Biosensors and Bioelectronics 99 (2018) 251–258

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

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

Carbon dots based photoelectrochemical sensors for ultrasensitive detection MARK

of glutathione and its applications in probing of myocardial infarction
Zhengping Lia,1, Jun Zhangb,1, Yunxiao Lia, Shuang Zhaoa, Peixin Zhangb, Yue Zhangc,
⁎
Jinshun Bid, Guohua Liua,e, Zhao Yuea,e,
a
Department of Microelectronics, Nankai University, Tianjin 300350, China
b
Department of Biology, Nankai University, Tianjin 300071 China
c
Department of Physiology & Pathophysiology, Tianjin Medical University, Tianjin 300070 China
d
Key Laboratory of Microelectronic Devices & Integrated Technology, Institute of Microelectronics, Chinese Academy of Sciences, Beijing 10010 China
e
Tianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Tianjin 300350, China

A R T I C L E I N F O A BS T RAC T

Keywords: In this work, photoelectrochemical (PEC) sensors based on carbon dots (CDs) were developed for ultrasensitive
Carbon dots detection of glutathione (GSH) without additional catalysts. In this PEC sensing system, CDs exhibited both
Photoelectrochemistry photoelectric and catalytic properties. Silver nanoparticles (AgNPs), graphene oxide (GO), and mesoporous
Glutathione silica (MS) were introduced in order to enhance the sensing properties of CDs for GSH. Among the different
Hybrid nanocomposites
hybrid nanocomposites, CDs@MS based PEC sensors exhibited the best sensing properties: the sensitivity and
Myocardial infarction
Mesoporous silica
limit of detection (LOD) for GSH were found to be 57.6 nA μM− 1 and 6.2 nM (S/N = 3), respectively, in the
linear range 0.02–4 μM. In addition, the developed PEC sensors showed a high selectivity for GSH even with
interferences of other biological thiols and amino acids. The PEC sensor was successfully applied for GSH
detection in human serum and probing of myocardial infarction (MI) conditions by estimating the amount of
GSH in the myocardial cells of mice, which had been treated with different ischemia/ischemia-reperfusion
times. These results indicated that the CDs based hybrid nanocomposites are promising candidates for the
development of PEC biosensors with enhanced sensing performances.

1. Introduction Compared to bulk materials, nanomaterials exhibit unique physical

Glutathione (GSH, γ-glutamyl-cysteinyl-glycine) is an important
tripeptide thiol, which acts as an antioxidant and whose intracellular
concentration is an indicator of oxidative stress. It is well known that
abnormal levels of GSH are closely related to ailments, such as cancer,
liver damage, AIDS, and cardiovascular disease (Han et al., 2008; Mills
et al., 2000; Townsend et al., 2003; Vairetti et al., 2007). As the changes
in cellular concentration of GSH are normally very small, a highly
sensitive and selective detection of GSH in biological samples is of great
importance for the early diagnoses of the above mentioned diseases (Hao
et al., 2015). Myocardial infarction (MI), a common heart disease, is
induced by continuous myocardial ischemia and hypoxia, which can
trigger free radical production. The substantial production of free radicals
following myocardial ischemia usually leads to the depletion of GSH.
Therefore, it may be concluded that heart healthy is closely associated
with the level of GSH, which can be used for indirect diagnosis of acute
myocardial infarction in its early stages (Li et al., 2009).
and chemical properties because of their small size, and can be used for
the detection of biological substances such as GSH with better sensing
performances. Among the different analytical methods available for
this purpose, PEC sensors based on fluorescent nanomaterials are
excellent detection tools because of their significant advantages such as
lower limits of detection, lower potential consumption, easier opera-
tion, and easier extension to light-addressable sensors for multichannel
detections (Yue et al., 2013; Zhang et al., 2013; Zhao et al., 2014; W.W.
Zhao et al., 2016). However, most fluorescent nanomaterials that are
currently used are quantum dots (QDs) and these often contain Cd or
Pb heavy metal ions. Thus, the applications of QDs based PEC sensors
are limited because of their toxicity in fabrication processes and
biological detections, especially in cell detections. Moreover, owing to
complex fabrication processes and limitation of material properties,
these nanomaterials based PEC sensors do not exhibit good enough
sensing performances, and the limit of detection is also not very high
(Yue et al., 2013). Therefore, new fluorescent nanomaterials with lower

⁎
Corresponding author at: Department of Microelectronics, Nankai University, Tianjin 300350, China.
E-mail address: yuezhao@nankai.edu.cn (Z. Yue).
1
Equal contribution to this work.

http://dx.doi.org/10.1016/j.bios.2017.07.065
Received 12 February 2017; Received in revised form 16 July 2017; Accepted 26 July 2017
Available online 27 July 2017
0956-5663/ © 2017 Elsevier B.V. All rights reserved.
Z. Li et al.
Fig. 1. (A) Absorption spectrum of CDs. (B) Fluorescence spectra of CDs, CDs/AgNPs, CDs/GO, and CDs@MS with optimized concentrations of AgNPs, GO, and MS, respectively. (C)
AFM topography image of CDs with the height profile. (D) HRTEM image of the CDs@MS nanocomposite. The inset shows the HRTEM image of individual carbon dots on the surface of
MS.
toxicity and better sensing properties are required for developing PEC
sensors for GSH detection in cells.
Carbon dots (CDs) are “green” fluorescent nanomaterials and have
been utilized in this work to build a PEC sensor for ultrasensitive GSH
detection. CDs display unique optical properties, high chemical stabi-
lity, low environmental risk, and excellent biocompatibility. Owing to
these advantages, CDs based nanomaterials are now being used to
develop biosensors for different analytical methods (Zuo et al., 2015;
Wang and Qiu, 2016). To date, several CDs based optical sensors, such
as fluorescent, electrochemiluminescent, and colorimetric sensors, for
the detection of GSH have been developed (Cai et al., 2015; Niu et al.,
2015; Shi et al., 2014; Y. Xu et al., 2016). All these sensors follow the
same sensing mechanism: MnO2, gold nanoparticles, or other metal
ions either affect or quench the fluorescence intensity of CDs. After
GSH is added to the test solution, it either reacts or competes with
these nanomaterials, leading to a change of fluorescence intensity of
CDs. The concentration of GSH can be determined by measuring the
changes in optical properties of the CDs. In these detection systems,
CDs were only used as optical probes, which implies that the photo-
electric and catalytic properties of CDs were not utilized efficiently.
However, to the best of our knowledge, there are only a limited number
of studies that have used CDs for building PEC sensors, besides a PEC
cytosensor that utilizes gold nanoparticle-enhanced CDs (CDs-AuNPs-
Cys/chitosan/folic acid/bovine serum albumin) for the detection of
Hela cells (Liu et al., 2015). In this PEC biosensor, CDs played the role
of fluorescent materials to provide photoelectric properties. The
252
Biosensors and Bioelectronics 99 (2018) 251–258
sensing mechanism on Hela cells was mainly due to folic acid rather
than CDs. Moreover, no CDs based PEC sensors have been reported for
the detection of GSH. Therefore, by combining the advantages of CDs
and PEC, CDs based PEC sensors were developed for ultrasensitive
detection of GSH with a better limit of detection (LOD) and higher
selectivity. This is the first report where CDs exhibit both photoelectric
and catalytic properties in a PEC sensing system. Moreover, silver
nanoparticles (AgNPs), graphene oxide (GO), and mesoporous silica
(MS) were introduced as hybrid nanocomposites in order to enhance
the catalytic properties of CDs and obtain better GSH sensing
performances. It has been previously shown that these three nanoma-
terials can improve the catalytic properties of other nanomaterials
(Liang et al., 2014; Zhao et al., 2017; Fang et al., 2013).
Finally, the practicality of our PEC sensors was proven by the
estimation of GSH concentration in human serum and in myocardial
cells of mice, which had been treated with different ischemia/ischemia-
reperfusion times.
2. Materials and methods
2.1. Reagents and apparatus
All chemicals were of analytical grade and are listed in the
Supplementary Information (SI). All aqueous solutions were prepared
using ultrapure water with electrical resistance of 18.2 MΩ cm (25 °C).
The absorption and fluorescence measurements of the different
Z. Li et al.
nanomaterials were recorded by a Shimadzu RF-5301 spectrofluor-
ophotometer (Japan). Atomic force microscopy (AFM) images were
recorded by Bruker Dimension Icon (Bruker, USA). Ultra-high resolu-
tion transmission electron microscopy (HRTEM) images were obtained
on an ultra-high resolution JEM-2100 (JEOL Ltd., Japan) instrument
operating at 200 kV.
2.2. Synthesis processes and characterizations of different
nanomaterials
CDs were synthesized following a slightly modified literature
procedure (Li et al., 2013). The concentration of the CDs solution
was 1 mg mL− 1. The solution of AgNPs was prepared with [Ag(NH3)2]
OH reagent as a precursor (Ye et al., 2014) and treated with poly-L-
lysine hydrobromide (PLL). The final concentration of the AgNPs
solution was 0.023 mg mL− 1. GO was prepared from 20 nm carbon
in diameter fibres following the KMnO4 oxidation method (Chou et al.,
2012) and subsequent treatment with ethylenediamine. The final
concentration of GO was 2 mg mL− 1. MS was synthesized according
to the literature reported method (Xiong et al., 2014). CDs/AgNPs and
CDs/GO nanocomposites were also synthesized by typical procedures
(Marsich et al., 2012; S. Zhao et al., 2016; Wu et al., 2014). The
electrostatic blending methods of CDs and either AgNPs or GO to form
hybrid nanocomposites have been reported by other research groups
(Marsich et al., 2012; S. Zhao et al., 2016, 2017). The CDs@MS
nanocomposite was synthesized by using electrostatic interactions,
according to a slightly modified literature procedure (Croissant et al.,
2016). Further details of the synthesis of different nanomaterials are
given in the SI.
The characterization results of CDs, AgNPs, GO, MS, CDs/AgNPs,
CDs/GO and CDs@MS hybrid nanocomposites have also been de-
scribed in manuscript and SI. The absorption and fluorescence
measurements of different nanomaterials are shown in Fig. 1A and
B, and S1 and S2. The AFM images of CDs, and different hybrid
nanocomposites are shown in Fig. 1C and Fig. S3. The HRTEM images
of CDs@MS are given in Fig. 1D and Fig. S4.
2.3. Fabrication of the photoelectrochemical sensor chips
In the first step to prepare the sensors, the Au electrodes (200 nm
Au/20 nm Ti/500 µm glass, contact area of 0.5 cm2) were sonicated in
acetone, ethanol and ultrapure water for 5 min sequentially as cleaning
processes. In the second step, the cleaned Au electrodes were incubated
in a 20 mM solution of sodium 3-mercapto-1-propanesulfonate (MPS)
in 16 mM H2SO4 for 30 min, in order to form a self-assembled
monolayer of MPS on the surface of the Au electrodes (MPS/AuE).
The self-assembled method used in this study is based on the
electrostatic blending methods. For immobilization of CDs, CDs/
AgNPs and CDs/GO, the negatively charged MPS/AuE was then
incubated in poly(diallyldimethylammonium chloride) (PDDA, posi-
tively charged polyelectrolyte) solution for 15 min. After rinsing with
ultrapure water and drying with N2, PDDA formed a monolayer on the
surface of the Au electrodes (PDDA/MPS/AuE), which switched the
surface charge to positive. The modified positive electrode was then
covered with negatively charged CDs, CDs/AgNPs or CDs/GO hybrid
nanocomposites solutions for 30 min. Subsequently, the solvent was
removed by spin coating to form different sensor chips, namely, CDs/
PDDA/MPS/AuE, CDs/AgNPs/PDDA/MPS/AuE, and CDs/GO/
PDDA/MPS/AuE. The electrodes were then rinsed with water and
dried with N2. Since positive charges already existed on the surface of
CDs@MS, the as-prepared MPS/AuE was covered directly with CDs@
MS for 30 min and the solvent was removed by spin coating. After
rinsing with water and drying with N2, the CDs@MS sensor chips
(CDs@MS/MPS/AuE) were well-fabricated. A schematic diagram of
the different fabrication processes is illustrated in Fig. S5.
253
Biosensors and Bioelectronics 99 (2018) 251–258
2.4. Photoelectrochemical measurement system
A potentiostat CHI 604D (CH Instruments, USA) and lock-in
amplifier (SR830, USA) were used for the photoelectrochemical
experiments with a home-made electrochemical cell. Ag/AgCl and Pt
wire were used as the reference and counter electrodes, respectively.
The fabricated sensor chips described above were used as the working
electrodes. A mounted high-power LED (M365L2: 365 nm and
190 mW, Thorlabs, USA) was used as the light source. Fetched from
transient I–t curves of lock-in amplifier, all data points are the mean
values of the largest photocurrent amplitude from three independent
measurements and the corresponding standard deviations are given as
error bars. The schematic diagram and other details of the home-built
photoelectrochemical measurement system are shown in Fig. S6.
2.5. Biological samples preparation
In order to prove the viability of our proposed PEC sensors for
practical applications, human serum was selected as a more realistic
environment. The serum samples were spiked with additional known
concentrations of GSH (0.2, 1, and 2 μM). The GSH content in the
serum samples was determined from the regression equation of a
standard calibration curve for GSH of CDs@MS based PEC sensors.
For further validation, the MI model of mice was introduced. Mice
were anesthetized by isoflurane and their hearts were exposed via a left
thoracotomy. MI was induced by permanent ligation of the left anterior
descending artery. Mice were randomly assigned to five groups as
follows: myocardial ischemia for 30 min, myocardial ischemia for
60 min, myocardial ischemia for 30 min followed by 2 h of ischemia-
reperfusion, myocardial ischemia for 60 min followed by 2 h of
ischemia-reperfusion, and a blank control group. At the appropriate
time points, mice were sacrificed and their hearts were removed and
treated according to the following procedure. The hearts were first
washed with physiological saline, snap frozen in liquid nitrogen, and
then fully homogenized with the M reagent from the GSH detection kit.
The resulting mixture was centrifuged at 10,000 rpm for 10 min at 4 °C
in order to remove plasma proteins and other molecules. After
myocardial infarction, part of the GSH from myocardial cells may have
already entered into the blood. MI conditions can be evidenced by the
presence of GSH in myocardial cells as well as in blood. For better light
absorption of PEC sensors, the transparent supernatant was collected
for GSH measurement. Besides using our designed PEC sensors, the
level of GSH in each group were also determined by using the
commercial GSH detection kit and measured with a multimode plate
reader at the wavelength of 412 nm as references.
3. Results and discussions
3.1. Characterizations of the different hybrid nanocomposites
The absorption spectrum of CDs shows a peak at ca. 360 nm, as
illustrated in Fig. 1A. Thus, an LED with 365 nm wavelength was
chosen as the excitation light source for PEC sensing. A strong
fluorescence peak of CDs was observed at 455 nm when they were
excited at 365 nm. This peak at 455 nm does not exhibit any shift when
optimal concentrations of AgNPs, GO, and MS were mixed with CDs to
obtain CDs/AgNPs, CDs/GO and CDs@MS hybrid nanocomposites, as
shown in Fig. 1B. Furthermore, the fluorescence intensity does not
show any obvious variation from that of CDs to different hybrid
nanocomposites, which confirms that the introduced nanomaterials
do not show any quenching processes towards CDs. Further character-
ization details of CDs mixed with different concentrations of AgNPs,
GO, and MS are provided in Fig. S1 and S2.
The AFM topography images of CDs and CDs@MS (given in Fig. 1C
and S3) show that these nanomaterials were 2.4 nm and 68 nm in
height, respectively. In addition, the HRTEM image of CDs@MS reveal
Z. Li et al.
Fig. 2. (A) I–V curves of CDs based PEC sensors in phosphate buffer. (B) Transient I–t curves of CDs/SAM/AuE chips in presence of GSH at + 500 mV. (C) A PEC model diagram of the
CDs@MS/SAM/AuE sensing system for GSH. (D) The dependency of photocurrents on GSH concentrations and applied potentials. The inset is the fitted curve in the linear range of
0.02–4 μM at + 500 mV.
that they are well-distributed without apparent aggregation and show
an average diameter of 70 nm (Fig. 1D). These results are consistent
with the AFM images. The image in Fig. S4 was obtained from a single
CDs@MS and shows a uniform dispersion of CDs inside MS in the as-
prepared CDs@MS hybrid nanocomposites. The average diameter of
CDs was found to be 2.4 nm with a standard deviation of 0.4 nm. The
inset of Fig. 1D shows the lattice fringes of CDs with a spacing of
0.34 nm, which is close to the (002) facet of graphite. The electrostatic
self-assembled monolayer (SAM) method can enforce a good connec-
tion between AgNPs/GO/MS and CDs, which was confirmed by TEM
and AFM results after centrifugation. Based on the AFM and TEM
characterization results, it can be concluded that CDs were well-formed
into hybrid nanocomposites with AgNPs, GO and MS, in addition to
being well-ordered and dispersed on the electrodes.
3.2. CDs based photoelectrochemical sensing mechanism for the
detection of GSH
In order to study the photoelectric properties of PEC sensors based on
CDs, the photocurrents of these devices at different applied potentials in
phosphate buffer were measured, as shown in Fig. 2A. The magnitude of
the generated photocurrent was found to be linearly dependent on the
applied potentials with high stability, which suggested that the CDs
provided a stable energy band and could replace quantum dots to achieve
photoelectric conversion for sensing. The sensing mechanism of fluor-
escent nanomaterials based PEC sensors has been previously discussed in
detail (Yue et al., 2013; Zhang et al., 2015). As shown in Fig. 2C, CDs can
be excited to generate photoinduced electron-hole pairs under illumina-
tion. The electrons present in the LUMO can cross the SAM to the gold
electrode (Ke) via tunnelling processes. Meanwhile, the electrons from
AuE can tunnel to the HOMO of CDs to fill the photoexcited holes (Kb).
254
Biosensors and Bioelectronics 99 (2018) 251–258
The differences between the two electron transfer processes generate an
observable photocurrent in a phosphate buffer in the absence of oxidizing
and reducing substances. As the positions of HOMO and LUMO are fixed
and the Fermi level of AuE is determined by the applied potential, Ke and
Kb can be determined directly by the applied potentials (Zhang et al.,
2015). No photocurrent appears when these two processes are in
dynamic equilibrium (Ke = Kb). This applied potential is termed as
formal potential, which is about − 100 mV for the CDs based PEC system.
When a potential larger than − 100 mV (Ke > Kb) is applied, more
electrons tunnel from the LUMO of CDs to the gold electrode, leading
to a positive photocurrent. When a potential smaller than − 100 mV is
applied (Ke < Kb), the photocurrent will be negative for the same reason.
These CDs based PEC sensors were used for the detection of GSH.
The photocurrent-time response curves (I–t) under different concen-
trations of GSH at + 500 mV and − 500 mV are shown in Fig. 2B and
S7, respectively. The positive photocurrents promptly increased with
increasing concentration of GSH from 0 to 3 μM at + 500 mV. As far as
we know, in a PEC sensing system based on REDOX (reduction and
oxidation) reactions, molecules have to be catalysed first before they
are oxidized or reduced by photoexcited hole-electron pairs formed in
the fluorescent nanomaterials. Thus, the result of enhanced photo-
current amplitude with increasing GSH concentrations at positive
potentials indicated that GSH undergoes catalytic oxidation by our
designed PEC sensors. GSH is the most abundant antioxidant among
the various biological molecules. It possesses a thiol group in its
structure and with proper catalysts, these thiol groups can be oxidized
to complete dehydrogenation. This reaction converts GSH into glu-
tathione disulfide (GSSG) (Ndamanisha et al., 2009), detailed explana-
tions described in SI. However, there is no increase in negative
photocurrent at − 500 mV in presence of different concentrations of
GSH (Fig. S7). These results are consistent with the fact that GSH is
Z. Li et al.
difficult to be reduced and its detection can only be achieved by
oxidation. As shown in Fig. 2C, reducing agents such as GSH can
undergo catalytic oxidation by the photoexcited holes in the HOMO of
CDs into GSSG (termed as Kv). The Kv process competes with the Kb
process for photoexcited holes. The tunnelling processes of Kv and Ke
with same direction are predominant and both produce positive
photocurrent. Therefore, the photocurrent generated can be calculated
as Kv + Ke – Kb. Larger concentrations of GSH can speed up Kv, leading
to a larger amplitude of the photocurrent. Thus, the concentration of
GSH can be determined by the enhancement of positive photocurrent
amplitude at an applied voltage larger than − 100 mV. It should be
noticed that not all fluorescent nanomaterials based PEC sensing
system can oxidize GSH directly, such as TiO2/CdS hybrid nanomater-
ials based PEC sensors (Wang et al., 2009). As GSH cannot be catalysed
by TiO2/CdS, it is not oxidized by the photoinduced holes. In our
proposed PEC sensors, CDs exhibited catalytic activities toward GSH.
They were able to directly catalyse and open GSH to be oxidized by the
photoexcited holes of CDs. According to the structure, the catalytic
properties of CDs will be from ligands or carbon core. Previous reports
have shown that hydroxyl and carboxyl groups, which are the main
surface groups of CDs, were not responsible for their catalytic proper-
ties toward GSH (Zhao et al., 2015). By comparisons, the only
explanation for the GSH sensing ability of our designed CDs based
PEC sensors is the catalytic properties of carbon core. It has been
reported that carbon dots show catalytic activity in water oxidation
(Datta et al., 2016), which implies that CDs can be used as catalysts.
Carbon nanomaterials such as carbon nanotubes (Tang et al., 2006),
graphene oxide (Yuan et al., 2013), reduced graphene oxide (Li et al.,
2016), graphene quantum dots (Vinoth et al., 2017), and ordered
mesoporous carbon (Ndamanisha et al., 2009) with different ligands
showed catalytic activity toward GSH. Based on these references and
our PEC experimental results, it can be concluded that CDs showed
catalytic activity toward GSH because of the carbon nanomaterials core
and not the ligands around CDs. Thus, CDs exhibit both photoelectric
and catalytic properties toward GSH simultaneously in our designed
PEC sensors.
In view of the formal potential (–100 mV) determined from Fig. 2A,
the oxidation reaction can occur when the applied potential is above −
100 mV (Ke > Kb). In this work, potentials of + 100, + 300, and +
500 mV were applied for oxidizing GSH. The relationship between
photocurrents and concentrations of GSH from 0 to 10 μM at different
potentials is shown in Fig. 2D. It is clear that the sensitivity at +
500 mV is far better than that at + 300 mV or + 100 mV. The reason for
this trend is described in SI. After overall considerations, + 500 mV is
chosen as the optimal potential.
A saturation phenomenon of the photocurrent is observed when the
concentration of GSH is larger than 4 μM. A good linear relationship
was obtained for concentration of GSH ranging from 0.02 to 4.0 μM
with the sensitivity of 35.8 nA μM− 1 at + 500 mV (insert in Fig. 2D).
The determination coefficient R is 0.9965 and the limit of detection
estimated as 12.6 nM (S/N = 3) was better than that previously
reported for other CDs based GSH sensors (Table S1).
3.3. Enhanced sensing properties by hybrid nanocomposites
Based on the sensing mechanism described above, GSH was
catalysed by CDs and oxidized into GSSG by holes from photoexcited
CDs. Therefore, the PEC sensing properties of CDs towards GSH are
determined by both photoelectric properties of CDs and the catalytic
performances of CDs toward GSH. In this work, different nanomater-
ials with different dimensions (AgNPs, GO, and MS) were mixed with
CDs to obtain the hybrid nanomaterials CDs/AgNPs, CDs/GO, and
CDs@MS, respectively, with enhanced sensing properties of CDs based
PEC sensors.
The CDs/AgNPs, CDs/GO, and CDs@MS hybrid nanomaterials
were first tested in phosphate buffer to study their photoelectric
255
Biosensors and Bioelectronics 99 (2018) 251–258
properties. Fig. S8 shows that all photocurrent amplitudes did not
show an obvious change with increasing concentrations of AgNPs, GO,
and MS. This suggests that AgNPs, GO, and MS cannot enhance the
photoelectric properties of our designed CDs based PEC sensors. The
fluorescence spectra of nanomaterials in Fig. 1B also confirmed that
there was no obvious diminishing in fluorescence intensity when
different concentrations of either AgNPs, GO, or MS were introduced
into CDs. These results are contrary to previous reports of CDs/GO,
TiO2/AgNPs, and AuNCs/GO nanomaterials (Yu et al., 2014; Bian
et al., 2013; S. Zhao et al., 2016): GO and AgNPs enhanced the
photoelectric properties of CDs, TiO2 and AuNCs. According to the
synthesis method used (Li et al., 2013), the reason why there was no
obvious decrease in fluorescence intensity was because of the presence
of the ligands of protein surrounding the surface of the CDs. The
protein could change the properties of the surface groups and enlarge
the distances between CDs and other nanomaterials, which impeded
the electron transfer processes to some extent. Therefore, the ligands
around CDs had a big influence on the enhancement of photoelectric
properties.
However, Fig. 3 shows that different hybrid nanomaterials provided
better sensitivities than CDs, which suggests that the sensitivity toward
GSH had obviously been improved. As described above, the sensing
properties of PEC sensors were related to two parts: photoelectric
properties and catalytic performances (S. Zhao et al., 2016). As AgNPs,
GO and MS failed to enhance the photoelectric properties, the only
reason for the improvement of the sensing properties of CDs toward
GSH was because the catalytic properties of CDs toward GSH were
enhanced. Previous reports have shown that AgNPs can enhance the
catalytic performance of other nanomaterials because of a large
surface-to-volume ratio (Gao et al., 2015). GO is capable of catalysing
GSH via electrochemical methods (Yuan et al., 2013) as GO provides a
large amount of oxygen-containing functional groups, which play an
important role in the electrocatalytic oxidation (Dreyer et al., 2009).
Meanwhile, GO can improve the catalytic effect of other nanomaterials
such as gold nanoclusters toward H2O2 (Tao et al., 2013; S. Zhao et al.,
2016). Unlike GO, MS cannot catalyse GSH directly (S. Xu et al., 2016).
However, MS has an enormous surface-to-volume ratio because of the
large number of pores on the surface of silicon dioxide. Incorporation
of nanoparticles into MS has been persistently pursued in order to
develop novel hybrid nanocomposites for a wide range of applications
such as catalysis, separation, and imaging (Lee et al., 2011). These
pores not only provide ideal places for CDs to adhere, which can
significantly increase the contact area between CDs and analytes, but
also effectively protect them against aggregation, which helps enhance
their catalytic properties and stabilities. Moreover, it should be noted
that MS is almost transparent and therefore does not disturb the light
absorption capability of CDs in the PEC sensing processes.
From these results, it can be concluded that the enhancement in
sensitivity enhancement is due to the enhanced catalytic properties, no
matter if these introduced nanomaterials show direct catalytic activity
for GSH or they enhanced the catalytic performance of CDs towards
GSH. It also can be seen that the concentrations of AgNPs, GO, and MS
greatly affect the sensitivity of the PEC sensors toward GSH in
Fig. 3(A)–(C): the sensitivity increased with the increasing concentra-
tion of AgNPs, GO, and MS and then decreased. The reasons for this
decrease in sensitivity were discussed in SI and previous works (Pu
et al., 2013; S. Zhao et al., 2016). The optimized concentrations of
AgNPs, GO, and MS in CDs for the best sensing properties were
determined to be 1.84 μg mL− 1, 0.08 mg mL− 1, and 2.5 mg mL− 1,
respectively. With optimized concentrations, sensitivities of PEC
sensors were enhanced from 35.8 to 49.9 nA μM− 1 by CDs/AgNPs,
46.9 nA μM− 1 by CDs/GO, and 57.6 nA μM− 1 by CDs@MS at +
500 mV in the range 0.02–4 μM. Owing to their porous structure,
CDs@MS based PEC sensors exhibited the largest sensitivity, as shown
in Fig. 3D. Moreover, the porous structure enhanced the stability of
CDs based PEC sensors against aggregation, leading to smaller devia-
Z. Li et al. Biosensors and Bioelectronics 99 (2018) 251–258

Fig. 3. The dependency of sensitivities on the concentrations of (A) AgNPs, (B) GO, (C) MS in linear range of 0.02–4 μM for GSH detection. (D) The dependency of photocurrents of
different PEC sensors based on hybrid nanocomposites on the GSH concentration. Optimum concentrations of AgNPs, GO, and MS were added to CDs.

tions in the measurement. Hence, the CDs@MS based PEC sensors

exhibited the best LOD of 6.2 nM (S/N = 3) compared to 9.0 nM and
9.6 nM for CDs/AgNPs and CDs/GO, respectively.

3.4. Selectivity, stability and reproducibility of CDs and CDs@MS

based PEC sensors for GSH detection

In order to evaluate the selectivity of the designed PEC sensors,

they were studied in the presence of different electroactive reducers,
such as a ten-fold amount of homocysteine (Hcy), cysteine (Cys),
tyrosine (Tyr), threonine (Thr), valine (Val), phenylalanine (Phe), and
aspartic acid (Asp), and a 100-fold amount of amino acids (AA),
dopamine (DA), leucine (Leu), glutamic acid (Glu), arginine (Arg),
lysine (Lys), tryptophan (Trg), histidine (His), and serine (Ser). As
shown in Fig. 4 and Fig. S9, both CDs@MS and CDs based PEC sensors
showed no obvious photocurrent enhancement due to the potential
interference of different reducing agents. The good selectivity of CDs
toward GSH can be explained by their catalytic properties. CDs do not
show catalytic activity toward AA, DA, or other reducing substances,
and these compounds cannot be catalysed and oxidized by the
photoexcited CDs as GSH. Previously reported carbon nanomaterials
also did not show any catalytic activity toward these reducers (Tang
et al., 2006; Yuan et al., 2013; Li et al., 2016; Vinoth et al., 2017).
These results reflect the outstanding selectivity of CDs and CDs@MS
based PEC sensors for GSH.
The reproducibility and stability of the prepared sensors were also
investigated. The relative standard deviation (RSD) was calculated to
be 5.5% from five different sensors at 1 μM of GSH, which indicated
good fabrication reproducibility of the sensors. In order to estimate the
long-term stability, the photocurrent change toward 1 μM GSH at +
500 mV was measured every day. The sensors were stored in phosphate
Fig. 4. The photocurrent responses of CDs@MS based PEC sensors in the presence of
different biologically relevant species (1 μM for GSH, 10-fold and 100-fold concentra-
tions of other potential interferences).
buffer at room temperature when not in use. The CDs@MS based
sensors retained 93% of the initial photocurrent response after two
weeks, while the CDs based PEC sensors retained 90% of the initial
photocurrent response. These results suggested that MS enhanced both
the sensing properties of CDs and the stability of the PEC sensors.
Thus, it was evident that the PEC sensors could be potentially used for
practical purposes.
3.5. Preliminary applications in human serum samples and probing
MI using CDs@MS based PEC sensors
In order to demonstrate the analytical performances of the pro-
posed PEC sensors in complex biological environments, GSH concen-

256
Z. Li et al. Biosensors and Bioelectronics 99 (2018) 251–258

Table 1
Analytical results of the determination of GSH in human serum using CDs@MS based
PEC sensors.

Sample Added (μM) Found (μM) Recovery (%) RSD (%, n = 3)

1 0 0.31 – 4.6
2 0.2 0.47 92.2 5.3
3 1 1.28 97.7 4.4
4 2 2.18 94.4 5.0

trations in human serum were determined by PEC sensors based on

CDs@MS hybrid nanocomposites. The diluted serum samples were
spiked with GSH concentrations of 0.2, 1, and 2 μM. Recoveries of the
known spiked amounts of GSH in human serum were between 92.2%
and 97.7% with ~5% RSD (shown in Table 1). These results confirmed
the reliability of the proposed sensors for the determination of GSH in
human serum. They also illustrated that the output changes of the
prepared PEC sensors were because of GSH and not the other
substances present in human serum.
In order to further prove the practicability of our designed CDs
based PEC sensors, a myocardial ischemia/ischemia-reperfusion model
of mice was designed. The GSH concentrations were determined in the
supernatants of homogenized samples from the hearts of mice that had
been treated with different times of myocardial ischemia and ischemia-
reperfusion. It is known that the concentrations of GSH can keep
decreasing with prolonged ischemia time. Furthermore, reperfusion
after ischemia can also decrease the GSH concentration in myocardial
cells (Braunwald and Kloner, 1985). Therefore, it is expected that with
longer myocardial ischemia and ischemia-reperfusion time duration,
the GSH concentration in the cells will be smaller. Fig. 5A clearly
indicates that the concentrations of GSH from CDs@MS based PEC
sensor kept decreasing as the time of myocardial ischemia increased (0,
30, and 60 min). Moreover, the GSH concentrations of samples where
the mice had been subjected to ischemia for 30 or 60 min and followed
by 2 h of ischemia-reperfusion showed a clear downtrend via CDs@MS
based PEC sensing (Fig. 5B and C), suggesting that the GSH concen-
tration decreased again after ischemia-reperfusion. These results were
in good agreement with previous biological studies and proved that the
PEC sensors designed in this work can be used to investigate MI
conditions accurately by measuring the GSH concentrations. In
comparisons with a commercially available GSH detection kit, our
designed CDs based PEC sensors showed more advantages. First, the
GSH detection kit shows a very large error range, which produces a big
overlap among different MI periods. Thus, the results from a GSH
detection kit are often quite inaccurate. Second, owing to the low
sensitivity of the GSH detection kit, small GSH concentration changes
cannot be monitored very well, especially as shown in Fig. 5B and C.
Meanwhile, the PEC sensors were much more capable of distinguishing
among the different GSH concentrations from different MI conditions.
Therefore, it can be concluded that our PEC sensors displayed higher
sensitivities and better stabilities for the detection of GSH compared to
a commercial optical GSH detection kit. These CDs based PEC sensors Fig. 5. (A) The dependency of GSH concentrations on different myocardial ischemia
demonstrate the potential for applications in probing of early-stage time periods from a commercial GSH detection kit and the developed PEC sensors. The
myocardial cells by an accurate detection of GSH. Moreover, these dependency of GSH concentrations on different ischemia-reperfusion time after (B)
sensors are easy to operate, exhibit rapid detection, low power 30 min and (C) 60 min of myocardial ischemia.

consumption, and can be recycled, as opposed to a commercial GSH

detection kit. This may enable an easy diagnostic probing of MI via
GSH detection in blood.
4. Conclusion
In summary, CDs based PEC sensors were successfully designed for
the detection of GSH which combined the advantages of PEC and CDs.
The PEC sensing mechanism of CDs on GSH was analysed in detail
based on experimental results. In these designed sensors, CDs ex-
hibited both photoelectric and catalytic properties for the first time,
eliminating the need for additional catalyst. Several different nanoma-
terials such as AgNPs, GO, and MS were introduced into the CDs to
obtain hybrid nanocomposites for enhancement of the sensing proper-
ties for GSH. It was demonstrated that the catalytic properties of CDs
toward GSH were enhanced by these nanomaterials. The ultrasensitive
PEC sensors with CDs@MS hybrid nanocomposites showed the highest
sensitivity (57.6 nA μM− 1), best limit of detection (6.2 nM), and very
good selectivity toward GSH with interfaces of biological thiols and
other reducing agents. The practicality of our designed sensors was

257
Z. Li et al.
verified by the determination of GSH in human serum and myocardial
cells of mice that had been treated for different ischemia/ischemia-
reperfusion times. It was found that the developed sensors exhibited
much better sensing performances than the commercially available
GSH kit. These results proved that CDs based hybrid nanocomposites
are competitive tools for developing “green” PEC sensors with en-
hanced sensing properties. As a new sensing strategy, our proposed
sensor system has already exhibited its feasibility and capability of
biodetection, even in complicated biological environments. Therefore,
biologists can utilize these sensors to study the MI model with complex
parameters by detecting GSH in blood from different parts of the
human body. In order to achieve long-term stability of these sensors for
commercial uses in hospitals, the reliability of the fabrication processes
still needs to be improved.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (Nos. 61471207 and 81471799) and the
Opening Project of Key Laboratory of Microelectronic Devices &
Integrated Technology, Institute of Microelectronics, Chinese
Academy of Sciences.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.bios.2017.07.065.
References
Bian, H., Shu, X., Zhang, J., Yuan, B., Wang, Y., Liu, L., Xu, G., Chen, Z., Wu, Y., 2013.
Chem. Asian J. 8 (11), 2746–2754.
Braunwald, E., Kloner, R.A., 1985. J. Clin. Investig. 76 (5), 1713–1719.
Cai, Q.Y., Li, J., Ge, J., Zhang, L., Hu, Y.L., Li, Z.H., Qu, L.B., 2015. Biosens. Bioelectron.
72, 31–36.
Chou, S.S., De, M., Luo, J., Rotello, V.M., Huang, J., Dravid, V.P., 2012. J. Am. Chem.
Soc. 134 (40), 16725–16733.
Croissant, J.G., Zhang, D., Alsaiari, S., Lu, J., Deng, L., Tamanoi, F., AlMalik, A.M., Zink,
J.I., Khashab, N.M., 2016. J. Control. Release 229, 183–191.
Datta, A., Kapri, S., Bhattacharyya, S., 2016. J. Mater. Chem. A 4 (38), 14614–14624.
Dreyer, D.R., Park, S., Bielawski, C.W., Ruoff, R.S., 2009. Chem. Soc. Rev. 43 (15),
5288–5301.
Fang, X., Zhao, X., Fang, W., Chen, C., Zheng, N., 2013. Nanoscale 5 (6), 2205–2218.
Gao, L., Gan, W., Xiao, S., Zhan, X., Li, J., 2015. RSC Adv. 5 (4), 52985–52992.
Han, Y.H., Kim, S.H., Kim, S.Z., Park, W.H., 2008. Int. J. Oncol. 33 (33), 1099–1105.
Hao, M., Li, C., Liu, R., Jing, M., 2015. Spectrochim. Acta A 149, 600–606.
Lee, J.E., Lee, N., Kim, T., Kim, J., Hyeon, T., 2011. Acc. Chem. Res. 44 (10), 893–902.
258
Biosensors and Bioelectronics 99 (2018) 251–258
Li, S., Guo, W., Liu, R., Du, J., 2016. Int. J. Electrochem. Sci. 11 (9), 7690–7704.
Li, S., Zheng, M.Q., Rozanski, G.J., 2009. Exp. Physiol. 94 (7), 815–824.
Li, W., Yue, Z., Wang, C., Zhang, W., Liu, G., 2013. RSC Adv. 3 (43), 20662–20665.
Liang, M., Su, R., Huang, R., Qi, W., Yu, Y., Wang, L., He, Z., 2014. ACS Appl. Mater.
Interfaces 6 (7), 4638–4649.
Liu, S., Cao, H., Wang, Z., Tu, W., Dai, Z., 2015. Chem. Commun. 51 (75), 14259–14262.
Marsich, L., Bonifacio, A., Mandal, S., Krol, S., Beleites, C., Sergo, V., 2012. Langmuir 28
(37), 13166–13171.
Mills, B.J., Weiss, M.M., Lang, C.A., Liu, M.C., Ziegler, C., 2000. J. Lab. Clin. Med. 135
(5), 396–401.
Ndamanisha, J.C., Bai, J., Qi, B., Guo, L., 2009. Anal. Biochem. 386 (1), 79–84.
Niu, W.J., Zhu, R.H., Cosnier, S., Zhang, X.J., Shan, D., 2015. Anal. Chem. 87 (21),
11150–11156.
Pu, Y.C., Wang, G., Chang, K.D., Ling, Y., Lin, Y.K., Fitzmorris, B.C., Liu, C.M., Lu, X.,
Tong, Y., Zhang, J.Z., Hsu, Y.J., Li, Y., 2013. Nano Lett. 13 (8), 3817–3823.
Shi, Y., Pan, Y., Zhang, H., Zhang, Z., Li, M.J., Yi, C., Yang, M., 2014. Biosens.
Bioelectron. 56, 39–45.
Tang, H., Chen, J., Nie, L., Yao, S., Kuang, Y., 2006. Electrochim. Acta 51 (15),
3046–3051.
Tao, Y., Lin, Y., Huang, Z., Ren, J., Qu, X., 2013. Adv. Mater. 25 (18), 2594–2599.
Townsend, D.M., Tew, K.D., Tapiero, H., 2003. Biomed. Pharmacother. 57 (3–4),
145–155.
Vairetti, M., Ferrigno, A., Rizzo, V., Richelmi, P., Cillo, U., Imberti, R., 2007. Transplant.
Proc. 39 (6), 1768–1770.
Vinoth, V., Rozario, T., Wu, J., Anandan, S., Ashokkumar, M., 2017. ChemistrySelect 2
(17), 4744–4752.
Wang, G.L., Xu, J.J., Chen, H.Y., Fu, S.Z., 2009. Biosens. Bioelectron. 25 (4), 791–796.
Wang, J., Qiu, J., 2016. J. Mater. Sci. 51 (10), 4728–4738.
Wu, X., Li, R., Li, Z., Liu, J., Wang, G., Gu, Z., 2014. RSC Adv. 4 (48), 24978.
Xiong, L., Du, X., Shi, B., Bi, J., Kleitz, F., Qiao, S., Xiong, L., Qiao, S., 2014. J. Mater.
Chem. B 3 (8), 1712–1721.
Xu, S., Li, Y., Chen, Z., Hou, C., Chen, T., Xu, Z., Zhang, X., Zhang, H., 2016a. Mater. Sci.
Eng. C. 59, 258–264.
Xu, Y., Chen, X., Chai, R., Xing, C., Li, H., Yin, X.B., 2016b. Nanoscale 8 (27),
13414–13421.
Ye, W., Yu, H., Zhang, W., Li, W., Liu, G., Wang, C., 2014. Micro & Nano Lett. 9 (4),
294–296.
Yu, P., Wen, X., Toh, Y.R., Lee, Y.C., Huang, K.Y., Huang, S., Shrestha, S., Conibeer, G.,
Tang, J., 2014. J. Mater. Chem. C. 2 (16), 406–412.
Yuan, B., Zeng, X., Xu, C., Liu, L., Ma, Y., Zhang, D., Fan, Y., 2013. Sens. Actuat. B-Chem.
184 (9), 15–20.
Yue, Z., Lisdat, F., Parak, W.J., Hickey, S.G., Tu, L., Sabir, N., Dorfs, D., Bigall, N.C.,
2013. ACS Appl. Mater. Interfaces 5 (8), 2800–2814.
Zhang, J., Tu, L., Zhao, S., Liu, G., Wang, Y., Wang, Y., Yue, Z., 2015. Biosens.
Bioelectron. 67, 296–302.
Zhang, X., Guo, Y., Liu, M., Zhang, S., 2013. RSC Adv. 3 (9), 2846–2857.
Zhao, S., Li, Z., Li, Y., Yu, J., Liu, G., Liu, R., Yue, Z., 2017. J. Photoch. Photobio. A 342,
15–24.
Zhao, S., Zhang, J., Li, Z., Zhang, P., Li, Y., Liu, G., Wang, Y., Yue, Z., 2016a. Microchim.
Acta 184 (3), 677–686.
Zhao, W.W., Wang, J., Zhu, Y.-C., Xu, J.-J., Chen, H.Y., 2015. Anal. Chem. 87 (19),
9520–9531.
Zhao, W.W., Xu, J.J., Chen, H.Y., 2014. Chem. Rev. 114 (15), 7421–7441.
Zhao, W.W., Yu, X.D., Xu, J.J., Chen, H.Y., 2016b. Nanoscale 8 (40), 17407–17414.
Zuo, P., Lu, X., Sun, Z., Guo, Y., He, H., 2015. Microchim. Acta 183 (2), 519–542.