1. Introduction

Global warming is one of the significant challenges faced in the modern world. The current Earth's surface temperature is rising at about 20 times faster than its natural climate change and 2016 was recorded as the hottest year since the 1950s [1]. Although CO₂ is a minor atmospheric gas component, it has the most prevalent GHG effects, causing the rise of global atmospheric temperature, sea level and other environmental issues [2]. Its increase is strongly correlated to the growing global population, which subsequently elevated the world energy demand.

The world's population and energy demand is projected to reach 8.6 billion by 2030 and 48% increase by 2040, respectively [3-5]. Currently, about 86% of the global energy needs is met through the combustion of oil, gas and coal [6]. Therefore, the huge amount of CO₂ is continuously discharging to the environment and an immediate solution to tackle global warming is of utmost priority.

For the past few decades, there has been an increased interest in the carbon mitigation plan. One of the solutions is to utilise CO₂ as raw material to produce advantageous industrial chemicals, chemical intermediates and/or hydrocarbon fuels. The source of CO₂ can come from the pre-established carbon capture and sequestration technology. The fuel produced is combusted to produce energy and the CO₂ generated is captured and recycled for synthesis, forming an overall zero-carbon loop as per Fig. 1.

Carbon dioxide can be reduced to various chemicals, such as methane (CH₄), methanol (CH₃OH), formic acid (HCOOH), carbon monoxide (CO) and others. Methanol (CH₃OH) is the main product considered in this project due to its vast industrial use and applicability in energy-related sector [8, 9]. It is the precursor to most industrial chemicals including formaldehyde, MTBE and acetic acid, which are used in adhesives, subfloors, solvents, washer fluids, etc. Methanol is also a low emission and high-octane efficient fuel when compared to conventional gasoline and is therefore, a suitable fuel alternative. Furthermore, the production of CH₃OH rather than CO and CH₄ is more advantageous as it is liquid at atmospheric temperature and pressure, making it desirable for energy storage and transport [10].

The global market demand of methanol has been increasing over the past decades. The IHS Markit [8, 11] projected the annual market demand of methanol to increase by 25%, from 80 million metric tonnes (MMT) in 2016 to just below 100 MMT is 2020. Methanol demand in the newly emerging energy sector as direct blending into gasoline, production of biodiesel and dimethylether (DME) has been growing rapidly since the past 3 years. According tothesame source, about40% of the current methanol isconsumed in the energy sector and approximately 200, 000tonnes of methanol is used as a chemical feedstock or as vehicle fuel per day. Furthermore, China's direct blending of methanol into the gasoline pool is projected to grow from 7 MMT in 2017 to 10 MMT by 2026 [11]. Hence, the production of methanol to satisfy the global market demand which synergistically reduce the anthropogenic CO₂ and provide an alternative green fuel is a prime consideration to solve the aforementioned issues.

The main challenge tothis solution is that CO₂ is very stable and significant energy is required to break the bond prior to reducing it tousefulproducts.Infact, due toits highstability, theconversion of CO₂ is not possible without any catalyst and input energy [12]. Furthermore, methanol exhibits high selectivity only under extreme conditions when compared to CH₄ and CO [10]. As such, the current industrial scale methanol production is predominantly done through the steam reforming or coal gasification of natural gas. However, this remains an issue as it consumes large amount of fossil fuel and thus, generates significant CO₂ per tonne of methanol produced [9].

Electrochemical reductionmethod, wherebyelectrical energyis applied across two electrodes, enables CO₂ conversion under mild conditions [13, 14]. However, substantial development is still required before it is applicable in industrial-scale due to its high operating cost. Furthermore, significant amount of electricity, which primarily comes from fossil fuel incineration is required to powerthereaction.Therefore, thedevelopmentof arenewableand cost-effective methanol production method is paramount to ensure the sustainability of the CO₂ conversion.

The proposed solution is to utilise solar energy to power CO₂ reduction into methanol as it is abundant, renewable and environmentally safe. Although the current achievable photoconversion efficiencies are mostly unfeasible for industrial-scale operations when compared to thermal production process, the prospective of solar irradiation is a promising one. This is because the sunlight consists of wide range of energy; UV–vis-IR, which enables CO₂ reduction. An energy-effective, prompt and highly selective conversion of the stable atmospheric CO₂ is possible by the addition of catalyst to increase the rate of reaction.

The most commonly used metal-based photocatalyst such as TiO₂ and most group Ⅷ metals have limited availability and some are in rising threat of depletion due to its increased usage [15]. Additionally, these photocatalysts alone have poor light-utilisation and high recombination rate which reduces the overall quantum efficiency [16]. Another limitation of the preceding metal-based photocatalyst research is that they mostly focused only on the use of UV or IR energy for conversion. Very few research emphasised the visible light region which has shorter wavelength and thus higher energy than the infrared region.

Carbon quantum dots (CQDs) have attracted tremendous attention due to their wide availability, excellent physicochemical properties and vast potentials inphotocatalytic systems. They have low toxicity, chemically inert, cheap, have simple synthetics routes and are adaptive to surface functionalisation, which could further increase their designated function [17]. Several studies have reported CQDs' superior performance under light irradiation. CQDs are able to convert low-energy photons from the infrared (IR) and visible light (Vis) region into high-energy. This is highly beneficial, particularly when sunlight is used as only 4% of the solar spectrum is in the UV (high energy photon) region.

Thus, the ultimate solution to the climate change, rising energy demand and depleting reserves issues is the optimised utilisation of renewable sunlight energy to power CO₂ conversion into sustainable fuel with the addition of CQDs catalyst. The objective of this paper is to provide a thorough literature review on the proposed solution, whilst comparing the preceding CO₂ reduction research. Outlined in this paper are the possible methods of CO₂ conversion into methanol, the reaction mechanism, available catalyst and the properties and application of CQDs based photocatalyst along with their performance and synthesis method.

2. Summary of catalytic methods of methanol production 2.1. Thermal catalysis

Thermal catalytic reaction is the most commonly used reaction method in industrial scale due to its efficiencies in overcoming the endothermic nature of the CO₂ conversion. The current methanol production is primarily done through the classic Fisher-Tropsch synthesis of syngas (CO with H₂), which is generated by steam methane reforming (SMR) and gasification (CG) [9].

Methanol can also be produced by catalytic hydrogenation of CO₂ as per reaction (1) [18]. Cu-based catalysts are the most commonly used for this thermal reaction. However, its instability due to catalyst sintering and poor metal dispersion are still major issues. The addition of catalyst support such as reduced graphene oxide (rGO) or ZnO support were reported to improve performance [18, 19].

(1)

However, thermal reduction is still a fossil-fuel intensive process and generates significant amount of GHG throughout the process [10, 20]. This major drawback kills the process sustainability and would cause a net increase of CO₂ rather than the initial purpose of reducing it.

2.2. Electro-catalysis

Electro-catalysis is considered as one of the promising methanol production process due to its in-situ hydrogen generation in the aqueous solution for CO₂ reduction. Methanol is then formed by the hydrogenation of CO₂ as it is bubbled into the solution [21].

Over the years, there has been increasing interest on reducing GHG emission by integrating electro-catalysis with renewables such as wind, nuclear and solar cells to supply the electricity. However, the advancement of just the electro-catalysis itself is still hindered due to its poor energy efficiency, selectivity and high over-potentials for CO₂ reduction at low temperatures [21, 22]. The integration with renewables would only be efficient when moderate to high electro-catalysis performance is obtained. Furthermore, large-scale CO₂ reduction with electro-catalysis remains a challenge due to its low efficiency with high energy requirement and capital cost of electrolyser unit when compared to the Fisher-Tropsch process [20, 23]. In addition, the use of precious-metal electrode catalyst prevents its widespread application [24].

2.3. Photo-catalysis

The rising concerns from the use of non-renewable fossil fuel energy sources, such as coal and natural gas, can be overcome by the utilisation of light. The earth has abundant average solar irradiation, at approximately 120, 000 TW [7]. The efficient utilisation of this energy to power CO₂ conversion is the key to economical, affordable and environmentally sustainable source of renewable energy.

3. Photocatalysis for CO₂ reduction 3.1. Basic principles

The solar-driven reduction reaction can be thought as an artificial photosynthesis which generates energy while reducing or recycling waste products such as CO₂, simultaneously. Fig. 2 illustrates the artificial photosynthesis concept. The photosynthetic reaction can be summarised into three processes; (1) generation of charge carriers or electron-hole pairs upon suitable light-harvesting, (2) charge carrier separation and transportation, and (3) catalytic reactions between surface species and charge carriers [7, 25]. The sum and synergy of the thermodynamics and kinetics of these processes give the net efficiency of the system.

The redox reaction in photocatalysis can only be initiated when the photon energy is equal to or greater than the band gap energy received by the catalyst material [25, 26]. The solar energy first excites the electron from the valence band (VB) to the conduction band (CB), which then produces highly energetic and active charge carriers for reaction. VB is the highest energy band occupied by the electrons, whereas CB is the lowest energy band where no electron is at its ground state. In order for reduction reaction to occur, the CB potential has to be more negative than the required potential [12]. The opposite applies for oxidation reaction. This is because CB electrons reduce adsorbed species whereas VB electrons oxidise adsorbed species or cause indirect oxidation via surface-bound hydroxyl radicals [26].

Ideally, the photo-generated electrons will migrate on the surface, combine at shallow or deep trapping sites, then react with chemisorbed species such as CO₂ [12, 25]. However, this can only occur if the recombination rate is slower than the transition reactions. Charge recombination (CR) happens either when electrons fail to find any trapped sites or when the band gap energy of the catalyst material is too small. As a result of CR, small amount of thermal energy is released at the surface of bulk photocatalyst. A simplified photocatalysis schematic is illustrated in Fig. 3.

CO₂ reduction is a 4–8 electrons process [27]. Studies have stated the importance to ensure the equal number of photogenerated holes and electrons to prevent charge recombination and photocorrosion of photocatalysts [7, 27]. The addition of electron donors (catalyst pairing) is able to both scavenge these holes and supply hydrogen for the methanol synthesis. Since the photosynthetic reactions are comprised of several steps, catalyst pairing to create multiple proton-electron excitations pathway from physical contact cascade or Z-scheme is a common solution. Fig. 4 compares the single, cascade and Z-scheme photocatalytic reaction. Physical contact cascade is when photocatalyst pair is connected by junctions whereas Z-scheme system is comprised of H₂-evolving photocatalyst, an electron mediation and an O₂- evolving photocatalyst [28].

For example, Iwase et al. [28] demonstrated the Z-schematic for water splitting and CO₂ reduction composite catalyst under visible light. The study combined metal sulfides as an H₂-evolving photocatalyst, reduced graphene oxide (RGO) as electron mediator and a visible-light driven BiVO₄ as an O₂-evolving photocatalyst as illustrated in Fig. 5. The limitation to the study is the low selectivity on CO₂ reduction.

It is important that the multiple proton-electron transfer steps for CO₂ photoconversion and water oxidation are synergistically met during methanol synthesis. Water oxidation reaction in heterogeneous system is much slower (rate limiting step) than the proton reduction and CO₂ conversion is still kinetically challenging due to the high charge recombination rate, which inhibits the methanol formation [7]. This can be overcome by the addition of catalyst support and/or doping. By exploiting the different properties of various elements and its interaction with both the reagent and main catalyst, the overall reaction performance can be significantly improved.

3.2. Synthesis medium

The conversion of CO₂ into methanol can be done in gaseous or aqueous media. Literatures have shown that CO₂ tends to form CH₄ instead of CH₃OH under gaseous condition due to the reaction between carbon radicals and atomic hydrogen [29]. Studies by Mao et al. [30] reported the use of BiVO₄ catalyst in a gas photoreactor showed that the reaction is only selective to methanol under carbon-deficient system. It was also postulated that gaseous medium could inhibit methanol and CO desorption from catalyst surface, and re-oxidise the intermediate or product back into CO₂ from the molecular oxygen [12].

In contrast, aqueous CO₂ photocatalysis was reported to have higher methanol production rate relative to the other side products and is more advantageous due to the possibility of its direct use as liquid fuel [12]. The higher methanol yield is most likely due to the formation of formic acid (HCOOH), an intermediate to methanol synthesis. CH₄ and CO are formed concurrently and are usually present in trace amounts. The amount of H₂O to CO₂ ratio for methanol production is 2:1, excessive water was reported to suppress the reaction [31]. The fundamental reactions in aqueous media are listed below [30].

(2)

(3)

(4)

Furthermore, water in the aqueous media can be considered as an ideal electron donor and could prevent CB as previously mentioned. Studies reported the drawbacks of utilising pure water are its large oxidation potential, competition with CO₂ reduction and low solubility of CO₂ in water [12, 27]. However, experiments revealed the addition of NaOH or other alkaline solution is able to dissolve more CO₂, improve CO₂ reduction with more OH^- ions and promote longer decay time of CB electrons [30, 32]. Higher yield was due to the increased acidic CO₂ adsorption on the photocatalyst surface in the alkaline solution. OH^- ions serve as stronger hole-scavenger compared to water.

3.3. Photocatalyst 3.3.1. Types of photocatalysts

Photocatalysts can be categorised based on their structure or chemical nature. Three categories as described by Nikokavoura & Trapalis [26] are (1) Inorganic: single-metal oxides, mixed-metal oxides, metal oxide composites, layered double hydroxides (LDH) and salt composites, (2) carbonaceous: graphene (GR), carbon nanotubes (CNTs) and g-C₃N₄ composites and (3) hybrid organicinorganic photocatalytic materials. Each of the category's irradiation and experimental conditions requirement as reported are listed in Table 1.

TiO₂ is the most extensively studied photocatalysts due to its non-toxicity, high activity, performance and stability. It has remarkable performance in photoelectrochemical water splitting and CO₂ photoreduction. The issue with TiO₂ is its large band gap (~3.2 eV) which makes it active only under UV light, and the high electron hole recombination rate [32]. Similar issue occurs with SrTiO₃, ZnO, SiC and most metal oxides [7]. For example, studies of BiVO₄ by Mao et al. [30] showed CO₂ reduction rate increased by 31% under the irradiation of the whole solar spectrum compared to just Vis. The poor Vis-light performance was attributed to the narrow absorption band with absorption edge at ~550 nm, making its performance is highly dependent on the UV-light.

Copper is another popular catalyst for CO₂ reduction. However, due to the weak catalytic performance and adsorbed molecule binding, the major product formed at low potentials is CO [7]. This prevents the stabilisation and protonation of adsorbed CO to form CHO, which is one of the key intermediates for hydrocarbon formation [33]. Experiments by Gusain et al. [34] with CuO nanorods concluded CuO with the very low activity due to the fast recombination rate.

Most photocatalytic studies focus on the reduction of CO₂ under ultraviolet (UV). The disadvantage is its high energy, which only presents about 4% of the sunlight [35]. The IR region is 53% of the solar spectrum, but its low thermal energy often renders its impracticality for photocatalytic reduction. Vis contains higher when compared to IR and is more abundant than UV, representing 43% of the solar spectrum. The utilisation of Vis light for CO₂ conversion will be extensively elaborated in this review.

3.3.2. Properties of photocatalysts

Clearly, photocatalyst with low bandgap is more preferable as it is readily excitable under Vis light. It is common to apply the topdown approach whereby dopant or catalyst pairing was introduced to the base wide bandgap material to form visible-responsive catalyst. For example, Abdullah et al. [32] reported the stabilised TiO₂ over the presence of CeO₂ under Vis light. CeO₂ was found to decrease crystallite size, increase surface area, lower band gap energy and electron-hole recombination rate. Methanol yield was 3 times higher when compared to using the bare TiO₂. The XRD results from the study also reported the anastase phase with high crystallinity in CeO₂/TiO₂ has fewer defects for recombination and is therefore, more preferred for photocatalytic reaction.

In addition to doping and support addition, physical alterations to the catalyst and its support could be done to tune the photocatalytic performance. Maxwell et al. [36] reported the enhanced Vis absorption of nitrogen-doped anastase TiO₂ microsheets compared to the pure TiO₂ due to the heterojunctions formation on the microsheet surface. The study mentioned the dark colour inherited from the TiN precursor gives its strong absorption in the whole Vis spectrum.

Gusain et al. [34] compared the performance between CuO nanorods, reduced graphene oxide (rGO)-Cu₂O and rGO-CuO under Vis light. The pure CuO nanorods has very low activity and yield (5 and 7 times less) compared to other two, respectively. The rGO skeletons were able to prevent charge recombination and allow efficiency electrons transfer for better performance.

Sharma & Lee [37] investigated nanocomposite of Ni-TiO₂ immobilized on activated carbon fibers (ACFs). They found that Ni in the TiO₂ inhibited the grain growth of anastase crystallised, suppressed phase transformation and forms new adsorption site, thus improving optical activity. The addition of porous and large surface area of ACF decreased the rate of charge recombination by providing trap sites. Several photocatalyst reported by papers with their synthesis method, performance and reaction conditions are listed in Table 2.

3.3.3. Carbon quantum dots based photocatalysts

Despite the catalyst-pairing, doping and support, most inorganic catalysts still have poor solar irradiation sensitivity. Contrariwise, carbonaceous photocatalysts, are mostly active and exhibit good performance under Vis light irradiation as per Table 1. Carbon quantum dots (CQDs) are a new class of carbon nanomaterials comprised of discrete, quasi-spherical carbon nanoparticles with sizes below 10 nm [16]. Over the years, studies have witnessed the superior performance of CQDs-based photocatalyst owing to their unique microstructure and optical properties. Although the synthesis method may be facile, they are low cost, abundant, easy to functionalise and eco-friendly [26].

Several advantages of nanocarbons CQD as reported by literatures are: (1) good capability in the separation of photoexcited electron-hole pairs and thus, reducing CR, (2) ability to transfer electrons into an inorganic compound, (3) unique up-conversion photoluminescence (UCPL) property such that they can function as photosensitizer which extends light absorption range, meaning it can lower catalyst bandgap, (4) act as electron donor and reservoir. CQD can accept photogenerated electrons or act as a template to stabilize smaller catalyst particles, which could increase conversion rate, (5) able to transfer electrons into an inorganic compound, (6) conduct excess heat during catalytic reaction, and (7) tunable and easy to functionalise, including the introduction of heteroatoms to control pore structures, increase surface area, changing electronic and chemical properties [16, 38, 39].

One of the most denoted strengths is their UCPL property. CQDs are able to emit shorter wavelength (higher energy) than the excitation wavelength [16, 17]. Miao et al. [39] described UCPL as a two-photon absorption process. As electrons in the π orbital are excited by low energy photons, the electrons move to higher energy state. They then drop down into the σ orbital to emit shorter wavelength light by radiative relaxation. Li et al. [40] stated that the strong PL properties is attributed to the quantum confinement effect of emissive energy traps on the CQD surface.

Miao et al. [39] mentioned that the CQD emission wavelength is excitation-dependent. Hiswork observeda red-shift inthe emission peaks with excitation wavelengths above 550 nm. The opposite was found at lower excitation wavelengths as depicted in Fig. 6.

Zhang et al. [46] reported the excitation of CQDs with Vis light generate wavelengths at 330–530 nm, some of which lies in the UV region as per Fig. 7. This makes CQDs effective photosensitizers and photocatalyst as (1) It is able to activate co-catalyst with high band gap by supplying the high energy emitted from the UCPL and (2) easily-activated under Vis light irradiation, hence, lowering energy supply requirement, making solar energy CO₂ reduction more effective and feasible.

Pure CQDs alone were found to exhibit excellent photocatalytic activity. Kang et al. [47] reported the high conversion (92%) and selectivity (100%) of CQDs with sizes of 1–4 nm in the photocatalytic oxidation of benzyl alcohol to benzaldehyde in the presence of H₂O₂ under NIR. The OH^● radicals formed by the decomposition of H₂O₂ on the CQDs surface is highly oxidative in forming benzaldehyde or benzoic acid. In addition, the π–π interactions between the CQDs and benzyl alcohol or benzaldehyde were able to facilitate the adsorption of reagent on the catalyst surface for conversion. Under the same oxidation process, Hu et al. [48] postulated the surface oxygenated functional groups are the main contributor affecting the photocatalytic CQDs performance. The carbonyl and carboxylic acid groups reduce CR by forming high upward band bending, whereas, the hydroxyl groups that bring about low upward bending have no CR effect. The role of CQDs in this process is to produce the radical OH^● to form the CHO- or COOH- functional group molecules as illustrated in Fig. 8. Another work by Kang's group with 5–10 nm sized CQDs showed high Vis-light-induced photon generation in solution [25]. The exited photons were equivalent to acid catalyst and the results indicated relatively high conversion at 34%–46.2% in various reactions including esterification, Beckmann rearrangement and aldol condensation.

The challenge to CQDs in photocatalytic reactions is their relatively low quantum yield (QYs), typically less than 10% [40]. Quantum yield is the ratio of the number of photons emitted to those absorbed. Although CQDs alone exhibit reasonable photocatalytic performance, Handoko et al. [7] stated that noble-metal co-catalysts will always be required for decent photocatalytic performance, as agreed with various literatures. For example, Zheng et al. [49] enhanced the QYs of CQDs from 2% to 24% by reduction with NaBH₄. The reduction shifted the maximum emitted wavelength from 520 nm to 450 nm due to the decreased CQDs sizes, which improves QYs (Fig. 9).

Depending on the aqueous medium conditions, the role of CQDs doped or paired with another material changes as CQDs can function as electron acceptor or donor, depending on the media they are exposed in. This was demonstrated by Li et al. [47] as per Fig. 10. The broad luminescence peak indicate that the emission intensities were quenched by the known electron acceptors and electron donors.

Yu et al. [50] studied CQDs/TiO₂ nanosheet (TNS) composites for the degradation of rhodamine B under Vis irradiation. The addition of CQDs into TNS was found to increase the transient photocurrent response by 3 folds, aid the transfer of photogenerated electrons and reduce CR. Martindale et al. [51] investigated CQDs/NiP for hydrogen generation in aqueous EDTA solution under solar irradiation. The results showed high activity in Vis region and high quantum efficiency. In this reaction, he postulated that CQDs do not serve as a sacrificial electron donor for auto-oxidation, rather they promote the electron transfer to NiP to reduce the aqueous protons. EDTA is the electron donor and quenches the photoinduced holes in CQDs to prevent CR. The mechanism as shown in Fig. 11.

Miao et al. [39] reported that CQDs/mesoporous TiO₂ was able to degrade methylene blue (MB) by 98% under Vis light. The author found that when the CQDs loading is above 5 wt%, the dye degradation efficiency decreases as excessive CQDs loading impeded the reaction of photo-induced electrons and holes with the chemisorbed oxidants and reducers. Additionally, they postulated that the lower energy reaching the composite is because the CQDs self-absorbed the excitation or emission energy. Work by Zhang et al. [52] on Fe₂O₃/CQDs for photodegradation of toxic gas (benzene and methanol) under Vis light showed that the addition of CQDs were able to increase the efficiency and activate Fe₂O₃ from the unique PL properties as seen in Fig. 12. In this case, CQDs provided electron-storage for the photon-excited electrons from Fe₂O₃ to prevent CR. The π–π interaction of the conjugated structure of CQDs and benzene also promote benzene adsorption on the surface.

The effect of CQDs doping can be clearly seen from Zhang et al. [53] experiment. The N-doped CQDs (NCQDs) showed broad UV–vis absorption peaks below 400 nm. The 240-nm peak was attributed to the π–π^* electronic transitions of the aromatic sp² domains and the n-π^* transition of C＝O bond. The 350-nm peak is the typical adsorption pattern of carbon nitrides semiconductor. The author also reported the role of NCQDs is to prevent CR by providing the photogenerated TiO₂ electrons trap site, and ease electron transfer and charge separation. These roles are consistent to those reported by Martins et al. [54] in their work on NCQDs/ TiO₂ for NO photo-oxidation under UV–vis irradiation, and by Chen et al. [55] using CQDs/BiOI microsphere for the degradation of methylene orange (MO) (Fig. 13).

CQDs also improve catalyst stability. For example, Zhu et al. [56] showed that the addition of N-CQDs on pure Ag₃PO₄ results in consistent performance in the repeated experiments for the degradation of MB. In this process, N-CQDs act as electron trap to prevent CR that would otherwise happen very promptly as pure Ag₃PO₄ photogenerated electrons are unstable (Fig. 14).

Li et al. [57] synthesised Cu₂O/CQDs for Vis light driven CO₂ reduction into methanol. They initially identified the issue with the low stability of Cu₂O being easily oxidised or reduced by photogenerated charge carriers despite its low bandgap at 2.2 eV. The results showed considerably higher methanol yield of 167mmol/g-cat after 3 h (rate of 56 mmol g-cat^-1 h^-1) when compared to pure Cu₂O and CQDs. This performance was attributed to the enhanced light absorption as evidenced by having broader and stronger peaks from the DRS reflectance data, Fig. 15a. The bandgap energy of CQDs/Cu₂O is lower than the pure Cu₂O, at 1.96 eV, suggesting CQDs lowered the apparent bandgap of the composite. It was reported that the accumulation of MeOH in the solution limits the yield as noted by the black line in Fig. 15b. The repeated experiments in Fig. 15b showed no performance decay after 5 cycles, showing constant methanol formation.

Fig. 16 illustrates the postulated conversion mechanism. As visible light was irradiated, electron-hole pairs were generated from the Cu₂O, the electrons were consumed to produce methanol on the Cu₂O surface. Water is oxidised into O₂ by the holes on CQDs surface. In this case, CQDs are electron acceptor. Hence, in the overall process, the roles of CQDs are to enhance Vis absorption, lower composite band gap, and activate Cu₂O as electron acceptor.

Table 3 lists CQDs based photocatalysis based on their reaction purpose. As can be seen, despite its high photocatalytic activity, the conversion of CO₂ into hydrocarbons is scarcely studied, with Li et al. [57] being the only paper reporting its performance in CO₂ reduction into methanol. This prompts the need to further study the noble material in photocatalysis.

4. Conclusion

To conclude, the photocatalytic reduction of CO₂ using various catalyst, whether pure, with doping and/or support have been evaluated. The conventional metal oxides catalysts, modified or un-modified, typically exhibits high activity only under the UV region. CQDs as a new subset of the carbonaceous catalyst have been acknowledged for its high photocatalytic performance under Vis light due to the excellent up-conversion PL and optical properties. CQDs also have multifaceted roles including electron mediators, spectral converter, photosensitizers and can function as good photocatalyst on its own.

The overall high photocatalytic performance of CQDs in different types of reactions indicate their promising outlook for the CO₂ conversion reaction. However, CQDs in photocatalytic reactions has yet to be fully exploited and the research is still on early stage relative to the other carbon family. For example, smaller CQDs produced by reduction reaction have higher QYs, but there are no available literatures on precisely tuning and manipulating the physicochemical properties of the CQDs to further improve performance. Additionally, majority of the papers reported CQDs excellent absorption of Vis region. However, very minimum literatures strive to fabricate and investigate possible heteroatom doping or surface modification to improve absorbance in the NIR region. There is also no clear explanation on how the synthesis method and properties such as crystallinity, size, surface area, doping and functionalisation affect the optical properties of CQDs.

Thus, it is hoped that this review can stimulate further research of the CQDs-based photocatalysis for CO₂ reduction into useful hydrocarbons, particularly methanol. Surface and band gap modification of CQDs via doping and functionalisation to further improve performance and efficiency is expected to be the future of sustainable photocatalysis and green chemistry.

Acknowledgements

Tjandra gratefully acknowledges the scientific communication in photocatalysis with Scientia Professor Rose Amal at the University of New South Wales (Australia).