ANALYTICAL CHEMISTRY
1494 METAL ANALYSES
Am. SOC.Testing Materials, ”Stmdards on Petroleum Produrts
and Lubricants,” A.S.T.M. Committee D-2. Clark, G . L., “Applied X-Rays,” 3rd ed., New York, McGrawHill Book Co., 1940. Conigton, A. H., and Allison, S. K., “X-Rays in Theory and Experiment,” 2nd ed., New York, D. Van Nostrand Co., 1943. Gross, W., and Staab, F., German Aeronautical Research, U. S. Dept. Commerce, Research R e p t . 1476 (1941). Kehl. W.L., and Hart, J. C., paper presented at 28th Annual A.P.I. Meeting, Chicago, November 8, 1948. Liebhafsky, H . A., Smith, H. M., Tanis, H. E., and Winslow, E. R., ANAL.CHEM.,19, 861 (1947). Michel, P. C., and Rich, T. A , , Gen. Eke. Rea., 50, 45 (February 1947). Sullivan, M. V., and Friedman, Herbert, IND.EXG.CHEM., AXAL.ED.,18, 304 (1946). Trost, A , , Phys. J . , 115, 456 (1940). Zemany, P. D., Winslow, E. H., Poellmita, G. S., and Liebhafsky, H. A., ANAL.CHEM.,21,493 (1949).
The manufacture of metal-organic derivatives requires a great deal of inorganic analytical testing to arrive at the desired metal content. All the metals involved are of considerably higher atomic number than carbon and are frequently present in comparatively large amounts. I t is possible to prepare standards of varying metal content and calibrate to a sensitivity as close as the usual routine analytical accuracy. The proper choice of sample weight is an important factor on samples of high metal content, for beyond a certain mass the x-rays are so completely absorbed that insufficient impulse remains to activate the amplifier. 4CKNOWLEDGMENT
The authors wish to express appreciation for the helpful criticisms and suggestions received from John Y. Beach of the California Research Corporation. LITERATURE CITED
(1) Aborn, R. H.. and Brown, R. H., IND. ESG.CHEM.,AN*L. E.D ,
1,26 (1929)
RECEIVEDApril 19, 1949. Presented before t h e Division of Petroleum Chemistry a t the 115th Meeting of the AMERICAN CHEMICAL SOCIETI-, San Francisco, Calif.
Preparation of Sugar liquors and Sirups for Color Determinations P. F. ME.4DS
AND T.
K. GILLETT
California and Hawaiian Sugar Rejining Corporation, Ltd., Crockett, Calif. Definite procedures for clarification and pH and density adjustments, essential for accurate determination of colors of sugar solutions, have been investigated in some detail in connection with a program of replacing visual color methods with photoelectric determinations. The results of this investigation led to the selection of an optimum method of preparing sugar liquors and sirups for routine color determinations. The sugar products to be tested usually require adjustment in density, to permit filtration and to give a color reading that will fall within the effective range of the color instrument. A schedule of standard dilutions has
T
H E accurate measurement of the colora of sugar liquors and sirups is of considerable importance in the sugar industry for controlling operations and maintaining product quality. This measurement depends primarily upon two factors: an accurate and reliable instrument for the determination of color, and a standardized procedure for the preparation of solutions for color determination. The first factor has been investigated in considerable detail in this laboratory and several photoelectric instruments have been developed for mearmring colors of process liquors (8), refined white sugars ( 7 ) , and soft sugars ( 6 ) . The development of these instruments has resulted in the replacement of all the visual color methods formerly used in this laboratory with photoelectric methods which are murh more accurate and reliable. The second problem of preparing solutions properly for color determinations arises principally in connection with the routine analysis of sugar sirups and liquors. The colors of a wide variety of refinery products are measured in the photoelectric colorimeter developed for this purpose (8). These products vary from the lightrcolored washed raw sugar liquor, on the one hand, to the dark-colored affination green sirup on the other hand. There is
been established for determining the color of liquor and sirup samples normally encountered in refinery practice. Based on experimental work, a standard filtration procedure has been established to provide a clear sample and j e t avoid removal of color by the diatomaceous filter medium employed. istandard pH of 7.0 has been selected and all samples are adjusted to this pH value in order to avoid the effect of variations in color due to the original acidit? or alkalinity of the sample. Color readings are determined for solutions prepared in the prescribed manner and calculated to a standard reference basis (100q~solids). This calculation is brieflj discussed.
considerable difference in the densities, pH values, and clarities of the products. Obviously, a standardized method for preparing these materials for color determination is essential if comparable results are to be obtained. The present paper describes the results of an investigation which was carried on to determine the optimum procedure. PREVIOUS WORK IN T H I S FIELD
A considerable body of literature has been developed relative to the proper preparation of sugar solutions for color determinations. The most detailed of these procedures have generally involved rather technical and complex techniques which are difficult to apply t o routine laboratory analytical work. Most previous investigators agree that color determinations on sugar solutions should be made a t the highest possible density, preferably 55” to 65” Brix (3, 9, 14). In diluting sugar solutions below these high densities, colloidal material is apt to be precipitated, which, in some cases, can be removed only with great difficulty. Consequently, the sugar solutions must be maintained a t the higher densities to avoid these problems. How-
V O L U M E 21, NO. 1 2 , D E C E M B E R 1 9 4 9 ~.
1495
____.
during the course of this investigation: dilution, clarification, pH adjustment, and, finally, calculation (Water a n d colorless sugar solution used as diluents) of the resulting color to a uniform basis. Colorless Solids of Color Calcd. t o Dilution. Ii order that color readings will fall Dark Sugar Distilled Final Color 100% Solids of Samula Liauor Solution Water Solution Reading D a r k Liquor within the most sensitive range - of the instrument, MI. MI. MZ. % O Stammer Stammer adsorotion vessels a t different thicknesses mav be 29.5 43.7 1A 10 90 1.55 employed or the samples may be diluted. Some 29.6 2A 3.11 43.7 20 80 30.2 70 4.76 investigators have suggested the use of a heavy 43.7 3A 30 30.8 6.39 43.6 4A 60 40 density colorless sugar solution in order to ac31 .O 8.08 43.4 80 50 5A 31.3 9.8 23.5 60 40 6A complish the required dilution. Obviously, the use 31.3 1.58 90 4.95 1B 10 of such a product makes it very difficult to filter 32.4 3.30 80 9.80 2B 20 32.3 4.95 14.5 70 30 3B the solutions. Consequently, distilled water wati 32.0 4B 40 60 6.59 19.1 32.2 8.29 23.4 .. 50 considered preferable in this laboratory, as it not 50 5B only provided the necessary dilution but also greatly ___ simplified filtration. In view of the fact that the color of a given product does not ever, Zerban and Sattler (19) have pointed out that precipitation vary greatly from time to time, it has been found possible to of colloidal matter may occur even in the high density solutions. dilute each sample to some predetermined point to bring color Generally, the use of these high density solutions necessitates readings n-ithin the desired instrument range. Such dilutions filtering while hot. Brewster and Phelps developed a procedure give solutions varying from 30" to 40" Brix for washed raw sugar wherein the solutions were heated to 80" to 90" C. (3). More liquors to 2 to 5' Brix for affination sirup samples. Dilutions are recently, the Bureau of Standards has suggested a temperature made prior to filtration, so that possible precipitated colloidal of 50" C. ( 2 ) . Obviously, the use of high temperatures is undematerial is removed. This procedure has been satisfactory and sirable in a procedure for color determination because of the there has been no evidence of precipitation following filtration. development of color in heating. This comment is equally In Table I is shown a comparison between the colors of various applicable to the recommended practice of dissolving solid sugars samples of 44' Brix sugar liquor, obtained by diluting in one in boiling hot water ( 4 ) . case with distilled water and in the other case with a 44' Brix With regard to the filtration of sugar solutions for the removal solution of confectioners' sugar. In this comparison, there is of suspended matter, there have been some differences of opinion better agreement, between the solutions which were diluted with on the proper filter medium to be used. Peters and Phelps (9) and, water. In view of these and similar results, distilled water was later, Brewster and Phelps ( 3 ) established a procedure involving adopted as the diluent in the procedure developed. the filtration of the solution through a mat of specially prepared Clarification. 4 number of filtering media have been tested asbestos which they indicated gave a very clear solution without in this laboratory in order to find a suitable method for clarifying any removal of color. I n their opinion, the use of other materials the diluted sugar solutions. Some time was spent in attempting caused a loss of some color. Balch ( I ) , on the other hand, to develop the asbestos filtration method as recommended by filtered the bolution through kieselguhr, which he found did not abPeters and Phelps and improved by Brewster and Phelps for the sorb color. However, it is likely that no medium can remove all purpose of routine color determinations. However, this method the turbid matter without some effect on the color of the solution. was very time-consuming, did not, give clew solutions, and therek h a n and Sattler (12, 13) after an extended investigation fore did not seem suitable for routine analytical control. Consefound that results obtained when ashestos was used as the filterquently, it appeared that a filter aid such ns one of the diatomaing medium varied considerably, drpending upon how the ashestos pad ~ v pached. a ~ nhereas results with kieselguhr \\ere much ceous earths would be required. more uniform Their conclusions were that, of the filteiing In order to ascertain the most sat ctory material, R number media tested, kieqelguhr nas probably the most desirable, but that of filter aids of different types were tested. Table I1 sumselection n a s a matter of personal preference. This subject of marizes the results of one test series in which a number of diafiltration is w l l summed up by Spencer and IIeade (10): "In tomaceous filter aids JTere employed in different quantities, to view of the diveigenc~of opinion on asbestos and the tediouancss clarify solutions for color determinations. The filter aids used of the procpdure, it iq probable that kieselguhr is preferable for were as follows: routine work and, possibly, for all typey of work." Filter ;\id SO. Deacriiltion pH has a considerable effect on color: the higher the pH the 1 Uncaloined (natural) kieselguhr darker the appearance of the sugar solution. Consequently, 2 Calcined kieselguhr there is considerable recognition that the p H of sugar qolutions 3' Kieselguhr calcined with alkaline should be standardized before color determinations ma? be '. earth compounds: with increas!, ing flow rate characteristics made ( 4 ) , although in some instances, it may l x desirable to 7 .inalytical grade kieselguhr determine color at the original pH of the sample. 8 Factory regenerated kieselguhr Table I.
Comparison of Colors of a Sugar Liquor
INVESTIGATION
There are various objections to the methods for preparation of sugar products for color determination, as described in the literature, particularly from the standpoint of use for routine control. The present work has been directed toward providing a simplified, rapid, and accurate procedure. All the factors that influence the color readings of sugar solutions have been considered
Table 11.
Clarification of Washed Raw- Sugar Solution with Filter Aids"
Turbidity Color, O Stammer per 100% Solids 1 e./ 2 g./ 3 B./ 0.5 g./ 1 g./ 2 E./ 3 g./ 250 nil. 250 ml. 250 ml. 250 ml. 250 nil. 250 ml. 250 ml. 1.8 1.3s 1.1 1.0 8.07 8.0Ob 7.60 7.23 2.5 2.5 2.8 3.0 8.42 8.27 8.22 8.29 2.0 1.9 1.7 1.5 8.29 8.18 8.04 7.79 2.4 2.1 , 2.0 2.0 8.52 8.45 8.35 8.25 5 2.6 2.5 2.4 2.3 8.79 8.85 9.07 9.07 6 2.8 2.9 3.1 3.3 8.7.5 8.80 9.03 9.14 7 2.3 2.3 2.3 2.2 8.47 8.35 8.22 7.95 8 2.7 3.1 3.2 3.4 8.67 8.67 8.67 8.61 2.9 8.59 ~ ? ~ ~ : ~ ~ ~ l ~ 4.4 a p e r 8.85 a Washed raw liquor diluted to 35' Brix, filtered, a n d then adjusted t o pH 7.0. First 50 ml. of filtrate discarded a n d next 100 ml. collected for test. b Present standard method. Filter .iid 1 2 3 4
0 ,5 g , / 250 ml.
1496
ANALYTICAL CHEMISTRY
large number and represented median characteristics. In view of the arbitrary nature of the Stammer unit, it is: ultimately planned to adopt the more basic -log T unit, when this receives more general usage in the sugar industry. The color readings obtained on the photoelectric colorimeter, as described above, are calculated to a 100% solids basis, This involve. dividing the color reading by the solids content and specific gravity and multiplying by 100. Inclusion of the specific gravityisnecessary in the calculation to obtain comparable colors regardless of density. This is indicated by Figure 1. Effect of pH on Color of Typical Refinery Products Table 111, which shows the colors of a series of samples a t different densities calculated with and without the specific gravity factor. There is On the basis of these tests, uncalcined kieselguhr appeared to be the only filtering medium that would give a clear filtrate as considerable variation in the colors calculated without taking determined by a photoelectric turbidity indicator, based on specific gravity into account. measurement of the Tyndall effect ( 5 ) . Some kieselguhrs, particularly the calcined materials actually gave increasing turbidities as the quantity of filter aid was increased. Other materials Table 111. Colors of Washed Raw Sugar Liquor added color as the amount of filter aid was increased, although Determined at Various Densities the clarity was somewhat improved. Color, Stammer per 100% Solids __ The selection of a particular filter aid and the quantity of that Color With Kithout filter aid to be used represents, of course, a compromise among a Reading, s p . gr. s p . gr. % Solidi Stammer factor factor number of factors, including the resulting clarity which is obtained, the apparent removal of color, and the length of time required to obtain n suitable quantity of filtered material. In this laboratory, it has been determined that the use of 1 gram of natural kieselguhr per 250 ml. of diluted liquor offers the optimum filtration characteristics. Tests have indicated that this quantity of kieselguhr absorbs very little color and gives satisfactory PROCEDURE clarity. I n practice, the first 50 ml. of filtrate are discarded, inasmuch as usually this material m a r be somewhat cloudy. On the basis of the foregoing, the procedure established is as The next 100 ml. are collected for color determination. follows: pH Adjustment, The color of a sugar solution changes with The sample is diluted to a predetermined point to give a suita change in alkalinity or acidity. Thr extent of this variation in able reading within the range of the instrument. color with the pH of three sugar products is shown in Figure 1. One gram of filter aid is mixed with 250 ml. of diluted sample. The sample is filtered through a coarse filter paper (rapid gray It is obvious from this figure that some standard p H value must sugar paper S o . 1241/2, Geo. D. Feidt & Co., Philadelphia, Pa.) be selected if color results are to have any meaning or are to be in a stemless funnel; the first 50 ml. are discarded and the nest suitable for comparative purposes. A p H value of 7.0 has been 100 ml. saved for color determination. selected in this laboratory. After the samples have been filtered, The filtered sample is adjusted to pH 7.0, using 0.0357 hi sulfuric acid or sodium hydroxide as the adjusting reagent. the p H is adjusted to the neutral point by titration with either The color is read in the colorimeter. 0.0357 N sodium hydroxide or sulfuric acid as required. The solids content of the sample is determined with the reCalculation of Color, The colorimeter used in this investigafractometer. tion is the photoelectric device described previously (8). This The color is calculated in accordance Kith the folloning formula : instrument was designed to read transmittance directly, and was therefore calibrated in terms of -log T . Color in degrees Stammer per 100% solids = The photoelectric colorimeter was also calibrated in Stammer color reading, O Stammer X 100%. units (II), inasmuch as the Stammer colorimeter had been in prior % solids X specific gravity use in this laboratory. This secondary calibration was possible because the Stammer scale is directly_ proportional to -log T. _ . SUMMARY Sumerous comparisons between visual Stammer readings and The procedure described has been in satisfactory use in this photoelectric color determinations developed the following laboratory for a number of years. Reliable and consistent rerelationship: Degrees Stammer = 11.0 (-log T ) where T is in sults have been obtained even by nontechnically trained laboratransmittance of light of 535 millimicrons dominant wave length tory personnel. The method is relatively simple and rapid, and through an absorption vessel of 3-cm. thickness. consequently has proved very suitable for process color control. This relationship holds only for the particular Stammer glasses It is of especial value for investigative purposes. Previous used; however, these glasses were carefully selected from among a
V O L U M E 21, NO. 12, D E C E M B E R 1 9 4 9 methods did not give consistent or reliable color data and, consequently, much time and effort were uselessly expended in attempting to interpret color information. The situation has heen completely changed and color results are now accepted without question. This has saved a great deal of time and manjioivPr in plant and laboratory investigations. LITERATURE CITED
(1) Balch, It. T., IND.ESG.CHEM.,. 1 ~ . 4 ~ED., . 3,124 (1931). (2) Bates, F. J., and associates, Natl. Bur. Standards, Circ. C-440, 2G5 (1942). (3) Brewster, J. F., and Phelps, F. P., B u r . Standards J . Research, 10, 365 (1933) : Research P a p e r 536. 14) Hrowne, C. A , , and Zerban, F. W.,“Physical and Chemical llethods of Sugar Analysis,’’ 3rd ed., New York, John TYiley R- Sons, 1941.
1497 ( 5 ) Gillett, T. R., and Holven, 9.L., Inrl. Eng. Cheni., 28, 391
(1936).
(6) I b i d . , 35, 210 (1943). (7) Gillett, T. R., Meads, P. F.,and Holven, -\. L.. U. S.Patent 2,356,288 (Aug. 22, 1944). ( 8 ) Holven, A . L., and Gillett, T. R., Facts about S u g a r , 30, 169
(1935). (9) Peters, H. H., and Phelps, F. P., B u r . Standards Tech. P a p e r , 338 (March 12, 1927). (10) Spencer, G. L., and Meade, G. P., “Cane Sugar Handbook,” 8th ed., p. 474. New York, John Wiley BE Sons, 1945. (11) I b i d . . D. 479.
(12) Zerban, F. W., and Sattler, Louis, ISD.ESG.CHEM.,ANAL.ED.. 8, I68 (1936). (13) I b i d . , 9 , 229 (1937). RECEIVED April 15, 1919. Presented before the Division of Sugar Chemist r y a n d Technology a t the 115th Meeting of the .I\IERICAX CHEMICAL SoCIETY, Ran Francisco, Calif.
Rapid Measurement of Cellulose Viscosity by the Nitration Method W. .J. ALEXANDER
AND
K. 1,. JIITCHELL, Rayonier Incorporated, Shelton, IVash.
A siniplified technique is described for rapidly determining the nitrate viscosity of cellulose. Convenience in procedure is achieved by forming the pulp into soft, thin disks which may be easily handled in the various stages of nitration and stabilization. Degradation is kept to a minimum by using a small sample of open structure and further by using an optimum acid composition under favorable conditions of time and temperature. Data are presented to show the effecl of various changes in method on the level of calculated degree of polymerization.
T
HE nitration method for measuring cellulose viscosity has in recent years been widely used in both fundamental and development research dealing with the structure and properties of wlluloeic materials. It has not, however, been accepted by the pulp and paper industry as a control method for mill operations. .Uthough Berl ( 1 ) has suggested such use and has outlined a proi d u r e for obtaining rapid measurement of viscosity, it has usu:~llyhwn thought, and rightly so, that the method was highly coniI>lic,ztcdand n-ould yield precipe data only in the hands of a highly .killed analyst,. In :idapting the measureniei:t of nit,rate viscosity fur more general and practical use, the present method provides a much Yimplified technique n-hich gives highly reproducible data and t~liniinatcsmany of the difficult, tedious, or dangerous steps that h:ive formed the basis for most objections t o previous methods The nitrate viscosity method offers certain advantages which other viscosity methods do not afford: Because the solution is made in a simple organic solvent, a representative sample may be readily dissolved, aliquoted, and diluted t o any suitably ]om- con(.cntr;ition for viscometric determination of degree of polynierization. The degree of polymerization of the carbohydrate portion of unbleached pulps containing large quantities of lignin, or even o f wood itself, may be reliably measured, inasmuch as xvith suitable nitration conditions, the lignin is substantially removed in the nitration step and therefore does not interfere with the viscaosity measurement on the nitrated cellulosic constituents ( 6 ) . The nitration method is based on conversion of the cellulose sample into cellulose nitrate by a nitration step which does not appreciably degrade the cellulose (2, a), measurement of the viscosity of a dilute ethyl acet,ate solution of the nitrate, and calculation of the average chain length (3,4). The method in its present, status is an accurate and fairly rapid
means for nieasuring cellulose intrinsic viscosity and may well replace the very slow cupranimoniuni method or serve as a check and mean3 of calibration for the very fast cupriethylenediamint method. The time requirement is about 2 to 3 hours and the rpproducibility of degree of polymerization values is \Tithin 1%. The values a9 calculated are not presumed t o be absolute. They serve, nevertheless, as a convenient means for expressing the relative significance of measured tliffer~iicrsin solution viscositv IIETHOD
Preparation of Cellulose. The cellulose sample is prepared by remaking it into a thin soft sheet that \vi11 permit ready penetration of the nitrating acid. For wood cellulose this is best accomplished by dispersing the pulp fibers in water and flowing the slurry onto a sheet mold. About, 20 grams of pulp are used for a 13 X 13 inch (32.5 X 32.5 cm.) sheet which is lifted from the sheet mold and, without pressing, dried at 50” C. on a stainless wire screen support. Enough disks, 40 mm. in diameter, are cut from the dried sheet to give about 1.0 gram of pulp (three or four disks). The disks are further softened by flexing between the fingers, placed in a weighing bottle, and redried to a constant weight at 50” C. (about 99% bone dry) in a mechanical convection oven. Preparation of Nitrating Mixture. The nitrating acid mixture is prepared by adding cautiously, Tvith a spatula, 404 grams of‘ phosphorus pentoxide very slowly to 1000 grams of cold 90% nitric acid (907, fuming acid but not red fuming) contained in a 2-liter Erlenmeyer flask. The acid is kept ice-cold by imniersion in an ice water bath and is swirled continuously during addition of the phosphorus pentoxide. This produces a mixture with the composition: . nitric acid, 647,; phosphoric acid, 26% : phosphoric pentoxide, 10%. With occasional gentle shaking. solution is complete in a few hours and the acid mixture is then filtered t,hrough glass wool into a glass-stoppered bottle and stored in a cool dark place. Nitration. I n the actual nitration, a 40-gram portion of the prepared nitrating acid mixture is weighed out into a tall weighing bottle of about 100-ml. capacity, and placed in a constant
