ARTICLE pubs.acs.org/EF
Use of High-Pressure Impregnation in Preparing Zn-Based Sorbents for Deep Desulfurization of Hot Coal Gas Xianrong Zheng,*,†,‡ Weiren Bao,† Qingmai Jin,† Liping Chang,*,† and Kechang Xie† †
Key Laboratory of Coal Science and Technology, Taiyuan University of Technology, Ministry of Education and Shanxi Province, Taiyuan 030024, People’s Republic of China ‡ College of Electrical and Power Engineering, Taiyuan University of Technology, Taiyuan 030024, People’s Republic of China ABSTRACT: A new preparation method of zinc oxide (ZnO) sorbents for removing H2S from hot coal gas was explored. Highpressure impregnation with semi-coke as the support and zinc nitrate as the active component precursor was used to prepare Znbased sorbents (HZnSC). The desulfurization activity of the sorbents was studied in a fixed-bed reactor under the simulated coal gas mixture of CO (33 vol%), H2 (39 vol %), H2O (5 vol %), H2S (300 ppmv), and N2 (balance) at 300550 °C. The results show that the desulfurization activity of the HZnSC sorbents is higher than that of the sorbents prepared using atmospheric pressure impregnation. The surface area of the semi-coke expanded from 16.65 to 272.59 m2/g. Nanoscale ZnO was acquired, and the active component was uniformly dispersed on the support with the high-pressure impregnation method, which afforded H2S removal from hot coal gas. HZn20SC (prepared with 20 wt % zinc nitrate solution) reduced H2S from 300 to less than 0.1 ppmv at 300550 °C, and its sulfur capacity was 4.5 g of S/100 g of sorbent at 500 °C. It also maintained good heat stability during three desulfurization/ regeneration cycles. The findings suggest that semi-coke with a low surface area can be used as a support of sorbents to remove H2S from hot coal gas using high-pressure impregnation.
’ INTRODUCTION The polygeneration system of syngas production from the combination of gasification gas and oven gas (dual-gas resources) for alcohol ether synthesis1,2 is one of the most prospective coalbased generation technologies. CO2 emission from coal gas can be reduced by the reforming reaction of CO2 from gasification gas and CH4 from oven gas to produce CO and H2 in this system. However, S-containing gases (H2S, >90%), which may influence the effective operation of this technique by polluting the environment, corroding equipment, and poisoning catalysts, exist in dual-gas resources and syngas. H2S in these gases must be removed. Desulfurization technology under the middle temperature range of 300550 °C is primarily used to avoid heat loss and save energy, and it is being widely developed to remove H2S from hot coal gas. For this purpose, the studies on single metal oxide and mixed metal oxides as sorbents, such as zinc-, cerium-, copper-, and manganese-based sorbents, have been performed. Among these metal oxides, zinc oxide (ZnO)37 has favorable thermodynamics for H2S removal and is considered to be a major desulfurizer with high precision, stability, and reliable performance. H2S in hot coal gas must be reduced to 0.0510 ppmv, and the sorbent adsorbing H2S must have high precision; therefore, ZnO is chosen as the active component. In sorbent preparation, ZnO is typically loaded onto supports with large surface areas for dispersing the active components and increasing the surface area of the sorbents. ZnO sorbents have also been prepared with activated carbon as the support using the matrix-assisted method, and nanosized ZnO sorbents with excellent performance for H2S removal have been acquired.8,9 Activated carbon is used as a sorbent support because of its large surface area and rich pore structure. Semi-coke is a porous substance similarly structured to activated carbon (both are made of carbon materials). In comparison to activated carbon, semicoke is cheaper and more easily obtained. However, it cannot be r 2011 American Chemical Society
directly used as a sorbent support and requires further processing because its surface area and pore volume are smaller. Highpressure hydrothermal synthesis is an effective method for improving the pore structure and surface area of semi-coke.10 Preparing zinc-based sorbents by high-pressure impregnation is thus feasible and may even improve the pore structure of the semicoke support as well as simultaneously impregnate the active component on the support. In the current study, the desulfurization performance of zinc-based sorbents prepared by highpressure impregnation was explored.
’ EXPERIMENTAL SECTION Preparation of Zn-Based Sorbents. Zn-based sorbents were prepared using the high-pressure impregnation method, with the semicoke particle as the support and zinc nitrate as the active component precursor. The semi-coke had a low surface area of 16.65 m2/g, and the concentration of zinc nitrate solution was approximately 524 wt %. The semi-coke particle and zinc nitrate solution were simultaneously placed into an autoclave by high-pressure impregnation at 28 atm for 5 h. The sample was then filtered after being cooled to room temperature, dried at 50 °C for 5 h, dried at 100 °C for another 5 h, and finally calcined at 500 °C in pure N2 for 5 h based on the optimizing preparation conditions. The particle size of sorbents is 2.383.36 mm (68 mesh). The flow diagram of the sorbent preparation is shown in Figure 1. The sorbents for comparison were also prepared using the atmospheric pressure impregnation method. The sorbents obtained in the experiments were marked as HZnxSC and AZnxSC for the high-pressure and atmospheric pressure impregnation methods, respectively, where “x” indicates a Zn(NO3)2 solution concentration of 5, 10, 20, or 24 wt %. Received: April 13, 2011 Revised: May 27, 2011 Published: May 31, 2011 2997
dx.doi.org/10.1021/ef200565p | Energy Fuels 2011, 25, 2997–3001
Energy & Fuels
ARTICLE
Figure 1. Flow diagram of the sorbent preparation. In addition, H means high-pressure impregnation, and A means atmospheric pressure impregnation. A 150 mL autoclave (Shanghai) was used for high-pressure impregnation. Its design temperature and pressure were 250 °C and 100 atm, respectively. Desulfurization Tests of Zn-Based Sorbents. Desulfurization tests were carried out in a fixed-bed reactor, and a vertical quartz reactor (2.0 cm inner diameter) was placed in an external heating furnace. Approximately 10 g (20 mL) of sorbent was placed in the middle of the reactor with the quartz wool packed and the bed temperature at 300500 °C. The simulated coal gas was composed of CO (33 vol %), H2 (39 vol %), H2O (5 vol %), H2S (300 ppmv), and N2 (balance), with the total flow rate being 667 mL/min. Gas compositions were analyzed with a gas chromatograph (GC-950) equipped with a flame photometry detector (FPD) and a thermal conductivity detector. The testing precision for H2S of FPD is 0.001 ppmv in our GC analyzer. The schematic diagram of the experimental system is shown in Figure 2. Sorbent regeneration was carried out in two stages: First, the sorbent after desulfurization was regenerated at 560 °C with 2% O2 as well as 98% N2 and with the space velocity of 800 h1. When the outlet SO2 could not be detected, the sorbent was purged in pure N2 at 700 °C with the space velocity of 2000 h1. H2S breakthrough time and sulfur capacity are the evaluation standards of sorbent desulfurization ability. The H2S breakthrough time is defined as the time from the beginning of the desulfurization experiment to the point at which the outlet H2S concentration reaches 0.1 ppmv. The sulfur capacity in the breakthrough time is the mass of sulfur adsorbed by 100 g of sorbent. When the inlet H2S concentration and the total flow rate remain unchanged, the longer breakthrough time and the higher sulfur capacity indicate better desulfurization ability. The computational formulas for the efficiency of H2S removal, the sulfur capacity, and the utilization rate of the sorbent used in this study are as follows: the efficiency of removing H2 S, η ð%Þ :
η ð%Þ ¼
C0 C 100 C0
the sulfur capacity, S ðg of S=100 g of sorbentÞ : 32vsp Vbj C0 100 S ¼ Σηi ti 22:4G the utilization rate, Y ð%Þ : ¼
Y ð%Þ actual sulfur capacity 100 theoretical sulfur capacity
where νsp is the space velocity of gas (h1), Vbj is the volume of sorbent in the reactor (L), ti is the adsorption time in the i time sampling before breakthrough (h), G is the weight of sorbent in the reactor (g), C0 is the inlet H2S concentration (ppmv), and C is the outlet H2S concentration in the i time sampling before breakthrough (ppmv). Characterization of Zn-Based Sorbents. The surface area and pore volume of sorbents were measured with ASAP 2000 Micromeritics
Figure 2. Schematic diagram of the experimental system: 1, cylinder; 2, valve; 3, gas flow meter; 4, mixing chamber; 5, water saturation; 6, furnace; 7, sorbent; 8, thermocouples; 9, temperature controller; 10, quartz reactor; 11, gas chromatography; 12, readout; and 13, vent. (Micromeritics Instrument Corp.) using nitrogen adsorption at 196 °C. Crystalline structures of Zn-based sorbents before and after the reaction were determined using X-ray powder diffraction (XRD 6000, Daoding) with Cu KR radiation. The applied current and voltage were 30 mA and 40 kV, respectively. The diffraction patterns were recorded from 15° to 75° using a scan rate of 8° min1. The morphology of the sorbents was observed under a JSM-6700F scanning electron microscope (JEOL, Japan). The uploading amount of the components on the support was measured using an atomic absorption spectrum instrument (TAS-990 Super, Beijing, China).
’ RESULTS AND DISCUSSION Desulfurization tests were performed in a fixed-bed reactor at 500 °C and at atmospheric pressure to investigate the performance of the HZnSC and AZnSC series sorbents in removing H2S. The desulfurization performance curves of the sorbents are illustrated in Figure 3. A good sorbent is characterized by a high sulfur capacity and a long breakthrough time; the values obtained for the sorbents tested in this study are shown in Table 1. The HZnSC and AZnSC sorbents significantly differed in their desulfurization performance, with the former sorbents having higher precision before breakthrough compared to the latter sorbents. When the outlet H2S reached 0.1 ppmv and the removal rate was 99.97%, the sorbent HZn20SC demonstrated the best desulfurization ability; its breakthrough time was as long as 23 h, and its sulfur capacity was 4.46 g of S/100 g of sorbent. On the other hand, the sulfur capacity of the AZn5SC sorbent was only 0.68 g of S/100 g of sorbent. These results clearly show that the sorbent preparation method plays a critical role in the desulfurization performance of sorbents. The sorbents prepared using different methods but with the same materials had distinct desulfurization abilities. Because the different preparation methods could have led to different kinds of active components, XRD was performed on fresh HZn20SC and AZn5SC sorbents. The results are shown in Figure 4. ZnO and SiO2 diffraction peaks were observed for the two fresh sorbents (Figure 4). SiO2 occurred in the raw semicoke support, whereas ZnO was an active component of HZn20SC and AZn5SC. New phase ZnS diffraction peaks were observed in the sorbent after desulfurization, confirming the conversion of the ZnO sorbent into ZnS as a result of its reaction with H2S in hot coal gas. The desulfurization reaction is ZnO þ H2S = ZnS þ H2O. 2998
dx.doi.org/10.1021/ef200565p |Energy Fuels 2011, 25, 2997–3001
Energy & Fuels
ARTICLE
Figure 3. Desulfurization performance curves of the HZnSC and AZnSC series sorbents.
Table 1. Breakthrough Time and Actual Sulfur Capacity of Sorbents
Table 2. Loading Amount and Utilization Ratio of Sorbents theoretical sulfur utilization ratio of capacity (g of S/100 g active component
breakthrough
actual sulfur capacity
sorbent
time (h)
(g of S/100 g of sorbent)
sorbent
(mol)
of sorbent)
(%)
AZn5SC
4.0
0.68
AZn5SC
0.013
2.40
28
AZn10SC
2.0
0.22
AZn10SC
0.011
2.32
10
AZn20SC
2.0
0.28
AZn20SC
0.011
2.31
12
AZn24SC
2.0
0.34
AZn24SC
0.010
2.14
17
HZn5SC
4.5
0.78
HZn10SC
7.0
1.21
HZn5SC HZn10SC
0.013 0.023
2.40 5.23
32 25
HZn20SC
23.0
4.46
HZn20SC
0.056
11.72
38
HZn24SC
12.0
2.13
HZn24SC
0.032
7.23
29
Figure 4. XRD patterns of HZn20SC and AZn5SC sorbents.
The utilization rate and loading amount of the active component were calculated to determine the factors affecting the different desulfurization performances of the HZnSC and AZnSC series sorbents. The results are shown in Table 2. The uploading amount of the components on the support was acquired by subtraction. The zinc nitrate content of the residual impregnation liquid was measured using an atomic absorption instrument. The actual uploading amount was derived from the original zinc nitrate amount minus the surplus amount, whereas the utilization rate of the active component was calculated according to the desulfurization test results. Assuming that
loading amount
uploading Zn(NO3)2 on the support is entirely decomposed into the active component ZnO and all ZnO sorbents fully react with H2S, the mass of sulfur adsorbed by 100 g of sorbent is defined as the theoretical sulfur capacity of the sorbent. The uploading amount hence refers to the theoretical sulfur capacity. The results show that the loading amount greatly changed with an increasing concentration of zinc nitrate solution for HZnSC sorbents. Previous research has shown that the pore structure of the raw semi-coke can be effectively improved by the high-pressure hydrothermal reaction.10 Tar was detected in the crack between microcrystallines of semi-coke, which results from low-temperature carbonization. Liquid pressure is very high in the autoclave,11 giving the liquid great penetrating ability and dissolving capacity under high temperature and pressure. The hot solution can open some blocked pores of the raw semi-coke, improving its pore structure. In the current study, the uploading amount of the HZnSC sorbents was greater than that of the AZnSC sorbents. According to the substance transmission principle, the increase in the concentration of the impregnation solution means that more pores are blocked and the porosity is reduced, which will certainly increase mass transmission resistance, in turn leading to an optimal impregnation concentration. The results of the present study show that the optimal concentrations were 20 and 5% for the HZnSC and AZnSC sorbents, respectively. The desulfurization capacity was not proportional to the impregnating solution concentration (Figure 3). When porous semicoke is impregnated in zinc nitrate solution, the impregnation components can be uniformly dispersed in the pore and surface of the semi-coke support at low concentrations. However, with an increasing concentration, the porosity rapidly 2999
dx.doi.org/10.1021/ef200565p |Energy Fuels 2011, 25, 2997–3001
Energy & Fuels
ARTICLE
Table 3. Structural Property of Sorbents Determined by N2 Adsorption sample
BET surface area (m2/g)
pore volume (cm3/g)
AZn5SC
32.18
0.02
AZn10SC AZn20SC
46.72 145.85
0.04 0.08
AZn24SC
190.12
0.10
HZn5SC
73.09
0.04
HZn10SC
117.01
0.09
HZn20SC
272.59
0.17
HZn24SC
152.71
0.08
declines and impregnating components are accumulated on the support surface. The pore size of sorbents has also been reported to be directly influenced by its component particle sizes and the gathering pattern of particles. The pores become narrow when the primary particles gather, and the pores become coarse when subprime particles gather.12 These findings indicate that the active components cannot be used effectively. In the current study, the maximum utilization ratio of the active component was 38% for the HZnSC sorbents, higher than that of the AZnSC sorbents (28%). The desulfurization activity of the HZnSC sorbents was significantly higher than that of the AZnSC sorbents, and HZn20SC removed H2S from hot coal gas most efficiently. The analysis of scanning electron microscopy (SEM) and the textural properties were carried out for different sorbents to identify the essential difference between the high-pressure impregnation and atmospheric pressure impregnation methods. The textural properties of the sorbents were detected by N2 adsorption experiments, and the characterization results are shown in Table 3. The sorbents prepared using high-pressure impregnation had greater surface area and pore volume compared to the sorbents prepared using atmospheric pressure impregnation. This result shows that the high-pressure impregnation method can increase the pore area of semi-coke. Furthermore, the high-pressure impregnating solution can dissolve the remaining tar, thereby enlarging the pore. The surface area and pore volume of sorbents under different zinc nitrate solution concentrations varied, with HZn20SC having the largest surface area (272.59 m2/g) and pore volume (0.17 cm3/g). As shown in Table 3, the high-pressure impregnation method can significantly improve the pore structure of the semi-coke support. Therefore, the active component of HZnSC can be evenly dispersed in the semi-coke support. Together with the sulfur capacity and desulfurization performance curve, the pore volume and surface area directly affect the loading amount and utilization ratio of active components. As a result, these properties of sorbents directly influence their ability to remove H2S from hot coal gas. In addition, the high-pressure impregnation method effectively disperses active components on the support and reduces its particle size. The SEM results for HZnSC and AZnSC series sorbents are illustrated in Figure 5. The ZnO active component of HZnSC was evenly dispersed in the semi-coke support, and the sizes of ZnO particles are presented in the range of 10300 nm. ZnO of AZnSC was unevenly heaped, and its particle size was larger. The dispersion degree and particle size of the active component are the principal factors affecting the desulfurization activity of
Figure 5. SEM images of sorbents (a) HZn5SC, (b) AZn5SC, (c) HZn20SC, (d) AZn20SC, (e) HZn24SC, and (f) AZn24SC.
sorbents. In the desulfurization process, H2S molecules diffuse to the surface of ZnO sorbents and ZnS molecules are immediately generated during the desulfurization reaction. The surface reaction, however, is far from sufficient, and it is more important for H2S molecules to enter the internal surface of sorbent particles.13 When the ZnO particles of 200500 nm are reduced to 2471 nm, their dynamic reaction capacity can improve nearly 10 times.14 Nanometer particle size ZnO of HZnSC series sorbents possesses this advantage compared to the AZnSC series sorbents. The sorbents prepared using the high-pressure impregnation method can efficiently remove H2S from hot coal gas because of their large surface area and high desulfurization activity. All of these results show that it is feasible to use semicoke as the support and ZnO as the active component of sorbents for hot coal gas desulfurization by high-pressure impregnation. HZn20SC was selected for further research on the effects of the desulfurization temperature. The desulfurization experiment was carried out at temperatures of 300, 350, 400, 450, 500, and 550 °C, and the results are shown in Figure 6 and Table 4. The findings indicate that HZn20SC has similar breakthrough behaviors at different desulfurization temperatures. The desulfurization performance of the sorbent was better at 450 and 500 °C, with longer breakthrough time and higher sulfur capacity. When the desulfurization temperature was increased to 550 °C, the desulfurization performance significantly dropped. The breakthrough time was shorter, and the sulfur capacity was lower. Meanwhile, the elemental zinc component was found at the outlet of the reactor at this temperature. The ZnO active 3000
dx.doi.org/10.1021/ef200565p |Energy Fuels 2011, 25, 2997–3001
Energy & Fuels
ARTICLE
Figure 6. Desulfurization performance curves of HZn20SC at different desulfurization temperatures.
Table 4. Breakthrough Time and Sulfur Capacity of HZn20SC at Different Desulfurization Temperatures temperature
breakthrough
actual sulfur capacity
(°C)
time (h)
(g of S/100 g of sorbent)
300
21.5
4.17
350
21.5
4.17
400
21.5
4.17
450
23.0
4.46
500 550
23.0 20.5
4.46 3.98
’ CONCLUSION The high-pressure impregnation method is characterized by impregnation of the active component precursor on the support while simultaneously improving its pore structure. In this study, the desulfurization activity of the HZnSC series sorbents was better than that of the AZnSC samples. Semi-coke with a low surface area expanded from 16.65 to 272.59 m2/g using highpressure impregnation, supporting its potential use as a support of sorbents. A ZnO active component with a nanometer particle size formed and was uniformly dispersed on the semi-coke support under the proposed method. HZn20SC exhibited the best desulfurization ability, reduced the concentration of H2S from hot coal gas from approximately 300 to less than 0.1 ppmv at 300550 °C, with the maximum sulfur capacity of 4.46 g of S/ 100 g of sorbent, and had good heat stability during the three desulfurization/regeneration cycles at 500 °C. These findings indicate that high-pressure impregnation is a suitable preparation method of sorbents for removing H2S from hot coal gas. ’ AUTHOR INFORMATION Corresponding Authors
*Telephone: þ86-351-6010482. Fax: þ86-351-6010482. E-mail:
[email protected] (X.Z.);
[email protected] (L.C.).
’ ACKNOWLEDGMENT The authors gratefully acknowledge the financial support of the National Basic Research Program of China (2005CB221203), National Natural Science Foundation of China (20976117), Shanxi Province Natural Science Foundation (2010011014-3), and Shanxi Province Basic Conditions Platform for Science and Technology Project (2010091015). ’ REFERENCES
Figure 7. Desulfurization performance curves of HZn20SC in three desulfurization/regeneration cycles.
component was reduced to elemental zinc in a reductive atmosphere, causing the loss of active components. The results of the three desulfurization/regeneration cycle experiments for HZn20SC at 500 °C are shown in Figure 7. The HZn20SC sorbent had a breakthrough time of 17 h and a sulfur capacity of 4.03 g of S/100 g of sorbent after the first regeneration; the values of sorbent after the second and third regenerations were similar. Although its efficiency in removing H2S was slightly worse than that of the fresh sorbent, HZn20SC maintained good heat stability during the three desulfurization/ regeneration cycles at the desulfurization temperature of 500 °C. The results confirm that HZn20SC is an excellent sorbent for removing H2S from hot coal gas.
(1) Xie, K. C. Proceedings of the 16th International Symposium on Alcohol Fuels (ISAF XVI); Rio de Janeiro (RJ), Brazil, 2006; plenary lecture. (2) Xie, K. C. Proceedings of International High-technology Symposium on Coal Chemical Industry and Coal Conversion; Shanghai, China, 2004; pp 107112. (3) Wang, X. H.; Jia, J. P.; Zhao, L.; Sun, T. H. Appl. Surf. Sci. 2008, 254, 5445–5451. (4) Wang, X. H.; Sun, T. H.; Yang, J.; Zhao, L.; Jia, J. P. Chem. Eng. J. 2008, 142, 48–55. (5) Zeng, Y.; Zhang, S.; Groves, F. R.; Harrison, D. P. Chem. Eng. Sci. 1999, 54, 3007–3017. (6) Garcia, E.; Palacds, J. M.; Alonso, L. Energy Fuels 2000, 14, 1296–1303. (7) Alonso, L.; Palacds, J. M. Energy Fuels 2002, 16, 1550–1556. (8) You, J. L.; Park, N. K.; Gi, B. H.; Si, O. R.; Tae, J. L. Curr. Appl. Phys. 2008, 8, 746–751. (9) Park, N. K.; You, J. L.; Gi, B. H.; Si, O. R.; Tae, J. L. Colloids Surf. 2008, 313314, 66–71. (10) Zheng, X. R. Master’s Thesis, Taiyuan University of Technology, Taiyuan, Shanxi, China, 2003. (11) Fu, S. L.; Cai, S. Z.; Zhang, X. J.; Wang, W. W.; Wei, Z. R. J. Synth. Cryst. 2006, 5, 1016–1021. (12) Wang, G. R. Textbook of Catalyst and Catalysis; Dalian University of Technology Press: Dalian, China, 2007; pp 910. (13) Feng, X. J. Chem. Ind. Eng. 2008, 29, 31–35. (14) Novochinskii, I. I.; Song, C. S.; Ma, X. L.; Liu, X. S.; Shore, L.; Lampert, J.; Farrauto, R. J. Energy Fuels 2004, 18, 576–583.
3001
dx.doi.org/10.1021/ef200565p |Energy Fuels 2011, 25, 2997–3001