Lat-"“5 .1 M " :r -_.____.. -....;. I . . ’- Midllgan 53:3 !_.0.. , .14... L“; €251.47 This is to certify that the thesis entitled A NOVEL THERMOBALANCE FOR HIGH PRESSURE GAS-SOLID REACTION STUDIES presented by Michael Henry Treptau has been accepted towards fulfillment of the requirements for M. S. 4133113.31" Chemical Engineering . o \ Major pro essor Date August 7, 1985 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU LIBRARIES ”— RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. A NOVEL TEERHOBALANCE FOR HIGH PRESSURE GAS-SOLID REACTION STUDIES By Micheel Henry Trepteu A THESIS Submitted to Michigan State University in partial fulfill-ent of the require-eats for the degree of MASTER OF SCIENCE Department of Chenicel Bn‘ineering 1985 ABSTRACT A NOVEL TRERMOBALANCE FOR HIGH PRESSURE GAS-SOLID REACTION STUDIES By Michael R. Treptau A thermobalance has been developed which is capable of operating up to 900°C and 600 psi. The apparatus, a variation of the hanging reactor thermobalance, has several advantages over traditional hanging basket designs. The primary advantages are operation as a packed bed reactor, and the placement of thermocouples directly in the sample. Several reactions have been run at 825°C and 300 psi, and the apparatus has shown the proper trends of weight loss vs. time. The accuracy, however, has thus far been unsatisfactory. The error in calculated reaction rates has been in the range of 0.1-0.2 grams/hour for reactions whose rates are 0.2-0.3 grams/hour based on actual total weight loss. The primary source of error appears to be a baseline weight shift which occurs when switching from purge gas to reactant gas. This shift is assumed to be due to the difference in the thermal conductivities of the two gases. To my parents, who have continuously supported and encouraged me throughout my education. ii ACKNOWLEDGEMENTS The author wishes to thank Dr. Dennis J. Miller for his guidance and support during the course of this project; and Mr. Kevin Nichols, for his excellent graphics program and for translating the curve-fitting programs into BASIC. iii TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES . . . . . . . . . . . . . . . . . . . viii CHAPTER I. INTRODUCTION . . . . . . . . . . . . . . . . 1 1.1 General Introduction . . . . . . . . . . l 1.2 Current Thermobalance Technology . . . . 1 1.3 Limitations of Hanging Basket Designs . . 5 1.4 Hanging Reactor Thermobalance . . . . . . 10 1.5 Limitations of the Banging Reactor Thermobalance . . . . . . . . . . . . . 12 II. NEW DESIGN CONCEPT . . . . . . . . . . . . . . 15 2.1 Conceptual Design of Reactor . . . . . . . 15 2.2 Evolution of Balance System Design . . . . 17 2.2.1 Configuration I . . . . . . . . . . 17 2.2.2 Configuration II . . . . . . . . . . 18 2.2.3 Configuration III . . . . . . . . . 20 2.2.4 Configuration IV . . . . . . . . . . 21 iv Page CHAPTER III. DETAILED DESCRIPTION OF APPARATUS . . . . . . 25 3.1 Reactor Design . . . . . . . . . . . . . . 25 3.1.1 Pressure Vessel . . . . . . . . . . 25 3.1.2 Insulation . . . . . . . . . . . . . 29 3.1.3 Reactor Tube . . . . . . . . . . . . 29 3.1.4 Heater . . . . . . . . . . . . . . . 30 3.1.5 Temperature Measurement . . . . . . 31 3.2 Balance System . . . . . . . . . . . . . . 31 3.2.1 Balance Arm . . . . . . . . . . . . 32 3.2.2 Weight Measurement . . . . . . . . . 32 3.3 Gas Flow System . . . . . . . . . . . . . 33 3.4 Attachment of Tubes and Nires to Reactor . . . . . . . . . . . . . . . . 35 3.5 Temperature Monitoring and Control . . . . 36 IV. OPERATING PROCEDURE . . . . . . . . . . . . . 39 4.1 Sample Loading and Preparatory Steps . . . 39 4.2 Experimental Procedure . . . . . . . . . . 41 4.3 Shutdown . . . . . . . . . . . . . . . . . 43 Page CHAPTER V. RESULTS AND DISCUSSION . . . . . . . . . . . . 44 5.1 Balance Behavior . . . . . . . . . . . . . 44 5.1.1 Sensitivity . . . . . . . . . . . . 44 5.1.2 Noise . . . . . . . . . . . . . . . 45 5.2 Reactions . . . . . . . . . . . . . . . . 46 5.2.1 Description of Reaction Conditions and Sample . . . . . . . . . . . . 46 5.2.2 Reactor Heatup . . . . . . . . . . . 46 5.2.3 Raw Data . . . . . . . . . . . . . . 49 5.2.4 Baseline Shift . . . . . . . . . . . 49 5.2.5 Final Results . . . . . . . . . . . 55 5.2.6 Discussion of Final Results . . . . 59 VI. CONCLUSIONS AND RECOMMENDATIONS . . . . . . . 61 6.1 Conclusions . . . . . . . . . . . . . . . 61 6.2 Recommendations for Modifications . . . . 62 6.3 Recommendations for Suitable Reactions . . 64 REFERENCES . . . . . . . . . . . . . . . . . . . . . . 66 APPENDIX A . . . . . . . . . . . . . . . . . . . . . . 67 APPENDIX B . . . . . . . . . . . . . . . . . . . . . . 72 APPENDIX C . . . . . . . . . . . . . . . . . . . . . . 76 vi TABLE C1. CZ. C3. LIST OF TABLES Sensitivity Test Results . . . . . . Summary of Reaction Results . . Summary of Data Aquisition Program Functions . . . . . . . . . . . . Data Aquisition and Graphics Program Data Aquisition Program for Tecmar Interface Board . . . . . . . vii Page 45 59 77 78 83 FIGURE 1. 10. 11. 12. 13. 14. 15. LIST OF FIGURES Hanging Basket Thermobalance (Taken from Reference 2) DuPont 950 TG (Taken from Reference 1) DuPont 950 TG with High Pressure Modification (Taken from Reference 1) Cold Hall High Pressure Thermobalance (Taken from Reference 6) . Thermobalance of Sears, et a1. [7] (Taken from Reference 7) . . . . . Ranging Reactor Thermobalance (Taken from Reference 2) . . . . . . . . . . . . . Sketch of Reactor . . . . . . . . Schematic of Configuration I Schematic of Configuration II Schematic of Configuration III . Schematic of Configuration IV . . Photograph of Thermobalance Schematic of Cover Assembly Gas Flow System Temperature Control System . . . . . . . . viii Page 11 16 19 19 22 22 24 28 34 37 FIGURE 16. 17. 18. 19. 20. 21. 22. 23. Page Typical Reatup Curve . . . . . . . . . . . . . 47 Raw Data for Run 1 . . . . . . . . . . . . . . 50 Raw Data for Run 2 . . . . . . . . . . . . . . 51 Raw Data for Run 3 . . . . . . . . . . . . . . 52 Typical Baseline Shift . . . . . . . . . . . . 53 Final Results of Run 1 . . . . . . . . . . . . 56 Final Results of Run 2 . . . . . . . . . . . . 57 Final Results of Run 3 . . . . . . . . . . . . 58 CHAPTER I INTRODUCTION 1.1 General Igtrodgctigg One of the most straightforward methods for the analy- sis of the kinetics of gas-solid reactions is to directly measure the change in weight of the solid sample as it re- acts. The technology for accomplishing this is highly de- veloped for reactions which take place at room temperature and atmospheric pressure or lower. Only in the past few de— cades, however, have advances been made in the development of laboratory scale equipment (thermobalances) to study the many important reactions which take place at high tempera- tures and pressures. 1.2 Current Thermobalance Technology Most efforts in thermobalance technology to date have focused on modifying commercially available microbalances by enclosing them in pressure containment vessels and iso- lating the balance mechanism from the high temperature zone by various means. Dobner, et a1. [1] review the early chro- nology of developments in this area. Gardner, et a1. [2] review some of these devices in more detail and discuss some of their associated problems and limitations. In order to understand the rationale behind this project, however, it is beneficial to review again some of the thermo- balances currently in use. The most prevalent thermobalance is currently the hanging basket thermobalance developed by Feldkirchner and Johnson [3], and later modified by Gardner, et al. [4]. With this configuration, the sample is placed in a small basket which is then suspended into a tube surrounded by a furnace. A sketch of the apparatus is shown in Figure 1. The balance assembly is isolated from the high temperature environment, and is protected from corrosive reactant gases by a counterflow of purge gas through the chamber contain- ing the assembly. The force on the balance arm is measured by a small mass transducer. The apparatus shown in Figure 1 is designed to operate up to 1000°C and 2000 psi. A second, somewhat more complex thermobalance has been developed by modifying the DuPont 950 or 951 Thermograv- imetric Analyzer [1,5]. This configuration eliminates the long tube necessary to isolate the balance from the high temperature zone by using a cooling coil underneath the bal- ance assembly and a purge stream of cold nitrogen. A sketch of the DuPont 950 TG is shown in Figure 2, and the pressure containment modification is shown in Figure 3. The particu- lar apparatus shown in Figure 3 is designed to operate up to 1100°C and 880 psi. Forgac and Angus [6] attempted to avoid the problems with mechanical weakening of the metal wall which occur when the pressure vessel is externally heated. They con— structed a cold wall thermobalance, in which an induction heating coil was placed inside the pressure vessel, and Ir [III/III], WINCH ASSEMBLY WElGHT TRANSDUCER 9 5 I ELECTRICAL FEEDS PURGE AND V : PRESSURE ENTRY ACCESS PORT NICHROME WIRE ‘fi'.\\\\\\\\‘\\\\\\\.‘-.‘s§“-.‘ . \\\< S ‘ g g COOL-OFF SECTION Ellllé 'L|!||; REACTOR TUBE PRODUCT GAS OUT--——— '- IIIIIIIIIII’: I 'l’ /IIII I’ll ’IIIIIIIIIIII’IIIIIIII /L_/D|P TUBE / / CE //FURNA / rug/com. BASKET. I 7 THERMOCOUPLES l I AA.\\\\\ INERT HEAT TRANSFER PACKING§ FEED GAS AND BOLTED CLOSURE 1 /THERMOCOUPLE ENTRY Figure 1. Hanging Basket Thermobalance (Taken from Reference 2) a S RUARTT sEAM STANDARD QUART [mm "”5““ FURNACE TUDE ”ELL JAR -\ fis L. }—7L—-T\ WIGHT ”mu“ SAMPLE PAN SLOTS FOR BEAM Figure 2. DuPont 950 TC (Taken from Reference 1) 4 \ WE FLOW —.:/.Z_ N Momma! —.r££:.‘ N {'[fiE '33:"? nusrauamKuos 411:." bun-UH?"— “NM? “fi? 4 C ”M? 76A \ \ T‘ J], — ----- J J Figure 3. DuPont 950 TC with High Pressure Modification (Taken from Reference 1) cooling water flowed through grooves in the walls of the pressure vessel, thus keeping it at or near room tempera- ture. Again, the balance assembly is protected by a water cooled plate. Temperature measurements were made using a two-color pyrometer which was focused on the sample through a viewing port in the lid of the pressure vessel. Figure 4 shows a sketch of the reactor. The particular apparatus shown was designed to operate up to 1700°C and 2000 psi in reducing atmospheres. Most recently, Sears, et al. [7] attempted to improve on the cold wall design by placing high temperature insu- lation between the heating elements and the pressure vessel walls. A sketch of the apparatus is shown in Figure 5. This apparatus can achieve temperatures of 1700°C at atmospheric pressure and pressures up to 1000 psi at lower tempera- tures. The maximum attainable simultaneous conditions were 1300°C and 450 psi due to altered heat transfer character- istics at high pressures. Operation in both reducing and oxidizing conditions is possible with this design. .3 imitations of Ban in Basket Des ns Differences in pressure containment, mass measurement, and heater type notwithstanding, all of the previously dis- cussed thermobalances fall into the category of hanging basket reactors. While they all are able to measure changes Figure 4. Cold Wall High Pressure Termobalance H O Lid Viewing Port 316 Stainless Steel Outer Shell A286 Inner Steel Liner Cooling Hater Grooves Electronic Balance Water Cooled Plate Gas Inlet Tube Induction Beating Coil ommamo’bww H Water Cooling Inlet H H O Be-Cu Insert H N Teflon Liner l0 NZ" 1 g a 4 I/2" (La/J F1 F1 U—UULILJLILJ jL lj[*\ EVE/iii r1 r1 Fl,f1 r1 r1 l IRFT ULJLJLJLJ Figure A. Cold Wall High Pressure Thermobalance (Taken from Reference 6) ' “MC“ .-, IUPTURE DIsc Ls. LOAD . I ' I 6- us INLET (PURGE) DALANCE LOAD l . ‘1. 6 ELECTRIC LEADS '"Lu' ‘ . E' - -- sALANCE SENSOR . ill I I \ SANPLE LDNERIND 5°z=gfl_.__:°‘ NICRoNoToR \ol. 0, ’0‘ I E . - GRAYLOC FLANGE "o I" SANPLE LOADING '0'" l I I . I 1.. _ 0 III ' v 1 ll / Y , I [E If __1 W '0'“ L005 _% 9 F REATIND ELENENTs STEAN INLET __ / ‘ rLATINUH NESN SANRLE NOLDER V‘ 6; % THERHOCOUPLES —G W 0A5 OUTLET INSULATION / \ _ LEVELING PADS Figure 5. Thermobalance ofSears, et a1. [7] (Taken from Reference 7) in weight of a small sample very accurately, they never- theless have some limitations inherent to the design concept. First, it is very difficult to directly measure the sample temperature. Because of the nature of the hanging basket design, a thermocouple clearly can never be placed directly into the solid sample undergoing reaction. Either a thermocouple is placed very close to the sample or optic- al pyrometry is used. In most cases, pyrometry is not a viable option from either a technical or economic stand- point, due to the difficulty in designing a high pressure reactor with an optical path to the sample. The other alter- native, placing a thermocouple very close to the sample, does not always give an accurate indication of the sample temperature, as was shown by Dobner, et al. [1]. Inaccuracy can result when the gas stream is colder than the sample, or when radiative hotspots are present. An elevated sample temperature could occur either from inadequate gas pre- heating, or if the solid is reacting rapidly in an ignited state because of the exothermicity of the reaction. The second major limitation of the hanging basket de- sign is the fact that the reactant gas flows around the basket and the sample, rather than directly through the sample. This means that one has little control over the magnitude of the external mass transfer resistances between the flowing gas and the sample basket. In addition, the reactor cannot be operated as either an integral or 10 differential bed reactor, since the majority of the gas bypasses the sample without ever coming near it. The third concern, which has received very little attention, is the possibility of catalytic effects from the basket itself. Baskets have been made of stainless steel or platinum, but often the material of construction is not even mentioned. Both platinum and many of the components of stainless steel, such as nickel, are catalysts. Their ef- fects on many reactions cannot be neglected. 1.4 gauging Reactor Thermobalance Gardner, et al. [2] attempted to overcome some of the limitations of the hanging basket design by constructing a hanging reactor thermobalance. With this design, the entire reactor and its contents are weighed during the course of a reaction. A schematic of the system constructed by Gardner, et al. [2] is shown in Figure 6. Their reactor weighs 300 pounds, the bulk of which is tared off by a counterweight suspended from the other end of a balance arm. A 0-1 1b. force transducer measures the difference in weight between the reactor and the counterweight. The sample size is there- fore limited to around one pound. A large furnace surrounds and heats the entire reactor, and the apparatus is designed to operate up to 1100°C and 1470 psi. This design presents a new set of problems, since the reactor must be weighed with the gas lines and thermocouple wires attached, but it SAMPLE INJECTOR SOLID SAM PLE INERT PACKING RE ACTOR REACTANT GAS SUPPLY Figure 6. BALANCE D ~0~ I TARE I 11 bO-I L8. FORCE TRANSDUCER DATA ACOUISTION AND PROCESS CONTROL 40 THERMOCGJPLES 6 PRESSURE TRANSDUCERS 5 ANALOG CHANNELS I MASS TRANSDUCER CONTROL 32 ON-OFF CONTROL 6 PROPORTIONAL CONTROL REAL TIME DISPLAY 5 CHANNELS L ....... 4 rC-J I O PNEUMATIC : I rtf‘ I . :II‘ I I' ' II I i' .I I: "II . IL- I I I I; ‘#1 PRODUCT sis LjNERNocoUPLEs FROM REACTOR To DATA ":0 AOUISITION GAS I I I —-I :L. Hanging Reactor Thermobalance (Taken from Reference 2) GRAPH CM ROMATO' IF. II TO DATA ACQUISITION O O N 12 also eliminates many of the problems associated with the hanging basket design. First, thermocouples can now be placed directly into the solid sample without adversely affecting the weight measurement, providing accurate sample temperature measure— ment with fast response time. This is only true, of course, if the external thermocouple wires also do not affect the weight measurement. Secondly, the reactor is now essen- tially a packed bed tubular reactor. All reactant gas must flow directly through the sample, which means external mass transfer resistances can be reduced by increasing the gas flow rate. Also, product gas analysis can be performed to extract kinetic information on the basis that the system is functioning either as a differential bed reactor or a plug flow reactor. 1.5 Limitations of the Ranging Reactor Thermobalance As previously stated, a new set of problems arises with the type of design presented by Gardner, et al. [2]. Most of the problems are associated either with the large size of the reactor or the fact that the entire reactor, in- cluding the sample, the gas in the reactor, and the vessel itself, must be weighed. The combination of these condi- tions means that a small change in weight of a large mass must be measured to extract kinetic information. As Gardner mentioned, this requires a balance with high sensitivity at large loads. With all gas lines and thermocouple wires 13 attached, the hanging reactor thermobalance of Gardner, et a1. [2] can detect a change in mass of approximately 0.5 gram. This means that for adequate accuracy, a large sample must be used. The large sample size leads to problems with concentration and temperature gradients within the sample, as well as prohibitive materials costs when expensive reactants are used. Also as a result of the large reactor volume, fluctua- tions in gas density, especially at high pressure, become important. This is because all the gas contained within the reactor is being weighed. Gardner and coworkers [2] re- ported that for CO: at 500°C and 1000 psi, a fluctuation of one psi caused an apparent mass change of 20 mg, and a fluc- tuation of 1°C caused an apparent mass change of 40 mg. This problem is clearly directly related to the reactor vol- ume, and means that the system requires very precise temper- ature and pressure control to avoid large disturbances in the weight measurement. The combination of external heating and weighing the entire reactor also leads to problems. Oxidation of the hot reactor walls which are exposed to air could cause signifi- cant errors in the weight measurement, although proper material selection should keep this effect under control. Gardner and coworkers used Inconel 617 with an aluminum/ chromium diffusion coating. Noise in the mass measurement due to convective cur- rents between the furnace and the hot walls of the reactor 14 is also a problem. Data from Gardner, et al. [2] show noise with magnitude on the order of one gram. Care must be taken to provide appropriate baffling in order to eliminate weight fluctuations from convective currents. The final problem is the possibility of either contamin- ation or catalytic effects from the reactor walls for cer- tain reactions. As was the case with the baskets used in hanging basket thermobalances, the reactor walls in Gard- ner’s balance may not be inert to all the types of reac- tions one might wish to study. Gardner and coworkers assumed that the large sample size would make wall effects negligible, but it is not clear whether they were referring to the fluid mechanics at the wall or its catalytic effects. This summarizes the various thermobalances currently in use, along with their associated strengths and limitations. Looking at the features of these balances leads to the con- clusion that the potential exists for combining some of these strengths along with some new ideas into a new thermo- balance which will eliminate the major difficulties associ- ated with collecting solid reaction rate data at elevated temperatures and pressures. CHAPTER 11 NEW DESIGN CONCEPT 2.1 Conceptual Design of Reactor It was desired to design an apparatus to carry out reac- tions simultaneously at 600 psi and 1000°C and accurately measure the sample weight. The new thermobalance design set forth in this work is intended to improve upon the design of Gardner, et al. [7] by accomplishing the following: A) Reducing the size of the reactor and the balance arm necessary to support it, thus increasing sensitivity and reducing the amount of solid sample needed. B) Employing a cold wall design by using internal heating surrounded by high temperature insulation to protect the pressure vessel. C) Providing inert reactor walls of nonporous alpha- alumina. A sketch of the reactor is shown in Figure 7. Materials of construction were chosen with the EzCOa-catalyzed carbon hydrogasification reaction specifically in mind, and are discussed in Chapter III. The gas flows in through fittings located on the cover of the pressure vessel. The gas then flows down through an annular space between the heater and insulation, where it is preheated. It then turns around and flows up through the alumina reactor tube containing the solid sample. The gas flows out through another set of fittings in the cover. The inside diameter of the reactor 15 16 Gas Inlet and Gas Outlet and Power Leads Thermocouple Probes -SJ'L_J Flange Alumina Reactor Tube \ Insulation Heater Stainless Steel Pressure Vessel Thermocouple Probes / //// Figure 7. Sketch of Reactor l7 tube is 1/2 inch and the heated zone is three inches long, allowing samples of up to three grams to be heated uniformly. The size of the reactor is limited by the capacity of the balance arm used in the weighing system. This was set at 30 pounds. On the other hand, the need for adequate insu- lation between the heater and pressure vessel walls limited the minimum size of the reactor. One inch of insulation was used, thus setting the outside diameter of the pressure ves- sel to be four inches. At steady state, this amount of insu— lation keeps the walls of the pressure vessel at a suffi- ciently low temperature so that significant weakening of 316 stainless steel does not occur (see Appendix B for heat transfer calculations). 2.2 Evolution of Bglance System Design 2.2.1 Configuration I Initially, the balance system was set up as shown in Figure 8. The reactor rested on a Mettler electronic bal- ance, and the difference in weight between the reactor and counterweights was measured directly. This design had two difficulties, both associated with the fact that the counterweights were hanging. First, it was difficult to make fine adjustments to the amount of weight which was tared off by the counterweights. Since they were hanging, it was not possible to simply move them closer or farther away from the fulcrum of the balance 18 arm. Fine adjustment could only be accomplished by adding or removing small weights from the counterweight. The second and more serious problem was that slight swinging of the counterweights caused undamped oscillations in the balance. This eventully resulted in the reactor bouncing rather violently on the Mettler balance. 2.2.2 Configuration II The problems from Configuration I were solved by set- ting up the balance system as shown in Figure 9. The coun- terweights rested on a plate mounted on the balance arm, rather than being suspended freely underneath the balance arm. This design gave accurate and stable weight readings until the reactor was heated. As the the pressure vessel walls heated slowly to a steady state temperature of about 190°C, they also expanded slightly, causing an apparent mass shift of up to 15 grams over the course of the four hour heatup time. The pen of the Mettler balance also be- came quite warm, even with a 3/8 inch thick piece of insula- tion between the pan and the reactor. After the weight stabilized as the reactor walls reached a steady state temperature, and the reactant gas was fed to the reactor, the weight reading began shifting again, causing large errors in the results of the experi- ment being run. The weight shift was on the order of 1-2 grams/hour. This phenomenon was assumed to be due to the difference in heat capacity and thermal conductivity Fulcrum-\\\ Balance arm I— I /7\'7Z'>77 Upper Platform Counterweight J T Reactor i'Mettler Balance /7777//////////////7/ Lower Platform Figure 8. Schematic of Configuration 1 Fulcrum——\\ I Counterweight L- \ J £‘> Balance arm Upper Platform Reactor Mettler Balance Lower Platform Figure 9. Schematic of Configuration 11 20 between the purge gas (helium) and reactant gas (hydrogen or carbon dioxide). Switching gases was thus suspected to cause the temperature of pressure vessel walls and top to shift to a new steady state value. 2.2.3 Configuration III The next step in the search for a workable balance con- figuration, in which the difficulties associated with Con- figuration II were eliminated, was to interchange the posi- tions of the reactor and counterweights. This configuration is shown in Figure 10. The reactor was isolated from the Mettler balance, preventing it from overheating, and the reactor could expand without exerting a force directly on the Mettler balance. With the tubing and wiring attached, however, expansion of the reactor was still detected by the balance. Two to three hours were required for the weight to stabilize after the heat was turned on. This was undesirable, since 1200: reacts with the carbon at elevated temperatures. When the reactor was switched from purge gas to reactant gas, a fair- ly reproducible weight shift occured. This apparent weight shift was much less severe than for the previous balance configuration, on the order of 0.1-0.2 grams per hour. A problem associated with placing the counterweights on the Mettler balance, rather than the reactor itself, is that a change in weight registered on the Mettler balance does not correspond directly to the change in weight of the 21 reactor and its contents. This is due to the fact that the two arms of the balance have different lengths. Correcting for this difference can be accomplished by placing a series of small weights on the reactor and noting the correspond- ing weight change registered by the Mettler balance. A scal- ing factor can then be obtained for converting the apparent weight change data into an actual weight change. The scal- ing factor for this configuration is 0.6-0.7 (apparent change/actual change), depending on the exact position of the reactor on the balance arm. 2.2.4 Configuration IV In an effort to eliminate the error due to weight shifts which occured during heatup and the switching of gases, the reactor was suspended underneath the balance arm. This configuration is shown in Figure 11. This con- figuration allowed the reactor to expand freely as the tem- perature of the pressure vessel walls increased. Unfortunately, a new problem developed with this configura- tion which made this goal difficult to achieve. The problem was similar to that of the first configura- tion, in that a slight disturbance introduced undamped oscillations into the system. The system appeared to have a resonant frequency of about 2 Eerz, with an amplitude of several grams. At room temperature, this effect was controlled by placing a foam rubber pad between the weights and the Mettler balance. At elevated temperatures, however, 22 Fulcrum—\\ Reactor /7%7_7 Balance arm Upper Platform I Counterweight Mettler Balance /7/////////////////// Lower Platform Figure 10. Schematic of Configuration III Fulcrum—~\\ Balance arm I \ , I ”9977 Upper Platform r—I Reactor Counterweight _ Mettler Balance Lower Platform Figure 11. Schematic of Configuration IV 23 convective currents made it extremely difficult to prevent oscillations, even with the use of baffles around the reactor. It was then decided that the most promising method for operating the thermobalance successfully was to use Configu- ration III, where the reactor rests on the balance arm. The photograph in Figure 12 shows the reactor in this configura- tion, and gives a clearer idea of the spatial arrangement of the various components. In this configuration, a correction must be made for the shift in the baseline weight which occurs when switch- ing from purge gas to reactant gas. The magnitude of this shift can be determined by running a blank experiment using a sample of inert alpha-alumina rather than carbon. The results of the experiments performed with this configura- tion are presented in Chapter V. Figure 12. Photograph of Thermobalance 24 CHAPTER III DETAILED DESCRIPTION OF APPARATUS 3.1 Reactor Design A sketch of the reactor is shown in Figure 7. Impor- tant dimensions will be mentioned as each particular compo- nent is discussed. Fully assembled, the reactor weighs 22 pounds. 3.1.1 Pressure Vessel The external pressure vessel consists of a seamless cylindrical shell, a bottom piece and flange welded on to the shell, and a cover which is bolted to the flange and sealed with an o-ring. The material of construction was chosen to be 316 stainless steel. This was preferred over other stronger and more expensive alloys because of its re- sistance to attack by hydrogen and its lower nickel content than alloys such as Inconel. The mechanical strength of 316 stainless steel decreases rapidly with increasing temperature [8], but this problem was mitigated by the use of proper insulation, to be described shortly. The design pressure of the pressure vessel, excluding the flange, is 1100 psi for a 200°C wall temperature. The required wall thickness can be determined by using the design equation for thin-walled cylindrical shells [8]. t = P2 (1) 25 26 where t = shell thickness, in. P = pressure, psi 8 = allowable stress, psi. S=13300 psi at 200°C [8]. E = joint efficiency, assumed to be 1.0 for a seamless tube. R = inside radius, 1.75 in. Therefore, t (1100)(l.75) = 0.15 in. 13300 — 0.6(1100) A 1/4 inch wall thickness was chosen for its commercial availability. The required thickness for the bottom of the pressure vessel is given by [8] t = d 0.13P (2) where d = inner diameter, 3.5 in. Therefore, t = 3.5 [0.13)]1100) = 0.40 in. (13300(0.8) Here, the joint efficiency E was assumed to be 0.8 in order to take the welds into account. A bottom thickness of 1/2 inch was chosen for its commercial availability and to provide an adequate safety factor. The flange has a thickness of 1/2 inch and is one inch wide. The design calculations are quite involved and are shown in detail in Appendix A. Due to an error in 27 calculation, the maximum operating pressure of the flange, and thus the pressure vessel, is 600 psi. For design purposes, the cover can also be considered a type of flange. The minimum thickness, however, was not set by stress considerations, but rather by the need for ade- quate thickness in order to properly install the various fittings. For this reason, a one inch cover was chosen. Based on the relative size of the cover to the bottom of the pressure vessel and the flange, this is clearly thick enough to withstand the maximum operating pressure of 1100 psi. The cover is fastened down by eight 5/16 inch high— strength bolts. The calculations for the bolt load are shown in Appendix A. The seal for the vessel is provided by a 4 3/8” ID x 1/8 inch thick Viton o-ring. All gas and electrical feed-throughs are located on the cover. A Conax multiple probe seal is used for the thermo- couple probe feed-through, and a Conax power lead gland provides the feed-through for the heater power wires. Swagelock fittings of various sizes and types provide access for gas flow. Due to the stringent space con- straints, the various fittings are stacked on top of one another such that the electrical leads enter the pressure vessel through the same holes in the cover as the gas inlet and outlet flows. Figure 13 shows a schematic of the cover assembly. 28 Thermocouple Probes Gas Inlet L. Conax Multiple Probe Seal Power Leads rJl I Conax Multiple \ I l ; Wire Seal r-- ‘..J L i——1 Gas Outlet .. l afl“. ‘7' Tee Tee I .4: L‘s J‘C—L‘ L R r' O—ring [_E:j V\Retaining Ring Reactor Tube Figure 13. Schematic of Cover Assembly 29 3.1.2 Igsulatign A suitable insulating material for this application must have very low thermal conductivity in order to meet the space and power requirements, while maintaining the stainless steel pressure vessel at a moderately low tempera- ture. It must also hold its shape and be able to withstand temperatures to 1000°C. The insulating material which meets these requirements is standard E-20 firebrick, suitable for use to 1100°C. It can be easily formed to fit snugly into the pressure vessel, and has been coated with a high temperature ceramic impreg- nant manufactured by Aremco Products, Inc. This provides a durable, nonporous surface for the otherwise fragile and dusty firebrick. The coating also prevents absorption of gases into the insulation. The heat transfer calculations shown in Appendix B indi- cate that for a reactor temperature of 760°C and one inch of insulation, the surface of the pressure vessel can be expected to reach approximately 130°C at steady state. Ex- perimental results using a foil thermocouple for measuring surface temperatures show that the surface temperature actu- ally approaches 175—190°C, depending on ambient conditions. 3.1.3 Reactor Tube The reactor tube is a 5 1/2 inch cylinder made of 99.78 pure, nonporous alpha-alumina. The inner diameter is 1/2 inch and the outer diameter is 3/4 inch. Alumina was chosen 30 for its inertness to a wide variety of materials and condi- tions, including attack by molten alkali metal salts. Con- clusive literature on the latter, however, is difficult to find. A stainless steel ring is cemented to the top of the tube with a high temperature ceramic adhesive, also from Aremco. The ring with tube attached is then pressed up against the bottom of the pressure vessel cover by a nut which slips over the outside of the tube and is screwed into the threaded hole on the underside of the cover (see Figure 13). A Ealrez o-ring provides a seal between the ring and cover, preventing short-circuiting of the gas flow. Inside the tube, the solid sample rests on a bed of 28- 48 mesh alpha-alumina powder. A plug of quartz wool sup- ports the contents of the tube. The height of the alumina layer can be adjusted to assure that the thermocouple probes lie directly within the reacting sample. The alumina also prevents molten alkali metal salts from coming into contact and reacting with the quartz wool. In order to prevent entrainment of fine particles, another layer of alumina is placed on top of the sample. 3.1.4 Heater The heater is a 750 watt Mightyband heater manufactured by the Tempco Electric Heater Corporation. It is three inches long with an inner diameter of 3/4 inch, thus 31 fitting snugly over the reactor tube walls. Heat transfer calculations given in Appendix B show that at steady state, roughly 100 watts are required to maintain the reactor at 750°C. Thus, the 750 watt heater was deemed to have ade- quate power to provide reasonably rapid heatup without being oversized from a temperature control standpoint. Experimental results show that the time required to heat a sample to 825°C is approximately four minutes, and the power required at steady state is about 200 watts. 3. 5 Tem erature Measurement Two Chromel-Alumel thermocouple probes are used inside the reactor to measure the sample temperature. Each is pro- tected by a .040 inch sheath made of Inconel 600, which has a maximum operating temperature of 1150°C. The probes are placed at differing depths within the sample in order to determine whether axial temperature gradients exist. It is also possible, although more difficult, to control the radial position of the probes with the use of alumina guide tubes inserted into the sample. 3.2 Balance Sgste; The photograph in Figure 12 shows the spatial arrange- ment of the balance system with the reactor in place and all lines attached. 32 3.2.1 Balance Arm The balance arm is part of a two-pan Toledo Scale with a 30 lb. capacity for each pan. All excess parts were re- moved until only the balance arm and its base remained. The balance arm is arrested by placing a small jack under each end of the balance arm. An aluminum plate is fastened to one end of the arm, on which the reactor or counterweights can be placed. Weights can also be suspended from either end of the arm through holes in the platform. Brackets extend from the fulcrum on both sides of the arm and con- tain fittings to bring the gas lines in at the fulcrum and then along the balance arm to the reactor. The balance assembly rests on a platform consisting of a 1/2 inch wood— en plate on top of a 1/4 inch steel plate bolted onto two triangular brackets of 1 1/4 inch angle iron. The entire assembly is mounted on a one foot thick concrete wall (part of the radiation shielded chamber in room 265) in order to minimize the effect of building vibrations. 3. . Ne' ht Measure ent A Mettler PE 360 top-loading electronic balance is used to measure the change in weight of the reactor’s contents with time. It has a capacity of 360 grams with measurement to the nearest 0.01 gram. Within this capacity, it also has a fine (Delta) range of 0-60 grams with a sensitivity of 0.001g. 33 The Mettler balance has a digital output which is interfaced to the RS-232 port of a Zenith personal compu- ter. The BASIC program shown in Appendix C samples the weight at a user-specified time interval and writes the data to a disk file for future plotting and analysis. The maximum sampling rate with this program is 1 Hz, which is sufficiently rapid to provide an essentially continuous weight vs. time curve of slow gas-solid reactions. In order to minimize the effect of air currents, the thermobalance is located in the corner of a small, 8 1/2 x 7 1/2 foot room. The two exposed sides and the top are enclosed with clear Lexan sheets which also serve as pro- jectile shields. 3.3 Gas Flow System Figure 14 shows a schematic of the gas flow system. The reactant gas is Grade 5 Hydrogen, with a maximum impurity level of 10 ppm. The hydrogen flows through a bed of 1/8 inch Linde 3A molecular sieves in order to further purify it. The pressure is controlled by the cylinder regulator, and the flowrate is controlled by a Linde Mass Flow Con- troller with a flow capacity of 200 SCCM. The flow con- troller has the capability to be interfaced with a personal computer for remote control and data aquisition. It also has four channels for blending applications. For rapid pressure relief, there is a line upstream of the reactor which leads directly to a vent. During normal 34 Emumxm 30gb mmo .OH ouswwm caucufiso Leuomom mmo Owuzm O>fim> o>~m> xmzlomuch Muensxm xmwo musuasm . ll o>~m> Heumfinwem fionucoo 3ofim enammmum 00 oh no A umsmcxm _ . _ . O>Hm> Loccwfizo _ mmo ucmuomom . Loggom on N. 3:828 Saga fl ‘ 3ofim cemecumz mm>ofim umfisowfioz LAMB newswaxo OAQEmm 3S operation, this line is closed by a shutoff valve. The system also contains a Fike rupture disk upstream of the reactor to protect the system from excess pressure due to regulator failure or operator error. Currently, the exit gas from the reactor flows directly to the vent, but this can be easily routed to a gas collec- tion system for analysis by gas chromatography. 3.4 Attachment of Tubes and Hires to Reactor The most critical factor in the successful operation of this particular thermobalance is the manner in which the tubes and wires are attached to the reactor. Since the tubes and wires will inevitably exert forces on the balance arm, it is essential that the system be designed in such a manner that these forces remain constant during operation. The gas lines, power, and thermocouple wires never come in contact with the balance arm itself, only with the bracket which extends from the fulcrum. One-sixteenth inch copper or stainless steel tubing is used in this portion of the system in order to minimize the force that the tubing may exert on the balance arm. The inlet tube, supported by a bracket located two feet from the balance arm, travels along the axis of the fulcrum until it reaches the bracket attached to the balance arm. There, it passes through fit- tings and then continues on to balance arm. The tube then makes a right angle turn down the length of the arm and then another turn up to the inlet of the reactor. The 36 outlet tubing follows a similar pattern on the other side of the balance. The power and thermocouple wires are attached in basically the same manner as the tubing, rest- ing only on the bracket and not touching the balance arm itself. 3.5 Temperature Monitorigg and Control The temperature of the sample is controlled by an Omega CN-2013 programmable temperature controller with PID con- trol. A schematic of the temperature control and electrical system is shown in Figure 15. The output of the controller is a one amp solid state relay. This relay operates a 10 amp solid state relay which switches the heater on and off. When the controller is properly tuned (see Chapter IV for typical tuning parameters), it is able to maintain the temperature within 11°C of the setpoint temperature. In order to conveniently display both temperatures with- in the reactor, the thermocouples are connected to an Omega Model 199 ten channel thermocouple thermometer. This model has built-in cold junction compensation, as does the temper- ature controller. This eliminates the need for any ice bath reference junctions. The thermocouple thermometer has an accuracy of il°C. For data acquisition purposes, the thermocouples are also connected to a Tecmar Lab Master multiple-channel in- terface board installed in an expansion slot of the Zenith personal computer. The board has 12 bit resolution, and 37 Emumxm ~0nucoo ouzumuOQEOH .nfi muswwm Aom Cu usauso Leaguezv Ecuomom zmfimm fimcLOucH U<> ONH nTllllllllAf/l/ LoflfioLucoo ensemLOQEOH wwwEo mmfiom Oumuwlcwfiom mouflz Lozom mecw3 LOquoELOLH OLQDOUCELezk OHQDOUOELOLH g..........- 0.. Um o<> DON chcm oumeOucH umEumH 38 converts the thermocouple voltage to a digital signal, which is then sampled and converted to a temperature with software (see program in Appendix C). This summarizes the various components of the thermo- balance and their functions. A detailed operating procedure for the apparatus is presented in Chapter IV. CHAPTER IV OPERATING PROCEDURE The detailed operating procedure for the thermobalance is outlined in the following set of steps. Also refer to the operating manuals of the various components. 4,1 Sample Loading and Preparatory Steps With the reactor vessel removed from the balance assem- bly, open the vessel and place the cover assembly on a ring stand. Remove the heater from the reactor tube, and then re- move the reactor tube from underneath the cover. Support the tube on a stand using a test tube clamp. Weigh and load the quartz wool plug (0.06-0.lg), bottom alumina layer (1.5— 2.5g), sample (2-3g), and top alumina layer (~2.0g), respec- tively. The quartz wool is loaded from the bottom, while all other materials are loaded from the top with a long- stem funnel. The bottom layer of alumina is added to a depth such that the tips of the thermocouple probes will be located within the reacting sample. Reattach the tube to the underside of the cover, sliding the thermocouple probes into the powder sample. Make sure that the o-ring is in place, and firmly tighten the nut finger tight. Do not twist the reactor tube, since the thermocouple probes are now imbedded in the solid sample. Replace the heater, ro- tating it into position such that the power wires are coiled underneath the nut. Close the pressure vessel, making sure that the top end of the heater is aligned into 39 40 the slot in the insulation to allow the cover to close freely. At the same time, make sure that the bolt holes are aligned. If not, rotate the heater until alignment is achieved. Make sure that the balance arm is fully arrested on both ends. Place the reactor assembly onto the balance arm and attach all gas lines, power and thermocouple wires. Make sure that no wires touch anything until they reach the bracket two feet from the apparatus. Turn on the Mettler balance. Remove the jack from un- der the counterweights and replace it with the Mettler bal- ance, making sure that the Mettler is centered under the counterweights. Lower the weights onto the Mettler balance very slowly in order to prevent bouncing. This is very im- portant, since only very small oscillations are damped. If the Mettler does not register 250 $20 grams, raise the weights off of the Mettler balance and arrest the balance arm. Adjust the position of the reactor by sliding it away from or toward the fulcrum so that when the weights are lowered, the Mettler registers 250 :20 grams. After the weight reading has stabilized sufficiently, switch the Mettler into Delta Range. Close the draft shield and begin the helium purge flow at 5.0 SCCM and 50 psig, following the instructions found on page 18 of the oper- ating manual for the Linde Flow Controller. Make sure the reactant gas pressure at the regulator is below the intended operating pressure. If not, partially 41 open the relief valve shown in Figure 14. Switch the three-way valve to reactant gas in order to bleed down the pressure, then switch back to helium and close the relief valve. Allow the system to sit overnight for 12-18 hours in order to purge all oxygen and stabilize the weight. After the weight has stabilized, place a 100 mg cali- bration weight on the reactor and remove it several times to determine the average weight change detected by the Mettler balance. This is the calibration factor to be used in conjuction with the data collected from this run. 4.; Experipental Procedure Turn on the power to the temperature controller and allow it to warm up. Make sure that the output is turned off and the controller is in AUTO mode. Also turn on the power to the thermocouple thermometer. The flow controller and Mettler balance will already have been running during the stabilization period. Set the purge gas flow rate to the desired value, typically 100-200 SCCM. Raise the pressure at the regulator to approximately 100 psi below the intended operating pressure to allow for pressure rise during heatup. If the setpoint temperature is to be different from that of the previous experiment, it can be changed by pressing either the up or down arrow on the keypad of the temperature controller. If the setpoint is much different from that of the previous run, the controller may have to 42 be retuned. This can be accomplished by following the directions found in Section 3.10 of the operating manual for the temperature controller. Tuning parameters for 825°C are: proportional band, 88; reset, 0.87 R/M; rate, 0.18 M; cycle time, 15 sec. Since tuning is accomplished experimen- tally, this should be performed before a reacting sample is loaded into the reactor. Turn on the heater by pressing the START/STOP key and turning on the OUTPUT switch. After the temperature reaches the setpoint, raise the pressure at the regulator the rest of the way to the desired operating pressure. Also make sure the reactant gas is at the desired operating pressure. It will now take 2 1/2 to 3 hours for the weight to stabilize. During this period, prepare the computer for data aqui- sition. Turn on the computer and load the BASIC program ”GRAPHM". Choose the data aquisition mode from the initial menu and follow the instructions given. After the weight has stabilized, begin data aquisition by pressing the "Return” key and switch to reactant gas by turning the three-way valve. Change the flow rate if de- sired. If the pressure begins to drop, adjust the regulator accordingly. Allow the experiment to run until the desired time or extent of reaction has been reached. 43 4.3 Shutdown After the experiment has been completed, terminate data aquisition by pressing capital ”S" on the keyboard. Turn off the heater by following the reverse procedure for turning it on, and switch back to purge gas. When the re- actor has cooled and the pressure has been relieved, the reactor can be removed from the balance by using the re- verse procedure for placing it on the balance. It can then be opened and the contents removed and weighed. After the sample has been removed, reattach the reactor tube and heat in air to approximately 750°C in order to remove the sample residue sticking to the walls of the tube. The reactor is now ready for the next experiment. CHAPTER V RESULTS AND DISCUSSION The following results, with the exception of the sensi- tivity tests, were obtained while the balance was in Config- uration III, which was determined to be the best configura- tion of the four tested. The reasons for this choice will be discussed in the next chapter. 5.1 Balance Behavior 5.1.1 Sensitivity Sensitivity tests were performed while the thermo- balance was still in Configuration II (see Chapter II for a description). The results should also be applicable to Con- figuration III, however, since the positions of the reactor and counterweights were simply interchanged. The sensitivity tests were performed by placing a cali- bration weight on the reactor and recording the change in weight detected by the Mettler balance. The calibration weight was then removed, and the change in weight recorded once again. Since there was a considerable amount of noise present, the procedure was repeated several times and the results averaged. The results for the sensitivity tests performed with all tubing and wiring attached to the reactor are summar- ized in Table 1. These results show that the intrinsic error involved in detecting weight changes with this system is less than as for changes as small as 20 mg. In 44 45 Configuration III, part of this error is taken into account when the scaling factor to account for unequal balance arm lengths is determined. Table 1. Sensitivity Test Results Actual Weight Average Apparent Weight Std. Deviation 100 mg 94.5 mg 3.5 mg 50 mg 49.9 mg 1.8 mg 20 mg 19.1 mg 1.0 mg 5.1.2 Noise On a short time scale on the order of seconds, the mag- nitude of noise in the data is on the order of 20 mg for a sample temperature of 825°C. This is noticeably more noise than is encountered at room temperature, and is probably due to convective currents set up by the relatively hot walls (~180°C) of the reactor. On a longer time scale, the data exhibit noise of a larger magnitude. Excursions of 0.1-0.2 grams over the course of several minutes occasionally occur in what would otherwise be a fairly smooth curve. These excursions can be seen clearly in the data to be dicussed shortly. During the course of many tests, the thermobalance has shown itself to be very sensitive to small fluctuations in room tempera- ture. This was most dramatically demonstrated when the air conditioning was turned off during a run, and the reading on the Mettler balance shifted several tenths of a gram 46 over the course of the next half hour. This is assumed to be the primary cause of the long term noise. 5.2 Reactions 5.2.1 Description of Reactiop;Conditions and Sample Three reactions were performed at 825°C, 300 psi and a hydrogen flow rate of 100 SCCM while the thermobalance was in Configuration III. In each case, the solid sample con- sisted of 2.00 grams of 50-200 mesh activated cocoanut char- coal impregnated with 22.96X by weight K2C03. For clarity, the three reactions are referred to as Runs 1, 2, and 3, in the order in which they were performed. Runs 1 and 2 lasted two hours, while Run 3 lasted three hours. The actual total weight loss of sample during a reaction was determined by weighing the carbon sample, alumina powder, and quartz wool plug before and after the reaction. All runs were performed according to the procedure detailed in Chapter IV. 5.2.2 Reactor Heatup Before switching to hydrogen, the reactor was heated in helium at 825°C and 300 psi for 2 1/2 hours to allow the reading on the Mettler balance to reach a steady state value. The helium flow rate was approximately 100 SCCM. A typical heatup curve showing the reading of the Mettler balance vs. time is shown in Figure 16. 47 Weight (9) \_1 o I l i I 0 3O 60 90 120 150 Time (min) Figure 16. Typical Heatup Curve In order to know the initial weight of the sample at the time hydrogen is introduced into the reactor, it is necessary to determine how much of the sample is lost dur- ing the heatup period. To accomplish this, experiments were carried out where the reaction was stopped after the heatup period, rather than switching to hydrogen. The sample was then removed and weighed. It was found that 0.43 grams of the sample is lost during this time. The following mechanisms are assumed to contribute to the weight loss which occurs during the heatup period: 48 A) Reaction of the catalyst with the carbon by the reaction K2003 + ZC = 2H + 300. (3) If all the catalyst reacts, and all the free potassium deposits on the reactor tube walls, the maximum measured weight loss due to this mechanism would be 0.28 grams. This is not believed to occur, however, because cooled samples which are exposed to air after reaction become quite hot. This heating is most likely caused by the reaction of free potassium with oxygen, and is evidence that some potassium remains on the carbon. B) Volatilization of the water of hydration present in the potassium carbonate. This amounts to 0.06 grams per water molecule of hydration, or a maximum of 0.18 grams if the E200: is all in the form of [2003-3H20 (the highest hydrate form). C) Reaction of carbon with the water and oxygen pre- sent as impurities in the helium purge gas. Theoretically, this should be negligible, since the maximum amount of water plus oxygen in Grade 5 helium is 2.5 ppm. D) Vaporization of adsorbed water or other impurities from the activated carbon. E) Vaporization of adsorbed water from the alumina bed. The weight loss due to the last two mechanisms was determined by heating an uncatalyzed sample in helium at 49 30 psi for 3 1/2 hours. The helium flow rate was 100 SCCM. Under these conditions, it was found that 0.09 grams were lost. Assuming that a negligible amount of carbon reacts with impurities, this amount corresponds to the devolatili- zation of water or other impurities adsorbed on the carbon and the alumina. 5.2.3 Raw Data The raw data for the three reactions are shown in Fig- ures 17-19. The absolute values of the weights are arbi- trary, and depend on the exact time after placing the re- actor on the balance arm that the Mettler balance was switched into Delta Range. Since the reactor rests on the opposite end of the balance arm from the Mettler balance, positive slope indicates weight loss in the reactor. 5.2.4 Baseline Shift In addition to the reactions, blank runs were per- formed in order to determine the shape of the baseline shift caused by the switch from purge gas to reactant gas. Inert alpha-alumina was used in place of the KecOa/carbon sample for these runs. Figure 20 shows a typical baseline shift which occurs when switching from purge gas to reactant gas. This blank run is one of four which were performed between Runs 2 and 3. All of them were nearly identical in shape. By a 50 Aswev oEwH H 53% MOM wumo 3mm ONH co Nu me em 0 n . q _ _ _ o . . N.o I a ¢.O . I a l Coo 1 a I l ®.O _ _ p _ L _ va .AL auswha Seweas 51 ONH 2:: mm weak we N cam new mama —1 3mm .wH epswwm Awe uzwflaz 52 owfi OOH Acwev DEHH wo~ m cam Lem mumo 3mm .oH euawfim om - _ 53 Sufism Dansamam Sadaaxe .oN Shawna ACLEV DEAF Ow“ JO“ wOH NB on O a . n 1 . Li. J R T o I ‘\% o . i ~.o v . ¢ 1 «.0 A ... . . . \Si 4 Amy ucwmm3 . r . 0.0 .. A t A w.O I l _ _ h _ _ . L L _ 54 least-squares fit, the third order polynomial which best fits this curve was determined to be w = 9.0x10’7t3 - 1.06x10"t2 + 9.23x10’3t - 0.0334 (4.a) Equation 4.a was then subtracted point by point from the data of Runs 2 and 3. A different polynomial was used as the baseline shift of Run 1. In this case, a polynomial was constructed which reflected the average baseline shift of three blank runs which were performed prior to Run 1. This polynomial is w = 3.28x10’7t3 - 1.06x10“t2 + 9.23x10‘3t - 0.0334 (4.b) As mentioned in Chapter II, this baseline shift, which appears to be the major source of error, is apparently caused by differences in the physical properties of the two gases being used. Degassing of some sort from the reactor or insulation can be ruled out, since the magnitude of the shift is larger than the maximum weight of gas which the re- actor could contain. In Chapter VI, the majority of conclu- sions and recommendations will be focused on minimizing this shift and making it more reproducible. 55 5.2.5 Final Results In order to arrive at a final plot of weight vs. time, several steps of data manipulation are required. These steps are outlined as follows: A) The ordinate is shifted vertically so the curve begins at the point (0,0). D) From the resulting curve, the polynomial which corresponds to the baseline shift caused by switching from purge gas to reactant gas is subtracted. C) The ordinate is rescaled to account for the fact that the actual weight change is ~68X of the recorded weight change (see Chapter II for discussion). D) The weight values are subtracted from the initial weight of 1.57 grams in order to convert to a curve which represents weight of sample vs. time. The value of 1.57 grams is the initial weight of the sample after subtracting the amount lost during heatup (0.43 grams). The final results of the three reactions are presented in Figures 21—23 and summarized in Table 2. 56 H com “0 muflsmem answm .HN euswfim ACMEV mewe ON“ 00 NB we ON 0 . J 4 _ _ a A 1 q o r l 1 J .V.O 1 1 1 . w.o r 1 va ucwww3 Oocmfimm mmmz oh . mewcuooo< ncwoavcm n N.H I l t . o.H . v I p _ _ _ _ _ _ . _ 57 ONH ACHEV DEMH N cam mo mu~3mom Hmcfim oucmdmm man: 03 wcficuoou< hawoaccm co we me am 0 fi _ a W _ — — q 0 I L f . «.0 u . . A To L .NN muswha Awe Dzwhaz 58 m com “0 mafismom Hmch .mm ouswwm ACHEV OEWH owfi OOH wofl mm on o H A d — _ J1 AI _ a O V l 1 0.0 E l s a w.O % 1 n va ucwwoz n A N.H E l v Oucmdmm mmmz oh . o.~ mcwcpcoo< ucHoaccm fi- 1 _ _ h P _ P _ _ _ 59 Table 2. Summary of Reaction Results Run 1 2 3 Sample weight (g) 2.00 2.00 2.00 Heatup time (min) 145 144 150 Height loss during heatup (g) 0.43 0.43 0.43 Reaction time (min) 120 119 180 Total weight loss (E) 0.78 0.75 1.11 Actual amount reacted (g) 0.35 0.32 0.68 (Total weight loss - Loss during heatup) Apparent amount reacted according to thermobalance (g) 0.2 0.6 1.2 5.2.6 Discussion of Final Results The weight vs. time curve of Run 1 is virtually flat for the first 90 minutes, but this does not necessarily mean that no reaction occured until 90 minutes had elapsed. A more plausible conclusion is that the baseline shift which was subtracted from the raw data differs signifi- cantly in shape from the baseline shift which actually occured. In fact, the curve for Run 1 looks quite reason- able after simply applying the scaling factor of 0.68 to the data. The curves for Runs 2 and 3 are virtually identical for the first two hours, the entire length of Run 2. The actual total weight loss for two hours, however, is about 0.32 grams, or a factor of two less than the weight loss indicated by the curves. After three hours, the error in 60 the total weight loss increases to 0.57 grams for Run 3. Again, the magnitude of these errors suggests that they are caused by the uncertainty in the magnitude of the baseline shift. CHAPTER VI CONCLUSIONS AND RECOMMENDATIONS 6.1 Conclusions Several reactions have been successfully carried out at 825°C and 300 psi, and the reactor has proven to be very durable. The thermobalance has recorded the proper trends in weight loss, but the accuracy has thus far been unsatisfactory. In spite of its disadvantages, Configuration III appears to provide the best opportunity for success among the various configurations tested. This configuration has the shortest heatup time and the smallest baseline shift, and there is no chance of overheating the Mettler balance. The primary disadvantage is that the reactor must be re- moved from the balance arm in order to open it and change samples. The reason that this is such a problem will be discussed shortly. The 2 1/2 hour wait required to allow the weight to stabilize during heatup is unacceptable for reactions such as the primary one for which this reactor was designed, namely, EzCOa catalyzed hydrogasification of carbon. At elevated temperatures, E2003 reacts with carbon, meaning the catalyst loading and weight of carbon at the point where the gasification reaction is begun will not be the same as their initial values. While the thermobalance is sensitive enough to detect weight changes on the order of 10 mg, the mass vs. time 61 62 trace of an actual reaction still has a large amount of error associated with it. The error in the overall mass balance is on the order of 0.1-0.2 grams over the course of a two hour reaction, giving an uncertainty in the reaction rate on the order of 0.1—0.2 grams/hour. The bulk of this error is assumed to be due to the uncertainty in the shape of of the baseline shift which occurs after switching from purge gas to reactant gas. This can only be measured by performing blank runs before the sample is loaded and after the sample has been removed following a reaction. Unfortunately, all gas lines, power and thermocouple wires must be detached from the reactor, and the reactor must be removed from the balance arm each time a sample is changed. This means that for any new sample, the baseline shift cannot be determined exactly, since the exact position of the reactor and attached lines cannot be consistently reproduced. Fluctuations in room temperature also seem to affect the shape of the baseline shift. 6.2 Recommendations for Modifications Work will continue in an effort to improve the accu- racy of the thermobalance. The following recommendations, where feasible, will be implemented. Most of these recommen- dations are directed at reducing the magnitude of the base- line shift and increasing its reproducibility. 63 A) The thermocouple and power wires should be made of the smallest guage wire allowable for the current involved. This includes using uninsulated wire wherever possible in order to minimize the force the wires exert on the balance. The fine wires would have to travel directly from the wall bracket to the reactor in order to avoid touching anything. B) The 1/16 inch stainless steel tubing leading away from the fulcrum of the balance on both sides should be re- placed with more flexible copper tubing. C) The temprature and air flow in the room where the thermobalance is located must be more precisely con- trolled in order to eliminate some of the long term fluc- tuations in the weight reading. D) In order to maintain the walls of the pressure vessel at a lower temperature, the outer portion of the high temperature insulation should be replaced with an insulation which has a much lower thermal conductivity. E) The placement of radiation shields on the inner surface of the insulation and the underside of the cover may help to reduce the outer wall temperature of the reactor. F) The reactor should be longer in order to provide a longer cool-off section between the top of the reactor and the heated zone. This would of course require the construc- tion of a new pressure vessel. The benefits of doing this may not be as pronounced as expected, however, because at 64 the low gas flow rates involved, the dominant mode of heat transfer appears to be conduction. G) Using nitrogen instead of helium as a purge gas may reduce the magnitude of the baseline shift since, like hydrogen, nitrogen is a diatomic molecule. 6.3 Recommendations for Suitable Reactions The accuracy of the thermobalance would be improved considerably for the case of faster reactions. Assuming the error in guessing the shape of the baseline shift remains unchanged, the relative error in the measured rate of a reaction which goes to 100* conversion over the course of half an hour or so would be an order of magnitude less than for the reaction which was used for testing this apparatus. Reactions which take place at lower temperatures should also be more suitable for this apparatus. Lower oper— ating temperatures imply lower steady state wall tempera- tures, which would tend to reduce the magnitude of the base- line shift. Since such a long heatup time is involved, reactions whose reactants are stable at high temperature in an inert atmosphere are more suitable for study with this apparatus. If EacOa-catalzyed carbon hydrogasification is performed, a sample which has been placed in helium for the length of the heatup period should be analyzed to determine the ac- tual amount of catalyst remaining at the time the switch to reactant gas is made. This can be accomplished by 65 radioactive labelling of the potassium. If it is found that nearly all of the E2003 reacts with the carbon during the heatup period, it would be pointless to further study this particular reaction with this apparatus. In conclusion, this thermobalance has not yet produced results with an acceptable amount of accuracy. With further modifications, however, the potential still exists for ex- tracting accurate kinetic information for certain gas-solid reactions. LIST OF REFERENCES 66 LIST OF REFERENCES [l] Dobner, 8., Han, G., Graff, R. A., and Squires, A. M., Thermochimica Acta, lg, 251 (1976). [2] Gardner, N. C., Leto, J. J., Lee, S., and Angus, J. C., NBS Special Publication, 580, 235 (1980). [3] Feldkirchner, H. L., and Johnson, J. L., Rev. Sci. Instr., 29, 1227 (1968). [4] Gardner, N., Samuels, E., and Wilkes, E., ACS Advances in Chemistry Series, 1 l (1974). [5] Li, E., and Rogan, F. H., Thermochimica Acta, 26, 185 (1978). [6] Forgac, J. M., and Angus, J.C., Ind. Eng. Chem. Fundam., 22, 416 (1979). [7] Sears, J. J., Maxfield, E. A., and Tamhankar, S. S., Ind. Eng. Chgp. Fundam., 2;, 474 (1982). [8] Megyesy, E. F., Pressure Vessel Handbook, 6th Ed., Pressure Vessel Handbook Publishing, Inc., Tulsa (1983). [9] Shigley, J. E., Mechanical Engineering Design, 3rd Ed., McGraw-Hill, New York, Chapter 6 (1977). [10] Waters, E. 0., and Taylor, J. H., Mech. Eng., 49:5a, 531 (1927). [11] Holmberg, E. 0., and Axelson, E., Trans. Am. Soc. Mech. Eng., 51, 13 (1932). [12] Ereith, F., and Black, W. 2., Basic Heat Transfer, Harper & Row, New York (1980). [13] Carnahan, B., Luther, H. A., and Wilkes, J. 0., Applied Numerical Methods, John Wiley A Sons, New York (1969). APPEND I X A APPENDIX A PRESSURE VESSEL DESIGN CALCULATIONS Bolt Load Calculations The bolt load is the sum of the load due to pressure and the load required to seal the o-ring. The load due to pressure is the pressure multiplied by the exposed area. Load due to pressure =n(2.25 in)? (1100 psi) = 17500 lbf. (A1.a) The load to seal the o-ring equals the o-ring circumference multiplied by the force per unit length required to seal the o-ring. Load to seal =n(4.5 in)(100 lbt/in) = 1410 lbs. (A1.b) Therefore, the total bolt load is 18900 lbt. The proof strength of high strength (Grade 8) bolts is 120 kpsi [9]. To provide a safety factor, use 50000 psi as the working strength. Total bolt area required = 18900 lb: = 0.378 inz. (A1.c) 50000 psi The tensile strength area for a 5/16 inch (nominal) diameter bolt is 0.052 in? [9]. Therefore, eight 5/16 inch bolts are sufficient. 67 68 Flangp Cplculatiops The bolt circle diameter was chosen to be 5 1/4 inches, in order to ensure that the bolt holes are at least 2 1/4 diameters apart [10]. The width of the flange was then chosen to be one inch, in order to allow full seating of the bolt heads and nuts. The flange thickness was chosen to be 1/2 inch, but due to an error in the calculation, the thickness was underdesigned. The following calculations were redone in order to determine the maximum operating pressure using the 1/2 inch thick flange. Method of Waters and Taylor [10]: Pto = 2.858Wb Elzani + 0.1169 (A2) t2(ri-ro) Klz-l where b = Radial distance from bolt circle to center of o-ring, 3/8 in. E1 = ri/ro = 1.5. pto = Maximum stress in flange, psi. ro, r1 = Inside and outside radius of flange, respectively, in. t = Thickness of flange, 1/2 inch. W = Bolt load, lb. In the development of Equation A2, the modulus of elasticity was taken as 3x107 psi, and Poisson’s ratio as 0.3. 69 According to this equation, in order for the maximum stress in the flange to be less than the working strength of stainless steel at 200°C, which is 13300 psi [8], the bolt load must not exceed 7150 lbs. This corresponds to a maximum pressure of 450 psi. Holmberg and Axelson [11] showed that the maximum stress actually occurs at the junction of the flange with the cylinder wall. Since the flange is welded on, the fillet will distribute some of this stress, allowing a higher maximum pressure. The following calculations determine the stresses in the flange at 600 psi: s = VG? = 0.9682 (A3.a) T1 = t3(3a2 + 5d?) = 1.266 (A3.b) h3(d2-a2) 3 858t3 d2 b 2 2 T = ' [- ln(—)+ 0.1(b -a g = 0.1099 (A3.c) 2 3 2 2 3 a h (d —a ) 2 h3 R = (s -2? T1)(t+0.23255T1)p-2T2(h+0.5377s)P #L 1.865t + T [h‘(2+o.iibsr )+1.6103sh + 0.86652] 1 t 1 (A3.d) : -586.5 lb/in 2 2 h3 {h T1+1.865t)R+hT2P—O.5tp(s ~32T1) K = (A3.e) 1.5hT1—3.464t 7O Stresses in cylinder wall at junction with flange: 32350 psi (A3.f) 0= 6K/t2 1125 psi (A3-8) OQ= tip/2 Total stress in cylinder wall = <7+3 THEN 50 70 ON M GOSUB 220,2730,4000 100 GOSUB 600 110 GOSUB 1030 120 IF FLAG-0 THEN GOSUB ZOSOIGOSUB 2560 140 ZScINKEYS 150 IF 238"S" THEN 170 160 GOTO 140 170 CLSIINPUT "RESCALE GRAPH";B£ 180 IF 83¢"Y" THEN 100 190 INPUT ”RETURN TO MENU";BS 200 IF B$="N" THEN SCREEN 0,0,0,0,0:END 210 GOTO 40 220 REM DATA AQUISITION SUBROUTINE 225 CLSIPRINT “Data aquisition subroutine tor Mettler PE360 balance.“IPRlNT 230 TIBOIN-OIINPUT "ENTER SAMPLING INTERVAL IN SECONDS:",T1NTS 235 TlNTchNTS/bo 240 INPUT "ENTER NAME OF DATA FILE AND PRESS RETURN TO BEGIN.”,FILE$ 250 OPEN FILES FOR OUTPUT AS 02 260 TIME8=“OO:OO“ 270 ON ERROR SOTO 280 280 CLOSE l 290 OPEN "COM1:2400,E,7,1,LF" AS 01 300 INPUT 01,53 310 IF LENIS$><14, THEN 300 320 CLOSE 1 330 IF N>O THEN 350 340 TIMEzVALtLEFTs(TIMEs,2))sbo + VALIMIDSITIME3,4,2)) + VALIRIGHTSITIME8,2))/60 350 WEIGHT-VAL (MIDSISS,S,12)) 360 PRINT RIGHT:(S$,10),TIME$ 370 PRINT 02, NEIGHTI",";TIME 380 N=N+l 390 XT(N)=TIME 400 YTIN)-NEIGHT 410 IF N-IOOO THEN 480 420 T13T1+TINT 430 TlME=VAL(LEFT$(TIMEs,2))Nbo + VALIMID$(TIME$,4,2)) + VALIRIGHT$(TIME$,2))/60 440 IF TIME>8TI THEN 290 450 XSBINKEYS 460 IF xs="s" THEN 480 470 GOTO 430 480 CLOSE 490 BEEP:PRINT "DATA AOUISITION TERMINATED"IPRINT 500 PRINT N;"DATA POINTS COLLECTED":PRINT 510 INPUT ”PRINT LIST OF DATA";A$ 520 IF A38"Y" THEN GOSUB 284OTPRINT 540 INPUT "PLOT DATA”:AS 550 IF As="v" THEN 590 560 GOTO 190 590 RETURN 79 Table C2 (cont'd). CU” 610 020 635 act 670 51) 70:. 714') 720 730 740 750 760 771') 890 1030 1070 1041') 1070 1080 10%) 1100 1120 1130 1140 1150 1160 1170 1180 1190 1200 1210 1220 1230 1240 1250 1260 1270 1280 12‘?“ 1:04;) 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 14611 1 471‘) 1481) REM SET UP GRAPH BEEF INPUT "ENTER XHIN, XMAX, YHIN, YMAX TO SET UP AXES.“,XHIN,XHAX,YMIN,YMAX INPUT "ENTER ORDER OF CURVE FIT (PRESS RETURN FOR NO CURVE FIT):",M IF Mzn THEN FLAG=1 ELSE FLAG=0 GOSUF 2910 CLS LINE (59,0) - (59.190): LINE «59,01 - (639.0) LINE (59,190) - (639,190): LINE (639,0) - (639,190) FOR I = 0 TO 10 LINE (59,19*I) — (64.19%!) LINE «634,19§I) — (639,19.I) LINE (59+I*58,0) — (59*1*58.3) LINE (59+I§58,187) -(59+Ifi58,190) NEXT 1 RETURN RETURN 'plot x — y values FOR I=1 TO N XIZ=(PPS THEN GOTO 1430 SIMUL=0 RETURN 1R0w1= IRUW(F) JCOLI=JCOL(I) DETER=DETER¢PIVOI FOR J=1 TO MAX AJCOLk THEN A=0 THEN B(JCOLI)=A(IRONI,MAX) 1640 NEXT I 1650 ICZ=0 1660 NM1=M-1 1670 FOR I=1 TO NMI 1680 IP1-I+1 1690 FOR J=IP1 TO M 1700 IF JORD(J)>=JORD(I) THEN GOTO 1750 1710 JTEMP=JORD(J) 1720 JORD(J)=JORD(I) 1730 JORD=JTEMP 1740 ICZ=1C2+1 1750 NEXT J 1760 NEXT 1 1770 IF INT(IC%/21§2 <: ICZ THEN DETER=-DETER 1780 IF INDIC€=0 THEN GOTO 1810 1790 SIMUL=DETER 1800 RETURN 1810 FOR J-l TO M 1820 FOR I = 1 TO M 1830 IRONI=IRON 1840 JCOLI=JCOL(I) 1850 YZ(JCOLI)-A(IRONI,J) 1860 NEXT 1 1870 FOR I I 1 TO M 1880 A(1,J)=Y2(I) 1890 NEXT I 1900 NEXT J 1910 FOR I=1 TO M 1920 FOR J-I TO M 1930 IRONJ=IRON SYY = SYY + Y(I)”2 DUM=1 FOR J= 1 TO M DUM = DUM§X(I) SX(J)=SX(J) + DUM SYX(J)=SYX(J)+Y(I)§DUM NEXT J FOR J: MP1 TO MTWO DUM = DUM*X(I) SX=SX(J) + DUM NEXT J NEXT I FMSN CYY I SYY - SY*SY/FM FOR I = 1 TO M CYX(I)-SYX(I)-SY§SX(I)/FM A(I,MP1)-CYX(I) FOR J= 1 TO M IPJ=I + J A(I,J)'SX(IPJ)-SX(I>‘SX(J)/FM NEXT J NEXT I PPS=9.999999E-21 INDIC=1 GOSUB 1140 IF DETER 4} 0 THEN GOTO 2460 REGR=0 RETURN DUM 8 SY TEMP 8 CYY FOR I a 1 TO M DUM c DUM - B(I>*SX(I) TEMP = TEMP - B(I)§CYX(I: NEXT I AA=DUM/FM DENOM=N-M-1 REGR=S RETURN 82 Table C2 (cont'd). 2560 2570 2580 2590 2600 2610 2620 2630 2640 2650 2660 2670 2680 2690 2700 2710 2720 2730 2740 2750 2760 2770 2775 2780 2790 2800 2810 2820 2830 2840 2850 2860 2880 2890 2900 2910 2920 2930 2940 2950 2960 2970 4000 4010 4030 4035 4040 4050 4060 4070 4080 4090 4100 4110 4115 4117 4120 4180 4190 4200 4210 4220 'plot fitted line XX2=XMIN:YY2=AA:FOR J= 1 TO M:YY2=8(J)GXX2’J+YY2=NEXT J FOR I 3 1 TO 580 XX1=XX2 XX2=1*(XMAX-XM1N)/580 YY1=YY2 YY2=AA FOR J=1 TO M YY2=YY2+B(J)£XX2“J NEXT J Y12=190-(YY1-YMIN)*190/(YMAX-YMIN) IF YIZ 0 THEN YIZ=0 Y2%=190-(YY2-YMIN)4190/(YMAX-YMIN) IF Y2Z"0 THEN Y2Z=0 LINE (58+I,YIZ) — (59+I,Y2Z) NEXT I RETURN REM FILE READING SUBROUTINE CLS:PRINT "Subroutine to load file from disk into memory and plot data." PRINTilNPUT “INPUT NAME OF FILE TO BE READ.",FILE$ INPUT "PRINT LIST OF DATA“;D$ OPEN FILES FOR INPUT AS 01 N=1 INPUT «1, WEIGHT,TIME YT(N)=NEIGHT: XT(N)=TIME IF DS="Y" THEN PRINT HEIGHT,TIME IF EOF(1) THEN CLOSE: NTOT=N:RETURN N=N+1 GOTO 2780 REM PRINTING SUBROUTINE PRINT “NEIGHT(g) TIME(MIN)“:PRINT FOR I=1 TO N PRINT XT(I),YT(I) NEXT I RETURN REM SUBROUTINE TO REMOVE POINTS OUTSIDE SET LIMITS. J=O FOR I=1 TO NTOT IF XT(I)>=XMIN AND XT(I)<8XMAX THEN J=J+11X(J)-XT(I):Y(J)=YT(I) NEXT 1 N=J RETURN 'Subroutine to subtract baseline shift from raw data. CLS PRINT ”Subroutine to subtract baseline shift from raw data.”:PRINT INPUT "Enter coefficients of third order polynomial:",AA,BI,82,83 INPUT "Name of data file";FILE8 INPUT "Name of output file";NFILE$ OPEN FILE: FOR INPUT AS 01 FOR N=1 TO 1000 INPUT .1, YT(N),XT(N) IF EOF(1) THEN CLOSE: GOTO 4115 NEXT N OPEN NFILE! FOR OUTPUT AS 02 SHIFT = YT(1) FOR I=1 TO N YCOR(I)=(YT(I)-(AA+BI¢XT(I)+B2*XT(I>"2+830XT(I)'3)-SHIFT)/.6B URITE 82, YCOR(J),XCOR(J) NEXT 1 CLOSE RETURN 83 Table C3. Data Aquisition Program for Tecmar Interface Board 10 20 31:1 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 345 350 360 370 380 390 400 410 Data Aquisition Program for TECMAR Interface Board. KEY OFF:CLS PRINT ”Data Aquisxtion Program for TECMAR Interface Board.":PRINT 'Coeffic1ents of third order polynomial used to convert voltage to temperature. AA=15.985:BI=25.704:B2=~.0994:83=.001567 INPUT "ENTER GAIN(1,10,100,500):",GAIN IF GAIN=1 THEN B=128:GOTO 130 IF GAIN=10 THEN 8=129:GOTO 130 IF GAIN=100 THEN B=130zGOTO 130 IF GAIN=500 THEN B=131zGOTO 130 GOTO 60 'Input starting address of board in deCimal and the number of the channel to convert. ADDRESS=1808!:CHANNEL-1 ’Disable auto-incrementing, external start conversions, and all interrupts. OUT ADDRESS+4, 8 ‘Output channel number. OUT ADDRESS+5, CHANNEL 'Reset counter. SUMB0:N=O ‘Start a conversion. OUT ADDRESS+6, 0 ’Nait until bit 7 of status byte is a 1 Signaling done converting. 1F INP(ADDRESS+4) < 128 THEN 230 'Read in the data. LONIINPtADDRESS+51 HIGH-INP(ADDRESS+6) 'Convert from twos complement to a number between -10 and 10. X3256eHIGH+L0w IF X>32767 THEN XIX-65536! VOLTAGE-X01000/(204.8eGAIN) 'CONVERTS TO mV SUMsSUM + VOLTAGE N-N+1 IF N<50 THEN 210 VAVISUM/N Convert voltage to temperature TEMP-AA+B1*VAV+B2*VAV“2+B3*VAV”3 'Set up online display on screen. LOCATE 12,30,01PRINT “VOLTAGE =" LOCATE 14,26,0:PRINT ”TEMPERATURE = ” LOCATE 12,391PRINT USING "00.00_mV";VAV LOCATE 14,39:PRINT USING "000_ C “:TEMP X$=INKEY3:IF X$="S" THEN END ELSE GOTO 190