ANALYTICAL CHEMISTRY
1172
tions of this solution to such :t volume that 0.004 to 0.08 mg. pc'r ml. of the 2-thenyl salicylate will be present. If the 2-thenyl salicylate coneentrat,ion is complet,ely unknown, prepare several dilutions until the right range is obtained. In each case, place 10 ml. of the diluted solution in a calibrated Klett-Summerson h h e , add 0.20 ml. of the ferric alum solution, mix well, and allow to stand for a t least, 1 minute (if the solution is cloudy, filter int,o the Klett tube). (The color developed after about 1 minute undrr these conditions has been found t.o be stable for a t least 5 days i n the absence of strong sunlight.) Take the reading in a KlettSummerson photoelectric colorimeter, using green filter 54. Prepare a curve relating colorimeter reading with 2-t,heny1salicylate concentration using known solut,ions treated as described above, and make the quantitative determination of the unknown by spplying t,he reading obtained.
Table I.
?-Thew1 Salicylate, % Bromination Colorimetric ferric procedure alum method
Sample NO.
3.49 5.49 5.92 6.35 4.25 2.61 1.52 0.9.; 0 38 0.26 0.12 0 07 0.06 toncloth, fixatives such as chlorinated paraffin, or plasticizers such as dioctyl phthalate. A slight blank correction is necessary nit,h sizod-cloth samples. The colorimetric method is also rapid and accurate for the estimation of amounts down to 0.05 nig. of 2-thcnyl salicylate. S o interferences are experienced from undyed or olive drab pure-finish or sized cloth, chlorinated paraffin, and dioctyl phthalate. The saponification method W ~ uscd only a8 a check method and is accurate for amounts down to about 30 mg. of the miticidt.. Sacconol NR interferes slightly: a special correction must he made in the presence of chlorinated paraffin; and the method is not applicable when dioctyl phthalatc is preeent.
Determination of 2-Thenyl Salicylate in Cloth Impregnated in a Pilot Plant
5.55 5.53 5.81 6.35 4.15
2.i4 1.46 0.93 0.37 0.27 0.12 0.0s 0.07 0.03
LITERATURE CITED
(1) Cross, H. F., J . Econ. Entoinol., 41, 731-4 (1948). (2) Cross, H. F., and Snyder, F. M., Soap Sanit. Chemicals, 25, S o . 2 ,
P
1 3 5 4 9 (1949). (3) Francis, A. W., and Hill, A. J.. J . A m . Chem. Soc., 46, 2498 (1924). (4) King, W. V.,Am. J . Trop. M e d . , 28, 487-97 (1948). (5) Mehlig, J. P., IND. ENG.CHEM.,ANAL.ED.,10, 136 (1938). (6) Snell, F. D., and Snell, C. T., "Colorimetric Methods of Analysis," 1st ed., Vol. 11,pp. 163, 369, 372, New York. D. Van Nostrand Co., 1937. (7) Sprung, M. hf.. ISD. ESG. CHEM..Asar,. ED.,13,35 (1941).
RECEIVED September
1. 19.50.
Determination of Glucose in Presence of Maltose Rapid and Convenient Semimicromethod LOUISE LANG PHILLIPS AND M . L. CALDWELL Columbia University, New York,N . Y .
N T H E investigation of amylase action, it is often important
1 to determine small concentrations of glucose in the presence
of maltose and other reducing products. The method developed to meet this need is essentially a modification of the method of Zerban and Sattler ( 7 ) , but adjusted for the determination of much smaller concentrations of glucose. Instead of 20 to 80 mg. of glucose in volumes of 20 ml., it permits the accurate determination of 2 to 6 mg. of glucose in 2 ml. of sugar solutions or of starch hydrolyzates in the presence of maltose and other products of amylase action. The present report introduces no new theoretical principles, but unites in one convenient rapid procedure conditions that have been found essential to accurate and precise results. The method of Zerban and Sattler ( 7 ) and that reported here employ the Soxhlet modification of Fehling's copper solution, and an excess of sodium acetate (6). The cuprous oside formed in the reduction is determined by the iodometric method of Shaffer and Hartman ( 5 ) . The method depends primarily upon the equations: IC103
+ SKI + 3HeSOi ---+3KtSO1 + 31, + 3H2O
The difference between the blank determination of the iodine released (by thiosulfate titration) and the unknown, in which
wme of the potassium iodate ha8 been used to osidiae the cuprouq oxide formed, is a measure of the glucose present in the misturc. REAGENTS
Copper Sulfate (Soxhlet modification of Fehling's solution). Dissolve 69.28 grams of copper sulfate pentahydrate in water and dilute to 1 liter. Use large clear crystals showing little or no efflorescence. Sodium Acetate Trihydrate, reagent grade. Dissolve 500 grams in about 800 ml. of hot distilled water, cool, and dilute to 1 liter. This solution should give pH values of 8.7 t o 8.8. If the solution gives lower pH values it should be discarded. Potassium Iodide-Iodate. Dissolve 5.400 grams of reagent grade potassium iodate and 60.0 grams of reagent grade potassium iodide in about 500 ml. of distilled water, add 0.25 gram of sodium hydroside, and dilute to exactly 1 liter. Sulfuric Acid, approximately 4 N . Add 114 ml. of eoncentrated sulfuric acid to 500 ml. of distilled water, cool, and dilute t o 1 liter. Potassium Oxalate. Dissolve 330 grams of potassium oxalate crystals in 1 liter of hot distilled water (saturated solution). Cool. Sodium Thiosulfate, 0.01 .V. Dilute 0.1 N solution prepared by dissolving 24.8 grams of reagent grade sodium thiosulfate pentahydrate and 3.0 grams of sodium tetraborate in 1 liter of water. The dilute solution should be freshly prepared and standardized. Dextrose, Bureau of Standards, 1% solution for preparing ralibration charts. Maltose Hydrate, 2% solutions for preparing calibration curves.
1173
V O L U M - E 23, NO. 8, A U G U S T 1 9 5 1
I t is desirable for each operator to prepare calibration curves Table I. Glurwse" , M g . / 2 111. 1
2 3 4 5
6
Data for Calibration Curves Maltose, M g . per 2 M I . None 5 19 Titer of 0 01 S Sodium Thiosulfate, A. A4altoseb 1 7.17 3.37 5.19 10.55 8.93 7.38 13.28 11.24 12.19 15.45 13.80 14.56 16.97 15.83 16.40 18.57 17.90 17.43 B.
1
2
3 4 5
6
7.65 11.56 14.28 16.06 17.53
Maltosec 2 5.38 9.26 12.22 14.66 16.49 17.90
7.21 10.61 13.23 15.47 17.06 18.60
for use with the reagents and equipment a t hand to compensste
for differences in technique and in the quality of the reagentq. 20
MI
CORRECTION FOR REDUCING 4CTIOY OF .\l.iLI'oSE 10.37 12.70 14.59 16.59 17.91 19.55 9.99 12.36 14.26 16.38 17.66 19.11
C. hfaltosed 3 10.30 11.65 7.65 13.18 14.17 11.56 15.04 15.87 14.28 5 16.06 ... 6 17.53 ... l8:24 18:il 11 Glucosr Vational Bureau of Standards. b Maltose: *[cz]y= 129.9. Reducing value 93% determined by iodometrir iriethod ( I ) . c Maltose [m]%6 = 130.1. Reducing value 95% determined by iodometric method ( I ) . d Maltose, highly purified, prepared from starch by action of P-anwla% and recrystallized; [a]%6 = 131.25. Reducing value 98.5% determined by iodonietric method ( I ) . 2 3 4
PROCEDURE
Portions of 2 ml. of the sugar solution or the hydrolyzates to be examined (or smaller volumes diluted to a total volume of 2 in].) are added to 1 ml. of the copper sulfate solution and 2 ml. of the sodium acetate solution in a test tube. Tubes thus prepared and similar blank tubas containing 2 ml. of water instead of the sugar solution are covered with glass bulbs and placed in 1)riskIy boiling water for 20 minutes. It is important that the tribes be placed in a rack, so that they do not touch the bottom or sides of the container and that the level of the boiling water be above that of the solution in the test tubes. The test tubes should also be of the same size. At the end of 20 minutes the tubes are cooled in running water for 4 minutes. The following solutions are then added rapidly: :i ml. of the potassium iodide-iodate solution, 1 ml. of the dilute sulfuric acid solution, and 2 ml. of the potassium oxalate solution. All reagents are added rapidly to one tube a t a time and mixed before going on to the next tube. The solutions are then mixed thoroughly and allowed to stand a t room temperature for 30 minutes or longer with occasional additional mixing. All traces of red cuprous oxide or white cuprous iodide precipitate must be dissolved. The greenish blue cupric oxalate precipitate which sometimes forms does not, interfere. After standing a t room temperature for 30 minutes or longer, the contents of the tubes are transferred to 125-ml. Erlenmeyer Hasks and the tubes are washed six times with 2.5-nil. portions of distilled water. The resulting Eolutions are then titrated with 0.01 A' sodium t,hiosulfate. CALIBRATION CURVES
With solutions of pure glucose, the difference between the titer of sodium thiosulfat,e for the blank and for the glucose solution is due to t,he reducing actiori of glucose under the conditions employed. Therefore, a calibration curve is set up for the values obtained with solutions containing known weights of pure glucose and the reagents and equipment employed. For this curve, the milliliters of 0.01 N thiosulfate are plotted against milligrams of glucose per 2 ml. of glucose solution. Solutions containing 0.5 to 8 mg. of glucose per 2 ml. are used. Maltose causes a small but often significant' reduction of the cupric ion in copper sulfate, even in the presence of acetatr. Therefore, it is necessary also to set up glucose-maltose ca1ibr:~tion curves for solutions containing know-n small weights of maltase in addition to known weights of glucose. When plott.ed as above, a separate glucose-maltose calibration curve is obtained for each concentration of maltose with glucose. The calibratioii curves used here were based upon solutions containing 0.5 to 8 ing. of glucose and 5 , 10, or 20 mg. of nialtose hydrate.
\Vith unknown sug:Lr solutions the differonc~c~ tiet ivc~cnt,lic titer of sodium thiosulfate for the Iilank :ind that given by th(, sugar solution may be influenced by the pi'esoricc of nialtosca. Thcrefort,, this difference represents the apparent rather than the tiue glucose. The value for t,his :ipparent glucose is read from the glucose calibration curve and must 1)c correcttd for any maltose i i i the sugar solut~iori. I n order to make this corrcctioii, the tutal reducing valuc: of the sugar solution is dotermined by an iodoinetric met,hod ( 1 ) . This value is stoichiometric and includes both glucose and nialtose. I t is calculated to its glucose equiva1ent.s. The apparent weight of maltose is then deterniincd in turn by taking the difference between t,he total reducing value as glucose equivnlents and t,he apparent weight, of glucose taken from the glucose c-alibration curve. This difference, expressed as glucose equivLtlents, must, be multiplied by t w o to convert it, back to m:ilt,cise on which the glucostxnaltose calibration curves are bascd. The true glucmw value of the solution is then read from t.lie glucose-maltose calibration curve representing the weight of admixed maltose nearest to the apparent value for malt.o~eobtained above or, when necessary, by iiiterpolation t)etween tw-o glucowmaltose curves. If the glucose value thus obtained differs significantly from tbc apparent value obtained above, it may be necossarp to recalculate the apparent maltose value on the basis of the new gluccw value and then to redetermine the t,rue glucore value from t h r recalculat,ed maltose value. RESULTS
Scope and Accuracy of Method. The average data given i n Table I were used to prepare calibration curves for glucose alone and in the presence of known weights of maltose. These data were plotted on large size graph paper, from which thc glucosc could be read easily to 0.01 mg. The three samples of maltose employed in t.his work did not differ markedly from each other in their optical activities. However, the reducing values obtained by the iodometric method ( 1 ) differed considerably. Table I shows that the t h e e samples of maltose gave different reducing values with copper acetate in the presence of the same ooncentrations of the same glucose under similar conditions. Two samples of maltose, 1 and 2, were found later by amylase act,ion ( 3 ) to be contaminated with small coiicentrations of both glucose and dextrins in proportions to compensate for each other in measurements of optical rotation. Therefore, subsequent work was carried out with glucose-maltow calibration curves based on measurements made with the third highly purified maltose that had been obtained by the action of beta-amylase on starch. Comparison of the data for glucose alone given in Table I , A and B, shows different reducing values for the glucose. Thew> data were obtained by the same operator with the same pure glucose and the same equipment but with different samples of sodium acetate. Both samples of sodium acetate gave values of pH 8.7 to 8.8 for 50% solutions. These data emphasize t,he importance of the recommendat,ion that each operator prepare his own calibration curves. The data given in Table I1 represent a study of the influence of differences in the proportions of glucose to maltose in the sugar solutions upon the accuracy of the method. These solutions contained different concentrations of glucose in the presence of 4 and 16 mg. of maltose hydrate per 2 ml. These two Icvc~lso f maltose were used to determine whether therc. is a ' 1'( dt,er error when large quantit,ies of maltose are present than when the pro'5
ANALYTICAL CHEMISTRY
1174
portion of glucose to maltose is more nearly equal. With 4 mg. of maltose hydrate, the average error was 0.08 mg. of glucose per 2 ml. or 1.6%. With 16 mg. of maltose hydrate, the average error was 0.11 mg. of glucose per 2 ml. or 2.7%. Comparison of Chemical Method with Selective Fermentation with Yeast. The data summarized in Table I11 compare the results obtained by the method described here with those obtained by selective fermentation with yeast No. 2019, as recommended by Schultz et al. ( 4 ) . With solutions of known concentrations of glucose, the method described gave an average error of 0.19 mg. of glucose per 2 ml. in solutions containing from 0 to 16 mg. of glucose per 2 ml. The maximum error was 0.42 mg. per 2 ml. The yeast fermentation method, using Fleischmann's yeast S o . 2G19, gave an average error of 0.23 mg. of glucose per 2 ml. on the same fourteen sugar solutions. With the fermentation method, the maximum error was 0.58 mg. per 2 ml. The two methods were also compared with hydrolysates obtained a t 40' from 1% Lintner's soluble potato starch or 1% maltose by the action of a glucose-forming amylase (3) produced by the mold, Rhzzopus delemar (8). This comparison is summarized in Table IV, which shows an average difference of 5.0% when the substrate was starch, and 10.3% when maltose was the substrate. The data given in Tables I1 and I11 for solutions of known concentrations of glucose or of glucose and maltose show that glucose can be determined by the method described here with an error of 0.1 to 0.2 mg. of glucose per 2 ml. in glucose solutions that contain 2 to 6 mg. of glucose per 2 ml. and in the presence of 20 mg. or less of maltose. DISCUSSION
Certain precautions in the use of the copper reduction method should be observed.
Influence of Proportion of RIaltose to Glucose upon Accuracy of Method
Tahle 11.
RIaltose Hydrate Added, AlgLpec-AIl. 4
Glucose Glucose RecovDifferError, Added, ered. ence. Mg./2 111. 11g./2 MI. M&/2 111. 7c
4 5 5 6 6 6
1.97 1.97 2.91 2.99 4.08 3.94
0.03 0.03 0.09
0.08 0.06
1.5 1.5 3.0 0.3 2.0 1.5
4:95 4.88 6.21
0.05 0.12 0.21
1.0 3.0 3.5
0 08
1 0
..
A \,
0.01
2.09 2.12 2.96 2.93 3.86 3.90 3.84 4.75 4.79 6.02 6.05 6.10
...
... ...
~
_____
Glllcose RecovDifferered, ence, Error. AIg./2 M I . Mg./2 111. %
...
...
16
...
0.09 0.12 0.12 0.07 0.14 0.10 0.16 0.25 0.21 0.02 0.05 0.10 0 11
2 5 6 0 1 3 2.3
3.6 2.5 4 0 5 0 4 2 0 3 0 8 1 7 2 7
l a b l e 111. Comparison of Chemical hlethod and Selective Fermentation with Yeast for Determination of Glucose ( X g . per 2 ml.)
Solution Examined Sugar solution No. 1
Starchb substrate 0
1
0
3
0 8
4
8 8
8 0
n IO I1 12
13 11 .4v. a
b
0
0 4 0 8 0 16 0
Maltose 21.06 18.90 16.80 4.2 4.2 4.2 0 13.44 16.85 10.16 5.04 10.16 10.16 5 04
__
Chemical Met= Glucose recovGlucose ered Error 0 2 4 2 4 8 0 8 0 4.8 10 0 16.0 6.1 3.2 0
0.06 1.75 3.58 1.91 3.64 7.80 0.13 7 98 4.53 9.76 15.76 6.16 3.31 0
t 0 06 - 0 25 - 0 42 -0 o(, - 0 36 - 0 20 t 0 13 -0 02 -0.27 -0.24 -0.24 -0.24 -0 Ob 0 0 I!?
Selective Fermentation with Yeasta Glucose rerovered Error 0.42 2 16 4 20 1.81 2.62 i 42 0 24 8 02 4.74 9.68
f0.42 i.0.16 +0.20 -0.19 -0.38 -0.58 +0.24 10.02 -0 06 -0.32
6 25 - 0 . 1 5 3 . 1 3 -0.07 -0.18 -0.1s 0 23
Yeast 2019, kindness of h. Y. Sctiiiitz and I'ieisctiniann 1,ahoratories. Lintner's soluble potato starch adjusted to 0.01 L1 acetate. IJH 4 . 5 (a).
Table I V . Comparison of Chemical Method and Yeast Fermentation upon Analysis of Amylase Hydrolyzates for Glucose Hydrolysis Time, Hours 1
2 3
Glucose Produced by Amvla-e. Mg Der 2 111.of Hydrolh rate Starch" as Substrate Maltose" a s Substrate Chemical Yeast Chemical Yeast method fermentation method fermentation 5.84 5.23 2.10 2.25 9.58 11.10
9.31 11.21
3.00 4.46
3.65 4.74
a h glucose-forming amylase reacted with 1% Lintner's soluble potato starch or with 1%maltose, p H 4.5; 0.01 .M acetate and 40' ( 3 ) .
It is absolutely essential that a good grade of sodium acetate be used. The 50% solution of sodium acetate should give a value of approximately pII 8.8. Some solutions of so-called reagent grade sodium acetate give values as low as pH 6.7 a t 50% concentrat.ions. Such solutions are entirely unsatisfactory. Thorough mixing of reagents in the test tubes is essential both before boiling and after the second group of reagents has been added. Insufficient mixing before boiling will give incomplete precipitation of cuprous oxide. Insufficient mixing after addit,ion of the second group of reagent's permits the formation of difficultly soluble cuprous iodide rrhich shifts the equilibrium of the reaction:
+
~ C U + + 41-
2CuI
+ 1,
to the right and thus prevents an accurate determination of t,he amount of cuprous ion formed by the reduction with glucose. I n any event, a t least 30 minutes must be allowed after the second group of reagents has been added before the solutions are titrated. If less time is permitted for the reaction to come to equilibrium, a fading end point is obtained and the results cannot be checked. S o t more than 10 tubes should be prepared and boiled a t one time. Otherwise, the time that elapses before the addition of all reagents is completed will cause poor results, which may be due to oxidation of the cuprous oxide by air before the reagents are added. When the above precautions are observed, the use of the copper acetate reduction method in conjunction R-ith the iodometric method (1) gives a convenient procedure for the determination of glucose in the presence of maltose, which compares favorably with the accuracy obtained by yeast fermentation. The chemical method is more easily and rapidly carried out than the ferment,at,ion methods which are in more general use, and it is more adaptable for routine activity measurements. ACKNOW LEDG,\IE.YT
The authors wish to thank the Takamine Laboratory, Inc., far generous grants in aid of this investigation. The authors wish
to thank A. S. Schultz and the Fleischmann Laboratories for generous gifts of yeast and for helpful suggestions. LITERATURE CITED (1) Caldwell, M .
L., Doebbeling, S. E., a n d M a n i a n , S. H., ISD.
ENG.CHEM.,Aiv.4~.ED., 8, 181 (1936). ( 2 ) C o r m a n , b., and L a n g l y k k e , A. F., Cereal Chem., 25, 190 (1948). (3) Phillips, L. L., a n d Caldwell, 51. L., J . Am. Chem. Soc., 73, 3559 (1951). ( 4 ) Schultz, A. S.,F i s h e r , R. A . , A t k i n , L., a n d F r e y . C. S . , ISD. ENQ.CHEM.,ANAL.ED.,15, 496 (1943). ( 5 ) Shaffer, P. A., a n d H a r t m a n , A. F., J . Bid. Chem., 45, 349 (1920). ( 6 ) Steinhoff, G., Spiritusind., 56, 65 (1933). t i ] Zerban, F. IT., a n d S a t t l e r , L., ISD. ENC.CHEM.,h.41.. ED., 10, 699 (1938). RECEIVED October 20, 1950. Data taken from a dissertation submitted by 1,ouise Lang Phillips i n partial fulfillment of the requirements for the degree of doctor of philosophy in chemistry under the Faculty of Pure Scienre of C'uluinbia University.