INTERMOLECULAR HYDROGEN BONDING phate diester anions. The comparisons between our experimental results and the results of both sets of calculations are not as direct as we should like. The theoretical calculations of Collin and of Boyd were carried out with model compounds which are not the same as any in our series but several in each set bear a strong resemblance to ones in our present set. We conclude from our experimental results that Boyd’s
3313 set of calculations6 is more valid than the other with the regard t o the net charge on the phosphoryl oxygen. We are more satisfied that our previous explanation* of the inversion of the values of K L for the methyl- and ethylpiperidinium cations in the case of the oxygen-containing ligands is a correct one in view of the excellent correlation of the values of the ratios KL(Et)/KL(Me) and VPO.
Intermolecular Hydrogen Bonding. I. Effects on the Physical Properties of Tetramethylurea-Water Mixtures1 by K. R. Lindfors,* S . H. Opperman, M. E. Glover, and J. D. Seese Department of Chemistry, Central Michigan University, Mt. Pleasant, Michigan
48868
(Received March $1, 1971)
Publication costs assisted by Central Michigan University
Vapor pressures, viscosities, densities, surface tensions, heat of mixing, and molar refractivities of watertetramethylurea mixtures were measured at temperatures from 25 to 85’. The data show that the tri- to pentahydrates are the most stable water-tetramethylurea species formed.
Introduction Recently there has been considerable interest in the interactions in hydrogen bonded binary mixtures. 2 , 3 The powerful interactions which occur between a protic and a dipolar-aprotic liquid produce marked changes in many chemical3band physical4properties. Tetramethylurea (TMU) and water form such a binary pair which has not been extensively studied. We report here studies on the TMU-water system using classical physical-chemical techniques. The effects of the TMU-water hydrogen bonds on density, viscosity, surface tension, vapor pressure, refractive index, and heat of mixing of these solutions were measured. The main purpose of the investigation was to determine the most prominent TMU-water complex species present in solution at various temperatures. When TMU and water are mixed, some of the hydrogen bonds between the water molecules are broken. New hydrogen bonds form between the water molecules and the TAW molecules. Thus the deviations in properties of the TMU-water mixtures from a linear interpolation between the properties of the pure components result not only from the formation of TMUwater complexes but also from the disruption of the water structure. However, large deviations in physical properties are strong evidence for the existence of interactions. The presence of a hydrogen bond donor,
the water, and of possible acceptors, the oxygen and nitrogens of the TMU, suggest that at least a major part of these interactions will be hydrogen-bond in nature. The chemical and physical properties of TMU have been reviewed in considerable detail previo~sly.~It is a clear, polar, aprotic liquid which is miscible in all proportions with water and all common organic solvents. TMU has been suggested as a useful reaction medium. I n its solvent properties, it resembles pyridine and dimethylformamide except that it has a higher boiling point (176.5’ (760 Torr)).6 It is an important medium for aryl deaminations’ and is the solvent of choice for higher temperature Ullmann reactions.$ TMU also increases the rate of alkylation reaction^.^ (1) Work supported in part by The Ott Chemical Co., Muskegon, Mich. (2) A. K . Covington and P. Jones, Ed., “Hydrogen-Bonded Solvent Systems,” Proceedings of a Symposium on Equilibrium and Reaction Kinetics in Hydrogen-Bonded Solvent Systems, University of Newcastle upon Tyne, Jan 10-12, 1968, Taylor and Francis Ltd., London, 1968. (3) (a) J. F. Coetzee and C. D. Ritchie, “Solute-Solvent Interactions,” Marcel Dekker, New York, N. Y., 1969; (b) A. J. Parker, Chem. Rev., 69, 1 (1969). (4) G. C. Pimentel and A. L. McClellan, “The Hydrogen Bond,” W. H. Freeman and Co., San Francisco, Calif., 1960. (6) A. Lllttringhaua and A. W. Dirksen, Angew. Chem. Int. Ed. Engl., 3, 260 (1964). (6) W. Mischler and C. Escherich, Chem. Ber., 12, 1162 (1879). (7) K. G. Rutherford and W. Redmond, Org. Syn., 43, 12 (1963). The Journal of Physical Chemistry, Vol. 76, No. 21, 1971
3314
K. R.
LINDFORS,
S. H. OPPERMAN, M. E. GLOVER,AND J. D. SEESE
Experimental Section The TAIU was obtained from the Ott Chemical Co., Muskegon, Mich. It was purified by the following procedure. Ten per cent by volume of dry benzene was added to the T N U . After stirring, the benzene was distilled off at atmospheric pressure. The dissolved water and dimethylamine codistill with the benzene. The remainder was vacuum distilled. The fraction boiling at 62-64" at 13 Torr was collected and stored in glass-stoppered flasks in the dark. The formation of hydrogen-bonded complexes can be inferred by a number of different experimental metho d ~ .We ~ report here the measurements of several intensive physical properties at temperatures from 25 to 80" made on TMU-water mixtures ranging in mole fraction from 0 to 1. Systematic deviations of experimental values from calculated ones are evidence of association. Vapor pressures of TJIU-water mixtures were measured in a greaseless vacuum apparatus similar in design 7 ~ entire to the one described by Christian, et ~ 1 . ~ The apparatus was immersed in a constant temperature water bath at 45.00 f 0.02" and the pressures were measured manometrically. Samples xere introduced through the mercury seal, using a syringe with a blunted needle. The amounts added were determined by weight. Densities mere measured pycnometrically at 25, 45, and 80" in baths regulated to f0.02". The pycnometers were calibrated with water and the calibrations were checked several times at each temperature.1° Viscosities were determined using a standard Ostwald viscometer in the same regulated baths as above. Water was used to calibrate the viscometer at the various temperatures.l 1 Surface tensions were measured using the capillary rise technique. Measurements were done in the above baths. The liquid levels were read with a cathetometer. Water was used for the calibrations.12 The refractive index measurements of the solutions were made with an Abbe refractometer through which water from the 25" bath was circulated. The heat of mixing determinations were made in a calorimeter of our own design. It consisted of a small Dewar flask fitted with a stirrer, heater, and thermistor. The thermistor made up the fourth leg of a Wheatstone bridge. The bridge balance mas monitored on a variable sensitivity recorder. As can be shown, for small temperature changes, the signal to the recorder is directly proportional to the temperature change from the balance condition. A weighed amount of TRIU or water was placed in the Dewar and the bridge was balanced. Weighed portions of the other ingredient, adjusted to the same temperature, were then added. The change in bridge output was noted. After each addition, a known amount of electrical energy was added and the pen shift was again noted. In this way the The Journal of Physical Chemistry, Vol. 76,JVo. 21, 1971
.I
0
.2
.4
.3
.5
.6
MSLE FRACTION
-
.7
.B
.P
l.C
TMU
Figure 1. Vapor pressures of TMU-water mixtures at 45'.
heat of mixing was determined over the entire concentration range.
Results and Discussion Figure 1 shows a plot of the observed vapor pressures of water-TJIU mixtures at 45". As can be seen, the deviations from Raoult's law are small but always positive. Apparently the change from water-water hydrogen bonding to water-TAIU association produces little change in the vapor pressures of such mixtures. However, apparent agreement with Raoult's law is not necessarily indicative of ideality. Many systems are known where a cancellation of effects leads to apparent ideality whereas association is known to occur.13 The pyridine-ethanol system is an e~amp1e.l~ The densities of TMU-mater mixtures show marked deviations from additivity at the three temperatures studied, 25, 45, and 80". At the two lower temperatures, the densities go through a maximum near mole fraction of 0.2 indicating a TMU. (HzO), complex where n is between 3 and 6. At 80" the densities no longer show this maximum and in fact show a negative deviation from linearity. The densities for pure TMU were 0.9622,0.9458, and 0.9131 g/ml at 25, 45, and 80", respectively. Figure 2 is a plot of Ad us. XTwhere Ad = d(expt1) - d(ca1cd) and XT is the mole fraction of TMU. dcalod
where
dTO
= XTdTO
+ Xwdw'
and O d, are the densities of pure TRIU and
(8) S. D. Christian, H. E. Affsprung, and C. Lin, J. Chem. Educ.7 40, 323 (1963).
(9) A. A. Taha, R. D. Grigsby, J. R. Johnson, S. D. Christian, and H. E. Affsprung, ibid.,43, 432 (1966). (10) R. C . TVeast, Ed., "Handbook of Chemistry and Physics," 49th ed, The Chemical Rubber Co., Cleveland, Ohio, 1969, p FS. (11) See ref 10, p F45. (12) See ref 10, p F30. (13) G. C. Pimentel and A. L. McClellan, "The Hydrogen Bond," IT.H . Freeman and Co., San Francisco, Calif., 1960, p 38. (14) A. Blackburn and J. J. Kipling, Nature, 171, 174 (1953).
INTERMOLECULAR HYDROQEN BONDING
3315
--
o
..01
10
.1
.2
.I
.4
.5
rnLE FRICTION
-
.6
a 7
.8
.I
.a
,3
,I
.5
.6
.7
.8
.9
.03
-.04
1.0
J .9
1.0
TW
Figure 4. Molar refractions and their deviations from linearity for TMU-water mixtures a t 25'.
Figure 2. The differences between experimental and calculated densities for TMU-water mixtures at 26, 45, and 80".
The molar refractions and the differences between experimental and calculated molar refractions at 25" are shown in Figure 4.
R, R, (calcd)
c - 30
0
.1
.2
.I
.I
,5
.6
MOLE FRACTION
;I
.e
.9
1.0
- Tm
Figure 3. The differences between experimental and calculated surface tensions for TMU-water mixtures: 0, 25'; 0 , 45'; 80'.
0,
pure water, respectively. This plot more clearly shows the positive and negative deviations. The formation of hydrogen bonds leads to an increase in density and vice versa. Since both the formation and rupturing of hydrogen bonds is occurring in this system, the complex results are not unexpected. For the three temperatures, the differences between calculated and experimental values of surface tension fell along the same curve, as is shown in Figure 3. Ycalcd
= XTYT'
+
Xw')'wo
The values of surface tension observed for pure TMU at the three temperatures were 34.60, 31.64, and 28.30 dyn/cm. The surface tension of binary mixtures has been used to determine the form of hydrated species.15 The fairly sharp break in Figure 3 near a mole fraction of TAIU of 0.16 supports a hydrate with the formula TMU * 5Hz0.
+
n2 - 1 (XTMT XwMw) a n2 2
+
= _____
- 1 XwMw + nZw nZw+ 2 d,
n2T - 1 X T M T n2T 2 dT
+
= ___
___
where n, nT, and n, are the refractive indices of the solution, of pure TMU, and of pure water; MT and Mw are the appropriate molecular weights; and d, d ~ and , d, are the densities of the solutions, pure TMU, and pure water. The experimental molar refraction data lie on a very nearly linear curve. There are, however, systematic deviations as can be seen from the AR, plot. These deviations are small and go through an extreme near 0.5 mole fraction TMU. The heat of mixing data for the TMU-water system are shown in Figure 5 . The heat evolved per mole of TMU passes through a minimum of -13.7 kcal/mol at mole fraction TMU of 0.07. Since there can be many contributions to the heat of mixing, the location of this maximum is not necessarily indicative of the form of the complex. The fluidities (the reciprocals of the viscosities) a t the various temperatures of the solutions all pass through minima near mole fraction TMU of 0.15. Figure 6 shows the difference between experimental and calculated fluidities. The calculated values were computed using the KendalllBequation log
= XT log P T
+
x w
log P w
Large deviations from the Kendall equation are seen to occur a t all three temperatures. The measured values (15) P. Dunn and J. B. Polya, Red. Trav. Chim., 69, 1297 (1950). (16) J. Kendall and K. P. Monroe, J. Amer. Chem. Soc.,t39, 1787 (1917). The Journal of Physical Chemistry, Vol. 76, N o . 81, 1971
K. R. LINDFORS, S. H. OPPERMAN, M. E. GLOVER,AND J. D. SEESE
3316
Table I: Deviations in Physical Properties of TMU-Water and DMSO-Water Mixtures from Ideality Property
7-TMU-water--XT % dev
0.175 0.50 0.19
Y
Rm 926 d264
AHmix
VP
mu
FCTION
-
0.19 0.07 0.35
--DMSO-waterXDMSO % dev
61 0.17 810 19
0.51
0.23 0.30
0.7 203 5.6
11
deviations in the various properties for TMU-water mixtures. TWJ
dev
Figure 5. The heat of mixing per mole of TMU for TMU and water.
Bu
0
Figure 6. The differences in fluidities of TMU-water mixtures a t 25, 45, and 80" from those predicted by the Kendall equation.
of fluidity for pure TMU at the three temperatures were 73.7,101.6, and 161.3 P-l. The fluidity is a reliable and sensitive probe of association. The formation of complexes, either strong or weak, leads to large deviations from the Kendall e q ~ a t i o n . ' ~ ~Schottls '~ has argued that fluidity is the preferred method in that the deviations from ideality are much larger than for other techniques. I n Figure 6 the fluidity data indicate that the tri- to pentahydrate is the major complex formed between TMU and water. I n Table 1 are listed the COmpOSitiOnS of maximum deviations, values of the properties, and the percentage
The Journal of Physical Chemistry, Vol. 76, N o . 21, 1971
=
1OOlexptl - calcdl/exptl
For comparison, dimethyl sulfoxide-water mixture properties are also included. Dimethyl sulfoxide is known t o hydrogen bond with water to form a trihydrate. From the large systematic differences between the experimental and calculated values for the physical properties of TAN-water mixtures, the existence of strong interactions can be inferred. These are most likely hydrogen-bonding in nature. Infrared and nuclear magnetic resonance studies (to be published) also support the existence of hydrogen bonds in this system. To assign a formula to the TMU-water complex unambiguously on the basis of these data is impossible. I n fact, there may well be more than a single complex present. Since many types of interactions exist in these solutions, such as dipole-dipole, the maximum deviations of experiment and calculations should not all occur at the same composition. The most stable complex appears, however, to be between the tri- and pentahydrate.
Acknowledgments. The authors whish t o thank Dr. D. X. West and Dr. Paul D. Cratin for many helpful discussions, the Kumerical Analysis Laboratory of Central Michigan University for computer time, NIrs. A. Fallon for computer assistance, and the Faculty Research and Creative Endeavor Committee for partial financial support. We also thank The Ott Chemical Co. for the gift of the TJlU and for partial financial support of this project. (17) S. Glasstone, "Textbook of Physical Chemistry," 2nd ed, Van Nostrand, New Yo&, N. y . , 1951, 500. (18) H. Schott, J . Pharm. sci., 58, 946 (1969)
