5. Record

(a) c=grams of dry substance per milliliter of final filtrate. (b) b=thickness of the absorption cell used.

(c) Tor-log T for each wave length as read.

6. Reduce the readings obtained to unit basis as regards concentration and thickness by calculating specific absorptive index,-log t from the equation, log t=1/cb (-log T).

When color is diluted record also:

7. For volumetric method, procedure (a):

(a) Grams of dry substance per milliliter (=c) of turbid solution.

(b) Volume of turbid solution.

(c) Volume of diluent solution.

(d) Grams of colored dry substance per milliliter in mixture. 8. For gravimetric method, procedure (b), for correction of -log t (a) Refractometric Brix of diluent sirup.

(b) Weight of diluent sirup or solid sugar.

(c) Specific absorptive index, -log to of diluent.
(d) Refractometric Brix of colored sugar product.
(e) Weight of sugar product taken.

After filtration with asbestos, record:

(f) Refractometric Brix of final filtrate.

For qualitative and quantitative investigation of sugar color, readings are made at wave-length intervals of 10 or 20 mu over that part of the spectrum included between 420 and 700 mp. For routine. technical purposes a reading at the single wave length 560 mμ is sufficient, which, with the use of the filters designated for the purpose, may be done directly with a simple instrument or by calculation from readings at X546.1 and X578 mμ of the mercury arc (see paragraph 10, below).

9. For some purposes it is convenient to reduce log t at 560 mμ to n units of coloring matter (equivalent to n sugar color degrees). This is to be done by dividing -logt by the absorption unit, -log t, of the standard at the same wave length. This value (-log t1) for the provisional standard adopted is 0.00485 at X560 mμ for all types of coloring matter usually occurring in sugar products. It is now known as the Peters and Phelps color unit. One absorption unit at 560 mμ is equivalent to one unit of coloring matter which evokes a color sensation of one color degree. The value, n, absorption units, represents the sum total effect of absorption at all wave lengths. It is therefore a measure of n units of coloring matter in 1 g of sugar-dry substance and may be used as the final value to be regarded as characteristic of the sample measured.

In all cases additional measurements at other wave lengths than 560 mp, especially the shorter wave lengths (for example, the Hg lines 546 mμ and 436 m), may be made to yield valuable information as to changes taking place in various processes of manufacture, especially when studied as absorption ratios.

10. When the mercury arc is used as light source, a reading at 560 mμ may be obtained by a process of interpolation between readings at 546.1 and 578 mu or by applying a correction factor as follows: Deduct 48 percent of the difference between -log t at λ546.1 and X578 from --log t at X546.1; the result is -log t at X560.

903232 - 50 - 23

9. COLOR STANDARDS FOR SOLID SUGARS

Reference has already been made in this chapter to the measurement of reflectance from a smoothed surface of solid sugar, using certain of the photometric instruments described in previous sections. In some countries duty on imported sugar is charged according to color as well as to polarization, i. e., a higher duty is paid on a sugar lighter than a certain standard than on one that is darker. As a basis for color comparison, the well-known Dutch Standards are used.

The Dutch color standards consist of a series of samples of crystalline sugars of approximately equal gradations of color ranging from No. 7, which is very dark, to No. 25, which is almost white. Number 16 Dutch Standard is the technical color distinction in some countries between refined and raw sugars. The color grade of any given sample is determined by comparison with the standard series. The standards contained in sealed bottles are prepared for the sugar trade by a firm in Holland. They should be renewed once a year because of the tendency of the colors to change.

Color standards for soft sugars were prepared from ground glass by Wills [61]. The glass specimens (designated as colorless, light brown, dark amber, and uranium yellow) after grinding, were mixed in varying proportions and graded to give gradually increasing color. In all, 26 samples were formed. Later, the reflectance of these samples was measured with a Keuffel & Esser color analyzer and any necessary adjustments were made.

Raw-sugar color standards were prepared by Spengler and von Heyden [62] by coating white sugar crystals with yellow ochre (dark), cadmium yellow, and Norit, in various proportions. The pigments were first thoroughly mixed with a 67-percent solution of white sugar. The crystalline sugar, after mixing with the pigment suspension, was centrifuged and the coated sugar first air-dried, then heated in the oven for 1⁄2 hour at 70° to 80° C. The cooled samples were then placed in containers having Cellophane windows.

10. DECOLORIZING EFFICIENCY OF CARBONS

(a) FINELY DIVIDED CARBONS

(1) DETERMINATION OF MOISTURE. Decolorizing carbons capable of retaining several percent of moisture, depending upon humidity conditions, and the main sample should therefore be stored in a tightly stoppered bottle. In order to maintain a uniform basis of comparison among different lots of carbon, it is necessary that the results of decolorizing tests be calculated to moisture-free material. Approximately 2 g of carbon is weighed in a large, tared weighing bottle provided with a stopper, and heated 12 hours or overnight in an oven maintained at 150° C. At the end of the heating period, the bottles are stoppered and transferred to a desiccator while still warm. The bottle with contents is weighed after cooling. The loss of weight of the sample is calculated to percentage of moisture.

(2) DECOLORIZING EFFICIENCY.-The removal of sugar color from solutions by carbon has been shown by several investigators [43, 44, 47, 48, 49, 50] to follow the general adsorption equation, or isotherm, expressed by Freundlich [51, 52] as follows:

m

where r is the amount of color removed by m grams of carbon, C is the amount of color left in the treated solution at equilibrium, and k and 1/n are constants, which for any given solution are characteristic of the carbon.

Writing the Freundlich equation in logarithmic form, we have

the equation of a straight line. If log x/m as ordinates is plotted against log C as abscissas on rectangular coordinate paper, log k is determined as the value of the ordinate, where C-1 (or 100, where percentage of decolorization is used), while 1/n represents the slope of the isotherm. The factor 1/n is equal to the tangent of the angle formed by the isotherm with the log C axis. The values for C and a are best expressed in terms of percentage of the color of the original solution before treatment. This in turn may be in terms of log t, -log T, or any of the usual color units. The value of m may be expressed as grams of carbon, or as percentage carbon on sugar-dry substance.

The extent of decolorization of sugar solutions by a carbon is dependent upon several factors, among which are time of contact, temperature, concentration, pH, and nature of the sugar product [49]. In comparing carbons, therefore, it is necessary that these conditions be kept uniform. Also, it is generally agreed that the test material should be similar to that to be treated on the large scale.

As test materials, diluted molasses or high-density solutions of raw sugar are most frequently used in the sugar industry. The former permits the more rapid work, whereas the latter is capable of affording more accurate results which may be made directly applicable to largescale practice. Washed raw sugar, dried at the ordinary temperature and well mixed, may be stored in large quantity and is recommended for the isotherm test.

Using washed raw sugar, 500 g is dissolved in 350 ml of water at 90° C, and 20 g of kieselguhr (or more if necessary) is added and mixed. The mixture is filtered in a large Büchner funnel fitted with two layers of paper (Whatman No. 42 or equivalent) which has been moistened and the excess water removed by suction. Small quantities of the mixture are filtered at first, and, if the filtrate appears to be perfectly clear, the main portion is filtered. In case of cloudy first-runnings, the mixture may be added little by little until a sufficient amount of kieselguhr has become deposited to give a clear filtrate. The funnel is then transferred to a clean receiver and the filtration completed, including the cloudy first-runnings. filtrates may be improved by a second filtration, through asbestos, using a 120-ml Jena No. 1 funnel. The solution is cooled, thoroughly mixed, and the Brix is determined with the refractometer.

The

For convenience, such amounts of the solution are taken for carbon treatment as to contain 100 g of dry substance each (found by dividing 100 by the Brix as a decimal, thus 100/0.60-166.67 g). This amount is weighed to within 0.1 g in each of four tared 250-ml Erlenmeyer flasks. These are placed in a bath held at 90° C and allowed to warm. In the meantime, 4 portions of the carbon under test are weighed out (to 1 mg) equivalent to 0.5, 0.7, 1.0, and 1.5 g (or percent

on sugar solids), moisture-free basis, each weighed portion being placed in a scoop or trough made by bending a 3- by 4-inch sheet of thin aluminum or brass. It may be necessary to vary the quantities of carbon from those given, depending upon the color of the solution and the decolorizing ability of the carbon. A weighed portion of the carbon is transferred to each flask, which is shaken continuously in the heating bath for 15 minutes. The carbon is best removed by filtration on an asbestos pad, using a Jena No. 1 filter. For paper filtration, approximately 1 g of kieselguhr is mixed with the contents of each flask, which is filtered by suction through a double layer of Whatman No. 42 paper, 7.5 cm in diameter, in a Büchner funnel. A battery of four filters with clean, dry receivers may be arranged and an extra receiver used for starting the filtration of each solution, in turn. When the filtrate becomes perfectly clear, the filter is placed in its receiver and the filtration completed. The decolorized solutions are cooled and thoroughly mixed, the refractometer Brix of each is determined, and the solutions, including the original undecolorized liquor, are evaluated photometrically, preferably at 560 mu, the thickness, b, of the light-absorbing layer being governed by the color of the solution. From the Brix, layer thickness, and photometer readings (giving c, b, and T for each solution), -log t is calculated as already described. These values of log t, without conversion to color units, are used for computing the values for plotting the isotherm.

Table 40 illustrates the manner of computing the isotherm data. The values of C are found by (-log t decol./-log t orig.) X100 to give a whole number and x=100-C. The values of m as given may be called the percentage of carbon on saccharine dry substance.

TABLE 40.--Computation of data for the isotherm in figures 78 and 79

The values of log r/m may now be plotted against log Cas rectangular coordinates, or as recommended by Sanders [45], r/m and Care plotted on double logarithmic paper, as shown in figure 79, the data being those of table 40.

Diluted molasses is preferred by many as a material for the decolorization test, since the solutions are easily filtered without suction. Diluted molasses is unstable in color, and stock solutions should therefore be prepared at least daily. This stock solution should not be too dark for ease of color determination or so light that it is too easily decolorized.

The procedure as outlined by Sanders [44] is as follows: Equal volumes (150 ml) of the molasses are pipetted into five 250-ml Erlenmeyer flasks. These are placed in a water bath and heated to 90° C. On each of five strips of white glazed paper 1 g of Filter-Cel is placed. Four portions of carbon are weighed, using, for instance, 0.1, 0.2,

0.4, and 0.7 g weighed to the nearest mg, allowance being made for the moisture content of the carbon. Each weight of carbon is placed on the Filter-Cel on each of the strips. The Filter-Cel alone is added to one of the flasks and the mixture of Filter-Cel and carbon to the other four. The heating is continued during 10 minutes after the addition, the flasks being shaken continuously. The flasks are removed from the bath and their contents filtered through paper on separate 60° funnels. Cloudy first-runnings are discarded or returned to the filter, only brilliantly clean filtrates being accepted for colorimetry. After cooling, the filtrates are measured photometrically, preferably at wave length 560 mu. Concentration of dry substance is not taken into account in this procedure. Thicknesses

FIGURE 78.-Adsorption isotherms of three decolorizing carbons. C Concentration of coloring matter remaining in solution plotted against x/m, where x is the amount of coloring matter removed by m grams of carbon.

(d) are reduced to 1 cm by dividing -log T as found or calculated from the photometer reading, by the length in centimeters of the cell used. If Meade-Harris color units are used, the number of units found, corresponding to the photometer scale reading, are divided by the cell thickness in centimeters.

Depending upon the system used, either-log T or the MeadeHarris color unit of the untreated molasses solution is taken as the basis for calculating the percentage of original color remaining, (C), and the percentage of color removed, (x). Knowing m from the weight of carbon, x/m is computed, and the results are plotted on double logarithmic paper, as recommended above.