124
ANALYTICAL EDITION
Vol. 3, No. 2
Measurement of Turbidity with a Spectrophotometer’” With Especial Reference to Sugarhouse Products R. T. Balch BUREAUOX CHEMISTRY AND
SOILS, WASHINGTON, D.
c.
As it is frequently desirable to measure the turfactors cannot be applied exROM the early days of bidity of sugar liquors during the course of sugar cept under rather ideal condithe sugar industry it manufacture or refining, a number of methods have tions. I n nephelometric anhas been known that previously been proposed. Adoption of these methods alysis (la),for instance, it is s u s p e n d e d solids in sugar has not, however, been extensive because, it is believed, possible to control conditions juices and sirups have detritoo large errors are introduced owing to lack of color sufficiently to obtain a conm e n t a l effect upon the compensation. In the method described, transmission siderable degree of precision, quality of the sugar produced of light at a definite wave length is taken as a measure but such is not the case when t h e r e f r o m . I n raw sugar of turbidity, and color is compensated by using as a attempting to determine the manufacture it has a bearing standard a portion of the same sugar solution from quantity of suspended mateon the cleanliness and filterawhich suspended solids are removed. For the preparial in a sugar liquor from a bility of the sugars, and in ration of this standard, filtration through paper with measure of its turbidity. The the manufacture of direct conthe aid of a slow-filtering grade of commercial kieselmost disturbing factors are sumption and refined sugars guhr is recommended in place of asbestos or sand. the varying composition and suspended material may be Turbidities are expressed in -log t values as being the the varying size of particles, responsible for the producmost convenient method, in spite of the fact that which covers the entire range tion of off-colored or dull-log T values are not strictly proportional to the confrom coarse to molecular disappearing sugars. centration or the depth of the suspension. persion. It is therefore frequently This is, perhaps, not as desirable to introduce a turbidity control of various clarification processes, but owing to serious as it appears since in practice the comparative turthe lack of adequate methods or standards, such a control has bidities and not the actual quantity of material present is the not been widely adopted. The principal objection to the chief concern. It does not matter greatly whether the parmethods recommended for use in sugar technology (4, 5, 8) ticles are relatively large or small, heavy or light in density, is that usually no provision is made, or it is inadequately except to suggest means for their removal. It is desirable, made, to compensate for the color of the product, which however, to have a method which gives comparable valuesmay introduce considerable error. Some of the methods that is, one which gives with sufficient accuracy the relative for determining turbidity require, too, the use of turbidity turbidities of sirups or juices of the same general type. As already indicated, one of the principal difficulties with standards which are frequently unstable and cannot, therefore, be highly recommended. The ideal method is one which existing methods is the lack of color compensation. Since so will permit the determination of turbidity in either colored many shades and intensities of color are encountered, it is or colorless solutions with equal accuracy, and in which the not very practicable to attempt the preparation and use of results may be expressed in numerical values with or, pref- synthetically colored solutions or glass screens. I n the erably, without reference to synthetic turbidity stan$ards. method which the writer proposes, color compensation is The method which is to be described is believed to approach accomplished by using a portion of the sugar solution to be tested from which material responsible for visible turbidity this ideal more closely than any yet proposed. has been removed. This immediately brings up the question Theory and Brief Description of Methods Used for whether it is possible to remove suspended material without Determining Turbidity altering the color and, if so, the best method by which this When a turbid solution is illuminated by a beam of light, may be accomplished. a portion of the incident light is reflected, a portion is abI n accurate colorimetric work, particularly in spectrosorbed by the suspended particles and by the solution, .par- photometric analysis, it is required that all material causing ticularly if colored, and a portion of the light is transmitted turbidity be removed from the solution before making the through the solution. One group of methods measures the color determination. For this purpose, various methods intensity of the scattered light or the Tyndall beam in rela- have been proposed. Peters and Phelps (7) recommend tion to the incident or standard light. Another group deter- filtration through specially prepared asbestos, whereas mines the masking effect or absorption of light by the par- Lunden (6)claims that ultracentrifuging or filtration through ticles (and color), the so-called extinction methods, or, in- sand are the only proper methods, and Spengler and Landt versely, the intensity or the percentage of the incident light (10) that special filters of other types are required. Honig that is transmitted. There are so many factors such as the and Bogstra (5) also recommend sand filtration but apnumber, size, shape, and other physical characteristics, parently adopt filtration through kieselguhr in spite of the of the particles which affect the intensity of either the Tyn- fact that various investigators claim that this agent has a dall or the transmitted light, that the mathematical expres- selective adsorption action on color. sions which have been developed (11) relating some of these It is very doubtful whether it is actually possible to remove 1 Received December 10, 1930. Presented before the Division of all the material causing turbidity, if judged by the Tyndall Sugar Chemistry at the 80th Meeting of the American Chemical Society, beam, without changing the color of the solution. It is Cincinnati, Ohio, September 8 to 12, 1930. well known that small particles within the range of colloid 2 Contribution No. 100, Carbohydrate Division, Bureau of Chemistry dimensions (5 to 200 mp) and even those which exceed this and Soils.
F
INDUSTRIAL AND ENGINEERING CHEMISTRY
April 15, 1931
size to some extent exhibit color phenomena by both reflected and transmitted light. Thus, if the clarification process is sufficiently efficient to remove suspended particles within this range of size, the apparent color of the solution would, without question, be changed to some extent. This fact accounts, it is believed, for the diversity of opinions as to the best method for clarifying sugar solutions for colorimetric analysis and the belief of some investigators that kieselguhr selectively adsorbs color. It is the author’s opinion that this agent removes only by mechanical means, and not by true adsorption, particles within the range of size which affects color to a slight extent. These particles are not completely removed by either asbestos or sand under the procedures generally recommended for preparing sugar solutions for spectrophotometric analysis. The results of some preliminary experiments conducted by the author substantiate this view. For example, a raw sugar solution, clarified with kieselguhr, had the same -log t value regardless of whether 1, 2, or 3 per cent of filter aid (based on sugar solids) was used or whether the solution was filtered immediately or after it had stood for some time. From our knowledge of adsorption, considered in its broader sense, one would expect differences in the color of the solution treated in these different ways unless it were assumed that maximum adsorption occurred with the smallest quantity filter aid used and in the minimum time period, which are very doubtful assumptions to make. Kieselguhr unquestionably removes particles of smaller size than is possible by filtration through asbestos or sand. However, the minimum size of particles which are removed by these various agents has never been determined. The experiments conducted by the writer indicate that true adsorption of color is lacking or extremely slight a t the most, so that it is permissible to use kieselguhr for the preparation of sugar solutions for either colorimetric analysis or as a standard in compensating for color in turbidity measurements.
125
a Keuffel and Esser color analyzer. Between the high and the low concentrations, a maximum variation of 15 per cent in the -log t value was observed, whereas with different length tubes the maximum variation in -log t values was about 10 per cent. These variations are probably somewhat exaggerated by the inability to obtain a correct aliquot of the stock suspension owing to the rapid settling of the larger particles. I n spite of these divergences from Lambert and Beer’s law, it is considered advisable to express turbidities as -log t values, as this affords a fairly close index of relative turbidities with sugar liquors of most types and undoubtedly better than by any other method so far proposed. In order to ascertain whether the method of color compensation is satisfactory and whether turbidities can be measured in colored solutions as accurately as in colorless solutions, two series of bentonite suspensions in sugar sirups were prepared. One sirup was colorless and the other was colored strongly with caramel to simulate a sirup prepared fr6m a dark raw sugar. The concentration of bentonite in each series ranged from 0.25to 2.00 grams per 100 cc. of final solution. The transmission of light was measured at 560 mp through a 1-cm. cell, again with a Keuffel and Ewer color analyzer. The standards for the two solutions were, of course, a portion of the colorless sirup for the one series of determinations, and a portion of the colored sirup for the other series. The data given in Table I show conclusively, it is believed, that when color is properly compensated, the transmission of light gives a very satisfactory measure of turbidity regardless of the color of the solution. Table I-Transmission
of Light through Bentonite Suspensions (At 560 mfi in 1-cm. cell)
BENTOSITE CONCENTRATION Gvams/100 cc.
TRANSMISSION THROUGH BENTONITE Colorless sirup
%
Colored sirupo
%
Description of Spectrophotometric Method for Measuring Turbidity
The method proposed for measuring turbidity is based upon the principle of measuring the transmission of light through the turbid solution, using as L? standard to compensate for color a portion of the same sugar solution from which material causing turbidity has been removed. The principles involved and the apparatus used are the same as for spectrophotometric color analysis (7, 9). Lambert and Beer’s law, which is used so extensively in spectrophotometric analysis, correlating transmission of light and concentration of color and thickness of solution, unfortunately does not hold exactly for turbid solutions. Until it is possible to develop more adequate relationships it is proposed to assume the validity of Lambert and Beer’s law and to express turbidities as -log t values in accordance with the following equation: -log t =
1 (-log T) cb
where t = transmittancy reduced to unit conditions as regards concentration and thickness of solution, c = concentration of the solution expressed as grams of dry substance per 1 cc.; b = thickness of solution expressed in cm.; and T = the transmitted fraction of the incident light. I n order to determine how closely this relationship might be expected to hold in the case of suspensions, a series of bentonite suspensions were prepared whose concentrations were made to range from 0.25 to 2.00 grams per 100 cc. (coarse particles included) by aliquoting from a 5 per cent suspension. The -log t values of these suspensions were then determined a t 560 mp in cells of different length with
a
Colored with caramel.
Preparation of Samples for Turbidity Determination
It is well known that dilution of many impure sugar solutions causes an increase in turbidity; hence it is important that the procedure for preparing the solutions for turbidity measurements be carefully standardized. Juices and factory sirups can, as a rule, be filtered with kieselguhr without dilution, a procedure which is recommended by all means. Difficulties exist, however, when dealing with solid and semisolid products, and undoubtedly differences of opinion will arise as to the selection of the density to which dilutions should be made. Since it is necessary to filter a portion of the sugar solution, it will probably not be practicable to exceed 60 per cent solids to any material extent except for a very few high-purity products; but for comparative purposes it is believed to be permissible to prepare a more dilute solution such as one of a 50 per cent solids concentration. Such details will have to be worked out by the person making the tests, keeping in mind, however, that whatever concentration is decided upon, all similar products must be diluted to the same concentration if comparative values are to be obtained. This is a rather unsatisfactory procedure to recommend, but it cannot very well be avoided. Clarification of that portion of the sample to be used for color compensation is accomplished by filtration with the aid of kieselguhr. Since it is desired to remove as much of the suspended matter causing turbidity as is possible, the
*
126
ANALYTICAL EDITION
Vol. 3, No. 2
slower filtering grades of kieselguhrs are more satisfactory solution which are removable by kieselguhr and which conthan the faster ones. The latter grades very frequently tribute to the color. To determine this point definitely, yield filtrates which are not entirely free even from visible however, would require a painstaking study of the size of turbidity. Even the best quality filtrates usually show a particles and of the relationship between the size of the pardistinct Tyndall beam when an intense source of light is ticles and the color of the solution, which is beyond the used and the observation is made in a dark room, so that this scope of this paper. criterion cannot be used and it seems necessary to adopt a As previously indicated, it seems preferable to calculate purely arbitrary method of filtration. The method to be the readings to -log t values in order to have a common described will, in the writer's opinion, give very satisfactory basis for comparing sugar products of different types. If relsults with practically all sugarhouse products with the the instrument is calibrated to read directly in -log T possible exception of those of low purity or those which have units, this value is simply divided by the product of the conundergone partial fermentation. The measurement of tur- centration of the solution (grams solids per cc.) and length bidity of these products will probably never be desired, and of cell (cm.) to obtain -log t; on the other hand, if the intherefore need not be considered. strument gives the percentage of light transmitted or T, A 100- to 200-gram portion of the sugar solution to be tested -log T is first calculated from the equation is filtered with 2 to 10 per cent of kieselguhr (based on sugar 1 -log T z log ?; solids). The quantity of filter aid to be used depends upon the quantity and character of the suspended material affect- which value, when treated as above, gives -log t. ing filtration. The more resistant to filtration this material is, the greater the quantity of kieselguhr required. I n genResults on Sugars eral, the minimum quantity stated will be sufficient to filter A number of granulated and raw sugars were examined high-purity products satisfactorily but, for cane juice or lime-defecated juices and sirups, the maximum will undoubt- by the spectrophotometric method for determining turbidity, edly be necessary. After mixing the kieselguhr very thor- and the results are given in Table 11. oughly with the sample, it is filtered at room temperature Table 11-Turbidities (-log t ) of Granulated a n d Raw Sugars under diminished pressure through paper in a Buchner funnel SUGAR c b T - log T - log; which is attached to a fractional distillation receiver. The source of vacuum may conveniently be a water-jet pump. Gram/cc Cm. After 25 to 50 grams of the solution have filtered, the re- Granulated 10 0.650 1 0.615 0,18709 0.030 mainder is received in another container, the change of con2 0.615 10 0.505 0.29671 0.048 3 0.615 10 0.685 0.16431 0.027 tainers being made without disturbing the vacuum. This 4 0.615 10 0.425 0.37161 0.059 last portion should appear brilliantly clear by transmitted 5 0.615 10 0.320 0.49485 0.081 10 0.000 6 0.615 0.22185 0.036 light of ordinary intensity. For more precise determinations 10 0.020 7 0.615 0.20761 0.034 10 0.325 8 0.615 0.48812 0.079 of the -log t values, it will be desirable to obtain as accurately 10 0.090 9 0.615 0.16115 0.026 as possible the solids content of this filtered sirup, but for most purposes, the per cent solids by refractometer will Cube 10 0.615 10 0.880 0.05562 0.009 suffice. Raw cane 0.615 1 0.62893 14 0.235 1.023 Measurement of Turbidity 0.46852 0.615 1 0.340 0.762 2" The clarified portion of the sugar solution is placed in one cell of the spectrophotometer as a standard, and in the other cell is placed the turbid portion freed from coarsely suspended material by straining through approximately 200-mesh bolting silk and freed from air bubbles by subjecting the sample to vacuum, conveniently in the fractional distillation receiver a t the time the filtration is being made. The length of cell to be used depends upon the color and turbidity of the sample. With the type of instrument used by the writer, 10-cm. cells were used for the 50 per cent solutions of granulated sugars, whereas for products as dark and turbid as raw sugar solutions of the same concentration, 1-cm. cells were used. The transmission measurement may be made a t any convenient wave length, such as a t 560 mp. As in all optical measurements which depend upon personal observation, the accepted reading is the average of a number of settings of the instrument. It was interesting to note that the per cent transmission of light through turbid sugar solutions, and to a lesser degree through bentonite suspensions, varied with the wave length of light, greater transmission occurring in the red end of the spectrum and less in the violet end. This probably indicates the presence of particles of such a size as to contribute to the color of the solution, a portion of them being removed by kieselguhr. The difference in percentage of light transmitted a t certain definite wave lengths at each end of the spectrum might conceivably be an approximate measure of the average or typical size of particles present in the sugar Standard Filter-Cel, a product of the Johns Manville Corp., has proved satisfactory.
0.615 0.615 0.615 0.615 0.615 0.615 0.615
3"
40 5b
;8C: 9C 5
Poor filtering sugars.
1 1
0.120 0.005 0.360 0.380 0.400 0.205 0.455
1 1 1
1 1
b Fair filtering sugar.
0.92082 2.30103 0.44370 0.42022 0.39794 0.53018 0.34199
1.497 3.742 0.722 0.683 0.647 0.862 0.556
c Good filtering sugars.
As was to be expected, cube sugar showed distinctly lower turbidity than standard granulated sugars. I n the case of raw cane sugars there is a general relationship between turbidity and filtering quality, as determined with the Dawson filter (6)but, since the specific physical character of the suspended matter has such a pronounced effect upon filtration ( I ) , a high degree of correlation could not be expected. I n Table I11 is shown a comparison between -log t values and percentage of suspended material contained in the sugar. The latter values are based on the differences in the weight of alcohol precipitates (50 per cent alcohol concentration) before and after filtration of a 50 per cent solution of the sugar with kieselguhr. Table 111-Turbidity
in Relation t o Suspended Solids i n Raw Sugar
~~
SAMPLE
SUSPENDED SOLIDS ON SUGAR
AT
-log t 560 r n N
% 4
0.131 0.084 0.084 0.065 0.042 0.036 0.030 0.027 0.026
3.742 1.497 1.023 0,762 0.862 0.683 0.556 0.722 0.647
INDUSTRIAL AND ENGINEERING CHEMIXTRY
April 15, 1931
The correlation in these instances is about as good as could be expected in view of the variation in the specific character of the material causing turbidity and the errors involved in the data, particularly the percentage of suspended solids. Conclusions
Although it has not been possible to test the spectrophotometric method for estimating turbidities very extensively, this method appears to have interesting possibilities. It is believed that the principle on which the method is based is satisfactory and that the method affords comparable results with various types of sugarhouse products. From a practical point of view, however, the method has a t least one drawback. The instrument used in this study is rather expensive, and as a>resultonly a few sugar laboratories are equipped with one but, since the light transmission needs
127
to be measured only a t a single wave length, it is possible that a.simplified apparatus would give satisfactory results. Literature Cited (1) Balch, Unpublished report. (2) Dawson, “Testing Apparatus and Method for Determining the Filterability of Raw Sugars,” paper presented before the Division of Sugar Chemistry at the 69th Meeting of the American Chemical Society, Baltimore, Md , April 9, 1925. (3) Honig and Bogstra, Facts About Sugar, 2 3 , 446, 470, 494 (1928). (4) Horne and Rice, IND. EKG.CHEM., 16, 626 (1924). (5) Ingersoll and Davis, Ibid., Anal. Ed., 2 , 248 (1930). (6) Lunden, Z. Ver. deut. Zuckerind., 77, 709 (1927). (7) Peters and Phelps, Bur. Standards, Tech. Paper 338 (1927). (8) Smith, Intern. Sugar J . , 26, 381 (1925). (9) Spencer-Meade, Handbook for Cane Sugar Manufacturers and Their Chemists, p. 293, Wiley, 1929. (10) Spengler and Landt, Z . Ver. deut. Zuckerind., 77, 431 (1927). (11) Wells, Chem. Rev., 3, 331 (1927). (12) Yoe, “Photometric Chemical Analysis. XI-Nephelometry,’’ Wiley, 1929.
Separation of Calcium and Magnesium by Molybdate Method’R. C. Wiley UNIVBRSITY
OF
MARYLAND, COLLEGE PARE, MD.
Calcium molybdate is quite insoluble in a solution calcium in solution was deterthat is nearly neutral. m i n e d by the m o l y b d a t e ring minerals contain Calcium may be determined accurately by precipitatmethod, and also by the percalcium and magneing as the molybdate. The presence of residual ammanganate method. The sosium. It is customary to monium molybdate in the washings from the calcium lution was about 0.1 N . separate the calcium from the molybdate does not prevent the precipitation of mag(2) A solution of magnemagnesium by making a nesium as the magnesium ammonium phosphate, and sium nitrate made by disdouble precipitation of the the presence of ammonium salts does not Interfere solving magnesium nitrate in calcium as the oxalate, evapowith the determination of calcium as molybdate. distilled water. The magr a t i n g the filtrates to a Calcium and magnesium may be determined in nesium in solutions was desmaller volume and then makmuch less time when the molybdate method for caltermined by precipitating as ing a double precipitation of cium is used than when the calcium is precipitated by magnesium ammonium phosthe magnesium. There is no the oxalate method. phate, and i g n i t i n g and question that this method of weighing in the usual way. d e t e r m i n i n g- calcium and magnesium is accurate. However, while it is accurate, it is The solution was about 0.115 N . (3) A solution of ammonium molybdate about 0.4 N , likewise tedious and time-consuming. The object of the present investigation was to find, if possible, a method whereby slightly acid with acetic acid. (4) A solution of ammonium molybdate about 0.4 h’ the double precipitations could be eliminated, thus saving a slightly alkaline with ammonium hydroxide. considerable amount of time. Smith and Bradbury (1) found that calcium molybdate Plan of Procedure for Precipitation in Alkaline Solution was only slightly soluble in water. Magnesium molybdate, on the other hand, is quite soluble in water. For these The calcium chloride solution and the magnesium nitrate reasons the writer thought it might be possible to separate these two elements by precipitating the calcium as the solution were carefully pipetted into the same beaker. The molybdate, washing the calcium molybdate well, and then solution was heated to boiling, a drop or two of concentrated precipitating the magnesium from filtrate and washings as ammohium hydroxide added, and the molybdate solution magnesium ammonium phosphate in the usual way. It was added at the rate of about one drop per second until an excess hoped that the presence of the molybdate in the filtrate would was present. The solution was then boiled until the supernot interfere with the precipitation of magnesium as mag- natant liquid was clear, and allowed to stand until cool. The precipitate was filtered from the solution by means of a tared nesium ammonium phosphate, as does an oxalate. Gooch crucible in which the layer of ordinary asbestos had been reinforced by a little ground asbestos. The suction on Solutions Used the crucible was carefully regulated. The precipitate was careThe solutions used, aside from the ordinary laboratory fully washed with ten portions of hot distilled water of about reagents, were as follows: 10 cc. each. The crucibles were then placed in an oven at (1) Carefully standardized solutions of calcium chloride 130” C. and dried for 30 minutes, then ignited and weighed. made by weighing out calcite, dissolving it in hydrochloric The filtrate was used for the determination of magnesium acid, and evaporating to dryness over a water bath, and then without evaporation. Its volume was about 100 to 125 cc. making up to the required volume with distilled water. The The magnesium was determined without double precipitation in the usual way; that is, by precipitating as the magnesium 1 Received August 16, 1930.
M
ANY naturally occur-
