342
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
Vol. 9, No. 4
ORIGINAL PAPERS FERRO-URANIUM’
been widely advertised as “the last word in high-speed tool steel,” and it has also been advertised t h a t “ferroUranium steel apparently dates from about 1897, uranium used in high-speed steels greatly increases when i t was stated2 t h a t t h e French Government was strength, toughness and durability, producing a steel trying t o make use of uranium steel in guns. Merck’s t h a t will stand up on t h e job.” 1907 index (p. 4 5 2 ) states t h a t the only technical However, t h e reports of Hoffman and Johnson’ use for uranium is in t h e form of an alloy in t h e were not so favorable, the former stating t h a t a uramanufacture of gun barrels, nium steel with 5 per cent W a n d 3 t o 4 per cent Cr Escard3 states t h a t it is reported t h a t Krupp uses made a very good tool and did good work for say two uranium steel in armor plate. There have been various grindings, b u t after t h a t did not hold its efficiency rumors, naturally not capable of direct proof, t h a t and had t o be rehardened, and t h e latter, t h a t a 40Germany is using uranium steel liners in big guns, point carbon steel with 0.3 per cent U was disappointin t h e present war. ing, being red short a t ordinary forging heat and alFischer4 says, “a German firm is putting on the together uninteresting from a practical point of view. market ferro-uranium t o be used in the manufacture It is also understood t h a t in most attempts it has of steel. England is also interested in uranium steel been found very difficult t o produce uranium higha n d as a permanent supply of ferro-uranium is guar- speed steel free from streaks or seams. anteed by Messrs. Geo. G. Blackwell Sons and Co., It will be necessary t o have more definite data t h a n Ltd., of Liverpool, trials on a large scale will be made have yet been published before t h e real value or lack in t h e steel industry. The properties of uranium and of value of uranium in steel can be determined. All tungsten steels are similar. Fischers says, ‘‘some of t h a t can be said a t present is t h a t uranium deserves t h e large (American) steel companies have tried t o use a careful trial both in tool steel and in ordnance, though uranium in their line of work, but with little or no suc- t h e former is probably t h e more promising field. cess.” Although the production of uranium steels was Tourchinsky6 includes without further comment attempted ten years ago, t h e literature is singularly 0.23 per cent in t h e composition of steels made a t the barren of really definite information on them, and is Sonoritz works in 1913. even more meager in regard t o t h e manufacture of Commercial use of uranium steel in the United ferro-uranium. States is quite recent. Keeney’ states t h a t in 1915 The electrolysis* of fused mixtures of CaO and t h e t h e Standard Chemical Company worked on t h e use oxide of the metal t o be produced has been suggested of uranium in steel and put ferro-uranium on the as a method of making ferro-alloys. I n a private market. He says, ‘‘although t h e applications of ferro- communication, Mr. Beckman has outlined his prouranium have not been completely solved, t h e results cess for making ferro-uranium, in which equimolecular are encouraging and indicate t h a t in high-speed steel proportions of CaO and rather impure Us08 were a small percentage of uranium may be substituted for fused in a magnesite-lined furnace and. electrolyzed a very large percentage of tungsten without injuring with an iron cathode and carbon anode. The furt h e cutting qualities of t h e steel. A high-speed steel nace was not arranged t o pour or t a p , and t h e product showing excellent cutting qualities contained C 0.78 was chiseled out after cooling. Mr. Beckman stated per cent, M n none, Si 0.16 per cent, P 0 . 0 2 per cent, t h a t he had thus made carbon-free ferros of 60 per W 8.15 per cent, Cr 3.62 per cent, V 1.81 per cent, cent U, and 9 per cent V with t h e balance mainly F e U 1.02 per cent.” Two or three other American and Si. He kindly sent t h e writers a small sample firms are contemplating t h e commercial production of whose analysis he did not have, which was made by ferro-uranium. the above process. Comparative tests of uranium steels whose comThe analysis3 of this sample, the composition of t h e position is not given, against other high-speed steels, ferro produced by the Standard Chemical Company, TABLEI also of unstated composition, have been given by t h e RATIO -PERCENTAGECOHPOSITION- TO 10% U Standard Chemical Company.8 Uranium steel has Ferro U C Si V A1 C Si 1 T o be read at the Kansas City Meeting of the American Chemical Electrolytic ....,... . . . . . 5 7 . 2 4 . 2 4 . 3 Traces 1 . 5 0.73 0.76 By H. W. GILLETTAND E. . I MACK
Society. Published by permission of the Director of the Bureau of Mines. 2 Dennis, L. M., “Uranium,” “Mineral Industry,” 6 (1897), 654. a J. Escard, “Sur les differents procedes de preparation de l‘uranium metallique pur ou a 1’ etat de fonte.” Rev. chim zndustrielle, 18 (1907). 81. 4 S. Fischer, “Uranium and Vanadium,” “Mineral Industry,” 22 (1913). 773. 6 S. Fischer, “The Carnotite Industry.” Trans. A m . Elecfrochem. SOC., a9 (i913), 374. K. Tourchinsky, “Nathusius Electric Furnace in the Steel and Tube Works at Sonoritz,“ Rev. Russ. SOC.of metallurgy, through Rev. de m e t , 12 (1915). extraits, p. 180. 7 R . M. Keeney, “Uranium and Vanadium,” “Mineral Industry,” 24 (1915). 706. See also J. M . Flannery, U. S. Patents 1,201,625; 1,201,626; 1,201.627, January 2. 1917. 8 Standard Chemical Co.. “Uranium in High-speed Steel,” Met. and Chcm. Eng.. 15 (1916), 160; Iron Age, 97 (1916), 952.
................
Keeney 50.0 Commercial No. 1 41.2 Commercial No. 2 ....... 4 0 . 0
.......
3.0 4.9 3.5
1.0 2.4 2.0 to 3.0
2.0 2.3
. ..
...
.... ..
0.60 0.20 1.12 0 . 5 8 0.88 0 . 5 0 to 0.75
as given by Keeney,4 and the analyses of two shipments of commercial ferro made in 1916 a n d reported by t h e purchasers, are given in Table I. The absolute percentage of carbon and silicon in t h e ferro is not SO 1 Hoffman and C. M . Johnson, in discussion, “Symposium on Electric JOURNAL, 8 (1916). 949; Met. and Chem. Eng., 15 (1916). 448. Steel,” THIS 2 J. W. Beckman, “An Electrolytic Furnace Method for Producing Metals,” U. S. Pat. 973.336, Trans. .4m. Electrochem. Soc.. 19 (1911), p 1 7 1 . a B y E. L. Mack. 4 R . M. Keeney, LOG. cil.
Apr., 1917
T H E JOURNAL O F INDUSTRIAL A N D ENGINEERING CHEMISTRY
important as their ratio t o t h e uranium content, so, for comparison, t h e percentages of these impurities for each I O per cent U have been calculated and inclu de d. Johnson1 states t h a t he has encountered so-called ferro-uranium containing I 5 t o 2 0 per cent aluminum, and t h a t vanadium was always present, from 2 t o 3 per cent up t o 28 per cent; he states also t h a t one ferro-uranium analyzed b y him contained 15 per cent silicon. One steel company, according t o a private communication, has made ferro-uranium on a n experimental scale, in a n Acheson graphite crucible, under a n indirect arc, using silicon or ferrosilicon as reducing agent, a n d has produced ferro of 15 t o 8 j per cent U, with carbon averaging about 4 . j per cent in all lots. Quite a little silicon was also left in t h e ferro. On t h e 8 j per cent ferro, this would give 0.53 per cent C for each I O per cent U. The price of American ferro-uranium in February 1917 was $7.50 per lb. of contained U, i. e . , $3.75 per lb. for a 5 0 per cent ferro. A German product was quoted2 in 1914 at 450 Marks per kilo for a ferro of about 50 per cent U. ,4 recent quotation3 on uranium oxide is $3.60 per lb. of 96 per cent UaOs, with special prices on t o n lots, equivalent t o $4.40 per lb. of contained U at t h e price for small lots. This leaves a margin of $3.10 per lb. U t o cover loss of U in reduction, cost of iron, coke, flux, power, labor, interest, depreciation a n d profit in t h e manufacture of t h e ferro. Partly on account of t h e price, experiments on uranium steels seem t o have been confined t o those with a maximum of about I per cent U. While i t is probable t h a t uranium steels with such high percentages as t h e 1 2 t o 20 per cent W in some tungsten steels may not be commercially desirable, i t would be well t o know what t h e properties are of steels really high in U. I n present practice in t h e addition of ferro-uranium t o steel, to t h e U is lost. Experience will probably reduce this loss materially, b u t reports so far indicate t h a t with a ferro much below 40 per cent U t h e U is not readily taken u p a n d t h a t with a very high U ferro-say 9 j per cent, or practically a crude metallic U, t h e metal burns u p so rapidly t h a t much is lost before i t can get into t h e steel. I t seems probable t h a t 45 t o 6 5 per cent U will be about t h e proper percentage. Those who have attempted t o use ferrouranium find t h a t it must be added just before pouring or during pouring, as, if added a n y length of time before t h e steel is cast, no U is found in t h e steel, possibly because of reaction with slag as well as oxidation. The steel should be very hot. The C a n d Si ratios in t h e ferros whose compositions have been given are so high t h a t , with t h e to 1 / * loss of U,these ferros, even if added t o a carbonfree iron, would reach t h e usual limit of carbon for a tool steel when about j per cent U remains in t h e 1 C. ?iI Johnson, “Chemical Analysis of Special Steels,” 1914 ed., pp 288, 299. * De Haen’s price list, May, 1914. 8 Foote Mineral Co. Mineral Foote-notes, January 10, 1917.
343
steel, and in most cases t h e silicon would also be over t h e usual limit. On t h e other hand, if experiments prove t h a t only very small amounts of U are desirable, ferros of t h e composition given, or those even higher in carbon, could be used. But until a purer ferro is produced, experiments on steels really high in U a n d of normal C a n d Si content will be impossible. The uranium oxide used in t h e work described below was produced b y t h e National Radium Institute. It was mainly U 0 2 , with some U303. This runs about 83 per cent metallic U. The oxide contained about 2 per cent F e z 0 3 ,0.1 per cent A1203, 0.1 t o 0.25 per cent V 2 0 s , 0 . 2 0 t o 0.35 per cent SiOz, 0.3 per cent moisture, 1.3 per cent NaCl, 0 . I j per cent carbon. Of these impurities, for use in a ferro-uranium made b y reduction with carbon, only t h e A1203, PZOSand Si02 can introduce impurities into t h e ferro. On t h e basis of metallic U V Si Al, t h e total impurities t h a t can go into t h e ferro are less t h a n 0.5 per cent. Beside t h e greater purity of this UOZ t h a n most commercial ‘CT303, t h e lower oxide is advantageous in t h a t t h e first stage of t h e reduction, from Us03 t o UO2, has already been accomplished, hence t h e further reduction, from U O z t o U, v;iii not require as much energy as would be t h e case when u308 is used. I n order t h a t t h e other materials used might approach this standard, a pure ingot iron, a n d a low-ash coke were used. U O n is not reduced by carbon below 1500’ C.,l a n d according t o temperature measurements on t h e surface of t h e slag a t t h e end of successful runs, a temperature of at least 1700’ is required for efficient reduction, requiring a n electric furnace. UOZ has a specific gravity of 1 0 . 2 , t h a t of iron is a little under 8, a n d t h a t of metallic uranium is about 18.7. Hence if one melts U 0 2 , carbon and iron together without a slag, t h e iron will s t a y on t o p a n d will not collect t h e uranium. So it is essential t o have some flux present which will combine with UOZ t o form a slag lighter t h a n iron. This slag should also be a good arc-supporter, in order t o use t h e direct arc type of furnace, since t h e bulk of t h e reduction seems t o go on directly under t h e arc itself. The slag must not introduce undesirable impurities into t h e ferro. T h e furnace must be provided with a lining t h a t will not be strongly attacked by t h e slag or introduce undesirable impurities, a n d which will stand up at t h e high temperature needed. Attempts t o produce t h e ferro in a n indirect arc type of furnace soon showed t h a t there would be an excessive power consumption t o get t h e required temperature, a n d t h e direct arc type was then taken as t h e most promising. The Rennerfelt type, where t h e arc is deflected onto t h e charge, might serve, b u t was not tried, as only single-phase power was available, and there was no difficulty in getting slags t h a t were good arc supporters for t h e direct arc type. I n preliminary experiments i t was found t h a t made a very fluid slag with UOs, and as i t was a t first
+ +
+
1 H. C. Greenwood, “Reduction of Refractory Oxides by Carbon.” J . Chern. SOC..98 (1908). 1483.
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
344
thought t h a t A1 would not be reduced rapidly enough t o produce a very impure ferro, this was tried. I n order first t o produce a ferro sufficiently high in U without regard t o carbon content, experiments were made with both Girod and Hkroult type furnaces with carbon or graphite hearths. Some of the first of these were not tilting or tapping and t h e product was taken out when cold, b u t it was soon found t h a t even for preliminary work, a tilting furnace was desirable, since, by tilting t h e furnace t o and fro slightly during the run, a fresh charge can be brought directly under t h e arc and far better results obtained.
Vol. 9 , No. 4
over t h e graphite hearth, in t h e hope t h a t it might keep t h e ferro away from t h e graphite, but while some of this layer remained unfused, t h e ferro would break through and touch t h e graphite, and high carbon ferros resulted, giving a ratio of from 0.9 t o 1.3 per cent C for each I O per cent U,save in Expt. 49, where, after t h e furnace was cold, a small separate regulus was found t h a t had been kept out of contact with carbon by the UO? layer. This piece, 49 B, analyzed jo per cent U, 0.62 per cent C , or 0.12 per cent C per I O per cent U. It is evident t h a t ferro-uranium can be readily pro-
TABLE 11-RUNS Expt. No. 39 40 41 44
45
Lbs. Fe 3.0 4.0 4.0 3.5 3.5
Lbs. Slag-Formers Lbs. Old Charged with Pe Slag Ala08 CaO UOz Charged 1.0 0.25 1.0
. ........... I
0.5
0.3
0.5
.
IN TILTINGFURNACES, GRAPHITE H E A R T H - A ~ O I IN S b A G Experiments 47, 48. 49, 50 and 65: Layer of UOz on Hearth Furnace Lbs. PERCENTAGE ANALYSIS Lbs. Lbs. Lbs. Time at Kw. h. Ferro OF PRODUCT Si02 UOI Coke Hrs. Min. Start Used Poured U C Si AI . . . 4 . 0 1 . 2 5 1 8 Cold 51 4.75 47.5 4 . 2 0 . 3 Not det. ... 5 . 2 5 1 . 5 ( b ) 1 12 Cold 6 0 . 5 8.1 42.5 4 . 5 0 . 3 Not det.
i : O (No.43)
...
2 . 0 (No. 43) 0 . 5
1
6.0 6.5 6.5
1.25 1 1.5 1 O.Z5(c)
3 0 48
1.5
Cold Cold Hot
58.5 59 30
n/n
l6:i"' 6.0 7.0
45.0 4.2 52.0 5.8 39.0 5 . 2
0 . 4 Not det. 0 . 5 N o t det. 1.6 2.0
Ratio to 10% C 0.89 0.06 1.06 0.07
gi
0.94 0.10 1.12 0 . 1 0 1.33 0.41
....
8.5 55.0 5.9 . . . . . . 1.07 6.6 4 5 . 0 4.85 . . . . . . 1.08 3.0 29.5 2 . 5 5 0.87 . . . . l . 5 V ) 5 0 . 0 0 . 6 2 ( i ) 013 1:: 0.12 0.06 50 4.0 1.0 10.6 0.9(d) 1 Cold 0.5(g) 5 2 . 0 5 . 3 . . . . . . 1.04 . . . . 4.0(h) 6 0 . 0 5 . 3 0.4 3.4 0.90 0.17 1 40 . , . ( e ) 50.0 3.2(j) 3.6 0.64 .... 8.0 38.0 5.8 1.0 ... 1.53 0 . 2 6 65 4.0 3.0(a) 8.75 0.6(d) 1 12 Cold 50.75 ( a ) Plus 1.0 CaF2. ( b ) Charcoal. (c) For reduction of S O ? . (d) Coke at end, after adding UOz not mixed with it. ( e ) Total. cf) Taken from furnace when cold. (9) Sample poured 3 min. after coke was charged. ( h ) Second pour. ( i ) This piece not in contact with graphite. ( j ) This piece probably not in contact with graphite.
47 48 49
'
4.0 4.0 4.0
0.9
6.4
8.0 3.75
3.5
1.5 0.7 0.5
1 1
Hence a small tilting, single-phase HCroult type furnace, with a graphite hearth, was built. This furnace, and those later used, all took 600 t o 7 j o amperes a t 60 t o go volts; with 90 t o g j per cent power factor, i. e., some 30 t o 60 kw. It was regulated either by t h e length of the arcs or by regulating t h e voltage. The iron was charged into the furnace, slag formers added, t h e furnace heated till iron and slag were fluid, then the charge proper ( U 0 2 mixed with coke) added slowly a n d t h e furnace then heated 2 0 t o 30 minutes more before pouring the ferro. Results in this furnace are shown in Table 11.
0 10 32
Cold Cold Cold
40 40.5 36 Sl(e) 45.5 67.5
....
...
duced on a carbon hearth with very little-loss of U, b u t t h a t there will be 4 t o j per cent carbon in t h e ferro, and t h a t without water-cooling of the hearth a layer of U 0 2 cannot be satisfactorily preserved. A magnesite hearth was tried in the same furnace shell as was used with t h e graphite hearth, with t h e A1203 slag and with excess carbon in the charge and ferros of 33 t o 40 per cent U, 4 t o j per cent C,produced, but the magnesite hearth was ruined in one t o three heats. A zirkite hearth was ruined in one heat. Small scale tests of silica and chromite were made, but neither stood up and they introduced, respectively,
TABLE111-RUNS IN STATIONARY WATER-COOLED FURNACE Experiments 54. 56, 57 and 58 on Magnesite Hearth. Experiments 61, 62, 63 and 64 on SIC Hearth Lbs. FerroLbs. Ferro Silicon U-free in PERCENT ANALYSIS Ratio to Sep- RegOF REGULUS (50% Time Lbs. Lbs. Oid Slag in Hearth Lbs. Lbs. Lbs. Lbs. Lbs. Expt. T; C Si A1 &oo70~ Si) Hrs. Min. Kw. h. arable ulus Coke CaCz Lbs. Fe CaFz A1203 CaO UOz NO. 54 2 8 . 0 ( a ) ( S o s . 45, 49and 50) 4 . 0 . . . . . . . 1 . 0 0.5 l.O(e) . . , 1 20 4 2 . 7 5 2 . 0 3.5 4 6 . 0 1 . 3 1 . 3 Trace 0 . 2 8 0 . 2 8 56 1 8 . 0 (No. 55) 3.0 .... 0.4 . . . . 3 . 0 O.iS(fl ... 1 37 Much 1 . 4 2 8 . 0 3 . 4 1 . 8 4 . 9 1 . 2 1 0 . 6 5 4 . 0 1 . 2 5 . . . . . . . . . . . 0.6(e) . . . . . . . 1 is 40.75 2 . 2 5 2 . 2 5 3 3 . 0 2 . 7 1 . 4 0.7 0 . 8 2 0 . 4 2 57 2 3 . 0 (Nos: 55 & 56) 6.0 1.6 . . . . . . . . . . . . . . . . . 4.0 O.5(e) . . . . . . . 53 40 58 8 . 0 (No. 57) l.'iS(b) lO.O(c) 0 . 6 ( e ) 1 18 40 4.5 45.0 3.7 0.3 0 . 8 2 0.07 61 1.5 1 6.5(d) 62 1 1 . 0 (No. 61) ll.O(c) None 2.O(g) 42 6 . 2 5 12.0 3 . 5 0 . 7 2.92 0.58 63 25.0 (No. 62) 3 . 9 . . . . . . . . . . . . . . . Sone 0 . 9 0.5 1 40.5 5 . 2 5 9 . 0 2 . 5 4 . 5 ... 2 . 7 8 5 . 0 6 4 1 9 . 0 (No. 63) 3.0 . . . . . . . . . . . . . . . 0.5 0.75 1 40 4.0 35.0 3.9 1.4 1.12 0 . 4 0 (a) Plus 1 lb. fine magnesite. ( b ) Also 0.35 Bz03. ( c ) In bottom. ( d ) In charge. ( e ) Alone. V) Mixe h UOz. ( 9 ) Also 0.3 NazC03.
.......
....
...
...
I n Expts. 39-41, 16.75 lbs. UOZ were charged, and 24.65 lbs. ferro averaging 44.j per cent U were obtained. The U 0 2 charged was equivalent t o about 14 lbs. metallic U a n d 1 1 lbs. metallic U were obtained in t h e ferro, or nearly an 80 per cent recovery, not counting t h e U in t h e metal and slag (about 3 lbs.) left in t h e furnace which bring t h e loss down t o about I O per cent. I n other words, t h e loss of uranium is very small. There is a slight orange sublimate, apparently U 0 3 , given off in small amounts, some of which condenses on the electrodes. I n Expts. 47-50 inclusive, a layer of U 0 2 was fused
.. .. ..
.... ...
...
... ...
large amounts of Si and Cr. Hence, a magnesite hearth, and later a carborundum hearth, water-cooled in order t o maintain a layer of frozen U02 or slag over it, was tried in a stationary furnace. The product was dug out when the furnace was cold. This was usually in two parts, one which had not gotten under t h e arcs, as t h e furnace could not be rocked back and forth t o stir t h e charge, and which had only a trace of U, t h e other the regulus, in the center. The results are given in Table 111. The A1203 in t h e slag introduced A1 into the ferro. This may be more of an apparent t h a n a real impurity,
Apr.7
T H E J O U R N A L OF I N D U S T R I A L A N D E d G I N E E R I N G C H E M I S T R Y
1917
shell 18 in. X 18 in. X 14l/? in. high in t h e body portion, a n d with a hearth portion 11 in. X 11 in. X 4 l / 2 in. deep extending downward from t h e bottom of t h e body portion, was made up, mounted on trunnions, provided with pouring spout, a n d t h e hearth portion arranged for water-cooling b y being surrounded with a perforated spray pipe, t h e holes in which were a t such an angle t h a t t h e streams hit t h e main or upper bottom just outside t h e junction with t h e sides of t h e hearth portion, causing sheets of water t o cover t h e junction and flow down t h e sides. A pan 1 5 in. X I j in. X 3 in. deep suspended by corner posts extending down from t h e main bottom was hung with its bottom I in. below t h e hearth bottom. -4 hole was cut in t h e center of t h e bottom of t h e p a n of such size t h a t i t did not drain t h e cooling water all off, b u t some ran over the edges. This covers t h e bottom and t h e lower z in. of t h e
as A1203a n d uranium oxides form a fluid slag, a n d as small amounts of A1 might be expected t o be oxidized before t h e bulk of t h e U is, t h e A1 might even t o some extent protect t h e ferro from loss of U a n d t h e A1203 be eliminated as a fluid slag instead of held as infusible ALO3 inclusions, as when A1 is used alone. I t is possible t h a t a Fe-U-A1 alloy t h a t would give a fluid slag on oxidation might be found t o reduce t h e trouble from streaks a n d seams in uranium high-speed steel or a t least prove a good deoxidizer and scavenger for steel. However, t h e present purpose is t o produce a ferro-uranium as free as possible from all impurities. Experiments were made with other slags: one of B 2 0 3 - C a 0 - U 0 2 was not satisfactory, b u t a CaO-CaF2-U02slag worked fairly well. For t h e preparation of pure uranium from t h e oxide, Kuzel a n d Wedekindl suggest t h e use of metallic calcium, while another patent2 suggests t h e use of CaC?
-
TABLE IV-TILTIXG FURNACE. WATER-COOLED HEARTH
-CHARGE PROPER IN POUNDSDisregarding Slag Left in Furnace Mill Old CaO CaFz Scale Slag UOr 5 .... 5 5 5 o:i b.i 5:o .... 2.5 5.2
Exp t . KO.Fe 66 4 67 4 68 4 69 4 70 4 73
... ... ... . . . . ... ... ... ... ... ... ....
... ... ... .... .... ...
74
...
75 76 77
4.0 4.0 4.0
78 79
4.0 4.0
80 81 82
4.0 4.0 4.0
0.2
. ., o:i
0.2
o:i o : z 0.3 0.2 0.1
0.3 0.2 0.1
4.0 4.0
0.1
...
0.6 0.6
87
4.0
...
1.4 2.7
... 4.0 ... . . . ... . . , .. . . . . ...
89 92 93 94 95
96 97
.
,
,
.
.. , ., . . . . ...
5.0
.. . . .. 0:2 ...
85 86
88 4 . 0
345
0.9 0.75 1.25 1.25 2.40
11.0 (No. 74) 4.5 (No. 74) 5 . 0 (No. 76) & 3 . 5 (No. 71)
... . . . .... .... . , 6 . 0 (No. 79) 7.25(No.79) . . . ....
,...
... ,...
.., .... ... .... ... ..._
5.5 5.5 5.5 6.0
1.SO 6 . 5 1.50 6.5
.7.0 .,.
9 . 0 (No.91) 9 . 0 (No. 91) 9 . 0 (No.91)
....
.. ....
,
,
0.63
3.0
0.85 0.8 0.8
10.0 12.0
0.8 0.8
6.0 5.0 8.75(No.79) 4.0
...
,,.
Furnace Time at Coke Hrs. Min. Start Kw. h. 0.9 1 10 Cold 53.5 0.8 45 Hot 33 0.7 1 10 Cold 53 0.6 40 Hot 28.25 0.6 32 Hot 18.25 1.0 1 10 Cold 47.5 0.63 5.2 1.0 40 Hot 28.25
0.8 0.8 0.8
1 1
I
Ratio to 10% u C Si 0.42 0.17 0 . 5 1 0.23 0 . 5 8 0.80 0.47 0.54 1.64 0 . 9 8 0.25 0.55
5.9
60.0
2.4
0.40
0.40
0.40 0.35
00.49 .37
2.4
5 40 0
Cold 50.0 Hot 3 0 . 0 Cold 46.75
8.9 6.5 6.8
51.5 48.0 45.5
2.05 2 . 5 1.65 1.75 4.0 2.0
0.44
0.88
40 40
Hot Hot
27.521. I >
6.9 3.0
46.0 33.0
2.1 2.6
1.8 1.4
0.46 0.79
0.39 0.43
48.75 25.0 23.5
5.0(u) 5.2 67.0 1.1 6.8 53.5 2 . 4 8.0 45.0 3 . 6 2.0(0)
0.9 1.2 1.2
0.17 0.14 0.45 0.23 0 . 8 0 0.27
2 40 32
Cold Hot Hot
9.0 0.8 10.0 0 . 8
1
10 Cold 50 Hot
13.0 1 . 0 10.0 0 . 8
1
10
1
Lbs. Percentage Analysis Ferro of Product C Si Poured U 4.0 5 8 . 0 2.4 1.0 7.3 44.5 2.25 1.0 3.0 6.5 37.5 2.1 5.4 37.5 1.75 2 . 0 2.5 6.5 25.5 4.2 3.8 5.8 70.0 1 . 7
7
50 33
5.2 14.9
64 3.25 68.5 5.00
0.35 0.25
0.51 0.06 0 . i 3 0.04
4.2 6.6
65.5 59
1.5 1.3
0.45 0.60
0 . 2 3 0.07 0.22 0.10
32.5 9.2 50.5 9.5 38.25 12.5 30.75 10.6 45.00 3.2
56.0 63.5 63.0 72.0 58.5
1.2 2.15 1.75 2.4 3.65
0.65 0.65 0.90 0.60 1.20
0.22 0.12 0.34 0.10 0 . 2 8 0.14 0.32 0.08 0.62 0.21
35.50
63.0
2.15
0.75
43.0
2.15
0.70
Cold 50.5 Warm52
9.0 4.0 6.0 8.0 9.0
0.8 2.0 2.0 2.0 2.0
1 0 0 1
12.0
2.0
1
3
Hot
13.0
2.0
1
11
Warm41.25
42 Hot 10 Cold 53 Hot 55 Hot 15 Cold
(a) Taken out when furnace was cold.
or CaC? plus ferrosilicon. At arc temperatures CaCz will be decomposed into C a n d Ca, both of which should take p a r t in t h e reduction. According t o Expts. 62-64 (Table 111),coke appears a more effective reducing agent t h a n CaC2, with or without ferrosilicon. As t h e CaO.CaFz slag was fairly satisfactory a n d as t h e water-cooling preserved a frozen layer of slag over t h e lining of t h e stationary furnace, t h e next step was t o combine these in a tilting furnace. An iron 1 H. Kuzel and E. Wedekind, French Patent 419,043, Class VIII, 2. Application Oct. 15, 1909, granted Oct. l i , 1910, published Dec. 24, 1910. 2 Electric Furnace and Smelters, Ltd., London, German Patent 247,993. Class 40c. Group 12, Application Apr. 8, 1911, patented June 11. 1912.
7.5
i
t
REMARKS S i c hearth & sides. Little CaO & CaF? used or needed. Ferro touched S i c sides. Magnesite hearth, carbon sides in this and all subsequent runs. 5 lbs. mill scale 1 lb. coke charged before UO? and rest of coke.
+
No. 71 slag contaminated with Sic.
Too much slag in furnace, slag too stiff, magnesite spalled from roof.
* A . C. generator temporarily out of commission. Used D. C. a t 50 volts 800 amps. on these two. This voltage would hold only one arc, so one electrode alwaystouched slag or ferro. giving high carbon. 10 in. on each electrode used in the two heats. Fresh lining of U O L a O CaF? put in before No. 85 (not contaminated with S i c ) . Charge all added in s center from No. 85 on. Furnace cooled 21/4 hrs. between Expts. 87 and 88.
h'o. 91 slag contained some emulsified No. 91 ferro. Too much CaFz. Slag too fluid. Ferro spattered against the carbon sides, giving a high carbon mrtal. 0.34 0.12 h-os. 96 and 9 i : entire charge added together. 0.50 0.16 Ferronot hot enough to pour properly.
sides of t h e hearth portion continually with quite a volume of water, while the upper zl/'? in. of t h e sides are cooled b y sheets of water. The water falls from this pan into a large pan beneath t h e furnace, connected t o t h e sewer. A neater job could have been made b y completely enclosing t h e hearth position and providing inlet a n d outlet pipes, b u t t h e open system was used instead of a closed one for greater safety in case something should go wrong a n d t h e bottom cut through, and, in order t h a t t h e cooling might be watched. An apron on t h e front of t h e furnace below the pour-
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
346
ing spout (which is I ’ / ~ in. above t h e main bottom) prevents t h e water from splashing into t h e ladle or mold, when t h e furnace is tilted t o pour. A carborundum lining was first used, both in t h e hearth and on t h e sides, b u t the sides were attached by spattered slag, crumbled off, a n d contaminated t h e ferro with Si. H a d sides a s well as hearth been water-cooled, this lining would have served. Then split magnesite brick ( I I / ~in. thick) were used in t h e hearth, and carbon on the sides. For sometime t h e charge was added between the electrodes and t h e outside of furnace, with t h e idea of making sure t h a t t h e slag did not melt clear t o t h e magnesite. It was found, however, t h a t this was not necessary, and all charging was finally done in t h e center, between t h e arcs. Until a high CaF2 content of slag was maintained, and until center charging was begun, t h e slag built up in t h e sump and after a few heats t h e ferro was brought up so far t h a t it touched t h e carbon sides and gave high carbon ferros. T o prevent this, some of t h e slag had t o be removed from time t o time a n d added t o subsequent charges. After chiseling out some slag t h e rest was roughly shaped into a sump and the furnace heated without charge till t h e slag was melted, t o form a tight frozen lining. I n a run, the slag was first melted, t h e iron charged and melted, then t h e charge proper slowly added, and heating continued till t h e CO flame grew weak. The slag never poured out of t h e furnace with the ferro, but all remained in t h e furnace. The electrode consumption averaged I in. on each 2 in. diameter graphite electrode for each 26 kw. h. used, but much of this was due to oxidation in the air while cooling, so the electrodes were pulled out of t h e furnace while pouring, which would not be t h e case in a steadily-run commercial furnace. The results in t h e water-cooled tilting furnace are given in Table IV on page 345. Better results were obtained when t h e iron was charged in t h e metallic state, t h a n as oxide (mill scale). Including UOz for slag and lining, there was charged in Exnts. 8q-07 ,. (see . Table I V ) : I
Lbs. F e 2.1
U .......... Lbs. 125.7 ....................... .......................... . . . . .. ... .. .. .. .. ... ............ 9 . 7 ..........................
lbs. UOt equivalent t o . . lbs. CaFz.. 1bs.CaO lbs. Fe equivalent t o . . lbs. mill scale equivalent t o . . lbs. Fe for remelting equivalent to (see Table V) 16.2 lbs. coke
151.5 18.65 1.6 20.0 35.5 15 . O
135.4 There was obtained 107.45 1b3. F e equivalent to and 85 .O lbs. slaa to.. - eauiv. .
..................... ...
..
20:o 25.7 4.7
..
Lbs. U
52.5
63.4 60.0 48.7
- 123.4 -
123.4
Lbs. Fe 40.9
7.8 48.7
Loss, l b s . . .............. 1 2 . 0 3.8 Loss calculated oa 75.4 lbs. U actually used, 16.5 per cent; on 44 lbs.
Fe actually used, 8.5 per cent. Loss calculated on total U,9 per cent; on total Fe. 7.5 per cent.
Most of t h e loss was due t o fine particles carried out by t h e CO flame (this might be reduced by briquetting), some t o spatter, some t o volatilization of UOa or of a fluorine compound, and some t o mechanical loss in charging a n d pouring. Two of t h e thirteen heats included above were on remelting, and t h e loss therein makes t h e loss calculated above, too high, 1 5 per cent being probably nearer t h e true figure.
Vol. 9, No. 4
The average Si in t h e last 11 runs in Table IV (all others in t h a t table being contaminated by S i c ) was under 0.7 per cent. The average carbon in all 26 runs, good and bad, of Table IV was under 2 . 5 per cent. Excluding those where t h e ferro touched or spattered against t h e carbon sides, which would not occur if t h e furnace had water-cooled magnesite sides, t h e average of 2 1 runs is under 2 per cent carbon. T h a t of t h e three runs 87-89, in which t h e best conditions obtained, was 1.33 per cent C. The power consumption in 6 runs, 87-89 and 9294 was under 4 kw. h. per lb. for ferro averaging 63 per cent U, 1.8 per cent C, and on heats 89 and 94, with t h e furnace fully hot, about 3 kw. h . per lb., pouring about g lbs. per heat. I n a large commercial furnace, t h e power consumption should be much lower t h a n in the little experimental furnace. For a commercial furnace, assuming a capacity of zoo lbs. ferro per heat, 2 0 0 kw. (say I O O volts, 1 2 0 0 amperes per phase), on a three-phase tilting direct arc furnace of the HBroult type, would probably not be too much, as t h e best results on t h e ferro come when it is heated very hot, though such a furnace would be overpowered for steel. T o withstand the high temperature, t h e thin magnesite walls and bottom (say Z’/Z in. thick) should be strongly water-cooled so as t o maintain a solid UOz lining within them. The roof should be of carbon bricks on t h e inside, as t h e usual silica roof would drip and contaminate the ferro. T h e center of t h e roof should be open for charging. A three-phase direct arc furnace like the HBroult is preferable t o a single-phase like t h e Girod, and rather large carbon electrodes should be used, t o give as many and as large arcs as are practical, since there is little reduction outside t h e arc itself. Fluorspar alone is t h e best flux tried, if A1 is t o be kept out of t h e ferro. I t s amount should be so regulated t h a t the slag is not so stiff as t o hold ferro in emulsion, and not so fluid as t o spatter badly under the arc. The slag was good when 60 lbs. UOZ(excluding t h a t reduced and poured out asferro, b u t including golbs. used as original lining) and I O lbs. CaFz had been charged. Bleeckerl suggests first making ferro alloys high in carbon, by reduction with carbon, then crushing the ferro, mixing i t with iron oxide or oxide of the other metal in t h e ferro, and remelting, in order t o decarbonize. Or, he states, one may first make a high carbon alloy and decarbonize by later adding t o t h e molten ferro iron oxide or t h e oxide of t h e other metal. Remelting iron high in carbon with iron oxide t o refine it is of course steel-making routine, and R?oissan2 and Escard3 both long ago described decarbonizing metallic uranium high in carbon, by heating i t with uranium oxide, Table V gives some results on refining high carbon ferros. The finely crushed ferro emulsified badly with t h e slag, particularly in Expts. 83 a n d 84 when the slag was extra stiff because of adventitious magnesite. W. F. Bleecker, U.S. Pat. 1,094,114, Apr. 21, 1914. H. Moissan, trans. by V. Lenher, “The Electric Furnace,” 1904 ed., pp. 167, 170 8 J. Escard, LOC.cif. 1 2
T H E J O U R i V A L O F I Y D C ‘ S T R I A L A N D ENGINEERING C H E M I S T R Y
Apr.9 = 9 = 7
347
TABLEV-REMELTINGFERROSFOR REFINING Tilting Furnace, Water-cooled Hearth CHARGE PROPER IN POUNDS Disregarding Slag in Furnace Furnace ANALYSIS FERRO Mill Time at Expt. FBRRO U C Si Slag UOz Scale CaFz Hr. Min. Start NO. 7 1 .................. 6 . 0 (No.7 0 ) ( a ) 2 5 . 5 4 . 2 2 . 5 None 5.0 1 0 Cold 83 9 . 0 (No. 82)(a) 4 5 . 0 3 . 6 1 . 2 9 . 0 (No. 82) 5 . 0 0:;s 1 20 Cold 84 7 . 0 (No. 65)(a) 3 8 . 0 5 . 8 1 . 0 7 . 0 (hTo. 82) ... 1 . 7 5 0 . 7 5 50 Hot 90 7 . 0 (No. 89)(a) 5 6 . 0 1 . 2 0 . 6 5 . . . 3.0 .... 1.2 1 .. Cold 91 .................. 9 . 0 (No.86)(a) 6 8 . 5 5 . 0 0 . 2 5 . . . 6.0 . . . . 2.5 . . 50 Hot
.... ....
.................. .................. ..................
Expt.
No.
71..
.........
83... 83
RATIOSPER 10% U Original Rifined Ferro CFe’roSi C Si 1.64 0 . 9 8 0 . 2 5 1.28
........ 0.80
........... 1 . 5 3 90 ........... 0 . 2 2 91 ........... 0 . 7 3
0.27 0.26 0.12 0.04
0.30 0.52 0.22 0.27
0.29 0.30
0.12 0.11
--RECOVERY I n Ferro Charged U C Si 1.50 0.25 0.15 4.05 2.67 3.56 6.17
0.25 0.41 0.07
0.45
0.11
0.07 0.04 0.02
(POUNDS)In Ferro Recovered U C Si 1.25 0 . 0 3 0.16 2.70 0.07 1.90 0 . 1 0 0 . 9 7 0.024 0.126 4 3 ,. 36 02 o , 0 9
0.06 0.06
..
Lbs. Percentage Analysis Ferro of Product Kw.-H. Poured U C Si 49 3.5 36.5 0 . 9 4.7 55 4.0 56 1.7 1.6 31.25 4.0 47.5 2.45 1.4 49 2.25 43.0 0.95 1.S5 36 7.0 66.0 1 . 8 0.75 5.0(b)
REMARKS S i c hearth &walls, walls crumbled and contaminated Ferro
...---.
mith .Si
Magnesite hearth Magnesite hearth 0.035 Magnesite hearth‘ Increase in O.O53(c) 0.038(d) siliceous impurities in CaFz
si
due to
( a ) Crushed to pea size and smaller.
Plus any ferro left in furnace from preceding run. ( b ) Taken from furnace in regulus, when cold. (6) Poured. ( d ) Regulus.
I n all of these runs t h e recovery of t h e ferro charged was poor, much less being poured t h a n was charged, There was always some spatter, small globules of ferro being shot u p into t h e air above t h e hearth, oxidizing, a n d dropping back into t h e hearth, t h e Fez03t h u s produced tending t o decrease t h e percentage of U in t h e recovered ferro, as metallic U in t h e ferro will be oxidized b y Fez03. The experiments indicate t h a t i t is probably desirable t o produce a low carbon ferro in one operation, rather t h a n first t o make a high carbon ferro a n d t h e n t o refine it. Vanadium has not been determined in most of t h e ferros. It cannot average as high as 0 . 2 5 per cent, though individual ferros, made with a large amount of fresh UOz, will run a trifle higher t h a n t h e average because V is more readily reduced t h a n U, while one made from a charge consisting largely of old slag from which t h e V is already largely extracted will run lower t h a n t h e average. A few analyses of t h e ferros indicated t h a t prabably 0.4 per cent a n d certainly not over 0 . j per cent is t h e maximum. Traces of V were present in all. 9 s V is used in almost all tool steels, t h e very small amount of V t h a t would be introduced into steel by t h e ferro-uranium certainly would not be classed as a harmful impurity, though for experimental work on t h e value of U in steel it is desirable to have a ferrouranium as low as possible in V in order not t o have another variable t o contend with in t h e V. It appears t h a t b y using a pure UOz, a low-ash coke, and a pure iron as raw materials, with C a F z as slag former, a n d using a tilting direct arc t y p e furnace with water-cooled magnesite hearth and sides, i t should be possible t o produce commercially, without a second refining operation, ferro-uranium of any desired U content, say 40-70 per cent, with carbon averaging below 2.0 pei cent, silicon below 0.7 j per cent, vanadium below 0 . j per cent, a n d with aluminum, sulfur, phosphorus and manganese all so low as t o be negligible. If experiments with such a ferro show t h a t uranium steels high in uranium are not valuable, b u t t h a t a little uranium is useful, a n d if t h e amount required is so low t h a t t h e carbon introduced b y a high carbon ferro is harmless, then t h e furnace might have a n uncooled carbon hearth, a n d t h e ferros would contain 4 - j per cent carbon, If uranium is found useful only as a deoxidizer or
scavenger of oxygen and nitrogen, aluminum would not be harmful a n d might be advantageous, a n d t h e slag former might be wholly or in p a r t A1203. Grateful acknowledgment is made t o t h e Department of Chemistry a t Cornel1 University for t h e use, under a cooperative agreement, of its laboratory facilities, which are particularly well adapted t o a problem of this nature. ITHACA, NEWYORK
INFLAMMABILITY OF CARBONACEOUS DUSTS IN ATMOSPHERES OF LOW OXYGEN CONTENT By H. H. BROWNAND J. K. CLEMENT Received November 20, 1916 INTRODUCTION
As was stated in a previous paper,’ if dust could be entirely confined within t h e machinery of a mill in which combustible dust is produced, a n d a method could be found for preventing explosions in these machines, a long step would be taken in t h e prevention of dust explosions in mills. T o keep a dust cloud from forming in t h e machines appears t o be almost, if not utterly, impossible. It is possible, however, by proper cleaning t o remove foreign material from t h e grain and t h u s lessen t h e possibility of a spark being formed in t h e machine which might ignite t h e dust. But cleaners a n d separators do not always take out all t h e foreign material, so t h a t even under t h e best conditions foreign materials may get into t h e machines, or other conditions develop which might cause sparks, or other sources of heat, t o be formed within t h e machine, thus creating a very dangerous condition. However, if there were present within t h e machine a n a t mosphere which would not support combustion, t h e dust could not ignite and a n explosion could not take place. For some time t h e possibility of preventing explosions within machines b y t h e use of inert gases has been under consideration. I n a n article on “Coal Dust Explosions and Their Prevention,” J. Harger2 recommends a low oxygen content in t h e atmosphere of t h e mines as a preventive of explosions. The writer states in part as follows: “ T h e only way t o absolutely prevent dust explosions is t o reduce t h e oxygen percentage below t h e lower limit, which varies with t h e different coal dusts. 1 “Inflammability
of Carbonaceous Dusts,” THIS JOURNAL, 9 (1917).
269. 2
J . SOL.Chem. Ind.. 31 (1912). 413.
