I
J. D. BATCHELOR, EVERETT GORIN, and C. W. ZIELKE Consolidation Coal Co., Lib1aryl Pa.
Desulfurizing Low Temperature Char Kinetic data presented here can b e used to design commercial plants for production of low-sulfur char useful in metallurgical applications
Rate and Total inhibition Data in Batch Systems T H E present work constitutes an extension of an earlier investigation of desulfurization of char with hydrogen and hydrogen-steam mixtures ( 75). The basic kinetics of the desulfurization of char with hydrogen and of the inhibition by hydrogen sulfide is further elucidated here. I t is known (7-77) that the reaction of hydrogen sulfidehydrogen mixtures with carbonaceous materials is reversible. However, this is not a true equilibrium process, because the results depend upon the thermal history and origin of the carbonaceous material. Characteristic isotherms were developed for low temperature chars used in this work, which define the point of total inhi bition of the desulfurization process. The results obtained in desulfurizing low temperature char in a continuous unit with both pure hydrogen and recycle gas are interpreted in terms of basic data.
R a w Materials The chars used were produced in a 50-ton-per-day fluidized low temperature carbonization pilot plant at 900' to 950" F. Their sulfur content varied from 1.3 to 4.5 weight 7 0 (Table I). The Moundsville and Alexander Mine chars are interchangeable.
Table 1.
VENT WTM
U
N, CYLIKWR
MULLITE TUBE WINDING
A bed of char is fluidized
INSULATION
with premixed hydrogen and hydrogen sulfide to establish equilibrium distribution of sulfur between gas and char
Yl
METER
Total inhibition Data The data reported here were obtained by a stmdard equilibrium method. Chars Of Several sulfur contents were brought in contact with hydrogen Sdfide-hYdrOgen mixtures of several concentrations for runs 10% enough to approach equilibrium. A single series of experiments with a constant gas composition caused sulfiding of the lower sulfur chars and desul-
Analysis of Feed Chars Shows Variation in Sulfur Content (Pittsburgh Seam coal) Raw
Arkwright Ia Raw Pretreatedb
H C
N 0
s
Ash Sulfide S
VYCOR REACTOR
Raw Arkwright IIe
Raw Montour 10
Alexander Mine 3.25 70.46 1.25 5.89 4.51 14.64
2.99 77.69 1.74 4.93 1.96 10.69
0.68 84.70 1.27 0.68 1.43 11.24
3.22 78.29 1.72 5.01 2.53 9.23
3.34 80.46 1.79 6.56 1.33 6.52
0.13
0.39
0.39
...
furization of the higher sulfur chars. The sulfur content corresponding to total inhibition was thus bracketed by approaching it from both directions. The chars employed for developing a particular isotherm covered a range of sulfur values. They were prepared by desulfurization of the standard nitrogenpretreated char with hydrogen. The pretreatment and desulfurization temperatures were the same as those for which the isotherm was developed.
Table II. Typical Bed Residues from Deep Bed Desulfurization at 1600' F. (6 atm. Hs)
wt. of Feed, G.
860'
l72Ob
s
0.98 85.25 1.09 0.85 0.42
0.79 85.45 1.00 0.50 0.33
0.49 05.49 1.11 0.43 0.89
Ash
11.41
11.93
C
N 0
...
Feed char to deep bed, desulfurization runs. Pretreated at 1600' F. for 1hour in nitrogen. Feed char to shallow bed and total inhibition experiments.
430"
H
60-minute treatment time. treatment time. a
VOL. 52, NO. 2
11.59
* 30-minute
FEBRUARY 1960
161
Table 111.
Sulfur Reduction Is Lower in Deep Beds for Same Treatment Time
(Conditions. Pressure 6 atm. abs. Temperature 1600" E'. LTC CHAR MINE R
Bed Weight, Grams Approx. After VOl. Raw Feed drying and Hydrogen Hz/Lb' Charged, pretreatTreatment Char G. ment Final Time, Min. SCF/Lb. 500
1000
2000 0
08 WT %
16
SULFUR
430
860
1720
24 32 IN SOLID
Figure 1. Total inhibition curves for chars at 1350' F. lowest sulfur value is reached by char derulfurization and highest value b y char sulfurization
T h e reactor was a Vycor tube, 22 mm. in inside diameter about 20 inches long, with a standard-taper joint at the upper end for the take-off tube and thermowell assembly. A mixture of hydrogen sulfide and hydrogen was made up and stored in the gas mixture bomb. Hydrogen sulfide was determined by passing a portion of the mixture after drying with Anhydrone through a Nesbit bottle containing Ascarite and measuring the volume of residual hydrogen passed. T h e hydrogen sulfide content was obtained from the weight gain of the Ascarite. The
416 411 405
30 60 120
200
835 822 812
30 60 120
1671 1650 1633
30 60 120
analysis was made before and after a run and the average used for the calculations. A 10-gram char sample was used for an actual run. The Vycor reactor was brought up to temperature with nitrogen passing through the char bed. The nitrogen was replaced with the hydrogen sulfide-hydrogen mixture after the temperature was lined out. T h e reactor was operated as a fluidized bed at a superficial velocity of 0.4 foot per second. Treatment of the char was continued generally for from 2 to 4 hours. The residue char was analyzed for total and sulfide sulfur. The observed sulfur level of the char was then associated as near equilibrium with the gas composition used. Total sulfur was determined by the Eschka method (73), while sulfide sulfur was determined by the evolution method (73). All the data obtained in runs of this kind are not reproduced, but are given graphically in condensed form in Figures 1, 2, and 3. Deep Bed Runs
I" I
a
0
P
Figure 2. Effect of temperature on total inhibition isotherm for Arkwright char
162
The procedure used in these runs was similar to that of earlier gasification (76) and desulfurization (75) studies. The main difference was the use of a pressure-balanced reactor 4 inches in internal diameter instead of the Uniloy reactor 11/* inches diameter (76). All runs were made at 1600" F. and 6-atm. pressure using 35- to 100-mesh Arkwright I char in a batch fluid bed fluidized with pure hydrogen a t 0.2 foot per second. The char was introduced when the reactor temperature reached 950' F., and was fluidized in nitrogen and preheated at 1600' F. for 1 hour (Table I). The reactor was then pressurized to 6 atm. and the hydrogen turned on. A series of nine runs was made with hydrogen treatment times of 30, 60, and 120 minutes and initial bed weights
INDUSTRIAL A N D ENGINEERING CHEMISTRY
50
Inlet HI rate 97.7 SCFH at 70° F.)
Sulfur Content of Residue Organic Sulfide Total 0.34 0.29 0.12
0.14 0.13 0.07
0.50 0.42 0.19
100
0.38 0.23 0.14
0.18 0.10 0.09
0.56 0.33 0.23
12 25 50
0.47 0.31 0.21
0.42 0.24 0.12
0.89 0.55 0.33
100 25 50
of 430, 860, and 1720 grams. The loss in bed weight and in sulfur due to the nitrogen pretreatment was determined by separate experiments. Only proximate and ultimate analyses of bed residues were made in this series of runs (Table 11). The complete summary of run conditions and sulfur analyses of the solid residues is given in Table 111.
Shallow Bed Runs The aim of these runs was to obtain differential data-Le., the rate of desulfurization in pure hydrogen and in the absence of any significant concentration of hydrogen sulfide. The runs were made using a batch fluid bed and the Uniloy reactor system. The char bed was, however, contained in a stainless steel liner which was lifted out of the reactor at the end of each run. One important change in procedure was made. The initial pretreatment for 1 hour in nitrogen before desulfurization was eliminated by preheating the reactor to desulfurization temperature in a hydrogen atmosphere before admitting the char. A 10-gram char sample was usually used, although all the work with the Alexander char and some with the other chars were done with a 5-gram sample. The weighed char sample was charged, and the reactor sealed and brought to full pressure and temperature as quickly as possible. This took between 4 and 10 minutes. The small char bed was fluidized with hydrogen a t a velocity of 0.4 foot per second for 1 hour after full pressure was reached. The reactor was quickly depressurized and purged with nitrogen and the bed residue was removed, cooled in nitrogen, and recovered for weighing and analysis. Only the sulfur content of the bed residues was determined. Three chars of widely different sulfur content were used in this work. Each char was treated at five temperatures
D E S U L F U R I Z I N G LOW TEMPERATURE C H A R from 1000" to 1600" F., inclusive, and a t three pressures-1, 3, and 6 atm. absolute. The total hydrogen used in these runs per pound of char treated varied from about 140 to 280 standard cubic feet, depending on whether a 5- or IO-gram bed was used in the runs a t 1 atm., to about 600 to 1200 standard cubic feet in the 6-atm. runs. I n a pair of runs a t 1450" and 1600" F. and 1 atm. the hydrogen rate was increased by a factor of 3-i.e., from about 140 to 400 standard cubic feet per pound. T h e sulfur content of the residue a t 1450" F.-i.e., 0.66 weight %-was unaffected by this increase in hydrogenchar ratio, while at 1600" F. a decrease from 0.72 to 0.58 weight 70was noted. The data therefore approach those which would be obtained in a true differential reactor. The reference temperature for standard cubic feet used here and elsewhere in this work is 70" F.
than the desulfurization kinetics. T h e result is conservative, because the total inhibition isotherms are representative for chars that have received several hours of thermal pretreatment at the temperature in question. Powell's data as well as the authors' experience suggest that the sulfur is somewhat more labile in the raw char. DESULFURIZATION IN AN ISOTHERMAL BATCHFLUIDIZED BED. I n the limiting totally inhibited case, the gas leaving the bed will be in equilibrium with the solids in the bed a t any instant. The equilibrium conditions may be developed mathematically in terms of the total inhibition isotherms by forcing a sulfur balance between the sulfur in the gas and in the solid. By this procedure we find
t
24
- X
:/
- ARKWRIGHT LTC CHAR AT 1350'F.
22
I
f-
FOR SUGAR CHAR
0
12
08
04
16
20
Y T % SULFUR IN MAF SOLID
Figure 3. Total inhibition isotherms for organic sulfur depend only upon nature and prior history of char
or
2 4 4 ,
I , , ,
I
I
,
,
I
- EXPERIMENTAL DATA 1720 gm BED ) 3 - EXPERIMENTAL DATA 860 am BED ) - EXPERIMENTAL DATA 430 gm BED )
,
i
A
Interpretation
of Results
Total Inhibition Data. The data of Figures 1 and 2 show the position of the sulfide plateau in the correct place as reported by Powell (8- 77) and are in accord with the equilibrium data of Alcock and Richardson (7). The plateau corresponds to the equilibrium in the reduction of iron sulfide : FeS
Equation 2 may be further simplified if the devolatilization or loss in weight of the char due to passage of hydrogen through the bed can be neglected-i.e., a = 0. I n this case Equation 2 reduces to
14 12
10
08 06
04
+ Hz = HZS + Fe
The length of the plateau is determined by the amount of ferrous sulfide in the char. The amount of organic sulfur "locked" below the sulfide plateau is also in general agreement with Powell's ( 9 ) data for coke. The present data, however, d o not permit distinguishing between the forms of organic sulfur as discussed by Powell-Le., solid solution and adsorbed sulfur. The comparison between the data for Arkwright char and Alexander Mine char shows that in the latter case the curve does not go through the origin, because the Alexander char contains about 0.25 weight 70of sulfide sulfur in the form of calcium sulfide, which is substantially irreducible. The total inhibition data for the organic sulfur on a moisture- and ashfree basis for Arkwright char are given in Figure 3. A comparison is also made with Powell's data for sugar char ( 7 7 ) . The inhibition of organic sulfur removal, like that of sulfide sulfur, becomes more marked as the temperature is lowered. Deep Bed Runs. Under conditions where the rate of drsulfurization is rapid, the total inhibition isotherms conservatively determine the amount of desulfurization that can be achieved rather
AT
02
0
Where one is interested only in rLiatively high sulfur levels-Le., above about 1.2 to 1.3 weight 7 0 sulfur for the Arkwright case-it is a sufficiently good approximation to treat f(S) as a linear function : f ( S ) = k ( S - S,)
(4) Using Equation 4, Equations 2 and 3 reduce to the simplified Equations 5 and 6, respectively. 1
log
(
F
H
cy
)
=
(32k-rr)
10
20
30
TOTAL HYDROGEN &ED
40
53
60
SCF/Lb
Figure 4. Application of Equation 3 shows amount of desulfurization to b e expected in batch fluidized beds when equilibrium i s established. Comparison with experimental results in deep beds shows that desulfurization i s controlled b y total inhibition and not rate -.-e-.
----
1 3 5 0 ' F. equilibrium curve, 2.43% initial S 1600' F. equilibrium curve, 1.43% initial S 1600" F. equilibrium curve, 2.43% initial S
(32k-a)So-32kSi - a ) S - 32 k S ,
log [ ( 3 2 k
These equations can be used to determine the "maximum" amount of desulfurization that can be achieved in batch fluidized beds. The first rather remarkable result worth noting is the rather high degree of desulfurization achieved in batch desulfurization of the feed char. The desulfurization from 1.96 to 1.43 weight yo sulfur was accomplished simply by devolatilization in a stream of nitrogen a t 1600" F. Results reported below show that under these conditions the yield of devol-
atilized char is 88.7 weight Yo of the feed char, while about 2.6 standard cubic feet of hydrogen is evolved per a = 16.2. pound of feed char-i.e., O n the assumption that all the hydrogen was evolved at 1600" F. and using Equation 5, one calculates that the sulfur content of the devolatilized char should be 1.35 weight 7 0 as against the observed value of 1.43 weight Yo. Thus, very close to equilibrium conditions are maintained during devolatilization of the char in spite of the very low average partial pressure of hydrogen. The extraordinary lability of the sulfur in low temperature char is emphasized. VOL. 52, NO. 2
0
FEBRUARY 1960
163
I ATM I
0
14
0
1003
1100
1200 TEMP
1300
- ‘F.
-
1400
1500
1
1600
Figure 5. Desulfurization of char in shallow beds A. B. C.
“1
o
!mo
1100
1200 1300 1400 TEMPERATURE -‘F
1500
1
16W
The degree of desulfurization is also clearly greater than in the decolatilization of coal in a coke oven. The reason for this can be traced back to the total inhibition isotherm. The staging effect obtained by batch desulfurization in a fluidized bed is now absent, as the coke sees on the average a gas corresponding
Medium sulfur (Arkwright) l o w sulfur (Montour 10) High sulfur (Alexander Mine)
in sulfur content to the whole coke oven gas. For example, if the Arkwright char were equilibrated with the whole devolatilization product gas, the sulfur content of the product char would rise from 1.43 to 1.65 weight 70. After the char is devolatilized, further weight loss in the hydrogen treatment in deep beds is rather minor (Table 111). As a good approximation it is possible to assume CY = 0 in establishing the equilibrium desulfurization curves for devolatized chars. It is also necessary to assume that the total inhibition isotherms are independent of the pressure. Some prior published data using
hydrogen-steam mixtures for char desulfurization were treated in the same way. A comparison of the sulfur values observed with those calculated on the above equilibrium basis shows the same results-at partial pressures of hydrogen above 1.5 atm. and for beds larger than 70 grams the desulfurization is equilibrium-controlled. The steam under these conditions thus plays the role of an inert diluent with respect to desulfurization. Shallow Bed Runs. For shallow char beds, kinetics and not equilibrium determines the amount of desulfurization. Thus, from Figure 4 it is clear that the hydrogen-char ratio used in shallow bed runs-Le., greater than 140 standard cubic feet per pound-would be sufficient to achieve very much lower sulfur values than are apparent from Figure 5. All the results with the three chars are plotted on a percentage sulfur elimination basis in Figure 6. Thus the sulfur elimination achieved in shallow beds is, as a first approximation, a function only of the conditions employed and independent of the sulfur content of the feed char. The sulfur elimination shown is the percentage of the total sulfur in the feed char that is rejected and thus takes account of the measured weight loss of the chars during desulfurization. The above correlation is consistent with the assumption that sulfur elimination is roughly a first-order reaction independent of the sulfur level of the feed char: (7)
Desulfurizing l o w Temperature Char
Rate Data in a Continuous System Raw Arkwright Char I1 (Table I) was fed without prior drying. The moisture and elemental analysis was determined for the feed char to each run. Run 58 was made with a char which had previously been desulfurized to 1.477, sulfur. Equipment The continuous bench scale fluidized reactor had been designed to study gasification reactions at temperatures up to 1800’ F. and pressures up to 50 atm. The char was fed from a pressurized hopper by a rubber roll feeder, at a feed rate regulated by a variablespeed drive which turns the feeder shaft. A pressure-equalizing line equalizes the pressure between the feed hopper and the discharge from the rolls. Solids were picked up with process gas-fresh hydrogen or recycle gas-and transported into the fluidized reactor. For removal of oxygen and moisture, the hydrogen was passed through a purifying train which consisted of a
1 64
nickel-alumina catalyst chamber and a silica gel tower in series (not shown). The hydrogen was then metered into the reaction system by means of a rotameter. The recycle gas runs were conducted by recycling the product gas, after removal of hydrogen sulfide, with a Cast rotary pump enclosed within a pressure housing. The hydrogen sulfide was removed by passing through a catalyst tube packed with l/*-inch pellets of a sulfur acceptor containing 80% Cu-lO% V-10% Cr. This material was prepared and regenerated according to the procedure recommended by the U. S. Bureau of Mines ( 1 2 ) . The fluidizing vessel was a cast Duraloy (287, Cr-4Y0 Ni) tube 4 inches in diameter and approximately 55 inches long, with a 60’ cone bottom. It was supported by means of a ring joint flange to an upper section attached to the top flange. This alloy was resistant to attack by hydrogen sulfide. The reactor tube was surrounded by a concentric furnace which contained
INDUSTRIAL AND ENGINEERING CHEMISTRY
four independently controlled heater circuits. The tube was maintained in pressure balance with the furnace acket, both were enclosed in a large water-jacketed pressure vessel. The fluidized bed level was controlled by an adjustable J-tube which entered through a packing gland in the top flange. The solid and gaseous reaction products left the reactor through the J-tube and entered a cyclone 11/2 inches in internal diameter. The solids knocked out of the gas in this cyclone flowed by gravity to either of two char receivers selected b y the char diversion valve mounted above them. The partially cleaned gas passed on to one of two cyclone, l/z inch in internal diameter each of which was mounted inside a fines receiver. A fines diversion valve was placed upstream for switching from one receiver to another. Final cleanup of the exit gas from the fines receiver was accomplished by the dust filter which was packed with a Fiberglas blanket material.
