Carbonization of Typical Bituminous Coals The effects of the variation of final maximum temperature from 5400 to 1OOO" C. and of rate of heating from 1.4" to 21.8" C. per minute on the compositions and magnitudes of the yields of products and on the hardnesses of the cokes in the carbonization of a series of three coals of similar type and increasing rank, as well as one coal of intermediate rank but of different type, are reported. In the cases of two of the coals it is shown that the greater part of the changes in magnitude of yields with change in rate of heating is related to the rate of heating used in heating the coal through a limited temperature range, which has been named the "sensitive range," and that this range lies either wholly or in greater part below the plastic ranges of these coals. A theory of the mechanism of carbonization, based on these facts, is presented.
Effect of Rate of Heating and Final Maximum Temperature WM. B. WARREN Carnegie Institute of Technology, Pittsburgh. Pa.
P
REVIOUS studies of the carhonization of a typical coking coal showed that the yields of solid, liquid, and gaseous products were proportional to the logarithm of the rate of heating (IO) and that the effects caused by a change in the rate of heating were determined ahnost entirely by the rate of heating through a limited temperature range just below the temperature of initial plasticity of that coal (11). Similar studies have now been made on three other bituminous coals of different rank and typc. Table I gives analyses of these coals and, for comparison, data for the Pittsburgh coal previously reported. The data given here permit conclusions to be drawn regarding the mechanism of coal carbonization which have more general significance than those presented in the two earlier papers. The apparatus and general procedure described in the earlier papers were used. The samples of Illinois No. 6 and Pocahontas No. 3 coals were obtained a t the mines and brought to the laboratory. The eo51 was immediately mixed, ground, and sifted in natural gas, and the sized coal was stored in glass bottles which had been flushed out with natural gas and sealed with paraffin. To keep the oxidation of the Illinois No. 6 coal a t a minimum, the air in the drums was replaced with coke-oiTen gas during shipment. The High Splint coal was shipped by the mine operator to the laboratory in a wooden barrel and, on a r r i d , received the same preparation as the other two samples. All the work reported here was done with the 20- to SO-mesh fraction for which the analyses in Table I are given. 136
FEBRUARY,1938
INDUSTRIAL AND BNGINEERING CHEMISTRY
137
Constant Heating Rate The rates of heating and the final maximum temperatures used in the work with constant rates of heating of the four coals are given in Table 11. The actual data obtained are not shown but were a n a l y z e d by the methods described previously. The yields of coke, liquor and “free carbon,” gas, and tar for each final maximum t e m p e r a t u r e were plotted against the logarithm of the rate of heating; the best straight line fitting the data was d e t e r m i n e d by the method of least squares. The deviations of the individual observations from the line were of the same magnitude as the experimental errors determined from the earlier work (IO), indicating that for all four coals the yields of carbonization products are proportional to the logarithm of the rate of heating, or expressed mathematically, Y
where Y
=
YI
=
R
= =
a
=
Y l + alog1oR
yield on b a s i s o f c o a l charged, % a constant which equals percentage yield at 1”C. per min. rateof hegting, C.permin. a constant which e q u a l s change in yield for tenfold change in R O
The constants Y1 and a, which summarize the experimental data, are recorded in Table I11 in Roman type. It is evident that both the yields of products a t unit rate of heating ( Y 1 values) and the effects of rate of heating a t constant temperature (a values) are dependent on the coal used. The average deviation of the experimental points from the values determined from the best straight line is recorded in Table I11 and, as stated above, the values for the Pocahontas KO.3, Illinois No. 6, and High Splint coals compare favorably in magnitude with those for the Pittsburgh coal. According to mathematical theory, if the straight lines were the best representations of all the data for each coal a t each temperature, the values of I: in Table I11 should be 100.0and 0.0 for the values of yield a t l o C. per minute and of change in yield for a tenfold change in rate, respectively. The relatively small deviations from these figures lend confidence to the conclusions drawn from the data.
Courtesy, K o p p e r s Company
MODERNBY-PRODUCT COKEPLANT Coke ovens are shown in the upper section.
Since it seems logical to regard the mineral matter and moisture in the coals as inert diluents and since the amounts of these inert diluents differ for each coal (Table I),recomputation of the values of Y1 and a was made on an ash- and moisture-free basis as follows: (1) the percentage of ash in the coal is subtracted from the experimental Y1 value for
TABLE I. DESCRIPTION OF COALS~ =-.-
.^_.-I
Seam
Mine
Location
Type
Bituminous Rank
Moia- Volatile Fixed ture matter carbon
% Illinois No. 6 Pittsburgh Pocahontas No. 3 High Splint
0
Orient No. 1 Franklin, Ill. Bright High-volatile B Edenborn Fayette P a Bright High-volatileA Pinnacle MoDow’ell W. Va. Low-volatlle Bright Clo-splint Dull Harlan, K;. High-volatile A Proximate analyses of coals on ass-received basis, courtesy of U. S. Bureau of iMines, Pittsburgh
%
%
0.7 33.9 50.3 1.9 33.6 57.0 0.8 15.3 78.3 2.9 36.4 57.3 Experiment Station.
Ash
Sulfur
%
%
9.1 7.5 5.6 3.4
0.9 1.0 0.5 0.6
Heat content B.t.u. 12,380 13,910 14,750 14,050
VOL. 30, NO. 2
INDUSTRIAL AND ENGINEERING CHEMISTRY
138
tion, if the column is white a t the base line, increased rate of heating decreases the yield, and if the column is black a t the base line, increased rate of heating increases the yield. In considering the data presented for the coals, their relative rank must be borne in mind. On the bases of moist mineral-matter-free B. t. u. and of dry mineral-matter-free fixed carbon, used in the classification of coals by the A. S. T. M. (I), the order of increasing rank of the coals is 11linois No. 6, High Splint, Pittsburgh, and Pocahontas No. 3. Although on these bases the High Splint is about halfway between the Illinois No. 6 and the Pittsburgh coals, it differs from both and also from the Pocahontas No. 3 in type (Table I). The type names "dull" and "bright" have no official standing. They are used here as descriptive terms. The dull coal from the High Splint seam is predominantly durain or attritus; the other three coals characterized as bright are predominantly vitrain or anthraxylon. A t constant heating rate, the yield of coke decreases and that of gas increases for all four coals with increased final temperature of carbonization. The yield of coke is less a t each temperature, the lower the rank of the bright coals. Except a t 540" C. and 10' C. per minute, the dull High Splint coal gives the least coke and the most tar-less and more, respectively, than would be expected on a rank classification disregarding differences in type. The yield of gas as a percentage of the weight of the moisture- and ash-free coal is not a simple function of rank. The tar recovered is greater, the lower the rank of the bright coal, except a t 700" C. with low heating rates, where the relative positions of the Illinois No. 6 and the Pittsburgh coals are interchanged. The yield of liquor and free carbon from the High Splint coal is essentially independent of the temperature of carbonization between 540" and 1000° C.; that from Illinois No. 6 coal increases with increase in temperature. With increase in
coke, and the resulting figure is multiplied by the ratio of 100to 100 minus the sum of ash and moisture in the'coal; (2) the percentage of moisture in the coal is subtracted from the experimental YIvalue for liquor and free carbon, and the result is multiplied by the same ratio; and (3) the experimental Yl values for gas and tar are multiplied by the same ratio without other correction. The a values are corrected simply by multiplication by the same ratio. The values so TABLE11. EXPERIMENTAL CONDITIONS Coal Illinois No. 6
Final Temp., -Heating " C . 0 . 7 1.4
.. ... . . x. .
540 700
......
1000 Pittsburgh Pocahontae No. 3 High Splint
540 700 1000 640 700 1000 640 700 1000
x
x x
x
x
x
Rate,
2.7
x x
a
C./Min.-
6 . 5 10.9 21.8
.x. . . x. .
x x x
......
x
x x
x
x x
x
x
x
x x
.x. . .. .. .. ..... . . .x . ... x ...... x ...... x x ... x x ... x x x x x x x x ...
x x
x x
x x
x x
-
obtained are given in italics in Table 111, and are shown graphically in Figure 1; they permit a comparison of data obtained for the four coals on a common basis. In Figure 1 the value of the yield a t a heating rate of 1' C. per minute is represented by the height from the base line to the top of the black portion of each column; the corresponding value a t a heating rate of 10" C. per minute is represented by the distance from the base line to the top of the white column. I n effect, the shortest column has been placed in front of the longest one in every case. With this method of representa-
TABLE111. RESULTS OF COALCARBONIZATION AT CONSTANT HEATINQ mms* Illinois No. 6
Yl** Coke, %
+ free C, %
Liquor
Gmts, % Tar,
%
73.7
76.6
13.1 8.3 7.7 9.1
5.6
6.6 -
z
*
0.3
=t
0.6
* *
0.2
-4.2 -5.0 -0.6 -0.7 0.0 0.0
0.8
100.1 100.6
Pittsburgh
Yl T o 540° C.
a***
4.9
78.7 78.6 5.7 4.8 9.6 10.6 4.8
=t
0.4
*
0.3
*
0.8
*
0.6
6.8 -
6.3 -
$0.1 +0.1
98.8 08.7
-3.1 -3.7 -0.2 -0.2 -1.1 -1.3 4.4 6.2
75.3 74.Q 6.7 6.3 11.3 19.4 4.7
Q
-1.7 -1.8 0.1 0.1 -0.6 -0.6
2.3
Pocahontas No. 3 Y1 a 90.7 00.9 3.2
8.6 5.6 6.0
0.6
2.5 -
0.6 -
$0.1
100.1 100.1
-2.7
86.3 86.2 3.2
fO.8
To 700' C.
Coke, % Liquor
+ free C , %
69.6 71.4 14.8
IO.?
=t
0.3
*
0.4
11.6 =t 0 . 8 1s.7 4.0 * 0.4 4.7":
GW % Tar. %
__
z
100.0 100.0
Coke. % Liquor
om, % Tar, %
z
+ free C, %
66.9 68.2 16.0 11.7 13.4 16.8 4.0 4.7
100.3 100.4
0.0 0.0
*
0.4
=t
0.4 0.5
*
0.1
-3.0 -3.6 -1.4 -1.7 -1.2
-1.4 5.3 6.8
0.3
*
0.3
*
0.3
0.4 I
5.8 -
-3.0
0.8 0.0 0.3
0.s
1.4 1.6 -
98.0 07.8
-0.2 -0.3 T o 1000° C . 73.2 * 0 . 3 -1.6 72.6 -1.7 6 . 6 =t 0.3 0.6
6.9 16.1 16.7 3.8
-
4.2 -
-0.3
98.7 QS.7
-0.3
* *
*
0.6
*
0.4
0.7 -0.8 -0.Q 2.3 6.6
2.6 8.6 Q.1 2.1
2.2 -
* * * *
* * * *
0.1
0.1
0.1
0.4 0.4 0.2 0.8
100.1 100.1
-
4-0.5
100.0
4-0.6
89.0
-1.1 -1 . 2 0.8 0.0 2.2 3.4
-
f0.2 f0.S
100.2 100.2
-0.8 -0.8 -0.9 -1.0 0.2 0.9 1.5 1.6
66.5 67.3 9.0
0.0
* *
72.8 74.1 8.8 6.3 7.9 8.4 7.9 8.4
-
-0.1
85.0 1 . 0 . 1 84.8 4.1 A 0 . 1 3.6 10.2 0.1 10.9 0.7 0.1 0.7
-
-1.7
-1.8 0.3
-0.8
-0.Q -1.0 -1.1 -0.2 -0.0
1.9 8.0
High Splint
Y1
6.6 11.3 11 .Q
13.3
14.2 -
* *
0.4
*
0.3
*
0.6
0.2
-
-
-0.1 -0.9
100.0 100.1
* Figures in Rvman type are tsken direot, from experimental data: figures in italics have been corrected for ash and moisture. ** Calculated yields at l o C nun. *** Calculated change in y i e l d for a tenfold increase in rate of heating.
-3.6 -3.8 -0.2 -0.2 0.6 0.6 3.1 3.5
-0.1 -0.1
* *
0.5 0.3
A
0.6
*
0.8
100.1 90.9 64.2 66.0 8.6 6.1 12.4 13.0 14.8 16.8
a
-0.6 -0.6 -0.6
-0.6 0.6 0.7
0.8
0.8 -
f0.2
+o.s
* *
0.4
*
0.5
0.2 0.3
-0.9 -1 . o -0.3 -0.3 0.2 0.9 1.0 1.1
0.0 0.0
FEBRUARY, 1938
INDUSTRIAL AND ENGINEERING CHEMISTRY
rank the yields of liquor and free carbon from all the bright coals decrease a t all three temperatures. At constant h a 1 carbonization temperature, the yield of coke decreases and that of tar increases for all four coals as the heating rate increases. For the three bright coals the effect of heating rate generally on yield of coke decreases and that on yield of tar is relatively unchanged as the final temperature of carbonization increases from 540" to 700" to 1000" C.; the effect of heating rate on yields of both coke and tar is greatest at 540" and least a t 700" C. for the High Splint coal. In general, the higher the rank of the bright coal, the less responsive are the yields of coke and tar to changes in heating rate; with this as a criterion the High Splint coal is intermediate in rank between the Illinois No. 6 and the Pittsburgh coal a t 540" but a t 700" C. is nearly the same rank as the Pocahontas No. 3 coal, and a t 1000" the apparent shifting in rank is necessarily attributed to the difference in type. The effect of heating rate to the three
DIAGRAM OF COKEIN BY-PRODUCT AND BEEHIVE OVENS
Courtesy, K o p p e m Company
maximum temperatures of carbonization on yield of gas and on that of liquor plus free carbon is generally small and there is little correlation with rank. However, a t all three temperatures, the yield of liquor plus free carbon is decreased for Pocahontas No. 3, Illinois No. 6, and High Splint coals, and increased for Pittsburgh coal by increased rate of heating. The yield of gas from High Splint coal is increased and from Illinois KO. 6 coal is decreased a t all three temperatures by an increase in heating rate, except that at the lowest temperature there is no effect of heating rate on yield of gas from the Illinois No. 6 coal. In the previous paragraphs the yield of gas as a percentage by weight of the coal was considered. I n Table IV the average gas compositions are given, from which it is evident that in general the gas becomes richer in hydrogen and poorer in saturated hydrocarbons (largely methane) both as the temperature of carbonization is increased at constant heating rate, and a t constant temperature as the heating rate is increased. For the three bright coals a t constant temperature of carbonization and constant heating rate, the percentage of hydrogen in the gas generally increases with increase in rank. The cokes obtained in these studies were tested for hardness by a standardized grinding procedure (IO). Samples of 60- to 100-mesh coke were wed in all cases, but the cokes from the Illinois No. 6 coal were so soft that a modification of the standard procedure for preparation of the 60- to 100-mesh sample was necessary. These samples were ground under standard conditions, and the percentages remaining on the 100-mesh screen were determined and used to characterize the hardness. As in the earlier work on the Pittsburgh coal (IO) the hardness of the cokes, H , was found to be proportional to the logarithm of the rate of heating, R, within the reproducibility of the tests:
100-
eo -=
1
80 -
70. 6050
-
40
-
V, 3020 IO.
e-
n
10.
139
I
FIGURE1. EFFECTOF VARIATION OF RATEOF HEATINQ UPON Y I ~ L DOF S PRODUCTS IN THI CARBONIZATION OF POCAHONTAS No. 3, PITTSBURQH, ILLINOIS No. 6, AND HIQRSPLINT COALS
H =clogR+d where d = a constant which equals the hardness at unit heating rate of 1" C./min. c = a constant which equals the change in hardness for a tenfold change in R
-
Yield at 1 g. min = height from base line to top of black column. Yield at 10 &./min. height from base line to tpp pf white cqlumn. If column is blaok at base line, increased rate of heating increases yield. If column is white at base line, increased rate of heating decreases yield. Magnitude of tenfold change in rate is proportional to height of upper portion of each column.
TABLEIV. TABLEOB AVERAQEGASCOMPOSITIONS Final max. temp.: H e a t h Rate, Illinois s o . 6
C./min. Hn CnHnn + z Pittsburgh HI CnHm + z Pocahontas No. 3 Hn CnHan + P High Splint Hr CnHm + 1
0.7
....
1.4
.. ..
13.0 13.1 69.7 69.8
.... .. ..
.. ..
16.9 59.7
540' C. 2.7 5.5 10.9 21.8 21.6 28.1 55.6 48.4 17.7 21.8 20.6 25.4 63.1 58.3 63.1 50.7 20.8 42.0 65.0 . . 44.9 21.0 24.5 . . 54.0 . . 51.3
..
.. ..
.. ..
..
.. .. ..
0.7
1.4 32.5 .. 50.7 32.4 37.5 53.1 48.8 51.7 38.4 36.7 42.2
..
.... .. ..
700' C. 2.7 5 , 5 37.8 42.2 46.2 41.4 44.5 48.7 42.8 39.1 ,. 57.1 33.7 40.3 45.0 40.4 36.7
..
10.9 48.3 35.6 50.0 35.6
21.8 49.2 34.6 51.2 36.0 64.0 27.2 51.0 50.9 32.7 30.0
....
... .
0.7
..
1.4
..
51.0 53.3 37.6 36.2
.. ,.
.. ..
..
51.3 32.0
..
1000~c. 2.7 5.5 53.7 33.5 56.6 47.9 33.3 30.9 65.3 27.7 . 55.8 57.9 29.0 26.1
.. ..
. ..
10.9 21.8 . 58.1 .. 26.6 59.0 60.1 30.2 29.1 70.3 . 22.1 57.7 58.2 25.7 25.0
.
...
Table V, which summarizes the experimental data, shows that in general the hardness of the cokes increases with increase in final temperature of carbonization; from the constants given in Table V this statement can be shown to be strictly true a t heating rates of 3" to 30" per minute. At a final carbonization temperature of 540" C. an increase in heating rate produces a weaker coke from each coal except the Illinois No, 6. At 700" C. an increase in rate of heating gives stronger cokes except for the cokes from the Pittsburgh coal. At 1000" C. there appears to be no effect on cokes from the Pocahontas No. 3 and High Splint coals, while the effects on the cokes from the Pittsburgh and the Illinois No. 6 coals are in opposite directions; in the former case rapid heating produces a weaker coke. The advantages of rapid heating of Illinois No. 6 coal for production of stronger coke have been recognized previously (7), and the best commercial cokes made with this coal have been produced in an oven in which the heating rate is higher than usual in by-product carbonization (9). TABLEV. COKEHARDNESS
Illinois No. 6 Pittsburgh Pocahontas No. 3 High Splint
VOL. 30, NO. 2
INDUSTRIAL AND ENGINEERING CHEMISTRY
140
Hardness,No. at Unit Heating Rate 540' 700' 1000° c. c. c. 0.7 3.8 1.6 8 29 6 4 15 6 6 12 10
Change in Hardness No. for Tenfold Increase in Heatins Rate 540' 700 lOOO", c. c. C. 0 +1.9 11.7 -4 -2 -12 -1 +3 0 -5 +1 0
Nonuniform Heating Rate A previous study with the Pittsburgh Seam coal (11) showed that the range within which rate of heating is a factor in determining the relative yields of solid, liquid, and gaseous products lies just below the plastic range of that coal and probably above 200" C. This temperature interval for the Pittsburgh coal is 200' to 393" C. and was called the "sensitive range." In the present paper results of a similar study to determine the limits of the sensitive range of Illinois No. 6 coal are reported. The method first employed in this series was the same used previously with the Pittsburgh seam coal (11); that is, the whole temperature range from 20" to 700" C. was considered to be divided into the preplastic, plastic, and postplastic ranges. A series of experiments was made in each of which two rates of heating were used-a slow one of 1.4" C. per minute and a rapid one of 21.8" C. per minute. These rates may be represented by S and R, respectively, and the conditions of any particular experiment may be represented by a series of three letters; the first letter indicates the rate through the preplastic, the second that through the plastic, and the last that through the postplastic range. The plastic range of Illinois No. 6 coal by a modified Layng-Hathorne test is 405" to 450" C. according to Fieldner and co-workers (6). This is a short plastic range; yet the interval between initial contraction (275" C.) and initial expansion (414" C.) is unusually long. Moreover, this coal is reported to show no resistance in the Davis plastometer and differs markedly in its plastic properties from Pittsburgh seam coal. The complete series of eight experiments, including all the possible combinations of rates over the three intervals, was made in duplicate to a final temperature of 700" C. with the standard one-hour hold a t 700" C.; the intervals were 20405", 405-450°, and 450-700" C. The results from the two coals are plotted in Figure 2. It cannot be said that the rate of heating through the preplastic range alone determines the yields from this coal. It is true that all runs of the form Rxx,where X is either
S or R, yield more liquids and less solids than does any run of the form SXX, but it is not true, as in the work with Pittsburgh seam coal, that treatment subsequent to the preplastic range has no effect. The effect of a change in the rate of heating on gas composition is not confined to the preplastic range, as indicated in Table VI, and no significant correlation between coke hardness and rate of heating through any particular temperature range could be found. The hardness of the Pittsburgh seam cokes was previously reported to be affected chiefly by the rate of heating through the plastic range and to a lesser degree by the rates through the pre- and postplastic ranges. In order to determine the true location of the sensitive range of this coal, a series of runs was made to a maximum temperature of 700" C. with a one-hour period of constanttemperature heating a t 700" C. as usual. Each run was started a t a rate of 21.8" C. per minute, but successive runs were changed to 1.4" C. per minute at 200°, 300°, 405", 450", 500", 550", and 600" C . Those runs can also be included in this series which were made with a constant rate of 1.4' and 21.8" C. per minute. When the four items coke, liquor, tar, and gas were plotted against the temperature a t which the rate of heating was changed from 21.8" to 1.4" C. per minute (Figure 3), it became clear that an abrupt change occurs above 300" C., and that after 450" little further change takes place. As a first approximation it was considered that there was no real difference between any of the runs in which the rate was changed a t any temperature up to 300" C. and also for all runs in %
I
I
a
I
I
I
1
I
YIELDS
-
75
I
~
I
74
1 T
I
8I
1
15
I 14
I
I
I
1
I
I
I
0
I
I
I
I
e
I O
PITT.
6 LIQUID I
A
I
I 1
J
I1
SSS
SRS
SSR
FIGURE 2. EFFECT ON YIELDS OF VARIATION O F RATEOF HBATI'NQTHROUQH PRBPLASTIC, PLASTIC, AND POSTPLASTIC RANGES OF ILLINOIS No. 6 AND PITTSBURGH SEAM
COALS
FEBRUARY. 1938
INDUSTRIAL AND ENGINEERING CHEMISTRY
which the rate change was made between 450" and 700" C. This leaves out of consideration only those runs in which the rate change was made at 405' C. Each of the two classes was considered to be a group of measurements of the same thing, and the averages and probable errors of the averages TABLEVI.
EFFECTOF VARIATION IN HEATINQ RATE ON GAS OF ILLINOIS No. 6 COAL COMPOSITION Av. %
Av. %
cn4+ C
H I
RSR RRR SRR SSR RRS RSS SRS
sss
53.6 48.2 48.2 47.7 38.6 38.3 37.8 32.5
SSS
SRS RRS RSS SSR SRR RRR RSR
50.7 44.2 43.5 43.0 37.4 37.0 34.9 30.8
~ H ~ XXR XXS
Av. %
n2
49.4 36.8
12.6
RXX SXX
44.7 41.5
3.2
XRX
43.2 44.5
1.3
xsx
were determined for each of the four items. As the probable errors were no greater than those obtained in a series of runs in which all conditions were held constant, i t was considered that the first approximation cited was correctthat is, that there was no effect of rapid heating up to 300" or after 450" C. In other words, for Illinois No. 6 coal, the range within which rapid heating is effective in changing the relative yields of products lies between 300" and 450" C. This work indicates that the sensitive range is not necessarily directly related to the plastic range, as might have been concluded from the work with Pittsburgh seam coal previously reported. It is interesting to note that the points representing the runs in which the rate was changed while the temperature was within the sensitive range are no farther from the line joining the two averages than might be expected from the experimental error of the determinations, indicating that the magnitude of the variation in yield is related to the portion of the total sensitive range rapidly traversed. The figure shows that tar increases at the expense of coke whenever the charge is heated rapidly through the range 300" to 450" C.; and that gas and liquor are little affected by any change of rate within the range studied.
Mechanism of Coal Carbonization The facts observed in the course of this work may be qualitatively explained by considering that coals, on heating, undergo in sequence different types of reaction which have different temperature coefficients, and that the first action of heat is to depolymerize the coal to relatively small units of structure in accordance with the point of view developed through the work of several members of the staff of this laboratory (2, 3, 4,6, 8). This depolymerization may be followed, or accompanied, by a mild decomposition as a result of which water and simple gases are released from the periphery of the depolymerizedunits of structure. At slightly higher temperatures the sensitive range is entered during which the depolymerized molecules react, probably by condensation, to form larger ones. The extent of these reactions will depend on the length of time available-i. e., the rate of heating-since it is probable that the reactions are relatively slow. With rapid heating the number of small molecules which survive until a high enough temperature is reached for them to escape through mild decomposition and distillation is larger than at slow rates of heating where reaction between the small units has largely converted them to relatively large molecules. These large molecules will not distill without active decomposition and consequent conversion in greater part to coke and gas. A t temperatures above the sensitive range, distillation is extensive and the distribution of the carbonization products among the solid, liquid, and
141
gaseous forms will P I I I 1 depend, a s i n d i cated above, upon the extent of condensation w h i c h has occurred during the time the coal w a s i n t h e s e n s i t i v e range. It seems probable that for each coal there is a maximum p o t e n t i a l yield of distillation products which is obtainable o n 1y when the rate of heating is so high t h a t appreciable ! I condensation of 4 , the depolymerized molecules does not o c c u r w h i l e the temperature of the coal is within the sensitive r a n g e . High heating rates 1 I I 1 600 200 I 400 I would not increase 0 300 450 further the yield of TEMP OF RATE DECREASE distillation p r o d FIGURE 3. EFFECT ON YIELDS OF THE ucts to a practical TEMPERATURE AT WEICESLOW HIOATextent. O n t h e ING WASSTARTED, ILLINOIS No. 6 COAL other hand. there should be some minimum yield of distillation products which is obtainable when the heating rate is so low that the greatest possible amount of condensation takes place while the temperature of the coal is within the sensitive range; reducing the rate below that a t which complete condensation takes place will result in no further decrease in yields of distillation products. i
I
I
Acknowledgment It is a pleasure to acknowledge the fine cooperation of the following firms and individuals in connection with securing the samples of coal used in this work: W. E. E. Koepler, Pocahontas Operators Association, Wm. Buery, Algoma Coal & Coke Company, Thomas Moses, H. C. Frick Coke Company, G. B. Harrington, Chicago, Wilmington & Franklin Coal Company, and H. N. Eavenson, Clover Splint Coal Company. Thanks are aleo due John Robertson for aid in preparation of samples and testing of the cokes, J. M. Scott for gas analyses, and H. G . Landau for help with the mathematical analysis of the data.
Literature Cited (1) Am. Soc. Testing Materials, Proceedings, 35, Part I , 847-53 (1935). (2) Asbury, R. S.,IND. ENG.CHEM.,26, 1301-6 (1934); 28, 687-90 (1936). (3) Biggs, B. S.,J . A m . Chem. Soc., 58, 484-7, 1020-4 (1936). (4) Biggs, B. S., and Weiler, J. F., Ibid., 59, 369-72 (1937). (5) Fieldner, A. C.,Davis, J. D., Thiessen, R., Kester, E. B., Selvig, W. A., Reynolds, D. A., Jung, F. W., and Sprunk,G. C., U.S. Bur. Mines, Tech. Paper 519 (1932). (6) Howard, H. C.,J . Phys. Chem., 40, 1103-12 (1936). (7) McBride, R.S., and Selvig, W. A.. Natl. Bur. Standards, Tech. Paper 137 (1919). (8) Smith, R. C., and Howard, H. C., J . Am. Chha. Soc., 58. 740-2 (1936). (9) Thiessen, Gilbert, IND. ENQ.CHEM.,29,606-13 (1937). (10) Warren, W.B.,Ibid., 27, 72-7 (1935). (11) Ibid., 27, 1350-4 (1935). RECEIVED July 15, 1937.