3.1 Morphological, Elemental and Mapping analysis
SEM analysis is a powerful tool to study surface morphological features of the photocatalyst. Figure 1(a-d) represents the lower and higher magnification SEM micrographs of the as-collected uncalcined and calcined iron rust, respectively. After comparing the morphologies, it is clear that both samples are showing significant agglomerations, with highly irregular morphologies. Moreover, SEM micrographs indicate that the individual particle sizes in both cases are in the nanometer range, which agglomerated into large blocks of irregular shape particles. We assume that other parameters could promote the photo- and photoelectrocatalytic performance of calcined Rust compare to uncalcined Rust, under similar conditions. Therefore, we have characterized these samples further with XRD, BET, and TGA to confirm those parameters, which are contributing to enhanced catalytic performance.
Figures 2 and 3 illustrated the elemental analysis and mapping study of uncalcined and calcined Rust, respectively. Figures 2a show the EDX spectra of uncalcined Rust with %elemental composition in tabular form. The representing peaks for Fe, C, Zn, Si, Al, and O are observed with respective %compositions of each element. As expected, the %compositions indicate that the Fe and O are present in higher concentration compared to Zn, Si, and Al, which are existed in minuscule quantities. The presence of Zn, Si, and Alelements can be justified as the Rust was collected from a steel source, that contains these metals in proportions (to enhance their strength). Figure 2b-d presented the mapping results of Fe, O, and Si, respectively. The mapping study revealed the presence of Si, which is due to earthy impurities as Rust is obtained from the ground water pipes. Figure 3 represents the EDX spectra and % elemental composition (tabular form) of calcined iron rust, displaying peaks for C, Zn, K, Si, Ca, Al and O. The percentage of Fe and O were found higher than other elements. Furthermore, the percentage of oxygen in the case of calcined Rust is lower than uncalcined Rust NPs, which might be due to the elimination of any attached water and other volatile matter. The quantity of C is also reduced after calcination due to its oxidation to CO2. Moreover, n the case of calcined Rust, we have observed K and Ca in trace amounts, which could be possibly added from the ceramics where we have calcined the Rust at a high temperature, i.e. 700 oC. Figure 3b-d presented the mapping results of Fe, O, and Si, respectively. It is essential to state that we deliberately not removed the various elemental impurities to see their overall effect on photocatalysis and also to look after the high cost.
3.2 Structural analysis of uncalcined and calcined Rust
FT-IR spectroscopy is known for its high sensitivity, especially in the detection of inorganic and organic species with low content. Figure 4 shows the FT-IR spectra of uncalcined and calcined Rust representing various peaks in different regions. The uncalcined Rust NPs show peaks at about 1033cm− 1, which might be due to the lepidocrocite (ferric mineral) (Veneranda et al. 2018).The peaks at about 903 and 799 cm− 1 might be due to the goethite (hydrated iron oxide) (S and A 1993). The absorption peak in the absorption band of uncalcined Rust at 1630 cm− 1 is the characteristic peak of -OH, indicating a large amount of crystal water in the corrosion products(Wang et al. 2020). This peak almost disappears in the calcined Rust due to high-temperature calcination. Similarly, the calcined iron rust NPs also gives peaks at about 1033, and 903 and 799 cm− 1,which might be due to the lepidocrocite and goethite, respectively. The spectrum of calcined iron rust also presented prominent peaks at about 526 and 436 cm− 1 is due to the stretching vibration of Fe-O (Rahim et al. 2011).
The chemical composition, phase purity, and crystallinity of the uncalcined and calcined were identified via XRD analysis. The XRD patterns of the uncalcined and calcined iron rust are shown in Fig. 5a. The XRD pattern of the uncalcined rust clearly indicates that this contains a mixture of FeOOH and iron oxide (γ-Fe2O3). On the other hand, the XRD pattern of the calcined rust shows diffraction peaks balanced at 30.18°, 57.28°, and 62.94° corresponding to the (220), (511) and (440) planes of γ-Fe2O3 iron (JCPDS file nos. 75 − 1594). The peaks located at 24.26°, 33.17°, 40.82°, and 49.57° match well with the (012), (104), (113), and (024) planes of α-Fe2O3 (JCPDS file no. 33 − 0664). This means that calcined rust is the combination of γ-Fe2O3 and α-Fe2O3 iron. The quantification of XRD patterns indicates that the sample contains ~ 75% α-Fe2O3 and ~ 25% γ-Fe2O3. The crystallite particle size is also calculated from XRD patterns bya widely reported Debye-Scherrer equation as below (Salavati-Niasari et al. 2008; Khan et al. 2017).
(c)
Where D is showing the crystallite size (nm), K is crystallite shape factor, approximated to be 0.9, λ is x-ray wavelength, β is full width at half the maximum (FWHM) in radians and θ is Braggs angle. By averaging the values of all peaks, the crystallite size of the calcined rust was about 374.5 nm, which is less than the uncalcined sample. The calcination temperature is efficient for the phase transformation of thermodynamically unstable iron oxides to Fe2O3 phase by losing water molecule (Sarif et al. 2020; Ashraf et al. 2020; Gurav et al. 2020b). To further see the effect of calcination on the particle size we also performed the DLS analysis. The DLS particle size distribution of uncalcined and calcined rust is represented in Fig. 5b. The uncalcined rust has a broad peak representing its agglomeration and the average particle size of uncalcined was found to be as high as ~ 487 nm. Comparatively, the calcine rust sample demonstrates a significantly sharp peak area, which indicates a sharp particle distribution. The calculated average particle size of the calcined sample is measured to be ~ 379 nm. These XRD and DLS results indicate that the calcination reduced the particle size significantly, which can add incremental characteristics to the photocatalytic behavior of rust. Dehydration and phase conversion due to calcination at high temperature cause a decrease in attractive forces thus the dispersed and displayed small particle size than uncalcined rust. Thermogravimetric analysis (TGA) measurements of both samples were carried out to find their thermal stability and the results are represented in Fig. 5c. The isotherms indicate the uncalcined iron rust lost about 21 % weight loss at a maximum temperature of 800°C. The attached and adsorbed water is removed at above 120°C. Maximum weight loss occurs between 200 to 300 °, however, no significant weight loss is observed above 500°C. In the case of calcined rust, only 2 % weight was lost during its thermal analysis, which is showing the thermostable nature of calcined samples as their calcination is already performed at a high temperature of 700°C. Finally, the effect of calcination over the surface area is also performed through the Brunauer-Emmett-Teller (BET) adsorption method with nitrogen gas (Fig. 5d). It can be seen in Fig. 5d, that both isotherms are of type IV. The calcination increased the specific surface area (SSA) of calcined rust to 98.84 m2/g, which was 29.30 for an uncalcined sample. The three times enhancement in the SSA indicated that after calcination the crystalization and phase transformation tool place, which also reduces the particle size as indicated by the XRD and DLS results.
3.3 Photodegradation of Methylene blue
The obtained Fe2O3·nH2O and prepared Fe2O3 NPs were utilized as photocatalysts for the PD of MB dye in an aqueous medium under visible light. Figure 6a represents the UV-Vis spectra of MB dye before reaction and after different visible light reaction time in the presence of calcined Fe2O3 NPs, displaying that dye degradation increase with increasing irradiation time. Figure 6b shows the %degradation of MB dye in the presence of Fe2O3·nH2O, Fe2O3 NPs, and without photocatalysts (photolysis). Without photocatalyst, and in the presence of Fe2O3·nH2O no considerable degradation was observed while Fe2O3 NPs significantly degraded MB dye in a short time. The uncalcined Rust have little adsorption capability and thus display less photocatalytic efficiency. The increased photocatalytic efficiency of Fe2O3 NPs than Fe2O3·nH2O is due to its high-temperature calcination, which makes the materials porous and ultimately adsorb the dye efficiently followed by drastic PD. It was reported that calcination enhances catalysts activity(Al-Fatesh and Fakeeha 2012). The result shows that Fe2O3 NPs rapidly degraded 82.5% dye within 1 min and then beyond this time a small increase observed in the efficiency of PD, and finally degraded about 94% dye within 11 min. Initially, the degradation efficiency is much faster and slows down upon increasing the irradiation time. This is because at the beginning formation of •OH radicals is faster and the more availability of the active site for dye adsorption. After a particular reaction time, the remaining active site is not easy to fill because of the repulsive force between the molecules on the surface with the bulk phase, and thus after sometime the %degradation tends to be constant (Aziztyana et al. 2019). The sustainability of Fe2O3 NPs is evaluated by utilizing recovered and re-recovered Fe2O3 NPs under the same experimental conditions. The photocatalyst was recovered by washing with distilled water and oven drying at 100°C. Figure 6c represents the comparison of %degradation of MB dye PD by fresh, recovered, and re-recovered Fe2O3 NPs. The results revealed that the fresh Fe2O3 NPs (1st run) degraded about 94% MB dye within 11 min, while the recovered (2nd run) and re-recovered (3rd ) run degraded about 71.7% and 62% dye, respectively within the same irradiation time. The decrease observed in the photocatalytic activity of the recycled photocatalysts might be due to the blockage of active surface sites by deposition of the photo insensitive hydroxides (Khan et al. 2019).
The photodegradation of MB dye occurs as visible light adsorbed by Fe2O3 NPs that result in excitation of electrons (e−) from valence band (VB) to the conduction band (CB), creatingpositively charged holes (h+) in the valence band. The hole in the VB reacts with H2O molecules and produces hydroxyl radicals (•OH), while the e− present in the CB reacts with an oxygen molecule and produces superoxide anion radical (•O2−). These generated radicals are highly reactive and degraded MB dye molecules into intermediates products and finally into more unaffected species such as CO2 and H2O as shown in Fig. 7. The possible reaction steps in this mechanism are summarized in the following equations (Khan et al. 2019).
NPs → NPs (e− + h+) (I)
O2 + e− → •O2− (II)
H2O/OH− + h+ → •OH (III)
Dye + •OH + •O2−→ degraded products (IV)
The effect of pH on the PD of MB was also evaluated as various discharge their effluents at different pH. Figure 8a shows the effect of pH of the medium on the photodegradation efficiency of MB dye keeping irradiation time (1 min) and initial dye concentration (15 ppm) constant. It was observed that the efficiency of photodegradation of MB is much faster in the basic medium as compared to the acidic medium. It might be due that in the basic medium the photocatalysts tend to acquire a negative charge that results in increased adsorption of positively charged MB (cationic) dyes because of the rising electrostatic attraction (Nguyen Thi Thu et al. 2016). The Figure revealed that from acidic medium to pH 8 the degradation efficiency is almost constant and then enormously increased at pH 9. The results confirmed that at pH 3 the Fe2O3 NPs degraded about 54% dye which increases to 95% by increasing the pH of the medium to 9. Similarly, the effect of photocatalyst dosage and initial dye concentration was also evaluated, and the results are represented in Figs. 8b,c, respectively. Figure 8b represents that efficiency of photodegradation increases rapidly by increasing Fe2O3 NPs dosage up to a specific limit (optimum amount) and then level off beyond that limit. The results verified that till 0.02g photocatalyst dose, the PD is rapidly increasing and a maximum of 82.5% MB dye degraded in 1 mi. However, further dose increments showed very little enhancement in the PD performance of the photocatalyst. At 0.035g catalyst dose, the PD efficiency increases to ~ 93% within the same time. The leveling in PD with increasing catalyst dosage might be the increase in solution opacity which decreases the penetration of the photon flux and ultimately decreases the photocatalytic degradation efficiency (Khan et al. 2020). The effect of initial MB dye concentration is consolidated in Fig. 8c, displaying that maximum PD is achieved at a lower concentration. According to the results, a maximum PD of 92% occurs at an initial concentration of 5 ppm, which frequently decreases to 72.9% by increasing dye concentration to 25 ppm. Such decreases in PD with increasing dye concentration are due to the more adsorption of dye molecules on the catalyst surface and occupy its active sites. The adsorb dye molecules absorb a significant amount of light rather than a catalyst which decrease the generation of hydroxyl radical and hence reduces the photocatalytic efficiency (Reza et al. 2017).
Production of H2 attracts attention because it is essential for fuel and chemical reactions(Yamada et al. 2020). Both calcined and uncalcined were also utilized for water splitting. The PEC performance of the calcined and uncalcined rust photoelectrodes was explored via chronoamperometry and linear-sweep voltammetry (LSV). Figure 9a,b indicates the dosing effect on the overall water oxidation for calcined and uncalcined photoanodes. It can be seen that increasing the catalyst dose from 0.1 to 0.3 mg/50mL ethanol has a nonsignificant effect in the case of uncalcined samples. However, the calcined samples indicate enhancement in the activity until 0.3 mg/50mL ethanol, after which saturation took place and no further photocurrent density enhancement observed. Figure 9c displays the LSV measurements under regular solar illumination. For both the samples, the photocurrent remained significantly lower under dark, while under the simulated light, the photocurrent densities increased enormously beyond 0.7 V vs RHE with the voltage sweep, as indicated by the I-V spectra. The spectra represent that both the materials are photoactive, but the photocurrent density of calcined Rust (0.42 mA cm− 2) is much higher than uncalcined Rust (0.34 mA cm− 2), which can be attributed to the good crystallinity and porosity of the calcined sample. A significant dark current appeared above 1.23 V vs RHE for the calcined rust photoelectrodes, which can be attributed to electrochemical processes (Khan and Qurashi 2017). The photocurrent densities observed at the thermodynamic potential (1.23 Vvs RHE) of the water oxidation reaction were ~ 0.40 and ~ 0.32 mA/cm2 for calcined and uncalcined, respectively. Moreover, it is important to sate the onset potential of calcined rust shift cathodically to 0.96 V vs RHE from 1.07 V vs RHE (for uncalcined sample). The excellent photodegradation efficiency, photostability, and photocurrent density, with good cathodic onset potential shift of calcined Rust, indicate that it could be a feasible photocatalytic as well as photoelectrocatalytic materials for future energy and environment applications. Figure 9d shows the long-term I-t photostability curve obtained from photoanodes under the light with an illumination period of 80 min under calibrated simulated 1-SUN power. The results revealed that the stability curve of calcined Rust decreased insignificantly and showed small photocorrosion decay and hence offered considerable resilience and stability during the irradiation period. Similarly, the stability curve of uncalcined Rust significantly decreased with more considerable photocorrosion decay, which can be attributed to the photocorrosion of this material due to structural faults and charge high recombination rate. The uncalcined rust lost its PEC activity entirely after 57 min, due to uncontrolled photodecay.
The optoelectrical characteristics are useful to realize the photoactivity of the samples. For this purpose, we have performed the UV/Vis-diffuse reflectance analysis of the powder sample. The samples were measured using a quartz cell, from wavelength ranges from 200 to 800 nm at the scan rate of 50 nm/min. As indicated in Fig. 9e, the calcined samples are showing superb enhancement in the visible light absorption. The onset value of absorption in the range of 580–600 nm for the calcined sample indicates that the bandgap value is well in the order of visible light region and suitable for PEC water splitting application. The samples are expected to show improved Photocatalytic water splitting and dye degradation performance, which is the case as discussed above. To further support over claim we also measured the resistivity and surface transport behavior of the samples via electronic impedance spectroscopy (EIS) measured at 0.2 V vs. SCE under AM 1.5 G irradiation as shown in the Fig. 9f. Based on the semicircle diameter of the Nyquist plots one can realize the interfacial charge transport behavior (Khan and Qurashi 2018; Khan et al. 2021). The larger the semicircle means lower the transportation and vice versa. By comparing the circles it can be concluded that the calcined rust sample has more interfacial transport than the uncalcined. This means that the generated electrons/holes are facilely transferred in case of calcined rust, hence minimizing the charge recombination and enhance the over all photo and photlectrocatalytic activity. A comparison table (Table 1) is also provided which is showing some existing hemtite based materials with their photocurrent denisties and photostablities.
Table 1
Photocurrent densities and photostability of some reported hematite based photoanodes
S/No
|
Materials
|
Photocurrent [email protected] vs. RHE
|
Photostability [email protected] vs. RHE
|
Ref.
|
1
|
α-Fe2O3/TiO2
|
1.05 mA/cm2
|
2500 s
|
(Khan and Qurashi 2018)
|
Ag/α-Fe2O3/TiO2
|
2.59 mA/cm2
|
3600 s
|
2
|
Fe2O3
|
1.55 mA/cm2
|
---
|
(Wei et al. 2020)
|
FeOOH/Fe2O
|
2.40 mA/cm2
|
5 h
|
3
|
Ta:Fe2O3@Fe2O3
|
2.45 mA/cm2
|
5 h
|
(Zhang et al. 2020)
|
NiFe(OH)x/Ta:Fe2O3@Fe2O3
|
3.22 mA/cm2
|
5 h
|
4
|
Fe2O3
|
0.12 mA/cm2
|
2 h
|
(Tang et al. 2019)
|
Fe2O3/Fe2TiO5
|
0.90 mA/cm2
|
2 h
|
Fe2O3/Fe2TiO5/CoFe-PBA
|
1.25 mA/cm2
|
2 h
|
5
|
Uncalcined rust
|
0.34 mA cm− 2
|
20 min
|
This work
This work
|
Calcined rust
|
0.42 mA cm− 2
|
70 min
|