Recovery of Uranium

In the extraction process practically all of the uranium in the ore is dissolved in the nitric acid, a little remaining in the insoluble residue because of incomplete washing. An average of 2.3 per cent of the uranium oxide in the ore has remained in the residue, varying in different carload lots from a mere trace to about twice the average. A more thorough washing would remove practically all of this uranium, but would considerably increase the volume of liquid to be handled and evaporated.

As already indicated (p. 50) the loss in the iron-calcium precipitate precipitate has varied, with the amount of sodium carbonate used and the length of time of heating the solution. In the early stages of the work most of the precipitate contained 1½ to 2 per cent U3O8 on the dry weight, or 15 to 20 per cent of the oxide in the original ore. This loss was afterwards considerably reduced, mainly by boiling the solution for a longer time and washing the precipitate more thoroughly. The iron-calcium precipitate from some of the carload lots contained as little as 0.48 to 0.50 per cent U3O8, representing about 5 per cent of the oxide in the original ore. The U3O8 content of the iron-calcium precipitate obtained from the last six carload lots treated averaged 0.7 per cent.

The original plan (see p. 52) called for a double precipitation in order to recover this uranium and the associated radium and vanadium, but it was found that the cost exceeded the value of the products obtained. All efforts were then bent toward raising the yields without reprecipitation.

If, after the precipitation of the sodium uranate, the solution is heated for one hour, practically all of the uranium is precipitated, only a trace appearing with the iron vanadate.

The recovery of the uranium as sodium uranate has, of course, varied with the losses, the extremes being between 75 and 94 per cent. The average on the last 10-carload lots treated has been 84.4 per cent.

Recovery of Vanadium

The process described in this bulletin could not be recommended were the recovery of the vanadium in the ore the main object. Hydrochloric acid, under plant conditions, will extract the vanadium from carnotite more efficiently than will nitric acid, because the reducing action of hydrochloric acid prevents the vanadium from separating as vanadic acid. In this respect even sulphuric acid is better. The chief recommendation for using nitric acid is, therefore, the recovery of radium rather than vanadium. Indeed, it was found in the early operations that any attempt to extract the larger part of the vanadium almost invariably resulted in the precipitation of vanadic acid, retarding filtration and reducing the recovery of radium. Carnotite itself is readily soluble in nitric acid, but the other vanadium minerals present, especially the silicates, are decomposed only after long boiling. A considerable loss of vanadium, therefore, takes place at the start, owing to the insolubility of the vanadium minerals other than carnotite. In order to prevent the separation of vanadic acid, the ore has been heated only as long as the solution remained green, heating being discontinued at once if the solution has showed any tendency to become brown. The main object, therefore, has been to recover radium even at the sacrifice of some of the vanadium. The vanadium content of the ore used has varied from 3.46 to 5.43 per cent V2O5, and ore carrying less than 5 per cent presents no difficulties in treatment; however, if this proportion is largely exceeded, the rapidity of filtration is affected, resulting in some loss of radium.

Under the above conditions only about 45 per cent of the vanadium in the ore goes into solution. The iron-calcium precipitate involves a further loss. If sodium carbonate is run into a carnotite acid leach, so that the liquid is acid during most of the time, the vanadium losses will be large even though the carbonate is finally added in excess. Iron vanadates are seemingly formed near the neutral point and after precipitation are redissolved only slowly by the carbonate. In a small experimental plant that used this method to separate the uranium and vanadium from the iron and calcium the iron precipitate obtained carried 15 to 20 per cent V2O5. However, if the acid liquor is run into a boiling solution containing an excess of sodium carbonate, as already described, the iron-calcium precipitate obtained need not average more than 2 per cent V2O5, and the proportion may be reduced even to 1 to 1.25 per cent under favorable conditions. As explained on page 55, the ruling factor is the length of time the solution is boiled and how much excess carbonate is used. Of course, a point is finally reached where the additional recovery does not pay for the increased expense.

The sodium uranate has carried on an average 8.1 per cent V2O5, the limits being 5.2 to 9.4 per cent. If the salt fusion method is used to refine the sodium uranate, practically all of this vanadium is recovered, as the refined sodium uranate on an average carries less than 0.2 per cent V2O5. The iron vanadate obtained in the recovery of this vanadium is of exceedingly high grade, most of it containing about 45 per cent V2O5. Seemingly, in the melt the vanadium is completely oxidized to meta-vanadate.

The sodium nitrate filtrate from the iron vanadate may or may not contain vanadium, depending on the method of precipitation. If the precipitation is properly done, the sodium nitrate will carry only a mere trace of vanadium, hardly enough to give a qualitative test.

Lately, on an average 55.5 per cent of the vanadium in the ore remains in the residue, and 13.6 per cent remains in the iron-calcium precipitate, while 8.1 per cent appears in the sodium uranate and 21.4 per cent in the iron vanadate. The total average recovery in vanadium, including that from the sodium uranate is therefore a little less than 30 per cent.

Recovery of Radium

When an element exists in an ore in the proportion of 1 part to 200,000,000, its extraction and recovery present difficulties not ordinarily encountered in metallurgy. A recovery of 60 to 70 per cent or even 50 per cent might, under such conditions, appear to be satisfactory. A much larger recovery than 70 per cent is undoubtedly exceptional. The unusually high recovery of 90 per cent and over of the radium present gives the nitric acid method its real value.

Tabulated Data

The table following gives the results of the extraction of the first 21 carloads of carnotite ore received at the plant of the National Radium Institute:

Discussion of Tabulated Data

The table shows that with the exception of the ore from two cars, the radium recovered as sulphate has varied from 80.4 to 96 per cent of the radium in the ore with an average of 89.6. If the two cars mentioned are included, the average is 88.3. The average recovery on the last ten carloads treated is 91.5.

In tabulating these results it has been assumed that the radium in the ore is in equilibrium with the uranium, as shown by Lind and Whittemore, “ and the radium in each carload has been calculated from the uranium oxide content. The radium in the sulphates has always been determined by actual measurement by the emanation method. Results have been accepted only when duplicates gave check results. Although individual sulphate batches were frequently tested, the actual determinations on carload recoveries were made on composites in which the quantity of sulphate used from each batch was proportionate to the total weight of that batch.

During the early part of the work the sulphates were ground and mixed by hand, and a small error in sampling was possible. Later, small ball mills were installed for grinding and thoroughly mixing the sulphates, eliminating the possible error mentioned.

Most of the filtrate from the first sulphate precipitate has carried a little more than 2 per cent of the radium in the ore, as is shown in the column giving the proportion of radium recovery as second sulphates, the filtrate from the second precipitate containing less than 0.1 per cent of the original radium. With four carloads the proportion given was exceeded, the second sulphate from the ore in car P-3 containing 6 per cent of the radium in the ore. The larger part of this was found in three batches and was undoubtedly due to incomplete settling of the first sulphates, a portion of the precipitates being siphoned over and finally appearing with the second sulphate. Indeed, a large part of the average 2 to 2½ per cent loss in the filtrate from the first sulphate is mechanical, not chemical. Efforts are being made to reduce this loss and have already partly succeeded, so well that after the ore in car P-9 had been treated no second sulphates were made.

The ore from cars 1 and 2 gave abnormal results. As regards the ore from car 1, the result can easily be explained, as this ore was part of five cars purchased by the National Radium Institute and had been ground in a ball mill to over 100-mesh fineness instead of being ground to 20 to 30 mesh fineness as is required. Filtration was therefore slow, with a resulting loss of radium. This explanation, however, can not be applied to the ore from car 2, which was readily filtered. Many extraction tests of the ore from this car were made in the laboratory, and only 70 to 75 per cent of the radium could be obtained. Although the amount of sulphate in the ore was fairly high, being 0.21 per cent calculated as sulphuric acid, it was no higher than in the ore from car 4 on which the recovery as sulphate was nearly 85 per cent. The delivery during September, 1915, was 407.5 milligrams, making a total to October 1, 1915, of 2,355 milligrams of radium element.

Up to September 1, 1915, 1,947.5 milligrams of radium element had been delivered as high-grade chloride or bromide out of the 4,774 milligrams of radium produced as sulphate. All of this was delivered between February 1 and September 1, 1915, as experimental work on methods of fractionation was not commenced until about December 1, 1914. The grade of the material delivered varied, some containing as much as 87.8 per cent radium. As a rule deliveries were about equally divided between material containing 6 to 15 per cent radium and a higher-grade product containing more than 50 per cent. The lower-grade product is used as a source of emanation for cancer treatment, and the higher-grade product may be used for direct radiation in connection with such work.

A total of 58.2 milligrams of radium from the fractionation process was returned to the plant as barium (radium) chloride. This was low-grade material, the radium content of which was too high to justify its discarding. The large quantity returned in the case of the ore from car P-4 was due to an accumulation from the ore from preceding cars. The total amount of radium discarded in connection with, the fractionation of the first 1,646.26 milligrams of element crystallized was 5.13 milligrams, or 0.31 per cent. The radium-barium ratio in this discarded material averaged 25.2 parts per billion.

Except for the discarded liquor mentioned, there are no definite losses that can be indicated quantitatively in refining the radium from the sulphate to the finished product. And yet, there are, of course, small unavoidable losses. The size of these losses will depend largely on the care of the workmen and chemists who do the refining, the personal factor being quite as important as the equipment. In handling and drying the sulphates some small loss through loss of material as dust takes place. In reducing the sulphate with charcoal the evolved gases carry a small amount of sulphate with them. Liquors are occasionally spilt, and “creeping” sometimes takes place in the porcelain and silica vessels in the laboratory. All of these factors can be more or less controlled, and with care the total losses may be kept down to between 1 and 2 per cent. With less careful work they may rise as high as 3 per cent. The loss of one or two crystals of high-grade salt will, of course, have a much greater effect than the careless handling of a large quantity of low-grade material.

The total refining losses at the plant of the National Radium Institute can be only estimated at present. A check on all radium under treatment has been made on more than one occasion, but as nearly 60 determinations were required in each instance the possible total experimental error was much larger than the probable losses. It is believed that the latter have been less than 2 per cent.

Recovery of Sodium Nitrate

During the time when crystallization of sodium nitrate from sodium sulphate was necessary, the recovery as usable nitrate was not more than 60 per cent, and sometimes went as low as 50 per cent. With the substitution of nitric for sulphuric acid for a neutralizing agent in the precipitation of the uranium and vanadium, the recovery in nitrate immediately rose to 70 to 75 per cent. More recently the yield has greatly increased, during the months of May, June, and July, 1915, averaging 87.5 per cent, with a minimum of 84.87 per cent and a maximum of 90.04 per cent.

Cost of Production

During the early stages of the work cost data were kept for all operations, but owing to changes and for other reasons the figures were not subdivided among the several departments in sufficient detail to give the exact data for these departments. Later, this omission was remedied but the departmental figures obtained do not represent an average of the whole operations but rather an average of the results, since the combined plants have been running in connection with the nitric acid plant.

The cost figures include not only expense for labor and materials in connection with the particular operations in question but also the cost of repairs and the actual expense in connection with boiler- room operation, water, electricity, office, laboratory, and superintendence. The figures by departments do not, however, include amortization, insurance, experimental work, cooperation of the Bureau of Mines, and other overhead costs.

The operations have been divided to cover the following items: Leaching, sodium uranate, iron vanadate, sodium nitrate recovery, nitric acid, radium refining, uranium refining, and boiler room.

Leaching

Leaching includes operations up to the separation of the barium (radium) sulphate previous to the addition of the acid liquor to an excess of sodium carbonate. The average cost has been $22.62 per (ton of ore treated. This item includes the cost of the recovery of the radium as barium (radium) sulphate. If the recovery of uranium and vanadium was not desired, operations could stop at this point. Such a procedure would vary the conditions and connection with the recovery of sodium nitrate; and as the cost of the nitric acid used is based on a definite recovery of sodium nitrate the above figure of 122.62 would be increased if nitric acid or a large proportion of the sodium nitrate for the manufacture of nitric acid had to be purchased.

Sodium Uranate

The figure for sodium uranate includes the cost of all operations in connection with the precipitation, filtering, and drying of the uranium as sodium uranate. The average cost has been 31.49 cents per pound of U3O8 in the dry sodium uranate.

Iron Vanadate

The figure for iron vanadate includes the cost of all operations in connection with the precipitation, filtering, and drying of the iron vanadate. The average cost has been 59.1 cents per pound of V2O5 in the dry iron vanadate. This part of the work is therefore conducted at an actual loss, but in combination with the sodium-nitrate recovery, which is based on the extraction of the vanadium, is worked at a profit.

Sodium Nitrate Recovery

The recovery of sodium nitrate involves the evaporation of the sodium-nitrate solution and the crystallization of the nitrate. The cost per pound of sodium nitrate in the recovered salt has averaged 0.3429 cent.

Nitric Acid

The cost of nitric acid has varied with the recovery of sodium nitrate, the market price of nitrate used to make up losses, and the cost of repairs. During the past three months (up to Aug. 1, 1915) it has averaged 2.411 cents per pound of 100 per cent acid, and during the preceding two months 2.050 cents per pound on a basis of 80 per cent recovery of the nitrate.

Radium Refining

Radium refining includes all operations in the plant in the reduction of the barium (radium) sulphates and the fractionation of the chloride liquors up to the point where deliveries are made to the laboratory. The cost has been $2.44 per milligram of radium element delivered to August 1, 1915. At first it was necessary to get the fractionation systems established both at the plant and at the laboratory. Consequently more radium went into the systems than came out. The relative cost of refining at the start, based on the quantity of material delivered, was therefore high. More recently the cost has been reduced to $1.02 per milligram. The cost of refining in the laboratory has been $1.03 per milligram of radium element. This figure includes the salaries of the men engaged in the fractionation work and the cost of chemicals, gas, etc., but does not include the cost of analytical work or supervision or other overhead charges.

Expenditures to August 1, 1915

The expenditures to August 1, 1915, exclusive of those for ore and for Bureau of Mines cooperation, are given below. As already stated, the proper distribution of costs on power, water, etc., was not made to departments during the early experimental part of the work, all such costs being placed under “plant operation.” Although this was done later on, as already detailed (pp. 113-115), it has been thought better, in giving the figures below, for the sake of uniformity, to still summarize such costs under “plant operation.”

Total Costs per gram of Radium Element

The first ore used in the plant was purchased. This gave an opportunity to test out the process before mining operations were begun. Details of costs of ore mined will be given in a bulletin, being prepared by the Bureau of Mines, on the mining, milling, and concentration of carnotite. These costs cover mining, hauling, freight, grinding, sampling, 70 per cent amortization of equipment, Bureau of Mines cooperation, royalty, and other overhead expenses.

In figuring the total cost of the radium produced, certain overhead expenses, in addition to the cost of the ore and operating costs, that are not included in the operating costs must be charged against the radium. As the plant is in excellent condition and with ordinary repairs, which are taken care of under operating costs, will be perfectly serviceable at least 10 years, an amortization charge for plant and equipment of 20 per cent per annum against each unit from the time it started operation until August 1, 1915, has been made and should be more than ample. The time given to the radium work by the technical men and chemists of the Bureau of Mines has been charged at the full salary rate, a monthly record having been kept of the proportionate amount of his time given by each man. Traveling expenses are also included.

As shown in the table on page 110, 4,258 milligrams of radium element was produced in the form of sulphate from the ore in the first shipments up to and including car P-16, or August 1, 1915. Of this amount, 1.646.56 milligrams of radium element had been delivered, the rest being in the form of untreated sulphate or in process of fractionation. In the table below the cost of refining for delivery these untreated sulphates has been figured at the average refining cost during the last two months considered.

In figuring the cost of the radium, the uranium and vanadium products may be either included or excluded. All the uranium and vanadium has been recovered by the National Radium Institute plant as sodium uranate or uranium oxide and iron vanadate, the whole plant being designed and constructed with this object in view. By subtracting the actual costs connected with the production of the uranium and vanadium compounds from the total cost, a close approximation of the cost of the radium, provided the uranium and vanadium were not recovered, will be obtained, but such a figure will not be exact, as under the changed conditions various factors would enter in, the effects of which can only be estimated. On this basis the average cost of 1 gram of radium element, including the more expensive early treatment, may be determined as follows:

As already stated, refining losses have almost certainly been less than 2 per cent and probably less than 1 per cent. In order, however, to be on the safe side, an allowance of 3 per cent is made for such losses; 4,258 milligrams less 3 per cent is 4,131 milligrams, which represents the radium finally recovered as high-grade salts. The average cost of 1 gram of radium element has therefore been $37,599. It should be remembered that this cost includes the much higher operating costs of the smaller experimental plant and that the first 2 grams of radium extracted cost considerably more per gram than the last 3 grams; also there have been extracted 31,650 pounds of uranium oxide and 11,528 pounds of vanadium oxide. This material has all been contracted for and in part delivered. The returns from its sale will considerably more than cover the cost of its production, and this profit, together with other credits, will ultimately lower by several thousand dollars the cost per gram of radium.

As the price of ore is variable, the question will naturally arise as to the influence of the price of ore on the figures given above. A simple calculation will show at once that the cost of extracting radium, exclusive of the cost of the ore, has been $20,710 per gram. The ore included in the figure of $69,767.99 cited for the cost of ore used in the investigation herein outlined was partly purchased before the war and partly mined by the National Radium Institute at Long Park, Colo. As 723.97 tons was used, the average cost was $96.36 per ton. If the ore had cost $120 per ton, the cost of radium would have been $41,742 per gram, and for every additional amount of $20 per ton above these figures the cost of radium would increase approximately $4,000 per gram.

As it is the hope of the bureau and one of the purposes of this investigation that the miners shall receive a more adequate return for their ore, the figures presented herein will enable anyone interested to determine the approximate cost of radium by the methods outlined under any market price for ore that may prevail.

A limited supply of the following publications of the Bureau of Mines is temporarily available for free distribution. Requests for all publications can not be granted, and to insure equitable distribution applicants are requested to limit their selection to publications that may be of especial interest to them. Requests for publications should be addressed to the Director, Bureau of Mines.