Journal of Chromatography A, 1429 (2016) 30–39

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Carbon nanotubes@silicon dioxide nanohybrids coating for

solid-phase microextraction of organophosphorus pesticides followed
by gas chromatography–corona discharge ion mobility spectrometric
detection
Mohammad Saraji ∗ , Mohammad Taghi Jafari, Mehdi Mossaddegh
Department of Chemistry, Isfahan University of Technology, Isfahan 84156-83111, Iran

a r t i c l e i n f o a b s t r a c t

Article history: A high efficiency solid-phase microextraction (SPME) fiber coated with porous carbon
Received 8 July 2015 nanotubes–silicon dioxide (CNTs–SiO2 ) nanohybrids was synthesized and applied for the deter-
Received in revised form 2 December 2015 mination of some organophosphorus pesticides (OPPs) in vegetables, fruits and water samples. Gas
Accepted 3 December 2015
chromatography–corona discharge ion mobility spectrometry was used as the detection system. Glucose,
Available online 9 December 2015
as a biocompatible compound, was used for connecting CNT and SiO2 during a hydrothermal process.
The electrospinning technique was also applied for the fiber preparation. The parameters affecting
Keywords:
the efficiency of extraction, including stirring rate, salt effect, extraction temperature, extraction
Carbon nanotubes@SiO2
Nanohybrids
time, desorption temperature and desorption time, were investigated and optimized. The developed
Solid-phase microextraction CNTs@SiO2 fiber presented better extraction efficiency than the commercial SPME fibers (PA, PDMS, and
Gas chromatography–ion mobility PDMS–DVB). The intra- and inter-day relative standard deviations were found to be lower than 6.2 and
spectrometry 9.0%, respectively. For water samples, the limits of detection were in the range of 0.005–0.020 ␮g L−1 and
Water the limits of quantification were between 0.010 and 0.050 ␮g L−1 . The results showed a good linearity in
Fruit and vegetable samples the range of 0.01–3.0 ␮g L−1 for the analytes. The spiking recoveries ranged from 79 (±9) to 99 (±8). The
method was successfully applied for the determination of OPPs in real samples.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction Carbon nanotubes (CNTs) were first discovered in 1991 by Iijima

Solid-phase microextraction (SPME) is a simple and solvent-
free sample preparation technique introduced by Pawliszyn and
co-workers in 1990 [1]. SPME is based on partitioning the analyte
between a phase immobilized on a SPME fiber and the matrix [1].
It has been widely applied in many researches such as pharma-
ceutical [2], biochemistry [3], food [4] and environmental samples
[5]. One of the most important research fields in SPME relates
to the synthesis and preparation of new sorbent materials as the
fiber coating. There are many review articles which have mainly
discussed the development of new SPME coatings [6–8]. Despite
introducing the huge number of SPME coatings in the past decade,
some of the coatings suffer from complicated synthesis procedures,
time-consuming processes, low thermal and mechanical stability,
short lifetime, and non-reproducible preparation conditions [9].
∗ Corresponding author. Fax: +98 31 33912350.
E-mail addresses: saraji@cc.iut.ac.ir, msaraji@ymail.com (M. Saraji).
[10]. CNT can be described as a graphite sheet rolled up into a
nanoscale tube. CNTs have been applied in many fields such as
fuel cell, tissue engineering, biosensors, hydrogen storage media
and catalyst [11]. Their remarkable properties like large surface
area (typically > 1000 m2 g−1 ), unique tubular structure, and strong
␲–␲ interaction with analytes, make them an excellent adsorbent
material for the SPME fiber coating [12].
To enhance the physical and chemical characteristics of CNTs-
based composites, nanohybrids of CNTs/metal oxides have been
used in important applications like catalysis, energy storage, water
desalination and pollution control [13]. There has also been one
report on the use of CNT–TiO2 as SPME coating [14]. Among various
metal oxides, SiO2 is a good choice to combine with CNTs, because
of its high surface area, large pore volume, structural stability, high
thermal resistance and tunable pore size [15].
Most of the methods used for the synthesis of CNTs@SiO2
nanohybrids need functionalizing CNTs with a strong acid (nitric
acid and sulfuric acid) [16]. The oxidation of CNTs with acid is
a time-consuming reflux process (24 h) leading to loosening the
mechanical strength and destroying the structure of CNTs [17]. To

http://dx.doi.org/10.1016/j.chroma.2015.12.008
0021-9673/© 2015 Elsevier B.V. All rights reserved.
 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39
avoid this tedious process, a non-covalent mild approach using ben-
zyl alcohol has been applied for the coating of TiO2 on the CNTs
[18,19]. However, the use of additional chemical reagents increases
the cost and complexity of the process and poses a toxic effect for
human and environment.
Very recently, Wang et al. [17] utilized glucose as a structure-
directing agent to form CNTs@TiO2 nanohybrids as an anode for
lithium-ion batteries. The use of glucose in a simple, green and
one-step hydrothermal process could solve the interfacial incom-
patibility between CNTs and TiO2 without any prerequisite for
covalent/non-covalent functionalization of CNTs. Glucose is an
abundant, biocompatible and inexpensive compound applied as
a carbon source in hydrothermal reactions [20]. At a tempera-
ture near 180 ◦ C, glucose is polymerized and carbonized without
any glycosidation process, thereby leading to producing hydrochar
[21,22]. This core/shell product is composed of multiple aromatic
rings (hydrophobic core), with hydroxyl groups remaining from the
original glucose structure (hydrophilic shell). It can be a connective
agent by ␲–␲ interactions of the aromatic rings with the CNTs and
chemical bonding between the hydroxyl groups and tetraethoxysi-
lane (TEOS). Actually, glucose acts as a bridge to attach intact CNTs
to the silica network.
Ion mobility spectrometry (IMS) is a fast, cost effective and
highly sensitive analytical technique for the analysis of volatile and
semi volatile compounds such as drugs (especially illicit drugs),
explosives and environmental pollutants [23]. Separation in IMS
is performed based on mass, charge, size and shape differences of
compounds. Despite these advantages, when two or more analytes
are simultaneously analyzed by IMS, the analyte interference may
have an undesirable influence on the analyte quantification [24].
Therefore, IMS is not suitable for the analysis of multi compounds
samples. In addition, due to its poor selectivity, it is not appropri-
ate for the analysis of samples with complex matrixes. Coupling
gas chromatography (GC) with IMS can solve the above mentioned
problems, and improve method capability for the analysis of real
samples [25].
In this study, for the first time, a simple and environmentally
friendly approach was developed for the synthesis of CNTs@SiO2
nanohybrids with high surface area as a SPME coating. CNTs@SiO2
coating was fabricated using pristine CNTs, tetraethoxysilane, and
glucose (as a biocompatible material) in a one-step hydrothermal
process. In this way, the use of toxic materials and organic sol-
vents were avoided. The electrospinning technique was applied to
immobilize CNTs@SiO2 nanohybrids on the surface of a stainless-
steel wire. Polyvinyl alcohol (PVA) as a green high molecular weight
polymer was used in the electrospinning process. Electrospin-
ning experimental conditions including feed solution flow rate,
applied voltage, fiber rotating speed, and the distance between the
stainless-steel needle and the fiber were studied and optimized.
Four organophosphorus (OPPs) compounds (diazinon, fenthion,
parathion and chlorpyrifos) were used as the model to investi-
gate the extraction efficiency of the fiber. The extracted compounds
were quantified using gas chromatography–corona discharge ion
mobility spectrometry (GC–CD-IMS). Various effective parameters
on the extraction efficiency, such as extraction time, extraction
temperature and ionic strength, were investigated. The analytical
performance data of the method for the determination of analytes
in real samples (pear, grape, eggplant, river water and agricultural
wastewater) was investigated too.
2. Experimental
2.1. Chemicals and materials
Diazinon and fenthion were purchased from Accustandard, Inc.
(St. Louis, USA) and Dr. Ehrenstorfer (Augs-burg, Germany), respec-
31
tively. Parathion, chlorpyrifos, PVA (Mw = 78000, 98% hydrolyzed)
and glucose were obtained from Sigma–Aldrich (St. Louis, USA).
TEOS (synthesis grade), sodium chloride (99.5%) and methanol
(HPLC grade) were supplied by Merck (Darmstadt, Germany).
Ethanol was purchased from Bidestan Co. (Qazvin, Iran). Pure water
was prepared by OES (Overseas Equipment & Services) water purifi-
cation system (OK, USA). Multi-walled CNTs (10–20 nm OD, length
range < 1–2 ␮m and purity > 95%), was purchased from Ionic Liquids
Technologies (Heilbronn, Germany). Stock standard solutions of the
analytes (1000 mg L−1 ) were prepared in methanol and stored in a
refrigerator at 4 ◦ C. A mixture of standard solution of the analytes
was prepared in methanol at 100 mg L−1 concentration level. Fresh
working standard solutions of the OPPs were prepared by appropri-
ate stepwise dilution of the standard mixture solution using water.
2.2. Instrumentation
The GC–CD-IMS used for this research has been described previ-
ously [25]. The prim zone of IMS instrument is a cell containing the
corona discharge needle, two high voltage power supplies, a pulse
generator, an analog to digital converter, and a computer. The first
section of IMS cell is ionization region (2 cm in length, 40 mm i.d.)
and the second section is drift region (11 cm in length). The ioniza-
tion region contained two separated rings, uniform hole (21 mm
i.d.) and cone shaped hole (small diameter 10 and large diame-
ter 21 mm). The Bradbury–Nielsen shutter grid with an orthogonal
field relative to the drift field, removed for a short period of time
(200 ␮s) by the pulse generator to admit an ion pulse to the drift
region. Preheated nitrogen was used as the drift gas with the flow
rate of 700 mL min−1 . Another nitrogen flow (20 mL min−1 ) as a
makeup gas was passed around the capillary column inside the nee-
dle. The IMS instrument was equipped with a Faraday plate detector
inclusive 21 mm diameter stainless-steel plate positioned ∼1.0 mm
behind the aperture grid. A homemade preamplifier with the gain
of 1 × 1010 V A−1 was used, and then the signal was further ampli-
fied up to 1000 times using a tunable-gain amplifier. Afterward,
the signal was processed using a 12-bit analog to digital interface
card (PCI-1711, Advantech, Taipei, Taiwan). Spectra were obtained
and recorded using especially developed software in LABVIEW.
Gas chromatographic analysis was carried out using a Shimadzu
(model 14A, Kyoto, Japan), fitted with a split/splitless injector and
the IMS detector. The injector was equipped with a low-volume
insert designed for the analysis by SPME (Restek, Bellefonte, PA,
USA). GC separation was achieved with a capillary column (HP-5,
30 m × 0.32 mm i.d., 0.5-␮m film thickness, Agilent Technologies,
Palo Alto, CA, USA). Nitrogen was used as the carrier gas and was
set at 1 mL min−1 . The temperatures of the injector and detector
(IMS) were set at 260 and 230 ◦ C, respectively. The GC oven was
held at the initial temperature of 70 ◦ C for 1 min followed by a linear
thermal gradient of 15 ◦ C min−1 to 250 ◦ C (held for 1 min), result-
ing in a run time of 14 min. The instrumental conditions of the IMS
for all experiments are summarized in Table S1 (Supplementary
information).
To disperse solid particles in solutions, an ultrasonic bath
(Citizen, model SW1500, Tokyo, Japan) was used at a fre-
quency of 42 kHz. The surface characteristic of the fabricated
CNTs@SiO2 coating was studied by scanning electron microscopy
(SEM) (Philips XL400 SEM, Eindhoven, The Netherlands). Thermo-
gravimetric analysis (TGA) was performed using a Rheo-metric
Scientific TGA 1500 instrument (STA, Piscataway, NJ, USA). The
TGA curve was obtained under argon atmosphere at a heating
rate of 20 ◦ C min−1 . The SPME holder and commercial fibers, poly-
acrylate (PA 85 ␮m), polydimethylsiloxane (PDMS 100 ␮m), and
polydimethylsiloxane–divinylbenzene (PDMS–DVB 65 ␮m) were
supplied by Supelco (Bellefonte, PA, USA). A home-made electro-
spinning apparatus consisting of a 4-mL polypropylene syringe
32 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39

equipped with a 25 gauge stainless-steel needle, a syringe pump

(FNM, Tehran, Iran), a high voltage power supply (FNM, model

This study
Reference
HV50POC, Tehran, Iran), and a rotating metal drum collector (Hei-

[38]
[39]

[41]
[42]
[43]
[34]
[40]
dolph, model RZR50, Mannheim, Germany) were used for the
immobilization of CNTs@SiO2 on the surface of a stainless-steel
wire (0.25 o.d.).

2.3. Preparation of CNTs@SiO2

Recovery (%)

52–117
80–105

106–120
87–89

95–98

80–98
77–89
89–99
79–98
A typical hydrothermal method for preparing CNTs@SiO2 was
used as follows: 50 mg of CNTs was added to 10 mL of ethanol/water
(90:10) solution containing 10 mg of glucose. The mixture was
ultrasonicated for 1 h. The produced black mixture was heated at
180 ◦ C for 5 h and transferred into an autoclave to perform the poly-
merization and carbonization of glucose. After that, 60 mg of the

0.1–13.7
dried solid was added to 10 mL of absolute ethanol with ultrasonic

1.6–2.4
RSD (%)

7–14

8–10
7–10
5–10
5.1–6.3
5.9–7.8
3–6
agitation for 20 min. A volume of 0.7 mL of TEOS was added to the
solution during stirring. After 150 min, the solution was centrifuged
at 4000 rpm for 10 min and the precipitate was dried at 75 ◦ C.

8.1–12.3 (␮g kg−1 )

0.01–0.2 (␮g kg−1 )

6–10 (␮g kg−1 )

2.4. SPME fiber preparation

10–15 (ng L−1 )

20–35 (ng L−1 )
1–5 (␮g kg−1 )

5–20 (ng L−1 )

20 (ng L−1 )

0.05–0.30
(␮g kg−1 )
CNTs@SiO2 fiber was prepared using electrospinning technique.
First, an aqueous solution of PVA (10% w/v) was prepared under

LOD
mild stirring at 80 ◦ C for 2 h. A solution of CNTs@SiO2 (10% w/v)
in absolute ethanol was dispersed under ultrasonic agitation for
3 h. Then, the PVA solution (cooled to the room temperature) was

Sample volume
added to the dispersed CNTs@SiO2 solution and the mixture was
ultrasonicated for 1 h. Finally, the viscoelastic solution was stirred
at room temperature for 24 h to achieve a spinnable solution. The

(mL)
solution was loaded into the 4 mL polypropylene syringe. The solu-

10
10
10
5
5
10
10
4
tion flow rate and the applied voltage were 45 ␮L min−1 and 15 kV,
respectively. The wire was spun at 100 rpm using electromagnetic
rotating motor and the distance between the stainless-steel needle
Extraction time

and the wire was 15 cm. After the electrospinning procedure, the
fiber was thermally conditioned inside the GC injection port under
the nitrogen atmosphere at 270 ◦ C for 2 h.
(min)

30
30
30
45
45
60
30
30
2.5. SPME procedure
River water, natural waters

A 4-mL standard solution of the OPPs containing 1.2 g of NaCl

Cucumber, strawberry
Comparison of CNTs@SiO2 fiber with other fibers used for the determination of OPPs.

was placed into a 5-mL glass vial. The glass vial was placed on a
magnetic stirrer (IKA, model BIG SQUID, Freiburg, Germany). The
Natural waters

temperature of the sample was set at 45 ◦ C using a water bath. Then,

Sample type

wastewater
River water

Pear, grape,

CNTs@SiO2 SPME fiber was immersed into the sample solution for
Mangoes

eggplant
Olive oil

a fixed time of 30 min under the stirring rate of 1000 rpm. After
Water,
Apple

extraction, the fiber was withdrawn into the needle and imme-
diately inserted into the GC injection port. A desorption time of
5 min at 260 ◦ C in splitless mode was applied. Prior to the first
CAR–PDMS, PDMS–DVB, PDMS

extraction, a blank analysis was carried out to ensure the absence

of contaminants.
PA, PDMS–DVB, PDMS

2.6. Real samples

River water and agricultural wastewater were collected from

CNTs@SiO2
Fiber type

PA, PDMS

Zayandeh-rood river (Isfahan, Iran) and Asgharabad village

PDMS
PDMS

PDMS

(Khomeini shahr, Iran), respectively. The samples were filtered

PA

through a 0.45 ␮m nylon membrane filter (Millipore, Bedford, MA,

USA). Pear, grape and eggplant were purchased from a local super-
market. Pesticide-free samples were obtained from the faculty of
DI–SPME–GC–CD-IMS

agricultural engineering (Isfahan University of Technology, Isfa-

HS–SPME–GC–ECD

HS–SPME–GC–FTD

HS–SPME–GC–FTD
HS–SPME–GC–MS

HS–SPME–GC–MS
DI–SPME–GC–MS

DI-SPME-GC–MS

han, Iran). The samples were washed with tap water, chopped
and homogenized using a blender. A portion of sample (2 g) was
placed in a glass centrifugation tube, spiked with the analytes and
Method
Table 1

homogenized with 4 mL of water. The sample was heated in a

water bath (45 ◦ C) for 20 min and then centrifuged at 4000 rpm for
 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39
Fig. 1. Scanning electron micrographs of (a) typical image of the prepared fiber,
(b) homogenously distributed CNTs@SiO2 nanohybrids on the surface of the porous
structure fiber, (c) SiO2 nanoparticles with uniformly decorated on the surface of
the carbon nanotubes.
10 min. The clear supernatant diluted twice with water to reduce
the matrix effect [26]. The diluted sample was extracted using the
above described procedure.
The limit of detection (LOD) and limit of quantification (LOQ)
were calculated based on the signal-to-noise ratio (S/N) of 3:1 and
10:1, respectively. The enrichment factor (EF), was defined as the
ratio of the concentration of analyte after extraction to the initial
concentration spiked in the solution [27].
33
The spiking recovery was calculated using the following equa-
tion:
(Cfound − Creal )
Spiking recovery (%) =
Cadded
here Cfound , Creal , and Cadded are the concentration of analyte after
the addition of a known amount of standard in the real sample,
the concentration of analyte in the real sample, and the concen-
tration of a known amount of standard spiked to the real sample,
respectively.
3. Results and discussion
3.1. Characterization of the CNTs@SiO2 coating
The morphology of the CNTs@SiO2 fiber was investigated by
SEM under different magnifications (Fig. 1). The SEM image of the
fiber coating as shown in Fig. 1(a), revealed that the fiber had a layer
by layer structure produced using the electrospinning technique.
Comparing the diameter of the coated and uncoated fibers showed
that the coating had a thickness of about 35 ± 2 ␮m. Fig. 1(b) shows
the homogenously distributed CNTs@SiO2 nanohybrids on the sur-
face of the porous structure fiber. The non-uniform particles seen
on the PVA network were probably macrosized silica particles. The
magnified SEM image, as shown in Fig. 1(c), revealed that the SiO2
nanoparticles had been uniformly decorated on the surface of the
carbon nanotubes. The diameter of the CNTs@SiO2 nanohybrids
was about 100 nm. Besides the original high extraction capacity
of CNTs@SiO2 nanohybrids, the porous and rough structure of the
modified nanotubes, also increased the mass adsorption capacity
and mass transfer rate.
Since the analytes were desorbed in GC injection port at 260 ◦ C,
the thermal stability of the fiber coating had to be evaluated. The
TGA curve is shown in Fig. S1 (Supplementary information). As can
be seen, the coating was thermally stable without any significant
weight losses up to 480 ◦ C. After that, a mass loss of 10% occurred
from 480 to 800 ◦ C. Therefore, it can be said that the CNTs@SiO2 has
a very high thermal stability due to the existence of CNTs and SiO2
in the structure of coating.
3.2. Optimization of SPME method
Several experimental parameters such as desorption time and
temperature, extraction time and temperature, solution ionic
strength and stirring rate could affect SPME efficiency. To achieve
high sensitivity and good efficiency using the fiber, these factors
were investigated and optimized. All experiments were performed
three times.
3.2.1. Stirring rate
Sample agitation is an important factor in the extraction pro-
cess because it increases the sample movement to fiber surface and
reduces thermodynamic equilibrium time [28]. The effect of the
stirring speed on the extraction efficiency of the OPPs compounds
was investigated from 200 to 1100 rpm. The results showed that
the peak area of the analytes was increased with the stirring rate
up to 1000 rpm. After that, the peak area almost remained constant.
Thus, 1000 rpm was selected as the optimum stirring rate.
3.2.2. Ionic strength
Generally, the presence or addition of salt to sample solution
enhances the extraction efficiency by reducing the solubility of
hydrophilic analytes, which is referred to as the salting out effect
[29]. The ionic strength effect was optimized by adding different
amounts of NaCl into the solution (Fig. S2, Supplementary infor-
mation). The extraction efficiency was enhanced with increasing
34 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39

Fig. 2. Comparison of fiber type on the extraction efficiency of OPPs (sample volume, 4 mL; concentration of analytes, 2 ␮g L−1 ; stirring rate, 1000 rpm; extraction time,
30 min; temperature, 45 ◦ C; salt addition, 0.30 g mL−1 ; desorption temperature, 260 ◦ C and desorption time, 5 min).

Table 2
Analytical results of OPPs in real samples for SPME–GC–CD-IMS.

Sample Compound Amount measured inthe real samplea Amount addeda Amount founda Relative errorb (%) Spiking recoveryc (%)
d e
River water Diazinon ND 0.10 0.09 (7) 5.0 99 (8)e
Parathion ND 0.10 0.09 (4) 3.0 97 (4)
Fenthion ND 0.10 0.09 (3) 3.0 99 (6)
Chlorpyrifos ND 0.10 0.09 (4) 4.0 96 (5)
Diazinon ND 0.75 0.72 (5) 4.0 96 (6)
Parathion ND 0.75 0.73 (6) 2.7 97 (6)
Fenthion ND 0.75 0.72 (4) 4.0 96 (5)
Chlorpyrifos ND 0.75 0.74 (4) 1.3 99 (4)

Agricultural wastewater Diazinon ND 0.10 0.09 (5) 8.0 97 (6)

Parathion ND 0.10 0.09 (7) 7.0 95 (5)
Fenthion ND 0.10 0.09 (6) 5.0 96 (7)
Chlorpyrifos ND 0.10 0.09 (6) 7.0 94 (6)
Diazinon ND 0.75 0.69 (5) 8.0 92 (5)
Parathion ND 0.75 0.67 (6) 10.7 89 (6)
Fenthion ND 0.75 0.68 (5) 9.3 90 (4)
Chlorpyrifos ND 0.75 0.70 (5) 6.6 93 (5)

Pear Diazinon 2.91 1.75 4.31 (7) 20.0 80 (5)

Parathion 0.62 1.75 2.18 (6) 10.8 89 (6)
Fenthion ND 1.75 1.54 (4) 12.0 88 (6)
Chlorpyrifos ND 1.75 1.50 (5) 14.3 86 (7)
Diazinon 2.91 4.25 6.56 (6) 14.1 86 (6)
Parathion 0.62 4.25 4.47 (8) 9.4 90 (8)
Fenthion ND 4.25 3.69 (4) 13.2 87 (6)
Chlorpyrifos ND 4.25 3.59 (5) 15.5 84 (7)

Grape Diazinon 1.04 1.75 2.45 (7) 19.4 85 (8)

Parathion 0.41 1.75 1.68 (5) 16.0 84 (8)
Fenthion ND 1.75 1.54 (3) 12.0 88 (9)
Chlorpyrifos ND 1.75 1.55 (4) 11.4 89 (6)
Diazinon 1.04 4.25 4.73 (7) 13.1 87 (9)
Parathion 0.41 4.25 4.37 (6) 2.1 98 (6)
Fenthion ND 4.25 3.61 (8) 15.0 85 (6)
Chlorpyrifos ND 4.25 3.69 (6) 13.1 87 (5)

Eggplant Diazinon 0.25 1.75 1.71 (6) 16.5 83 (7)

Parathion 0.44 1.75 1.83 (5) 20.5 79 (9)
Fenthion ND 1.75 1.41 (8) 19.4 80 (6)
Chlorpyrifos ND 1.75 1.39 (6) 20.5 79 (9)
Diazinon 0.25 4.25 4.02 (4) 11.3 89 (6)
Parathion 0.44 4.25 4.12 (9) 13.4 87 (5)
Fenthion ND 4.25 3.61 (5) 15.0 85 (8)
Chlorpyrifos ND 4.25 3.51 (7) 17.4 82 (7)
a
The concentration unit is ␮g L−1 for water and wastewater samples, and ␮g kg−1 for pear, grape and eggplant samples.
b
The relative error is the absolute error divided by the actual measurement.
c
Spiking recovery was calculated by analyzing real samples spiked with 0.1 and 0.75 ␮g L−1 for water and, 1.75 and 4.25 ␮g kg−1 for pear, grape and eggplant samples.
d
Not detected.
e
Relative standard deviation, RSD (%).
 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39 35

Fig. 3. (a) Typical chromatogram obtained after the extraction of OPPs solutions (2 ␮g L−1 ) with the CNTs@SiO2 fiber and (b) the ion mobility spectra obtained after injection
of the standard solutions of OPPs. (D) diazinon; (P) parathion; (F) fenthion; (C) chlorpyrifos.
36 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39

Fig. 4. The GC–CD-IMS chromatograms obtained for samples extracted from river water (above) and grape (bottom); (a) blank and (b) spiked sample (0.5 ␮g L−1 for river
water and 4 ␮g kg−1 for grape sample). (D) diazinon; (P) parathion; (F) fenthion; (C) chlorpyrifos.
 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39 37

Table 3
Analytical parameters of the selected OPPs with SPME–GC–CD-IMS using the CNTs@SiO2 fiber.

Sample Compound Linear rangea Determination coefficient (r2 ) Limit of quantificationa Limit of detectiona Repeatability (%)b

Intra-dayc Inter-dayc

River water Diazinon 0.025–3.0 0.996 0.025 0.010 5.8 8.5

Parathion 0.050–3.0 0.993 0.050 0.020 5.4 6.9
Fenthion 0.010–2.0 0.992 0.010 0.005 6.1 7.3
Chlorpyrifos 0.025–2.0 0.997 0.025 0.010 4.9 7.8

Agricultural wastewater Diazinon 0.025–3.0 0.994 0.025 0.010 6.3 8.9

Parathion 0.050–3.0 0.996 0.050 0.020 5.8 7.2
Fenthion 0.010–2.0 0.994 0.010 0.005 5.1 6.9
Chlorpyrifos 0.025–2.0 0.992 0.025 0.010 6.1 8.1

Pear Diazinon 0.75–20 0.990 0.75 0.30 6.4 9.3

Parathion 0.50–15 0.995 0.50 0.20 5.9 8.7
Fenthion 0.50–15 0.994 0.50 0.20 6.2 8.1
Chlorpyrifos 0.75–20 0.991 0.75 0.30 6.7 8.5

Grape Diazinon 0.50–15 0.993 0.50 0.20 5.9 8.0

Parathion 0.30–15 0.994 0.30 0.10 6.5 8.8
Fenthion 0.50–15 0.991 0.50 0.20 7.0 9.3
Chlorpyrifos 0.75–20 0.991 0.75 0.30 6.3 9.1

Eggplant Diazinon 0.20–10 0.990 0.20 0.07 7.1 9.3

Parathion 0.15–10 0.992 0.15 0.05 7.5 9.0
Fenthion 0.50–10 0.994 0.50 0.20 7.8 10.1
Chlorpyrifos 0.50–15 0.995 0.50 0.20 6.9 10.3
a
The concentration unit is ␮g L−1 for water and wastewater samples, and ␮g kg−1 for pear, grape and eggplant samples.
b
Relative standard deviation.
c
Intra-day and inter-day precision were calculated by analyzing water samples spiked at 0.5 ␮g L−1 and pear, grape and eggplant samples spiked at 3.0 ␮g kg−1 within one
day (n = 3) and over a period of three days (n = 3), respectively.

salt concentration up to 0.3 g mL−1 for diazinon, parathion and fen-
thion. A small decrease in the amount of the extracted analytes
was observed at the saturated solution. However, for chlorpyrifos a
very small decrease was found with increasing salt concentration.
Log Kow of chlorpyrifos is almost one unit higher than the rest of
the analytes implying that chlorpyrifos is more hydrophobic than
diazinon, parathion and fenthion. Therefore, the ionic strength of
the solution had no significant effect on the extraction efficiency
of chlorpyrifos. Based on the results, sample solution containing
0.3 g mL−1 was used for the further experiments.
3.2.3. Extraction temperature
The extraction efficiency is enhanced at higher temperatures,
because the value of analyte diffusion coefficient is increased, thus
leading to the decrease in the equilibrium time. On the other
hand, increasing extraction temperature decreases the partition
coefficient between the fiber coating and sample matrix, thereby
decreasing the adsorption capacity of the coating [30]. The temper-
ature influence on the SPME extraction was studied at temperatures
ranging from 25 to 65 ◦ C. The results (Fig. S3, Supplementary infor-
mation) showed that the maximum peak area for all analytes except
parathion was obtained at 45 ◦ C. The parathion had a maximum
response at 35 ◦ C. Based on the results, 45 ◦ C was chosen as the
extraction temperature for the subsequent experiments.
3.2.5. Desorption condition
The effect of the desorption temperature on the response of the
target compounds was investigated at temperatures ranging from
200 to 280 ◦ C while the desorption time was kept constant (5 min).
The peak area for all analytes was increased as the desorption tem-
perature was increased up to 260 ◦ C. The analytes were completely
desorbed at 260 ◦ C without observing any memory effect. To find
out the minimum time required for the complete desorption of
the analyte from the fiber, the desorption time was investigated
in the range of 1–6 min. The maximum response was observed at
5 min desorption time. Therefore, desorption at 260 ◦ C for 5 min
was chosen as the optimum desorption conditions.
3.3. Fiber durability
An important parameter considered for practical application of
the fiber coating is the lifetime of the fiber. Organic solvent, high
temperature, strong acidic or basic solution can harm the fiber
coating [32]. In this work, after 120 extraction–desorption cycles,
no significant decline on the extraction efficiency of the fiber was
observed. It demonstrated that the new fiber had good mechanical
and thermal stability, as compared to the commercial fibers.
3.4. Comparison of CNTs@SiO2 fiber with other fibers and SPME
methods

3.2.4. Extraction time
The extraction time is another important parameter that influ-
ences the extraction efficiency. SPME technique is a non-exhaustive
approach employed on the basis of equilibrium principles. The
amount of extracted analytes is increased up to equilibrium time,
according to Nernst’s partition Law [31]. The extraction time was
investigated between 10 and 40 min. The results (Fig. S4, Supple-
mentary information) showed that equilibrium was obtained after
30 min and after that, no significant changes in the extraction effi-
ciency occurred. Therefore, 30 min was selected as the extraction
time for the subsequent experiments.
To evaluate the extraction performance of the fiber, the EF val-
ues were calculated. The EFs obtained by CNTs@SiO2 fiber was
compared with three commercial SPME fibers (PA, PDMS and
PDMS/DVB) and CNTs fiber, used for the extraction of OPPs. Fig. 2
shows the EF of CNTs@SiO2 , the three commercial and CNTs fibers.
The commercial fibers were used under previously reported extrac-
tion conditions [33]. Among the commercial fibers, PA fiber showed
the most extraction performance. This was in accordance with the
previous researches on the extraction of OPPs by SPME technique
[34,35]. As shown in Fig. 2, despite the large thickness of PA coating
(85 ␮m), as compared to CNTs@SiO2 fiber (35 ± 2 ␮m), the extrac-
38 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39
tion efficiency of CNTs@SiO2 fiber was twice more than that of PA
fiber. Because of the porous structure of coating, the surface area
of CNTs@SiO2 was very large, as compared to PA. In addition, ␲–␲
interactions between the CNTs and OPPs enhanced analyte adsorp-
tion on CNTs@SiO2 . Also, Fig. 2 shows that the presence of SiO2 in
the structure of coating led to the better extraction efficiency of
CNTs@SiO2 , as compared with CNTs coating.
Table 1 shows some analytical characteristics of the present
method, such as LOD, relative standard deviations (RSD) and recov-
ery values, as compared with the commercial SPME fibers used
for the determination of OPPs. As can be seen in Table 1, the
present method had lower LODs, and RSDs comparable to the other
reported methods.
The commercial fibers used a fragile fused silica support,
whereas the electrospinning technique provided the possibility
of using unbreakable and mechanically stable solid supports like
stainless-steel wire.
3.5. Method validation
It was necessary to validate the performance characteristics of
the present method to confirm the method capability for the anal-
ysis of OPPs. The analytical figures of merit of the method such as
LOD, LOQ, dynamic range and precision were evaluated under the
optimized extraction conditions (Table S2, Supplementary infor-
mation).
Fig. 3(a) shows typical GC–IMS chromatogram obtained after
the extraction of a water sample spiked with the OPPs at 2 ␮g L−1 .
The ion mobility spectra of the analytes can be observed in Fig. 3(b).
Drift times of product ions between 11.0 and 13.8 ms were selected
for the quantification of diazinon and parathion using GC–IMS
chromatogram (Fig. 3b). However, due to overlapping fenthion
and chlorpyrifos peaks in the chromatogram (Fig. 3a), single ion
drift times were used only by IMS. Fenthion and chlorpyrifos were
separated in second dimensional (IMS detection system), with
respect to their different mobilities. Single ion drift times in the
range of 12.0–13.4 and 13.4–15 ms were chosen to obtain the
chromatogram containing only IMS response to fenthion and chlor-
pyrifos, respectively (Fig. S5, Supplementary information).
To find the dynamic range of the method, pure water sam-
ples spiked with the OPPs at different concentrations from 0.01
to 3.0 ␮g L−1 , were investigated. The method linearity for diazi-
non, parathion, fenthion and chlorpyrifos was in the range of
0.025–3.0, 0.050–3.0, 0.010–2.0 and 0.025–2.0 ␮g L−1 , respectively.
The results (Table S2, Supplementary information) showed a good
linearity relationship (r2 > 0.990) between the corresponding peak
areas and the concentration levels for all the analytes. The LODs
were in the range of 0.005–0.020 ␮g L−1 for all OPPs.
In order to estimate the intra-day and inter-day repeatability
values, three replicate determinations (spiked at 0.5 ␮g L−1 for each
target analyte) were carried out using a single fiber within one
day (n = 3) and over a period of three days (n = 3), respectively. The
RSDs for intra-day and inter-day were in the range of 4.3–6.1 and
6.5–8.9%, respectively. Also, the fiber-to-fiber reproducibility for
three different fibers produced under the same conditions ranged
from 8.4% to 12.6%.
3.6. Real sample analysis
Pear, grape, eggplant, river water and agricultural wastewater
were collected and analyzed to confirm the reliability and appli-
cability of the method in real sample analysis. The analytes were
not detected in the river water and wastewater samples, while
diazinon and parathion were detected in pear, grape and egg-
plant (Table 2). The multiple standard addition technique was
used for the determination of the analytes in these samples. The
amount of diazinon found in pear, grape and eggplant was 2.91, 1.04
and 0.25 ␮g kg−1 , respectively. Concentration of parathion in pear,
grape and eggplant was also found to be 0.62, 0.41 and 0.44 ␮g kg−1 ,
respectively. In order to study the method trueness, the spiking
recovery was calculated. For this purpose, water and wastewater
samples were spiked at two different concentration levels (0.10 and
0.75 ␮g L−1 ). Also, eggplant and fruit samples were spiked at 1.75
and 4.25 ␮g kg−1 . The quantitative results are listed in Table 2. The
results clearly demonstrated that the method had good trueness
without significant matrix effects. Therefore, the analysis of four
target OPPs was practically applicable in real samples. Fig. 4 shows
the chromatograms of spiked and unspiked river water and grape
samples. The river water and grape samples were spiked with the
analytes at 0.5 ␮g L−1 and 4 ␮g kg−1 , respectively.
The method figures of merits including LOD, LOQ, linear range
and precision were calculated using spiked real samples. For
pear, grape and eggplant, pesticides-free samples were used.
The results are shown in Table 3. The LODs were in the range
of 0.005–0.020 ␮g L−1 for water and wastewater samples, and
0.05–0.30 ␮g kg−1 for pear, grape and eggplant. The LOQ values
were in the range of 0.010–0.050 ␮g L−1 and 0.15–0.75 ␮g kg −1
for water and solid samples, respectively. The European Union
has set a limit value of the individual pesticides in drinking water
at 0.1 ␮g L−1 [36]. Also, the European Commission has reported a
maximum residue limit of about 10–500 ␮g kg−1 for the selected
organophosphorus in fruits and vegetables [37]. Considering the
acceptable concentration of the analytes in real samples, the
present method could be suitable for the analysis of the selected
organophosphorus pesticides in water, fruit and vegetable sam-
ples. The precision of the method expressed as intra- and inter-day
RSD were in the range of 5.1–7.8% and 6.9–10.3%, respectively.
4. Conclusions
In this work, for the first time, CNTs@SiO2 was synthesized
as a SPME coating based on a hydrothermal reaction using glu-
cose for CNTs activation. The traditional multistep time-consuming
procedure and the use of toxic reagents were avoided. Electro-
spinning, as a simple and reproducible technique, was applied for
the fabrication of CNTs@SiO2 SPME fiber. The fiber coated on a
stainless-steel wire had excellent thermal and mechanical stabil-
ity. The method was used for the extraction of four OPPs from pear,
grape, eggplant, river water and agricultural wastewater. Separa-
tion and determination of the extracted analytes were carried out
using gas chromatography corona discharge–ion mobility spec-
trometry as a hyphenated instrument. The fiber was durable for
at least 120 extraction–desorption cycles. The high specific sur-
face area, due to the porous structure coating of the fiber, and
␲–␲ interaction between the CNTs and OPPs, provided the excel-
lent extraction performance. Also, the extraction efficiency of the
CNTs@SiO2 fiber was better than that of the commercial fibers (PA,
PDMS and PDMS/DVB).
Acknowledgments
The authors would like to acknowledge the Research Council of
Isfahan University of Technology (IUT) and Center of Excellence in
Sensor and Green Chemistry for financial support of this work.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.chroma.2015.12.
008.
 M. Saraji et al. / J. Chromatogr. A 1429 (2016) 30–39
References
[1] C.L. Arthur, J. Pawliszyn, Solid phase microextraction with thermal desorption
using fused silica optical fibers, Anal. Chem. 62 (1990) 2145–2148.
[2] H. Peltenburg, F.A. Groothuis, S.T.J. Droge, I.J. Bosman, J.L.M. Hermens,
Elucidating the sorption mechanism of mixed-mode SPME using the basic
drug amphetamine as a model compound, Anal. Chim. Acta 782 (2013) 21–27.
[3] D. Poli, M. Goldoni, M. Corradi, O. Acampa, P. Carbognani, E. Internullo, A.
Casalini, A. Mutti, Determination of aldehydes in exhaled breath of patients
with lung cancer by means of on-fiber-derivatisation SPME-GC/MS, J.
Chromatogr. B 878 (2010) 2643–2651.
[4] C. Murat, K. Gourrat, H. Jerosch, N. Cayot, Analytical comparison and sensory
representativity of SAFE, SPME, and purge and trap extracts of volatile
compounds from pea flour, Food Chem. 135 (2012) 913–920.
[5] A. Sarafraz-Yazdi, M. Mosadegh, A. Amiri, Determination of volatile organic
compounds in environmental water samples using three solid-phase
microextraction fibers based on sol–gel technique with gas
chromatography–flame ionization detector, Anal. Methods 3 (2011)
1877–1886.
[6] C. Dietz, J. Sanz, C. Cámara, Recent developments in solid-phase
microextraction coatings and related techniques, J. Chromatogr. A 1103
(2006) 183–192.
[7] A. Spietelun, M. Pilarczyk, A. Kloskowski, J. Namieśnik, Current trends in
solid-phase microextraction (SPME) fibre coatings, Chem. Soc. Rev. 39 (2010)
4524–4537.
[8] A. Mehdinia, M.O. Aziz-Zanjani, Recent advances in nanomaterials utilized in
fiber coatings for solid-phase microextraction, Trends Anal. Chem. 42 (2013)
205–215.
[9] M. Saraji, B. Farajmand, Microporous silica with nanolayer structure coated
with renewable organic solvent film as a novel extracting phase: a
combination of solid- and liquid-phase microextraction, Anal. Chim. Acta 721
(2012) 61–67.
[10] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58.
[11] M. Endo, M.S. Strano, P.M. Ajayan, Potential applications of carbon nanotubes,
Top. Appl. Phys. 111 (2008) 13–62.
[12] F. Ghaemi, A. Amiri, R. Yunus, Methods for coating solid-phase
microextraction fibers with carbon nanotubes, Trends Anal. Chem. 59 (2014)
133–143.
[13] H. Chu, L. Wei, R. Cui, J. Wang, Y. Li, Carbon nanotubes combined with
inorganic nanomaterials: Preparations and applications, Coord. Chem. Rev.
254 (2010) 1117–1134.
[14] M. Sun, J. Feng, H. Qiu, L. Fan, X. Li, C. Luo, CNT–TiO2 coating bonded onto
stainless steel wire as a novel solid-phase microextraction fiber, Talanta 114
(2013) 60–65.
[15] Y. Fan, Y.Q. Feng, Z.G. Shi, J.B. Wang, Ordered mesoporous silica coated
capillary for in-tube solid phase microextraction coupled to high performance
liquid chromatography, Anal. Chim. Acta 543 (2005) 1–8.
[16] D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato, Chemistry of carbon nanotubes,
Chem. Rev. 106 (2006) 1105–1136.
[17] B. Wang, H. Xin, X. Li, J. Cheng, G. Yang, F. Nie, Mesoporous CNT@TiO2 -C
nanocable with extremely durable high rate capability for lithium-ion battery
anodes, Sci. Rep. 4 (2014) 3729–3736.
[18] D. Eder, A.H. Windle, Carbon–inorganic hybrid materials: the
carbon-nanotube/TiO2 , Adv. Mater. 20 (2008) 1787–1793.
[19] D. Eder, A.H. Windle, Morphology control of CNT–TiO2 hybrid materials and
rutile nanotubes, J. Mater. Chem. 18 (2008) 2036–2204.
[20] X. Qi, Y. Lian, L. Yan, R.L. Smith Jr., One-step preparation of carbonaceous solid
acid catalysts by hydrothermal carbonization of glucose for cellulose
hydrolysis, Catal. Commun. 57 (2014) 50–54.
[21] G.C.A. Luijkx, F. van Rantwijk, H. van Bekkum, M.J. Antal Jr., The role of
deoxyhexonic acids in the hydrothermal decarboxylation of carbohydrates,
Carbonhydrate Res. 272 (1995) 191–202.
[22] Q. Wang, H. Li, L. Chen, X. Huang, Novel spherical microporous carbon as
anode material for Li-ion batteries, Solid State Ionics 152 (2002) 43–50.
[23] S. Armenta, M. Alcala, M. Blanco, A review of recent, unconventional
applications of ion mobility spectrometry (IMS), Anal. Chim. Acta 703 (2011)
114–123.
[24] M.T. Jafari, M. Saraji, S. Yousefi, Negative electrospray ionization ion mobility
spectrometry combined with microextraction in packed syringe for direct
analysis of phenoxyacid herbicides in environmental waters, J. Chromatogr. A
1249 (2012) 41–47.
39
[25] M.T. Jafari, M. Saraji, H. Sherafatmand, Design for gas chromatography-corona
discharge-ion mobility spectrometry, Anal. Chem. 84 (2012) 10077–10084.
[26] M. Saraji, B. Rezaei, M.K. Boroujeni, A.A.H. Bidgoli, Polypyrrole/sol–gel
composite as a solid-phase microextraction fiber coating for the
determination of organophosphorus pesticides in water and vegeTable
samples, J. Chromatogr. A 1279 (2013) 20–26.
[27] Z. Huang, P.E. Chua, H.K. Lee, Carbonized polydopamine as coating for
solid-phase microextraction of organochlorine pesticides, J. Chromatogr. A
1399 (2015) 8–17.
[28] W.A. Wan Ibrahim, H. Farhani, M.M. Sanagi, H.Y. Aboul-Enein, Solid-phase
microextraction using new sol–gel hybrid
polydimethylsiloxane-2-hydroxymethyl-18-crown-6-coated fiber for
determination of organophosphorous pesticides, J. Chromatogr. A 1217
(2010) 4890–4897.
[29] J. Feng, M. Sun, H. Liu, J. Li, X. Liu, S. Jiang, Au nanoparticles as a novel coating
for solid-phase microextraction, J. Chromatogr. A 1217 (2010) 8079–8086.
[30] M.H. Banitaba, S.S.H. Davarani, S.K. Movahed, Comparison of direct,
headspace and headspace cold fiber modes in solid phase microextraction of
polycyclic aromatic hydrocarbons by a new coating based on
poly(3,4-ethylenedioxythiophene)/graphene oxide composite, J. Chromatogr.
A 1325 (2014) 23–30.
[31] A. Kumar, Gaurav, A.K. Malik, D.K. Tewary, B. Singh, A review on development
of solid phase microextraction fibers by sol–gel methods and their
applications, Anal. Chim. Acta 610 (2008) 1–14.
[32] M.F. Alpendurada, Solid-phase microextraction: a promising technique for
sample preparation in environmental analysis, J. Chromatogr. A 889 (2000)
3–14.
[33] M.T. Jafari, M. Saraji, H. Sherafatmand, Polypyrrole/montmorillonite
nanocomposite as a new solid phase microextraction fiber combined with gas
chromatography–corona discharge ion mobility spectrometry for the
simultaneous determination of diazinon and fenthion organophosphorus
pesticides, Anal. Chim. Acta 814 (2014) 69–78.
[34] A. Menezes Filho, F.N. dos Santos, P.A.P. Pereira, Development, validation and
application of a method based on DI-SPME and GC–MS for determination of
pesticides of different chemical groups in surface and groundwater samples,
Microchem. J. 96 (2010) 139–145.
[35] P.G. Su, S.D. Huang, Determination of organophosphorus pesticides in water
by solid-phase microextraction, Talanta 49 (1999) 393–402.
[36] European Commission, Directive 98/83/EC of the Council of 3 of November
1998 on the quality of water intended for human consumption, Off. J. Eur.
Communities L330 (1998) 32 (05.12.98).
[37] European Commission, Council Directive 91/414/EEC of the Council of 23
February 2005 on maximum residue levels of pesticides in food and feed of
plant and animal origin and amending, Off. J. Eur. Communities L70/1 (2005)
(16.3.05).
[38] L.B. Abdulrauf, G.H. Tan, Multivariate study of parameters in the
determination of pesticide residues in apple by headspace solid phase
microextraction coupled to gas chromatography–mass spectrometry using
experimental factorial design, Food Chem. 141 (2013) 4344–4348.
[39] C.M. Kin, T.G. Huat, Headspace solid-phase microextraction for the evaluation
of pesticide residue contents in cucumber and strawberry after washing
treatment, Food Chem. 123 (2010) 760–764.
[40] A. Menezes Filho, F.N. dos Santos, P.A.P. Pereira, Development, validation and
application of a methodology based on solid-phase micro extraction followed
by gas chromatography coupled to mass spectrometry (SPME/GC–MS) for the
determination of pesticide residues in mangoes, Talanta 81 (2010) 346–354.
[41] D.A. Lambropoulou, V.A. Sakkas, D.G. Hela, T.A. Albanis, Application of
solid-phase microextraction in the monitoring of priority pesticides in the
Kalamas river, J. Chromatogr. A 963 (2002) 107–116.
[42] D.A. Lambropoulou, T.A. Albanis, Optimization of headspace solid-phase
microextraction conditions for the determination of organophosphorus
insecticides in natural waters, J. Chromatogr. A 922 (2001) 243–255.
[43] C. Tsoutsi, I. Konstantinou, D. Hela, T.A. Albanis, Screening method for
organophosphorus insecticides and their metabolites in olive oil samples
based on headspace solid-phase microextraction coupled with gas
chromatography, Anal. Chim. Acta 573–574 (2006) 216–222.