Photocatalytic Hydrogen Production Under Near-UV Using Pd-Doped Mesoporous TiO₂ and Ethanol as Organic Scavenger

Abstract

Photocatalysis can be used advantageously for hydrogen production using a light source (near-UV light), a noble metal-doped semiconductor and an organic scavenger (2.0 v/v% ethanol). With this end, palladium was doped on TiO₂ photocatalysts at different metal loadings (0.25 to 5.00 wt%). Photocatalysts were synthetized using a sol-gel method enhancing morphological properties with a soft template precursor. Experiments were carried out in the Photo-CREC Water II reactor system developed at CREC-UWO (Chemical Reactor Engineering Centre- The University of Western Ontario) Canada. This novel unit offers hydrogen storage and symmetrical irradiation allowing precise irradiation measurements for macroscopic energy balances. Hydrogen production rates followed in all cases a zero-order reaction, with quantum yields as high as 30.8%.

1. Introduction

Hydrogen is a key energy carrier that will likely play an important role in the transportation sector by 2050 [1]. It is considered an environmentally friendly energy vector due to its zero CO₂ and zero noxious gas emissions when combusted [2]. Through a photocatalytic water-splitting process, hydrogen can be produced using water and a light source as primary resources [3]. Furthermore, a sacrificial organic agent is required to allow the photocatalytic reaction to occur, forming the desired products [4].

Common sacrificial agents include methanol, triethanolamine, ethanol, acids and inorganic compounds [5]. Ethanol, as one of the most-investigated sacrificial agents, provides high quantum efficiencies and will be used in this work as scavenger. The use of ethanol as a sacrificial agent is advantageous given that it can be easily produced from renewable biomass (fermentation processes), making it available and inexpensive [6].

Photocatalytic hydrogen production with sacrificial organic agents proceeds as follows: (a) absorbed photons surpass the energy band gap and generate excited electron-hole pairs [7], and (b) photoexcited electron-hole pairs can be separated due to the sacrificial agent presence. This allows the formation of hydrogen with minimum electron-hole pair recombination [8] and (c) hydroxyl groups from dissociated water lead OH· radical formation and contribute to the conversion of the scavenger [9]. The “in-series-parallel” reaction network was described in detailed by our research team in [10].

Titanium dioxide (TiO₂) is a well-known photocatalyst capable of absorbing light and producing electron-hole pairs to accelerate the rate of a water-splitting reaction [11]. TiO₂ has been the most used material due to its stability, resistance to corrosion, cleanliness (no pollutant), availability in nature and inexpensiveness compared to other semiconductors [12]. It can be found in three allotropic phases—anatase, rutile, and brookite—where the anatase phase is the most photoactive phase reported [13]. Doping noble metals on TiO₂ allows (a) increasing the efficiency of the hydrogen evolution reaction, (b) narrowing the band gaps, and (c) improving the optoelectronic semiconductor properties [14]. Furthermore, doping TiO₂ with noble metals such as Pd helps to promote energy levels near the band edges or mid-gap states, which may reduce the effective bandgap energy [15]. Therefore, the effect of Pd on TiO₂ can be attributed to the shift of Fermi levels in the composite material [16,17,18,19,20].

One of the most important parameters in photocatalytic reactors and photocatalysts is the quantum yield (QY). This parameter establishes the process efficiency relating the photogenerated radical rate over the absorbed photons rate [21]. Using the QY, the doped photocatalysts of the present study were evaluated.

Previous studies by our research team considered Pt doped on TiO₂ obtaining modest quantum yield efficiencies in the 0.7–8.0% range when using undoped DP-25 and 1.0% Pt-impregnated DP-25, respectively [22]. However, the present work emphasizes the photocatalytic hydrogen production using doped Pd, as a much less expensive dopant on mesoporous TiO₂. This semiconductor material, Pd–TiO₂, is employed under near-UV light, in the Photo-CREC Water II reactor unit developed at CREC (Chemical Reactor Engineering Centre). Results obtained are of significant value due to the high quantum yields obtained in the system.

2. Results and Discussion

2.1. Photocatalyst Characterization

2.1.1. Brunauer–Emmett–Teller (BET) Surface Area

Using a BET surface area unit (Micrometrics, ASAP 2010), the photocatalysts were analyzed with nitrogen. Furthermore, the adsorption–desorption isotherms of type IV were generated and the V_p total photocatalyst pore volume was also calculated with the liquid nitrogen adsorbed at the P/P_o relative pressure of 0.99 [23].

Table 1. Surface area and pore diameter using template Pluronic F-127.

Photocatalyst	SBET (m2 g−1)	DpBJH (4VpBJH/SBET) (nm)	VpBJH (cm3g−1)
Degussa P-25	59	7.5	0.25
F-127–TiO2 500 °C	140	17.5	0.61

reports the specific surface area, the average pore diameter and the specific pore volume for mesoporous TiO₂ prepared using F-127 template. It is shown that when using this template, the morphological properties of TiO₂ are improved as follows: (a) specific surface areas are increased, (b) average pore diameters are augmented, and (c) specific pore volumes are increased. Furthermore, one can also notice that the F-127–TiO_2, displays both higher porosity and specific surface area than the TiO₂ itself.

According to

Table 2. Surface area and pore diameter using palladium.

Photocatalyst	SBET (m2 g−1)	Dp BJH (4VpBJH/SBET) (nm)	VpBJH (cm3g−1)
Anatase	11	7.3	0.05
Rutile	5	4.7	0.05
Degussa P-25	59	7.5	0.25
F-127–TiO2-500 °C	140	17.5	0.61
F-127–0.25 wt% Pd–TiO2 500 °C	131	16.5	0.53
F-127–0.50 wt% Pd–TiO2 500 °C	124	16.8	0.52
F-127–1.0 wt% Pd–TiO2 500 °C	123	21.2	0.65
F-127–2.5 wt% Pd–TiO2 500 °C	122	19.9	0.60
F-127–5.0 wt% Pd–TiO2 500 °C	119	18.9	0.56

, the best results in specific surface area were obtained with the mesoporous TiO₂ photocatalysts. These photocatalysts display a clear increment of the specific surface area and specific pore volume (D_p) when compared to Degussa P-25 (commercial titania photocatalyst that is commonly used in photocatalytic reactions). However, when using doped Pd on TiO₂, there was a modest reduction in specific surface area and a mild change in pore diameter attributed to a possible and moderate blocking of the TiO₂ pores with Pd [24].

The Barrett–Joyner–Halenda (BJH) method was also used to determine the pore size distribution, by utilizing N₂ as an adsorbate and as a desorption isotherm. In all cases, a bimodal pore volume distribution was observed. The largest pore sizes in the 16–35 nm range were achieved with F-127–1.0 wt% Pd–TiO₂-500 °C.

2.1.2. Pulse Hydrogen Chemisorption

Table 3. Chemisorption analysis: metal dispersion.

Photocatalyst	Metal Dispersion (%)
F-127–0.25 wt% Pd–TiO2 500 °C	75
F-127–0.50 wt% Pd–TiO2 500 °C	27
F-127–1.0 wt% Pd–TiO2 500 °C	26
F-127–2.5 wt% Pd–TiO2 500 °C	12
F-127–5.0 wt% Pd–TiO2 500 °C	8

reports hydrogen chemisorption showing the effect of metal loading on metal dispersion. When Pd is used as a dopant, it is shown that higher metal loadings lead to reduced metal dispersion.

2.1.3. X-Ray Diffraction (XRD)

Figure 1 reports XRD diffractograms for TiO₂ synthesized using the F-127 template and with varying palladium loadings. Anatase and rutile XRDs are reported as references. In this respect, one can observe that anatase peaks are at the 2θ diffraction angles of 25°, 38°, 48°, 54°, 63°, 69°, 70.5° and 75° corresponding to the planes (101), (004), (200), (105), (204), (116), (220) and (215) [JCPDS No. 73-1764]. For rutile, there are a couple of noticeable peaks at 54° and 67° corresponding to the planes (201) and (301) [JCPDS No. 34-0180].

Furthermore, Figure 1 also shows that the XRD diffractograms for Pd-doped TiO₂ were consistent where a significant anatase XRD peak was observed. The nature of the desirable anatase peaks in this semiconductor was confirmed with a 99.7% anatase from Aldrich reference sample [25]. Anatase for all photocatalyst was the dominant TiO₂ crystalline phase assumed as 100% with no rutile being present.

Pd peaks were also identified and recorded at 40.12° (111) and 46.66° (200) at the 2θ angles [JCPDS No. 87-0638]. One should observe that, in principle, a third peak at 2θ = 68.1° (220) might be recorded when using Pd as a dopant. However, this peak may overlap with anatase and as a result cannot be used for Pd identification [26].

On the other hand, the crystallite sizes for each photocatalyst were determined using the Scherrer equation. On this basis, the mesoporous photocatalysts displayed crystallite sizes between 9 and 14 nm. Lattice parameters of the tetragonal anatase unit cell were also calculated and are reported in Appendix B.

2.1.4. Band Gap

Figure 2 reports that increasing the Pd content slightly augments the band gap. However, and regarding the observed results, one can see in all cases a significant reduction in the band gaps for TiO₂ doped with lower Pd loadings versus the band gaps for undoped TiO₂. The best band gaps achieved were 2.51 for 0.25 wt% Pd and 2.55 eV for 0.50–1.00 wt% Pd–TiO₂.

By applying the Kubelka–Munk (K–M) model and following the Tauc plot methodology, the band gaps were determined. Figure 3 reports the changes of the “(αhv)^1/2” function versus the photon energy “hν”, with α representing the absorption coefficient, h being the Planck constant (6.34 × 10³⁴ J s/photon) and v denoting the radiation frequency. It should also be noted that v = c/λ, where c is the speed of light under vacuum (3.00 × 10⁸ m/s²). If the straight-line methodology is applied for the band gap calculation as shown with the red line, one can see that the intersection of this line with the abscissa provides the wavelength corresponding to the semiconductor band gap. Furthermore, the Tauc plots (Figure 3) were developed for Pd-doped TiO₂ photocatalysts using the F-127 template and at a 500 °C calcination temperature.

2.1.5. X-Ray Photoelectron Spectroscopy (XPS)

The 1.0 wt% Pd–TiO₂ photocatalyst was analyzed using the XPS technique. Figure 4 shows the XPS for Pd(0) and PdO. In each case, one can see double peaks as follows: (a) Pd (0) 3d5/2, with 334.70 eV and 3d3/2 at 339.96 eV binding energies, (b) PdO 3d5/2 with 336.46 eV and 3d3/2 at 341.72 eV. This yields a consistent 48.6% of Pd (0) and 51.4% of PdO, at the two binding energy ranges considered.

It was also observed that titanium and oxygen were present as major components in the photocatalyst mesoporous support as TiO₂ species. Titanium was detected at a binding energy position of 454.45 eV, while oxygen was identified at 525.85 eV. These bands fell outside the Pd and PdO binding energies as shown in Figure 4, avoiding any possible inadequate band assignment.

In conclusion, the XPS data of Figure 4 shows the significant Pd (0) availability, and points towards possible future improvements of the synthesized photocatalyst via enhanced Pd reduction.

2.2. Macroscopic Radiation Energy Balance (MREB)

Photocatalytic reactors operate based on emitted photons. These photons are absorbed by a circulating semiconductor slurry suspension. To be able to establish the absorbed radiation in the Photo-CREC Water II Reactor, one must develop a macroscopic radiation balance for accurate energy efficiency calculations [27].

The macroscopic balance estimates the photons absorbed as the difference between the incident photons and the combined scattered and transmitted photons [28].

$$P_a=P_i-P_{bs}-P_t$$

(1)

where Pa is the rate of absorbed photons, for which it is desired to be as high as possible; P_i is the rate of photons reaching the reactor at the inner Pyrex glass surface and is calculated according to Equation (1) in Einstein s⁻¹; P_bs is the rate of backscattered photons; and P_t is the rate of transmitted photons. All these variables can be expressed using the Einstein s⁻¹ units.

Furthermore:

$$P_i=P_0-P_{a-wall}$$

(2)

with P₀ in Einstein s⁻¹ being the rate of photons emitted by the lamps as per P_a-wall in Einstein s⁻¹, which accounts for the rate of backscattered photons absorbed by the Pyrex glass walls.

In addition, P₀ can be calculated as:

$$P_0={\displaystyle {\int }_{{\lambda }_1}^{{\lambda }_2}\lambda }{\displaystyle {\int }_0^L{\displaystyle {\int }_0^{2\pi }q(\theta ,Z,\lambda )r}d\theta dzd\lambda }$$

(3)

where q (θ, z, λ) is the radiative flux (J s⁻¹ m⁻³), λ represents the wavelength (nm), r stands for the radial coordinate (m), z denotes the axial coordinate (m), h is the Planck’s constant (J s), and c represents the speed of light (m s⁻¹). The term q (θ, z, λ) is determined using the spectrometer.

Furthermore, when photocatalytic experiments are performed in the Photo-CREC Water II (PCW-II) reactor, photons are absorbed and scattered in the reacting medium. As a result, a backscattering has to be accounted for. A possible approach to calculate backscattering is to establish the difference between P_i and the rate of photons transmitted when the catalyst concentration approaches zero (P_t|_C→0+)_:

$$P_{bs}=P_i-P_t{|}_{c\to 0^{+}}$$

(4)

Equation (4) assumes that photons are backscattered on the TiO₂ particle layer close to the inner surface of the transparent Pyrex walls surface. Equation (4) also assumes that no other backscattered photons contribute to P_bs.

Additionally, for P_t determination, Equation (5) considers that transmitted radiation can be defined as the addition of normal scattered photons and forward scattered photons:

$$P_t=P_{ns}-P_{fs}$$

(5)

One should note that (P_fs + P_ns) can be measured by employing aluminum polished collimators, which capture radiation reaching the measuring point, with large view angles [28].

Thus, to assess P_a as in Equation (1), macroscopic balances using near-UV light were established at the central axial position using a 0.15 g/L photocatalyst concentration. Figure 5 reports measurements for various TiO₂ photocatalysts with different metal loadings.

According to

Table 4. Absorbed photon rates on TiO₂ photocatalysts at 0.15 g/L of photocatalyst concentration.

Near-UV Light	Pa (Einstein/s)
TiO2	3.11 × 106
0.25 wt% Pd	3.18 × 10−6
0.50 wt% Pd	3.52 × 10−6
1.00 wt% Pd	5.11 × 10−6
2.50 wt% Pd	3.77 × 10−6
5.00 wt% Pd	3.76 × 10−6

and Figure 5, one can observe that additions of Pd on TiO₂ show that (a) lower Pd levels (0.25 to 1.00 wt%) lead to an increased P_a and high absorption efficiencies compared to undoped TiO₂, and (b) higher Pd levels (2.50 and 5.00 wt% Pd) give smaller P_a and reduced absorption efficiency. These findings are in line with an increased rate of transmitted photons when using low Pd loadings as well as incremental photon backscattering when using high Pd loadings.

2.3. Hydrogen Production

2.3.1. Effect of Palladium Loadings

Palladium was used as co-catalyst to dope the structure of the TiO₂ photocatalyst. This metal enhances the hydrogen production, as compared to the undoped mesoporous TiO₂. Nobel metal crystallites reduce the band gap and facilitate electron capture [29]. As a result, Pd reduces the recombination between holes and electrons, promoting better photocatalytic water-splitting performances [30].

Figure 6 reports the influence of Pd on TiO₂ in terms of cumulative hydrogen volume.

Figure 6 shows there is a maximum volume of 140 cm³ STP (standard temperature and pressure) of hydrogen produced in six hours when using 1.00 wt% Pd on TiO₂. This volume is slightly higher than the maximum volume of hydrogen produced when using 0.25 wt% and 0.50 wt% Pd–TiO₂, and three times the volume of hydrogen obtained for undoped TiO₂. One should also note as well that this volume is close to the 128 cm³ STP of hydrogen produced when platinum is used as a dopant under the same reaction conditions but with a much larger metal loading (5.00 wt% Pt) on TiO₂ [31].

Furthermore, one should note that the 140 cm³ STP of hydrogen produced in six hours with 1.00 wt% Pd on TiO₂ decreased up to 60 cm³ STP when using higher Pd loadings (2.50 wt% Pd and 5.00 wt% Pd). The macroscopic radiation energy balance provides an explanation showing that at the higher Pd loadings, there is increased irradiation backscattering, with greater irradiation being reflected and, as a consequence, light absorption being reduced. This is in contrast with the lower than 1.00 wt% Pd loadings evaluated, where the absorption efficiency, as well as the rate of transmitted photons, increases. Thus, a diminished irradiation absorption given by 2.50 wt% Pd and 5.00 wt% Pd negatively affects the photocatalyst performance [32].

In agreement with this, at the lower palladium loadings studied (0.25, 0.50 and 1.00 wt%) good metal dispersion, mildly affected specific surface area and pore structure were achieved [33]. On the other hand, for 2.50 and 5.00 wt% Pd–TiO₂, poorer metal dispersion with larger metal crystallite sizes were observed, with this being in line with the lower photocatalytic activity [34].

In all cases, palladium-doped TiO₂ showed a consistent steady linear trend. The hydrogen production rate displayed consistent zero-order kinetics, with no noticeable photocatalytic decay. This material is stable for extended irradiation periods and no apparent deactivation for 24 h following an “in series-parallel” reaction mechanism shown in detail in [22].

These results show that palladium at 1.00 wt% loading can produce valuable hydrogen yields, with this being an excellent replacement for platinum. As well, Pd can be considered more advantageous than Pt, given that Pd is less expensive (only 20–25% of the cost of platinum). Furthermore, and given the premise of nominal 1.00 wt% Pd–TiO₂, photocatalyst X-ray fluorescence spectrometry (XRF) was used to confirm the nominal loading. The observed XRF value was 1.17 wt% Pd on mesoporous TiO₂.

2.3.2. Effect of Catalyst Concentration on Hydrogen Production

Considering that 1.00 wt% Pd–TiO₂ showed the best performance in terms of hydrogen production, additional experiments were carried out to determine the influence of the catalyst concentration during photoreaction.

Figure 7 displays four different slurry concentrations of the 1.00 wt% Pd-TiO₂ photocatalyst: 0.15, 0.30, 0.50 and 1.00 g/L. These experiments were studied during 6 h of irradiation. One can observe that the runs with 1.00 g/L showed the highest hydrogen production. Thus, given these results, it can be considered that when higher photocatalyst slurry concentrations are used, more photocatalyst electron-holes are provided, with this promoting better hydrogen production.

As a result, it was observed that the hydrogen production rate increased 54% when photocatalyst concentration was augmented seven times from 0.15 to 1.00 g/L. However, despite this hydrogen production increase, this could be considered a modest improvement only, given that the photocatalyst needed and the related cost was significantly augmented. Therefore, a photocatalyst concentration of 0.15 g/L was considered as a best choice and was selected for further experimentation.

2.3.3. Effect of Photo-CREC Water II Atmosphere using Argon and CO₂

Before starting water-splitting runs, the reactor gas chamber was purged with an inert gas to remove the oxygen from the air, avoiding combustion reactions. Argon was used initially as the inert gas given this is heavier than oxygen facilitating its displacement [35]. On the other hand, CO₂ was also used in separate runs in the reactor gas chamber to determine its possible influence on water dissociation reactions.

According to Figure 8, using argon as an inert gas and utilizing 1.00 wt% Pd–TiO₂, yielded 140 cm³ STP of hydrogen after six hours of irradiation. On the other hand, when a CO₂ atmosphere was employed, only 80 cm³ STP of hydrogen was obtained. One should note that under a CO₂ atmosphere and due to the competition of the CO₂ photoreduction with the hydrogen production, a lower net hydrogen formation can be explained.

Regarding CO₂ during the six h of irradiation, it was observed that it steadily augmented under an argon atmosphere reaching 0.4 cm³ STP. On the other hand, when the runs were performed under a CO₂ atmosphere, the CO₂ increment was limited to 0.01 cm³ STP. These findings support the view that there is competition between CO₂ photoreduction and CO₂ formation via ethanol OH· radical scavenging. It is assumed that these gas phase CO₂ findings could be also be influenced by the enhanced CO₂ solubility in water–ethanol [36].

2.3.4. Effect of Sacrificial Agent Concentration

As a scavenger, ethanol offers important advantages, such as the photogeneration of electron-holes, limiting electron-site recombination and improving photocatalytic activity. Ethanol can donate electrons to scavenge the valence holes and suppresses the reverse reaction [37].

Experiments were performed at 1.00 wt% Pd and three ethanol concentrations (1.0, 2.0, 4.0 v/v%) under an argon atmosphere and with 0.15 g/L of photocatalyst concentration. This was done to evaluate the effect of the ethanol concentration on hydrogen production. Figure 9 reports the influence of increasing ethanol from 2.0–4.0% on hydrogen production rates.

As shown in Figure 9, the highest hydrogen formation rate was obtained at the highest ethanol concentration. However, these important ethanol concentrations changes did not influence hydrogen production significantly. This was particularly true between 2.0 and 4.0 v/v% ethanol concentration. Therefore, 2.0 v/v% was considered fully adequate and was the selected concentration of the ethanol scavenger used for further studies.

2.3.5. By-Products Formation

There are several by-products generated from the water-splitting reaction in the gas phase. Detected by-products include methane, ethane, acetaldehyde and CO₂. To quantify these by-products, gas samples were taken hourly from the gas port located in the storage tank. They were analyzed using a Shimadzu gas chromatograph (GC) unit. All the experiments were repeated at least three times to secure reproducibility.

One can thus see that as soon as the photo-redox reaction starts, all these by-products, together with hydrogen, increase progressively as is shown in Figure 10. In the liquid phase, ethanol was also measured using a Shimadzu HPLC. One can observe in Figure 11 a balanced consumption-formation of ethanol, with a net stable ethanol concentration. This occurs when hydrogen is being produced using the 1.00 wt% Pd–TiO₂ photocatalyst.

The observed trends could be considered a promising result, showing that none or little additional scavenger is required in subsequent runs once the initial ethanol is fed to the Photo-CREC Water II reactor unit.

2.4. Quantum Yield (QY) evaluation

The quantum yield (QY) is the most important parameter to establish the energy utilization efficiency in photocatalytic reactors [38]. In terms of hydrogen production, quantum yield can be defined as the hydrogen radical production rate over the absorbed photon rate on the photocatalyst surface. According to this definition, QY can be determined as follows:

$$Q{Y}_{H•}=\frac{molesofH•/s}{molesofphotonsabsorbedbythephotocatalyst/s}$$

(6)

Equation (6) is equivalent to:

$$\%QY=\frac{\left[\frac{d{N}_H}{dt}\right]}{P_a}\times 100$$

(7)

where $\frac{d{N}_H}{dt}$ represents the rate of moles of hydrogen radicals formed at any time during the photocatalyst irradiation.

To use Equation (7) the assessment of P_a or the moles of absorbed photons is required. This can be accomplished by using the macroscopic radiation energy balance (MREB) in the Photo-CREC Water II reactor as proposed by Escobedo et al. [39]. Appendix C provides a calculation sample to assess the QY.

2.4.1. Effect of Pd Addition on Quantum Yields

The quantum yield evaluation for different TiO₂ photocatalysts involves rigorous macroscopic radiation energy balances. These calculations require the assessment of the P_t transmitted, the P_i incident, and the P_bs backscattered photons using the macroscopic radiation energy balance as described in Section 2.2. With this information and using Equation (1), the P_a was calculated.

Furthermore, for every experiment and once the lamp is turned on, the rate of moles of hydrogen can be established. On this basis, QY% can be calculated using Equation (7).

Table 5. Quantum yield (QY) for the Pd–TiO₂ photocatalyst when using 0.15 g/L. All reported data are average values of three repeats.

Semiconductor	QY (%)
F–127 TiO2	5.0
F-127–0.25 wt% Pd–TiO2	13.7
F-127–0.50 wt% Pd–TiO2	12.8
F-127–1.00 wt% Pd–TiO2	10.9
F-127–2.50 wt% Pd–TiO2	9.6
F-127–5.00 wt% Pd–TiO2	8.5

and Figure 12 report QY% for the mesoporous photocatalysts doped with palladium at different metal loadings (0.25, 0.50, 1.00, 2.50 and 5.00 wt%) under the following conditions: (a) photocatalyst slurry concentrations of 0.15 g/L, (b) 2.0 v/v% ethanol, (c) pH = 4 ± 0.05 and (d) near-UV light.

There is a significant increase of QY% with 0.25, 0.50 and 1.00 wt% Pd–TiO₂, whereas higher Pd loadings led to a decrease of QY%. These results are in line with the QY% of 8% reported by Escobedo when Pt addition proceeds [39].

Figure 12 reports that QY% displays consistent trends for Pd-doped TiO₂ photocatalysts: (a) during the first hour of irradiation, QY% increased progressively until it reached a stable value; and (b) during the following six hours of irradiation, QY% remained unchanged, with this showing a steady performance of the photocatalysts under study.

2.4.2. Effect of Catalyst Concentration on Quantum Yields

Considering the QY% observed for the 1.00 wt% Pd–TiO₂ during hydrogen production, further QY% evaluations were developed by changing the photocatalyst concentration in the slurry.

Table 6. Quantum yield for 1.00 wt% Pd–TiO₂ photocatalyst at different photocatalyst concentrations in the slurry.

Catalyst Concentration (g/L)	QY (%)
0.15	10.9
0.30	14.5
0.50	22.4
1.00	30.8

and Figure 13 report the QY% obtained, by augmenting the photocatalyst concentration, under the following conditions: (a) 2.0 v/v% ethanol as scavenger organic compound, (b) pH = 4 ± 0.05 and (c) near-UV light irradiation.

Figure 13 provides QY% for different photocatalyst concentrations. Here, it was again observed that there was a noticeable increase of the QY% in the first hour of irradiation, followed by a stable QY% in the next 5 h of irradiation. Constant QY% during the 1 to 6-h irradiation period was assigned to the steady hydrogen formation rate, linked to consistent zero-order reaction kinetics with no photocatalyst activity decay observed in all cases.

3. Experimental Methods

The photocatalysts of the present study were synthesized using the sol–gel methodology and doped with palladium. Different techniques were utilized to characterize the doped semiconductors as follows: (a) BET for specific surface area, (b) chemisorption for crystallite size, (c) x-ray diffraction for crystallographic structure and (d) UV-vis absorption for band gaps. The prepared semiconductors were evaluated in a Photo-CREC Water II reactor unit.

3.1. Photocatalyst Synthesis

The sol–gel method can be used for photocatalyst synthesis by converting monomers into colloids (sol phase), and thus promoting a gel structure formation [40]. The sol–gel method for TiO₂ synthesis can be modified, leading to improvements in photocatalyst structural properties such as particle diameter and surface area. Therefore, this also leads to improved photocatalytic activity [41].

Some copolymers, such as Pluronic^® F-127 and Pluronic^® P-123, formed by chains of ethylene oxide and propylene oxide, can be used for TiO₂ synthesis as soft templates. These templates optimize the pore structure network during semiconductor preparation, enhancing pore size distribution, enlarging the surface area, controlling the purity, homogeneity, and morphology of mesoporous materials [42].

Rusinque shows that the Pluronic F-127 template has a greater impact than the Pluronic P-123 template on TiO₂ photoactivity, increasing the hydrogen production up to 86% [31]. Thus, considering the Pluronic F-127 advantage over Pluronic P-123 for hydrogen production, further experiments were carried out using only copolymer Pluronic F-127.

The sol–gel method adopted used the following reagents: (a) ethanol USP (C₂H₅OH) from commercial alcohols, (b) hydrochloric acid (HCl, 37% purity), (c) Pluronic F-127, (d) anhydrous citric acid, (e) titanium (IV) isopropoxide, and (f) palladium (II) chloride (PdCl₂, 99.9% purity). All the reagents were obtained from Sigma Aldrich, with photocatalyst preparation effected according to the methodology proposed by Guayaquil et al. [43].

Figure 14 describes the sol–gel synthesis as follows: (a) Step 1: In 400 mL of ethanol, 33 g of hydrochloric acid and 20 g of Pluronic F-127 were added until dissolution, under continuous stirring for 1 h. (b) Step 2: 6.30 g of citric acid were dissolved in 20 mL of water for posterior addition to the initial suspension to mix them together for 1 h in order to set the pH at 0.75. (c) Step 3: 28.5 g of titanium (IV) isopropoxide was dissolved in ethanol and added dropwise to the mixture. Finally, palladium (II) chloride was incorporated at different loadings (0.25 to 5.00 wt% Pd). (d) Step 4: The resulting sol–gel suspension was stirred for 24 h and then calcined at 500 °C for 6 hours under an air atmosphere. The copolymer was evaporated during the thermal treatment and an ordered mesoporous titanium framework was formed [23].

3.2. Equipment

The Photo-CREC Water II (PCW-II) reactor is a novel unit used for water splitting reactions and therefore, hydrogen production. It is a 5.7 L slurry batch reactor configured with two concentric tubes: (a) an inner tube made from transparent borosilicate (Pyrex) and (b) an outer tube made from opaque polyethylene. The fluorescent lamp is placed inside this inner Pyrex tube. Furthermore, the suspended photocatalyst flows in the annular space between the outer polyethylene tube and the inner Pyrex transparent tube which only absorbs 5%) of the near-UV light emitted by the lamp [44]. See Appendix A for a detailed lamp characterization.

The PCW-II unit is equipped with a storage feed tank where the photocatalyst suspension is always kept sealed under agitation. This tank has 2 ports for periodic liquid and gas phase sampling. Figure 15 describes the main components of PCW-II: (a) the Photo-CREC Water II Reactor, (b) the centrifugal pump, (c) the sealed storage tank, and (d) the electrical circuit powering the near-UV light lamp.

The emitted radiation spectra of the lamp used inside the Photo-CREC Water II was established using a Stellar Net EPP2000-25 spectrometer (StellarNet Inc.). The light source is a polychromatic black light blue (BLB) Ushio UV lamp (15 W, 0.305 A, 55 V) with a spectral peak at 368 nm in the 300–420 nm emission range [45].

3.3. Photocatalyst Characterization

Photocatalyst specific surfaces areas were determined using a BET surface area analyzer (Micrometrics, ASAP 2010) at −195 °C. Each photocatalyst was degassed at 300 °C during a period of 3 h. The BET analysis was developed using nitrogen to generate the adsorption–desorption equilibrium isotherms and to establish the isotherm inflection point. The BJH (Barrett–Joyner–Halenda) method was used to determine the pore size distribution, by utilizing the desorption isotherm with N₂ as an adsorbate.

By using the Micromeritics AutoChem II Analyzer for pulse chemisorption, one can calculate the fraction of dispersed metal and average active metal crystallite size [46]. Furthermore, to identify the phases of a crystalline material, X-ray diffraction (XRD) was used [47]. The XRD spectra were analyzed in a Rigaku Rotating Anode X-Ray Diffractometer (Rigaku) perated at 45 kV and 160 mA. The scans were taken between 20–80°, with a step size of 0.02° and a dwell time of 2 s/step.

In order to determine the characteristic band gap associated to each photocatalyst an UV-VIS-NIR spectrophotometer (Shimadzu UV-3600) was used [48]. BaSO₄ was utilized as a reference sample. Kubelka–Munk (K–M) developed a Tauc plot methodology that was followed to establish the corresponding band gaps [49]. X-ray photoelectron spectroscopy (XPS) analysis was also used to identify the elemental composition and the chemical state of each element in the synthesized photocatalyst [50].

3.4. Hydrogen Production

Pd-doped TiO₂ photocatalysts were evaluated using the Photo-CREC Water II reactor equipped with the BLB near-UV lamp for 6 hours of continuous irradiation. This lamp was turned on 30 min before initiating the photoreaction. The hydrogen storage/mixing tank was loaded with 6000 mL of water. Ethanol was used as an organic scavenger and the pH was adjusted to 4 ± 0.05 with H₂SO₄ [2M] keeping the photoreaction under acidic conditions, which favours available H⁺ for water splitting process [39].

Following this step, the photocatalyst was loaded at a specific weight concentration ensuring that most of the radiation was absorbed in the slurry medium. The photocatalyst was subjected to sonication, which reduces the formation of particle agglomerates and promotes homogeneous mixing. Argon gas was circulated to guarantee an inert atmosphere at the beginning of the reaction.

3.5. Analytical Techniques

The gas phase was analyzed with a Shimadzu GC2010 gas chromatograph using argon (Praxair 99.999%) as gas carrier. It has 2 detectors, a flame ionization detector (FID) and a thermal conductivity detector (TCD). This unit was equipped with a HayeSepD 100/120 mesh packed column (9.1 m × 2 mm × 2 μm nominal SS) used for the separation of hydrogen from air. This equipment detects hydrogen (H₂), carbon monoxide (CO), carbon dioxide (CO₂), methane (CH₄) and other hydrocarbon organic species.

A Shimadzu HPLC model UFLC (ultra-fast liquid chromatography) system was utilized to characterize the liquid phase. This analytical technique allows the liquid mobile phase (0.1% H₃PO₄) to transport the sample through a column (Supelcogel C-610H 30cm × 7.8mm ID) containing a stationary phase. It selectively separates individual compounds (i.e., ethanol) from water for further detection. This quantitative analysis is performed by employing the RID (refractive index detector) 10A due to polar nature of ethanol.

Both the GC and the HPLC analytical techniques were used simultaneously. Samples were taken at different irradiation times.

4. Conclusions

(a)

The TiO₂ mesoporous photocatalysts of the present study were prepared using a F-127 template and following a sol–gel methodology. It was found that the mesoporous prepared using a F-127 template displayed a good photocatalytic performance.

(b)

The prepared Pd–TiO₂ photocatalysts were characterized using BET, XRD, UV-VIS and XPS. On this basis it was proven that energy band gaps were significantly affected with Pd addition, and that binding energies showed significant contribution of the Pd (0) on the doped-palladium TiO₂.

(c)

Macroscopic radiation energy balances were successfully employed to establish photon absorption rates and radiation absorption efficiencies in the PCW-II unit. For the Pd–TiO₂ semiconductors, photon absorption efficiencies were in the 45 and 60% range under near-UV light.

(d)

The formation of hydrogen using Pd–TiO₂ photocatalysts followed, in all cases, steady zero-order kinetics with no apparent photocatalyst activity decay.

(e)

The prepared Pd–TiO₂ photocatalysts under near UV-light were shown to be adequate for hydrogen production reaching up to 210 cm³ STP when using the 1.00 wt%-Pd on TiO₂. This photocatalyst showed a best QY% of 30.8%.