Ultrasound-Assisted Synthesis of Nonmetal-Doped Titanium Dioxide Photocatalysts for Simultaneous H2 Production and Chemical Oxygen Demand Removal from Industrial Wastewater.

A series of nonmetal-doped titanium dioxide (Nm x /TiO2, where x is the weight fraction of nonmetal elements) photocatalysts was prepared via ultrasonic-assisted impregnation for simultaneous hydrogen (H2) production and chemical oxygen demand (COD) removal from industrial wastewater. Three types of Nm elements, carbon (C), silicon (Si), and phosphorus (P), were explored. The P1/TiO2 exhibited a higher photocatalytic activity for H2 production and COD removal than the C1/TiO2 and Si1/TiO2 photocatalysts. Approximately 6.43 mmol/g photocatalyst of H2 was produced, and around 26% COD removal was achieved at a P1/TiO2 loading of 4.0 g/L, a light intensity of 5.93 mW/cm2, and a radiation time of 4 h. This is because the P1/TiO2 photocatalyst exhibited lower point of zero charge values and a more appropriate band position compared with other Nm x /TiO2 photocatalysts to produce H+, which can consequently form H2, and reactive oxygen species (HO· and O2 · -), which serve as oxidizing agents to degrade the organic pollutants. Increasing the content of the P element doped into the TiO2-based material up to 7.0% by weight enhanced the H2 production and COD removal up to 8.34 mmol/g photocatalyst and 50.6%, respectively. This is attributed to the combined effect of the point of zero charge value and the S BET of the prepared photocatalysts. The photocatalytic activity of the P7/TiO2 photocatalyst was still higher than the TiO2-based material after the fourth use.

Introduction

Due to the problems of global warming and poor air quality that are associated with greenhouse gases (GHG) and particulates emitted from fossil fuel combustion, together with the fast depletion of nonrenewable fossil-based energy sources, a number of research studies have been conducted to develop more eco-friendly energy sources.[1,2] Currently, hydrogen (H2) is considered as an environmentally friendly energy carrier for the near future development due to its toxic-free emission and high energy capacity.[3] In addition, it is widely used in various sectors, such as transportation, industries, and residences.[1] For example, it is used as a hydrogenating agent in industrial processes, as a direct feed to fuel cells to produce electricity, and also as a feedstock for internal combustion engines, for turbines, and thermal energy.[4] Almost all H2 is currently produced from carbon-based materials, such as natural gas via steam reforming (∼50%), oil/naphtha reforming (∼30%), coal gasification (∼18%), and other sources (∼2%), which leads to massive emissions of GHG2. Within the next decade, the steam reforming of natural gas and catalytic biomass gasification will still be the principal processes to produce H2. However, in the future, it is believed that the massive transition of H2 production from fossil resources (current stage) will be complemented with H2 produced from biomass (midterm stage) and finally green H2 from renewable resources (e.g., water and sunlight) by the year 2050.[4] Thus, many research activities have been performed to develop a clean and environmentally benign method for producing H2 from renewable feedstocks in order to achieve a zero-carbon footprint in all sectors.

Based on various H2 production processes, water splitting using a photocatalyst is a promising environmental benign process to produce H2 from water and sunlight. This process involves the dissociation of water (H2O) molecules to H2 and oxygen (O2) by the generated electron (e–)–hole (h+) pairs in a semiconductor when it absorbs appropriate energetic light.[5,6] Currently, the most utilized photocatalyst is titanium dioxide (TiO2) due to its low cost, high photochemical and thermal stability, and environmentally friendly nature.[7,8] However, its wide bandgap (∼3.2 eV) and fast rate of e––h+ recombination serve to limit its applications under visible light, which lower its photocatalytic activity.

Many strategies have been attempted to shorten the bandgap energy of TiO2 to promote more visible-light absorption and slow down the rate of e––h+ recombination of TiO2. These include the modification of the TiO2 structure by metal doping,[9−16] nonmetal (Nm) doping,[17−19] metal and Nm codoping,[20,21] and coupling with other semiconductors,[22−25] as well as tuning oxygen defects.[26,27] Moreover, in order to mitigate the fast rate of e––h+ recombination, various organic and inorganic substances have been extensively used as h+ scavengers (or e– donors), such as organic compounds (e.g., alcohols, amine compounds, organic acids, etc.),[28−31] inorganic compounds (e.g., sodium sulfide and sodium sulfite mixtures),[31] and paper.[32] These added substances are able to react irreversibly with the photogenerated h+, leaving the strong reducing photogenerated e– to react with H+ to form H2 molecules.[6]

For the purpose of green energy production and environmental remediation, the utilization of organic and/or inorganic substances in wastewater as h+ scavengers instead of synthetic chemical addition is beneficial in terms of both economic and environmental points of view. This is because in addition to H2 being produced via a green process, the pollutant level in the wastewater is reduced in the photocatalytic process.[6] Therefore, many attempts have focused on the production of H2 by a photocatalytic process from various wastewaters such as those from glycerol,[33] ammonia,[34,35] palm oil mill,[36] olive mill,[37,38] fruit juice,[39] and sulfide[40] production.

Throughout 2017–2020, our research group has focused on modifying commercial TiO2 (P25) for H2 production together with pollutant removal from biodiesel wastewater. This is because this wastewater contains a high content of glycerol, soap, and some saturated/unsaturated fatty acids (e.g., methyl caprate, methyl laurate, lauric acid, methyl palmitoleate, and methyl 9-octadecenoate), as well as residual biodiesel.[41] Various modification methods have previously been conducted to promote a higher photocatalytic activity of TiO2, such as thermal treatment,[42,43] metal doping with gold (Au), nickel (Ni), palladium (Pd), and platinum (Pt),[44] and heterojunctions with other semiconductors (Bi2O3, Nb2O5, and WO3).[45] This work attempted to modify the TiO2 structure with a combined process of thermal treatment and Nm doping using ultrasound-assisted impregnation. Three types of Nm elements, carbon (C), silicon (Si), and phosphorus (P), were employed to promote the photocatalytic activity for simultaneous H2 production and pollutant removal, the latter measured in terms of the chemical oxygen demand (COD) from biodiesel wastewater.

Results and Discussion

Effect of Pollutant Loading

As mentioned previously, the organic molecules in wastewater can act as the h+ scavengers or e– donors and so can prolong the e––h+ pair life as well as the photocatalytic activity.[44] Prior to the photocatalytic reaction, they must be properly adsorbed onto the photocatalyst surface under a dark condition. Thus, the appropriate pollutant loading should be determined to get an effective operating condition since a too low pollutant loading would give ineffective e––h+ separation, while a too high pollutant loading may saturate the surface of the photocatalyst and become a shading barrier to absorb the photon energy from irradiated light leading to a low photocatalytic efficiency.[46]

In this study, the pollutant loading was designed as a function of the COD level as a high COD level (∼7322 ± 736 mg/L, obtained from pretreated wastewater) and a low COD level (∼1574 ± 275 mg/L obtained from a dilution of the pretreated wastewater). As shown in Figure , the wastewater with a high COD level provided a higher quantity of H2 production and COD removal for all selected photocatalysts. This might be attributed to the presence of adequate organic molecules in the high COD level-containing wastewater, to attach to the generated h+ in the VB and consequently slow down the rate of e––h+ recombination.

Figure 1

Effect of wastewater concentrations on H2 production and COD removal using the TiO2 and Nm1/TiO2 photocatalysts at a loading of 4.0 g/L, a light intensity of 5.93 mW/cm2, and an irradiation time of 4 h.

Effect of the Type of Doped Nm

Using ultrasound-assisted impregnation facilitated the decoration of each of the C, Si, and P Nm elements into the calcined TiO2 (Figure ) to a final level close to the target value (Table ). Doping of the respective Nm element into TiO2 did not markedly affect the crystal phases and structures of TiO2, as seen in Figure a. All the Nm1/TiO2 photocatalysts exhibited almost similar XRD peaks, which were the main characteristic peaks of the calcined TiO2-based material. That is, they exhibited the characteristic peaks of the anatase phase at 2θ angles of 25.30, 37.91, 48.00, 54.00, 55.04, 62.67, 68.87, 70.31, and 75.07°, related to the crystal planes of A(101), A(004), A(200), A(105), A(221), A(204), A(116), A(220), and A(215), respectively. In addition, they also displayed the three characteristic peaks of TiO2 in the rutile phase at 2θ angles of 27.37 and 36.08°, corresponding to the crystal planes of R(110) and R(101), respectively. This was probably attributed to the presence of the respective Nm on the surface of the TiO2-based material at a low quantity and/or a high dispersion level.[47] By using the Spurr and Myers and Debye and Scherrer equations,[48] the Nm1/TiO2 photocatalysts were found to exhibit almost similar anatase contents (∼91%) and crystal sizes (∼21 nm), as summarized in Table . This indicated that the doping of Nm elements via ultrasound-assisted impregnation at the investigated content did not change either the anatase content or the crystallite phase of the calcined TiO2.

Figure 2

Representative EDS of the Nm1/TiO2 photocatalysts.

Table 1

Properties of the Prepared Photocatalysts

type of photocatalyst	target content (wt %)	actual contenta (wt %)	anatase contentb (%)	crystal size (nm)c	bandgap energyd (eV)	PZC	SBET (m2/g)
TiO2			90.12	22.19	3.32	6.77	83.19
C1/TiO2	1.0	1.44 ± 0.09	91.10	21.98	3.21	5.83	78.41
Si1/TiO2	1.0	0.98 ± 0.03	90.85	22.06	3.26	4.65	55.48
P1/TiO2	1.0	0.98 ± 0.01	91.92	21.75	3.22	2.98	79.38
P3/TiO2	3.0	3.32 ± 0.02	91.09	22.03	3.22	2.39	57.83
P5/TiO2	5.0	4.91 ± 0.02	91.18	22.45	3.22	2.23	46.79
P7/TiO2	7.0	6.88 ± 0.02	91.56	22.84	3.22	2.19	45.02
P9/TiO2	9.0	9.31 ± 0.03	91.39	22.66	3.22	2.15	44.57

Estimated from the SEM-EDS analysis.

Calculated from the A(101) and R(110) peaks of the XRD pattern using the Spurr and Myers equation.

Calculated from the A(101) peak of the XRD pattern using the Debye and Scherrer equation.

Calculated from the UV–vis spectra using the Tauc equation.

Figure 3

Representative XRD analysis of (a) Nm1/TiO2 and (b) P/TiO2 photocatalysts.

Regarding the light absorption capacity of the Nm1/TiO2 photocatalysts (Figure ), both the C1/TiO2 and Si1/TiO2 photocatalysts exhibited a higher UV light absorption capacity than that of the calcined TiO2 at identical wavelengths (λ < 415 nm), but they cannot promote visible-light adsorption (λ > 415 nm). This is not the case for the P1/TiO2 photocatalyst because it exhibited an obvious visible-light absorption ability at wavelengths above 415 nm. Quantitatively, according to the Tauc plot, the explicit values of the bandgap energy of the TiO2, C1/TiO2, Si1/TiO2, and P1/TiO2 photocatalysts were 3.32, 3.21, 3.26, and 3.22 eV, respectively.

Figure 4

Representative UV–visible-light absorption spectra of (a) Nm1/TiO2 and (b) P/TiO2 photocatalysts and their Tauc plots (inset figures).

According to the XPS analysis, all Nm1/TiO2 demonstrated the spectra of O 1s, Ti 2p, and C 1s peaks of O, Ti, and C (from carbon tape) (Figure a). In addition, Figure b displays the Si 2s and P 2p peaks of Si and P at binding energies of 99.7 and 135 eV, respectively. The presence of the C 1s peak of C was overlaid with that of carbon tape. Nevertheless, the peaks can indicate the replacement of the O atomic lattices by the doped C, Si, or P elements, which usually affected the effects of the electronic structure of the TiO2-based material.[47,49] That is, the doped elements can cause the formation of an interstage energy level above the VB edge of TiO2 (O in 2p). The excitation of electrons from this interstage energy level to the conductance band (CB) induces the redshift of UV–vis spectra.[50] Typically, the interstage energy level of the bands introduced by the Nm dopants increased with decreasing electronegativity.[51] In this case, the electronegativities of the C, Si, and P elements were 2.5, 1.8, and 2.1, respectively, so the distance of the interstage energy level with respect to the VB was ranked in the order of Si > P > C. Too short or too long distance between the interstage energy level and the VB may lessen the capability of electron excitation from the VB to the interstage level and/or from the interstage level to the CB. The P1/TiO2 photocatalyst exhibited a greater redshift than the other prepared photocatalysts, which is probably due to its more appropriate interstage energy level. In addition, at a low binding energy, the TiO2, C1/TiO2, Si1/TiO2, and P1/TiO2 photocatalysts exhibited the maximum edge position of the VB at energies of around 2.60, 3.00, 1.89, and 2.43 eV, respectively (Figure ). This indicated that the minimum CB of the respective prepared photocatalysts would occur at −0.72, −0.21, −1.37, and −0.79 eV. Interestingly, the valence band XPS spectra of Si1/TiO2 and P1/TiO2 exhibited additional diffusive electronic states above the valence band edge of the TiO2-based material. These confirmed the presence of mid-band states coming from the fulfillment of Si and P species as impurities above the VB. A similar feature was also reported in the C-, S-, and N-doped TiO2.[52,53] The band position of the Nm1/TiO2 photocatalysts was roughly sketched (Scheme ).

Figure 5

Representative XPS spectra at (a) whole range and (b) low ranges of binding energy of each Nm1/TiO2 photocatalyst.

Figure 6

Representative XPS spectra at a low range of binding energy of the Nm1/TiO2 photocatalysts.

Scheme 1

Sketch of the Band Position of the Nm1/TiO2 Photocatalysts with 5.93 mW/cm2 for 4 h

It is well-known that the dark adsorption of active substances proceeds with the photocatalytic reaction. The quality and quantity of absorbed molecules depend upon the relationship between the PZC of the photocatalytic materials and the pH of the aqueous medium.[46] Due to electrostatic interactions, when the solution pH is higher than the PZC value, the photocatalyst exhibits negative charges, which strongly prefer to adsorb cationic molecules. However, when the solution pH is lower than the PZC value, the photocatalyst is positively charged and readily adsorbs anionic molecules.[54] In practical operation, a solution pH close to the PZC value will induce aggregation of photocatalyst nanoparticles (NPs) due to the reduced electrostatic repulsion among the solid NPs within the liquid. This, consequently, reduced the photocatalytic efficiency through the reduction of the exposed surface area and the shading effect.[54] As demonstrated in Figure a, all the photocatalysts exhibited a buffer behavior during the initial pH of 4. The P1/TiO2 exhibited the lowest PZC compared to the TiO2 and other Nm1/TiO2 photocatalysts (Table ). This might be attributed to the surface modification of TiO2 via the (PO4)3+, which can induce the formation of a large number of negatively charged sites,[55−57] resulting in the marked decrease in the PZC value as well as prolongation of the lifetime of photogenerated charge carriers and also the improvement of the photocatalytic performance. From the effect of the three Nm types (C, Si, and P) on the qualitative rate of e––h+ recombination, the PL analysis was then performed over a wavelength range of 350–550 nm. As shown in Figure , three sharp peaks of PL spectra of all photocatalysts appeared at the wavelengths of 420, 486, and 530 nm, representing the backward transfer of electrons from the CB to the VB. Among all prepared Nm1/TiO2, only the P1/TiO2 exhibited an intensity lower than that of the TiO2-based material, indicating its lower e––h+ recombination rate compared with other Nm1/TO2.

Figure 7

Relationship between the initial and final pH of (a) Nm1/TiO2 and (b) P/TiO2 photocatalysts.

Figure 8

PL spectra of all photocatalysts with peaks appearing at wavelengths of 420, 486, and 530 nm.

For the textural properties, shown in Figure , the Nm1/TiO2 photocatalysts exhibited typical H4-shaped hysteresis loops of mesoporous materials, similar to those of M/TiO2, where M is various decorated metals, including Ni, Au, Pt, and Pd,[44] and S/TiO2, where S is a semiconductor, including Bi2O3, Nb2O5, and WO3.[45] Moreover, all the Nm1/TiO2 photocatalysts exhibited a lower BET surface area (SBET) than the original TiO2 (Table ). This was attributed to the partial loss of the TiO2 surface area caused by occupation of the doped Nm elements. However, the C1/TiO2 and P1/TiO2 photocatalysts exhibited comparable SBET values of 78–79 m2/g, which were higher than that of the Si1/TiO2 photocatalyst. This might be attributed to the large atomic size of Si (1.11 Å) compared with C (0.67 Å) and P (0.80 Å), and so, it required a large contact surface area.

Figure 9

Representative N2 physisorption isotherm and (inset) pore size distribution of the P1/TiO2 photocatalyst.

Figure a displays the photocatalytic activity of all the prepared Nm1/TiO2 photocatalysts, in terms of the simultaneous H2 production and COD removal, from pretreated-biodiesel wastewater in the presence of a 4.0 g/L photocatalyst loading and UV irradiation at 5.93 mW/cm2 for 4 h. All the Nm1/TiO2 photocatalysts exhibited a higher H2 production than the TiO2-based material and were ranked in the order of P1/TiO2 > Si1/TiO2 > C1/TiO2. Both C1/TiO2 and P1/TiO2 exhibited a higher COD removal level than TiO2, while Si1/TiO2 provided a broadly comparable level of COD removal. Here, it seems that only the PZC value played a role in either H2 production or COD removal.

Figure 10

Effect of (a) nonmetal types (Nm1/TiO2) and (b) phosphorus loading levels (P/TiO2, x = 1, 3, 5, 7, and 9 wt %) on the simultaneous H2 production and COD removal from the pretreated-biodiesel wastewater via photocatalysis with a 4.0 g/L photocatalyst and a light intensity of 5.93 mW/cm2 for 4 h.

In the photocatalytic process, when the photocatalyst absorbs light with a photon energy equal to or higher than its bandgap energy, the electron is excited from the VB to the CB, leaving a h+ at the VB.[5] Subsequently, several reactions can proceed via the photogenerated h+ and e– depending on their redox potentials. The generated h+ is able to oxidize H2O to the hydroxyl radicals (HO·) and O2 together with photons (H+) at potentials of +2.27 and +1.23 V/NHE (pH 7) as reactions and R2, respectively. In the meantime, the photogenerated e– is capable of reducing the dissolved O2 to form superoxide radicals (O2·–) and hydrogen peroxide (H2O2) at potentials of −0.28 (reaction ) and +0.28 V/NHE (reaction ), respectively.[5,6,58,59] In addition, the photogenerated e– can react with the H+ generated from H2O dissociation with h+ to form H2 according to reaction

In the presence of organic substances, the HO·, O2·–, and h+ are able to degrade the contained substances to form various intermediate species.[45,58] Generally, the types of generated intermediate species depend upon the type of organic substance. For example, the RCH2O–, RCH2O·, R′C·HOH, R′CHO, [R′COOH]−, and H+ can be generated if glucose (RCH2OH or R′CH2OH) is used as the source of organic molecules.[60] The generated H+ is readily reduced via the photogenerated e– according to reaction to form H2, while the generated intermediate species can be further oxidized to form lower-molecular-weight intermediate species, as well as CO2 in the case of complete oxidation, as shown in reactions –R8

Based on all the above-mentioned reactions, a high quantity of the produced oxidizing agent can promote a high reforming efficiency of organic substances. This was typically based on the band position of the prepared Nm1/TiO2 photocatalysts. According to the sketch of the band position of the Nm1/TiO2 photocatalysts (Scheme ), the TiO2 and P1/TiO2 photocatalysts exhibited a more positive VB position than the oxidation potential of H2O to HO· radicals and also displayed a more negative CB position than the reduction potential of O2 to O2·– radicals. Thus, both HO· and O2·– radicals would be formed in the presence of either photocatalyst and were effective to degrade pollutant molecules.

A high COD removal level was observed via the P1/TiO2 photocatalyst compared with TiO2 due to its shorter bandgap energy to absorb a high quantity of light. However, only the HO· radicals could be produced in the presence of C1/TiO2, while only the O2·– radicals could be generated in the presence of Si1/TiO2. The high COD removal level by the C1/TiO2 photocatalyst was due to the fact that HO· radicals have a stronger oxidizing power than O2·– radicals. For the H2 production, the bottom of the CB was more negative than the reduction potential of H+ in all the photocatalysts, indicating their ability to produce H2 via the photocatalytic reaction. The quantity of H2 produced was greater with P1/TiO2 than with Si1/TiO2, which was probably due to its shorter bandgap energy. The C1/TiO2 exhibited a lower H2 production than the P1/TiO2 even though it exhibited a shorter bandgap energy, which was probably due to the limited H+ quantity available to proceed with reaction . Notably, only H2 was detected via our gas analysis system without a trace of CO2. It can be implied that complete degradation of organic substances as reactions –R8 was not accomplished via the synthesized photocatalyst. Thus, at this stage, it is worth noting that the P1/TiO2 photocatalyst exhibited the best photocatalytic activity for both H2 production and COD removal due to its short bandgap energy and the presence of appropriate VB and CB band positions in its structure. In addition, it possessed a lower PZC value, which can induce more contact of photocatalyst NPs and promote an effective transfer of e– and h+ to the adjacent NPs[61] as well as a prolongation of the e––h+ lifetime. In summary, the H2 production process together with the decontamination mechanism can be roughly proposed as the Scheme .

Scheme 2

Sketch of H2 Production and Decontamination Mechanisms in Wastewater

Effect of the P Loading

To enhance the photocatalytic activity of P/TiO2 for simultaneous H2 production and COD removal, different contents of P were doped into the structure of the TiO2-based materials via ultrasound-assisted impregnation. As shown in Table , the actual P element contents in the respective P/TiO2 photocatalysts, as measured using SEM-EDS analysis, were close to the target weight content. Doping different amounts of the P element did not alter the crystal structure of the calcined TiO2 (Figure b) nor the bandgap energy, which remained at 3.22 eV (Figure ), suggesting that doping with the P element at up to 9.0 wt % did not affect the internal structure and optical property of the P/TiO2 photocatalysts. However, increasing the content of the P element doped into the calcined TiO2 increased the amount of surface negative charges on the TiO2 and led to a marked decrease in the PZC value (Figure and Table ). With respect to the textural properties, all the prepared P/TiO2 photocatalysts exhibited typical H4-shaped hysteresis loops of mesoporous materials (figure not shown). However, increasing the amount of the doped P element on the TiO2-based material reduced the SBET of the P/TiO2 photocatalysts.

The photocatalytic activity of the P/TiO2 photocatalyst for simultaneous H2 production and COD removal is summarized in Figure b. All the P/TiO2 photocatalysts were clearly more effective in producing H2 simultaneously with COD removal compared with calcined TiO2. Increasing the P content from 1.0 to 7.0 wt % increased the H2 production and COD removal from 6.43 to 8.34 mmol/g photocatalyst and from 26.0 to 50.6%, respectively. However, further increasing the P element content from 7.0 to 9.0 wt % diminished both the H2 production and COD removal efficiency. This might be attributed to the combined effects of the altered surface charge (as the PZC value) and the reduced SBET of the P/TiO2 photocatalysts. At low P element contents (1.0–7.0 wt %), increasing the P content decreased both the SBET and the PZC values of the respective P/TiO2 photocatalysts (Table ). The increased photocatalytic activity with the decreased SBET of the photocatalysts indicated the insignificant effect of the SBET on the photocatalytic activity. However, a low PZC value might induce high adjoining between PZC values and the pH of wastewater, which will promote contact or aggregation of the adjacent photocatalyst NPs in aqueous systems.[54] The appropriate contact of solid NPs might in turn promote an effective transfer of e– and h+ to the adjacent NPs under the employed thorough stirring condition,[61] resulting in a prolonged e––h+ lifetime as well as an enhanced photocatalytic activity of the P/TiO2 photocatalysts. However, in the presence of a too high P element content (P > 7.0 wt %), the obtained P/TiO2 photocatalyst exhibited an even lower SBET that could then restrict the amount of absorbed light. In addition, it displayed a low PZC value close to the pH of the wastewater, which would induce high contact or aggregation of adjacent photocatalyst NPs, forming a shading behavior and so limiting the capacity to absorb incident light. Consequently, the photocatalytic activity for H2 production and COD removal was reduced.

Reusability of the P7/TiO2 Photocatalyst

The reusability of the P7/TiO2 photocatalyst was evaluated without any chemical or thermal treatment. After the first use, the photocatalyst NPs were separated from the processed wastewater by filtration and washed thoroughly with deionized water until the pH of the filtrate was close to that of deionized water. The ready-to-reuse P7/TiO2 photocatalyst was then obtained after drying at 105 °C for 3 h. The H2 production and COD removal were found to decrease from 8.34 to 5.39 mmol/g photocatalyst and from 50.6 to 16.2%, respectively, from the first use to the fourth use (Figure ). This decline was due to the loss of the P element from around 6.88 to 2.65% (figure not shown). Nevertheless, the photocatalytic activity of P7/TiO2 after the fourth use was still higher than that of the freshly calcined TiO2.

Figure 11

Reusability of the P7/TiO2 photocatalyst for simultaneous H2 production and COD removal from pretreated-biodiesel wastewater via a photocatalytic process with a 4.0 g/L photocatalyst and irradiation.

The properties of the processed-biodiesel wastewater after photocatalysis using the P7/TiO2 catalyst (4.0 g catalyst/L loading and 5.93 mW/cm2 for 4 h) are summarized in Table . It was clearly demonstrated that the quality of the biodiesel wastewater was improved by the photocatalytic process with P7/TiO2. Although the levels of some pollutants, measured in terms of the COD, oil and grease, and TDS levels, were still higher than the standard level set by the Thai Government for discharge into the environment, this work reclaims the use of a nonprecious photocatalyst to produce clean energy (H2) by using an eco-friendly process.

Table 2

Properties of the Fresh- and Processed-Biodiesel Wastewaters

property	Thai standard	fresh wastewater	pretreated wastewatera	processed wastewaterb
pH	5.5–9.0	4.61 ± 0.01	2.03 ± 0.02	2.21 ± 0.01
soap (wt %)		2010 ± 11.0	1829 ± 10.0	1721 ± 4.25
FFA (wt %)		1.50 ± 0.03	2.58 ± 0.08	1.13 ± 0.07
COD (mg/L)	≤400	15,083 ± 1060	7322 ± 736	3790 ± 494
BOD (mg/L)	≤60	75.50 ± 1.50	15.0 ± 9.0	6.00 ± 1.06
oil and grease (mg/L)	≤15	358.3 ± 25.0	163.3 ± 10.0	110.0 ± 36.7
TDS (mg/L)	≤3000	1592 ± 15.0	2402 ± 11.7	2396.2 ± 3.9
TSS (mg/L)	≤150	308 ± 18.3	100 ± 13.3	55.0 ± 18.3

Fresh-biodiesel wastewater was pretreated with concentrated H2SO4 to a pH of around 2.

Wastewater after photocatalytic processing using P7/TiO2 at a loading of 4.0 g/L, a light intensity of 5.93 mW/cm2, and a radiation time of 4 h.

Table shows the comparative results of H2 production and COD removal from biodiesel wastewater via commercial TiO2 (P25) after being modified by heat treatment,[43] metal decoration,[44] heterojunctions with a semiconductor,[45] and Nm doping (this work) at otherwise identical operating conditions (photocatalyst loading of 4.0 g/L and light intensity of 5.93 mW/cm2 for 4 h). It clearly demonstrated that all the modified TiO2 photocatalysts gave a higher H2 production and pollutant removal than the calcined TiO2. Among all the modified TiO2 photocatalysts, the Pd-decorated TiO2 exhibited the highest H2 production level of up to 540 mmol (1350 mmol/g photocatalyst). Although Pd is too expensive compared with other modified substances, further evaluation of its reusability should be performed. In addition, a kinetic study in a batch reactor should be examined to get the data for the design of a pilot reactor in order to compare the H2 production efficiency and the economic break-even point.

Table 3

H2 Production and COD Removal from Biodiesel Wastewater via a Photocatalytic Reaction at Otherwise Identical Operating Conditions (Photocatalyst Loading of 4.0 g/L, Light Intensity of 5.93 mW/cm2, and Irradiation Time of 4 h)

		soap		FFA		COD		oil and grease
photocatalyst	H2 production (mmol)	initial (mg/L)	final (mg/L)	initial (mg/L)	final (mg/L)	initial (mg/L)	final (mg/L)	initial (mg/L)	final (mg/L)	author
T400	0.228	npa	0.09–1.07	np	0.02–0.04	28,600 ± 141	24,738	693 ± 83	235 ± 15	(43)
Pd3/T400	540	np	11.0 ± 0.09	np	2.21 ± 0.01	25,460 ± 688	19,856 ± 4189	200 ± 24	84 ± 16	(44)
B5/TiO2	3.79	19.6 ± 0.50	16.9 ± 0.20	8.23 ± 0.09	7.01 ± 0.10	27,008 ± 1016	18,831 ± 115	313 ± 95	235 ± 15	(45)
P7/TiO2	3.34	1829 ± 10.0	1721 ± 4.25	2.58 ± 0.08	1.13 ± 0.07	7322 ± 736	3790 ± 494	163.3 ± 10.0	110 ± 36.7	this work

Not reported.

Conclusions

In this work, the photocatalytic activity of the calcined TiO2 was improved by doping with C, Si, or P via ultrasonic-assisted impregnation. The P1/TiO2 photocatalyst exhibited the highest photocatalytic activity for H2 production and COD removal compared to the C1/TiO2, Si1/TiO2, and calcined TiO2 photocatalysts. This is because the doped P element furnished new optical properties to TiO2 by shortening the bandgap energy from 3.32 to 3.22 eV, so allowing more visible-light absorption. In addition, the positions of the VB and CB of P1/TiO2 still covered the potential to produce H2 and oxidizing agents in the system. In addition, it exhibited an appropriate surface charge, which can induce an effective transfer of e– and h+ as well as a prolongation of the e––h+ lifetime. Varying the amount of the doped P element revealed that the P7/TiO2 gave the highest photocatalytic activity for simultaneous H2 production and COD removal from biodiesel wastewater due to its appropriate PZC value and SBET. A 35% decreased H2 production and 68% reduced COD removal were obtained after the fourth use compared to the fresh P7/TiO2 catalyst, but these values were still 1.5- and 2.0-fold higher, respectively, than those obtained with the pristine calcined TiO2.

Materials and Methods

Preparation of the Nm/TiO2 Photocatalyst and Characterizations

The preparation of the Nm/TiO2 photocatalyst was adopted from the previous works.[62,63] Three Nm elements (C, Si, and P) were individually doped into TiO2 by ultrasonic-assisted impregnation to form the Nm/TiO2 photocatalysts, where x is the content of Nm in the range of 1.0–9.0% by weight (wt %). Initially, TiO2 was prepared from the calcination of commercial TiO2 (P25; 99.5%, Sigma Aldrich) in a muffle furnace (PLF160/9B, Protherm) at 400 °C for 3 h.[42] Next, approximately 2.97 g of calcined TiO2 was dispersed in 40 mL of 50% (v/v) ethanol (99.9%, Ajax) in distilled water. Meanwhile, approximately 0.0657 mL of phosphoric acid (85.0%, QRec) was mixed with 10 mL of distilled water by rigorous mixing for 15 min and then added into the above TiO2 slurry. The obtained acid slurry was stirred at 500 rpm for 1 h and successively sonicated in an ultrasonic bath (NXPC-2010(P)) for 30 min. The excess water was then eliminated from the solid substances via drying in an oven (BE-100) at 110 °C for 24 h. The remaining solid portion was ground using a ceramic mortar and subsequently calcined at 350 °C for 3 h, yielding a 1 wt % phosphorus-doped TiO2 (P1/TiO2) photocatalyst. A similar procedure was repeated to obtain a phosphorus content in the range of 3.0–9.0 wt %. For the C-doped TiO2 (C/TiO2) and Si-doped TiO2 (Si/TiO2), the same method was still used, except that glucose (99.5%, Ajax) and tetraethoxysilane (99.0%, Sigma Aldrich) were used as the C and Si precursors, respectively.

The point of zero charge (PZC) of all photocatalysts was measured via the pH drift method as previously reported.[64] In brief, the initial pH value of the 0.1 M potassium nitrate (KNO3; Carlo Erba Reagent) solution was adjusted as required between pH 2 and 12 using either 0.1 M nitric acid (Qrec) or 0.1 M sodium hydroxide (Qrec). Then, approximately 0.4 g of the respective photocatalyst was dispersed in 20 mL of the respective pH-adjusted KNO3 solution. The obtained slurry was mechanically shaken at 120 rpm for 24 h. The aqueous solution was separated from the solid portion by filtration, and the final pH was measured using a pH meter (IQ150-77, IQ Scientific). The PZC value was estimated from the intersection point between the line curve of the initial versus final pH values and the line passing through the origin (initial pH = final pH).

The morphologies and optical properties of the prepared photocatalysts were examined as follows. The actual content of the Nm element doped on the TiO2 surface was examined via scanning electron microscopy (SEM) combined with energy-dispersive X-ray spectroscopy (EDS; JEOL). The anatase content and crystal size of TiO2 in each prepared photocatalyst were calculated from the corresponding X-ray diffraction peaks obtained from the X-ray diffraction (XRD) analysis (D8 Advance, Bruker) using the Spurr and Myers and Debye and Scherrer equations, respectively. The bandgap energy was determined from the linear portion of the Tauc plot estimated from the UV–vis absorbance obtained using a UV–visible–near-infrared spectrometer (Lambda 950, PerkinElmer).

The textural properties of the prepared photocatalysts, including the surface area and pore size distribution, were examined via a surface area analyzer (Quantachrome, Autosorb-1) using the Brunauer–Emmett–Teller (BET) and BJH methods, respectively. The rate of e––h+ recombination was qualitatively monitored using photoluminescence (PL) spectrometry (LS-55, PerkinElmer). The electron density state at the valence band (VB) of the Nm/TiO2 photocatalyst was evaluated using X-ray photoelectron spectroscopy (XPS; Axis Supra, Kratos).

Photocatalytic Activity of Nm/TiO2 Photocatalysts

The photocatalytic activity of each Nm/TiO2 photocatalyst was evaluated in terms of the simultaneous H2 production and COD removal at room temperature (∼30 °C) and atmospheric pressure using industrial biodiesel wastewater (Table ) as the feedstock (hydrogen source). Prior to performing each photocatalytic assay, the raw biodiesel wastewater, collected from a biodiesel production plant in Thailand that uses waste cooking oil as the raw material, was acidified to a pH of around 2,[41] by the addition of concentrated sulfuric acid (98%, QRec).

The initial properties of the raw and acid-treated wastewater were analyzed as follows. The pH was measured using a pH meter (IQ150–77), while the levels of COD, biological oxygen demand (BOD), oil and grease, total dissolved solids (TDS), and total suspended solids (TSS) were measured according to standard methods[65] and the soap and free fatty acid (FFA) contents by potentiometric titration with 0.01 M hydrochloric acid and 0.1 M potassium hydroxide, respectively.[66,67] Next, 100 mL of either pretreated or diluted pretreated wastewater plus 0.4 g of the respective photocatalyst was placed in a cylindrical glass reactor, located centrally in a UV-protected box equipped with a 120 W high-pressure mercury lamp (100–600 nm, RUV533 BC) on the top. Argon (Ar; 99.999%, Linde) was supplied to the photoreactor at a flow rate of 500 mL/min to sweep air from the reactor. At the same time, the liquid slurry was stirred thoroughly at 400 rpm for 1 h to allow the equilibrium adsorption of pollutants on the surface of photocatalysts. The Ar supply was then terminated, and both the inlet and outlet gas valves were closed to initiate a closed system with an inert environment. The photoreactor was then irradiated via the UV lamp at a constant light intensity of 5.93 mW/cm2 for 4 h. Throughout the experiment, the temperature of the photoreactor was maintained at 30 °C using an automatic temperature controller (HS-28A). When the experiment was finished, the generated gas was harvested from the reactor using Ar as the carrier gas and analyzed for its composition and concentration using gas chromatography (Shimadzu 2014) with a thermal conductivity detector. The processed wastewater in the presence of the solid catalysts was filtered in the filtration unit using an electrical pump (66688, Suoka) to separate the filtrate from the solid portion, and then, the properties of the processed wastewater were analyzed.