MARCH, 1974 THE ANALYST Vol. 99, No. 11 76 Coulometric Microtitration of Arsenic(II1) and Isoniazid Using a Vitreous Carbon Generating Electrode BY V. J. JENNINGS, A. DODSON AND A. HARRISON (Department of Chemistry and Metallurgy, Lanchester Polytechnic, Priory Street, Coventry, C V l 5FB) A small-scale coulometric titration cell and apparatus is described in which a vitreous carbon rod anode is used to generate bromine. The amount of the latter required to titrate from 3 to 15 pg of arsenic(II1) or isoniazid with a precision of 2 per cent. is found by using an integrating digital milli- ammeter. The titration end-points were obtained by use of a form of differential electrolytic potentiometry. PREVIOUS work has shown that bromine can be generated at a vitreous (or glassy) carbon electrode with a current efficiency of at least 99.9 per cent.1 In the present work results obtained by using a vitreous carbon electrode to generate bromine in the coulometric micro- titration of arsenic( 111) and isoniazid (isonicotinylhydrazide) are reported.There is general interest in analytical methods for determining microgram amounts of both substances, arsenic because of its toxicity and isoniazid because of its use as a chemo- therapeutic agent for the treatment of tuberculosis.2 Conventional oxidimetric titration methods have been reported for the determination of arsenic(I1I)s and isoniazid4 in which the oxidising agent is potassium bromate, occasionally in the presence of added bromide ions. For the latter condition the chemical reactions are- for arsenic(II1) and for isoniazid.Coulometric titrations with electrolytically generated bromine at a platinum anode have been reported for determining a r s e n i ~ ( 1 I I ) ~ ~ ~ and isoniazid.' A major difficulty in micro-titrations in which the initial concentration of the titrand is approximately 10-5 M is in the location of the end-point, as the detection system must be sensitive to concentration levels of about lo-' M in the vicinity of the equivalence point if an end-point precision of 1 per cent. is to be achieved. In the present work, a form of differential electrolytic potentio- metry has been used to locate the end-point because it is known to provide an end-point detection method of high sensitivity in, for example, the coulometric determination of hydrazine with bromate.* In differential electrolytic potentiometry, the potential difference between two platinum electrodes that have been polarised by the passage of microampere currents is measured.For arsenic(II1) and isoniazid titrations it is expected that the complete titration graph of potential versm amount of bromine generated would show a relatively high potential plateau region and a sharp fall in this potential at the end-point. However, in order to decrease the time required to complete a titration in the present work, it was decided that instead of attempting to plot a complete graph, the amount of titrant required to produce an arbitrarily chosen decrease in potential of 1oOmV from the initial start potential would be taken as the end-point value.In a conventional coulometric titration, the current is maintained at a constant value and the elapsed time from the start to the end-point is used to determine (from Faraday's laws of electrolysis) the amount of titrant generated. After using some commercially available coulometric titration apparatus, we have found that it is not always a simple matter to main- tain a constant current and there are difficulties in synchronising the current flow with the 145; As3+ + Br, = As5+ + 2 B r .. .. * * (1) (2) C5H,N.CONHNH, + 2Br, + H,O = C,H,N.COOH + 4 B r + 4H+ + N, @ SAC and the authors.146 [Analyst, Vol. 99 timing system. There have been attempts to use a low-inertia integrating meter as a coulo- nieter in series with the titration cell9 in order to measure the charge directly. Integrating electronic digital milliammeters have recently become commercially available and such an instrument was used successfully in this work.JENNINGS et al. : COULOMETRIC MICROTITRATION OF ARSENIC(III) EXPERIMENTAL REAGENTS- Reagents of analytical-reagent grade quality were used when possible. The isoniazid (isonicotinylhydrazide pure, Koch-Light Laboratories Ltd.) was recrystallised from solution in ethanol and dried at 110 "C for 2 hours. Its melting-point after recrystallisation was 170 to 171 "C. E 0- F' P -N -M H A B C D E F G H J K L & M N 0 P Q R Griffin constant-current, 1 - to 10-mA supply unit (S576-130) Time Electronics Digital Integrator Type TS1 OOA (1 00-mV range) Beckman Research pH meter (mV mode) (also used to measure the potential across L and M) ALMA Type SIT 50sZ f 0.05 per cent, resistor Le Carbone (Great Britain) Ltd.3-mm diameter vitreous carbon rod anode (V25) Johnson Matthey Ltd. thermopure platinum wire cathode, 0.5-mm diameter Glass cathode compartment Porous plug (coil of Whatman No. 541 filter- Magnetic stirrer bar Bottom half of glass cell, capacity 3 ml D.E.P. 0.5-mm diameter p!atinum wire elec- trodes Sovirel bored cap for sliding joint, size 22 mm PTFE wrapped silicone rubber sealing ring (22 x 16) Top half of glass cell White-spot nitrogen inlet Nitrogen outlet (also used for adding sample with hypodermic syringe) paper) Fig. 1. The cell and coulometer circuitAIarCh, 197-11 water and making the volume up to 500 ml in a Grade A calibrated flask.previously described.1 AKD ISOXIAZID USING A VITREOUS CARBON GENERATING ELECTRODE 147 Isoniazid solution, 0.01 &I-This was prepared by dissolving 0.2 g of the solid in distilled Arscnic(II1) oxide solzttioiz, 0.01 nf-This was prepared in a similar manner to that APPARATCS- The cell and couloiiieter circuit are sliowii in Fig. 1. The integrator display, which was a six-digit totalising magnetic counter with maiiual zero re-set, was calibrated by passing a known (2 mA) constant current for a given length of time (measured on a stopwatch to h O . 1 s). The mean value of 1 integrator unit was equivalent to a charge of 6-652 x C (theoretical value, 6.666 x lo-* C) and the relative standard deviation from the mean value was 0.1 per cent. for timed periods oi 100 s.The surface area of the vitreous carbon rod exposed to the electrolyte solution was 0-16 cm2 with a current density of 12 mA cm-2. A for the differential electrolytic potentiometric platinum electrodes was supplied by a 90-V Exide Dymax dry battery (DM 256) in series with a W7elwyn Electric Ltd. 100 MQ (&5 per cent.) carbon film resistor. The current density a t the differential electrolytic potentiometric electrodes was about 12 x A cm-2. The potential across the electrodes was measured on a Beckman Research pH meter used in its millivolt measuring mode. The polarising current of about PROCEDURE- For carrying out titrations tlie method of successive aliquot additions, as described previously,1° was used. To the cell was added 2 ml of electrolyte solution that was 0.2 M in potassium bromide and 1-0 nr in sulpliuric acid.White-spot nitrogen was bubbled through the cell for 5 minutes so as to remove any titratable volatile impurity, then an aliquot of titrand was added by use of a 10-pl precision syringe (S.G.E. Pty. Ltd.). The potential across the two differential electrolytic poteiitionietric indicator electrodes was measured and the amount of charge, in integration units, required to decrease this potential by 100 mV was found. ,4 further aliquot of titrand was tlieii added and the above procedure repeated. Naturally, the result for the first aliquot included any blank value, that is, the charge required to cause the 100-mV potential change in the absence of sample. This result was therefore high and was rejected.With the isoniazid titrations it was found that the result for the second aliquot could also be high so that this result was again rejected. It is believed that some pre-conditioning of the diff ereniial electrolytic potentiometric electrodes is necessary for isoniazid titration, as has been found by other workers in the coulonietric titration of sulphur dioxide with 1)romine.ll The results are given in Tables I to IV. It was observed that there was a slow downward drift of the initial differential electrolytic potential after the addition of each aliquot. TABLE I COULOXETRIC TITRATIOKS OF 10 p1 o b 0.01 M ARSENIC(III) OXIDE SOLUTION 10 p1 of solution zz 15.13 pg of arsenic(II1) Experiment series . . .. * . .. A r, C D E Nuiiiber of individual results in series ..8 8 8 1Ican amount 01 arscnic(II1) fouIid/pg . . 15.16 15.31 15.19 15.39 15.18 Relative staiidard deviation, per cent. . . 2.1 1-8 1.0 1-8 1-9 Error, per cent. . . .. . . .. . . t o e 1 f l - 2 +0*4 +1.4 + 0-4 c 10 DISCUSSION AND CONCLUSION A RS E x IC ( I I I) - The results in Tables I and 11 show that the accuracy and precision are of the order of 1 and 2 per cent., respectively, except for the srnallest amount of sample (3 pg). At that level there is an accumulation of random errors, for example, the fact that there is an experimental limit to the minimum incremental amount of bromine that can be generated. A general limit on tlie precision level is imposed by the use of a microsyringe, with which the volume delivered is subject to a probable random error of 1 per cent.148 JENNINGS, DODSON AND HARRISON TABLE I1 COULOMETRIC TITRATIONS OF 2 TO 8 pl OF A 0.01 M ARSENIC(II1) OXIDE SOLUTION Volume of hs,O, solution taken/pI .. . . 2 4 G 8 Number of individual results in series .. 8 8 8 8 Amount of arsenic(II1) presentlpg . . . . 3-03 6.05 9-08 12-11 Mean amount of arsenic(II1) found/pg . . 3.15 6.14 9.04 12.22 Relative standard deviation, per cent. . . 2.5 1.5 0.8 1.9 Error, per cent. .. .. .. .. . . f 4 . 2 +1*4 -0.4 + 1.0 A least-squares fit of the results given in Tables I and I1 shows that if x pg of arsenic(II1) are taken, then the y pg of arsenic(II1) found is given by the equation- I SON I AZI D- The results in Tables I11 and IV are similar to those for arsenic(II1) although the precision is lower and there is an appreciable increase in the error.A least-squares fit of the results given in Tables I11 and IV shows that if xpg of isoniazid are taken, then the ypg of isoniazid found is given by the equation The error may be partly due to the possibility that the solid, although recrystallised, was not absolutely pure. However, it is also likely that at these concentration levels the rate of reaction of isoniazid with bromine and the instrumental response of the end-point detection system limit the precision and accuracy that can be attained. TABLE I11 COULOMETRIC TITRATIONS OF 5 pl OF 0-01 M ISONIAZID SOLUTION 5 pl of solution = 6.88 pg of isoniazid Experiment series . . .. .. .. A B C D E Relative standard deviation, per cent. . . 2.0 1.6 4.0 3.1 2.2 Error, per cent... .. . . .. . . +1*4 4 3 . 5 $7.3 +5*6 +5.5 It is concluded that the apparatus and technique described in this paper are suitable for the coulometric titration of down to 3-pg amounts of arsenic(II1) and isoniazid and that the method could be applied to the determination of large numbers of samples as a result of the speed with which these titrations can be carried out. y = 1.002~ + 0.07 y = 0.981~ + 0.390 Number of individual results in series .. 8 8 8 9 8 Mean amount of isoniazid found/pg . . . . 6-97 7.1 1 7-35 7.24 7.23 TABLE IV COULOMETRIC TITRATIONS OF 2 TO 10 pl OF 0.01 M ISONIAZID SOLUTION Volume of isoniazid solution talrenlpl . . 10 8 6 4 Number of results in series . . .. .. 8 7 8 8 Amount of isoniazid takenlpg .. . , 13.75 11-00 8.25 5.50 Mean amount of isoniazid found/pg . . . . 13.85 11-22 8.58 5.56 Relative standard deviation, per cent. . . 1.6 1.3 1.9 2.8 Error, per cent. .. .. . . . . . . +0*7 +2-0 +4.1 +1-1 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES 2 8 2.75 3.16 2.7 + 14.8 Jennings, V. J., Dodson, A., and Atkinson, A. M., Analyst, 1972, 97, 923. Ellard, G. A., Gammon, P. T., and Wallace, S. M., Biochem. J., 1972, 126, 449. Bishop, E., Analyst, 1958, 83, 212. Vulterin, J., and Z$ka, J., Talanta, 1963, 10, 891. Myers, R. J., and Swift, E. H., J . Amer. Chem. Soc., 1948, 70, 1047. Lee, J. K., and Adams, R. N., Analyt. Chem., 1958, 30, 240. Kawamura, F.. Momoki, K., and Suzuki, S., Bull. Fac. Engng Yokohama Natn. Univ., 1955,4, 123. Bishop, E., Mikrochim. Acta, 1960, 803. Smythe, L. E., AnaZyst, 1957, 82, 228. Jennings, V. J., Dodson, A., and Tedds, G., Talanta, 1973, 20, 681. Bailey, P. L., and Bishop, E., Analyst, 1972, 97, 311. Received Octobev 5th. 1973 Accepted November 12th, 1973