A review of recent advances in water-gas shift catalysis for hydrogen production

The water-gas shift reaction (WGSR) is an intermediate reaction in hydrocarbon reforming processes, considered one of the most important reactions for hydrogen production. Here, water and carbon monoxide molecules react to generate hydrogen and carbon dioxide. From the thermodynamics aspect, pressure does not have an impact, whereas low-temperature conditions are suitable for high hydrogen selectivity because of the exothermic nature of the WGSR reaction. The performance of this reaction can be greatly enhanced in the presence of suitable catalysts. The WGSR has been widely studied due do the industrial significance resulting in a good volume of open literature on reactor design and catalyst development. A number of review articles are also available on the fundamental aspects of the reaction, including thermodynamic analysis, reaction condition optimization, catalyst design, and deactivation studies. Over the past few decades, there has been an exceptional development of the catalyst characterization techniques such as near-ambient x-ray photoelectron spectroscopy (NA-XPS) and in situ transmission electron microscopy (in situ TEM), providing atomic level information in presence of gases at elevated temperatures. These tools have been crucial in providing nanoscale structural details and the dynamic changes during reaction conditions, which were not available before. The present review is an attempt to gather the recent progress, particularly in the past decade, on the catalysts for low-temperature WGSR and their structural properties, leading to new insights that can be used in the future for effective catalyst design. For the ease of reading, the article is divided into subsections based on metals (noble and transition metal), oxide supports, and carbon-based supports. It also aims at providing a brief overview of the reaction conditions by including a table of catalysts with synthesis methods, reaction conditions, and key observations for a quick reference. Based on our study of literature on noble metal catalysts, atomic Pt substituted Mn3O4 shows almost full CO conversion at 260 °C itself with zero methane formation. In the case of transition metals group, the inclusion of Cu in catalytic system seems to influence the CO conversion significantly, and in some cases, with CO conversion improvement by 65% at 280 °C. Moreover, mesoporous ceria as a catalyst support shows great potential with reports of full CO conversion at a low temperature of 175 °C.


Introduction
Fossil fuels have been continuously used as the primary source of energy for a long time. These resources are not renewable and pollute the atmosphere through greenhouse gases emission, which is linked to global warming. The depletion in fossil fuel reservoirs as well as the knowledge of their adverse environmental impacts has gained researchers attention to think about alternative and clean sources of energy [1]. Hydrogen is considered as a promising renewable alternative that minimizes CO 2 emissions and produces only water as a byproduct upon combustion [2]. Traditionally, among the non-fossil fuel-based methods, hydrogen has been produced by mixing iron (in the metal state) with strong acids, or using the well-established water electrolysis process. However, both techniques have some drawbacks, such as being costly to operate and not adequate to meet the growing market demands, which resulted in methane steam reforming being the most popular method for industrial-scale hydrogen production [3]. Recent research and development in the direction of alternative and sustainable energy production, such as wind energy, solar energy, and geothermal energy, indicates that future energy demands will be met by a mixture of sources to decrease our dependence on fossil fuels. As the renewable energy becomes cheaper with time and starts to gradually dominate the energy market, diverse applications of existing hydrocarbons are being sought by converting them into more valuable chemicals and products rather than combusting them for energy production, which is not environmentally friendly and expected to be more expensive in the future.
The abundance of water and sunlight as affordable sources offers a good alternative for hydrogen generation. Hence, photocatalytic water splitting is considered to have a potential for sustainable production of hydrogen that can be employed in different scales of hydrogen generators [4]. For achieving high conversion efficiency in solar water splitting, photocatalysts should have some critical features such as high absorption of visible light, low charge recombination, and high surface reaction [5]. It is assumed that photocatalytic hydrogen is conceivable as a commercial fuel in the future, yet a considerable research is required in this area, addressing the aforementioned key points for manufacturing. The enhancement in photocatalytic performance is attributed to the presence of oxygen vacancies, which act as trapping sites via taking photogenerated holes or electrons that could be beneficial in the carrier's separation [6]. Furthermore, during the synthesis of photocatalysts, nanostructures and morphologies can be controlled to facilitate the generation of oxygen vacancies that efficiently raise the adsorption capacity and consequently make gas molecules (including O 2 , N 2 , H 2 O, and CO 2 ) active [7]. Oxygen vacancies help promote hydrogen evolution reaction performance, due to their beneficial role in charge transfer and water molecule dissociation [8].
Hydrocarbons are converted to other chemicals, including hydrogen, by means of reforming processes in the presence of suitable catalysts. The WGSR, ubiquitous in reforming processes, is known to play a central role in adjusting hydrogen and carbon monoxide amount. In the WGSR, water reacts with CO to form hydrogen and carbon dioxide (CO + H 2 O ↔ CO 2 + H 2 ), where the CO 2 can be separated from the stream to get pure hydrogen [9]. The WGSR is relevant to various industrial sectors, directly or indirectly, such as the fertilizer industry for the production of ammonia, petroleum refineries, and fuel cell for generating power as well as transportation applications [10]. This reaction is exothermic, which makes it thermodynamically suitable at lower temperatures [11]. Being exothermic in nature may lead the system to a thermally runaway condition at higher temperature if appropriate cooling arrangements are not made. The heat released will increase the temperature further accelerating the reaction and generating more heat. Hence, control and monitoring the temperature is mandatory in order to prevent such instability. Since the WGSR is an equilibrium-limited exothermic reaction, lower temperatures will favor more hydrogen production, however with slower kinetics [12]. Higher temperatures are more desirable for reforming reactions. Therefore, if a high ratio of H 2 to CO is needed and CO 2 in the reaction amount is larger than the stoichiometric value, higher temperatures are not favorable [13]. With the objective of enhancing the reaction performance, the WGS reaction has been studied by many groups using different catalysts, resulting in a substantial literature elucidating the catalysis and reactor design aspects. Some review articles are also available summarizing the effect of various parameters, such as reactants flow rate, temperature, and pressure on WGSR in the presence of catalysts [14][15][16]. Catalyst composition, metal-support interactions, porosity, and surface area effects have been widely reported; however, in our knowledge, a study relating the catalyst nanostructures with WGSR performance is still lacking. The microscopic imaging techniques have seen a remarkable improvement in the past few decades, enabling us to image and monitor the changes at nanoscale structures while performing catalytic reactions, with structural details leading to single-atom precision. This enhancement resulted in revealing valuable nanoscale insights that tend to affect the catalytic reactions significantly, leading to the emergence of novel research areas such as single-atom catalysts, tandem catalysts, and shape-size selective catalysts. We believe that a study is needed to collect and correlate the key developments relating the nanostructure information with catalytic performance, specifically for a ubiquitous reaction like WGSR with a high industrial significance. Herein, our aim is to provide a review of the recent progress in catalysts for lowtemperature water-gas shift reaction over the last decade, with a primary focus on catalyst nanostructure analysis. In addition, we also aim to provide an overview of the recent development in catalyst preparation techniques, reaction conditions, supported catalysts design for WGSR.

Mechanism
Despite the simplicity of the reaction and extensive studies over the years, the WGS reaction mechanism is still controversial and complicated due to the sensitivity of the catalysts to minor changes in operational conditions. There are two prominent mechanisms for the WGS reaction: the regenerative (redox) mechanism and the associative (Langmuir-Hinshelwood) mechanism. Generally, at high temperatures, the WGS reaction is accepted to follow the redox mechanism, where CO molecule adorbs on the catalyst surface and abstracts one oxygen from the metal-oxide support to form carbon dioxide. Loss of oxygen from metal-oxide creates an oxygen vacancy that is fulfilled by dissociating a water molecule to generate hydrogen and oxygen atoms. Hydrogen atoms combine and desorb as hydrogen gas whereas oxygen atom is captured by the oxygen deficient metal-oxide [17]. The associative mechanism, on the other hand, has been reported at low as well as high temperatures, usually proceeds by the adsorption of CO and H 2 O on the catalysts surface leading to a reactive intermediate that subsequently decomposes to produce CO 2 and H 2 [18]. Nonetheless, the redox mechanism is still the most widely accepted mechanism for WGSR. Theoretically, Sun et al. combined DFT calculations with micro-kinetic analysis on Au/TiO 2 catalysts to conclude that the redox mechanism was the most plausible route for WGS [19]. Ammal and Heyden also reached the same outcome from their investigation on Pt/TiO 2 (110) catalysts, that both Pt clusters and single cationic Pt follow the redox mechanism [20]. This mechanism involves CO oxidation through obtaining atomic O from H 2 O, via either two consecutive steps of H abstraction, or one H abstraction followed by two OH species disproportionation [21,22]: In this mechanism, the surface of the catalyst is successively oxidized by H 2 O and then reduced by CO. Low temperatures are reported to be more favorable for hydrogen production, which is the primary objective and focus of this review article.

Catalysts in low-temperature water-gas shift catalysts
In order to facilitate the WGS reaction for a large scale of hydrogen production, choosing a suitable catalyst is important, which plays a central role in accelerating the reaction. A good catalyst would facilitate the adsorption of reactants on the surface of catalyst, their reaction, and desorption of the products to regenerate the active sites for the cyclic process of adsorption-reaction-desorption with new reactant molecules. An optimum interaction between gaseous molecules and catalyst surface is required as a weak interaction would result in low conversion due to quick desorption of the reactants, whereas a strong interaction could also result in low conversion and catalyst deactivation as the adsorbed molecules are too strongly bound to the surface leading to catalyst poisoning. These interactions are best understood in terms of volcano plots highlighting the optimum interactions leading to highest conversion of reactants. Figure 1 shows a 3D volcano plot for WGSR rate over the surfaces of transition metals as a function of the binding energies of atomic oxygen and carbon monoxide, formed on chemisorption-energy amounts on step sites. The presented model offers a good qualitative estimate of the relative order of activity of selected transition metal catalysts compared to the experiments. The figure reveals Cu to be the most active among the investigated catalysts, by being at the center with optimum adsorption energies for O and CO interactions [23].
Availability of oxygen vacancies, optimum strength for CO adsorption, and activity for water dissociation are some common characteristic roles of WGS catalysts [24]. In general, an ideal catalyst used for driving low-temperature WGS reactions should be cost-effective and CO-tolerant for large-scale uninterrupted operations as CO is known to have a poisoning effect on catalysts, even at an insignificantly low concentrations [25]. Besides, in terms of energy efficiency, the appropriate catalysts must be thermally stable and active at temperatures below 250°C to provide reasonably high reaction rates. According to the literature, a wide variety of catalysts have been employed to reduce the overall cost of hydrogen production and enhance the low-temperature WGS reaction catalytic activity and stability. In the following sections, a comprehensive summary of the recent development on different types of catalysts used in low-temperature WGS reaction is presented. It includes the catalyst preparation method, the operation conditions, nature of catalyst support, along with the efficiency for hydrogen production. Our particular focus is on the nanostructure of the catalyst, which has not been extensively covered in previously reported reviews. The catalysts are classified based on the metals used as well as the type of support.

Noble metal based
For commercial applications, any WGS catalysts should be free from some inherent disadvantages such as low stability and activity, along with the need for specific pretreatment and regeneration [26]. Chemically speaking, noble metals are well-known to be most resistant against corrosion and oxidation. Platinum (Pt), rhodium (Rh), gold (Au), palladium (Pd), and silver (Ag) are the commonly known noble metals, which have been extensively studied in recent years for WGS applications. The catalyst activity of the noble metals depend on many variables such as the method for catalyst preparation, noble metal amount, the support used, the amount, and characteristics of the prompter in case a promoter is used [27]. In multicomponent catalysts, the complex interactions between the metals and support also have a significant impact on the catalytic activity. Moreover, in the operation phase, the experimental variables like the space time, the reaction temperature, and the feed composition are influential [28]. The WGS catalysts activity of this group is known to depend on crystal size, dispersion, metal loading, metal oxide support, and the nature of the dopant used [27]; however, some contrary reports also exist indicating the activity of Pt/TiO 2 and Ru/TiO 2 not dependent on the metal loading or the crystallite size [29,30].
Reducible oxides are reported to significantly improve the catalyst activity at low temperatures. Vecchietti et al. discussed that water dissociation seems to be a critical step for reaction to move forward [31]. They studied the activation of water on oxygen vacancies on both pure as well as galliumdoped ceria supported platinum catalysts. The results showed a reverse correlation between the activity of the WGS catalyst and oxygen vacancies number [31]. In another study conducted by Tiwari et al., Pt nanoparticles with nanocrystalline CeO 2 support were investigated for low-temperature WGS reaction [32]. The 0.9Pt-CeO 2 (with 0.9 wt% Pt) catalyst showed the best activity (85.1%) with 1.66 × 10 10 mol cm −1 s −1 CO 2 formation rate at 140°C and 20,000 GHSV [32]. The nanostructure of the catalyst is shown in Fig. 2a-d, with HRTEM in Fig. 2c showing Pt crystallite size of 0.8 nm on spherical CeO 2 support of size between 40 and 50 nm [32]. Increasing the wt% of Pt resulted in an increase in Pt crystallite size as well as a lower surface dispersion of Pt of CeO 2 , negatively affecting the overall catalytic performance, providing conclusive evidence of higher interaction between Pt and CeO 2 resulting in improved WGS activity [32].
Nguyen-Phan et al. investigated the performance of both Au-and Pt-based metal catalysts on hydrogen production through various conversion processes like WGS reaction [37]. The activity of Pt was concluded to be much higher compared to Au catalysts, where Pt was also susceptible to quick deactivation, possibly by CO poisoning or oxidation by water molecules [37]. Palma et al. presented a strategy to alleviate the challenges of Pt supported on Ce/Zr, suffering from deactivation caused by sintering and coking as well as high selectivity of methane [38]. They used wet impregnation method to synthesize supported bimetallic catalysts (Pt/M/ CeZrO 4 (M = Na, Mo, Sn, Cu, Zn)) and compared their performance with monometallic ones. According to the result, in 473-675 K temperature range, 1Pt/1Sn/CeZrO 4 and 1Pt/1Na/ CeZrO 4 showed similar CO conversion as Pt-based monometallic catalysts, however, with higher hydrogen production and reduced methane formation than the bimetallic systems. In general, the 1Pt/1Sn/CeZrO 4 catalyst had the best conversion and hydrogen selectivity at 480-673 K. At lower temperatures, however, the monometallic 1Pt/CeZrO 4 catalyst exhibited the best performance [38].
The use of supercritical fluid deposition (SCFD) is considered a promising catalyst preparation method, which has been recently highlighted in some studies. Deal et al. synthesized platinum catalysts on alumina support, both in the presence and absence of ceria by this method and tested them for WGS reaction [39]. The result of the analysis showed Pt to be in a highly dispersed form on the support. The CeO x /Al 2 O 3 pretreatment in H 2 before Pt deposition resulted in more Pt nucleation on Ce than in the non-pretreated support. However, the non-pretreated Ce-containing catalyst showed a more uniform particle size with the Pt nanoparticles enclosed in the crystallites of ceria. No methanation or deposition of carbon was witnessed in any of the experiments. Measuring the reaction rate confirmed that the catalyst was more active for WGS, with the rates per Pt mass exceeding the values reported in most literature for WGS reaction on the Pt-CeO x catalysts. The high catalytic activity was attributed to the considerable amount of interfacial platinum/ceria contacts due to unique nanoscale interactions [39]. Kaftan et al. investigated KOHcoating Pt/alumina catalytic performance on WGS reaction [40]. They found that both activity and selectivity of the catalyst improved compared to the uncoated catalyst. Besides, the selectivity of CO 2 increased in the coated sample because of hydroxides and carbonates film formation on the catalyst surface [40]. González-Castaño et al. introduced a precatalytic buffer layer of CeO 2 /Al 2 O 3 oxides in Pt/CeO 2 / Fig. 1 3D volcano plot of the turnover frequency for the lowtemperature WGSR vs O and CO adsorption energies [23] Al 2 O 3 catalytic system for WGS reaction [41]. The inclusion of the buffer layer did not affect the CO oxidation; however, it increased the number of engaging sites during the reaction in the water dissociation phase leading to a catalytic activity enhancement in real conditions [41].
Recently, Pt-Re bimetallic clusters on TiO 2 support were reported to be a superior catalyst than Pt cluster alone for the low-temperature WGSR [42]. In order to understand the impact of both the Pt-Re clusters composition as well as Re oxidation, Duke et al. investigated a model system with bimetallic clusters grown on the surface of a rutile TiO 2 (110). The results showed that pre-oxidized clusters had lower activity in comparison to both unoxidized Pt-Re and pure Pt, which indicated that ReO x sites are not active in the WGS reaction. Reduced CO poisoning in Pt-Re clusters demonstrated a higher WGS activity of the bimetallic clusters [42]. Synergistic effects at the interface of the metal-carbide are reported to cause a chemical activity enhancement for pure Pt, TiC, and MoC [43]. Rodriguez et al. found a correlation between Pt/MoC and Pt/TiC (001) water dissociation capability and their activity in the low-temperature WGS reaction. They found that both of these catalysts are highly active, stable, and selective with excellent catalytic performance. Moreover, small Pt coverage on the carbide substrates had the highest activity among the tested catalysts [43].
Pt nanoparticles with TiO 2 support have also been considered as an appropriate catalyst for the low-temperature WGS reaction, which exhibited the potential to be employed in a fuel cell system [33]. Zhao et al. published a synthesis strategy that affords uniform particles of Pt/TiO 2 with Pt nanostructures tunable size of around 1.0 to 2.6 nm incorporated in TiO 2 nanoshells of a thickness of ∼ 3-5 nm and diameter of ∼ 32 nm (Fig. 2e, f). These yolk/shell nanostructured catalysts were found to have remarkable stability, and their activity monotonically increased with the Pt nanoparticles size for the low-temperature WGS [33]. Rivero-Crespo et al. used water clusters to isolate and stabilize single atom Pt 1 1+ that showed activity for WGSR at a low temperature of 50°C [44]. They suggested that the water cluster could help in regulating the metal charge that facilitates the adsorption of reactants resulting in efficient low-temperature WGS catalysis, where a double water attack mechanism generates CO 2 with both oxygen coming from water [44].
The performance of Pt catalysts loaded on different supports, Mo 2 C/η-Al 2 O 3 , Mo 2 C/γ-Al 2 O 3 , or Mo 2 C, were tested by Osman et al. for low-temperature WGS reaction [34], where Pt/Mo 2 C/η-Al 2 O 3 was found to be a promising catalyst with 44% CO conversion at 180°C. However, 4 wt% Pt-Mo 2 C (Fig. 2g) presented the highest catalytic activity and stable performance with 50% conversion at 180°C and no  [32]. e, f HAADF-STEM images of the Pt@TiO 2 yolk/shell nanospheres and the Pt nanoparticles size distribution in the hollow TiO 2 nanospheres. Along with synthesis of the Pt@TiO 2 yolk/shell nanospheres [33]. g SEM image of the pure Mo 2 C catalyst [34]. h, i Bright-field and HRTEM images of Mn 2.94 Pt 0.06 O 4-δ [35]. j, k AC-HAADF-STEM images of Pt/Fe-0.05 and HRTEM image of Pt/Fe-0.05 [36] deactivation over 85 h of time on stream (TOS) at 250°C temperature, maintaining over 70% CO conversion [34]. This improved performance was attributed to a uniform distribution of smaller Pt nanoparticles on Mo 2 C structure compared to other supports. Recently, sonochemical synthesis has emerged as one of the useful methods for synthesizing metal oxides [45]. Anil  , with the possibility of single-atom active sites (Fig. 2h, i), revealed excellent activity, with 99.9% conversion at 260°C and no methane formation, owing to a strong interaction existing between Pt and Mn 3 O 4 [35].
In order to design a metal-based catalyst with a high performance, identifying the effect of size at the atomic level is important. A comparative study on a series of Pt/FeO x catalysts with two types of Pt, one as nanoparticles and the other as single atoms, was conducted by Chen et al. to evaluate the catalytic performance for low-temperature WGS reaction [36]. The results exhibited lower activation energy by Pt downsizing, and Pt single atom with 0.05 wt% loading (Fig. 2j, k) was found to be the most active catalyst with a CO conversion of~65% at 300°C [36]. In a theoretical study conducted by Fajín et al., the platinum nanotube performance was analyzed for the WGS reaction [46]. The authors suggested that Pt nanotube showed improved performance compared to the extended platinum surfaces or Pt nanoparticles deposited on metal oxides. Pt nanotubes were identified to be structurally more stable and avoided catalysts sintering in WGS reaction [46].
Pd is another well-known noble metal widely investigated for WGS reaction, as a single metal or in alloyed forms with other metals. Pd with different Co 3 O 4 catalysts was investigated by Kono et al. for water-gas shift (WGS) reaction to produce hydrogen [48]. Palladium alone on cobalt oxide (Pd/Co 3 O 4 ) exhibited low activity and higher methane selectivity; however, the Pd/K/Co 3 O 4 as a catalyst with potassium loading > 0.78 wt% showed high activity for the WGS reaction. The potassium loading affected the CO species adsorption state, which caused Pd/K/Co 3 O 4 catalyst to have high activity [48]. Sun et al. synthesized a series of Pd catalysts supported on FeO x for WGS reaction and used Pd/Al 2 O 3 as a reference for comparison [47]. They found that Pd/FeO x showed a higher CO conversion compared to Pd/Al 2 O 3 (Fig. 3a, b), and with high stability, even in the presence of CO 2 and H 2 . Moreover, Pd single atoms can greatly enhance FeO x reducibility and promote oxygen vacancies formation (Fig. 3c, d) [47].
Bulk gold is thought to be chemically inert; nonetheless, gold nanoparticles have revealed a high catalyst activity in the WGS reaction [49]. Bimetals of gold (Au-M, M = metals) on CeO 2 support were evaluated for WGS reaction where 3 wt% Au-Pt/CeO 2 showed high activity within a temperature range of 100-300°C [50]. The WGS activity correlated well with the reducibility of ceria, that in turn was affected by the bimetallic cluster promoted local electronic band causing the stabilization of germinal OH groups, which were supposed to be responsible for enhanced WGS activity [50]. A comparative study was conducted by Castaño et al. on the performance of gold and platinum catalysts on ceria support [51]. Gold-based catalysts indeed showed a higher reaction rate at 180°C; however, the activity of the catalysts was sensitive to temperature with Pt showing more activity at the higher temperatures. The gold catalysts had a high dispersion of 32% with an average crystallite size of 4 nm (Fig. 4a), that enabled Au to influence the electronic properties of the ceria support, which correlated with the activity for WGS reaction [51]. Later, the same group carried out a comparative study between Pt and structured gold catalysts, both supported on Fe doped ceria mixed oxide inserted onto metallic micro size monolithic structures (Fig.  4b, c) in order to find their performance in WGS reaction with O 2 assistance [52]. Pt was the most active metal in conventional conditions; nevertheless, in O 2 assisted WGS reaction, Au-based catalysts performed superior to Pt-based catalysts [52].
To understand the effect of the preparation method and the amount of Au on the catalyst activity in low-temperature WGS reaction, Soria et al. synthesized a series of nano-sized Au/Fe 2 O 3 catalysts with several Au loadings and different synthesis methods [53]. The deposition-precipitation (DP) method with the Au content of 5 wt% (the highest tested loading) (Fig. 4d, e) was found to have the highest conversion of CO. Among the Au/Fe 2 O 3 liquid phase reductive deposition (LPRD) samples, the most active catalyst was the one with the lowest quantity of Au (1.5 wt%) (Fig. 4f, g). Based on H 2 -TPR analysis, the authors concluded that gold could promote Fe 2 O 3 support reducibility, which seems to be a key factor in enhancing redox WGS reaction activity. In both, the DP and the LPRD methods, high dispersion of gold over the iron support with 2.2-3.1 nm crystallites was observed, and only an insignificant increase in the average particle size of used catalyst was witnessed after the reaction [53].
In another study, Reina et al. synthesized multicomponent catalysts including Au/Ce 1-x Cu x O 2 /Al 2 O 3 and tested for WGS reaction to understand the role of each element; Au, Ce, and Cu; in the reaction [55]. The synergy between Cu-Ce and Ausupport interactions greatly influenced the catalytic properties, and oxygen storage capacity. Though copper is an active catalyst for WGS reaction, gold was required to achieve a high conversion of CO. CeO 2 also played an important role, as both Cu and Au showed remarkable improvement in activity in the presence of CeO 2 with reasonable long-term stability [55]. Some studies have discussed that using bimetallic systems could improve both the activity and stability of the catalyst in the WGS reaction. For example, adding rhenium (Re) to Au catalysts with CeO 2 support significantly improved the catalytic activity [56]. Ozyonum et al. conducted an investigation on monolithic Au-Re/CeO 2 over microchannel cordierite [57]. They realized a relatively lower performance with Au/CeO 2 catalyst in comparison to the considerable improvement with Re addition as a promoter [57]. For achieving a high activity with Au-based catalysts for low-temperature WGS reaction, Yao et al. incorporated atomic layer deposition technique to grow clusters of layered gold on a molybdenum carbide (α-MoC) substrate to create an interfacial catalytic system for ultra-low-temperature WGS (Fig. 4h, i) [44]. This catalytic system was capable of dissociating water efficiently on the MoC surface as well as activating the CO on neighboring Ag clusters, resulting in high WGS activity at low temperature. Interestingly, even after the reaction, the sample still included both single-atom Au and layered Au clusters (Fig.  4j). The use of sodium cyanide (NaCN) solution has been effective in selectively leaching the layered Au clusters from in such systems (Fig. 4k) [54,58].
Santos et al. examined the activity of gold nanoparticles catalyst combined with Cu/ZnO/Al 2 O 3 for hydrogen production through WGS reaction [59]. This system could successfully achieve a full CO conversion at 180°C due to the perfect synergy of gold and copper [59]. Fu et al. synthesized two different kinds of Au/Ce WGS reaction catalysts supported on the same nanorods of ceria, one with < 2 nm disordered clusters, and the other with 3-4-nm particles. The results exhibited that the disordered gold clusters had superior activity than the nanoparticles of gold [60]. Stere et al. utilized nonthermal (cold) plasma activation to facilitate low-temperature WGS reaction over an Au-based/CeZrO 4 catalyst, where they concluded that the catalyst activity was comparable to that achieved through heating the catalyst up to 180°C [61]. Santos et al. performed an investigation to optimize the performance of the Au/Cu-ZnOAl 2 O 3 catalytic system by applying various ratios of active components as well as the Au preparation methods [62]. This system not only had a high carbon monoxide conversion but also showed stability during the start/stop operations as well as for long-term TOS, which is a prerequisite in real applications. It was concluded that combining Au with commercial-like Cu-Zn-Al compounds by hydrotalcite process can generate a new series of catalysts for WGS reaction with the capability to develop large scale hydrogen production technologies [62]. Au-based catalyst supported on Ce-Zr has been considered as one of the most active low-temperature WGS catalysts recorded to date; however, in the reaction conditions, a rapid deactivation occurs [63,64]. Carter et al. evaluated a series of Ce-Ti support that were synthesized by the sol-gel method. They supported Au on these mixed metal oxides via a deposition-precipitation method, intending to resolve the deactivation issue. The   [47] results of low-temperature WGS reaction showed high activity and stability for Au/Ce 0.2 Ti 0.8 O 2 compared to ceriazirconia due to the highest specific surface area represented by amorphous Ce 0.2 Ti 0.8 O 2 and well-dispersed of Au species [65].
The use of a suitable Rh single-atom catalyst (SAC) could make the WGS performance very active and selective in terms of adsorption and reaction. SAC is now considered a new dimension in heterogeneous catalysis that presents a very high dispersion of metal species, in atomic form, and reported to display optimum catalytic properties in many cases [67]. Guan et al. reported a Rh 1 /TiO 2 SAC with a loading amount of 0.37 wt%, resulting in an overall 95% CO conversion with no methanation at 300°C, which is viable even in a WGS stream rich with CO 2 and H 2 [66]. The result suggested that Rh single atoms enhanced the oxygen vacancies formation on TiO 2 support in order to generate more H 2 than Rh clusters. Besides, the SAC activity was about four times higher than the cluster catalyst with no formation of methane. The single atoms of Rh and nanoclusters were verified by HRTEM, as shown in Fig. 5 [66]. Mandapaka et al. synthesized Rh and aluminum co-doped ceria catalysts by a single-step solution combustion synthesis technique and tested its reactivity for WGS reaction and CO oxidation. The results showed that the bimetallic Rh and Al doped catalyst yielded a higher activity and surface area than the single metal doping of Rh/Ce and Al/Ce synthesized by the same method [68].
The catalytic performance in heterogeneous catalysts is affected by the dispersion of the active precious metal, and frequently, the active sites coalesce and grow during the reaction that result in lower activity over a longer time of analysis. Homogeneous catalysts, on the other hand, are known for high turnover frequency due to availability of atomic level sites. Ruthenium-based homogeneous catalysts have been particularly investigated for WGS reaction. Metal carbonyls, e.g., ruthenium-carbonyl, catalysts are low-cost homogeneous catalysts, which have been reported to operate at a lower temperature than the heterogeneous catalyst systems resulting in higher equilibrium conversion [70]. However, these systems require a high CO pressure, above 10 bar, and in this regard, volatile solvents and ionic liquid (IL) have been considered. Huang et al. reported that Ru-iongel catalysts show better activity and stability for low-temperature WGS reaction in comparison to the catalysts with supported ionic liquid (IL) phase, possibly because of a strong interaction between silica support and IL. The iongel catalyst activity increased with Fig. 4 a HAADF-STEM images of Au/CeFeAl [51]. b, c SEM microphotographs of the micro monoliths Au/CeFe catalyst: b front; c crosssection view [52]. HRTEM images of d, e the 5Au/Fe 2 O 3 DP fresh and used (f, g) 1.5Au/Fe 2 O 3 LPRD fresh and used catalysts, respectively [39]. HAADF-STEM images of h, i fresh (2%) Au/α-MoC, j used (2%) Au/α-MoC, and k the (2%) Au/α-MoC catalyst after NaCN leaching, representing predominantly Au single atoms in which most of them overlapped with Mo sites in the support lattice (very bright features were created by overlapping MoC particles. (Blue circles: Au single atoms, yellow highlights: layered Au structures) [54] increasing the IL loading, and the larger anions iongel catalysts showed higher activity in the WGS reaction [71]. The effect of adding K 2 CO 3 on the Ru/C catalysts performance ( Fig. 6a) was investigated by Liu et al. through the positive reduction method by three different agents, NaBH 4 , ethylene glycol (EG), and H 2 , in which the Ru/C catalyst reduced by H 2 exhibited the highest catalytic activity for WGS reaction (Fig.  6b). The alkali additives showed a strong adsorbed gas bonding that improved the catalytic activity of the Ru/C. It also resulted in a concentration balance of water and CO on the active sites [69]. Queiroz et al. likewise studied the performance of 2% Ru/C for the low-temperature WGS reaction in a fixed bed reactor that resulted in 80% CO conversion under the steady-state condition at 553 K [72]. Later, they examined the catalytic performance of Ru during the low-temperature WGS reaction with two different supports, TiO 2 and Al 2 O 3 . The result showed that at 573 K, CO conversion on both Ru/ Al 2 O 3 and Ru/TiO 2 catalysts stayed 85% in a steady-state condition [73]. Queiroz et al. developed a WGS process with a membrane reactor using 2 wt% Ru on TiO 2 catalyst at 453-573 K and atmospheric pressure [74]. The result of the experiments at 573 K, both with and without permeation, showed 90 and 75% CO conversion, respectively. Besides, the results showed a 75-96% hydrogen recovery in the permeation operation, which was~5 times higher than the operation without hydrogen permeation [74]. Supported ionic liquid phase (SILP) materials have been used to perform reactions at a low-temperature range [75,76]. Stepic et al. successfully used this approach to enable a highly effective, homogeneously dispersed Ru-based catalyst to perform WGS reaction in the temperature range of 100°C and 150°C [77].
From this section, it is clear that noble metals have been extensively studied for WGSR. The dispersion of the active metal and the reducibility of the metal oxide support seem to be the key parameters affecting a catalyst's performance. The use of the SCFD method, as well as sonochemical preparation techniques, seems to offer a good control over nanoscale dispersion of the noble metals on the support leading high turnover numbers. In the case of Pt-containing supported catalyst, washing with a base, like KOH, helps in improving the catalyst performance for WGSR. The presence of gold, alone or with other noble metals, is shown to affect the reducibility of CeO 2 as well as FeO x , which in turn positively affects their performance. Single-atom catalysts of Pt and Au supported on CeO 2 perform exceptionally well, which can further be improved by promoting the catalyst with single-atom sites of Re.
Next, we review the progress in transition metal catalysts, which offer economic alternatives to the noble metals.   [66] Supported catalysts composed of Cu, Zn, Ni, and Co, alone or alloyed, have been investigated broadly for WGSR and offer to have good activity and selectivity for the low-temperature reaction.

Transition metal catalysts
Transition metals are elements from groups 3 to 12 in the periodic table, and they are known for their catalytic properties for many reforming reactions. Inclusion of transition metals such as Fe, Cu, and Zn as dopants in CeO 2 for gold supported ceria catalysts (Au/CeO 2 /Al 2 O 3 ) is reported to improve the performance for WGSR [78]. Transition metals tend to increase the redox properties and promote structural stability. Herein, we present some of the transition metal-based catalysts for WGSR.
Previous sections indicate that Au nanoparticles act as suitable catalysts for the oxidation of CO; however, some authors argue that water dissociation and adequate WGS reaction rate on pure Au are challenging to achieve. Copper, on the other hand, is reported to be more active for WGS reaction, although a significant barrier in Cu activation is required to be overcome for water dissociation [79]. Saqlain et al. studied the behavior of two different catalysts, Cu (100), and bimetallic Cu-Au (100) [80]. Results indicated that the dissociation of water for the Cu (100) and bimetallic surfaces were spontaneous up to 229 K and 520 K, respectively. In terms of reactivity, the study suggested that the bimetallic surface was far more reactive compared to the Cu (100) surface [80]. In another study, Wijayapala et al. used catalyst systems Mo/Co/K/ ZSM-5 and Mo/Ni/K/ZSM-5 (ZSM-5 = zeolite), alone and with a copper-based WGS catalyst, for CO/H 2 ratios conversion in a batch reactor into aromatic hydrocarbons. The presence of Cu in the WGS catalyst system significantly improved CO conversion from 25 to 90% at 280°C [81]. Jeong et al. reported a strong effect of support on catalytic activity in copper-based WGS catalyst [82]. In this regard, they tested Cu/CeO 2 catalysts prepared by incipient wetness and coprecipitation methods and optimized the copper loading in order to get a highly active Cu/CeO 2 catalyst. Results indicate that the catalysts prepared by co-precipitation ( Fig. 7a) had superior performance with highest CO conversion and 100% CO 2 selectivity in addition to easier reducibility [82]. Camara et al. investigated the low-temperature WGSR mechanism on CeO 2 /Cu catalysts using operando SSITKA-DRIFTS-mass spectrometry, where they identified a specific type of carbonate intermediate (bi or tridentate), which was believed to be the rate-limiting intermediate controlling the formation of CO 2 [83]. Moreira et al. studied the sorption enhanced WGS reaction at low-temperature (125-295°C) over Cu-CeO 2 / HTlc catalysts, where they reported a 70% volume enrichment in hydrogen and improved CO conversion upon efficient removal of CO 2 from the effluent stream [84]. Cu supported on nanoparticles size polyhedral ceria, with 87.6% CO conversion, was found to have the highest activity in comparison to other catalysts reported in this work [84].
Jeong et al. prepared Cu-CeO 2 , Cu-ZrO 2 , and Cu-CeO 2 -ZrO 2 to obtain feasible catalysts for low-temperature (200 to 400°C) WGS reaction. The cubic Cu-Ce 0.8 Zr 0.2 O 2 catalyst showed the highest CO conversion maintaining a stable TOS performance for 100 h with no significant loss in the activity [85]. Price et al. proposed a design strategy to overcome the limitations of large residence time requirement for Cu-based catalysts in WGS reaction [86]. They synthesized a series of Cu-ZnO catalysts with different supports and evaluated their performance. The results indicate that CeO 2 -Al 2 O 3 support can work under medium to high space velocities while maintaining a high activity and long-term stability [86]. Lang et al. introduced copper/ceria/foam catalysts to perform WGS reaction at a temperature range of 150 to 300°C and GHSV between 3600 and 9500 h −1 at the outlet of a biomass gasifier with a limited pressure drop. The optimum amount of Cu and Ce resulting in the highest catalytic activity were 5.5 and 9.0 wt% respectively on a 30 ppi porosity foam [87]. Dalin Li et al. developed Cu/ZnO/Al 2 O 3 catalysts with various compositions of Cu-Zn-Al layered double hydroxides (LDHs) for WGS reaction [88]. The formation of 2-6 nm size Cu with a dispersion percentage of 18-48% was observed after the reduction. The results showed that the 30%Cu/Zn 1 Al catalyst had the highest activity, thermal stability, and long-time catalytic stability compared to the commercial catalyst of Cu/ZnO/ Al 2 O 3 [88]. Zhang et al. used uniform nanocrystals of Cu to investigate the Cu actives site for low-temperature WGSR [89]. The results indicated that Cu cubes with (100) exposed facets, in contrast to Cu octahedral enclosed with (111) facets, are active up to 548 K. Therefore, it was concluded from this study that Cu cubes-supported on ZnO had an extremely high activity for low-temperature WGSR [89]. Yan et al. reported a bulk promoted Cu-Fe 3 O 4 catalysts for low-temperature WGS reaction, where the interaction between Cu and Fe helped in stabilizing and significantly improving the Cu 0 dispersion that resulted in an increased activity for WGS CO conversion. The Fe addition promoted both CO and H 2 O adsorption. Different compositions of Cu and Fe were investigated and the Cu 0.3 Fe 0.7 O x catalyst showed a high initial activity that deactivated after 26 h; thereafter, Cu 0.7 Fe 0.3 O x surpassed the Cu 0.3 Fe 0.7 O x activity. Moreover, the inclusion of alumina in the multicomponent Cu-Fe 3 O 4 -Al 2 O 3 catalyst resulted in superior performance, improved activity, and long-term stability compared to the catalysts without alumina in the structure (Fig. 7b) [90].
Coal gasification, which inherently involves WGS, is considered a clean hydrogen production method to convert coal resources into chemical energy; however, the process is still costly. In this regard, Zhao et al. adopted a solar-driven WGS method to significantly decrease the consumption of energy [91]. They found that the CuO x /Al 2 O 3 (Fig. 7c) (Fig. 7d), which showed good activity for high-temperature WGS reaction but did not show much activity at low temperatures [92]. The authors conducted ambient pressure XPS (AP-XPS) analysis and identified Cu + as the dominant, and possibly active phase, for WGSR.
To investigate the role of preparation method on the structure and catalytic activity in WGSR, Farzanfar et al. synthesized Cu-Mn/SiO 2 catalysts with three different methods [93]. The results determined that thermal decomposition of the inorganic complex was more suitable compared to impregnation and co-precipitation methods for developing active and stable Cu-Mn/SiO 2 WGS catalysts for 180-320°C temperature range [93].
Kowalik et al. studied the impact of the hydroxycarbonateassisted precipitation media on the physicochemical properties and activity of Cu/ZnO/Al 2 O 3 system in low-temperature WGS reaction [95]. They employed water, glycol, and aqueous ethanol solution as the reactive medium. The results revealed that the alcohol assisted method could lead to a Cu/ ZnO/Al 2 O 3 catalysts with higher catalytic activity and stability than the other conventional techniques [95]. A comprehensive investigation was performed by Yin et al. to illustrate the beneficial characteristics of a zeolitic imidazolate framework-8 (ZIF-8) reactor (Fig. 7e, f) and Cu/Zn/Al 2 O 3 catalysts for a low-temperature WGS reaction [94]. The membrane reactor exhibited 9.2 × 10 −7 mol/m 2 s·Pa hydrogen permeance, which was much higher than the conventional metal and zeolite membranes, making it more attractive for practical hydrogen production. In addition, at a low temperature of 120-220°C, the CO conversion using the ZIF-8-based membrane reactor was 13.5% higher than the conventional packed reactor [94]. Despite the high gas permeance of ZIF-8 membrane with considerable H 2 perm-selectivity, due to the low hydrothermal stability of materials used in this membrane, the system needs further improvements to make it more suitable for WGS reaction conditions [96,97].
To evaluate the Fe-doping effects on the catalyst performance, Chen et al. synthesized Fe-Ce-O x composite oxides, consisting of CeO 2 nanorods and Fe 2 O 3 nanoparticles, and applied them in the WGS reaction [98]. The results of the analysis revealed that introducing Fe 3+ into CeO 2 crystal lattice structure increased the catalytic activity for the WGS reaction. It was also concluded that the reducing gas atmosphere synthesis method was more favorable for enhancing the doping effect. However, the catalyst was more active at high temperature and the CO conversion increased when the temperature was increased from 300 to 500°C [98]. The effect of Fe-  [92]. e, f The cross-sectional SEM images of the ZIF-8 membrane e before and f after use [94] Ce-O x composite oxides had also been studied by other groups showing high catalytic activity towards WGS reaction [99].
With the objective of studying the impact of alkali-I metals promotion, such as Li, Na, K, Rb, and Cs, on both activity and stability of Co 2 C for the low-temperature WGS reaction, Gnanamani et al. performed a study at 453 to 573 K and atmospheric pressure and compared the results with the unpromoted Co 2 C reference catalyst [99]. The WGS reaction results indicated the unpromoted cobalt carbide to be active, although the catalyst deactivated over longer TOS due to the chemical transition of carbides to metallic form during the WGS reaction. However, the potassium promoted catalyst presented higher activity and better stability compared to the rest of the alkali supported Co 2 C catalysts [100].
Chen et al. studied the intrinsic activities of Co-promoted MoS 2 and unpromoted-MoS 2 on Co-MoS 2 /Al 2 O 3 for WGSR, and reported a lower activation energy for Co-promoted sites [101]. The Co-promoted MoS 2 /Al 2 O 3 is widely used in hydrogen production industries as a catalyst for WGS reaction in the presence of sulfur [101].
Recently, Alamolhoda et al. worked on understanding the synergetic effect between nickel and cerium as catalysts in WGS reaction, and reported the use of low percentage loading (0-3 wt%) of cerium and nickel as dispersed phases on MFI framework [102]. According to the results, the cerium only catalysts had shown no sign of reaction, whereas the single metal Ni-MFI series catalysts successfully activated the CO. The prepared Ni-Ce-MFI succession had together speeded up the WGSR even at a low temperature of 503 K, and in some cases, the equilibrium conversion was achieved at temperatures less than 548 K. The presence of ceria is believed to provide the oxygen required for promoting nickel activity, which was clear from the TPR profile requiring lower temperature for nickel reduction in Ni-Ce-MFI catalysts. The lowest reduction temperature was observed when Ni and Ce were present in equal weight percentages [102]. Iriarte-Velasco et al. used calcined pork bone, composed mainly of hydroxyapatite (HAp), as support for transition metals for WGSR [103]. Both activity and selectivity were studied and correlated with the catalytic properties. The catalytic activity was observed to follow a trend: Ni > Co > Cu > Fe. Moreover, compared to the synthetic hydroxyapatite, the natural support showed a lower Ni methanation activity and increased longterm stability [103].
From this section, it is clear that transition metals can be a good choice for low-temperature WGS reaction. Bimetals and alloys, particularly of Cu and Zn, in some cases are reported to offer better performance than single metal catalysts. Transition metals normally enhance the redox properties of the catalyst that helps in improving the activity and longterm stability. Addition of alkali metals is also reported to enhance the activity by facilitating the adsorption and dissociation of the water molecule, which tend to be an energy-consuming step for most of the transition metals. Transition metal carbides (e.g., Co 2 C), sulfides (MoS 2 ) are also reported to perform well for low-temperature WGS reaction. Throughout the previous section related to catalysts, the function of support is found to be quite instrumental in defining the performance of the catalyst. The next section presents a review on the role and recent development in catalyst supports for WGSR. This part is divided into two categories, metal oxide supports and carbon-based supports, for the ease of reading.

Type of metal oxide support
Ceria has been broadly investigated as a support to disperse the noble metals active sites in WGS reaction. CeO 2 supported catalysts have shown effective performance with effluent gas streams resulting in low CO concentrations. Cerium is known to have dynamic oxidation states which can change easily during reaction conditions. Besides, it also has a high oxygen movement and vacancy [104]. Jain et al. investigated the potential of three distinct Pt/Ceria catalysts at 150-450°C with a GHSV of 13,360 h −1 to determine the impact of processing techniques and surface area, as well as porosity and crystallite size of ceria on the rate of WGS reaction [105]. In this work, the mesoporous ceria with a crystallite size of 5.8 nm and synthesized via sol-gel method was found to be the best catalyst with a complete conversion of CO at 175°C (Fig. 8a-c) [105]. Byun et al. prepared a series of Cu-ZnO-CeO 2 catalysts to investigate the influence of CeO 2 addition on lowtemperature WGS reaction performance [106]. These catalysts included a fixed amount of Cu (50 wt%) and different ceria content (from 0 to 40%), which affects the Cu dispersion and binding energy. The result showed that a 10 wt% cerium could promote catalyst reduction and CO conversion at low temperatures of 200-400°C [106]. Petallidou et al. also investigated the effect of three synthesis methods (sol-gel, pechini, and urea co-precipitation) on 0.5 wt% Pt/Ce 0 . 5 La 0.5 O 2-δ (Ce: La = 1:1) catalyst and used different characterization techniques, such as in situ Raman, temperature-programmed techniques (TPD-H 2 , TPD-NH 3 , TPD-CO 2 ), powder XRD, and oxygen storage capacity (OSC), to thoroughly analyze the properties of the synthesized catalysts. The results revealed that Pt/Ce 0 . 5 La 0.5 O 2-δ synthesized by urea co-precipitation had the highest CO conversion activity compared to the other two methods [107].
Schilling and Hess investigated CeO 2 support for Au catalyst using operando Raman spectroscopy to understand the role of bulk and surface oxygen in the WGS reaction [111]. They concluded that the presence of 0.5 wt% Au/CeO 2 catalyst could successfully reduce the Ce support under WGS conditions. Results also revealed the dominant mechanism of the reaction to be the redox-type [111]. The use of mesoporous ceria, both as a catalyst and as a support, is widely seen for different reactions due to a large amount of specific surface area, well-developed nanocrystalline framework, and excellent surface oxygen vacancy compared to the bulk form of ceria [112]. In a study conducted by Li et al., an ordered mesoporous series of ternary metal oxide Cu-Mn-Ce were synthesized using a nano-casting method [113]. It was found that among the tested catalysts, the Cu 0.18 Mn 0.02 Ce 0.8 O 2 mesoporous catalyst had the highest activity with almost 100% CO conversion for low-temperature WGS reaction without any methane production. Moreover, the catalyst was tested for 16 h TOS with no significant activity loss [113]. In order to understand the effect of modifying ceria by Y 2 O 3 rare earth on the WGS reaction activity of Au/ceria, a study was conducted by Tabakova et al., in which all the designed catalysts, except for Au 2.5 YCe synthesized by co-precipitation method, showed > 90% CO conversion at 180-220°C (Fig. 8d) [108].
Ceria with less than 10 nm particle size with exposure to WGS environment containing H 2 O show high densities of adsorbed -OH group resulting in hydroxylated ceria nanoparticles [114]. Huang et al. showed that hydroxylated ceria nanoparticles are active for WGS reaction without requiring any metal cocatalyst and have a small activation energy of 0.5 eV [115]. Na et al. proposed a single-step reverse precipitation preparation technique to control the physicochemical properties of nano-sized CeO 2 as support for low-temperature WGS reaction (Fig. 8e), which resulted in a high surface area support of 162.8 m 2 /g [109]. They compared the catalytic activity with a normal precipitation method synthesized CeO 2 by supporting 5 wt% Cu on both the supports. The results indicated that the one with the support prepared by reverse precipitation exhibited higher CO conversion than the CeO 2 prepared by normal precipitation method [109].
The ceria support morphology has a significant impact on the properties of the catalysts and can affect the catalytic dispersion, water dissociation, the metallic-base stability, and the metal particle size upon reduction during WGS reaction. In this regard, Yao et al. synthesized three different shapecontrolled CeO 2 nanostructured supports with CuO to catalyze WGSR [110]. The result showed that the nanospheres (NS) ceria support (Fig. 8f, g) had the highest activity and long-term stability in comparison to the nanorods (NR) and nanocubes (NC) [110].
Cámara et al. studied the effect of a novel manganesedoped inverse CeO 2 /CuO catalyst under WGS reaction conditions [116]. The presence of Mn along with Cu strongly influenced the physical and chemical properties of the ceria nanoparticles leading to high catalytic activity towards WGS reaction [116]. They also examined CeO 2 /CuO, Mn-doped CuO, and Zn-doped CuO for WGS catalytic performance, and the effect of O 2 and/or H 2 present in the reaction mixture. The result of this work demonstrated a positive impact of Mn and Zn doping on WGS performance; whereas for the undoped catalyst, CO conversion enhancement was seen in the presence of some oxygen at low temperature and high CO/ O 2 ratio [117].
Tabakova et al. conducted a study aiming to introduce an efficient, cost-effective catalyst of alumina-supported Cu-Mn mixed oxides (Fig. 9a, b) synthesized by wet impregnation, where Au nanoparticles promoters were prepared by the deposition-precipitation approach [118]. According to the Fig. 8 a-c Electron microscopy analysis of Pt/ceria catalysts prepared by sol-gel after WGS testing [104]. d Temperature dependence of the CO conversion over Au catalysts on Y-doped Ce synthesized by CP [107]. e TEM images of the Cu/CeO 2 catalysts prepared by reverse precipitation method [108]. f TEM and HRTEM images of CuO/CeO 2 (ns), g STEM and HRSTEM images of CuO/CeO 2 (ns) [109] result, the Au promotional effect was more demonstrated in a sample with a higher amount of Cu (Cu:Mn molar ratio of 2:1) due to the active presence of the two dispersed metallic phases. The high surface area of alumina modified by Cu-Mn mixed oxides surface fraction also favored the stabilization of dispersed gold nanoparticles. The high activity of this catalyst system makes it capable of practical usages, primarily due to economic viability as it consists of 80% alumina [118]. Sagata et al. investigated the effect of aluminum oxide and the pore size of the support on the catalytic activity during the low-temperature WGS reaction over Cu catalyst [119]. They concluded that by decreasing the mesopore size of Cu/Al 2 O 3 , the catalytic activity increased only when the steam/carbon (S/C) ratio was 2.2, whereas the catalytic activity increased in the S/C ratio of 4.6 by increasing the mesopore size [119]. One of the important considerations in an industrial sulfur tolerant WGS process is that the reaction conditions, not only catalysts, dictate many of the outcomes such as improvements in production capacity, decreasing production costs, and change in the catalytic activity [120,121]. For example, Liu et al. developed Mo-Co/alkali/Al 2 O 3 for sulfur tolerant WGS reaction that resulted in a steady operation even at 0.2-0.3 steam to gas ratio. Plant efficiency improvement, enhanced safety, and reduction in steam consumption during H 2 production were some of the key achievements [122].
Nano-structural composites of ceramometals (or cermets) can be attractive owing to reported catalytic property improvements compared to the common solid oxide supports. The ceramometals have excellent thermal conductivity and high strength as well as regulated mesoporous structure suitable for catalytic reactions to take place without much transport limitations [123]. Tikhov et al. synthesized porous CuAlO/ CuAl ceramometals and tested for WGS reaction, where they reported the granulated ceramometals catalytic activity to be comparable with granulated CuZnAl oxide, owing to a high diffusion permeability (Fig. 9d) [124]. In another study based on low-temperature WGS, they compared catalysts prepared by co-precipitation method and ceramometal CuAlO/CuAl catalysts that were made from CuAl alloy powder (Fig. 9c). The catalysts activity was tested at one bar, using a mixture of CO: Н 2 О:Н 2 = 8:42:50 and at a steam/gas proportion of 0.6-0.7 in both small (0.14-0.25 mm) and large (3.2 × 3.2 × 5 mm) portions. The result accordingly showed that the large fraction ceramometal catalyst had doubled the efficiency owing to a greater inner dissemination [125].
Longlong et al. studied the aluminum doping effect in Cu/ ZnO/Al 2 O 3 complex synthesized via modified coprecipitation method. The catalyst produced with this method containing 2.5% co-precipitated Al 3+ of total Cu and Zn atoms exhibited higher activity and stability in the WGS reaction compared to the commercial sample [126].
Silva et al. studied the impact of H 2 S on the platinum catalysts performance in WGS reaction. In this regard, they tested Pt-based catalysts with various oxides supports under real conditions, both with and without the presence of H 2 S, thereafter extended the experiment further in the absence of H 2 S to see if the catalyst performance can be recovered [127]. They found that zirconia incorporation to CeO 2 showed increased Fig. 9 a, b HRTEM micrograph of a fresh Au/CuMn (2:1) catalyst [117]. SEM images of c CuAl ceramometal and d CZA oxide catalyst [124] sulfur tolerance as well as preserved high activity. They concluded that in the presence of H 2 S, the Pt/Ce 0.25 Zr 0.75 O 2 showed the best performance in terms of lower deactivation, higher activity, and complete recovery, whereas under contaminated conditions, Pt/CeO 2 was not suitable anymore because of a higher deactivation rate [127].
Some reports indicate that the reducible oxides and solids providing oxygen vacancies can enhance the employed metal activity for the WGS reaction [128]. For example, García-Moncada et al. presented a multicomponent catalyst designed using different active phases (Cu, Pt, or Au) and different ionic conductors for WGS reaction [129]. The results of the study demonstrated that the ionic conductor, along with an appropriate catalyst, is necessary to the water supply as it potentiates activation of water as well as overall performance. The influence trend of ZrEu > MoEu > NbEu was observed for the ionic conductors, no matter the nature of the catalyst (Fig. 10a-c). The ZrEu ion was the most suitable conductor for the WGS reaction owing to the more oxygen vacancies provided by having a sub-stoichiometric structure [129]. A similar result was also presented by González-Castaño et al. in 2017 by using a bi-doped of Zr and Fe [130]. Yun and Guliants synthesized ZrO 2 -supported Mo sulfide (Mo-S/ ZrO 2 ) WGS catalysts using incipient wetness impregnation approach with different MoO 3 surface coverage (Fig. 10d-f). The CO conversion in WGS reaction over these catalysts increased with the presence of Mo content up to the monolayer coverage that showed a good correlation with the Mo sulfidation extent [131].
Molybdenum (Mo) and Molybdenum oxide (MoO x ) as promoters are reported to improve the performance of Ptbased WGS reaction catalysts (Fig. 11a) [132]. Pt-Mo/SiO 2 prepared with the impregnation-reduction technique have been used as effective catalysts in low-temperature (below 300°C) WGS reaction. Pt nanoparticles with Mo promotion showed higher activity than Pt/SiO 2 due to an improvement in Pt dispersion as well as a synergistic effect in Pt-Mo system that were found to correlate with the atomic ratio of Mo/Pt (Fig. 11b) [132]. Marras et al. used two different types of porous silica, namely aerogel and cubic mesostructured, as supports for Cu-based nanoparticles that were tested for WGS reaction at 200-350°C [133]. The catalytic performance was found to be influenced by the ability of the support matrices to disperse the copper nanoparticles. The results showed that the silica aerogel was a highly stable matrix throughout the catalytic cycles with a constant CO conversion [133].
TiO 2 is considered a good choice for catalyst support on the basis of as low-cost, non-toxic, and chemically stable nature. Some supports such as CeO 2 with variable oxidation states tend to be more effective only under specific conditions [134]. During the low-temperature WGSR, some metals, e.g., gold nanoparticles, tend to agglomerate even on rigid support like TiO 2 which has a higher ability to disperse them; nonetheless, this effect is reported to depend on the catalyst synthesis method as well [135]. For example, in a study conducted by Pérez et al., Au/TiO 2 was synthesized by different methods and Au loading was also controlled by three methods (double impregnation method (DIM), deposition-precipitation (DP), and liquid phase reductive deposition (LPRD)). The catalysts were evaluated for low-temperature WGSR and compared with the commercial sample provided by the World Gold Council (WGC) (Fig. 12a). The results revealed that the Au/ TiO 2 synthesized by the DP (Fig. 12b)  different Pd molar ratio and tested their catalytic activity towards the low-temperature WGS reaction [136]. All the samples showed almost zero methane formation, whereas the inclusion of K in Pd caused a reduction in sintering, and supporting the Pd on smaller TiO 2 particle size resulted in better catalytic performance [136].
Apart from the well-known oxides, such as Al 2 O 3 , SiO 2 , ZrO 2 , CeO 2 , and TiO 2 , manganese oxides have also been reported as supports for low-temperature WGS reaction. The role of oxygen vacancies in manganese oxides has been critical, effective, economically favorable, and in some cases competitive to CeO 2 [137]. In this regard, Shan et al. reported a relatively good catalytic activity of Pd and Pt nanoparticles supported on α-MnO 2 nanorods (termed as Pd/α-MnO 2 and Pt/α-MnO 2 ) prepared by precipitation deposition method at 300°C for WGSR with low-temperature (140-350°C), which  Fig. 11 a Schematic of Pt nanoparticles along with MoO x nano-patches. b Temperature dependence of the CO conversion over different catalysts and Mo/Pt atomic ratios [131] were comparable to Pd/CeO 2 and Pt/CeO 2 . The surface oxygen vacancies density of the catalysts generally remained at 10-15% during the WGS reaction [137]. Poggio-Fraccari et al. evaluated copper and nickel separately as monometallic catalysts supported on cerium-manganese mixed oxides through a WGS reaction, where they found that both Cu and Ni performed best on MnO x -CeO 2 oxides support with 30% Mn:Ce molar ratio [138]. Pd/CeO 2 catalysts are believed to show irreversible deactivation in the WGS reaction owing to losses in the metal surface area. Some reports suggest the use of transition metals as additives in Pd/CeO 2 to counter the irreversible deactivation [139]. Xiao [140]. They reported an enhancement in the WGS activity by using manganese oxide, which possibly could be due to an improvement in the redox property in presence of Pd-Mn and particularly Mn-Ce interactions. The other reason for catalytic performance enhancement was attributed to the formation of metallic Pd species because of the electron-donating effect of MnO x neighbor, which facilitates water dissociation [140]. Lanthanum is also reported to have a unique electronic structure and oxygen storage capability to act as co-catalysts, where it can store as well as release oxygen during oxidation-reduction reactions. In a study conducted by Runxia et al., the effects of La loading on Cu-Mn catalysts towards low-temperature WGS reaction were investigated, and the results showed significant improvement in the catalyst activity by using 0.5 mol% La [141]. The introduction of La resulted in an increase in the dispersion of Cu on the catalyst surface, as well as a more uniform distribution of Cu and Mn in the catalyst [141].

Carbon-based supports
Carbon, in different forms, has gained significant research interest as a support for active sites in catalysis [142]. An investigation, by Zugic et al., on Pt supported on oxygenated multi-walled carbon nanotubes (MWNTs) indicates that platinum can be activated by controlled addition of alkali promotor (Na) on MWCNTs (Fig. 13a) [143]. Pt on MWCNs without Na was however reported to be not active for the WGS reaction. The sample prepared by suspending MWNTs (C N ) in the HNO 3 solution for 2 h and then dissolving in 1 M sodium acetate with a heat-treatment at 800°C (Pt/ 800-Na-2hr-C N ; Fig. 13b, c) was found to have the highest WGS activity (Fig. 13d) [143]. In another study, they also reported a direct promotional impact of sodium on the WGS platinum activity and stability when supported on oxygen-free C N . The co-impregnation of 1 wt% Pt and Na on the annealed nanotubes (1 wt% Pt 1 Na 6 /1000-2h-C N ; Fig. 13e, f) resulted in an exceptionally high catalytic activity in WGSR [144]. Dongil et al. examined two commercial Ni/CeO 2 /CNT catalysts for low-temperature WGSR; moreover, they also used activated carbon for comparison [145]. They demonstrated that CNTs improved the dispersion of ceria, leading to a better CO activity than the ceria used with activated carbon. Furthermore, in conditions where CO 2 and H 2 exist in the feed, the activity decreased slightly, and methanation was discovered at temperatures only above 573 K [145].
Carbon nanofibers (CNF) is another promising alternative to metal oxide-based supports due to the unique structural characteristics that enhance the metal and support interactions. They also have high mechanical strength, high electrical conductivity, and specific surface area, which is suitable for formation and anchoring of smaller metal nanoparticle [146]. Oliveira, Valençaa, and Vieira studied both Cu/CNF and Cu/Al 2 O 3 catalysts for the WGS reaction at a temperature range of 398-573 K [147]. In the case of the lowest evaluated water partial pressure and H 2 O:CO ratio of 3.1, Cu/CNF exhibited the highest activity. In the WGS reaction, however, 5% Cu/Al 2 O 3 exhibited better results owing to the support hydrophobicity of Cu/CNF that prevents water adsorption [147].
Graphitic oxides are also reported to enhance the catalytic performance due to their surface chemical functions and confinement effects [149]. Dongil et al. successfully investigated the low-temperature WGS reaction using both binary Ni/ Graphene and ternary Ni-CeO 2 /graphene systems with 5 wt% Ni and 20 wt% CeO 2 and compared them with NiCeO 2 catalyst [148]. The catalytic activity improved with sodium addition in the presence of Ce (Ni-CeO 2 /G-Na), possibly due to the promotional effect of Na in facilitating ceria reduction (Fig. 14) [148]. Pastor-Perez, Buitrago-Sierra, and Sepulveda-Escribano dispersed different loadings of cerium oxide over activated carbon and investigated their catalytic behavior [150]. The synthesis resulted in small ceria particles, high surface area, and an easier reduction of ceria in the presence of Ni. The catalysts, Ni supported on different loading of CeO 2 /C as well as Ni/CeO 2 (Fig. 15a) were examined for the low-temperature WGSR. The optimum amount of CeO 2 loading was determined to be 20 wt% (Fig. 15b); however, in the realistic conditions, in the presence of CO 2 , the catalyst with 10 wt% CeO 2 showed a better performance. The Ni10CeO 2 /C catalyst was found to have a 40% CO conversion after a 150-h stability test at 493 K (Fig. 15c, d) [150]. In a comparative study conducted by Pastor-Pérez et al. on the WGS performance using Ni/CeO 2 bulk and Ni20CeO 2 /C, the carbonsupported catalyst showed far more activity than the one without the carbon support which can be due to its capacity to store oxygen. Moreover, the Ni-CeO 2 /C stability (Fig. 15e, f) was evaluated under practical conditions, revealing a considerable catalyst activity loss, owing to the Ni particles sintering [151].
A summary table is provided below to enlist the different catalysts used in the review. The synthesis methods and   [143]. e, f HAADF-STEM images of e fresh Pt 1 Na 6 / 1000-2h-C N and f used Pt 1 Na 6 /1000-2h-C N (red circles are atomically scattered species) [144] experimental conditions under which the catalysts were evaluated for WGS reaction are also provided along with the key results in a convenient way to compare their performance (Table 1).
To understand the effect of nanostructure and shape of active metals at nanoscale, following table is constructed by coming the literature reports on similar materials with different nanostructures and morphology (Table 2).

Summary and future outlook
WGSR is one of the commonly used industrial reactions for hydrogen production, and by some estimates, approximately 10% of the annual world energy consumption by 2030 will be linked to the WGS reaction [153]. The WGSR is a wellestablished procedure in conventional chemical/hydrocarbon industries for the production of ammonia, methanol, hydrogen, saturated hydrocarbons, and many other chemicals and petrochemicals. Thermodynamically, WGSR should be conducted at low temperatures for higher hydrogen selectivity; however, low temperature reduces the kinetics rate and thereby increases the required amount of catalysts to achieve high CO conversions [154]. The recent few years have seen considerable growth in nanomaterials and chemical characterization techniques, leading to unique details that were unheard of before. The catalysis community has benefitted significantly with these developments in understanding the dynamic changes in the structure of the catalysts in situ during the reaction conditions, in a more realistic environment than under ultra-high vacuum. Herein, we provided a review of the recent development in the catalysts used for low-temperature WGSR focusing on the nanostructure of the catalysts along with other key factors influencing catalytic performance, such as synthesis technique, nature of active site/phase and support, and reaction environment.
Introducing support materials, especially those with controlled and uniform nanostructured morphology, onto various metal, non-metal, and carbon-based catalysts can be beneficial in achieving catalytic systems with larger surface area and abundant active sites resulting in higher CO conversion activity, and long-term stability. Supports with oxygen storage capacity, such as CeO 2 , are particularly favorable for lowtemperature WGSR. Synergistic effects of combining different catalysts or adding various alkali promoters not only improve the performance of the existing catalysts but also help in reducing the overall cost by introducing earth-abundant elements in developing new types of highly active and stable catalysts. Notably, the introduction of promoters leads to a high chemical and thermal stability for these catalysts by limiting the sintering of smaller particles under severe reaction conditions. Many of the recent work focused on single-atom catalysis to achieve a very high dispersion of the active sites encapsulated on the support structure, which seems to achieve higher turnover rates compared to bigger nanoparticles. Fig. 15 a, b TEM micrographs of a Ni/CeO 2 , b Ni20CeO 2 . c, d TEM micrographs of the Ni10CeO 2 /C catalyst c before and d after the stability reaction test at 220°C [149]. e, f TEM micrographs of the Ni-CeO2/C catalyst e before and f after the stability reaction test at 250°C [150]  /min -Catalyst amount: 250 g -The monolithic catalyst synthesized by the HDP approach had superior performance than the DP technique.
-The presence of Re as a promoter and modifying the HDP method made CO conversion over the monolithic catalyst higher than the particulate ones.
-The monolithic Au-Re/CeO 2 performance was determined to be promising for both ideal and realistic feed conditions. [57] Au/α-MoC Precipitation followed by sequential temperature-programmed ammonization and carburization. The iongel catalyst exhibited high activity at a temperature range of 120 to 160°C.
-The higher the IL loading is, the higher catalytic activity the iongal has.
-The iongel with more anions has higher activity. [71] Ru/C and Ar as balance -Catalyst amount: 0.5 g -SV: 6400 NmL/g·h -Steam/gas volume ratio: 0.6 -Temperature:150-300°C -The structure and properties of the studied catalysts were closely associated with the La content, and 0.5% La loading showed the best performance at low temperature for the WGS reaction. [141] CNT Pt/Na-C N Step-wise manner using carbon-modification Application of ionic conductors in the support also provided encouraging results by enhancing the mobility of oxygen and vacant sites. The search for a low-cost catalyst with high activity, good hydrogen selectivity, and long-term stability will remain the focus from an economic point of view. Though many reports are available on the WGSR mechanism, the reaction pathway is still not very clear on many catalytic surfaces, particularly on catalysts with a low number of active sites. Efficient reactor design, along with catalyst development, will also play a critical role in this thermodynamically limited reaction. An optimal design will ensure quick removal of the products to push the reaction further by avoiding the reaching of thermodynamic equilibrium conditions.
Abbreviations AC, activated carbon; ACP, aromatic conductive polymer; ASSA, aerosol spray self-assembly; CNF, carbon nano fibers; CNT, carbon nanotube; DFT, density functional theory; DI, double impregnation; EDX, energy dispersive X-ray; EG, ethylene glycol; GHSV, gas hourly space velocity; GO, graphene oxide; HAADF-STEM, high-angle annular dark-field scanning transmission electron microscopy; HAp, hydroxyapatite; HDP, homogeneous deposition precipitation; HRSTEM, high-resolution scanning transmission electron microscopy; HRTEM, high-resolution transmission electron microscopy; HTlc, hydrotalcite; IL, ionic liquid; I-R, impregnation-reduction; IWI, incipient wetness impregnation; LDHs, layered double hydroxides; LPRD, liquid phase reductive deposition; MFI, mordenite framework inverted; MWNTs, multiwalled carbon nanotubes; NC, nanocube; NR, nanorod; NS, nanospheres; OSC, oxygen storage capacity; PD, precipitation deposition; RM, reverse microemulsion; SAC, single-atom catalyst; SCFD, supercritical fluid  deposition; SEM, scanning electron microscopy; SFD, supercritical fluid deposition; SILP, supported ionic liquid phase; SV, space velocity; TEM, transmission electron microscopy; TOS, time on stream; TPD, temperature-programmed desorption; WGC, world gold council; WGS, watergas shift; WGSR, water-gas shift reaction; WHSV, weight hourly space velocity; XRD, X-ray diffraction; ZIF-8, zeolitic imidazolate framework-8; ZSM-5, zeolite socony mobil-5 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.