1 Introduction

Over the past decades, the continuous rise in CO2 levels has made it one of the most problematic greenhouse gases (GHGs) [1]. Carbon Capture and Utilization (CCU) technologies have emerged as highly effective solutions for reducing CO2 concentrations and harnessing its potential [2]. These technologies enable the transformation of CO2 into valuable chemicals like CO or CH4, facilitating a circular economy that promotes a smarter utilization of natural resources [3]. Among CCU paths, syngas production from CO2 is a very promising route since syngas is a precursor for many fuels (i.e., methanol, formic acid or light olefins) [4]. Syngas can be produced from CO2 and H2 via the reverse water–gas shift (RWGS) reaction [5]. This path allows to valorise large quantities of CO2 due to the high CO2 conversion and low H2 consumption of the RWGS compared with other CO2 hydrogenation reactions [6].

The RWGS consists of the hydrogenation of CO2 to form CO and H2O (Eq. 1) [7]. The RWGS is an endothermic reaction. In other words, CO production via RWGS is favoured at high temperatures. Nonetheless, there are also other side reactions competing with the RWGS[8]. These are the Sabatier reaction (Eq. 2) and the CO methanation reaction (Eq. 3). Both are exothermic, being therefore favoured at low temperatures [9]. Besides these unavoidable side reactions, the high stability of CO2 makes necessary the synthesis of a high selective and active catalyst.

$${{\text{CO}}}_{2}\left({\text{g}}\right)+ {{\text{H}}}_{2}\left({\text{g}}\right)\to \text{CO}\left({\text{g}}\right)+ {{\text{H}}}_{2}{\text{O}}\left({\text{g}}\right) \quad {\Delta {\text{H}}}_{2}^{\circ }\;42.1\text{ kJ/mol}$$
(1)
$${{\text{CO}}}_{2}\left({\text{g}}\right)+ {4{\text{H}}}_{2}\left({\text{g}}\right)\to {{\text{CH}}}_{4}\left({\text{g}}\right)+ {2{\text{H}}}_{2}{\text{O}}\left({\text{g}}\right)\quad {\Delta {\text{H}}}_{2}^{\circ }\; - 65 \text{kJ/mol}$$
(2)
$${\text{CO}}\left({\text{g}}\right)+3{{\text{H}}}_{2}\left({\text{g}}\right)\to {{\text{CH}}}_{4}\left({\text{g}}\right)+ {{\text{H}}}_{2}\text{O}\left({\text{g}}\right)\quad {\Delta {\text{H}}}_{2}^{\circ }-206 \text{ kJ/mol}$$
(3)

Using the RWGS reaction at low temperatures is especially appealing due to its clear energy-saving benefits and its compatibility with downstream processes, such as methanol production, which typically occurs at temperatures around 300 °C [10]. Several attempts to explore the RWGS at low temperature have been described in literature [11,12,13,14]. Among the different supports that have been studied for this process (i.e., Al2O3 [15], CeO2 [16], SiO2 [17]), TiO2 stands out for its redox properties, providing oxygen vacancies that facilitate the activation and reduction of CO2. Another advantage of TiO2 is the strong metal–support interaction (SMSI), which enhances the number of reduction sites on the TiO2 surface [18]. Many noble metals have been studied as the active phase for RWGS catalysts due to their high catalytic activity and selectivity (i.e., Rh [19], Pd [20], Au [21]). However, the high cost associated with these metals has limited their industrial-scale application. Cu has also been studied as an active phase in this reaction, offering a more economically viable alternative to noble metals. Even though Cu stands as a very promising alternative, Cu-based catalysts exhibit lower catalytic activity, particularly for low-temperature reactions. Additionally, long-term experiments have revealed surface sintering of the catalyst, diminishing its lifespan [22]. To overcome these issues, research on new Cu catalytic systems is needed.

Despite the recent efforts of the catalysis community working on the RWGS reaction, there are still many challenges when facing low temperature reaction performance. To shade a bit of light to this intriguing reaction, in this work we explore the use of effective multicomponent catalysts using Cs as promoter that our team identified as a very promising dopant [23]. Indeed, Cs possesses a large ionic radius, low ionization potential, and excellent basic properties. Its electropositive character facilitates CO2 activation [23]. To examine the influence of Cs, we adopt the following approach. Firstly, we synthesize Cu/TiO2 and Cu–Cs/TiO2 catalysts and evaluate their catalytic performance. Subsequently, Pt/TiO2 and Pt–Cs/TiO2 catalysts are also synthesized and studied in order to compare the catalytic performance of Cu with a well-known noble metal for this process [24, 25]. Furthermore, we explore the impact of Cs on Ni-based catalysts. As Ni exhibits high activity in methanation reaction [26], we aim to ascertain the ability of Cs to inhibit methanation and promote CO production. To achieve this, we prepare Ni/TiO2 and Ni–Cs/TiO2 catalysts for comparison.

Under these premises, this work presents a comparative study of transition metals (Cu, Ni) vs noble metal (Pt) based catalysts for the low-temperature RWGS analysing the impact of Cs as a dopant. Activity–structure/chemical properties correlations are discussed along with lessons learned in the design of low-temperature RWGS catalysts that could play a role in the urgent challenge of industrial decarbonisation.

2 Experimental

2.1 Catalyst Preparation

The TiO2 supported catalysts were prepared by the sequential wet impregnation method using a commercial TiO2-P25 (Degussa) support. First, the TiO2 was washed with deionized water, filtered and dried for 48 h at 115 ºC. Then, the washed TiO2 was calcinated at 550 ºC during 3 h. Once the support was calcinated, the active phases were impregnated. The right amount of H2PtCl6, Ni(NO3)2·6H2O and Cu(NO3)2·5H2O was dissolved in excess of water to obtain 1 wt% Pt, 10 wt% Ni and 10 wt% Cu respectively. Different loads of the active metals in the catalyst were selected due to the different properties of those. For instance, only 1 wt.% of Pt was used due to its high reactivity and high cost compared to Cu and Ni. Afterwards, the samples were dried for 48 h at 115 ºC and calcinated for 3 h at 550 ºC. Finally, for the Cs-doped samples, the right amount of Cs2CO3 was impregnated on the Pt–TiO2, Ni–TiO2 and Cu–TiO2 samples to get a 2.5 wt% of Cs. Then, the samples were dried and calcinated under the same conditions (i.e., 3 h and 550 °C).

2.2 Catalyst Characterization

The as-prepared catalysts were characterized by X-ray diffraction (XRD) with an X’Pert Pro PANalytical at room temperature using Cu-K (40 mA, 45 kV) over a 2-theta range of 10°–90°. The textural characterization of the catalysts was carried out by N2 adsorption at − 196 °C with an AUTOSORB-6 equipment (QUANTACHROME INSTRUMENTS). Catalysts were first outgassed at 250 ºC for 2 h.

The H2 temperature-programmed reduction (TPR) was performed using a conventional U-shaped quartz reactor connected to a thermal conductivity detector (TCD). A gas flow of 50 mL/min consisting of 5% H2 diluted in Ar was passed through the reactor. TPR experiments were conducted using approximately 100 mg of catalysts for Cu and Ni, while 150 mg was used for Pt due to its lower weight percentage. The heating rate was set at 10 ºC/min, gradually increasing the temperature from room temperature to 900 ºC.

The investigation of particle size and morphology of the active phases was conducted using transmission electron microscopy (TEM) on a JEOL 2100Plus microscope operating at 200 kV. The microscope was equipped with an Energy Dispersive X-Ray analysis system (EDX X-Max 80 T, Oxford Instruments) and a CCD camera for image recording. High-resolution TEM (HR-TEM) analysis was performed on a FEI Talos F200S electron microscope, utilizing an acceleration voltage of 200 kV, a field emission filament, and a side-mounted Ceta 16 M camera.

The XPS experiments were conducted on a SPECS photoelectron spectrometer equipped with a PHOIBOS 150 MCD analyzer operating at a constant pass energy of 35 eV and a resolution of 0.1 eV. The X-ray radiation source used was Kα emission generated by bombarding an aluminum target with electrons at an energy of hν = 1486.6 eV and a bandwidth of 0.85 eV, operating at 250 W and maintaining a potential of 12.5 kV. Due to significant overlap with CsMNN, Cu catalysts utilized a Mg source with an energy of hν = 1253.6 eV and a bandwidth of 0.7 eV, operating at 150 W and maintaining a potential of 10 kV. The analysis chamber operated under ultra-high vacuum conditions, with a pressure of 10−10 mbar.

2.3 Catalytic Activity

The RWGS reaction was carried out in a vertical continuous fixed-bed reactor where 0.2 g of each catalyst was placed in the middle of a quartz reactor. Initially, the catalyst was first activated by heating the reactor up to 500 ºC under a H2/N2 gas mixture (5/45). Subsequently, the temperature was reduced to 300 ºC using a N2 gas flow. Following this, the catalyst was exposed to a feed gas mixture of H2:CO2 in a ratio of 4:1, while maintaining a weight hourly space velocity (WHSV) of 30,000 mL/g h The catalytic tests were performed at temperatures of 300, 400 and 500 ºC for 1 h at each temperature. The products of the reaction were analysed using gas chromatography integrated with flame ionization (FID) and thermal conductivity (TCD) detectors. CO2 conversion (Eq. 4), CO selectivity (Eq. 5) CH4 selectivity (Eq. 6) and specific activity towards CO (Eq. 7) and CH4 (Eq. 8) were utilized as parameters to compare the catalytic activity of the different samples.

$${{\text{CO}}}_{2}\text{ Conversion }\left(\%\right)=\left(\frac{{\left[{{\text{CO}}}_{2}\right]}_{in}- {\left[{{\text{CO}}}_{2}\right]}_{out}}{{\left[{{\text{CO}}}_{2}\right]}_{in}}\right)\times 100$$
(4)
$$\text{CO Selectivity }\left({\%}\right)=\left(\frac{{ \left[{\text{CO}}\right]}_{out}}{{\left[{CH}_{4}\right]}_{out} +{ \left[{\text{CO}}\right]}_{out} }\right)\times 100$$
(5)
$${{\text{CH}}}_{4}\text{ Selevticity }\left({\%}\right)=\left(\frac{ {\left[\text{CH}_{4}\right]}_{out}}{{\left[\text{CH}_{4}\right]}_{out} +{ \left[{\text{CO}}\right]}_{out} }\right)\times 100$$
(6)
$$Specific\; activity\;to\;CO=\left(\frac{ mol CO/h}{{\text{gcat}}}\right)$$
(7)
$$Specific\;activity\;to\;CH4=\left(\frac{\text{mol}{CH}_{4}/\text{h }}{{\text{gcat}}}\right)$$
(8)

2.3.1 Equilibrium Limits Calculation

ChemStations’ ChemCad software package was used to observe the thermodynamic limits of the RWGS reactions over a range of temperatures. The Soave–Redlich–Kwong equation of state was used in a Gibbs reactor. Material flows into the reactor are identical to those intended to be used for experimentation.

3 Results and Discussion

3.1 XRD

XRD analysis was conducted to determine the crystalline phases present in our catalytic systems. Figure 1 displays the XRD results for the six catalysts as well as the calcined support. For sake of clarity, the crystalline phases that are common to all XRD profiles are marked in the support profile, while the newly observed phases are indicated in the pattern above, in Fig. 1.

Fig. 1
figure 1

Powder XRD patterns of calcinated samples: A TiO2, Cu/TiO2 and CuCs/TiO2; B TiO2, Ni/TiO2 and NiCs/TiO2; and C TiO2, Pt/TiO2 and PtCs/TiO2

Full size image

Upon calcination of TiO2 at 550 ºC, both anatase and rutile crystalline phases are observed, being anatase more predominant. The presence of these two phases in Pt/TiO2 catalysts has also been previously identified by other researchers [27]. Previous works point out that the lack of diffraction lines corresponding to the Pt phase in the 1% Pt/TiO2 catalyst indicates that Pt is indeed well dispersed throughout the catalyst. This fact can also be confirmed by the HR-TEM results (see Sect. 3.3), where the crystalline Pt particles on TiO2 cannot be clearly distinguished. This can also be ascribed to the low Pt loading which is below the detection limit of the equipment. It is noteworthy that the addition of the active phase and the promoter leads to an increase of the rutile phase in all catalysts. This could be explained by the repeated calcination of the samples due to the nature of the synthesis, as the conversion from anatase to rutile initiates at temperatures exceeding 500 ºC [28]. The absence of Cs peaks in the XRD patterns of the different Cs-doped catalysts suggests that the Cs species are either in an amorphous state or very well dispersed on the prepared catalysts. This could also be due to the low concentration of Cs.

Apart from the anatase and rutile phases from the TiO2 support, the XRD patterns for the Ni and Ni-Cs catalysts reveal the presence of NiO2 and NiTiO3 peaks. The NiTiO3 phase is more predominant in the Cs-doped catalyst, as clearly observed in Fig. 1B indicating the impact of Cs addition to the final catalysts structure. As for the Cu family, both Cu and Cs-doped Cu catalysts present a similar XRD pattern with evidences of CuO formation.

3.2 Textural Properties

Table 1 presents the specific surface area, pore volume and pore diameter of the commercial calcined TiO2–P25 and the six synthetised catalytic systems [29]. TiO2 is known for its relatively high surface area, which can be advantageous for catalytic reactions, providing more active sites for reactant adsorption, enhancing the catalytic activity. The addition of the active phase results in a decrease in the specific surface area of TiO2 for Ni and Cu based catalysts. For Pt based catalysts surface area is maintained. This can be correlated to the lower Pt loading compared to that of Cu and Ni. The increased in pore diameter for all the samples could be attributed to the incorporation of the metals within the pores of the support. Cs does not seem to affect the surface area. This could suggest a high dispersion and low particle size of the promoter which is in good agreement with previous reports [18]. It can be observed that the pore diameter decreases significantly for the Pt-based catalysts while maintaining the highest pore volume. This could be explained by the fact that Pt presents the highest atomic diameter and is deposited at the pores mouth, reducing the pore diameter while preserving the pore volume. The pore diameter of the Cu-based catalysts is the largest, while the pore volume is the lowest, indicating that copper particles are likely deposited inside the pores. The situation for Ni catalysts is similar to that of copper. The six samples display similar type-IV isotherms (Fig. S1), which are characteristic of mesoporous solids [30]. These mesopores can offer improved mass transport and diffusion properties, enabling better reactant access and product desorption. To gain further insights into the textural properties, the crystallite sizes of anatase and rutile of the various catalysts have been depicted in Fig. 2. It can be observed that there is no clear trend among the samples. In the case of the Pt-based catalyst, the Cs-doped system exhibits a significant decrease in anatase crystallite size. Additionally, this catalyst is the only sample that presents a smaller crystallite size than the bare support. For the Ni catalysts, the doped one shows an increase in rutile crystallite size. As will be discussed in the catalytic activity section, the Cu/TiO2 catalyst is the least active sample. This may be correlated with the fact that these catalysts exhibit the highest crystallite size for the two phases indicating some sort of correlation between activity and crystal size and perhaps suggesting such large crystallite size marks a threshold for activity depletion.

Table 1 BET, pore volume and pore diameter of the calcinated samples
Full size table
Fig. 2
figure 2

Crystallite size of anatase and rutile phases for the six catalysts and the support

Full size image

3.3 HR-TEM

Figure 3 provides HR-TEM images that confirm the presence of CuO, NiO and NiTiO3 as discussed in the XRD results. In addition, it appears that interplanar distances characteristic of mix copper oxide (I, II) named paramelaconite (Cu4O3) is present in the Cs-doped catalyst. This occurrence can be explained by the electron-donor factor of Cs, that reduces Cu species in the surfaces of the catalyst. In both the Pt/TiO2 and Pt–Cs/ TiO2 samples, a remarkable dispersion of Pt particles is achieved. The particle size distribution (Fig. S2) reveals that the Pt particles in the Cs doped catalyst are about 0.1 nm, while in the non-doped catalyst, they range between 0.1 and 0.2 nm. However, in the case of the Cu and Ni catalysts, the opposite effect is observed. The Cu/TiO2 catalyst exhibits a significant number of metallic Cu particles with a medium size smaller than 5 nm, which are absent in the Cs-doped catalyst where the medium particle size is around 22 nm (Fig. S2). A similar effect is observed in the Ni catalysts. While Ni/TiO2 shows a medium particle size distribution between 5 and 10 nm, in the Cs-doped catalysts, it is about 21 nm. This adverse effect can be attributed to the larger loading of Cu and Ni compared to Pt catalysts. 10% of Cu and Ni is enough for Cs to make it difficult for the particles to distribute homogenously in the support surface. This effect is not observed in the PtCs/TiO2 catalyst due to the lower Pt loading (1 wt%). The larger size of Cs particles helps Pt particles to be separated over the TiO2 surface.

Fig. 3
figure 3

HR-TEM images: A Cu/TiO2; B CuCs/TiO2; C Ni/TiO2; D NiCs/TiO2; E Pt/TiO2 and F PtCs/TiO2

Full size image

3.4 H2-TPR

Redox behaviour is relevant for a redox process like the RWGS. Figure 4 shows the H2-TPR of the six catalysts providing insights in the redox properties of the multicomponent systems. In the Ni/TiO2 catalyst, three distinct reduction zones (low, medium, and high temperature zones) are observed, being the maximum temperatures located at 404, 508 and 602 ºC respectively. The low temperature reduction process is attributed to bulk NiO located on the surface of the support. The medium temperature reduction zone corresponds to the reduction of NiO significantly interacting with TiO2 while the high temperature zone is associated to the reduction of NiTiO3 species to Ni0 [31]. In the case of the Cs-doped nickel catalyst, these three zones appear at higher temperatures, being the maximum temperatures at 433, 542 and 636 ºC respectively. This may indicate a stronger interaction between the NiO and NiTiO3 species with the TiO2 in the presence of Cs doping which is consistent with previous studies [32]. In addition, the presence of Cs could favour the formation of nickel hydroxide during the synthesis. This nickel hydroxide would be the precursor for the formation of NiTiO3 which could explain the higher intensity of this specie in XRD. In other words, the addition of Cs impacts on the final structure and active phase of the Ni-based system [32].

Fig. 4
figure 4

TPR profile of calcinated samples

Full size image

Similarly, in the Cu catalyst, two reduction zones are observed with the maximum temperatures located at 140 and 208 ºC. The first reduction zone corresponds to highly dispersed copper oxide nanoparticles, while the second zone is attributed to larger copper clusters. This would be in good agreement with particle size distribution from HR-TEM, where half of the copper particles size were 1 nm for the Cu/TiO2 sample. Normally, CuO reduction peaks appear above 300 ºC. However, the interaction between CuO and TiO2 shifts these peaks to lower temperatures. This shift is caused by the formation of oxygen vacancies in the TiO2 surface, leading to the generation of TiOx species. These species attract oxygen, facilitating the reformation of TiO2 due to its inherent instability [33]. In the Cs-doped catalysts, these two reduction zones appear at higher temperatures (220 and 293 ºC). The first zone is ascribed to the reduction of CuO species interacting with TiO2, while the second one is likely due to bulk CuO with weaker interaction with the support. This would be also in good agreement with the particle size distribution as for the CuCs/TiO2 catalysts, copper particle size is bigger in comparison with the non-doped sample.

As for the Pt/TiO2 catalyst, two reduction zones are observed centred at 335 and 562 ºC respectively. The zone at the lower temperature corresponds to the reduction of surface oxygen of TiO2, while the zone at the higher temperature is associated with the reduction of bulk TiO2 [34]. The presence of Pt in the catalyst seems to facilitate the reduction of oxygen species in TiO2, which is consistent with previous works [34]. It is reported that the addition of noble metals to oxide supports can enhance the reduction of the support because of H spill-over effect. This is explained because of the ability of H2 dissociation of noble metals. This reduction associated with H spill-over effect could reduce TiO2 to TiOx promoting the SMSI effect [35]. In the Cs-doped catalyst the reduction zones associated with the reduction of surface and bulk TiO2 are also slightly shifted to higher temperatures (385 and 579 ºC). An additional reduction zone appears at lower temperatures. This zone is related to the reduction of PtOx species to Pt0 and TiO2 species interacting strongly with Pt [18, 36]. The addition of Cs seems to promote the reduction of PtOx to metallic Pt, as this peak is not present in the non-doped catalyst.

3.5 XPS

Figure 5 shows the XPS measurements of the six studied catalysts. Figure 5c presents the Ni2p spectra of Ni/TiO2 and NiCs/TiO2 samples. It can be observed that there is no electronic change in the presence of Cs. Ruling out the electronic effect opens the question of how Cs significantly affects the selectivity trends on Ni (and also Pt and Cu). In this sense building upon our recent operando study, we see that Cs changes the reaction pathway by neutralizing the hydroxyl groups of the TiO2 surface, partially inhibiting methane formation [37]. When it comes to Cu and Pt, Cs does affect their surface electronic features as observed in our XPS results. Figure 5a shows Pt4d spectra for the two Pt based catalysts. The Pt/TiO2 sample presents a maximum in binding energy at 318.3 eV that can be attributed to PtO2 [38]. It can be observed a decrease in the binding energy of the PtCs/TiO2 sample to 317.8 eV, indicating an increase in electronic density. It can be deduced that this electron density increase is due to Cs electropositive behaviour. Also as mentioned before, our recent study demonstrates how the presence of Cs neutralizes hydroxyl surface group and provides new active sites that favours CO production rather than CH4 [37]. In addition, another potential consequence of this increase in electronic density could be to facilitate electron transfer from the catalyst to the reactants, thus enhancing their adsorption and activation [23].

Fig. 5
figure 5

XPS spectra of Pt/TiO2–PtCs/TiO2 (a), Cu/TiO2–CuCs/TiO2 (b) and Ni/TiO2–NiCs/TiO2 (c)

Full size image

Finally, Fig. 5b shows the Cu2p spectrum for Cu/TiO2 and CuCs/TiO2. In the case of Cu/TiO2, the characteristic spectrum of CuO can be observed [38]. However, when the sample is doped with Cs, the signal related to Cu2p3/2 is broader. This suggest that Cs is helping to stabilize another type of copper with higher binding energy. This information can be correlated with the HR-TEM images of CuCs/TiO2 sample where Cu4O3 type species can be observed but are not present in the non-doped catalyst. The interplay among the different copper species reflects Cu redox behaviour which is also helpful to undertake CO2 activation and push forward the RWGS reaction [39].

3.6 Catalytic Activity

The catalytic performance of the six samples is presented in Figs. 6 and 7. The Ni/TiO2 catalyst presents the highest CO2 conversion but exhibits poor CO selectivity due to its tendency to form methane, in fair agreement with previous studies [40]. Ni/TiO2 catalyst have been reported before, achieving remarkable CO2 conversions and CH4 selectivity [41]. However, the addition of Cs significantly alters the selectivity towards CO, particularly at low temperatures. This reinforces our hypothesis that Cs suppresses methanation in Ni-based catalysts opening a new route for typical Sabatier catalysts to become switchable Methanation/RWGS systems as reported by our team [9]. Cs addition results in an electron-rich surface favouring CO desorption and hence the selectivity towards CO [42]. CO2 conversion is also affected being significantly lower in comparison to the non-doped sample. This may be explained by the lower dispersion and higher particle size of Ni, as it was discussed in the HR-TEM section.

Fig. 6
figure 6

CO2 conversion of the studied catalysts at 300, 400 and 500 °C

Full size image
Fig. 7
figure 7

Selectivity for A CO and B CH4 of the six catalysts at 300, 400 and 500 °C

Full size image

On the other hand, both the Cu/TiO2 and CuCs/TiO2 catalysts exhibit nearly 100% CO selectivity. However, it is noticeable that the non-doped Cu catalyst presents low CO2 conversion, while the addition of Cs remarkably enhances it. The low activity for the non-doped catalysts can be related to Cu small particles not being accessible for reactants. As can be seen in the particles size distribution from HR-TEM, half of the particles measured for this samples are 1 nm size. It can be thought that encapsulation of these particles by the TiO2 due to SMSI effect is taking place, avoiding reactants to enter in contact with copper active sites. This has been previously observed in surface reconstruction phenomena on Cu/TiO2 nanotubes for CO2 conversion [43].

In the case of CuCs/TiO2 catalyst, the CuO clusters are bigger and even though SMSI effect can be also occurring in this sample, the total encapsulation of the active phase is not taking place. In fact, SMSI effect can improve the catalytic activity of the system. Zhang et al. [44] studied the SMSI effect by the introduction of oxygen vacancies in Cu/TiO2 catalyst for CO2 hydrogenation reactions, concluding that controlling this effect can promote CO2 adsorption and activation because of the new interface generated between Cu and TiO2. In addition, the presence of Cu4O3 species observed in HR-TEM is supported by XPS, as the addition of Cs seems to help the stabilization of a different type of copper species with lower electron density. The presence of this new specie could be the reason of the difference in the catalytic activity of the two copper base catalyst, enhancing CO2 adsorption and activation [42].

Selectivity is the key factor in the RWGS specially at low-temperatures where methanation is favoured. Specific activities for products formation provides a clear basis for performance comparison among the studied multicomponent systems. Figure 8 shows specific activity towards CO and CH4. The positive promotion effect of Cs is more clearly seen in these graphs. When Cs is added to Cu catalyst, its specific activity follows the trend of the Pt-Cs catalysts. On the other hand, it should be remarked again how the CH4 production significantly drops upon Cs addition in the Ni catalysts. Overall Pt-systems present the best activity/selectivity balance, but it is worth nothing that the Cs-promoted Cu catalysts display comparable and even higher CO production capacity with very similar CO2 conversion values making the multicomponent CuCs/TiO2 an economically appealing option for low-temperature RWGS units. To further set the context of our multicomponent catalysts in the RWGS scenario, Table 2, provides a comparison of the catalytic performance of our systems versus reported catalysts in literature tested under comparable conditions. Overall, the table showcases the potential of our multicomponent catalysts that present remarkable performance standing out within the so far reported systems for this reaction.

Fig. 8
figure 8

Specific activity towards CO for Pt/TiO2, Cu/TiO2 Ni/TiO2 (a) and Cs-doped catalysts (c) and towards CH4 for Pt/TiO2, Cu/TiO2 Ni/TiO2 (b) and Cs-doped catalysts (d)

Full size image
Table 2 Catalytic activity comparison with literature findings
Full size table

4 Conclusion

This study showcases an effective strategy to design low-temperature RWGS catalysts using Cs-doped multicomponent catalysts. The addition of Cs to all catalysts resulted in improved catalytic performance at low temperatures. The Cs promoter effectively inhibited the methanation reaction, favoring the production of CO. This effect can be clearly observed in the Ni/TiO2 catalyst. While the non-doped catalysts present a 100% selectivity towards CH4, the addition of Cs undoubtedly shift this behavior, reducing this value to less than 50%. Addition of Cs to Cu/TiO2 catalyst doubled its CO2 conversion reaching almost the activity of Pt/TiO2. Overall, the observed trends in the catalytic activity of the different catalysts may be attributed to Cs modifying the redox and structural properties of the catalytic systems. Based on TPR and HR-TEM results, the electron-donor capacity of Cs towards the different species is evidenced. This effect can be better analyzed by XPS over Pt and Cu based catalysts. An increase in electron density over Pt species due to the presence of Cs can be observed. This could facilitate de activation an adsorption of the reactants. For copper catalysts, the addition of Cs seems to stabilize a different type of copper active sites, enhancing remarkably its catalytic performance. Additionally, Cs may favor stronger support-metal interactions and could also facilitate the adsorption and activation of CO2 molecules.

Outperforming Pt as champion active phase is hard and our data indicate that Pt-systems display the best activity/selectivity balance. Nonetheless, the Cs-promoted Cu catalysts reach comparable and even higher CO production specific activity with very similar CO2 conversion values making the multicomponent CuCs/TiO2 an economically appealing alternative for low-temperature RWGS units. Hence, further refinement of our CuCs/TiO2 could lead to promising results and direct application on advanced CO2 conversion units that will be crucial in modern net-zero industrial environments.