US2926993A - Process for the concentration of zirconia - Google Patents

Description

March 1, 1960 D. w. MARSHALL ETAL PROCESS FOR THE CONCENTRATION OF ZIRCONIA 3 Sheets-Sheet 1 Filed Dec. 5, 1956 INVENTORS DOUGLAS W MARSHALL J. SCOTT JOHN ATTOEN'Y March 1, 1960 0 w. MARSHALL ETAL 2,926,993

PROCESS FOR THE CONCENTRATION OF ZIRCONIYA Filed Dec. 5, 1956 a sheets-sheet 2 INVENTORS DOUG LAS w. MARSHALL,

JOHN J. SCOTT A 7' TORNEY March 1 1960 D. w. MARSHALL ETAL PROCESS. FORTHE CONCENTRATION OF ZIRCONIA 3 Sheets-Sheet 3 Fiied Dec. s 1956 INVHVTORS DOUGLAS W. MARSHALL,

Fig/Z JOHN .1. SCOTT 4 1 HTTOE EY welded together.

Unitgd g rent PROCESS non THE CONCENTRATION on ZIRCONIA Y Application December'S, 1956, Serial No. 626,467

3 Claims. (or. 23-16) The invention relates to the concentration of Zirconia. This application is a continuation in part of our copending application, Serial No. 357,726, filed May 27, 1953, and now abandoned.

One object of the invention is to provide an inexpensive process for the concentration of Zirconia. Another object is to eliminate the use of iron borings in the process offusing ores to concentrate zirconia. Another object is, in the process of beneficiating zirconia containing ores, to eliminate most of the silica by reduction and conversion to a mixture of gas and fine dust, which passes out of the furnace. Another object is to provide a process of the nature indicated wherein no metal is formed in the furnace.

Other objects will be in part obvious or in part pointed out hereinafter. I v

lathe accompanying drawings illustrating one type of arc furnace by means? of which the process of the invention can be carried out,

Figure l is a front elevation of the furnace shell;

Figure 2 is plan view of the furnace shell,

Figure 3 is a sectional view of the furnace shell, the section being taken along the line 33 of Figure 1,

Figure 4 is a sectional view of the wall of the furnace shell, the section being taken along the line 44 of Figure 2,

Figures 5 and 6 are fragmentary sectional views taken on the lines 55 and 6--6 respectively of Figure 1,

Figures 7 and 8 are plan views of water-cooling pipes,

Figure 9 is a front elevation of a furnace bottom truck,

Figure 10 is a plan view of the furnace bottom truck,

Figure 11 is a sectional view taken along the line 11--11 of Figure 9,

Figure 12 is a front elevation of the furnace completely assembled, illustrating the carbon or graphite electrodes and the track upon which the wheels of the truck rest.

Referring now to Figures 1, 2, and 3, the furnace shell can be made out of two pieces of steel plate its shape is oval, tapering from the bottom to the top, and the shape is sufficiently indicated in Figures 1, 2, and 3.- As shown in Figures 1, 4, and 5 at the top of the shell 20 is welded a lip 21 to which is welded, a depending skirt 22. At the bottom of the shell 20 is a large flange 23, which also can be made of steel plate and welded to this flange 23 is a depending skirt 24. As clearly shown in Figure 3 the outer contour of the flange 23.is circular while the inner contour is oval; the flange 23 is welded to the shell 20. Steel ribs 25, 26, 27,. and 28 are welded to the shell 20 and to the flange 24 and serve to give this unit great strength.

Hooks '30 made of steel plate are welded to the outside of the shell 20, near the top, at opposite ends of the long axis of the oval. These hooks 30 can be engaged by the hooks of chain falls connected to an overhead hoist, not shown, for the purpose of lifting the unit comprising the shell 20 off the ingot of zirconia formed bythe fusion, when the ingot has cooled sufiiciently. ..A pair of bent pipes32, one of which is'shown in Figure 7 and the other of which is identical therewith except that it is of the opposite hand are providedwith ery fine holes 33' which' are preferably located (when the pipes 32 are in place as shown in Figures 4 and 5) tothrow water upwardly'and inwardly; for example they can be 30 from the top and on the inside as illustrated. At one end of each pipe 32 is a pair of elbows 34 and 35 connected bya connecting nipple 36. At the other end of each pipe 32 is a cap 37 to plug the free end of the pipe whereby to compel the water to issue through the holes 33. The pipes 32 can be supported under the lip 21 in any manner; as illustrated in Figures 2 and 5 it will suffice to provide four bolts 40 extending through the lip 21 upholding bars 41 by means of nuts 42, and the pipes 32 simply rest uponthe bars 41.

Figure 8 illustrates one of a pair of bent pipes 45, of which one is as shown and the other is of the opposite hand, and these pipes 45 have holes 46 and are equipped with elbows 47. and 48 and a connecting nipple 49 on one end, and caps 50 on the other end. As shown in Figures 1 and 6 these pipes can be supported close to the shell 20 by means of lugs 52 Welded to the outside of the shell 20 together with bent bolts 53 passing through pipe sleeves 54 welded to the lugs 52, and with nuts 55 to draw the bent bolts firmly against the pipes 45. Illustratively the holes 46 can be oriented as shown in Figure 6, but this is not critical so long as they are located to throw the water inwardly since at the locus of pipe 45 there is already a cascade of water.

The pipes 32 and 33- are, of course, connected to water supply as by means of hoses 57, 58, 59, and illustrated in Figure 12, so that atall times when the power is on, the shell 20 is cooled by a cascade of water. Although zirconia melts at the remarkedl'y] high temperature of about 2700" C. and steel plate melts at around 1500 C;, the cascade of water over the shell 20 coupled with the fact that steel is highly thermally conductive protects the shell 20 from melting. The cascade of water hugs the surface of the shell 20 and covers all parts of it from the top down and is assisted in so doing by the taper of the shell. Also it may be remarked that the shell 20 quickly acquires a coating of rust, which is easily understood since it is alternately wet and dry and frequently quite hot, and the rust makes it much more wettable than clean, unrusted steel.

Referring now to Figures 9 and 10, the furnace bottom truck comprises a circular table made of steel plate to which is welded a depending skirt 66 also made of steel plate. To the upper surface of table 65 is welded an oval bottom container 67 made of steel plate. This oval bottom container 67 has the same shape as the bottom of the shell 20 as shown in Figure 3, but the oval container 67 is slightly smaller in size so that the shell 20 and flange 23 can be placed over the container 67 and on the table 65. As shown in Figure 12, when the shell 20 is placed on the bottom with the flange 23 on the table 65, a hose 69 isinterposed to act as a seal. This hose 69 is a plain piece of rubber hose without couplings or other fittings and simply acts as a cushion and sealing means, and extends all the way around the table 65 with the ends thereof overlapping.

Referring now to Figures 9 and 11 a pair of parallel steel I beams 70 are welded to the bottom of the table 65 and a plurality of I beams 71, for example three of them, are also welded to the bottom of the table 65. These I beams 71 are desirably perpendicular to the I beams 70. To the outer I beams 71 are welded U-shaped bars 72, for the purpose of connecting a hook of a chain to the truck to draw it along the track 73 illustrated in Figure 12. v

Restingupon the track 73 are four flanged wheels 75 TABLE I Percentage by weight Zirconia, ZrO 65 to 67 Silica, SiO 32 to 34 Iron oxide, calculated as Fe O 0.01 to 0.30 Titania, TiO 0.02 to 0.30

Alumina, A1 Less than 1.0 Other matter, apart from hafnia, HfO Trace For most uses, the hafm'a content of zirconia is reported as zirconia, since hafnia reacts the same as zirconia in most chemical reactions and has many similar physical characteristics. Almost all zirconia containing ores have a small proportion of hafnia therein, perhaps of the order of two percent. The presence of hafnia, even up to or more in the ore and in the final product is immaterial in our process which functions relative to the hafnia the same as it does to the zirconia. We do not know of any ores that are richer in hafnia than they are in zirconia, but if there were any we believe our process would purify the hafnia. We shall hereinafter refer to the zirconia content with the understanding that this includes the hafnia content also.

Before assembling the shell 20 and the furnace bottom 67 on the truck, we fill the bottom container 67 with zircon ore which can be the same ore which is going to be concentrated or it could be other ore of generally similar characteristics, for example it could be baddeleyite, which is relatively pure zirconia of alluvial origin. We generally fill the bottom 67 to the brim. Next we assemble the furnace as shown in Figure 12, with the shell 20 on the truck table 65 and surrounding the container 67, and then we shovel the same ore as fills the container '67 (or similar ore) into the annular space between the shell 20 and the outside of the oval container 67, filling this space to the top.

Next we load the shell 20 with a level layer about eight inches deep of the mixture to be concentrated.

This figure eight inches is for furnaces of various sizes since the depth. of the first layer is a function of the melting point of the mixture. Of course the depth does not have to be exactly eight inches, but could be more or less over a fairly wide range, yet about eight inches is our best appraisal of the proper depth.

The mixture is the ore above illustratively described in Table I, plus from 45% to 90% of the stoichiometrical quantity of carbon required completely to reduce the SiO in the ore to silicon. The amount of coke to use depends upon the quantity of fixed carbon in the coke, which is determined beforehand by sample analysis in any desired manner, such analysis being well understood to those skilled in this art. The coke can be metallurgical coke or petroleum coke or pitch coke. The

coke for the mixture can be calcined or uncalcined,

but we prefer-a dry coke (not wet with water) for ease of feeding. The partice .size of the coke can vary, usually we' prefer sizes from lumps of one-half inch size down to fine particles as fine as 100 mesh size, but usu ally 75% of the coke should be no larger than one quarter inch.

The zircon or other ore should have a considerable quantity of fines and as zircon ore comes in particle sizes from about 72 mesh grit size to under 255 mesh grit size, we can use this material as it comes. An analysis of a typical lot of zircon ore is as follows.

4 TABLE 11 by Weight 208 Through 25S Loss Next we build a bridge of coarse coke on top of-the layer of mix. The coke should be electrically conductive coke; one way to make coke electrically conductive is to calcine it to drive off the volatiles or most of the volatile matter. The lumps of coke may be as large as two inches as as small as half an inch; in general we prefer lumps of about one inch size. The bridge of coke lumps should be about as wide as the diameter of the electrodes and need not be wider; it should be about two inches deep everywhere and a trench in the mixture is preferably made to receive it. I It should extend from the locus of one electrode to to the locus of the other electrode; from the outer edges of each.

Now we can move the truck on the track 73 to place the furnace shell 20 in symmetrical position under the electrodes (see Figure 12) and then we lower the electrodes to make contact with the coke bridge. The electrodes 80 are supported by overhead mechanism involving an automatic controller to lift the electrodes when the kva. (kw.) increases and to lower them when the kva. decreases. The electrodes continually hunt for position responsive to the automatic control. Such supporting and controlling mechanism is well known but is classified as electrical apparatus and is in a different field of science from the present invention.

We cannot give the parameters of the current in the furnace, nor of the resistance of the bridgenor of the bath of zirconia which is formed. The former varies for different sizes of furnaces and from time to time in a given furnace, while the latter is widely variable even .in a single processing (furnace run). However, we can give the'parameters of the electromotive force which should be from about 80 volts to about volts. As the is increased the reducing conditions are decreased to the point where less of the SiO- in the ore which is combined with ZrO as zircon, ZrSiO is separated so that the final product contains too much Si0 component. Higher voltage results in longer arcs more gaseous turbulence and therefore more atmospheric air under the electrodes which, at about 160 volts, so inhibits the reducing effect of the coke in the mix as to increase the amount of Si0 in the final product thus to lower its value commercially to the point where the process becomes uneconomical.

Furthermore, the electrodes themselves are strongly reducing when in the molten bath and if they are pretty close to it on the average they make frequent contact with it because of the hunting efiect and the splashing of the liquid. But if the electrodes are high, on the average, as in the case when the voltage is high, they actually contact the liquid less and at a certain voltage value (other conditions being equal) they will not contact the liquid at all. So therefore, for practical commercial operation of the process, the should not be greater than about 160 volts.

On the other hand, as the is decreased','- the reducing effect is increased, to the point where not only is the SiO;. separated from the ZrSiO but is also reduced 7 spots there.

to metallic silicon, otherwise called elementary silicon Si.- At a low voltage not all of this will pass off as gaseous silicon and in fact, at less than 80 volts, at great deal of it will be found in the pig produced by the process, some of 'which will recombine with oxygen after the furnace shell is stripped. However, we want neither SiO nor Si nor ZrSiO in the pig which is the product of the process, that is we want as little of any of them as possible. Lowering the voltage to below 80 volts insures frequent or continuous contact of the electrodes with the bath greatly increasing the reducing effect. The electrodes are, of course, made of graphite or carbon.

A further explanation of why the conditions should be reduced to a certain degree but no more so is that at high temperatures (and we attain them) the SiO splits off from the ZrO -SiO but that the temperature cannot everywhere in the turbulent bath be high enough to volatilize all or nearly all of the SiO sothat wemust have the conditions sufficiently reducing to reduce Si to Si or to a lower oxide or oxides of silicon. a large part of the SiO of the zircon is reduced to silicon we find a considerable quantity of silicon in the pig which, in the absence of any appreciable amount of metallic iron in the melt, does notgo to the bottom but is found well distributed throughout the pig. Now one of our objects was to eliminate the use of iron borings which are expensive, but we certainly do not want to make an inferior zirconia, and zirconia containing free silicon is inferior for many purposes.

We believe that by keeping the reaction zone from being too reducing we reduce the SiO;,; component to silicon monoxide or other oxide of silicon which is volatile at relatively low temperatures. Anyway by feeding the mixture to zones right under the electrodes, we get rid of the partially reduced silica almost entirely. It is very hot under the electrodes and there are no cool We believe that the Zircon ZrO -SiO breaks down into ZrO +SiO or possibly Si O, the freed oxygen unites with the carbon of the coke and a fume of CO gas, SiO gas, and reoxidized condensed SiO particles goes up the hood and is carried. away by the ventilators. 'For carrying out the process a hood and a ventilation system should be used but this is well known apparatus. 1

The limitation of not more than 90% of the stoichiometric quantity of carbon plus the feeding of the mixture to right under the electrodes has made a great change in this art and eliminates the use of iron which is expensive in the first place and hard to get rid of in the second place.

We are aware that in US. Patent No. 1,427,816, Otis Hutchins described fusing in an electric arc furnace an ore having the analysis 64.9% ZrO 19.0% SiO 7.5% Fe O 1.1% TiO ignition loss 2.8% and he says that, using 13.2% of cokethe product had the following oxide analysis: 97.7% 'ZrO .3% SiO .2% Fe O and .5% TiO But he does not account for the elementary silicon in the pig of which hecertainly had a great deal because he used much more coke than the maximum according to our formula. Hutchins reduced 7.3% of Fe O which requires 1.64% carbon.v He reduced .6% TiO which requires .18% carbon. This totals 1.82%. If this is increased by to convert to coke, the calcu- Butif' lation is perfectly fair because coke usually has 90% or of the silica. He had 19% SiO in the ore and stoichio- 'metrically this requires only 7.6% of carbon or roughly 8.0% of coke. Therefore it is shown that Hutchins used considerably more than the stoichiomctric proportion of carbon or coke required whereas we have found it is critical that no more than 90% thereof be used,

Example 1 Using the furnace illustrated in which the height of the shell 20 was four feet and six inches (the other dimensions in proportion) we filled the bottom container 67 as previously described, placed the shell 20 on the table 65, filled the space between the shell 20 and oval bottom container as described and loaded the assembled furnace with eight inches of the zircon sand and coke mixture to be concentrated according to the following analysis:

TABLE III Percentage by weight Zirconia, ZrO 60.00 Silica, SiO 30.10 lron oxide, calculated as Fe O 0.04 Titania, TiO 0.24 Alumina, A1 0 -About 0.50 Coke (88.8% carbon) 8.88 Other matter, apart from hafnia, Hf0 About 0.25

We then built a coke bridge four inches deep between the loci of the electrodes as described and, lowering the electrodes 80 onto the bridge, we applied the power at volts. The mix was continuously fed at the rate of 500 pounds per hour in steady streams through two large steel pipes 81 and 82 whose location and angular relationship to the electrodes 80 are shown in Figure 12 and in dotted lines in Figure 2. These steel pipes 81 and 82 feed the mixture right to the electrodes and the mixture follows along the electrodes and enters the pool of molten matter right under the electrodes thus satisfying our requirements in this respect. The axes of the pipes 81 and 82 are in the same plane as the axes of the electrodes 80.

The furnace was continued in operation for ten hours fifteen minutes with a continuous steady feed at a uniform rate of the mixture of Table Ill. At the end of that time the electrodes were lifted, the furnace was allowed to cool and after sixteen hours the shell 7 was stripped from the pig. Of course during the operation of the furnace the water was turned on at all times and during the cooling of the pig in the shell the pipes 32 were cascading water over the shell.

As the result of this furnacing operation, a pig having the following analysis was produced:

TABLE IV lercentage by Weight Silica, Si0 0.14 Iron oxide, calculated as Fe O 0.10 Titania, TiO 0.47 Alumina, Al O Less than 1.0

Other matter, apart from zirconia and hafnia Less than 0.5 Zirconia and hafnia Balance to detail the same. Our process is quite economical and yields a crystalline zirconia of the purity indicated. Large furnaces built substantially in accordance with the accompanying drawings can be built and used in this process; for example in furnace shells. 20 which are about nine feet high. We feed the mixture through the pipes 81 and 82 at such a slow rate that We maintain a molten pool of zirconia under each electrode. In.

other words the incoming mixture fuses almost instantaneously which is why We are able to get rid of the silica content as reduced gaseous oxide of silicon, having less oxygen than SiO The incline of the pipes 81 and .82 is such that the incoming mixture has a trajectory which carries it into contact with the electrodes 80, which contact causes its subsequent travel to be vertically downward thus getting the mixture to right under each electrode.

The electrodes can be carbon or graphite electrodes, both kinds being in current use in electric arc furnaces. The word carbon has two meanings, being generic to amorphous carbon, graphite and diamond and being specific to amorphous carbon. Graphite electrodes do not burn away as fast as amorphous carbon electrodes .but the latter are cheaper per piece. But the melting point of zirconia is about 2700 C. and for zirconia graphite electrodes are cheaper in the long run because they last longer. Also graphite is more conductive than carbon and thus smaller electrodes can be used. Furthermore graphite electrodes are purer in C than are carbon electrodes.

The pig produced does not occupy the entire volume of the shell 20, as about 40% to 60% will be partially concentrated material which is easily separated from the pig of concentrated zirconia, and such partially concentrated material is mixed with ore for charging a furnace in a subsequent run. (This is a higher yield of fully concentrated material than used to be achieved.) Therefore the charge for subsequent runs will be higher in zirconia and lower in silica than in Table 1, within the range, so far as the oxides are concerned and apart from the carbon, within the following table:

TABLE V Percentage by weight Zirconia, ZrO 67 to 90 Silica, Si mm 32 Iron oxide, cal. as re o, .01 to- .30 Titania, Ti0 .05 to 0.40 Alumina, A1 0 Less than 1.0 Other matter, apart from HfO and carbon Trace Thus the charge for the furnace can be anywhere within the combined limits of Tables I and V.

Example II Using the furnace illustrated, as in the case of Example I, we proceeded as set forth in Example I to concentrate zirconia from zircon sand, coke and lime, CaO,

The lime was added in accordance with the teachings in the patent to Ballard and Marshall, No. 2,535,526, patented December 26, 1950, to make stabilized zirconia for reasons set forth in that patent and the percentage of lime in the zirconia extracted should be within the limitation set forth in that patent, namely from 3% to 6% of CaO on the total ZrO in the product. All the steps in this example of the invention were, except for the provision of the mixture of Table VI, the same as set forth in Example I.

As a result of this furnacing operation, a pig of the following analysis was produced:

TABLE VII Percentage by weight Silica, SiO, 0.40 Iron oxide, calculated as l e- 0 0.16 Titania, TiO; 0.28 Alumina, A1 0 Less than 1.0 Lime, CaO 3.98 Other matter, apart from zirconia and hafnia Trace Zirconia and hafnia (by difference) 94.18

It thus appears that the mixture fed under the electrodes in accordance with this invention is zircon ore and coke the mixture being substantially free of free iron but it may or may not have lime in the amount of from 3% to 6% of the amount of Zr0 in the ore depending upon whether stabilized or unstabilized zirconia is wanted. Another way to say it is that the mixture which is fed to the furnace essentially consists, apart from any stabilizing agent, of only zircon ore and coke which mixture is substantially free of free iron.

It will thus be seen that there has been provided by this invention a process for the concentration of zirconia in which the various objects hereinabove set forth together with many thoroughly practical advantages are successfully achieved. As various possible embodiments might be made of the mechanical features of the above invention and as the art herein described might be varied in various parts, all without departing from the scope of the invention, it is to be understood that all matter hereinbefore set forth or shown in the accompanying drawings is to be interpreted as illustrative and not in a i limiting sense.

We claim:

1. Process for the concentration of zirconia to produce a product low in silica and low in silicon from zircon ore in an electric arc furnace having carbon or graphite electrodes, the zircon being combined zirconia and oxide of silicon, which comprises building a carbon bridge between the loci of the electrodes, feeding to the furnace substantially continuously right under the electrodes a mixture comprising zircon ore and coke substantially free of free iron and containing not more than 0.30% of combined iron calculated as Fe O in an amount such that the fixed carbon content of the coke is from 45% to 90% of the stoichiometric proportion required completely to reduce the oxide of'silicon content of the ore to silicon, energizing the electrodes with an of from volts to 160 volts thus forming a molten bath out of the mixture, the mixture of zircon ore and coke being fed directly to said molten bath, heating the bath to such a high temperature that the silica of the zirconia is driven off as gas and silica particles without forming a metallic button, feeding the mixture slowly enough so that a molten pool of zirconia is maintained under each electrode, breaking down the zircon into material including oxide of silicon free from the zirconia while volatilizing some of said oxide of silicon free from the zirconia thereby forming zirconia low in silicon values by the elimination from the zircon of oxide of silicon with which the zirconia was combined, and recovering said zirconia low in silicon values.

2. Process for the concentration of zirconia to pro duce a'product low in silica and low in silicon from zircon ore in an electric arc furnace having carbon or graphite electrodes, the zircon being combined zirconia and oxide of silicon, which comprises building a carbon bridge between the loci of the electrodes, feeding to the furnace substantially continuously right under the electrodes a mixture which, essentially consists of zircon ore, lime and coke which mixture is substantially free of free iron and containing not more than 0.30% of combined iron calculated as Fe O in an amount such that the fixed carbon content of the coke is from 45% to of the stoichiometric proportion required completely to reduce the oxide ofsilicon content ofthe ore to silicon, energiz- 9 ing the electrodes with an E.M.F. of from 80 volts to 160 volts thus forming a molten bath out of the mixture, the mixture of zircon ore and coke being fed directly to said molten bath, heating the bath to such which the zirconia was combined and recovering said zirconia low in silicon values.

3. Process according to claim 2 in which the lime in the mixture that is fed to the furnace is from 3% to6% a high temperature that the silica of the zirconia is driven 5 of the zirconia in the mixture by weight.

olf as gas and silica particles without forming a metallic button, feeding the mixture slowly enough so that a molten pool of zirconia is maintained under each electrode, breaking down the zircon into material including oxide of silicon free from the zirconia while voltailizing some of said oxide of silicon free from the zirconia thereby forming zirconia low in silicon values by the elimination from the zircon of oxide of silicon with References Cited in the file of this patent UNITED STATES PATENTS

Claims (1)

1. PROCESS FOR THE CONCENTRATION OF ZIRCONIA TO PRODUCE A PRODUCT LOW IN SILICA AND LOW IN SILICON FROM ZIRCON ORE IN AN ELECTRIC AND FURNACE HAVING CARBON OR GRAPHITE ELECTRODES, THE ZIRCON BEING COMBINED ZIRCONIA AND OXIDE OF SILICON, WHICH COMPRISES BUILDING A CARBON BRIDGE BETWEEN THE LOCI OF A ELECTRODES, FEEDING TO THE FURNACE SUBSTANTIALLY CONTINUOUSLY RIGHT UNDER THE ELECTRODES A MIXTURE COMPRISING ZIRCON ORE AND COKE SUBSTANTIALLY FREE OF FREE IRON AND CONTAINING NOT MORE THAN 0.30% OF COMBINED IRON CALCULATED AS FE2O3, IN AN AMOUNT SUCH THAT THE FIXED CARBON CONTENT OF THE COKE IS FROM 45% TO 90% OF THE STOICHIOMETRIC PROPORTION REQUIRED COMPLETELY TO REDUCE THE OXIDE OF SILICON CONTENT OF THE ORE TO SILICON, ENERGIZING THE ELECTRODES WITH AN E.M.F. OF FROM 80 VOLTS TO 160 VOLTS THUS FORMING A MOLTEN BATH OUT OF THE MIXTURE, THE MIXTURE OF ZIRCON

Images