Introduction

Photoelectrochemical (PEC) assays have attractive advantages such as low cost, simple instrumentation and high sensitivity [1,2,3,4,5,6]. The PEC process refers to the conversion of photon-to-current which is caused by the electron excitation and subsequent charge transfer of a photoactive material after absorbing photons under illumination [1, 7]. For the construction of PEC biosensor, “signal-off” type was a classic photocurrent signal response pattern [7, 8]. In general, a quencher with high quenching efficiency for the initial photocurrent was necessary in “signal-off” model PEC biosensor [9, 10]. The reported photocurrent quenchers are usually materials with poor conductivity which can enhance the steric hindrance and suppress the electron transfer [4]. However, the quenching efficiency was very limited and needed further improvement. Therefore, it is meaningful to find other approach to regulate the photocurrent signal. As well known, photo-absorption was very important in a PEC process [11]. The intensity of the photocurrent signal can be tuned by adjusting the light harvesting efficiency of the photoelectric material. Therefore, it is possible to develop PEC biosensor based on the change of photo-absorption efficiency induced by light-absorption materials. Regrettably, until now, this kind of materials has been rarely reported.

Carbon nanomaterials have attracted much attention due to their wide range of applications, such as biological imaging [12], green processing [13], and thin film transistors [14,15,16], originated from their advantages of excellent chemical stability, electrical and optical properties [17]. Despite they are widely applied in these area, exploiting them with high light absorption efficiency is just beginning [18, 19]. Since carbon nanomaterial exhibited high surface area, excellent dispersity, and good light absorption efficiency, it was a good candidate for photocurrent signal quencher in PEC assay and was expected to be an effective signal quencher.

The vascular endothelial growth factor (VEGF) is a kind of cell product which simulates the formation of new blood vessels of tissues [20,21,22]. The abnormal expression of VEGF is closely related to many diseases [23]. Therefore, the sensitive detection of VEGF165 can be a promising approach in clinical diagnoses. To date, there are various analytical methods for detecting VEGF165, such as field-effect transistor (FET) [24], fluorescence (FL) [22] and electrochemical [25], enzyme-linked immunosorbent assay (ELISA) [26]. However, the analytical methods are complicated operation, high cost and time-consuming. Lately, the aptasensors on 2D nanomaterial have attracted much attention due to their excellent property, such as stability, specificity and easy operation [27, 28]. Herein, we proposed a novel “signal-off” PEC biosensor for the detection of VEGF165 by using g-C3N4 as the signal indicator and porous carbon spheres as the efficient signal quencher. As shown in Scheme 1, photoactive material g-C3N4 can be filmed on the bare electrode surface, which provided a high initial PEC signal. The DNA strand S0 and S1 are two split parts of a whole VEGF165 aptamer [7, 29]. Firstly, S0 was incubated on the electrode, when target VEGF165 and S1 modified porous carbon spheres were incubated on the electrode, a sandwich structure can be fabricated by specific recognition of the aptamer and VEGF165 to form aptamer-VEGF165 complex, thus the carbon material got close to the electrode. Therefore, the original photocurrent signal was efficiently quenched by porous carbon spheres via competing with g-C3N4 to absorb light and competing electron donors to reduce the separation efficiency of electron and hole, as depicted in Scheme 1c. The change of the photocurrent should be related to the concentration of VEGF165, which was used for the detection of target VEGF165. The “signal-off” method based on adjusting the light absorption efficiency will provide new direction for design of other PEC biosensors.

Scheme 1
scheme 1

a Preparation and modification of porous carbon spheres; (b) Schematic illustration for the preparation of PEC biosensor, (c) The mechanism of the photocurrent response in graphitic carbon nitride

Full size image

Experimental section

Materials and reagents

Polyacrylic acid (PAA) was purchased from Sigma-Aldrich (Beijing, China, www.jk-scientific.com). hydrogen peroxide (H2O2) was bought from Kelong Chemical Inc. (Chengdu, China, webmaster53029.company.lookchem.cn). N-(3-(Dimethylamino)propyl)-N′-ethylcarbodiimide hydrochloride (EDC), N-hydroxy succinimide (NHS), gold chloride (HAuCl4), hexanethiol (HT), cetyltrimethyl-ammonium bromide (CTAB, 98%), ammonium hydroxide (NH3·H2O, 28% by weight in water), poly(vinyl pyrrolidone) (PVP, Mw~40,000), tetraethyl orthosilicate (TEOS, 99%), resorcinol, ethanol, formaldehyde (37%) and tetrabutyl orthotitanate (TBOT, 99%) were obtained from Sigma-Aldrich (St. Louis, MO, USA, www.sigmaaldrich. com). 0.1 M Na2HPO4, 0.1 M KCl and 0.1 M KH2PO4 were dissolving to obtain phosphate buffered saline (PBS, pH 7.0). Potassium ferrocyanide and potassium ferricyanide were dissolving with 0.1 M PBS solution (pH 7.0) to prepare [Fe(CN)6]3−/4- solution (5.0 mM). Vascular endothelial growth factor (VEGF165) and the oligonucleotides in this work were ordered from Sangon Inc. (Shanghai, China, www. sangon.com), which were shown as follows:

  • S0: 5’-AAGAGTGCAGGGTTTTTTTTTT-SH-3’

  • S1: 5’-NH2-ACCGTCTTCCAGAC-3’

Instrumentation

The PEC workstation (Ivium, P Netherlands) was used for PEC measurement, which contained a three-electrode system: the glassy carbon electrode was the working electrode, platinum wire electrode was the counter electrode, and Ag/AgCl (saturated KCl) electrode was the reference electrode. The CHI660D electrochemical workstation (Shanghai Chenhua Instrument, Shanghai, China) was used for cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) detection. The characterization of the synthesized carbon materials was obtained from transmission electron microscope (TEM, HT7700, Hitachi, Japan).

Synthesis of graphitic carbon nitride (g-C3N4)

The g-C3N4 was obtained based on previously reported literature with some modifications [30, 31]. First, 15 g of urea was dried in a crucible at 80 °C and held for 12 h, next, the dried sample was calcined at 550 °C for 4 h using a heating rate of 10 °C min−1 in a muffle furnace. The powders were with slight yellow color.

Preparation of porous carbon spheres

Porous carbon spheres were prepared as follows (Scheme 1a), firstly, the resorcinol-formaldehyde (RF) spheres were synthesized by mixing formaldehyde solution (0.07 mL), aqueous ammonia solution (28%, 0.05 mL), deionized water (10 mL), resorcinol (0.05 g) and ethanol (4 mL), which were stirred for 12 h to get the RF sample by centrifuging out, washing three times with ethanol, and redispersing in 10 mL of ethanol [32, 33]. The above RF particles were then treated with cetyltrimethyl ammonium bromide (CTAB, 5 mg·mL−1) for allowing CTAB adsorption on the surface of RF spheres. The above particles were then centrifuged to separate from the solution, and redispersed in ethanol (16 mL). The solution was mixed with water (32 mL), 0.4 mL of ammonia (28%). After stirring for 30 min, TEOS (0.8 mL) was added and stirring overnight. The resulting RF@SiO2 core-shell composites were centrifuged by washing with ethanol for three times and dried in a vacuum. To obtain the C@SiO2, the dried RF@SiO2 samples were heated at 800 °C in N2 with a heating rate of 2.5 °C min−1 for 2 h and then cooled to room temperature. Finally, to prepare the porous carbon spheres, 100 mg of the sample were dispersed in 30 mL of the aqueous NaOH solution (5 M), following by stirring for 5 h at 80 °C to etching the SiO2 layer. Then, the porous carbon spheres samples were centrifuged and washed with ethanol and ultrapure water for several times, followed by dispersal in ultrapure water.

Preparation of S1 modified porous carbon spheres

The porous carbon spheres were modified by polyacrylic acid (PAA). Firstly, 1 mL of the samples was acidified with 0.072 g PAA in 10 mL water by stirring overnight. The product was then tcentrifuged three times and redispersed in ultrapure water.

The S1 modified porous carbon spheres composite was obtained as follows: Firstly, 500 μL of porous carbon spheres was dissolved in 500 μL of PBS. Next, 0.0388 g EDC and 0.0058 g NHS dissolved in 1 mL PBS solution (PH 7.0), then the solution were dropped to the mixture and stirred for 1 h for activating the carboxyl group on the porous carbon spheres. 20 μL of S1 (50 μM) was then added to the mixture solution, stirring for 4 h to obtain S1 modified porous carbon spheres, the product was washed and redispersed in PBS solution (1 mL, pH 7.0).

Fabrication of the PEC biosensor

Scheme 1b represented the stepwise fabrication process for the biosensor. First of all, the glassy carbon electrode (GCE) (Φ = 4 mm) was sonicated in ultrapure water and anhydrous ethanol for three times after polished with alumina powder and dried at room temperature for 10 min. 10 μL solution of g-C3N4 (1 mg·mL−1) was dropped onto the GCE to form a homogeneous film after drying for 1 h. Next, Au NPs were electrodeposited on GCE by using HAuCl4 (1%) solution under −0.2 V for 10 s. After that, 10 μL of 2.0 μM S0 was coated on the electrode to incubate at 4 °C for 12 h. After blocking the nonspecific adsorption sites with HT (1.0 mM) for 40 min, VEGF165 and S1 modified porous carbon spheres solution were incubated on the electrode to hybridize with the S0 for 2 h at 37 °C. The electrode should be rinsed with ultrapure water after each step.

PEC measurement

The PEC measurement was carried out under optimal experimental conditions of 5 mL 0.1 M PBS solution (PH = 7.0) containing 0.1 M electron donor H2O2, the light-emitting diode (LED) light source acted as excitation light source with switching off-on-off for 10–20-10 s under the potential of 0.0 V.

Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS)

The CV measurement was recorded in 0.1 M PBS solution (pH = 7.0), which contained 5.0 mM [Fe(CN)6]3−/4- at the potential between −0.2 V and 0.6 V, with a scan rate of 50 mV·s−1. And the EIS measurement was performed in 0.1 M PBS solution (PH = 7.0) containing 5.0 mM [Fe(CN)6]3−/4-, with a scan rate of 100 mV·s−1.

Results and discussion

Morphology and characterization of porous carbon spheres

To obtain carbon materials with good dispersion and facilitate the DNA modification, the carbon spheres were obtained by a protected calcination procedure (Scheme 1a), which involved the synthesis of RF core and coating of the cores with porous SiO2 layer, followed by the protected calcination procedure and selective etching the SiO2 layer by NaOH solution. The morphology of the C@SiO2 and porous carbon spheres was characterized by TEM. As presented in Fig. 1a ,c, the C@SiO2 exhibits regular spherical structures with a uniform size. After removing the outer SiO2 layer by aqueous NaOH, the sample shows a porous structure with the size of 540 ± 39 nm (Fig. 1b and d). The surface area of the carbon spheres is 345 m2·g−1 (Fig. S-1), implying that the carbonization of RF resulted a mesoporous structure.

Fig. 1
figure 1

TEM images of C@SiO2 (a, c) and porous carbon spheres (b, d)

Full size image

Quenching mechanism of the porous carbon spheres

To investigate the quenching mechanism, the quenching efficiency of the porous carbon spheres was adjusted to be similar to C@SiO2. As evaluated by PEC, the decreases of photocurrent (ΔI) caused by porous carbon spheres and C@SiO2 are 1.319 μA and 1.407 μA, respectively (Fig. 2a and Fig. S2A). Cyclic voltammograms (CV) and electrochemical impedance spectroscopy (EIS) of g-C3N4, g-C3N4/C@SiO2 and g-C3N4/porous carbon spheres were then measured to compare the quenching mechanism. The voltammetric results confirm that when g-C3N4 immobilized on GCE, there is a well-defined redox peak (Fig. S2B and Fig. 2b). A significant decline can be obtained when C@SiO2 nanomaterials are further modified on the electrode (curve b in Fig. S2B). This is ascribed to the inhibition of electron transfer by the SiO2 layer. The conductivity is further investigated by EIS detection. As shown in Fig. S2C, the nyquist plots reveal a remarkable increase in the charge-transfer resistance (Ret) from g-C3N4 (curve a) to g-C3N4/C@SiO2 (curve b), which indicates an increase in reaction resistance. However, an increase of the redox peak is found after porous carbon spheres immobilized on the g-C3N4 electrode (curve b, Fig. 2b), owing to the conductivity of the porous carbon spheres. Similarly, compared to the Ret of g-C3N4 (curve a), a decrease of the Ret is obtained after the modification of porous carbon spheres (curve b, Fig. 2c). This is due to the accelerated electron transfer caused by carbon spheres. These results strongly supported the idea that carbon spheres were electroconductive material which can quench the photocurrent signal due to the competitive light absorption and competitive electron donors, which decreased the light generated electrons and holes from g-C3N4 and also reduced the charge separation efficiency.

Fig. 2
figure 2

a The PEC responses of (a) g-C3N4 and (b) g-C3N4/porous carbon sphere with 365 nm irradiation in 0.1 M pH 7.0 PBS containing 0.1 M H2O2. (b) CV of (a) g-C3N4 and (b) g-C3N4/porous carbon sphere in 0.1 M pH 7.0 PBS solution containing 5.0 mM [Fe(CN)6]3−/4− at the potential between −0.2 V and 0.6 V. (c) Nyquist diagrams for (a) g-C3N4 and (b) g-C3N4/porous carbon sphere modified GCE in 0.1 M pH 7.0 PBS containing 5.0 mM [Fe(CN)6]3−/4−

Full size image

Photoelectrochemical and electrochemical characterization of the modified electrode

The construction of the biosensor was investigated by PEC and CV measurements. The PEC characterization of the stepwise-modified electrode is shown in Fig. 3a, there is nearly no photocurrent of the bare GCE (curve a), obviously photocurrent response is observed when g-C3N4 was coated onto the bare GCE, which provided a high initial PEC photocurrent signal (curve b). Subsequently, a further enhancement of photocurrent was obtained after Au NPs was modified on the electrode (curve c), because the Au NPs facilitated the electron transfer. After incubating with S0 and HT, the photocurrent decreased, which was probably caused by the poor charge transfer ability (curve d and curve e). Finally, with the immobilization of VEGF165 and S1 modified porous carbon spheres, the photocurrent decreased significantly according to the reduced light absorption induced by porous carbon spheres.

Fig. 3
figure 3

a PEC responses of (a) GCE, (b) g-C3N4/GCE, (c) AuNPs/g-C3N4/GCE, (d) S0/AuNPs/g-C3N4/GCE, (e) HT/S0/AuNPs/g-C3N4/GCE, (f) S1-carbon/target VEGF165/HT/S0/AuNPs/g-C3N4/GCE with 365 nm irradiation in 0.1 M pH 7.0 PBS containing 0.1 M H2O2. (b) CV responses of (a) GCE, (b) g-C3N4/GCE, (c) AuNPs/g-C3N4/GCE, (d) S0/AuNPs/g-C3N4/GCE, (e) HT/S0/AuNPs/g-C3N4/GCE, (f) S1-carbon/target VEGF165/HT/S0/AuNPs/g-C3N4/GCE in 0.1 M pH 7.0 PBS containing 5.0 mM [Fe(CN)6]3−/4− at the potential between −0.2 V and 0.6 V

Full size image

CV was investigated to characterize the step-by-step construction process of the modified electrode. As exhibited in Fig. 3b, a well redox peak of the bare GCE was observed (curve a). The peak current apparently decreased when g-C3N4 was modified on the GCE (curve b). After electrodepositing with Au NPs, the peak current increased (curve c), because Au NPs possessed well conductivity. The redox peak declined after incubating with S0 (curve d), which was attributed to that DNA with negative charge hindered the electron transfer. Subsequently, the peak current further decreased (curve e) after blocking with HT, owing to that HT greatly reduced the electronic transmission. Subsequently, there was an enhancement of the redox peak after S1 modified porous carbon spheres and VEGF165 were assembled onto the electrode (curve f), which confirmed the excellent electrical conductivity of the porous carbon spheres.

Analytical performance of the biosensor for VEGF165 detection

The assembled PEC aptasensor was applied for determination of VEGF165 as a model. VEGF165 with various concentrations was detected. As exhibited in Fig. 4, the calibration plot shows a nice linear relationship between the decrease of photocurrent (ΔI) and the logarithm of the VEGF165 concentration from 10−5 nM to 102 nM with the detection limit of 3 fM. The linear equation for VEGF165 detection is ΔI = 0.1920lgc + 2.739, where c is the concentration of VEGF165, with a correlation coefficient of 0.9939. Besides, a comparison of the analytical performance between the PEC biosensor and the previous reported methods was shown in Table 1, it can be observed that the “signal-off” PEC biosensor displays wider linear range and higher sensitivity compared to the reported methods, showing an excellent analytical performance.

Fig. 4
figure 4

The linear relationship between the change of photocurrent and the concentration of VEGF165 from10−5 nM to 102 nM

Full size image
Table 1 Comparison for VEGF165 detection between our proposal method and other reported detection methodologies
Full size table

Selectivity and stability of the PEC biosensor

In order to investigate the specificity of the PEC biosensor, alpha-fetal protein (AFP), carcinoembryonic antigen (CEA), bovine serum albumin (BSA) and prostate specific antigen (PSA) were chosen as the interfering agents. As shown in Fig. 5a, there is an obvious photocurrent decrease (ΔI = 2.041 μA) with the addition of 100 fM VEGF165. However, there are negligible photocurrent changes with the addition of 1 pM AFP, CEA, BSA and PSA (ΔI = 0.362 μA, 0.469 μA, 0.310 μA, 0.593 μA). The result confirmed that the PEC biosensor displayed high specificity to VEGF165. Moreover, the stability was investigated by continuous cyclic measuring for 10 cycles under continuous off-on-off light. As shown in Fig. 5b, the constructed PEC biosensor shows negligible photocurrent decay in cyclic experiments, with RSD of 1.87%, indicating good stability for the biosensor.

Fig. 5
figure 5

a Selectivity of the PEC biosensor with interferences: AFP, CEA, BSA, and PSA. b Stability of this PEC biosensor incubated with 1 pM VEGF165 under continuous off-on-off light for 10 cycles

Full size image

Analysis of clinical serum samples

To research the potential application of the PEC aptasensor, the human blood serum samples (provided by the Ninth People’s Hospital of Chongqing, China) were used for research the spiked recovery experiment. Firstly, the serum samples were diluted 10 times with 0.1 M PBS (pH 7.0). The target VEGF165 was diluted with serum samples to different concentrations for further detection. The recovery rate of VEGF165 was between 87.30% and 107.47%, as shown in Table S-1, suggesting that the PEC aptasensor exhibits great potential for real sample analysis.

Conclusions

We show in this work a construction of a novel “signal-off” biosensor by using g-C3N4 as signal indicator and porous carbon spheres as signal quencher. The biosensor displayed higher sensitivity and wider linear range for the detection of VEGF165. The photocurrent of g-C3N4 is quenched due to the superior light absorption capacity of porous carbon spheres. This is different from previously reported quenching mechanisms. We are able to confirm that tuning the light absorption is an effective method to regulate photocurrent in the construction of PEC biosensor. We also believe that the “signal-off” strategy developed here can provide a promising platform and offer more opportunities for producing other biosensors with further enhanced performance.