Electrochemical detection of baicalein based on a three-dimensional micromixer

The performances of the three passive micromixers, as shown in figure 1, were numerically simulated using the COMSOL Multiphysics 5.3 a software. The incompressible Navier–Stokes and continuity equations were executed to describe the single-phase and laminar flows. The governing equation of the solver is defined as

$$\begin{equation}\rho ({\boldsymbol{u}} - \nabla ){\boldsymbol{u}} = \nabla \cdot \left[ { - p\boldsymbol{I} + \mu \left( {\nabla {\boldsymbol{u}} + {{\left( {\nabla {\boldsymbol{u}}} \right)}^T}} \right)} \right] + {\boldsymbol{F}},\end{equation} \tag{ 1 }$$

$$\begin{equation}\rho \nabla \cdot {\boldsymbol{u}} = 0,\end{equation} \tag{ 2 }$$

where ρ denotes the fluid density, u is the fluid velocity, p is the pressure, $\mu$is the dynamic viscosity, F is the body force, I denotes the identity matrix, and T represents the viscous stress tensor.

The concentration distribution in the steady-state problem is described by the convection–diffusion equation:

$$\begin{equation}\nabla \cdot \left( { - D\nabla c} \right) + {\boldsymbol{u}} \cdot \nabla c = R.\end{equation} \tag{ 3 }$$

To quantify the concentration distribution in a cross-section of the proposed microchannels, mixing uniformity M is defined as follows [29]:

$$\begin{equation}M = 1 - \sqrt {\frac{1}{N}{{\sum\limits_{n = 1}^N {\left( {\frac{{{c_n} - {c^*}}}{{{c^*}}}} \right)} }^2}} ,\end{equation} \tag{ 4 }$$

where N represents the total number of concentration sampling points obtained in a microchannel cross section, c_n is the mass fraction of the sampling point on node n, c^* denotes the expected mass fraction of the sampling point, and c* = 0.5. M = 1 implies that the mixture is completely mixed and M = 0 implies that it is completely unmixed. Each calculation had more than 300 sampling nodes in the z= 0.2 mm section lines on the y–z cross-sections within a microchannel.

The driving force required for the fluid transport in the passive micromixer was affected by the pressure drop. Pressure drop $\Delta {\text{p}}$ [30] is the pressure difference between the inlet and outlet sections of the microchannel.

The mixing performance index, I_MP, is defined to comprehensively evaluate the mixing property of a micromixer:

$$\begin{equation}{{\text{I}}_{{\text{MP}}}}{\text{ = }}\frac{{\text{M}}}{{\Delta {\text{p}}}}.\end{equation} \tag{ 5 }$$

The microchannels were first modelled using the SolidWorks software, which was later imported into the COMSOL 5.3 a software for simulation. The initial conditions of the simulation were set as follows. Water was used as the fluid material, and the initial concentrations of the two inlets were 0 mol l⁻¹ and 1 mol l⁻¹. The wall was set to incompressible nonslip boundary conditions. Other parameters and boundary conditions were set as shown in table 1. As the substance concentration is nanoscale in this simulation, the transfer of diluted substances module was coupled to model mixing, which is helpful to simulate the concentration distribution of dilute substance. The dynamic viscosity and the molecular diffusivity are the same as H₂O (water)[liquid]'s, shown in table 1. The models were meshed with a tetrahedral grid by the meshing tool in the COMSOL software, and fluid dynamics was used as the grid calibration model. The simulation results are presented in section 4.1.

Baicalein (>98%) and baicalein aluminium capsules were used as the reagents. Britton–Robinson (B–R), Citrate buffer saline (CBS), and Phosphate buffer saline (PBS) were prepared using analytical grade reagents. Other reagents included absolute ethanol, potassium chloride, sulphuric acid, and sodium bicarbonate. The experiment also used deionised water (DI water).

Preparation of baicalein stock solution: Dissolve 20 mg of baicalein in absolute ethanol and dilute it to 50 ml to obtain 1.48 × 10⁻³ mol l⁻¹( 0.4 mg ml⁻¹) baicalein stock solution, mixed well, and refrigerated at 4 °C.

Baicalein extraction from baicalein Al capsules: Dissolve a capsule (net weight 0.2 g) in DI water, and then filter it through a filter paper. Dry the filter residue and stir to dissolve it with absolute ethanol. Filter the solution and collect the filtrate, which should be diluted to 25 ml with absolute ethanol, mixed well, and refrigerated at 4 °C.

The apparatus included a dual-channel syringe pump (Wenhao Co., Ltd. Suzhou), a CHI660E electrochemical workstation (Chenhua Co., Shanghai), a precision electronic analytical balance (Ohaus Co., Ltd. Shanghai), micromixers (School of Mechanical Engineering, Jiangnan University), a screen-printed electrode (the working and counter electrodes were carbon while the reference electrode was made of silver/silver chloride), and a UV 1800 ultraviolet-visible spectrophotometer (Shimadzu Co., Japan).

CV determination of baicalein: The baicalein stock solution was diluted with DI water. The baicalein and buffer solutions were pushed into the micromixer using a dual syringe pump. Further, the influences of the buffer varieties, buffer pH values, scanning speeds, and inlet flow rates on the redox currents of baicalein were studied in the 0–0.8 V potential window to obtain suitable experimental operating parameters. The calibration curve of baicalein was determined under the aforementioned conditions and used to detect the baicalein content in the baicalein Al capsules. After each set of experiments, the electrodes were replaced with new ones to ensure the reliability of the experimental results.

Spectrophotometric (SP) determination of baicalein: Transfer 0 ml, 0.25 ml, 0.5 ml, 0.75 ml, 1.0 ml, 1.25 ml, 1.5 ml, 2.0 ml of baicalein solution with the mass concentration of 0.4 mg ml⁻¹ in 10 ml colorimetric tubes. The baicalein was treated using a sodium nitrite-aluminium nitrate-sodium hydroxide colour system. A UV 1800 ultraviolet-visible spectrophotometer was used to detect samples at 510 nm, and a calibration curve was obtained. The absorbance of the medicine was measured under the same conditions and the baicalein content in the medicine was calculated. The experimental results are presented in section 4.3.

The mixing uniformity in different wave microchannels was simulated under Reynolds numbers of Re = 0.1–40. Figures 2–4 show the simulation results in SCW microchannel, RCW microchannel and 3D-ECW microchannel respectively, in which the (a) shows concentration distribution and the (b) shows velocity streamlines of two species. It can be learned from figure 2(a) that with the increase of Reynolds numbers, the concentration distribution became more and more uniform. The concentration cloud image mainly changed when the fluid flew through the corner structures. In figure 2(b) When Re < 10,the streamlines were basically paralleled to each other without crossing; when Re = 20, the streamlines overlapped on the contact interface; when Re = 40, vortex current could be observed but not obvious. In the RCW microchannel, as shown in figure 3, the simulation results had the same trend as figure 2. However, even when Re = 40, the streamlines in RCW microchannel did not appear vortex phenomenon. In figure 4(a), when Re > 10, the concentration distribution changed obviously in 3D-ECW michrochannel, which shows that the expansion and contraction structure has a significant disturbing effect on microfluids. In figure 4(b), the streamlines also changed obviously. The streamlines overlapped on the contact interface when Re = 10; vortex current could be observed when Re = 20 and streamlines overlapped almost in the entire microchannel. When Re = 40, an extremely obvious vortex phenomenon can be observed. After all, the expansion and contraction structure can disturb microfluids and increase the contact area of the microfluids at a relatively low Reynolds number, leading to an effectively improvement of mixing uniformity.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 5 shows the influences of Re on the mixing uniformity of microfluids. As shown, in different wave microchannels, as Re increases, the mixing uniformity of the two species first decreases sharply, then increases, and finally stabilises. The worst mixing uniformity of species was observed when Re = 3. For Re < 3, the flow velocities of the microfluids were low, resulting in a long residence time in the microchannel. The species mixing was mainly dominated by molecular diffusion. When Re > 3, the uniformity of the species gradually increased with increase in Re. The mixing uniformity of the RCW microchannel was always the lowest when Re > 10; thus, this structure was not considered in later experiments. In the 3D-ECW microchannel with an EC ratio of 7/3, the mixing uniformity was always excellent when Re < 28 and was constant when Re = 20. The micromixing uniformity in the other wave microchannels was not stable until Re = 30. In summary, the 3D-ECW microchannel with an EC ratio of 7/3 could achieve good species mixing uniformity at a relatively low Reynolds number.

Zoom In Zoom Out Reset image size

In a passive microchannel, the flow of microfluids are driven by external energy. The structures of microchannels and velocity of microfluids not only affect the mixing uniformity but also affect the inlet pressure in microchannels. Therefore, suitable microchannel structures and inlet flow rates can save driving energy as well as improve the performance of micromixers. The pressure distribution diagram in the SCW microchannel, ECW microchannel and 3D-ECW microchannel are demonstrated in figure 6. It can be learned that with the increase of Reynolds number, the inlet pressure increase gradually. Compared with the other two microchannels, the corner structure is the sharpest in the SCW microchannel, leading to an increase in pressure drop during fluid flow at the same Reynolds number. While in the RCW microchannel, the corners are relatively smooth, resulting in a lower local pressure drop. In the 3D-ECW microchannel, inlet pressure did not increase due to the continuous and smooth change of extraction and contraction structures.

Zoom In Zoom Out Reset image size

Figure 7 shows the pressure-drop diagram of different structural wave microchannels with Re in the range of 0.1–40. As shown, the pressure drops in the SCW and 3D-ECW microchannels with an EC ratio of 4/3 were relatively higher than in the other microchannels. In the RCW and 3D-ECW microchannels with an EC ratio of 8/3, the pressure drops were relatively low. The higher pressure drops might be due to the sudden change of the sidewall structures in the SCW microchannel and the continuous structural changes of the bottom surfaces and smaller flow space in the 3D-ECW microchannel with an EC ratio of 4/3, which led to an increase in the flow resistance of the microfluids. The lower pressure drop could be associated with the smooth and relatively simple structure of the RCW microchannel and the large flow space in the 3D-ECW microchannel with an EC ratio of 8/3. Therefore, we conclude that compared with the SCW microchannel, the use of rounded corners in the design could decrease the pressure drop and weaken the mixing uniformity simultaneously. With the optimal structure design of the 3D-ECW microchannels, both the mixing uniformity and pressure drop were affected to varying degrees with the variation of the EC ratio. Thus, a new evaluation standard must be derived to obtain an optimal microchannel structure.

Zoom In Zoom Out Reset image size

Figure 8 describes the effect of the EC ratio on the mixing performance index in the 3D-ECW microchannels for the Reynolds numbers of 20, 30, and 40. As shown, the peak mixing performance occurred when the EC ratio was α= 7/3 and Re = 20. This is because in a microchannel, with the increase in the Re numbers (Re > 20), the mixing uniformity increased gradually while the pressure drop in the microchannel increased sharply, resulting in a decrease in the mixing performance.

Zoom In Zoom Out Reset image size

Figure 9 shows the velocity streamline diagrams of the 3D-ECW microchannel with EC ratio of 7/3. Figure 9(a) shows the velocity vector diagram on the y–z sections at corners of the microchannel (S1–S4 in figure 9). As the Reynolds numbers increased, the Dean flow phenomenon on the cross sections became increasingly obvious. The colours in figure 9(a) represent the concentration distribution of species in the microchannel. Figure 9(b) shows that when Re = 0.5, the streamlines of two species did not interfere with each other, and the diffusion mixing of species only occurred on the contact surface. When Re = 10, the streamlines appeared staggered and turbulent. In contrast, for Re = 20, the chaotic convection of the two species was obvious, and this greatly improved the mixing efficiency.

Zoom In Zoom Out Reset image size

The redox curve of baicalein is shown in figure 11. Curve (a) in figure 11 was detected through SPCEs in the blank BR buffer (pH 2.5), and curve (b) shows the redox curve of the 1.184 × 10⁻⁴ mol l⁻¹ baicalein solution mixed with BR buffer (pH 2.5). As shown, the substances in the BR buffer solution did not affect the CV behaviour of baicalein. Because the baicalein oxide on the electrode flowed out of the detection area with the fluid, the reduction peak was not observed, and therefore the oxidation peak was the object of research in the subsequent experiments.

Zoom In Zoom Out Reset image size

The SPCEs activation method: First, scan 0.25 mol l⁻¹ sodium bicarbonate solution for 300 s by using the amperometric i–t method with 2 V scanning potential. Then, scan the 0.1 mol l⁻¹ dilute sulphuric acid circles (40) by using the CV method to remove impurities on the electrode surfaces. This activation method could enhance electrode reversibility, increase electron transfer speed, and promote electrode activity.

In the baicalein solution with a concentration of 1.184 × 10⁻⁴ mol l⁻¹, the influences of buffer varieties, including BR, CBS and PBS buffers in the pH range of 2.5–7 on the oxidation peak currents of baicalein were studied. The optimal result for each buffer solution is shown in figure 12. The highest oxidation peak current of baicalein for all three buffers appeared at pH = 2.5. Among the three buffers, baicalein showed the optimum redox peak current in the BR buffer at pH of 2.5.

Zoom In Zoom Out Reset image size

Figure 13 shows the effect of the pH of the BR buffer on the redox curves of baicalein. As shown in figure 13(a), for the pH values of 2.5, 3, 4, 5, 6, and 7 of the BR buffer, the oxidation peak potentials gradually decreased with increase in the pH value, and the peak current reached its maximum value at the pH of 2.5. As shown in figure 13(b), the oxidation peak potentials of baicalein were linearly related to the pH values; ${{E}_{\text{p}}} = - 0.05607pH + 0.70849$(R²= 0.9923). The slope of 56.07 mV pH⁻¹ is close to the theoretical Nernst equation slope of 59.2 mV pH⁻¹. Therefore, we can conclude that the number of electrons involved in the reaction of the electrode oxidation process was equal to the number of protons.

Zoom In Zoom Out Reset image size

Figure 14 shows the effect of the scanning speed on the redox curves of baicalein. As shown in figure 14(a), with increasing scanning speeds of 50 mV s⁻¹, 80 mV s⁻¹, 100 mV s⁻¹, 130 mV s⁻¹, 150 mV s⁻¹, 170 mV s⁻¹, and 200 mV s⁻¹, the oxidation peak currents increased. These peak current values and scanning speeds are linearly related, as shown in figure 14(b). The linear regression equation is given as ${{i}_{\text{p}}} = 0.02325{v} + 1.39513$(R² = 0.99333). Therefore, we conclude that the electrode reaction process is mainly controlled by adsorption.

Zoom In Zoom Out Reset image size

To investigate the influence of the inlet flow rates on the oxidation peak currents of baicalein, we analysed the rates in the range of 100 μl min⁻¹ and 1500 μl min⁻¹, as shown in figure 15. With an increase in the inlet flow rates, the oxidation peak currents first increased and then decreased. For the inlet flow rate of 960 μl min⁻¹, the peak current reached its maximum value, and the corresponding Reynolds number was Re = 20. In the micromixer, the inlet flow rate showed a significant influence on the mixing uniformity. As the inlet flow rates increased in the range of 100 μl min⁻¹ and 960 μl min⁻¹, the peak currents increased significantly. When the inlet flow rate was higher than 960 μl min⁻¹, the peak current decreased; this may be because the flow rates were too fast, thus reducing the residence time of the reagents. The reagents were not mixed perfectly because of insufficient contact; this might have led to the decrease in peak currents. For the inlet flow rate of 960 μl min⁻¹, the residence time of the reagents in the microchannel was less than 200 ms, and thus a millisecond-level micromixing of reagents was realised; this is beneficial for rapid detection and reduced sample consumption.

Zoom In Zoom Out Reset image size

A series of baicalein solutions with different concentrations were prepared and the baicalein content was determined through the CV method under the conditions of 2.5 pH of the BR buffer, scanning speed of 100 mV s⁻¹, and inlet flow rate of 960 μl min⁻¹. As shown in figure 16(a), the oxidation peak currents increased with the increase in baicalein concentration. Figure 16(b) shows that the oxidation peak current had a good linear relationship with the baicalein concentration in the range of 3.55 × 10⁻⁶ mol l⁻¹ and 5.92 × 10⁻⁵ mol l⁻¹, with the linear equation of ${{i}_{\text{p}}} = 0.02454{c} + 0.25565$(R² = 0.99711) and the detection limit of 1.861 × 10⁻⁸ mol l⁻¹( S/N = 3).

Zoom In Zoom Out Reset image size

Under the appropriate experimental conditions, as mentioned in section 4.3.6, the reproducibility and stability of the detection system were determined. The detection of the baicalein solution with a concentration of 7.104 × 10⁻⁶ mol l⁻¹ was measured five times, and the relative standard deviation of the oxidation currents was measured as 2.86%, indicating that the detection method had good reproducibility. The detection system was used to detect the baicalein solution with a concentration of 7.104 × 10⁻⁶ mol l⁻¹ after the device was stored in a dark place at room temperature for 1 week. The relative standard deviation of the oxidation peak current was 4.15%, indicating that the new detection method has good stability.

To evaluate the application of the micromixing electrochemical detection system, it was used to determine the baicalein solution extracted from baicalein Al capsules. Before the determination, the spectrophotometry method was employed as a control method.

A series of baicalein solutions with different mass concentrations were detected through spectrophotometry. The absorbance of baicalein was linearly related to the mass concentration in the range of 0.01–0.08 mg ml⁻¹. The linear equation is written as ${A}{\text{ = 0}}{\text{.60479}}{{\rho }_{\text{b}}}{{ - 0}}{\text{.00322}}$(R² = 0.99745), as shown in figure 17. The baicalein solution extracted from the baicalein Al capsules was later detected through spectrophotometry. The results showed that the baicalein content in one capsule was 12.86 mg. The results of the sample recovery test are listed in table 2, and the recovery was in the range of 98.9%–103.6%, indicating that the detection results were reliable.

Zoom In Zoom Out Reset image size

The extracted baicalein solution was then diluted 5000 times and tested using the CV method. The diluted sample concentration was measured at 9.489 μmol l⁻¹, which can be converted to 12.82 mg of baicalein in one capsule. This value differs from the aforementioned results by 0.31%, confirming that the micromixing electrochemical method was effective, and the rapid and accurate detection of real baicalein samples was thus realised.