3.1 Catalyst characterization
By choosing a suitable precursor, good nanofiber morphology can be obtained using the electrospinning technique. If an improper precursor is selected, the good morphology of the initial electrospun nanofibers can be destroyed during the calcination process. During the calcination of the electrospun nanofibers in air, the utilized polymer is fully eliminated and the metallic precursor decomposes into its most stable form. The main aspect of a proper precursor is polycondensation during sol–gel preparation. The high polycondensation tendency and hydrolysis reactions for alkoxides explain their distinct performance in forming an integrated network [44-46].
The organometallic family comprises compounds having one or more metal atoms in the molecule, and metal alkoxides are members of that family. Metal alkoxides (R-O-M) were obtained by replacing the hydrogen atom in the hydroxyl group of alcohols (R-OH) with a metal atom M, which are the class of chemical precursors most widely used in sol–gel formation because of their condensation behavior and tendency to combine together to form a gel network. The gel network can be formed using metal salts such as chlorides, nitrates, and acetates besides alkoxides. The acetates showed the most convenient polycondensation behavior for gel network formation. The excellent morphology and bead-free nanofibers shown in Fig. 1A indicate that the electrospinning working parameters and the composition of the utilized sol–gel were properly selected. As a result of the polycondensation behaviors of all utilized precursors, the calcination process had no effect upon the nanomaterials’ morphology.
It is notable that the addition of cadmium acetate did not affect the general morphology of the obtained nanofibers, however, the addition of cadmium distinctly affected the obtained nanofibers’ morphology. The nanofiber sample chosen for SEM imaging is that having a composition of 0.5 wt% Cd-doped TiO2 nanofibers, presenting the best photocatalytic performance under visible light irradiation. As observed in Figs. 2B and C, cadmium incorporation results in the breaking of nanofibers. The produced nanofibers length decrease by increasing the cadmium content in the initial electrospinning solution and acetate ions decomposition. Typically results in evolving CO2 and methane gases leaving behind holes on the surface, potentially explaining the nanofiber breakdown. Comparing Fig. 2B to 2A, the addition of cadmium can be seen to result in the creation of rough surface nanofibers. By contrast, the cadmium-free nanofibers have a smooth surface. A high surface area is a preferable characteristic for a photocatalyst, as it leads to enhanced photon absorption and consequently increases the catalytic activity.
Figure 2D shows the TEM image of the Cd-doped TiO2 nanofiber, the appearance of transparent CdTiO3 bubble-like nanoparticles attached to the surface of the TiO2 nanofiber is notable. As CdTiO3 nanoparticles exhibit several catalytic, optical, and electrical properties, their presence over the surface of the nanofiber can enhance the TiO2 nanofiber’s photocatalytic performance by functioning as electron trapping sites for the hydrogen evolution reaction, as will be discussed below.
Figure 3 displays the TEM and the EDX results for the Cd-doped TiO2 nanoparticles. The CdTiO3 phase clearly appears in the TiO2 matrix. As shown in the Figure, CdTiO3 has a different structure than that of TiO2 (Fig. 3B), which confirms the formation a Cd-doped TiO2 nanostructure. The EDX result (Fig. 3C) confirms the presence of Cd in the TiO2 matrix.
X-ray diffraction analysis is a typical technique for determining the compositions of crystalline materials. Basically, brookite, anatase, and rutile are common TiO2 phases from a crystal structure point of view, the latter two are abundantly found. Figure 4 represents the impact of cadmium doping on the crystal structure of the produced nanofibers. Cd-free electrospun nanofibers consist entirely of the anatase phase, the diffraction peaks appear at 2Ɵ values of 25.09o, 37.65o, 38.44o, 47.89o,53.89o, 55.07o, 62.40o, 68.70o, 70.04o, and 75.00o and correspond to the (101), (004), (112), (200), (105), (211), (204), (220), (220), and (215) crystal planes, respectively. This suggests the formation of pure anatase TiO2 according to the XRD database (JCPDS card no 21-1272). One peak refers to the rutile phase at a 2Ɵ value of 27.4 detected in the 0% Cd nanofiber sample. It is well-known that metal oxides can form chemical compound alloys. As can be seen in Fig. 4, CdAc addition enhanced the rutile phase formation, which compared to anatase, has a greater tendency for visible light absorption [25, 26]. The tetragonal rutile phase (JCPDS 21-1276) at 2Ɵ values of 27.4o, 36.1o, 39.2o, 41.2o, 44.1o, 54.3o, 56.6o, 62.7o, 64o, 69o, and 69.8o corresponds to the (110), (101), (200), (111), (210), (211), (220), (002), (310), (301), and (112) crystal planes, respectively. These results indicate that the amount of rutile phase increases with increasing cadmium content in the produced nanofibers. The rutile phase peak intensity decreases by increasing Cd content. The reason for this may be that the addition of foreign dopant atoms into the lattice weakens its structure, thereby decreasing crystallinity [46]. The rutile (110) peaks shift towards lower angle for the samples with the addition of Cadmium in addition to Cadmium smaller radii compared to radius of Ti+4 indicating the incorporation of Cadmium metal ions in the TiO2 matrix.
As observed in Fig. 4, Cd was dissolved in TiO2 to form cadmium titanate (CdTiO3). Diffraction peaks indicating CdTiO3 formation were observed at 31.1o, 34.2o, 46.9o, and 59.3o, corresponding to the (101), (104), (110), (024), and (214) crystal planes, respectively (JCPDS card no. 29-0277). Those peaks appeared at high concentrations of Cd-doped nanofibers (1 and 2%), while the CdTiO3 phase cannot be observed in 0.5 wt% Cd nanofiber. This may be due to polymer encapsulation at low Cd content. The XRD result for Cd-doped TiO2 nanoparticles shows a similar pattern to the nanofibers, as the XRD result does not depend upon the nanostructure’s morphology (Fig. 3B).
3.2 Photocatalytic activity investigation
Figure 5 represents the effect of cadmium content upon the amount of hydrogen evolved under visible light irradiation using the prepared Cd-doped nanoparticles and nanofibers. The hydrogen production rates for the 2 wt% Cd samples were 27 and 250 ml H2/gcat. for nanoparticles and nanofibers, respectively. The nanofibrous morphology distinctly enhanced its photocatalytic activity. As observed in the Figure, the rate of hydrogen production was increased greatly using nanofibers compared to nanoparticles at the same composition. The structure that provides one dimension for electron motion may be the reason for the considerable enhancement in the nanofibers’ photocatalytic activity, however, in the nanoparticles with 0D structure, full electron confinement takes place, which favors the e/h recombination reaction [47]. Moreover, the results show the effect of cadmium doping upon the photocatalytic activity. As shown in Fig. 5, a remarkable enhancement in the photocatalytic activity of TiO2 was detected after the addition of small amounts of cadmium. In fact, cadmium particles act not only as visible light absorption centers with an absorption coefficient dependent on Cd density, but also as recombination sites between photogenerated electrons and holes. Moreover, doping by Cadmium nanoparticles yields a higher Schottky barrier that increases electron capture [24]. This electron capture results in an increased e/h—pair separation lifetime, which consequently suppresses the recombination of e/h pairs. Decimating e/h-pair recombination enhances the hole transfer and perhaps O2 adsorption on the TiO2 surface. The observed decrease in the photocatalytic activity of the nanofibers for Cd concentrations above 0.5 wt% is related to the decrease in rutile phase intensity, as confirmed by XRD results (Fig. 4).
Basically, the photocatalytic activity is intensely related to the optical properties of the light source used in the experiments, like the light intensity and irradiation area. Hence, the catalyst activities cannot be compared with each other if the reaction conditions are different from each other. Therefore, determination of an apparent quantum yield (AQY), which rules the effect of light source out, is essential. Accordingly, It is necessary to determine the efficiency (AQY) of the present system and compare with other published works. To judge the feasibility of a proposed photocatalyst, the quantum efficiency of H2 evolution is calculated as follows:

where η is the energy conversion efficiency, Ne is the number of electrons involved in the hydrogen evolution reaction, and Nv is the number of incident photons in the reaction area. Ne can be calculated from the rate of hydrogen evolution (from Fig. 5, production rate= 0.63 mmol H2 g−1catalystmin−1 for Cd-doped TiO2 nanoparticles).
Through those rates, the actual number of electrons disappearing due to the hydrogen evolution reaction can be obtained, meanwhile, the number of incident photons (Nv) is calculated as

where I is the solar intensity in the El Minia Governorate, Egypt (Latitude: 31o 25`, I= 258–266.7 Wm−2 [48]) and Ev is the photon energy (3.97×10−19, J photon−1). Accordingly, the photon flux intensity is 6.51×1020 photons/s m2 solar. This calculation indicates that the prepared catalyst has achieved a 20 % conversion of photons to electrons, showing that the Cd-doped TiO2 nanostructure photocatalyst is capable of increasing the absorbance of photons and prolonging the lifetime for e/h pairs.
Instead, researchers used another practical standard to calculate the efficiency of the photocatalystic water splitting, namely solar-to-hydrogen (STH) efficiency. It can be calculated using the following formula:

where
is the hydrogen production rate in mol/s, ΔG is the Gibbs free energy associated with hydrogen gas in J/mol, Is is the solar energy flux in (W/cm2), and Ar is the surface area of the photocatalytic reactor. Using the STH formula, the STH efficiency of the prepared Cd-doped TiO2 nanoparticles is calculated to be 0.64% and 9% for Cd-doped TiO2 nanoparticle and Cd-doped TiO2 nanofiber, respectively. Table 1 shows the rate of hydrogen evolution of this study compared with other works. It is important to note that comparison of the rate of hydrogen evolution should be done for experiments that are conducted using same light source and same sacrificing agent. However, it is hard to fulfill that condition to compare our results with other works. For nanostructures containing TiO2, it is apparent that the moles of hydrogen produced by the Cd–TiO2 nanofibers prepared in this study is relatively higher than that obtained previously by other scholars (Table 1), the exception is Mohamed et al. [20], who indicated that a Cd-doped TiO2 nanotube can achieve 24 mmol H2/gcat. min using methanol as a scavenger agent. This reveals that the Cd-doped TiO2 nanofiber may act as an effective photocatalyst candidate for the photon-induced water-splitting reaction. However, the stability and recyclability of this substance must be considered in greater detail before nominating it as a viable photocatalyst.
Table 1: A comparison of the hydrogen evolution rate for different nanocatalysts
No
|
Photocatalyst
|
Light source
|
Scavenger agent
|
Hydrogen production rate
mmol H2/gcat. min
|
Ref./year
|
1
|
Pt/ TiO2 nanosheet
|
Xenon Arc lamb 350 W
|
Ethanol
|
0.0056
|
[49]/2010
|
2
|
Graphene modified TiO2 nanoparticles
|
Xenon Arc lamb 350 W
|
Methanol
|
0.0123
|
[50]/2011
|
3
|
TiO2 nanoparticles
|
Xenon lamb 150 W
|
Methanol
|
0.1
|
[51]/2014
|
4
|
(Pt/HS-TiO2)
|
Mercury lamb 400 W
|
Methanol
|
0.017
|
[52]/2016
|
5
|
Pt-doped TiO2–ZnO |
Mercury lamb 400 W
|
Methanol
|
0.0034
|
[53]/2017
|
6
|
Pt-TiO2 particles
|
Mercury lamb 450 W
|
Methanol
|
0.444
|
[54]/2005
|
6
|
Cd-doped TiO2 nanotube
|
Mercury lamb 1000 W
|
Methanol
|
24
|
[20]/2018
|
7
|
CdS/TiO2 mesoporous core-shell
|
Sunlight
|
Na2S/Na2SO3
|
1.13
|
[21]/2018
|
8
|
Ni/TiO2 nanotube
|
Solar simulator Xenon lamb
|
-
|
0.433
|
[22]/2019
|
9
|
Ni/GO-TiO2 nanoparticles
|
Sunlight
|
Methanol
|
3
|
[55]/2019
|
10
|
Ag-TiO2 NF
|
Mercury lamb 1000 W
|
Na2S/Na2SO3
|
2
|
[56]/2020
|
11
|
NiCo2S4/[email protected]
|
Sunlight
|
-
|
0.00125
|
[57]/2020
|
12
|
Cd-doped TiO2 nanoparticles
Cd-doped TiO2 nanofibers
|
Mercury lamb 1000 W
|
Na2S/Na2SO3
|
0.7
16.5
|
This study
|
3.3 Influence of reaction temperature
The kinetic energy can be measured by the temperature of the system, so higher temperature leads to a higher average molecular kinetic energy and more collisions per unit time. Therefore, in most chemical reactions, the temperature has a positive effect upon the reaction rates, however, nanostructures frequently show unexpected behaviors.
Figure 6 shows water photo splitting at temperatures of 298, 308, and 318 K. The results indicate that the hydrogen production rate decreases with increasing reaction temperature using Cd-doped TiO2 nanofibers (0.5 wt% sample) as a photocatalyst. This behavior may be because of the negative impact of the temperature upon the metal’s electrical conductivity as increasing temperature increases the electrons’ kinetic energy and disordering. This kinetic energy can move the reactant molecules away from the active zones. In addition, other researchers have noted that a surface plasmon is remarkably observed at low temperatures [58]. Moreover, increasing the temperature can increase the possibility of recombination between the charge holders, which consequently decreases the semiconductor’s photoactivity.
Therefore, we theoretically project that increasing the temperature is not preferred in the water-splitting reaction. This hypothesis has been verified experimentally (Fig. 6). Numerically, the hydrogen production rates are 350, 300, and 100 ml/ gcat at reaction temperatures of 298, 308, and 318 K, respectively. This finding confirms that the water-splitting reaction over the proposed catalyst surface does not follow the Arrhenius equation. Obtaining a high yield at low temperatures is an economically preferred characteristic in industrial settings.