. .3 . v I ‘ v. 7:... .upuflflflnmrafifltlfit 1!. vgiolhq: v1 9‘79vuv11mtéc. . hawtfi . WW... . g r c ' z. 7' ,: . 9,7; ! a “7,3" 1 ‘ :‘ V t .. ‘ i . . I :2. ‘ . 1.... HQ. ‘ A . ‘3. N33! £1.12... l..2sl.h5bm.4fi% 1.....2... z vafifsvabo. U: .3 - fififlfiWg .éhfi? . .. bflbfifififivfinuhmwi 93,” 9413:}, 1!. . » +3.)? l~r8b§fiu€hwba§§uoih, . 2. .. a .1. y . Ll~zzth¢Ki {5.5.59.4 is 1‘ 2th.) a; .213»? .3 hzflfluwaibzziiti . ~ . , 2.. ‘3. x I a . 12.2.): ‘ ) 1|! wx \. . 12¢ 1. viii I.):Pkal$fil§.§x§ V112. l! .1. gur“: | \ a... y. .I . {42 I75! 11“ Vg..£ i1 .3 A , :1 \ ll... \3‘85. 2.211.131... )Kalgixéhgfi behWH-.2 » .fifix if? til... ’1‘ ,. ’00. , . t J .i .v (J: ((1)2..xf \t. ‘ kw}; .‘I 15:..3‘ » .,. rial); .. 7 .. :Is... £4,ng 9;.7Errrfvlnri , ’1 Ill.\’)‘ 11 \ 5:11... Jgip‘ligg‘é p. 3? .ithLJL.......n-1"AVAVO»| I. £27}!\v)}1\\fiugliyu§\‘1jll.1.)a\59|dfli}fiq{1‘zva\ zilggi I ‘ 1‘ l 1 {in}; f ‘ C K 1"}, .dflfinitfiUmnruivJohyflfihulluuzubio if; ‘ '.u.f..lfz)un . . (URN? .(aZsJuut‘we .. .‘Hrgipvassgii in. t! .l . 1);er .9»?! 5,.le In! .(rflu! ’3’» Vii‘rti. ‘vaziri‘phu y I 2‘ hfqiullnl’.‘ % 51: I)?” .‘IvXIV lav... K .. .315. :IS’. !. I..." c . (ll .7? . :Nfl 5-: a»). ...:u\?.asnl.4r ..... igirroflnfifl . . a)... lit! ! .... 1...? .x!:§.ni10¢-8ni 2.3.: aia\uu;¢»1‘v,¥.S 71.9.3012 r‘gvt ‘1‘;ka . 1 52g}; 111‘); .H. ‘ hand-II... wK .. r3W>l5p2¢V£fi¢f riff \gilflddlsxaaiatuiifighgzzixl) Lal..i»21:r4mv.mwm u . 2... ~ 2ft)“. audit}; {5?}? ’23.:31‘311‘51311’ :2 1H1.“ 5%)}: 5 . . It»! 1. . . . nigh): . , git. Ll» Dial}?! {.1011 .11..) I. J“?!Iv§f)|blfi\“fib“fl 3:. .V £3.38. Hunvv Ir?) 1 .. saucy. -. . {113m refit. . 1.1., If; . ‘ [-1.11 I . ‘ Falfuxtllt?‘ Lyifilvir: .f . , 4. {-5 I“: .215. glut!!! r ..L . 41 r- I, y . .1).lJ.I.l’I.‘I..,, fx‘gg‘llffxf’filfilto’) ... :34 .. : . ill». , .. A :HII?LLL{4)..}|~QI.3¢!.. .\.{er.ll.5 . . . .I I. l)!!r£;l3.w.t\uiftLanufilltiiflh 4.1).! fair}. ; I”: LHHuHuJflqfiry.‘chx . itiiiiiliizilijjlfliiilmimWm“ LIBRARY Michigan Sta :0 University THESfS This is to certify that the thesis entitled PYROLYSIS AND CATALYTIC GASIFI CATION 0F PORLAR SPP presented by Mark Roger Boyd has been accepted towards fulfillment of the requirements for Master's degreein Chemicai Engineering Major professor October 4, 1979 Date _________ 0-7639 OVERDUE FINES ARE 25¢ PER DAY g PER ITEM . 7 Return to book drop to remove my this checkout from your record. PYROLYSIS AND CATALYTIC GASIFICATION OF POPLAR SPP BY Mark Roger Boyd A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemical Engineering 1979 ABSTRACII PYROLYSIS AND CATALYTIC GASIFICATION OF POPIAR SPP By Mark Roger Boyd Two salts, Na2C03 and K2C03, have been investigated to determine their catalytic effects on pyrolysis and gasification of Poplar SPP and char derived from Poplar SPP. A differential reactor was used with temperatures ranging from 550° to 700°C, space velocities of 2 to 7.25 5-1, and steam partial pressures of 45 to 100 kPa. Pyrolysis has been found to be nondestructive to the basic cell structure of wood, except when impregnated with an alkali salt. 2, and CH 4. The char gasification rate law observed from steam partial Pyrolysis products are CO, CO pressure data is -moles C ‘ min Gasification rate = k exp (-Ea/RI‘) pH20 g Preexponential factors of 10.7 and 16.4 g — moles/g C min kPa and activation energies of 89.2 and 94.2 kJ/mole were observed for the catalytic gasification of char using Na2C03 and K2C03 , respectively. Char gasification products were H2 , C0, C02, and CH 4. Rates were compared considering methane as derived from carbon monoxide as opposed to the product of direct hydrogenation of carbon . Intraparticle diffusion becomes substantial for particles over 1.0 cm. Conversion of steam is not sensitive to mixing conditions within a tubular flow reactor . "gg‘ 4-”... To my parents ii ACKNOWLEDGEMENTS The author would like to express his appreciation to his academic advisor, Dr. Martin C. Hawley, for his encouragement and assistance. Also to Dr. Antonio Devera for his advice and guidance during the study. The author also wishes to thank Mr. Don Childs for his invaluable assistance in constructing and assembling the experimental apparatus, and Dr. John Young of Argonne National Labor- atory for timely advice on engineering the experimentation. Appreciation is also extended to Dr. James Hanover for his valuable expertise, and Robert Konopacz for obtaining the electromicrophotographs. The encouragement and support of the author's wife, Anita, is also sincerely appreciated. TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . LIST OF FIGURES . . . . . . . . . NOMENCLATURE . . . . . . . . . . INTRODUCTION . . . . . . . . . . PHYSICAL AND CHEMICAL PROPERTIES OF POPLAR PYROLYSIS AND GASIFICATION BACKGROUND . Pyrolysis . . . . . . Gasification of Carbon and Char . . Early Pyrolysis and Gasification Experiments EXPERIMENTAL CONDITIONS . . . . . . Introduction . . . . . . . . Reactant Gas Production . . . . . Reactor Assembly . . . . . . . Product Gas Processing . . . . . Gas Analysis . . . . . . . . Sample Preparation . . . . . . Operating Procedure . . . . . . EXPERIMENTAL RESULTS . . . . . . . Effects of Pyrolysis . . Salt Catalyzed Wood Char Gasification MODELING . . . . . . . . . . . Single Particle Effectiveness Factor Tubular Reactor Axial Dispersion Model iv SPP Page vi vii ix l4 14 24 32 39 39 4O 42 44 45 46 47 49 49 53 64 64 71 Page SUMMARY AND CONCLUSIONS . . . . . . . . . . 81 Conjectures on Catalytic Activity . . . . . 8l Observations and Comments . . . . . . . . 82 APPENDICES . . . . . . . . . . . . . . 86 APPENDIX A: Future Work . . . . . . . . . 87 APPENDIX B: Sample Rate Calculation . . . . . 90 APPENDIX C: Effectiveness Factor Calculation . . 94 APPENDIX D: Summary of Rate Data . . . . . . 97 BIBLIOGRAPHY . . . . . . . . . . . . . lOl LIST OF TABLES Table Page 1. Summary of pyrolysis investigation . . . . 15 2. Summary of carbon gasification investigation . 25 3. Summary of rate data . . . . . . . . . 98 vi Figures 10. 11. 12. 13. 14. 15. 16. LIST OF FIGURES Microphotograph of Poplar at 200 magnification Schematic of section A in Figure 1 showing components of cell wall structure . Chemical structure of (a) cellulose and (b) hemicellulose . . . . . . Chemical structure of lignin . Reaction mechanism of char gasification . Observed gas flow from pyrolysis of K2C03 impregnated wood . . . . . Schematic of experimental apparatus . Picture of reactor assembly showing main reac- tor tube and steam preheater . Picture of reactor assembly . Electromicrophotograph of Poplar Electromicrophotograph of Poplar minute of pyrolysis . . . . Electromicrophotograph of Poplar minutes of pyrolysis . . . . Electromicrophotograph of Poplar minutes of pyrolysis . . . . Effect of steam partial pressure rate calculated from C0 and C02, Effect of steam partial pressure rate calculated from C0, C02 and SPP . SPP after 1 O O O SPP after 4 SPP after 8 on reaction case A on reaction CH4, case B Activation energy plot, in rate versus 1/T vii Page 7 ll 13 28 36 41 43 43 50 50 51 51 57 58 6O Figures Page 17. Effect of space velocity on reaction rate at 625°C for uncatalyzed material only . . . . 62 18. Drawing of flat slab geometry for a particle with external bulk film resistance . . . . 65 19. Plot of effectiveness factor versus particle size, k = 1.519 s-l, De = 3.19 cmZ/s . . . 68 20. Comparison of effectiveness factor for cata- 1yzed (k = 1.519 5-1) and uncatalyzed (k = 1.14 s'l) char, De = 3.19 cmZ/s . . . . . 7o 21. Drawing of a plug flow reactor showing a differential element . . . . . . . . . 72 22. Plot of 1 - X versus Peclet number. De = 3.19 cm /sec, k = 1.519 seC'l, L = 40 cm, V = 40 cm/s, Dp = 0.5 cm . . . . . . . . . . 74 23. Plot of l - X versus Peclet number. De = 3.19 cmZ/sec, k 1.519 sec-l, L = 40 cm, v = 40 cm/sec, Dp 1.0 cm . . . . . . . . . 75 24. Plot of l - X versus Peclet number. De = 3.19 cmz/sec, k = 1.519 5-1, L = 40 cm/sec, D = 2.0 cm . . . . . . . . . . .P . . 76 25. Plot of l - X versus Peclet number, De = 3.19 cmZ/sec, k = 4.464 5-1, L = 40 cm, v = 40 cm/ sec, Dp = 1.0 cm . . . . . . . . . . 77 26. Effect of residence time of steam on conversion, De 3.19 cmZ/sec, k = 1.519 s-l, L = 40 cm, Up 1.0 cm . . . . . . . . . . . . 79 27. Effect of effective diffusivity on conver- sion. k = 1.519 s’l, L = 40 cm, v = 40 cm/sec, D = 1.0 cm . . . . . . . . . . . . 80 P viii Bi GW‘G NOMENCLATURE k L Biot mass number, dimensionless,-Tg— e Concentration of species 1, gr - mole/cm3 Bulk diffusion coefficient of component i in j, cmZ/sec Effective diffusivity coefficient of component i, cmZ/sec . . L Damkohler number, dimen51onless, nk 5 Dimensionless concentration . -1 reaction rate constant, 5 Length, cm . . L Peclet number, dimen51onless = 5— z Dimensionless distance within the reactor conversion of steam Mole fraction of component 1 distance, cm Dimensionless concentration Dimensionless distance Thiele modulus, dimensionless Effectiveness factor Void fraction INTRODUCTION Gasification technology can be applied to biomass to produce synthesis gas, a mixture of CO and H2, which can be converted to a variety of fuels and chemicals. The objectives of this research are to determine and model physical and chemical changes which occur during high temperature pyrolysis and gasification of biomass. The kinetic parameters determined by the study can be coupled with mass transfer knowledge to describe and design the characteristics of a biomass gasifier. The size, pore structure, and composition of the biomass feedstock will all have significant effects on the rate of gasification and thus, design of commercial gasification units. Bio- mass gasification may become an important alternative for utilizing solar radiation for production of gaseous and liquid hydrocarbons for fuels and chemicals. "Biomass potential in 2000 put at 7 quads," is a current concensus among recent estimates (1). This amounts to 10 percent of the current energy consumption rate. Current DOE funding was noted to amount to $26.9 million in 1979 for biomass conversion, which is projected to grow to $118.9 million by 1981. Most economically feasible projects at this time are processes which produce expen- sive chemicals. Thermochemical gasification and lique- faction are currently important areas for study. This study is aimed toward defining unknowns related to appli- cation of gasification technology to biomass. Specifically this study is concerned with moni- toring physical and chemical changes of Poplar SPP under a variety of gasification conditions. Poplar SPP is a hybrid tree which has, due to its very rapid growth, been singled out as a likely candidate for cultivated biomass plantations. From a few preliminary experiments it was observed that two important processes occur in the gasi- fication of raw biomass, pyrolysis and gasification. From a process standpoint, gasification of raw biomass into synthesis gas will involve several steps(15): initial devolatilization, pyrolysis, and gasification. Upon being subjected to a high temperature environment (600° to 800°C) a temperature gradient begins to form, up to 100°C, within the particle which quickly devolati- lizes any adsorbed water and gases. As the intrapellet temperature gradient continues to increase (100° to 500°C) chemical degradation begins. This process of pyrolysis is the thermal degradation of the wood components into simpler molecules. Wood is composed predominately of hemicellulose, cellulose, lignin, and water. Typically the water content of green wood is 50 percent by weight. Cellulose, hemicellulose, and lignin are polymeric in nature. Poplar SPP is a typical hardwood which is chemi- cally composed of 48.8 percent cellulose, 19.3 percent lignin, and 29.7 percent hemicellulose, and 1 percent ash. The remaining 1.2 percent is composed of extractives. Decomposition involves first the breakdown of the polymers into monomer units, then the further degradation of the monomer units into an array of decomposition products. The degree of complexity of the final products is a func— tion of the heating conditions, the initial sample size, porosity, and the ash content. The pyrolysis products may also interact during the decomposition to form an even greater variety of products. Termination of this stage of the reaction sequence leaves a residual char composed mainly of carbon. Further heating of char will not produce any more gaseous or liquid components. The resi- dual char can be gasified or reacted with steam at 500°C and higher to produce synthesis gas. This process for wood char gasification is similar to coal gasification, but is somewhat different due to pore structure and ash content. Important parameters for the steam carbon reaction are steam partial pressure, temperature, and catalyst impregnation. Previous workers (17, 21, 22) have observed that impregnation of the biomass with potassium and sodium salts catalyzes the pyrolysis and gasification reactions. The experimental work of this study includes pyrolysis and gasification of nonimpregnated and salt impregnated biomass samples. Catalyst activity was investigated to determine effects on morphological structure during pyro- lysis, pyrolysis rate, and gasification rate. Pyrolysis experiments were performed at tempera- tures ranging from 400° to 700°C. Space velocity of the pyrolysis gas was 4 sec.1 based on a sample volume of approxi- mately 60 cm3 (five to seven grams of finely ground parti- cles) and a volumetric reactant gas flow of 240 cm3/sec. The final set of experiments performed were the measurements of reaction rates of catalyzed pyrolysis char. For these experiments, finely divided one to two mm par- ticles of Poplar SPP were subjected to pyrolysis for 10 minutes to produce char for the gasification experiments. The resultant char was then impregnated with a catalyst, and gasification reaction rates of the resultant material determined for values of steam partial pressure between 45 and 100 kPa, and steam temperatures between 550°C and 700°C. These experiments were modeled and the kinetic constants were determined. The kinetic constants along with estimates of mass transfer parameters were used to model more complex reactor systems. The original objective of this study was to gasify raw wood and model the process. However, after a series of preliminary experiments it was found that the liquid products of the pyrolysis process were too difficult to handle in a quantitative manner with our experimental apparatus. Therefore it was decided to examine the gasi- fication of the residual char, which is a much cleaner process. The experimental work was done jointly with Mr. Craig Anderson. Experiments were done on both salt impregnated and nonimpregnated wood char samples. Results of the nonimpregnated experiments are reported in Mr. Anderson's thesis. This thesis reports details of gasi— fication experiments for salt impregnated wood char and comparisons are made with the nonimpregnated results. This report also deals with some qualitative results of the preliminary pyrolysis experiments. PHYSICAL AND CHEMICAL PROPERTIES OF POPLAR SPP The cellular structure of poplar wood is shown in Figure l. The major distinguishable features are the parallel, closely packed cells. The intracellular sub- stance is composed predominately of amorphous lignin. The cell walls are divided into two main parts as can be seen in Figure 2. The primary wall (P) is an amorphous collection of microfibrils which are essentially noncrys- talline. The secondary wall is subdivided into the following three parts: (a) the outer layer (81) which contains a large number of cross hatched fibrils, (b) the middle layer (82) of the inner wall which constitutes the largest percentage of the cell wall volume, and (c) the inner layer of the secondary wall (S3). The 82 layer is composed of a large number of parallel fibrillar units oriented at a slight angle to the cell axis. The S3 layer is composed of parallel fibrils formed in a flat helix, nearly perpendicular to the fibrils of the 82 layer. The microfibrils are the basic physical entity in the cell wall construction. They are thin rods Section A Figure l.—-Microphotograph of Poplar at 200 magnification. S3 S 2 51 '~'\‘|V"0'g‘ 9". 9'. 30.379. “\‘. ....... .... é."’ “‘......‘.Q. ..o..” P .A.. .“’A““.. ' Figure 2.--Schematic of section A in Figure 1 showing components of cell wall structure. approximately 5 to 10 nm in diameter. Theories from as recently as 1970 have postulated the physical structure of the fibrils as alternating crystalline and amorphous regions. The highly crystalline regions are believed to be more pure in D-glucose than the amorphous regions which may contain high percentages of other sugar mole- cules. The actual polymeric compositions of the three different walls do not vary substantially. The primary distinguishing feature of the walls is the contrasting orientation of the microfibrils. X-ray data have shown (21) that there must be some type of crystalline imper- fections in the fibrils, but whether or not they are in discrete locations or homogeneous throughout the structure is not yet known. The cellulosic fibrils imbedded in a matrix of lignin and hemicellulose should provide a relatively high permeability barrier against radial diffusion. The ratio of axial to radial permeability of raw wood is over 100 (23). This property is important during pyrolysis and gasification of wood in a reactor, especially during the stages of decomposition in which the original physical structure remains primarily intact. As was stated previously, the polymeric composi- tions of the three different cell walls do not vary sub- stantially. This leads to the fact that the chemical 10 nature of wood is quite uniform throughout the cellular parts. Although uniform in distribution, the chemical makeup of the individual polymers are quite complex. On a dry basis, Poplar SPP is composed of 48.8 percent cellulose (C 19.3 percent lignin 6H1005)’ (C29H36Oll), 29.7 percent hemicellulose (C5H1005)' and 1 percent ash. The remaining 1.2 percent is comprised of extractives such as resins, tannins, and other miscel- laneous compounds. Ash content is 70 percent potassium, calcium, and magnesium. Molecular structures ofcellulose and hemicellulose are illustrated in Figure 3. Madorsky (9) has shown that under slow heating conditions of 5° to 25°C per minute, less thermally stable C-O bonds which abound on the monomer, break with the result of a breakdown of the individual monomer units which yields water, carbon monoxide, carbon dioxide, and char from the polymer. Under rapid heating conditions, above 25°C temperature rise per minute, more selective breaking of the glycosidic bonds connecting the monomer units occurs leading to the formation of levoglucosan, the primary component of tar or smoke. Hemicellulose is chemically similar to cellulose. Instead of glucose units, xylans constitute the backbone of the polymer. Some of the xylose units carry a single side chain of a glucuronic acid residue. This is illustrated in Figure 3b. 11 H2 OH H HZCOH o o o H H H o— H o H H 0 0 H OH 2 H OH P1COH (a) Cellulose H O H -'O OH H H (b) Hemicellulose Figure 3.--Chemical structure of (a) cellulose and (b) hemicellulose. 12 Pyrolysis products are similar to those of cellulose except that no levoglucosan formation is possible due to the more compact chemical structure of the monomer units. Therefore hemicellulose does not contribute to production of volatile products significantly. Instead, less gas evolution occurs with a large portion of the material being converted to char. Lignin (Figure 4) is a polymer composed of highly substituted propylbenzene monomer units. This is the most stable polymer in wood, and is postulated to be the sub- stance which bonds the other polymers together to give wood its strength. Upon decomposition, a wide variety of substituted phenolics are produced. Much of this material volatilizes to tar, but a greater portion contributes to the formation of char. 13 c a H Hé————1) l HCOH HZOOH 00H3 Hc———-o HéOH OCH3 HZC/O\CH H$-—-CH H ca\o/4£H2 00H3 Figure 4.--Chemical Structure of Lignin PYROLYSIS AND GASIFICATION BACKGROUND Pyrolysis Several workers (2-8, 10, 12, 13) have studied pyrolysis of wood and and wood components. Highlights of this work are summarized in Table 1. Brenden (2) discusses briefly the effects of rapid heating (0.5 gram samples of wood subjected to 400° to 800°C) on wood decomposition and found (a) weight loss can be as great as 80 percent, and (b) achieving such weight loss provides gaseous pyrolysis products of high heat content (2312 to 4025 cal/gram). Although no details of the heating periods or rate of temperature increase were presented concisely, this communication supports the fact that high temperature isothermal conditions produce high weight loss and yields products of complicated chemical makeup. Much work has been done on low temperature pyrolysis in the range of 275° to 500°C (3, 4, and 5). Roberts and Glough (3) try to predict the behavior of wood pyrolysis (activation energies and rate constants) over the range 275° to 435°C. Specimens in the study were 2 cm in diameter and 15 cm in length. Experimental procedure involved thermocouple readings at various radii, and 14 5:03 A 303300 A $33 mfiwfimcm 3.39565 Oooom mmHQEmm pouomHu $0020: .950 COD: No .5335ng mo 3mm nom.mm 955m ammobflz can H2 L8 can umsgmw .HE mmaqsoo 8mm: 553m 303300 9.8an 88.500 $033865: J3 H3 Uooooanomm moan—5mm L639: was down 8 Emawu 333m 5.53 63398 3595485953 022.5 9. cm $553 @38th 8.39. 83$ 3QO 53.0.68 8.. 89mm 8322 695m .539: .0..on BS? :3 80m: .380 umfim 50309.0. 38H? bgadfimg noonm .mflmadmcm 352M465 p5... P595 305m 683:8 paucoo new $30390 9:» fig .530 mcofifipcoo $8: umHC 3 Guam Sum moan? cosmic 3am 3.93th 5390.539: $3 ”5303 9*me Ego :0 m .mmogfioolo udxumm tam 9.de 353% mafia mo fig 93 mo Ba cums oomuon 8.8% [mom—50 m ma @003 m0 sag grandma maniacs Huang 335.338 .muficogbo was was man :Bmamm :wad 30:me H “69% E: omN uooomuooov 038 0? oooomuomm 1:2 .803 mo 3388 En imam 13598 .Uooovlomhm um 930m 05096 Honm: .mflmimcw oflbmfiwuono Hag—85mg 5cm: .wmoHsflmofig .wmodpzmo :3: zoom memcwuu mmmE H9535 hm 5 .ucoo 05% Uooowloooa 98310 .36 $33 mo coflogm Wucwflpmnm 030.893 H2835 498 oaHugxrmqu—HQU 335.833 5 H .uwncflxo $03300 conned his: tam goflm mung—mafia: AFB 330.3 .95» H5535 mo :ofloca ma mama Hmduwpfi paw Hmumpfiuflwcwp >munx Mono—=36 EU mud .932on 10.83 .mHQSHmox S H mm 8305.0. “mono umHE .ncoo mwmxaohfl conga :08.“qu $03300 mommoum 92.8003 333 Scum 23895 oommTomR .5882 some 2968 mum 95308... $008 .8323 page any: 5. mflmfiohfl owfiwfixfihmgfiwfi. 0:3 EU ma imam—EH0 ab a £80m Em mange 5383 2329 no .5383 oommmuoovm .8852 =65 m} .82 5 H ma 388nm $940.63 NC coflmgcbo mo umo: 5. mflmmaonhm Unsung—mg? fidugflp 5 {a .wfim mmouwficom 2.me :mpcwum mcoU§8\wucflEnd gouging 8am n5... on? 398m 00.8ng .mcoflmwflmwxfiw mflmwlalohnm Me Submit; mam/B. 16 continuous weight loss measurements. No specific conclu— sions were reached on the physical or chemical nature of pyrolysis. They treated their data according to first order kinetics with respect to weight loss. , t in -fl——¥Ew—= -k J exp (—E/RT) dt (1) w - w o o w' = final weight of specimen wO = initial weight of specimen k = rate constant Using such sample sizes the author showed that simple pyrolysis kinetics as shown in Equation 1 are too simple a treatment, "since values obtained for the activation energy and heat of reaction vary with experimental conditions." For the early stages of pyrolysis the activation energy was found to be 25 kcal/mole and the rate constant equal to 9.1 x 104 minutes-l. Kanury (4) also examined the limitations of studying pyrolysis of large scale specimens. In-situ measurements and X-ray density measurements were adequate for determining pyro- lysis kinetics and a consistent activation energy along the radius of the specimen. Significant observations based on this data are (a) layers of solid lying deeper from the exposed surface exhibit a progressively lower maximum pyrolysis rate and a progressively wider range of temperatures in which pyrolysis occurs, (b) the ratiocflfthe 17 maximum pyrolysis rate to the temperature at which it occurs is roughly constant, (c) the rate curves near the surface and those near the axis are smooth and regular; those at intermediate locations however, exhibit a con— stant rate for a certain duration, the length of which increases with depth beneath the exposed surface and then decreases, and (d) the rate curves fall nearly exponen- tially with time at some radial locations close to the axis. Blackshear and Murty (5) performed weight loss, and surface and internal temperature measurements on 2.54 cm diameter pressed cellulose cylinders. Accordingly, mass and heat transfer for burning and pyrolyzing solids are coupled with a driving force originating from the decomposition of the cellulose within the sample. The driving force is a function of heating history and sample size. They state that a more elaborate treatment of the transient pyrolysis of cellulose material would result in some sort of size dependency. Experiments were conducted on internal time-temperature histories. The process their data suggests is one in which gas is evolved below the surface first endothermically (500°C). The same process occurs near the surface as well, but is augmented by further heat release when vapors from lower regions undergo a secondary pyrolysis, this time an exothermic 18 one, as they pass through the hot outer char layer. Such a picture is in agreement with the results of Roberts and Glough. A significant conclusion in their paper states that internal energy transport (probably in the form of pyrolysis products) will be directly affected by diffusion and convection of mass within the pyrolyzing solids and must be considered when formulating a mathematical description of the process. The chemical interaction of the pyrolysate gases and the char should also be considered. All previous papers (2, 3, 4, 5) point out the significance of heating conditions, and the coupled effects of heat and mass transfer on predicting the nature of wood pyrolysis on a large scale. Intrinsic phenomenoncxfthermal degradation of wood and its components has also been the subject of intense interest, especially to those involved in the studies of flame prevention. Beall (6) observed two major features of the aforesaid temperature phenomenon. One endothermic peak was apparent at 375° to 400°C and a larger exothermic peak from DTA was observed from 400° to 800°C. These were accounted for as the thermal breakdown of the primary material, and chemical interaction of the pyrolysis products, respectively. Aside from a few insignificant aspects, the thermogram of wood is largely a product of the individual components (lignin, hemicellulose, and l9 cellulose). Arseneau (7) examined the thermal breakdown of balsam fir and its components in air from 50° to 420°C. His conclusion states that "the differential thermogram for balsam fir when heated in air is simply a composite of the individual thermograms of the various components of the wood, with little, if any, interaction between the components." Based on this information, data on the pyrolysis of wood components may shed some information on wood pyrolysis. Lipska and Parker (8) analyzed kinetics of cellulose in the range of 250° to 298°C. Different types of rate dependence were found for different time inter- vals: (a) an initial period of rapid decomposition and weight loss, (b) a range in which both the volatilization and decomposition are of zero order, and (c) a region in which the volatilization follows a first order rate, leaving a char deposit which does not undergo further pyrolysis. An additional observation states that the degree of decomposition and volatilization occuring during the zero order phase increases with increasing temperature. An activation energy of 175.8 kJ/mole applies to the decomposition rates in the zero order range. Analysis of the remaining glucosan content shows no readily hydro- lyzable material remaining after 75 percent weight loss. Weight loss data show a consistent value of 15 percent 20 residual material remaining. An important point to note is the elemental compositions of the remaining char heated at 500°C, 98 percent carbon and 2 percent hydrogen. Ele- mental analysis was not done as a function of time, though it might be instructive to determine the rate of oxygen loss as compared to carbon. This information might infer something about the processcflfmolecular breakdown. A majority of the weight loss observed in cellu- lose can be explained by an observation made by Madorsky (9) where 37 percent of a cellulose sample was converted to levoglucosan. Madorsky accounted for this by random scission of 1,4 glycosidic linkages to selectively produce levoglucosan. lOOS)n > n C6H1005 (2) levoglucosan (C6H Other workers have also reported high percentages of levoglucosan yields (Conrad and Chatterjee, 10). Weight loss may actually be controlled by larger morphological influences according to ideas postulated by Golova, Krylona, and Daklady (11). They note such influences as degree of crystallinity, length of polymer constituents, and other macroscopic properties in the structure of the cell wall. ChatterjeeanuiConrad (10) studied the kinetics of cellulose decomposition in the range of 270° to 310°C. 21 They observed a complex reaction rate behavior up to 60 percent conversion between the temperature range of 270° to 350°C. Beyond this a slow weight loss was observed beginning at 360°C and lasting until 450°C. This latter stage was considered the decomposition of the residual material. Two activation energies were found for the initial degradation (levoglucosan formation) of 138 kJ/ mole for adsorbent cotten, and 157 kJ/mole for ball milled cotton. These differences also suggest that morphological structure of the polymeric material may affect the decom- position process. The authors make an important note concerning their investigations as opposed to other studies. In their technique the temperature was gradually raised from ambient to the specified, thereby avoiding the problem of a time lag while a temperature profile forms within the material. The author believes that the thermal lag inherent in isothermal experiments introduces error in the rate data during the initial period of exposure. It is certainly true that divorcing the aspects of kinetics from temperature-time histories will certainly lead to different results, since they are directly related. As for the other components of wood, Beall (12) observed the maximum temperatures in which lignin is thermally stable is 450°C. Major products formed are largely substituted phenolics from the monomer units 22 comprising the polymeric lignin. Stamm (13) observed a lower degradation temperature for hemicellulose of 260°C, and an activation energy of 111.7 kJ/mole in a range of 110° to 220°C. He also stated that hemicellulose degrades four times faster than wood, and lignin only one half as fast as wood. In any case, the decomposition of wood will be somewhat more complex than that of cellulose. Shafizadeh (14) presented a schematic of the general pyrolysis reactions of cellulose. (C6H1005) ———> levoglucosan (3) \4 ' combustible volatiles Also, very rapid high temperature heating is responsible for the following variety of products reported by Shafizadeh. CH CO 0 (4) hemicellulose > CO, H H, CH 2' 3 2 2 lignin —> substituted phenolics, OO 00, CH4, C2H4, C2H6’ HCHO, CH3CHO (5) 2’ As pointed out by Shafizadeh (14) "the difficult task of extending the data obtained on cellulose to wood and plant materials is further complicated by the paucity 23 of information on the pyrolytic reactionsumeOIl.o onsmflm NIW/m ‘MO'H sve 37 material. No microphotographs were taken, but it appeared as though no original fibrous wood structure remained after the initial reaction period for the catalyzed samples. The remaining char is a highly porous and brittle material, which appeared to have had some foaming action occuring during pyrolysis. Some of the sample had become encrusted on the exterior of the sample basket in the reactor. 5. For runs 12 and 13, which were performed at temperatures of 625° and 700°C, respectively, nearly all of the char was consumed. The remaining material was highly porous and brittle. A small amount in water produced a basic solution. It appearedtx>be some form of potassium oxide. Experimental data on simultaneous pyrolysis and gasification of wood is difficult to analyze. Some quan- titative information on product gas composition and gas production flow rates was obtained, however smoke and tar production were not determined quantitatively. The original objective of this study of the gasification of the wood, was redefined to overcome problems related to production of tars and other liquid products which could not be determined quantitatively and plugged the reactor 38 exit several times. Also some products, CO and CO were 2! common to both pyrolysis of raw wood and gasification, thus interpretations<1fresults were complicated. For these reasons gasification of pyrolyzed wood char was investigated. EXPERIMENTAL CONDI TI ONS Introduction Char samples were prepared from dried samples of Poplar SPP. Char samples were impregnated with Na2C03 and K2C03 and gasified at various temperatures, steam partial pressures, and space velocities. The following range of experimental conditions were used to study the kinetics of pyrolysis and gasification of char: (a) steam partial pressure of 45 to 100 kPa, (b) space velocity range of 2 to 7.25 sec—l, and (c) temperature range of 500° to 700°C. These experiments were conducted in a differential fixed bed reactor. Small sample size of l to 2 mm in diameter was chosen to maintain temperature and concentration gradients to a minimum. High gas velocities of 10 to 35 cm per second were chosen to mini- mize reaction of product gases with the reacting char due to the short contact times involved. By observing gas flow, gas composition, temperature, and total sample weight loss, evaluation of the kinetic parameters for the system was accomplished. Additional information was also observed on the effects of 12 percent by weight of Na CO 2 3 and K2CO3 on the gasification rates. 39 40 Reactant Gas Production Figure 7 is a schematic diagram of the reactor system. A large water reservoir (A) is suspended six feet over the reactor assembly. A needle valve provides a constant supply of feed water between 0 and 12 cm3 per minute and is mixed with a set flow of nitrogen. The nitrogen is controlled by a tank regulator, valve, and flow meter arrangement (C). Combined nitrogen and water stream is then fed to a vaporizing furnace (E), where water is vaporized and the gases are heated to high temper- atures before entering the reactor. The water and nitrogen stream is heated in the gas fired furnace by flowing through a stainless steel cylinder (D) 130 mm long and 51 mm in diameter, with combustion gases impinging on the outside. The hot gases flow out of a vaporizing heater into a 6.35 mm O.D. heavily insulated stainless steel tube that delivers the hot gas mixture to the reactor (H). Temperature control of the reactant gas is accomplished by adjustment of the furnace flame and the flow rates of nitrogen and water. The temperature of the reactor system was allowed to stabilize for approximately 10 to 15 minutes before making a run. Gas temperatures between 400° to 800°C were obtained for steam partial pressures between 45 and 100 kPa and reactor space velo— cities of 2 to 7.25 sec-l. 41 .msumummmm Hmucmefluwmxw mo oamemcomll.h madman 80.86338 apatuoz mam £88 “mama umwfi uwz Sham mfiascmm m8 Hmmmm> coauooaaoo mummcwpcoo 9580 wfiuwflo fiofiwmmm H93 €3ch wwwcwucoo uwHHouucoO musumuwghwa oonosz mwfimsoootswne Hmesamo: uonommm B n35 wcflq mmo amusumz mafia HH< momcusm uwumocoum Hmummnwum mmw Hmumeonh 0cm m>~m> cmmouuaz “884.35 :68sz HHo>ummmm mammsm uwumz w u Amv Ev 8v EV 8V 2 9: Si O: E A: :3 g E: :5 Ev CV 3: A5 42 Reactor Assembly The reactor assembly, Figuresiiand 9, is con- structed entirely of type 304 stainless steel. The reactor tube (H) is 336.5 mm in length and 44.5 mm O.D., and 38.1 mm I.D. is placed within the heavy-duty electric furnace (I) such that the bottom 100 mm of the tube are heated by means of the furnace heating elements (J). Furnace capacity ranges up to 560 watts. Temperature is monitored by means of three chromel- alumel thermocouples (K). One thermocouple is centered in the reactor cavity 75 mm above the heating zone of the reactor to monitor the incoming gas temperatures. This thermocouple is removable to facilitate addition and removal of the sample. The second thermocouple is located at the bottom of the sample compartment (i.e., the heating zone) to provide data on the sample and exit gas temper- atures. Both thermocouples are mounted by welded 6.35 mm O.D. stainless tubing and Swaglok fittings. A third thermocouple is situated near the heating elements of the electric furnace to provide a temperature signal for a Wheelco temperature controller (L). The top of the reactor is removable by a spring arm arrangement to provide rapid manual sample addition and withdrawal. Asbestos is used as a seal between the top of the reactor and the reactor cap. 43 .manfiwwmm Houommu mo mHSDOflmII.m wnomflm .kumwnmum Emmum Cam ondu H0pomou :HmE mcfl30£m mfinfiwmmm Houommu mo whopoflmnl.w musmflm 44 A 100 mm long by 36 mm O.D. stainless steel screen basket fits snugly into the reactor to serve as a sample holder. This is raised and lowered into place to initiate and terminate the run, respectively. Product Gas Processing Product gases exit the reactor and flow into a water immersed copper coil (M) which cools the gases to approximately 20°C, and condenses any nonvolatile products and water vapor in the product stream. Liquid water separation is accomplished in a small vessel of 50 cm3 in volume (N). The liquid level was maintained approximately half full. The liquid was continuously drained into a beaker at a rate which would maintain a constant level in the separator. The total volumetric flow of gas ranged between 160 and 500 cm3 per minute, thus the gas residence time in the separator was less than 0.1 seconds. The effect of gas mixing in the separator was neglected. After leaving the reservoir the gases flow into a Drierite column (0) which adsorbs any remaining water. Gas sampling is done following the last water removal step (Q). Gas sampling was accomplished using 2 cm3 capacity Pressure-Lok gas syringes equipped with side port needles and valves. Based on calculations using the calibration data, accuracy with 0.5 cm3 samples was within 5 volume 45 percent. After gas sampling product gases then pass con- tinuously through a Precision wet test meter (R) which is used to measure the flow rate and total gas production with time. Additional equipment required for the experimental apparatus included a Wheelco pyrometer (L) for furnace temperature control, Leeds and Northrup potentiometer (S) for monitoring the reactor temperature via chromel-alumel thermocouples, and a Perkin-Elmer model 154 Vapor Fracto— meter with accompanying Photovolt 43 strip chart recorder. Gas Analysis Syringes were used to take 2.0 cm3 aliquots of gas periodically during a run, and the gas samples were later analyzed with a Perkin-Elmer model 154 Vapor Fracto- meter. In order to flush the needle and valve assembly, 1.5 cm3 of the sample was purged into the atmosphere, and the remaining 0.5 cm3 was injected into the chromatograph column. This procedure was utilized to reduce possible contamination due to the other gases in the needle. The packing material used in the column is carbosieve S 100/ 120 mesh. Glass column dimensions are, 1830 mm in length by 6.35 mm O.D. with an I.D. of 3.175 mm (6 ft. x .25 in O.D. x 0.125 in I.D.). A thermal conductivity detector provides adequate response for all the gases analyzed, 46 provided neon is used as a carrier gas. It is especially important to note that using helium as a carrier gas does not allow sufficient sensitivity to hydrogen. A column temperature of 136°C and 15 cm3 per minute of neon flow provided an adequate retention time for the required peak separation. Calibration was achieved by separate analysis of pure 0.5 cm3 samples of each of the component gases. Response of hydrogen was assumed linear over the concen- tration range observed. Sample Preparation For pyrolysis experiments Poplar SPP was ground in a hammer mill then oven dried and stored in a dessi- cator. For the salt impregnated experiments the wood CO was pretreated with 12 percent Na and K CO3 solutions, 2 3 2 then oven dried. Average particle size was 1 mm in diameter and 2 to 4 mm in length. No change in size was obtained after salt pretreatment. Sample sizes for the pyrolysis experiments ranged between 5 and 7 grams. For all the gasification experiments the finely divided wood was pyrolyzed at 700°C in flowing nitrogen for 10 minutes. The pyrolysis time of 10 minutes was based on previous observations that no additional weight loss of wood was observed after 5 minutes at those condi- tions. Char gasification sample sizes were 2.67 grams 47 ($0.03 grams), except in the case of the salt impregnated char, for which the total carbon content of the sample was approximately 2.67 grams. Total sample weight for these runs are higher due to the added weight of the alkali salt. All samples are thoroughly dry before use. OperatinggProcedure The char addition which initiates the run was begun after the system had been maintained at the desired operating temperature and flow conditions. Correct operating conditions were achieved when temperature readings from thermocouples above and below the sample chamber were equal, and constant water and nitrogen flow rates were observed. At this time the wet test meter was zeroed and the tOp of the reactor removed. The sample was quickly inserted into the reactor and the reactor cap and condensate vessel were simultaneously put in place. The beginning of the reaction sequence required a total time of three seconds. Conditions were maintained con— stant over the 5 to 10 minute period of the run, during which three to five gas samples were taken usually three minutes apart. The temperature of the reaction environ- ment was monitored during the reaction as was the gas flow rate with time. Since analysis time for each gas sample is 12 minutes, part of the gas analysis was completed 48 after the run had ended. All gas samples were analyzed within one hour. EXPERIMENTAL RESULTS Effects of Pyrolysis A microphotograph of Poplar SPP is shown in Figure 10. From this photograph it can be seen that the struc- ture is quite regular and that there are essentially two sizes of pores, the vessels were measured to be 70 pm in diameter while smaller tracheids were measured to be 20 pm in diameter. The cellular structure has cell walls of 2 to 4 pm in thickness. Details of the aforementioned cellular structure as described in Figures 1 and 2 are apparent, particularly the large 82 layer which comprises a majority of the cell wall. The highly porous nature of the wood is readily apparent. A series of electromicrophotographs were taken of wood after having been exposed to a high temperature gasi- fication environment. Three millimeter diameter samples were pyrolyzed in 600°C steam for periods of l, 4, and 8 minutes, and electromicrophotographs of these samples are shown in Figure 11, 12, and 13, respectively. Photographs of the material after just one minute of exposure shows two striking morphological changes. It can be seen that there is a slight reduction in size of the tracheids and 49 50 I Figure ll.—-Electromicrophotograph of Poplar SPP after 1 minute of pyrolysis. 51 Figure 12.——Electromicrophotograph of Poplar SPP after 4 minutes of pyrolysis. Figure l3.--Electromicrophotograph of Poplar SPP after 8 minutes of pyrolysis. 52 enlargement of the vessels. Secondly, a reduction occurs in the thickness of the individual cell walls. After one minute of gasification, the cell walls are reduced to 0.5 to 0.8 pm in thickness, corresponding to an 80 percent loss in size. Weight loss measurements for the samples indicate 85 percent weight loss after 8 minutes of gasi- fication. Therefore the initial density of the cell walls of the unpyrolyzed wood of 1.5 grams per cm3 remains essentially constant during pyrolysis while the apparent bulk density of the char particles becomes 0.06 grams per cm3. The cell walls of the residual material appear to be uniform in structure, with fragile cellular components remaining quite intact. There is no apparent fissuring or cracking of the cell walls except in the areas adjacent to the vessels, and they appear to remain quite impermeable, despite the fact that 85 percent weight loss has occured. Photographs of the samples exposed to the same high tem— perature for periods of 4 and 8 minutes appear very similar to the photographs after one minute of gasifi— cation. These observations indicate that during the early periods of gasification and pyrolysis of wood, the fine structure of the cell walls is decomposed where as the macropore structure of the solid residue is essentially the same as the macropore structure of the original dried wood before reaction. 53 Salt Catalyzed WOod'Char Gasification The experimental work was done jointly with Mr. Craig Anderson. Experiments were done on both salt impregnated and nonimpregnated wood char samples. Results of the nonimpregnated experiments are reported in detail in Mr. Anderson's thesis. This thesis reports details of gasification experiments for salt impregnated wood char and comparisons are made with results of experiments on nonimpregnated samples. Reaction conditions and techniques were duplicated for the catalytic and noncatalytic runs. Table 3 contains a summary of experimental conditions and results. The catalytic activities of two different salts, Na CO 2 3 chosen since they have been shown (20) to be effective in and K2C03 were investigated. These salts were reducing tar formation in pyrolysis and the enhancement of gasification. Details of the preparation of the salt impregnated samples are presented in the experimental section. Samples contain approximately 12 percent by weight. Ten experiments were conducted for partial pres- sures of steam between 45 and 100 kPa for each catalyst, temperatures of the reaction environment of 550°, 625°, and 700°C for each catalyst, and all at a space velocity of 4 5'1. 54 Gasification reaction rates were calculated for each run basedcxitwo different assumptions. In case A, it is assumed that methane is produced by direct hydro- genation of char, thus the reaction rate of char gasified is proportional to the rate of production of carbon oxides, whereas in case B it is assumed that methane is produced by reaction of carbon monoxide and hydrogen, thus the rate of char gasified is proportional to the sum of the rates of carbon oxides and methane. Therefore reaction rates were calculated for each run based on (a) total production of CO and CO in the effluent, and (b) total 2 2, and CH4. See Appendix B for sample calculation of reaction rate. production of CO, CO Carbon monoxide is the direct result of gasifi- cation. C (s) + H20 (9) > co (9) + H2 (9) (7) But due to high temperature and high concentration of steam in the system, the shift reaction also takes place. CO (9) + H20 (9) > C02 (9) + H2 (9) (8) Previous workers have usually considered the shift reaction to be fast and thus always in equilibrium. Due to stoichi- ometry the total moles of carbon oxides in the effluent 55 stream will not change, and will be equal to the total moles of CO formed from the gasification reaction. Three other reactions are important at the experimental condi- tions. c (s) + €02 (g) -——> 2 co (9) (9) c (s) + 2 H2 (g) —> CH4 (9) (10) CO (g) + 3 H2 (g) —> CH4 (g) + H20 (9) (11) Reaction 9 is the Boudouard reaction and occurs at a very slow rate at experimental temperatures. Reaction 10 is favored by equilibrium at low temperatures. The other possibility of methane formation within the system is reaction 11. If reaction 11 is the source of methane in the product gas, then the total moles of methane should be included in the determination of the reaction rate. Due to stoichiometry, the total moles of methane formed by reaction 11 is equal to the moles of carbon monoxide consumed. Depending upon the source of methane, the rate may be calculated in two different manners. Assuming reaction 11 is occurring, then the production of methane must be included in calculating the rate of gasification. If the source of methane is carbon, perhaps by catalyzed direct 56 hydrogenation, then methane must be neglected in calcu- lating the rate. Both cases are considered in the fol— lowing steam partial pressure experiments. To evaluate the affect on steam concentration of the reaction rate, the space velocity was maintained at 4 s-1 and reaction temperature was held constant at 625°C, while steam partial pressure was varied from 45 to 100 kPa. Figure 14 is a plot of experimental results assuming methane to be produced by direct hydrogenation of char. Due to scatter of the experimental points it is difficult to determinetflmareaction order with respect to steam with much precision. As can be seen in Figure 14 a first order in steam rate equation is appropriate for both uncatalyzed and catalyzed material. Figure 14 is a plot of experimental results assuming methane to be produced from carbon monoxide and hydrogen reacting. Comparison of Figures 14 and 15 indicates for Na2C03 catalyzed runs there appears to be no appreciable difference in methods for treating the data, since methane production was quite small. K2C03 catalyzed char exhibits higher reac- tion rates since the methane product gas contributions were high for these runs, but the data still appears roughly first order. It is apparent that first order kinetics for the gasification reaction are consistent with both methods for treating the data for both salts. 57 .4 mmmu .NOO pam 00 Eoum cmumasoamo mumu cofluommu :0 whammmum Hmflpumm Emwum mo uommwMIl.vH musmflm . been... mmwammmma .253 243m 0. mo *0 who II 1 d I d <1 N 901x 31w E] r‘o .o 3 MW ‘0 was IT—Z-H o"'w"-—"a 9 85.3.20 noo~xo . 85538 .862 o o ENS/4:82: 4 . m 58 . .m ammo .amo mam .Noo oo Eoum Umumasoamo mama COHuommH co madmmwum Hmfluumm Emwpm mo pomMMMII.mH musmam ..o_§n_x.mm:mmmmn_ ESE/a 23.5 0.. mo mo No mo mo v.0 no «I 4 J u . . H V I.- .N 3 wk: 0 1m Aw. OW .fiV WW W" W am D ‘ . w . . . Ln 0354.20 00 oz 0 m 8qu20 ”8.310 .m 59 The following equation is suggested for the rate of gasi- fication. moles reacted (12) gr C, minute Gasification rate = k(T) pH20 Activity of char is assumed to be unity, therefore it does not appear in the rate expression. These results are in contrast to the results for the noncatalyzed experiments, where it appears that Langmuir-Hinshelwood kinetics fit the data well for case B methodology. Assuming a typical Arrhenius dependence on the rate constant, activation energies were calculated. This was accomplished by determining the reaction rate versus the inverse of the absolute temperature for 550°, 625°, and 700°C, see Figure 16. The resulting activation energy is 94.2 for Na CO CO 2 3 2 3 kJ/mole. Uncatalyzed char had an observed activation is 84.9 kJ/mole, while that of K energy of 157 kJ/mole. Assuming a first order reaction rate and an Arrhenius temperature dependence, frequency factors were calculated and then averaged. Average values were then used for calculated reaction rates. Averaged frequency factors found were 10.7 and 16.4 moles/gr C minute kPa for Na CO and KZCO 2 3 3' data are summarized in Table 3 in the Appendix. respectively. All rate 60 .B\H msmuo> mpmu ca 2.x mo. x H) ON. 0... Il‘ d 4 w ENS/EB noouoz 0 855420 .00.; a 85.2202: 4 O... .pOHQ mmuwcm coflum>fluo mommm mo noomwmll.na Guzman :on E03“; Noqam m m \u m m 4 m N u N OIx Elva r0 4/ . 1 NIW‘O wvas . o mow-39 2 1 ID 63 slower gas velocities. This is the most plausible explan- ation for the observed experimental behavior. MODELING Single Particle Effectiveness Factor The observed intrinsic kinetic information has been combined with estimated diffusion coefficients to model a reacting particle. Figure 18 is a schematic diagram of the reacting particle. A flat slab geometry is justified by observations of pyrolyzed wood that indi- cate the original structure of parallel closely packed cells is retained in the char. The high impermeability of the cell walls is the justification for assuming one dimensional transport within the particle. Effective diffusivity within the particle has been calculated based on the Random Pore Model (25), Equation 13. 01 = Tl + 51— (13) e AB K where De = Effective diffusivity DAB = Bulk diffu51v1ty DK = Knudsen diffusivity The ratio of the pore diameter to the mean free path of steam at 700°C is over 100. This is a sufficient 64 65 X X+AX Figure l8.-—Drawing of flat slab geometry for a particle with external bulk film resistance. 66 condition for neglecting the contribution of Knudsen Diffusion. DAB was calculated from Equation 14. 2 DAB = DAB 8 (14) where DAB = Effective bulk diffusivity DAB = Bulk difquiVity e Void fraction A void fraction of 0.7 was calculated for the residual char. A bulk diffusion coeffecient for steam in the reacting mixture can be calculated using Equation 15. n D . = (l - Y )/ 2 (Y /D. ) (15) H20, mixture H20 j=l j 3, H20 where Y = mole fraction of steam H20 Y. = mole fraction of the other species 3 in the reacting mixture Experimental binary diffusion coefficients for Equation 15 were extrapolated to higher temperatures using the Chapman Enskog formula. Using first order kinetics and a mass balance on the geometry in Figure 18, we have De ———-= kC (16) 67 Equation 16 in dimensionless variables becomes, 3% = %f5 w = ¢w (17) e 2 where ¢2 = %:§ The boundary condition at the center of the particle is 31P_ _ fi—O,€—O (18) while the boundary condition at the surface becomes 8w _ . _ _ 5:.— Elm (1 w). a — 1 <19) for mass transfer through a stagnant boundary layer. Solving the differention Equation 17 with the boundary conditions 18 and 19, yields w _ cosh ¢ (20) $_§££Q_$ + cosh ¢ Blm Effectiveness factor is defined as follows, _ rate observed rate (co) (21) where CO = bulk concentration 68 and for this problem (see Appendix C), tanh ¢ = 22 ” ¢(l+¢tanh¢) () Bi m therefore Global rate = Intrinsic rate at bulk conditions taghl ¢ ] 4) (23) d) (l + T) m The influence of particle size and finite external mass transfer rate on the isothermal effectiveness factor for the catalyzed reaction rate is shown in Figure 19. A value of k = 1.519 5.1 (Na2C03 catalyzed gasification reaction rate at 625°C) was used for the isothermal effectiveness factor calculations. It can be seen that mass transfer becomes a significant factor for particles larger than 1.0 cm. This substantiates the fact that intrinsic kinetics were observed in the study, since experimental particle size was less than 0.3 cm. Figure 20 illustrates a comparison of effectiveness factor for catalyzed and uncatalyzed char at 625°C over a range of particle sizes. This result shows that diffusional resistances are more significant at higher reaction rates. 69 m 9 s H .m\NEo ma.m m mam.a u x .wNHm maoflpuma m5mu®> Houomm mmmcm>fluommmw mo uonll.mH musmflm so .NNa Noemi on 0m QN m... 0.. no . (l . Hi|1|||||||1|||||||4 [ll _.0n_m nu. #~ 00. .. 0.0 0.. 70 m .m\NEo afim n 988:0 Sim. 4E: Umwhamumocs cam AHIm mam.a n xv pmmmamumo How Hepomm mmmcw>fluommmw mo COmflHmmEounl.om musmflm so .NNa 39:15 On mm 0N m._ 0.. no . N.O QwN>u_<._.<0 .. v.0 mOONoZ C . 0.0 DUNrJ/EKOZD . 0.0 .O._ 71 Tubular Reactor Axial Dispersion Model Considering an axial dispersion model for a tubular flow reactor, containing wood particles a mass balance for first order kinetics over a differential element as shown in Figure 21 yields D ——§-- V 52 - nkC = O. (24) -9 _£ f_C'S-L 0 Equation 24 becomes, 3.9.33-1£_Df=0 (25) P 2 as a ’ e as where -21. = E Pe _ D ’ Da nk v Two boundaries conditions are required to solve Equation 25 analytically. lst B.C. g: = Pe (f — 1), s = o S 2ndB.c.§—f-=o ,s=1 c) U) 72 .pcmEmHm Hmflucmummwflp WCNBOQm Honommu 30am msHm mo mcfl3muonl.am musmflm O u Pd I-HN N<+N \N 0 1EN 1 00 73 The solution of 25 is Pe /’ 2 f (l) = 7r ( Pe + 4 PeDa) exp (Pe) (26) where A = Ri exp (R+) - RE exp (R_) (27) Ri = Pe i /Pe2+ 4 PeDa (28) and l - f (l) = X (conversion) (29) Figures 22, 23, and 24 show the effect of a range of Peclet numbers from 0 to 14 on the conversion of steam in a tubular reactor for char particle sizes of 0.5, 1.0, and 2.0 cm. Biot mass numbers from 0.1 to 100 were used as a parameter to evaluate the effect of external mass transfer. The following values were used: k = 1.519 s-l, De = 3.19 cmZ/s, L = 40 cm, and v = 40 cm/s. Comparison of Figures 25, shows the effect of increasing reaction rate on steam conversion. The 625°C K2CO3 catalyzed rate constant of 4.464 s—1 was used in Figure 25, with all other parameters the same. Conversion of steam is improved from 0.55 in the NaZCO3 catalyzed case to 0.75 in the K2CO3 catalyzed case for a well mixed reactor 74 Q WED m.o M 0 .EU 04 u > .60 ov M Q .Hum mHm.H n x .omm\mfio oa.m u o .uones: om msmuw> x I H m0 uonuu.mm wusmsm an. 4. N. 0. m 0 4 N L N0 H 00. . . _ _.o .5 / 6.0 X . m0 . 0.. 75 . EU . . I . M ED . l 0 Hum mam H u x om \N NH m 1 a Q o.H n o .omm\Eo ow n > .50 ow n A .HOQEC Hmflomm WDWHQ> X I H 00. ..0n.m No uonun.mN wnsmam 0.0 0.. 76 Q 0.80 o.m n 0 .omm\Eo 04 n > .80 04 u A .Hlm mam.a n x .omm\NE0 ma.m u 0 .umnfisq uoaoom m5mH®> x u H mo ponll.vm musmflm ma 4. N. 0. m 0 v N . N0 00. fwd 0. fl _ . X . 0 0 _ 1 6.0 ..0 Am .0. 77 m 0.60 0.H u 0 .omm\Eo ow u > .50 ow n H .Hlm 404.4 N x .omm\NEo mH.m u 0 .H0355: umHoom msmum> x I H mo HOHmII.mm wusmHm on. N. O. m m .4 N 0. l . No - / . 4.0 ..| _ . VA .00 ..0u.m .00 .0. 78 and no external mass transfer resistance. Figure 26 shows the effect of residence time on the conversion of steam. Figure 27 demonstrates conversion as a function of effec— tive diffusivity for a 1.0 cm particle and a rate of 1.519 s-l. A limiting value of approximately 0.70 con— version is reached for values of effective diffusivity above 3.0 cmZ/s. Results of this modeling show that there is little sensitivity on conversion by axial mixing, since conver- sion is independent of Peclet number above a value of 2 for nearly all the plots. 79 Q m .80 o.H n 0 .80 ow n H .H:m mHm.H u x .omm\NEo mH.m n 0 .conum>:oo co Emmum mo 08H» mocomemu mo Hommmmln.0m ousmHm mozoowm m2... mozmofimm m N . W J a 80 m . .Eo o.H n 0 .owm\Eo 04 n > .80 ow H Hum mHm.H u x .COHmHm>soo so muH>Hmsme0 m>Huomwmw mo pommwmnu.>m musmHm 0wm\ N2 0 .60 m N d d a N 0 0.. SUMMARY AND CONCLUSIONS Conjectures of Catalytic Activity The influence of alkali carbonate salts on the pyrolysis of cellulosic materials and wood has been studied by several investigators (l7, 19, 20). The observed effects during rapid pyrolysis of wood, of less tar production and greater char production, have also been observed in this study. It appears evident from results of this study that the salts actively degrade the chemical nature of the polymer constituents. The residual char product of the salt catalyzed wood does not have the structure of the original material. This is in definite contrast to the pyrolysis char of uncatalyzed wood which retains its shape, size, and microstructure throughout pyrolysis. The catalyzed char product looks as though a foaming action occurred throughout the entire sample. Some of the char actually becomes encrusted on the outside of the sample basket. This degradation of the wood may occur during the pretreatment step, while soaking in the salt solution, or while drying at 100°C for 24 to 48 hours. Or the degradation may occur just prior to the pyrolysis period as the material is undergoing a rapid 81 82 temperature rise. Sufficient chemical or physical degra- dation occurs such that the material flows. Foaming of the material may be the result of volatile alkali com— pounds occuring at the experimental temperatures above 600°C. The increase in the gasification rate for char impregnated with salt as opposed to untreated char may possibly result from two different effects. The graphitic carbon structure of char provides active sites to the reacting steam molecules. The actual enhancement might possibly be the result of increasing the rate of water adsorption at the active sites, or increasing the number of active centers for the gasification reaction to occur. Increased rate of water adsorbtion may be the result of a changed charge distribution on the surface. The number of active sites would depend on how the cation chemically reacts with the surface. The alkali metal is most likely responsible for the rate increase, since the carbonate would be thermally unstable above 500°C. Observations and Comments The initial objective of this research was to study the catalytic gasification of Poplar SPP. Prelim- inary gasification experiments of dried wood over the temperature range of 400° to 700°C produced gaseous, 83 liquids, tars and char products. Experimental difficul- ties were encountered related to quantitative analysis of liquid and tar products and deposition of tars on various equipment components. Therefore the objectives of this study were revised to study the intrinsic gasification kinetics of catalyzed Poplar SPP char prepared from pyro— lysis of raw wood in 700°C nitrogen for a period of 10 minutes. The catalytic activity of two different salts was investigated. Experiments were conducted for partial pressures of steam between 45 and 100 kPa for each cata— lyst, and temperatures of the reaction environment of 550°, 625°, and 700°C for each catalyst. Experiments were also conducted on uncatalyzed material to determine the effect of external mass transfer by observing space velocities of 2, 4, and 7.25 5.1. Product gases were C0, C02, and CH4, and H2, which were analyzed by gas chromatography. Based on results of the space velocity experiments it was concluded that external mass transfer is not controlling for the experimental conditions of 4 s-1 space velocity and l to 2 mm diameter particles. Reaction rates were calculated in two manners. For case A, it was assumed that methane was produced by direct hydrogenation of char, thus the gasification rate was calculated based on the total production of carbon oxides. In case B, it 84 was assumed that methane was produced by reaction of carbon monoxide and hydrogen, therefore the reaction rate is proportional to the sum of the carbon oxides and methane. Both reaction rate calculations indicate the following rate law. moles 0 gr C, min (12) Gasification rate = k (T) pH2 These results are in contrast to the results of the noncatalyzed experiments, where it appears that Langmuir- Hinshelwood kinetics may apply. Activation energies of 84.9 kJ/mole and 94.2 kJ/ mole were obtained for NaZCO3 and K2CO3 catalyzed gasifi- cation reactions, respectively. Uncatalyzed char had an activation energy of 156 kJ/mole. Averaged frequency factors found were 10.7 and 16.4 moles/gr C, min, kPa for NaZCO3 and KZCO3 catalyzed gasification reactions, respectively. Using the observed intrinsic kinetics and an estbmated effective diffusivity of steam, a particle model was developed to calculate an effectiveness factor versus particle size. The results of the calculation show the internal mass transfer becomes significant for particles larger than 0.5 cm when no external mass transfer limi- tations are included. The effect of external mass transfer 85 can be significant for any particle over 0.2 cm in width. The results of the effectiveness factor calculation were then used in an axial dispersed plug flow reactor model to estimate conversion of steam. Using estimated diffu- sion characteristics, a range of Peclet numbers were analyzed for particle sizes of 0.5, 1.0, and 2.0 cm in size. The results of this modeling show that there is generally little sensitivity of the system to mixing conditions within the reactor, since conversion is inde— pendent for values of the Peclet number larger than two for all particle sizes. APPENDICES 86 APPENDIX A FUTURE WORK 87 FUTURE WORK 1. A series of microphotographs of salt impreg- nated char, prior to and during pyrolysis or gasification, to examine possible structural damage to the cell walls. 2. Examination of the catalytic activity of a series of salts formed from the same alkali metal. This might give some insight on the specific chemical nature of the char catalyst interaction. 3. The affect of a range of concentrations of one salt on catalytic activity. This might provide data on the number of active sites generated on the carbon surface by the catalyst. 4. Quantitative examination of initial wood and the residual material, particularly for ash content. 5. Test a greater variety of metal compounds for possible catalytic activity on the gasification reaction. 6. Examination of the gasification reaction rate behavior for known concentrations of H2 and C02, to analyze possible inhibition of the rate by these gases. To do this safely a significant change in the experimental apparatus would be required. 88 89 7. A theoretical treatment of pyrolysis reactions. This could be accomplished by lumping populations of functional groups and bonds within the material to develop a series of hypothetical reactions. 8. Extension of the experimental apparatus to enable the quantitative analysis of liquid pyrolysis products. 9. Experimental study of the gasification reac- tion rate kinetics of gaseous and liquid pyrolysis products derived from wood. 10. Physical characterization study of the char residue to determine accurately the pore distribution, porosity, and diffusional properties. 11. More elaborate experiments with the existing experimental apparatus, to determine the effect of space velocity on mass and heat transfer within the reaction chamber. 12. Extension of the modeling to a nonisothermal treatment, particularly for pyrolysis modeling. 13. Extension of experiments to larger particles to evaluate the previous modeling studies. APPENDIX B SAMPLE RATE CALCULATIONS 9O SAMPLE RATE CALCULATIONS The following presents sample calculations for the experimental run with a temperature of 550°C, steam partial pressure of 100 kPa, and a space velocity of 4 5.1. Calculation of product gas composition.* Component Calibration factor, fi Peak weight, wi H2 2.41 0.07475 N2 1.00 0.18703 CO 0.96 0.00600 CH4 0.82 0.00726 C02 1.53 0.02826 *The product gas compositions were averaged over all of the gas samples taken. Only the data for one gas sample is presented above. mol fractioni = i i 91 92 M01 fraction on Component Mol fraction nitrogen free basis H2 .243 .618 N2 .607 -- CO .019 .051 CH4 .039 .099 C02 .092 .234 Calculation of material balance: Average nitrogen flow 0.105 fi/min at 25°C, 1 atm Total measured gas flow 2.78 2 at 25°C, 1 atm - Total nitrogen flow 1.58 2 at 25°C, 1 atm fhmv Total gas flow - N’ 1.20 1 at 25°C, 1 atm 2 Total gas - N2 (1) l 1.0 (mol) l 12 (g) | 24.451 (4) I 1 (mol) Total carbon in product gas x mol fraction of (CO2 + C0 + CH4) 1.20 (1.0) (12) (.238 + .060 + .102) 24.451 0.24gnams tkkalcmmbmucomamed==(L24gnams 93 = Total carbon in product gas - Total carbon consumed % error Total carbon consumed _ 0.24 - 0.24 0.24 0.0% Calculation of gasification rate: rate = Average product gas flow rate (IL/min) — Average nitrgen flow rate (ll/min) Original carbon present (9) x 1.0 1101 I mol fraction of (C0 + (I32) 24.451 (2) | = 3.652 x 10'4 39-1— gmin APPENDIX C EFFECTIVENESS FACTOR CALCULATION 94 EFFECTIVENESS FACTOR CALCULATION Given, k C dx and w = cosh ¢£ 9—§%%§—i + cosh ¢ m Since rate = De ——5 then, 8x L L 2 L rate dx = D 2—9 dx = D 39 e 2 e 3x 0 0 3x 0 making this dimensionless, l l D C C D eL O %% rate dx = 0L e (g: - 3g ) o o l 0 since L rate dx = C0 De fii L 3g 1' 95 96 then n: ”e 32:12. sz 35 1 0 fig _ ¢ sinh ¢ 8E — ¢ sinh ¢ 1 ———EEf——-+ cosh ¢ m then substituting, n = ;L_ o sinh ¢ = tanh ¢ (1)2 ———-————¢ Sggh¢+cosh¢ ¢(1+——————¢ E?“ ‘1’) m m APPENDIX D SUMMARY OF RATE DATA 97 98 mH NH.O HO.N O.Om ONO O OO.m OO.m m.OO ONO m H~.m OO.m ~.OO ONO O OO.OH OO.OH OOH OOO N. mv.m mO.m N.OO mmO O. OO.H OO.H OOH Omm moomm O.O OO.N OO.N O.HO ONO O.m O0.0 Hm.O O.OO ONO O.O mm.O OO.m N.OO ONO Hm mm.OH OO.O OOH mOO Om OO.O OO.m N.OO ONO O Om.H HO.H OOH Omm moommz g “Baum SHE Ego—u m an: 9308 0 mg 00 mmmpcmoumm OH x mama OH x N m m CHI mfiapmfiXEmB MMDMdRJwUM%WUME< meangameo mHSmmmHm Hwfluumm .wumo mumm mo mnmEESmII.m mamda ma 904‘ mhjv m.mm mmm 99 H O0.0 ON.O 0.00 ONO ON ON.O O0.0 N.OO ONO O ON.OH HN.OH OOH OON O. N0.0 O0.0 N.OO ONO O NO.H HO.H OOH OOO mooNO ON OO.N O0.0 O.HO ONO NH N0.0 ON.O 0.0N ONO NH O0.0 NH.O N.OO ONO ON N0.0H O0.0 OOH OOO ON N0.0 NH.O N.OO ONO O OO.H OO.H OOH OOO Ooonz "aw “Bum CHE 3308 m 5.8 939: m mmx Uo mmmucmoumm OH x wumm OH x N m m OOO wnnmawéme oopfldoamu mmmum>¢ 8mm Motowno mHommmHnH ngm . GmDGHuGOUII . m mam/Ha. 100 Ow Om.m 5min mNK O mm .m mm .m O max OOH u whsmmmum HmHuthm 8899 N O .O OO .O N AooONO u mange. £9588 EUQE mUmNHm HQHHm OHHE m\m.HoE m :HE m\mHoE m m m 3688 OOH x Bum OOH x 728% CREHSOHMU mmwuwéa mumm EEmQO huHUOHm> mommm O.NH HN.OH O0.0 OOH mOO m.O MHw. mOm. OOH omm OH 22H OO.H OOO ONO HH Om.m NH.N 9mm mNO O om.m 3.6 0.00 mNO OH mn.m brim 0.00 mmO ”mzam QmNMHANBdUZD HDHHm CHE m\®HoE m 5H5 939: m mmx Do $3888 OH x 3mm O x N m m H o m whopmuwgme BOEHHHUHMU mmmumée 8mm 8588 mammwhnm“ HMHMTHMAH . @mSCHpcOUII . m mqmflfi BIBLIOGRAPHY 101 10. BIBLIOGRAPHY Anonymous, "Biomass Potential in 2000 put at 7 quads," Chem. Eng. News, February 12, 1979, pp. 20-22. Brenden, J.J., "Caloric Values of the Volatile Pyro- lysis Products of Wood," Combustion and Flame, ll, 4, pp. 437—439, 1967. Roberts, A.F. and Glough G., "Thermal Decomposition of Wood in an Inert Atmostphere," Ninth Symposium (International) on Combustion, pp. 158-166, 1963. Kanury, A.M., "Thermal Decomposition Kinetics of Wood Pyrolysis," Combustion and Flame, 18, pp. 75-83, 1972. Blackshear, Jr., P.L. and Murty, K.A. "Heat and Mass Transfer to, from, and within Cellulosic Solids Burning in Air," Ninth Symposium (International) on Combustion, pp. 910-923, 1965. Beall, F.C., "Differential Calorometric Analysis of Wood and Wood Components," Wood Science and Technology, 5, pp. 159-175, 1971. Arseneau, D.F., "The Differential Thermal Analysis of Wood," Can. J. Chem., 39, pp. 1915-1919, 1961. Lipska, A.E. and Parker, W.J., "Kinetics of the Pyrolysis of Cellulose in the Temperature Range 250°-350°C," J. of Applied Polymer Science, 10, pp. 1439-1453, 1966. Madrosky, S.L., Hart, F.G., and Straus, S., "Pyrolysis of Cellulose in a Vacuum," J. Res. Natl. Bur. Std., 56, 343, 1956. Chatterjee, P.K. and Conrad, C.M., "Kinetics of the Pyrolysis of Cotton Cellulose," Textile Research Journal, 36, pp. 487-494, 1966. 102 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 103 Golova, O.P. and Krylona, R.G., "Thermal Decompo- sition of Cellulose and its components," Dokl. Akad. Nauk SSR, 116, pp. 419, 1957. Chem. Abstracts, 52, pp. 4979, 1958. Beall, F.C., "Thermogravimetric Analysis of Wood Lignin and Hemicelluloses," Wood and Fiber, 1, pp. 215-226, 1969. Stamm, A.J., "Thermal Degradation of Wood and Cellulose," Ind. and Eng. Chem., 48, pp. 413— 417, 1956. Shafizadeh, F., "Pyrolysis and Combustion of Cellu- lose Materials," Advan. Carbohydrate. Chem., 23, pp. 419-74, 1968. Browne, F.L., "Theories of the Combustion of Wood and its Control," Forests Products Laboratory Report 2136, 1958. Blackwood, J.D. and McGrory, F., "The Carbon-Steam Reaction at High Pressure," Aust. J. Chem., 11, pp. 16-33, 1958. Lewis, W.K., Gilliland, E.R., and Hipkin, H., "Carbon- Steam Reaction at Low Temperatures," Ind. Eng. Chem., 45, pp. 1697-1703, 1953. Stern, E.W., Logiudice, A.S., and Heinemann, H., "Approach to Direct Gasification of Cellulosics," Ind. Eng. Chem., Process Design and Development, 4. Pp. 171-173, 1965. Garten, R.L., "Catalytic Conversion of Biomass to Fuels," Paper presented at the Second Annual Fuels from Biomass Symposium, Troy, New York, June 20, 1978. Walkup, P.C., Mudge, L.K., Cox, J.L., Sealock, L.J., and Roberts, R.J., "Investigation of Biomass in the Presence of Multiple Catalysts," Paper pre- sented at Second Annual Fuels from Biomass Sympo- sium, Troy, New York, June 20, 1978. Shafizadeh, F. and McGinnis, "Morphology and Biogen- isis of Cellulose and Plant Cell Walls," Advan. Carbohydrate Chem. Biochem, 307-312, 1971. 22. 23. 24. 104 Bird, R.B., Stewart, W.E., and Lightfoot, E.W., Transport Phenomena, Wiley, New York, 1960, Eg. 184-25, p. 571. Panshin, A.J. and DeZeeuw, C., Textbook of Wood Technology, McGraw—Hill, New York, 1970, p. 217. Wakoo, N. and Smith, I.M., "Diffusion in Catalyst Pellets," Chem. Eng. Sci., 17, p. 825, 1962. T '41::11. .. in Iri 5119.51. 4 .. 111111111111111111111111111111111111111111111111111111111111