26 PYRIDINE HOMOLOGUES APPLICATIONS OF THE ELECTRONIC SPECTRA OF PYRIDINE HOMOLOGUES TO QUANTITATIVE ANALYSIS AND TO THE MEASUREMENT OF DISSOCIATION CONSTANTS BY E. F. G. HERINGTON Received 27th June, 1950 The difficulty of preparing pure samples of pyridine bases is indicated. The absorption spectra of highly purified samples of pyridine, a-, P-. and y-picoline and z : 6-lutidine obtained under different conditions are discussed from the point of view of chemical analysis. The spectra of the vapours over a wide range of pressures a t 20° C using cells of length 14-5 cm., 61.2 cm. and 153 cm. are compared. Special attention has been paid t o the effect of small amounts of impurities. The spectra of solutions of these bases in 0-1 N sulphuric acid, 0-1 N sodium hydroxide and cyclohexane are recorded.A method for the analysis of a mixture of 8-picoline, y-picoline and z : 6-lutidine based on the differences in absorption coefficient of acid and alkaline solutions is described. The dissociation constants of these bases have been determined from the spectra of buffered solutions. Compounds of the pyridine series are particularly difficult to prepare in a state of purity, partly because there are no convenient synthetic methods, partly because certain members have boiling points near othersE. F. G. HERINGTON 27 and partly because they are difficult to free from water. The modern technique of fractional distillation will readily separate at-picoline and pyridine from a mixture of bases which was impossible when much of the earlier work on the spectra was undertaken (e.g.see Herrman l), but certain homologues such as jl-picoline (b.p. 144.0~ C), y-picoline (b.p. 145.3~ C) and z : 6-lutidine (b.p. 144.0' C) have boiling points which are so little different that even now special methods have to be used in their separation. a Extensive fractional distillation fails to remove hydrocarbon impurities from pyridine and a-picoline so that chemical treatment is a necessary step in the purification of these bases.a In view of the difficulty of preparing pure samples, it is not surprising that Herrmanl thought that the bands at 2789-4, 2832-4 and 2876.7A obtained by him in the vapour spectra of a-picoline were due to the presence of pyridine in his sample although Purvis at an earlier date had commented upon the large number of near coincidences in the wave- lengths of bands in pyridine and at-picoline.The present work shows that some of these bands do occur in the spectra of pure at-picoline. A more recent example of the difficulty of obtaining pure samples of these bases is afforded by the investigation of the Raman spectra by Herz, Kahovec and Kohlrausch 6 who showed that their jl-picoline sample contained 2 : 6-lutidine even after extensive treatment. have suggested that the ultra-violet spectra of cc-picoline and pyridine vapour should be compared in order to establish whether a band at 33824 cm.-l observed by them and by Henri and Angenot * in pyridine was due to a small quantity of a-picoline. Highly purified samples of these bases have been prepared at the Chemical Research Laboratory in connection with other work, and in view of the uncertainties concerning published spectra it was decided to record the absorption of these authenticated specimens.Interest at the Chemical Research Laboratory centred on the possibility of devising spectroscopic methods for the analysis of mixtures and no vibrational assignments have been attempted. The ultra-violet absorption spectra of pyridine in the vapour and in solutions have frequently been described, e.g. vapour spectra by Henri et aL8 and by Sponer et aZ.% 7 ; heptane solution by Spiers and Wibaut ; aqueous solution by Marchlewski and Wqtrobek and by Hughes, Jellinek and Ambrose.11 The spectra of a-picoline in the vapour has been reported by Purvis and by Herrnan,l and in solution by PuMs and by Baker and Baly.13 The vapour spectra of 2 : 6-lutidine was de- scribed by Purvis 4 and that of jl-picoline by Herrman.l but in view of the spectra reported here it appears probable that his jl-picoline specimen was composed largely of z : 6-lutidine.A note by Sponer and Rush la on the near ultra-violet spectra of a-, jl- and ypicoline appeared when our preliminary work on the spectra had been carried out. Experimental Sponer et d.'* General.-The purities of the samples of bases used were established by a freezing point technique 14 and the following values found : pyridine 99-85 f Herrman, 2. Wiss. Photog., 1919, 18, 233. Coulson and Jones, J . SOL Chem. Ind., 1946, 65, 169. Report Chem. Res. Board (19481, 33. 4Purvis, J . Chem.SOC., 1910, 97, 692. 5 Herz, Kahovec and Kohlrausch, 2. physik. Chem. B, 1943, 53, 124. 6 Sponer, Rev. Mod. Physics, 1944, 16, 224. 7 Sponer and Stticklcn, J . Chem. Physics, 1946, 14, 101. * Henri and Angenot, J . Chim. Phys., 1936, 33, 641. Spiers and Wibaut, Rec. trav. chim., 1937, 56, 573. lo Marchlewski and Wyrobek, Bull. Acad. Polanaise, 1929, A 93 ; 1934. A 22. llHughcs, Jellinck and Ambrose, J . Physic. Chem., 1949, 53, 410. l2 Baker and Baly, J . Chern. SOC., 1907. 91, 1122. l3 Sponcr and Rush, J . Chem. Physics, 1949, 17, 587. l4 Hcrington and Handley, J . Chem. SOC., 1950, 199.28 PYRIDINE HOMOLOGUES 0.07 yo, a-picoline 99-89 0.06 yo, p-picoline 99-97 f 0.02 yo, y-picolinc. 99-75 f 0.13 yo, z : 6-lutidine 99-93 f 0.04 "/o. The water content was less than 0-02 yo by volume.Sensitive colour reactions l6 gave results consistent with the freezing point determinations, thus, for example, the 7-picoline con- tents of the ,9-picoline and 2 : 6-lutidine samples were shown to be less than 0-1 yo by volume. The vapour spectra were obtained by means of a Zeiss quartz spectrometer, slit width 0.01 mm., reciprocal dispersion 7.4 A per mm. in the wavelength range 2400-3000 A, using Ilford Selochrome 30' backed plates. The light source was a high tension, 2500-V hydrogen lamp consuming I kW and exposure times varied from 4 to 16 min. The vapour absorption cells of lengths 14-5 cm.. 61.2 cm. and 153 cm. were of Pyrex glass with quartz windows and carried a side tube containing the liquid. The cells were evacuated with a mercury diffusion pump backed by a rotary oil pump.The pressure of the vapour of the base in the cell was varied by immersing the liquids in the side tube in suit- able baths. The temperature of the vapour in the light path was that of the room. The wavelengths of the absorption bands were measured on photo- graphic enlargements of the spectra using an iron arc as standard l6 and em- ploying the Hartman formula for interpolation. The wavelengths in air recorded in the tables are expressed in International angstrom and are believed t o be correct t o a t least 0.5 A. The letters given in the tables following the wavelengths designate the appearance of various types of band : thus, a, strong band with sharp edge on violet side ; b, strong band most intense in centre ; c, weak narrow band with sharp edges ; d, diffuse band ; e, centre of very broad band ; f, violet edge of very broad band.Some of the solution spectra were measured on a Beckman quartz spectro- photometer and others on a Unicam quartz spectrophotometer. The 0.1 N sulphuric acid used as solvent had a transmission of 95 % a t 2450 A compared with conductivity water while the value for the 0.1 N sodium hydroxide solu- tion was 97-2 yo. The cyclohexane which was used as a solvent was purified by acid treatment followed by percolation through silica gel l 7 and had a trans- mission of go % at 2450 A compared with conductivity water. The appropriate solvent was employed in the blank cell in every instance. Vapour Spectra.-PYRnxNE.-Henri and Angenot 8 who studied the spectra of this material over a wide range of temperatures and pressures recorded 255 bands.Such a wide range of conditions were not examined here and it was not expected that so many bands would be observed, but nevertheless 109 bands were measured and agreement found with those recorded by Henri in most cases. The only bands recorded by the French workers which were not observed al- though it was anticipated that these bands would be found from the recorded intensities and from the respective resolving powers of the spectrometers were bands at 2925.5, 2932.2, 2923.0, 2goo.0, 2886.3, 2833-1 and 2750.8 A. o-Xylene has been separated from some impure samples of pyridine but it was shown that these bands which we have failed to detect could not have been due t o the presence of this hydrocarbon in the pyridine used by other workers.The only bands found in this work which were additional t o those recorded by Henri and Angenot were weak and diffuse bands a t 2497, 2487, 2466, 2447 and 2427 A. Plate I is a photograph of the spectra. The absence of diffuse bands of the type shown by a-picoline (Plate 11) should be noted. These pyridine spectra exhibited a band a t 33824 cm.-1 (29556 A) which was not due to the presence of picoline (see below). experienced diffi- culty in assigning notably because Kline and Turkevich18 had reported a mediumly strong band in the infra-red a t 936 cm.-1 (mean 941 cm.-l). Measure- ments of the infra-red spectra of a sample of pure pyridine carried out a t the Chemical Research Laboratory failed t o reveal the presence of a mediumly strong band a t 941 cm.-l.a-PICOLINE.-The spectra of this base shows fewer bands than pyridine (Plate 11). Com- parison of the wavelengths of these bands with those recorded by Henri8 for pyridine shows a large number of near coincidences. The general intensitv of This is the band which Sponer and Stiicklen The 34 bands which were observed are listed in Table I. l5 Herington (in preparation). l 6 Iron Charts (Adam Hilger Ltd.), 3rd edn. l 7 Maclean, Jencks and Acree. J . Res. Nut. Bur. Stand., 1945, 34, 271. Kline and Turkevich, J . Chem. Physics, 1944. 12, 300.PLATE I.-Vapour spectra of pyridine at 19O C. A Iron arc. D 14-5 cm. cell, liquid at 14' C. cell, liquid at -15" C. B 153 cm. cell, liquid at 19" C. c 61 cm. cell, liquid at 19' C.F 14-5 cm. E 14-5 cm. cell, liquid at 0" C. G 14-5 cm. cell, liquid at -30" C. A B F G PLATE 11.-Vapour spectra of a-picoline at zoo C. B 153 cm. cell, liquid a t 20" C. E '14.5 cm. cell, liquid at 0" C. G 14.5 cm. cell, liquid at -22" C. A Iron arc. D 14-5 cm. cell, liquid a t 18" C. cell, liquid at -17" C. c 61 cm. cell, liquid'at zoo C. F 14.5 -cm. [To face page 28A B C G PLATE 111.-Vapour spectra of P-picoline at I ~ O C. A 1-53 cm. ccll, liquid at 19" C. cell, liquid at oo C:. -30' C. B 61 cm. cell, liquid at 19' C. c 61 cm. E 61 cm. cell, liquid a t n G I cm. cell, liquid at - - I ~ O C,. F 61 cm. cell, liquid at -45" C. G Iron arc. PLATE 1V.-Vapour spectra of y-picoline at 19" C . A 153 cm. cell, liquid at 19" C . B 61 cm.cell, liquid a t 19" C . c 61 cm. cell. liquid a t oo C. D 61 cm. cell, liquid at -15" C. E Iron arc.PLATE V.-Vapour spectra of 2 : 6-lutidine at 19" C. A Iron arc. B 153 cm. cell, liquid at 1 9 ~ C. c 61 crn. cell, liquid at 19" C. D 14.5 cm. cell, liquid at 15" C. E 14.j cm. cell, liquid at - 5" C .E. F. G. HERINGTON 2826.8 d 2823-3 a 2813.3 d 2805-9 d 29 the a-picoline spectra is less than that of pyridine. region 2663-3-25 16.6 A are particularly striking. spectra of p-picoline vapour has not been studied within recent times. The diffuse bands in the ~-PICOLINE.-EXCept. for the note by Sponer and Rush13 the aksorption TABLE ~.-WAVELENGTHS OF ABSORPTION BANDS IN OL-PICOLINE VAPOUR A l 1 - 2928-8 c 2914'7 c 2905.1 c 2898.7 c 2891.7 c 2887-9 G 2921'3 2 2879-2875.9 d A 2875.9 b 2869.7 b 2866.1 c 2862.1 d 2861.3 d 2858.8 d 2856.0 d 2853.8 c A I A 2850.0 c 2847.6 c 2844.1 c 2836.4 G 2823-3 b 2826-2882 d 2816.5 b 2839.7 c 2806.9 c 2803.8 c 2795'9 c 2789.0 b 2783'5 c 2775'5 d 2747'2 d 2736.9 c A Plate I11 shows the spectra and the wavelengths of 50 bands recorded in Table 11. The tendency for bands to occur in small groups gives to this spectra its characteristic appearance.The bands become more diffuse towards shorter wavelengths but wide diffuse bands of the type shown by a-picoline or by 2 : 6- lutidine are absent. TABLE II.-WAVELENGTHS OF ABSORPTION BANDS IN ~-PICOLINE VAPOUR A 2938'2 C 2934'4 c 2925.6 c 2922-6 c 2918.5 C 2912'5 C 2904.5 c 2896.8 c 2892.3 c 2888.9 c 2884.5 b A 2880.8 b 2872.9 a 2867.0 c 2859-6 a 2856.8 b 2851.1 c 2848.1 c 2840.1 a 2837.1 b 2853'5 c 2845.2 c A 2833.0 C 2828-3 c 2823.0 c 2816.8 b 2813.3 a 2809.1 c 2795'2 d 2784.2 c 2779'2 c 2775'3 c 2762-7 c A A 2635'5 d 2594'9 d 2579'9 d 2450-2550 d 2618.1 d 2606.8 d - yPICoLINE.-The absorption bands in the spectra of this base (see Plate IV) are displaced towards shorter wavelengths compared with those in a- and pTpicoline and are much more diffuse particularly below 2800 A although wide diffuse bands of the type which appear in a-picoline and 2 : 6-lutidine are absent.Table I11 gives the wavelengths of 25 bands measured. TABLE III.-WAVELENGTHS OF ABSORPTION BANDS IN Y-PICOLINE VAPOUR A 2884.1 d 2864.4 c 2851.2 c 2841.9 a 2834.6 c A A A 2748.0 d 2720.6 d 2713.6 d 2694.2 d 2727'5 d A 2 : ~-LuTIDINE.-T~~ spectra of this material exhibits very few bands, The wavelengths of the bands measured were : 2866.7 c, 2824-9 c, 2693.2 f, see Plate V.2702.8 el 2651-8 d , 2627.2 d, range 2500-2620 A weak diffuse bands.3 0 PYRIDINE HOMOLOGUES General Comparison of the Vapour Spectra.-Pyridine exhibits the largest number of bands, and successive substitution of methyl groups in the a-positions causes a progressive suppression of fine bands and an enhancement of diffuse bands. Amongst the mono-methyl substituted bases, the /3- and y-homologues exhibit most fine bands and of the five bases studied only a-picoline and z : 6- lutidine exhibit broad diffuseibands. Vapour Spectra of Binary Mixtures of the Bases .-Experiments were carried out with the object of establishing the smallest percentage of one base that can be detected in the presence of another by means of the spectra of the vapours.Pyridine could be detected in a-picoline down t o a concentration of 3 yo by the appearance of a band a t 2788-582 using the 14-5 cm. cell with the liquid in the side arm a t room temperature. /3-Picoline or 2 : 6-lutidine in y-picoline could be detected down t o a concentration of 5 yo by a change in the position of the ‘ I cut off ” using the 14-5 cm. cell and liquid a t room temperature although no F-picoline or z : 6-lutidine bands could be detected. The presence of 50 yo of z : 6-lutidine by volume in p- and y-picoline does not produce any marked change in the spectra of the picolines so that clearly the vapour spectra do not provide a convenient means for the detection of small amounts of these bases in the presence of homologues.Absorption Spectra of Solutions.-Fig. I, 2 and 3 show the molar extinction E plotted against wavelength expressed in A for solutions of these bases in 0-1 N sulphuric acid, 0-1 N sodium hydroxide and cyclohexane respectively. Certain relationships can easily be seen from these figures. Thus the molar ex- tinction coefficient of these bases in acid solution is approximately twice that in alkaline solution and is least in the hydrocarbon solvent. Hartley l9 lsng ago com- mented on the sharper spectra of an aqueous solution of pyridine hydrochloride compared with pyridine, while Purvis 2o reported that acid caused a marked per- sistence in the spectrum of pyridine and or-picoline.The sequence of the main peaks arranged in the order of decreasing wavelength is approximately z : 6- lutidine, a-picoline, j3-picoline, pyridine and y-picoline in all the three solvents, and the magnitudes of the maximum extinction coefficients follow the .same order. All the spectra studied except z : 6-lutidine show the greatest amount of fine structure in 0.1 N sodium hydroxide solution while the lutidine exhibits most structure in cyclohexane. The shift in the position of the main peak t o longer wavelengths on substitution in the ring in the sequence, pyridine, a- picoline, 2 : 6-lutidine, is similar t o that shown by the series benzene, toluene, dimethyl benzenes, but I-picoline is an exception as i t does not exhibit a shift compared with pyridine.Estimation of @-, y-picoline and 2:6-lutidine in Mixtures .-A method for estimating pyridine based on the measurement of the absorption spectra of acid solutions has been described by Hofmann 21 and by Le Rosen and Wiley.22 Study of Fig. I reveals that 2 : 6-lutidine can readily be estimated in a j3-picoline fraction by measuring the absorption at 2800 A of an acid solution. The values of the differences in extinction coefficient (g./l.) in 0-1 N acid and 0-1 N alkali solutions for these bases are plotted in Fig. 4. A very simple method of analysis of the j3-picoline fraction which takes advantage of the occurrence of isobestic points has been developed based on the data in Fig. 4. Thus if measurements of the absorption are made a t 2416A the difference between the extinction co- efficients of acid and alkaline solutions so found is independent of the /3-picoline concentration while at 2780 8, it is independent of the y-picoline concentration.As this technique of analysis which employs the difference between the ab- sorption of the unknown in acid and alkaline media does not appear t o have been very widely used, a short account of the procedure as applied t o mixtures of the three bases j3-, y-picoline, and 2 : 6-lutidine will be given. Analyses can be made rapidly by this spectroscopic method and t o take full advantage of the speed of this method it was decided to carry out measurements on a volume basis, To obtain the necessary calibration points 0.1 ml. of each pure base was made up t o IOO ml. with distilled water and each of these solutions were further diluted by taking I ml. of the appropriate dilute solution and dilut- ing t o roo ml.with 0-1 N sulphuric acid. The alkaline solutions were prepared similarly except that z ml. of the dilute aqueous solution were made up t o IOO ml. 19 Hartley, J . Chem. SOC., 1885, 47, 685. 2o Purvis, ibid., 1909, 95, 294. 21 Hofmann, Arch. Hyg. Baht., 1942, 128, 169. zp Le Rosen and Wiley, Anal. Chem., 1949, 21, 1175.E. F. G, HERINGTON FIG. I .-Solution spectra, solvent 0.1 N sulphuric acid. Pyridine- - - - - - - - - - a-Picoline , y-Picoline- . .- . . -. . - 2 : 6-Lutidine- - - - p-Picoline .............. , \ A A 2400 2500 zqoo 2 ~ 0 0 >\ FIG. 2.-Solution spectra, solvent 0.1 N sodium hydroxide. Pyridine- - - - - - - - a-Picoline I 8-Picoline ...............Y-Picoline- . . - . . - . .- 2 : 6-Lutidine- - - - ;Q5 40 35 FIG. 3.-Solution spectra, solvent c yclohexane. I d i n e - - - - - - - - - - - m: /3-Picoline .............. , a-Picoline s 7-Picoline- . . - . . -. . - z : 6-Lutidine- - - - FIG. 4. 8-Picoline .............. y-Picoline- . .- . .- . .- z : 6-Lutidine- - - -32 B-picoline y-picoline PYRIDINE HOMOLOGUES 2 : 6-lutidine with 0-1 N sodium hydroxide. The absorptions were then measured a t 2416, 2646 and 2780 A and the observed optical densities were multiplied by the factors IOO and 50 for the acid and alkaline solutions respectively t o calculate the ex- tinction coefficients equivalent t o a base concentration of I ml./l. of solution. The differences between these extinction coefficients for acid and alkaline solu- tions so obtained are recorded in Table IV.To minimize errors it is desirable Optical Density Optical Density Acid Soln. Acid Soln. TABLE ~V.-~AVELENGTHS AND THE DIFFERENCES BETWEEN THE EXTINCTION COEFFICIENTS (CONC. I ml./l.) IN ACID AND ALKALINE SOLUTIONS Difference in Extinction Coefficient (I ml.11.) A 0.140 0.412 0.162 2416 2646 2780 0.103 3'7 0.241 17.1 0.027 13'5 0 25'55 0.55 I 4-2 - 2'2 0 -2-8 27'9 40.0 I I I t o carry out this calibration with each spectrometer employing the same volu- metric apparatus that is t o be used for the estimation of the unknown. In a typical analysis the unknown mixture was diluted t o a concentration equivalent t o 0.01 ml. of base per litre in the acid and alkaline solutions and the optical densities measured and the results shown in Table V obtained.TABLE V A ~~ 2416 2646 2780 The concentration of the three bases in the unknown sample were calculated by successive approximation. Let x, y and z be the volume fractions of ,8- picoline, y-picoline and 2 : 6-lutidine respectively and let xl, yl. z1 and x,, yz. 2% be the first and second approximations t o these quantitives. The first approxim- ation to the 2 : 6-lutidine concentration 2, is given by 13-5/40-0 since a t 2780 A, the y-picoline has a zero difference value and the /3-picoline difference is so small (0.55) that it can be ignored to a first approximation, hence z,'= 0-3375. At 2416 A, 8-picoline has a zero difference so that eqn. (I) applies ( I ) At 2646 A, eqn. (2) applies 3-7 = 14.2 yl - 2-8 z,, hence y1 = 14.2 yl - 2-8 x 0-3375 = 0.3274 17.1 = 25'55 X i + 27'9 21 - 2'2 y1 = 25-55 X i + 27'9 X 0.3375 - 2'2 X 0'3274, - - (2) hence X, = 0.3287. centration a, is calculated by the equation hence 2, = 0'334.Using this value of x1 the second approximation t o the 2 : 6-lutidine con- 13.5 = 40.0 Z2 f 0.55 Xl = 40.0 2% + 0.55 X 0.3287, By substituting 2, in eqn. (I) in the place of z1 the second approximation t o the y-picoline concentration, y,, is found to be 0.327. Substituting z2 = 0.-334, y, = 0.327 in the place of z1 and yl in eqn. (2) yields x, = 0-333. There 1s no necessity t o carry the approximations further and these values-are accepted as the final analysis. Table VI gives the results obtained in the analysis of various synthetic mixtures by this method.The solutions where one component is in large excess were chosen to serve as a fairly exacting test of the method, and as can be seen the results for these solutions appear to be as accurate as for the other mixtures. The wavelengthsE. F. G. HERINGTON r-piwbe -0.6 -1.8 -0.7 -0.6 -2.3 -1.3 - - 33 L2 utidme :. 6: -- f o e ~ +I-I +o-4 3-0-8 - - +0'3 -0'1 2416, 2646, 2780 A were selected as being most suitable for mixtures containing approximately equal proportions of each base but other wavelengths may with advantage be used if certain components are in excess. Thus while the wave- length 2780 A is robably always the best choice for the estimation of 2 : 6- lutidine yet 2620 may be more suitable for the estimation of /3-picoline when 2 : 6-lutidine is in excess, while a wavelength of 2491 A may be the most suitable for the determination of y-picoline under these conditions.TABLE VI.-ANALYSIS O F p-, Y-PICOLINE, 2 : 6-LUTIDINE MIXTURES. ALL PERCENTAGES ARE BY VOLUME Synthetic Mixtures (%) #?-picoline 33'3 60.0 20.0 20'0 90.0 90.0 10.0 10'0 y -picoline 33'3 60.0 20.0 20'0 10'0 90-0 - - 2: 6- Lutidine 33'3 20'0 20'0 60.0 - I 10'0 90.0 Analytical Results (%) 33'3 59'1 19.0 18.7 go' I 9.6 91'3 8.3 32'7 18-22 59'3 19'4 7'7 88.7 - - 33'4 20.4 60.8 21'1 - - 10-3 89.9 Differences (%) Lpicoline 0'0 -0.9 - 1-3 +O.I -0.4 +I'3 - 1-7 - 1'0 Pyridine and a-picoline will interfere with the determination of p- and y- picoline by this method and these bases should be removed by fractional dis- tillation before carrying out an analysis of the 8-picoline fraction.Dissociation Constants from the Absorption Spectra.-Table VII taken from a recent paper by Brown and Barbaras z3 shows some of the values which have been reported for the dissociation constants of these bases. TABLE VII Base Pyridine . a-Picoline . . p-Picoline . . y-Picoline . . K B x 109 I I 2-24 2'4 3'0 10-5 45 32 I1 I1 Ref. - 24 25 26 24 25 26 26 26 I The dissociation constants have now been redetermined by studying the change of absorption spectra with pH.* The appropriate expression for the calculation of the logarithm of the thermodynamic dissociation constant KB from absorption measurements is where el, E ~ , and c3 are the extinction coefficients for the base in acid, buffered and alkaline solution respectively, KW is the ionic product of water and yB+ is 23 Brown and Barbaras, J .Amer. Chem. Soc., 1947, 69, 1137. s4 Barron, J . Biol. Chem., 1937, I21,.313. * Goldschmidt and Salcher, 2. physzk. Chem., 1899, 29, 114. 96 Constram and White, Amer. Chem. J., 1903, 29, 46. * See Flexser, Hammett and Dingwall e7 also Hughes, Jellinek and Ambrose 11 87 Flexser, Hammett and Dingwall, J . Amer. Chem. Soc., 1935, 57, 2103. who obtained a value of 1-32 X I O - ~ for pyridine at 20 f zo C. B34 PYRIDINE HOMOLOGUES the activity coefficient of the base ion, from the approximate Debye-Hackel expression. The value of log yB+ was calculated 1'074 1.065 1-077 1.378 1.406 1.407 1.005 0.985 1.000 1.726 1.710 1'753 3.207 3'330 3-356 where I is the ionic strength, z the valency and A has the following values for water, 0.500 a t 15" C, 0.509 a t 25O C and 0.514 a t 30' C.Solutions were made up by volume and the pH of the solution buffered with a sodium acetate + acetic acid mixture (approximately 0.01 M) was measured with a glass-saturated calomel electrode. The pH meter was standardized with M/zo potassium hydrogen phthalate and sodium borate buffers. The temperature of the buffered solution was determined directly after measuring the extinction coefficient and the appropriate values of A and log,, K , (see Harned and Owen 28) were used in the calculation. In order t o obtain the highest accuracy the pH was chosen so that the value of (eZ - cS)/(el - e2) was near unity. Substitution of methyl groups in this series clearly increases the dissociation constant. The results are given in Table VIII. TABLE VIII.-THERMODYNAMIC DISSOCIATION CONSTANTS OF SOME PYRIDINE HOMOLOGUES 5-16 5-16 5-16 5-84 5'84 5-72 5-72 5-72 5.86 5.86 5.84 5-86 6.06 6-06 6.06 Base Pyridine . a-Picoline /3-Picoline y-Picoline 2 : 6- Lutidine I wish Wave- length A 2506 2609 2600 2650 2710 2547 2610 2660 2558 2485 2589 2519 2646 2720 2780 Temp. of Soh. "C 25'5 25'3 25'3 25-2 25-2 25'2 24-6 24.6 24.6 24'3 24'3 24'3 26.7 26.7 26-7 81 - 82 I pH -- I r 0'010 0'0 I 0 0'0 I0 0.013 0.013 0.0 I 3 0.0125 0.0125 0.0125 0.013 0.013 0.013 0.014 0.014 0.014 KB x 14 1-44 1-43 1-45 8.65 8-83 8-83 4'5 8 4'56 4'49 10.59 10.50 10.76 37'3 38.8 39.1 Mean Thermo- dynamic Constant KB x 10' 1-44 8-77 4'54 10.62 38-4 I thank Dr. E. A. Coulson, 1. B. Ditcham, E. C. Holt, A. Sleven for supplying the pure base samples, R. Handley and A. J. Cook for the estimations of the purity by freezing point determinations and J. L. Hales for the determinations of the water contents. Mr. R. Handley carried out much of the preliminary investigations on the vapour spectra while W. Kynaston made the recorded measurements. The work described has been carried out as part of the research pro- gramme of the Chemical Research Laboratory and this paper is pub- lished by permission of the Director, Chemical Research Laboratory, Teddington, Middlesex. Chemical Research Laboratory, Teddington, Harned and Owen, The Physical Chemistry of Electrolytic Solutions (Rein- Middlesex. hold Publishing Corporation, New York, 1g43), p. 485.