Research on adsorptive desulfurization of H 2 S from natural gas over Mn x O y -Co x O y supported on activated carbon

Abstract

Mn_xO_y, Co_xO_y, and Mn_xO_y-Co_xO_y supported on activated carbon (AC) adsorbents have been successfully prepared and characterized by specific surface area (BET), X-ray photoelectron spectroscopy, X-ray diffraction analysis, and Fourier-transform infrared spectroscopy. The results show that Mn is present in the form of Mn³⁺ and Mn⁴⁺, and Co is present in the form of Co²⁺ and Co³⁺. For all of the modified samples, the functional groups of the AC were well retained and the specific surface areas followed the order: AC > Mn_xO_y-Co_xO_y/AC > Co₃O₄/AC > Mn_xO_y/AC. For Mn_xO_y-Co_xO_y/AC, Mn and Co mainly exist in the forms of MnO₂, Co₃O₄, and Mn₃Co₂O₈. Mn_xO_y-Co_xO_y/AC was found to exhibit the best H₂S desulfurization performance. At 20℃, the breakthrough time and saturated adsorption time were 115 min and 185 min, and the saturated sulfur capacity and desulfurization rate were 347.7 mg/g and 74.2%. The H₂S removal capacities of the different adsorbents decreased in the order: Mn_xO_y-Co_xO_y/AC > Mn_xO_y/AC > Co_xO_y/AC > AC.

Keywords

activated carbon adsorptive desulfurization

Introduction

Natural gas holds an irreplaceable position among energy resources (Zou et al., 2013). The deep desulfurization of H₂S from natural gas is a very important gas clean-up process. Currently, wet and dry desulfurization methods are mainly applied to natural gas. Wet desulfurization is used for high concentrations of hydrogen sulfide. It has the advantage of large capacity, but the disadvantage of a complicated operation process. Dry desulfurization employing solid adsorbents is used for low concentrations of hydrogen sulfide. It has the advantages of less energy consumption and a small footprint. Activated carbon (AC), molecular sieves, and silica are the solid adsorbents usually used for dry desulfurization (Yu et al., 2009).

AC has a developed pore structure and a high specific surface area. However, adsorption on unmodified activated carbon is mainly by physical adsorption, which is poor in desulfurization. Recently, the modification of activated carbon has become a hot topic. As reported by Hui-bin Fang (Fang et al., 2013), metal oxides supported on activated carbon adsorbents showed much higher desulfurization performances for H₂S. The desulfurization performances of various adsorbents decreased in the order: Mn/AC > Cu/AC > Fe/AC > Ce/AC > Co/AC > V/AC. It has been shown (Li et al., 2011) that Co_xO_y/AC has electrocatalytic activity for oxygen reduction reaction under acidic conditions. An adsorbent composed of Mo-Co supported on activated carbon exhibited good carbon dioxide removal performance for flue gas (Wang et al., 2012).

In the present work, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC have been prepared, and their adsorptive desulfurization performances toward a simulated low H₂S concentration natural gas have been studied.

Experimental

Preparation of adsorbents

Activated carbon-supported manganese oxide (Mn_xO_y/AC) was synthesized by reduction of KMnO₄ using hydrogen peroxide with AC. The solid was dried to constant weight to obtain Mn_xO_y/AC precursor (PMA). The PMA was calcined at 300℃ for 2 h to obtain Mn_xO_y/AC adsorbent with an Mn:AC mass ratio of 1:1.

Activated carbon-supported cobalt oxide (Co_xO_y/AC) was synthesized by impregnating AC with cobalt nitrate solution (8 wt%). After filtration and drying in air, the solid was calcined at 300℃ for 2 h to obtain Co_xO_y/AC adsorbent with a Co:AC mass ratio of 0.8:1.

Activated carbon-supported manganese oxides and cobalt oxides (Mn_xO_y-Co_xO_y/AC) were synthesized by impregnating PMA with cobalt nitrate solution (8 wt%). After filtration and drying in air, the solid was calcined at 300℃ for 2 h to obtain Mn_xO_y-Co_xO_y/AC adsorbent with an Mn:AC:Co mass ratio of 1:1:0.8.

Characterization of adsorptions

XPS experiments

X-ray photoelectron spectra (XPS) were obtained by using an Κ-Alpha (Thermofisher Scientific) with Al Ka X-radiation (the resolution of the instrument is 1.15 eV).

X-ray diffraction experiments

X-ray diffraction (XRD) were obtained by using an D/max-2200PC (Rigaku Corporation) at 40 kV and 30 mA using monochromatized Cu-Ka radiation. The patterns were measured over the 2θ range from 10° to 80° at a scan step of 10°/min.

Fourier transform infrared spectroscopy Experiments

Fourier transform infrared spectroscopy (FTIR) were obtained using the KBr disk technique on a TENSOR 27 (Bruker) FTIR spectrophotometer (the resolution of the instrument is 4 cm⁻¹). Background and sample were scanned in a minute by using 24 DigiTect TM detector system and ROCKSOLIDTM interferometer.

BET experiments

BET surface areas were determined by nitrogen adsorption at 77 K by using a Micromeritics ASAP 2020 (USA) instrument. Prior to N₂ physisorption, the adsorptions were degassed at 120℃ for 4 h.

Adsorption experiments

Measurements of adsorbent activity were conducted in a fixed-bed reactor. The fixed-bed reactor, which consisted of a quartz glass tube of internal diameter 10 mm and length 300 mm, was positioned vertically in a clam shell furnace. A 0.2 g of the prepared adsorbent was loaded into the center region of the reactor. The up and down sides of the adsorbent were filled with quartz sand. The model fuel gas from a feed tank passed through a reducing valve and entered into the reactor at a flow rate of 20 mL min⁻¹. Any unadsorbed H₂S gas in the effluent gas would be absorbed by the liquid in the absorption bottle, and the H₂S content of the absorption liquid was determined by iodometry at intervals of 5 min. The breakthrough concentration was defined as that at which the H₂S concentration in the effluent gas reached 10% of the initial gas concentration, that is, 2 ppm. The H₂S capacity, saturated H₂S capacity, and H₂S removal rate were calculated according to the concentration of sulfur in the absorption liquid after the experiment.

The H₂S breakthrough capacity is defined as the H₂S capacity in the adsorbent when the H₂S concentration in the effluent gas reaches 10% of the initial concentration (2 mg/L). It can be calculated as follows:

q = \frac{c_{o} Qt - Q \int_{0}^{t} c dt}{m}

(1)

where q is the H₂S breakthrough capacity, mg·g⁻¹; c_o is the initial concentration of H₂S, mg/L; Q is the gas flow rate of the model fuel gas, mL·min⁻¹; c is the H₂S concentration in the effluent gas, mg/mL; t is the adsorption time, min; and m is the mass of adsorbent, g.

H₂S saturation capacity is defined as the amount of adsorbed H₂S per gram of adsorbent when the H₂S concentrations in the effluent and in the initial gas are the same. It can be calculated as follows:

q_{m} = \frac{V \times ω_{H_{2} S} - (c_{I_{2}} v_{I_{2}} - 0.5 c_{{Na}_{2} S_{2} O_{3}} v_{{Na}_{2} S_{2} O_{3}}) \times 34}{m}

(2)

where q_m is the H₂S saturation capacity, mg·g⁻¹;

ω_{H_{2} S}

is the H₂S concentration in the sample gas, mg/L; V is the total gas volume, L; and 34 is the molar mass of H₂S, g·mol⁻¹.

The H₂S removal rate η can be expressed as:

η = \frac{V \times ω_{H_{2} S} - (c_{I_{2}} v_{I_{2}} - 0.5 c_{{Na}_{2} S_{2} O_{3}} v_{{Na}_{2} S_{2} O_{3}}) \times 34}{V \times ω_{H_{2} S}} \times 100 %

(3)

Results and discussion

Adsorbent characterization

XPS analysis

The Mn_xO_y/AC and Mn_xO_y-Co_xO_y/AC adsorbents were studied by XPS and the Mn 2 p spectra are shown in Figure 1(a). It is well known (Di and Park, 2012; Li et al., 2009) that Mn³⁺ can be confirmed by an Mn 2p3/2 binding energy within the range 641.3–641.7 eV, and that Mn⁴⁺ can be identified by a binding energy of the XPS contribution in the range from 642.2 to 643 eV. The Mn 2p3/2 line for Mn_xO_y-Co_xO_y/AC was chemically shifted by 0.4 eV and 0.5 eV to lower binding energy compared to that for Mn_xO_y/AC. The Mn³⁺/Mn⁴⁺ mass ratio of Mn_xO_y/AC was 0.19/0.81. Compared with Mn_xO_y/AC, the Mn³⁺/Mn⁴⁺ mass ratio of Mn_xO_y-Co_xO_y/AC was increased to 0.29/0.71, showing an increase in the amount of Mn³⁺ by introducing the Co_xO_y. The Mn 2 p content in Mn_xO_y-Co_xO_y/AC was decreased from 26.4% to 13.4% by introducing Co_xO_y (Table 1). The Co 2p3/2 spectra of Co_xO_y/AC and Mn_xO_y-Co_xO_y/AC are shown in Figure 1(b). The XPS patterns of Co_xO_y/AC and Mn_xO_y-Co_xO_y/AC show two peaks and the shake-up satellite. For Co_xO_y/AC, the binding energy at 786.1 eV can be attributed to Co²⁺ species, whereas that at 781 eV can be attributed to Co³⁺ species (Cheng et al., 2004; Machida et al., 2000). For Mn_xO_y-Co_xO_y/AC, the Co 2p3/2 line was chemically shifted by 0.2 eV and 0.6 eV to lower binding energy compared to that for Co_xO_y/AC. The Co²⁺/Co³⁺ mass ratio of Co_xO_y/AC was 0.11/0.89. Compared with Co_xO_y/AC, the Co²⁺/Co³⁺ mass ratio of Mn_xO_y-Co_xO_y/AC was increased to 0.20/0.80, showing an increase in the amount of Co²⁺ by introducing Mn_xO_y. The Co 2p content was increased from 2.1% to 3.8% (Table 1).

Figure 1.

XPS spectrum of Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC absorbents.

Table 1.

Spectral parameters of AC, Co_xO_y/AC, Mn_xO_y/AC, and Mn_xO_y-Co_xO_y/AC by XPS analysis.

Adsorbent	Co_2p		Mn_2p		Percentage/%
	Binding energy (eV)		Binding energy (eV)		Percentage/%
	Co²⁺	Co³⁺	Mn³⁺	Mn⁴⁺	C_1s	N_1s	O_1s	Co_2p	Mn_2p
AC	–	–	–	–	84.7	3.7	11.6	–	–
Co_xO_y/AC	786.1	781.0	–	–	77.6	1.6	18.7	2.1	–
Mn_xO_y/AC	–	–	641.5	642.8	32.5	2.7	38.4	–	26.4
Mn_xO_y-Co_xO_y/AC	785.9	780.4	641.1	642.3	41.1	2.1	39.6	3.8	13.4

The electronic absorption abilities of the different species decrease in the order: Co³⁺> Mn⁴⁺> Mn³⁺> Co²⁺. For Mn_xO_y-Co_xO_y/AC, the shift in the peak positions of the Mn and Co species can be attributed to an increase in the electron cloud density, which is caused by sharing electron pairs through covalent bonds because of interaction between these four species.

XRD results

The XRD patterns of AC, Co_xO_y/AC, Mn_xO_y/AC, and Mn_xO_y-Co_xO_y/AC are shown in Figure 2. The diffuse peaks at 2θ = 20°–30° (002) and 2θ = 40°–50° (101) can be assigned to activated carbon, and indicate an irregular microcrystalline graphite structure therein. For MnO_x/AC, the peak at 2θ = 12.8° (110) can be attributed to birnessite MnO₂, and the peak at 2θ = 36.1° can be attributed to birnessite Mn₃O₄. In addition, the peaks at 2θ = 20°–30° (002) and 2θ = 40°–50° (101) attributable to the microcrystalline graphite structure of activated carbon were no longer seen. For Co_xO_y/AC, the peaks at 2θ = 36.9° (311), 56° (422), 60.1° (511), and 65.2° (440) can be attributed to birnessite Co₃O₄, and the peaks at 2θ = 27.9° and 44.7° can be attributed to birnessite carbon, indicating that the regularity of the activated carbon structure was greatly enhanced. For Mn_xO_y-Co_xO_y/AC, peaks arising from the surface of the microcrystalline graphite structure of the activated carbon were no longer seen, and the peaks at 2θ = 12.8° (101) 18° (200), 28.7° (301), and 49.9° (411) could be attributed to cubic phase MnO₂ (JCPDS Card No. 44-1386) (Tan et al., 1991). The peaks at 2θ = 37.4° (311), 41.9° (400), 56° (422), 60.1° (511), and 65.2° (440) could be assigned to cubic phase Co₃O₄ (JCPDS Card No. 42-1467) (Jiang et al., 2013; Li et al., 1999). The peak at 2θ = 37.5° could be attributed to birnessite Mn₃Co₂O₈, showing the interaction between Mn and Co. In conclusion, Mn and Co mainly exist in the forms of MnO₂, Co₃O₄, and Mn₃Co₂O₈ in Mn_xO_y-Co_xO_y/AC. These findings are consistent with the XPS results. Compared with Mn_xO_y/AC, the peak at 2θ = 12.8° for Mn_xO_y-Co_xO_y/AC becomes sharp, showing an increased degree of crystallinity of MnO₂.

Figure 2.

XRD patterns of AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC.

FTIR results

Figure 3 shows the FTIR spectra of the AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC adsorbents. For each of the samples, a peak at 3420 cm⁻¹ was observed, which could be attributed to the O–H stretching vibration of hydroxyl, carboxyl, and surface-adsorbed water. The peaks at 2926 and 1629 cm⁻¹ could be attributed to the symmetric and asymmetric stretching vibrations of C–H and C = O of the lactone base, respectively. The peaks at 1383 and 1039 cm⁻¹ could be attributed to asymmetric and symmetric COO⁻ stretching vibrations and skeleton C–O stretching vibrations of the activated carbon, respectively. In conclusion, the functional groups of AC remained unchanged for Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC. Compared with AC, a peak at 601 cm⁻¹ was detected for Mn_xO_y/AC, which could be attributed to the characteristic absorption of Mn-O (Wang et al., 2014). For Co_xO_y/AC, peaks at 581 and 681 cm⁻¹ could be attributed to the characteristic absorptions of Co–O (Bahaa et al., 2014). For Mn_xO_y-Co_xO_y/AC, a strong and sharp characteristic peak at 570 cm⁻¹ could be attributed to overlapping absorptions of Mn-O and Co-O, showing the strong interaction between Mn and Co. These findings were in accordance with the results obtained by XRD and XPS.

Figure 3.

FTIR spectra of AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC.

BET results

The textural parameters of AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC from BET are summarized in Table 2. The surface area and pore volume of AC were evaluated as 710.5 m²g⁻¹ and 0.36 mL g⁻¹, respectively. The surface areas and pore volumes of the Mn_xO_y- and Co_xO_y-modified adsorbents decreased significantly, showing channel plugging. However, Mn_xO_y-Co_xO_y/AC showed the highest surface area and pore volume among the modified samples. The surface areas of the various adsorbents decreased in the order: AC (710.5 m²g⁻¹) > Mn_xO_y-Co_xO_y/AC (603.6 m²g⁻¹) > Co_xO_y/AC (564.3 m²g⁻¹) > Mn_xO_y/AC (430.4 m²g⁻¹).

Table 2.

Pore properties of AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC.

Adsorbent	BET surface area	Pore volume	Average pore
Adsorbent	A/(m²·g⁻¹)	v/(mL·g⁻¹)	diameter d/nm
AC	710.5	0.36	3.1
Co_xO_y/AC	564.3	0.18	4.1
Mn_xO_y/AC	430.4	0.15	4.8
Mn_xO_y-Co_xO_y/AC	603.6	0.41	3.8

Desulfurization performances of the adsorbents

All of the adsorbents were evaluated at room temperature (298 K) with a gas velocity of 20 mL/min using a fixed-bed plug flow reactor. The results are shown in Figure 4 and Table 3. For AC, the desulfurization performance was poor, with a breakthrough time of 20 min and a saturated adsorption time of 70 min. All of the modified adsorbents showed better desulfurization performances compared to that of AC. This was presumably because H₂S is mainly physisorbed on AC, whereas it is both physisorbed and chemisorbed on the modified adsorbents. Compared to those of AC, the breakthrough time (115 min) and saturated adsorption time (185 min) for Mn_xO_y-Co_xO_y/AC were increased five-fold and 2.5-fold, respectively.

Figure 4.

Breakthrough curves of AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC absorbents.

Table 3.

Desulfurization performance of AC, Mn_xO_y/AC, Co_xO_y/AC, and Mn_xO_y-Co_xO_y/AC.

Adsorbent	Breakthrough time (min)	The saturated adsorption time (min)	Saturated sulfur capacity (mg/g)	The desulfurization rate (%)
AC	20	70	37.7	25.1
Co_xO_y/AC	41	121	169.8	52.8
Mn_xO_y/AC	74	149	289.3	70.2
Mn_xO_y-Co_xO_y/AC	115	185	347.7	74.2

Table 3 shows the saturated sulfur capacities and desulfurization rates for the samples. For AC, the saturated sulfur capacity (37.7 mg/g) and desulfurization rate (25.1%) were poor. Compared to those of AC, the saturated sulfur capacity (347.7 mg/g) and desulfurization rate (74.2%) for Mn_xO_y-Co_xO_y/AC were increased nine-fold and three-fold, respectively. The highest desulfurization performance of Mn_xO_y-Co_xO_y may be attributed to the higher surface area (BET result) and greater degrees of crystallinity of MnO₂ and Co₃O₄ (XRD results). The desulfurization performances of the various adsorbents decreased in the order: Mn_xO_y-Co_xO_y/AC > Mn_xO_y/AC > Co_xO_y/AC > AC.

Conclusions

Adsorbents consisting of Mn_xO_y, Co_xO_y and Mn_xO_y-Co_xO_y supported on activated carbon have been successfully prepared. The results showed that Mn was present in the forms of Mn³⁺ and Mn⁴⁺, and that Co was present in the forms of Co²⁺ and Co³⁺. The functional groups of activated carbon were well retained in all of the modified samples. The specific surface area of Mn_xO_y-Co_xO_y/AC was evaluated as 603.6 m²/g. Compared with the original activated carbon, the specific surface areas of all of the modified samples were decreased, diminishing in the order: AC > Mn_xO_y-Co_xO_y/AC > Co₃O₄/AC > Mn_xO_y/AC. For Mn_xO_y-Co_xO_y/AC, the contents of Mn and Co were 13.6% and 3.9%, respectively, and the mass ratios of Co²⁺/Co³⁺ and Mn³⁺/Mn⁴⁺ were increased compared with the monometal-modified AC samples, possibly because of interaction between the active components. The crystal regularities of Mn_xO_y and Co_xO_y in Mn_xO_y-Co_xO_y/AC were enhanced. The results of fixed-bed adsorption desulfurization experiments indicated that Mn_xO_y-Co_xO_y/AC showed the best H₂S desulfurization performance. The H₂S removal capacities of the various adsorbents decreased in the order: Mn_xO_y-Co_xO_y/AC > Mn_xO_y/AC > Co_xO_y/AC > AC.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Bahaa

MAZ

Soliman

Sarah

(2014) Pure and Ni-substituted Co₃O₄ spinel catalysts for direct N₂O decomposition. Chinese Journal of Catalysis 35(7): 1105–1112.

Cheng

Yang

Yan

(2004) Preparation of Co₃O₄ nanosheet supported on Ni foam and its catalytic performance for H₂O₂ electroreduction. Chemical Journal of Chinese Universities 35(1): 110–114.

Park

(2012) Influence of nickel oxide on carbon dioxide adsorption behaviors of activated carbons. Fuel 102: 439–444.

Fang

Zhao

Fang

(2013) Selective oxidation of hydrogen sulfide to sulfur over activated carbon-supported metal oxides. Fuel 108: 143–148.

Jiang

Zhu

(2013) Au/γ-MnO₂ catalyst for solvent-free toluene oxidation with oxygen. Chinese Journal of Catalysis 34(9): 1683–1689.

Zhu

Qiu

(2011) Electro catalytic activity of Co₃O₄/C for oxygen reduction and the reaction mechanism. Chinese Journal of Catalysis 32(4): 624–629.

Lin

Zhang

(2009) Low-temperature selective catalytic reduction of NO with NH₃ over MnO_x-CeO₂/ZrO₂-TiO₂ monolith catalyst. Chinese Journal of Catalysis 30(2): 104–110.

(1999) Fabrication of Co₃O₄ ultrafines by a liquid-control-precipitation method. Chinese Journal of Catalysis 20(4): 519–522.

Machida

Uto

Kurogi

(2000) MnO_x−CeO₂ binary oxides for catalytic NO_x orption at low temperatures. Sorptive removal of NO_x. Chemistry of Materials 12(10): 3158–3164.

10.

Tan

Kenneth

Peter

MAS

(1991) XPS studies of solvated metal atom dispersed catalysts. Evidence for layered cobalt-manganese particles on alumina and silica. Journal of the American Chemical Society 113: 855–861.

11.

Wang

Bing

Guo

(2012) Preparation of activated carbon-supported Co-Mo bimetallic catalyst by impregnation-precipitation method. Journal of Fuel Chemistry and Technology 40(10): 1252–1257.

12.

Wang

Zhang

(2014) The effects of Mn loading on the structure and ozone decomposition activity of MnO_x supported on activated carbon. Chinese Journal of Catalysis 35(3): 335–341.

13.

Fang

(2009) Selectivity of oxidative desulfurization of gasoline and diesel fuel using catalytic oxidation method with activated carbon. CIESC Journal 14(160): 1007–1016.

14.

Zou

Zhao

(2013) Oxidation/adsorption desulfurization of natural gas by bridged cyclodextrins dimer encapsulating polyoxometalate. Fuel 104: 635–640.