Efficient double-quenching of electrochemiluminescence from CdS:Eu QDs by hemin-graphene-Au nanorods ternary composite for ultrasensitive immunoassay

A novel ternary composite of hemin-graphene-Au nanorods (H-RGO-Au NRs) with high electrocatalytic activity was synthesized by a simple method. And this ternary composite was firstly used in construction of electrochemiluminescence (ECL) immunosensor due to its double-quenching effect of quantum dots (QDs). Based on the high electrocatalytic activity of ternary complexes for the reduction of H2O2 which acted as the coreactant of QDs-based ECL, as a result, the ECL intensity of QDs decreased. Besides, due to the ECL resonance energy transfer (ECL-RET) strategy between the large amount of Au nanorods (Au NRs) on the ternary composite surface and the CdS:Eu QDs, the ECL intensity of QDs was further quenched. Based on the double-quenching effect, a novel ultrasensitive ECL immunoassay method for detection of carcinoembryonic antigen (CEA) which is used as a model biomarker analyte was proposed. The designed immunoassay method showed a linear range from 0.01 pg mL−1 to 1.0 ng mL−1 with a detection limit of 0.01 pg mL−1. The method showing low detection limit, good stability and acceptable fabrication reproducibility, provided a new approach for ECL immunoassay sensing and significant prospect for practical application.

applications. Therefore, another coreactant consumption probe enzyme mimics with high electrocatalytic reactivity and convenient storage and treatment attracted increasing interests. Among these enzyme mimics, for the reason that graphene oxide (GO) provides large specific surface areas and other specific properties, it has been widely employed for construction of devices involving catalytic processes. With its rapid development, more and more soluble GO-based hybrid peroxidase mimetics were developed and applied due to its good catalytic activity 29,30 . However, rare attention has been paid to the applications of mentioned GO-based hybrid peroxidase mimetics in smart ECL sensing. A ternary complex H-RGO-Au NRs was first synthesized and reported during our previous work 31 and recently we found that it showed excellent performance during ECL biosensing which has not been reported. Peroxidase-like activity of H-RGO-Au NRs could consume ECL coreactant (H 2 O 2 ) during enzymatic reactions on the immunosensor surface, resulting ECL signal diminution. Furthermore, with the combination of our previous interesting discovery that ECL resonance energy transfer (ECL-RET) could generate between Au NRs and the CdS:Eu QDs 32 , the introduce of the ternary complex is expected to realize further ECL quenching with the abundant Au NRs on its surface.
Here, we first presented a novel ultrasensitive ECL immunoassay method based on the high electrocatalytic activity of ternary complexes hemin-reduced graphene oxide-Au nanorods (H-RGO-Au NRs) for the reduction of H 2 O 2 as the coreactant of QDs-based ECL, and the ECL resonance energy transfer (ECL-RET) strategy between the abundant Au NRs on the graphene triple complex surface and the CdS:Eu QDs (Fig. 1). The designed immunoassay method showed a linear range from 0.01 pg mL −1 to 1.0 ng mL −1 with a detection limit of 0.01 pg mL −1 . Compared former reported material 33 , this novel H-RGO-Au NRs showed many unique and interesting properties including large specific surface, with abundant Au NRs for ECL-RET and peroxidase-like activity, all of which were favorable conditions for ECL biosensor. The sensing platform provided a feasible way to detect different cancer biomarkers by changing antibody. Also, the method provided a new approach for ECL immunoassay sensing and significant prospect for practical application.

Results and Discussion
Characterization of the materials. TEM images of employed materials were shown in Fig. 2. Therein, Fig. 2A demonstrates the surface topography of doped CdS:Eu QDs with about 6 nm diameter. And for pure GO, a thin layer morphology was found in Fig. 2B. And after modification with hemin by heating a hemin and GO mixture with reflux in ammonia in the presence of hydrazine, the formed hemin-modified GO (H-RGO) was used as a substrate material for growth of Au NRs in situ through reduction of HAuCl 4 by NaBH 4 and ascorbic acid (AA). From Fig. 2C, we could find that Au NRs were successfully generated and grew on the surface of the H-RGO. The TEM image also showed that Au NRs were homogeneous with size of 20 × 10 nm and covered on the surface of H-RGO densely. The abundant decoration of Au NRs not only enlarged the surface area of the composite and also could greatly quench ECL signal through effect of energy transfer efficiency (ECL-ET) due to the overlapped spectrum between the ECL spectrum and Au NRs's absorption spectrum. Energy-dispersive analysis of X-rays (EDX) was further used to characterize the element component of the as-synthesized H-RGO-Au NRs (Fig. 2D).
UV-vis absorption spectra was also employed on water-dissolved GO, hemin, H-RGO and H-RGO-Au NRs respectively and the results were shown in Fig. 3A. From carve b, it can be clearly seen that GO exhibits its obvious absorption peak at 233 nm and shoulder peak at 300 nm corresponding to π-π* transitions of the aromatic C= C band and n-π* transitions of the C= O band, respectively. Besides, carve c shows that hemin has an obvious absorption peak at 388 nm which is attributed to the Soret band of hemin. Also another group of weak peaks between 500 and 700 nm ascribed to the Q-bands of hemin. However, after hemin was added on GO, the 388 nm peak red shifted to 413 nm (carve d), which was deemed to the reason of adsorption of hemin molecule on GO driven by π-π stacking interactions between the porphyrin moiety and RGO. After continuous addition of HAuCl 4 and corresponding reduction the appearance of two new absorption peaks respectively at approximately 520 and 620 nm were consistent with the surface plasmon absorption of Au NRs. And this phenomenon indicated the generation of H-RGO-Au NRs complex. For the application of ECL-RET in biosensor, the most important point is to choose an effective ECL donor and a suitable acceptor, with an essential condition that the donor's ECL spectrum should be overlapped with the acceptor's absorption spectrum. As shown in carve a of Fig. 3A, two spectral bands could be in the ECL emission spectrum of the CdS:Eu film: the one which was from 450 to 550 nm came out of the host CdS, and the other one from 600 to 700 nm was attributed to the energy transfer from host CdS to Eu 3+ ions 32 . Exhilaratingly, from carve e of Fig. 3A, we can see that the Au NRs from the prepared ternary complex have two absorption peaks which completely overlap the ECL emission spectrum of CdS:Eu film. This great spectrum overlap between donor and  Electrocatalytic reactivity mechanism of ECL biosensor. In our design, glassy carbon electrode (GCE) was decorated by drop-coating of 10 μ L CdS:Eu QDs which was used as ECL emitter. During the cathodic potential scan, the CdS:Eu QDs were reduced to CdS:Eu −• , and the coreactant H 2 O 2 could react with CdS:Eu −• to obtain an excited state (CdS:Eu* ). This state could produce stable and high ECL signal in the aqueous solution and emit light at the same time (Fig. 4). The mechanism of this ECL process was listed as Eqs (1-3) 34 . Upon the addition of Hemin-RGO-Au, the ECL intensity decreased due to the occurred electrochemical reduction of hemin-RGO-Au. The reduced hemin-RGO-Au then chemically reduced the H 2 O 2 and resulted to the efficient consumption of the coreactant. The process could be described by the following mechanism (Eqs (4) and (5) The optimization of experimental conditions. The quenching efficiency of I E = 1 − I/I 0 (I, I 0 are the ECL intensity treated with or without CEA respectively.) is used to study the effect of experimental conditions on ECL intensity. The experimental conditions mostly containing the concentration of H 2 O 2 , the temperature and time of Ab-CEA incubation, the volume of H-RGO-AuNRs/Ab 2 tags. The concentration of H 2 O 2 is important to the detection of biosensor. As can be seen in Fig. 5A, 7.5 mM H 2 O 2 is the best concentration. Temperature of CEA incubated with GCE-CdS:Eu QDs/Ab 1 are an important factor of the incubation efficiency which could also affect the further incubation between CEA and H-RGO-Au/Ab 2 . Figure 5B shows that the ECL intensity increase with the increasing temperature and reach the maximum at 37 °C which means 37 °C could be chosen as incubation temperature. Also, the incubation time is studied and shown in Fig. 5C, the ECL intensity associated with the time increase until 60 min, and then ECL intensity keeps invariant after 60 min. The concentration of CEA is 1.0 × 10 −10 mg/mL during detection. Besides, Fig. 5D shows the volume of H-RGO-AuNRs/Ab 2 tags could also influence the ECL intensity. The ECL intensity reach maximum when the volume increased to 20 μ L, however, the signal remain unchanged when the volume continue to increase. So, 20 μ L H-RGO-AuNRs/Ab 2 tags is the optimized condition in this experiment. Characteristic of biosensing platform. ECL signals at each immobilization step were also recorded to characterize the fabrication process of the biosensing platform. Figure 6A are signal-potential and signal-time curve of GCE-CdS:Eu QDs after added a scanning potential from − 1.3 V to 0 V, respectively. And we can see  At the same time, every step of the fabrication process of this ECL biosensor was continuously monitored by electrochemical impedance spectroscopy (Fig. 6B). The bare electrode displayed an almost straight line (curve a), which was characteristic of a diffusion process. While CdS:Eu QDs were assembled on the electrode surface, the electron− transfer resistance (Ret) increased obviously (curve b), which indicated that the CdS:Eu QDs were immobilized on the electrode surface and decreased the electron-transfer efficiency. Ab 1 and following incubation of CEA resulted in a larger Ret (curve c and d). And this is mainly due to the poor electroconduction of biomolecules which results in steric hindrance and then inhibits the interfacial charge transfer. However, after following conjugated nanoprobe of H-RGO-Au/Ab 2 onto the surface of Ab 1 through specific interaction, an obvious of Ret decreases was appeared, which indicates that the synthesized H-RGO-Au/Ab 2 possessed high conductivity, good electron transfer efficiency and excellent biocompatibility (curve e).

Analytical performance.
Under the optimization of experimental conditions, the sensitivity of the ECL biosensor based on CdS:Eu QDs and H-RGO-Au NRs were studied by changing the concentration of CEA. From Fig. 7, ECL intensity is found to be very sensitive and decreased with increased CEA concentration form 1.0 × 10 −14 to 2.0 × 10 −9 g/mL. Besides, the inset of Fig. 7 show that quenching efficiency (I E ) was also logarithmic related to the CEA concentration with R = 0.998 and limit of detection (LOD) was 0. 01 pg/mL, which was comparable with that of 0.06 pg/mL CEA by hybrid gold/silica/CdSe-CdS quantum-dot nanostructure based ECL immunosensor 36 . And it much lower than that of Au-polydopamine functionalized carbon encapsulated Fe 3 O 4 magnetic nanocomposites based biosensor (0.33 pg/mL CEA) 37 , polybead carried gold nanoparticles assistant biosensor (0.12 pg/mL CEA), electrochemical immunoassay with branched electrode system (0.05 ng/mL CEA) 38 , and photoelectrochemical immunoassay (2.1 pg/mL CEA) 39 or fluorescent sensors (5.0 pg/mL CEA) 40 . This low LOD was supposed to be the reasons as following: the strong and stable ECL intensity of CdS:Eu QDs, the high electro-catalysis effect of H-RGO-Au NRs to coreactant and the energy transfer quenching effect of Au NRs.
The selectivity, repeatability and stability of ECL biosensor. The practicality ECL biosensor was measured by the selectivity, repeatability and stability. The detection signal of 1.0 pg/mL CEA, 50 pg/mL AFP, 50 pg/mL PSA, 100 pg/mL BSA mixture was compared with pure CEA. And the result shows that the signal caused by mixture has no obvious difference with the signal caused the pure CEA, and the relative standard deviation between them is 7.26% which means the designed ECL biosensor has an excellent selectivity to detect CEA. Meanwhile, the repeatability was also been studied. 1.0 pg/mL CEA solution was detected by the same biosensor for five times with the relative standard deviation of 3.28%. Besides, 1.0 pg/mL CEA solution was also detected business five electrodes and got a relative standard deviation of 5.40%. Those result indicated the designed ECL biosensor has a favorable repeatability. Finally, the stability of the biosensor was studied. The biosensor has been kept in pH = 7.4 PBS buffer at 4 °C for a week before detection. And the ECL response of treated electrode is almost the same as the electrode without be treated which means the nice stability of the biosensor.

Application in analysis of samples. The practical application of the immunoassay was investigated by
analyzing the recoveries of different concentrations of CEA in human serum samples. In order to confirm the practicality and reliability of this ECL biosensor in clinical application, the standard addition method was used to detect the content of CEA in human serum. The result shows that the recovery of the ECL biosensor was between 97.5-110% which indicates this designed biosensor has a potential to be used in CEA clinical detection (Table 1).

Conclusion
In summary, a novel ECL biosensor for ultrasensitive CEA detection was built based on the high electrocatalytic activity of ternary complexes for the reduction of H 2 O 2 , acted as the coreactant of QDs-based ECL, and the ECL-RET strategy between the large amount of Au NRs on the graphene triple complex surface and the CdS:Eu QDs. This simple and ultrasensitive ECL biosensor combined the advantage of the high electrocatalytic activity of H-RGO-Au NRs and the superiority ECL-RET strategy between the Au NRs and the CdS:Eu QD. Besides, a dynamic detection for monitoring the concentration of CEA was also achieved in this biosensor based on the electrochemical method. The designed biosensor showing simple operation, wide linearity range, high sensitivity and excellent selectivity, provided a new approach for ECL immunoassay sensing and significant prospect for practical application.

Materials and Methods
Chemicals and reagents. Poly (sodium 4-styrenesuifonate) solution (PSS) was obtained from Sigma-Aldrich. Carcinoembryonic antigen (CEA) was purchased from Zhenglong Biochem.Lab and used without further purification. Graphene oxide was gotten from Nanjing XFNANO Materials Tech Co., Ltd. Cadmium nitrate tetrahydrate (Cd(NO 3 ) 2 ), Europium(III) nitrate hexahydrate (Eu (NO 3 )  Construction of ECL Biosensor. The CdS:Eu QDs modified GCE was treated with 1.5 mM MPA for 3 h at 4 °C, then the terminal carboxylic acid groups of the MPA on GCE surface were activated using 20 mg of EDC and 10 mg of NHS which were dispersed in 1.0 mL of 0.1 M imidazole-HCl buffer (pH 6.8) for 1 h at room temperature. After rising the electrodes with 0.1 M PBS (pH 7.4), 100 μ g mL −1 Ab 1 solution was introduced to the electrode and incubated at 4 °C for 12 h. Following this, the above electrode was immersed in 2 wt% BSA solution for 40 min to block the nonspecific active binding sites. Finally, the electrode (marked as GCE/CdS:Eu QDs/Ab 1 ) was washed by PBS and stored in 4 °C when not used.

Preparation of Hemin-Reduced Graphene Oxide (H-RGO).
Hemin-reduced graphene oxide (H-RGO) was synthesized according to our previous reported procedure 41 . First, 5 mL GO (2 mg/mL) was dispersed in 35 mL H 2 O in a 100 mL dried three-necked flask. After stirring and ultrasonication for 30 min, ammonia solution (60 μ L) and hemin (10 mg) were added to the mixture, followed by the addition of hydrazine solution (10 μ L). The solution was heated in a water bath at 60 °C for 4 h. The final product was collected after 10000 rpm centrifugation for 25 min and then washed and dried in vacuo for further use.

Preparation of H-RGO-Au NRs.
According to the provious work 30 , the seed solution was first prepared for Au NRs. 1 mL 0.12 mM HAuCl 4 was mixed with 1 mL 0.048 mM CTAB solution. Then fresh prepared NaBH 4 (2.4 mM, 0.12 mL) was added into above Au (III)-CTAB solution under vigorous stirring for 2 min. until the solution color changed from yellow to brownish-yellow. The prepared seed solution was aged at 27 °C for 2 h before use. Then the growth solution was prepared using 50 mL 0.2 mM HAuCl 4 together with 50 mL 0.048 M CTAB, 2.5 mL of 0.96 mM AgNO 3 and 3.4 mg H-RGO product under gently mixing at room temperature. Then 0.7 mL 0.02 M AA was added in to above mixture. The color of the solution could be found to fade to colorless from dark-yellow. Finally, 0.12 mL prepared seed solution was gently added into this solution at 27 °C for 20 min and then was kept at 30 °C for 12 h for NR growth on the surface of H-RGO. The final products were collected by centrifugation at 8500 rpm for 25 min and were then re-dispersed in water.
Preparation of H-RGO-Au NRs/Ab 2 . The Ab 2 -labeled H-RGO-Au NRs were prepared as follows: H-RGO-Au NRs (0.1 mg mL −1 , 0.5 mL) were mixed with PSS solution (200 μ L, 30%), ultrasonic for 5 min, and vibrated for 2 h, and then centrifuged to remove the excess PSS, and scattered in 400 μ L PBS. Subsequently, a solution of Ab 2 (50 μ g L −1 , 100 μ L) was added. The mixture was gently stirred at room temperature for 1 h and then centrifuged. After washing with PBS, the pellet was resuspended in PBS (500 μ L) containing 2% BSA and stored at 4 °C for further use.