RESPONSE SURFACE OPTIMISATION OF A RICE HUSK-DERIVED FE-PROMOTED ZEOLITE CATALYST SUPPORT FOR BIOGAS STEAM REFORMING

Reforming catalysts for biogas-derived hydrogen are almost entirely imported into sub-Saharan Africa, and the cost of that dependence is one reason small-scale reforming has not taken hold there. This study asks whether two waste streams already abundant in Nigeria, rice husk and post-consumer aluminium, can together supply a zeolite catalyst support on a quantitative rather than an empirical footing. Silica was recovered from rice husk ash by alkaline dissolution and acid precipitation, and alumina by calcining lathe-turned aluminium scrap at 750 °C. A rotatable central composite design of 20 runs related the charges of silica, alumina and sodium hydroxide to the silica-to-alumina ratio of the product. The fitted second-order model was significant (F = 136.55, p < .001; R2 = 0.9919) with a non-significant lack of fit (F = 0.73, p = .634). Canonical analysis placed the stationary point inside the design region at 11.02 g SiO2, 3.47 g Al2O3 and 3.25 g NaOH, with all three eigenvalues of the second-order matrix negative, confirming a true maximum at a predicted silica-to-alumina ratio of 5.72, equivalent to Si/Al = 2.86 on an atomic basis. Promotion used Fe2O3 at a 3 : 1 mass ratio by solid-state contacting, avoiding the nitrate precursor and solvent of wet impregnation. X-ray diffraction identified Zeolite A as the dominant phase (26.9 wt%) beforehand and a surviving Zeolite 4A framework alongside hematite afterwards, while X-ray fluorescence recorded 26.11 wt% iron. Diagnostic problems are reported rather than smoothed over. Residuals depart from normality (W = 0.794, p < .001) because one of six centre-point replicates is discrepant, though a leave-one-out refit moves the optimum by at most 0.047 g and leaves its classification unchanged. Phase quantification of the promoted sample cannot be reconciled with its elemental composition, which bounds what the diffraction data support. The route and the optimum are sound; performance claims await reactor and textural work