Effect of Process Parameters on the Microstructure and Performance of TiO2-Loaded Activated Carbon.

In this study, the visible-light-driven photocatalytic regeneration performance of TiO2-loaded activated carbon (TiO2/AC) was effectively improved. By carefully controlling the activation condition at 700 °C for 2 h with a 60% H3PO4 concentration and 3:1 TBT (tetrabutyl titanate) impregnation ratio, 90.5% of methylene blue (50 mg/L) was removed within 2 h by a low-dose TiO2/AC (0.5 g/L), which was much higher than those obtained in previous studies on TiO2/AC. Moreover, the effects of process variables on the microstructure and performance of TiO2/AC were systematically investigated. The results showed that (1) the long period of activation time effectively inhibited the photogenerated charge carrier recombination and enhanced the regeneration performance of samples; (2) the photogenerated charge carrier recombination rate was lowered initially and then increased as the temperature ascended, whereas the pore volume showed an opposite variation tendency, and thus the adsorption and regeneration performances of samples were improved at 500-700 °C and then weakened at 800 °C; (3) the increase of H3PO4 concentration effectively inhibited the charge carrier recombination and had an improvement in the adsorption and regeneration performances of samples; and (4) the photogenerated charge carrier recombination rate and bandgap value of samples decreased initially and then increased with increasing TBT mass ratio, so the regeneration performances of samples were improved initially and then lowered.

Introduction

Activated carbon (AC) is widely used for water purification due to its excellent adsorption performance and fast adsorption kinetics.[1,2] However, AC adsorbs contaminants mainly by physical action; once AC is saturated with the adsorptive species, it is mostly discarded in landfills. Therefore, regeneration of saturated AC is indispensable to minimize operational costs and product waste. In previous studies, thermal,[3] chemical,[4] electrochemical,[5] biological,[6] steam,[7] and microwave irradiation methods[8] have been used for AC regeneration. Compared with these methods, photocatalytic regeneration presents great superiority because of its nontoxic products and features of low cost and less secondary pollution. In recent years, TiO2 as an ideal photocatalyst has usually been used for the photocatalytic regeneration of AC. During photocatalytic regeneration, the TiO2 interacts with light of sufficient energy to produce reactive oxidizing species (ROS), which can cause the degradation of pollutants and then achieve the regeneration of AC.

However, due to the large energy bandgap and rapid electron–hole recombination rate characteristic of TiO2 and other factors such as the poor stability of TiO2 and adsorption property, the regeneration efficiency of TiO2/AC is still not satisfactory. To solve this problem, the doping modification with metal or nonmetal materials seems to be a common approach and is widely adopted by previous studies.[9,10] Indeed, its regeneration performance is improved by these treatments, but the TiO2/AC synthesis cost is further increased. Moreover, this doping modification only aimed to enhance the photocatalytic activity, while the adsorption performance of TiO2/AC is weakened in most cases. According to the previous studies, the regeneration performance of TiO2/AC is mainly determined by its adsorption and photocatalytic properties,[11,12] where the properties show a tight relationship with the microstructure of samples. On the one hand, the pore structure and surface chemistry structure are considered to be the key factors affecting the adsorption property of TiO2/AC. The more developed the pore structure is, the higher is the adsorption performance. Meanwhile, Bandosz[13] discovered that there was a significant difference in the adsorption capacity for adsorbents with a similar pore structure; this is because the oxygen-containing functional groups at the TiO2/AC surface played a major role in defining its hydrophilicity, hydrophobicity, polarity, acidity, and reactivity.[14,15] For example, an increase of hydrophilic groups can enhance the polar compound adsorption and reduce the nonpolar compound adsorption;[14,15] the acidic groups favor the alkali compound adsorption, while the alkali groups favor the acidic compound adsorption; C=O and O=P functional groups represent a great adsorption driving force toward cationic dyes due to their π–π conjugation structure.[16] On the other hand, the photocatalytic activity of TiO2/AC is highly determined by the band gap and electron–hole pair recombination rate. The narrow band gap can behave as a sensitizer to increase visible-light absorption capacity and improve visible-light responsive photocatalytic activity,[17] which is often influenced by factors such as phase structure,[18] crystal size,[19] and oxygen vacancy.[20] High recombination rate of electron–hole pairs, which have faster kinetics than the surface redox reactions, can significantly reduce the quantum efficiency of photocatalytic oxidation.[21] Therefore, TiO2/AC with a low electron–hole pair recombination rate would have a strong photocatalytic activity.

Considering the above microstructure factors influencing the regeneration performance of TiO2/AC, we infer that selecting a suitable preparation process is a promising and efficient method to enhance its regeneration performance. This is because the preparation process greatly affects the microstructure of TiO2 and AC, respectively. Pinjari et al. discovered that with an increase in the calcination time, there was an increase in the average crystallite size and rutile content.[22] Mamaghani et al. demonstrated that the charge separation efficiency of TiO2 steadily improved while the surface porosity and OH density diminished with increasing calcination temperature from 300 to 800 °C.[23] Nayak et al. had confirmed that a long heating time of greater than 1 h was seen to have an adverse effect on the surface area of AC.[24]

To the best of our knowledge, the effect of process parameters on the microstructure and performance of TiO2/AC has not yet been systematically studied. Therefore, the purposes of the present work are to investigate the influences of different operational parameters such as activation time, activation temperature, H3PO4 concentration, and TBT impregnation ratio on the microstructure and regeneration performance of TiO2/AC. Meanwhile, we synthesized TiO2 in the wooden pores and spontaneously doped C–N–P atoms into the TiO2 lattice to increase the stability and photocatalytic activity of TiO2. Finally, the enhancement reasons for the regeneration performance were revealed by examining the crystal structure, chemical component, pore structure, bandgap value, and electron–hole pair recombination rate.

Results and Discussion

Effect of Process Parameters on Crystal Structure

The XRD diffraction patterns of the prepared samples are presented in Figure a–d, and crystal structure parameters are shown in Table . All the samples exhibit diffraction peaks at 25.31, 37.81, 48.01, 55.11, and 65.71°, corresponding to the (101), (004), (200), (211), and (204) reflection planes of anatase TiO2. The peaks observed at 27.41 and 54.31° are attributed to (110) and (211) planes of rutile TiO2,[25] and the peak at 22.51° is assigned to the (600) plane of TiP2O7 that was synthesized by the reaction between pyrophosphoric acid and TiO2.[26]Figure a exhibits XRD patterns of samples prepared for different periods of activation time. The similar peak shape and intensity indicate that the activation time variation had less influence on the crystal structure of samples. Figure b shows XRD patterns of samples calcined at different activation temperatures. All peaks become sharp and obvious, and relative anatase crystallinity increases with increasing activation temperature. As listed in Table , from 500 to 800 °C, WA increases from 65.1 to 77%, SA enlarges from 14 to 18 nm, RC grows from 2.88 to 5.36%, and more TiP2O7 crystals are synthesized. In Figure c, samples prepared at higher phosphoric acid concentrations exhibit more TiP2O7 diffraction peaks, which are assigned to the (511), (600), (660), (1022), (690), (1230), (1260), (1182), (1442), (11111), and (12120) planes, respectively. The result indicates that the TiP2O7 crystal structure became more ordered and integrated with increasing phosphoric acid concentration. In addition, WA, RC, SA, and TiP2O7 content generally increase with increasing H3PO4 concentration. Figure d shows XRD patterns of samples prepared by different TBT mass ratios. Table reveals that both WA and RC increase with increasing TBT mass ratio.

Figure 1

X-ray diffraction patterns of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d).

Table 1

Crystal Structural Characteristics of Different Samples

	phase weight fraction (%)
sample	anatase (WA)	rutile (WR)	relative crystallinity (RC)	crystalline size (SA, nm)
AT-0.5-600-10-2:1	69.9	30.1	3.40	15
AT-1.0-600-10-2:1	65.1	34.9	3.80	14
AT-1.5-600-10-2:1	68.7	31.3	2.88	13
AT-2.0-600-10-2:1	70.4	29.6	3.70	15
AT-1.0-500-10-2:1	65.2	34.8	2.88	14
AT-1.0-600-10-2:1	65.1	34.9	3.80	14
AT-1.0-700-10-2:1	69.5	30.5	4.12	15
AT-1.0-800-10-2:1	77.0	23.0	5.36	18
AT-1.0-600-10-2:1	65.1	34.9	3.80	14
AT-1.0-600-30-2:1	55.9	44.1	6.92	29
AT-1.0-600-40-2:1	52.7	47.3	10.58	33
AT-1.0-600-60-2:1	52.2	47.8	12.34	52
AT-1.0-600-10-1:1	56.7	43.3	4.59	16
AT-1.0-600-10-2:1	65.1	34.9	3.80	14
AT-1.0-600-10-3:1	64.9	35.1	4.38	15
AT-1.0-600-10-4:1	68.7	31.3	5.75	16

It is generally accepted that the mixed-phase (anatase/rutile) TiO2 exhibits a higher photocatalytic activity compared to the pure anatase TiO2 because of its efficient charge separation due to the band edge alignment at the anatase/rutile interface. For the pure phase, anatase TiO2 has enhanced photocatalytic activity compared to rutile TiO2.[27] The crystallization of TiO2 is an important factor influencing the photocatalytic activity. A higher crystallization of TiO2 is conducive to the separation of charge pairs and photocatalytic activity.[28] Furthermore, a small TiO2 crystal size can accelerate the transfer of charge carriers on the surface, thus reducing the chance of photo-electron–hole pair reorganization and improving the photocatalytic activity.[29] As the TiP2O7 content increases, the content of the P atom incorporated into the lattice of TiO2 will also increase, which could markedly reduce the band gap and maintain the stability of anatase at high temperatures. Based on the above results, it is found that activation time had no significant effect on crystal structure including crystal size, phase transformation, relative crystallinity, and anatase and TiP2O7 content, whereas the other process parameters (such as activation temperature, phosphoric acid concentration, and TBT mass ratio) would significantly influence the photocatalytic activity of TiO2 by changing its crystal structure.

Raman spectroscopy is further conducted to characterize the crystalline structure of TiO2, and the results are displayed in Figure a–d. The observed peaks at ca. 625 cm–1 (Eg), 516 cm–1 (A1g), 396 cm–1 (B1g), 200 cm–1 (Eg), and 146 cm–1 (Eg) are characteristic peaks of the predominant anatase phase.[30] The Raman peaks of B1g (145 cm–1) and the multiphoton process (232 cm–1) are observed for rutile nanoparticles.[31] The peaks at the 1041 cm–1 bands can be assigned to the TiP2O7 stretching vibrations,[32] which are in accord with the XRD results. The Raman spectra confirmed the formation of P5+ after P-doping in samples. In the high wavenumber range, two vibration peaks from carbon appeared in all samples. The peak at 1586 cm–1 is a characteristic G-band for graphite carbon, while the other peak at 1354 cm–1 assigned to the D-band is often associated with bond-angle disorders and asymmetric lattices in the graphitic structure.[33] As shown in Figure a, the activation time had no significant effect on the crystal structure of samples. When the activation temperature rose, the anatase peak intensities enhanced gradually with the developing TiO2 crystallization, while the D-band and G-band declined due to the removal of bond-angle disorder. The intensities of TiP2O7 peaks significantly increased at 40–60% H3PO4 concentration, which are attributed to the increasing H3PO4 content. In Figure d, with increasing TBT mass ratio, the intensities of anatase peaks significantly increased, indicating that high TiO2 content produced more anatase crystals.

Figure 2

Raman spectra of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d).

Effect of Process Parameters on Chemical Constitution

To investigate the influence of different process parameters on chemical constitutions of samples, XPS analysis was conducted, and the corresponding surface element content is calculated in Table . As the activation time, activation temperature, and H3PO4 concentration increased, the content of C element generally decreased, while those of O and P elements rose. It is well known that the carbonization or pyrolysis of raw materials takes place at the first calcination process (450 °C for 2 h).[34] When the activation process was carried out for a longer time or at a higher temperature/H3PO4 concentration, some C elements of samples were involved in the activation reaction and then released in the form of CO, CO2, CH4, etc. Moreover, because of the absence of inert gas, these gases produced at the previous stage would lead to a new physical activation reaction and further oxidized the C element. The rise of O and Ti content at higher TBT mass ratio is attributed to the increasing Ti source.

Table 2

Surface Element Content of Samples

sample	C (at. %)	O (at. %)	N (at. %)	Ti (at. %)	P (at. %)
AT-0.5-600-10-2:1	53.98	34.43	1.45	3.19	6.95
AT-1.0-600-10-2:1	44.61	39.41	1.58	4.49	9.9
AT-1.5-600-10-2:1	55.13	33.23	1.58	3.28	6.78
AT-2.0-600-10-2:1	43.61	40.84	1.22	3.95	10.93
AT-1.0-500-10-2:1	63.12	28.43	1.31	2.13	5.00
AT-1.0-600-10-2:1	44.61	39.41	1.58	4.49	9.9
AT-1.0-700-10-2:1	42.87	40.77	1.13	4.61	10.62
AT-1.0-800-10-2:1	23.14	53.36	1.44	6.29	15.76
AT-1.0-600-10-2:1	44.61	39.41	1.58	4.49	9.9
AT-1.0-600-30-2:1	44.17	40.16	1.13	4.36	10.19
AT-1.0-600-60-2:1	40.29	44.33	1.03	3.92	10.43
AT-1.0-600-10-1:1	46.65	38.33	1.35	4.29	9.39
AT-1.0-600-10-2:1	44.61	39.41	1.58	4.49	9.9
AT-1.0-600-10-3:1	34.60	45.13	2.67	6.17	11.43
AT-1.0-600-10-4:1	33.64	46.73	1.62	6.10	11.91

Effect of Process Parameters on Surface Functional Groups

The XPS fitting results of functional groups of all samples are shown in Table . Because the C 1s, O 1s, Ti 2p, N 1s, and P 2p spectra of all samples are similar, only AT1.0-600-10-2:1 is exhibited as an example. The C 1s high-resolution spectra are deconvoluted into four peaks in Figure a: the peak at 284.6–284.8 eV is ascribed to the C–C bond; the peaks centered at 286.0–286.4 and 288.3–288.8 eV are assigned to the C–O bond and C=O bond, respectively;[35] and the peak with low intensity at 289.9–291.1 eV is attributed to the π–π* shake-up satellite peak.[36] The absence of the Ti–C bond with a binding energy of 282 eV indicates that oxygen atoms in the TiO2 lattice were not replaced by the C element.[37] Thus, it is inferred that the C element might exist at the interstitial or surface position in the TiO2 lattice.[38] In Table , the C–C content generally reduced with increasing period of activation time, activation temperature, H3PO4 concentration, and TBT mass ratio, while the oxygen-containing functional group content showed an opposite changing trend. This is because graphite carbon could be oxidized in the process of activation and TiO2 synthesis. First, some graphite carbon could react with H3PO4 to form the oxygen-containing functional groups in the activation process. On the other hand, the reaction between TiO2 and water could produce hydroxyl radicals with strong oxidation, and some graphite carbon reacted with these •OH and formed the C–O, C=O, or π–π* groups. Thus, as the activation time, activation temperature, H3PO4 concentration, and TBT mass ratio rose, increasing graphitic carbon was oxidized to oxygen-containing functional groups. Moreover, the previous studies demonstrated that the more the oxygen-containing functional groups there are, the higher is the adsorption performance.[39] Therefore, it is inferred that the samples prepared by a long period of activation time, high activation temperature, large H3PO4 concentration, and great TBT mass ratio would represent a better adsorption.

Table 3

Surface Functional Groups of the C 1s, O 1s, Ti 2p, and N 1s Region

	C 1s (%)				O 1s (%)			Ti 2p (%)		N 1s (%)
sample	C–C	C–O	C=O	π–π*	C–O	O–P/O–N/O–H/C=O	Ti–O	Ti3+	Ti4+	N–	Ti–O–N
AT-0.5-600-10-2:1	66.56	18.05	8.42	6.97	28.62	61.84	9.54	37.22	62.78	20.17	79.83
AT-1.0-600-10-2:1	67.50	18.92	7.67	5.92	21.56	72.10	6.34	39.31	60.69	24.04	75.96
AT-1.5-600-10-2:1	65.88	19.17	8.15	6.79	27.00	63.18	9.82	50.51	49.49	0	100
AT-2.0-600-10-2:1	64.94	18.53	8.95	7.57	24.24	69.86	5.90	55.59	44.41	0	100
AT-1.0-500-10-2:1	68.60	16.74	9.44	5.22	34.68	55.04	10.28	47.55	52.45	0	100
AT-1.0-600-10-2:1	67.50	18.92	7.67	5.92	21.56	72.10	6.34	39.31	60.69	24.04	75.96
AT-1.0-700-10-2:1	60.13	17.14	10.39	12.34	26.30	62.09	11.61	33.31	66.69	71.69	28.31
AT-1.0-800-10-2:1	61.46	28.08	10.46	0.00	20.93	72.30	6.77	25.70	74.30	0	100
AT-1.0-600-10-2:1	67.50	18.92	7.67	5.92	21.56	72.10	6.34	39.31	60.69	24.04	75.96
AT-1.0-600-30-2:1	67.45	11.24	9.60	11.71	23.55	69.61	6.84	29.58	70.42	0	100
AT-1.0-600-40-2:1	66.62	21.94	7.30	4.15	21.26	76.14	2.60	19.84	80.16	0	100
AT-1.0-600-60-2:1	63.14	19.12	9.44	8.30	26.36	67.36	6.28	17.74	82.26	0	100
AT-1.0-600-10-1:1	68.78	17.96	6.85	6.41	22.85	69.64	7.51	33.64	66.36	23.34	76.66
AT-1.0-600-10-2:1	67.50	18.92	7.67	5.92	21.56	72.10	6.34	39.31	60.69	24.04	75.96
AT-1.0-600-10-3:1	67.60	17.16	8.14	7.10	20.90	67.03	12.07	52.09	47.91	25.02	74.98
AT-1.0-600-10-4:1	67.71	21.01	8.24	3.04	20.50	67.71	11.79	50.51	49.49	15.53	84.47

Figure 3

(a) The C 1s high-resolution spectra of AC1.0-600-10-2:1. (b) The O 1s high-resolution spectra of AC1.0-600-10-2:1. (c) The Ti 2p high-resolution spectra of AC1.0-600-10-2:1. (d) The N 1s high-resolution spectra of AC1.0-600-10-2:1. (e) The P 2p high-resolution spectra of AC1.0-600-10-2:1.

Figure b displays XPS spectra corresponding to the O 1s region. For all samples, the peaks at 530.3–530.9 eV can be ascribed to the Ti–O bond in the TiO2 crystalline, which is a surface hydroxyl group binding with Ti atoms.[40] This peak exhibits a shift to high binding energy, revealing that Ti3+ (or oxygen vacancy) was formed in TiO2 lattice.[41] The peaks at 531.7–531.9 eV are related to the surface O–P/O–N/O–H bond or C=O bond.[42] The peaks at 533.1–533.3 eV are derived from surface oxygen in the C–O bond. In comparison, it is found that the O–P/O–N/O–H/C=O bond was the predominant component. As interpreted in this context, the hydroxyl groups play an important role in photocatalysis and adsorption, as they can react with photogenerated holes to generate •OH for photodegrading pollutants and also serve as adsorption sites for pollutant molecules.[29] Therefore, we speculate that samples with abundant hydroxyl groups would have better regeneration performance.

In Figure c, the two peaks located at 460.5–460.9 and 466.4–466.8 eV can be ascribed to the binding energy of Ti4+ 2p3/2 and Ti4+ 2p1/2, respectively; the other two Ti 2p peaks exhibit a slight shift to a low binding energy at 459.2–459.6 and 465.1–466.1 eV, separately, also confirming the existence of Ti3+ ions.[43,44] Binding energy shifting to low energy will lead to an increase of electron density, which is opposite to the shift to high binding energy.[45] To satisfy the requirement of electrostatic balance, the oxygen vacancies around Ti3+ must exist, and thereby, the surface oxygen vacancies will be generated on the TiO2 surface.[46,47] The Ti3+ content gradually declined as the activation temperature and H3PO4 concentration increased but showed an opposite variation tendency with the increase of activation time and TBT mass ratio, as listed in Table . It was reported that Ti3+ could extend the visible light absorption and improve the visible light photocatalytic activity.[48] Thereby, it is expected that samples with abundant Ti3+ ion would exhibit a better photocatalytic activity.

The N 1s XPS spectra of all samples are given in Figure d. The major peak around 400.1–400.8 eV is assigned to O–Ti–N,[49] and the other peak around 401.5–402.2 eV is attributed to oxidized nitrogen in the form of Ti–O–N.[50] Both of these species come from the interstitial N. There was no obvious peak at 396 eV assignable to Ti–N bonds produced by the replacement of oxygen in the TiO2 lattice.[51] Both the interstitial N and substitution N could lead to the formation of a new mid gap energy state and eventually decrease the band gap of TiO2.[52] However, in comparison, the decrease of bandgap obtained by interstitial N was larger than that achieved by substitution N, which would effectively improve the photocatalytic performance of samples under visible light irradiation in this study.

The XPS spectrum of P 2p is shown in Figure e. The P 2p binding energy appears at 134.2–134.5 eV, demonstrating that P atoms in the sample exist as the pentavalent oxidation state (P5+) in the Ti–O–P linkage rather than as PO43– in a tetrahedral environment.[53] This is attributed to the result that P5+ replaced part of Ti4+ in the crystal lattice of TiO2.

The FTIR spectra for all the prepared samples are presented in Figure a–d. The broad peaks at 400 to 800 cm–1 for all samples are ascribed to bending vibrations of Ti–O or O–Ti–O bonds.[54] The broad band at 920–1300 cm–1 (peaks at 960 and 1080 cm–1) is ascribed to C–O stretching in acids, alcohols, phenols, ethers, and/or ester groups.[55] The broad peaks near 1606 and 3380 cm–1 can be assigned to the flexural vibrations of H–O–H groups due to the chemisorbed surface water and stretching vibrations of −OH bonds from hydroxyls.[56] The absorption band at 1710 cm–1 is attributed to C=O stretching vibrations of carboxyl or carbonyl groups.[57] Although FTIR analysis is usually suitable for qualitative assessment of functional groups, it is reasonable to deduce the relative content of the groups according to the corresponding peak intensity.[58] In Figure a, the similar peak intensity indicates that activation time had no significant effect on functional groups. In Figure b–d, the intensity of the peak corresponding to C–O groups increased gradually with the increase of activation temperature, H3PO4 concentration, and TBT mass ratio, which were consistent with the XPS results. Moreover, the peak of the Ti–O group is gradually broadened with the TBT mass ratio ascending due to the proportion increase of the Ti source.

Figure 4

The FTIR spectra of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d).

Effect of Process Parameters on Thermal Stability

The thermogravimetric analysis (TGA) was used to understand the thermal stability of the samples. As shown in Figure b–d, the first stage of weight loss from room temperature to 200 °C was attributed to the loss of desorption of physisorbed water,[59] which had been confirmed by the FTIR measurement. The second stage (≥500 °C) had a significant weight loss signifying the carbon residue degradation as well as TiO2 crystal phase transformation processes.[60] The samples prepared for different periods of activation time have similar weight change trends. The maximum weight loss is observed at 0.5 and 1.5 h activation time. This is attributed to the relatively high carbon content (confirmed by XPS). In Figure b, the weight loss of samples decreased gradually as the temperature rose. This is because high activation temperature led to more organic compounds and carbon degradation during sample preparation and a better thermal stability. When the H3PO4 concentration rose from 10 to 30%, the weight loss was reduced due to high carbon content and TiO2 crystal phase transformation from anatase to rutile, while at 40–60%, the increasing organic compounds (confirmed by FT-IR) resulted in the increase of weight loss. As the TBT mass ratio rose, the weight loss decreased gradually, which is ascribed to the increment of TiO2 content.

Figure 5

The TG curves of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d).

Effect of Process Parameters on Pore Structure

Pore structures of samples were determined by N2 adsorption–desorption isotherms, which are shown in Figure S1a–d (see Appendix). The details of pore structures calculated by N2 adsorption–desorption isotherms are quantified in Table . The pore parameters of all samples obtained by different calculation models have similar variation trends. As the activation time and temperature increased, the specific surface area and pore volume of samples increased initially and then decreased. This is because the longer duration of activation time (≥1 h) caused some of the pores to enlarge or even collapse.[61] On the other hand, during activation temperature at 500–700 °C, the carbonization of hemicellulose, cellulose, and lignin components of SP caused the pore opening along with the development of new pores, while with increasing temperature (>800 °C), the realignment of the carbon structure played a destruction and enlargement effect on most existing pores.[24] Unlike the activation time and temperature, the increasing H3PO4 concentration and TBT mass ratio led to a decrease of the specific surface area and pore volume for samples. This is because the existing pores were destroyed at high H3PO4 concentrations, as well as the excessive TiO2 aggregation at large TBT mass ratios.[24] The pore size distributions of samples are displayed in Figure a–d. The pore sizes of all samples are mainly concentrated in 0.4 and 2.0 nm, which confirm their highly developed micropores. The variation trend of pore size was similar to that of the specific surface area and pore volume with the process parameter changing.

Table 4

Surface Area and Pore Volume of Samples

	total surface area (m2/g)			total pore volume (cm3/g)		micropore surface area (m2/g)		micropore volume (cm3/g)		mesopore surface area (m2/g)		mesopore volume (cm3/g)
sample	SBET	SLan	Stotal	Vtotal	VDFT	St-Plot	SD-A	Vt-Plot	VD-A	SBJH	SMeso	VBJH	VMeso
AT-0.5-600-10-2:1	181	219	165	0.096	0.077	160	219	0.073	0.084	38	21	0.026	0.023
AT-1.0-600-10-2:1	221	252	203	0.133	0.101	159	242	0.072	0.098	83	62	0.059	0.061
AT-1.5-600-10-2:1	157	188	143	0.087	0.075	133	169	0.061	0.068	38	24	0.027	0.026
AT-2.0-600-10-2:1	165	196	151	0.090	0.068	143	199	0.065	0.076	36	22	0.027	0.025
AT-1.0-500-10-2:1	177	215	162	0.093	0.071	156	210	0.071	0.081	37	21	0.024	0.022
AT-1.0-600-10-2:1	221	252	203	0.133	0.101	159	242	0.072	0.098	83	62	0.059	0.061
AT-1.0-700-10-2:1	251	298	228	0.129	0.110	219	289	0.100	0.116	64	32	0.036	0.029
AT-1.0-800-10-2:1	7	5	6	0.014	0.013	3	4	0.001	0.002	8	4	0.014	0.013
AT-1.0-600-10-2:1	221	252	203	0.133	0.101	159	242	0.072	0.098	83	62	0.059	0.061
AT-1.0-600-30-2:1	185	216	169	0.117	0.089	157	224	0.072	0.088	47	28	0.047	0.45
AT-1.0-600-40-2:1	142	161	129	0.091	0.070	121	157	0.055	0.065	44	21	0.040	0.036
AT-1.0-600-60-2:1	138	154	125	0.078	0.057	109	135	0.049	0.057	59	29	0.036	0.029
AT-1.0-600-10-1:1	272	323	248	0.146	0.123	230	321	0.105	0.126	74	42	0.045	0.041
AT-1.0-600-10-2:1	221	252	203	0.133	0.101	159	242	0.072	0.098	83	62	0.059	0.061
AT-1.0-600-10-3:1	206	244	188	0.110	0.084	176	234	0.080	0.092	56	30	0.034	0.030
AT-1.0-600-10-4:1	102	113	94	0.070	0.061	66	108	0.030	0.044	46	36	0.038	0.040

Figure 6

Pore size distributions of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d).

Effect of Process Parameters on Charge Carrier Separation Efficiency

It is accepted that PL can investigate the efficiency of charge carrier trapping, migration, and transfer. The fate of photogenerated electron–hole (e––h+) pairs in semiconductors can be qualitatively determined by PL because PL results from the recombination of photogenerated charges.[62−64] A lower PL peak indicates less e––h+ pair recombination rate. It is generally accepted that the recombination efficiency of the photogenerated charge carriers is influenced by the following factors: first, the higher the Ti3+ defect number is, the higher is the charge separation efficiency, and this is because the Ti3+ defect being generated from the oxygen vacancy could increase the scavenging of electrons;[28] second, the mixed TiO2 phases are beneficial in improving charge carrier separation due to the interfacial transfer of electron between anatase and rutile;[20] third, the high crystalline TiO2 with rapid movement of charge carriers can retard the recombination of photoinduced carriers;[65] fourth, the smaller the particle size of TiO2 is, the faster the electron holes diffuse to the surfaces of grains, which cause a lower probability of charge carrier recombination;[29] and fifth, AC with developed pores served as an electron storage that can capture electrons emitted from the conduction band (CB) of TiO2, thereby improving the efficiency of charge carrier separation.[66] As seen in Figure a–d, the peak positions of all samples are basically the same, signifying that the different treatment conditions have not induced new photoluminescence. As the activation time and H3PO4 concentration increased, the PL peak intensity of samples reduced gradually in the following sequence: AT2.0-600-10-2:1 < AT1.5-600-10-2:1 < AT0.5-600-10-2:1 and AT1.0-600-10-2:1 < AT1.0-600-30-2:1 < AT1.0-600-60-2:1 < AT1.0-600-90-2:1. This result indicated that the longer period of activation time and higher H3PO4 concentration were able to effectively inhibit the recombination of the photogenerated charge carriers. According to the results of XPS, BET, and XRD, it is concluded that the increasing Ti3+ content was the main reason leading to the decline of PL peak intensity with activation time prolongation. The TiO2 crystallization gradually grew with increasing H3PO4 concentration, causing the decline of PL signal intensity. Unlike the activation time and H3PO4 concentration, the PL peak intensity of samples decreased initially and then increased with increasing activation temperature and TBT mass ratio. As the activation temperature rose from 500 to 600 °C, the decreasing PL peak intensity was mainly due to the developing of AC pore structure; at 700–800 °C, the decrease of AC pores/Ti3+ content and the enlargement of TiO2 particle size finally resulted in the increase of PL peak intensity. The sample prepared at the TBT mass ratio of 2:1 had smaller TiO2 particle size and developed AC pores, thereby leading to its low PL peak. Therefore, the photogenerated charge carrier recombination of samples could be effectively inhibited when the preparation condition was carried out at 600–700 °C activation temperature, 2:1–3:1 TBT mass ratio, longer period of activation time, or higher H3PO4 concentration.

Figure 7

Photoluminescence spectra of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d).

Effect of Process Parameters on Bandgap Value

As is known to all, the conduction band value of TiO2 is about −0.2eV.[67] Therefore, the bandgap value of samples was calculated from the results obtained from the VB-XPS and constant CB value. Figure a–d reveals that the calculated bandgap value of all the samples is significantly lower than that of the reported pure TiO2 (∼3.20 eV). Because a narrow bandgap is beneficial to enhance the TiO2 response to visible light, we infer that the photocatalytic activity of TiO2/AC samples in this study will be stronger than that of pure TiO2 in the visible-light irradiation. The reason for the decrease of bandgap value can be attributed to the C, N, and P tri-doping. C and N dopants narrow the bandgap by introducing midgap/surface states into the electronic band structure of TiO2.[68,69] The P doping reduces the bandgap by mixing the P 3p states with the O 2p states.[70] In addition, the microstructure of TiO2 also has a great influence on its bandgap value: first, the smaller the TiO2 particle size (quantum-size effect) is, the larger is the bandgap value;[53] second, the crystal phase transformation from anatase to rutile can lead to a reduction in the band gap because the rutile phase has a narrower bandgap than anatase by ca. 0.2 eV;[18] and third, the increasing Ti3+ defects being generated from the oxygen vacancy can decrease the bandgap value.[71] The preparation conditions of samples play an important role on regulating these microstructure. As seen in Figure a, because the activation time had less influence on the crystal structure of samples, it had no significant effect on the bandgap value, while with increasing activation temperature, the bandgap value gradually increased in Figure b, which contributed to the decrease of rutile (confirmed by XRD) and Ti3+ defect contents (confirmed by XPS). Figure c shows that when the phosphoric acid concentration rose from 10 to 30%, the bandgap value increased due to the decreasing Ti3+ defect content, while at 40–60%, the enlarged rutile content and TiO2 particle size lowered the bandgap value. As the TBT mass ratio rose from 1:1 to 3:1 in Figure d, the bandgap value narrowed owing to the increasing Ti3+ defect, whereas at 4:1, the bandgap value was expanded because of the decreased Ti3+ defect content.

Figure 8

XPS valence band spectra of samples prepared for different periods of activation time (a), different activation temperatures (b), different H3PO4 concentrations (c), and different TBT mass ratios (d). The band-edges were calculated based on the standard CB value and XPS valance band information.

Effect of Process Parameters on Regeneration Performance

The effect of process parameters on the adsorption performance of samples is shown in Figure a–d through examining the A48/A0 value (A0 refers to the initial absorbance of MB solution, and A48 refers to the absorbance of the MB solution after being adsorbed for 48 h in the dark). In Figure a,c, as the activation time and H3PO4 concentration increased, the adsorption performance of samples was enhanced due to the increasing oxygen-containing functional groups, which provided more adsorption site for MB removal. In Figure b,d, it is found that the adsorption performance was improved initially and then lowered with activation temperature and TBT mass ratio ascending. From 500 to 700 °C, the increasing pores and oxygen-containing functional groups were formed in samples, whereas these pores were tremendously damaged by the high temperature. Thus, the samples prepared at 700 °C showed a better adsorption performance. Similarly, the samples prepared by low TBT mass ratios (≤2:1) possessed more oxygen-containing functional groups to adsorb MB, while at TBT mass ratios ≥3:1, the excessive amount of TiO2 blocked the sample pores, which finally led to the decrease of adsorption performance.

Figure 9

Regeneration performance of samples at different periods of activation time (a), activation temperatures (b), H3PO4 concentrations (c) and TBT mass ratios (d).

The regeneration performance of samples was judged by the At/A0 value (At refers to the absorbance of the MB solution at time t, and t ≥ 48 h after visible-light illumination) in Figure a–d. At the visible-light irradiation, all the samples still maintained a certain MB removal effect after reaching the adsorption–desorption equilibrium, which demonstrated that these samples had a regeneration performance. As shown in Figure a,c, the regeneration performance was improved by the increasing period of activation time and H3PO4 concentration. This is because both the adsorption performance and photocatalytic performance were enhanced with the increase of activation time and H3PO4 concentration, where the photocatalytic performance was judged by the analysis results of charge carrier separation efficiency. The recombination of photogenerated charge carriers was effectively inhibited for samples prepared by a long period of activation time. Figure b reflects that the regeneration performance was gradually strengthened with the temperature rising from 500 to 700 °C and then lowered at 800 °C, which shows a similar variation trend with adsorption and photocatalytic performances. Therefore, the sample prepared at 700 °C represented the best regeneration performance because it possessed a narrower bandgap value and lower probability of charge carrier recombination. Coincidentally, as the TBT mass ratio increased, the regeneration performance of samples was improved initially and then became weak in Figure d. Excellent regeneration performance was exhibited by the sample prepared at the TBT mass ratio of 2:1 due to its high adsorption and photocatalytic performances.

The kinetics of MB degradation were analyzed using second-order kinetic equations to get the best process parameters. We carried out a correlation analysis of zero-order, first-order, and second-order kinetic models. After comparison with each other, the degradation of the MB molecules was well represented by the second-order kinetic model, and the other kinetic model with figures and parameters had been omitted. The second-order kinetic equation is expressed in eq :[72]where A0 and A are the absorbance of MB at equilibrium and various times, t is the irradiation time, and k (h–1) is the rate constant of the kinetic model.

A regression analysis based on the second-order reaction kinetics for the MB degradation process was conducted, and the results are shown in Figure a–d. The values of the correlation coefficient (R2) and rate constants (K) were mentioned in Table . The fitting curves also confirmed the effect of different processes on the MB degradation process mentioned above. By comparing the degradation rate (K), it was found that the best process for MB degradation was 2 h activation time, 700 °C activation temperature, 60% H3PO4 concentration, and 3:1 TBT impregnation ratio. Moreover, the high correlation coefficient values revealed that the photocatalytic degradation of MB followed the second-order kinetic model.

Figure 10

Plots for photocatalytic degradation kinetics of samples at different periods of activation time (a), activation temperatures (b), H3PO4 concentrations (c), and TBT mass ratios (d).

Table 5

Kinetic Parameters of Samples

sample	K (h–1)	R2
AT-0.5-600-10-2:1	0.097 × 10–1	0.96
AT-1.0-600-10-2:1	0.16 × 10–1	0.90
AT-1.5-600-10-2:1	0.13 × 10–1	0.85
AT-2.0-600-10-2:1	0.16 × 10–1	0.77
AT-1.0-500-10-2:1	0.073 × 10–1	0.84
AT-1.0-600-10-2:1	0.16 × 10–1	0.90
AT-1.0-700-10-2:1	0.68 × 10–1	0.76
AT-1.0-800-10-2:1	0.14 × 10–1	0.89
AT-1.0-600-10-2:1	0.16 × 10–1	0.90
AT-1.0-600-30-2:1	0.15 × 10–1	0.75
AT-1.0-600-40-2:1	0.85 × 10–1	0.88
AT-1.0-600-60-2:1	2.17 × 10–1	0.75
AT-1.0-600-10-1:1	0.50 × 10–1	0.88
AT-1.0-600-10-2:1	0.16 × 10–1	0.90
AT-1.0-600-10-3:1	2.16 × 10–1	0.87
AT-1.0-600-10-4:1	0.085 × 10–1	0.99

In Table , the photocatalytic degradation performances of TiO2/AC in this work and the literature are compared. It is very encouraging to see that the visible-light-driven photocatalytic degradation performance of TiO2/AC in this work was significantly higher than that of the other TiO2/AC in the literature. Therefore, the present work provided a promising process method to regenerate AC.

Table 6

Comparison of the Photocatalytic Performance of Carbon–TiO2 Composites from the Literature

material	sample dosage (mg)	dye concentration	light source/irradiation time	photocatalytic degradation efficiencya (%)	ref
TiO2/AC	50	50 mg/L MB	visible light/2 h	90.5	this work
TiO2/AC	5	20 mg/L acid orange	UV light/3 h	57.6	(73)
C-TiO2	400	20 mg/L MB	UV light/4 h	52	(74)
TiO2/AC	50	12.9 mg/L 4-chlorophenol	visible light/2 h	89.7	(75)
TiO2/AC (microwave)	37.5	5 mg/L paracetamol	UV light/6 h	80	(76)

The photocatalytic degradation efficiency was calculated using , where A0 is the initial absorbance and At is the final absorbance at given time after irradiation.

Conclusions

The effect of process parameters on microstructure and performance was systematically studied in this work. The main conclusions were listed as follows:

The activation time variation had less influence on the crystal structure and bandgap value of samples. However, the long period of the activation time resulted in a rise of the oxygen-containing functional group and effectively inhibited the recombination of photogenerated charge carriers with the increase of Ti3+ content. Meanwhile, the pore volume of samples increased initially and then decreased with increasing period of activation time. Therefore, the adsorption and regeneration performances were enhanced as the activation time was prolonged.

From 500 to 800 °C, the bandgap was broadened with the decreasing rutile and Ti3+ defect contents, and more oxygen-containing functional groups were formed in the samples. Moreover, the photogenerated charge carrier recombination rate was lowered initially and then increased as the temperature ascended, whereas the pore volume showed an opposite variation tendency. Thus, the adsorption and regeneration performances were improved at 500–700 °C and then weakened at 800 °C.

The increase of H3PO4 concentration facilitated the growth of the TiO2 crystal structure and oxygen-containing functional groups, as well as the destruction of pore structure. Thereby, the recombination of the photogenerated charge carriers was effectively inhibited at high H3PO4 concentration with the developing TiO2 crystallization. Also, the TiO2 bandgap was expanded initially with an increase in H3PO4 concentration for the range of 10 to 30% due to the decreasing Ti3+ defect content and finally lowered as the H3PO4 continued to rise with the increasing TiO2 particle size. All of the above results led to the improvement of adsorption and regeneration performances of samples at high H3PO4 concentration.

The enlargement of the TBT mass ratio gave rise to the growth of the rutile phase and the decline of pore volume. The recombination efficiency of the photogenerated charge carriers of samples decreased initially and then increased with increasing TBT mass ratio. As the TBT mass ratio rose from 1:1 to 3:1, the bandgap value narrowed, whereas at 4:1, the bandgap value was expanded. Therefore, the adsorption and regeneration performances of samples were improved initially and then lowered with TBT mass ratio ascending.

Experimental Section

Materials

The 9 mm diameter Salix psammophila (SP) with a cutting length of 20 mm was collected from Erdos in the Inner Mongolia Autonomous Region, China. Before preparing samples, the SP bark was peeled off. Tetrabutyl titanate (TBT) functioned as a Ti source and was purchased from Tianjin Huihang Chemical Technology Co. Ltd. (China). Methylene blue (MB) used as the simulated pollutant was received from Yun Gong Synthetic Technology Co. Ltd. Absolute ethanol and phosphoric acid were purchased from Sinopharm Chemical Reagent Co. Ltd. (China). All chemicals were used as received without further purification.

Pretreatment of Salix psammophila

Firstly, the extract in the SP cell lumen was removed by conducting the hydrothermal treatment in a thermostat water bath for 3 h at 100 °C. To extend the pore diameter in the cell wall, the obtained SP was wrapped by a tin paper and placed in a microwave field for 5 min treatment. Finally, the treated SP was dried in the air. After this pretreatment, TBT could access the SP cell wall much easier.

Preparation of the TiO2/AC Composite

The pretreated SP (10 g) was immersed in a TBT and absolute ethanol solution with mass ratios of 1:1, 2:1, 3:1, and 4:1 for 24 h. The immersed SP was placed in a constant relative humidity temperature chamber at 85 ± 2% relative humidity (RH) for 24 h to trigger TBT hydrolyzation. The calcination process was carried out at 450 °C for 2 h to produce TiO2–carbon (the TiO2 and carbon composite). The obtained samples were soaked in 10, 30, 40, or 60% H3PO4 solution for 12 h and then dried at 100 °C until a constant weight was reached. The composites were activated at 500, 600, 700, and 800 °C for 0.5, 1, 1.5, and 2 h, respectively, and then washed with distilled water until neutral pH. The obtained samples of different process parameters were labeled as AT-t-a-c-x, where t, a, c, and x indicated the activation time (h), activation temperature (°C), H3PO4 concentration (%), and TBT mass ratio, respectively. The fabrication procedure of the TiO2/AC composites is illustrated in Figure .

Figure 11

Schematic illustration of the TiO2/AC fabrication process.

Characterization

An X-ray diffractometer (XRD, Shimadzu, Japan) and Raman spectroscopy (inVia, Renishaw, Britain) equipped with a 532 nm laser source were used to analyze the crystal structure of samples. The average crystalline size of anatase TiO2 was represented by SA and calculated using Scherrer’s equation from the (101) diffraction peak. The relative anatase crystallinity (RC) was estimated via the relative intensity of the diffraction peak from the anatase (101) plane.[77] The weight fraction of anatase (WA) and rutile (WR) were calculated from eqs and 3:where AA, AR, and AB are the integrated intensity of the anatase (101), rutile (110), and brookite (121) peaks, respectively. KA and KB are two coefficients and their values are 0.886 and 2.721, respectively.

The specific surface area (SSA) and pore volume were determined by using a surface area and porosity analyzer (ASAP-2460, Micromeritics, USA). The specific surface area (SBET, SLan, and Stotal) was calculated according to the BET equation, Langmuir model and single point surface area at P/P0 = 0.300000000, respectively. The total pore volume was determined by the DFT method (VDFT) and single point adsorption total pore volume of pores less than 194.2574 nm diameter at P/P0 = 0.990041377 (Vtotal). The t-plot method (St-Plot, Vt-Plot) and D-A method (SD-A, VD-A) were used to calculate the micropore area and volume. The mesopore area and volume were estimated by the BJH method (SBJH, VBJH) and as the difference between the total pore and micropore area and volume (SMeso, VMeso). The pore width distribution was determined by the DFT method.

Photoluminescence (PL) spectra of the samples were obtained by a fluorescence spectrophotometer (FLS1000, British Edinburgh, U.K.) to analyze the recombination information of the photogenerated electron–hole pairs.

The chemical component, electronic state, and valence band position were analyzed using X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher Scientific, USA) with an Al Kα irradiation source. The change of functional groups of the samples was judged by Fourier-transform infrared spectra (FTIR) using a spectrometer (IS10, Nicolet, USA) between the frequency ranges of 400 and 4000 cm–1 using KBr as the diluent.

The thermal stability of samples was analyzed using a thermal gravity analyzer (STA 449 F3/F5, NETZSCH, Germany). The process was essentially monitored from room temperature to 800 °C at a flow rate of 10 °C/min under a nitrogen gas atmosphere.

Regeneration Measurement

The regeneration performance of samples was evaluated by adsorbing and degrading a methylene blue (MB) solution. A 0.05 g sample and 100 mL of the MB solution with a concentration of 50 mg·L–1 were added to a cylindrical glass reactor. To ensure that the adsorption–desorption equilibrium is reached, the mixture was shaken in the dark at 25 °C for 48 h. Then visible light illumination treatment was carried out for 24 h under a 500 W Xe lamp irradiation. At a given time interval, 4 mL of the suspension was collected by a syringe and filtered through a 0.22 mm nylon syringe filter. The filtrate was then measured by a UV–vis spectrophotometer (TU-1950, Beijing Purkinje General Instruments Co., Ltd., China) to evaluate the regeneration performance of the samples.