V O L U M E 20, NO. 3, M A R C H 1948 cides containing labile chlorine. Controlled dehydrochlorination might possibly be used in developing methods for the analysis of these complex insecticides in the presence of one another. n70rk i s being continued along this line of investigation. The presence of sulfur, whether part of a formula or contamination, in a hexachlorocyclohexane dust causes some difficulty in the analysis for the gamma isomer. The sulfur is converted to sulfide during the dehydrochlorination reaction a t 0" C., and when silver nitrate is added t o the acidified solution a precipitate of silver sulfide is formed. This was overcome by adding a large enough excess of 0.1 X silver nitrate solution t o the 15- and 50-minute reaction flasks, following dehydrochlorination a t 0' C., and acidification with nitric acid, t o precipitate completely both sulfide and chloride. The flasks were then placed on a hot plate and most of the alcohol was boiled off, 2 ml. of concentrated nitric acid were added t o each flask, and the silver sulfide Itas decomposed by continued boiling. After cooling, the excess silver nitrate was titrated in the regular n-ay and the gamma isomer per cent calculated as described. Sample 3, Table 11, shows the results obtained on a dust containing sulfur. The average accuracyobtained on commercial samples should be in the neighborhood of 5y0,though greater accuracy has been obtaincd. The method tends to give high results for the gamma isomer, as indicated in Tables 1-1and TII.
245 Some typical results obtained in analysis of conimercial dust mixtures are presented in Table VIII. ACKNOWLEDGMENT
The author wishes to thank R. H. Kimball and L. E. Tufts of Hooker Electrochemical Company, Niagara Falls, N. Y., for the samples of purified alpha, beta, gamma, and delta isomers, infrared analyzed technical samples, and many helpful suggestions. Thanks are also extended t o Norman Wright and F. N. Alquist of the Dow Chemical Company, hlidland, Mich., for valuable information and a sample of purified epsilon isomer. LITERATURE CITED
(1) C.aldn-el1, J. R., and -Moyer, H. V., IND.EXG.CHEM.,ANAL. 1 b .ED., 7, 3 8 (1935). (2) Cristol, S. J., J.Am. Chem. Soc., 6 9 , 338 (1947). (3) Kauer, K. C., Du Tall, R. O., and Alquist, F. N., I n d . Eng. Chem., 3 9 , 1 3 3 5 (1947). (4) Kimball, R. H., and Tufts, L. E., IND. ENG.CHEM., ANAL.ED., 10, 5 3 0 (1938). (5) Kimball, R. H., and Tufts, L. E., private communication.
(6) Parr Instrument Co., Moline, Ill., Direction Booklet 116. ( 7 ) Slade, R. E., Chemistry & I n d u s t r y , 40, 314 (1945). (8) Tufts, L. E., and Kimball, R. H., Physical Chem. Lab., Hooker Electrochemical Co., Niagara Falls, N. Y., Rept. 4706 (March 17, 1947). (9)
Tan der Linder, L., Ber., 45, 236:(1912).
R E C E I V EAugust D 4, 1947.
ANTHOCYANIN PIGMENTS Colorimetric Determination in Strawberries and Strawberry Products ERNEST SONDHEIMER AND Z. I. KERTESZ S e w Y b r k S t a t e Agricultural Experiment S t a t i o n , Cornell University, Geneca, S. 1.. The amount of red anthocyanin pigment in strawberry products can be quantitatively determined by measuring the light absorption in extracts made therefrom at 500 mp and at pH 3.40 and 2.00. The measurement consists of subtracting the optical density at 500 mp of an anthocyanin solution at pH 3.4 from its optical density at pH 2.0, at known concentration. The increase in color intensity under
I
i\; CONSECTION with work in this laboratory on the changes
of strawberry preserves during storage, a quantitative method for the determination of the red anthocyanin pigments of strawberries was desired. During the storage of strawberry products at temperatures above 0' C., two main types of nonenzymatic color reactions occur: the red anthocyanin color is lost and secondary brown pigments develop. None of the existing methods has been found satisfactory for the quantitative determination of the actual loss of red pigment. Usually data are obtained with the Lovibond-type colorimeter ( 1 , 9 ) , which give a n indication of the degree of color deterioration in the juice Brhen the ratio of red t o yellow light absorption is calculated. Yet these measurements are not strictly quantitative or even directly related t o t h e loss of anthocyanin. Attempts to separate the anthocvanin quantitatively from other colored materials present in the juice by chromatographic adsorption and various solvents 1% ere unsuccessful. The pH of the medium strongly affects the light absorption of anthocyanin (3, 8). For instance, when the p H of fresh strawberry juice is lowered from 3.5 to 2.0 the absorption at 500 mp, expressed as optical density, is more than doubled. Figure 1 shows the effect of decreasing pH on the absorption a t 500 mp
such conditions is proportional to the concentration of anthocyanin in the solution. A solution of Congo red is suggested as a standard of color intensity. Examples illustrate applicability of the method, which may be used for objective comparison of the red color in different samples of strawberries and in strawberry products of differentorigin, and to follow color deterioration in the product after manufacture. of pure anthocyanin pigment a t constant concentration. (The anthocyanin used in these experiments was isolated in crystalline form from Dresden strawberries. Work is now in progress on the characterization of this pigment .) Further experimentation showed that this absorption change can be used for the quantitative determination of anthocyanin in strawberries and strawberry products. A4nthocyaninsare amphoteric substances which form oxonium salts with mineral and organic acids. It is highly probable that in neutral or nearly neutral solutions the pigment exists in the free state ( 5 ) and upon acidification the equilibrium between the color base and the oxonium salt is shifted, resulting in a molecule of higher resonance and therefore exhibiting greater light absorpt ion. METHOD OF MEASUREMENT IN STRAWBERRY PRODUCTS
TVhen the p H of a solution containing anthocyanins is lowered, the light absorption in the visible range increases. Under the experimental conditions used, this increase is proportional t o the concentration of anthocyanin. The readings are taken a t 500 mp, since this is the highest peak of absorption in the visible range and therefore maximum sensitivity is obtained a t this wavc length. Furthermore, the absorption peaks of ex-
ANALYTICAL CHEMISTRY
246
tracts of more than 300 varieties and seedlings of stram-berries (6) ivere all found to be a t this wave length. Since Schou ( 7 ) has shoivn that the absorption peaks of the natural anthocyanidins are a t different wave lengths, it seems safe to assume that the anthocyanins in different species and varieties of strawberries are similar. At the concentrations and pH range used all solutions follow Beer’s law. Since this s h o w that the optical density
0.40 F
t
2
W
a
2 0.20
0
c
6 I
I
2
3 PH
4
Figure 1. Effect of pH on Absorption of Strawberry Anthocyanin Chloride a t Constant Concentration (1.4 Mg. %) and a s Measured a t 500 mw
.
is proportional to the concentration of anthocyanin a t a given pH, one can conclude that the absorption increase obtained on loi+ering the p H is also proportional t o the concentration of anthocyanin. Working with purified anthocyanin solutions, the authors found this relationship to hold true in all concentrations tested (Figure 2). A pH change between 3.4 and 2.0, with purified anthocyanin, affects only the intensity of absorption and, as shown in Figure 3, the position of the peakwith respect t o wave length is unchanged. The measurement consists of subtracting the optical density at 500 mp of an anthocyanin solution a t pH 3.4 from its optical density a t pH 2.0, a t knom-n concentration. To facilitate the use of various colorimeters and spectrophotometers this reading is expressed as milligram per cent of Congo red which may be used conveniently as a standard. I n Table I the optical density data are also given as milligram per cent of anthocyanin as well as of Congo red. However, it is the authors’ opinion that the general applications of this method do not warrant the use of pure anthocyanin as a standard nor do they feel that much can be gained by expressing the results in milligram per cent of anthocyanin. PROCEDURE
This procedure is applicable to fresh and stored strawberries, strawberry juice, and strawberry preserves. Preparation of Material. Eighty grams of the sample are homogenized in a Waring Blendor for 0.5 minute with 100 ml. of Sorensen’s citrate-hydrochloric acid buffer of pH 3.40. This mixture is then filtered through sharkskin filter paper, or centrifuged and filtered a second time through Whatman No. 2 filter paper. This is solution A, dilution factor 2.25. This step is unnecessary when preparing strawberry juice for analysis. At times it is difficult to obtain optically clear solutions from fresh strawberries. Adding a pinch (about 25 mg.) of Pectinol 1 O M (obtainable from the Rohm and Haas Co., Philadelphia, Pa.) a t the time of mixing in the Blendor followed by filtration through the sharkskin paper will give clear solutions. Freezing the berries solid for several hours will also eliminate the difficulties in obtaining solutions suitable for colorimetric measurements.
A known volume of solution -1is dilut,ed with the above buffer to an extent to give an optical density within the optimum range of the instrument used. If the pH of this solution is not 3.40 * 0.05, another sample should be prepared; in the dilution small quantities of sodium citrate or citric acid should be used to give the desired final pH value. -1 Beckman pH meter with glass electrode has been used throughout these investigations. The dilution factor is noted for this solution B. To a knoivn volume of solution A sufficient dilute hydrochloric acid of such concentration as to produce a final pH of 2.00 * 0.05 is added. The dilution should again be adjusted to give a reading at 500 mp in the most sensitive range. This is solution C. I t should be allowed to stand for an hour before the readings are taken in order to allow full color development. Preparation of Color Standard. For a standard of color intensity Congo red (Congo Red Special, obtained from the Kational Aniline Division, Allied Chemical and Dye Corporation) is dissolved in 0.01 S sodium carbonate to give a 20 mg. % solution. The calibration curve is obtained by diluting this solution with 0.01 S sodium carbonate and plotting the optical density against concentration. Measurement and Calculations. A11 measurements are made a t 500 mp. The per cent transmission is determined for solutions B and C, and expressed as optical density, log,dO,’I. I n solutions where the dilution factors of B and C are not the same, the optical densit,y readings should be equalized with respect to dilution. On subtracting the equalized optical density reading of solution B from the optical density reading of solut,ion C the net reading is obtained. This net reading is converted to milligram per cent of Congo red equivalents, using the calibration curve; on multiplying b+ the total dilution factor the milligram per cent of Congo red equivalent of the original sample is obtained. The color intensity of a solution containing 0.825 mg. of Congo Red Special in 0.01 .V sodium carbonate equals the increase in absorption which occurs when the pH of a purified strawberry anthocyanin chloride solution containing 1.0 mg. % pigment is changed from pH 3.40 to 2.00. Thus the Congo red values multiplied by 1.2 will give the anthocyanin equivalent of the observed absorption. The data in this paper were obtained with a Beckman D-U quartz spectrophotometer using 1-em. Corex cells. With such instruments as the Lumetron and the Coleman spectrophotometer results roughly within 10% of thr Beckman spectropho-
0.6 0
> b
gvr 040 -1
4
c
0
0 20
29 MG.%
1.0
Figure 2. Effects of Concentration and pH on Absorption of Strawberry Anthocyanin Chloride at 500 m p 1. pH 2.00.
2.
pH 2.50.
3.
pH 3.42.
4.
pH 3.62
V O L U M E 20, NO. 3, M A R C H 1 9 4 8
241
I 2-
c
hum the same (homogenized) strawberry preserve and diluting them all to a total volume of 360 cc., the Congo red color equivalents were 2.6, 2.8, and 2.7 mg. %, respectively, indicating the reliability of the method of extraction. It is clear from these data that the change in pH has no appreciable effect on the absorption at 500 mg of the brown color developed during storage in strawberry preserves. Furthermore, these results indicate the absence of such interfering factors as co-pigments or Pffects attributable to other components of strawberry pre+rves.
0.40
si z
U
a J
< 0 0.20 c
F4CTORS AFFECTING METHOD O F MEASUREMEhT
a 0
600 Figure 3. pH 3.40 (1.Y8
500
400 MY
lbsorption Curves of Strawberry Anthocyanin Chloride mg. %) lower curve, pH 2.00 (0.84 m g . 70) upper curve
tometer readings might be obtained. (Further work on thib compariqon of various instruments is now in progress. j ACCURACY O F DETERMINATION
Although the solutions should be optically clear, filter aids iuch as Celite Super-Cel must not be used, as these materials adsorb some of the pigment and produce changes in the pH of the wlution. The method cannot be used in any products which contain alcohol, such as strawberry wines, since different solvents produce great changes in the absorption curves of the anthocyanin-for example, there is a shift from 500 to 520 mp when strawberry anthocyanin is dissolved in 95yo ethanol instead of water. Care must be taken to exclude heavy metals from contaminating the solutions to be analyzed. Iron in concentrations as low as 2 p.p.m. produces "discolorations" in strawberries a t room temperature ( 2 ) . The authors are not prepared as yet to state the usefulness of the method for products containing other fruits in addition to strawberries. According to Fear and Kerenstein (4j, measurements of anthocyanin solutions must be standardized with respect to pH, temperature, and time of contact with reagents. A temperature range of 18' to 34" C. was found to have no effect on the light absorption of the solutions used in these experiments. When the solutions are stored a t 0" C., they are stable for a t
To study the accuracy of the method recovery experiments ivere conducted. Anthocyanin chloride diwolved in Sorensen's citrate-hydrochloric acid buffer was added to strawberry preserve extracts prepared as solution A . Heatrid preserves Table I. Recovery o f , Anthocyanin Chloride Added to Strawberry Preserves (14 Lveeks a t 38" C.) were Congo Congo Ket Reading Red Red chosrn for these expL'rimentr ' Equivalent (Added Equivalent sincr in these samples the anOptical Density X Dilution Anth,oof S e t At p H Differ- Factor, cyanin Reading, Recovery, At p H thocyanin content is very low Sample 3.40 2.00 ence Llg. % Chloride) Ma. % 70 and the brown pigment content Anthocyanin chloride, considerably higher than in 4determinations 0 045 + 0 , 0 0 1 0 175 f 0 . 0 0 0 4 0 . 1 3 0 1 24 preserves stored a t 0" C. They Strawberry preserves stored 14 weeks a t are therefore well suited for 38O C . , 7 determinations 0 218 1 0 . 0 0 7 0 240 & 0 . 0 0 6 0.022 0 21 studying the effects of brown pigments on the anthocyanin strawberry preserves stored 14 weeks a t determination. The anthocy38O C . , plus anthocyan+ chloride, 7 deanin was diluted t o the same terminations 0.277 & 0 , 0 0 6 0 , 4 2 1 1 0 . 0 0 9 0.145 l 38 0.126 1.18 95 i. 4 . 4 volume as that added to thr preserves and the determination made as outlined in the Table 11. Anthocyanin Concentration in Strawberries and Strawberry Products procedure, dilution factor 4.5. Congo Red Loss The results are given in Equivalent An+ in Table I. Optical Density DiluOriginal cyanin AnthoA t p H At p H Differtion Observed, sample, Chloride, cyanln, For the seven determinations Sample 3.40 2.00 ence Factor mg. % mg. 70 Alg. % ' "0 an average deviation of 7.1'3 Frozen strawberries from a theoretical recovery Dresden 0.234 0,772 0,538 20.25 1.140 23.1 27.9 Clermont 0,170 0.490 0,320 30.00 0,676 20.3 24.6 was found. The experimental 0 , 1 4 9 0 , 4 8 2 0 , 3 3 3 3 0 . 0 0 0 , 7 0 2 2 1 . 1 25.5 Pathfinder .. Brightmore 0,200 0.586 0.386 30.00 0.815 24.5 29.6 errors caused by pipetting, reproducibility of colorimetric Preserves 0.350 0,894 0,544 9.00 1,150 10.4 12.5 1 readings, etc., vary from 2.2 to 4.8 5,s 4.50 1.060 0.502 0,448 0.950 2 2.8 3.4 1.135 2.50 0..537 0.538 1.075 3 3.1yGfor each series of determinations. Since operationc Preserve 3 12 3.0 2.5 0.545 4.50 0.530 0,268 0.272 Stored 6 weeks a t 15' C . entailing these errors are re50 1.7 1.4 0.318 4.50 0.380 0.147 0.233 Stored 21 weeks a t 15' C . 62 1.3 0.240 1.1 4.50 peated several times during 0 , 2 3 6 0 , 3 4 8 0.112 Stored 3 1 weeks a t 15' C. 15 2.9 2.4 0.950 2.50 0.450 0.940 0,490 Stored 1 week a t 38' C. the determination, an average 50 1.4 1.7 0.570 2.50 0.270 0,450 0.720 Stored 3 weeks a t 38' C. 65 1 . 2 0 . 4 5 0 1 . 0 2 . 5 0 0 . 2 1 2 0 . 6 2 0 0 . 4 0 8 Stored 4 weeks a t 3 8 O C. deviation of 7.1% is not exces85 0.5 0.4 0.170 2.50 0.080 0.358 0.438 Stored 8 weeks a t 38' C. 94 0.2 0,049 0.2 4.50 sively high. On taking sam0,240 0.023 0.217 Stored 14 %eeksa t 38' C. ples of 20, 40, and 80 grams
ANALYTICAL CHEMISTRY
*
least 24 hours. In the preserves tested the pigment concentration in the fruit was the same as in the surrounding jelly. Some examples of the application of the method are shown i n Table 11. The total pigment, contents of the four varieties of strawberries (measured in frozen samples stored a t -18’ C. for 6 months) were rather similar. Strawberry preserve sample 1 was unusually high in pigment, content, the color intensity bein’g about what one would expect from the results on frozen fruit, after allowing for the effects of dilution by other components of the preserve and for some color loss during manufacture. Preserve 3 was obtained from the same manufacturer as S o . 1, and showed only about 27y0 as much red anthocyanin as KO. 1. The color difference was obvious upon visual inspection of the preserves. Sample 2 was purchased in March on the open market. Preserve 3 when stored at 15’ and 38” c.shon-ed rapid loss of pigment, especially at the higher temperature.
LITERATURE CITED
Beattie, H. G . , Wheeler, K. A , , and Pederson, C. S.,Food Research, 8 . 395 (1943).
Bryant, J. XI., and Morris, T. N., Canning Trade J., 3, 252 (1933).
Buxton, B. H., and Darbishire, F. l‘., J . Genetics, 21, 71 (1921). Fear, S.M., and Nierenstein, M.. B w c h e m . J.,22, 615 (1928). Gilman, Henry, “Organic Chemistry,” p. 1316, New York, John Wiley & Sons, 1943. Robinson, W. B., Division of Food Science and Technology, N. T. State Agricultural Experiment Station, unpublished work.
Schou, 9. H., Hela. C h i m . Acta, 10, 907 (1927). Smith, E. P.. Protoplasma, 18, 112 (1933). Thompson, H. H.. Cecil, S. R.. and Woodroff, J. W., Food Induotries, 18, 1341, 1510 (1946). RECEIVED May 23, 1947. Journal Paper 709, New York State Agricultural Experiment Station, Geneva. Investigations in p a r t supported by a grant from the Sational Preservers Association.
Determination of Safrole in the Oil of Ocotea cymbarum A Cryoscopic Method ANTHONY J. SHUKIS AND HERMAN WACHS Dodge & Olcott, Inc., Bayonne, N . J . A cryoscopic method employing congealing temperatures has been developed for the determination of safrole in the oil of Ocotea cymbarum of commerce. A graph is included, by means of which the safrole content of the oil as determined by the mercuric chloride method, is correlated with the congealing temperature.
S
AFROLE is of considerable technical importance as a flavoring material and as a raw material for the preparation of piperonal, and recently it has come into use as a starting material for the synthesis of insecticidally active materials (7). Prior to the recent war, Japanese camphor oil was the only available material from which safrole could .be obtained in commercial quantities. It was a fortunate coincidence that a new source was made available in Brazil by the steam distillation of the oil of Ocotea c y m b a r u m ( 2 ) . Since safrole is the only valuable constituent of the oil of Ocotea c y m b a r u m , a simple and quick method for its determination was desired. A cryoscopic method appeared to be most suitable. However, for this purpose it was necessary to obtain correlation of physical constants, congealing temperature, and concentration. To determine the concentration of safrole in samples of the oil of Ocotea c y m b a r u m the addition compound of safrole and mercuric chloride (1, 6) yielding hydroxychloromercuri dihydrosafrole wm used. The methods described in the literature (3, 4) were modified as described below, to minimize the errors due to transfer and solubility of the hydroxychloromercuri dihydrosafrole in water, and thereby give more readily reproducible results. The method outlined by Huzita and Nakahara (3) was modified as follows: The reaction is carried out in one vessel to avoid a troublesome transfer. The reaction is allowed to proceed in a homogeneous solution for 30 minutes, using mercuric acetate solution. Under the conditions of homogeneity the reaction readily proceeds to completion. Conversion t o the chloride is then effected. Allowance is made for the solubility of hydroxychloromercuri dihydrosafrole in water a t 0 ” C. by determining the solubility of the material at that temperature and applying a correction factor. By these means the error in general is reduced
from i l . 5 to *0.770 and readily reproducible results are obtained. Safrole “drainings” were prepared by freezing a sample of the oil of Ocotea c y m b a r u m . The congealed fraction was used as a source of safrole, which was subjected to further purification. The noncongealed fraction was taken as safrole drainings. Samples of safrole drainings, oil of Ocotea c y m b a r u m , and safrole, the latter purified by freezing, were doubly distilled and center sections were taken as standards for analysis. Congealing points on each were taken, and repeated if necessary until checks within 0.1’ C. were obtained. Each sample was then analyzed by the mercuric chloride method. Summarized in Table I are these data, with additional data taken to round out the information,
Table I. Sample Safrole drainings Oil of Ocotea
cymbarum
Safrole
Properties of Analytical Standards
Congealing Point, O C.
Safrole Content Weight 7c. Mercury Analysis
Specific Gravity, 25O C.
2.4
1.5299
1.0541
8’8 11.0
1.5352
1.0843
1.5382
1.0987
99,518::
$
A plot of weight per cent of safrole versus the congealing points was prepared; the points in Table I fell on a straight line. To ensure conformity and check the linearity of the plot, samples differing in safrole content by approximately 1 weight yo,in the range 85 to 95%, were prepared from mixtures of the safrole drainings and safrole mentioned above and their congealing points were determined. When plotted, these points fell on the straight line plotted using the standards. All the values fell within
