(b) ABSOLUTE ROTATION OF NORMAL SUCROSE SOLUTION

The rotation of the normal sugar solution for λ=5461 A was found by direct measurement.

Normal sugar solution=100° sugar-40.763°.

(21)

Since the rotation ratio for the normal solution for λ=5892.5 A and λ=5461 A is shown by eq 20 to be 0.84922, the rotation of the normal solution for λ=5892.5 A is

Normal sugar solution=100° sugar=34.617°.

(22)

(c) ROTATORY DISPERSION CURVES OF QUARTZ AND NORMAL SUCROSE SOLUTION

The difference between the rotations of the normal quartz plate and the normal solution for λ=5892.5 A is shown to be 0.003° and for λ=5461 A 0.073°. The values indicate that the rotatory dispersion curves of plate and solution cross at about λ=5850 A. The reading of the normal solution on the true saccharimeter scale with the source λ=5892.5 A has been calculated to be 99.99°S.

(d) ROTATION DIFFERENCE, IN SUGAR DEGREES, FOR NORMAL SUCROSE SOLUTION BETWEEN λ=5461 A AND λ=5892.5 A

The difference in rotation in sugar degrees, for the normal solution on the saccharimeter, for the sources X=5461 A and λ=5892.5 A, was calculated from the absolute rotations, with the following result:

Saccharimeter reading (X=5461 A)-saccharimeter reading

(X=5892.5 A)=0.192S.

(23)

An independent experimental determination was made of this difference and the value 0.185° obtained.

(e) THICKNESS OF THE NORMAL QUARTZ PLATE

Inasmuch as the value of the conversion factor, i. e., the rotation of the normal quartz plate, is found to be 34.620° for λ=5892.5 A and 40.690° for X=5461 A, the old value of 1.5958 mm for the thickness of the normal plate is no longer applicable. Gumlich [20], as the result of a painstaking investigation, found the rotation of 1 mm of quartz for X=5892.5 A (the light traveling parallel to the optic axis) to be 21.7182° ±0.0005 at 20° C. Recently Lowry [21, 22] has made a number of measurements on the rotation of quartz and finds at 20° C 21.7283° per mm for (λ=5892.5 A) and 25.5371° per mm for λ=5461 A. The values of the thickness of the normal plate calculated from the above data are given in table 4. The argeement between the second and third values in column 4 is very satisfactory in view of the fact that two independent values of the rotation per millimeter are used. The agreement between Gumlich's and Lowry's values for sodium light is not satisfactory.

Of all the polarimetric constants relating to the sugars, none has received the thorough study by numerous investigators that has been given to the specific rotation of sucrose and its variations with concentration. The formulas of Tollens [23] and of Nasini and Villavecchia [24] giving the values at different concentrations have been generally accepted as the most accurate. Landolt [25] combined the two, giving [a]3892.5A=66.435+0.00870c-0.000235c2 (c=0 to 65), where c is the number of grams per 100 ml of solution. This equation gives a specific rotation of 66.502° for 26.016 g per 100 ml (vacuo).

From a critical survey of the work involving the specific rotation of sucrose, performed by prominent investigators in various parts of the world, it appears that the most likely value for this constant is very close to 66.53° for 26 g of sucrose in 100 ml of solution and for sodium light (weighings in air with brass wts.).

In the light of this work Landolt's formula has been adjusted to give 66.53° at 20° C for 26.016 g per 100 ml (weighed in vacuo). The adjusted equation follows:

[a] 892 5.4 66.462° +0.00870c-0.000235c2

(24)

This equation gives [a]=66.53° for 26.016 g of sucrose in 100 ml of solution and 66.54° for 16.280 g in 100 ml.

Bates and Jackson [18] in their investigation on the constants of the quartz-wedge saccharimeter made a determination of the specific rotation for two wave lengths. They found that the rotation of the normal solution for λ=5892.5 A is 34.617°, and for λ=5461_A_is 40.763°. Since this solution contains 26.016 g of sugar (weighed in vacuo) in 100 ml at 20° C,

Dextrose may be determined upon the saccharimeter, the readings. being directly in percent dextrose, provided the correct normal weight for this sugar is used. This saccharimetric constant, namely the weight of dextrose, which when dissolved in 100 ml of solution and

read in a 200-mm polariscope tube with white light and bichromate filter, gives a reading of 100° S on the International Sugar Scale (conversion factor, 34.620), has been carefully determined by Jackson [26] at the National Bureau of Standards and found to be 32.231 g weighed in air with brass weights, and 32.2515 g weighed in vacuo.

For concentrations less than normal, the rotations deviate considerably from proportionality. It is therefore necessary to correct for this deviation. Table 74, page 562, gives the corrections to be applied to the scale readings in order to obtain the true percentage of dextrose. These corrections are based upon the use of a standard 200-mm tube; hence, if any other tube-length is used, this fact must be taken into account. For example, if a 400-mm tube is used, giving twice as large a scale reading as the standard 200-mm tube for the same concentration of dextrose, the observed scale reading obviously must be halved before entering the table to obtain the proper correction.

(h) ROTATION OF NORMAL SOLUTION AND THE SPECIFIC ROTATION OF DEXTROSE Jackson found for the rotation of the normal dextrose solution the value 40.897° for the wave length λ=5461 A. The corresponding value for sucrose is 40.763° and the rotation of the normal quartz plate is 40.690°. There is thus a considerably greater difference between the rotatory dispersion curves of dextrose and quartz than between sucrose and quartz. This difference between dextrose and quartz is not as thoroughly eliminated by the bichromate filter as is the corresponding difference between sucrose and quartz (see fig. 19). A slight difference in color between the two halves of the field results when the quartz-wedge saccharimeter is set for a photometric match. This necessarily causes a lower degree of reproducibility for dextrose than for sucrose solutions. The difficulty is partially overcome by an increased number of settings or by increased experience on the part of the observer.

The specific rotation of dextrose solutions varies with the concentration according to the formula

[a]30=62.032+0.04257c,

where c is grams of anhydrous dextrose weighed in vacuo and contained in 100 ml of solution, or the formula

[a] =62.032+0.04220p+0.0001897p2,

A

where p is percentage dextrose by weight in vacuo.

(i) SPECIFIC ROTATIONS OF OTHER SUGARS

See tables 73, 75, and 76, beginning on page 562.

(j) CONVERSION AND SCALE COMPARISON FACTORS

Table 5(a) gives equivalents of various types of sugar scales in circular degrees.

Table 5(b) gives figures based upon the magnitude of the circulardegree rotations given in table 5(a) and therefore are useful only in giving an idea of the relative physical size or length of the different scales. These values cannot be used for converting from one scale into another, as the normal weight must also be taken into account.

However, they would hold as a conversion factor if, for example, 26 g were used as the normal weight on an instrument graduated on the French Sugar Scale.

TABLE 5.-Saccharimeter scale

[Normal weight 26.000 g, International Sugar Scale; 16.269 g, French Sugar Scale; 10.000 g, Wild Sugar Scale]

It has been unfortunate in the development of saccharimetry that more consideration was not given from the beginning to the question of suitable light sources and particularly to the influence of the source on the saccharimeter reading. It has been a more or less common practice among the users of saccharimeters to employ whatever source happened to be most convenient without adequate consideration of its effect upon the reading.

The light originally used in setting the 100° point of the saccharimeter was light from the Auer or Welsbach incandescent gas mantle, filtered through a proper bichromate filter. This has been largely displaced today by the electric incandescent lamp in various forms, which is far more satisfactory from the standpoint of constancy and ease of operation. A proper filter should be used, however, in every

case.

(b) BICHROMATE FILTER

While the saccharimeter is based upon the use of a quartz compensating system originally devised by Soleil, whose function is to balance out the rotation of the substance being measured by the insertion of a layer of quartz whose rotation is exactly the same in amount but opposite in direction to that of the substance being measured, yet the compensation is never absolutely complete. For white light containing all wave lengths of the visible spectrum, the compensation can be exact for only one substance, namely quartz.

Owing to the fact that the rotation dispersion curves of optically active substances are not identical, the quartz-wedge system does not completely return the polarization planes of all waves to their original

positions from which they had been rotated by the substance being tested. In the case of a solution of sugar the rotatory dispersion is nearly the same as that of quartz, but the divergence is sufficient in the green and blue end of the spectrum (see fig. 19 curves S, Q, and S-Q) to cause the halves of the field to appear of different tints. The field must appear uniform in color if the readings by different observers are to agree. Obviously the same instrument will also give different readings with different sources inasmuch as the luminosity curves of the sources are different. In testing sugar, the field may be made almost uniform in color for all incandescent sources of ordinary intensity by placing a cell of potassium bichromate solution between the polarizing system and the lamp. The function of this filter is to eliminate by absorption most of the shorter waves from the visible spectrum. Some saccharimeters are provided with a cell, fitted in the metal tube which houses the polarizing system, for containing the absorbing solution. Owing to their small diameter and also to the possibility of leakage inside the instrument, this Bureau has not found these cells satisfactory. Better results have been obtained by using a cell with a diameter of 40 or 50 mm placed between the saccharimeter and the light source. Any thickness may be used, but the optical path in the bichromate solution should, however, always be equivalent to a layer of liquid 15 mm thick for a 6-percent solution. If the cell is not 15 mm in thickness, the concentration of the solution must be changed accordingly. A simple rule, satisfactory for sucrose, is always to have the product of the thickness in millimeters and the percentage concentration equal to 90. In some instances where the rotation dispersion differs to a greater extent from that of quartz, as in the case of dextrose and some other sugars, it is found advisable to restrict the short-wave end of the spectrum still more by using 2 cm of a 9-percent solution, or the equivalent.

Figure 19 shows the importance of the use of a bichromate filter for white-light saccharimetry. The curve marked 6 gives the transmission (uncorrected for reflection) of a layer 1.5 cm thick of a 6-percent potassium bichromate solution, while that marked 9 gives the transmission of a 2.0-cm layer of a 9-percent solution. It will be noted that wave lengths to the left of these curves are absorbed while those to the right are transmitted. The curve marked Q shows the rotation, plotted against wave length, of the normal quartz control plate, while the curve marked S is the rotation of the normal sucrose solution. Because of the impossibility of plotting these two curves on a sufficiently large scale, they appear to be identical over most of their length, diverging slightly only in the blue, yet there are small systematic differences between them which are of great importance in saccharimetry. These differences, S-Q, have been plotted as a function of the wave length on a magnified scale, shown on the right, 100 times as great as that upon which S and Q separately are plotted. The dotted curve gives approximately the corresponding differences D-Q between the normal dextrose solution and the normal quartz plate.

The curve S Q shows the residual amount of rotation at each wave length produced by sucrose which is not balanced out by the quartz-wedge compensating system. It will be seen that there is considerable unbalance in the blue and even in the green but that light of these wave lengths is effectively eliminated by the bichromate