Comparison of a plant based natural surfactant with SDS for washing of As(V) from Fe rich soil

JOURNAL OFENVIRONMENTALSCIENCES

ISSN 1001-0742

CN 11-2629/X

www.jesc.ac.cn

Available online at www.sciencedirect.com

Journal of Environmental Sciences 2013, 25(11) 2247–2256

Comparison of a plant based natural surfactant with SDS for washing of As(V)from Fe rich soil

Soumyadeep Mukhopadhyay1, Mohd. Ali Hashim1,∗, Jaya Narayan Sahu1,Ismail Yusoff2, Bhaskar Sen Gupta3

1. Department of Chemical Engineering, University of Malaya, Kuala Lumpur 50603, Malaysia2. Department of Geology, University of Malaya, Kuala Lumpur 50603, Malaysia

3. School of Planning, Architecture and Civil Engineering, Queen’s University Belfast, BT9 5AG, UK

Received 01 January 2013; revised 02 April 2013; accepted 08 April 2013

AbstractThis study explores the possible application of a biodegradable plant based surfactant, obtained from Sapindus mukorossi, for washinglow levels of arsenic (As) from an iron (Fe) rich soil. Natural association of As(V) with Fe(III) makes the process difficult. Soapnutsolution was compared to anionic surfactant sodium dodecyl sulfate (SDS) in down-flow and a newly introduced suction mode for soilcolumn washing. It was observed that soapnut attained up to 86% efficiency with respect to SDS in removing As. Full factorial designof experiment revealed a very good fit of data. The suction mode generated up to 83 kPa pressure inside column whilst down-flowmode generated a much higher pressure of 214 kPa, thus making the suction mode more efficient. Micellar solubilisation was found tobe responsible for As desorption from the soil and it followed 1st order kinetics. Desorption rate coefficient of suction mode was foundto be in the range of 0.005 to 0.01, much higher than down-flow mode values. Analysis of the FT-IR data suggested that the soapnutsolution did not interact chemically with As, offering an option for reusing the surfactant. Soapnut can be considered as a soil washingagent for removing As even from soil with high Fe content.

Key words: soil washing; soapnut; Sapindus mukorossi; arsenic; plant based surfactant

DOI:10.1016/S1001-0742(12)60295-2

Introduction

Accumulation of arsenic (As) in soil due to unsafe agri-cultural practices, mining, smelting, coal burning, woodpreservation and illegal waste dumping activities continueto be a serious threat to human health and environment(Tokunaga and Hakuta, 2002). The non-biodegradability ofAs and its variable mobility under different geochemicalprocesses and soil redox conditions ensure its transfor-mation and continued presence in the soil matrix fora long period of time (Cheng et al., 2009). A numberof methods have been reported for the treatment of Ascontaminated soils (Wang and Zhao, 2009). Soil washingby acids, alkaline reagents, phosphates and Bureau ofReference three-step sequential extraction procedure arewell researched (Jang et al., 2005, 2007; Ko et al., 2005).Biosurfactants synthesised by living cells for the removalof toxic metals from soil matrix is also currently beingassessed (Chen et al., 2008; Mulligan and Wang, 2006;

* Corresponding author. E-mail: [email protected]

Polettini et al., 2009). In this research work, a naturalsurfactant obtained from Sapindus mukorossi plant orsoapnut has been used to wash low level of arsenic(V) fromsoil and its performance has been compared with sodiumdodecyl sulphate (SDS), an inorganic anionic surfactant.The soil matrix used here has a high level of Fe richmineral maghemite, which has a good affinity for As(V)(Yamaguchi et al., 2011). The pollutant removal becomesdifficult at lower concentrations (Sundstrom et al., 1989).

The fruit pericarp of Sapindus mukorossi or soapnut is asource of saponin, an effective plant based surfactant (Chenet al., 2008; Song et al., 2008). Soapnut tree is common inIndo-Gangetic plains, Shivaliks and sub-Himalayan tracts.The soapnut fruit pericarp contains triterpenoidal saponin;a natural surfactant that has been used as an environ-ment friendly detergent and medicine for many decades(Suhagia et al., 2011). Previously, saponin was used forremoval of Cd, Zn, Ni and a number of organic pollutantswith success (Chen et al., 2008; Kommalapati et al., 1997;Polettini et al., 2009; Roy et al., 1997; Song et al., 2008).The mechanism for pollutant removal involved increase

2248 Journal of Environmental Sciences 2013, 25(11) 2247–2256 / Soumyadeep Mukhopadhyay et al. Vol. 25

of wettability of the soil by surfactant solution, sorptionof the surfactant molecules onto the soil surface, physicalor chemical attachment with the pollutant, detachment ofthe pollutant particle or molecule from the soil into thesurfactant solution, and subsequent association with sur-factant micelles. However, saponin extracted from soapnuthas never been used for removal of soil arsenic, whichhas entirely different chemical characteristics from theheavy metals. As(V) was used in this study as the pH andredox values in the sampling site as well as in the columnfavoured the presence of As(V) over As(III) (Dobran andZagury, 2006). Speciation was confirmed by a solventextraction process. Moreover As(V) is more difficult toremove than As(III), from Fe(III) bearing minerals of soilcomponents (Yamaguchi et al., 2011). The soil sampleused in this work contains maghemite, which has a highaffinity for As(V) (Chowdhury and Yanful, 2010). Arsenicpolluted sites such as mine tailings and agricultural fieldscontain high level of iron. The intension of using highiron containing soil in this study is to test the efficiencyof soapnut solution in such challenging conditions.

Sodium dodecyl sulphate (SDS), an anionic inorganicsurfactant has been used by various research groups for soilwashing to remove heavy metals and organics (Torres etal., 2012). In this study, SDS has been used to comparethe efficiency of soapnut with respect to this widely usedanionic surfactant.

The objectives of this research were to study; (1) theperformance of the natural surfactant soapnut solutioncompared to commonly used SDS in washing arsenic froma soil column; (2) the desorption kinetics and the mecha-nism of arsenic removal by soapnut solution; and (3) theadvantage of a newly introduced suction mode comparedto traditional down-flow mode of column washing.

1 Materials and methods

1.1 Soil sample, surfactants and analytical methods

A composite soil sample was collected from the first layeraquifer in Hulu Langat area, Selangor, Malaysia. The soilwas dried in an oven overnight at 105°C and then crushedand passed through a 2 mm sieve (Roy et al., 1997). All soilparameters are measured according to standard proceduresas shown in Table 3. XRD analysis was performed bya Panalytical Empyrean diffractometer using HighscorePlus software. As(V) salt (Na2HAsO4·7H2O) was used forspiking the soil matrix depending on the Eh and pH ofthe unspiked soils (Tokunaga and Hakuta, 2002). AlthoughAs(V) salt is soluble in water, it binds strongly with Fe(III)minerals and cannot be removed by water alone. Thesoil was spiked by 50 and 200 mg/L concentrations ofsodium arsenate solution, at room temperature by mixingit for 7 days at a weight: volume ratio of 3:2. In orderto increase field relevance and wash away the looselybound water soluble arsenic, the spiked soil samples were

leached with 2 pore volumes of artificial rainwater ofpH 5.9 consisting of 5 × 10−4 mol/L CaCl2, 5 × 10−4

mol/L Ca(NO3)2, 5 × 10−4 mol/L MgCl2, 10−4 mol/LNa2SO4, and 10−4 mol/L KCl (Oorts et al., 2007). Porevolume for a 300 g soil column was calculated to beapproximately 80 mL which was evaluated by measuringthe weight difference between dry and water saturated soilcolumns (Dwarakanath et al., 1999). After this stage, thesoils were allowed to drain overnight, then air dried at25°C for 24 hr and sieved through a 2-mm mesh. Theywere digested following US EPA method 3050B to mea-sure metal contents by ICP-OES (Perkin-Elmer Optima7000DV). All the samples were analyzed in triplicates andthe results were reproducible within ± 3.5%. Based onsome preliminary experiments, 20 mmol/L of SDS, 0.5and 1% (W/V) of soapnut extractions were selected for thestudy and were compared against a standard blank sample.All of the surfactant concentrations greatly exceeded thecritical micelle concentration (CMC) of the respectivesurfactants.

The surfactant was extracted from the soapnut fruitpericarp by water (Roy et al., 1997). The CMC and surfacetension of surfactants were measured by a ring type surfacetensiometer (Fisher Scientific Manual Model 20 SurfaceTensiometer). The functional groups present in the soapnutextract and the effluent solution were characterized byFT-IR spectroscopy (PerkinElmer Spectrum 100 Series)collected in the range of 400–4000 cm−1 wavenumbers.Zetasizer Nano ZS series (Malvern Instruments Ltd., UK)was used to measure zeta potential of the soil particles inthe presence of different surfactant solutions.

1.2 Speciation of arsenic in spiked soil by solvent ex-traction

Arsenic in the soil spiked with sodium arsenate wasspeciated by the solvent extraction process (Chappell et al.,1995). This extraction was performed in three steps; first,total As was extracted following Method 3050B, secondly,any trivalent As was extracted from an aliquot of this totalAs extract in CHCl3 in a separating funnel and again backextracted into aqueous layer, and thirdly, total inorganicAs was determined by adding 50% (W/V) KI solution withthe As extract and then extracting inorganic As by CHCl3in a separating funnel. The concentration of As for eachsolution was determined by ICP-OES.

1.3 Design of experiments by full factorial design

A number of factors influencing the soil column washingprocess have been investigated; viz. the type of surfactantand its concentration, level of As contamination in the soiland the washing mode. A full factorial design was followedto include all possible combinations of the levels acrossall of these control factors. Design Expert 7.0.0 was usedto plan the experiments and to analyze the results. Thefull factorial design reveals the effect of each factor on

No. 11 Comparison of a plant based natural surfactant with SDS for washing of As(V) from Fe rich soil 2249

the response variable, as well as the interactions betweenfactors on the response variable. In total, 96 experimentswere conducted in duplicate in three identical experimentalsetups. The levels and ranges of the studied process param-eters are given in Table 1, and the experimental design ispresented in Table S1.

Response is recorded in terms of percentage of Asremoval (r, %) from the column in each pore volume (PV),defined by Eq. (1).

r =Aseffluent × PV × 100

Ascolumn × 1000× 100% (1)

where, Aseffluent (mg/L) is the concentration of As in theeffluent, Ascolumn (mg) is the total amount of As in the 300gm of soil inside the column initially.

1.4 Statistical analysis

Analysis of variance (ANOVA) is a general statisticalmethod used for testing the hypothesis that the meansamong two or more groups are equal. It was used foranalysing the data to obtain the interaction among differentcontrol factors. After the data was gathered from theexperiments, a square root transformation was applied tothe data which was then fitted into the 2FI model. Squareroot transformation was applied on the data because with-out any transformation, the coefficient of variance (CV)came up to be very large, which is unacceptable. Someof the important factors in the ANOVA method are sumof squares (SS), R2, adjusted R2, P-value and adequateprecision (AP). The SS of each control factor quantifiesits importance in the process. As the value of the SSincreases, the significance of the corresponding factor inthe undergoing process also increases. A high R2 value,close to 1, is desirable to ensure a satisfactory adjustmentof the mathematical model to the experimental data. Areasonable agreement of R2 value with adjusted R2 value isalso necessary. Model terms were evaluated by the P-value(P < 0.05 are potentially significant) with 95% confidencelevel. AP is a type of signal to noise ratio and comparesthe range of the predicted values at the design points to theaverage prediction error. AP value greater than 4 indicateadequate model discrimination.

1.5 Column washing procedure

The contaminated soil was packed in a 10 cm long plexi-glass column having 5.5 cm internal diameter (Roy et al.,1997). Approximately 300 gm of soil was packed in each

column. Circular plexiglass discs with perforations wereinserted at 3 cm intervals to distribute the liquid flow and toavoid preferential flow. Arsenic extraction was induced bypumping 6 PV of flush solution. The packed column wasflooded with water from the bottom at the rate of 5 mL/minto remove air spaces. Then flushing solution was pumpedinto the saturated soil column from top to provide washin a down-flow mode. In the suction mode, flush solutionwas introduced from the top of the saturated column whilea peristaltic pump was attached at the outlet to suck thewash liquid out of the column. The effluent was collectedfor each PV and As concentrations were measured by ICP-OES. The cumulative As removal in Nth PV (R, %) wasmeasured after each PV (r, %) by Eq. (2).

R = rN−1 + rN (2)

The pressure drop across the soil column was monitoredin both cases. The experimental setup is shown in Fig. 1.

1.6 Desorption kinetics

The cumulative concentration of As remaining in the soilcolumn (mg As/kg soil) was used to plot the kinetics(Sparks et al., 1980). Desorption of As varied with vari-ation in the flow rate of the surfactant. Column washingwas performed in continuous mode by 6 pore volumes ofwash liquid and effluent samples were collected at differenttime. The time intervals were noted down accordingly andhas been shown in Table S2. The time taken for differentmodes vary due to development of pressure inside thecolumn. The apparent desorption rate coefficient (kd

′) andthe order of the processes was determined using Eq. (3)(Sivasubramaniam and Talibudeen, 1972).

ln(Ast/As0) = −k′d(t) (3)

which can also be written as

Ast = As0exp(−k′dt) (4)

where, Ast (mg) is the quantity of As on soil exchange sitesat time t (min) of desorption or amount of As (mg) in soilcolumn at time t (min), As0 (mg) is the quantity of As onexchange sites at zero time of desorption or the amount ofAs in soil column at initial stage before column washingwas initiated. The ln(Ast/As0) vs. t relationship is linear ifthe rate of release of As follows first-order kinetics.

Table 1 Control factors and their levels for the experimental design

Control factor Level1 2 3 4 5 6

Wash solution (A) Water SDS 20 mmol/L Soapnut 0.5% Soapnut 1%Washing mode (B) Down-flow SuctionSoil contamination (C) High LowPore volume (D) PV1 PV2 PV3 PV4 PV5 PV6


10 cm

Conta

min

ated

soilPeristaltic

pump

Surfactantsolution

a

10 cm

Peristaltic pump

Surfactantsolution

b

Peristaltic pump

Effluent sample

Effluent sample

Conta

min

ated

soil

Fig. 1 Schematic diagram of the column washing experiment (a) down-flow mode (b) suction mode.

2 Results and discussion

2.1 Soil and surfactant characterization

The soil particle size distribution was as following: sand(< 50 µm) 92.66%; silt (50–2 µm) 5.2%, and clay (> 2µm) 2%, and the soil was classified as sandy soil accordingto USDA soil classification. XRD analysis of both spikedand unspiked soils revealed that the soil samples con-tained silicon dioxide as quartz (SiO2, XRD displacement0.158), magnesium aluminium silicate hydroxide as mica((Mg, Al)6(Si, Al)4O10(OH)8, XRD displacement 0.119),sylvine, sodian (Cl1K0.9Na0.1, XRD displacement –0.171),maghemite Q (Isometric Fe21.333O32, XRD displacement0.001), feldspar albite (AlNaO8Si3, XRD displacement –0.053). The XRD spectrum of the spiked soil is shownin Fig. 2. Arsenic was not detected in the mineral phaseas expected in such low levels as 52.5 mg/kg. The char-acteristics of the natural soil sample are as following:pH 4.5 (USEPA SW-846 Method 9045D); specific gravity2.64 (ASTM D 854 water pycnometer method); CEC(meq+/100 g) 5 (ammonium acetate method for acidic soil,Chapman, 1965); organic matter content 0.14% loss ofweight on ignition (Storer, 1984); and bulk density 2.348mg/cc. The elements were: total arsenic 3 mg/kg; totaliron 3719 mg/kg; total silicon 390,000 mg/kg; aluminium2400 mg/kg; total manganese 185 mg/kg; magnesium 635mg/kg; lead 11 mg/kg; and zinc 18 mg/kg.

900

400

100

020 30 40

Alb

ite

Mag

nes

ium

alu

min

um

sili

cate

hy

dro

xid

e

Sil

ico

n o

xid

e

mag

hem

ite

Q (

sub

cell

)

Sy

lvin

e so

dia

n

Position (2θ) (copper (Cu))

Fig. 2 XRD spectra of the As spiked soil sample (52.5 mg/kg).

Extracted As from spiked soil was as following: total Aswas 22.6 mg/kg, As(III) was 1.7 mg/kg, and As(V) was20.9 mg/kg as low contaminated soil, while total As was52.5 mg/kg, As(III) was 3.4 mg/kg, and As(V) was 49.1mg/kg as high contaminated soil. The characterization ofextractant are summarized in Table 2.

2.2 Arsenic sorption in soil

The soil spiked with 50 mg/L As solution is found to retain22.6 mg/kg of As (low contaminated soil) after washingwith artificial rain water of pH 5.9, whilst soil spikedwith 200 mg/L As solution retains 52.5 mg/kg As (high

Table 2 Characterization of extractants

Extractant Empirical Molecular Critical micelle Concentration Surface pH Viscosityformula weight concentration at 25°C tension (mN/m) at 25°C (cP)

Water H2O 18 – – 71.2 7 0.89Soapnut C52H84O21·2H2O 1081.24 0.1% 0.5% 41 4.33 1.1Soapnut C52H84O21·2H2O 1081.24 0.1% 1% 40 4.26 1.2SDS NaC12H25SO4 288.38 8.2 mmol/L 20 mmo/L 34 7.5 1.4


1.5

1.0

0.5

0

-0.5

-1.0

Eh

(V

)

0 2 4 6 8 10 12 14pH

H3AsO4

H3AsO3

H2AsO4

-

H2AsO3

-

HAsO4

2-

AsO4

3-

Native As

Stating point

Ending point

Fig. 3 Initial and final pH and Eh values of contaminated soil duringcolumn washing experiment reported in the Eh-pH diagram for the systemAs-O-H at 25°C and 1 bar with activities of soluble As species = 10−6

mol/L (Lu and Zhu, 2011). Gray shaded areas denote solid phases.

contaminated soil) (Table 1). Arsenic is retained in thesoil matrix mostly by hydrous oxides of Fe(III) and Al(III)(Jacobs et al., 1970). Arsenic adsorption by soil organicmatter and silica are negligible (Wasay et al., 1996; Weng

et al., 1997).The unspiked soil has a pH value of 4.5 and Eh value

of 260 mV. According to the revised Eh-pH diagramsfor the As-O-H system at 25°C and 1 bar (Lu and Zhu,2011), arsenic is expected to exist in +5 state under theseconditions in aqueous matrices. Hence, aqueous solution ofNa2HAsO4·7H2O was used to spike the soil. After spikingand washing the soil, the pH and Eh values of soil werefound to be 5.2 and 210 mV respectively. In Fig. 3, the Eh-pH diagram illustrates this scenario; highlighting the initialand final Eh and pH values of both the unspiked and spikedsoil samples. A slight decrease in Eh value in the spikedsoil was accompanied by an increase in soil pH value.The As speciation in high and low contaminated soils bysolvent extraction confirmed the presence of As(V) species(up to 94% of total As).

2.3 Cumulative As removal in down-flow mode andsuction mode

The data for cumulative As removal from both low andhigh contaminated soil columns are provided in Fig. 4.The cumulative As removal efficiency was the highestin the presence of 20 mmol/L SDS solution in all fourscenarios. SDS 20 mmol/L solution was succeeded by1% soapnut solution. Conversely, the As removal by 0.5%soapnut solution was considerably low. To account for thewater soluble As in the column, it was washed with waterwhich removed a maximum of 6.4% of As in the low

55

50

45

40

35

30

25

20

15

5

055

50

45

40

35

30

25

20

15

5

0

-

-

-

-

-

-

-

-

-

-

Cu

mu

lati

ve

As

rem

ov

al (

%)

Cu

mu

lati

ve

As

rem

ov

al (

%)

Water floodSoapnut 0.5%Soapnut 1%SDS 20 mmol/L

a b

c d

PV1 PV2 PV3 PV4 PV5 PV6 PV1 PV2 PV3 PV4 PV5 PV6

Fig. 4 Cumulative As removal from (a) down-flow mode+low contaminated soil, (b) down-flow mode+ high contaminated soil, (c) suction mode+lowcontaminated soil, (d) suction mode+high contaminated soil.


contaminated soil in 6 PVs. Therefore, it is evident thatvery little water-soluble loosely bound As was present inthe soil column, perhaps in As(III) state as indicated by thespeciation study. From low contaminated soil in down-flowmode, 20 mmol/L SDS, 1% and 0.5% soapnut solutionremoved 8.8, 7.2 and 3.8 times more As respectively thanthat recovered with water. The corresponding values forhigh contaminated soil in down-flow mode are a factorof 9.8, 7.7 and 4.8 times greater than that recovered withwater. The trend was similar in suction modes as well.For the low contaminated soil washed in suction modewith 20 mmol/L SDS, 1% and 0.5% soapnut solution, therespective As removal was 8.7, 7.3 and 4 times more thanthat recovered with water. The corresponding values forhigh contaminated soil in suction mode are a factor of 9.7,8 and 5.2 times greater than that recovered with water.

Arsenic removal increased significantly when the con-centration of soapnut was increased from 0.5% to 1%.For low contaminated soil in down-flow mode, after 6PVs using 0.5% solution, approximately 1.66 mg As wasrecovered compared to 3.1 mg using a 1% solution and 3.6mg using 20 mmol/L SDS solution. In contrast, the waterflood recovered only 0.44 mg As under similar conditions.For high contaminated soil in down-flow mode, 0.5% and1% soapnut, 20 mmol/L SDS and water flood removed 3.4,5.36, 6.85 and 0.70 mg of As from the column respectively.From the low spiked soil in suction mode, SDS removed3.78 mg As, 1% soapnut solution removed 3.17 mg Aswhilst 0.5% soapnut solution recovered 1.8 mg of Ascompared to a 0.44 mg with water flood. In the case ofthe highly spiked soil in suction mode, the correspondingvalues are 6.74, 5.65, 3.6 and 0.7 mg respectively.

It is clear that with an increase in the concentrationof natural surfactant, there is a significant increase in theremoval of As from the soil column. This can be explainedby the increased solubility of As in the surfactant micelles.Other studies observed that, with an increase in surfactantconcentration above CMC, the number of micelles usuallyincrease, resulting in enhanced solubilisation of pollutants,which are easily mobilized and washed from the soilmatrix (Mulligan, 2005).

The square root of the percentage of arsenic removal ineach pore volume for both down-flow and suction modeshas been shown in Fig. S1. The As concentration in the ef-fluent was the highest during the initial PVs and continuedto decrease in the subsequent washings. Roy et al. (1995)also observed high initial removal of residual transmissionfluid from the soil column. This was attributed to any freephase pollutant in the column. Any loosely bound As(V)may get easily detached from the soil particles by theinitial spurge of the surfactant having enhanced wettabilitythan water and high micelle forming ability. The laterPVs experience extraction of strongly bound As. However,opposite observations were made by other researchers whoexperienced less removal in initial PVs (Kommalapati et

al., 1997; Roy et al., 1997; Wang and Mulligan, 2009). Thismight be due to the presence of strongly attached pollutantswhich required higher desorption time to leach out fromthe soil particles.

The soapnut concentration in the effluent increasedafter the 3rd or 4th PV, signifying that during the initialstages the extractant solutions underwent adsorption on thesoil particles thereby extracting the pollutant by micellarsolubilisation. As the washing process progressed, thesurfactant absorbance in the soil reached saturation. Atthe 5th and 6th PV, the concentration of effluent soapnutsolution resembled that of influent solution. Roy et al.(1997) also observed that the soapnut concentration wasless than the influent in the initial effluents and increasedgradually with each pore volume.

Figure 5 compares the overall performance of all thesurfactants. The anionic surfactant SDS was the betterextractant. However, soapnut at 1% concentration achieved78%–86% of the performance of SDS under all condi-tions. Considering that soapnut is an environment friendlybiodegradable non-ionic surfactant, this performance isencouraging and it merits further investigation. Washing1 ton of low level arsenic contaminated soil in the ratioof 300:480 (corresponding to washing of 300 g soil with6 PV liquid of 80 mL each) by 1% SN and 20 mmol/LSDS solutions under similar conditions will require 22.86kg of soapnut and 9.23 kg of SDS, both of which will costroughly 30 USD at the current market price. Preliminaryexperiments and previously published work indicate thatsoapnut concentration higher than 1% developed excesspressure inside the column and the process slowed down(Roy et al., 1997).

2.4 Statistical analysis

Analysis of variance (ANOVA) is shown in Table 3. TheF value for a term is the test for comparing the varianceassociated with that term with the residual variance. It is

Low contaminated soil+

Down-flow mode

Low contaminated soil+

Suction mode

High contaminated soil+

Down-flow mode

High contaminated soil+

Suction mode

0

10

20

30

40

50

60

70

80

90

Cum

ula

tive

arse

nic

rem

oval

(%

)

Water flood

Soapnut 0.5% Soapnut 1%

SDS 20 mmol/L

Fig. 5 Comparison of As removal performance by surfactants underdifferent flow modes.


Table 3 ANOVA for selected factorial model for cumulative As removal

Source Sum of squares DF Mean square F value p value

Model 75.57 35 2.159 38.484 < 0.0001 SignificantA (wash solution) 47.89 3 15.962 284.492 < 0.0001B (washing mode) 0.068 1 0.068 1.216 0.2745C (contamination) 1.07 1 1.068 19.033 < 0.0001D (pore volume) 14.17 5 2.835 50.525 < 0.0001AD 10.37 15 0.691 12.322 < 0.0001BD 0.92 5 0.183 3.265 0.0114CD 1.13 5 0.226 4.031 0.0033Residual 3.31 59 0.056Lack of fit 3.082 58 0.053 0.232 0.9575 Not significantPure error 0.229 1 0.229Cor Total 78.88 94Std. Dev. 0.237 R-squared 0.958Mean 2.015 Adj R-squared 0.933CV (%) 11.757 Pred R-squared 0.891PRESS 8.575 Adeq precision 28.537

Run no 62 was ignored during the analysis due to bad data points.

the mean square for the term divided by the mean squarefor the residual. The Model F-value of 38.484 implies theproposed model is significant. There is a 0.01% chancethat a “Model F-value” this large could occur due to noise.Model terms were evaluated by the p-value (probability)with 95% confidence level. This is the probability valuethat is associated with the F value for this term. It isthe probability of getting an F value of this size if theterm did not have an effect on the response. In general,a term that has a probability value less than 0.05 would beconsidered a significant effect. A probability value greaterthan 0.10 is generally regarded as not significant. A, C,D, AD, BD, CD are significant model terms. However, thefactor B was also included to maintain the heirarchy of thedesign. The “Pred R-squared” of 0.891 is in reasonableagreement with the “Adj R-squared” of 0.933. The AdeqPrecision value of 28.537 indicates an adequate signal fitto be used for navigating the design space. The interactiongraphs for sqrt of As removal at different PVs are given inFig. S1. From the normal probability plot of the residuals(Fig. S2) it can be seen that the data points are fairly closeto the straight line indicating that the experimental resultsconform to a normally distributed population (Antony,2003). Diagnostic plots of predicted versus actual valuesaids in judging the quality of the model. Table S1 enliststhe actual and predicted values of the experimental datapoints. Figure S3 indicates a good agreement betweenactual data and data predicted by the model which is notshown here. The overall statistical analysis indicates therobustness of the experimental data.

2.5 Pressure build-up in soil column

Table 4 illustrates the variation in pressure developmentfor the surfactant runs for low contaminated soils inboth down-flow and suction modes. The pressure build-up was the highest for 1% soapnut solution, followedby 20 mmol/L SDS solution, 0.5% soapnut solution andwater. Previous studies also experienced the development

of pressure in down-flow mode in the soil column (Kom-malapati et al., 1998; Roy et al., 1995). However, forthe suction mode, the pressure build-up was much lowerthan down-flow mode. The maximum pressure build-uprecorded by 1% soapnut at the sixth PV was only 83kPa compared to 214 kPa at down-flow mode, resultingin a much faster operation in down-flow mode. Therefore,a fast process with low pressure development makes thesuction mode more advantageous than down-flow mode.The development of higher pressure in down-flow moderesulted from clogging of the soil pores due to a dispersionof colloids and trapping of air bubbles inside the soil pores,which obstructed the flow of flushing solution through thecontaminated area, reducing the efficiency of As removalfrom the soil matrix (Nash, 1987; Roy et al., 1995). Suctionmode also produced channelized flow and pore clogging,however the entrapment of air bubbles was negligibledue to the suction pressure provided at the outlet. Thesurfactant solutions easily flowed through the channels tothe outlet point without building up undesirable pressure,making the operation easier. The cumulative As removalwas similar in both suction and down-flow modes.

2.6 Kinetics of As extraction from soil column

Whilst performing the column washing procedure, thetime taken for each PV to pass through the column wasrecorded. The time increased with each subsequent PV.Ast was calculated from the samples collected at the endof each PV. Ast values are provided in the appendix inTable S2. ln(Ast/As0) was plotted against time and thegraph showed linearity. This confirmed that the kineticsof desorption phenomenon to be of first order for theentire period of column washing. Table 5 summarizesthe kd

′ values which ranged from 0.001 to 0.010 min−1.Satisfactory fits were obtained and, with a few exceptions,R2 values were above 0.9. The kd

′ values for suction modeswere much higher than down-flow mode due to a faster rateof washing. Whilst down-flow mode took up to 170 min for


Table 4 Pressure build-up in low contaminated soil column

Pressure build-up (kPa)Pore volume Water flood Soapnut 0.5% Soapnut 1% SDS 20 mmol/L

Down-flow mode 1 0 7 21 142 7 21 55 283 10 48 103 554 14 69 131 1175 21 90 159 1456 34 110 214 193

Suction mode 1 0 7 14 72 7 14 17 143 8 28 28 214 10 31 48 345 12 38 69 626 14 48 83 76

Table 5 Apparent desorption rate coefficients (kd′) and R2 for As desorption from the soil columns

Low contamination soil High contamination soilkd′ (min−1) R2 kd

′ (min−1) R2

Down-flow mode Soapnut 0.5% 0.001 0.997 0.001 0.964Soapnut 1% 0.002 0.978 0.001 0.979SDS 20 mmol/L 0.004 0.970 0.003 0.926

Suction mode Soapnut 0.5% 0.006 0.967 0.005 0.990Soapnut 1% 0.008 0.903 0.006 0.900SDS 20 mmol/L 0.010 0.948 0.007 0.916

passing 6 PV, suction mode took only up to 75 min of timeat the maximum. In general, SDS demonstrated the highestkd′, followed by 1% and 0.5% soapnut solutions.

2.7 Zeta potential, FT-IR spectral data and mechanismof As removal by soapnut

The zeta potential values of the soil particles were mea-sured in de-ionized water, 20 mmol/L SDS, 0.5% and1% soapnut solutions and were found to have values of–34.3, –61.8, –17.1 and –11.8 mV respectively. Therefore,in the case of both SDS and soapnut, zeta potential valuesunderwent significant change. Compared to water, thezeta potential decreased significantly for 20 mmol/L SDS,which indicates adsorption of the anionic surfactant SDSon the surface of soil particles. A similar decrease in zetapotential of kaolinite was observed when it sorbed SDSon its already negative basal plane, because of the originalnegative kaolinite charge plus the negative charge due tosorbed SDS head groups (Ko et al., 1998). However, thezeta potential value was much higher for soapnut dueto the non-ionic tails of saponin molecules, which wereadsorbed on the soil particles, thereby reflecting higherzeta potential values. It was postulated that surfactant ad-sorption is essential for the removal of soil contaminants,and surfactants that adsorb at the soil-water inter-phasesare better detergents (Raatz and Hartel, 1996). Therefore,soapnut and SDS, both were adsorbed on soil particle andwere effective detergents.

FT-IR spectral data as exhibited in Fig. 6, displayed thedifferences in average absorbance spectra for the influent

and the effluent soapnut solutions, together with the ab-sorption range of different molecular vibrations present inphenolic-OH at 3435 cm−1, carbonyl groups of carboxylicacid at 2090 cm−1 and alkene groups at 1640 cm−1, similarto earlier observations (Pradhan and Bhargava, 2008). Noshifting of peaks in FT-IR spectra was observed in thesoapnut solution in presence of As in the soil column.Similar analysis in UV-Visible frequency range also didnot show any shift in the peaks. The UV-Vis spectra arenot shown here due to absence of significant observations.Thus in this study, no suggestion of chemical interactionof As with soapnut was obtained. However, earlier workssuggested complexation of saponin molecule with heavymetals (Hong et al., 2002; Song et al., 2008). A mechanismfor arsenic removal by nonionic soapnut can be proposedas in Fig. 7. The nonionic surfactant gets adsorbed ontothe soil surface and gets attached to the arsenic by physicalforce. Arsenic which is in turn adsorbed on soil particlegets detached and goes into the solution and subsequentlygets associated with surfactant micelles.

3 Conclusions

The performance of 1% soapnut solution reached an effi-ciency of 86% of the performance of 20 mmol/L solutionof anionic SDS both in low and high-contaminated soilin down-flow and suction modes. Considering the factthat soapnut is non-ionic, this performance is satisfac-tory. Arsenic desorption was found to occur mostly bymicellar solubilisation following first order kinetics. The


70

60

50

40

30

20

10

0

-

-

-

-

-

-

-Tra

nsm

itta

nce

(%

)

70

60

50

40

30

20

10

0

-

-

-

-

-

-

-

Tra

nsm

itta

nce

(%

)

4000 3000 2000 1500 1000 550

4000 3000 2000 1500 1000 550

Wave number (cm-1)

Wave number (cm-1)

3435

2370

2090

1640

770

Phenolic-OH Carbonylstretching

Soapnut

Soapnut + As wash solution

Fig. 6 FT-IR spectra of soapnut solutions before and after washing.

Fig. 7 Mechanism of arsenic removal from soil by soapnut solution.

low contaminated soil column in suction mode had higherrate constants. A model was proposed for As desorptionfrom each PV by Design expert software and the data fitswere satisfactory indicating robustness of the experimentalobservations. The As removal during initial PVs were high,whilst it decreased during the later PVs in agreement withpublished literature (Roy et al., 1995). The performanceof extractant solutions was similar in both down-flow andsuction modes. Suction mode generated up to 50% lesspressure inside the column and was more advantageousthan the traditional down-flow mode, which experienced

significant pore clogging, and air bubble entrapment (Royet al., 1995). Zeta potential measurements confirmed veryweak ionic charge in the hydrophobic tails of soapnutmolecules, thus eliminating the ionic interaction mech-anism behind As removal by soapnut. Absence of anychemical structure change in the soapnut solution, asevident from the FT-IR and UV-Vis spectra, opens upthe possibility of reusing the same soapnut solution afterseparation of As from the wash liquid. A mechanism forAs desorption has also been proposed. From economicperspective, it is estimated that soapnut will incur similarcost as that of SDS under similar conditions of washing.

Acknowledgments

The authors acknowledge the funding provided by Uni-versity of Malaya, Kuala Lumpur (No. PV102-2011A,UM-QUB6A-2011) for carrying out this research. MissSumona Mukherjee from University of Malaya and MissAntonia O’Neill from the Environmental EngineeringResearch Centre, Queen’s University Belfast, are acknowl-edged for proofing the article for English language.

References

Antony J, 2003. Design of Experiments for Engineers andScientists. Butterworth-Heinemann, New York.

Chapman H D, 1965. Cation-exchange capacity. In: Methodsof Soil Analysis-Chemical and Microbiological Properties(Black CA, ed.). Agronomy, 9. 891–901.

Chappell J, Chiswell B, Olszowy H, 1995. Speciation of arsenicin a contaminated soil by solvent extraction. Talanta, 42(1):323–329.

Chen W J, Hsiao L C, Chen K K Y, 2008. Metal desorptionfrom copper(II)/nickel(II)-spiked kaolin as a soil compo-nent using plant-derived saponin biosurfactant. ProcessBiochemistry, 43(5): 488–498.

Cheng H F, Hu Y N, Luo J, Xu B, Zhao J F, 2009. Geochemicalprocesses controlling fate and transport of arsenic in acidmine drainage (AMD) and natural systems. Journal ofHazardous Materials, 165(1-3): 13–26.

Chowdhury S R, Yanful E K, 2010. Arsenic and chromiumremoval by mixed magnetite-maghemite nanoparticles andthe effect of phosphate on removal. Journal of Environmen-tal Management, 91(11): 2238–2247.

Dobran S, Zagury G J, 2006. Arsenic speciation and mobilizationin CCA-contaminated soils: Influence of organic mattercontent. Science of The Total Environment, 364(1-3): 239–250.

Dwarakanath V, Kostarelos K, Pope G A, Shotts D, Wade WH, 1999. Anionic surfactant remediation of soil columnscontaminated by nonaqueous phase liquids. Journal ofContaminant Hydrology, 38(2): 465–488.

Hong K J, Tokunaga S, Kajiuchi T, 2002. Evaluation of remedia-tion process with plant-derived biosurfactant for recovery ofheavy metals from contaminated soils. Chemosphere, 49(2):379–387.

Jacobs L W, Syers J K, Keeney D R, 1970. Arsenic sorption bysoils. Soil Science Society of America Journal, 34(5): 750–


754.Jang M, Hwang J S, Choi S I, 2007. Sequential soil washing

techniques using hydrochloric acid and sodium hydroxidefor remediating arsenic-contaminated soils in abandonediron-ore mines. Chemosphere, 66(1): 8–17.

Jang M, Hwang J S, Choi S I, Park J K, 2005. Remediation ofarsenic-contaminated soils and washing effluents. Chemo-sphere, 60(1): 344–354.

Ko I, Chang Y Y, Lee C H, Kim K W, 2005. Assessment of pilot-scale acid washing of soil contaminated with As, Zn and Niusing the BCR three-step sequential extraction. Journal ofHazardous Materials, 127(1-3): 1–13.

Ko S O, Schlautman M A, Carraway E R, 1998. Effects ofsolution chemistry on the partitioning of phenanthrene tosorbed surfactants. Environmental Science and Technology,32(22): 3542–3548.

Kommalapati R R, Valsaraj K T, Constant W D, Roy D, 1997.Aqueous solubility enhancement and desorption of hex-achlorobenzene from soil using a plant-based surfactant.Water Research, 31(9): 2161–2170.

Kommalapati R R, Valsaraj K T, Constant W D, Roy D, 1998.Soil flushing using colloidal gas aphron suspensions gen-erated from a plant-based surfactant. Journal of HazardousMaterials, 60(1): 73–87.

Lu P, Zhu C, 2011. Arsenic Eh-pH diagrams at 25°C and 1 bar.Environmental Earth Sciences, 62(8): 1673–1683.

Mason R L, Gunst R F, Hess J L, 2003. Statistical Design andAnalysis of Experiments, Eighth Applications to Engineer-ing and Science, Second ed. Wiley, New York.

Mulligan C N, 2005. Environmental applications for biosurfac-tants. Environmental Pollution, 133(2): 183–198.

Mulligan C N, Wang S, 2006. Remediation of a heavy metal-contaminated soil by a rhamnolipid foam. EngineeringGeology, 85(1-2): 75–81.

Nash J H, 1987. Field Studies of In-Situ Soil Washing,EPA/600/2–87/110. U. S. Environmental Protection Agen-cy, Cincinnati, Ohio.

Oorts K, Ghesquiere U, Smolders E, 2007. Leaching and agingdecrease nickel toxicity to soil microbial processes insoils freshly spiked with nickel chloride. EnvironmentalToxicology and Chemistry, 26(6): 1130–1138.

Polettini A, Pomi R, Calcagnoli G, 2009. Assisted washingfor heavy metal and metalloid removal from contaminateddredged materials. Water, Air, and Soil Pollution, 196(1-4):183–198.

Pradhan M, Bhargava P, 2008. Defect and microstructural evolu-tion during drying of soapnut-based alumina foams. Journalof the European Ceramic Society, 28(16): 3049–3057.

Raatz S, Hartel G, 1996. Application of surfactant combinationsfor cleaning clays contaminated with polycyclic aromatichydrocarbons. Aufbereitungs-Tensidk, 37(2): 57–62.

Roy D, Kommalapati R R, Mandava S, Valsaraj K T, ConstantW D, 1997. Soil washing potential of a natural surfactant.

Environmental Science and Technology, 31(1): 670–675.Roy D, Kommalapati R R, Valsaraj K T, Constant W D, 1995.

Soil flushing of residual transmission fluid: application ofcolloidal gas aphron suspensions and conventional surfac-tant solutions. Water Research, 29(2): 589–595.

Sivasubramaniam S, Talibudeen O, 1972. Potassium-aluminiumexchange in acid soils I. Kinetics. Journal of Soil Science,23(2): 163–176.

Song S, Zhu L Z, Zhou W J, 2008. Simultaneous removalof phenanthrene and cadmium from contaminated soilsby saponin, a plant-derived biosurfactant. EnvironmentalPollution, 156(1): 1368–1370.

Sparks D L, Zelazny L W, Martens D C, 1980. Kinetics ofpotassium desorption in soil using miscible displacement.Soil Science Society of America Journal, 44(6): 1205–1208.

Storer D A, 1984. A simple high sample volume ashing procedurefor determination soil organic matter. Communications inSoil Science and Plant Analysis, 15(7): 759–772.

Suhagia B N, Rathod I S, Sindhu S, 2011. Sapindus mukorossi(Areetha): An overview. International Journal of Pharma-ceutical Sciences and Research, 2: 1905–1913.

Sundstrom D W, Weir B A, Klei H E, 1989. Destruction ofaromatic pollutants by UV light catalyzed oxidation withhydrogen peroxide. Environmental Progress, 8(1): 6–11.

Tokunaga S, Hakuta T, 2002. Acid washing and stabilization ofan artificial arsenic-contaminated soil. Chemosphere, 46(1):31–38.

Torres L G, Lopez R B, Beltran M, 2012. Removal of As, Cd,Cu, Ni, Pb, and Zn from a highly contaminated industrialsoil using surfactant enhanced soil washing. Physics andChemistry of the Earth, Parts A/B/C, 3739: 30–36.

Wang S L, Mulligan C N, 2009a. Arsenic mobilization frommine tailings in the presence of a biosurfactant. AppliedGeochemistry, 24(5): 928–935.

Wang S L, Mulligan C N, 2009b. Rhamnolipid biosurfactant-enhanced soil flushing for the removal of arsenic and heavymetals from mine tailings. Process Biochemistry, 44(1):296–301.

Wang S L, Zhao X Y, 2009. On the potentialof biological treatment for arsenic contaminated soilsand groundwater. Journal of Environmental Management,90(8): 2367–2376.

Wasay S A, Haron M J, Tokunaga S, 1996. Adsorption offluoride, phosphate, and arsenate ions on lanthanum-impregnated silica gel. Water Environment Research, 68(1):295–300.

Weng H, Liu Y, Chen H, 1997. Environmental geochemical fea-tures of arsenic in soil in China. Journal of EnvironmentalSciences, 9(2): 385–395.

Yamaguchi N, Nakamura T, Dong D, Takahashi Y, Amachi S,Makino T, 2011. Arsenic release from flooded paddy soilsis influenced by speciation, Eh, pH, and iron dissolution.Chemosphere, 83(7): 925–932.

Comparison of a plant based natural surfactant with SDS for washing of As(V) from Fe rich soil

Documents

Transcript of Comparison of a plant based natural surfactant with SDS for washing of As(V) from Fe rich soil