Using Rice-husk-derived Porous Silica Modified with Recycled Cu from Industrial Wastewater and Ce to Remove Hg0 and NO from Simulated Flue Gases

Resource-recovered CuOx-CeOx/SiO2 samples were prepared by using rice-husk-derived silica modified with recycled copper ion from panel industry wastewater for controlling Hg0 and NO emissions from simulated flue gases. By using the silicate exfoliation method for sample preparation, the presence of Cu and Ce oxides could increase the specific surface area (SBET) of SiO2. 50%Cu-10%Ce/SiO2 having the largest SBET showed significant NO removal. XRD results indicated that significant CuO diffraction peaks were not detected among all the CuOx/SiO2 samples, suggesting that CuOx was highly dispersed on the surface. SEM and TEM images showed that the uniform spherical SiO2 particles have changed into platelike structure, further confirming the occurrence of structural rearrangement after incorporated with Cu/Ce oxides via silicate exfoliation. XPS results showed that Cu2+ and Ce4+ were the major valence states presenting in the samples. H2-TPR and NH3-TPD indicated that the 50%Cu/SiO2 and 50%Cu-10%Ce/SiO2 had greater redox ability and stronger acidity as compared to those containing smaller amounts of CuOx and CeOx. Cu and Ce modification was shown to successfully improve the NO removal efficiency. 50%Cu-10%Ce/SiO2 had the best NO conversion efficiency of 70–85% with a broad temperature window of 150–300°C. 50%Cu/SiO2 exhibited the greatest total Hg removal efficiency of 88.2% among all the tested samples at 150°C and remained almost the same removal efficiency at 250°C. These results suggest that the recycled Cu modified rice-husk-derived SiO2 is feasible for not only controlling Hg0 and NO emissions but also recovery of agricultural and industrial wastes.


INTRODUCTION
Mercury (Hg) and NO x discharged from coal-fired power plants (CFPPs) have both received special concern owing to the high toxicity and long retention time in the environment of Hg and the formation of acid rain, photocatalytic smog and secondary PM 2.5 from NO x UNEP, 2018;U.S. EPA, 2019). Hg 0 has been known hardly removed due to its high stability and volatility as compared to Hg 2+ and particulate Hg (Hg p ). Instead of using activated carbon to capture Hg 0 (Chou et al., 2018), catalytic transformation of Hg 0 to Hg 2+ and subsequently to Hg p by metal-oxide selective catalytic reduction (SCR) catalysts is beneficial, not only the formed Hg 2+ and Hg p can be removed by downstream flue gas desulfurization device(FGD) but also NO x reduction could be simultaneously achieved. Catalytic oxidation of Hg 0 using different metal-oxide catalysts in CFPP gases has been examined (Wang et al., 2013(Wang et al., , 2014Xiong et al., 2017). Numerous studies have also been conducted to comprehend the simultaneous removal of Hg 0 and NO (Li et al., 2011;He et al., 2013;Chang et al., 2015;Li et al., 2015;Zhao et al., 2016;Song et al., 2018;Lin et al., 2019). The influencing factors on the removal effectiveness of Hg 0 and NO x include flue gas composition and temperature, the physiochemical characteristics of catalysts, and the dosage of NH 3 for NO x reduction that could limit Hg 0 removal by competing the adsorption sites.
A summary of selected research on novel metal oxides for Hg 0 and NO removal is listed in Table S1. Although V 2 O 5 -WO 3 /TiO 2 and V 2 O 5 -MoO 3 /TiO 2 are the most widely applied industrial catalysts for the SCR process, several kinds of transitional metal oxides, including CuO x and CeO 2 , are also active for NH 3 -SCR reactions and have attracted great attention in these years Chiu et al., 2015;Chen et al., 2018;Fan et al., 2018;Jiang et al., 2018;. Sullivan et al. (2005) indicated that two different copper precursors, Cu(NO 3 ) 2 and CuSO 4 impregnated on oxide-supported TiO 2 , Al 2 O 3 , and SiO 2 demostrated good NH 3 -SCR activity in the presence of H 2 O. Chiu et al. (2017) reported that CuO x /SiO 2 with high copper content (50 wt%) and large surface area/mesoporosity showed greater NO and Hg 0 removal performance as compared to VO x and MnO x /SiO 2 . CeO x has also received considerable attention for NH 3 -SCR in recent years due to its excellent oxygen storage capacity and high redox ability via Ce 4+ to Ce 3+ transition . CeO 2 mixed with other oxides, such as Mn-Ce/TiO 2 , CeO 2 -TiO 2 , Ce x Ti 1-x O 2 , CeO 2 -WO 3 /TiO 2 , CeO 2 -WO 3 , and CeO 2 /Al 2 O 3 , had great SCR activity but with poor resistance to sulfur and water poisoning (Liu et al., 2012). Gao et al. (2010) indicated that compared with Ce-Ti (CET) oxide catalyst, the Ce-Cu-Ti (CCT) oxide catalyst showed better performance at temperatures lower than 350°C. CCT also exhibited higher SO 2 -resistant ability than CET.
Mesoporous SiO 2 has been shown to successfully remove NO or Hg 0 as adsorbents or catalytic supports. Liu et al. (2012) reported that adding SiO 2 into TiO 2 successfully broadened the activity temperature range. The Ce/TiO 2 -SiO 2 catalyst with a Ti/Si mass ratio of 3/1 exhibited the best NH 3 -SCR activity and its NO x conversion was greater than 90% at the temperature range of 250-450°C. Peng et al. (2013) showed that the doping of CeO 2 -WO 3 /TiO 2 catalyst with SiO 2 increased the reaction rate of reducing NO with NH 3 at relatively low temperatures. Zhao et al. (2014) indicated that SiO 2 -TiO 2 nanocomposite catalyst had the greatest Hg 0 removal efficiency in the simulated flue gas when Ti: Si ratio was 2:1. Tan et al. (2012) also indicated the potential of Fe 2 O 3 -SiO 2 composite for Hg 0 removal via heterogeneous oxidation in flue gases.
Rice husk is abundantly available agricultural waste in rice-producing countries (Adam et al., 2012). According to statistics from the Council of Agriculture in Taiwan, approximately six hundred thousand tons of waste rice husk were produced each year. The white ash obtained from the combustion of rice husk at moderate temperature contains 87-97% silica in an amorphous form and some amount of metallic impurities (Yalçin and Sevinç, 2001). Therefore, rice husk ash has been utilized as a silicon source to prepare high-performance products for diverse applications due to its large specific surface, high activity, and abundant silicon content (Deng et al., 2016).
In this research, rice-husk-derived SiO 2 incorporated with copper ion recycled from panel industrial wastewater was prepared via silicate-exfoliation method to understand the samples' removal effectiveness for Hg 0 and NO. Importantly, because the SiO 2 support and Cu ion are both recycled from wastes, the resulting materials are considered as environmentfriendly catalysts and adsorbents. Furthermore, silicateexfoliation synthesis leads to reformation of the SiO 2 surface structure, which results in uniform distribution and high loading of Cu and Ce oxides on the silica surface and high surface area and pore volume (Lin et al., 2019). The improvement in the abovementioned properties of materials could greatly enhance the removal of Hg 0 and NO x , but still not yet be thoroughly comprehended.

Fabrication of Metal-oxide-incorporated SiO2 via Silicate Exfoliation
To synthesize rice-husk-derived SiO 2 , 160 g of rice husk and 40 g of citric acid were first mixed in 1500 ml H 2 O to hydrolyze organic matter and remove metal ions. The resulted solution was then hydrothermally treated at 100°C for 24 h. Filtration, drying, and calcination at 600°C were then employed for 6 h in air to yield the raw SiO 2 sample.
To prepare CuO x -CeO x /SiO 2 , the weight percentages of Cu in CuO x /SiO 2 were set at 10, 25, and 50 wt%; the weight percentages of Cu and Ce in CuO x -CeO x /SiO 2 were set at 50 and 10 wt%, respectively. For example, to prepare 50%Cu-10%Ce/SiO 2 , 10 ml industrial wastewater with 75,000 ppm Cu and 0.51 g Ce(NO 3 ) 2 were dissolved in water to form a solution. The rice-husk-derived SiO 2 was then added into the metal salt solution. Afterward, the mixture was neutralized with 0.6 M Na 2 CO 3 aqueous solution and increased pH to 9.0. After stirring for 2 h at 40°C, the resulted solution was hydrothermally treated at 100°C for 24 h. The metal oxide would incorporate with the rice-husk-derived SiO 2 during the hydrothermal process via silicate exfoliation. Filtration, drying, and calcination at 400°C were employed for 4 h in air to yield the 50%Cu-10%Ce/SiO 2 sample. The resulting metal oxide-incorporated SiO 2 was ground into powder, passed through a 50-mesh sieve, and stored for subsequent sample characterization and Hg 0 and NO removal tests.

Characterization of Metal-oxide-incorporated SiO2
The specific surface area (S BET ), total pore volume (V t ), and average pore size of raw and metal-oxide-incorporated SiO 2 samples were measured by N 2 adsorption at 77K (Micromeritics ASAP-2020). Surface morphology and sample size were obtained using scanning electron microscopy (SEM; Hitachi SU8010) and transmission electron microscopy (TEM; JEOL JEM-1400). The crystalline structures of samples were examined by X-ray diffraction (XRD; Shimadzu 7000S) with Cukα radiation (λ = 1.5405 Å). The instrument was operated at 45 kV and 40 mA target current. Continuous scans were performed from 2θ = 30° to 80° with a 0.03° step size and a counting time of 4 s/step. X-ray photoelectron spectroscopy (XPS; VG Scientific ESCALAB 250) was used to examine the surface chemical compositions and valence states of metal oxides in SiO 2 . For XPS examination, all binding energies were referenced to C1s peak at 285 eV. Temperature-programmed reduction of H 2 (H 2 -TPR) in 10% H 2 /Ar was conducted using 10 mg of sample with a total flow rate of 50 mL min -1 . The sample was pretreated in a pure Ar flow at 300°C for 0.5 h and cooled to 50°C before measuring the H 2 -TPR. The sample was placed in dilute H 2 , and the H 2 consumption was monitored using a Micromeritics Autochem 2920 by increasing the temperature to 800°C at a rate of 10°C min -1 . Temperature-programmed desorption of ammonia (NH 3 -TPD) was also performed by a Micromeritics Autochem 2920 using 10 mg sample. The powder sample was first pretreated in 30 mL min -1 N 2 at 150°C for 1 h. Subsequently, the sample was cooled down to room temperature and saturated with 10% NH 3 /Ar for 1 h. Adsorbed NH 3 was removed using Ar flow (50 mL min -1 ) for 40 min before starting the TPD experiments. After saturation, the sample was flushed in a pure Ar flow for 30 min at 100°C. Finally, the sample was heated up to 750°C with a heating rate of 10°C min -1 . The amount of NH 3 desorption from the samples was quantified by a thermal conductivity detector.

Hg 0 /NO Removal Tests
The Hg 0 removal tests were carried out in a fixed-bed reactor with 10 mg sample using a simulated flue gas containing 30, 65, and 100 ± 5 µg m -3 Hg 0 at 150, 250, and 350°C. The simulated flue gas also contained 12% CO 2 , 10% H 2 O, 6% O 2 , 50 ppm HCl, 200 ppm SO 2 , 200 ppm NO, and balanced N 2 prepared from standard gas cylinders with a flow rate of 1.5 L min -1 at 25°C. Detailed descriptions regarding the experimental apparatus, flue gas generation, data acquisition, and removal efficiency calculation can be found in the Supplementary Material, Fig. S1, and elsewhere (Chiu et al., 2015(Chiu et al., , 2017Liu et al., 2017;Lin et al., 2019). Notably, the reason to test the Hg 0 removal at three different temperatures in this study was to better understand the removal dependence on flue gas temperature in coal-fired utilities, for which adsorption of Hg 0 is dominant at lower temperature and catalytic oxidation is dominant at higher temperature with presence of porous catalyst. Additionally, the flue gas temperature is typically around 150°C before the particle control device like electrostatic precipitator or fiber filter, it is the zone suitable for injecting powder adsorbents for Hg 0 removal. The deNO x SCR catalyst was often operated at around 350°C, at which Hg 0 can be oxidized into Hg 2+ and the downstream wet flue gas desulfurization can successfully remove Hg 2+ . For the NO removal experiment, the same fixed-bed reactor with 1.0 g sample was used, and the flue gas of 1.5 L min -1 contained 6% O 2 , 200 ppm NH 3 , 200 ppm NO, and balanced N 2 at 150 to 400°C. The test sample was placed in the reactor and heated to 200°C under 500 mL min -1 N 2 flow for 1 h to remove the water vapor inside the pore before conducting the removal experiment.

Characterization of Raw and Metal-oxide-incorporated SiO2
The physical properties of raw and modified rice-huskderived silica are listed in Table 1. For the raw SiO 2 , the S BET was 160 m 2 g -1 and V t was 0.298 cm 3 g -1 . The S BET and V t increased significantly after the incorporation of copper oxide, indicating that the presence of copper oxide did not block or shrink the pore structure of the raw SiO 2 ; instead, the structure might reconstruct during the hydrothermal synthesis process. Among all the copper-incorporated SiO 2 , 50%Cu/SiO 2 demonstrated higher S BET (210 m 2 g -1 ) and V t (0.473 cm 3 g -1 ) than 10% and 25%Cu/SiO 2 . Moreover, the S BET was further improved to 260 m 2 g -1 after CeO x addition, indicating that there might be synergistic effects between copper oxides and ceria oxides on building up new structure on the SiO 2 surface. In contrast, 50%Cu-10%Ce/SiO 2 had the lowest V t (0.278 cm 3 g -1 ) and the smallest average pore volume (5.2 nm), which might be due to the shrinkage of the original SiO 2 mesopores. This result is consistent with one previous research, in which the CuCeTi catalyst showed higher specific surface area than the CuTi catalyst, suggesting that the introduction of Ce led to highly dispersed copper species on the catalyst surface .
The X-ray diffraction patterns of raw SiO 2 and metal oxides-incorporated SiO 2 (Fig. 1) showed that no significant peaks corresponding to the crystal phases of SiO 2 were observed among all the test samples, which indicated that the rice-husk-derived silica was amorphous. For 10%Cu/SiO 2 and 25%Cu/SiO 2 , only weak diffraction peaks of cubic CuO (2θ = 35.29°, 38.49°) were found (Wang et al., 2013). The diffraction peaks became even weaker as the copper content increased. For 50%Cu/SiO 2 , no obvious peak attributed to CuO was detected. This indicated that the copper oxides were amorphous and/or highly dispersed on the surface of SiO 2 . Highly dispersed active sites on the surface of the catalyst can result in an enhancement of catalytic activity (Xiang et al., 2015). This result also suggested that there were interactions between copper oxides and SiO 2 and these interactions might lead to more amorphous species formed in SiO 2 . Some earlier studies also demonstrated the presence of an amorphous or poorly crystalline state in SCR or mercury removal catalysts (Shu et al., 2012;Xiang et al., 2015;Chi et al., 2017). These amorphous transition metal oxides showed larger Hg 0 catalytic activity than its crystalline phase on the MnO x -CeO 2 /TiO 2 catalyst . Fig. 2 shows the SEM images of raw and metal-oxide SiO 2 , indicating that raw SiO 2 comprised uniform spherical particles with sizes within 0.1-0.15 µm. Changes in surface morphology for SiO 2 were found after the incorporation of metal oxides. The SEM images clearly demonstrated that the particles turned into plate-like structure, confirming that the presence of metal oxides could change the surface morphology of SiO 2 particles. Compared to Fig. 2 Fig. 2(c) shows that 50%Cu/SiO 2 had more plate-like particles and the size of the particles was smaller and more uniform. This result can be attributed to its higher copper content. As the copper loading increased, the atomic arrangement of CuO x /SiO 2 may be closer, which can result in the smaller particle sizes. As shown in Fig. 2(d), the 50%Cu-10%Ce/SiO 2 possessed some layer  structures with many fine metal oxide aggregates attached on them. This result was similar to our previous study (Chiu et al., 2017) that the aggregation and particle growth did not decrease the specific surface area, which indicated that wellstructured mesopores were created. The TEM image of raw SiO 2 (Fig. 3(a)) shows that the typical nanoparticles were well dispersed with the size in the range of 10-15 nm. The porous structure had advantages for dispersion of metal oxides. As the copper content increased, the nanoparticles turned into silk-like with a uniform size and the precipitant of copper oxides started to appear on the SiO 2 surface. The image of 50%Cu-10%Ce/SiO 2 (Fig. 3(d)) shows that great amounts of copper oxide and ceria oxides were present on the surface of SiO 2 and these metal oxides were formed as clusters. This is consistent with the XRD result that uniformed CuO and CeO x crystallinities were detected on the surface of 50%Cu-10%Ce/SiO 2 .
Four samples were further analyzed by XPS in order to investigate the chemical state of species on the surface, and the obtained O1s, Cu2p, and Ce3d spectra are shown in Figs. S2 and S3 and Table 2. The spectra of O1s for the four samples were divided into three single peaks, lattice oxygen at 529.6-530.6 eV (O α ), chemisorbed oxygen at 531.0-532.0 eV (O β ), and oxygen in hydroxyl form and/or surface adsorbed water at 532.0-533.0 eV (O γ ) . Results showed that as the copper content increased, the amount of O γ decreased and the amount of chemisorbed oxygen increased. Zhang et al. (2017) indicated that chemisorbed oxygen played an important role in redox reactions and resulted in enhancing Hg 0 removal efficiency because of its high mobility. In addition, the high ratio of O β was also helpful for the NO removal efficiency since more NO can be oxidized to NO 2 and enhanced the "fast SCR" reaction: 2NH 3 + NO + NO 2 → 2N 2 + H 2 O (Chi et al., 2017). Although the ratio of chemisorbed oxygen decreased after the addition of ceria oxide, the combination of CuO x and CeO x might facilitate the conversion of chemisorbed oxygen and gasphase O 2 and lead to the high NO removal efficiency .
The Cu2p XPS spectra for the metal oxide-incorporated SiO 2 are shown in Fig. S3. The binding energy peaks at 932.2-932.6 eV could be attributed to Cu + while the peaks at 933.2-934.4 eV could be regarded as Cu 2+ (Chi et al., 2017). The ratio of Cu 2+ obviously increased as the copper content increased from 10% to 50%; however, the ratio of Cu 2+ decreased after the addition of ceria oxides. Cu 2+ was considered to play an important role in NO removal since it can adsorb, activate NO molecules and favor the SCR reaction by the redox cycle between Cu 2+ and Cu + . Therefore, the higher Cu 2+ ratio on the 50%Cu/SiO 2 can be a reason for its greater NO removal efficiency.
The deconvolution of XPS core-level O1s, Cu2p, and Ce3d is summarized in Table 2. The results indicated that the relative ratio of Ce 3+ calculated by Ce 3+ /(Ce 3+ + Ce 4+ ) on the 50%Cu-10%Ce/SiO 2 was 41.07% and Ce 4+ was 58.93%. Ceria oxide has great ability to store and release oxygen through the redox shift between Ce 3+ and Ce 4+ (Kwon et al., 2015). In addition, Ce 3+ could create charge imbalance, vacancies, and unsaturated chemical bonds on the surface of catalyts, which could result in more oxygen defects. Therefore, the higher ratio of Ce 3+ could promote the oxidation of NO to NO 2 , thus increasing the NO removal efficiency at low temperature (Shu et al., 2012). 50%Cu-10%Ce/SiO 2 with the high content of Ce 3+ might provide its best NO removal efficiency among all the samples.
In order to investigate the redox property of the metal oxide-incorporated SiO 2 , H 2 -TPR was conducted and shown in Fig. S4. For 10, 25, and 50%Cu/SiO 2 , two different peaks were found. The low temperature peak at around 225°C could be attributed to the reduction of isolated Cu 2+ → Cu + , while the high temperature peak at around 325°C could be    (Niu et al., 2016). After the addition of ceria oxide, the peak at higher temperature disappeared, and one large peak at about 266°C was observed. This result suggested that the interaction between copper and ceria oxides might lead to an easier reduction of ceria and CuO (Mrabet et al., 2012). NH 3 -TPD was carried out to further investigate the acid sites on the metal oxide-incorporated SiO 2 . The results are shown in Fig. S5. All the samples had a broad peak that started at 150°C and ended at around 450°C, which could be regarded as ammonia desorbed from Brönsted acid sites (Pan et al., 2014), and the peak of 50%Cu/SiO 2 at higher temperature (about 600 to 700°C) could be assigned to the NH 3 desorbed from Lewis acid sites (Boxiong et al., 2014). Since Lewis acid sites are the primary acid sites evolved at high temperatures, the great amount of Lewis acid sites on 50%Cu/SiO 2 might improve the NO removal efficiency at higher temperature. The peak area of the 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 was much larger than that of the other two samples. Because the quanity of NH 3 is proportional to the surface acid amount, this result could be concluded as that the 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 samples had stronger acidity. These results were also consistent with the H 2 -TPR results and might lead to the better catalytic activity of 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 . Fig. 4 clearly shows that the raw SiO 2 did not demonstrate any NO removal efficiency from 200 to 300°C and the removal efficiency became negative with further increase in temperature, which was mainly due to the direct reaction of NH 3 with O 2 to form nitrogen oxides (Piubello et al., 2016). In contrast, the NO removal efficiency of CuO x /SiO 2 first increased with the increasing temperatures and then decreased when temperature was higher than 300°C, which might also be due to the NH 3 oxidation at high temperature. Similar test results can be found in previous research that the NO conversion efficiency of 8%Cu/SCR catalyst started to decrease when the temperature was above 300°C (Chi et al., 2017).

NO Removal Efficiency of Metal-oxide SiO2 NO Removal of Metal-oxide SiO 2 at Various Metal Components and Temperature
The NO removal efficiency also increased as the copper content increased. 25%Cu/SiO 2 exhibited similar NO removal efficiency with 10%Cu/SiO 2 at temperature below 300°C but became greater at higher temperature. 50%Cu/SiO 2 showed a significant improvement in NO removal efficiency at all test temperatures, which might relate to its largest S BET and a greater amount of Cu 2+ and surface chemisorbed oxygen.
The influence of CeO x modification on the improvement of NO removal efficiency for CuO x /SiO 2 was also investigated. The addition of ceria oxides successfully improved the NO removal efficiency at low temperature. 50%Cu-10%Ce/SiO 2 showed the best NO conversion efficiency of 70-85% and a broad temperature window of 150-300°C. This result can be attributed to its large S BET (i.e., 260 m 2 g -1 ) and great amount of Ce 3+ ( Table 2) that can create charge imbalance, vacancies, and unsaturated chemical bonds on the catalyst surface , thus increasing the ability to adsorb, store, and transfer the surface oxygen and further enhancing the deNO activity at low temperature (Sohot et al., 2012). Li et al. (2015) also showed that the NO reduction efficiency of CuCeTi catalysts was significantly higher than that of CuTi and CeTi catalysts in low-temperature range (150-250°C), which was similar to our test results. Because 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 had the superior NO removal efficiency among all the test samples, these two samples were futher investigated under different flue gas conditions.

Effect of H 2 O and SO 2 on NO Removal
To clarify the influence of H 2 O on the NO removal test over 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 , 10% H 2 O was initially injected into the gas mixing chamber and then stopped injection after 2 h. Fig. 5(a) shows the effect of H 2 O on deNO x efficiency of the samples at 250°C. The NO removal efficiency of 50%Cu/SiO 2 rapidly decreased to about 28% after the addition of H 2 O, but the deactivation was reversible since the removal efficiency recovered to its original level after stopped injecting H 2 O. In contrast, the deactivation caused by H 2 O was less obvious for the 50%Cu-10%Ce/SiO 2 sample, which might be due to its larger S BET . This phenomenon has been found in several studies showing that H 2 O has a reversible negative effect on the NO conversion efficiency at lower temperature (Du et al., 2013;Pan et al., 2013). The inhibition effect by H 2 O can be attributed to competitive adsorption between H 2 O and NH 3 on the active sites (Pan et al., 2013).
The effect of H 2 O at 300°C is shown in Fig. 5(b). The inhibition caused by H 2 O became less significant as the temperature increased to 300°C. Moreover, the addition of H 2 O showed a promotion on NO removal over 50%Cu-10%Ce/SiO 2 at 300°C, and the NO removal efficiency slightly decreased after H 2 O was removed. This result is similar to previous research that the NO conversion over Cebased catalyst was improved after the introduction of H 2 O at higher temperature and the promotional effect could be attributed to the inhibition of NH 3 oxidation (Du et al., 2013;Xiao et al., 2016). Fig. 6 shows the effect of SO 2 on the NO removal efficiency of 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 at 250 and 300°C. 200 ppm of SO 2 was injected into the gas mixing chamber for 900 min under this condition. The NO removal efficiency was calculated after the 15 h test. As shown in Fig. 6(a), the NO removal efficiency of both two samples significantly decreased after the injection of SO 2 at 250°C. The NO reduction activity of 50%Cu/SiO 2 decreased from 70.7% to 24.6%. 50%Cu-10%Ce/SiO 2 showed a better SO 2 resistance since the NO removal efficiency decreased from 77.9% to 42.1% after the addition of SO 2 at 250°C. The reduction in NO removal efficiency at low temperature might cause by the formation of ammonia sulfate or sulfite salts such as (NH 4 ) 2 SO 3 , (NH 4 ) 2 SO 4 , and NH 4 HSO 4 . These species could deposit on the surface of the samples and blocked the active sites, thus restrained NO and NH 3 adsorption and further decreased the SCR activity (Ma et al., 2013). Xie et al. (2004) also indicated that the formation of CuSO 4 during the reaction could deactivated the catalyst by either pore filling and/or plugging or by its low NO removal efficiency compared to the CuO. The role of ceria on SO 2 resistance has also been investigated. Kwon et al. (2015) indicated that the V/Sb/Ce/Ti catalyst showed an excellent SO 2 resistance compared to V/Sb/Ti catalyst since the formation of NH 4 HSO 4 over the V/Sb/Ce/Ti catalyst could be inhibited by consuming SO 2 in the formation of Ce 2 (SO 4 ) 3 . Fig. 6(b) shows the effect of SO 2 on NO removal efficiency at 300°C. The results revealed that the NO reduction activity over 50%Cu/SiO 2 almost remained constant after the injection of 200 ppm SO 2 . The addition of 200 ppm SO 2 even led to a promotion effect over 50%Cu-10%Ce/SiO 2 at 300°C since the NO removal efficiency slightly increased from 71.5% to 76.3%. Xie et al. (2004) indicated that ammonia sulfate salt such as (NH 4 ) 2 SO 4 and NH 4 HSO 4 , which were considered to be the deactivating agents for SCR catalysts at low temperatures, would decompose at temperatures of 250-450°C. Xiao et al. (2016) also showed that the introduction of SO 2 caused a promotion effect of NO reduction over the Ce/TiO 2 catalyst at higher temperature, and the promotion could be attributed to the sulfation of catalyst. The Ce/TiO 2 catalyst demostrated a great NH 3 adsorption ability and the the catalytic oxidation of NH 3 to NO was inhibited after the sulfation at high temperature.

Hg Removal Test Hg Removal Efficiency of Various Metal-oxide SiO 2
The total Hg (T-Hg) removal efficiency was tested with Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 . The results indicated that none of these four samples reached adsorption equilibrium during the test duration of 900 min, indicating that the adsorption did not obtain complete breakthrough. Fig. 7(a) shows that 50%Cu/SiO 2 had a higher average T-Hg removal efficiency (88.2%) than 10%Cu/SiO 2 (85.3%) and 25%Cu/SiO 2 (82.5%). This test result is expected since 50%Cu/SiO 2 had the largest S BET and the greatest amount of CuO x highly dispersed on the surface. Furthermore, the high ratio of Cu 2+ and chemisorbed oxygen can also lead to its greatest mercury removal efficiency. This result is similar to our previous research that a very large amount (approximately 50%) of CuO x was introduced into mesoporous SiO 2 and presented an excellent performance for Hg 0 removal under the flue gas condition (Chiu et al., 2017). Fig. 7(b) shows that the T-Hg removal efficiencies of 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 were almost the same during the first 400 min. However, the T-Hg removal efficiency of 50%Cu-10%Ce/SiO 2 slightly decreased to about 55% at the end of the Hg 0 removal test. This result indicated that the modification of CeO x did not enhance the T-Hg removal efficiency and the durability of the sample was also decreased. Although 50%Cu-10%Ce/SiO 2 had the largest specific surface area, the CuO and CeO x crystallinity and the lower ratio of Cu 2+ and surface chemisorbed oxygen might be the reason for its poor durability.

Effect of Inlet Hg 0 Concentration and Temperature on Hg
Removal Efficiency of Metal-oxide SiO 2 Table 3 shows that 50%Cu/SiO 2 had great T-Hg removal efficiency when the inlet Hg 0 concentration increased to 65 µg m -3 . However, the average removal efficiency decreased to 63.1% at 100 µg m -3 inlet Hg 0 concentration. The corresponding adsorption capacity was 3330.6, 7769.9, and 8989.2 µg g -1 at 30, 65, and 100 µg m -3 Hg 0 , respectively. Table 4 shows that the T-Hg removal efficiency for 50%Cu/SiO 2 remained almost the same as the temperature increased from 150 to 250°C. When the temperature increased to 350°C, the removal efficiency decreased to 55.3%. Futhermore, the Hg adsorption capacity was 3330.6, 3421.5, 2524.4 µg g -1 at 150, 250, and 350°C, respectively. These results indicated that the 50%Cu/SiO 2 still possessed an excellent adsorption performance even when temperature increased to 250°C, and the decrease of T-Hg removal efficiency at 350°C was expected since the adsorption of Hg was thermodynamically unfavorable at higher temperature (Chiu et al., 2015).

CONCLUSIONS
This study investigated the effects of metal oxide incorporation with rice-husk-derived silica on the physicochemical properties and NO/Hg removal efficiency of CuO x -and CuO x -CeO x /SiO 2 from simulated flue gases.  The goal was to develop a new SCR catalyst and Hg adsorbent from two wastes: copper-containing wastewater and rice-husk-derived silica, and to understand the effects of flue gas conditions on removal of Hg 0 and NO. The primary conclusions are summarized as below: (1) The specific surface area of rice-husk-derived SiO 2 significantly increased after the surface modification by metal oxides. XPS results showed that Cu 2+ and Ce 4+ were the major valence states presenting in metaloxide SiO 2 . H 2 -TPR and NH 3 -TPD indicated that 50%Cu/SiO 2 and 50%Cu-10%Ce/SiO 2 had better redox ability and stronger acidity.
(2) The incorporation of Cu and Ce oxides successfully improved the NO removal efficiency. 50%Cu-10%Ce/SiO 2 showed the best NO removal efficiency and a broad temperature window, which may be due to its highest specific area and great redox ability and acidity.
(3) 50%Cu/SiO 2 exhibited greatest T-Hg removal efficiency among all the tested samples under flue gas condition at 150°C, which may be due to its large specific surface area and great amount of CuO x highly dispersed on the surface. (4) Rice-husk-derived SiO 2 incorporated with copper ion recycled from industrial wastewater has been shown successfully fabricated via silicate-exfoliation hydrothermal processes and is verified with good removal effectiveness for both Hg 0 and NO. As noted earlier, because the SiO 2 support and Cu ion are both recycled from agricultural and industrial wastes, the resulting products can be considered as environmentfriendly catalysts and adsorbents.