LIBRARY OF THE UNIVERSITY OF CALIFORNIA, RECEIVED BY EXCHANGE Class tfje Specific (irabttteg of anb tantalum MAURICE ALLISON LAMME On the Specific Gravities OF Niobium and Tantalum Pentoxides BY MAURICE ALLISON LAMME, B.S.. A.M. DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN THE FACULTY OF PURE SCIENCE OF COLUMBIA UNIVERSITY LANCASTER, PA. STEINMAN & FOLTZ, 1909 ? ACKNOWLEDGMENT This investigation was undertaken at the suggestion of and carried out under the direction of Dr. Floyd J. Metzger, to whom I here wish to express my appreciation. M. A. L. QUANTITATIVE CHEMICAL LABORATORY, HAVEMEYER HALL, COLUMBIA UNIVERSITY, MARCH 1, 1909. Specific Gravities of Niobium and Tantalum Pentoxides PURPOSE THE original purpose of this investigation was to find a method for the determination of tantalum and niobium, without separa- tion, based on the specific gravity of the combined oxides. Until recently, 1 the only available method for their determination was by fractional crystallization of the potassium double fluorides. In this method the mixed hydroxides of the elements are obtained by fusion with potassium bisulphate and boiling the melt with water. After removal of impurities and thorough washing, these hydroxides are dissolved in hydrofluoric acid, potassium fluoride added and the solution evaporated until the more insoluble tan- talum salt separates. It is a difficult matter to free this salt from the mother liquor containing the niobium without, at the same time, dissolving some of the tantalum salt. That the method is inaccurate and unreliable, is shown by Metzger and Taylor. 2 Penfield and Ford, 3 in an article on Stibiotantalite, in which tantalum and niobium were determined by a method based on the specific gravity of the combined oxides, say "As there is no satisfactory method for separating tantalum from niobium, the attempt has been made to determine the proportions of the two oxides by taking the specific gravity of the mixed oxides as ob- tained from the analysis, and comparing the results with those of pure Ta 2 O 5 and Nb 2 5 , which are quite widely separated. This method has been tested sufficiently to prove that it gives fairly satisfactory results. . . . The greatest variation of the determination of Nb20s from these curves is not over one per cent., which is well within the errors of analysis." It was thought that if the specific gravities of the oxides were *"A New Rapid Volumetric Method for the Determination of Niobium in the Presence of Tantalum," Metzger and Taylor, to be published soon. 2 Loc. Cit. 3 "Am. Jour. Science," 1906, p. 72. 211037 4 Specific Gravities of Niobium and Tantalum Pentoxides accurately determined under conditions which gave constant results, 9 curve could be used which would give the percentage of each oxide in a mixture by its specific gravity. This would obviate the separation of the elements in the analysis and greatly shorten it. In order to do this and have the results accurate to within 0.25 per cent., it was found that the specific gravity must be accurate to the third decimal. This meant even greater accuracy in the weights of the specific gravity flask and the liquid used. It was necessary that these weights should check within 0.0005 gm. This was one of the most difficult problems. The weights did not, as a rule, check within this limit, notwithstand- ing the fact that various forms of apparatus were employed. While the work, primarily, had this object in view, it was found necessary later to change the purpose somewhat, and, therefore, the investigation has been divided into three parts. PREPARATION AND PROPERTIES OF THE MATERIAL The niobium and tantalum oxides employed in this work were prepared from South Dakota columbite by fusion with potassium bisulphate, boiling with water, removing impurities as usual, leaving the mixed hydroxides of niobium and tantalum, which were then converted to oxides and separated as follows: The mixed oxides were treated in small portions by fusion with crys- tallized potassium bisulphate (the crystallized bisulphate is much more effective in decomposing the mixed oxides than is the fused pyrosulphate), sulphuric acid added to the melt to make it pasty, the mass poured into hot water, boiled and washed several times with hot water by decantation, the hydroxides transferred to a filter (using a hard rubber funnel) and the washing with boiling water continued until thoroughly washed. The hydrox- ides so obtained were dissolved through the filter paper with hydrofluoric acid. When about a liter of this hydrofluoric acid solution had been obtained, somewhat more than the required amount of potassium fluoride to form the double salt was added and the solution evaporated in a large silver dish until crystals of the tantalum double fluoride, K 2 TaF 7 , began to form on the surface. These crystals are long, fine needles. However, if the solution contains too much hydrofluoric acid, a niobium salt Specific Gravities of Niobium and Tantalum Pentoxides 5 separates in short prisms, not unlike the tantalum salt, but larger, and more insoluble than potassium niobium oxyfluoride, K2NbOFs.H20. The amount of acid present should be only sufficient to keep the tantalum salt in solution, i. e., about 1 to 2 cc. per 100 cc. of solution. As soon as crystals of tantalum began to form, the solution was allowed to cool slowly, the crystals filtered off, washed two or three times with cold water and allowed to drain. The solution was then further concentrated and a second crop of crystals obtained, which was kept separate from the first. On the third concentration, the solution was tested from time to time by taking out a drop or two of the liquid and allowing to cool to note the presence or absence of the niobium crystals. The niobium double fluoride crystallizes in lustrous plates easily distinguished from the tantalum salt, When the proper concentration had been reached, as shown by the above test, the solution was allowed to cool. The crystals obtained at this point, being a mixture of both salts, were rejected. The mother liquor from these mixed crystals was concentrated until two crops of the potassium niobium oxyfluoride were obtained. These two frac- tions of the niobium salt were kept separate. The mother liquor was rejected. These operations were repeated until all the oxides were con- verted into the double fluorides and the mixed fluorides separated. All the " first crop" of tantalum salt was combined and recrystal- lized once. To the mother liquor from this the entire " second crop" of crystals was added and one crystallization made. The crystals from the last mother liquor were rejected. Finally, the two crops of crystals just obtained were united and recrystal- lized and the mother liquor rejected. This final lot of crystals constituted the material used in the subsequent work on tan- talum. For the K^NbOFs.H^O the same general method of purification was employed. There seems to be considerable lack of agreement among investigators with regard to the specific gravities and properties of the oxides of niobium and tantalum, as may be seen in the following paragraphs. 6 Specific Gravities of Niobium and Tantalum Pentoxides Abegg 1 gives the properties of tantalum pentoxide as a taste- less and odorless white powder which suffers no chemical change on heating. By continued heating, it becomes crystalline. Ebelmen 2 obtained well-formed orthorhombic prisms by heating Ta2(>5 in a muffle furnace with phosphorous salt, which were crystallographically investigated by Mallard 3 and by Norden- ^kjoid and Chydenius. 4 The specific gravity varies according to the method or prepa- ration and is especially influenced by the length of time and in- tensity of heating, as is shown in the following example from Rose:' 5 The specific gravity of the oxide was 7.109. Heated 1 Hour at Moderate Heat. 4} Hours. 9 Hours. 15 Hours. 7.274 7.383 7.529 7.536 And after a further heating of eleven hours at a high temperature, 7.914. On still further heating, it rose to 7.994. By heating strongly in a furnace, it sank to 7.652, because the Ta20s became amorphous. On fusion with potassium bisulphate and again heating the hyrdoxide, the specific gravity was 8.257. The values, given in the literature, vary between 7.03 and 8.26. 6 It is to be remarked, however, that many of the old determina- tions are quite unreliable on account of the use of impure Ta20g. In all cases in which the decomposition of the tantalum com- pounds has been effected by the use of potassium bisulphate, the product contains sulphuric acid, which cannot be removed by washing. To get rid of this, the hydroxide must be heated with ammonium carbonate. Niobium pentoxide can be obtained by heating E^NbOFs.H^O with sulphuric acid in a platinum crucible until all the hydro- 1 "Handbuch der Anorganischen Chemie," Vol. Ill, part III, p. 853. 2 "Ann. Chim. Phys." (3), 33, 34. 3 Compt. Rend., 105, 1260. 4 "Oefvers af. k. Vet. Ak. Forh," 1860, 3; "Pogg. Ann.," 110, 642. 5 "Pogg. Ann.," 74, 285. 6 Rose, "Marignac, Ann. Chim. Phys." (4), 9, 249; Deville and Troost, "Compt. Rend.." 60, 1121. Specific Gravities of Niobium and Tantalum Pentoxides 7 fluoric acid is driven off, boiling the residue with water, washing and heating. The precipitate must be washed with ammonia before it is heated, to remove the sulphuric acid. Obtained in this manner, it usually contains traces of alkali. The specific gravity of the oxide varies between 4.37 and 4.46 when prepared by means of potassium bisulphate, and between 4.51 and 4.53 when prepared by heating ammonium niobium oxyfluoride. 1 Strongly heated, the oxide becomes crystalline. APPARATUS EMPLOYED A pycnometer of about 30 cc. capacity, with a thermometer ground in, was first tried, but it was found that the evaporation of the water from the ground glass joint gave such variable re- sults that this form could not be used. Then, a specific gravity flask with a long neck, fine ground, 20 cc. capacity and provided with a cap ground in the same manner, was used. The bottle itself, but not the neck, was surrounded with a vaccuum jacket. This was to guard against changes in temperature in handling, etc. The object in having such a long neck was to insure a good, tight joint, so there would be the least possible difference in the volumes of the liquid in the flask in separate determinations, caused by slight differences in pressure exerted in inserting the stopper. The cap, ground in the same way, fitted over the stopper, was to prevent any evaporation from the surfaces, between the neck and the stopper or from the top of the capillary tube in the stopper. This combination was found somewhat too heavy and cumbersome, and a similar flask, having a capacity of 10 cc., was substituted later. This was used in most of the work with water, but when chloroform was used, the flask was essen- tially like that above, except that there was no vacuum jacket. It was found extremely difficult to obtain weights that checked within the desired limits. The total variation in the weights is great, but those taken within short periods of time checked much better. METHOD The method used in the first part of the work is essentially as iMarignac, "Ann. de Chim. et Phys." (4), 8, 19. 8 Specific Gravities of Niobium and Tantalum Pentoxides given below. Some changes were made from time to time to suit new conditions. The flask was weighed filled with water, the water poured out and a weighed amount of oxide in fine powder introduced, the air removed as explained later, the flask filled and weighed again. The difference in weight between the flask with the oxide and with water only was subtracted from the amount of oxide, and the amount taken divided by this value, giving the specific gravity. In preparing the flask for weighing, the stopper was inserted each time with as near as possible the same pressure. Then all superfluous moisture was removed, that from the top of the capillary tube removed last, and the cap put on. It was found that much better weights could be made if the flask were now dipped in cold water and carefully dried with a clean cloth. After standing ten minutes in the balance case, the weight was taken. The oxide must be in a fine powder to reduce to a minimum the possibility of included air. As a further precaution, after in- troducing the oxide the flask was only partially filled with water, shaken with a rotary motion and the filling completed. In completing the filling it was necessary to run the water in slowly in order to leave a clear portion above that which had been shaken up, otherwise some of the fine oxide held in suspension would be lost when the stopper was inserted. The oxides were prepared from the double fluorides as follows: A portion of the double fluoride, corresponding roughly to 1.00 gram of the oxide, was heated in a platinum dish with sulphuric acid until all the hydrofluoric acid had been driven off and copious fumes of SOs evolved. After cooling, the solution was slowly, and with vigorous stirring, poured into about 700 cc. of hot water. This was then boiled, the water poured off through a filter after allowing the hydroxide to settle, boiled up with about the same quantity of water twice more, and then filtered. The hydroxide, while still moist, was placed in a large platinum crucible and gently heated until dry. After this, the heat was raised, the paper burned off and the resulting oxide pulverized in an agate mortar to approximately 80 mesh. On trying to filter the oxides after a determination, it was Specific Gravities of Niobium and Tantalum Pentoxides 9 found that nearly all of the finer portions ran through the paper. This solution, if left undisturbed, would remain cloudy for several weeks. On account of the difficulty of filtering, not more than one value was obtained from each portion in the first part of the work. It is possible that some of this finer portion was rehy- drated to some extent (?). 10 Specific Gravities of Niobium and Tantalum Pentoxdies EXPERIMENTAL PART I In this part of the work the flask was filled with boiled dis- tilled water at room temperature, temperatures being recorded by a thermometer, graduated to read to 0.2 C., and the usual precautions taken in weighing. It was found that the weight of the flask filled with water must be taken before each determin- ation, since the variation in the temperature of the water from day to day caused such a variation in the weight that a constant value could not be used. In many cases the weights obtained on one day did not check with those obtained on the next, though the temperature was the same. The flask was handled as little as possible after taking the temperature in order to avoid any change. A number of weighings was made to determine the error in filling and in inserting the stopper and are given here: Temperature. Weight. 20.8 65.2000 20.8 65.1992 20.8 65.1983 20.8 65.1980 20.8 65.2002 ] 20.8 65.2002 | 20.8 65.1987 [ b 20.8 65.1978 | 20.8 65. 1972 J In a series of weighings the values varied in this manner, then, after a period of rest, as allowing to stand over night, the weight returned practically to the original value. Series (a) was obtained one day and series (b) on the following day. The variation was sometimes in one direction and sometimes in the other. A number of determinations was made to ascertain if constant Specific Gravities of Niobium and Tantalum Pentoxides 11 results could be obtained by washing the hydroxide with water alone. It was thought that the sulphuric acid could be so far removed as not to cause appreciable error, even though it is practically impossible to remove all of it. The table with the results is given and the detailed description of the conditions of each experiment follows: Weight Flask Weight Flask Weight Tempera- No, and Water. and Oxide. Oxide. ture. Gravity. 1 65.1918 65.6066 0.5411 23.0 4.284 2 65.1945 65.8617 0.8659 22.2 4.357 3 65.2020 65.9425 0.9570 21.2 4.420 4 65.1893 66.1641 1.2661 23.2 4.344 5 65.2003 66.1697 1.2774 20.8 4.147 6 65.2026 66.2168 1.3258 20.4 4.254 7 65.1913 66.1797 1.2996 23.0 4.176 8 65.1978 66.1852 1.2847 21.4 4.322 9 65.1870 66.0078 1.0662 22.9 4.344 10 65.1834 66.0211 1.0738 23.7 4.459 11 65.1974 66.0128 1.0513 20.5 4.456 (1) About 1 gm. K 2 NbOFs.H 2 was heated with sulphuric acid until fumes of 80s were given off, and the solution poured, with stirring, into 700 cc. hot water. This was boiled ten min- utes, allowed to settle and boiled with fresh portions of water for the same length of time twice more. The hydroxide was filtered and dried. A moderate blast was then applied for thirty minutes. (2) The same method of preparation was followed as in (1), except that the blast was used only twenty minutes. (3) The blast was here applied forty minutes and the method of preparation the same as in (1). (4) About 3 grams K 2 NbOF 5 .H 2 O were treated with 20 cc. sulphuric acid and the solution poured, with stirring, into two liters of hot water, boiled, and washed on the filter with one liter of boiling water. The hydroxide did not settle well. It was then dried, powdered and heated for one hour in a platinum crucible at a full red heat. (5) The same amount of K 2 NbOF5.H 2 was treated as in (4). 10 cc. sulphuric acid were used and the hydroxide washed with 12 Specific Gravities of Niobium and Tantalum Pentoxides two liters of hot water on the filter, after being boiled with one liter. The hydroxide settled well and washed easily. (6) The same amount of K2NbOF 5 .H 2 O as in (4) was treated with 10 cc. sulphuric acid, poured into two liters of hot water, boiled, filtered and washed with three liters of hot water. Dried and heated one hour in a platinum crucible at a full red heat. (7) Three grams K 2 NbOF 5 .H0 2 were treated with 12 cc. sul- phuric acid, hydrolyzed in three liters of hot water, boiled, decanted and boiled with the same amount four times, then fil- tered and dried. A moderate blast was applied to the oxide for ten minutes. (8) Three grams K 2 NbOFs.H 2 treated in the same manner as (7). (9) Two and a half grams K 2 NbOF 5 .H 2 treated with 12 cc. sulphuric acid, hydrolyzed in two liters of hot water and boiled three times with the same amount. The filter paper was burned off at a low temperature and a higher blast used than in (7) and (8) for ten minutes. (10) Two and a half KoNbOFs.H^O treated in the same manner as in (9). (11) Two and a half grams of K 2 NbOF5.H 2 treated in the same manner as in (9). In the foregoing experiments the oxide increased in weight on standing on the balance pan. Notwithstanding the thorough washing of the hydroxide and the prolonged and intense heating it is interesting to note that the oxides, when ground in an agate mortar, were invariably slightly sticky and adhered to the mortar, due, undoubtedly, to the attraction of moisture by a small amount of sulphuric acid still present in the oxide, which opinion seems to be further borne out in the subsequent experiments. Another series was made, using ammonium hydroxide to neu- tralize the sulphuric acid present. No. Weight Oxide. Temperature. Gravity. 12 1.0520 20.5 4.974 13 1.0772 20.6 4.936 14 1.0619 22.2 5.047 15 1.0635 21.5 4.914 16 .. . 1.0620 22.0 4.482 Specific Gravities of Niobium and Tantalum Pentoxides 13 (12) Two and a half grams K 2 NbOF 5 .H 2 O were treated with 12 cc. sulphuric acid, poured into one liter of hot water, boiled and allowed to settle. Boiled twice more with one liter of water, the last portion being made slightly alkaline with ammonium hydroxide before filtering. The oxide was heated with a Chad- dock burner until the paper was burned off, usually about forty- five minutes. The crucible was surrounded with a clay cylinder and the temperature regulated so that the bottom of the crucible was red. (13) Same as (12). (14) Two and a half grams K 2 NbOF 5 .H 2 treated with 12 cc. sulphuric acid, poured into one liter of hot water and washed twice by decantation with the same amount of hot water, boiling each time. No ammonium hydroxide was used in the last wash water. The same low heat was applied until the paper was incinerated, the oxide ground and placed in a crucible with concentrated ammonium hydroxide and heated again for twenty minutes at a low red heat. (15) Same as (14) except that the partially dried hydroxide was mixed with ammonium carbonate and heated in the usual manner. (16) Same as (7) except that after filtering the hydroxide was washed on the filter with strong ammonia and on account of the gelatinous character, the precipitate probably did not receive a thorough treatment with the alkali. In these experiments the oxide did not gain in weight, except that of (16), which apparently still contained some sulphuric acid. In the following work the last portion of water used in washing the hydroxide was always made distinctly alkaline before filtering. In the following determinations about 2.5 grams K 2 NbOFs.H 2 were treated with 12 cc. sulphuric acid, poured into one liter of hot water, boiled, washed twice by boiling and decantation, the last wash water made alkaline with ammonium hydroxide. The hydroxide, after drying at a low temperature, was heated for one hour at a low red heat. 14 Specific Gravities of Niobium and Tantalum Pentoxides No. Weight- Oxide. Temperature. Gravity. 17 1.0870 22.7 4.954 18 1.0221 20.6 4.973 19 1.0163 21.6 4.936 20 1.1354 23.3 4.932 21 1.1461 23.5 4.950 22 1.1051 25.3 4.962 23 1.1097 24.55 4.960 24 1.1133 24.8 4.952 The average of these determinations is 4.949, which repre- sents the value of the specific gravity of Nb20s under these con- ditions. For the corresponding value of the tantalum oxide the same conditions were used in the preparation of the oxide, ignition, etc., as in the case of the niobium oxide. No. Weight Oxide. Temperature. Gravity. 25 ...1.1745 .22.9 7.909 26 1.1883 23.3 7.909 27 1.1600 23.4 7.854 28 1.2207 22.9 7.916 29 1.1910 23.7 7.856 30 1.2060 23.6 7.805 31 1.1543 22.7 7.857 These results give an average of 7.872 as the specific gravity of Ta20s under these conditions. On account of the smaller displacement, the variation is greater in these determinations than that in the case of niobium. Knowing that difference in the length of time and intensity of the heat applied makes a difference in the specific gravity, it was decided to make a series of determinations, using a higher temperature for the same length of time as above. The prepa- ration of the sample was the same as before, but the oxide was heated with the full flame of the Chaddock burner for one hour. This gave a bright orange heat. No. Weight Oxide. . Temperature. Gravity. 32 1.1573 22.2 8.362 33 1.1588 21.9 8.277 34 1.1574 22.1 8.362 Specific Gravities of Niobium and Tantalum Pentoxides 15 No. Weight Oxide. Temperature. Gravity. 35 ............. 1.1485 22.3 8.310 36 ........... . . 1.1512 22.4 8.294 37 ............. 1.1592 22.5 8.303 38 ............. 1.1263 21.9 8.373 39 ............. 1.1604 22.0 8.371 40 ............. 1.1718 21.8 8.281 41 ............. 1.1754 21.9 8.353 This gives an average of 8.328 as the specific gravity of at a higher temperature. An exactly similar series was made on niobium pentoxide to find the value under the above conditions. No. Weight Oxide. Temperature. Gravity. 42 ......... ____ 1.1209 23.2 4.844 43 ............. 1.0951 23.1 4.849 44 ............. 1.0778 23.0 4.807 45 ............. 1.1094 23.8 4.802 46 ............. 1.0865 24.2 4.835 47 ............. 1.0867 23.2 4.806 48 ............. 1.1215 23.4 4.852 These results give an average of 4.828. This would seem to indicate that with increasing heat the specific gravity of the tantalum oxide increases, whereas that of the niobium oxide decreases. It was thought possible that there might be slight differences in the hydrolysis in the separate experiments above, which would affect the results, therefore, two larger portions of about double the quantity of salt were taken and treated as usual. After the hydroxides had been obtained, each of these was divided into two parts. Each pair of hydroxides was ignited separately under the same conditions, i. e., the first pair at a full red heat for one and a half hours, the second pair at the same temperature for two and a half hours. No. Weight Oxide. Temperature. Gravity. 49 ............. 1.1017 23.6 4.702 Second half ..... 1.1134 23.6 4.770 50 ............. 1.1713 22.9 4.370 Second half.. .1.1505 22.9 4.329 16 Specific Gravities of Niobium and Tantalum Penioxides Each pair shows no greater disagreement than obtains where the portions were prepared separately. The length of time of heating, however, has a marked influence on the values. Mixtures Using the average values obtained for tantalum and niobium at a full red heat, 8.328 and 4.828, a curve was plotted and mix- tures made to determine whether the per cent, of tantalum and niobium could be obtained in this manner. The pure oxides, or mixtures of the same, are usuall)' very slightly grayish in color when prepared by fusing the oxides with potassium bisuiphate, hydrolyzing, etc., but those prepared from the double fluorides are white. This was thought due to the presence of a slight amount of platinum. An effort was made to make them whiter by dissolving the hydroxides in hydrofluoric acid, filtering, taking down to fumes again with sulphuric acid and rehydrolyzing, but the result ^as the same. The color is nearly always the same, and if there is any error due to this, it would be constant. In preparing the mixtures, weighed amounts of the oxides were fused with potassium bisuiphate, sulphuric acid added to make the melt more fluid, hydrolyzed and washed by decantation through a filter, etc., as already described. After drying, the hydroxides were heated for one hour at the full temperature of the Chaddock burner, pulverized, weighed and the specific gravity taken. The mixture had the following proportions: Nb2C>5, 0.5897 gm., and Ta205, 0.6899 gm., corresponding to 46.08 per cent. Nb20s and 53.92 per cent. Ta2O5- The specific gravity was 6.345, which, according to the curve, gave 56.50 per cent. Ta20s, an error of 2.58 per cent. 1 PART II On account of many difficulties encountered when water was used as the liquid in which the specific gravity was taken, some other liquid was sought. The purpose of heating of the oxides is to dehydrate the hydroxides, and on being immersed in water, !The curve is not shown here for this single experiment. UNIVERSITY OF Specific Gravities of Niobium and Tantalum Pentoxides 17 it is probable that a second determination on the same sample would vary on account of differing degrees of rehydration. If some liquid were used which had no effect of this kind, and in which the finely powdered oxides settled readily and were easily filtered, this effect could be avoided and a number of determina- tions made upon the same portion, the period of heating in each determination being simply a continuation of the preceding and the effect cumulative. A number of such liquids were tried, including bromoform, benzol, xylol, chloroform, etc. Chloroform was finally selected, having a comparatively low boiling point and easily handled. On account of the low boiling point, the oxides were very easily dried. A number of methods were tried to obtain the most con- stant results. The co-efficient of expansion of chloroform being larger than that of water and its specific gravity greater, it was necessary to maintain as constant temperature conditions as possible. One method used was to employ a large flask filled with chloro- form at room temperature. From this the chloroform was forced by air pressure into the gravity flask, in order to avoid handling; a thermometer was kept in the chloroform at all times, but the variations in weight were too great for even as accurate work as in the case where water was used, and this method was discarded. A second method was to use a constant temperature bath, as follows: A large, three-liter beaker served as the bath, being well protected by cloth and felt. A smaller beaker, of tall form, about 700 cc. capacity, was suspended in the center of the larger one by a wooden cover, provided with a hole of proper size. The smaller beaker was nearly filled with chloroform and the larger beaker filled with water. Before each determination, the gravity flask was immersed in the chloroform and the liquid agitated until the temperature was constant. The flask was filled and the stopper inserted while still in the chloroform vapor. It was found impossible to obtain concordant weights for the flask filled with chloroform in this way, and the method was aban- doned. A third method was decided upon and used throughout the 18 Specific Gravities of Niobium and Tantalum Penioxides remainder of the work. To obtain a constant temperature a bath of boiling chloroform was used and the flask placed in its vapor. The chloroform was boiled in a Victor Meyer vapor density tube and the gravity flask suspended in a wire basket about an inch and a half above the boiling chloroform. A ther- mometer, graduated to hundredths of a degree C., was suspended in the tube to record temperatures. Platinum wire was used to insure uniform boiling and the heat was supplied by an electric hot plate. No reflux condenser was needed, as the vapor was all con- densed about three or four inches from the top of the tube The temperature of the vapor varied about 0.6 on different days (from 60.40 to 61.03), depending upon the barometric pressure. This, however, did not affect the accuracy of the weights, since the two required for a determination were taken in comparatively short intervals of time, and the barometric change in that time was negligible. A gravity flask with a long and well-fitted stopper, but with- out a vacuum jacket, was used. Determinations were made as follows: The flask filled with chloroform was heated for ten minutes in the bath, allowed to cool ten minutes and weighed, the cap being placed on the flask as soon as the small amount of chloroform on the outside had evaporated. This was prac- tically accomplished by the time the flask had been removed from the tube. The flask was then emptied and a weighed amount of oxide introduced. It was partially filled with chloro- form and again heated. This served to remove any air that might be included in the powdered oxide. The flask was then re- moved from the bath, shaken with a rotary motion, filled, and the stopper replaced, again suspended in the bath and the heat- ing continued for ten minutes longer. The weight was taken after allowing to cool ten minutes. It is important that the flask be only partially filled when suspended in the bath for the first time, for if completely filled, vapor invariably forms inside the flask, causing serious errors. It is equally important that the lower surface of the stopper be rounded and well polished. It is also required that the oxide be dried on the filter paper and then removed before reheating Specific Gravities of Niobium and Tantalum Pentoxides 19 for another determination. If the paper is incinerated with the oxide, vapor always forms in the flask when this oxide is again employed for a gravity determination. The average of four weights of the flask filled with chloroform and an average of the same number of weights with the flask filled with water in the same bath was made. The weight of the flask alone being known, the ratio of chloroform to water at this temperature (60.80 C.) was found to be 1.432. The material for the following experiments was prepared in a somewhat different manner from that of the first part. A large quantity of niobium double fluoride was treated in small lots until about 50 grams of the pure oxide was prepared. This was then placed in a large crucible and heated for one hour at a full red heat, avoiding thereby the probability of separate portions differing in specific gravity on account of slight but unavoidable differences in preparation. A number of determinations was made on the oxide prepared as just described. It was thought at first that a weight could be determined upon by a series of weighings which would be constant for the flask filled with chloroform. But it was found that a separate weighing must be made for each determination, as before. This was due to the difference in barometric pressure on different days and the same gradual increase in weight of the flask for a time and after a period of rest, a return to nearly the original weight, as mentioned in the case where water was used (page 10). Gravity Weight Flask Weight Flask Weight Referred No. and CHC1 3 . and Oxide. Oxide. to CHC1 3 . 51 30.2674 30.9878 1.0220 3.388 52 30.2692 30.9826 1.0127 3.383 53 30.2683 30.9818 1.0134 3.380 54 30.2683 30.9955 1.0330 3.375 55 30.2674 30.9792 1.0160 3.340 56 30.2684 30.9806 1.0176 3.340 57 30.2670 30.9757 1.0123 3.334 The last three results were made upon a different lot of oxide r prepared, however, as nearly as possible in the same manner. Another sample of oxide, prepared as nearly as possible as before, gave the following results: 20 Specific Gravities of Niobium and Tantalum Pentoxides Gravity Referred No. Weight Oxide. to CHC1 3 . 58 1.0138 3.238 59 1.0192 3.248 60 1.0151 3.227 61 1.0400 3.224 62 1.0337 3.236 63 1.0349 3.244 64 1.5238 3.234 65 2.0475 3.224 These results show a maximum difference of 0.024, which was regarded sufficiently accurate for this work. The results on the three different samples (Nos. 51-54, 55-57 and 58-65) differing as they did, showed beyond question that the method of prepa- ration and the length of time of heating the oxide were ver} r important factors in obtaining a constant specific gravity. This led to a series of determinations on one sample, heating a specified time before each determination, to find if any point could be reached where the gravity remained practically con- stant. For this purpose a quantity of the niobium pentoxide was prepared in the usual way in small quantities and the entire lot thoroughly mixed and heated for one hour at a full red heat. From this lot the portions for the following experiments were taken. Each group represents the determinations on one sample until the gravity remained practically constant. The temperature employed for these ignitions was a full red heat. Weight Time of Gravity Referred No. Oxide. Heating. to CHC1 3 . 66 1.2230 20 mins. 3.186 67 1 .2008 20 " more. 3. 178 68 1.1956 25 " " 3.166 69 1.1822 25 " " 3.152-> 70 1.1829 25 " " 3.158/ A 71 1.2339 20 mins. 3.188 72 1.2298 20 " more. 3.171 73 1.2114 30 " " 3.166 74 1.2055 25 " " 3 ' 159 \D 75 .. . 1.2026 20 " " 3.159J B Specific Gravities of Niobium and Tantalum Pentoxides 21 Weight No. Oxide. 76 1.2361 77 1.2415 78 1 . 2354 79 1.2243 80 1.2223 81 1.2145 82 1.2094 83 1.1987 84 1 . 1851 85 .. . 1.1861 Time of Heating. 50 mins. 20 " more. 20 " 30 " 20 " 20 mins. 20 " more. 30 " 20 " 20 " Gravity Referred to CHC1 3 . 3.162 3.159 3.161 3.156 j 3.160J 3.183 3.178 3.159^ 3.162 3.159' C In each of these groups the gravity gradually diminished until it became practically constant. In the next experiments two portions of about three grams each of K^NbOFs.H^O were taken and the oxide prepared in the usual way. Determinations were made on these to see if the gravity would reach approximately the same value obtained above. Gravity Referred to CHC1 3 . 3.300 3.190 3.161 > 3.159 IE 3.159J Weight No. Oxide. 86 1.2855 87 1.2658 88 1 . 2555 89 1.2473 90 1.2375 91 0.9570 92 0.9304 93 0.9079 94 0.9175 95 .. . 0.9141 Time of Heating. 30 mins. 35 " more. 20 " 20 " 20 " 20 mins. 25 " more. 30 " 30 " 20 " 3.346 3.242 3.174 3.161 3.158 The final values obtained in the last two groups agree fairly well with those obtained above and were included in the general average. The results marked A, B, C, D, E and F were assumed to be constant and the average gave 3.159. Later in the work the gravity of niobium was re-determined to see if the results obtained above could be duplicated under the same conditions. An average of three closely-agreeing values gave 3.161. 22 Specific Gravities of Niobium and Tantalum Pentoxides Tantalum Pentoxide The method employed for the preparation of the tantalum pentoxide was the same as that described for the niobium pent- oxide. Each of the groups shown below represents determina- tions on a separate portion of oxide taken from one stock supply. Weight Time of Gravity Referred No. Oxide. Heating. to CHC1 3 . 96 1.2236 40 mins. 6.191 97 1.1985 20 " more. 6.127 98 1.1900 20 " " 6.140 99 1.1852 20 " " 6.0971 100 1.1775 20 " " 6.084J A 101 1.2010 40 mins. 6.062 102 1.1943 20 " more. 6.093 103 1.1895 20 " " 6.205 104 1.2841 20 " " 6.0851 105 1.1760 20 " " 6.090/ B 106 1.2117 20 mins. 6.098 107 1 .2001 20 " more. 6.097 108 1.1828 20 " " 6.079 109 1.1824 20 " " 6.0671 110 1.1680 20 " " 6.055/ C 111 1.2158 40 mins. 6.018 112 1.2004 20" more. 6.096 113 1.1872 20 " " 6.057 114 1.1789 20 " " 6.077 115 1.2201 40 mins. 6.061] 116 1.2019 20 " more. 6.070 117 1.1855 20 " " 6.061 [ E 118 1.1784 20 " " 6.074 j 119 1.1766 20 " " 6.087 | 120 1.1644 20 " " 6.064J The agreement is not as satisfactory as in the case of the niobium, but those marked A, B, C, D and E are taken as con- stant and averaged. This gives 6.073 as the value of tantalum pentoxide under these conditions. This value, with that of nio- bium pentoxide, 3.159, was used to plot a curve. (Curve I.) Specific Gravities of Niobium and Tantalum Pentoxides 23 Mixtures of the two pure oxides were now made, fused with potassium bisulphate, and the oxides obtained in the usual manner. Determinations were made on these mixtures in the same way as those above until the specific gravity became more or less constant. No. Weight Ta 2 5 . Weight Nb 2 5 . Per Cent. Per Cent. Nb 2 5 Weight Oxide. Time of Heating. Gravity Referred to CHCls Nb 2 5 from Curve. Error. 121 0.1224 1.0826 10.15 89.85 1 . 2044 30 mins. 3.575 1.1892 30 " more. 3.565 1.1732 30 " 3.517 1.1246 40 " 3.481 89.00% 0.85% 122 0.2417 0.9617 20.09 79.91 1.0515 1.0289 40 mins. 20 " more. 3.8361 3. 844 I 76.60% 3.31% 123 0.3614 0.8416 30.05 69.95 1 . 1952 30 mins. 3.962 72.50% 2.55% 124 0.4813 0.7211 40.03 59.9? 1.1518 1.1293 30 mins. 30 " more. 4.209) 4.201J 64.20% 4.23% 125 0.1218 1.0808 10.12 89.88 1.1348 20 mins. 3.431 1 . 1054 30 " more. 3.367 92.80% 2.92% 126 0.1219 1.0813 10.14 89.86 1.2032 80 mins. 3.343 1 . 1845 20 " more. 3.369 1.1757 40 " 3.352 93.40% 3.54% 127 0.4059 0.7962 33.77 66.23 1.2021 80 mins. 3.794 1.1868 20 " more. 3.807 77.60% 11.37% (121) Before these determinations were made, the mixture of oxides was heated for one hour at a full red heat. (122) This mixture was prepared as in (121) with the excep- tion that after heating for one and a half hours the oxides were ground and boiled with water, to remove any possible soluble impurity. On filtering, the powder was too fine and ran through the filter to some extent. It was washed with alcohol, dried and taken from the paper. Heat was applied for twenty min- CURVE I One Division = 0.02 Specific Gravity or !.0 per cent. Specific Gravities of Niobium and Tantalum Pentoxides 25 utes, the oxides cooled and placed in a desiccator for four hours, heated again for twenty minutes, and the first determination made. (123) This mixture was treated as before until the oxides were again obtained. These were then heated for one hour at a full red heat, cooled in a desiccator, again heated for twenty minutes and the determination made. (124) In this case, after obtaining the hydroxides the first time, they were dissolved in hydrofluoric acid, filtered, the solu- tion taken down to fumes with sulphuric acid and rehydrolyzed. The oxides were dried, heated one and a half hours, pulverized, heated again for thirty minutes and the gravity taken. (125) After the hydroxides had been obtained by fusion with potassium bisulphate and hydrolyzing they were dissolved in hydrofluoric acid, filtered, taken down to fumes with sulphuric acid, again hydrolyzed, washed and dried as usual. Heat was applied for one and a half hours, the oxides powdered and heated again for 20 minutes and the first determination made. (126) In this case the oxides were merely mixed and not fused with potassium bisulphate. This should have given results according to the curve, being a simple mixture. Before the first determination they were heated for one hour and twenty min- utes. (127) This experiment was of somewhat the same character as 126, but in this case the oxides were heated separately for one hour and twenty minutes, weighed, mixed, weighed again and the specific gravity taken. In the second determination the two. mixed, were heated for twenty minutes and the specific grav- ity taken. These results are unreliable from a quantitative point of view and could not be depended upon for the estimation of Nb20s and Ta2C>5 in mixtures. In order to determine if any change in weight took place on ignition of the oxides, quantities of each were taken, ignited one hour and weighed. They were then ignited one hour longer (Chaddock burner) and the loss in weight determined. 1.0026 grams Nb 2 O 5 lost 0.0028 gram and 1.0011 grams Ta 2 O 5 lost 0.0012 gram. The crucible itself showed no loss in weight. 26 Specific Gravities of Niobium and Tantalum Pentoxides Similarly, 4.0357 grams Ta20s ; which had been ignited with the blast lamp for fifteen minutes, lost 0.0010 gram on blasting for twenty-five minutes longer, and 5.0083 grams Nb20s lost 0.0045 gram on blasting for twenty-five minutes. III It was next decided to try the effect of the blast lamp on the gravity of the pure oxides, in order to establish a new basis for a curve to use with a series of mixtures, hoping by this means to obtain constant results. Determinations were made on two portions of niobium pent- oxide, blasting each time, until the gravity became practically constant. Weight Time of Gravity Referred No. Oxide. Heating. to CHC1 3 . 128 ............ 1.4951 20 mins. 3.170 129 ............ 1.4752 20 " more. 3.230 130 ............ 1.4646 20 " " 3.227 131 ............ 1.4545 20 " " 3.262] 132 ............ 1.4479 20 " " 3.2621 133... ......... 1.4396 20 " " 3.268 [ 134. . 1.4357 ' 20 " " 3.283'' 135 1.4270 20 3.262 136 1.4273 20 " " 3.280J 137 1.4945 20 mins. 3.228} 138 1 .4848 20 " more. 3.255 > B 139 1.4715 20 " " 3. 255-* 140 1.4629 20 " " 3.263 The average of those marked A and B is 3.265. Determinations were made in the same manner on tantalum pentoxide and gave the following results: Weight Time of Gravity Referred No. Oxide. Heating. to CHC1 . 141 1 .4528 20 mins. 6.2061 142 1 .4309 20 " more. 6. 189 143 1.4211 20 " " 6.212 144 1.4122 20 " " 6.226 j-A 145 1.4064 20 " " 6.236 146 1.3935 20 " " 6.204 147 1.3909 20 " " 6.195j Specific Gravities of Niobium and Tantalum Pentoxides 27 No. 148 Weight Oxide. 1 4971 149 1 . 4735 150 1 4695 151 1 . 4622 152 . . . ... 1 4562 153 1 4504 154 155 . . 1.4471 . 1.4404 Time of Heating. 20 mins. 20 " 20 " 20 " 20 " 20 " 20 "' 20 " Gravity Referred tO*CHCl 3 . 6.183 6.155 6.166 6.1961 6.194 I 6.193 I-B 6.218 j 6.203J Those marked A and B were averaged and gave 6.201. The values just obtained for tantalum and niobium oxides were used in plotting Curve II. A series of mixtures was now made, nine in all, from 10 per cent, to 90 per cent, tantalum. These were heated with the full heat of the Chaddock burner until a practically constant gravity was obtained, then to the same mixture the blast was applied for definite periods until another constant value was reached. The first part of each group represents the determina- tions with the lower heat and the second part the results obtained by blasting. No. Weight. Per Cent. Weight Oxide. Time of Heating. Gravity Referred to CHC13 Aver- age. Ta 2 5 Nb 2 5 Ta 2 5 Nb 2 5 156 0.1228 1.0812 10.20 89.80 1.1637 90 mins. 3.500 .1083 30 " more. 3.467 .0942 40 " 3.304 .0825 110 " 3.240 .0677 20 " 3.220 .0725 30 " 3.225 Blast 1.0715 15 " 3.356 157 1.0699 1.0562 30 " 15 " 3.3151 3.317 J 3.316 0.2415 1.0812 20.06 79.94 1.1497 90 mins. 3.890 1.1286 60 " more. 3.758 1.1192 20 " 3.699 1 . 1029 60 " " 3.523 1.0935 30 " 3.516 28 Specific Gravities of Niobium and Tantalum Pentoxides Weight. Per Cent. Gravity No. Weight Time of Referred Aver- Ta 2 5 Nb 2 5 Ta 2 5 Nb 2 5 Oxide. Heating. to CHC1 3 age. 1.0995 60 " 3.488 1.0858 30 " 3.664 1.0726 40 " 3.448 1.0410 20 " 3.418 0.0454 20 " " 3.453 1.0451 30 " " 3.432 1.0487 30 " 3.432 Blast 1.0480 15 " 3.414 1.0487 15 " 3.438 1.0459 15 " 3.464 0.0447 15 " 3.488 1.0377 15 " 3.497 1.0367 1.0257 15 " 15 " 3.4811 3.487 1 3. 434 158 0.3627 0.8420 30.10 69.90 1.1540 120 mins. 4.157 1.1396 40 " more. 4.200 1.1330 50 " 4.177 1.1231 20 " 4.182 1.1217 20 " 4.154 1.1117 20 " 4.184 1 . 1028 20 " 4.186 Blast 1.0931 15 " 3.378 1.0849 15 " 3.648 1.0747 15 " 3.656 1.0696 1.0682 15 " 15 " 3.6931 3.680/ 3.686 159 0.4816 0.7209 40.05 59.95 1.1513 90 mins. 4.440 1.1303 20 " more. 4.422 1 . 1231 60 " 4.420 Blast 1.1103 15 " 4.017 1.0961 15 " 3.904^ 1.0852 15 " 3.919 I 3.908 1.0808 15 " 3.901 J Specific Gravities of Niobium and Tantalum Pentoxides 29 Weight. Per Cent. 1 Gravity No. Weight Time of Relerred Ta 2 5 Nb2Os T 2 aOs Nb 2 5 Oxide. Heating. to CHCl.s age. 160 0.6016 0.6015 50.02 49.98 1 . 1404 120 mins. 4.583 1.1182 30 " more. 4.642 1.1017 30 " 4.642 1.0910 20 " 4.644 Blast 1.0884 15 " 4.366 1.0875 15 " 4.308 1.0881 15 " 4 . 282 1.0835 15 " 4 . 148 1.0781 15 " 4.169 1.0792 15 " 4.163 1.0780 15 " 4.142 1.0736 15 " 4.080 1.0721 1.0715 15 " 15 " 4.0741 4.074 / 4.074 161 0.7210 0.4810 59.98 40.02 1.1592 1.1414 120 mins. 20 " more. 4.786 4.857 1.1293 20 " 4.878 1.1129 20 " 4.868 Blast 1.1136 15 " 4.848 1.1146 15 " 4.903 1 . 1090 20 " 4.926^, 1 . 1006 1.0988 15 " 15 " 4.926 i 4.916 f 4.925 1.0947 15 " 4.93lJ 162 0.8411 0.3616 69.94 30.06 1.1323 90 mins. 5.123 1.1163 20 " more. 5.189 1 . 1055 20 " 5.202 1.0992 30 " 5.194 1.0895 60 " 5.271 1.0821 30 " 5.253 1.0781 20 " 5.266 Blast 1.0704 15 " 5.221-j 1.0697 15 " 5.223[ 5.227 1.0637 15 " 5.237^ 30 Specific Gravities of Niobium and Tantalum Pentoxides No. 163 Weight. Per Cent. Weight Oxide. Time of Heating. Gravity Referred to CHC13 Aver- ' age. Ta 2 5 Nb 2 5 Ta2Os Nb2Os 0.9612 0.2410 79.95 20.05 1.1102 90 mins. 5.178 1.0883 20 " more. 5.301 1.0814 20 " 5.337 1.0796 50 " 5.363 1.0681 80 " 5.460 1.0588 40 " 5.443 1.0573 30 " 5.469 Blast 1.0484 15 " 5.454 1.0452 15 " 5.486 1.0416 15 " 5.522 1.0354 30 " 5.475 1.0342 15 " 5.530 5.530 164 1.0823 0.1229 89.80 10.20 1.1360 120 mins. 5.601 1.1126 30 " more. 5.816 1.1015 20 " " 5.773 1.0967 30 " 5.808 1.0857 30 " 5.774 Blast 1.0803 15 " 5.827 1.0819 1.0788 1.0709 15 " 15 " 15 " 5.845} 5.838h 5.845 J 5.843 Although the values obtained on these mixtures do not fall on the curve obtained from the pure oxides, nevertheless, they are quite regular, the greatest variation among themselves being about one per cent. It is interesting to note the sudden break in the curve between the 40 and 50 per cent, mixtures. Whether or not this break is due to the formation of a compound of the two oxides the author was unable to determine. In view of the fact that the curve obtained on these mixed oxides was so regular, it was hoped that satisfactory results might be obtained on the analysis of a sample of columbite. The mineral used was the same "Columbite A" analyzed volu- metrically by Metzger and Taylor, 1 which gave an average value 1 To be published soon. 3.465 3.265 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% CURVE II One Division = 0.02 Specific Gravity or 1.0 per cent. 32 Specific Gravities of Niobium and Tantalum Pentoxides of 54.21 per cent, niobium pentoxide and 22.75 per cent, tan- talum pentoxide, which means that the mixed oxides would be composed of 70.43 per cent, niobium pentoxide and 29.57 per cent, tantalum pentoxide. Three different portions were treated, as follows: The ore was fused with potassium bisulphate and the melt boiled with about one liter of water. The impure hydroxides so obtained were filtered, washed with boiling water until the filtrate gave only a faint test for sulphuric acid, treated on the filter with yellow ammonium sulphide, washed again and treated with hot dilute sulphuric acid. The washed hydroxides were dissolved through the filter paper with hydrofluoric acid, the solution taken down to furnes with sulphuric acid, rehydrolyzed, boiled with water and washed by decantation twice, the last wash water being made alkaline with ammonium hydroxide. The hydroxides were then filtered, dried, blasted for definite periods and the gravity taken. Amount Weight Per Cent. Gravity No. Ore of oi Weight Time of Referred Aver- Taken. Oxides. Oxides. Oxides. Heating. to CHCls- age. 165 2.0023 1.5458 77 . 25 1 . 5266 30 mins. 3.487 1.5011 20 " more. 3.514 .4929 20 " 3.560 .4857 20 " 3.578 .4764 20 " 3.587 .4692 40 " 3.592 .4656 20 " 3.610] 1.4620 1.4586 1.4568 1.4502 20 " 25 " 20 " 20 " 3.618 3.616 } 3.600 I 3.618J 3.612 1G6 2.0019 1.5446 77.15 1.5279 20 mins. 3.479 1.5159 40 " more. 3.519 1.5052 20 " 3.553 1.5049 20 " 3.574 1.4981 20 " 3.599 1.4927 20 " 3.599 1.4938 1.4910 1.4872 25 " 20 " 20 " 3.617} 3.612 [ 3.620^ 3.616 Specific Graviteis of Niobium and Tantalum Pentoxides 33 Amount Weight Per Cent. Gravity No. Ore of of Weight Time of Referred Aver- Taken. Oxides. Oxides. Oxides. Heating. to CHC1. age. 167 2.0000 1.5382 76.91 1.5271 20 mins. 3.282 1.5081 20 " more. 3.462 1.4985 20 " 3.543 1.4908 75 " " 3.572) 1.4918 20 " 3.580 f 3.577 1.4840 40 " 3.578' (165) According to the curve, the value 3.612 gives 74.4 per cent. Nb2O5, and calculating from this, the ore contains 57.43 per cent. Nb2O5, an error of 3.22 per cent. (166) This value, 3.616, is very close to that of 165 and the per cent, of Nb20s cannot be read closer than 74.4 per cent, on the curve. This gives 57.47 per cent. N^Os in the ore. an error of 3.24 per cent. (167) The value 3.577 gives 76.4 per cent, as the amount of Nb2Os in the mixture. This is equal to 58.75 per cent. Nb2Os in the ore, an error of 4.54 per cent. CONCLUSIONS. The mixtures of oxides, as prepared by fusion with potassium bisulphate, hydrolysis, etc., do not follow the straight line of the curve. The specific gravity of the oxides varies slightly with the method of preparation and greatly with the length of time and intensity of the heating. As a method for the determination of the percentages of niobium and tantalum in a mixture of the two oxides, it can only be relied upon to an accuracy of about five per cent. In general, the specific gravity of niobium pentoxide decreases with increase in temperature and length of time of heating, while that of tantalum pentoxide increases under the same conditions. The average values of the specific gravity of niobium and tan- talum pentoxides under the different conditions tried are: 34 Specific Gravities of Niobium and Tantalum Pentoxides Referred to Referred to NIOBIUM PENTOXIDE. water. chloroform. At a low red heat 4 . 949 At a bright red heat 4 . 828 3 . 159 At temperature of blast 3 . 265 TANTALUM PENTOXIDE. At a low red heat 7 . 872 At a bright red heat 8 . 328 6 . 073 At temperature of blast 0.201 THE UNIVERSITY THIS f-XSSft jjf: V^ 20m--