*2.6. Catalyst Stability*

Catalyst longevity is one of the cornerstones in the study of reforming catalysts. Sintering of metal particles, loss of support surface area, and extensive coke formation can cause catalyst fouling at the harsh conditions necessary to perform reforming reactions. Spinel supported metals (especially Ir and Rh) have been shown to maintain physical stability under these harsh reaction conditions with minimal coke formation [20,21]. In this section, we analyze the e ffect that the use of a simulant natural gas has in catalyst performance, focusing on longer term stability. In the prior sections we demonstrated that C2+ steam reforming activity is limited over Ir-based catalysts, presumably due to a reduced capacity of breaking ethane's C–C bond becoming rate limiting at low reaction temperatures. In the following section, we present results for the Rh catalyst (chosen as benchmark) and evaluate the influence that reaction conditions have on catalyst performance. The aim is to find reaction conditions that allow stable catalytic operation even with fast throughputs and relatively low S/C ratios.

For the reforming experiments shown in Figure 3, methane was fed at a rate of 5.3 mmol/min and the reaction was carried out on similar metal surfaces (23.5 and 25.2 μmol of Rh and Ir respectively). The catalytic rate was reported for each temperature after 1 h where steady state was reached. Using these reaction conditions, especially with short reaction times (1 h time on stream [TOS]), no deactivation was seen; however, if the reaction was allowed to proceed for longer times, deactivation was evident. The results shown in Figure S6 (for the reaction at 850 ◦C) show how conversion starts to decline slightly after 8 h of reaction. Under ethane reforming conditions, a more pronounced deactivation was found after a few hours on stream for the reaction at 600 ◦C (see Table 1). It is clearly evident that the presence of C2+ hydrocarbons, and especially at a high carbon throughput, has a negative e ffect on catalyst life, which largely depends on the reaction temperature and feed rate.

Catalyst durability experiments for methane steam reforming over the Rh catalyst are presented in Figure 7a. After over 100 h at both 850 and 900 ◦C (τ = 2.3 ms, S/C = 2), no deactivation was observed. Figure 7b depicts results when using a simulated natural gas feed mixture containing methane, ethane, propane, and butane. For simplicity, only methane conversion results are shown in Figures 7b and 8. After 100 h, catalysts showed continued deactivation and retention of only part of the initial activity (91%, 78%, and 53% activity retained for the Ir, Rh, and Ni catalysts, respectively). These results illustrate how the presence of C2+ hydrocarbons are detrimental to catalyst durability. However, we note that these results were obtained at a very low molar S/C ratio of 1.5. We attribute catalyst deactivation to increased carbon fouling (coke buildup) of the higher hydrocarbons on the catalyst surface. As shown in Figure 7c when the molar S/C feed ratio was increased to 2.0 the catalyst was quite stable at 850 and even 900 ◦C for 120 h duration. Although we note that prolonged duration testing at 900 ◦C revealed some deactivation, at 850 ◦C the reaction was quite stable.

**Figure 7.** Methane conversion vs. time on stream (TOS) using (**a**) pure methane feed over Rh at 850 and 900 ◦C (S/C = 2.3 ms, τ = 2.3 ms), and using natural gas simulant feed over (**b**) Rh, Ir, and Ni at 850 ◦C (S/C = 1.5 ms, τ = 2.3 ms), and (**c**) Rh at 850 and 950 ◦C (S/C = 1.9 ms, τ = 2.0 ms).

**Figure 8.** Methane conversion vs. TOS over Rh/MgAl2O4 catalyst as a function of steam/carbon molar feed ratio at 850 ◦C (**A**), and at 750 ◦C for 500 h TOS (**B**).

There are two common strategies for avoiding/suppressing coking on reforming catalysts: (1) increase the amount of steam fed and (2) conduct the reaction at lower temperatures. On one hand, increasing the amount of steam increases the formation of oxidation products (CO, and CO2), while hydrocarbon adsorption/activation rates remain unaffected. On the other hand, the main effect of a lower reaction temperature is a decrease in the rates of hydrocarbon activation; as a result, reforming reaction rates supersede the coke rate formation, finally leading to a more stable long-term operation. Results for these two situations are presented in Figure 8 for the Rh catalyst. At 850 ◦C, the S/C molar feed ratio was varied. Increasing the S/C ratio enhanced catalyst durability. Figure 8a shows how running with an S/C molar feed ratio of 3 catalytic activity was largely maintained after 100 h of testing (some initial deactivation appeared to have occurred). By comparison, only 78% of catalytic activity was maintained with a lower S/C molar feed ratio of 1.5. Thus, increasing the S/C ratio improves catalyst durability. Figure 8b shows that a reduced operation temperature has a beneficial effect on catalyst life. Even after a 500 h test duration, the Rh catalyst retained its activity when operated at 750 ◦C.

## **3. Experimental Methods**

## *3.1. Catalyst Preparation*

As reported elsewhere [20], a series of catalysts were prepared by incipient wetness impregnation of MgAl2O4 (Puralox 30/140 from Sasol) with a solution of Rh nitrate (10 wt.% Rh in nitric acid), Ir nitrate (19.3 wt.% Ir in nitric acid), and Ni nitrate hexahydrate salt (Sigma-Aldrich, St. Louis, MO, USA). The resulting metal loadings were 5 wt.% Rh, 5 wt.% Ir, and 15 wt.% Ni. After impregnation, the catalysts were dried at 120 ◦C for 8 h and calcined at 500 ◦C for 4 h under static air.
