Fe³⁺ and Ce³⁺ modified nano-TiO₂ for degradation of exhaust gas in tunnels

2.1 Experimental reagents and experimental apparatus

The nano-TiO₂ and modified nano-TiO₂ were prepared by the sol–gel method [14,15,16,17,28]. The primary reagents were butyl titanate (analytical purity), anhydrous ethanol (superior purity), glacial acetic acid (analytical purity), deionized water (analytical purity), cerium chloride (analytical purity), ferric chloride (analytical purity), etc. The central apparatus are constant temperature magnetic stirrer (HS-19), vacuum drying oven (DZF-6050), muffle furnace (SX2-2.5-10), etc. The costs of various reagents are shown in Table 1.

2.2 Experimental steps

The specific steps for the preparation of nano-TiO₂ and modified nano-TiO₂ by the sol–gel method are as follows:

1. 25 mL of anhydrous ethanol and 5 mL of glacial acetic acid was poured into a beaker at room temperature. Then, 17 mL of butyl titanate was slowly added with uniform stirring using a magnetic stirrer and stirred for 30 min to obtain a uniform and transparent light yellow solution A.

2. 20 mL of anhydrous ethanol, 5 mL of glacial acetic acid, and 10 mL of deionized water was thoroughly mixed at room temperature and a few drops of hydrochloric acid was added to control the pH of the solution between 2 and 3 to form solution B. Then, place it in a separatory funnel and set aside. Add 0.1%, 0.5%, and 1% of Fe³⁺, Ce³⁺ relative to the molar fraction of Ti atoms to make modified nano-TiO₂.

3. The solution B was slowly added (1–2 drops per second) to the solution A to obtain the sol under the action of a thermostatic magnetic stirrer at a controlled temperature of 30°C. Attention was paid so that the speed of the dropping was not be too fast to prevent the butyl titanate from hydrolyzing too fast, and the gel was obtained by aging for a while.

4. The gel was dried in a vacuum drying oven at 80°C for 24 h to get a dry gel.

5. The dry gel was crushed in a mortar and then placed in a muffle furnace and calcined at 500°C for 2 h to obtain nano-TiO₂. Nano-TiO₂ sols and nano-TiO₂ particles are shown in Figure 1.

Figure 1

(a) The sol of TiO₂ and (b) the particles of TiO₂.

The yield of nano-TiO₂ in the theoretical case is 4 g. The orthogonal experiments of nano-TiO₂ obtained under different conditions are given in Table 2.

The dissolution speed of butyl titanate is too fast and white flocculent is easily formed, while the dissolution speed is too slow and it is not easy to form a sol. Generally, the sol contains many tiny bubbles after it is made, and a series of chemical reactions will occur subsequently. Hence, it needs to be placed for a while to make the butyl titanate fully hydrolyzed. Calcination temperature and calcination time mainly affect the crystalline structure. When the temperature is 400°C, the crystalline form is mostly anatase phase, but the crystallization is not complete; when the temperature is 500°C, the crystalline form is completely anatase phase; when temperature increases, the material will be transformed from anatase phase to rutile phase, and the photocatalytic activity is reduced. With close to theoretical yield as the standard, the dissolution time of butyl titanate should be selected for 30 min, gel placed for 10 days, and calcination for 2 h at 500°C.

2.3 Photocatalytic principle

As a semiconductor material, the electronic energy level of nano-TiO₂ is discontinuous, in which there are empty energy level regions, so there are both full and empty bands. The bottom full band is called the valence band, the top empty band is called the conduction band, and the region from the top of the valence band to the bottom of the conduction band is called the forbidden band or bandgap. When the photoelectrons’ energy is greater than or equal to the width of the forbidden band, the electrons in the valence band will be excited by light and jump to the conduction band, while the electrons in the valence band will generate holes due to the jump, forming electron-hole pairs. The electron-hole complex pair will be compounded under the action of an electric field and thus lose its activity. The electron-hole complex pair can react with H₂O and O₂, which are on the surface of nano-TiO₂.

It can be seen from the above reactions that the electron-hole complex pair eventually generates superoxide ion radicals and hydroxyl radicals, which can oxidize CO and NO in the air into CO₂, various carbonates, nitrates, etc., due to the strongly oxidizing properties of these radicals, so as to achieve the purpose of degrading the car exhaust to cure the environment, and the specific catalytic principle is shown in Figure 2 [18].

Figure 2

Schematic diagram of photocatalysis.

2.4 Test systems

The P750 gas catalytic reaction device, which was developed by the Cusson Company in the United States, calculates the catalytic conversion efficiency by testing the change in the gas concentration (mainly refers to the harmful gas of automobile exhaust gas) flowing through the reactor. However, the equipment could not perform to its full efficiency when it is applied to test the catalytic decomposition performance of asphalt mixture on automobile exhaust gas, and the reason is that the equipment cannot provide the light conditions needed for the photocatalytic reaction [20]. The auto-exhaust reaction test equipment was developed independently by Tohoku Forestry University [21], which contains an automobile exhaust gas supply system, computer-aided test system, gas chamber, and other multi-functional components and structures, with a variety of sensors (such as oxygen sensor, carbon dioxide sensor, hydrocarbon sensor, carbon monoxide sensor, pressure box, temperature sensor, etc.). However, due to the instability of the sensor, the device has the disadvantages of unstable reading and short service life. In this study, the actual situation in the tunnel is combined with the disadvantages of poor lighting conditions and poor ventilation in the tunnel, so a closed chamber is used to simulate the tunnel, and the light is a high-pressure sodium lamp, which is commonly used in such low-light conditions in the tunnel due to its high luminous efficiency, low power consumption, long life, and high penetration [22]. The test device consists explicitly of three parts: a standard exhaust gas tank, a closed reaction chamber, and an exhaust gas analyzer, as shown in Figure 3.

Figure 3

(a) Degradation system design drawing and (b) exhaust gas degradation test system; (1) standard exhaust gas tank, (2) air inlet pipe, (3) small fan, (4) high pressure sodium lamp, (5) UV lamp, (6) sealed door, (7) exhaust gas cleaning device, and (8) exhaust gas analyzer.

2.5 Photocatalytic material specimen preparation

The coating type in this study uses a water coating solution. In the process of preparing the coating solution, it is easy to precipitate in the process of preparation as the specific gravity of nano-TiO₂ is 4.26, and nano-TiO₂ is easy to disperse unevenly in water and produce agglomeration. Therefore, during the preparation, a certain amount of dispersing stabilizer needs to be added, and the addition amount is 10–20% of the mass of nano-TiO₂. In the coating solution, water:nano-TiO₂:dispersion stabilizer = 1%:2%:0.4%. The specific laboratory preparation steps are as follows: (1) Weigh a certain amount of water, dispersant, and stabilizer. Add the dispersant and stabilizer to the aqueous solution, and shear at 500 rpm for 5 min with a high-speed shearer. (2) Add nano-TiO₂ to the above solution, and the water-based nano-TiO₂ coating solution can be produced by shearing at a high speed of 2,000 rpm for 30 min with a high-speed shearer [23,24].

In this study, relying on the actual needs of the project, the AC-13 asphalt mixture is used, fine aggregate is limestone, coarse aggregate is basalt, oil to stone ratio is 5.0%, 70# matrix asphalt is used, and the gradation design is as shown in Table 3.

The nano-TiO₂ coating solution and the specimens covered with photocatalytic materials are shown in Figure 4.

Figure 4

(a) The coating solution of nano-TiO₂ and (b) specimen covered with photocatalytic material.

2.6 Evaluation indicators

Among the evaluation indices currently used to evaluate the performance of photocatalytic decomposition of vehicle exhaust, the cumulative decomposition rate (i.e., decomposition efficiency) is the most common evaluation index [25,26]. The cumulative decomposition rate is calculated in Eq. 7:

where η – cumulative decomposition rate, %; M ₀ – initial concentration, ppm; and M ₁ – final concentration, ppm.

The cumulative decomposition rate is one of the essential evaluation indicators for the reaction photocatalytic decomposition of automobile exhaust gas, as well as a quantitative indicator, but there are shortcomings that the indicator requires a fixed initial gas concentration of the test as well as a fixed closed reaction chamber.

In the process of photocatalytic degradation of automobile exhaust, the rate of gas concentration reduction is not always constant, which means that the gas concentration is not linearly related to time, so the decomposition rate is different at different periods. The average decomposition rate is an important index to react to different decomposition rates at different periods. The average decomposition rate is calculated in Eq. 8 [27]:

where v – average decomposition rate, ppm·min⁻¹; and t – reaction time, min.

In this study, the average decomposition rate is differentiated based on the average decomposition rate, i.e., it is more time-sensitive. The concentration-time relationship graph is made with time as the horizontal coordinate and gas concentration as the vertical coordinate, which was then fitted by Origin to obtain the exhaust gas degradation curve, as shown in Figure 12. The slope of this curve indicates the decomposition rate.

3.1 XRD

XRD uses the position and intensity of XRD diffraction angles to analyze the composition of minerals. It explores the phases of the substance based on the spacing of each crystal plane and the relative intensity of diffraction lines against PDF cards. By comparing the diffraction peaks of unknown substances with those of known substances, the various phases in the sample are derived. It can be derived from Figure 5 that the diffraction angles 2θ = 25.3, 37.9, 48.2, 54.1, 55.7, and 63.4, so the nano-TiO₂ is anatase crystalline type, and these diffraction peaks belong to (101), (004), (200), (105), (211), and (204), respectively. The determination of XRD crystal size is related to the peak width of the diffraction peak and the material grain size, which is calculated in Eq. 9:

where D – grain size, nm; K – Scherrer constant, generally taken as 0.89; λ – wavelength of the rays, nm; B – half-peak width of the diffraction peak, rad; and θ – Bragg diffraction angle, °.

Figure 5

X-ray diffraction analysis.

This study used a DB-type X-ray diffractometer with the following technical parameters: Cu target K-α radiation (λ = 1.5406 nm), operating voltage of 40 kV, operating current of 150 mA, scan range of 5–90°, and scan rate of 10°·min⁻¹. The calculation results of the relevant lattice parameters for different ion doping are displayed in Table 4.

As can be seen from Figure 5, the doping of Fe³⁺ and Ce³⁺ did not cause fundamental changes in the positions of the diffraction peaks of nano-TiO₂, which indicates that the modified nano-TiO₂ is still dominated by anatase. With the doping of ions, the diffraction peaks show specific regular changes. From Table 4, it can be concluded that the average grain size of Ce³⁺, Fe³⁺ co-doping is significantly smaller than that of pure TiO₂. The ionic radius of Fe³⁺ is 0.064 nm and that of Ti⁴⁺ is 0.068 nm, and the difference between the ionic radii is not large, indicating that part of Fe³⁺ will replace Ti⁴⁺ and exist in the crystal lattice during the doping modification process. The entry of Fe³⁺ into the nano-TiO₂ lattice will destroy the periodicity of the crystal plasmon arrangement, thus inducing crystal distortion and inhibiting the growth of TiO₂ crystals. The ionic radius of Ce³⁺ is 0.102 nm, which is much larger than that of Ti⁴⁺, so the possibility of Ce³⁺ entering the TiO₂ lattice to replace Ti⁴⁺ is negligible, but it may exist at the TiO₂ grain boundary, i.e., the interface between two grains or microcrystals. On the one hand, it induces crystal distortion and form lattice defects. On the other hand, it will increase the diffusion energy barrier of atoms between grains and prevent direct contact between grains. Both of them will inhibit lattice growth, form lattice defects to trap photogenerated electrons, and reduce the probability of electron-hole complexation, thus improving photocatalytic activity.

3.2 UV-visible (UV-Vis) spectroscopy analysis

UV-Vis diffuse reflectance spectroscopy is the UV-Vis spectrum generated by using the degree of absorption of UV or visible light by the sample. This study used a UV 3600 Plus UV-Vis spectrophotometer with a scanning range of 200–900 nm.

The wavelengths of various kinds of light are composed of ultraviolet wavelength (below 400 nm), visible wavelength (400–760 nm), and infrared wavelength (above 760 nm). It can be seen from Figure 6 that the absorbance of nano-TiO₂ decreases with the increase in the wavelength of light, which decreases dramatically between the wavelengths of 350–450 nm. In this interval, the rate of decline first increases and then decreases, that is, there is an inflection point, making a tangent line at the point. The intersection of the tangent line and the horizontal coordinate can roughly represent the maximum value of the absorbance wavelength. The tangent line drawn for each spectrum in Figure 6 concludes that the doping of ions will increase the wavelength of TiO₂ response to visible light so that its absorption range is expanded from ultraviolet light to visible light. The co-doping effect is better than Ce³⁺ single doping and Fe³⁺ single doping because the doping of ions inhibits the growth of TiO₂ nanocrystals and generates lattice defects, so that electron leap can occur even at low energy. The redshift of the sample’s absorbed optical band is beneficial to excite the photocatalytic activity and form an electron-hole, thus improving the utilization of visible light.

Figure 6

UV-Vis light diffuse reflection.

To further investigate the effect of different ion doping on the forbidden bandwidth of nano-TiO₂, the optical band gap energy value of TiO₂ can be calculated after the Kubelka–Munk transformation of the UV-Vis diffuse absorption spectrum of TiO₂. The specific expression of the change is given by Eq. 10:

where R _∞  – reflectance; s – scattering coefficient; and k – molar absorption coefficient.

The relationship between the forbidden bandwidth of semiconductor materials and the wavelength of light absorption is given by Eq. 11:

where E _g – bandwidth and λ _g – light absorption threshold.

After the K–M transformation of the UV-Vis diffuse reflectance absorption spectrum, the tangent line intersects with the horizontal axis and then the K–M transformation curve graph was obtained. It can be seen from Figure 7 that the absorption sidebands of the samples gradually move from the shortwave direction (390 nm) to the longwave direction (520 nm) with the doping of different ions, and the forbidden bandwidth decreases from 3.22 to 2.71 eV. This is related to the lattice distortion produced by Fe³⁺ and Ce³⁺, which causes defect traps on the TiO₂ surface and achieves the purpose of trapping photogenerated electron pairs. And according to the semiconductor energy band theory, the low-valent Fe³⁺ replaces Ti⁴⁺ in the lattice and forms a new d-energy band in the energy band gap of TiO₂. The d-energy band creates separated impurity energy levels in the forbidden band, leading to a decrease in the position of the bottom of the TiO₂ conduction band. The position of the TiO₂ conduction band bottom decreases so that the less energetic photons are excited to the impurity energy levels to trap electrons. This is consistent with the findings that ion doping can cause the UV-Vis absorption edge of TiO₂ to be redshifted.

Figure 7

K–M conversion curve graph.

3.3 PL analysis

Photoluminescence (PL) reflects the separation of electrons and holes. In general, the lower the intensity, the more efficient the separation of electrons and holes and the lower the probability of compounding. In this experiment, an FL4500 type fluorescence photometer was used for characterization and analysis. The PL spectra in the range of 375–525 nm were measured at an excitation wavelength of 365 nm, and as shown in Figure 8, the intensities of modified nano-TiO₂ were all lower than those of pure nano-TiO₂, indicating that the doping of modified ions induced crystal distortion and formed lattice defects, which produced traps for trapping photogenerated electrons and reduced the probability of electron-hole complexation. Among them, Fe³⁺ and Ce³⁺ co-doping has a lower electron-hole complex probability, indicating that the co-doping of two metal ions is more favorable to improve the catalytic efficiency of nano-TiO₂.

Figure 8

Fluorescence spectroscopy analysis.

3.4 BET analysis

In chemical materials research, the biggest advantage of nanoscale photocatalytic materials over other macroscopic photocatalytic materials is that nanomaterials have a huge surface area and surface structure advantages. The specific surface area test referred to as the BET test method, has become the standard for qualitative analysis of adsorption properties of materials and microscopic particles in the field of materials research at home and abroad. Before the BET test analysis of composite nano photocatalytic materials, the sample needs to be subjected to nitrogen adsorption–desorption isothermal test and then the data are fitted and analyzed according to the BET formula, and the specific surface area and pore size of nano-TiO₂ and modified nano-TiO₂ are derived from the distribution data. As shown in Figure 9a, the modified nano-TiO₂ changed the linear type from type II to type IV compared with the pure TiO₂. It can be observed that the inflection point of modified nano-TiO₂ occurs between 0.55 and 0.75 relative pressure. After the inflection point, the rise rate decreases, and both show the absorption–desorption hysteresis phenomenon, indicating that the modified nano-TiO₂ belongs to the mesoporous material structure. This hysteresis phenomenon is caused by the presence of a large number of capillary pores on the surface, indicating that the doping of ions makes the catalyst surface more rough and porous with increased capillary structures. Figure 9b also reflects that the average pore size of modified nano-TiO₂ is smaller than that of pure TiO₂ and the specific surface area is larger than that of pure TiO₂, which further confirms the XRD analysis.

Figure 9

(a) N₂ adsorption/desorption isotherms and (b) pore size distribution curves.

3.5 Electron spin resonance (ESR) analysis

In this thesis, ESR was used to monitor the formation of ^•OH. As shown in Figure 10, the area ratio of the four characteristic peaks was 1:2:2:1 under the high-pressure sodium lamp irradiation for 30 min, which belonged to the typical characteristic signal peaks of ^•OH, and all samples had obvious ^•OH signal peaks. The ESR signals of modified nano-TiO₂ are stronger than those of pure TiO₂, which should be related to the number of separated electron-hole pairs. In general, more reactive radicals are beneficial to photocatalytic performance, which provides the possibility to subsequently enhance the tail gas degradation efficiency.

Figure 10

Results of electron spin resonance analysis.

3.6 Exhaust degradation test

The source of exhaust gas for the test was a standard exhaust gas tank, and the composition ratio of the gas mixture was NO (40,560 ppm), CO (9.96%), CO₂ (5%), and N₂. The volume of the exhaust gas tank is 8 L, the length, width, and height of the sealed chamber are 60 cm, converted to a volume of 216 L. Using a micro-flow meter to control the volume of gas passing each time, the volume of gas mixture passed in the test is 0.5 L, and the initial gas concentration control range is shown in Table 5.

As can be seen from Table 5, the actual concentrations of CO and NO are less than the theoretical values, but the concentration of CO₂ is greater than the theoretical value. The reason for this is that in the process of the mixed gas entering the pipe, the connection between the pipes is not completely dense and part of the gas is left in the pipe, resulting in the actual concentration being less than the theoretical concentration. The increase in CO₂ concentration is due to the fact that the closed chamber is not a vacuum and contains air, which contains CO₂ in its composition, and when this content is greater than the loss value, it causes an increase in CO₂ content.

Before starting the test, the tightness of the reaction chamber should be checked to prevent large test errors. The airtightness was checked by not putting any specimen in the reaction chamber and then introducing 0.5 L of the gas mixture. The main objects of this test are CO and NO, and the concentration changes in CO and NO are tested within 60 min. To avoid the error brought by one group of tests, two groups of each test are done in this study, and the specific test results are shown in Figure 11.

Figure 11

Graph of CO and NO concentration changes over time (blank test).

As can be seen from Figure 11, the concentration of CO slightly decreases with the growth of time, while the concentration of NO significantly reduces with the growth of time. The reason is that part of NO has an oxidation reaction with O₂ in the air, so the concentration of CO slightly decreases and can be concluded that the airtightness of the confinement chamber is good and meets the requirements of the test.

After the airtightness check, debug and preheat the exhaust gas analyzer, then put the photocatalytic specimen into the sealed chamber, close the sealing door, start the exhaust gas analyzer, and open the pressure reducing valve of the exhaust gas tank after the analyzer indicates a stable number. The volume of the gas mixture is controlled by the micro-flow meter, and the rate chosen by the micro-flow meter is 2 L·min⁻¹, i.e., the pressure reducing valve is closed after 15 s. Turn on the small fan and wait for 5 min to make the mixture evenly fill the sealed chamber, then turn on the high-pressure sodium lamp, set the parameters of the exhaust gas analyzer for a total duration of 60 min, because the exhaust gas analyzer is equipped with a data export function, the data export interval is set to 6 min per time. At the end of the test, turn off the fan, open the sealed door, and take out the test specimen after the exhaust gas is exhausted from the test system to prepare for the next test.

In the exhaust gas degradation test, the surface of the AC-13 asphalt mixture was pretreated first. Then, the prepared water-based nano-TiO₂ coating solution was sprayed onto the rutting plate at a dosage of 556 mL·m⁻², and it could be used in the exhaust gas degradation test after the water evaporated. A total of 20 groups were tested. The specific experimental data are shown in Table 6.

Because the airtightness of the closed chamber body is not completely sealed as well as the reason that NO will be naturally oxidized in the air, a blank test group is needed to make the correction comparison. The specific compensation method is: the initial concentration of the blank group is C ₀, the concentration of the blank group is C ₁, and the concentration of the test group is C ₂, then the actual concentration of the test group after degradation is given by Eq. 12:

After compensating the data of each group, the average value of two groups A and B was taken as the final test data, as shown in Figure 12.

Figure 12

Graph of NO and CO concentration changes over time.

It can be noted that the doping of ions will improve the degradation efficiency of the exhaust gas of nano-TiO₂ under high-pressure sodium lamp irradiation, and the degradation effect of co-doping is better than cerium ion single doping, better than iron ion single doping better than pure TiO₂, and the degradation efficiencies are shown in Tables 7 and 8.

The concentration curve of exhaust gas degradation was obtained by fitting with Origin software. The maximum value of the slope of the curve was found to represent the maximum value of the degradation rate of exhaust gas, and the categories from left to right in Table 8 were represented by S₀–S₁₀. The exhaust gas degradation rates of NO and CO are shown in Figure 13. The maximum values of the degradation rates of NO and CO were obtained as shown in Tables 9 and 10.

Figure 13

Exhaust gas degradation rate of NO and CO.

It can be noted from exhaust gas degradation test that the degradation effect of Ce³⁺, Fe³⁺ co-doping is better than Ce³⁺ single doping, Fe³⁺ single doping and pure TiO₂, where the maximum degradation efficiency and maximum degradation rate are Ce³⁺, Fe³⁺ co-doping, and the doping amount is 0.5%. This is because in the process of doping, Fe³⁺ with a small ion radius replaces part of Ti⁴⁺ and thus changes the crystal structure, while Ce³⁺ with a large ion radius exists at the grain boundary position and increases the diffusion energy barrier between the grains, both of which can induce lattice distortion, form lattice defects, reduce the probability of electron-hole complexation, improve photocatalytic activity, and thus improve the exhaust gas degradation efficiency. But too much ion doping will cause the remaining ions to accumulate on the particle surface of the catalyst, resulting in a reduced contact area between nano-TiO₂ and visible light, thus decreasing the catalytic activity.