Janiinry 15, 1932
INDUSTRIAL AND ENGINEERING CHEMISTRY
107
TABLEI. ACCURACY OF JELLY-STRENGTH MEASUREMENTS PRODUCT Apple jelly Cranberry sauce A Cranberry sauce B Cranberry aauce C Cranberry sauce D
INVESTI- SAMPLEMEANJELLYSTREKQTH Av. DEYIATION FROM MEAN QATOR No. Top Bottom TOP Bottom Grams % Grams %, 4.-2 4.0 5.8 34 86.1 68.5 3.6 CRF JAC 32 88.3 71.6 3.6 4.1 4.0 5.6 CRF 70 122.6 3.8 3 ..I JAC 70 123.9 4~. .5 3.7 CRF 60 174.1 5.5 3.4 JAC 60 172.8 6.6 3.8 CRF 10 233.1 176.4 7.7 3.3 9.7 5.5 JAC 10 234.0 155.9 8.4 3.6 3.9 2.5 CRF 10 251.4 161.8 6.9 2.7 9.6 5.9 JAC 10 256.0 162.5 8.2 3.2 5.1 3.1
It was found that the best method of testing jellies was to make three or more measurements on a sample. The mean of the three determinations is then taken as the jelly strength, In order to determine the accuracy of this method, samples of jelly and strained cranberry sauce were measured by different investigators and the probable error of the mean calculated in each case. ERROR OF MEANWHENTHREE TABLE11. PROBABLE MEASUREMENTS AREMADEON JELLY SAMPLE No. XHVESTIQATOR 1 CRF 1 JAC 1 CCR 2 CRF 2 JAC 3 CRF 3 JAC 4 CRF 4 JAC
JELLYSTRENQTHS 255, 238, 256 247 257 250 237’ 260’ 245 84: 88: 82 85, 75. 85 160 180 170 165’ 179’ 170 129: 120: 115 125, 124, 130
MEAN 249.7 251.3 247.3 84.7 51 7 170.0 172.3 121.3 126.3
PROBABLE ERROROF .MEAN &3.6 ztl.6 13.7 10.97 11.8 13.2 &?.8
&2.3 *l.O
Table I1 shows the probable error of the mean of three measurements on the same sample to vary from 0.97 to 3.7 where the jelly strengths varied from 82 to 260 grams. The
PROBABLE ERROROF MEAN TOP Bottom 0.55 0.78 0.36 0.46 0.61 0.65 1.74 1.96 2.12 1.99
0.62 0.77
APPARENT CONSISTENCY Medium Medium soft Firm
2.88 1.70 2.26 0.92
Veryfirm Too firm, rubbery
personal error was likewise slight. Because of the small probable error, the mean of three determinations on the same sample may be considered reasonably reliable. The new jelly-strength tester combines low cost, adaptability, ease and speed of operation, mobility, and simplicity. It may be purchased from John Chatillon & Sons, 85 Cliff St., New York, N. Y.
LITERATURE CITED Raker, G. L., IND.ENQ.CHFIM.. 18,89-93 (1926). Blake, M .A., N. J. Agr. Expt. Sta., Circ., 212 (1929). Bloom, 0. T., U. S. Patent 1,540 979 (June 9, 1926). Fellers, C. R., and Griffiths, F. P., IND.ENQ.C H ~ M 20, . , 85762 (1928). ( 5 ) Gavett, G. I., “A First Course in Statistical iMethod,” 1st ed., p. 181, McGraw-Hill, 1926. (6) Magness, J. R., and Taylor, G. F., U. S. Dept. Agr., Circ. 350 (1925). (7) Paine, H. S., Am. Food J.,17, No. 3, 11-13 (1922). (8) Richardson, W. D., Chem. Met. Eng., 28, 551 (1923). (9) Sucharipa, R., “Die Pektinstoffe,” p. 81, Serger and Hempel, Braunschwcig, 1925 (10) Tarr, L. W., Del. Agr. Expt. Sta., Bull. 142 (1926). (1) (2) (3) (4)
RECEIYED August 12, 1931.
Determination of Phosphates in Waters JASONE. FARBER AND GUY E. YOUNGBURG, University of Bufflo Medical School, Buffalo,N . Y
T““
need for a more rapid, reliable, and simplified method for the determination of phosphates in certain waters is evident. In the biological field limnologists are giving attention to the relation of phosphates to plankton content, and although primarily interested in this field, the authors have noted the recent method of Scarritt (8) for the determination of phosphates in boiler water in the presence of silicates. This method has been tried and found to have some outstanding disadvantages which can be practically entirely eliminated by the method prescribed. Scarritt’s main disadvantages are that the amount of color is not sufficient; that some of the reagents are unstable; and that the color due to silicates is not obviated, the final solution not being acid. Truog and Meyer (3) have improved Denigb’ method for phosphorus analysis but have not presented it in detail for water, After considerable experience with phosphorus determinations, the Deniges principle, as outlined by Kuttner and Cohen (1) and further elaborated by Youngburg and Youngburg ( d ) , has been chosen for biological work. It is the colorimetric method whereby phosphomolybdate is reduced by stannous chloride to give an intense blue color. The concentration of the acid and the molybdate is much different from that used by Truog and Meyer, and sodium
molybdate is preferred to the ammonium salt, although either may be used. This method, here presented for waters, involves no deteriorating reagents except the dilute stannous chloride solution; the preparation of it, however, is extremely simple. The color development is immediate, the fading is slow (see Figure l), and the amount of color obtained is greater than has been found for any other phosphorus method. The method of Truog and Meyer, however, gives only 12.5 per cent less color. The reason for this is the lower concentration of molybdate and of acid which they use. Al-
Time, Min.
Tims,
Hrs.
OF BLUECOLOR FIGURE1. RATEOF FADIXG
though 12.5 per cent less color is not a major objection in itself, the greater the acidity of the solution the less interference of silica, and in the determination of total phosphorus,
ANALYTICAL EDITION
108
as here indicated, the organic matter must be oxidized, and therefore the more sulfuric acid present the better. Truog and Meyer's concentration of acid is 0.4 N , the authors' 1.0 N . By reducing the concentration to that used by Truog and Meyer, the oxidation could not be done so simply. Since the solution is strongly acid (1 N sulfuric acid), silica interferes practically not a t all. This is shown in Table I. There can be present 3500 p. p. m. before it interferes as much as 1 p. p. m. of Pod. TABLEI. PRODUCTION OF BLUE COLORBY SODIUM SILICATE COLOR DEVELOPED BY SILICATE CORRESPONDING TO Po4 P. P. m. 0 0
SILICATE PRESENT P. p. m. 100 200 600 800 900
0 0
0.05
1000 1200 1500 2000 2500
0.10 0.15
3000 4000 5000 6000
0.78 1.25 2.00 3.50
0.25 0.35 0.50
The effect of variations of temperature on the color reaction is small. After heating the final colored solution to boiling and then cooling to 40" C., there is an increase of color of only 4 per cent. If the reagents are warmed to 50" C. before mixing, there is a loss of about 8 per cent of color. It is evident then, that nothing would be gained by heating, and that ordinary fluctuations in temperature can be disregarded.
INFLUENCE OF INTERFERING SUBSTANCES Experiments on the influence of ferrous and ferric iron are in agreement with those of Truog and Meyer (3). Ferrous iron shows no influence up to 50 times as much iron as Pod. Ferric iron interferes when present in more than 6 p, p. m. of iron, The interference is essentially an increase in color and the development of a greenish shade which superimposes on the blue to make reliable color comparison impossible. Copper, added as cupric sulfate, shows little effect up to 30 to 50 times as much as Pod. Beyond this, the solution begins to show a greenish tint which increases proportionately with the copper added. Carbonates and sulfates do not interfere in even as much as 400 times the amount of Po4 present, and aluminum up to 50 times the amount of Po4 does not interfere. Nitrates may be present in considerable quantity without appreciably depressing the formation of the blue color, as shown in Table 11. TABLE11. INTERFERENCE OF NITRATES Po4 P. a. m.
N
AS
NaNOs
%
15 25 50 75 100
100
200 300
93 85 83 82 71
500 1000
For this work a Spencer Duboscq colorimeter was used and the matching of the colors was done under very good conditions. TABLE111. RECOVERY OF PHOSPHATE FROM W x r m Po4
IN
ORIGINAL PO4 WATER ADDED P,p.m. P.9.m. 2 0.5 2 2 2 5 2 8 2 18 2 2 2 2
23 28 38 48
TOTAL EXPECTEDRECOVERED LOSS P,g.m. P.p.m. P.0.m. 2,475 0.025 2.5 4 3.94 0.060 6.82 7 0.180 0.160 10 9.84 19.80 0.200 20
25 30 40 50
98 96 96 95
50
0 0 0 0
55 04
25
% 99 0 98.5 97.4 98.4 99.0 9s 98 J 97 6 98 5
500 450 960 750
I
PROCEDURE. The method is prescribed for a total volume of 50 cc. Place 35 cc. of the water in a large test.tube (200 by 25 mm.) graduated a t 35 cc. and 50 cc. For less accurate work a graduated cylinder may be used. To two similar tubes add 1 cc. and 10 cc., respectively, of standard phosphate solution. Then make up the contents of the two tubes to the 35-cc. mark with water. To each of the three tubes add 10 cc. of molybdate-sulfuric acid solution, then 5 cc. of the dilute stannous chloride solution, and mix the contents well a t once. Match the colors with a colorimeter or with the naked eye, depending on the accuracy desired. Standards of intermediate concentrations may, of course, be prepared by taking amounts of phosphate solution between 1 cc. and 10 cc. CALCULATION STANDARD PHOSPHATE SOLN.MATCHED BY WATER
PO1 IN NATURAL
WATER
cc.
P. P. %a.
6 7 3 9 10
6 7 8 9 10
The interference of various other substances has been ascertained by Kuttner and Cohen, and Truog and Meyer. ACCURACY OF METHOD One part of PO4 can be detected in 100 million parts of water. The quantitative recovery is shown in Table 111.
24 29 39 49
RECOVERY
SPECIALREAGENTS REQUIRED.1. Ten N sulfuric acid. Add 450 cc. of concentrated c. P. sulfuric acid to 1100 cc. of water. Titrate the solution and dilute to make 10 N. If not too great accuracy is required, it is sufficient to mix 450 cc. of concentrated acid with 1156 cc. of water and to omit the titration. 2. Molybdate-sulfuric acid solution. Mix 500 cc. .of 7.5 per cent sodium molybdate, c. P. (phosphorus-free), with 500 cc. of 10 N sulfuric acid. Ammonium molybdate may be used. With the samples employed, a 12.8 per cent solution of either (NH4)nMo04or the hepta salt, (NH4)sRIoi02~4H20, was equivalent to 7.5 per cent of sodium molybdate, NalMoOq2H2O. 3. Stannous chloride solutions. For the stock solution, dissolve 10 grams of stannous chloride, SnCIZ,c. P., in 25 CC. of concentrated hydrochloric acid, c. P, (warming is permissible) Store in a brown, glass-stoppered bottle. For the dilute solution, dilute 1 cc. of the above stock solution to 200 cc. with water. This reagent is safe to use for a week or until a turbidity forms, after which a new dilution should be made. 4. Standard phosphate solution. Dissolve 2.506 grams of pure, dry, monopotassium phosphate (KH~POI)in water to . . make 100 cc. For natural water analysis, dilute 2 cc. of this stock solution to 1000 cc. (1 cc. when diluted to 35 cc. = 1 p. p. m. of Poi). For boiler water analysis, dilute 10 cc. to 1000 cc. (1 cc. when diluted to 35 cc. = 5 p. p. m. of Pod.
RECOVERY
P. 9. m.
400
Vol. 4, No I
Reading Of standard Reading of water
-
POI I N BOILER
WATER P. P . m. 5 10 15 20 25 30
36 40
45 50
x standard in p. p, m. = p. p. m. of PO4 in water
January 15, 1932
INDUSTRIAL AND ENGINEERING CHEMISTRY
109
If 5 cc. of phosphate solution (25 P. P. m. the organic matter oxidized by digesting with sulfuric acid were the Of the standard was 2o m m . ~ and perhydrol according to Youngburg and Youngburg (4). and that of the water 24 mm., then 20/34 X 25 = p. p. m. of POc in the water. LITERATURE CITED IkmPLE.
Of
The procedure is the Same for any of the water analyses, except that for natural waters the more dilute standard phosphate is used. For the determination of total phosphorus in plankton water, the sample must be evaporated to near dryness and
(1) Kuttner, T.,and Cohen, H. R., J. Bid. Chem., 75, 517 (1927). (2) Scarritt, E. W., IND. ENG.CHEM.,Anal. Ed., 3, 23 (1931). (3) Truog, E., and Meyer, A. H., Ibid., 1, 136 (19ag). (4) Youngburg, G . E., and Youngburg, M. V., J. Lab CLin. Med.. 16, 158 (19301. R ~ C ~ I V B DJ
~ ! YR, 1931.
Potentiometric Titration of Acidity in Oils ROBERTR. RALSTON,LTniversily of Michigan, Ann Arbor, Mich., C . H. FELLOWS AND K. S. WYATT,T h e , Detroit Edison Company, Detroit, Mich.
T
THE PLATINUM-CARBON electrode Blank determinations on the solvent were made f r e q u e n t l y I n e n d a t i o n (1) for the couple is shown to possess distinct advantages and required only 0.10 to 0.15 cc. determination of acidity over others previously suggested for the acidimetric of alkali. of oils, t h o u g h e x t e n s i v e l y titration Of Oils. used, is subject to several limiAPPARATUS An explanation of the second inflection point t a t i o n s . The use of phenolphthalein and other indicators in the difSerential plot of results on oil titrations S e v e r a l d i f f i c u l t i e s encountered in the use of the agar($) for detecting the end point is posfulaied as being due to the presence two of an oil titration is obviously agar bridge of Seltz and Mcdifferent classes of organic acids. Kinney (6) have been experiu n s a t i s f a c t o r y in the case of A system is described which permits acidity dark-colored oils b e c a u s e of e n c e d in preliminary work in determinations on oil samples as small as 0.5 the difficulty of observing the this laboratorv. chief a m o n g indicator change. Another disgram. which was a t e n d e n c y of t h i agar-agar to dehydrate slowly a d v a n t a g e is the f a c t that most oils are not completely soluble in the hot alcohol solu- and permit leakage. Isolation of the reference electrode in a tion. A two-phase system results, and the success of the separate vessel through use of a bridge was found to be imdetermination of the acidity of the oil is dependent upon practical because of the relatively high resistance of the the distribution of the acidic ingredient between the two solution. The silver-silver chloride electrode, when used phases. Seltz and McKinney (6) have shown the possi- as described by Seltz and Silverman (7), proved quite satisbility of the detection of the potentiometric end point for factory if freshly prepared, but after some days ceased to oil titration in isoamyl or n-butyl alcohol solutions. Later function as a standard. Such a deterioration was shown Seltz and Silverman ( 7 ) suggested improvements in the by the failure of the electrodes to give any potential break apparatus by substitution of a silver-silver chloride elec- a t the end point. In both of these systems (6, 7) a quinhytrode for the quinhydrone reference electrode and its agaragar bridge. BURET The work to be described was undertaken in connection with the research on the deterioration of high-tension cable now in progress in this laboratory. It represents an attempt to develop the potentiometric method as a practical laboratory means of determining acidity in small samples of oil obtained a t different points in the cable insulation. HE A. S. T. M. reconi-
-i-l
SOLVENTS For use in titration work and especially in potentiometric titrations, a solvent should be capable of effectively dissolving the sample, the titrating agent, and the “conducting salt,” and should possess a dielectric constant sufficiently high to permit the desired ionic reaction to take place rapidly, as well as to permit ionization of the conducting salt in order to increase the conductivity of the solution. Both isoamyl and n-butyl alcohol are excellent oil solvents, are satisfactory solvents for potassium hydroxide and lithium chloride (used as conducting salt), and have dielectric constants of about 15 to 18 (5). I n the work to be described the solution used as a solvent for the oil was isoamyl alcohol saturated with lithium chloride in order to give higher conductivity. The neutralizing solution consisted of a 0.025 to 0.050 N solution of potassium hydroxide in isoamyl alcohol standardized against Bureau of Standards benzoic acid.
CARBONROD
FIGURE 1. CELLFOR POTENTIOMETRIC TITRATIONS
drone electrode served as the indicator electrode. This quinhydrone electrode itself was found to be slow in coming to equilibrium potential in addition to its potential being notably unreliable in alkaline solutions.