conducted upon the sole addition of hydrogen peroxide. The results presented in Fig. 3 give evidence that the addition of ferrous iron to the solid suspension had no significant effect as compared to the sole addition of hydrogen peroxide. This observation suggests that iron oxides present in the solid matrix may be involved in the oxidation process, as already reported (Watts et al., 1999). Indeed, in presence of a large excess of hydrogen peroxide, Lin and Gurol (1998) assumed that this reagent reacts with iron oxides

Control 140

Fe2+/H2O2

H2O2

120 Recovery (%)

100 80 60 40 20

(a)

0

Fla

BbF

Control

Fe2+/H2O2

120

BaP H2O2

100 Recovery (%)

contents of the matrix, some assumptions can be made to explain these results. Thus, the limited degradation of the PAHs in the sludge sample could be due to two effects. Firstly, insufficient OH concentrations could be available for PAH oxidation, due to their consumption with the sludge natural organic matter (benzo[a]pyrene being slightly more degraded than the other two compounds due to its easiest oxidation as observed previously), as natural organic matter is well-known to inhibit PAH oxidation by the FentonÕs reagent (Lindsey and Tarr, 2000). Secondly, it seems that the fraction of undegraded benzo[a]pyrene was related to the unavailable fraction of this pollutant, as extraction of the sludge sample with ethanol enabled only 50% benzo[a]pyrene removal (while complete removal of the other two PAHs could be achieved) (data not shown). The higher removal of benzo[b]fluoranthene as well as the limited fluoranthene and benzo[a]pyrene degradation in the sediment were attributed to the low content of these native PAHs, and thus to lower required OH content for oxidation (in case of benzo[b]fluoranthene) or to their non-availability (for the latter two compounds). In case of the soil, the absence of significant oxidation was attributed to the low levels of both benzo[b]fluoranthene and benzo[a]pyrene, as well as on the low organic matter of the soil. Hence, a recent study showed that for soils containing less than 5% organic matter, the pollutants were adsorbed in the micropores, being less available than for soils with higher organic matter content for which pollutants are mainly sorbed onto the organic matter (Bogan and Trbovic, 2003). This could explain the absence of fluoranthene oxidation observed for our soil sample. Consequently, it is obvious that the efficiency of the Fenton treatment is strongly dependent on the solid matrix characteristics and the contaminant availability. In particular, in the case of low levels of contamination, large amounts of hydrogen peroxide are required to enable oxidation of sorbed PAHs due to strong matrixPAH interactions, which may reduce the cost-efficiency of this remediation process. So it seems that the Fenton treatment would be more dedicated in practice to highly contaminated environmental matrices, such as the remediation of industrial sites (EPA, 1998).

1435

80 60 40 20 0

(b)

Fla

BbF

BaP

Fig. 3. Influence of the addition of reagents on the oxidation of sorbed native PAHs from real solid matrices: (a) sediment and (b) sludge. Vertical bars represent the 95% confidence interval of the mean on triplicates. Experimental conditions: [Fe2+]0 = 102 mol l1; [H2O2]0 = 4.9 mol l1; no pH adjustment; 24 h oxidation.

1436

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437

of the solid matrix, enabling the formation of hydroxyl radicals at the surface of the solid, which can further oxidize sorbed pollutants that are in their vinicity. They proposed as the initial mechanism the formation of a surface complex between hydrogen peroxide and iron oxides. This step would be followed by an electron transfer inside the complex, leading to Fe(II) and the radical HO2 . This complex further dissociates, and due to the high reactivity of the radical this dissociation should be non-reversible. Then excess hydrogen peroxide further reacts with Fe(II) with a mechanism similar to that of the Fenton reaction, except that this is an heterogeneous reaction in that case, leading to the generation of OH at the surface of the solid matrix. Due to their high reactivity, such radicals would oxidize only pollutants that are near their site of production.

4. Conclusion The Fenton oxidation of sorbed PAHs from solid environmental samples can be achieved, using large excess of hydrogen peroxide, without any pH adjustment. Besides, the presence of iron oxides in the matrix may avoid the addition of ferrous iron to the system. However, the level and nature of the contamination appear as important factors for the success of the oxidation process, as well as the organic matter content of the matrix. In particular, experiments conducted with spiked samples are insufficient to assess the efficiency of an oxidative remediation process. Also, reagent addition should be optimized as competition between PAH and natural organic matter for reaction with hydroxyl radicals was evident, as well as competition between PAHs. In particular, benzo[a]pyrene was found more reactive towards  OH than PAHs containing a five-atom carbon cycle. So, further studies should be conducted to assess the extent of oxidation, especially the nature and toxicity of possible by-products formed. Acknowledgement The authors gratefully thank Dr. L. Debrauwer (INRA, Toulouse, France) and Dr. J. Einhorn (INRA, Versailles, France) for their precious advices and help in conducting HPLC-MS experiments and proposing identification of by-product P3.

References AFNOR NF X31–47, 1996. Qualite´ des sols. Sols, se´diments. Mise en solution totale par attaque acide. Berset, J.D., Ejem, M., Holzer, R., Lischer, P., 1999. Comparison of different drying, extraction and detection techniques

for the determination of priority aromatic hydrocarbons in background contaminated soil samples. Anal. Chim. Acta 383, 263–275. Bogan, B.W., Trbovic, V., 2003. Effect of sequestration on PAH degradability with FentonÕs reagent: roles of total organic carbon, humin, and soil porosity. J. Hazard. Mater. B 100, 285–300. EPA US, Report EPA 542-R-98-008, 1998. Field applications of in situ remediation technologies: chemical oxidation. Office of Solid Waste and Emergency Response, Technology Innovation Office, Washington, DC, US, pp. 1–31. Flotron, V., Delteil, C., Padellec, Y., Camel, V., 2003. Remediation of matrices contaminated by polycyclic aromatic hydrocarbons: use of FentonÕs reagent. Polycycl. Aromat. Compd. 23, 353–376. Henner, P., Schiavon, M., Morel, J.L., Lichtfouse, E., 1997. Polycyclic aromatic hydrocarbon (PAH) occurrence and remediation methods. Anal. Mag. 25 (9–10), 56–59. Kang, Y.W., Hwang, K.-Y., 2000. Effects of reaction conditions on the oxidation efficiency in the Fenton process. Water Res. 34 (10), 2786–2790. Kawahara, F.K., Davila, B., Al-Abed, S.R., Vesper, S.J., Ireland, J.C., Rock, S., 1995. Polynuclear aromatic hydrocarbon (PAH) release from soil during treatment with FentonÕs reagent. Chemosphere 31 (9), 4131–4142. Kelley, R.L., Gauger, W.K., Srivastava, V.J., 1991. Application of FentonÕs reagent as a pretreatment step in biological degradation of polyaromatic hydrocarbons. Gas Oil Coal Environ. Biotechnol. 3, 105–120. Kuo, W.G., 1992. Decolorizing dye wastewater with FentonÕs reagent. Water Res. 26 (7), 881–886. Lee, B.-D., Nakai, S., Hosomi, M., 2002. Application of Fenton oxidation to remediate polycyclic aromatic hydrocarbons-contaminated soil. J. Chem. Eng. Japan 35, 582–586. Lin, S.H., Lo, C.C., 1997. Fenton process for treatment of desizing wastewater. Water Res. 31 (8), 2050– 2056. Lin, S.-S., Gurol, M.D., 1998. Catalytic decomposition of hydrogen peroxide on iron oxide: kinetics, mechanism, and implications. Environ. Sci. Technol. 32, 1417–1423. Lindsey, M.E., Tarr, M.A., 2000. Inhibited hydroxyl radical degradation of aromatic hydrocarbons in the presence of dissolved fulvic acid. Water Res. 34 (8), 2385–2389. Manoli, E., Samara, C., 1999. Polycyclic aromatic hydrocarbons in natural waters: sources, occurrence and analysis. Trends Anal. Chem. 18, 417–428. Martens, D.A., Frankenberger Jr., W.T., 1995. Enhanced degradation of polycyclic aromatic hydrocarbons in soil treated with an advanced oxidative process—FentonÕs reagent. J. Soil Contamin. 4, 175–190. Mustranta, A., Viikari, L., 1993. Dewatering of activated sludge by an oxidative treatment. Water Sci. Technol. 28, 213–221. Nam, K., Rodriguez, W., Kukor, J.J., 2001. Enhanced degradation of polycyclic aromatic hydrocarbons by biodegradation combined with a modified Fenton reaction. Chemosphere 45, 11–20. Neyens, E., Baeyens, J., 2003. A review of classic FentonÕs peroxidation as an advanced oxidation technique. J. Hazard. Mater. B 98, 33–50.

V. Flotron et al. / Chemosphere 59 (2005) 1427–1437 Neyens, E., Baeyens, J., Weemaes, M., De Heyder, B., 2002. Advanced biosolids treatment using H2O2-oxidation. Environ. Eng. Sci. 19, 27–35. Schulte, P., Bayer, A., Kuhn, F., Luy, Th., Volkmer, M., 1995. H2O2/O3, H2O2/UV and H2O2/Fe2+ processes for the oxidation of hazardous wastes. Ozone Sci. Eng. 17, 119– 134. Sedlak, D.L., Andren, A.W., 1994. The effect of sorption on the oxidation of polychlorinated biphenyls (PCBs) by hydroxyl radical. Water Res. 28 (5), 1207–1215. Tang, W.Z., Huang, C.P., 1996. 2,4-Dichlorophenol oxidation kinetics by FentonÕs reagent. Environ. Technol. 17, 1371– 1378. Venkatadri, R., Peters, R.W., 1993. Chemical oxidation technologies: ultraviolet light/hydrogen peroxide, FentonÕs reagent, and titanium dioxide-assisted photocatalysis. Hazard. Waste Hazard. Mater. 10, 107–131.

1437

Wanpeng, Z., Zhihua, Y., Li, W., 1996. Application of ferrous-hydrogen peroxide for the treatment of H-acid manufacturing process wastewater. Water Res. 30 (12), 2949–2954. Watts, R.J., 1992. Hydrogen peroxide for physicochemically degrading petroleum-contaminated soils. Remediation 2, 413. Watts, R.J., Dilly, S.E., 1996. Evaluation of iron catalysts for the Fenton-like remediation of diesel-contaminated soils. J. Hazard. Mater. 51, 209–224. Watts, R.J., Stanton, P.C., Howsawkeng, J., Teel, A.L., 2002. Mineralization of a sorbed polycyclic aromatic hydrocarbon in two soils using catalyzed hydrogen peroxide. Water Res. 36 (17), 4283–4292. Watts, R.J., Udell, M.D., Kong, S.G., Leung, S.W., 1999. Fenton-like soil remediation catalysed by naturally occurring iron minerals. Environ. Eng. Sci. 16 (1), 93–103.