JOHNSON O N FOR.WATION OF WHITE SCALE ON STEEL. gravimetrically by mercurous nitrate precipitation with the following results: Vanadium, per cent.: 1.36, 1.39, 1.34, 1.37, 1.35, 1.36, 1.36, 1.36. Then as above, titrating cold, per cent.: I .3j, I .33, I .36,I .33,I .29, I .3I,I .29. Then with iron present equivalent to a sample of steel: I .37,I .38,1.39,1.38, 1.37,I . 3 8 , 1.36, 1.38,1.39. Then without iron, titrating hot: 1.37, 1.37. The point made by Treadwell that strong hydrochloric acid reduces V,05 to a variable mixture of V,O, and V,03 is a point that i t appears has been entirely ignored by other analysts. The above results show no indication of any such thing as declared by Treadwell, but the writer did get a series of three results-all obtained at the same time-that do corroborate Treadwell's statement, namely, 1.48, I . 5 2 , I .55. But evidently the error is one that occurs but rarely. To obviate it altogether, the writer can think of nothing better than to evaporate to pastiness a second time before titration, in order to give the V,O, if present plenty of opportunity to oxidize up to lTzO,. Whether or not this precaution serves its purpose, he cannot say. H e has not had any high results since its adoption.
THE FORMATION OF WHITE SCALE ON STEEL AND THE SURFACE DECARBONIZATION OF PIPE-ANNEALED STEEL.' By CHARLESMORRIS JOHNSON. Received April 10, 1909.
When bars of steel are annealed in pipes, with charcoal, to produce a scale-free, frosted, metallic finish, there is frequently found a t the surface of the metal a coarsely crystalline structure (H), (G) Fig. 2 , that is much lower in carbon content than the remainder of the bar. Such steel will not harden file-proof on the outside. It is rejected for that reason by makers of twist drills, though this defect be so slight as to require a magnifying glass for its detection. In pipe-annealing, the bars are put in a steel tube that is welded shut a t one end. The spaces between the pieces are filled in with wood charcoal. The open end of the pipe is plugged with fire-brick, fire-clay and a disc of plate steel. A small vent hole is located at one end of the pipe to permit the escape of the large quantity of carFrom a paper read at the March, 1909, meeting of the Pittsburg Section of the American Chemical Society.
459
bon monoxide that is generated by the reaction between charcoal and the air yet remaining in the vessel. As apparently dry charcoal often holds considerable moisture in its pores, some water vapor must also be liberated. The superficial decarbonization (G), (H), (I), Fig. 2 , is a t times much more pronounced than a t others. The writer became interested to investigate the process with a view to discovering the primary cause of this very objectionable feature of annealing in a closed tube. In the first place, i t was soon noted that the condition frequently exists in steel before it reaches the annealer, due to forging a t too high temperatures. Again, i t was deemed possible that the scale, always existing on the steel when i t is put in the pipes, might react with the charcoal to form CO,. Further, that the latter gas would, under the existing conditions, decarbonize the steel by the reaction CO, C = aCO. To test this theory some 5,"" Rd., high carbon steel rods were placed in a porcelain tube and heated for 18 hours with a slow stream of pure, dry carbon dioxide passing through the enclosure. The following points were noted: 1st. A glittering black scale was produced on the fractured or otherwise unpolished surfaces of the bars. On fracturing the latter, a distinct ring of coarse crystals was found to exist a t the margin of the fractures. This scale has a curious property of adhering in a thick, sparkling black mass on rough fractured surfaces but when polished steel is exposed, at a red heat, to the attack of CO,, only a black discoloration resulted. The scale referred to, proved, on analysis, to be Fe,O,. This experiment showed that carbon dioxide may cause '' bark" (surface decarbonization) b u t not to a marked enough extent to offer a satisfactory explanation. 2nd. Some pieces of the same bar were heated in a stream of pure, dry hydrogen. A frosted metallic surface was produced and also a slight '' bark." Hydrogen, therefore, will decarbonize steel by forming hydrocarbons. 3rd. Next, a piece of the bar was heated in the closed porcelain tube, packed loosely with charcoal. First air was expelled, but, as the heat attained slight redness, large quantities of CO escaped a t the outlet end. After an 18-hour heating, a t about j j o ° C . to 780' C., a handsome frosted, metallic surface had displaced the black oxide and much decarbonization was noted. The carbon
+
460
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 . July, 1909
content of the bar before annealing was I .08 per cent. The decarbonized zone yielded b u t 0.84 per cent. 4th. Thinking that CO gas might be the active agent in (3), a small clay boat was filled with about a tablespoonful of charcoal. The boat was then placed in the tube with a piece of the steel and the usual period of heating followed. The result was the same as in the third experiment. h’ote: Annealing in a stream of natural gas produced a sooty exterior and a heavy “bark” (see J, Fig. 2 ) . 5th. The fourth trial suggested annealing a piece of steel in an EMPTY, CLOSED TUBE with NO CHAR-
The writer felt that he had now reached the first goal and that the most active agent in surface decarbonization is the rust or scale which is actually reduced to metal a t the expense of the carbon in the steel to which it adheres: 4C Fe,O, = 3Fe 4CO. I n this connection the question arose, would steel with a n extra heavy, loosely adhering scale perform in like manner? Or, in other words, Was close contact necessary for reduction? To settle this query a small piece of quarter octagon of saw analysis was given a prolonged heating in an opes muffle at about 8.50’ C. This treatment blistered
+
I
+
I
Fig. 1. COAL or other reducing substance! No attempt was made to remove the air except as it was partially driven out a t the vent end by expansion. The only precaution taken was to prevent indrawing of more air, a t any time, while the tube was hot. The same result was obtained as is shown in the white bars a t (C), Fig. I , that is, an aluminumlike surface with great decarbonization underneath ! This experiment was repeated with steel containing but 0 . IO per cent. carbon and also with steel containing large quantities of chromium and tungsten. The 0.10carbon steel is given a t (B), Fig. I .
the bar with a thick scale that was so loose that the sample had to be transferred to the annealing tube quite carefully to prevent the scale from being jarred off. After 15 hours’ heating no change was noted, that is, the scale was still black. After a second heating of 18 hours the surjace of the scale presented a slight grajish caste. After a third period of 18 hours’ heating i t was found that the heavy black scale was gone and, in its place, a white, loosely adhering, aluminum-like scale existed on the steel. Here, apparently, the black scale had begun to reduce on its top surface first.
JOHN,SON O N FORMATIOiV OF W H I T E SCALE ON STEEL. This fact pointed to the existence of a reducing gas in the tube. 6th. The writer then placed a piece of high carbon steel in the tube, together with a porcelain boat containing some hard, semi-fused iron oxide obtained from a carbon combustion of steel drilliiigs in oxygen. This hard, baked mass (D), Fig. 2 , was given 18 hours' heating at a temperature of about 8jo-gooo C. The porcelain boat was removed and found to contain, instead of a dense baked mass, a loose, friable substance, and that the volume of i t exceeded the original about 2 % times. The steel that had been in the tube was then fractured.
461
(the writer used concentrated sulphuric acid for a seal), together with a piece of low sulphur, high carbon tool steel. Before turning off or lowering the heat i t is, of course, necessary to close all vents perfectly, otherwise air will be drawn in the tube and metallic iron surfaces will lose the aluminumlike luster and become blued. Metallic iron so prepared should certainly be free of occluded hydrogen, which constitutes a n objection to electrolytic iron. Some specular iron ore was ground to a red powder and then reduced to a gray powder in this way. A piece of steel (A), Fig. I , that had been lying
H
I
J
Fig. 2
It was exceedingly tough and disclosed a heavy surface decarbonization (B), Fig. 2 . The loose sponge of oxide was put back in the tube with a fresh piece of steel and given a second heating. This time the substance (E), Fig. 2 , in the boat had become light gray in color and was no longer friable but was now adherent and almost sticky in its clinging fibers. This material assayed 98 per cent. metallic iron and, on being cut with a knife blade, presented a metallic luster. I t occurs that here is a means of preparing pure metallic iron from pure oxide by heating it in a closed tube with a sealed vent
in water for weeks and was covered with both black and yellow oxides was heated in an empty, sealed tube. The result was a white metallic-surfaced sample. A piece of blue steel was heated four hours in a n empty tube and the blue surface was replaced by a white aluminum-like one. 7th. Further experiments developed the fact that this scale forms much more rapidly at high temperatures, that is, those above 700'. The higher the heat the more rapid the transformation from rust and black scale to the white and metallic scale. This white scale takes on the white ap-
462
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHE*VISTRY . July, 1909
pearance long before it is entirely reduced to metal. The author has had white scale t h a t would be brittle and grind to a black powder. However, when the reduction is complete the scale is no longer brittle and cannot be powdered, but is entirely metallic in its properties. 8th. The corollary from the fact t h a t white scale forms more slowly below 750’ C., is that, at still lower ranges, perhaps below 650°, it may not form at all. Further experiments covering this point will be made. 9th. By annealing steel in a closed tube with a small vent to permit egress of gases b u t sealed against ingress of air, at temperatures close to jooo C., the surface decarbonization is so slight t h a t no ring of coarser crystallization can be detected. Only a cupped effect can be noted around the margin of the fracture (see F, Fig. z ) , yet such steel will take on a suggestion of the aluminum-like finish. Here the surface decarbonization is confined to the thinnest skin. Such steel hardens file-proof immediately under this extremely thin zone. 10th. B y annealing steel t h a t had been polished free of all rust and scale, in a tube from which ALL OXYGEN had been expelled by CO, no surface decarbonization was noted and the steel hardened file-proof. The CO was generated by heating wood charcoal. 11th. A rod of polished steel was dipped in a solution of copper sulphate until i t was plated with metallic copper. After heating this rod in a closed tube, without expulsion of the air, for a few hours, the rod was removed from the tube and was found to be coated with a handsomely appearing metallic copper. During this experiment the tube was sealed against ingress of oxygen.
who is seeking generally for a desulphurizing agent, which barium sulphate does. Because of the high atomic weight of barium, one per cent. of sulphur combined with it means 7 . 2 8 per cent. of the whole, while one per cent. of sulphur in combination with calcium represents only 4 . 2 5 per cent. of the whole. The determination of the sulphur alone, therefore, does not tell the whole story. It therefore became a matter of first importance to devise a rapid means of properly valuing this material while i t is still on the cars.
BARIUM AND SULPHUR IN FLUORSPAR. B y HENRYG. MARTIN.
Received April 29, 1909.
The author has recently analyzed samples of fluor spar containing quantities, u p to ten per cent., of barium sulphate. The presence of this impurity is a more reasonable cause for condemnation of the material than either high silica or high percentage of carbonates, for while the presence of the latter substances, b y lowering the content of calcium fluoride, and thereby affecting the value of the material, gives legitimate cause for argument before paying the bill, neither one proves t h a t bane to the open-hearth furnace manager,
SL/NDST‘RO! - BOMB. SULPHUR /N FLUOR SPAR. BY h/.G.M?nw
For this purpose recourse were had to the method given by S.,117. Parr’ for the determination of sulphur in mineral matter, using a “Sundstrom” bomb2 for the combustion and oxidation. The bomb (see figure) was made in our machine shop out of seven per cent. nickel steel and was heavily nickel-plated. Jour. A m . Chem. Svc.. 30,764. Ibid.,25, 184.
