The concentrator is shown diagrammatically in figure 98. The entire assembly, with the exception of the aluminum vessel, B, and condensing coils, C, is of glass; and with the exception of the small asbestos filter, the sirup never comes in contact with any other substance. The evacuating is done by a power-driven vacuum pump or an aspirator connected at D. When it is desired to concentrate a sugar sirup, the procedure is as follows: The solution, usually approximately of 50 percent concentration, is placed in the flask, A, and the entire system evacuated up to the cock, E. This cock is then carefully opened and the solution slowly driven through the asbestos filter into the boiling flask, F, capacity 13 liters. Here it is gently warmed by the water bath, and the temperature of the sirup noted on the thermometer, G. In order to obtain any desired boiling point, it

is only necessary to regulate the pressure, as indicated by the mercury gages H, H. The efficiency of the assembly is such that sirup is rapidly brought to the desired concentration of about 80 percent at a temperature below 32° C. This is made possible by the high efficiency of the condensing system and by having all joints carefully ground to fit. The stopcocks are lubricated with water or sugar sirup. The condensing coils, C, are of copper, eight in number, connected in parallel, and immersed in ice water. To prevent vapor reaching the vaccum pump, the pipe, I, is joined to the base of the condenser. As rapidly as the vapors condense they pass into vessel J and subsequently, by closing cock L and opening M, are expelled into K from which the liquid is eventually driven into the waste. By noting the quantity of this liquid, the concentration of the sirup in F is known at all times.

The question of size of crystals is of the first importance. In general the smaller the crystals the less the "included solvent" or entrapped mother liquor. When the solution in the boiling flask, F, has reached the desired concentration, the vacuum is broken at N and the solution poured out to crystallize. Crystallization in general does not begin in these pure solutions until they are seeded with a few fine crystals of sugar. This is not done until after the solution is removed from the boiling flask. Two methods of crystallization have been tried. In the first the concentrated solution is transferred to a precipitation jar. It is then carefully stirred with a glass rod provided with a glass shield. This procedure gives satisfactory results so far as the control of the crystal growth is concerned, but it is laborious. In the second, the liquid is transferred to a crystallizer, a cross section of which is shown in figure 99. By this means it was found practicable

GROUND JOINT

FLASK FOR

SUGAR SOLUTION

FIGURE 99.-Crystallizer.

to crystallize from a larger quantity of liquid and with no danger of contamination. The device is very simple and consists of a hardwood box mounted on bearings and driven by an electric motor. The supports, A, A, which carry the container, are adjustable in order to keep the center of gravity on the axis of rotation. The box is rotated slowly so that the contents are subjected to a gentle and continuous agitation.

The solid is separated from the mother liquor by a motor-driven centrifugal machine. The centrifuge shown in cross section in figure 100 was built in cooperation with the Bureau, and has met all requirements. Its height over-all is 2 ft, and it requires but 4 square feet of floor space. All the inside walls of the metal housing, in fact all surfaces with which either the crystals or the mother liquor can come in contact, are silver- or nickel-plated. The basket is carried on the end of the vertical shaft of a three-fourths horsepower motor. It has an inside diameter of 91⁄2 inches, and is capable of carrying 10 pounds of sugar. Its heavy cover is held in place by a number of set screws, A, A, and

is readily removable in order to facilitate the removal of the centrifuged material. The problem of a suitable lining has given considerable trouble. To be satisfactory it must retain very small crystals, permit free drainage of the mother liquor, and be able to stand the severe strains incidental to high speeds. It is obvious that no single lining is available that will meet all three requirements. A built-up lining was accordingly resorted to. It consists of two layers. The outer one is the regular copper centrifuge lining with elongated conical holes. The inner one is of 200-mesh brass gauze. Both linings are silver-plated.

In order to hold the flimsy gauze in place, it was necessary to fold it over the edge of the heavy copper lining, both above and below, and

FIGURE 100.-Centrifuge.

also around the ends where the vertical edges meet, in every case allowing it to overlap by 1 or 2 cm. At the junction of the two edges an extra piece of gauze was fastened to close the crack. Inasmuch as the wear on the gauze lining occurs at the top, it was protected by a piece of thin copper tape which was bent into place. This combination has given excellent results. The mother liquor finds an outlet through the drain, B. The small space left between the lid and the frame was closed by stretching a rubber band tightly around the whole

machine.

It will be observed that when the centrifuge is in operation with the hinged lid, C, closed, the contents of the basket, as well as the

mother liquor, are safe from contamination by the air of the room. The speed of rotation of the basket is controlled by a rheostat in series with the motor.

In order to secure a proper distribution of the crystals and insure smooth running of the basket, the crystal mass is introduced while the machine is stationary or running at very low velocity. The speed is gradually increased as the mother liquor runs off, until finally the maximum speed of 3,000 revolutions per minute is attained.

The substance is then recrystallized in a similar manner, but a second filtration is accomplished through the layer of asbestos in a glass filter, shown in figure 98, connected directly to the boiling apparatus which supplies the necessary vacuum. This asbestos filter effectively removes the shreds of filter paper which are almost invariably in the filtrate after a filtration through paper. After the solution has passed this filter, it comes in contact with nothing but clean glass, and all manipulation of the solution or crystals is carried out under glass cabinets, to reduce possible contamination with dust. The airdried crystals are powdered to dust in an agate mortar, of which the pestle carries a shield to prevent contamination, and are placed in a vacuum desiccator over lime.

(e) MODIFICATION OF METHOD OF BATES AND JACKSON [5]

A modification of the method of Bates and Jackson has been tried at this Bureau. A 50 percent solution is clarified with alumina cream, filtered and boiled below 35° C as described above. The boiling is continued to a concentration of 70 to 73 percent of sugar. poured into a crystallizing jar and precipitated by the addition of an equal volume of pure alcohol. The precipitate is separated centrifugally, washed with alcohol, and air-dried. This process is repeated. Before bottling, it is dried over lime in a vacuum.

The alcohol used for the precipitation of pure sugar should be highly purified with respect to acids or aldehydes. It is not essential that it be dry or free from other members of the alcohol group. The method of purification described by Dunlap [6] meets these requirements. Dissolve 1.5 g of silver nitrate in 3 ml of water, add to a liter of 95 percent alcohol in a glass-stoppered cylinder and shake. Dissolve 3 g of caustic potash in 10 ml of warm alcohol, and after cooling pour slowly into the alcoholic AgNO3 solution. Do not shake. Allow to stand overnight. Siphon off and distill.

(1) ANALYSIS

If sugar is to be used for the purpose of standardizing instruments, its purity should be ascertained. This is particularly true of samples required for polarimetric standards and for scientific data of all kinds. There are many sorts of impurities whose presence, if undetected, may lead to false conclusions in the interpretation of data. These impurities may be grouped into classes for the sake of convenience: (a) Soluble inorganic impurities, (b) soluble organic impurities which reduce alkaline copper, (c) soluble organic impurities which do not reduce alkaline copper, and (d) moisture.

Inorganic impurities may readily be recognized by an ash determination. To perform this, a quantity of the sugar should be weighed into a very carefully weighed platinum capsule and burned to a char.

The char should then be ignited in a muffle furnace at a low red heat, properly below 550° C, until the carbon has been consumed. The ash from a sample of properly purified sugar should not amount to 0.01 percent. It must be ascertained that the dish itself is of constant weight during a similar period of heating. Samples prepared at this Bureau have an ash content of less than 0.1 mg for a 5-g sample.

The estimation of very small quantities of reducing substances in the presence of the very large quantity of sucrose requires the employment of special methods. The usual methods in which copper sulfate is dissolved in caustic alkali are unsuitable because of the destructive action of the reagent upon the sucrose.

Bates and Jackson [5] have made a detailed study of the reducing action of pure sucrose on other alkaline copper reagents in which the concentration of the hydroxyl ion was diminished. They found that after several recrystallizations of the sucrose a minimum value for the reduced cuprous oxide was obtained with each of the reagents investigated, and that further recrystallizations failed to lower this value. It appeared, therefore, that either a constant quantity of reducing sugar was present, distributing itself in a constantratio between crystals and mother liquor, or sucrose itself effected the slight reduction of the copper. From this indirect method they concluded that if any reducing substance other than sucrose itself was present in their purified samples it was of the order of 0.001 percent and entirely negligible for most purposes. Of the several methods studied, Bates and Jackson found a modification of the Soldaini method to yield the most satisfactory results. The modified method is as follows:

Dissolve 297 g of KHCO3 and 1 g of CuSO4.5H2O in water and dilute to 1 liter. The potassium bicarbonate should not contain much K&CO. Transfer 50 ml of this solution to a beaker, boil for 1 minute, add a solution containing 10 g of the sugar, bring the whole to boiling, and continue the boiling for exactly 2 minutes. At the end of this period add 100 ml of cold, recently boiled distilled water and collect the precipitate on a Gooch-Monroe crucible or on a very closely packed asbestos mat. Under this procedure it has been shown that 10 g of pure sucrose produces 1.1 mg of Cu2O, while each 0.01-percent impurity of reducing substance, estimated as invert sugar, increases this precipitate by 1.9 mg.

Another method applicable to the estimation of very small quantities of reducing sugars in pure sucrose is that of Ofner, see page 193.

The soluble organic impurities, which do not reduce alkaline copper, are not so easily detected, but investigations at this Bureau have shown that recrystallization from aqueous solution or by precipitation with alcohol results, in every case examined, in a purification of the sucrose. This fact can only be shown by a polarimetric study.

In order to dry pure, powdered sucrose, care should be exercised not to subject the sample to a prolonged high temperature. Under the influence of an elevated temperature, sucrose undergoes a decomposition which is similar to a process of "caramelization." In table 44 is given the time at each temperature required to form "caramel" equivalent in reducing power to 0.01 percent of invert sugar. Caramel can be detected by its reducing action on the alkaline copper solution. Finely pulverized sugar can be dried at 70° C in a vacuum in 4 hours.

903232 - 50-27