1 Introduction

The global environment has been polluted by harmful substances of many kinds including volatile organic compounds (VOCs), causing damage to human health. It is extremely important to remove these harmful compounds and to purify the living environment. Purification by photocatalysis is an effective method to decompose harmful airborne organic contaminants [1, 2]. The decomposition mechanism is the following: when light is irradiated on a photocatalyst and the band gap is smaller than the light energy, the photocatalyst absorbs the light. Carriers (electrons and holes) are generated. The holes usually have strong oxidizing capability [3] to mineralize organic compounds, leading to production of much cleaner air.

Anatase-type TiO2 photocatalyst with low cost has been used for outdoor applications [4,5,6]. The semiconductor works as a photocatalyst under irradiation of UV light, a minor component of solar light. However, TiO2 is not useful for indoor applications because TiO2 can absorb only in the UV region. Furthermore, indoor lighting has very weak UV illumination. Visible-light-sensitive photocatalyst with high activity is promising for indoor air cleaning because indoor lighting can emit intense visible light. Consequently, the development of novel visible-light-sensitive photocatalysts and low-cost earth-abundant photocatalysts has been anticipated and explored [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24].

Yellow semiconductor WO3, with a band gap of 2.6–2.8 eV, is recognized as a promising candidate among visible-light-sensitive photocatalysts. Its potential at the bottom of the conduction band (CB) is more positive than that for one electron oxygen reduction [25]. When WO3 is used as a photocatalyst, its photogenerated electrons cannot be consumed well. Pure WO3 is deactivated easily. Research related to loading of cocatalysts such as Pt, PtPb, Cu ion, and NaOH on WO3 has been undertaken to promote electron consumption and to increase its photocatalytic activity [25,26,27,28,29,30,31,32]. However, Pt is a precious metal that cannot be used sufficiently because of its high cost and likely exhaustion of the Pt resources. Other reported cocatalysts are harmful, or the loading process is complex. A safe and abundant cocatalyst is still needed. Therefore, in this study, we tried to develop a novel mixture photocatalyst consisting of WO3 and ubiquitous cocatalyst; we also sought a simple method to prepare the mixture.

Apparently, soil is an abundant and ubiquitous material that might be a promising cocatalyst. From results of preliminary investigation, we selected red clay among soils. Results showed that red clay is a suitable cocatalyst. Moreover, red-clay-loaded WO3 exhibits remarkable photocatalytic activity.

2 Experimental

2.1 Material preparation

Red-clay ball (Akadama soil; Tachikawa Heiwa Nouen Co. Ltd., Japan) preheated to 673 K was used for this study. The clay was crushed using a mortar and pestle to prepare fine particles. The crashed material and WO3 were mixed with proper mixing ratios using a mortar. Samples were obtained after the mixtures were dried at 343 K for 5 h. The weight ratios for the red clay and WO3 were, respectively, 1:100, 5:100, 10:100, and 50:100. Furthermore, percentage of loading weight in this study was obtained by dividing cocatalyst weight by WO3 weight and expressed as the divided amount per one hundred. For example, 1 wt% red clay-loaded WO3 was the sample with the ratio (1:100).

2.2 Characterization

Samples were characterized using several analytical devices. Crystal structures were measured using an X-ray diffractometer (XRD, X’pert Pro; PANalytical B.V., Netherlands) with Cu Kα radiation. Optical absorption spectra were evaluated using a UV–Vis spectrophotometer (UV-2500PC; Shimadzu Corp., Japan). Reflectance spectra were first measured using BaSO4 as a reference. Then, the obtained data were converted into optical absorption spectra using Kubelka–Munk theory. Surface chemical states of the samples were measured using an X-ray photoelectron spectroscope (XPS, AXIS-HS; Kratos Analytical Ltd., UK) with monochromatic Al Kα radiation. Binding energy in the XPS spectra was calibrated using C 1s peak, of which the binding energy is 284.5 eV. The specific surface area was evaluated at 77 K using a surface analyzer (Gemini 2360; Micromeritics Co., USA) with Brunauer–Emmett–Teller (BET) method. The clay chemical composition was estimated using an inductively coupled plasma optical emission spectroscope (ICP-OES, ICPS-8100; Shimadzu Corp.).

Photocatalytic activity was evaluated at room temperature in a 500-mL cylindrical glass reactor [33]. Photocatalytic decomposition of gaseous 2-propanol (IPA) into CO2 was selected as a model reaction because IPA is often used as a model of an organic gas in this field [34]. Details of the photocatalytic evaluation procedures were the following: We spread powder photocatalyst with weight of 0.4 g uniformly on a Petri dish with 8.5 cm2 base area. Then, the dish was set on the center of the base in the reactor. After the inside atmosphere was replaced with pure air, concentrated IPA gas was injected to produce concentrations of IPA in the reactor of about 600–800 ppm. The reactor was kept in the dark until the adsorption–desorption equilibrium state was confirmed. Visible light (400 < λ < 530 nm) was irradiated using 300 W of Xe lamp equipped with Y-44, B390, HA-30 filters, and a water filter. The IPA concentration was measured using a gas chromatograph (GC-14B; Shimadzu Corp., Japan) with a flame-ionized detector (FID). The CO2 concentration was estimated using the chromatograph (GC-14B) with FID and a methanizer (TN-1; Shimadzu Corp.). The light intensity of the visible-light irradiation, set to about 1 mW cm−2, was measured using a spectroradiometer (UV-40; Ushio Inc., Japan). Maximum CO2 evolution rate was evaluated from the rate of zero order reaction [r = k (reaction rate constant), and C = k t] in the range where CO2 concentration linearly increased against time.

Reduction of O2 into H2O2 might occur during the photocatalytic IPA oxidation. The H2O2 generation on the sample was evaluated qualitatively using dimethylphenanthroline (DMP) method [35]. After the reactor was kept in the dark for 10 min after visible-light irradiation, the sample was washed with distilled water. An aqueous solution possibly containing H2O2 can be prepared. Phosphate, DMP, and CuSO4 solutions were added to the aqueous solution. The color turns yellow if the solution contains H2O2. Its absorbance at 454 nm in wavelength was measured using UV–Vis to calculate the H2O2 amount.

3 Results and discussion

3.1 Crystal structure

Red clay has been used since ancient times for many applications such as bricks and horticulture soils. The red-clay ball (akadama) selected for this study is a volcanic product produced in eastern Japan. The clay is reportedly composed of quartz (SiO2), diaspore (AlOOH), hematite (Fe2O3), and goethite (FeOOH) [36, 37]. The clay components were ascertained using XRD in reference to these data.

Figures S1 and 1 depict XRD patterns of the red clay, WO3, and the mixture. Results confirmed that the clay consisted mainly of quartz (PDF No. 33-1161). Also, diaspore (PDF No. 5-355), hematite (PDF No. 33-664), and goethite (PDF No. 29-713) might be present in the clay. The XRD data in Fig. 1 imply that the WO3 was a well crystallized oxide with a monoclinic structure (PDF No. 43-1035) [38]. The XRD pattern of the mixture was similar with that of the WO3, suggesting that crystal structure of the WO3 was not changed by the loading.

Fig. 1
figure 1

XRD patterns of WO3 and a mixture consisting of red clay (10%) and WO3

Full size image

3.2 Chemical composition

The red clay contains silicon (Si) and metals of many kinds. The atomic ratio in the clay was evaluated using ICP. The evaluated ratios of Si to principal component metals are presented in Table 1. Results show that the clay was composed mainly of Si, Al, and Fe. These metal elements are abundant principal components in the crust of the Earth. Furthermore, these three metal-containing compounds were the main components of the chemical compounds including quartz and diaspore, which were detected using XRD (Figure S1). Accordingly, the ICP data support the XRD results. The clay also contained some alkali (K) and alkali earth (Mg) metals.

Table 1 Weight ratios of metals to Si in red-clay ball
Full size table

3.3 XPS analysis

The red clay mainly contains Si, Al and Fe, as described above. The valence states of Si and Al were expected to be +4 and +3, respectively. Two kinds of valence states of Fe ions such as Fe2+, and Fe3+ might exist in the clay. The valence state was assessed carefully using XPS. Figure 2 exhibits the XPS spectra of Fe 2p in the clay and Fe2O3, which was used as a reference sample. Two peaks are apparent at the binding energy of 711 eV and 725 eV, respectively, corresponding to Fe 2P3/2 and Fe 2P1/2 peaks. Yamashita et al. [39, 40] reported that the peaks centered at 711 eV and 709 eV originated, respectively, from Fe3+ and Fe2+. We infer that the clay contained Fe3+ ion (Figure S2). This result was also evidenced by the fact that XPS data of Fe2O3 resembled those of the clay.

Fig. 2
figure 2

XPS spectra of a red clay and b Fe2O3. The arrow indicates the satellite peak of Fe 2P3/2

Full size image

3.4 Optical absorption spectra

The visible-light absorption property was evaluated using UV–Vis spectrophotometry. Figure 3 depicts the optical absorption spectra of WO3, the clay, and the mixtures. WO3 can absorb light partially in visible light region. Its absorption threshold is about 480 nm. The red clay, which was light brown, can absorb light more widely. Its weak absorption tail extends beyond about 600 nm. Visible-light absorption between 400 and 600 nm might derive dominantly from Fe3+-containing oxides in the clay [41].

Fig. 3
figure 3

Optical absorption spectra of red clay, WO3, and their mixture (10%)

Full size image

The mixture can absorb visible light well, but its onset absorption edge was blue-shifted by 10 nm compared with that of pure WO3. Similar shifts were observed in other mixtures of NaOH-loaded WO3 and NaBiO3-loaded WO3, suggesting that it was not relevant to species of the loaded cocatalysts. Absorption spectra are reported to be affected by many factors including size [42], crystallinity [42], morphology [43], surface functional groups [30, 44], and grain boundary [45]. Because its size, crystallinity, and morphology did not change by mixing, this blue shift might be related mainly to changes in the amounts of surface functional groups [30, 44] and grain boundary of WO3. The grain boundary may change because of stress from the mixing on a mortar.

3.5 Photocatalytic oxidation property

In photocatalytic oxidation of gaseous IPA, acetone is first generated as an intermediate in the gas phase [33]. Then, acetone is oxidized further into CO2 via carboxylic acids such as formate (–COOH) and acetic acid (Figure S3) [46]. Organic compounds of many kinds might be generated as intermediates. The final product (CO2) was monitored carefully for this evaluation of photocatalytic activity.

Figure 4 presents the irradiation time dependence of variation of CO2 evolution over WO3 or the mixture under visible-light irradiation. Carbon dioxide was rarely evolved on pure WO3 with less than 200 h of irradiation. In contrast, the mixture consisting of WO3 and the clay (10%) showed remarkable photocatalytic activity. Most of the IPA was oxidized into the final product, CO2 within 34 h (Fig. 4). The quantitatively estimated maximum rate (zero order reaction rate: about 130 ppm/h, which was calculated from CO2 concentrations between about 10 h and 23 h) of CO2 evolution for this sample, which was not affected by diffusion limitation, was much higher than that for pure WO3 (0.24 ppm/h, calculated from the concentrations between about 110 h and 170 h). Moreover, the mixture (10%) showed over twice higher activity than the reported high-activity photocatalyst Pt-loaded WO3 [27] (62 ppm/h, calculated from the concentrations between about 11 h and 24 h), as shown in Figure S4. This result means that the photocatalytic activity can be enhanced sufficiently without using expensive chemicals such as those including precious metals.

Fig. 4
figure 4

Irradiation time dependence of changes of CO2 generation from IPA decomposition under visible-light irradiation in the presence of WO3, red clay, and the mixtures: a WO3, b 1 wt% of red-clay-loaded WO3, c mixture(clay 5 wt%), d mixture (10 wt%), e mixture (50 wt%) and f red clay

Full size image

We investigated the relation between activity and loading amount of the cocatalyst (clay) to ascertain an optimal amount of the clay loading. Figure 5 exhibits the amount of photocatalytic CO2 evolution for every sample after about 21 h of visible-light irradiation. The photocatalytic activity increased concomitantly with increasing amounts of the clay between 0 and 10 wt% loading. This tendency might be attributed to cocatalyst loading effects, as described in the next section. In contrast, a wider range of the WO3 surface was covered with the clay. The WO3 absorbed the irradiated light less when the mixture contained increased loading amounts, especially exceeding approx. 10%. The activity decreased on the mixture loaded by over 10 wt% of the clay with increasing loading amount (Fig. 5). Results show that the highest activity was exhibited by 10 wt% of the clay-loaded WO3 because of competition between the cocatalyst effect and absorption effect.

Fig. 5
figure 5

Amount of CO2 evolution for photocatalytic IPA decomposition at about 21 h after visible-light radiation: a WO3, b 1 wt% of red-clay-loaded WO3, c mixture (5 wt%), d mixture (10 wt%), e mixture (50 wt%), and f red clay

Full size image

3.6 Mechanism

IPA was oxidized mainly by photogenerated holes. Electrons might contribute importantly to the oxidation. In this section, we specifically examine the role of photogenerated electrons, which were first consumed by reduction of O2 into H2O2 (Reaction (1)) for pure WO3. The H2O2 reportedly decomposed over some time, eventually saturating on the WO3 surface, leading to deactivation and lower activity of WO3 [30]. It is crucially important to promote electron consumption, including H2O2 decomposition, for the maintenance of WO3 photocatalytic activity. H+ in Reaction (1) derived from water adsorbed on the photocatalyst.

$${\text{O}}_{2} + 2{\text{H}}^{ + } + 2{\text{e}}^{ - } = {\text{H}}_{2} {\text{O}}_{2}$$
(1)

Figure 6 shows that the amount of H2O2 formation was measured experimentally on pure WO3 and the mixture during IPA oxidation. The amount of H2O2 first increased gradually and was then saturated on pure WO3, indicating almost total deactivation within 6 h. However, H2O2 on the mixture was generated much less than that on pure WO3. Photogenerated electrons were consumed using the same reduction process (Reaction (1)), even for the mixture. Therefore, it seems that H2O2 was consumed much more quickly on the mixture photocatalyst.

Fig. 6
figure 6

Effect of irradiation time on H2O2 generation for WO3 and the mixture (10%)

Full size image

The chemical composition of the clay was checked again to elucidate this superior H2O2 consumption property. The clay contained relatively large amounts of highly dispersed Fe3+ (Table 1). Apparently, Fe in the clay can play an important role in H2O2 consumption. These three reactions (2)–(5) might occur in the consumption process. In reaction (2), part of Fe3+ was reduced into Fe2+ by photolysis [47]. Furthermore, reaction (3) might occur. Light was absorbed by complexes consisting of Fe3+ and carbonic acids (ligand), which were intermediates in the photocatalytic IPA oxidation; Fe3+ was reduced into Fe2+ by this light absorption [48]. Except for the above reactions, Fe3+ might be directly reduced into Fe2+ by photogenerated electrons (Reaction (4)) [49]. These formed Fe2+ reacted further with H2O2 in reaction (5) (Fenton reaction), resulting in the promotion of H2O2 consumption on the mixture [48] (Fig. 7). Radicals of •OH might be generated and be used for IPA decomposition during this H2O2 consumption. Consequently, photogenerated electrons can be consumed easily by three reactions. Furthermore, recombination between electrons and holes can decrease, leading to high activity of the mixture.

$${\text{Fe}}^{3 + } + {\text{H}}_{2} {\text{O}} + h\nu \left( {\lambda < 600\,{\text{nm}}} \right) = {\text{Fe}}^{2 + } + {\text{H}}^{ + } + \cdot {\text{OH}}$$
(2)
$$\left[ {{\text{Fe}}^{3 + } {\text{L}}_{\text{n}} } \right] + h\nu = \left[ {{\text{Fe}}^{2 + } {\text{L}}_{{{\text{n}} - 1}} } \right] + \cdot {\text{L}}$$
(3)
$${\text{Fe}}^{3 + } + e^{ - } \left( {photogenerated} \right) = {\text{Fe}}^{2 + }$$
(4)
$${\text{Fe}}^{2 + } + {\text{H}}_{2} {\text{O}}_{2} = {\text{Fe}}^{3 + } + {\text{OH}}^{ - } + \cdot {\text{OH}}$$
(5)

Here, L in Eq. (3) represents organic ligands such as –OCH3, –COOH, and –CH2COOH.

Fig. 7
figure 7

Schematic illustration of photocatalytic 2-propanol (IPA) oxidation mechanism for red-clay-loaded WO3

Full size image

Photoluminescence (PL) data were measured (Fig. 8) because weaker intensity of PL spectra from band-to-band transition indicates better separation of photogenerated charges (holes and electrons) [50]. The PL spectrum due to band-to-band transition for WO3 appears at around 450–480 nm of wavelength [51]. The PL intensity around at around 450–480 nm for the clay-loaded WO3 was weaker than that for pure WO3, indicating that charge separation was promoted by addition of the clay. This result may be explained as follows: On pure WO3, photogenerated electrons were consumed by oxygen reduction into H2O2. However, H2O2 was accumulated gradually on the WO3 surface, leading to suppress of H2O2 formation and electron consumption. On the other hand, on the red-clay-loaded WO3, H2O2 was decomposed by Fe2+, leading to less inhibition of further O2 reduction (H2O2 formation) and electron consumption. As a result, recombination between electrons and holes was more difficult to occur on the mixture.

Fig. 8
figure 8

Photoluminescence (PL) spectra of WO3 and the mixture (10%) (excitation wavelength: 310 nm)

Full size image

To confirm the Fe effect further, photocatalytic activity for 5 wt% of Fe3+-doped, zeolite-loaded WO3 was evaluated (Figure S4). The activity for the Fe-doped-zeolite-loaded WO3 also showed relatively high activity. These results showed that loading of a Fe3+-containing compound gave a positive effect for WO3 photocatalysis.

We examined the effects of high surface area for the cocatalyst because a larger surface area often positively affects photocatalytic activity. The BET surface areas are presented in Table 2. The surface areas of the WO3 and the mixture (10% of cocatalyst loading) were estimated, respectively, as 5.8 and 12 m2/g. The area of the mixture was about twice that of the WO3, which suggests that high surface area of the cocatalyst also contributed to the high photocatalytic activity of the mixtures. However, the effect of larger surface area on the activity’s difference is very limited (Figure S5) because the surface area was just twice higher.

Table 2 Surface area of the samples
Full size table

4 Conclusion

Mixtures of WO3 and red clay were prepared using a simple kneading method. The samples absorbed visible light well and showed photocatalytic activity under visible-light irradiation. The photocatalytic activity was evaluated from IPA decomposition into CO2. Relation between photocatalytic activity and mixing ratios (red clay to WO3) exhibited that 10 wt% of red-clay-loaded WO3 showed the highest activity. Furthermore, the activity of the mixture was higher than that of Pt-loaded WO3 with relatively high activity. This higher activity of the mixture should derive from good H2O2 consumption efficiency of the red clay. H2O2 that was generated from photocatalytic reduction of O2 on WO3 might accumulate on the WO3 surface and suppress WO3 photocatalytic activity.

Salient benefits of this mixture photocatalyst are its high activity and its usage of ubiquitous red clay as a cocatalyst. The clay is remarkably cheap and abundant. The mixture photocatalysts and preparation process are eco-friendly because no wastes, even wastewater, are produced during its preparation.