~~~~~
Table II.
Analyses of Ores b y Different Methods
Method Aa - Method Ba, Method C’, Mg. U3O*b Ore Type Sample wt. per sample %u308 ?& u308 % U3OJ 0 39 0 41 0 34 3 OG 12 Pitchblende 0 31 0 34 0 32 4 02 12 Carnotite 0 020 0 05 0 035 10 06 5 0 Fe,A1 silicate 0 65 0 68 17 5 0 73 2 51 Carnotite 3 31 3 36 67 3 35 2 00 Pitchblende 0 025 0 029 1 5 0 067 4 85 Phosphate 40 0 052 7 97 Carnotite 11 06 21 0 20 0 19 0 18 6 37 12 0 18 Carnotite 10.23 11 0.14 0.11 0.11 17.66 19 0.11 0 A, manganese dioxide-sulfuric acid. B, hydrofluoric-nitric acid, folloffed by sodium carbonate fusion; ion exchange separation. C, standard opening and separation of Xew Brunswick Laboratory. b Estimate from NeLv Brunswick analysis.
1954. Workers at the University of Nevada have also recently reported (7) its use in the assay of over 3000 ore samples. An experienced analyst is able to perform 20 separations per day, starting with the solution of the ore. ACKNOWLEDGMENT
The authors are indebted to the U. S. Atomic Energy Commission, Yew Brunsn-ick, N. J., for their cooperation in analyzing the ores used in this work. They acknowledge the capable assistance of Diana Kenny, who performed the majority of the analyses reported herein. LITERATURE CITED
(1j Arnfelt, A,, Acta Chem. Scand. 9, 1484
small. This laboratory has not performed sufficient ore analyses to calculate the over-all accuracy of the method starting with the ore. On solutions a n accuracy of =t2Y0 of the uranium content is expected in routine analpis. DISCUSSION
Although a definite procedure has been specified, the method outlined may be adapted to materials of a wide range of uranium contents by varying the sample size and the size of the resin bed. Because the resin serves to con-
centrate as well as to separate the uranium, materials of low uranium content may be analyzed by increasing the sample size or by decreasing the volume of resin and of the perchloric acid used in the elution. Similarly, larger amounts of uranium may be taken for analysis if the volume of resin and of eluting agent are increased, keeping them in the same ratio as those recommended in the above procedure. The ion exchange separation has been used for routine control, not only in this laboratory but also in the uranium purification plants in South Africa since
(1955). (2) Fisher, S., U. S. Atomic Energy Comm. Document RMO-2530 (1954). (3) Zbid.,RMO-2531 (1954). (4) Fisher, S., Kunin, R., Proc. Intern. Conf. Peaceful Uses of Atomic Energy 8 , United Nations, Sew Ynrk f,-195A’I (5) GrimGdi, F. S., May, I., Fletrher, At. H., Titcomb, J., U. S. Geol. Survey Bull. 100: (1954). (6) Rodden, C. J., Analytical Chemistry of the hlanhattan Project,” hlcGraw-Hill, New York, 1950. (7) Seim, H. J., Morris, R. J., Frerv, D. W., U. S. Atomic Energy Comm. Document UN-TR-5 (1956). - I
RECEIVED for review July 28, 1956. Accepted December 27, 1956.
Use of Thymol-Sulfuric Acid Reaction for Determination
of Carbohydrates in Biological Material M. R. SHETLAR and YANNA F. MASTERS Research laboratory, Veterans Administration Hospifal, and Department o f Biochemistry, University of Oklahoma School of Medicine, Oklahoma City, Okla.
b Different absorption curves were given by nearly all sugars when they were made to react with thymol in the presence of strong sulfuric acid. Glycosidic linkages did not affect the reaction either qualitatively or quantitatively. The presence of protein had a slight effect on absorption curves, but this appeared to b e due to the reaction of sulfuric acid with protein and did not involve thymol. The reaction, when used to estimate the carbohydrate bound to protein in the sera of a number of patients, correlates closely with an accepted method. 402
ANALYTICAL CHEMISTRY
T
HE REACTIOS of thymol with C ~ I bohydrates in the presence of strong sulfuric acid, as described by Udransky (4, has been modified for use as a method for the estimation of blood sugar by Alonzo and Bruna (1) and by Schmor ( 2 ) . As more specific methods are needed for the determination of carbohydrate in the presence of protein and other biological material the following work was undertaken to investigate this reaction.
Sulfuric acid solution, 77% by volume. Add 770 ml. of concentrated sulfuric acid (Du Pont reagent grade, specific gravity, 1.84 at 15’ c.)> to 230 ml. of distilled water. An American Optical Model 1A rapid scanning spectrophotometer and Beckman D U spectrophotometer were used for the work involving absorption curves. A Coleman Model 14 spectrophotometer was used for quantitative colorimetric work.
REAGENTS A N D APPARATUS
T o 1 cc. of sugar solution in a 15-ml. glass-stoppered test tube, 7 cc. of 77% sulfuric acid was added a t room tem-
PROCEDURE
Thymol reagent, USP, 10% in absolute ethyl alcohol.
pended in 1 ml. of distilled water. The mixture was subjected to the procedure described above for sugar solutions. DISCUSSION
MICROGRAMS OF GALACTOSE AND MANNOSE
Figure 1. Standard sugar curves for determination of serum glycoprotein Equimolar mixture of mannose and galactose used to prepare standard solutions -Beckman Model DU .Coleman 1 A
.. . . .
B
A
Figure 2.
Absorption curves after reaction with thymol-sulfuric acid reagent Glucose curves indicated by arrows Galactose and glucose 6. Glucose and fructose
A.
perature. The tube was left a t room temperature for 10 minutes without mixing, and was then placed in a n ice bath for 15 minutes. From pipets 0.1 ml. of the thymol solution and 0.9 nil. of water lvere added; the contents of the tubes \yere mixed by inversion and the tubes placed in a boiling water bath for 20 minutes. The tubes were removed from the miter bath and inimediately placed in a n ice bath, where they were left for 5 minutes. They were then kept for 25 minutes at room temperature before readings were taken in the spectrophotometer.
Serum protein samples were prepared by diluting human serum with five times the volume of 0.9% saline. Aliquots of 0.2 ml. were pipetted dropwise into glass-stoppered 15-ml. test tubes containing 10 ml. of absolute ethyl alcohol. The' stoppers were inserted and the contents of each tube mixed by inversion of the tube. The stoppers and sides of the tubes were washed down with 5 nil. of ethyl alcohol. The precipitate was centrifuged, mashed by suspension in 10 nil. of absolute ethyl alcohol, and centrifuged again. The alcohol was drained off and the precipitate sus-
Heating Time. Solutions of a n equimolar mixture of galactose and mannose and suspensions of serum samples prepared as described above were subjected t o t h e color reaction with thymol and sulfuric acid, in which t h e reaction time n a s yaried betneen 10 and 30 minutes. S o qualitative differences in t h e curves n e r e noted a t different times. As might be predicted, the niaxiniuni color development occurred niore slowly in the serum protein samples than in the sugar solutions; however, optimum color development occurred in 20 minutes for both. The absorptions of the reaction complex froni either pure sugar solutions or from serum protein were not decreased when heated for 25 minutes. Effect of Sugar Concentration. The reaction follon s the Lambert-Beer lan reasonably well in the 10- to 100-1 range when the Beckman DU spectrophotometer is used, but not when the Coleman 14 is used (Figure 1). This necessitates the preparation of a concentration curve Khen the Coleman 11 is used. I n work involving the quantitative detection of unknowns such a curve Ivas used and two standards a t different concentrations were run with each set of unknowns. Absorption Curves of Different Sugars. Curves of several monosaccharides are shown in Figure 2. These curves, niade with t h e A 0 scanning spectrophotometer, are superimposed photographically for easier comparison. A didymium calibration curve appears in both photographs. The wave length of maximum absoibance and relative absorption of a number of sugars or sugar derivatives are given in Table I. Curves of aldohexoses, ketohexoses. pentoses and methyl pentoses all differ froni each other. Absorption curves of the aldohexoses, galactose, mannose, and glucose differ slightly from each other, as do those of the ketohexoses, fructose, and sorbose and those of the methyl pentoses, fucose, and rhamnose. The curves of the pentoses, ribose. and arabinose, however, are nearly identical qualitatively. These data indicate that pentoses are degraded to a coninion derivative in the reaction. However, the curves of the pentoses are not identical with that obtained with fuifural. Apparently the reaction with pentoses is niore complicated than a simple degradation to furfural followed by coupling with thymol. Glucuronic lactone gives a curve similar to that of furfural. VOL. 2 9 , NO. 3, MARCH 1957
403
A
B
C
D
Table I. W a v e Length of Maximum Absorbance and Relative Absorptions of Sugar-Thymol-Sulfuric Acid Reaction Mixtures
(Determined on Beckman Model DU spectrophotometer) Mas. Relative Absorbance, AbsorpSugar 11pR tion Glucose 509 1 00 >Tannose 513 1 14 Galactose 511-12 0 76 Fructose 512-13 1 65 Sorbose 512-13 1 18 493- 4 1 15 Brabinose Ribose 4947 5 1 18 Fucose 503- 4 0 50 504 1 46 Rhamnose Galaheptose 494- 6 0 39 Glucoheptose 495- 9 0 44 Glucuronic acid lactone 49 1 0 85 Furfural 49 1 4 00 Glvcogen 509-10 0 (18 Inulin 513 1 63 Turanose 513 1 38 1 13 Raffinose 512 a -411 samples read against reagent blank.
Effect of Glycoside Linkage. The effect of t h e glycoside linkage was studied by comparing t h e glucose absorption curve with those of glucose-lphosphate, salicin (suligenin P-D-glucoside), trehalose (a-D-glucosido-a-Dglucoside), cellobiose [4-(P-~-glucosido)~-glucose],and glycogen. Inulin curves were compared with fructose; lactose m s compared with an equimolar mixture of galactose and glucose; and turanose was compared y i t h a mixture of glucose and fructose. I n all cases, the absorption curves derived from the complexes are those predictable from the monosaccharide components (Figure
3 )*
Effect of Protein. Colorimetric reactions for sugars are usually influenced by t h e presence of protein. An absorption curve of t h e thymol reaction with a serum protein sample precipitated as described above is shown in Figure 4. Human serum protein is known t o contain galactose, mannose, glucosamine, and fucose. Consequently, the serum curve was compared with a mixture of 25 y of mannose, 25 y of galactose, 5 y of fucose, and 30 y of glucosamine, As compared to the simple sugar curve, the protein curve absorbs more in the 420to 440-mp range of the curve. Absorption maxima are the same, however. Hydrolysis of serum protein (with 4N hydrochloric acid for 4.5hours) before carrying out the reaction resulted in only a slight change of the curve. Addition of crystalline pepsin (which contains very little carbohydrate) to the sugar solution results in absorption curves similar to serum protein curves; however, crystalline bovine
404
ANALYTICAL CHEMISTRY
Figure 3. acid
Effect of glycosidic linkage on sugar after reaction with thymol-sulfuric:
Absorption curves of complexes superimposed on monosaccharide component; all curves made a t concentration of 50 y of hexose or equivalent A. Glucose-1 -phosphate, glucose B. Inulin, fructose C. Glycogen, glucose D. Turanose, equimolar mixture of glucose and fructose
albumin had less effect. As the absorption curve obtained when sulfuric acid reacts with protein (in the absence of thymol) has some absorption at 420 mh (Figure 4), most of the effect is apparently due to the reaction of sulfuric acid with protein, and does not involve the reaction with thymol. Determination of Serum Glycoprotein. The thymol method was used as described above for t h e estimation of serum glycoprotein on a series of 22 human Serum samples. An analvsis by t h e tryptophan method of Shetlar, Foster, and Everett (3) was made on the same samples. The following results indicate good agreement between the methods. s o . of
Method Tryptophan Thymol Method Trvptophan Th”ymo1
Sera 22 22 Max. Diff., Mg. 70 22
Range, Mg. 70 122 to 248 125 to 257
Figure 4. Effect of protein on sugarthymol-sulfuric acid reaction All curves made with A 0 scanning spectrophotometer Curve indicated by a r r o w Reaction carried out on mixture of 25 y of galactose, 25 y of mannose, 5 y of fucose, and 30 y of glucosamine Upper curve: Blank of same sugars without thymol Curve next to top: Serum protein blank without thymol Lost curve: Reaction carried out on serum protein
Correlation Coefficient 0.971
By statistical methods, the difference betiyeen averages by the two methods
was not significant, and the correlation coefficient betneen the two methods. n-as highly significant.
The use of thymol-sulfuric acid appears t o have an advantage over tryptophan for quantitative work with biological samples, in that protein has less influence on the absorption curve.
LITERATURE CITED
(1) Alonzo, L. P., Bruna, J . JI., Laboratorzo 15, 301 (1953). ( 2 ) Schmor, J., Klzn. Vochschr. 33, 449 (1955). ( 3 ) Shetlar, M. R., Foster, J. Y.,Everett,
M. R., Proc. Soc. Exptl. Biol. X e d . 67, 125 (1948). (4) Udransky, L., Hoppe-Seyler's 2. phys201. Chem. 12, 355 (1888). RECEIVED for review July 16, 1956. Accepted October 22, 1956.
Estimating Total Absolute Activity of Small Radioactive Precipitates on Filter Paper PAUL
T. WAGNER, LOUIS R. POLLACK,
and CLARENCE G. DONAHOE, Jr.
Industrial laboratory, Mare Island Naval Shipyard, Vallejo, Calif.
b The total absolute activity of a small amount of a radioactive precipitate on filter paper, containing a simple, low-energy beta emitter, is estimated from the counts obtained from both sides of the paper. By means of a chart based on exponential precipitate distribution, a relationship between the two counting rates and the total activity (counts per minute) is obtained. This relationship is dependent upon the product of the absorption coefficient and the thickness of the paper (including precipitate). For purposes of counting, a close geometry is stipulated-e.g., a windowless flow counter. Total activity is converted to total absolute activity (d.p.m.1 by multiplying by a geometrical factor. This factor is the ratio of 4n to the solid angle subtended by the sensitive volume of the counter based on average precipitate position, including a correction factor for radiations absorbed by the walls of the counter. For radiocarbon precipitates, an accuracy within 10% of the absolute value is expected.
W
a small amount of a precipitate containing a simple, lowenergy beta emitter is filtered on paper, the material becomes embedded within the paper and the activity appearing at the surface is reduced by absorption of radiations by the paper and precipitate. This article shows that the total activity (activity which would have been observed in the absence of any absorption losses) can be obtained from the ratios of the observed activities of the top and the bottom of the paper. ,4 chart showing this relationship (Figure 1) is based on exponential precipitate distribution and close-geometry conditions of counting. As shown by Suttle and Libby ( 5 ) , activities obtained under close-geometry conditions can HEK
conveniently be converted to absolute activities. CLOSE-GEOMETRY AND ABSOLUTE ACTIVITY
For close-geometry conditions, the absolute specific radioactivity was shonm to be related to the observed activity of either an infinite or a finite thickness of precipitate (6). This relationship is based on the values for the absorption coefficient, the surface area of the sample, and the factor G. G is the ratio of 4r to the solid angle subtended by the sensitive volunie of the counter from the point source on the precipitate being considered. These same arguments apply also to closegeometry placement of filter paper samples in end-window positions, as well as to the cylindrical placement used by Suttle and Libby in their screen-n-all counter. I n this case the filter paper with precipitate represents a finite sample thickness. I n end-window placement. the geometrical factor, G, can be considered from a standpoint of average precipitate position, with a correction for the radiation absorbed by the walls of the counter. To convert total activity counts per minute as obtained from Figure 1,to total absolute activity (disintegrations per minute), we need only to multiply the total activity obtained by the value of G. I n a 2n counter, G would be expected to be only slightly greater than 2. EXPONENTIAL DISTRIBUTION
I n a filtered precipitate the particles tend to be concentrated on top of the paper, the concentration (expressed as activity per unit thickness of paper) decreasing with increasing depth. Such a distribution of particles can readily be envisaged as exponential. This mathematical relationship satisfies a large number of precipitate distributions, while a given proportionality
constant will define a particular distribution. I n an exponential distribution of precipitate:
- dc dl
E
kc
and c = c0ebk1
where c
(1)
activity per unit thickness at depth 1 which would be observed if there were no absorption of beta rays co = activity per unit thickness at zero depth k = proportionality constant, sq. cm. per mg. 1 = depth within filter paper, mg. per sq. cm. =
If we consider a sample of filtered precipitate with an exponential distribution, the total activity can be considered to be the sum of the activities of an infinite number of infinitesimally thin layers. Using Equation 1, the total activity, z , is as follows:
where g
=
total thickness of paper with precipitate, mg per sq. em.
Kunierous authors (1-5) have shown the applicability of self-absorption equations relating measured activity to total activity in solid radioactive samples. These relationships are based on exponential beta-ray absorption for homogeneous samples where the activity from the top is measured. For exponential precipitate distribution, and exponential absorption of beta rays, the measured activity, 2, from the top side of filter paper becomes:
VOL. 29,
NO. 3,
MARCH 1957
405
