Magnetic poly(β-cyclodextrin-ionic liquid) nanocomposites for micro-solid phase extraction of selected polycyclic aromatic hydrocarbons in rice samples prior to GC-FID analysis

Article abstract of DOI:10.1016/j.foodchem.2018.10.145

Poly(β-cyclodextrin functionalized ionic liquid) immobilized magnetic nanoparticles (Fe₃O_4@βCD-Vinyl-TDI) as sorbent in magnetic μ-SPE was developed for the determination of selected polycyclic aromatic hydrocarbons (PAHs) in rice samples coupled with gas chromatographic-flame ionization detector (GC-FID). The nanocomposite was characterized by various tools and significant parameters that affected the extraction efficiency of PAHs were investigated. The calibration curves were linear for the concentration ranging between 0.1 and 500 μg kg^?1 with correlation determinations (R²) from 0.9970 to 0.9982 for all analytes. Detection limits ranged at 0.01–0.18 μg kg^?1 in real matrix. The RSD values ranged at 2.95%–5.34% (intra-day) and 4.37%–7.05% (inter-day) precision for six varied days. The sorbents showed satisfactory reproducibility in 2.9% to 9.9% range and acceptable recovery values at 80.4%–112.4% were obtained for the real sample analysis. The optimized method was successfully applied to access content safety of selected PAHs for 24 kinds of commercial rice available in Malaysia.

Full text of DOI:10.1016/j.foodchem.2018.10.145

Accepted Manuscript
Magnetic poly(β-cyclodextrin-ionic liquid) nanocomposites for micro-solid
phase extraction of selected polycyclic aromatic hydrocarbons in rice samples
prior to GC-FID analysis
Boon Yih Hui, Nur Nadhirah Mohamad Zain, Sharifah Mohamad, Hasnah
Osman, Muggundha Raoov
PII:
S0308-8146(18)31928-9
DOI:
https://doi.org/10.1016/j.foodchem.2018.10.145
FOCH 23806
Reference:
To appear in:
Food Chemistry
Received Date:
Revised Date:
Accepted Date:
14 August 2018
30 October 2018
31 October 2018
Please cite this article as: Hui, B.Y., Mohamad Zain, N.N., Mohamad, S., Osman, H., Raoov, M., Magnetic
poly(β-cyclodextrin-ionic liquid) nanocomposites for micro-solid phase extraction of selected polycyclic aromatic
hydrocarbons in rice samples prior to GC-FID analysis, Food Chemistry (2018), doi: https://doi.org/10.1016/
j.foodchem.2018.10.145
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Magnetic poly(β-cyclodextrin-ionic liquid) nanocomposites for micro-solid phase
extraction of selected polycyclic aromatic hydrocarbons in rice samples prior to
GC-FID analysis
a
a
b,c
d
Boon Yih Hui , Nur Nadhirah Mohamad Zain , Sharifah Mohamad , Hasnah Osman ,
Muggundha Raoovb,c*
,
a
Integrative Medicine Cluster, Advanced Medical & Dental Institute, Universiti Sains
Malaysia, Pulau Pinang 13200, Malaysia
Email: boonyihhui@hotmail.com; nurnadhirah@usm.my
b
Department of Chemistry, Faculty of Science,Universiti Malaya, Kuala Lumpur
5
0603, Malaysia
c
Universiti Malaya Centre for Ionic Liquids (UMCiL), Department of Chemistry,
Faculty of Science, Universiti Malaya, Kuala Lumpur 50603, Malaysia
Email: sharifahm@um.edu.my; muggundha@um.edu.my
d
School of Chemical Sciences, Universiti Sains Malaysia, 11800, Pulau Pinang,
Malaysia
Email: ohasnah@usm.my
*
Corresponding Author. E-mail: muggundha@um.edu.my (M. Raoov)
Phone: +6003-79677022 (ext. 2544)
Postal Address: Department of Chemistry, Faculty of
Science, University of Malaya, 50603 Kuala Lumpur,
Malaysia.
1

Graphical Abstract
2

Abstract
Poly(β-cyclodextrin functionalized ionic liquid) immobilized magnetic nanoparticles
(
Fe O βCD-Vinyl-TDI) as sorbent in magnetic µ-SPE was developed for the
3 4@
determination of selected polycyclic aromatic hydrocarbons (PAHs) in rice samples
coupled with gas chromatographic-flame ionization detector (GC-FID). The
nanocomposite was characterized by various tools and significant parameters that
affected the extraction efficiency of PAHs were investigated. The calibration curves
-1
were linear for the concentration ranging between 0.1 and 500 μg kg with correlation
2
determinations (R ) from 0.9970 to 0.9982 for all analytes. Detection limits ranged at
-1
0
.01-0.18 μg kg in real matrix. The RSD values ranged at 2.95%-5.34% (intra-day)
and 4.37%-7.05% (inter-day) precision for six varied days. The sorbents showed
satisfactory reproducibility in 2.9% to 9.9% range and acceptable recovery values at
8
0.4% -112.4% were obtained for the real sample analysis. The optimized method was
successfully applied to access content safety of selected PAHs for 24 kinds of
commercial rice available in Malaysia.
Keywords:
Magnetic poly(β-cyclodextrin-ionic liquid) nanocomposites, Magnetic micro-solid
phase extraction, Polycyclic aromatic hydrocarbons, Rice samples
1
.
Introduction
Polycyclic aromatic hydrocarbons (PAHs) refer to a group of environmental
contaminants with significant mutagenic and carcinogenic properties (Sarafraz-Yazdi,
Ghaemi, & Amiri, 2012). Although many PAHs have been listed in the European
Community (EC) and in the Environmental Protection Agency (EPA) as priority
3

pollutants, several uncertainties still lurk over the significance of PAHs pertaining to
their transfer from the natural environment into the human food chain. In fact, the
existence of PAHs in environmental samples, such as air particulates, sediments, and
water has been investigated substantially, leaving out food samples, particularly rice
grain that is considered as primary source in Asian diets.
Due to the toxicity properties of these compounds toward human health World Health
Organization (WHO) had regulated the levels of listed PAHs. (Registry, 2008)
However, not all the PAHs enter food legislation especially those compounds with low
molecular weight Based on Commission Regulation (EU) No 835/2011 as regards
maximum levels for PAHs in foodstuffs, 4 heavy molecular PAHs namely PAH4
(benzo[a]pyrene, benz[a]anthracene, benzo[b]fluoranthene and chrysene) had been
recommended as suitable markers in food analysis with their maximum levels at 1 to
-1
3
0 ug kg (Food Safety Authority of Ireland, 2015). Nevertheless, the needs for reliable
data on the occurrence of others PAHs in food still increasing since they are not covered
in any existing PAHs regulatory limit as reported by European Food Safety Authority
(EFSA) (Simon, Ruiz, Von Holst, Wenzl, & Anklam, 2008).
Although various analytical methodologies have been developed in different food
matrices, few studies that monitored the PAHs in rice grain have been published
(Escarrone et al., 2014; Xiaoxing Liu & Korenaga, 2001; Matin, Biparva, & Gheshlaghi,
2
014). For this reason, the development of an accurate and reliable analytical method
for the analysis of multiple PAHs in order to get a clearer picture of their levels in rice
grains is of prime importance.
4

From the view of analytical chemistry, extraction of PAHs from different matrices such
as environmental and food samples prior to instrument analysis requires complex
procedures. Although traditional sample preparation techniques such as liquid-liquid
extraction (LLE) and solid phase extraction (SPE) have been widely applied for the
analysis, however, they involved multiple-steps that consume large amounts of solvent
lead to organic waste and take up to much time which may cause analytes loss, as well
as blockage of column bed during the extraction process (Mukhtar & See, 2016).
Therefore, at recent times, researchers have begun developing novel sorbents with high
extraction efficiency and anti-interference ability towards an exciting direction in food
and environmental analysis.
Numerous formats of new adsorbent-based extraction techniques have been introduced,
for example stir-bar sorptive extraction (SBSE) (Camino-Sánchez, Rodríguez-Gómez,
Zafra-Gómez, Santos-Fandila, & Vílchez, 2014), magnetic solid phase extraction
(MSPE) (Wan Ibrahim, Nodeh, Aboul-Enein, & Sanagi, 2015), solid phase
microextraction (SPME) (Biziuk, 2006), stir-cake sorptive extraction (SCSE) (Huang,
Chen, & Yuan, 2013), and fabric phase extraction (FPE) (Aznar, Úbeda, Nerin, Kabir,
&
Furton, 2017).Among these, MSPE seems to be the most popular analytical approach
in sample preparation based on the use of magnetic sorbent. This method does not
require the sorbent to be packed into the cartridge and the phase separation can be easily
isolated by applying external magnetic field (Ebrahimpour, Yamini, Seidi, & Tajik,
2
015) without additional filtration or centrifugation of the sample, hence easing both
collection and sampling. In MSPE, the contact area between adsorbents and analytes is
large since the adsorbent can be uniformly dispersed into the sample solution, thus,
rapid the mass transfer. This leads to shorter extraction time with low toxicity profile.
5

Some nanomaterials varieties that are frequently used as absorbents in MSPE are
magnetic nanoparticles (MNPs) based on their different functional groups that suit the
varied analytes. A good modified magnetic material should be able to recover the high
extraction capability of Fe O nanoparticles, apart from displaying high reusability and
3
4
good stability toward acidic and alkaline solutions, as well as organic solvent (Mehdinia,
Khodaee, & Jabbari, 2015).
Cylodextrins, (CDs) belong to the series of macro cyclic oligosaccharides which are
composed of hydrophobic inner cavity and hydrophilic exterior (Boon, Raoov, Zain,
Mohamad, & Osman, 2017). Since it can encapsulate varied organic guest molecules
into its cavity to form a stable inclusion complexes without forming any chemical bond,
therefore, CDs allow their removal from aqueous solution when organic pollutants are
bound into their cavities (Mahlambi, Malefetse, Mamba, & Krause, 2010). Nonetheless,
parental CDs are water-soluble and is readily being modified through polymeric solid
supports or bifunctional cross-linking reagent (e.g. toluene-2,4-diisocyanate (TDI),
epichlorohydrin (EP), 1,6-hexamethylene diisocyanate (HMDI), and succinyl chloride
(SC) to enhance its water insolubility and stability properties (W. Zhang et al., 2017).
This modification not only enable us to ascertain the roles of the cyclodextrin cavities,
but also the effect of crosslinking networks in the sorption mechanism.
On the other hand, CDs are also functionalized to enhance both their physical and
chemical properties, so as to suit with various types of applications. In this case, ionic
liquids (ILs) is an example of excellent functional group that has been studied
extensively mainly due to their fascinating properties, such as non-volatility, non-
flammability, and high thermal stability, which allow fine-tuning of ILs for various
applications. As such, ionic liquid functionalized CDs with cross-linked polymer, as a
supramolecular material, does not only retain the hydrophobic cavity of CDs, but the
6

modified groups can also enhance their functionality (Chen, Qin, Gu, & Zhu, 2016).
Hence, this study proposes that CD based functional composite may be an effective
sorbent that extract hydrophobic PAHs due to its hydrophobic cavity and the interaction
between the analytes with functional group of IL and the polymer cross-linker.
To the best of author’s knowledge, studies regarding the preparation of magnetite CD-
based functionalized nanocomposite are in scarcity, particularly those that determine
PAHs in food samples. By considering the cooperating effect of the as mentioned
materials, herein, this study reports the preparation and the characterization of
Fe O βCD-Vinyl-TDI, along with its applications in MSPE. In order to assess the
3
4@
feasibility of the proposed method, five representative light molecular weight PAHs
compounds (listed in 16 EPA PAHs) had been selected as the model compounds,
including Acenapthalene (Ace), Fluorene (Flu), Phenanthrene (Phe), Fluoranthene
(Fla), and Pyrene (Pyr), which are frequently found in environmental monitoring
samples.
With that, the integration of Fe O βCD-Vinyl-TDI-MSPE with gas chromatography
3
4@
flame ionization detector (GC-FID), should be able to deliver a highly reliable and
effective Fe O βCD-Vinyl-TDI-MSPE-GC-FID analytical method.
3
4@
2
.
Experimental
2
.1.
Chemicals and reagents
All PAHs standards (Ace), (Flu), (Phe), (Fla), and (Pyr) were purchased from Dr.
Ehrenstorfer GmbH (Augsburg, Germany). Toluene-2,4-diisocyanate (TDI) and 1-
vinylimidazole (99%) were obtained from Sigma Aldrich (St. Louis, MO, USA), while
iron (II) chloride tetrahydrate (FeCl .4H O) and iron (III) chloride hexahydrate
2
2
(
FeCl .6H O) were purchased from R&M Chemicals (Essex, UK). Acetonitrile,
3 2
7

methanol (HPLC grade, 99.7%), acetone technical grade, and ammonia solution (25%)
were supplied from Friendemann Schmidt (Parkwood, Australia). Ethyl acetate, hexane,
anhydrous N,N-dimethylformamide (DMF) and dichloromethane, were purchased
from Merck (Darmstadt, Germany). β-cyclodextrin (βCD, 99%) was commercially
available and purchased from Acros (Hungary). Ultrapure water was prepared by using
the Sartorius Milli-Q system (Aubagne, France). All other reagents were of analytical
grade and used without further purification.
-1
Acetonitrile was used to prepare stock solutions (1000 mg L of each analyte). The
working standard solutions were freshly prepared by dilution of stock solutions with
ultrapure water to desired concentration. All the standard solutions were stored in
refrigerator (4℃) to avoid degradation.
2
.2.
Instruments
Several imminent aspects of the prepared sorbents, such as size, morphology,
composition, and magnetic property were studied by using a scanning electron
microscope (SEM, Quanta FEG 650), a transmission electron microscope, (TEM,
Phillips, CM12 Version 3.2), and a vibrating sample magnetometer, (VSM, lake
Shore/7404, Lake Shore Cyrotonics, Ohio, USA). Meanwhile, infrared analysis was
performed with a Perkin Elmer Fourier Transform Infrared (FT-IR, MA, USA)
spectrometer. The Brunauer-Emmett-Teller (BET) analysis was conducted to identify
the specific surface area, the pore volume and the size of the sorbents (Quantachrome,
Boynton Beach, FL, USA), whereas elemental analysis was performed by using CHN
Analyzer Perkin Elmer 2400 Series II in order to identify the composition of C, H, and
1
13
N. H and C NMR spectra were recorded by using a Bruker's 700 MHz ASCEND™
spectrometer with DMSO-d as the solvent at 25℃.
6
8

2
2
.3
Synthesis of Fe O βCD-Vinyl-TDI
3 4@
.3.1. Synthesis of mono-6-deoxy-6-(p-tosyl)-β-cyclodextrin (β-CDOTs)
β-CDOTs refers to a crucial intermediate product that can be employed to functionalize
the primary hydroxyl groups (C ) of β-CD with IL. A solution of β-CD (2.86 g, 2.53
6
mmol), and p-toluenesulfonic anhydride (Ts O) (1.24 g, 3.8 mmol) in 60 mL of distilled
2
water was stirred at room temperature for 2 h under inert atmosphere. After that, a
solution of NaOH (1.25g, 31.3 mmol) in water (13 mL) was added, and after 10 min,
the reaction mixture was filtered through Celite sintered glass funnel to remove
unreacted TsO . The filtrate was adjusted to pH 8 by adding ammonium chloride (3.36
2
g) and stored overnight in the refrigerator. White and solid powder was formed after
drying under reduced pressure and the product was used without further purification
through chromatography. The synthesis was prepared in adherence to the methods
depicted in the literature (Zhong, Byun, & Bittman, 1998).
1
The degree of tosylation (DT) was determined via H NMR spectroscopy, which had
been based on the ratio of the proton areas.
퐴푟/4
퐷푇 (%) =
푋 100
(1)
퐻1 ‒ 퐻6
[
]푥 7
7
where DT(%) refers to the degree of tosylation, AR denotes the integral area of aromatic
protons at chemical shift (δ) 7.8-7.4 ppm, and H1-H6 reflect the integral areas of the
CD protons at δ 5.9-3.2 ppm. The DT (%) appeared to be less than 1, hence suggesting
the successful tosylation at primary hydroxyl of β-CD.
IR/KBr, cm-1: 3374 (OH), 2928 (C–H), 1654 (C=C), 1364 (SO Assy), 1169 (SO ,
2
,
2
Sym), 1028 (C-O).
9

1
H NMR/ppm, DMSO- : H8 (7.74, d), H9 (7.43, d), OH-2-OH-3 (5.5–5.8, m), H1
d6
(
4.84, s), H1* (4.76, s), OH-6 (4.20–4.55, m), H3, H5, H6 (3.21–3.65, m), H2–H4
2.88–3.21, m), H11 (2.42, s).
(
1
3
C NMR/ppm, DMSO- : C7 (145.8), C10 (133.1), C9 (128.5), C8 (128.0), C1 (102.3),
d6
C4 (82.0), C2 (73.5), C3 (72.5), C5 (72.0), C6* (69.4), C6 (59.9), C11-CH (21.6).
3
DT (%) = 0.84, Percentage yield: 39.0%, Melting point: 202-204 ℃
TLC: Eluent: 5:4:3 butanol / ethanol (95%) /water, R = 0.53
f
2
.3.2. Synthesis of mono-6-deoxy-6-(1-vinylimidazole)-β-cyclodextrin tosylate (β-CD-
Vinyl-OTs)
In order to prepare CD-IL complex, reaction was further carried out. Typical 1-
vinylimidazole (10 equiv.) was dropwised to a stirred solution of fresh β-CDOTs (1.0
g, 0.88 mmol) dissolved in 40 mL of anhydrous DMF. Stirring was continued at 90 ℃
under nitrogen for another 48 h. Acetone was added into the medium after cooling to
room temperature in order to precipitate the product before being stirred for 30 min.
Then, followed by filtration and washing in excess amount of acetone, followed by
vacuum dried to give a white powdery solid.
-1
IR/KBr, cm : 3348 (OH), 2932 (C–H), 1020 (C-O), 1668 (C=C), 1149 (C–N).
1
H NMR/ppm, DMSO- : Hd (7.95, s), He (7.74, d), H8 (7.49, d), Hf (7.47,d), (H9 (7.13,
d6
d), OH-2–OH-3 (5.50–5.90, m), Hc (5.52, d), H1 (4.80,s), OH-6 (4.48–4.66, m), Ha
(4.31, d), Hb (4,3, d), H6* (3.92, s), H3,H5,H6 (3.38–3.63, m), H2–H4 (3.21–3.38, m),
H11 (2.08, s)
1
0

1
3
C NMR/ppm, DMSO- : Cc (162), C7 (145), C10 (138), C8, Ce, Cb, C9, Cd (125.9-
d6
1
28.6), C1 (102.7), Ca (102.4), C4 (81.9), C2 (73.5), C3 (72.5), C5 (72.3), C6* (70.2),
C6 (60.0), C11 (21.3)
Percentage yield: 81.0%, Melting point: 204-207 ℃.
TLC: Eluent: 5:4:3 butanol / ethanol (95%) /water R = 0.75
f
2
.3.3. Synthesis of Fe O βCD-Vinyl-TDI
3 4@
The synthesis of Fe O βCD-Vinyl-TDI involved two stages. Initially, Fe O
4
3
4@
3
nanoparticles were prepared vial chemical co-precipitation method from Fe(II) and
Fe(III) salts under non-oxidizing conditions. Briefly, 40 mL of deionized water was
used to dissolve 0.86 g, 4.33 mmol of FeCl .4H O and 2.34 g, 8.66 mmol of FeCl 6H O
2
2
3.
2
and stirred for 30 min under nitrogen gas at room temperature. Next, the solution was
stirred at 80 ℃ and 5 mL of 25% NH OH was added dropwise to the solution. The
4
resulting black mixture was stirring for 1 h. After that, the obtained Fe O nanoparticles
3
4
were isolated from the reaction medium using external magnet and washed copious
times with deionized water. Finally, the Fe O nanoparticles were vacuum dried for 6 h
3
4
at 65 ℃.
The β-CD-Vinyl-TDI was grafted on the surface of Fe O nanoparticles by employing
3
4
the polymerization-induced self-assembly (PISA) method with slight modification
Karagoz et al., 2014). Briefly, 0.4 g, 0.29 mmol of β-CDOTs-Vinyl complex was
dissolved and definite 1.0 g of Fe O nanoparticles was dispersed into 12 mL of
(
3
4
anhydrous DMF. After that, 0.42 mL, 2.9 mmol of toluene-2,4-dissocynate (2,4-TDI)
was added dropwise and the reaction was stirred for 24 h at 80℃ under nitrogen
atmosphere. The resultant was precipitated with excess acetone. The product formed
was stirred and allowed to settle down in acetone for 10 min to removal DMF residual.
1
1

After that, the reaction had been filtered and washed with copious acetone for few times
in order to remove the unreacted cross-linker and later dried in a vacuum oven at 65 ℃
for 6 h. The synthesis of Fe O βCD-Vinyl-TDI is illustrated in Fig. 1.
3
4@
Fig. 1
2
.4.
Chromatographic Analysis
A Shimadzu gas chromatographic (Shimadzu-GC-20, Tokyo, Japan) equipped with a
split/split less inlet and flame ionization detector (FID) had been applied to determine
the PAHs. The temperature for injector and detector were 300 ℃ and 330℃ ,
respectively at spiltless mode. N (99.999%) served as the carrier gas with a constant
2
-1
flow of 30 mL min . Chromatographic separation was carried out with a HP5-MS fused
silica capillary (5%-Phenyl)-methylpolysiloxane at 30 m length, 25 mm I.D, and 0.25
μm film thickness. The GC oven was set as followed: 90 ℃ (5 min hold), and ramped
-1
to 290 ℃ at 10 ℃ min .
The gas chromatography-mass spectrometry (GC-MS) analysis was performed by
using the model 7890A GC-Agilent 5975C Series mass selective detector system
equipped with a splitless injector and a MSD chemstation software program on a non-
polar capillary column of Agilent HP5-MS column (30 m x 0.25 mm x 0.25 μm film
-1
thickness). Helium served as the carrier gas at a flow rate of 1.0 mL.min . The
temperature of both injector and the detector was set at 280 °C each. Meanwhile, the
column temperature was fixed as follows: 90 ℃ for 3 min, increased to 190 ℃ at 15.0
-1
-1
℃
min , and raised to 290 ℃ at a rate of 10.0 ℃ min and then held for 5 min. Data
were collected and quantified in the selected ion monitoring (SIM) mode by using mass
values that corresponded to the molecular ions of the PAHs: Ace, m/z 150.0, 151.0,
1
2

1
2
52.1, 153.0, and 154.1; Flu, m/z 165.0, and 166.1; Phe, m/z 176.0, and 178.1; Fla, m/z
00.0, and 202.1; as well as Pyr, m/z 200.0, and 202.1.
2
.5.
Rice Samples
A total of 24 rice samples, including the local brands consumed in Malaysia and
imported brands from Cambodia, India, Pakistan and Thailand were purchased from a
local supermarket (Penang, Malaysia). All the samples were stored at -4 ℃ after being
homogenized with a rice grinder and screened through 65-mesh sieve before being
analyse.
The rice samples were extracted by applying centrifugation method as described with
some slightly adjustments (Li et al., 2016). Briefly, 3 mL of acetonitrile was added into
a 15 mL centrifuge tube that contained 0.5 g of spiked grinded rice sample, and
followed by centrifuged for 3 min at 4000 rpm. Next, the supernatant was transferred
into a 40 mL glass vial before being top up to 25 mL of sample volume.
2
.6.
Magnetic solid phase extraction (MSPE) procedure
Initially, 20 mg of Fe O βCD-Vinyl-TDI was dispersed into 25 mL of sample solution
3
4@
and shaken for 30 min by an orbital shaker incubator to attain adsorption equilibrium.
Next, the analytes-adsorbed Fe O βCD-Vinyl-TDI was magnetically retrieved and the
3
4@
supernatant was decanted. After that, the analyte was desorbed in 200 μL of acetonitrile
under shaking for 15 min and Fe O βCD-Vinyl-TDI was again magnetically isolated
3
4@
from the desorption solution. The resulting desorption solution was collected and
filtered with nylon membrane filter 0.45 μm and later, collected into a 1.5 mL vials. 1
μL of the sample was injected into the GC-FID system for further analysis.
1
3

As for reusability tests, the residual sorbent was washed with 1 mL of acetonitrile
thrice) for 15 min and dried for 20 min before it was used for the next extraction
process.
.7. Solid phase extraction (SPE)
(
2
Slight modifications of the reported optimum SPE protocols (Sun, Littlejohn, & David
Gibson, 1998) has been adopted in this work by using Chromabond columns C18, 3
mL, 200 mg, Macherey-Nagel, Germany. In brief, 5 mL of acetone was used to
condition the SPE cartridge and equilibrated with 5 mL of 40% (v/v) acetone. Next, 25
-1
mL of spiked sample at a flow rate of 0.5 mL min was loaded into the cartridge.
Washing step was employed using 5 mL of deionized water and the cartridge was
vacuumed dried for 5-10 min to eliminate water content. Subsequently, 3 mL of
tetrahydrofuran: acetone (1:1) was applied to elute the adsorbed analytes. Finally, 1 μL
of the eluent was injected into the GC-FID system for quantification.
3
3
3
.
Results and discussion
.1.
Characterization
.1.1. FT-IR analysis
Table S1 summarizes the FT-IR wavenumbers of Fe O Fe O βCD-TDI and
3
4,
3
4@
Fe O βCD-Vinyl-TDI. All the spectra displayed in Fig. 2(A) exhibits an intense
3
4@
-1
characteristic band at 580 cm , which is attributable to Fe-O-Fe vibration of the
magnetic core (Zhai, He, Han, & Duan, 2012). After functionalization of imidazole
groups, the resultant Fe O βCD-Vinyl-TDI portrayed three peaks around 1469, 1575
3
4@
-1,
and 1388 cm which reflected N-H bending vibration, C=C, and C=N (Bi, Wang, Yang,
Row, 2014; Galán-cano, Alcudia-león, Lucena, Cárdenas, & Valcárcel, 2013)
signifying that the IL was successfully introduced to the material. The slightly weak
&
1
4

-1
bands around 2850-2950 cm can be assigned to the C-H stretching (C. Liu, Liao, &
Huang, 2017) from the -CH=CH groups of IL. The total disappearance of the peak at
2
-1
2
270 cm (in corresponding to isocyanate group) and the vivid presence of carbamate
-1
linkage NCHO at peaks 1653 and 1535 cm signifies the complete polymerization
process (Raoov, Mohamad, & Abas, 2014). The broad OH stretching band around 3335
-1
cm in bare Fe O nanoparticles reflected the stretching vibration that derived from the
3
4,
surface hydroxyl groups of Fe O nanoparticles, whereas it pointed out the combination
3
4
of multiple OH functional groups of β-CD and surface hydroxyl groups of Fe O
3
4
nanoparticles in Fe O βCD-TDI. Nonetheless, the intensity of the peaks for
3
4@
Fe O βCD-Vinyl-TDI increased after the IL was introduced, mostly due to the
3
4@
combination effect of amide N-H absorption, hydroxyl groups of β-CD, and surface
hydroxyl groups of Fe O nanoparticles. The outcomes of FT-IR spectra proved that all
3
4
the studied materials had been successfully prepared.
.1.2 Elemental analysis
3
Elemental analysis was conducted to determine the composition of C, H, and N for all
the developed materials wherein the findings are tabulated in Table S1. The outcomes
revealed that the content of the three elements increased significantly from Fe O
3
4,
Fe O βCD-TDI, and Fe O βCD-Vinyl-TDI, respectively. This shows that βCD-TDI
3
4@
3
4@
and βCD-Vinyl-TDI had been successfully coated onto the Fe O nanoparticles,
3
4
particularly the molar ratio of N contributed by the isocyanate group and the imidazole
ring. Consequently, the elementary analysis seemed to have supported the formation of
the developed materials, which are in agreement with the expected outcomes.
1
5

3
.1.3 BET analysis
The porous structures of Fe O Fe O βCD-TDI, and Fe O βCD-Vinyl-TDI
3
4,
3
4@
3
4@
nanoparticles were studied by using nitrogen adsorption-desorption isotherms. The
2
-1
unmodified Fe O nanoparticles yielded a BET surface of 96.99 m .g . Nevertheless,
3
4
the surface area for both Fe O βCD-TDI and Fe O βCD-Vinyl-TDI decreased from
3
4@
3
4@
2
-1
5
4.1 to 8.45 m .g . This decreasing could be attributed to the increment in particle
diameter due to the presence of coating layer from the organic moieties (polymeric β-
CD) on the surface of Fe O nanoparticles.
3
4
By comparing the adsorption capacity of the liquid nitrogen, Fe O βCD-Vinyl-TDI
3
4@
3
-1
3
-1
(
35 cm .g ) exhibited the lowest capacity, when compared to Fe O (180 cm .g ) and
3 4
3
-1
Fe O βCD-TDI (140 cm .g ), which is a conclusive evidence for the successful
3
4@
preparation of Fe O βCD-Vinyl-TDI nanocomposite since the N gas molecules had
3
4@
2
been hindered by IL from passing through the molecular dimension of the materials.
Barrett–Joyner–Halenda model (BJH) was used to calculate the pore size distribution
that gave pore diameters of 24.9 nm, 34.6 nm and 38.2 nm for Fe O Fe O βCD-TDI,
3
4,
3
4@
and Fe O βCD-Vinyl-TDI, respectively, indicating mesoporous structure. These
3
4@
results are in agree with the IUPAC definition that, which classifies the mesoporous
materials between diameter of 2–50 nm. The escalating trend of the pore size may be
linked to the presence of coating layer from polymeric β-CD and IL, which
dimensionally increased the diameter of particles, hence further verifying the successful
synthesis of the studied materials.
The Fe O nanoparticles displayed type IV isotherm with H1-type hysteresis loop
3
4
indicative of its mesoporosity. The outcomes are in line with the SEM analysis,
whereby porous material that exhibited type H1 contained uniform sphere aggregates.
1
6

However, Fe O βCD-TDI demonstrated type IV while Fe O βCD-Vinyl-TDI
3
4@
3
4@
showcased type III with H3-type hysteresis loop. The type H3 loop rendered that both
0
materials did not displayed any limiting adsorption at high p/p . Type IV isotherm
displayed by Fe O βCD-TDI could related to monolayer-multilayer adsorption on the
3
4@
same surface area of the adsorbent in non-porous form (mesoporous). On the other hand,
Fe O βCD-Vinyl-TDI with type III isotherm showed a gradual curvature in the
3
4@
adsorbate-adsorbate interactions (Sing, 1985). Table S1 presents the BET results of
Fe O Fe O βCD-TDI and Fe O βCD-Vinyl-TDI
3
4,
3
4@
3
4@
3
.1.4. Magnetization Studies
The principal feature of a magnetic material is its magnetic properties as sufficient
magnetism ensure rapid separation of materials from an aqueous medium practically in
MSPE application. VSM was used to access the magnetic behavior of the as-prepared
materials, as displayed in Fig. 2(B). The magnetism value of Fe O Fe O βCD-TDI,
3
4,
3
4@
-1
and Fe O βCD-Vinyl-TDI nanoparticles were 82, 80 and 46 emu.g , respectively. As
3
4@
portrayed in the figure, the three hysteresis loops showed almost remanence and zero
coercivity, implying that all the samples exhibited superparamagnetic behavior.. The
decrement in saturation magnetization value of Fe O βCD-TDI and Fe O βCD-
3
4@
3
4@
Vinyl-TDI nanocomposites can be contributed by the shielding of IL and cross-linker
due to the modification process. Nevertheless, the magnetic strength of the developed
material was still adequate for magnetic separation, as isolation was rapidly carried out
within 30 seconds from the medium dispersion using external magnetic field. Therefore,
the outstanding redispersibility and magnetic responsivity rendered Fe O βCD-Vinyl-
3
4@
TDI as a potential sorbent for MSPE application.
Fig. 2
1
7

3
.1.5. SEM and TEM analyses
The microscopic morphology of all nanostructures was investigated via SEM and TEM
and the representative surface analyses are displayed in Fig. S1. Bare Fe O
3
4
nanoparticles showed uniform agglomeration surfaces. Comparatively, the
Fe O βCD-TDI exhibited fluffy, porous and unorganized beads that covered the
3
4@
surface of Fe O nanoparticles, while, Fig. S1(A) (c) portrayed that the
3
4
functionalization of IL increased the number of amount small and well-organized
porous particles of Fe O βCD-Vinyl-TDI with apparent sphere-like shape that
3
4@
decorated uniformly on the surface of Fe O nanoparticles. Besides, low surface tension
3
4
of IL can serve as a steric stabilizer, thus lowering particle growth and preventing them
from agglomeration (Xiaodi Liu, Duan, Qin, Wang, & Zheng, 2013). Therefore, it is
indeed to obtain the surface morphology as described.
The TEM images of Fe O displayed nearly quasi-spherical in shape and mono-
3
4
dispersed. Little or no variance was observed between Fe O and Fe O βCD-TDI, thus
3
4
3
4@
indicating that polymerized βCD has no obvious effect on the morphology of the
support material. In Fig. S1(B) (c), the TEM image clearly shows that a thin layer of
organic phase that refers to IL layers were coated on the surface of dark magnetite
particles, hence presenting a core–shell structure of Fe O βCD-Vinyl-TDI
3
4@
nanoparticles. At the same time, functionalization of IL also greatly enhanced the
dispersion of Fe O nanoparticles in aqueous solution.
3
4
Fig. 2
Optimization of MSPE conditions
In order to comprehensively assess the performance of Fe O βCD-Vinyl-TDI, some
3
.2.
3
4@
essential experimental parameters, including sorbent dosage, extraction and desorption
1
8

time, type and volume of desorption solvent, ionic strength, sample pH, organic
modifier and sample volume had been investigated. All the optimization parameters
-1
were evaluated in real blank rice sample with spiking concentration of 300 μg kg and
all analyses were performed in triplicate. The standard deviations were calculated,
while analysis of variance (ANOVA) was applied for data analysis with IBM® SPSS®
Statistics 20, IBM Corp, USA at 95% confidence level.
3
.2.1. Comparison between different modified sorbents
In the attempt to determine presence of IL modification on the extraction performance
of the sorbents, the extraction capability of Fe O nanoparticles, Fe O βCD-TDI,
3
4
3
4@
Fe O βCD-Vinyl-TDI, βCD-Vinyl-TDI, and βCD-TDI had been compared using
3
4@
similar extraction conditions (n=3), which were 20 mg of sorbent, 20 mL of sample
solution, 30 min for both extraction and desorption, and 1.5 mL of acetonitrile as
desorption solvent. The findings depicted in Fig. S2 indicate that bare Fe O
3
4
nanoparticles exhibited limited extraction capability towards PAHs. On the other hand
the extraction efficiency exemplified by Fe O βCD-Vinyl-TDI sorbent was
3
4@
significantly superior, in comparison to the rest based on ANOVA (one way analysis).
This is perhaps due to the presence of IL in Fe O βCD-Vinyl-TDI that offers another
3
4@
excellent merit in terms of the structural design and the interaction system.
Vinylimidazole refers to an alkenyl IL that possesses -CH=CH vinyl group and
2
imidazolium ring. These features may enhance interactions with PAH analytes, such as
휋-휋 and hydrophobic interactions, as compared to those in Fe O and Fe O βCD-TDI,
3
4
3
4@
apart from the host guest interaction and the Van der Waal forces. Therefore, this study
concluded that the exceptional enrichment ability displayed by Fe O βCD-Vinyl-TDI
3
4@
to PAHs is a result of IL functionalization.
1
9

3
.2.2. Effect of sorbent dosage
In order to determine the relationship between the sorbent quantity and extraction
capability for desired PAH compounds, varied amounts of nanocomposite sorbents in
the range of 10-50 mg had been applied during the MSPE procedure. Fig. S3(a)
portrayed the extraction efficiency of PAHs that raised with the addition of 10-20 mg.
This had been due to the increment in the adsorption sites at higher amounts of sorbents
(Rezvani-Eivari, Amiri, Baghayeri, & Ghaemi, 2016). Light PAHs that comprised of
fewer aromatic rings used in this study could be extracted effectively as the amount of
sorbents increased. Nonetheless, further increment in sorbents seemed to slightly
decrease and remain unchanged in the efficiency. Higher amount of sorbents may not
be dispersed or separated in the same extraction time, which could reduce the surface
area that affects the efficiency. Thereafter, (>20 mg), some active sites still remained
unsaturated as not many analytes were adsorbed and adsorption process reached
equilibrium. This is associated with the fact that higher dosage of sorbent at specific
sample volume and contact time had reduced the efficiency of elution process
(
Tahmasebi, Yamini, Moradi, & Esrafili, 2013). Hence, 20 mg of the nanosorbents was
chosen for subsequent experiments.
.2.3. Effect of extraction time
3
Extraction time facilitates the analyte-particles interaction, which is linked to the
extraction performance. Sufficient time should be given to nanosorbents and sample
solution for the desired analytes to reach adsorption equilibrium, besides attaining
optimum extraction efficacy. Nonetheless, prolonged extraction time might reduce the
sorbent performances due to the effects of back diffusion (Nyi, Fong, Li, & Kee, 2016)
and dissolution of the sorbent onto the matrix. Fig. S3(b) illustrates the effect of
2
0

extraction time investigated within the range of 10 until 50 min. The peak area raised
rapidly for the first 30 min before reaching its plateau. No obvious change could be
observed after 40 and 50 min as the peak area were insignificant. Therefore, an
optimum of 30 min was selected for the extraction of PAHs.
3
.2.4. Effect of ionic strength
In order to facilitate the transfer of PAHs into Fe O βCD-Vinyl-TDI, sodium chloride
3
4@
as the model electrolyte was used to study the effect of ionic strength within the
concentration interval of 0-5% (w/v). Accordingly, the results are portrayed in Fig.
S3(c). The extraction efficiency of PAHs seemed to increase with the NaCl
concentration up to 3% due to salting out effect where the solubility of the analyte in
aqueous solution reduced, thus facilitating the adsorption of analyte towards
Fe O βCD-Vinyl-TDI. Kuo and co-workers stated that inorganic ions that present in
3
4@
the sample may disturb the diffuse double layer and notably reduce the repulsive energy
between the adsorbents, apart from enhancing the analyte adsorption implicitly (Kuo,
Wu, & Wu, 2008). Comparatively, with further increment in concentration, the
extraction efficiency appeared to decrease. It is interesting to note that sample viscosity
could be increase as the ionic strength increase, hence led to difficulty for the transfer
of analyte from bulk sample solution to the adsorbents. In addition, at high
+
-
concentration of salt solution, Na and Cl ions reversed the metathesis reaction to make
Fe O βCD-Vinyl-TDI more polar, aside from changing the Nernst diffusion film and,
3
4@
slower the diffusion rate of analyte from aqueous to sorbents since the solvation cage
of analyte is disrupted. As such, the extraction performance declined, therefore, further
experiments were carried out with 3% (w/v) of NaCl.
2
1

3
.2.5. Effect of pH
Solution pH has a crucial function in separation science as it controls both the
dissociation of functional groups and degree of ionization of the materials present in
the solution. Therefore, the effect of sample pH on extraction performance of the
targeted PAHs was evaluated by adjusting the pH value with hydrochloric acid or
sodium hydroxide solution in the range of 5.0-13.0. As expected, the optimum pH for
the extraction of PAHs was achieved in neutral pH (figure not shown). This is because
PAHs exist as neutral molecules and no other functional group may undergo
protonation or deprotonation under varied pH conditions attached to them (Han et al.,
2
012). Hence, for convenient analysis, the proposed method avoids fussy sample
adjustment.
.2.6. Effect of organic modifier
3
In order to reduce adsorption of PAHs on the glass vials (namely “wall effect”) (García-
Falcón, Cancho-Grande, & Simal-Gándara, 2004), the modifier effect was investigated.
For this purpose, various amounts of organic modifier, which were made up of 1% to
2
0% of acetic acid in elution solvent (acetonitrile), had been as added. As illustrated in
Fig. S3(d), increment in modifier content appeared to decrease the extraction efficiency
for PAHs. This negative effect may result due to the association between PAHs and the
organic modifier. The Fe O βCD-Vinyl-TDI sorbents failed to adsorb PAHs due to
3
4@
the electrostatic repulsion between the sorbents and acetic acid (S. Zhang, Niu, Hu, Cai,
Shi, 2010), in which the presence of acetic acid caused early breakthrough of PAHs.
Hence, organic modifier was excluded added in the following experiments.
&
2
2

3
.2.7. Effect of sample volume
Enrichment factor (EF) is directly proportional to sample volume. In order to get higher
EF, the influence of sample volume on the extraction efficiency was studied from 10 to
-1
3
5 mL of spiked 300 μg kg of sample solution in the presence of 20 mg of
Fe O βCD-Vinyl-TDI nanosorbents. Accordingly, the results are displayed in Fig.
3
4@
S3(e). Higher analytes indicate that all the PAHs were observed with increment volume
from 10 to 25 mL, but seemed to decrease significantly with further increased to 35
mL, as expected. This is because higher volumes caused the adsorption sites of the
sorbent completely saturated with analytes and it could not be processed as the amount
of sorbent in volume unit of the sample decreased due to high dispersion on the sorbent
in aqueous media (Pichon, 2000). This makes adsorption of analytes to become more
difficult. Based on the results, 25 mL was selected as the breakthrough volume and
deemed as the most favorable sample solution for the proposed method.
3
3
.2.8. Desorption Conditions
.2.8.1 Effect of desorption/elution solvent
In MSPE, desorption solvents have a significant role in ensuring that high extraction
efficiency could be obtained. Six solvents with varied polarity (polar, mid polar to non-
polar), namely, methanol, acetonitrile, acetone, ethyl acetate, dichloromethane, and
hexane had been applied to assess their desorption capability. The literature review
depicts that the choice of desorption solvent often reflects the solubility of analyte
(Galán-cano et al., 2013). The highest chromatographic response was achieved when
methanol (Fig. S3(f)) was used as the eluent, followed by acetonitrile and acetone. The
strong hydrophobic interaction between PAHs and Fe O βCD-Vinyl-TDI, as well as
3
4@
effective solubility of PAHs in methanol, is attributable to this observation since all the
2
3

studied PAHs indicated relatively higher hydrophobic character with log Pow > 3.5
Sahu & Pandit, 2003).
(
Mid polar to less polar solvents (ethyl acetate, dichloromethane, and hexane) showed
lower desorption efficiency because the developed sorbent may have not been well
dispersed in these solvents, thus leading to low extraction performance. Since the
prepared Fe O βCD-Vinyl-TDI nanocomposite displayed mid hydrophilic, its
3
4@
dispersive capacity in less polar solvent appeared to be poor, and it is reasonable to
record such outcomes.
Pure methanol caused peak shape and size issues, which led to poor resolution of the
peaks obtained. As such, further investigation was carried out by employing the mixture
of acetonitrile: methanol (1:1) and acetonitrile: acetone (1:1). Unfortunately, this
mixture of ACN:MeOH generated poor peak resolution, while the mixture of ACN:
Acetone yielded lower extraction ability. Hence, after weighing in chromatographic
response and peak resolution, acetonitrile was finally selected for desorption of PAHs
from the proposed material which revealed exceptional solubility of PAHs in ACN.
3
.2.8.2. Effect of desorption time
The effect of desorption time was also examined between 10 and 35 min. As illustrated
in Fig. S3(g), the analytical signals increased with the shaking time from 10 to 15 min.
The analytical signal of the analytes, however, did not significantly increase after 15
min of shaking. Hence, an optimum desorption time of 15 min apparently to be
acceptable for further experiments. Nevertheless, no further improvement (slight
increment in the peak area) was noted in some PAHs after 20 min, but not to the extent
of that recorded reached at 15 min in the enriched solvent phase. The decrease observed
in the peak area beyond this desorption time could be due to the re-desorption of the
2
4

target analytes to the sorbent upon when increased duration (Nyi et al., 2016). Therefore,
it is crucial to consistently apply desorption time upon attaining optimization. The
conclusion drawn here is that 15 min seemed sufficient and optimal for a complete mass
transfer of the analytes from Fe O βCD-Vinyl-TDI nanosorbents toward the organic
3
4@
solvent.
.2.8.3. Effect of desorption volume
3
The effect of desorption volume on the elution of analytes was examined to determine
the optimum analytical signal. Fig. S3(h) portrays that the smaller the volume used, the
higher the extraction performance. Finally, 200 μL of desorption solvent has been
proved sufficient to completely elute the PAHs from the sorbent to achieve satisfactory
efficiency.
Based on the overall outcomes, the optimum extraction parameters were finally selected
as follows: 20 mg of nanosorbents in 25 mL sample solution without pH adjustment,
3
% (w/v) of salt addition, extraction time of 30 min, 200 μL of acetonitrile without
organic modifier as desorption solvent and desorption time of 15 min under shaking
condition.
3
.2.9. Sorbent reutilization study
The possibility of reusing Fe O βCD-Vinyl-TDI nanosorbents was studied after
3
4@
considering the environmental importance of saving chemicals and to prevent wastes.
This reusability study was carried out by employing the same Fe O βCD-Vinyl-TDI
3
4@
nanosorbents, which were being washed thrice with 1.5 mL of acetonitrile, respectively
and dried at 65 ℃ for 30 min. The results (Fig. S4) illustrated that the sorbents could
be reused as many as six times with its signal remaining practically constant, and no
carryover of analytes during the MSPE procedure. This demonstrated the stability and
2
5

the possibility of reusing Fe O βCD-Vinyl-TDI nanosorbents in the real MSPE
3
4@
application.
.3. Analytical performance/Validation of the proposed MSPE method
3
In order to validate the performance of the proposed MSPE coupled with GC-FID
method, a series of experiments with regard to linearity, limit of detection (LOD), limit
of quantification (LOQ), precision, enrichment factor (EF), and sorbent-to-sorbent
reproducibility, had been performed under optimal conditions. Additionally, the
standard SPE procedures were established as well for comparison purpose.
3
.3.1. Linearity, limit of detection (LOD), and limit of quantification (LOQ)
-1
A series of spiked samples that ranged between 0.1 and 500 μg kg had been applied
to plot the working curves (n=3). The validation parameters for the target analytes via
MSPE under optimized conditions are listed in Table 1. All the analytes exhibited
2
exceptional linearity with satisfactory determination coefficients (R ) ranging from
0
.9970 to 0.9982. The LODs were defined as LOD=3.3 s/m while the LOQs were
denoted as LOQ=10 s/m. (s is standard deviation of blank residuals, and m is slope of
-1
calibration graph). The LODs were found to be in the range of 0.01-0.18 μg kg ,
-1
whereas, the LOQs ranged at 0.05-0.55 μg kg . It is interesting to note that the value
of LODs and LOQs for some PAHs analytes were comparable or appeared to be better
-1
-1
than those achieved by standard SPE (0.01-0.22 μg kg , 0.1-0.68 μg kg ), as presented
in Table S2. Under optimal MSPE-GC-FID conditions, the PF values for Ace, Flu, Phe,
-1
Fla, and Pyr at spiked level of 10 μg kg were obtained as 196, 122, 133, 147, and 126,
respectively.
Table 1
2
6

3
.3.2. Precision
The reproducibility of the method was determined by measuring the intra-day
repeatability) and inter-day (intermediate) precisions at spiking level of 300 μg kg-1
(
rice samples. The intraday precision was determined on the same day and consisted of
six replicates. Interday precision was calculated with three replicates on six different
days. Satisfactory precisions were obtained with relative standard deviation (RSD)
values in the range between 2.95% and 5.34% for intra-day, as well as between 4.37%
and 7.05% for inter-day precision, illustrating the good reproducibly achieved by the
developed method.
3
.3.3. Sorbent reproducibility study
The preparation reproducibility of the Fe O βCD-Vinyl-TDI nanosorbents was
3
4@
evaluated in terms of lot-to-lot or batch-to-batch reproducibility. Three batches of
sorbents were prepared under the same conditions, but varied timings were used in the
-1
extraction study that examined fortified 500 μg kg of rice sample. The RSD values
ranged from 2.9% to 9.9%, and these seemed to be better than the acceptable values
recommend by the AOAC (<16%) (AOAC, 2002). This indicates undertaken to prepare
of Fe O βCD-Vinyl-TDI nanosorbent in this study displayed has a good
3
4@
reproducibility.
.3.4. Matrix effect
3
Matric effect due to co-extracting and co-eluting matrix substances can greatly
influence the chromatographic signals of the analytes. Commonly, the matrix effect
could be determined by using following equation (Li et al., 2016):
퐴푚 + 푠 ‒ 퐴푚
푀푎푡푟푖푐 퐸푓푓푒푐푡 (%) =[
] ‒ 1 푥 100
(2)
퐴0
2
7

Where Am+s = analytical signal of an analyte in spiked rice sample, A = analytical
m
signal of an analyte in blank rice sample and A = analytical signal of an analyte in
0
standard solution. In this study, all PAHs compounds were investigated and the
resulting absolute values had been less than 3%, as tabulated in Table 1, hence
revealing that the matrix effect of the present method is very slight and negligible (Dong
&
Xiao, 2017; Dong, Xiao, Xian, Wu, & Zhu, 2019; Dong, Zeng, & Bai, 2018)
3
.4.
Relative recovery study
The performance of extraction for the proposed approach based on relative recovery
was assessed by fortifying the analytes at three varied concentration levels (5, 50, and
-1
3
00 μg kg ) into the blank rice samples. Relative recovery was calculated as the
percentages of the mean values of extracted PAHs peak area (n=3) against the peak area
of PAHs derived from the calibration curve. The outcomes as summarized in Table 2,
ranged between 80.4% and 112.4% with RSD values falling from 4.2 to 8.9%. The
recovery is positively acceptable, as the results illustrated that the extraction efficiency
is stable and unaffected by the variances in PAHs concentration found in the rice
samples. The figure displayed the chromatogram of the separation of PAHs fortified
into rice samples after treated with Fe O βCD-Vinyl-TDI nanosorbents. All the peaks
3
4@
in the spiked rice sample extracts exhibited good separation along with good resolution.
Fig. 3 demonstrates the GC chromatograms of the separation of PAHs fortified into rice
samples after treated with Fe O βCD-Vinyl-TDI-MSPE under optimized conditions.
3
4@
Fig. 3
Table 2
2
8

3
.5.
Analysis of real rice samples
In the attempt to access the applicability of the proposed method in real matrices, PAHs
were analyzed in real rice samples purchased from a local market (Penang, Malaysia).
A total of 24 different rice samples, including 14 types of white rice, 2 types of glutinous
rice, 5 types of brown rice, and 1 type of black, red and baby rice, respectively, were
tested to determine the contents of PAHs. From the quantitative results (Table S3), only
-1
two PAHs compounds were detected in all the tested rice samples (μg kg dry weight):
Phe and Fla with Phe emerging as the most abundant element. The concentration of Phe
-1
-1
detected in the sample ranged from 4.82 μg kg to 22.3 μg kg , whereas 7.68- 8.62 μg
-1
kg for Fla. The found concentration (expressed in Toxic Equivalent Factor (TEQ) also
reflected the carcinogenic potency) of the tested rice groups are portrayed in Fig. S5
and was determined by using the following equation (Nisbet & LaGoy, 1992):
푇퐸푄 = ∑ 푃퐴퐻 푥 푇퐸퐹
(3)
푖
푖
Where PAH is concentration of PAH congener; TEF is toxic equivalent factor for PAH
i
i
congener and TEQ is the toxic equivalents of reference compound. Sample 9 was found
i
to have the least TEQ of 0.00482, while Sample 15 displayed the highest TEQ of 0.0223
in which the values are direct correlation to the found concentration, respectively for
both samples. Coherently, a GC-MS analysis was performed to confirm the gas
chromatographic results among the tested rice samples, wherein the findings verified
the presence of the two analytes, which were Phenanthrene (Phe) and Fluoranthene (Fla)
(Table S3), and the obtained MS spectra were compared with the library (Fig. S6). The
findings observed in this study mirror those of survey data and pass studies that
examined the presence of PAHs in rice samples via various extraction techniques, in
which low molecular weight PAHs compounds, particularly Phenantherene,
2
9