1937
Neutralization of Alkylamines by Hydrogen Halides
Thermodynamics of Neutralization of Alkylamines by Hydrogen Halides in Benzene F. Grauer and A. S. Kertes* Institute of Chemistry, The Hebrew University,Jerusalem, Israel (Received January 2 1, 1976)
Thermodynamic functions, pK, AH, AS, and AG, of neutralization of normal octyl-, dioctyl-, trioctyl-, tridecyl-, and tridodecylamine were determined calorimetrically at 303.15 K in benzene. Hydrogen chloride and bromide served as reference acids. The base strength of the amines in benzene decreases from primary to secondary to tertiary amines, and increases with the length of the alkyl chains in the tertiary amines. Alkylammonium bromides are more stable than the corresponding chlorides. The data suggest that the apparent base strength of the amines is governed primarily by the difference in solvation between the amine base and the corresponding ammonium salt in solution.
The most characteristic reaction of an uncharged alkylamine base is that with an acid to form the salt. The energetics of this neutralization process have received considerable attention in the last 2 decades, when the key part in these studies was played by the factors affecting the basicity of the uncharged aliphatic amine.l The concensus appears to be that the experimentally observed order of base strength of the three amine classes in any given solvent depends on the extent of proton transfer between the two possible extremes, that of mere hydrogen bonding to that of complete protonation. Consequently,the order of base strength is as much a function of the nature and structure of the reference acid (the proton donor), as that of the proton acceptor, the amine class, and the steric factors associated with the length of the alkyl chain and its branching. In water, and to some extent also in ionizing polar organic solvents, the experimentally determined enthalpy changes of neutralization are bound to be affected by the heat effects due to the dissociation of the ion pairs formed and their solvation. Consequently, a correlation between heat of neutralization and the base strength of the alkylamines might be misleading. On the other hand, in low-dielectric constant, nonpolar hydrocarbon solvents, no dissociation of the alkylammonium ion pairs is likely to be significant.2 The differences in the thermodynamic functions of the protonation reaction are thus expected to be more straightforwardly related to the polar effect of the substituents bound to the amine nitrogen. With this rationalization in mind, and in the frame of a long-term project on the thermodynamics of high-molecular weight alkylammonium salts,2 has this study been undertaken. The immediate aim is to obtain reliable thermodynamic data, generated by solution calorimetry, on the possible effect of two steric factors upon basicity: that of the number of alkyl chains (amine class), and their length, when attempting to minimize the number of possible additional factors affecting the equilibrium. This was hoped to be achieved by (i) selecting alkylamines with long enough aliphatic chains so that the difference in the electronegativity of the nitrogen atom? and the dipole moment of the salts formed: are less pronounced; (ii) employing benzene as the hydrocarbon medium, which exhibits some solvation properties both toward the amine bases5 and the ammonium salts,2preventing thus the molecular association of the latter;2and (iii) using hydrogen chloride and bromide, the simplest possible reference acids in terms of their steric interference (F and B strain$).
Experimental Section Materials. n-Octylamine (Fluka), di-n-octylamine (Fluka), tri-n -0ctylamine (Fluka), tri-n -decylamine (Eastman), and tri-n-dodecylamine (Schuchardt), of the highest purity commercially available, were purified by fractional distillation under reduced pressure. The small middle cut collected gave in all cases a single peak when gas chromatographed on a 1.5-m long column of SE (15%)at temperatures between 150 and 300 "C, depending on the amine. The estimated purity was not lower than 99.9%. The densities of the purified products were as follows: n-octylamine dq5 = 0.7787 (0.7790 in the literatureca), di-n-octylamine d:5 = 0.8000 (0.80036a),tri-n-octyiamine dj5 = 0.8086 (0.80886a), tri-n-decylamine di5 = 0.8167 (0.81656a),and tri-n-dodecylamine dz5 = 0.8217 (0.821gca). Benzene (Malinckrodt) was dried over sodium for several days prior to distillation, and after it. The purity, checked by gas chromatography using a 2-m long Apiezon L 10% column at 150 "C, was better than 99.85%. The density was diO= 0.8687 (0.8685in the literaturecc). The water content of the amines and benzene was determined by Karl Fischer titration and found that the purified compounds contained less than 0.01% of water. The gaseous hydrogen chloride and bromide (Matheson) had a purity of better than 99.99 mol %. Titrant solutions were prepared by passing the hydrogen halides through drying columns containing anhydrous magnesium perchlorate and silica gel, and introducing into dry benzene under stirring. The titrants were made up to a concentration of about 0.06-0.12 mol dm-3, and were standardized by diluting aliquots with water and ethanol, and titrating with standard aqueous sodium hydroxide. Frequent standardization was necessary because of the volatility of the hydrogen halides. Apparatus and Procedure. The enthalpies of the neutralization reactions were determined at 303.150 f 0.001 K using a Tronac Model lOOOA continuous automatic titration calorimeter. The equipment, its calibration, and the general procedure employed for the determination of heat capacities have been d e ~ c r i b e d . ~ For the determination of the heat of reaction in a typical run, 50 ml of freshly prepared -0.005 mol dm-3 benzene solution of the amine is introduced into the reaction vessel of similar capacity, the free vapor space above the solution being thus at a minimum. When thermal equilibrium is reached, usually after 3 h, the titrant, a -0.1 mol dm-3 benzene solution of the hydrogen halide, is delivered from the burette at a constant rate of 0.8220 ml/min. A t least 20, but usually as The Journal of Physical Chemistry, Vol. SO, No. 17, 1976
1938
F. Grauer and A. S. Kertes
TABLE I: Thermodynamic'Functionsof Neutralization of Alkylamines by Hydrogen Halides in Benzene (303.15 K)
Concn, lo2mol dm-I System
Amine
(CsH17)H2N-HCl 0.447 (C&I17)2HN-HCl 0.777 ( C E H I ~ ) ~ N - H C ~ 0.268 ( C I O H ~ ~ ~ N - H C I 0.350 ( C I Z H ~ ~ ~ N - H C ~0.431 (CsH17)H2N-HBr 0.487 (CsH17)ZHN-HBr 0.487 (Ce"7)3N-HBr 0.487 (CloH21)3N-HBr 0.487 (C12H&N-HBr 0.487
-AH,
Acid
PKa
6.53
6.75 f 0.20 6.26 f 0.21 5.51 f 0.18 5.61 f 0.22 5.66 f 0.21 7.37 f 0.25 6.74 f 0.17 6.22 f 0.20 6.55 f 0.15 6.79 f 0.18
12.88
4.21 7.97 10.50 6.13 6.84 6.21 6.59 6.59
kJ mol-1 126.5 f 0.8 117.7 f 0.8 93.1 f 0.8 98.0 f 0.6 102.0 f 0.6 134.1 f 1.2 125.1 f 1.1 106.4 f 0.8 113.2 f 1.3 117.6 f 1.4
- AG,
-AS,
-m,a
kJ mol-I
J mol-l deg-l
kJ mol-1
39.2 36.3 32.0 32.6 32.8 42.8 39.1 36.1 38.0 39.4
288.0 268.5 201.8 215.9 228.2 301.4 283.6 231.9 248.3 257.9
127.8 119.8 91.6 100.9 107.3 136.5 126.8 107.3 115.9 120.8
Calculated by Wadso's compensation pro~edure.~ much as 42, data points are automatically recorded at 10-s intervals. Every titration was repeated at least three times, the reproducibility always being better than 1%,being expressed in terms of uncertainties in the thermodynamic values given in Table I. The standard experimental procedure calls for the correction of the total heat effect for the nonchemical heat effect and those due to heats of dilution of the titrant and the titrate. Heats of dilution were determined in separate experiments, identical with those of neutralization except for one of the solutes absent at a time, and every point corrected accordingly. An explicit set of experiments revealed that some vaporized hydrogen halide diffused from the titrant during the period of thermal equilibration. Since repeated attempts to eliminate the phenomenon were not fully successful, corrections had to be introduced for the loss of the escaped acid. Based on the analytical determination of the amount of vaporized acid, the corrections involved recalibration and subsequent standardization of the initially introduced volumes of the titrant by setting the starting point of the thermogram to the differential concentration change of the extrapolated temperature corresponding to the time 281'0.7 This extrapolation procedure, estimated to introduce an error of less than 0.3% of the heat effect, was sufficient to account for the small amount of the amine neutralized by the diffused hydrogen halide. The complete array of the numerical data involved in all corrections is recorded elsewhere in detail.8 Calculation. The general method to calculate the equilibrium constant and the enthalpy change of the neutralization reaction from thermometric titration data using an iterative least-squares analysis has been described in a series of publications? In the nonpolar and nonionizing hydrocarbon medium, and at low total solute concentration employed in this study, the heat produced in the calorimeter due to the neutralization reaction A
+ B e AB
K , = [AB][A]-l[B]-l
is given by Qcorr,p =
(2)
for any given point p , where the number of moles of the salt AB, A n = [AB]V, [AB] being the concentration of the product in mol dm-3, and Vis the volume of the solution in dm3, and AH the corresponding enthalpy change. Since the ratio of protonated to unprotonated amine varies from point to point along the thermogram, the species distribution, expressed by The Journal of Physical Chemistry, Vol. SO, No. 17, 1976
mass-action law equilibria through the constant K,, can be used to calculate the AH values. Assuming that AH is constant, in the dilute Concentration range employed, the values of the two unknowns, K , and A H , which best describe the experimental titration curve are found by least-squares analysis of the error squares sum
U(Ka,AH) =
5
p= 1
(Qcarr
- HAnp)'
(3)
where q is the number of experimental points along the thermogram. An additional and independent calculation of the AH values has been made by the comparison of the heat effect measured during the reaction with that observed when a known amount of heat is evolved in the same calorimetric system by supplied electrical heat, developed by Wadso.9 Results The calculated neutralization constants, K,, and the corresponding thermodynamic functions, including the entropy and free energy changes derived simultaneously by the computer program used, are summarized in Table I, with the uncertainties expressed as twice the mean deviation. The AH values calculated by Wadsokg compensation procedure are given in the last column. These values are generally higher than those calculated in conjunction with the equilibrium constants, the difference between the two sets of AH values being about twice the corresponding standard deviation. Some comments concerning the accuracy of the numerical data compiled in Table I are in order. Careful analysis of calorimetric titration data in generaVbvcindicate that there is a definite relationship between the accuracy (and precision) of the AH values and their absolute magnitude: the larger the enthalpy value the less it is affected by the value of the equilibrium constant. When K , is of the order of lo4or higher, the corresponding AH becomes almost completely independent of K,. Under the present experimental conditions, and the magnitude of the K , and AH values listed, calculations show7b that the enthalpies should be accurate to about f0.12 kJ molm1. We are somewhat less comfortable with the accuracy of the equilibrium constants derived because of the way they have been computed. Equation 1 has not been taken as a strictly thermodynamic definition since molarities were used rather than activities. In other words, the assumption was made that the ratio of the activity coefficients of all species remains constant in the concentration range studied. Based on frag-
Neutralization of Alkylamines by Hydrogen Halides mentary information available, it is believed that the assumption is justified. Activity coefficients of tri-n-dodecylammonium chloride and bromide in benzene a t 25 and 37 "C have been determined previously by differential vaporpressure osmometry.2a Extrapolation of these data to the concentrations of the solutes in this study (-4.5 X mol dm-3) shows that their activity coefficients vary between unity and about 0.985. In the same concentration range, the parent alkylamine base (employed as reference standard for the calibration of the osmometer2a),exhibits an ideal behavior.2aJ0 No similar data are available for the rest of the systems reported here, but it seems reasonable that the ratio of the activity coefficients, YAB/YAYB, in other systems will not vary by more than about 1.5%. Such a close-to-the-ideal behavior of the systems can be explained by the absence of dissociation of the alkylammonium salts into ions, and also their negligible molecular aggregation into dimers a t the present low solute concentrations. Indeed, in the low dielectric constant and zero polarity of the hydrocarbon medium employed, no ionization of the ion pairs is to be expected according to the theory.ll On the other hand, previously determined dimerization constants2* indicate that the tri-n -dodecylammonium bromide-benzene system at 25 "C, the fraction of the dimerized salt does not exceed 0.02 under the experimental conditions of the present study. The enthalpy of dimerization of this and similar alkylammonium salts has been found2aJ2 to be about 15 k J mol-1, which is lower by a factor of 7 or more than the heat of neutralization. This is enough, however, to increase the uncertainties in the AH values listed in Table I by about 0.3%. Finally, a comparison should be made with the calorimetrically determined enthalpies of neutralization of tri-noctylamine with hydrogen chloride and bromide in benzene at 25 "C of 148.64 and 169.25 kJ mol-l reported by Frolov et al.,13 the only relevant data located in the literature. Except for their ten times higher solute concentration range, and the small temperature difference, the data should be directly comparable to the present ones. Since no experimental details were recorded, it is rather difficult to assess the reliability of their data. However, their significantly higher enthalpies are unlikely to be due solely to the higher solute concentrations employed. Discussion The basicity of alkylamines in the gaseous phase appears to be governed solely by the inductive effect of the aliphatic chains attached to the nitrogen.14 The base strength decreases from primary to secondary to tertiary amines, and increases, though to a much lesser extent, with increasing number of carbon atoms in the chain. As to the alkylammonium salts, simple electrostatic considerations require that the bond energy increase with decreasing distance between the charges of the ions constituting the ion pair.ll Thus, due to more favorable steric relationships, the distance between the ions increases in the order primary < secondary < tertiary, which is then also the order of the energy of formation of the salts. Hydration (solvation) and steric considerations, of both the amine and the ammonium salt, have been invoked to explain the order of basicity of the alkylamines in so1ution.l The anomalous order of primary < tertiary < secondary observed in water has been attributed entirely to hydration effects.lb However, more recent calorimetric datal5 on all three classes of amines using hydrochloric acid as the reference acid in
1939
water reveal that the order of base strength is very much temperature dependent. On the other hand, the order in acetonitrilele is determined by steric requirements when hindrance increases with the number and length of alkyl chains, with little contribution from solvation. In chlorobenzenelc and nitromethanelf solutions, inductive effects have been invoked to explain the order primary < secondary < tertiary. The above generalizations concerning the order of basicity of aliphatic amines were based on either AH or pK values of neutraliFation, but not on both. To the best of our knowledge, this is the first report in which a complete set of thermodynamic functions has been determined for the process of neutralization of long-chain alkylamines in hydrocarbon solvents. Prior to analyzing the numerical data, it should be helpful to define the state of the two solutes, the alkylamine base and the alkylammonium salt, in benzene solution. Benzene Solutions of Alkylamine Bases. As mentioned before, normal tertiary amines are nonassociated liquids, exhibiting an ideal behavior in a variety of nonpolar and low-dielectric constant hydrocarbons, including benzene.2J0 In the same solvents, primary and secondary alkylamines are known16 to associate through a system of intermolecular hydrogen bonds. The extent of association is more pronounced with the primary amines having two hydrogen molecules available for bonding. However, as spectral data suggest, the tendency toward association of both classes of amines decreases with increasing chain length, becoming insignificant a t high dilution in hydrocarbon solvents. Recent calorimetric data5 on enthalpies of mixing of symmetrical n-tertiary alkylamines with benzene a t 30 "C have shown that there is an n-ir interaction between the lone-pair electrons of the nitrogen atom and the a electrons of the benzene ring. The excess enthalpies of mixing are endothermic, increasingly so with increasing chain length of the alkyl groups. Invoking a statistical thermodynamic model the trend has been explained5 in terms of the effect of the hydrocarbon chain on the screening of the lone-pair nitrogen electrons from the sight of benzene's a electrons. Similar, though less extensive, enthalpies of mixing data on primary and secondary amines1' are, a t least qualitatively, in line with the trend observed for the tertiary amines. In addition, these data suggest that primary and secondary amines interact with benzene more than tertiary amines, since the screening is bound to be less effective by fewer, one or two, aliphatic chains. We rationalize by saying that in benzene solutions and in the concentration range employed in this study (-0.005 mol kg-l), primary, secondary, and tertiary alkylamines are nonassociated. They are all solvated by benzene molecules, though to a different extent. Due to steric hindrance toward n-a solvation, the extent of solute-solvent interaction decreases both with the number of alkyl substitution on the nitrogen, and the number of carbon atoms in the normal alkyl chain. Benzene Solutions of Alkylammonium Salts. High-molecular weight alkylammonium salts, due to their amphiphilic character, undergo aggregation in both aqueous and organic s o l ~ e n t s . In ~ J hydrocarbon ~ media the molecular association is essentially due to electrostatic interactions of which the dipole-dipole attraction is by far the most important. Specific hydrogen bonds of the type N+-H-X- have also been found to play a contributing role in the case of primary and secondary alkylammonium salts.18 These forces, thus the formation of aggregates and their size, depend primarily on the dipole moment of the salt (which in turn is a function of the size of The Journal of Physical Chemistry, Vol. 80, No. 17. 1976
1940
the ions and the overall symmetry of the molecule), its concentration, and the nature of the solvent. Quantitative information on these factors is almost entirely restricted to tertiary alkylammonium compounds.2Js The dipole moment of trialkylammonium halides is practically constant, 8.50 D for the chlorides and 9.15 D for the bromides4 (see ref 2b), and independent of the chain length C S - C ~ ~The . average aggregation number of these salts increases with concentration in b e n ~ e n e , the ~ ~ bromides ,~~ having slightly higher R values than the chlorides. While no comparable numerical data are available on the fi values of n-octyl- and di-n-octylammonium halides in benzene, it appears that the secondary salts exhibit generally a more pronounced tendency toward aggregation than either primary or tertiary ammonium salts under comparable experimental conditions.lg The solvent, its physical and chemical properties, governs the competition between solute-solvent (solvation) and solute-solute (aggregation) interactions to a considerable extent. Ionic dissociation of these salts is unmeasurably small in benzene even a t increased temperatures.lsS20 Benzene is not an ideally inert solvent toward the high-dipole moment alkylammonium salts. The pronounced solvation capacity is reflected by an overall lower aggregation number of the salts in benzene when compared to that in aliphatic hydrocarbons.18 It is thus fair to assume that at the low salt concentrations employed in this study, the process of aggregation has no misleadingly large effect upon the thermodynamic functions of neutralization determined here. Base Strength of Alkylamines. Turning now to the analysis of the thermodynamic functions listed in Table I, we propose to discuss the effect of (i) the length of the alkyl chains, (ii) the amine class, and (iii) the reference acid in terms of factors affecting the order of basicity of alkylamines in benzene solutions. (i) Enthalpy and entropy changes of the process of neutralization of the members of the homologous series of tertiary alkylamines appear to be considerably more pronounced than those of the corresponding pK, values. Both AH and A S become increasingly negative as the chain length increases from octyl to decyl to dodecyl. The pK, values increase in the same order, though only slightly. The base strength thus increases with the number of carbon atoms in the normal chains. The much smaller differences in the pK, values as compared to those of the corresponding AH values suggest that the inductive effect of chain length on the basicity of the tertiary alkylamines must be relatively small. Similar were the observations for the homologous series of normal primary amines in benzene.lg Changes in the entropy can be explained by the differences in the extent of solvation of the amines. As discussed above, the amine with the longer chains is less stabilized by solvation with benzene. The extent of solvation of the corresponding alkylammonium salts is probably not significantly different (same dipole moment), thus the overall loss of rotational entropy must be due primarily to the differences in the solvation of the amine bases. (ii) The stability of the salts formed, expressed through the pK, values, both chlorides and bromides, decreases from primary to secondary to tertiary octylammonium. There is more than one order of magnitude difference in the pK, values between the primary a n d t h e tertiary salts, and about 20% difference in the corresponding free energy changes. The heat of neutralization decreases in the same order, roughly to the Same extent, but the entropy decreases by more than 40%, with the main decrease occurring in going from the secondary The Journal of Pbysical Cbemistry, Vol. 80, No. 17, 1976
F. Grauer and A. S.Kertes
to the tertiary ammonium salt. The basicity of n-octylamines in benzene increases in the order tertiary < secondary < primary. Since the A S values change more abruptly than the corresponding AH values, becoming more negative as the stability of the salts formed increases, it is apparent that the change of the entropy rather than the enthalpy is the primary factor contributing to the free energy changes. This is then a further indication of the important role which the solvation of the amines plays in determining their basic strength. (iii) The last comment concerns the effect of the reference acid on the thermodynamic functions of neutralization. In all systems investigated the alkylammonium bromides are more stable than the corresponding chlorides. From purely electrostatic considerations of bond energies between the ions in the salt, the opposite order could have been expected: the larger the anion the weaker the electrostatic interaction, thus the pK, of the salt. Obviously, the size effect is overcompensated by solvation effects of the two types of halides. It is perhaps interesting to note that the differences in solvation between the chlorides and bromides are much less pronounced with the primary and secondary ammonium than with the tertiary ammonium cation. The difference in the AG values between the chlorides and bromides of the primary and secondary salts is about 8%,as compared to 16% with the three tertiary alkylammonium compounds. Very similar are the enthalpy and entropy change relations: about 6% between C1and Br- for the octyl- and dioctylammonium radicals, and 13%with the trialkylammonium halides. Conclusions In concluding this discussion, we rationalize that the differentiation between steric, inductive, and solvation effects in their influence on the order of base strength of alkylamines in nonpolar hydrocarbon media might not be as straightforward as previously suggested on the basis of more limited thermodynamic data.l While the various effects cannot be easily resolved on the basis of thermodynamic evidence alone, the present set of data suggests that solvation in benzene, perhaps because of its T electrons, is considerably more important than either steric or inductive effects. The thermodynamic functions of neutralization are governed primarily by the difference in solvation between the amine base and the ammonium salt, which determines the base strength of the alkylamines.21 Acknowledgment. The authors are indebted to Drs. G. Y. Markovits and 0. Levy for helpful discussions. References and Notes (1) (a) R. Spitzer and K. S. Pitzer, J. Am. Cbem. SOC.,70, 1261 (1948); (b) A. F. Trotman-Dickenson, J. Cbem. SOC.,1293 (1949); (c) R. P. Bell and J. W. Bayies, ibid., 1518 (1952); J. W. Bayles and A. Chetwyn, ibid., 2328 (1958); (d) H. K. Hall, J. Am. Cbem. Soc., 79,5441 (1957): (e) E. J. Forman and D. N. Hume, J. Phys. Cbem., 63, 1949 (1959); (f) T. E. Mead, ibid., 66, 2149 (1962); (g)A. Rieure, M. Pumeau, and B. Tremillon, Bull. SOC.Cbim. Fr., 1053 (1964); (h) E.M. Arnett and E. J. Mitchell, J. Am. Cbem. SOC.,93, 4052 (1971). (2) (a) A. S. Kertes and G. Markovits, J. Phys. Cbem., 72, 4202 (1968); (b) A. S. Kertes, 0. Levy, and G. Markovits, J. Pbys. Cbem., 74,3568 (1970); (c) 0. Levy, G. Markovits, and A. S. Kertes, ibid., 75, 542 (1971): (d) J. David-Auslaender, H. Gutmann. A. S. Kertes, and M. Zangen, J. Solution Cbem., 3, 251 (1974); (e) A. S. Kertes, J. Iflofg. Nucl. Chem., 27, 209 (1965). (3) M. Tamres, S. Searles, E. M. Leighly, and D. W. Mohrman, J. Am. Chem. Soc., 76, 3983 (1954). (4) A. V. Ochkin and R. A. Zagorets, Russ. J. Pbys. Chem., 46, 1656 (1972). (5) A. S. Kertes and F. Grauer, J. Phys. Cbem., 77,3107 (1973). (6) (a) R. R. Dreisbach, Adv. Cbem. Ser., No. 29, Vol. 3 (1961); (b) "Dictionary
1941
Air Diffusion and Interface Aging of WaterIAir Systems of Organic Compounds", Vol. 5,Oxford University Press, New York, N.Y., 1965;(c) J. A. Riddick and W. B. Bunger, "Organic Solvents", Wiley-lnterscience, New York, N.Y., 1970. (7)(a) A. S.Kertes and E. F. Kassierer, inorg. Chem., 11, 2108 (1972);(b) J. J. Christensen, D. P. Wrathai, J. 0. Oscarson, and R. M. izatt, Anal. Chem., 40, 175,1713 (1968);(c) P.W. Carr. Thermochim. Acta, 3, 427 (1972); Crit, Rev. Anal. Chem., 2,491 (1972). (8)F. Grauer, Ph.D. Thesis, The Hebrew University, Jerusalem, 1975. (9)i. Wadso, Sci. Tools, 13, 33 (1966). (IO) C. Kiofutar, S. Paljk, and D. Kremser, J. Inorg. Nuci. Chem., 37, 1729
(1975). (11) R. M. Fuoss and F. Accascina, "Electrolytic Conductance", Interscience, New York, N.Y., 1959;C. W. Davies, "Ion Association", Butterworths, London, 1962. (12)K. H. Stern and E. A. Richardson, J. Phys. Chem., 64, 1901 (1960);G. S. Denisov, H. P. Oya, E. V. Ryachev, and D. N. Suglobov, Teor. Eksp. Khim.,
5,254 (1969). (13)Yu. G. Frolov. T. N. Vinetskaya, and V. V. Sergievskii, Radiokhimiya, 12, 753 (1970). (14)W. G. Henderson, M. Taagepera, D. Hoitz, R. T. Mclver, J. L. Beauchamp, D. H. Aue, H. M. Webb, and R. W. Taft, J. Am. Chem. SOC.,94,4728(1972);
and M. T. Bowers, ibid., 94,4726 (1972). (15)S.Bergstrom and G. Olofsson, J. Solution Chem., 7,507 (1975). (16)H. Wolff and D. Starschewski, 2.Eiektrochem., 65,840 (1961);H. Wolff and R. Wurtz, Berichte, 72, 101 (1968). (17)T. M. Letcherand J. W. Bayies, J. Chem. Eng. Data, 16, 266(1971);J. S. Afr. Chem. inst., 25,53 (1972);R. Siedler, L. Grote, E. Kauer, U. Werner, and J.-H. Bittrich, Z.Phys. Chem., 241,203 (1969). (18)A. S.Kertes and H. Gutmann in "Surface and Colloid Science", Vol. 8,E. Matijevic, Ed., Interscience, New York, N.Y., pp 193-295. (19)H. Gutmann and A. S. Kertes, J. Coiioidinterface Sci., 51,406(1975);A. S.Kertes, H. Gutmann, 0. Levy, and G. Markovits, in "Chemie, physikalische Chemie und Anwendungstechnik der grenzflachenaktiven Stoffe", Carl Hanser Verlag, Munich, 1973,pp 1023-1033. (20)H. Gutmann and A. S.Kertes. isr. J. Chem., 6, 947 (1970),and references therein. (21) in a preliminary report by the authors (Proceedings of the international Solvent Extraction Conference, 1974,Voi. 2,Society of Chemical industry, London, 1974,p 1441)the pKa values for the three trialkylammonium chlorides given were miscalculated due to a flaw in the computer program used at that time. The conciusions reached there are thus obviously untenable in view of the present data.
Effect on the Water/Air Surface Tension of Air Diffusion and Interface Structuring HOctor Sobol," Javier Garfias, and Jaime Keller Facuitad de Ouhica, Universidad Nacionai Autonoma de Mexico, Mbxico 20, D.F. Mexico (Received November 2 1, 1972: Revised Manuscript Received February 17, 1976) Publication costs assisted by Division de €studios Superiores, Facuitad de Quimica, U.N.A.M.
Experimental values of the water/air surface tension were obtained by the "pendant bubble method" and showed a large time dependence. This effect is related to the amount of dissolved air initially present in the water and to the change in the 0 2 / N ~ratio of the air bubble. Measurements of surface tension vs. time (0.25, 90 min) a t 25 "C and 583 mmHg were made on water/air systems, under different aeration and deaeration conditions of the water phase. While the initial surface tension was always within a fraction of 1%of the generally reported values, i t was found that after a few minutes the surface tension increased or decreased in reproducible patterns. We believe this should be expected because the air content of the surface region changes with time. The maximum increase (aerated water/air system) is of the order of 2% and the largest decrease (deaerated water/air system) is of the order of 15%.These large changes are thought to be related to a pronounced influence of the presence of polarizable molecules (02,N2) in the local molecular organization of the interfacial region of the water/air system. T o estimate the order of magnitude of the energy changes involved, a dynamic model is proposed based on experimental evidence and preliminary considerations of the factors involved. The phenomena described in this paper will be more difficult to observe, if at all possible, with other surface tension measuring methods; our experiments refer to an interface created in a two-phase system which is not in initial thermodynamic equilibrium. This implies that in some cases final equilibrium can never be reached, because the bubble disappears absorbed in the water or will detach itself because of volume growth. The high accuracy required to measure the profile of the bubble was attained by us with improved photographic techniques using a monochromatic parallel beam of h 6943 A pulsating ruby laser light source. Changes in the bubble's 02/N2 ratio of the air phase (deaerated water/air system) due to aging were observed by means of a gas chromatographic technique.
I. Introduction
A review of current literature concerning the structures of liquid water systems, both in the bulk and in the interfacial region,l-I2 reveals that many different types of association of water molecules have been related to the properties of these systems. The present experimental work on the structure and structuring processes of the water/air interface has again made evident that the various types of water molecular "organiza-
tion" that can be present in the bulk and in the interfacial region affect the properties of the latter to a large extent, and are responsible for high sensitivity of the surface tension toward certain foreign molecules. For example, the initial air content of the liquid water phase has an influence on the surface aging, observed as changes in surface tension as large as 15%of the initial value with characteristic times of the order of 103 s. It is worth remembering that the air content of the bulk water and its effect on the structuring of groups of water The Journal of Physical Chemistry, Voi. 80, No. 17, 1976
