Reactor Design for CO₂ Photo-Hydrogenation toward Solar Fuels under Ambient Temperature and Pressure

Abstract

Photo-hydrogenation of carbon dioxide (CO₂) is a green and promising technology and has received much attention recently. This technique could convert solar energy under ambient temperature and pressure into desirable and sustainable solar fuels, such as methanol (CH₃OH), methane (CH₄), and formic acid (HCOOH). It is worthwhile to mention that this direction can not only potentially depress atmospheric CO₂, but also weaken dependence on fossil fuel. Herein, 1 wt % Pt/CuAlGaO₄ photocatalyst was successfully synthesized and fully characterized by ultraviolet-visible light (UV-vis) spectroscopy, X-ray diffraction (XRD), Field emission scanning electron microscopy using energy dispersive spectroscopy analysis (FE-SEM/EDS), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), and Brunauer-Emmett-Teller (BET), respectively. Three kinds of experimental photo-hydrogenation of CO₂ in the gas phase, liquid phase, and gas-liquid phase, correspondingly, were conducted under different H₂ partial pressures. The remarkable result has been observed in the gas-liquid phase. Additionally, increasing the partial pressure of H₂ would enhance the yield of product. However, when an extra amount of H₂ is supplied, it might compete with CO₂ for occupying the active sites, resulting in a negative effect on CO₂ photo-hydrogenation. For liquid and gas-liquid phases, CH₃OH is the major product. Maximum total hydrocarbons 8.302 µmol·g⁻¹ is achieved in the gas-liquid phase.

1. Introduction

Nowadays, it is important to develop an efficient and effective method for utilizing carbon dioxide (CO₂) greenhouse gas. The CO₂ captured from the atmosphere will be stored in the ground or the ocean [1]. Another option, which is a more attractive method for CO₂ consumption, is to convert it into useful hydrocarbon fuels directly [2]. In industry, CO₂ is widely used for Sabatier reaction [3,4] and reverse water-gas shift (RWGS) reaction [5,6]. Regarding the Sabatier reaction, CO₂ reacts with H₂ to form CH₄ and H₂O at 300–400 °C. As for RWGS reaction, CO₂ firstly reacts with H₂ to form CO and H₂O; then CO can be easily converted into hydrocarbons by the Fischer-Tropsch reaction. However, both reactions require high temperature and high pressure to reduce CO₂ into hydrocarbons. Thus, it is inevitable to develop another method to conduct CO₂ reduction.

The possible CO₂ conversion processes, including biological [7], catalytic [8,9,10,11], photocatalytic [12,13,14], and electrocatalytic [15,16] conversion are summarized in

Table 1. Summary of different processes of CO₂ reduction.

Conversion Process	Main Products	Advantages	Disadvantages	Ref.
Biological	HCOOH: conversion efficiency of (22 ± 9) × 10−2%	Biological capability to synthesize liquid fuels	Complicated and cumbersome biological processes	[7]
Catalytic (Heterogeneous catalysis)	CH4: 3.8 × 104 μmol·h−1·g−1 catalyst	High efficiency	High temperature requirement	[8]
	CH3OH: 8.8% of CO2 conversion with the corresponded selectivity of 63%			[9]
	CH3OH: 193.9 g·kgcat−1·h−1			[10]
Catalytic (Homogeneous catalysis)	CH3OH: 79% yield			[11]
Photocatalytic (Heterogeneous catalysis)	CH4: 0.56 μmol h−1·g−1 catalyst	Storage of solar energy	Low efficiency	[12]
	CH3OH: 4.6 μmol h−1·g−1 catalyst			[13]
Photocatalytic (Homogeneous catalysis)	CO: 12.66 h−1 of TOF			[14]
Electrocatalytic	Liquid fuels	Converting CO2 directly to liquid fuels (long-chain molecule)	High energy barrier needs overcoming	[15,16]

. Among these processes, photocatalytic CO₂ reduction is one of the most promising technologies and has received much attention recently [14,17,18,19,20,21]. CO₂ as a C1-feedstock could be reduced to desirable and sustainable solar fuels at ambient conditions under light irradiation.

Ideally, photocatalytic CO₂ reduction could provide an alternative and sustainable pathway to producing desired hydrocarbon products from renewable energy and CO₂. However, it is noted that their efficiency is still relatively low. To enhance the photocatalytic performances, several factors—such as light harvesting, loss of photons, product separation, and charge carrier recombination—are considered in our group to design and developed in a new photo-reactor. In 2007, a circulated photocatalytic reactor was developed, which could provide a large specific surface area and uniformity of gas concentrations in the reactor [22]. In 2008, the optical-fiber photo-reactor, which could deliver light efficiently and uniformly to the surface of a photocatalyst, was also successfully designed [23]. It is important to note that there are two main sources to conduct CO₂ reduction, H₂O_(g) and H_2(g). In a previous study, Abbott et al. calculated a series of changes of enthalpy (ΔH⁰) and changes of Gibbs free energy (ΔG⁰), respectively, for CO₂ reduction to form hydrocarbons [24]. The result clearly shows that adding H₂ largely decreases these values. That is, the CO₂ reducing reaction in which the involved H₂ requires lower energy and becomes more spontaneous. Therefore, in 2013, a novel twin reactor was successfully developed to hydrogenate CO₂ into CH₃OH [20]. However, it is noted that its efficiency is still relatively low.

In this study, 1 wt % Pt/CuAlGaO₄ was prepared using a well-known solid-state fusion and photo-deposition method, respectively. Detailed characterization of photocatalyst was conducted to reveal its structure. Three kinds of reactors—including gas phase reactor, liquid phase reactor, and gas-liquid reactor—were designed and employed to study the influence of hydrogen (H₂) and carbon dioxide (CO₂) on the photo-hydrogenation of CO₂. Additionally, a possible reaction pathway of CO₂ photo-hydrogenation is also proposed based on the knowledge of products presented during the photocatalytic reaction.

2. Results and Discussion

2.1. Characterization of Photocatalysts

All photocatalysts were fully characterized by several techniques. Brunauer-Emmett-Teller (BET), ultraviolet-visible light (UV-vis) spectroscopy; X-ray diffraction (XRD), transmission electron microscopy (TEM), and Field emission scanning electron microscopy (FE-SEM) using energy dispersive spectroscopy (EDS) analysis were used to reveal their structure and surface morphology of photocatalysts. X-ray photoelectron spectroscopy (XPS) was used to reveal the chemical state of the species.

Figure 1 displays the UV-vis light absorption spectra of CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄ photocatalysts, respectively. Clearly, all the photocatalysts have the absorption band in the range of 250–670 nm, which is consistent with that reported previously [25]. It notes that loading of the Pt neither significantly improves the light absorption nor changes the band gap. A possible reason comes from the fact that Pt loading is very small and highly dispersed on the supporting surface.

Figure 2 shows the XRD patterns of CuAlGaO₄, 1 wt % Pt/CuAlGaO₄ photocatalysts and their starting materials of CuO, Al₂O₃, and Ga₂O₃. Clearly, no peaks can be assigned to either CuO, Al₂O₃, or Ga₂O₃, indicating that complete reaction of the precursors was achieved by solid-state fusion a mixture of Al₂O₃, Ga₂O₃, and ZnO at 1125 K for 12 h. The loading of Pt obviously retains the structural features of CuAlGaO₄ photocatalyst, but its crystallites appear to be slightly decreased. Both photocatalysts have patterns similar to that of the single-phase spinel type structure of the CuAlGaO₄ (JCPDS file, card No. 26-0514). However, there is no noticeable crystalline phase observed in the XRD pattern of Pt/CuAlGaO₄ photocatalyst that can be attributed to Pt element. As discussed above, the Pt loading might be very small and highly dispersed on the supporting surface, which is consistent with the UV-vis spectrum.

Figure 3 shows the SEM images with corresponded elemental spectra of CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄ photocatalysts, respectively. For CuAlGaO₄ photocatalyst, it has cube-like particles and smooth surfaces. Most importantly, loading of Pt on CuAlGaO₄ could not change the shape and morphology of the photocatalyst. Additionally, a uniform distribution of the Pt particles was found with a narrow size range of 4–20 nm.

Table 2. Energy dispersive spectroscopy (EDS) analysis for element compositions of CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄ photocatalysts.

Element	Atomic Ratio (%)
	CuAlGaO4	Pt/CuAlGaO4
O	61.9	61.6
Al	23.0	22.0
Cu	12.8	13.0
Ga	2.3	3.2
Pt	N/D 1	0.2

¹ N/D—Not detection.

shows elemental analysis of CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄, respectively. It clearly points out that only 1 wt % Pt/CuAlGaO₄ showed the Pt signal (Figure 3). On the other hand, the other elemental signals (O, Al, Cu, Ga) were very similar between CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄ photocatalysts.

Figure 4 reveals the TEM images, which were conducted to further understand the presenting of Pt loading on 1 wt % Pt/CuAlGaO₄ photocatalyst. It has several well-dispersed Pt nanoparticles with corresponded size of near 4–20 nm are located on the surface of Pt/CuAlGaO₄ photocatalyst.

XPS Pt4f spectra of CuAlGaO₄ and 1 wt %Pt/CuAlGaO₄ photocatalysts are shown in Figure 5. Most importantly, there are two possible peaks at 73.6–74.1 and 77.0–77.4 eV have been observed for 1 wt % Pt/CuAlGaO₄ photocatalyst, which may be attributed to 4f_7/2 and 4f_5/2 of Pt oxide, respectively [26].

Lastly, the BET surface area of CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄ was approximately 0.65 and 0.86 m²·g⁻¹, respectively. Both values were not very high due to the solid-state fusion method which operates at a high temperature, causing the photocatalyst aggregation.

2.2. Photocatalytic CO₂ Reduction with Gas, Liquid, Gas-Liquid Phase Reactors

In this study, several blank tests in all type of reactors (gas phase, liquid phase, and gas-liquid phase reactors) were conducted to answer the question whether the formation of the reaction products come from photo-hydrogenation and not from CO₂ contaminations or the photocatalyst itself. The result of tests was shown in

Table 3. Summary of blank tests for photo-hydrogenation of CO₂.

No.	Type of Reactors	Experimental Conditions				Yield of Products (µmol·g−1)
		CO2	H2 (0.01 atm)	Photocatalyst (1 wt % Pt/CuAlGaO4)	Light Source
1	Gas phase reactor	X 1	O 2	O	O	BDL 3
2		O	O	X	O	BDL
3		O	O	O	X	BDL
4	Liquid phase reactor	X	O	O	O	BDL
5		O	O	X	O	BDL
6		O	O	O	X	BDL
7	Gas-liquid phase reactor	X	O	O	O	BDL
8		O	O	X	O	BDL
9		O	O	O	X	BDL

¹ X—absent in the photoreactor; ² O—present in the photoreactor; ³ BDL—below the detection limit of gas chromatography.

. The result clearly shows that if CO₂ was not introduced to the reacting system, no product was generated even with the presence of the hydrogen, photocatalyst, and light irradiation. It is important to remember that the photocatalyst synthesis process did not use any carbon-containing precursor. This blank test result is further confirmation that the photocatalyst itself contains nearly no carbon residue. Table 3 also clearly shows that, without any of three essential elements in performing the photo-hydrogenation reaction, including (a) CO₂, (b) photocatalyst and (c) light source, no product could be detected.

For CO₂ reduction, there are several possible C₁ products, such as formic acid, formaldehyde, methanol, and methane. Because of the high detection limit of formaldehyde by HPLC (3.3 ppm), we used the Nash reagent [27] to react with formaldehyde first, then analyzed the colored product diacetyldihydrolutidine (DDL) by UV-vis at 414 nm to quantify the formaldehyde. The detail was described in the previous literature [27,28]. Even the detection limit of this method was 0.17 µmol·L⁻¹, which was about three orders lower than HPLC, formaldehyde was not observed under all investigated conditions. However, it is noted that although we did not detect formaldehyde, some of the formaldehyde might also be formed on the surface of the catalyst [29,30]. For the gas phase reactor, there is also no formic acid and methanol could be detected, and only methane evolution would be discussed in this case. The CH₄ yield under gas phase conditions was shown in Figure 6.

Most interestingly, H₂ plays an active role on CH₄ yield.

CO_2(g) + 4H_2(g) → CH_{4 (g)} + 2H₂O_(l)
∆H⁰ = −259.9 kJ/mol; ∆G⁰ = −132.4 kJ/mol

(1)

It is evident that the conversion of CO₂ into CH₄ has negative ∆H⁰ and ∆G values, meaning that the process is spontaneous, equilibrium favorable, and exothermic. In this study, the enthalpy (∆H⁰) and Gibbs free energy (∆G⁰) values at 298 K for several interesting reactions were calculated by Aspen [31]. At first 2 h, the initial rates of CH₄ yield were 0.01, 0.57, 0.73, and 0.40 µmol·g⁻¹·h⁻¹ that correspond to H₂ partial pressure = 0, 0.01, 0.1, 0.2 atm, respectively. It is worth noting that increasing the partial pressure of H₂ would enhance the yield of the product. Among the conditions of H₂ partial pressure, 0.1 atm shows the best performance. However, an extra supply of H₂ might compete with CO₂ for occupying the active sites, bringing a negative effect on CO₂ photo-hydrogenation. As shown in Figure 6, CH₄ evolution became stable after two hours in reaction, implying that the reactions in the system reached a dynamic equilibrium.

A comparative experiment was also conducted in the liquid phase. The correlation between the amount of CH₄, CH₃OH, and HCOOH formed and the reaction time for the liquid phase are shown in Figure 7.

The main products of the liquid phase reactor were CH₄, CH₃OH, and HCOOH, respectively.

CO_2(g) + 3H_2(g) → CH₃OH_(l) + H₂O_(l)
∆H⁰ = −137.8 kJ/mol; ∆G⁰ = −10.7 kJ/mol

(2)

CO_2(g) + H_2(g) → HCOOH_(l)
∆H⁰ = −31.0 kJ/mol; ∆G⁰ = +34.3 kJ/mol

(3)

It is clear that all exothermic reactions exhibit negative ∆H⁰ values. Moreover, many reactions are also associated with negative ∆G⁰ values and, as a consequence, the reaction will be thermodynamically favorable. Most importantly, CH₃OH was found to be the major product under current condition. This fact is that CH₃OH requires only six photoelectrons for the photo-hydrogenation of CO₂ while CH₄ needs to have eight photoelectrons for reaction. Last but not least, an extra amount of H₂ provided may also suppress the yield of products, which is consistent with the observation in the gas phase.

The correlations between the amount of CH₄, CH₃OH, and HCOOH formed and the reaction time for the combined gas-liquid phase are shown in Figure 8.

Similar to the result of the liquid phase, CH₃OH still is the main product for gas-liquid phase. On the other hand, the CH₄ yield has a tendency to be suppressed after conducting the experiment for two hours. It is possible that an intermediate product of CO₂ photo-reduction, CO, could be generated as the reaction proceeds (Equation (4)). Following is the competing reaction pathways of photo-reduction of CO to CH₃OH and HCOOH via H₂ and H₂O, respectively (Equations (5) and (6)). This result implies that the presence of photocatalyst in the gas phase will promote the converting of CH₄ into another compound, such as CH₃OH or HCOOH, resulting in a decrease of CH₄ yield.

CO_2(g) + CH_4(g) → 2CO_(g) + 2H_2(g)
∆H⁰ = −247.5 kJ/mol; ∆G⁰ = +170.8 kJ/mol

(4)

CO_(g) + 2H_2(g) → CH₃OH_(l)
∆H⁰ = −131.6 kJ/mol; ∆G⁰ = −29.9 kJ/mol

(5)

CO_(g) + H₂O_(g) → HCOOH_(l)
∆H⁰ = −24.8 kJ/mol; ∆G⁰ = +15.1 kJ/mol

(6)

In brief summary, the possible reaction pathways of CO₂ photo-hydrogenation is illustrated in Scheme 1. This mechanism is proposed based on the knowledge of products presented during the photocatalytic reaction. Our observation indicates that CH₄ is the only product under gas phase condition. On the other hand, CH₄, CH₃OH, and HCOOH are found in the liquid and gas-liquid phase reactions. Although we could not measure an intermediate CO product quantitatively, we expect that CO might be generated during the reaction. In the gas-liquid phase reactor, the photocatalyst is well packed and dispersed on the quartz plate in the gas phase. Hence, it might promote the transformation of CO into CH₃OH and HCOOH products. We do believe that conducting the experiment in different phases (such as gas, liquid, and gas-liquid phases) could not change the mechanism of the photo-hydrogenation of CO₂. However, the presence of photocatalyst in different phases might accelerate different pathways of the reaction.

Table 4. Summary of the product yields at 2 h in three kinds of reactors over 1 wt % Pt/CuAlGaO₄ under different H₂ partial pressures.

Entry	Phase	H2 Partial Pressure (atm)	Product Yields (µmol·g−1)
			CH4	CH3OH	HCOOH	Total HCs 1
1	Gas	0.00	0.012 ± 0.010	BDL 2	BDL	0.012
2		0.01	0.400 ± 0.100	BDL	BDL	0.400
3		0.20	0.780 ± 0.300	BDL	BDL	0.780
4	Liquid	0.00	0.010 ± 0.010	0.235 ± 0.100	BDL	0.245
5		0.01	0.149 ± 0.050	BDL	0.094 ± 0.045	0.243
6		0.20	0.112 ± 0.030	0.340 ± 0.200	BDL	0.452
7	Gas-Liquid	0.00	0.014 ± 0.010	0.285 ± 0.100	BDL	0.299
8		0.01	0.480 ± 0.200	7.352 ± 2.100	0.470 ± 0.100	8.302
9		0.20	0.666 ± 0.120	0.445 ± 0.110	0.145 ± 0.010	1.255

¹ Total hydrocarbons yield = CH₄ yield + CH₃OH yield + HCOOH yield); ² BDL—below the detection limit of gas chromatography.

summarizes the product yields at two hours in three kinds of reactors under different H₂ partial pressures. We see that H₂ plays an active role in the photocatalytic reduction of CO₂. Additionally, a remarkable synergetic activity was clearly observed when the experiment was conducted in gas-liquid phase under 0.01 atm of H₂. In more detail, the total products in the gas and liquid phase are 0.400 and 0.243 µmol·g⁻¹, respectively. Interestingly, about 8.302 µmol·g⁻¹ is achieved in the gas-liquid phase.

Lastly, the quantum efficiencies in three kinds of reactors under 0.01 atm of H₂ are also calculated. The highest quantum efficiency was in the gas-liquid phase reactor for 0.0011%, while the gas and liquid phases were about 0.0001% and 0.0005%, respectively. The possible reason for higher quantum efficiency in the gas-liquid phase reactor is that both H₂ in the gas phase and the proton in liquid phase could be utilized simultaneously, affording more chances to conduct CO₂ photo-reduction.

3. Materials and Methods

3.1. Preparation of Photocatalysts

CuAlGaO₄ photocatalyst was firstly prepared by solid-state fusion method, which is mentioned in previous studies [19,25]. Firstly, copper oxide (CuO, Showa, Tokyo, Japan), aluminum oxide (Al₂O₃, Type A-5, Sigma-Aldrich, St. Louis, MO, USA), and β-gallium trioxide (β-Ga₂O₃, ≥99.9%, Sigma-Aldrich) powders were mixed in the molar ratio of Cu/Al/Ga = 1:1:1 and pulverized in a mortar. Subsequently, the resulting mixture was calcined at 1150 °C for 12 h, and then cooled to room temperature and further pulverized to obtain a CuAlGaO₄ powder.

1 wt % Pt/CuAlGaO₄ photocatalyst was prepared by the photo-deposition method as described in the same reference. The required amount of chloroplatinic acid hydrate (H₂PtCl₆·xH₂O, ≥99.9%, Sigma-Aldrich) solution was mixed with the as-prepared CuAlGaO₄ powder. Herein, the loading of Pt on CuAlGaO₄ is 1 wt %. The mixed solution was irradiated by a UV source (320–500 nm, EXFO S1500, EXFO Inc., Quebec City, QC, Canada) for 90 min to perform the photo-deposition process. After that, the solid product was centrifuged and washed with deionized water several times. Finally, the washed material was dried at 80 °C for 8 h to obtain 1 wt % Pt/CuAlGaO₄ powder.

3.2. Characterization of Photocatalysts

X-ray diffractometer (XRD, Ultima IV, Rigaku, Tokyo, Japan) equipped with Cu Kα (1.5418 Å) was used to verify the crystalline structure of photocatalysts. The UV-vis diffuse reflectance spectrum of the photocatalyst was fully recorded over the range 300–800 nm by a Cary 100 UV-visible spectrometer (UV-vis, Varian Cary 100, Agilent Technologies, Santa Clara, CA, USA). BaSO₄ was used as the reflectance standard. A field emission scanning electron microscope (FE-SEM, Nano SEM 230, FEI, Hillsboro, OR, USA) equipped with energy dispersive spectroscopy (EDS) was used to directly reveal the presence of the atomic elements in the photocatalysts. Transmission electron microscopy (TEM, Hitachi H-7100, Hitachi Inc., Tokyo, Japan) was performed to check the surface morphology of CuAlGaO₄ and 1 wt % Pt/CuAlGaO₄. The state of Pt loaded on CuAlGaO₄ was determined by X-ray photoelectron spectroscopy (XPS, Thermo Theta Probe, Thermo Fisher Scientific Inc., East Grinstead, UK) equipped with a Mg Kα (1253.6 eV) X-ray source. The Brunauer-Emmett-Teller (BET) surface area of photocatalysts was measured by a specific surface area analyzer (ASAP 2020, Micromeritics, Norcross, GA, USA).

3.3. Photo-Hydrogenation of CO₂ Reaction

In this study, we perform the experiments in three kinds of batch reactor systems: gas phase reactor, liquid phase reactor, and the gas-liquid phase reactor (Scheme 2).

For the gas phase reactor, the cylindrical tube reactor was made of Pyrex with a volume of 260 mL (Scheme 2a). A quartz plate, 92 × 40 × 2 mm, was placed inside the reactor and inclined at an angle of 20° to face the light source, 300 W Xenon lamp (Model 66902, Newport, Irvine, CA, USA). During the photoreaction, the light source was placed at an equal-distance from the reactor to the lamp so that the catalyst could receive the same amount of light intensity (mostly visible light), which was around 270 mW·cm⁻². Powder photocatalyst (0.30 g) was evenly packed on the quartz plate. For the liquid phase reactor, the reactor was also made by Pyrex with a volume of 385 mL (Scheme 2b). 0.30 g photocatalyst was well dispersed in the liquid phase with stirring. A 300 W Xenon lamp was also provided as the light source. The liquid phase solution, which contained 180 mL 0.1 M NaOH (Alfa Aesar, Ward Hill, MA, USA), was used to increase the solubility of CO₂. For the gas-liquid phase reactor (Scheme 2c), it was almost the same with the liquid phase reactor that combined with a Teflon plate. 0.10 g and 0.20 g 1 wt % Pt/CuAlGaO₄ photocatalysts were well packed and dispersed on the Teflon plate and in the solution, respectively. Before introducing the CO₂ (99.999%, Air Products, Taipei, Taiwan) and H₂ (99.99%, Air Products) into the reactor, Argon (99.9995%, Air Products) was purged to bring out any impurity contained within.

In this study, all the samples were taken and analyzed every two hours. Both methane (CH₄, gas phase) and methanol (CH₃OH, liquid phase) were detected by individual GC-FID (China GC 2000, China Chromatography Co., Taipei, Taiwan) equipped with a packed column (molecular sieve 5A, 3.5 m) with a carrier gas of Ar (99.9995%, Air Products). Formic acid (HCOOH, liquid phase) was detected by High Performance Liquid Chromatography (HPLC) (Shodex RI-2000, Shoko Co., Tokyo, Japan) equipped with a column (ICSep ICE-ORH-801, 300 mm, Interchim, Montluçon, France).

4. Conclusions

A potentially clean, low-cost CO₂ photo-hydrogenation method to convert the solar energy into desirable and sustainable energy—such as methanol, methane, and formic acid—has been demonstrated. In this study, CO₂ as a C1-feedstock was photo-reduced in the gas phase, liquid phase, and gas-liquid phase, respectively, over 1 wt % Pt/CuAlGaO₄ photocatalyst under different H₂ partial pressures. Interestingly, a remarkable result has been observed in the gas-liquid phase. The result also shows that increasing the partial pressure of H₂ could improve the yield of products. However, it notes that H₂ might compete with CO₂ for occupying the active sites when an extra amount of H₂ is provided, bringing about the adverse effect on CO₂ photo-hydrogenation. For liquid and gas-liquid phases, CH₃OH is the major product while CH₄ is the only product for the gas phase. The quantum efficiencies in gas, liquid, and gas-liquid phase under 0.01 atm of H₂ are about 0.0001%, 0.0005%, and 0.0011%, respectively. It is worthwhile to mention that this direction not only can potentially depress atmospheric CO₂, but also weaken dependence on fossil fuel.