ISDUSTRIAL A N D ENGINEERIi$'G CHELVIISTRY
84
were undertaken to lay the foundation for the design of such a furnace. Acknowledgment
The writer wishes to express his thanks to J. JV, Turrentine suggestions given and Royster for many freely in connection with this work.
Vol. 23, No. 1
Literature Cited (1) Chance, J . SOL.Chem. I n d . , 37, 222-230 (1918). ( 2 ) Jackson and Morgan, J. IRD. EX. C H E M .15, , 110 (1921). (3) Jackson and Morgan, Ibid., l 5 292 ~ (lg2I). (4) Rhodin, J . SOC.Chem. I n d . , 20, 939 (1901). ( 5 ) Ross, 8th Intern. Cong. d p p l i e d Chem., 15, 217 (1912). (6) Wells, U. S. Geol. Survey, Prof. Paper 98-D,37-40 (1916)
Smelting of Wyomingite and Phosphate Rock in the Blast Furnace' T. P. Hignett and P. H. Royster FERTILIZER AHD F I X E DNITROGEN INVESTIGATIONS, BUREAUO F CHEMISTRY A N D SOILS,WASHINGTON, D. C .
The Bureau of Chemistry and Soils has made atthat is, the more than 65 per N IMPORTANT part tempts to smelt both phosphate rock and Wyomingite, of the work on fertilcent of the heat not usable with a view to recovery of soluble potash salts and phosizer technology a t the for phosphate reduction-all phates from the resulting fume. The results are quite goes to waste. As the hot Bureau of Chemistry and encouraging, volatilization of more than 90 per cent gas from the reduction zone Soils is concerned with the of both KnOand P206having been achieved. The volaascends the furnace shaft, it possibility of applying blasttilization of the PnOaseems to depend on high-temmust preheat the descending furnace technic to the properature heat, while volatilization of KzO is dependent charge and melt the slag. duction both of phosphoric on the addition of chlorides. I n many cases, however, the acid and of p o t a s h . I n a Rock Springs non-coking soft coal was investigated heat required for slagging the paper read before the AMERIas a possible blast-furnace fuel, with promising results. charge is considerably less CAN CHEMICAL SOCIETYa t Indications are that a commercial blast furnace operthan the shaft heat, and in Minneapolis, R o y s t e r and ating on Wyoming leucite, western phosphate rock, Turrentine described an atc on s e q u e n c e the over-all limestone, and Rock Springs coal should be able to prothermal efficiency of the proctempt to smelt p h o s p h a t e duce K 2 0 and PnOaat plant cost of 525 per ton. The ess is relatively low. rock in the bureau's experim e n t a l blast furnace. In product might suitably be in the concentrated form It was in an"effort to utilize spite of the small size of the of potassium phosphates. some of this waste shaft heat for the volatilization of potunit, no operating difficulties were encountered and the process appeared technically simple. ash that the experiments described in the present paper were The commercial success of the phosphate furnace was found undertaken. Some such process is uniquely applicable in to be a question solely of coke consumption, and this coke the Wyoming potash field, where high-grade phosphate rock consumption in its turn was found to be largely, if not exclu- is available both from Wyoming and from southern Idaho. Although in a strictly logical sense the potash produced in sively, a matter of preheating the blast. this process is a by-product of a phosphate furnace, economiUtilization of Shaft Heat cally the phosphoric acid is more nearly a by-product in Although the misused expression "volatilization of phos- the production of potash. I n smelting high-grade phosphate rock in any furnace it is phoric acid" is rather firmly fixed in the literature, in actual fact that PzOsin phosphate rock is not volatilized a t any at- necessary to add to the charge some form of siliceous material tainable furnace temperatures. The removal of P205in a as a flux. The use of a potash-bearing silicate for fluxing phosphate rock and recovering both phosphorus and potash has furnace can be accomplished only by the reduction of phosbeen suggested a number of times (1, t?,6 , B ) . Recently Pike phoric oxide to elementary phosphorus. This reduction re(4) has described the smelting of Idaho phosphate rock and action is strongly endothermic and does not take place with any great rapidity below 1300" C. I n a blast furnace blown Wyomingite in an experimental blast furnace. I n the present with air preheated to 750" C. gas is produced in the combus- experiments, however, the potash silicate was added to utilize tion zone a t about 2000" C. As it passes through the phos- the shaft waste heat, and the amount of silicate used in each phate charge this gas can supply the reduction reaction with case was more than enough to flux the phosphate rock, so heat only while its temperature is greater than 1300" C. I n that a further addition of limestone or burnt lime was necessary to flux the excess silica in the charge. I n other words, other words, only the fraction or 35 per cent a potash furnace was operated to smelt Wyomingite, and the of the heat, is available for reduction. By heating the blast lime contained in phosphate rock was used as a flux for the to a higher temperature the combustion-zone temperature silica in the charge up to the limit of the heat available for can be raised and the fraction of the heat usable in phosphate phosphate reduction and limestone was used for the rest of reduction can be increased. It is obvious, therefore, that the needed flux. any important reduction in the fuel consumption of the phosOperation of Experiment Blast Furnace phate furnace must come through an improvement in the dcThe blast furnace used in the experiment is shown in Figsign of the hot blast stoves. ure 1. It is 80 inches tall and has a hearth diameter of 13 It should not be thought that the so-called "shaft heat"inches, a bosh diameter of 19 inches, and a total volume of 12 1 Received September 22, 1930. Presented before the Division of cubic feet. For comparison a modern blast furnace is 90 Fertilizer Chemistry at the 80th Meeting of the American Chemical Society, feet tall, has a 25-foot hearth and a total volume of over Cincinnati, Ohio, September 8 t o 12, 1930
A
200\&01300,
January, 1931
ILVDUSTRIAL AhTD EA$-GIA%ERINGCHEMISTRY
30,000 cubic feet. The experimental furnace is therefore 0.04 per cent of a commercial unit. The furnace was blown with about 60 to 70 cubic feet of air per minute preheated in an iron pipe stove. The stove operated on city gas, and temperature in excess of 700" C. could not be readily maintained. The preheated air entered the furnace through one or both of two tuy6res. These tuy6res, as well as the brick work in the tuyere breast and bosh walls, were protected by
85
found suspended in the slag emerging from the furnace and remaining imbedded in the slag when i t solidified. The excessive length of time required for the calcination of limestone is due to the endothermic reaction involved in the thermal dissociation of calcium carbonate, involving as it does 420 calories per gram. When limestone containing even a small fraction of its original carbon dioxide content reaches the reduction zone above 1300" C., the carbon dioxide is evolved vigorously and to the 955 calories per gram of carbon dioxide absorbed in the endothermic reaction CaCCs
= CaO
+ COz - 42,000 calories
is added a second 945 calories per gram of carbon dioxide absorbed in the equally endothermic producer gas reaction COZ
+ C = 2CO - 41,500 calories
with a net loss of heat of 1900 calories per gram of carbon dioxide. Of the nine experiments in which phosphate rock and Wyomingite were charged, limestone was used in three and burnt lime in six. I n an actual furnace project of this kind one would naturally burn the stone before introducing it into the furnace, using blast-furnace waste gas for calcination. This is the customary procedure in English iron blast-furnace practice, where low-grade carbonate iron stoves are used. I n the experiments with burnt lime the low reduction of P205 was not observed. The results of fifteen runs on the experimental blast furnace are shown in Table I. The materials used were wyomingite containing 12.76 per cent K20 and 1.97 per cent P206, and a washed Florida rock containing 31.63 per cent P206 and 0.26 per cent K20. The coke was of poor grade, conFigure I-Blast Furnace Used for S m e l t ing Wyomingite and P h o s p h a t e Rock
cooling coils, and the heat loss through the furnace walls was measured throughout each experiment by observing the temperature rise and the amount of the cooling water. I n a large furnace this heat loss is a matter of a very few per cent, but in the experimental furnace it was an importantly large part of the total heat and \\-as lost from the hottest part of the furnace-that is, from the heat available for phosphate reduction. Except for this loss of available heat by conduction through the walls, the furnace is as satisfactory for phosphate smelting as a large commercial unit. A photograph of the furnace in operation is shown in Figure 2 . A larger experimental furnace unit is being constructed at the bureau and with it the loss of heat through the bosh walls will be relatively unimportant. Figure 3 shows the larger furnace under construction and the smaller furnace in operation. I n a production plant the furnace gases would, of course, pass through an electrical precipitator for the recovery of phosphorus and potash, and the gases after being cleaned in the precipitator would be used as fuel in the hot blast stoves, and for the generation of power for blowing and for other plant purposes. I n the experiments, however, the gases were exhausted to the air, and the reduction of phosphorus and the volatilization of potash calculated from the weight and the KzO and P2OScontent of the slag. The time of passage of the stock through the furnace mas about 13/4 hours., Although this is much longer than is necessary for the heating and slagging of the charge for the reduction of P20s,or for the volatilization of potash, it was found not to be sufficiently long to calcine the limestone completely. I n every case where limestone was used the reduction of Pz05 fell off to about one-third of the value calculated from the available hearth heat. I n each case numerous lumps of limestone in a partly calcined condition but still containing a substantial fraction of its carbon dioxide content were
,
I
I
I
Figure I-Blast
Furnace in Operation
taining 79.5 per cent fixed carbon and 15 per cent ish. The charge in each case is given in terms of the interesting constituents, KzO and P2O6, and the carbon in the fuel. I n tests 1 and 2 phosphate alone was smelted and in tests 3, 4, 5, and 6 potash alone; the last nine experiments contained a binary charge of phosphate and potash. I n four of the potash runs (3, 7 , 11, and 13) no chlorides mere used. The per cent K20 volatilized was 43.2, 54.1,
INDUSTRIAL A N D ENGINEERING CHEMISTRY
86
Vol. 23, No. 1
Table I-Operating Data of Experimental Furnace Run ........................... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 ~. 15 Ratio K,O/C charged.. . . . . . . . . . . 0.162 0.162 0.162 0.306 0.110 0.110 0.110 0.191 0.218 0.218 0.245 0.-2!20 0.220 Ratio PaOs/C charged.. . . . . . . . . . . O:ZQ3 0:328 . . . . . . . . . . . . 0.173 0.173 0.173 0.264 0.187 0.187 0.370 0.268 0.268 Mol ratio CIz/KzO.. . . . . . . . . . . . . . . . . ... 0 0.60 1.20 1.10 0.00 0.73 0 . 1 4 5 0 . 8 4 0.00 0.77 0.00 1.41 2.62 810 442 324 cold Blast temp., C . . . . . . . . . . . . . . . . . 654 704 520 650 522 717 585 557 648 552 684 Top temp., C . . . . . . . . . . . . . . . . . . . . 210 248 281 287 346 343 174 284 264 Rate of combustion, grams/second carbon.. ...................... 4.13 3.94 6.51 9.04 9.74 7.96 6.845 6.809 7.437 6.908 6.100 6.202 6.307 6.450 6.278 Heat loss cal./gram carbon,.. . . . . 442 534 286 311 677 616 562 701 505 976 907 843 950 975 Grams &/gram carbon.. . . . . . . . . . . . ... 2.68 2.43 2.19 3.37 1.83 1.86 1.89 2.52 2.56 2.62 3.19 2.73 2.83 PiOr in slag, Yo... . . . . . . . . . . . . . . . 2.80 5.51 5.07 4.92 6.37 3.51 2.70 3.44 6.75 3.94 6.28 P30, reduced, %. . . . . . . . . . . . . . . . . 90.5 93.5 . . . . . . . . . . . . 46.42 49.43 30.68 66.58 63.70 52.65 38.74 60.40 46.09 KrO in slag, %. . . . . . . . . . . . . . . . . . . . . 3.45 1.20 3.36 3.24 2.77 2.47 1.59 0.77 4 80 3.38 4.49 1 5 4 0.35 KaO volatilized, %.. . . . . . . . . . . . . . . . . 43.2 82.8 54.9 64.4 54.14 58.42 72.90 89.85 44:OO 59.64 42.68 81:03 95.35
... ...
...
.........
............
... ...
44.0, and 42.6 for an average of 45.9 per cent. I n the two experiments given in Pike's paper no chlorides were used and he reports volatilizations of 41.0 and 46.9 per cent This result would be more or less expected from the results announced by Madorsky (3). The theory that the addition of lime to potash silicates will cause the volatilization of KzO seems in the actual case to be slightly less than half true (44 per cent). I n experiments 3, 4, 5, and 6 sodium chloride
The last five runs in Table I show in a somewhat remarkable way the influence of calcium chloride on the potash volatiliza-
tion. All five runs were made with burnt lime, phosphate rock present, and variable amounts of calcium chloride added. With no chloride (runs 11 and 13) the potash yield was 44.00 and 42.58 With 0.77 molecular equivalent of chloride added the potash recovery is 59.64 (test 12) with twice that amount (test 14) it is 81.03 and with 2.62 molecular equivalent i t is 95.35. These results are plotted in Figure 4 with potash yield as ordinate and molecular equivalent of chloride as abscissa; the heavy line connects the points. I n the same figure are similar data taken from the paper by Madorsky (represented by the broken curve). The general shape of -the two curves is surprisingly similar, when the number of variables that were known not to be constant are considered. The percentage of PZOSreduced in these five last runs are quite low. Little significance should be attached to the amount of Pzos reduced in per cent of that charged. The actual number of grams of Pz05reduced per gram of fuel carbon is the important figure, which is readily arrived a t by multiplying the charge by the per cent reduced. I n operating the furnace one estimates ahead of time how much hearth heat will be available for reduction and then puts in somewhat more PzO5 to be sure of using up all the available heat. The low percentage recovery then means, if anything, that the writers were Door a t messinp. The analysks of t h i r a w miterials and of the slag are given in Table 11. Table 11-Analyses
of R a w Materials a n d S l a g
CONSTITUENTWYOMINGITE CaO Si02 AliOi MgO K20 P20J FeO FeaO; TiOa NazO
PHOSPHATE ROCK
TYPICAL SLAG
%
%
%
5.07 51.34 11.68 6.92 12.76 1.97
46.50 6.90 2.13 0:26 31.63
35.13 34.16 10.2? 4.89 0.77 3.51 2.86
4:89 2.12 1.42
0:+2
1:91 1:69 5.95
Economic Considerations Figure 3-Large
Furnace under Construction a n d S m a l l Furnace in Operation
was used and no phosphate rock. The use of rock salt is not of great practical interest to us, in that calcium chloride has been found to be much better and is more readily obtainable. It might be pointed out that run 5 on cold blast with more than one molecular equivalent of sodium chloride gives only 54 per cent volatilization in comparison with 82.8 per cent in run 4 and 64.4 per cent in run 6. Although extreme blast temperatures do not seem necessary for potash volatilization, unh'eated blast seems to show worse results. It is perhaps to be expected that the results obtained with a furnace too small properly to calcine limestone will be rather variable when limestone is in the charge. The results are instructive, however, and the runs 4 and 10, which show potash yields of 82.8 and 89.7 per cent, show that substantial volatilization can obtain with rock salt and unburned limestone.
As mentioned above, the process of producing potash and phosphoric acid is applicable to the Green River Valley of Wyoming, where phosphate rock and Wyomingite can be brought together a t the same furnace with a reasonable freight haul, Unfortunately no source of metallurgical coke appear8 locally available, nor can coke be shipped into the Green River district except a t prohibitive cost. Natural gas and non-coking bituminous coal are abundant and cheap a t the potash deposits and a priori there is no good reason for believing that this coal will not serve as furnace fuel. A number of runs on the experimental furnace were made with Rock Springs coal as a fuel, and the results are encouraging. It would not be wise a t this time to state that the coal has been shown to be in every way satisfactory, because of the limited amount of coal that was shipped from Wyoming to Washington. On the basis of the evidence in hand, however, i t is not anticipating the expected results too far to indicate
ISDUSTRIA L A S D E.VGISEERISG CHEMISTRY
January, 1931
87
coke would be operated with the same charge, recovery, fuel consumption, and heat losses as are shown in experiment 14 of Table I. The Rock Springs coal has 47.3 per cent fixed carbon, and 1.83 tons of this coal would be equivalent to 1 ton of 87 per cent coke. The operating costs per day would be:
E a
Coal, 1370 tons at $ 0 . 8 7 . . . . . . . . . . . . . . . . . . . . . . . . . $1192 Wyomingite, 1637 tons at $0.90. . . . . . . . . . . . . . . . . . . 1473 Phosphate rock, 556 tons at $ 3 . 2 5 . .. . . . . . . . . . . . 1807 Limestone. 557 tons at $ 2 . 7 5 . . . . . . . . . . . . . . . . . . . . . . . . 1.532 ~ Labor, per d a y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600 Daily investment charge on $1,250,000 at 21 per cent per annum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 719
.
-
l
I
01
I
Figure 4-Effect
2 MOL RATIO, C ~ C I ,
I
Phosphate credit, 106 tons PzOj at 9 4 0 . , . . . . . . . . . . . . . . . . . 4240 Daily cost chargeable to potash.. . . . . . . . . . . . . . . . . . . . . . . 83083 Cost of Kz0 per ton, = $18.25 $169
3
rO n,o of C a l c i u m Chloride on P o t a s h Volatilization
briefly what the costs would be in a commercial furnace project using Wyomingite, Idaho rock, Rock Springs coal, burnt lime, and calcium chloride. The phosphorus from the precipitator would be burned in air and the resulting phosphoric acid caught in a second precipitator. The potassium chloride from the first precipitation would be used to produce potassium phosphate with the generation of hydrochloric acid. The hydrochloric acid reacting on limestone would regenerate calcium chloride for use in the furnace. The product shipped would be potassium phosphate. I n calculating the costs of this furnace project, the assumption will be made that a large furnace burning 750 tons of
$3083
It should be remembered that the above cost is calculated for a commercial blast furnace supposed to operate with a thermal efficiency no greater than that realized in the small experimental furnace. I n actual fact a considerably lower cost is to be expected. Literature Cited Hachenbleikner, U . S. Patent 1,103,910 (1914). Haff, U. S. Patents 789,438; 789,439; 789,440 (1905). Madorsky, IND. ENG.CHEM.,as, 78 (1931). Pike, I b i d . , 22, 243, 344 (1930). Ross, Mehring, and Jones, I b i d . , 16, 563 (1924). ( 6 ) Wilson and Haff, U . S. Patent 1,018,186 (1912).
(1) (2) (3) (4) (5)
Studies in the Development of Dakota Lignite IV-Critical Oxidation Temperature of Lignite1,z W. C. Eaton, G. A. Brady, A. W. Gauger,s Irvin Lavine,‘ and C. A. Manns UNIVERSITY OF NORTH DAKOTA, GRANDFORKS,N. D A X .A~N D UNIVERSITY OF
hfIKSESOTA, MINXEAPOLIS, h I l N X ,
The methods of Wheeler and Parr for studying the taneous ignition which i s a t HE School of Mines of ignition temperature of coals have been adapted for a least sufficiently accurate for the U n i v e r s i t y of study of North Dakota lignite. purposes of comparison of North Dakota is enThere appears to be no material difference in the processed with unprocessed gaged in a study of the propcarbon dioxide index and the critical oxidation temlignite. erties and treatment of lignite peratures of lignite from different localities in North This experimental work for the purpose of aiding in Dakota. was begun a t the University thedevelopment of the vast The carbon dioxide index and critical oxidation temof North Dakota. Through deposits with which the westperatures are affected by moisture content of the a cooperative arrangement era half of the state is undersample, size of particles, rate of gas supply, and prewith the Division of Chemical laid, The original lignite a;; vious history of the lignite. Engineering, School of Chemmined contains upwards of Drying of lignite by the Fleissner process does not istry, University of Minne30 per cent water and is not materially affect the carbon dioxide index and critical sota, it became feasible to particularly liable to sponoxidation temperature beyond the effect due to decontinue the study of the taneous combustion. Upon exposure to atmospheric drycreased moisture content. problem simultaneously a t both u n i v e r s i t i e s . This ing the material slacks badly and the residue frequently ignites spontaneously with ease. It paper sums up the results of the experiments carried out is obvious that any treatment for the purpose of increasing the during the past year a t both institutions. heating value must not do so a t the expense of an increased tendThe problem of spontaneous combustion of coal while in ency towards spontaneous combustion. It becomes impor- storage, transit, or in the mine has engaged the attention tant to develop a method of estimating this liability to spon- of scientists for a t least seventy-five years. The general method of procedure has been to heat the coal in a furnace 1 Received August 11, 1930. Presented before the Division of Gas and
T
Fuel Chemistry a t the 80th Meeting of the American Chemical Society, Cincinnati, Ohio, September 8 t o 12, 1930. Abstracted from theses presented by W. C. Eaton t o the University of Minnesota and G. A. Brady t o the University of North Dakota in partial fulfilment of the requirements for the degree of master of science in chemical engineering. f
.
Director, Division of Mines and Mining Experiments, University of North Dakota. 4 Assistant professor of Chemical Engineering, University of North Dakota. * Chief, Division of Chemical Engineering, University of Minnesota.
~
