Unique selectivity of Au/TiO 2 for ammonium formate decomposition under SCR-relevant conditions
3.3 Results and Discussion
Anatase Au/TiO2 catalysts were synthesised via a facile modified incipient wetness impregnation method[131] as described in Chapter 2. Figure 3.1 presents the XRD patterns of the fresh and 10 h hydrothermally aged catalyst which had been exposed to 10 vol% water at 800 °C. Both the catalysts exhibited peaks corresponding to anatase phase. The low loading (0.5 wt%) and small particle size (5-7 nm) of gold rendered its detection impractical. There was a slight anatase peak broadening which corresponded to an increase in the anatase crystallite size from 25 nm to 32 nm. However, no phase transformation of titania from anatase to rutile was evident from the XRD patterns. The BET surface areas of the fresh, 5 h and 10 h hydrothermally aged catalysts are presented in Table 3.1. A gradual decrease in surface area was witnessed with increasing hydrothermal aging time. After 10 h of hydrothermal aging, the BET surface area dropped by ca.
35% to 49 m2 g-1. Figure 3.2 depicts the nitrogen sorption isotherms of the fresh and 10 h hydrothermally aged catalysts. Both the catalysts exhibited Type IV isotherms characteristic of H2-hysteresis loops associated with capillary condensation taking place in mesopores. The hydrothermally aged catalyst exhibited sintering-induced lowering of adsorbed volume. Figure 3.3 presents the SEM and STEM images of the fresh and 10 h hydrothermally aged catalyst. It appears that while the support (titania) morphology exhibited signs of sintering upon 10 h hydrothermal aging at 800 ºC, the particle sizes of the visible gold particles were only subtly
Unique selectivity of Au/TiO2 for ammonium formate decomposition under SCR-relevant conditions
27
affected. Titania particles of the aged catalyst are in an aggregated state upon sintering which is in consensus with the lowering of BET surface area.
Figure 3.1 XRD patterns of fresh and 10 h hydrothermally aged 0.5 wt% Au/TiO2 catalysts showing the absence of any major peaks associated with the rutile phase. The TiO2 crystallite sizes were determined to be 25 nm and 32 nm in the fresh and 10 h hydrothermally aged samples, respectively.
Table 3.1 BET surface area of fresh and hydrothermally aged 0.5 wt% Au/TiO2 catalysts.
Sample BET Surface Area (m2 g-1)
Fresh 75
5 h Aged 51
10 h Aged 49
Figure 3.2 Nitrogen sorption isotherms of fresh and 10 h hydrothermally aged 0.5 wt% Au/TiO2 catalysts. Closed symbols indicate the adsorption branch while the open symbols represent the desorption branch.
0.0 0.2 0.4 0.6 0.8 1.0 0
50 100 150 200
Aged 10 h
Volume adsorbed, cm3 g-1
Relative pressure, P/P0
Fresh 10 20 30 40 50 60 70
AA A
A A
A
A A
0.5Au/TiO2-Aged 0.5Au/TiO2-Fresh
A A : Anatase
Intensity, a.u.
2, degree
Figure 3.3 SEM and HAADF-STEM images of 0.5 wt% Au/TiO2, fresh (left column) and 10 h hydrothermally aged (right column) catalysts, showing sintering of TiO2 particles upon aging without any significant change in Au particle size.
Figure. 3.4 Product yields for carbon dioxide, carbon monoxide, ammonia, formic acid, methanamide and nitrogen oxide (NO+NO2) yields obtained upon the complete decomposition of 40% AmFo over 0.5 wt% Au/TiO2. (0.5 g∙L-1 washcoat loading on the monolith, GHSV = 19,490 h-1; feed gas: 5 vol% water, 10 vol% oxygen, 85 vol% N2 and 0.05 mL∙min-1 liquid spray of 40 wt% AmFo).
200 220 240 260 280 300 0
20 40 60 80 100
Yield, %
Temperature, C
N-balance C-balance NH3 CO2 CO HCOOH HCONH2 NOx
Unique selectivity of Au/TiO2 for ammonium formate decomposition under SCR-relevant conditions
29
The carbon and nitrogen balances were closed using the molar feed and reactor outlet concentrations as quantified by FT-IR spectroscopy.[149] Washcoat loading refers to the amount of the catalyst deposited on the monolith per unit volume.[162] At a 100 g∙L-1 washcoat loading, which is approximately two-thirds of the typical catalyst loadings in automobiles, Au/TiO2 reliably converted AmFo into ammonia and carbon dioxide in feed containing 10 vol% oxygen and 5 vol% water (Figure 3.4). Ammonia did not react under such conditions rendering negligible NOx and methanamide yields, the latter being a side-product arising from the reaction between formic acid and ammonia.[23] Methanamide may dehydrate further to form hydrogen cyanide (HCN); however, under these reaction conditions, no HCN was formed over Au/TiO2. To determine the stability of the catalysts at partial conversion, the washcoat loading was reduced by 99.5% and tested under identical conditions of GHSV and feed composition. Additionally, the fresh catalyst was subjected to two incremental hydrothermal aging steps each lasting for 5 h at 800 °C in air containing 10% water, and the activity tests were repeated.
Figure. 3.5 Product yields and AmFo conversion during AmFo decomposition over (a) fresh (b) 5 h aged, and (c) 10 h aged 0.5 wt% Au/TiO2 catalysts (0.5 g∙L-1 washcoat loading, GHSV = 19,490 h-1; feed gas: 5 vol% water, 10 vol% oxygen, 85 vol% N2 and 0.05 mL∙min-1 liquid spray of 40 wt% AmFo), showing nearly 100% ammonia yields between 200 °C and 300 °C.
Figures 3.5 (a), (b) and (c) depict the yield of all products formed from the decomposition of 40 wt% AmFo over fresh, 5 h and 10 h hydrothermally aged 0.5 wt% Au/TiO2, respectively. 100%
conversion for AmFo decomposition was achieved with all catalysts. Ammonia, formic acid and
200 250 300
carbon dioxide were the major products, while carbon monoxide and methanamide formed in low yields. The carbon monoxide yield decreased from ~10% over fresh catalyst to ~3% over aged catalysts. Methanamide, which was not observed in the case of 100 g∙L-1 catalyst, started to appear when using 0.5 g∙L-1 catalyst and accounted for ~3% yield or lesser over both the fresh and aged catalysts across all temperatures. Carbon dioxide yields were decreased due to aging from ~43% over the fresh catalyst to ~20% and ~17% over 5 h and 10 h aged catalysts, respectively, at 300 °C. The product yield patterns over the 5 h and 10 h aged catalysts suggest a kind of stabilisation of the catalytic activity after the first aging step. Moreover, ammonia yields always remained close to 100%, which further reiterates the discriminative oxidation behaviour of our catalyst, and contrasts the typical ammonia oxidation behaviour of precious metal containing catalysts.[163,164]
Table 3.2 Mass-based rate constants at various reaction temperatures calculated using pseudo-first-order kinetics.
Pseudo-first-order kinetic constants were calculated to assess the relative activities of catalysts tested under identical operating conditions.[19,152,154] Table 3.2 lists the mass based rate constants calculated for the fresh, 5 h aged, 10 h aged 0.5 wt% Au/TiO2 and fresh bare titania catalysts. The rate constants progressively decreased for all catalysts with decreasing temperature, indicative of generic Arrhenius behaviour. Hydrothermal aging at 800 °C for 5 h decreased the rate constants by over 60%, while 5 h further treatment resulted in a marginal drop of ~15% when comparing the activities at 300 °C. At lower temperatures, the rate constants tend to nearly identical values for 5 h and 10 h aged catalysts. To elucidate the influence of gold, a control experiment using similar washcoat loading of bare titania was performed under identical conditions. The resultant rate constants were found to be a magnitude of at least 4 lower than that of the fresh Au/TiO2 catalysts. Importantly, over bare titania, the conversion of formic acid selectively produced carbon monoxide; there is no carbon dioxide formation, under these conditions.
Catalyst sample Temperature (K) Rate constant kmass (L∙g-1s-1)
Unique selectivity of Au/TiO2 for ammonium formate decomposition under SCR-relevant conditions
31
Because, formic acid undergoes secondary reactions to form carbon monoxide and carbon dioxide, the evolution of their selectivities as a function of temperature over the fresh and aged catalysts is interesting. Figure 3.6 (a) and (b) show that the temperature dependence of the selectivities for carbon monoxide and carbon dioxide formation and formic acid conversion remained essentially unchanged even after 10 h aging. A decrease in carbon dioxide selectivity was accompanied by an increase in the carbon monoxide selectivity with increasing temperature for all catalysts. Formic acid conversion was significantly affected by the hydrothermal aging translating to a drop in conversion from ~57% to ~17% after 10 h treatment in 10% water.
Figure 3.6 Relationship between selectivity and conversion with temperature for fresh (a) and 10 h aged (b) 0.5% Au/TiO2 catalysts and (c) Time on stream plot portraying long term stability of fresh (blue) and 10 h hydrothermally aged (red) catalysts at 190 °C (0.5 g∙L-1 washcoat loading on the monolith, GHSV = 19,490 h-1; feed gas: 5 vol% water, 10 vol% oxygen, 85 vol%
N2 and 0.05 mL∙min-1 liquid spray of 40 wt% AmFo).
The time on stream activities of the fresh and 10 h hydrothermally aged catalysts were evaluated over 48 h (Figure 3.6 (c)). A low temperature (190 °C) was chosen to examine the low conversion stability. Both the fresh and aged catalysts exhibited stable activities, selectivities and nearly 100% ammonia yields with no signs of deactivation with time-on-stream.
200 240 280
3.4 Conclusions
Au/TiO2 catalysts demonstrating unique selectivity against ammonia oxidation in a highly oxidizing environment during ammonium formate decomposition is reported in this chapter. The catalysts display excellent stability giving close to 100% ammonia yields for 48 h of time on stream showing no signs of deactivation. The observations evidenced in this work, reflecting the exceptional selectivity against ammonia oxidation could entail new applications of monometallic gold in exhaust gas catalysis.