Wearable Noninvasive Glucose Sensor Based on CuxO NFs/Cu NPs Nanocomposites

Designing highly active material to fabricate a high-performance noninvasive wearable glucose sensor was of great importance for diabetes monitoring. In this work, we developed CuxO nanoflakes (NFs)/Cu nanoparticles (NPs) nanocomposites to serve as the sensing materials for noninvasive sweat-based wearable glucose sensors. We involve CuCl2 to enhance the oxidation of Cu NPs to generate Cu2O/CuO NFs on the surface. Due to more active sites endowed by the CuxO NFs, the as-prepared sample exhibited high sensitivity (779 μA mM−1 cm−2) for noninvasive wearable sweat sensing. Combined with a low detection limit (79.1 nM), high selectivity and the durability of bending and twisting, the CuxO NFs/Cu NPs-based sensor can detect the glucose level change of sweat in daily life. Such a high-performance wearable sensor fabricated by a convenient method provides a facile way to design copper oxide nanomaterials for noninvasive wearable glucose sensors.


Introduction
To control the evolution of diabetes mellitus and its dangerous consequences, it is crucial to constantly monitor the in-vivo and in-vitro glucose concentration in the blood or other bodily fluids [1,2]. Beyond using human blood as the analyte for invasive glucose sensing, sweat is also considered a promising target to indicate the blood glucose level owing to the non-invasive epidermal sensing [3,4]. Due to their excellent selectivity, strong anti-interference performance and high compatibility, glucose oxidase and glucose dehydrogenase electrochemical sensing technologies have been the primary foundation for human glucose sensors up to this point [5]. But pH, temperature, and dissolved oxygen all played a significant role in how active enzymes were.
Due to their stability and superior detection capability, several metal materials, especially gold [1,[6][7][8], platinum [9], copper [10,11] and nickel [12][13][14], have gained a lot of emphasis in non-enzymatic glucose sensors [15]. Noble metal materials are pricey, which has hindered their utilization in non-enzymatic glucose sensors [16]. Despite the fact that inexpensive transition metals and their alloys had glucose-sensing properties, surface poisoning proved a significant issue [17]. Copper oxides deliver excellent electrochemical activity [18] and own a sufficient potential for non-enzymatic glucose oxidation [19][20][21][22][23]. Cu(II) oxide and Cu(I) oxide are widely known copper oxides that have outstanding sensing capacities. An electron pair reaction involving the redox couple ions Cu(III) and Cu(II) formed the foundation for the sensing mechanism [24,25]. The continuous transformation of electrons was available by the redox pair ions, which serve as an intermediary mediator [26][27][28][29]. The copper oxide function layer allows the glucose to transfer its released electrons to the current collector [30][31][32].
An abundance of active sites for the electrocatalytic reaction can be provided by a large surface area without the cover of surfactants, thus improving the sensor's sensitivity. This is even more valuable when considering the low glucose concentration in human Sensors 2023, 23, 695 2 of 11 sweat [33]. Our previous work used H 2 O 2 as the oxidant to directly grow the high surface area Cu 2 O/CuO nanosheets on the surface of a copper foil [34,35]. The surface of these nanosheets is active for glucose sensing, as no organic surfactant was involved in the reaction. However, the capability for H 2 O 2 to oxide the Cu foil is limited and a high temperature (180 • C) is required to enhance the oxidation reaction [34]. When the Cu foil was replaced by the Cu nanowires, the generation of nanosheets can be performed at room temperature, but the reaction requires an extended duration [35]. The novel strategy was required to improve the oxidation capability of H 2 O 2 to produce highly glucose-sensitive copper oxide materials for human sweat detection.
In this work, we synthesized Cu x O NFs/Cu NPs nanocomposites with abundant active sites with the help of CuCl 2 and then evaluated their ability for noninvasive wearable glucose sensing. The schematic representation of the synthesis process and its application mechanism in noninvasive glucose detection was demonstrated in Figure 1. The involvement of CuCl 2 can strengthen the oxidation of Cu nanoparticles to Cu x O nanoflakes at room temperature. The Cu x O NFs/Cu NPs nanocomposites based noninvasive wearable glucose sensor delivers high sensitivity, good stability, anti-inferring capability, and durability to mechanical deformation, enabling its application in monitoring glucose level change in human sweat. Our work demonstrates that the as-prepared Cu x O NFs/Cu NPs nanocomposites deliver high performance in sweat glucose sensing, enriching the application of metal oxide in wearable electrochemical sensors. as an intermediary mediator [26][27][28][29]. The copper oxide function layer allows the glucose to transfer its released electrons to the current collector [30][31][32]. An abundance of active sites for the electrocatalytic reaction can be provided by a large surface area without the cover of surfactants, thus improving the sensor's sensitivity. This is even more valuable when considering the low glucose concentration in human sweat [33]. Our previous work used H2O2 as the oxidant to directly grow the high surface area Cu2O/CuO nanosheets on the surface of a copper foil [34,35]. The surface of these nanosheets is active for glucose sensing, as no organic surfactant was involved in the reaction. However, the capability for H2O2 to oxide the Cu foil is limited and a high temperature (180 °C) is required to enhance the oxidation reaction [34]. When the Cu foil was replaced by the Cu nanowires, the generation of nanosheets can be performed at room temperature, but the reaction requires an extended duration [35]. The novel strategy was required to improve the oxidation capability of H2O2 to produce highly glucose-sensitive copper oxide materials for human sweat detection.
In this work, we synthesized CuxO NFs/Cu NPs nanocomposites with abundant active sites with the help of CuCl2 and then evaluated their ability for noninvasive wearable glucose sensing. The schematic representation of the synthesis process and its application mechanism in noninvasive glucose detection was demonstrated in Figure 1. The involvement of CuCl2 can strengthen the oxidation of Cu nanoparticles to CuxO nanoflakes at room temperature. The CuxO NFs/Cu NPs nanocomposites based noninvasive wearable glucose sensor delivers high sensitivity, good stability, antiinferring capability, and durability to mechanical deformation, enabling its application in monitoring glucose level change in human sweat. Our work demonstrates that the asprepared CuxO NFs/Cu NPs nanocomposites deliver high performance in sweat glucose sensing, enriching the application of metal oxide in wearable electrochemical sensors.

Preparation of the Cu x O NFs/Cu NPs Nanocomposites
In a typical procedure, 0.1 g of Cu NPs was dispersed in DI water (10 mL) by 30 min ultrasonication. 1 mL of CuCl 2 ·2H 2 O aqueous solution (0.1 M) was added to the suspension and stirred for 15 min. The obtained products were then centrifuged, washed with absolute alcohol and deionized water three times and dried at a 60 • C vacuum atmosphere for 2 h. Then 0.1 g of dried product was added in DI water (5 mL) and stirred for 30 min. 0.1 M H 2 O 2 aqueous solution (10 mL) was added to the suspension and with further string for 2 h. The final products were then centrifuged, washed with absolute alcohol and deionized water three times, then dried at a 60 • C vacuum atmosphere for 10 h.

Preparation of the CuO NFs/Cu NPs Nanocomposites
0.1 g of Cu NPs was added into DI water (5 mL) and stirred for 30 min. 0.1 M H 2 O 2 aqueous solution (10 mL) was added to the suspension and follower by 2 h stirring. The obtained products were then centrifuged, washed with absolute alcohol and deionized water three times, then dried at a 60 • C vacuum atmosphere for 10 h.

Preparation of the Cu x O NFs/Cu NPs Nanocomposites-Based Wearable Sensor
The Cu x O NFs/Cu NPs nanocomposites-based noninvasive wearable sensor was fabricated on the screen-printed electrodes with Ag/AgCl as the reference electrode, and carbon as both the counter electrode and the working electrode. The substrate is a polyethylene terephthalate film. Typically, 1 mg samples were dispersed in a solution composed of 250 µL isopropanol, 250 µL DI water mixed and 15 µL Nafion solution (isopropanol solution, 5%). The suspension was further mixed under ultrasonic agitation for 2 h and then 2 µL of the mixture was dropped onto the working electrode (diameter is 3 mm), followed by drying at room temperature for 6 h. For wearable application, the intergrade electrode was attached to the BOPP waterproof tape and following covered by PU film.

Electrochemical Measurements
All the electrochemical experiments were conducted using a three-electrode electrochemical system. The electrolyte is a 100 µL aqueous solution of NaOH (0.1 M, pH~13) and NaCl (0.1 M). Cyclic Voltammograms (CVs) were recorded between 0 and 0.7 V. CVs were collected between 0 and 0.1 V for ECSA measurement. The electrolyte used in Electrochemical Impedance Spectroscopy (EIS) measurements is a combination of 0.1 KCl, 0.1 M K 3 [Fe(CN) 6 ] and 0.1 M K 4 Fe(CN) 6 .

Characterization of Materials
After the successive treatment of CuCl 2 aqueous solution and H 2 O 2 aqueous solution, Cu NPs were transferred to the intermediate product of Cu 2 (OH) 3 (111) and (111) planes of Cu 2 O and CuO, respectively. As a comparison, the CuO NFs/Cu NPs nanocomposites without the involvement of CuCl 2 treatment show a similar morphology with that of Cu x O NFs/Cu NPs nanocomposites (Figure S1d-f), suggesting that the H 2 O 2 is the key step to generate the nanoflakes structure. In addition, as shown in Figure S2a-c, the intermediate product of Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites still show the solid particle shape.

Characterization of Materials
After the successive treatment of CuCl2 aqueous solution and H2O2 aqueous solution, Cu NPs were transferred to the intermediate product of Cu2(OH)3Cl/CuCl/Cu NPs nanocomposites and the final products are CuxO NFs/Cu NPs nanocomposites. The morphology of CuxO NFs/Cu NPs nanocomposites was characterized by FESEM (  (111) and (111) planes of Cu2O and CuO, respectively. As a comparison, the CuO NFs/Cu NPs nanocomposites without the involvement of CuCl2 treatment show a similar morphology with that of CuxO NFs/Cu NPs nanocomposites ( Figure S1d-f), suggesting that the H2O2 is the key step to generate the nanoflakes structure. In addition, as shown in Figure S2a-c, the intermediate product of Cu2(OH)3Cl/CuCl/Cu NPs nanocomposites still show the solid particle shape.  To further investigate the valence state of the synthesized materials, XPS spectra (Figure 2f) of Cu x O NFs/Cu NPs nanocomposites were carried out. In Cu x O NFs/Cu NPs nanocomposites, the peaks located at 934.6 eV and 953.7 eV were assigned to the Cu 2p 3/2 and Cu 2p 1/2 of CuO, suggesting the presence of CuO [36]. The existence of shake-up satellite peaks at higher binding energy around 941.2 eV also confirmed the CuO state. Meanwhile, there were doublet peaks at 932.84 eV and 955.2 eV which were assigned to the Cu 2p 3/2 and Cu 2p 1/2 of Cu 2 O [34]. Only peaks belonging to Cu (932.3 eV and 952.1 eV) and CuO (933.1 eV and 953.0 eV) were found in the XPS spectra of commercial Cu NPs and the sample without the involvement of CuCl 2 [23]. The XPS results further suggest that the CuO/Cu 2 O layer was formed on the surface of Cu NPs.

Electrochemical Measurements
The electrochemical measurement was performed on flexible sensor chips, as shown in Figure 3a. The CV characteristics of the bare Nafion, Cu NPs, Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites, Cu x O NFs/Cu NPs nanocomposites and CuO NFs/Cu NPs nanocomposites were measured to compare their electrocatalytic performance ( Figure S3a-f). As shown in Figure 3b, the current response of curves of Cu x O NFs/Cu NPs nanocomposites is much higher than that of bare Nafion, Cu NPs, Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites and CuO NFs/Cu NPs nanocomposites with the presence of 1.0 mM glucose. The CuCl 2 transfer more Cu to generate Cu 2 O/CuO NFs and thus the response was much higher than that of the CuO NFs/Cu NPs nanocomposites. The response of the Cu x O NFs/Cu NPs nanocomposites-based sensor gradually raised towards the oxidation of glucose when the concentration was increased to 2.5 mM and the catalytic current peak ranges from 0.40 V to 0.60 V (Figure 3c). The CV property of Cu x O NFs/Cu NPs nanocomposites with different scan rates (v) was tested from 5 mV s −1 to 50 mV s −1 with the presence of 1.0 mM glucose concentration. The CVs of Cu NPs, CuO NFs/Cu NPs nanocomposites, Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites, Cu x O NFs/Cu NPs nanocomposites with varied scan rates were shown in Figure S4a-d for studying the electrochemical active surface area (ECSA). The extracted calibration curves of current response versus the scan rates with different samples for ECSA was shown in Figure S4e. The ECSA of Cu NPs, Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites, CuO NFs/Cu NPs nanocomposites and Cu x O NFs/Cu NPs nanocomposites were 118.51 µF·cm −2 , 167.07 µF·cm −2 , 50.8 µF·cm −2 and 140.05 µF·cm −2 , respectively. In addition, the Electrochemical Impedance Spectroscopy (EIS) of Cu NPs, Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites, CuO NFs/Cu NPs nanocomposites and Cu x O NFs/Cu NPs nanocomposites was measured and fitted in Figure S4f. As shown in Figure S4f, the transfer resistance value of Cu NPs, Cu 2 (OH) 3 Cl/CuCl/Cu NPs nanocomposites, CuO NFs/Cu NPs nanocomposites and Cu x O NFs/Cu NPs nanocomposites were 506.2 ohm, 339.6 ohm, 787.3 ohm and 198.7 ohm, respectively. This result demonstrated that, after the CuCl 2 aqueous solution treatment and H 2 O 2 aqueous solution, the ECSA was notably increased and transfer resistance value was highly reduced. As shown in Figure S5a,b, the oxidation peak current of the as-prepared sensor rises with the scan rate and the corresponding fitting equation is I = 11.34 × v 0.5 − 35.49. The linear relationship between the response current intensity and the square root of the scan rate suggests that the adsorption of glucose determined the redox reaction [34].
glucose. The detection limit was calculated by 3σ/s and the σ and s are defined as the background current's standard deviation and the slope of the calibration curve, respectively. The detection limit of the CuxO NFs/Cu NPs nanocomposites-based sensor was 79.1 nM. Table 1 shows the performance comparison among the CuxO NFs/Cu NPs nanocomposites-based sensor and several reported non-enzyme glucose sensors. The asprepared CuxO NFs/Cu NPs nanocomposites-based sensor exhibits better overall performances than other sensors for more abundant active sites from nanoflake structure.  The amperometric curves of the Cu x O NFs/Cu NPs nanocomposites-based sensor at different potentials upon increased glucose concentration are shown in Figure S6a-e. Figure 3d shows the extracted curves of a current response versus the glucose concentration at different potentials. It is found that the current responses at +0.40 V, +0.45 V, +0.5 V, +0.55 V and +0.6 V are lower than that at +0.55 V. So, the following amperometric curves were tested at +0.55 V. To calculate the sensitivity of the sensor, the amperometric responses of Cu x O NFs/Cu NPs nanocomposites-based sensor at +0.55 V with the increased glucose concentration were measured and the calibration curve between glucose concentration and the current was shown in Figure 3e,f, respectively. The sensitivity of the as-prepared sensor was 779 µA mM −1 cm −2 based on the linear fitting equation (I(µA) =9.93 + 0.055 × C (µM)), which was 4.0-fold that of CuO NFs/Cu NPs nanocomposites (196 µA mM −1 cm −2 ) ( Figure S7). The relationship between current and glucose concentration follows the Michaelis-Menten equation. The response current intensity j could be expressed by j = K cat × [S]/(Km + [S]) [34], in which the K cat , [S], K m are substrate concentration, Michaelis constant, apparent catalytic turnover rate, respectively. In an extreme case, if the glucose concentration was at a low range and the K m was larger than [S], the response current intensity j exhibits linear relation with the concentration of glucose. The detection limit was calculated by 3σ/s and the σ and s are defined as the background current's standard deviation and the slope of the calibration curve, respectively. The detection limit of the Cu x O NFs/Cu NPs nanocomposites-based sensor was 79.1 nM. Table 1 shows the performance comparison among the Cu x O NFs/Cu NPs nanocomposites-based sensor and several reported non-enzyme glucose sensors. The as-prepared Cu x O NFs/Cu NPs nanocomposites-based sensor exhibits better overall performances than other sensors for more abundant active sites from nanoflake structure.  Figure 4a shows the amperometric response stability of Cu x O NFs/Cu NPs nanocomposites sensor at 1.0 mM glucose during 1000 s testing and no significant attenuation was detected. The amperometric responses of the sensor were measured 6 times and no remarkable deviation among each measurement, suggesting the great repeatability of our Cu x O NFs/Cu NPs nanocomposites-based sensor (Figure 4b). The anti-interference performance was measured by injecting 1.0 mM glucose first and then subsequently adding 10 mM urea, ascorbic acid (AA) and uric acid (UA). As shown in Figure 4c, signal responses of the sensor toward interfering chemicals were quite weak. In comparison, the sensor generated a strong current response when glucose was reinjected. The intensity of the second responding current was close to that of the first one. Figure 4d shows the normalized current response of different interfering species and the current response to 1.0 mM glucose was set as 100%. Cu x O NFs/Cu NPs nanocomposites-based sensors exhibited excellent anti-interference performances. These results suggest that the Cu x O NFs/Cu NPs nanocomposites-based sensor was reliable and suitable for the following noninvasive wearable glucose sensing.   Figure 4a shows the amperometric response stability of CuxO NFs/Cu NPs nanocomposites sensor at 1.0 mM glucose during 1000 s testing and no significant attenuation was detected. The amperometric responses of the sensor were measured 6 times and no remarkable deviation among each measurement, suggesting the great repeatability of our CuxO NFs/Cu NPs nanocomposites-based sensor (Figure 4b). The antiinterference performance was measured by injecting 1.0 mM glucose first and then subsequently adding 10 mM urea, ascorbic acid (AA) and uric acid (UA). As shown in Figure 4c, signal responses of the sensor toward interfering chemicals were quite weak. In comparison, the sensor generated a strong current response when glucose was reinjected. The intensity of the second responding current was close to that of the first one. Figure 4d shows the normalized current response of different interfering species and the current response to 1.0 mM glucose was set as 100%. CuxO NFs/Cu NPs nanocomposites-based sensors exhibited excellent anti-interference performances. These results suggest that the CuxO NFs/Cu NPs nanocomposites-based sensor was reliable and suitable for the following noninvasive wearable glucose sensing.

Real Sample Analysis
The Cu x O NFs/Cu NPs nanocomposites-based sensor not only shows superior electrochemical performance but also delivers good flexibility, making it suitable for wearable applications. To evaluate the flexibility of our sensor, we test the sensing performance after repeated bending and twisting conditions. In our experiment, the sensor was twisted and bent to 90 • , and then unfolded to restore to the original geometry, as shown in Figure 5a,b. After the deformation was completed, the sensor returned to its original state. The durability of mechanical deformation was assessed by checking the response to a glucose solution. After 100, and 500 cycles, respective amperometric tests were performed. Figure 5c,d show the amperometric responses of the sensor toward 0.5 mM glucose before and after cycling twisting and bending, respectively. The amperometric responses demonstrate that the sensor exhibits robust performance against repeated mechanical deformation, with insignificant departure from the original condition. The flexibility ensures good contact between the sensor and human skin and durability to mechanical deformation guarantees reliable sensing.

Real Sample Analysis
The CuxO NFs/Cu NPs nanocomposites-based sensor not only shows superior electrochemical performance but also delivers good flexibility, making it suitable for wearable applications. To evaluate the flexibility of our sensor, we test the sensing performance after repeated bending and twisting conditions. In our experiment, the sensor was twisted and bent to 90°, and then unfolded to restore to the original geometry, as shown in Figure 5a,b. After the deformation was completed, the sensor returned to its original state. The durability of mechanical deformation was assessed by checking the response to a glucose solution. After 100, and 500 cycles, respective amperometric tests were performed. Figure 5c,d show the amperometric responses of the sensor toward 0.5 mM glucose before and after cycling twisting and bending, respectively. The amperometric responses demonstrate that the sensor exhibits robust performance against repeated mechanical deformation, with insignificant departure from the original condition. The flexibility ensures good contact between the sensor and human skin and durability to mechanical deformation guarantees reliable sensing. To further demonstrate the effectiveness of the Cu x O NFs/Cu NPs nanocompositesbased glucose sensor for wearable monitoring, the sensor was mounted on the arm of one volunteer to detect glucose concentrations in human sweat during the exercise (Figure 5e). In this study, a healthy subject (32 years old) who has no medical history of diabetes participated in the experiment. Before the mounting of the sensor, disinfection and cleansing were performed on the arm skin. The PU film was used to protect the sensor and keep the measurement condition with a steady wire connection. Then the volunteer was asked to do 30 min rope skipping to secrete sweat. The sweat sample was measured in situ without any treatment. As shown in Figure 5f, the sweat glucose concentrations of the volunteer changed before (−30 min) and after (+30 min, +3 h and +5 h) remarkably. The measured current response before a meal was lower than after a meal, suggesting that the glucose level in the subject's sweat increased. Although the detection current has slight fluctuation, the obtained results show that the sensor signal increased when the glucose concentration of sweat was changed, promising the application of wearable Cu x O NFs/Cu NPs nanocomposites-based sensors in detecting the glucose concentration in human sweat sensors.

Conclusions
In conclusion, the Cu x O NFs/Cu NPs nanocomposites with a high surface area were successfully fabricated through the involvement of CuCl 2 . Its application as non-enzyme noninvasive wearable glucose sensor was evaluated. The as-prepared sensor electrode exhibits high sensitivity (779 µA mM −1 cm −2 ), wide linear range (higher than 2.5 mM), low detection limit (79.1 nM), durability to twisting and bending deformation and good selectivity. The sensitivity of Cu x O NFs/Cu NPs-based sensor was 3 times higher than that of the sensor based on CuO NFs/Cu NPs nanocomposites synthesized without CuCl 2 . The detection of glucose concentration in human sweat samples was successfully measured, indicating its potential application as a noninvasive wearable glucose sensor.  Institutional Review Board Statement: Ethical review and approval were waived for this study due to the use of non-invasive measures. The participants provided their written informed consent to participate in this study.