Photoelectrochemical device based on Mo-doped BiVO4 enables smart analysis of the global antioxidant capacity in food

An ultrasensitive Mo-doped BiVO4 composite was used to engineer a photoelectrochemical platform for the direct analysis of the global antioxidant capacity. Using this principle, an integrated device was successfully exploited for the “smart” monitoring of antioxidant-rich foodstuffs.


Physicochemical characterization
The structure of Mo-doped BiVO 4 was investigated by a D/MAX 2500V/PC X-ray diffraction (Cu Kα radiation, λ = 0.15405 nm), operated at 40 kV and 30 mA. X-ray photoelectron spectroscopy (XPS) was carried out on an ESCALAB-MKII250 photoelectron spectrometer with Al Kα X-ray radiation as the X-ray source for excitation. High-resolution transmission electron microscope (HRTEM), selected area electron diffraction (SAED), element mapping and energy dispersive X-ray (EDX) spectrum were recorded with a TECNAI G2 microscope operating at 200 kV. Scanning electron microscope (SEM) images were taken on a FE-SEM XL30 ESEM-FEG at the accelerating voltage of 10.0 kV. Raman spectra were collected on a Renishaw Raman system model 2000 spectrometer using a 514 nm argon ion laser and calibrating referenced to the 520 cm -1 line of silicon. The UV-visible diffuse reflectance spectra (DSR, using BaSO 4 as the reference) were obtained from a Hitachi U-3900 spectrometer. Fluorescence emission spectra were performed on a Hitachi F-4600 fluorescence spectrophotometer. Except in an integrated device, all the electrochemical tests were measured at room temperature using a conventional three-electrode cell, comprising modified ITO or a glass carbon electrode (GCE, d = 3 mm) as working electrode, a platinum wire as the auxiliary electrode and a Ag/AgCl (3 mol L -1 KCl) as the reference electrode on a CHI920C electrochemical workstation. The PBS buffer as supporting electrolyte was bubbled with N 2 for at least 15 min and kept over a N 2 atmosphere during the experimental process. LED light (470 nm, Beijing Perfectlight Technology) was used as excitation source of the photoelectrochemical sensor. Electrochemical impedance spectroscopy (EIS) measurements were obtained using a Solartron 1255 B frequency response analyzer (Solartron Inc., UK) in a mixed solution of 1mmol L -1 [Fe(CN) 6 ] 3-/4and 0.1 mol L -1 KCl aqueous solution (amplitude 10 mV, 10 -1 to 10 5 Hz). A Mott-Schottky plot was performed in 1 mol L -1 Na 2 SO 4 with frequencies of 1000, 2000 and 5000 Hz.

Antioxidant Capacity (AC) Detection.
Fruits, and drinks were bought from the local supermarket. As for fruits, juice was collected from their flesh and then the supernatant liquid was further treated with centrifuge. As to teas, firstly, dry leaves (0.25 g) of each tea were suspended in 30 mL of boiling water. Then the tea was kept infusion for 30 min at room temperature. Finally the tea extracts were obtained via a normal funnel. All the commercial drinks were used as received without further purification. The resultant samples were stored at a 4 ° C refrigerator.
The ITO electrode was cleaned with NaOH (1 mol L -1 ) and H 2 O 2 (30%), followed by sonicating in acetone and water, and then dried under ambient conditions. After that, 100 μL of bare BiVO 4 or Mo-doped BiVO 4 suspension (1 mg mL -1 ) was cast onto the pretreated electrode within defined area. Through natural drying, the ITO electrode was further dried at 100 °C for 12 h to improve adhesion.
As illustrated in Figure 1a, the AC assay process on PEC sensor is clearly observed. In brief, once the buffer or sample was injected into the PEC cell, with light irritating from the back side of modified ITO electrode, the system instantly acquired electrical signal. Note that the final photocurrent was recorded and calculated as average value of three measurements at the same concentration. The PEC current was defined as following: I = I sample -I blank (where I sample is the photocurrent generated in the presence of sample, I blank is the photocurrent generated in the absence of sample). Similarly, experimental procedures in our integrated device are nearly commensurate with the PEC sensor, which only has some little details different from it. Or to be more precise, a two-electrode cell is employed in this designed integrated device instead of the traditional three-electrode system. And the static cell used in PEC sensor is upgraded by the flow cell, where the injection rate is controlled with peristaltic pump at 10 r min -1 .
As control experiments, classical spectroscopic methods like DPPH and F-C approaches were also introduced to evaluate global AC in food. DPPH radical scavenging activity was determined as reported by Band-Williams et al. with minor modifications. 1 In short, 100 μL of Trolox standard solution or practical sample diluted solution was added into 900 μL of DPPH reagent (0.04 mg mL -1 , in absolute methanol). After reacting fifteen minutes, the change in absorbance at 514nm was immediately monitored by UV-vis spectroscopy. For F-C method, as described by Ainsworth, 2 briefly to the volume of 100 μL of GA solution or practical sample diluted solution, 200 μL (1/10 dilution) of F-C reagent was added. Then 800 μL of Na 2 CO 3 (0.7 mol L -1 ) was quickly mixed with the above solution. The mixture was incubated at 25 °C for 2h before reading the absorbance at 765 nm. All the results were replicated for three times. It is worth mentioning that DPPH assay is expressed as Trolox equivalents, while both F-C and PEC assays take GA as calibration standard.

Supplementary Figures and Table
A tiny distinction in width of band gap will entail a profound influence on photocatalytic activity. According to Fig. S1     Current intensity or charge measurement is susceptible to change by the applied potential. Fig.   S4 depicts that the photocurrent intensity gradually rose from -0.3 V to 0.3 V, followed by steady state from 0.3V to 0.6V. The positive potential is conducive to electron transfer through CB of semiconductor towards ITO electrode. The higher the potential applied, the greater the photocurrent expressed until the effect of electron concentration gradient is negligible. Since the current at 0 V accounted for 64% of 0.3 V, it is considerable for sensitivity performance.
Meanwhile, the open circuit potential is also convenient for development of the integrated photoelectrochemical device. Thus at the comprehensive respect, 0 V was a suitable choice for PEC transducer.  Once hydroxyl radicals generate in the experiment, the special fluorescence signal will appear at 425 nm. 3 In reality, taken as a whole, any change was hardly discovered in 425 nm (Fig. S6). It could be inferred that antioxidants might react directly with the trapped holes rather than hydroxyl radicals. The flat band potential (E fb ) of semiconductor is closely related to the bottom of the conduction band (E cb ). As for many n-type semiconductors, E fb is roughly at 0.1 V below the E cb . 4 Hence, the position of conduction band is approximately equal to the flat band. Mott-Scottky analysis is a powerful tool to measure E fb . Just as Fig. S7, the extrapolation Mott-Scottky plot intersect at -0.57 V (vs Ag/AgCl, namely, -0.363 V vs NHE), which sheds a brilliant light on E cb value. To some degree, the redox potential is on behalf of reducing ability of an antioxidant. When no corresponding return cathodic peak is produced (as for some antioxidant molecules), the formal potential can be estimated by the potential halfway between E p/2 and E pa of the first peak. 5 (E pa means the potential of anodic peak; E p/2 means the half potential of anodic peak from the cathonic direction). The final results are listed in Fig. S8h.   S9. The inner structure of the integrated device.
As shown in Fig. S10, it is observed that the amount of electrolytes hardly had any impact on the photochemical examinations towards the antioxidant (e.g. upon a 74.44 μmol L -1 GA solution).