V O L U M E 2 6 , NO. 5, M A Y 1 9 5 4
931 tillation, eicept to ~ a i until t the proper temperature is reached. This entire determination should take about 1 hour to complete n ith results comparable to the best of the aforemeritionecd methods.
Table I. .4nalysis of SJnthetic 3Iixtures Containing Eth: lene Gl\col, 1,2-Propylene Glycol, and Borax I P - P r o p l lene Glycol,
i n samgle
5
round
Error, %
09 04 09 09
+o -0 +o
+o
01 04 01 01
' 95
10 10 10 07 10 00 10 00
i o +o +o +o
15 12 05 05
;o
14 70 14.80 14.60 14 90
+'ri.'lO - 0 10 +o 20
3 08
:4
8 3 3 3
.4CIiSOULEDG3IENT The author is indebted to -4ndrew Davis and n'athan Ingher, \\.hose n o r k and ideas proved invaluable to the successful completion of this investigat,ion. LITERATURE CITED (1) Cannon, IT.A , , and Jackson, L. C., .INAL. CHEM.,24, 1053 (1952). ( 2 ) Dal Sogare, S., Sorris, T. O., and Mitchell, J., Jr., I b i d , , 23,
1473 (1951).
(3) Fuson, R. C., "Advanced Organic Chemistry," p. 378, Sew Tork, John Wiley & Sons, 1950. (4) Jordan, C . B., and Hatch, V. O.,ASAL. CHEM.,25, 636 (1953).
(5) Reinke, R. C., and Luce, E. X., IND.ENG.CHEM.,ANAL.ED., 18,244 (1946). (6) Shell Development Co., Utrariolet Spectral Data, d.P.1. Project 44, Satl. Bur. Standards, Serial 326, Sept. 30, 1949. (7) Shriner, R. L., and Fuson, R. C., "Identification of Organic Compounds," p. 105, Yea- Tork. John Wiley & Sons, 1948. (8) STarshoc\-sky,B., and Eli-ing, P. ,J., ISD. ESG. CHEX, ; ~ I C A L ED., . 18, 2-33 (1946).
nisCussIoN Representative (lata oil eynthetic antifreeze mixtures are s1ion.n in Table I. These mixtures contain 2.5% boraxu,which is used as a corrosion inhibitor. Its presence did not interfere with the determination. There can l w no interferences from po!ypropylene glycol with this method. :is only the vicinal glycols are oxidized by periodic acid and onl>- their osihtion products come over in the distillate. S Ofurther c~hemicalnianipulation is neceqeary after dip-
RECEIT-ED for reriew J u n e 12, 1$1.3. l c c e g t e d February 10, 1054. The oiinions expressed in this paper are those of the author a n d are not necessarily official opinions of t h p t-,S . Sal-a1 Esperiinent Station or the S a v y Deiiartnient.
Aluminum Powder as a Binder in Sample Preparation For X-Ray Spectrometry ISIDORE ADLER and 1. M. AXELROD U. S. Geological Survey, Washington 25, D.
C.
S
.4MPI,E preparation is an important, ever-present problem in x-ray fluorescence spectrometric analysis of powders. In particular, the density of packing, the surface smoothness, and the position of the sample may have a marked effect on the measured intensities. I t was felt that briquets or compacts, eompressed to the same extent, offered one obvious solution. This approach has been tried and found to be eminently satisfactory.
niobium in an iron oxide matrix, a briquet with a 1 to 1 dilution resulted in no more than a 5 to 10% loss of absolute intensity over the use of the sample powder alone. This is, of course, a favorable case because of the relatively short wave length of the niobium line. It is to be expected that the attenuation would be greater ior elements of low atomic numbers, but even here the use of the aluminum powder briquets offers advantages. Homogeneity of the specimens was also investigated. Five briquets n-ere prepared containing 57, each of niobic and molybdic oxides. The ratio of intensity of the molybdenum K a line to the niobium K a line was measured ten times for each briquet, using a multi-wave-length spectrometer built in this laboratory (1). One goniometer was set for the molybdenum K a line and the other was set for the niobium K a line. The briquets were changed after each measurement. Each side of each briquet was measured five times and the positioning was random. Be-
PROCEDURE One half gram of minus 200-mesh sample or standard is intimately mixed with 0.5 gram of minus 270-mesh aluminum metal dust. The diist, obtained from commercial sources, required further sieving. A convenient method of mixing the powder and aluminum dust is to grind the mixture under ether, a technique commonly employed in some emission spectrographic laboratories. Enough ether to make a thin mud is added to the mixture, which is ground until the ether evaporates. This process repeated three or four times gives good mixing ( a well ventilated hood is required). The mixed sample is then compressed in a briquetting press with a mold 1 inch in diameter a t approximately 30,000 pounds per square inch. This results in a disk 1 Table I. inch in diameter and approximately 0.05 inch thick. These briquets have Briquet S o . the appearance of aluminum disks, are A Sidea ~ t r o n g and , can be handled and stored 1.35 with ease. 1.35 TESTS FOR INTENSITY AND HOMOGENEITY
Lose of interlsity due to dilution by aluminum powder was investigated.
~~~
Comparison of the Ratio 1
2
-__
3IOKcP - for Five Briquets NbKa 3
4 5 B -4 B I B 1.36 L37 1.40 1.38 1.37 1.39 1.38 1.34 1.34 1.38 1.33 1.35 1.34 1.34 1.33 1.36 1.32 1.36 1.38 1.31 1.36 1.85 1.38 1.3H 1.36 1.37 1.37 1.33 1.32 1.36 1 . 3 8 1.33 1 34 1 34 1.35 1.33 1.32 1 31 1.32 1.33 1.32 1.34 1 34 1 28 1.30 1.30 1.34 1 28 dverage 1.344 1.348 1.352 1 . 3 6 4 1.356 1.338 1.346 1.348 1.338 1.324 Because a multi-ware-length spectrometer, featuring separate optics for each line. was used, there need not be 1 to 1 correspondence between ratios of concentrations and intensity ratios observed. X b K a count = 6400. b Labeling is random.
n
A
n
4
ANALYTICAL CHEMISTRY
932 cause single-slit collimators were used, radiation from only a small portion of the sample was measured each time and it was felt that this would afford a good test of homogeneity. Table I shows the individual measurements and the average? for the ratio
MoKa NbKa'
ADVANTAGES OF T H E BRIQUETTINC TECHNIQUE
-Istatistical analysis of the data for the presence of honiogeneity is summarized in Table 11. Table 11.
quets. However, the greatest variation found for two sides of one briquet was 1.3% and this is an acceptable precision for ore and mineral analysis.
The advantages of this briquetting technique are ease of mixing; convenience of handling and storage; nonhygroscopicity of aluminum, .permitting good mixing and making of stable briquets; reproducibility of sample preparation; and simple method for incorporating an internal standard. (The internal standard is easily incorporated into the aluminum powder.)
Summary of Statistical Test for Homogeneity
Briquets Between sides of briquets Between determination on same side
df
.\lean Sgt:ares
F
ACKNOW LEDGM E S T
4
5
0.000673 0.000371'
2 5
40
0,000101
The authors wish to thank IT.J. 'I-ouden and llarvin Zelen, Xational Bureau of Standards, for their aid in the preparation and interpretation of the statistics.
? i
The analysis reveals that determinations from different sides made on the same briquet show less agreement than deterniinations made on a single side and that the briquet averages do not differ as judged by the agreement shown by the two sides of bri-
LITERATURE CITED
(1) Adler, I., and .Ixelrod, J. LI.,J . O p t . Soc. A i w r . , 43,7G9 ( 1 9 S ) .
RECEIVED for review August 25, 1953. lcceyred February 5 , 1954, lication authorized by the Director, C . s;.Geologicai Siin-ey.
Pub-
Determination of Methanol in Biological Fluids by Microdiffusion Analysis MILTON FELDSTEIN and NlELS C. KLENDSHOJ Division of Toxicology, University of Buffalo, School of Medicine and the Biochemistry Laboratory, Buffalo General Hospital, Buffalo, N. Y.
T
HE methods that have been proposed for the detection and estimation of methanol in biological fluids have all been based on an initial distillation to isolate the volatile alcohol (3, 4). This step necessitates the use of special equipment and fairly large samples of biological material, and is thus unrvieldy for routine determinations. In the method described here, the alcohol is separated from the biological material by diffusion in a standard Conway microdiffusion cell (2, 6 ) , thus eliminating the necessity for distillation and permitting the use of small aliquots of material. The methanol is absorbed by a solution of sulfuric acid in the center well of the Conway unit, and then determined quantitatively by oxidation to formaldehyde and subsequent reaction with chromotropic acid (1,8-dihydroxpnaphthalene-3,6-disulfonicacid) ( 1 ). The specificity of the reaction between formaldehyde and chromotropic acid is indicated by the fact that alcohols, ketones, and the following aldehydes do not react to give a colored solution: acetaldehyde, propionaldehpde, butyraldehyde, isobutgraldehyde, isovaleraldehyde, chloral, glyoxal, benzaldehyde, and phthalaldehyde. Glyccraldch) de gives a yellow color ( 1 ). REAGEATS
Sulfuric acid. Dilute 10 ml. of concentrated sulfuric acid to 100 mi. with distilled water. Potassium carbonate, saturated aqueous solution. Chromotropic acid, 0.570 aqueous solution. Prepare fresh weekly and keep under refrigeration. Potassium bisulfite, saturated solution. Potassium permanganate, 5 % aqueous solution. Sulfuric acid, concentrated, reagent grade. White petroleum vaseline lubricant. PROCEDURE
Pipet 2.2 ml. of the 10% sulfuric acid solution into the center well of a Conway unit. I n the outer compartment place a n aliquot of biological material (0.5 ml. of blood or urine). Apply the vaseline lubricant to the surface of the ground-glass cover plate and place on the unit P O that only a small portion of the outer compartment remains unroveietl. .\(Id to the outer compartment
1.0 ml. of the potassium carhonate solution and *ea1 the unit. Carefully tilt the cell several times to allow the fluids in the outer compartment to mix thoroughly. .Illow the cell to stand at room temperature for 2 hours. Pipet 1.0 ml. of the sulfuric acid solution from the center \vel1 into each of two 25-ml. test tubes (19 X 150 nini.). Add to one of the tubes 1 drop of potassium permanganate solution and allon to stand for 5 minutes, mixing occasionally. At the end of this time, add potassium bisulfite dropwise until the excess permanganate is completely decolorized. The second tube serves as a control to detect the presence of formaldehyde and is not treated with permanganate. Prepare a third tube containing 1.0 nil. of distilled water as a blank. To each, add 0.2 ml. of the chromotropic acid solution and place in an ice bath. Add 4.0 ml. of concentrated sulfuric acid to each tube, mix well, and place the tubes in a boiling water bath for 15 minutes. Cool to room temperature, transfer quantitatively to a IO-ml. volumetric flask, and dilute to mark. Cool again, and redilute to the mark. The absorbance is then determined in a spectrophotonieter a t a wave length of 580 mp, with the blank set a t zero density. calibration curve is prepared by oxidizing known amounts of methanol in 10% sulfuric acid according to the procedure described above. The Beer-Lambert laws apply over the range of concentration 0.004 t o 0.08 mg. of methanol. This represents a range of concentration in 0.5 ml. of biological sample of 1.8 to 35 mg. %. If the concentration is greater than 35 mg. %, the determination is repeated on 0.5 nil. of a diluted aliquot of biological material. If the concentration is less than 1.8 mg. yo (0.004 mg. methanol in 1 ml. of absorbing solution), the determination is repeated on a larger volume of biological material. However, in this case it is necessary to determine a new correction factor for the diffusion equilibrium as described in the experimental section. I n the absence of formaldehyde-i.e., control tube remains colorless-unknowns are read directly from the calibration curve, and corrected for incomplete diffusion as described below. When small amounts of formaldehyde are present, the ahsorbance of the unoxidized control sample is subtracted from the absorbance of the oxidized sample. If the sample is grosslv contaminated with formaldehyde, the method is not appllcuble. EXPERIMCVT 4 L
Known amounts of methanol were added io I1100d a n d urine samples which weix shown to he free of methanol and were as-
