2.3.1. Water Poisoning

As seen in the above CO2 hydrogenation reactions, the dissociation of CO2 produces oxygen atoms, which in turn results in the formation of water. This byproduct is necessary for the thermodynamic favorability of the entire process, but can be unfortunately detrimental to catalytic performance. It is because of this unavoidable mechanistic absolute that the mitigation of water poisoning must be part of all catalytic investigations [66].

Wu et al. [67,68] examined the effect of the produced water on the stability of Cu/ZnObased catalysts in methanol synthesis from the high temperature hydrogenation of CO2. Specifically, Cu/ZnO/ZrO2/Al2O3 (40/30/25/5) was subjected to a CO2-rich feed, which produces water, and a CO-rich feed, which does not produce water. The examination of the catalysts by X-ray powder diffraction (XRD) after 1 h and 500 h time-on-stream of a CO2-rich feed containing steam showed that the Cu and ZnO crystallized more rapidly when compared to identical catalysts exposed to a CO-rich feed not containing steam. In particular, the Cu particle size in the catalyst used with the CO2-rich feed containing steam grew from 94 Å to 166 Å from 1 h to 500 h. The particle size growth under steam might be the key reason causing catalyst deactivation.

Huber et al. observed the rapid deactivation of Co/SiO2 during an FTS reaction at high water partial pressure, and the loss of activity was attributed to the support breakdown byproduct water accompanied by the formation of stable, inactive cobalt-silicates and the loss of the BET surface area [69]. van Steen et al. stated that metallic cobalt crystallites with a diameter less than 4.4 nm are more susceptible to oxidation by water to form Co(II)O [70]. This is in agreement with Iglesia's work showing that small Co metal crystallites (<5–6 nm diameter) appear to re-oxidize and deactivate rapidly in the presence of a water reaction product in typical FTS conditions [71].

Water poisoning has the most dramatic effect on zeolite-based CO2 hydrogenation catalysts for which the acidic sites of the zeolite are essential for catalysis. Recently, Zhang et al. investigated the water effect over zeolite-based catalysts at high temperatures, and found that water caused the loss of crystallinity and modified acid sites, thereby deactivating the catalyst [72]. Their studies show, by functionalization with organosilanes, that the tolerance of defective zeolites to hot liquid water can be greatly enhanced. This method renders the zeolite hydrophobic, which prevents the wetting of the surface. At the same time, the organosilanes act as a capping agent of Si−OH species, reducing their reactivity. Both aspects are important for the prevention of water attack [72].

It appears that there are several analogies of Fe catalysts for CO2 hydrogenation and CO catalysts for conventional FT synthesis. Kliewer et al., for example, showed that for a supported CO catalyst, water can oxidize the surface of the CO to an inactive oxide phase, and it also plays a large role in sintering. With a high water partial pressure in the Fe system, it appears that this can also oxidize iron carbides to inactive surface oxide phases and also promotes particle growth sintering [51].

#### 2.3.2. Carbonaceous Deposits (Coke)

Coke is produced by the decomposition or condensation of hydrocarbons on the surfaces of catalysts, and is primarily is comprised of polymerized hydrocarbons. There have been several books and reviews that describe the formation of coke on catalysts, and the resulting deactivation [73–78].

These deposits are most problematic for catalysis involving zeolites, because the active sites of the zeolites become blocked or fouled by the coke deposits. The deactivation of MTO reactions over zeolites due to coke deposition results in a reduction in both the catalyst activity and product selectivity [79–81].

Nishiyama et al. [82] studied the effect of the SAPO-34 crystal size on the catalyst lifetime, and found that the amount of coke deposited on the deactivated SAPO-34 catalyst increased with the decreasing crystal size, indicating that for larger crystals, the reactants were unable to penetrate further into the larger crystals to reach other acidic sites. Because MTO reactions and coke formation take place simultaneously in the same pores, it seems likely that the effectiveness of the catalyst increased with the decreasing crystal size. Their studies demonstrated that the coke formation was inhibited in small-crystal SAPO-34 due to reduced diffusive resistance.

The work of Wei et al. on CO2 hydrogenation found that the deactivation of the zeolites HMCM-22 and HBeta was the result of coke formation, which deposited in the zeolites' cavities and channels. The deposition blocked the reactants' access to the zeolites' acid sites, leading to the deactivation [11]. Muller et al. investigated the MTO process on H-ZSM-5 catalysts in plug-flow (PFR) and fully back-mixed reactors (CSTR). They found that the catalysts deactivated under the homogeneous gas phase in the CSTR. It was shown unequivocally that, in the early stages of the reaction, the zeolite deactivates via Brønsted acid site blocking, and not by coke-induced deposition restricting the pore access. The deactivation of H-ZSM-5 in the CSTR occurred at first rapidly, and then at a much slower rate (Figure 5). The rapid deactivation was observed in a PFR due to the

formation of a larger fraction of the oxygen-containing carbon species. The larger fraction of oxygen-containing carbon species increases the reaction with the desired olefins, which results in a strongly adsorbed aromatic molecule. The formation of aromatic coke proceeds mostly by hydride transfer between olefins and carbon growth via multiple methylations of such aromatic species [83].

**Figure 5.** MTO reaction over the H-ZSM-5-S catalyst in the PFR at ambient pressure and the H-ZSM-5-S catalyst in the CSTR at 6.5 bar, T = 723 K and pMeOH = 178 mbar. Adapted with permission from ref. [83]. Copyright 2015 Elsevier.

Zeolite-based catalysts that show promise for high olefin selectivity are unfortunately typically limited by mass transfer, suffering from rapid deactivation due to carbon deposition and water poisoning [83]. The issues with coke deactivation on the zeolite catalysts involved in MTO reactions are seen in classical MTO chemistry. The directed transformation of coke into active intermediates in a methanol-to-olefins catalyst was reported to boost the light olefin selectivity [84]. Another strategy to mitigate the deactivation was to synthesize nanozeolites, which have shortened diffusion paths, or mesoporous hierarchical zeolites, which exhibit longer catalyst lifetimes because of their larger pores and improved mass transfer [85–87].

With an Fe catalyst, the deactivation by coke is not related to the constriction of narrow pores, but several authors have reported the formation of carbonaceous residues on the active sites. Lee et al. investigated the deactivation behavior of an Fe–K/-Al2O3 catalyst, and found that the deactivation pathway was different according to the reaction position and reaction time. The main deactivation reason was the phase transformation at the top of the reactor. Conversely, the main factor at the bottom of the reactor was the deposited coke generated by secondary reactions. In particular, the produced olefins may have been adsorbed on acidic sites, and thus the olefins served as major precursors to coke. The SEM micrographs of the used catalysts clearly showed that most of the surface was covered by deposited graphite and graphite clusters protruding on the surface, mixed with some fine filamentous carbon (Figure 6) [57].

With these multiple deactivation pathways having been identified, it now becomes a critical issue to find ways of modifying the catalyst to become more stable. In the section directly below, we will describe the approaches that multiple researchers have examined in an attempt to mitigate deactivation.

**Figure 6.** SEM image of the Fe–K/γ-Al2O3 catalysts after the CO2 hydrogenation reaction: (**a**) 100 h, (**b**) 300 h, and (**c**) 500 h. Adapted with permission from ref. [57]. Copyright 2009 Elsevier.

#### **3. Recent Progress on the Mitigation of the Catalyst Deactivation**

We will discuss the effect of promotors, metal oxide support, bifunctional composition and structure on the catalyst design to minimize the catalyst deactivation. We will summarize reports published within the last five years showing that promoters, supports, and novel morphology designs have mitigated the deactivation effects.
