8

6

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

Research Article, Issue 2

Analytical Methods in Environmental Chemistry Journal

Journal home page: www.amecj.com/ir

AMECJ

Determination of fenthion in environmental water samples

by dispersive liquid–liquid microextraction coupled with

spectroﬂuorimetric and chemometrics methods

a

b,

c

Tahereh Eskandari , Ali Niazi , Mohammad Hossein Fatemi and Mohammad Javad Chaichi

a

Department of Chemistry, Arak Branch, Islamic Azad University, Arak, Iran

Department of Chemistry, Central Tehran Branch, Islamic Azad University, Tehran. Iran

Department of Analytical Chemistry, Faculty of Chemistry, Mazandaran University, Babolsar, Iran

b

c

A R T I C L E I N F O :

Received 15 Feb 2021

A B S T R A C T

In the present study, a simple, rapid and efﬁcient dispersive liquid–

liquid microextraction (DLLME) coupled with spectroﬂuorimetry

Revised form 24 Apr 2021

Accepted 20 May 2021

Available online 30 Jun 2021

(SFM) and chemometrics methods have been proposed for the

preconcentration and determination of fenthion in water samples.

Box–Behnken design was applied for multivariate optimization

of the extraction conditions (sample pH, the volume of dispersive

solvent and volume of extraction solvent). Analysis of variance

was performed to study the statistical signiﬁcance of the variables,

their interactions and the model. Under the optimum conditions, the

------------------------

Keywords:

Fenthion,

Pesticides,

Organophosphoruse pesticides,

Dispersive liquid–liquid icroextraction,

Box–Behnken design,

Spectroﬂuorimetry

-1

calibration graph was linear in the range of 5.0–110 ng mL with

the detection limit of 1.23 ng mL (3S /m). Parallel factor analysis

-1

b

(PARAFAC) and partial least square (PLS) modelling were applied

for the multivariate calibration of the spectroﬂuorimetric data. The

orthogonal signal correction (OSC) was applied for preprocessing of

data matrices and the prediction results of model, and the analysis

results were statistically compared. The accuracy of the methods,

evaluated by the root mean square error of prediction (RMSEP) for

fenthion by OSC-PARAFAC and OSC-PLS models were 0.37 and

0

.78, respectively. The proposed procedure could be successfully

applied for the determination of fenthion in water samples.

1

. Introduction

phosphorothioate ) is a contact and stomach or-

The organophosphorous pesticides (OPPs) have

been widely used in agriculture for crop production

and fruit tree treatment, but many of them are iden-

tiﬁed as highly toxic compounds [1–3]. They are

released into the environment from manufacturing,

transportation and agriculture applications. OPPs

have been found in ground waters, surface waters,

lagoons and drinking water. Fenthion (O,O-Di-

ganophosphorous pesticide widely used in the con-

trol of many sucking, biting pests, especially fruit

ﬂies, stem borers and mosquitoes on crops such as

alfalfa, rice, sugar, vegetables and forests. Fenthion

is toxic for the human and animal health [4–6]. The

toxicological effect of fenthion, is almost entirely

due to the inhibition of acetylcholinesterase in the

nervous system, resulting in respiratory, myocar-

dial and neuromuscular transmission impairment

methyl

O-[3-methyl-4-(methylsulfanyl)phenyl]

[5, 7]. Due to the low concentration of the analytes

*

Corresponding Author: Ali Niazi

and the complex matrix of the samples, a prelim-

inary sample preconcentration and a separation

Email: ali.niazi@gmail.com, ali.niazi@iauctb.ac.ir

https://doi.org/10.24200/amecj.v4.i02.138

Determination of fenthion by DLLME-SFM

Tahereh Eskandari et al

87

technique are required. Thus, different extraction

processes have been used for separation and

pre-concentration of trace pesticide residues, such

as solid phase extraction method (SPE) [8–11], sol-

id phase microextraction (SPME) [12], [13], single

drop microextraction (SDME) [14] and dispersive

liquid–liquid microextraction (DLLME) [15–17].

In the last decades, liquid–phase microextraction

matical and statistical method. The main advantage

of RSM is that it reduces the number of experiment

because several factors can be varied simultaneous-

ly for optimization and as a result saves time, en-

ergy, and chemicals [16,37,38]. Box–Behnken de-

sign is the most common and efﬁcient design used

in RSM. Box–Behnken design is a second order

multivariate technique based on three level partial

factorial designs. Box–Behnken is a spherical, ro-

tatable or nearly rotatable that consists of a central

point and with the midpoints of the edges of the

variable space [15–17], [33, 34]. Two dimensional

excitation emission (EEM) ﬂuorescence data can

be obtained by measuring the emission spectra at

various excitation wavelengths. In recent years, ap-

plication of multi-way data analysis techniques has

increased signiﬁcantly in the analytical chemistry.

There are several multivariate calibration proce-

dures that can be used for the treatment of EEM ﬂu-

orescence data, in order to quantify the compounds,

present in a mixture [39]. In ﬂuorescence analysis,

parallel factor analysis (PARAFAC) [28], [40–44]

and partial least-squares regression (PLS) [34, 43],

[45–47] has been mostly applied for the analyses

of three-way data obtained as series of emission

spectra measured for different excitations. PLS is a

factor analysis method that has been used in multi-

component quantitative analysis from several spec-

tral data, such as IR, UV-visible or ﬂuorescence

[47]. Partial Least Squares (PLS) regression is a

method to predict the response variable based on

predictor variables and to describe their common

structure. The main advantage of PLS calibration

procedures is that they can model a system even

in the presence of interfering signals, provided that

they are included in the calibration step. PARA-

FAC is a multi-way decomposition method that

has investigated to be useful for the analysis of

second-order calibration. The main advantages of

the PARAFAC model are the uniqueness, simplic-

ity of its solutions and quantiﬁcation of an analyte,

even in the presence of unknown interferences (the

second-order advantage) [40, 44].The orthogonal

signal correction (OSC) is a useful pre-process-

ing step that improves the chemometrics model

(LPME), based on the miniaturization of traditional

LLE technique by greatly reducing the use of or-

ganic solvent has been reported as an alternative

for sample preparations. One of the most popular

LPME techniques is dispersive liquid-liquid mi-

croextraction (DLLME) which is widely used as

a preconcentration method [18-21]. DLLME was

developed by Assadi and co-workers [16]. By con-

sisting of the formation of a cloudy solution pro-

moted by the fast addition in the aqueous sample

of a mixture of extractor and dispersive solvents.

The tiny droplets formed and dispersed among the

aqueous sample solution are further joined and

sedimented in the bottom of a conical test tube by

centrifugation. This method provides many advan-

tages including rapidity, simplicity of operation,

high recovery and enrichment factor. After sample

preparation, the determination of OPPs in different

sample matrices was carried out by using gas chro-

matography mass spectrometry (GC-MS) [9, 22],

gas chromatography (GC) [23–25] and high-per-

formance liquid chromatography (HPLC) [26,27].

Fluorescence spectrometry is a sensitive, selective

and relatively low cost method for the quantitative

analysis of pesticides and other pollutants [28–30].

Different experimental variables can affect the ex-

traction yield in the DLLME procedure; therefore,

a multivariate approach has been widely used for

their optimization. Statistical methods are useful to

determine the effects of variables on the extraction

procedure. The response surface methodology

(

RSM) based on statistical design of experiments

DOEs) has been extensively used for modelling

(

and optimization in various analytical procedures

31–36]. Response surface methodology (RSM)

[

is powerful multivariate technique that used for

building empirical model via collection of mathe-

8

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

by ﬁltering systematic variation in the spectra not

associated with the concentration [40].

2.2. Apparatus and software

The pH was determined with a model 780 Metrohm

pH-meter with combined glass–calomel electrode.

A centrifuge (Sigma) was used to accelerate the

phase separation process. A PerkinElmer, LS 45

Spectroﬂuorimeter enhanced by 150 W Xe lamp

was coupled with a computer and equipped with a

300 µL quartz microcell which was used for record-

ing the spectra using Windows 7 operating system.

All the measurements were done at the exciting

wavelength of 200-300 nm for every 10 nm, and at

the emission wavelength in the 300-500 nm range

for every 1 nm. Box–Behnken design and statisti-

cal analysis were performed with Minitab Version

16. The programs for PLS, PARAFAC, and OSC

calculation were written in MATLAB 2018 and run

on a personal computer (CPU 3.0 GHz and RAM 4

GB) equipped with the Windows 7 operating sys-

tem. The applied OSC version is based on the Wold

et al. algorithm.

The aim of this paper is to develop a fast, sensi-

tive and inexpensive spectroﬂuorimetric method

coupled with PARAFAC modelling for the deter-

mination and preconcentration of fenthion in envi-

ronmental water samples using DLLME procedure.

Also, the effects of various experimental variables,

including sample pH, the volume of extraction sol-

vent and volume of dispersive solvent were inves-

tigated and optimized using Box–Behnken design.

2

. Experimental

2

.1. Chemicals and reagents

All chemicals and solvents, such as methanol, eth-

anol, acetonitrile, chloroform, carbon tetrachloride,

chlorobenzene, and dichloromethane were pur-

chased from Sigma–Aldrich and Merck. fenthion

standards were obtained from Dr. Ehrenstorfer

(Augsburg, Germany). All of the reagents used in

this work were of analytical grade. Chloroform,

carbon tetrachloride, chlorobenzene and dichloro-

methane purchased from Sigma, Germany. Abuffer

solution was prepared using universal buffer solu-

tion. Universal buffer solutions were prepared by

mixing phosphoric, acetic, and boric acid. A stock

2.3. Experimental procedure

10 mL of sample solution containing 5.0–110.0 ng

-1

mL of Fenthion, and 1.0 mL of buffer solution

(pH was adjusted to 10.0) was poured into a test

tube with a conical bottom. Then an appropriate

mixture of disperser solvent (methanol, 600 µL)

and extraction solvent (chlorobenzene, 220 µL)

was rapidly injected into the sample tube. In this

step, a cloudy solution was immediately formed in

the test tube and then, it was centrifuged for 1 min

at 3000 rpm to separate the phases. Finally, the up-

per aqueous solution was removed by syringe, and

the sediment phase was used for subsequent mea-

surement by spectroﬂuorimetric which was shown

in Fig.1.

-1

solution of Fenthion (C H O PS , 1000 mg L )

10

5

3

2

was prepared by dissolving appropriate amounts of

analyte in methanol, stored under dark conditions

in refrigerator (Schema 1) and synthesized by con-

densation of 4-methylmercapto-m-cresol and di-

methyl phosphorochloridothionate. Working stan-

dard solutions were obtained daily by appropriately

diluting this stock solution with ultrapure water.

3

. Result and discussion

3

.1. Selection of extraction and dispersive

solvent

The selection of an appropriate extraction solvent

is very important for a DLLME procedure. It must

have some properties, such as higher density than

water, good extraction efﬁciency of the analytes,

and low solubility in water. Chloroform, carbon

Schema1: Picture of Fenthion as an

organothiophosphate insecticide

Determination of fenthion by DLLME-SFM

Tahereh Eskandari et al

89

Fig.1. Dispersive liquid–liquid microextraction (DLLME) procedure.

tetrachloride, chlorobenzene and dichloromethane

3.2. Effect of pH and salt addition

were studied as extraction solvents. The results

showed that among the solvents tested, chloroben-

zene has the highest recovery in comparison with

the other tested solvents. Therefore, chlorobenzene

was chosen for further experiments. The disper-

sive solvent should be miscible with the organic

extraction solvent and the aqueous phase. Suitable

dispersive solvent can increase the surface area for

transferring the analyte from sample to extraction

solvent. Thereby, ethanol, methanol, acetonitrile

and acetone are selected for this purpose. The re-

sults indicated that the best recovery was obtained

by using methanol. Thus in this study methanol

was selected as suitable disperser solvent.

The extraction efﬁciency for analyte can be affect-

ed by adjusting the pH of the aqueous solution.

The effect of pH variation on extraction efﬁcien-

cy was investigated in the range of 1-12 and the

optimal pH was found to be 10 (Fig. 2). The in-

ﬂuence of salt addition is also an important factor

for extraction. Salt addition can improve extraction

yield in DLLME, especially for those analytes with

a lower solubility, as a result of a salting out ef-

fect. Therefore, NaCl in the concentration range of

1–15% (w/v) was studied as a salting agent and no

signiﬁcant effect on the extraction efﬁciency was

observed. Considering the obtained results, no ad-

dition of salt was chosen in the further analysis.

Fig. 2. The effect of pH variation on extraction efﬁciency

9

0

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

3

.3. Effect of extraction and centrifugation

lationship between the response and the vari-

time

ables can be presented by the ﬂowing equation

In DLLME, extraction time is deﬁned as the in-

terval time between injection of the extraction

mixture into the aqueous sample and starting

to centrifuge. The effect of the extraction time

was studied in the range of 1–10 min. The ex-

perimental results showed that time has no im-

pact on extraction efﬁciency. This means that

the transfer of the analyte from aqueous phase

to the extraction solvent was fast, which was the

advantage of DLLME procedure. Therefore, 1

min was deﬁned as extraction time. Centrifuga-

tion is a critical step in the DLLME technique,

in order to achieve the phase separation of ex-

traction phase from the aqueous phase. So, the

effect of centrifugation time was also examined

in the ranges of 1–5 min. It was observed that by

increasing the centrifugation time, the response

remained constant. Therefore, the time and rate

of centrifugation had no signiﬁcant effect on the

extraction efﬁciency. According to this result, 1

min was selected as the optimum centrifuge time,

in the following study.

(Eq.1):

2

i

Y = β + β X + β X + β X X + ε

0

∑

i

∑

i

∑

ij

i

j

(Eq. 1)

Where Y is the predicted response and X repre-

i

sented the effect of the independent variables.

2

Thus, X and X X represented the quadratic, and

i

j

interaction terms respectively[7]. β , β and β

ij (i≠j)

i

ii

were the coefﬁcient of linear, quadratic and inter-

action, respectively. β and ε represented the con-

0

stant and the random error, respectively.

Experimental data were ﬁtted to a second-order

polynomial mathematical equation. Analysis of

variance (ANOVA) was applied to the analysis of

experimental data at 95% conﬁdence interval so

that the signiﬁcance of each term was evaluated by

their corresponding p-values which are presented

in Table 3.

According to Table 3, it was concluded that all the

2

linear (X , X , X and X ) and quadratic terms ( X₁

1

2

3

2

4

2 2

,

X₂

, X₃and X₄), were signiﬁcant at 5% proba-

3

.4. Box–Behnken analysis

bility level. As shown in Table 3, the interaction

between sample pH and volume of extraction sol-

vent (X X ) and the volume of extraction solvent

Box–Behnken experimental design was used

to optimize and evaluate the main effects and

interaction effects of the process variables on

1

2

and dispersive solvent (X X ) were signiﬁcant. The

2

3

the recovery. The sample pH (X ), the volume

polynomial model is represented in Equation 2:

1

of extraction solvent (X ) and volume of dis-

2

persive solvent (X ) were selected as the three

3

independent variables as showed in Table 1.

The number of experiments (N) required for

the development of Box– Behnken design was

The 3D response surface graphs of the effect be-

tween each factor were shown in Figure 3. The

plot in Figure 3 displayed that the recovery in-

creased with the increase of initial solution pH

ranging from 9 to 10. However, pH values higher

than 10 reduced the recovery. Also, the response

ﬁrst increased with extraction solvent volume ap-

proximately 220 µL, and thereafter decreased. The

optimum values of the tested variables were ob-

tained as follows: X = 10.0, X = 220.0 µL and X =

defined as N = 2k

the number of factors and C is the number of

(

k −1

)

+ C₀, (where k was

0

central points). Thus, a total of 15 runs were

carried out for optimizing these three vari-

ables at three levels (low, medium and high).

The Box–Behnken design matrix and the re-

covery are presented in Table 2. According to

Box–Behnken matrix, a total of 15 tests con-

taining 3 replicates at the center point were

performed in random order.An empirical re-

1

2

3

600.0 µL.

Determination of fenthion by DLLME-SFM

Tahereh Eskandari et al

91

Table 1. Variables and their levels in Box-Behnken design.

Level

Factors

Symbol

Low (-1)

Medium (0)

High (+1)

pH

X₁

X₂

X₃

9

10

11

1

5

00.0

200.0

300.0

V_e_x_t(µL)

V_d_i_s_p(µL)

600.0

700.0

Table 2. Box-Behnken design matrix with obtained result.

Recovery

Actual level of factors

(%)

Run No.

X₁

X₂

X₃

1

2

3

4

5

6

7

8

9

10.0

9.0

100.0

300.0

200.0

300.0

200.0

100.0

200.0

100.0

300.0

200.0

500.0

700.0

600.0

700.0

600.0

500.0

600.0

500.0

600.0

500.0

700.0

59.2

78.4

97.1

77.6

69.5

73.2

71.2

60.4

95.7

68.5

70.1

81.3

96.5

65.2

71.5

11.0

10.0

9.0

10.0

11.0

10.0

9.0

1

0

1

2

3

4

5

9.0

9

2

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

Table 3. Analysis of variance evaluation of linear, quadratic, and interaction terms for each response variable.

a

b

c

MS

Variables

DF

SS

F-values

p-value

Model

X₁

9

1

2092.19

820.22

379.47

374.70

828.46

536.50

518.85

79.21

232.465

820.222

379.466

374.700

828.463

536.503

518.848

79.210

0.640

50.44

177.97

82.33

81.30

179.76

116.41

112.58

17.19

0.14

0.000

0.009

0.725

0.018

-----

X₂

1

X₃

1

2

X₁

1

2

X₂

1

2

X₃

1

X X₂

1

X X₃

1

0.64

1

X X₃

1

55.50

55.502

4.609

12.04

-----

2

Residual

Lack-of-Fit

Pure Error

Total

5

23.04

3

22.06

7.352

14.90

-----

0.064

-----

2

0.99

0.493

14

2

R = 98.91; Adjusted R = 96.95.

a

DF: degree of freedom.

SS: sum of squares.

MS: mean square.

b

c

Fig.3. Three dimensional response surface plots representing the effect of process variable on recovery

(

A): (A) pH–volume of extraction solvent (V ); (B) pH–volume of dispersive solvent (V ); (C) volume

ext

dis

of extraction solvent–volume of dispersive solvent.

Determination of fenthion by DLLME-SFM

Tahereh Eskandari et al

93

3

.5. Statistical analysis

each 10 nm, while emission wavelength was in

the range of 300-500 nm for every nm. 15 sam-

ples were used for calibration set and ﬁve samples

not used for building the PLS calibration model

were selected as a validation test. Using the PLS

and OSC-PLS methods, the concentration of fen-

thion in the validation set were calculated. The

predicted concentrations of analyte with these

methods are shown in Table 4. In the PLS mod-

el, the number of factors was determined by the

cross-validation (leave-one-out) method com-

ponents and the predicted residual error sum of

squares (PRESS) was calculated. As shown in

Table 4, the optimum number of factors of PLS

for fenthion (N.F. = 5) was larger than the theo-

retically expected value of 1.

The suitability of the model was analyzed by ANO-

VA and the results shown in Table 3. The analy-

sis of variance of regression model demonstrated

that the model was highly signiﬁcant at probabil-

ity level (p-value is below 0.05). Also, the quality

of the ﬁtted model was studied by the coefﬁcients

of determination and adjusted the determination

2

coefﬁcient. The coefﬁcient of determination R

2

and adjusted R values were 0.9891 and 0.9695,

respectively. In other words, the model could ex-

plain 98.91% of the variability in the response. The

validation of the goodness of ﬁt was evaluated by

the lack of ﬁt test. The lack of ﬁt p-value of 0.064

implies the lack of ﬁt is not signiﬁcant and it means

that the quadratic polynomial model ﬁt the data

well.

OSC is a preprocessing technique that improves

the calibration model by removing the information

from the spectroﬂuorimetric data that unrelated

to target variables based on constrained principal

component analysis. Therefore, the spectral data

were preprocessed by OSC method. The results

(Table 4) showed that OSC preprocessing has re-

duced the number of factors (N.F. = 3).

3

.6. Analytical ﬁgures of merit

The linear range, repeatability, reproducibility and

limit of detections (LODs) for fenthion were inves-

tigated under the optimized conditions to evaluate

the proposed procedure performance. The Linearity

˗1

was obtained over the range of 5.0–110.0 ng mL

with a calibration curve (I=0.024+0.3745C) and

2

with a correlation coefﬁcient (r ) of 0.9870. The lim-

3.8. Parallel factor analysis

it of detection which is deﬁned as 3S /m (where S

The data was then arranged in a 10 × 11 × 251

three-dimensional array consisting of 11 solutions

with different fenthion concentrations in the rows

b

is the standard deviation of the blank signals for ten

replicate, and m is the slope of the calibration curve

-

-1

after extraction) was calculated to be 1.23 ng mL

(5.0-110.0 ng mL ), 200 emission wavelengths in

1

.

Precision, accuracy and stability were evaluated

the columns, and 10 excitation wavelengths in the

slices. For the evaluation of the predictive ability

of a multivariate calibration model, the root mean

square error of prediction (RMSEP) and relative

standard error of prediction (RSEP) were also

applied. The obtained results were summarized

in Table 4. The RMSEP and RSEP values with

OSC-PARAFAC were 0.37, 0.45% for fenthi-

on, respectively. These results conﬁrmed that the

OSC-PARAFAC method provided high predic-

tion ability with low RMSEP values with respect

to PLS method. Statistical parameters of the lin-

ear relationship between the proportion loadings

calculated by PARAFAC and OSC-PARAFAC are

shown in Table 5.

by repeatability (intra-day) and reproducibility (in-

ter-day) analyses. The precision of the method was

determined by analysing 5 samples on the same day

(intra-day) or 5 samples on consecutive days (in-

ter-day), and represented as RSD%. The intra-day

precision was 2.74 % and the inter–day precision

was 3.82 %. Finally, the enrichment factor (EF)

(calculated from the ratio of the slopes of the cal-

ibration curves obtained with and without pre-con-

centration) of 80.12 for Fenthion were determined.

3

.7. Partial least squares analysis

PLS model was prepared and recorded for an ex-

citation wavelength in the 200-300 nm range for

9

4

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

Table 4. Added and obtained results of the prediction set of fenthion using different methods (ng mL ).

-1

Founded fenthion (ng mL )

Added fenthion (ng mL )

PLS

OSC-PLS

PARAFAC

OSC-PARAFAC

1

3

5

7

9

5.0

15.7

36.2

56.3

74.2

92.2

5

15.4

35.8

56.1

74.5

93.6

3

14.7

34.6

55.8

75.8

94.5

2

15.1

35.5

55.3

75.5

95.4

1

Number of factor

PRESS

2.81

0.92

1.46

1.23

0.78

1.21

-

RMSEP

0.64

0.74

0.37

0.45

RSEP

Table 5. Statistical parameters of the linear relationship between the proportion loadings calculated

by PARAFAC and OSC-PARAFAC.

a

b

Parameters

PARAFAC

OSC-PARAFAC

Number of data point

Intercept

11

0.0841

0.4531

0.1241

0.2140

0.9412

0.0103

0.1381

0.3681

0.0985

0.9826

Standard deviation of intercept

Slope

Standard deviation of slope

Correlation coefﬁcient

a

PARAFAC: parallel factor analysis;

OSC-PARAFAC: orthogonal signal correction parallel factor analysis.

b

3

.9. Application of the method in synthesis and

results are summarized in Table 6. Moreover, the

OSC-PARAFAC model was better than OSC-PLS

model in terms of the determination of fenthion in

complex matrices, without considerable error. The

results demonstrated that satisfactory recovery for

fenthion could be obtained using the proposed pro-

cedures. Hence, the OSC-PARAFAC model was

able to predict the concentrations of fenthion in the

real matrix samples.

real matrix samples

In order to investigate the applicability of the opti-

mized methods for real samples, it was used to the

preconcentration and determination of the fenthion

in spiked water sample and real samples including

three water samples (tap, river and waste water).

The concentrations of fenthion were determined

by the OSC-PLS, and OSC-PARAFAC and the

Determination of fenthion by DLLME-SFM

Tahereh Eskandari et al

95

Table 6. Application of the proposed method for the determination of fenthion in real samples.

-1

Found (ng mL )

Sample

Added

OSC-

PLS

OSC-

Recovery (%)

PARAFAC

N.D.

a

-

N.D.

-

Tap water

5

0.0

52.3

104.6

50.6

101.2

-

N.D.

48.7

N.D.

-

97.4

-

N.D.

49.3

N.D.

-

98.6

-

River water

5

0.0

-

Waste water

5

0.0

51.7

103.4

48.6

97.2

a

N.D.: Not Detected.

4

. Conclusions

cides: their effects on biosentinel species and

humans, control and application in Chile, Int.

J. Morphol., 35 (2017) 1069–1074.

A simple and efﬁcient DLLME coupled with spec-

troﬂuorimetry was developed for the extraction and

determination of fenthion in water samples. The

proposed method has numerous advantages such as,

simplicity and rapidity of extraction and analysis that

reduced the organic solvent consumption within a

short time. In this study, the RSM based on the BBD

was successfully used for optimization of variable

the DLLME method that led to a saving of experi-

mental time and materials. PLS and PARAFAC mul-

tivariate calibration models, with and without OSC

pre-processing, were used for Modelling second-or-

der ﬂuorescence signals and quantiﬁcation of fenthi-

on. The predicted values obtained by application of

OSC-PARAFAC model showed the high predictive

ability compared with OSC-PLS method, which ex-

plained that the tolerance limit of three-way calibra-

tion methods for the matrix effect was better than that

of the two-way methods. Therefore, the proposed

procedure can be successfully applied for analysis

and monitoring of fenthion in water samples.

[2] K. Vala Ragnarsdottir, Environmental fate and

toxicology of organophosphate pesticides, J.

Geol. Soc., 157 (2000) 859–876.

[3] P. Sánchez Lizardi, M. K. O’Rourke, R. J.

Morris, The effects of organophosphate pes-

ticide exposure on hispanic children’s cog-

nitive and behavioral functioning, J. Pediatr.

Psychol., 33 (2007) 91–101.

[4] A. S. H. Derbalah, H. Wakatsuki, T. Yamaza-

ki, H. Sakugawa, Photodegradation kinetics

of fenitrothion in various aqueous media and

its effect on steroid hormones biosynthesis,

Geochem. J., 38 (2004) 201–213.

[5] R. A. Cheke, Effects of the organophosphate

fenthion for control of the redbilled quelea

Quelea quelea on cholinesterase and haemo-

globin concentrations in the blood of target

and non-target birds, Ecotoxicol., 21(2012)

1761–1770.

[

6] T. Galeano Díaz, A. Guiberteau Cabanillas,

M. D. López Soto, J. M. Ortiz, Determina-

tion of fenthion and fenthion-sulfoxide, in

olive oil and in river water, by square-wave

adsorptive-stripping voltammetry, Talanta, 76

(2008) 809–814.

5

. Acknowledgment

The authors are grateful to the Department of Chem-

istry, Arak Branch, Islamic Azad University, Central

Tehran Branch, IslamicAzad University and Faculty

of Chemistry, Mazandaran University, Iran

[

7] Y. Zhou, An experimental design approach

using response surface techniques to obtain

optimal liquid chromatography and mass

spectrometry conditions to determine the al-

6

. reference 37 (2013)

[1] O. Espinoza-Navarro, C. Ponce-LaRosa, E.

Bustos-Obregón, Organophosphorous pesti-

9

6

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

kaloids in Meconopsi species, J. Chromatogr.

A, 1216 (2009) 7013–7023.

[15] A. Niazi, N. Khorshidi, P. Ghaemmaghami,

“Microwave-assisted of dispersive liquid-liq-

uid microextraction and spectrophotometric

determination of uranium after optimization

based on Box-Behnken design and chemom-

etrics methods, Spectrochim. Acta - Part A:

Mol. Biomol. Spect., 135 (2015) 69–75.

[

8] S. Boulanouar, S. Mezzache, A. Combès, V.

Pichon, Molecularly imprinted polymers for

the determination of organophosphorus pesti-

cides in complex samples, Talanta, 176 (2018)

4

65–478.

[

9] T.-K. Ly, T.-D. Ho, P. Behra, Determination

of 400 pesticide residues in green tea leaves

by UPLC-MS/MS and GC-MS/MS combined

with QuEChERS extraction and mixed-mode

SPE clean-up method, Food Chemistry, 326

[16] F. Assadian, A. Niazi, Application of response

surface modeling and chemometrics methods

for the determination of oﬂoxacin in human

urine using dispersive liquid-liquid microex-

traction combined with spectroﬂuorimetry, J.

Brazilian Chem. Soc., 28 (2017) 2291–2300.

[17] A. Niazi, S. Habibi, M. Ramezani, Spectro-

photometric determination of bismuth in

water samples by dispersive liquid-liquid

microextraction after multivariate optimiza-

tion based on box-behnken, J. Chilean Chem.

Soc., 58 (2013) 1899–1901.

(2020) 126928.

[

10] F. Tokay, S. Bağdat, Solid phase extraction

and preconcentration of some metal ions using

Schiff base immobilised silica gel followed

by ICP-OES, Int. J. Environ. Anal. Chem., 00

(2019)1–12.

[

11] P. Sun, Y. L. Gao, C. Xu, Y. F. Lian, Determi-

nation of six organophosphorus pesticides in

water samples by three-dimensional graphene

aerogel-based solid-phase extraction com-

bined with gas chromatography/mass spec-

trometry, RSC Adv., 8 (2018) 10277–10283.

12] X. C. Huang, J. K. Ma, R. X. Feng, S. L. Wei,

Simultaneous determination of ﬁve organo-

phosphorus pesticide residues in different

food samples by solid-phase microextraction

ﬁbers coupled with high-performance liq-

uid chromatography, J. Sci. Food Agric., 99

[18] L. Mousavi, Z.Tamiji, MR. Khoshayand,Ap-

plications and opportunities of experimental

design for the dispersive liquid-liquid micro-

extraction method - A review, Talanta, 190

(2018) 335-356

[

[19] C. M. Monzón, C. M. Teglia, M. R. Delﬁ-

no, H. C. Goicoechea, Chemometric optimi-

zation and validation of a novel dispersive

liquid-liquid microextraction-HPLC method

for gliclazide, glibenclamide and glimepiride

quantitation in serum samples, Microchem. J.,

127 (2016) 113–119.

(2019) 6998–7007.

[

13] S. Manaﬁ Khoshmanesh, H. Hamishehkar, H.

Razmi, Trace analysis of organophosphorus

pesticide residues in fruit juices and vegeta-

bles by an electrochemically fabricated sol-

id-phase microextraction ﬁber coated with a

layer-by-layer graphenized graphite/graphene

oxide/polyaniline nanocomposite, Anal.

Methods, 12 (2020) 3268–3276.

[20] S. Veyseh, A. Niazi, Talanta A novel aera-

tion-assisted homogenous liquid – liquid mi-

croextration for determination of thorium and

uranium in water and hair samples by induc-

tively coupled plasma-mass spectroscopy, Ta-

lanta, 147 (2016) 117–123.

[21] J. I. Cacho, N. Campillo, P. Viñas, M. Hernán-

dez-Córdoba, In situ ionic liquid dispersive

liquid-liquid microextraction coupled to gas

chromatography-mass spectrometry for the

determination of organophosphorus pesti-

cides, J. Chromatogr. A, 1559 (2018) 95–101.

[22] H. Malekzadeh, M. H. Fatemi, Analysis of ﬂa-

vor volatiles of some Iranian rice cultivars by

[

14] N. S. Pano-Farias, S. G. Ceballos-Magaña, R.

Muñiz-Valencia, J. M. Jurado, Á. Alcázar, I.

A. Aguayo-Villarreal, Direct immersion sin-

gle drop micro-extraction method for multi-

class pesticides analysis in mango using GC–

MS, Food Chem., 237 (2017) 30–38.

Determination of fenthion by DLLME-SFM

Tahereh Eskandari et al

97

optimized static headspace gas chromatogra-

phy-mass spectrometry, J. Iran. Chem. Soc.,

[31] W. M. Marget, M. D. Morris, Central compos-

ite experimental designs for multiple respons-

es with different models, Technometrics, 61

(2019) 524-532.

1

2 (2015) 2245–2251.

[

23] V. Kostik, Development and validation of a

method for the simultaneous determination

of 20 organophosphorus pesticide residues in

corn by accelerated solvent extraction and gas

chromatography with nitrogen phosphorus

detection, Am. J. Appl. Chem., 2 (2014) 46.

24] A. Bidari, M. R. Ganjali, P. Norouzi, M. R.

M. Hosseini, Y. Assadi, Sample preparation

method for the analysis of some organophos-

phorus pesticides residues in tomato by ultra-

sound-assisted solvent extraction followed by

dispersive liquid-liquid microextraction, Food

Chem.,126 (2011)1840–1844.

[32] N. Khorshidi, A. Niazi, Optimization of pyro-

catechol violet biosorption by Robinia pseu-

doacacia leaf powder using response surface

methodology: Kinetic, isotherm and thermo-

dynamic studies, J. Water Reuse Desal., 6

(2016) 333–344.

[

[33] Z. Zhang, H. Zheng, Optimization for decolor-

ization of azo dye acid green 20 by ultrasound

and H2O2 using response surface methodolo-

gy, J. Hazard. Mater., 172 (2009) 1388–1393.

[34] F. B. Shahri, A. Niazi, Synthesis of modi ﬁ ed

maghemite nanoparticles and its application

for removal of Acridine Orange from aqueous

solutions by using Box-Behnken design, J.

Magn. Magn. Mater., 396 (2015) 318–326.

[35] M. H. Dehghani, Production and application

of a treated bentonite–chitosan composite

for the efﬁcient removal of humic acid from

aqueous solution, Chem. Eng. Res. Design,

140 (2018) 102–115.

[

25] J. Sherma, Review of thin-layer chromatogra-

phy in pesticide analysis: 2014–2016, J. Liq-

uid Chromatogr. Related Technol., 40 (2017)

2

26–238.

26] J. Hassan, M. Sarkouhi, Miniaturized counter

current liquid-liquid extraction for organo-

phosphorus pesticides determination, Arab. J.

Chem., 9 (2016) 38–42.

27] T. T. Hu, C. M. Lu, H. Li, Z. X. Zhang, Y.

H. Zhao, J. Li, Determination of Eleven or-

ganophosphorus pesticide residues in textiles

by using HPLC-HRMS, Anal. Sci., 33 (2017)

[36] S. G. Tuncel, T. Topal, Multifactorial optimi-

zation approach for determination of polycy-

clic aromatic hydrocarbons in sea sediments

of Turkish Mediterranean Coast, Am. J. Anal.

Chem. 2 (2011) 783–794.

1

027–1032.

[

28] N. Ferretto, M. Tedetti, C. Guigue, S. Mou-

nier, R. Redon, M. Goutx, Identiﬁcation and

quantiﬁcation of known polycyclic aromatic

hydrocarbons and pesticides in complex mix-

tures using ﬂuorescence excitation-emission

matrices and parallel factor analysis, Chemo-

sphere, 107 (2014) 344–353.

[37] R. Bhaumik, N. K. Mondal, S. Chattoraj, J.

K. Datta, E. T. Al, Application of Response

Surface Methodology for Optimization of

Fluoride Removal Mechanism by Newely

Developed Biomaterial, Am. J. Anal. Chem.,

4 (2013 404–419.

[38] M. Z. Mohamad Zulhelmi, A. Rasyidah, S.

Faraziehan, K. MohamadAnuar, Response sur-

face methodology approach for optimization of

biosorption process for removal of binary met-

als by immobilized saccharomyces cerevisiae,

Appl. Mech. Mater., 661(2014) 51–57.

[

29] J. Dong, P. Li, S. Wang, J. Wei, Y. Yang, Flu-

orometric determination of pesticides and

organophosphates using nanoceria as a phos-

phatase mimic and an inner ﬁlter effect on car-

bon nanodots, Microchim. Acta, 186 (2019)

6

6.

[39] X. Yan, H. Li, X. Wang, X. Su, A novel ﬂuo-

rescence probing strategy for the determina-

tion of parathion-methyl, Talanta, 131(2015)

88–94.

[

30] H. C. Liang, N. Bilon, M. T. Hay, Analytical

methods for pesticide residues, Water Envi-

ron. Res., 86 (2014) 2132–2155.

9

8

Anal. Methods Environ. Chem. J. 4 (2) (2021) 86-98

[

40] M. L. Nahorniak, K. S. Booksh, Optimizing

the implementation of the PARAFAC method

for near-real time calibration of excitation–

emission ﬂuorescence analysis, J. Chemom.,

1

7 (2004) 608–617.

[

41] S. Elcoroaristizabal, A. De Juan, J. Antonio,

I. Elorduy, N. Durana, L. Alonso, Chemomet-

ric determination of PAHs in aerosol samples

by ﬂuorescence spectroscopy and second-or-

der data analysis algorithms, J. Chemom., 28

(2014) 260–271.

[

42] A. Niazi, M. Sadeghi, PARAFAC and PLS

applied to spectrophotometric determination

of tetracycline in pharmaceutical formulation

and biological ﬂuids, Chem. Pharm. Bull., 54

(2006) 711–713.

43] J. Ghasemi, A. Niazi, Two- and three-way

chemometrics methods applied for spectro-

photometric determination of lorazepam in

pharmaceutical formulations and biological

ﬂuids, Anal. Chim. Acta, 533 (2005)169–177.

44] K. Kumar, A. Kumar Mishra, Parallel factor

(PARAFAC) analysis on total synchronous

ﬂuorescence spectroscopy (TSFS) data sets

in excitation-emission matrix ﬂuorescence

(

EEMF) layout: Certain practical aspects,

Chemom. Intell. Lab. Syst., 147 (2015) 121–

30.

1

[

45] J. U. Ghasemi, A. Niazi, Simultaneous deter-

mination of cobalt and nickel, comparison of

prediction ability of PCR and PLS using orig-

inal, ﬁrst and second derivative spectra, Mi-

crochem. J., 68 (2001) 1-11.

[

46]A. Niazi, M. Goodarzi, Orthogonal signal

correction–partial least squares method for

simultaneous spectrophotometric determina-

tion of cypermethrin and tetramethrin. Spec-

trochim. Acta Part A, 62 (2008)1165-1169.

[47] S. Wold, M. Sjöström, L. Eriksson, PLS-re-

gression: A basic tool of chemometrics, Che-

mom. Intell. Lab. Syst., 58 (2001) 109–130.