R. F. Johnson assisted in conducting tests. A. Attari and J. E. Neuzil supervised the analytical work. Literature Cited
(1) Birch. T. J.. Hall, K. R.. Urie. R. I V . . J . Inst. Fuel 33, 422-35 (1 960). (2) Blackwood. J. D.. Australzan J . Chem. 12, 14-28 (1959). (3) Blackwood. J. D., McGrory, F.. Ibid.,11, 16-33 (1958). (4) Feldkirchner. H. L.. Lee. A . L.. Johnson. J. L., Eakin, B. E.. 52nd National Meeting. American Institute of Chemical Engineers, Memphis, Tenn., Feb. 3-5, 1964. (5) Feldkirchner. H. L.. Linden. H . R.. Division of Fuel Chemistry, 143rd Meeting, ACS. Cincinnati. Ohio. Jan. 13-18, 1963. (6) Feldkirchner, H . L.. Linden. H. R.. IND. EX. CHEY.PROC. DESIGN DEVELOP. 2 , 153-62 (1963). (7) Goring. G. E.. Curran. G. P.. Tarbox, R. P., Gorin. E., Ind. E n e . Chem. 44. 1051-65 119521. (8) Goring, G. E.. Curran. G.‘ P.. Zielke. C. \V..Gorin, E.. Ibzd., 45, 2586-91 (1953). (9) Hodge. E. S., Boyer. C. B.. Orcutt. F. D.. Ibzd.. 54, 31-5 (1 0 6 3I) ’ \ A ’ - -
(10) Hunt, E. B.. Mori, S., Katz, S., Peck, R. E.. Ibid.: 45, 677-80 (1953). (11) Linden, H. R.. 48th National Meeting, American Institute of Chemical Engineers, DenL-er; Colo.. Aug. 26-29, 1962; Preprint S o . 14, 17-52. von Fredersdorff. C. G., Inst. Gas Technol. Res. Bull. 19 (May (l%55). (13) IVasilewski, J. C.. Ind. Eng. Chem. 52, 60.6 61.4-64.4 (1960). (14) IVen, C. Y., Huebler, J., 56th National Meeting, American Institute of Chemical Engineers, Houston. Tex.. Dec. 1-5. 1963. (15) Zielke, C. LV., Gorin, E.: Ind. E n g . Chem. 47, 820-25 (1955). (16) Ibid.. 49, 396-403 (1957).
Figure 11. Effect of feed gas steam content and conversion on ratio of methane formation to carbon oxides formation at 1000 p.s.i.g.
this work. Therefore, since there may have been more steam reforming of exit methane in these tests, inhibition by hydrogen may have been obscured.
RECEIVED for review May 21. 1964 ACCEPTED h-ovember 24. 1 9 3 Division of Fuel Chemistry, 147th Meeting, .4CS, Philadelphia, Pa., April 1964. M‘ork supported by the Gas Operations ILesearch Committee, American Gas Association. as part of the Promotion-.4dvertising-Research plan of the association. Il’ork guided by Project PB-23a Supervising Committee under the chairmanship of B. J. Clarke.
Acknowledgment
Thanks are due to H. A . Dirksen, W. G. Bair, and S. I’olchko for their helpful suggestions in operation and design of the equipment. E. J. Pyrcioch, F. C. Rac, A. E. Richter, and
KINETIC S T U D Y OF COAL CHAR HYD ROGAS I FICAT IO N Rapid Initial Reaction C
.
Y
.
J A C K
W E
N ,
We5t Virginia C’niuerstty. M o r g a n f o x n . T1.’ C’a.
H U EBL E R
,
Instztute of Gas Technology. Chlcapo. Ill
The coal char hydrogasification reaction may be regarded as two simultaneous reactions dittering considerably in their rates. The rate of coal hydrogasification, particularly that for the initial rapid reaction, was studied. The initial reaction rates in semiflow tests were analyzed and found to b e proportional to the amount of unreacted carbon remaining in the particles as well as the effective partial pressure of hydrogen. The rate constant for small conversion of carbon was found to be 2.0 (ah-hr.)-’. The reaction model used for semiflow test analysis was satisfactorily applied to continuous free-fall reactor tests.
HE production of synthetic methane irom coal by hydroT g a s i f i c a t i o n has considerable economic potentiality. A number of experimental studies of coal hydrogasification have been reported during the ’last few years. Only a few studies were concerned with reaction kinetics. Many problems associated with coal hydrogasification must be solved before the process can become economically feasible.
142
l&EC
PROCESS DESIGN A N D DEVELOPMENT
O n e of these concerns the properties of coal that affect the hydrogasification rates. From a reaction-kinetic point of view, carbon in coal consists roughly of two types which differ greatly in reactivity-the portion associated with the volatile matter, and the remainder, which corresponds to the residual carbonaceous matter, coke (The term “volatile matter“ is used very loosely in this paper. This portion of the coal
corresponds only roughly to that material designated as volatile matter in the standard procedures for coal analysis. I t is hoped that research now in progress will allow a more accurate definition of this highly reactive portion of coal to be made.) Under the high temperatures and pressures required for hydrogasification, coals tend to soften and agglomerate, preventing free flow through the reactor. Pretreatment of coal, therefore, has usually been required in moving-bed or fluidizedbed reactor operation. Roughly, 20 to 40y0 of the carbon in the pretreated coal char used in this study is in the form of volatile matter which reacts rapidly with hydrogen to form methane a t temperatures ranging from 1300’ to 1800’ F. and pressures ranging from 50 to 200 atm. T h e remainder of the carbon reacts very slowly and the reaction rate is subject to more severe equilibrium hindrance. During the pretreatment of coal, which usually involves oxidation or charring of the coal surface by air or other gases to produce a protective film, a great portion of the reactive volatile matter is lost. I n many cases, as much as one half of the volatile matter carbon content is lost in pretreatment. I t is appropriate, therefore, to search for means of utilizing unpretreated coals in order to utilize fully the highly reactive portion of coal. O n e such process is to conduct the hydrogasification of coal in a so-called “free-fall” reactor. Coal particles pass through the top of the reactor in a dilute suspension, and hydrogen gas is passed either countercurrently or cocurrently. I n such a’ process the contact time between the coal particle and the gas must be, by necessity, shorter t h a n in a moving-bed or fluid-bed operation. This paper presents a kinetic study of the rapid reaction during the initial contacting of coal and hydrogen. Mechanism of Hydrogasification Reaction
It has been postulated that in the initial period of hydrogasification, pyrolysis of the char occurs, resulting in devolatilization of light components, which then undergo hydrogenolysis. T h e devolatilization produces reaction intermediates that are derived from aliphatic hydrocarbon side chains and oxygenated functional groups ( 3 ) . T h e remainder of the organic matter in the coal is converted more slowly to methane, apparently almost according to the carbon-hydrogen reaction. For convenience, the following two abbreviated simultaneous reactions are employed to represent the mechanism of hydrogasification of the different coal components. FIRST-PHASE REACTION C,H,Oi
+ qH2 S aCH4 + dC2He + eH2O + bCOa $I c C 0
This reaction is characterized by very rapid methane formation, probably by pyrolysis and hydrogenolysis reactions not too different from those in petroleum hydrogasification. SECOND-PHASE REACTION C
+ 2H2 $ C H I
This reaction is characterized by a relatively slow rate of methane formation, which approaches the rate of the graphite carbon-hydrogen reaction a t nearly complete conversion. T h e reaction is limited by equilibrium hindrance. Of course, the actual reactions are much more complex than indicated by the above simple schemes. However, these two reactions are believed to be sufficient to characterize the two greatly different rate periods observed in coal hydrogasification ( 7 ) . If X,X I , and x2 are, respectively, the total carbon gasified, the. fraction of reactive carbon gasified in the first-phase reac-
tion, and the residual carbon gasified in the second-phase reaction, then
X = fXl
+ (1 - f) xp
(1)
where f is the fraction of carbon in the coal char which potentially can react according to the first-phase reaction. For very short reaction times, 8, the first-phase reaction prevails, and
x = fx1 Coal char hydrogasification data, obtained in a semiflow system by Feldkirchner and Linden ( Z ) , indicate that the rate of carbon conversion in the first-phase reaction is proportional to the amount of unreacted carbon and to the effective hydrogen partial pressure. Accordingly,
where PH2and P13,* are the partial pressure of hydrogen in the system, and the partial pressure of hydrogen a t equilibrium, respectively, and k l is the proportionality constant for the first-phase reaction expression. Combining Equations 1 and 2:
d X / d 8 = ki(f - X)(PH2 - PH2*)
=
kiP,(f
-
X ) b - y*)
(3)
where P, is the total pressure of the system, and y and y* are the mole fractions of hydrogen in the system and a t equilibrium a t time 8, respectively. Integration of Equation 3 between the initial and final conditions for media of constant hydrogen concentration gives In (1 - X / f ) where y
>> y*
=
-kl P, (y
-j*)
8
(4)
(for pure hydrogen feed and small conversion), In (1 - X i f )
=
(5)
-kl P,y 8
A plot of In (1 - X / f ) us. P,y8 can be made from the experimental data, and the slope of the line, kl, may be evaluated. T o obtain the reaction constant kl from Equation 4, the equilibrium composition of the hydrogen-char system must be available. Equilibrium Reactivity of Bituminous Coal Char-Hydrogen System
A batch system study \vas made a t the Institute of Gas Technology of the bituminous coal char-hydrogen reaction determine the methane-hydrogen equilibrium concentration a t various char conversions. T h e equipment consisted of a reactor of 2-inch inner diameter, 8 inches deep. A sample of char was dropped into the reactor, which contained hydrogen or a mixture of hydrogen and methane held a t 1300’ F. and 2000 p.s.i.g. T h e temperature was controlled so that the system \vas essentially isothermal throughout the run. Approximately 20 hours were required for the system to reach a n apparent equilibrium. T h e carbon content of the reactor char was analyzed to determine the degree of conversion. T h e amount of char used was varied from run to run so as to produce varying amounts of conversions. since it was believed that the equilibrium constant would be a function of the char conversion level. Figure 1 shows that as a function of time in a typical run, rises from below the @-graphite the value of the ratio PCH4/PH2P equilibrium value to well above it, and then begins to fall. This indicates that the apparent chemical activity of the residual carbon is high and holds for a reasonably long time at T h e maximum value of the ratio PcH4/P1122is 1300’ F taken to be the value of the equilibrium constant, K,. VOL. 4
NO. 2
APRIL
19 65
143
04
0.3
T
I
G
02
N N ; a
\
I*
no
0.I
0
rl 1 I I I 30
50
70
I I I 5001 1 1 I IIO00
1
100
200 300
700
I
2000
TIME, MIN.
Figure 2 indicates the general trend of the K , values with respect to the fraction of carbon converted. T h e value of K p a t 0% conversion was obtained by extrapolation of data from pilot plant countercurrent moving-bed operation a t very low carbon conversions; the value a t complete conversion was calculated from equilibrium compositions of the p-graphite and hydrogen system at 1300" F. Since more extensive data were not available, the effect of temperature on the value of K p was obtained by taking the ratio of K P for @-graphite a t the temperature of the system, and that based on 1300' F., and applying this ratio to the char-hydrogen reaction. An equation is thus obtained which relates the equilibrium constant a t any temperature, T , to the constant a t 1300' F. and a t the same char conversion levels.
Figure 1 . Apparent equilibrium constant obtained a t 1275 minutes and 1300" F.
Figure 2. Approximate trend of equilibrium constant as a function of conversion for hydrogen-char reaction at 1300" F. and 2000 p.s.i.g. total pressure
o i ! l l I l 1 ~ I 1 l 1 ! 1 l l I l l l I I
0.I
09
0.3
0.5
0.4
07
0.6
0.8
1.0
0.9
CONVERSION, FRACTION
1.0
1.0
0.7
0.7
0.5
0.5
0.3
z-
E
0.2
0 0.2
z 3
LL
LL 0
0.1
;
l--+-----i----t---i---i 0.07
0.07
0.05
0.05 INITIAL SLOPE =2.0(ATM;Hx11
0.03 0
20
40
60 TI ME,SEC.
80
100
0.03 0 120
I
20
1 40
I
I
60 TIME, SEC.
I
I- ++--
1
rtl I J
80
-.
-
100
120
Figure 3. Linear relationship of logarithm of first-phase conversion function with time, A, and with time at two different pressures, B i44
l & E C PROCESS D E S I G N A N D D E V E L O P M E N T
Analysis of Semiflow Tests
Feldkirchner and Linden (2) measured coal char hydrogasification rates by dropping char particles into a heated reactor. through ivhich gas was passed a t high velocity. Their semiflow study covers the pressure range u p to 2500 p.s.i.g. and the temperature range u p to 1700' F . They carried out their experiments in such a \yay that both char heatup and product gas residence times were very short. This enabled them to follow directly the course of the reactions during the initial high-rate period. Their data \yere examined in light of the mechanism presented above. Figure 3, A and B. gives plots of In (1 - X l y ) us. 0 for Montour No. 10 char and North Dakota lignite under a variety of conditions sho\vn. Since the feed gas flow rate was high, the change in composition of the gas was small; P,y was approximately constant through the run. As is seen, the initial portion of the data for lignite and bituminous coal char can be represented by straight lines by adopting approximate, arbitrary values of f. From the slopes of these lines, k l is calculated to be 2.0 (atm.hr.)-l. T h e displacement of the intercept a t 8 = 0 from unity is due to the initial heatup of the char to reaction temperature. It is found that the higher the volatile matter content. the larger the value o f f (the fraction of carbon \vhich reacts by the first-phase reaction) must be made if a straight line is to be obtained. Thus. the value o f f for Montour No. 10 char is 0.25a t 1700' F. and that of lignite is 0.48 at 1700O F. These materials have chemical analyses with volatile contents of 0.179 and 0.412, respectively. A slight effect of particle size may be seen in Figure 3 A . However, the slope of the line is independent of the particle size, indicating that the reaction rate is not dependent on
particle diameter in the range studied but the heatup time is; this! in turn, affects the lateral position of the lines. Figure 3 B sho\vs the linear relation of the data from t\vo tests at different total gas pressures. T h e displacement at 0 = 0 is again due to particle heatup time. T h e slope of the lines is again 2.0 (atm.-hr.)-'. A detailed comparison of all available results of tests on Montour No. 10 char with different feed gas compositions is given in Figure 4. Feeds include hydrogen? hydrogennitrogen mixtures: and hydrogen-methane mixtures. As indicated in Figure 2, a slight equilibrium hindrance of the reaction exists even during the rapid, initial reactions. but the effect is so small under most of the experimental conditions. particiilarly when pure hydrogen feed i q employed. that this can be neglected. Figure 5 shows the effect of temperature on the rate constant, k l >and on f. For relatively small temperature changes. the value of k l may be kept constant with a small adjustment of the value off. T h e increase of the value off as temperature is increased may be explained thus: ,4s the temperature of reaction is increased, the rate of devolatilization of reactive carbon becomes larger; consequently, the recondensation of the reactive carbon during pyrolysis to form stable ring compounds (as residual carbon) becomes less. In terms of the amount of carbon available for the first-phase reaction, this means the value of f increases as temperature increases. However. the rate of pyrolysis should also increase as the temperature is increased. Therefore, the extrapolation to the higher temperature range beyond the experimental points should be avoided. Analysis of Continuous Free-Fall Reactor Test Results
I .oo
0.90
0.80
T o apply the above analysis of semiflo\\ test results to continuous-flojv reactors. a material balance betneen the carbon in the solid phase and carbon in the gaseous phase must be made.
0 70 1.0
0 60 0.7 0.5 XI+
I
2-
0.3
i-
o
2
:0.2 3
0 W
3
3 0.1
0.07 0.05 0
20
40
60
a0
100
120
TIME, SEC.
Figure 4. Effect of hydrogen partial pressure and product P l ( y - y*)O on logarithm of first-phase function for hydrogasification of Montour No. 10 char
Figure 5. Effect of temperature and time on logarithm of first-phase conversion function for Montour No. 10 char
VOL. 4
NO. 2
APRIL
1965
145
Table I.
T,' F.
Run No. SD56 SD57 SD58
Results of Countercurrent Free-Fail Reactor Runs
P t , Atm. 139.8 137.6 137.0
1270 1245 1240
G,
F, Lb./Hr. 4.53 7.68 11.75
For plug flow of both gas and solids:
X (aFIC./12)E G (yo - Y)
(7)
y = yo - ( 4 / 1 2 ) ( F / G )X = yo
or
- mX
m = (aii./12)(F/G)
where
Substituting y into Equation 3,
dX/dO = klP,(f
- X ) ( y , - mX - y*)
(8)
Since y * may be taken to be nearly constant for a small conversion taking place in the first-phase reaction, Equation 9 can be integrated between the inlet and exit of the reactor. Hence :
kleP,(jo - j*
- mf) = In
[jO0- y* - m X ) / (Yo
-y*)(f
f, yo
SCFIHr. 46.53 50.21 105.24
- X)l
(9)
IC. 0.78 0.782 0.783
21.7 21.5 21.4
L , Ft. 9 9 7
e, Sec. 8.57 6.67 4.12
kl,
(Atm.-Hr.)-' 3.10 2.10 1.8
The results of several continuous pilot plant free-fall test conducted a t the institute were used to verify the above analysis. T h e values df (Y = q/n, computed from inlet and exit gas analysis of the tests. seem to be affected only slightly by conversion, as is shown in Figure 6. Since residence times could not be readily obtained from the gasification runs, separate experiments were conducted. Terminal velocities of Montour No. 10 particles char were found by dropping the char from a hopper through a pipe 2 inches in inside diameter and approximately 10 feet long. T h e gas medium was air a t 70' F . and 1 a t m . T h e holdup of particles in the pipe was measured by simultaneously closing plug valves at the inlet and exit. T h e terminal velocities of the particles were calculated from the solids holdups and solids flow rates. If W is the weight of solid particles trapped in the pipe and L is the length of the pipe through which the particles fall, the average solid terminal velocity, U,, is
or
k l = [l/PcO(yo- y*
Uc= FL/W
- ( ~ $ F f / l 2 G )In] [ f ( j o ' - y* - a$XF/12G) (Yo
-y*)(f
- X)l
(10)
X, FRACTION CARBON CONVERTED Figure 6. Effect of carbon conversion on stoichiometric ratio of hydrogen to carbon in first-phase reaction
d $0 25
>2.0
2W 1.5
z
(11)
Terminal velocities of the solids a t reactor conditions were calculated from those a t ambient conditions in air by use of Stokes' law. Both the terminal velocity in air and that calculated a t reactor conditions are shown as a function of solids feed rate in Figure 7 . The residence time is
where U, is gas velocity. The results of three countercurrent free-falling reactor tests were used to compute the values of k l and are listed in Table I. In view of the uncertainties involved in calculating the values of particle residence times, the agreement of the k l values from continuous reactor tests with those obtained from semiflow tests [kl = 2.0 (atm.-hr.)-'] is considered good. Because of the high rate of the first-phase reaction, the particles must undergo deformation, reduction in size, and reduction in density during free fall. T h e actual particle residence times must be affected by these changes which did not exist in the residence-time measurement tests which were used to obtain the values of 0. Because of the assumptions and uncertainties introduced for such a complex system, instead of using Equation 10, simplified equations may be employed for practical application without introducing significant errors. This involves the use of an average driving force, ( y - Y * ) ~ " . From Equation 4 :
I- 1.0 IW
- X/f)/Ptb -Y*)~J
(13)
x = f { I - e x p [ - k l P c b - y*).+91}
(14)
kl
= -In
(1
v)
and
0.5
t B o
e
0
2
4
6
8
IO
12
FEED RATE, LWHR.
14
16
18
20
Figure 7. Effect of char feed rate of terminal settling velocity of Montour No. 10 char in a 2-inch inside diameter tube \46
I & E C PROCESS D E S I G N A N D D E V E L O P M E N T
Nomenclature
D, F
= particle diameter, feet or inches = char feed rate, lb. per hour
f
=
G
= =
k,
Kp L P,
P P* T
C, L’t IV
X
= = = = = =
= = = =
x1 = xp
=
y
=
yo =
y* = 0 = I$ a
= =
fraction of carbon that reacts according to first-phase reaction gas feed rate, SCF per hour or moles per hour reaction rate constant in first-phase reaction (atm.hr.)-l thermodynamic equilibrium constant = PcH4*/’(PH,*)2 length of reactor, feet total pressure. a t m . partial pressure. atm. equilibrium partial pressure, a t m . absolute temperature, O R . gas velocity, feet per sec. terminal velocity of particles, feet per sec. \veight of bed, lb. over-all fractional conversion of carbon fraction of carbon converted in first-phase reaction fraction of carbon converted in second-phase reactior mole fraction of hydrogen mole fraction of hydrogen in inlet gas mole fraction of hydrogen a t equilibrium time, seconds or hours fraction of carbon in coal q ‘n. ratio of hydrogen t o carbon stoichiometric coefficients for first-phase reaction
literature Cited
(I) Dent. F. J., Gas. J . 244, 502 (1944). 12) \ , Feldkirchner. H. L.. Linden. H. R.. IND.ENG.CHEM.PROC. DESIGNDEVELOP. 2, 153 (1963j. (3) Linden, H. R.. “Conversion of Solid Fossil Fuels to High B.t.u. Pipeline Gas,” 48th national meeting, American Institute of Chemical Engineers, Denver. Colo.. August 1962. Preprint Paper 14. RECEIVED for review May 21. 1964 ACCEPTED December 16. 1964 Di\-ision of Fuel Chemistry. 147th Meeting. ACS. Philadelphia. Pa., April 1964. Study sponsored by the Gas Operations Research Committee. American Gas Association. as part of i t s Promotion-Advertising-Research Plan. \Vork guided by the Project PB-23a Supervising Committee under the chairmanship of B. J . Clarke. and by the research department of the Consolidated Natural Gas System (now Con-Gas Service Carp.) under the guidance of H. E. Benson and F. E. Vandaveer. Material supplementary to this article has been deposited as DocumeTt No. 8283 with the AD1 Auxiliary Publications Project. Photoduplication Service: Library of Congress, \Vashingtoii. D. C. A copy may be secured by citing the document number and by remitting S1.25 for photoprints or $1.25 for 35-mm. microfilm. Advance payment is required. Make checks or money orders payable to Chief, Photoduplication Service, Library of Congress.
KINETIC STUDY OF COAL CHAR HYDROGASIFICATION Second- Phase Reaction Y
.
J A C
K
C
.
W EN
,
West Virginia Cniisersity. Morgantown, 14.. Va.
H U EB L E R
,
Institute of Gas Technology, Chicago. lli
Reaction kinetics and the mechanism of coal char hydrogasification in a pilot plant reactor under cocurrentand countercurrent-flow systems with both fluidized- and moving-bed contacting were studied. The rate of relatively slow second-phase reaction was investigated. Diffusional resistance was the major controlling factor at low gas velocities. Correlations for different controlling regimes are given, and the effect of major parameters on the rate of reaction i s discussed. The good agreement between the correlated results and the experimental data indicates that the correlations and the proposed reaction model are satisfactory for design purposes. With previous study on the initial rapid reaction available, it i s now possible to predict the carbon conversion as well as the composition and heating value of the product gas a t the exit of the reactor under widely varying conditions.
IN
R E C E ~I years. hydrogenation of coal has shown substantial potentiality for augmenting supplies of natural gas and crude petroleum. A number of research programs are now under \va) all over the world to test and develop efficient and economical production of high heating value gas from coal. Under American Gas Association sponsorship, a n intensive studv of coal hydrogasification has been made a t the Institute of Gas Technology particularly under high-temperature and high-pressure conditions. More recently, the Office of Coal Rrsearch has joined with AGA as a cosponsor of this work. Various investigators (7, 37, 39, 44) have demonstrated that coal consists roughly of two portions that differ greatly in reactivity: a highly reactive portion, possibly corresponding i n some way to the amount of volatile matter present, and a portion of relatively low reactivity, residual carbonaceous matter-coke. T h e authors divided the reaction of these portions into a first and a second phase (46). T h e first-phase
reaction was discussed in detail and a reaction mechanism proposed which adequately describes experimental results. In the study of the chemical reaction of a complex substance such as coal with hydrogen. analyses often fail to reveal the true mechanism because of the great number of variables involved. consequently, any theory of the kinetics of the coal hydrogenation reaction must be based on a number of simplifying assumptions. T h e comparison bet\veen the kinetic curves calculated from a conceptual model and those determined experimentally over a wide range of conditions \vi11 then show to what extent the simplifying assumptions are justified. T h e purpose of the work reported here was to analyze the experimental d a t a obtained from cocurrent and countercurrent operation of both moving and fluidized beds in the IGT pilot units, in light of a reaction y c h a n i s m for the second phase based on thermodynamic and kinetic considerations. and to compare the results of the analysis to the experimental results. VOL. 4
NO. 2
APRIL
1965
147
