Catalytic Activity and Deactivation Mechanisms of Supported NiO in CH₄ Oxidation Citation Phichitkul, Charunya (1981) Catalytic Activity and Deactivation Mechanisms of Supported NiO in CH₄ Oxidation. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/kmv4-kb66. https://resolver.caltech.edu/CaltechETD:etd-10042006-154422 Abstract The partial oxidation of CH4 was studied using a commercial Ni/α-Al 2 O 3 catalyst. The results show that two operating regimes can be maintained: a low-conversion regime where CO 2 and H 2 O are formed and a high-conversion regime where CO, CO 2 , H 2 and H 2 O are formed. The latter regime is established by "catalytic light-off" which among other conditions depends on temperature, feed composition and catalyst activity. Prior to light-off, the catalyst is in the NiO form. Following light-off, the catalyst is in NiO form near the inlet and metallic form further downstream. A series of NiO/α-Al 2 O 3 , and NiO/ZrO 2 catalysts were prepared by impregnation and calcination in air between 750 and 1050°C. These catalysts were characterized by the BET method, O 2 chemisorption, X-ray diffraction, electron microscopy and ESCA. Test analyses indicate that the supported NiO particles are made up of a large number of small crystallites which become more dispersed after reduction and reunite after oxidation. The activity of these catalysts for CH 4 oxidation under pre-light-off conditions was found to decrease dramatically with increasing calcination temperature. However, the calcined catalysts can be reactivated by reduction in H 2 prior to reaction. The change in activity is attributed mainly to the change in specific activity of NiO which is explained in terms of the change in excess oxygen content. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemical Engineering) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Bailey, James E. Thesis Committee: Gavalas, George R. (chair) Weinberg, William Henry Voecks, Gerald E. Bailey, James E. Defense Date: 22 April 1981 Funders: Funding Agency Grant Number Union Oil Corporation UNSPECIFIED JPL UNSPECIFIED Caltech UNSPECIFIED Record Number: CaltechETD:etd-10042006-154422 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-10042006-154422 DOI: 10.7907/kmv4-kb66 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 3922 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 12 Oct 2006 Last Modified: 25 Feb 2025 22:05 Thesis Files Preview PDF - Final Version See Usage Policy. 4MB Repository Staff Only: item control page

CATALYTIC ACTIVITY AND DEACTIVATION MECHANISMS OF SUPPORTED NiO IN CH 4 OXIDATION Thesis by Charunya Phichitkul In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 91125 1981 (Submitted April 22, 1981) ii ACKNOWLEDGEMENTS I would like to thank my advisor, Professor George R. Gavalas, and Dr. Gerald E. Voecks for providing help and counsel through the course of this investigation and creating a climate in which my own efforts could prosper. I am grateful to the Union Oil Corporation, the Jet Propulsion Laboratory and the Chemical Engineering Department for providing financial support during my tenure as a graduate student. Over the years, I have enjoyed interaction with my colleagues Mr. Mark Siddoway, Mr. Nicholas Kaffes, Dr. Basil Iatrides and Dr. Miretta Stephanopoulos. I am also thankful to my family for believing in me. Finally, I must thank Nora Quezada together with her sister Nelia Quezada for putting up with me during the frantic period of typing this thesis into its final form. iii ABSTRACT The partial oxidation of CH 4 was studied using a commercial Ni/a-Al 20 3 catalyst. The results show that two operating regimes can be maintained: a low-conversion regime where C0 2 and H 20 are formed and a high-conversion regime where CO, C02 , H 2 and H 2 0 are formed. The latter regime is established by "catalytic light-off" which among other conditions depends on temperature, feed composition and catalyst activity. is in the NiO form. Prior to light-off, the catalyst Following light-off, the catalyst is in NiO form near the inlet and downstream. metallic form further A series of Ni0/a-Al 2 0 3 and Ni0/Zr0 2 catalysts were prepared by impregnation and calcination in air between 750 and lOSo 0 c. These catalysts were characterized by the BET method, 0 2 chemisorption, X-ray diffraction, electron microscopy and ESCA. Test analyses indicate that the supported NiO particles are made up of a large number of small crystallites which become more dispersed after reduction and reunite after oxidation. The activity of these catalysts for CH4 oxidation under pre-light-off conditions was found to decrease dramatically with increasing calcination temperature. However, the calcined catalysts can be reactivated by reduction in H2 prior to reaction. The iv change in activity is attributed mainly to the change in specific activity of NiO which is explained in terms of the change in excess oxygen content. v TABLE OF CONTENTS Page ACKNOWLEDGEMENTS i ABSTRACT ii TABLE OF CONTENTS v CHAPTER 1. INTRODUCTION 1 CHAPTER 2. CATALYTIC PARTIAL OXIDATION OF CH1t 6 Abstract Introduction Experimental Results Discussion 7 8 10 CHAPTER 3. EXAMINATION OF NiO CATALYSTS 13 21 SUPPORT~D Abstract Introduction Experimental Results Discussion CHAPTER 4. CA'i'ALYTIC ACTIVITY FOR CH4 OXIDATION OF SUPPORTED NiO CATALYS'.f'S Abstract Introduction Experimental Results Discussion APPENDIX I 33 34 35 36 44 67 87 88 89 90 94 119 127 1 CHAPTER 1 INTRODUCTION 2 Industrial production and use of hydrogen has increased by more than three orders of magnitude since the 1930's. Approximately 3 X 10 12 cubic feet of hydrogen was used in the United States during 1973. Petroleum refining is the primary consumer of hydrogen in which hydrocarbon upgrading in the production of fuels and petro-chemicals is of prime importance. Ammonia production for the large fertilizer industry constitutes the second largest consumer. Methanol synthesis, treatment of foodstuffs, reduction of metals and other minor uses comprise the remaining industrial application of hydrogen. In the past the hydrogen sources in the U.S. have been petroleum and natural gas contributing in a ratio of ~1:3. European countries in contrast have had to rely on petroleum and coal as the primary sources of hydrogen because of the lack of natural gas availability. The process predominately used for hydrogen production from petroleum and natural gas is steam reforming because of the high hydrogen content in the effluent. However, for the heavier components in the petroleum refineries a process called thermal partial oxidation is employed. Steam reforming of hydrocarbons which are heavier than a No. 2 fuel oil, for example, is not commonly practiced because the temperatures are too high for standard tube materials, the steam requirement becomes too high to maintain 3 an efficient operation and the rate of catalyst degradation becomes unacceptable for continuous operation. Thermal partial oxidation (TPO), sometimes known as the Shell or Texaco process (1,2), involves fuel-rich oxidation of hydrocarbons with air or oxygen. The result is a conversion of the hydrocarbon to hydrogen and carbon monoxide. This synthesis gas can be used for petroleum refining, ammonia production and even methane generation. However, so~~ soot is also formed and in industrial applications of TPO an extensive part of the operation is dedicated to soot removal. Use of a nickel catalyst was found (3) to allow operation near the soot limit region where TPO still produced soot. Other properties of catalytic partial oxidation (CPO) such as controllability, start up, and stability, in addition to improved hydrogen yield, with no soot make CPO a more efficient process than TPO. More recently the incorporation of CPO with fuel cells has gained prominence. The feasibility ·Of using the CPO process for the production of H2 has been studied at JPL for the past seven years. It has been found in various experiments that the activity of these alumina supported catalysts decreased slowly with time and reduce the performance in extent of conversion and maximum throughput. 4 The present study is addressed to basic problems in CPO which would help in the optimization of reaction conditions and the design of better catalysts. While the CPO of hydrocarbons poses many fundamental problems, the present study focused mainly on the relationship between activity and physicochemical properties of catalysts differing by the pretreatment procedure. Having established such relationship, the mechanism of catalyst deactivation were identified. 5 LITERATURE CITED 1. Noyes, R., Ammonia and Synthesis Gas, Chemical Process Monograph, #26 2. Kuhre, C.J. and Reed, C.L., Oil and Gas Journal Jan. 12, 1976, 3. 1967 p 110 Houseman, J. and Cerini, D.J., "Onboard Hydrogen Generation for Automobiles" Jet Propulsion Laboratory, NASA Contract NAS 7-100 6 CHAPTER 2 Catalytic Partial Oxidation of CH 4 7 ABSTRACT The partial oxidation of CH4 was studied using a coprecipitated Ni/a-Al 20 3 catalyst. Two distinct operating regimes were observed in the experiments. At low conversions oxidation resulted in a product of C02 and H20 and at high conversions a mixture of CO, H2, C02 and H20 was produced. The latter regime is established after "catalytic light-off" which among other conditions depends on temperature, feed composition and catalyst activity. catalyst is in the NiO form. Prior to light-off, the Following light-off, the catalyst is in the NiO form near the inlet and the metallic nickel form further downstream. 8 INTRODUCTION Hydrogen can be produced by the thermal or catalytic partial oxidation of hydrocarbons with oxygen or air. Most of the work in partial oxidation involved the thermal partial oxidation (TPO) processes that have been developed by Shell and Texaco (1,2) and have been applied industrially. However, a fair amount of soot is produced in the process and a considerable amount of processing equipment is required to remove the soot. The presence of a nickel catalyst has been found to permit operation at lower temperatures and without soot formation at the soot limit region (3). In light of this, hydrogen generation for fuel-cell and other small scale portable uses is being studied at JPL using the catalytic partial oxidation (CPO) of hydrocarbons. The selective oxidation of various hydrocarbons to petrochemicals such as ethylene oxide, maleic or pthalic anhydride, butene, and formaldehyde have been studied extensively and are well covered in literature (4,5). The catalytic oxidation of hydrocarbons in low concentrations by noble metals such as Pt and Pd which is the basis for automobile emission control devices has been the subject of several fundamental studies (6-10). The catalytic partial oxidation of hydrocarbons for the production of hydrogen or synthesis gas has received only limited attention (11-15). 9 It differs from other oxidation processes in the catalyst (Ni) and the temperature range used (>700°C). However, these limited studies have not elucidated the fundamental aspects of the CPO of hydrocarbons. In the present study, the CPO of methane is investigated by utilizing the high-activity catalysts which have been employed in the CPO process at JPL. The effect of process variables such as temperature and 0 2 /CH 4 ratio on the catalyst activity and product distribution are examined. 10 EXPERIMENTAL A coprecipitated Ni/a-Al 20 3 G-56B catalyst, containing 25% Ni, was supplied by the Girdler Catalyst Company (now United Catalyst Inc.). was 58 m2/g The catalyst total (BET) surface area About O.lg of crushes catalyst (60-80 mesh) was used for each loading. The reactor consists of an alumina tube with a 0.47 cm. i.d. The temperature is measured by a movable axial chromel-alumel thermocouple in a 0.15 cm. quartz thermowell located in the center of the catalyst bed. The system has two independently controlled heating sections of 15 and 30 cm. length rated at 500 and 1000 watts repectively. The temperature in the catalyst bed is very close to the furnace temperature at low conversions. The reaction mixture contained H2, 02 and N2 as diluent. CH4 (Linde, 99.997% purity), N2 (Linde, 99.996% purity) and 02 (Amweld, 99.95% purity) were purified by passing, through a dehydrated molecular sieve 13X. A schematic of this set up is shown in Fig. 1. The product gases are passed through a drierite column and the dry product gases are analyzed by an HP 5830A gas chromatograph with a 10' carbosieve (100/120 mesh) column and a thermal conductivity detector. Using temperature programming from room temperature to 150°C (30°:/min.) CH 4 N 2 o2 I I I PLAIN ALUMINA MIXER • !!!!NJ .. I Fig. 1 ~FURNACE • ~ #2 TO G. C. FURNACE #1 THERMOWELL Diagram of Experi!11ent..al Apparatus MOLECULAR SIEVE 13X NEEDLE VALVE BACK-PRESSURE CATALYST BED ! ~!!I VALVE PRESSURE GAUGE PLAIN ALUMINA "H~•~ FLOW METER CERAMIC FILTER---!---!-itn PREHEATER \---THERMOCOUPLE I-' I-' 12 allows the separation of all products: CH4. 02, N2, CO, C0 2 , H2 , He (99.997% purity) was used as a carrier in the chromatography. Certified calibration gases (Scotty) were used for calibrating the gas chromatograph. 13 RESULTS Effect of Temperature In order to stabilize its activity, the catalyst was left under reactant flow at 700°c for several hours prior to data collection. Since the catalyst bed is not isothermal under the experimental conditions employed, the percent of CH 4 consumed is shown with respect to furnace temperature (Fig. 2). The CH 4 consumption increases slightly with increasing temperature up to a critical temperature at which it increases abruptly to a higher level. As the temperature is decreased after reaching the high level of activity, a hysteresis is observed. The dry products composition is shown in Table 1 for a series of progressive reaction temperatures. At the "pre-light-off" conditions, only complete oxidation to C02 and H20 takes place. Following light off, 02 is completely consumed and CO and H2 are the predominant products with some C0 2 and H20 also formed. Prior to light-off, the temperature profile in the catalyst bed is broad, but following light-off, the temperature rises across a thin region (Fig. 3). This region is probably where most of the reaction takes place. 14 TABLE 1 Dry Products Composition (mol %) co H2 7.27 3.89 16.55 25.49 1.06 4.47 15.81 24.43 55.70 1. 72 5.00 15.29 22.29 670 58.12 2.85 5.95 14.08 19.00 654 60.76 4.07 7.04 12.68 15.45 Tfurnace (oC) T exit (oC) 02 N2 CH 4 C0 2 696 716 5.92 68.15 19.39 6.54 718 740 730 755 4.82 68.94 18.98 53.49 0.58 700 725 54.24 675 699 640 612 Feed Composition: (0 2) = 0.16, (CH 4 ) = 0.23, (N 2} = balance; SV = 100,000 hr-l 60 N = 61% BY VOLUME 2 SV = 100,000 hr-l o2JcH 4 = 0. 7 G - 56B CATALYST --- -- - - -700 I I ~ 550 Fig. 2 650 FURNACE TEMPERATURE (°C) Effect of Tempera ture on % CH 4 Consump tion 600 750 OL--~~~~~--L~~~~~~-L-~~~~~~-'-~~~~~___. 201- ;~ ~~ z Vl :::> ~ I- 0 -a.. z 80 100 GAS FEED DIRECTION 800 0 TFURNACE = 718 C 750 700 CATALYST BED 650 before light-off 0 GAS FEED DIRECTION TFURNACE = 7oo C 750 700 ° 0 TFURNACE = 56 C 650 600 light- ff afte 55QL-------_l_-------1..---------L-.------_L_----------1.---------.l..-------~ 2 0 3 4 5 6 REACTOR BED LENGTH (cm.) Fig. 3 Temperature Profile in the Catalyst Bed 7 17 Effect of 02/CH4 Ratio In this experiment, the furnace temperature was fixed at 675°C. The mole fraction of 02 at the inlet was varied while keeping the mole fraction of CH4 fixed at 0.24. The N 2 flow was adjusted so that the space velocity stayed the same at all conditions. The percent of CH 4 consumed and the product yield are shown as a function of 0 2 concentration in Figs. 4 and 5. Al low 0 2 /CH 4 ratios, the catalyst activity and the product distribution are similar to those following light-off. Upon increasing the 0 2 /CH 4 ratio, the CO/C02 in the product decreases (Fig. 5) and the position of the steep temperature gradient region is moved further down the bed (Fig. 6). At the 02/CH4 ratio of 1.43, this region is driven out of the catalyst bed and the CH 4 consumption drops suddenly from almost 100% to about 7%. At this point, only complete oxidation to C0 2 is observed similar to the pre-light-off condition. bed (Fig. 6) increased with increasing 02/CH4 ratio. This indicates that the sudden drop in conversion was not due to thermal extinction in which the reaction rate drops suddenly when the catalyst temperature is decreased below the extinction temperature. -a.. z # u ':l:.~ u 0 z ::;:) V') := t- 0 I I Fig. 4 0 I 2or- 40 60 80 100 2 4 o tcH RATIO -1 1 Effect of 0 2 /CH4 Ratio on % CH 4 Consumption 0.5 ----- SV • 60, 000 hr (CH ) = 0. 24 4 (N ) = BALANCE 2 FURNACE TEMPERATURE= 675°C G- 568 CATALYST ,, .... ,, ,, I 1. 5 . 6 I co ~ a::: a_ ·O 0 ::::> (.) I- ->-w _J 0 ~ ~ 0 l...L. u :c""'" u 0 z V> ::::> ~ Ll.J 0 0.5 • • • • 1.0 • • I • SV = 60, 000 hr - l •:co •= co2 FURNACE TEMPERATURE= 675°C (N 2) = BALANCE (CH ) = 0.24 4 G - 56 B CATALYST Fig. 5 Effect of 0 2 /CH 4 Ratio on the Product Yield 02 CH RATIO 4 1.5 O.__~~-J.--~~~~~__,_~~--~-4---L-~--~~___, 20~ 40 ._ 60~ so~ 100 ._ 120 \.0 I-' GAS FEED DIRECTION 750 02 CH = o.68 700 4 CATALYST BED 6500 I 2 3 4 5 6 7 900 800 750 02 700 CH = 1. 07 4 650L-~~-1-~~~L-~~-L-~~~L--~~-'-~~---J~~----' 0 I 2 3 4 5 6 GAS FEED DIRECTION 750 02 a:r = 1.25 4 REACTOR BED LENGTH (cm.) Fig. 6 Temperature Profiles at Different 0 2 /CH 4 Ratios 7 21 DISCUSSION The role of a supported nickel catalyst on the reaction of methane-rich/oxygen mixtures has been investigated. The sudden change in the carbon oxide product distribution upon reaching a critical condition was observed. It is anticipated that this behavior is due to an in situ modification of the catalyst properties. For the nickel oxidation reaction: 2 Ni + 02 (1) 2 NiO + the dissociation pressures of NiO at various temperatures are show11 in Table 2. Prior to light-off, 0 2 partial pressure at the catalyst bed exit is much higher than the dissociation pressure (see Table 1). The 02 partial pressure next to the catalyst can be calculated from: rate where rate k02 02 02 = k0 2 (02 D - 02 n \rt/ ) = rate of ox ya en consumption = oxygen mass transfer coefficient - 02 b 0 2* = oxygen concentration in the bulk gas = oxygen concentration next to the catalyst surface Preliminary calculation showed that 02 * was about the same order of magnitude as 0 2 b for the reaction under the pre-light-off condition. Therefore, the catalyst under the pre-light-off condition was in the NiO form throughout the bed. 22 TABLE 2 Dissociation Pressure of NiO as a Function of Temperature T(°C) Dissociation Pressure (a tm. ) 900 x 10-19 2.52 x 10-16 2.86 x 10-14 2.66 x 10-12 1000 1. 22 x 10-10 1100 3. 19 x 10- 9 1200 5.37 x 10- 600 700 800 1. 46 8 23 From equation (A) , the 02 * is lower at higher rate of reaction. Since the rate of reaction is, among others, a function of the CH 4 partial pressure, the 02 * will be a function of 0 2/CH4 ratio. Following light-off, the equilibrium composition of CH 4 , C02, CO and H2 were calculated from the equilibrium constants of the following reactions: CH4 + H2 0 -+ CO + 3 H2 ( 2) In Table 3, the CO/C0 2 ratios (from Table 1) are compared with the thermodynamic equilibrium ratios at the catalyst bed exit temperature. Although the space velocity in the present study was extremely high (100,000 hour -1 j, the CO/C0 2 ratios in the product following light-off was close to the equilibrium values. Under cothi'aercial operation in which the space velocity is much lower, the product composition should always be at equilibrium. The reaction between hydrogen and oxygen: ( 4) is known to be much faster than reactions (2) and (3) and is most likely at equilibrium. The equilibrium partial pressures of 0 2 have been calculated from the equlibrium concentrations of H 2 and H 20 and the equilibrium constants of the reaction (4). The equilibrium partial pressures of 02 were several orders of magnitude lower than the NiO dissociation 24 TABLE 3 Comparison with Theoretical Thermodynamic Equilibrium Tfurnace (oC) T exit (oC) 740 755 4.25 5.25 700 725 3.54 4.53 675 699 3.06 3.88 640 670 2.37 3.15 612 654 1.80 2.75 Feed Composition: (0 2 ) = 0.16, (CH 4 ) = 0.23, (N 2 ) = balance; SV = 100,000 hr-l 25 pressure (Table 4). Therefore, following light-off, the catalysts downstream of the maximum temperature was in the metallic nickel form (Fig. 7). The light-off phenomenon can be explained in terms of changes in the catalyst-oxygen interaction. Under the pre- light-off conditions, it can be seen from the stoichiometry of complete CH4 oxidation: CH2 + 202 --+ C02 + 2H20 that the 02/CH4 ratio decreases with increasing extent of reaction. Thus the 02 * decreases along the catalyst bed. Under the pre-light-off conditions, the 02 * was higher than the dissociation pressure throughout the bed. Upon increasing the furnace temperature, the rate of reaction was increased. Consequently, the 0 2 * was decreased. When the temperature was raised above the critical temperature, the 0 2 * near the catalyst bed exit (where the 0 2/CH4 ratio is the smallest) became less than the dissociation pressure and NiO was reduced to metallic nickel. The activity of the metallic nickel is higher than that of the NiO and the consumption of CH 4 increased suddenly. This resulted in the temperature profile at the beginning of the metallic nickel zone becoming very steep. A similar explanation can be given for the effect of 02/CH 4 ratio at the inlet as Shown in Figs. 4 and 5. light-off occured at low 0 2/CH4 ratios. The When the 02/CH4 ratio at the inlet was increased, the extent of reaction had 26 TABLE 4 Equilibrium Partial Pressure of 02_ as a Function of Temperature Texit 755 725 699 670 654 Feed Composition: ( 0 c) 02 equil i bri um ( atm.) 1. 31 x 10- 21 x 10- 22 5.32 x 10-23 24 9.15 x 103.11 x 10- 24 2.24 (02) = 0.16, (CH 4 ) = 0.23, (N 2 ) =balance Fig. 7 REACTANTS FLOW Ni The Nature of Catalys t After Light-O ff NiO TEMPERA TU RE PROF ILE 1 N -..J 28 to increase before the 02 * equaled the dissociation pressure. Consequently, the position of the steep temperature profile zone was moved further down the catalyst bed. When the 0 2/CH 4 ratio at the inlet was too high, the 0 2 * near the exit of the bed was still higher than the dissociation pressure. As a result, the metallic nickel zone could not exist in the bed and the reaction rate drops dramatically. The observation that the CO/C0 2 ratio was less than the equilibrium value (Table 3) was consistent with the fact that C0 2 was the primary product formed in the NiO zone. In order to better understand the reaction mechanism in the metallic nickel zone, the % C0 2 conversion [ (co 2) I (CH 4) . 1 t x 10 0 ] in e was calculated. In Table 5, it can be seen that the amount of C02 formed prior to light-off was higher than that after light-off. Thus, some C0 2 that had been formed in the NiO zone was consumed in the zone of metallic nickel. Finally, it can be seen that the critical conditions for light-off are strongly dependent on the activity of NiO (upstream of the maximum temperature). When the activity of NiO is increased, the 02/CH4 ratio decreases more rapidly along the bed and the 02* is lower for a fixed 02/CH4 ratio. Accordingly, the light-off occurs at lower temperature and/or higher 0 2 /CH 4 ratio. After light-off, it can be seen 29 TABLE 5 Comparison of% C0 2 conversion before and after Light-off TFurnace ( oc) before light-off after light-off - - - x 100 (CH 4 )inlet 696 25.2 718 26.8 740 18.51 700 20.6 675 22.7 640 26.0 612 29.6 30 that the maximum temperature also depends on the activity of NiO. 31 LITERATURE CITED 1. Noyes, R. Ammonia and Synthesis Gas, Chemical Process Mongraph, #26, )967 2. Kuhre, C.J. and Reed, C.L. Oil and Gas ~ournal, Jan. 12, 1976 p. 110 3. Houseman, J. a.~d Cerini, D. J. "O!lboard Hydrogen Generation for Automobiles", Jet Propulsion Laboratory NASA Contract NAS 7-100 4. Emmett, P.H., Catalysis Vol. VII, Rheinhold Publishing Corp • 5. ( 19 6 0 ) Dadyburjor, D.B., Jewur, S.S. and Ruckenstein, E., Catal. Rev. Sci. Eng. 19 (2), 293 (1979) c., J. 6. Patterson, W.R. and Kemball, 7. Mezaki, R. and Watson, Ind. Eng. Chern. Process Design Catal. 2,465 (1965) Develop. 5,62 (1966) 8. Accomazzo, M.A. and Nobe, K., Ibid. 4,425 (1965) 9. Cullis, C.F. and Keene, D.E., J. Catal. 19,378 (1970) 10. Wei, J. Advan. Catal. 24, 57 (J.975) and references therein 11. Yoshitomi, s., Morita, Y. and Yamamoto, K., Bull. Japan Petro. Inst. 4,15, (1962) 12. Schmulder, Brennstaff-Chernie, 46,23 (1965) 13. Prettre, M. et al. Trans. Faraday Soc. 42,335 (1946) 14. Peter, K. et al. Brennstoff-Chemie, 36,257 (1955) 32 15. Huszar, K., Racz 1 Gy, and Szekely, Gy, Acta. Chim. (Budapest) 70,4,287 (1971) 33 CHAPTER 3 Examination of Supported NiO Catalysts 34 ABSTRACT A series of NiO/a-Al20 3 and NiO/Zr02 catalysts were prepared by impregnation and calcination in air between 750 and 1050°c. Examination of the catalysts by means of the BET method, 0 2 chemisorption, X-ray diffraction, electron microscopy and ESCA showed marked variations in physical properties as a function of calcination temperature. The results indicate that NiO does not interact with Zr0 2 whereas NiO-Al20 3 interaction is extensive above about 8S0°c. The results also indicate that the supported NiO particles are made up of a large number of small crystallites which disperse after reduction and agglomerate after reoxidation. 35 INTRODUCTION The partial oxidation of hydrocarbons for the production of hydrogen has been demonstrated at the Jet Propulsion Laboratory (1) with a process development reactor utilizing Ni/Al203 catalysts. However, the activity of these catalysts was found to decrease slowly with time with an attendant decrease in reactor performance. To better understand how to maintain catalyst activity and reduce the severity of catalyst degradation, we have begun to study the relationship between the physicochemical properties of the catalyst and the activity for CH4 oxidation of supported NiO catalysts. An initial study of partial oxidation of CH 4 utilizing the Girdler nickel steam reforming catalyst G-56 B was reported in the previous chapter. This catalyst, however, was too active and the reaction temperature could not be controlled to maintain isothermal conditions. In the present study, several catalysts were prepared by impregnation of low surface area supports. These catalysts were then calcined in air at high temperatures commensurate with those typical of hydrocarbon partial oxidation. The study reported in this chapter deals with the physicochemical properties of these catalysts. The next chapter will report the activity of the same catalysts for CH4 oxidation. 36 EXPERIMENTAL Catalyst Preparation and Pretreatment Three commercially available supports were crushed and screened to 60-80 U.S. mesh. Table 1 lists the supports and their properties. The supports were impregnated with 2M Ni(N0 3 ) 2 solution at room temperature. After 15 minutes of impregnation, the solution was drained off and the sample was dried at 80-l00°c for 24 hours. The sample was then heated in air at 400°c for 1 hour to decompose the Ni(N0 3 ) 2 . This procedure was repeated several times to obtain the desired nickel weight percent. These catalysts which had not been heated above 4oo 0 c were labeled as "non-calcined". in air at 750 to Subsequent heating 10S0°c of the non-calcined catalysts prepared from the T-61 a-Al203 and Zr02 supports provided the calcined catalyst samples. Chemical Analyses The amount of nickel as NiO was determined by acidic extraction of nickel from the support followed by dimethylglyoxime precipitation (2). The first step in the chemical analvsis was to dissolve the NiO in SN HCl. Since it is known (3-8) that NiO can react with the Al203 support to form NiAl20~ (nickel aluminate spinel) at high temperatures, the nickel 37 TABLE 1 Properties of Supports Support Norton SZ5464 Zr02 * Data not available Crystalline Total Surface Area Compound (m2/g) Pore Volume (cc/g) 0.11 0.015 3.88 0.27 0.39 * 38 content which was available as nickel oxide was analyzed after the high temperature calcination. It has been shown in our laboratory that the nickel in the NiAl204 matrix is not extracted by this method. Hence, the difference in the nickel weight percent before and after the high temperature calcinations is the amount of nickel in the NiA1 20 4 form. For comparison purposes, similar analyses were carried out for the zirconia-supported catalysts. Total Surface Area Measurements A Quantachrome Corp. Sorption System was used to measure the total (BET) surface area by N2 adsorption by a continuous flow method (9). (BET) The application of the Brunauer-Emmet-Teller (10) equation to calculate the amount of monolayer adsorption is discussed in all modern texts on surface chemistry. In the present work, the effective cross sectional area of N2 was taken as 16.2 A 2 . Nickel Surface Area Measurements A pulse flow method based on chemisorption of oxygen at room temperature (11) was used to measure the nickel surface area. Ultra-pure helium (Linde, 02<0.5 PPM) was used as the carrier gas. flowing H2 all the catalysts were reduced in (20 cc/min) at 450°c for 15 hours and then the 39 H2 was flushed out with the He carrier gas at the same temperature for l hour. After the catalyst was cooled to room temperature, pulses of 0.5 to 2 cc of 1% 02 in He were injected. The nickel surface area was calculated by assuming 1.3 X 10 19 molecules of oxygen adsorbed on 1 m2 of nickel surface (11) which is equivalent to two layers of epitaxial NiO (See Appendix I). X-ray Diffraction Analyses The crystalline compound and the average crystallite size of NiO were determined by X-ray diffraction analysis (XRD) using a G.E. XRD spectrometer with Ni filter and CuKa radiation. The mean dimension (D) of the crystallites is related to the pure X-ray broadening (S) by the Scherrer formula (12) : D = K).JS cose In the present work, D was defined as (volume)l/ 3 and this leads to the value of K~0.95 as the half-maximum line width. (13-15) when B is defined The half-maximum line width from NiO (220) and NiO (111) reflecting planes were employed for the alumina-supported and zirconia-supported catalysts respectively to eliminate overlap with support-derived peaks. The instrumental line broadening is determined from the half-maximum line width of a single silicon crystal. 40 SEM Studies The catlysts were studied using an AMR 900 scanning electron microscope which has a resolution of about 5001000 A 200 A 0 0 • The samples were coated with gold to about 100- thickness prior to observation. ESCA Studies ESCA spectra were obtained using a HP 5950 B ESCA spectrometer. The sample chamber was pumped to a pressure of 5 X 10 -9 torr and a monochromatized AlKa exciting radiation was used. All samples were dusted onto a double-sided adhesive taped onto a gold-plated copper sample holder. binding energies for the Ni(2P 312 The electron ), O(lS) and Al(2P) were recorded and compared with those for the NiO and NiAl204 standards. Since the standards and the catalysts built up a charge from 1-3 e.v. due to their insulating nature, the displacement of the binding energy of the contaminant C(lS) peak from a standard value of 284.6 e.v. was used for correction. of 284.6 e.v. was used for correction. The NiO standard used was a commercially available, high purity NiO powder (Ventron Corp., 99.99% purity). The NiA1 20 4 standard was prepared by calcination of the non-calcined NiO/T-375 a-Al20 3 catalyst in air at 1370°C 41 for 50 hours. After calcination, the XRD spectrum (Fig.l) and the chemical analysis showed that NiA1 2 0 4 was the only compound of nickel. are shown in Table 2. The properties of the NiA1 2 0 4 standard ·~· 11 I I I I . 1 I. I '1' . I "" - I I I 1 ! , ' ' I' I ' I ' : : ' : I 1 I il I' I . I i . -· i 11 1 I ' 11 ' ' '' I I H- . 1 .•.• : I : , r ·- '.I ' :' : I i I 1 ·1 - . I i ' 'I . ' I · I1 I i ' 1 • ' I 67 ( 124) , 1 , !, .1 I I r I -- , I -. • ·. · : "" II . ,. ! 1 1 i 1. ' ,. iI 11 · ~ ' I (440) I I t ,1 NiAl204 1 I I I , ". ~j - - I. I i 1 1 II I 'I l' 11 1J ]" 1 j I :I I i ... :I 1 I I, . ., 11' i ' i - ' I i- . I I'. I ' I I I I I 65 '1 I I_ "' : ' ' : ' 11 I .l .l i rl J 1I j: : i 11 Ii ! • I I I I 1·' ' i.t I ' !, I I , I ·n: I 'ii :',II ' 1 • l, ' I I 1 I r I . 1 "' "" I , I11· 1· I I l II I 'ti !: -10- : ' I 'I I' i : ;.., ! r11J T111 11 1Ii1I11:: ; I 111 I:Ii' I I I ' J. I !: i':' ' ·H-l+t+H-f++l+l++++-t++++++++++++++t++I.., I •• 1 f 1 I: ! 11 1 i ' 1 1 l) Ji1 64 63 62 28 ~ ff11tttH1ttIttit Hitt ItHtl tttttttt tHiitttttttttttttHHtflfHtt 1' ' 11 I1 , . - . , :' ' , - ' ' I i :'i::! I i .- 1 ' ;I ll{J [)111 J- 11111 ~, , ~ , . ,: 1 1 I XRD Spectrum of the NiA1 20 4 Standard 66 :' 11 i ,, l11 ii I ! : : - I ' I I . rr I II .I ' ]' · I 1ttttttt1ffi 11mtttttt1+ti• .! 11 I It I. ,[I, I r11 ! · ·1+1 ' : I.' Ii 111';' --r- ·t-r ''1' ' ' ! iI ' , .. , . . " j ' I 'I IJi~1i \ti~ 'I I 1 "' ' ' :. i , I' '' --·: ir' ' . , ... · 1 • ... , " 'i l1i Ii : I . '~l I !, 1 111 : Ii I '.lrn i , , H, ,1,\ ' . II"'.~:: '.11 ~+;_~;~+1-1~.· Fig. 1 N °' •H 1111111 I I I ' i I' :' 1' i' 1 1, 1 " " " "· ..H+H++H+H II II II llH+++Hr T " ! 11 11: I: , '1' I j 11' ' II '·1' ! 1 ·II t - ---j·I•• I· I I ' ' I '. ' ,, j. ' I ,j3o., IV 43 TABLE 2 Properties of NiAl204 Standard Support T-375 a-Al203 Ni Wt. % a 17.3 Total Surface Area 2.37 m2 /g Average Crystallite Size of Ni Al 204 630 A0 XRD Crystalline Compound Color Blue a Chemical analysis prior to calcination 44 RESULTS The calcination procedures and properties of various catalysts are shown in Table 3 to 6. Chemical analyses have shown that the nickel weight percent of the NiO/Zr02 catalysts before and after calcination are the same in all cases. Further evidence of any chemical reaction between NiO and Zr02 to form Zro.s1Nio.220o.11 (Zirconium Nickel Oxide) was sought with XRD but no characteristic lines were observed in the spectra. On the contrary, chemical analyses of the NiO/a-Al20 3-950 and NiO/a-Al20 3-1050 catalysts show that a significant amount of NiO reacts with the Al203 support to form NiAl204 and the interation is more extensive at higher temperature. Since the XRD spectra of these catalysts do not show the lines that are attributed to the NiAl204, it is assumed that the NiAl204 must be in an amorphous state (crystallite size <50-100 A 0 ) • The XRD spectrum of the NiO/a-Al203-l050 is shown as an example in Fig. 2. The average crystallite size of NiO on both the NiO/a-Al203 and NiO/Zr02 catalysts, estimated from X-ray line broadening, is not affected by the calcinations. A slight decline in the average crystallite size of NiO is observed in the NiO/a-Al203-950 and NiO/a-Al203-l050 catalysts for which the NiAl204 formation is extensive. SZ5464 Zirconia SZ5464 Zirconia SZ5464 Zirconia SZ5464 Zirconia SZ5464 Zirconia Ni0/Zr0 2 (Non-calcined) Ni0/Zr0 2 -750 Ni0/Zr0 2 -850 Ni 0/Zr0r950 Ni 0/Zr0 2 -1050 1050°C, 15 hours 950°C, 15 hours 850°C, 15 hours 750°C, 15 hours None Ca lei nation in air 6.6 6.6 6.6 6.6 6.6 Ni Ht. % (After Calcination) a Average NiO crystallite size estimated form X-Ray line broadening 6.6 6.6 6.6 6.6 6.6 Support Ni Wt. % (Prior to Ca lei nation) Catalyst 440 440 440 440 440 a DNiO Properties of Zirconia-Supported Catalysts TABLE 3 NiO, Zr0 2 NiO, Zr0 2 Zr0 2 NiO, NiO, Zr0 2 NiO, Zr0 2 XRD (Crystalline Compound) Grayishyell ow Gray Gray Gray Black Color U1 ""' T-61 a-Al203 T-61 a-Al203 T-61 a:-Al203 T-61 a-Al203 T-61 a-Al203 NiO/a-Al 20 3 (non-calcined) NiO/a-Al20 3-750 r~iO/a-Al 20r850 NiO/a-Al20 3-950 NiO/a-Al20 3-1050 2.20 2.20 2.20 2.20 2.20 Ni Wt. % (prior to calcination) 2.20 2.20 2.20 1.54 1.29 750 c, 2 hours 850 c, 3hours 950 c, 15 hours 1050 c, 15 hours Ni t1t. % (after calcination) None Calcination in air a Average NiO crystallite size estimated from X-Ray line broadening Support Catalyst Physical Properties of Alumina-Supported Catalysts TABLE 4 290 290 300 300 300 a DNiO (Ao) NiO, a-Al203 NiO, a-Al203 NiO, a-Ai203 NiO, a-Al203 NiO, a-Al20 3 XRD (crystalline compound) Greenishblue Green Gray Gray Black Color ~ 0\ 0.49 0.61 0.85 NiO/Zr02-850 Ni0/Zr0 2-950 Ni 0/Zr0 2-1050 a In air at 500°C, 3 hours * Data not available 0.45 Ni Surface Area (m2/g) Ni0/Zr0 2-750 Catalyst 0.34 0.42 0.47 0.53 Before Reduction 1.13 0.93 0.81 0.78 After Reduction 0.59 0.53 0.49 * After Reoxidation Total Surface Area (m2/g) Ni and Total Surface Area of Zirconia-Supported Catalysts TABLE 5 a .i::o. .......i 0.30 0.15 0.12 Ni0/a-Al 203-850 Ni0/a-Al 20 3-950 Ni 0/ a-A 120 3-1050 a In air at 500°C, 3 hours * Data not available 0.24 Ni Surface Area (m2/g) NiO/a-Al203-750 Catalyst 0.10 0.15 0.24 0.38 Before Reduction 0.20 0.24 0.45 0.44 After Reduction Total Surface Area (m2/g) Ni and Total Surface Area of Alumina-Supported Catalysts TABLE 6 0.10 * 0.25 * After Reoxidation a ~ 00 I 1• Fig. 2 r 1 ' · (124) o-H--+ ' ' llttll~lillllJJ, II I II I I llH*C i11l1llll11lllij1f~~llli~:l~ · Jl I ! I ] 1111 IIIll IIIll IIIIII 11111Jl1JIJI llllt I 65 64 67 J 11 :~,I~' 1 I 1 I I I 29 I 1 I I JI • Hllr I. . u NiO ( 22 0) ~oUJJilllJJJJJJJLUlliU~J..J..il.LLL~ lJ IIIIJJ J~.2 " 111111111111 f' -ol 1111111i11111111111 HI I I I I I I l-11111111 t-l-l-l-+-l..OJ..U..H-11: I I I I I I I IJJJ1illJ11--:+++t XRD Spectrum of the Ni0/a-Al 2 0 3 -1050 Catalyst 66 ',,i . 111111111111111111111111111t l l l l "fj~ll11 !II Jll@i mc~·I HI i1*1-...W#t4~ l« ~ rn Jo« f!;i 1"*'1 :µ.1;P!4cHH-1 '" 11:: I .. 1 i j j j j j j j j j I I I II I II l':'l I I I I I I I I I ltttffi ! I I I I I I I lttt1itl 11111111111':'11111111i1111 i 11111i11111 l+++H- '. · Al203 liW-1-U+H-++ttftflit r ~~in ~~· :::::: '() "-I I I I I I I I I iJl+H+~tjitH j ... t .. . f'' 1rnrG-iI 7'. 1.: /1 ii::. \0 50 The nickel surface area of Ni0/Zr0 2 catalysts increases with increasing calcination temperature. The nickel surface area of NiO/a-Al20 3 catalysts increases with increasing calcination temperature up to about 850°c. 850°c at which temperature Above about the NiAl204 formation becomes significant, the nickel surface area decreases with increasing calcination temperature. As expected, the total surface areas of NiO/a-Al20 3 and NiO/Zr02 catalysts decrease with increase in calcination temperature. In order to better understand the properties of catalysts after reduction, the total surf ace area of the reduced catalysts was also measured. It has been found that a monolayer chemisorption of 02 does not affect the total surface area of the reduced catalysts. Thus, the total surface areas of the reduced catalysts were measured in situ after the 02 chemisorption experiments. The total surf ace areas after reduction are also much higher than those prior to reduction. An increase in the total surface area after reduction, however, is not observed with the Zr02 and the Al20 3 supports that have been previously calcined at 1050°c for 15 hours. Similar to the nickel surface area, the total surface area of reduced NiO/Zr02 catalysts increases with increasing calcination temperature. The total surface area of reduced NiO/a-Al203 catalysts 51 increases with increasing calcination temperature up to about 850°c and then decreases. reoxidized in air at The reduced catalysts were soo 0 c for 3 hours. nickel oxidation rates From the reported (16), these catalysts were completely reoxidized by this procedure. The total surface areas were again measured and it was found that after reoxidation they are much lower than after reduction but slightly higher than prior to reduction. The non-calcined NiO/Zr02 catalyst is black. After calcination between 750 and 950°c, the catalyst becomes gray and a grayish-yellow catalyst is obtained at 105o 0 c. Similarly, the non-calcined NiO/a-Al203 catalyst is black. After calcination at 750 and 850°c, the catalyst becomes gray. However, after calcination at 950 and 1050°c green and greenish-blue catalysts, respectively, are obtained. Although these catalysts have been kept in an air atmosphere at room temperature for several months, their color remained unchanged. However, after the reduction and reoxidation as indicated earlier, all catalysts became black. The significance of catalyst color will be discussed in some detail later on. 52 SEM The surface of the T-6la-Al20 3 support has been examined by SEM and found to be quite smooth as shown in Fig. 3a and 3b. The catalysts on this support were also examined by SEM. For each catalyst, several catalyst particles were examined and appeared to be quite uniform in terms of NiO coverage and particle size. Thus~ the following electron micrographs are representative of the respective catalyst. In the micrographs of the Ni0/a-Al 2 0 3 750, 850 and 1050 catalysts (Fig. 4 to 6), irregular NiO particles of about 1000-2000 A 0 sizes can be identified. These NiO particle sizes are much larger than the average crystallite size of 300 A 0 estimated from X-ray line broadening. The surface coverage of the support by NiO is very extensive and a large number of NiO particles are contacting each other. For the sample that has been calcined at higher temperature, the NiO particles are more rounded, and contacts of the NiO particles with each other and with the Al20 3 support are more intimate. As a result, the surface appears smoother. The NiO particle size appears unaffected by the calcinations. 53 - 0 50000 A a) 5000 A0 - b) Fig. 3 M1cr~grctphs of T-61 a-Al 2 0 3 at 2000X (a) and 20000X (b) 54 0 a) 20000 A ~ 0 b) 10000 A ~ Fig. 4 Micrographs of Ni0/a-Al 2 0 3-750 catalyst at 5000X (a) and lOOOOX (b) 55 20000 A0 a) ~ 0 20000 A b) ~ Fig. 5 Micrographs of Ni0/a-Al 2 0 3 -850 catalyst at 5000X 56 ........ 0 20000 A a) b) 5000 A0 .......... Fig. 6 Micrographs of Ni0/a-Al 2 0 3 -1050 catalyst at 5000X (a) and 20000X (B) 57 ESCA The NiO/a-Al203-850 and NiO/a-Al20 3-1050 catalysts were In order to search for amorphous NiA1 20 4 studied by ESCA. by ESCA, the NiO/a-Al20 3-1050 catalyst from which the NiO has been dissolved in a dilute acid was also studied. The surface elemental compositions are shown in Table 7. They reveal only Ni, O and Al for the NiO/a-Al20 3-850 catalyst and additional traces of Ca and Na for the NiO/a-Al203-l050 catalyst. Thus, the catalysts prepared and calcined in our laboratory are free from any foreign metals that can affect the overall catalytic activity. The electron binding energies for the standard samples and the catalysts are shown in Tables 8 and 9. show the spectra of interest in all samples. Figs. 7 to 9 The results, corrected according to the C(lS) peak position, are described below: a) NiO standard ) spectrum has a principal peak at 312 854.4 e.v. and is accompanied by a satellite shoulder at The Ni(2P 855.9 e.v. The nrigin of the shoulder is still a matter of debate (17). Wertheim and Hufner (18) have suggested that the shoulder is due to the shake-up transition d 8 ~d 8 *• Kim and Davis (19) have suggested that the shoulder is due to the exchange interaction of 2P and 3d electrons. A broad satellite peak at 861.4 e.v. is also observed. 15. 15. 8.9 NiO/a-Al20 3-1050 NiO/a-Al203-l050 (NiO extracted) Note: 24. rv14. NiO/a-Al 20r850 51. 48. 42. 52. 43. 0 18. 21. 20. 19. 23. c 1.1 0.9 - - - - - 0.2 - Ti 0.5 - - Ca Na - - - - 1.2 Cl Values followed by(-) indicates no observation of signal; an approximate value. 21. 24. 4.0 NiAl204 - Al 33. Ni NiO Sample Surface Elemental Compositions (At. %) TABLE 7 (ru) indicates 0.42 1.00 rul. 71 0.17 Ni I Al U1 00 59 TABLE 8 Electron Binding Energies for Standard Sa~ples (e.v.) Sample Ni(2P 312 ) 0( IS) NiO 854.4(P) 529.4(P) 855.7 531.4 A1(2P) s 325 861.4 NiA1 2o4 S Ni(2P 856.7(P) 531. 3 74.5 325.4 )-0(1S) binding energy separation (principal peaks) 312 P principal peaks 50 TABLE 9 Electron Binding Energies of the Catalysts (e.v.) O(lS) Al (2P) s 854.9(P) 530.0(P) 73.5 324.9 856.5 b 73.8 324.8 73.8 325.4 Catalyst Ni(2P Ni0/a-Al 203-850 312 ) 861.8 Ni0/a-Al 203-1050 855.2(P) 856.6 530.4(P) b 861.8 Ni0/a-Al 203-1050 856.l(P) (NiO extracted) 862.1 530.7 S Ni(2P 312 )-0(1S) binding energy separation (principal peaks) P principal peak b A broad sate 11 ite s hou 1der I: 534 530 526 Binding Energy (e.v.) Fig. 7 The O(lS) core level peak for the catalysts and the standard samples. All binding energies are corrected for charging (C(lS) = 284.6 e.v.). N(E) .; r--- --=~+- '"t• ~~~._l_~~-+--·~-:~1_·_·,·.·--+--~~+--~~-~--1' ------+· -+- -1--·- ·.:+==.:.:;~;.~--'------L-··-f-HLL ::: :: ~:~:-:-:--L ---+- -,-----~- --- -- - • -1- - - • --1- - E==:~~:=~:=-=~:==~=-r-~:: :: . ---1- .....'-;~~-t-"~---!~.,,;,J---.-'-'-~-+~---'-'.:.tt~-'-'-'.c.:+'-'..-'--.c.:+-'-'-'-'--_;_:..:._ ____ ~~~ +:e--.--:~:-~.:~:j~-~~~.~~-;--. ,. Binding Energy (e.v) Fig. 8 The Ni(2P 312 ) core level peak for the catalysts and the standard samples. All binding energies are corrected for charging ( C(lS) = 284. 6 e. v. ). N(E) ~ -I Binding Energy (e.v.) Fig. 9 The A1(2P} core level peak for the catalyst and the standard samples. All binding energies are corrected for charging (C(lS) = 284.6 e.v.). 64 The O(lS) spectrum has a principal peak at 529.4 e.v. and is accompanied by a satellite shoulder at 531.3 e.v. The spectra are similar to those reported from other laboratories (17-21). b) NiA1 20 4 Standard The Ni(2P 312 ) spectrum has a principal peak at 856.7 e.v. without an accompanying satellite and a broad peak at 863.3 e.v. The spectrum of O(lS) is also a singlet with a binding energy of 531.3 e.v. ) and O(lS) 312 binding energy separation is 0.4 e.v. higher than that observed for NiO. Hence, the Ni(2P The Al(2P) binding energy is observed at 74.5 e.v. c) Ni0/a-Al 20 3 -1050 (NiO extracted) The XRD spectrum has confirmed a complete extraction of surf ace NiO from this sample but portrays no evidence of NiAl O • Th Ni(2P 312 ) and O(lS) spectra are similar to those observed for the NiA1 20 4 both in overall line shape and line width. The Ni(2P 312 ) and O(lS) binding energies are 0.6 e.v. lower than those observed for the NiAl204. Consequently, the Ni(2P ) and O(lS) peak separation for this sample is 312 identical to that of the NiA1 20 4 • Since the separation of peaks is independent of charging, the spectrum of this sample is consistent with the presence of nickel in the NiA1 20 4 form. The result clearly verifies the existence of an X-ray amorphous 65 NiAl204 layer on the NiO/a-Al203-l050 catalyst in agreement with the result from chemical analysis. d) NiO/a-Al203-850 and NiO/a-Al203-l050 The Ni(2P 312 ) spectra for both catalysts exhibit a structural characteristic similar to that observed for the NiO. These spectra have broader line widths than that observed for the NiAl204 due to an accompanying satellite shoulder. The O(lS) spectra also appear quite similar to those observed for NiO, with a broad weak shoulder on the higher binding energy side. Although the observed Ni(2P 312 ) and O(lS) peaks for these two catalysts are shifted to about 1 e.v. higher than those for NiO, their Ni(2P 312 ) and O(lS) binding energy separations agree quite well witr those of NiO. The peak separation, which is a more realiable parameter, indicates that the surface nickel on both catalysts is present primarily as Nio. In addition, there is no detectable difference in the overall line shape between the two catalysts and the NiO. contribution to the Ni(2P negligible. 312 Thus, the ) spectra by the NiAl204 is Evidently, the NiAl204 on the Ni0/a-Al 2 o 3 -1050 is shielded from the ESCA analyzer by the surface NiO. In some earlier studies (20,21), it was suggested that the supported NiO particles might be covered by an spinel overlayer. a~orphous The present data show that only a very small fraction of the surf ace could be covered by the 66 NiAl204 overlayer. The above differences in charging shift observed earlier are probably caused by an uncertainty in the binding energy calibration technique due to the following reasons: differences in the forms of adsorbed carbon species (22), possible carbon reactions with the surface to form a new carbon species, and uneven charging on different parts of the sample (22). 57 DISCUSSION For a typical catalyst, the surf ace coverage by active metal (or oxide) is usually a small fraction of the The active metal (or oxide) particles are support surface. single crystals (crystallites) and the average particle size can be estimated from X-ray line broadening. In the present study, however, the SEM studies of calcined NiO/a-Al203 catalysts have shown that most of the NiO particles are about 1000-2000 A 0 in size, much larger than the average crystallite size of NiO (300A broadening. 0 ) estimated from X-ray line Although the particle size distribution of supported NiO prepared by impregnation is usually broad and some NiO particles of <1000 A 0 size which were not observed by SEM probably exist, it is unlikely that the X-ray line broadening analysis would grossly underestimate the average size of mostly 1000-2000 A as 300 A0 • 0 single crystal particles We believe that the NiO particles that are observed by the SEM are aggregates of several NiO crystallites which are crystallographically misoriented, termed as (23). An idealized sketch of NiO particles is shown in Fig. 10. It is reasonable to assume "mosaic single crystals" that these mosaic single crystals are also formed in the calcined Ni0/Zr0 2 catalysts inasmuch as they have been 68 (/) Q) ,....; u ·.-l +.l H rO iii 0 ·.-l :z; 4-1 0 ~ > c.. 0 f- ..c: u +.l Q) ~ (/) ] N ·.-l ,....; rO Q) '"d H 0 ,....; 69 prepared by the same procedure and have about the same nickel loading per unit surface area of support as that of the NiO/a-Al203 catalysts. The X-ray line broadening established that the contacting NiO crystallites on both NiO/a-Al203 and NiO/Zr02 catalysts did not coalesce completely during the calcination. Under this treatment, however, the necks between contacting NiO crystallites (both within and between NiO particles) grow (Fig. 11) resulting in lower NiO surface area and smoother surface. NiO Surf ace Area Although the NiO surface area has not been measured directly, it can be estimated from: NiO surface area = (total surface area) - (A) + (correction) (1) where A is the exposed surf ace area of the support after reduction. The exposed surf ace areas of the support after reduction are simply the differences between the total surface areas after reduction and the correspondinq nickel surface areas. They are shown in Tables 10 and 11. The correction in (1) is needed because the exposed surface area of the support in an oxidized state is different from that in a reduced state. The volume ratio (Pilling-Bedworth Ratio) of oxide to metal Fig. 11 CALCINED GRAIN BOUNDARY Idealiz ed Sketch of Neck Growth Betwee n NiO Crysta llites ro AIR 0 -...J 71 for nickel on a molar basis at room temperature is 1.6 (24). If the particle strain at the particle-support interface is assumed to be zero at all times, the relationship between the NiO/support and the Ni/support (after reduction) contact areas is: However, the friction at the interface hinders interfacial movement between particles and support and causes a finite strain. Hence: For convenience, we assume that ~iO = 1.2 ~i so that correction= 0.2 A~:· The detailed calculation of ANi is ~he Although the above assumption is shown in Appendix. probably invalid, it can be seen that the accuracies of the NiO surface areas are estimated to :s0.2 A!-:~ which are adequate for our purpose (Tables 10 and 11). As expected, the NiO surface area decreases with increasing calcination temperature (Table 10 and 11). The NiO surface area of the NiO/Zr0 2 -750 catalyst is a factor of three higher than that of tbe NiO/Zr0 2 -1050 catalyst where as the NiO surface area of the Ni0/a-A1 2 0s-750 catalyst is a factor of nine higher than that of the NiO/a-Al 2 0 3 -1050 catalyst. The NiO surface area is much lower than the 0.32 0.32 0.28 Ni0/Zr0 2-850 Ni0/Zr0 2-950 Ni 0/Zr0 2-1050 0.012 0.012 0.012 0.012 Correction (m2/g) 0.07 0.11 0.16 0.21 * Data not available 0.32 0.22 0.18 * After Reduction and Reoxidation NiO Surface Area a (m2/g) Before Reduction a The accuracy of the NiO surface areas is estimated to ~0.012 m2/g 0.33 Exposed Surface Area of the Support After Reduction (m2/q) Ni0/Zr0 2-750 Catalyst Estimated NiO Surface Area of Zirconia-Supported Catalysts TABLE 10 -..J t.J 0.15 0.09 0.08 Ni0/a-Al 20 3 -850 Ni0/a-Al 20 3 -950 Ni 0/ a-A 120r1050 0.006 0.006 0.006 0.006 Correction (m2/g) 0.02 0.06 0.09 0.18 * Data not available 0.02 * 0.10 * After Reduction and Reoxidation NiO Surface Area a (m2/g) Before Reduction a The accuracy of the NiO surface areas is estimated to $0,006 m2/g 0.20 Exposed Surface Area of the Support After Reduction (m2/g) NiO/a-Al 20r750 Catalyst Estimated NiO Surface Area of Alumina-Supported Catalysts TABLE 11 -.....) vJ 74 corresponding Ni surface area (Table 5 and 6). The NiO surf ace area after reduction and reoxidation is slightly higher that that prior to reduction. Nickel Redispersion The increase in the total catalyst surf ace area following reduction is very interesting. If one Ni particle is formed from one NiO particle, the change in the total surface area can be computed (assuming hemispherical crystallites) from: .6Area = .6Areacrystallite + .6Areasubstrate (2nr2Ni - 2nr2NiO) + = -n(r 2 NiO - r 2 (r2NiO - rzNi) .) Ni < 0 Hence, the total surface area should decrease with reduction. The observed increase in the total surf ace areas indicates that nickel redisperses during reduction of NiO. Similarly, if one Ni crystallite is formed from one NiO crystallite, the NiO surface area is related to the Ni surface area by the following relation: ANiO = (1.6)2/3 = 1.4 ANi Thus, the Ni surface area should be less than the corresponding NiO surface area. However, the opposite was observed. The 75 results further support the hypothesis that nickel is redispersed during reduction. As indicated earlier, the nickel surface areas are higher for the catalysts that have been calcined at higher temperature. This is contrary to the numerous reports (25-27) which indicate that the nickel surface area of supported catalysts decrease with increasing pretreatment temperature due to nickel particle-growth. In order to explain this paradox, we hypothesize that the redispersion is more extensive for catalysts that have been A calcined in air at higher temperature prior to reduction. further explanation for this phenomenon is given below. In the case of the NiO/a-Al203 catalysts, NiO is consumed in the interaction with the Al203 which becomes extensive above 850°c. The effect of increased redispersion is offset by the loss of nickel in the NiO - Al203 interaction and the surface area therefore decreases with increasing calcination temperature (~850 0 C). It is anticipated that the nickel redispersion involves actual fracture of nickel particles. When a NiO particle is completely reduced to a Ni particle, its volume decreases by about 40%. Accordingly, there is an increasing tensile stress on the particle and a balancing compressive stress on the support as the degree of reduction increases. At a critical point the particle can move along the particle/ substrate interface to relax this strain energy. The 76 critical stress (friction force) for the movement at the particle/substrate interface is dependent on the interfacial bonding and contact area. In the present study, the NiO particles have been postulated to consist of several NiO crystallites. Since grain boundaries which separate crystallites have incomplete bonding across the interface, the particle will fracture along the grain boundaries at a lower stress than the critical tensile stress for a crystallite splitting. When the friction force is strong and the stress is not relaxed by the interface movement a particle can fracture into smaller particles along the grain boundaries of the constituting crystallites (Fig. 12). As suggested earlier, the redispersion is more extensive for the catalysts that have been calcined at higher temperatures. Following the above interpretation, this is probably due to a stronger bonding between NiO and substrate after calcination at higher temperature. As a result, the particle/substrate interface movement is hindered and particle fracture is more probable. When a nickel crystallite is reoxidized, its volume increases and most of the crystallites that have been segregated from each other during reduction contact each other again. Accordingly, the total surface area that has previously been increased by reduction decreases after Fig. 12 Ni OXIDIZED Idealized Sketch of Nickel Particle Splitting REDUCED NiO -...J -...J 78 As indicated earlier, the NiO surface areas reoxidation. after reduction and reoxidation are higher than those prior to reduction. This is not totally unexpected. The neck growth between contacting crystallites which decrease NiO surface area, is more extensive after calcination at higher temperature. Since the reoxidation temperature (500°c) is much lower than the calcination temperature prior to reduction (~7so 0 c), the neck growth between crystallites that have been separated during reduction and reunited during reoxidation is less extensive than that prior to reduction. The Colors of Catalyst We would like to draw attention to the significance of the changes in catalyst colors after the calcination. These changes are similar to the results obtained by several authors (28-32) in the calcination of NiO in air. They are caused by the change in the concentration of excess oxygen in the lattice and on the surface of NiO (28-33). The 3+ concentration of lattice Ni ions (positive hole) is related to the concentration of lattice excess oxygen by the following reaction: ~02 (g) + 2Ni 2 + = 0 2 - + 2Ni 3+ + Ni[] where Ni[ I is a cation vacancy. Thus, two lattice Ni 3+ ions are formed from one excess atomic oxygen. ( This relationship, 79 however, is not accurate at the surface because some of the surface excess oxygen might exist in the Oi or o- form (30) Accordingly, the nonstoichiometric NiO is a p-type semicondutor in which the electrical conductivity increases with the increasing concentration of Ni 3 + ions (or excess oxygen). Bielan'ski et al. (32) have determined, by the Bunsen-Rupp method, the concentration of the surface and the lattice Ni 3+ ions of the NiO prepared by calcination of Ni(N03)2 in air at various temperatures. The results, shown in Table 12, indicate that the surface and the lattice Ni* ions decrease with increasing calcination temperature. After calcination above 800 C, the surface Ni 3+ 0 ions is the dominant part of the total Ni 3 + ions and the concentration of the lattice Ni 3+ ions is below the sensitivity of the analytical method. Prasad and Tendulkar (28) measured the specific resistance of NiO prepared by calcination of Ni(N03)2 in air and observed an increase in specific resistance with increasing calcination temperature up to above 900 0 C (Table 12) . According to Mitoff (34) the equilibrium concentration of Ni 3 + ions in the NiO lattice at l000°c is in on the order of 10 -4 ions/pair of ions and increases with increasing temperature. Hence, the concentration of Ni 3+ ions in the NiO lattice which decreases with increasing calcination temperature up to above 900°c are not at equilibrium (30). )• 80 TABLF 12 Properties of Nonstoichiometric NiO Ni 3+ ions a (At. %) Specific Resistance b (Ohms) Calcination Temperature (oC) Surface Lattice 400 * 3.0 4880 500 0.28 1.44 * 600 0.192 0.148 7598 700 0.170 0.10 9034 800 0.144 0.00 14772 900 0.122 0.00 very high 1000 0.084 0.00 very high a) From Ref. 32 b) From Ref. 28 * Data not available 81 From the results above, the concentration of Ni"+ ions in the catalysts that have been calcined at different temperatures cannot be observed by ESCA which is capable . of detecting only about 5-10% of a new species ( . 3+ . Ni ions as opposed to ~i 2 +). The black color of all catalysts reoxidation after reduction and indicates that the Ni 3 + concentration in the NiO is increased. Thus, it appears that the Ni 3 + concentration in the NiO that has been calcined at high temperature can be restored to that prior to the calcination by a reduction and reoxidation at low temperature. Appendix: Estimations Qf Ni/Support Contact Area The value of ANi can simply be obtained from the following relationship: ANi = (total surface area of the support - A) (2) It has also been shown that the average NiO (or Ni) crystallite size of both Ni0/Zr0 2 and NiO/a-AJ. 2 0 3 catalysts is not affected by calcination. Since the decrease in Ni/support contact area is caused mainly by crystallite growth, it is reasonable to assume that their Ni/support contact areas are not affected by calcination. This conclusion is supported by a constant value of A for the Zirconia-supported catalyst that have been calcined at different temperatures (Table 10). A slight decrease in A 82 for the NiO/Zr02-1050 catalyst is consistent with a slight decrease in the total surface area of the Zr0 2 support after calcination at 1050°c for 15 hours. On the other hand, the values of A for the Ni0/a-Al 20 3 -750 and Ni0/a-Al 20 3 -850 catalyst (Table 11) are higher th~n the total surface area of the T-61 a-Al 20 3 support (Table 1). This led us to speculate that some fine NiA1 20 4 crystallites are formed after calcination at 750 and 85o 0 c. This speculation is based on the results from several studies (1,5,6) which show that NiAl204 is formed on a coprecipitated NiAl204 catalyst even after calcination at 500°c. ~lt~o~ah the amount of NiAl204 on these two catalysts is minute and not detectable by the chemical analysis, these fine NiAl204 crystallites can considerably increase the total surface areas of the catalysts. Since it has been shown in our laboratory that NiAl204 is not reducible at 4S0°c (the reduction temperature prior to nickel surface area measurements) the value of A is therefore overestimated. After calcination at 950 and 1050°c, the values of A become less than the total surface area of the support. It has been found that the total surface area of the T-61 a-Al203 support is unchanged after calcination at 1050°c for 15 hours. Thus, it appears that these fine NiAl204 crystallites coalese with each other or with larger NiO particles after calcination at about 950QC~ Since the 83 total surface area of the NiO/a-Al20 3 -1050 catalyst (0.10 m 2/g) is about the same as that of the T-61 a-Al 20 3 support (0.11 m2/g), the increase in the total surface area from fine NiA1 20 4 crystallites is evidently negligible. Taking the above interpretation into account, the values of A from the NiO/Zr0 2-750 and NiO/a-Al 20 3 -1050 catalysts are used in equation (2). Hence, for the Zirconia-supported catalysts: 2 ANi = 0.39 - 0.33 = 0.06 m /g, for the alumina-supported catalysts: ANi = 0.11 - 0.08 = 0.03 m 2/g 84 LITERATURE CITED 1. Stephanopoulos, M.F., and Voecks, G.E., "Catalytic Autothermal Reforming of Hydrocarbon Liquid", Presented at AIChE National Meeting, Houston, Texas 2. April 5-9,1981 Welcher, F.J., "Organic Analytical Regents" Vol.III, Van Nostrand, New York, 1947, P 157-199 3. Milligan, W.O. and Merten, L., J.Phys. Chem. 50,465(1946) 4. Richardson, J.T. and Milligan W.O., J.Phys. Chem. 60,1223 (1956) 5. Rubinstein, A.M., Slinkin, A.A. and Pribytkova, N.A., IZV. Akad. Nauk. SSSR, Otd.Khim. N. 814 (1958) 6. Rubinstein, A.M., Akimov, V.A. and Kretalova, L.D., ibid 929 (1958) 7. Lo Jacono, M., Schiavello, M. and Cimino, A., J. Phys. Chem. 75,8,1044 (1971) 8. Simonova, L.G. et al., Kinet. Katal. 14,6,1566 (1973) 9. Nelson, F.M. and Eggerton, F.T., Anal. Chem. 30,8,1387 (1958) 10. Brunauer, S., Emmett, P.H. and Teller, E., J. Amer. Chem. Soc. 60, 309 (1938) 11. Buyanova, N.E. et al., Kinet. Katal. 8,868 (1967) 12. Scherrer, P., Gettinger Nachrichten 2,98 (1918) 13. Murdoch, C.C., Phys. Rev. 31,304 (1928) and 35,8 (1930) 85 14. Stokes, A.R. and Wilson, A.J.C., Proc. Cambridge Phil. Soc. 38, 313 (1942) 15. Klug, H.P. and Alexander, L.E., "X-ray Diffraction Procedures", Wiley, New York 1954 16. Gulbransen, E.A. and Andrew, K.F., J. Electrochern. Soc. 101,3,128 (1954) 17. Oku, Masaoki and Hirokawa,K. ,J. Electron. Spectros. 10,103 (1977) 18. Werthiem, G.K. and Hufner, S., Phys. Rev. Letters 28, 16,1028 (1972) 19. Kim, K.S. and Davis, R.E., J. Electron Spectros. 1,251 (1972/73) 20. Kim, K.S. and Winograd, W., Surf. Sci. 43,625 (1974) 21. Shalvoy, R.B., Rencroft, P.J. and Davis, B.H., J. Vac. Sci. Technol. 17 (1), 209 (1980) 22. Madey, T.E. and Wagner, C.D., J. Electron. Spectrosc. 10, 359 (1977) 23. Figlarz, M., Vincent, F., Lacaille, C. and Amiel, J., J. Powder Technol. 1,121, (1967) 24. CRC Handbook of Chemistry and Physics 25. Shephard, F.E., J. Catal. 14,148 (1969) 26. Richardson, J.T. and Crump, J.G., J. Catal. 57,417 (1979) 27. Kou, H.K. ,Ganesan, P. and De Angeles, R.J., J. Catal. 64,303 (1980) 28. Prasad, M. and Tendulkar, M.G., J. Chem. Soc., P 407 (1931) 86 29. Deren', J., Haber, J. and Stoczyn'ski, J., Bull. Acad. Pol. Sci. Ser. Sci. Chim. 9, 245 (1961) 30. Deren', J. and Stach, J., J. Catal. 18, 249 (1970) 31. Tourky, A.R., Hanafi, Z. and Salem, T.M., Z. Phys. Chem. (Leipzig) 243,145 (1970) 32. Bielan'ski, A., Deren', J. and Wolter, M., Kinet. i Katal. 5,849 (1964) 33. Krylov, O.V., "Catalysis by Nonmetal", Academic Press (1970)' p 66 34. Mitoff, S.P., J. Chem. Phys. 35,882 (1961) 87 CHAPTER 4 Cat~lytic Activity for CH 4 Oxid~tion Catalysts ot Supported NiO 88 ABSTRACT The activity of supported NiO cata~ysts for CH 4 oxidation decreases substantially with increasing calcination temperature. However, the calcined catalysts can be reactivated by reduction in H 2 prior to reaction. The changes in activity are attributed to changes in NiO surface area and the changes _in specific activity of NiO with the latter being more important. The change in specific activity of NiO is explained in terms of the change in excess oxygen content. 89 INTRODUCTION In the previous chapter, the physicochemical properties of supported NiO catalyst were examined and found to vary markedly with calcination temperature. In particular, both NiO surface area and concentration of excess oxygen in NiO decrease with increasing calcination temperature. The concentration of excess oxygen in the NiO that has been calcined in air at high temperature appears to be restored by reduction, and reoxidation at lower temperature than the calcination temperature. This cliapter reports measurements of the catalytic activities of these catalysts for CH4 oxidation. The main objective is to correlate the catalytic activity in terms of different physicochemical properties and, in particular, to determine the effect of reduction prior to reaction. 90 EXPERIMENTAL The catalysts used in the present work were obtained from the same batches as those examined in Chapter 2. ~able 1 and 2 list the catalysts and their NiO and total surface areas. The catalysts' activities for CH~ oxidation were measured with the same apparatus and procedure used previously (1). The reaction mixture contained CH 4 , 0 2 and diluent N2 The catalyst packing consisted of 10-100 mg. of catalyst • diluted with low surface area a-Al 20 3 to reduce the heat produced per unit volume. In order to achieve isothermal operations and obtain a uniform velocity profile, packing of 5 cm. length ~f low surface area a-Al 20 3 were placed before and after the catalyst bed. The reactor was operated in a differential mode with CH 4 conversion of 2% or less. The temperature gradient along the catalyst bed was always within ±2°c. On this basis, the reactor could be considered isothermal. The experlments were conducted at 0 2/CH 4 ratios which are substoichiometric relative to complete combustion. The reaction temperature was maintained within a range wherein the CH4 which reacted was oxidized completely to C02 and H20. Under these conditions, the nickel catalyst was present in the NiO form (1). 0.53 0.47 0.42 0.34 o. 39 Ni 0/Zr0 2- 750 Ni 0/ZrOz-850 Ni0/Zr0 2-950 Ni 0/Zr0 2-1050 Zr0 2 (Norton SZ5464) 0.07 0 .11 0 .16 0.21 * Before Reduction b Reduced in H2 at 450°C for 15 hours and reoxidized in air at 500°C for 3 hours 0.32 0.22 0.18 * * After Reduction b and Reoxidation NiO Surface Area a (m2/g) a The accuracy of NiO surface areas was estimated to ~J.012 m2/g (2) 0.74 Total Surface Area (m2/g) Ni0/Zr0 2 (non-calcined) Catalyst NiO and Total Surface Area of Zirconia-Supported Catalysts TABLE 1 ~ 1--' 0.38 0.24 0.15 0.10 3.88 Ni0/a-Al 20r750 Ni0/a-Al 20 3 -850 Ni0/a-Al 20 3 -950 Ni 0/ a-A 120 3 -1050 a-Al 20 3 (Girdler T-375) * 0.02 0.02 0.10 * * After Reduction b and Reoxidation NiO Surface Area a (m2/g) 0.06 0.09 0.18 * Before Reduction b Reduced in H2 at 450°C for 15 hours and reoxidized in air at 500°C for 3 hours a The accuracy of NiO surface areas was estimated to ~ 0.006 m2/g 0.45 Total Surface Area (m2/g) Ni0/a-Al 20 3 (non-calcined) Catalyst NiO and Total Surface Area of Alumina-Supported Catalysts TABLE 2 N \.D 93 The differential rate of CH4 consumption was calculated from: r where F .w x F = the CH4 flow rate, mol/sec. x = the fraction of CH4 consumed w = the catalyst loading, g. 94 RESULTS Catalyst Activity The activity of the non-calcined NiO/Zr02 catalyst at 7S0°c is shown in Fig. 1. The activity declines rapidly during the first few hours and more slowly thereafter. In Fig. 2, the same behavior is seen for the non-calcined NiO/a-Al203 catalyst. Also in Fig. 2, the activity of the Ni0/a-Al 2 0 3-850 catalyst is shown to be stable at a reaction 0 temperature of 70S c but the activity is measurably reduced in comparison. Therefore, the deactivation can be accelerated by calcination in air for a few hours above the reaction temperature. The catalytic activities and kinetic parameters of the catalysts that have been calcined at high temperatures showed little change during reaction. To avoid possible errors due to activity changes during reaction, verification of all the reaction rate measurements was established by making measurements under standard conditions before and after data collection. Preliminary tests showed that the extent of the gas phase reaction and reaction at the reactor wall were negligible. From the analytical criteria reviewed by Mears (3), the reaction rates were free of nass 2nd heat transfer effects. 5 Q:! I<( LU 0 l..L.. u :::c 'o::::::t z 0 u Vl :::> ~ a_ I- 0 • • • • • • NiO/ Zr0 2 (NON-CALCINED) (02) = 0.08 (CH ) = 0.12 4 (N2) = BALANCE = 1os 0 c T I 4 5 TIME (hr) 6 7 8 9 Deactivation of the Non-Calcined Zirconia-Supported Catalyst 3 Fig. 1 2 1 0 10 0---~--~--~----~_._~--~_._~_._~__._~---~-- I 2 3 - 4 E -z 0 ~ I O'l - V) - 6 7 l.O U1 a=: u.J t<( 0 LL.. u :c""" 0 z u V> :::::> ~ t0.. -0 E -z 0 ~ - Vl I 0 Fig. 2 0 I I'- 2L 3 -b 4 5 x 2 6 x 3 6 x TIME (hr) 4 6 x 5 6 x 6 Deactivation of the Non-Calcined Alumina-Supported Catalyst I 6 x x Ni0/a-Ai o (NON-CALCINED) 2 3 6 Ni0/a-Ai o -850 2 3 (02) = 0.15 (CH ) = O. 25 4 (N2) = BALANCE = 705°C T -------1 I 7 - - - - l.O O"I 97 In Fig. 3, the specific activities ( activity per unit surface area) of the a-Al203 and Zr0 2 supports are presented in Arrhenius plots and shown to be about the same. The activation energies (Table 3) of the two supports are also comparable. The specific activities (activity per unit NiO surface area) of various catalysts are shown in Figs. 4 and 5. The reaction rates of the catalysts that had been calcined at 950 and 1050°c were corrected for the contributions from the supports. The specific activities of both Nio/a-Al20 3 and Ni0/Zr0 2 catalysts decrease substantially with increasing calcination temperature. The activation energies were found to be independent of gas composition. After calcination at and below 950°c, the activation energies (Table 3) change very little and without any noticeable trend. is no compensation effect. Thus, there The activation energies of the Ni0/a-Al 2 0 3 catalysts are slightly higher than those of the Ni0/Zr0 2 catalysts. All of these activation energies are between those obtained under oxygen-rich conditions by Anderson et al. (4) [31.3] and Andrushkevich et al. (5) [20.0]. When the calcination temperature is increased to 1050°c, the activation enerqies increase dramatically (Table 3) • These activation energies are excessively high for a catalytic reaction. The dramatic increases in activation energy might be caused by errors associated with the correction indicated earlier. 0 At 700 c, the contributions from the supports of I I • a- Al 203 1.00 ............. 1. 06 BALANCE 0.25 0.15 SPECIFIC RATE 2 (mo1/m -sec) 1.04 1.02 _!_X 103 (oK-l) T """ .... (02) :s (CH ) ,.. 4 (N2) :s z: Arrhenius Plots of Alumina and Zirconia Supports 0.98 -161- x zr02 -15 Fig. 3 ~ c: er:: --141- R 1. 08 I ():) '° 99 TABLE 3 Kinetic Parameters Catalyst Ea (kcal/mol) Ni0/Zr0 2-750 23.6 Ni0/Zr0 2-850 21.0 Ni 0/Zr0r950 21. 7 Ni 0/Zr0 2-1050 53.3 NiO/a-Al20r750 28.2 Ni0/a-Al 20 3-850 29.3 Ni0/a-Al 20 3-950 26.1 Ni0/a-Al 20rl050 52.4 Zr0 2 36. 7 a-Al 20 3 35.9 a At 600°C a f3 0.43 0.43 a ~ c 0::: ~ e NiO/a -At.p - 1050 3 NiO/a -At.p - 850 3 N = BALANCE 2 6 o NiO la -Al .p 3 - 950 1 . o.1s o NiO/a -Al-fJ - 750 3 ICH l = 0. 25 4 m 2 R =SPECIFIC RATE (mol/m NiO - sl Fig. 4 Arrhenius Plots of Alumina-Supported Catalysts _!_ x lcY (oK-ll T 1. 20 -1s.__~___.'--~---'~~--1~~--1~~--l~~--1.~~-.1.~~-.l.~~--L~~-L~~-L~~-1..~~-' 0. 96 - 'IH -0. 1. 00 1.02 1.06 1. 04 1. 08 1.10 1.16 1.12 1.14 1. 22 1.18 -14 -13 -12 -11 -10~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- 0 0 ...... ~ n:: c 0. 98 1. 00 ·~ -~, NiO/Zr0 - 850 2 NiO/Zr0 - 950 2 NiO/Zr0 - 1050 2 IOzl = 0.15 (CH ) = 0. 25 4 (N zl = BALANCE l. 08 NiO/Zr0 - 750 2 R =SPECIFIC RAlE (mollm 2NiO - S) -----.,....___ 1.12 I I Fig. 5 ix 103 (oK-1) 1. 04 1.06 Arrheniu s Plots of Zirconia -Suppor ted Catalys ts l. 02 1.10 ~--~--1.~--~;;----~-----:-1:::----_l_ _ __J__ __J I I 0. 96 -16 -15 -14 -13 -12 -11 -10 -9 -----.-----------,..----------. -s..----~-----.r----~----.-----------.~--------.----------.----- ..... () ..... 102 the NiO/Zr0 2-1050 and the Ni0/a-Al 2 0 3-1050 catalysts are about 70% of total catalyst activity. It is plausible that activRtion energies of the supports that have been subjected to repetitive impregnation with Ni(N0 3 ) 2 solution during catalyst preparation and then calcination in air at 1050°c for 15 hours are higher than those of the respective non-calcined supports. Since the reaction rates from non-calcined supports were used in the correction, the activation energies were overestimated. The rate of CHi; consumption is fitted by a power law rate expression: r = k • (CH i;) a• ( o 2 ) B The reaction orders on the NiO/a-Al203-750 catalyst are shown in Table 3. The reaction order with respect to CH 4 was obtained by maintaining the 0 2 mole fraction at 0.15 and varying the CH 4 mole fraction from 0.10 to 0.40. Similarly, the reaction order with respect to 0 2 was obtained by maintaining the CH 4 mole fraction at 0.25 and varying the 0 2 mole fraction from 0.1. to 0.35. The reaction order with respect to CH 4 is comparable to that obtained by Andruskevich et al. (5). The fractional order with respect to both CHi; and 0 2 indicates that the rate determining step is the surf ace reaction between preadsorbed oxygen and methane. The specific activities of various catalysts at 700°c are compared in Table 4. The results sho•1 that the specific activity of supported NiO decreases 103 TABLE 4 Comparison of the Activity of Various Catalysts Ca ta lyst Specific Rate a (mol/m2 NiO-S) Ni0/Zr0 2-750 1.29 x 10- 4 Ni 0/Zr0r850 1.04 x 10- 5 Ni 0/Zr0r950 4.23 x 10- 6 1.86 x 10- 7 Ni O/Zr02-1050 (2.05 x 10- 7 ) b Ni0/a-Al203-750 NiO/a-Al203-850 1.46 x 10- 4 2.64 x 10- 5 Ni0/a-Al 20 3-1050 1. 73 x 10- 6 5.35 x 10- 7 a-Al 20 3 (2.50 x 10- 7 ) b Ni 0/ a-A 120 r950 a Measured at 700°C, (0 2 ) = 0.15 and (CH 4 ) = 0.25 b (mol/m 2-s) 104 substantially with increasing calcination temperature. For each increase of l00°c, the specific activity decreases by almost an order of magnitude. As a result, the specific activities of supported NiO after calcination at 10S0°c are over two orders of magnitude lower than those after 0 calcination at 7So c. Moreover, the specific activity of NiO that has been calcined at 10S0°c becomes comparable to those of the Zr02 and the Al203 supports. In comparing NiO surface areas (Tables 1 and 2) and the specific activities of NiO (Table 4) of the catalysts that have been calcined between 750 and 10S0°c, it can be seen that the dominant mode of deactivation is the decrease in the specific activity of NiO due to calcining. Inspection of used catalysts shows that the color of the non-calcined catalyst changes from black to grayish-black with time of operation, whereas that of the calcined catalysts remains the same. This seems to indicate that the concentration of excess oxygen in calcined NiO is not altered during activity measurements. Effect of Reduction To explain the relatively stable but less active character of the calcined catalysts which had undergone the high temperature treatment under oxidizing atmosphere, the catalysts were reduced in H2 prior to activity measurements. 105 Although most of the nickel is in metallic form after reduction, it is reoxidized again when the reactant flow is resumed. As shown in the last chapter, all catalysts have an average Ni crystallite size (500 A From the reported 0 based on XRD analysis. nickel oxidation rate (6), these catalysts should be completely reoxidized in less than 10 minutes under all reaction conditions employed. Thus, the activities after reduction are measured when nickel is in the NiO form. The activities of the Ni0/a-Al 2 0 3 -850 and Ni0/a-Al 2 0 3 -1050 catalysts before and after reduction at 650°c for 1 hour are shown in Fig. 6. The activity of the former is increased by more than an order of magnitude after reduction whereas that of the latter is increased by about 3 orders of ~agnitude. The activity of the NiO/a-Al 2 0 3 -850 catalyst after reduction is higher than that of the NiO/a-Al 2 0 3 -750 catalyst prior to reduction while the activity of the Ni0/a-Al 2 0 3 -1050 catalyst after reduction is comparable to that of the Ni0/a-Al 2 0 3 -750 catalyst prior to reduction. In Fig. 7, a similar behavior is observed with the Ni0/Zr0 2 -850 and Ni0/Zr0 2 -1050 catalysts. After reduction at 600°c for 1 hour, the activity of the former is increased by more than an order of magnitude while that of the latter is increased by about 3 orders of magnitude. As a result of reduction, activities of both catalysts are higl1er than those of the Ni0/Zr0 2 -750 catalyst prior to reduction. 106 x 10-s ~ l::i. x x x x • • • • - V) -cs I x x • • • O'> E ; 10-6 Q xx x I- a... l::i. ~ * :::> V> z NiO/ a -A.£ 203-750 x NiO/ a -A.£203-850 0 • NiO/ a-A.£203-1050 u :c"'3" I 0- 7 REACTION CONDITION: u LL.. 0 ~ <( e:::: 10-8 ••• 1 (02) = 0.085 (CH ) 4 (N2) = 0.14 T = 650°C = BALANCE *REDUCED IN H2 AT 650°C 2 3 4 5 6 7 8 9 10 TIME (hr) Fig. 6 Activation of Alumina-Supp orted Catalysts 107 ~ lo-5 - ~ ~ ~ • ~ •x •x •x x 6. 6. V') - I C"> 0 -z: lo-6 E * -0.a....... ~ :::> xx V'> z: u x NiO/ Z ro 2-850 0 oi:::::!" :::c u NiO/Zr02-750 6. • NiO/ Z ro2-1050 lo-7 u.. 0 REACTION CONDITION: LLJ ...... <( ~ 10-8 ••• (02) = 0.085 (CH ) 4 (N2) = 0.14 T = 610°C = BALANCE * REDUCED IN H2 AT 610°C 10-9 1 Fig. 7 2 3 4 5 7 8 TIME (hr) 6 9 10 11 12 13 14 Activation of Zirconia-Supported Catalysts 108 Although prior to reduction the activity of the Ni0/Zr0 2 -850 catalyst is much higher than that of the Ni0/Zr0 2-1050 catalyst, after reduction the activity of the latter becomes slightly higher than that of the former consistent with their NiO surface areas after reduction and reoxidation (Table 1). In all cases the catalyst activities that have been increased by reduction decline slowly with time on stream. It is reasonable to assume that the NiO surface areas of the catalysts that have been reduced and reoxidized under reaction (>600°c) are comparable to those after reduction and reoxidation (500°c) in Table 1. The increase in NiO surf ace area after reduction and reoxidation (Table 1) is relatively small compared to the increase in catalyst activity. For example, the NiO surface area after reduction and reoxidation of the NiO/a-Al203-850 catalyst is about the same as that prior to reduction whereas the activity increased by over an order of magnitude after reduction. Th:e NiO surface area after reduction and reoxidation of the NiO/Zro 2-1050 catalyst increases by a factor of two or three whereas its activity increases by about three orders of magnitude. Thus, the increase in catalyst activity is due mainly to the increase in specific activity of the NiO. In order to better understand the changes in the specific activity, rate measurements were made for an a:: <C ~ 0 LL. (..) r."'l::S (..) 0 z V) ::::> :2: c.. I- 0 z x ...-I E '-'b 0 en I - V) - LI - * -- Fig. 8 1- 1. 5 2 3 * - 4 6 0 •• •• 5 TIME (hr) I • • I - * - • I = 710°C BALANCE 0.16 0.12 - 7 8 I 9 I 10 REDUCED IN H2 AT 705°C T (N2) = = (CH ) 4 = (02) 2 3 REACTION CONDITION: • • NiO/ a -A.t o -850 •• • I Effect of Alternating Reduction and Oxidation on Catalyst Activity 1 I - ~ •• •o • •• • I l.O 0 ...... c:::: I=! <( 0 LL- u 'I.oe:::r 0 z u V') :::::> ~ z Q Ia.. ~ ........ >< 0 E C'> I - c.n - 0 1 • • • • I I = 0.16 = BALANCE = 110°c 2 I 3 I 4 I 5 I I TIME (hr) * REDUCED IN H2 AT 705°C T (CH ) 4 (N2) 2 3 REACTION CONDITION: (02) = 0.12 • • NiO/ a -Ai, 0 -850 • I 6 -- * -- 7 Fig. 9 Effect of Alternating reduction and Oxidation on Catalyst Activity o.o - -- * . 1. 0 ....... 1.5 • 8 I • I 9 I • • I • IO 0 ...... ...... 111 alternating reduction and reaction (oxidation) sequence. The results are shown in Fig. 8 and 9. The activity gain after reduction is less prominent after each additional cycle. It appears that a slow irreversible process is taking place. In an attempt to explain this phenomenon, the effect of reduction time and temperature were examined. In Fig. 10, a longer reduction time results in approximately the same loss in activity as if the catalyst was under reaction conditions. Thus, the irreversible process continues even in the presence of H2. Reduction at higher temperature (Fig. 11) has no beneficial effect on restoring the catalyst activity. The effect of reduction temperature is also shown in Fig. 12. The rate of CH 4 consumption after reduction (measured at the reduction temperature) decreases with increasing reduction temperature from 650 to 750°c. Without any correction for the differences in reaction temperature, it is evident that reduction for 1 hour at >650°c decreases the activity gain after reduction The effect of 02/CH4 ratio on the rate of decline is shown in Fig. 13. 0 At 7lo c the rate of activity decline appears to be independent of the 02/CH4 ratio in the range 0.2 to 1.2. It is important to stress that, thus far, the activities were measured under reaction conditions in which C0 2 was the only oxygenated carbon product and the catalyst LI cri I<( 0:::: LJ.J 0 u.. :::c u """ u 0 z Vl :::> :E o.. l- 0 z x r-1 '-'"b ~- r-- r-- 0.4 ,!< Fig. 10 0 0. 6 - 0.8 1 1. 2 - ';i 1. 4 :---1 E 1. 6 2 I •• • • • • • •• • ~:.c 4 6 4J • •• • • • • • • • • •• • I 12 -I 10 8 14 I ::.REDUCED IN H AT 705°C 2 I TIME (hr) * (N ) = BALANCE 2 T = 110°c 4 (CH ) = 0.16 (02) = 0. 12 REACTION CONDITION: NiO I a - Al o - 850 2 3 16 IV I-' I-' 1 ~ a:::: <( I- LLl 0 0. 4 • •• 3 • j. 4 TIME (hr) • • • 5 I • • • 6 I 7 I The Effect of Reduction Temperature on Catalyst Activity 2 I • (CH ) = 0.16 4 N =BALANCE 2 T = 110°c (02) = 0. 12 Fig. 11 • IN H AT 2 815°c 1 '"-- • REDUCED o - 850 2 3 REACTION CONDITION: NiO/a - Al 0 0. 6 :c u"""" 0.8 L.J.. z 0 u V) :;E :::::> a.. I- - 0 z .-I ~ ~ 1.4 ~ REDUCED o, E O> IN H AT ~ 2 x 1. 2 705°c 1.6 8 w I-' I-' ~ <( I- LLJ 0 LL. u r.«::j u 0 z (/) :::> 2= 0.... -I- 0 z x r-1 '°0 E 0 01 -- V 'l I ,...._ "- 1 • 650°C Fig. 12 6 3 2 6 • • L\ • • 4 I 5 I I 6 • • 7 6 TIME (HR) I *REDUCED IN H 2 6 • • 8 I 6 • • = 0. 085 9 I T 10 I 11 I 12 I TEMPERATURE = REDUCTION (CH ) = 0.14 2 (N2) = BALANCE (02) N10 /a-AL o - 850 2 3 Effect of Reduction Temperature on the Catalyst Activity I I 760°C 0 6 • • REDUCTION TEMPERATURE 2- - • 700°c ,__ * • 6 • 4 6 ,__ - ,...._ 8- 10 12 ,__ 13 - - 14 15 16 13 ~ ..... """" RIR 5 ~- 1 '- 1. 5 - 2 ~!~ ~ig. 1 13 2 I 0 n 3 I -..L.. 0 n 8 4 I H i 5 I ~ 6 I 0 7 I 8 I 2 9 L_ 10 I ·~REDUCED IN H AT 705°C R = RATE OF CH CONSUMPTION 4 5 AFf ER 5 HO URS ON STREAM T = 710°C 4 (N ) = BALANCE 2 (02) = 0. 055 (CH ) = 0. 25 0 : TIME(hr) ci} w2) = 0. 29 Effect of Oxygen Partial Pressure on Deactivation e o - 850 2 3 • : (02) = 0. 17 A : NI 0 Ia - AI 11 I-' I-' lJl 0::: <( I- LLJ 0 LL. u ~ - 00 I - -- ..; 3 ...._ 4 "I.""1",2 (.) 0 z (./') ::> :2: I0.. - 0 z x r-1 LC'\ 0 E - 0 01 I - 5 V) 6 1 Fig. 14 * 3 I • 4 I • 5 I • 6 I • = 7I0°C 7 I 705°C 8 I * REDUCED IN H2 AT Activity of NiO/a -Al 2 0 3 -850 After Reduction 2 • T (CH ) = 0.27 4 (N 2) = BALANCE N10 Ia - AL 0 - 850 23 REACTION CONDITION: (02) = 0.03 9 °' I-' I-' z 0 -E O:::'. Lt...I u 0 0 I<( > z 8J 1 0 LL. V> 0 - x 0 >- z x .--I (.'.) zLU I<( I 0 I ~ 2- en Vl I - L.LI Lt'\ Cl u <( O:::'. co 0 3 :!~ 1 I I 3 • (l 4 • 0 • (·) 6 ___L 5 TIME (HR) • 0 = 710°C co2 • = co 0 = 7 8 ·:'REDUCED IN H 2 AT 705°C T Fig. 15 Product Distribution at Low o /C Ratio 2 2 • 0 = 0.03 (CH ) = 0. 27 4 (N2) = BALANCE (02) 9 2 3 REACTION CONDITION: NI 0 la -AL o - 850 10 I ....... -..J ....... 118 was in the NiO form. However, when the 02/CH4 ratio at the inlet was lowered to 0.11, the activity (after reduction) did not decline (Fig. 14). Under these conditions, both CO and C0 2 were formed (Fig. 15) and the catalyst was in the metallic nickel form (1). Finally, the reduced catalysts were found to have a grayish-black color irrespective of their original colors. Thus, it appears that the concentration of excess oxygen in calcined NiO is regained after reduction and reoxidation under reaction (2). 119 DISCUSSION The results in this study have shown that the catalytic activity of supported NiO catalysts for CH 4 oxidation decreases dramatically with increasing calcination temperature. However, calcined catalysts can be reactivated by reduction in H2 prior to reaction. This rather interesting change in activity has two explanations that are not mutually exclusive: 1) A change in the specific activity of Nio; 2) A change in the effective NiO surface area. Catalytic Properties Of NiO It has been shown that the change in the specific activity of NiO is the main cause of the change in catalyst activity. This dramatic dependence of the specific catalyst activity of NiO for CH 4 oxidation on the pretreatment atmosphere and temperature requires some discussion. In general, the catalytic properties of semiconducto rs are influenced by the following: a) electronic interaction with the support b) change in surface structure with crystallite size c) collective electronic properties d) localized surface properties These apply to the system examined in the present study in the following ways; 120 a) Electronic interaction with the support The electronic interaction between the catalysts and support has been discussed in the literature for many years. It is known (7,8) that the electronic interaction is significant only with the crystallite consisting of just a few atoms. Since the catalysts in this study have an average crystallite size of greater than 300 A0 , it is anticipated that the electronic interaction with the support is negligible. b) Change in surface structure with crystallite size An important factor that can affect the catalytic properties of supported catalysts is related to the change in structure with particle size. Van Hardeveld and Hartog (9,10) have shown that the surface structure of a crystallite becomes drastically different from that of a large particle only when the crystallite size is less than about 50A 0 • Although the average NiO crystallite size determined from X-ray line broadening is greater than 300A 0 for all catalysts examined here, a small number of NiO crystallites of less than 50A 0 size may have existed prior to calcination at high temperature. These crystallites may have possessed higher activities than large particles due to different coordinations or arrangements on the surface. During calcination at higher temperature, the crystallites would coalesce, and the catalyst activity would decrease drastically. However, the loss in 121 catalytic activity by this mechanism could not be reversed by reduction because small nickel crystallites coalesce extensively both under oxidizing and reducing atmosphere (11-13). Thus, the change in specific activity of NiO could not be explained by a change in crystallite size. c and d) Collective electronic properties and localized properties. The nonstoichiometric NiO is a p-type semiconductor in which the electrical conductivity increases with increasing concentration of excess oxygen (or ~li 3 + ions). As indicated earlier, the concentration of excess oxygen in NiO decreases with increasing calcination temperature, appears to be stable under reaction and increases by reduction. Several studies (14-18) have interpreted the catalytic properties of NiO in terms of its semiconductivity which can be modified by introducing altervalent cations or excess oxygen into the lattice. According to the electronic theory of catalysis (19-21), whenever change transfer is the rate determining step, the overall rate of reaction depends on the electronic properties of semiconductors. Krauss (14) has shown that the activity of NiO for the oxidation of NH3 to N20 is proportional to the amount of excess oxygen in lattice. Hauffe (15) and Winter (16) have shown that the activity of lithium-doped NiO for N20 decomposition increases with the increasing conductivity. Parravano (17) and Schwab 122 and Black (18) demonstrated that the activation energies for CO oxidation on NiO doped with Li 2 0 or Cr 2 0 3 were dependent on its semiconductivity. Taking these prior facts into account, it appears that the changes in specific activity of NiO for CH 4 oxidation are related to the changes in concentration of excess oxygen. Similar to the studies mentioned above, we can phenomenologically explain the change in specific activity of NiO for CH 4 oxidation in terms of the change in its conductivity. However, the localized properties at the NiO surface must surely be affected in some manner by the chanqe in An obvious example is the concentration of conductivity. Ni 3 + ions in the surface layer. Moreover, it has been suggested (22) that the interaction of 02 with the NiO surface takes place at surface defects such as ion vacancies, steps, etc. where the Ni 2 + ions are more exposed. However, the relationship between the conductivity and the properties of the surface defects has not been established. Thus it is plausible to speculate that the specific activity of NiO for CH4 oxidation may be related to the localized interaction at the surface. Thus, we hypothesize that the specific activity of NiO for CH4 oxidation is dependent on the concentration of excess oxygen. However, based on the limited data available, the significance of the collective electronic properties 123 versus the localized surf ace properties cannot be determined but requires further investigation. Finally, it has been shown in chapter two that the concentration of excess oxygen in NiO is a function of calcination temperature. Based on the above interpretation, the specific activity of NiO after reduction is irrespective of the previous calcination temperature but dependent on the reoxidation temperature. It follows then that the difference in catalyst activity after reduction and reoxidation at the same temperature is due to the differences in NiO surface area. For example, the activity after reduction of the NiO/Zr0 2 -1050 catalyst is higher than that of the Ni0/Zr0 2 -850 catalyst (Fig. 7), whereas the activity after reduction of the NiO/a-Al 2 0 3 -850 catalyst is higher than that of the Ni0/a-Al 2 0 3 -1050 catalyst (Fig. 6), in agreement with their NiO surface areas after reduction and reoxidation (Table 1) • Increased Ni•) Surface Area It has been shown in the last chapter that the NiO surface area increases sightly after reduction and reoxidation due to nickel redispersion. Although the increase in NiO surface area after reduction is not the main cause of the increase in catalyst activity, this phenomenon can explain the effect 124 of reduction time and temperature on the catlyst activity. Since the nickel redispersion results from volume change during reduction (2), it must terminate when the reduction is completed. Preliminary calculations based on the results from a number of studies on NiO reduction (23-25), show that the reduction of all catalysts in the present study above 600°C is completed within an hour. Continued heat treatment in H2 will not redisperse nickel crystallites any further but will instead cause the redispersed crystallites to coalesce. This phenomenon accounts for the decreases in activity gain after reduction at 705°C for >l hour (Fig. 10) or for 1 hour at >650°C (Fig. 11 and 12). It is now clear that a slow decline in activity gain after reduction in the alternating reduction and oxidation sequence (Fig. 8) is due to the decrease in NiO surface area. In order to explain this phenomenon, we must keep in mind that the nickel redispersion is more extensive after calcination in air at higher temperature prior to reduction (2). Thus, it can be seen that the nickel redispersion during the first reduction is more extensive than those during additional reductions. Following the first reduction, the redispersed NiO crystallites coalesce during reaction and the coalescence appears to offset the redispersion during additional reductions. As a result, the activity gains after additional reduction are less prominent. 125 LITERATURE CITED 1. Chapter 2 2. Chapter 3 3. Mears, D.E., Ind. Eng. Chem. Process. Develop. 10,4, 541 (1971) 4. Anderson, R.B., Stein, K.C., Feenan, J.J. and Hofer, L.J.E., Ind. Eng. Chem. 53,809 (l961) 5. Andrushkevich,T.V., Popovskii, V.V. and Boreskov, G.K., Kinet. Catal. 6,5,777 (1965) 6. Gulbransen, E.A. and Andrew,. K.F., J. Electrochem. Soc. 101,3,128 (1954) 7. Dalla Betta, R.A. and Boudart, M., Proc. 5th Int. Congr. Catal., Palm Beach, 1972, 2,1329 (1973) 8. S in f e 1 t , J • H. , J . Ca ta 1 . 2 9 , 3 0 8 ( 1 9 7 3 ) 9. Van Hardeveld, R. and Hartog, F., J. Catal. 22,75 (1972) +o. Van Hardeveld, R. and Hartog, F., Surface Sci. 15, 189 (1969) 11. Shephard, F.E., J. Catal. 14,l48 (1969) 12. Richardson, J.T. ana Crump, J.G., J. Catal. 57, 417 (1979) 13. Kou, H.K., Ganesan, P. and De Angeles, R.J., J. Catal. b4,303 (1980) 14. Krauss, w., z. Elektrochem. 53,320 (1948) 15. Hauffe, K., Glang, R. and Engell, H.J., Z. Physik Chem. (Leipzig) 201, 223 (1952) 126 16. Winter, E.R.S,, Dis. Faraday Soc. 28, 183 (1959) 17. Parravano, G., J. Amer. Chern. Soc. 75,1452 (1952) 18. Schwab, G.M. and Block, J., z. Physik. Chern. (Frankfurt) 1,42 (1954) 19. Wagner, C. and Hauffe, K., Z. Elektrochern. 44,1972 (1938) 20. Wolkenstein, F.F., VSP. Fiz. Nauk. 60, 249 (1956) 21. Hauffe, K., 11 Semiconductor Surf ace Physics 11 , (Kingston, R.H. ed.), University of Pennsylvania Press, Philadelphia (1956) p.259 22. Charrnan, H.B., Dell, R.M. and Teale, S.S., Trans. Faraday Soc. 59,453 (1963) 23. Bandrowski, J., Bickling, R., Yang, K.H. and Hougen, O.A., Chern. Cng. Sci. 17,379 (1962) 24. Chiesa, F. and Rigand, M., Can. J. Chern. Eng. 49,617 (1971) 25. Szekely, J., Lin, C.I. and Sohn, H.Y., Chern. Eng. Sci. 28,1975 (1973) 127 Appendix I: o2 Chemisorption by the Pulse Flow Method Selective chemisorption is the most attractive method for determining the active nickel surface area and the dispersion of supported metal catalysts (1,2). Most of the previous studies have used H2 (3-12), CO (3,5,13-20) and o2 (21-26) as the adsorbates. The chemisorption is usually carried out in a static vacuum system. However, it has been found that the simpler and less time consuming continuous flow method (11,13) or pulse flow method (12,25,27) give essentially identical results with the static method. In the pulse flow method, pulses of adsorbate are injected downstream of the reference thermal conductivity detector into the flow of He carrier gas (Fig. 1). After passing over the reduced nickel catalyst sample, the amount of adsorbate left in the carrier gas stream is detected by the thermal conductivity detector. During the first few pulses, no chromatographic peak is detectable. Eventually, a partial peak appears and when the nickel surface is completely covered, the area of subsequent peaks becomes constant (Fig. 2). The amount of adsorbed gas is the sum of completely adsorbed pulses plus partially adsorbed pulses. This method determines only the amount of irreversibly adsorbed gas because the reversibly adsorbed portion is swept from the nickel surface by the carrier gas. The amount of the reversibly adsorbed part can only be determined by a relatively lenghty analysis of the chromatographic peak using the theory of 128 equilibrium chromatography (11,28). H2 chemisorption has been the most extensively investigated method for the determination of nickel surface area. The stoichiometry of H2 chemisorption on nickel is well established as one hydrogen atom per nickel surface atom (5,8,10), except at very low nickel loading and/or when the nickel-support interaction is prominent as in NiTi0 2 (3). However, H2 is adsorbed on nickel both reversibly and irreversibly (3,12,25,29) and the amount of reversibly adsorbed part depends on pressure, temperature (25) and nickel crystallite size (29). CO chemisorption has also been widely used to estimate the nickel surface area. However, CO is obviously not an ideal adsorbate because it can be adsorbed on nickel with the formation of linear and bridged forms depending on temperature, pressure, dispersion and preparation method (3,5,9,18,30,31). Recently (3,19,20), the presence of subcarbonyl 0 0 "c \ / Hi cu 0 ,, 0 II 0 4 c c \ I / c Ni species and adsorption on unreduced NiO has been suggested. This stoichiometric uncertainty can cause a serious ambiguity in the reported area. In addition, at room temperature Ni(C0) 4 is formed (3,9,13,17,32) and, if not corrected can lead to overestimating the nickel surface area. Finally, similar to H2 , CO is adsorbed on nickel 129 both reversibly and irreversibly (3,27). Published results on the chemisorption of 02 on clean nickel single crystal (33), supported nickel catalyst (21,22,23,25,34), nickel powder (21,22) and evaporated nickel film (26,34,35) have demonstrated that the adsorption mechanism consists of ti.'/o regions: first, a fast oxidation leading to two layers of epitaxial NiO and a slow oxygen incorporation into the nickel lattice. The rate of oxygen diffusion into the bulk nickel increases with increasing temperature. For all practical purpose, however, the amount of oxygen incorporation is negligible if the 02 chemisorption experiment is conducted for less than an hour at room temperature (22,23,25). In addition to a known stoichiometry, the rapid chemisorption of oxygen on nickel is irreversible and independent of pressure between <l torr and 100 torr (22,23,25). Thus. oxygen is apparently the most suitable adsorbate for the determination of nickel surface area by the pulse flow method. TCD Fig. 1 VENT CATALYST SAMPLE Diagram of the Flow Apparatus FLOWMETER -..-------... CARRIER GAS INJECTION 0 w I-' 131 0 ...-4 0- ~ 0 ·.-1 +l p_, 00 !--! I 0 Ul ·.-1 s (J) ..c:: u < "° ~ (J) LLJ b"l V') >:: LC"\ _J :::::::> a.. >-t 0 H 0 4-1 """' s '° H b"l 0 +l s'0° N H ..c:: u N . b"l ·.-1 0 ind1no ~OlJ3130 .J ·1 Ii-I 132 LITERATURE CITED 1. Muller, J., Rev. Pure &Appl. Chem. 2. Farranto, R. J., AIChe Symp. Ser. 3. Bartholomew, C.H. and Pannell, R.B., J. Catal. 65,390 (1980) 4. Bartholomew, C.H., J. Catal. 45, 41 (1976) 5. Brooks, C. S. and Christopher, G.L.M., Ibid 10,211 (1968) 6. Taylor, W.F., Sinfelt, J.H. and Yates, D.J.C., J. Phys. Chem. 19,151 (1969) 70 #143 PP 9-22 (1975) 69,3857 (1965) 7. Taylor, W.F., Sinfelt, J.H. and Yates, D.J.C., J. Phys. Chem. 68,2962 (1964) 8. Yates, D.J.C., Taylor, i~.F., and Sinfelt, J.H., J. Amer. Chem. Soc. 9. 86,2996 (1964) Vlasneko, V.M. et al., Kinetika i Kataliz 10. Selwood, P.W., J. Catal. 2,337 (1964) 42,148 (1976) 11. Hansen, A. and Gruber, H.L., J. Catal. 20,97-105 (1971) 12. Roca, F.F. et al. Journal of Gas Chromatography 6,161 (1968) 13. Hughes, T.R., Houston, R.J. and Sieg, R.P., I & EC, Proc. Design & Devel. 1,2 96 (1962) 14. Emmet, P.H. and Skow, N., J. Amer. Chem. Soc. 65, 1029 (1943) 15. Germain, J.E., Ostyn, M. and Beaufils, J.P., J. Chim. Phys. 61, 686 ( 1964) 16. Schuit, G.C.A. and Van Reijen, L.L., Advan. Catal. 10, 242 (1958) 17. Yates, J.T. and Garland, C.W., J. Phys. Chem. 65, 617 (1961) 18. 0 Neill, C.E. and Yates, D.J.C., J. Phys. Chem. 65, 901 (1961) 1 133 19. Rochester, C.H. and Terrel, R.J., J. Chem. Soc. Trans. Faraday I 73, 609 (1977) 20. Primet, M., Dalmon J.A. and Martin G.A., J. Catal 46, 25 (1977) 21. Ljubarskij, G.D., Avdeeva, L.B. and Kulkova, N.V., Kinetika i Kataliz 3, 123 (1962) 22. Muller, J., J. Catal. 6, 50 (1966) 23. Muller, J., Pour, V. and Regner, A., J. Catal. 11, 326 (1968) 24. Vlasenko, V.M., Rusov, M.T. and Juzefovic, G.E., Kinetika i Kataliz 2, 394 (1961) 25. Buyanova, N.E. et al. Kinetika i Kataliz 8, 868 (1967) 26. Brennan, D., Hayward, D. O. and Trapnell, B.M.W., Proc. Roy. Soc. (London) A256, 81 (1960) 27. Brooks, C.S. and Kehrer, Jr., V.J., Anal. Chem. 41, 1, 103 (1969) 28. Kiselev, V. and Yaskin, v.1., Gas-Adsorption Chromatography, Plenum Press (1969) 29. Slinken, A.A., Kucherov, A.V. and Rubinstein, A.M., Kinetika i Kataliz 19, (2), 415 (1978) 30. Eischens, R.P., Francis, S.A. and Pliskin, W.A., J. Phys. Chem. 60, 194 (1956) 31. Trapnel 1, B., 11 Chemisorption 11 (1958) 32. Goldberger, W.M. and Othmer, D.F., I & E.C., Process Design & Devel. 2, 3, 202 (1963) 33. Holloway, P.H. and Hudson, J.B., Surface Sci. 43, 123 (1974) 34. Roberts, M.H. and Wells, B.R., Dis. Faraday Soc. 41, 162 (1966) 35. Delchar, T. and Tompkins, F.C., Surface Sci. 8, 165 (1967)