Introduction Ammonia selective catalytic reduction of NOx with NH3 (NH3-SCR) is an effective technology for NOx removal. As a catalyst, V2O5-WO3 (MoO3) /TiO2 has a high denitrification activity, but suffers from the problems of poor low-temperature activity, high V-toxicity, and susceptibility to heavy metal toxicity at 300-400 ℃. Heavy metals such as Cd, Pb, Zn, etc. widely present in the flue gas of solid waste incineration and coal-fired power plants. The toxic effects on catalysts are manifested via covering the active sites on the catalyst surface, reducing the redox properties, acidifying the surface and chemically adsorbing oxygen, leading to severe catalyst deactivation. CeO2 as an effective catalyst has a higher oxygen release/storage capacity, which is characterized by a higher oxygen release/storage capacity, compared to the conventional catalysts. CeO2 is used as a main active component in the preparation of rare-earth-based denitrification catalysts due to its high oxygen release/storage capacity. Manganese oxide (i.e., MnOx) exhibits a good low-temperature catalytic activity, which is often used as a modified component to improve the low-temperature denitrification activity of catalysts. TiO2 interacts with the active component and has a higher thermal stability and a sulphur resistance. However, Ce-Mn/Ti catalysts are susceptible to poisoning by toxic substances in the flue gas, and there are more studies on the poisoning of Ce-Mn/Ti catalysts by SO2, H2O, etc.. However, a few studies on the poisoning of Ce-Mn/Ti catalysts by heavy metals are reported. Catalysts with three-dimensional ordered macroporous structures have superior redox properties, sufficient chemisorbed oxygen and suitable acid sites to promote its NH3-SCR denitrification reaction. In this paper, Ce-Mn/Ti catalysts with three-dimensional ordered macro-porous structures were thus prepared to increase the redox properties and acidic sites and promote the denitrification activity.Methods 17 mL of tetrabutyl titanate was added drop by drop into 30 mL of anhydrous ethanol under vigorous stirring, which was recorded as liquid A. Concentrated nitric acid (65%, in mass fraction) was added dropwise into the mixed solution of deionized water (1.5 mL) and anhydrous ethanol (10 mL) at pH values of 2-3, which was recorded as liquid B. The solution was then mixed into a mixture of deionized water (1.5 mL) and anhydrous ethanol (10 mL). Liquid B was slowly mixed into liquid A under vigorous stirring to obtain a homogeneous and transparent TiO2 solution. A certain amount of PMMA template was added to the TiO2 precursor solution, sealed and stayed for 12 h. The remaining precursor solution was removed by vacuum filtration and baked at 50 ℃ for 24 h. Finally, the samples were baked at 300 ℃ for 2 h, and then heated at a heating rate of 1 ℃/min at 500 ℃ for 5 h. The ordinary structure TiO2 carriers were prepared in the same process as 3DOM TiO2 without PMMA template. After the preparation of TiO2 carriers, Ce(NO3)3-6H2O and Mn(NO3)2 mixtures were prepared at different molar ratios, and impregnated with 3DOM TiO2 and normal TiO2, respectively, and stirred at room temperature for 4 h. Afterwards, the catalysts were transferred to an oil bath and stirred at 80 ℃ for 6 h. They were baked at 80 ℃ for 24 h, and finally, the catalysts were heated at 1 ℃/min at 500 ℃ for 5 h. The obtained samples were recorded as fresh-CMT-3DOM (i.e., fresh-CMT). The catalyst with heavy metals (i.e., Cd, Pb, Zn) were prepared by an excess impregnation method. A certain amount of cadmium nitrate, lead nitrate, and zinc acetate solution was prepared at different mass ratios of metal ions to catalyst (10% Cd, 10% Pb and 5% Zn), and fresh-CMT-3DOM and fresh-CMT catalysts were added to the heavy metal salt solutions, respectively. The morphology structure was detected by a scanning electron microscope (SEM, TESCAN MIRALMS Co., Czech Republic). The specific surface area and pore size were determined by specific surface area intruement based on BET (Micromeritics Co., USA). The crystalline phase structure of the catalysts was analyzed by an X-ray diffracometer (XRD, Bruker Co., Germany). The elemental valence states of the catalysts were analyzed by an X-ray photoelectron spectroscope (XPS, Thermo Scientific Co., USA) using Al Kα as X-ray sources. The redox properties of the catalysts were characterized using a FINESORB-3010 adsorption apparatus (Finetec Instruments Co., China). The ammonia adsorption experiments of the catalysts were carried out by a Nicolet 6700 instrument (Thermo Fisher Scientific Co., USA) to determine the type of acid on the catalyst surface. The NH3-SCR denitrification activity was analyzed in a fixed bed quartz reactor at 1 000 mg/L NO, 1 000 mg/L NH3, 5% (v/v) O2 and N2 as ab equilibrium gas. The denitrification efficiency of the catalyst was calculated according to the results that were determined by a flue gas analyzer. Results and discussion The NO conversion for 10% Cd-CMT-3DOM, 10% Pb-CMT-3DOM, and 5% Zn-CMT-3DOM exceeds 85% at 175-300 ℃. However, the denitrification activity decreases at 100-150 ℃. Compared to the activity of CMT and the three metals after poisoning, the maximum NO conversion of 10% Cd-CMT decreases to less than 85% at 100-300 ℃. The maximum NO conversion of 10% Pb-CMT is less than 100%, and fresh CMT is weakly poisoned by Zn. The three-dimensional ordered macro-porous structure enhances the CMT-3DOM catalyst's tolerance to Cd, Pb, and Zn to some extent. The ratio of Ce3+ can be determined via calculating the peak area ratio of Ce3+/(Ce3+ + Ce4+). The proportion of Ce3+ in fresh-CMT-3DOM (19.0%) is greater than that in fresh-CMT (15.7%), accounting for the favorable denitrification activity of CMT-3DOM catalyst. After adding Cd, Pb and Zn, the content of Ce3+ in the fresh-CMT-3DOM catalyst shows a better resistance to heavy metals, compared to the fresh-CMT catalyst. This indicates that the fresh-CMT-3DOM catalyst is more resistant to heavy metals. Based on the calculation results of Mn4+ content, which decreases similarly to Ce3+ content, fresh-CMT-3DOM and fresh-CMT display the greatest reduction in Mn4+ content after Cd poisoning, followed by Pb poisoning. Zn poisoning has a smaller impact on the Mn4+ content reduction. The Mn4+ content on the surface of fresh-CMT-3DOM is relatively higher after three heavy metals poisoning rather than that on the corresponding fresh surface. Also, the relative contents are higher than those of the corresponding fresh-CMT catalysts. This indicates that the Mn4+ content of fresh-CMT-3DOM catalysts is less affected by heavy metal poisoning, which is beneficial to improving the anti-heavy metal performance of fresh-CMT-3DOM catalysts. The relative content of Oα in the catalysts is obtained via the calculation. The amount of chemisorbed oxygen in fresh-CMT-3DOM catalysts (i.e., 35.2%) is greater than that of fresh-CMT. The relative content of Oα, Ce3+ and Mn4+ depict a consistent pattern of variation in these catalysts, which coincide with the outcomes of the NH3-SCR activity. This further shows the safeguarding ability of the catalysts active species through the three-dimensionally ordered macropore structure. Based on the H2-TPR reduction peak results, the reduction peaks of fresh-CMT-3DOM move to lower temperatures, compared to those of fresh-CMT catalysts. The relative peak areas of fresh-CMT-3DOM are greater, indicating that the 3DOM structure is more favorable for the redox reaction of the catalysts. When Cd, Pb, and Zn are introduced, the reduction peaks of both catalysts shift towards at higher temperatures, indicating that the heavy metals form bonds with the active species of the catalysts. The reduction peaks of all catalysts are analyzed to determine the relative peak areas. The influence of Cd, Pb, and Zn on the catalytic redox activity aligns with SCR denitrification performance. This indicated that the catalysts ability to undergo redox reactions is a crucial factor affecting their heavy metal resistance and SCR denitrification activity. The NH3-DRIFTS characterization results indicate that the Lewis acid is a primary acid type in both fresh-CMT-3DOM and fresh-CMT catalysts. Moreover, fresh-CMT-3DOM catalysts have a higher concentration of the Lewis acid sites rather than fresh-CMT. Fresh-CMT-3DOM after the heavy metal poisoning has the more Br?nsted acid sites, compared to fresh-CMT catalysts, and the influence of heavy metals on its acid sites is reduced. Fresh-CMT-3DOM catalysts possess a larger number of acidic sites, resulting in a greater resistance to heavy metals.Conclusions TiO2 carrier with a three-dimensional ordered macro-porous structure was prepared using PMMA template, and fresh-CMT-3DOM catalysts were obtained after loading with Ce and Mn. The activity and heavy metal resistance performance were investigated. The experimental results showed that the fresh-CMT-3DOM catalyst had the excellent activity with the NO conversion rate of more than 85% at 100-300 ℃. The low-temperature activity was higher than that of fresh-CMT, which still remained at 85% at 100-150 ℃. When Cd, Pb and Zn were introduced, NO conversion rate of fresh-CMT-3DOM at 175-300 ℃, while the fresh-CMT activity after Cd, Pb and Zn poisoning decreased, in which the maximum NO conversions of the catalysts corresponding to Cd and Pb poisoning were less than 85% and 100%, respectively. The characterization results indicated that the NO conversions of the catalysts corresponding to Cd, Pb and Zn poisoning were higher than those of fresh-CMT-3DOM and fresh-CMT-3DOM. 3DOM and fresh-CMT catalysts had the similar crystalline phases, but the diffraction peaks of fresh-CMT-3DOM were weaker in intensity and its active components were more dispersive. Compared with fresh-CMT catalysts, fresh-CMT-3DOM catalysts had excellent Mn4+/Mn3+/Mn2+ and Ce4+/Ce3+ redox cycling, higher redox capacity and more acidic sites, thus improving the anti-heavy metal performance of fresh-CMT-3DOM.