In Situ Hydrogen Uptake and NO x Adsorption on Bifunctional Heterogeneous Pd/Mn/Ni for a Low Energy Path toward Selective Catalytic Reduction.

Facilitating catalyst accessibility of H2 and NO x at the catalyst surface remains a great challenge in catalytic selective catalytic reduction (SCR). The efficient conversion of NO x into N2 under mild conditions is an attractive pathway as SCR usually requires high operating temperature which consumes extra operating energy and restricts the possible locations of an SCR device. The H2 supply concentration of conventional H2-SCR is relatively sparse (0.5-2%), which leads to a relatively high operating temperature to activate H. We developed a H2-SCR process with the monolithic catalyst which combined with localized rarefied hydrogen enrichment enhanced by porous nickel and adsorption of NO x on Mn oxide with only 0.08, 0.25, and 0.42% palladium can achieve over 80% NO removal efficiency at 120, 100, and 90 °C. Maximizing the role of nickel foam-fixed hydrogen and Mn oxide in combination with NO can provide enriched NO x and H2 atmosphere for adjustable valence state Pd to yield positive catalytic behavior.

Introduction

Aggravation of air pollution has motivated an intense research effort aimed at improving the state of the atmospheric environment, and NO has been regarded as an important environmental pollution source. SCR (selective catalytic reduction) is generally accepted as the dominant denitrification technology. The current mainstream technology NH3-SCR required a high operating temperature that restricts the location of the SCR device and must be placed before the flue gas desulfurization and the dust-extraction unit; as a consequence, the inevitable toxic effect of SO2 limits the lifetime of the catalyst. For many existing rectification factories, very little space is available for adding additional waste gas treatment equipment. Therefore, a lower operating temperature and environmentally friendly SCR technology H2-SCR is emerging as a promising technique in denitrification.[1−8] Its low operating temperature makes it possible for the flexible placement of the SCR device, for example, with the SCR device placed after the desulfurizer or the dust-extraction unit. This has a significant positive effect on the life of the catalyst and therefore, lowers the cost of denitrification.

Although, H2-SCR can operate at low temperatures and solve the problem of the energy dissipation, hydrogen storage is harsh and limited. Hence, various online hydrogen production technologies have emerged such as steam methane reforming and auto-thermal reforming of diesel fuel.[9] Meanwhile, the H2 supply concentration of conventional H2-SCR is relatively sparse (0.5–2%), which leads to a relatively high operating temperature to provide H activation. The efficient utilization of hydrogen resources is still a major challenge in the field that must be overcome.

Conventional H2-SCR catalysts are mainly based on noble metals, mostly using Pt and Pd.[10−13] Although the use of Pt or Pd as an active component meets the process requirements, the costly high operating temperature limits their practical application. Therefore, to support the implementation of H2-SCR, it is necessary to maximize the effective utilization of trace noble metal at lower temperatures and to search for a base metal substitute for Pt and Pd.

Earth-abundant nonnoble metal elements such as transition metals (e.g., Co, Fe, Cu, and Mn) have been reported to show extraordinary SCR activity at low temperatures.[10,14−23] The multivalent nature of these oxides could facilitate the redox activity of the catalyst which is attributed to the enhancement of efficiency. Nickel foam is attractive, given its desirable three-dimensional structure and excellent H adsorption capacity,[23−27] which has been widely used as an electrode substrate commercial material. With its favorable 3D cross-linked grid structure and capacity of H2 adsorption, it can not only localize rarefied hydrogen enrichment but also expose more active materials per unit. Hence, nickel foam can be the substitution of non-noble host structures of choice.

In this work, we developed a H2-SCR process with the monolithic catalyst which combined with localized rarefied hydrogen enrichment enhanced by porous nickel and adsorption of NO on Mn oxide with only 0.08, 0.25, and 0.42% palladium can achieve over 80% NO removal efficiency at 120, 100, and 90 °C under conditions of 2% O2 and gas hourly space velocity (GHSV) = 32 000 h–1. We emphasize the role of nickel foam and Mn oxide, which provide Pd with an enriched NO and H2 atmosphere for improving reaction efficiency. Via various characterization methods, such as morphology characterization and surface element chemical valence and composition detection, the likely reaction process of denitrification was elucidated. Because of its low operating temperature and high efficiency, our candidate catalyst has the advantage of a more flexible selection of the flue gas denitrification location.

Experimental Section

Catalyst Preparation

All analytical grade (AR) reagents were purchased from Sinopharm Chemical Reagent Co. LTD without further purification. Before nickel foam was put into use, it needed to be pretreated to ensure the removal of the probable oxidation layer and organic substances on its surface. Mn (NO3)2 and PdCl2 were deployed into a precise concentration solution, respectively.

In a typical synthesis, a calculated amount of preconfigured Mn (NO3)2 and PdCl2 solution were infused in a 100 ml beaker under magnetic stirring for 30 min. During stirring, 0.3 g NaOH, 500 μL H2O2, and 114 mg NaBH4 were added to make mixed solutions. The mixture was then transferred to a high-pressure reactor lined with a Teflon liner for 12 h of hydrothermal reaction at 150 °C. After the reaction process, the autoclave was allowed to cool down naturally. Then the products were leached with deionized water several times. The desired catalysts were obtained and denoted as the PdO–Mn3O4 catalyst.

As for the PdO–Mn3O4@nickel foam catalyst: 40 mg (19.35% loading percentage) of the prepared PdO–Mn3O4 catalyst was distributed in 5 mL isopropanol upon ultrasonic vibration followed by the impregnation method to obtain the desired catalysts. Besides, the purified single foamed nickel was used as the blank control group.

Catalyst Characterization

Different characterization methods were used to exhibit the internal structure of the catalyst: X-ray diffraction (XRD) patterns were recorded between 10 and 80° by using a D/max 2550 V X-ray diffractometer with a copper target (k = 1.5406 Å) and operating at 40 kV and 100 mA. The elementary composition and morphology characteristics of the samples were examined by energy-dispersive spectroscopy (EDS) using an energy-dispersive X-ray analyzer (Falion 60S), and scanning electron microscopy (SEM) equipped with EDS. Field-emission SEM (FESEM) was using to gain EDS color mapping. The transmission electron microscopy (TEM) and high-resolution TEM (HTEM) images were used to further demonstrate the topographic characteristics equipped with a JEM-2100 transmission electron microscope. X-ray photoelectron spectroscopy (XPS) analysis was conducted by point and neutral neutralization electron gun; ion beam EX06 device, and XR6 micro focusing monochromatized XPS (Al Kα) device, which determined the surface valence state of the catalyst. Nitrogen adsorption–desorption analysis was carried out using a Micromeritics Tristar II 3020 apparatus and the subsequent information about the specific surface area, pore volume, and average pore size of the samples were then computed using the Brunauer–Emmett–Teller model.

H2-TPR (temperature-programmed reduction by hydrogen) and NH3-TPD (temperature-programmed desorption of ammonia) were conducted by an automatic temperature-rising chemical adsorber (Micromeritics, Autochem II 2920). Before adsorption on the surface of the catalyst, the 50 mg dry catalyst samples need to be pretreated under 300 °C within 60 min to remove the impurities and moisture. In H2-TPR experiments, the sample was exposed to 10% H2/Ar reduction gas (40 ml/min), and the temperature was elevated continuously at 10 °C/min to 800 °C. In the NH3-TPD experiments: the sample was first exposed to 10% NH3/He mixture (40 mL/min) under 50 °C, to complete the adsorption of NH3 60 min; then switched to He gas (40 mL/min) purification within 60 min; finally the temperature was elevated continuously at 10 °C/min to 800 °C.

Catalytic Activity Procedure

The synthetic catalyst was placed in a fixed-bed flow reactor for H2-SCR. The simulated flue gas containing 900–1000 ppm NO, x % O2, 9000–10 000 ppm H2, and N2 as balance gas was used in the H2-SCR reaction. A GHSV of about 32 000 h–1 was maintained by a Brooks thermal mass flowmeter. During the reaction, the temperature was controlled within the scope of 50–150 °C by K-type thermocouples inside and outside the fixed bed. The concentration of gas in the effluent of the reactor was analyzed using the online chemiluminescence NO–NO2–NO analyzer (Thermofisher, 42i model). The final total conversion is calculated with the stable data. The eqs –3 used to calculate the NO conversion, H2 conversion, and N2 selectivity werewhere [NO]in, [H2]in, [NO]out, and [H2]out represents the inlet and outlet gas concentration of the system at each temperature. N2O concentration and H2 concentration were analyzed by an online gas chromatograph (GC 7890) equipped with a Hayesep Q column (i.d. 2 mm, length 35 cm), and a thermal conductivity detector.

Results and Discussion

Synthetic Method, Structural, and Morphological Characterization of the PdO–Mn3O4@Nickel Foam Catalyst

The fabrication of the PdO–Mn3O4@nickel catalyst is schematically described in Figure and involves a facile two-step route detailed in the Experimental Section: first, using the hydrothermal method to acquire PdO–Mn3O4 and then obtaining the loaded PdO–Mn3O4@nickel foam catalyst by impregnation (detailed in Experimental Section). According to our synthetic process (displayed in Figure c), the solution environment could form birnessite at ambient temperatures, and then during the hydrothermal method, Pd nanoparticles begin to grow dispersive, hence avoiding agglomeration of catalysts. The three-dimensional cross-linked grid structure of the foamed nickel can be observed by SEM images in Figure a,b, which provided high porosity to enlarge the more specific surface area and exposed more possible active sites. Comparing Figure b with 1a, impregnation by PdO–Mn3O4 resulted in a significant morphology change in the foamed nickel surface. As observed from the TEM image presented in Figure b, flaky, polyhedral Mn3O4 with uniformly dispersed Pd nanoparticles (marked by red circles) was revealed, forming a regular three-dimensional and hierarchical porous superstructure.

Figure 1

(a,b) SEM images of bare nickel foam and the loaded PdO–Mn3O4@nickel foam catalyst; (c) chemical synthesis step of the PdO–Mn3O4 catalyst; (d–f) HTEM images of the PdO–Mn3O4 catalyst [(e): built-in diagram displays the particle size of Pd nanoparticles; (f): red lines denote the Pd lattice spacing; yellow lines denote the lattice spacing of base Mn3O4; illustration denotes the SAED pattern of PdO and Mn3O4 crystals].

XRD patterns and the corresponding HTEM images of the samples were further obtained to characterize the composition of the samples. PdO–Mn3O4@nickel foam samples (shown in Figure S1) exhibited distinct XRD peaks at 29.0, 32.4, 36.0, and 60.2° of the (112), (103), (211), and (224) reflection planes, respectively, which correspond to Mn3O4 (hausmannite, PDF 18-0803). The grain size of the catalysts was calculated using Scherrer’s formula, and it was found that the average crystal size was about 30.6 nm, which matched well with the HTEM images. With the increasing of Pd, still, no obvious Pd or PdO peaks were observed, which is possibly because the uniformly dispersed PdO nanoparticles were too small to reach the detection limit. The material has a spinel mesostructure of tetrahedral crystals as determined by TEM. Pd nanoparticles were observed (marked by a blue ellipse in Figure e) and were highly dispersed on the parallelogram surface of Mn3O4. In more detail, the inset figure in Figure e showed that the Pd nanoparticles were anchored evenly over the Mn3O4 hosts and had a narrow distribution in the average size of 3 to 10 nm. The lattice fringes of the material displayed interplanar spacings of 1.74 and 1.60 Å in the particles, which matched well, respectively, with those of the Mn3O4 (312) crystal plane and the PdO (200) crystal plane, the inset figure in Figure f showed the SAED pattern of PdO and Mn3O4 crystals. The d-spacing of the Pd lattice stripes is smaller than that of pure PdO, which indicated that various Pd element forms exist. FESEM was used to reveal the elemental composition of the catalyst, and the results were in line with expectations (shown in Figure S2).

H2-SCR Evaluation, Valence States, and Redox Capacity of the PdO–Mn3O4@Nickel Foam Catalyst

As the Pd content increased, the denitrification efficiency of the PdO–Mn3O4@nickel foam catalyst improved. Notably, unlike for the other reported medium-temperature area Pd-based catalysts,[10] the temperature window of the de-NO reaction was advanced. Remarkably, trace Pd doping of 0.08% PdO–Mn3O4@nickel foam still enabled the removal of nearly 90% NO at 120 °C. The enhancement of denitrification efficiency may be related to the change in valence, the increase in the oxygen vacancy content, and the variety in the acid content on the surface.

X-ray photoelectron spectra of the samples were measured to gain insight into the element chemical valence at the surface. As the main catalytically active center of the reaction, the Pd nanoparticles show a significant splitting between the Pd 3d orbital peaks. Two different photoelectron peaks that were ascribed to the metallic Pd species and Pd2+ were observed at approximately 335.1 and 340.8 eV, respectively. According to the synthesis steps, the addition of NaBH4 should reduce some of the Pd atoms from the divalent states to the metallic state, giving rise to a partial split peak. With the increasing of Pd, the split peak also showed an increasing trend (shown in Figure S2c), while loading on nickel foam, a more obvious split peak was emerged (shown in Figure b). Because foamed nickel could enhance the ability of ion transport and maintain very smooth electron channels,[28] so it can provide Pd a place to transform electrons, which attributed to the increase of the split peak of the low valence state Pd.

Figure 2

(a) H2-SCR performance of the prepared PdO–Mn3O4@nickel foam catalyst with different Pd contents tested in a fixed bed. The simulated flue gas including 1070 ppm of NO, 10 700 ppm of H2, 2% O2, and a GHSV of 32 000 h–1, using N2 as balance gas; XPS patterns of (b) Pd 3d spectra of the prepared PdO–Mn3O4 and PdO–Mn3O4@nickel foam catalyst; (c) Mn 2p spectra of the prepared PdO–Mn3O4 and PdO–Mn3O4@nickel foam catalyst; and (d) O 1s spectra of the prepared PdO–Mn3O4 and PdO–Mn3O4@nickel foam catalyst.

The splitting of the Mn 3s peaks was because of the interaction between the nonionized 3s electrons and the three-dimensional valence band electrons, showing the difference in the actual Mn oxidation state. Generally, the Mn 3s peak splitting width varies from 4.5 to 6.0 eV. MnO2 peaks are found at approximately 4.5 eV, while those of Mn2O3 and Mn3O4 are observed at approximately 5.5 eV, and those of MnO are observed at approximately 5.9 eV.[29] Upon Pd ion doping, Mn 3s splitting increased from 5.58 to 5.67, and when Mn3O4 was placed on nickel foam, Mn 3s peak division increased to 5.71 eV (shown in Figures c and S2a). This indicates that the average Mn oxidation state decreased and was between +3 and +4. This is in good agreement with the Mn 2p level binding energy change (shown in Figure.S2b). Differences in the Pd content of the PdO–Mn3O4@nickel foam catalyst appeared to increase the MnIV content, which may have led to enhanced denitrification.

The strongest O 1s photoelectron peak (Figure d) with a binding energy of approximately 529.3 eV was attributed to lattice oxygen. The PdO–Mn3O4@nickel foam catalyst O 1s orbitals showed an extra splitting peak. The presence of Mn–O–Mn (529.3–530.3 eV), Mn–OH (530.5–531.5 eV), Ni–O (529.0–531.0 eV), Ni–OH (530.8–531.7 eV), and H–O–H (531.8–532.8 eV) could be attributed to various chemisorbed hydroxyl (OH−) groups. These chemical adsorptions of oxygen could provide more oxygen vacancies that contributed to NO adsorption and dissociation.[30]

To further illustrate the possible reasons at the enhancement of deNO performance by our candidate, H2-TPR coupled with NH3-TPD technology XPS methods were further performed, which could obtain more information in materials’ redox capability. Figure a shows the characterization results of catalyst H2-TPR. Both the catalysts had negative peaks at 73 and 75 °C that indicated that palladium oxide was reduced to Pd metal. The reduction temperatures of 196 and 213 °C as shown in Figure a are much lower than expected. H2 reduction peaks on MnO/TiO2 reported at 282 °C were attributed to the transformation from MnO2 to Mn2O3; the peak at 348 °C was attributed to the transformation of Mn2O3 to Mn3O4 and the peak at 432 °C was attributed to the transformation of Mn3O4 to MnO.[31] The activated hydrogen can be absorbed on the surface of Pd particles by dissociation and reduce Mn species, making the position of reduction peak shift to low temperature.[32,33] The coexistence of different reduction peaks further indicate the presence of a mixture of manganese oxides Pd nanoparticles, loaded on the above foamed nickel supporter, exposed more active sites for hydrogen dissociation, so the consumption of the activated H atom accelerated the adsorption rate of fresh hydrogen, thus enhancing the activity. The synergistic effects between Pd, Mn, and Ni were prone to changing the oxidation state of a catalyst under the reaction conditions, which intensively altered the species of chemisorbed oxygen. Added oxygen vacancies on the catalyst activated the adsorption of NO and oxygen and therefore facilitated the H2-SCR reaction.

Figure 3

(a) H2-TPR and (b) NH3-TPD profiles of the PdO–Mn3O4 catalyst and the PdO–Mn3O4@nickel foam sample.

The ammonia temperature-programmed method (NH3-TPD) (shown in Figure b) confirmed the increase in the number of acid sites. Based on the data in the literature,[34] the strong physical adsorption of NH3 and surface chemical adsorption of the NH3 stripping peak correspond to the alpha peak in the temperature range of 100–150 °C. Moreover, the NH3 stripping peak attributed to the defects and weak Brønsted and Lewis acid was the beta peak, and its stripping temperature was between 150 and 300 °C. According to the NH3 desorption temperature, the acid can be defined as strong acid (250–350 °C), super strong acid (>500 °C), and weak acid (<250 °C).[35] Therefore, the different peak patterns of the PdO–Mn3O4 catalyst and PdO–Mn3O4@nickel foam catalyst corresponded to ammonia and the combination of the catalyst surface acid sites. It is observed that the PdO–Mn3O4@nickel foam catalyst displayed two obvious stripping peaks at 385 and 685 °C, respectively, and the increasing number of acid sites is strong acid sites. Nickel oxide can broaden the distribution of acid sites, which make the catalyst more advantageous for the reduction of nitric oxide.

Optimized H2-SCR Catalyst Performance

Unlike the bare nickel foam and PdO–Mn3O4, the PdO–Mn3O4@nickel foam catalyst exhibited 90% denitrification conversion at an ultra-low-temperature of 90 °C, which was much better than the results obtained for other known Pd-based catalysts (Figure f). Different impact factors were discussed in this section.

Figure 4

H2-SCR performance of the prepared PdO-Mn3O4@nickel foam catalyst tested in a fixed bed. The simulated flue gas including 1070 ppm NO, 10 700 ppm H2, 2% O2, and a GHSV of 32 000 h–1, using N2 as balance gas; (a) effect of different O2 feed concentrations on NO conversion over the prepared PdO–Mn3O4@nickel foam catalyst; (b) effect of GHSV on NO conversion by H2-SCR over the prepared PdO–Mn3O4@nickel foam catalyst; (c) effect of the H2/NO ratio on NO conversion by H2-SCR over the prepared PdO–Mn3O4@nickel foam catalyst; (d) H2 conversion over the prepared PdO–Mn3O4@nickel foam catalyst; (e) N2 selectivity over the prepared PdO–Mn3O4@nickel foam catalyst; and (f) H2-SCR performance comparison of different known catalysts at 90 °C.

For most catalysts, excess O2 will strongly inhibit the reduction of NO, as the oxidation of the H2 reducing agent will greatly reduce the effect of the reducing agent.[11,36,37] As for our candidate, H2 conversion was exhibited on a high level within a tested scale (Figure d), which indicated that the unexpected side reaction of H2–O2 combustion can be effectively avoided at low temperatures. Moreover, O2 can also act as a promoter; however, NO oxide was found in the adsorbed state of nitrate or nitrite, and H2 then has a stronger effect for the reduction of nitrate or nitrite, thus increasing the strong H2 reducing capability of the SCR technology.[38] In Figure a, with the increase in the O2 concentration, the PdO–Mn3O4@nickel foam catalyst also showed less obstruction of NO conversion.

The whole denitrification conversion curve is a typical volcanic curve that was repeatedly observed in previous work.[10,39] Two maximum removal efficiency temperatures were found. With GHSV increasing, the conversion of NO dropped from 92 to 73%. Owing to the three-dimensional space structure inside nickel foam, a large amount of the reaction gas could be absorbed and stored in this extremely large space. Hence, the reaction has a certain buffer effect, improving its ability to resist to a certain degree, so when the gas space velocity increased within a certain range, there was less obstruction for the catalyst H2-SCR denitrification effect. However, when the gas space velocity increased to 43 000 h–1 because of the large gas space velocity, the nickel foam cushion was not sufficient to maintain the full reaction, and the reaction contact time between the catalyst and the simulated gas was too short, leading to an obvious decrease in the NO removal efficiency.

It was found that an increase in the H2/NO ratio was advantageous to the H2-SCR denitrification. Generally, the NO removal efficiency of the catalysts was comparatively high for different H2/NO ratios. For H2/NO = 5, only 58% removal efficiency was maintained, indicating that the H2/NO ratio had little influence on the catalyst. The effect of the H2/NO ratio on the catalyst was attributed mainly to the effect of nickel foam because nickel foam has very high adsorption ability, it sets the role of hydrogen, and the effect of the H2/NO ratio on the catalyst was weak within certain limits.

N2 selectivity has also been tested in our project, thus the main byproduct N2O is verified by our experiments. Pd content showed a significant effect on N2 selectivity, and coupled with better NO conversion, our candidate had an increasing trend on N2 selectivity with the increase in the testing temperature.

Conclusions

In summary, we developed a H2-SCR process which combined localized hydrogen enrichment enhanced by a porous nickel with NO adsorption on the modified Mn oxide catalyst. Doped with 0.08, 0.25, and 0.42% palladium, the monolithic catalyst can achieve over 80% NO removal at 120, 100, and 90 °C, respectively. Mn oxide prefers the combination with NO and could anchor well-dispersed Pd nanoparticles on the surface, so the catalyst achieved two functions simultaneously: (1) enrichment of H2 by nickel foam via physical and chemical adsorption and (2) conversion of NO by Pd. The possible reaction process on the monolithic-modified manganese catalyst was revealed by XPS, NH3-TPD, and H2-TPR. We proposed that a variety of active substances contained in the catalyst may have contributed to the enhanced deNO activity. As a consequence, the results achieved in the current study support the further exploration of such an approach toward the economical and rational design of the catalysts for efficient H2-SCR and are also beneficial to practical applications.