Biosensors and Bioelectronics 86 (2016) 741–747

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios

“Sign-on/off” sensing interface design and fabrication for propyl gallate

recognition and sensitive detection
Yunlong Dai a,b,c, Xueyan Li a,b,c, Limei Fan a,b,c, Xiaojing Lu a,b,c, Xianwen Kan a,b,c,n
a
College of Chemistry and Materials Science, Anhui Normal University, Wuhu, 241000 PR China
b
The Key Laboratory of Functional Molecular Solids, Ministry of Education, PR China
c
Anhui Laboratory of Molecule-Based Materials, Anhui Key Laboratory of Chemo-Biosensing, PR China

art ic l e i nf o a b s t r a c t

Article history: A new strategy based on sign-on and sign-off was proposed for propyl gallate (PG) determination by an
Received 4 May 2016 electrochemical sensor. The successively modified poly(thionine) (PTH) and molecular imprinted poly-
Received in revised form mer (MIP) showed an obvious electrocatalysis and a good recognition toward PG, respectively. Fur-
18 July 2016
thermore, the rebound PG molecules in imprinted cavities not only were oxidized but also blocked the
Accepted 21 July 2016
electron transmission channels for PTH redox. Thus, a sign-on from PG current and a sign-off from PTH
Available online 22 July 2016
current were combined as a dual-sign for PG detection. Meanwhile, the modified MIP endowed the
Keywords: sensor with recognition capacity. The electrochemical experimental results demonstrated that the pre-
Electrochemical sensor pared sensor possessed good selectivity and high sensitivity. A linear ranging from 5.0
10  8 to
Propyl gallate
1.0
10  4 mol/L for PG detection was obtained with a limit of detection of 2.4
10  8 mol/L. And the
Molecular imprinted polymer
sensor has been applied to analyze PG in real samples with satisfactory results. The simple, low cost, and
Poly(thionine)
Dual-sign effective strategy reported here can be further used to prepare electrochemical sensors for other com-
pounds selective recognition and sensitive detection.
& 2016 Elsevier B.V. All rights reserved.

1. Introduction
Antioxidants are of significant in food industry because they are
usually used as additives to prevent or retard oxidation reactions
in food products (Andreu-Navarro et al., 2011). Propyl gallate (PG),
as one of the most widely used synthetic phenolic antioxidants, is
often used in dehydrated food stuffs, edible fats and oils (Cui et al.,
2015). The use of PG is subject to very strict regulation owing to its
potential toxic effects. The acceptable PG daily intake (ADI) re-
commended by the Joint FAO/WHO Expert Committee on Food
Additives (JECFA), expressed in milligrams of additive per kilogram
body weight, is 0–1.4 mg/kg, and the maximum level re-
commended in foodstuffs is 200 mg/kg (Andre et al., 2010).
Moreover, PG strengthens the antioxidative action of tert-butyl-
hydroxyanisole so that these two additives can be found together
in a lot of foodstuffs (Morales et al., 2005). And the hygienic
standards of using food additives in China (GB2760-2011): PG can
be used in edible oil, fried food, dried fish products, biscuit, instant
noodles, quick-cooking rice, canned nut, curing meat products. The
maximum amount of PG can't be more than 0.1 g/kg.
n
Corresponding author at: College of Chemistry and Materials Science, Anhui
Normal University, Wuhu 241000, PR China.
E-mail address: kanxw@mail.ahnu.edu.cn (X. Kan).
So far, many methods have been employed for PG detection. For
example, capillary electrophoresis (Hall et al., 1994) was employed
to separate PG and other three food grade antioxidants with the
same concentration of 1 mmol/L (2 pmol). Liquid chromatography
(Perrin and Meyer, 2002) and gas chromatography (Wyatt, 1981)
were also used for PG detection with the concentration ranges of
2.4
10  6–2.4
10  4 mol/L and 8
10  6–1.02
10  5 mol/L, re-
spectively. Although these methods are useful for the selective
determination of the analytes, they require for a long time as well
as expensive instrument. Therefore, it is necessary to develop a
convenient and inexpensive analytical method for PR detection
with high sensitivity and selectivity.
Molecular imprinted technique has been extensively developed
for template molecule recognition due to the specific adsorption
capacity of the prepared molecular imprinted polymer (MIP)
(Wulff, 1995; Vlatakis et al., 1993). The advantages of easy pre-
paration, low cost, and stability over harsh conditions enable MIP
to be applied in many fields, such as solid phase extraction,
chromatography separation, and electrochemical sensoring (Cai
et al., 2014; Xiao et al., 2013; Bai et al., 2015; Wang et al., 2016;
Ahmad et al., 2015). The MIP based electrochemical sensors have
exhibited good selectivity toward template molecule. However,
the sensitivity of the MIP based sensor has been limited due to the
poor conductivity of polymer matrix. Many kinds of nanomaterials
or nanocomposites are involved for MIP based electrochemical

http://dx.doi.org/10.1016/j.bios.2016.07.072
0956-5663/& 2016 Elsevier B.V. All rights reserved.
742 Y. Dai et al. / Biosensors and Bioelectronics 86 (2016) 741–747
sensors preparation to enhance the sensitivity of the sensors. MIP
and bimetallic Fe/Cu magnetic nanoparticles were modified on
pencil graphite electrode for pyridoxine and pyridoxal-5-phos-
phate detection. The limits of detection (LOD) of both compounds
were 2.4
10  10 mol/L and 1.7
10  10 mol/L, respectively (Patra
et al., 2015). Electrode modified with MIP/CuO was used for do-
pamine detection with a LOD of 8
10  9 mol/L (Li et al., 2015a).
An electrochemical sensor based on nanoporous gold leaf and MIP
was prepared for metronidazole detection with a LOD of
1.8
10  11 mol/L (Li et al., 2015b). And graphene has been ex-
tensively involved for MIP based sensors preparation due to its
fascinating electrical, thermal and physical properties. Recently, a
review has reported the graphene based MIP sensing platform
fabrication and application (Roy et al., 2016). All these sensors
showed much high sensitivity for template molecules detection.
And yet, the over-loaded nanomaterials should influence the sta-
bility of the prepared sensor.
Conductivity polymers have been employed for electro-
chemical sensor preparation to enhance the sensitivity of the
sensor. Poly(aniline), poly(pyrrole), Prussian blue (PB), and other
polymers with high conductivity have been modified on electrode
surface to detect target molecules with high sensitivity (Zhai et al.,
2013; Mao et al., 2015; Kong et al., 2015). Meanwhile, PB is also
reported for MIP based electrochemical sensor preparation (Li
et al., 2013). PB and MIP were successively modified onto bare
electrode surface. The inner layer of PB film was used as a redox
mediator to directly produce an electrochemical signal since the
imprinted cavities in outer MIP layer can be used as channels for
electron transmission of PB on the electrode surface. With the
adsorption of template molecules, the current signal decreased
continuously since the adsorbed template molecules blocked the
cross of probe. Thus, this sign-off strategy has been used for non
electroactive template molecule recognition and detection (Yang
et al., 2015).
Due to good redox reversibility and stability, poly(thionine)
(PTH) has been used as an excellent electron mediator for elec-
trochemical sensors preparation. PTH modified glassy carbon
electrode showed catalytic activity and high sensitivity for hy-
droquinone and catechol simultaneous detection (Ahammad et al.,
2011). Gao fabricated PTH-modified screen-printed carbon elec-
trodes by electrooxidative polymerization of thionine in neutral
phosphate buffer, which was found to give stable and reproducible
electrocatlytic responses to NADH (Gao et al., 2003).
In the present work, a new strategy based on the combination
of sign-on and sign-off has been proposed for PG recognition and
detection. PTH and MIP were successively electropolymerized on
multi-walled carbon nanotubes modified glassy carbon electrode
surface. After the removal of PG, the underlying PTH was used as a
redox mediator to directly produce a high electrochemical signal.
The adsorbed PG in low concentration could not cause obvious
current, which block the channels for electron transmission of PTH
in turn. Thus, the obvious decrease of PTH current could be found.
With the increase of PG concentration, the peak current of PG
could be observed, which increased with the increase of its con-
centration. Meanwhile, PTH current decreased gradually. Sign-on
from PG and sign-off from PTH were combined as a dual-sign for
PG sensitive and selective detection with a very wide linear range.
2. Experimental
2.1. Chemicals
o-phenylenediamine (OPD) and thionine (TH) were purchased
from Sigma-Aldrich Co., LLC. Propyl gallate (PG, 98%), gallic acid
(GA, 98.5%), ethyl gallate (EG, 98%), butylated hydroxytoluene
(BHT, 99.9%), butylated hydroxyanisole (BHA, 98%), and tert-bu-
tylhydroquinone (TBHQ, 99%) were ordered from Aladdin (Alad-
din, China). Multi-walled carbon nanotubes (MWCNTs, diameters
of 10–30 nm and lengths of 1–2 mm) were purchased from
Shenzhen Nanotech Port Co. All other reagents were of at least
analytical-reagent grade. And double-distilled deionized water
was used for all solutions.
2.2. Apparatus
Cyclic voltammetry (CV) and differential pulse voltammetry
(DPV) were performed on CHI 660 C workstation (Chenhua In-
struments Co., Shanghai, China) with a conventional three-elec-
trode system. A bare or modified glassy carbon electrode (GCE)
was served as a working electrode. A saturated calomel electrode
and a platinum wire electrode were used as a reference electrode
and a counter-electrode, respectively. Field emission scanning
electron microscope (FE-SEM) images were obtained on an S-4800
field emission scanning electron microscopy (Hitachi, Japan).
2.3. Fabrication of MWCNTs modified electrode (MWCNTs/GCE)
2.5 mg MWCNTs were immersed into 5 mL ultrapure water
under ultrasonic dispersion for 30 min. The clean GCE was im-
mersed into the MWCNTs-COOH suspension and treated with a
constant potential of þ1.7 V for 400 s, obtaining the MWCNTs/GCE
(Kan et al., 2012).
2.4. Fabrication of poly(thionine) (PTH) modified MWCNTs/GCE
(PTH/MWCNTs/GCE)
The PTH/MWCNTs/GCE was prepared according to literature
(Liu et al., 2008; Dempsey et al., 2004) with little modification as
follows: CV method was carried out for 30 cycles with the po-
tential range from  0.4 V to 0.4 V at a scan rate of 50 mV/s by
immersing MWCNTs/GCE in 5 mL 5 mmol/L thionine solution
phosphate buffer solution (PBS, 0.1 mol/L, pH 6.0).
2.5. Fabrication of MIP modified PTH/MWCNTs/GCE (MIP/PTH/
MWCNTs/GCE)
The electropolymerization of MIP film on PTH/MWCNTs/GCE
was performed in a pre-polymerization solution containing the o-
phenylenediamine (OPD, 0.0162 g, 30 mmol/L ) and propyl gallate
(PG, 0.0106 g, 10 mmol/L). The polymerization cycles and scan rate
were  0.35 V  0.85 V and 50 mV/s, respectively. Afterwards, the
modified electrode was incubated in the solution of hydrochloric
acid solution (0.1 mol/L) for 15 min to extract the template mole-
cules. Then the modified electrode was washed by ultrapure water,
which was labeled as MIP/PTH/MWCNTs/GCE. The procedure for
the construction of the MIP/PTH/MWCNTs/GCE was illustrated in
Scheme 1.
As a comparison, a non-molecular imprinted polymer (NIP)
modified PTH/MWCNTs/GCE was prepared in the same process
just in the absence of PG to get NIP/PTH/MWCNTs/GCE. Other two
modified electrodes of MIP/MWCNTs/GCE and NIP/MWCNTs/GCE
were also fabricated in the same way except that no PTH layer was
electropolymerized in the sensor preparation process.
2.6. Electrochemical measurements
All electrochemical measurements were performed in PBS
(0.1 mol/L, pH 5.0). Chronoamperometry was performed at the
potential of 1.7 V and the equilibrium time was set at 400 s. The CV
method was performed from  0.6 V to þ0.8 V at a scan rate of
100 mV/s. The DPV method was performed from 0.6 V to þ0.6 V
 Y. Dai et al. / Biosensors and Bioelectronics 86 (2016) 741–747
Scheme 1. The schematic for MIP/PTH/MWCNTs/GCE preparation.
at pulse amplitude and width of 50 mV and 50 ms, respectively.
Selectivity of the sensor was investigated by choosing GA, EG, BHT,
BHA, and TBHQ as analogs.
2.7. Treatment of samples
All samples were treated according to literature (Guan et al.,
2006) with a little modification. Firstly, 1 mL oil sample (or 1 g
cookies) mixed with 2 mL methanol for 0.5 h in an ultrasonic bath
with subsequent centrifugation. Then, the supernatant of each
sample was diluted 100 times with PBS. Each sample was spiked
with 1.0
10  6 mol/L and 1.0
10  5 mol/L PG. The known
amount of PG was spiked into each samples for the determination
on MIP/PTH/MWCNTs/GCE.
3. Results and discussion
3.1. Design principle of the sensor
In this work, a novel route for the preparation of MIP/PTH/
MWCNTs/GCE was proposed, as illustrated in Scheme 1. Firstly, the
MWCNTs were electrodeposited onto the surface of GCE to get a
large surface area and high electric conductivity, which can serve
as the electronic conducting channels for the further modification.
Then the PTH was subsequently electropolymerized on the surface
of the modified electrode to form conducting polymer film. Due to
PTH film contains free –NH2 group (Xu et al., 2005), which could
form hydrogen bonds with the –OH groups of the PG. This es-
sentially contributes to weakening of the hydroxyl bond energies
to facilitate electron transfer through O–H  N (Ding et al.,
2005). Moreover, the combination functions also probably can be
represented as hydrogen bonds between the –OH groups of the PG
and –NH2 group of functional monomers (o-phenylenediamine,
OPD), which are very important for the identification property of
the prepared MIP sensor. After the removal of PG, the underlying
PTH was used as a redox mediator to directly produce a high
electrochemical signal. The adsorbed PG molecules in low con-
centration could not cause an obvious current of its own, which
block the channels for electron transmission of PTH in turn to
decrease the current of PTH. With the increase of PG concentra-
tion, the peak current of PG could be observed, which increased
with the increase of its concentration. Meanwhile, MIP film en-
dowed the sensor with good recognition capacity to PG. Thus, a
signal-on/off composed of the increase of current intensity of PG (|
ΔIPG|) and decrease of current intensity of PTH (|ΔIPTH|) can be used
for the sensitive detection of PG.
743
3.2. Characterization of the sensor
The morphologies of the MWCNTs/GCE, PTH/MWCNTs/GCE,
and MIP/PTH/MWCNTs/GCE were investigated by SEM, as shown
in Fig. 1. MWCNTs have been successfully electrodeposited onto
the surface of GCE with a three dimensional network structure,
which can provide a large surface for the further modification
(Fig. 1A). After the polymerization of PTH on MWCNTs/GCE, the
shape of MWCNTs/GCE became obscure and some nanoculsters
appeared (Fig. 1B). The energy dispersive spectroscopy (EDS)
profile in Fig. 1(D) revealed the presence of S and N, which are the
component elements of the PTH layer. The morphology and EDS
results indicated the successful preparation of the PTH/MWCNTs/
GCE. Fig. 1(C) showed the morphology of the MIP/PTH/MWCNTs/
GCE. An obvious film was found on the surface of PTH/MWCNTs/
GCE, indicating the successful preparation of MIP/PTH/MWCNTs/
GCE.
Cyclic voltammetry (CV) curves were used to investigate the
behavior of different modified electrodes. As shown in Fig. 2(A), a
pair of redox peaks of  0.13 V and  0.20 V was found on both
PTH/GCE and PTH/MWCNTs/GCE (Fig. 2Aa) in PBS (0.1 mol/L, pH
5.0), which were the typical redox peak of PTH (Liu et al., 2008).
Compared with PTH/GCE, the current response remarkably in-
creased on PTH/MWCNTs/GCE (Fig. 2 Ab). This could be attributed
to the large surface area and high conductivity of MWCNTs, which
improved the sensitivity of the modified electrode. When CV was
performed on MWCNTs/GCE in 2.0
10  5 mol/L PG solution, only
a oxidation peak at 0.29 V was found (Fig. 2 Ac). An obviously
negative shift of the oxidation peak of PG was found on PTH/
MWCNTs/GCE (ΔE ¼21 mV) and the current response of PG was
almost 2 times as that on MWCNTs/GCE (Fig. 2Ad), indicating the
good electrocatalysis of PTH toward PG. More importantly, the
redox peaks of PTH were far apart from the oxidation peak of PG.
This non-overlapping of oxidation peak between PTH and PG
provides a possibility for the investigation of sign-on and sign-off
strategy.
Fig. 2(B) showed the electrochemical behavior of different
modified electrode in the absence or presence of PG. Compared
with MIP/MWCNTs/GCE (Fig. 2Bc), a pair of redox peaks of PTH
could be observed, indicating that the left imprinted cavities
provided electron transmission channels for PTH redox (Fig. 2Ba).
The oxidation peak of PG was found on both MIP/PTH/MWCNTs/
GCE (Fig. 2Bb) and MIP/MWCNTs/GCE (Fig. 2Bd), implying that PG
molecules have been rebound into imprinted cavities. It's also
interesting to see that the current of PTH distinctly decreased on
MIP/PTH/MWCNTs/GCE, which probably because that the channels
for electron transmission to PTH film had been partly blocked by
744 Y. Dai et al. / Biosensors and Bioelectronics 86 (2016) 741–747

Fig. 1. SEM images of MWCNTs/GCE (A), PTH/MWCNTs/GCE (B), and MIP/PTH/MWCNTs/GCE (C). And the EDS of PTH/MWCNTs/GCE (D).

Fig. 2. (A) CV curves recorded on PTH/GCE (a) and PTH/MWCNTs/GCE (b) in the absence of PG; CV curves recorded on MWCNTs/GCE (c) and PTH/MWCNTs/GCE (d) in the
presence of 2.0
10  5 mol/L PG. (B) CV curves recorded on MIP/PTH/MWCNTs/GCE in the absence (a) and presence of 1.0
10  5 mol/L PG (b); MIP/MWCNTs/GCE in the
absence (c) and presence of 1.0
10  5 mol/L PG (d), and NIP/PTH/MWCNTs/GC in the absence (e) and presence of 1.0
10  5 mol/L PG (f).

the rebound PG molecules. Thus, the sign-on from PG and sign-off
from PTH can be combined as a dual-sign for sensitive detection of
PG. As for NIP/PTH/MWCNTs/GCE, both oxidation peak of PG and
redox peaks of PTH could not be found although the modified was
immersed 1.0
10  5 mol/L PG due to no imprinted cavities left in
polymer film.
3.3. Optimization the preparation of experimental conditions
Different influencing facts for MIP film preparation, such as the
molar ratio between template molecule and function monomer,
the polymerization cycles, the scan rate, and the pH value of the
electropolymerization solution were investigated. And the ad-
sorption time for PG detection also was optimized. Upon the sign-
on and sign-off strategy, the dual-signal of current changes (|
ΔIPG| þ|ΔIPTH|), caused by the decrease of PTH and the increase of
PG on modified electrode were recorded and calculated. In the
optimized experiments, the concentration of PG used for current
measurement was 5.0
10  6 mol/L.
 Y. Dai et al. / Biosensors and Bioelectronics 86 (2016) 741–747
The molar ratio between template molecule and function
monomer would affect the amount of imprinted cavities. Molar
ration between PG and OPD from1:1 to 1:5 were used to prepare
five different sensors. As shown in Fig. S1A, the current response
increased with the increase of molar ration from 1:1 to 1:3. And
then the current decreased when the molar ration further in-
creased. This phenomenon can be explained that less PG mole-
cules were wrapped during the electropolymerization when the
concentration of OPD was lower. However, high concentrations of
OPD would result in deep embedding of PG, which further caused
the difficulty for PG removal. Therefore, the optimal molar ratio
between the template molecule and function monomer was 1:3.
The pH value of the electropolymerized solution for MIP pre-
paration would affect the sensitivity of the modified electrode. As
shown in Fig. S1B, the peak currents increased rapidly with the
increase of pH from 4.0 to 6.0. A reason might be due to a proto-
nated form (NH3 þ ) of the nitrogen atoms of PTH films at low pH
(Yang et al., 2009). With the increase of pH, the formation of NH3 þ
decreases and more protic aromatic molecules, PG, can interact
with nitrogen atoms of the polymer. However, at neutral pH or in
an alkaline solution, deprotonated PG molecules become anions to
be oxidized (Qi and Zhang, 2005). Also, nitrogen containing
functional groups at the PTH modified electrode may become
deprotonated and possess negative charges at higher pH. There-
fore, electrostatic repulsion between the analytes and the PTH
might be one of the reasons in the decrease of the peak currents of
PG and PTH (|ΔIPG| þ |ΔIPTH|) with the pH increasing from 6.0 to 8.0.
Thus, the optimal pH value of electropolymerization solution was
chosen as 6.0.
The thickness and the compactness of the polymer film would
affect the sensitivity of the modified electrode, which was con-
trolled by the polymerization cycles and scan rate, respectively. As
shown in Fig. S1C, the current response increased with the in-
crease of the polymerization cycles, which could be attributed to
the formation of less imprinted cavities. However, high poly-
merization cycles would produce too thick MIP film to extract
template molecules. It was found that the highest current re-
sponse of PG was obtained when the electropolymerization cycle
was 30, which was chosen as the optimized polymerization cycle.
Fig. S1D showed the effect of scan rate for the sensor preparation.
When the scan rate was low, the tight film was produced, leading
to the difficulty of PG removal. Due to the loose and rough polymer
film, a low recognition capacity would produce when the sensor
was prepared at a high scan rate. Thus, the optimal polymerization
Fig. 3. (A) DPV curves recorded on MIP/PTH/MWCNTs/GCE with the successive addition of PG (a to k are 5.0
10  8,1.0
10  7, 5.0
10  7, 1.0
10  6, 3.0
10  6,
6.0
10  6,1.0
10  5, 2.0
10  5, 4.0
10  5, 6.0
10  5, 1.0
10  4 mol/L). (B) The calibration plot of the concentration of PG versus peak current on MIP/PTH/MWCNTs/GCE
(a) , MIP/MWCNTs/GCE(b). Inset: The calibration plot of the concentration of PG versus peak current on MIP/PTH/MWCNTs/GCE in low concentrations.
745
scan rate was found to be 50 mV/s.
Since the MIP film was electropolymerized on the surface of
PTH/MWCNTs/GCE, adsorption rate is expected to be rapid. As
shown in Fig. S2, the current changes of PG and PTH increased
rapidly in 2 min. After 2 min, the current changes tend to be
equilibrium, indicating the adsorption saturation of MIP film. The
result also demonstrated that the prepared sensor possessed rapid
adsorption rate. Therefore, the optimal adsorption time was 2 min
3.4. Analytical performance of the sensor
Under the optimal conditions, the dependence of the oxidation
peak of PG and PTH on MIP/PTH/MWCNTs/GCE was recorded by
DPV, as shown in Fig. 3. With the successive addition of PG into the
solution, peak current of PTH obviously decreased. However, no
obvious oxidation peak current of PG could be found until the
concentration of PG increased to 1.0
10  5 mol/L. Although the
peak current of PG slowly appeared, it subsequently increased
with the increase of concentration. Meanwhile, PTH current de-
creased gradually. A dual-sign (|ΔIPG| þ|ΔIPTH|) combined by sign-
on from PG and sign-off from PTH was proportional to the con-
centration of PG from 5.0
10  8 to 1.0
10  4 mol/L (Fig. 3Ba), and
the linear regression equations were |ΔIPG| þ|ΔIPTH| (μA)¼
3.2461
10  6 þ0.4896 cPG (μM) (R¼0.99908). And the limit of
detection (LOD) was detected and calculated to be 2.4
10  8 mol/
L (S/N¼ 3). Compared with MIP/MWCNTs/GCE (Fig. 3Bb, ΔIPG
(μA) ¼  2.4941
10  6 þ 0.3323 cPG (μM) (R¼ 0.99839)), the sig-
nal-on/off strategy based sensor exhibited an obviously higher
sensitivity and wider linear range for PG detection. Also, compared
with the linear range and LOD reported in previous studies
(Dwiecki et al., 2016; Vikraman et al., 2013; Aguilar-Caballos et al.,
1997), the present sensor showed an excellent detection capacity
to PG. And compared with other electrochemical methods for PG
detection, the imprinted sensor in the our work showed a wider
detection range or a lower LOD, as summarized in Table S1. The
result implied that the sensor had a high sensitivity with the as-
sistance of PTH and MWCNTs.
3.5. Selectivity of MIP/PTH/MWCNTs/GCE
In order to verify the recognition capacity of the sensor toward
template molecule, GA, EG, BHT, BHA and TBHQ were selected as
the analogs for the comparison experiments. The structural for-
mulas of template molecule and each analog were shown in Fig. 4
746 Y. Dai et al. / Biosensors and Bioelectronics 86 (2016) 741–747

Fig. 4. (A) The structural formulas of PG and its analogs. (B) DPV current response recorded on MIP/PTH/MWCNTs/GCE and NIP/PTH/MWCNTs/GCE in PG or each analog
solution with the same concentration of 1.0
10  5 mol/L.

(A). The current responses of each compound recorded on MIP/
PTH/MWCNTs/GCE or NIP/PTH/MWCNTs/GCE were compared in
Fig. 4(B). The current response caused by PG were higher than that
of any other non-template molecules, which can be attributed to
the microenvironment and functional group complementary be-
tween imprinted cavities and the template molecules. The re-
bound PG molecules into imprinted cavities not only could be
oxidized on the sensor but also blocked the electron transmission
to PTH film, resulting in the increase of PG current and the de-
crease of PTH current, respectively. While each compound, in-
cluding PG, displayed weak current response on NIP/PTH/
MWCNTs/GCE due to no imprinted cavities in the NIP film. Thus,
no obvious increase of current of each compound and decrease of
current of PTH could be observed. This phenomena demonstrated
that the prepared MIP/PTH/MWCNTs/GCE could specifically re-
cognize the template molecules from its analogs.
3.6. Reproducibility, regeneration, and stability
be used under room temperature with a relative standard devia-
tion (RSD) of 4.3%. As for relative humidity, no obvious change of
current can be observed whether under 0 °C or 25 °C (Fig. S3B),
which demonstrated the good stability of the sensor under humid
environment. All these results demonstrated that the prepared
sensor possessed acceptable reproducibility, regeneration, and
stability.
3.7. Analysis of PG in real samples
In order to investigate the practical application of the fabricated
sensor, PG in two kinds of edible oil and a kind of cookies were
analyzed by the standard addition method. The results summar-
ized in Table 1 showed the good recoveries varied from 94.7% to
106.3%, indicating that the present sensor was reliable and effec-
tive for the determination of PG in real sample.

The reproducibility of the modified electrode was investigated 4. Conclusions

by determination 1
10  5 mol/L PG solution with five modified
electrodes, which were fabricated under the same condition. The In summary, a novel imprinted electrochemical sensor based
current responses showed a relative standard deviation (RSD) of on sing-on and sign-off strategy was designed and fabricated for
1.2% for five independent measurements. To investigate the re- PG selective recognition and sensitive detection. The sensor could
generation of the prepared sensor, MIP/PTH/MWCNTs/GCE was
recognize PG from its analogs due to the specific rebinding of PG
used to detect 1.0
10  5 mol/L PG for five times with subsequent
molecules into imprinted cavities. The rebound PG caused the
cycles of extraction and measuring operations. According to the
increase of PG current response and the decrease of PTH current
current responses, the RSD was calculated to be 2.3%. Stability is
response. The sign-on from PG and sign-off from PTH were com-
another important parameter for a sensor, several sensors (MIP/
bined as the dual-sign for PG sensitive detection with a wide linear
PTH/MWCNTs/GCE) were put into constant temperature and hu-
range and a low LOD. And the sensor was applied for real samples
midity box. Under the constant relative humidity of 60%, the
determination with satisfactory results.
temperature was set as 0 °C, 4 °C, 25 °C, 30 °C, and 40 °C, respec-
tively. In the same way, when the temperature was kept at 4 °C or
Table 1
25 °C, the relative humidity was regulated to be 20%, 40%, 60%, Application of the sensor for PG determination in real samples.
70%, and 90%, respectively. After storing for two days, these sen-
sors were used to detect 1.0
10  5 mol/L PG under the same Samples Added Detected content by the Recovery (%) RSD (%)
conditions. The results as shown in Fig. S3. It was found that the (mmol/L) sensora (mmol/L)

current remained about the same when the sensor was stored at Oil 1 1 0.96 96.9 3.2
0 °C, 4 °C, and 25 °C. And the current decreased when the tem- 10 10.38 103.8 2.9
perature was set at 30 °C and 40 °C. Under these high tempera- Oil 2 1 1.02 102.6 2.4
tures, we found that the current change of PTH decreased, which 10 9.58 95.8 3.5
Cookies 1 0.99 99.8 3.8
probably was caused by the instability of PTH film under the high
10 9.71 97.1 1.6
temperature. Although the prepared sensor can not keep the
constant current response under high temperature, the sensor can a
Average value of three determination.
 Y. Dai et al. / Biosensors and Bioelectronics 86 (2016) 741–747
Acknowledgments
We greatly appreciate the support of the National Natural Sci-
ence Foundation of China (21005002, 21575003) and Foundation
for Innovation Team of Bioanalytical Chemistry of Anhui Province.
Appendix A. Supplementary material
Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.bios.2016.07.072.
References
Aguilar-Caballos, M.P., Gomez-Hens, A., Perez-Bendito, D., 1997. Anal. Chim. Acta
354, 173–179.
Ahmad, R., Griffete, N., Lamouri, A., et al., 2015. Chem. Mater. 27, 5464–5478.
Ahammad, A.S., Rahman, M.M., Xu, G.R., 2011. Electrochim. Acta 56, 5266–5271.
Andre, C., Castanheira, I., Cruz, J.M., et al., 2010. Trends Food Sci. Technol. 21,
229–246.
Andreu-Navarro, A., Fernández-Romero, J.M., Gómez-Hens, A., 2011. Anal. Chim.
Acta 695, 11–17.
Bai, M., Huang, S.B., Xu, S.Y., et al., 2015. Anal. Chem. 87, 2383–2388.
Cai, X.Q., Li, J.H., Zhang, Z., et al., 2014. ACS Appl. Mater. Interfaces 6, 305–313.
Cui, M., Huang, J.D., Wang, Y., et al., 2015. Biosens. Bioelectron. 68, 563–569.
747
Dempsey, E., Diamond, D., Collier, A., 2004. Biosens. Bioelectron. 20, 367–377.
Ding, Y.P., Liu, W.L., Wu, Q.S., et al., 2005. J. Electroanal. Chem. 575, 275–280.
Dwiecki, K., Tomczyk, Ł., Nogala-Kałucka, M., et al., 2016. Eur. J. Lipid Sci. Technol. .
http://dx.doi.org/10.1002/ejlt.201500453
Gao, Q., Cui, X., Yang, F., et al., 2003. Biosens. Bioelectron. 19, 277–282.
Guan, Y., Chu, Q., Fu, L., et al., 2006. Food Chem. 94, 157–162.
Hall, C.A., Zhu, A., Zeece, M.G., 1994. J. Agric. Food Chem. 42, 919–921.
Kan, X., Zhou, H., Li, C., et al., 2012. Electrochim. Acta 63, 69–75.
Kong, B., Selomulya, C., Zheng, G.F., et al., 2015. Chem. Soc. Rev. 44, 7997–8018.
Li, B.B., Zhou, Y.S., Wu, W., et al., 2015a. Biosens. Bioelectron. 67, 121–128.
Liu, H., Wang, G., Chen, D., et al., 2008. Sens. Actuators, B 128, 414–421.
Li, Y.C., Liu, Y., Liu, J., et al., 2015b. Sci. Rep., 5. http://dx.doi.org/10.1038/srep07699.
Li, Y., Li, Y.J., Hong, M., et al., 2013. Biosens. Bioelectron. 42, 612–617.
Mao, H., Liang, J.C., Zhang, H.F., et al., 2015. Biosens. Bioelectron. 70, 289–298.
Morales, M.D., González, M.C., Reviejo, A.J., et al., 2005. Microchem. J. 80, 71–78.
Patra, S., Roy, E., Das, R., et al., 2015. Biosens. Bioelectron. 73, 234–244.
Perrin, C., Meyer, L., 2002. Food Chem. 77, 93–100.
Qi, H., Zhang, C., 2005. Electroanalysis 17, 832–838.
Roy, E., Patra, S., Tiwari, A., et al., 2016. Biosens. Bioelectron. . http://dx.doi.org/
10.1016/j.bios.2016.02.056
Vlatakis, G., Andersson, L.I., Mosbach, K., 1993. Nature 361, 645–647.
Vikraman, A.E., Rasheed, Z., Rajith, L., et al., 2013. Food AnalMethod 6, 775–780.
Wang, Q.L., Chen, M.M., Zhang, H.P., et al., 2016. Sens. Actuators, B 222, 254–269.
Wulff, G., 1995. Angew. Chem. Int. Ed. Engl. 34, 1812–1832.
Wyatt, D.M., 1981. J. Am. Oil Chem. Soc. 58, 917–920.
Xiao, D.L., Dramou, P., Xiong, N.Q., et al., 2013. J. Chromatogr. A 1274, 44–53.
Xu, Y., Jiang, Y., Yang, L., et al., 2005. Chin. J. Chem. 23, 1665–1670.
Yang, P., Zhu, Q., Chen, Y., et al., 2009. J. Appl. Polym. Sci. 113, 2881–2886.
Yang, Y.K., Cao, Y.Y., Wang, X.M., et al., 2015. Biosens. Bioelectron. 64, 247–254.
Zhai, D.Y., Liu, B.R., Shi, Y., et al., 2013. ACS Nano 7, 3540–3546.