A feasible approach to synthesize Cu₂O microcrystals and their enhanced non-enzymatic sensor performance

Abstract

We have introduced potassium bromide (KBr) as an additive to synthesize cuprous oxide (Cu₂O) microcrystals with various well-defined shapes. Here, the bromide ions play a pivotal role in controlling the shape of the Cu₂O microcrystals, from concave cubic into short hexapod shapes. As a typical representative, the obtained Cu₂O microcrystals were further utilized in a non-enzymatic amperometric glucose sensor. And the sensor constructed by the extended hexapod Cu₂O microcrystals show the best performance, exhibiting remarkable sensitivity (97 μA mM⁻¹ cm⁻²), significant selectivity and a wide linear response (up to 14.3 mM) towards glucose detection. Compared with the previous sensors that were constructed by the Cu-based materials, this detection range is much closer to the glucose range in human serum. The wide range can be ascribed to the “clean surface” (with no organic capping agent adsorbed on the surface) and more rich {111} facets exposed for the extended hexapod structure, which maximize the accessible electroactive surface for the efficient transfer of electrons, as well as the product molecules. This work provides a green and feasible approach to enhance the Cu₂O sensor performance, which can be extended to other applications such as solar-energy conversion and catalysis.

---

Introduction

As nano/microtechnology evolves, tailoring the architecture of nano/microcrystals has recently been of great interest.^1,2 Different crystallographic planes of a certain crystal usually possess distinctive electronic structures, bondings, surface energies, and chemical reactivities, which determine its shape-dependent chemical and physical properties. Therefore, nano/microcrystals enclosed by single-crystal planes are being more and more anticipated to maximize the special properties of such special planes. In order to enlarge the area of such special planes for nano/microcrystals, a special polyhedral structure with an identical lattice plane as the external surface may be the best choice.³ From the thermodynamic viewpoint, the shape of a polyhedral nano/microcrystal is strongly determined by the different free energies for various facets. Introducing a capping agent is an effective approach to the shape control of polyhedral structures by manipulating the free energy of different facets effectively.⁴ Nevertheless, this approach would restrict the performance of polyhedral nano/microcrystals in areas such as catalysis and sensors. This may be derived from the capping agent covered on the external surface of the polyhedron, which hampers the efficient electron transfer. Thus, it is desirable to find a feasible approach to obtain the polyhedral nano/microcrystal with more desired planes, as well as a “clean surface”.

In recent years, diabetes mellitus has become one of the major diseases worldwide. Various materials, including noble metals or their alloys,^5–9 transition metal and transition metal oxides,^10–12 have been fabricated to construct non-enzymatic glucose sensors with good performance for their importance in clinical diagnosis.¹³ Particularly, Cu-based nano/micromaterials (Cu, Cu₂O, and CuO) have gained much attention due to their low cost and the electrocatalytic effect mediated by the Cu(III)/Cu(II) redox couple.^14–17 Although they have reasonable sensitivity toward glucose, some Cu-based nano/micromaterials obtained by conventional routes cannot be well applied in practice because of their confined detection ranges.^18–21 Up to now, multiple efforts have been directed to enhance the sensor performance by synthesizing Cu-based composite materials.²² However, composite materials with a wide detection range and high sensitivity, are confronted with complicated synthesis processes, high costs and low yields. Therefore, it is especially important to find a facile strategy to construct a Cu-based sensor with enhanced performance. In this study, we have successfully synthesized polyhedral Cu₂O microcrystals, which exhibit their morphological evolution from a concave cubic, multiple-branching, extended hexapod, and finally to a short hexapod shape by using the additive, KBr. Experimental results suggest that the bromide ions could control the growth of Cu₂O crystals observably. Furthermore, the four kinds of Cu₂O microcrystals obtained were used to construct non-enzymatic glucose sensors, and their performance varied with the different surfaces of the microcrystals. The extended hexapod Cu₂O shows the widest linear response range (up to 14.3 mM) towards glucose oxidation. This can result from the rich exposed {111} facets of the extended hexapod shape, which increase the exposure of ‘Cu’ atoms with dangling bonds.²³ Cu₂O obtained by our method presents a wider detection range than the same shape mediated by organic surfactants, which can be applied to the detection of human serum.

Experimental

Reagents

Sodium carbonate decahydrate (Na₂CO₃·10H₂O), trisodium citrate dihydrate (C₆H₅Na₃O₇·2H₂O), copper sulfate pentahydrate (CuSO₄·5H₂O), and hydrogen peroxide (H₂O₂) were purchased from Beijing Chemical Works (Beijing, China). Glucose, potassium bromide (KBr), polyvinyl pyrrolidone (PVP), potassium sulfate (K₂SO₄), uric acid, L-ascorbic acid and fructose were obtained from the Guoyao Chemical Reagent Company (Shanghai, China). All the chemicals were of analytical grade and used without further purification.

Synthesis of Cu₂O microcrystals

In a typical synthetic process, 2.0 g Na₂CO₃·10H₂O, 1.5 g C₆H₅Na₃O₇·2H₂O, 0–14.0 g KBr, and 72 mL of deionized water were mixed in a round-bottomed glass flask. After the reagents had dissolved completely, 4 mL of 0.68 M CuSO₄ aqueous solution was added in. After stirring for 10 min, 4 mL of 1.5 M glucose solution was added to the solution. The mixture was then transferred into a water bath and kept at a temperature of 80 °C for 15 min. The resulting precipitate was centrifuged at 2000 rpm min⁻¹ for 5 min, washed by rinsing and centrifugation with distilled water and then absolute alcohol several times, and finally dried at 60 °C for 6 h.

Instrumentations

Powder X-ray diffraction (XRD) data was collected on a Shimadzu XRD-6000 instrument, which used Cu Kα radiation at 40 kV and 40 mA. The shape and size of the as-obtained samples were characterized by a Magellan-400 Field emission scanning electron microscope (FESEM, FEI Company). The mid-IR absorption spectrum was collected on a Bruker Vertex80 V FT-IR spectrometer, with a resolution of 4 cm⁻¹.

Electrochemical measurements

All electrochemical measurements were carried out on a CHI 660E electrochemical workstation (Shanghai Chenhua Apparatus, China) under an ambient temperature. A traditional three-electrode system was set up with a catalyst-coated glassy carbon electrode (GCE) (d = 3 mm) as the working electrode, a Pt wire with a diameter of 1 mm as the counter electrode and a AgCl electrode as the reference electrode.

Firstly, the GCE was polished with 0.05 μm alumina slurry and rinsed thoroughly with deionized water. To modify the working electrode, the obtained Cu₂O microcrystals were dispersed into the deionized water to form a 2 mg mL⁻¹ suspension. 6 μL of this suspension was added onto the GC electrode surface, and 5 μL of 0.5 wt% Nafion (Nf) solution was dropped onto the GC electrode when the suspension had dried naturally. The modified electrode (Nf/Cu₂O/GCE) could be obtained when the Nf solution dried. To ensure the durability of the modified electrode, 100 cycles of cyclic potential sweeps at a sweep rate of 0.25 V s⁻¹ were been carried out before each measurement.

Results and discussion

Investigation of the formation mechanism

As a representative, Fig. 1 displays the XRD pattern of the extended hexapod shaped Cu₂O and the standard card (JDPDS 05-667), indicating that all the diffraction peaks are readily indexed to the cubic phase Cu₂O without any impurities, such as metallic copper or cupric oxide. The strong and sharp peaks demonstrate that the obtained Cu₂O microcrystals are highly crystalline. Field emission scanning electron microscopy (FESEM) images clearly demonstrate the interesting shape evolution of Cu₂O achieved by simply mediating the amount of KBr. As shown in Fig. 2, the architecture of Cu₂O evolved from a concave cubic shape, a multiple-branched shape, an extended hexapod shape and finally to a short hexapod shape with increasing amounts of KBr. Moreover, the average edge length of the Cu₂O microcrystals increased gradually in this evolution process.

Fig. 1 XRD pattern of the extended hexapod Cu₂O and the standard data for Cu₂O (JDPDS 05-667).

Fig. 2 FESEM images of the Cu₂O microcrystals synthesized at 80 °C by adding various amounts of KBr: (a) 0 g; (b) 0.5 g; (c) 5.0 g; (d) 14.0 g.

As presented in Fig. 2a, the obtained Cu₂O takes on a concave cubic shape with an average edge length of about 700 nm when no additive was added. Multiple-branched Cu₂O microcrystals can be observed when a small amount of KBr was added (Fig. 2b). This shape could be caused by the growth of the initial small nanoparticles via Ostwald ripening. And the ripening time is thought to be extended in the presence of KBr. However, when 5.0 g KBr was added, the initial aggregation of these nanoparticles changed and developed into the extended hexapod shape (Fig. 2c). As documented in the literature, chloride-ions can be strongly adsorbed on the {100} facets of the Cu₂O nanocrystals and thus reduce their surface energies.²⁴ Similarly, when the mass of KBr was increased to 14.0 g, short hexapod Cu₂O can be observed (Fig. 2d). This observation suggests that the growth of the [100] axis was restrained in the presence of a large quantity of bromide ions. In order to prove this viewpoint, the growth processes of each shape were monitored by the FESEM images.

Fig. 3 demonstrates the growth processes of the extended hexapod Cu₂O. We can find that the length of each branch of the microcrystal extends from 1.0 to 2.9 μm as the reaction proceeded. This length did not increase further after the 6th min, indicating that the final shape had formed. To distinguish the geometrical shape of the extended hexapod shape from the short hexapod shape, the parameter C was defined by b/a (Fig. S3c†). The C value for the extended hexapod Cu₂O was about 4.60 (Fig. 3), while it was about 2.38 for the short hexapod (Fig. S3b†). Growth processes of other morphologies are shown in Fig. S1–S3,† and a similar regularity can be observed. Based on the above analysis, the bromide ion is considered to play a dual role in mediating the morphology of the Cu₂O microcrystals. One role is to manipulate the initial small nanoparticles aggregation, and the other is to extend the ripening time and restrain the growth along the [100] axis. To demonstrate that the potassium ion provided by KBr is not responsible for the shape evolution of Cu₂O, we replaced KBr with K₂SO₄. The morphology of Cu₂O had no significant change in the presence of K₂SO₄ (Fig. S4†). This result also demonstrates that different types of anions play diverse roles in controlling the shape of nano/micromaterials. To make a brief illustration, the whole morphological evolution of our system is shown as Scheme 1.

Fig. 3 FESEM images of the Cu₂O formed at: (a) 2 min; (b) 4 min; (c) 6 min; (d) 90 min, when 5.0 g KBr was added.

Scheme 1 Schematic illustration for the growth of the Cu₂O microcrystals as a function of the KBr amount.

Non-enzymatic sensor performance

To determine the electrochemical catalytic performance of the as-obtained samples, glucose sensors were constructed. Typically, 0.1 M KOH was chosen as the supporting electrolyte to analyze the sensing glucose at Nf/Cu₂O/GCE. For clarity, samples A, B, C, and D were designated, which were the concave cubic, multiple-branched, extended hexapod and short hexapod Cu₂O, respectively. Fig. 4 exhibits the cyclic voltammograms (CVs) of the four electrodes. An obvious reduction peak can be found at about +0.6 V (vs. Ag/AgCl), which may be assigned to the Cu(III)/Cu(II) redox couple. The distinct oxidation peak at +0.43 V (vs. Ag/AgCl) may correspond to the oxidation of the carbohydrate–Cu(I) chelate.^25–27 The Cu(III) species are crucial for the oxidation of glucose and thought to act as an electro-transfer mediator.^21,28 Visually, the reduction peak at a potential of +0.6 V is weakened, while the oxidation peak at about +0.43 V appeared gradually with increasing glucose concentration, which confirms that the Cu(III) species are consumed in the glucose oxidation process. During the reductive sweep, the current density at +0.6 V did not climb when the glucose concentration exceeded a certain limitation. This behaviour can be attributed to the chemisorption of poisoning intermediates on the electrode surface under high glucose concentrations and thus bury the active sites.⁶ Compared with the other samples, sample C is the last poisoned, for which the current density at +0.6 V shows a sustainable increasing trend before the concentration reached 14.5 mM (Fig. 4c). This result reveals that the extended hexapod Cu₂O may exhibit a higher electro-catalytic activity to glucose oxidation. In addition, the CVs at different scan rates in alkaline conditions, demonstrate that the electron transfer of the modified electrode is controlled by surface adsorption (Fig. S5†).^29–31

Fig. 4 CVs of the four samples (Nf/Cu₂O/GCE) under different glucose concentrations: (a) concave cubic Cu₂O; (b) multiple-branched Cu₂O; (c) extended hexapod Cu₂O; (d) short hexapod Cu₂O. The scan rate is 0.25 V s⁻¹. Insets: the relationships between the current density at +0.6 V (reductive sweep) and the glucose concentration.

A further comparative study was conducted by the amperometric responses to different concentrations of glucose, for which +0.6 V was chosen as the applied potential. As illustrated in the Fig. 5, the response current density and glucose concentration show similar linear relationships to the four obtained electrodes. However, the linear ranges of these electrodes are different, implying their different sensor performances in the detection range. To ensure the reliability of the measurement, for one individual sample, we fabricated two other electrodes and evaluated the average sensing performance. The sensitivity and linear range of each sample for glucose detection are summarized in Table 1. The result for each measurement is shown in Table S1 and S2 .† Moreover, the insets of Fig. 4 and 5, with a high resolution, are shown in the Fig. S6 and S7,† respectively.

Fig. 5 a–d) Amperometric responses of the obtained electrodes (Nf/Cu₂O/GCE) by successive addition of glucose at the applied potential of +0.6 V. (e) The response current density vs. glucose concentration for different samples. The dashed lines indicate the calibration curves for each sample. Insets: shape model of the corresponding sample.

Table 1 Summary of the sensitivities and detection ranges for the four electrodes for glucose detection

Sample	A	B	C	D
Range (mM)	10.3	8.7	14.3	10.6
Sensitivity (μA mM^−1 cm^−2)	94.3	116.1	97	92.3

---

As seen from the data in Table 1, the four samples show no big difference in sensitivity. This is mainly due to the fact that all the samples were prepared by the same approach. However, the detection range varies with the shape of the Cu₂O microcrystals. According to an integrated analysis, sample C shows the widest detection range up to 14.3 mM and a remarkable sensitivity toward glucose. This conclusion is consistent with the result obtained in Fig. 4, which further demonstrates the credibility of the data. The detection limit for the sensor constructed by sample C is 0.33 μM (with a signal to noise ratio of 3), which is eligible in the field of non-enzymatic sensors.

In a Cu₂O typical unit, as is well known, each ‘O’ atom is surrounded by a tetrahedron of ‘Cu’ atoms, while each ‘Cu’ atom has two ‘O’ atom neighbors. For the {111} planes, every two ‘Cu’ atoms have a dangling bond perpendicular to the {111} planes, whereas for {100} and {110} planes, ‘O’ and ‘–O–Cu–O–Cu–’ terminated surfaces are present, respectively.²³ Obviously, the unsaturated ‘Cu’ in the {111} facet is more active. As reported in the previous literature, the hexapod structure exhibited a more active and effective catalytic performance than the octahedral structure, which is because of their increased {111} facets.³² The wide detection range of the extended hexapod shape may be ascribed to the more rich exposed {111} facets, which would increase the exposure of ‘Cu’ atoms with dangling bonds and facilitate the formation of Cu(III) in the alkaline condition.

To make a contrast, the performances of other non-enzymatic glucose sensors that were constructed by the Cu-based materials are summarized in the Table 2. Compared with the previous reports, sample C also demonstrates a wider detection range, which would be much closer to human physiological levels. And the sensor in our work, from the point of view of preparation costs or environmental friendliness, is practical and rational.

Table 2 Comparison of the fabricated non-enzymatic glucose sensor performance in our work with other reported works

Electrode material	Linear range (up to mM)	Sensitivity (μA mM^−1 cm^−2)	Detection limit (μM)
1 CuO nanoparticles^18	0.17	246	0.91
2 Flower-like CuO^19	1.6	5368	1.20
3 CuO/graphene nanocomposite^33	2.0	2939	0.09
4 CuO/carbon nanotube^34	3.0	2190	0.80
5 Stalactite-like Cu micropillar^14	4.7	2432	0.19
6 Cu/graphene sheets^17	4.5	157	0.50
7 Cu nanocubes/carbon nanotube^35	7.5	1096	1.00
8 Porous Cu2O microcubes^20	0.5	1002	0.8
Our work extended hexapod Cu2O	14.3	97	0.33

---

As is known to all, the detection of glucose will often be affected by some interferences, such as uric acid, fructose, and L-ascorbic acid in human serum. Selectivity of the sensor constructed by the electrode modified by sample C was investigated by the amperometric responses. It has been reported that the normal physiological level for glucose is 3–8 mM, while for interferences this is about 0.05–0.1 mM.¹⁶ Here, interference tests of the fabricated non-enzymatic sensor (constructed by sample C) were carried out by adding 1.0 mM uric acid, 0.5 mM fructose and 0.5 mM L-ascorbic acid to 0.1 M KOH. The interferences we added were 10 times more than normal physiological levels, which aims to exclude the special habitus. As shown in Fig. 6, the response signals to the interferences are much weaker than to 3.0 mM glucose. The result demonstrates that the interferences in human serum are negligible in this detection.

Fig. 6 The response signals for interferences and glucose at a potential of +0.6 V in 0.1 M KOH.

Repeatability is an important factor for sensor performance. We evaluated the repeatability by monitoring the amperometric current response signals to 0.3 mM glucose on an electrode modified by sample C. The relative standard deviation is 2.88% via four parallel measurements, revealing the repeatability of this sensor. Remarkably, Cu₂O microcrystals still remain in their original shape after glucose detection (Fig. S8†), suggesting their good stability.

To investigate the main reason for the wide detection range for glucose in this work, Cu₂O microcrystals with hexapod shapes were synthesized by using PVP as the surfactant (Fig. S9†).³⁶ It is noteworthy to mention that the synthesizing process of the hexapod shape Cu₂O microcrystals is similar with our method, except that PVP was used as the surfactant. Here, these Cu₂O microcrystals were designated as sample E. The amperometric response of the electrode modified by sample E at +0.6 V with successive addition of glucose is shown in Fig. S10.† The response signal remains unchanged when the concentration of glucose exceeded 1.0 mM, indicating a narrow detection range. This contrast, when compared with the electrode modified by sample C, may be because the surface of sample E is capped by PVP. And the capped surface cannot provide enough electroactivity for the transfer of electrons and interaction sites for the oxidation of glucose. From the figures above, we can conclude that the main cause for this wide detection range in our work can be attributed to the “clean surface”, with no organic or polymers surfactant adsorbed on the surface (verified in Fig. S11†). The “clean surface” enhances the glucose sensor performance by maximizing the accessible electroactive surface and affiliating the transfer of electrons.

Conclusions

In this work, Cu₂O microcrystals with a unique surface and well-defined morphology have been synthesized by inducing bromide ions via a facile solution-phase route. The correlation between the shape evolution of the Cu₂O microcrystals and the role of bromide ions has been investigated systematically. Compared with other non-enzymatic glucose sensors constructed by the Cu-based materials, the sensor constructed by the “clean surface” Cu₂O microcrystals not only exhibit a wide detection range but can also reduce the preparation cost effectively.

Acknowledgements

This work is supported by NSFC (No. 91227202 and 51202086), RFDP (No. 20120061130006), and China Postdoctoral science foundation (No. 2013T60325 and 2014M561281).

References

1. Y. Cui and C. M. Liebe, Science, 2001, 291, 851–853.
2. Y. Sun and Y. Xia, Science, 2002, 298, 2176–2179.
3. J. Nai, Y. Tian, X. Guan and L. Guo, J. Am. Chem. Soc., 2013, 135, 16082–16091.
4. M. H. Huang and P.-H. Lin, Adv. Funct. Mater., 2012, 22, 14–24.
5. J. H. Yuan, K. Wang and X. H. Xia, Adv. Funct. Mater., 2005, 15, 803–809.
6. M. Q. Guo, H. S. Hong, X. N. Tang, H. D. Fang and X. H. Xu, Electrochim. Acta, 2012, 63, 1–8.
7. Y. Bai, Y. Sun and C. Sun, Biosens. Bioelectron., 2008, 24, 579–585.
8. P. Holt-Hindle, S. Nigro, M. Asmussen and A. Chen, Electrochem. Commun., 2008, 10, 1438–1441.
9. Y. Fu, F. Liang, H. Tian and J. Hu, Electrochim. Acta, 2014, 120, 314–318.
10. Y. Ding, Y. Wang, L. Su, H. Zhang and Y. Lei, J. Mater. Chem., 2010, 20, 9918–9926.
11. Y. Ding, Y. Wang, L. Zhang, H. Zhang and Y. Lei, J. Mater. Chem., 2012, 22, 980–986.
12. Y. Ding, Y. Wang, L. Su, M. Bellagamba, H. Zhang and Y. Lei, Biosens. Bioelectron., 2010, 26, 542–548.
13. E. H. Yoo and S. Y. Lee, Sensors, 2010, 10, 4558–4576.
14. M.-M. Guo, Y. Xia, W. Huang and Z. Li, Electrochim. Acta, 2015, 151, 340–346.
15. G. Wang, Y. Wei, W. Zhang, X. Zhang, B. Fang and L. Wang, Microchim. Acta, 2009, 168, 87–92.
16. Z. Zhuang, X. Su, H. Yuan, Q. Sun, D. Xiao and M. M. Choi, Analyst, 2008, 133, 126–132.
17. J. Luo, S. Jiang, H. Zhang, J. Jiang and X. Liu, Anal. Chim. Acta, 2012, 709, 47–53.
18. Y. Li, Y. Wei, G. Shi, Y. Xian and L. Jin, Electroanalysis, 2011, 23, 497–502.
19. K. Li, G. Fan, L. Yang and F. Li, Sens. Actuators, B, 2014, 199, 175–182.
20. L. Zhang, H. Li, Y. Ni, J. Li, K. Liao and G. Zhao, Electrochem. Commun., 2009, 11, 812–815.
21. S. K. Meher and G. R. Rao, Nanoscale, 2013, 5, 2089–2099.
22. X. Zhang, G. Wang, W. Zhang, Y. Wei and B. Fang, Biosens. Bioelectron., 2009, 24, 3395–3398.
23. D.-F. Zhang, H. Zhang, L. Guo, K. Zheng, X.-D. Han and Z. Zhang, J. Mater. Chem., 2009, 19, 5220–5225.
24. M. H. Kim, B. Lim, E. P. Lee and Y. N. Xia, J. Mater. Chem., 2008, 18, 4069–4073.
25. E. Reitz, W. Z. Jia, M. Gentile, Y. Wang and Y. Lei, Electroanalysis, 2008, 20, 2482–2486.
26. P. Luo, S. V. Prabhu and R. P. Baldwin, Anal. Chem., 1990, 62, 752–755.
27. J. M. Marioli and T. Kuwana, Electrochim. Acta, 1992, 37, 1187–1197.
28. H. Wei, J. J. Sun, L. Guo, X. Li and G. N. Chen, Chem. Commun., 2009, 2842–2844.
29. S. Li, Y. Zheng, G. W. Qin, Y. Ren, W. Pei and L. Zuo, Talanta, 2011, 85, 1260–1264.
30. H. Y. Zhao, W. Zheng, Z. X. Meng, H. M. Zhou, X. X. Xu, Z. Li and Y. F. Zheng, Biosens. Bioelectron., 2009, 24, 2352–2357.
31. F. Xu, M. Deng, G. Li, S. Chen and L. Wang, Electrochim. Acta, 2013, 88, 59–65.
32. J.-Y. Ho and M. H. Huang, J. Phys. Chem. C, 2009, 113, 14159–14164.
33. T. Alizadeh and S. Mirzagholipur, Sens. Actuators, B, 2014, 198, 438–447.
34. J. Yang, L. C. Jiang, W. D. Zhang and S. Gunasekaran, Talanta, 2010, 82, 25–33.
35. J. Yang, W. D. Zhang and S. Gunasekaran, Biosens. Bioelectron., 2010, 26, 279–284.
36. Y. Sui, W. Fu, H. Yang, Y. Zeng, Y. Zhang, Q. Zhao, Y. Li, X. Zhou, Y. Leng, M. Li and G. Zou, Cryst. Growth Des., 2010, 10, 99–108.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed information on the FESEM images of the Cu₂O formed at different stage, mid-IR absorption, electron transfer type, and amperometric responses of the electrode E. See DOI: 10.1039/c5ra08586a

---