Refining in the Plant

The first treatment of the radium-barium sulphates is of great importance, as the whole capacity of the plant, as far as the radium goes, depends upon this first treatment. If the sulphates can not be handled as rapidly as produced, the actual amount of radium refined each year must necessarily be much below the capacity of the plant itself. The tediousness of the process used by Haitinger and Ulrich (see p. 20) can be readily appreciated on reading the account of their work. In order to produce 5 to 10 grams of radium metal per year, it is absolutely necessary to have shorter methods of handling the sulphates than were used in the early days of radium refining.

The first step in this direction is to obtain a high-grade sulphate, as sulphates containing a considerable amount of silica and other impurities are not nearly as readily treated as those that are comparatively pure. Therefore, any process for the treatment of carnotite or pitchblende that involves the precipitation of a high-grade, clean sulphate has a great advantage over other processes involving the precipitation of the sulphate along with silica or allowing the sulphate to remain in the residue with silica and other impurities. This is one of the advantages possessed by the process described in this report. As a rule, the sulphates obtained contain at least 85 per cent BaSO4, and in much of the product the proportion of barium sulphate present is as high as 90 per cent. The proportion of radium in these sulphates has averaged about 1 milligram of radium metal per kilogram of total weight.

Old Method of Treating Radium-Barium Sulphates

The old method of boiling the sulphates with a solution of sodium carbonate in order to obtain radium carbonate and sodium sulphate is one that can be applied to any radium barium sulphate. It involves, however, considerable labor and much filtering and washing, as the radium-barium carbonate obtained must be washed completely free from traces of sulphate. The carbonate is then dissolved in chemically pure hydrochloric acid and filtered, and any insoluble material is worked up with the next batch of sulphate. If three parts of sodium carbonate dissolved in water is boiled with one part of radium-barium sulphate, a considerable part of the sulphate is converted into carbonate at one treatment. It is claimed that this conversion can be improved if the heating is done under pressure in an iron tank, although the authors have not tried this method. Any method, however, that will shorten this somewhat tedious process should be welcomed.

Results of Ebler’s Experiments

The results obtained by Ebler are suggestive. In his experiments he used two sulphates, one low grade and the other high grade. The low-grade sulphate contained 37.1 per cent barium sulphate, 21.8 per cent silica, and 23.9 per cent total sulphate, with traces of vanadium, iron, aluminum, etc. The radium content was only 0.032 milligram per kilogram. The analysis of the high-grade sulphate was as follows:

On heating these sulphates with carbon to a bright red heat for two to four hours he obtained a 32 per cent reduction with the low-grade and a 68 per cent reduction with the high-grade sulphate. With calcium carbide the following reaction takes place:

MeSO4 + 4 CaC2 = 4CaO + MeS + 8C

The temperature and time of heating was the same as for carbon. With the low-grade sulphate he obtained an 83 per cent reduction and with the high-grade sulphate a 70 per cent reduction, showing an improvement in carbon with the low grade, but not with the high grade.

With calcium hydride a spontaneous reaction takes place as follows:

MeSO4 + 4CaH2 = MeS + 4CaO + 4H2

The reduction with the low-grade material was 75 per cent, and with the high grade 80.5 per cent.

Ebler also tried a mixture of calcium hydride and carbide. The low-grade sulphate required three parts of the former to one of the latter to produce a 64 per cent reduction. The high-grade sulphate gave a 60 per cent reduction with one part of hydride to two parts of carbide.

In general, he found that the crude sulphate was better reduced by the more vigorous reducing agent. Although the reactions with calcium hydride and carbide are immediate and spontaneous, the cost and the difficulty of obtaining calcium hydride make the desirability of their use doubtful.

PART 1: Radium, Uranium & Vanadium Extraction & Recovery from Carnotite

PART 2: URANIUM, RADIUM & VANADIUM Ore Processing

PART 3: VANADIUM & URANIUM Extraction and Recovery

PART 4: RADIUM Extraction & Recovery

PART 5: Processing, Extraction & Recovery of RADIUM 

Reduction of Barium Sulphate by Coal Gas in an Electric Furnace

It is a well-known fact that barium sulphate can be reduced by heating in a current of coal gas. This method promised good results, although others had tried it and had reported failure without giving details. As temperature control within reasonably narrow limits was necessary in order to get reduction, and at the same time to get the material in a nonclinkering condition so that it could be readily removed from the tube in which it was heated, the use of an electric furnace seemed more desirable than heating by coal, oil, or gas.

The reduction chamber of the electric furnace consisted of a cast-iron tube 5 feet long and 6 inches in internal diameter. Around this was wound 120 feet of No. 21 nichrome ribbon, the insulation being asbestos cloth and a double layer of mica, the pieces being 4 inches by 5 inches in size.

Reduction begins at 575° to 600° C., the actual working temperature being 725° to 740° C. For raising the temperature a current of 40 to 50 amperes was required, and for maintaining it during the reduction, 15 to 20 amperes from a 220-volt circuit. The gas consumption averaged 200 to 300 cubic feet per run on 25 to 28 pounds of sulphate.

The heat of reduction was positive at high temperatures, producing reaction clinkers, some of which were friable and some glassy and hard to break. The formation of the clinkers did not seem to affect the reduction, which varied from 75 to 90 per cent.

The main advantage of a reduction with coal gas lies in the fact that no foreign material is introduced. The principal disadvantage is the difficulty of maintaining the insulation owing to the deterioration of the mica. Silicon-carbide tubes were also tried in the electric furnace instead of cast-iron tubes without success. As a rule, after 8 or 12 runs a tube would have to be rewound and the insulation renewed, and during reduction the furnace needed constant attention in order to maintain the correct temperature. The iron tube was also gradually attacked by the charge.

A constant-feed furnace of much smaller size, having an alundum tube, was also built. The difficulty was in getting a proper feed through the furnace and a sufficiently high reduction. When the tube got hot it became sticky, and the rate at which the sulphate went through could not be controlled.

Reduction with Charcoal

While the work was being done with the electric furnace, the reduction of radium-barium sulphates with charcoal was tried. This was done by mixing the sulphate with about one-fifth its weight of powdered charcoal and heating for 7 or 8 hours in a graphite crucible in an oil furnace, the temperature being about 800° C. Reduction takes place readily, according to the equation BaSO4 + 4C = BaS + 4CO, with carbon monoxide burning at the spout of the crucible. When the evolution of gas has stopped, the crucible is allowed to cool with the lid on, and the radium-barium sulphide is removed and broken up. Under such conditions, with a sulphate reasonably free from silica, it is usual to get a 90 per cent reduction of the sulphate to sulphide, the ratio of the reduction of the radium being the same. The sulphide is then dissolved in dilute chemically pure hydrochloric acid. This step should be performed out-of-doors or with a hood having a good draft, as the evolution of hydrogen sulphide, of course, is large. The gas also catches fire spontaneously at times, owing to the possible formation of small quantities of phosphides, although it is difficult to see how these could be formed under the conditions. If the work is conducted in open vessels in the open air this danger is minimized. The residue obtained on filtration consists of undecomposed radium barium sulphate, a small amount of charcoal, and some silica. The amount of radium in this residue is usually about the same as that in the original sulphate. The residues are stored and re-treated in the same manner, although filtering them after leaching with acid is more difficult than filtering a sulphide obtained from an original sulphate.

After the re-treatment of these residues, a third residue is, of course, also obtained. There is a tendency for the accumulation of lead in the second and third residues, especially the latter, and, therefore, it is better to fuse the third residue with sodium carbonate than to reduce it once more with carbon in the usual way. Alter fusion, during which most of the lead is eliminated, the material is thoroughly washed and dissolved in chemically pure hydrochloric acid. As the weight of the second residue is only 1 to 2 per cent of that of the original radium barium sulphate treated, it is necessary to carry out this treatment only once or twice a year.

The reduction with carbon is the method that has been used almost exclusively at the plant of the National Radium Institute. An oil furnace (23, Pl. III) holding three No. 100 graphite crucibles handles all of the sulphate made in the plant, the furnace being used only every other day. The actual amount of charcoal required is a little more than one-fifth the weight of the sulphate treated. The charcoal and sulphate are ground in one or two small ball mills (66, fig. 2), a short piece of 1-inch steel rod being placed in each mill instead of balls. This treatment for 10 or 15 minutes is sufficient to intimately mix the sulphate and the charcoal. The ground material is then carefully removed from the ball mill and is placed in the crucibles, which are then transported to the furnace room.

Treatment of Sulphide

After reduction, the sulphide is ground in a similar manner in the other mill and is then leached with moderately dilute chemically pure hydrochloric acid in an earthenware pot placed just outside of the building. No particular precautions are necessary to prevent oxidation of sulphide to sulphate. If such oxidation does take place, the amount of radium and barium precipitated will be proportionate to the relative quantities of these elements present—roughly, 1 part per million. The material in the pot is then transported in crocks to an earthenware suction filter, and the filtrate, which consists of an almost saturated solution of radium-barium chloride, is ready for fractionation. In actual practice it has been customary to use an amount of acid that will not be saturated with barium chloride when hot, but will be almost saturated when cold. This precaution reduces the boiling afterwards required to get the first batch of crystals.

Equipment for Fractionation

The actual arrangement for fractionation as indicated in figures 2 and 3 is not the arrangement that has been used up to the time of writing (September, 1915), but is what will be used after additional acid-proof ware has been received from France. In the past it has been necessary to use an inadequate equipment, so that the refining capacity has been less than that of the rest of the plant. With the equipment shown in figures 1 and 2, however, it will be possible to refine considerably more than the total output of the plant, now about 400 milligrams per month.

Procedure in Fractionation

The method of treatment with this new equipment is to be as follows:

The radium-barium sulphides are dissolved in an earthenware pot, 63 (figs. 2 and 3). The solution and the insoluble material are run onto the upper part of a suction filter, 64. After filtration and washing, the liquor is run into the upper steam-jacketed vessels, 66. These vessels are made of silica-lined acid-proof ware, as previously described (p. 44), the upper three each having a capacity of 250 liters and the lower two each having a capacity of 200 liters. The evaporators are covered by two wooden hoods, in the flue of which is placed a small steam jet to produce a sufficient draft to carry off the hydrogen chloride fumes. The barium chloride solution in the first evaporator, 65, is then so concentrated that on cooling one-half of the barium chloride crystallizes out. The liquor is siphoned into evaporator 65a and the process repeated, one-half of the remaining barium chloride being allowed to crystallize. This procedure can be continued all the way down the series, or a fresh batch of liquor may be introduced into 65a after the crystals have been removed.

The system may be established in several ways, either by adding fresh material to 65 each day and gradually working the mother liquors down the evaporators to the bottom of the series, or by crystallizing the first batch of liquor introduced into 65 all the way down the series before the introduction of fresh material. Whichever method is used, the procedure afterwards adopted is the same. Half of the barium chloride in evaporator 65d is crystallized out and removed from the mother liquor. This liquor should then be discarded entirely. After the crystals have been removed from the evaporator, the liquor obtained from a similar crystallization in 65c is siphoned into 65d. The liquor from a similar crystallization is then siphoned from 65b into 65c, and the crystals from 65d are dissolved with this liquor in 65c. In other words, the crystals from each evaporator ascend, and the mother liquors at the same time descend, the crystals from a given evaporator being dissolved in the liquor from the second evaporator above.

For example, the crystals from 65d will be dissolved in the liquor from 65b in the evaporator 65c, with the addition of the minimum amount of water required, the crystals from 65c will meet the liquor from 65a in 65b, and will be dissolved in this liquor, and so on up the series. The new material is always introduced into 65. The crystals obtained from 65 are then refractionated twice in smaller pots (62) of about 80-liter capacity, the liquor from the first crystallization coming back into 66 in the same manner as already described. Evaporator 65, therefore, may be considered as the “zero” pot, with the barium going in one direction and the radium in the other. After the second recrystallization in the smaller pots, the material is ready to go to the laboratory, the ratio of radium to barium varying from 4 to 10 parts per million. The whole method is more fully described on pages 76-86.

The crystallization factor of hydrochloric-acid solution is about 1.5 to 1.6; that is to say, if 50 per cent of the barium chloride is removed, there will be 50 per cent more radium in the crystals removed than in the liquor left behind. Owing to lack of equipment only two “minus” crystallizations have in the past been given in acid solution. The average radium content of the liquor thus obtained has been too high to discard, averaging during the past few months, 0.0679 milligram of radium per liter of solution, or a ratio of radium to barium of 187.5 parts per billion. With three, or even two more “minus crystallizations, as planned, the radium content of this material could readily be reduced so that the final liquor could be discarded.

In actual practice up to September, 1915, this solution has been neutralized with ammonia, preferably filtered, and fractionated in the manner already described, in a series of seven bathtubs shown at 52 in Plate VII. The head batch of crystals is then put back at a suitable point in the acid crystallizing system and the tailings either discarded or used as barium chloride solution for the precipitation of fresh batches of barium sulphate in the plant. During the past few months, the average radium content of the tailings discarded, or used as described, has been 29 parts per billion. Ordinarily, material having 40, or even 50, parts of radium per billion might well be discarded.

At all necessary stages in the fractionation, samples are taken for the determination of radium, so that a check may be kept on the progress of the work.

Laboratory Refining

Radium of any desired degree of purity may be obtained by fractional crystallization of the barium solution, first as chloride and later as bromide. The radium continues to be enriched in the crystal fractions and impoverished in the successive mother liquors, as already described.

The salts of most metals, such as iron, aluminum, and vanadium, that may occur with the radium-barium salt as impurities, pass into the mother liquors, and give no difficulty in the enriching radium fractions. Lead is an exception, however, and requires special treatment. Where barium (radium) sulphide, after reduction from sulphate is dissolved in hydrochloric acid solution with copious evolution of hydrogen sulphide (see p. 74), one might expect the almost complete removal of the lead content. This by no means occurs, no matter what precautions are taken in the way of final acid concentration or other conditions, such as have been prescribed by other authors. The treatment necessary to remove all lead and the reasons for its seemingly abnormal action are discussed in a subsequent section.

Figure 5 shows the steps of the concentration system employed in the laboratory.

Treatment of Chlorides

The crude barium (radium) chloride received from the plant, containing 4 to 10 parts of radium element per million, is dissolved in water in large porcelain dishes (Pl. XIII, A), and hydrochloric acid

is added to precipitate as much lead as possible. After the solution has stood over night, it is filtered to remove lead chloride and whatever carbon and barium sulphate may have escaped previous filtration. For this purpose a Schleicher & Schull 50-cm. folding filter is used repeatedly (10 to 12 times) until clogged, when it is washed, dried, and ignited together with the residue, which is held for treatment to recover the radium.

The filter ash and residue, after the passage of about 1,500 milligrams of radium element, weighed 9.5 kilograms and contained 37 milligrams of radium, or about 2.5 per cent of the total put through. The residue would therefore contain about 4 milligrams of radium per kilogram, being considerably richer in radium than the original first sulphates. During part of the operations, however, the residue ran as low as 1 milligram per kilogram (about the same as the original first sulphates), which shows that thorough washing can reduce the radium ratio at this point to that in the original insoluble material.

The filtrate is made up to an 8-liter volume, and, after a sample of 0.01 c. c. has been taken for analysis, the filtrate is introduced into the chloride crystallizing system in the vessel designated ClO in figure 5. The vessels used for ClO, Cl—1, and Cl —2 (PI. XIII, A) are all cast-iron ware lined with silica and were obtained from the Danto-Rogeat Co., of Lyons, France. At the time of writing (September, 1915) they have withstood continuously for 9 months hydrochloric-acid solutions which were boiled for several hours daily and yet show little or no deterioration. The pots contain, in the order named above, 16.6, 13.6, and 10 liters, and each is suited to handling about 2½ to 3 kilograms of barium chloride in one daily crystallization. The pots are mounted on iron-ring tripods and heated with perforated-cap gas burners 5 inches in diameter (see Pl. XIII, A).

As the factor of enrichment of radium chloride from barium chloride and also of radium bromide from barium bromide is more favorable in acid than in neutral solutions, a fair acid concentration is maintained throughout the chloride and bromide systems.

As indicated in figure 5 in both systems the crystals moving forward are combined with mother liquors moving backward from pots two places removed from each other in the system. Another advantageous feature of the procedure employed consists in introducing the new material, not into terminal but into intermediate vessels, so that the system may be shortened or lengthened at either end without disturbing the rest of the system. For vessels +1 and +2 in the chloride system either porcelain or fused silica vessels may be conveniently used.

The mother liquor from Cl —2 is collected in 100-liter lots, which are analyzed for radium and barium before being returned to the crystallizing system at the plant. The solutions, which are saturated with barium chloride, contain 0.16 to 0.32 gram of barium chloride per cubic centimeter, the exact quantity depending upon the degree of acidity maintained, and have varied in radium content from

0.00001 to 0.00020 gram of radium per liter, the exact content depending on the quantity of radium accumulated in the system and also the acidity. The total quantity of radium thus returned to the plant out of 1,500 milligrams put into the system was 60 milligrams, or about 4 per cent.

The chloride crystals from Cl + 2 are dissolved in water without acid, and the solution from two daily batches is brought into a large glass precipitating jar 16 inches high and 8 inches in diameter and made slightly ammoniacal. Hydrogen sulphide is passed in until the precipitation of lead is complete as lead sulphide. This sulphide is filtered and collected according to the procedure for lead chloride and barium sulphate already described, and analyzed. The amount of lead sulphide thus collected during the treatment of 1,500 milligrams of radium was 700 grams, which contained 3 milligrams of radium or 0.2 per cent of the total radium involved. The lead sulphide is stored for treatment to recover the radium. Some lead sulphide was collected, after sulphuric acid free from lead had been used at the plant, in order to preserve a sample of lead coming exclusively from a uranium mineral.

Removal of Lead

The difficulty of removing lead has already been mentioned. Attempts to precipitate it as sulphide even from slightly acid solution usually result in obtaining an orange-red precipitate, which is a sulphochloride of composition (PbS)x.PbCl2, and far more soluble in acid than lead sulphide; hence complete precipitation of the lead is possible only in ammoniacal solution. Explanation of this action is to be sought in the high concentration of barium chloride in the solution producing such a high concentration of Cl’ ions that the dissociation PbCl2↔PPbCl+ + Cl’↔Pb++ + 2Cl’, is almost entirely arrested at the intermediate step. The absence of Pb++ ions and the readiness of Pb2Cl2S to form according to the equation 2PbCl+ + S” = Pb2Cl2S, explains the impossibility of obtaining lead sulphide. Precipitation of Pb2Cl2S is only partial so far as lead is concerned and does not take place unless the acid concentration is low. Complete removal of lead as normal lead sulphide is readily accomplished, however, by passing hydrogen sulphide in ammoniacal solution. The loss of radium here has already been shown to be only 0.2 per cent.

Conversion to Bromide

Into the filtrate from the lead precipitate powdered ammonium carbonate is introduced gradually with vigorous stirring until all the barium has been precipitated as carbonate. Where a slight excess of reagent is added, the whole is allowed to stand over night for the settling of the barium carbonate and the thorough precipitation of the radium carbonate. The supernatant solution is siphoned off as far as possible, and the rest goes to a 9-inch Buchner funnel, where it is filtered and washed several times with suction. The filtrate, together with the wash water, designated “laboratory carbonate liquor, ” is stored in 50-liter lots and returned into the acid leach liquor at the plant to recover radium and assist in neutralization. The amount of radium in the liquor is surprisingly low, usually 0.001 to 0.003 milligrams per liter, and out of 1,500 milligrams of radium only 6.4 milligrams, or 0.4 per cent, remained in this liquor. In most lots the amount was much lower than this average.

Treatment of Bromides

By inverting the Buchner funnel the barium (radium) carbonate precipitate is removed, and after being separated from the filter paper is dissolved in chemically pure hydrobromic acid (20 to 35 per cent) in the precipitating jar already used. The Buchner funnel and the filter paper are also washed with 1 to 1 hydrobromic acid. If necessary, this solution is again filtered and washed through a Buchner funnel and then introduced directly into the BrO vessel.

All the vessels used in the bromide crystallizing system (Pl. XIII, B) are of fused silica ware, which can be more safely heated by free flame than porcelain and the solution in silica ware “crawls” less. For BrO the 12-inch or 15-inch dish is convenient; for Br — 1 and Br — 2 the 12-inch dish is sufficient, and the sizes fall off rapidly in the plus direction to the smallest silica dishes obtainable.

The degree of acidity to be maintained both in the chloride and bromide crystallizing systems can not be stated definitely, as it depends on operating conditions and must be left somewhat to the judgment of the operator. In general, it may be stated that the acid concentration should be maintained as high as consistent with the size of the crystal crop to be obtained and the quantity of acid fumes that can be conveniently liberated.

The silica vessels in the bromide system are heated on tripods with bare gas flames until the richer fractions are reached, when the evaporation is carried out on an electrically heated water bath in which only distilled water is used, which is shielded from the heating coils by a block-tin protector. In case of an accidental loss of radium solution into the bath, all the water can be drawn off and returned into the system just before the treatment with hydrogen sulphide.

The evaporation required to obtain a suitable batch of crystals may be generally regulated by concentrating the solution until vigorous fanning just begins to cause the formation of crystals on the surface of the hot solution. Of course, the higher the acid concentration the more generous the crystal batch will be. All vessels in the chloride system are allowed to cool over night, so that only one crystallization a day is made in each vessel in this system. In the bromide series, in which the vessels are smaller, cooling and crystallization proceed rapidly enough to permit several batches being put through in a day if desired. In warm weather the use of an ice box to contain the vessels of the bromide system facilitates operation. The vessels of the bromide system, when not being heated, are kept in order on a board with holes of sizes suited to the sizes of the various dishes. (Pl XIII, B.)

The barium-chloride crystals invariably form a mat on the bottom of a vessel, from which the mother liquor is drained with some difficulty, whereas the bromide crystals form in heavy individual needles, from which the mother liquor is poured off clean with great ease. Only when the bromide system is disturbed by the presence of lead or chloride salts or when it contains too little acid do the bromide crystals fail to form in their characteristic habit. Both disturbing influences should be carefully avoided, as radium concentration is at once retarded if the bromide crystals do not retain their type.

Owing to the high factor of enrichment as bromide, the radium content of the mother liquor from Br — 2 is extremely low. Out of about 1,400 milligrams of radium treated in the bromide system, only 3 milligrams, or 0.2 per cent, went back into this final bromide liquor. In four out of six lots the radium content was only 20 to 30 parts per billion of salt. As it is not profitable to recover radium at a concentration below 40 parts per billion parts of barium salt, the liquors were treated for the recovery of hydrobromic acid only. In the two lots exceeding this concentration the barium-radium sulphate obtained in the acid recovery was stored, to be handled as second sulphates.

The method of recovering hydrobromic acid from the final bromide liquor just mentioned consists in adding a slight excess of sulphuric acid to precipitate all barium, then in adding a slight excess of the original barium-bromide liquor to remove any excess of sulphate, filtering on a large earthenware Buchner funnel, and distilling off hydrobromic acid from the filtrate, a large round-bottomed glass flask and a glass condenser being used. The distillation is repeated until at least a 20 per cent acid is obtained.

Number of Fractions Employed

The number of fractions employed in the plus direction in the bromide system varies somewhat with conditions, 10 to 12 being the usual number. The crystallization is conducted in such a way that the barium (radium) bromide collected in the final fraction should not fall below 1 per cent of radium bromide, and sometimes is as high 4 per cent. The total weight of the fraction should be 1 to 2 grams. The factor of concentration for each step in the acid bromide system is about 2 to 2.2, and in the acid chloride system about 1.5 to 1.6.

The procedure in collecting the final fraction is as follows: The mother liquor having been poured back, the crystals are dried in the silica crystallizing dish, first on the steam bath, and finally in an air-drying oven at 105° to 110° C. until thoroughly dry. The crystals are then collected in a previously-weighed glass tube sealed at one end, 4 to 6 millimeters in internal diameter, and of sufficient length to contain the salt and permit the other end to be sealed with the blast. The total weight of salt is determined by weighing the sealed tube and contents together with the detached end. The salt is kept stored in the tube, and gamma-ray measurements of the radium content are made at intervals of a few days until the determination is thoroughly established. When a number of tubes have been collected, they are opened and the contents combined for further fractionation.

In opening tubes that have been closed for several weeks, care must be taken that the contents shall not be scattered by the release of gas that may have been generated by the chemical action of the alpha rays, especially if the salt was not thoroughly dried when placed in the tubes. Two precautions may be taken—one end may be drawn out to a capillary, which may be opened first, thus relieving the pressure; or a scratch may be made near one end, which may then be inserted well into a beaker lying horizontally and the tube opened by a hot piece of glass applied to the scratch. If a capillary end is provided, it is preferable to have this so narrow that the salt crystals can not enter and the salt will remain confined in a definite length of tube, one not too great for accurate gamma-ray measurements. It is also desirable that the drying and sealing of a tube shall follow crystallization as promptly as possible in order to furnish a definite zero point from which the accumulation of gamma radiation may be dated. This detail is more fully discussed in a subsequent description of measurements (p. 89).

Procedure in Higher Fractionation

After the collection and measurement of the salt in a sufficient number of tubes (usually 10 to 20), the higher fractionation may be commenced. If the percentage of radium is approximately the same in all tubes, they may be combined into one solution; if not, two or three solutions are made, which later fall into the crystallization series in their proper places.

As a guide for procedure in making this special fractionation, it has been found convenient to assume a concentration factor of 2.0 for the radium enrichment, which means briefly that if one-half of the salt is crystallized from a given fraction, two-thirds of the radium is contained in the crystals and one-third in the mother liquor. The advance fractions are then made and the lower ones recombined as suggested by this factor until it is thought that the desired products have been attained, when the crystals are collected and sealed as already described for the lower-grade salt. If the products prove on measurement not to have the desired percentage of radium, the tubes are opened again and either the salts are suitably recombined, or possibly further fractionation is canned out.

As regards the higher grade salts, one additional precaution is observed in sealing the tubes, namely, a small platinum wire is sealed through one end of the tube to conduct away the unipolar charge that may collect in the interior, attaining voltages that could cause destructive sparking. Reports are on record of serious radium losses having resulted through neglect of this precaution.

Results of Higher Fractionation

The following data from a series of higher fractionations will serve to illustrate the procedure:

As will be noted, the total weight of salt at the start was 11.86 grams, and after recrystallization 11.64 grams. The difference was due to material adhering to crystallizing dishes in the poorer fractions. The total Ra at the start was 221.7 milligrams, and after recrystallization 222.67 milligrams. The difference was due to a difference in measurements.

Tubes C-1, C-2, and C-3 were the fractions preserved, the others were returned into the crystallizing system at suitable points.

Comments on Results

In beginning this series, tubes 33, 34, and 35 were combined and recrystallized in twelve “head” fractions directly to tube C-1. To the first mother liquor from tubes 33, 34, and 35 were added the contents of the other tubes, Nos. 31, 32, 36, and 37. This series was carried in about 15 recrystallizations up to tube C-2. The other lower fractions in the C series were obtained by further fractionation and combination of the mother liquors, as suggested by the factor 2.0.

In general, the fractionation of radium bromide from barium bromide by means of crystallization is simple and sure. No especial skill of the operator is required, as many authors lead one to believe, nor does it seem possible that the substitution of other processes, such as the adsorption of radium by means of colloidal gels or the fractionation of radium in concentrated alkaline solution would prove advantageous. The applicability of such methods should be regarded as doubtful until demonstrated on an actual working scale. There is a wide difference between commercial operation and the extraction of a small quantity of radium in the laboratory.

The time of only one man continuously is required to carry on the laboratory refining through the system indicated in figure 5, and the time of one additional man is required for four to six days each time the higher fractionation is performed.

Adsorption of Radium

Contrary to rather widely entertained views, the production of a precipitate, or the presence of one, in a radium-barium solution does not necessarily involve the adsorption of considerable quantities of radium from solution. As has already been pointed out, the precipitation of 700 grams of lead sulphide in ammoniacal solution carried with it only 3 out of 1,500 milligrams of radium. The recent principle of adsorption established by the work of Paneth points out that any serious removal of radium should be expected only in case the radium forms an insoluble salt with the negative radical of the adsorbent. Without entering at present into discussion as to whether this removal is really due to adsorption or to ordinary chemical precipitation, it is interesting to note that the failure of radium to be removed as sulphide, together with lead, accords well with its chemical nature, as one would expect radium sulphide to be soluble. On the other hand, the completeness of its removal as carbonate, which is to be expected, has been shown by the fact that only 6.4 out of 1,500 milligrams of radium element passed into 300 liters of filtrate.

Even if barium sulphate is present in or is precipitated in a solution containing radium, the removal of the radium is small provided a large excess of barium is present. A discussion of the “protective” action of barium appears in the section on measurements (p. 92).

Accumulation of Radium in Laboratory Crystallizing System

It is desirable to allow several hundred milligrams of radium to accumulate in the crystallizing system. This accumulation tends to increase gradually unless special preventive means are taken, such as crystallizing out special fractions without corresponding introduction of fresh material. The following table serves to show the results of operating the system during a period of about one month:

Changes in the richness of the deliveries into the system affect concentration in the higher fractions with a certain lag. The distribution of radium in the different vessels of the system is somewhat accidental, dependent on the size of fractions taken, acidity, and other conditions that may vary from time to time, but in general, radium tends to accumulate in the —1, 0, and +1 fractions both in the chloride and bromide systems. The total accumulation in the bromide is greater than in the chloride system on account of its greater number of fractions, the higher solubility of barium bromide, and the higher degree of concentration of the radium.

All the refining operations described in this section are carried out in a room 12 by 20 feet. The vessels of the chloride system are handled over a large, shallow copper trough to prevent possible loss of solution by “crawling,” possible breaks, or other accidents. Incidentally, it may be mentioned that the refining work up to the present has been notably free from misfortunes, and that no serious losses of radium have ever occurred. Although a hood is not necessary, it would sometimes prove advantageous in carrying off acid fumes. Usually, however, solutions are not sufficiently concentrated for the acid fumes to become very objectionable.