INDUSTRIAL AND ENGINEERING CHEMISTRY Meinert, R. N., a n d Hurd, C. D., J. Am. Chem. SOC.,53,4840 (1930). Midgley, T. G., a n d Hochwalt, C. A., U. S. Patent 1,713,236 ( M a y 14, 1929). Norris, J. F., a n d Reuter, R., J . Am. Chem. Soc., 49, 2624 (1927). Pease, R. N , Ibid.. 52, 1158 (1930). Pease, R. N., Ibid., 53,613 (1931). Peters, W.A., a n d Baker, T., IND. ENG.CHEW,18,69 (1926). Routala, O., Petroleum (Chicago),5, 321 (1909).
(28) (29) (30) (31)
Vol. 24, No.
Stobbe, H . , a n d Posnjak, G., Ann., 371, 259 (1909). Van Winkle, R., J . Am. Pharm. Assoc., 17, 544 (1928). Voorhees a n d Eisinger, Oil and Gas J.,27, No. 31, 152 (1929). , 1823. Wheeler, R. V., a n d Wood, W. L., J . Chem. ~ o c . 1930,
RECEIVED November 13, 1931. Contribution 217 of the Department ot Chemistry, University of Pittsburgh. This paper is an abatract of a thesis presented t o the Graduate School of the University of Pittsburgh by C. R. Payne in partial fulfilment of the requirements for the degree of doctor of philosophy.
Impervious Crucibles of Magnesium Oxide PAULS. ROLLERAND DAVIDRITTENBERG U. S. Bureau of Mines, Nonmetallic Minerals Experiment Station, Rutgers University, New Brunswick, N. J.
F
OR t h e e x p e r i m e n t a l
possible. Owing to the nonRefractory crucibles of magnesium oxide have melting of high-temperaplastic nature of the moistened been prepared that are recrystallized to transture alloys, magnesium powder, there is no flow under lucence and impervious to air under pressure at oxide is particularly suitable as pressure, so that, in pressing in room temperature. a refractory because of its high one operation, some parts are The crucibles are formed by pressing the melting point, 2800" C. (4), and highly compressed and others its resistance to reduction. I n relatively s l i g h t l y . In order moistened refractory powder in two steps in melts which involve slags as in to overcome this difficulty, the such a manner that the base is welded onto the the metallurgical studies carried p r e s s i n g w a s effected in two I n this way a structure of uniform walls. out at the Pittsburgh Station operations, first of the walls and density and strength is obtained. Firing is of the B u r e a u of M i n e s , not then of the base; in other words. carried out in the high-frequency induction only is the chemical resistance the b a s e w a s welded to the to the slag to be considered, but walls after the latter had been furnace with graphite inductor at a temperature equally important is the texture c o n s o l i d a t e d . In a general of about 2600" C. of the crucible. way the effectiveness of such Significant variables are the relative pressure It is commonly known that a a procedure will depend upon the in forming the base and walls, the fineness disporous refractory is much more extent to which the p h y s i c a l tribution of the refractory powder, and the duranature of the powdered material rapidly attacked by slag than a dense one. As it was desired to is altered by pressing. If the tion of firing above 1800" C. alteration is great, conditions at employ magnesia c r u c i b 1e s in Owing to their dense structure, these crucibles the surface to be welded will, slag studies, attention was conare much more resistant to attack by slags than because of the first pressing, be centrated on p r e p a r i n g s u c h are porous crucibles. different from those obtaining crucibles so that they would be vitrified and impervious. throughout the r e m a i n d e r of Tritton ( 7 ) rindered the surface of magnesia and other the powder, and a discontinuity willresult at this place. If, on refractories impervious by playing an arc in the interior of the the other hand, the effect of pressing is merely to force the rotating crucible shape. Adcock and Turner (1) have formed particles into closer juxtaposition, then the weld will be impervious tubes of alumina by glazing the outside surface indistinguishable from the remainder of the pressed body. Hagen (5') has classified solid substances, according to the by means of the oxyhydrogen flame. I n the present work it was found possible to prepare impervious crucibles of physical properties of their compressed powders, into six magnesium oxide by direct heating a t temperatures around categories, ranging from plastic to pulverulent and nonpress2600" C. in the Ajax-Northrup high-frequency induction able. I n the latter instance, pressure exerts no physical effect other than to bring the grains closer together. Pulfurnace. verulent substances are hard and have a high melting point. FORMATION Fortunately these are precisely the properties possessed by The crucibles were made by pressing a t loads up to 60,000 the nonplastic refractories, and this accounts for the success pounds in a manually operated hydraulic press. Two of the observed in the two-step method of pressing the crucible. As will be discussed below, an important factor to be concommon difficulties met in this procedure are sticking of the crucible to the walls of the die during removal and nonuniform sidered in the two-step formation of the raw crucible is the density of the product. To prevent sticking, the usual microscopic particle-size distribution of the refractory powder. It has been found that the addition of a little phosphoric or expedients of oiling and greasing the die or of lining it with paper, etc., were tried, but with indifferent success. It was boric acid in the water used to moisten the powder helps to finally found that the combination of keeping the walls of the overcome any tendency toward the formation of small, fine core and die scrupulously clean and bright, and moistening cracks during firing. PROCEDURE OF FORMATION. Figure 1 shows the details of the refractory powder, not with water, but with dilute hydrochloric acid solution rendered the removal of the crucible the steel mold. Die A is split in two halves, each of which is easy. The hydrochloric acid also served to eliminate air-dry- case-hardened on the interior. The steel core, B, has a taper of about 2 in 1000. Plus and minus signs indicate clearances ing cracks completely. I n order successfully to fire the crucible in the high-fre- of about 0.003 inch. quency furnace a t the required temperatures, it was found The refractory powder is moistened with about 10 per cent that the structure of the green crucible must be as uniform as of its weight of water that is approximately one-quarter molar
April, 1932
INDUSTRIAL AND ENGINEERING CHEMISTRY
in hydrochloric acid and one-quarter molar in o-phosphoric acid. With core B in place in die A , the refractory powder is well tamped into the annular space to form the walls of the crucible. Strong tamping is essential in order to eliminate entrained air that would otherwise give rise t o a weak structure. Cylinder D is now set on B, and the walls formed by applying pressure to annular ring C and to core B. By means of a templet of brass, 0.003 inch thick, the excess wall is broken down with an L-shaped chisel t,o the level of the templet.
A
B C f7 FIGURE 1. DETAILS OF MOLD
t-
437
I n forming the crucible, curves, such as those shown in Figure 2, serve as guide to the relative pressure required for base and walls. I n practice, an excess in pressure of 10 per cent over that required for equal density of compression is applied to the base to consolidate more thoroughly the weld of base and walls. Any wide variation of relative pressure id1 result in cracking of the fired crucible. Aside from the relative pressure, it is found that increase in absolute pressure increases the strength of the green crucible and promotes the ease of recrystallization to an impervious structure in firing. However, it is conceivable that, by compressing and crushing individual grains, too high pressure might tend to weaken the weld, although this condition has not been perceptible at the pressures employed, i. e., up to about 40,000 pounds per square inch on the walls and 15,000 on the base.
FIRIPI'G While temperatures of 3000" C. can, without much trouble, be reached in the high-frequency furnace, there are certain other considerations that must be borne in mind in regard to the application of this furnace to firing crucibles.
With the templet removed, the material for the base is now tamped in and the base formed by applying pressure to cylinder E and t o B. To remove the finished crucible, the die bolts are loosened and made finger-tight. Core D is removed by twisting and pulling. The green crucible, which is fairly strong, is then removed from the die by forcing up at the base.
As noted, the use of dilute hydrochloric acid prevents sticking of the crucible to the mold during withdrawal. However, hydrochloric acid does not appear to be effective a t the shoulder of B or on the surface of cylinder E , both of which are subjected to direct pressure. This may be due to the fact that some pitting takes place a t these surfaces which could probably be overcome by case-hardening these parts. I n lieu of this, the protection of these places by means of' two pieces of paper is quite suitable for avoidance of sticking. The green crucible is air-dried for 24 hours and then dried a t 120" C. for several hours. When the proper procedure is observed, it is found that the weld a t the base and walls is about as strong before and after firing as the rest of the crucible. FIGURE4 PRESSURE INFORMFIGURE 3 IPI'G C R U C I B L E .I n HIGH-FREQUENCY FURNACE SET-UP forming the crucible Since a graphite cylinder is used as the inductor, a reducby p r e s s u r e in two ing atmosphere will necessarily be present. Fortunately operations, it must be 2 50 b o r n e in mind that magnesia does not appear to form a carbide. However, a t k the d e g r e e of com- about 1800" C., reaction between the carbon and the magnesia & N 5 pression depends not begins and increases rapidly with increasing temperature. only on the pressure To this must be added the sublimation of magnesia itself above $240 but also on the depth 2000" C., although the reaction with carbon appears to be of 9 235 9 of the column of pow- greater significance. As a result of these effects, the crucible d e r ( 8 ) . Conse- that has been fired to about 2600" C. will, depending on the 2 30 quently, in order to furnace set-up and duration of firing above 1800" C., weigh effect the same den- 65 to 90 per cent of the original weight of the dried crucible. 2 25 M /5 20 4'5 30 Another point to be considered is the fact that highs i t y of compression throughout the cru- frequency currents are induced solely in the vertical walls of f oad (tons) FIGURE2. DENSITYOF COMPRESSIONcible, the pressure on the graphite inductor so that heat is radiated to the refractory the base a n d w a l l s crucible as from a cylindrical surface. Consequently, the vs. LOAD will differ. Figure 2 base of the crucible tends to lag in temperature with respect to shows the density of the walls and base of the cricible, the rest of the crucible. The detrimental effects of this dried a t 120" C., against the load applied in pressing. It condition appear to be greatly counteracted by a uniformly is seen that the density of compression varies practically dense structure of green crucible. linearly with the load, both for the base and for the walls. On the other hand, the lower temperature at the base is This is the result obtained by Hagen ( 3 ) for the so-called advantageous in firing the magnesia crucibles to recrystallizapulverulent materials. Owing to the greater length of wall, tion because it eliminates the excessive interaction that would 2.25 inches against 0.25 inch, the pressure required to com- otherwise take place between the magnesia and carbon when press it to a given density is about three times that required in contact. for the base. To reduce the initial high rate of heating of the furnace, a
4%
I N D U S I R I A L A N D ENGINEIiIIIiLG
secoridary coil diose reactance was altered by means of a water-cooled iron core was placed in series with the furnace coil. FUHNACE SET-UP. The furnace (coil diameter, ti inches) is siiowri in Figures 3 and 4.
Vol 2 1 , N o 4
required for uniform heating as shown in Figure 5 . Table I gives a comparison of the uniform and actual time during which the furnace was maiiitained above 1800" C. TARLE
I. P o w m
INPI:T 08.
Uilifoii"
INPDI.
W C k
CYCh
h-w.
Min.
6.5
12
Mrn. 5
9
5
11.5
13 11.5 0 6.5
TIMEABOVE 1800" C .
T ~ MOP E 11,:~~
Poxrn
9
PROCEDTTRE I N FIRING.h i l l g to the fact that fumes Of magnesia appeared at 1800" C., temperatures could not be read directly wit.h the optical pyrometer above this point. Consequently, measurements of temperature against power iiiput and time were made on tlie empt.y furnace. By varying tlie power input direetly a t tlie gap and by nieaiis of the. secondary inductance, a firing cycle with the desired uniform lieating nnd cooling was established. Such n cyclc is sliown in
CNIiMIS1'RY
10 20 10 12 11
AOt"G1
Stl.M,,*ns Teiiii,. 1800" c. st i,w,,r,~,tg, and iurnrs cnmnreirer Fumes
s
F*mW
Punier Fumes Fumes Fumes decreasing
5
3 8 5
In the actual cycle, corresponding to the first arid third columns of T d h I, the temperature was still increasing ~vlieii the gomer was dropped from 13 to 11.5 kw., reaching a marimum of about 2600" C. as determined directly on the empty furnace and also on the furnace with crucible therein by extrapolation or by plugging the graphite sight tube so as t o c u t off t h e fumes. A correction m u s t be made to the actual cycle above 1800" C. because of deterioration of the carbon black insulation with time. As a measure of t h e i i i s u l a t i o n efficiency, t h e t i m e required for the Eursiace temperature to drop, with power off, from 1250" to 1150" C. w n s n o t e d . I n t,crms of this time, an empirical correction was made to the duration of the snaximuiii f u r n a c e power, iii this case 1 3 kw. FlGUnE
6.
CRUCIBLE1 I L L ~ I M I N ~ T E D
SIZE. A parIWOM 1WFr:nI"R ticularly significant variable in this work has been the particle size of the refractory powder. The powder was passed entirely tlirougli the 2OO-mcsh sieve; thus the particles correspoiided to a maximum size of 100 microns. If coarser grains are present, it has been foiind that, although the crucibles are well crystallized and exhibit the same porcelain-like translucence (compare Figure ti), nevertheless they are not impervious. This is understandable from the fact t.hat the distance of separatioii of the grains increases with particle size so that for very coarse grains the structure of the fired crucible, even if well recrystallized, will contain many unsealed spaces. 7;n
