The structure of the microreactor used in this study is shown in Figure 1; it has a 5-level tree-shaped microchannel and comprises a PDMS plate as the cover and glass slides as the substrates. The microchannel and thin film are fixed in the middle of the microreactor to form a square reaction chamber of volume 7.2 μL (12 mm × 12 mm × 50 μm). The glass substrate is coated with the Y, Yb/TiO₂ film (5 μm thick), which covers the entire reaction chamber. Preparations for the porous TiO₂ film experiment are shown in Supplementary Figure S1. The tree-branch-shaped microchannels are used to uniformly distribute the solution throughout the reaction chamber so as to achieve maximum contact with the film.

P25 anatase TiO₂, methylene blue (MB) reagent, acetylacetone, Triton X-100, and polyethylene glycol-2000 were purchased from Aladdin. Y(NO₃)₃·6H₂O and Yb (NO₃)₃·6H₂O were purchased from Chemical Reagent Co. Except for the MB reagent, all other chemical reagents were analytically pure. MB solutions of different concentrations were prepared in this study by dissolving in deionized water.

The synthesized Y, Yb/TiO₂ nanoparticles were characterized by scanning electron microscopy (SEM, Thermo Fisher Scientific, Helios G4 CX), X-ray diffraction (XRD, Rigaku, MiniFlex 600), X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific, ESCALAB 250), and UV-visible diffuse reflectance spectroscopy (UV-vis DRS, Agilent, Cary 7000). The absorption spectra between the original and degraded MB solutions were analyzed using a UV-vis spectrophotometer (Beifen-Ruili UV-2601).

Some of the main characterizations are achieved through setting the appropriate experimental parameters. The XRD parameters are as follows: Cu target, Kα ray, angle 10°–80°, and scattering slit 1°. The UV-vis instrument parameters are as follows: wavelength 200–800 nm, wavelength accuracy ≤ ±0.08 nm, and photometric accuracy ±0.00025 A.

The photocatalytic microfluidic system consists of five parts: the syringe pump, light source, collection bottle, temperature controller, and planar microreactor, as shown in Figure 2. The fluid volumes and flow rates are controlled precisely using the syringe pump. The planar microreactor connects with the syringe pump and collector via pipes. The light source is placed 10 cm directly above the microreactor to provide uniform light. The temperature controller maintains the reaction temperature. All components except the microreactor are set so as to avoid light exposure.

In the photocatalytic microfluidic system, the flow rate and MB solution concentration are the major factors affecting the photocatalytic reaction efficiency. The solutions are pumped at 33, 42, 55, 83, and 167 μL/min to investigate the flow rate effects in different planar microreactors. The initial concentrations of the MB solution are maintained at 3 × 10⁻⁵, 4 × 10⁻⁵, and 5 × 10⁻⁵ mol/L. The photocatalytic experimental parameters of the planar microreactor are shown in Table 1.

The aqueous MB solution was introduced into the microreactor through the syringe pump that also controls the reaction flow rate, and degradation was carried out under 15 mW/cm² of ultraviolet light (<390 nm) and visible light (390–770 nm). After a single set of reactions, the samples were retrieved from the collection bottle at the end of the system for detection.

To investigate the potential applications of the reactor in industry, the microreactor was designed to be used both in batch and continuous reactor modes for a long period of time. In the batch mode, the reaction stops, and the microreactor resets after each 5-mL sample generation. The reset procedure means that the microreactor is washed with deionized water at a flow rate of 50 μL/min for 5 min and that the light is maintained to remove any residual MB solution in the microfluidic system.

As a In the continuous mode, the reaction is continuous, and reactor is not cleaned at intervals. After every 5 mL of sample generation, the degradation products were retrieved for absorbance detection in each set of reactions. After each sampling, the reaction continued in the continuous reaction mode.

For data analyses, the degradation and reaction rate constants can be calculated from the concentration data of MB through the following relationships (Meng et al., 2013; Lamberti, 2015): x = C 0 − C C 0 × 100 % , (1) k ∝ − ln 1 − x t , (2) where k is the reaction rate constant, t is the effective residence time, C ₀ and C represent the initial concentration and the concentration after degradation, respectively, and x is the photocatalytic degradation rate. The microreactor system shows continuous and intermittent reactions during research progress.

The microreactors for the batch reactions that were cleaned using deionized water and ethanol after each reaction (approximately 15–150 min) were defined as the intermittent planar microreactors (IPRs), and the microreactors for continuous reactions (>12 h) that were not cleaned at intervals were defined as continuous planar microreactors (CPRs).

The apparent morphology of Y, Yb/TiO₂ cross-sectional morphology of the catalyst film, and qualitative analysis of the electron probe attached to the scanning electron microscope by an energy spectrometer are shown in Supplementary Figure S2. The 5-μm catalyst film shows good uniformity as a whole and has a loose structure and micropores in the depth direction.

The phase compositions and valences of the samples were characterized by XRD and XPS, as shown in Figure 3. The diffraction peaks agree well with those of the pure anatase TiO₂ according to JCPDS card no. 21–1272, as shown in Figure 3A. This shows that the synthesized Y, Yb/TiO₂ has a pure anatase phase.

The optical absorption of the catalyst is a critical factor for photoreactivity (Dong et al., 2016). The DRS characterization spectrum of the sample in the UV-vis region is expressed in Figure 3A. After doping with elemental Y and Yb, the absorption band of TiO₂ shifts to the infrared region. The bandgap of TiO₂ doped with Y and Yb becomes narrower, and Y, Yb/TiO₂ shows enhanced photon absorption effect in the spectral range of 390–800 nm. Therefore, an expansion in the absorption wavelength range of the incident photons can promote photocatalytic performance.

The highly symmetric shapes of the titanium XPS lines in Figure 3B suggest that titanium is present only as Ti⁴⁺. In fact, the presence of titanium with any oxidation state lower than +4 should have been evidenced by the presence of shoulders on the binding energy sides of the peaks, which was not observed. Because the radii of the Y and Ti atoms are similar, it is easier for the Y atoms to enter the catalyst lattice. There is one electron missing in the crystal lattice, which generates oxygen vacancies to balance the overall electron composition. The fresh oxygen vacancies are bound to titanium dioxide, so the remaining O atoms move to form additional energy levels to narrow the bandgap transitions of the titanium dioxide electrons conveniently to realize effective broadening of the spectral absorption range.

For an overall comparison of different reactors, Figure 9A plots the reaction rate constant, which is derived from formula 2. The reaction rate constants under UV light are compared, where the slope represents the degradation rate. The analytical data points are the experimental averaged values. A linear fit of the tested data for the contrast experiment shows that the bulk reaction yields k = 0.0109 min⁻¹. The Y, Yb/TiO₂ IPR yields k = 0.586 s⁻¹, which is nearly two times higher than that of other reported microfluidic reactors (Li et al., 2018) (Figure 9B) and more than that of the TiO₂ IPR in this work. Both IPRs show more efficient degradation performances than the bulk reactors, and the reaction rate constants are higher by at least two orders of magnitude. In the CPR, the degradation of the TiO₂ film decreases by more than 60% in the presence of attached reactant residue problems. The reaction rate constant of the Y, Yb/TiO₂ CPR is calculated as 0.290 s⁻¹, which is almost two times that of the TiO₂ CPR. These results show that the Y, Yb/TiO₂ film is more efficient and stable. Compared with the bulk reactor, both the IPR and CPR have better degradation efficiencies, but the CPR’s degradation efficiency is lower than that of the IPR, especially at high flow rates. This is because in the continuous high-flow-rate degradation process, the difference in flow rates between the central and edge areas of the microreactor leads to the formation of dead zones at corners, resulting in reactant residues that affect the IPR’s degradation efficiency.

The degradation rates of the MB solution by Y, Yb/TiO₂ and TiO₂ under UV and visible irradiation with different flow rates in the CPR are shown in Figure 10A. The degradation rates decrease obviously as the flow rates increase. The photocatalytic degradation effect of Y, Yb/TiO₂ on MB solution under visible light is lower than that under UV light but higher than that of TiO₂ under UV irradiation. This difference in degradation under UV and visible light is influenced by the range of spectral absorption of the Y, Yb/TiO₂ nanoparticles. It is confirmed that enhancement of Y, Yb/TiO₂ photodegradation exists in both the continuous and intermittent reactions. Therefore, TiO₂ doping of Y, Yb to expand the light response range is an effective method of achieving enhanced photocatalytic effects in materials research.

The influence of the concentration of the MB solution on the Y, Yb/TiO₂ CPR is analyzed in Figure 10B. The degradation effect is better at low flow rates considering first-order kinetic factors, but the initial concentration has a certain effect on the photodegradation efficiency. With the increase in the initial concentrations of the reactants, the photodegradation efficiency increases, but the overall effect is not large. When the flow rate increases to 167 μL/min, the overall degradation rate decreases, and there are marked differences in photodegradation between different concentrations. The photodegradation degree is 69.37% at C₀ = 3 × 10⁻⁵ mol/L and increases further at C₀ = 5 × 10⁻⁵ mol/L. Through fitting the reaction rate constant at different concentrations, the value of k is found to be similar.

The performances for photocatalytic degradations of the MB solution at 5, 25, and 50°C are shown in Figure 11; it is seen that the reaction rate increases with the increase in time and temperature. When the flow rate is high, the degradation performance of the MB solution at 50°C is more efficient than that at 5 or 25°C because of the increase in the catalyst activity at a suitably high temperature. The effects of the reaction temperature mainly influence the quantum effects, i.e., electron–hole separation and recombination (Chen and Hsu, 2021). Moreover, the diffusion of MB and its reaction products increased with the increase in temperature, and the mass transfer limit was weakened further.

In bulk reactors, H₂O₂ with strong photocatalytic activity is generally used to enhance the photodegradation effect. The UV degradation experiment for 12.8 ppm MB was carried out to explore the effect of addition of H₂O₂ on the Y, Yb/TiO₂ CPR, as shown in Figure 12. However, it is seen that the decrease in degradation rate may be due to the effect of scavengers. The scavenger effect refers to the fact that during photocatalysis, some substances (often referred to as scavengers) may react with the generated reactive oxygen species, thereby slowing or hindering effective reaction between the oxygen radicals and target pollutants (Múčka et al., 2013). As an oxidation scavenger, H₂O₂ neutralizes the activities of hydrogen- and oxygen free radicals by trapping them. This trapping effect can make the hydroxyl radicals lose their ability to react, thus slowing the degradation reaction in the photocatalytic process and reducing the photodegradation performance of MB.