Sanders [45, 46] has applied the adsorption isotherm to calculate the amount of carbon to be used to reach a desired degree of decolorization and also to compute the decolorization and the carbon required in countercurrent work.

Although one may read the values of x/m and C from the isotherm plot and may calculate m from the relation x+C-1 (if color is expressed as a fraction), this computation is done more easily by reference to the nomograph shown in figure 80.

This nomograph solves the equation Y/M=K(1-Y)". This is the same as r/m=kC", where the color is a fraction of the original

FIGURE 79.-Adsorption isotherms of the same three decolorizing carbons shown in

figure 78 (plotted on a logarithmic scale).

k, A constant (value of ordinate, where log C=1); 1/n, a constant (slope of the isotherm or tangent of the angle formed by the isotherm with the log Caxis).

color, x=y and C-1-y. To exemplify the use of the chart, let us suppose that the values for K and 1/n were found to be 3 and 0.610, respectively, and that it is desired to know how much carbon is necessary to remove 70 percent of the color. On figure 80 draw a straight line from 1/n=0.61 to Y=70, and extend this line to the MK axis. The line extended cuts this axis at 1.45. Since K=3, then M=1.45/3=0.483 g (or pounds) of carbon per unit volume of solution. In describing a carbon in terms of K and 1/n, it is necessary to adopt standards of mass and volume of carbon and solution, respectively.

(b) GRANULAR BONE BLACK

The sample of bone black may be pulverized and tested for decolorizing efficiency in the same manner described for the char powders. However, since the two kinds of carbon differ so greatly in grain size

and in other properties, and since the mode of application in practice is different, the refinery chemist prefers to use a method that is more nearly representative of practical working conditions.

Experimental methods of evaluating bone chars are described by Wayne [53] and by Knowles [54], both of whom apply the char to the sugar solution on a weight basis. Blowski and Bon [49] apply the carbon on a volume basis. They place 170 ml of bone char in a 500-ml flask containing 10 ml of Filter-Cel and add 200 ml of a 47.5-Brix solution of crystallizer remelt sugar. The mixture is shaken and digested 3 hours on the water bath with 30-second periods of shaking every half hour. The mixture is decanted into paper filters and allowed to filter overnight. A blank and an arbitrary standard

FIGURE 80.-Nomograph for calculating, from the isotherms, the amount of carbon required to reach a desired degree of decolorization.

char are submitted to the test in parallel. The color remaining in each filtrate is determined and the decolorizing efficiencies are compared with that of the arbitrary standard.

In the Spencer-Meade handbook [55] is described a procedure in which the weighed bone black is placed in cylindrical copper funnels, diameter 4 inches and height 15 inches, cach provided with an outlet pipe and cock at the bottom. The char rests on a perforated copper plate covered with cloth. The funnels are placed in a water bath, provided with suitable openings for the outlet cocks and are filled to within a few inches of the top with the chars to be compared, the weight of char in each filter being the same.

A 55-Brix solution of a molasses sugar clarified by filtration with kieselguhr is used for the test. The liquor is heated to 165° F (73.9° C) and equal portions are added, little by little, to avoid air pockets. After the char is covered, the remainder of the liquor is poured into the filter. The temperature of the water bath is maintained at 160° to 170° F (71.2° to 77.2° C) for several hours, and the liquor is then drawn off through the outlet cocks. The color of the filtrates and of the untreated liquors are compared.

In this procedure it may be necessary to clarify the char filtrates, if char particles are present, by further filtration with kieselguhr, and changes in the concentration of dry substance may be detected by means of the refractometer.

11. REFERENCES

[1] K. Ventzke, Z. Ver. deut. Zucker-Ind. 10, 381 (1860); 11, 495 (1861). [2] K. Stammer, Z. Ver. deut. Zucker-Ind. 11, 41 (1861).

[3] F. J. Bates and associates, Int. Sugar J. 22, 654 (1920) (abstract).

[4] H. H. Peters and F. P. Phelps, Tech. Pap. BS 21, 261 (1927) T338.

[5] L. Eynon, Committee on Publication (1936), 7 & 8` Idol Lane, London, E. C. 3, England.

[6] K. S. Gibson, J. Opt. Soc. Am. & Rev. Sci. Instr. 10, 169 (1925); 11, 359 (1925).

[7] K. S. Gibson and H. J. McNicholas, Tech. Pap. BS 12 (1919), T119.

[8] K. S. Gibson, H. P. T. Tyndall, and H. J McNicholas, Tech. Pap. BS 13 (1920), T148.

[9] K. S. Gibson, J. Opt. Soc. Am. 24, 234 (1934).

[10] H. J. McNicholas, BS J. Research 1, 793 (1928) RP30.

[11] Spectrophotometric Equipment (Bausch & Lomb Optical Co., Rochester, 'N. Y.)

[12] Optical Instruments of Recent Design, Bul. 126 (The Gaertner Scientific Corporation, 1201 Wrightwood Ave., Chicago, Ill.)

[13] C. W. Keuffel, J. Opt. Soc. Am. & Rev. Sci. Instr. 11, 403 (1925).

[14] Advertising Leaflets, Palo-Myers, Inc., 81 Reade St., New York, N. Y.

[15] G. P. Meade and J. B. Harris, J. Ind. Eng. Chem. 12, 686 (1920).

[16] F. J. Bates, Bul. BS 2, 239 (1906) S34.

[17] J. F. Brewster, J. Research NBS 16, 349 (1936) RP878.

[18] J. F. Brewster, Facts About Sugar 28, 228 (1933) (abstract.)

[19] K. S. Gibson, J. Research NBS 14, 543 (1936) RP785; J. Opt. Soc. Am. 25,

131 (1935).

[20] C. Pulfrich, Z. Instrumentenk. 26, (1929). [21] Pulfrich Photometer, Advertising booklet.

New York, N. Y.)

(Carl Zeiss, Inc., 485 Fifth Ave.,

[22] E. Landt, Z. Ver. deut. Zucker-Ind. 84, 954 (1934).
[23] R. Müller, Ind. Eng. Chem., Anal. Ed. 11, 1 (1939).
[24] K. S. Gibson, NBS Letter Circular LC 473 (1936).

[25] K. Sandera, Z. Zucker-Ind. čechoslovak. Rep. 31, 261 (1927–28).
[26] P. Jackuschoff, Deut. Zucker-Ind. 56, 859 (1931).

[27] A. L. Holven and T. R. Gillett, Facts About Sugar 30, 169 (1935).

[28] J. C. Keane and B. A. Brice, Ind. Eng. Chem., Anal. Ed. 9, 258 (1937).

[29] E. Landt and H. Z. Hirschmüller, Z. Wirtschaftsgruppe Zucker-Ind. 87, 449 (1937).

[30] Th. Herke and N. Rempel, Deut. Zucker-Ind. 59, 379, 397 (1934).

[31] Brukner and Becker, Deut. Zuckerind. 59, 692 (1934).

[32] A. R. Nees, Ind. Eng. Chem., Anal. Ed. 11, 142 (1939).

[33] B. Lange, Z. Instrumentenk. 53, 344, 379 (1933).

[34] H. Hirschmüller, Dessertation, Univ. Berlin, June 10, 1938.

[35] B. Lange, Photoelemente und ihre Auwend ing (J. A. Barth, Leipzig, 1936). [36] H. Krefft and E. Summerer, Das Licht 4, 1, 23, 86, 105 (1934).

[37] K. S. Gibson, J. Opt. Soc. Am. 2, 23 (1919).

[38] R. B. Withrow, C. L. Shrewsbury, and H. R. Kraybill, Ind. Eng. Chem., Anal. Ed. 8, 214 (1936).

[39] J. F. Brewster and F. P. Phelps, Ind. Eng. Chem., Anal. Ed. 2, 373 (1930). [40] H. H. Peters and F. P. Phelps, Tech. Pap. BS 21, 271 (1927) T338.

[41] Jena Fritted Glass Filters, Advertising booklet, Fisch-Schurman Corporation, 230 E. 45th St., New York, N. Y.

[42] J. F. Brewster and F. P. Phelps, BS J. Research 10, 365 (1933) RP536. [43] T. Tadokoro, J. Chem. Ind., Japan, 21, 405 (1918). [44] M. T. Sanders, Chem. Met. Eng. 28, 541 (1923). [45] M. T. Sanders, Ind. Eng. Chem. 15, 781 (1923). [46] M. T. Sanders, Ind. Eng. Chem. 15, 785 (1923).

[47] M. T. Sanders, Ind. Eng. Chem. 20, 791 (1928).

[48] J. E. Teeple and P. Mohler, Ind. Eng. Chem. 16, 498 (1924).

[49] A. A. Blowski and J. H. Bon, Ind. Eng. Chem. 18, 32 (1926).

[50] J. Dedek, Sugar 29, 255, 307 (1927).

[51] H. Freundlich, Z. physik. Chem., 57, 385 (1907).

[52] H. Freundlich, Kapillarchemie, p. 232 (Akadem. Verlags Ges. M. B. H. Leipzig, 1922). English translation by H. S. Hatfield, Colloid and Capillary Chemistry (Mathuen & Co., Ltd., London).

[53] T. B. Wayne, Ind. Eng. Chem. 20, 933 (1928).

[54] H. I. Knowles, Ind. Eng. Chem. 19, 222 (1927).

[55] G. L. Spencer, Handbook for Cane Sugar Manufacturers and Their Chemists, 7th ed. Revised by G. P. Meade (John Wiley & Sons, New York, N. Y., 1929).

[56] D. B. Judd and K. S. Gibson, J. Research, NBS 16, 261 (1936).

[57] R. T. Balch, Ind. Eng. Chem., Anal. Ed. 3, 124 (1931).

[58] F. W. Zerban and L. Sattler, Ind. Eng. Chem., Anal. Ed. 8, 168 (1936).

[59] V. Mastalir, Z. Zuckerind. čechoslovak. Rep. 56, 337 (1931-32).

[60] P. Honig and W. Thomson, Mededeel. Proefsta. Java-Suikerind vol. 38, p. 1417 (1930).

[61] L. A. Wills, Facts About Sugar 21, 1114 (1926).

[62] O. Spengler and E. von Heyden, Z. Ver. deut. Zucker-Ind. 81, 693 (1931).

XX. MEASUREMENT OF TURBIDITY

1. METHOD OF BALCH

The method proposed by Balch [1] for the measurement of turbidity in sugar products is based upon the principle of measuring the transmittancy of light through the turbid solution, matched against a portion of the same solution from which turbidity-causing material has been removed by filtration. The color of the turbid solution is thus compensated by the filtered portion.

The procedure described here is taken mainly from Balch's article. The instrument used was a Keuffel & Esser spectrophotometer, all measurements being made at wave length 560 mμ.

Juices and many factory sirups may be filtered without dilution. In the case of high-density or solid products, a density greater than 60 percent of solids is not recommended. A 100- to 200-g portion of the turbid sugar solution to be tested is filtered with 2 to 10 percent of kieselguhr (based on sugar solids). The quantity of filter aid to be used depends upon the quantity and character of the suspended material affecting filtration. After mixing the kieselguhr very thoroughly with the sample, it is filtered at room temperature under diminished pressure through paper in a Büchner funnel which may be attached to a fractional distillation receiver whereby containers may be changed without breaking the vacuum. The filter apparatus of Zerban and Sattler (fig. 77, p. 328) serves the same purpose. After 25 to 50 ml of the solution has filtered, the remainder is received in a clean container. This last portion should appear brilliantly clear by transmitted light.

The turbid untreated portion of the solution is freed from coarse suspended material by straining through 200-mesh bolting silk, and freed from air bubbles by subjecting the sample to vacuum, con

veniently in the fractional distillation receiver at the time of filtration. The percentage of solids is determined with a refractometer.

In one of a matched pair of absorption cells is placed a portion of the clarified solution and in the other a portion of the strained and bubble-free turbid solution. The choice of thickness of the absorbing layer to be chosen depends upon the depth of color and turbidity of the solutions. For white sugars, cells 10 cm long or longer are needed, while for raw sugars a thickness of 1 cm or less is sufficient. The cells are mounted on the spectrophotometer and the transmittancy readings are taken in the usual manner. The accepted transmittancy value (the average of a number of readings) is then reduced by means of the Lambert-Beer formula to the specific absorptive index,

1
—log t=-(—log T).
cb

Since the color of the turbid solution is compensated by that of the clarified solution, log t is taken to represent the turbidity. By replacing the turbid solution in its absorption cell with distilled water, the transmittancy of the filtered solution may now be measured and calculated to log t to give a measure of color.

Since, in the Balch method, the turbidity value is taken as the difference in log t of a filtered and an unfiltered solution, this value may be similarly obtained with any instrument capable of yielding transmittancy readings at the same wave length. For instruments of the Duboscq type, as previously described (p. 315), where thickness of the absorbing layer is the quantity observed, reliance cannot be placed upon direct matching with a standard, as in filtered solutions, because of the uncertainty of the Lambert law applying to turbid solutions. A procedure may be used with the Duboscq wherein -log T for 1-cm thickness of the filtered solution is first found as prescribed, and the latter is then matched against the turbid solution, which is held at fixed thickness. This may perhaps be made more clear by the following example.

The solution (washed raw sugar) was filtered through a loose plug of absorbent cotton and a portion was filtered with asbestos. The absorption of the glass standard, -log T, at 560 mμ was 0.31069, and the thickness of the filtered solution at match was 1.821 cm. Then -log T for b=1 cm is 0.31069/1.821 0.17062. After replacing the filtered solution in one cup with turbid solution, the thickness of the latter in the colorimeter was set at 1 cm, and an intensity match was obtained by adjusting the filtered solution, no standard plate being used. The observed average thickness found was 3.313 cm. Deducting 1 cm, allowance for ordinary absorption in the turbid solution, we have 2.313 cm. Multiplying by -log T for b=1 cm in the filtered solution, we have 2.313X0.17062=0.39464, the value to be called -log T for a thickness of 1 cm in the turbid solution. The values for -log t are then calculated in the manner prescribed.

2. METHODS EMPLOYING THE ZEISS NEPHELOMETER

(a) DESCRIPTION OF THE APPARATUS

An early design of the Zeiss nephelometer was described by Sauer [2]. The following is a description of a later model, and a diagram of the instrument is shown in figure 81. A more detailed description is to be found in the Zeiss advertising pamphlet [11]: