THE PRODUCTION OF LIQUID HYDROCARBONS BY THE FISCHER-TROPSCH SYNTHESIS by F. A. Mackinnon Buck, B. A. Sc. A Thesis submitted in partial fulfilment of the requirements for the Degree of MASTER OF APPLIED SCIENCE in the Department of CHEMICAL ENGINEERING The University of British Columbia April, 1944. TABLE OF CONTENTS Page I Introduction (A) (B) II Theory (A) (B) (C) (D) (E) Object of Research Historical 3-Principal Reactions Equilibrium Calculations (1) Discussion (2) Calculations for Methane (3) Extension to higher homologues The Catalyst Influence of Magnetic Field Other Factors (1) Composition of reactant gas (2) Size of reaction tube (3) Influence of catalyst concentration (4) Influence of flow velocity III Apparatus 15. (A) General Description (B) Reaction Chamber (0) Solenoid (D) Temperature Control (E) Condensers IV Experimental Work 19. (A) Operation (B) The catalyst Y Appendix (A) References (B) Heats of formation (C) Free energies of formation (D) Out-throw diagram THE PRODUCTION OF LIQUID HYDROCARBONS BY THE FISCHER-TROPSH SYNTHESIS I Introduction (A) Object of Research This research was undertaken i n an attempt to investigate a few of the factors influencing the hydrogenation of carbon monoxide to liquid hydrocarbons by the so called Fischer-Tropsch Synthesis. The investigations originally planned were: (1) The influence of different catalysts (2) The influence upon the rate of reaction and the products obtained of a superimposed magnetic f i e l d . It is evident, however, that before any conclusions regarding the effect of these two factors may be made, some sort of reaction standard must be adopted to act as a blank against which results may be compared. In other words i t was necessary f i r s t of a l l to design and build an apparatus so that a l l process variables except the two under consideration could be maintained constant; and secondly to make enough runs under normal operating conditions to establish some sort of a standard of comparison for reaction rate, products obtained, catalyst l i f e , thermal efficiency, and so forth. (B) Historical It has long been known as a result of the 1 pioneering work of Sabatier that finely divided nickel 1. w i l l catalyze the hydrogenation of carbon monoxide to methane, ffischer and Tropsch were among the f i r s t to show that with the proper selection of the catalyst the methane reaction of carbon monoxide and hydrogen may be modified so drastically that the principal product is a complex mixture of higher paraffin hydrocarbons originally known as "Synthen" but more recently given the appropriate name "Kogasin". Like the Berguis process, the Kogasin process provides a new method for the chemical conversion of coal to liquid products, and is looked upon as a partial solution to the diminishing resources of crude petroleum. Kogasin consists almost wholly of straight chain hydrocarbons. It i s unsuitable for motor fuel unless i t is blended with antiknock agents or re-formed by crack-ing. The higher fractions, however, are eminently suited to use as Diesel fuels and lubricants. It is not surpris-ing, therefore, that technological progress has been rapid, and that the Fischer-Tropsch synthesis i s already a well established industrial process in Germany. 3. II Theory. (A) Principal Reactions The Kogasin synthesis is , complicated by the fact that a single set of reactant gases will give a wide multiplicity of products, depending upon the control of the operating conditions. The following reactions w i l l indicate some of the more important possible syntheses, a l l of which have been more or less established industrially: (1) CO + H2 HCHO (2) CO + 2H CH OH ' 3 (3) CO + 3H0 CH + H O 2 4 2 (4) nCO -f- (2n)H2 ^ Cn H2n+1QH -r (n-l)H20 (5) nCO + (2n + l)H 0 =?=^ C H -t- nH?0 2 n 2n-h 2 * Fortunately the catalysts required and the equilibrium conditions for the formation of oxygenated compounds, such as formaldehyde and methanol, are such that these do not take place to any appreciable extent in the range of the Kogasin synthesis. The actual problems in this process are how to keep the formation of the Kogasin fraction (b.p. 50 - 200° C ) at a maximum and the formation of the volatile gases (i.e. methane, ethane, Sec. ) and the heavy waxes at a minimum. 4. (B) Equilibrium Calculations (l) Discussion Thermodynamic calculations w i l l be applied to the formation of hydro carbons by the hydrogenation of carbon monoxide. In particular , free energy equations are to be investigated for reactions of the following type: (1) CO + 3H 2 , CH4-t-H20(g) (2) 2C0 H~ 5H2 C -t- 2H20(g) (3) 3C0 •+ 7H2 ^ C?H8 ~t 3H20(g) or, in general, (4) nCO -+- (2m-l)H2 ^ C nH 2 n ^ 1 - H n^OCg) Carrying out a reaction of this type involves the following problems: (1) How fast can the reaction be made to take place; and (2) What is the maximum possible yield under given conditions of temperature and pressure? The rate of the reaction depends largely on the catalyst used, and hence must be determined experimentally. But before the efficiency of the catalyst can be tested i t i s necessary to find out under what donditions of temperature and pressure the reaction is thermodynamically possible of taking place, and what the maximum yield, would be given a favourable rate. It is also helpful to know the thermo-dynamics of other possible reactions, so that conditions may be adjusted to exclude undesirable side reactions. 3. (2) Calculation of Equilibrium Constant for Methane Reaction . CO -4- 3H0 CH H 0(g) 2 4 2 To calculate the equilibrium constant of the above reaction i t i s necessary to calculate the free energy change ( <3 ) as a function of the temperature. The method employed w i l l be the integration of the specific heat equations for the substances involved. The equations 2 used were determined from spectroscopy data by Bryant : 6.88 -+• .000066T -h .000000279T2 6.25 -h .002091T - .000000459T2 2 3.38 -+- .017905T - .O00004188T 2 6.89 -+- .003283T - .000000343T Combination of these equations in the usual manner leads to the following expression: = - 16.62 -+• .018 899 T •+ .000 004 909T2 To obtain an equation for the heat of reaction, recall (a) CO: ~ Cp s (b) H2: Cp 3 (c) C H C = 4 P (d) H20: Cp = and by integration * H - ^ r a T i - ^ ^ -r 3 to where <4M0is the constant of integration. To get the free energy equation, substitute for ^ H in the expression «/ (J = — J^=i. ^ ^ ~ which gives = _ // A , ^ . 4 / ; 7-^7- , 7 4 ^ = £g? -^/z^~r ~ ^7 7 - - ^JS ^ 1^ = -^1/7 TVw 7"- T - 2 - T " * 7 (2) o where I is another constant of integration. The final equation now contains two constants. These are evaluated in the following manner: (1) To evaluate ^ /L\ is calculated for the reaction from the heats of formation of the products and reactants at 18°C and 1 atm. pressure. (see appendix 1 ) ^ z i > l = - 5 0 , 5 0 0 cal. per mol. This value is substituted in equation (l) for heat of reaction, giving = - 4 6 , 5 0 0 cal per mol. (2 ) To evaluate integration constant I The free energy change for the reaction i s calculated for 25°G by the summation of the standard free 7. energies of formation (see appendix 2 }: & & Z S 8 ~ ~ 33,blO cal. per mol. This value and the value of are substituted in equation (2), giving T(29Q) " - 33, €>/o - -+- A Sg (^S>S)(23) /oS(zs>e) - - 14,460 I = - 48.5 The final free energy equation then becomes: To get the equilibrium constant from the free energy change, use the, relation Z±6 - - RT In K or, In K - (4) (2.303) RT Example: To get the equilibrium constant for the reaction at 200°C (473° A): Erom equation (3) = - 23,200 cal per mol. Substituting this value in equation (4) K = 5.4 x 1 0 1 0 8. The variation of the equilibrium constant with temperature is shown by the following table, using results calculated from equations (3) and ( 4 ) . Temp, deg. c K 25 1.4 x 102^ 100 1.02 X 1 0 1 7 200 5.4 X 1 0 1 0 300 2.8 x 105 (3) Extension to higher homologues. Oalculations of the equilibrium constants of several different hydrocarbons have been made by 3 Smith in a paper published in 1927. Although the cal-culations for methane made in this thesis were from more recent data, the results obtained were in close enough agreement to Smith's results to justify accepting his values for the higher homologues. Eigure 1 shows graphically the variation with temperature of the calculated values of the equil-ibrium constant for the reaction. nCO + (2n-*-l)H — ^ C n H 2 n 2 ^ 2 ° Analysis of Figure 1 shows that i t becomes increasingly easier to form the higher paraffin hydrocarbons than the lower members at a l l temperatures. It should be Figure 1 Variation of Equilibrium Constants with Temperature > 10. possible, given suitable catalysts, to form any of the paraffin hydrocarbons at atmospheric pressure. The tendency to form these compounds f a l l s off rather rapidly with increased temperature. (C) The Catalyst The catalysts most suited for the kogasin synthesis are the metals iron, cobalt, and nickel - these metals have the common property of forming carbides. As 4 outlined by Fischer the reaction is catalysed by the formation of metal carbides of the formula Me C at the active centres of the catalyst. In the presence of hydrogen the carbide is converted again into metal while (CELj) rad-icals are formed. The reaction in the case of cobalt and nickel below 200°C is as follows: 2H0 -+- CO -t~Ni C Ni,C0 -f- H0 HO c y p <- c d hi C -h H. >• Ni C -h (CH } 3 2 2 3 2 Additions to the true catalytic metals are materials that have characteristic chemical action. The addition of copper to cobalt and iron catalysts activates the catalyst since i t lowers the reduction temperature for oxides; however i t acts adversely in the presence of nickel. Other additions, such as manganese and aluminum, which are simultaneously precipitated with the* catalytic metal, appear to play a part in obtaining certain configurations or lattice structures of the metallic alloy i t s e l f . The use of . 11. kieselgur greatly aids In the distribution of the actual catalyst by an increase in the surface area of the contact material. In another article , Fischer claims the best catalyst to be a mixture of Ni - Mn - Al prepared by precipitating the oxides on kieselgur. This was the catalyst chosen for the present investigation. Details for preparation w i l l be given later. (D) The Influence of the Magnetic Field According to Bhatnagar there are theoretical reasons why an impressed magnetic f i e l d should influence chemical reactions. In chemical reactions, even i f no change of valence i s involved, there may be a change in the magnetic moments of the i n i t i a l and final products. If during the reaction the magnetic moments tend to decrease, then the presence of an external f i e l d w i l l try to "conserve" the magnetism, and so retard the reaction. On the other hand, i f the magnetic moment increases during the course of reaction, then the presence of a f i e l d w i l l further help the increase and so accelerate the rate of reaction. Also, in a magnetic fie l d the atoms tend to assume certain definite positions relative to cue another. Thus the collisions with-in the f i e l d should take place in a more orderly and directed manner, and the rate of reaction should change. It seems logical, therefore, that a chemical transformation from diamagnetic to paramagnetic (or less 12. diamagnetic) state should be accelerated, and that from a paramagnetic to diamagnetic or from feebly diamagnetic to more strongly diamagnetic state should be retarded. This result can be expressed symbolically by the following: (1) X/w > ]E! X ^ then reaction is accelerated M* by f i e l d . (2) 2! X ^ ^ ^ (reactants) From data in Chemical Rubber Co. handbook (27th. ed.) AH.u = - 19,100 - 76,926 1- 26,428 - 50,498 A/S2S>, - - 50,500 cal per mol. (C) Free energy change from free energies of formation. CO + 3H2 *- CH4 -h H20(g) /^G - A G° (products) — IE. ^ &° (reactants) From Chem. Rubber Co. handbook, (27th. ed.) AG%^ - 11,617 - 66,124 i- 32,510 - 33,614 •AG2gg 5 - 33,610 caliper, mol. AH. CO AK - 26,428 0 - - 32,510 Co = 0 APPARATUS see p. 17