Cu-Doped Carbon Dots as Catalysts for the Chemiluminescence Detection of Glucose.

Development of metal-doped carbon dots (CDs) to effectively modulate their electronic properties and surface chemical reactivities is still in its early stage. In this paper, a facile solid-phase synthesis strategy was developed to synthesize Cu-doped CDs (Cu-CDs) using citric acid as the carbon source and Cu(NO3)2·3H2O as the dopant, respectively. The as-prepared Cu-CDs exhibited superior peroxidase-like activity to horseradish peroxidase and were stable under a wide range of pH and temperatures. Consequently, the Cu-CD-based chemiluminescence sensing was applied to sensitively detect glucose with a low detection limit of 0.32 μM, and the recoveries and the relative standard deviation of the serum sample are 87.2-112.2 and 8.16% (n = 6), respectively. Notably, the proposed chemiluminescence sensing was also successfully applied for label-free detection of glucose in complex biological samples, which envisioned its potential applications in clinical diagnosis and other analytical assays.

Introduction

As a versatile analytical technique, chemiluminescence (CL) assays have received considerable attention because of their fastness, very high sensitivity, and a broad range of analytical applications with no monochromator required.[1−3] For a traditional CL immunoassay strategy, a natural enzyme, such as horseradish peroxidase (HRP), was extensively used as a catalyst to achieve highly sensitive determinations. However, the natural enzyme are often limited for its some serious shortcomings (including poor stability, limited sources, and lack of long-term stability under environment changes).[1,4] Therefore, development of an artificial synthetic peroxidase-like nanomaterial acting as the catalyst mimicking the native enzyme attracts growing interest.

On the basis of this consideration, various types of peroxidase-like nanomaterials, such as metal nanoparticles (NPs),[5] metallic oxide NPs,[6,7] metal–organic frameworks,[8] and carbon NPs,[9,10] have been reported for strengthening the highly sensitive CL detection. Among these artificial enzymes, carbon dots (C-dots) have attracted growing attention due to its unique electron transfer, a large specific surface area, and broadband light-absorbing abilities.[11,12] For instance, Jiang et al. reported that CDs could effectively enhance the NaIO4–H2O2 CL reaction.[13] Shi et al. demonstrated that CDs have an intrinsic peroxidase-like activity to oxidation of 3,3,5,5-tetramethylbenzidine (TMB) by hydrogen peroxide.[14] Guo et al. reported that the catalytic activity of CDs is mainly affected by the surface states. Specially, carbonyl-functionalized CDs have better enhanced ability for the luminol CL reaction, and after being reduced, the CL intensity decreased significantly.[9] It is worth to propose a strategy for the facile synthesis of carbonyl-functionalized CDs as peroxidase mimetics in the CL reaction.

Recently, some reported strategies indicated that the intrinsic properties of CDs can be effectively tuned by doping with heteroatoms and/or surface passivation.[15,16] Doping and/or surface passivation could provide a means for emerging chemical reactivities and potential applications of CDs because of additional chemical groups associated with CDs.[17] Consequently, some efforts, to enhance their photocatalytic ability, have been devoted to the preparation of doped CDs.[18−20] More interestingly, metal atom doping can improve the optical properties as well as novel functionalities of the CDs because of its different band structures.[21] For example, Ni-doped CDs can greatly enhance the efficiency of hydrogen production because of the easy electronic transfer between Ni-doped CDs and solution.[22] A fluorescent probe based on Zn-doped CDs was proposed for highly sensitive detection of glucose because of the heteroatom-directed, oxidized carbon-based surface passivation.[23] After Cu and N co-doping, CDs were used as a novel photocatalyst due to their electron-accepting and donating ability enhanced.[18]

Inspired by the above observations, herein, a facile one-pot synthesis strategy was proposed to prepare Cu-doped CDs (Cu-CDs) using citric acid (CA) and Cu(NO3)2·3H2O as precursors (Scheme ). During pyrolysis, Cu2+ was used not only as a doped atom but also as an oxidant to form carbonyl-functionalized CDs. The as-prepared Cu-CDs exhibited superior peroxidase-like activity to HRP. Meanwhile, the catalytic activity of as-obtained Cu-CDs was not obviously changed under a wide range of pH and temperatures. More importantly, CL sensing based on Cu-CDs could be used for sensitive and selective determination of glucose.

Scheme 1

Synthesis Strategy of Cu-CDs and the Proposed CL Sensing for Glucose Detection

Results and Discussion

Characterization of Cu-CDs

The morphology and structure of the as-prepared Cu-CDs are shown in Figure . According to the transmission electron microscopy (TEM) image (Figure a), the size distribution of Cu-doped CDs was mainly in the range of 6.0–16 nm with an average size of 11.0 nm (Figure b), which was comparable to that of the reported Cu-CDs (around 10.5 nm).[21] The high-resolution TEM image (Figure a, inset) shows the lattice parameter of the as-prepared Cu-CDs, which is measured to be 0.21 nm, which correspond to the diffraction planes of sp2 graphitic carbon.[24] In comparison to weaker photoluminescence of Cu-CD, the photoluminescence quantum yield of bare CD calculated using quinine sulfate as a reference is up to 16.8%.

Figure 1

TEM image (a) and the diameter distribution (b) of Cu-CDs.

X-ray photoelectron spectroscopy (XPS) was performed to reveal the composition of the as-synthesized Cu-CDs. Figure a is the XPS survey spectrum of Cu-CDs, which indicated that C 1s, O 1s, and Cu 2p signals appeared at 287.3, 530.3, and 936.6 eV, respectively. It illustrated the successful doping of Cu atoms in Cu-CDs. Significantly, a new peak at 530.5 eV could be observed in the high-resolution O 1s spectrum of Cu-CDs (Figure S1), which was ascribed to O–Cu after Cu doping.[25] These results further proved that Cu atoms were perfectly doped in Cu-CDs. As shown in Figure b, the C 1s spectrum consisted of four types of peaks at 284.0, 284.6, 285.8, and 287.6 eV, which were well-fitted into C–C/C=C, C–O, C=O, and C(O)–O, respectively.[26] Three peaks at 529.8, 531.1, and 531.6 eV in the O 1s spectrum (Figure S1) can be ascribed to C–O, C=O, and C–O–C, respectively.[27] The results mentioned above confirmed the existence of −OH and −COOH groups on the surface of Cu-CDs.

Figure 2

XPS survey spectrum (a) and XPS high-resolution survey of C 1s (b) of Cu-CDs.

The functional groups of CA are prone to coordinate with Cu2+, which might result in electron transport between CA and Cu2+ in the process of synthesizing Cu-CDs. To bear out this hypothesis, the UV–vis spectrum of the as-prepared CDs and Cu-CDs was recorded. Most surprisingly, the peaks at 340 nm disappeared after the doping of Cu2+ (Figure S2). We speculated it may result from the oxidation of nitrogen-doped graphene quantum dots.[28] Inspired by this, Fourier-transform infrared (FTIR) was used to investigate the change of functional groups on the surface of the bare CDs and Cu-CDs. As observed in Figure a, obvious absorption bands at 3444, 1462, and 1277 cm–1 were ascribed to −OH, COO–, and C–O stretching vibrations, respectively. These observations confirmed the existence of −OH and −COOH groups on the surface of Cu-CDs, which was consistent with the XPS results of Cu-CDs. Compared to bare CDs, the absorption band of the C–O group at 1277 cm–1 decreases in the FTIR spectrum of the as-prepared Cu-CDs, while a increase in the intensity of −OH groups at 1462 cm−1 increases choosing the intensity of peak at 1400 cm−1 as a reference. These changes implied that the −OH groups on the surface of Cu-CDs were oxidized by Cu2+. To further reveal Cu2+ as an oxidant, the XPS high-resolution survey of Cu 2p was used to reveal the composition of copper valence. As shown in Figure b, the peaks at about 934.0 and 953.7 eV correspond to the spin–orbit splitting of Cu 2p3/2 and Cu 2p1/2, respectively. The obvious peaks at 932.2 were attributed to Cu+, which had further illustrated the electron transport between CA and Cu2+.[29] It was deduced that the electronic properties and surface chemical reactivities of Cu-CDs might be improved because of the introduction of Cu atoms.

Figure 3

FTIR spectrum of the bare CDs and Cu-CDs (a) and XPS high-resolution survey of Cu 2p (b).

Peroxidase-like Activity of Cu-CDs

In view of the peroxidase-like activity of CDs, whether the as-prepared Cu-CDs possess the peroxidase-like activity was explored. As shown in Figure S3, a clear absorption band at 425 nm could be observed, which belonged to the characteristic absorption peak after oxidation of luminol. These results illustrated that Cu-CDs also had the properties for catalysis. Additionally, in the absence of Cu-CDs, the intensity of proposed CL sensing was relatively weak. Therefore, it was assumed that the enhanced intensity of CL may be ascribed to Cu-CD interaction with the reactants to form the intermediates, resulting in the oxidation of luminol. On the basis of the results of Merényi and co-workers,[30] the possible enhanced mechanism is summarized in Figure . First, luminol molecules transformed into luminol anions under alkaline conditions. Second, luminol anion and H2O2 were absorbed on the surface of Cu-CDs, and H2O2 is activated by Cu2+/Cu+ via electron exchange to produce hydroxyl radicals (•OH). The formed •OH reacted with HO2– and the luminol anion to facilitate the formation of superoxide radicals and luminol radicals. Meanwhile, an activated transition complex (Cu-CDs-luminol anion-O2•–) was formed through the charge transfer reaction between luminal as the donor and Cu-CDs as the acceptor. Dissolved oxygen was also absorbed at the surface of activated transition complex and decomposed into O2•– by activated Cu-CDs-luminol anion-O2•–. Then, the resultant Cu-CDs-luminol anion-O2•– further oxidize luminol radicals to yield electronically excited 3-aminophthalate anions via electron exchange. Subsequent formation of the excited 3-aminophthalate anions produces a strong CL emission when it is relaxed to the ground state.[9,31]

Figure 4

Possible Mechanism for the Cu-CDs-Luminol-H2O2 CL system.

Encouraged by the above observation, the catalytic efficiency of the as-prepared Cu-CDs under different pH and temperatures was explored. After being treated in a wide range of pH values (5.0–11.0) (Figure S4a) and temperatures (20–80 °C) (Figure S4b), the relative catalytic activity of Cu-CDs did not change much. More importantly, the as-prepared Cu-CDs exhibit superior peroxidase-like activity to HRP (Figure S4a,b), whose HRP activity dramatically declined after incubation at temperatures greater than 40 °C for 2 h or at pH values less than 5.0.[32]

CL Biosensor for H2O2

To achieve the best performance of Cu-CDs for the H2O2 assay, the effects of pH value, concentration of luminol, and dosage of Cu-CDs were studied. As shown in Figure S5, the optimal pH value, concentration of luminol, and dosage of Cu-CDs were measured to be 10.0, 0.75 mM, and 0.5 mg/mL, respectively. Under the optimal conditions, an increased CL intensity could be observed with increasing H2O2 concentrations (Figure a). As illustrated in Figure b, there was a good linear correlation between the CL intensity at 425 nm and the H2O2 concentration in the range of 7.5–150.0 μM, and the linear equation is Y = 1.39X – 5.13, R2 = 0.998, where Y stands for the CL intensity of the system and X stands for the different concentrations of H2O2, and it was estimated that the detection limit was about 1.48 μM (S/N = 3).

Figure 5

CL spectra of the Cu-CD-based biosensor in the presence of different concentrations of H2O2. The concentrations of H2O2 were 0, 7.5, 15.0, 30.0, 45.0, 75.0, 150.0, 225.0, 300.0, 450.0, 600.0, and 1125.0 μM (a). The linear plot of the CL intensity at 425 nm vs different concentrations of H2O2 (b).

CL Biosensor for Glucose

On the basis of the above results, sensitive detection of glucose through the H2O2-mediated oxidation reaction by Cu-CD-based CL sensing is proposed. As shown in Figure a, the intensity at 425 nm of the CL system increased gradually with an increase in the concentration of glucose. The CL intensity is linearly proportional to the glucose concentration in the range of 1.0–48.0 μM (Figure b), and the linear equation is A = 1.35B – 1.42, R2 = 0.999, where A is the CL intensity of the system and B is the concentration of glucose. At the same time, it was estimated that the detection limit was as low as 0.32 μM (S/N = 3), which is much lower than those previously reported CL methods (Table ).

Figure 6

CL spectra of the Cu-CD-based biosensor in the presence of different concentrations of glucose, which were 0, 1.0, 4.0, 8.0, 12.0, 16.0, 24.0, 36.0, 48.0, 60.0, 80.0, and 110.0 μM (a). The linear plot of the CL intensity at 425 nm vs different concentrations of glucose (b).

Table 1

Comparison of Different CL Methods for the Determination of Glucose

system	linear range (μM)	detection limit (μM)	references
graphene oxide-luminol-H2O2	100–2000	82	(33)
hemin@HKUST-1-luminol-H2O2	7.5–750	7.5	(34)
CuII/[bmim][Br]-lucigenin-H2O2	50–4000	6.5	(35)
CuO NPs-luminol-H2O2	5–60	2.9	(31)
PtCox@graphene nanocomposite-luminol-H2O2	3.33–27.75		(36)
Fe3O4-chitosan NPs-luminol-H2O2	0.85–100	0.43	(37)
TiO2-electrochemiluminescence	400–3600	5	(38)
Au NPs–MWCNT-electrochemiluminescence	1–1000	0.5	(39)
N,S-doped CDs-colorimetry	200–2500	30	(40)
B-doped carbon quantum dots-fluorescence	8–80	8.0	(41)
CDs-fluorescence	100–8000	100	(42)
CDs-fluorescence	10–300	2.9	(43)
Cu-doped CDs-luminol-H2O2	1–48	0.32	this work

Selectivity of the CL Biosensor

To test the selectivity of the proposed CL sensor based on Cu-CDs for the glucose assay, the competition and control experiments were conducted. As illustrated in Figure , even when the concentration of K+, Mg2+, mannose (Man), saccharose (Sac), xylose (Xyl), fructose (Fru), and galactose (Gal) was 400 μM, no obvious change was observed. Obvious changes can be observed after the addition of Fe2+, Mn2+, and ascorbic acid (Asc), which suggested that Fe2+, Mn2+, and ascorbic acid may influence the performance of glucose detection. Figure S6 shows that the interferences of the above interfering ions and Asc were eliminated by adding 0.5 mM EDTA and 4-hydroxy-2,2,6,6-tetramethyl-N-oxygen-piperidine, respectively. Therefore, the proposed Cu-CD-based CL sensor has promising selectivity for the detection of glucose.

Figure 7

Selectivity of the proposed strategy for glucose sensing (the concentrations of glucose and other interfering substances were 40.0 and 400.0 μM, respectively).

CL Sensing of Glucose in Serum Samples

On the basis of the proposed CL sensor based on Cu-CDs for the sensitive and selective assay for glucose, the applicability of the Cu-CD-based CL sensor in complex matrixes was evaluated. Dilute human serum samples were spiked with different concentrations of glucose and detected by the proposed CL sensing. Table S1 shows that the obtained recoveries of serum samples were in the range of 87.2–112.2%. In addition, the relative standard deviation was less than 8.16% (n = 6). All observations further approved the reliability and feasibility of developed CL sensing based on Cu-CDs for detecting glucose in biological samples.

Conclusions

In summary, we proposed a facile and high-output solid-phase synthesis strategy for synthesizing Cu-doped CDs (Cu-CDs) using Cu(NO3)2·3H2O as the dopant. The electronic properties and surface chemical reactivities of as-prepared Cu-CDs had been substantially improved because of Cu doping, which was indeed beneficial for the peroxidase-like activity of Cu-CDs. Moreover, the prepared Cu-CDs exhibited good stability and outstanding peroxidase-like activity under a wide range of pH values and temperatures. As a consequence, the as-synthesized Cu-CDs were used as novel CL sensing catalysts for the highly sensitive detection of glucose with a detection limit as low as 0.32 μM. It is envisioned that the Cu-CD-based CL sensing would be widely applied in sensing target analytes in the future.

Experimental Section

Materials

Glucose, mannose, saccharose, xylose, fructose, galactose, and glucose oxidase were purchased from Sigma-Aldrich Co., Ltd. CA (99.5%), hydrogen peroxide (H2O2, 30%, v/v), Cu(NO3)2·3H2O, Na3PO4, NaH2PO4, NaOH, and HCl were purchased from Aladdin Chemistry Co., Ltd. (Shanghai, China). All reagents are of analytical grade and were used without further purification. Ultrapure water was prepared using a Millipore water purification system (≥18 MΩ, Milli-Q, Millipore) and was used in all of the runs.

Preparation of Cu-CDs

Typically, CA (1.0 g) and Cu(NO3)2·3H2O (0.1 g) were mixed in a 25 mL round-bottom flask and heated to 230 °C for 20 min under vigorous stirring. The obtained mixture solution was transferred into a little beaker and dissolved in 10 mL of water after naturally cooling to room temperature. The pH of the above solution was adjusted to 7.0 by adding NaOH. The obtained mixture was dialyzed with a MD34 (3500 Da) dialysis tube for 48 h to remove the small fragments. The Cu-CD powder was obtained using a rotary evaporator. Then, the obtained Cu-CDs were soluble in water and preserved at 4 °C for the following experiments.

For the preparation of CDs, CA (1.0 g) was mixed in a 25 mL round-bottom flask and heated to 230 °C for 20 min under vigorous stirring. Other experimental steps were conducted similar to the preparation of Cu-CDs as mentioned above.

Quantum Yield Measurements

The photoluminescence quantum yield of fluorescent CDs was calculated using the following equation

Quinine sulfate (ΦR = 0.54) was dissolved in 0.1 M H2SO4 [refractive index (η of 1.33)], and CDs were dissolved in deionized water (η = 1.33). Here, Φ and I are the quantum yield and integrated emission intensity and η and A are the refractive index and optical density. The subscript R refers to the reference fluorophore of the known quantum yield.

Characterization

A Cary 60 UV–vis spectrometer (Agilent Technologies, USA) was used for absorption measurement. The CL spectra were recorded using a Cary Eclipse fluorescence spectrophotometer (Agilent Technologies, USA) under the optimal conditions: a voltage of 650 V and an emission slit of 10 nm, respectively. FTIR spectroscopy study was conducted in KBr pellets using a PerkinElmer FTIR spectrophotometer (PerkinElmer, USA). TEM images were recorded using a Tecnai G2 F20TEM (FEI, USA) operating at 200 kV. XPS spectra were recorded with a Thermo ESCALAB 250Xi Multitechnique Surface Analysis (Thermo, USA). X-ray diffraction analyses were carried out on a Rigaku D/max 2500 v/pc X-ray powder diffractometer (Rigaku, Japan) with Cu Kα radiation (λ = 0.154 nm).

CL Biosensor for H2O2 and Glucose Detection

In a typical experiment, 100 μL of Cu-CDs (1.0 mg/mL) and 100 μL of luminol (5 mM) were added in 700 μL of PBS buffer solution (25 mM, pH = 9.0). Then, H2O2 with concentrations of 7.5, 15.0, 30.0, 45.0, 75.0, 150.0, 225.0, 300.0, 450.0, 600.0, and 1125.0 μM were mixed with the above mixture, respectively. The CL was recorded immediately, and the emission intensity at 425 nm was used to evaluate the assay performance.

For glucose detection, 50 μL of glucose oxidase (1 mg/mL) was incubated with 50 μL of glucose (the concentrations of glucose were 0, 1.0, 4.0, 8.0, 12.0, 16.0, 24.0, 36.0, 48.0, 60.0, 80.0, and 110.0 μM, respectively) at 37 °C for 30 min. Then, the obtained solution was added to the mixture containing 100 μL of Cu-CDs (1 mg/mL), 700 μL of PBS buffer (25 mM, pH = 9.0), and 100 μL of luminol (5 mM). The CL generated at 425 nm was used to evaluate the assay performance.

All the measurements in this section were performed three times, and the standard deviation was plotted as the error bar.

CL Biosensor for the Detection of Glucose in the Real Sample

To evaluate the practicality of the presented CL method, the CL biosensor-based Cu-CD was applied to determine the level of glucose in the human serum sample obtained from the No. 5 Hospital of Guilin (Guangxi, China). The obtained serum was diluted with 20 μL of H2O, and then 500 μL of Ba(OH)2 (0.08 M) and 500 μL of ZnSO4 (0.1 M) were added. After vortex oscillations and centrifugation for 10 min, these samples were diluted in PBS buffer (25 mM, pH 9.0), and then they were spiked with standard solutions containing different concentrations of glucose. The final solution was analyzed by the CL biosensor-based Cu-CDs.