Study of Short-Term Catalyst Deactivation Due to Carbon Deposition

Nov 24, 2015 - Vivek Pawar , Srinivas Appari , Dayadeep S. Monder , and Vinod M. Janardhanan. Industrial & Engineering Chemistry Research 2017 56 (30)...
1 downloads 0 Views 1MB Size
Article pubs.acs.org/EF

Study of Short-Term Catalyst Deactivation Due to Carbon Deposition during Biogas Dry Reforming on Supported Ni Catalyst Vivek Pawar,† Debjyoti Ray,‡ Ch. Subrahmanyam,‡ and Vinod M. Janardhanan*,† †

Department of Chemical Engineering and ‡Department of Chemistry, Indian Institute of Technology, Hyderabad, Telangana 502 285, India ABSTRACT: Dry reforming experiments are performed for five different CH4/CO2 ratios at three different operating temperatures. The focus of the work is on the short-term catalyst deactivation due to carbon deposition and analysis of the nature and structure of deposited carbon at CH4/CO2 ratio of 2, which is typical for biogas. The dry reforming experiments indicate that the short-term deactivation is insignificant at an equimolar ratio of CH4 and CO2. The rate of carbon deposition becomes faster at higher CH4/CO2 ratios and higher temperatures. CH4 and CO2 conversions and the exit mole fraction of CO and H2 confirm the occurrence of the reverse water−gas shift reaction. Characterization of the spent catalyst and deposited carbon is performed using XRD, TGA, FTIR, and Raman spectroscopies, which indicates the presence of graphitic carbon and carbon nanostructures.



CO disproportionation reaction.6,7 Thermodynamically CO disproportionation is favored at temperatures below 710 °C.1 According to the reaction stoichiometry CH4 + CO2 ⇌ 2H2 + 2CO, an equimolar mixture of CH4 and CO2 results in the same conversion for both reactants and the same ratio for the products. Although the global representation of dry reforming leads to the above conclusions, in reality the reaction proceeds through a number of steps involving adsorption, disproportionation, recombination, and disorption. These microscopic elementary kinetics can lead to the occurrence of other reactions such as reverse water−gas shift (RWGS), which leads to a gas composition different from the stoichiometrically predicted one to which the catalyst is exposed.6 The RWGS reaction is also helpful in the gasification of any coke that may be formed during the dry reforming process.6 Several researchers hinted on the occurrence of RWGS at temperatures higher than 700 °C, particularly because of higher CO2 conversions than CH4 and lower H2/CO ratio in the product mixture.6,8 The occurrence of RWGS also leads to H2/CO ratios less than unity.6,9 Processing of biogas requires one to deal with different CH4/CO2 ratios that are most often higher than 1.0 if additional CO2 is not supplied to maintain the stoichiometry at 1.0. In addition to the difficulties associated with coke formation due to the high CH4/CO2 ratio, biogas may also contain H2S, which leads to poisoning of catalysts, particularly those based on Ni.10,11 Although the use of noble metal catalysts suppresses coke formation,12−14 Ni is generally preferred for reforming reactions to noble metals. Since coke formation affects the catalyst longevity, activity, and selectivity, several researchers have focused on identifying ingenious catalysts and reactors that suppress coke formation.15 For instance, Ce has been identified as a good support for Ni to resist coke formation. The oxygen storage capacity of Ce mitigates coke deposition on

INTRODUCTION Today most of the hydrogen is produced from reforming of natural gas. As petroleum feed stocks are depleting daily it is imperative to find other sources for the production of H2. Electrolysis of water to produce H2 is a highly energy-intensive process, and hydrogen economy based on water electrolysis at the moment looks like a distant future. Biogas containing 50− 75% CH4, 50−25% CO2, 0−10% N2, and 0−3% H2S is another versatile and renewable source for the production of H2 and synthesis gas. Biogas can be converted to synthesis gas by steam reforming, dry reforming, a combination of both, or partial oxidation. However, compared to the other routes dry reforming leads to syngas with higher CO content, which is more appropriate for production of liquid hydrocarbons using the Fischer−Tropsch process.1 One of the other major attractions for biogas dry reforming is that it is a route for CO2 recycling. Although biogas dry reforming is attractive there are two major bottlenecks for practicing it industrially: (i) the reaction is highly endothermic and therefore requires significant energy input; (ii) catalyst deactivation due to a high CH4/CO2 ratio, which leads to deposition of inactive carbon.2 Although dry reforming is more endothermic than steam reforming, the cost may be offset if the energy required for production of steam is taken into consideration while doing an economic analysis. The stability of operation and carbon deposition depends largely on the CH4/CO2 ratio, operating temperature, catalyst employed, catalyst formulation, catalyst support interaction, and type of reactor employed. Thermodynamic analysis shows that carbon deposition increases with decreasing temperature and increasing CH4/CO2 ratio, and maximum H2 production occurs at CH4/CO2 a ratio of 1.0. Lowering the ratio can lead to a further drop in carbon deposition, however, at the expense of lower process efficiency. Therefore, from an industrial stand point it is desirable to work near a CH4/CO2 ratio of unity. Stable operation without deactivation has been demonstrated for CH4/CO2 = 1 at temperatures higher than 750 °C.3−5 While high temperature and low pressure favors CH4 decomposition, low temperature and high pressure operation favors carbon deposition by the © XXXX American Chemical Society

Received: August 17, 2015 Revised: November 11, 2015

A

DOI: 10.1021/acs.energyfuels.5b01862 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

°C for 12 h and further calcined in air at 800 °C for 6 h. The pellets were approximately 5 mm in length and 3 mm in diameter; however, they are further broken into smaller pieces (6−8) before loading into the reactor. Smaller pieces were used in order to limit any diffusion resistance that may be preset during the reaction. The catalysts were then loaded into a quartz tube reactor having a 18 mm inner diameter. In order to maintain uniform temperature over the bed of the catalyst, the catalyst particles were mixed with quartz beads. The reactor was placed in a three-zone heating furnace (Applied Test Systems Inc., USA) and heated to the desired temperature in N2 flow. K-type thermocouples were used to measure the temperature of the heating furnace as well as the catalyst bed. Brockhorst mass flow controllers were used to meter the flow of gases, and the product gas composition was analyzed using a micro-GC (Agilent 490) equipped with a TCD detector. A schematic representation of the reactor is shown in Figure 1.

Ni catalyst to some extent during reforming reactions.7 Forced periodic cycling between dry reforming and coke gasification using CO2 has also been reported to improve the time on stream conversion of CH4.7 Alkaline earth metal support for Ni has been shown to provide excellent stability for long-term operation on stoichiometric mixtures of CH4 and CO2.12,16 The addition of basic metal oxides to Ni increases the active surface area; however, it suppresses the activity of Ni for CH4 dehydrogenation and CO2 reforming.16 Fluidized bed reactors are generally more superior to fixed bed reactors for carrying out dry reforming of CH4.17 The catalyst support also plays a major role in deciding the coking propensity. When metal is supported on metal oxide with Lewis basicity, it helps attenuate carbon deposition.5,16,18 On any support the carbon deposition is more pronounced when catalyst particles are larger.5 A summary of some of the works that have been done on CO2 reforming of CH4 is given in Table 1. It is evident from Table 1. Existing Studies for CH4 Dry Reforming on Supported Ni Catalysts CH4/CO2

T (K)

p (atm)

0.28−0.86 1.0

1213 1063

1.0 1.0

1.0

673−823

1.0

1.0 1.0

673−823 773−1073

1.0 1.0

0.4−1.0 1.0 1.0 1.0 1.0 1.0−1.5 2.0

997−1173 673−823 923−1023 1073 1023 823−973 573−1173

1.0 1.0 1.0 1.0 1.0 1.0 1.0

catalyst NiO/Al2O3 NiO on MgO, CaO, BaO, SrO NiO on SiO2, C, TiO2, MgO NiO on SiO2, TiO2, MgO NiO on MgO, CaO, K2O, Na2O NiO on Al2O3 Ni/Al2O3, Ni−Co/Al2O3 La−NiMgAlO NiO on SiO2 Ce−Co−Ni/Al2O3 Ni/Al2O3 NiSrZrO3

ref 1 12

Figure 1. Schematic representation of the reactor used in the experiments.

8 19 16

Although the reverse water−gas shift reaction produces water, in our experimental set up the gases are cooled to 0 °C without using a gas liquid separator before sending to the GC; thus, the experimental set up does not allow the detection of water. The gas liquid separator is purposefully not used to avoid any delay in recording the reactor response by GC that may arise because of the accumulated gas volume in the gas liquid separator. We observed 5% error in closing H2 balance, which probably arises from the undetected water vapor present in the product stream. The weight hourly space velocity (WHSV) is maintained constant at 12.0 m3/kg·h for all sets of experiments. The flow rates of N2 and CO2 are maintained constant in all experiments at 110 and 13 mL/ min, respectively, and the flow rate of CH4 is adjusted to obtain different CH4/CO2 ratios (1−2). The gases (CH4, CO2, and N2) were purchased from PraxAir (99.99%). Catalyst Characterization. XRD patterns for the catalyst samples were collected between 2θ = 10° and 90° by (X’Pert PRO PANalytical, Netherlands) using Cu Kα radiation (λ = 0.15418 nm) at a scan rate of 0.01670 s−1 . The BET surface area was measured using a Micromeritics ASAP 2020 surface and porosity analyzer. The samples were degassed at 200 °C for 6 h, and the pore size distribution was calculated using the BJH method. TPR studies of the prepared catalysts were performed using a Micromeritics Autochem II-2920 chemisorption analyzer. H2 pulse chemisorption was performed to determine the active metal surface area and metal dispersion. Thermogravimetric analysis (TGA) (TA analytical SDT Q600) was performed in O2 atmosphere in order to analyze the nature and stability of the carbon deposited, whereas Raman data (BRUKER Sentera micro-Raman spectrometer with 532 nm excitation source) was used to confirm the presence of carbon nanostructures.

17 20 3,5,6 4 7 9 15

the table that most of the work focuses on a CH4/CO2 ratio of 1. However, depending on the substrate the CH4/CO2 ratio in biogas may vary anywhere between 1 and 3. Therefore, direct reforming of biogas without the supply of excess CO2 than the stoichiometric requirement and catalyst modification is susceptible to carbon formation. Since there is significant interest in utilizing biogas as a renewable source of energy, the focus of the present work is the analysis of carbon deposition during its dry reforming and short-term stability of the catalyst at different operating temperatures and CH4/CO2 ratios. Moreover, the number of studies that looks into carbon deposition and analysis of the nature and structure of deposited carbon during biogas dry reforming is rather limited in the literature. CH4/CO2 ratios ranging from 1 to 2 are considered in this study, and the temperature is varied from 600 to 800 °C. Furthermore, catalyst deactivation as a result of carbon deposition is reported in time on stream.





EXPERIMENTAL SECTION

RESULTS AND DISCUSSION Dry reforming experiments are carried out for five different CH4/CO2 ratios (1.0, 1.25, 1.5, 1.75, and 2.0) and three different temperatures, and the exit compositions are measured using a micro-GC every 8 min. After each set of experiments, the catalyst sample and deposited carbon are collected; however, characterization of spent catalyst and carbon

The γ-Al2O3 pellets were calcined at 800 °C in air for about 4 h to remove any volatile matter. Ni was loaded onto the pellets by the wet impregnation method. Measured quantities of nickel(II) nitrate hexahydrate (Merk 99% purity) were dissolved in deionized water with a volume slightly higher than the pore volume of the pellets. Subsequently, the pellets were added to the precursor solution and stirred continuously for almost 2 h. The pellets were then dried at 80 B

DOI: 10.1021/acs.energyfuels.5b01862 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

higher at 600 °C compared to 800 °C, which is in very good agreement with thermodynamic predictions.

deposited is done only for reactions performed at 800 and 600 °C with CH4/CO2 = 2.0 and CH4/CO2 = 1.0. Figure 2 shows the XRD pattern of Ni/γ-Al2O3 before and after dry reforming reaction at 800 °C. XRD patterns of both

Table 2. Amount of Carbon Formed at Different Temperatures and CH4/CO2 Ratios Per Gram of Sample CH4/CO2

wt (g) at 600 °C

wt (g) at 800 °C

1 2

0.39 0.61

0.04 0.52

Figure 4 represents the Raman spectra of carbon deposited on Ni/γAl2O3 at 800 °C and CH4/CO2 = 2. The band which

Figure 2. XRD pattern for Ni/γAl2O3 before and after reaction at 800 °C for CH4/CO2 = 2. Data in black represent sample before reaction, and data in red represent the sample after reaction.

catalysts showed signature peaks at 2θ = 44.520°, 51.770° corresponding to (111), (200) planes of Ni, whereas the peaks at 2θ = 67.290° and 76.580° represent the (220) and (311) planes of Al2O3. For the spent catalyst, the additional peak observed at 2θ = 26.10° (0.341 nm) is due to the (002) plane of graphitic carbon. Ideal graphite has an interlayer spacing of 0.3348 nm, while turbostatic graphite has an interlayer spacing of 0.344 nm. The XRD study indicates the presence of graphitic carbon.21,22 Since XRD indicates the presence of graphitic carbon, thermogravimetric analysis of fresh and spent catalyst at 800 °C for CH4/CO2 = 2 is performed to understand the nature and stability of deposited carbon. Figure 3 shows the TGA trace recorded in O2 atmosphere. The weight loss for the fresh catalyst (Figure 3a) below 500 °C is because of the desorption of gases as a result of burning of the binder present in the catalyst. The subsequent weight gain is probably because of NiO formation in O2 atmosphere. The weight loss of the spent catalyst after 550 °C is due to oxidation of crystallographic carbon and typical for multiwalled carbon nanotubes. The amount of carbon formed in each case is shown in Table 2. The table clearly shows that the amount of carbon formation is

Figure 4. Raman spectra of carbon deposited on catalyst at 800 °C for CH4/CO2 = 2.

appears at 1571 cm−1 corresponds to planar vibrations of carbon atoms present in graphite-like materials, whereas the D band which appears at 1335 cm−1 is due to structural defects in graphite-like carbons. The band observed at 2678 cm−1 is due to the overtone of the D band, whereas the band at 300 cm−1 corresponds to RBM because of radial expansion and contraction of nanotube, indicating formation multiwalled carbon nanotube (MWCNT). Typical ID/IG of 0.84 indicates the formation of MWCNTs during the course of the reaction.23,24 The FTIR spectra of the coke deposited during the reaction for CH4/CO2 = 2 for the three temperatures considered in this study are shown in Figure 5. The characteristics of the coke can be identified from the different absorption bands found between 700 and 3700 cm−1. The frequency at around 2850 and 2960 cm−1 corresponds to the symmetric and asymmetric

Figure 3. Percentage weight loss of spent catalyst at 600 and 800 °C. (a) TGA trace for fresh catalyst. C

DOI: 10.1021/acs.energyfuels.5b01862 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

aromatic rings. The peak at 1460 cm−1 is assigned to asymmetric deformation of aliphatic compounds. Peaks between 700 and 1400 cm−1 are attributed to skeleton vibrations of CH in CH, CH2, or CH3 aliphatic groups.25−27 The fresh calcined and reduced samples of Ni/γ-Al2O3 gave a BET surface area of 151 and 106 m2/g, respectively. On the basis of TPR studies the maximum reduction temperature of the catalyst occurred at 790 °C. The pulse chemisorption analysis gave 3% metal dispersion and 19.9 m2/g metallic surface area. The particle size (34.2 nm) is calculated according to the Scherrer equation. The CH4 and CO2 conversion for three different ratios is, respectively, shown in Figures 6 and 7. The conversions are calculated according to Xk = 1 − nk,out ̇ /nk,in ̇

(1)

Here ṅk,in and ṅk,out are, respectively, the inlet and exit moles of species k, which is either CH4 or CO2. The conversions are maximum at early times and falls down gradually. The deactivation is faster at higher temperatures and higher CH4/ CO2 ratios. Figure 6a shows that the CH4 conversion approaches steady state faster at higher temperature compared to lower temperatures. Furthermore, at 800 °C the CH4 conversion follows a precipitous drop after the initial slow decrease as compared to 700 and 600 °C. Following the trend in Figure 6a one can anticipate that CH4 conversion at 600 °C will assume a steady state conversion lower than that at 700 °C. Figure 6d shows that short-term deactivation is insignificant at stoichiometric condition, and CH4 conversion remains almost stable for all temperatures considered. Although the measurements at stoichiometric conditions do not show any sign of catalyst deactivation, deposited carbon is found in the spent catalyst and also confirmed by the TGA analysis. Although CH4 conversion drops sharply at higher temperature and higher CH4/CO2 ratios, the drop in CO2 conversion at the same ratio is rather slow. At stoichiometric ratio, CO2 conversion is slightly lower than that of CH4. The quick drop in methane conversion at 800 °C indicates that most of the carbon is formed during the initial stages of reaction. The

Figure 5. FTIR spectra of carbon deposited during the reaction: (a) 600, (b) 700, and (c) 800 °C.

stretching frequency of aliphatic −C−H groups. The strength of the peak at 2960 increases slightly with temperature, indicating that unsaturated aliphatic compounds are deposited at higher temperatures. The absorption bands at 1580−1590 and 1630 cm−1 represent the stretching vibration of CC

Figure 6. CH4 conversion at 800, 700, and 600 °C for different CH4/CO2 ratios. D

DOI: 10.1021/acs.energyfuels.5b01862 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 7. CO2 conversion at 800, 700, and 600 °C for different CH4/CO2 ratios.

Figure 8. Exit mole fractions of H2 at 800, 700, and 600 °C for different CH4/CO2 ratios.

trends and remains steady right from the beginning of the experiment for CH4/CO2 = 1. This means the dry reforming reactions immediately reach steady state in the absence of significant carbon deposition. Unlike H2, the CO mole fraction remains steady (Figure 9) for all ratios considered, which indicates the occurrence of RWGS reaction and carbon oxidation. The higher conversion of CO2 compared to CH4 also confirms the occurrence of RWGS reaction. These inferences indicates that at later time scales the state of the surface allows the participation of all species but CH4 in surface reactions. In all cases the H2/CO ratio is higher than 1 at the beginning of the experiment and approaches 1 at steady state. All exit mole fractions are well within the equilibrium predicted limits, which are calculated using DETCHEM software.

carbon thus formed can block the pores, and the catalyst surface becomes inaccessible for surface adsorption. However, more than 50% CH4 conversion is observed after the initial stages of poisoning at 800 °C. This leads to the conclusion that carbon formation approaches saturation after a certain time probably due to the occurrence of the RWGS reaction.6 The water formed by means of the RWGS reaction can gasify C according to the following mechanism 0 H 2 + CO2 ↔ H 2O + CO ΔH298 = 41 kJ mol−1 0 C + H 2O ↔ H 2 + CO ΔH298 = 131 kJ mol−1

The appreciable conversion of CO2 at 800 °C also indicates the occurrence of RWGS. The exit mole fractions of H2 and CO on time on stream are shown Figures 8 and 9. From Figure 8 it can be clearly observed that the H2 mole fraction follows CH4 conversion E

DOI: 10.1021/acs.energyfuels.5b01862 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 9. Exit mole fractions of CO at 800, 700, and 600 °C for different ratios of CH4/CO2.



(2) Wang, S.; Lu, G. Q. M.; Millar, G. J. Energy Fuels 1996, 10, 896− 904. (3) Serrano-Lotina, A.; Daza, L. J. Power Sources 2013, 238, 81−86. (4) Chen, X.; Jiang, J.; Tian, S.; Li, K. Catal. Sci. Technol. 2015, 5, 860. (5) Serrano-Lotina, A.; Daza, L. Appl. Catal., A 2014, 474, 107−113. (6) Serrano-Lotina, A.; Daza, L. Int. J. Hydrogen Energy 2014, 39, 4089−4094. (7) Alenazey, F. S. Int. J. Hydrogen Energy 2014, 39, 18632−18641. (8) Bradford, M. C. J.; Vannice, M. A. Appl. Catal., A 1996, 142, 97− 122. (9) Rathod, V.; Bhale, P. V. Energy Procedia 2014, 54, 236−245. (10) Appari, S.; Janardhanan, V. M.; Bauri, R.; Jayanti, S. Int. J. Hydrogen Energy 2014, 39, 297−304. (11) Appari, S.; Janardhanan, V. M.; Bauri, R.; Jayanti, S.; Deutschmann, O. Appl. Catal., A 2014, 471, 118−125. (12) Ruckenstein, E.; Hu, Y. H. Appl. Catal., A 1995, 133, 149−161. (13) Bradford, M. C. J.; Vannice, M. A. J. Catal. 1998, 173, 157−171. (14) Carvalho, D. C.; Silvester, H.; de Souza, A.; Filho, J. M.; Oliveira, A. C.; Campos, A.; Milet, E. R. C.; de Sousa, F. F.; Padronhernandez, E.; Oliveira, A. C. Appl. Catal., A 2014, 473, 132−145. (15) Evans, S. E.; Good, O. J.; Staniforth, J. Z.; Ormerod, R. M.; Darton, R. J. RSC Adv. 2014, 4, 30816−30819. (16) Horiuchi, T.; Sakuma, K.; Fukui, T.; Kubo, Y.; Osaki, T.; Mori, T. Appl. Catal., A 1996, 144, 111−120. (17) Chen, X.; Honda, K.; Zhang, Z.-g. Appl. Catal., A 2005, 288, 86−97. (18) Ma, J.; Sun, N.; Zhang, X.; Zhao, N.; Xiao, F.; Wei, W.; Sun, Y. Catal. Today 2009, 148, 221−231. (19) Bradford, M. C. J.; Vannice, M. A. Appl. Catal., A 1996, 142, 73−96. (20) Haag, S.; Burgard, M.; Ernst, B. J. Catal. 2007, 252, 190−204. (21) Kim, Y.; Luzzi, D. E. J. Phys. Chem. B 2005, 109, 16636−16643. (22) Chen, Y.; Chen, P.; Hong, C.; Zhang, B.; Hui, D. Composites, Part B 2013, 47, 320−325. (23) Huang, W.; Wang, Y.; Luo, G.; Wei, F. Carbon 2003, 41, 2585− 2590. (24) Srikanth, I.; Kumar, S.; Kumar, A.; Ghosal, P.; Subrahmanyam, C. Composites, Part A 2012, 43, 2083−2086. (25) Guisnet, M.; Magnoux, P. Appl. Catal., A 2001, 212, 83−96. (26) Zhang, H.; Shao, S.; Xiao, R.; Shen, D.; Zeng, J. Energy Fuels 2014, 28, 52−57. (27) Dhanala, V.; Maity, S. K.; Shee, D. RSC Adv. 2015, 5, 52522− 52532.

CONCLUSIONS Dry reforming experiments were performed for five different CH4/CO2 ratios and three different temperatures. The spent catalyst samples were characterized using XRD, TGA, FTIR, and Raman to understand the nature and structure of deposited carbon at a CH4/CO2 ratio of 2. XRD analysis of the spent catalyst showed the presence of graphitic carbon, and the TGA trace of the spent catalyst showed the presence of carbon nanostructures. The presence of carbon nanostructures was further confirmed by the Raman spectra of the spent catalyst. The FTIR spectra of the carbon deposited at three different operating temperatures indicated the presence of unsaturated aliphatic compounds at higher reaction temperatures. The reforming experiments showed a faster deactivation rate at higher CH4/CO2 ratios. However, for a given ratio, toward the steady state the conversions were higher at higher operating temperatures. Our results showed that the short-term deactivation was insignificant at stoichiometric conditions; however, carbon deposition was observed at stoichiometric conditions as well. Higher CH4 mole fractions and lower temperature led to higher carbon deposition in accordance with thermodynamic prediction. The occurrence of the reverse water−gas shift reaction was confirmed by the appreciable CO2 conversion compared to CH4 conversion at later time scales and the steady CO exit mole fractions.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +91 (0)40 2301 6073. Fax: +91 (0)40 2301 6032. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS We acknowledge funding received from CSIR under project 22(0635)/13/EMR-II. REFERENCES

(1) Gadalla, A. M.; Bower, B. Chem. Eng. Sci. 1988, 43, 3049−3062. F

DOI: 10.1021/acs.energyfuels.5b01862 Energy Fuels XXXX, XXX, XXX−XXX