Self-assembly of porous copper oxide hierarchical nanostructures for selective determinations of glucose and ascorbic acid

The simple design of CuOmicro-/nanostructures has recently attracted tremendous interest particularly for the enzyme-less sensing of biological molecules due to their intrinsic electronic and catalytic properties. Consequently attention has been directed to the development of new CuO nanomaterials that have multi-interdisciplinary applications. Herein, we report for the first time the fabrication of hierarchical porous CuO micro-/nanostructures with flowerand hollow sphere-like morphology via a facile hydrothermal method. Our experimental findings clarify that the source of the copper-ions effectively control the assembly of CuO nano-building blocks via the one-step hydrolysis of [Cu(NH3)4(H2O)2]SO4 and [Cu(NH3)4(H2O)2]Cl2 precursors, which produce hollow sphere and flower-like morphologies for sensitive and selective determination of ascorbic acid and glucose, respectively. Moreover, such unique properties of macro-/mesoporous CuO with defined dimensions and topologies offer minimized diffusive resistance for the dispersion of active sites. The best performance of the glucose and ascorbic sensor can be obtained at +0.55 V in 0.1 M sodium hydroxide solution. The as-prepared CuO modified (drop-casted) screen-printed electrodes (SPE) exhibit a fast electroactive response with high sensitivity within a wide concentration range of glucose and ascorbic acid in real samples. Significantly, the aniondependent approach might be used to control effectively the expansion and features of other metal oxide micro-/nanostructures.

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal's standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

Introduction
The synthesis of transition metal oxide nanostructures with controlled shapes and sizes has been developed over recent decades due to their shape/size-dependent physical, chemical, electrical, optical, and catalytic properties. 1,2 They exhibit versatile functions of their energy gap, numerous valence states, and tunable performances for electron transport. Since, their complex morphology and high dimensionality reveal advanced geometric structure and atom arrangement on the specific facets which provide novel characteristics. Up-to-date, hierarchical metal oxide micro-/nanostructures have attracted considerable attention, because of their technological and fundamental scientific importance and potential applications; including chemical and electrochemical sensing, catalysis, nanoelectronics and nanophotonics. [3][4][5][6] These hierarchical porous micro-/nanostructures offer minimized diffusion resistance, high surface area and good conductivity. 7 Therefore, numerous hierarchical metal oxides micro-/nanostructures have been recently reported. [8][9][10][11][12] Direct hydrothermal via hydrolysis of metal salts and sacrificial templates strategies have been widely used to control fabrication of hierarchical structures, yet technical challenges in terms of costintensive and time-consuming synthetic conditions remain. In this respect, development of suitable synthetic strategies is highly desirable.
properties. Such properties have allowed for its wide application within many disciplines ranging from conversion of solar energy, 13 lithium ion batteries, 14 supercapacitors, 15 heterogeneous catalysis, 16 and electrochemical sensor. 12,17 Among these potential applications, the exploration of enzyme-less electrochemical sensor especially for glucose has been intensively investigated. 17,18 Successful synthesis strategies have been employed to control the morphology of CuO nanostructures as nanowires, 19 nanobelts, 20 nanoplatelets, 21 nanospheres, 22 nanorods, 23 and nanofibres. 24 Recently, a new class of hierarchical CuO micro-/nanostructures has been studied for the extension of its sensing applications. 12,19,25 Sun et al., 18 reported the fabrication of hierarchical CuO nanoribbons using a green water/ethanol solution-phase hydrolysis of Cu x (OH) 2x−2 (SO 4 ) precursors. Such nanoporous CuO ribbons show a high sensitivity, low detection limit, and a good selectivity towards glucose. However, the relationships between the copper ion source, additives and morphology are not clear yet.
Diabetes is a chronic disease that occurs usually when the pancreas does not produce sufficient insulin. Unfortunately, about 387 million over the world living with diabetes, more than 77 % of diabetes deaths occur in developing countries. 26 Certainly, several efforts have focused upon developing of a cheap, stable and reliable sensor for self-monitoring of blood glucose levels as one modality that can help people to improve their glycemic control. 27 Furthermore, ascorbic acid known as vitamin C, is a powerful antioxidant, which naturally occurs within citrus fruits and vegetable products. Significantly, the determination of ascorbic acid within biological fluids is very important to assess the amount of oxidative stress in the human metabolism, a process that has been linked to cancer, diabetes mellitus and several liver diseases. 28 Over recent decades, the allure of electrochemical techniques has provided a broad array within environmental monitoring and clinical diagnostics. The development of an enzyme-less electrode has attracted considerable attention due to its simple, low fabrication cost, long term stability and good reproducibility which are the limitations of those enzymatic biosensors. However, the selectivity of the electrode sensor for target analytes is still great challenge. 29 In this manuscript, we report a novel, effective design structure based on the 'bottom-up' self-assembly of elementary nanobuilding blocks of CuO nanostructures. The effect of copper ions source upon the CuO nanostructures assembly is explored for the first time. These unique micro/-nanostructures are then utilised to produce disposable screen-printed electrodes. Additive-free, largescale production and simple fabrication strategy of hierarchical CuO micro-/nanostructures were developed via one-step hydrolysis of [Cu(NH 3 ) 4 (H 2 O) 2 ]SO 4 and [Cu(NH 3 ) 4 (H 2 O) 2 ]Cl 2 precursors without any post-thermal treatment (Scheme 1). Such results show that this anion-dependent approach would effectively optimize the selfassembly of CuO nanoneedle and nanoplate building blocks to construct hallow sphere-and flower-like morphology, respectively. Interestingly, the nano-building block assembly of CuO micro-/nanostructures is of great significance in green design of simple, low-cost, and sensitive enzyme-less SPE for selective determination of glucose and ascorbic acid in real samples.

Scheme 1:
A schematic representation of the hydrothermal pathway of the copper-ion source hydrolysis approach.

Synthesis of CuO micro-/nanostructures.
The fabrication of CuO flower-and sphere-like morphologies were carried out utilising a typical hydrothermal procedure, with 1.70 g of copper chloride dihydrate and 1.59 g of anhydrous copper sulphate precursors respectively dissolved (under constant stirring) in 50 ml deionized water. The solutions were introduced into a 100 ml Teflon-lined, stainless steel autoclaves. Then, the solutions were adjusted to pH 10 using ammonium solution (NH 4 OH, 28%). The autoclaves were sealed and maintained at 160 °C for 12 hours. Subsequently, these solutions were kept at room temperature. The black precipitates were collected, washed several times with water and ethanol to remove the remaining agents, and then dried at 45 °C for 6h hours.

Fabrication of Screen-Printed Electrodes (SPEs).
The screen-printed graphite electrodes were fabricated at Manchester Metropolitan University utilising appropriate stencil designs with a microDEK 1760RS screen-printing machine (DEK, Weymouth, UK). For each of the screen-printed sensors a carbongraphite ink formulation (Product Code: product code: C2000802P2; Gwent Electronic Materials Ltd, UK) was first screenprinted onto a polyester flexible film (Autostat, 250 µm thickness). 30

Fabrication of CuO electrodes.
The CuO micro/nanostructure "paste" was prepared as previously reported by Grätzel et al. 33 Briefly, 30.0 mg of CuO sample was mixed and sonicated for 10 min with a mixture of 5 µL acetic acid, 5 µL deionized water and 17.5 µL ethanol in glass beaker (1). A mixture (2) containing 100 µg of ethyl cellulose dissolved in 150 µL ethanol and 135 µL water were introduced into glass beaker (1). Then, 5 µL of the paste was drop-casted onto the screen-printed electrode (SPE) surface and left it to dry in oven at 50 o C for 24 hours.

Amperometric measurements
The CuO-SPEs were activated in 0.1 M NaOH solution by successively sweeping from −1.0 to 0.8V unƟl the electrochemical signal became stable. A holding potential of + 0.55 V was applied to the CuO-SPEs and the background current was allowed to decay to a steady state. The amperometric response was recorded as a function of glucose or ascorbic acid concentrations in the cell containing 20 ml of 0.1 M NaOH solution.

Real sample analysis
Human blood and urine samples were collected from volunteers in hospital of Sohag University. The blood samples were centrifuged (3500 rpm/ 5min), The supernatant of which was blood serum was obtained. 200 µL blood serum sample was injected into 19.8 mL of a stirring solution of 0.1 M NaOH. The amperometric current response was recorded using CuO MF-SPE and a standard addition method was used to determine the glucose concentration. As well as, 200 µL urine sample was injected into 19.8 mL of a stirring solution of 0.1 M NaOH. The amperometric current response was recorded using CuO HS-SPE and a standard addition method was used to determine the ascorbic acid concentration.

Characterization of CuO micro-/nanostructures
The morphology of the as-synthesized CuO samples were investigated using field emission scanning electron microscopy (FE-SEM, JEOL model 6500). Before insertion into the chamber, the CuO powders were ground and fixed onto a specimen stub using doublesided carbon tape. Then, a 10 nm Pt film was coated via anion sputtering (Hitachi E-1030) at room temperature to obtain highresolution micrographs. For better recording of FE-SEM images, the SEM was operated at 15 KeV.
High-resolution transmission electron microscopy (HR-TEM), electron diffraction (ED), and energy dispersive X-ray spectroscopy for elemental analysis (EDS) were performed using a JEOL JEM model 2100F microscope. HR-TEM was conducted at an acceleration voltage of 200 kV to obtain a lattice resolution of 0.1 nm. The HRTEM images were recorded using a CCD camera. In the HRTEM, ED and EDS characterization, the CuO samples were dispersed in ethanol solution using an ultrasonic bath, and then dropped on a copper grid. Prior to inserting the samples into the HRTEM column, the grid was vacuum dried for 20 min.
Wide-angle powder X-ray diffraction (XRD) patterns were measured X-ray diffractometer (Model FW 1700 series, Philips, Netherlands) using with monochromated CuKα radiation (λ = 1.54 Å), employing a scanning rate of 0.06°/min and 2θ ranges from 4° to 80°. The diffraction data were analysed using the DIFRAC plus Evaluation Package (EVA) software with the PDF-2 Release 2009. The Fourier transform infrared (FTIR) spectra of the CuO samples were recorded using Bruker Alpha FTIR instrument.
The textural surface properties and pore size distribution was determined by N 2 adsorption/desorption isotherms at 77 K with a BELSORP36 analyzer (JP. BEL Co., Ltd.). The specific surface area (S BET ) was calculated using the Brunauer-Emmett-Teller (BET) method with multipoint adsorption data from the linear segment of the N 2 adsorption isotherm. The pore size distribution was determined from the analysis of desorption branch of isotherm using the nonlocal density functional theory (NLDFT).
The electrochemical experiments were performed using Autolab 302N potentiostat/galvanostat workstation. All measurements were conducted using a screen-printed electrode configuration. During the development of the protocol, the 3 mm CuO modified graphite working electrode of the screen-printed electrode was used with a platinum counter electrode and Ag/AgCl as reference electrode. Connectors for the efficient connection of the screen-printed electrochemical sensors were purchased from Kanichi Research Services Ltd (UK). 34

Formation mechanism of hierarchical CuO micro-/nanostructures
The copper oxide micro/-nanostructures were fabricated as described in the experimental section. Control over the particle morphology of CuO was clearly evident by using one-pot, simple hydrothermal synthesis (Scheme 1). Interestingly, it was noted that, the copper-ion precursors were effectively controlling the selfassembly of CuO nanoparticles within the hierarchical structures.  Figure 1 (A-D), show the FE-SEM images of hierarchical CuO microstructures prepared without any additives. The hydrolysis of CuCl 2 in the presence of NH 4 OH at pH 10 reveals uniform microbunches of CuO nanoplate clusters assembled in flower-like morphology (Fig. A). The bunches of these microflowers (MF) have an average diameter of ∼ 5 µm, which the thickness of the aggregated plates ∼ 220 nm. Such high magnification images reveal the formation of densely aggregated nanoplate clusters, with smooth fine surfaces that were aligned according to uniform flower morphology (Fig. B). Moreover, it is clearly evident that there is an interlayer spacing as macropores with slit-shape structure between the nanoplate arrays, enhancing the diffusion of ions and molecules to their respective active sites of the CuO crystals. While, the hydrolysis of CuSO 4 in the same conditions reveals the formation of a hallow sphere (HS) consisting of closely packed nanoneedle arrays (Fig. 1C). The average size of aggregated needles is ~ 230 nm, the diameter of sphere is ~ 5.60 µm, and open pores are ~ 1.66 µm. Interestingly, the needles are obviously rough and is an assembly of a large number of nanoneedles and voids, leading to the formation of hierarchical hallow sphere-like morphology. Overall, the SEM images demonstrate the versatility of metal-ion source approach over effectively control of the nanoparticle shape and their assembly in hierarchical microstructures.
The High-resolution transmission electron microscopy (HR-TEM) images [ Fig. 2] clearly indicates that, the CuO MFs and HSs are composed form assembled plates and needles nanoblocks respectively with disordered porous network. These HR-TEM images showed the formation of CuO nanoplates and nanoneedles firstly and then assembled together to form CuO microstructures with flower-and hollow sphere -like morphology. According to the distinct color contrast of TEM images associated with flower structure (Fig. 2(a-b)), we could detect that the triangular nanoplate in the core region of flower were more compact than those in the exterior. Fig. 2b is a typical HR-TEM image of an individual CuO nanoplate, and it can be interestingly found that the plate is very thin and obviously rough with size about 270 nm. Moreover, The most prominent feature was that the plate showed uniform arrangements and continuous ordering of lattice fringes over largescale regions without distortion. The TEM image (Fig.2c), in general, revealed well-organized lattice fringe arrays over a large area with a distance between two lattice fringes of 0.23 nm that features the interplanar space of (111) facets of monoclinic crystal lattices.   Figure 3a shows the wide angle XRD patterns of the hierarchical CuO microstructures utilising different Cu-ion precursors. It is clear that they exhibit typical diffraction peaks of monoclinic Tenorite CuO phase of (JCPDS, No. 02 -1040) and no impurities are detected, indicating that the CuO microstructures are pure and well crystalline. Although, the anion in copper ion source has largest effects on CuO morphology, it cause greatly affect on the crystallographic orientation of CuO nanocrystals.  The porous network of hierarchical CuO structures were investigated using nitrogen adsorption/desorption isotherms (Figure 4). The specific surface areas S BET were calculated by employing the Brunauer-Emmett-Teller (BET) method and the pore size distributions were obtained by means of the nonlocal density functional theory (NLDFT) equation using the adsorption isotherm branch. The CuO MFs feature type IV isotherms with a H 2 -type hysteresis loop for typical mesocage materials with large pore cavities. However, CuO HSs show type II isotherms without any hysteresis loops. It is worth noting that the porous CuO samples featured a specific surface area of S BET = 60.37 m 2 g -1 , and 60.24 m 2 g -1 , for CuO MFs and HSs respectively. The pore size distribution show bimodal distribution centred at 4.4 nm and 108 nm for CuO MFs and 5.4 nm and 108 nm for CuO HSs. These results are in accordance with SEM and HR-TEM micrographs ( Figs. 1 and 2). In general, the mesopore within the hierarchical CuO microstructures is probably related to the pores present inside the nanoplates or nanoneedles, which were formed between primary crystallites. However, the macropore is related to assembly of secondary crystals of plates or needles. Such control of hierarchical surface morphologies with well-developed porous network offer great advantageous in electrochemical applications since macropore channels permit fast ions transport, while the mesopores provide more electroactive sites for redox reactions.

Electrochemical characterization of hierarchical CuO nanostructures
The  CuO microstructures possessed variable electrochemical reactivity, and thus, the investigation of the electrochemical sensing platforms will be promising.

Enzyme-less features of hierarchical CuO microstructures
The electrochemical performance of the hierarchical CuO micro-/nanostructures modified SPE were investigated by CV in 20 mL NaOH (0.1 M) solution as the supporting electrolyte. Figure 6a depicts CV curves of CuO microstructures in absence and presence of 5 mM glucose. There is no obvious current signal observed in the absence of glucose for both CuO-SPE electrodes. After injection of glucose, the oxidation current was increased remarkably at + 0.55 V (vs. SCE) for CuO MFs however, in the case of the CuO HSs did not show any significant change. Although they were synthesized via the same route, CuO MFs have a good electrocatalytic activity towards the oxidation of glucose. Interestingly, the aniondependent approach would effectively control the electrocatalytic activity of CuO microstructures. In addition, the previous experiments were conducted within phosphate buffer solution (pH 7.5) instead of an alkaline medium (Fig. S2 within ESI). The CuO microstructures did not show any current response, which indicates that the OHions have a significant role within the electrocatalytic oxidation process. The electrocatalytic oxidation mechanism of glucose onto CuO MFs modified SPE can be described in the following equations (4-6); CuO + OH -→ CuOOH +e -(4) Cu(III) + glucose + e -→ gluconolactone + Cu (II) (5) Gluconolactone →Gluconic acid (6) Significantly, CuO MF was electrochemically oxidized to strong oxidizing agent Cu (III) species (i.e. CuOOH) and then glucose is catalytically oxidized to gluconolactone, which subsequently hydrolysed to gluconic acid. The oxidation of glucose has been also investigated at different scan rates (10 -100 mV/s) (Fig. 6b).
Noticeably, the oxidation current is linearly correlated with square Please do not adjust margins Please do not adjust margins root of scan rate with slightly positive potential shifts; I p (A) = 2.94 × 10 -6 A/(Vs -1 ) 0.5 + 31.5 × 10 -6 A, R 2 = 0.98). This observation reveals that, electrocatalytic oxidation of glucose is governed by a diffusional process.   Furthermore, a chronoamperometric method was utilised to obtain better sensitivity because enhanced analytical signals can be achieved by optimizing the working potential of lower background currents. The effect of applied potential was tested by successive amperometric measurements of 0.5 mM glucose additions into a stirring 0.1 M NaOH at different potentials ranging from + 0.4 V to + 0.6 V (Figure 7b). The largest sensitivity was observed at + 0.5 V which also used for the subsequent amperometric measurements. Figure 8 is a typical amperometric response of the hierarchical CuO MF-SPE electrode with successive additions of a certain concentration of glucose into continuously stirred 0.1 M NaOH solution at an applied potential of + 0.55 V (vs. SCE). Since there was no obvious amperometric response to dissolved oxygen within the NaOH solution upon a CuO MF -SPE, the glucose can be detected in the presence of oxygen. It was clearly seen that the modified electrode exhibits a fast and sensitive response towards the electrocatalytic oxidation of lower and high concentrations of glucose (Figure 8 a, and b). The response time for detection of glucose using CuO MF -SPE is ~ 2.0 seconds (in-set Figure 8b). The amperometric oxidation current increased linearly with increasing concentration of glucose over the concentration ranges 0.05 -3 mM; I p /µA = 0.027 µA µM -1 + 0.42 µA, R 2 = 0.99, N=3; with the detection limits of 10 µM (S/N = 3σ) and the sensitivities were estimated to be 383 µA mM -1 cm -2 , indicating superior electrocatalytic oxidation behaviour of CuO MF toward lower and higher concentration of glucose. Such electrodes show high sensitivities, fast responses, and linear ranges within the realm of blood glucose concentration within the human body (i.e. 2-10 mM) as shown in Figure S3, which is promising in online glucose monitoring compared with other sophisticated sensors reported in the literature.

Reproducibility and Selectivity of enzyme-less CuO nanostructures
The reproducibility of CuO MF -and CuO HS -SPE were investigated by determining 100 µM glucose and ascorbic acid respectively in a stirred solution of (0.1 M) NaOH. Five CuO MF-and CuO HS -SPEs were prepared and tested independently, the relative slandered deviation (RSD) of 2.374 % and 1.79 % were obtained. Moreover, successive amperometric measurements of 100 µM were repeated on the same electrode which yielded a RSD of 3.7 % and 2.67 % for CuO MF-and CuO HS -SPE respectively. Next, attention was turn to perform selectivity experiments as one of the significant analytical factors for an amperometric sensor. The selectivity experiments of glucose in a stirred solution of NaOH were carried out using CuO MF and CuO HS -SPEs in the presence of 1mM of possible interferes such as lactose, maltose, and ascorbic acid. Interestingly, we found that, the CuO MF -and CuO HS -SPE are highly selective for glucose and ascorbic acid respectively at + 0.55 V (vs. SCE). As shown in Figure 9 (a and b), the current responses of CuO MF-and CuO HS -SPEs to the oxidation of glucose and ascorbic acid remain almost constant. A good linear relationship can be obtained up to 6 mM of glucose and ascorbic acid even in presence of co-existing electroactive species. Indicating excellent specific selectivity of these CuO MF-and CuO HS-SPEs towards glucose and ascorbic acid respectively (Fig. 9 (c and d)). Such promising selectivity of CuO microstructures are related to anion-dependent approach which effectively optimize the active surface of self-assembled CuO nano-building blocks in hierarchical structures (Table S4). The selectivity of CuO micro-/nanostructures are related to their surface properties. Ascorbic acid is a well-known reducing agent both in vivo and in vitro. 38 Hence, the oxidation of ascorbic acid is preferentially onto CuO HSs due to relatively higher oxygen-surface content as shown in EDS data. The oxidation process of ascorbic acid can be described in the following steps; 1) the ascorbic acid molecules are firstly diffused and then adsorbed onto the active site of CuO HSs. 2) The adsorbed ascorbic acid molecules are oxidized to dehydroascorbic acid catalysed by the conversion from CuO to Cu 2 OOH on the surface.
Next, our attention was focused on the determination of ascorbic acid using CuO HS in a stirred solution of 0.1 NaOH.( Figure  S5). The successive amperometric additions of ascorbic acid in a stirred (0.1 M) NaOH was shown in Figure 10 (a and b). The amperometric currents increased linearly with ascorbic acid concentrations over the concentration ranges 0.1 -7 mM; I p /µA = 37.24 µA mM -1 + 10.59 µA, R 2 = 0.99, N=3; with the detection limits of 90 µM (S/N = 3σ) and the sensitivity was estimated to be 533 µA mM -1 cm -2 . The CuO HS-SPE shows wider concentration ranges and excellent sensitivity towards the oxidation of ascorbic acid than those reported previously. 39

Real application of glucose measurements
In order to verify the applicability and feasibility of the CuO MF-and CuO HS -SPE for determination of glucose and ascorbic acid respectively. 200 µL blood serum sample or urine sample was added into 19.8 mL 0.1 M NaOH (see experimental section). The amperometric current response was recorded at + 0.55 V (vs. SCE) and a standard addition method was used to determine the glucose and ascorbic acid concentrations. Eight 100 µM of standard solutions of glucose or ascorbic acid were successfully added to the cell for slandered addition determination. Each sample was measured three times to overcome systematic error. The calculated results are summarized in Table 1 and 2. Therefore, the proposed CuO MF-and CuO HS-SPE sensors are efficient glucose and ascorbic acid detection and could be used in practical applications.

Conclusions
In this manuscript, the hierarchical CuO micro-/nanostructures with sphere and flower-like morphology were successfully synthesized via a green solution-phase transformation of Cu 3 (OH) 4 SO 4 and Cu 2 (OH) 3 Cl units in strong alkaline solutions, respectively. A novel hierarchical CuO microstructures modified screen-printed electrodes (SPE) has been designed upon an anion-dependent approach which effectively optimized the catalytic active surface for an enzyme-less determination of glucose and ascorbic acid. The CuO-SPE exhibited high electrocatalytic capabilities with an excellent selectivity and sensitivity, as well as wide linear range, low detection limit, good stability and reproducibility. In addition, the proposed sensors were successfully applied for glucose and