EFFECT OF DILUTION IX ETAECTRO-TITRIMETRIC ANALYSES. 55 IT V. -The Efect qf Dilution in Blectro-titmhetric Analyses. By GILBERT ARTHUR FREAK. THE first application of conductivity measurements to analysis is that due to Kiister and Griiters (Zeitsch. anorg. Chem. 1903 35, 454) who showed that acids could be titrated accurately by these means. Later Kiister Griiters and Geibel (ibid. 1904 42 225) proved the accuracy of the method even when such substances as potassium dichromate and potassium permanganate were present in the acid solution. The estimation of acetic acid in vinegar of total acid in red wine of magnesia and o,f various alkaloids was also shown to be possible. Further work has demonstrated that the method is capable of very varied application. Amongst the uses to which i t has been put may be noticed the analysis of wines by Duboux (C'hem.Zed., 1913 37 879) and by Duboux and Dutoit (Compt. rend. 1908 56 FREAK THE EFFECT OF DILUTION IN 147 134) the preparation of neutral ammonium citrate solutions by Hall (J. Ind. Eng. Chem. 1911 3 559) and the analysis of soil solutions by Van Suchtelen and Itano (52nd ,4nn. Rep. Mich. B w d of ,4gm'c. 1913 49). The lastlnamed workers have also pub-lished ( J . Amer. C'hem. SOC. 1914 36 1793) the results of experi-ments on the estimation of chlorides sulphates nitrates phosphates, potassium calcium ferrous iron strong and weak acids and of chlorides and phosphates in urine. Quite recently Harned ( J . Amer. Chem. SOC. 1917 39 252) has shown that certain bivalent metals in the form of their sulphates can be determined accurately by the conductivity method by ticration with barium hydroxide.Meerburg (Versl. w. h. Centr. Lab. t. b . h. v. h. Staatsoez. 0. d. Vdksgehondh. 44-54 1917 ; Chern. TT/'eel;blad 1917 14 1054) has reported adversely on the method as applied t o the estimation of sulphates by barium acetate and of calcium by oxalic acid but mentions that good results may be obtained in the determination of alkalinity in potable waters. It is noteworthy that although the method has been applied to SO many reactions no attention has been paid to the1 lower limit of concentration at which accuracy may still be obtained. Most of the experiments have been carried out on relatively concentrated solutions (seldom weaker than O-llV) the only reference to results with very dilute solutions being one by Van Suchtelen and Itano (loc.cit.) who state that the titration of as little' as 5 C.C. of 0.001N-sulphuric acid with O.OIAT-sodium hydroxide can be per-formed accurately. I n those titrations involving the precipitation of a salt the solubility of which would be expected to limit the sensitiveness of the method this p i n t has not been touched upon, It appears therefore to the author that an investigation of this nature was desirable. If the method is capable of yielding accurate results a t very low concentrations many estimations f o r example, those carried out in the analysis of potable waters could be made without previous concentration of the solutions. The present com-munication is concerned with the limits of the method as applied to the estimation of sulphates chlorides calcium and magnesium.Rigid accuracy was not aimed at the object being t o find t o what extent the method could compete with ordinary gravimetric or volumetric processes without the introduction of troublesome pre-cautions. To that end beyond the use of standardised measuring vessels and of " purest " commercial reagents no special precautions were taken. No attempt was made to keep the temperature of the s0l;utions constant during the titrations the duration of which was usually fifteen t o twenty minutes ELECTRO-TITRIMETRIC ANALYSES. 57 F x P E I M E N T A L. Except f o r the use of a double receiver telephone which proved very convenient i n minimising the interference of external noises, the apparatus employed was of the usual nature and therefore calls for no special comment.The liquid t o be1 titrated was placed in a beaker of such a size that thorough mixing could be effected by giving the vessel a rotatory motion the stationary electrodes serving as a stirrer. The reagent was ddivere'd from a burette capable of being read to 0.01 C.C. The curves were plotted with conductivity as ordinates and volume FIG. 1. 0.024 0.023 0.016 0.01 4 C.C. N/10-BaCI2. of reagent as abscissz and in order to eliminate experimental errors a t least three and usually more readings were taken on each limb' of the curve. Estimation of Sulpha tes. A stock solution of Xerck's purest potassium sulphate was pre-pared and the SO estimated gravimetrically in duplicate as barium sulpliate.From this solution weaker solutions were prepared by dilution. For the titration of these two solutions of barium chloride approximately IT/ 10 and iV/ 25 respectively were prepared and silrilarly standardised. Table I shows that the results are accurate only down to a con-centration of about 200 milligrams of SO per litre a typical curve being shown in Fig. 1. When the concentration of SO is 100 milli 58 FREAK T’HE EFFECT OF DILUTION IN grams o r less per litre tlie precipitation of barium sulphate is so incomplete that the resulting curve changes entirely in character, exhibiting irregularity but no definite break such as appears in Fig. 1. Van Suchtelen and Itano (loc. cif.) added to their solutions a certain amount of the salt that was to be precipitated with the object of avoiding errors due to solubility.I n order t o test this procedure a suspension of fine precipitated barium sulphate in dis-tilled water was added to the sulphate solution in the beaker some little time prior to titration. The results are given in Table 11. TABLE I. Concsntration of SO, Strength of so found. mg. per litre. BaCl,. Per cent. 990 N/10 100.0 100.4. 99.4 198 99 49.5 Y Y 99.2; 99.2 .99-2 N,%5 -l f No end-point. TABLE 11. Concentration of SO, Strength of SO found. mg. per litre. BaCl,. Per cent. 99 N/10 99-7 99.7 49.5 24.7 9.9 N j25 99.5; 99.2 99-6 ” } End-point indefinito. Reference to table I1 will show that this method succeeded to a certain extent as by means of it good results were obtained down t o a concentration of about 50 milligrams of $0 per litre.How-ever it was not effective when only 25 milligrams of SO per litre were present giving curves with an indefinite end-point of tlie type shown in Fig. 2. Titratioii a t boiling temperature did not alter the character of this curve. I n such cases it is possible by taking only those points well remote from the curved portion to arrive a t an approximate value for the end-point (see dotted lines). For instance, figures derived in this manner from experiments on solutions containing 25 milligrams of SO per litre were about 5 per cent. in excess of the correct value. Estimation of Chlorides. The salt chosen for this purpose was a sample ob Merck’s purest fused sodium chloride.A stock solution of this together with the approximately N / 10- and N j 25-silver nitrate solutions used for titration was standardised by duplicate gravimetric estimations as silver chloride BLECTRO-TITRIMETRIC ANALYSES. 59 The results obtained with varying concentration of chlorine are shown in table 111. TABLE 111. Concentration of C1 Strength of C1 found. mg. per litre. AgNO,. Per cent. 1000 N/10 99-7 99.3 200 100.1 99.3 100.8 50 N725 100.2 100.9 10 7 100.6 100-8 5 Y End-point indefinite. As in the case of sulphates a limit of concentration is reached a t wbich the method fails the figure in this instance being 10 milli-FIG. 2. C.C. N/25-,EaCl2. grams cjf chlorine per litre. The addition of precipitated silver chloride before titration did not render the end-point sharp a t lower concentrations.The type of curve obtained in this estimation is similar to that shown in Fig. 1. Estimation of Calcium. A stock solution of calcium chloride was prepared by dissolving pure calcite in hydrochloric acid and eliminating the excess of acid by repeated evaporations on the water-bath. Both this solution and the approximately N J 10-ammonium oxalate solution employe 60 EFFECT OF DILUTION IN ELECTRO-TITRIMETRIC ANALYSES. in the titrations were staridardised by means of potassium pennan-ganate. Table IV shows the results obtained. TABLE IV. Concentration of Ca Ca found. mg. per litrc. Per cent. 500 100.0 100.4 99.2 100.4 200 99.1 100-6 99.1 100 End-point indefinite.As in the case of the determination of chlorides previous addi-tion of the precipitated salt dirt not lead to sharp end-points a t the last-mentioned concentr8tion. I n three such experiments figures given by producing the straight portions of the curves gave 97.0, 97.6 and 95.8 per cent. respectively of the amounts taken. 3 s tinta tioia of Magnesium. A solution of Merck’s purest magnesium sulphate was employed, standardisation being effected by duplicate estimations as mag-nesium pyrophosphate. An appro,ximately N / 10-sodium hydroxide solution standardised by means of sulphuric acid and pure sodium carbonate was used f o r titration. Variation of the concentration of magnesium gave results shown in table V. The typical curves for these cases have no minimum, but exhibit a definite greak.TABLE V. Concentration of Mg Mg. found. mg. per litre. Per cent. 639 99.2 99.7 99.7 269.5 99.7 100.1 100’4 100.1 202 100.8 99.9 100-8 100.6 134.5 End-point indefinite. Addition of magnesium hydroxide prior to titration lead to no improvement a t the last-mentioned concentration. I n connexion with the estimation of magnesium by this method i t is interesting t o note that Harned (Zoc. c i t . ) says “This titra-tion gives only a fairly easily detectable end-point for the change in direction of the plot before and after the end-pint is not great. A reagent must therefore be sought which will increase; the difference in the slopes a t the end-pint.” For this reason he employed barium ’hydroxide t o titrate solutions containing magnesium sul-phate.Apart from the fact that the use of this reagent t o b THE OPTICALbY ACTIVE NEOMETHYLHYDRINDAMINES. 61 effective demands that the magnesium shall be present as sulphate and the b t a l elimination of carbon dioxide from the solution com-parison of Harned’s figures with those recorded above shows that the use of barium hydroxide does not present any definite advan-tage. It appears that errors from other sources are of greater magnitude than that derived from difficulty in reading the inter-section of the two limbs of the curve. Su?n,nnnry of Results. (1) The determination by means of conductivity measurements, of sulphates chlorides calcium and magnesium has been studied a t low concentrations. (2) It has been shown that in relatively weak solutions very small quantities of each of the above1 mentioned may be estimated, withoat any attempt a t temperature control with an error not exceeding + I per cent. (3) In each case a limit of dilution is reached a t which the results cease t o be accurate smooth conductivity curves being obtained. With the exception of the1 case of sulphate estimations, saturation of the solution with the salt to be precipitated does not lead to an improvement in this respect. WELLCOME TROPICAL RESEARCH LABORATORIES, GORDON MEMORIAL COLLEGE, KHARTOUM. [Received August Cith 1918.