254
M. J. COPLEY,EMANUEL GINSBERG, G. F. ZELLEIOEFERAND C. S. MARVEL
Vol. 63
5.3 X sec. under slightly different conditions. the same viscosity under the conditions of tem3. The use of the lead guard-mirror technique perature and pressure used, and under similar conhas established that free hydrogen atoms certainly ditions of illumination appear to produce approxicomprise less than 2% of the total number of mately equal concentrations of alkyl radicals. mirror-active particles produced in the photolysis The significance of this result in reference to the of propionaldehyde and are probably completely mechanism of the photolysis is discussed. absent. The result supports the view that the 5. The possibility is also discussed that there free formyl radical is stable up to temperatures may be competing free radical decompositions of 100O. in the photolysis of propionaldehyde. 4. Propionaldehyde and acetone vapors have NEWYORSC, N. Y. RECEIVED SEPTEMBER 23, 1940
[CONTRIBUTION FROM
THE
CHEMICAL LABORATORY OF THE UNIVERSITY OF ILLINOIS AND HEATING CORPORATION]
THE
WILLIAMS OIL-0-MATIC
Hydrogen Bonding and the Solubility of Alcohols and Amines in Organic Solvents. XIII BY M. J. COPLEY,’E. GINSBERG, G. F. ZELLHOEFERAND C. S. MARVEL A study2 of the solubilities of the volatile haloforms in donor solvents, free from hydroxyl or amide hydrogens, has shown that large negative deviations from Raoult’s law are consistently observed. This behavior has been explained by the assumption that on mixing, intermolecular association of unlike molecules occurs through the mechanism of C - H c N or 0 bonds. Recent infrared absorption studies by Buswell, Roy, and Rodebushs and by Gordy‘ have confirmed this picture. It is also desirable to make studies of the solubilities of volatile alcohols and amines in a variety of types of organic liquids containing donor atoms. Such studies would furnish further information on the influence of solute and solvent association on solubility. A comparison of the solubilities of an alcohol in a series of solvents containing different functional groups should give the relative strengths of the donor atoms. It is of interest to determine whether this relative order is the same as was observed for the haloforms in the same type of solvents. Gordy6 has determined the shift of the OD fundamental frequency when CHsOD is mixed with a series of (1) Present address, Eastern Regional Research Laboratory, U. S. Department of Agriculture, Chestnut Hill Station, Philadelphia, Pennsylvania. (2) (a) Zellhoefer, Copley, and Marvel, THIS JOURNAL, 60, 1337 (1938); (b) Zellhoefer and Copley, ibid., 60, 1343 (1938); (c) Copley, Zellhoefer, and Marvel, ibid., 60, 2666 (1938); (d) Copley, Zellhoefer, and Marvel, ibid., 61, 8550 (1939). (3) Buswell,Roy, and Rodebush, ibid., 60, 2528 (1938). (4) (a) Gordy, ibid., 60, 605 (1938); (h) J , Cham. Phys., 7 , 163 (1939). ( 5 ) (a) Gordy, ibid., 7, 93 (1989); (b) Gordy and Stanford, ibid., 8, 170 (1940).
different donor solvents. A relation should exist between his data and the relative solubilities of alcohols in similar solvents. Experimental The method used in making the solubility measurements is the same as that described in a previous paper8 except that the lower vapor pressures involved made it advisable to substitute an oil or mercury manometer for the pressure gage. The materials used were all purified carefully by chemical means and fractional distillation where feasible. Solubility measurements were made over a range of pressures at a temperature of 32.2’. For comparative purposes, the solubilities at a partial pressure corresponding to the vapor pressure of the solute a t 4.5’are used in Tables I1 and I11 and in the discussion. The vapor pressures of a number of the solutes at temperatures other than their boiling points which had not been previously reported in the literature were measured experimentally. Vapor pressures were determined at a number of different temperatures and a plot of logarithm of the pressure against l/T’K. was made. The vapor pressures tabulated in Table I were calculated from these plots. TABLEI VAPORPRESSURES OF SOLUTES V. p . at 4 S 0 ,
V. p. at 32.2’.
5.5 223
31 743 397 316 287.5 199 105
mm.
s-Butyl alcohol i-Prop ylamine n-Propylamine Diethylamine s-Butylamine i-Butylamine n-Butylamine
106
88 56.5 45.4 24
mm.
In Tables I1 and I11 the “ideal” or theoretical mole fraction solubility was Calculated using Raoult’s law, and is the ratio of the vapor pressure of the solute at 4.5’ to its (6) G. F. Zellhoefer, I r d . E r g . Cham., 39, 584 (1937).
Jan., 1941
HYDROGEN BONDINGAND SOLUBILITY OF ALCOHOLSAND AMINES
value at 32.2’. The ratio of observed to “ideal” mole fraction solubilities given in the last column of Tables I1 and I11 is used for comparison rather than absolute value of solubility, since the “ideal” mole fraction solubilities are different for the different solutes. The vapor pressures of the solvents are for the most part negligible, but where it was appreciable an approximate correction based on Henry’s law was applied.
HzO
Glycerol 0.177
,239
1.35
HzO
Carbitol ,177
.126
0.71
Dimethyl ether of tetraethylene glycol HzO .177 .099 .56
TABLE I1 Hz0 SOLUBILITY OF WATERAND ALCOHOLS INORGANICSOLVENTS CHaOH AT 4.5’ Ratio GHiOH Mole fraction “Ideal’ Observed
Observed/ “Ideal”
Triethylenetetramine 0.588 3.33 Hz0 0.177 CHsOH .223 .605 2.71 CzHsOH ,200 .508 2.54 i-CsHiOH .164 .357 2.17 n-CsH7OH .148 .362 2.45 t-C4HpOH .183 .376 2.05 s-C~H~OH .178 .360 2.02 n-CdHoOH ,126 ,345 2.74 Tetraethylene pentamine HzO .177 .622 3.51 Methylated triethylene tetramine I” HzO .177 .398 2.24 CHsOH .223 .588 2.63 GHsOH .200 .469 2.35 Methylated triethylene tetramine IIh CHaOH ,223 .523 2.35 CZH~OH ,200 .416 2.08 Triacetyl trimethyl triethylene tetramine@ .177 .610 3.45 He0 CHsOH .223 .520 2.33 Hexamethylenediamine 2.08 .177 .382 Ha0 CHsOH 2.59 .223 .578 2.02 CzHsOH * 220 .445 i-CsH;OH 2.15 .164 ,362 n-CaHjOH 2.50 .148 .370 t-C4HpOH 1.94 .183 .364 s-C~H~OH 2.15 .178 .385 n-CdHoOH 2.55 .126 .321 N,N-Dimethylacetamide .177 .232 1.31 HnO CHsOH ,223 .369 1.65 .200 .314 1.57 CzH60H N-Methylacetamide .223 .297 CHsOH 1.33 Adiponitrile .223 .lo9 C-OH 0.49 Sebaconitrile 0.51 .223 .114 CHaOH Ethylene glycol 1.17 .177 .207 HD CHsOH 0.84 .223 .188 .200 .llO 0.55 4HsOH
255
Triethyl phosphate ,177 .lo4 .223 .390 .200 .363
.59 1.75 1.81
n-Tributyl borate 0.83 CHsOH .223 .186 a Prepared by methylation of triethylene tetramine with dimethyl sulfate. B. p. 97-107’ (2-3 mm.). Anal. Calcd. for C&IN~ (trimethyl): C, 57.5; H, 12.8; N, 29.8. Found: C, 57.9; H, 12.6; N, 29.6. Prepared by methylation of triethylene tetramine with formaldehyde and formic acid, b. p. 99-103’ (2 mm.). Anal. Calcd. for ClOHz6N4(tetramethyl) : N, 27.7. Calcd. for CltHasNl (pentamethyl) : N, 25.9. Found: N, 26.3. Hence, the compound is a mixture containing an average of 4.7 methyl groups. Prepared by acetylating methylated triethylenetetramine I with an excess acetic anhydride. The product is a viscous brown liquid with a green fluorescence, b. p. 200-225’ (2-4 mm.). Anal. Calcd. for ClsH?oOsNc (trimethyltriacetyltriethylenetetramine): N, 17.8. Found: N, 17.6.
TABLE I11 SOLUBILITY OF A ~ E IN S ORGANICSOLVENTS AT 4.5O Mole fraction “Ideal” Observed
Water” 0.355
Ratio OSjserved/ Ideal”
0.457
1.29
n-Octyl alcohol .267 .408
1.53
Ethylene glycol .355 .662 .300 ,488 ,267 .465 .279 ,371 .227 ,381 .229 .400 .238 .397
1.86 1.63 1.74 1.33 1.68 1.75 1.67
1,3-Butylene glycol .267 .460
1.72
Glycerol .355 .344 .267
.e52 .572 .497
1.84 1.66 1.86
Diethylene glycol .355 .653 .300 .517 .267 ,510 .229 .384
1.84 1.73 1.91 1.68
256
M. J. COPLEY, EMANUEL GINSBERG, G. F. ZELLHOEFERAND C. S. MARVEL TABLE I11 (Concluded)
Vol. 63
On the other hand, when three of the methyl groups of the completely methylated triethylenetetramine are replaced by highly polar acetyl Triethylene glycol groups, the ratio (3.45) for water is greatly eni-CsH,NHz 0.300 .552 1.84 hanced. n-CsH,NHs .267 .519 1.94 The high solubilities observed in the tertiary wC~HC”~ .229 .405 1.77 amides2c indicates that an amide nitrogen is as Tetraethylene glycol strong a donor as an amide nitrogen. Quantita~Z-C~HQNHZ .229 .410 1.79 tively this is best illustrated by comparing the Hexamethylenediamine ratios obtained using N,N-dimethylacetamide n-CaHpNHz ,229 .144 0.63 with that obtained for the reciprocal case, the Triethylenetetramine solubility of n-propylamine in n-octyl alcohol, n-&4HgNH2 .229 .145 0.63 shown in Table 111. However, all nitrogen atoms a Solubility values interpolated from curves given by are not necessarily good donors as shown by the Mehl, Z.ges. Kulte Ind.,42, 13 (1935). fact that less than theoretical solubilities were obDiscussion served for methyl alcohol in adiponitrile and In Table I1 are given the solubilities of a num- sebaconitrile. These results are similar to those’ ber of alcohols in a few different polyethylene obtained previously with haloform in nitriles. The solubilities of water, methyl alcohol, and amines and their derivatives. For each alcohol ethyl alcohol were measured in some non-volatile the observed solubility was greatly in excess of the polyhydric alcohols and esters. Some excess “ideal” solubility calculated by use of Raoult’s solubility for water in ethylene glycol and glycerol law. The pure water and alcohols are strongly was observed whereas less than normal solubilities associated through 0 - H t O bonds. Some aswere observed for the two alcohols. The unexsociation may exist in the pure amines but prepectedly high solubilities of methyl and ethyl vious work has shown that while the nitrogen atom alcohols in triethyl phosphate may be connected of an amine is a good donor, the hydrogens are not with the presence of the semipolar oxygen. sufficiently activated to act readily as acceptors. In Table I11 the solubilities of a number of Reasoning from analogy with the behavior of volatile amines are shown for several different solhaloforms,2c the simplest explanation of the exvents. These combinations are reciprocals of the treme solubilities is obtained by assuming that ones shown in Table 1 1 . Hence, as one would ex0-H+N bonds are much stronger than 0 - H c O bonds. When equilibrium is reached after mix- pect, extremely high solubilities were observed for ing, there are relatively few intermolecular com- the amines in non-volatile polyhydric alcohols plexes between like molecules but instead many whereas they exhibited low solubilities in polybetween unlike molecules. Gordy6bobserved that ethyleneamines. the shift of the fundamental frequency of OD was Summary a maximum in mixtures of CHIOD and aliphatic The solubilities of a number of volatile alcohols amines. The solubilities of a series of eight alcohols were and amines are given for a variety of solvents inmeasured in triethylenetetramine. The extreme cluding non-volatile polyhydric alcohols, polyvalue (3.33) of the ratio of observed to ideal solu- ethylene glycols, polyethylene amines, amides, bility obtained with water may result from partial esters, and ethers. Extremely high solubilities were observed for ionization of the complex formed. It is noticeable that larger values of the ratio are obtained alcohols in polyethylene amines and tertiary with the normal or straight chain alcohols than amides and for the amines in polyhydric alcohols with secondary or tertiary alcohols where one and polyethylene glycols. might predict that steric hindrance may play a A plausible explanation of these results is obtained by applying to the data the hydrogen bond role due to branching of the chain. Progressive methylation of triethylene tetra- picture of solubility. mine leads to lower solubilities of water and alco- URBANAAND BLOOMINGTON, ILLINOIS RECEIVED OCTOBER 26, 1940 hols. However, the solubility of water decreases more on methylation than that of the alcohols. (7) Copley,Zellhoder and Marvel, Tars JOURNAL, (IS 227 (1940). Mole fraction “Ideal” Observed
Ratio Otserved/ Ideal”
