Home

Bibliography - Ruhr-Universität Bochum

1. Acknowledgments Financial support by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center SFB 558 Metal Substrate Interactions in Heterogenous Catalysis are gratefully ac knowledged Bibliography 1 J B Hansen in Handbook of Heterogenous Catalysis VCH Verlagsgesellschaft Wein heim 4 1997 1856 2 H Wilmer T Genger and H Hinrichsen J Catal 215 2003 188 3 J D Grunwaldt A M Molenbroek N Y Topsge H Topsge and B S Clausen J Catal 194 2000 452 4 A Auroux Topics in Catalysis 19 2002 205 5 M G Cutrufello I Ferino R Monaci E Rombi and V Solinas Topics in Catalysis 19 2002 225 6 B E Spiewak and J A Dumesic Thermochim Acta 290 1996 43 7 S Vollmer G Witte and C W ll Catal Lett 77 2001 97 8 H G Karge and W Nie en Catal Today 8 1991 451 9 B Bems M Schur A Dassenoy H Junkes D Herein and R Schl gl Chem Eur J 9 2003 2039 10 O Hinrichsen T Genger and M Muhler Chem Eng Technol 11 2000 956 11 H Bielawa M Kurtz T Genger and O Hinrichsen Ind Eng Chem Res 40 2001 2793 12 T Genger O Hinrichsen and M Muhler Catal Lett 59 1999 137 13 H Wilmer and O Hinrichsen Catal Lett 82 2002 117 14 E Giamello B Fubini and V Bolis Appl Catal 36 1988 287 15 R Naumann d Alnoncourt M Kurtz H Wilmer E L ffler V Hagen J Shen and M Muhler J
2. Copper catalysts are industrially used for the synthesis of methanol These catalysts are ternary systems containing copper Cu zinc oxide ZnO and alumina Al2O3 1 Several recent studies 2 3 indicate there are strong metal support interactions SMSI between copper and zinc oxide in these catalysts Grunwaldt et al 4 presented a model for these effects Under the reducing conditions of the methanol synthesis the metallic copper surfaces are covered by zinc and oxygen species Under more severe conditions surface and bulk alloying leads to the formation of brass In literature many studies are found investigating copper binary e g Cu ZnO Cu Al203 or ternary copper catalysts e g Cu ZnO Al203 by microcalorimetry temperature programmed desorption TPD or Fourier transform infrared FTIR spectroscopy 5 6 7 8 9 10 Carbon monoxide CO carbon dioxide CO2 and hydrogen H2 are often used probe molecules In most cases a comparison of these studies is difficult as the samples are prepared and pretreated following different procedures For example Giamello et al 8 detected in a microcalorimetric study the presence of Cu l species in a Cu ZnO sample after hydrogen reduction and postulated 28 3 Part I The reduced catalyst these species to be dissolved in the ZnO matrix They assumed that these species might play a role in the synthesis of methanol The authors also pointed out that Boccuzzi et al 9 using FTI
3. Catal 220 2003 249 16 N Cardona Martinez and J A Dumesic Adv Catal 38 1992 149 17 J Strunk R Naumann d Alnoncourt M Bergmann O Hinrichsen and M Muhler in preparation 18 J A Konvalinka J J F Scholten and J C Rasser J Catal 48 1977 365 19 M J Sandoval and A T Bell J Catal 144 1993 227 20 J A Dumesic D F Rudd L M Aparicio J E Rekoske and A A Trevi o ACS Profes sional Reference Book Washington DC 1993 21 K I Hadjiivanov and G N Vayssilov Adv Catal 47 2002 307 3 Part I The reduced catalyst Abstract Our goal is a detailed understanding of the strong metal support interactions between copper and zinc oxide in copper catalysts These interactions are significantly influenced by the pre treatment The adsorption of carbon monoxide is used as investigation tool in microcalorimet ric temperature programmed desorption and infrared spectroscopy experiments Many efforts were taken to assure that the pretreatment conditions are identical for all investigation methods The results in the present contribution refer to the state of the catalysts after reduction by hy drogen at 513 K All investigation methods confirm that the presence of zinc oxide lowers the initial heat of adsorption while catalysts containing alumina have higher fractional coverages The adsorbed carbon monoxide species are at low coverages less mobile on samples free of zinc oxide 3 1 Introduction
4. Figure 5 1 The mass spectrometry traces of hydrogen carbon monoxide carbon dioxide water and methanol recorded while flushing the sample with pure helium at 498 K The flushing out of educts and products is almost completed after 10 min but is carried on for at least 20 min more on line mass spectrometry For the microcalorimetric experiments 100 mg of the sieve fraction of 250 355 um were pre treated in a specially designed pretreatment reactor and then sealed in a pyrex capsule Next the pyrex capsule was placed into the sample receptacle of the microcalorimeter C80 II Se taram The calorimetric set up was degassed and the capsule was broken After reaching thermal equilibrium at 303 K the adsorption measurement was started Small doses of car bon monoxide were subsequently admitted to the sample and while the heat of adsorption was measured calorimetrically the amount of adsorbed species was measured volumetrically In order to test the reversibility of the observed processes the sample was evacuated overnight and the experiment was repeated The employed measurement technique was adopted from the pi oneering work by Spiewak and Dumesic 16 The technique allows to investigate air sensitive samples unimpaired by poisoning A detailed description of the experimental procedure and the set up is given elsewhere 17 72 5 Part III The state of the catalyst after methanol synthesis The TPD experiments were carried out in a
5. Figure 5 7 FTIR spectra obtained with CA2 after after methanol synthesis in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for a pressure of 0 and 100 Pa of CO formaldehyde and methoxy species Formate formaldehyde and methoxy species should be removed by the purging of the sample at elevated temperatures in pure helium that followed each pretreatment Atomic oxygen is therefore the most likely adsorbate to be left on the copper surface Unfortunately the removal of the other adsorbate species can not be confirmed by the FTIR spectroscopy results in this contribution as the noise is to intense in the region of the vibrational frequencies of these adsorbates below 1100cm above 2500cm and 1650 1350cm 3 19 20 A reaction of CO with atomic oxygen forming carbon dioxide followed by desorption or the formation of carbonates on the supporting material can explain the increase of the heat of adsorption found in the experiments after methanol synthesis The formation of carbon dioxide and carbonates are exothermic processes that would increase the measured heat and thus lead to an apparent increase of the heat of adsorption of CO In the case of desorption of carbon dioxide without formation of carbonates
6. Higher values of 58 kJ mol were found for kinks steps and defect structures In an FTIR study Dulaurent et al 19 found a band at 2092 cm after the adsorption of CO on a completely reduced 4 7 Cu Al203 at 300 K for pressures of 1 and 10 kPa They reported isosteric heats of adsorption of 82 kJ mol at low coverages and 57 kJ mol at high coverages Boccuzzi et al 21 investigated the adsorption of CO on Cu ZnO catalysts by FTIR spec troscopy They observed a strong and narrow band at 2098 cm and a broad and weak band at 2070 cm after the adsorption of CO on the reduced catalyst Giamello et al 22 presented a microcalorimetric study of the adsorption of CO on Cu ZnO They assigned heats of adsorp tion of about 300 110 kJ mol to the adsorption of CO on Cu II species heats of adsorption in the range from 110 70 kJ mol to the adsorption on Cu I species and heats of adsorption of 70 40 kJ mol to the adsorption on Cu 0 1975 Pritchard et al 23 presented an infrared spectroscopy study of carbon monoxide chemisorbed on copper The authors compared own data and literature data concerning copper single crystals and copper catalyst samples with various supports The frequency of the stretch ing vibration of adsorbed CO is significantly different for different copper single crystal planes The frequency is lower for low indexed single crystal planes than for high indexed planes 1 e 2085 2076 and 2093 cm for Cu 100 Cu 111 and
7. Pressure Pa Figure 2 8 FTIR spectra obtained with Cu ZnO A1 03 10 60 30 left and adsorption isotherms of CO obtained with Cu ZnO Al203 10 60 30 and Cu ZnO Al 03 50 35 15 right determined in the pressure range of 0 100 Pa and at room temperature mum shifting from 2090 to 2086 cm with increasing coverage can be seen This band can be ascribed to the vibration of CO adsorbed on metallic copper 21 The band gets broader and more asymmetric with increasing coverage It decreases instantly when the cell is evacuated There is no band in the spectrum recorded 1 min after evacuation A qualitative adsorption isotherm can be derived from the IR data by integrating the peak areas Fig 2 8 right The adsorption isotherm of the industrial ternary catalyst measured by mi crocalorimetry is included for comparison The shape of the qualitative adsorption isotherm of the ternary catalyst 10 60 30 fits well to the adsorption isotherm of the industrial catalyst mea sured by microcalorimetry While the absolute coverages umolco amp at Will surely be different for the two catalysts the relative coverages umolco pMOl Cx surface appear to follow the same adsorption isotherm This confirms that the copper surfaces of the two ternary catalysts are in 24 2 The combined application of microcalorimetry TPD and FTIR spectroscopy the same state after the hydrogen reduction pretreatment 2 4 Conclusions The microcalorimetric results
8. The gas supply system consists of two gas lines and a vacuum line CO is used as probe molecule in the adsorption experiments Nitrogen is used as inert gas e g for the calibration of the volume of the dosing section or the measurement of the volume of the measuring cells Nitrogen is preferable to helium as nitrogen is more similar to CO in terms of thermal con ductivity and viscosity The vacuum line can be used to evacuate the gas lines or the dosing section 101 linear motion feedthrough dummy NSS 99 linear motion NN feedthrough SS double sided CF flange wi SS II tee piece steel rod lt sample and reference receptacle pyrex capsule with sample Figure 6 13 A schematic diagram of the microcalorimetric measuring cells Both gas lines consist of a gas bottle a pressure reducing valve and a shut off valve The purity of the used gases is 99 997 and 99 99990 for CO and nitrogen respectively The 101 CO bottle is stored in a vented gas cupboard equipped with an additional pressure reducing valve and a solenoid shut off valve for safety reasons fig 6 14 The 501 nitrogen bottle is fixed to the wall directly behind the set up fig 6 15 The pressure of the two used gases should be set to low values i e about 0 bar excess or bar absolute to allow the dosing of small gas quantities The vacuum line is connected to a rotary vane pump Fig 6 16 shows a pa
9. in the region 800 6000 cm were recorded The pressure of CO was varied stepwise between 0 100 Pa 0 0 5 1 2 5 5 10 20 40 80 100 Pa and evacuation in order to investigate the adsorption of CO as a function of coverage Details about the experimental conditions and the set up are given elsewhere 17 3 3 Results The heat of adsorption was measured calorimetrically using only the samples with a high cop per content Fig 3 1 shows the results left differential heat of adsorption right adsorption isotherm obtained with the samples CZA1 CZ1 and CA1 The adsorption of CO is strongly influenced by the presence of ZnO The initial heat of adsorption of CO on zinc containing catalysts CZA1 CZ1 is about 10 kJ mol lower than on the ZnO free sample BA1 The heat of adsorption monotonously decreases in all cases with increasing coverage In the observed pressure range the final heat of adsorption is significantly lower in the case of the ZnO free sample The absolute coverage of the samples for a given equilibrium pressure e g 60 Pa is higher for ZnO containing catalysts CZA1 the sample most active for the production of methanol see table 3 1 shows the highest coverage fig 3 1 right A summary of the calori metric data is given in table 3 2 Differences between the data published here and in ref 19 3 Part I The reduced catalyst 31 are due to a reprocessing of the raw data The results in ref 19 are calculated u
10. 1 e if carbon dioxide stays in the gas phase the partial pressure calculated for CO would be overestimated while the amount adsorbed CO would be calculated too low This would result in a flattening of the shape of the adsorption isotherm Such a flattening is not found in the calorimetric results after methanol synthesis A comparison 80 5 Part III The state of the catalyst after methanol synthesis of the adsorption isotherms obtained after methanol synthesis and hydrogen reduction leads to the conclusion that the state of the copper is similar after both pretreatments The observed differences can be correlated with changes of the free copper surface area The measured heats of adsorption are in direct contradiction to this conclusion This contradiction strongly indicates the occurrence of additional processes such as the formation of carbon dioxide and carbonates The assumption that a certain pressure of CO or a certain surface coverage with CO has to be exceeded to start the reaction results in the conclusion that the initial heats of adsorption are measured correctly The initial heats of adsorption are in all cases comparable to the initial heats measured after hydrogen reduction This also confirms the conclusion drawn from the TPD data the FTIR data and the adsorption isotherms that the copper content of the catalysts is completely reduced after methanol synthesis and in a state that is comparable to the state after hydrogen reduct
11. 30 1995 99 4 K Klier Adv Catal 31 1982 243 5 I Nakamura T Fujitani T Uchijima and J Nakamura J Vac Sci Technol A 14 1996 1464 6 I Nakamura T Fujitani and T Uchijima and J Nakamura Surf Sci 400 1998 387 7 Y Kanai T Watanabe T Fujitani M Saito J Nakamura and T Uchijima Energy Con vers Mgmt 36 1995 649 G C Chinchen K C Waugh and D A Whan Appl Catal 25 1986 101 T S Askgaard J K N rskov C V Ovesen and P Stoltze J Catal 156 1995 229 2 8 9 10 11 Part I The reduced catalyst chapter 3 M Kurtz H Wilmer T Genger O Hinrichsen and M Muhler Catal Lett 86 2003 77 12 Part II The state of the catalyst after pretreatment in CO chapter 4 13 B Bems M Schur A Dassenoy H Junkes D Herein and R Schl gl Chem Eur J 9 2003 2039 14 O Hinrichsen T Genger and M Muhler Chem Eng Technol 11 2000 956 15 H Bielawa M Kurtz T Genger and O Hinrichsen Ind Eng Chem Res 40 2001 2793 16 B E Spiewak and J A Dumesic Thermochim Acta 290 1996 43 17 The combined application of microcalorimetry TPD and FTIR spectroscopy chapter 2 18 H G Karge and W Nie en Catal Today 8 1991 451 19 S G Neophytides A J Marchi and G F Froment Appl Catal A 86 1992 45 20 J F Edwards and G L Schrader J Phys Chem 89 1985 782 21 J Shen J M
12. 6 shows the results of the CO adsorption experiments on the binary Cu ZnO catalyst 50 50 The sample was reduced by hydrogen as described above Integration of the mass cat D E o D 20 e first adsorption amp e first adsorption O second adsorption gt O second adsorption 0 0 20 40 60 80 0 15 30 45 60 Coverage umol g Pressure Pa Figure 2 6 Differential heat of adsorption and adsorption isotherms of CO on Cu ZnO 50 50 determined at 300 K spectrometry traces confirmed the complete reduction of the copper content The differential heat of adsorption is found to decrease almost linearly from 68 kJ mol at very low coverage to 57 kJ mol at a coverage of about 80 umolco Z ar followed by a steep decrease to nearly 45 kJ mol The saturation coverage of the sample with CO was about 90 umolco geat at a pressure of 60 Pa The low initial value of the heat of adsorption confirms the absence of Cu l species and the 20 2 The combined application of microcalorimetry TPD and FTIR spectroscopy complete reduction of copper 14 There are only negligible differences between the first adsorption and the second adsorption after evacuating at room temperature indicating that the adsorption process is fully reversible even at room temperature This can be seen in the plot of the differential heat as well as in the plot of the adsorption isotherms Fig 2 6 The repeated adsorption experiment shows also that th
13. A Dumesic Langmuir 11 1995 2065 23 M Kurtz H Wilmer T Genger O Hinrichsen and M Muhler Catal Lett 86 2003 TI 24 J J F Scholten and J A Konvalinka Trans Faraday Soc 65 1969 2465 25 J Strunk R Naumann d Alnoncourt M Bergmann O Hinrichsen and M Muhler to be published 26 J A Dumesic D F Rudd L M Aparicio J E Rekoske and A A Trevi o ACS Profes sional Reference Book Washington DC 1993 27 F Boccuzzi and A Chiorino J Phys Chem 100 1996 3617 28 K I Hadjiivanov and G N Vayssilov Adv Catal 47 2002 307 29 N Y Tops e and H Tops e Top Catal 8 1999 267 4 Part II The state of the catalyst after pretreatment in CO Abstract The influence of the strong metal support interactions between copper and zinc oxide on the adsorption of carbon monoxide are studied in this contribution Samples containing only cop per and alumina as an inert supporting material are included for comparison The samples are pretreated under strongly reducing conditions prior to the adsorption of carbon monoxide The adsorption of carbon monoxide is investigated in microcalorimetric temperature programmed desorption and infrared spectroscopic experiments A comparison of the results with already reported results obtained after the reduction of the samples with hydrogen shows that the ad sorption of carbon monoxide is strongly influenced by the interactions between copper and zin
14. Bc 117 Exporting A ZOnE cis re ed 118 The logged pressure data wo a na a a a SL Ee BLS Vf ALE HOE 119 Curriculum Vitae Pers nliche Daten Name Geburtsdatum Geburtsort Schulbesuch 1978 1978 1980 1980 1982 1982 1983 1983 1985 1985 1988 1988 1991 24 Mai 1991 Wehrdienst Aug 1991 Sep 1992 Hochschulausbildung 20 Sep 1992 23 Okt 1995 26 Okt 1995 Jan Juli 2001 21 Aug 2001 Dez 2001 Dez 2004 Raoul Naumann d Alnoncourt 10 Juli 1972 G ttingen Grundschule Neu Aubing M nchen Montessorischool Amsterdam Holland Grundschule Utting St dtisches Gymnasium Weilheim Collegium Josephinum Bonn Reuchlin Gymnasium Pforzheim Carl Humann Gymnasium Essen Abitur Zivildienst Alfried Krupp Krankenhaus Essen Immatrikulation an der Universit t Dortmund Vordiplom in Chemie Immatrikulation an der Universit t zu K ln Diplomarbeit zum Thema Untersuchungen zur Wechselwirkung von Wirkstoffen und Polymeren mit L sungsmitteln und Additiven in Formulierungen Diplom in Chemie Dissertation am Lehrstuhl f r Technische Chemie der Ruhr Universit t Bochum Lehrstuhlinhaber Prof Dr M Muhler
15. Clausen J Catal 194 2000 432 5 J Pritchard T Catterick and R K Gupta Surf Sci 53 1975 1 6 S Vollmer G Witte and C W ll Catal Lett 77 2001 97 7 O Dulaurent X Courtois V Perrichon and D Bianchi J Phys Chem B 104 2000 6001 8 E Giamello B Fubini and V Bolis Appl Catal 36 1988 287 9 F Boccuzzi G Ghiotti and A Chiorino Surf Sci 156 1985 933 10 S Bailey G F Froment J W Snoeck and K C Waugh Catal Lett 30 1995 99 11 T Genger O Hinrichsen and M Muhler Catal Lett 59 1999 137 12 H Wilmer and O Hinrichsen Catal Lett 82 2002 117 13 B Bems M Schur A Dassenoy H Junkes D Herein and R Schl gl Chem Eur J 9 2003 2039 14 O Hinrichsen T Genger and M Muhler Chem Eng Technol 11 2000 956 15 H Bielawa M Kurtz T Genger and O Hinrichsen Ind Eng Chem Res 40 2001 2793 16 B E Spiewak and J A Dumesic Thermochim Acta 290 1996 43 17 The combined application of microcalorimetry TPD and FTIR spectroscopy chapter 2 18 H G Karge and W Nie en Catal Today 8 1991 451 19 R Naumann d Alnoncourt M Kurtz H Wilmer E L ffler V Hagen J Shen and M Muhler J Catal 220 2003 249 20 N Cardona Martinez and J A Dumesic Adv Catal 38 1992 149 21 J C Tracy J Chem Phys 56 1972 2748 46 Bibliography 22 G D Borgard S Molvik P Balaraman T W Root and J
16. Hill R M Watwe B E Spiewak and J A Dumesic J Phys Chem B 103 1999 3923 6 Conclusions Microcalorimetry TPD experiments and FTIR spectroscopy were successfully combined in the investigation of the adsorption of CO on copper catalysts after various pretreatments The results obtained after hydrogen reduction confirm that all samples were pretreated under identi cal conditions All presented data show clearly that applying identical pretreatment conditions leads to reproducible and fully comparable states of the catalyst surface To the best of our knowledge the present contribution shows for the first time results measured by fundamen tally different investigation methods concerning samples both prepared and pretreated under identical conditions All results support a grouping of copper catalysts in three classes as proposed by Hinrichsen and co workers 1 ternary catalysts consisting of Cu ZnO and AlO Cu ZnO Al203 bi nary catalysts consisting of Cu and ZnO Cu ZnO and binary ZnO free catalysts such as Cu supported on Al203 Cu Al203 After a specific pretreatment samples of one class showed similar behavior in the adsorption of CO while all three classes were differently influenced by the pretreatment conditions The interaction of copper catalysts with hydrogen was not sig nificantly influenced by the supporting material 2 3 while a strong influence of the catalyst support on the adsorption of CO was found after all
17. Mgmt 36 1995 649 11 B S Clausen and H Topsge Catal Today 9 1991 189 12 B S Clausen G Steffensen B Fabius J Villadsen R Feidenhans l and H Tops e J Catal 132 1991 524 13 T H Fleisch and R L Mieville J Catal 90 1984 165 14 G C Chinchen P J Denny D G Parker M S Spencer and D A Whan Appl Catal 30 1987 333 15 T S Askgaard J K N rskov C V Ovesen and P Stoltze J Catal 156 1995 229 16 R A Beebe and E L Wildner J Am Chem Soc 56 1934 642 17 J C Tracy J Chem Phys 56 1972 2748 18 P Hollins and J Pritchard Surf Sci 89 1979 486 19 O Dulaurent X Courtois V Perrichon and D Bianchi J Phys Chem B 104 2000 6001 20 S Vollmer G Witte and C W ll Catal Lett 77 2001 97 21 F Boccuzzi G Ghiotti and A Chiorino Surf Sci 156 1985 933 8 Bibliography 22 E Giamello B Fubini and V Bolis Appl Catal 36 1988 287 23 J Pritchard T Catterick and R K Gupta Surf Sci 53 1975 1 24 P Hollins Surf Sci Rep 16 1992 53 25 S J Tauster S C Fung and R L Garten J Am Chem Soc 100 1978 170 26 J D Grunwaldt A M Molenbroek N Y Tops e H Tops e and B S Clausen J Catal 194 2000 452 27 P L Hansen J B Wagner S Helveg J R Rostrup Nielsen B S Clausen and H Topsge Science 295 2002 2053 28 H Wilmer and O Hinrichsen Catal Lett 82 2002 117 2
18. Pa After cooling the calorimeter to 303 K and the volumetric dosing section to 313 K overnight the capsule is broken via a linear motion feedthrough The pressure of the released helium is reduced to about 80 Pa After reaching thermal equilibrium and a steady baseline of the heat flow signal the adsorption measurement is started The period of time between the breaking of the capsule and the start of the measurement typically amounts to less than 60 min Fig 2 3 shows a schematic diagram of the specially designed pretreatment reactor used for the sample pretreatment It consists of a glass lined stainless steel U tube a pyrex metal 14 2 The combined application of microcalorimetry TPD and FTIR spectroscopy pyrex tube gas inlet exhaust heating element Figure 2 3 U tube reactor used for the pretreatment of microcalorimetric samples joint GA 050P S Caburn MDC with a pyrex NMR tube welded to it a manometer 1000 to 2000 hPa Millipore and a four way valve 4UWE Valco VICI The complete re actor is metal tightened Samples are placed into the U tube and kept in place by a quartz wool plug The reactor can be heated by a vertically movable heating element The thermocouple which controls the heating element is fixed onto the outside of the reactor at the position of the sample The reactor can be connected by Cajon VCR connectors with the flow set up used for CO TPD experiments see section 2 2 2 The pressur
19. a a Tae hee ee ee te B AA WUSCUSSIOM 2 24 A Go ec Sato Ga ie Br theta he pete Bo bee hp ete BS BGS AD COMCIUSIONS Xela ya u Weve A EZ hawk B Dea has Bday Beet Soy 11 11 15 16 17 18 18 20 22 24 27 27 28 30 35 43 IV Contents 5 Part III The state of the catalyst after methanol synthesis 69 3 1 gt Introduction auc eeno ta Redd Re Mid ha we ee a anna Binnen 69 5 2 VExXPerimental 2 SA Ge etd Bae ead Sete RA 70 Dee Resulta ra Sra e ad Seo al oaks Bo a tg Ml gt AM eA ol GA a ca Mt 12 DA DISCUSSION tr ii a Va a eine re AE 77 5 3 O O fe ert ta ed haps tat gest teh dar arate a anak da Ariat Pas te ath dat 81 6 Conclusions 85 6 1 Introduction 4 44 Skee Yate u ee BPS Be PES BS Bhd 89 6 2 Experimental a sia A det Mee Ae a ae CG ae Gi gh At a te 90 6 2 1 Connectors and flanges vun war aa ee Sc BW AS a SLE 90 6 2 1 1 Swagelok connectors and tube fittings 90 6 2 1 2 The Cajon VCR connection 22 04 Dede 91 6 2 L 3 Th CF Blancs id 93 6 2 1 4 The KF Flange ana aa Oe wh ie SR OS 94 6 2 2 The adsorption microcalorimetry set up 94 6 2 2 1 The microcalorimeter 24 2 232er wee a a 95 6 2 2 2 The volumetric dosing section 96 6 2 2 3 Controlling the temperature of the set up 98 6 2 2 4 The measuring cells 2 2 2 22 22 le ee eS 99 6 2 2 5 The gas supply system E A eh ee 100 6 2 2 6 The vacuum system lt lt a eS cae BETS Ewe 102 6 2 2
20. a higher reduction potential surface and bulk alloying leads to the formation of brass Hansen et al 27 gave experimental evidence of dynamic shape changes of copper nanocrys tals supported on ZnO The changes were directly observed by in situ transmission electron microscopy under high pressures The effects were caused by changes of the reduction po tential of the surrounding gas phase and were fully reversible The authors concluded that the processes were due to adsorbate induced changes of the Cu ZnO interfacial energies They as sumed that oxygen vacancies in the ZnO play a role in the observed processes Only negligible shape changes were found for copper nanocrystals supported on SiO Wagner et al 29 investigated the SMSI between copper and ZnO by applying in situ electron energy loss spectroscopy They reported that the support induced a tensile strain in the Cu nan oclusters The degree of this strain was dependent on the reduction potential of the surrounding gas phase They found no indication for strained Cu nanoclusters in Cu SiO samples Hinrichsen and co workers 28 30 investigated the interaction of hydrogen and nitrous oxide with Cu Al203 Cu ZnO and Cu ZnO A1 O3 samples after different pretreatments The applied pretreatment conditions included reduction in hydrogen methanol synthesis and a strongly reducing pretreatment in CO The authors found that Cu Al O3 is hardly influenced by the pretreatment while dynamical chan
21. calorimetrically using the samples with a high copper content CA1 CZ1 and CZA1 Fig 4 1 4 2 and 4 3 show the results left differential heat of adsorption right adsorption isotherm respectively The results of the experiments after hydrogen reduction presented in 8 are included for comparison The adsorption of CO is strongly influenced by the pretreatment Significant differences are found for catalysts with CZA1 CZ1 and without ZnO CA1 Fractional coverage 0 00 0 08 0 16 0 24 0 32 cat o X Q kJ mol Coverage umollg Fractional coverage CO pretr 1 adsorption 4 CO pretr 2 adsorption t reduction by hydrogen CO pretr 1 adsorption CO pretr 2 adsorption reduction by hydrogen 0 15 30 45 60 0 20 40 60 80 Coverage mol g Pressure Pa Figure 4 1 Differential heat of adsorption and adsorption isotherms of CO on CAl at 303 K after CO pretreatment The results obtained after hydrogen reduction are included for com parison The sample was evacuated overnight between the first and the second adsorption experiment In the case of the ZnO free sample CA1 the heat of adsorption decreases monotonically with increasing coverage from 71 33 kJ mol first adsorption with a plateau around 55 kJ mol The range of the heat of adsorption is comparable to the results after hydrogen reduction only that the initial and final in the observed coverage range heat of adsorption are slightly lowe
22. careful handling Even light scratches will cause a poor seal Several types of gaskets are available made of stainless steel of silver plated stainless steel 92 Figure 6 3 Installation of a Swagelok tube fitting 3 and of copper Harder gaskets give lower leakage rates while the use softer gaskets avoids damaging of the sealing faces during assembly Therefore copper gaskets should be used for frequently disassembled connections Sidadcad Female mit Gand retainer gasket land Male miut i i also ee F y we OR Ferree mf Gand Gasket Body Figure 6 4 The Cajon VCR connection 4 The installation instruction for a Cajon VCR connection is shown in fig 6 5 First place a suitable gasket between the glands and tighten the two nuts finger tight The gasket may not be touched without gloves otherwise it will be contaminated by skin fat and cause a poor seal 93 Then use two wrenches to tighten the connection 45 using stainless steel gaskets 90 using copper gaskets Figure 6 5 The installation of a Cajon VCR connection 4 6 2 1 3 The CF Flange CF flanges are used at frequently disassembled positions especially if tubes of a large diameter are connected To connect two CF flanges place a copper gasket against the knife edge seal of one of the flanges preferably the flange most likely to hold and support the gasket from falling With the gasket in place arrange the s
23. catalyst after pretreatment in CO 0 1 CZA2 2063 cm 100 Pa CO en Extinction Extinction 1228 cm 1246 cm 2000 1875 1750 1275 1200 1125 Wavenumbers cm Wavenumbers cm Figure 4 5 FTIR spectra obtained with CZA2 after pretreatment in CO in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for pressures of 0 and 100 Pa of CO and after evacuation band as a function of the pretreatment are summarized in table 4 3 v is the wavenumber of the peak maximum A is the shift of the peak maximum relative to the maximum after hydrogen reduction The given peak areas are normalized to the area of the peak after hydrogen reduction respectively 4 4 Discussion Wilmer et al 6 4 investigated the interaction of hydrogen with the catalyst systems Cu Al 03 and Cu ZnO A1 O3 after hydrogen reduction and CO pretreatment Both samples were influ enced by the pretreatment but the changes in the hydrogen TPD profiles of Cu Al03 were small compared to those in the profiles of Cu ZnO Al 03 After the CO pretreatment of the Cu ZnO A1 03 sample an additional peak maximum was found and the free copper surface area was drastically decreased while its size was essentially constant in the case of Cu Al203 However
24. coverage experiment c 0 34 experimental simulated 0 154 experimental simulated 0 2 0 10 Effluent mole fraction Effluent mole fraction 0 1 Temperature K Temperature K Figure 3 9 Simulation of the TPD peaks obtained with CZA1 and CA1 3 Part I The reduced catalyst 4 Table 3 4 Simulated data Sample Peak AH AH Aads Ades kJ mol kJ mol s s7 CZA1 55 6 56 4 755 10 57 3 58 4 681 10 CAI 71 7 65 17 111 10 73 6 73 1 12 10 75 3 79 61 2 10 C D E 58 9 61 7 377 10 c d e f 75 9 80 99 2 10 ov ot a CY a ot OT AH is estimated from the peak maximum following the method of Konvalinka and ads Scholten AH is the integral molar heat of the adsorption of CO derived from the mi crocalorimetric data 0 84 molco gca1 0 62 umolco gcar 0 33 umolcolZcat 0 42 umolc0 8ca1 0 27 umoloo geat f 0 11 umolcolgcat 2 0 4 umolco geat In summary the CO TPD experiments confirm the results by microcalorimetry The experimen tal heats of adsorption could be used to simulate the CO TPD experiments in good agreement The simulation also reflects the differences in the entropies of adsorption in the case of CZA1 and CA1 The higher fractional coverage of CA1 at a given equilibrium pressure of CO can also be found in the TPD data Pritchard et al 5 presented a detailed IR study of the adsorption of CO on copper single crystal surfaces a
25. decreased by 45 65 The FTIR results of the experiments using CZA2 42 mg 2 cm and CA2 31 mg 2 cm are shown in fig 4 5 and 4 6 In the case of CZA2 mainly one broad and to lower wavenumbers asymmetric band with a peak maximum at 2063 cm is observed The band becomes broader and more asymmetric with increasing coverage Additionally weak bands can be observed below 2000 cm at about 1690 1246 and 1228cm No bands appear above 2100 cm Due to intense noise no information can be gained in the range 1650 1300cm and above 4 Part II The state of the catalyst after pretreatment in CO 55 Table 4 2 CO TPD data obtained with CZA1 and CA1 Sample Peak Tisi step Tmar FWHM coverage Tmar FWHM coverage K K K umol g K K umol g CZA1 B 275 333 85 78 315 112 145 C 300 338 62 53 346 87 102 D 325 350 44 26 357 61 62 E 350 355 44 11 367 49 33 CA1 b 275 340 100 56 336 98 58 c 300 348 76 41 345 76 42 d 325 354 59 27 355 59 27 e 350 365 50 12 364 46 11 f 375 373 44 3 369 40 2 a after CO pretreatment after hydrogen reduction 2200cm It is found that the main band at 2065 cm decreases instantly when the cell is evacuated while the weak bands change only little in intensity The adsorption of CO induces 1 Shifts to lower extinction a baseline shift to higher extinction in the range above 1700 cm are observed at lower wavenumbers The shifts are only partially reversible af
26. dioxide Carbon monoxide is part of the water gas shift reaction which is important in order to remove water formed by the synthesis of methanol from 70 5 Part III The state of the catalyst after methanol synthesis the copper surface thus keeping copper in a reduced state In this contribution the adsorption of carbon monoxide on copper catalysts is studied by mi crocalorimetry temperature programmed desorption TPD experiments and Fourier transform infrared FTIR spectroscopy The samples are pretreated by a reduction in hydrogen followed by the synthesis of methanol The results in ref 10 show that the applied techniques and the used equipment allow to study the samples in reproducible states with the different investiga tion methods Results obtained with catalysts of different copper content are directly compared based on the classification of copper catalysts postulated by Hinrichsen and coworkers 11 5 2 Experimental The investigated samples are binary and ternary catalysts containing copper zinc oxide and alumina The samples are identical to those in 10 12 The methods of preparation and char acterization of the samples are described in detail elsewhere 13 14 15 Table 5 1 summarizes the main characteristics of the samples Table 5 1 Characterization and catalytic data Catalyst CZAI CZA2 CZI CAI CA2 BET surface area m Zcat 73 64 51 124 Cu content wt CuO 47 7 68 76 20 Specific amount of Cu
27. ex C2 dy 5 H NR H Y Tut Q H O P lt NR 2 oO O e Ox gt 5 x CO He Ey yy de oy e SA E l O ex Cx A O gt gt CO He 3 pa H He xD ex exhaust He E 7 Yh ty Figure 2 4 Flow scheme of the set up used for CO TPD experiments and sample pretreatment reactor is a glass lined stainless steel U tube reactor of 3 8 mm inner diameter with two Cajon VCR connectors The set up can be operated at pressures of up to 6 MPa and reactor temper atures of up to 873 K A personal computer equipped with the software package LabView is used to control the set up All tubings are made of glass lined stainless steel to avoid adsorp tion or reaction of any gas components on the inner tube walls Tubings between the reactor and the mass spectrometer are heated to a temperature of 366 K to prevent the condensation of products such as water or methanol The gas lines of the gas supply unit are all of similar design They consist of a gas cylinder a pressure reducing valve and a mass flow controller 0 100 Ncm min A pneumatically actuated shut off valve is installed before and after each 16 2 The combined application of microcalorimetry TPD and FTIR spectroscopy mass flow controller The gas lines are connected with four way valves 4UWE Valco VICI Only the gas of one line can flow through the reactor at the same time The employed gases are helium a 4 CO2 He mixture a 10 CO He mixture hydrogen a 1 N20 He mix ture
28. identical receptacles for the sample and 12 2 The combined application of microcalorimetry TPD and FTIR spectroscopy La T heated box T e co TEENS heating element Lf N cells in N calorimetric INN block x calorimeter gt Figure 2 1 Flow scheme of the adsorption microcalorimetry set up the reference sample are connected to the two double sided flanges A bellows sealed linear motion feedthrough E LMD 133 2 Caburn MDC and a dummy resembling the form of the half expanded feedthrough are connected to the other side of the flanges The linear motion feedthrough can be used to crush the pyrex capsules in the sample receptacle via a steel rod The standard flange is used to connect the cells to the volumetric dosing section All parts of the cells are made of stainless steel and are UHV ultrahigh vacuum tight The volumetric dosing section is made of four completely metal tightened bellows valves and a Baratron capacity manometer range 0 100 Pa The complete dosing section is placed in a heated box The four valves are connected to the microcalorimetric cells a CO gas line a helium gas line and a turbomolecular drag pump TMU 071 P Pfeiffer The turbomolecular pump is equipped with a membrane pump as backing pump and a compact full range gauge PKR 251 Pfeiffer to monitor the pressure The helium gas line can be switched to a mem brane pump which is used as a rough pump fo
29. in a stainless steel U tube reactor connected to a flow set up and the desorption of CO into a stream of helium was monitored by on line mass spectrometry CO is a often used probe molecule in Fourier transform infrared FTIR spectroscopy studies CO adsorbs onto copper surfaces at low coverages non dissociatively and linearly in on top position The v C O stretching vibration of adsorbed CO is sensitive to the strength of the bond to the adsorbent and thus sensitive to the state of reduction of copper In transmission IR 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 11 studies relative coverages can be determined according to the Lambert Beer law thus allowing to determine adsorption isotherms In this contribution the adsorption of CO was investigated using a modified transmission IR cell designed by Karge et al 8 The pressure of CO was varied stepwise between 0 100 Pa in order to investigate the adsorption of CO as a function of coverage The interaction of Cu catalysts with hydrogen has been studied extensively by the Bochum group 12 13 2 The catalysts had been investigated in different states directly after hy drogen reduction after a period of 12 h of methanol synthesis and after a pretreatment of 64h in CO He In order to compare the results of the studies investigating the interaction of the copper catalysts with hydrogen and with carbon monoxide the same reduction procedure was applied in t
30. irreversible A possible explanation for a reversible reaction of carbon monoxide with atomic oxygen is given in fig 4 9 If the presence of Zn and O species on the copper surface stabilizes a state that can CO 1 20 2 adsorption site on the copper surface 9 2 9 O Zn Zn and oxygen species on the Cu surface Potential energy I kJ mol Figure 4 9 The potential energy diagram shows the reaction of CO and oxygen on a copper surface The presence of Zn species on the copper surface could inhibit the desorption of carbon dioxide thus leading to the reverse reaction including the desorption of CO be formally described as carbon dioxide bound to copper and zinc the reverse reaction may 4 Part II The state of the catalyst after pretreatment in CO 63 be favorable to the desorption of carbon dioxide This can be rationalized by the fact that a desorption of CO from the described state would result in the reduction of the Zn atom If no surface reaction of CO and atomic oxygen occurs an unknown exothermic process must be as sumed in order to interpret the calorimetric results The process has to be reversible and should be related to ZnO even though not necessarily to the Zn and oxygen species on the copper surface The process does not occur after hydrogen reduction and is stronger in the absence of alumina There are indications in the FTIR results for an influence of the CO pretreatment on the ZnO support The single b
31. methanol synthesis CZA1 CA1 Effluent mole fraction 75 150 225 300 375 450 Temperature K Effluent mole fraction Effluent mole fraction 150 225 300 375 450 Temperature K Temperature K Figure 5 5 CO TPD spectra obtained with CZA1 and CA1 fully covered A a and partially covered B E b f with CO after methanol synthesis Tist step 275 K B b 300K C c 325 K D d 350 K E e and 375 K f width at half maximum is given in the column FWHM The initial coverages were calculated by integrating the mass spectrometry traces over the complete range of the desorption peak The results obtained with CA1 after the two different pretreatment are essentially the same In the case of CZA1 only small differences in Tmar and FWHM were observed while the initial coverages of corresponding experiments are significantly decreased after methanol synthesis The FTIR results of the experiments using CZA2 42 mg 2 cm and CA2 31 mg 2 cm are shown in fig 5 6 and 5 7 In the case of CZA2 mainly one broad and to lower wavenum 1 was observed The band became bers asymmetric band with a peak maximum at 2084 cm broader and more asymmetric with increasing coverage Additionally weak bands were ob served below 2000 cm at about 1685 1228 1246 and 1164cm No bands appeared above 2100 cm Due to intense noise no information can be gained in the range 1650 1300 cm and above 2200 cm It is found
32. or differently pretreated samples are compared A complete description of the set up and all experimental conditions is given in the second chapter Analogous to the work in ref 28 30 the catalyst samples were studied after three different pretreatments after a complete reduction of the copper content of the sample in hydro gen after a 12 h period of methanol synthesis and after a 50h pretreatment in strongly reducing CO gas The results obtained with all samples after a specific pretreatment are presented in one chapter respectively The final chapter summarizes the conclusions of each chapter Bibliography 1 J B Hansen in Handbook of Heterogenous Catalysis VCH Verlagsgesellschaft Wein heim 4 1997 1856 2 K Kochloefl in Handbook of Heterogenous Catalysis VCH Verlagsgesellschaft Weinheim 4 1997 1831 3 K Klier Adv Catal 31 1982 243 4 Y Okamoto K Fukino T Imanaka and S Teranishi J Phys Chem 87 1983 3747 5 G C Chinchen K C Waugh and D A Whan Appl Catal 25 1986 101 6 J L Robbins E Iglesia C P Kelkar and B DeRites Catal Lett 10 1991 1 7 J C Frost Nature 334 1988 577 8 I Nakamura T Fujitani T Uchijima and J Nakamura J Vac Sci Technol A 14 1996 1464 9 I Nakamura T Fujitani and T Uchijima und J Nakamura Surf Sci 400 1998 387 10 Y Kanai T Watanabe T Fujitani M Saito J Nakamura and T Uchijima Energy Con vers
33. pdf Sept 22 2004 5 http www caburn co uk resources downloads pdfs sec1 1 1 pdf Sept 22 2004 6 http www caburn co uk resources downloads pdfs sec1 2 1 1 pdf Sept 22 2004 7 B E Spiewak and J A Dumesic Thermochim Acta 290 1996 43 List of Tables 2 1 3 1 3 2 3 3 3 4 4 1 4 2 4 3 5 1 5 2 CO TPD data obtained with Cu ZnO AlL20O3 50 35 15 21 Characterization and catalytic data o o e 29 Calorimetric results o vo s soste ore AA AA 31 CO TPD data obtained with CZAlandCAl 2 2 2 2 22er 34 Simulaled data sat a a ea A a a ok 41 Characterization and catalytic data 2 o o 49 CO TPD data obtained with CZAl and CAL o o ooo o 55 Influence of the pretreatment on the main IR band 60 Characterization and catalytic data o o 70 CO TPD data obtained with CZAl and CAl 77 List of Figures ZA 22 2 3 2 4 2 5 2 6 2 7 2 8 3 1 3 2 3 3 3 4 Flow scheme of the adsorption microcalorimetry set up 12 Schematic drawing of the microcalorimetric cells 13 U tube reactor used for the pretreatment of microcalorimetric samples 14 Flow scheme of the set up used for CO TPD experiments and sample pretreat MENL a ee e e Be WSS we a y ada y 15 Flow scheme of the transmission IR cell 17 Differentia
34. powdered samples in an agate mortar sieving the powder through a piece of cotton fabric directly into the pressing tool and then pressing the powder for 60s at a pressure of less than 1 MPa The resulting wafer is cut to fit into the sample holder and is transferred into the IR cell For the pretreatment the wafer is lifted into the pretreatment section by the magnetic manipulator It is then pretreated under conditions resembling those of the microcalorimetric and CO TPD experiments After the pretreatment the wafer is lowered into the IR beam and the IR cell is evacuated 2 2 4 Investigated samples The investigated samples are binary and ternary catalysts containing copper zinc oxide and alumina with the following molar compositions Cu ZnO 50 50 Cu ZnO A1 03 50 35 15 and Cu ZnO Al203 10 60 30 The samples were prepared by coprecipitation using solutions of nitrates and sodium carbon 18 2 The combined application of microcalorimetry TPD and FTIR spectroscopy ate and calcined afterwards In order to obtain reproducible results the preparation follows a standardized procedure under controlled conditions e g pH temperature reaction and age ing time The samples were characterized by physisorption measurements BET surface area temperature programmed reduction copper content N2O reactive frontal chromatography re duced copper surface area tests of activity for the methanol synthesis reaction and X ray diffract
35. pretr 1 adsorption A CO pretr 2 adsorption reduction by hydrogen reduction by hydrogen 0 20 40 60 80 0 20 40 60 80 Coverage ymollg Pressure Pa Figure 4 3 Differential heat of adsorption and adsorption isotherms of CO on CZA1 at 303 K after CO pretreatment The results obtained after hydrogen reduction are included for com parison The sample was evacuated overnight between the first and the second adsorption experiment 84 umol g at While the adsorption isotherms shown in fig 4 3 confirm the reversibility of the adsorption processes the differences in the heat of adsorption between the first and the second adsorption experiment indicate non reversible changes of the sample The CO TPD experiments were performed with CA1 and CZAl Fig 4 4 shows the results of the TPD experiments The desorption from the fully covered copper surface of CA1 ex periment a results in an intense peak at 110K and a broad signal in the temperature range 200 400 K with a maximum at 283 K and a shoulder at about 345 K The TPD profile is similar to the profile obtained after hydrogen reduction 8 The TPD profile of the fully covered cop per surface of CZA1 experiment A is significantly different to the corresponding profile after hydrogen reduction 8 There is an intense peak at 105 K and a broad signal in the temperature range 200 400 K with maxima at 204 and 275 K and a shoulder at about 330 K The amount of desorbing
36. prove the efficiency of the applied technique for the sample trans fer into the calorimeter without contamination by air The results also show that the leakage rate of the microcalorimetry set up is sufficiently low to investigate the air sensitive copper surfaces of the reduced samples The CO TPD data confirms the microcalorimetry results Similar surface coverages were ob tained by both investigation methods The decrease of the heat of adsorption with increasing coverage measured by microcalorimetry is also observed in the TPD experiments The results obtained by FTIR spectroscopy support the conclusions of the microcalorimetric experiments copper is completely reduced to its zero valent state after hydrogen pretreatment even under the non ideal flow conditions of the IR cell and the adsorption of CO is fully reversible at room temperature The adsorption isotherms measured by the two different in vestigation methods fit to each other The results prove that the FTIR set up is suitable to investigate air sensitive copper samples without contamination by air before or during the CO adsorption All results are in good agreement The three different investigation methods were successfully combined by applying identical pretreatment conditions which is further demonstrated by the modelling of the isotherm and the CO TPD data in 17 Publications covering other model catalysts and the additional pretreatments described above are in preparation
37. rate The coverage dependence of the heat of adsorption of CO measured by microcalorimetry can also be found in the CO TPD experiments The peak maximum shifts with decreasing initial coverage to higher temperatures indicating an increase of the heat of adsorption with decreas ing coverage The higher fractional coverage for a given equilibrium pressure of CA1 in com parison to CZA1 can also be found in the TPD results The fractional coverage measured for the experiments b d is significantly higher than for the corresponding experiments B D The integral heat of adsorption can be roughly estimated from the peak maximum following the method of Konvalinka and Scholten 24 Values for the Arrhenius parameters of adsorp tion A as and desorption Ages are needed to apply this method Tab 3 4 shows the values obtained by this method and the integral heats of adsorption for the corresponding coverages calculated from the microcalorimetric data A detailed description of the calculations is given elsewhere 25 In the case of CZA1 standard values were used for Aaqs 10 s and Ages 101 s71 as given by Dumesic et al 26 The estimation fits well with the experimental val 40 3 Part I The reduced catalyst ues To estimate the integral heat of adsorption on CA1 Ages was kept at 10 s but Agas was lowered to 10 s in order to take into account the lowered mobility of the adsorbed CO species found by microcalorimetry The esti
38. tape that can be easily detached for disconnecting the measuring cells With exception of the calorimetric block each heating unit is controlled by a programmable temperature controller Eurotherm 2416 The calorimeter includes a controller Setaram CS32 that controls the temperature of the calorimetric block During an adsorption experiment the temperature of the calorimetric block typically is set to 303 K while the three additional heating elements are set to a slightly higher temperature 99 heatflux transducers ns reference y sample Ces thermostated calorimetric block Figure 6 11 A schematic diagram of a Tian Calvet sensor 313 K A temperature difference of 10 K between the calorimetric block and the rest of the set up has been found suitable to avoid measurement artifacts when the gas dose expands into the calorimetric cells Prior to each adsorption experiment the complete set up is degassed During degassing all temperatures typically are set to 418 K The temperature controllers of the calorimeter the heating element fitted into the calorimeter and the heated box may be set directly from room temperature to 418 K as these heating elements rise the temperature slowly without any overshooting The temperature controller of the heating tape attached to the connection between the dosing section and the calorimetric cell may only be set stepwise and carefully to higher temperatures The heating tape
39. that the main band at 2065 cm decreased instantly when the cell was evacuated while the weak bands changed only little in intensity The baseline 5 Part III The state of the catalyst after methanol synthesis 71 Table 5 2 CO TPD data obtained with CZAl and CAl Sample Peak Tis step Tmar FWHM coverage Tmar FWHM coverage K K K umol g K K umol g CZA1 B 275 331 100 108 315 112 145 C 300 345 71 78 346 87 102 D 325 355 58 50 357 61 62 E 350 365 46 24 367 49 33 CA1 b 275 338 104 58 336 98 58 c 300 345 74 39 345 76 42 d 325 354 57 25 355 59 27 e 350 365 47 11 364 46 11 f 368 373 43 2 369 40 2 after methanol synthesis after hydrogen reduction was shifted to lower extinction in the region below 1350cm The spectrum of CA2 shows only one broad and asymmetric band with a peak maximum at 2091 cm The band became also broader and more asymmetric with increasing coverage No additional bands were found The baseline was shifted to higher extinction with increasing coverage in the region below 1 1350cm Due to intense noise no information can be gained in the range 1650 1350 cm and above 2200 cm 5 4 Discussion The TPD experiments and the FTIR experiments indicate that the copper content of the samples CZA1 and CZA2 and CAl and CA2 are after methanol synthesis in nearly the same state as after hydrogen reduction 10 while the calorimetric results after these two pretreat
40. the pressure but the 0 10 V DC signal generated by the Baratron pressure gauge connected to the Voltmeter vide supra The data is stored in a two column ASCII file The stored signal can be easily converted to a pressure value after importing the ASCII file into an Origin or Excel file Depending on the range of the monitored Baratron pressure gauge 10 V equals 100 00 Pa or 100 00 kPa A power failure of the computer that is used to run the LabView programs might have severe 111 i VENT PULS EDT ZEN Fa Y Figure 6 24 The user interface of the Vent Puls Edit program consequences During the boot process of the computer random signals are sent to the parallel port If the box controlling the solenoid valves is attached to the computer during booting the valves of the dosing section will be actuated randomly To prevent this a magnetic safety switch is inserted in the power supply of the computer After a power failure the valve controller is disconnected from the computer the clearance button of the safety switch is pressed and the computer is switched on Only after starting the Vent Puls program the valve controller is connected to the computer 6 2 3 3 Volume calibration As the amount of adsorbed species is volumetrically measured accurate values of the volumes involved is needed Two volumes are of interest the volume of the dosing section and the volume of the measuring cells The volume of the dosing section i
41. there are indications that the heat of adsorption of CO on copper is lowered after the CO pretreatment and that the measured increase of the heat of adsorption may be due to the adsorption on oxygen vacancies in the ZnO The presence of oxygen vacancies after a strongly reducing pretreatment is also mentioned in liter ature 2 3 5 Further experiments are needed for an interpretation of the microcalorimetric results 4 5 Conclusions The results obtained using CAl and CA2 confirm the absence of SMSI between copper and alumina The differences found in the comparison of the results after hydrogen reduction and after CO pretreatment are only small compared to those in the case of ZnO containing samples The observed differences can be explained without considering an influence of the support All experiments using ZnO containing samples show that the adsorption of CO on copper is strongly influenced by the SMSI between copper and ZnO The TPD and FTIR results are in 4 Part II The state of the catalyst after pretreatment in CO 65 qualitative agreement with theoretical studies in literature The microcalorimetric experiments are not fully understood but there are indications for a lowered heat of adsorption of CO on copper The measured increase of the heat of adsorption may be due to CO strongly bound to oxygen vacancies in the ZnO In summary the results obtained after CO pretreatment in the present contribution together with those mea
42. to eq 6 2 l qu ai 6 2 The differential heat of adsorption can be measured by giving small amounts An as of ad sorptive subsequently to the adsorbent Depending on the experimental conditions the heat Q equals a change in the internal energy U or a change in the enthalpy H If the adsorption exper iment is carried out in a twin type calorimeter e g a Tian Calvet calorimeter the measured heat Q equals the change in the enthalpy H as the work of expansion is compensated qe Ane 6 3 A comprehensive description of the thermodynamics of adsorption can be found in ref 1 Tian Calvet microcalorimeters are differential scanning calorimeters DSC A DSC consists of two thermally decoupled measuring cells the sample cell and the reference cell The measured signal is a differential signal measured between the two cells Thus all external influences on the measurement are compensated as they should have the same effect on the sample and 90 reference cell assumed the cells are completely identical Tian Calvet microcalorimeters can be further classified as heat flux DSCs In a heat flux DSC a defined exchange of the heat to be measured with the environment takes place via a well defined heat conduction path with given thermal resistance The primary measurement signal is a temperature difference which determines the intensity of the heat exchange The temperature difference AT can be converted into the heatflow Q using a cali
43. using the rotary vane pump After degassing the calorimeter is cooled to 303 K while the rest of the set up is cooled to 313 K The fans cooling the solenoid valves are switched off and the leakage rate is measured The leakage rate is derived from the measured increasing rate of the pressure under static vacuum conditions and the volume of the dosing section and the microcalorimetric cells app 100cm After measuring the leakage rate the set up is evacuated for several minutes While only the valve connecting the volumetric dosing section and the measuring cells is opened the pyrex capsule is crushed via the linear motion feedthrough The complete set up is evacuated and then filled with nitrogen at a pressure of about 80 95 Pa When the recorded heatflow signal shows a stable baseline the adsorption experiment itself is started Typically about fifty doses of CO are admitted to the sample while the heatflow and the pressure are recorded When the experiment is finished no further detectable heatflow or an equilibrium pressure of more than 80 Pa of CO the complete set up is degassed overnight under static vacuum conditions at room temperature After degassing the volume of the measuring cells is measured using about twenty doses of nitrogen vide supra After the measurement of the cells volume the sample may be removed or a second adsorption experiment is started to investigate the reversibility of the adsorption processes All valves should be clos
44. 5 1 50 Energy a u Energy a u 5000 4000 3000 2000 1000 5000 4000 3000 2000 1000 Wavenumbers cm Wavenumbers cm Figure 3 6 Single beam spectra obtained at room temperature with CZA1 CZA2 and CA2 in the calcined state cal and after reduction red The spectra of the calcined samples CZA2 and CA2 are shifted to higher energy to allow a better comparison The single beam spectra served as background for the spectra in fig 3 4 3 5 and 3 7 respectively of CO A study which can more directly be compared with our data was presented by Dulau rent et al 7 They derived isosteric heats of adsorption of CO on a reduced 4 7 Cu Al203 sample from FTIR spectroscopy experiments in the temperature range from 298 to 740 K The reported isosteric heats of 82 kJ mol initial value and 57 kJ mol equilibrium coverage at a pressure of 1 kPa of CO are in excellent agreement with our differential heats of adsorption of CO on CAl In summary the obtained differential heat of adsorption of CO on CAl is in good agreement with literature results obtained with various methods on Cu single crystals and supported Cu catalysts including SiO and Al O3 as non interacting supports Giamello et al 8 measured the differential heat of adsorption of CO on Cu ZnO samples The samples were prepared by coprecipitation from a solution of nitrates They measured values of 70 40 kJ mol for the adsorption of CO on zero valent copper in a sample w
45. 7 6 28 6 29 6 30 6 31 The adsorption microcalorimetry set up personal computers and pressure dis plays amp aa ara San me uen Seer Sa Oe Sa des ee eS Ee we Oe oie es 98 A schematic diagram of a Tian Calvet sensor o 99 A pneumatic valve Swagelok SS 4BG series o o 100 A schematic diagram of the microcalorimetric measuring cells 101 The CO gas bottle stored in a vented gas cupboard 102 The nitrogen gas bottle at the backside of the set up 103 Part of the gas supply iis sir a a ee eS 104 The toolbar of the Setsoft 2000 program 00 105 Choosing the measurement device 2 o o 000000000 105 Description of the experiment 1 4 api id Be es 106 Experiments consist of zones and sequences 00 107 Defining the acquisition period and the number of data points fora zone 108 Monitoring of the experiment in the Real time drawing window 109 The user interface of the Vent Puls program 2 2 2 2 2222 110 The user interface of the Vent Puls Edit program 111 The user interface of the program M4660A mV C3 112 Calibration of the volume of the dosing section 113 Th pretreatmentreaci y oa geag ea ii eh Sig hee hy Beg ee 115 The special glass container and glass funnel 115 Processing an experiment Yesa o as a cor aes Boge a
46. 7 The Baratron pressure gauges ooa 103 6 2 3 Operation of the adsorption microcalorimetry set up 104 6 2 3 1 Controlling the calorimeter via the Setsoft 2000 software 104 6 2 3 2 Controlling the volumetric dosing section via LabView pro 6 2 3 3 6 2 3 4 6 2 3 5 6 3 Data processing Frams 2 a ae a A A a a ole ae a N aN 107 Volume calibration 5 tends Ghee et a a 111 Sample pretreatment o o e Se kw a HS 113 Measurement of heats of adsorption 115 Se Bette a Di leah eri Beatle St DESEE DES il De yee a yore 116 Contents 6 3 1 Calculation of the adsorption isotherm 6 3 2 Calculation of the differential heat of adsorption 1 Introduction Methanol is together with ammonia and sulphuric acid one of the three most important products synthesized industrially Copper catalysts are widely used for the synthesis of methanol The industrially applied catalyst is a ternary system containing copper Cu zinc oxide ZnO and alumina Al203 1 2 Although this catalyst system has been used for about 40 years and in spite of the importance of the methanol synthesis process the nature of the active site the role of the different catalyst components and the mechanism of the methanol synthesis are still subject of investigations Klier 3 proposed copper species incorporated in interstitial and substitutional sites in ZnO to be the active site in methanol synthesis Okamoto et al 4 cor
47. 9 J B Wagner P L Hansen A M Molenbroek H Tops e B S Clausen and S Helveg J Phys Chem B 107 2003 7753 30 H Wilmer T Genger and H Hinrichsen J Catal 215 2003 188 31 M Kurtz H Wilmer T Genger O Hinrichsen and M Muhler Catal Lett 86 2003 74 2 The combined application of microcalorimetry TPD and FTIR spectroscopy Abstract The strong metal support interactions between Cu and ZnO are strongly influenced by the pre treatment of the Cu ZnO catalysts The objective of this contribution is to demonstrate that by pretreating binary Cu ZnO and ternary Cu ZnO Al203 samples under the same conditions the Cu metal surface in identical states is accessible to adsorption microcalorimetry TPD exper iments and transmission FTIR spectroscopy Carbon monoxide is used as probe molecule to investigate the state of the Cu surface after H reduction All results show that a fully reduced and adsorbate free Cu surface is obtained after a reduction pretreatment in flowing hydrogen The adsorption of CO on these surfaces is fully reversible at room temperature with heats of adsorption ranging between 70 kJ mol at low coverages and 45 kJ mol at high coverages 2 1 Introduction Copper catalysts are widely used for the industrial methanol synthesis These catalysts are ternary systems containing copper zinc oxide and alumina 1 Several recent studies indicate there are strong metal support interactions SMSI b
48. CO in total is lower In the case of CA1 and CZAI the desorption from the partially covered surface results in a broad peak that is asymmetric to lower temperatures The peak maximum is shifted to higher temperatures with decreasing initial coverage Tab 4 2 summarizes the results of the TPD experiments after CO pretreatment and hydrogen reduction The final temperature of the first 54 4 Part II The state of the catalyst after pretreatment in CO CZA1 CA1 Effluent mole fraction 75 150 225 300 375 450 Temperature K Effluent mole fraction Effluent mole fraction 150 225 300 375 450 150 225 300 375 450 Temperature K Temperature K Figure 4 4 CO TPD spectra obtained with CZA1 and CA1 fully covered A a and partially covered B E b f with CO after CO pretreatment Tist step 275 K B b 300 K C c 325K D d 350 K E e and 375 K P heating step is Tist step The temperature of the absolute peak maximum is Tmas The full width at half maximum is given in the column FWHM The initial coverages are calculated by integrating the mass spectrometry traces over the complete range of the desorption peak Only small differences in Tmar FWHM and initial coverage are observed in the case of CA1 A comparison of the corresponding peaks obtained with CZA1 shows that the peak maxima are shifted to lower temperatures after the CO pretreatment the peaks are narrower and the initial coverage is significantly
49. Cu 110 and 2100 2102 and 2098 cm 1 Introduction 3 for Cu 211 Cu 311 and Cu 755 It is interesting to note that only the high frequency bands were observed in the case of supported copper evaporated films and dispersed metal even though other characterization methods indicated the presence of low indexed surface planes The authors stated that this effect was not due to incomplete reduction of the copper or tem perature differences in the experimental conditions 77 K in the case of copper single crystal studies about 300 K in the case of supported copper In 1992 Hollins 24 gave an explana tion of this apparent contradiction Dipole dipole coupling of chemisorbed species leads to an intensity shift from low frequency to high frequency bands In the case of CO chemisorbed on copper this intensity shift leads to an increase of the bands assigned to CO adsorbed on defect sites high indexed planes high frequency bands and a decrease of the bands assigned to CO on terrace sites low indexed planes low frequency bands Hollins showed that the resulting spectrum can be completely dominated by the bands assigned to defect sites even when the defect sites are a minority species of less than 10 In 1978 Tauster et al 25 reported that the state of noble metal catalysts using TiO as support depended on the reduction conditions The investigated samples showed different behavior in the adsorption of hydrogen and carbon monoxide after l
50. R spectroscopy did not find any evidence for Cu I species investigating the same catalyst sample after a similar pretreatment In our studies we compare different catalyst samples after identical pretreatment using a com bination of calorimetry TPD experiments and FTIR spectroscopy By doing so we hope to get a deeper understanding of the nature of the SMSI between copper and ZnO and its importance for the synthesis of methanol The interaction of Cu catalysts with hydrogen has been studied extensively by Hinrichsen and co workers 11 12 2 The catalysts have been investigated in different states directly after hydrogen reduction after a period of 12h of methanol synthesis and after a pretreatment of 64 h in CO He The present study is the first of three parts investigating the same catalyst systems after anal ogous pretreatments and using the adsorption of CO as investigation tool All results in this contribution refer to the state of the catalyst after hydrogen reduction 3 2 Experimental The investigated samples are binary and ternary catalysts containing copper zinc oxide and alumina Industrial and model catalysts are included in this study The samples were prepared by coprecipitation using solutions of nitrates and sodium carbonate followed by ageing wash ing drying and calcining The catalysts were characterized by physisorption measurements BET surface area temperature programmed reduction copper content N2O reactiv
51. The coverage dependent adsorption of carbon monoxide on copper catalysts studied by a combination of adsorption microcalorimetry temperature programmed desorption and Fourier transform infrared spectroscopy DISSERTATION zur Erlangung des Grades eines Doktors der Naturwissenschaften vorgelegt von Raoul Naumann d Alnoncourt aus Gottingen Ruhr Universit t Bochum Lehrstuhl fiir Technische Chemie 2005 Die vorliegende Arbeit wurde in der Zeit vom Dezember 2001 bis Dezember 2004 am Lehrstuhl fiir Technische Chemie der Ruhr Universit t Bochum angefertigt Vorsitzender der Pr fungskommission Prof Dr Ch W ll Referent Prof Dr M Muhler Korreferent Prof Dr W Griinert Dritter Priifer Prof Dr R Fischer Tag der miindlichen Priifung 31 Januar 2005 Danksagung Mein besonderer Dank gilt Prof Dr Martin Muhler fiir die wissenschaftliche Leitung dieser Arbeit fiir sein stetes Interesse und seine Geduld Prof Dr Wolfgang Griinert danke ich fiir die Ubernahme des Korreferats Den Mitgliedern der Methanol Gruppe m chte ich fiir die gute Zusammenarbeit danken Elke Hagen Jenny Lamy Maurits Melanie und Olaf Vielen Dank dem technischen und administrativen Personal des Lehrstuhls ohne deren Hilfe die Doktorarbeit nicht m glich gewesen w re Astrid Heinz Manuela Sandy Sigrid und Sigrid Susanne und die Herren Otto Ich m chte an dieser Stelle auch den Werkst tten der Fakult t Chemie der RUB dan
52. The results given for the different pretreatments of CA1 in table 4 2 show only small differences The mi crokinetic modelling needed for a detailed interpretation of these small changes has yet to be done but it can be said that no fundamental changes of the copper content after the CO pre treatment were detected A similar result is found by Wilmer et al 6 investigating the same sample after CO pretreatment in H TPD experiments A comparison of the results obtained after CO pretreatment by microcalorimetry FTIR spec troscopy and TPD experiments using CAl and CA2 leads to the conclusion that the copper content is in the same oxidation state as after hydrogen reduction but the morphology of the copper particles is slightly changed so that there are less surface defects after the CO pretreat ment This effect can be compared to the annealing of copper single crystal surfaces often carried out e g after sputtering The results of the microcalorimetric and the FTIR experiments are in good agreement The TPD experiments give no valuable information in the case of CAl but are in general agreement with the results of the microcalorimetry and FTIR spectroscopy In summary analogue to literature data concerning Cu SiO2 2 3 5 and in good agreement with the data presented by Hinrichsen and co workers 6 no indications were found for SMSI between copper and alumina after CO pretreatment 4 Part II The state of the catalyst after pretreatmen
53. a methanol synthesis gas mixture and a 2 H2 He mixture The methanol synthesis gas consists of 72 H 10 CO 4 CO and 14 He The gases are all of ultra high purity gt 99 9995 Three four way valves 4UWE Valco VICI are used to connect a guard reac tor filled with ZnO the standard U tube reactor and the transmission IR cell described below to the flow set up All reactors can be switched on line or by pass to the flowing gas The U tube reactor fits into an aluminium block with heating elements and a gas line for cooling air which can be controlled by the LabView software The gas analysis is performed by an on line quadrupole mass spectrometer GAM 400 Balzers connected via a heated capillary and a tee piece to the exhaust side of the reactor Data evaluation is performed with the software package Quadstar 2 2 3 Transmission FTIR spectroscopy The FTIR experiments are performed using a Nicolet Nexus FTIR spectrometer equipped with a MCT A detector Experiments can be performed in transmission using a further developed IR cell originally designed by Karge et al 8 Its design is shown in Fig 2 5 The original IR cell designed by Karge et al 8 is connected with a sample pretreatment section and a CO dosing system The cell itself consists of a bronze body soldered to a stainless steel tube of 25mm diameter The bronze body is heatable to 403 K and has 9x 18 mm sized openings for the IR beam The openings are closed by CaF
54. ailable 0 84 umolco gcar 0 62 molco 8car E 0 33 wMolco ear AH is ads the integral molar heat of the adsorption of CO on Cu ZnO AlO 50 35 15 derived from the microcalorimetric data in 15 first heating step is Tist step The temperature of the absolute peak maximum is Tmar The peak maxima are shifted by more than 100 K to higher temperatures compared with a thermal desorption spectroscopy TDS study by Vollmer et al 7 These experiments included low and high indexed copper single crystal surfaces and polycrystalline copper The temperature shift between experiments using single crystals and fixed catalyst beds is due to re adsorption occurring in the catalyst bed The full width at half maximum is given in the column FWHM The TPD peaks are asymmetric to lower temperatures indicating a first order desorption pro cess and are broadened by re adsorption An additional broadening of the largest peak a is 22 2 The combined application of microcalorimetry TPD and FTIR spectroscopy due to contributions from a second desorption site occupied at lower temperatures The initial coverages are calculated by integrating the mass spectrometry traces over the complete range of the desorption peak AH gt designates a rough estimate of the heat of adsorption based on the results of the TPD experiments The estimation follows the method of Konvalinka and Scholten 18 as applied by Sandoval and Bell 19 in a study of the adsor
55. ails about the experimental conditions and the set up are given elsewhere 17 5 3 Results The heat of adsorption was measured calorimetrically using the samples with a high copper content CA1 CZ1 and CZA1 Fig 5 2 5 3 and 5 4 show the results left differential heat of adsorption right adsorption isotherm respectively The results of the experiments af ter hydrogen reduction presented in 10 are included for comparison Please note that the fractional coverages given in fig 5 2 5 3 and 5 4 are calculated using the specific amount of copper surface atoms as measured by nitrous oxygen reactive frontal chromatography for the sample in the state after hydrogen reduction respectively The sorption capacities of the three catalyst samples were differently influenced by the methanol synthesis Compared to the state after hydrogen reduction the equilibrium coverage of the sample at a pressure of 60 Pa of CO was significantly increased in the case of CA1 20 nearly the same for CZ1 and signif 5 Part III The state of the catalyst after methanol synthesis 73 Fractional coverage 0 00 0 08 0 16 0 24 Q kJ mol Fractional coverage Coverage mol g MeOH synthesis 1 adsorption MeOH synthesis 2 adsorption t reduction by hydrogen MeOH synthesis 1 adsorption MeOH synthesis 2 adsorption reduction by hydrogen 0 20 40 60 Coverage mol g Pressure Pa Figure 5 2 Differential heat of adsorpti
56. and computational simulation kinetics on these two samples and that the microcalorimetric and the FTIR experiments were conducted under near equilibrium conditions In the case of CAl and CA2 the adsorption isotherm measured by microcalorimetry and the simulated Temkin type adsorption isotherm fit very well There is only arough agreement with the adsorption isotherm derived from the FTIR results This can be rationalized by the lower quality of the spectra obtained with CA2 The high noise leads to a high uncertainty of the integrated peak areas However the conclusions made for CZAl and CZA2 can be also drawn for CAl and CA2 The shifting of the baseline observed during the adsorption of CO on CZA2 is not yet under stood Further investigations of this effect are needed In summary the results by FTIR spectroscopy are in good agreement with the microcalori metric and the TPD data even though the copper content of the investigated samples differs Experiments with catalysts of the same class yield analogous results The influence of ZnO on the heat of adsorption and the higher fractional coverage of Cu A1 O3 at a given equilibrium pressure were also found in the FTIR data 3 5 Conclusions The results demonstrate that the use of the TPD set up for all sample pretreatments leads to well defined and reproducible catalyst states which are accessible to fundamentally different 44 3 Part I The reduced catalyst investigation methods as t
57. applied pretreatment conditions After hydrogen reduction the copper content of all samples was completely reduced to its zero valent state The copper surfaces were clean and free of adsorbates but the FTIR results suggest that there was a significant amount of defect sites The initial heat of adsorption of CO was lowered significantly by the presence of ZnO Catalysts containing alumina showed higher fractional coverages for a given pressure of CO Although ZnO influences the adsorption of CO the effects observed after hydrogen reduction cannot be classified as SMSI as described by Tauster et al 4 SMSI between copper and ZnO can be clearly seen in the results after CO pretreatment The results support the dynamic alloying model presented by Grunwaldt et al 5 The free copper surface area of ZnO containing samples was drastically decreased due to Zn and O species migrated onto the copper surface The microcalorimetric results showed that the initial heat of adsorption of CO was lower than after hydrogen reduction but the heat increased with increas 86 6 Conclusions ing coverage The increase of the heat of adsorption can be rationalized by Zn and O species on the copper surface or by oxygen vacancies in the supporting ZnO providing sites with higher heat of adsorption The population of these sites seems to be a slightly activated process The TPD data give no indication of an increase of the heat of adsorption with increasing coverag
58. better comparison 55 kJ mol at a fractional coverage of 0 1 The isosteric heat was found to be constant in the cov erage range 0 1 0 5 The initial value and the coverage dependence are in good agreement with the results obtained with CZA1 More recently Vollmer et al 6 derived site specific adsorp tion energies of CO on single crystal faces and poly crystalline copper by thermal desorption spectroscopy TDS For the close packed surfaces Cu 111 and Cu 110 binding energies of 47 and 51 kJ mol respectively were determined Higher values of 58 kJ mol were found for kinks steps and defect structures Keeping in mind that TDS gives only integral heats of adsorption the two single crystal studies confirm the experimental data obtained with CA1 Borgard et al 22 presented calorimetric data concerning the adsorption of CO on a supported catalyst in 1ts reduced state They reported values of 64 46 kJ mol for the adsorption of CO on Cu SiO2 These values were measured for the adsorption of CO on CAI in the coverage range of 14 35 umol Zcat A more detailed comparison of these measurements is not possible due to the lack of any characterization data in Ref 22 However the total amount of adsorbed CO on the Cu SiO catalyst is less than 14 ymol g suggesting a very small specific Cu surface area Therefore the initial heat of adsorption could not be measured reliably with the first dose 3 Part I The reduced catalyst 37 3 00 2 2
59. bration factor k eq 6 4 2 GERAT 6 4 6 2 Experimental 6 2 1 Connectors and flanges The choice of connectors used to build a vacuum system is a very important issue as the connectors have a strong influence on the tightness of the system Three types of connectors are present in the adsorption microcalorimetry set up Swagelok fittings Cajon VCR fittings and CF flanges A KF flange is only used to connect a rough vacuum pump to the set up 6 2 1 1 Swagelok connectors and tube fittings Swagelok connectors and tube fittings are used to connect two metal tubes or a metal tube to a special part like a valve A Swagelok fitting consists of a body a nut and two ferrules Fig 6 1 shows its working mechanism The elements of the fitting are depicted in cross section prior to make up the fitting nut top the back ferrule left the front ferrule center and the fitting body right The tube wall section is shown below the ferrules and body During make up the front ferrule center is driven into the body of the fitting right and the tube bottom to create primary seals tube and body while the back ferrule left hinges inward to create a strong grip on the tube A Swagelok tube fitting can be reusable until the edge of the front ferrule nearly reaches the edge of the back ferrule If that happens cut a small segment of the tube with the ferrules and use a pair of new ferrules Swagelok tube fittings are specified according t
60. c oxide while alumina has only negligible effects on the adsorption behavior of the samples 4 1 Introduction Copper catalysts are industrially used for the synthesis of methanol These catalysts are ternary systems containing copper Cu zinc oxide ZnO and alumina Al203 1 Several recent studies 2 3 4 5 6 indicate there are strong metal support interactions SMSI between cop per and zinc oxide in these catalysts In 1978 Tauster et al 7 reported that noble metal catalysts using TiO show different behavior in the adsorption of hydrogen and carbon monoxide after low temperature reduction and high temperature reduction The adsorption capacity of the samples decreased to nearly zero after high temperature reduction but was fully restored by an oxidizing treatment Using electron microscopy and x ray diffraction the authors showed that this loss of adsorption capacity was not due to metal agglomeration or encapsulation They concluded that the loss of adsorption capacity should be related to the formation of bonds between the noble metal atoms and ti tanium atoms or cations of the support thus changing the electronic properties of the metal clusters They referred to these processes as strong metal support interactions SMSI 48 4 Part II The state of the catalyst after pretreatment in CO Based on in situ EXAFS and XRD experiments Grunwaldt et al 2 presented a model for the SMSI between copper and ZnO Under the redu
61. can produce fast temperature increases leading to a large overshooting The temperature of the thermostated box may not be risen above 418 K otherwise the Baratron pressure gauges will be seriously damaged 6 2 2 4 The measuring cells The microcalorimetric cells are shown in Fig 6 13 They fit exactly into the calorimeter and the inset heating element vide supra The cells are designed as symmetrically and simply as possible They are made of five parts the central part is a tee piece with two double 100 inlet for actuating gas i a bil P aa Figure 6 12 A pneumatic valve Swagelok SS 4BG series sided DN40 CF vacuum flanges and one standard DN40 CF vacuum flange Two identical receptacles for the sample and the reference sample are connected to the two double sided flanges A bellows sealed linear motion feedthrough and a dummy resembling the form of the half expanded feedthrough are connected to the other side of the flanges The linear motion feedthrough can be used to crush a pyrex capsule in the sample receptacle via a steel rod The standard flange is used to connect the cells to the volumetric dosing section All parts of the cells are made of stainless steel and are UHV tight After a microcalorimetric experiment the cells are disconnected from the dosing section To clean the sample receptacle only the flange connecting the receptacle and the tee piece has to be opened 6 2 2 5 The gas supply system
62. cing conditions of methanol synthesis metallic copper particles spread on the support and their surfaces are covered by zinc and oxygen species Under more severe conditions surface and bulk alloying leads to the formation of brass Hansen et al 3 using in situ transmission electron microscopy gave experimental evidence of dynamic shape changes of copper nanocrystals supported on ZnO The changes were in duced by changes of the reduction potential of the surrounding gas phase and were fully re versible The authors concluded that the changes were caused by adsorbate induced changes of the Cu ZnO interfacial energies The authors assumed that oxygen vacancies in the ZnO play a role in the observed processes Only negligible shape changes were found for copper nanocrystals supported on SiO Wagner et al 5 investigated the SMSI between copper and ZnO by applying in situ electron energy loss spectroscopy They found that the support induces a tensile strain in the Cu nan oclusters The degree of this strain is dependent on the reduction potential of the surrounding gas phase They found no indication for strained Cu nanoclusters in Cu SiO2 samples Hinrichsen and co workers 4 6 investigated the interaction of hydrogen and nitrous oxide with Cu Al 03 Cu ZnO and Cu ZnO Al 03 samples as a function of the pretreatment The authors found that Cu Al203 is hardly influenced by changes of the pretreatment while dy namical changes of the copper conte
63. corded while purging with a nitrogen gas cylinder Prior to the reduction of the samples CO was adsorbed on the calcined samples Fig 3 7 shows the range of the CO stretching vibration for CZA1 CZA2 and CA2 at a pressure of 100 Pa of CO The resulting broad and symmetric band has a peak maximum at 2112 cm in the case of CZA1 and at about 2100 cm for CZA2 and CA2 3 4 Discussion The differential heat of adsorption of CO on CAl is in good agreement with data in literature In 1972 Tracy 21 measured the isosteric heat of adsorption of CO on a Cu 100 single crystal plane in the temperature range from 77 to 300 K as a function of the coverage An isosteric heat of adsorption of about 70 kJ mol was reported for very low coverages decreasing to about 36 3 Part I The reduced catalyst CA2 2090 cm Exti Extinction Extinction 2150 2100 2050 2000 1950 A Wavenumbers cm Figure 3 5 FTIR spectra obtained with CA2 after reduction in the pressure range of 0 100 Pa of CO and at room temperature The left side shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right side shows the range below upper right and above lower right the CO stretching vibra tion for a pressure of 0 and 100 Pa of CO The upper right and the lower right spectrum at 0 Pa CO are shifted by an extinction of 0 04 and 0 75 respectively to allow a
64. d to 450 K in two steps e g the sample is first heated to 275 K again cooled to 78 K and finally heated to 450 K In the first step all loosely bound CO species are desorbed The final temperature of the first step defines the initial surface coverage of the second step In order to investigate the influence of the coverage on the desorption the complete experiment is repeated at different final tem peratures of the first step Tist step Typical temperatures are 275 300 325 and 350 K This technique leads to surface coverages of CO which are comparable to the coverages obtained in the microcalorimetric experiments 2 1 Fig 2 7 shows the results of the experiments with varying initial surface coverage The char 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 21 oO N o 0 15 Effluent mole fraction of CO o o o al o 0 00 200 250 300 350 400 450 Temperature K Figure 2 7 CO TPD spectra obtained with Cu ZnO Al203 50 35 15 with varying initial cov erage of CO Tist step 275 K a 300 K b 325 K c and 350 K d acteristics of the four peaks a d are summarized in table 2 1 The final temperature of the Table 2 1 CO TPD data obtained with CW ZnO ALO 50 35 15 Peak Tist step Tmar FWHM coverage AH A IK IK IK umolco gear kJ mol kJ mol a 275 315 112 145 50 8 7 b 300 346 87 102 55 6 56 4 c 325 357 6l 62 57 3 58 4 d 350 367 49 33 58 9 61 74 a not av
65. e Compared to the state after hydrogen reduction the state of copper in Cu Al203 samples can be described as completely reduced with less surface defects as indicated by all investigation methods The data measured after methanol synthesis show that copper was completely reduced to its zero valent state although the copper surface was partly covered with atomic oxygen The atomic oxygen reacted with adsorbed CO and resulted in the formation of carbon dioxide and carbonates The observed reversibility of the adsorption of CO can be rationalized by the existence of a second oxygen species This second species can restore the consumed atomic oxygen on the copper surface The results show that the SMSI observed after CO pretreatment of the ZnO containing samples play only a small or even no role under the methanol synthesis conditions of the present study In summary the results support the classification of copper catalysts proposed in ref 1 and the dynamic alloying model presented in ref 5 SMSI between copper and ZnO were observed after CO pretreatment but not after hydrogen reduction or methanol synthesis Copper was partly covered by atomic oxygen after methanol synthesis but was in a completely reduced state More investigations are needed for a thorough understanding of the increase of the measured heat of adsorption after CO pretreatment The exact role of ZnO under these conditions has not yet been determined The increase of the fractiona
66. e 3 3 CO TPD data obtained with CZAl and CAl Sample Peak Tis step Tmax FWHM coverage fractional K K K umolcolZcat Coverage CZA1 B 275 315 112 145 0 28 C 300 346 87 102 0 20 D 325 357 61 62 0 12 E 350 367 49 33 0 06 CAl b 275 336 98 58 0 32 c 300 345 76 42 0 23 d 325 355 59 27 0 15 e 350 364 46 11 0 06 f 375 369 40 2 0 01 a calculated by dividing the total amount of desorbed CO by the number of copper surface atoms The FTIR results of the experiments using the reduced samples CZA2 42 mg 2 cm and CA2 31 mg 2 cm are shown in fig 3 4 and 3 5 In the case of CZA2 mainly one broad and to lower wavenumbers asymmetric band with a peak maximum shifting from 2090 to 2086 cm with increasing coverage is observed The band becomes broader and more asymmetric with increasing coverage Additionally weak bands can be observed below 2000cm at 1689 1246 1228 and 1166cm No bands appear above 2100 cm It is found that the main band at 2086 cm decreases instantly when the cell is evacuated while the weak bands change only little in intensity The adsorption of CO induces a strong baseline shift to lower extinction in the range above 3000 cm Only minimal shifts are observed at lower wavenumbers The spectra of CA2 show only one broad and asymmetric band with a peak maximum at 2090 cm The band becomes also broader and more asymmetric with increasing coverage No further bands and no baseline shift is
67. e contamination of the sample with air by leakage is without measurable effect The reversibility of the adsorption on the ternary catalysts can be safely assumed considering the results presented above and in 15 Microcalorimetric results of Cu Al2O3 85 15 Cu ZnO 70 30 and Cu ZnO ALO 50 35 15 samples pretreated by hydrogen reduction have been presented in 15 The differential heat of adsorption of CO on Cu ZnO A1203 50 35 15 ranges from 68 50 kJ mol with a saturation coverage of ca 85 umolco gcat Following the argumentation of Cardona Martinez and Dumesic in 16 leads to the same conclusion The adsorption isotherms can be described by simple models such as the Temkin or Freundlich isotherm This will be shown elsewhere 17 2 3 2 CO TPD results For CO TPD experiments with the ternary catalyst 50 35 15 100 mg of the sieve fraction of 250 355 um are used After the catalyst pretreatment CO is adsorbed at 300 K in a flowing mixture of CO in He 10 CO 10 Nem min The sample is then cooled to 78 K rapidly in flowing CO He by pouring liquid nitrogen into the heating block It is purged with pure He 10 Ncm min for 10 min and then heated to 450K with a heating rate of 6 K min in flow ing helium The effluent mole fraction of CO is monitored by the mass spectrometer This procedure yields the CO TPD data from the fully covered catalyst surface To vary the initial coverage of the CO TPD experiments the catalyst samples are heate
68. e extinction at a CO pressure of 100 Pa is significantly higher in the case of CA2 In transmission FTIR experiments the extinction is only a function of the concentration and the extinction coefficient of the absorbing species As the latter should be similar for CO adsorbed on CZA2 and CA2 this indicates a higher fractional coverage in the case of CA2 It is possible to calculate a qualitative adsorption isotherm from the FTIR data by integrating the peak areas Fig 3 10 compares the qualitative adsorption isotherms derived from the FTIR data with the adsorption isotherms calculated from the microcalorimetric results and a simulated adsorption isotherm of the Temkin type The Temkin type adsorption isotherms were simulated using the differential heats of adsorption measured by microcalorimetry details concerning the simulation are given in ref 25 The adsorption isotherms are in excellent agreement for the samples CZA1 and CZA2 indicating that the adsorption of CO follows the same adsorption 3 Part I The reduced catalyst 43 Fractional coverage Peak area a u Peak area a u O CZA2 FTIR CZA1 calorimetry Temkin simulated O CA2 FTIR e CA1 calorimetry Temkin simulated 0 20 40 60 80 100 Pressure Pa Pressure Pa Figure 3 10 Comparison of adsorption isotherms derived from microcalorimetric results for samples of high copper content CZAl CAl FTIR experiments with samples of low copper content CZA2 CA2
69. e frontal chromatography reduced copper surface area tests of activity for the methanol synthesis re action and X ray diffraction measurements before and after calcination The preparation and the characterization of the samples are described in detail elsewhere 13 14 15 Table 3 1 summarizes the main characteristics of the used samples The adsorption of carbon monox ide on these samples was studied by adsorption microcalorimetry TPD experiments and FTIR spectroscopy in transmission mode In the case of all three applied investigation methods the set up used for the TPD experiments was also used for the sample pretreatment The fact that the pretreatment was always carried out using the same gas supply and on line gas analysis 3 Part I The reduced catalyst 29 Table 3 1 Characterization and catalytic data Catalyst CZAl CZA2 CZI CAI CA2 BET surface area m g 73 E 64 51 124 Cu content wt CuO 47 7 68 76 20 Specific amount of Cu surface atoms 513 134 513 176 139 umMol g eat Specific Cu surface area m g t 21 6 21 7 6 Specific methanol production rate 0 112 0 065 0 077 0 015 0 012 umol s Zeat Turnover frequency 107 s7 21 8 48 5 150 83 9 0 1 Derived by NO RFC Assuming that 1 m of Cu surface area equals 24 41 umol Cu atoms Obtained at ambient pressure using 100 mg catalyst in synthesis gas 72 H 10 CO 4 CO and 14 He and a volumetric flow rate of 50 cm min STP g
70. e in the reactor may not exceed atmo spheric pressure and the temperature can be raised up to 873 K After pretreating the sample and purging in helium the four way valve of the reactor is closed and the reactor is discon nected from the flow set up The heating element is lowered and the reactor is cooled down In the following step the pressure of helium in the reactor is decreased to 200 500 hPa using a rotary vane pump DUO 2 5 Pfeiffer The reactor itself is then turned vertically by 180 Thus the sample falls out of the U tube into the pyrex tube while the quartz wool plug stays in the U tube Using a small torch fuelled by liquid gas the pyrex tube is welded to yield a sealed 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 15 capsule of 80 90 mm length containing the pretreated sample in helium The sealed capsule is finally placed into the calorimeter All pretreatment procedures that can be applied to samples in the standard U tube reactor can be carried out in the pretreatment reactor except for high pressure treatments and are monitored by mass spectrometry 2 2 2 CO TPD experiments A schematic diagram of the set up is given in Fig 2 4 It includes a gas supply unit with seven gas lines a heated reactor and a quadrupole mass spectrometer for on line gas analysis The ex high pressure e quadrupole unit l mass spectrometer PIC O a DA syn gas O oy A oy e ex Cx r e N O He yde O a
71. e lower CF flange of the tee piece and the VCR connector that is connected to the four way valve The pyrex metal joint equipped with a new NMR tube is connected to the upper CF flange of the tee piece The thermocouple used for the regulation of the heating element is fixed on the outside of the reactor at the position of the sample using metal wire The heating element is lifted until the reactor touches the bottom of the heating element The gas inlet and exhaust of the four way valve the thermocouple and the power supply of the heating element are connected to a suitable set up and the four way valve is opened The desired pretreatment steps e g reduction oxidation or synthesis reactions are carried out After the pretreatment the reactor is purged with helium and cooled down to room temperature The heating element is lowered and the four way valve is closed The complete reactor is disconnected from the set up The reactor is turned vertically by 180 The pretreated sample falls into the pyrex tube while the quartz wool plug stays in the U tube The gas inlet and exhaust of the closed four way valve are connected to the rotary vane pump of the microcalorimetry set up The pressure of helium in the reactor is lowered to 200 500 hPa by shortly opening the four way valve Using a small torch fuelled by liquid gas the pyrex tube is welded to yield a sealed capsule of 95 mm length containing the pretreated sample in helium The sealed capsule is f
72. e stretching vibration of CO is active in FTIR spectroscopy One single investigation method cannot give all desired information For example FTIR spectroscopy can be used to identify chemically different adsorption sites but the strength of the bond between adsorbate and adsorption site cannot be directly measured On the other hand microcalorimetry can be used to measure the heat of adsorption of a specific adsorbate on a substrate but cannot distinguish between chemically different sites with the same heat of adsorption Therefore a combination of microcalorimetry temperature programmed desorption experiments and FTIR spectroscopy was applied to study the adsorption of CO In order to obtain a high degree of comparability with the work of Hinrichsen and co workers 28 30 31 the same samples were included in the present study In addition the samples were pretreated using the same set up and following the same procedures applied in ref 28 30 The building of a microcalorimetry set up including a special pretreatment reactor that can be attached to the flow set up used for the TPD experiments was one of the main tasks in this study In addition an already existing FTIR cell was modified to allow pretreatments in flowing gas and the dosing of CO at pressures below 100 Pa The FTIR cell was also connected to the flow set up used for the TPD experiments All these efforts were made to avoid the problems that arise when data produced with different
73. e upper right corner 6 2 3 2 Controlling the volumetric dosing section via LabView programs The valves of the volumetric dosing section can be operated via a program Vent Puls based on the software package LabView The program uses the parallel port LPT1 to operate the valves The pneumatic valves of the dosing section are actuated via solenoid valves that are connected to the parallel port via a controller box The controller box is equipped with LED lights indicating the state of the solenoid valves and switches that allow manual controlling of the solenoid valves Fig 6 23 shows the user interface of the Vent Puls program The program 108 0 Standard zons Acgasilon parod in sl 24 Humber of pois res Le Cancel Figure 6 21 Defining the acquisition period and the number of data points for a zone can be used to control up to eight valves V1 V8 The valves can be opened and closed by clicking on the grey square buttons A red box indicates an opened valve The valves can also be opened for a defined period Clicking on the grey buttons labelled PULS opens the valves for the time specified in the field next to the buttons respectively Pulses shorter than 0 20 s are to short to affect the pneumatic valves although the solenoid valves are actuated The program can be used to control the valves in an automatic mode using sequence files created with the LabView program Vent Puls Edit vide infra To load a f
74. eam spectrum of CZA2 after CO pretreatment differs strongly from the spectrum after hydrogen reduction while the single beam spectrum of CA2 is hardly influenced by the pretreatment The loss in transparency observed for CZA2 after CO pre treatment could be explained by oxygen vacancies in the ZnO The oxygen vacancies would result in partially filled valence bands of the semiconducting ZnO The electrons in the par tially filled valence bands would be excited by the IR beam and thus increase the absorption of the light Fig 4 10 shows preliminary results derived from microcalorimetric experiments using coprecipitated ZnO Al203 samples after different pretreatments The results after hy ZnO AI O e CO pretreatment H reduction ZnO AI O CO pretreatment O H reduction Q kJ mol Coverage umol g Coverage umollg Pressure Pa Figure 4 10 Preliminary results of experiments measuring the differential heat of adsorption and adsorption isotherms of CO on ZnO A1 O3 at 303 K after CO pretreatment and hydrogen reduction drogen reduction are similar to literature values 25 obtained with commercial available ZnO Kadox25 and precipitated ZnO The increase of the heat of adsorption of CO on ZnO after CO pretreatment is not reported in literature The results suggest that CO is strongly bound to 64 4 Part II The state of the catalyst after pretreatment in CO oxygen vacancies in the ZnO but a certain th
75. econd flange in a suitable bolt hole orientation After hand tightening all bolts use two wrenches for further tightening The tightening process must be done gradually in 1 4 to 1 2 turns of each bolt in an alternating crisscross pattern until 94 the desired torque ratings have been achieved fig 6 6 This procedure ensures a reliable seal due to even gasket compression and deformation CF flanges are suitable even for vacuum conditions below 1x107 Pa They can be opened and closed for over 5000 times without a loss in sealing quality Hex head data CF targes Figure 6 6 The installation of a CF connection 5 6 2 1 4 The KF Flange A KF connection consists of two flanges a viton O ring with a centering ring and a clamp fig 6 7 It can quickly be connected and disconnected The viton O ring can be re used and has a long life time KF flanges can be used under vacuum conditions down to 1x10 Pa 6 2 2 The adsorption microcalorimetry set up The adsorption microcalorimetry set up consists of the microcalorimeter itself a volumetric dosing section in a thermostated box additional heating elements the microcalorimetric mea suring cells a gas supply two vacuum pumps and two pressure gauges It is mainly used to measure the heat of adsorption of CO The set up is based on the works of Spiewak and Dumesic 1 The main modification is the change from glass to stainless steel as construction material The com
76. ed to its zero valent state even after the mild reduction conditions applied No bands can be detected above 2100cm Bands in that region indicate cationic copper species 28 The reason for the poor quality of the spectra obtained with CA2 compared to the results with CZA2 can be explained by the copper content of 20 wt The high copper content severely reduces the transparency of the sample A comparison of the single beam spectra of the samples CZA1 and CZA2 supports this conclusion A lower transparency results in a lower signal noise ratio This can be clearly seen in the spectra of the reduced samples CZA2 and CA2 in the region 1400 1600 cm and above 3000cm Thus no information can be gained about the formation of carbonates during the adsorption of CO on the support of these samples The results obtained with CZA2 are similar to the results presented by Topsge and Topsge 29 for a Cu ZnO sample reduced at 493 K A comparison of these results confirm the strong influ ence of the ZnO on the heat of adsorption of CO found by microcalorimetry while the alumina seems to have no effect on the heat of adsorption The differences in the fractional coverage for a given equilibrium pressure of CO found by microcalorimetry for the catalysts CZA1 and CAl are found analogous in the FTIR experiments CZA2 and CA2 have nearly the same free copper surface area Although the mass of the investigated CZA2 wafer is higher than that of the CA2 wafer th
77. ed when the sample is removed and the set up should be filled with nitrogen not with CO 6 3 Data processing The collected pressure and heatflow data is best processed using the software Origin Micro Cal and Excel Microsoft While the pressure data is directly stored in an ASCII format the heatflow data needs to be exported from the Setsoft 2000 software To export data from the Setsoft 2000 software click the Data processing button in the tool bar see fig 6 17 The appearing window contains the newest collected experiment see fig 6 29 Any experiment may be opened by clicking the open menu item in the Experiment drop down menu Each zone has to be opened in an own window The signals to be displayed and exported have to be added to the window by drag zdrop actions Right click on the desired signal and drop it into one of the fields labelled Y1 Y6 The run time of the zone is always 117 assigned to the x axis To export the zone in the active window click the export menu item in the Zone drop down menu In the appearing window see fig 6 30 define a file name select an ASCII file format typically fixed length click ASCII in file and select the signals to be exported time and heatflow The type of the decimal separator and the number of figures can be individually defined for each signal after clicking the format button The file is exported after clickin
78. encies of CO adsorbed onto the Cu 111 single crystal plane Their investigations include the influence of Zn adatoms and of ZnO deposited on the copper surface At low coverages with CO the calculated values for the binding energy and the vibrational frequency are similar for the case of Zn adatoms and ZnO deposited on the copper surface Compared to the adsorbate free copper surface Cu 111 as reference state the authors predict a lowered binding energy and vibrational frequency of CO The predicted decrease of the binding energy can be found in the results of the TPD experiments carried out using CZA1 A comparison of the data obtained after hydrogen reduction and CO pretreatment shows that the peak maxima are shifted to lower temperatures for experiments with the same Tist step after the CO pretreatment This temperature shift can be interpreted as a decrease of the heat of adsorption of CO The increase in the temperature of the peak maximum B Tist step 275 K is due to the fact that the peak is not well resolved after hydrogen reduction but has a large contribution from the peak with its maximum around 285K After the CO pretreatment peak B is well resolved with nearly no contributions from the low temperature peak which is shifted to 275 K The decrease in the vibrational frequency predicted in ref 23 is found in the FTIR data obtained with CZA2 The shift of the vibrational frequency is much 62 4 Part II The state of the catalyst after pretreatm
79. ent in CO stronger for CZA2 than for CA2 Although the shift might be partially due to a decrease of defect sites as in the case of CA2 the main part of the shift should be caused by the effect of the zinc and oxygen species on the copper surface as indicated by the DFT study The microcalorimetric results concerning the heats of adsorption of CO obtained with the ZnO containing samples CZ1 and CZA1 cannot be fully understood As the TPD results are qualita tively confirmed by the DFT study of Greeley et al 23 it seems likely that the heats measured by microcalorimetry are not solely the heats of adsorption of CO on copper An increase of the heat of adsorption with increasing coverage can be rationalized by surface reactions with consecutive desorption of products Such surface reactions may need a certain threshold cov erage in order to proceed with measurable rates In such a case the amount of adsorbed species is calculated to low and the combined heats of adsorption reaction and desorption are misin terpreted as heat of adsorption This leads to an overestimation of the heat of adsorption An example of such a process is described in 24 A possible reaction in our experiments would be the oxidation of CO by the oxygen species present on the copper surface after CO pretreat ment However this seems unlikely as the observed processes are reversible while a reaction of CO and oxygen with consecutive desorption of carbon dioxide should be
80. er than the measuring range of the used Baratron pressure gauge gt 100 Pa After several pulses of the turbo pump valve the pressure of CO in the dosing section is lowered to a suitable pressure p1p of 85 100 Pa pie is the total pressure of nitrogen and CO after opening the valve connecting the dosing section to the measuring cells p q is the total pressure of nitrogen and CO in the dosing section and the measuring cells at tiena the difference between p c and pia is due to the adsorption of CO and leakage Based on the universal gas equation the partial pressures of nitrogen and CO can be calculated if the volume of the dosing section and the measuring cells and the leakage rate are known The volume of the dosing section is calibrated prior to the experiment the volume of the measuring cells is measured after the experiment and the leakage rate Lrate can be calculated from the increase of the pressure between tostar and toena The amount of nitrogen in the set up originates from the nitrogen filled into the measuring cells after the crushing of the capsule and from leakage The amount of nitrogen is lowered with each dose due to the evacuation of the dosing section between two doses and increases 119 i waka ge rate cakad mea gute g here r pue Logged doma vessel pressure inert gas posal pesare mn mesunng al parta pros sar m measar ng cell Figure 6 31 The logged pressure data linearly with time accordi
81. es as indicated by the entropy values for Saas vib Shown in fig 3 2 Many authors found a linear correlation between the methanol synthesis activity and the free copper surface area of a catalyst Hinrichsen and co workers 23 confirmed this correlation but postulated three classes of catalysts Cu Al203 lowest activity area Cu ZnO slightly higher activity area and Cu ZnO A1 O3 highest ac tivity area The here presented results give a possible interpretation for the different activities of the different catalyst classes The TPD experiments cannot be directly compared to the TDS study of Vollmer et al 6 Readsorption phenomena cannot occur under the UHV conditions of a TDS experiments but have a strong influence on the results of a TPD experiment The peak maxima in a TPD experi 3 Part I The reduced catalyst 39 Fractional coverage 20 40 60 80 Pressure Pa Figure 3 8 Comparison of adsorption isotherms derived from microcalorimetric results using the samples CZA1 CZ1 and CA1 The fractional coverage is calculated by dividing the amount of adsorbed CO by the specific amount of copper surface atoms respectively ment are broadened and shifted to higher temperatures by readsorption processes compared to the TDS experiment A comparison with TPD studies in literature is also difficult as the TPD experiment is influenced by many parameters like the flow rate of the inert gas the length of the catalyst bed or the heating
82. etr 2100 2000 Wavenumbers cm Wavenumbers cm Figure 4 8 The influence of the pretreatment on the main band The spectra are recorded after hydrogen reduction and CO pretreatment during 1h 18h and 50h at a pressure of 100 Pa of CO The results for CZA2 are in the right figure for CA2 in the left figure bands are due to CO adsorbed on defect sites The magnitude of the intensity shift depends on the ratio of the adsorbed species A defect terrace site ratio of 1 10 can lead to the nearly complete absence of the low frequency band The spectrum is dominated by the low frequency band if the ratio is less than 1 100 The shift of the peak maximum to lower wavenumbers ob served after the CO pretreatment of CA2 can be caused by a decrease of the defect terrace site ratio The results presented in 8 confirm the complete reduction of the copper content of CA1 and CA2 to its zero valent state after hydrogen reduction Thus the differences in the results measured after hydrogen reduction and CO pretreatment cannot be explained by changes of the oxidation state of the copper content There are no reports in literature of SMSI between copper and alumina A change of the morphology of the copper particles of CA1 and CA2 after CO pretreatment that leads to less defect sites can explain the results by microcalorimetry and FTIR spectroscopy It has to be noted that the FTIR results give no indication of an increased free copper surface area The pea
83. etreatment on all samples is generally the same but significant differences between the samples of different classes are found by all investigation methods 5 Part III The state of the catalyst after methanol synthesis 81 5 5 Conclusions All results indicate that copper is completely reduced to its zero valent state under the condi tions of methanol synthesis regardless of the supporting material The SMSI observed after CO pretreatment of the ZnO containing samples are not found after methanol synthesis There are strong indications for the presence of atomic oxygen on the copper surface after methanol synthesis Atomic oxygen reacts with adsorbed carbon monoxide thus increasing the measured heat of adsorption Further investigations are needed to explain the reversibility observed in the microcalorimetric experiments The results support the classification of copper catalysts postulated by Hinrichsen and cowork ers 11 Acknowledgments Financial support by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center SFB 558 Metal Substrate Interactions in Heterogenous Catalysis are gratefully ac knowledged Bibliography 1 J B Hansen in Handbook of Heterogenous Catalysis VCH Verlagsgesellschaft Wein heim 4 1997 1856 2 M Muhler E Tornqvist L P Nielsen B S Clausen and H Topsge Catal Lett 25 1994 1 3 S Bailey G F Froment J W Snoeck and K C Waugh Catal Lett
84. etric and fractional inch fittings 3 Installation of a Swagelok tube fitting 3 The Cajon VCR connection 4 2 2 6 2 u Se ion ee The installation of a Cajon VCR connection 4 The installation of a CF connection 5 The KF connection 6 2 2 2 2 mo mn The adsorption microcalorimetry set up 1 CO gas bottle 2 nitrogen gas bot tle 3 5 pressure gauges 4 6 pressure reducing valves 7 8 12 shut off valves 9 filter 10 rotary vane pump 11 Pirani gauge 13 turbomolecular pump 14 membrane pump 15 control unit 16 full range pressure gauge 17 heating element 18 20 pneumatic valves 21 23 fans 24 power sup ply unit 25 26 Baratron pressure gauges 27 heater fan 28 30 heating elements 31 33 temperature controllers 34 CF flange 35 linear motion feedthrough 36 37 measuring cells 38 insulation 39 calorimetric block 40 CS 32 controller 41 power supply unit 42 Linear motion device 43 44 signal conditioners 45 power supply and pressure display unit 46 Voltmeter 47 control unit 48 power supply unit 49 51 personal computer 50 safety e a a Sa ae ARSS E A RA The adsorption microcalorimetry set up front view 78 79 91 91 92 92 93 94 95 130 List of Figures 6 10 6 11 6 12 6 13 6 14 6 15 6 16 6 17 6 18 6 19 6 20 6 21 6 22 6 23 6 24 6 25 6 26 6 2
85. etween copper and zinc oxide in these catalysts Under the reducing conditions of the methanol synthesis the metallic copper surfaces are covered by zinc and oxygen species 2 Under more severe conditions surface and bulk alloying leads to the formation of brass 3 The adsorption of CO can be applied as a tool to investigate the nature of the SMSI effect In this contribution three different techniques are used to study the adsorption of CO on copper catalysts which are linked by the same gas supply and gas analysis units for sample pretreat ment The goal of this contribution is to demonstrate that by pretreating samples under the same 10 2 The combined application of microcalorimetry TPD and FTIR spectroscopy conditions the metallic copper surface is accessible to the different investigation methods in identical states The heat of adsorption can be derived from temperature programmed desorption TPD experi ments using the Redhead equation or from adsorption isotherms using the Clausius Clapeyron equation A more direct approach is to measure the differential heat of adsorption as a func tion of coverage by adsorption microcalorimetry The value of this tool for the characterization of catalyst surfaces is demonstrated in numerous examples in literature focusing e g on acid sites in zeolites or on metal oxide surfaces 4 5 Spiewak and Dumesic 6 presented a tech nique which allows to study reactive catalyst surfaces unimpa
86. evacuated overnight between the first and the second adsorption experiment ences The initial heat of adsorption was slightly higher at 60 kJ mol decreasing to a plateau at 54 kJ mol in the coverage range 10 24 umol g The heat of adsorption increased to a maxi mum of 89 kJ mol and dropped then to 34 kJ mol The start of the increase and the maximum were measured at slightly lower pressures The CO TPD experiments were performed with CA1 and CZA1 Fig 5 5 shows the results of the TPD experiments The desorption from the fully covered copper surface of CA1 and CZA1 experiment a and A resulted in an intense peak at 112 K and a broad signal in the temperature range 200 400 K with a maximum at about 280 K and a shoulder at about 340 K similar to the TPD profiles obtained after hydrogen reduction 10 respectively For both samples the amount of desorbing CO in total was slightly lower compared to the experiments after hydrogen reduction In the case of CA1 and CZA1 the desorption from the partially covered surface resulted in a broad peak that is asymmetric to lower temperatures The peak maximum was shifted to higher temperatures with decreasing initial coverage Tab 5 2 summarizes the results of the TPD experiments after methanol synthesis and hydrogen reduction The final temperature of the first heating step is Tist step The temperature of the absolute peak maximum is Tmar The full 76 5 Part III The state of the catalyst after
87. f 106 IL SETSOFT 2000 Acquection BE Cobection Display Window 7 S omal alala A Parenter Tas Cor e Co Om do E 1 Standard zone E 2 Standard zone E 3 Standard mne E 4 Standard zone E 5 Standard zone fel 6 Standard mne fel 7 Standard sore fel 8 Standard zone fel 3 Standard mne Cea fel 10 Sided son Ciuihcel r E 1 Sendo son aon Masz mg Molar masz g mol 7 m lal 12 Stardsed zon s ae Pram tmb He CO S on Procede nane Colbeston2 Empenment gop co Echo Figure 6 19 Description of the experiment at least one sequence For each sequence an initial and final temperature a heating rate max 2 K min and the duration of the sequence have to be specified see fig 6 20 The calorimeter is equipped with a fan to increase the speed of cooling rates up to 1 K min can be achieved The operation of the fan can be defined for each sequence the switch controlling the fan has to be set to automatic see the C80 II user manual Data storage can be activated for each zone The acquisition time and thus the number of data points per zone Standard 4966 Points has to be defined for each zone individually by filling in the desired values into the pop up window see fig 6 21 appearing after clicking acquisition period in the Zone drop down menu To add a zone to an experiment just right click the last zone in the Experiment explorer area and select Add zone Standard
88. ferential heat of adsorption and adsorption isotherms of CO on CZAI at 303 K after methanol synthesis The results obtained after hydrogen reduction are included for comparison The sample was evacuated overnight between the first and the second adsorption experiment 04 CO TPD spectra obtained with CZA1 and CA1 fully covered A a and par tially covered B E b f with CO after methanol synthesis Tist step 275 K B b 300 K C c 325 K D d 350 K E e and 375 K f 2 2 2 List of Figures 129 5 6 5 7 6 1 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 FTIR spectra obtained with CZA2 after after methanol synthesis in the pres sure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for pressures of 0 and 100 Pa of CO FTIR spectra obtained with CA2 after after methanol synthesis in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for a base line or the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for a pressure of O and 100 Pa of CO The Swagelok connection before left and after right make up 3 M
89. flow rates were always fixed to 10 Ncm min All pretreatments were monitored by on line mass spectrometry For the microcalorimetric experiments 100 mg of the sieve fraction of 250 355 um were pre treated in a specially designed pretreatment reactor and then sealed in a pyrex capsule Next 50 4 Part II The state of the catalyst after pretreatment in CO the pyrex capsule was placed into the sample receptacle of the microcalorimeter C80 II Se taram The calorimetric set up was degassed and the capsule was broken After reaching thermal equilibrium at 303 K the adsorption measurement was started Small doses of car bon monoxide were subsequently admitted to the sample and while the heat of adsorption was measured calorimetrically the amount of adsorbed species was measured volumetrically In order to test the reversibility of the observed processes the sample was evacuated overnight and the experiment was repeated The employed measurement technique was adopted from the pi oneering work by Spiewak and Dumesic 13 The technique allows to investigate air sensitive samples unimpaired by poisoning A detailed description of the experimental procedure and the set up is given elsewhere 12 The TPD experiments were carried out in a stainless steel U tube reactor connected to a flow set up Typically 100 mg of the sieve fraction of 250 355 um were investigated in situ directly after the pretreatment The samples were cooled from room
90. found The band decreases instantly when the cell is evacuated Fig 3 6 shows single beam spectra of CZA1 CZA2 and CA2 For CZA1 only a spectrum of the calcined sample can be achieved The transparency of the sample is reduced to zero upon reduction For CZA2 and CA2 samples with a low copper content spectra of the calcined 3 Part I The reduced catalyst 35 CZA2 zoge cm Extinction oe 0 24100 Paco j 0 3 2150 2100 2050 2000 1950 5250 4500 3750 y 3000 2250 Wavenumbers cm Wavenumbers cm Figure 3 4 FTIR spectra obtained with CZA2 after reduction in the pressure range of 0 100 Pa of CO and at room temperature The left side shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right side shows the range below upper right and above lower right the CO stretching vibra tion for a pressure of 0 and 100 Pa of CO The upper right spectrum at 100 Pa CO is shifted by an extinction of 0 025 to allow a better comparison and the reduced state are presented The transparency of the samples decreases strongly after reduction This effect is stronger for CA2 which has a copper content three times higher than CZA2 The single beam spectra showing gas phase contributions of carbon dioxide and water were recorded while the spectrometer was purged by a commercial air dryer the spectra without gas phase contributions were re
91. g in the case of CZ1 The fact that CZ1 is affected by similar effects as CZA1 but already at much lower coverages is analogous to the findings after CO pretreatment 12 After the CO pretreatment the effects on the adsorption of CO on CZ1 were much stronger than the effects concerning CZA1 This was rationalized by the free copper surface areas of the samples After CO pretreatment the free copper surface area was significantly smaller in the case of CZ1 than in the case of CZA1 while the opposite was found after methanol synthesis This might explain why CZA1 is more strongly influenced as CZ1 The assumption that the effect on the heat of adsorption after methanol synthesis is caused by only one process leads to the conclusion that the effect is not directly related to the supporting material However the results obtained after hydrogen reduc tion 10 show that even in the absence of the strong metal support interactions SMSI found after CO pretreatment 12 there are influences of the supporting material on the adsorption of CO on copper Askgaard et al 9 presented a detailed kinetic model of the methanol synthesis Possible adsorbates left on the catalyst surface from the methanol synthesis are atomic oxygen formate 5 Part III The state of the catalyst after methanol synthesis 79 CA2 2091 cm 100 Paco 100 Paco Extinction Extinction 2125 2000 1875 1750 1300 1250 1200 1150 Wavenumbers cm Wavenumbers cm
92. g the OK button N SETSOFT 2000 Emerrert Zoe Processing Window 7 S 5 alalis md helam 2 Standard mre 2 Standard ore Standard rre E 13 Star tand sone LI HealFlow 5 14 Starddand sons HealFlow Figure 6 29 Processing an experiment 6 3 1 Calculation of the adsorption isotherm The adsorption isotherm is calculated from the collected pressure data only using an excel spreadsheet The spreadsheet includes a template that shows directly the adsorption isotherm The equations in the spreadsheet are based on the following considerations The measured pressure is the total pressure in the set up either in the volumetric dosing section alone or in the dosing section and in the measuring cells It consists of two partial pressures see fig 6 31 the partial pressure of CO and the partial pressure of inert gas nitrogen The partial pressures of nitrogen and CO and the amount of CO adsorbed for each dose can be calculated directly from the collected pressure data 118 N SETSOFT 2000 Processing Emermert Zune Prcesirg Window 7 Preferences on all signals Decimal segerstor Thousands separ 5 een s Figure 6 30 Exporting a zone Poa is the pressure of nitrogen after the crushing of the capsule at tostar in the dosing section and the measuring cells poy is the pressure of nitrogen at toena the increase with time is due to leakage pi indicates a pressure of CO in the dosing section high
93. ges of the copper content were found in the case of ZnO containing samples The strongly reducing conditions of the pretreatment in CO led to a loss in free copper surface area The free copper surface area was derived from hydrogen TPD experiments The authors concluded that zinc and oxygen species migrated onto the copper surface thus decreasing the free surface area The presence of the zinc species was confirmed by nitrous oxide reactive frontal chromatography It is difficult to draw conclusions from a comparison of all these studies as the samples are never prepared and pretreated following identical procedures For example Giamello et al 22 found in their microcalorimetric study the presence of Cu I species in a Cu ZnO sample af ter hydrogen reduction and postulated that these species were dissolved in the ZnO matrix Following the authors their results supported the role of Cu l species as active site in the syn thesis of methanol However the authors also pointed out that Boccuzzi et al 21 using FTIR spectroscopy did not find any evidence for Cu I species investigating the same catalyst sample 1 Introduction 5 after a similar pretreatment The goal of this study is to reach a better understanding of the SMSI between copper and zinc oxide For several reasons the adsorption of carbon monoxide was chosen as investigation tool The adsorption of carbon monoxide on copper catalysts is non dissociative and not activated and th
94. his present manuscript All results refer to the state of the catalyst samples after hydrogen reduction The investigated samples are a Cu ZnO catalyst with a molar Cu Zn ratio of 50 50 an industrial catalyst with a copper content of 50 and a ternary catalyst with a molar Cu Zn Al ratio of 10 60 30 2 2 Experimental 2 2 1 Adsorption microcalorimetry The adsorption microcalorimetry set up consists mainly of three sections the calorimeter the microcalorimetric cells and the thermostated volumetric dosing section A schematic diagram of the set up is shown in Fig 2 1 The calorimeter is a commercial Tian Calvet heat flux microcalorimeter C80 II Setaram It can be operated from room temperature up to 573 K The calorimetric resolution is 0 1 uW and the detection limit is 2 5 uW The calorimeter is equipped with a homemade heating element fitted into the upper part of the calorimeter The calorimeter is connected to the volumetric dosing section by specially designed microcalorimetric cells The heating element keeps all parts of the cells that are not in the calorimetric block or the volumetric dosing section thermostated The microcalorimetric cells are shown in Fig 2 2 They fit exactly into the calorimeter and the heating element The cells are designed as symmetrically and simply as possible They are made of five parts the central part is a tee piece with two double sided DN40 CF vacuum flanges and one standard DN40 CF vacuum flange Two
95. ied pressure The set up includes two Baratron pressure gauges MKS 121AA01000B and MKS 121AA00001B with a range of 0 100 00 kPa and 0 100 00 Pa respectively The gauges are connected to a power supply and pressure display unit MKS PR4000 The MKS PR4000 unit is connected to a personal computer via a voltmeter equipped with a RS 232 output Voltcraft 104 Figure 6 16 Part of the gas supply M 4660A A program based on the software package LabView is used for data logging Typi cally the data measured by the gauge with the lower range is stored by the LabView program The signal generated by a Baratron pressure gauge is sensitive to temperature changes There fore reliable pressure measurement can only be achieved after reaching thermal equilibrium of all thermostated parts Prior to each adsorption experiment the zero point of the pressure gauges needs resetting when evacuated to pressures lower than 10 Pa and thermally equilibrated The maximum working temperature of the gauges is 423 K and their burst pressure is 0 24 MPa both values may never be exceeded 6 2 3 Operation of the adsorption microcalorimetry set up 6 2 3 1 Controlling the calorimeter via the Setsoft 2000 software A detailed description of the Setsoft 2000 software is given in the user manual A short de scription only of the parts necessary for a typical experiment is given below The Setsoft 2000 program can be run under Microsoft Windows NT It is de
96. ile click the button labelled Auto and choose the appropriate file The fields next to the Auto button indicate the run time of each programmed step its total duration the number of the step and the total number of steps in the sequence file Operating the volumetric dosing section in the automatic mode requires thorough knowledge of the complete set up and careful planning The turbomolecular pump can be easily damaged by running an unsuitable sequence file If the valve connecting the turbomolecular pump to the volumetric dosing section is opened for a period longer than 0 20 s while the pressure in the dosing section exceeds 10 hPa the pump will be irreversibly damaged If the pressure exceeds 100 hPa even a short pulse might damage the turbomolecular pump In such a case lower the pressure using the rotary vane pump not possible in the automatic mode Fig 6 24 shows the user interface of the Vent Puls Edit program To create a file each step has to be programmed individually A step is programmed by pressing the buttons analogous to using the Vent Puls program The duration of the step has to be defined in the field Zeit sec The field Satznr shows the number of the present step A step can be finished by pressing the replace button A step can be copied by pressing the C button A copied step can be inserted by pressing the T button followed by the replace button If the replace button was n
97. inally placed into the sample receptacle of the measuring cells All pretreatment procedures that can be applied to samples in a standard U tube reactor can be carried out in the pretreatment reactor except for high pressure treatments due to the glass part Before the next sample pretreatment can be carried 115 out a new pyrex tube has to be welded to the glass metal joint glass lined steel tube Figure 6 27 The pretreatment reactor Figure 6 28 The special glass container and glass funnel 6 2 3 5 Measurement of heats of adsorption Prior to the measurement itself the measuring cells containing the sealed pyrex capsule are put into the calorimeter and connected to the dosing section The calorimeter has to be lowered down completely to introduce the measuring cells To connect the measuring cells to the dosing section the calorimeter has to be lifted until the CF flanges to be connected nearly touch A copper gasket is placed between the flanges and the flanges are connected The heating tape used to thermostat the connection between the calorimetric cells and the dosing section is put in place A thick layer of insulating material is wrapped around the heating tape and the connection In the next step the complete set up is degassed for at least 72 h under static vacuum 116 conditions at elevated temperatures 418 K Before the valve connected to the turbomolecular pump is opened the set up is first evacuated
98. ing the amount of adsorbed CO by the specific amount of copper surface atoms FESPECUVE Ne AAA Simulation of the TPD peaks obtained with CZAl and CAL Comparison of adsorption isotherms derived from microcalorimetric results for samples of high copper content CZA1 CA1 FTIR experiments with samples of low copper content CZA2 CA2 and computational simulation Differential heat of adsorption and adsorption isotherms of CO on CA1 at 303 K after CO pretreatment The results obtained after hydrogen reduction are included for comparison The sample was evacuated overnight between the first and the second adsorption experiment 04 38 39 List of Figures 127 4 2 4 3 4 4 4 5 4 6 4 7 4 8 Differential heat of adsorption and adsorption isotherms of CO on CZ1 at 303 K after CO pretreatment The results obtained after hydrogen reduction are in cluded for comparison The sample was evacuated overnight between the first and the second adsorption experiment 2 2 2 nn nme 52 Differential heat of adsorption and adsorption isotherms of CO on CZAI at 303 K after CO pretreatment The results obtained after hydrogen reduction are included for comparison The sample was evacuated overnight between the first and the second adsorption experiment 53 CO TPD spectra obtained with CZA1 and CA1 fully covered A a and par tially covered B E b f with CO after CO p
99. into the appropriate cells 6 3 2 Calculation of the differential heat of adsorption The differential heat of adsorption AH can be calculated for each dose by dividing the heat generated during the dose by the amount of CO adsorbed during the dose The heat generated during each dose can be determined by integrating the heatflow measured during the dose The integration of the heat flow data is best done using the software Origin MicroCal The ASCII file containing the exported heatflow data can be easily imported into Origin files After creating plots of the heatflow versus the time the heatflow can be integrated using Origins tools The first step is to subtract a baseline from the heatflow signal In the next step markers are used to define the part of the signal that should be integrated The last step is the integration itself The yielded value is the generated heat in units of mJ The value of the generated heat can be directly inserted into the Excel spreadsheet for each dose An included spreadsheet displays the differential heat of adsorption versus the coverage of the sample Bibliography 1 N Cardona Martinez and J A Dumesic Adv Catal 38 1992 149 2 H hne G W H W F Hemminger and H J Flammersheim Differential Scanning Calorimetry Springer Berlin 2003 3 http www swagelok com downloads webcatalogs MS 01 140 pdf Sept 22 2004 4 http www swagelok com downloads webcatalogs MS 01 24
100. ion An example of an adsorption process followed by a surface reaction is given in ref 21 The authors reported that a certain threshold pressure or surface coverage was needed to start the reaction Therefore the heat of adsorption was measured correctly at low coverage At higher coverages the measured heat of adsorption increased significantly Unfortunately the authors did not include the measured adsorption isotherm The reaction of CO with adsorbed oxygen and the formation of carbon dioxide or carbonates are irreversible processes Further investigations are needed to explain the apparent reversibil ity of the adsorption processes It has to be considered that the samples are evacuated overnight between the adsorption experiments Slow processes like the diffusion of atomic oxygen dis solved in the copper subsurface or bulk onto the copper surface thus restoring the initial coverage with oxygen could be the reason for the apparent reversibility In summary the results show that the copper content of all investigated samples after methanol synthesis is in a state which is comparable to the state after hydrogen reduction The mi crocalorimetric results are yet not fully understood but a surface reaction of adsorbed carbon monoxide and atomic oxygen is a reasonable assumption The classification of copper cata lysts into three classes by Kurtz et al 11 is supported by the presented results The effect of the methanol synthesis pr
101. ion measurements before and after calcination The preparation and the characteriza tion of the samples are described in detail elsewhere 9 10 11 The hydrogen reduction is carried out in two steps The samples are first treated for 12h in a mixture of H2 He at 448 K and then for 30 min in pure hydrogen at 513 K The synthesis pretreatment starts with the hydrogen reduction The catalyst is then treated additionally for 12h in a methanol synthesis gas mixture see section 2 2 2 The CO pretreatment also starts with the hydrogen reduction and the catalyst is treated additionally for 60h in a mixture of CO He at 498 K All samples are flushed for at least 30 min at elevated temperatures in a flow of pure helium after the pretreatment The flow rates are always fixed to 10 Ncm min all heating rates are set to 1 K min 2 3 Results and discussion 2 3 1 Calorimetry In order to measure the differential heat of adsorption of carbon monoxide on the sample small doses of CO ca 1 umol are sequentially admitted to the sample until the saturation of the sample surface at an equilibrium pressure of ca 80 hPa is reached The admission of each dose of CO follows the same cycle consisting of four steps The first step is the evacuation of the volumetric dosing section to a pressure of less than 10 Pa while the valve to the microcalorimetric cells is closed In the second step the dosing section is filled with ca 80 Pa of CO The admission of CO i
102. ired by poisoning For example they applied this technique to determine the heat of adsorption of nitrogen on reduced and extremely air sensitive iron catalysts Samples were pretreated ex situ in ultra pure flowing gases and then sealed in pyrex capsules in inert gas These capsules were transferred into the calorimeter and were broken after degassing the microcalorimetric cells Small doses of the adsorptive gas were subsequently admitted to the adsorbent and while the heat of adsorption was measured calorimetrically the amount of adsorbed species was measured volumetrically A crucial point is the required very low leakage rate of the complete set up This procedure prevents the contamination of the sample with oxygen or moisture and yields data of the clean catalyst surface A further developed version of the adsorption microcalorimetry set up de scribed in 6 was used in this contribution to study the adsorption of CO on copper catalysts as a function of coverage and pretreatment The adsorption of CO can be indirectly investigated by TPD experiments For non activated adsorption the activation energy of the desorption equals the heat of adsorption as is the case for CO adsorption on metallic Cu surfaces 7 Consequently the TPD peaks are shifted to higher temperatures due to readsorption within the fixed bed The coverage dependence of the heat of adsorption can be investigated by varying the initial coverage The experiments were carried out
103. is Maca ews ew Meow Pe i Dene er 59 128 List of Figures 4 9 4 10 5 1 5 2 5 3 5 4 5 5 The potential energy diagram shows the reaction of CO and oxygen on a copper surface The presence of Zn species on the copper surface could inhibit the desorption of carbon dioxide thus leading to the reverse reaction including the GeSOMUON OTE Oi 5 445 tas es RE SSA Preliminary results of experiments measuring the differential heat of adsorption and adsorption isotherms of CO on ZnO Al20 at 303 K after CO pretreatment and hydrogen reduction 2 2 Cm The mass spectrometry traces of hydrogen carbon monoxide carbon diox ide water and methanol recorded while flushing the sample with pure helium at 498K The flushing out of educts and products is almost completed after 10 min but is carried on for at least 20 Min more 0 Differential heat of adsorption and adsorption isotherms of CO on CAl at 303 K after methanol synthesis The results obtained after hydrogen reduction are included for comparison The sample was evacuated overnight between the first and the second adsorption experiment 0 Differential heat of adsorption and adsorption isotherms of CO on CZ1 at 303 K after methanol synthesis The results obtained after hydrogen reduction are included for comparison The sample was evacuated overnight between the first and the second adsorption experiment 04 Dif
104. ith a copper content of 30 wt These values are in perfect agreement with the results obtained using CZ1 The fact that the authors measured slightly lower values at high coverages can be rationalized by the higher equilibrium pressure 5 33 kPa and the resulting higher fractional coverage 0 23 which were not investigated in our study The adsorption of CO on CZA1 yields results which are very similar to the results obtained with CZ1 The heat of adsorption is slightly lower over 38 3 Part I The reduced catalyst Extinction 2150 2100 2050 1 Wavenumbers cm Figure 3 7 FTIR spectra obtained with CZA1 CZA2 and CA2 in the calcined state at a pres sure of 100 Pa of CO and at room temperature Only the spectrum of CZA2 is corrected for a baseline and the gas phase CO spectral contribution the complete coverage range and the fractional coverage at 60 Pa of CO is significantly higher in the case of CZA1 In summary the microcalorimetric experiments indicate that ZnO containing catalysts show lowered initial values of the heat of adsorption compared to unsupported copper literature data or Cu Al203 catalysts literature data and this study Al203 containing samples show higher fractional coverages for a given equilibrium pressure of CO as shown in fig 3 8 The lowered heat of adsorption in combination with the higher fractional coverage for a given equi librium pressure result in a higher mobility of the adsorbed CO speci
105. ith the measurement of V cetls 6 2 3 4 Sample pretreatment Typically experiments are carried out using 100mg of powder samples with a particle size of 250 355 um Pellets and powders consisting of too large particles are ground in an agate mortar The resulting powder is sieved to obtain the desired particle size Too fine powders are pressed to form tablets The tablets are ground and the powder is sieved The tools used for 114 grinding pressing and sieving have to be completely free of impurities which could change the adsorption behavior of the sample e g catalyst poisons or compounds with a very high adsorption capacity Fig 6 27 shows a schematic diagram of the specially designed pretreatment reactor used for the sample pretreatment It consists of a glass lined stainless steel U tube a pyrex metal joint with a NMR tube welded to it a manometer and a tee piece with two CF flanges and one connection to a four way valve The reactor can be heated by a vertically moveable heating element The complete reactor is metal tightened Samples are placed into the U tube and kept in place by a quartz wool plug To pretreat a sample it is first weighed in a special glass container see fig 6 28 A quartz wool plug is put in the longer part of the U tube and its position is adjusted to about 2cm above the bending The sample is put on top of the quartz wool plug using a small glass funnel see fig 6 28 The U tube is connected to th
106. ives access to the copper surface of the samples in identical states with all three investigation methods All results in this contribution refer to the state of the catalyst samples after hydrogen reduction The reduction was carried out in two steps The samples were first treated for 12 h in a mixture of H2 He at 448 K and then for 30 min in pure hydrogen at 513 K All samples were flushed for at least 30 min at elevated temperatures in a flow of pure helium after the pretreatment The flow rates were always fixed to 10 Ncm min all heating rates were set to 1 K min The complete reduction of the copper content to its zero valent state was confirmed for each experiment by on line mass spectrometry For the microcalorimetric experiments 100 mg of the sieve fraction of 250 355 um were pre treated in a specially designed pretreatment reactor and then sealed in a pyrex capsule Next the pyrex capsule was placed into the sample receptacle of the microcalorimeter C80 II Se taram The calorimetric set up was degassed and the capsule was broken After reaching thermal equilibrium at 303 K the adsorption measurement was started Small doses of the adsorptive gas were subsequently admitted to the adsorbent and while the heat of adsorption was measured calorimetrically the amount of adsorbed species was measured volumetrically The employed measurement technique was adopted from the pioneering work by Spiewak and Dumesic 16 The technique allows to i
107. k area of the IR band is decreased by about 40 after the CO pretreatment It is hard to say whether this decrease is caused by changes in the apparent absorption coefficient of the adsorbed CO or a decrease of the concentration of the adsorbed species which would indicate a loss in free copper surface area It also has to be considered that the microcalorimetric experiments were performed using fresh samples for each experi 60 4 Part II The state of the catalyst after pretreatment in CO Table 4 3 Influence of the pretreatment on the main IR band Sample CZA2 CA2 Pretreatment Y Av peak area Y Av peak area lem cm norm cm 7 cm norm Hy red 2086 1 2090 1 1h CO 2073 13 0 45 18h CO 2070 16 0 36 50h CO 2065 21 0 24 2086 4 0 62 a the areas are normalized relatively to the area of the peak after hydrogen reduction after hydrogen reduction after 1 h of CO pretreatment after 18h of CO pretreatment after 50 h of CO pretreatment ment while the FTIR experiments after hydrogen reduction after methanol synthesis 21 and after CO pretreatment were performed using the same wafer Therefore sintering effects should be stronger in the FTIR experiments than in the microcalorimetric ones The TPD experiments after the different pretreatments were also carried out with only one sample The TPD experi ments concerning CAl give no indication of an increased free copper surface area
108. ken Ich bedanke mich bei meinen Eltern die mich w hrend meines gesamten Studiums immer voll und ganz unterst tzt haben Vielen Dank an Silke f r Hilfe und Unterst tzung im Alltag wie in besonderen Zeiten Vielen Dank auch an Euch Wilma Markus Obi Wan Volker die Ethanolgruppe Contents 1 Introduction 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 2 1 Tntroduction ai A o ti o 22 Experimental x s r a d enou A A A ein 2 2 1 Adsorption microcalorimetry 2 2 2 2 nn nennen 222 CODO experiments 2 Da a Ye ae e Y R G 2 2 3 Transmission FTIR spectroscopy o 2 2 4 Investigated samples u a sr se a 2 3 Results and discussion 2 4 228 4 Oak Se dD ek Ye a wet che 23 1 2Calorimelry 2 45 4 os ee Sed See Ded ee Sed ESS Se 2 32 O SCA a a a a a A A 2 3 3 Results by FTIR spectroscopy o 2A CONCIUSIONS 5 terns ra o he eK ee ee ee 3 Part I The reduced catalyst 3 1 Introduction a ce ect wa acest in feet ca Pia Pas ee hee ce Hee htt te ca ret Pe 32 Experimental add dc a ee dde a 3 3 RESUS yoni wee a a Bio ser dr a 34 DISCUSSION Sana sa HAVER de RIA Ae ee we PS we Sa a 3 5 MCONCIUSIONS ue oa a e ee we ele ee Be ES WR Be MD BO BONS 4 Part II The state of the catalyst after pretreatment in CO AA Introduction ee A we NR es om ae ag NE A RO Oe ee Be A 4 2 TBEXDEHIMENILAN es pr Bn e Beene ee ah Selle A A A A BES RESUS ca a Te 2 a Dar Fa a a
109. l coverage in the presence of alumina observed after hydrogen reduction also needs further inquiry At last the observed reversibility of the adsorption of CO after methanol synthesis is not yet understood Bibliography 1 M Kurtz H Wilmer T Genger O Hinrichsen and M Muhler Catal Lett 86 2003 717 2 H Wilmer and O Hinrichsen Catal Lett 82 2002 117 3 H Wilmer T Genger and H Hinrichsen J Catal 215 2003 188 4 S J Tauster S C Fung and R L Garten J Am Chem Soc 100 1978 170 5 J D Grunwaldt A M Molenbroek N Y Tops e H Tops e and B S Clausen J Catal 194 2000 452 Appendix A detailed description of the adsorption microcalorimetry set up and its operation 6 1 Introduction The most direct way to study the strength of the interaction between adsorbed species and the adsorbent is to measure the heat of adsorption by microcalorimetry Adsorption is a sponta neous process and is therefore generally an exothermic process The heat evolved during the process is the heat of adsorption The heat of adsorption can be determined as integral heat of adsorption or as differential heat of adsorption The integral heat of adsorption q is defined by eq 6 1 ae g 6 1 Nads Q is the total amount of heat evolved during the adsorption of the complete amount Naas of the adsorptive on the adsorbent The differential heat of adsorption q f is related to q according
110. l heat of adsorption and adsorption isotherms of CO on Cu ZnO 50 50 determined at 300K ee o 19 CO TPD spectra obtained with Cu ZnO Al20 50 35 15 with varying initial coverage of CO Tist step 275 K a 300 K b 325 K c and 350K d 21 FTIR spectra obtained with Cu ZnO Al203 10 60 30 left and adsorption isotherms of CO obtained with Cu ZnO A1 03 10 60 30 and Cu ZnO Al 03 50 35 15 right determined in the pressure range of 0 100 Pa and at room temperature 2 aie oe its Gace tt a en an BE en IS a ae Dee 23 Differential heat of adsorption and adsorption isotherms of CO on CZA1 CZ1 and CA1 determined at 303 K volar Ae oe Ros Deas De Beat Bas 31 The entropy of adsorption of CO on CZAl and CA1 at 303 K and its contributions 32 CO TPD spectra obtained with CZA1 and CA1 fully covered A a and par tially covered B E b f with CO Tist step 275 K B b 300 K C c 325K D d 350 K E eand 319 KT ea en Be ie 33 FTIR spectra obtained with CZA2 after reduction in the pressure range of 0 100 Pa of CO and at room temperature The left side shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right side shows the range below upper right and above lower right the CO stretching vibration for a pressure of 0 and 100 Pa of CO The upper right spectrum at 100 Pa CO is shifted by an extinction of 0 025 to allow a better c
111. ll a controlling unit CS 32 and a power module The CS 32 controller is used to operate the calorimeter via a personal computer It controls the oven and the fan of the calorimeter and collects the measured data The calorimetric block is mounted on a linear motion device as it has to be lowered to introduce or remove the calorimetric measuring cells see fig 6 9 6 2 2 2 The volumetric dosing section The volumetric dosing section consists of three pneumatic valves Swagelok SS 4BG VCR 3C see fig 6 12 and two Baratron pressure gauges MKS 121AA01000B and MKS 121A A00001B connected by stainless steel tubing 1 4 The pneumatic valves are connected 97 Figure 6 9 The adsorption microcalorimetry set up front view to the gas supply the turbomolecular pump and the measuring cells The internal volume of the dosing section is mainly determined by the volume of the two Baratron pressure gauges ca 15cm per gauge The pneumatic valves are operated via solenoid valves The solenoid valves are connected directly to the actuator of the pneumatic valves in order to minimize the opening time of the pneumatic valves Helium is used as actuator gas for the same reason due to its lower viscosity compared to compressed air The helium is taken from the central gas supply of the laboratory A pressure of about 0 6 MPa is needed to operate the pneumatic valves Minimum opening times are typically 20 ms All parts of the dosing sec
112. mated heat is too high for trace c and too low for trace e and f only for trace d the values are in good agreement This may be rationalized by the microcalorimetric results The mobility of the adsorbed CO strongly depends on the coverage in the case of CAl but changes only little for CZA1 Microkinetic modelling was used to confirm the agreement between the TPD experiments and the microcalorimetric results quantitatively The TPD peaks were simulated using the integral heats of adsorption as measured by microcalorimetry Fig 3 9 shows the results of the calcu lations for the peaks C E and c f Values for Aga and Age were optimized in order to give the best fit between simulated and experimental TPD peaks The used values are included in table 3 4 The peak maxima are given correctly by the simulations but the peaks are to narrow This is due to the fact that the heat of adsorption was kept constant for the simulation of each peak The peaks simulated E e and f fit best as the integral heat of adsorption differs only slightly from the differential heat of adsorption in the regarded coverage range A detailed de scription of the simulation is given elsewhere 25 The simulation confirms also qualitatively the entropies of adsorption measured by microcalorimetry There are only small changes of Aads in the simulation of the experiments C E using CZAI in the case of CA1 Aja ranges from 2 s at low coverage experiments e and f to 111 s at high
113. ments show some differences If the differences in the microcalorimetric results are not related to differences in the state of the copper content as indicated by the two other investigation methods they must be related to differences in the state of the supporting material or to the presence of adsorbates left on the catalyst surface from the synthesis of methanol The microcalorimetric results obtained with CZA1 and CAl indicate that the effects on the heat of adsorption caused by the methanol syn thesis pretreatment are not directly related to ZnO The results obtained with CZA1 and CA1 are comparable although CA1 is free of ZnO A comparison of the microcalorimetric results 78 5 Part III The state of the catalyst after methanol synthesis 100 Pa CO 0 00 1164 cm 1228 cm Extinction Extinction 2125 2000 1875 1750 1275 1200 1125 Wavenumbers cm Wavenumbers cm Figure 5 6 FTIR spectra obtained with CZA2 after after methanol synthesis in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for pressures of 0 and 100 Pa of CO obtained with CZAl and CZ1 leads to the conclusion that the observed effects of the methanol synthesis might have the same cause but the effects are less stron
114. metric block As the calorimetric block is considered to have a constant temperature profile the temperature difference between the cells can be measured by measuring the thermopiles in difference The high sensitivity of a Tian Calvet sensor is due to the amplifying effect of the thermopiles and due to the fact that the thermopiles are the main heat conducting path Thus the amount of heat lost undetected is minimized The complete calorimeter consists of the calorimetric block including the sensor in an insulat 96 50 51 as line Cables Boader of heated box Figure 6 8 The adsorption microcalorimetry set up 1 CO gas bottle 2 nitrogen gas bottle 3 5 pressure gauges 4 6 pressure reducing valves 7 8 12 shut off valves 9 filter 10 rotary vane pump 11 Pirani gauge 13 turbomolecular pump 14 membrane pump 15 con trol unit 16 full range pressure gauge 17 heating element 18 20 pneumatic valves 21 23 fans 24 power supply unit 25 26 Baratron pressure gauges 27 heater fan 28 30 heat ing elements 31 33 temperature controllers 34 CF flange 35 linear motion feedthrough 36 37 measuring cells 38 insulation 39 calorimetric block 40 CS 32 controller 41 power supply unit 42 Linear motion device 43 44 signal conditioners 45 power supply and pressure display unit 46 Voltmeter 47 control unit 48 power supply unit 49 51 personal computer 50 safety switch ing she
115. nd on supported copper samples They assigned bands at 2093 2085 and 2076 cm to the stretching vibration of CO adsorbed onto the low indexed single crystal faces 110 100 and 111 respectively Dulaurent et al 7 investigated the adsorption of CO on a 4 7 Cu A1 03 catalyst prepared by impregnation They found a broad and symmetric band at 2120 cm after adsorbing CO on the oxidized sample After the complete reduction of the copper content in flowing hydrogen at 713 K the adsorption of CO resulted in a broad and to lower wavenumbers asymmetric band at 2092cm A 3 Cu ZnO sample was studied by Boccuzzi and Chiorino 27 They found only one broad and asymmetric band at 2090 cm after adsorbing CO on the reduced sample Topsge and Topsge 29 investigated the influence of the reduction temperature on the IR band of adsorbed CO using 1 and 5 Cu ZnO samples The maximum of the broad and asymmetric band shifted with increasing reduction temperature from 2096 cm 453 K to 2067 cm 573 K After reduction at 493 K the band was found at 2085 cm The authors reported also that no shifts were found in the case of Cu Al 03 42 3 Part I The reduced catalyst samples Our FTIR results are in good agreement with the data in literature even though the weak bands found in the case of reduced CZA2 could not be found in literature The results obtained with CA2 confirm that the copper content of the sample is completely reduc
116. ng cycle consists of up to 15 steps The first step is the evacuation of the dosing section for about 120 s the valve to the turbomolecular pump is open the valves to the measuring cell and the CO gas line are closed In the next step a small dose of CO is admitted into the dosing section by pulsing the 110 mH VENT PULS ae e o MOT 0 05 0 0 0 0 0 0 0 1 DAS DS Figure 6 23 The user interface of the Vent Puls program valve 0 20 s connected to the gas line while all other valves are closed If the resulting CO pressure is unsuitably high but lower than 10 kPa to lower the pressure in the dosing section the valve connected to the turbomolecular pump is pulsed several times 0 20 0 30 s while all other valves are closed When the pressure is in the desired range typically 85 100 Pa as the last step of the cycle the valve to the measuring cell is opened for about 1h with all other valves closed A complete cycle should last exactly one hour The change of the pressure in the dosing section versus time can be monitored and stored by the LabView program M4660A mV C3 Fig 6 25 shows the user interface of the M4660A mV C3 program The name of the data file to be stored can be defined after clicking the button New file Data is only stored when the button store is activated The field next to the button displays the run time of the stored file The signal stored as a function of time is not directly
117. ng to the leakage rate The partial pressure of nitrogen at the end of dose n Pinert n can be calculated according to eq 6 7 Pinert n Pinert n 1 x Veenay Veetls Weisse T Lrate x brand iia 6 7 The equilibrium partial pressure of CO Paa n eq 18 the difference between the measured total pressure and the calculated partial pressure of nitrogen at the end of the dose n The amount of CO adsorbed during dose n Naa n can be calculated from the pressure difference between Pad n eg and the theoretical partial pressure of CO Paan start q 6 9 Pad n start 18 calculated under the assumption that no adsorption of CO occurs eq 6 8 Pad n start Danses x Vdose Pad n 1 eq x Veet Veen Vose 6 8 Nad n Pod nst rt E Diab x Veet Vs x T 6 9 Plotting the sum of the amounts adsorbed of the doses 1 n versus the equilibrium pressure of CO at the end of dose n yields the adsorption isotherm For each dose the values of pn and pna and the length of the dose in minutes have to be inserted into the Excel spreadsheet In addition to the pressure data the Excel spreadsheet needs the following experimental values inserted to calculate the adsorption isotherm the sample mass the active sample area the temperatures of the dosing section and the calorimeter the dosing 120 section volume and the leakage rate The volume of the measuring cells is also calculated from pressure data that have to be inserted
118. nt were found in the case of ZnO containing samples The strongly reducing conditions of a pretreatment in CO led to a loss in free copper surface area due to zinc and oxygen species migrating onto the copper surface The present contribution investigates the influence of a strongly reducing CO pretreatment on the adsorption of carbon monoxide on copper catalysts The investigated copper catalysts are the same samples studied by Hinrichsen and co workers in ref 4 6 The adsorption of CO is investigated by microcalorimetry CO TPD experiments and FTIR spectroscopy in transmission mode Many efforts were taken to assure that the pretreatment conditions are identical for all investigation methods thus giving all experimental techniques access to the copper surface in an identical state The results in ref 8 refer to the state of the catalysts after hydrogen reduction The published data confirm that the results by microcalorimetry CO TPD experiments and FTIR spectroscopy are all related to copper surfaces in an identical state for 4 Part II The state of the catalyst after pretreatment in CO 49 each catalyst system respectively 4 2 Experimental The investigated samples are binary and ternary catalysts containing copper zinc oxide and alumina The samples are identical to those in 8 The methods of preparation and character ization of the samples are described in detail elsewhere 9 10 11 Table 4 1 summarizes the main characteri
119. nto the microcalorimetric cells is step three The fourth step is to measure the heatflow and the pressure drop for 1 h A complete experiment can consist of up to 50 cycles The differential heat of adsorption and the adsorption isotherm can be derived from the experimental data For each cycle the evolved heat is determined by integrating the measured heat flow and the amount of the adsorbed CO is determined from the measured pressure drop using the universal gas equation The volume of the dosing section is constant 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 19 and was determined by the expansion of He into a calibration chamber The volume of the microcalorimetric cells changes with the position of the linear motion feedthrough and the sample volume This volume was measured after each experiment by the expansion of He into the evacuated microcalorimetric cells To distinguish between reversible and irreversible adsorption the complete set up is evacuated overnight without raising the temperature After filling the microcalorimetric cells with ca 80 Pa of helium and reaching a steady baseline of the heatflow signal the adsorption experiment is repeated Any differences between the first and the second adsorption experiment are related to irreversible adsorption In total one experiment can last longer than one week This explains why the leakage rate of the set up has to be less than 107 Pam min Fig 2
120. nvestigate air sensitive samples unimpaired by poison ing A detailed description of the experimental procedure and the set up is given elsewhere 17 The TPD experiments were carried out in a stainless steel U tube reactor connected to a flow 30 3 Part I The reduced catalyst set up Typically 100 mg of the sieve fraction of 250 355 um were investigated in situ directly after the pretreatment The samples were cooled from room temperature to 78 K in a flow of a 10 CO He mixture and afterwards heated to 450 K in a flow of ultra pure He The desorption of CO into the stream of helium was monitored by on line mass spectrometry The coverage dependence of the CO desorption was investigated by varying the initial coverage This was achieved by heating the samples to 450 K in two steps i e after dosing CO the sample was first heated to Tist step 275 300 325 350 and 375 K again cooled to 78 K and finally heated to 450 K The experimental conditions and the set up are fully described elsewhere 17 The infrared spectroscopy experiments were performed using a modified transmission IR cell designed by Karge et al 18 in a Nicolet Nexus FTIR spectrometer The cell was connected to a sample pretreatment section and a CO dosing system The investigations were carried out using wafers of an area of about 2 cm and a mass of less than 50 mg After the pretreatment the wafer was brought into the IR beam and spectra 250 scans resolution 2cm
121. o the outer diameter o d of the used tube Tube fittings for tubes of metric and fractional inch o d may not be interchanged All metric tube fittings have a stepped shoulder on the body hex fig 6 2 Shaped fittings such as elbows 91 Figure 6 1 The Swagelok connection before left and after right make up 3 crosses and tees are stamped MM for metric tubing instead of having a stepped shoulder No shoulder Swagelok Swagslok matne tube and fractional Swagelok meine tube ends N i tube stub T _ DA y a stapped shoulder Figure 6 2 Metric and fractional inch fittings 3 To install a new Swagelok tube fitting for o d of less than 1 in 25 mm insert the tubing into the Swagelok assembled tube fitting Make sure that the tubing rests firmly on the shoulder of the tube fitting body and that the nut is finger tight Then mark the nut at the 6 o clock position While holding the fitting body steady tighten the nut 1 1 4 turns to the 9 o clock position fig 6 3 In the case of reusing a Swagelok tube fitting just tighten it for 1 2 turn 6 2 1 2 The Cajon VCR connection A Cajon VCR connection consists either of two glands a gasket a female and a male nuts or a gland a body a gasket and a nut fig 6 4 Many special parts e g valves or pressure gauges are commercially available with Cajon VCR connectors The sealing face of the glands is polished very finely and demands
122. omparison 35 126 List of Figures 3 5 3 6 3 7 3 8 3 9 3 10 4 1 FTIR spectra obtained with CA2 after reduction in the pressure range of 0 100 Pa of CO and at room temperature The left side shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right side shows the range below upper right and above lower right the CO stretching vibration for a pressure of 0 and 100 Pa of CO The upper right and the lower right spectrum at 0 Pa CO are shifted by an extinction of 0 04 and 0 75 respectively to allow a better COMPArISON gt at a Wee rar SE ee a be et Single beam spectra obtained at room temperature with CZA1 CZA2 and CA2 in the calcined state cal and after reduction red The spectra of the calcined samples CZA2 and CA2 are shifted to higher energy to allow a better compar ison The single beam spectra served as background for the spectra in fig 3 4 3 5 and 3 7 respectively 2 4 tat Sera A BAL gs Rata Rae a tle gaat al FTIR spectra obtained with CZAl CZA2 and CA2 in the calcined state at a pressure of 100 Pa of CO and at room temperature Only the spectrum of CZA2 is corrected for a baseline and the gas phase CO spectral contribution Comparison of adsorption isotherms derived from microcalorimetric results us ing the samples CZA1 CZ1 and CA1 The fractional coverage is calculated by divid
123. on and adsorption isotherms of CO on CAl at 303 K after methanol synthesis The results obtained after hydrogen reduction are included for com parison The sample was evacuated overnight between the first and the second adsorption experiment icantly decreased for CZA1 12 The effect of the pretreatment on the heat of adsorption was roughly similar for the three catalysts The initial value of the heat of adsorption indicated the complete reduction of the copper content of the samples With increasing coverage the heat of adsorption increased reached a maximum and steeply decreased A comparison of the adsorption isotherms measured for the first and second adsorption indicates the reversibility of the adsorption processes only in the case of CZ1 a slight decrease of the adsorption capacity 8 was observed Differences of the measured heat of adsorption between the first and the second adsorption experiment indicate that not all processes are fully reversible and that the catalyst surface was changed after the first adsorption experiment In the case of CA 1 the heat of adsorption first adsorption decreased with increasing coverage from 59 41 kJ mol followed by an increase to a maximum of 77 kJ mol After reaching the maximum the heat dropped to 34 kJ mol The start of the increase can be correlated by the adsorption isotherm to a CO pressure of about 5 Pa The maximum was measured at a CO pressure of about 32 Pa The adsorption isotherm was e
124. opy chapter 2 13 B E Spiewak and J A Dumesic Thermochim Acta 290 1996 43 14 H G Karge and W Nie en Catal Today 8 1991 451 15 J C Tracy J Chem Phys 56 1972 2748 16 J Kessler and F Thieme Surf Sci 67 1977 405 17 P Hollins and J Pritchard Surf Sci 89 1979 486 18 S Vollmer G Witte and C W ll Catal Lett 77 2001 97 19 G E Parris and K Klier J Catal 97 1986 374 20 P Hollins Surf Sci Rep 16 1992 53 21 22 M Kurtz H Wilmer T Genger O Hinrichsen and M Muhler Catal Lett 86 2003 Tak Part III The state of the catalyst after methanol synthesis chapter 5 68 Bibliography 23 J Greeley A A Gokhale J Kreuser J A Dumesic H Topsge N Y Topsge and M Mavrikakis J Catal 213 2003 63 24 J Shen J M Hill R M Watwe B E Spiewak and J A Dumesic J Phys Chem B 103 1999 3923 25 E Giamello and B Fubini J Chem Soc Farraday Trans 1 79 1983 1995 5 Part III The state of the catalyst after methanol synthesis Abstract The adsorption of carbon monoxide was used to probe the state of copper after the synthesis of methanol The investigated samples were binary and ternary catalysts containing copper zinc oxide and alumina All samples were pretreated under identical conditions resulting in repro ducible states of the catalysts A comparison of the results with a study on the same sample
125. ot pressed all changes are lost when the step number is changed The step number can 109 N SETSOFT 2000 Acquicition Lobecton Display Window CBOLI Eu e file el Mansal programming CB0_It Erle Sampie lemper due C Figure 6 22 Monitoring of the experiment in the Real time drawing window be changed by pressing the buttons lt and gt To save all steps up to the step displayed press the write button and define a file name To alter an existing file press the read button choose the file alter the steps as desired and save it press the write button A sequence file is stored as a simple text file Each line represents a step and consists of 10 numbers The first number indicates the state of all valves following a binary code The next eight numbers are the numbers in the fields of the Vent Puls Edit user interface indicating the opening time for each valve even if the valves are not operated in the pulse mode The last number is the duration of the step The binary code of the first number consists of 16 bits The first eight bits indicate opened and closed valves 0 closed 1 open the last eight bits indicate which valves are operated in the pulse mode O closed 1 pulse e g a value of zero means all valves are closed three means valve 1 and 2 are open 1024 means only valve 3 is open pulse mode A typical adsorption experiment consists of up to 50 dosing cycles Each dosi
126. ow temperature reduction and high temperature reduction After low temperature reduction all samples adsorbed hydrogen and CO in reasonable agreement to literature data The adsorption capacity of the samples decreased to nearly zero after high temperature reduction The effect was reversible as the adsorption ca pacity was fully restored by an oxidizing treatment and subsequent low temperature reduction Using electron microscopy and X ray diffraction the authors showed that the loss of adsorption capacity was not due to metal agglomeration or encapsulation They concluded that the loss of adsorption capacity should be related to the formation of bonds between the noble metal atoms and titanium atoms or cations of the support thus changing the electronic properties of the metal clusters They referred to these processes as strong metal support interactions SMS Several recent studies 26 27 28 29 30 using different investigation methods indicate that there are strong metal support interactions SMSI between copper and zinc oxide in these catalysts Based on in situ EXAFS and XRD experiments Grunwaldt et al 26 presented a model for the SMSI between copper and ZnO as a function of the surrounding atmosphere Under the reducing conditions of methanol synthesis metallic copper particles spread on the support and 4 1 Introduction their surfaces are covered by zinc and oxygen species Under more severe conditions 1 e in an atmosphere of
127. perature K Effluent mole fraction Effluent mole fraction 150 225 300 375 450 150 225 300 375 450 Temperature K Temperature K Figure 3 3 CO TPD spectra obtained with CZA1 and CA1 fully covered A a and partially covered B E b f with CO Tist step 275 K B b 300K C c 325K D d 350K E e and 375K f The CO TPD experiments include the two samples CZA1 and CA1 Fig 3 3 shows the results of the TPD experiments The desorption from the fully covered copper surface experiments A and a results in an intense peak at 115 K and a broad signal in the temperature range 200 400K with a maximum at about 285 K and a shoulder at about 345 K in the case of CZA1 and CA1 The desorption from the partially covered surface experiments B E and b f results in a broad peak that is asymmetric to lower temperatures The peak maximum is shifted to higher temperatures with decreasing initial coverage Tab 3 3 summarizes the results of the TPD experiments The final temperature of the first heating step is Tist step The temperature 34 3 Part I The reduced catalyst of the absolute peak maximum is T 4z The full width at half maximum is given in the column FWHM The initial coverages are calculated by integrating the mass spectrometry traces over the complete range of the desorption peak The fractional coverage is calculated by dividing the total amount of desorbed CO by the number of copper surface atoms Tabl
128. plete set up is metal tightened and made of components suitable for ultrahigh vacuum UHV conditions The set up is also thermostated and suitable for experiments above room temperature A schematic 95 Coating ring Drg Figure 6 7 The KF connection 6 diagram of the set up is shown in fig 6 8 Fig 6 9 is a photography of the front side of the set up the photography in fig 6 10 shows the personal computers used for data logging and controlling the set up and the rack that houses the turbomolecular pump and several pressure displays 6 2 2 1 The microcalorimeter The central piece of the set up is the calorimeter itself It is a commercially available Tian Calvet heat flux microcalorimeter Setaram C80 II It can be operated from room temperature up to 573 K The Tian Calvet sensor gives a high calorimetric resolution of 0 1 yW and a low detection limit of 2 5 uW The high sensitivity of the calorimeter is needed to detect the usu ally low heats evolved during adsorption processes Fig 6 11 shows a schematic diagram of a Tian Calvet sensor The sample cell and the reference cell are surrounded by thermopiles con necting the cells with the calorimetric block The thermopiles consist of many thermocouples connected in series The voltage generated by a thermopile is a function of the number of joined thermocouples The resulting voltage can be used to measure the mean temperature difference between each cell and the calori
129. ption of CO on Cu SiO03 AH pie is derived using standard values for the Arrhenius parameters of adsorption and desorption 10 Pa s and 10 s7 respectively given by Dumesic et al 20 The values of AH od are slightly lower than AH the integral molar heat calculated from the microcalorimetric data in 15 for corresponding coverages The CO TPD data confirm the decrease of the heat of adsorption with increasing coverage measured by microcalorimetry in good quantitative agreement It can be assumed that the TPD experiments are carried out under thermodynamically con trolled near equilibrium conditions This assumption is based on the following arguments the desorption is very fast see section 2 3 3 there is non activated re adsorption over the length of the catalyst bed and the heating rate of the experiments is sufficiently moderate A modelling of the peaks b d using the integral molar heat of adsorption measured by microcalorimetry and considering re adsorption is in progress 17 The modelling intends to validate the data obtained by the different investigation methods 2 3 3 Results by FTIR spectroscopy The industrial ternary catalyst 50 35 15 has a very low transparency in the mid IR region when completely reduced due to its high copper content Therefore a ternary catalyst system 10 60 30 with a lower copper content was investigated in the transmission FTIR experiments A wafer of ca 45 mg was prepared as described abo
130. r The amount of CO adsorbed at an equilibrium pressure of 60 Pa of CO is increased by roughly 50 compared with the results after hydrogen reduction Please note that the fractional coverages given in fig 4 1 4 2 and 4 3 are calculated using the specific amount of copper surface atoms as measured by nitrous oxygen reactive frontal chromatography for the sample in the state after hydrogen reduction respectively The reversibility of the adsorption is indicated by the results 52 4 Part II The state of the catalyst after pretreatment in CO Fractional coverage 0 00 0 03 0 06 0 09 cat Q kJ mol e t att Coverage umol g Fractional coverage e CO pretr 1 adsorption 4 CO pretr 2 adsorption t reduction by hydrogen 0 20 40 60 80 Coverage mol g y Pressure Pa e CO pretr 1 adsorption A CO pretr 2 adsorption reduction by hydrogen Figure 4 2 Differential heat of adsorption and adsorption isotherms of CO on CZ1 at 303 K after CO pretreatment The results obtained after hydrogen reduction are included for com parison The sample was evacuated overnight between the first and the second adsorption experiment of the second adsorption The only differences are a slight decrease of the coverage and a higher initial value of the heat of adsorption 94 kJ mol The microcalorimetric results obtained with CZ1 show completely different trends The initial value of the heat of adsorption is
131. r the turbomolecular pump The set up completely consists of UHV tight components Adsorption processes can be in vestigated at constant temperatures between room temperature and 423 K The leakage rate of the set up including the microcalorimetric cells is less than 107 Pa m min after degassing for more than 72h at 423 K The leakage rate is derived from the measured increasing rate 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 13 linear motion feedthrough dummy NW NW linear motion feedthrough Ls SST DK S Y ZE Ly Y NN double sided CF flange VL SEL tee piece steel rod Figure 2 2 Schematic drawing of the microcalorimetric cells of the pressure under static vacuum conditions and the volume of the dosing section and the microcalorimetric cells app 100 cm Typically 100 mg of the sieve fraction of 250 355 um is used for the investigations Samples are pretreated ex situ in a specially designed pretreatment reactor and then sealed in a pyrex capsule of 5mm diameter and 80 90 mm length under a reduced pressure of helium 200 500 hPa The pyrex capsule is then placed into the sample receptacle of the microcalorimetric cells The cells are placed into the calorimeter and are connected to the volumetric dosing section The complete set up is evacuated at a temperature of 418K for at least 72h This leads to a dynamic vacuum of less than 107
132. ransient temperature programmed experiments optical spectroscopy and calorimetry The comparison of the microcalorimetric experiments using samples of high copper content with the FTIR results obtained with samples of lower copper content supports the classification of copper catalysts as proposed by Hinrichsen et al 23 All experiments show that the studied copper catalysts are fully reduced by the mild pretreat ment conditions chosen in this study The strong influence of ZnO on the adsorption of CO was found by all applied investigation methods Alumina has no effect on the heat of adsorption but increases the fractional coverage for a given CO pressure The results can give a preliminary explanation for the different methanol synthesis activities of the different catalyst classes Further investigations are needed to confirm this thesis Acknowledgments Financial support by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center SFB 558 Metal Substrate Interactions in Heterogenous Catalysis are gratefully ac knowledged Bibliography 1 J B Hansen in Handbook of Heterogenous Catalysis VCH Verlagsgesellschaft Wein heim 4 1997 1856 2 H Wilmer T Genger and H Hinrichsen J Catal 215 2003 188 3 J B Wagner P L Hansen A M Molenbroek H Topsge B S Clausen and S Helveg J Phys Chem B 107 2003 7753 4 J D Grunwaldt A M Molenbroek N Y Topsge H Topsge and B S
133. re damage of the full range gauge and the turbomolecular pump To prevent severe damaging of the turbomolecular pump the pneumatic valve between the dosing section and the turbomolecular pump may never be opened while the pressure in the dosing section exceeds 10hPa The turbomolecular pump should be switched off only for maintenance rea 103 Figure 6 15 The nitrogen gas bottle at the backside of the set up sons to prolong the lifetime of the pump The exhaust ports of the rotary vane pump and the turbomolecular pump always need to be directly connected to the air ventilation system of the laboratory Otherwise considerable quantities of carbon monoxide may flow into the atmo sphere of the laboratory 6 2 2 7 The Baratron pressure gauges A Baratron pressure gauge is an absolute capacitance manometer of high accuracy about 0 1 of reading It includes a signal conditioning unit that is part of the calibrated sensor The sen sor consists of rigidly attached capacitive electrodes on the reference side of a flexible metal diaphragm Applied pressure changes the deflection of the diaphragm and thus produces a change in the distance between the electrodes and the metal diaphragm and the resulting capac itances The capacity changes are converted into an AC voltage signal via a high impedance bridge circuit The voltage signal is amplified and demodulated to give a stable 0 10 V DC output signal that is directly proportional to the appl
134. related the activity of Cu ZnO catalysts to Cu Cut species in a two dimensional epitaxial layer on ZnO Other authors 5 6 identified copper in its metallic state as the active site and ZnO as inert supporting material that stabilizes a high copper surface 5 Frost 7 assumed that the active sites are located at the Cu ZnO interface Nakamura and co workers 8 9 10 reported that ZnO species mi grated under reducing conditions onto the Cu surfaces forming Cu Zn alloys and stabilizing Cu l species Clausen et al 11 12 found no experimental evidence for Cu I species when investigating Cu ZnO Al203 samples in an in situ X ray diffraction XRD and an in situ X ray absorption fine structure XAFS study under the conditions of methanol synthesis while measuring the catalytic activity using on line gas chromatography X ray photoelectron spec troscopy XPS data presented by Fleisch and Mieville 13 included also no evidence for Cu l species under methanol synthesis conditions Chinchen et al 14 investigated the synthesis of methanol from C labelled reactants over Cu ZnO A1 03 The authors concluded that methanol is predominantly produced from carbon dioxide CO2 and not from carbon monoxide CO for a wide range of CO2 CO ratios in the feed gas Askgaard et al 15 presented a detailed kinetic model of the methanol synthesis over copper catalysts based on surface science studies In their model methanol is synthesized from carbon dio
135. reshold pressure of CO has to be exceeded The threshold pressure would result in CO being first loosely bound to sites on the copper surface and then strongly bound to the oxygen vacancies The stronger influence of these processes on the heat of adsorption of CO on CZ1 can be rationalized by the fact that the free copper surface area after CO pretreatment is much smaller in the case of CZ1 compared to CZA1 The initial value of the heat of adsorption is slightly lower on CZ1 and significantly lower on CZA1 compared to the values after hydrogen reduction respectively There is even a further decrease of the heat of adsorption during the first five doses of CO in the case of CZA1 This agrees qualitatively with the results of the CO TPD experiments from CZA1 and the calculations of Greeley et al 23 The measured increase of the heat of adsorption is already found for the second adsorptive dose in the experiments using CZ1 while the increase occurs only after five doses in the case of CZA1 Please note that the measured equilibrium pressure of CO is similar in both cases when the increase of the heat of adsorption begins about 20 Pa In summary although the results by microcalorimetry and CO TPD experiments obtained with CZAI are not in good agreement all results of the experiments using ZnO containing sam ples indicate an influence of the SMSI between Cu and ZnO on the adsorption of CO The microcalorimetric results are not fully understood yet but
136. retreatment Tist step 275 K B b 300 K C c 325 K D d 350 K E e and 375K nananana aaa 54 FTIR spectra obtained with CZA2 after pretreatment in CO in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for pressures of 0 and 100 Pa of CO and after evacuation 56 FTIR spectra obtained with CA2 after pretreatment in CO in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contribution The right figure shows the range below the CO stretching vibration for a pressure of O and 100PaofCO 57 Single beam spectra obtained at room temperature with CZA2 and CA2 after different pretreatments hydrogen reduction and CO pretreatment during 1h 18h and 50h The single beam spectra after 50h CO pretreatment served as background for the spectra in fig 4 5 and 4 6 respectively 58 The influence of the pretreatment on the main band The spectra are recorded after hydrogen reduction and CO pretreatment during 1h 18h and 50h at a pressure of 100 Pa of CO The results for CZA2 are in the right figure for CA2 1i the left TY UTS os Ya ee Dene Deen Nev
137. rt of the gas supply system i e the shut off valves of the gas lines and the vacuum line the Pirani pressure gauge of the vacuum line and the rotary vane pump The gas supply is connected to the dosing section via a filter in order to prevent damaging of the pneumatic valves by dust particles 102 Figure 6 14 The CO gas bottle stored in a vented gas cupboard 6 2 2 6 The vacuum system Two pumps are included in the set up a rotary vane pump and a turbomolecular pump The rotary vane pump Pfeiffer DUO 2 5 can be used to evacuate the gas lines the dosing section or the complete set up A final pressure of less than 2 5 Pa in the dosing section can be achieved using the rotary vane pump depending on the experimental conditions The rotary vane pump is also used to lower the pressure of helium in the process of sealing a pretreated sample vide infra The operation of the pump is monitored by a Pirani gauge The turbomolecular pump Pfeiffer TMP 071P is used to evacuate the dosing section and the measuring cell to pressures below 10 Pa UHV The turbomolecular pump uses an oil free membrane pump as backing pump The operation of the turbomolecular pump is monitored by a full range pressure gauge Pfeiffer PKR 251 105 107 Pa In order to achieve UHV conditions the connecting line between the dosing section and the turbomolecular pump is always thermostated at 348 K Raising the temperature above 350 K will lead to seve
138. s after hydrogen reduction shows that copper is reduced to its zero valent state after methanol synthesis in the case of all investigated samples The microcalorimetric results indicate that the adsorbed carbon monoxide reacted with adsorbates on the copper surface left from the synthesis of methanol 5 1 Introduction Copper catalysts are industrially used for the synthesis of methanol These catalysts are ternary systems containing copper Cu zinc oxide ZnO and alumina Al203 1 Numerous studies investigating these catalysts in situ under the conditions of methanol syn thesis are found in literature e g 2 3 Important questions are inter alia the state of copper during methanol synthesis and the reaction mechanism of the synthesis of methanol Klier 4 proposed that copper species are incorporated into the ZnO in interstitial and substitutional sites Nakamura and coworkers 5 6 7 reported that Cu I species are stabilzed under reduc ing conditions by ZnO species migrated onto the copper surface Chinchen et al 8 identified metallic copper as the active component in methanol synthesis and ZnO as inert support that stabilizes a high copper surface area Askgaard et al 9 presented a detailed kinetic model of the methanol synthesis based on surface science studies The model includes the water gas shift reaction and a route to formaldehyde as by product The main carbon source for the syn thesis of methanol on copper is carbon
139. s constant The volume needs to be calibrated only after parts of the dosing section are exchanged e g for repairing or maintenance The volume of the measuring cells is influenced by the amount of sample and the position of the linear motion feedthrough This volume has to be measured after each adsorption experiment 112 ECT MET ray noha CAMI A Tero Ie Figure 6 25 The user interface of the program M4660A mV C3 A special calibration chamber built of UHV tight components and a stainless steel cylinder of exactly known volume V are used to calibrate the volume of the dosing section The cylinder can be placed into the calibration chamber The calibration chamber is connected to the gas supply system and can be shut off by a valve fig 6 26 In order to determine the volume of the dosing section V4ose the dosing section the empty calibration chamber and part of the tubing of the gas supply system are filled with nitrogen and the pressure p is recorded The calibration chamber and the tubing together form the volume Vow of unknown value In the next step only V 4ose is evacuated by the turbomolecular pump In the last step the gas contained in V ou is expanded into the dosing section and the pressure p2 is recorded The ratio p pa is calculated The three steps are repeated several times to give a mean value of the pressure ratio with sufficient accuracy The complete procedure is repeated with
140. signed to control many different devices including the C80 II microcalorimeter The Setsoft 2000 consists of five modules Of these modules only the acquisition module is needed during a calorimetric experiment After starting the programme and logging in with a user name and password the 105 acquisition module can be started by moving the mouse pointer to the left side of the active program window A tool bar with five buttons will appear see fig 6 17 Pressing the first button will activate the acquisition module A Acquisition SS Data processing Tools Priteni Apparatus Figure 6 17 The toolbar of the Setsoft 2000 program Figure 6 18 Choosing the measurement device To start an experiment click on the New Collection menu item in the Collection drop down menu In the appearing pop up window see fig 6 18 select C80_IT in the SETARAM ap paratus drop down menu and click the OK button In the appearing window see fig 6 19 the experiment has to be described Only defining an experiment name that serves also as file name is mandatory all other information is only used for file management Each calorimetric experiments consists of at least one zone Up to 15000 data points can be collected in each zone Given a frequency of one data point per second a zone lasts about 4h The duration of the complete experiment defines the number of zones of the experiment Each zone consists o
141. similar to the value measured after hydrogen reduction but the heat of adsorption increases with increasing coverage to a maximum at about 115 kJ mol followed by a steep drop to about 60 kJ mol The final value is similar to the final heat of adsorp tion after hydrogen reduction but it is measured with a large uncertainty of about 10 kJ mol Compared to the results after hydrogen reduction the amount adsorbed at an equilibrium pres sure of 60 Pa is drastically decreased from 64 to 28 umol g ar The second adsorption experi ment implies the complete reversibility of the observed processes In the first adsorption experiment using CZA1 the heat of adsorption first decreases from 58 to 44 kJ mol then increases to 60 kJ mol and finally decreases to 50 kJ mol In the second experi ment the heat increases from about 40 kJ mol to 64 kJ mol and the decreases to 50 kJ mol The final values are similar to the values measured after hydrogen reduction but are measured with a large uncertainty The equilibrium coverage at 60 Pa of CO is significantly lower after the CO pretreatment 56 uumol g compared to the corresponding coverage after hydrogen reduction 4 Part II The state of the catalyst after pretreatment in CO 53 Fractional coverage 0 00 0 04 0 08 0 12 0 16 80 cat fez o oO N Q kJ mol gt o A o Fractional coverage Coverage umol g N is 204 CO pretr 1 adsorption A CO pretr 2 adsorption e CO
142. sing the mass of the reduced sample as weighed directly after the experiment The characterization data in ta ble 3 1 is correlated to the mass of the calcined sample prior to reduction The calorimetric data was reprocessed using the sample mass prior to reduction in order to obtain a better correlation with the characterization data cat Q kJ mol Coverage umol g 0 20 40 60 80 Coverage umollg Pressure Pa Figure 3 1 Differential heat of adsorption and adsorption isotherms of CO on CZA1 CZ1 and CAl determined at 303K Table 3 2 Calorimetric results Catalyst CZAl CZ1 CAI Initial heat of adsorption kJ mol 68 71 81 Final heat of adsorption kJ mol 50 60 38 Equilibrium coverage umoloo Zcat 84 64 36 Fractional coverage 0 16 0 12 0 20 at a CO partial pressure of 60 Pa calculated by dividing the amount of adsorbed CO by the number of copper surface atoms The reversibility of the adsorption of CO under the conditions of the adsorption microcalorime try was experimentally confirmed 17 Following the argumentation of Cardona Martinez and Dumesic 20 a near equilibrium state between the adsorbed CO and the gas phase can be safely assumed Therefore it is possible to calculate the entropy of adsorption AS as from the heat 32 3 Part I The reduced catalyst 1 1 A Entropy JK mol A Entropy JK mol 0 00 0 04 0 08 0 12 0 00 0 05 0 10 0 15 0 20 Fractional coverage Frac
143. stainless steel U tube reactor connected to a flow set up Typically 100 mg of the sieve fraction of 250 355 um were investigated in situ directly after the pretreatment The samples were cooled from room temperature to 78 K in a flow of a 10 CO He mixture and afterwards heated to 450 K in a flow of ultra pure He The desorption of CO into the stream of helium was monitored by on line mass spectrometry The coverage dependence of the CO desorption was investigated by varying the initial coverage This was achieved by heating the samples to 450 K in two steps i e after dosing CO the sample was first heated to Tist step 275 300 325 350 and 375 K again cooled to 78 K and finally heated to 450 K The experimental conditions and the set up are fully described elsewhere 17 The infrared spectroscopy experiments were performed using a modified transmission IR cell designed by Karge et al 18 in a Nicolet Nexus FTIR spectrometer The cell was connected to a sample pretreatment section and a CO dosing system The investigations were carried out using wafers of an area of about 2 cm and a mass of less than 50 mg After the pretreatment the wafer was brought into the IR beam and spectra 250 scans resolution 2cm in the region 800 6000 cm were recorded The pressure of CO was varied stepwise between 0 100 Pa 0 0 5 1 2 5 5 10 20 40 80 100 Pa and evacuation in order to investigate the adsorption of CO as a function of coverage Det
144. stics of the samples Table 4 1 Characterization and catalytic data Catalyst CZA1 CZA2 CZI CAI CA2 BET surface area m 8 at 73 64 51 124 Cu content wt CuO 47 7 68 76 20 Specific amount of Cu surface atoms 513 134 513 176 139 umol g eat Specific Cu surface area m cat 21 6 21 7 6 Specific methanol production rate 0 112 0 065 0 077 0 015 0 012 umol s Zeat Turnover frequency 107 s7 218 485 150 83 9 0 1 Derived by N O RFC Assuming that 1 m of Cu surface area equals 24 41 umol Cu atoms Obtained at ambient pressure using 100 mg catalyst in synthesis gas 72 H 10 CO 4 CO and 14 He and a volumetric flow rate of 50 cm min STP This contribution includes results from adsorption microcalorimetry TPD experiments and FTIR spectroscopy in transmission mode All investigations focus on the adsorption of CO on the copper catlysts The experiments were performed analogous to the experiments in 8 Only the pretreatment conditions were changed The samples were first reduced by flowing hy drogen as described in 8 The reduction was followed by a treatment in a flowing 10 CO He mixture at 498 K The duration of this treatment was 50h in the case of the microcalorimetric and TPD experiments while the duration was varied in the case of the FTIR experiments 1 18 and 50h All samples were flushed for at least 30 min at elevated temperatures in a flow of pure helium after the pretreatment The
145. sured after hydrogen reduction in ref 8 support the dynamic alloying model proposed by Grunwaldt et al 2 for ZnO containing copper catalysts Acknowledgments Financial support by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center SFB 558 Metal Substrate Interactions in Heterogenous Catalysis are gratefully ac knowledged Bibliography 1 J B Hansen in Handbook of Heterogenous Catalysis VCH Verlagsgesellschaft Wein heim 4 1997 1856 2 J D Grunwaldt A M Molenbroek N Y Topsge H Topsge and B S Clausen J Catal 194 2000 452 3 P L Hansen J B Wagner S Helveg J R Rostrup Nielsen B S Clausen and H Topsge Science 295 2002 2053 4 H Wilmer and O Hinrichsen Catal Lett 82 2002 117 5 J B Wagner P L Hansen A M Molenbroek H Topsge B S Clausen and S Helveg J Phys Chem B 107 2003 7753 6 H Wilmer T Genger and H Hinrichsen J Catal 215 2003 188 7 S J Tauster S C Fung and R L Garten J Am Chem Soc 100 1978 170 8 9 B Bems M Schur A Dassenoy H Junkes D Herein and R Schl gl Chem Eur J 9 2003 2039 10 O Hinrichsen T Genger and M Muhler Chem Eng Technol 11 2000 956 11 H Bielawa M Kurtz T Genger and O Hinrichsen Ind Eng Chem Res 40 2001 2793 Part I The reduced catalyst chapter 3 12 The combined application of microcalorimetry TPD and FTIR spectrosc
146. surface atoms 513 134 513 176 139 umol gcat Specific Cu surface area m g at 21 6 21 7 6 Specific methanol production rate 0 112 0 065 0 077 0 015 0 012 umol s Zeat Turnover frequency 107 s7 218 485 150 83 9 0 1 Derived by N O RFC Assuming that 1 m of Cu surface area equals 24 41 umol Cu atoms Obtained at ambient pressure using 100 mg catalyst in synthesis gas 72 H 10 CO 4 CO and 14 He and a volumetric flow rate of 50 cm min STP This contribution includes results from adsorption microcalorimetry TPD experiments and FTIR spectroscopy in transmission mode All investigations focus on the adsorption of CO on the copper catlysts The experiments were performed analogous to the experiments in 10 Only the pretreatment conditions were changed The samples were first reduced by flowing hydrogen as described in 10 The reduction was followed by 12h of methanol synthesis in a 5 Part III The state of the catalyst after methanol synthesis 71 gas mixture consisting of 72 H2 10 CO 4 CO and 14 He at 498 K All samples were flushed for at least 30 min at elevated temperatures in a flow of pure helium after the pretreat ment As an example in fig 5 1 the helium flushing is shown after methanol synthesis using CZ1 The flow rates were always fixed to 10 Ncm min All pretreatments were monitored by Purging with helium started at time 0 Effluent mole fraction Time s
147. t a pressure of about 12 Pa of CO and the maximum was measured at a pressure of about 45 Pa The results concerning the heat of adsorption were reproduced in the second adsorption experiment with only small differences up to a coverage of about 65 ymol g For higher coverages the heat of adsorption varied irregularly between 93 and 63 kJ mol The adsorption isotherm measured in the second adsorption experiment indicates irreversibly adsorbed species In the first adsorption experiment using CZA1 the heat of adsorption was constant at about 52 kJ mol in the coverage range 0 25 umol Zcat then increased to 100 kJ mol and finally de creased to 45 kJ mol The increase started at a pressure of about 4 Pa of CO and the maximum was measured at a pressure of about 20 Pa In the second adsorption experiment the adsorption isotherm was exactly reproduced The heat of adsorption was reproduced with small differ 5 Part III The state of the catalyst after methanol synthesis 75 Fractional coverage 0 00 0 04 0 08 0 12 0 16 cat Q kJ mol Fractional coverage Coverage umol g MeOH synthesis 1 adsorption A MeOH synthesis 2 adsorption t reduction by hydrogen 0 25 50 75 Coverage mol g y Pressure Pa Figure 5 4 Differential heat of adsorption and adsorption isotherms of CO on CZA1 at 303 K after methanol synthesis The results obtained after hydrogen reduction are included for com parison The sample was
148. t in CO 61 Hinrichsen and co workers 4 6 concluded that after the strongly reducing CO pretreatment the copper surface of ZnO containing catalysts is partially covered by zinc and oxygen species Their study included the same catalysts as those in the present contribution The conclusions of the authors are based inter alia on the decreased capacity of CZA1 for the adsorption of hydrogen after CO pretreatment The adsorption isotherms yielded from the microcalorimetric results are in good agreement with the results of Hinrichsen and co workers 4 6 The capacity for the adsorption of CO is significantly decreased for both ZnO containing samples The fact that the decrease is stronger for CZ1 than for CZA1 may be due to alumina being a structural promotor as found by Kurtz et al 22 The CO TPD experiments using CZA1 also indicate a decreased copper surface area The initial coverages in the TPD experiments from the partially covered surfaces are comparable to the coverages in the microcalorimetric experiments using CZA1 The decrease of the peak area of the main band in the FTIR experiments using CZA2 also indicates a decrease of the free copper surface area The decrease in the peak area is much stronger for CZA2 than for CA2 It can be safely assumed that the decrease might be only partially due to a change in the absorption coefficient of the adsorbed CO species Greeley et al 23 presented a DFT study on binding energies and vibrational frequ
149. temperature to 78 K in a flow of a 10 CO He mixture and afterwards heated to 450 K in a flow of ultra pure He The desorption of CO into the stream of helium was monitored by on line mass spectrometry The coverage dependence of the CO desorption was investigated by varying the initial coverage This was achieved by heating the samples to 450 K in two steps i e after dosing CO the sample was first heated to Tist step 275 300 325 350 and 375 K again cooled to 78 K and finally heated to 450 K The experimental conditions and the set up are fully described elsewhere 12 The infrared spectroscopy experiments were performed using a modified transmission IR cell designed by Karge et al 14 in a Nicolet Nexus FTIR spectrometer The cell was connected to a sample pretreatment section and a CO dosing system The investigations were carried out using wafers of an area of about 2 cm and a mass of less than 50 mg After the pretreatment the wafer was brought into the IR beam and spectra 250 scans resolution 2cm in the region 800 6000 cm were recorded The pressure of CO was varied stepwise between 0 100 Pa 0 0 5 1 2 5 5 10 20 40 80 100 Pa and evacuation in order to investigate the adsorption of CO as a function of coverage Details about the experimental conditions and the set up are given elsewhere 12 4 Part II The state of the catalyst after pretreatment in CO 51 4 3 Results The heat of adsorption was measured
150. ter evacuation of the cell The spectrum of CA2 also mainly shows one broad and asymmetric band with a peak maximum at 2086 cm The band becomes also broader and more asymmetric with increasing coverage Weak bands can be observed at 1246 and 1228 cm The baseline is shifted to higher extinction with increasing coverage in the region below 1400 cm Due to intense noise no information can be gained in the range 1650 1350 cm and above 2200 cm Fig 4 7 shows single beam spectra of CZA2 and CA2 The figures include spectra after hy drogen reduction and after CO pretreatment The transparency of the ZnO containing sample CZA2 decreases strongly after the CO pretreatment as a function of the pretreatment time The single beam spectra of CA2 are not significantly influenced by the pretreatment The sin gle beam spectra showing gas phase contributions of carbon dioxide and water were recorded while the spectrometer was purged by a commercial air dryer the spectra without gas phase contributions were recorded while purging with a nitrogen gas cylinder The influence of the pretreatment on the frequency of the main band is depicted in fig 4 8 The CO pretreatment has a significant influence on the main band The peak maximum shifts to lower wavenumbers and the peak area decreases The effects are much stronger for the ZnO containing sample and increase with the pretreatment time The characteristics of the main 56 4 Part II The state of the
151. the TPD from the copper surface of Cu Al O3 after the CO pretreatment was broader and shifted to lower temperatures indicating changes in the state of the copper content The authors concluded that under the more reducing conditions of the CO pretreatment zinc and 4 Part II The state of the catalyst after pretreatment in CO 57 0o5ICA2 2085 cm 100 Pa CO Extinction 1228 cm 0 0 1248 cm 2125 2000 1875 1750 1350 1275 1200 1125 Wavenumbers cm Wavenumbers cm Figure 4 6 FTIR spectra obtained with CA2 after pretreatment in CO in the pressure range of 0 100 Pa of CO and at room temperature The left figure shows the range of the CO stretching vibration The spectra are not corrected for a baseline or the gas phase CO vibrational contri bution The right figure shows the range below the CO stretching vibration for a pressure of 0 and 100 Pa of CO oxygen species migrated onto the copper surface in the case of Cu ZnO Al203 thus de creasing the free copper surface area while the hydrogen reduction led to fully reduced and adsorbate free copper surfaces Kurtz et al presented a study on the deactivation behavior of supported copper catalysts for the synthesis of methanol Based on their findings the authors proposed a classification of the investigated copper catalyst in three classes Cu Al203 Cu ZnO and Cu ZnO A1 03 The three classes differ significantly in their activity for the synthesis of methanol while the ac
152. the cylinder placed in the calibration chamber yielding a mean value of the ratio of the pressures pj and p3 Based on the universal gas equation V ose can be calculated according to eq 6 5 Vo Vdos iad EBENE p pa pt 03 6 5 The main contribution to V gose 1s the volume of the Baratron pressure gauges V dose is typically 113 oO Chamber where the metal block can be pul in e e oo a Volume of this section Vase including chamber Vou Figure 6 26 Calibration of the volume of the dosing section around 40 cm when two Baratron pressure gauges are installed With V ose known the volume of the measuring cells Veens is easily accessible After evac uation of the complete set up first step Vaose is filled with nitrogen second step and the pressure is recorded p In the third step the nitrogen contained in V4 is expanded into V e s and the pressure is recorded p2 The ratio p p is calculated The three steps are re peated several times to give a mean value of the pressure ratio with sufficient accuracy Based on the universal gas equation V ose can be calculated according to eq 6 6 Vis Vise p1 pa 1 6 6 Depending on the experimental conditions it might be necessary to degas the complete set up over night after the adsorption experiment and prior to the measurement of Veens Slowly desorbing adsorbate species would interfere w
153. tion are reported for steps kinks and defect sites 18 This leads to the conclusion that the morphology of the copper content of CA1 was changed by the CO pretreatment The copper surface has less steps kinks and defects after the CO pretreatment The increased capacity for the adsorption of CO after the CO pretreatment can rather be rationalized by an increase of the free copper surface area than an increase of the fractional coverage The fractional coverage of the copper surface at room temperature and low pressures of CO should not exceed 0 25 19 An increased copper surface area also indicates a change in the morphology of the copper particles The FTIR results obtained with CA2 support the microcalorimetric results Hollins 20 inves tigated the influence of surface defects including steps and kinks of copper surfaces on the infrared spectra of adsorbed CO The author demonstrated that dipolar coupling between dif ferent adsorbate species causes strong effects on the measured spectra Intensity is shifted from low frequency to high frequency bands In the case of copper the low frequency bands can be assigned to terrace places on low indexed copper surface planes while the high frequency 4 Part II The state of the catalyst after pretreatment in CO 59 2086 cm H red 2090 cm H red 2073 cm 1 h CO pretr Extinction Extinction 2086 cm 2070 cm 50 h CO pretr 18h CO pretr 2065 cm 50h CO pr
154. tion contributing to the internal volume are placed in a thermostated box Only the actuators of the pneumatic valves with the attached solenoid valves are placed outside of the box The maximum working temperature of the solenoid valves is 321 K while the parts in the box can be heated up to 418 K Three fans are used to cool the solenoid valves if the temperature of the box is raised above 320 K e g during degassing of the set up vide infra If possible the fans should not be operated during a calorimetric measurement as the flowing air disturbs the heat flow measurement 98 Figure 6 10 The adsorption microcalorimetry set up personal computers and pressure dis plays 6 2 2 3 Controlling the temperature of the set up A constant temperature profile of the set up during a measurement is absolutely necessary as the amount of adsorbed gas is determined volumetrically Therefore the set up includes sev eral controlled heating units The calorimetric block in the calorimeter is a controlled heating unit in itself The parts of the measuring cells which are not in contact with the calorimetric block are thermostated by an additional heating unit that fits exactly into the upper part of the calorimeter replacing part of the insulation The volumetric dosing section is thermostated in an insulated box equipped with a fan and a heating element The connection between the dosing section and the measuring cells is thermostated by a heating
155. tional coverage Figure 3 2 The entropy of adsorption of CO on CZAl and CA 1 at 303 K and its contributions of adsorption AH 4 using eq 3 1 A Hads ASads 3 1 The entropy of the adsorbed CO Saas can be calculated with the eq 3 2 3 3 and 3 4 sy is the entropy of the gas phase the superscript refers to the standard state ASads Sads Sg 3 2 Sg 7 Rin 3 3 A Hads ASis p Rin 3 4 Sads can be interpreted as a sum Of Sads config aNd Saas vib eq 3 5 Sads con fig 18 the configura tional contribution of the entropy and is only a function of the fractional coverage 0 eq 3 6 Sads vib 18 the vibrational contribution and can be correlated with the mobility of the adsorbed species Sads Sads con fig Sads vib 3 5 1 9 7 3 6 Sads config Rin Fig 3 2 shows Sg Sads gt Sads configs ANd Sads vip AS a function of the fractional coverage for the samples CZAl and CA1 s can be correlated to the coverage by the adsorption isotherm 3 Part I The reduced catalyst 33 Only small changes of Sads wi can be observed in the case of CZAl Suds vip increases from about 65 J K mol to 85 K mol with increasing coverage The CO species on CA1 are less mobile at low coverage Sads vip 15 J K mol and similar mobile at higher coverages Sads vib 95 J K mol compared to the CO species on CZA1 CZA1 CA1 Effluent mole fraction 75 150 225 300 375 450 Tem
156. tivity is linearly correlated with the free copper surface area within each class The microcalorimetric results support the postulated classification The adsorption of CO on CA1 CZI and CZA1 is differently influenced by the CO pretreatment CZ1 is most affected while the changes are weakest for CA1 The ZnO containing samples show similar tendencies but to different extents The results obtained for CA1 show that the energy site distribution of the sample was altered by the CO pretreatment The differential heats of adsorption are essentially of the same magnitude as after hydrogen reduction but the abundance of the different adsorption sites has changed There are less sites with heats of adsorption of more than 55 kJ mol and more sites with lower 58 4 Part II The state of the catalyst after pretreatment in CO 1h CO pretr 18h CO pretr Energy a u 50h CO pretr 2500 2000 1500 2500 2000 1500 Wavenumbers cm Wavenumbers cm Figure 4 7 Single beam spectra obtained at room temperature with CZA2 and CA2 after different pretreatments hydrogen reduction and CO pretreatment during 1h 18h and 50h The single beam spectra after 50h CO pretreatment served as background for the spectra in fig 4 5 and 4 6 respectively heats of adsorption The lower values of the heats of adsorption are found in literature for different copper single crystal surfaces 15 16 17 18 while the values of the higher heats of adsorp
157. ve The sample was pretreated by the hy drogen reduction pretreatment The reduction of the copper to its zero valent state could only be qualitatively confirmed by mass spectrometry After the IR cell had been cooled down to room temperature a single beam spectrum was recorded under dynamic vacuum conditions This spectrum served as background spectrum for the adsorption experiment The background spectrum was recorded with 2000 scans at a resolution of 2cm The surface coverage of the sample with CO was then increased in steps by increasing the partial pressure of CO stepwise 1 At each pressure step a spectrum with 500 scans at a resolution of 2cm was recorded Typ 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 23 ically the pressure steps were 0 0 5 1 2 5 5 10 20 40 80 and 100 Pa A period of 10 min between each CO dosing and the recording of the spectrum ensured near equilibrium adsorp tion desorption conditions After recording the spectrum at 100 Pa pressure of CO the cell was evacuated and two spectra were recorded after 1 min 250 scans and 10 min 2000 scans Fig 2 8 shows the CO stretching vibration in the region 1900 2300 cm during the adsorption of CO in the pressure range 0 100 Pa Only one broad and asymmetric band with a peak maxi cat after 1 min of evacuation Extinction Peakareala u Coverage mol g 2250 2100 1950 1800 0 25 50 75 100 Wavenumbers cm
158. windows of 43 mm diameter sealed with viton O rings The sample is positioned into the IR beam in a sample holder made of tantalum which fits exactly into the bronze body The sample holder can be moved to the pretreatment section by a magnetic manipulator The pretreatment section is a stainless steel tube of 120 mm length and an inner diameter of 25mm Three Swagelok tube stubs with 1 8 fittings are welded to the tube It is connected to the cell and the dosing section by DN40 CF vacuum flanges One Swagelok fitting is used to introduce a thermocouple into the pretreatment section The other two fittings are connected to the flow set up described above via a four way valve 4UWE Valco VICI The sample holder is placed directly between the gas inlet and the outlet during pretreatment The pretreatment section can be heated by a heating tape to 673 K All 2 The combined application of microcalorimetry TPD and FTIR spectroscopy 17 CO dosing Pretreatment section IR cell Figure 2 5 Flow scheme of the transmission IR cell pretreatment procedures can be monitored by on line mass spectrometry but due to the large dead zones of the IR cell only qualitative results can be obtained CO is admitted to the sample via a mini leak valve E MLV 22 Caburn MDC All investigations in transmission are carried out using wafers of an area of about 2 cm and a mass of less than 50 mg The wafers are prepared by grinding
159. xactly reproduced in the second adsorp tion experiment while the measured heat of adsorption was different The heat of adsorption first decreased from 41 37 kJ mol then increased to a maximum of 97 kJ mol and dropped to 74 5 Part III The state of the catalyst after methanol synthesis Fractional coverage 0 00 0 04 0 08 0 12 0 16 cat Q kJ mol Fractional coverage Coverage umol g MeOH synthesis 1 adsorption MeOH synthesis 2 adsorption t reduction by hydrogen MeOH synthesis 1 adsorption MeOH synthesis 2 adsorption reduction by hydrogen 0 20 40 60 80 0 20 40 60 80 Coverage mol g Pressure Pa Figure 5 3 Differential heat of adsorption and adsorption isotherms of CO on CZ1 at 303 K after methanol synthesis The results obtained after hydrogen reduction are included for com parison The sample was evacuated overnight between the first and the second adsorption experiment 45 kJ mol The start of the increase and the maximum are found at lower pressures The initial value of the heat of adsorption of CO on CZ1 was equal to the value measured after hydrogen reduction The heat of adsorption stayed at a constant level for coverages up to 42 umMOol Zgcat but the heat measured in the second pulse is already increased compared to the results after hydrogen reduction The plateau was followed by an increase to a maximum of 83 kJ mol and a steep drop to about 35 kJ mol The increase started a
160. xide only Adsorbed carbon dioxide is hydrogenated to methanol and water via formate as stable intermediate product The model includes the water gas shift reaction of carbon monoxide and water and a route from formate to formaldehyde as by product 2 1 Introduction Carbon monoxide is an often used probe molecule In literature many studies are found investigating copper binary e g Cu ZnO Cu Al20O3 or ternary copper catalysts e g Cu ZnO A1 03 by microcalorimetry temperature programmed desorption TPD or Fourier transform infrared FTIR spectroscopy In 1934 Beebe and Wildner 16 measured calorimetrically the heat of adsorption of CO on reduced copper granules They observed a decrease with increasing coverage from an initial value of about 85 kJ mol to a value of about 57 kJ mol at a pressure of about 50 Pa of CO Tracy 17 measured the isosteric heat of adsorption of CO on the Cu 100 single crystal plane and reported values of about 70 kJ mol at low coverage and 56 kJ mol at high coverage Hollins and Pritchard 18 reported isosteric heats of adsorption of CO on Cu 111 of 50 and 38 kJ mol at low and high coverage respectively More recently Vollmer et al 20 derived site specific adsorption energies of CO on single crystal faces and polycrystalline copper by thermal des orption spectroscopy TDS For the close packed surfaces Cu 111 and Cu 110 integral heats of adsorption of 47 and 51 kJ mol respectively were determined
161. zone A typical zone of an adsorption experiment consists of only one sequence in isothermal mode initial final temperature 303 K with 14400 data points an acquisition period of 1 s and a duration of 4h in total The first zone may differ if degassing of the set up is required After all the zones are set click Start experiment in the Collection drop down menu To monitor the experiment in progress click Real time drawing in the Display drop down menu The appearing window see fig 6 22 displays the temperature and heat flow change 107 Y Neda DI 00000 0000 0000 0000 0000 0000 0000 0900 A v r ra u i we we ws wi EE wi F wi F F F F eee eeee eee nn nm mn mn m m u ION Figure 6 20 Experiments consist of zones and sequences versus time typically for the last 10 min The admission of gas into the measuring cells i e the adsorption experiment itself should only be started after the heat flow reaches a stable baseline This might take up to 1 h after starting the programmed experiment due to the onset of the heating action An additional window displays the parameters used for controlling the temperature of the calorimetric block including the PID parameters for an explanation refer to the C80 II user manual and the measured signals The window Manual programming can be used to end an experiment manually by clicking the red square button in th

Bibliography - Ruhr-Universität Bochum

Contents

Download Pdf Manuals

Related Search

Related Contents

Bibliography - Ruhr-Universit&auml;t Bochum

Contents

Download Pdf Manuals

Related Search

Related Contents

Bibliography - Ruhr-Universität Bochum