Global warming, caused by an increase in atmospheric CO 2 concentrations, and limited fossil-fuel resources have stimulated the search for environmentally sustainable energy sources with zero or low CO 2 emission. [1] The use of biomass as a CO 2 -neutral fuel has drawn widespread attention, since biomass as part of the global carbon ecocycle is an increasingly important energy resource. [2] An associated issue, which is often encountered in the control of NO x emissions from the alkali-rich stack gas of power plants (co-)fuelled by biomass, is the severe deactivation of conventional V 2 O 5 -WO 3 /TiO 2 catalysts in the selective catalytic reduction of NO by NH 3 (NH 3 -SCR). [3,4] This deactivation is predominately caused by the strong interaction of alkali-metal ions with catalytically active sites. [4][5][6][7][8][9][10][11] For example, Zheng et al. [8] proposed a deactivation mechanism by proton exchange at catalytically active Brønsted acid sites with alkali-metal ions. The exchange reactions occur readily at normal NH 3 -SCR operating temperatures because alkali species, such as KCl and K 2 SO 4 , are mobile in view of their low Tamman temperatures (T Tam ). [12] On the basis of this deactivation mechanism, a possible approach to the development of an alkali-resistant catalyst is to increase the number of catalytically active sites and "alkaliactive" sites (for trapping alkalis), for example, by using a support with the property of high or super acidity or by increasing the concentration of active elements. [9-11, 13, 14] However, this approach can not completely alleviate alkali poisoning because, to a great extent, alkali-metal ions and reactants compete for the same sites. Therefore, the development of a highly alkali resistant catalyst remains a great challenge. Herein we describe the successful development of a hollandite manganese oxide (HMO) catalyst with separate catalytically active sites (for NH 3 -SCR) and alkali-active sites (for alkali trapping). With the combined action of the two types of active sites, this bifunctional HMO catalyst not only protects catalytically active sites against alkali poisoning, but also preserves high catalytic activity at the original level until all the alkali-active sites are occupied.We used K + as a probe to clearly identify the catalytically active sites of HMO in NH 3 -SCR reactions, as K + is a wellknown catalyst poison. [7][8][9][10][11] Figure 1 shows the catalytic activity of HMO with different K + loadings in NH 3 -SCR reactions at 350 8C as expressed in terms of the rate constant for the conversion of NO. [9][10][11] The K + loading had almost no influence on the catalytic activity of HMO, and all HMO catalysts tested exhibited almost the same activity.For reference, a conventional deNO x catalyst (V 2 O 5 -WO 3 / TiO 2 ) with different K + loadings was also tested in the NH 3 -SCR reactions under the same conditions (Figure 1). As expected, the V 2 O 5 -WO 3 /TiO 2 catalyst exhibited rather poor alkali resistance and underwent severe and continuous d...