From CO2 to DME: Enhancement through Heteropoly Acids from a Catalyst Screening and Stability Study

Abstract

The direct synthesis of dimethyl ether (DME) via CO₂ hydrogenation in a single step was studied using an improved class of bifunctional catalysts in a fixed bed reactor (T _R: 210-270 °C; 40 bar; gas hourly space velocity (GHSV) 19,800 NL kg_cat ^-1 h^-1; ratio CO₂/H₂/N₂ 3:9:2). The competitive bifunctional catalysts tested in here consist of a surface-basic copper/zinc oxide/zirconia (CZZ) methanol-producing part and a variable surface-acidic methanol dehydration part and were tested in overall 45 combinations. As dehydration catalysts, zeolites (ferrierite and β-zeolite), alumina, or zirconia were tested alone as well as with a coating of Keggin-type heteropoly acids (HPAs), i.e., silicotungstic or phosphotungstic acid. Two different mixing methods to generate bifunctional catalysts were tested: (i) a single-grain method with intensive intra-particular contact between CZZ and the dehydration catalyst generated by mixing in an agate mortar and (ii) a dual-grain approach relying on physical mixing with low contact. The influence of the catalyst mixing method and HPA loading on catalyst activity and stability was investigated. From these results, a selection of best-performing bifunctional catalysts was investigated in extended measurements (time on stream: 160 h/7 days, T _R: 250 and 270 °C; 40 bar; GHSV 19,800 NL kg_cat ^-1 h^-1; ratio CO₂/H₂/N₂ 3:9:2). Silicotungstic acid-coated bifunctional catalysts showed the highest resilience toward deactivation caused by single-grain preparation and during catalysis. Overall, HPA-coated catalysts showed higher activity and resilience toward deactivation than uncoated counterparts. Dual-grain preparation showed superior performance over single grain. Furthermore, silicotungstic acid coatings with 1 KU nm^-2 (Keggin unit per surface area of carrier) on Al₂O₃ and ZrO₂ as carrier materials showed competitive high activity and stability in extended 7-day measurements compared to pure CZZ. Therefore, HPA coating is found to be a well-suited addition to the CO₂-to-DME catalyst toolbox.

Figure 1

Synthesis scheme of the CtM catalyst (left) and MtD catalysts (right). Materials used for the preparation of the bifunctional catalysts are highlighted in blue (CtM pre-catalyst) and orange (MtD catalysts 1 and 2).

Figure 2

Results of pure CZZ following the standard measurement protocol of five consecutive temperature steps (T₁ = 250 °C, T₂ = 230 °C, T₃ = 210 °C, T₄ = 270 °C, and T₅ = 250 °C; cf. small graph) at 40 bar and GHSV 19,800 NL kg_cat^–1 h^–1 with CO₂/H₂/N₂ 3:9:2.

Figure 3

NH₃-TPD results of (a) AlO_x, (b) ZrO_x, (c) FERH and FERL, and (d) ZeoH and ZeoL. Measurements were performed twice, once with and on a new sample without NH₃ loading (“w/o NH3”). The area between both curves is the amount of desorbed NH₃.

Figure 4

Synthesis scheme of the bifunctional catalysts (CtD catalysts). Homogeneous preparation leads to combined secondary particles or “single grain” bifunctional catalyst systems with high intra-particular interfaces and intensive contact between basic CtM and acidic MtD catalysts (“CZZ + MtD-cat”). Heterogeneous preparation leads to separate secondary particles or “dual grain” bifunctional catalysts with low contact due to separated basic and acidic particles (“CZZ//MtD-Cat”).

Figure 5

Results of all tested catalysts. (a) T₁ = 250 °C sorted by Y_DME from low to high. X_CO2 is marked as squares including error bars (3 σ = 6%). Selectivity for all carbon-based products is shown as bar diagram: S_CO (red), S_CH4 (yellow), S_MeOH (green), and S_DME (blue). Vertical dashed lines indicate the separation in the three main regions: (left) low Y_DME, (middle) medium Y_DME, and (right) high Y_DME. Results are from the first of five consecutive temperature steps T₁ = 250 °C at 40 bar, CO₂/H₂/N₂ 3:9:2, and GHSV 19,800 NL kg_cat^–1 h^–1. (b) Results for Y_DME of T₁ (250 °C, red dots), T₂ (230 °C, blue triangle), T₃ (210 °C, green triangle), and T₄ (270 °C, black squares). (c) Results for Y_Me of T₁ (250 °C, red dots), T₂ (230 °C, blue triangle), T₃ (210 °C, green triangle), and T₄ (270 °C, black squares). (d) Difference of values of T₅ (250 °C) and T₁. ΔX_CO2 (black squares), ΔS_CO (red triangles)_, ΔS_CH4 (yellow hexagons)_, ΔS_MeOH (green squares), ΔS_DME (blue dots), and ΔY_Me (orange diamonds). For orientation, a line at zero and a dotted line at ΔX_CO2(CZZ) = −1.4% are added.

Figure 6

Heat map representation of catalysis results for X_CO2 and S_DME at T₁ = 250 °C tested for homogeneously (a, b) and heterogeneously (c, d) prepared CtD catalysts. Values are colored as follows: high/good (green), middle/mediocre (yellow), and low/bad (red). Results are from the first of five consecutive temperature steps T₁ = 250 °C at 40 bar, CO₂/H₂/N₂ 3:9:2, and GHSV 19,800 NL kg_cat^–1 h^–1. P-HPA catalysts are not shown due to lack of activity.

Figure 7

SEM and EDX analysis of Si-HPA@AlO_x 58 (a–d) and Si-HPA@ZrO_x 31 (e–h). Carrier material (AlO_x or Zr) is colored green, and tungsten (W) of Si-HPA is colored pink.

Figure 8

Absolute change in X_CO2 visualized as a heat map for homogeneously (a) and heterogeneously (b) prepared bifunctional catalysts. The pure CZZ and less active P-HPA-containing catalysts are not shown. CZZ-level (ΔX_CO2: absolute = −1.4%; relative, −8%) deactivation is set to light green, less deactivated bifunctional catalysts are marked in dark green, and stronger deactivation is marked yellow to red.

Figure 9

Extended measurement of 7 days (160 h), 40 bar, GHSV 19800 NL kg_cat^–1 h^–1. Shown are X_CO2 (black squares), S_DME (blue dots), S_MeOH (green squares), and S_CO (red triangles) of (a) CZZ (250 °C), (b) CZZ//FERH (250 °C), (c) CZZ//Si-HPA@AlO_x 58 (250 °C), (d) CZZ//Si-HPA@ZrO_x 31 (250 °C), (e) CZZ//FERH (270 °C), (f) CZZ//Si-HPA@AlO_x 58 (270 °C), and (g) CZZ//Si-HPA@ZrO_x 31 (270 °C).