Improved Technique for the Microdetermination of Silver by Linear Null-Point Potentiometry

Richard A. Durst, Erick L. May,' and John K. Taylor Dioision of Analytical Chemistry, Institute for Materials Research, National Bureau of Standards, Washington, D . C. 20234

IN the determination of submicrogram amounts of silver by linear null-point potentiometry ( I ) , a five-fold decrease in the error has been achieved by the use of several modifications of the technique. Except for external pregeneration of the standard silver solutions, the apparatus is identical to that previously described (I).

EXPERIMENTAL

Electrodes. The two silver indicator electrodes were annealed in a reducing flame to remove surface oxides, washed in dilute nitric acid, and rinsed with distilled water. This pretreatment of the electrodes was necessary to achieve the best matched indicating system. When electrodes treated in this way were placed in the same silver solution, the bias potential was typically less than 0.1 mV. In addition, it was found unnecessary to seal the electrodes into glass tubes with paraffin, thus simplifying their construction.

Instead of coulometrically generat- ing silver by increments in the titration half cell, a series of standard silver solutions are coulometrically prepared in separate 100-ml volumes of the inert electrolyte, 1N H2S04. By the use of such pregenerated standard solutions, the analysis time for a set of test solutions is reduced by a factor of about 4 or 5 to approximately 10 to 15 minutes per deter- mination.

The standard solutions are prepared in accordance with the concentration of the unknown silver solution. The approximate concentration of the unknown is first determined by a direct emf measurement or by a preliminary null-point determination using standard solutions encompassing a wide concentration range. A test solution approximating the unknown concentration and a minimum of two solutions above and below this concentration are then coulometrically prepared. These five standard solutions permit the deter- mination of the equilibrium null-point potential and provide two data points before and after the equivalence point for a reliable graphical or computer fit of the data to a straight line.

A 0.1-ml aliquot of the end point solution is taken from the standard solution (titrant) half cell (100 ml) and pipetted into the null-point (test solution) half cell. The porous Vycor null-point half cell is then immersed in the

Solution Preparation.

Procedure.

1 Present address, Department of Chemistry, Howard Uni- versity, Washington, D. C.

(1) R. A. Durst and J. K. Taylor, ANAL. CHEM., 39,1374 (1967),

standard solution, and the matched indicator electrodes are inserted into each half cell. The cell potential is measured with a high impedance recording system as before ( I ) . There is an initial potential drift, caused by the dissolution of the silver electrodes, which is more pronounced in the more dilute silver solutions. After about 5 minutes, a limiting potential is reached which remains essentially constant (<0.1 mV/10 minutes) for the remainder of the time required for a complete determination. This potential is used as the null point in subsequent determinations thereby compensating for the effect of silver dissolution. Once this null-point potential is established for the end point solution, the analyses of the unknown samples can be performed.

The 0.1-ml sample of the unknown is pipetted into the null-point cell, and the cell containing the standard silver solution of lowest concentration is used as the titrant half cell. After a stable emf is reached and recorded, the elec- trode is rinsed, dried, and the titrant half cell is replaced by the cell containing the next higher silver concentration. This substitution procedure is repeated until all of the standard solutions have been used and the data points recorded.

The data are evaluated as before, using both a graphical semilogarithmic plot and a linear least squares fit by com- puter. In the graphical analysis, the equivalence point is obtained by a linear interpolation to the equilibrium null- point potential. The equivalence point is obtained by com- puter techniques from the null-point intercept of a linear least squares fit of the data.

RESULTS

Standard 0.1-ml samples containing 0.054 to 54 pg of silver in 1N H2SO4 were analyzed. Several determinations were made at each concentration level and all showed good Nern- stian behavior with the slopes of the titration curves approach- ing the theoretical slope of 59.2 mV/pAg unit. Also, the deviation from linearity previously observed at the lowest concentrations was not evident in this work, presumably because of the establishment of the dissolution equilibrium prior to the actual measurement of the cell emf.

The summary of results in Table I gives the amount of silver present in the standard samples in terms of concentra- tion and micrograms/O.l ml. The average amount of silver experimentally found and the 95 % confidence limits for the average based on the indicated number of determinations are given in column 3. The error is given both as a percentage

Table I. Results of the LNPP Determination of Silver

Ag+ taken Ag+ found" No. of Error Slope Molarity pgl0.l ml pgl0.l rnl detns 73 (mV/pAgY

5 x 10-8 53.94 54.07 f 0.28 7 0.24 0.13 58.9 i= 0 . 3 5 x 10-4 5.394 5.410 0.038 9 0.30 0.016 59.1 f 0 . 3 5 x 10-1 0.5394 0.5412 f 0.0114 9 0.33 0.0018 59.1 f 0 . 5 5 x 10- 0.05394 0.05536 f 0.00205 7 2.64 0.00142 59.1 f 0 . 9

a The indicated uncertainties are the 95 73 confidence limits for the reported value.

VOL 40, NO. 6, MAY 1968 977

and in micrograms. An average five-fold decrease in the error over the previous study is observed at concentrations from to 5 X 10-6 M in silver. This improvement in accuracy is presumably due to the partial compensation of the dissolution error which had resulted in significantly higher results. A reduction in the standard deviation of the determinations was also effected.

DISCUSSION

The advantages of greater linearity of the titration curves, shorter analysis time, improved accuracy, and a reduction in

the standard deviation of the measurements, are achieved by pregeneration of the standard solutions and the establish- ment of an equilibrium null-point potential to reduce the error caused by dissolution of the indicator electrode in the null-point cell, In this study, 54 ng of silver were determined with an error of less than 2 ng, while at higher concentrations (above 10 p M ) errors of less than 1 can be easily obtained using the improvements in the LNPP technique discussed.

RECEIVED for review January 2, 1968. Accepted March 6, 1968.

Microdetermination of Diphenylhydantoin in Biological Specimens by U I traviolet Spectrophotometry Jack E. Wallace Forensic Toxicology Branch, USAF Epidemiological Laboratory, Lackland AFB, Tex. 78236

A NEED EXISTS for a rapid, sensitive chemical procedure capable of determining diphenylhydantoin at therapeutic levels in blood. The lack of a well defined ultraviolet ab- sorption spectrum for diphenylhydantoin makes its determina- tion by direct ultraviolet spectrophotometry difficult ; never- theless, the literature describes several methods which de- pend upon the spectrum of the unchanged drug ( I ) . Methods that require the conversion of the drug to benzophenone and recovery of the ketone by steam distillation have been re- ported (2, 3). For determining elevated levels of the drug, these procedures are adequate, but extensive concentration of the distillate is required for accurate analysis of therapeutic levels of the drug in biologic specimens. With the concen- tration technique, subtherapeutic to therapeutic levels often appear elevated and exhibit a significant degree of variability. This report describes a method which permits rapid and ac- curate analysis of even subtherapeutic levels of the drug. The method does not require separation of diphenylhydantoin from other acidic drugs with which it is frequently associated. Hydrolysis of the hydantoin ring in strong alkali is followed by permanganate oxidation of the resulting amide to benzo- phenone, which is subsequently extracted into heptane. In contrast to diphenylhydantoin, benzophenone has a well defined absorption spectrum and exhibits a high molar ab- sorptivity.

EXPERIMENTAL

Apparatus. A Beckman DK-2A ratio-recording spectro- photometer with linear wavelength presentation was used for the ultraviolet absorption measurements. The sample path length was 10 mm. Infrared spectrograms were prepared by means of a Beckman IR-4 spectrophotometer. A Barber- Colman Model 5000 gas chromatograph with a 6-ft column consisting of 2z carbowax 20 M on gas Chrom Q, 100-120

(1) 0. Svensmark and P. Kristensen, J. Lab. Clin. Med., 61, 501

(2) J. E. Wallace, J. D. Biggs, and E. V. Dahl, ANAL. CHEM., 37,

(3) J. E. Wallace, J . Forensic Sci., 11, 552 (1966).

(1963).

410 (1965).

0.80 7

810 LZO m MO zm ~m 270 280 zoo xo at0 azo pa0 aw w am MILLIMICRONS

Figure 1. Ultraviolet absorption spectra of diphenylhydantoin in water and of the diphenylhydantoin reaction product in n-heptane both corresponding to a concentration of 10 Mg per ml of diphenylhydantoin

mesh, was used for gas chromatographic analysis. The col- umn was operated at 200" C and the carrier gas was nitrogen flowing at a rate of 50 ml per minute.

Procedure. Ten milliliters of blood, serum, or urine are adjusted to a pH between 6 and 7 by the dropwise addition of 0.5N HCI and placed in a 250-ml separatory funnel to which is added 100 ml of chloroform. This mixture is shaken vigorously for three minutes. After separation, the chloroform layer is removed by filtration through Whatman No. 541 or equivalent filter paper into a graduated cylinder. The volume of recovered chloroform is recorded. Five millimeters of 1.ON NaOH are added to the chloroform and the mixture is shaken for three minutes. After separation, 4 ml of the aqueous layer are placed in a 250-1111 round bottom flask, which is then connected to a vacuum rotary evaporator assembly and placed in a water bath of 50"-60" C. With vacuum applied, the flask is rotated until the volume is re-

978 m ANALYTICAL CHEMISTRY