G A S P H A S E M E T H A N O L S Y N T H E S I S F O R C A R B O N - 1 1 R A D I O P H A R M A C E U T I C A L S by E r i k van L i e r B . E n g . , M c G i l l University, M a y 2001 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Chemical and Bio log ica l Engineering) T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A August 2007 © E r i k van Lier , 2007 Abstract Carbon-11 radiopharmaceuticals are gaining an increasing importance in positron emission tomography due to their importance in diagnostic medicine. The most wide spread method of production of these radiopharmaceuticals is by methylation of an appropriate precursor with the highly reactive [ n C ] m e t h y l iodide. Conventional synthesis o f this intermediate involves l iqu id phase synthesis of [ n C]me thano l , which is the step that l imits the specific activity of the final product. A catalytic gas phase methanol synthesis process was eval-uated, that promises to avoid the loss of specific activity. In this procedure, [ n C ] c a r b o n dioxide produced in the target is first trapped and purified, then converted to [ n C ] c a r b o n monoxide using molybdenum and finally reduced to [ n C ] methanol using a copper zinc oxide catalyst in the presence of hydrogen. In this study a device to trap and purify [ n C ] c a r b o n dioxide was developed and opti-mized. [ n C ] Carbon dioxide produced in target was quantitatively trapped at -20° C on a carbon molecular sieve column and quantitatively released in less than 3.5 minutes. A reactor to convert 50 ppm carbon dioxide to carbon monoxide, based on the reaction with molybdenum, was developed. A commercial ly available process simulator was used to assist the optimization of operating conditions and molybdenum pretreatment methods. Under optimal conditions, carbon dioxide was converted to carbon monoxide wi th over 70% yie ld . A reactor to catalytically convert 50 ppm carbon monoxide to methanol was developed. A copper zinc oxide catalyst was prepared by a co-precipitation method. The catalyst was i i activated by reduction with hydrogen and passivated with compressed air prior to methanol synthesis. The effect of temperature, pressure and flowrate on the conversion of carbon monoxide to methanol were studied and results were used to create a kinetic model. This model was used to determine optimal operating conditions for this reactor and predicts 60% conversion of [ n C ] c a r b o n monoxide to [ n C]me thano l . These findings suggest that gas phase [ n C ] methanol synthesis is a viable alternative to the conventional l iqu id phase method, meriting further studies with carbon-11. i i i Contents Abstract «' Nomenclature x Acknowledgements xv 1 Introduction 1 1.1 Background 1 1.1.1 Positron emission tomography 2 1.1.2 Appl ica t ion of P E T 7 1.1.3 Carbon-11 radiopharmaceuticals 8 1.2 Objective o f the study 9 2 Literature review 11 2.1 Radiochemistry of carbon-11 11 2.1.1 Production of C-11 12 2.1.2 Recovery and purification of n C 0 2 and n C H 4 14 2.1.3 Reduction of n C 0 2 and oxidation of U C H 4 17 2.1.4 Synthesis of u C H 3 O H 20 2.1.5 Synthesis of carbon-11 radiopharmaceuticals 22 2.2 Molybdenum compounds for the reduction of C 0 2 to C O 23 2.2.1 History 23 iv 2.2.2 Reduction of molybdenum oxides . 24 2.2.3 Oxidat ion of molybdenum 26 2.2.4 Interaction of C 0 2 and C O with M o 2 C 27 2.3 Methanol synthesis 30 2.3.1 History 30 2.3.2 Catalyst preparation 31 2.3.3 Catalyst poisoning / deactivation 32 2.3.4 Thermodynamics and kinetics 32 3 Experimental: Trapping and purification o f 1 1 C 0 2 34 3.1 Introduction 34 3.2 Materials and methods 35 3.3 Experimental set-up . ; '. 36 3.4 Experimental procedure 36 3.4.1 Production o f 1 1 C 0 2 . 37 3.4.2 Trapping and release o f 1 1 C 0 2 39 3.5 Results and discussion 40 3.5.1 Effect of carbon molecular sieve mass on 1 1 C 0 2 trapping efficiency 41 3.5.2 Effect of trap temperature on C 0 2 trapping efficiency 42 3.5.3 Overal l trapping efficiency 43 4 Experimental: Reduction of C 0 2 to CO 44 4.1 Introduction 44 4.2 Preliminary experiments: reaction of n C 0 2 and M o 44 4.2.1 Materials and methods 44 4.2.2 Experimental set-up 45 4.2.3 Experimental procedure 46 4.2.4 Results and discussion 46 v 4.3 Equi l ib r ium computer simulations of systems containing molybdenum . . 48 4.3.1 Reduction of M o 0 3 wi th H 2 49 4.3.2 Oxidat ion of M o with 0 2 50 4.4 C o l d experiments: reaction of C 0 2 and M o 54 4.4.1 Introduction 54 4.4.2 Materials and methods 55 4.4.3 Experimental set-up 56 4.4.4 Experimental procedure 57 4.4.5 Results and discussion 59 5 Experimental: Methanol synthesis 65 5.1 Introduction 65 5.2 Materials and methods 66 5.2.1 C u / Z n O catalyst preparation 67 5.3 Experimental procedure 68 5.3.1 Preliminary experiments: C u / Z n O catalyst for C H 3 O H synthesis . . 68 5.3.2 Effect of flowrate and temperature on CH3OH synthesis 69 5.4 Results and discussion: 71 5.4.1 Preliminary experiments: C u / Z n O catalyst 71 5.4.2 Effect of flowrate and temperature on C H 3 O H synthesis 73 5.4.3 Kinet ic M o d e l 76 6 Conclusion and Recommendations 80 6.1 Conclusions 80 6.2 Recommendations for future work 81 Bibliography 83 vi Appendices 91 Appendix I: Flowmeter calibration curves 91 Appendix II: Calibration curves 93 Appendix III: Surface molybdenum 96 Appendix IV: Temperature profile 98 Appendix V: Process simulator 100 Appendix VI: Carbon dioxide estimate 105 v i i List of Tables 3.1 C - l 1 target yields, 14 M e V 40 3.2 n C 0 2 trapping/releasing Sequence 41 3.3 n C 0 2 trapping efficiencies 41 3.4 Comparison of overall trapping efficiency 43 4.1 U C 0 2 Reduction sequence 47 4.2 Summary of u C O yields, 17.5-70 cm 3 /min , 2 bar 47 5.1 Effect of flowrate on conversion of C O to methanol (50bar, 180°C) . . . . 74 5.2 Effect of temperature on conversion of C O to methanol at 55bar, 126 cm 3 /m in (STP) 74 5.3 Kine t ic parameters for Leonov 's model 76 6.1 M o surface moles rough estimate 96 6.2 M o total moles 97 6.3 C 0 2 concentration estimates 105 v i i i List of Figures 1.1 Conceptual visualization of a cyclotron 3 1.2 Interior view of the P E T T R - 1 9 cyclotron 4 1.3 Annih i la t ion of a positron . . 5 1.4 Commerc ia l P E T scanner 6 1.5 1 8 F D G - P E T scan of a breast cancer patient 7 2.1 Separation of air, C O , C H 4 and C 0 2 using gas chromatography 17 2.2 Separation of H 2 , air, C H 4 and C 0 2 using gas chromatography 18 2.3 Decay corrected n C O yields as a function of temperature 28 2.4 C O formation rate from temperature programmed reaction of C 0 2 wi th M o 2 C 29 3.1 1 1 C 0 2 Trapping experimental flow diagram 37 3.2 1 1 C 0 2 activity vs. time, for different trap temperatures 42 4.1 n C 0 2 Reduction experimental flow diagram 45 4.2 Equi l ib r ium compositions of M o - M o O - M o 0 2 - M o 0 3 - H 2 - 0 2 - H 2 0 from 300 to 1500 K 50 4.3 Equi l ib r ium mole fraction of M o 0 2 for different H 2 : M o 0 3 ratios from 300 to 1500K 51 4.4 Equi l ib r ium of composition of M o - M o 0 2 - M o 2 C - C 0 2 - C O 52 4.5 Equ i l ib r ium molar fraction of C 0 2 from eqs. 4.5 53 4.6 Equi l ib r ium molar fraction of C O from eqs. 4.6 53 ix 4.7 Conversion of C 0 2 to C O vs. temperature for different C 0 2 concentration, based on Aspen equil ibrium data 54 4.8 Experimental flow-diagram for the reduction of C 0 2 57 4.9 Effect of temperature on reaction of C 0 2 and M o , at 4 c m 3 / m i n and 2 bar . 60 4.10 Effect of flowrate on reaction of C 0 2 and M o at 8 2 5 ° C and 2 bar 61 4.11 Total carbon oxides as a function of temperature 62 4.12 Total carbon oxides as a function o f flowrate 62 5.1 Experimental flow diagram for methanol synthesis 68 5.2 Aspen generated equil ibrium conversion of 50 ppm C O to methanol, in hydrogen 69 5.3 Methanol produced for continuous feed of H 2 / C O 72 5.4 Semi-batch methanol produced, 180°C, 50 bar for different flowrates . . . 73 5.5 Semi-batch methanol produced, 120 c m 3 / m i n , 50 bar for different temper-atures 75 5.6 Conversion of carbon monoxide to methanol vs. temperature 77 5.7 Conversion of carbon monoxide to methanol vs. flowrate 78 6.1 Mass flow control valve calibration curve 92 6.2 Correlated ball flowmeter calibration curve 92 6.3 Gas chromatogram of C O containing sample 94 6.4 C 0 2 , C O and C H 3 O H calibration curves 95 6.5 Reactor temperature profile, at 7 4 0 ° C 99 6.6 Equi l ib r ium relationships at atmospheric pressure M o - C - M o 2 C - H 2 - C H 4 . . 101 6.7 Aspen generated equil ibrium relationships at atmospheric pressure M o - C -M o 2 C - H 2 - C H 4 102 x Nomenclature a Alpha particle: 2 protons and 2 neutrons

t:x' + p e ( i . i ) } 4 N + p -> I'C + a (1.2) The production of carbon-11 (eq. 1.2) involves the bombardment of nitrogen gas with an incident beam of protons, which creates a very unstable oxygen-15 intermediate [13] containing 8 protons and 7 neutrons. This intermediate instantaneously stabilizes by emit-ting an alpha particle (2 protons, 2 neutrons), to yie ld the radionuclide carbon-11, which has 6 protons and 5 neutrons. C o m m o n radionuclides produced with a medical cyclotron are n C (ti/2 = 20.4 min), 1 3 N (ti/2 = 10 min), 1 50 ( t 1 / 2 = 2 min) and 1 8 F ( t 1 / 2 = 110 min). These radionuclides decay by pure positron emission, as illustrated by the decay o f carbon-11 in eq. 1.3. The half life is a characteristic constant for a given radioisotopes and corresponds to the time 4 required for half of a radioactive sample to decay [14]. nC^l°B + P+ (1.3) In addition to their short half lives, the principal positron emitters n C , 1 3 N , 1 5 0 and 1 8 F are useful for clinical PET since their stable isotopes are the building blocks for most organic molecules ( 1 8 F is used in place of H or OH). The radionuclides are chemically processed to prepare a radiopharmaceutical that wil l target the desired metabolic process or tissue in the body. P E T C a m e r a s Positrons ((3+) travel at most a few millimeters before they are annihilated with nearby electrons to give two gamma rays of 511 keV emitted in opposite directions (Figure 1.3). Before • + electron positron Best mass of each is 0 51 MeV Figure 1.3: Annihilation of a positron The PET camera uses rings of detectors that register the 511 keV rays from a single annihilation. The detectors are arranged in a circle around the patient, or in a hexagonal array where the detectors are grouped into cassettes. The content of each cassette varies 5 0.51 MeV Gamma photon 0.51 MeV Gamma photon according to the manufacturer, but usually these consist of 256 scintillating crystals being viewed by 16 photomultiplier tubes (four blocks composed of 64 crystals and four photo-multiplier tubes) wi th as many as three rings of 16 buckets in a circle (12,288 crystals with 768 photomultiplier tubes) [15]. A n alternative design to the bucket, is one that uses rings of detectors with 11 crystals in a staggered array being exposed to 6 photomultiplier tubes [15]. A s can be seen in Figure 1.4, once the patient is placed inside the scanner, he/she slowly moves through the rings of detectors to provide a whole or partial body scan. Figure 1.4: Commerc ia l P E T scanner W h e n two detectors at opposite sides of the camera simultaneously detect a gamma ray, the data processor assumes that on the virtual line connecting the two detectors a positron was annihilated. Once sufficient data is collected, a 3-dimensional image o f the radiation distribution can be generated. For example, Figure 1.5 shows the 1 8 F D G uptake in a patient wi th breast cancer, revealing that numerous active metastatic cancer sites are present. 6 Courtesy Dr. F. Benard Figure 1.5: 1 8 F D G - P E T scan of a breast cancer patient 1.1.2 Application of PET P E T is currently used c l in ica l ly in cardiology, neurology, and oncology. The most common radiopharmaceutical used in P E T is [ 1 8 F]fluoro-deoxy-glucose ( F D G ) [16], which allows the measurement of glucose consumption in vivo and detects organs and tissues wi th high metabolic activity. Applicat ions of P E T in cardiology also include measurement o f blood flow wi th 1 3 N H 3 and/or [ n C]acetate , in addition to the studies on the glucose consumptions of the heart with 1 8 F D G [17]. Applicat ions in neuroscience include glucose metabolism with 1 8 F D G , protein synthesis rate with [ n C] ty ros ine and blood flow with H 2 l s O [18]. In oncology, 1 8 F D G - P E T is a major tool for diagnostic and treatment follow-up. In addition to F D G , [ J ^ m e t h i o n i n e , [ n C ] t h y m i n e , [ J ^ c y t o s t a t i c s are used for imaging of more specific cancers. Positron emitter labeling with carbon-11, nitrogen-13 and oxygen-15, can be used to investigate the catalytic processes and to quantify reaction phenomenon, yie lding insight in 7 elementary reaction mechanisms useful as input to mathematical simulations [19]. 1.1.3 Carbon-11 radiopharmaceuticals Carbon-11 radiopharmaceuticals are finding an increasing number of applications in car-diology, neurology, and oncology [18]. M a n y applications, such as brain imaging, require high specific activity radiopharmaceuticals. For example, [ n C]rac lopr ide , which is used to visualize dopamine receptors in the brain, requires a high specific activity since only a small number of dopamine receptors are available. L o w specific activity, i.e. the presence of many [ 1 2 C]raclopride molecules, may result in receptor saturation and poor image qual-ity. M a n y [ n C]radiopharmaceuticals are synthesized by the methylation of an appropriate precursor with [ x ^ m e t h y l iodide, which in turn is generally prepared from [ U C]methano l . Production of carbon-11 radiopharmaceuticals Carbon-11 is produced in a cyclotron by the bombardment of a gas target filled wi th n i -trogen, using a proton beam (typically 10-19 MeV), v ia the 1 4 N ( p , a ) n C nuclear reaction. In the presence of 0 2 , the carbon-11 is produced as n C 0 2 , whi le in the presence of H 2 , n C H 4 is produced. A sequence of chemical reactions is carried out to convert these pre-cursors to a carbon-11 radiopharmaceutical, typically v ia the [ n C ] C H 3 O H and [ n C ] C H 3 I intermediates. The conventional synthesis of [ n C ] m e t h y l iodide most widely used over the past 30 years [20] involves two main reactions: the reduction of n C 0 2 to n C H 3 O H , fol lowed by iodination to y ie ld n C H 3 I (eq. 1.4) [20]. LiAlHA HI nC02 -> nCH3OH _> nCH3I (1.4) In the l iquid phase synthesis, n C 0 2 is bubbled through a solution of l i th ium aluminum 8 hydride (L iA lH 4 ) for the production of u C H 3 O H , which is then treated with HI to yield n C H 3 I . The L i A l H 4 is a strong reducing agent and is readily contaminated with C 0 2 from the atmosphere during preparation of the reagent. Consequently, relatively low specific activity [ nC]methyl iodide is generally produced. The quantity of L i A l H 4 utilized in the reaction has a great influence on the yield and specific activity of n C H 3 O H [13]. Large quantities of L i A l H 4 in the reactor give rise to high yields with low specific activity. On the other hand, small amounts of L i A l H 4 lead to low yields but with higher specific activity. Thus, the main drawback of this method is that a sufficient yield cannot be obtained while simultaneously achieving a high specific activity. Alternative methods, include the gas phase production of [ n C ] methyl iodide by react-ing U C H 4 with I2[21, 22]. This synthesis route avoids the use of L i A l H 4 and consists of recirculating n C H 4 with iodine vapor through a hot reactor. 1.2 Objective of the study The conventional method to produce many carbon-11 radiopharmaceuticals involves the L i A l H 4 reduction of U C 0 2 and is not suitable for production of high specific activity ra-diopharmaceuticals. The main objective of this study was to determine the feasibility of an alternative gas phase process involving catalytic methanol synthesis, which providing a process for the production of large quantities of high specific activity carbon-11 radiophar-maceuticals. The proposed gas phase reaction route consists of the reduction of C 0 2 to C O followed by catalytic reduction of C O to C H 3 O H . The steps leading to methanol synthesis wil l be developed and optimized. Detailed objectives of the present study include: 1. To design / implement / optimize performance of C 0 2 trapping device 2. To design / implement / optimize performance of reduction of C 0 2 t o C O 9 3. To design / implement / optimize performance of reduction of CO to CH 3 OH 4. To develop an empirical rate equation for methanol synthesis to predict performance using carbon-11 10 Chapter 2 Literature review This chapter is divided in three sections. The first section deals wi th the radiochemistry of carbon-11. The second section provides in depth information of the reactions of carbon oxides and molybdenum compounds. The third section includes history and background information on industrial methanol synthesis. 2.1 Radiochemistry of carbon-11 Since the early 1970's, when medical cyclotrons where introduced, an increasing number of compounds are labeled with the positron emitting radionuclide carbon-11 [23]. The most widely used method to produce carbon-11 radiopharmaceuticals involves the methy-lation of an appropriate precursor wi th [ n C ] m e t h y l iodide [2]. Based on this method, a large number of receptor-ligands have been investigated [24]. These radiopharmaceuticals require a high specific activity and it is thus important that each step in their production minimizes contamination with carrier carbon-12. The conventional method for the preparation of I1 methyl iodide is the reduction of [ 1 1 C]carbon dioxide to [ n C ] m e t h a n o l by l i thium aluminum hydride, fol lowed by conver-sion to [ n C ] m e t h y l iodide with hydroiodic acid [25]. A n alternative method used by many P E T centers involves a recently developed gas phase method, which consists of reacting 11 [nC]methane with iodine vapor to give [uC]methyl iodide [21, 22, 26]. Throughout this work the terms yield and decay corrected yield are frequently used. The yield is defined as the ratio of the amount of radioactive product to the amount of radioactive starting material, as shown in eq. 2.1. Although synthesis time does not show up directly in this equation, it is incorporated in the yield since a certain amount of C-11 is lost due to decay during the synthesis. However, in order to quantify the efficiency of the various steps in the process, it is desirable to calculate the decay corrected yield, also referred to as the radiochemical yield. It is calculated by multiplying the yield by a decay correction factor, which depends on the time elapsed and the half life of the radionuclide in question, as shown in eq.2.2. The decay corrected yield refers to the yield assuming no radioactive decay occurred, and is also referred to as the chemical yield. Amount Product (mCi) ^ ^ Amount of Reagent (mCi) y' = y e x p 0 - 6 9 3 ' ' / ' ^ (2.2) For example, consider a process which has a 30% yield ( U C 0 2 to 1 1 CH 3 OH) with synthesis time of 20.4 minutes. Since the half life of carbon-11 is 20.4 minutes, the decay corrected yield is 60%. Thus the actual chemical conversion of [nC]carbon dioxide to [nC]methanol is 60%, but since half of the carbon-11 decayed, the actual yield is only 30%. 2.1.1 Production of C - l l Although several nuclear reactions can be used to produce carbon-11, the most convenient is the 1 4 N(p,a) n C reaction using natural occurring nitrogen gas. Carbon-11 can be pro-duced in the gas target as n C 0 2 or n C H 4 ; in the case of n C 0 2 0.1-2% oxygen is added to the nitrogen, while for n C H 4 production 5-10% hydrogen is added to the nitrogen gas. The natural abundance of C 0 2 in the air is 330 ppm, whereas that of methane is 1.6 ppm. 12 This means that much precaution must be taken to exclude air from synthesis modules and reagents during synthesis with n C 0 2 . However, specific activities of n C 0 2 and n C H 4 produced in target are reported to be similar. For in target production of n C H 4 , a specific activity of 5 Ci/u-mol was reported [27, 28]. For in target production of n C 0 2 , specific activities up to 16.5 Ci/iamol were reported [29]. A n independent study demonstrated that only a negligible amount of carrier 1 2 C 0 2 originates from the target [24]. [ n C ] M e t h a n e production yields are typical ly less than 65% of [ n C ] c a r b o n dioxide yields [28]. These factors are among the reasons that many P E T centers routinely produce 1 1 C 0 2 as opposed to n C H 4 . Target for C- l l production Pure aluminum, aluminum alloys and stainless steel are suitable materials for target con-struction. H a v a r ® , titanium or stainless steel are suitable materials for the target window foils. Meta l gaskets are preferable to rubber O-rings [13]. In order to reduce potential car-rier carbon and increase specific activity, ultra-pure materials wi th very low carbon content should be used. Thus, ultra pure aluminum would be suitable as target body and ultra pure titanium foi l for the target window. For good recovery of C - l 1 from the target, it is recom-mended that newly constructed targets be carefully washed with phosphoric acid fol lowed by water, and dried under vacuum. The target gas should be of high purity and in particular as free as possible from carbon-containing impurities. A n additional precaution against [ L 2 C]carbon dioxide contamination would be to introduce a carbon dioxide trap in between the nitrogen gas cylinder and the target and also to use stainless steel tubing to minimize entry of carrier 1 2 C 0 2 by diffusion [13]. n C Q 2 Produced(EOB) y U C ° 2 " I- (1 -exp (-0.693-t / t 1 / 2 )) ( 2 3 ) 13 An important characteristic for a target is the n C0 2 yield at the end of bombardment (EOB). As can be seen in eq. 2.3, the n C0 2 yield depends on the amount of n C0 2 pro-duced in the target, the half life of C-l 1, the beam intensity (I) and the duration of irradi-ation (t). For an energy beam of 14 MeV, C-ll experimental target yields vary between 93 and 135 u-Ci/uA [30, 31, 32]. The theoretical target yield, calculated from 14N(p,n)nC excitation function, was determined to be 169 (a.Ci/(J.A at 14 MeV [31]. As can be seen from this data, the reported experimental yields are at least 25% less than the theoretical yield. 2.1.2 Recovery and purification of N C 0 2 and n C H 4 The target gas is either a mixture of N2/ H2 or of N2/ 02. The short half life of C-11 makes it necessary to concentrate it rapidly from a large volume of target gas in order to prepare the C-ll radiopharmaceutical [33]. In addition, the carbon-11 produced in the target as n C0 2 or n C H 4 must be separated from the target gas, in order to remove the 02 or the H2, which can affect downstream processing of the C-ll radiopharmaceutical [34]. Two common methods exist for recovery of C-l 1, which are the use of a cryogenic trap and/or the use of an adsorbent trap (such as molecular sieve) [13]. The performance or overall trapping efficiency (£T) depends on the amount of C-l 1 released from the target, the amount of C-l 1 released from the trap and the time required for this process. The trapping efficiency ( £ T 0 is defined as the ratio of the amount of C-l 1 trapped to the amount of C-l 1 entering the trap, as shown in eq. 2.4 and indicates how well the trap retains the n C0 2 . The overall trapping efficiency is defined as the ratio of the amount of C-ll recovered from the trap to the amount of C-l 1 released from the target, as shown in eq. 2.5. The overall trapping efficiency can also be related to the trapping efficiency and the time required for trapping 14 and releasing the C - l 1, as shown in eq. 2.6. t f = C - l l trapped (mCi) ' T C - l l entering trap (mCi) C - l l released from trap (mCi) C - l l released from target (mCi) £ T = exp-°-693'tT/tl/2 (2.6) Cryogenic n C 0 2 trap: The trap typically consist of a small stainless steel tube, (40 cm x d> 0.02 mm). The tube is ini t ial ly immersed in l iqu id nitrogen or l iqu id argon. The target gas is then flushed through the cold trap with an inert carrier gas such a hel ium, thus trapping the [ u C ] c a r b o n dioxide. The [ n C ] c a r b o n dioxide is recovered simply by passing a slow stream of inert carrier gas while heating the trap to room temperature. This requires removing the trap from the l iqu id nitrogen bath. Cryogenic n C 0 2 traps typically have a trapping efficiency of 90% using a stainless steel tube immersed in l iqu id nitrogen or argon with overall trapping/releasing time of approximately 5 minutes [13]. This corresponds to an overall trapping efficiency of 76%. This method has been further improved, by the use of a stainless steel frit trap instead of a tube. The trap consist of a 2.25 cm long stainless steel check valve cartridge (for | tubing) which is packed with eight 20 urn stainless steel frits. The trapping efficiency was improved to 96% and overall trapping/releasing time was reduced to 3.9 minutes [35]. This corresponds to an overall trapping efficiency of 84%. Molecular sieve U C 0 2 trap: The basic principle behind the molecular sieve trap is chromatography. Chromatography is a method of separation that relies on differences in partitioning behavior between a flowing 15 mobile phase and a stationary phase to separate the components in a mixture. Molecular sieves have been used as packing for chromatographic recovery of nCG*2 [13]. The trap typically consists of a column packed with zeolites molecular sieves. The target gas is un-loaded through the trap, then flushed with inert gas, and heated to above 200°C for thermal desorption of the n C 0 2 . For a column of the following dimensions, I = 6cm, i.d. = 9mm, packed with activated 60-80 mesh molecular sieve 4A, the trapping efficiency is above 98% [13]. With an overall trapping/releasing time of approximately 5 min, the trap releases 90% of the carbon dioxide when heated to 230°C. This corresponds to an overall trapping effi-ciency of 74%. The tenacious affinities of zeolites for both CO2 and water require careful activation be-fore use and rather high thermal desorption temperatures. These inherent characteristics of zeolites lead to the development of an improved molecular sieve trap, with the use of carbon molecular sieves. In contrast to zeolites based materials, carbon sieves are quite hy-drophobic in nature and have little affinity for water and small polar molecules. However, they avidly and selectively retain CO2 then release it freely upon the application of modest heat. This is clearly illustrated when comparing the chromatograms of a carbon molecular sieve column at room temperature and at 140°C, as shown in Figure 2.1. A similar trapping procedure has been used, with a 65 cm long trap, 2 mm i.d., packed with 1 g of 80-100 mesh carbon molecular sieve. Trapping efficiency was essentially 100%, and 98% of the n C 0 2 was released from the trap heated to 100°C, with an overall trap/release time of 5 minutes [34]. This corresponds to an overall trapping efficiency of 83%. Control experi-ments showed that most of the carbon sieves tested did not significantly contribute 12C02 carrier to the final product [34]. 16 0 2 4 15 Min. 0 2 4 Min. A ) 1. 0 2 , 2. N 2 , 3. C O , 4. C H 4 B ) 1. A i r , 2. C O , 3. C H 4 , 4. C 0 2 2 0 ° C , 45 cm 3/min He 140°C, 45 cm 3/min N 2 Chromatogram, using carbon molecular sieve column, obtained from Al l t ech Inc. [36] Figure 2.1: Separation of air, C O , C H 4 and C 0 2 using gas chromatography Cryogenic/chromatographic n C H 4 trap In order to increase the retention time of n C H 4 in a chromatographic column and obtain separation from target gas, a combination of cryogenic trapping and chromatography are used. The trap consists of a tube packed with Poropak N (80-100 mesh). The trap is ini t ial ly submerged in l iqu id nitrogen, after which the target is unloaded under controlled flow. A s can be seen in Figure 2.2, heating to room temperature is sufficient to release the [ n C]methane . Trapping methane on Poropak N at - 1 9 6 ° C is less efficient than trapping car-bon dioxide on molecular sieves [22], wi th overall trapping releasing time approximately 6.0 minutes. Assuming that the trapping efficiency is 90%, this would correspond to an overall trapping efficiency of 73%. 2.1.3 Reduction of n C 0 2 and oxidation of n C H 4 Reduction of U C 0 2 to n C O [ n C ] C a r b o n monoxide was one of the first tracers used for blood flow measurement in hu-mans [37]. Since it is less reactive than other [ n C ] l a b e l i n g agents, it has found little appli-17 T = 35°C, 30 c m 3 / m i n He, 1. H 2 , 2. Air, 3. C H 4 , 4. C 0 2 Chromatogram, using poropak Q column, obtained from Alltech Inc. [36] Figure 2.2: Separation of H 2 , air, C H 4 and C 0 2 using gas chromatography cation as an intermediate in radiopharmaceutical syntheses. However, recently palladium catalyzed cross-coupling reactions for the direct preparation of ketones have been reported [38, 39]. [ n C]Carbon monoxide can be produced in situ or by reduction of [ 1 1C]carbon dioxide on the surface of various transition metal elements. A n established method for [ n C]carbon monoxide preparation is the reaction of [ n C]carbon dioxide with zinc metal, as illustrated in eq. 2.7. Highest yields are obtained at 400°C which 4 is very near the melting point of zinc (420°C), thus rigorous temperature control is required to avoid melting the zinc. The zinc granules must be thoroughly cleaned before use and require frequent replenishment. Yields are under 50% since a significant amount of ra-dioactivity remains in the reactor [37]. A more recent method for [ 1 1C]carbon monoxide preparation is the reaction of [ 1 1C]carbon dioxide with molybdenum metal (eq. 2.8). The reactor consisted of a 25 m m inner diameter quartz tube and 150 m m long, packed with 2000 m of 0.05 m m diameter molybdenum wire (99.97% Mo). A helium stream of 25 c m 3 / m i n (STP) was used to transfer the [ n C]carbon dioxide through the molybdenum reactor, which was kept at a preset temperature. Decay corrected yields of up to 81% were reported with reactor temperature of 800° C [40], with 18 a gas transfer and reaction time of approximately 3 minutes, resulting in a yield of 73%. n C 0 2 + Z n ( s ) - n C O + Z n O ( s ) (2.7) 400°C 2 1 1 C 0 2 + M o ( s ) _> 2 n C O + M o 0 2 ( s ) (2.8) 800° C Reduction of n C 0 2 to n C H 4 The reduction of U C 0 2 to n C H 4 has been carried out for several applications, includ-ing preparation of [nC]methyl iodide, [nC]methyl triflate and [nC]cyanide [2]. Although [uC]methane can be produced in target, many PET centers choose to first produce [nC]carbon dioxide and then react it with hydrogen using a nickel catalyst to produce [ n C] methane, as illustrated in eq. 2.9. Ni n C 0 2 + 4 H 2 _ , n C H 4 + 2 H 2 0 (2.9) 450° C In gas chromatography, flame ionization detectors (FID) are not very sensitive to trace amounts of carbon oxides, but are much more sensitive to methane. Accordingly, nickel based catalyst are frequently used to convert trace amounts of carbon oxides to methane in order to improve the FID sensitivity. In a unique setup, [uC]carbon dioxide is reduced to [nC]methane for production of [nC]methyl triflate. The reactor consists of a nickel/alumina/silica (64% Ni) powder mixed with glass wool packed into a 4 mm inner diameter borosilicate glass tube. The purified 19 [nC]carbon dioxide is swept from the trap with 10% hydrogen in nitrogen. The reactor oven is held at 450°C and the gas flowrate is approximately 50 cm 3/min (STP). Yields of [nC]carbon dioxide to [nC]methane routinely exceed 95% [41] in approximately 2 minutes, corresponding to a decay corrected yield of over 99%. 2.1.4 Synthesis of n C H 3 O H The established method to prepare [nC]methanol involves the reduction of U C 0 2 to n C H 3 O H using LiAlH 4( a q). This method produces low specific activity n C H 3 O H due to contamina-tion of LiAlH 4 with 1 2 C 0 2 . To improve the specific activities of n C H 3 O H , alternate liquid phase and gas phase methanol synthesis routes have been investigated. Liquid phase n C H 3 O H synthesis The conventional process for the reduction of [nC]carbon dioxide to [uC]methanol uses a solution of L iAlH 4 in tetrahydrofuran (THF), as shown in eq. 2.10. The n C 0 2 recov-ery from the trap is done with a gas stream of helium or nitrogen at flow rate of 10-100 cm 3/min (STP). The gas is dried over a dehydrating agent (MgC104 • H 2 0 or P 2 0 5 ) and then the gas is bubbled through a solution of LiAlH 4 in THF The solvent is then evaporated under a stream of nitrogen at approximately 100°C, leaving behind a dry radioactive com-plex from which [nC]methanol is generated by the addition of water. The [nC]methanol can be distilled for downstream use or alternatively hydroiodic acid can be added for in-vial preparation of [nC]methyl iodide. An increase in the amount of L iAlH 4 leads to an increase in the yield, but a decrease in the specific activity. While a decrease in the amount of L iAlH 4 leads to an increase in the specific activity but a decrease in the yield. The prepa-ration of methanol typically takes 3-4 minutes with decay corrected yield in the order of 75%, and corresponding yield of approximately 66% [1]. Reported specific activities for [nC]methyl iodide produced via this procedure range from 0.1-1.7 Ci/u.mol [1, 13]. In attempts to reduce the amount of LiAlH 4 , a variation of the conventional method has 20 been developed, using L i A l H 4 adsorbed on alumina. The alumina cartridge is pretreated at 2 0 0 ° C under helium flow, for 1 hour, to desorb water and possible traces of carbon dioxide and then cooled to room temperature. It is then impregnated with 50 u l of 1 M L i A l H 4 diluted with 200 u.1 diethyl ether. The [ u C ]carbon dioxide is trapped on the cartridge, after which it is heated up to 160°C under a 50 c m 3 / m i n flow of helium to remove the solvent. The dry complex is then hydrolyzed by injecting 0.01 M phosphoric acid to form [ 1 1C]methanol. Decay corrected yields of 95% are reported with [ n C ]methanol preparation time of approximately 5 minutes, which corresponds to a y ie ld of 80%. The specific activity o f the downstream produced [ n C ]me thy l iodide has been reported to be 2-2.5 C i / u - m o l [42, 13]. l ) L i A l H 4 l T H F n C 0 2 _> nCH3OH 2 ) H 2 0 H20 n C 0 2 ( g ) C H 3 0 ~ ( a q ) —> C H 3 O H Gas phase nCH3OH synthesis Only a few publications report the catalytic conversion of [ u C ]carbon dioxide to [ n C ]methanol according to eq. 2.12. C u / Z n O / A l 2 0 3 n C 0 2 + 3 H 2 _> n C H 3 O H + H 2 0 (2.12) 1 5 0 - 2 5 0 ° C (2.10) (2.11) 21 J.T. Patt studied the synthesis of [ n C]me thano l from [ n C ] c a r b o n oxides using Pd/Al 2C>3 and C u / Z n 0 7 A l 2 0 3 catalysts [43]. Us ing either catalyst, and a mixture of u C O and/or n C 0 2 , A r , H 2 as feed gas, produced negligible amounts of methanol at 200-240° C and at 2-3 bar. The highest methanol yie ld obtained was 7%, with synthesis time o f 25 min , at 240°C and at 2 bar, by adding N 2 0 to the feed stream [43]. A patent describes a pre-conditioning method of a catalyst prior to use for [ n C ] m e t h a n o l synthesis [44]. A copper-zinc oxide catalyst, supported on alumina and/or si l ica, is first re-duced and then preconditioned with a stream of C O : C 0 2 : H 2 (1:1:8). Prior to [ 1 1 C]methanol synthesis, the catalyst is heated to 200°C and then gas containing [ n C ] c a r b o n dioxide and hydrogen is passed through the reactor at a pressure of 50 bar. The [ n C ] m e t h a n o l may remain adsorbed on the catalyst, and can be removed by addition of a catalyst poison, such as hydrogen sulphide or by increasing the temperature of the reactor to 280-320°C. The y ie ld of [ n C]me thano l is 57% and the specific activity ranging from 4-20 C i / u m o l [44]. A n alternative method uses 1 1CH4 as starting material instead of 1 1 C 0 2 . A mixture of n C H 4 , C l 2 and H 2 is passed over C r 2 0 3 on pumice stone at 700°C in order to oxidize the n C H 4 to n C H 3 O H with yields up to 4 5 % [45]. Synthesis time is 2 minutes decay corrected yie ld of 48%. The specific activity of the subsequently prepared n C H 3 I was 1 C i / u - m o l [45]. 2.1.5 Synthesis of carbon-11 radiopharmaceuticals Most procedures to synthesize carbon-11 radiopharmaceuticals involve the methylation of an appropriate precursor. The reactivity of n C H 3 O H is insufficient for most applications and accordingly the n C H 3 0 H needs to be converted to a more reactive intermediate, such as n C H 3 I . The preparation of [ n C ] m e t h y l iodide is done by reacting [ n C ] m e t h a n o l with a source of iodine, a step that does not affect the specific activity of the product. Traditionally, [ n C ] m e t h y l iodide has been prepared by reaction of [ n C]me thano l with hydrogen iodide under reflux [13]. The yie ld is above 90% [45, 42]. A more recent vari-22 ation of this reaction route involves the use of aqueous H I impregnated on alumina, for which the yie ld was above 97% at optimal conditions [42]. Alternative iodination agents, diphosphorous tetra-iodide and triphenylphosphine diiodide, have been investigated for the production of [ n C ] m e t h y l iodide[46, 47]. The yields are similar as for the H I procedure, however by avoiding the use of volatile H I , the solid reagents al low a cleaner operation. A n alternative procedure, based on the iodination of [ n C ] methane with iodine, has been pioneered by Larsen et al . [22, 21]. The reaction is carried out in a reactor in which [ n C]methane , helium and iodine vapors are mixed and heated. The formed [ n C ] m e t h y l iodide is continuously removed from the reactor whi le the unreacted [ n C]methane is re-circulated through the reactor. The synthesis time is 10.5 minutes and reported yields are 58% wi th specific activity of 15 C i / ixmol . 2.2 Molybdenum compounds for the reduction of CO2 to CO Molybdenum commonly occurs in nature as the mineral molybdenite, M0S2, in quartz rock. For this study molybdenum w i l l be used to reduce carbon dioxide to carbon monoxide, by oxid iz ing the molybdenum. In this section, the history and chemistry of molybdenum, molybdenum oxides and molybdenum carbides are presented. 2.2.1 History Molybdenum was discovered in 1778, but for the next hundred years, molybdenite was merely a laboratory curiosity. The first major use came during Wor ld War I when it was dis-covered that addition of molybdenum produced steels with excellent toughness and strength at high temperatures, suitable for use as tank armor and in aircraft engines [48]. Molybdenum is mainly used as an al loying element in steel, cast iron, and super-alloys to increase hardenability, strength, toughness, and corrosion resistance. However, 23 it has found many other applications in l ighting, electronics, vacuum coating and nuclear medicine. It is also extensively used as a catalyst or as a component thereof [48]. 2.2.2 Reduction of molybdenum oxides Reduction of molybdenum oxides with hydrogen Molybdenum metal powder is produced industrially by reducing high-purity molybdenum compounds, such as molybdenum trioxide, ammonium hexamolybdate or ammonium d i -molybdate, wi th hydrogen [48]. Molybdenum metal powder can be produced industrially by reducing M0O3 powder with hydrogen between 500-1150° C [49]. Molybdenum tri-oxide, a gray-green powder is reduced by hydrogen at 500-600°C to M0O2, which is fur-ther reduced at 900-1050°C to molybdenum metal. Since the first reduction to M o 0 2 is exothermic, this step is performed at 600°C to prevent caking due to the melting of M0O3 ( 800° C). The red-brown M o 0 2 [50] is reduced to metallic molybdenum powder at 1050°C. The powder has a particle size of 2-10 u.m, a specific surface area of 0.1-1 m 2 /g, and an oxygen content of 100-500 mg/kg (partly adsorbed and partly as oxide) [48]. Thermal decomposition and reduction of molybdenum trioxide under different reduc-ing conditions has been extensively studied [51, 52, 53]. Molybdenum trioxide powder was reduced in pure hydrogen with gradual temperature increase from 300 to 800°C at approximately 6 °C/min and isothermally at 600°C. Under these conditions, the reduc-tion of molybdenum trioxide to molybdenum dioxide took place at 387-615°C whi le the reduction to molybdenum metal took place slowly at 623-740°C[51] . Lee et al . [54] performed temperature programed reduction (TPR) of high purity molybdenum trioxide, from 300 to 750 °C, in pure hydrogen, with a heating rate of 1 °C /min . The gas products obtained during T P R were monitored by gas chromatography, equipped with thermal conductivity detector for the detection of water. Reduction of M0O3 to M o 0 2 occurred between 430-620°C, while complete reduction to molybdenum occurred above 700°C. 24 Iwasawa et al . [55] found that molybdenum dioxide on alumina was reduced with hydrogen to molybdenum at 6 0 0 ° C . In contrast to the one-step reduction mechanism of molybdenum trioxide to molybdenum dioxide, Burch [56] suggested that M04O11 is an in-termediate product of the reaction. However, Ressler et al . [53] found that M o 4 O n was also formed by reaction of different molybdenum oxides. Temperature programmed reduction of M 0 O 3 wi th 5-100% H 2 was studied for temperatures ranging from 3 0 0 - 8 0 0 ° C . Between 350 and 4 2 5 ° C , the reduction of M 0 O 3 to M o 0 2 is a one-step process without formation o f crystalline intermediates. A t temperatures above 4 5 0 ° C , M o 4 O n can be obtained and its formation was explained as the product of a parallel reaction between molybdenum diox-ide and molybdenum trioxide. There is a general agreement between various researchers that at reduction temperatures above 5 0 0 ° C and hydrogen concentrations of at least 10%, metallic molybdenum is produced as the final product according to a two step reduction, as illustrated in eqs. 4.2 and 4.3 [53]. Reduction of molybdenum oxides with carbon monoxide Reduction of molybdenum trioxide with carbon monoxide at 400° C gave carbon dioxide, the reddish-brown molybdenum dioxide and unreacted molybdenum trioxide [57]. The reduction was presumed to proceed according to the eq. 2.13. A t higher temperature (565°C) , molybdenum trioxide in a stream of carbon monoxide gave carbon dioxide and a dark-gray almost black material, which was assumed to be molybdenum carbide and free carbon. The overall reaction is illustrated in eq. 2.15, which proceeds through intermediate eqs. 2.13 and 2.14. Further experiments were carried out wi th carbon dioxide (19%) and carbon monoxide (81 %) and molybdenum dioxide at roughly 8 0 0 ° C , which yielded grayish molybdenum carbide, presumably according to the reversible eq. 2.14. The addition of carbon dioxide was to prevent build-up of carbon on the surface, which was confirmed by 25 the color and the carbon content of the molybdenum carbide [57]. 4 0 0 ° C C O + M o O ; '3 C 0 2 + M o 0 2 (2.13) 565°C 6 C O + 2 M o 0 2 5 C 0 2 + M o 2 C (2.14) 565°C 8 C O + 2 M o O ; '3 7 C 0 2 + M o 2 C (2.15) Hexagonal Mo 2 C is the only molybdenum-carbide phase of commercial interest and is the only phase that is stable below 1100°C [58]. It is produced as a powder in the micron range, with colors varying from white-gray to black [48, 59, 60]. It is stable in hydrogen, but it oxidizes in air above 5 0 0 ° C [48]. 2.2.3 Oxidation of molybdenum Oxidation of molybdenum with oxygen gas Molybdenum retains its luster almost indefinitely in air, particularly when it has been drawn to fine wire. On prolonged heating in air below 6 0 0 ° C , the metal becomes covered with its trioxide; at 6 0 0 ° C the oxide sublimes and rapid oxidation occurs. Molybdenum burns in oxygen at 5 0 0 - 6 0 0 ° C [48]. Oxidation of molybdenum with carbon dioxide Vandenberghe reported that carbon dioxide reacts with molybdenum to form carbon monox-ide and molybdenum trioxide above 7 0 0 ° C [61]. Spencer et al. [57] later pointed out that 26 these authors made no analysis of the solid product, but assumed that the molybdenum was oxidized to molybdenum trioxide. The progression of the oxidation was monitored by observations of the appearance of the solid phase. The colors suggested that the oxidation was carried only to the dioxide, according to eq. 2.16 [57, 61]. 700° C Mo + 2 C 0 2 — M0O2 + 2CO (2.16) Hilpert et al . [62] indicated that at 1000°C , carbon monoxide passed over finely divided molybdenum leads to dimolybdenum carbide. A t 8 0 0 ° C they obtained a carbide of high carbon content by reacting carbon monoxide with molybdenum trioxide. Smith et al . [63] previously observed no reaction when molybdenum was heated to a over 1000°C in an atmosphere of carbon monoxide. L i a n et al. [64] demonstrated that at room temperature and at 8 0 ° C , H 2 , N 2 , C H 4 and C O showed no reactivity with molybdenum atoms, whi le 0 2 and C 0 2 both reacted with molybdenum. Th in molybdenum wire is reported to react with n C 0 2 between 700 and 1000°C for the formation of n C O , and it was concluded to proceed according to eq. 2.8. Chemica l conversions of up to to 80% were reported, over a narrow temperature range, 8 4 0 - 8 6 0 ° C , as illustrated in figure 2.3. A t 8 2 5 ° C n C 0 2 decay corrected yie ld was 35% and decreased with decreasing temperature, while at 8 7 5 ° C , n C 0 2 decay corrected y ie ld was 4 5 % and decreased with increasing temperature. 2.2.4 Interaction of C 0 2 and CO with Mo 2 C M02C is used in special cemented carbide grades containing T i C and nickel metal. Mos t M o 2 C is used in steel alloys, where it forms during melting. Al though M0O3 or M0O2 can be carburized with carbon at 1500°C, a carbide with the correct carbon content and a 27 90 -Temperature (*C) Radiochemical yie ld , i.e. decay corrected to E O B , from Zeilser et al . [40] Figure 2.3: Decay corrected n C O yields as a function of temperature low oxygen content is difficult to obtain. Pure M o 2 C is best made by heating molybdenum metal powder with carbon in hydrogen at 1500°C [65]. Molybdenum carbide is also used extensively as a catalyst, such as for aromatization of ethane, ethylene, propane and in the oxidative dehydrogenation of these compounds using carbon dioxide as an oxidant. The catalyst is prepared by heating molybdenum trioxide in a stream of methane and hydrogen, from 500 to 8 0 0 ° C [54]. Dur ing preparation, surface contamination by carbon generally occurs. The latter can be removed by treatment with hydrogen under controlled environment. Once the excess carbon was completely removed, it was observed that the B E T surface area of the catalyst and C O chemisorption were high-est. For a catalyst treated with hydrogen, the surface area was found to be 60 m 2/g, wi th a C O uptake at room temperature of 220 u.mol/g [54]. Lee et al . [54] showed that the molybdenum metal has a much lower surface area than the M o 2 C catalyst. The M o 2 C had a B E T surface area o f 60-100 m 2/g whi le T P R of M0O3 with hydrogen yielded molybdenum with B E T surface area of 3 m 2/g. Prior 28 500 600. 700.. 800 900 1000 Temperature [KJ F r o m Solymosi et al . [66] Figure 2.4: C O formation rate from temperature programmed reaction of C 0 2 wi th M o 2 C to removal from the reactor, the dimolybdenum carbide was passivated with 1% oxygen at room temperature [54]. Molybdenum can be passivated by oxidation, especially by electrolytic oxidation, becoming chemically unreactive [48]. Iwasawa et al . [55] observed deposition of small amounts of carbon on molybdenum fixed on alumina, after it had been reduced with hydrogen at 5 0 0 ° C for 5 h. Solymosi et al . [66] studied the reaction between carbon dioxide and supported M o 2 C . Temperature programmed reaction of carbon dioxide with supported M o 2 C forms carbon monoxide, as illustrated in figure 2.4. Carbon monoxide was first detected at 3 0 0 ° C , and a more extensive decomposition of carbon dioxide to carbon monoxide occurred above 6 0 0 ° C [66]. Us ing 1 3 C 0 2 as supply gas, it was demonstrated that over 90% of the C O was formed from the supply gas and not from the carbide [66]. It was demonstrated that the C O comes mainly from decomposition of C 0 2 , according to eq. 2.17, and that the contri-bution of the reaction of carbon in M o 2 C with the O atom formed in the C 0 2 dissociation, according to eq. 2.18, is l imited [66]. 29 C 0 2 ^ C 0 ( g ) + 0 ( a ) (2.17) C ( 8 ) + 0 ( a ) ^ C 0 ( g ) (2.18) The reaction between 25% C 0 2 and M o 2 C was carried out for several hours at 8 0 0 ° C and complete oxidation to M 0 O 3 d id not occur [66]. 2.3 Methanol synthesis 2.3.1 History The history of industrial methanol synthesis covers over three quarters o f a century, wi th the first barrel of synthetic methanol produced at B A S F , Germany, in 1923. The first industrial methanol synthesis process is known as the high pressure process, which operated at 250-300 bar and 3 2 0 - 4 5 0 ° C , and was the dominant technology for 45 years. The feed syngas was based on coal/lignite, which generally contained a significant amount of poisons, such as chlorine and sulfur. Accordingly , a relatively poison-resistant catalyst was developed, based on zinc oxide / chromium oxide. However, further developments found that copper increased the selectivity to methanol, and that zinc was an efficient dispersant of copper [67]. The copper based catalyst is quite easily deactivated by poisons found in coal/lignite, though this problem was overcome by ut i l iz ing synthetic gas originating from natural gas and removal of catalyst poisons prior to methanol synthesis. Accordingly , a low pressure process was developed in the late 1960's, which operates at much milder conditions, i.e. 35-55 bar and 2 0 0 - 3 0 0 ° C . The low pressure process is still the dominant process being used today [68]. 30 2.3.2 Catalyst preparation M a n y different types of catalysts have been studied, including co-precipitated copper-zinc oxide, co-precipitated copper-zinc oxide alumina, Raney copper, intermetallic compounds and precious metals such as palladium [69]. The current catalysts used in industrial low pressure methanol synthesis are based on copper / zinc oxide / alumina with possible addi-tives such as magnesium oxide [67]. The ratios of components vary from one manufacturer to another, but typically the copper oxide ranges between 25-80%, zinc oxide between 10-50% and the alumina between 5-10% [68]. Commerc ia l ly available low pressure methanol catalysts have a methanol selectivity above 99% [68]. The low pressure catalysts are obtained as metal hydroxycarbonates or nitrates by co-precipitation of a multi-metal nitrate solution with a solution of sodium carbonate [70, 71]. Preparation parameters" such as p H , temperature, composition, duration, play an important role in the quality of the produced catalyst. A typical sequence for co-precipitation is the fol lowing: 1. Prepare solution of zinc, copper and aluminum nitrates to desired ratio 2. Co-precipitate metal ions using a solution of sodium carbonate 3. Fil ter metal carbonates and wash with water 4. Dry metal carbonates at 120°C 5. Calcinate the metal carbonates in air at 300-500° C 6. Pelletize metal oxides 7. Activate the resulting catalyst by reduction in 2% hydrogen at 250°C Prior to reduction, the commercial catalysts have a B E T surface area of 60-100 m2/g, wh ich is reduced to 20-30 m2/g after reduction [68]. 31 Activat ion of freshly prepared industrial catalyst is generally carried out by reduction in a 1-5% H2/N2 at 1 bar for several hours by ramping up the temperature to 240°C(~30 °C/h) and holding at this temperature for several hours. The reduction of C u O / Z n O and C u O / Z n O / A l 2 0 3 catalysts, wi th 2% hydrogen in nitrogen at 250°C and 1 bar has been demonstrated to reduce the C u O to C u metal, v ia the intermediate C u 2 0 [70]. A t these conditions, the copper oxide is fully reduced to copper metal, without any residual copper oxide remaining [67]. 2.3.3 Catalyst poisoning / deactivation Halogen compounds are known to poison the copper surface, so methanol feed gas needs to be halogen free [69]. Other impurities in the catalyst itself, such as s i l icon compounds, nickel carbonyls or iron carbonyls, can cause damage to the catalyst [68]. The catalyst can also be deactivated when the reactor is operated in the absence of oxid iz ing agent such as carbon dioxide [70]. Experiments using a mixture consisting of only H 2 / C O rapidly and irreversibly deactivated the catalyst [71]. Methanol yields have been enhanced by the presence o f carbon dioxide, water, and/or oxygen [72]. Experiments performed by pulsing oxygen to a C O / H 2 feed indicated that up to a 15 fold increase of methanol yields could be obtained [72]. Act ive sites for methanol synthesis can be deactivated when the catalyst is operated for extended periods of time at elevated temperatures above 2 4 5 ° C [72, 69]. 2.3.4 Thermodynamics and kinetics Methanol synthesis, typically involves five components: hydrogen, carbon dioxide, carbon monoxide, water and the product methanol. The role of carbon dioxide in the reaction mechanism has been an ongoing debate [73]. U n t i l the early 1980's, mechanistic considerations were based mainly on the reaction between carbon monoxide and hydrogen, illustrated in eq. 2.19 [68]. The higher yields achieved by the addition of carbon dioxide was attributed to the displacement of the re-32 verse water gas shift equilibrium 2.20. In addition, carbon dioxide was believed to affect the oxidation state of the active sites in the catalyst [70]. Alternatively, it was proposed that methanol was formed uniquely due to the reaction of hydrogen and carbon dioxide, illustrated in eq. 2.21 [74]. Recent experiments with isotope labelled reactants showed that both reactions 2.19 and 2.21 are possible [68]. CO + 2H 2 ^ C H 3 O H (2.19) C 0 2 + H 2 ^ H 2 0 + CO (2.20) C 0 2 + 3H 2 ^ C H 3 O H + H 2 0 (2.21) Reaction enthalpies can be determined from the standard enthalpies of the reactants and products, and are -91 kJ/mol for eq. 2.19, -49 kJ/mol for eq. 2.21. Both reactions for methanol synthesis are exothermic and consequently methanol synthesis is favored by low temperature and high pressure. Additionally, the reaction of carbon dioxide with hydrogen, know as the reverse water-gas shift reaction, shown in eq. 2.20, must be taken into account. The enthalpy for the reverse water-gas shift reaction, is 41 kJ/mol. Consequently, this endothermic reaction is favored by high temperatures and low pressure. 33 Chapter 3 Experimental: Trapping and purification of nCC>2 The key intermediate in the radiosynthesis of many carbon-11 radiopharmaceuticals is [ n C]methano l . This study focuses on the feasibility of a gas phase methanol synthesis, the optimization of the main steps involved and it's applicabili ty to the radiosynthesis of [ n C]me thano l . The three principal steps evaluated in this study include: 1. Trapping and purification of [ n C ] c a r b o n dioxide 2. Conversion of carbon dioxide to carbon monoxide using molybdenum 3. Methanol synthesis from carbon monoxide and hydrogen over a copper zinc oxide catalyst 3.1 Introduction The UC02 is produced in a cyclotron and subsequently concentrated in a carbon molecular sieve trap. The effect of temperature and mass of carbon molecular sieves on the trap performance was examined by computing the trapping efficiency, illustrated in eq. 2.4. The 34 overall trapping efficiency, illustrated in eq. 2.6, is based on the irradiation conditions for production of carbon-11 (proton current, proton energy, irradiation time) and on trapping data (amount n C 0 2 not trapped, amount of n C 0 2 released from trap and time required). 3.2 Materials and methods Irradiations were performed with the Ebco Technologies T R I 9 cyclotron of the C H U S P E T facilities, Sherbrooke, Q C . Capintec dose calibrators where used for quantitative measure-ment of carbon-11. The target used is an Ebco Technologies gas target, consisting of a water-cooled aluminum cyl indrical body, which contains the target gas. Two Havar w in -dows, which separate the high pressure gas target from the high vacuum cyclotron, are cooled with hel ium. A trap was designed such that it could be cooled to -20° C and rapidly heated to 110°C. The prototype trap was built at Ebco Technologies. Thermo-electric coolers ( T E C ' s ) were selected for their ability to be remotely controlled by computer and their relatively rapid cooling time. They have the advantage over conventional cryogenic traps that no l iqu id n i -trogen is required and thus no moving parts are necessary. The T E C ' s require a D C voltage input, therefore the trap cool ing can be remotely controlled by computer. This increases the level of automation in the final synthesis module and enables repeated productions with-out having to access the unit. To provide fast heating, a 325 W cartridge heater were used, combined with the T E C ' s operated in reverse mode, which provided an additional 100 W heating power. The final trap design implemented consisted of a copper block, through which a U shape was machined and onto which two 35 m m long 6.35 m m outer diameter copper tubes were brazed. The U shape trap was then packed with carbon molecular sieve (Carbosphere 60-80 mesh) and both ends were plugged with quartz wool . Addi t ional ly , a 6.35 m m diameter hole was bored through the copper block, in which the 325 W cartridge heater was placed. 35 A K-type thermocouple was mounted on the copper trap to monitor temperature and allow temperature control. The copper block was screwed onto an aluminum heat sink block, wi th 2 thermo-electric coolers ( T E C ' s ) placed in between. A hole was dri l led through the aluminum block and connected to a supply of cool ing water, to remove the heat gener-ated by the T E C ' s during the cool ing step. After trap assembly, the carbospheres where conditioned by heating to 2 5 0 ° C and under flow of hel ium for 2 hours. Nitrogen/oxygen ( U H P 99.5% N 2 , 0.5% 0 2 ) for the production of carbon-11 and hel ium (Hel ium U H P ) were available on site at the C H U S facilities. The temperature was controlled by a manual thermostat, obtained from Omega. The flow rate of the helium sweep gas was controlled by a mass flow controller, obtained from M K S . The mass flow controller was calibrated using the soap bubble method, for which calibration curves are illustrated on page 91. Solenoid valves, 2-way normally closed, were obtained from Precision Dynamics . A check valve, wi th cracking pressure of 0.3 bar, was obtained from Swagelok. A control panel, to enable remote control o f valves, reactor temperature and mass flow controller was built by Ebco Technologies. 3.3 Experimental set-up Figure 3.1 illustrates the experimental setup used to determine the trapping efficiencies. The [ n C ] c a r b o n dioxide trap, the solenoid valve for helium gas supply, the mass flow control valve for hel ium flow control, the check valve and Ascarite column were located in a lead shielded hot ce l l . The cyclotron was located in a concrete shielded vault. The target was installed on the cyclotron target selector, which is local ly shielded. 3.4 Experimental procedure The trapping efficiencies of a copper tubular trap, wi th 4 different amounts of carbon molec-ular sieves, were measured with the experimental configuration illustrated in figure 3.1, for 36 99.5% N. 0.5% O 2 Alphas Protons 15.4 MeV *| Target He N2, a,,1 SV2 0) O SV3 MFC Dose Calibrator Fume Hood Figure 3.1: n C 0 2 Trapping experimental flow diagram temperatures varying from -20° C to 100°C. The experimental procedure can be separated in to two distinct majors steps; production of n C 0 2 and 1 1 C 0 2 trapping, purification and release. [ n C ] C a r b o n dioxide was produced by proton irradiation of the nitrogen/oxygen filled gas target. The gas target and all valves to control the filling/unloading were already in place at the C H U S P E T facilities. Thus the target loading and unloading were done using the current setup implemented at C H U S for [ n C]aceta te production, wi th the unload l ine re-plumbed to connect it to the experimental setup located in the hot ce l l . The 22 cm 3 (STP) target was filled to 17 bar. Once the cyclotron was in operation at a few U .A, the target was positioned so that at least 80% of the beam bombarded the tar-get. The current was then increased to approximately 10 u-A, whi le monitoring the target pressure, ensuring that it d id not exceed 25 bar. A saturated y ie ld of C - l l is produced in roughly 40 min of bombardment. However, for these tests, only small amounts o f ra-dioactivity were desired, mainly to keep exposure to radiation below allowable levels set by Health Canada. Consequently 10 m i n of bombardment was sufficient. The cyclotron energy was not modified, and remained at the preset value of 15.4 M e V . Due to energy drop across the Havar foils, the beam energy that actually irradiates the target gas was about 3.4.1 Production of U C 0 2 37 14 MeV. Several control runs were performed in order to determine the yie ld of the target. The target was irradiated as above, and its contents were directly emptied in an Ascarite trap lo-cated in the Capintec dose calibrator. This allows measurement of total amount of [ n ] C 0 2 produced in target as a function of irradiation conditions, namely beam intensity and irradi-ation duration. Ten control runs were performed and the target yie ld was determined using eq. 2.3. For subsequent experiments, the amount of n C 0 2 produced in target was com-puted based on the target y ie ld , proton beam intensity and an irradiation duration, using eq. 2.3. The steps involved in the production of n C 0 2 are the fol lowing: 1. Turn on cyclotron (Ion source, main magnet, vacuum, utilities, target window cool -ing, target cool ing, R F ) 2. Open the main valve on the nitrogen/oxygen gas cylinder 3. Set the pressure regulator to read 17 bar 4. Open the target valve 5. Close the target valve once pressure sensor read-back is roughly 17 bar 6. Set the N 2 / 0 2 pressure regulator to 8 bar(for target rinse) 7. Remove ion beam stop (IBS) (now beam w i l l be on target) 8. Adjust cyclotron main magnet 9. Adjust position of target 10. Increase proton beam to 10 u.A 11. Irradiate for 10 min 38 3.4.2 Trapping and release of N C 0 2 Prior to production of the n C 0 2 , the trap water cool ing valve was opened and the T E C ' s were turned on. A stable trap temperature of - 2 0 ° C was reached in less than 10 min. The target gas was unloaded through the pre-cooled trap and rinsed with 8 bar N 2 / 0 2 mixture. The trap was then rinsed with U H P H e at flow-rates up to 70 cm 3 (STP) . The trap was closed and heated to 110°C. The n C 0 2 was released with a stream of H e and then trapped in a sodium hydroxide column. The sodium hydroxide column was placed in a dose calibrator, to al low continuous measurement of n C 0 2 . This enabled quantification of the amount of n C 0 2 not trapped by the trap, and also the amount released from the trap. These quantities were decay corrected to end of the bombardment ( E O B ) (eq. 2.2), to enable determination of the overall trapping efficiency (eq. 2.5). In order to ensure that no activity remains on the trap, it was disassembled and placed in the dose calibrator. The steps involved in the n C 0 2 trapping and releasing are the fol lowing: 1. Turn on cool ing water and T E C 2. Open valve on hel ium gas cylinder and set regulator to 2 bar 3. Set mass flow-control valve to desired set point 4. Place a fresh sodium hydroxide column in dose calibrator 5. Open valve to unload target 6. M a k e sure target pressure is below 8 bar 7. Open valves on carbon molecular sieve trap 8. Measure amount of n C 0 2 breakthrough throughout unloading step 9. Close valve to unload target 10. Open valve to fill target wi th N 2 / 0 2 to rinse target and close after 10 s 39 11. Open valve to unload target 12. W h e n the target pressure is ~ atmospheric, close a l l valves 13. Open H e l i u m supply valve to rinse trap for 30 s 14. Heat trap (until release temperature is reached) 15. F lush helium through cold trap 16. Note radioactivity reading every 15 s 3.5 Results and discussion The target yie ld was determined to be 83 ± 4 m C i / u . A , for an extracted beam energy of 15.4 M e V and energy on target of approximately 14 M e V . A s can be seen in table 3.1, this value is lower than other reported experimental and theoretical yields. The target was subsequently redesigned to give a yie ld of about 150 m C i / u A . Table 3.1: C - l l target yields, 14 M e V Target Y i e l d ( m C i / ^ A ) Ebco improved target 150 Ebco target 83 ± 4 Vandewalle et al.[32] 135 ± 7 Casel la e t a l . [30] 92 Theoretical yield[31] 169 The time required for each step of the trapping sequence was determined in a system-atic fashion. The unload time was set to 30 s since at that time the pressure in the target had dropped to nearly atmospheric. The time required for the rinse target step was set to 15 s, corresponding to the time required to manually perform the sequence of valve actuation. The time for the second target unload was determined by monitoring the radioactivity co l -lected in an Ascarite column, with the n C 0 2 trap bypassed. A maximum value was seen after 45 s. Addi t ional runs were done with 2 target rinses, however no significant amount 40 Table 3.2: U C 0 2 trapping/releasing Sequence Step Description Time elapsed (min:s) 1 Unload target ( E O B ) 0:00 2 Rinse target 0:30 3 Unload target 0:45 4 He Rinse trap 1:30 5 Heat trap 2:00 6 Release n C 0 2 2:30 Table 3.3: n C 0 2 trapping efficiencies Trap Carbosphere (g) Trapping efficiency ( ± S T D E V ) 1 1.0 25% ± 2% (n=4) 2 1.3 4 5 % ± 2 % (n=4) 3 1.9 92% ± l%(n=4) 4 2.8 99% ± l%'(n=4) of additional C-11 was recovered. The H e rinse of the trap was set to 30 s, to al low suf-ficient time to remove the target gas from the trap [34]. The time for heating the trap to 110°C was 30 s. Table 3.2 illustrates the final sequence used, wi th the time (from E O B ) at which each step was executed. 3.5.1 Effect of carbon molecular sieve mass on n C 0 2 trapping effi-ciency The amount of carbon molecular sieves necessary for optimal trapping was determined by comparing the trapping efficiencies for different amounts of carbon molecular sieves. The pre-cooled trap temperature was typically -20° C and the amount of 60-100 mesh carbon molecular sieves, was varied from 1.2 g to 2.8 g . Table 3.3 illustrates the trapping efficiencies for four different amounts of carbon molecular sieve. A s the amount of carbon molecular sieves increases, the trapping efficiency increases. However, wi th increasing amount of carbon molecular sieves, the time required for n C 0 2 recovery w i l l l ike ly increase and the volume in which the n C 0 2 is delivered to downstream process may be larger. Thus the amount o f carbon molecular sieves used for subsequent 41 300 — T = 100 °C ^ - T = 20°G -«-T.= -20 °C . ^ ' t ' — i . , . , . 1 , , 4 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 Elapsed time, from E'OB (min) Figure 3.2: 1 1 C 0 2 activity vs. time, for different trap temperatures. experiments was set to 2.0 g , i.e. the lowest amount required to obtain above 95% trapping efficiencies. The optimal amount w i l l vary with target volume, target pressure, length and size of tubing from the target to the trap and trapping temperature. 3.5.2 Effect of trap temperature on CO2 trapping efficiency The temperature required for optimal trapping was determined by comparing the trapping characteristics at different trapping temperatures. The trap temperature was kept constant for each run, and the sequence shown in table 3.2 was used, with steps 5 and 6 omitted. Thus, the H e rinse step was continued until 10 minutes E O B . Figure 3.2 illustrates the effect of temperature on the retention of 1 1 C 0 2 in the trap. A t 100°C, the trap did not retain any significant amount of n C 0 2 . A t room temperature, there was a slight delay before the activity was seen in the ascarite column. A t - 2 0 ° C , the activity remains on the trap even after 10 minutes E O B . The initial amount of n C 0 2 at E O B was approximately 250 m C i in each case. 42 Table 3.4: Comparison of overall trapping efficiency Trap Overal l trapping efficiency Cryogenic [13, 35] 76-84% Molecula r sieve 13x[ l3] 74% Carbon molecular sieve[34] 83% Carbon molecular sieve (this work) 88% ± 2 % 3.5.3 Overall trapping efficiency For a trapping efficiency of 99% obtained using 2.8 g carbon molecular sieve at - 2 0 ° C , the time required to deliver the n C 0 2 was 3.5 minutes. This corresponds to an overall trapping efficiency of 88%, which is competitive with reported results (see table 3.4). This would be sufficient for incorporating into a system for routine production o f carbon-11 radiopharmaceuticals. 43 Chapter 4 Experimental: Reduction of CO2 to CO 4.1 Introduction The fol lowing step in the overall process for 1 1 C H 3 O H preparation is the reduction of n C 0 2 to n C O . Yields up to 80% were reported at 8 5 0 ° C for the reduction of n C 0 2 to n C O using molybdenum [1], as described in section 2.1.3. N o special treatment to the molybdenum was done prior to reaction, and the reaction proceeds by direct oxidation of molybdenum to molybdenum dioxide as shown in eq. 2.8 [75]. 4.2 Preliminary experiments: reaction of n C 0 2 and Mo Preliminary experiments for the reduction of carbon dioxide were performed at the C H U S P E T facilities, Sherbrooke, Q C , using n C 0 2 . 4.2.1 Materials and methods The n C 0 2 was produced and purified using procedure described in Chapter 3 on page 34. The carbon dioxide/carbon monoxide reactor consisted of a 9.5 m m outer diameter cy l in -drical quartz tube with 6 m m inner diameter, packed uniformly with 2.3 g of molybdenum 44 wire, 0.05 mm diameter and 500 m (Goodfellow Corp). The quartz tube was horizontally mounted in a 400 W ceramic tubular heater (Omega). The temperature was measured with a K-type thermocouple (Omega) and controlled with a manual thermostat (Omega). A mass flow controller, obtained from M K S , was used to control the flow rate o f the hel ium sweep gas. It was calibrated by varying the input voltage from 0-5 V , and the output flow of helium was measured using the soap-bubble technique. The flowrates were corrected to S T P using the ideal gas law. The resulting calibration curve is illustrated on page 92. Solenoid valves, 2-way normally closed, were obtained from Precision Dynamics . A control panel, to enable remote control operation of valves, reactor temperature and mass flow controller, was built by Ebco . 4.2.2 Experimental set-up The flow diagram for the experiments using n C 0 2 is illustrated in figure 4 .1. The same set-up was used as described in section 3.3 for production and release of n C 0 2 . The molybdenum reactor was added to the outlet of the n C 0 2 trap, fol lowed by an Ascarite (Aldr ich) trap and a s i l ica (Aldr ich) trap, which was cooled with l iqu id nitrogen. A t the outlet of the reactor, a solenoid valve was placed to avoid exposure to atmosphere. The setup was located in a lead shielded hot ce l l , wi th all valves / flowmeter controlled remotely. 99.5% N . 0.5% Q , SV1) Alpha s Protons 15.4 MeV Target N ^ . O ^ ' . C O , He SV2 I c v 1 j - X — t & H 1 !SV3 MFC 11co2 Trap He. " C 0 „ " C 0 , SV5 fx3—1——-2 ^ He SV4 He, N „ 0 . Mo > Reactor 8 CD C/3 H e . 1 , C O Dose Calibrator ; Fume Hood Figure 4 .1: n C 0 2 Reduction experimental flow diagram 45 4.2.3 Experimental procedure The production of n C 0 2 was carried out as in the n C 0 2 trapping experiments, described in section 3.4.1. The irradiation condition were noted to enable quantification of the total amount of n C 0 2 produced in target. The n C 0 2 trap was conditioned by flowing hel ium at 250°C for 2 h. Initial experiments were carried out with molybdenum as-supplied. Sub-sequently, experiments where carried out after pre-reducing the molybdenum with 10 % hydrogen in helium at 850°C. The si l ica gel trap was ini t ial ly conditioned by heating to 650°C while purging with helium. The molybdenum oven was preheated to 800, 850, 900 and 950°C, under flow of helium. The n C 0 2 was released from the n C 0 2 trap at 17.5, 35 and 70 cm 3/min (STP) helium, through the molybdenum reactor, the Ascarite column and the si l ica column submerged in l iqu id nitrogen. The co ld s i l ica column was placed in the dose calibrator to measure the amount of u C O . W h e n the activity on the si l ica co l -umn reached a maximum, the amount o f n C 0 2 on the Ascarite was l ikewise measured by placing the column in the dose calibrator. This enabled quantification of the amount of un-reacted n C 0 2 that had been released from the reactor. To minimize exposure to radiation, long thongs were used to place the Ascarite trap in the dose calibrator. The decay corrected conversion of n C 0 2 to 1 1 C O was computed based on the amount o f activity produced in target and the amount of n C O measured. Table 4.1 illustrates the sequence that was used, with the time from E O B at which each step was executed. 4.2.4 Results and discussion Initial experiments resulted in n C 0 2 conversion to u C O of about 5%. A t the end of the experiments, it was observed that the molybdenum surface was no longer a shiny metallic color but a dul l grayish color. Since the industrial compound M o 0 3 is described as a gray-green powder [48], it was suspected that the molybdenum was oxidized. Molybdenum metal powder is produced industrially by reducing compounds such as M 0 O 3 powder with 46 Table 4.1: U C 0 2 Reduction sequence Step Description Time elapsed (min:s) 1 Unload target ( E O B ) 0:00 2 Rinse target 0:30 3 Unload target 0:45 4 H e rinse trap 1:30 5 Heat trap 2:00 6 Release U C 0 2 2:30 7 1 1 C O measurement 6:30 8 1 1 C 0 2 measurement 7:30 Table 4.2: Summary of n C O yields, 17.5-70 cm 3/min, 2 bar T ( ° C ) Y i e l d (decay corrected) S.D. 800 6% 2% 850 8% 2% 900 9% 4% 950 8% 2% hydrogen between 5 0 0 - 1 1 5 0 ° C [49]. In order to reduce the molybdenum trioxide back to the metallic state, 10% hydrogen in argon was flowed through the reactor at 8 5 0 ° C , for 30 min. Thereafter, experiments were resumed. Decay corrected yields were calculated based on eq. 2.2. The amount of [ n C ] c a r b o n dioxide produced in the target at the end of bombardment was computed using eq. 2.3, based on the target y ie ld (see table 3.1), beam current and irradiation duration. A summary of the decay corrected yields obtained during hot experiments are given in table 4.2. For each temperature tested, at hel ium flow rates of 17.5, 35 and 70 cm 3/min (STP) , the average conversion of n C 0 2 to n C O was at most 9%. Based on these results, there is no apparent correlation between flow and/or temperature with conversion. Conversion in the order of 80% was expected based on published data [40]. After completing the hot experiments, the molybdenum was again a dul l grayish color indicating that oxidation may have occured. It is suspected that the helium flush o f the n C 0 2 trap was insufficient, and that some of the oxygen from the target gas remained in the n C 0 2 trap. W h e n attempting to reduce the U C 0 2 wi th the molybdenum reactor, the 47 residual oxygen may have oxidized the molybdenum. Molybdenum retains its luster almost indefinitely in air, particularly when it has been drawn to fine wire. However, when heated in air, oxidation occurs and renders the surface chemically unreactive [48]. Consequently, for future use it is recommended to ensure that a l l oxygen is removed from the n C 0 2 trap. This can be practically implemented by flowing additional hel ium through the n C 0 2 trap at room temperature or during the heating step, prior to n C 0 2 desorption. Another interesting observation is that on average, only 53% of the C - l 1 was accounted for during runs performed immediately after reduction of the molybdenum. U s i n g a hand held Geiger counter, most of the unaccounted for carbon-11 was found to be on the molyb-denum reactor. For subsequent runs (i.e. not immediately after reduction), an average of 96% of the expected carbon-11 was recovered from the reactor. It appears that freshly re-duced molybdenum either reacts wi th carbon oxides or that carbon oxides remain adsorbed on the surface. Clearly the state of the molybdenum plays an important role in the reduction of carbon dioxide to carbon monoxide. To gain a better understanding o f surface pretreatment and the effect of flow and temperature on C 0 2 reduction by molybdenum, further experiments were carried out wi th cold 1 2 C 0 2 at U B C . In parallel to these experiments, a study o f equil ibrium using Aspen was carried out for various systems containing molybdenum, molybdenum oxides, molybdenum carbides, oxygen, hydrogen, carbon, carbon oxides and water. 4.3 Equilibrium computer simulations of systems contain-ing molybdenum Equi l ib r ium of molybdenum oxidation with oxygen and molybdenum trioxide reduction with hydrogen reduction were examined, as was the equil ibrium of molybdenum oxidation with carbon dioxide using a process simulator. Two process simulators were available in the Department of Chemica l & Bio log ica l Engineering; Aspen [76] and Hysys [77]. Aspen was 48 selected due to it's broad component database which includes solids such as molybdenum. In order to obtain equil ibrium data, a Gibbs reactor was setup using Aspen. A l l relevant components available from the database were included for each simulation. U s i n g the sensitivity study feature o f Aspen, equi l ibr ium data was generated for a wide range of temperatures. Further details about setting up the Aspen simulations are described in the appendix, on page 100. A good overall agreement between Aspen produced simulation data with published equi l ibr ium data for the system containing M o , C H 4 , C , H 2 and M o 2 C is demonstrated. 4.3.1 Reduction of M0O3 with H 2 Aspen generated equi l ibr ium data for the system containing H 2 , 0 2 , H 2 0 , M 0 O 3 , M o 0 2 , M o O and M o is presented in figure 4.2, based on an init ial H 2 : M 0 O 3 ratio of 15:1, and at atmospheric pressure. M o 0 3 is reduced to M o 0 2 at room temperature. A t 800 K , M o 0 2 begins to form M o , until it is completely reduced at -1000 K . Based on simulation re-sults, it appears that the reduction of M o 0 3 to M o proceeds through two reactions, v ia the intermediate molybdenum dioxide, as shown in eqs. 4.2 and 4.3. In practice, a large excess of H 2 w i l l be present, since a continuous stream of hydrogen would flow over a fixed amount of molybdenum. Figure 4.3 illustrates the effect of increas-ing the amount of Ff 2 from stoichiometric feed amounts (for complete reduction according to eq. 4.1) up to an excess with H 2 : M 0 O 3 molar ratio of 240:1. The effect of increas-ing the excess of hydrogen is to shift the equil ibrium to metallic molybdenum at a lower temperature. 3H 2 + M0O3 — Mo + 3H 2 0 H 2 + M0O3 — M o 0 2 + H 2 0 2H 2 + M o 0 2 — Mo + 2H 2 0 (4.1) (4.2) (4.3) 49 900 Temperature (K) -MO — -MOO -H2 1500 02 Initial ratio of H 2 : M 0 O 3 of 15:1, at atmospheric pressure Figure 4.2: Equi l ib r ium compositions of M0-M0O-M0O2-M0O3-H2-O2-H2O from 300 to 1500 K 4.3.2 Oxidation of Mo with 0 2 Equilibrium of oxygen and molybdenum Using the same component database and Gibbs reactor as for M 0 O 3 reduction, equi l ibr ium data for M o oxidation was obtained. The data generated from Aspen indicates that M o would be completely oxidized to M 0 O 3 , as illustrated in eq. 4.4, at temperatures above 273 K , for stoichiometric and larger amounts of 0 2 . 3 0 2 + 2 M o - * 2 M o 0 3 (4.4) Equilibrium of carbon dioxide and molybdenum Similar to above, simulations were carried out using Aspen to obtain equi l ibr ium curves to better understand the effect of temperature on interaction between M o and C 0 2 . To cover all possible reactions, the components added to the simulation were the fo l lowing: H e , A r , 50 300 500 ?00 900 1,100 1,300 1.500. Temperature (K) H , : MoO, Ratio | 240:1 60:1. 15:1 7.5:1 -iiT] Figure 4.3: Equ i l ib r ium mole fraction of M o 0 2 for different H 2 : M o 0 3 ratios from 300 to 1500K C O , C 0 2 , C , M o , M o O , M o 0 2 , M 0 O 3 , M o C and M o 2 C . The gas feed stream was set to contain 50 ppm C 0 2 in argon, with an excess molybdenum, at atmospheric pressure. The resulting data is illustrated in figure 4.4. The compounds present at equil ibrium are M o , C O , C 0 2 , M o 2 C , M o 0 2 . Three distinct stages can be observed. Between 500 K and 800 K , equimolar amount of M o 0 2 and M o 2 C are formed, most probably according to eq. 4.5, and no carbon is present in the form of carbon oxides. A s the temperature increases above 800 K , the amount of M o 2 C relative to M o 0 2 decreases, until no M o 2 C is present at 910 K . A t this point the amount of C O is the highest, corresponding to 73% conversion o f C 0 2 . A b o v e 910 K , it appears that the reaction proceeds according to eq. 4.6. C 0 2 + 3 M o M o 2 C + M o 0 2 (4.5) 2 C 0 2 + M o _> 2 C O + M o 0 2 (4.6) To verify these reactions independently, Aspen simulations were carried out for the 51 1.0 500 BOO 700 BOO 900 1000 1100 1200 1300 1400 1500 T ( K ) j C02 CO MO -MOQ2 - M02C | Figure 4.4: Equ i l ib r ium of composition of M o - M o 0 2 - M o 2 C - C 0 2 - C O same components as above, but using an equil ibrium reactor, for which the chemical re-action can be specified. The feed gas stream was set to contain 50 ppm C 0 2 in argon, with an excess molybdenum in the solid feed stream, both at atmospheric pressure. For an equil ibrium reactor based on eq. 4.5 only, the resulting equil ibrium curve is illustrated in figure 4.5. For temperatures below 800 K , no carbon dioxide is present, the only com-pounds present are dimolybdenum carbide and molybdenum dioxide, which confirms that the lower temperature plateau in figure 4.4 proceeds according to eq. 4.5. For an equilib-r ium reactor based on eq. 4.6, the resulting equil ibrium curve is illustrated in figure 4.6. A t 500 K , carbon monoxide is present wi th a molar fraction just below 0.90. A s temperature increases, the molar fraction of carbon monoxide gradually decreases to 0.66 at 1500 K . The high temperature plateau in figure 4.4 follows the same exponential decrease, which confirm that the high temperature stage proceeds according to eq. 4.6. Further simulations were carried out to explore the effect of varying the C 0 2 concen-tration on the equil ibrium amounts of C O present, as represented in figure 4.7. The equi-l ibr ium data is represented in terms of conversion of C 0 2 to C O , which was calculated by 52 Figure 4.5: Equilibrium molar fraction of CO2 from eqs. 4.5 1.00 0.90 0.80 0.70 0 0 c 0:60 0 ati 0:50 LL U 0 40 O S 0.30 0.20 0.10 500 700 900 1100 1300 1500 Temperature (K) Figure 4.6: Equilibrium molar fraction of CO from eqs. 4.6 53 80% Q70% 5 60% §50% ^40% c30% o 0 20% j$ 10% 0% r / ™ j ,< / : I ; / i / ! ; I i / I 1 / J 500 700 900 1100 T(K) 1300 1500 1 ppm C02 1% C02 10% CQ2 50 ppm Figure 4.7: Conversion of C 0 2 to C O vs. temperature for different C 0 2 concentration, based on Aspen equil ibrium data dividing the equil ibrium amount of C O by the initial amount of C 0 2 . The shape of the curve follows a similar pattern for each C 0 2 concentration. The three stages described above are present in al l cases. The temperature at which C 0 2 conversion reaches a maxi -mum increases with increasing carbon dioxide concentration. For 1 ppm carbon dioxide, the maximum C 0 2 c o n v e r s i o n to C O occurs at 850 K , while for 10% carbon dioxide the maximum conversion occurs at 1250 K . C o l d experiments were performed with 50 ppm carbon dioxide. Accord ingly , the ex-perimental temperature range was selected to cover the three stages described above, i.e. from 600 to 1150 K . 4.4 Cold experiments: reaction of CO2 and Mo 4.4.1 Introduction C H U S "hot" testing facilities provide a relatively simple method to quantify conversion of trace amounts o f n C 0 2 . However, due to l imited availability of the C H U S facilities and the financial burden associated with their use, " c o l d " testing with non-radioactive C 0 2 was 54 done to better understand the factors affecting the reduction of C 0 2 wi th molybdenum. Temperature programed reduction of carbon dioxide with molybdenum and the effect of flow rate on the reduction of carbon dioxide with molybdenum were studied. 4.4.2 Materials and methods During typical productions of carbon-11 compounds, the concentration of C 0 2 ranges from 1-2000 ppm in the feed gas, as demonstrated in the appendices on page 105. A 50 ppm C 0 2 feed gas was selected, consisting of certified 51.4 ppm C 0 2 in hel ium, which was contained in a 100 bar standard laboratory size gas cylinder. The gas flow rate was controlled by a mass flow controller ( M K S instruments) for flow rates below 70 cm 3/min (STP) and by an adjustable needle valve correlated flowmeter (Cole-Parmer) for higher flowrates. Calibrat ion of both flow control devices was done using the soap-bubble method, and calibration curves are included in the appendices, on page 91. Manual 2-way valves (Swagelok) were placed at the inlet and at the outlet of the reactor. The reactor consisted of a 400 W ceramic tubular heater (Omega Heating) in which a glass tube was placed, which contained 2.3 g molybdenum wire previously described, covering a length of 10 cm. A K-type thermocouple was placed in the middle of the ceramic heater, and a second K-type thermocouple was place inside the reactor, on the molybdenum wire. Temperature was controlled with a manual operated thermostat (Omega). The temperature profile along the reactor was obtained by measuring the temperature at 2 cm intervals from the entrance of the reactor. Temperature profile of the reactor showed a variation of over 160°C over the molybdenum covered length of the reactor for a setpoint o f 7 4 0 ° C , as illustrated in the appendices, on page 98. The addition of quartz woo l plugs on either end of the ceramic heater improved the temperature profile along the reactor, reducing the temperature variation of the molybdenum covered length to less than 6 0 ° C . The product gas stream was connected to a 6-port sampling valve (VICI ) . The 6-port 55 valve was also connected to a waste stream, to a 2 cm 3 ( S T P ) sample loop, to a hydrogen supply and to the the inlet of the gas chromatograph. A Varian 3600 gas chromatograph was used, equipped with a a flame ionization detector (FID) . The column used on the gas chromatograph was a 3 m long, 3.2 mm outer diameter column, packed with 60-80 mesh carbon molecular sieve. F I D detectors are not very sensi-tive to carbon monoxide, and are not useful for detection of carbon dioxide. A wel l known method of analyzing trace amounts of carbon oxides using F I D is to first convert the oxides to methane using a nickel catalyst. Near quantitative conversion of carbon monoxide and carbon dioxide to methane can be accomplished by flowing the sample gas containing the carbon oxides over a nickel catalyst with excess hydrogen, at 3 7 0 - 4 5 0 ° C . Accordingly , a methanizer was built. A 15 cm long aluminum heater block was built at Ebco to accom-modate a 6.3 mm diameter 175 W cartridge heater (Omega), a 3.2 mm diameter tube and a 1.6 mm diameter K-type thermocouple (Omega). A 3.2 mm diameter stainless steel tube was packed with 0.34 g of a mixture of 15% commercial ly available nickel on alumina powder (< 60 u.m) and 85% activated alumina, 100-120 mesh size. The excess nickel pow-der was shaken off prior to packing the tube, after which quartz wool plugs were placed at both ends of the tube. The nickel packed tube was placed in the aluminum heating block. Swagelok fittings were placed on either end of the tube. The methanizer was placed in the gas chromatograph, in between the separation column and the flame ionization detector. The temperature was controlled with a manual thermostat (Omega). Calibrat ion curves for carbon monoxide and carbon dioxide are given in the appendices, on page 93. 4.4.3 Experimental set-up The flow diagram for the experimental setup used for the reduction of C 0 2 to C O is i l lus-trated in figure 4.8. 56 He/C0 2 /CO He/ CO 2 M F C SV1 Reactor Mo He/CO,/CO — { X SV2 GC MV1 NV1 Figure 4.8: Experimental flow-diagram for the reduction of C 0 2 4.4.4 Experimental procedure Preliminary experiments, described in section 4.2, gave less than 10% conversion of [ n C ] c a r b o n dioxide to [ n C ] c a r b o n monoxide, most probably due to the presence of oxygen which ox-idized the molybdenum to molybdenum trioxide. Aspen equil ibrium data suggests that the temperature at which molybdenum oxides are reduced to molybdenum metal depends on the molar ratio of hydrogen to molybdenum. For a ratio of H 2 : M o of 240, complete reduction to molybdenum occurs at 400° C . For a ratio of F f 2 : M o of 7.5, complete reduction to molybdenum occurs at 9 0 0 ° C . In the reactor used, 2.3 g of molybdenum wire was placed in a length o f 10 cm within the 6.3 mm inner diameter quartz reactor. Assuming only the exposed surface atoms are available for reaction, this corresponds to a molar ratio for H 2 : M o of 130. A t this molar ratio, Aspen generated data indicates that 4 2 5 ° C would be sufficient for reduction to molybdenum. There is a general agreement that the reduction of molybdenum trioxide with hydrogen to molybdenum metal is a two step process, forming molybdenum dioxide as intermediate, though the exact temperature at which reduction is reported varies. The reduction of molyb-denum trioxide to molybdenum dioxide is reported to occur between 350-600°C while the reduction of molybdenum dioxide to molybdenum is reported to occur between 500-1150°C . Lee et al. [54] prepared molybdenum by T P R of 0.5 g M o 0 3 powder with pure hydrogen at 50 c m 3 / m i n , from 3 0 0 ° C to 7 0 0 ° C , increasing temperature 60 ° C / h . 57 Carbon monoxide is more strongly adsorbed on freshly reduced molybdenum carbide than unreduced molybdenum carbide [54]. Iwasawa et al. [55] observed deposition of small amounts of carbon on molybdenum fixed on alumina, after it had been reduced with hydrogen at 5 0 0 ° C for 5 h. Al though no data was provided for adsorption of carbon monoxide and carbon dioxide on freshly reduced molybdenum, there is a chance that ad-sorption is higher on freshly reduced molybdenum as wel l . Lee et al . [54] suggested to passivate dimolybdenum carbide with 1 % oxygen at room temperature prior to removal of carbide from the reactor for surface analysis. Fo l lowing Lee 's [54] procedure, temperature programmed reduction was performed on the molybdenum wire, from 300 to 7 0 0 ° C wi th a temperature increase of 6 0 ° C / h , wi th an additional 12 hours of isothermal reduction at 7 0 0 ° C . U H P hydrogen, which contained 50 ppm of C O and 50 ppm CO2 was used as reduction media. Fo l lowing cool ing, the molybdenum wire was passivated with U H P compressed air, at room temperature, for one hour. Experiments were carried out using the experimental set-up illustrated in figure 4.8. The experiments were separated in two parts, temperature programed reaction and flowrate programed reaction. For the temperature programed reaction experiments, the gas stream was fed through the reactor at a fixed pressure of 2 bar and a fixed flowrate of 4 cm 3/min (STP) . The temperature of the molybdenum wire was varied from 331 and 8 8 3 ° C , in increments of approximately 6 0 ° C . A t each temperature, 2 cm 3 samples of the gas product were analyzed using G C for quantification of carbon dioxide and carbon monoxide. For the flowrate programmed experiments, the molybdenum wire was kept at a fixed temperature of 8 2 5 ° C and the feed gas was kept at constant pressure o f 2 bar. The flowrate of the feed stream containing 50 ppm C 0 2 in hel ium was varied from 5 to 1642 cm 3/min (STP). A t each flowrate, a 2 cm 3 sample of the gas product was analyzed using G C for quantification of carbon dioxide and carbon monoxide. 58 4.4.5 Results and discussion During the reduction of molybdenum, water vapor was visible at the outlet of the reactor, indicating that molybdenum oxides were present initially. A t the end of the isothermal re-duction at 7 0 0 ° C , water vapor was no longer visible. There was some concern that molyb-denum carbide may have been formed, due to the presence of 50 ppm carbon monoxide and carbon dioxide in the hydrogen gas. G C analysis of the product stream showed a sin-gle peak, methane, indicating that carbon oxides reacted with hydrogen to form methane. Addi t iona l Aspen equil ibrium simulations indicated that the only components in the prod-uct stream at equil ibrium at 7 0 0 ° C are molybdenum, methane, water and hydrogen. The molybdenum wire had a shiny metallic finish, characteristic of molybdenum metal. It was concluded that molybdenum oxides were present and that they were reduced to molybde-num. The effect of temperature on the ratio of C O to total carbon oxides in the product stream from a feed stream containing 50 ppm C 0 2 reacting with molybdenum from 331 and 8 8 3 ° C is illustrated in figure 4.9. Three distinct stages can be observed: 1. from 331 to 6 0 0 ° C , a plateau with average C O / ( C 0 2 + C O ) of 20%, 2. from 600 to 7 0 0 ° C , an abrupt increase from 20% to over 70% C O / ( C 0 2 + C O ) 3. from 700 to 8 8 3 ° C , a plateau with average C O / ( C 0 2 + C O ) of 71 %. A similar abrupt increase in carbon monoxide generation was observed by Zeisler et al. [40] at 8 2 5 ° C (see figure 2.3), for reaction of n C 0 2 wi th molybdenum, and by Solymosi et al. [66] at 6 0 0 ° C (see figure 2.4), for reaction of carbon dioxide with molybdenum carbide. Aspen generated equil ibrium data shows an abrupt increase in the conversion of C 0 2 to C O , as illustrated in figure 4.7. The most notable difference between Aspen generated equil ibrium data and experimental data is the fact that Aspen data suggests the absence of carbon oxides in the lower temperature plateau. However, experimental data from this study 59 80% -| 70% -60% -6 50% -o + o 40% -o o 30% -o 20% -10% 0% * X X x x X X X X K X. 500 600 700 800 900 1000 1100 1200 1300 T(K) Figure 4.9: Effect of temperature on reaction of C 0 2 and M o , at 4 cm 3/min and 2 bar and others indicate that carbon monoxide is present at the lower temperature plateau. Aspen equil ibrium data suggests that in the lower temperature plateau all the carbon dioxide reacts with molybdenum to form dimolybdenum carbide and molybdenum dioxide, as illustrated in figure 4.5. However, experimental results show that both carbon dioxide and carbon monoxide are present between 331 to 6 0 0 ° C . Overal l , Aspen appears to be a good tool to determine the general profile for equil ibrium conversions of carbon dioxide to carbon monoxide as a function of temperature for the conditions of this study, in particular for the second and third temperature stages described above. The effect of flowrate on the ratio of C O to total carbon oxides in the product stream from a feed stream containing 50 ppm C O 2 reacting with molybdenum at 8 2 5 ° C is i l lus-trated in figure 4.10. Between 5 and 70 c m 3 / m i n (STP), there is no significant change in the ratio of carbon monoxide to total carbon oxides, averaging 76%. A t 70 c m 3 / m i n (STP) , the ratio of carbon monoxide to total carbon oxides decreases exponentially from 74% to less then 2% at approximately 1250 c m 3 / m i n (STP) . The total amount of carbon oxides in the product stream exceeded the 51.4 ppm carbon 60 250 500 750 1,000 1,250 1,500 1,750 Flowrate (seem) Figure 4.10: Effect of flowrate on reaction of C 0 2 and M o at 8 2 5 ° C and 2 bar dioxide contained in the feed stream. For this reason, both figure 4.9 and 4.10 were plotted as a ratio of carbon monoxide to total carbon oxides. Figures 4.11 and 4.12 show the total amount of carbon oxides in the product stream, as a function of temperature and flowrate respectively. In figure 4.11, the data follows a similar trend as the curve in figure 4.9. There is an init ial increase in total carbon oxides in the product stream from 50 ppm to around 120 ppm, from 340 to 4 0 0 ° C . Between 400 and 6 0 0 ° C , the amount of carbon oxides gradually decreases from 120 ppm to 90 ppm. Between 600 and 8 0 0 ° C , there is a much larger increase from 90 to 450 ppm carbon oxides. F rom 800 to 8 8 0 ° C , the amount remained constant at 450 ppm. In figure 4.12, the total amount of carbon oxides decreases exponentially from 75 ppm down to around 50 ppm. Considering that the flow experiment were done after the temper-ature experiments, this decrease from 75 to 50 ppm appears to be the tail end of the larger peak seen in figure 4.11, beginning at 6 0 0 ° C . There are two possible explanations for the source of this additional carbon. The first possibility is that molybdenum carbide may have formed at the lower temperature during 61 100 90 80 10 250 500 750 1,000 1,250 1,500 1,750 Flowrate (seem) Figure 4.12: Total carbon oxides as a function of flowrate. 62 start-up of the apparatus and reacted with carbon dioxide and/of molybdenum dioxide to form carbon monoxide. The second possibili ty is that carbon dioxide and/or carbon monox-ide remained adsorbed on the molybdenum surface prior to experiments and were released during experiments. Aspen generated equi l ibr ium data suggests that M o 2 C and M o 0 2 are formed by react-ing molybdenum with carbon dioxide, from room temperature to 500° C. However, pub-lished data on formation of molybdenum carbide from carbon monoxide with molybdenum dioxide, according to the reversible eq. 2.14 on page 26, proceeds at temperatures above 565°C [57]. The formation of molybdenum carbide by reaction of C H 4 in hydrogen with molybdenum trioxide was reported to occur above 500°C. Thus, regardless of how molyb-denum carbide is formed, it appears that the temperature must be over 500°C and that the carbon source must react wi th a molybdenum oxide. In Figure 4.11, there is always more than 51.4 ppm total carbon in the product stream, indicating that carbon dioxide probably did not react wi th molybdenum to form molybdenum carbide. Thus, the possibility that molybdenum carbide was the source of additional carbon seems highly improbable. Considering the second possibility, numerous studies have found C O and carbon ad-sorption on molybdenum compounds is significant, in particular after reduction with hy-drogen [54, 55]. Dur ing preparation of the molybdenum, after T P R reduction, some carbon dioxide and carbon monoxide may have remained adsorbed on the freshly reduced molyb-denum while cool ing to room temperature. Dur ing start-up of the experiments, hel ium containing 51.4 ppm carbon dioxide was flowing through the reactor. Since up to 120 ppm carbon dioxide was measured in the product stream, it is highly probable that some carbon dioxide was previously adsorbed on the molybdenum surface. The adsorption of carbon monoxide and carbon dioxide on the molybdenum surface during start-up and pretreatment seems to be the most l ike ly source of additional carbon measured during experiments. The components molybdenum, molybdenum oxides, molybdenum carbides and carbon oxides constitute a complex system. The equil ibrium data generated by Aspen provided 63 a good prediction of the achievable yie ld of carbon monoxide. For the high temperature plateau, the reduction of C 0 2 to C O proceeds according to eq. 4.6, ox id iz ing molybdenum to molybdenum dioxide. 64 Chapter 5 Experimental: Methanol synthesis 5.1 Introduction In order to estimate the requirements for the production o f [ n C ] m e t h a n o l v ia a gas phase catalytic process, a series of experiments have been conducted with non-radioactive carbon monoxide. Al though commercial methanol production involves a catalytic gas phase pro-cess, no practical procedure is currently available for the production of [ n C]me thano l . The industrial production o f methanol uses catalysts such as C u / Z n O / A l 2 0 3 that are prepared by co-precipitation and activated with 1-3% hydrogen at 2 5 0 - 2 9 0 ° C prior to use. The feed gas is normally syngas, which contains C O , C 0 2 , C H 4 and H 2 , wi th the carbon dioxide acting as an ox id iz ing agent. This study aimed at determining the suitability of the catalytic gas phase procedure for the production of [ n C]methano l . Act ivat ion and precondition of a copper zinc oxide catalyst were evaluated. The effect of pressure, temperature and flowrate of the feed gas on the methanol yie ld were studied and a kinetic model was generated. The model was used to establish optimal condition and estimate the potential yields of [ 1 1 C]methanol production v ia a catalytic gas phase method. 65 5.2 Materials and methods In order to be consistent with conditions encountered during carbon-11 labelling proce-dures, a 50 ppm CO feed gas was selected (see page 105). It consisted of certified 50 ppm CO in hydrogen, which was supplied in a 100 bar standard laboratory size gas cylinder. UHP helium gas, contained in a 100 bar standard laboratory cylinder, was used as inert sweep gas. The gas flow rate was controlled by a mass flow controller (MKS instruments) for flow rates below 70 cm 3/min and by an adjustable needle valve correlated flowmeter (Cole-Parmer) for higher flowrates. Calibration of both flow control devices was done using the soap-bubble method, for which calibration curves are included in the appendices, on page 91. A manual 3-port valve (Swagelok) was placed at the inlet of the reactor, to select either helium or the feed gas. The reactor consisted of a 400 W ceramic tubular heater (Omega Heating) in which a 6.3 mm inner diameter stainless steel tube was placed, which contained the copper based catalyst, covering a length of 6 cm. A K-type thermocouple was placed in the middle of the ceramic heater, and a second K-type thermocouple was placed inside the reactor, in contact with the catalyst. The reactor temperature was controlled with a manual operated thermostat (Omega). Quartz tubes, 6.3 mm outer diameter with 1 mm inner diameter, where placed on both ends of the reactor to reduce the void space and the time required for the product gas to flow through the remaining portion of the reactor. Glass wool plugs were placed on both ends of the heater, to improve temperature uniformity. The resulting temperature profile of the reactor is included in the appendices. The outlet of the reactor was connected to a pressure regulator to maintain downstream pressure below the rating of the flow control devices. The product stream was also connected to a 6-port sampling valve (VICI). The 6-port valve was connected to a waste stream, a 2 cm 3 sample loop, a hydrogen supply and the inlet of a varian 3600 gas chromatograph, equipped with a flame ionization flame detector (FID). The GC was equipped with a 3 m long stainless 66 steel column (1.6 mm inner diameter) packed with 60-80 mesh poropak. U s i n g a stream of helium saturated with methanol, a calibration curve was generated and is illustrated in the appendices, on page 93. 5.2.1 Cu/ZnO catalyst preparation Copper nitrate pentahemihydrate, C u ( N 0 3 ) 2 • 2 . 5 H 2 0 , and sodium carbonate were obtained from A l d r i c h . Z inc nitrate hexahydrate, Z n ( N 0 3 ) 2 • 6 H 2 0 was obtained from JT Baker. The procedure used for the preparation of the copper zinc oxide catalyst was based on the method described by Herman et al . [71]. A 1.0 M solution of copper nitrate was prepared by dissolving 21.6 g copper nitrate pentahemihydrate in 92.9 mL distil led water. Similar ly, a 1.0 M zinc nitrate solution was prepared by dissolving 64.5 g in 216.7 m l distilled water. These nitrate solutions were mixed to obtain a 30:70 molar ratio of C u : Z n . A 1.0 M solution of sodium carbonate was prepared by dissolving 50 g N a 2 C 0 3 in 472 mL distil led water and was added dropwise to the 310 mL metal nitrates solution at 90° C , until the p H was raised from 2 to 7.0. The titration took 5 h and consumed approximately 360 mL sodium carbonate. Fo l lowing a 12 h digestion at room temperature, the turquoise precipitate was filtered over glass frit and dried overnight at 8 5 - 1 0 5 ° C . The subsequent calcination o f the copper/zinc carbonates was carried out in air by heating from 150 to 3 5 0 ° C in 2 h, with the maximum temperature maintained for 4 h. The resulting copper/zinc oxides were pelletized from an aqueous slurry, dried at ambient temperature and crushed and screened to a uniform size of 650 ± 200 |j.m. Samples of 2 g were placed in the reactor, covering a length of 6 cm. The catalyst was reduced under flow of 2% hydrogen in hel ium, for 12-16 h at 2 5 0 ° C and 1 bar. A standard B E T method was used for the determination of surface area from argon adsorption at - 1 9 6 ° C and the surface area of the reduced, used catalyst, was 30.3 m 2 /g. This is in the same range as measurements made by Herman et al. , who reported a surface area of 37.1 m 2 /g for catalyst prepared by a similar method, and used under commercial methanol synthesis conditions [71]. 67 5.3 Experimental procedure The flow diagram for the experimental setup used for the methanol synthesis reactor is illustrated in figure 5.1. H 2 / CO He MV1 -fc