Introduction

With industrialization and population growth, the environmental contamination, especially caused by organic pollutants, becomes an increasingly severe problem around the world. And the possession of clean water determines whether an area is suitable for human habitation. In the past decades, various techniques, such as chemical precipitation, adsorption, biological removal, reverse osmosis, photocatalysis and membrane, have been studied for the removal of organic pollutants from wastewater [1–3]. Among these approaches, adsorption and photocatalytic degradation are the two most popular routes.

As the easiest method of decontamination, adsorption has been considered as one of the most promising technologies for organic removal in terms of cost, simplicity of design, and operation. A wide range of adsorbent materials such as activated carbon, clay, silica gel, and cellulose have been reported. Activated carbon stands out among these adsorbents owing to its extensive pore surface area, interconnected pore structure, unique surface chemistry [4–6]. Unfortunately, merely physical adsorption is insufficient for pollutants removal because the toxic pollutants cannot be effectively converted to other non-toxic substances. The adsorbents need to be replaced when they reach the saturation adsorption, which leads to a new problem that how to deal with these adsorbents. However, due to its unique performance, the adsorption process could be considered as a necessary factor to increase the encounter probability between the pollutants and catalysts when the pollutants are degraded by catalysis [7, 8].

As a green technology, heterogeneous photocatalysis by semiconductor materials has gained increasing interest because it is widely applied in environmental purification (non-selective process) and selective organic transformations to fine chemicals in both gas and liquid phases [9–11]. Among the various semiconductor materials, titanium dioxide (TiO2) has been extensively studied and commonly applied as a photocatalyst in environmental cleaning because of its superior photocatalytic activity, high photo-stability, chemical resistance, low cost, abundance, and environmentally benign nature [12, 13]. However, the photocatalytic activity of TiO2 is greatly restricted by its low quantum efficiency in photocatalytic reactions and ineffective utilization of visible light, which result from its high recombination of photo-generated electron–hole pairs and wide band gap (3.2 eV for anatase), respectively [14–18]. Moreover, although nanosized TiO2 catalyst has high photocatalytic activity among the current popular catalysts, it still presents a drawback that the powders are not easy to precipitate and recover from water, which are detrimental to their regeneration and reuse [19]. In order to realize photocatalytic degradation of pollutants used in large-scale water treatment facilities, the immobilized problem of TiO2 must be solved. In recent years, many researchers have been devoted to immobilize TiO2 nanostructures on different substrate materials to hinder agglomeration [20]. For example, Wang et al. have demonstrated carbon fibers (CFs) can be used as substrate for growing TiO2 nanosheets because CFs are flexible, conductive, and stable in corrosive conditions [21]. In addition, synergistic effect between TiO2 and carbon could greatly retard the recombination of photo-induced electrons and holes, leading to the enhanced photocatalytic performance of TiO2 [22]. To utilize this synergistic effect, many related hybrid structures have been investigated, such as graphene/TiO2 composites, [23] carbon nanotube (CNT)–TiO2 composites, [24] TiO2/AC composites (TiO2-loaded on activated carbon) [25, 26], and so on [27–29]. In view of the high cost of graphene and CNT, carbon fiber as an industrial product is an excellent candidate material to support the catalysts. But CF has lower surface area than graphene, CNT and AC due to no significant nanostructures on its surface, which restricts its potential as catalyst supporter.

In this paper, we report a facile preparation method of TiO2/porous CFs composite based on self-template method. The highlights of this work are that the CFs would be etched into the porous structure by the oxygen in air at high temperature when the TiO2 particles were loaded on its surface. This TiO2/porous CFs composite shows unique characteristics such as excellent adsorption and high photocatalytic activities. These findings afford the convenient optimization for the design of TiO2-based photocatalytic device, which would achieve the application of the photocatalytic degradation of the organic pollutants in wastewater.

Experimental section

Synthesis of samples

All reagents were of analytical grade and used without further purification. The carbon fibers calcined at 1600 °C in argon atmosphere (Blue Star (Group) Co., Ltd., Lanzhou) were used in our experiment and the length of the CFs was unified to about 1 mm. The preparation process of the composite is described as below. P25 powders (1 g, Degussa) and additives (polyethylene glycol or acetic acid) were dissolved into 20 mL deionized water with stirring for 1 h to get a homogeneous P25 solution. Subsequently, 0.5 g prepared CFs were added into the above solution and magnetically stirred for another 1 h. Then the solution was dried and the P25/CFs composite was obtained. Then the P25/CFs composites were annealed with a ramp rate of 2 °C min−1 for 2 h at 400, 450, 500, 550, and 600 °C, respectively.

Characterization of samples

Morphologies were examined by field emission scanning electron microscopy (FESEM, Hitachi S-4800). Raman spectra were recorded on JY-HR800 Raman spectroscopy with a 532-nm laser line from the yttrium aluminum garnet laser. The crystal phases were determined by X-ray diffraction analyzer (XRD, PANalytical X’Pert) with Cu-Kα radiation (λ = 1.5418 Å). X-ray photoelectron spectra (XPS) were recorded with a Kratos AXIS Ultra DLD XPS instrument equipped with an Al Kα source at 10−9 Torr. The surface area and porosity of the products were measured by the nitrogen adsorption–desorption isotherm and Barrett–Joyner–Halenda (BJH) methods on a Micrometitics ASAP 2020M accelerated surface area and porosimetry system.

Catalytic activity measurements

The photocatalytic activity was evaluated by the time evolution of photodegradation of rhodamine B (RhB). During the measurement of the photocatalytic performance, the amount of photocatalyst was kept at the same in all tests. 50-mg photocatalyst and 50 mL of RhB solution (10 mg L−1) were placed into a crystallizing dish (100 mL). The suspension was magnetically stirred in dark conditions for 60 min to ensure the adsorption/desorption equilibrium. After the dark treatment, the solution was subsequently irradiated by a 175 W tungsten halogen lamp with magnetic stirring at ambient temperature. During the irradiation, 4 mL of the solution was taken from the reactor at a scheduled interval and filtered with pore size = 1 μm cellulose acetate membrane filters to remove the residual catalyst. The filtrate was analyzed by recording variations at the wavelength of maximal absorption (λ = 554 nm) in the UV–visible spectrum of RhB using a TU-1901 spectrophotometer.

Results and discussion

Morphology, components, and phase analysis

Scheme 1 shows a schematic representation of the formation process of the TiO2/porous CFs composite by the self-template method. The SEM images of the original P25 nanoparticles and CFs are shown in Fig. S1 (Supporting information). It can be seen that the P25 nanoparticles with average diameter of 25 nm are agglomerate in a certain degree, and the CFs have uniform size of 8 μm in diameter and smooth surface. Figure 1 shows the SEM images of the CFs with the P25 particles loaded on its surface. In order to make the P25 particles on the CFs surface more dispersed and reduce the agglomerations, a common dispersant (polyethylene glycol, PEG) was added into the P25 aqueous solution. Figure 1a shows the morphology of the P25/CFs composite with PEG as additive. Huge P25 clusters are formed on the CFs surface because P25 can easily adsorb on PEG molecules and then adhere to CFs, which leads to more serious agglomeration than that without additive (as shown in Fig. 1b). Considering the above problem, acetic acid (CH3COOH) was added into the P25 solution to replace PEG, because of the change of zeta potential on the TiO2 surface in acid solution and the existence of titanium acetate, which can enhance the dispersion of P25 particles in aqueous solution [30, 31]. Figure 1c and d displays the SEM images of P25/CFs composites with CH3COOH as additive. Although few P25 clusters are still present, the distribution of P25 particles is obviously more uniform than those without additive and with PEG as additive. Figure 2 demonstrates the high-resolution SEM images of the P25/CFs composites with PEG and CH3COOH as additives, respectively. Due to the PEG coating, the P25 nanoparticles are not very clear, as shown in Fig. 2a and b, which confirms the inference mentioned above. Compared with the sample using PEG as additive, the P25 particles of the sample using CH3COOH as additive are distributed more uniformly, as shown in Fig. 2c and d. Therefore, a TiO2/CFs composite is obtained with the relatively uniform distribution of P25 particles.

Scheme 1
scheme 1

Schematic representation of the formation process of the TiO2/porous CFs composite by a simple self-template method

Full size image
Fig. 1
figure 1

SEM images of TiO2/CFs composites with different additives were added: a PEG, b no additive, c and d CH3COOH

Full size image
Fig. 2
figure 2

The high-resolution SEM images of TiO2/CFs composites: a and b used PEG as additive, c and d used CH3COOH as additive

Full size image

To further enhance the interaction between P25 and CFs, the as-prepared P25/CFs composites were calcined for 2 h at 400, 450, 500, 550, and 600 °C, respectively. The corresponding samples are labeled as S400, S450, S500, S550, and S600, respectively. For the S400 sample (Fig. 3a), the surface of CFs still remains the original morphology while it becomes smooth after calcination at 450 °C (Fig. 3c). A unique porous structure appears when the calcination temperature reaches 500 °C (Fig. 3e) or higher (550 °C of Fig. 3g, 600 °C of Fig. 3i). The contour of their pores coincides with the contour of P25 clusters, as shown in Fig. S2 (Supporting information). In order to verify the role of TiO2 particles, the CFs without P25 particles attached were also calcined at above mentioned temperatures. From their SEM images (400 °C of Fig. 3b, 450 °C of Fig. 3d, 500 °C of Fig. 3f, 550 °C of Fig. 3h, 600 °C of Fig. 3j), it can be observed that the diameter of the CFs is decreased and few pores appear on the surface as the temperature increases, and only some flat pits are formed when the calcination temperature is 600 °C. Therefore, we can conclude that the TiO2 particles can act as catalyst to enhance the etch reaction between CFs and oxygen. At the same time, it can act as template to determine the shape of the pores. So far, a TiO2/porous CFs composite is obtained, and the formation mechanism will be discussed later.

Fig. 3
figure 3

SEM images of TiO2/CFs composites and CFs after calcined at high temperatures: a, b 400 °C; c, d 450 °C; e, f 500 °C; g, h 550 °C; and i, j 600 °C, respectively

Full size image

The characteristics of the TiO2/porous CFs were further investigated with the N2 adsorption/desorption analysis. The typical result is shown in Fig. S3 (Supporting information), and the relative BET data are shown in Table 1. From Table 1, it can be seen that the special surface area of the original TiO2/CFs composite is about 41.38 m2 g−1. After annealing at 400 and 450 °C, the surface area of the composite dropped obviously because the surface of CFs is etched by the oxygen in air and becomes very smooth, which is corresponding to the SEM images of S400 (Fig. 3a) and S450 (Fig. 3c) samples. But when the temperature reaches 500 °C, the surface area reaches to 57.05 m2 g−1 dramatically, which can be ascribed to the appearance of the pores on the surface of the CFs. However, when the temperature continues to raise, the surface area decreasees inversely, which is caused by the increase of the pore diameter. All these results are consistent to the variation of pore size, as shown in Fig. S3.

Table 1 BET data of all the samples
Full size table

Figure 4a shows the XRD patterns of all the samples. The peaks at 2θ values of 25.3, 37.8, 48.0, 53.9, 55.1, 62.7, 68.8, 70.3, and 75.0° correspond to the crystal planes of (101), (004), (200), (105), (211), (204), (116), (220), and (215) of the anatase phase TiO2 (JCPDS No. 21-1272), respectively. In addition, characteristic diffraction peaks at 27.4, 36.1, 41.2, and 56.7° are also observed, which are attributed to the (110), (101), (111), and (220) planes of the rutile phase TiO2 (JCPDS No. 21-1276). It can be obviously seen that the crystallinity of P25 is improved after calcination treatment. The weight ratio of anatase (f A) in the mixture can be obtained from the Eq. (1): [32]

$$ f_{\text{A}} = \frac{1}{{1 + 1.26\frac{{I_{\text{R}} }}{{I_{\text{A}} }}}} $$
(1)
Fig. 4
figure 4

XRD patterns and Raman spectra of all the P25/CFs composites

Full size image

where I R is the intensity of the strongest rutile reflection, and I A is the intensity of the strongest anatase reflection. For the original P25 particle, f A is 0.75, which agrees well with the characteristic of commercial P25. For the S400, S450, S500, and S600 samples, the f A is 0.64, 0.69, 0.71, and 0.65, respectively, which are all lower than the value of P25, demonstrating that the rutile phase increased after calcination treatment. However, the f A of S550 sample (0.76) is higher than that of P25, implying that the carbon plays an important role in the phase transition process caused by calcination treatment.

Raman spectrum is another useful technique to analyze the change of composition and phase. Here we used the JY-HR800 microdomain Raman spectroscopy to measure the Raman spectra of all the samples. Due to the small spot (diameter less than 1 μm), the test area can be effectively positioned on the P25 clusters, which can achieve the individual observation of the change of TiO2. The characteristic modes of anatase TiO2 are clearly observed in the Raman spectra of all the samples (Fig. 4b). The peaks centered at around 150, 398, 511, and 620 cm−1 are associated with Eg(1), B1g, A1g, and Eg(2) mode of anatase TiO2, respectively [33, 34]. The band at 260 cm−1 is corresponding to the rutile TiO2. Another two weak peaks centered at around 1348 and 1575 cm−1 are consistent with the D and G peaks of the carbon. Based on the above Raman results, it can be confirmed that the phase has little change.

In order to confirm that physical adsorption is not the only action form between TiO2 and CFs, simple ultrasonic cleaning tests were carried out for the S400, S550, and original P25/CFs samples. The corresponding SEM images of the samples before and after test are shown in Fig. 5a, c, e and b, d, f, respectively. After ultrasonic cleaning, only few P25 particles attach on the CFs for the original P25/CFs (Fig. 5b) and S400 (Fig. 5d) samples, while lots of P25 particles still exist on the surface of CFs for the S550 sample (Fig. 5f), especially on the bottom of the pits. To further shed some light onto the surface chemical composition of the samples, the interface between TiO2 and carbon was investigated in detail using XPS measurements. Figure 6 shows the XPS spectra of the S400 and S550 sample. Figure 6a depicts the full survey spectra of the samples. The elements of C, Ti, and O can be clearly identified in the S550 sample but there are no obvious Ti element-related peaks in the S400 sample due to the low amount of TiO2 after ultrasonic cleaning. High-resolution scans show that the exact binding energies of these elements are slightly different, as revealed in panels from b to d. From Fig. 6b, for the S400 sample, there are three major peaks centered at 285, 285.4, and 286 eV, which may be assigned to the 1 s electrons of sp3 hybridized carbon (C–C) and sp2 hybridized carbon (C=C), and the formation of C–OH bond, respectively [35]. For the S550 sample, the peak of C=C is still existed at 285.4 eV, and other two peaks appear at around 283.6 eV and 288.2 eV, which may be related to the formation of Ti–O–C bond and C=O bond. Both the samples have two peaks at 458.4 and 464 eV corresponding to Ti4+ atoms in TiO2 (as shown in Fig. 6c) [36]. For the O 1s electrons (as shown in Fig. 6d), both the samples exhibit a peak at around 532 eV corresponding to the hydroxyl O on the TiO2 surface while the S550 sample has a another strong peak at 529.6 eV, which is assigned to oxygen bonded to Ti [37].

Fig. 5
figure 5

The SEM images of TiO2/CFs composites calcined at different temperatures before and after ultrasonic cleaning: a, b unannealed sample; c, d sample S400; and e, f sample S550, respectively

Full size image
Fig. 6
figure 6

a XPS survey spectra of TiO2/CFs composites calcined at 400 and 550 °C, and high-resolution scans of the binding energies of b C 1s, c Ti 2p, d O 1s electrons in the composites

Full size image

From the above results, the formation mechanism of TiO2/porous CFs composite can be deduced. The oxygen molecules in air would be activated on the surface of TiO2 particles at relatively low temperature because the oxygen vacancy can trap an electron, which has been reported in many literatures [38, 39]. The activated oxygen molecules react with carbon and TiO2 leading to the formation of C–O bond and Ti–O bond. Then the C–O bond would further react with the activated oxygen molecules to form the C=O bond as the calcination temperature increases, which can be observed from the changes of XPS spectra. Finally, the carbon dioxide forms and escapes from the interface, as schematically illustrated in Scheme 2. Because the amount of activated oxygen formed at the interface between TiO2 and CF is larger than other places on the CFs, the etching rate of carbon attached with TiO2 is increased significantly, leading to the formation of the pores. Naturally, the pore size is determined by the size of the TiO2 clusters. In the formation process, the TiO2 clusters act as either catalyst or template. In summary, we prepared a TiO2/porous CFs composite by a simple self-template method.

Scheme 2
scheme 2

A schematic showing the etching process of the CFs at high temperature with TiO2 particles as catalyst

Full size image

Photocatalytic activity

Photocatalytic activities of the TiO2/porous CFs composites were evaluated by the photo-degradation of RhB aqueous solutions. The dye adsorption rates of all the samples are displayed in Fig. 7. It can be observed that the amount of residual RhB in the sample solutions after dark treatment for 60 min decreases with the increase of the calcination temperature, except the S600 sample. This result reflects that the porous structure could improve the adsorption ability of the TiO2/CFs composites. The time evolution of the dye concentration in different catalysts solution is shown in Fig. 8a, where the entire reaction time does not include the time for adsorption equilibrium. The RhB concentration (%) after several intervals of time could be estimated using the Eq. (2):

$$ \% {\text{ RhB concentration}} = \left( {C/C_{0} } \right) \times 100\% $$
(2)
Fig. 7
figure 7

The dye adsorption rate of all the TiO2/CFs composites

Full size image
Fig. 8
figure 8

a Photocatalytic degradation of rhodamine B with TiO2/CFs composites as photocatalysts under UV light and b kinetics of RhB decolourisation with different photocatalysts

Full size image

where C 0 is the initial RhB absorbance at 554 nm and C is the absorbance after several intervals of time (t). As shown in Fig. 8a, the removal rate of dye is enhanced after the calcination of TiO2/CFs composites. Due to the effect of adsorption, the amount of residual dye in solution cannot be regarded as a direct evidence to demonstrate the catalytic efficiency. The apparent rate constant can be chosen as the basic kinetic parameter for the comparison of different photocatalysts’ photocatalytic activities. The obtained apparent rate constants of different samples are 0.05076, 0.06038, 0.06542, 0.07765, 0.04735, and 0.06296 min−1 for the P25/CFs, S400, S450, S500, S550, and S600 samples, respectively (as shown in Fig. 8b). It can be seen that all the samples show enhanced activity compared to pure P25 (0.03302 min−1, as shown in Fig. S4 (Supporting information)). The enhancement can be ascribed to the synergistic effect between TiO2 and carbon, which can hinder the recombination of photo-generated electrons and holes. Scheme 3 illustrates the process diagram of RhB degradation caused by the TiO2/porous CFs composites. Theoretically, the dye in solution cannot sufficiently contact with P25 particles due to their immobilization on the CFs surface, and the photocatalytic efficiency would be seriously weakened. But in the experiment, the dye would be adsorbed on the surface of the composite due to the existence of the porous structure, which can increase the encounter rate of dye and hydroxyl radicals (•OH). But too much adsorbed dye would cause the shielding effect and decrease the light absorption of TiO2 [40]. This may be the major reason that the S550 sample has the worst catalytic activity. The catalytic activity of the S600 sample may be influenced by the disappearance of the porous structure caused by the excessive etching, as shown in Fig. S5 (Supporting information). The role of CFs in the photocatalysis maybe is as not only a substrate to support the TiO2 particles, but also an electron collector to retard the recombination of photo-induced charge carries and prolong electron lifetime.

Scheme 3
scheme 3

Schematic illustrations of the synergistic effect in the photocatalysts of TiO2/porous CFs composites due to the unique structure of porous CFs

Full size image

Conclusions

In summary, TiO2/porous CFs composites were fabricated by a simple self-template method. During the calcinations process, the P25 clusters as catalyst can enhance the etch reaction rate between carbon and oxygen. It can also be considered as a template, which would determine the size and sites of the pores. Due to the unique porous structure, the TiO2/porous CFs composites exhibit excellent adsorption ability, and enhanced photocatalytic activity. But too much adsorbed dye also reduces its activity because of the shielding effect. Combining the advantages of TiO2 and CFs, the composite can be applied in many fields, such as water splitting, lithium ion batteries, especially the application of the photocatalytic degradation of organic pollutants.